ABSTRACTS

3rd Annual MGH Research Fellows’

Poster Celebration

Tuesday, June 24, 2008 4:00-6:00pm, Bulfinch Tent

Posters ~ Awards ~ Food & Drinks

MGH Center for Faculty Development Bulfinch 370 | 617.643.1606 | www.massgeneral.org/facultydevelopment/orcd

TABLE OF CONTENTS Agenda ………………………………………………………………………3 Reviewers ………………………………………………………………………4 Mass. General Postdoctoral Association ………………………………………5 Abstracts ………………………………………………………………………6 Author Index ………………………………………………………………………62 Notes ………………………………………………………………………64

2

AGENDA 10:00 am Posters on display 4:00 pm Event opens 4:15 pm Authors present at odd-numbered posters 4:45 pm Authors present at even-numbered posters 5:30 pm Awards presentation Welcoming remarks: Tayyaba Hasan, PhD Mass General Postdoc Association Announcement of winners: David MacLaughlin, PhD; Hensin Tsao, MD, PhD 5:45 pm Refreshments and networking Posters may remain on display until 6:30 pm

3

REVIEW COMMITTEE We are most grateful to the members of the Review Committee who spent a great deal of time reading abstracts and reviewing posters to determine this year’s winners. The review process this year occurred in two steps based on feedback from fellows who attended last year’s poster celebration. First, each abstract was carefully read and ranked by a sub-group of the committee. Based on these rankings and a discussion that roughly followed the format of NIH study sections, we chose 30 abstracts as finalists. Secondly, the committee reviewed the posters of the finalists for clarity of message and research significance to determine the final winners.

David MacLaughlin, PhD, Co-chair

Hensin Tsao, MD, PhD, Co-chair

Murat Bastepe, MD, PhD

Oksana Berezovska, PhD

Wei Chao, MD, PhD

Giulia Fulci, PhD

Christene Huang, PhD

James Kobler, PhD

Madhu Malo, MD, PhD

Giudo Musch, MD

Catherine Nutt, PhD

Hyle Park, PhD

John R. Sims, MD

David Sweetser, MD

Xun Zhang, PhD

4

MGH POSTDOCTORAL ASSOCIATION The Office for Research Career Development thanks the MGH Postdoctoral Association for its help in planning the 3rd Annual Poster Celebration, and we especially thank the members of the MGPA Board who volunteered their time to make this a successful event. We encourage all MGH Research Fellows to learn more about the MGPA and join the organization by visiting their web site at: https://www.massgeneral.org/mgpa/index.asp. 2008 MGPA Board Members Adnan Abu-Yousif, PhD Albena Kantardzhieva, PhD Joseph Todd Blois, PhD Sarah Bowman, PhD Jonas Dyhrfjeld-Johnsen, PhD Eric B. Finkelstein, PhD Pawel Mroz, MD Yuichi Niikura, PhD Rosanna M. Rahman, PhD Nadege Roche, PhD Ayguen Sahin, PhD Joao C. Seco, PhD Kaisa L. Selesniemi, PhD (Administrator) Olga Syrkina, MD (Co-Chair) Julie E. Tetzlaff, PhD (Co-Chair) Sarika Verma, PhD (past Co-Chair) Guoqi Zhang, PhD

5

Poster Board Number: 1 Optimization of combinatorial therapy using EGFR inhibition and photodynamic therapy in ovarian cancer models. Adnan O. Abu-Yousif, Marcela G. del Carmen, Imran Rizvi, Anne C. Moor, Thomas Stepinac, and Tayyaba Hasan Wellman Center for Photomedicine, Massachusetts General Hospital, Harvard Medical School. Boston, Massachusetts, USA Purpose: The high mortality associated with epithelial ovarian cancers is largely attributed to the development of resistance to standard treatments. Overall response rates to systemic chemotherapy in platinum resistant disease are dismal, ranging from 12-25%, thus creating a pressing need for the development of target-based combinatorial therapies that are effective in resistant disease. Photodynamic therapy (PDT) is an emerging photochemistry-based molecular and biophysical modality for the treatment of selected cancers that has proven effective at treating cancers that have developed resistance to chemo- and radiotherapies. Our group has shown the combination of benzoporphyrin derivative monoacid A (BPD)-based PDT and C225 (erbitux), a monoclonal antibody that inhibits the receptor tyrosine kinase activity of EGFR proved to be well tolerated, effective, and synergistic in mice with a 33% cure rate in disseminated murine model of ovarian cancer (1). In the present study, two combinatorial therapeutic regimens were designed to investigate the synergistic relationship between BPD and C225 treatments. Materials & Methods: The strategy used in the current study involved development of a BPD-C225 photoimmunoconjugate (PIC) that allows for simultaneous delivery of both agents. In order to optimize the two-step combinatorial therapeutic regimen used in previous studies we investigated the effects of BPD-PDT on localization and activity of EGFR and downstream effector molecules. Both approaches were investigated in the ovarian cancer cell line, OVCAR-5, and will be explored in 3D cultures grown on reduced growth factor matrigel based on models developed by Joan S. Brugge and Results: Confocal laser microscopy analysis of EGFR-expressing cells revealed that free BPD was found predominantly in the mitochondria compared with the lysosomal localization of BPD-C225. Furthermore, the PIC retained the Mab's target recognition and effectively blocked EGF-induced phosphorylation of the EGFR in OVCAR-5 cells. Incubation with BPD-C225 also inhibited the phosphorylation of Akt-1 and MAPK/ERK, two downstream molecules involved in growth arrest, chemosensitivity, and angiogenesis. Next, we investigated the effects of BPD on EGFR signaling.BPD-based PDT induced a biphasic response of EGFR. Immediately after PDT EGFR is internalized, however the receptor is recycled to the cell surface 24-48 hours post PDT treatment. Interestingly, once recycled to the cell surface the EGFR was much more sensitive to EGF stimulation, suggesting that timing of administration of C225 may be important to the synergistic relationship between BPD-based PDT and C225. Future studies aim to evaluate the efficacy of the BPD-C225 PIC in 3D ovarian cancer models and in vivo models, as well optimizing the timing of administration of C225 in the two-step combination protocol. Conclusion: Our results suggest that photoimmunotargeting could be a useful dual strategy for selective destruction of tumor cells, given that this PIC directs the photosensitizer specifically to the target cells and also exerts the receptor blocking function of the Mab. 1. del Carmen MG, Rizvi I, Chang Y, et al. Synergism of epidermal growth factor receptor-targeted immunotherapy with photodynamic treatment of ovarian cancer in vivo. Journal of the National Cancer Institute 2005;97(20):1516-24. 2. Duska LR, Hamblin MR, Miller JL, Hasan T. Combination photoimmunotherapy and cisplatin: effects on human ovarian cancer ex vivo. Journal of the National Cancer Institute 1999;91(18):1557-63. Author Contact: Adnan O. Abu-Yousif, Wellman Center for Photomedicine, 617-726-1297, aabu-yousif@partners.org

6

Poster Board Numer: 2 Vaccination with autologous apoptotic macrophages diminishes Leishmania major infection by involvement of Foxp3-positive regulatory T-cells Oleg E. Akilov, Yongzhu Jin, Zhuoyan Zhou, Mei X. Wu, Tayyaba Hasan Wellman Center for Photomedicine, Massachusetts General Hospital, Boston, MA, USA Purpose: Parasites of the genus Leishmania cause substantial mortality and morbidity in the developing world, with the infection at endemic levels in 88 countries. While numerous treatment modalities exist, there is no ideal therapy for leishmaniasis. The purpose of or study was elucidate the mechanisms of regulation of the intracellular parasite clearance in macrophages (Mф) during cutaneous leishmaniasis (CL). Materials & Methods: BALB/c mice were vaccinated subcutaneously three times with 2-week intervals in between with cell lysate obtained from non-infected bone marrow-derived Mф after photodynamic therapy with photosensitizer EtNBSe (PDT). The PDT regimen was chosen in this way that death of Mф occurred through predominantly apoptotic pathway. The control group of mice received PBS injections. Two weeks after the last vaccination mice were infected intradermally in the ear with 106 metacyclic promastigotes of L. major (MHOM/IL/80/Friedlin). Results: Mice of the vaccinated group showed a delayed onset of disease and smaller lesions that contained ~ 3.1-logs less parasites than did the controls (1.1 x 103 parasites/ear vs. 3.0 x 106 in control group, p < 0.05). Mice that received a vaccine displayed lower levels of cytokines TNF-α and IL-6 (cytometric beads array), and lower levels of chemokines CCL2/MCP-1 and CCL22/MDC (ELISA) in the ear tissues than did the controls. A percentage of IFN-γ-producing T cells in the draining lymph nodes of the vaccinated mice was higher than in the control group. We found that a Mф death in CL lesions was a result of the attack of the immune system (caspase-3 activation on IHC) than cytopathic effect of the intracellular parasitism. Flow cytometric analysis revealed higher number of F4/80 (Mф) and CD4+CD25+Foxp3+ T regulatory (T reg) cells in the lesions of the vaccinated mice than did the controls. Correlation analysis showed that the percentage of Mф in the skin directly correlates (r = 0.93) with number of T reg. Depletion of CD25+ cells (three subcutaneous injections of anti-CD25 antibody derived from PC61 5.3 hybridoma within the two-week interval) lead to an increase of the number of parasites and the number of CD8 T cells. An increase of CD3+CD8+ cells in the leishmanial granuloma was accompanied by an increase of IL-10 production. Number of apoptotic Mф did not increase dramatically in the group of αCD25 mice which means that CD8 T cells did not affect Mф directly. Conclusion: These results demonstrate that T reg may play a role in a reduction of the parasite burden through inhibition of CD8 T cell-attack on the infected Mф. Author Contact: Oleg E. Akilov, MD, PhD, Wellman Centre for Photomedicine, 617-724-7131, oakilov@partners.org

7

Poster Board Number: 3 Internalization Of Epidermal Growth Factor Receptor (EGFR) In Response To Photochemically Induced Oxidative Stress In An In Vitro Ovarian Cancer (OVCAR-5) Model Humra Athar, Thomas Stepinac and Tayyaba Hasan Department of Dermatology, Wellman Center for Photomedicine, Massachusetts General Hospital / Harvard Medical School, Boston, MA 02114

Purpose: The Epidermal Growth Factor Receptor (EGFR) is frequently over expressed in many types of cancer and serves as a target for cancer therapy. The efficacy of single modality anti-EGFR therapy based on Tyrosine kinase inhibitors is limited by the fact that some patients harboring EGFR mutation subtypes have developed resistance to TKI (e.g. Iressa). It is increasingly getting evident that combination therapies would be required for more effective management of cancer treatment. In this study, we have used an ovarian cancer cell line (OVCAR-5) to investigate the possible mechanism of synergism between EGFR blockade by Cetuximab, (an EGFR blocking antibody) and a cytotoxic treatment of cancer like PDT.

Materials & Methods: Photodynamic treatment: Cells (~ 0.6x 106) were plated in triplicates in a 60 mm dish (~ 50% confluent) and were allowed to attach overnight. After ~ 24 hr, the cell medium was changed and the cells were incubated with 140 nM BPD (QLT, Vancouver, BC, Canada) (0.1 μg / ml) for 90 min at 370C / 5% CO2. All the work involving BPD was performed in subdued light. After 90 min. of incubation, the BPD laden media was replaced with the fresh media and the cells were irradiated at 690 nm at a fluence rate of 10 mW/cm2 using a laser source (model # 7401, High Power Devices, North Brunswick, NJ, USA) following our established laboratory protocol. A light dose of 100 mJ/cm2 induced about 35-40% cell killing. After PDT, the cells were either lysed for Western analysis or processed for Immunofluorescence. The conditioned media of the cells was used for VEGF ELISA studies.

Results: In the current in-vitro study, we report that sublethal dose of PDT with the photosensitizer, Verteporfin induces a biphasic response of EGFR (a) EGFR gets internalized in the small vesicles during the first few hours following PDT. (b) However, twenty-four to forty-eight hours post PDT, the EGFR gets recycled back to the cell surface and gets much more sensitive to EGF stimulation than before the treatment. (c) Sublethal PDT treatment increased VEGF production by a factor of four. Cetuximab treatment alone had little effect on VEGF secretion in non-PDT treated cells. However, we observed a subtle decrease (~ 20%) in VEGF secretion after Cetuximab treatment in cells treated with PDT. Conclusion: The results presented herein, provide us with a better understanding of the implications in the administration of combination treatment of PDT with EGFR blockade. These results are validated in our laboratory in an orthotopic in vivo mouse model of ovarian cancer, which demonstrated 33% survival in a combination modality of receptor tyrosine kinase inhibition with Cetuximab and BPD-PDT. Author Contact: Humra Athar, Dermatology, Wellman Center for Photomedicine, 617-726-6169, hathar@partners.org

8

Poster Board Number: 4 Selective Loss of MEG3 Expression and IG-DMR Hypermethylation in the MEG3/DLK1 Locus in Human Clinically Non-functioning Pituitary Adenomas of a Gonadotroph Origin Dalia L Batista, Roger Gejman, Ying Zhong, Ali Mahta, Wendy Chen, Yunli Zhou, Xun Zhang, Brooke Swearingen, E. Tessa Hedley-Whyte and Anne Klibanski Massachusetts General Hospital and Harvard Medical School. Boston, MA Purpose: MEG3 is a human maternally expressed imprinted gene that maps to chromosome 14q32 and its gene products function as non-coding RNAs. We previously reported that MEG3 is highly expressed in normal human pituitary gonadotrophs but not expressed in clinically non-functioning pituitary tumors. Whether MEG3 is expressed in other cell types of the normal pituitary gland, or whether loss of MEG3 expression occurs in other human pituitary tumor phenotypes is unknown. Materials & Methods: We examined the expression of MEG3 in normal human pituitary cells and investigated if the loss of expression of MEG3 was associated with the methylation status of the intergenic differentially methylated region (IG-DMR), which controls imprinting of the MEG3 gene. Thirty-nine human pituitary adenomas (17 clinically non-functioning, 11 GH-producing, 7 PRL-producing, and 4 ACTH-producing) were evaluated. Expression of MEG3 was determined by a combination of mRNA in situ hybridization, immunofluorescence staining, RT-PCR and quantitative RT-PCR. Bisulfite sequencing was performed to analyze the DNA methylation status of the IG-DMR region. Results: Our study shows that all normal human pituitary cell types express MEG3. Importantly, loss of MEG3 expression occurs only in non-functioning pituitary adenomas of a gonadotroph origin. In contrast, all functioning pituitary tumor phenotypes examined express MEG3. Bisulfite sequencing shows hypermethylation of the IG-DMR region in non-functioning pituitary adenomas. We conclude that MEG3 expression is lost exclusively in non-functioning pituitary tumors of gonadotroph origin and that hypermethylation of the IG-DMR region, controlling the imprinting of the MEG3 gene, is associated with lack of MEG3 expression in such tumors. Conclusion: Therefore, our data demonstrate that MEG3 is the first identified gene whose expression is selectively lost in human non-functioning pituitary tumors of a gonadotroph origin and suggest that the loss of MEG3 contributes to their tumorigenesis. Author Contact: Dalia L Batista; Neuroendocrine Unit; 617-724-0087; dlbatista@partners.org

9

Poster Board Number: 5 The Mucus Shaver, a new device to keep the endotracheal tube free from secretions: a clinical investigation L. Berra, MD1; L. Bigatello, MD1; T. Kolobow, MD2; A. Pesenti, MD3. 1Anesthesia and Critical Care, MGH, Boston; 2NIH, NHLBI, Bethesda, 3Anesthesia and Intensive Care, University of Milan-Bicocca, Italy

Purpose: ICU-patient prolonged ventilated have high risk of endotracheal tube (ETT) occlusion from secretion accumulation. Also, ETTs become easily colonized and may spread bacteria to the lungs causing hospital-acquired pneumonia. To date there are no effective means to clean ETTs. Materials & Methods:

We fabricated and evaluated a new medical device (“Mucus Shaver”=MS) designed to clean the endotracheal tube (ETT). It was studied in a prospective, randomized trial of patients ventilated for more than 72h (control group: n=12patient, MSgroup: n=12 patients). The MS is a concentric, inflatable catheter for the removal of mucus and secretions from the interior surface of the ETT. The MS is designed to be advanced to the distal ETT tip, the distal balloon inflated, and the device subsequently withdrawn. Each ETT was cleaned every 2hours.

Schematic representation

Results: No adverse event related to the use of the MS was observed. At extubation, 1 ETT from the MS group was colonized, while in the control group 10 ETTs were colonized (8% vs. 83%; p<0.001). Scanning Electron Microscopy did not show any bacterial biofilm in ETTs from the study group, while thick bacterial deposits were present on all the ETTs from the control group (p<0.001). The nursing staff was satisfied by the feasibility, comfort and efficacy of the MS.

of the MS.

The MS inflated.

MS inflated after being introduced into an

ETT.

Conclusion: 1) The MS seems to be a safe, comfortable, feasible and efficient device for ETT cleaning. 2) The MS may be helpful in preventing ETT colonization by multi-drug-resistant-ICU microorganisms. Author Contact: Lorenzo Berra, Anesthesia and Critical Care, lberra@partners.org

(A) ETT appearance at extubation in a patient from the MS group. Note the

absence of any mucus or secretions deposit on it.

(B) ETT from another patient from the MS group: no secretions were detected

even at the tip of the ETT.

(C) Electron micrograph of the lumen of an ETT where MS was used. Few small

drops (few microns-100microns) are noted on the lumen of the ETT.

(D) Electron micrograph of ETT lumen of a patient from control group. Note the

tick, continuous deposit of secretions on the lumen of the endotracheal tube.

10

Poster Board Number: 6 Silver Sulfadiazine Coated Endotracheal Tubes: From Bench to Bedside L. Berra, MD1; L. Bigatello, MD1; T. Kolobow, MD2; A. Pesenti, MD3. 1Anesthesia and Critical Care, MGH, Boston; 2NIH, NHLBI, Bethesda, 3Anesthesia and Intensive Care, University of Milan-Bicocca, Italy

Purpose: After only few hours of mechanical ventilation, bacteria might form biofilm on endotracheal tubes (ETTs), that might enter the lungs causing pneumonia. We coated ETTs with silver-sulfadiazine (SSD-ETT) to prevent bacterial colonization, and tested those ETTs in an in-vitro study, in an animal model, and in a clinical study. Materials & Methods: First, 6 SSD-ETTs and 6 standard ETTs (St-ETT) were challenged in-vitro every 24h with 104–106 Pseudomonas aeruginosa/mL. ETTs were evaluated at 6, 24, and 72h. Secondly, 8 sheep were randomized to receive either a SSD-ETT or a St-ETT, and ventilated for 24h. Finally, a prospective, randomized phase I-II trial was performed in 46 adults (SSD-ETT versus St-ETT) expected to need 12-24h of intubation, in a Italian university ICU. Results: In-vitro studies: St-ETTs showed heavy gorwth of Pseudomonas aeruginosa throughout the study, bacteria adhered to the ETT wall after only 6h. SSD-ETT remained bacteria-free for at least 72h (p<0.01 vs. St-ETT), no bacteria were detected on the ETT wall at any time. In sheep mechanically ventilated, SSD-ETT prevented bacterial colonization of the lower respiratory tract (p<0.01 vs. St-ETT). In the clinical study, colonization of the lower respiratory tract and of the lumen of the ETT was prevented by the use of SSD-ETT (p<0.01 vs. St-ETT). Mucus accumulation in the SSD-ETT appeared to be reduced compared to the St-ETT.

a b

c d

a, b, c, d: Micrographs of the lumen of the ETT at extubationstained with BacLight Live-Dead and imaged with confocalmicroscopy. a: St-ETT sample, cross-section is 746 µm thick at its maximum. b: SSD-ETT sample, cross-section is 64 µm thick. c: Lumen of St-ETT with scanning electron microscopy showing diplococci, macrophages, and epithelial cells on the amorphous deposits. d: Lumen of a SSD-ETT, note absence of any deposit. Only a thin layer of silver-sulfadiazine can be seen

Conclusion: SSD-ETT could be safely used in preventing bacterial colonization and narrowing of the ETT in patients intubated for a brief period of 24h of mechanical ventilation. Author Contact: Lorenzo Berra, Anesthesia and Critical Care, lberra@partners.org

11

Poster Board Number: 7 Thalamic injection of AAV-β-galactosidase results in global correction of storage in the adult GM1-gangliosidosis brain Marike L. D. Broekman1,2,3, Jill C. Narak5, Nancy R. Cox5, Laryssa A.Tierney1, Carlos E Sanchez4, Bob S Carter4, Douglas R. Martin5, Miguel Sena-Esteves1*. 1. Molecular Neurogenetics Unit, Department of Neurology, Massachusetts General Hospital, 149 13th Street, Charlestown, MA 02129, USA 2. Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, The Netherlands 3. Department of Neurosurgery, University Medical Center Utrecht, The Netherlands 4. Department of Neurosurgery, Massachusetts General Hospital, Boston, MA 02114, USA 5. Auburn University College of Veterinary Medicine and the Scott-Ritchey Research Center, Auburn, AL, USA Purpose: Lysosomal storage diseases (LSD) are characterized by the accumulation of undigested macromolecules in the lysosomal compartment. Up until now, over 40 LSDs have been described and affect 1/7,000 live-born infants. Almost two thirds of these heritable human disorders exhibit some degree of neurological impairment, due to global storage in the central nervous system (CNS). Correction of storage in the CNS remains a challenge as most treatment options are hindered by the blood-brain barrier or have the disadvantages associated with life-long repeated administrations.

Direct infusion of adeno-associated virus (AAV) vectors encoding lysosomal enzymes into the brain parenchyma has emerged as a viable strategy to create an in situ source of normal enzyme in the brain. The rationale for this approach is the release of lysosomal enzymes from genetically modified cells to the extracellular space and subsequent uptake by untreated cells by receptor-mediated endocytosis. This process of enzyme transfer between normal or genetically modified cells and enzyme-deficient cells is known as cross-correction. Multiple studies in a number of animal models of lysosomal storage diseases have shown that genetic modification of a relatively small number of cells in the brain is sufficient to deliver corrective levels of enzyme to large regions of the brain.

Recently, we have demonstrated that in the neonatal GM1-gangliosidosis mouse brain storage could be corrected by AAV mediated delivery of the enzyme that is deficient in these mice, the lysosomal acid beta galactosidase (mβgal). This enzymatic deficiency leads to accumulation of GM1-ganglioside and its asialo derivative GA1 in the central nervous system (CNS) and progressive neurodegeneration. The most severe form of this disease (Infantile or type I) has a very early onset, and it is characterized by rapid neurological decline with death occurring usually before 2 years of age. Materials & Methods: Since most cases of GM1-gangliosidosis in humans are diagnosed after birth, the aim of this study was to develop a strategy to achieve global distribution of βgal in the adult brain with a minimal amount of injections. Lysosomal enzymes are distributed in the CNS from vector-transduced cells by diffusion, axonal transport within neurons, and by CSF flow. We reasoned that we could use these properties to achieve widespread distribution of the enzyme by targeting a region in the brain with many connections. Since the thalamus is such an area of the brain, we injected an AAV vector encoding for mouse βgal into this structure of adult GM1-gangliosidosis mice. Results: One month post-injection, we observed exceptional distribution of βgal and global correction of storage in the brain. To investigate whether this distribution was due to the size of the mouse brain rather than to the connectivity of the target structure, we stereotactically infused an AAV vector encoding for feline βgal into the thalamus of cats. One month after injection, widespread distribution of the enzyme was observed. Conclusion: These results confirm the notion that the thalamus is an appealing target for genetic modification and can be viewed as the central node in a ‘built-in’ network for widespread distribution of lysosomal enzymes throughout the CNS. Author Contact: Marike Broekman, Neurogenetics, 617-724-8917, Broekman.Marieke@mgh.harvard.edu

12

Poster Board Number: 8 An oncolytic herpes simplex virus armed with a xenogeneic homologue of prostatic acid phosphatase enhances anti-tumor efficacy in prostate cancer Pedro Castelo-Branco, Jason S. Buhrman, Cécile Zaupa, Brent J. Passer, Samuel Rabkin, and Robert L. Martuza Department of Neurosurgery, Massachusetts General Hospital 1These authors contributed equally to this work Purpose: Oncolytic herpes simplex viruses (oHSV) harbour deletions/mutations in genes that cause them to selectively replicate in proliferating tumor tissue and that safely attenuate the virus for therapeutic use. HSV-BAC technology allows for rapid integration of transgenes into a viral genome by manipulating the viral DNA in E. coli. We hypothesize that arming oHSVwith human prostatic acid phosphatase (hPAP), a xenogenic antigen whose homologue is expressed only in prostate and prostate tumor cells, will improve the efficacy of viral treatment in prostate cancer by combining the lytic effect of an oncolytic virus with tolerance-breaking autoimmunity directed to prostate. By employing the immune system to aid in treating the tumor by priming it to recognize endogenous proteins as foreign antigens, effective treatment of the tumor may be possible with fewer administrations of the oncolytic virus. Materials & Methods: To manipulate HSV in E. coli, a Bacterial Artificial Chromosome (BAC) sequence was inserted into the HSV genome. This BAC contains an E. coli origin of replication as well as LoxP and FLP site-specific recombination sequences. A shuttle vector containing a LoxP site was integrated into the BAC sequence with Cre-recombinase. The bacterial shuttle vector sequences were removed byFLP site-specific recombination after HSV-BAC DNA was transfected into mammalian cells, and virus (bPat∆6-hPAP) was isolated. Animal studies were conducted in C57/BL6 mice injected subcutaneously with the TRAMP C2 mouse prostate cancer cell line. Oncolytic virus was administered at differing concentrations once the tumors reached 100mm3 in size. Results: Western blot analysis revealed that bPat∆6-hPAP infected TRAMP C2 cells produce hPAP. RT-PCR reveals that TRAMP C2 cells express endogenous PAP. After intratumoral viral administration, local hPAP was detected by ELISA in the tumor, indicating that the virus is infecting and expressing the transgene in the tumor. A larger decrease in subcutaneous tumor growth was observed after bPat∆6-hPAP treatment than with bPat∆6-empty, a virus that does not express hPAP, suggestive of an increased immune response against the implanted subcutaneous tumor. Conclusion: These findings suggest that arming a virus with a xenogenic transgene may be a useful strategy to treat prostate cancer. Further studies are necessary to understand the immune mechanisms involved in this tumor regression. Author Contact: Jason S Buhrman, Department of Neurosurgery, MGH, buhrmanj@gmail.com

13

Poster Board Number: 9 CLN3 and CLN6 encode proteins that function at distinct points in a membrane trafficking pathway commonly disrupted in the neuronal ceroid lipofuscinoses Yi Cao, Janice Espinola, Nicole Smith, Vijaya Ramesh, Marcy MacDonald, Susan Cotman

Molecular Neurogenetics Unit, Center for Human Genetic Research, Massachusetts General Hospital, Boston, MA Purpose: The neuronal ceroid lipofuscinoses (NCL) are a genetically heterogeneous group of devastating neurodegenerative disorders that strike children, causing blindness, epilepsy, loss of motor and cognitive function, and premature death. Mutations in at least 8 mostly novel genes (CLN genes) are known to cause NCL, yet the disease pathway remains poorly understood. The overlapping disease features of the different forms of NCL suggest that the proteins encoded by the CLN genes function in a common pathway critically affected in NCL neurons. To test this hypothesis, we established precise genetic models for two forms of NCL, variant late-infantile NCL (vLINCL/CLN6) and juvenile NCL (JNCL/CLN3), for genotype-phenotype studies. Materials & Methods: CbCln6nclf cells, an immortalized cerebellar neuronal precursor cell culture model of vLINCL, were generated from Cln6nclf mice, to complement previously established CbCln3∆ex7/8

neuronal cell lines, vLINCL and JNCL patient lymphoblasts, and Cln3∆ex7/8 and Cln6nclf mouse models. To first further characterize the novel CLN6-encoded protein, we performed organelle marker co-immunostaining studies in wild-type and mutant CbCln6nclf cell lines and patient lymphoblasts. Membrane trafficking phenotypes, including endocytosis (dextran-Alexa488 uptake) and lysosomal trafficking assays (Lysotracker and LC3 turnover assays), were examined and compared across the precise genetic cellular models using automated image analysis, confocal microscopy, and immunoblotting. Results: Co-immunostaining for the CLN6 protein with a newly established monoclonal antibody and organelle markers revealed a predominantly ER subcellular localization. Previous studies have demonstrated that the CLN3 protein is predominantly observed in endosomes and lysosomes. Marker immunostaining also detected ER alterations in the homozygous CbCln6nclf cells that were not observed in homozygous CbCln3∆ex7/8 cells. Intringuingly, although the CLN6 protein localizes to ER, like Cln3 mutation, Cln6 mutation caused altered endocytosis, lysosome distribution, and disruptions in autophagy trafficking. However, we also noted differences in the trafficking defects caused by the vLINCL and JNCL mutations. Endocytosis was more severely compromised with Cln3 mutation, compared to Cln6 mutation, and the appearance of the lysosomes and endosomes were distinct in each genetic model. Conclusion: These results support the hypothesis that the CLN3 and CLN6 encoded proteins function at different steps in a common pathway. The preponderance of membrane trafficking abnormalities observed in the NCL genetic model systems suggests that altered membrane homeostasis may be a key factor in the NCL disease process. The establishment of neuronal culture and mouse models that precisely mimic the genetic defects causing human NCL provides an ideal system for further study of the importance of membrane homeostasis and trafficking to NCL. Moreover, the establishment of quantitative cell-based assays provides the opportunity to further probe the NCL disease pathway through chemical genetics and siRNA technologies, which may identify novel disease modifiers that can be further tested in mice and may ultimately lead to novel disease therapies. Author Contact: Yi Cao; Center for Human Genetic Research; 617-724-9677; YCAO1@partners.org

14

Poster Board Number: 10 Minimally invasive fluorescence endoscope for detection and treatment monitoring of ovarian cancer Jonathan Celli, Wei Zhong, Imran Rizvi, Conor Evans, Zhiming Mai, Johannes deBoer, Seok-Hyun Yun, Tayyaba Hasan Wellman Center for Photomedicine, Massachusetts General Hospital, Harvard Medical School

Purpose: The goal of this work is to develop a minimally invasive imaging system with sufficiently high resolution and selectivity at the cellular level to detect ovarian cancer (OVCA) micrometastases at an early stage and monitor their response to treatment. The dismal mortality rate for patients with ovarian cancer is often associated with these small, disseminated metastases that pose serious challenges to current diagnostic and treatment assessment techniques such as white light laparscopy. These modalities lack adequate sensitivity for detecting OVCA at the sub-clinical stage or for detecting recurrent disease.

Materials & Methods: We have developed a custom micro-endoscope which achieves 9 μm spatial resolution. This instrument employs epi-fluorescence illumination to image through a flexible optical fiber coupled to a microscope objective onto a low light sensitive CCD camera with on chip multiplication gain. Using this system we conducted imaging experiments on a murine animal model of OVCA previously established in the laboratory. In these studies the photosensitizer BPD-MA, used here both as a fluorescent contrast agent and in some cases additionally for photodynamic therapy (PDT) treatment, was injected into mice 1 hour prior to imaging. Images of tumor nodules were compared to histology in order to establish the specificity and sensitivity of the instrument. In some mice PDT treatment was conducted at the time of the initial imaging. These mice were imaged again 5-7 days following treatment to monitor the treatment efficacy. A recent modification to the micro-endoscope system includes the addition of optical coherence tomography (OCT) imaging, a highly sensitive, cross-sectional imaging technique (analogous to ultrasound) capable of visualizing the microscopic details of biological tissues to provide additional 3D structural information.

Results: We demonstrate that the endoscope can detect in vivo tumor nodules tens of microns in size, an order of magnitude smaller than previously reported. By comparison of imaging data to histology we have determined the system’s sensitivity and specificity for detecting tumor nodules, to be 86% and 53%, respectively. In experiments in which diseased mice were imaged with the endoscope immediately before and seven days following PDT treatment, a significant decrease in tumor burden is observed relative to control mice and confirmed by surgical resection and determination of tumor burden by weight.

Liver

Small Bowel

Tumor100 μm100 μm

Left: Fluorescence image of the peritoneal cavity of an OVCA mouse 1 hour after injection of BPD-MA. Center: H & E stain of a 5 μm section from the same region. Arrows indicate tumor nodules visible in each image. Right: en bloc resection showing advanced metastatic disease in OVCA mouse model.

Conclusion: We have developed and characterized a fluorescence endoscope for minimally invasive imaging and demonstrated its capability for detecting OVCA tumor nodules in vivo, an order of magnitude smaller than those reported previously.

Author Contact: Jonathan Celli, Wellman Center for Photomedicine, 617-726-3991, jcelli1@partners.org

15

Poster Board Number: 11 Dynamic In Vivo VEGF Monitoring Strategy for Mechanism-based Optimization of Anti-angiogenic Cancer Therapy Sung K. Chang, Imran Rizvi, Nicolas Solban, Zhiming Mai, Arshi Malik, Tayyaba Hasan Wellman Center for Photomedicine, Massachusetts General Hospital, Boston MA Purpose: A major challenge facing effective cancer management is timely identification of patients with recurrent or resistant disease requiring more aggressive treatment. Current efforts to address this challenge are spearheaded by the use of biomarkers. Because of its critical role in tumor growth and metastasis, vascular endothelial growth factor (VEGF) is being actively investigated as a prognostic biomarker. Despite the biological significance of VEGF, the results of these investigations show mixed success, which highlights the limitations of the current approach: (i) current studies often measure a snapshot of VEGF status, although it is becoming increasingly clear that temporal VEGF dynamics can provide more sensitive information, (ii) current ex vivo assays have limited sensitivity and/or specificity for measuring VEGF. In an effort to address these limitations and design a more innovative strategy, this study aims to (i) investigate the significance of temporal VEGF dynamics in cancer treatment and (ii) develop a novel technology to monitor the VEGF dynamics with high sensitivity and specificity. Materials & Methods: In order to investigate the temporal VEGF dynamics following cancer therapy, VEGF levels from orthotopic prostate tumors were assessed at various timepoints (1-, 3-, 7- and 14-days) following PDT using an ex vivo assay. To investigate the therapeutic significance of the dynamic response, a separate study was conducted in which Avastin, an anti-VEGF therapy, was delivered at the different timepoints (1-, 3-, 7- OR 14-days) following PDT. In order to translate the significant findings from the preclinical study into the clinic, technologies capable of monitoring the temporal VEGF dynamics are needed. For this purpose, a novel optical molecular imaging strategy was investigated in a proof-of-principle study. The imaging strategy utilizes (i) fluorescently-tagged Avastin for in vivo VEGF labeling and (ii) a hyperspectral imaging system for sensitive detection. The specificity of the novel imaging strategy was tested in detail using both empirical and analytical approaches. Results: VEGF response following PDT was transient with a peak response at 1-day post-PDT. This temporal VEGF dynamics had significant implications in cancer therapy, as combination anti-VEGF therapy was most effective in controlling metastasis when delivered at the time of peak VEGF response. The novel optical molecular imaging strategy was capable of interrogating the temporal VEGF dynamics in vivo with high sensitivity and specificity. Conclusion: This is the first report to demonstrate the significance of the temporal VEGF dynamics in optimizing combination therapies and enhancing the overall treatment outcome. This observation points to the need to develop technologies to monitor dynamic molecular response to cancer therapy. To address this need, we also developed the first optical molecular imaging strategy to continuously monitor secreted protein dynamics in vivo. The results of this study have important implications in personalized treatment strategies, as in vivo VEGF monitoring technology can help inform the need as well as the optimal timing of combination anti-angiogenic therapy in individual patients. Author Contact: Sung Chang, Dermatology, 617-726-6139, SKCHANG@PARTNERS.ORG

16

Poster Board Number: 12 A novel molecular mechanism for the attenuation of NFκB signaling Vasant Chellappa, Stewart Moran, Sven Diederichs, Sheila Le, Cristian Boboila, and Shiv Pillai

MGH Cancer Center and Harvard Medical School, Boston, Massachusetts, USA Purpose: PKK is a member of the RIP family of serine/threonine kinases, which is expressed primarily in transit amplifying/progenitor cells and is required for proper differentiation of the epidermis. PKK is aberrantly expressed at high levels in certain cancers and is required for the survival of a subset of human B cell lymphomas. PKK contains multiple ankyrin repeats and activates NFκB, when over-expressed in mammalian cell lines. This kinase activates NFκB in a unique IKKγ and IKKα independent manner, distinct from the known “canonical” and “alternative” pathways for NFκB activation. We have sought to understand the mechanism by which PKK activates NFκB in an IKKγ independent manner. Apart from obtaining biochemical insights regarding this “third” pathway of NFκB activation, we have discovered a novel molecular mechanism by which NFκB signaling is attenuated. Materials & Methods: HEK 293T and U2OS cell lines, epitope-tagged cDNA expression constructs for PKK and IKKβ, site directed mutagenesis, and the proteasome inhibitor, MG-132. Results: We show that PKK associates with IKKβ and thus contributes to IκBα phosphorylation and degradation, similar to that observed in the “canonical” NFκB pathway. PKK mediated IKKβ activation requires the conserved 177S and 181S residues on IKKβ. Intriguingly, the association of PKK with IKKβ in over-expression studies leads to the rapid proteasomal degradation of IKKβ. This degradation event requires catalytically active PKK and 177Sxxx181S motif in IKKβ that is phosphorylated by upstream kinases during NFκB activation. Indeed, we find that this rapid 177S and 181S dependent turnover of IKKβ also occurs in MEKK1 induced “canonical” NFκB activation. Conclusion: While the physical association of PKK with IKKβ and the requirement for 177S and 181S residues on IKKβ have provided insights about PKK mediated NFκB activation in the absence of IKKγ, a far more far-reaching conclusion has been obtained from these studies. Attenuation of NFκB signaling, is known to be mediated in part by the feedback re-synthesis of IκBα. Here, we show that 177S and 181S residues on IKKβ that are phosphorylated during NFκB activation, also targets the active phosphorylated form of this kinase for rapid proteasomal degradation. This constitutes a previously unidentified but, potent mechanism for the rapid attenuation of “canonical” as well as PKK-induced NFκB signaling. Author Contact: Vasant Chellappa (vchellappa@helix.mgh.harvard.edu) and Shiv Pillai (pillai@helix.mgh.harvard.edu), MGH Cancer Center, Boston, MA.

17

Poster Board Number: 13 Intestinal Alkaline Phosphatase Protects Against E Coli Invasion In Vitro Kathryn Chen, Madhu Malo, Golam Mostafa, Skye Zeller, Sayeda Alam, Elizabeth Hohmann, Richard Hodin Purpose: The gut barrier performs the critical function of protecting the host from commensal and pathogenic bacteria. Intestinal alkaline phosphatase (IAP) is a brush border enzyme that has the ability to dephosphorylate and detoxify lipopolysaccharide. We hypothesized that IAP protects intestinal cells from invasion from certain bacteria. Materials & Methods: Varying amounts of LPS were incubated with and without calf intestinal alkaline phosphatase (CIP) and applied to HT29 cells for 24 hours. Il-8 secretion by the intestinal cells in response to LPS stimulation, and the effect of CIP on that response, was measured by ELISA Human intestinal HT29 cells were stably transfected with an IAP expression plasmid, and IAP enzyme activity was confirmed biochemically. A commensal Escherichia coli was isolated from wild type mouse stool. Wild type and IAP-expressing cells were incubated with E. coli. At various times, gentamicin was applied to kill the extracellular bacteria; cells were lysed, and intracellular bacteria enumerated by plating and counting colony forming units (cfu). Il-8 secretion into tissue culture medium was determined by ELISA. The mouse commensal E. Coli was grown with and without CIP in BHI. OD600 of the culture was taken each hour. Results: HT29 cells exposed to LPS showed significantly elevated levels of Il-8 at LPS concentrations ranging from 0.5 ug/mL to 1 ug/mL. Il-8 production was blocked by the presence of CIP with LPS at all concentrations. Wild type cells showed significantly higher levels of E. coli invasion as compared to IAP-expressing cells. The increase in intracellular bacteria was time and dose dependent. At the longest incubation time (3 hours) and multiplicity of infection (104 bacteria per cell), wild type cells contained 3-fold more E. coli than did IAP cells. Il-8 secretion was also up to 4-fold greater in wild type cells, and was proportional to the level of intracellular E. coli. Mouse commensal E. coli grown aerobically without CIP exhibited a standard growth curve. The presence of CIP killed the mouse commensal organism in log phase growth. Other E. coli tested were unaffected by presence of CIP. Conclusion: IAP is able to protect HT29 colonic cells from LPS induced Il-8 secretion. IAP appears to prevent invasion of E Coli within HT29 cells, even at a high multiplicity of infection. Il-8 secretion, a marker of bacterial invasion, is altered in parallel. Furthermore, the CIP appears to inhibit E. Coli during log phase growth We conclude that IAP within the gut may protect the host from over proliferation and invasion of commensal bacteria, and from consequent inflammatory reaction within the gut. Author Contact: Kathryn Chen, GI unit, 614-595-5352, Kathryn.chen@gmail.com

18

Poster Board Number: 14 A Point Process Framework to Assess Cardiovascular Functions Zhe Chen, Emery N. Brown, and Riccardo Barbieri Neuroscience Statistics Lab, Dept of Anesthesia and Critical Care, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA Purpose: To estimate and monitor the instantaneous heartbeat dynamics and hemodynamics involved in cardiovascular control, which are important for evaluation of a large spectrum of cardiovascular diseases and disorders (e.g., hypertension and congestive heart disease). To provide a potential real-time noninvasive assessment for ambulatory monitoring in both laboratory research and clinical practice. Materials & Methods: We have devised a novel statistical framework where heartbeats, once detected from continuous ECG signal, are treated as discrete events that can be modeled by a stochastic point process. An inverse Gaussian probability distribution can be used to model the heartbeat interval, whereas its mean is modulated by previous inter-beat intervals. To test if highly nonlinear dynamics are present, we have included the option of extending the autoregression with a Volterra-Wiener series expansion. Additionally, we have modeled respiratory influence with a further linear regression, and the interactions between heartbeat intervals and systolic blood pressure measurements with a bivariate autoregressive structure to mimic the closed-loop cardiovascular system. Various statistical indices of interest, such as the instantaneous heart rate (HR), heart rate variability (HRV), respiratory sinus arrhythmia (RSA), as well as the baroreceptor-cardiac reflex (Baroreflex) can be rigorously derived within a parametric framework and instantaneously updated with an adaptive algorithm. Results: We report findings from three different experimental settings. First, comparison across 15 subjects in an "autonomic blockade assessment of the sympatho-vagal balance and RSA'' protocol indicated that RSA gain is greater in supine than upright posture, and greatest in the control condition relative to other pharmacological conditions where atropine, propranolol, or both are administered. All these differences are statistically significant. Second, comparison of Baroreflex gain between rest and tilt conditions in a “tilt-table'' protocol indicated the gain in the rest condition is statistically greater than that in the tilt condition, accompanying with higher HRV. Third, comparison between 7 congestive heart failure patients and 10 healthy subjects (all retrieved from Physionet database) indicated the heartbeat dynamics in healthy subjects exhibit stronger nonlinearity, confirmed by a nonlinearity test as well as a lower spectrum/bispectrum power ratio. Conclusion: A novel point process framework was proposed to simultaneously assess linear and nonlinear indices of HRV, together with RSA and Baroreflex, under a wide range of experimental protocols. Results confirm established findings regarding the most important physiological and pathological mechanisms involved in cardiovascular control, and they also reveal interesting dynamic trends across different posture/pharmacological/age/heart disease conditions. Author Contact: Zhe Chen; Anesthesia; 617-724-1061; zhechen@neurostat.mgh.harvard.edu

19

Poster Board Number: 15 Pd-1 Exerts Its Inhibitory Effect On Primary Human T Cells In Hiv By Interfering With Proximal Tcr Signaling And Targeting Downstream Signals Associated With Proliferation And Survival. Quentin Eichbaum1, Mohammed El-Far2, Francesco Procopio2, Zhong He2, Chao Qui2, Andre Tanel2, Simone Fonseca2, Mark Brockman1, Rachid Boulassel5, Jean-Pierre Routy5, Gordon Freeman3, Michael L. Dustin4, Elias Haddad2, Rafick-Pierre Sekaly2, Bruce Walker1. 1Partners AIDS Research Center, Massachusetts General Hospital, Boston; 2Laboratoire d’Immunologie, Centre de Recherche CHUM Saint Luc, Montreal, Quebec, Canada; 3Dana Farber Cancer Inst, Boston. 4Skirball Inst, NYU, New York; 5 McGill University Health Centers, Montreal, Canada. Purpose: Programmed death-1 (PD-1) - a member of the CD28 family of immune modulators - has two ligands (PD-L1 and PD-L2) which, upon engagement, inhibit the immune response of the T cell. Recently, it was shown that PD-1 is upregulated on ‘exhausted’ T cells during chronic HIV infection. Of potential therapeutic value, is the additional finding that blockade of this PD-1-ligand pathway reinvigorates these HIV+ exhausted T cells and allows them once again to produce effector cytokines and to proliferate. However, the molecular mechanisms by which PD-1 exerts its inhibitory effect on T cell immune function remains unknown. Since PD-1 is dependent on the TCR for effective signaling, we hypothesized that PD-1 achieves its inhibitory effect by interfering with proximal TCR signaling. The aim of this study was to determine the pathway(s) by which PD-1 functions as an inhibitor of the T cell immunity, particularly in ‘exhausted’ T cells from chronically-infected HIV+ individuals. Materials & Methods: To examine the inhibitory signaling pathway of PD-1, we developed an APC model to stimulate human T cells using magnetic beads coated with anti-CD3 and anti-CD28 antibodies together with either an agonistic anti-PD-1 antibody or isotype control antibody. We used IL-2 secretion as a readout assay (by ELISA) and examined responses in both PD-1hi (stably transfected) Jurkat cells as well as in PD-1hi healthy and HIV-infected human CD4+ and CD8+ cells. We applied this model (which was also validated in 293 cells transduced with the natural ligands PD-L1/L2 in lentiviral constructs) to determine the signal transduction pathway of PD-1 by Western blot, phosflow, immunoprecipitation as well as by dynamic imaging using confocal and TIRF microscopy. To verify the inhibitory potential of specific signaling molecules in the PD-1 pathway, we used the approaches of RNAi knockdown and molecular mutation to demonstrate ‘rescue’ of IL-2 function from inhibition. Finally, BrDU/Ki67 and cell cycle assays were used to identify downstream survival and proliferation signals implicated in PD-1 signaling. Results: We show that PD-1/TCR co-ligation leads to increased recruitment of the phosphatases SHP-1 and SHP-2 to the PD-1 cytoplasmic tail. We provide evidence by several methods that PD-1 signals through p56Lck and abrogates TCR signaling by increasing phosphorylation of the pLck505 - the ‘inactive’ phosphorylated form of p56lck. In addition, we show that Csk, the immediate upstream regulatory kinase of pLck505, may be recruited to PD-1. Target signals immediately downstream of these TCR effector molecules are attenuated (CD3, ZAP70, Slp76, vav, Erk). Using RNAi and molecular mutation, we demonstrate that inhibition of SHP1/2, Csk or p56Lck can ‘rescue’ IL-2 function in T cells. We corroborate these biochemical and molecular findings through dynamic imaging, using confocal and TIRF microscopy together FRET constructs, to show how these signals may mediate the inhibitory effect of PD-1 in the immunological synapse. Finally, we identify downstream signals associated with cell survival and proliferation that may be implicated in mediating T cell ‘dysfunction’ in chronic HIV. Conclusion: These findings suggest that PD-1 exerts its inhibitory effect by interfering with TCR signaling through the proximal negative regulators SHP1/2, Csk and pLck505 - which may also impact downstream signals associated with proliferation and survival. These results are significant in that they reveal novel signaling mechanisms through which PD-1 exerts its inhibitory effects on immune function in T cells. Furthermore, these findings suggest potential strategies for therapeutic intervention via manipulation of these inhibitory signals to restore T cell function in diseases such as chronic HIV. Author Contact: Quentin Eichbaum; Partners AIDS Research Center; 617-697-9556; qeichbaum@partners.org

20

Poster Board Number: 16 Development of an In Vitro Model to Study Vascular Injury Eric B. Finkelstein, Ph.D1,2., Gwen E. Owens, B.S1., David M. Hoganson, M.D1,2., Mark Keegan, Ph.D4., James Hsiao, B.S4., Irina Pomerantseva, M.D., Ph.D1,2., Craig M. Neville, Ph.D1,2., Jeffery Bornstein, Ph.D4., and Joseph P. Vacanti, M.D1,2,3. 1Center for Regenerative Medicine, Massachusetts General Hospital; 2Harvard Medical School; 3MassGeneral Hospital for Children, Boston, MA USA; 4Charles Stark Draper Laboratories, Cambridge, MA USA. Purpose: Drug-induced vascular injury (DIVI) is a common complication during preclinical testing of diverse pharmacological compounds. DIVI injuries have been observed in the mesenteric arteries of rats and the coronary arteries of dogs, but not in humans. The vascular injury is characterized by extravasation of red blood cells (RBCs) from the vasculature into the surrounding smooth muscle layer. It is believed that the RBCs leave the vessels through “holes” that form in the endothelial cells through an uncertain mechanism. Additionally, there are no definitive biomarkers of this vascular injury. Lack of a mechanism and biomarkers makes it difficult to transfer the observable injury in the dog and rat to potential effects in humans. To address the need for a better understanding of the mechanisms of DIVI and to evaluate potential biomarkers, we have developed an in vitro model of the rat mesenteric vasculature. Materials & Methods: Using microfabrication technology, a microvascular channel was manufactured from polydimethylsiloxane (PDMS) to mimic the mesenteric vasculature. The device is configured such that endothelial cells (ECs) are in co-culture with smooth muscle cells (SMCs), separated by a polycarbonate membrane. Flow can be induced in this model to mimic the arterial shear stress observed in the mesentery. Conditions for seeding rat aortic ECs and SMCs on the polycarbonate membrane were developed using Transwells fitted with the same polycarbonate membrane. From this, we have developed optimal conditions for cell seeding in devices with rat aortic ECs and SMCs in co-culture. Fluorescent microspheres were used as an extravasation test to assess the cellularity of devices. Results: Using Transwell co-cultures, we show that seven days of SMC culture prior to the addition of ECs to the opposite side of the membrane is optimal for consistent EC growth. We have adapted this to the seeding of devices, and subjected them to flow conditions. Device seeding has also been optimized to account for the low surface area to volume ratio within the cell-seeding areas. Here we show that devices can be cellularized, with rat ECs and SMCs in co-culture for 14 days. Microsphere extravasation tests were carried out and when the EC layer was intact, there was little extravasation into the SMC layer. Conclusion: This microfluidic co-culture model will be useful to examine the extravasation induced by potential DIVI compounds. The cells in the devices can also be used to explore potential biomarkers of the drug-induced injury. In addition, this model can be used to examine other vascular injuries and disease that is not drug-induced, including those of the human vasculature. Author Contact: Eric B. Finkelstein, Ph.D., Surgery/Center for Regenerative Medicine, 617-643-3377, ebfinkelstein@partners.org

21

Poster Board Number: 17 Graft Endothelium is the Dominant Driver of Natural Killer Cell Stimulation and Chronic Allograft Vasculopathy in Heart Grafts Jay A. Graham, Tsutomu Hirohashi, Catharine M. Chase, Robert B. Colvin, Paul S. Russell, Joren C. Madsen The MGH Transplant Center, Massachusetts General Hospital, Boston Purpose: We have demonstrated that natural killer (NK) cells, members of the innate immune system, contribute to the development of chronic allograft vasculopathy (CAV) in the heart. This finding uncovered a previous unknown pathway toward CAV and assigned a novel role to NK cells in solid organ allograft rejection. In an attempt to isolate NK cell activity and eliminate the alloresponse of B and T cells we used the parental to F1 transplantation model. While parental to F1 grafts develop profound CAV lesions, much to our surprise NK cell depletion failed to abrogate their formation. Interestingly only NK and T cell inhibition could prevent CAV lesion pathogenesis. These finding had many implications, one of which was that NK cells alone could not cause CAV. Our latest experiments look to explore the role of donor antigen presenting cells (APCs) and T cells on the activation of NK cells and development of CAV. Materials & Methods: To accomplish this goal, we created full bone marrow chimeras by lethal irradiation and bone marrow transplantation. Two different bone marrow chimers were created for the experimental groups. B6 (H2b) mice were lethally irradiated and reconstituted with bone marrow (1x10e7 cells) from CB6F1 (H2bxd) mice. We refer to B6 mice receiving CB6F1 bone marrow as B6(CB6F1) chimeras. Conversely, CB6F1 mice that received B6 marrow are referred to as CB6F1(B6) chimeras. Chimeras were deemed fully replaced with donor-type APCs after phenotypic confirmation with flow cytometry by staining peripheral blood with PE anti-mouse H2Db (KH95) and FITC anti-mouse H2Dd (34-2-12). Moreover, immunohistochemical analysis on the spleen with biotinylated anti-mouse H2Dd (34-2-12) further confirmed our flow cytometric analysis. Previous studies have demonstrated that the degree of T cell and APC replacement in the spleen accurately reflects the degree of lymphocytic replacement in the heart. Results: Nine of ten B6(CB6F1) hearts transplanted into untreated CB6F1 mice developed CAV. This result suggested that NK cells are being activated by the graft endothelium because the graft APCs and T cells are isogeneic. Conversely, only one of seven CB6F1(B6) hearts transplanted into CB6F1 recipients developed CAV, further suggesting that the endothelium is the dominant stimulator of CAV lesion formation. Conclusion: While a definitive role of NK and T cells in CAV production still remains unclear, our experiment suggests that the donor population of lymphocytes does not participate in this lesion formation. This striking finding implicates a graft endothelium directed recipient response, as the chimeric donor hematopoetic population mimics the recipient pool. Given the sheer bulk of the vascular endothelium and the fact that endothelial cells can act as potent APCs for T cells in vivo, we surmise that graft endothelium is the main driver of NK cell stimulation and CAV formation. Author Contact: Jay Graham, Transplant Center, MGH, 617-726-6536, jgraham8@partners.org

22

Poster Board Number: 18 PRINTING A 3D BLADDER MODEL TISSUE Syed K Hasan*1, SangJun Moon*1, Sohan Mikkilineni1, Pei-Ann Lin 1, Fahim Manzur1, Young S. Song1, Jiro Nagatomi3, Ali Khademhosseini 2, Utkan Demirci 1,2ׁש

1 Bio-Acoustic- MEMS in Medicine (BAMM) Laboratory, HST-Center for Biomedical Engineering, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston MA USA 2 Harvard – Massachusetts Institutes of Technology Health Sciences and Technology, 65 Landsdowne Street, Cambridge MA 02139 USA 3Department of Bioengineering, Clemson University,313 Rhodes Engineering Research Center, Clemson, SC 29634-0905 Purpose: Many common diseases involve the injury, loss, or death of organ tissues. Bioengineering models allow understanding/elucidating mechanisms to create 3D Tissues. The ability to engineer 3D tissues is a powerful approach to treat diseases or to develop in vitro tissue culture models. To date we have bioengineered rat bladder patches (5cm x 5cm) consisting of smooth muscle cell layers as observed in native rat bladder tissue. Materials & Methods: Our novel and powerful new cell printing technology allows for rapid (3000 cells/sec) printing of bladder smooth muscle cells encapsulated in collagen hydrogel droplets accurately, reliably, and with high viability (>90%) and functionality. Moreover, our printing technology provides spatial and temporal control of printing viable and functional cells. This technology enables microscale level control of the smooth muscle 3D cell position and orientation. Printed 3D bladder tissue constructs are five layers thick, allowing for proper diffusion of oxygen and nutrients. Results: 3D bladder patches were made by printing smooth muscle cells (5.6 x105 cells) in a cross over pattern and cultured over 7 days. Smooth muscle cells were characterized for viability, functionality, and morphology using histological stains (e.g., DAPI, eosin, hemotoxylin). Dynamic characterization of the constructs includes mechanical (e.g. Biaxial elongation) testing. Gap junctions were characterized with appropriate histological stains (e.g. Connexin 43, and cadherin), for both cell viability and cellular architecture. Characterization of the phenotype of layered SMC’s within tissue patches was done by immunohistochemistry for proteins specific to smooth muscle cell (e.g. Rho A, CPI-17) Conclusion: 3D Bladder patches can be created with proper cellular directionality and architecture using our 3D cell printing technology. Author Contact: Syed K Hasan MD, 617-768-8315, docsy@mit.edu

23

Poster Board Number: 19 Enhanced neurogenesis and neuritogenesis by docosahexaenoic acid in fat-1 transgenic mice Chengwei He, Jingdong Wang, Jing X. Kang Cardiovascular Research Centre, Massachusetts General Hospital, Harvard Medical School 149 13th Street, Charlestown, Massachusetts 02129 Purpose: Docosahexaenoic acid (DHA), an n-3 long chain polyunsaturated fatty acid (LC-PUFA) which is highly enriched in nerve system, is critical for brain development and function. Deficiency of DHA results in impaired cognitive performance, while supplement of DHA improve the condition. However, little is known about how DHA affects the brain structure in terms of neurogenesis and neuritogenesis directly associated with cognitive activity. The purpose of this study is to determine whether DHA enhances neurogenesis and neuritogenesis in vitro and in vivo and the relevance to cognitive activity in a well-controlled fat-1 transgenic mouse model. Materials & Methods: A fat-1 mouse model was used in this study. Fat-1 mouse is a transgenic mouse that expresses C. elegans fat-1 gene, whose product catalyses the conversion of n-6 to n-3 PUFA. The absolute amount of n-3 PUFA increases significantly with the decrease of corresponding n-6 PUFA endogenously all the time in all tissues of fat-1 mice fed with the same diet as wild mice do. Thus, confounding factors such as heavy metals contamination, variation in content and composition, and oxidation of the supplemented oil can be avoided. Lipid profile was tested by gas chromatography. The hippocampal neurogenesis and neuritogenesis were measured by BrdU incorporation assay and Golgi-Cox staining, respectively. Learning and memory performance was determined by Morris water maze. Differentiation of mouse embryonic stem (ES) cells into neuronal cells was applied to evaluate the effects of DHA on neurogenesis and neuritogeneis in vitro. Results: Fat-1 mice had a significant higher level of DHA in the brain compared to wild type mice. The hippocampal neurogenesis and neuritogenesis were significantly promoted in fat-1 mice compared to wild type mice. Concurrently, these mice exhibited improved performance of learning and memory. Furthermore, DHA promoted differentiation and neurite outgrowth of neural cells differentiated from ES cells in vitro. This was probably due to an enhanced transition from totipotential cells to neural stem cells (NSC) and/or from NSC to mature neuronal cells. DHA also increased the proliferation of cells undergoing differentiation into neuronal lineages from ES cells which may be owing to an anti-apoptotic property of DHA-derived docosatrienes such as neuroprotectin D1. Conclusion: These results together provide direct evidence for a positive effect of DHA on neurogenesis and neuritogenesis and suggest that this effect may be a mechanism underlying its beneficial effect on cognitive function. Author Contact: Chengwei He, Cardiovascular Research Centre, 617-726-6492, che3@partner.org

24

Poster Board Number: 20 Impact of myelomonocytic MHC class I receptors on HIV-1 infection Jinghe Huang1, Daniel Kavanagh1, Thai Cung1, Luis Borges2, Mathias Lichterfeld1, Xu G. Yu1

1Partners AIDS Research Center, Massachusetts General Hospital and Harvard University Center for AIDS Research, Boston, MA, USA 2Amgen Inc. Seattle, WA, USA Purpose: Myelomonocytic MHC class I receptors represent important immunoregulatory molecules that can either stimulate or inhibit functional properties of professional antigen-presenting cells, and may thus have a critical impact on the qualities of HIV-1-specific T cells. Here we analyzed the surface expression of a panel of different myelomonocytic receptors in HIV-1 infection and assessed the impact of DC maturation cytokines on expression of various ILTs. Materials & Methods: Flow cytometry was used to determine the surface expression intensity of myelomonocytic MHC class I receptors (ILT1, ILT2, ILT3, ILT4, and ILT5) on monocytes, dendritic cells, as well as B and T lymphocytes in HIV-1-infected and -uninfected persons. To detect the influence of DC maturation factors on surface expression of ILTs, monocyte derived dendritic cells were stimulated to mature under various conditions and ILT expression assessed by flow cytometry. Results: In comparison to uninfected persons, the stimulatory myelomonocytic receptor ILT1 was down-regulated in HIV-1 infected persons (p=0.01), while the inhibitory receptors ILT2, ILT3, and ILT4 were strongly up-regulated (p<0.01) on myelomonocytic cells. Expression of ILT2 was specifically elevated on CD8+ T cells (p=0.001). Finally, we observed that polyinosinic-polycytidylic acid double-stranded RNA (Poly I:C), induced down-regulation of the inhibitory receptors such as ILT2, ILT4, and ILT5 together with up-regulation of stimulatory receptor ILT1 as compare to the control cytokine combinations Mimic (IL1β, TNFα, IL-6, PGE-2), CL097 (TLR7 and TLR8 ligands), and sLPS. Conclusion: These data suggest that the decrease of stimulatory MHC class I receptors and increase of inhibitory MHC class receptors in HIV-1 infection may represent a critical mechanistic factor contributing to immune deficiency in HIV-1 infection. Poly I:C is a strong regulator of decreasing inhibitory myelomonocytic MHC class I receptor as well as increasing stimulatory myelomonocytic MHC class I receptor and may therefore be particularly suitable for dendritic-cell based immunotherapeutic interventions to strengthen T cell mediated immune responses against HIV-1. Author Contact: Jinghe Huang, PARC, 6176059567, hjinghe@partners.org

25

Poster Board Number: 21 Effect of Ferumoxtran-10 Induced Susceptibility on Nodal Size on T1-, T2- and T2*- Weighted Lymphotropic Nanoparticle-Enhanced MRI Tina Islam ; Amol S Katkar ; Ravi T Seethamraju ; Peter F Hahn ; Mukesh G Harisinghani1,2 1 3 2 1,2

1 Center for Molecular Imaging Research, 2 MGH Department of Radiology, 3 Siemens Medical Solutions, USA Inc. Purpose: Lymphotropic nanoparticle-enhanced MRI (LNMRI) using the superparamagnetic contrast agent ferumoxtran-10 is an accurate technique for lymph node characterization in various primary malignancies. The purpose of this study was to compare the effect of ferumoxtran-10 induced susceptibility on nodal short axis diameter (SAD) on pre- and postcontrast T1-weighted gradient echo (GRE), T2-weighted fast spin echo (FSE) and T2*-weighted GRE images. Materials & Methods: 24 patients (16 males, 8 females, mean age 66.34 years) with various primary malignancies [prostate cancer (n=13); bladder cancer (n=6); testicular cancer (n=3); breast cancer (n=2)] underwent T1-weighted GRE, T2-weighted FSE, and T2*-weighted GRE (dual echo using short (12.6 ms) and long (21 ms) TE values) 1.5 T MR imaging before and 24-36 hours following the intravenous administration of ferumoxtran-10 (Combidex; AMAG Pharmaceuticals Inc, Cambridge, MA). The SAD was measured on the pre- and postcontrast MR images and paired t-test was used to compare the mean sizes. All evaluated lymph nodes underwent pathological examination by surgical resection or image guided biopsy. Results: 65 lymph nodes were seen on imaging that were unequivocally correlated with histopathology: 45 (69.2%) were benign and 20 (30.8%) were malignant. Mean values of SAD before and after ferumoxtran-10 administration did not differ significantly for malignant nodes. For benign nodes, mean SAD appeared significantly larger on the postcontrast T2* GRE sequence both for a TE of 12.6 ms (p<0.0001) and 21 ms (p<0.0001), with this effect being significantly larger at the higher TE (p<0.0001). Postcontrast mean SAD was significantly larger (p<0.0001) for T2*-weighted GRE when comparing it with postcontrast T2-weighted FSE. No differences were observed for pre- and postcontrast T1-weighted GRE and T2-weighted FSE. Conclusion: Susceptibility induced by ferumoxtran-10 leads to a blooming artifact and significantly changes the mean nodal short axis diameter on postcontrast T2* weighted gradient recalled echo in benign lymph nodes, even at short echo times. If only postcontrast sequences are used with ferumoxtran-10 for nodal characterization, the size of the lymph nodes should be measured on the T1-weighted GRE or T2-weighed FSE pulse sequences. Author Contact: Tina Islam, M.D., Ph.D., CMIR/Department of Radiology, 6177244266/ 6176992703 Islam.Tina@mgh.harvard.edu

26

Poster Board Number: 22 Decreased uptake of ferumoxtran-10 in inguinal lymph nodes in lymphotropic nanoparticle-enhanced MRI Tina Islam ; Peter F Hahn ; Mukesh G Harisinghani1,2 2 1,2

1 Center for Molecular Imaging Research, 2 MGH Department of Radiology Purpose: Lymphotropic nanoparticle-enhanced MRI (LNMRI) has been shown to cause a homogeneous drop in signal intensity of benign lymph nodes. In patients with primary prostate cancer we evaluated morphological characteristics of inguinal lymph nodes and their signal intensity behavior on LNMRI. This patient population was selected because their inguinal lymph nodes are likely to be benign. Materials & Methods: 29 patients with prostate cancer (mean age 63 years) underwent T2-weighted fast spin echo and T2*-weighted gradient echo MR imaging before and 24-36 hours following the intravenous administration of ferumoxtran-10 (Combidex®; AMAG Pharmaceuticals Inc, Cambridge, MA). On the precontrast images shape (oval or round), border contour (well- or ill-defined), hilum status (presence or absence of fatty hilum), signal homogeneity (homogeneous or heterogeneous) were described and nodal short axis diameters were measured for inguinal nodes. On the postcontrast T2*-weighted images the percentage (<30%, 30%-50%, or >50%) of high signal intensity within the benign inguinal nodes were estimated. Results: 233 inguinal lymph nodes were examined in 29 patients. 93.6% were oval, 57.9% showed a fatty hilum, 90.6% showed a well-defined border, and 92.3% a homogeneous signal. Mean short axis nodal diameter was 0.42 cm (+/- 0.17). Estimated high signal intensity of <30% was found in 31.3% of all groin nodes, indicative of benign differentiation. Estimated high signal intensity region between 30%-50% and >50% was found in 23.2% and 45.5% respectively, mimicking malignant features. Conclusion: In a patient population with unlikely tumor involvement of inguinal lymph nodes, 68.7% of the presumed benign nodes show a high signal intensity region larger than 30% on LNMRI, similar to that observed in malignant lymph nodes. Benign inguinal lymph nodes take up nanoparticles less avidly than benign deep nodes do. Consequently, different standards must be used in interpreting the appearance of inguinal nodes on LNMRI performed for tumors that typically spread to the groin. Author Contact: Tina Islam, M.D., Ph.D., CMIR/Department of Radiology, 6177244266/ 6176992703 Islam.Tina@mgh.harvard.edu

27

Poster Board Number: 23 Human Glioblastoma Stem Cell and Endothelial Cell Co-cultures as a model of the tumor perivascular niche and the therapeutic role of Oncolytic Herpes Simplex Virus Deva Sanjeeva Jeyaretna, MD, Hiroaki Wakimoto, MD, Samuel Rabkin, PhD and Robert Martuza, MD. Department of Neurosurgery, Massachusetts General Hospital Purpose: Glioblastoma multiforme, the most malignant primary brain tumor, carries a median survival of 12 to 14 months. In human glioblastomas, glioma stem cells (GSCs) are thought to exist in close proximity to and interact closely with endothelial cells (EC). These GSCs are thought to be responsible for mediating resistance to standard chemotherapeutic and radiotherapy regimes. We hypothesize that the interaction between GSCs and ECs can be modeled in vitro. Oncolytic Herpes Simplex Viruses (oHSV) are genetically engineered viruses that may be designed to selectively target rapidly dividing cells such as tumor cells while sparing normal cells. We hypothesize that oHSV can disrupt the interaction between GSC and ECs. Materials & Methods: We established two primary culture GSC lines from surgical specimens of human glioblastomas. Stem-like attributes of the GSCs including sphere formation, CD133 positivity and in vivo tumorigenicity were determined. The ability of oHSV to infect GSCs was analyzed using a GFP expressing oHSV. GSC and EC susceptibility to oHSV in co-culture was assessed using direct-contact and non-contact models. In the direct contact model, GSC and ECs were co-cultured on Matrigel where ECs differentiate to form tubes. To determine if any effect was due to secreted factors the experiment was repeated with GSC conditioned media. Once the co-culture was established, the wells were infected with oHSV. The ability of oHSV to infect and kill GSCs in co-culture was also tested in a non-contact model. A transwell co-culture assay was used in which GSCs and ECs were separated by a permeable membrane that prevented cell contact but allowed secreted factors to be exchanged. After 24 hours, GSCs were removed and pulsed with a dose of oHSV. After two hours the medium was changed to reduce the viral load and the GSCs returned to co-culture. Results: G47Δ our latest generation oHSV was used in all the experiments. oHSV is able to efficiently kill proliferating ECs. We also demonstrated that GSCs are susceptible to oHSV infection and killing. Furthermore, the cytotoxic response was dose dependent and consistent across GSC cell lines. In co-cultured wells, GSCs preferentially attach to EC tubes on Matrigel and prolong the survival of the EC tubes when compared to control wells. Conditioned medium from GSC culture produced a similar increase in EC tube survival implying that although direct contact occurs in this model, it is not necessary for imparting a survival benefit. Conditioned medium from GSCs was analyzed for VEGF expression. Under hypoxic conditions GSCs were noted to increase the levels of VEGF secreted. The co-cultured tubes were then infected with oHSV which effectively disrupted tube formation. In the transwell assay, oHSV was able to infect and kill ECs and CD133 positive and negative GSC populations. Conclusion: We show that EC and GSC crosstalk can be modeled in vitro. oHSV is able to infect and kill both components of the perivascular niche in a co-culture model. This may aid in the development of oncolytic therapies targeted toward the stem cell fraction of glioblastomas. Author Contact: Deva Sanjeeva Jeyaretna, Department of Neurosurgery, 617-3197387, djeyaretna@partners.org

28

Poster Board Number: 24 An Assay System for Quantitative Analysis of Autophagy Robin Ketteler and Brian Seed. Center for Computational and Integrative Biology, Massachusetts General Hospital Department of Genetics, Harvard Medical School Purpose: Autophagy is an essential cellular process for the degradation of proteins and organelles that has been associated with neurogenerative diseases, cancer and infection. Although autophagy is currently widely investigated, the systematic identification of molecular events in autophagy has been hampered by the lack of suitable assays. We have devised a simple luciferase-based system that specifically detects key steps in autophagy. Materials & Methods: The method is based on the unexpected discovery of a non-conventional pathway for secretion of a potent enzymatic reporter, the luciferase from Gaussia princeps (GLUC). GLUC rapidly exits the cell without benefit of a secretory leader peptide, but can be anchored in the cell by fusion to β-actin. By including protease cleavage sites in a linker between GLUC and β-actin, events that increase or decrease proteolysis can be readily detected. Results: We present here a cell-based assay that specifically measures proteolytic cleavage of a tripartite sensor protein by the autophagy protease ATG4B. Expression of ATG4B induces a strong induction of luciferase secretion. We also demonstrate that rapamycin and shRNA-mediated knockdown of AKT1, treatments that lead to induction of autophagy, can efficiently activate this reporter. Conclusion: Proteolytically-activated secretion assays provide novel sensitive tools for the study of protease regulation and function in vivo. The luciferase release assay will be suitable for genome-wide functional screens and allows a quantitative analysis of autophagy. Further studies that contribute to the understanding of autophagy may help identify novel targets for therapeutic approaches. Author Contact: Robin Ketteler; Center for Computational and Integrative Biology; 617-642-3342; rketteler@ccib.mgh.harvard.edu

29

Poster Board Number: 25 IL-6 Cytoprotection In Hyperoxic Acute Lung Injury Occurs Via SOCS-1-Induced ASK-1 Degradation Narasaiah Kolliputi and Aaron B. WaxmanPulmonary and Critical Care Unit, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA. Purpose: Therapeutic strategies that improve lung cell resistance to reactive oxygen species may be superior to and certainly can complement anti-inflammatory and anti-oxidant strategies currently in use. We propose a novel strategy, where the by interleukin-6 (IL-6) induced Suppressor of Cytokine Signaling-1 (SOCS-1) has remarkably potent protective actions dramatically reducing lung cell death and profoundly increasing survival in response to 100% oxygen, in our mouse model of hyperoxia acute lung injury. Materials & Methods: The present study is the first report showing that SOCS-1 has direct cytoprotective actions against oxidant-induced injury in vitro and in vivo. We characterized the expression of SOCS-1 and downstream apoptosis signal-regulating kinase 1 (ASK)-1-Jun N-terminal kinase (JNK) signaling molecules in small airway epithelial cells (SAEC) in the presence of H2O2, which induces oxidative stress and in wild-type and lung-specific IL-6 Tg(+) mice exposed to 100% oxygen for 72 hours. Results: In control SAEC exposed to H2O2 or in wild-type mice exposed to 100% oxygen, a marked induction of ASK-1 and pJNK was observed. Both IL-6-stimulated endogenous SOCS-1 and SOCS-1 overexpression abolished H2O2-induced ASK-1 activation. In addition, IL-6 Tg(+) mice exposed to 100% oxygen exhibited reduced ASK-1 levels and enhanced SOCS-1 expression compared with wild-type mice. Interestingly, no significant changes in Tumor Necrosis Factor Receptor 1/Tumor Necrosis Factor Receptor-Associated Factor 2 (TNFR1-TRAF2), a key activator of ASK-1, activation were observed between wild-type and IL-6 Tg(+) mice. Furthermore, the interaction between SOCS-1 and ASK-1 promotes ubiquitin-mediated degradation both in vivo and in vitro. Conclusion: These findings suggest that SOCS-1 is an important regulator in IL-6-induced cytoprotection against hyperoxia induced acute lung injury and may provide a novel therapeutic approach to treat oxidant-mediated cellular injury. Author Contact: Narasaiah Kolliputi; Pulmonary and Critical Care Unit; 617-726-6567; nkolliputi@partners.org

30

Poster Board Number: 26 Cross Sectional Analysis of the Iliopsoas Tendon and Its Relationship to the Acetabular Labrum: An Anatomic Study Joshua Alpert M.D., Michal Kozanek M.D., Guoan Li Ph.D., Brian Kelly M.D., Peter Asnis M.D. Sports Medicine Service and Bioengineering Laboratory, Department of Orthopaedic Surgery, Massachusetts General Hospital, Boston, MA Purpose: To evaluate the anatomic relationship of the iliopsoas tendon to surrounding structures as well as the cross-sectional area of the tendon at the level of the hip labrum and at the level of the lesser trochanter. Materials & Methods: A controlled laboratory design was performed using 4 fresh-frozen cadaveric pelvic specimens for a total of 8 hip dissections. All specimens were examined with direct visualization and fluoroscopy prior to testing. There was no evidence of degenerative joint disease, prior trauma, or prior surgery. Using digital calipers and image analysis software, cross sectional measurements of the insertion of the psoas tendon at the lesser trochanter, the psoas tendon at the level of the labrum, and the psoas muscle tendon complex at the level of the labrum were performed. Additional measurements included the longitudinal distance of the psoas tendon at the level of the lesser trochanter to the tendon at the level of the labrum. Results: The overall length of the iliopsoas tendon from the lesser trochanter to the acetabular labrum was 75.4 ± 0.9 mm. The circumference of the Iliopsoas tendon at the lesser trochanter was 25.5 ± 2.6 mm, the iliopsoas tendon at the level of the labrum was 28.4 ± 2.8 mm, and iliopsoas tendon-muscle belly complex at the level of the labrum was 63.8 ± 7.4 mm. At the level of the labrum, the iliopsoas is composed of 44.5% tendon, and 55.5% muscle belly. The iliopsoas tendon in all specimens was located directly anterior to the antero-superior capuslo-labral complex. Conclusion: Knowledge of the cross sectional anatomy of the psoas tendon and the surrounding structures will better assist surgeons in treating pathology associated with psoas injury. The close anatomic relationship of the psoas tendon to the anterior capsulo-labral complex suggests that psoas pathology at this level may lead to labral injury. When indicated, surgical release of the iliopsoas tendon may be more effective at the level of the labrum as apposed to the lesser trochanter. Furthermore, this data suggest that 44% tendon-muscle belly complex should be released in order to release the tendinous portion. Author Contact: Michal Kozanek, Department of Orthopaedic Surgery, Phone: 617-724-3246, email: mkozanek@partners.org

31

Poster Board Number: 27 Computational Network Genetics for the Identification of Genes Involved in Congenital Anomalies

Kasper Lage1,2,3; Kristin Noonan1,2; Mauro Logoni1,2; Meaghan Russell1,2; Charles Lee2,5, 6; Mark J. Daly1,2,5; Vamsi K. Mootha1,2,5; Søren Brunak3; Barbara R. Pober1,2,4; Patricia K. Donahoe1,2,6

1 Massachusetts General Hospital, Boston, MA; 2 Harvard Medical School, Boston, MA; 3 Center for Biological Sequence Analysis, DTU, DK-2800 Lyngby, Denmark; 4 Children’s Hospital Boston, MA; 5 Broad Institute of MIT and Harvard, Cambridge, MA; 6 Brigham and Women’s Hospital, Boston, MA Purpose: In studying the genetics of many rare congenital disorders an important issue is the lack of large sets of patient cases for the study, due to reduced reproduction and survival in the patient groups. This suggests that strategies designed to enhancing statistical power drawn from relatively small patient samples will have to be implemented to identify the genes and pathways involved in this important subset of diseases. To address this problem we will work towards developing a number of approaches that combine heterogeneous data such as genomic and proteomic data for the identification of genes, and networks of genes, involved in congenital anomalies. We use the term network genetics to describe the integration of functional networks with patient derived genomic data. One of these strategies is the integration of copy number variation data (CNV) with protein - protein interaction (PPI) data. Combining such data types seems reasonable since proteins involved in congenital diaphragmatic hernia (CDH) have been hypothesized to be part of the same or related pathways (Kantarci, et al 2007, Kantarci et al., 2007, Yang et al., 2003), which can be analyzed using PPI data. Materials & Methods: Copy Number Variations: We used the Agilent 244 K oligonucleotide CGH Array Platform on 50 patients with CDH we have found a preliminary set of 68 CNVs, 47 in coding regions. These are currently being validated by quantitative PCR. Protein – protein interactions: We made a human protein interaction network by pooling human interaction data from the largest protein-protein databases, described in (Lage et al, 2007). A probability distribution for a network with given parameters is determined using random sampling and creating 10,000 random networks. This distribution is used to calculate a P value of the network, described in (Bergholdt et al., 2007; D'Hertog et al., 2007). Results: To test the hypothesis of pathway relatedness in CDH we created an interaction network based on the interactions of 51 proteins known to be involved in CDH which and subjected the network to statistical testing. This analysis shows that the CDH network is significantly enriched for interactions between the CDH proteins compared to random networks (P<1.0e-5), and indicates i) that CDH proteins are indeed part of one or several cellular pathways and ii) that a integration of PPI and genomic data could be a feasible way of identifying novel genes and pathways involved in CDH and perhaps other congenital anomalies. We then proceeded by combining PPI network and CNV data. Specifically, we identified the regions showing CNV in 50 patients with CDH and identified the proteins in these regions. Then we investigated whether these proteins could be placed in one or several significant PPI subnetworks which would suggest shared function. Hereby, we identify 5 significant subnetworks spanning 15 proteins from the CNV regions (adjusted p values of subnetworks are 0.0036, 0.005, 0.008, 0.011 and 0.021 respectively). Amongst these 15 proteins we identify two out of two (GATA4 and COUP TF2) known CDH related proteins that are in the CNV regions identified in patients with a CHD phenotype. Conclusion: Our results indicate that CDH proteins are part of one or several cellular pathways and suggest that network genetics are a powerful tool for identifying genes involved in congenital anomalies Author Contact: Kasper Lage, Pediatric Surgical Research Laboratories, MassGeneral Hospital for Children, Massachusetts General Hospital, Boston, MA. Tel: 617 378 2583. Email: klage@partners.org

32

Poster Board Number: 28 Preoperative Analysis Of Adjacent Segment Kinematics In A Patient With Idiopathic Adult Scoliosis: Application Of A Novel Imaging Technique By Gang Li, Michal Kozanek, Shaobai Wang, Peter Passias, Kirkham Wood and Guoan Li Purpose: Adjacent segment degeneration following spine fusion remains a widely acknowledged problem, but there remains insufficient knowledge regarding the factors that contribute to its occurrence. Little is known about the kinematic characteristics of adjacent segments prior to surgical correction. We investigated the applicability of a novel imaging technique to measure the six degrees of freedom (6DOF) kinematics of adjacent segments prior to spondylodesis and instrumentation in a patient with idiopathic adult scoliosis.

Materials & Methods: Kinematics of a 33 year old female with idiopathic scoliosis (major curve: T6-T12, 55° by the Cobb method) with planned arthrodesis and instrumentation (T4-L2) were measured preoperatively in the adjacent segments (L2-L5) using a combined dual fluoroscopic imaging system (DFIS) and computer tomography (CT). First, parallel axial images of the lumbar spine were obtained using computer assisted tomography (CT). Each image was processed using a Canny edge filter programmed in a commercially available software package (Matlab, Mathworks, Canton, MA). Subsequently, the patient was fluoroscopically imaged at varied postures representing extremes of the spinal range of motion in three rotational planes: 1) maximal left-right twist, 2) left-right bending and 3) flexion-extension of the upper body. The vertebral models were then oriented in a solid modeling software (Rhinoceros®, Robert McNeel & Associates, Seattle, WA) to match the fluoroscopic images in order to reproduce their in-vivo positions at each posture. The kinematics were then determined by the relative positions and orientations of the proximal vertebra with respect to the distal vertebra.

Results: The L2-3motion segment had the greatest contribution to motion in the sagittal plane (11.9°) during flexion and extension of the spine (Table 1). The L4-5 motion segment was responsible for the majority of the rotation in the coronal plane during right and left bending (16.3°). Range of motion in the transverse plane during right and left twist of upper body was relatively small (less then 2° in each segment) and fairly equally distributed among the studied segments in the patient. We also observed that coupled translations in all primary rotations. Specifically, rotation in the sagittal plane was coupled with anterior-posterior translation whereas rotation in the coronal plane was coupled with translations in the left-right direction. These pilot data suggest that the motion of the patient with idiopathic adult scoliosis was not equally distributed among the vertebral segments.

Conclusion: These data are the first attempt to measure the in vivo kinematics in a patients with adult scoliosis using the DFIS and CT imaging technique. The data indicates that this technique is a potentially powerful tool that can effectively be applied to study the effects of surgery on the kinematics of adjacent vertebral motion segments in patients with adult idiopathic scoliosis. Knowledge of kinematic changes before and after surgical correction in the vertebral segments adjacent to surgical arthrodesis is critical for understanding of the mechanism of adjacent vertebral degeneration and may ultimately help improve the surgical treatment of idiopathic adult scoliosis and provide insight into the etiology of post-operative degeneration of the adjacent segments. This novel imaging technique has great potential for scoliosis research.

Author Contact: Gang Li, Department of Orthopaedic Surgery, 617-726-1346, email: gli6@partners.org

33

Poster Board Number: 29 Characterization of Interactions Between the Vitamin D Receptor (VDR) and the Canonical Wnt Signaling Pathway Hilary F Luderer, Ph.D., Marie B Demay, M.D. Endocrine Unit, MGH/HMS Purpose: Ligand-independent functions of the Vitamin D Receptor (VDR) in keratinocyte stem cells are essential for normal hair follicle regeneration in mice. Previous studies demonstrated that the VDR is necessary for synergistic activation of Wnt reporter genes by LEF-1 and β-catenin in primary keratinocytes. Furthermore, when co-expressed in COS-7 cells, the VDR is present in a complex with β-catenin and LEF-1. While these data suggest that the VDR acts with β-catenin and LEF-1 to activate Wnt signaling in keratinocytes, the specific protein-protein interactions involved remain undefined. The purpose of this study was to identify and characterize the protein-protein interactions between the unliganded VDR and key effectors of the canonical Wnt signaling pathway. Materials & Methods: To determine if β-catenin and/or LEF-1 bind the VDR directly, a GST-fusion system was employed. An N-terminal GST-fusion of the wild type VDR was engineered in pGEX-5x-1 and expressed in bacteria. GST-VDR was bound to glutathione sepharose beads in the absence of ligand and then incubated with nuclear extracts isolated from COS-7 cells expressing endogenous β-catenin with or without LEF-1-HA. Proteins were eluted using reduced glutathione and subjected to Western analysis for detection of LEF-1-HA, endogenous β-catenin, and the VDR. Results: While LEF-1 co-eluted with the GST-VDR, β-catenin did not. These data suggest that the unliganded VDR interacts with LEF-1 independently of β-catenin. To test whether the interaction between the VDR and LEF-1, was indeed ligand independent, similar experiments were performed using GST-L233SVDR, a VDR mutant unable to bind ligand. The L233SVDR mutant retained the ability to bind to LEF-1, demonstrating that ligand is not necessary for this interaction. This is in contrast with previous investigations demonstrating that liganded VDR interacts with β-catenin. Conclusion: Based on the observation that the effects of the VDR on the hair follicle are ligand-independent, it is likely that a VDR-LEF1 interaction plays a significant role in maintaining hair follicle homeostasis. Additional studies will be directed at determining which domains of the VDR are required and whether other effectors of canonical Wnt signaling also contribute to this interaction. Author Contact: Hilary Luderer, Endocrine Unit, 726-3966, hluderer@partners.org

34

Poster Board Number: 30 Preventing growth of brain tumors by creating a zone of resistance Casey A Maguire1, Dimphna H Meijer1,2, Stanley G LeRoy1, Laryssa A Tierney1, Fabricio F Costa1,5, Xandra O Breakefield1, Anat Stemmer-Rachamimov3,4, Miguel Sena-Esteves1

1Department of Neurology, Massachusetts General Hospital, and Neuroscience Program, Harvard Medical School, Boston, Massachusetts, USA 2Department of Neuroscience and Pharmacology, Rudolf Magnus Institute of Neuroscience, UMC Utrecht, Utrecht, the Netherlands Purpose: Glioblastoma multiforme (GBM) is a devastating form of brain cancer for which there is no effective treatment. Here, we report on a novel approach to brain tumor therapy through genetic modification of normal brain cells with an adeno-associated virus (AAV) vector to block tumor growth and effect tumor regression. Previous studies have focused on the use of vector-based gene therapy for GBM by direct intratumoral injection with expression of therapeutic proteins by tumor cells themselves. However since anti-tumor proteins are generally lethal to tumor cells, the therapeutic reservoir is rapidly depleted, allowing escape of residual tumor cells. Moreover, it has been difficult to achieve consistent transduction of these highly heterogeneous tumors using viral vectors. Materials & Methods: Our approach utilizes the innate ability of AAV to efficiently transduce normal brain cells with a transgene encoding an anti-tumor protein, interferon beta (IFN-β) with the purpose of setting up a barrier to tumor growth and spread. We utilized an AAV vector encoding human IFN-β to transduce brain parenchyma to evaluate whether this interfered with growth of orthotopic human glioma tumors in a nude mouse. Results: In this study, we tested whether creating a “zone of resistance” in the brain by AAV-mediated expression of human interferon-β (hIFN-β) in normal brain tissue could prevent the growth of tumors in orthotopic xenograft models of GBM. First, in a proof of concept experiment, we found that pretreating normal brain with an AAV vector encoding hIFN-β could effectively prevent the establishment of intracranial tumors using two different human glioma cell lines, in contrast to mice pretreated with a control AAV vector. In fact, we observed a robust anti-tumor effect against tumors located distally to the vector injection site. Second we demonstrated, via neuronal-restricted expression of IFN-β, that complete regression could be achieved without the need for transduction of tumor cells. Conclusion: Our results represent a new paradigm for GBM therapy, as well as other malignancies, based on AAV-mediated genetic engineering of normal tissue to manipulate the tumor microenvironment. Furthermore, the efficiency of tumor regression in the brain (obtained using one injection) is currently unrivaled with existing preclinical GBM-therapy strategies using AAV vectors. This therapeutic approach could be translated into clinical trials by multiple injections of vector into the surrounding brain parenchyma at the time of tumor resection to create a zone of resistance to tumor recurrence. Author Contact: Casey Maguire, Department of Neurology, Phone:585-943-9201, cmaguire@partners.org

35

Poster Board Number: 31 Nanotechnology Platforms in Targeted Photodynamic Therapy and Imaging of Ovarian Cancer Zhiming Mai*, Benjamin A. Teply† and Tayyaba Hasan* * Wellman Center for Photomedicine, Massachusetts General Hospital, Harvard Medical School. Boston, MA, USA; and †Department of Chemical Engineering, Harvard–MIT Center of Cancer Nanotechnology Excellence, Massachusetts Institute of Technology, Cambridge, MA, USA Purpose: Poor survival rates associated with advanced ovarian cancer disseminated in the abdominal cavity as well as resistance to standard chemotherapeutic drugs and radiation necessitate the need for a drastic change of approach to the disease management. The overall goal of this study is to develop nanoparticle platforms utilizing targeted photodynamic therapy and imaging modalities to achieve high selectivity combined with the most effectiveness against disseminated chemo- and/or radiation-resistant ovarian cancer, as well as to perform early diagnosis and monitor treatment outcomes. Materials & Methods: Figure 1 cartoons the concept of a nanoparticle platform for targeted photodynamic therapy. Targeted nanoparticles synthesized here are multicomponent structures with a

carrier system that forms the core and contains the therapeutic or imaging payload, surface modifiers to reduce reticuloendothelial system uptake and enhance biodistribution, and a targeting component. Polymers used in the preparation of nanoparticles were poly (lactic-co-glycolide)-poly(ethylene glycol) (PLGA-PEG), biodegradable and FDA-approved drug carriers. The photosensitizer-bearing naoparticles

encapsulation of the chlorin e6 monoethylene diamine disodium

salt (CMA) into PLGA-PEG. Finally, the targeting CMA-NP was obtained through conjugation of CMA-NP with targeting ligands called aptamers for EGFR-family receptors overexpressed on ovarian cancer cells. Using such nanoparticle platforms, we have initiated photosensitizer uptake, photodestruction, and imaging assays to characterize their multifunctional capacities in management of ovarian cancer. Results: The synthesized nanoparticles wer

(CMA-NP) were made by

e in monodisperse with an average size of 202.0 nm and

rticle platform is the first using nanobiotechnology and

negative surface charge at -30 mV. The payload of photosensitizers in nanoparticles was at 0.41 mM or 0.279 mg/ml of equivalent CMA. Results of photosensitizer uptake measurement suggested that lower CMA-NP accumulation in normal tissues/cells was due to their overall negative charge which mediates the repulsion between the nanoparticles and cellular membranes innately bearing negative charge. This feature comprises one of the aspects in selection, thus minimizing side effects of drug toxicity to normal tissue/cells. Confocal fluorescence microscope imaging indicated that specific ligand-bearing CMA-NP had a high selectivity for OVCAR-5 and OVCAR-3 ovarian cancer cells as compared with control cells, consisting of another aspect in selection. Also, the fluorescent property of CMA in the nanoparticle complex formed the basis of imaging analysis for both diagnosis and treatment monitoring. Photodestruction studies showed that the photosensitizers in nanoparticle core have a higher potency in cancer cell photodynamic killing than free CMA. Conclusion: To our knowledge, this novel nanopaaptamer-based targeting PDT and imaging. The nanoparticle complex facilitates the delivery of therapeutic photosensitizers, and will provide a model for a highly selective, effective, and predicable therapy, an approach that can be adapted for other targets as reagents become available. Author Contact: Zhiming Mai; Dermatology; 617-726-6995; zmai@partners.org

36

Poster Board Number: 32 V-ATPase Interacts with Regulatory Elements of Arf-Family Small GTPases Activator: Mapping the Interaction Sites between the a2-isoform and ARNO Maria Merkulova and Vladimir Marshansky Program in Membrane Biology, Center for Systems Biology and Division of Nephrology, Simches Research Center, Massachusetts General Hospital and Harvard Medical School, Boston, MA, 02114,USA Purpose: Trafficking within the endosomal/lysosomal protein degradative pathway plays an important role in protein homeostasis and depends on V-ATPase-driven endosomal acidification. In kidney proximal tubule epithelial cells, the V-ATPase a2-isoform is specifically targeted to early endosomes and directly interacts with ARNO (which is a GTP/GDP exchange factor and an activator of Arf GTPases) in an acidification-dependent manner. This acidification-dependent interaction between V-ATPase and small GTPases is crucial for trafficking between early and late endosomes. We propose that the V-ATPase a2-isoform functions as an endosomal pH-sensor and regulates the endosomal/lysosomal pathway by scaffolding and recruiting small GTPases. However, the molecular mechanism of this novel function of the a2-isoform and its interaction with ARNO remains obscure. Materials & Methods: GST and peptide pull-down assays were used to map interaction sites between the V-ATPase a2-isoform and ARNO. 22 peptides covering the complete sequence of the cytosolic N-terminal tail of the a2-isoform (a2N), and 3 peptides corresponding to Sec7/PH-linker and unphosphorylated or phosphorylated (at serine392) polybasic (PB) C-terminal regions of ARNO were synthesized and purified in MGH Peptide/Protein Core Facility. Cloning, expression and purification of GST-tagged ARNO constructs and a2N were performed by conventional methods. Results: We have mapped the interaction sites between the a2-isoform and ARNO. First, peptide pull-down assays demonstrated that ARNO strongly interacts with one (a2N-03) and weakly interacts with five (a2N-01, a2N-11, a2N-12, a2N-18 and a2N-22) peptides within the a2N. Secondly, using recombinant constructs containing various ARNO domains we demonstrated that a2N specifically interacts with both Sec7 and plekstrin-homology (PH) domains of ARNO. Thirdly, another peptide pull-down assay showed that a2N directly interacts with two important regulatory elements of ARNO: the Sec7/PH-linker region and the PB/C-terminus. Importantly, interaction of the a2N with the PB/C-terminus was dependent on the phosphorylation state of serine392 and, therefore, could be regulated by protein-kinase C activity. Thus, we have shown that the cytosolic N-terminus of the V-ATPase a2-isoform contains multiple high and low affinity sites of interaction with ARNO. We propose that some of these sites interact with regulatory regions on ARNO and suggest an involvement of the V-ATPase in modulating ARNO activity. Conclusion: Our findings have revealed a complex and multiple-site interaction of V-ATPase with ARNO during its pH-dependent recruitment to endosomal membranes. Based on our data, we suggest the following two-step mechanism of a2-isoform interaction with ARNO during which some conformational changes take place. At neutral/mild acidic intra-endosomal pH, the a2-subunit of the V-ATPase is in an inactive conformation with possible weak association (most likely through peptide a2N-22) with the PH domain of auto-inhibited ARNO. Upon intra-endosomal acidification, the a2-isoform undergoes a conformational change and adopts an active conformation, which gives rise to tight binding to ARNO in the following manner: N-terminally located peptides a2N-01, -03, -11 and -12 together with peptide a2N-18 interact with the Sec7 domain and regulatory elements (Sec7/PH-linker and PB/C-terminus) of ARNO leading to activation of the auto-inhibited ARNO. Thus, our data raise the possibility that the V-ATPase may not just recruit and scaffold small GTPases to their target membranes as part of its pH sensing function, but also could have a novel role as a possible activator of small GTPase-regulating proteins, such as ARNO, during this process. Author Contact: Maria Merkulova; Program in Membrane Biology, Center for Systems Biology and Division of Nephrology; 617-724-0063, mmerkulova@partners.org

37

Poster Board Number: 33 Quantitative Assessment of Dermal Fibroblasts in Nephrogenic Systemic Fibrosis Rosalynn Nazarian, Rajni Mandal, Anna Kagan, Jonathan Kay, Lyn Duncan

Unit and Department of Medicine

nction who have received gadolinium-containing radiographic contrast. Skin biopsies from patients

.

a

-

ver

n:

Department of Pathology, Dermatopathology Purpose: Nephrogenic systemic fibrosis (NSF) is a disorder that affects patients with impaired renal fuwith NSF most commonly reveal the following histopathologic findings: fibroblast cell proliferation, increased mucin deposition, and thickened collagen bundles with adjacent clefts within the dermis The current study has the following aims: 1. Does the dermal fibroblast cell count in skin biopsies from patients diagnosed with NSF differ significantly from that of non-NSF patients? 2. Do the cutaneous histopathological findings correlate with duration of skin lesions, level of plasmcreatinine, and symptom-preceding dose or cumulative dose of gadolinium? Materials & Methods: Retrieved from the MGH dermatopathology files were 17 histologically confirmed cases of NSF and 17 non-NSF controls. Controls were matched for age (mean: 63, range: 3679 years), sex (11 male, 6 female), and site (forearm, thigh, and lower leg skin biopsies). Five high-power fields (hpf) using a 10x ocular and 40x objective magnification (240 mm2 round field area) from H&E stained 5mm thick sections of formalin-fixed paraffin-embedded skin biopsies were digitally

phed and printed. The total number of dermal fibroblast nuclei (raw number of spindphotogra le-shaped cells per hpf) was then manually quantified by three independent observers blinded to the diagnosis andthen averaged over 5 hpf per patient. Patient records were reviewed to obtain demographic and clinical data, specifically, plasma creatinine, gadolinium dose, and date of last gadolinium exposure. Results: The dermal fibroblast cell count was significantly greater for all NSF cases (mean: 68.87/hpf, SD: 18.24) in comparison to negative controls (mean: 14.03/hpf, SD: 8.81). There was high inter-obseragreement for cases and controls with Pearson correlation coefficients of 0.945 (p<0.01) and 0.726 (p<0.01), respectively. No significant correlation between dermal fibroblast cell count and time from last gadolinium exposure (mean: 358, range: 3-959 days), symptom-preceding dose (mean: 73, range: 20-190 cc Magnevist®) orcumulative gadolinium dose (mean: 109, range: 40-220 cc Magnevist®), and plasma creatinine (mea5.8, range: 3.5-9.4 mg/dL) at time of skin biopsy was identified. Conclusion: This is the first study to quantitatively assess the extent of dermal fibroblast proliferation inNSF. The finding of a mean dermal fibroblast cell count greater than 50 per hpf is consistent with NSF and may be utilized to identify cases at the time of diagnosis. Our results substantiate previous reports ofincreased dermal fibroblast number as a diagnostic histologic feature in skin biopsies of patients with NSF with high interobserver correlation. The extent of dermal fibroblast proliferation does not correlate with increased age of skin lesion, dose of gadolinium administered, or plasma creatinine at the time of skin biopsy. Author Contact: Rosalynn M. Nazarian, Department of Pathology, 617-724-1407, rmnazarian@partners.org

38

Poster Board Number: 34 Fast Fourier Fluorescence Lifetime Excitation and Emission Spectrometer Leilei Peng, Brett E. Bouma, and Guillermo J. Tearney Harvard Medical School and the Wellman Center for Photomedicine, Massachusetts General

40 Blossom St, BAR 703, Boston, Massachusetts 02114

s a nce

ission tion.

herent

rophores, and conducted

eiLei Peng, Wellman Center for Photomedicine, 617-643-2346,

Hospital, Purpose: Full characterization of a sample’s fluorescence requires measuring its spectral intensity atwo-dimensional function of both excitation and emission wavelength. We report a fluorescespectrometer that utilizes principles of Fourier transform spectroscopy to measure the excitation emmatrices (EEM) rapidly and with high spectral resolu Materials & Methods: For this intensity and lifetime EEM spectrometer (Figure 1), incoexcitation light is first input into a path-length scanning Michelson interferometer. Light from the output port excites sample fluorescence. The fluorescence remitted from the sample is directed a

cond Michelson interferometer, whose path-length scanning is synchronized with the first seinterferometer. The first interferometer produces spectrally encoded modulation on the fluorescence excitation source. The second interferometer resolves the emission wavelength via Fourier transformation. By synchronizing two scanning interferometers, EEM can be obtained by two-dimensional Fourier analysis.

Results: We measured lifetime and intensity EEM of mixtures of different fluo

Fig. 1 Fourier fluorescence lifetime excitation emission spectrometer

preliminary study of fluorophores undergoing resonant energy transfer by the Fourier Fluorometer. Results obtained corresponds well with results measured by traditional methods. Conclusion: we have demonstrated a Fourier spectrometer for fast fluorescence excitation emission matrix measurement, which has a broad spectral range, variable spectral resolution andrelatively low equipment cost. Author Contact: Llpeng1@partners. org

39

Poster Board Number: 35 Longitudinal cellular imaging of murine internal organs by in vivo endomic

2 3roscopy

t , Hiroshi Yamashita , Dai Fukumura3,

nd Harvard Medical School

as a powerful tool for visualizing the cellular

e into a 15-gauge tube that was designed to hold the tissue locally pled the endoscope into the real-time video-rate confocal and

multiphoton microscope system we have developed previously. The image resolution was 1 μm in transverse and 15 μm in depth. To measure the kinetics of antigen presenting cells trafficking (from donor or recipient origin), hearts were harvested from MHC class-II:GFP+ mice and transplanted into the abdomen of wildtype and aly/aly mice, and the heart, spleen, and draining lymph nodes were imaged by laparotomy from day 0 till several weeks after transplantation. To study the angiogenesis in colorectal cancer, green fluorescent protein (GFP) expressing human colon adenocarcinoma cells were surgically implanted to the wall of colon in athymic nude mice, and the tumor and vasculature were imaged by rotational-pullback colon-microscopy at post-operative day 6, 14, and 34, Results: Initially, right after heart transplantation, a large number of MHC class-II+ cells were observed in the heart. After 3 days, in a wildtype recipient mouse, 93% of these cells in heart were no longer present in the graft, and donor-derived MHC class II+ cells were observed in various lymph nodes and spleen. In contrast, in of donor MHC mber of MHC-class II+ cells as the wildtype recipient was observed in the spleen. At 2 weeks post operation, we found

observed the increase in

a novel microendoscopy system for longitudinal in vivo imaging of arious viscera organs in mice. Using home-built endoscopes, we have demonstrated the first cellular

aging in a beating heart, quantitative immune cell trafficking following organ transplantation, and prove useful

edical investigations using experimental animal models. Author Contact: Pilhan Kim, Wellman Center for Photomedicine, 617-643-2908, pkim5@partners.org

Pilhan Kim1, Georges Tocco2, Euiheon Chung3, Cavit D. KanRakesh K. Jain3, Gilles Benichou2, and Seok H. (Andy) Yun1

1 Wellman Center for Photomedicine, Massachusetts General Hospital a2 Surgery / Transplantation Unit, Massachusetts General Hospital 3 Edwin L. Steele Lab., Department of Radiation Oncology, Massachusetts General Hospital and Harvard Medical School

urpose: Intravital fluorescence microscopy has emergedPprocesses within small animals. However, due to the large size of standard microscope lenses, imaging the internal organs of small animals, such as mice, remains a challenge, requiring invasive procedures. The first specific aim of our research is to develop a miniature confocal and multiphoton endoscopic imaging system. The second specific aim is to apply this novel technique to visualize and quantify immune cell trafficking in chronic heart rejection and angiogenesis in orthotopic colorectal tumors, following the same animals over weeks. Materials & Methods: Using gradient-index (GRIN) lenses, we fabricated front- and side-viewing types of imaging probes that are 1 mm in diameter and 15 to 40 mm in length. To avoid motion-induced image blurring, we integrated the probby gentle vacuum pressure. We then cou

aly/aly recipient mice (lacking lymph nodes) after 1 week, we observed that 96% class II+ cells are still present in the heart graft, and interestingly a similar nu

10 times increased number of MHC class II+ cell in the spleen of same aly/aly recipient mouse, while 92% of donor MHC class II+ cells are still present in the grafted heart, suggesting the antigen presentation in the spleen of aly/aly recipient mouse. In the colon cancer xenograft model, we

lanted tumor with time. Supplying peri-tumor vessels for tumor grosize of the imp wth and the elevated leakage at tumor sites were visualized at the boundary of tumor. Dilated and distorted vessels compared to normal vessels in healthy region were observed at tumor sites. Conclusion: We have developedvimwide-area tumor imaging in the colon mucosa. We expect in vivo mouse endomicroscopy toin a variety of other biom

40

Poster Board Number: 36 The Role of Desmosomes in the Regulation of Cardiomyocyte Differentiation Eva Plovie1, Shannon Coy1, Daniela Panáková1, Andreas Werdich1, Patrick Ellinor1 and Calum MacRae1

1Cardiovascular Research Center, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA

both nd function have been observed in many

t mutations in genes encoding proteins within the desmosome, one of the three major nctional complexes within the ICD, cause an inherited form of human heart muscle disease known as

e

me luding

the h n

a2+

yocyte

ore the

d ts

n acellular dense midline of

ect

02129, USA Purpose: Cardiac myocytes rely on specialized structures known as intercalated discs (ICD) formechanical and electrical coupling. Perturbations of ICD structure aforms of heart disease, but until recently were thought to be secondary phenomena. We, and others, have demonstrated thajuArrhythmogenic Right Ventricular Cardiomyopathy (ARVC), characterized by a dystrophy of cardiac musclewith loss of myocyte differentiation, so called fatty replacement of myocardium, arrhythmias and contractilfailure. Preliminary work has raised the possibility of Wnt/beta-catenin signaling as a potential disease mechanism, but the precise pathogenesis remains unclear, largely as a result of the lack of model organisms for efficient pathway analysis. We have developed a zebrafish model in which perturbations of the demsosocan be studied in an intact myocardial syncytium in the presence of all of the other native cues, incmechanical and electrical signals. Material and Methods: We first cloned, and knocked down with antisense morpholino oligos, each of major components of the desmosomal protein complex in the zebrafish. The use of the embryonic zebrafisallowed us to study the effects of these manipulations from the earliest stages of cardiomyocyte specificatiothrough initiation of sarcomere assembly and on to definitive cardiomyocyte differentiation. Using electronmicroscopy, immuno-histochemistry, high resolution physiologic assays for the zebrafish, cellular Cimaging and optical voltage mapping we defined the effects of desmosomal perturbations on cardiomdifferentiation and coupling. Using zebrafish mutant in the Wnt/beta-catenin pathway and in the non-canonical Wnt pathway, as well as additional morpholinos or small molecule inhibitors, we have begun to explpathways downstream of abnormal desmosomal structure and function. Results: We identified the gene encoding the desmosomal protein desmocollin in the zebrafish and confirmeits expression in the heart using whole-mount in situ hybridization. Morpholino (MO) knockdown experimendemonstrated a dose-dependent bradycardia, abnormal contractility and progressive cardiac edema. Electromicroscopy (EM) of the desmocollin morphant zebrafish revealed that the extrnormal desmosomes was absent in the myocardial tissue of the ventricle. We were able to rescue the morpholino knockdown phenotype with human wild-type desmocollin mRNA but not with a mutant human mRNA, confirming the specificity of the knockdown itself, but also suggesting that it will be feasible to dissthe correlation between desmosomal structure and function in vivo. Activating the Wnt/beta-catenin pathway with the small molecule GSK inhibitor BIO we were unable to rescue the cardiac defects observed with desmocolling knockdown, indeed we observed synergistic interaction with the desmocollin morpholinos. These data suggest that if there is downregulation of the Wnt/beta-catenin pathway in ARVC, it occurs downstream of beta-catenin itself or through other pathways. Using the wnt5 and wnt11 morphants, we have observed subtle but significant synergy between wnt5 and desmosomal gene knockdown for both cardiac and extracardiac phenotypes, However, marked synergies between Wnt11 null alleles and desmosomal knockdown, strongly suggest that the major effect of desmosomal mutation in myocardial differentiation is mediated through the planar cell polarity pathway.

Conclusion: Taken together these data establish the zebrafish as a model system for the efficient study intercellular junction structure and function in the context of all of the native cues. These results implicate only the canonical, but also non-canonical Wnt signaling, in particular the planar cell polarity pathwaymyocardial morphogenesis and differentiation. Author Contact: Eva Plovie Buys, Cardiovascular Research Center, 617-726-6471,

of not in

lovie@partners.orgep

41

Poster Board Number: 37 Thermal bystander effect: Medium-mediated induction of DNA double strand breaks and

, 3 %

es

2 fold cells

e

res

xtent.

Author Contact: Martin Purschke, Wellman Center for Photomedicine, 617-726-1591, mpurschke@partners.org

apoptosis after thermal stress via intercellular communication Martin Purschke, Hans J. Laubach and Dieter Manstein Wellman Center for Photomedicine, Massachusetts General Hospital, Harvard Medical School. Boston, Massachusetts, USA Purpose: Thermal exposure causes protein degradation and DNA damage, which can leads to genetic alteration and cell death. In comparison, little is known about the surrounding non-heated tissue. So far, the loss of viability in non-heated cells is explained due to heat diffusion. Our results demonstrate, for the first time that in the absence of temperature diffusion surrounding non-heated cells can also develop DNA damage and loss of viability. This so called “thermal bystander effect” might be of importance for clinical applications. Materials & Methods: We used human fibroblasts cells (HFF1), grown in a transwell insert system, where heated 37-65 C) and non-heated cells shared the same medium, but had no cell-to-cell contact and no heat diffusion. As a viability assay we used the MTT assay. Cells were heated, co-incubated for 24 h, incubated with MTT for 2 h and solved in DMSO for 30 min. Absorption was measured at 570 nm by using an ELISA plate reader. DNA damage and apoptosis was analyzed morphologically after fluorescence staining. Cells were heated, co-incubated for 72 h, fixed with Methanol/Acetic acid (3:1) and stained with DAPI. 500 cells for each sample were analyzed. Results: The cell viability of heated fibroblast cells 24 h after heat exposure, decreased significantly at temperatures above 40°C and slowed down at 50°C. We also observed a significant increase of DNA damage and apoptosis in heated cells. Interestingly, non-heated cells, which shared the same medium with the heated cells, also showed, although less extensivea loss in viability and an increased DNA damage/apoptosis. The viability decreased up to 1starting at a temperature of 40 °C, and came back to control levels for higher temperatur(55°C). In addition, we observed approximately a two-fold increase in the level of DNA damage in both heated and non-heated “bystander” cells. The induction of apoptosis in heated cells was much more pronounced (up to 8 times compared to control) than in bystander cells (about compared to control). At higher temperatures greater than 50°C, the effect on the bystander is similar to the control cells, which are kept at 37°C. Conclusion: Our data indicate the first time that both heat exposed and bystander cells, whilsharing the same medium, can generate DNA damage and undergo apoptosis, which results inloss of viability. The disappearance of this thermal bystander effect at higher temperatuindicates that the heat exposed cells need to be viable to produce the signal to damage thesurrounding cells. Nothing is known so far about the time course, how long does the bystander signal get produced and which molecules (ROS, or protein) are responsible for the active thermal bystander effect. Therefore the bystander effect needs further investigation to determine, the involved mechanisms, whether this effect is also applicable to other cell lines and if so, to what e

42

Poster Board Number: 38 High-Throughput Chemical Screening Implicates the Wnt/β-Catenin Pathway in the Renewal And Differentiation Of Isl1+ Cardiovascular Progenitors Yibing Qyang,1, 7 Martin-Puig, S.,1, 7 Chiravuri M.,1, 7 Chen S.,2 Xu, H.,1 Bu L.,1 Jiang X.,1 Lin, L.,3

Laugwitz, K.,1 Moon, R.,6 Gruber, P.,4

arch Center, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114-2790, 2 Department of Chemistry, The Scripps Research Institute, La Jolla, CA 92037,

Granger, A., Moretti, A., Caron, L., Wu, X., Taketo, M.,4 1 1 2 5

Evans, E.,3 Ding, S.,2 and Kenneth R. Chien1

1 Cardiovascular Rese

3 Skaggs School of Pharmacy, UCSD, 9500 Gilman Drive, La Jolla, CA 92093, 4 Childrens Hospital of Pennsylvania, University of Pennsylvania, Philadelphia, PA 19104, 5 Department of Pharmocology, School of Medicine, Kyoto University, Kyoto 606-8501, Japan, 6 HHMI, and Department of Pharmacology, University of Washington, Seattle, WA 98109, 7 authors contributed equally

Purpose: The recent discovery of cardiovascular progenitor cells (CPC) in adult hearts has opened a

im

study is the first report utilizing a high-throughput chemical

proliferation of the OFT myocardial cells is

new exciting avenue for research into cardiac repair mechanisms. CPCs marked by Isl1, a LIM-homeodomain transcription factor, have the potential to form all cardiovascular cell types in the heart. Our laboratory has recently discovered that the cardiac mesenchymal cells are able to provide cues renewing isl1+ cardiovascular progenitors, but mechanisms governing their self-renewal and differentiation remain largely unknown. Through a novel high-throughput small molecule screen, we ato identify signaling pathways modulating the renewal and differentiation of isl1+ cardiovascular progenitor cells. Materials & Methods: Ourscreening strategy to identify biological activities modulating the diverse functions of cardiovascular progenitor cells. Additionally, murine embryonic stem cells were utilized to establish an assay to isolate these cardiovascular progenitor cells at large scale for mechanistic studies. Furthermore, gain- and loss-of-function of murine models were investigated to demonstrate the in vivo relevance of our stem cell-based in vitro studies. Results: Through a high-throughput small molecule screen coupled with murine embryonic stem cell studies, we have discovered that the Wnt/β-catenin pathway promotes the renewal while negating the pre-specification and differentiation of isl1+ cardiovascular progenitors. In vivo activation of β-catenin in isl1+ second heart field progenitors results in their marked expansion, inhibition of differentiation, and abnormal outflow tract (OFT) morphology. Furthermore, thesignificantly reduced in murine embryos with a loss-of-function of β-catenin in isl1+ precursors. In addition, a Wnt/β-catenin agonist is able to markedly expand isl1+ cardiovascular progenitors from human neonatal hearts. Conclusion: The current study suggests that canonical Wnt signals are a major component of the cardiac mesenchymal microenvironment that controls the renewal and differentiation of isl1+

cardiovascular progenitors. More importantly, it implicates that the manipulation of Wnt signals might have a direct effect on the isolation, cloning and expansion of rare human isl1+ cardiovascular progenitors from either ES cells or intact human heart tissue Author Contact: Yibing Qyang; Cardiovascular Research Center; 617-412-5000; yqyang@partners.org

43

Poster Board Number: 39 The role of common mitochondrial genetic variants in stroke and radiographic white matter diseasePurpose: Stroke is the 3

hite

n

McaS ation.

riates in model. A p value of ≤5x10-4 was considered significant.

plex

a e SNPs

WMD and stroke. We

rd leading cause of death worldwide but has few proven preventative and acute treatments. Novel genetic risk factors for stroke may help to elucidate new therapies. Radiographic wmatter disease (WMD) is a potent risk factor for stroke and the most prevalent form of cerebrovascular disease. Rare variants in mitochondrial DNA (mtDNA) are associated with the presence of WMD and aincreased risk of stroke in mitochondrial encephalomyopathy, lactic acidosis and stroke-like episodes patients. Our study objective was to determine whether common mtDNA variants may contribute to severity of WMD and increased risk of stroke in the general population.

aterials & Methods: We selected 64 tagging single-nucleotide polymorphisms (SNPs) that efficiently pture all common mtDNA variation and an additional 16 haplogroup (phylogenetic branch) defining

NPs in order to determine each subject’s maternal ancestry and account for population stratificaplogroup assignments were derived from Hernstadt and colleagues detailed mapping of the H

mitochondrial tree (Figure). Genotyping was performed by primer extension of multiplex products with detection by MALDI-TOF mass spectroscopy. Subjects included 1149 individuals with symptomatic stroke and 749 stroke-free controls, prospectively recruited through the MGH inpatient and outpatient wards (03/1997 to 07/2007). SNP association with WMD was assessed in a 703-subject subset of caseswhose WMD burden was quantified on magnetic resonance imaging. The case/control analysis was stratified by hemorrhagic (489) and ischemic stroke (660). The logistic regression analysis for the case/control study was performed using a Cochran-Mantel-Haenszel (CMH) stratified test of association by haplogroup. For analysis of WMD, a continuous trait, haplogroup and age were treated as covaa linear regressionResults: Preliminary results identified multiple mtDNA SNPs in subunits 1 and 4 of the gene for com1, which influenced risk of stroke and WMD. No SNP reached genome-wide significance in our case/control analysis due to our low control to case ratio. To test this hypothesis we performedcase/control analysis that included an additional 4,000 subjects recruited for other studies. The samwere significantly associated with stroke (<1x10-6 or -7). These exploratory results suggest that the associations we found are robust. Conclusion: We found several mtDNA SNPs to be suggestively associated with

e currently performing identical genetic analyses in a larger case-control cohort in order to confirmarthese findings.

-1

00 0 20 40 60 80

-5

-4

-3

-2

1

2

3

4

5

6

-120 -100 -80 -60 -40 -2

AsianAfricanH1H2+IJKpreHV-VTUWX

igure. Clustering of Subjects by mtDNA Haplogroup. Subjects were F assigned into European (H1, H2+, I, J, K, nstrated that our rrected for with

by haplogroup.

preHV-V, T, U, WX), Asian or African haplogroups by their genotype. Clustering of subjects demopopulation was non-homogenous despite >95% of subjects self-reporting as Caucasian. This was cothe CMH-stratified analysis

44

Author Contact: Rosanna Rahman; Neurology; 617-643-2469; rahman@chgr.mgh.harvard.edu

45

Poster Board Number: 40 Imaging Drug Penetration and Metabolic Activity in 3D Ovarian Cancer Models Imran Rizvi1,2, Conor L. Evans1, Daina Burnes1, Jonathan Celli1, Johannes de Boer1 and Tayyaba Hasan1

1Wellman Center for Photomedicine, Massachusetts General Hospital, Boston, MA 02114 2Thayer School of Engineering, Dartmouth College, Hanover, NH 03755 Purpose: Our lab has published promising pre-clinical results using photodynamic therapy (PDT), an emerging light-based modality, in combination with Erbitux®, an antibody targeted against the epidermal growth factor, to synergistically treat advanced ovarian carcinomatosis. As this therapeutic strategy moves into clinical trials, we are investigating the mechanisms underlying the observed synergism to optimize this combination regimen. The purpose of the current study is to characterize heterogeneities in drug penetration and metabolic activity using 3D ovarian cancer models, and optical imaging technologies, to elucidate the mechanisms of synergistic treatment response.

Materials & Methods: Based on breast cancer models described by Drs. Mina Bissell and Joan Brugge, we have developed 3D cultures for ovarian carcinoma as our in vitro research platform to conduct these mechanistic studies. Ovarian cancer cells seeded on Growth Factor Reduced (GFR) Matrigel™ beds spontaneously form 3D acinar structures that more closely resemble in vivo tumor nodules than cells in monolayer. We are using confocal microscopy of benzoporphrin derivative (BPD) and Erbitux® along with multiphoton microscopy of intrinsic fluorophores to nonpurtubatively investigate the patterns of drug penetration and metabolic activity in these acinar structures.

Results: As with other treatments, preliminary experiments suggest that the 3D acini are less sensitive to PDT than cells in monolayer. This differential treatment response could be attributed to heterogeneity in drug penetration as well differences in signaling and architectural cues between the two systems. In this study, we have shown that BPD and Erbitux® are unevenly distributed in 3D acinar structures (Figure 1). Furthermore, an unexpected distribution pattern of metabolic and apoptotic activity is observed in the ovarian cancer acini following multiphoton microscopy of intrinsic fluorophores. Conclusion: Development of novel research platforms to complement, or replace, traditional model systems could significantly impact our ability to conduct reliable mechanistic studies and predict the efficacy of therapeutic strategies. Our current research demonstrates heterogeneity in drug distribution

and metabolic activity in 3D ovarian cancer acini, which could have implications for optimizing PDT + Erbitux®-based combination treatment of ovarian carcinomatosis. Figure 1 Confocal and multiphoton fluorescence images of 3D acinar cultures of ovarian cancer: A) BPD and B) Erbitux distributions; C) Intrinsic fluorescence excited at 720nm; D) Volumetric rendering of BPD and intrinsic fluorescence. Figures C and D are presented in false-color intensity scale. The scan volume for Figure D is 110µm x 105µm x 100µm.

Author Contact: Imran Rizvi,Wellman Center for Photomedicine, 617-726-3991 IRIZVI@PARTNERS.ORG

A B C D

46

Poster Board Number: 41 Dominant-negative inhibition of Ets 1 suppresses tumor growth, invasion and migration in rat C6

, Virginie Nicolas Wernert1

Pathology, University of Bonn, 53011 Bonn, Germany. ce.

ery,

ption factor the prototype of the Ets

k

well of a

ion, mparing

)

assays were performed ough Ets-

es

and the anchorage-independent growth up to 80.2%. s reduced by 49% in fibronectin, and 19% ype I coated Boyden chambers. Invasion was inhibited by 51%, and DB-tumors were significantly smaller (84.5%) than Neo and C6 control tumors. Furthermore, 4 out of 10 Ets-DB expressing cells were unable to develop a tumor. The expression of Ets-DB further showed decrease in mRNA transcripts of mmp-9 (one of Ets-1 target genes) by 87%. Ets-DB expression led to an up-regulation of several genes which are involved in cell metabolism, such as eef1a1, eef1b, serpine2, sparc, txn, atp5b, grp58, tpt1, enpp1, ldha, nne and lrp. Our data bank analyzes showed that these genes contain binding sites for Ets 1, and for the Ets family members Elf-1 and Fli-1. Our RT-PCR results

,

ul

@mgh.harvard.edu

glioma cells and reveals differentially expressed Ets 1 target genes Ayguen Sahin1, 5, Chantal Vercamer2, Annette Kaminski1, Tanja Fuchs1, Jens Claus Hahne1

Mattot3, Albin Pourtier-Manzanedo2, Torsten Pietsch4, Véronique Fafeur2 and 1 Institute of 2 CNRS UMR 8117, Institut de Biologie de Lille, Institute Pasteur de Lille, B.P. 447, 59021 Lille, Fran3 CNRS UMR 8161, Universite de Lille II, Institut Pasteur de Lille, 1 rue Calmette, 59021 Lille Cedex, France. 4 Department of Neuropathology, University of Bonn, 53105 Bonn, Germany. 5 Current address: MGH-HMS Center for Nervous System Repair, Department of NeurosurgMassachusetts General Hospital, Harvard Medical School, Boston, MA 02114, USA.

Purpose: Our research is focused on the molecular and cellular mechanisms of the transcriEts 1 in the deadly disease Glioblastoma multiforme (GBM). Ets 1 istranscription factor family and, as previously described, participates in tumor invasion involving both thetumor stroma and the neoplastic cells. Expression of Ets 1 correlates with a poor prognosis and a high risfor lymph node metastases in several cancers. However, the role of Ets 1 in gliomas has not been described. Here, we evaluated a blockade of Ets 1 in rat C6 glioma cells through a retroviral expression the Ets 1 DNA-binding domain (Ets-DB). We investigated whether Ets-DB expression in rat C6 gliomcells affects major tumor cell properties such as proliferation, anchorage-independent growth, migratinvasion as well as in vivo tumor development. In addition we searched for Ets target genes by cogene expression between Ets-DB expressing and control cells. We finally identified further Ets familymembers which are expressed in rat C6 glioma cells in addition to Ets 1. Materials & Methods: Rat C6 glioma cells were transduced with Ets 1 DNA-binding domain (Ets-DBexpressing retrovirus lacking its transactivation domain. Cells expressing an empty Neo-vector as well as untransduced rat C6 glioma cells were used as controls. Proliferation assays were performed using MTT assays and anchorage-independent growth through soft agar assays. Migration assays were analyzed using wound assays and either fibronectin or collagen type I Boyden chambers. Invasion using extracellular matrix (ECM)-coated Boyden chambers and mmp-9 expression was analyzed threither semi-quantitative or quantitative RT-PCR. In vivo tumor growth was carried out by implantingDB expressing rat C6 glioma cells onto the chicken chorioallantoic membrane (CAM). Ets target gen

showed that rat C6 glioma cells express in addition to Ets 1 further Ets family members, such as Elk-1Elf-1, Fli-1 and Etv-1.

were investigated using suppression subtractive hybridization (SSH). The expression of further Ets members was analyzed using semi-quantitative RT-PCR. Results: Our results indicate that the retroviral expression of Ets-DB in rat C6 glioma cells inhibits cell proliferation in soft agar Migration wa

in collagen t

Conclusion: Our results indicate that Ets transcription factors play important roles for basic properties ofrat C6 glioma cells. Targeting of Ets 1 and/or other Ets-family members might therefore become a usefexperimental tool for therapeutic strategies against malignant gliomas. Author Contact: Ayguen Sahin, PhD; Neurosurgery; 617-726-1012; sahin.ayguen

47

Poster Board Number: 42 Improved strategies for adoptive immunotherapy against glioblastoma through “3rd generation” chimeric receptor engineered T cells Ayguen Sahin, Carlos E. Sanchez, Steven Maxfield, Szofia Bullain, Beth Baratta, Bob S. Carter Department of Neurosurgery, MGH-HMS Center for Nervous System Repair, Massachusetts General

d l with a specificity

on mutant protein

Hospital, Harvard medical School Purpose: Artificial T cell receptors, also known as chimeric immune receptors (CIR), are engineeremolecules, which when expressed by T cells, redirect the T cells to kill a target celdictated by the artificial receptor. Using this technology, we previously engineered a first generation CIR targeting EGFRvIII mutant protein (G1-MRI-CIR). EGFRvIII is the most commexpressed in the deadly disease Glioblastoma multiforme (GBM, 30-50%). CD3ζ signaling alone in G1 is sufficient for initiation of target cell killing and Interferon-γ release, but ζ signaling alone fails to induce the full activation and proliferation of T cells. This causes limitations in current adoptive immunotherapy strategies. To overcome the problem of reduced activation and survival of the first generation T cells, we created a “3rd generation” (G3), tripartite CIR (MRI-CD8TM-CD28-OX40-CD3), which contained both co-stimulatory factors CD28 and OX40. Here, we investigated cell proliferation in G1 and G3-T cells grown on two different artificial Antigen Presenting Cell lines (aAPC) expressing additional co-stimulatory molecules CD32, CD80, 41bbl, and for antigen specific stimulation, EGFRvIII lacking its intracellular domain (EGFRvIIIΔ654). Materials & Methods: Peripheral Blood Mononuclear Cells (PBMCs) were isolated from healthy donors and nucleofected with either G1 or G3-MRI-CIR vector plasmids. G1 or G3-T cells were cultured under Hygromycin drug selection. T cells were re-stimulated every 12-14 days by co-culturing with two different aAPC lines KJ (K562-CD32-CD80-41bbl) or KJEAS (K562-CD32-CD80-41bbl-EGFRvIIIΔ654 and in presence of IL-2. aAPCs were loaded with anti-CD3 and anti-CD28 antibodies prior to T cell co-culturing. Cell counts were obtained via GUAVA EasyCyte (GUAVA Technologies). Results: Our results indicate that the additional co-stimulatory factors expressed in G3-T cells improved cell growth compared to G1-T cells in vitro. Although G3-T cells co-cultured with either KJ or KJEAS showed no difference in cell growth, G1-T cells co-cultured with KJEAS showed increased cell growth compared to G1-T cells co-cultured with KJ, in later time points. Conclusion: 3rd generation (G3)-CIR-MRI-T cells with additional co-stimulatory factors can improve T cell growth. Further investigations of cytokine expression, survival and anti-tumor activity will help to better understand G1 and G3-T cell biology and allow better control over adoptive immune responses in glioblastoma patients. Author Contact: Ayguen Sahin, PhD; Neurosurgery; 617-726-1012, sahin.ayguen@mgh.harvard.edu

48

Poster Board Number: 43 An Enzyme Activated Photodynamic Prodrug for the Photodynamic Therapy of Drug Resistant Bacterial Infections Ulysses W. Sallum, Xiang Zheng, Sarika Verma, Conor L. Evans, Humra Athar, Tayyaba Hasan Wellman Center for Photomedicine, Massachusetts General Hospital, Department of Dermatology, Harvard Medical School, Boston, MA, 02114

β-LEAPP was determined for various strains of S. aureus

ectrometry and nuclear magnetic resonance spectrometry. A high level of purity was verified by high performance

not produce any significant activation of β-LEAPP as detected by fluorescence spectroscopy.

β-Lactamase Enzyme Activated Photodynamic Prodrug has been shown

supported by the efficient cleavage of β-LEAPP by whole cell suspensions of bacteria induced through incubation with penicillin G and not by non-induced reference cultures. Similar specificity was observed for the photodynamic antibacterial effect of β-LEAPP, indicating that the principle of exploiting a pathogen specific enzyme for the activation of a photodynamic prodrug, in fact, works. Importantly, β-LEAPP was not hydrolyzed by cultures of human fibroblasts and keratinocytes further substantiating this general conclusion. In cocultures of S. aureus and human fibroblasts a great difference in the level of fluorescence was observed between S. aureus and human fibroblasts. Although a dramatic difference in the fluorescent emission was not observed between induced MRSA and noninduced S. aureus a striking difference in the subcellular localization of fluorescence was observed. Author Contact: Ulysses W. Sallum, Ph.D.; (617)-724-5394; USallum@partners.org

Purpose: To lead a proof of principle effort towards the creation of enzyme activated photodynamic prodrugs for the treatment of drug resistant localized infections. Specifically the design, synthesis and characterization of β-Lactamase Enzyme Activated Photodynamic Prodrug (β-LEAPP) for the photodynamic therapy of β-Lactam resistant bacterial infections.

Materials & Methods: • All chemicals and supplies were purchased from Sigma-Aldrich and Fisher Scientific, respectively. • β-LEAPP was synthesized using a multistep reaction sequence. The chemical structure of β-LEAPP

features a cephalosporin core conjugated to two phenothiazinium photosensitizers. • The ability of B. cereus penicillinase to hydrolyze and activate β-LEAPP was assayed in vitro and the

Michaelis Constant (Km) was determined. • Whole bacterial cell suspensions of S. aureus, MRSAs, & E. coli with various minimum inhibitory

concentrations for penicillin G were also assayed for the enzymatic activation of β-LEAPP. The antibacterial photodynamic efficacy of •

in vitro. • Human fibroblasts (HFF-1) were cultured in the presence or absence of; (i) β-LEAPP, and (ii) S.

aureus 29213 & MRSA and imaged by confocal microscopy using the red β-LEAPP fluorescence (670 nm).

Results: • β-LEAPP was successfully synthesized and the structure was confirmed by mass sp

liquid chromatography. • B. cereus penicillinase activated β-LEAPP efficiently with a Km value of 1.76x10-6 M. • Four strains of MRSA and one strain of ESBL-producing E. coli were found to activate β-LEAPP

while noninduced reference strains of S.aureus and E. coli did

• β-LEAPP exhibited a high level of selectivity for β-Lactamase producing strains of S. aureus. • β-Lactamase producing and nonproducing S. aureus produced distinctly different cellular effects in

HFF-1 cells as observed via confocal laser scanning microscopy.

onclusion: This first-of-a-kind Cto be highly effective in selectively targeting drug resistant bacteria in vitro. This conclusion is further

49

Poster Board Number: 44 In vivo Biomechanical Microscopy of Tissue and Biomaterials Giuliano Scarcelli and Seok Hyun Yun

ogical tissues and biomaterials are closely related to their

he ll motility. However, measuring such

remains a significant challenge due to a dearth of non-invasive technologies.

Wellman Center for Photomedicine, Massachusetts General Hospital and Harvard Medical School Purpose: The mechanical properties of biolfunctional abilities, and thus play significant roles in many areas of medicine. For example, hardened coronary arteries by calcification can cause heart problems; changes in the elasticity of crystalline lens and cornea are thought to be central in the development of ocular disorders such as cataracts, presbyopia and corneal ectasia; biomechanical compatibility is crucial in tissue engineering procedures; and, ttiffness of extra-cellular matrix influences drug delivery and ces

biomechanical propertiesOur goal is to develop a novel diagnostic technique, Brillouin confocal microscopy[1], to probe the

. Ma n light and sound

d red

w

ig 1b). e

e

suring nt

e

uthor Contact: Giuliano Scarcelli, Dermatology, 617-724-6798, gscarcelli@partners.org

biomechanical properties of tissue in vivo without contact, quantitatively, and with high spatial resolutionterials & Methods: Brillouin light scattering arises from the interaction betwee

waves inside material. Such interaction induces a small frequency shift in the scattered light that is i ctly related to the viscosity and elasticity of samples. We have developed a high-resolution optical

spectrometer that is able to measure such tiny frequency shift with unprecedented detection efficiency ane integrated it with a home-built confocal microscope.

Results: Using Brillouin microscopy, we have obtained the first 3D images that use elasticity as contrast mechanism (Fig.1a); we monitored fast dynamic changes in elastic modulus during polymer cross-linking(F .1c); and we performed the first in vivo measurement of crystalline lens in the mouse eye (Fig.W have also obtained the first in vivo demonstration of age-related stiffening of lenses (not shown).

Conclusion: Brillouin confocal microscopy can indeed characterize in vivo the biomechanics of tissuand biomaterials non-invasively and with micron-scale resolution. The first areas of biomedicalapplications we are exploring are in ophthalmology where Brillouin microscopy may enable meachanges in corneal and lens elasticity by aging, by the progression of disease, or in response to treatmeand drugs; and in tissue engineering for the optimization of procedures by mapping and monitoring in situand in real time the micromechanical properties of host and implanted tissue. [1] G. Scarcelli & S. H. Yun, Brillouin confocal microscopy for 3D mechanical imaging, NaturPhotonics 2, 39-43 (2008). A

g. 1: (a) Brillouin images of an intraocular lens (ellipse) immersed in fluFi id (blue). (b) In situ measurement of

the elastic modulus of a mouse crystalline lens. (c) Real-time monitoring of elasticity variation in a polymer due to UV-induced cross-linking.

50

Poster Board Number: 45 Deletion of CXCR3 on effector T cells prevents rejection and enhances tolerance in a model

enter for Immunology and Inflammatory Diseases, Massachusetts General Hospital, Harvard

0-OVA mice). Adoptive transfer of OVA-

Results: We were able to inhibit early Teff recruitment into the lung and induce long-term survival with transfer of CXCR3-/- OT-I Teff cells into the CC10-OVA mice. Analysis of the lung from surviving CC10 mice revealed CXCR3-/- Teff cells were present, but unresponsive to restimulation with OVA, suggesting that the early reduction in Teff recruitment into the lung ereth ficant increase in Foxp3 Treg cells compared to control mice. We

splant rejection.

uthor Contact: Edward Seung, CIID, 6-6149, eseung@partners.org

of acute lung rejection Edward Seung, Benjamin D. Medoff, Andrew D. Luster CMedical School, Charlestown, MA Purpose: Lung transplantation remains the only effective therapy for patients with end-stage pulmonary diseases. Unfortunately, acute rejection of the lung remains a frequent complication and is an important cause of morbidity and mortality. Both the rejection process and induction of transplantation tolerance are now thought to be dependent on a delicate balance between effector T lymphocytes (Teff) and Foxp3+ regulatory T cells (Tregs). We explored a novel approach to tip the balance in favor of Tregs and tolerance induction by modulating chemokine activity. Materials & Methods: We have developed a novel transgenic model of lung rejection in which mice were generated that express a membrane bound form of chicken egg albumin (OVA) exclusively in the airway lining cells of the lung (CC1specific CD8+ Teff cells (OT-I) into these mice leads to rapid rejection of the lungs and death of the animals.

nhances the induction of tolerance. Recovered CXCR3-/- Teff cells from the spleen, however, sponded vigorously to OVA restimulation. Consistent with this, analysis of the lungs from ese mice revealed a signi +

examined the importance of Treg cells by demonstrating that CC10-OVA mice deficient in Tregs were not able to generate tolerance with reduced Teff recruitment and resulted in death. Conclusion: We propose a new paradigm in which manipulation of Teff trafficking into allografts can facilitate the induction of tolerance in the target organ by influencing regulatory functions. This may provide a new therapeutic approach to prevent tran

CC10-OVA Mouse

DEATH

5-7 daysInjected i.p.

In Vitro ActivatedAnti-OVA CD8+ T Cells

(OT-I cells)

A

51

Poster Board Number: 46 The IFN Antagonist SOCS3 Downregulates Hepatitis C Virus Replication in Cell Culture Run–Xuan Shao, Wenyu Lin, Andrew W. Tai, Lee F. Peng, Sabina Sabharwal, Woo Jin Chung, Raymond

,

and

o en its

sought

& Methods: We used OR6 cells harboring a full-length HCV genotype 1b replicon pressing luciferase as a measure of HCV replication. A genotype 2a HCV full-length infection system

CV replication. wer in OR6 cells and JFH1 infected Huh7.5.1

in OR6 cells, HCV replication was significantly se assay (rluc/cell viability, SOCS3: 1.6, control cells: 3.3,

p=0.0012), and by RT-PCR (HCV/actin, SOCS3: 2.1, control cells: 5.1, P=0.028). Similarly, HCV core protein levels were significantly decreased. Moreover, in JFH1 infected Huh7.5.1 cells, SOCS3 overexpression also lowered HCV-RNA level (HCV/actin, SOCS3: 1.0, control cells: 2.4, P=0.043) and

re protein. In contrast, SOCS3 knockdown in JFH1 infected cells using shRNA was associated with creased HCV core protein. Interestingly, IFN-induced P-STAT1 levels were not increased by

not

T. Chung, Gastrointestinal Unit, Massachusetts General Hospital, Harvard Medical School. Boston, MassachusettsUSA Purpose: Chronic infection with HCV is associated with defects in interferon (IFN) signaling activation. The protein suppressor of cytokine signaling 3 (SOCS3) is thought to block type I IFNsignaling by impairing STAT1 activation. We and others have observed that levels of hepatic SOCS3 aresignificantly higher in chronic hepatitis C (CHC), particularly among persons who are nonresponders tsubsequent IFN treatment. However, the effects of SOCS3 on HCV itself are not known. Givputative role, we hypothesized that SOCS3 would be antagonistic to viral replication. We therefore to dissect the relationship between SOCS3 and HCV using HCV replication models. Materialsexwas also established in Huh 7.5.1 cells using the JFH1 infectious strain. Levels of SOCS3 and HCV core protein were determined by Western blotting. HCV replication in OR6 cells was monitored by measuring Renilla luciferase activity (RLUs). HCV RNA was quantified by real time RT-PCR. We transientlytransfected pCDNA3-human-SOCS3 into OR6 cells and JFH1 infected Huh7.5.1 cells, and observed theimpact on HCV replication. We also created stable SOCS3 knockdown using shRNAs in Huh7.5.1 cells,and assessed HResults: SOCS3 protein expression was significantly locells. Surprisingly, when SOCS3 was overexpressed lower than in mock-transfected cells by lucifera

coinoverexpression of SOCS3 in OR6 or JFH-1 infected cells. Conclusion: Our data surprisingly suggest that SOCS3 actually suppresses HCV replication, and does impair IFN signaling in the context of HCV infection. They instead suggest a model by which increased levels of SOCS3 observed in chronic IFN nonresponders may reflect a compensatory host response to persistent infection and impaired defenses elsewhere in the innate antiviral pathway. Efforts to enhanceSOCS3 function may be a useful strategy to control HCV infection. Author Contact: Run-Xuan Shao, Gastrointestinal Unit, 617-726-2061, rshao@partners.org

52

Poster Board Number: 47 Discovery and Initial validation of novel markers for myocardial injury Xu Shi1, Terri Addona2, Dongxiao Shen1, Marc Sabatine3, Steve A. Carr2, Robert e. Gerstzen1

1Cardiology Division and Center for immunology and Inflammatory Diseases, Massachusetts General Hospital/Havard Medical University ; 2Broad Institute of MIT and Harvard University; 3Brigham and Women’s hospital and TIMI studygroup

ve

onclusion: This study identifies changes in circulating proteins as soon as 10min after ated and

paves the way for novel protein biomarkers to be studied for the diagnosis of myocardial ischemia. Validation in larger, heterogeneous cohorts is ongoing. Author Contact: Xu Shi, Medicine, 617-726-0725, xshi@partners.org

Purpose: Presently available biomarkers do not have sufficient sensitivity and specificity early after the onset of myocardial injury. Emerging proteomics discovery tools have raised the possibility of establishing novel plasma signatures of myocardial injury Materials & Methods: An LC-MS-based proteomics discovery platform was applied to 3 patients undergoing alcohol septal ablation treatment for hypertrophic obstructicardiomyopathy, a human model of “planed” myocardial infarction (PMI). Serial blood samples from the coronary sinus (the main vein that drains the heart) were obtained before and at 10min and 60min after induced myocardial infarction. For initial validation, peripheral blood samples were obtained before and after 10, 60, 120, 240 minutes and 24 hours after induced myocardial infarction in 5 patients. Additional patients undergoing elective diagnostic coronary angiography and patients with spontaneous MI served as negative and positive controls, respectively. Results: The LC-MS-based platform identified changes in known markers of myocardial injury, including creatine kinase and myoglobin. In addition, we identified hundreds of novel plasma protein changes following myocardial injury. 2 of these candidates (Midkine and CCL21) were further validated in both cs blood and peripheral blood and found they all increased from 10min till 120/60min (Midkine/CCL21) and back to baseline level at 240/120min (Midkine/CCL21), the validation results were consistent with that found by LC-MS/MS. Cmyocardial injury, a time frame in which no currently used plasma biomarkers are elev

53

Poster Board Number: 48 JNK I knockout mice protected from lung injury due to smoke inhalation from fire.

, .

ouse injury to show that

oke inhalation induces airway apoptosis through activation of the JNK pathway and that

Cell Mol

e burning cotton for 15 minutes.

arles River Laboratory) and JNK I knockout to smoke, allowed to recover and sacrificed 4

later.

IL-6 TNF-α

O. L. Syrkina, J .L Beagle, D. A. Quinn, C. A. Hales. Pulmonary and Critical Care Unit, Department of Medicine, Massachusetts General HospitalHarvard Medical School, Shriners Burn Hospital, Boston Purpose: In United States each year more than 4,000 people die and more than 25,000 are injured in fires. (FEMA, 2006). Smoke inhalation injury is a serious threat to victims of hfires, explosions. We used our established rat model of smoke inhalationsmtreatment with JNK-inhibitor diminished airway apoptosis and prolonged animal survival. (Quinn DA et al. J Burn Care Rehabil 2003. Syrkina OL et al. Am J Physiol Lung Physiol. 2007). We hypothesized that JNK knockout mice would bused our mouse model of exposure to smoke fromMethod: Mice (22-27g), wild type C57 BL6 (Ch(Jackson Laboratory), were anesthetized, exposed(S/4,4 hrs post smoke) and 24 (S/24) hoursResults: Group Total number neutrophils MIP-2

e protected from smoke inhalation injury. W

(x104) (pg/ml) ( pg/ml ) (pg/ml) in BAL S/4 hrs S/24 hrs S/4 hrs S/24 hrs S/4 hrs S/24hrs S/4 hrs S/24hrs WT 109.1± 10.0 233.0±54.1 20.7 ±2.6 36.7±2.8 1.7±0.1 0.73±0.1 47.5±17.6 57.9± 10.1 JNK I 30.21 ± 0.7* 36.7 ± 0.5 * 19.6± 1.8 26.1 ±1.6* 0.59±0.1* 0.61±0.1 32.9±5.6* 28.5±2.1* After smoke animals had higher level of PMN and cytokines than non smoked control but knockouts had a lower level of PMN, MIP-2 and TNF-α at 24 hrs. IL-6 peaked earlier at 4 hrs but was lower in KO mice. *p< 0.05 JNK I KO vs. WT at the same time point. JNK knockout mice showed no positive staining for apoptosis and have less inflammation in major airways. Conclusion: We conclude that JNK I knockout mice were protected from the inflammatory component of smoke inhalation injury. This gives a new focus for the approach to the treatment f smoke induced lung injury. o

Supported by: Shriners Grant # 8620

Author Contact: Olga L Syrkina, MD; e-mail: osyrkina@partners.org; phone: 617-726-5630

54

Poster Board Number: 49 Rapid determination of arginine, citrulline and ornithine in human plasma by liquid chromatography-mass spectrometry

s in

t

t and ons

thods: a LC-MS instrument equipped with ESI was used to establish the method.

g

ance issued by FDA: Guidance for Industry, al Methods Validation. In addition, Plasma samples from 11 human volunteers were analyzed

Hongyun Wang, Yong-ming Yu Shriners Hospital for Children, Department of Surgery, Massachusetts General Hospital, Harvard Medical School. Boston, Massachusetts, USA. Purpose: Arginine is a nutritionally essential amino acid in burn patients, and it serves multiple rolethe patho-physiological response to burn injury. To fully explore the status of arginine metabolism and conversion rate of citrulline to arginine, and disposal rate of arginine via its conversion to, and subsequenoxidation of ornithine, a robust and precise method for the determination of arginine, cirtrulline and ornithine in biological fluids would be of great value. The purpose of this study is to develop a robusaccurate method for further study of the metabolic pathways and potential patho-physiological functiconcerning urea cycle. Materials & MeConcentrations of arginine, citrulline and ornithine were quantified within 3.5 min, using only 25μl of plasma sample, pretreated by protein procipitation. Matrix effect, the phenomena interferindetermination of amino acids in biological matrix was systemically evaluated. To ensure the accuracy of this method, the method was validated for selectivity, linearity, lower limit of quantification, extraction recovery, accuracy and precision according to the guidBioanalyticusing this method.

Figure 1. Typical SIM chromatogram of extracted human plasma sample: (A) ornithine, (B) arginine, (C) citrulline Results: Plasma concentrations of arginine,citrulline and ornithine can be simultaneously determined within 3 min(Figure 1). Results of validation met the strict requirements from FDA. The values of arginine, citrulline and ornithine in human volunteers obtained with this LC-MS method were well comparable to the previously published reports. Conclusion: The LC-MS method is reliable and efficient for the rapid determination of arginine,citrulline and ornithins, providing a facile access to acquiring comprehensive data on metabolism of the three urea cycle intermediates. Author Contact: Hongyun Wang; Surgery; 617-823-5791; hwang19@partners.org

55

Poster Board Number: 50 Increased Cardiovascular Risk Markers in Patients with Growth Hormone DeficiencyAcromegaly

1

after Cure of

l, C Beauregard,1 L Nachtigall,1 B Swearingen,2 J Loeffler,3 Z Omer,1

gery, Department of Radiation Oncology, Division of ardiology Massachusetts General Hospital and Harvard Medical School, Boston, MA, United States,

r risk markers and truncal adiposity compared to patients with prior acromegaly

), III.

l

eral Clinical Research Center. Results: Mean age and BMI were 45.6

TL Wexler,1 KK Miller,1 L GunnelLC Hemphill,4 BMK Biller1 and A Klibanski.11Neuroendocrine Unit, 2Department of Neurosur 3 4

C02114. Purpose: Growth hormone deficiency (GHD) is associated with increased visceral adiposity, cardiovascular morbidity, and cardiovascular (CV) risk markers. Following cure of acromegaly, GHDmay develop; however, its effects on CV risk and body composition in this population are not established. We hypothesized that the development of GHD in patients with prior acromegaly would result in increased cardiovasculawho remained GH sufficient. Materials & Methods: We investigated CV risk markers and visceral fat in 4 groups of subjects (n=94): I. Cured acromegaly with GHD (n=26), II. Cured acromegaly with GH sufficiency (n=19Active acromegaly (n=11), and IV. GHD with no history of acromegaly (n= 38). Fasting serum was collected at 0800h for hsCRP, total cholesterol (TC), HDL, LDL and triglycerides. Carotid intima-mediathickness (IMT), total and visceral fat determined by DXA and cross-sectional CT at L4, and insulinprofile via oral glucose tolerance test and HOMA-IR, were measured. This cross-sectional study was conducted in the Gen

1.1 y (SEM) and 29.8 0.66 kg/m2, respectively, and were not different among the groups. Mean IGF-1 SDS values were: I. -2.0 0.1; II. -1.3 0.2; III. 4.7 1.1; IV. -1.70.1. Visceral adipose and total fat mass were higher in both GHD groups (I and IV) than in II and III (p<0.006). Abdominal subcutaneous fat mass was also higher in I and IV than in III (p<0.02). Mean hsCRP differed among all groups, with the highest levels in group IV (6.5 0.9 mg/L), followed by I (4.00.9 mg/L), and lowest in III (0.8 0.3 mg/L) (p<0.04). IMT was higher in III than all other groups (p<0.03), as were baseline glucose, insulin, AUC glucose and HOMA-IR (p<0.02). Groups did not differ in TC, HDL, LDL, or triglycerides. hsCRP differs with GH status

sis of

7) 726-1347,

Conclusion: Our data demonstrate that patients who develop GHD following cure of acromegaly have increased visceral adiposity and hsCRP compared with those who remain GH sufficient. The diagnoGHD is rarely pursued or treated in patients with a history of acromegaly, and these data suggest that it will be important to determine whether GH replacement will improve body composition and CV risk in this group. Author Contact: Tamara Wexler, Neuroendocrine Unit, Div of Endocrinology, phone (61email twexler@partners.org

0GHD Non- GHD Post- GH Sufficient Active

0

1

23

45

67

8

1 2 3 4

*

*

*

Acromegaly

*

hsC

RP

mg/

L

1

2

3

4

5

6

7

8

Acromegaly

IV

Acromegaly

I

Post-Acromegaly

II III

0GHD Non- GHD Post- GH Sufficient Active

0

1

23

45

67

8

1 2 3 4

*

*

*

Acromegaly

*

hsC

RP

mg/

L

1

2

3

4

5

6

7

8

Acromegaly

IV

Acromegaly

I

Post-Acromegaly

II III

56

Poster Board Number: 51 L-DOPA-induced dyskinesia is attenuated in the absence of adenosine A1 and A2A receptors in hemiparkinsonian mice

General Hospital, Charlestown, MA 02129 o

sia (LID). Initial clinical trials of A2A antagonists targeted PD patients who had already odels

icantly over the 21 ays (p<0.05, two-way ANOVA) in A1 KO, A2A KO nd (A1-A2A) double KO mice (see figure). PPE RNA was found to be reduced in all KO mice and

ced significantly in 2A KO mice. In addition, neuronal integrity assessed y striatal dopamine content was unaltered on both e intact and lesioned sides of all KOs, which had

ver 90% loss of dopamine observed in the arkinsonian side.

inergic therapy for treating LID, and the possibility of targeting either the A1 or A2A denosine receptor subtypes.

Xiao D., Cassin J., McAleavey A., Xu Y-H and Schwarzschild M.A. Department of Neurology, Massachusetts Purpose: Adenosine A2A receptor antagonists provide a promising nondopaminergic approach tthe treatment of Parkinson’s disease with L-3,4-dihydroxyphenylalanine (L-DOPA)-induced dyskinedeveloped LID in an effort to improve symptoms while reducing existing LID. In mouse mof LID, we found that the development of LID is significantly but only partially reduced in forebrain A2A receptor knockout (KO) mice. The goal of this study is to explore the effects of genetic depletion of the A1 as well as the A2A subtypes of adenosine receptor on the development of sensitized responses to L-DOPA including abnormal involuntary movement (AIM), a measurement of LID, on molecular markers of LID including striatal enkephalin (PPE) and dynorphin, and on dopaminergic nigrostriatal neuron integrity. Materials & Methods: Hemiparkinsonian A1, A2A and double (A1-A2A) KO mice (unilaterally lesioned with 6-hydroxydopamine) were treated daily for 3 weeks with a low dose of L-DOPA preceded by benserazide (a dopa decarboxylase inhibitor).

A

Following daily L-DOPA treatment, the hemiparkinsonian mice developed behavioral sensitization recorded by contralateral rotations and AIMs. Three days after the last treatment, the striata of the mice were analyzed for their expression of PPE and prodynorphin mRNA by in situ hybridization and their content of dopamine by HPLC. Results: Chronically, rotational sensitization and

tal AIMs were attenuated signif Btodamprodynorphin mRNA was reduAbthopConclusion: Here we show that the adenosine A1 as well as A2A receptor contributes to LID in a mouse model of Parkinson’s disease. Indirect pathway labeled by PPE may underlie the role of adenosine receptor in LID. Lack of adenosine receptor has no effect on nigrostriatal neuron integrity. This result supports the development of anti-adenosaAuthor contact: Danqing Xiao; Neurology; 617-726-1243; dxiao@partners.org

57

Poster Board Number: 52 Reversal of established autoimmune diabetes by H60-loaded immature dendritic cells plus the natural killer T cells ligand -Galactosylceramide Wen Yang, Xiaoping Song, KyungMee Chung, Ed Paek, Michael Cla

αre-Salzler, and S. Brian

t lineage fate mapping transgenic mice that have recovered glycemic

lophos mide (CY)-induced diabetic female -loaded immature DC (2 x 106/mouse) plus α-

received an insulin injection daily until they were e levels once weekly.

ia.

s derived from innate immune receptors

to OD mice with H60-loaded DC and

re ion of 64% of cyclophosphamide tes r ires both iNKT-based therapy and % of recovery of glycemic control was

Z/AP bi-transgenic NOD mice. We find ts of non-diabetic bi-transgenic NOD

ts of reversed bi-transgenic

e and subsequent β-cell function can be and

Wilson Diabetes Research Laboratory, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Cambridge, Massachusetts, USA Purpose: Our proposal is based on the hypothesis that an integral misstep in the development of islet-specific autoimmunity is ineffective iNKT/DC cross talk and the failure to generate “tolerogenic” DC. Published and preliminary results support the notion that controlling the iNKT/DC axis can rectify this misstep. To test this idea the following aims are proposed: 1. Determine if the combination of iNKT cell activation and transfer of antigen-loaded DC canreverse autoimmune diabetes in the NOD mouse model. 2. Investigate the mechanisms responsible for the recovery of euglicemia. 3. Identify cells regen

he possible sources of pancreatic beta eration in the

control. Materials & Methods: New onset (< 1 week) cycNOD mice were injected intravenously with H60GalCer (2ug) once per week for 5 weeks. Micereversed. Mice were evaluated for glucosuria daily and blood glucosComplete remission was defined as the disappearance of glycosuria and a return to euglycem Results: To examine whether class I restricted antigencan be used with DC, an immunodominant H60 minor histocompatibility (H) antigen-derived peptide presented by H-2Kb MHC class I was loaded into bone marrow-derived DC and used

pha

treat diabetic NOD mice. Notably, concurrent treatmenthe iNKT superagonist α-GalCer resulted in long-term(CY)-induced diabetic NOD mice. Reversal of diabeCD1d expression by the infused DC because only 10noted when CD1d(-/-)-DC where infused or iNKT activation was not included. To investigate the resource of islets regeneration, we use RIP-CreER;that most insulin positive cells express HPAP in the islemice, but very few insulin positive cells express HPAP in the isleNOD mice.

t of Nmissequ

Conclusion: These findings suggest that self-tolerancrestored in overtly diabetic mice via the modulation of antigen-loaded DC and iNKT cells, islets regeneration may be mainly the result of differentiation of pancreatic progenitor cells. Author Contact: Wen Yang; Medicine; 617-768-8489; wyang@mgh.harvard.edu

58

Poster Board Number: 53 Breathing Nitric Oxide Prevents Pulmonary Vasoconstriction Caused by Infusion of Hemoglobin-based Oxygen Carrier in Awake Lambs Binglan Yu1, Gian Paolo Volpato 1, Kenneth D. Bloch 1,2, Warren M. Zapol 1From the 1Anesthesia Center for Critical Care Research of the Department of Anesthesia and Critical Care and the 2Cardiovascular Research Center of the Department of Medicine, Massachusetts GeneHospital, Harvard Medical School, Boston, Massachusetts, USA 02114 Purpose: Two primary concerns hampering the clinical acceptanc

ral

e of acellular, hemoglobin(Hb)-based

201 this

al

ntravenous administration of HBOC-201 (12 ml/kg) in healthy, awake lambs induced prolonged

ncentration of NO during and after administration of HBOCs may enable administration of acellular Hb

oxygen carriers (HBOC) are the occurrence of systemic and pulmonary vasoconstriction and the maintenance of the heme-iron in the reduced state (Fe+2). We recently demonstrated that pretreatment with inhaled nitric oxide (NO) prevents the subsequent systemic hypertension induced by HBOC-(polymerized bovine Hb) infusion in awake mice and sheep without causing methemoglobinemia. Instudy, we investigated the pulmonary and systemic hemodynamic effects of breathing NO both beforeand after the administration of HBOC-201 in awake lambs. Materials & Methods: HBOC-201 was obtained from Biopure (Cambridge, MA). Mean arteripressure, mean pulmonary arterial pressure, central venous pressure, pulmonary arterial occlusion pressure, heart rate and cardiac output were measured. Methemoglobin (MetHb) levels in whole bloodand plasma were determined by the cyanomethemoglobin method.

esults: IRsystemic and pulmonary vasoconstriction. Pretreatment with inhaled NO (80 parts per million (ppm), 1 h) prevented the HBOC-201-induced increase in mean arterial pressure, but not pulmonary arterial pressure, systemic vascular resistance, or pulmonary vascular resistance. Pretreatment with inhaled NO (80 ppm, 1h) followed by breathing a lower concentration of NO (5 ppm, 2 h) during and after HBOC-201 infusion prevented systemic and pulmonary vasoconstriction without causing methemoglobinemia. Conclusion: These findings demonstrate that pretreatment with inhaled NO followed by breathing a low cosubstitutes without vasoconstriction while preserving their oxygen-carrying capacity. Author Contact: Binglan Yu , Department of Anesthesia and Critical, 617-726-8825, byu1@partners.org

59

Poster Board Number: 54 Phosphate Regulates Chondrocyte Differentiation, Proliferation and Apoptosis in a Model of Embryonic Endochondral Bone Formation

ic chondrocytes during postnatal

l. This

ml ascorbic acid and 0.25% heat-inactivated e ulture

g showed no

tivated caspase-9 in pertrophic chondrocytes of metatarsals cultured with 7 mM phosphate compared to 1.25 mM.

reversed the growth inhibition observed with 7 mM phosphate. Conclusion: Our studies demonstrate that embryonic hypertrophic chondrocytes are susceptible to phosphate mediated apoptosis by activation of the caspase-9-mediated mitochondrial pathway in the metatarsal culture system. Author Contact: Alena A. Zalutskaya, Endocrine Unit. 617-724-7790, azalutskaya@partners.org

Alena A. Zalutskaya and Marie B. Demay Endocrine Unit, Massachusetts General Hospital, Boston, MA, USA Purpose: Phosphate is required for terminal differentiation of hypertrophgrowth plate maturation. To determine whether extracellular phosphate plays a role during embryonic endochondral bone formation, we performed investigations in the mouse metatarsal culture modemodel recapitulates in vivo bone development in culture and circumvents the homeostatic changes thatoccur in the intact animal. Materials & Methods: Metatarsals were isolated from day 15.5 C57BL6/J embryos and cultured in phosphate-free high glucose DMEM (1% Pen/Strep, 0.05mg/FBS) for 4, 8 and 12 days (37ºC, 5% CO2) with 1.25 mM (control) and 7 mM phosphate (high). Thlength of each metatarsal was measured at the time of isolation, as well as at the end of the 12 day cperiod. Frozen sections were obtained to permit histology, immunohistochemistry, TUNEL and in situ hybridization analyses with digoxigenin-UTP labeled probes. Results: Metatarsals cultured with 7 mM phosphate showed a decrease in growth and proliferation compared to 1.25 mM phosphate. Histologic analysis and in situ hybridization with Collagen II, Collagen X and Osteopontin revealed an early enhancement of hypertrophic chondrocyte differentiation in metatarsals cultured with 7 mM phosphate (4 days), followed by a decrease in osteopontin expression (8 and 12 days), compared to 1.25 mM phosphate. In contrast to the control, Von Kossa staininmineralization in metatarsals cultured with 7 mM phosphate. TUNEL assays revealed a dramatic increase in apoptosis of hypertrophic chondrocytes in metatarsals cultured with high phosphate. Immunohistochemistry with anti-cleaved caspase-9 antibody showed an increase in achyTreatment of metatarsals with 20 μM Z-LEHD-FMK caspase-9 inhibitor inhibited apoptosis and partially

60

Poster Board Number: 55 Survivin Expression Is Associated with Prostate Cancer Metastasis Purpose: The major clinical challenge in prostate cancer is metastatic disease. Therefore, the

tastasis in prostate cancer may be portant for treating this disease. Survivin is a 16.5 kDa member of the inhibitor of apoptosis family

gly s

by assachusetts General Hospital who have been clinically followed for 10 years

r longer. We also used preclinical prostate cancer models to explore the potential mechanisms by which

age

ed ned tumor material for Survivin analysis.

edian follow-up was 102 months (range 5 to 127 months). Distant failure was determined on the basis Soft

od. In

at Massachusetts

s (Figure

ostate cancer in vivo model (Figure 2). urthermore, attenuation of Survivin resulted in changes in the microtubule cytoskeleton, loss of cellular

ts with T1/T2 prostate cancers treated at Massachusetts General Hospital by efinitive radiotherapy. In prostate cancer cells, Survivin seemed to control cytoskeletal remodeling, cell

tastasis. Survivin may be a potentially important prognostic marker and therapeutic target in metastatic prostate cancer. Author Contact: Min Zhang; Radiation Oncology; 617-643-3424; mzhang5@partners.org

identification of target proteins responsible for the control of meimthat has been implicated in both control of apoptosis and regulation of cell division. Survivin has not been found to be expressed in normal secretory epithelium of the prostate, but has been found to be stronexpressed in prostate cancer cells. In the present study, we examined whether Survivin expression waassociated with an increased risk of distant metastasis in men with T1/T2 prostate cancers treated definitive radiotherapy at MoSurvivin controls metastatic potential in prostate cancer cells. Materials & Methods: Between 1991-1993, 205 patients with T1 (23%) and T2 (77%) prostate cancer were treated with external beam radiation therapy at the Massachusetts General Hospital. The medianat diagnosis was 74 years (range 51 to 90). The median dose to the prostate was 68.4 Gy (ranging from 64 to 72.6 Gy). No patient received adjuvant androgen deprivation therapy. Tissue blocks were obtainfrom many, and 62 patients had adequate and suitable-staiMof clinical criteria, with confirmation of bony metastases by a positive bone scan or radiograph.tissue metastases were confirmed by computed tomography. Actuarial analyses were performed using the Kaplan-Meier method, and differences between curves were assessed using the log-rank methpreclinical studies, replication-deficient adenovirus encoding phosphorylation-defective SurvivinThr34→Ala dominant-negative mutant pAd-S(T34A) or short hairpin RNA (shRNA) was used to inhibitSurvivin in prostate cancer models, and the cell motility, morphology, and metastasis were investigated. Results: Our correlative data on men with early-stage (T1/T2) prostate cancers treatedGeneral Hospital by definitive radiotherapy indicated that higher levels of Survivin expression was associated with a significantly increased risk for the subsequent development of distant metastasi1). The inhibition of Survivin dramatically inhibited invasiveness of prostate cancer cells in the in vitro invasion assay and spontaneous metastasis in the Dunning prFpolarity (Figure 3), and loss of motility. Conclusion: High Survivin expression appeared to be associated with a significantly increased risk of distant metastasis in patiendpolarity, and cell motility, and inhibition of Survivin dramatically inhibited spontaneous me

Fwith a significantly increased risk of distant metastasis.

igure 1. Higher levels of Survivin expression was associated F e microtubule cytoskeleton, loss of cellular polarity in AT6 cells.

igure 2. Attenuation of Survivin resulted in changes in th

61

Poster Board Number: 56 Inhibition of p21-activated Kinase 6 (PAK6) Increases Radiosensitivity of Prostate Cancer Cells

ed, and involve alterations in cell cycle distribution and impaired

Min Zhang, Michael Siedow, Gregory Saia, Kamalakannan Palanichamy, Tim Lautenschläger, Arnab Chakravarti Department of Radiation Oncology, Massachusetts General Hospital/Harvard Medical School Purpose: For patients with high-risk localized prostate cancer, biochemical and clinical relapse after radiotherapy remains a significant problem. New therapeutic target for radiation sensitization is needed. p21-activated kinase 6 (PAK6), a serine/threonine kinase belonging to the p21-activated kinase (PAK) family, was first identified as an androgen receptor (AR)-interacting protein able to inhibit AR-mediated transcriptional responses. In this study, we investigated the role of PAK6 in radiation-induced cell death in human PC-3 and DU-145 prostate cancer cells. Materials & Methods: We used a short hairpin RNA (shRNA) strategy to stably knock down PAK6 in PC-3 and DU-145 cells. Radiation sensitivities were compared in PAK6 stably knockdown cells and in scrambled shRNA-expressing cells. Results: 6 Gy of irradiation upregulated PAK6 mRNA and protein levels in PC-3 and DU-145 cells. Transfection with PAK6 shRNA resulted in a significant decrease of PAK6 mRNA and protein expression. After irradiation, an increased percentage of apoptotic cells and an increased cleaved caspase 3 level in parallel with a decreased cell viability and a reduced clonogenic survival was shown. Furthermore, transfection with PAK6 shRNA blocked cells in a more radiosensitive G2/M phase and increased levels of DNA double-strand breaks as indicated by a higher amount of γ-H2AX in irradiated cells. We further explored the potential mechanisms by which PAK6 mediates resistance to radiation-induced apoptosis. Inhibition of PAK6 caused a decrease in Ser112 phosphorylation of BAD, a proapoptotic member of the Bcl-2 family, which led to enhanced binding of BAD to Bcl-2 and Bcl-XL and release of cytochrome c culminating into caspase activation and cell apoptosis. Conclusion: The combination of PAK6 inhibition and irradiation resulted in significantly decreased survival of prostate cancer cells. The underlying mechanisms by which targeting Pak6 may improve radiation response seem to be multifacetDNA double-strand break repair as well as relieved BAD phosphorylation. Additional studies are ongoing to validate PAK6 as a target for radiation sensitization in animal models of prostate cancer. Author Contact: Min Zhang; Radiation Oncology; 617-643-3424; mzhang5@partners.org

62

AUTHOR INDEX

Akilov, Oleg ............................................................ 2 7

.................... 10 15

.............................................. 14 19 Eichbaum, Quentin ............................................................ 15 20

Lage, Kasper ............................................................ 28 33

Peng, Leilei ............................................................ 35 40

Purchke, Martin ............................................................ 37 42 Qyang, Yibing ............................................................ 38 43 Rahman, Rosanna ............................................................ 39 44 Rizvi, Imran ............................................................ 40 45 Sahin, Ayguen ............................................................ 41, 42 46, 47 Sallum, Ullysses ............................................................ 43 48 Scarcelli, Giuliano ............................................................ 44 49 Seung, Edward ............................................................ 45 50 Shao, Run Xuan ............................................................ 46 51 Shi, Xu ............................................................ 47 52 Syrkina, Olga ............................................................ 48 53 Wang, Hongyun ............................................................ 49 54 Wexler, Tamara ............................................................ 50 55 Xiao, Danqing

............................................................ 51 56

Author Name Board # Page # Abu-Yousif, Adnan ............................................................ 1 6

Athar, Humra ............................................................ 3 8 Batista, Dalia ............................................................ 4 9 Berra, Lorenzo ............................................................ 5, 6 10, 11 Broekman, Marieke ............................................................ 7 12 Buhrman, Jason ............................................................ 8 13 Cao, Yi ............................................................ 9 14 Celli, Jonathan ........................................Chang, Sung ............................................................ 11 16 Chellappa, Vasant ............................................................ 12 17 Chen, Kathryn ............................................................ 13 18 Chen, Zhe ..............

Finkelstein, Eric ............................................................ 16 21 Graham, Jay ............................................................ 17 22 Hasan, Syed ............................................................ 18 23 He, Chengwei ............................................................ 19 24 Huang, Jinghe ............................................................ 20 25 Islam, Tina ............................................................ 21, 22 26, 27 Jeyaretna, Deva ............................................................ 23 28 Ketteler, Robin ............................................................ 24 29 Kim, Pilhan ............................................................ 25 30 Kolliputi, Narasaiah ............................................................ 26 31 Kozanek, Michal ............................................................ 27 32

Li, Gang ............................................................ 29 34 Luderer, Hilary ............................................................ 30 35 Maguire, Casey ............................................................ 31 36 Mai, Zhiming ............................................................ 32 37 Merkulova, Maria ............................................................ 33 38 Nazarian, Rosalynn ............................................................ 34 39

Plovie, Eva ............................................................ 36 41

63

Author Name Yang, Wen ..............

Board # Page #

............................................................ 53 58

............................................................ 54 59

............................................................

.............................................. 52 57 Yu, Binglan Zalutskaya, Alena Zhang, Min 55, 56 60, 61

64

NOTES

65

66

Massachusetts General Hospital 55 Fruit Street, Bulfinch 370

Boston, MA 02114

Peggy Ryan Staff Assistant Bulfinch 370 (617) 643-1606 mryan11@partners.org

Ann Skoczenski, PhD Program Manager Bulfinch 360E (617) 643-1170 askoczenski@partners.org

Tayyaba Hasan, PhD Director Bulfinch 370F (617) 643-2589 thasanORCD@partners.org

www.massgeneral.org/facultydevelopment/orcd