home | clinical cancer research - evaluation of novel prostate … · 1 evaluation of novel...

1

Evaluation of Novel Prostate-Specific Membrane Antigen-Targeted Near Infrared Imaging

Agent for Fluorescence-Guided Surgery of Prostate Cancer

Sumith A. Kularatne,1, *

Mini Thomas,1 Carrie H. Myers,

1 Pravin Gagare,

1 Ananda K.

Kanduluru,1 Christa J. Crian

2, Brandy N. Cichocki

2

1 On Target Laboratories, 1281 Win Hentschel Blvd, West Lafayette, IN, 47906 USA

2 Department of Veterinary Clinical Sciences, Purdue University, Lynn Hall, 625 Harrison St.,

West Lafayette, IN 47907 USA

* correspond: Sumith A. Kularatne, Ph.D. On Target Laboratories, 1281 Win Hentschel Blvd,

West Lafayette, IN, 47906, Tele: 765-558-4547, Fax: 765-598-4452, email:

[email protected]

Keywords: cancer surgery, fluorescence-guided surgery, tumor-specific dyes, targeted-near-

infrared dyes, and fluorescence-guided radical prostatectomy

Statement of Translational Relevance: We developed a Prostate-specific membrane antigen

(PSMA)-targeted near infrared (NIR) imaging agent (OTL78) that: (i) binds to PSMA+ tumors

with high affinity and specificity, (ii) allows to use sub-nanomolar concentration to visualize

small tumors, (iii) clears rapidly from PSMA-negative tissues with half-life of 17 min, (iv)

retains tumor fluorescence for over 48 hours, allowing visualization throughout FGS, and (v)

allows to accomplish negative surgical tumor margins, (vi) has an excellent safety profile in

animals. OTL78 has proven to be a clinical candidate to yield sharp tumor boundaries with

negative tumor margins within 1 - 2 hours of infusion during radical prostatectomy. With its

recent entry into Investigational New Drug (IND)-enabling studies, OTL78 has a potential to

Research. on August 7, 2020. © 2018 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on September 10, 2018; DOI: 10.1158/1078-0432.CCR-18-0803

mailto:[email protected]

http://clincancerres.aacrjournals.org/

2

become the first PSMA-targeted NIR agent to enter into the clinic for use in fluorescence-guided

radical prostatectomy.

Abstract

Purpose: The ability to locate and remove all malignant lesions during radical prostatectomy

leads not only to prevent biochemical recurrence (BCR) and possible side effects but also to

improve the life expectancy of prostate cancer (PCa) patient. Fluorescence-guided surgery (FGS)

has emerged as a technique that uses fluorescence to highlight cancerous cells and guide

surgeons to resect tumors in real-time. Thus, development of tumor-specific near-infrared (NIR)

agents that target biomarkers solely expressed on PCa cells will enable to assess negative tumor

margins and affected lymph nodes.

Experimental Design: Since PSMA is overexpressed in PCa cells in > 90% of PCa patient

population, a PSMA-targeted NIR agent (OTL78) was designed and synthesized. Optical

properties, in vitro and in vivo specificity, tumor-to-background ratio (TBR), accomplishment of

negative surgical tumor margins using FGS, pharmacokinetics (PK) properties, and preclinical

toxicology of OTL78 were then evaluated in requisite models.

Results: OTL78 binds to PSMA-expressing cells with high affinity, concentrates selectively to

PSMA-positive cancer tissues, and clears rapidly from healthy tissues with a half-time of 17 min.

It also exhibits an excellent TBR (5:1) as well as safety profile in animals.

Conclusions: OTL78 is an excellent tumor-specific NIR agent for use in fluorescence-guided

radical prostatectomy and FGS of other cancers.




3

Introduction

Prostate cancer (PCa) continues to present significant medical challenges affecting a sizeable

portion of the male population. According to the American Cancer Society, over 164,690 men

will be diagnosed with PCa in the US during 2018, leading to over 29,430 deaths (1). PCa also

has a major impact on the US economy with cumulative treatment costs estimated at $10

billion/year (2). Radical prostatectomy remains the primary therapeutic modality for patients

with localized PCa (3-4). With more than 90,000 patients undergoing prostatectomies in every

year in the US (5), 32-38% of them have BCR within 5 years (6). In fact, 20–48% of men with

PCa leave the surgery room with positive tumor margin that directly correlates to BCR and

cancer management (7-8). Therefore, it is important to excise all cancerous tissues with negative

tumor margins to improve the quality of life and life expectancy of the patient.

While removal of malignant tissues completely depends on accuracy of prognosis, there are

major limitations in current standard of care for accomplishing negative tumor margins in radical

prostatectomy. In PCa, the most common sites for BCR and positive tumor margins are known to

be the posterolateral prostate and prostatic apex (9). These are the areas closely associated with

nerves responsible for erectile function and urinary control. Since erectile dysfunction and

urinary incontinence are the major possible side effects of prostatectomy, surgeons may tend to

preserve tissues and nerves around posterolateral prostate and prostatic apex to maintain quality

of life of the patient (9). Consequently, the caution exercised in sparing these areas may lead to

diseased tissue being left behind. The standard practice for prostatectomy relies on visual

inspection and palpation during classical open surgery. Since the naked eye is more often limited

in its ability to differentiate cancer cells verses healthy cells, visual localization of tumor cells

and margins using abnormal color and/or morphology is not reliable (10). Moreover, the naked




4

eye cannot detect smaller and early stage tumors, especially tumors obscured under other healthy

tissues. Palpation, on the other hand, lacks the sensitivity to feel and distinguish the texture of

cancerous versus healthy tissue. Since robotic surgery has gained popularity (i.e. >80% of all

prostatectomies in the US are performed robotically) due to its minimally invasive and fast

recovery process (1-2 days) when compared to open surgery (10-12 days), use of palpation as a

diagnostic tool is in decline (3-4). Moreover, complete removal of affected lymph nodes using

conventional visual or palpation is unreliable as they often look or feel normal. Therefore, better

methods for assessing negative tumor margins and affected lymph nodes are needed.

In response to this unmet clinical demand, FGS has emerged as a technique that uses

fluorescence to highlight cancerous cells and guide surgeons to resect tumors in real-time.

Currently, while the field is still in its infancy and the industry slowly developing better dyes that

selectively accumulate in PCa with improved tumor-to-back ground ratios (TBR). FDA approved

indocyanine green (ICG) has been used in prostatectomy to detect PCa tissues, lymph nodes, and

vascularization of prostate (11). However, ICG has shown significant limitations with respect to

sensitivity, specificity, poor TBR, and higher liver as well as GI tract uptake due to its non-

targeted nature of the molecule (12-13). In order to overcome deficiencies of non-targeted NIR

dyes, tumor specific NIR agents that target biomarkers solely expressed on cancer cells have

been evaluated in preclinical stages. One example is PSMA that is overexpressed on PCa cells in

> 90% of PCa patient population (14). Examination of post-prostatectomy specimens has shown

that the expression level of PSMA not only correlated with tumor grade, pathologic stage, PSA

level, and aneuploidy but also with BCR (15). PSMA is also expressed in neovasculature of solid

tumors developed in organs such as liver, lung, breast, colon, renal, brain, sarcoma, gastric, and

oral (16-17). It allows internalization of PSMA-targeted agents into an endosomal compartment,




5

thereby maneuvering PSMA as an excellent biomarker for use in FGS (18). Therefore,

antibodies or small molecule ligands-targeted NIR agents that are specific for PSMA are being

currently evaluating in the pre-clinical research stages (19-26). Since each of these molecules has

their own limitations, we developed a PSMA-targeted NIR agent (referred herein as OTL78)

with superior optical, pharmacokinetic, and biological properties for use in FGS. In this paper,

we describe the synthesis & characterization, optical properties, preclinical evaluation in cancer

cell in culture and in subcutaneous & orthotopic tumor models, accomplishment of negative

surgical tumor margins using FGS, and pharmacokinetics (PK) properties of OTL78. We then

provide pre-clinical evidence of its remarkable safety profile in mice.

Materials & Methods

In vitro binding: For OTL78 relative affinity (IC50), 22Rv1 or PC3 cells were plated into a T75

flask and allowed to form a monolayer over 48h. After trypsin digestion, released cells were

transferred into centrifuge tubes (1 × 106 cells/tube) and centrifuged. Spent medium in each tube

was replaced with 100 nM DUPA-FITC in the presence of increasing concentration (0.001 nM –

10 μM) of OTL78 in fresh medium (0.5 mL). After incubating for 30 min at 4°C, cells were

rinsed with culture medium (2 x 1.0 mL) and saline (1 x 1.0 mL) to remove any unbound DUPA-

FITC. Cells were then re-suspended in saline (0.5 mL) and cell bound fluorescence was

quantified using a flow cytometer. The relative affinities were calculated using a plot of percent

cell bound fluorescence versus the log concentration of OTL78 using GraphPad Prism 6.

For OTL78 binding affinity, 22Rv1 or PC3 cells were seeded into a T75 flask and allowed to

form a monolayer over 48h. After trypsin digestion, cells were transferred into centrifuge tubes

(1 × 106 cells/tube) and centrifuged. The medium was replaced with fresh medium containing

increasing concentration of OTL78 and incubated for 30 min at 4 °C. After rinsing with fresh




6

medium (2×1.0 mL) and saline (1×1.0 mL), cells were lysed with 1% SDS with in saline (1.0

mL) and cell bound fluorescence was analyzed using a fluorometer (Cary Eclipse, Agilent

Technologies). The binding affinity (Kd) was calculated using a plot of percent cell bound

fluorescence versus concentration using GraphPad Prism 6.

Confocal Microscopy: 22Rv1, LNCaP, or PC3 cells (50,000 cells/well in 1 mL) were seeded

into poly-D-lysine microwell Petri dishes and allowed cells to form monolayers over 12 h. Spent

medium was replaced with fresh medium containing OTL78 (100 nM) and cells were incubated

for 1 h at 37 °C or 4°C. After rinsing with fresh medium (2 × 1.0 mL) and saline (1× 1.0 mL),

fluorescence images were acquired using an epi-microscopy.

Whole body Imaging & Tissue biodistribution: Seven-week-old male nu/nu mice were

inoculated subcutaneously with 5.0x106 22Rv1, LNCaP, PC3 or A549 cells/mouse in 50% high

concentrated matrigel with RPMI1640 medium on the shoulder. Growth of the tumors was

measured in perpendicular directions every 2 days using a caliper (body weights were monitored

on the same schedule), and the volumes of the tumors were calculated as 0.5×L×W2 (L=longest

axis and W=axis perpendicular to L in millimeters). Once tumors reached approximately 300 –

400 mm3 in volume, animals (3-5 mice/ group) were intravenously injected with appropriate

dose of OTL78 in saline.

For orthotopic tumors, 2 x 105 22Rv1 cells/mouse in 10% high concentrated (HC) matrigel with

RPMI1640 medium were surgically implanted in the prostate of seven-week-old male SCID

mice. Briefly, seven-week-old male SCID mice were given 1-5% isoflurane for anesthesia and

subcutaneous injection of 5 mg/kg meloxicam preoperatively for analgesia. The mice were

placed dorsal side up and washed above the prostate with a chlorhexidine scrub to ensure a




7

sterile area for incision. After an insertion was made using scalpel through the skin, the

peritoneal lining was lifted to make a small incision using a scissor and widened using forceps.

Dorsal lobes were exteriorized and gently stabilized with a wet (PBS) cotton swab. 22Rv1 cells (in 10

μL of 10% HC-matrigel) were injected the prostate using a 28-gage needle. After placing the

prostate back into the peritoneum, the abdominal wall was sutured, the body wall was closed using

3-0 or 4-0 vicryl and the skin was closed using staples. Animals were monitored until use them

for the studies. After one month, the animals were administered with OTL78 (10 nmol in 100 µL

saline per mouse), euthanized after 2 h by CO2 asphyxiation, and imaged using AMI image

system.

For whole body imaging and biodistribution studies, animals were euthanized after 2 h of

administration of OTL78 by CO2 asphyxiation. For time dependent studies, animals were imaged

under anesthesia using isoflurane. Imaging experiments were then performed using IVIS or AMI

image systems. Following whole body imaging, animals were dissected and selected tissues were

analyzed for fluorescence activity using IVIS or AMI image system and ROI of the tissues were

calculated using Living Image 4.0 software or AMIView Image Analysis Software.

For ImageJ analysis, whole body imaging was acquired in gray scale and processed in ImageJ

software. Either a line across the tumor or box around the tumor was drawn to define the

fluorescence to be quantitated. The tumor-to-muscle ratio was analyzed using a plot of the

fluorescence gray value versus distance.

Tumor surgeries: Seven-week-old male nu/nu mice were inoculated subcutaneously with

5.0x106 22Rv1 cells/mouse in 50% high concentrated matrigel with RPMI1640 medium on the

shoulder. Growth of the tumors was measured as previously mentioned. After one month, the




8

animals were mixed and divided into 2 groups (n=5 mice/group). Two hours after administering

OTL78 (10 nmol in 100 µL saline per mouse), animals were given 1-5% isoflurane for

anesthesia and imaged using AMI image system. After an insertion was made using scalpel

through the skin, surgical removal of the tumors was performed either following conventional

technique (e.g. visualization under white light or palpation) or with the aid of fluorescence

guidance (FGS: debulking of visible tumors under conventional method followed by resection of

residual fluorescence tissues under image-guided method). After the surgery, the skin was closed

using staples and imaged the mice using AMI image system. After imaging, the residual

fluorescent tissues from selected mice of conventional surgery group and tissues samples from

the tumor beds of selected mice of FGS group were submit for pathological (IHC) analysis.

Response to surgical treatment was monitored for over 30 days by imaging using AMI image

system 2 h after injecting OTL78 (10 nmol/mouse) and by measuring the growth of the tumor

volume using a caliper. Any animal with tumor volume ≥ 1000 mm3 were euthanized. Tumor-

free survival of the mice was documented as %survival vs. time using GraphPad Prism 6. IHC

studies were done as explained bellow in the Safety Studies.

Pharmacokinetic Study: For serum clearance, 10 nmol of OTL78 was administered to male

nude mice (n=3 mice) as a single bolus intravenous injection. Blood was collected at regular

intervals (0 – 90 min) and serum bound OTL78 was quantified by measuring the fluorescence

using IVIS imager. The half-life of OTL78 was calculated as %serum bound fluorescence vs.

time using GraphPad Prism 6.

For tissue clearance, 10 nmol of OTL78 was administered to male nude mice bearing 22Rv1

tumors (n=5 mice/group) as a single bolus intravenous injection. Animals were euthanized at 2,

4, 8, 24, and 48-hour time points and selected tissues were analyzed using IVIS imager. The




9

tissue clearance was determined as % tissue bound fluorescence vs. time using GraphPad Prism

6.

Note: Procedures for in vitro binding of DUPA-FITC, Human Serum Binding Studies, Safety

Study, and Tolerability Studies can be found in the Supplementary Data and Methods.

Results

Design and Synthesis of OTL78

In an effort to improve limitations in current clinical practice of radical prostatectomy and tumor-

specific imaging agents that are in the preclinical stages, OTL78 was assembled using: (i) a high

affinity PSMA-targeting ligand (coined DUPA) (19, 26), (ii) a rationally designed 14 atoms long

polyethylene glycol-dipeptide linker, and (iii) an inexpensive NIR dye (>$400/gram) named

S0456 (see SI Fig. 1a for the chemical structure). The dipeptide consisting of phenylalanine-

tyrosine was designed to fit to the contours and chemistry of the tunnel accessing the binding

pocket of the PSMA protein (27). Upon conjugation of DUPA-PEG-dipeptide to S0456, the

dipeptide is not only improved the binding affinity of OTL78 but also enhanced the fluorescence

of S0456 by ≥2 at the same concentration (see SI Fig. 1b showing excitation & emission spectra

of OTL78 at1 μM and S0456 at1 μM in 1 mL of PBS obtained using fluorometer). Synthesis of

OTL78 was conducted as shown in the SI Scheme 1 by following procedure in the SI Materials

& Methods. OTL78 was obtained as a dark green solid with >99% purity at 774 nm

wavelengths and with ~82% yield. It was characterized using 1H- &

13C-NMR, HPLC, and

HRMS (SI Fig. 2 & 3).

Evaluation of in vitro affinity and specificity of OTL78




10

In an effort to evaluate specificity of OTL78 for PCa, PSMA expression in LNCaP, 22Rv1, PC3

(three human PCa cell lines) and A549 (a human alveolar basal epithelial carcinoma cell line), as

a negative control, was first examined by flow cytometry. PSMA expression was highest in

LNCaP followed by 22Rv1 and negligible in PC3 and A549 (SI Fig. 1c). These results were

agreeing with number of PSMA molecules per LNCaP, 22Rv1, and PC3 cells reported by Wang

and colleagues (35). Due to moderate PSMA expression levels and better tumorigenic capacity

with low necrosis, 22Rv1 was selected as the primary cell line to characterize OTL78.

The affinity of OTL78 for PSMA was first screened by competing with DUPA-FITC. The

absolute binding affinity (Kd) and specificity of DUPA-FITC for PSMA (Kd = 6 nM, SI Fig. 1d)

was first established using PSMA+22Rv1 and PSMA-negative PC3 cells as described in the SI

Materials & Methods. OTL78 was able to compete with DUPA-FITC for PSMA on 22Rv1

cells with IC50 of 7 nM (SI Fig. 1e). The affinity and specificity of OTL78 was then evaluated by

incubating increasing concentrations of OTL78 with either 22Rv1 or PC3 cells and analyzing for

cell bound fluorescence by Fluorometer. OTL78 was able to bind to PSMA on 22Rv1 cells with

very high affinity (Kd = 4.7 nM) whereas it did not bind to PSMA-negative PC3 cells confirming

specificity of OTL78 to PSMA (Fig. 1b).

PSMA-mediate internalization of OTL78 was next evaluated by incubating OTL78 with 22Rv1

and PC3 cells. Analysis of fluorescence microscopy images indicate that OTL78 was able to

efficiently label 22Rv1 and LNCaP cells [(Fig. 1c (i & ii)] but not PC3 cells [(Fig. 1c (iii & vi)]

indicating PSMA-mediated uptake of OTL78. Fluorescence was detected throughout the

cytoplasm of 22Rv1 and LNCaP cells at 37 ºC. Moreover, we also observed that OTL78 is

highly concentrated and entrapped in the certain regions of 22Rv1 and LNCaP cells. Labelling of

22Rv1 and LNCaP cells with OTL78 in the presence a nuclear staining dye (DAPI) at 4 ºC was




11

also conducted to decrease the endocytosis and recycling of PSMA [SI Fig. 1 f(i & ii)]. Epi-

fluorescence images from this study indicated that OTL78 binds to PSMA on the cell surface.

Therefore, we assume that OTL78 first binds to PSMA on the cell surface and then it undergoes

receptor-mediated endocytosis. We further assume that OTL78 is entrapped in the acidic

endosomes within PCa cells.

Evaluation of in vivo efficacy and specificity of OTL78

The ability of OTL78-mediated imaging of PCa was next established by conducting a series of

experiments in mouse models. First, the optimal dose for tumor imaging was determined by

administering increasing concentrations of OTL78 (0.3–120 nmol/mouse) to mice bearing 22Rv1

tumor xenografts followed by ex vivo tissue biodistribution analysis. The IVIS image analysis

obtained at 2 hour time point indicated that OTL78 provided excellent TBR at dose range

between 1-30 nmol per mouse with the best TBR occurring at ~3-10 nmol/mouse (SI Fig 4a-b).

We next evaluated in vivo tumor specificity of OTL78 by administering 10 nmol of OTL78 to

mice bearing subcutaneous 22Rv1, LNCaP, PC3 or A549 tumor xenografts followed by

conducting whole body imaging and ex vivo tissue biodistribution using either IVIS or AMI

image systems. Both studies demonstrated that OTL78 accumulated predominantly in PSMA

expressing 22Rv1 (Fig. 2a, d and Fig. 3a, d) and LNCaP (Fig. 3b, e) tumors, with no substantial

fluorescence activity in other tissues except kidneys. Although tumor accumulated fluorescence

was not seen in PC3 and A549 tumors at higher threshold (Fig. 2b-c & e-f), uptake of OTL78

was observed in both tumors at lower threshold (Fig. 2 e-f: Lower panel). While fluorescence

intensities of PC3 and A549 tumors were ~6 folds less compared to 22Rv1 tumors (Fig. 2a, d),

fluorescence accumulation in PC3 and A549 tumors was higher than rest of the tissues except

kidneys and skin (SI Fig. 5a-b). We therefore assume that the observed fluorescence in PC3 and




12

A549 tumors maybe due to accumulation of OTL78 via PSMA in the neovasculature of PC3 and

A549 solid tumors. This further suggests that OTL78 will be able to detect tumors with low

PSMA expression levels. OTL78 also had a significant kidney uptake due to high PSMA

expression in murine kidneys and clearance of OTL78 through the kidneys. More importantly,

fluorescence in the kidneys were clearly visible in whole body images collected from AMI

imager demonstrating penetrating ability of OTL78 to locate buried PSMA+ tissues. We assume

that observed skin uptake maybe due to non-specific uptake of S0456 moiety of OTL78

molecule. Although skin uptake clears within 4 – 5 h, skin will not be interfered with open or

robotic surgery because the camera will be directly focusing to the prostate in both techniques.

We then examined the ability of OTL78 to detect primary tumors in the prostate and regional

metastasis in seminal vesicles. In that case, 22Rv1 cells were surgically implanted in the prostate

of SCID mice as described in the SI Materials & Methods. Once tumors grow, the animals were

imaged using AMI image system 2 h after administering of 10 nmol of OTL78. Orthotopic

imaging studies also demonstrated that OTL78 mainly accumulated in prostate tumors with no

fluorescence observed in other tissues except kidneys (Fig. 3c, f & SI Fig.5c-d). Moreover,

OTL78 was able to detect local regional metastasis in seminal vesicles in the presence of primary

tumor (Fig. 3 g & SI Fig. 5d) indicating ability of OTL78 locate tumors and lymph nodes that

are buried under the prostate.

Following biodistribution studies, specificity of OTL78 for PSMA was quantitated by calculating

TBR. In both subcutaneous and orthotopic tumor models, OTL78 displayed excellent TBR (SI

Fig. 6a) ranging from 19:1 – 25:1 (tumor:muscle), 11:1-14:1 (tumor:lung), 11:1-15:1

(tumor:liver), 14:1-23:1 (tumor:heart), 19:1 (tumor:intestine), 11:1-20:1 (tumor:spleen), 4:1

(tumor:prostate), and 4:1-10:1 (tumor:skin). Observed better TBRs, especially tumor:skin, in




13

orthotopic model compared to subcutaneous model may be due: (a) higher accumulation of

OTL78 due to better tumor angiogenesis and (b) less non-specific skin uptake of NIR dye moiety

in SCID mice.

Finally, the ability of OTL78 to define the tumor/healthy tissue boundaries was evaluated using

ImageJ software analysis. The whole-body image of mice injected with 10 nmol of OTL78 was

acquired as fluorescence in a gray scale and either a line or box (SI Fig. 6b) was drawn to

quantitate the fluorescence to be defined in the tumor boundaries. As shown in SI Fig. 6c-d,

OTL78 was able to define tumor boundaries precisely with a TBR of 5:1 suggesting its

capability to guide surgeons to accurately detect the tumor margins (acceptable TBR for image-

guided surgery is considered to be >1.5) (28).

Fluorescence-Guided Surgery Using OTL78

The ability of OTL78 to guide surgeons to excise all cancerous tissues with negative tumor

margins was next investigated by performing image-guided surgery in tumor bearing mice.

Briefly, 10 nmol of OTL78 was administered into mice bearing 22Rv1 tumor xenografts and

comparative study was conducted by performing surgeries under conventional (e.g. visualization

under white light or palpation) or fluorescence-guided technology (i.e. debulking of visible

tumors under conventional method followed by resection of residual fluorescence tissues under

image-guided method) at 2 h time point. Preoperative fluorescence images of tumor bearing

mice demonstrated that OTL78 able to localized 22Rv1 tumors with high contrast within 2h

(Fig. 4a: first column and SI Fig. 7). Postoperative fluorescence imagers indicated presence of

residual fluorescent in the tumor bed of the conventional cohorts, whereas no significant

fluorescence was observed in the FGS cohorts (Fig. 4a: middle column and SI Fig. 7).

Pathological analysis of residual fluorescent tissues from the conventional surgery confirmed




14

that the fluorescence is due to cancer cells (Fig. 4a-b: middle panel). More importantly, no

residual tumors were identified in tissues from tumor margin/bed from the FGS cohorts (Fig. 4b:

right column). Following surgeries, biochemical recurrence (BCR) of the cancer was assessed

by monitoring animals for over a month using fluorescence imaging. As anticipated, only the

conventional surgery cohort had recurrence at the primary tumor site and no sign of BCR was

observed in the FGS cohort during the study (Fig. 4a & SI Fig. 7). As shown in the survival

curve (Fig.4c), the FGS cohorts were survived during the study with no BCR whereas all mice in

the conventional surgery group had to euthanize within 3 weeks. Although the observed BCR

rate is higher than the reported values for human and mice (6, 36), this proof of concept study

highlights the importance of excising all cancerous tissues with negative tumor margins to

improve the quality of life and life expectancy of the patient.

Evaluation of Pharmacokinetic Properties of OTL78

Having demonstrated in vivo specificity, we then evaluated the PK profile of OTL78 in animal

models as described in the Methods. OTL78 was able to generate excellent tumor images within

1h of post-injection, with TBR ratios remaining excellent throughout the experiment time of 48

hours (Fig. 5a & SI Fig. 8). Moreover, time dependent tumor clearance studies demonstrated

that ~50% of fluorescence was retained in the tumor at 48 hours post-injection (Fig. 5b),

suggesting that OTL78 fluorescence was indeed entrapped in the tumor. OTL78 was mainly

excreted through the kidneys, and fluorescence of the kidneys became negligible after 8 hours

(Fig. 5b). We also elected to monitor the skin clearance due to observed high fluorescence in the

skin at the 2 hour time point. While skin uptake may not affect the outcome of open or robotic-

assisted FGS, as shown in the Fig. 5b, OTL78 cleared from the mice skin between 4 - 5 hours of

post-injection.




15

Finally, serum clearance of OTL78 was examined by injecting 10 nmol/mouse intravenously,

collecting blood samples at regular intervals (0–90 min), and measuring serum-bound OTL78

using IVIS imager. OTL78 reached a peak concentration in the circulation ∼10 min post-

injection and cleared with a serum half-life of ~17 min (Fig. 5c). Human serum binding studies

conducted using LC/MS analysis also indicated that OTL78 has very low affinity (32%) for

human serum proteins (SI Fig. 9). Taken together, OTL78 clears rapidly from nonmalignant

tissues to allow tumor visualization within 1-2 hours of administration, avoiding the requirement

for a prolonged hospital stay.

Evaluation of Safety Profile of OTL78

Motivated by the specificity and PK properties described above, safety profile of OTL78 was

then evaluated using ex vivo and in vivo models. The acute maximum tolerance dose of OTL78

was initially determined by injecting 6 μmol/mouse (600x of normal dose) to healthy Balb/c

mice. Body weights and clinical observations were monitored during the study and

histopathological analysis on selected tissues was then conducted on day 14 of post-injection.

The animals were active after administration of OTL78 and behaved normally throughout the

study. As shown in the Fig. 6a, body weights over the course of the study remained unchanged

(<5% increase), suggesting that OTL78 is not grossly toxic to the animals. Moreover, no obvious

pathological changers were detected in hematoxylin and eosin (H&E) staining conducted on any

of the tissues (Fig. 6b and SI Fig. 10). No noticeable toxicities were also noticed in clinical

pathology analysis on blood samples collected from mice injected with OTL78 (6 μmol/mouse).

Possible OTL78-related hypersensitivity in human was next examined using basophil activation

assay. Drug related hypersensitivity is mainly due to immune response caused by cross-linking

of immunoglobulin E (IgE) expressed on mast cells and basophils (Fig. 6d) resulting in




16

activation and subsequent degranulation to release vasoactive amines, prostaglandins, and

cytokines (29). Since cross-linking of IgE can be due to aggregates, concentration dependent UV

spectrometric studies were conducted to determine higher order aggregates of OTL78. As shown

in the Fig. 6c, there were no noticeable higher order aggregates observed with OTL78 whereas

the positive control (i.e. OTL38) (30) exhibited concentration dependent aggregation peak at λmax

~700 nm at 75 μM in saline.

Since basophils are readily available from blood samples when compared to tissue-resident mast

cells, we then evaluated drug-related hypersensitivity due to monomer and low order aggregates

(if present) of OTL78 using basophil activation test in human blood samples as described in the

Method section. Briefly, 75 μM of OTL78 was first added to a tube containing whole blood

from donors and stimulating buffer. After labeling with anti-CCR3 (CD193)-phycoerythrin and

anti-CD63-CD203c-PE-DY647, the percentage of activated basophils was quantitated using flow

cytometric analysis (31). As shown in the Fig. 6e and SI Table 1, no obvious differences in

percentage activated basophils were seen between the OTL78 treated sample and negative

control resulting in stimulated index of 1, whereas stimulated index is defined as the ratio

of %basophil activation by the allergen: %basophil activation by background and stimulated

index ≥ 2 considered as positive response (31). However, when similar assays were conducted

using fMLP (a non-specific basophil activator) or anti-FcεR antibody, positive response of

73.5% (stimulated index = 29.5) or 6.49% (stimulated index = 2.6) was observed.

Discussion

The objective of the study was to introduce a PCa-specific NIR agent that could assist surgeons

to conduct radical prostatectomies with negative tumor margins. Conventional prognostic

modalities such as visual inspection and palpation may lead to BCR as a result of incomplete




17

tumor removal or to erectile dysfunction and/or urinary incontinence due to nerve damages.

Thus, we developed a PSMA-targeted NIR imaging agent (OTL78) that: (i) binds to PSMA+

tumors with high affinity and specificity, (ii) allows to use sub-nanomolar concentration to

visualize small tumors, (iii) clears rapidly from PSMA-negative tissues with half-life of 17 min,

allowing for real time FGS within 1 - 2 hours of post-injection, (iv) retains tumor fluorescence

for over 48 hours, allowing visualization throughout FGS, (v) allows to accomplish negative

surgical tumor margins, and (vi) has an excellent safety profile in animals. Moreover, OTL78

can be readily synthesized in multigram quantities at low cost (<$500/gram) with high purity and

excellent yield.

While tumor-specific NIR dyes have not yet entered to the clinic for use in radical

prostatectomy, several PSMA-targeted NIR dyes are currently under preclinical development

(20-25). Unfortunately, antibody-targeted dyes such as J591-ICG have taken ~2 days to obtain

acceptable tumor contrast in PSMA transfected PCa tumor xenograft model (20). The observed

slow clearance from the nonmalignant tissues and longer optimal imaging time necessitate for a

prolonged hospital stay or multiple hospital visits leading to higher cumulative cost. On the other

hand, NIR dyes are often non-specifically conjugated to antibodies using Lys or Cys sites that

lead to heterogeneous chemical entities resulting in variable affinities, efficacies, PK and safety

profiles. Moreover, it’s well established that Cys-S-maleimide (i.e. thio-ether) bond is unstable

during the circulation and tend to undergo retro-Michael reaction (β-elimination) and oxidation

leading to release thiol and maleimide adducts that eventually lead to poor TBR (32-33).

Therefore, production cost of these antibod-NIR conjugates can be higher when compared to

small molecular ligands. In contrast, small molecule ligands (Mr >0.5 Da) penetrate solid tumors




18

rapidly, clears from PSMA-negative tissues in < 2 hours, show high TBR, easy to synthesis, and

stable during the synthesis and storage.

Despite all the advantages those small molecule ligands have, development of NIR dye that

maintains the properties of an ideal fluorescent contrast agent can be challenging. While

investigators claimed to observe acceptable tumor contrast within 6 hours, IR800CW-YC-27 (a

small molecule ligand-targeted IR800CW) has taken ~ 20 hours to obtain optimal tumor images

in PSMA transfected PCa tumor xenograft model and 72 hours to clear from non-targeted tissues

(22). This will also cause for an extended hospital stay and to higher cumulative cost.

Furthermore, IR800CW-YC-27 has also demonstrated a substantial amount of non-specific

fluorescence uptake in PSMA-negative tumor xenograft. On the other hand, IR800CW is an

expensive asymmetric NIR dye could cost $30,000/ gram whereas cost of S0456 is <$400/gram.

Moreover, as established in the literature, IR800CW has additional disadvantage over other NIR

dyes: (a) hydrolysis of N-hydroxysuccinimide (NHS) ester during the synthesis, and (b)

formation of unwanted enamine byproduct due to replacement of 4-hydroxybenzensulfonate

during the reaction with ligand-linker-amine (i.e. in the presence of amines) (34). Formation of

undesired byproducts may result to perform complex purifications, higher production cost, and

longer waiting period for clinical translation.

Due to its high affinity and specificity for PSMA expressing tumors, rapid clearance from

PSMA-negative healthy tissues, excellent safety profile, and ability to synthesize in gram scale at

low cost, OTL78 has proven to be a clinical candidate to yield sharp tumor boundaries with

negative tumor margins within 1 - 2 hours of infusion during radical prostatectomy. With its

recent entry into IND-enabling studies, OTL78 has a potential to become the first PSMA-

targeted NIR agent to enter into the clinic for use in fluorescence-guided radical prostatectomy.




19

Acknowledgements: The authors thank Xin Liu at Purdue Drug Discovery Center for insights

on docking images. The authors also thank Dr. Tiffany Lyle at Purdue University College of

Veterinary Medicine for insights on pathological analysis and Dr. Gert J. Breur for insights on

orthotopic tumor implantations.

Author contributions: S.A.K conceived the concept, designed the experiments, analyzed data,

and wrote the manuscript. MT and C.H.M conducted in vitro and in vivo studies, acquired and

analyzed data. P.G. and A.K.K optimized manufacturing procedures, synthesized and

characterized OTL78. C.J.C. conducted orthotopic tumor surgeries and monitoring tumor

development. B.N.C. performed fluorescence-guided surgery and monitoring tumor recurrence.

Competing financial interests: S.A.K., C.H.M., M.T., P.G., A.K.K are employed by On Target

Laboratories. C.J.C. and B.N.C. are employed by Purdue University. No competing financial

interests were disclosed by all authors.

Supplementary Data and Methods: Supporting Information for this article including

supplementary Figures & Legends, Tables, Materials & Methods, synthetic procedures,

quantities and yield, and additional in vitro & in vivo experimental procedures are available in

the online version of the paper.




20

Reference

1) Siegel RL, Miller KD, Jemal A. Cancer statistics. 2018. CA Cancer J Clin. 2018, 68: 7-30.

2) Roehrborn CG, Black LK. The economic burden of prostate cancer. BJU Int. 2011, 108: 806-

13.

3) Rocha R, Fiorelli RK, Buogo G, Rubistein M, Mattos RM, Frota R, Coelho RF, Palmer K,

Patel V. Robotic-assisted laparoscopic prostatectomy (RALP): a new way to training. J

Robot Surg. 2016, 10: 19-25

4) Laviana AA, Williams SB, King ED, Chuang RJ, Hu JC. Robot assisted radical

prostatectomy: the new standard? Minerva Urol Nefrol. 2015, 67: 47-53

5) Strassberg DS, Zavodni SM, Gardner P, Dechet C, Stephenson RA, Sewell KK. Quality of

Life Following Prostatectomy as a Function of Surgery Type and Degree of Nerve Sparing.

Curr Urol. 2017, 11: 16-20.

6) Grossfeld GD, Latini DM, Lubeck DP, Mehta SS, Carroll PR. Predicting recurrence after

radical prostatectomy for patients with high risk prostate cancer. J Urol 2003, 169: 157–63.

7) Iczkowski KA, Lucia MS. Frequency of positive surgical margin at prostatectomy and its

effect on patient outcome. Prostate Cancer. 2011, 2011: 673021.

8) Servoll E, Vlatkovic L, Sæter T, Nesland JM, Axcrona U, Waaler G, Axcrona K. The length

of a positive surgical margin is of prognostic significance in patients with clinically localized

prostate cancer treated with radical prostatectomy. Urol Int. 2014, 93: 289-95.

9) Sopko NA, Burnett AL. Erection rehabilitation following prostatectomy--current strategies

and future directions. Nat Rev Urol. 2016, 13: 216-25.

10) Nagaya T, Nakamura YA, Choyke PL, Kobayashi H. Fluorescence-Guided Surgery. Front

Oncol. 2017, 7: 314.




21

11) Mangano MS, De Gobbi A, Beniamin F, Lamon C, Ciaccia M, Maccatrozzo L. Robot-

assisted nerve-sparing radical prostatectomy using near-infrared fluorescence technology

and indocyanine green: initial experience. Urologia. 2018, 85: 29-31.

12) Choi HS, Nasr K, Alyabyev S, Feith D, Lee JH, Kim SH, Ashitate Y, Hyun H, Patonay G,

Strekowski L, Henary M, Frangioni JV. Synthesis and in vivo fate of zwitterionic near-

infrared fluorophores. Angew Chem Int Ed Engl. 2011, 50: 6258-63.

13) Kobayashi H, Ogawa M, Alford R, Choyke PL, Urano Y. New strategies for fluorescent

probe design in medical diagnostic imaging. Chem Rev. 2010, 110: 2620-40.

14) Evans JC, Malhotra M, Cryan JF, O'Driscoll CM. The therapeutic and diagnostic potential of

the prostate specific membrane antigen/glutamate carboxypeptidase II (PSMA/GCPII) in

cancer and neurological disease. Br J Pharmacol. 2016, 173: 3041-3079.

15) Chang SS. Overview of prostate-specific membrane antigen. Rev Urol. 2004, 6 Suppl 10:

S13-8.

16) Mahalingam D, Wilding G, Denmeade S, Sarantopoulas J, Cosgrove D, Cetnar J, Azad N,

Bruce J, Kurman M, Allgood VE, Carducci M. Mipsagargin, a novel thapsigargin-based

PSMA-activated prodrug: results of a first-in-man phase I clinical trial in patients with

refractory, advanced or metastatic solid tumours. Br J Cancer. 2016, 114: 986-94.

17) Kiess AP, Banerjee SR, Mease RC, Rowe SP, Rao A, Foss CA, Chen Y, Yang X, Cho SY,

Nimmagadda S, Pomper MG. Prostate-specific membrane antigen as a target for cancer

imaging and therapy. Q J Nucl Med Mol Imaging. 2015, 59: 241-68.

18) Liu H, Rajasekaran AK, Moy P. Constitutive and antibodyinduced internalization of prostate-

specific membrane antigen. Cancer Res. 1998, 58: 4055–4060




22

19) Kularatne SA, Wang K, Santhapuram HK, Low PS. Prostate-specific membrane antigen

targeted imaging and therapy of prostate cancer using a PSMA inhibitor as a homing ligand.

Mol Pharm. 2009, 6: 780-9.

20) Nakajima T, Mitsunaga M, Bander NH, Heston WD, Choyke PL, Kobayashi H. Targeted,

activatable, in vivo fluorescence imaging of prostate-specific membrane antigen (PSMA)

positive tumors using the quenched humanized J591 antibody-indocyanine green (ICG)

conjugate. Bioconjug Chem. 2011, 22: 1700-5.

21) Watanabe R, Sato K, Hanaoka H, Harada T, Nakajima T, Kim I, Paik CH, Wu AM, Choyke

PL, Kobayashi H. Minibody-indocyanine green based activatable optical imaging probes: the

role of short polyethylene glycol linkers. ACS Med Chem Lett. 2014, 5: 411-5.

22) Chen Y, Dhara S, Banerjee SR, Byun Y, Pullambhatla M, Mease RC, Pomper MG. A low

molecular weight PSMA-based fluorescent imaging agent for cancer. Biochem Biophys Res

Commun. 2009, 390: 624-9.

23) Wang X, Huang SS, Heston WD, Guo H, Wang BC, Basilion JP. Development of targeted

near-infrared imaging agents for prostate cancer. Mol Cancer Ther. 2014, 13: 2595-606.

24) Bao K, Lee JH, Kang H, Park GK, El Fakhri G, Choi HS. PSMA-targeted contrast agents for

intraoperative imaging of prostate cancer. Chem Commun (Camb). 2017, 53: 1611-1614.

25) Kelderhouse LE, Chelvam V, Wayua C, Mahalingam S, Poh S, Kularatne SA, Low PS.

Development of tumor-targeted near infrared probes for fluorescence guided surgery.

Bioconjug Chem. 2013, 24:1075-80.

26) Kularatne SA, Zhou Z, Yang J, Post CB, Low PS. Design, synthesis, and preclinical

evaluation of prostate-specific membrane antigen targeted (99m)Tc-radioimaging agents.

Mol Pharm. 2009, 6: 790-800.




23

27) Mesters JR, Barinka C, Li W. Structure of glutamate carboxypeptidase II, a drug target in

neuronal damage and prostate cancer. EMBO J. 2006, 25: 1375–1384.

28) Gulec SA. PET probe-guided surgery. J Surg Oncol. 2007, 96: 353-7.

29) Johansson SG. The History of IgE: From discovery to 2010. Curr Allergy Asthma Rep. 2011,

11: 173-7.

30) Boogerd LSF, Hoogstins CES, Gaarenstroom KN, de Kroon CD, Beltman JJ, Bosse T,

Stelloo E, Vuyk J, Low PS, Burggraaf J, Vahrmeijer AL. Folate receptor-α targeted near-

infrared fluorescence imaging in high-risk endometrial cancer patients: a tissue microarray

and clinical feasibility study. Oncotarget. 2017, 9: 791-801.

31) Boumiza R, Debard AL, Monneret G. The basophil activation test by flow cytometry: recent

developments in clinical studies, standardization and emerging perspectives. Clin Mol

Allergy. 2005, 3: 9.

32) Fontaine SD, Reid R, Robinson L, Ashley GW, Santi DV. Long-term stabilization of

maleimide-thiol conjugates. Bioconjug Chem. 2015, 26: 145-52.

33) Shen BQ, Bumbaca D, Saad O, Yue Q, Pastuskovas CV, Khojasteh SC, Tibbitts J, Kaur S,

Wang B, Chu YW, LoRusso PM, Girish S. Catabolic fate and pharmacokinetic

characterization of trastuzumab emtansine (T-DM1): an emphasis on preclinical and clinical

catabolism. Curr Drug Metab. 2012, 13: 901-10.

34) Hyun H, Owens EA, Narayana L, Wada H, Gravier J, Bao K, Frangioni JV, Choi HS, Henary

M. Central C-C Bonding Increases Optical and Chemical Stability of NIR Fluorophores.

RSC Adv. 2014, 4: 58762-58768.




24

35) Wang X, Ma D, Olson WC, Heston WD. In vitro and in vivo responses of advanced

prostate tumors to PSMA ADC, an auristatin-conjugated antibody to prostate-specific

membrane antigen. Mol Cancer Ther. 2011, 10:1728-39.

36) Neuman BP, Eifler JB, Castanares M, Chowdhury WH, Chen Y, Mease RC, Ma R,

Mukherjee A, Lupold SE, Pomper MG, Rodriguez R. Real-time, near-infrared

fluorescence imaging with an optimized dye/light source/camera combination for surgical

guidance of prostate cancer. Clin Cancer Res. 2015, 21:771-80.




25

Figures Legends

Figure 1. In vitro binding and specificity of OTL78. (a) Excitation (Ex) & emission (Em) spectra

of OTL78. (b) Dose dependent binding of OTL78 to prostate-specific membrane antigen

(PSMA)+ 22Rv1 cells and PSMA-negative PC3 cells in culture (n=2). (c) Binding and

internalization of OTL78 to (i) 22Rv1, (ii) LNCaP, or (iii) PC3 (fluorescence image) and (iv)

PC3 (DIC image) cells by epifluorescence (epi) microscopy. Note: OTL78 is highly concentrated

in the acidic endosomes of 22Rv1 and LNCaP cells. DIC = Deferential Interference Contrast

Images.

Figure 2. In vivo efficacy and specificity of OTL78 in subcutaneous tumor models using IVIS

image system. Representative fluorescence images from IVIS imager showing mice bearing (a)

22Rv1 (n=5 mice/group), (b) PC3 (n=5 mice/group), and (c) A549 (n=3 mice/group) tumors 2 h

after administering 10 nmol of OTL78. Tissue biodistribution analysis of the same mice with (d)

22Rv1, (e) PC3, and (f) A549 tumors at 2 h post-injection. Note: * Representative fluorescence

images of PC3 and A549 after lowering threshold to ~ 1 x 108 [(p/sec/cm

3/sr)/(µW/cm

2)].

Figure 3. In vivo efficacy and specificity of OTL78 in orthotopic and subcutaneous tumor

models using AMI image system. Representative fluorescence images from AMI image system

showing mice bearing (a) 22Rv1 subcutaneous (n=3mice/group), (b) LNCaP subcutaneous (n=3

mice/group), and (c) 22Rv1 orthotopic (n=5 mice/group) tumors 2 h after administering 10 nmol

of OTL78. Tissue biodistribution analysis of the same mice with (d) 22Rv1, (e) LNCaP, and (c)




26

22Rv1 tumors at 2 h post-injection. Note: *Primary tumor is in the prostate in Fig (f) and K =

Kidneys. Note: PT = Primary Tumor, SC = Secondary Tumor, & SV = Seminal Vesicle

Figure 4. Comparison of surgeries performed under conventional and fluorescence-guided

techniques. (a) Representative fluorescence images of tumor beds of mice before and after

surgically removing 22Rv1 tumor xenografts by conventional (n=5 mice/group) or fluorescence-

guided (n=5 mice/group) techniques. Mice were administered with OTL78 (10 nmol/mouse) 2 h

before imaging with AMI image system. (b) Representative H&E staining of 22Rv1tumor (left

column) after surgical resection, the residual fluorescent tissues after conventional surgery

showing positive tumor margins (middle column), and tumor bed tissues after FGS showing

negative tumor margins. (c) Survival curve of the same mice (n=5 mice/group) over 30 days.

Growth of tumors was monitored during the study and any animal with tumor volume ≥ 1000

mm3 were euthanized.

Figure 5. Pharmacokinetics profile of OTL78. (a) Representative time dependent whole body

fluorescence images over white light images of a mouse bearing 22Rv1 tumor after injecting 10

nmol of OTL78 and image with IVIS imager at different time intervals (n=5 mice/group). (b)

Clearance of OTL78 from tumor, kidney, and skin from time dependent biodistribution analysis.

(c) Determination of half-life of OTL78 by time dependent serum analysis. Error bars represents

SD (n=3 mice/group). Note: K = Kidneys

Figure 6. Safety profile of OTL78. (a) Assessment of body weight change after administering 6

μmol (i.e. 600x of normal dose) of OTL78 to healthy balb/c mice and (b) representative H&E

staining of kidney and prostate of mouse injected with 6 μmol of OTL78 at 14 days post-




27

injection (n = 5 mice/group). (c) UV spectra of OTL78 showing no aggregates whereas the

positive control (OTL38) demonstrating >50% higher aggregates at 75 μM concentration in

saline. (d) Possible mechanism for drug related hypersensitivity reactions due to activation of

basophils and mast cells. (e) Evaluation of drug- related hypersensitivity in human blood samples

using basophil activation assay by flow cytometry. fMLP: N-formylmethionyl-leucyl-

phenylalanine is a non-specific cell activator, anti-FcεR: a high affinity monoclonal antibody

binding to IgE, CCR3 (CD193): specific biomarker on basophils, CD63 and CD203c: receptors

that upregulated upon activation of basophils, PE: phycoerythrin, background: negative control,

and CD63-CD203c-PE-DY647+/ CCR3-PE

+ (Q2) cell population considered as the positive

response for basophil activation.




Figure 1




Figure 2




Figure 3




Figure 4




Figure 5




Figure 6




Published OnlineFirst September 10, 2018.Clin Cancer Res Sumith A Kularatne, Mini Thomas, Carrie H Myers, et al. Fluorescence-Guided Surgery for Prostate CancerAntigen-Targeted Near-Infrared Imaging Agent for Evaluation of Novel Prostate-Specific Membrane

Updated version

10.1158/1078-0432.CCR-18-0803doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2018/09/08/1078-0432.CCR-18-0803.DC1

Access the most recent supplemental material at:

Manuscript

Authoredited. Author manuscripts have been peer reviewed and accepted for publication but have not yet been

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/early/2018/09/08/1078-0432.CCR-18-0803To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-18-0803

http://clincancerres.aacrjournals.org/content/suppl/2018/09/08/1078-0432.CCR-18-0803.DC1

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/early/2018/09/08/1078-0432.CCR-18-0803


home | clinical cancer research - evaluation of novel prostate … · 1 evaluation of novel...

Documents