Cytotherapy, 2014; 16: 17e32

ORIGINAL PAPERS

Conditioned medium from amniotic membrane-derived cells preventslung fibrosis and preserves blood gas exchanges in bleomycin-injuredmice—specificity of the effects and insights into possible mechanisms

ANNA CARGNONI1, ESTER COTTI PICCINELLI1, LORENZO RESSEL1,2,DANIELE ROSSI1, MARTA MAGATTI1, IVAN TOSCHI3, VALENTINA CESARI3,MARIANGELA ALBERTINI4, SILVIA MAZZOLA4 & ORNELLA PAROLINI1

1Centro di Ricerca E. Menni, Fondazione Poliambulanza-Istituto Ospedaliero, Brescia, Italy, 2School of VeterinaryScience, University of Liverpool, Neston, United Kingdom, 3Dipartimento di Scienze Agrarie e Ambientali, Università diMilano, Milano, Italy, and 4Dipartimento di Scienze Veterinarie e Sanità Pubblica, Università di Milano, Milano, Italy

AbstractBackground and aims. We recently demonstrated that injection of conditioned medium (CM) generated from cells of themesenchymal region of human amniotic membrane (AMTCs) reduces bleomycin-induced lung fibrosis in mice, suggesting acrucial role of paracrine factor(s) secreted by AMTCs in these beneficial effects. We further investigated this hypothesis, themechanisms involved, the effects on some lung functional parameters and whether AMTC-secreted effector(s) are specificto these cells and not produced by other cell types, extending the time of analysis up to 28 days after treatment. Methods.Bleomycin-challenged mice were either treated with AMTC-CM or CM generated from human skin fibroblasts, human pe-ripheral blood mononuclear cells or Jurkat cells, or were left untreated. Mouse lungs were analyzed for content of pro-inflammatory and pro-fibrotic molecules, presence of lymphocytes andmacrophages and for fibrosis level (through histologicalsemi-quantitative evaluation and quantitativemeasurement of collagen content). Arterial blood gas analysis was also performed.Results.Up to 28 days after delivery, AMTC-CM-treatedmice developed reduced lung fibrosis with respect tomice treated withother CM types. AMTC-CM-treated mice had comparatively better preservation of blood gas parameters and showed lowerlung content of interleukin-6, tumor necrosis factor-a, macrophage inflammatory protein-1a, monocyte chemoattractantprotein-1 and transforming growth factor-b associated with reduced lungmacrophage levels. Conclusions. AMTC-CM preventslung fibrosis in bleomycin-challenged mice, improving survival and preserving lung functional parameters such as blood gasexchanges. The specificity of AMTC-CM action was indicated by the absence of fibrosis reduction when other CM types wereused. Finally, we provide some insights into the possible mechanisms underlying AMTC-CM-mediated control of fibrosis.

Key Words: amniotic membrane-derived cells, amniotic mesenchymal tissue cells, conditioned medium, human term placenta, lungfibrosis, mesenchymal stromal cell

Introduction

We and others have previously described in vitrostudies that show the modulatory ability of cellsderived from the amniotic membrane both on T lym-phocytes (1e3) and on antigen-presenting cells (4).

In vivo studies have also demonstrated anti-inflammatory and anti-fibrotic properties of placenta-derived cells (5e7). In particular, we have shown thathuman placental fetal membrane-derived cells—specifically, a mixture composed of cells from themesenchymal layer of the human amniotic and

Correspondence: Ornella Parolini, PhD, Centro di Ricerca E. Menni, FondaziItaly. E-mail: ornella.parolini@tin.it

(Received 18 December 2012; accepted 8 July 2013)

ISSN 1465-3249 Copyright � 2014, International Society for Cellular Therapy. Phttp://dx.doi.org/10.1016/j.jcyt.2013.07.002

chorionic membranes and cells from the amnioticepithelial layer—were able to reduce the progression oflung fibrosis when transplanted into bleomycin-chal-lenged mice (8). Furthermore, in a similar animalmodel, Moodley et al. (9) reported that treatmentwith human amniotic epithelial cells (hAECs) reducedlung inflammation and fibrosis and that these cells,when engrafted in host lungs, differentiated into cellswith an alveolar epithelial phenotype. More recently,Murphy et al. (10) confirmed the anti-inflammatoryand anti-fibrotic effects of hAECs, although these

one Poliambulanza-Istituto Ospedaliero, via Bissolati 57, I-25124 Brescia,

ublished by Elsevier Inc. All rights reserved.

18 A. Cargnoni et al.

authors did not detect significant cell engraftment inthe lungs of transplanted mice; meanwhile, Vosdo-ganes et al. (11) found that hAECs that had beeninjected into a sheep fetus affected by intra-amnioticlipopolysaccharide-induced pulmonary inflammationwere able to cause a reduction in levels of lung in-flammatory cytokines.

Despite these evident beneficial effects ofplacental cells on lung fibrosis, there have been noclear reports to date on the mechanism(s) throughwhich these cells act. The fact that cell treatment ef-ficacy has also been observed in the absence of sig-nificant levels of placenta-derived cell engraftment(8,10) suggests that these cells could work througha paracrine mechanism by releasing soluble factorsthat are able to exert trophic actions on host lungcells and modulatory actions on inflammatory cellsrecruited to the injury site. Indeed, our grouphas reported strong evidence in support of such amechanism by administering conditioned mediumderived from the culture of cells isolated from themesenchymal region of human amniotic membrane(amniotic mesenchymal tissue cells, AMTCs) tobleomycin-challenged mice, rather than adminis-tering the cells themselves (12). Even in the absenceof cells, mice treated with AMTC-derived condi-tioned medium (AMTC-CM) developed less pro-nounced lung fibrosis in terms of fibrosis distribution,fibroblast proliferation, collagen deposition and alve-olar obliteration when compared with untreated mice.

However, several questions remain open in re-gard to the beneficial effects exerted by AMTC-CMon pulmonary fibrosis, such as their ultimateoutcome on pulmonary function, the mechanism(s)through which these effects are exerted and whetherthese positive effects are specific to AMTC-CM, or ifthey are instead also exerted by CM obtained fromother cell types of different nature.

In the current study, with the use of the samemurine model of bleomycin-induced lung fibrosisdescribed previously (8,12), we addressed some ofthese open questions, thus going beyond our previ-ous results (8,12), as well as further exploring theproperties of AMTC-CM. In particular, besidesconfirming the ability of AMTC-CM to reduce theseverity and progression of lung fibrosis, weexplored these effects for up to 28 days after deliveryand examined the potential of AMTC-CM in pro-tecting pulmonary function from impairmentinduced by fibrotic lesions. In addition, in attempt-ing to elucidate the mechanism(s) that may underliethe beneficial effects of AMTC-CM, we also quan-tified the inflammatory and pro-fibrotic cytokinelevels in the lungs of bleomycin-challenged micethat had been treated with AMTC-CM and evalu-ated T-lymphocyte and macrophage infiltration in

the mouse lungs. Finally, for all of these experi-ments, we investigated whether the paracrine effec-tor(s) secreted by AMTCs are specific to these cellsby comparing the effects of AMTC-CM with thoseexerted by conditioned media generated fromother cell types sharing a similar mesodermal originwith AMTCs and that were representative of bothnon-adherent hematopoietic and adherent non-hematopoietic (stromal) cell lineages, namely, hu-man peripheral blood mononuclear cells (PBMCs),human T-leukemia cells (Jurkat cells) and humanskin fibroblasts.

Methods

Ethics statements

Human term placentas and skin biopsies werecollected after obtaining written informed consentaccording to the guidelines of the Ethical Commit-tee of the Catholic Hospital (CEIOC) and of theEthical Committee of the hospital FondazionePoliambulanza-Istituto Ospedaliero (Brescia, Italy),respectively.

Animal experiments were carried out in accor-dance with the guidelines established by the Euro-pean Community (No. 2007/526/CE) and by theItalian law 116/92 on the accommodation and careof animals used for experimental purposes. Theexperimental protocol was approved by the Com-mittee on the Ethics of Animal Experiments of theUniversity of Milano.

Cells used to prepare CMs

CMs were generated from the culture of thefollowing four cell types:

(i) Amniotic mesenchymal tissue cells are cellsisolated from the mesenchymal region of hu-man amniotic membrane, following a well-established protocol of our laboratory (13,14).As previously described (14,15), at passage 0,their phenotype (by fluorescence-activated cellsorting analyses) is CD90 (82%� 3%), CD73(66% � 6%), CD13 (89% � 2%), CD44(57% � 10%), CD105 (6% � 4%), CD166(14% � 4%), CD45 (6% � 3%), HLA-DR(6% � 3%) and CD14 (6% � 3%) and nega-tive for CD34. AMTCs used in this study wereobtained from six different placentas.

(ii) Human fibroblasts were taken from a dermalhuman fibroblast cell line, from skin biopsies,established in our laboratory (15).

(iii) Human peripheral blood mononuclear cellswere obtained from heparinized blood samples

AMTC-CM exerts specific fibrosis-preventing action 19

of four different healthy donors through theuse of density gradient centrifugation (Lym-phoprep; Axis-Shield, Oslo, Norway).

(iv) Jurkat cells were taken from a human T-cellleukemia cell line obtained from CentroSubstrati Cellulari, Istituto Zooprofilattico ofBrescia (Italy).

Preparation of CMs

CMs were generated from in vitro cultures ofAMTCs (AMTC-CM), as well as of human fibro-blasts (Fibro-CM), PBMCs (PBMC-CM) andJurkat cells (Jurkat-CM). The primary AMTCs andPBMCs were used immediately after isolation.

The culture conditions were the same for all celltypes: cells were plated in 24-well plates (CorningInc, Corning, NY, USA) at a density of 1 � 106 cellsper well in 1 mL of serum-free culture medium[UltraCulture medium, (Lonza, Basel, Switzerland)]supplemented with 100 U/mL penicillin and 100mg/mL streptomycin (both from Euroclone, Whe-therby, UK). Cells were cultured for 5 days at 37�Cand 5% CO2. At the end of this culture period, cellviability was >85% for all cell types used. The cellculture supernatants, constituting of the CM from thefour different cell types, were harvested, centrifugedat 300g, filtered through a 0.8-mm sterile filter(Sartorius AG, Goettingen, Germany) and oppor-tunely aliquoted under sterile conditions and frozenat �80�C.

Non-conditioned medium (control-CM) wasused as a control and consisted of serum-freeUltraCulture medium supplemented with 100 U/mLpenicillin and 100 mg/mL streptomycin incubated(1 mL/well) without cells for 5 days at 37�C and 5%CO2. Afterward, the medium was collected andprocessed in the same way as the other CMs.

Preparations generated from cultures of AMTCsobtained from the six different placentas andcultured PBMCs obtained from the four differentdonors were pooled.

Each CM preparation, before and after beingpooled, was tested for its ability to inhibit lymphocyteproliferation in vitro, as previously described (14).

Finally, all CMs were lyophilized and stored at 4�Cbefore being reconstituted to be injected into bleo-mycin-challenged mice, as previously described (12).

Induction of lung injury in mice

All experimental procedures were performed onanesthetized animals [by intramuscular injection ofTiletamine chloride þ zolazepam chloride (Zoletil-100, Virbac, Milano, Italy, at 0.3 mL/30 mg/kg, andxylazine, Bayer Schering Pharma AG, Leverkusen,

Germany, at 2 mg/kg)], and all efforts were made tominimize animal suffering.

A total of 288 8- to 9-week-old female C57BL/6mice (Charles River, Calco, Italy), a strain reportedto be bleomycin-sensitive (16), was used. Amongthese, 269 were instilled intratracheally with 50 mL ofbleomycin (Nippon Kaiaku Co, Tokyo, Japan) at aconcentration of 4 U/kg, as previously described (8),whereas the remaining 19 were instilled intra-tracheally with 50 mL saline, thus representing thecontrol, non-pathological animals (saline group).

Experimental groups

After bleomycin (Bleo) instillation, animals wererandomly divided further into six experimentalgroups as follows: Bleo group: bleomycin-instilledmice that received no further treatments (n ¼ 52);Bleoþcontrol-CM group: bleomycin-instilled micetreated with control CM (n ¼ 43); BleoþAMTC-CM group (n ¼ 47); BleoþFibro-CM group(n ¼ 44); BleoþPBMC-CM group (n ¼ 40) andBleoþJurkat-CM group (n ¼ 43).

Animals from each experimental group, includingthose from the saline group, were euthanized 14, 21 or28 days after bleomycin/saline instillation.

Treatment with CM

Fifteen minutes after bleomycin instillation, 100 mLof reconstituted CM was injected into the rightthoracic cavity of each treated mouse, as previouslydescribed by our group (12). Specifically, the anes-thetized mice were placed in the left lateral decubitusposition, and CM was injected percutaneouslythrough the right fifth intercostal space with the useof a 27-gauge needle. The needle was advancedapproximately 8 mm into the thorax and quicklyremoved after injection.

Mouse body weights

The body weight of mice from all experimentalgroups was measured before performing any proce-dure on the animals (day 0) and at days 14, 21 and28 after saline/bleomycin instillation. The change inbody weight of each animal was evaluated by sub-tracting the weight at day 0 from the weight at time ofdeath.

Blood gas and electrolytes analysis

At day 14 or 28 after bleomycin or saline instillation,mice were anesthetized and blood was drained fromthe abdominal aorta into a heparinized syringe andimmediately used for analysis. Blood analysis was

Figure 1. Lung sampling procedure. After euthanasia of the mice,lungs extracted from each mouse were separated into the left (L)and the four right lung lobes, namely, the right superior (RS), rightmedian (RM), right inferior (RI) and right postcaval lobes (PC).Each lobe was further sub-sectioned into hemilobes (representedby the solid lines), which allowed us to use representative tissuesamples from each lung for the indicated examinations. Theremaining hemilobes were further sub-sectioned into two series ofhemi-hemilobes, as shown by dotted lines. Sub-samples from eachlobe of any individual mouse were either put into formalin forsubsequent paraffin inclusion and microscopic analysis, or werecombined into a single sample, snap-frozen in liquid nitrogen andstored at �80�C for subsequent assessment of collagen, cytokineand PGE2 levels.

20 A. Cargnoni et al.

performed with an IL1640 blood gas analyzer(Instrumentation Laboratory System, Lexington,MA, USA). For each mouse, arterial pH, arterialoxygen and carbon dioxide tensions (PO2 and PCO2),arterial bicarbonate (HCO3

�) concentration andhematocrit (Hct) were measured.

Lung injury evaluation

Lung sampling.After euthanasia, lungs were extractedand sectioned on a bed of ice as reported in Figure 1.Sections were representative of the entire lung tissueand were used for different analyses. To ensure cor-rect sampling of tissues, lungs were separated into thefive individual lobes (Figure 1), namely, the left lobeand the four right lobes (superior, median, inferiorand postcaval). Each lobe was further sectioned intotwo equivalent hemilobes to give a total of two hem-ilobes for each of the five lobes. One series of ran-domly selected hemilobes (ie, one hemilobe of eachlung lobe) from each mouse was processed for histo-logical analysis.

The remaining hemilobes were further sub-sectioned into two equivalent parts, obtaining two se-ries of hemi-hemilobes. The two hemi-lobes (one

hemi-hemilobe for each lung lobe) were pooledtogether to constitute a single sample, snap-frozen inliquid nitrogen and stored at �80�C for subsequentassessment of lungcollagen (onehemi-hemilobe series)and levels of cytokines and prostaglandin (PG) E2

(the other hemi-hemilobe series), as described below.

Histological analysis

Histological processing. Each lung hemilobe wasformalin-fixed (10% neutral formalin from BiOptica,Milan, Italy) for 48 h at room temperature. Allhemilobes from each individual mouse were thenembedded in a single paraffin block. Consecutive4-mm-thick sections of each sample representative ofa section plane of each hemilobe were cut andmounted on Superfrost slides (Thermo Scientific,Menzel GmbH & Co, KG, Braunshweig, Germany)and dried overnight at 37�C. All of the histologicalanalyses were performed in a blinded manner by aveterinary pathologist, as described below.

Immunohistochemical evaluation of T-lymphocyte andmacrophage levels in mouse lung tissues. Lung sectionswere de-waxed in xylene, passed through a graded se-ries of alcohols and rehydrated in deionized water.Antigen retrieval was performed with a citrate buffer atpH 6.0 in a microwave oven for 15 min at 650 W, fol-lowed by an incubation of 20 min at room temperature(RT). After endogenous peroxidases were blocked withthe use of 0.5% hydrogen peroxide for 30min, followedby three washes in 0.05% Tween tris buffered salinesolution (TBST) at pH 7.6, unspecific labeling wasreduced by 30 min of incubation at RT with normalserum (diluted 1:10 in TBST) from the same speciesfrom which the secondary antibodies were derived.After three washes, the sections were incubated over-night at 4�Cwith the following primary antibodies: anti-CD3 (rabbit polyclonal anti-human/mouse CD3,A0452, Dako, Glostrup, Denmark; diluted 1:100 inTBST), for detecting T lymphocytes; anti-F4/80 (ratpolyclonal anti-mouse F4/80, MCA497G, Serotec,Oxford, England, UK; 1:200 in TBST), for detectingmouse macrophages; anti-iNOS (inducible nitric oxidesynthase) (rabbit polyclonal anti-mouse iNOS, 06-573,Millipore, Billerica, MA, USA; 1:500 in TBST) andanti-arginase 1 (rabbit polyclonal anti-human/mousearginase 1, PA5-22009, Thermo Scientific, MenzelGmbH & co, kG, Braunshweig, Germany; 1:750 inTBST), as markers associated with classically (M1) andalternatively (M2) activated macrophages, respectively.After three washes, sections were incubated with ananti-rabbit or anti-rat horseradish peroxidase conju-gated secondary antibody (Vectastain, Vector Labs Inc,Cambridgeshire, UK) for 30 min at RT. Peroxidase

AMTC-CM exerts specific fibrosis-preventing action 21

reaction was developed for 10 minutes with the use ofdiaminobenzidine as the chromogen (Impact DAB,Vector Labs Inc) and blocked with deionized water.Substitution of the primary antibody with a rabbit- orrat-unrelated primary polyclonal serum served as anegative control.

All of the lung tissue sections present in the slidewere scanned and analyzed in sequential �200 fields.A semiquantitative histological score was used tograde the presence of lymphocytes and macrophagesin the lung tissues as follows: grade 0 ¼ no positivecells; grade 1 ¼ low number of discrete positive cells(between one and 10 cells) scattered throughout thefield or a single small aggregate of positive cells(composed of three to 10 cells); grade 2 ¼ moderatenumbers of discrete positive cells (between 10 and30) or presence of multiple small aggregates of pos-itive cells (10e15 cells); grade 3 ¼ moderate to highnumber of discrete positive cells (30e60) or multiplemedium-sized aggregates of positive cells (15e20cells); grade 4 ¼ high presence of discrete positivecells (>60) or presence of multiple large-sized ag-gregates of positive cells (>20 cells). Scores obtainedfrom each �200 field examined of each lung werethen averaged to give a final score, which was rep-resentative for the lungs of the single mouse.

Evaluation of lung fibrosis. Consecutive sections werestained with hematoxylin and eosin and withMasson-Goldner trichrome (BiOptica). Histologicalgrading of fibrosis was performed with the use of abright-field microscope. Specifically, each consecu-tive field (�100 objective) of each hemilobe sectionwas individually assessed for severity of intersti-tial fibrosis and scored by use of grades rangingfrom 0e8, by means of the semi-quantitativemethod and criteria reported by Ashcroft (17). Afterthe whole section of each hemilobe was examined,the mean score of all of the fields was taken as thefibrosis score for that hemilobe. For each animal,the lungs’ mean score of fibrosis (the score of wholelungs) was the average of all of the hemilobes’ meanscores.

Assessment of lung collagen content

The frozen lung tissue of each mouse (one series oflung hemi-hemilobes, one hemilobe for each lunglobe) was homogenized (100 mg/3.5 mL) with theuse of a tissue lyser (RETSC Qiagen, Milan, Italy)in 0.5 mol/L acetic acid to which 0.1 mg/mL pepsin(Sigma Aldrich, St Louis, MO, USA) had beenadded. The total content of collagen was measuredin lung homogenates by performing the SircolSoluble Collagen Assay (Biocolor, Newtownabbey,

Northern Ireland) according to the manufacturer’sinstructions.

Assessment of lung cytokine and chemokine levels

The frozen lung tissues (from the second series oflung hemi-hemilobes, one hemilobe for each lunglobe) of each animal were homogenized (100 mg/1.5 mL) in a buffer containing 50 mmol/L Tris-HCl,150 mmol/L NaCl, 1% Triton X-100, 0.5% Tween-20, 0.1% SDS, 1 mmol/L ethylenediamine tetra-acetic acid and protease inhibitor cocktail (Roche,Mannheim, Germany) (9), followed by the additionof 10 mmol/L indomethacin (Sigma Aldrich). Thehomogenates were then centrifuged for 15 min at14,000g, and the supernatant was opportunely ali-quoted and stored at �80�C until use for cytokineand PGE2 detection.

Levels of interleukin (IL)-2, IL-4, IL-13, IL-12,IL-6, IL-10, interferon-g (INF-g), tumor necrosisfactor-a (TNF-a), macrophage inflammatory pro-tein-1a and 1b (MIP-1a and MIP-1b) and monocytechemoattractant protein-1 (MCP-1) were measuredwith the use of the Cytometric Bead Array (CBAFlex Set, BD Biosciences, San Jose, CA, USA) ac-cording to the manufacturer’s instructions.

Transforming growth factor-b1 (TGF-b1) wasmeasured in the homogenate aliquots, after acidifi-cation, by enzyme-linked immunosorbent assay.Microtiter plates were coated overnight at RT with arat anti-mouse antieTGF-b1 monoclonal antibody(BD Pharmigen) (3 mg/mL in carbonate-bicarbonatebuffer, pH 9) and washed with phosphate-bufferedsaline containing 0.05% Tween-20 (Sigma Aldrich).Nonspecific binding sites were blocked with 1%casein and 5% Tween-20 in phosphate-buffered sa-line (60 min at RT). After the wash, standards andlung homogenates were plated and biotinylatedantibody was added at a concentration of 2 mg/mLper well (biotinylated rat anti-mouse TGF-b1, BDPharmigen). After a 2 h incubation at RT, plateswere washed and alkaline phosphataseeconjugatedextravidin (Sigma Aldrich) was added and incubatedfor 60 min at RT. After the wash, the alkalinephosphatase substrate (Bio-Rad Laboratories, Seg-rate, Milan, Italy) was added and incubated for 120min. Plates were read at 490 nm by an automatedDV990/BV6 microplate reader (GDV, Rome, Italy).

Determination of lung PGE2 levels

PGE2 content was measured in lung homogenates ofmice belonging to the saline group, the Bleo groupand the BleoþAMTC-CM group. Detection wasperformed through enzyme immunoassay, whichwas carried out according to the manufacturer’s

22 A. Cargnoni et al.

instructions (Prostaglandin E2 EIA Kit, CaymanChemical Company, Ann Arbor, MI, USA).

Figure 2. Mouse survival functions after bleomycin and CMdeliveries. Mouse deaths were recorded in the different experi-mental groups throughout the study. Day zero represents bleo-mycin or saline instillation. The mouse survival functions arerepresented for each experimental group as Kaplan-Meier curves.Each tick mark on the curves indicates the censored cases within

Statistical analysis

All of the parameters obtained from blood gas ana-lyses and from lung tissue evaluations were expressedas median values and the relative interquartile range(IQR). Kruskal-Wallis nonparametric analysis ofvariance was performed to evaluate the statisticalsignificance of the differences among experimentalgroups. Differences between experimental groups atspecific time points were then assessed by use of theMann-Whitney test with Holm-Bonferroni correc-tion for multiple comparisons.

The comparison of mortality between theexperimental groups was carried out with the useof a c2 test.

A P value of <0.05 was considered statisticallysignificant. Statistical analysis was performed withSPSS Advanced Statistics 13.0 (SPSS Inc, Chicago,IL, USA).

each experimental group, representing mice killed at 14, 21 and28 days after bleomycin or saline instillation.

Results

Animal survival and body changes after bleomycin andCM delivery

Within 24 hours of bleomycin instillation andCM delivery, probably as the result of the treatmentprocedures, we observed early death in the Bleogroup (one death), the BleoþFibro-CM group (onedeath), the Bleoþcontrol-CM group (seven deaths),the BleoþAMTC-CM group (three deaths), theBleoþJurkat-CM group (three deaths) and theBleoþPBMC-CM group (four deaths). Between 8and 14 days, as also reported by others (18e20), deathwas also observed. In particular, the highest number oflate deaths was recorded in the Bleo group (26 deaths/51 mice ¼ 51.0%), whereas the lowest rate was in theBleoþAMTC-CM (2 deaths/44 mice ¼ 4.5%)(Figure 2). The other experimental groups showedintermediate levels of late death: 11.1%, 20.9%,16.7% and 25.0% in Bleoþcontrol-CM, BleoþFibro-CM, BleoþPBMC-CM and BleoþJurkat-CMgroups, respectively. The mortality rate of mice inthe BleoþAMTC-CM group was significantly lowerthan that recorded for the Bleo group (P < 0.0001;byc2 test) andwith respect to the cumulativemortalityof animals belonging to all of the other treatmentgroups (P ¼ 0.017, by c2 test). On the contrary,none of the other treatment groups (Bleoþcontrol-CM, BleoþFibro-CM, BleoþPBMC-CM andBleoþJurkat-CM) showed any significant differenceswith respect to the cumulative mortality of the othergroups (P ¼ 0.196 for the Bleoþcontrol-CM group;

P ¼ 0.360 for the BleoþFibro-CM group; P ¼ 0.265for the BleoþPBMC-CM group; P ¼ 0.167 for theBleoþJurkat-CM group). No death was observed inmice of the saline group (Figure 2).

Therefore, after 14 days from the beginning of theexperiments, the number of remaining mice distrib-uted in the experimental groups was as follows:Bleo (n ¼ 25); Bleoþcontrol-CM (n ¼ 32); BleoþAMTC-CM (n ¼ 42); BleoþFibro-CM (n ¼ 34);BleoþPBMC-CM (n ¼ 30) and BleoþJurkat-CM(n ¼ 30).

Assessment of the effects of bleomycin and CMdelivery on body weight revealed that after 14 days,as expected, mice from the Bleo group showed asignificant loss in body weight [�1.80 g (IQR: 5.70)]when compared with mice of the saline group, whichshowed an increase in body weight [1.30 g (IQR:0.60); P < 0.01 versus Bleo group] (Figure 3). Micetreated with AMTC-CM showed an increase in bodyweight [0.60 g (IQR: 1.33), P < 0.02 versus Bleogroup]. Animals that had been treated with the otherCM types decreased in body weight, with theexception of mice of the BleoþJurkat-CM group,which had a slight increase in weight, althoughthis did not reach statistical significance [0.40 g(IQR: 4.03)].

Twenty-one days after bleomycin instillation,mice from the Bleo group regained body weight[1.30 g (IQR: 1.27)], which did not increase anyfurther up to day 28 [1.20 g (IQR: 6.25)]. Mice ofthe BleoþAMTC-CM group progressively gained

Figure 3. Changes in body weight at 14, 21 and 28 days after bleomycin/saline instillation. Body weight changes were evaluated for eachanimal by subtracting the baseline weight from that monitored at the time of death. A negative change indicates a decrease in body weight; apositive change indicates an increase. White bars represent values collected at day 14, light gray bars at day 21 and dark gray bars at day 28after bleomycin/saline instillation. Numbers in bars indicate the number of animals. þþþP < 0.01, Bleo group versus saline group; **P <

0.02 and *P < 0.05, CM-treated groups versus Bleo group.

AMTC-CM exerts specific fibrosis-preventing action 23

weight, and by day 28 had undergone an increase inbody weight [2.00 g (IQR: 1.73)] that was higherthan that of the Bleo group and very close to thatshown by mice of the saline group at this same timepoint (Figure 3).

In time, the mice of all of the other groupsregained body weight, albeit to a lesser extent thanmice from the BleoþAMTC-CM group (Figure 3).

Figure 4. Lung fibrosis severity and progression. Lung fibrosis was microin the Methods section. Fibrosis was evaluated in each lung hemilobe (luthe four right lung lobes, namely, the right superior [RS], right median [Rof the hemilobe scores represents the fibrosis grade for each animal. Whitand dark gray bars at day 28 after bleomycin/saline instillation. Numbersversus saline group; **P < 0.02, CM-treated groups versus Bleo group.

Effect of AMTC-CM and other CMs on lung fibrosis

At day 14, mice of the Bleo group developed lungfibrosis that was scored at a grade of 3.15 (IQR: 2.18)score units (Figure 4), which, by Ashcroft’s criteria,indicates “increased fibrosis with definite damage tolung structure and formation of fibrous bands orsmall fibrous masses.” Twenty-one and 28 days afterbleomycin instillation, lung fibrosis in mice of this

scopically graded by applying Ashcroft’s scoring system as reportedngs extracted from each mouse were separated into the left [L] andM], right inferior [RI] and right postcaval lobes [PC]); the averagee bars represent scores evaluated at day 14, light gray bars at day 21in bars indicate the number of animals. þþþP < 0.01, Bleo group

Figure 5. Collagen content in mouse lungs. Lung fibrosis was quantitatively evaluated as lung collagen content, as described in the Methodssection. White bars represent collagen content measured after 14 days, light gray bars after 21 days and dark gray bars after 28 days afterbleomycin/saline instillation. Numbers in bars indicate the number of animals. þþþP < 0.01, Bleo group versus saline group; ***P < 0.01and *P < 0.05, CM-treated groups versus Bleo group.

24 A. Cargnoni et al.

group retained severity score grades resembling thoseobserved at day 14 [3.53 (IQR: 1.84) and 3.47 (IQR:2.44) score units, at days 21 and 28, respectively](Figure 4).

Meanwhile, lung fibrosis observed in mice treatedwith AMTC-CM was significantly lower than thatfound in mice of the Bleo group. In particular, at day14, fibrosis severity was scored at 1.72 (IQR: 1.10)score units (P < 0.01 versus Bleo group) and 1.61(IQR: 1.73) and 1.25 (IQR: 0.91) score units (P <0.01 versus Bleo group) at days 21 and 28, respec-tively (Figure 4).

Mice treated with control-CM, Fibro-CM,PBMC-CM and Jurkat-CM showed lung fibrosislevels similar to or higher than those observed inmice of the Bleo group (Figure 4).

The histological evaluations of fibrosis wereconfirmed by the detection of collagen in the lungs ofthe mice. In mice of the Bleo group, the developmentof lung fibrosis was accompanied by significantlyhigher collagen deposition with respect to miceinstilled with saline, and mice of this group main-tained stable lung collagen levels at days 14, 21 and28 after bleomycin instillation (Figure 5).

AMTC-CM was the only treatment that was ableto significantly reduce lung collagen deposition.Indeed, AMTC-CMetreated animals exhibited lungcollagen levels that were significantly lower, both atday 14 [2087.3 (IQR: 970.9) versus 2690.1 (IQR:1279.3) mg; P < 0.01] and at day 28 [1592.4 (IQR:1070.2) versus 2481.9 (IQR: 889.3) mg; P < 0.05],than the levels observed in mice from the Bleo group(Figure 5). Mice treated with control-CM, Fibro-CM, PBMC-CM and Jurkat-CM showed lung

collagen levels that were not significantly different orwhich were even higher than those measured in theBleo group at days 14, 21 or 28 (Figure 5).

We also found that the data obtained from theapplication of the two different methods used toevaluate lung fibrosis (the Ashcroft’s histologicalscore system and the lung collagen quantitativecontent measurement) are significantly correlated(Pearson coefficient ¼ 0.230, P ¼ 0.020).

Analysis of blood gas and electrolytes

To establish a possible correlation between fibroticlung injury and consequent pulmonary functionalimpairment, as well as to evaluate the effects of CMtreatment on pulmonary function, we performed ablood gas analysis and assessed any alterations in gasexchange parameters.

Among the parameters monitored, arterial pH didnot show significant variations in mice from thedifferent experimental groups over any of the timepoints studied (data not shown). On the contrary,differences in oxygen (PO2) and carbon dioxide ten-sions (PCO2), bicarbonate blood concentrations(HCO3

�) and hematocrit (Hct) were observed be-tween the different treatment groups (Figure 6AeD).Specifically, at day 14, when compared with mice ofthe saline group, mice from the Bleo group showedreduced PO2 [33.50 (IQR: 19.75) versus 47.50 (IQR:13.50) mm Hg] (Figure 6A) and increased PCO2

[67.00 (IQR: 16.50) versus 58.00 (IQR: 15.00) mmHg] (Figure 6B). In response to the lung’s impairedability to eliminate CO2, this group also showed anincrease in arterial HCO3

e [26.80 (IQR: 6.05) versus

Figure 6. Blood gas analysis in mice of different treatment groups. At day 14 and day 28, blood was collected from anesthetized mice beforeeuthanasia and were immediately used for gas analysis (see Methods section). Arterial oxygen and carbon dioxide tension (PO2 and PCO2)assessments are reported in A and B, respectively. Bicarbonate concentration and Hct measurements are reported in C and D, respectively.White bars represent parameter values assessed at day 14 and dark gray bars at day 28 after bleomycin/saline instillation. Numbers in barsindicate the number of animals. þþþP < 0.01, Bleo group versus saline group; *P < 0.05, **P < 0.02 and ***P < 0.01, CM-treated groupsversus Bleo group.

AMTC-CM exerts specific fibrosis-preventing action 25

21.40 (IQR: 6.30) mmol/L; P < 0.05] (Figure 6C).Furthermore, Hct was higher in the Bleo groupmice compared with saline-treated mice [41.00 (IQR:4.25) versus 37.00 (IQR: 6.75) %] (Figure 6D),which suggests that a compensative mechanism toblood hypo-oxygenation had been activated in theseanimals.

Compared with day 14, mice of the Bleo groupshowed slightly increased PO2 values at day 28(Figure 6A), with a parallel significant increasein Hct levels (P < 0.01 versus saline group)(Figure 6D), whereas slightly increased PCO2 levelsand stable levels of arterial HCO3

� concentrationthat remained higher with respect to the saline groupwere also observed (Figure 6B,C).

With the exception of mice of the BleoþAMTC-CM group, animals of all treatment groups showedPO2, PCO2, HCO3

� concentration and Hct valueswhich were largely similar or not statistically different

to those measured in the Bleo group, both at day14 and day 28.

In mice treated with Fibro-CM at day 14, PCO2

and HCO3� concentration, as well as Hct, were lower

than in the Bleo group, but only bicarbonate levelswere statistically different in these animals comparedwith the Bleo group [18.70 (IQR: 7.03) mmol/L;P < 0.02] (Figure 6C). Instead, at day 28, all pa-rameters in the BleoþFibro-CM group mice weresimilar to those observed in mice of the Bleo group.

In addition, whereas mice treated with PBMC-CM showed PO2, PCO2 and HCO3

e concentrationvalues similar to those observed in the Bleo group,the Hct values in these animals were highly variableand displayed median values that were even lowerthan those of the saline group.

Meanwhile, mice treated with AMTC-CM showedsignificant improvements at both time points ofanalysis. At day 14, when compared with mice of the

26 A. Cargnoni et al.

Bleo group, these animals showed lower levels ofPCO2 [55.00 (IQR: 11.00) mm Hg; P < 0.02];HCO3

� concentration [20.00 (IQR: 3.65) mmol/L;P < 0.01] and Hct [39.00 (IQR: 6.50) %; P < 0.05],the levels of which were all closer to those observedin the saline group. At day 28, mice of the BleoþAMTC-CM group maintained these parameters atlevels similar to those recorded at day 14, which weresignificantly lower with respect to those observed inthe Bleo group, with the exception of PO2, which wasat a higher level than that observed in Bleo groupmice, although this difference did not reach statisticalsignificance.

Cytokine and chemokine levels in mouse lungs

To gain some insights into the possible mechanism(s)underlying the beneficial effects of AMTC-CM, weanalyzed the levels of both pro-inflammatory andpro-fibrotic cytokines, which have been shown to becentral to the pathogenesis of lung injury, in lunghomogenates from mice that had been treatedwith the different CM types. At day 14, we observedthat mice of the Bleo group showed higher levelsof IL-6, TNF-a, MIP-1a, MCP-1 and TGF-b1with regard to animals of the saline group (Table I).Meanwhile, significantly lower levels of all of thesecytokines were observed in mice of the BleoþAMTC-CM group when compared with mice ofthe Bleo group: IL-6 [8.82 (IQR: 5.02) versus 28.20(IQR: 28.59) pg/mL; P < 0.01], TNF-a [2.96 (IQR:2.12) versus 8.55 (IQR: 3.38) pg/mL; P < 0.01],MIP-1a [65.42 (IQR: 17.42) versus 118.27 (IQR:23.84) pg/mL; P < 0.01], MCP-1 [161.49 (IQR:121.80) versus 372.92 (IQR: 140.96) pg/mL;

Table I. Cytokine and chemokine levels in mouse lungs.

TreatmentDays aftertreatment

IL-6 (pg/mL) TNF-a (pg/mL)

M IQR M IQR

Saline 14 3.85 2.02 1.63 1.7628 3.15 2.80 1.50 1.95

Bleomycin 14 28.20c 28.59 8.55b 3.3828 8.94 5.71 3.34 2.21

Bleoþcontrol-CM 14 28.04 15.22 9.80 3.7028 10.15 4.23 2.63 3.17

BleoþAMTC-CM 14 8.82f 5.02 2.96f 2.1228 6.89 2.70 2.95 2.86

BleoþFibro-CM 14 23.62 19.56 5.20e 3.5028 10.10 5.46 4.90 4.00

BleoþPBMC-CM 14 20.55 21.92 8.55 3.6728 12.42 4.24 3.15 3.13

BleoþJurkat-CM 14 19.93 9.24 9.21 5.6828 12.30 3.61 3.20 6.25

Median (M) values with IQR obtained from animals of the same groupaP < 0.05; bP < 0.02; cP < 0.01 Bleo group versus saline group; dP < 0

P < 0.01] and TGF-b1 [1121.50 (IQR: 567.73)versus 2106.04 (IQR: 1349.32) pg/mL; P < 0.02](Table I). At this time point, significantly lowerlevels of TNF-a, MIP-1a and TGF-b1 were alsoobserved in the BleoþFibro-CM group. Mean-while, all of the other treatment groups showedcytokine levels similar to those of the Bleo group(Table I).

Other cytokines, such as IL-2, IL-4, IL-12,IL-13, IL-10 and INF-g, did not reach the minimumdetection limit of the applied cytometric array in anyof the lung samples analyzed.

At day 28, the levels of all cytokines/chemokineswere low in mice of all experimental groups, with theexception of TGF-b1, which showed levels that weresignificantly lower in mice of BleoþAMTC-CM andBleoþFibro-CM groups with respect to animals ofthe Bleo group.

The lungs of mice from all groups showed nosignificant changes in MIP-1b levels, at either day 14or day 28 (data not shown).

T-lymphocyte and macrophage detection in mouse lungs

To investigate whether the reduction in pro-inflam-matory cytokines after treatment with AMTC-CMwas associated with lower T-cell/macrophage levelsin the lung, we analyzed the lung tissues from miceof the saline, Bleo, Bleoþcontrol-CM and BleoþAMTC-CM groups for the presence of T lym-phocytes and macrophages.

After 14 days from bleomycin challenge, lungsections of mice of the Bleo group showed moderatelevels of T lymphocytes (CD3þ cells) and macro-phages (F4/80þ cells) (Figure 7A), compared with

MIP-1a (pg/mL) MCP-1 (pg/mL) TGF-b1(pg/mL)

M IQR M IQR M IQR

35.00 29.90 93.17 30.63 346.48 194.8024.59 11.49 112.36 31.68 345.06 235.82

118.27b 23.84 372.92b 140.96 2106.04b 1349.3252.09 5.71 118.26 96.45 1568.85a 1123.4393.04 58.47 340.04 237.03 1575.35 860.6647.89 24.29 150.55 55.58 1160.82 458.4165.42f 17.42 161.49f 121.80 1121.50e 567.7343.07 21.81 119.11 81.10 649.83d 310.0274.20e 28.45 226.81 419.82 1025.28d 620.9837.08 26.21 139.43 131.19 469.88d 269.27

110.27 44.40 231.36 252.37 1622.74 616.0462.54 29.65 223.16 181.43 1390.55 719.0894.83 41.71 282.82 345.04 879.69 1093.0542.88 51.12 183.17 181.64 1361.93 840.05

at each time point are reported..05; eP < 0.02; fP < 0.01 CM-treated groups versus Bleo group.

Figure 7. Macrophage levels in mouse lungs. Semiquantitative score analysis of lung sections marked with antibodies: (A) anti- F4/80 (pananti-macrophages), (B) anti-iNOS and (C) antieArg-1 obtained from mice of the saline group (n ¼ 2; white bars), Bleo group (n ¼ 5; darkgray bars), Bleoþcontrol-CM group (n ¼ 5; light gray bars) and BleoþAMTC-CM group (n ¼ 5; light gray bars), at day 14 after bleomycin/saline instillation, are reported. *P < 0.05 versus Bleo group. (DeG) Representative photomicrographs from lung sections stained for iNOS(D, E) and Arg-1 (F, G) of mice of Bleo and BleoþAMTC-CM groups, respectively. Insets represent higher magnification of positivemacrophages. The indirect immunoperoxidase method was applied. Scale bar ¼ 50 mm.

AMTC-CM exerts specific fibrosis-preventing action 27

mice of the saline group. Inmice of the BleoþAMTC-CM group, the lung levels of T lymphocytes weresimilar (data not shown), whereas macrophages(F4/80þ cells) levels were reduced, although not to astatistically significant level, with respect tomice of theBleo group (Figure 7A). In particular, we observedreduced levels of both iNOSþ macrophages [1.06(IQR: 0.49) versus 1.50 (IQR: 0.42) score units; P ¼0.047] (Figure 7B,DeE) and Arg-1þ macrophages[from 1.07 (IQR: 0.23) versus 1.32 (IQR: 1.34) scoreunits; P ¼ 0.076] (Figure 7C,FeG). Instead, mice ofthe Bleoþcontrol-CM group showed lymphocyte andmacrophage lung levels similar to those in mice of theBleo group (Figure 7).

PGE2 levels in mouse lungs

Considering the implication of PGE2 in controllinglung fibroblast activation in normal lungs (21), wedetermined the PGE2 content in lung homogenatesof mice of the saline group and of the Bleo andBleoþAMTC-CM groups.

After 14 days, the levels of PGE2 detected in thelungs of mice of the Bleo group were significantlyhigher than those of the saline group [101.20 (IQR:

104.78) versus 25.49 (IQR: 3.42) ng/mL; P < 0.02].Treatment with AMTC-CM did not significantlychange PGE2 levels with respect to those observed inthe Bleo group [69.82 (IQR: 136.27) ng/mL].

At day 28, mice of Bleo group and of the BleoþAMTC-CM group still showed similar lung levelsof PGE2 [224.81 (IQR: 208.52) and 211.33 (IQR:275.27) ng/mL, respectively], even though the levelsobserved were higher than those measured at day 14in both groups.

Discussion

In the present study, we further dissected the bene-ficial effects of a cell-free treatment on the basis ofthe use of conditioned medium derived from cellsisolated from the mesenchymal region of humanamniotic membrane in mice with bleomycin-inducedpulmonary fibrosis. First, we confirmed our previousevidence regarding the AMTC-CM’s ability tomaintain lung fibrosis at levels lower than thoseobserved in untreated mice (12) and we extended thetime of analysis to 28 days; furthermore, throughcomparative studies, we demonstrated the specificityof AMTC-CM’s action with respect to CMs derived

28 A. Cargnoni et al.

from other cell types (fibroblasts, PBMCs and Jurkatcells). Secondly, we showed that AMTC-CM medi-ated improvement of functional lung parameters suchas blood gas exchanges. Finally, we provide insight onthe possible key players that may be responsible forthe observed reduction in fibrosis.

Before an in-depth discussion of our main results,we feel that it would be useful to first discuss someconsiderations regarding our methodological ap-proaches. In regard to the CM-delivery route, as pre-viously reported (12), in the present study CMs wereinjected intrathoracically rather than by a systemicroute, in order to deliver the CM as closely as possibleto the site of damage, and therefore probably deliv-ering the highest possible CM concentration to thissite, which is a strategy also used in humans to supplytreatments locally (22,23). We confirm that althoughthe intrathoracic injection was performed only intothe right cavity, the fibrosis was equally reduced inboth left and right lungs of each mouse treated withAMTC-CM, as in our previous study (12), and also inaccordance with findings of others for drug distribu-tion after intrathoracic administration (24).

Furthermore, we also assessed AMTC-CM’sability to inhibit in vitro T-cell receptoredependentlymphocyte proliferation as an in vitro control testof their potential immunomodulatory capacity. Inter-estingly, although consistent with our recent findings(14), each AMTC-CM preparation (regardless ofwhether alone or pooled) significantly inhibitedlymphocyte proliferation in a dose-dependent manner,the in vitro immunosuppressive effects observed withthe other CMs tested were absent or very low and didnot reach a statistical significance (data not shown).

It is also noteworthy that the results that we haveherein reported were obtained by introducing a sys-tem for lung sampling and sub-sampling, whichallowed us to perform each of our analyses in astrictly representational manner, ensuring uniformityof analysis for the accurate comparison of resultsbetween experimental groups, all together making usvery confident of the findings obtained.

Turning now to our main results, first of all, weobserved that AMTC-CMetreated mice had a lowermortality rate with respect to mice treated with theother CM types. The finding that treatment with thedifferent CM types caused differing rates of reduc-tion in mouse mortality may be due to the fact thatinjection of any type of solution may “dilute” thebleomycin concentration at the alveolar level, therebyreducing the acute toxicity of the drug. Alternatively,this could also be a consequence of the hypertonicityof all of the injected CM types, which were recon-stituted to a concentration that is 4-fold higher thanthe original CM, as in our previous publication (12).As already observed in the case of treatment with

hypertonic saline in a model of acute pulmonaryinjury induced by oleic acid (25), the hypertonicity ofCM could reduce acute lung inflammation andedema induced by bleomycin instillation. However,it is noteworthy that only mice of the BleoþAMTC-CM group displayed a reduction in mortality ratethat was accompanied by a parallel reduction in lungfibrosis, as assessed through both a semi-quantitativescoring system and a quantitative analysis of the lungcollagen content: that is, reduction in the toxicity ofbleomycin induced by any of the CM types deliveredwas not sufficient to cause a reduction in lungfibrosis, but rather, some specific action of theAMTC-CM was necessary to achieve such an effect.

Consistent with the presence of fibrotic lesions,which were less severe than those observed in un-treated mice or in mice treated with other CM types,animals that were treated with AMTC-CM showedcomparatively the less severe pulmonary dysfunction,with better blood gas exchanges. Whereas in the miceof all the other experimental groups, the pathologicalrespiratory imbalance (ie, increase in PCO2 andhypo-oxygenation) was accompanied by metaboliccompensatory changes that resulted in increases inHCO3

� concentration and in Hct (increase in thered cell number), in AMTC-CM-treated mice weobserved only minor levels of PCO2 with, accord-ingly, a lower concentration of HCO3

� in the bloodand also a lower Hct. Plethysmography was alsoperformed to evaluate mouse breathing patterns, butwe obtained variable results (probably as a resultof the low sensitivity of this technique when appliedto very small animals, as in our experimental condi-tions), with no significant difference among theexperimental groups (data not shown).

In the present study, we also attempted to addressan aspect that is surely the most challenged in thisfield, that is, to reveal the mechanisms that underliethe observed effects.

Considering that AMTC-CM was administeredalmost concomitantly to bleomycin injury induction,we first assessed whether the AMTC-CM treatmentcould influence the lung’s inflammatory response tobleomycin-induced alveolar epithelial damage,reducing inflammation and subsequent fibrosis.Indeed, although this represents a much-debatedtopic, many studies support a crucial role of inflam-mation in fibrosis development (26,27).

Although we did not find significant differencesbetween mice of the Bleo group and mice that hadbeen treated with the other CM types, the AMTC-CMetreated animals, at 14 days after treatment,showed significantly lower lung content of cytokinesand chemokines, which, in bleomycin-inducedlung injury, trigger the inflammatory response [IL-6and TNF-a, (28)], and propagate and perpetuate

AMTC-CM exerts specific fibrosis-preventing action 29

inflammation [MIP-1a and MCP-1; (29e32)],therefore supporting an early anti-inflammatory ac-tion of AMTC-CM. Moreover, considering thatthe cytokines/chemokines recalled above are alsoresponsible for initiation of the fibrotic process bystimulating the production of TGF-b, which is a keymediator of pathological fibrosis [ie, it induces pro-liferation and activation of fibroblasts/myofibroblasts,thereby stimulating collagen-producing machineryand the building of excessive extracellular matrixcomponents (33,34)], the early anti-inflammatoryaction of AMTC-CM may also explain the reducedlevels of TGF-b that we found in animals of theBleoþAMTC-CM group with respect to mice of theother groups.

It is interesting to note that similar anti-inflam-matory mechanisms have also been suggested toexplain the fibrosis-reducing actions of other amni-otic membrane-derived cells, for example, amnioticepithelial cells, when transplanted into a similar an-imal model (9,10).

In addition, we found that AMTC-CM treatmentreduced the levels of macrophages in the lung tissues,whereas it appears not to influence lung lymphocytelevels. Because it has been recognized that macro-phages play a pivotal role in bleomycin-induced lunginflammation and fibrosis (35,36) and in all stages ofthe fibrotic process in general (37,38), it is possiblethat AMTC-CM may prevent lung fibrosis bylimiting macrophage recruitment to the lung, with aconsequent decrease in secretion of pro-inflamma-tory cytokines and TGF-b, which is predominantlyproduced by macrophages (39).

Moreover, preliminary data indicate that AMTC-CM treatment reduced both the levels of M1(iNOSþ) and M2 (Arg-1þ) macrophages in mouselungs. Therefore, possible hypotheses regarding thepotential effect of AMTC-CM could be both in thelimitation of the early inflammatory burst and ofthe release of pro-inflammatory cytokines, as sug-gested by the reduction of M1 macrophage subsetlevels, and perhaps the limitation of fibrosis devel-opment by reducing the number of Arg-1þ macro-phages in lung tissues. Indeed, even though the roleof M2 macrophages in the development and resolu-tion of the fibrotic process has recently been a sourceof much debate (38), a few lines of evidence indicatethat reduction of lung fibrosis is associated withdiminished M2 macrophage levels (36,40). Thecontradictory nature of the above-mentioned dataregarding the involvement of M2 in fibrosis estab-lishment and progression is likely related to the dif-ficulty of matching in vitro definitions with thephenotype of in situ/in vivo macrophages, which ismade somewhat more complex by a continualswitching in vivo from one functional phenotype to

another in response to tissue micro-environmentalconditions (38,41). Therefore, taking into accountthese contradictory data, in order to obtain conclu-sive results in regard to the effects that AMTC-CMcould exert on macrophages during the differentphases of the lung inflammatory and fibrotic pro-cesses, both a time point analysis and an extendedmacrophage phenotypic signature to better charac-terize the M1 and M2 macrophage subsets are stillneeded.

We also noted that although treatment withFibro-CM did not result in significant fibrosisreduction, this treatment did result in reduced lunglevels of TNF-a, MIP-1a and TGF-b. These effectsmight be attributable to the immunosuppressiveproperties of fibroblasts (42,43) and the ability ofsoluble factors released by these cells to inhibitlymphocyte activation (44). However, unlike AMTC-CM, it is noteworthy that Fibro-CM did notcause a reduction in lung levels of IL-6 and MCP-1.This suggests that IL-6 and MCP-1 not only stim-ulate TGF-b production, but may also themselveshave pro-fibrotic actions that do not require media-tion by the TGF-b pathway, as indeed supported bymany lines of evidence, reported both in vitro (45,46)and in vivo (31,47,48). On the basis of these obser-vations, it is therefore plausible that AMTC-CMmayact both by reducing TGF-b levels in lung tissuesthrough early abrogation of the cytokine network,while also acting through a TGF-beindependentmechanism, by inhibiting the actions of some cyto-kines, such as IL-6 and MCP-1, which display directpro-fibrotic activities.

Considering the fact that (i) the suppressive ac-tions of PGE2 on fibroblast activation has beensuggested to be diminished in patients with idio-pathic pulmonary fibrosis (21,49) and (ii) a syntheticanalogue of PGE2 is able to reduce bleomycin-induced lung fibrosis (50), we also investigatedwhether AMTC-CM can increase PGE2 productionin lung tissue. We found that AMTC-CMetreatedmice displayed lung PGE2 levels that were just ashigh as those observed in mice of the Bleo group,indicating that AMTC-CM was not able to furtherincrease PGE2 levels in lungs under bleomycinchallenge. However, as also reported by others (51),we observed that bleomycin increases the PGE2

content of lung tissues, probably resulting from amechanism aimed at counteracting bleomycin-induced downregulation of PGE2 receptors, whichare expressed on lung fibroblasts (EP2 receptors)(52). Therefore, even though we did not observedifferences between PGE2 levels in mice of theBleo group and those of the AMTC-CM group, wecannot completely exclude a possible effect ofAMTC-CM on the lung PGE2 pathway, probably by

30 A. Cargnoni et al.

suppressing/reducing the bleomycin-induced down-regulation of EP2 receptors on lung fibroblasts,although this hypothesis was not addressed in thisstudy.

Taken as a whole, our data strongly support thenotion that the true mediators of the actions of am-niotic membrane-derived cells are the factors whichthey release. Thus far, through in vitro characteriza-tion of the AMTC-CM, we found that it containsseveral inflammation- and fibrosis-related factors(14), whereas previous reports show that placentalmembrane-derived cells are able to secrete manyknown mediators of tissue repair, including growthfactors, cytokines, chemokines and tissue inhibitorsof metalloproteinases (53e56). However, determi-nation of the specific soluble paracrine moleculescontained in AMTC-CM and responsible for itsanti-inflammatory actions and its fibrosis-reducingability in vivo still deserves further investigation.Likewise, linking of the information on cellularsecretome components obtained in vitro with theirspecific roles and the array of their potential thera-peutic effects in vivo remains a major task for helpingus to understand paracrine-mediated repair pro-moted not only by placenta-derived cells, but also byother stem/progenitor cells (57,58), which is mademore complex by the likelihood that in vivo there isno single paracrine mediator at play, but rather, acomplex of molecules, with possibly pleiotropic ef-fects, which may also act at distinct/overlappingphases of the reparative process.

In conclusion, this study strengthens our previousobservations on the therapeutic properties of AMTC-CM and is the first report to demonstrate the speci-ficity of its actions on lung functional parameters.

Considering that in our experiments, treatmentwas delivered almost simultaneously with thefibrosis-causing agent, at present we can concludethat AMTC-CM clearly displays fibrosis-reducingaction by affecting the early inflammatory phase andthe initiation of the fibrotic process, whereas theeffect of AMTC-CM in conditions characterized byestablished fibrosis remains to be evaluated. How-ever, given that inflammation plays a key role inmany degenerative pathologies, it is tempting tospeculate that such a cell-free treatment could beused to favour reparative/regenerative processes inother diseases, at least when applied during the in-flammatory phase.

Acknowledgments

The authors thank the physicians and midwives ofthe Department of Obstetrics and Gynaecology ofFondazione Poliambulanza-Istituto Ospedaliero,Brescia, Italy, and all of the mothers who donated

placentas and volunteers who donated blood. Wethank Marco Costanzi for technical assistance forpulmonary function evaluations and Dr FiammettaAdamo (IZO S.p.A., Brescia) for media lyophiliza-tion. We also sincerely thank Dr Maddalena Carusofor critically reviewing the manuscript and MarcoEvangelista for help in editing the manuscript.

This study was supported by Fondazione Car-iplo; MIUR 5x1000-2009 (Italy) and by MIUR(contributo DM 44/2008).

A European patent application has been filed withthe application number PCT2008-004845.

Disclosure of interests: Dr Parolini is member ofthe scientific board of Auxocell Laboratories, Inc.No competing financial interests exist for otherauthors.

References

1. Bailo M, Soncini M, Vertua E, Signoroni PB, Sanzone S,Lombardi G, et al. Engraftment potential of human amnionand chorion cells derived from term placenta. Trans-plantation. 2004;78:1439e48.

2. Wolbank S, Peterbauer A, Fahrner M, Hennerbichler S, vanGriensven M, Stadler G, et al. Dose-dependent immuno-modulatory effect of human stem cells from amniotic mem-brane: a comparison with human mesenchymal stem cellsfrom adipose tissue. Tissue Eng. 2007;13:1173e83.

3. Magatti M, De Munari S, Vertua E, Gibelli L, Wengler GS,Parolini O. Human amnion mesenchyme harbors cells withallogeneic T-cell suppression and stimulation capabilities.Stem Cells. 2008;26:182e92.

4. Magatti M, De Munari S, Vertua E, Nassauto C,Albertini A, Wengler GS, et al. Amniotic mesenchymaltissue cells inhibit dendritic cell differentiation of peripheralblood and amnion resident monocytes. Cell Transplant.2009;18:899e914.

5. Parolini O, Alviano F, Bergwerf I, Boraschi D, De Bari C, DeWaele P, et al. Toward cell therapy using placenta-derivedcells: disease mechanisms, cell biology, preclinical studies, andregulatory aspects at the round table. Stem Cells Dev. 2010;19:143e54.

6. Parolini O, Caruso M. Review: preclinical studies on placenta-derived cells and amniotic membrane: an update. Placenta.2011;32(Suppl 2):S186e95.

7. Manuelpillai U, Moodley Y, Borlongan CV, Parolini O.Amniotic membrane and amniotic cells: potential therapeutictools to combat tissue inflammation and fibrosis? Placenta.2011;32(Suppl 4):S320e5.

8. Cargnoni A, Gibelli L, Tosini A, Signoroni PB, Nassuato C,Arienti D, et al. Transplantation of allogeneic and xenogeneicplacenta-derived cells reduces bleomycin-induced lungfibrosis. Cell Transplant. 2009;18:405e22.

9. Moodley Y, Ilancheran S, Samuel C, Vaghjiani V, Atienza D,Williams ED, et al. Human amnion epithelial cell trans-plantation abrogates lung fibrosis and augments repair. Am JRespir Crit Care Med. 2010;182:643e51.

10. Murphy S, Lim R, Dickinson H, Acharya R, Rosli S,Jenkin G, et al. Human amnion epithelial cells prevent bleo-mycin-induced lung injury and preserve lung function. CellTransplant. 2011;20:909e23.

AMTC-CM exerts specific fibrosis-preventing action 31

11. Vosdoganes P, Hodges RJ, Lim R, Westover AJ, Acharya RY,Wallace EM, et al. Human amnion epithelial cells as a treat-ment for inflammation-induced fetal lung injury in sheep. AmJ Obstet Gynecol. 2011;205:156.e26e33.

12. Cargnoni A, Ressel L, Rossi D, Poli A, Arienti D,Lombardi G, et al. Conditioned medium from amnioticmesenchymal tissue cells reduces progression of bleomycin-induced lung fibrosis. Cytotherapy. 2012;14:153e61.

13. Soncini M, Vertua E, Gibelli L, Zorzi F, Denegri M,Albertini A, et al. Isolation and characterization of mesen-chymal cells from human fetal membranes. J Tissue EngRegen Med. 2007;1:296e305.

14. Rossi D, Pianta S, Magatti M, Sedlmayr P, Parolini O.Characterization of the conditioned medium from amnioticmembrane cells: prostaglandins as key effectors of its immu-nomodulatory activity. PLoS One. 2012;7:e46956.

15. Magatti M, De Munari S, Vertua E, Parolini O. Amnioticmembrane-derived cells inhibit proliferation of cancer celllines by inducing cell cycle arrest. J Cell Mol Med. 2012;16:2208e18.

16. Gharaee-Kermani M, Hatano K, Nozaki Y, Phan SH.Gender-based differences in bleomycin-induced pulmonaryfibrosis. Am J Pathol. 2005;166:1593e606.

17. Ashcroft T, Simpson JM, Timbrell V. Simple method ofestimating severity of pulmonary fibrosis on a numerical scale.J Clin Pathol. 1988;41:467e70.

18. Degryse AL, Tanjore H, Xu XC, Polosukhin VV, Jones BR,Boomershine CS, et al. TGFbeta signaling in lung epitheliumregulates bleomycin-induced alveolar injury and fibroblastrecruitment. Am J Physiol Lung Cell Mol Physiol. 2011;300:L887e97.

19. Zhu Y, Liu Y, Zhou W, Xiang R, Jiang L, Huang K, et al.A prostacyclin analogue, iloprost, protects from bleomycin-induced pulmonary fibrosis in mice. Respir Res. 2010;11:34.

20. Genovese T, Cuzzocrea S, Di Paola R, Mazzon E,Mastruzzo C, Catalano P, et al. Effect of rosiglitazone and 15-deoxy-Delta12,14-prostaglandin J2 on bleomycin-inducedlung injury. Eur Respir J. 2005;25:225e34.

21. Wilborn J, Crofford LJ, Burdick MD, Kunkel SL,Strieter RM, Peters-Golden M. Cultured lung fibroblastsisolated from patients with idiopathic pulmonary fibrosis havea diminished capacity to synthesize prostaglandin E2 and toexpress cyclooxygenase-2. J Clin Invest. 1995;95:1861e8.

22. Zhou A, Guo L, Tang L. Effect of an intrathoracic injection ofsodium hyaluronic acid on the prevention of pleural thick-ening in excess fluid of tuberculous thoracic cavity. Clin ExpPharmacol Physiol. 2003;30:203e5.

23. Mitamura T, Hosaka M, Takeda M, Watari H, Sakuragi N.Intrathoracic injection of paclitaxel for a patient with stage IVserous ovarian cancer: a case report. Cancer ChemotherPharmacol. 2009;64:169e70.

24. Chang CC, Chuang ST, Huang MC. Effects of chronictreatment with various neuromuscular blocking agents on thenumber and distribution of acetylcholine receptors in the ratdiaphragm. J Physiol. 1975;250:161e73.

25. Kennedy MT, Higgins BD, Costello JF, Curtin WA,Laffey JG. Hypertonic saline reduces inflammation andenhances the resolution of oleic acid induced acute lunginjury. BMC Pulm Med. 2008;8:9.

26. Wynn TA. Integrating mechanisms of pulmonary fibrosis.J Exp Med. 2011;208:1339e50.

27. Todd NW, Luzina IG, Atamas SP. Molecular and cellularmechanisms of pulmonary fibrosis. Fibrogenesis TissueRepair. 2012;5:11.

28. Smith RE, Strieter RM, Phan SH, Lukacs N, Kunkel SL.TNF and IL-6 mediate MIP-1alpha expression in bleomycin-induced lung injury. J Leukoc Biol. 1998;64:528e36.

29. Smith RE, Strieter RM, Phan SH, Lukacs NW, Huffnagle GB,Wilke CA, et al. Production and function of murine macro-phage inflammatory protein-1 alpha in bleomycin-inducedlung injury. J Immunol. 1994;153:4704e12.

30. Zhang K, Gharaee-Kermani M, Jones ML, Warren JS,Phan SH. Lung monocyte chemoattractant protein-1 geneexpression in bleomycin-induced pulmonary fibrosis.J Immunol. 1994;153:4733e41.

31. Inoshima I, Kuwano K, Hamada N, Hagimoto N, Yoshimi M,Maeyama T, et al. Anti-monocyte chemoattractant protein-1gene therapy attenuates pulmonary fibrosis in mice. Am JPhysiol Lung Cell Mol Physiol. 2004;286:L1038e44.

32. Chaudhary NI, Schnapp A, Park JE. Pharmacologic differ-entiation of inflammation and fibrosis in the rat bleomycinmodel. Am J Respir Crit Care Med. 2006;173:769e76.

33. Wilson MS, Wynn TA. Pulmonary fibrosis: pathogenesis,etiology and regulation. Mucosal Immunol. 2009;2:103e21.

34. Kottmann RM, Hogan CM, Phipps RP, Sime PJ. De-terminants of initiation and progression of idiopathic pulmo-nary fibrosis. Respirology. 2009;14:917e33.

35. Misharin AV, Scott Budinger GR, Perlman H. The lungmacrophage: a Jack of all trades. Am J Respir Crit Care Med.2011;184:497e8.

36. Gibbons MA, MacKinnon AC, Ramachandran P, Dhaliwal K,Duffin R, Phythian-Adams AT, et al. Ly6Chi monocytes directalternatively activated profibrotic macrophage regulation oflung fibrosis. Am J Respir Crit Care Med. 2011;184:569e81.

37. Wynn TA, Ramalingam TR. Mechanisms of fibrosis: thera-peutic translation for fibrotic disease. Nature Med. 2012;18:1028e40.

38. Boorsma CE, Draijer C, Melgert BN. Macrophage heteroge-neity in respiratory diseases. Mediators Inflamm. 2013;2013:Available at: http://www.hindawi.com/journals/mi/2013/769214/cta/.

39. Khalil N, Bereznay O, Sporn M, Greenberg AH. Macrophageproduction of transforming growth factor beta and fibroblastcollagen synthesis in chronic pulmonary inflammation. J ExpMed. 1989;170:727e37.

40. Murray LA, Rosada R, Moreira AP, Joshi A, Kramer MS,Hesson DP, et al. Serum amyloid P therapeutically attenuatesmurine bleomycin-induced pulmonary fibrosis via its effectson macrophages. PLoS One. 2010;5:e9683.

41. Murray PJ, Wynn TA. Protective and pathogenic functions ofmacrophage subsets. Nat Rev. 2011;11:723e37.

42. Haniffa MA, Collin MP, Buckley CD, Dazzi F. Mesenchymalstem cells: the fibroblasts’ new clothes? Haematologica. 2009;94:258e63.

43. Cappellesso-Fleury S, Puissant-Lubrano B, Apoil PA,Titeux M, Winterton P, Casteilla L, et al. Human fibroblastsshare immunosuppressive properties with bone marrowmesenchymal stem cells. J Clin Immunol. 2010;30:607e19.

44. Donnelly JJ, Xi MS, Rockey JH. A soluble product of humancorneal fibroblasts inhibits lymphocyte activation: enhance-ment by interferon-gamma. Exp Eye Res. 1993;56:157e65.

45. Moodley YP, Scaffidi AK, Misso NL, Keerthisingam C,McAnulty RJ, Laurent GJ, et al. Fibroblasts isolated fromnormal lungs and those with idiopathic pulmonary fibrosisdiffer in interleukin-6/gp130-mediated cell signaling andproliferation. Am J Pathol. 2003;163:345e54.

46. Gharaee-Kermani M, Denholm EM, Phan SH. Costimulationof fibroblast collagen and transforming growth factor beta1gene expression by monocyte chemoattractant protein-1 viaspecific receptors. J Biol Chem. 1996;271:17779e84.

47. Saito F, Tasaka S, Inoue K, Miyamoto K, Nakano Y,Ogawa Y, et al. Role of interleukin-6 in bleomycin-inducedlung inflammatory changes in mice. Am J Respir Cell MolBiol. 2008;38:566e71.

32 A. Cargnoni et al.

48. Moore BB, Paine R 3rd, Christensen PJ, Moore TA,Sitterding S, Ngan R, et al. Protection from pulmonaryfibrosis in the absence of CCR2 signaling. J Immunol. 2001;167:4368e77.

49. Keerthisingam CB, Jenkins RG, Harrison NK, Hernandez-Rodriguez NA, Booth H, Laurent GJ, et al. Cyclooxygenase-2deficiency results in a loss of the anti-proliferative responseto transforming growth factor-beta in human fibrotic lungfibroblasts and promotes bleomycin-induced pulmonaryfibrosis in mice. Am J Pathol. 2001;158:1411e22.

50. Failla M, Genovese T, Mazzon E, Fruciano M, Fagone E,Gili E, et al. 16,16-Dimethyl prostaglandin E2 efficacy onprevention and protection from bleomycin-induced lunginjury and fibrosis. Am J Respir Cell Mol Biol. 2009;41:50e8.

51. Moore BB, Coffey MJ, Christensen P, Sitterding S, Ngan R,Wilke CA, et al. GM-CSF regulates bleomycin-induced pul-monary fibrosis via a prostaglandin-dependent mechanism.J Immunol. 2000;165:4032e9.

52. Moore BB, Ballinger MN, White ES, Green ME,Herrygers AB, Wilke CA, et al. Bleomycin-induced E pros-tanoid receptor changes alter fibroblast responses to prosta-glandin E2. J Immunol. 2005;174:5644e9.

53. Steed DL, Trumpower C, Duffy D, Smith C, Marshall V,Rupp R, et al. Amnion-derived cellular cytokine solution: aphysiological combination of cytokines for wound healing.Eplasty. 2008;8:e18.

54. Wolbank S, Hildner F, Redl H, van Griensven M, Gabriel C,Hennerbichler S. Impact of human amniotic membranepreparation on release of angiogenic factors. J Tissue EngRegen Med. 2009;3:651e4.

55. Li H, Niederkorn JY, Neelam S, Mayhew E, Word RA,McCulley JP, et al. Immunosuppressive factors secreted byhuman amniotic epithelial cells. Invest Ophthalmol Vis Sci.2005;46:900e907.

56. Denison FC, Kelly RW, Calder AA, Riley SC. Cytokinesecretion by human fetal membranes, decidua and placenta atterm. Hum Reprod. 1998;13:3560e5.

57. Burchfield JS, Dimmeler S. Role of paracrine factors in stemand progenitor cell mediated cardiac repair and tissue fibrosis.Fibrogenesis Tissue Repair. 2008;1:4.

58. Boomsma RA, Geenen DL. Mesenchymal stem cells secretemultiple cytokines that promote angiogenesis and have con-trasting effects on chemotaxis and apoptosis. PLoS One.2012;7:e35685.