Conditioned medium from amniotic mesenchymal tissue cells reduces progression of bleomycin-induced lung fibrosis

Background and aims We have demonstrated recently that transplantation of placental membrane-derived cells reduces bleomycin-induced lung fibrosis in mice, despite a limited presence of transplanted cells in host lungs. Because placenta-derived cells are known to release factors with potential immunomodulatory and trophic activities, we hypothesized that transplanted cells may promote lung tissue repair via paracrine-acting molecules. To test this hypothesis, we examined whether administration of conditioned medium (CM) generated from human amniotic mesenchymal tissue cells (AMTC) was able to reduce lung fibrosis in this same animal model. Methods Bleomycin-challenged mice were either treated with AMTC-CM or control medium, or were left untreated (Bleo group). After 9 and 14 days, the distribution and severity of lung fibrosis were assessed histologically with a scoring system. Collagen deposition was also evaluated by quantitative image analysis. Results At day 14, lung fibrosis scores in AMTC-CM-treated mice were significantly lower (P<0.05) compared with mice of the Bleo group, in terms of fibrosis distribution [1.0 (interquartile range, IQR 0.9) versus 3.0 (IQR 1.8)], fibroblast proliferation [0.8 (IQR 0.4) versus 1.6 (IQR 1.0)], collagen deposition [1.4 (IQR 0.5) versus 2.0 (IQR 1.2)] and alveolar obliteration [2.3 (IQR 0.8) versus 3.2 (IQR 0.5)]. No differences were observed between mice of the Bleo group and mice treated with control medium. Quantitative analysis of collagen deposition confirmed these findings. Importantly, AMTC-CM treatment significantly reduced the fibrosis progression between the two observation time-points. Conclusions This pilot study supports the notion that AMTC exert anti-fibrotic effects through release of yet unknown soluble factors.


Introduction
Recent fi ndings indicate compelling benefi ts of cellbased treatments in animal models of lung diseases associated with infl ammatory and fi brotic processes (1). For example, transplantation of adult multipotent cells [e.g. bone marrow (BM)-derived mesenchymal stem cells (MSC) (2,3) and committed alveolar progenitor cells (4,5)], as well as transplantation of cells of fetal origin [such as placenta-derived cells (6,7) and human umbilical cord Wharton ' s jelly -derived cells (8)], has been shown to reduce lung fi brosis in animal models of bleomycin-induced pulmonary injury.
The mechanisms whereby these cell-based treatments are effective in reducing fi brosis in the lung remain poorly defi ned. Although initial interest focused on the level of engraftment of transplanted cells in host lungs and their potential for regenerating lung tissue by tissue-specifi c differentiation, the infrequency with which either occurrence has been documented makes it likely that additional mechanisms must account for the observed improvements (1). In this regard, we have reported previously that transplantation of either allogeneic or xenogeneic placental membrane-derived cells signifi cantly reduced the severity of bleomycin-induced lung fi brosis in mice, despite a rare persistence of donor cells in the lungs of recipient animals (6). In addition, we found that human amniotic membrane patches reduced postischemic cardiac scars and liver fi brosis in rat models of coronary artery and bile duct ligation, respectively (9,10). Importantly, we did not observe detectable levels of amniotic membrane-derived cells in organs treated using these approaches. In other studies, we have shown that amniotic membrane-derived cells, and in particular cells isolated from the mesenchymal region of this membrane, exert immunomodulatory actions through the release of soluble factors (11,12). In addition, placental membrane-derived cells are able to secrete a number of biologically active molecules with potentially multifaceted activities, such as cytokines (13,14), angiogenic factors associated with wound healing (15) and growth factors related to cell proliferation and differentiation (15 -17).
Based on these fi ndings, we have put forward the hypothesis that the anti-fi brotic effects observed after transplantation of placental cells are mainly related to paracrine actions exerted on host tissues by bioactive molecules secreted by these cells. An important role for soluble factors in stem/progenitor cell-mediated reparative effects has been supported by animal experiments using conditioned culture medium (CM) as an effective treatment for different tissue injuries. Indeed, in a porcine model of myocardial ischemia and reperfusion, intravenous and intracoronary injection of CM obtained from human MSC cultures reduced infarct size and improved cardiac performance (18). Furthermore, systemic infusion of human MSC-CM has been shown to reduce apoptosis and stimulate proliferation of hepatocytes in a rat model of acute liver injury (19). In addition, intramuscular injection of CM derived from human endothelial progenitor cells has resulted in tissue revascularization and functional recovery in a rat model of chronic hindlimb ischemia (20).
In this study, we evaluated the effects of administrating CM generated from cells derived from the mesenchymal region of human amniotic membrane (amniotic mesenchymal tissue, AMTC) (11) on bleomycin-induced lung fi brosis, and observed a reduction of fi brosis progression similar to that observed previously after allogeneic and xenogeneic placental membrane-derived cell transplantation. These proof of the principle results open the way to consideration of cell-free treatment approaches for fi brotic lung diseases.

AMTC isolation
Human term placentas ( n ϭ 8) were obtained following spontaneous vaginal delivery or Caesarean sections, with informed maternal consent according to the guidelines of the Institutional Ethics Committee (CEIOC, Ethics Committee of Catholic Hospital Institutions). After mechanical separation from the chorion, the amniotic membrane was washed extensively in phosphate-buffered saline (PBS; Sigma, St Louis, MO, USA) containing 100 U/mL penicillin and 100 μ g/mL streptomycin (both from EuroClone, Wetherby, UK), and then cut into 3 ϫ 3-cm fragments. Cells from the mesenchymal region of amniotic membrane were then isolated with a well established protocol that was set up at Centro di Ricerca E. Menni, Brescia, Italy, which is based on separation of the amniotic membrane from the chorionic membrane and subsequent enzymatic digestion, as described previously (11,21). We refer to these cells as AMTC (11,12). After passage (P) 2 in culture, these cells display a typical mesenchymal stromal phenotype, with the expression of markers such as CD90, CD73 and CD105 Ͼ 90% and the expression of hematopoietic markers, including CD45 and HLA-DR, Ͻ 2% (21,22). Freshly isolated AMTC display the following phenotype: CD90 (82 Ϯ 3%), CD73 (66 Ϯ 6%), CD105 (6 Ϯ 4%), CD44 (47 Ϯ 18%), CD166 (14 Ϯ 7%), CD45 (6 Ϯ 3%) and HLA-DR (6 Ϯ 3%) (similar to that previously reported by Magatti et al. ) (11). Data are expressed as the mean of the percentage positive cells Ϯ SD of at least 10 independent experiments.

AMTC cultures and preparation of CM
Freshly isolated AMTC were plated in 24-well plates (Corning Inc., Corning, NY, USA) at a density of 1 ϫ 10 6 cells/well in serum-free culture medium (UltraCulture, Lonza, Basel, Switzerland) supplemented with 100 U/mL penicillin and 100 μ g/mL streptomycin.
To generate AMTC-CM, cells were cultured for 5 days at 37 ° C in a humidifi ed atmosphere of 5% CO 2 . Supernatants from each plate were then collected, pooled, centrifuged at 700 g , fi ltered (0.2 μ m) to remove cellular debris and stored at -80 ° C. This procedure was performed for cells obtained from eight different placentas. Finally, all of the collected supernatants were pooled, lyophilized and stored at 4 ° C until use, when they were dissolved in sterile water to one-quarter of the initial volume. Control, non-conditioned medium (non-CM) was generated in the same way as above, except that no cells were cultured in the plates.

Experimental groups and CM injection
Animal experiments were carried out in accordance with current Italian and European regulations and laws on the use and care of animals for research (DL. 116/27 January 1992). All animal procedures were performed under general anesthesia (intraperitoneal injection of 2 mg/kg xylazine and 100 mg/kg ketamine).
Fifteen minutes after bleomycin instillation, which was the same time-point at which we transplanted placental fetal membrane-derived cells in our previous study (6), 100 μ L reconstituted medium was injected intrapulmonarily, in order to deliver the CM as close as possible to the site of damage. In detail, the anesthetized mice were placed in the left lateral decubitus and CM was injected percutaneously into the right fi fth intercostal space using a 27-gauge needle. The needle was advanced approximately 8 mm into the thorax and quickly removed after injection. The quantity of medium injected corresponded with the volume of CM generated by 4 ϫ 10 5 cultured cells, a number comparable with the number of placental mesenchymal cells that were transplanted into each mouse in previous experiments (6).

Lung fi brosis evaluation
Animals of each experimental group were euthanized at day 9 or day 14 after bleomycin instillation. Whole lungs of each mouse were formalin-fi xed for 48 h and embedded in paraffi n. Each block was cut at two different coronal section planes (at approximately 500 μ m distance from each other), representative of two regions of the maximum observable lung area. Consecutive 4-μ m thick sections were cut and mounted on Superfrost slides (Thermo Scientifi c, Menzel GmbH & Co. KG, Braunshweig, Germany) and dried overnight at 37 ° C. For each lung, two consecutive sections for each plane were stained with hematoxylin -eosin (HE) and Masson -Goldner trichrome (MGT).
Histologic grading of fi brosis was performed by a veterinary pathologist, who was blinded to the experimental conditions, using a bright-fi eld microscope and following a semi-quantitative method adapted from Hagood et al. (24). The fi brotic process was assessed by scoring four parameters: fi brosis distri-bution, collagen deposition (both on MGT-stained sections), fi broblast proliferation and alveolar obliteration (both on HE-stained sections). Fibrosis distribution evaluates semi-quantitatively the amount of lung area across the whole section that is affected by the fi brotic process, and is expressed as 1, 2, 3 or 4, when 1 -25%, 26 -50%, 51 -75% or 76 -100% of the lung area is fi brotic, respectively. Severity parameters (fi broblast proliferation, collagen deposition and alveolar obliteration) were scored as 1, 2 or 3 when pathologic alterations were graded as mild, moderate or severe, respectively. Specifi cally, fi broblast proliferation evaluates the presence of spindle cells, morphologically indicative of fi broblasts. Collagen deposition evaluates the presence of green-stained collagen areas, and alveolar obliteration evaluates the amount of area characterized by the thickening of interalveolar septa that reduces alveolar spaces. For each animal, the fi brosis distribution score and the fi brosis severity parameter mean score (of six fi brosis-representative, randomly chosen, nonoverlapping high power fi elds for each section) were represented by the average of the scores obtained from the two examined sections. The overall fi brosis score for each animal was obtained by taking the sum of each mean score of the fi brosis severity parameters and multiplying this result by the fi brosis distribution mean score.

Collagen deposition image morphometric analysis
To evaluate collagen deposition quantitatively, we adopted a previously described image analysisbased system (10). Briefl y, images of MGT-stained specimens were captured with a digital camera (Olympus Camedia C-4040 ZOOM) and digitalized at 1024 ϫ 768 pixel, 24-bit/pixel resolution with a global magnifi cation of ϫ 400. Digital images were processed to identify, isolate and measure areas occupied by green-stained collagen deposition as a percentage of the total lung parenchyma area included in each fi eld. The mean value of measurements obtained from three non-overlapping fi elds per section showing fi brotic processes was assigned to each animal.

Statistical analysis
Data are expressed as median values and the relative interquartile range (IQR). In the fi gures, the data are represented by box-plot, where the band inside the box represents the median value, the box represents the IQR and whiskers show the distribution of the data. Comparisons between values collected at day 9 with those collected at day 14 for the same experimental group was performed by using the Mann -Whitney test (Table I)

. Comparisons
In contrast, in samples from mice treated with AMTC-CM, the scores for all of these parameters remained similar to those observed at day 9 and were signifi cantly lower than those observed in both control groups, with the exception of the alveolar obliteration parameter, where the difference between the samples obtained from AMTC-CM-treated animals and the non-CM-treated controls did not reach statistical signifi cance ( Figure 1F -H). Specifically, when compared with the other two groups, AMTC-CM-treated mice showed lower fi broblast proliferation [0.8 (IQR 0.4) versus 1.6 (IQR 1.0) for the Bleo group ( P Ͻ 0.05), and versus 1.5 (IQR 1.2) for the Bleo ϩ non-CM group ( P Ͻ 0.01); Figure 1F], collagen deposition [1.4 (IQR 0.5) versus 2.00 (IQR 1.2) for the Bleo group ( P Ͻ 0.05), and versus 2.2 (IQR 1.0) for the Bleo ϩ non-CM group ( P Ͻ 0.01); Figure 1G] and alveolar obliteration [2.3 (IQR 0.8) versus 3.2 (IQR 0.5) for the Bleo group ( P Ͻ 0.05), and versus 3.1 (IQR 1.2) for the Bleo ϩ non-CM group; Figure 1H]. No differences were observed among scores for any of the parameters measured in lung samples from mice that were left untreated or were treated with non-CM ( Figure 1F -H).
Quantitative image morphometric analysis (10) of fi brotic areas confi rmed the above fi ndings and showed a signifi cant reduction in collagen deposition at day 14 in samples from the mice treated with AMTC-CM compared with the two control groups, with a median value of 6.6% (IQR 5.5) versus 13.2% (IQR 7.5) for the Bleo group ( P Ͻ 0.05) and 13.2% (IQR 9.6) for the Bleo ϩ non-CM group ( P Ͻ 0.05) (Figure 2A, B).
Comparison of scores for each lung fi brosis parameter at the two observation time-points revealed signifi cant worsening of all parameters in samples from mice from the two control groups, with the single exception of alveolar obliteration, which did not increase signifi cantly from day 9 to day 14 in the mice that were left untreated. Conversely, samples from lungs of mice treated with AMTC-CM showed no signifi cant increase in any of the four measured lung fi brosis parameters (Table I).
These results were also confi rmed with the analysis of the overall fi brosis score, which indicated signifi cant reduction in severity and prevention of progression of lung fi brosis in mice treated with AMTC-CM 14 days after bleomycin instillation. These mice showed a score of 4.9 (IQR 6.6) compared to a score of 20.8 (IQR 15.4), ( P Ͻ 0.01) and 16.7 (IQR 19.3) ( P Ͻ 0.05) observed in untreated and non-CM-treated control mice, respectively (Figure 3).
The experimental design used in our studies was based on data collection 9 and 14 days after intratracheal bleomycin instillation. The use of these time-points avoids the confounding aspects of the between more than two groups (in order to compare results obtained at each time-point for the three experimental groups) were performed by using the Kruskal -Wallis non-parametric analysis of variance. In cases of statistical signifi cance, the Mann -Whitney test using Holm -Bonferroni ' s correction for multiple comparisons was applied to assess the differences between experimental groups at the specifi c timepoints (at day 9 or at day 14). A P -value Ͻ 0.05 was considered statistically signifi cant. Statistical analysis was performed with SPSS Advanced Statistics 13.0 (SPSS Inc., Chicago, IL, USA).

Results and discussion
The aim of this study was to investigate the potential benefi cial effects of AMTC-CM on bleomycininduced lung fi brosis. We analyzed the effects of AMTC-CM administration on fi brosis distribution, a parameter that reveals the amount of lung area affected by the fi brotic process across the entire section examined. At day 9, about one-third of the lung area was affected by the fi brotic process in all three experimental groups, with no differences observed among the groups ( Figure 1A). At day 14, in mice left untreated and those treated with non-CM, the fi brotic process had developed further and occupied about two-thirds of the total area examined. The scores for fi brosis distribution in these two groups did not differ ( Figure 1E). In contrast, samples obtained at this time-point from the group of mice treated with AMTC-CM showed that only one-third of the lung area remained affected by the fi brotic process, which resulted in a fi brotic score that was signifi cantly reduced compared with the two control groups, with a median score of 1.00 (IQR 0.9) versus 3.0 (IQR 1.8) for the Bleo group ( P Ͻ 0.05) and 2.5 (IQR 2.0) for the Bleo ϩ non-CM group ( P Ͻ 0.05) ( Figure 1E).
To investigate further the potential anti-fi brotic actions of AMTC-CM, each lung area that had previously been identifi ed as fi brotic was analyzed and scored for fibroblast proliferation, collagen deposition and alveolar obliteration, the three main pathologic hallmarks used to describe the severity of bleomycin-induced fi brotic alterations (24). At day 9, the median scores for each of the three parameters were not signifi cantly different among the three experimental groups ( Figure 1B -D), with the exception of the score for fi broblast proliferation, which was lower in the samples from mice treated with non-CM compared with mice left untreated ( Figure 1B). This difference was no longer evident at day 14 ( Figure 1F). However, at day 14, samples from mice of the control groups showed worsening fi brosis, with more severe fi broblast proliferation, increased collagen accumulation and alveolar obliteration ( Figure 1F -H).
show signifi cant differences in the fi brotic process compared with that described at day 14 (A. Cargnoni et al. , unpublished observations).
In addition to the same exhaustive histologic assessment of fi brosis on the whole lung tissue early infl ammatory response and allows scoring of the lung fi brotic process at its intermediate and maximum levels, as we and others have demonstrated previously (2,3,6,25 -27). Indeed, studies performed 21 and 28 days after bleomycin instillation do not  It is noteworthy that, even though AMTC-CM was injected into the right lung, a reduction in pulmonary fi brosis was observed in both the left and right lung fi elds, with no detectable differences between the two lungs in each mouse. It is conceivable that AMTC-CM injected into the right lung can reach the left lung through the exchange of fl uid that occurs between communicating airways during respiratory function or, possibly, via the systemic circulation after absorption by alveolar capillaries, as we have detected donor cells in the left lungs of mice after injection of amniotic membrane-derived cells in the right lung (Cargnoni et al., unpublished observations).
Because virtually no effects on the fi brotic process were observed in mice receiving administration of non-CM, we conclude that AMTC-CM contained soluble molecules released by AMTC that were responsible for the reduction in severity and progression of the fi brotic reaction. Considering that increasing evidence has highlighted recently that MSC isolated from various sources produce bioactive molecules, which are potentially able to exert several types of paracrine effects (e.g. anti-scarring, anti-infl ammatory, anti-apoptotic) on target cells (29), and that injection of CM obtained from MSC is an effective experimental treatment for different tissue injuries (e.g. 18,19), interesting results will no doubt be obtained from comparative studies in the same animal model that will investigate whether CM generated from cells derived from sources other than the amniotic membrane can also reduce lung fi brosis in a manner similar to AMTC-CM.
Determination of the nature and characteristics of the paracrine soluble molecules involved in AMTC-CM-mediated fi brosis reduction, as well as their mechanism(s) of action and their target cells, (without subsampling the lungs) of experimental animals that we have applied in previous studies (6), the analysis of our current experiments was enhanced by the addition of a quantitative evaluation of collagen deposition performed by a digital image morphometric analysis (28). Our fi ndings show that the administration of AMTC-CM prevents the progression of bleomycin-induced lung fi brosis and results in a signifi cant reduction in the parameters related to the fi brotic process at day 14 after intratracheal bleomycin instillation. These results are consistent with the benefi cial effects of transplanted fetal membrane-derived cells that we have observed previously in the same animal model (6), and provides  A European patent application has been fi led with the application number PCT2008-004845.

Disclosure of interest:
No competing fi nancial interests exist.
is obviously a task of major importance that warrants further extensive investigation and is the object of ongoing studies. Similarly, trying to understand the mechanisms underlying the benefi cial effects of paracrine actions performed by stem cells in general remains a challenge for all researchers in the fi eld.
It is tempting to speculate that a cell-free treatment, based on the use of AMTC-CM, could potentially replace cell transplantation, particularly when tissue repair via paracrine bioactive molecules rather than tissue regeneration via cell replacement/differentiation is required, with this strategy also offering a series of added advantages. In particular, AMTC-CM can be produced easily and in large quantities; it can be stored effi ciently because it maintains its anti-fi brotic effi cacy after the lyophilization process; as a cell-free treatment, it can drastically reduce the risk of adverse immunologic reactions, infectious risks and other potential long-term negative effects caused by the presence of exogenous cells; fi nally, it is also conceivable that AMTC-CM could be administered safely via intravenous injection, avoiding clot formation and lung capillary entrapment.
In conclusion, the present study contributes to the body of knowledge regarding the properties of amniotic membrane-derived cells (and specifi cally of AMTC) and envisages a new potential use for the multifaceted immunomodulatory and trophic features of these cells through the application of cell-free preparations obtained from culture medium generated by these cells. Despite the fact that this study does not contribute to the identifi cation of which paracrine factor(s) produced by AMTC and released into the CM could be involved in the observed reduction in fi brosis, and also the fact that further studies are required to understand the paracrine mechanisms involved, as well as to investigate the effects of AMTC-CM on already established fi brosis and compare the effects of CM produced by other cell types, our results point to AMTC-CM as a potential new tool that deserves consideration in the development of novel strategies and clinical applications for lung fi brosis.