Amniotic MSCs reduce pulmonary fibrosis by hampering lung B‐cell recruitment, retention, and maturation

Abstract Growing evidence suggests a mechanistic link between inflammation and the development and progression of fibrotic processes. Mesenchymal stromal cells derived from the human amniotic membrane (hAMSCs), which display marked immunomodulatory properties, have been shown to reduce bleomycin‐induced lung fibrosis in mice, possibly by creating a microenvironment able to limit the evolution of chronic inflammation to fibrosis. However, the ability of hAMSCs to modulate immune cells involved in bleomycin‐induced pulmonary inflammation has yet to be elucidated. Herein, we conducted a longitudinal study of the effects of hAMSCs on alveolar and lung immune cell populations upon bleomycin challenge. Immune cells collected through bronchoalveolar lavage were examined by flow cytometry, and lung tissues were used to study gene expression of markers associated with different immune cell types. We observed that hAMSCs increased lung expression of T regulatory cell marker Foxp3, increased macrophage polarization toward an anti‐inflammatory phenotype (M2), and reduced the antigen‐presentation potential of macrophages and dendritic cells. For the first time, we demonstrate that hAMSCs markedly reduce pulmonary B‐cell recruitment, retention, and maturation, and counteract the formation and expansion of intrapulmonary lymphoid aggregates. Thus, hAMSCs may hamper the self‐maintaining inflammatory condition promoted by B cells that continuously act as antigen presenting cells for proximal T lymphocytes in injured lungs. By modulating B‐cell response, hAMSCs may contribute to blunting of the chronicization of lung inflammatory processes with a consequent reduction of the progression of the fibrotic lesion.


| INTRODUCTION
The pathogenesis of idiopathic pulmonary fibrosis (IPF) is still poorly understood, thus heavily contributing to the lack of an effective cure.
There is growing evidence suggesting a mechanistic link between inflammation, which endures to the end stage of IPF, and the development and progression of fibrosis. [1][2][3][4] Interestingly, we and others have reported that cells derived from the amniotic membrane of human term placenta, such as amniotic epithelial cells and amniotic mesenchymal stromal cells (hAMSCs) 5-10 and related secretome, [11][12][13] can prevent and reduce the progression of pulmonary fibrosis when injected in bleomycin-instilled rodents, a model which closely resembles human IPF. 14
Herein, we sought to explore whether hAMSC treatment affects the levels and phenotype of immune cell populations in alveolar spaces and in lung tissue during the course of bleomycin-induced lung injury, and its potential correlation to therapeutic outcome.

| hAMSCs isolation, culture, and characterization
Twelve amniotic membranes were collected and processed as previously described. 28 Briefly, the amniotic membrane was manually separated from the chorion, rinsed in saline solution containing 100 U/mL penicillin and 100 μg/mL streptomycin (both from Sigma-Aldrich, St. Louis, Missouri), fragmented (~3 × 3 cm 2 ) and then digested at 37 C with 2.5 U/mL dispase (Corning, New York). After 9 minutes, the amnion was washed in RPMI 1640 medium (Sigma-Aldrich) supplemented with 10% fetal bovine serum (FBS; Sigma), 2 mM L-glutamine (Sigma-Aldrich), and P/S (herein referred to as complete RPMI medium), and finally digested for 2.5 hours with 0.94 mg/mL collagenase and 0.01 mg/mL DNase (both from Roche, Basel, Switzerland). Isolated cells were filtered (100 μm strainer, BD Falcon, Bedford, Massachusetts), collected by centrifugation, resuspended in complete RPMI medium, and refiltered (70 μm strainer BD Falcon). All cell batches were used as individual preparations that were cryopreserved and/or in vitro cultured.
Cells whose viability after thawing was higher than 85% were used. Nonexpanded and in vitro expanded cell preparations were used in order to study the possible impact of in vitro culture on hAMSC therapeutic effects.

Significance statement
The immunomodulatory features of amniotic cells can create a microenvironment able to limit the evolution of chronic inflammation to fibrosis. However, the immune modulation induced by amniotic mesenchymal stromal cells (hAMSCs) in models of fibrosis has yet to be elucidated. For the first time, this study shows that in bleomycin-challenged mice, hAMSCs control pulmonary B-cell recruitment, retention, maturation, and reduce the formation and expansion of lung lymphoid aggregates. By modulating B-cell response, hAMSCs hamper the self-maintaining inflammatory condition promoted by B cells in injured lungs and may contribute to limiting the chronicization of lung inflammation that evolves into the fibrotic lesion.

| Sample collection
Each animal was anesthetized as described above and euthanized at the scheduled time point by exsanguination. Blood was drained from the abdominal aorta and bronchoalveolar lavage (BAL) was collected using a total volume of 2.0 mL of sterile PBS injected i.t. with a 20-gauge cannula. BAL fluids were centrifuged (1000g for 10 minutes, at 4 C) and cells were frozen in 90% FBS+10% DMSO for flow cytometry analysis.
Lungs were explanted and sectioned into the five individual lobes as previously described. 12 Each lobe was further sectioned into two equivalent hemilobes. One series of hemilobes was formalin-fixed (10% neutral formalin from Bio-Optica, Milano, Italy) for 48 hours at room temperature and processed for microscopic analyses. The other series of pooled hemilobes was snap-frozen in liquid nitrogen and stored at −80 C for real-time polymerase chain reaction (RT-PCR) analysis.

| Microscopy and image analysis
Lung hemilobes were paraffin-embedded and consecutive 4-μm-thick sections were cut. Sequential, nonoverlapping images were captured from whole hematoxylin and eosin or Masson's trichrome-stained sections with a digital camera (Olympus Camedia C-4040 ZOOM) in bright-field light microscopy (Olympus BX41, Tokyo, Japan) at 40× magnification. Color digital images obtained from each hemilobe were converted by the FiJi software (https://imagej.nih.gov/ij) to binary data, and the percentage of each alveolar hemilobe pixels to whole hemilobe pixels was calculated. The area occupied by alveoli of the entire lung was the sum of all hemilobe alveolar areas and was expressed as a percentage of total area of the entire lung section. 9,29 All analyses were performed in a blinded manner by a veterinary pathologist.   and relative gene expression was quantified using the 2 −ΔCT method.

| Flow cytometry analysis
Primer sequences are shown in Supporting Information Table S1.

| Statistical analysis
All data are presented as mean ± SE of three independent experiments.
Differences between groups were analyzed by one-way analysis of variance, followed by Tukey's multiple comparisons test. A P-value <.05 was considered statistically significant. Statistical analyses were performed using the Prism 6.05 software (Graphpad software Inc., La Jolla, California).

| hAMSCs ameliorate bleomycin-induced lung injury
To explore the potential benefits exerted by hAMSCs, we evaluated bleomycin-induced lung injury by measuring residual alveolar area and gene expression of markers indicative of alveolar integrity (eg, podoplanin) 33  . Density of CD80 expression was evaluated as median fluorescence intensity ratio between positive and negative cells. Data from mRNA levels analysis are expressed as log 2 of the fold change from the saline-instilled group. These data are reported as mean ± SE of n = 4 (day 4) and n = 5-7 (days 7, 9, and 14) samples. *P < .05, **P < .01 compared to control (Bleo+PBS) group 9 ( Figure 1F), and a decrease in expression of extracellular matrix proteins such as fibronectin ( Figure 1G) and collagen ( Figure 1H) occurred at all time points.

| hAMSCs alter antigen presenting cells in bleomycin-induced inflammation
To investigate a possible link between the immune-modulatory properties of hAMSCs and their ability to limit the progression of lung fibrosis, we studied the capacity of hAMSCs to reduce bleomycininduced recruitment of inflammatory cells into alveolar spaces and lung tissue, as well as their polarization and maturation.
Bleomycin challenge induced a recruitment of inflammatory cells into alveolar spaces. hAMSC treatment did not affect the alveolar recruitment of CD45 + cells at any time point (see Supporting Information Table S2); however, relative differences among CD45 + cell subpopulations were observed.
Bleomycin instillation produced marked alterations of immune populations present in alveolar spaces. In bleomycin-instilled animals, F I G U R E 3 Human amniotic mesenchymal stromal cells (hAMSCs) alter T lymphocyte subsets in bleomycin-induced inflammation. Bleomycin instillation produced recruitment of T lymphocytes in alveolar spaces (A, Bleo+PBS group), and increased both CD4 + and CD8 + T-cell subpopulations (B,C, Bleo+PBS group). hAMSC treatment slightly and transiently increased T lymphocyte recruitment (A-C). Panels D-F report lung mRNA levels of markers associated with distinct subsets of T lymphocytes including Tbet (related to Th1 cells), GATA3 (Th2), and ROR-γτ (Th17). Bleomycin challenge increased the expression of these markers (D-F, Bleo+PBS group). hAMSC treatment transiently increased expression of Tbet (D, day 4) and of GATA3 (E, day 7), while reducing ROR-γτ gene expression (F, day 4). Panels G-I report gene lung expressions of cytokines secreted by Th1 T-cell subset (IL-1β and IFN-γ, G,H) and Th2 subset (IL-4, I). Bleomycin challenge increased expressions of IL-1β and IFN-γ (G,H, Bleo-PBS group), except IL-4 which is reduced compared to saline instillation (I, Bleo+PBS group). Treatment with hAMSCs did not change lung expression of IFN-γ (G), reduced that of IL-1β (H), and increased that of IL-4 (I), suggesting a polarization toward Th2 T-cell subset. Panels J-L show alveolar levels of T regulatory (Treg) cells and the expression of markers related to an anti-inflammatory milieu. Compared to saline-instilled group, animals challenged with bleomycin displayed increased levels of alveolar Treg (J, Bleo+PBS group), accompanied by increased gene expression of lung Foxp3 (K, Bleo+PBS group) and IL-10 (L, Bleo+PBS group). This effect was further strengthened by treatment with hAMSCs, that increased lung expression of Foxp3 (K, days 4 and 7) and of IL-10 (L, days 7 and 9). Alveolar levels of T lymphocytes were analyzed by flow cytometry and expressed as percentage of viable CD45 positive cells in bronchoalveolar lavage collected from each mouse. Data from mRNA levels analysis are expressed as log 2 of the fold change from the saline-instilled group. These data are reported as mean ± SE of n = 4 (day 4) and n = 5-7 (days 7, 9, and 14) samples. *P < .05, **P < .01 compared to control (Bleo+PBS) group we observed a marked decline of relative amount of resident alveolar macrophages, identified as CD64 + CD11c + Siglec-F + cells, which represent the predominant cell population in normal conditions. Their percentage fell from 78.2% ± 2.4% in the saline control group, to 41.8% ± 5.8% and 8.6% ± 2.9% after 4 and 7 days from bleomycin instillation, respectively. In contrast, the percentage of not resident monocyte-derived macrophages (identified as CD64 + CD11c + Siglec- Altogether, these data suggest that hAMSC treatment promotes a shift toward immune phenotypes with lower inflammatory potential by decreasing CD80 expression and density on macrophages and dendritic cells, and by increasing expression of macrophage M2 markers.

| hAMSCs alter T lymphocyte subsets in bleomycin-induced inflammation
We also studied the ability of hAMSCs to counteract bleomycininduced alterations in alveolar and lung T-cell environment. Alveolar CD3 + T lymphocytes, which represent only 1.9% ± 0.9% of the total CD45 + population in saline-instilled mice; increased to 13.9% ± 3.0% (day 4) and to 47.9% ± 8.1% (day 14) in bleomycininstilled animals ( Figure 3A), with higher percentage of CD4 + cells compared to CD8 + cells ( Figure 3B,C). hAMSC treatment did not reduce T lymphocyte recruitment, rather at day 4 it slightly increased the early recruitment of CD3 + cells, including both CD4 + and CD8 + subpopulations ( Figure 3A-C).
To reveal whether hAMSC treatment differently affects distinct inflammatory subsets of CD4 + T cells, we determined the lung gene expression of markers associated to Th1 (Tbet), Th2 (GATA3), and Th17 (ROR-γτ) T-cell subsets and of cytokines with proinflammatory action (IFN-γ, IL-1β, and IL-4). As shown in  Figure 3D) and of GATA-3 at day 7 ( Figure 3E). Instead, they transiently reduced Th17 response, as indicated by declined expression of ROR-γτ at day 4 ( Figure 3F). Moreover, the treatment with hAMSCs decreased the expression of IL-1β ( Figure 3G) while increased the expression of IL-4 ( Figure 3H). The treatment did not affect IFN-γ expression ( Figure 3I).
We also investigated the ability of hAMSCs to promote an immune suppressive milieu, by detecting the levels of alveolar T regulatory (Treg) cells, lung expression of the Treg marker Foxp3 and of IL-10, a cytokine with important anti-inflammatory activity. hAMSC treatment did not consistently affect alveolar Treg cell levels ( Figure 3J), however it augmented expression of Foxp3 at days 4 and 7 ( Figure 3K) and of IL-10 at days 7 and 9 ( Figure 3L) in lungs.
It is of note that hAMSC treatment did not affect neutrophil alveolar recruitment, whose relative levels transiently peaked at day 4 after bleomycin instillation (data not shown).

| hAMSCs reduce bleomycin-induced recruitment and maturation of B cells
To refine the picture of the inflammatory environment in our mouse model of fibrosis, we next focused on the effects exerted by bleomycin , represented approximately 75% of the total aggregates (C, gray bars). Between days 9 and 14, the aggregates with a higher presence of B cells (from 41% to 70%) increased (C, black bars). hAMSC treatment prevented the enrichment in B cells in lung lymphoid aggregates (C; Bleo+hAMSC/P0 and Bleo+hAMSC/P2 groups). Data are reported as mean ± SE of n = 4 (for control and treated groups at day 9) and n = 5 (for control and treated groups at day 14). A, *P < .05 compared to control (Bleo+PBS) group. B, *P < .05 by comparing number of aggregates with small size (>5000 <10 000 μm 2 ) between control (Bleo+PBS) group and treated groups (Bleo+hAMSC/P0 and Bleo+hAMSC/ P2). #P < .05 by comparing number of aggregates with intermediate size (>10 000 <20 000 μm 2 ). C, *P < .05, **P < .01 by comparing number of aggregates with low presence of B cells (<10 %) between control (Bleo+PBS) group and treated groups (Bleo+hAMSC/P0 and Bleo+hAMSC/P2).

| hAMSCs compromise the formation and expansion of bleomycin-induced lymphoid aggregates
Given that intrapulmonary B-cell aggregates have been found in lungs of patients with inflammatory/fibrotic respiratory diseases, 4 we evaluated B lymphocyte distribution in lungs from bleomycin-challenged mice.
Hematoxylin and eosin analysis of lung sections evidenced the presence of numerous lymphoid aggregates mainly located within the peri-bronchial areas ( Figure 5A,B, arrows). These aggregates were primarily composed of CD3 + T and B220 + B lymphocytes mixed in variable proportions ( Figure 5C-F).
Next, we examined the ability of hAMSCs to influence the expansion of single lymphoid aggregates. At day 9, most of the aggregates ranged from 5000 to 10 000 μm 2 ( Figure 6B, upper panel), whereas at day 14 more than 25% of the aggregates had an intermediate size (from 10 000 to 20 000 μm 2 ) and some bigger aggregates appeared (higher than 20 000 μm 2 ; Figure 6B Figure 6C, gray bars). Between days 9 and 14, the aggregates with a F I G U R E 7 Human amniotic mesenchymal stromal cell (hAMSC) treatment reduces the expression of chemokines/cytokines involved in pulmonary recruitment and retention of B lymphocytes. Compared to saline-instilled mice, bleomycin challenge upregulated the gene expression of lymphotoxin (A, Bleo+PBS group), CCL21 (B), CXCL12 (C), and CXCL13 (D), which promote the recruitment and retention of lymphocytes into lung tissue. Expression of lymphotoxin, CCL21, and CXCL12 were early and transiently increased, whereas CXCL13 expression (the most effective chemoattractant for B cells) was high at all time points. hAMSC treatment accelerated the normalization of expression of lymphotoxin, CCL21 and CXCL12, and reduced CXCL13 expression at all time points. In addition, bleomycin challenge increased BAFF gene expression (E, Bleo +PBS group), a cytokine with a crucial role for B-cell survival. hAMSC treatment reduced BAFF expression at days 4 and 7 (E). Data are expressed as log 2 of the fold change expression from the saline-instilled group, and are reported as mean ± SE of n = 3 (day 2), n = 4 (day 4), and n = 5-7 (days 7 and 9). *P < .05, **P < .01, ***P < .01 compared to control (Bleo+PBS) group higher presence of B cells (from 41% to 70%) increased from 18.5% ± 2.3% (at day 9) to 26.6% ± 5.2% ( Figure 6C, black bars). Interestingly, at day 14, hAMSC treatment reduced the number of B cells into the lymphoid aggregates ( Figure 6C). Overall, these data indicate that hAMSCs blocked the formation, expansion, and B-cell enrichment of lung lymphoid aggregates.
To provide insight onto how hAMSCs affected B-cell lymphoid aggregates, we investigated B-cell recruitment, survival, and proliferation after treatment.
To investigate whether hAMSCs can affect the recruitment of lymphocytes in lung tissues, we performed gene expression analysis of the main chemokines involved in recruitment and retention of lymphocytes (lymphotoxin, CCL21, CXCL12, and CXCL13), and known to contribute to the generation of lymphoid aggregates in various organs. 34 In addition, the presence of lymphoid aggregates has been described in the lungs of mice challenged with bleomycin, 40 of mice genetically modified (CCR7−/− mice) showing an impaired homing of Treg cells, 41 and of mice with rheumatoid arthritis-related interstitial lung disease. 42 Here, we also detected intrapulmonary aggregates composed of T and B lymphocytes in bleomycin-instilled animals and we demonstrated that hAMSC treatment inhibited the formation and expansion of these aggregates.
Although the role of lung lymphoid aggregates is controversial, a growing body of clinical evidence strongly points at the contribution of intrapulmonary T/B-cell aggregates in the initiation and/or progression of a range of fibrotic diseases such as IPF 1,2,43,44 and nonspecific interstitial pneumonia. 45 In the lungs of IPF patients, B cells form focal aggregates together with T lymphocytes and their number has been correlated with the development and severity of lung fibrotic lesions. 2,45 In addition, as observed herein, B lymphocytes present in these aggregates do not proliferate, 1,2,40,43 suggesting that recruitment from systemic circulation likely represents the main mechanism leading to lymphocyte accumulation in lung parenchyma. Importantly, our data suggest that hAMSC treatment can reduce B-cell recruitment and homing by inhibiting the lung expression of homeostatic lymphoid chemokines, essentially CXCL13, a crucial chemoattractant for B cells, 36 and by reducing BAFF, that holds a major role in B-cell survival and maturation and can thereby promote lymphoid aggregate formation. 46 It is of note that in IPF patients, increased plasma concentrations of BAFF (or BLyS) correlated with clinical severity and negative outcomes of these patients 43  Considering that hAMSC treatment impaired the antigen presenting potential of dendritic cells and monocyte-derived macrophages, we hypothesize that this also may contribute to the ability of hAMSCs in controlling lymphoid aggregate formation/expansion and fibrosis progression in bleomycin-injured lungs.
Altogether, we propose that in bleomycin-challenged mice, hAMSCs create an anti-inflammatory microenvironment partially mediated by their ability to control recruitment, retention, and maturation of B cells in diseased lungs. Within the lymphoid aggregates, B cells can continuously act as antigen presenting cells for the adjacent T lymphocytes and can cause cytotoxicity by producing autoantibodies. hAMSCs can resolve this loop and hence break the selfmaintaining inflammatory condition promoted by B cells. 43,44,49 Finally, in an attempt to support the clinical application of hAMSC-based therapy, this study tried to address the important question on the impact of in vitro expansion on hAMSC therapeutic effects. In this study, indeed two different cell preparations were used: freshly isolated hAMSCs (hAMSC/P0) and hAMSCs expanded in vitro to passage 2 (hAMSC/P2). In bleomycin-induced pulmonary injury model, both treatments displayed similar effects on the immune populations investigated in this study, suggesting that at least a shortterm expansion in vitro does not alter cell activity.

| CONCLUSION
Our work provides key insights into the therapeutic potential of hAMSCs from an immunological perspective, providing further evidence for a potential clinical translation of hAMSCs in inflammation-related fibrotic diseases. In support of this, short-term in vitro expansion (passage 2) did not alter the activity of hAMSCs in comparison to freshly isolated cells (passage 0), thus further encouraging the translation of this therapeutic product into the clinic. However, given the need of high numbers of cells also for potential repeated treatments, studies elucidating the influence of in vitro long-term expansion on cellular functions are needed.

DATA AVAILABILITY STATEMENT
All relevant data are available from the corresponding author upon reasonable request.