Mesenchymal stem cell secretome alters gene expression and upregulates motility of human endometrial stromal cells.

hTERT-immortalized human endometrial stromal cells ((HESCs), ATCC) were grown in phenol red-free DMEM/F12 media (Gibco) supplemented with 10% charcoal stripped fetal bovine serum (FBS), 1% ITS+premix (Corning) and 500 ng/mL puromycin (Sigma) and maintained in a humidified incubator under standard culture conditions, as per manufacturer protocols. Human BM-derived mesenchymal stem cells (ATCC) were derived from a 35-year-old Hispanic female donor. These commercially obtained BM-MSCs, designated BM-MSC-1, were received at passage 2, expanded and cultured in MSC basal media (ATCC) supplemented with MSC Growth kit (ATCC) [hereafter referred to as “MSC media”] as per manufacturer protocols. Given cells received at passage 2 may remain contaminated with other nonhematopoietic cell types, adherent cells were serially passaged at 80% confluence until the fourth passage, at which time the cells were confirmed by phenotype and multilineage differentiation potential as described below. Confirmed MSCs were used for all experiments prior to passage 6. BM aspirates and UC were collected using a protocol approved by the Institutional Review Board of Rutgers Biomedical and Health Sciences. Three healthy women volunteers, age 20–30 years, provided written informed consent to undergo BM aspiration for use of their BM cells. Two women, age 20–30 years, provided written informed consent for the procurement and use of placental tissues. All subjects were generally healthy. Specific inclusion criteria for UC samples included cesarean delivery of a full-term male infant. Exclusion criteria for both samples included infectious disease, obese or underweight body mass index, chronic medical conditions, acute medical conditions at time of sample collection, and pregnancy-related complications for UC donors. Fresh BM aspirates were collected into heparinized syringes for transport to the laboratory and anonymized as BM-MSC-2, BM-MSC-3, and BM-MSC-4. MSCs were cultured from BM aspirates, as described ( Potian et al., 2003 ). Briefly, unfractionated BM was diluted 1:1 in DMEM with 10% FBS and added to vacuum gas plasma treated plates (BD Falcon). After three days, red blood cells and granulocytes were removed by Ficoll-Hypaque density gradient centrifugation, and the mononuclear cells were returned to the plates with fresh media. At weekly intervals 50% of the media was replaced with fresh media until the adherent cells reached 80% confluence. Adherent cells were serially passaged at 80% confluence until the fourth passage, at which time the cells were confirmed by phenotype and multilineage differentiation potential as described below. Confirmed MSCs were used through passage 6. Human placentas were collected after full-term cesarean delivery of healthy, full-term male infants. We used placental tissues from male infants to ensure MSCs of fetal rather than maternal origin, confirmed by PCR to determine SRY mRNA expression (data not shown). Placentas were transported to the lab in 1x phosphate buffered saline (PBS); placentas were processed within 1 hour or stored at 4°C for up to 18 hours prior to processing. Upon processing the samples were anonymized as UC-MSC-1 and UC-MSC-2. Within 24 hours of delivery the UC was excised from the placenta and washed with PBS to remove blood and mucus, transferred to petri dishes and cut into 1×1 cm pieces. UC-MSC isolation was perfomed as described by our laboratory previously ( Romagano et al., 2022 ). The outermost region of the UC was used for UC-MSC isolation; the Wharton’s jelly, and vein and artery region were discarded. Up to 15–20 UC pieces were transferred to tissue culture dishes, ensuring adequate spacing between each piece, and left without media for 10–15 min to attach. After that time, 10 mL media (Dulbecco’s modified Eagle’s media (DMEM, 1 g/L D-Glucose; Invitrogen) containing 10% fetal bovine serum (FBS; Hyclone) and 1% penicillin/streptomycin (Invitrogen) was added to each culture dish. Medium was fully replaced with fresh medium every 2–3 days until outgrowth from the UC pieces was visible. The sections were then transferred to a new culture dish. Up to 7 serial transfers were performed. Culture dishes with outgrowth were left for ~1 week until expansion of the cells was observed, this step was counted as passage 0. Adherent cells were serially passaged at 80% confluency and cryopreserved through passage 3. At the fourth passage cells were confirmed to be MSCs by phenotyping and multilineage differentiation capacity as described below. Confirmed MSCs were used through passage 6. Human BM-MSCs and UC-MSCs were phenotyped by flow cytometry. All conjugated antibodies were purchased from Becton Dickinson. MSCs were confirmed to be negative for CD45 and positive for CD44, CD90, CD73, and CD29 expression. All antibodies were used at 1:100 dilution and analyzed using FACScalibur flow cytometer and FlowJo software (Becton Dickinson). Multilineage differentiation protocols were performed as previously described ( Potian et al., 2003 ) and as described below. MSCs were seeded in a 6-well plate at 10,000 cells per cm 2 in MSC media, and incubated in a 37°C, 5% CO 2 humidified incubator overnight. The following day the cells were washed with PBS and media changed to human osteoblast differentiation media (Sigma-Aldrich). Osteoblast differentiation media was changed every 3 days for a period of 14 days as per the manufacturer’s guidelines. The mineral deposition of osteoblast cultures was visualized by Von Kossa staining for calcium. Briefly, cells were fixed with 4% PFA in PBS, incubated with 2% silver nitrate solution in the dark, washed with deionized water, and developed with 1% pyrogallol. Cells were counterstained with nuclear fast red (Vector). Osteoblast nodules containing calcium mineral stain black; less developed nodules stain yellow to brown. MSCs were seeded in wells of 6-well plates at 40,000 cells/cm 2 in MSC media and incubated in a 37°C, 5% CO 2 humidified incubator for 3 days. The cells were washed with PBS and media replaced with human adipocyte differentiation media (Sigma Aldrich). The medium was replaced with fresh human adipocyte differentiation media every 3 days as per the manufacturer’s protocol. Within 3–4 weeks cells were differentiated into adipocytes with large lipid droplets. Cells were fixed with 4% PFA in PBS, washed twice with water and incubated with 60% isopropanol. Cells were then incubated with Oil Red O Working Solution (Sigma Aldrich) and counterstained with hematoxylin prior to visualization with a light microscope. Cell proliferation assay was performed as described by our laboratory previously ( Perlman et al., 2021 ). A transwell co-culture system using 0.4μm pore size cell culture inserts was utilized to assess the impact of MSC secretome exposure on HESC proliferation. This small pore size was selected to allow for passage of the MSC secretome through the insert, while excluding the MSC ( Lipowsky et al., 2018 ). BM-MSCs, UC-MSCs, and HESCs were cultured in their respective media in 100 mm tissue culture dishes, grown to 70–80% confluency, and then harvested and washed with PBS. HESCs were transferred to 6-well plates containing 12mm diameter glass circles (Fisher) at 2×10⁵ cells per well. BM-MSCs and UC-MSCs were placed on 6-well plate formatted culture inserts with 0.4μm pore membrane (Corning) at 1×10⁵ cells per insert. BM-MSCs, UC-MSCs and HESCs were cultured independently for 24 hours in their respective media. All cells were washed with PBS and were then exposed to co-culture media (phenol red-free DMEM supplemented with 10% charcoal stripped FBS and 1% ITS+premix) in the transwell system for 24 hours. HESCs cultured in the transwell system without MSCs served as controls. After 24h, HESCs were fixed with 4% PFA, washed with PBS, permeabilized with PBST (1x PBS with 0.1% Triton-100) and incubated with anti-Ki67 antibody (Abcam, 1:500) for 2 hours at room temperature. Cells were incubated with a secondary antibody, Alexa Fluor 488 (Thermofisher, 1:1000 dilution) for 30 minutes at 4°C. Vectashield with DAPI (Vector) was used for mounting and nuclear visualization. Slides were imaged using the Leica DM2500 LED microscope and Leica Application Suite X (LAS X) software (Leica Microsystems). Two independent observers counted the number of Ki67 positive cells and the total number of cells per field in three 40x magnification fields per slide using ImageJ software. Percentage of proliferating cells was determined as the number of Ki67 positive cells per total cell count × 100. Transwell migration assays were performed as previously described ( Perlman et al., 2021 , Pijuan et al., 2019 ). 24-well transwell chambers (Corning) and Falcon transparent cell culture inserts (8 μm pore size, Corning) were used. The 8 μm pore size is routinely used in transwell migration assays to accommodate the size of eukaryotic cells, including HESCs ( Justus et al., 2014 , Pijuan et al., 2019 , Gellersen et al., 2013 ). The upper compartment contained 5×10⁴ HESCs in phenol red-free DMEM supplemented with 0.5% charcoal stripped FBS and 1% ITS+premix. The lower compartment of the chamber contained BM-MSCs or UC-MSCs in phenol red-free DMEM supplemented with 10% charcoal stripped FBS and 1% ITS+premix. HESCs exposed to DMEM supplemented with 10% charcoal stripped FBS and 1% ITS+premix served as the control. The extent of HESC migration was assessed at 24 hours, based on previous time course experiments to determine the optimal time of MSC culture prior to co-culture with HESC (data not shown). Non-migrated HESC in the upper compartments were removed with a cotton swab. Cells on the undersurface of the membrane were fixed with 4% PFA and stained with H&E. Slides were imaged using the Leica DM2500 LED microscope and Leica Application Suite X (LAS X) software at 2.5x magnification. Two independent observers counted the nuclei. The extent of cell migration was determined in five fields per insert (performed in triplicate), by counting nuclei of cells that migrated in a 1250 × 1250 pixel area of the insert using ImageJ software. Invasion assays were performed in 24-well transwell chambers using 8 μm pore size cell culture inserts (Corning) ( Perlman et al., 2021 , Pijuan et al., 2019 ). Cell culture inserts were coated with DMEM media with 0.3 mg/mL Matrigel (phenol red-free, growth factor reduced basement membrane matrix, Corning). Experiments were performed in the exact same manner as described for the transwell migration assays except 1× 10 5 HESCs were added to the upper compartment of the chambers. The extent of HESC invasion was assessed at 24 hours. Non-invaded cells in the upper compartments were removed with a cotton swab. Cells on the undersurface of the membrane were fixed with 4% PFA and stained with H&E. Slides were imaged using the Leica DM2500 LED microscope and Leica Application Suite X (LAS X) software at 2.5x magnification. Two independent observers counted the nuclei. The extent of cell invasion was determined in five fields per insert (performed in triplicate), by counting nuclei of cells that migrated in a 1250 × 1250 pixel area of the insert using ImageJ software. For mRNA-seq experiments, the transwell co-culture system utilized to assess HESC proliferation was used. 2 × 10 5 HESC cells per well were plated in 6-well dishes (Sigma-Aldrich). As performed for the proliferation assays, BM-MSCs were plated onto 0.4 μm pore size cell culture inserts, to allow for passage of the MSC secretome. Total RNA was isolated from HESCs with the RNeasy Plus Mini Kit (Qiagen) after 24 hours of co-culture with BM-MSCs and RNA integrity number (RIN) determined. Three biological replicates of control HESCs were compared to three biological replicates of HESCs exposed to either BM-MSC-1 or BM-MSC-2. This was performed to remove batch effect between sequencing rounds of the same biological samples( Sandiford et al., 2021 ). The 24h time point was chosen based on significant upregulation of HESC motility after 24h co-culture with BM-MSC-1 and BM-MSC-2 as demonstrated by the transwell migration and invasion assays. mRNA-Seq library construction was done using NEBNext UltraRNA Library Prep Kit for Illumina (NEB) at Novogene Corporation Inc. The multiplexed library was sequenced on an Illumina Novaseq 6000 platform (Illumina) in a 2 × 150 bp configuration to a target depth of 20–30 M reads per sample. Sequence analysis was performed following NF-Core RNA-Seq guidelines (v1.4.2) for rigor and reproducibility ( Ewels et al., 2020 ). The raw reads were aligned to the human genome (GRCh37) using STAR (v2.6.1), followed by transcript abundance and hit counts generation through StringTie (v2.0) and featureCounts (v1.6.4) ( Liao et al., 2014 , Pertea et al., 2015 ). Subsequent hit count normalization and pairwise differential gene expression comparisons were performed using the DESeq2 (v1.22.1) method with significance thresholds set at FDR adjusted p < 0.05 ( Love et al., 2014 ). Pathway term enrichment was performed via Ingenuity Pathway Analysis. mRNA-Seq data have been uploaded to the Genome Expression Omnibus/National Center for Biotechnology Information with accession number GSE217506 . Within four biological pathways which were enriched in HESCs exposed to both BM-MSC-1 and BM-MSC-2 secretomes ( Fig 4E ), only 2 genes ( CCL2 , HGF ) were >2-fold differentially expressed in HESCs that had been cocultured with both sources of BM-MSCs. We used qRT-PCR to confirm the increased expression of CCL2 and HGF in HESCs exposed the BM-MSC secretome. We applied the transwell system described for the proliferation studies for RT-qPCR. HESCs at 2 × 10 5 per well were plated in 6-well dishes (Sigma-Aldrich), with BM-MSCs or UC-MSCs in the inner wells with 0.4 μm pore size inserts to allow for passage of the MSC secretome. After 24 h, total RNA was isolated from HESCs using the RNeasy Plus Mini Kit (Qiagen). Total RNA (1 μg) was reverse transcribed using qScript cDNA Supermix (Quanta BioSciences). RT-qPCR was performed in triplicate using the QuantiNova SYBR Green PCR Kit (Qiagen) and the following primer sequences: 18s forward 5’-CCG GGC TTC TAT TTT GTT GGT-3’, 18s reverse 5’-TAG CGG CGC AAT ACG AAT G-3’, CCL2 forward 5’- GCT CAT AGC AGC CAC CTT CAT TC-3’, CCL2 reverse 5’ GCT CAT AGC AGC CAC CTT CAT TC-3’, HGF-1 forward 5’- GCA ATT AAA ACA TGC GCT GAC A-3’, HGF-1 reverse 5’-TCC CAA CGC TGA CAT GGA AT-3’. The relative expression level of each target gene was quantified using (2 −ΔΔCt ) comparing to 18s rRNA ( Livak and Schmittgen, 2001 ). Expression level was determined as a mean of the triplicates. To assess the effect of rCCL2 on HESC motility, transwell migration and invasion assays were performed using 24-well transwell chambers (Corning) and Falcon transparent cell culture inserts (8 μm pore size, Corning). A concentration of 100 ng/ml recombinant CCL protein (R&D Systems) was selected based on dose-response studies conducted by Li et al. in human endometrial stromal cells ( Li et al., 2012 ). To assess HESC migration after 24h, 5×10⁴ HESCs in phenol red-free DMEM/F12 media supplemented with 0.5% charcoal stripped FBS, 1% ITS+premix and 100 ng/ml recombinant CCL2 protein or 1% bovine serum albumin (BSA) for the control were placed on uncoated inserts in the upper compartment. To assess HESC invasion after 24h, 1×10 5 HESCs in phenol red-free DMEM/F12 media supplemented with 0.5% charcoal stripped FBS, 1% ITS+premix and 100ng/ml recombinant CCL2 protein or 1% BSA for the control were placed on Matrigel coated inserts in the upper compartment. Phenol red-free DMEM/F12 media supplemented with 10% charcoal stripped FBS and 1% ITS+premix was added to the lower compartment of the chamber. To assess HESC migration and invasion after 48 hours ( Li et al., 2012 ), 2×10⁴ HESCs in phenol red-free DMEM/F12 media supplemented with 1% charcoal stripped FBS, 1% ITS+premix and 100ng/ml recombinant CCL2 protein (R&D Systems) or 1% BSA for the control were placed on uncoated or Matrigel coated inserts in the upper compartment. Phenol red-free DMEM/F12 media supplemented with 5% charcoal stripped FBS and 1% ITS+premix was added to the lower compartment of the chamber. After 24 or 48 hours, non-migrated/non-invaded cells in the upper compartments were removed with a cotton swab. Cells on the undersurface of the membrane were fixed with 4% PFA and stained with H&E. Slides were imaged using the Leica DM2500 LED microscope and Leica Application Suite X (LAS X) software at 2.5x magnification. Two independent observers counted the nuclei. The extent of cell migration/invasion was determined by counting nuclei of cells that migrated in a 1250 × 1250 pixel area of the insert using ImageJ software. All experiments, with exception of mRNA-Seq, were performed at least three times with at least three technical replicates. Means were compared using unpaired t-test or ANOVA followed by Tukey’s multiple comparison test. Data are expressed as mean ± SEM. Statistical analyses were performed using Prism Version 8.0 (GraphPad). Statistical significance was defined as P < 0.05.

MSCs are mesoderm-derived, multipotent adult stem cells that can differentiate into a variety of cell types. We characterized BM-MSCs by phenotype using established surface markers ( Lv et al., 2014 ). Flow cytometry indicated dim fluorescence on 80% expressed CD44, CD73, CD29 and CD90 ( Fig. 1A – D ). Next, we performed multilineage differentiation assays and showed adipogenic differentiation ( Fig. 1E – H ) as well as osteogeneic differentiation ( Fig. 1J – M ). We used transwell coculture systems to assess the role of the BM-MSC secretome on HESC motility ( Fig. 2A ). To determine if exposure to the BM-MSC secretome enhances HESC migration toward a chemoattractant, transwell migration of HESCs across an 8 μm pore transwell chamber culture insert was determined. BM-MSCs were cultured in media supplemented with 10% FBS in the lower compartment. HESCs resuspended in media supplemented with 0.5% FBS were placed in the upper compartment of the transwell chamber, and after 24 h, the number of migrated HESCs was counted ( Fig. 2A and B ). Indirect coculture with BM-MSCs significantly increased HESC transwell migration, 3.9-fold for BM-MSC-1, 7.3-fold for BM-MSC-2, 5.8-fold for BM-MSC-3 and 3.4-fold for BM-MSC-4, relative to culture in media without BM-MSCs ( Fig. 2C and Supplementary Fig.1A ). We next asked if the secretome of BM-MSCs could enhance HESC invasion through a Matrigel matrix. As in the transwell migration assays, a 20-fold increase in FBS was used as a chemoattractant. 24h of indirect coculture with BM-MSC-1, BM-MSC-2, and BM-MSC-3 significantly increased HESC transwell invasion, 4.5-fold, 5.1-fold, and 4.2-fold, respectively relative to culture in media without BM-MSCs ( Fig. 2D and Supplementary Fig.1B ). HESC invasion after coculture with BM-MSC-4 was 2.8-fold greater than control, but this increase was not statistically significant ( Supplementary Fig.1B ). Together, these data show that BM-MSCs from all donors can enhance HESC motility. We used a transwell coculture system to determine if exposure of HESC to BM-MSC secretome increased HESC proliferation ( Fig 3A ). HESCs were placed in the lower compartment of the transwell chamber, and BM-MSCs in the inner chambers with 0.4 μm insert. After 24 h, the HESCs were assessed for proliferation based on the percentage of cells expressing Ki67 ( Fig. 3B and C ). The results indicated that BM-MSC-1 secretome showed a larger percentage of HESCs expressing Ki67 as compared to HECSs cultured in control media (53.8 ± 3.3% vs. 42.7 ± 2.8%, P =0.02) ( Fig. 3C ). These findings indicated that exposure to the BM-MSC-1 secretome increased HESC proliferation. In contrast, we noted no significant ( P >0.05) difference in the percentage of HESCs expressing Ki67 after exposure to BM-MSC-2 vs. control media (41.6 ± 3.6% vs. 46.1 ± 3.5%) ( Fig. 3C ). Together, these data show that the variable effect of the BM-MSC secretome on HESC proliferation might depend on heterogeneity in MSC culture. To examine the impact of the BM-MSC secretome on HESC gene expression, whole-genome transcriptome profiling (mRNA-seq) was performed in HESCs cocultured with BM-MSCs in a transwell system. We sought to identify changes in HESC gene expression after 24 h of transwell coculture with BM-MSC-1 (commercially purchased) or BM-MSC-2 (MSCs isolated in our laboratory) relative to HESCs cultured in control media ( Fig 4A ). We chose the time point of 24h based on significant upregulation of HESC motility after 24h co-culture with BM-MSC-1 and BM-MSC-2 as demonstrated by transwell migration and invasion assays ( Fig 2 ). PCA dimensionality reduction of the mRNA-seq data showed that the HESCs clustered into three groups ( Fig. 4B ). mRNA-seq analysis revealed 10,139 differentially expressed genes (DEGs) with FDR < 0.05 in HESC exposed to either BM-MSC-1 or BM-MSC-2, as compared to control HESC ( Fig 4D ). Of these, 4160 DEGs were identified in HESC exposed to BM-MSC-1, 1629 DEGs in HESC exposed to BM-MSC-2, and 4350 DEGs overlapped in HESC exposed to BM-MSC-1 and BM-MSC-2 ( Fig. 4C ). DEGs with Padj 1.5 or FC 1.3. After exposure to the BM-MSC secretomes, 16 pathways were enriched in HESC+BM-MSC-1 and 78 pathways were enriched in HESC+BM-MSC-2 ( Supplementary Tables 1 and 2 ). Four pathways, Coagulation system, NAD Signaling Pathway, Renal Cell Carcinoma Signaling and Tumor Microenvironment Pathway were enriched in HESCs exposed to the secretomes from both BM-MSC-1 and BM-MSC-2 ( Fig. 4E ). Within those 4 pathways, only 4 genes ( CCL2 , HGF , PLAU, and BDKRB2 ) were differentially expressed in HESCs that had been cocultured with both sources of BM-MSCs ( Fig. 4F ). We used qRT-PCR to confirm the increased expression of CCL2 and HGF in HESCs exposed the BM-MSC secretome. qRT-PCR showed significantly increased mRNA expression of CCL2 in both HESC+BM-MSC-1 (5-fold) and HESC+BM-MSC-2 (7.7-fold) compared to control HESC ( Fig. 4G ). In contrast, the increase in HGF expression was significant in HESC+BM-MSC-2 (1.8-fold), but not significantly different in HESC+BM-MSC-1 compared to control HESCs ( Fig. 4H ). Together, these analyses reveal DEGs and pathways that are similarly altered in HESCs exposed to BM-MSCs from different donors, with CCL2 as the gene that is consistently significantly upregulated. UC-MSCs are more readily accessible for therapeutic application and have been tested in the uterus ( Cao et al., 2018 ). We were interested in determining if the secretome of UC-MSCs had the same effects on HESC proliferation, migration, and invasion as the secretome of BM-MSCs. We isolated UC-MSCs from two donors, UC-MSC-1 and UC-MSC-2. First, we used flow cytometry to characterize surface expression of MSC markers. Fewer than 1% of the cells expressed CD45, and greater than 99% of the cells in both UC-MSC-1 and UC-MSC-2 expressed CD44, CD73, CD29 and CD90 ( Fig. 5A and B ). Next, we performed multilineage differentiation assays to demonstrate multipotency of the UC-MSCs. After 21–28 days of exposure of UC-MSCs to adipocyte differentiation media, differentiated cells contained red stained lipid vacuoles ( Fig. 5C and D ). After 30 days of exposure of UC-MSCs to osteoblast differentiation media, osteoblasts were identified by mineral deposition that was visualized by Kossa staining for calcium ( Fig. 5F and G ). We used transwell coculture systems to determine the impact of the UC-MSC secretome on HESC proliferation ( Fig. 3A ). We found that the percentage of HESCs expressing Ki67 was similar after 24h coculture with UC-MSC-1 or culture in DMEM+10% charcoal stripped FBS (48.4 ± 1.8% vs. 43.0 ± 3.2%) ( Fig. 6A ). In contrast, a larger percentage of HESCs expressed Ki67 after coculture with UC-MSC-2 (53.6 ± 1.9% vs. 43.0 ± 3.2%, p =0.01) ( Fig. 6A ). We used the transwell coculture systems with FBS as a chemoattractant to determine if the UC-MSC secretome could enhance growth factor induced HESC migration and invasion ( Fig. 2A ). We found that coculture with UC-MSCs significantly increased HESC transwell migration and invasion through Matrigel matrix, 3.8-fold for UC-MSC-1 and 3.4-fold for UC-MSC-2 for migration ( Fig. 6B ) and 3.0-fold for UC-MSC-1 and 4.3-fold for UC-MSC-2 for invasion ( Fig. 6C ). Thus, like our findings with BM-MSCs, UC-MSC secretome from both donors significantly increased HESC migration and invasion, whereas the secretome from only one donor led to a small, but significant increase in HESC proliferation. Given that BM-MSCs and UC-MSCs had a similar effect on HESC motility we sought to determine if gene expression changes were similar. We performed RT-qPCR to assess whether indirect coculture with UC-MSCs increased expression of CCL2 and HGF. Expression of CCL2 in HESCs was significantly increased (6.5-fold) by UC-MSC-2 but not by UC-MSC-1 ( Fig. 6D ). Expression of HGF in HESCs was significantly increased (1.6-fold) by UC-MSC-2, but not by UC-MSC-1 ( Fig. 6E ). Thus, UC-MSC secretome-induced and BM-MSC secretome-induced changes in CCL2 and HGF expression in HESCs were similar. The robust induction in CCL2 gene expression, consistently by secretome of BM-MSC from both BM donors tested, and by UC-MSC secretome (1 of 2 donors) suggests that induction of CCL2 is a key component conserved in paracrine signaling from BM-MSCs and UC-MSCs. The BM-MSC secretome is a rich source of many factors, including CCL2 ( Baberg et al., 2019 ). As an autocrine factor, CCL2 can regulate the biological behaviors of endometrial stromal cells in an autocrine manner ( Li et al., 2012 ). We hypothesized that CCL2 in the MSC secretome increases secretion of CCL2 by HESCs, resulting in increased HESC motility. To address this hypothesis, we cultured HESCs in media with and without recombinant CCL2 and determined the extent of HESC transwell migration and invasion through Matrigel towards a chemoattractant ( Fig 7 ). We found that exposure to CCL2 for 24 hours, with a 20-fold increase in FBS in the lower compartment as a chemoattractant, did not impact HESC transwell migration or invasion through Matrigel matrix ( Fig. 7A , B ). We then sought to determine if extended culture with CCL2 could impact HESC motility. For the extended 48 culture, we reduced the initial number of HESCs from 5×10 4 to 2×10 4 and used a 5-fold increase in FBS in the lower compartment as a chemoattractant. We found that exposure to CCL2 for 48 hours significantly increased HESC migration (1.2-fold) and invasion through Matrigel (1.4-fold) ( Fig. 7A , B ). Together, our data suggest that increased HESC motility induced by the MSC secretome appears to be mediated in part by CCL2.

The normal human endometrium is a highly dynamic tissue that undergoes massive tissue remodeling on a cyclic basis, characterized by repeated tissue breakdown followed by rapid post-menstrual regeneration. In this sense, the endometrium can be considered a “wound” at menses which undergoes cyclic repair, unique in its ability to do so without formation of a scar ( Henriet et al., 2012 ), but the cellular and molecular mechanisms underlying cyclic “scarless endometrial repair” are not completely known. Cyclic repair of the endometrium requires intact function of its multiple cell types, including parenchymal, endothelial and immune cells, as well as contributions from adult stem cell types, including mesenchymal stem cells which may originate from endometrial and/or extrauterine sources ( Maybin and Critchley, 2015 , Owusu-Akyaw et al., 2019 , Salamonsen et al., 2021 ). In many tissues, mesenchymal stem cells, including those derived from bone marrow, induce rapid repair of damaged tissue ( Gnecchi et al., 2016 , Hu et al., 2018 , Fu et al., 2019 ). Given the central role of endometrial stromal cell motility in the substantial tissue remodeling associated with endometrial regeneration ( Weimar et al., 2013 ), and the well-established role of MSCs in tissue regeneration/repair, we investigated the impact of the MSC secretome on endometrial stromal cell proliferation, motility, and invasive capacity, functions necessary for cyclic endometrial regeneration and repair. MSCs are non-hematopoietic, multipotent adult stem cells which traffic to sites of injury to repair damage by interacting with the niche and releasing secretome such as soluble factors (e.g. chemokines, cytokines, other growth factors) and small microvesicles such as exosomes ( Eleuteri and Fierabracci, 2019 , Gnecchi et al., 2016 ). The MSC secretome mediates tissue repair in a variety of diseased and injured tissues (e.g. cardiac, liver) by mediating angiogenesis and fibrosis pathways ( Ferguson et al., 2018 , Katsuda and Ochiya, 2015 , Zhang et al., 2016 ). We focused on effects of MSC derived from bone marrow (BM-MSC) as well as umbilical cord (UC-MSC) given relative ease in accessibility, distinct secretome profiles, and well-established roles for their respective secretomes in tissue remodeling and repair ( Vizoso et al., 2017 ). We found that exposure of HESCs to the secretome of both MSC types, from multiple donors, consistently and significantly increased HESC migration and invasive capacity. In contrast, effect of the UC- and BM-MSC secretome on increasing HESC proliferation was modest and donor-dependent. The BM-MSC secretome enriched HESC expression of genes involved in multiple cell signaling pathways involved in cell motility, with a robust and significant increase in CCL2 expression in HESCs exposed to MSCs from UC and BM. This finding, coupled with significant upregulation of HESC motility in response to recombinant CCL2, indicates the upregulation of HESC CCL2 expression as one mechanism by which the MSC secretome enhances HESC motility. Our findings have significant implications regarding mechanisms by which MSCs may support endometrial regeneration in both physiologic and pathologic settings. Although highly efficient in its physiologic repair processes after menses and parturition, a significant injury may result in intrauterine adhesions and/or atrophy due to inadequate repair. Severe intrauterine adhesions (Asherman syndrome) and endometrial atrophy contribute to female infertility and are among the most difficult causes of infertility to treat, given a lack of therapeutic options. BM-MSCs traffic to the endometrium ( Gil-Sanchis et al., 2015 ) and have been implicated in endometrial repair in rodent models of Asherman syndrome ( Gao et al., 2019 , Wang et al., 2016 ), but the precise mechanisms remain unknown. Although a number of studies provide evidence supporting the ability of BM-derived stem cells to differentiate into endometrial cell types in both rodent ( Ikoma et al., 2009 , Mints et al., 2008 , Morelli et al., 2013 ) and human models ( Mints et al., 2008 , Taylor, 2004 ), the overall number of endometrial cells of BM origin is very small. On the contrary Ong et al. ( Ong et al., 2018 ) found no evidence of transdifferentiation in a chimeric mouse model; rather, macrophages of BM-origin which weakly expressed CD45 and could be mistaken for BM-derived endometrial cells. The collective data point to a predominant role for the BM-MSC secretome, rather than differentiation capacity of MSCs, in supporting endometrial regeneration and/or repair. In addition to the bone marrow, mesenchymal stem cells can be isolated from multiple other tissues, such as umbilical cord, adipose tissue, placenta, endometrium and menstrual blood, among others ( Shin et al., 2021 , Vizoso et al., 2017 ). Although MSCs from different tissues share similar morphologic characteristics, the secretome profile of each differs significantly ( Shin et al., 2021 ), implicating distinct functional capacities dependent on tissue source. The secretome of MSCs derived from fetal tissues such as UC has been demonstrated to have a more diverse composition that that of MSC derived from adult tissues ( Shin et al., 2021 ). The non-invasive accessibility and distinct secretome profile of UC-MSCs highlights the need for further characterization of functional effects of UC-MSCs and their therapeutic potential in the endometrium. Our studies support a role for the UC-MSC secretome in enhancing endometrial stromal cell motility similar to that of BM-MSC, mediated in part by upregulation of CCL2 expression. That CCL2 is capable of mediating BM-MSC/UC-MSC-induced HESC motility is consistent with established roles for this cytokine in multiple tissues. Also known as monocyte chemoattractant protein (MCP-1), CCL2 is a pro-inflammatory chemokine known to recruit monocytes, memory T-cells, and dendritic cells ( Zhu et al., 2021 ). In cancer models, CCL2 has been implicated in stimulating tumor proliferation and metastasis by binding CCR2 and activating classic signaling cascades including MAPK/ERK signaling ( Wang et al., 2022 ). In the endometrium, CCL2 is secreted by endometrial stromal cells, facilitates macrophage recruitment to endometriotic cells, and upregulates invasion, migration, and survival of endometrial stromal cells through paracrine and autocrine regulation, implicated in the pathogenesis of endometriosis ( Gou et al., 2019 , Li et al., 2012 ). CCL2 regulates the biological behavior (invasiveness and proliferation) of endometrial stromal cells in an autocrine manner by activating Akt and/or MAPK/Erk1/2 signaling pathways to promote PCNA, survivin, and MMP2 expression ( Li et al., 2012 ). Our data demonstrate that exposure to the BM- and UC-MSC secretomes significantly upregulated CCL2 mRNA expression by HESCs and significantly increased T-HESC motility. Recombinant CCL2 similarly upregulated HESC migration and invasion. Since CCL2 has been identified in the MSC secretome ( Whelan et al., 2020 ), the MSC secretome is likely to upregulate HESC motility functions via both paracrine (i.e., MSC-derived CCL2 effects) as well as autocrine mechanisms. Like CCL2, hepatocyte growth factor (HGF) has also been implicated in endometrial regeneration. HGF binds c-MET and induces activation of multiple downstream signaling pathways involved in mitogenic and motogenic functions ( Kwon and Godwin, 2017 ). Previous studies have shown that HGF promotes proliferation, migration, and invasion of endometrial epithelial and stromal cells in vitro ( Sugawara et al., 1997 , Yoshida et al., 2004 ). In the murine uterus, HGF expression is differentially modulated by estradiol and progesterone and exhibits cell type- and estrous cycle-specific changes, supporting a role for this cytokine in cyclic endometrial remodeling via autocrine/paracrine mechanisms ( Zhang, 2010 ). Our RNAseq studies demonstrated significant upregulation of HGF mRNA expression in HESCs exposed to the secretome of BM-MSC. However, qRT-PCR studies indicated only a modest but significant increase in HGF mRNA expression when HESC were exposed to the secretome of MSC isolated from a single BM donor, and no difference when exposed to secretome of commercially obtained BM-MSC. Similarly, exposure to the secretome of MSC isolated from one of two UC donors resulted in a modest but significant increase in HGF expression. These variable results, which may be attributable to inherent differences between human donors, support the need for more definitive characterization of a role for this cytokine in MSC-mediated HESC motility functions.

Overall, the data presented herein demonstrate that the secretome of MSC isolated from human bone marrow and umbilical cord significantly increases HESC migratory and invasive capacities; cellular functions critical for cyclic regeneration and repair of this highly dynamic tissue. Functional effects of the MSC secretome on HESC motility appear to be mediated, in part, by upregulation of HESC CCL2 expression. Although emerging data increasingly support a potential role for mesenchymal stem cells in the treatment of defects in endometrial regeneration and function, the ability to safely and efficaciously administer these cells for therapeutic purposes remains to be established. Our data provide important implications for use of the MSC secretome as an attractive cell-free therapeutic strategy to enhance endometrial regeneration and repair after injury.

Cyclic regeneration and renewal of the uterine endometrium is required for human reproductive function. In normal menstrual cycles, demise of the ovarian corpus luteum and the subsequent decrease in circulating progesterone triggers an inflammatory response that results in shedding of the functionalis layer of the endometrium. Regeneration of the functionalis layer after menstruation, and parturition, involves inflammation, angiogenesis, re-epithelization, and tissue remodeling, which are processes integral to wound repair in non-reproductive organs ( Stone et al., 2016 ). Cyclic repair of the endometrium requires intact function of multiple cell types, including parenchymal, endothelial and immune cells, and contributions from adult stem cells which may originate from endometrial and/or extrauterine sources ( Maybin and Critchley, 2015 , Owusu-Akyaw et al., 2019 , Salamonsen et al., 2021 ). Mesenchymal stem cells (MSCs) are multipotent adult cells showing multilineage differentiation into cells such as adipocytes, chondrocytes, and osteoblasts. MSCs have been demonstrated to regenerate tissue and to migrate towards injured sites where they can secrete bioactive products to aid in tissue regeneration and repair ( Eleuteri and Fierabracci, 2019 ). These secreted factors include soluble factors such as cytokines and insoluble microvesicles ( Vizoso et al., 2017 ). The secretome regulates cellular processes critical for tissue regeneration by modulating biological processes such as immunomodulation, angiogenesis, cell migration and proliferation ( Liang et al., 2014 ). MSCs can be harvested from several anatomical sites including bone marrow (BM), adipose tissue, placenta, and umbilical cord. BM-derived MSCs (BM-MSCs) are the most frequently used stem cells in tissue regeneration and may be considered the “gold standard”, however, MSCs derived from other sources also possess the ability to home to injured sites and facilitate tissue repair ( Wu et al., 2007 , Wu et al., 2014 , Huang et al., 2020 , Fu et al., 2019 ). Asherman syndrome, characterized by intrauterine adhesions and inadequate endometrial regeneration, is a significant cause of female infertility with limited therapeutic options. Studies with rodent models and small clinical trials have shown that non-hematopoietic MSCs from different tissue sources can contribute to endometrial repair after injury ( Lv et al., 2021 ). In rat models of endometrial injury, BM-MSCs and umbilical cord (UC) MSCs aid endometrial repair through paracrine actions on the endometrium as well as differentiation into endometrial cell types ( Jing et al., 2014 , Zhang et al., 2018 , Zhang et al., 2022 , Zheng et al., 2020 ). The secretome and biological characteristics of MSCs vary based on tissue source ( Kozlowska et al., 2019 , Shin et al., 2021 ); however, both BM-MSCs and UC-MSCs promote endometrial regeneration in rodent models ( Lv et al., 2021 ). In a small trial of 26 women with intrauterine adhesions, transplantation of UC-MSCs into the uterus after adhesiolysis increased endometrial thickness and expression of factors associated with endometrial proliferation ( Cao et al., 2018 ). Together, these studies suggest that MSCs can be leveraged as potential primary or adjunct therapeutic approaches to enhance or restore endometrial regeneration ( Strug and Aghajanova, 2021 , Rungsiwiwut et al., 2021 ). However, the efficacy and safety of such cell-based therapies remains unproven, rendering use of the MSC secretome or its components as an attractive potential alternative for cell-free therapy. Whereas prior studies to determine the effects of MSCs on endometrial regeneration and repair have focused on endometrial cell proliferation, we tested the hypothesis that the MSC secretome enhances endometrial stromal cell motility and regulates expression of genes that promote cell motility. We used a coculture model with human MSCs from BM and UC. This model allowed for the secretome to mediate an interaction with human endometrial stromal cells (HESCs). Our first objective was to determine how exposure to BM-MSCs impacts proliferation, migration, and invasion of HESCs. To identify underlying molecular mechanisms associated with the observed changes in HESCs, we performed RNA sequencing (RNA-Seq) in HESCs exposed to BM-MSCs. Given the noninvasive, easy accessibility, and increasing utilization of human UC-MSCs in regenerative medicine, our second objective was to determine if UC-MSCs and BM-MSCs have a similar effect on HESCs. We found that exposure to the secretome of both UC- and BM-MSCs significantly increases HESC migration and invasion, with a modest effect on HESC proliferation. Increased HESC motility induced by the MSC secretome appears to be mediated by paracrine and autocrine mechanisms, in part by upregulation of HESC CCL2 expression. The potential to leverage the MSC secretome as a novel cell-free therapy in the treatment of disorders of endometrial regeneration is discussed.

Supplementary Figure 1. The secretomes of BM-MSC-3 and BM-MSC-4 increase HESC transwell migration and invasion. HESC transwell migration and invasion through Matrigel were assessed after coculture with BM-MSCs for 24h. ( A ) Transwell migration was significantly increased by coculture with BM-MSC-3 (**** P < 0.0001) and BM-MSC-4 (*** P =0.0005) compared to control HESC. ( B ) Transwell invasion through Matrigel was significantly increased by coculture with BM-MSC-3 (**** P < 0.0001), but not with BM-MSC-4 ( P =0.06) compared to control HESC. Data are expressed as means (of N=3) ± SEM. Statistical analysis was performed with an unpaired t test followed by Welch’s correction.