Restoration of spermatogenesis in azoospermic mice by bone marrow mesenchymal stromal/stem cells conditioned medium

One of the main cause of male infertility is non-obstructive azoospermia, which is not manageable medically. The rst aim of the current research was to show the effect of extracellular vesicle-contained conditioned media (CM) instead of mesenchymal stromal/stem cells (MSCs) for treatment of nonobstructive azoospermia. In the next step, we aimed to study the differentiation potential of MSCs into spermatocytes after injection of MSCs in mice seminiferous tubules. This study has provided an applied treatment for busulfan-induced azoospermia using adipose tissue-derived (AT-MSCs) and bone marrow-derived MSCs (BM-MSCs) and bone marrow CM (BMCM) in animal models. differentiation: C) BM-MSCs, D) AT-MSCs; Osteogenic differentiation: E) BM-MSCs, F) AT-MSCs; chondrogenic differentiation: G) BM-MSCs, H) AT-MSCs. Phenotype by ow cytometry analysis: I) Expression of BM-MSCs surface markers (CD44 and CD105) (>95%) in contrast to negative expression hematopoietic markers (CD34, CD45) (<10%) J) Expression of AT-MSCs surface markers (CD44 and CD105) (>95%) in contrast to negative expression hematopoietic markers (CD34, CD45) (<10%). Conrmation of isolation of EVs with scanning electron microscopy (SEM), transmission electron microscopy


Results
Histomorphometry and ow cytometry evaluation of testes showed normal morphology of most of the seminiferous tubules of therapy groups as well as successful recovery of spermatogenesis, but spermatogenesis was not observed in the azoospermia group. It is worth notable that the results of the BM-MSC therapy group were more favorable than other therapy groups.

Conclusions
AT-MSC, BM-MSC and BMCM can be strongly suggested as candidates in the therapy of azoospermia.

Background
Infertility is one of the highest serious global severe disabilities that in uenced the quality of life of many couples worldwide. In humans, male infertility accounts for 8-12% of reproductive-aged couples (1), which can be caused by congenital or acquired urogenital abnormalities, malignancies, urogenital tract infections, increased scrotal temperature (e.g. as a consequence of varicocele), endocrine disturbances, genetic abnormalities, immunological factors, and etc. (2). Azoospermia with two types of obstructive or non-obstructive azoospermia affects about 1.3% of all men (2). This static refers to a notable population in the eld of male infertility and shows the importance of efforts, which are being made today to treat non-obstructive azoospermia (3).
Although different therapies (for example hormone therapy and testicular sperm extraction) are recommended depending on the type and cause of azoospermia (4,5), about 67% of azoospermia is related to untreatable testicular disorders, which result in spermatogenic failure (6). Accordingly, researchers have recently tried to develop stem cell transplantation-based therapies for azoospermia (7).
These attempts arise from the view that on the one hand, spermatogenic epithelium contains germ cells (spermatogonia) with the ability to repair damages on their own, and on the other hand, if stem cells are placed in testis with defected spermatogonia, they will restore the function of these spermatogonia and induce differentiation/proliferation of them or differentiate to spermatogonia (7).
Various sources of stem cells are used, including embryonic stem cells or pluripotent stem cells, but the use of mesenchymal stromal/stem cells (MSCs) is of great interest as they are free from ethical problems of embryonic stem cells and can be obtained from various sources such as bone marrow, adipose tissue, dental pulp, and menstrual blood, all of which have a high ability to differentiate into different tissues (8).
Moreover, MSCs have been shown to secrete cytokines and growth factors that have anti-in ammatory, immune-modulating, anti-apoptotic, and proliferative effects which all are packed in extracellular vesicles (EVs) (9). Exosomes, a nano-scale EV, secreted from cells endosomally that mediate the fusion of endocytic, multicellular vesicles with the plasma membrane of cells (10). They are generally thought to be smaller than other EVs about 30 to 150 nm in diameter and can be said to be about the size of many lipoproteins but much smaller than cells (11). Exosomes could interact with macromolecules and that they can serve as distributors for proteins, lipids, mRNA, miRNA, and DNA (12,13). They also are capable of being drug carriers since they are composed of cell membranes. Moreover, many studies showed that the exosome could be used as the drug itself due to its anti-in ammatory and immune-modulatory effects (14). In addition, some exosomes from mesenchymal sources showed to have the proliferative effects in vitro and in vivo (15). Conditioned media (CM) of stem cell culture contains the EVs which are secreted from the cells (16). CM can induce tissue regeneration (17,18) or cell differentiation (19,20). In vitro effects of CM on germ cell regeneration has been shown (20). However, to the best of our knowledge CM has not been used for the treatment of azoospermia, yet.
On the other hand, bone marrow-derived MSCs (BM-MSCs) and adipose tissue-derived MSCs (AT-MSCs) have recently been considered by many researchers. BM-MSCs and AT-MSCs are used to produce male germ cells in vitro (21,22) and in vivo (23,24). Less invasive, less costly, powerful immunomodulatory effects, more proliferative potential, and more secretion of growth factors and cytokines are the most important priorities of AT-MSCs compared to BM-MSCs for cell therapy (25). As far as we know, there is no comparison between these two sources for cell therapy of azoospermia.
Despite the effectiveness of MSC therapy in the treatment of azoospermia (26), it has yet remained unknown whether transplanted stem cells differentiate to spermatocytes. Based on the existing data, this study is the rst attempt to compare the effect of allotransplanting of AT-MSC and BM-MSC and BMCM in seminiferous tubules of busulfan induced azoospermia in Balb/C mice. Therefore, the rst goal of the present study was to show the effect of EV-contained CM in comparison with MSCs for the treatment of non-obstructive azoospermia. The next aim of this study was to show the differentiation potential of inter-tubular injected MSCs to spermatocytes in mice seminiferous tubules.

Animals and ethics
Thirty male adult Balb/C mice (30±5 g) were provided from Research and Clinical Center for Infertility, Shahid Sadoughi University of Medical Sciences, and two female eGFP +/+ Balb/C mice were gifted from Royan Institute and were housed in the laboratory under the controlled condition of a temperature of 22±2°C, the humidity of 55±5% and lightening cycle of 12 h light/dark. They were fed ad libitum standard diet. All the manipulations were performed in accordance with the regulations of working with laboratory animals and approved by the Ethical Committee of Astana Medical University, Astana, Kazakhstan (Protocol No. 10).
Isolation and culture of BM-MSC and AT-MSC BM-MSCs were isolated from femurs of two eGFP +/+ Balb/C mice by modi cation of the previously reported method (24). In sterile condition and after soaking the separated thigh of mice in 70% ethanol, the skin has been removed. After cutting the two ends of the thigh bone, the thigh with the muscles was put in a blue-color sampler tip which its upper half parts had been cut. The tips were placed in 1.5 mL micro-tubes. Then all micro-tubes were centrifuged at 1500 ×g for 15 min at 37°C. The collected BM at the end of micro-tubes were mixed with 1 mL of the Dulbecco's modi ed eagle medium (DMEM) without fetal bovine serum (FBS, Gibco, U.S.A.) and penicillin-streptomycin antibiotics (Gibco, U.S.A.). Then, the BM suspension was transferred to a 25 cm 2 ask containing DMEM medium along with 10% FBS and 1% penicillin-streptomycin. The ask was incubated at 37°C, standard humidity, and 5% CO 2 concentration.
The culture medium was changed 24 h after the beginning of incubation, then every 72 h under sterile conditions, to remove unattached cells and debris. After increasing the density of the cells adhering to the ask to 80%, the cells were passaged using trypsin enzyme. For this purpose, the trypsin enzyme was added to the ask for 3 min to separate the cells sticking to the ask oor. Then, 2 mL of 10% FBS culture medium was poured in to the ask for neutralizing the trypsin enzyme effect. Collecting cells separated from the medium in the rst passage was performed and continued till the third passage.
AT-MSCs were provided using ovarian AT of the same mice used in the previous step. Brie y, AT was minced into tiny parts. The AT parts were explanted in a T75 ask. After 15 min of pasting the explants by drying the attachment surface of the explant and ask, they were gently covered with a drop of FBS.
The explanted AT were incubated for 48 h in an incubator containing 5% CO 2 at 37°C and saturated humidity. Then DMEM with 10% FBS and 1% penicillin and 1% streptomycin were gently added. By daily monitoring of the border of attachment surface of AT explants and ask, 5 days after explanting and observation of AT-MSCs, the explants were removed by tapping the bottom of ask and media replacement. Then, DMEM containing 10% FBS, 2 mM of L-glutamine (Invitrogen, Netherlands), 1% penicillin and 1% streptomycin were added and kept in a CO 2 incubator. The sub-culturing of cells was continued till the third passage.

Flow cytometry characterization of MSCs
To con rm the isolation of BM-MSCs and AT-MSCs and the non-appearance of hematopoietic stem cells, ow cytometry analysis was used. In this regard, the non-appearance of hematopoietic stem cell-speci c surface marker expression (CD34 and CD45) and the existence of speci c surface area marker expression on BM-MSCs and AT-MSCs (CD44 and CD105) in the isolated cells based on the method previously described were considered (27). Brie y, 5×10 5 cells of the third passage were incubated with speci c individual monoclonal antibodies, conjugated with uorescein isothiocyanate (FITC) or phycoerythrin (PE) in 250 mL phosphate-buffered saline (PBS) for half an hour at ambient temperature in the dark. CD34, CD45, CD44, and CD105 primary antibodies were used and cells were diluted using 4 mL PBS. The diluted cells were centrifuged, and re-suspended in 600 mL solution containing PBSformaldehyde 2%. Then, they were studied using a four-color FACS Calibur ow cytometer (BD bioscience), and obtained data were analyzed by FlowJo software (FlowJo LLC, Ashland, OR, USA). Immunoglobulin G (IgG)1 FITC and IgG1 PE monoclonal antibodies were applied as isotype controls.

Osteogenic, chondrogenic, and adipogenic characterization of MSCs
In order to determine the potential of osteogenic, chondrogenic, and adipogenic differentiation of AT-MSC and BM-MSC, the cells of third passage were used. MSCs were in 6-well plates. After fusion, 70% of the cells were grown in an osteogenic, chondrogenic, or adipogenic environment for 3 weeks. For osteogenic differentiation, the MSCs at 90% con uence were exposed to an osteogenic differentiation kit (Stem Cell Technology, Iran). Changing the medium was performed two times per week for 3 weeks. After 21 days, osteogenic differentiation was con rmed by Alizarin red staining (Stem Cell Technology, Iran). For chondrogenic induction, MSCs at the con uence of 90% were exposed to a chondrogenic differentiation kit (Stem Cell Technology, Iran). Induction continued 3 weeks and chondrogenic induction was con rmed by Alcian blue staining (Stem Cell Technology, Iran). For adipogenic induction, MSCs at the con uence of 90% were exposed to adipogenic differentiation kit (Stem Cell Technology, Iran). Induction continued 3 weeks and induction of adipogenic was con rmed by Oil Red O staining (Stem Cell Technology, Iran). MSCs without differentiation media were stained simultaneously with three protocols as control.

Preparation of BMCM
In order to obtain the BMCM, BM-MSCs at the third passage were cultured at a density of 10 6 cells in a T75 ask. The BM-MSCs at 80% con uence were rinsed three times with PBS and 10 mL of FBS-free DMEM media were replaced. Collection of the media was performed following 48 h incubation and they were ltrated using a 0.2-µm lter to eliminate cellular debris. The BMCM were stored at -80°C until injection or isolation and con rmation of EVs. EVs were evaluated by scanning electron microscopy (SEM) or transmission electron microscopy (TEM).

Isolation of EVs of BMCM by kit
The CM was collected and centrifuged in 2000 ×g and 4°C for 30 min. The supernatant was harvested and then added to precipitant reagent of Exovista kit (PerciaVista Co., Iran) with 1:1 ratio. The mixture was kept at 4°C for 14 h. To continue, the mixture was centrifuged in 10000 ×g and 4°C for 1 h. The EVs' pellet was observed at the bottom of the falcon and collected for further analysis.
SEM imaging of EVs of BMCM SEM imaging was performed to show EVs in the isolated BMCM (28). The isolated EVs were xed with 2% formaldehyde in PBS. Then, 10 L drop of the sample was placed on poly-l-lysine coated slides and was allowed to air dry for 30-45 min. The sample was washed with phosphate buffered saline (PBS). Then, the sample post-xed with 2.5% glutaraldehyde for 30 min. After that, the sample was washed another time for 10 min. Then the sample was stained with 1% osmium tetroxide for 20 min. The sample was washed again, immediately. At the next step, the sample was dehydrated with accelerated percentage of alcohol (25, 50, 70, 90, 100, 100, and 100% alcohol). Then the sample was placed into desiccator for drying. The sample was coated with layer of gold. The grid was examined on a FESEM TESCAN MIRA3 machine (TESCAN Co., Czech Republic) operating at resolution of 1.2 nm at voltage of 30 kV.
TEM imaging of EVs of BMCM TEM imaging was performed to con rm the presence of EVs in BMCM. The procedure was done based on the previous study (29). In details, 6 L drop of the xed sample was placed on a formvar Carbon lm coated on 200 mesh copper grid (EMS) for 10 min. Excess liquid was absorbed with lter paper. Then, the grid was washed quickly with 100 L MilliQ water and then blotted to remove excess liquid. At the nal stage, the grid was placed on 30 L of 1.5% uranyl acetate (w/v) for 12 s. The grid was examined by PHILIPS CM120 transmission electron microscope (Netherlands) operating at an accelerating voltage of 100 kV.

Induction of azoospermia
In order to model the experimental groups, the male adult Balb/C mice were randomly divided into six equal groups ( Figure 1A).
An intact control group (n = 6): the mice did not receive busulfan and treatment.
Azoospermia group (n = 6): the left testicles of mice received busulfan injections without treatment and were sampled 60 days following the last busulfan injection.
BMCM therapy group (n = 6): the right testicles of mice received busulfan injections and then treated with BMCM and were sampled 60 days following the last busulfan injection.
BM-MSC therapy group (n = 6): the right testicles of mice received busulfan injections and then treated with BM -MSC and were sampled 60 days following the last busulfan injection.
AT-MSC therapy group (n = 6): the right testicles of mice received busulfan injections and then treated with AT-MSC and were sampled 60 days following the last busulfan injection.
Spontaneous healing group (n = 6): the mice received busulfan injections and were left without any treatment and were sampled 120 days following the last busulfan injection.
Then, azoospermia was induced in mice of all groups except those of the intact control group. To induce azoospermia in mice, busulfan was used according to previous reports (24). Brie y, the azoospermia groups were received two doses of busulfan (10 mg/kg, Busilvex®; Pierre Fabre Medicament Boulogne, France) with 21 days' interval. Thirty-ve days following the second injection, their testes were removed and xed in 10% formalin buffer solution for histomorphometry studies. The epididymis was collected from all groups and after incubating in PBS, spermatozoa were evaluated by ow cytometry. The testes of control group mice, which were not received busulfan, were assessed through similar procedures.

BM-MSCs and AT-MSCs transplantation
Preparing an injecting set for a ne cell injection has been previously described (30). The mice in cell and CM therapies' groups (n=18, AT-MSC therapy, BM-MSC therapy, and BMCM therapy groups) were surgically allotransplanted (31). Cell suspensions or CM were mixed with vital stain, sterile trypan blue. The mice were anesthetized with xylazine and ketamine (Alfasan, Netherland) 35 days following the second injection of busulfan. After preparation of abdominal area in dorsal recumbence position, a 1 cm midline abdominal incision was made to reach the peritoneum and testicle. The fat pad and then right testicle were carefully pulled out using iris forceps under a stereomicroscope (Model SZN, Optika, Italy).
Simultaneously, a thin sterile plastic black card with a 30° V-neck as a holder was placed under the testicle. A polyethylene tube attached to a one-milliliter syringe, was lled with the cell suspension (10 6 cells in 100 µL). The tip of the pipette was carefully inserted into the efferent duct, and slowly threaded a few millimeters toward the testes ( Figure 1B). The blue suspension was injected as slowly as possible to avoid moving the pipette and the entrance of the blue suspension into the seminiferous tubules was observed ( Figure 1C). Finally, the testes were returned back into the abdomen and the abdominal wall was closed. The untreated testicles on the left side were considered as azoospermia groups.

Histological and histomorphometric assessments
Sixty days after treatment, the animals' testes were removed after euthanizing with ether. Next, they were put in para n for histological and histomorphometric assessments after xing in 10% formalin buffer solution and dehydrating using alcohol. For each testis, ve horizontal cross-sections with a thickness of 5 µm were made from appropriate regions and stained with hematoxylin-eosin for histological assessment. Then, they were studied carefully by means of a light microscope (Model CX21, Olympus, Tokyo, Japan) to evaluate the presence of spermatogonia, spermatocytes, and spermatids in all tubules. Five circular cross-sections were provided from different areas of the tubules and the inner, outer, and total diameters of all the tubules were determined using the Dinocapture software (version 2.0, Dino-Eye, San-Chung, Taiwan). Using the mean of two diameters (D1, D2) at right angles, the average diameter of the seminiferous tubules (D) was speci ed. Using diameter data, cellular (germinal epithelium), luminal, and cross-sectional areas were determined. In the seminiferous tubules, the cross-sectional area (A) was obtained by the equation A = πD2/4, where D is the average diameter of the tubules, and π is 3.142. The number of seminiferous tubule pro les per unit area was also considered. The area of cells was computed through subtraction of the luminal area from the cross-sectional area. In the seminiferous tubules, determination of the numerical density (Nv) was performed using the following equation: where N A indicates the number of pro les per unit area, D displays the average diameter of the seminiferous tubules, T shows the average section thickness, and indicates the number of tubules per unit volume.
The spermatogenesis index including the presence of spermatogenic cells throughout the testicular tissue, affecting the number of cell layers, cell types, and the existence of spermatids in the tubules was studied and the spermatogenic potential of testes was rated based on the modi ed scale of 0 to 7 (31) as following: 0, no spermatogonia; 1, the existence of only spermatogonia; 2, the presence of spermatocytes; 3, the presence of spermatids up to 25 in each tubule; 4, the presence of 25-50 spermatids per tubule; 5, the appearance of 50-75 spermatids per tubule; 6, the appearance of 75-100 spermatids per tubule; and 7, more than100 spermatids per tubule.
Flow cytometry of enhanced green uorescent protein expression (eGFP) Flow cytometry was conducted to prove the presence of eGFP spermatozoa. For this goal, epididymis was collected from all groups and was incubated in 1 mL PBS at 37°C for 15 min. Then, they were resuspended in a total volume of 1.8 mL, after adding 800 µL of 10% formalin buffer solution, and then stored on ice until analysis with a ow cytometry under illumination in the range of the 360-400 nm that is related to eGFP uorescence (32). A four-color FACS Calibur ow cytometer (BD bioscience) was used to collect the data and data were analyzed using the CellQuest Pro software package (BD bioscience).

Imaging of eGFP spermatozoa
The imaging of eGFP spermatozoa from the epididymis of treated mice was conducted using an epiuorescent microscope (XDS 3FL4, Optika, Italy) to prove the green uorescent nature of spermatozoa produced in cell therapy groups. For this purpose, a red lter (594 nm) and a green lter (498 nm) were used.

Statistical analysis
Kolmogorov-Smirnov test was conducted to study the normal distribution of data of histomorphometry indices of seminiferous tubules. All data were presented as means and standard error (mean ± SE) and analyzed using one-way ANOVA and Tukey post-hoc test (SPSS for Windows, version 20, SPSS Inc, Chicago, Illinois, USA). The Mann-Whitney U test was used to assess the spermatogenesis index. A Pvalue of less than 0.05 was considered statistically signi cant.

Isolated BM-MSCs and AT-MSCs show stem cells characteristics
Adherence to plastic and the existence of a homogeneous broblast-like population were observed for cultured BM-MSCs and AT-MSCs (Figures 2A and 2B). Additionally, studying the adipogenic, osteogenic, and chondrogenic differentiation abilities of BM-MSCs and AT-MSCs showed the successful differentiation of these cells to adipocytes with intracellular lipid droplets, chondrocytes, and osteoblasts, respectively ( Figures 2C to 2H). The diagrams obtained using FlowJo software for characterizing MSCs cell surface markers showed the presence of CD44 and CD105 and the absence of CD34 and CD45 in BM-MSCs and AT-MSCs ( Figures 2I to 2L).

Isolated CM from BM-MSCs contained EVs
In order to evaluate the quality of CM before injection, EVs visualization with SEM and TEM was performed. We isolated EVs with a commercial kit protocol that allows collecting EVs with a homogeneous diameter size. Furthermore, this protocol removed crystals, protein aggregates, and other residues obtained from CM. The exosomes were detected by SEM and TEM analysis ( Figure 2K). The size of exosomes was 20 to 140 nm in diameter (Figure 2k). In addition, the mean diameter of the exosomes was 74.8 nm. TEM images showed electron-lucent EVs with clear lipid bilayer.

Histologic ndings showed therapeutic effects of BMCM, BM-MSCs, and AT-MSCs
Normal intact mice had condensed germinal epithelium in their seminiferous tubules, and all of the epididymis tubes were lled by spermatozoa ( Figure 3A). However, the testes seminiferous tubules of the azoospermia group were empty and their spermatogenesis procedure was completely disrupted ( Figure  3B). Seminiferous tubules revealed degenerative variations including germinal epithelia degenerations and seminiferous tubular atrophy in all tubules and no spermatozoa was found in their epididymis. Moreover, thickening of the walls of some epididymal tubes was observed in this group. On the other hand, in the BM-MSCs, AT-MSCs, BMCM, and even the spontaneous healing groups, Sertoli cells, formation of cells of lower layers of germinal epithelium spermatozoa and spermatogonia were observed in some seminiferous tubules ( Figures 3C to 3F, respectively). Most of the tubules of these groups were nearly lled up with germinal cells including spermatogonia, primary spermatocytes, spermatids, and sperms. Most of the epididymis tubes of this group had spermatozoa, but less than the normal group (data not shown). It is evident that between the MSCs therapy groups, the results relating to BM-MSCs therapy were closer to normal intact mice compared to the AT-MSCs therapy group.

Histomorphometry ndings indicated the therapeutic effects of BMCM, BM-MSCs and AT-MSCs
Quantitative histomorphometry as well as qualitative histological analysis must be performed besides the qualitative histological analysis, which can be affected by operator-related bias and does not provide statistically assessable numerical data. Figure 4 indicates the histomorphometric indices of the seminiferous tubules in all studied groups. The cellular area chart shows that intact control, BMCM therapy, AT-MSC therapy, and spontaneous healing group had no signi cant difference ( Figure 4A). Moreover, the maximum and minimum values of this index were related to BM-MSCs therapy and azoospermia groups, respectively. The cross-sectional area chart shows that the maximum and minimum total area of seminiferous tubules were related to the BM-MSC therapy and azoospermia groups ( Figure   4B). Control, BMCM therapy, and spontaneous healing groups showed a statistically similar total area. This index had lower values in the ATMSC therapy group than the control group. The graph related to the number of tubules indicates the maximum amount of this index for the azoospermia group ( Figure 4C). This index was statistically the same in other groups except for the spontaneous healing group, which showed a slightly higher number of tubule than other groups. The spermatogenesis index of seminiferous tubules of the azoospermia group was minimum ( Figure 4D). The control and BMCM therapy groups had no signi cant difference in terms of this index. ATMSC therapy, BM-MSC therapy, and spontaneous healing groups showed a slightly higher value of this index compared to the control group.

Flow cytometry of eGFP sperms con rmed in vivo differentiation of BM-MSCs and AT-MSCs to spermatozoa
The results of ow cytometry have been shown in Figure 5. According to this gure, 4.07%, 61.0%, 61.5%, 20.9%, 67.4% and 1.60% eGFP spermatozoa found in intact control, GFP control, mixed of intact and the eGFP controls, the AT-MSC therapy, the BM-MSC therapy, and the spontaneously healed azoospermia groups, respectively ( Figure 5A). The uorescent-protein-based imaging of spermatozoa produced in cell therapy groups imaging proved the green uorescent nature of spermatozoa of MSCs therapy groups ( Figures 5B to 5E). Spermatozoa were clearly observed in the image obtained by using a green lter (498 nm), while the image taken using a red lter (594 nm) were not clear. In the ow cytometry analysis, it was expected to observe 0% eGFP subset in the control group, while this value was 4.07% that can be considered as noise. In contrast, in the eGFP control group, it was expected to observe 100% GFP subset, but this value was 61.0%. This difference can be assigned to those sperms whose direction was not appropriate.

Discussion
The CM and MSC therapies as promising therapy for non-obstructive azoospermia were investigated in mice. Our study for the rst time indicated the effects of BMCM for the healing of chemotherapy disrupted the germinal layer in mice. Consistent with our ndings, applications of Sertoli cell-CM on the azoospermia model induced spermatogenesis in mice (33). In addition, 10% of the CM from canine amniotic membrane-derived MSCs enhanced the quality of frozen-thawed sperm cells (34). Since EVs of MSCs contains cell paracrine factors (35), the application of BMCM for azoospermia clinical studies seems to be promising. MSCs-derived CM contains anti-apoptotic factors and growth factors, and it has been investigated in cell-free therapies to simulate the paracrine effects of stem cells on the ovary after chemotherapy (36). Several different cytokines and growth factors with therapeutic properties are secreting from stem cells into their CM, which may play key roles in the regeneration and angiogenesis of damaged tissues (37). Epidermal growth factor (EGF), broblast growth factor 2/basic broblast growth factor (FGF-2/bFGF), hepatocyte growth factor (HGF), insulin-like growth factor I (IGF-I), insulin-like growth factor II (IGF-II), platelet-derived growth factor (PDGF) and vascular endothelial-derived growth factor (VEGF) are the important growth factors in the CM of MSCs (38). Furthermore, the presence of antioxidants in the CM (34), in combination with cytokines and growth factors may have an effect on spermatogenesis of azoospermia mice.
Moreover, the ndings of our study showed the potential of BM-MSCs and AT-MSCs differentiation to spermatozoa. The ow cytometry graphs related to AT-MSCs therapy and BM-MSCs therapy showed the presence of eGFP sperms indicating the differentiation of transplanted MSCs to sperm. However, the higher percentage of eGFP subset in BM-MSCs therapy group (67.4%) than AT-MSCs therapy group (20.9%) revolves higher potential of BM-MSCs to differentiate to sperm compared to AT-MSCs.
Additionally, by comparing the graph of the BM-MSCs therapy group with the eGFP control group, it can be concluded that nearly all sperms in the BM-MSCs therapy are GFP+ that also con rms a higher ability of BM-MSCs to differentiate to spermatozoa than AT-MSCs. Germ cell line formation from pluripotent teratocarcinoma cells and successful in vitro generation of offspring mice from embryonic stem cellderived germ cells has been reported for the rst time (39,40). The successful differentiation of mice BM-MSCs into male germ cells in vitro was also reported (41). The trans-differentiation of BM-MSCs into germ cells, Leydig cells, and Sertoli cells, was shown in busulfan-treated infertile mice (42,43). Human fetal lung-and umbilical cord-derived MSCs were also shown to transdifferentiate into sperm-like cells (44,45). Coherently, recovery of fertility in azoospermia rats following injection of AT-MSCs and their trans-differentiation to spermatozoa have been shown (46). Moreover, the number of tubules except for the azoospermia group with remarkable number of tubules than control and other groups was comparable to the control. The total area and cellular area of all groups except azoospermia, which was remarkably lower than the control group, were comparable with the control group, con rming the successfulness of studied therapy in treating azoospermia mice. Moreover, it is notable that higher values of these indexes for the BM-MSCs therapy group in comparison to others demonstrate more signi cant effectiveness of this therapy method over other studies in the current work. However, induction of spermatogenesis with AT-MSCs and BM-MSCs has been shown in various animal models (23,24,31,(46)(47)(48)(49), although these two sources were compared by our ndings for the rst time and the priority of BM-MSCs for this purpose was shown.

Competing interests
Author Amin Tamadon is employed by PerciaVista R&D Co. The remaining authors declare that the research was conducted in the absence of any commercial or nancial relationships that could be construed as a potential con ict of interest.    area of the tubule, C) number of tubules per 5×5 mm 2 , D) spermatogenesis index of seminiferous tubules.
Lines above the columns show signi cant differences between groups (P<0.05).