Human Endometrial Reconstitution From Somatic Stem Cells: The Importance of Niche-Like Cells

Introduction

The presence of somatic stem cells (SSCs) is key to understanding the regenerative capacity that is retained in many adult tissues. ^{1 –6} Human endometrium is a unique tissue in which 1 layer, the functionalis, undergoes many cycles of growth, shedding, and regression during a women’s reproductive life, while another, the basalis, remains intact. ^7,8 Single dispersed human endometrial cells possess the ability to reconstitute endometrial tissue in an animal model, demonstrating the stemness ability of the whole endometrial fraction. ⁹ This extraordinary regenerative capability is attributed to the presence of endometrial SSCs, which preserve the endometrial lineage, give rise to differentiated cells, and interact with the endometrial environment throughout quiescence, proliferation, and regeneration. ¹⁰

Endometrial SSCs, first suggested in 2004, ^11,12 are probably located in the basalis layer of the endometrium. ^{13 –17} However, their characterization, identification, and isolation remain difficult due to the lack of validated, specific markers. Indeed, there are discrepancies in the identification of endometrial SSCs based on the expression of different markers. Nevertheless, markers defining the properties of bona fide stem cells could only be identified by in vivo assays that functionally define stem cells.

Initial studies indicated that human endometrial cells positive for both CD146 (a member of the immunoglobulin superfamily) and platelet-derived growth factor receptor β (PDGF-Rβ,; also called CD140b) show more clonogenicity than double-negative cells. ¹⁸ While both markers are expressed in basalis and functionalis, CD146 is located in endothelial cells around blood vessels and PDGF-Rβ is only present in stromal cells. ¹⁸ In addition, CD146⁺ cells also positive for sushi domain containing 2 (SUSD2) have greater regenerative capacity than CD146⁺ PDGF-Rβ⁺ cells. Moreover, SUSD2⁺ cells are able to generate stromal tissue in vivo, ¹⁹ suggesting that SUSD2⁺ cells might contribute to stromal vascular regeneration ²⁰ supporting the idea that this marker should be considered as an endometrial stem cell marker. Further, a beaded SUSD2⁺ endometrial cell population has a higher clonogenic efficiency than a flow-sorted SUSD2⁺ population, although the purity is greater by flow cytometry. ¹⁹ Probably due to the selection of SUSD2⁺ cells using magnetic beads, more cells with low expression of SUSD2 were included compared to sorter isolation, and the viability of beaded SUSD2⁺ cells were higher than flow-sorted SUSD2⁺ cells.

Recently, we identified the presence of the universal stem cell marker leucine-rich repeat–containing heterotrimeric guanine nucleotide-binding protein-coupled receptor 5 (LGR5) in the perivascular regions of the lower functionalis in both the epithelial and the stromal compartments in human endometrium. ²¹ The LGR5⁺ cells could be tissue-resident cells or transient-recruited macrophages contributing to components of the endometrial stem cell niche. ^21,22 Another proposed stem cell marker is epithelial cell adhesion molecule (EPCAM or CD326), an epithelial marker ²³ associated with endometrial progenitor cells due to their clonal efficiency and their capability to create gland-like structures in 3-dimensional cultures. ²⁴ Moreover, stage-specific embryonic antigen 1 (SSEA or CD15) surface marker is enriched in epithelial cells in the basalis layer, and these cells give rise to the functionalis layer. ²⁵

In the absence of specific stem cell markers, our group used the Side Population (SP) method ²⁶ to identify and isolate endometrial SSCs. ^14,15 These cells represent 1.68% to 2.7% of the epithelial fraction and 0.4% to 3.1% of the stromal fraction. ^15,27 Based on cloning efficiency method under hypoxic conditions and Hoechst 33342 methodology, we derived SP cell lines. ^{12,15,28 –30} Importantly, these cells (fresh or cultured) can reconstitute endometrial tissue in vivo when transplanted beneath the kidney capsule of immunocompromised mice. ^15,27,28 However, endometrial reconstruction capability remains low, likely due to the lack of an appropriate niche, which is essential for the development of stem cells. ^15,27 This hypothesis was confirmed by injecting these putative endometrial SSCs along with endometrial cells to mimic the uterine microenvironment. ³¹

We aimed to assess the endometrial reconstitution ability of different endometrial cell types isolated according to selected specific SSC markers (SUSD2 and intercellular adhesion molecule 1 [ICAM1]) and SP cell lines (ICE6 and ICE7 cell lines) in an established xenotransplantation model. To test functional properties of SSCs by xenotransplantation, injection was performed into the murine kidney capsule because it has a complex vasculature and strong blood supply, as well as a confined structure that is amenable to injecting cells. ^32,33 Moreover, previous studies have tested a similar approach, with endometrial total fraction (TF), ⁹ SP cells, ^27,28 and cultured SUSD2⁺ cells. ¹⁹ We also assessed several highly expressed markers derived from the endometrial SP ¹⁵ and related to endometrial functions that could be potential stem cell markers, including intercellular adhesion molecule 1 (ICAM1 or CD54), breast cancer resistance protein 1 (BCRP1 or ABCG2/CD338), tumor necrosis factor (TNF), and thrombospondin 1 (THBS1). The ICAM1 has been detected at perivascular locations and in stromal cells in human endometrium, ^34–36 and it is known to play an important role in immune response ³⁷ and has also exhibited high expression in mesenchymal stem cells. ^{38 –41}

Endometrial stem cell niche is still a concept and it is not defined yet; the term “niche-like cells” used in this study refers to cells similar to niche cells or cells that might reconstitute a stem cell niche-like scenario. The results of this work demonstrate the relevance of endometrial niche-like cells in providing an adequate microenvironment for the reconstitution of human endometrium in vivo by specific stem cell populations.

Materials and Methods

Experimental Design

The experimental conditions are detailed in Table 1. In every case, 500 000 cells were injected under the kidney capsule of immunocompromised and ovariectomized female non-obese diabetic–severe combined immunodeficiency (NOD–SCID) mice (n = 28). Positive control kidneys were injected with TF of human endometrial cells (500 000 cells), and noninjected kidneys represented the negative controls. The TF was also used as supplemental cells (400 000 cells) ³¹ in some animals treated with SUSD2⁺ (recognized by W5C5 antibody) and ICAM1⁺ cells (100 000 cells). Moreover, pure cells from SUSD2⁺, ICAM1⁺ as well as ICAM1⁻ and SUSD2⁻ cells were injected under the kidney capsule (500 000 cells). The SP cell lines (ICE6 and ICE7) were cultured, mixed (1/1), and injected (250 000 of each cell line) in the absence of TF. ²⁸

OPEN IN VIEWER

Table 1.

Xenotransplantation Experimental Design.^a

Total Fraction (TF)	ICAM1	SUSD2	SP Cell Lines
• 500 000 disaggregated cells from whole human endometrium (n = 3) • 500 000 disaggregated cells from whole human endometrium labeled with rhodamine (n = 5)	• 500 000 ICAM1+ (n = 3) • 100 000 ICAM1++  400 000 TF (n = 3) • 500 000 ICAM1− (n = 3)	• 500 000 SUSD2+ (n = 3) • 100 000 SUSD2+ +  400 000 TF (n = 2)b • 500 000 SUSD2− (n = 3)	• 250 000 ICE6+ 250 000 ICE7 (n = 3)

Abbreviations: ICAM1, intercellular adhesion molecule-1; SUSD2, sushi domain containing-2; SP, Side Population; n, number of mice per condition.

^aCell populations injected under renal capsule in immunocompromised mice.

^bOne mouse died during the experimental period.

Xenotransplantation Assays

Twenty-eight female NOD–SCID mice (strain code 394; NOD.CB17-Prkdc^scid/NCrCrl from Charles River Laboratories, Spain) were ovariectomized at 5 to 6 weeks of age and then used for xenotransplantation experiments (Figure 1A) following the guidelines and hormonal treatment described previously. ²⁸ Single-cell suspensions (500 000 per transplant) were resuspended in 30 μL of Hanks basal salt solution supplemented with 2% fetal bovine serum (Sigma-Aldrich, Spain) and 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) and then injected under the kidney capsule with a 27G needle.

OPEN IN VIEWER

Figure 1.

Workflow of xenotransplantation mouse model. A, Detailed diagram showing the animal model assay. One week of acclimation refers to 1-week period allowing animals to stabilize in a new environment and promoting animal welfare. B, Prussian blue staining of rhodamine-labeled cells in the kidney capsule. Assay to test the maintenance of injected cells at different times: 13, 32, 47, 50, and 57 days. Regions with iron deposit staining are magnified from the kidney pictures. The spleen is the positive control and the lung is the negative one. Scale bars: 20 µm. P₄ indicates progesterone; E₂, estradiol; OVX, ovariectomy.

Superparamagnetic iron oxide labeling of human endometrial cells

To trace human endometrial cells injected under the kidney capsule, 5 animals were treated with endometrial TF (epithelial and stromal cells) labeled with Molday ION Rhodamine B using our published protocol. ⁴³ Mice were euthanized at different time points after the injection (13, 32, 47, 50, and 60 days), and the kidneys, spleens, and lungs were collected to track these human cells. Spleen was used as the positive control due to the presence of iron deposits in the tissue, and lung as negative control (Figure 1B). Cell engraftment was assessed by intracellular iron deposit visualization through Prussian blue accustain iron stain (HT20; Sigma-Aldrich).

Mouse kidney removal and procedures

Xenotransplanted mice were euthanized 60 days after the cell injection experiments, and their kidneys were analyzed to assess the presence of endometrial cells and evaluate the endometrial tissue reconstitution. All harvested organs were fixed in 4% paraformaldehyde for 24 hours at 4°C. Fixed tissues were transferred to 70% ethanol overnight at 4°C and dehydrated in successive ethanol washes and xylenes. Finally, kidneys were paraffin embedded, serially sectioned (4 μm) in a microtome (HM 310; Microm, Spain), and mounted onto glass slides (Superfrost Plus; Thermo Scientific, Spain). Then, formalin-fixed, paraffin-embedded endometrial sections were deparaffinized with xylene and rehydrated with ethanol and water.

Histological analysis and immunostaining

Several serial sections were stained with hematoxylin–eosin. When immunostainings were performed, sections were treated for antigen retrieval and permeabilized with detergent. Then, nonspecific binding sites were blocked with BSA, detergent, and normal goat serum (Table 2). Slides were incubated with primary antibodies against human vimentin (VIM), human cytokeratin 18 (CK18), human progesterone receptor (PR), and fluorescent-labeled secondary antibodies. Finally, the slides were counterstained with 6-diamidino-2-phenylindole (DAPI; Prolong Gold Antifade reagent with DAPI, P-36931; Invitrogen, California). For Ki67, the Envision HRP system (Dako, Spain) was used for secondary antibody incubation, followed by detection with 3,3′-diaminobenzidine. Conditions, specifications, and concentrations for each antibody are detailed in Table 2. Immunolocalization was visualized under a Nikon Eclipse 80i (Spain).

OPEN IN VIEWER

Table 2.

Antibodies and Experimental Conditions Used for Immunohistochemistry.

	Antigen Retrieval	Membrane Permeabilization	Blocking Solution	Primary Antibody	Secondary Antibody	Positive Controls	Negative Controls
Vimentin	Citrate buffer 10 mmol/L 0.05% Tween 20 pH 6. 20 minutes, 95°C	PBS + Triton X-100 0.05%	PBS-BSA 15% + Triton X-100 0.05% + NGS 20%. 90 minutes, 37°C	Mouse anti-Vimentin (ab8069, Abcam) 1:200 dilution. RRID: AB_306239	Goat anti-mouse Alexa 488 Fluor (Invitrogen, A-21121), 1:500 dilution	Human endometrium	Mouse small intestine
Cytokeratin 18	Citrate buffer 10 mM 0.05% Tween-20 pH 6. 20 minutes, 95°C	PBS + Triton X-100 0.1%	PBS-BSA 5% + Tween-20 0.05% + NGS 4%. 30 minutes, RT	Rabbit polyclonal Anti-Cytokeratin 18 (ab52948, Abcam) 1:250 dilution RRID: AB_869873	Goat anti-rabbit Alexa 488 Fluor (Invitrogen, A-11034) 1:500 dilution	Human endometrium	Mouse kidney
Progesterone Receptor	Citrate buffer 10 mM 0.05% Tween-20 pH 6. 20 minutes, 95°C	PBS + Triton X-100 0.05%	PBS-BSA 15% + Triton X-100 0.05% + NGS 20%. 60 minutes, RT	Polyclonal rabbit antihuman progesterone receptor (A0098, Dako ) 1:50 dilution RRID: AB_2315192	Goat anti-rabbit Alexa 555 Fluor (Invitrogen, A-21429) 1:500 dilution	Human endometrium	Mouse kidney
Ki67	Citrate buffer 10 mM 0.05% Tween-20 pH 6. 20 minutes 95°C	PBS + NGS 1% + Triton X-100 0.4%	PBS + Triton X-100 0.4% + NGS 10%. 90 minutes, 37°C	Rabbit polyclonal anti-Ki67 (AB9260 Millipore). 1:100 dilution. RRID: AB_2142366	Envision HRP system (Dako )	Mouse small intestine	Mouse brain

Abbreviations: PBS, phosphate-buffered saline; BSA, bovine serum albumin; NGS, normal goat serum; RT, room temperature.

Engraftment efficiency

We calculated the engraftment efficiency of the injected cells by the reconstitution rate (RR). ²⁷ The RR is expressed as a percentagecorresponding to (the number of kidneys expressing all 3 antibodies (VIM, CK18, and PR) / number of kidneys transplanted per condition) × 100.

Results

Isolation and Immunocharacterization of Endometrial Stem Cells

The identification and isolation of possible stem cells were based on the expression of ICAM1, TNF, THBS1, EPCAM, SSEA1, BCRP1, and SUSD2. Flow cytometric histograms showing the immunophenotype of all tested stem cell markers are shown in Figure 2A (upper panel).

OPEN IN VIEWER

Figure 2.

Immunophenotype of human endometrium and SP cell lines. A, Percentage of positive cells for EPCAM, ICAM1, SSEA1, SUSD2, BCRP1, TNF, and THBS1 in dissociated human endometrial biopsies (n = 4 epithelium and n = 6 stroma) represented by flow cytometry. Data are shown as mean with error bars (standard deviation). The 2 selected endometrial SSC markers are boxed. In the bottom of (1A) panel, 2 representative fluorescence-activated cell sorter plots showing endometrial staining for SUSD2-APC and ICAM1-PE in whole human endometrium are shown. B, Representative fluorescence-activated cell sorter plot showing endometrial staining for SUSD2-APC and ICAM1-PE in SP cell lines (ICE6 and ICE7). A and B, Blue lines in cell sorter plots indicate positive staining for each marker compared to isotype control samples (red line). SP indicates Size Population; EPCAM, epithelial cell adhesion molecule; ICAM1, intercellular adhesion molecule-1; SSEA1, stage-specific embryonic antigen-1; SUSD2, sushi domain containing-2; BCRP1, breast cancer resistance protein-1; TNF, tumor necrosis factor; THBS1, thrombospondin-1; SSC, somatic stem cell.

The percentage of ICAM1⁺ cells was 3.5% ± 2.14% in the epithelial and 1.62% ± 0.97% in the stromal fractions. Similarly, SUSD2⁺ cells were found in 5.42% ± 2.97% and 2.95% ± 0.56% in the epithelial and stromal fractions, respectively. In contrast, TNF and THBS1 exhibited limited and inconsistent percentages of positive cell expression; therefore, we excluded these markers from further experiments. Highly variable expression of EPCAM, SSEA1, and BCRP1 was detected in all samples analyzed. We observed that the percentages of SUSD2⁺ cells are 3.89% in human endometrium (Figure 2A, lower panel), 0.169% in ICE6 cells, and 0.305% in ICE7 cells (Figure 2B, right panel). For ICAM1⁺ cells, the percentages were 2.1%, 96.2%, and 97.2% in the same populations (Figure 2A lower panel and B left panels).

Endometrial Reconstitution After Xenotransplantation

We administered endometrial TF with rhodamine under the kidney capsule and tracked cells using Prussian blue staining. Labeled cells were identified in the injected location up to 60 days after the intervention (Figure 1B). After confirming engraftment by the presence of Prussian blue staining, we assessed the reconstruction efficiency in the different experimental groups by the localization of 3 human endometrial markers (RR) and the observation of the kidney capsule thickness by hematoxylin–eosin staining (Supplemental Figure 1).

In all samples, the characterization of the newly formed endometrial tissue was assessed by immunostaining with human endometrial markers VIM, CK18, and PR (Table 3; Figure 3A shows TF injection and noninjected kidney as controls). Respective controls (positive and negative tissues for specific antibodies and kidneys with only secondary antibodies) are shown in Supplemental Figures 2 and 3.

OPEN IN VIEWER

Table 3.

Expression of Human Vimentin (VIM), Cytokeratin 18 (CK18), and Progesterone Receptor (PR) in the Capsule Kidney to Evaluate Endometrial Reconstitution After Xenotransplantation in a Murine Model.^a,b

Antibodies/Endometrial Cells	VIM	CK18	PR	RR (%)
TF	+	+	+	100
	+	+	+
	+	+	+
Noninjected kidney	−	−	−	0
	−	−	−
	−	−	−
SUSD2+	−	+	+	33
	+	−	−
	+	+	+
SUSD2+ + TF	+	+	+	100
	+	+	+
SUSD2−	+	+	+	66
	+	+	+
	−	−	−
ICAM1+	+	+	−	33
	+	+	+
	−	+	+
ICAM1+ + TF	+	+	+	100
	+	+	+
	+	+	+
ICAM1−	+	+	+	66
	+	+	+
	−	−	+
SP cell lines	+	+	+	100
	+	+	+
	+	+	+

Abbreviations: SUSD2, sushi domain containing 2; ICAM1, intercellular adhesion molecule 1; SP, Side Population.

^aEndometrial reconstitution was evaluated by the expression of 3 human endometrial antibodies (VIM, CK18, and PR) for each condition: total fraction (TF), SUSD2⁺, SUSD2⁺ + TF, SUSD2⁻, ICAM1⁺, ICAM1⁺ +TF, ICAM1⁻, and SP cell lines as well as noninjected kidney.

^bThe reconstitution rate (RR), represented as percentage, is indicated in the last row.

OPEN IN VIEWER

Figure 3.

Reconstitution rate (RR) based on the presence of human vimentin (VIM), cytokeratin 18 (CK18) and progesterone receptor (PR). The dotted lines indicate the border between human-reconstructed tissue in the kidney capsule (KC) and the mouse kidney. A, Kidneys with total fraction are the positive control of the experimental design and noninjected kidneys are the negative controls. B, Panel corresponding to SUSD2 marker, with positive (alone and supplemented) and negative fractions. C, Panel corresponding to ICAM1 marker, with positive (alone and supplemented) and negative fractions. D, Panel corresponding to SP cell lines. We considered 100% of RR when the 3 markers (VIM, CK18, and PR) are expressed in all kidneys analyzed per condition (see “Material and Methods” section). Scale bars: 10 µm. SUSD2 indicates sushi domain containing 2; ICAM1, intercellular adhesion molecule 1; SP, Side Population.

For SUSD2⁺ cells, in both negative and pure positive fractions, typical endometrial-like tissues were formed at different RRs (from 33% to 66%). Interestingly, SUSD2⁺ cells supplemented with TF had the largest RR (100%; Figure 3B). Similarly, xenotransplantation with ICAM1⁻ and ICAM1⁺ cells gave rise to lower endometrial reconstruction (RRs range between 33% and 66%) in comparison to the regenerative potential in the ICAM1⁺ supplemented fraction (Figure 3C). Unexpectedly, the ability to form endometrial tissue was greater in ICAM1⁻ and SUSD2⁻ transplants than in their positive counterparts.

The SP cell line transplants had the capacity to generate endometrium similar to TF and therefore more than cells identified by the markers described previously. Further, the new tissue generated in SP cell line transplants showed protein expression patterns similar to functional endometrium (Figure 3D).

Prussian blue-positive cells with human origin were located around blood vessels in injected kidneys (Figure 4A). Expression of Ki67 identified proliferative and active cells under the kidney capsule near the same location in all xenotransplanted kidneys (Figure 4B).

OPEN IN VIEWER

Figure 4.

Localization of human engrafted and proliferative cells in endometrial-like tissue. A, Prussian blue staining of xenotransplanted cells at perivascular location. Negative and positive control used here are mouse lung and mouse spleen, respectively. B, Proliferation assay by Ki67 showed the presence of dividing cells (indicated with arrows) around blood vessels in 2 representative regions (B1 and B2) within the kidney capsule. Negative and positive control used here are mouse brain and mouse small intestine, respectively. The dotted lines indicate the border between the human-reconstructed tissue and the mouse kidney. Scale bars: 20 μm.

Discussion

The main objective of this study was to identify bona fide markers for human endometrial SSCs using an established functional assay for human endometrial reconstruction. We demonstrated the heterogeneous formation of human endometrial-like tissue under the kidney capsule of immunocompromised mice following injection of different human endometrial cell subsets. We have deduced that injection of cells isolated using SUSD2 and ICAM1 as markers requires supplementation/support with total endometrial cells as niche-like cells due to the low endometrial reconstitution capability of SUSD2⁺ or ICAM1⁺ cells alone.

SUSD2 was previously described as a perivascular endometrial SSC marker. ¹⁹ This finding could enable the isolation of endometrial mesenchymal stem cells using a single marker; indeed, Masuda et al found that SUSD2⁺ cells could regenerate stromal tissue in vivo. However, in that study, the human endometrial cells used in xenotransplantation assays were “single-cell suspensions of cultured W5C5⁺ cells.” ¹⁹ In contrast, in our experimental design, we used fresh (noncultured) SUSD2⁺ and SUSD2⁻ cells, and surprisingly, neither combination led to significant results. Other factors that may influence the regeneration ability of these cells, such as the sorting procedure, may diminish the cell’s viability as described by Masuda et al. ¹⁹ Nevertheless, full endometrial reconstitution was observed with supplemented and noncultured cells. We propose that (1) the previous in vitro step could affect the final phenotype of isolated cells, promoting in vivo cell proliferation, and (2) the endometrial surrounding cells play an important role, acting as niche-like cells for tissue reconstitution in in vivo models.

We also analyzed endometrial epithelial and stromal cell fractions for cells expressing ICAM1 and concluded that the percentage of positive cells was consistent with the expected percentage of endometrial SSCs in thebwhole tissue. ^14,44 High expression of ICAM1 has also been detected in endothelial cells of veins, arterioles and capillaries, and stromal cells within human endometrium. ^{34 –36} This marker plays a significant role in immunological synapse formation, T-cell activation, leukocyte trafficking, and numerous cellular immune responses. ³⁷ Importantly, ICAM1 has exhibited high expression in mesenchymal stem cells. ^{38 –41} Together, this evidence suggested the potential of this molecule to be considered an endometrial SSC marker. However, consistent with our SUSD2 assays, ICAM1⁺ cells alone gave rise to low endometrial reconstitution efficiency, and ICAM1⁺ cells supplemented with total tissue promote efficient reconstruction.

Presumably, both negative fractions (for SUSD2 and ICAM1) resulted in greater endometrial regeneration in comparison to positive pure fractions. All together, these results reinforce the important role of endometrial niche-like cells as supportive population maintaining, guiding, and stimulating the endometrial SSC population for the reconstruction of human endometrium in an animal model. ⁴⁵

Finally, we tested established SP cell lines derived by endometrial stem cells, since this population has been proposed to include different stem/progenitor cells. In comparison to the other xenotransplant assays, the SP cells contributed to greater endometrial-like tissue reconstruction. Thus, we postulate that the SP from human endometrium could be enriched, at least in part, in several cells with stemness capabilities, corroborating previous findings. ^15,26,27,46 It is important to underline that the SP cells used here were derived from stable cell lines ¹⁵ (established karyotype, stem cell characteristics, and reconstitution potential), while the latter populations (SUSD2 and ICAM1) are isolated from primary samples. This fact could contribute to their proliferation activity.

To highlight the role of the niche or surrounding cells, we added total endometrial cells to supplement other cell types. The importance of the niche was described in detail by Lane and collaborators who highlighted cell components such as resident niche cells and direct cell contacts. ⁴⁷ Theoretically, stem cell niches contain both tissue-specific and generic cell populations having specialized roles in each context. The endometrial TF could act as a niche, resulting in an increased reconstruction efficiency, as described previously ³¹ and in our present work.

Moreover, in the proliferation assay, Ki67 staining indicated the presence of these human proliferating cells in the kidney capsule near blood vessels, where engrafted human cells were located. This fact indicates that human cells may drive the proliferation of surrounding cells and the differentiation to endometrial cell types such as epithelium or stroma. This is consistent with the plausible location of the endometrial niche, around blood vessels, corroborating the dialogue between this tissue and the bone marrow in endometrial restoration. ^{11,21,43,48 –50}

In conclusion, we assessed in a xenotransplantation model that pure SUSD2⁺ and ICAM1⁺ cells from human endometrium do not have the ability to efficiently reconstruct endometrium. Indeed, we demonstrated that isolated SUSD2⁺ and ICAM1⁺ cells require supplementation with total endometrial cells as specific niche-like cells. The use of whole endometrial tissue enabled a close association between putative endometrial stem cells with their niche cells. ⁵¹ In contrast, the SP appears to contain a stem cell population, indicated by its high reconstitution potential. As we postulated in prior work, the SP is a heterogeneous cell population containing a mixture of endometrial progenitors (epithelium, stroma, and endothelium).

Despite the results obtained, we are aware about the limitations of this study. Due to the low percentage of SP cells present in human endometrium, we used cultured cell lines, while the other endometrial cell populations analyzed in this study (SUSD2 and ICAM1) came from sorted primary samples. This last step (flow cytometry) could have a harmful effect on the cellular viability compromising probably their reconstitution efficiency in animal models. In any case, the reconstitution ability of the endometrial TF has been demonstrated previously ⁹ and confirmed along this work. We assessed here the capability of putative endometrial stem cells with and without TF to study their necessity as support or niche-like cells because it is widely known that the niche microenvironment is essential for the maintenance of stem cell features. ⁵² This endometrial TF should harbor niche-like/neighbor cells belonging to the stem cell niche that would enhance the regenerative potential of these postulated stem cells as described previously. ³¹ The identification of specific endometrial stem cell markers is essential for the study of the endometrial biology, stem cell isolation, and the understanding of endometrial pathologies.

Supplemental Material

Supplemental Material

Supplemental Material