Progesterone-driven stabilization of hybrid E/M states in amniotic epithelial cells enhances regeneration and immune modulatory capacities

TL;DR

Progesterone helps amniotic cells maintain a hybrid state that boosts their ability to regenerate tissues and control immune responses.

Contribution

This study reveals that progesterone stabilizes hybrid epithelial/mesenchymal states in amniotic cells, enhancing their regenerative and immunomodulatory functions.

Findings

Progesterone delays epithelial-mesenchymal transition and promotes hybrid E/M phenotypes in amniotic epithelial cells.

Hybrid cells show improved collective migration, stemness, and immunomodulation in vitro and in vivo.

Transplantation of hybrid cells accelerates tendon healing and modulates macrophage polarization in an ovine model.

Abstract

Epithelial-mesenchymal plasticity (EMP) plays a pivotal role in development, regeneration, and disease progression. In this study, we demonstrate that progesterone (P4) delays EMP in amniotic epithelial cells (AECs), promoting the emergence of hybrid epithelial/mesenchymal (E/M) phenotypes. These hybrid cells co-express E and M traits and exhibit distinct surface markers. Compared to fully M, hybrid AECs display enhanced collective migration, upregulation of stemness transcription factors, and enhanced immunomodulatory properties in vitro and in vivo. Their regenerative potential was validated by in vitro tendon differentiation on PLGA-fleeces and in an ovine tendon injury model, where the transplantation of hybrid AECs accelerated early regeneration. This effect was associated with a timely transition from inflammation to proliferation, mediated by macrophage polarization and extracellular matrix remodeling. Our findings reveal that P4 maintains AECs in a functionally hybrid E/M state, conferring regenerative and immunomodulatory advantages. These results offer insights into the physiological regulation of EMP and support the therapeutic relevance of hybrid AECs as promising candidates for cell-based regenerative medicine therapy.

•P4 delays EMT and promotes the emergence of AECs with hybrid E/M phenotypes•Hybrid AECs show enhanced stemness, differentiation, and collective migration•In vivo, hybrid AECs accelerate tendon healing and modulate macrophage polarization•Hybrid phenotypes offer a new framework for EMP-based cell therapies

P4 delays EMT and promotes the emergence of AECs with hybrid E/M phenotypes

Hybrid AECs show enhanced stemness, differentiation, and collective migration

In vivo, hybrid AECs accelerate tendon healing and modulate macrophage polarization

Hybrid phenotypes offer a new framework for EMP-based cell therapies

Molecular biology; Immunology; Developmental biology

Figures

Progesterone enhances hybrid phenotype acquisition in AECs during EMT(A) Representative flow cytometry histograms show the expression of epithelial markers (EpCAM, E-Cadherin), mesenchymal markers (CD73, CD90), and hybrid markers (CD51, CD61, CD106) in freshly isolated AECs (green), AECs cultured with progesterone (P4) at passages 1 (light green), 2 (yellow), and 3 (orange), and AECs undergoing spontaneous EMT during three passages (mAECs; shades of red). Each plot shows fluorescence intensity (x axis) versus cell count (y axis) for the indicated marker.(B) Quantitative analysis of surface mar

Dynamic progression of hybrid E/M states in P4-treated AECs defines a distinct transitional phase(A) Representative flow cytometry dot plots show the co-expression of epithelial (E-Cadherin or EpCAM) and mesenchymal (CD90 or CD73) surface markers in freshly isolated AECs (green), P4-treated AECs at passages 1 (light green), 2 (yellow), and 3 (orange), and AECs undergoing spontaneous EMT during three passages (mAECs; shades of red). The combinations analyzed include E-Cadherin^+^CD90^+^, E-Cadherin^+^CD73^+^, EpCAM^+^CD90^+^, and EpCAM^+^CD73^+^. Gates define double-positive populations represe

Hybrid AECs display morphological and molecular features intermediate between epithelial and mesenchymal states(A) Representative bright-field and immunofluorescence images of freshly isolated AECs, hyAECs (hyAECs), and mesenchymal AECs (mAECs). Freshly isolated AECs display a typical cobblestone-like epithelial morphology and are positive for E-Cadherin (green) while negative for α-smooth muscle actin (α-SMA, red). In contrast, mAECs exhibit a spindle-shaped mesenchymal morphology with strong α-SMA expression and reduced E-Cadherin. HyAECs show intermediate morphological features with mixed c

Transient upregulation of stemness factors Oct4 and Nanog defines the hybrid E/M state of AECsWestern blot analysis of pluripotency-associated transcription factors Oct4, Nanog, and Sox2 in amniotic membrane (AM), hyAECs (hyAECs), and mesenchymal AECs (mAECs). Tubulin was used as a loading control. Protein expression levels were quantified and normalized to tubulin. Data (mean ± SD) represent 3 independent sets of experiments (n = at least 3 biological replicates in each group per set; each biological replicate assayed in at least 3 technical replicates). Oct4 and Nanog expression were signifi

Hybrid AECs exhibit enhanced *in vitro* tenogenesis, collective migration and improved *in vitro* immunosuppressive capacity(A) Tendon differentiation potential of hyAECs (hyAECs) and mesenchymal AECs (mAECs) cultured for 48 h on PLGA fleeces. Gene expression of tendon markers Scleraxis (SCX), Tenomodulin (TNMD), and Collagen Type I (COL1) was assessed by qPCR. Western blot confirmed higher TNMD and COL1 protein expression in hyAECs versus mAECs engineered fleeces. Immunofluorescence staining revealed stronger TNMD and COL1 expression and enhanced the alignment of hyAECs along fleece fibers, i

Hybrid AECs promote early tendon healing and ECM remodeling in an ovine Achilles tendon injury modelRepresentative images show the *in vivo* regenerative effects of hyAECs (hyAECs) compared to mesenchymal AECs (mAECs) and control groups in a validated ovine Achilles tendon injury model. Animals were monitored for 14 days post-transplantation.Macroscopic evaluation of explanted tendons revealed reduced hemorrhage and edema in hyAEC-treated samples compared to mAEC-treated tendons and controls. PKH26-labeled AECs (red fluorescence) were detected within the lesion at days 7 and 14, predominantly

HyAECs promote macrophage polarization toward a pro-repair phenotype(A) Representative immunofluorescence images of ovine tendon sections at 7- and 14-day post-transplantation showing immunopositivity for CD86 (M1 macrophages) and CD206 (M2 macrophages) following hybrid AEC (hyAEC) or mesenchymal AEC (mAEC) treatment and in control tendons. CD206^+^ (green fluorescence) M2 macrophages were significantly more abundant in hyAEC-treated tendons, compared to mAEC-treated samples, but mostly to controls. CD86^+^ (green fluorescence) pro-inflammatory macrophages were reduced in both AEC-treated grou

HyAECs modulate inflammation by enhancing vascular remodeling in injured tendon tissueRepresentative immunofluorescence images show von Willebrand Factor (vWF; green) staining of blood vessels in tendon sections from control, mAEC-, and hyAEC-treated animals at 14 days post-transplantation. vWF was used as a marker to assess blood vessels and vascular organization within the regenerating tissue.Control tendons displayed disorganized and dense vascular networks, indicative of ongoing inflammation and poor structural integration. In contrast, AEC-transplanted tendons exhibited improved vascular

Full text

Introduction

Epithelial-mesenchymal transition (EMT) is a dynamic and reversible process by which epithelial cells lose their polarity and junctional integrity, acquiring mesenchymal features such as motility and invasiveness.1 While initially viewed as a binary switch, EMT is now recognized as a continuum of states, a concept defined as epithelial-mesenchymal plasticity (EMP).2 Within this spectrum, hybrid epithelial/mesenchymal (E/M) phenotypes have emerged as distinct and functionally relevant entities that co-express both epithelial (E) and mesenchymal (M) traits.2 These states contribute critically to physiological and pathological contexts, including embryonic development, wound healing, fibrosis, and cancer metastasis.3^,^4

Hybrid E/M cells exhibit functional attributes not found in fully E or M states, such as collective migration, enhanced plasticity, metabolic adaptability, and immunomodulatory capacity.5^,^6^,^7 During embryogenesis, they enable coordinated tissue morphogenesis, while in postnatal life, they have been implicated in tissue regeneration.8 In oncology, hybrid states drive collective invasion and dissemination of tumor clusters, suggesting they are not transient intermediates, but stable and biologically potent phenotypes.8

The acquisition and stabilization of hybrid states are regulated by environmental signals and intracellular pathways, including hormonal cues. Among these, progesterone (P4), a steroid ormone critical for pregnancy maintenance and tissue homeostasis, has been shown to modulate EMT in amniotic epithelial cells (AECs)9 and to affect amniotic membrane remodeling.10^,^11 AECs are a stem cell population capable of spontaneous EMT during in vitro expansion and endowed with regenerative and immunomodulatory properties. Prior studies have reported that P4 enhances these functional properties and supports tendon healing in vivo.12^,^13^,^14^,^15 However, whether these effects are linked to the stabilization of hybrid E/M phenotypes remains unclear.

In this study, we investigated whether P4 modulates the EMP trajectory of AECs by fostering the emergence of functionally potent hybrid E/M states. We performed a comprehensive phenotypic, molecular, and functional characterization of these hybrid populations, including their surface marker profile, stemness features, collective migratory behavior, immunoregulatory capacity, and regenerative efficacy both in vitro and in an animal model of tendon injury.

Results

Progesterone enhances hybrid phenotype acquisition in amniotic epithelial cells during epithelial-mesenchymal transition

To evaluate the effect of P4 on EMP in AECs, we analyzed the expression of E (EpCAM, E-Cadherin), M (CD73, CD90), and hybrid (CD51, CD61, CD106) surface markers by flow cytometry (Figure 1). Freshly isolated AECs (eAECs) served as the E reference, while fully transitioned M cells (mAECs), obtained after three passages in standard culture (without progesterone), represented the opposite end of the EMP spectrum.Figure 1. Progesterone enhances hybrid phenotype acquisition in AECs during EMT(A) Representative flow cytometry histograms show the expression of epithelial markers (EpCAM, E-Cadherin), mesenchymal markers (CD73, CD90), and hybrid markers (CD51, CD61, CD106) in freshly isolated AECs (green), AECs cultured with progesterone (P4) at passages 1 (light green), 2 (yellow), and 3 (orange), and AECs undergoing spontaneous EMT during three passages (mAECs; shades of red). Each plot shows fluorescence intensity (x axis) versus cell count (y axis) for the indicated marker.(B) Quantitative analysis of surface marker expression based on the mean fluorescence intensity (MFI) ratio. Data (mean ± SD) represent 3 independent sets of experiments (n = at least 3 biological replicates in each group per set; each biological replicate assayed in at least 3 technical replicates). Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparisons test (∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001).

In AECs, hybridness followed a clear passage-dependent pattern (Figure 1). In the presence of P4, E markers were transiently retained, with EpCAM and E-cadherin increasing 4.9-fold and 2.6-fold at passage 2, whereas in untreated cultures these markers declined markedly, showing a 4.0-fold and 13.2-fold reduction (Figure 1). Hybrid markers showed a similar trend. In P4-treated cells, CD51 increased 397.1-fold, CD106 increased 10.2-fold, and CD61 increased 5.8-fold at passage 2, while in the absence of P4, only CD51 showed a moderate rise of 67.9-fold, and CD61 and CD106 did not increase (Figure 1). Mesenchymal markers further reflected delayed EMT progression with P4, with CD73 and CD90 increasing 5.0-fold and 1.7-fold at passage 2, compared with much stronger increases of 14.4-fold and 12.4-fold in untreated cells (Figure 1). Together, these coordinated trajectories indicate that P4 promotes the early accumulation of hybrid E/M traits and stabilizes this intermediate phenotype through passages 2 and 3 before progression toward a mesenchymal state occurs.

Dynamic progression of hybrid epithelial/mesenchymal states in P4-treated amniotic epithelial cells defines a distinct transitional phase

To confirm the emergence of hybrid phenotypes observed in single-marker analysis, we evaluated the co-expression of E and M markers by flow cytometry (Figure 2A). Specifically, we quantified the percentages of cells double-positive for E-Cadherin^+^CD90^+^, E-Cadherin^+^CD73^+^, EpCAM^+^CD90^+^, and EpCAM^+^CD73^+^ across three passages in P4-treated cultures and compared them to eAECs and mAECs cultured without P4.Figure 2. Dynamic progression of hybrid E/M states in P4-treated AECs defines a distinct transitional phase(A) Representative flow cytometry dot plots show the co-expression of epithelial (E-Cadherin or EpCAM) and mesenchymal (CD90 or CD73) surface markers in freshly isolated AECs (green), P4-treated AECs at passages 1 (light green), 2 (yellow), and 3 (orange), and AECs undergoing spontaneous EMT during three passages (mAECs; shades of red). The combinations analyzed include E-Cadherin^+^CD90^+^, E-Cadherin^+^CD73^+^, EpCAM^+^CD90^+^, and EpCAM^+^CD73^+^. Gates define double-positive populations representing hybrid E/M phenotypes.(B) Quantification of the percentage of cells co-expressing paired E/M markers in each condition. Data (mean ± SD) represent 3 independent sets of experiments (n = at least 3 biological replicates in each group per set; each biological replicate assayed in at least 3 technical replicates). Peak co-expression was observed at passage 2, identifying this stage as enriched for hybrid epithelial/mesenchymal phenotypes. Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparisons test (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

As expected, cells co-expressing E and M markers were minimally represented in both eAECs and mAECs, and never exceeded 26% and 29%, respectively (Figure 2B). However, considering the lower expression recorded in single surface marker analysis (Figure 1), such values likely reflect residual E marker expression in mAECs and low native M marker expression in eAECs. In contrast, P4-treated cells exhibited a progressive increase in E/M co-expression, with values exceeding 38% by passage 1 (p < 0.001 vs. eAECs), and peaking at passage 2, where most hybrid combinations reached ∼46% (p < 0.001 vs. passage 1) (Figure 2B). By passage 3, these values declined (p < 0.05 vs. passage 2) but remained significantly higher than in mAECs (Figure 2B).

This temporal trend reveals a dynamic P4-induced enrichment of hybrid E/M phenotypes, with passage 2 representing the peak of co-expression. Based on these profiles, we defined three transitional phases: Late E (passage 1), Early Hybrid E/M (passage 2), and Late Hybrid E/M (passage 3). Early Hybrid E/M cells showed a more balanced E/M co-expression compared to the two late stages, which were predominantly E-like (Late E) or predominantly M-like (Late Hybrid E/M). The early hybrid E/M population was therefore selected for further characterization and is simply referred to as hyAECs throughout the study.

Hybrid amniotic epithelial cells display morphological and molecular features intermediate between epithelial and mesenchymal states

To characterize hyAECs at the morphological and molecular level, we compared them to eAECs and mAECs (Figure 3A). eAECs exhibited the typical cobblestone-like morphology and strong E-Cadherin expression, while lacking α-smooth muscle actin (α-SMA). In contrast, mAECs displayed a spindle-shaped M morphology with high α-SMA expression and reduced E-Cadherin (Figure 3A). Interestingly, hyAECs showed an intermediate phenotype, combining cobblestone-like and elongated features. Immunofluorescence confirmed the co-expression of E-Cadherin and α-SMA, supporting their hybrid identity (Figure 3A).Figure 3. Hybrid AECs display morphological and molecular features intermediate between epithelial and mesenchymal states(A) Representative bright-field and immunofluorescence images of freshly isolated AECs, hyAECs (hyAECs), and mesenchymal AECs (mAECs). Freshly isolated AECs display a typical cobblestone-like epithelial morphology and are positive for E-Cadherin (green) while negative for α-smooth muscle actin (α-SMA, red). In contrast, mAECs exhibit a spindle-shaped mesenchymal morphology with strong α-SMA expression and reduced E-Cadherin. HyAECs show intermediate morphological features with mixed cobblestone and elongated profiles and co-express both E-Cadherin and α-SMA, consistent with a hybrid epithelial/mesenchymal phenotype. Nuclei are counterstained with DAPI (blue). Scale bars, 25μm.(B) Western blot analysis of epithelial (E-Cadherin), mesenchymal (Vimentin, α-SMA), and hybrid-associated phenotypic stability factors (Jagged1, Nrf2) in amniotic membrane (AM), hyAECs, and mAECs. Tubulin was used as a loading control. Quantification of band intensities (normalized to tubulin) is shown in the accompanying graphs. E-Cadherin expression is highest in AM and decreases progressively across the EMP spectrum. Vimentin, α-SMA, Jagged1, and Nrf2 levels increase from AM to hyAECs and peak in mAECs, reflecting the transition from epithelial to mesenchymal identity. Data (mean ± SD) represent 3 independent sets of experiments (n = at least 3 biological replicates in each group per set; each biological replicate assayed in at least 3 technical replicates). Statistical significance was determined using one-way ANOVA with post hoc comparisons (∗p < 0.05, ∗∗p < 0.01).

To investigate the molecular mechanisms underlying P4-mediated EMT inhibition, we exposed P4-treated hyAECs to the effects of inhibitors targeting either the nuclear progesterone receptor PR (25 μM RU-486, Mifepristone) or the membrane-associated receptor PGRMC1 (AG-205, 5 and 10 μM) (Figure S1). Blocking either receptor consistently abrogated the effect of P4, enabling the cells to complete EMT and acquire an M phenotype (Figure S1).

Western blot analysis corroborated these findings (Figure 3B). E-Cadherin expression decreased progressively from eAECs to hyAECs to mAECs, while Vimentin and α-SMA increased along the same trajectory (Figure 3B). Notably, E-Cadherin remained significantly higher in hyAECs than in mAECs (p < 0.05), whereas Vimentin and α-SMA were moderately upregulated, consistent with a partial acquisition of M traits (Figure 3B).

We also examined the expression of phenotypic stability factors (PSFs) known to support hybrid E/M identity (Figure 3B). Nrf2 showed the highest expression in hyAECs, significantly increasing compared to eAECs (p < 0.01) and decreasing again in mAECs (p < 0.001 vs. hyAECs) (Figure 3B). Conversely, Jagged1 expression was elevated in hyAECs compared to eAECs (p < 0.05) but continued to increase in mAECs (p < 0.05 vs. hyAECs) (Figure 3B). These findings indicate that Nrf2 may play a key role in stabilizing the hybrid state, while Jagged1 expression may extend beyond the hybrid phase into full EMT.

Transient upregulation of stemness factors Oct4 and Nanog defines the hybrid epithelial/mesenchymal state of amniotic epithelial cells

To explore the relationship between EMP and stemness in AECs, we assessed the expression of pluripotency-associated transcription factors Oct4, Nanog, and Sox2 by Western blot (Figure 4).Figure 4. Transient upregulation of stemness factors Oct4 and Nanog defines the hybrid E/M state of AECsWestern blot analysis of pluripotency-associated transcription factors Oct4, Nanog, and Sox2 in amniotic membrane (AM), hyAECs (hyAECs), and mesenchymal AECs (mAECs). Tubulin was used as a loading control. Protein expression levels were quantified and normalized to tubulin. Data (mean ± SD) represent 3 independent sets of experiments (n = at least 3 biological replicates in each group per set; each biological replicate assayed in at least 3 technical replicates). Oct4 and Nanog expression were significantly upregulated in hyAECs compared to AM and mAECs, suggesting a transient activation of stemness programs during the hybrid E/M state. In contrast, Sox2 expression progressively increased along the EMP spectrum, peaking in mAECs. Statistical significance was assessed using unpaired Student’s t test (∗p < 0.05, ∗∗p < 0.01).

Oct4 and Nanog were significantly upregulated in hyAECs compared to eAECs (p < 0.05), suggesting a transient activation of stemness programs during the hybrid phase (Figure 4). Notably, their expression declined in mAECs, reaching levels comparable to eAECs, indicating that this upregulation is restricted to the hybrid state (Figure 4).

Sox2, by contrast, exhibited a distinct pattern. Its expression progressively increased across the EMP spectrum, peaking in mAECs (p < 0.05 vs. eAECs and hyAECs) (Figure 4). This suggests that, unlike Oct4 and Nanog, Sox2 may also contribute to stemness maintenance after full EMT, possibly supporting long-term plasticity in M cells.

These results demonstrate that the hybrid E/M state is defined by a transient upregulation of Oct4 and Nanog, potentially conferring increased plasticity and priming cells for lineage commitment before completing EMT.

Hybrid amniotic epithelial cells exhibit enhanced in vitro tenogenesis, collective migration, and improved in vitro immunosuppressive capacity

To determine whether the transient stemness observed in hyAECs translates into increased differentiation potential, we assessed their tenogenic response under inductive conditions. In detail, hyAECs and mAECs were cultured for 48 h on PLGA fleeces, a scaffold known to mimic tendon architecture and promote tenogenesis16^,^17 (Figure 5A).Figure 5. Hybrid AECs exhibit enhanced in vitro tenogenesis, collective migration and improved in vitro immunosuppressive capacity(A) Tendon differentiation potential of hyAECs (hyAECs) and mesenchymal AECs (mAECs) cultured for 48 h on PLGA fleeces. Gene expression of tendon markers Scleraxis (SCX), Tenomodulin (TNMD), and Collagen Type I (COL1) was assessed by qPCR. Western blot confirmed higher TNMD and COL1 protein expression in hyAECs versus mAECs engineered fleeces. Immunofluorescence staining revealed stronger TNMD and COL1 expression and enhanced the alignment of hyAECs along fleece fibers, indicating superior tenogenic commitment. Scale bars, 100μm.(B) Migration capacity assessed by wound healing assay at 24 and 48 h. Representative images show collective migration in hyAECs and individual migration in mAECs. Quantitative analysis revealed that mAECs migrated significantly faster than hyAECs (10.42 μm/h vs. 6.25 μm/h; ∗∗∗p < 0.001). Scale bars, 50μm.(C) Immunomodulatory activity assessed by measuring the proliferation of PHA-activated peripheral blood mononuclear cells (PBMCs) following exposure to conditioned media (CM) from hyAECs and mAECs, either under basal or LPS-stimulated conditions. Both cell types suppressed PBMCs’ proliferation, but LPS-stimulated hyAEC-CM induced significantly stronger suppression than mAEC-CM. PHA alone served as a positive control. Data (mean ± SD) represent 3 independent sets of experiments (n = at least 3 biological replicates in each group per set; each biological replicate assayed in at least 3 technical replicates). Statistical significance was calculated using unpaired Student’s t test (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001).

Gene expression analysis revealed that both cell types upregulated the tendon-associated markers Scleraxis (SCX) and Tenomodulin (TNMD) upon culture on the fleece (Figure 5A). However, TNMD expression was significantly higher in hyAECs than in mAECs (p < 0.05), indicating a more robust tenogenic response (Figure 5A). Collagen type I (COL1) was also significantly upregulated in hyAECs (p < 0.05), whereas mAECs, despite exhibiting higher basal COL1 expression, did not show further induction on the scaffold (Figure 5A).

These transcriptional findings were corroborated at the protein level. Western blot analysis confirmed the upregulation of TNMD and COL1 in hyAECs, with significantly stronger responses than in mAECs (p < 0.01) (Figure 5A). Immunofluorescence imaging further supported these data, showing intense TNMD and COL1 staining in hyAECs, along with marked alignment along the fleece fibers (Figure 5A). In contrast, mAECs displayed strong COL1 expression but weaker TNMD levels and reduced alignment (Figure 5A).

Altogether, these findings indicate that hyAECs exhibit superior responsiveness to teno-inductive cues and enhanced early tendon differentiation potential compared to fully M cells, suggesting that the hybrid E/M state may confer a plastic advantage during lineage commitment.

Collective migration is a functional hallmark of hybrid E/M phenotypes. To assess this behavior in hyAECs, we performed wound healing assays and compared their migratory dynamics to mAECs over a 48-h period (Figure 5B). As a result, hyAECs migrated as cohesive cell clusters, maintaining intercellular contacts as they advanced toward the wound area (Figure 5B). In contrast, mAECs exhibited individual, dispersed migration, consistent with their full M identity (Figure 5B). While hyAECs preserved collective organization, their overall migration velocity was significantly reduced compared to mAECs (6.25 μm/h vs. 10.42 μm/h; p < 0.001) (Figure 5B).

These results confirm that hyAECs adopt a coordinated, collective migratory behavior characteristic of hybrid E/M cells, albeit with a slower progression than mAECs, which display faster, individual motility typical of cells that have completed EMT.

To compare the immunomodulatory capacity of hyAECs and mAECs, we evaluated the proliferation of PHA-activated peripheral blood mononuclear cells (PBMCs) after exposure to conditioned media (CM) collected under basal or LPS-stimulated conditions (Figure 5C).

Under basal conditions, CM from both cell types significantly reduced PBMCs' proliferation, with no substantial difference between hyAECs and mAECs. This suggests that both phenotypes possess inherent immunosuppressive potential in the absence of inflammatory cues.

However, following LPS stimulation, CM from hyAECs exhibited a markedly stronger inhibitory effect on PBMCs proliferation compared to mAEC-derived CM (p < 0.05). Notably, mAECs did not enhance their immunomodulatory activity beyond basal levels in response to LPS. These findings indicate that hyAECs are more responsive to inflammatory stimuli and can release soluble factors with enhanced anti-inflammatory properties. This enhanced immunomodulatory response reinforces the therapeutic potential of hyAECs, particularly in inflamed or injury-associated microenvironments.

Hybrid amniotic epithelial cells promote early tendon healing and enhanced extracellular matrix remodeling in an ovine Achilles tendon injury model

To evaluate the regenerative potential of hyAECs in vivo, we employed a validated ovine Achilles tendon injury model.18 All animals well tolerated the surgical procedures, recovered mobility immediately after anesthesia, and showed no signs of pain during the 14-day observation period.

Both treatments improved tendon healing compared to controls (Figure 6).Figure 6. Hybrid AECs promote early tendon healing and ECM remodeling in an ovine Achilles tendon injury modelRepresentative images show the in vivo regenerative effects of hyAECs (hyAECs) compared to mesenchymal AECs (mAECs) and control groups in a validated ovine Achilles tendon injury model. Animals were monitored for 14 days post-transplantation.Macroscopic evaluation of explanted tendons revealed reduced hemorrhage and edema in hyAEC-treated samples compared to mAEC-treated tendons and controls. PKH26-labeled AECs (red fluorescence) were detected within the lesion at days 7 and 14, predominantly localized at the injury border near intact tissue. Transplanted cells retained a polyhedral or elongated morphology, which showed the co-colocalization of the red (PKH26) and green (COL1) fluorescent signals (arrows) and were frequently aligned along COL1-rich regions (green fluorescence).By day 14, COL1 fibers in hyAEC-treated tendons appeared more organized and aligned along the tendon’s longitudinal axis than in mAEC-treated tendons, reflecting enhanced extracellular matrix (ECM) maturation. No or blunt positivity to COL1 is visible in control tendons. White dotted lines show healthy portions of the tendon. Nuclei are counterstained with DAPI (blue). Scale bars, 20 μm; Scale bars, large image 100 μm.These findings demonstrate that hyAECs promote early tendon repair and support timely ECM deposition and structural organization.

Macroscopically, hyAEC-treated tendons displayed reduced hemorrhage and edema at explant (Figure 6). PKH26-labeled AECs were detected at the injury margins on days 7 and 14, where they retained a polyhedral or elongated morphology and were frequently aligned within collagen-rich regions (Figure 6, higher magnification). Immunofluorescence analysis revealed early COL1 expression by transplanted cells as early as day 7, suggesting rapid in situ differentiation and initiation of matrix deposition (Figure 6).

By day 14 post, COL1 fibers in hyAEC-treated tendons appeared more organized and longitudinally aligned compared to mAEC-treated tendons, reflecting more advanced ECM maturation (Figure 6). In contrast, control tendons showed faint or absent COL1 positivity. Furthermore, PKH26^+^-AECs were still visible at this time point, immersed within the aligned COL1 fibers (Figure 6).

These findings demonstrate that hyAECs enhance ECM deposition and organization during the early phases of tendon healing, underscoring their higher regenerative potential in vivo.

Hybrid amniotic epithelial cells promote macrophage polarization toward a pro-repair phenotype

To investigate whether hyAECs influence the inflammatory microenvironment during tendon repair, we analyzed macrophage polarization markers in injured tendons at 7- and 14-day post-transplantation (Figure 7A). Immunofluorescence revealed a higher density of CD206^+^ M2 macrophages in hyAEC-treated tendons compared to both mAEC-treated and control groups on day 7. This M2 enrichment persisted on day 14 in the hyAEC group, whereas it declined in the mAEC and control tendons (Figure 7A). Conversely, CD86^+^ M1 macrophages showed comparable distribution across groups on day 7 but significantly decreased by day 14 in both AEC-treated groups, with the most pronounced reduction in hyAEC-treated tendons (Figure 7A). Notably, PKH26^+^-AECs did not colocalize with either CD206 or CD86 signals.Figure 7. HyAECs promote macrophage polarization toward a pro-repair phenotype(A) Representative immunofluorescence images of ovine tendon sections at 7- and 14-day post-transplantation showing immunopositivity for CD86 (M1 macrophages) and CD206 (M2 macrophages) following hybrid AEC (hyAEC) or mesenchymal AEC (mAEC) treatment and in control tendons. CD206^+^ (green fluorescence) M2 macrophages were significantly more abundant in hyAEC-treated tendons, compared to mAEC-treated samples, but mostly to controls. CD86^+^ (green fluorescence) pro-inflammatory macrophages were reduced in both AEC-treated groups versus untreated controls, indicating the accelerated resolution of inflammation. Nuclei were counterstained with DAPI (blue fluorescence). Scale bars, 50μm.(B) Western blot analysis of CD206 and CD86 protein expression in tendon lysates at 7- and 14-day post-transplantation. Quantification revealed significantly higher CD206 levels in hyAEC-treated tendons compared to mAEC-treated groups (p < 0.05), and a corresponding reduction in CD86 expression, with the lowest levels in the hyAEC condition (p < 0.05 vs. mAECs).

Western blot analysis confirmed these observations (Figure 7B). CD206 expression was significantly elevated in hyAEC-treated tendons at both timepoints compared to mAEC-treated samples (p < 0.05), while CD86 levels were lower, with the most significant reduction observed in the hyAEC group at day 14 (p < 0.05 vs. mAECs) (Figure 7B). These findings suggest that hyAECs promote a sustained anti-inflammatory environment and favor macrophage polarization toward a pro-regenerative M2 phenotype, contributing to a microenvironment conducive to tissue repair.

HyAECs modulate inflammation by enhancing vascular remodeling in injured tendon tissue

To assess the effects of AEC transplantation on vascular remodeling, we analyzed blood vessel organization in tendon sections using von Willebrand Factor (vWF) staining at 14 days post-injury (Figure 8). Control tendons displayed a dense and disorganized vascular network, indicative of persistent inflammation and delayed tissue remodeling. In contrast, tendons treated with AECs exhibited improved vascular structure. In mAEC-treated tendons, vessels began to align along the longitudinal tendon axis. Notably, hyAEC-treated tendons showed a more advanced angiogenic response, characterized by a well-organized vascular network with vessels clearly aligned along the tendon’s main axis (Figure 8).Figure 8. HyAECs modulate inflammation by enhancing vascular remodeling in injured tendon tissueRepresentative immunofluorescence images show von Willebrand Factor (vWF; green) staining of blood vessels in tendon sections from control, mAEC-, and hyAEC-treated animals at 14 days post-transplantation. vWF was used as a marker to assess blood vessels and vascular organization within the regenerating tissue.Control tendons displayed disorganized and dense vascular networks, indicative of ongoing inflammation and poor structural integration. In contrast, AEC-transplanted tendons exhibited improved vascular remodeling by day 14. Blood vessels in mAEC-treated tendons began to align along the tendon axis, while hyAEC-treated tendons demonstrated a more advanced angiogenic response, characterized by a well-structured and clear longitudinal organized vascular network. Scale bars, 100 μm.

These results suggest that hyAECs not only promote ECM deposition and inflammation resolution but also support the formation of a structured and mature vascular network, further contributing to effective tissue regeneration.

Discussion

This study introduces a distinct biological model that enables a deeper exploration of EMP in AECs. Unlike late-term AECs, mid-gestation cells retain a fully E phenotype at isolation, express both nuclear and membrane progesterone receptors, and undergo spontaneous EMT in vitro.9 These features create a developmental context in which P4 does not merely maintain E identity19 but actively delays EMT progression, generating a defined window in which stable hybrid E/M states can arise. To our knowledge, this P4-induced stabilization of hybrid AEC phenotypes has not been previously reported.

The aim of our work was therefore to identify, expand, and functionally characterize this intermediate state as a therapeutically advantageous population for regenerative medicine. In particular, we investigated how P4 modulates EMP, how hybrid E/M cells can be selectively enriched, and how this phenotype enhances key functional properties, including immunomodulation, plasticity, collective migration, and tissue repair capacity.

In this regard, our findings help clarify the mechanistic basis of P4 action in mid-gestation AECs. Previous studies showed that nuclear PR activation suppresses EMT-inducing transcription factors (SNAIL, TWIST, ZEB) and reduces TGF-β secretion and SMAD2/3 phosphorylation, thereby limiting the autologous EMT program characteristic of AECs.9^,^13 To further dissect this mechanism, we inhibited either the nuclear progesterone receptor (RU-486) or the membrane-associated receptor PGRMC1 (AG-205). Blocking either receptor reduced the effects of P4 and allowed cells to complete EMT and acquire a M phenotype, indicating that both signaling pathways contribute to delaying EMT progression and enabling the emergence of hybrid states. These findings support a model in which nuclear PR and membrane PGRMC1 jointly modulate EMP in AECs. However, the relative contribution and potential crosstalk between these receptors remain unclear, and dedicated mechanistic studies will be required to fully elucidate how their coordinated signaling stabilizes hybrid E/M phenotypes.

Hybrid E/M states are typically stabilized by PSFs,20 a class of proteins that buffer cells from transitioning to fully M phenotypes. In this study, we observed Nrf2 as a key regulatory factor in AEC-derived hybrid cells. Its elevated expression appears essential for maintaining hybrid identity and preventing full EMT. This observation is consistent with several mathematical models and experimental studies conducted on cancer cells.21^,^22 Importantly, Nrf2 also activates the Notch signaling pathway, upregulating Jagged1, which promotes leader cell formation at the invasive front and collective migration.21^,^22 Although we observed Jagged1 upregulation in hybrid cells, it continues to increase in fully M cells, thus suggesting additional roles. Notably, in the amniotic membrane, Nrf2 is also critical to preserve cellular homeostasis by enhancing antioxidant response to reduce oxidative stress and inflammatory damage to the cell during pregnancy.11^,^23

Surface marker analysis further confirmed the hybrid identity of P4-treated AECs. Although CD51, CD61, and CD106 were first associated with hybrid EMT states in carcinoma models,24 these adhesion-related molecules participate broadly in integrin signaling and ECM remodeling, processes that are also active during EMT in AECs.13^,^25 For this reason, they can be appropriate indicators of intermediate adhesion states arising during EMP. In our study, their coordinated upregulation specifically at passage 2, corresponding to peak E/M co-expression, and their subsequent decline as cells adopted a fully M identity, further support their suitability as markers of AEC-derived hybrid states. Interestingly, only CD51 persisted in mAECs, potentially reflecting its role in TGFβ signaling, which is highly active in AECs and contributes to EMT progression.9^,^26^,^27

Unlike the patterns described by Pastushenko et al. (2019),28 EpCAM remained strongly expressed in hyAECs, suggesting a possible link to stemness maintenance. EpCAM is known to be co-expressed with M markers in several cancer cells29^,^30^,^31 and its level and function is context-dependent during EMP and metastasis.32^,^33^,^34 Moreover, EpCAM contributes to pluripotency by regulating the transcription factors such as Oct4, Sox2, and Nanog.35^,^36^,^37^,^38^,^39^,^40 Consistently, we observed transient upregulation of Oct4 and Nanog in hyAECs, which declined as cells completed EMT. Sox2 expression, by contrast, progressively increased, peaking in mAECs, possibly indicating a broader role in maintaining M stemness. Notably, Oct4 and Nanog overexpression has been shown to sustain pluripotency in stem cells, often acting synergistically, whereas their downregulation is closely associated with loss of pluripotency and the induction of differentiation.41^,^42^,^43^,^44^,^45

The functional relevance of this stemness-associated transcriptional profile was validated through tendon differentiation assays using PLGA fleeces.46^,^47 In this regard, hyAECs exhibited increased expression of tenogenic markers, such as TNMD and COL1, outperforming mAECs. These results support the hypothesis that the hybrid state represents a plastic and responsive window, particularly receptive to differentiation cues.48 Notably, previous studies have shown that the tendon differentiation of AECs involves EMT activation both in vitro49^,^50 and in vivo.18 Here, we further demonstrate that the transplantation of hyAECs leads to improved tissue regeneration, expanding upon previous evidence obtained with E cells at later stages of regeneration. Our findings indicate that cells maintaining a hybrid state possess enhanced plasticity compared to fully M cells.

In addition to their regenerative potential, hyAECs exerted stronger local immunomodulatory effects than mAECs. Upon transplantation, hyAECs promoted macrophage polarization toward the anti-inflammatory CD206^+^ M2 phenotype and reduced CD86^+^ pro-inflammatory cells. Paracrine signaling appears to be the predominant mechanism by which AECs modulate macrophage behavior. Our previous work showed that AEC-conditioned medium alone is sufficient to induce an M2-like phenotype in THP-1 macrophages,51 indicating that soluble factors can drive polarization independently of direct cell contact. Findings from human AEC studies support this view, as both transplantation experiments and hAEC-conditioned medium have been shown to promote M2 polarization.52 However, contact-dependent interactions may still enhance these effects, as direct co-culture systems often produce stronger macrophage modulation than transwell settings that prevent physical contact.53 Thus, while paracrine signaling predominates, a complementary contribution of contact-mediated mechanisms within the tissue microenvironment cannot be excluded.

This immunoregulatory effect was confirmed either in vitro, where conditioned media from hyAECs more effectively suppressed PBMCs proliferation following LPS stimulation. Importantly, while hybrid E/M states in tumors have been associated with immune evasion mechanisms linked to stemness,54^,^55 in our physiological, non-tumorigenic model, these immunosuppressive properties primarily reflect the intrinsic role of AECs in maintaining fetal-maternal immune tolerance during pregnancy. Nevertheless, we cannot exclude that regulatory mechanisms analogous to those described in cancer settings may partially contribute to AEC immune modulation. Moreover, the collective migration of hyAECs may significantly enhance their coordination during tissue regeneration.

Despite these promising findings, the identification and characterization of hybrid E/M cell states remains a significant challenge. A major limitation lies in the a priori selection of marker sets, which inevitably introduces bias into both identification and interpretation. Specifically, the number of hybrid phenotypes detected is largely dictated by the combinations of chosen markers, potentially skewing the analytical approach. A central question persists: How many hybrid E/M states exist? Theoretically, EMP represents a continuum rather than a series of discrete transitions, implying a virtually infinite spectrum of intermediate states. To address this complexity, we propose a framework based on the relative dominance of E versus M traits, rather than predefined hybrid subtypes. This spectrum-based approach more accurately captures the phenotypic dynamics that unfold during the transition from predominantly E−to predominantly M-like states and offers a more refined, less biased model for describing the complexity of EMP.

In conclusion, this study provides a comprehensive characterization of hyAECs and demonstrates their enhanced tissue regeneration and immunomodulatory potential. These findings shed light on physiological EMP regulation and highlight hybrid states as promising targets for next-generation cell therapies in regenerative medicine.

Limitations of the study

Although this study provides important insights into hybrid E/M states in AECs, limitations related to EMP classification should be acknowledged. EMP is a dynamic continuum, and population-level analyses may overlook cellular heterogeneity and transient intermediate states. In this work, hyAECs were defined based on the co-expression of selected E and M markers at the population level, which enabled functional characterization but does not fully capture the spectrum of hybrid phenotypes. Future single-cell approaches will be required to resolve EMP complexity and refine hybrid state classification.

Moreover, all experiments were performed using ovine mid-gestation AECs, and the findings are therefore restricted to this species and developmental context; nuclear progesterone receptor expression at comparable stages has not been reported in human amniotic cells, limiting direct extrapolation to humans. In addition, although membrane progesterone signaling was implicated, the relative contribution of different membrane progesterone receptors was not dissected, as PGRMC2 was not directly assessed. Finally, the exclusive use of animal-derived cells might limit the translational relevance of these findings and thus require future validation in human AEC-based models.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Verdiana Di Giulio (vdigiulio@unite.it).

Materials availability

This study did not generate new unique reagents or materials.

Data and code availability

• •All data reported in this article will be shared by the lead contact upon request.
• •This study does not report original code.
• •Any additional information required to reanalyze the data reported in this work article is available from the lead contact upon request.

STAR★Methods

Key resources table

REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesAnti-CD326 (EpCAM), PEAncellCat# 126-050; RRID: AB_29211773Anti-CD324 (E-Cadherin), FITCBioLegendCat# 324103; RRID: AB_3591549Anti-CD106 (VCAM-1), FITCAncellCat# 327-040; RRID: AB_3738382Anti-CD51, PEInvitrogenCat# A15458; RRID: AB_2534471Anti-CD61, FITCInvitrogenCat# 11-0619-42; RRID: AB_10667773Anti-CD73, PEBD BiosciencesCat# 550257; RRID: AB_393561Anti-CD90, PEBD BiosciencesCat# 555596; RRID: AB_395970Anti-CDH1/E-CadherinLSBioCat# LS-C204222; RRID: AB_3738381Anti-Jagged1Cell SignalingCat# 70109T; RRID: AB_2799774Anti-Nrf2Novus BiologicalsCat# NBP1-32822; RRID: AB_10003994Anti-α-SMAAbcamCat# ab7817; RRID: AB_262054Anti-VimentinDakoCat# M0725; RRID: AB_10013485Anti-TubulinCell SignalingCat# 3873; RRID: AB_1904178Anti-Oct4AbcamCat# ab18976; RRID: AB_444714Anti-Sox2AbcamCat# ab59776; RRID: AB_945584Anti-NanogSigma-AldrichCat# AB9220; RRID: AB_11213156Anti-TenomodulinAbcamCat# ab203676; RRID: AB_2722782Anti-Collagen IAbcamCat# ab292; RRID: AB_303415Von Willebrand FactorDakoCat# A0082; RRID: AB_2315602Anti-CD206RD SystemCat# AF2534; RRID: AB_2063019Anti-CD86Novus BiologicalsCat# NBP2-25208; RRID: AB_2923115Anti-rabbit Alexa Fluor 488AbcamCat# ab150077; RRID: AB_2630356Anti-mouse Cy3Sigma-AldrichCat# AP132C; RRID: AB_258785Goat anti-Mouse IgG Alexa Fluor 488InvitrogenCat# A28175; RRID: AB_2536161Goat anti-Rabbit IgG Alexa Fluor 488InvitrogenCat# A-11008; RRID: AB_143165ChemicalsTrypsin-EDTASigma-AldrichCat# T4049PBSSigma-AldrichCat# D8662FBSGibcoCat# 10270106α-MEMEurocloneCat# ECM0850LL-GlutamineEurocloneCat# ECB300DAmphotericin BEurocloneCat# EUM0009DPenicillin-StreptomycinLonzaCat# DE17-602EProgesterone (P4)Sigma-AldrichCat# P8783Mifepristone (RU-486)Sigma-AldrichCat# M8046AG-205Sigma-AldrichCat# A1487Ficoll-Paque PLUSCytivaCat# GE17-1440-02LPSSigma-AldrichCat# L2637PHA-LInvitrogenCat# 00-4977-93RIPA bufferSigma-AldrichCat# R0278Phosphatase InhibitorSERVACat# 39055Protease Inhibitor CocktailSigma-AldrichCat# P2714Nitrocellulose membranesBio-RadCat# 1620145EveryBlot Blocking BufferBio-RadCat# 12010020Clarity Max ECL SubstrateBio-RadCat# 1705062Trans-Blot Turbo 5x Transfer BufferBio-RadCat# 10026938Casein Blocker 1X TBSBio-RadCat# 1610782DAPISigma-AldrichCat# D9542FluoromountSigma-AldrichCat# F4680Triton X-100Sigma-AldrichCat# T8787Tween 20Carlo ErbaCat# 600481BSASigma-AldrichCat# A3059AssaysBradford 1x Dye ReagentBio-RadCat# 5000205eFluor™ 450 Cell Proliferation DyeInvitrogenCat# 65-0842Total RNA Purification KitNorgen BiotekCat# 17200OligodT PrimersBiolineCat# BIO-38029Tetro Reverse TranscriptaseBiolineCat# BIO-65050SYBR Lo-ROX kitBiolineCat# BIO-94050SoftwareCytExpert SRTBeckman Coulterv1.2.10004ImageJ 1.53kNIHv1.53kGraphPad Prism 10GraphPadVersion 10NIS-Element software 4.40Nikonv4.40Experimental modelsAECsUniversity of Teramo–PBMCsUniversity of Teramo–Ovine Achilles TendonUniversity of Bari–

Experimental model and study participant details

Ovine Achilles tendon injury model

All procedures involving animals were approved by the Italian Ministry of Health (authorization n° 771/2023-PR) and conducted in strict accordance with the recommendations of the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the ARRIVE guidelines. Eighteen female Lacaune sheep (2 years of age, ∼45 kg, not inbred) were used in this study. Animals were housed in group pens with ad libitum access to food and water, and their health status was regularly monitored under veterinary supervision. To ensure hormonal synchronization, all animals received progesterone-impregnated vaginal sponges (Chronogest® CR 20 mg, controlled release vaginal sponge for sheep, Chronogest CR Intervet International B.V., Boxmeer, The Netherlands) one week before and after surgery. Prior to surgical procedures, each animal underwent a comprehensive clinical and hematological evaluation to confirm overall health status. Radiographic and ultrasound examinations (Samsung HM70EVO Probe LA3-16AD; Samsung Medison Co., LTD. Bari, Italy) were performed to exclude pre-existing pathologies in the right hindlimb and the common calcaneal tendon. Body weight was recorded before surgery and monitored throughout the experimental period. Surgeries were performed under aseptic conditions and combined general and spinal anesthesia. Sedation was achieved by intravenous administration of 0.4 mg/kg diazepam (Midazolam-hameln 5 mg/ml, AIC 035325012, Hameln Pharma Plus GmbH, Hameln, Germany) and buprenorphine (10 μg/kg) (Bupaq Multidose 0,3 mg/ml, AIC 104356011, VetViva Richter GmbH, Wels, Austria) by using sterile syringes (2102691001M, Farmatex Softouch Med’s, FAMRAC-ZABBAN, Bologna, Italy). Propofol (Propofol Fresenius 10 mg/mL, AIC 036849065, Fresenius Kabi Deutschland GmbH, Bad Homburg vor der Höhe, Germany) was administered intravenously if necessary to maintain sedation. Spinal anesthesia was performed using 1 mL/10 kg of 2% lidocaine (Lidor 20 mg/ml, AIC 105118020, VetViva Richter GmbH, Wels, Austria) and 20G needles (405253; GIMA BD QUINCKE 20G, GIMA, Milan, Italy). A tendon injury was surgically created in the right common calcaneal tendon. Following a 4 cm skin incision above the calcaneal insertion, the subcutaneous tissue was bluntly dissected to expose the tendon. The epitenon was carefully opened to visualize the two components of the tendon (triceps surae and superficial digital flexor). A 3 mm biopsy punch (IMS-268C; BLife srl, Treviso, Italy) was used to create a partial-thickness circular lesion in the superficial digital flexor component, generating a pocket that was filled with 5 × 10^6^ PKH26-labeled AECs (hyAECs or mAECs) suspended in fibrin glue (Tisseel, #5500387; Baxter, Baxter Healthcare Ltd., Newbury, UK). The epitenon was closed over the lesion with 5-0 absorbable sutures (#748676; Biolong, BIOMED, Sanifarm Bolzano GmbH, Bolzano, Italy). Subcutaneous tissues and skin were closed in layers using standard surgical technique. Post-operative antibiotic prophylaxis consisted of intramuscular penicillin and streptomycin (Repen®, AIC 101775029, Fatro Industria Farmaceutica Veterinaria S.p.A., Ozzano dell'Emilia, Bolzano, Italy) for five days. Group allocation was randomized. All animals were monitored daily for signs of discomfort, including food intake, weight loss, fecal and urinary output, rectal temperature, and behavioral changes. Animals were euthanized on day 7 and 14 post-surgery, and tendons were harvested for histological and molecular analyses.

Ovine amniotic epithelial cells (AECs)

AECs were isolated from the amniotic membranes of three mid-gestation ovine fetuses: two male and one female. No ethical approval was required, as amniotic membranes were obtained as waste material from animals slaughtered for food production. Fetal sex was determined macroscopically. Cell authentication and mycoplasma testing were not performed.

Method details

Experimental design, randomization, and blinding

Animals were randomly assigned to treatment groups using a randomized block design to ensure even distribution across evaluation timepoints. Investigators performing imaging and histological analyses were blinded to the experimental group allocation.

Replicates and data inclusion

All experiments included a minimum of three biological replicates per condition. No samples or data points were excluded from analysis unless compromised due to technical failure (e.g., tissue degradation, loss of sample during processing).

Isolation and culture of ovine amniotic-derived AEC

AEC were isolated from amniotic membranes (AM) collected from mid-gestation pregnant Appenninica ewes (fetal length: 25–30 cm), as previously described. The uterus wall was opened under sterile conditions, and the AM was mechanically separated from the chorion using a stereomicroscope and then cut into 2–3 cm fragments. AM pieces were washed in Dulbecco’s PBS (DPBS; #D8662; Sigma-Aldrich, St. Louis, MO, USA) and incubated under gentle agitation with 0.25% Trypsin-EDTA (#T4049; Sigma-Aldrich, St. Louis, MO, USA) at 38.5 °C for 40 min. Enzymatic digestion was stopped by adding 10% FBS (#10270106; Gibco, Thermo Fisher Scientific, Waltham, MA, USA), and the cell suspension was filtered through a 40 μm mesh. Cells were pelleted by centrifugation at 500 × g for 10 min, and viable cells were counted using the LUNA-II™ Automated Cell Counter (Logos Biosystems Inc., Gyeonggi-do, Korea) with trypan blue staining. AEC were then seeded at 10^4^ cells/well in 6-well plates in α-MEM (#ECM0850L; Euroclone S.p.A., Milan, Italy) supplemented with 10% inactivated FBS (#10270106; Gibco, Thermo Fisher Scientific, Waltham, MA, USA), 1% L-Glutamine (#ECB300D; Euroclone S.p.A., Milan, Italy), 1% Amphotericin B (#EUM0009D; Euroclone S.p.A., Milan, Italy), and 1% penicillin/streptomycin (#DE17-602E; Lonza, Basel, Switzerland), referred to as complete medium, with or without 25 μM Progesterone (P4; 4-pregnene-3,20-dione, #P8783; Sigma-Aldrich, St. Louis, MO, USA) [Ref. 9 main text], and with 25 μM of Mifepristone (Sigma-Aldrich, Cat# M8046) or 5 μM and 10 μM of AG-205 (Sigma-Aldrich, Cat# A1487). Cells were cultured at 38.5 °C with 5% CO_2_ until ∼70% confluence, then detached using 0.05% Trypsin-EDTA for subsequent experiments.

AECs staining with the red fluorescent cell linker PKH26

Five million of cells (hyAECs and mAECs) per animal were thawed and stained with PKH26 Red Fluorescent Cell Linker Midi Kit for General Cell Membrane Labeling (#MIDI26; Sigma-Aldrich, St. Louis, MO, USA). PKH26 dye stably incorporates into lipid regions of the cell membrane. After thawing, cells were re-suspended into 1 ml of Diluents C and then added at 1ml of Dye Solution containing 4 μl of PKH26. The cellular suspension was incubated for 5 min at room temperature with periodic mixing. Cell staining was stopped with 2 ml of DPBS supplemented with 1% BSA (#A3059; Sigma-Aldrich, St. Louis, MO, USA) for 1 min and finally centrifuged at 400 x g for 10 min. Cells were suspended in completed growth medium and diluted 1:1 (v/v) in fibrin glue (Tisseel, #5500387; Baxter, Baxter Healthcare Ltd., Newbury, UK).

Conditioned medium (CM) production

As previously described for AEC, cells were starved in serum free medium for 4h and then treated with LPS (1 μg/mL; #L2637, Sigma-Aldrich, St. Louis, MO, USA) for 1 h. After treatment, all experimental groups (control and LPS-treated) were washed twice with serum-free medium and cultured for an additional 24 h in fresh serum-free medium. Conditioned medium (CM) was harvested by centrifugation at 500 × g for 10 min, and supernatants were stored at –80 °C until used in immunological assay.

Flow cytometry

For surface marker characterization, anti-human CD326 (EpCAM)/R-PE, clone ANC8D4 (1:50; #126-050, Ancell, Bayport, MN 55003-0087 USA), FITC anti-human CD324 (E-Cadherin) Antibody, clone 67A4 (1:50; #324103; BioLegend Way, San Diego, CA 92121 USA), anti-human CD106(VCAM-1)/FITC, clone 1.G11B1 (1:50; #327-040, Ancell, Bayport, MN 55003-0087 USA), CD51 Monoclonal Antibody, PE, clone 13C2 (1:50; #A15458, Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA), CD61 (Integrin beta 3) Monoclonal Antibody, FITC, eBioscience™, clone VI-PL2 (1:50; #11-0619-42, Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA), BD Pharmingen™ PE Mouse Anti-Human CD73, clone AD2 (1:50; #550257, BD Biosciences, San Jose, CA 95131, USA), BD Pharmingen™ PE Mouse Anti-Human CD90, clone 5E10 (1:50; #555596, BD Biosciences, San Jose, CA 95131, USA) were employed. Samples were thoroughly washed with Dulbecco’s PBS (DPBS; #D8662; Sigma-Aldrich, St. Louis, MO, USA) and resuspended in DPBS supplemented with 0.5% BSA (#A3059; Sigma-Aldrich, St. Louis, MO, USA). Flow cytometry analyses were performed using a CytoFLEX SRT system (Beckman Coulter, Brea, CA, USA), acquiring 10,000 events per sample. Excitation was performed using 405 nm, 488 nm, and 561 nm lasers, detecting fluorescence in the V450, B525, Y585, and Y780 channels. Fluorescence intensity was represented on a standard logarithmic scale. Data was analyzed and visualized with CytExpert SRT software (version 1.2.10004; Beckman Coulter, Brea, CA, USA).

Western blotting

Total protein was extracted from each sample using RIPA lysis buffer (#R0278; Sigma-Aldrich, St. Louis, MO, USA) supplemented with Phosphatase Inhibitor (#39055; SERVA Electrophoresis GmbH, Heidelberg, Germany) and Protease Inhibitor Cocktail (#P2714; Sigma-Aldrich, St. Louis, MO, USA), diluted according to the manufacturer’s instructions. Samples were incubated on ice for 30 minutes and centrifuged at 12,000 × g for 12 minutes at 4 °C. The supernatant was collected, and 5 μL was used to determine protein concentration using Quick Start™ Bradford 1x Dye Reagent (#5000205; Bio-Rad Laboratories, Milan, Italy). Proteins were separated on 4–15% gradient precast gels (#4568083; Mini-PROTEAN®, Bio-Rad Laboratories, Milan, Italy) and transferred onto nitrocellulose membranes (#1620145; Bio-Rad Laboratories, Milan, Italy) using the Trans-Blot® Turbo Transfer System and 5× Transfer Buffer (#10026938; Bio-Rad Laboratories, Milan, Italy). Membranes were blocked with EveryBlot™ Blocking Buffer (#12010020; Bio-Rad Laboratories, Milan, Italy) for 5 minutes. Primary antibodies against Polyclonal Goat anti-Human CDH1/E Cadherin Antibody (aa750-850, WB) (1:300; LS-C204222; LSBio, Newark, CA 94560 USA), Jagged1 (D4Y1R) XP® Rabbit mAb (1:300; #70109T; Cell Signaling technology, Danvers, MA 01923, USA), Nrf2 (1:300; NBP1-32822; Novus Biologicals, Centennial, CO 80112 USA), Anti-alpha smooth muscle Actin, clone 1A4 (1:300; ab7817, Abcam, Cambridge, UK), Vimentin, clone V9 (1:300; M0725; Dako, Santa Clara, CA 95051, USA), α-Tubulin, clone DM1A (1:300; #3873; Cell Signaling technology, Danvers, MA 01923, USA), Oct4 (1:300; ab18976; Abcam, Cambridge, UK), Sox2 (1:300; ab59776; Abcam, Cambridge, UK), Nanog (1:300; AB9220; Sigma-Aldrich, St. Louis, MO, USA), Tenomodulin (1:300; ab203676; Abcam, Cambridge, UK), Collagen I (1:300; ab292; Abcam, Cambridge, UK), CD86, BU63 (1:300; MCA1118SBV610; Bio-Rad Laboratories, Milan, Italy), MMR/CD206 (1:300; AF2535; R&D Systems, Minneapolis, MN 55413, USA) were diluted in 1× TBS containing 1% Casein Blocker (#1610782; Bio-Rad Laboratories, Milan, Italy) and incubated overnight at 4 °C. The HRP-conjugated anti-goat (1:5000; #SC-2354; Santa Cruz Biotechnology, Inc., Dallas, TX, USA), HRP-conjugated anti-mouse (1:10000; #NA931; Cytiva, Marlborough, MA, USA NA931), and HRP-conjugated anti-rabbit (1:10000; #NA934; Cytiva, Marlborough, MA, USA NA931) secondary antibodies were diluted in the same buffer and incubated for 1 hour at room temperature. Detection was performed using Clarity Max™ ECL Substrate (#1705062; Bio-Rad Laboratories, Milan, Italy), and signals were acquired with the ChemiDoc™ MP Imaging System (Bio-Rad Laboratories, Milan, Italy). Densitometric analysis was carried out using ImageJ software (v1.53k; National Institutes of Health, Bethesda, MD, USA), and protein expression was normalized to total protein content using Stain-Free™ technology (Bio-Rad Laboratories, Milan, Italy).

Immunocytochemistry (ICC)

HyAECs and mAECs were fixed in 4% paraformaldehyde for 15 min and then were permeabilized with PBS/Triton X-100 0.1% (#T8787; Sigma-Aldrich, St. Louis, MO, USA) for 10 minutes at room temperature (RT).

Harvested tendon samples were collected post-intervention and immediately snap frozen in liquid nitrogen, and 7 μm thick cryosections were prepared for analyses.

Following non-specific blocking with PBS/BSA 1% (#P3813; Sigma-Aldrich, St. Louis, MO, USA)-Tween 20 0.05% (#600481; CARLO ERBA Reagents S.r.l., Milan, Italy) for 1h at RT, sections were incubated overnight at 4°C with the following primary antibodies: Tenomodulin (1:200; ab203676; Abcam, Cambridge, UK), Collagen I (1:200; ab292; Abcam, Cambridge, UK), CD86, BU63 (1:200; MCA1118SBV610; Bio-Rad Laboratories, Milan, Italy), MMR/CD206 (1:200; AF2535; Minneapolis, MN 55413, USA), Von Willebrand Factor (1;200; #A0082, Dako, Santa Clara, CA 95051, USA). Anti-rabbit Alexa Fluor 488 (ab150077; Abcam, Cambridge, UK), Anti-mouse Cy3 (AP132C; Sigma-Aldrich, St. Louis, MO, USA), Goat anti-Mouse IgG (H+L) Superclonal™ Alexa Fluor 488 (A28175; Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA), Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Alex Fluor 488 (A-11008; Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) secondary antibodies, diluted 1:200 in PBS/BSA 1%-Tween 20 0.05% for 1h at RT was used to retrieve the antigen. The staining of the nuclei was performed with DAPI (diluted 1:2000 in PBS; #D9542; Sigma-Aldrich, St. Louis, MO, USA) for 10 minutes at RT. Then coverslips were mounted with Fluoromount (F4680; Sigma-Aldrich, St. Louis, MO, USA). The image acquisition was performed under a Nikon Ar1 laser confocal scanning microscope (Nikon, Dusseldorf, Germany) equipped with the NIS-Element software 4.40 (Nikon, Dü sseldorf, Germany).

RNA extraction

Total RNA from AECs was extracted using the Total RNA Purification Kit (#17200; Norgen Biotek Corp., Thorold, ON, Canada), following the manufacturer’s instructions. Subsequently, cDNA synthesis was performed starting from 1 μg of total RNA per sample using oligodT Primers (#BIO-38029; Bioline, Cincinnati, OH, USA) and Tetro Reverse Transcriptase (#BIO-65050; Bioline, Cincinnati, OH, USA), according to the manufacturer’s protocols.

Real-time qPCR

Real-Time qPCR was performed in triplicate using the SensiFAST SYBR Lo-ROX kit (#BIO-94050; Bioline, Cincinnati, OH, USA) on a QuantStudio™ 3 system (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA), following the manufacturer’s protocol. The thermal profile included an initial denaturation at 95 °C for 3 min, followed by 40 cycles at 95 °C for 5 s and 60 °C for 30 s. Primers (0.5 pmol/μL) were designed using NCBI Primer-BLAST and validated via melting curve analysis.

• •COL1 forward primer: 5’-CGTGATCTGCGACGAACTTAA-3’ (Tm = 63.4°C)
• •COL1 reverse primer: 5’-GTCCAGGAAGTCCAGGTTGT-3’
• •TNMD forward primer: 5’-TGGTGAAGACCTTCACTTTCC-3’ (Tm = 63.1°C)
• •TNMD reverse primer: 5’-TTAAACCCTCCCCAGCATGC-3’
• •SCX forward primer: 5’-AACAGCGTGAACACGGCTTTC-3’ (Tm = 65.6°C)
• •SCX reverse primer: 5’-TTTCTCTGGTTGCTGAGGCAG-3’
• •GAPDH forward primer: 5’-CCTGCACCACCAACTGCTTG-3’ (Tm = 64.9°C)
• •GAPDH reverse primer: 5’-TTGAGCTCAGGGATGACCTTG-3’

PLGA scaffold fabrication via electrospinning

Poly(lactide-co-glycolide) (PLGA, PLG8523, Purac Corbion, Gorinchem, The Netherlands) scaffolds with highly aligned fibers were fabricated through electrospinning using a commercial E-Spintronic electrospinning apparatus with climate control machine (Erich Huber, Gerlinden, Germany), as previously described (Ref. main text 14,15-46,47). PLGA solution (12% w/w) was prepared by dissolving the polymer in hexafluoro-2-propanol (HFIP, PC4750, Apollo Scientific Ltd, Manchester, UK) under magnetic stirring overnight. The electrospinning setup consisted of a 3 mL syringe connected to a polytetrafluoroethylene tube (PTFE, Intra Special Catheters, Rehlingen-Siersburg, Germany) and a stainless-steel needle (10438881, B Braun™ 466564/3, Fisher Scientific GmbH, Wien, Austria). The electrospinning process was carried out at 22.5°C with air humidity of approximately 65%. Process parameters included voltage of 33 kV and flow rate of 0.25 mL/h. Fibers were collected on baking paper using a rotating drum collector at 1000 rpm rotational speed. The resulting PLGA fleeces were obtained by electrospinning 250 μL of PLGA solution for each scaffold.

Physical characterization of the electrospun PLGA scaffolds was previously performed using scanning electron microscopy (SEM, Carl Zeiss AG, Jena, Germany), which confirmed the formation of highly aligned fiber architecture. Fiber morphology and orientation analysis revealed a homogeneous fibrous structure with consistent diameter distribution and minimal bead formation, confirming the reproducibility of the fabrication process across multiple batches (Ref. main text 14,15-46,47).

Tenogenic differentiation potential

To assess in vitro tenogenic differentiation ability of both hyAECs and mAECs, a validated in vitro approach using the produced poly(lactide-co-glycolide (PLGA) fleeces scaffold with highly aligned (ha) fibers was used (Ref main text 14). In detail, hyAECs and mAECs were seeded on validated teno-inductive PLGA fleeces scaffolds (Ref. main text 14,15-46,47), at 0.05x106 cells/fleece scaffold density and incubated for 48h at 38.5°C with 5% CO2. Before cell seeding, the ha-PLGA fleeces scaffolds were cut into rectangular pieces of 10mm×7mm each, sterilized with 70% ethanol and pre-incubated for 1 h at 38.5◦C and 5% CO2 in cell culture growth medium composed of α-MEM supplemented with 10% FBS, 1% L-glutamine, 1% amphotericin, and 1% penicillin/streptomycin. hyAECs and mAECs seeded on cover slides and incubated at the same conditions were used as control.

Isolation of ovine PBMCs

Ovine PBMCs were isolated by density gradient centrifugation from 16 mL of fresh peripheral blood collected at the slaughterhouse. The gradient was prepared using 12 mL Ficoll-Paque™ PLUS (#GE17-1440-02; Cytiva, Marlborough, MA, USA), following the manufacturer’s protocol. Isolated cells were cryopreserved in liquid nitrogen until used in immunological assays.

PBMCs activation test

To compare the immunomodulatory activity of hyAECs and mAECs, we evaluated the proliferation of PHA-activated PBMCs exposed to conditioned media (CM) from cells cultured under basal or LPS-stimulated conditions. PBMCs (2 × 10^5^ cells/well) were seeded in 96-well plates, stimulated with 10 μg/mL PHA-L (#00-4977-93; Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA), and cultured for 48 h in the presence of the various CM conditions. Cell proliferation was assessed using eBioscience™ Cell Proliferation Dye eFluor™ 450 (10 μM; #65-0842, Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA), following the manufacturer’s instructions. Data was normalized to PHA-stimulated controls.

Wound healing assay

To assess the cell migration property of AECs according to their phenotype, a wound healing assay was performed. In detail, cells were seeded at a density of 0.05x10^6^ cells/well in 12-well cell culture plates and grown until full confluence. Cells were then starved for 4 hours with serum-free medium and then washed two times in Dulbecco’s PBS (DPBS; #D8662; Sigma-Aldrich, St. Louis, MO, USA). Subsequently cell monolayer was scratched using a p200 pipette tip. A third PBS wash was done to remove detached cells and cell debris. Fresh serum-free medium was then added to each well and were monitored at different time points (0h, 24h, 48h, 72h). Images were taken using Nikon Eclipse Ti in time-lapse and quantified using the ImageJ software (ImageJ 1.53k, NIH, Bethesda, MD, USA).

Quantification and statistical analysis

A minimum of three biological replicates were used for AEC (fetal) and PBMC (animal) samples to assess inter-experimental variability. Each experiment was repeated three times to evaluate intra-experimental consistency. Sample sizes (n) and statistical significance are reported in the figure legends. Data is shown as mean ± S.D. Normality was tested using the D'Agostino and Pearson test. One-way ANOVA and appropriate unpaired t-tests were applied to normally distributed data, followed by Tukey’s post hoc test (GraphPad Prism 10, San Diego, CA, USA). Significance was set at p ≤ 0.05.

Acknowledgments

This project has received funding from the European Union-Next Generation EU, project code: ECS00000041, project CUP: C43C22000380007, project title: Innovation, digitalization and sustainability for the diffused economy in Central Italy-VITALITY. This project has received funding from the European Union-Next Generation EU: Mission 4 “Education and Research”-Component C2 Investment 1.1, “Fund for the National Program of Research and Projects of Significant National Interest (PRIN), funding number: 2022YXHEET.” This project has received funding from the European Union-Next Generation EU, M4C1, project CUP: C46E23000120006, Fellowship code: 39-411-A8-DOT13A8025-8428.

Author contributions

Conceptualization: A.C., Ant. C., V.R., M.M., and B.B.; methodology: A.C, A.M.C., and V.D.G.; formal analysis: A.C., V.D.G., G.P., M.E.K., L.P., M.G., O.D.G., and M.T.; investigation: A.C., V.D.G., G.P., M.E.K., L.P., M.G., O.D.G., and M.T.; writing – original draft: A.C., A.M.C., V.D.G., M.E.K., and L.P.; writing – review and editing: A.C., A.M.C., V.D.G., G.P., M.E.K., M.G., A.M., V.R., L.P., M.M., and B.B.; project administration: A.C., Ant. C., V.D.G., and B.B.; funding acquisition: B.B.

Declaration of interests

The authors declare no competing interests.