Decidualised endometrial stromal cell-derived extracellular vesicles induce bystander decidualisation and cAMP-mediated attenuation of natural killer cell cytotoxicity

Dear Editor,

Decidualisation, the process of differentiation of endometrial stromal cells (EnSCs) into secretory decidual cells, is fundamental to blastocyst implantation and endometrial immune modulation. Defective decidualisation has been closely linked to implantation failure and miscarriage. Decidualisation triggers a metabolic and immunomodulatory shift in EnSCs, enabling them to regulate uterine natural killer (NK) cells and T cells.¹

Understanding how decidual cell signalling influences neighbouring uterine cells is critical for elucidating key adaptations in early pregnancy, particularly maternal‒foetal crosstalk. A growing body of evidence highlights extracellular vesicles (EVs) as a novel axis of cell to cell communication, playing a pivotal role in tissue homeostasis and immune regulation.² Here, we sought metabolic reprograming of EnSCs during decidualisation and unrevealed new aspects of endometrial EVs by showing that EVs from decidualised endometrial stromal cells (D-EnSCs-EVs) induce decidualisation in neighbouring cells and modulate NK cell function.

Here, EnSCs were isolated from luteal-phase endometrial biopsies and characterised (Figure S1). Metabolome analysis of isolated cells demonstrated that decidualisation significantly alters the amino acid metabolome of EnSCs, effectively distinguishing undecidualised (uD) from decidualised (D) EnSCs by days 4 and 6 of decidualisation (Figure 1A,B). Cluster analysis revealed greater similarity in metabolomic profiles of late-stage D-EnSCs compared to earlier time points (Figure 1C). Metabolite analysis revealed methionine (Met) and phenylalanine (Phe) as key discriminators between uD-EnSCs and D-EnSCs (Figure 1D), and showed a coordinated shift in amino acid metabolism that may underpin the functional transformation of EnSCs during decidualisation (days 2–6; Figure 1E,F). Decidualised stromal cells are known to support decidual NK (dNK) cell generation from peripheral blood NK cells via secretion of transforming growth factor beta (TGF-β), interleukin (IL)-1β and IL-15.¹ Our finding that decidualisation induces methionine production provides a novel metabolic underpinning for this process, aligning with methionine's established epigenetic role in regulating IL-5 transcription and promoting endometrial receptivity.³ Based on GLUT1 expression, it is thought that decidualisation relies on glucose metabolism.⁴ Our findings, however, demonstrated that the process was characterised by a reduction in glucose consumption and lactate production (Figure 1G), consistent with lower proliferation capacity of D-EnSCs compared to uD-EnSCs (Figure S2). Notably, high cyclic adenosine monophosphate (cAMP) concentration (.5 mM, as used in this study) suppress GLUT1 expression.⁵ Furthermore, decidualisation led to sustained pro-inflammatory cytokines IL-6 and IL-8 secretion (Figure 1H), inversely correlated with the cells’ decidualisation capacity and prolactin secretion (Figure S2). These data align with the previous concept that decidualisation requires a transient inflammatory signal but is compromised by a prolonged inflammatory response.⁶

To explore the potential role of EnSC-derived EVs on induction of decidualisation, EVs were isolated (Figure S3 and Table S1) from D-EnSCs and uD-EnSCs and characterised (Figure 2A,B). Decidualisation significantly increased EV secretion (Figure 2C). Subsequent uptake kinetics assays showed a cell-type-specific pattern. EnSCs internalised EVs rapidly (within 4 h; Figure 2D,E) in contrast to NK cells, which displayed a slower uptake profile, requiring 24 h for significant incorporation (Figure 2F). This temporal difference may reflect variations in size-dependent endocytic activity between the two cell types.

To extend prior findings⁴ that EVs induce decidualisation genes in EnSCs, we next characterised the kinetics of this process at the protein level and under various culture conditions. These effects were evaluated both with EVs alone and in combination with various decidualisation inducers, across multiple EV collection time points. At all tested concentrations, EVs derived from decidualised EnSCs (DEVs) consistently demonstrated a robust capacity to induce decidualisation compared to EVs from undecidualised EnSCs (uDEVs), under various culture conditions on days 3 and 6 (Figure 3A,C). A marked effect was observed with EVs isolated on day 4 of decidualisation (DEV4) (Figure 3D‒F). These results underscore the time-sensitive nature of EV composition during decidualisation and point to DEV4 as an optimal pro-decidualisation signal. Interestingly, higher concentrations of differentiation stimuli paradoxically reduced decidualisation efficiency (Figure S2). Notably, prolactin secretion by DEVs was inhibited on days 3 and 6 under optimal concentration of cAMP and medroxyprogesterone acetate (MPA), indicating a threshold beyond which excessive signalling may disrupt cellular function (Figure 3B). Accordingly, similar to DEV4 derived from primary EnSCs, DEV4 isolated from an endometrial stromal cell line (ENSC) also induced decidualisation in primary EnSC cultures, albeit with lower potency (Figure 3G). Unlike uDEV from some donors, which modestly elevated prolactin secretion, uDEV4 derived from human foreskin fibroblasts (FSK-EV4) failed to induce decidualisation (Figure 3H), underscoring that EV functionality depends on cellular origin.

Although these experiments were conducted in vitro, the presence of EVs in human uterine fluid and their potential role in modulating the maternal‒foetal interface has been reported earlier. It is noteworthy that uterine fluid extracellular vesicles (UF-EVs) mirror the dynamic mRNA and miRNA changes of the endometrial tissue across the menstrual cycle.⁷ Therefore, the functional effects we observed, along with previous findings that EnSCs-EVs enhance vascular network formation and stimulate trophoblast differentiation,⁴ likely reflect the natural role of EVs at the maternal‒foetal interface. Consequently, analysing endometrial EVs shows significant promise as a basis for non-invasively assessing endometrial decidualisation potential.

Decidualisation and modulation of endometrial immune cells are closely linked processes. Within this context, modulating the function of NK cells, the most prevalent immune population in the early pregnant endometrium, is critically important, as they are essential regulators for maintaining pregnancy.⁸ Notably, menstrual stromal cells (MenSCs), commonly used as EnSC surrogates, can shift NK cells towards a dNK-like phenotype.⁹ Examining the effect of EVs on NK cell function, revealed that EVs derived from D-EnSCs did not significantly influence NK cell proliferation (Figure 4A), they notably, however, attenuated NK cell cytotoxicity against both EnSCs and K562 target cells (Figure 4B). Moreover, although treatment with these EVs did not significantly affect the expression of NK cell phenotypic and functional markers (Figure 4C), EVs from D-EnSCs increased the frequency of CD56^bright NK cells (Figure 4D), indicating a shift towards a ‘pregnancy-friendly’ phenotype and the role of D-EnSCs-EV in shaping uterine immune responses. Interestingly, decidualisation mediators (cAMP + MPA) exerted the same effects on the NK cell cytotoxicity and the frequency of CD56^bright NK cells (Figure 4B,D) suggesting that the impact of D-EnSCs-EV on NK cell cytotoxicity is in part mediated by decidualisation mediators (cAMP) that are packaged into the EVs during the decidualisation process. Interestingly, cAMP has been shown to modulate the function of NK cells by promoting the emergence of CD56^bright NK subsets, as we showed here, and this effect is mediated, at least in part, through the activation of the transcription factor FOXO1, a key regulator of NK cell differentiation within the decidual microenvironment.¹⁰

These findings are also consistent with our observation that decidualisation of EnSCs upregulated HLA-G expression, known to prevent NK cell cytotoxicity, following interferon gamma (IFN-γ) pre-treatment (Figure S4), and with the previous reports demonstrating that pro-inflammatory stimulation of stromal cells with IFN-γ and tumor necrosis factor alpha (TNF-α) enhances the anti-inflammatory and immunomodulatory properties of EVs.¹¹

To confirm these in vitro findings, in vivo studies are necessary. This would involve evaluating decidualisation markers in paired samples of UF-EVs and endometrial tissues, a crucial step for developing EV-based assessments of decidualisation capacity in women with infertility or miscarriage. Recent studies have also highlighted the diagnostic and therapeutic potential of EVs in endometrial pathologies such as endometriosis,¹² further supporting the relevance of EV-associated biomarkers in identifying endometrial dysfunction and guiding future interventions.

In conclusion, this research offers key insights into molecular interactions at the maternal‒foetal interface, emphasising the essential role of EVs in endometrial decidualisation and immune regulation, and proposes EVs as potential diagnostic agents for reproductive failures associated with impaired decidualisation.

Maryam Mousavi performed all experiments and wrote the first draft of the manuscript. Negar Vanaki contributed in performing the experiments and R data analysis. Kayhan Zarnani and Zahra Aghazadeh contributed in performing the experiments. Soheila Arefi and Jila Abedi-Aal acted as gynaecologist advisors and provided the endometrial biopsies. Fazel Shokri and Seyed-Aliraza Razavi critically read and edited the final version of the manuscript. Amir-Hassan Zarnani extensively contributed in conceptualisation, project administration, supervision, data validation, writing and critically editing the manuscript.

The authors declare they have no conflicts of interest.

The financial support was received for this research by grants from the Iranian Council for the Development of Regenerative Medicine and Stem Cell Technologies (https://stemcell.isti.ir/) (grant no. 11/104553) and Tehran University of Medical Sciences (https://en.tums.ac.ir/en) grant (no. 1401-1-99-57053).

All procedures carried out within the scope of this study received ethical approval from the Tehran University of Medical Sciences (TUMS) ethics committee under the reference code IR.TUMS.SPH.REC.1401.015. Written informed consent was obtained from all participants prior to inclusion in the study.