Amniotic MSC affect CD8 naive polarization toward SLEC/MPEC subsets by down-modulating IL-12Rβ1 and IL-2Rα signaling pathways

Summary Mesenchymal stromal cells (MSCs) are known for their immunomodulatory activity. Here, we report that MSCs isolated from the amniotic membrane of human term placenta (hAMSCs) impact CD8 T cell fate through a multifaceted mechanism. We observed that hAMSCs are able to impact the metabolism of naive CD8 lymphocytes by downregulating the phosphorylation of mTOR and AKT, thus blocking cell differentiation. This effect is due to the ability of hAMSCs to reduce the expression of two receptors, IL-12Rβ1 and IL-2RA, resulting in reduced phosphorylation of STAT4 and STAT5. In addition, hAMSCs reduce the expression of two transcriptional factors, Tbet and Eomes, directly involved in early effector cell commitment. Our results unravel an unknown feature of MSCs, offering alternative mechanistic insights into the effects of MSCs for the treatment of diseases characterized by an altered activation of memory subsets, such as autoimmune diseases and graft versus host disease.


INTRODUCTION
The formation of an adaptive T cell memory repertoire is a key feature of the primary adaptive immune response.
Shortly after activation, CD8 effector cells are differentiated into short-lived effector cells (SLECs), characterized by low survival and high proliferation potential, and long-lived memory precursor effector cells (MPECs) characterized by developing into memory cells. 1,2ntigen-specific memory CD8 T cells constitute the long-lived arm of immunity that rapidly protects the organism from secondary infections or tumors 3 and the inability to establish proper CD8 À driven immunological memory underlies numerous pathologies such as autoimmune diseases and allograft rejection [4][5][6] Mesenchymal stromal cells (MSCs) are versatile stromal components whose known role in modulating immunity and inflammation has been exploited in numerous clinical trials ranging from their use in the treatment or prevention of graft versus host disease (GvHD) to the treatment of autoimmune diseases such as Crohn's disease and neurodegenerative disorders. 7,8MSCs target a wide range of immune cells through cell-cycle arrest. 9,10One pivotal mechanism underlying this suppression involves interference with amino acid metabolism within the inflammatory microenvironment, in which the enzyme indoleamine 2,3-dioxygenase (IDO) has been shown to play a role. 11Other molecules with immunomodulatory properties that are secreted by MSCs include transforming growth factor-b1, hepatocyte growth factor, 12 prostaglandin E2, 13 and soluble human leukocyte antigen G (HLA-G). 14 Notably, tumor necrosis factor-a (TNF-a)-stimulated gene 6 protein (TSG-6) has emerged as a recent addition to the repertoire of anti-inflammatory factors secreted by human MSCs. 15ecently, efferocytosis has been reported as another mechanism responsible for the immune modulatory actions of MSCs.Upon infusion, apoptotic MSCs undergo phagocytosis by recipient phagocytes, leading to their functional polarization toward inhibitory phenotypes, and to the production of indoleamine 2,3-dioxygenase, thus triggering the immunomodulatory effect. 16SCs have been successfully identified and isolated from various tissues, encompassing not only adult sources such as bone marrow and adipose tissue, but also fetal sources such as the umbilical cord and placenta. 17This has propelled clinical research toward the use of different MSC subtypes due to their distinct characteristics, such as tissue availability, donor risk considerations, ease of isolation, which has led to a proliferation of clinical trials. 17,18mong the diverse types of MSCs, those derived from the amniotic membrane of the human term placenta (hAMSCs) have demonstrated distinct immunomodulatory properties.Indeed, unlike MSCs isolated from sources such as bone marrow, hAMSCs do not require priming to exert their immunomodulatory actions. 19,20In this context, hAMSCs have exhibited the capacity to influence the activation and differentiation of various components of adaptive immunity, including T cells, 21,22 whereby they have been shown to block the polarization toward inflammatory Th subsets, [21][22][23][24] and induce Treg regulatory cells. 22,25In addition, they have been shown to block the maturation and differentiation of monocytes into antigen-presenting cells (APCs) while promoting polarization and the acquisition of anti-inflammatory properties typical of M2 macrophages. 26,27][35] Here, we investigated a topic not yet explored, namely the potential impact of MSCs and in particular of hAMSCs and their secreted factors on the maturation and formation of the memory CD8 lymphocyte repertoire.
In this study we focused on the effect of hAMSC on the early, post-activation stages of naive CD8 lymphocytes, where the expression of two pivotal transcriptional factors, Tbet and Eomes, drive the differentiation of naive toward SLECs and long-lived MPECs CD8 T cells, respectively. 36,37We demonstrated that hAMSCs modulate the expression of Tbet and Eomes leading to reduced polarization toward the MPEC compartment, thus suggesting the capacity of hAMSC to counteract the formation of memory CD8 T lymphocytes.This effect is related to the capacity of hAMSC to directly impact CD8 T lymphocyte activation by reducing the phosphorylation of AKT and mTOR.Finally, we show that hAMSC downregulate the expression of IL-2Ra and IL-12Rb1 receptor resulting in reduced phosphorylation of STAT5 and STAT4, which are two key factors in governing the differentiation of CD8 naive lymphocytes. 38

hAMSCs harbor the potential to determine T cell fate
To investigate the impact of hAMSCs on immunological CD8 T cell memory commitment, we first analyzed the effect of hAMSCs on the activation and engagement of CD8 T lymphocytes within the peripheral blood mononuclear cell (PBMC) stimulated with anti CD3 mAbs over time.
Co-culture of PBMCs with hAMSCs, in cell-to-cell contact, reduced proliferation of both naive and differentiated CD8 T cell subsets (central memory (CM), effector memory (EM), and EM RA (TEMRA) at day 3 (Figure 1A).This effect was no longer detectable at day 7, since, at which time, CD8 proliferation was drastically reduced and the eventual effect of hAMSCs on CD8 proliferation was difficult to assess (Figure 1E).To determine whether the effect on T cell proliferation reflected changes in CD8 T cell memory commitment, we analyzed the proportion of different memory subsets generated from naive cells after treatment with hAMSCs.At day 3, differentiation of naive cells in CM, EM, and TEMRA was not affected by hAMSCs conditioning (Figure 1B).Conversely, at day 7, the EM pool was reduced and its reduction was counterbalanced by the preservation of naive and CM pools (Figure 1F).Accordingly, the flow cytometry analysis indicated a drastic reduction in the expression of Tbet and Eomes transcription factors after co-culture with hAMSCs.These are two transcription factors involved in the early stages of CD8 lymphocyte commitment to memory subsets.Notably, Tbet reduction was evident by day 3 and the reduction became statistically significant at day 7, particularly in the polarization toward the CM and EM subsets (Figures 1C and 1G).The influence of hAMSCs on Eomes transcription factor expression was notably more significant on day 3, resulting in a noticeable reduction in all examined subsets, although statistical significance was achieved only for the EM subset.By day 7, differences between cells co-cultured with hAMSCs and the control group had largely diminished for all the investigated subsets (Figures 1D and 1H).

hAMSCs directly orchestrate T lymphocyte commitment
To define the direct role of hAMSCs in CD8 T cell commitment, experiments were also performed with purified CD8 naive T lymphocytes stimulated with antiCD3 and antiCD28 mAbs in the presence of IL -12 and IL -2. 39,40fter activation, naive CD8 T cells progressively differentiate toward CM and EM subtypes (Figure 2A).Proliferation of naive, CM, and EM reached significant levels following activation, approximately three days after stimulation and peaks on day 7. Specifically, at day 7 approximately 80% of CMs and EMs proliferated (Figure 2B upper panel).When activation of naive CD8 T cells was performed in contact with hAMSC, the proliferation of naive T cells was not affected, instead the proliferation of both CM and EM was significantly reduced by hAMSC, throughout the period investigated (Figure 2B upper panel).Consequently, the commitment of cells to specific subpopulations was hindered, with a notable preservation of the naive cell fraction apparent at day 7.This was accompanied by a concomitant decrease in the polarization toward the CM and EM subpopulations (Figures 2A and 2B, lower panel).The observed effect, nonetheless, does not appear to be attributable to an increase in mortality (Figures S1A and S1B).To visualize the differentiation process over time and understand whether hAMSCs could induce the acquisition of distinctive features in CD8 T cells or differentiation into a different cellular subset, we conducted high-dimensional flow cytometry analysis, considering the expression of the activation marker CD183 and the proliferation marker Ki67.This approach has allowed for the identification of subclusters within different subpopulations that were actively proliferating.These subclusters indeed were characterized by their simultaneous association with a particular subset (Naive, CM, EM) and their positive expression of the proliferation marker Ki67.Once again, the impact of hAMSCs was evident in the reduction of the density of double-positive clusters, which was further reflected in a substantial accumulation of CD8 naive lymphocytes by day 10.These findings are in line with the results presented in Figure 2B, upper panel.This effect is lost in the control condition, in absence of hAMSCs, due to stimulus-induced differentiation.

hAMSCs influence early phases of CD8 naive T cell commitment
To further evaluate hAMSCs as determinants of T cell fate, we analyzed their ability to drive the differentiation of naive CD8 T cells toward the SLEC and MPEC subtypes.These two subtypes can be distinguished by the differential expression of two surface markers, CD183 and KLRG1.
SLECs are defined as CD183-KLRG1+, whereas MPECs are characterized by expression of CD183 and loss of positivity for KLRG1.We also examined the frequency of early effector cells (EECs), which function as master precursor effector T cells and are characterized by co-expression of the two markers mentioned previously (CD183+KLRG1+), 40 and the frequency of double-negative effector cells (DNECs), whose role in CD8 lymphocyte commitment remains unclear. 40aving confirmed the ability of activated naive CD8 T lymphocytes to differentiate into MPECs in the presence of exogenous IL-12 and IL-2 (Figure 3A), 39 we next examined the effects of hAMSCs on this early commitment phase.As shown in Figure 3A, treatment with hAMSCs resulted in a reduction in the MPEC compartment in terms of both relative frequency (Figure 3A left panel) and absolute number (Figure S2).In parallel, in presence of hAMSC the percentage of DNEC fraction increased (Figure 3A left panel), whereas no difference was observed in the absolute number of DNEC T cells (Figure S2).From the proliferation analysis of the different subsets, we observed that co-culture with hAMSCs leads to a significant reduction in the EEC and MPEC subsets at day 3 (Figure 3A, right panel).This difference persisted even on day 7 (Figure 3A, right panel).The transcription factors Tbet and Eomes promote the differentiation of naive CD8 lymphocytes. 37Here, we observed a substantial reduction in the mean fluorescence intensity (MFI) ratio of Tbet and Eomes on both CM and EM at day 3 from challenge with hAMSCs, a reduction that persisted until day 10, although not significantly (Figure 3B).
We next sought to identify the appearance of different CD8 T cell subsets or clusters over time.The unsupervised analysis was performed as aforementioned and shown in Figure 3C.Co-culture with hAMSCs allowed to retain a large amount of naive T lymphocytes (in brown) and increased the proportion of CD183 naive T lymphocytes (blue), suggesting that hAMSCs do not affect the early stages of activation of naive CD8 lymphocytes.However, hAMSCs strongly reduced the percentage of naive T lymphocytes positive for the expression of Tbet and Eomes (purple) as well as the commitment toward the CM (red and dark purple) and EM (green) memory subsets (Figure 3D).

hAMSCs affect CD8 T cell metabolism
The PI3K/mTOR pathway is involved in the activation, proliferation, and differentiation of CD8 T lymphocytes. 41Because mTOR and its two main complexes, mTORC1 and mTORC2, control the differential expression of Tbet and Eomes, 42,43 we next wanted to investigate the effect  of hAMSCs on these important regulators of cellular metabolism.After three days of co-culture with hAMSCs, the total and phosphorylated forms of AKT and mTOR were reduced in activated CD8 T lymphocytes (Figure 4A).These observations suggest that hAMSCs may negatively affect the metabolic switch of naive CD8 T lymphocytes from fatty acid to glycolytic metabolism, which is required for full maturation of CD8 T lymphocytes. 44epresentative estimation plots of the MFI ratio between total protein and relative phosphorylated protein demonstrate that hAMSCs reduce the phosphorylation of AKT, leading to a substantial reduction in mTOR phosphorylation and activation (Figure 4B).
When we compared the changes in the phosphorylation status of AKT and mTOR on day 7, we observed a drastic reduction in the phosphorylation of these two proteins, very likely because the commitment process is an early event, and therefore appreciable on day 3, and not on day 7 where the commitment process has already occurred (Figure 4C).This suggests a potential mechanism in which hAMSCs function to counteract the activation and differentiation of naive CD8 lymphocytes during the early days of commitment by inhibiting AKT and mTOR signaling.

hAMSCs modulate the transcriptional landscape of naive CD8 T lymphocytes
To outline the transcriptional landscape peculiar to the commitment of naive CD8 T lymphocytes upon hAMSCs stimuli, we performed a high throughput gene expression analysis focusing on several genes involved in differentiation and metabolic processes, as well as epigenetic and transcriptional regulation of naive CD8 lymphocytes.We performed an unsupervised hierarchical clustering analysis and identified genes whose expression is over 2-fold higher and statistically significant (p < 0.05) in the early (24 h) post-activation stages (Figure S3).
Having confirmed the upregulation of genes such as IL-12 receptor beta 1 (IL12RB1) and interleukin-2 receptor alpha (IL2RA), involved in the early stages of CD8 T cell activation, we next interrogated on determinants of cell fate.As shown in Figure 5A and Figure S3, the two main inflammatory cytokines, IFNG and TNF, typically induced in the early differentiation steps, were upregulated in activated cells. 38,45At the same time, also IRF4 and the two master regulators of SLEC and MPEC commitment, TBX21 (Tbet) and EOMES (Eomesodermin), respectively, were upregulated.Conversely, the expression of STAT genes (STAT3, 4 and 5), modulators of both activation and commitment of CD8 T lymphocytes, was higher in naive CD8 T cells compared to activated cells.Interestingly, when naive CD8 T lymphocytes were activated in the presence of hAMSCs, their gene expression profile was comparable to that of unstimulated naive CD8 T lymphocytes, with the exception of IL2RA and IRF4, whose increase upon activation was maintained also in the presence of hAMSCs.
Next, we focused our attention on genes involved in metabolic switching.As shown in Figures 5B and S3, SREBP1 and 2, AMPK, PDK1, RAPTOR, and RICTOR resulted downregulated after stimulation of naive CD8 T lymphocytes.Again, when naive CD8 T lymphocytes were activated in the presence of hAMSCs, their gene expression profile, and in particular SREBP1 and 2, PDK1, and RAPTOR genes, was comparable to that of unstimulated naive CD8 T lymphocytes.Similarly, the expression of genes involved in the epigenetic response such as BACH2 and ID3 or the transcription factor BCL6 46,47 was downregulated after activation and preserved in the presence of hAMSCs (Figures 5B  and S3).
The results obtained emphasize that hAMSCs block CD8 cells that keeps them transcriptionally similar to naive CD8 by inhibiting some of the signaling pathways responsible for the activation and commitment of CD8 lymphocytes.

hAMSCs regulate the downstream signaling cascade of IL-12 and IL-2 receptor
Given the transcriptional downregulation of IL12Rb1 on CD8 T cells after hAMSCs challenge (Figure 5A), we speculated on the possible involvement of this receptor in the mechanism of action of hAMSCs.To confirm our hypothesis, we examined the expression of the IL-2Ra and that of IL-12Rb1 at day 3 on both the surface and cytoplasm of activated naive CD8 T lymphocytes after complete and incomplete stimulation (i.e., in the absence of exogenous IL-12 and IL-2 stimulation).As shown in Figure 6A, the expression of IL-12Rb1 on the surface of activated naive CD8 lymphocytes increased after exogenous stimulation with the two cytokines, but no significant difference in IL-12Rb1 was observed when CD8 T lymphocytes were co-cultured with hAMSCs when compared with naive CD8 T lymphocytes.These results were also confirmed by the MFI data (Figure 6B).Instead, a strong decrease in IL-2Ra expression was observed in CD8 lymphocytes co-cultured with hAMSCs (Figures 6A and 6B).Whereas this has been observed previously for IL2-Ra in CD4, 48 this result is completely unknown for both IL2-Ra and IL-12Rb1 on CD8 T lymphocytes.Indeed, the expression levels of IL-12Rb1 in the membrane of naive CD8 cells co-cultured with hAMSCs (either incompletely or fully activated) were comparable to those of the control condition (naive CD8 T lymphocytes activated with the different combination of stimuli) (Figures 6A and 6B).This trend was not confirmed at the cytoplasmic level, where co-culture with hAMSCs resulted in a decrease in IL-12Rb1 expression.These results are functionally reflected in a reduction in downstream signaling of the two receptors.Indeed, activation of naive CD8 T cells in the presence of hAMSCs reduced phosphorylation of STAT4 (downstream the IL-12Rb1 receptor) and STAT5 (downstream the IL2-Ra receptor).

DISCUSSION
This study sheds light on the mechanism of action of MSCs isolated from the amniotic membrane (hAMSCs), by explaining how they affect CD8 T lymphocyte activation and differentiation and their commitment toward memory subsets.Indeed, it is well-established that MSCs possess the ability to inhibit the activation of T lymphocytes.Considerable knowledge exists regarding their capacity to direct the polarization of CD4 lymphocytes toward immunoregulatory subsets (Tregs), concomitantly leading to a reduction in the polarization of inflammatory subsets.However, not much is known about the capability of MSCs to influence the activation and differentiation of CD8 lymphocytes.Notably, in this study conducted on PBMCs, we observed that hAMSCs can attenuate the proliferation and the differentiation of stimulated T lymphocytes, promoting the preservation of T cells in a naive state.This effect may be partly attributed to the ability of hAMSCs to induce cell-cycle arrest, as previously reported by our group in the context of tumor cells. 49This cell-cycle arrest exerted by hAMSCs also influences the commitment of CD8 lymphocytes.Notably, hAMSCs induce a cessation in the expression of transcription factors T-bet and Eomes, both of which play critical roles in the differentiation process of T cells. 36,50,51Specifically, our study shows that hAMSCs block CD8 T cell differentiation and that this is related on the ability of hAMSCs to modulate the expression and downstream signaling of IL-12 and IL-2, receptors which are essential for the full activation and subsequent maturation of naive CD8 lymphocytes. 39emarkably, we observed that hAMSCs do not counteract the activation of CD8 T lymphocytes, because after stimulation naive CD8 T lymphocytes are equally capable of expressing CD183, a marker that is rapidly induced on naive T lymphocytes after activation, 52 but rather decrease commitment to the various memory subsets (CM, EM).This effect has been reported previously for both hAMSCs 25 and bone marrow-derived MSC, although in the latter it was observed only after stimulation with the peptide HY. 53t is well known that cell activation and metabolic activity are closely related.Therefore, we investigated whether hAMSCs are able to modulate the PI3K pathway and consequently the activation of mTOR.The PI3K pathway is involved not only in the activation of T lymphocytes, 41 but also in the regulation of metabolism, 54 and mTOR is considered a master regulator of memory CD8 + T cell differentiation. 55,56We observed that hAMSCs are able to decrease the phosphorylation levels of AKT and mTOR, suggesting a possible mechanism to modulate the activation of CD8 T lymphocytes.However, this result is in contradiction with the results of transcriptional analysis, where we observed increased expression of AKT, mTOR, LDHA, CS, and HK2 all genes whose expression is directly involved in the regulation of the various metabolic pathways underlying T cell activation and early commitment. 57In the context of T cell receptor (TCR) stimulation, signals stemming from growth factor cytokines such as IL-2 and the binding of co-stimulatory CD28 receptors can lead to an upsurge in glycolysis.This increase is a result of the PI3K-dependent activation of Akt. 58,59This, in turn, can promote the activation of the mTOR pathway. 60Additionally, it can stimulate glycolysis by elevating the activity of glycolytic enzymes and enhancing the expression of nutrient transporters.2][63] The involvement of fatty acid metabolism has also been known in the activation process.Sterol regulation during T cell activation has been linked to transcriptional responses mediated by SREBP and LXR. 64n our case, we observed levels of SREBP1 and 2 expressions comparable to those in naive CD8 lymphocytes.Interestingly, blocking the mTOR pathway gives effects similar to those we observed with hAMSCs.mTOR blockade, indeed inhibit cell proliferation while favoring the promotion of other forms of nutrient (protein) catabolism. 65n this regard, our findings suggest that hAMSCs may potentially impact CD8 naive T cell proliferation and activation by impeding cell cycle progression and even suggested by our Ki67 data, and by the reduced levels of total AKT and mTOR protein expression.
This effect may also be attributed to a metabolic reprogramming on multiple fronts.Supporting this hypothesis, for instance, is the expression level of AMPK, a protein involved in the regulation of lipid metabolism.In cells co-cultured with hAMSCs, AMPK exhibits a gene expression pattern akin to that of naive T cells.The sustained presence of AMPK could elucidate the decrease in mTOR phosphorylation.In fact, AMPK functions by phosphorylating TSC2, thereby impeding mTOR activation. 66ltogether, these findings suggest that hAMSCs do not completely block activation triggered by TCR and costimulatory molecules, but rather interfere with subsequent differentiation processes.By analyzing the expression of the two receptors involved in CD8 memory commitment, we observed that hAMSCs downregulate the expression of IL12RB1 transcript while upregulating the expression of IL2RA.
Analysis of membrane receptors, however, gave us unexpected results.While we expected a decrease in phosphorylation of STAT5 as a consequence of translational blockade of IL2RA, the similarity of the levels of IL-12Rb1 in naive CD8 T cells with or without hAMSCs conditioning suggests that there is no difference in STAT4 phosphorylation.Interestingly, we observed that STAT4 phosphorylation was reduced when naive CD8 T lymphocytes were co-cultured with hAMSCs.
This finding leads us to hypothesize that hAMSCs either induce a phenotype of CD8 T cells that is less responsive to exogenous stimulation or a potential blockade of epigenetic or cross-mediated activation pathways.Another possible explanation is related to the potential alteration in the expression of IL-12Rb2 whose heterodimerization with IL-12Rb1 is required for full activation of the receptor and subsequent phosphorylation of STAT4. 678][69] Moreover, hAMSCs may result in a non-fully functional IL -12Rb1 activation, as previously reported, 69 which can explain the observed reduction in signal transduction consequently impacting on the phosphorylation of STAT4.
Furthermore, the lack of IL-2Ra translation means that ligands, even when administered exogenously, do not trigger the transduction cascade, which in turn could explain the lower activation induced by hAMSCs.
The low activation status is also reflected in the lack of IFNg and TNFa expression.Indeed, the IFNg promoter is highly methylated in naive CD8 lymphocytes. 70hAMSCs likely retain this property in naive CD8 T lymphocytes, suggesting epigenetic regulation by hAMSCs that correlates with their immunomodulatory capacity.
Differentiation of naive CD8 lymphocytes is regulated by a dynamic and tightly coordinated process involving Tbet and Eomes 1,38,39 as well as other transcription factors such as IRF4, BCL6, and epigenetic factors such as BACH2, and ID3. 37,71,72We observed that hAMSCs modulate this innate cellular commitment and, in particular, affect T lymphocyte activation and differentiation into SLEC and MPEC subsets.TBX21 and EOMES, the two major regulators of SLEC/MPEC fate, are transcriptionally downregulated by hAMSCs as in nonactivated naive CD8 T cells.Consistently, protein expression of Tbet and Eomes was also decreased, which may be related to the observed reduction in STAT4 and STAT5 phosphorylation. 72,73urthermore, hAMSCs are able to affect the expression of BACH2 by decreasing the availability of the AP-1 complex at transcription sites thus inducing the repression of genes related to terminal differentiation of T lymphocytes. 74n addition, we observed that hAMSCs modulate the expression of ID2 and ID3, two genes that have been reported to be involved in the regulation of SLEC or MPEC commitment. 46IL-2, IL-12, and IL-21 have been described to upregulate ID2 and downregulate ID3 expression.We show that hAMSCs impact the expression of ID2 and ID3 by reducing IL-2Ra and IL-12Rb1 signal transduction, thus explaining the high level of ID3 expression comparable to that of naive CD8 T lymphocytes, and at the same time the reduction of ID2 expression.Conversely, ZEB2, a target of Tbet normally involved in differentiation processes and induction of effector CD8 T lymphocyte differentiation, 75 was upregulated by co-culture with hAMSCs.
Our work unravels an unknown immunological mechanism of hAMSCs that controls the transcriptional profile of CD8 T lymphocytes between that of naive and activated cells, thereby influencing the formation of an immunological memory.One of the major implications of our study is that it focuses on a poorly studied aspect that is mostly neglected in the field of MSC, namely the effects of MSC on CD8 lymphocytes.Although MSC remain one of the most studied cell therapies for the treatment of immune-related diseases due to their immunoregulatory capabilities, it remains critical to understand what will occur in a patient's immune system and if the treatment may have side effects.In addition, we report a mechanism of action for hAMSCs to explain the observed effect.
Moreover, it supports the possibility of administering multiple doses of MSC, even from different donors, and highlights their therapeutic potential in graft-versus-host disease 6 or autoimmune diseases in which CD8 T lymphocytes play an important role. 76,77gure 6.hAMSCs regulate the downstream signaling cascade of IL-12 and IL-2 receptor Purified naive CD8 T lymphocytes were either stimulated with anti CD3, anti CD28, and by the exogeneous administration of IL-12 and IL-2 or without the exogenous administration of the two cytokines (not fully activated) in the presence or not of hAMSCs.The total amount of IL-12Rb1 and IL-2Ra expressed on the membrane and in the cytoplasm of CD8 T cells (A) and the level of mean fluorescence intensity (B) were assessed by flow cytometry.The phosphorylation status of the STAT4 and STAT5 was assessed by flow cytometry on fully activated naive CD8 T lymphocytes co-cultured or not with hAMSCs (C).Results are displayed as violin plots showing median (dashed line), 25th and 75th quartiles (*p < 0.05, **p < 0.01****p < 0.001); N = 4 independent experiments performed starting from 4 purified CD8 naive donors and 6 different hAMSC preparation.

Limitations of the study
Our study, provides insights into the immunomodulatory capabilities of amniotic MSCs, particularly in modulating the activation and commitment of naive CD8 lymphocytes toward memory subsets.Although our study has unveiled alternative mechanisms, there are some limitations to consider.
Firstly, the study was conducted through PBMC/purified CD8 naive T lymphocytes and hAMSCs in co-culture in direct contact.While we have previously observed immunomodulatory effects by utilizing both hAMSCs in transwell system and the hAMSC secretome, suggesting the involvement of secreted factors, we cannot exclude the possibility that some of the observed effects may be the consequence of specific interactions with surface receptors expressed by hAMSCs.
Secondly, our study does not definitively identify the specific factor/s responsible for this immunomodulatory action.This is indeed a challenging task given the multitude of factors and molecules that could potentially contribute to such an action.Finally, another potential limitation could be that the study was conducted with a specific MSC population from the amniotic membrane, raising the possibility that the mechanism described here may be specific to these cells and not applicable to MSC from different tissues.Conducting a comparative study would be valuable in elucidating whether this mechanism is a shared characteristic among all types of MSCs.This clarification holds significance, especially in light of their widespread utilization in clinical trials, it could provide guidance for their application in the treatment of conditions with distinct characteristics.However, this research is crucial for advancing our comprehension of the therapeutic potential of these cells and for more precisely delineating their capacities in translational applications.Oligonucleotides
Proliferation of the T lymphocytes was assessed with Ki67+ Alexa 488 (clone B56, BD Biosciences) and Tbet APC (clone REA102, Miltenyi) and Eomes FITC (clone WD1928, Invitrogen) were assessed by staining with conjugated monoclonal antibody after intranuclear permeabilization used Transcription Factor buffer kit (BD).CD8 absolute count was performed by using BD Trucountä Trucount Absolute Counting Tubes IV (BD Bioscences) according to manufacturer instruction.
Cells were collected at day 3 and 7 (PBMC) and at day 3, 7, and 10 (naive CD8 T lymphocytes), and cells were stained with the eBioscien-ceTM Fixable Viability Dye eFluorTM 780 (Thermo Fisher Scientific) according to the manufacturer's instructions, in order to exclude dead cells during analysis.
For the analysis conducted on CD8 T lymphocytes to identify naive to memory subsets, the gating strategy was as follows: after initial gating based on morphological characteristics, we selected live cells as eFluorTM 780 dim/negative cells.Within the live cell population, we excluded doublets.Subsequently, we gated for CD3 and CD8-positive cells.The dissection of naive to memory subsets was performed based on the differential expression of CD45RO and CD197 markers, defining the subsets as follows: Naive cells: CD45RO-CD197+, Central Memory (CM): CD45RO+CD197+, Effector Memory (EM): CD45RO+CD197-, Terminally Differentiated Effector Memory (TEMRA): CD45RO-CD197-.
Unsupervised analyses on naive CD8 lymphocytes was performed to investigate cell fate development at different time points (day 3, 7, and 10).Multidimensional analysis, was performed by using FlowSOM meta-clustering in conjunction with a dimensionality reduction method such as Uniform Manifold Approximation and Projection (UMAP).The identification of specific cluster was performed by using Cluster Explorer plugin.
Samples were acquired on FACS Symphony A3 BD.Data were analyzed with FlowJo 10.8.

Determination of phosphorylation status
Three days after co-culture the cells were fixed by adding 1.5% methanol-free formaldehyde (ThermoFisher) for 10 min at room temperature (RT).Then, the cells were collected and stained for 20 min at 4 C with CD8 BV421 (RPA-T8, BD Biosciences).After cell permeabilization in cold 90% methanol for 2-3 days at À80 C, cells were stained in the dark with total AKT and total mTOR (both of Cell signaling Technology, Danvers, Massachusetts, USA) for 30 min at RT, then the cells were washed in stain buffer, consisting of 0.02% sodium azide and 0.1% bovine serum albumin in PBS (Sigma-Aldrich).Cells were then incubated for 30 min at RT with secondary antibody fluorochrome-conjugated anti-Rabbit Dylight 488 (Vector Laboratories, Newark, CA, USA).

Quantitative real-time PCR
Total RNA was extracted using EZ1 RNA cell Mini Kit protocol (Qiagen, Frederick, MD, USA), in a BioRobot EZ1 Advanced XL Workstation.The iScript Advanced cDNA Synthesis Kit for RTqPCR (Biorad, Hercules, California, USA) was used for cDNA synthesis, before quantitative realtime PCR cDNA was pre-amplified with SsoAdvanced PreAmp Supermix (Biorad).Real-time PCR was performed using the Biorad instrument CFX96 Quantitative Real-Time PCR.Data were analyzed with Biorad CFX Maestro 2.2 (Biorad).For molecular analysis, cells were collected at day 1 and CD8 + T cells were purified from hAMSCs by positive selection using CD8 microbeads and MACSâ separation columns (Miltenyi) according to the manufacturer's instructions.After separation, the cells were centrifuged and pellets were stored at À80 C for RNA extraction.

QUANTIFICATION AND STATISTICAL ANALYSIS
The sample size calculation will be set based on previous literature (53) in which the effect of hAMSC secretome on immunomodulatory activity was evaluated.In detail, in such a study the use of hAMSC secretome in PBMC reduced the lymphocytes proliferation mean from 4.6 (x10 4 cpm) (SD about 1.4) to 1.5 (SD = 1.2).Consequently, the sample size of the PBMC study was of R3 healthy donors for each readout performed.Data are shown as violin truncated plots with Tukey variations.The parameters were compared using two-way, one-way analysis of variance (ANOVA) and Student's t test.Normal distribution was analyzed by performing a Q-Q probability plot-and analytically by performing Shapiro-Wilk test, Kolmogorov-Smirnov test.Statistical details of every experiment can be found in the figure legends, median and quartile and value of n.Statistical analysis was performed using Prism 9 (GraphPad Software, La Jolla, CA, USA).A p value lower than 0.05 was considered statistically significant.

Figure 1 .
Figure 1.hAMSCs harbor the potential to determine T cell fate (A-H) PBMC stimulated with antiCD3 antibody were co-cultured with hAMSCs p1 (gray) for 3 (upper panels) or 7 (lower panels) days.Cell proliferation was evaluated as percentage of Ki67+ cells for the different CD8 + subsets investigated at day 3 (A) and 7 (E).Cells were then analyzed by flow cytometry to discriminate the CD8 naive T cell pool from the memory subsets as follows: naive T lymphocytes (CD197+CD45RO-), central memory (CM) (CD197+CD45RO+), effector memory (EM) (CD197ÀCD45RO+) and effector memory RA (TEMRA) (CD197-CD45RO-).CD8 T cell differentiation is represented as the percentage of the different subsets identified at day 3 (B) and 7 (F).The expression of transcription factors Tbet and Eomes within these subsets was quantified as mean fluorescence intensity (MFI) ratios at day 3 for Tbet (C) and Eomes (D) and at day 7 for Tbet (G) and Eomes (H).Results are displayed as violin plots showing median (dashed line), 25th and 75th quartiles (*p < 0.05, **p < 0.01, ****p < 0.001); stimulated PBMC alone represent the control group (white); N = 5 independent experiments performed starting from 5 PBMC donors and 6 different hAMSC preparation.

Figure 2 .
Figure 2. hAMSCs directly orchestrate T lymphocyte commitment Purified CD8 naive T lymphocytes were stimulated with antiCD3, antiCD28, and the exogeneous administration of IL-12 and IL-2, and cultured in the presence of hAMSCs (gray).(A and B) CD8 T cells were allowed to differentiate for 10 days and the degree of proliferation (as percentage of Ki67+ cells) and commitment (as percentage of the different T cell subsets) evaluated at day 3, 7, and 10 by flow cytometry.CD8 T cells were distinguished based on the differential expression of CD197 and CD45RO in naive (CD197+CD45RO-), central memory (CM) (CD197+CD45RO+) and effector memory (EM) (CD197ÀCD45RO+).(C) Uniform Manifold Approximation and Projection (UMAP) representation of the CD8 + T cell landscape obtained by Clusterexplorer plugin.Cells were stratified for CD197/CCR7, CD45RO and CD183 and subsequently clustered based on the expression or not of the proliferation marker Ki67, allowing to identify 6 clusters (Naive Ki67-, Naive Ki67+, CM Ki67-, CM Ki67+, EM Ki67-, and EM Ki67+ (C).Results in (A) and (B) are displayed as violin plots showing median (dashed line), 25th and 75th quartiles (*p < 0.05, **p < 0.01, ****p < 0.001); stimulated CD8 T cells alone represent the control group (white); N = 4 independent experiments performed starting from 4 purified CD8 naive donors and 4 different hAMSC preparation.

Figure 3 .
Figure 3. hAMSCs influence early phases of CD8 naive T cell commitmentPurified CD8 naive T lymphocytes were stimulated with antiCD3, antiCD28, and the exogeneous administration of IL-12 and IL-2, and cultured in the presence of hAMSCs (gray).CD8 T cells were allowed to differentiate for 10 days and the different cellular subtypes analyzed at day 3, 7, and 10 as relative frequency (left panel 3A or as % of Ki67 proliferating cells).(A) CD3 + CD8 + T lymphocytes were further analyzed for the expression of CD183 and KLRG1 and classified into four populations: SLECs (CD183-KLRG1+), EECs (CD183+ KLRG1+), MPECs (CD183+ KLRG1-), and DNECs (CD183-KLRG1-).(B) The expression of the transcription factors Tbet and Eomes was evaluated at all time points.(C) Uniform Manifold Approximation and Projection (UMAP) representation of the CD8 T cell landscape obtained by Clusterexplorer plugin.(D) Pie charts representing the distribution levels of the 6 identified clusters as a percentage of the total CD8 pool.Results in (A) and (B) are displayed as violin plots showing median (dashed line), 25th and 75th quartiles (*p < 0.05, **p < 0.01, ****p < 0.001); activated CD8 T cells alone represent the control group (white); N = 3 independent experiments performed starting from 3 purified CD8 naive donors and 4 different hAMSC preparation.

Figure 4 .
Figure 4. hAMSCs affect CD8 T cell metabolism Purified CD8 naive T lymphocytes were stimulated with antiCD3, antiCD28, and the exogeneous administration of IL-12 and IL-2, and cultured in the presence of hAMSCs.Cells were harvested at day 3 and analyzed by flow cytometry.Specifically, CD8 T cells were stained for phospho mTOR, phospho AKT, and for their total protein content.(A and B) Representative plots of the MFI ratio between total protein and relative phosphorylated protein.(C) Comparison of AKT and mTOR phosphorylation at the two different time points, 3 and 7 days.Results in (B) are displayed estimation plot with Tukey variation.Results in (C) are displayed as violin plots showing median (dashed line), 25th and 75th quartiles ***p < 0.001; activated CD8 T cells alone represent the control group; N = 3 independent experiments performed starting from 3 purified CD8 naive donors and 4 different hAMSC preparation.

Figure 5 .
Figure 5. hAMSCs modulate the transcriptional landscape of naive CD8 T lymphocytes Transcriptomic profile of naive, activated and activated in the presence of hAMSCs CD8 T cells 24 h upon activation.Panels show metabolic related genes (A), epigenetic and transcriptional factors (B), and genes strictly related to naive CD8 T cell differentiation (C).The mRNA expression profile is represented as heatmap graph.The genes were clustered according to their expression patterns.N = 3 independent experiments performed starting from 3 purified CD8 naive donors and 5 different hAMSC preparation.