Intracellular role of IL-6 in mesenchymal stromal cell immunosuppression and proliferation

Interleukin (IL)-6 is a pleiotropic cytokine involved in the regulation of hematological and immune responses. IL-6 is secreted chiefly by stromal cells, but little is known about its precise role in the homeostasis of human mesenchymal stromal cells (hMSCs) and the role it may play in hMSC-mediated immunoregulation. We studied the role of IL-6 in the biology of bone marrow derived hMSC in vitro by silencing its expression using short hairpin RNA targeting. Our results show that IL-6 is involved in immunosuppression triggered by hMSCs. Cells silenced for IL-6 showed a reduced capacity to suppress activated T-cell proliferation. Moreover, silencing of IL-6 significantly blocked the capacity of hMSCs to proliferate. Notably, increasing the intracellular level of IL-6 but not recovering the extracellular level could restore the proliferative impairment observed in IL-6-silenced hMSC. Our data indicate that IL-6 signals in hMSCs by a previously undescribed intracellular mechanism.

Transduction of shRNAs. shRNA expression vectors were constructed using standard cloning procedures.
The following shRNA sequences have been published previously 22 and were purchased from Sigma-Genosys (Oakville, ON, Canada): IL-6ia: AGA TGG ATG CTT CCA ATC TGG and IL-6ib: AAG GCA AAG AAT CTA GAT GCA. Both targeting sequences were purchased from the RNAi Consortium (www.broad insti tute.org/rnai). We used two different target sequences to avoid off-target effects. Oligonucleotides were annealed and cloned into the pSUPER plasmid carrying an H1 promoter using BglII-HindIII sites. The H1-shRNA expression cassette was then excised and cloned into pLVTHM (Addgene plasmid 12,247, www.addge ne.org) using EcoRI-ClaI sites 21 . Viral particles were produced as described by the Viral Vector Platform at Inbiomed Foundation 21 . hMSC transduction was carried out at a multiplicity of infection of ten in order to achieve 100% infection. When indicated, transduction was performed to obtain 50% infection to compare from the same population the effect of infection on GFP+ and GFP-cells.
Flow cytometry. Cell cycle analysis was performed as described Briefly, hMSCs were fixed and washed twice with PBS and resuspended in PBS containing 5 mg/ml propidium iodide (PI) and 10 μg/ml RNase A (Sigma-Aldrich). Cell cycle analysis was performed on GFP (530/30BP emission filter)-positive and living cells, excluding doublets 23 . IL-6 levels were measured in samples with a custom cytometric bead array kit (CBA; BD Biosciences, San Jose, CA) for IL-6 following the manufacturer's instructions 11 . Samples were incubated with the CBA during 30 min and were mixed with the combined cocktail of phycoerythrin (PE)-conjugated antibodies. IL-6 concentration was measured via quantification of PE fluorescence in reference to a standard curve.
Expression analysis. Total RNA was extracted using the RNAeasy Extraction Kit (Qiagen, Hilden, Germany). cDNA was obtained using the GeneAmp Reverse Transcriptase Kit (Applied Biosystems, Foster City, CA). Quantitative PCR was performed using the Power SYBRR Green PCR Master Mix (Applied Biosystems). IL-6, IDO, COX2 and GAPDH were amplified using the following oligonucleotide pairs: IL-6-AAC GCT CCT CTG CAT TGC CATT and GAG CAG CCC CAG GGA GAA; IDO-CTA CCA TCT GCA AAT CGT GAC TAA G and GAA GGG TCT TCA GAG GTC TTA TTC T; COX2-GAA TCA TTC ACC AGG CAA ATTG and TCT GTA CTG CGG GTG GAA CA; GAPDH-TGC ACC ACC AAC TGC TTA GC and GGC ATG GAC TGT GGT CAT GAG. Reactions were carried out in a Step One Plus Thermocycler (Applied Biosystems). Data were compared using the comparative CT method, normalizing all samples against hMSCs infected with the empty vector control (pLVTHM emp). GAPDH was used as a housekeeping gene control.
Detection of prostaglandin E2. PGE2 levels were determined using a commercial ELISA kit (R&D Systems) on supernatants of hMSC cells transduced with pLVTHM emp or pLVTHM IL-6i, treated or not, as indicated, with indomethacin or etoricoxib for 48 h.

Immunofluorescence.
To determine the intracellular localization of p65 or KI-67, cells were fixed with 4% paraformaldehyde solution (Pancreac, Barcelona, Spain) for 10 min and permeabilized with PBS-Triton X-100 (0.1%) for 10 min. Staining was performed using a rabbit anti-human p65 (Santa Cruz Biotechnology) or KI-67 (BD Biosciences) antibody, and revealed with a donkey anti-rabbit IgG secondary antibody conjugated with Cy-3 (Jackson Immunoresearch, West Grove, PA). Analyses were performed using an LSM 510 Meta Laser Scanning Microscope (Zeiss, Jena, Germany) at the Cytometry and Advanced Microscopy Platform at the Inbiomed Foundation.
Statistical analysis. Data are expressed as mean ± standard error of the mean. Student's t test was used for comparison between groups. When the distribution was not normal, the Mann-Whitney U test was used. Analysis of variance and the Kruskal-Wallis test were used to compare the means of more than 3 groups. Analy- www.nature.com/scientificreports/ ses were conducted with GraphPad Prism 8 software (GraphPad Software Inc., La Jolla, CA). Differences were considered statistically significant at p < 0.05 with a 95% confidence interval.
Ethics approval and consent to participate. The use of human cells and the project were approved by the Ethical committee of Inbiobank.

IL-6 is involved in the immuregulatory function of hMSCs.
We previously demonstrated that TNF-α/NF-κB priming/signaling regulates immunoregulatory profile of hMSCs, which could be inhibited by the presence of glucocorticoids such as dexamethasone 11 . Intrigued by the mechanisms involved in the modulation of hMSC properties, we focused our attention on IL-6, whose expression is under the control of NF-κB. Although hMSC constitutively express IL-6, TNF-αinduced a marked and statistically significant increase in IL-6 expression at the level of mRNA (Fig. 1A), protein (Fig. 1B) and secretion (Fig. 1C). Additionally, both alpha and beta subunits of IL-6R (CD126 and CD130, respectively) were also expressed in hMSC (Fig. 1D), suggesting that the IL-6 could have an autocrine role in the biology of these cells. Signalling with TNF-α priming did not significantly alter the abundance of the receptors in the membrane of hMSCs (Fig. 1D). Dexamethasone treatment of hMSCs inhibited both basal and TNF-α-induced IL-6 secretion (Fig. 1E), which correlated with a reduced capacity of hMSCs to impact T-cell proliferation (Fig. 1F).
To test whether the effect of dexamethasone on the hMSC immunoregulatory function was mediated by the decrease in IL-6 expression, we used a GFP-expressing lentiviral vector to transduce more than 95% of hMSCs with an empty vector or a shRNA targeting IL-6 (hereafter referred to as hMSC-emp and hMSC-IL6i, respectively) (Supp. Figure 1). Two silencing sequences were evaluated against IL-6 to account for off-target effects. As shown in Fig. 2A,B, the tested sequences significantly reduced the quantity of IL-6 mRNA and secretion, respectively. Both sequences were tested in the majority of the experiments, although for brevity only one sequence is shown in the figure panels (IL-6ia). We next analyzed whether inhibition of IL-6 expression impacted the capacity of hMSCs to regulate T-cell proliferation. As expected, hMSC-emp significantly impaired T-cell proliferation in a dose-dependent manner (Fig. 2C). By contrast, activated T-cells cultured with hMSC-IL-6i proliferated significantly greater than those cultured with hMSC-emp (Fig. 2C).
We next determined whether these differences were due to changes in the MSC:PBMC ratio. Analysis of spontaneous cell death measured by Annexin V revealed no significant differences between hMSC-IL-6i and control hMSC-emp, not even in the presence of exogenous IL-6 (20 ng/ml) (Supp. Figure 2).
Loss of IL-6 expression did not modify the adherence capacity of hMSCs, as the number of adhered cells did not change 24 h after plating (data not shown). Thus, we excluded differences in basal apoptosis or adherence capacity as a cause for the evident differences in the immunoregulatory capacity of hMSC-IL6i.
We next reasoned that the addition of exogenous IL-6 to cell cultures should be able to recover the loss of immunoregulatory capacity observed in hMSC-IL6i. Surprisingly, however, the addition of rhIL-6 to the cultures failed to reverse the immunosuppressive phenotype of hMSC-IL-6i (Fig. 2D). Moreover, inactivating extracellular IL-6 using a specific neutralizing anti-IL6 antibody 26,27 in non-transduced hMSC cultures had no effect on the immunosuppressive process, as lymphocyte proliferation was unchanged as compared with cultures in the absence of the anti-IL6 antibody (Fig. 2E). Overall, these experiments suggest that IL-6 does not play a canonical extracellular autocrine signaling role in this cell model.
Basal prostaglandin E2 secretion is enhanced in hMSC-IL-6i. Among the mechanisms proposed to mediate the immunosuppressive function of MSCs, cyclooxygenase-2 (COX-2) activity, through PGE2 production, is consistently reported as one of the most important mediators 6,28,29 . As the production of IL-6 is known to be differentially regulated by PGE2 in various cell types 30,31 , we next investigated PGE2 synthesis in hMSC-IL6i cells. As shown in Fig. 3A, basal levels of PGE2 were significantly higher in hMSC-IL6i than in control cells, correlating with a significant up-regulation of COX2 expression (Fig. 3B), and suggesting that hMSCs have a mechanism to control constitutive IL-6 expression by PGE2. The specificity of the PGE2-secretion was confirmed by blocking its production with indomethacin (a non-selective COX inhibitor) 32 or etoricoxib (a specific COX-2 inhibitor), which induced a decrease of basal PGE2 in hMSC-IL6i, reaching the level of control cells (Fig. 3A). These results suggest an increase of basal COX-2 activity in hMSC mediated by the reduction of IL-6 levels (hMSC-IL6i). However, when we examined the immunoregulatory capacity of hMSCs after treatment with COX-2 inhibitors, we observed that in both control and hMSC-IL6i cells, treatment with COX-2 inhibitors induced a similar decrease in immunoregulatory capacity (Fig. 3C). Overall, these data indicate that PGE2 is not related to the impairment of the immunosuppressive capacity of hMSC-IL6i.
Importance of IL-6 in the cell cycle progression. We next evaluated the effect of IL-6 on hMSC proliferation. To do this, 50% of hMSCs were transduced with the empty or IL6i viral vector and then re-plated at low density (500 cells/cm 2) , and cell growth rate was determined by analysis of the ratio of GFP+ to GFP-cells in the culture (Fig. 4A). As expected, hMSC-emp maintained the same ratio of GFP expression along the culture period of 14 days, indicating that viral integration had no effect on cell proliferation. By contrast, hMSC-IL6i failed to increase in number, which resulted in an overgrowth of non-transduced cells at confluence (Fig. 4A). This phenotype was unaffected by the presence of different concentrations of rhIL-6 (Fig. 4B). The finding that the unmodified hMSCs are able to overgrow hMSC-IL-6i even in the presence of high concentrations of rhIL-6 suggests that the proliferation of hMSCs is dependent of intracellular rather than extracellular IL-6. To confirm this, we analyzed the effect of IL-6 silencing on the cell cycle. Flow cytometry analysis showed that the percentage of cells in the active phases of the cell cycle (S/G2/M) was significantly lower for hMSC-IL-6i than for control www.nature.com/scientificreports/ cells, and the cell cycle was mainly blocked in G0/G1 in the former (Fig. 4C). These results suggest that the loss of IL-6 expression in hMSCs not only affects immunosuppression, but also impairs their normal proliferation. www.nature.com/scientificreports/ To explore this further, we analyzed KI-67 expression in hMSC-emp and hMSC-IL6i. KI-67 protein is present almost exclusively in the G1/S/G2/M phases, and is a very useful marker for recognizing dividing cells. As before, 50% of the hMSCs were transduced with the silencing/control vectors and the expression of KI-67 was analyzed by immunofluorescence in GFP+ and GFP-cells (Fig. 4D). We failed to observe differences in KI-67 expression between GFP-cells of hMSC-emp and hMSC-IL6i. However, we observed a significant decrease in KI-67 expression in GFP+ hMSC-IL6i cells, confirming the defect in proliferation. These results highlight the inability of hMSC-IL6i to progress through the cell cycle and this phenomenon does not involve extracellular IL-6, as all cells (both GFP+ and GFP-) were in the presence of the same amount of soluble IL-6 and only hMSC-IL6i cells showed impaired proliferation. Finally, we stimulated hMSC-IL6i with TNF-α (15 ng/mL) aiming to drive IL-6 expression (Fig. 4E). We found that this treatment allowed a significant recovery of hMSC-IL6i proliferation (Fig. 4F).

IL-6 silencing in hMSC-IL6i cells impedes ERK1/2/cyclin D1 pathway signaling. Cyclin D1 is
involved in the control of the cell cycle G0/G1/S transition and a defect in its expression in hMSCs alters the normal process of cell division 33 . Given our results, we next analyzed cyclin D1 levels in hMSC-IL6i cells by western blotting. As shown in Fig. 5A, the level of cyclin D1 was significantly lower in hMSC-IL6i than in hMSC-emp.
Remarkably, the addition of rhIL-6 (20 ng/ml) to the culture medium failed to recover cyclin D1 expression (Fig. 5A). We previously demonstrated that ERK2 (extracellular signal-regulated kinase 2) was a key transcription factor in the proliferation of hMSC, partly through its control of cyclin D1 expression 33 . To analyze the activation state of ERK2, we monitored the levels of ERK1/2 phosphorylation in hMSC-emp and hMSC-IL6i in the presence or not of rhIL-6 (20 ng/ml) or TNF-α (15 ng/ml). Results shown in Fig. 5B demonstrate a decrease  www.nature.com/scientificreports/ of ERK2 phosphorylation in hMSC-IL6i, in both conditions. To confirm that intracellular IL-6 plays a critical role in hMSC proliferation through cyclin D1, we investigated whether the increase in IL-6 stimulated by TNFα in hMSC-IL6i (see Fig. 1) was capable of recovering cyclin D1 levels. We observed a very significant increase in cyclin D1 expression in cells stimulated for 48 h with TNF-α (Fig. 5C), which was in sharp contrast to the failure of exogenous rhIL-6 to modify cyclin D1 level (Fig. 5A). These data allow us to conclude that IL-6 is related to the control of cell cycle progression through ERK1/2 and by an exclusively intracellular signaling pathway.

Discussion
hMSCs possess immunosuppressive capabilities, which endow them potential to treat inflammatory diseases 1 . hMSCs have capacity to regulate many aspects of T-cell response, such as proliferation, survival and differentiation. Numerous molecules secreted by hMSC and/or T-cells are known to be been involved in the regulation of hMSC immunoregulation 1 ; however, the controlling mechanisms are not fully understood. We recently described the importance of the NF-κB pathway in hMSC physiology. Here, we investigated the possible involvement of IL-6, a pleiotropic cytokine whose expression is under control of NF-κB transcription factor, in hMSC immunoregulation. Using a gene silencing approach, we show that the inhibition of IL-6 expression significantly impairs hMSC immunoregulatory functions. Interestingly, the effect of IL-6 is not due to its secretion into the extracellular milieu, since the addition of rhIL-6 (even concentrations up to 100 ng/ml) to the cultures could not reverse the observed phenotype. Numerous studies have investigated the involvement of soluble factors in the regulation of hMSC function, of which PGE2 is consistently described as one of the most important 34 . Accordingly, we were intrigued by the unexpected results showing that PGE2 secretion was consistently elevated in hMSC-IL6i cells, particularly since it has been described that the immunomodulatory function of hMSCs is partially attributed to IL-6-dependent secretion of PGE2, most likely through the positive regulation of COX-2 activity by IL-6 35 . Indeed, PGE2 secretion was significantly reduced in IL-6-deficient MSCs, which translated into a poor immunosuppressive ability in a collagen-induced experimental arthritis mouse model 35 . The discrepancy between these observations and ours is most likely due to the differences in the MSCs used. Moreover, in the aforementioned study by Bouffi et al 35 , C57BL/6 mice deficient for IL-6 were used, whereas in our study the expression of IL-6 was considerably reduced but not abolished. Several works have also implicated PGE2 in IL-6 production 36,37 . Therefore, we can hypothesize that the hMSC-IL6i cells, through a mechanism that remains to be determined, attempt to compensate for the loss or reduction in IL-6 expression by increasing COX2 expression and PGE2 secretion. In this line, we consistently observed a slight increase in NF-κB activity in hMSC-IL6i cells (data not shown). Nevertheless, these compensatory mechanisms do not seem to be sufficient to restore IL-6 expression and recover the immunoregulatory function of hMSCs. Further work will be necessary to determine the controlling mechanisms in the increase of PGE2 secretion. www.nature.com/scientificreports/ Our experiments demonstrate the involvement of IL-6 in the control of hMSC proliferation through an intracellular mechanism. As far as we know, this phenomenon has never been described in primary cell lines although other groups have shown similar phenotypes in tumor-derived cell lines including renal carcinoma 38 , choriocarcinoma 20 and melanoma 39 , in which reduced IL-6 expression slows their proliferation, whereas blocking IL-6 or gp-130 using specific antibodies do not affect cellular growth. We found that in hMSCs this proliferative effect is due to a defect of cell cycle progression that is rescued by partially recovering the levels of intracellular IL-6 expression through stimulation with TNF-α. In a previous study, we demonstrated that ERK2 is a key transcription factor in the proliferation of hMSCs, in part through its control of cyclin D1 transcription and, therefore, expression 33 . Cyclin D1 is a positive regulator of the cell cycle and promotes G1 to S phase transition in cooperation with CDK4 or 6. Since the protein level of cyclin D1 reflects cell cycle progression, the rates of protein production and degradation are strictly regulated at the level of the transcription and protein degradation 40 . Indeed, cyclin D1 is highly labile, with a half-life of 10-30 min, and undergoes polyubiquitination and proteasomal degradation 40 . Here, we demonstrate that in the absence of IL-6, the level of cyclin D1 is significantly decreased. This might also be due, at least in part, to inhibition of the MAPK and/or AKT pathways. Further work will be necessary to understand how the loss of IL-6 induces a decrease of ERK1/2 phosphorylation in hMSCs.
Overall, our findings reveal that IL-6 is a pivotal factor in the proliferation of hMSCs, but highlight that its effects occur at the intracellular and not at the extracellular level. Other studies have suggested that IL-6 could act as an autocrine/intracrine growth factor interacting with its specific receptors within the cell and not at the cell surface 20,38,41 . More work is needed to identify the intracellular receptors for IL-6. In this regard, our preliminary experiments suggest that small quantities of IL-6 receptor are present in the cytoplasm of some cells (data not shown). If confirmed, IL-6 may activate different transduction pathways through an interaction with its receptor in intracellular compartments. Our results open the door for further research into whether other interleukins can signal without the need for secretion, which would be a considerable advance in understanding the autocrine mechanisms of cellular regulation.

Conclusions
IL-6 is synthesized and secreted by many cell types including human mesenchymal stromal cells. It is often referred to as a pleiotropic cytokine that influences numerous cell types, with multiple biological activities. In this study, we have examined the role of this interleukin in the homeostasis of hMSC. Our results show that IL-6 is essential for the proliferation of stromal cells and their immunosuppression capacity. Nevertheless, our most relevant finding relates to a previously undescribed signaling mechanism of IL-6-driven MSC homeostasis. Our data show that changes to the extracellular levels of IL-6 do not resemble the phenotype observed when modifying intracellular expression, indicating that IL-6 signals intracellularly in hMSC. Further research is necessary to integrate the intracellular signaling described in this study with current knowledge on the role of this interleukin in vivo. Incorporating a new signaling mechanism into the understanding of IL-6 (or other interleukins) biology will be a huge step forward in unravelling cytokine-mediated intercommunication in human biology.