Immunomodulatory Drugs in the Context of Autologous Hematopoietic Stem Cell Transplantation Associate With Reduced Pro-tumor T Cell Subsets in Multiple Myeloma

Immunomodulatory drugs (IMiDs) are effective therapeutics for multiple myeloma (MM), where in different clinical settings they exert their function both directly on MM cells and indirectly by modulating immune cell subsets, although with not completely defined mechanisms. Here we studied the role of IMiDs in the context of autologous hematopoietic stem cell transplantation on the T cell subset distribution in the bone marrow of newly diagnosed MM patients. We found that after transplantation pro-tumor Th17-Th1 and Th22 cells and their related cytokines were lower in patients treated with IMiDs during induction chemotherapy compared to untreated patients. Of note, lower levels of IL-17, IL-22, and related IL-6, TNF-α, IL-1β, and IL-23 in the bone marrow sera correlated with treatment with IMiDs and favorable clinical outcome. Collectively, our results suggest a novel anti-inflammatory role for IMiDs in MM.


INTRODUCTION
Multiple Myeloma (MM) is a plasma cell neoplastic disorder primarily localized in the bone marrow (BM), where interactions between neoplastic cells and cells within the tumor microenvironment (i.e., mesenchymal stromal cells, endothelial cells, osteoblasts, osteoclasts, and immune cells) mediate disease development and progression (1)(2)(3).
Th cells play fundamental regulatory function in adaptive immune responses and in antitumor immunity being, according to their secretory cytokine profile, either anti-tumor or pro-tumor (4). In MM imbalanced Th cell polarization and subset distribution in the BM niche largely impact on disease progression (5).
Immunomodulatory drugs (IMiDs), namely thalidomide and its derivatives lenalidomide and pomalidomide, are major therapeutics in the treatment of MM (6), where their efficacy has been demonstrated in newly diagnosed patients either eligible or ineligible for autologous hematopoietic stem cell transplantation (ASCT), in the maintenance setting after ASCT and in refractory/relapsed disease (6).
In MM, in addition to anti-proliferative and pro-apoptotic effects on malignant cells, IMiDs exert immune regulatory function and interfere with tumor microenvironment interactions (7). In vitro IMiDs enhanced T cell proliferation, IL-2 and IFN-γ secretion and NK cell activation (8,9), lenalidomide improved immune checkpoint blockade-induced immune responses (10) and inhibited T regulatory cell (Treg) proliferation and suppressor function (11). In vivo lenalidomide augmented (i) vaccine responses and endogenous anti-tumor immunity (12), (ii) the number of central and effector memory CD8 + T cells, Tregs and CD14 + CD15 + myeloid derived suppressor cells in patients that received lenalidomide as monotherapy or in combination with other treatments (13), (iii) the number of Tregs in patients in consolidation/maintenance therapy (14,15), (iv) anti-myeloma specific T cell responses in patients that received lenalidomide as consolidation therapy after ASCT (16), and (v) the number of IFN-γ and IL-21 producing cells in the setting of maintenance therapy (17). Treatment with lenalidomide was also associated with impaired longterm thymic reconstitution and decreased number of CD4 + and CD8 + effector terminally differentiated T cells (14,15) and reduced PD-1 expression on T cells in the maintenance setting (18).
A prospective randomized trial comparing induction regimens prior to ASCT including or not thalidomide (i.e., bortezomib-thalidomide-dexamethasone) vs. bortezomibcyclophosphamide-dexamethasone) recently reported significantly higher overall clinical response rate in the bortezomib-thalidomide-dexamethasone arm (19).
In this study, we evaluated in newly diagnosed MM patients the impact of IMiDs used in the induction chemotherapy preceding ASCT on the distribution of T cell subsets and related cytokines in the BM after transplantation and whether changes in those immunological parameters correlated with the clinical outcome.

Subjects and Samples
Forty-four newly diagnosed MM patients, who had been hospitalized at the Hematology Department of our Institution and received ASCT as first line therapy, were selected for the study. The Institutional Ethics Committee (Comitato Etico Fondazione Centro San Raffaele, Istituto Scientifico Ospedale San Raffaele) had approved the study protocol and written informed consent was obtained from all donors. Clinical data and information on the induction chemotherapy received by each patient are reported in Table 1. BM mononuclear cells were isolated by density gradient centrifugation with Ficoll-Paque TM Abbreviations: ASCT, autologous stem cell transplantation; BM, bone marrow; IMiDs, immunomodulatory drugs; MM, multiple myeloma; Treg, T regulatory cells.
Plus (GE Healthcare, Uppsala, Sweden), frozen in fetal bovine serum (Lonza, Milan, Italy) + 10% DMSO (Sigma-Aldrich, Milan, Italy), and stored according to standardized operating procedures by the Institutional Biobank. BM mononuclear cells were used after thawing and viable cell counting. BM sera were taken and collected according to standardized operating procedures by the Institutional Biobank. Briefly, non-heparinized BM blood (5-7 ml) was incubated for 1 h at room temperature to achieve complete clotting. Then samples were centrifuged at 1,600 g without brake for 10 min at 4 • C and supernatants transferred in cryovials and stored in liquid nitrogen until use.

Statistical Analysis
Statistical significance was determined with Mann Whitney U test and Wilcoxon Signed-Rank Test, as reported in the figure legends. In addition, to evaluate the equality of variances for the values in the IMiD-treated vs. the IMiD-untreated group Frontiers in Immunology | www.frontiersin.org of patients we performed the Levene's test with no statistical evidence to support difference between the variance of the distributions. Statistical analyses were performed with an alpha level of 5% using GraphPad Prism version 5.0 for Mac (GraphPad Software), thus values of p < 0.05 were considered significant.

RESULTS
We studied forty-four newly diagnosed MM patients, who had received ASCT as frontline therapy and whose clinical features are detailed in Table 1. Patients were administered an induction chemotherapy, most often consisting of a bortezomib-based poly-chemotherapy (including or not thalidomide or, less often, lenalidomide) followed by stem cell mobilization with cyclophosphamide and G-CSF, high-dose melphalan and a single or tandem ASCT (see Table 1 for dosage and schedule details). BM mononuclear cells were analyzed by flow cytometry for intracellular cytokine staining of the T cell subset-distinctive cytokine patterns or directly for Treg specific markers within the T cell fraction (see Figure 1 for gating strategy).
Firstly, we measured the frequency of distinct effector cytokine-secreting T and Treg cell subsets in paired samples at diagnosis and 3 months after ASCT (i.e., a time point at which immune cells are mostly reconstituted), to evaluate potential changes of the T cell subset distribution occurring early after transplantation as a consequence of the complete therapy received by the patients (i.e., induction regimen including or not IMiDs, stem cell mobilization, chemotherapy and ASCT). We found that T cells secreting Th1 (i.e., IFN-γ and TNF-α) and Th2 (i.e., IL-13, IL-4, and IL-5) cytokines were significantly increased after ASCT compared to those at diagnosis as well as T cells coproducing IL-17 and IFN-γ, whereas T cells secreting IL-17, IL-22, and Tregs (CD4 + CD127 − CD25 + FoxP3 + ) were not (Figure 2). Secondly, to evaluate the impact of IMiDs used in the induction regimen, we focused on the frequency of the T cell subsets in the post-ASCT time point and performed analyses in patients grouped based on the absence or the presence of IMiDs in the induction chemotherapy. We found that IL-17 + IFN-γ + (i.e., mostly Th17-Th1, see Figure S1) and IL-22 + IL-17 − (i.e., mostly Th22, see Figure S1) T cells were significantly lower in the IMiD-treated compared to untreated patients ( Figure 3A). Remarkably, increased frequency of Th17-Th1 and Th22 cell subsets correlated with development of symptomatic MM and worse prognosis (25,26), respectively.  Indeed, poly-functional Th17-Th1 were found increased in the BM of MM patients compared to those of patients with pre-neoplastic gammopathy (25), and we previously reported increased frequency of Th22 cells in the blood and BM of patients with stage III disease at diagnosis and refractory/relapsed disease compared to those of asymptomatic patients or patients with stage I/II disease (26). Total IL-17 + and IL-22 + T cells were also lower within BM T cells from IMiD-treated compared to untreated patients ( Figure 3A). In agreement with previous reports (9,13), IFN-γ + secreting cells were higher in the IMiDtreated compared to the untreated group (Figure 3A), whereas T cells secreting the other cytokines (i.e., IL-13, IL-4, IL-5, and Frontiers in Immunology | www.frontiersin.org TNF-α) and Tregs were comparable between the two groups ( Figure 3A).
We then measured the levels of the same cytokines in the BM sera. We found that, as a result of the complete therapy received by the patients, in paired samples at diagnosis and at 3 months  after ASCT IFN-γ, IL-13, IL-4, IL-5, TNF-α, IL-17, and IL-22 were all significantly increased at 3 months after transplantation ( Figure S2A). In agreement with the T cell data, when we considered the levels of the cytokines in the BM sera after ASCT in patients grouped based on the absence or the presence of IMiDs in the induction chemotherapy, we found that the levels of IL-17 and IL-22 were significantly lower in the BM sera of IMiD-treated vs. untreated patients, and those of Th2 cytokines (i.e., IL-13, IL-4 and IL-5) were not (Figure 3B). On the contrary, the levels of IFN-γ and TNF-α were both significantly lower in IMiD-treated patients (Figure 3B), suggesting that the total amount of these cytokines in the BM sera after transplantation possibly depends on immune cells, other than T cells, which may be differentially targeted by the drugs used in the two induction chemotherapy regimens.
Next, we measured the BM serum levels of cytokines implicated in Th17-Th1 and Th22 cell polarization/expansion, namely, in addition to TNF-α, IL-6, IL-1β, and IL-23 (27). We did not find any significant change in the level of these cytokines when comparing paired BM sera at diagnosis and at 3 months after ASCT ( Figure S2B). Interestingly, these cytokines were all significantly reduced in IMiD-treated vs. untreated patients in the BM sera after ASCT ( Figure 3C).
Together, these data show that IMiD-including therapies are associated with a lower frequency in the BM at 3 months after transplantation of pro-tumor Th17-Th1 and Th22 cells and of the serum levels of their distinctive cytokines and cytokines implicated in their polarization/expansion.
Lastly, to assess potential correlations between the immunological changes observed, the clinical status and the use of IMiDs in the induction chemotherapy, we analyzed the levels of cytokines in patients grouped according to their clinical status at 3 months after ASCT ( Table 1). We found that IL-17, IL-22, TNF-α, IL-6, IL-23, and IL-1β were all significantly lower in patients with complete response compared to patients with refractory disease or partial response, whereas IFN-γ did not significantly change ( Figure 4A). IL-6, IL-23, and IL-1β were also significantly lower in patients with complete response compared to those with very good partial response ( Figure 4A).
As the cohort of patients that received IMiDs had better clinical outcome (complete response in 44 vs. 10% patients, respectively, see Table 1), to exclude that the immunological changes observed were simply the consequence of reduced tumor burden and not associated to IMiD-driven immune modulation, we compared the cytokine profile in IMiDtreated vs. untreated patients within patients with very good partial response (i.e., patients who had achieved very deep responses to therapy, yet still have some residual markers of disease). Although the number of patients compared is relatively small, we found that IL-17, IL-22, IL-6, and IL-1β were significantly lower in IMiD-treated compared to untreated patients ( Figure 4B).
Together, these data suggest that the better clinical responses observed in IMiD-treated patients depend not only on the wellknown anti-proliferative effects of IMiDs on myeloma cells (6) but possibly also on immune regulatory functions, among which the down-modulation of pro-tumor Th17-Th1, Th22 cells and their related cytokines.

CONCLUSIONS
In this study we report that (i) pro-tumor Th17-Th1 and Th22 cells and their related cytokines in the BM after transplantation are lower in patients treated with IMiDs during induction chemotherapy compared to untreated patients and (ii) the levels of the same cytokines are lower in patients with favorable clinical outcome (i.e., complete response) compared to patients with persistent disease (i.e., refractory disease and partial response).
In agreement with previous reports (9, 13) we found higher numbers of IFN-γ secreting T cells in the BM of patients treated with IMiDs in the induction chemotherapy: however, even though the levels of IFN-γ in the BM sera were increased after transplantation compared with those at diagnosis, we did not find a significant correlation between the levels of IFN-γ and the clinical outcome after transplantation.
Conflicting results are reported concerning the role of IMiDs on Tregs: while in vitro lenalidomide reduced the number of Tregs (11), in vivo studies showed an increase in the number of Tregs in different clinical settings (13)(14)(15). In our study we did not find significant changes in the numbers of Tregs neither between diagnosis and after transplantation nor between IMiDtreated and untreated patients. Factors such as the clinical setting in different studies (i.e., consolidation and/or maintenance vs. induction chemotherapy) and the site of investigation (i.e., peripheral vs. BM blood) may account for these differences.
Our findings are particularly relevant because of the protumor role of IL-17 and IL-22 in MM. IL-17 promotes MM cell growth/survival by interaction with the IL-17 receptor expressed on MM cells, and inhibits immune cell functions (28). In addition, expansion of Th17 cells correlates with development of bone lesions (29,30), possibly through IL-17-induced upregulation in BM cells of RANKL, which is a differentiation factor for osteoclasts [rev. in (31)]. A reduction in the RANKL/osteoprotegerin ratio and bone osteolysis was reported in MM patients with relapsed/refractory disease treated with thalidomide-containing regimen (32,33). In agreement with this report, we found that the percentage of nuclear magnetic resonance and/or positron emission tomography negative patients as from evaluation after transplantation was superior within the IMiD-treated compared to the untreated group (data not shown). Whether the effect on bone disease following treatment with IMiDs might at least partially depend on IL-17 modulation through the RANKL signaling as well as other mechanisms of bone remodeling [i.e., Wnt (34) and LIGHT/TNFSF14 (35) signaling pathways] might warrant further ad hoc investigation. Concerning IL-22, we previously reported that this cytokine directly increases MM cell growth and resistance to drug-induced cell death by binding to its IL22RA1 receptor, which is aberrantly expressed on a fraction of primary MM cells (26).
The immune regulation exerted by IMiDs has actually two sides: they co-stimulate activation of immune responses, as in hematologic cancers (6), but they can also dampen inflammatory reactions, as in autoimmune and inflammatory disorders (36). Such dual role conceivably depends on different immune cell types targeted, such as T and NK cells and antigen presenting cells, respectively (36). It is tempting to speculate that the use of IMiDs during induction chemotherapy might have favored the development and persistence of an anti-inflammatory BM milieu responsible for the reduced frequency of Th17-Th1 and Th22 cells, possibly through modulation of BM resident cells (i.e., mesenchymal stromal cells) and/or immune cells persisting in the BM or transferred within the autologous progenitor cell reinfusion.
In summary, our results contribute to the characterization of the immunomodulatory effects exerted by IMiDs in MM and should be taken into consideration for the implementation of new therapeutic strategies targeting IL-17 and IL-22 to be combined with drugs already used in MM treatment, especially in refractory/relapsed disease stages.

AUTHOR CONTRIBUTIONS
GDL designed research, performed research, analyzed data and wrote the manuscript. MM and FC contributed to patient selection and clinical data collection. SH performed research. CT contributed sample collection and storage. AMP contributed to statistical analysis design. CB contributed to data analysis. MPP designed research, analyzed data and wrote the manuscript.