Background
Treatment of cancer aims at the deletion of all malignant cells. Besides minimizing the bulk of tumor cells by conventional therapy, the cytotoxic ability of activated T cells is central to tumor eradication and cancer cure. T cell anergy, exhaustion, and/or senescence are characterized by elevated levels of multiple inhibitory molecules, impaired effector cytokine production, loss of proliferative capacity, impaired cytotoxicity, and altered use of key transcription factors including T-bet (reviewed in [
1‐
3]). T cell exhaustion occurs after chronic antigenic exposure and prevents optimal control of tumors. Modulation of molecules and pathways overexpressed in the exhaustion phase could reverse this dysfunctional state and reinvigorate immune response [
4,
5]. Senescent T cells are late differentiated memory/effector T cells which lack CD28 and gain CD57 and regulatory receptor expression. They rather express CD45RA than CD45RO, are in cell-cycle arrest, and have enhanced secretion of inflammatory cytokines [
6]. Most studied inhibitory molecules which function as negative regulators of lymphocyte activation are PD-1 (CD279) and CTLA-4 (CD152). Natural killer cell receptor 2B4 (CD244), glycoprotein CD160, TIM3, Lag3, and many others have also been shown to contribute essentially to the regulation of T cell activity [
2,
7]. Blocking immune checkpoint molecules achieved promising results in the treatment of solid cancers; however, targeting these molecules in hematologic malignancies is relatively understudied [
8‐
10]. First promising results with an anti-PD-1 antibody in the treatment of refractory Hodgkin’s lymphoma were recently published [
11‐
13], and studies are under way to investigate the effectiveness of combinations with low-dose glucocorticoids. Recently, data on immune checkpoints such as PD-1 expression of myeloma patients have been published. In addition, the potential of PD-1 blocking antibodies to reactivate diseased T cells were investigated in vitro and proved this molecule as promising candidate for immunotherapy [
14‐
16]. In vivo, however, these PD-1 inhibitors appear to work only in combination with immunomodulatory drugs in myeloma patients, and first positive results have been presented at the ASH 2015 meeting by J. San Miguel and A.Z. Badros.
Multiple myeloma is a hematologic cancer characterized by the accumulation of malignant plasma cells within the bone marrow. Despite the advances in therapy by the introduction of novel immunomodulatory agents and proteasome inhibitors, this cancer remains largely incurable. The immunogenic nature of multiple myeloma is underlined by the observation of disease remission after allogeneic stem cell transplantation or donor lymphocyte infusion, possibly caused by graft-versus-myeloma effects [
17,
18]. Delineating the tumor-driven defects of the immune system in myeloma in more depth will contribute to the development of novel immuno-therapeutical strategies.
Defects in T cell distribution and function have been consistently documented in this disease, including a reduction of peripheral blood CD4 and CD8 T cells, inversion of the CD4/CD8 ratio, abnormal Th1/Th2 ratio, downregulation of signal transduction components, and abnormal T cell responses [
19‐
22]. Intriguingly, the majority of studies so far focused on immunological alterations in the peripheral blood of myeloma patients despite the fact that myeloma cells almost exclusively grow and spread within the bone marrow. So far, little is known regarding the composition and activity of the immune system in myeloma bone marrow. In this study, we investigated the expression of inhibitory molecules on effector T cells as well as their function in response to in vitro stimulation by antibodies against CD3/CD28. We studied both T cells from peripheral blood and from the bone marrow of myeloma patients and healthy, age-matched donors in order to elucidate systemic and local tumor-mediated alterations.
We demonstrate that, especially at the tumor site, both T cell exhaustion and T cell senescence might play an important role in supporting tumor growth. Moreover, we show data on the T cell composition in refractory patients and how therapy might change T cell suppressive molecules.
Methods
Sample collection
Blood samples and bone marrow aspirates from the posterior iliac crest from the same patient were obtained from 16 newly diagnosed patients (MM naive) and 6 patients treated with immunomodulatory drugs and dexamethasone (MM Tx) at the Department of Internal Medicine, University Hospital of Innsbruck, Austria. Peripheral blood samples and bone marrow aspirates from the proximal femur of the same individuals were taken from age-matched healthy donors in the course of hip arthroplasty at the Department of Orthopedic Surgery, University Hospital of Innsbruck (
n = 12). According to a recent publication, T cell subset distribution is comparable between samples from the posterior iliac crest and proximal femur [
23]. Healthy subjects were screened for the presence of viral infections and any bone marrow abnormalities and did not receive immunomodulatory drugs or suffer from diseases known to influence the immune system, including autoimmune diseases or cancer.
Cell isolation from human PB and BM aspirates
Peripheral blood mononuclear cells (PBMC) and bone marrow mononuclear cells (BMMC) were separated by density gradient centrifugation (GE Healthcare Life Sciences) and either used immediately (for flow cytometric analyses) or cryopreserved for later functional analysis. Recovery rates from frozen T cells were above 85 %.
Immunofluorescence staining and flow cytometric analysis
Staining of surface antigens was performed in erythrocyte-free samples (blood samples and bone marrow aspirates) by multicolour staining using fluorescently labeled CD3 (clone UCHT1), CD4 (RPA-T4), CD8 (RPA-T8), CD28 (CD28.2), CD45 (2D1), CD45RA (HI100), CD57 (NK-1), CD62L (DREG-56), CD152/CTLA-4 (BNI3), CD160 (BY55), CD244/2B4 (2-69), CD279/PD1 (MI-H4) (all from BD, Heidelberg, Germany). For staining of intracellular proteins, PBMC and BMMC were stimulated with anti-CD3 (2 μg/ml) and anti-CD28 (5 μg/ml) (both from eBioscience/Affymetrix) for 6 or 24 h. Ten micrograms per milliliter Brefeldin A (Sigma-Aldrich) were added to the cell cultures for the last 4 h of incubation. Intracellular staining was carried out using CytoFix/CytoPerm kit, according to manufacturer’s instructions (BD). All antibodies against cytokines (IL-2, clone 5344.111; IFNγ, B27), T-bet (O4-46), and CD107a (H4A3) were purchased from BD. Flow cytometry was performed on a BD FACS Canto II flow cytometer with subsequent analysis using FACS DIVA Software 7.0. The analysis was performed after gating on single viable cells after 7-AAD staining.
T cell proliferation and degranulation
Cell proliferation was measured by culturing BMMC at a density of 2 × 106/ml in RPMI1640/10 % FCS together with anti-CD3 (2 μg/ml) and anti-CD28 (5 μg/ml) antibodies. 3H-thymidine (Hartmann Analytic) was added for the last 16 h of stimulation, cells were harvested, and 3H-thymidine incorporation was measured on a ß-counter (Beckman Coulter, USA). The proliferation index was calculated as ratio of stimulated T cells and unstimulated controls. For assessment of degranulation, cells were stimulated with anti-CD3 and anti-CD28 as above with the addition of 10 μg/ml anti-CD107a antibody for 6 h at 37 °C. For assessment of CD8+ T cell activation/proliferation, 5 × 106 cells/ml (BMMC) were washed and stained with 0.4 μM carboxyfluorescein succinimidyl ester (CFSE; eBioscience/Affymetrix) for 5 min at room temperature in the dark. After quenching with FCS and washing with RPMI1640 and once with PBS, the cells were diluted at 2 × 106/ml in RPMI1640/10 % FCS and stimulated with anti-CD3 and anti-CD28 antibodies for 72 h as described above.
Statistical analysis
For all data sets which could be accurately modeled by a Gaussian distribution an unpaired t test was used for analysis of differences between groups; otherwise, the Mann-Whitney U test was used. P values of less than 0.05 were considered statistically significant (significance levels *p < 0.05, **p < 0.01, ***p < 0.001, n.s. not significant).
Discussion
In this study, we investigated the nature of the T cell defects in multiple myeloma in more detail and compared our findings to healthy, age-matched donor samples. Such a thorough study of BM and PB samples from myeloma patients and healthy individuals has, to our knowledge, not been performed so far. Our investigation of systemic and local changes of the expression of checkpoint inhibitor molecules PD-1, CTLA-4, CD160, and 2B4 revealed significant upregulation of all tested markers especially in the bone marrow, suggesting that the immunosuppressive tumor microenvironment fuels immune escape in multiple myeloma.
Expression and ligation of immune-checkpoint molecules downregulate T cell responses and thereby maintain self-tolerance. The use of monoclonal antibodies (i.e., anti-PD-1) to disrupt receptor-ligand interactions has already shown remarkable results in the therapy of several solid tumors [
38‐
40] and is currently a fast progressing field also in hematological cancers including myeloma. A major obstacle is the profoundly suppressive tumor microenvironment which limits actions of immune cells against cancer [
41]. Coupled with the finding that myeloma cells express PD-L1 which can even be further upregulated by stroma cell contact, the PD-1 pathway has been shown to contribute to the aggressiveness of this disease in a subset of patients [
14]. Although in vitro results strongly support PD-1 inhibition as novel, effective therapy [
16], these drugs failed to induce major responses in myeloma in vivo so far, as compared to other B cell malignancies [
13]. Here, the tumor microenvironment emerges as important additional target and ongoing studies which combine PD-1 inhibitors with immunomodulatory agents (i.e. lenalidomide and pomalidomide) appear to be more successful. Lenalidomide has been shown to induce immuno-activating changes in the myeloma microenvironment [
15,
42] and to downregulate PD-L1 expression [
14]. However, since PD-L1 expression per se varies strongly in myeloma [
43], other checkpoint inhibitors might become equally important targets. Studies in the 5T33 murine myeloma model already showed the efficacy of lenalidomide in combination with blockade of other immune checkpoints (e.g. CTLA-4, LAG-3, TIM-3, and combinations thereof [
44]). Our analyses show that CTLA-4, 2B4, and CD160 expression remains high even after IMiD-containing therapies and thus could constitute additional novel targets. CTLA-4 competes with the immunostimulatory receptor CD28 for the binding of CD80/CD86 proteins. Tumor engagement of the CTLA-4 pathway may therefore dampen the immune response in the microenvironment resulting in an inappropriate T cell costimulation [
45]. In fact, ipilimumab, an antagonistic antibody of CTLA-4, was the first immune-checkpoint inhibitor studied in lymphoid malignancies with promising results [
46]. However, investigations in multiple myeloma are still pending. CD160, one of the five ligands of herpes virus entrance mediator (HVEM) has the potential to shift immune response towards exhaustion, and its expression has been shown to be independent of PD-1 [
26]. Moreover, CD160 blockade has not been investigated in myeloma so far. CD28 downregulation and 2B4 upregulation in the presence of CTLA-4 has been shown in a virus-specific T cell model [
47]. We here see a similar situation in myeloma bone marrow, i.e., a high proportion of CD8+ T cells which lost CD28 expression and concomitantly gained CTLA-4 expression and upregulated 2B4. These regulations could be the result of chronic stimulation leading to enhanced T cell exhaustion. The CMV status of the patients might play a role here, however, recent papers showed defects in T cell function irrespective of CMV serostatus [
48,
49].
Additionally, to our checkpoint molecule analysis, we detected an altered T cell subset distribution at the tumor site, leading to a significant increase of the effector T cell population at the expense of the memory T cell population. These CD8+ BM T cells would generally favor anti-cancer immune responses. However, these cells also displayed upregulation of checkpoint molecules, reduced cytokine production, reduced ability to degranulate in response to T cell stimulation, and reduced proliferative capacity. Thus, without further activations, these T cells are most likely ineffective in tumor immune surveillance. The loss of CD28 expression is a sign of T cell aging in healthy individuals [
30], and we found it further enhanced in myeloma patients. In addition, we observed that myeloma CD8+CD28− T cells concomitantly expressed CD57, pointing to a high accumulation of a late differentiated, senescent T cell population in the BM. CD8+CD28–CD57+ T cells from peripheral blood have been shown to play an active pro-tumor role via suppressing proliferation of responder T cells [
34], and removal of these cells in vitro restored T cell proliferative capacity [
50,
51]. Moreover, CD8+CD28–CD57+ T lymphocyte clones in the PB of myeloma patients have been associated with progressive and advanced stage disease [
52]. In our analyses, we found that the percentage of CD57+ T cells in the bone marrow of treated patients was reduced suggesting a window of opportunity for effective immune modulatory treatments. However, sample size needs to be extended to be able to draw solid conclusions about the potential role of CD8+CD28–CD57+ T cells in myeloma. Clearly, due to the severe impairment of T cell function, reactivation of T cells has to be built on several columns. An additional promising approach demonstrated that exogenous IL-7 added to T cells co-cultured with tumor cells inhibited loss of CD28, a feature of replicative senescence, and allowed normal proliferative capacity and IL-2 production [
53]. IL-7 can also induce telomerase activity [
54]. Interestingly, an ongoing phase I clinical trial using the non-glycosylated form of human IL-7 shows promise, in that it seems to cause the expansion of naive and memory CD4 and CD8 T cell populations (NTC00062049). This may be important for myeloma patients as especially memory T cells are reduced at the tumor site. Together with IL-15, which is a critical factor for development, proliferation and activation of natural killer cells and CD8+ memory T cells, it may effectively contribute to fight the tumor. In preclinical studies, this cytokine exhibited potent antitumor activities against established tumors in animal models, showing that CD8+ T cells play a pivotal role in the anti-myeloma effect of IL-15 agonists [
55,
56].
Acknowledgements
We thank Dr. Wolfgang Willenbacher and Mag. Roman Weger for participating in the retrieval of patient data.