Multiple myeloma (MM) cells acquire dormancy and drug resistance via interaction with bone marrow stroma cells (BMSC) in a hypoxic microenvironment. Elucidating the mechanisms underlying the regrowth of dormant clones may contribute to further improvement of the prognosis of MM patients. In this study, we find that the CD180/MD-1 complex, a noncanonical lipopolysaccharide (LPS) receptor, is expressed on MM cells but not on normal counterparts, and its abundance is markedly upregulated under adherent and hypoxic conditions. Bacterial LPS and anti-CD180 antibody, but not other Toll-like receptor ligands, enhanced the growth of MM cells via activation of MAP kinases ERK and JNK in positive correlation with expression levels of CD180. Administration of LPS significantly increased the number of CD180/CD138 double-positive cells in a murine xenograft model when MM cells were inoculated with direct attachment to BMSC. Knockdown of CD180 canceled the LPS response in vitro and in vivo. Promoter analyses identified IKZF1 (Ikaros) as a pivotal transcriptional activator of the CD180 gene. Both cell adhesion and hypoxia activated transcription of the CD180 gene by increasing Ikaros expression and its binding to the promoter region. Pharmacological targeting of Ikaros by the immunomodulatory drug lenalidomide ameliorated the response of MM cells to LPS in a CD180-dependent manner in vitro and in vivo. Thus, the CD180/MD-1 pathway may represent a novel mechanism of growth regulation of MM cells in a BM milieu and may be a therapeutic target of preventing the regrowth of dormant MM cells.
Significance: This study describes a novel mechanism by which myeloma cells are regulated in the bone marrow, where drug resistance and dormancy can evolve after treatment, with potential therapeutic implications for treating this often untreatable blood cancer. Cancer Res; 78(7); 1766–78. ©2018 AACR.
Multiple myeloma (MM) is characterized by deregulated growth of plasma cells (PC), terminally differentiated B-lymphocytes, in the bone marrow (BM). Initial treatments with proteasome inhibitors and/or immunomodulatory drugs (IMiD) significantly increased the remission rate of MM; however, MM is still one of the most intractable malignancies due to a high incidence of relapse (1, 2). It is widely accepted that relapse stems from dormant and highly drug-resistant clones in the MM compartment. Drug-resistant clones are kept dormant via interaction with bone marrow stroma cells (BMSC) in a hypoxic microenvironment (3–6). Elucidation of the mechanisms underlying the regrowth of dormant clones may prolong remission and ultimately improve the survival of MM patients.
Recently, we have identified a novel epigenetic mechanism of drug resistance in which BMSCs reduce the abundance of trimethylated histone H3 at lysine-27 (H3K27me3), a hallmark of condensed chromatin at silent loci, via EZH2 inactivation to derepress the transcription of several antiapoptotic genes in MM cells (5, 6). The derepressed gene products include CD180, a homolog of Toll-like receptor 4 (TLR4), originally identified as 105 kDa protein (RP105) that imparts the resistance to radiation and corticosteroids in B-lymphocytes (7, 8). CD180 was later demonstrated to mediate the response to bacterial lipopolysaccharide (LPS), similar to TLR4 (9).
TLRs play a pivotal role in sensing and initiating innate immunity (10). The TLR family is composed of 10 TLRs and CD180/RP105, each of which recognizes a specific pathogen-associated molecular pattern, leading to the activation of unique signaling pathways in immune cells. Some of them are also involved in oncogenesis and tumor growth (11). For example, MM cells express a broad range of TLRs and respond to specific ligands for proliferation and survival (12–15). Among the 10 TLRs, TLR4 is most prevalently expressed in MM cells. A study of a large panel of MM patients identified TLR4 overexpression in approximately 6% and its association with a poor treatment outcome (16). However, little is known about the role of CD180/RP105 in MM biology.
In the present study, we found that CD180 is robustly expressed by MM cells but not normal PCs in the BM microenvironment in an Ikaros-dependent manner and senses bacterial LPS to trigger the growth of dormant MM cells. Furthermore, the IMiD lenalidomide could prevent the LPS-triggered activation of dormant clones by targeting CD180, providing a possible rationale for continuous and maintenance therapies of MM with this agent.
Materials and Methods
The reagents used in this study and their sources are LPS (Wako Biochemicals), Pam3Csk (Novus), flagellin (Enzo Life Science), poke weed mitogen (Sigma-Aldrich), lenalidomide, pomalidomide (Selleck Chemicals), and SN-50 peptide (Santa Cruz Biotechnology; ref. 17). Anti-CD180 antibodies with a stimulatory function (MHR73-11) and without a stimulatory function (N2C1) were obtained from BioLegend and GeneTex, respectively (18, 19). LPS, Pam3Csk, flagellin, poke weed mitogen, and SN-50 were dissolved in 0.9% NaCl. Lenalidomide and pomalidomide were dissolved in dimethyl sulfoxide at appropriate concentrations and used at a final dilution of 1/1,000. The final concentration of dimethyl sulfoxide in the culture medium was <0.1%, a concentration that did not affect drug effects or cell growth per se.
Cells and cell culture
We used seven bona fide human MM cell lines, KMS12-BM, RPMI8226, KMS-21, KMS-26, KMS28-BM, KMS-34, and MM.1S, in this study. These were purchased from the Health Science Research Resources Bank, where cell line authenticity and Mycoplasma infection have been routinely checked by DNA fingerprinting and PCR. We used human BM-derived stromal cell lines UBE6T-7 and stroma-NK, which were immortalized by transducing with a telomerase catalytic protein subunit, as BMSCs (5). Primary BM cells were isolated from MM patients at the time of diagnostic procedure and used with or without positive selection of MM cells using CD138 MicroBeads and MACS separation columns (Miltenyi Biotech). Normal PCs were isolated similarly from healthy volunteers (purchased from LONZA). Written informed consent was obtained in accordance with the Declaration of Helsinki and the protocol was approved by the Institutional Review Board of Jichi Medical University.
In vitro coculture system with BMSCs to recreate the BM microenvironment
We used a culture system devised by a modified cell culture insert to analyze the functional interaction between MM cells and BMSCs (5). First, BMSCs were cultured on the reverse side of the polyethylene terephthalate track-etched membrane of a high pore density cell culture insert (35–3,495; Becton-Dickinson) in a 24-well plate (35–3,504; Becton-Dickinson). After obtaining a confluent feeder layer, MM cells were seeded on the upper side of the membrane where the cytoplasmic villi of BMSCs pass through the etched 0.4-μm pores. In another set of conditions, MM cells were seeded on the upper side of a low pore density cell culture insert (35–3,095; Becton-Dickinson). Under this condition, MM cells were physically separated from the stromal layer, providing an adhesion-negative control. We performed the coculture under hypoxic (5% O2) conditions to reproduce the BM microenvironment (20, 21).
Construction and production of lentiviral expression vectors
We used the lentiviral short-hairpin RNA/short-interfering RNA (shRNA/siRNA) expression vector pLL3.7 for knockdown experiments. Oligonucleotides containing siRNA target sequences are shown in Supplementary Table S1. Scrambled sequences were used as controls. We used the lentiviral vector CSII-CMV-MCS-IRES-VENUS (provided by Dr. Hiroyuki Miyoshi, RIKEN BioResource Center) containing the coding region of Ikaros cDNA for gain-of-function experiments. These vectors were cotransfected into 293FT cells with packaging plasmids (Invitrogen) to produce infective lentiviruses in culture supernatants. Lentiviruses were then added to cell suspensions in the presence of 8 μg/mL polybrene and transduced for 24 hours as previously described (5).
We amplified the promoter regions of the CD180 gene (−1,955 to +18, −1,547 to +18, −1,254 to +18, −1,040 to +18, and −384 to +18) using PCR and inserted them into the pGL4.17 firefly luciferase vector (Promega) to generate reporter plasmids. A mutation was inserted at nucleotide positions −1,076 to −1,070 and −1,059 to −1,052 by PCR-based site directed mutagenesis using wild-type (WT) reporter plasmid as a template (for primers, see Supplementary Table S2). We introduced reporter plasmids into KMS12-BM cells along with the pGL4.73 Renilla luciferase vector, which served as a positive control to determine transfection efficiencies, using electroporation. After 48 hours, firefly and Renilla luciferase activities were measured discriminately using the Dual-Luciferase Reporter Assay System (Promega). The promoterless pGL4.17-basic vector was used as a negative control. Luciferase activity was normalized with the internal standard and indicated as a relative ratio to negative controls.
Chromatin immunoprecipitation assays
We used the ChIP-IT Chromatin Immunoprecipitation Kit (Active Motif) to perform chromatin immunoprecipitation assays. In brief, cells were fixed with 1% formaldehyde at 37°C for 5 minutes, and sonicated to obtain chromatin suspensions. After centrifugation, supernatants were incubated with antibodies of interest at 4°C overnight. The mixture was then incubated with protein A-agarose beads at 4°C for 1 hour and centrifuged to collect the beads. DNA fragments bound to the beads were purified with vigorous washing and subjected to PCR. Detailed information on primers, including sequences, corresponding nucleotide positions, and PCR product sizes, is shown in Supplementary Table S3.
Xenograft murine MM model
For ex vivo tracing of tumors, we inoculated a luciferase-expressing subline of RPMI8226 with or without BMSCs in NOD/SCID mice (Charles River Laboratories). Tumor-derived luciferase activity was measured by the IVIS Imaging System with Living Image software (Xenogen; ref. 5). All animal studies were approved by the Institutional Animal Ethics Committee and performed in accordance with the Guide for the Care and Use of Laboratory Animals formulated by the National Academy of Sciences.
Other conventional techniques are described in Supplementary Materials and Methods.
The CD180/MD-1 complex is overexpressed in MM cells but not in normal PCs
In an initial effort to understand the role of CD180 in MM biology, we screened for the expression of TLRs in primary MM samples using the datasets in the Oncomine and GEO databases. We found a significant increase in transcription and DNA copy numbers of the CD180 gene in MM cells relative to normal PCs (Fig. 1A and Supplementary Fig. S1A). CD180 mRNA expression was significantly upregulated in CD138-negative stem-like MM cells (Supplementary Fig. S1B) and PC leukemia (Supplementary Fig. S1C). In contrast, no difference was observed in the expression of other TLRs between MM and normal PCs (Fig. 1A). Previous studies showed that CD180 is a homolog of TLR4 and requires the accessory molecule MD-1 to appear on the cell surface and recognize LPS. In addition, CD180 requires TLR4 or CD19 to transduce signals, because it lacks C-terminal Toll/interleukin-1 intracellular signaling domains (22–24). We therefore examined whether MM cells coexpress MD-1 and TLR4 with CD180 using CD138-positive cells isolated from the BM of MM patients and healthy volunteers. As anticipated, most of primary MM cells expressed CD180 and MD-1 very strongly as well as TLR4 moderately compared with normal peripheral blood mononuclear cells (PBMNC), whereas normal PCs showed only faint expression by semiquantitative RT-PCR (Fig. 1B). Next, we isolated BM mononuclear cells from six MM patients and two healthy volunteers, and immediately stained them with specific antibodies against CD138, CD180, MD-1 and TLR4. We confirmed the coexpression of CD180 and MD-1 in CD138-positive myeloma cells but not in CD138-negative nonmyeloma BM cells or normal PCs (Fig. 1C; Supplementary Fig. S2A). TLR4 was moderately expressed on CD138-positive MM cells (Fig. 1C) and its coexpression with CD180 was detected in a minor fraction of cells (Supplementary Fig. S2B). Similar expression patterns for CD180, MD-1 and TLR4 were obtained in six MM cell lines by RT-PCR (Fig. 1D, top), flow cytometry (Fig. 1D, bottom; Supplementary Fig. S2B), and immunocytochemistry (Fig. 1E). Overall, the CD180/MD-1 complex is highly and universally expressed in primary MM cells as well as MM cell lines. Notably, the signal intensity of CD180 is stronger in freshly isolated primary MM cells than MM cell lines, suggesting the positive impact of BM microenvironment on CD180 expression. In support of this view, the signal intensity of CD180 mRNA gradually declined after isolation from the BM in cultured MM cells (Supplementary Fig. S3A).
LPS enhances myeloma cell growth in positive correlation with the expression levels of CD180 in vitro
We reproduced the BM microenvironment, in which dormant MM cells physically interact with BMSCs under hypoxic conditions, using a coculture system to examine its regulatory effects on CD180 expression. In brief, we cultured three MM cell lines (KMS12-BM, RPMI8226, and KMS-21) with or without direct adhesion to BMSCs (UBE6T-7 and stroma-NK) in a cell culture insert under normoxic (20% O2) or hypoxic (5% O2) conditions. Real-time quantitative RT-PCR (qPCR) analyses revealed that the expressions of both CD180 and MD-1 were markedly upregulated via adhesion to BMSCs under hypoxic conditions (Fig. 2A, top; Supplementary Figs. S3B and S4A). This increase in mRNA expression resulted in the coordinated upregulation of the surface expression of CD180 on MM cells (Fig. 2A, bottom). In addition, CD180 expression, with regard to both mRNA and surface expression, increased in a time-dependent manner under adherent and hypoxic conditions (Supplementary Fig. S4B). In contrast, no obvious changes were detectable in the expression of other TLRs, including TLR1, 2, 4, 5 and 6, in adherent MM cells (Supplementary Fig. S4C). Therefore, it is highly likely that CD180 is a marker of dormant MM cells localized in the BM microenvironment.
Functionally, the overexpressed CD180/MD-1 complex may sense LPS to activate dormant MM cells, leading to regrowth and/or antiapoptotic response. To test this possibility, we screened for the effects of various TLR ligands, such as LPS (for TLR4 and CD180), anti-CD180 antibody (for CD180), Pam3Csk (for TLR1 and 2) and flagellin (for TLR5), as well as poke weed mitogen (TLR-independent B-cell stimulant) as a control on myeloma cell growth (Fig. 2B/C and Supplementary Fig. S4D). Among these, LPS and an activating antibody against CD180 (18) significantly enhanced the growth of KMS12-BM and RPMI8226 cells in a dose-dependent manner under adherent and hypoxic conditions, which was positively correlated with the expression levels of the CD180/MD-1 complex (Supplementary Fig. S4E). LPS failed to do so under nonadherent and normoxic conditions (Fig. 2B, Normo). Moreover, a nonactivating CD180 antibody did not potentiate the growth of CD180-positive MM cells (Supplementary Fig. S4F) and poke weed mitogen activated MM cells irrespective of culture conditions (Supplementary Fig. S4D), indicating the specificity of CD180-mediated signal transduction leading to the proliferative response. Because TLR pathways activate several intracellular signaling elements, including extracellular signal-regulated kinase (ERK), p38 MAP kinase, and c-Jun N-terminal kinase (JNK; refs. 9, 10), we examined their activation/phosphorylation status in CD180-engaged MM cells. Immunofluorescent staining revealed the activation of JNK and ERK, which coincided with CD19 but not TLR4 expression, in CD180-positive MM cells after stimulation with either LPS or anti-CD180 antibody (Fig. 2D/E; Supplementary Fig. S4G). These results strongly suggest that LPS enhances the growth of MM cells via the CD180/MD-1 pathway by the aid of CD19 (9, 24), which was expressed on dormant MM cells via the interaction with BMSCs (25, 26).
LPS enhances myeloma cell growth in positive correlation with the expression levels of CD180 in vivo
Next, we determined the effects of LPS on myeloma cell growth in vivo. For this purpose, we used a murine xenograft model, in which human MM cells were inoculated with BMSCs in immunodeficient mice to recapitulate human MM-stroma interaction in the BM milieu (5). As illustrated in Fig. 3A, a luciferase-expressing RPMI8226 subline (RPMI8226-Luc) and UBE6T-7 stroma cells were either injected separately into the right and left thighs, respectively (separate injection) or injected as a mixture into the right thigh (mixed injection) of immunodeficient mice. When tumors were measured by luciferase assay, either LPS (1 mg/kg) or vehicle (0.9% NaCl) was intraperitoneally administered twice a week for 3 weeks (n = 3 in each group). We compared the tumor sizes between vehicle-control and LPS-administered groups on day 21. Tumor growth was significantly accelerated by LPS when MM cells were coinjected with BMSCs (mixed-LPS), whereas there was no difference in case of separate injection (Fig. 3A and B). Histopathological examination also revealed tumor hypercellularity in mixed-injected mice administered with LPS, but not in the three other conditions (Fig. 3C). CD180 expression was readily detected in MM cells coinjected with BMSCs (mixed/Control), but not in the case of separate injection, and CD180/CD138 double-positive cells were markedly increased by LPS administration (mixed/LPS; Fig. 3D). These results suggest that LPS induces clonal expansion of CD180-positive MM cells in vivo.
Next, we performed knockdown experiments to investigate whether LPS-triggered growth is mediated via the CD180 pathway in MM cells. We established RPMI8226-Luc sublines lentivirally transduced with shRNA against CD180 (sh-CD180) and an ineffective control (sh-control). We selected two sublines (sh-CD180#1 and #2) in which CD180 expression was significantly downregulated relative to the control (Supplementary Fig. S5A). CD180 silencing itself did not affect the proliferative potential or viability of myeloma cells (Fig. 3E, control). First, we examined the effects of CD180 knockdown on LPS-triggered cell growth in vitro using the coculture system. As anticipated, LPS-enhanced cell growth was almost completely abrogated by CD180 silencing under adherent and hypoxic conditions (Fig. 3E, LPS). We then attempted to confirm the effects of CD180 knockdown in vivo. The sh-control and sh-CD180 cells were inoculated with UBE6T-7 cells as a mixture in the right thigh of immunodeficient mice. When measurable tumors were detected by luciferase assay, either LPS or vehicle was intraperitoneally administered twice a week for 2 weeks (n = 3 in each group). We compared the tumor sizes between vehicle-control and LPS-administered groups on day 18. LPS significantly enhanced the growth of sh-control sublines in mice, but not of CD180-knockdown sublines (Fig. 3F and G). In previous studies, LPS facilitates BMSCs to produce inflammatory cytokines, such as interleukin-6, TNF-α and IGF-1, which act in favor of MM cell growth (27, 28). Therefore, it is possible that LPS indirectly enhances the growth of MM cells by stimulating cytokine production from stromal cells. To exclude this possibility, we confirmed the absence of LPS receptors, TLR4 and CD180, in UBE6T-7 and stroma-NK cells using RT-PCR (Supplementary Fig. S5B). Moreover, the growth of RPMI8226 cells is known to be cytokine-independent (29). Taken together, we conclude that LPS directly enhances myeloma cell growth via the CD180/MD-1 pathway.
Regulation of CD180 promoter by IKZF1 transcription factor (Ikaros)
Next, we investigated the mechanisms of transcriptional activation of the CD180 gene via myeloma-stroma interactions. First, we performed reporter assays to determine the regulatory elements in CD180 promoter. We subcloned four promoter fragments, −1,955 to +18, −1,254 to +18, −1,040 to +18, and −384 to +18, into pGL4 luciferase vector to generate reporter plasmids (Fig. 4A). The analyzed regions were selected by database search for CD180 promoter (Supplementary Fig. S6). Reporter assays revealed that deletion of the segment between −1,254 and −1,040 significantly decreased CD180 promoter activity (Fig. 4B). This segment contains putative binding sites for the Ikaros family (IKZF), NF-κB and C/EBPs. Among these, we focused on IKZF-binding sites because IKZF1 (Ikaros) and IKZF3 (Aiolos) are known to be critical activators for myeloma master oncogenes IRF4 and c-myc as well as target molecules of lenalidomide, one of the most effective drugs for MM (30, 31). We constructed mutant promoters carrying a nonbinding mutation at the upstream IKZF-binding site (#1), the downstream IKZF-binding site (#2), and both (#1/#2; Fig. 4C). Promoter activities of the mutants #2 and #1/#2 were reduced to the background level in KMS12-BM cells, whereas the mutant #1 fully retained the activity (Fig. 4D). To examine the involvement of NF-κB in CD180 regulation, we treated KMS12-BM cells, in which the classical NF-κB pathway is constitutively activated for unidentified mechanisms (32), with a cell membrane-permeable peptide that inactivates NF-κB activity by inhibiting nuclear translocation of p50 (17, 33). The inhibition of NF-κB activity did not affect the expression of CD180 transcripts at all (Supplementary Fig. S7A). These results suggest that CD180 transactivation was mostly driven by Ikaros-family transcription factors through the downstream IKZF-binding site in the region between −1,254 and −1,040 of CD180 promoter.
To determine whether IKZF-family proteins are actually involved in CD180 transactivation, we first examined the expression of Ikaros and Aiolos in MM cells under adherent and hypoxic conditions. qPCR and immunoblot analyses clearly showed that cell adhesion increased Ikaros expression but decreased Aiolos expression in MM cell lines under hypoxia (Fig. 4E and F; Supplementary Fig. S7B for data quantification). In addition, we examined the status of EZH2 and its specific target site H3K27, because cell adhesion diminishes the abundance of H3K27me3 via EZH2 inactivation, leading to the transactivation of pro-survival genes in MM cells (5, 6). As anticipated, the methylation level of H3K27 was remarkably diminished in parallel with phosphorylation and downregulation of EZH2 in MM cells under adherent and hypoxic conditions (Fig. 4E and F; Supplementary Fig. S7C). No difference was observed in other histone modifications including H3K4me3 and H3K9me3 under the same conditions (Supplementary Fig. S7C). From these results, we reasoned that H3K27 demethylation and subsequent recruitment of Ikaros at the downstream IKZF-binding site may promote CD180 transcription in MM cells in direct contact with BMSCs. To confirm this hypothesis, we performed chromatin immunoprecipitation (ChIP) assays. H3K27 was readily trimethylated at the downstream IKZF-binding site (Fig. 4G, top), reflecting a relatively low baseline expression of CD180 in MM cells without cell adhesion. H3K27 was completely demethylated when MM cells were adhered to BMSCs under hypoxic conditions. Ikaros binding became detectable at this site concomitantly with H2K27 demethylation (Fig. 4G, bottom). Finally, we examined the effects of IKZF overexpression and knockdown on CD180 expression in MM cells. Consistent with the results of ChIP assays, Ikaros overexpression caused an approximately 3-fold increase in CD180 mRNA abundance (Fig. 4H), and reciprocally, its knockdown decreased CD180 expression to the same extent as Ikaros downregulation (Fig. 4I). In contrast, the modulation of IKZF3 (Aiolos) expression levels did not affect the abundance of CD180 expression in MM cells (Supplementary Fig. S7D). These data define IKZF1 (Ikaros) as a critical transactivator of the CD180 gene in MM cells.
IMiDs ameliorate LPS-triggered myeloma cell growth via downregulation of CD180 expression
The above notion prompted us to speculate that lenalidomide and its analogous IMiDs, such as pomalidomide, ameliorate LPS-triggered myeloma cell growth by targeting CD180 due to their ability of inducing cereblon-dependent degradation of Ikaros in MM cells (30, 31). To substantiate this assumption, we examined whether the two IMiDs repressed CD180 expression in parallel with the reduction of Ikaros expression levels in MM cell lines cocultured with BMSCs under hypoxia. Immunoblot analyses confirmed the dose-dependent decline in Ikaros expression by lenalidomide and pomalidomide (Fig. 5A and Supplementary Fig. S7E). With virtually identical kinetics, the two drugs reduced CD180 expression at the mRNA level and on the cell surface (Fig. 5B and Supplementary Fig. S7F). ChIP assays revealed that lenalidomide treatment readily decreased Ikaros binding to the IKZF-binding site of CD180 promoter with a reciprocal increase in H3K27 trimethylation in KMS12-BM cells (Fig. 5C). In parallel with CD180 suppression, lenalidomide treatment mitigated LPS-enhanced MM cell growth (Fig. 5D). Next, we verified the effects of lenalidomide on CD180 expression in primary MM cells. We isolated BM mononuclear cells from four MM patients and cultured them with or without lenalidomide for 24 hours. Flow cytometric analyses revealed that lenalidomide significantly downregulated CD180 expression in CD138-positive MM fractions (Fig. 5E and F). Taken together, IMiDs could suppress the LPS-triggered progression of MM by targeting CD180 expression.
Lenalidomide ameliorates LPS-triggered myeloma cell growth in vivo
We attempted to confirm the inhibitory effect of lenalidomide on LPS-triggered myeloma cell growth in vivo. The mixture of RPMI8226-Luc and UBE6T-7 cells was inoculated into immunodeficient mice as described above. When measurable tumors developed, mice were randomly assigned to 4 groups (n = 3 in each group) and treated with vehicle alone (0.3% DMSO in 0.9% NaCl), LPS alone (1 mg/kg, twice a week), lenalidomide alone (10 mg/kg, once a week), or both LPS and lenalidomide (single administration of lenalidomide followed by two injections of LPS in a week) for 3 weeks (Fig. 6A). The administration of LPS significantly enhanced tumor growth compared with vehicle control (Fig. 6A and B). Lenalidomide alone failed to retard the growth of MM cells at the dose and schedule used in this experiment, but completely cancelled LPS-induced growth enhancement in vivo. A histopathological examination confirmed the growth-promoting effect of LPS and its specific abrogation by lenalidomide (Fig. 6C). Immunofluorescent staining revealed a striking increase in CD138-positive cells along with CD180-positive as well as double-positive cells in LPS-treated mice. These cells were completely eradicated from tumor regions by lenalidomide treatment (Fig. 6D).
In the present study, we show that a noncanonical TLR complex, CD180/MD-1, is specifically expressed on MM cells and senses bacterial LPS to transduce proliferative signals, leading to the regrowth of dormant MM clones. MM cells obtain clonal dormancy via VLA-4–mediated adhesion to BMSCs, which in turn phosphorylates and inactivates the H3K27 methyltransferase EZH2 to derepress several genes conferring cell cycle arrest and drug resistance (5). We found that CD180 is an EZH2-regulated gene in MM cells and is transcriptionally activated by the recruitment of Ikaros, an important transcription factor for hematopoietic stem cell maintenance and lymphocyte development (34), to CD180 promoter upon H3K27 demethylation (Fig. 6E). Furthermore, we demonstrated that lenalidomide and its analog pomalidomide ameliorate LPS-triggered MM cell growth by silencing CD180 transcription to disrupt this circuit. In accordance with this model, disease progression occurred in 7 of 10 MM patients complicated with bacterial infections in our cohort (Supplementary Fig. S8 and Table S4). Furthermore, postinfectious disease progression was not observed in two patients under maintenance therapy with lenalidomide or pomalidomide. Because the sample size is too small to draw a firm conclusion, we have started a prospective study to verify this concept in a bigger cohort.
The CD180/MD-1 complex shares the architecture and cell surface localization with the TLR4/MD-2 complex, a canonical LPS receptor. The CD180/MD-1 complex has been proposed to play a role in fine-tuning of the cellular response to LPS, but there is still some controversy regarding its mechanisms (35). The CD180/MD-1 complex down-modulates the responses to LPS through direct interaction with the TLR4/MD-2 complex in dendritic cells and macrophages (36, 37), whereas LPS stimulates the proliferation of B cells via the CD180/MD-1 complex (22, 23, 38). The structural difference between MD-1 and MD-2 explains the low affinity of MD-1, which has a shallower cavity for direct binding to LPS (39). However, a recent crystallographic study demonstrates the structural basis for lipid and endotoxin binding to the CD180/MD-1 complex through a series of atomically detailed molecular simulations (40). The MD-1 cavity was expanded by a decrease in entropy to accommodate endotoxin binding, such that the CD180/MD-1 complex acts as a sink and source of LPS for TLR4. In this scenario, the CD180/MD-1 complex sequesters LPS from TLR4 to prevent overamplification of the TLR4 response, which ultimately leads to endotoxin shock (41). In MM cells, however, the CD180/MD-1 complex enhances the LPS response to stimulate cell proliferation because of the relatively lower expression of TLR4. This finding may increase our understanding of the functional diversity of TLRs.
From a mechanistic standpoint, our findings also provide insight into another enigmatic observation that the sensitivity of lenalidomide is positively correlated with the expression level of its target molecule Ikaros in MM cells (42, 43). We showed that Ikaros expression was increased in MM cells via adhesion to BMSCs under hypoxia. According to Jakubikova and colleagues (44), dormant MM cells, detected as SP-fraction cells by flow cytometry, are highly sensitive to lenalidomide especially in the presence of BMSCs. The causal link between dormancy and high Ikaros expression is well explained by our findings. Hence, lenalidomide can selectively eradicate dormant MM cells with high Ikaros expression in the BM microenvironment (45). In addition, we have shown that the combination of bortezomib and lenalidomide exerts additive cytotoxicity against MM cells adhered to BMSCs, whereas the same combination was antagonistic under stroma-free conditions (46). Cell adhesion increases the cytotoxic activity of lenalidomide by elevating Ikaros expression, which in turn yields additive effects with other drugs.
Recent studies with next generation sequencing disclosed the complex genomic architecture of MM (47, 48). A model combining “Big Bang” dynamics and Darwinian type of evolution has been put forward to explain the development and progression of the disease (48). According to this model, a “Big Bang” leads to the early establishment of intratumoral heterogeneity, followed by Darwinian evolution to generate different subclones with additional abnormalities (Supplementary Fig. S9). There are multiple clones with variable abilities to propagate descendants at each step of disease progression. An increase in these reservoir clones may strongly affect the biological behavior of the disease, including malignant phenotype and drug sensitivity. The entire process is strongly influenced by the interaction of each clone with the BM microenvironment. The present study discloses a previously unknown factor to affect the disease process. Bacterial infection may act as a driving force of disease progression to accelerate clonal heterogeneity, ultimately leading to clonal dominance of the selected ones. In addition, our findings suggest an indispensable role of IMiDs in maintenance (49) and/or continuous therapies (50) to improve the treatment outcome by inhibiting infection-triggered disease progression along with depletion of myeloma stem cells.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: J. Kikuchi, Y. Furukawa
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Kikuchi, Y. Kuroda, D. Koyama, N. Osada, T. Izumi, H. Yasui, T. Kawase
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Kikuchi, Y. Kuroda, D. Koyama, N. Osada, T. Ichinohe, Y. Furukawa
Writing, review, and/or revision of the manuscript: J. Kikuchi, H. Yasui, T. Ichinohe, Y. Furukawa
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Furukawa
Study supervision: T. Ichinohe, Y. Furukawa
This work was supported in part by the High-Tech Research Center Project for Private Universities: Matching Fund Subsidy from MEXT (to Y. Furukawa) and a Grant-in-Aid for Scientific Research from JSPS (to J. Kikuchi, D. Koyama and Y. Furukawa). J. Kikuchi and Y. Furukawa were funded by the Japan Leukemia Research Fund, Yasuda Memorial Cancer Foundation, Takeda Science Foundation, and Novartis Foundation Japan. J. Kikuchi was also funded by Mitsui Life Social Welfare Foundation and SENSHIN Medical Research Foundation. J. Kikuchi and Y. Furukawa received the Kano Foundation Research Grant and the Award in Aki's Memory, respectively, from the International Myeloma Foundation Japan.
The authors thank Dr. Akihiro Umezawa (National Research Institute for Child Health and Development, Tokyo, Japan) and Dr. Hirofumi Hamada (Sapporo Medical University, Sapporo, Japan) for providing UBE6T-7 and stroma-NK cell lines, respectively. We are grateful to Akiko Yonekura and Michiko Ogawa for their technical assistance.