Mechanistic Studies and a Retrospective Cohort Study: The Interaction between PPAR Agonists and Immunomodulatory Agents in Multiple Myeloma

Simple Summary Cereblon (CRBN) is a direct binding target of immunomodulatory drugs (IMiDs) that are commonly used to treat multiple myeloma (MM). Many patients with MM have comorbidities, including diabetes and/or dyslipidemia, and are treated with peroxisome proliferator-activated receptor (PPAR) agonists. This study aimed to further analyze the effects and mechanisms underlying the drug-to-drug interactions between IMiDs and PPAR agonists in MM. We found that PPAR agonists reduced CRBN expression by inducing DNA methylation and increasing protein degradation. PPAR agonists and IMiDs showed opposing metabolic effects in MM cells. Our retrospective study suggested an inferior response and outcome when PPARs and IMiDs were concurrently administered. Our study has important implications for the care of patients with MM and provides a foundation for exploring novel compounds or PPAR partial agonists/antagonists that can increase CRBN expression while retaining their lipid-lowering or insulin-sensitizing functions. Abstract Our previous study demonstrated that peroxisome proliferator-activated receptor (PPAR) agonists downregulated cereblon (CRBN) expression and reduced the anti-myeloma activity of lenalidomide in vitro and in vivo. We aimed to determine whether DNA methylation and protein degradation contribute to the effects of PPAR agonists. CRBN promoter methylation status was detected using methylation-specific polymerase chain reaction. The CRBN protein degradation rate was measured using a cycloheximide chase assay. Metabolomic analysis was performed in multiple myeloma (MM) cells treated with PPAR agonists and/or lenalidomide. Our retrospective study determined the effect of co-administration of PPAR agonists with immunomodulatory drugs on the outcomes of patients with MM. CpG islands of the CRBN promoter region became highly methylated upon treatment with PPAR agonists, whereas treatment with PPAR antagonists resulted in unmethylation. The CRBN protein was rapidly degraded after treatment with PPAR agonists. Lenalidomide and fenofibrate showed opposite effects on acylcarnitines and amino acids. Co-administration of immunomodulatory drugs and PPAR agonists was associated with inferior treatment responses and poor survival. Our study provides the first evidence that PPAR agonists reduce CRBN expression through various mechanisms including inducing methylation of CRBN promoter CpG island, enhancing CRBN protein degradation, and affecting metabolomics of MM cells.


Introduction
Multiple myeloma (MM) is an incurable cancer of terminally differentiated plasma cells and is the second most common hematological malignancy in the Western world [1]. Immunomodulatory drugs (IMiDs), including lenalidomide, pomalidomide, and thalidomide, are highly effective treatments for patients with MM [2]. Although the combination of

Protein Degradation Analysis (CHX Chase)
Protein degradation assays were based on the use of protein synthesis inhibitor cycloheximide (CHX). MM1.R and NCIH929 cells were treated with 5 µM PPAR agonist or DMSO for 3 h and then CHX (100 µg/mL) was added. Cells were collected at various time point (0, 2, 4, 6, 8 and 10 h) after CHX treatment, and whole cell lysate was prepared and analyzed for CRBN expression by Western blot analysis and the signal density was quantitated by Image J [27,28].

Western Blot Analysis
MM cells were harvested, washed with PBS, and re-suspended in lysis buffer containing 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton ×100, 1% Sodium deoxycholate, and 0.1% SDS. The cells were further lysed by brief sonication. The lysates were centrifuged at high speed for 15 min to remove the cell debris. Total protein was quantified using Dc protein estimation kit (Bio Rad) with BSA for standard curve. Approximately 20 µg protein was loaded and run on SDS PAGE. The proteins were transferred onto nitrocellulose membrane. The membrane was blocked with 5% BSA in Tris-Buffered Saline containing 0.05% of Tween 20 (TBST) and then incubated with indicated primary antibodies at 4 • C overnight. The membrane was then probed with HRP-conjugated secondary antibody and developed using Pierce ECL substrate.

Metabolomics
MM cells were treated with DMSO, fenofibrate, lenalidomide or the combination of fenofibrate and lenalidomide for 48 h. The cells were washed three times with ice-cold PBS and resuspended in 160 µL of ice-cold deionized water containing 0.6% formic acid. The cell suspension was then added 150 µL acetonitrile, mixed well with vortex, and stored at −80 • C until metabolomics analysis. Targeted metabolomics assays for acylcarnitines and amino acids were performed at Duke Metabolomics Core Facility using tandem mass spectrometry. The data were acquired using a Waters triple quadrupole detector equipped with Acquity UPLC system and controlled by Mass Lynx 4.1 software platform (Waters, Milford, MA, USA) [29].

Patient Population and Studies
The retrospective cohort study was to compare the treatment response, progression free survival (PFS), and overall survival (OS) between myeloma patients with co-existing type II diabetes and/or dyslipidemia managed with a PPAR agonist concurrently with an IMiD containing therapy for their myeloma and myeloma patients with type II diabetes and/or dyslipidemia who did not receive a PPAR agonist at any point but treated with other diabetes agents (metformin, insulin etc.) or lipid-lowering agents (Statins) during the course of their anti-myeloma treatment. Duke medical records were queried using parameters that included: (a) myeloma patients (ICD9 203.* or ICD10 C90.*) with Type II diabetes and/or hyperlipidemia (ICD9 250.*, 249.*, 272.* or ICD10 E08.*, E09.*, E10.*, E11.*, E13.*, E78.*) diagnosed between 2014 and 2020; (b) prescribed (at Duke) an IMiD (thalidomide, lenalidomide or pomalidomide) between 2014 and 2020; and with or without (c): prescribed or dispensed or administered an PPAR agonist, including fenofibrate, gemfibrozil, cirpofibrate, bezafibrate, pioglitazone, rosiglitazone, or troglitazone, between 2014 and 2020. Clinical characteristics of these patients were summarized in Supplementary Table S1.
Response to treatment was defined according to the International Myeloma working group (IMWG) treatment response criteria and classified as complete remission (CR), very good partial response (VGPR), partial response (PR), stable disease (SD), or progressive disease (PD) [30]. Cytogenetic risk stratification was performed based on the IMWG criteria and disease was staged according to the International Stage system (ISS) [31]. PFS was defined as the interval from the date of diagnosis to the last follow-up date or the date of disease recurrence. OS was defined as the duration between the date of treatment initiation to the date of death or last follow-up, with those alive censored on the date of last contact.

Statistical Analysis
Summary of patient characteristics as well as patient follow-up were tabulated. Kaplan-Meier estimation was used to determine median OS, as well as PFS, stratified by the presence of the PPAR agonist. In addition, multivariate Cox regression analysis was performed using the stepwise Cox regression model to assess the effect of multiple independent prognostic factors on survival outcome. The Log-Rank test was used to test for differences in OS and PFS among patients with or without PPAR agonists. Chi test and logistic regression analysis were used to study variables predictive of response.
Values reported and shown in graphical displays are the mean ± standard error of the mean, as indicated. Comparisons of mean expression across groups were made using twosample t-tests. It was decided to use a t-test with equal or unequal variances based on the distribution within groups. For all comparisons, p < 0.05 was used to denote significance.

High DNA Methylation at CpG Islands of the CRBN Promoter Region Was Associated with Acquisition of Resistance to Lenalidomide in MM Cell Lines
DNA methylation is a biological process by which methyl groups are added to DNA molecules. DNA methylation in a gene promoter typically represses gene transcription [32]. In our previous research, we checked the sensitivity to lenalidomide in multiple MM cell  [25]. To evaluate whether DNA hypermethylation correlates with resistance to lenalidomide, the DNA methylation status of CpG islands in the CRBN promoter region was evaluated in NCIH929 and RPMI8226 cells (lenalidomide-sensitive MM cell lines) as well as in MM1.R, MM1.S, and U266 cells (lenalidomide-resistant MM cell lines). We utilized the MethPrimer program [33] to predict CpG islands in the promoter region of CRBN ( Figure 1A). One CpG island was found from −1724 to −2014 bp and contained 10 CpG sites ( Figure 1B). Methylated and unmethylated DNAs in the CpG island of the CRBN promoter region were measured using methylation-specific polymerase chain reaction (MSP). MM1.R, MM1.S, and U266 cells were hypermethylated in CpG islands, while NCIH929 and RPMI8226 cells were partially methylated ( Figure 1C). This result is consistent with our previous results, which showed higher CRBN expression levels in NCIH929 and RPMI8226 cells and lower CRBN expression levels in MM1.R, MM1.S, and U266 cells. These data suggest that DNA methylation of CpG islands in the CRBN promoter region may be associated with lenalidomide resistance in MM cell lines.

Exposure to PPAR Agonists Induced DNA Methylation in the CpG Islands of the CRBN Promoter Region in MM Cell Lines
Our previous study showed that PPARs bind to the CRBN promoter region and directly repress CRBN transcription in MM cells [25]. Additionally, we demonstrated that fenofibrate (PPARα agonist), GW5015016 (PPARβ/δ agonist), and troglitazone (PPARγ agonist) enhance PPAR binding to the CRBN promoter and reduce CRBN expression in the mRNA and protein levels. In contrast, PPAR antagonists (GW6471, GSK3787, and GW9662) increased CRBN expression and augmented the anti-myeloma activity of lenalidomide.
When PPAR binds to PPREs in its target region, it recruits positive and negative coregulatory proteins (coactivators and corepressors, respectively) that covalently modify histone proteins, resulting in acetylation/deacetylation and methylation/demethylation [11,[34][35][36][37][38]. DNA methylation is a vital epigenetic modification that is involved in the regulation of numerous biological processes including embryogenesis, transcriptional regulation, and cell differentiation [39][40][41]. To determine the epigenetic mechanisms, we performed MSP, which would evaluate the methylation pattern of CpG islands in the CRBN promoter region upon exposure to PPAR agonists in MM cell lines, including the lenalidomide-sensitive and -resistant MM cell lines. We found that PPAR agonists induced methylation in the CpG island of the CRBN promoter region. We detected complete methylation in the CpG islands of the CRBN promoter region in the PPAR agonist-treated group, while only partial methylation was observed in the DMSO group. Similar results were observed across all tested MM cell lines ( Figure 2A).
In contrast, treatment with PPAR antagonists resulted in demethylation of the CpG island of the CRBN promoter region. We treated MM1.R, NCIH929, and RPMI8226 cells with PPAR antagonists (GW3787, GW9662, and GW6471). MSP results showed that PPAR antagonists reversed the methylation pattern from being methylated to unmethylated ( Figure 2B). These data demonstrate an important role of PPARs in regulating the methylation/demethylation of the CRBN promoter region.
Enhancer of zeste homolog 2 (EZH2), a histone-lysine N methyltransferase, catalyzes the addition of methyl groups to histone H3 at lysine 27, thereby repressing gene transcription and function [42]. Thus, we determined the effects of PPAR agonists and antagonists on EZH2 and H3K27me. MM1.R and NCIH929 cells were treated with PPAR agonists and antagonists for 48 h. Total protein was lysed for Western blotting. We found that PPAR agonists (fenofibrate, GW501516, and troglitazone) increased the levels of EZH2 and H3K27me, while PPAR antagonists (GW3787, GW9662, and GW6471) decreased their expression levels ( Figure 2C, Supplementary Figure S2). These results suggest that PPAR regulates the CRBN methylation status through the EZH2/H3K27me pathway, thereby affecting CRBN expression.

CRBN Is Rapidly Degraded in the Presence of PPAR Agonists
Proteasomal degradation is a key post-translational mechanism regulating protein levels and biological functions. In addition to serving as a transcription factor, PPARγ is an E3 ubiquitin ligase that causes ubiquitination and proteolytic degradation of NF-κB [43] and MUC1-C [44]. To evaluate whether CRBN is regulated by PPAR agonists via this mechanism, a cycloheximide degradation assay was performed, which would detect the degradation rate of CRBN in the presence or absence of PPARs agonists. Our previous data suggested that 5 µM of PPAR agonists effectively suppress CRBN expression in the mRNA and protein levels. Here, we treated MM1.R and NCIH929 cells with 5 µM of PPAR agonists, followed by cycloheximide and then measured CRBN protein level at different time points. We observed significant enhancement in CRBN protein degradation at 2-4 h with PPAR agonist treatment ( Figure 3A and Supplementary Figures S1 and S3). Furthermore, we tested whether PPAR agonist treatment promotes CRBN degradation in a proteasomedependent manner. To this end, MM1.R and NCIH929 cells were treated with only PPAR agonist, only proteasome inhibitor (bortezomib), or a combination of PPAR agonist and bortezomib. Bortezomib partially but reproducibly protected against CRBN degradation

Fenofibrate and Lenalidomide Have Opposite Effects on Lipid and Amino Acid Metabolism
PPARs play an important role in lipid oxidation by upregulating the gene expression of acyl-CoA oxidase and carnitine palmitoyl transferase I, which are enzymes involved in β-oxidation [45][46][47][48]. Fenofibrate, a PPARα agonist, is widely prescribed for the treatment of dyslipidemia, type 2 diabetes, and metabolic syndrome. Fenofibrate repairs liver injury by decreasing acylcarnitine levels and activating fatty acid β-oxidation genes that cause intracellular oxidative stress (e.g., Cpt1b, Cpt2, Mcad, and Hadha) [49]. However, the adverse effect of fenofibrate of decreasing amino acids by stimulating their catabolism has also been observed [49,50].
To further explore the mechanism of drug-to-drug interactions between PPAR agonists and IMiDs, a comprehensive metabolic analysis was performed using a combination of fenofibrate and lenalidomide, which would identify metabolite differences between the combination and control groups. Our data suggested that treatment with fenofibrate reduces the levels of C2 and C5 acylcarnitines ( Figure 4A), long-and medium-chain acylcarnitines ( Figure 4B,C), and several amino acids ( Figure 4D), while the combination treatment attenuates the metabolic modulatory effect of fenofibrate. The lenalidomide treatment group showed increased acylcarnitine and amino acid levels, which was the opposite to fenofibrate's effect. This result provides new insights into the molecular mechanism of the drug-to-drug interactions and could be relevant both for their therapeutic action and metabolic adverse effects. carnitines ( Figure 4B,C), and several amino acids ( Figure 4D), while the combination treatment attenuates the metabolic modulatory effect of fenofibrate. The lenalidomide treatment group showed increased acylcarnitine and amino acid levels, which was the opposite to fenofibrate's effect. This result provides new insights into the molecular mechanism of the drug-to-drug interactions and could be relevant both for their therapeutic action and metabolic adverse effects.

Retrospective Cohort Study: Co-Administration of PPAR Agonists with IMiDs Results in Worse PFS and OS in Patients with MM
We performed a retrospective cohort study to investigate the impact of co-administration of PPAR agonists with IMiDs on the clinical outcomes of patients with MM. A total of 196 patients with MM were included in the study-of these patients, 114 had MM

Retrospective Cohort Study: Co-Administration of PPAR Agonists with IMiDs Results in Worse PFS and OS in Patients with MM
We performed a retrospective cohort study to investigate the impact of co-administration of PPAR agonists with IMiDs on the clinical outcomes of patients with MM. A total of 196 patients with MM were included in the study-of these patients, 114 had MM with co-existing type II diabetes and/or dyslipidemia and received IMiDs but not a PPAR agonist during the course of their anti-myeloma treatment (they received other diabetes medications, such as metformin or insulin, and/or statins for dyslipidemia); 82 patients had MM with co-existing type II diabetes and/or dyslipidemia and received a PPAR agonist concurrently with an IMiD-containing therapy for MM management. Demographic and clinical features at baseline are shown in Supplementary Table S1. The median ages at diagnosis for the IMiD group and IMiD with PPAR agonist group were 64.5 years (range 40-88) and 64.7 years (range 38-85), respectively. There were no significant differences in age, sex, race, body mass index (BMI), International Staging System (ISS) stage, or cytogenetic risk between the two groups of patients. 54.4% and 54.9% of the IMiD and IMiD/PPAR agonist cohorts, respectively, received hematopoietic stem cell transplantation treatment (HSCT). The PPAR agonists comprised known antidiabetic drugs including gemfibrozil, fenofibrate, and pioglitazone.

Discussion
High CRBN expression in patients with MM is associated with a good clinical response to IMiDs, while inactivation of the CRBN gene in cell lines and low protein expression confer resistance to IMiD treatment [51,52]. Therefore, CRBN levels are expected to be a promising marker for the efficacy of IMiDs. However, it remains unclear how CRBN is regulated and whether different genomic approaches or drugs could affect sensitivity to IMiDs through the regulation of CRBN expression. We have previously shown that PPARs agonists (fenofibrate, GW501516, and troglitazone) reduce the anti-myeloma effect of lenalidomide by downregulating the transcription activity of CRBN, while PPARs antagonists (GW3787, GW9662, and GW6471) increase CRBN expression and improve lenalidomide activity [53]. Promoter binding analysis based on the Chip assay and luciferase assay suggested that there are different PPAR binding sites on the CRBN promoter region; therefore, PPARs regulate CRBN transcriptional activity by stimulating these sites [53]. In this study, we explored the exact mechanism of this transcriptional inhibition of CRBN by PPARs. We observed hypermethylation of CpG islands in the CRBN promoter in lenalidomide-resistant MM cell lines, while partial methylation was observed in lenalidomide-sensitive MM cell lines. These data are consistent with those of our previous study [25]. Importantly, our MSP analyses demonstrated that PPAR agonists affect the CRBN promoter CpG island methylation patterns and cause hypermethylation in CRBN promoter region. At the same time, PPAR antagonists changed the patterns from methylated to unmethylated. In addition, we found that CRBN was rapidly degraded upon exposure to PPARs agonists. These studies may help us fully elucidate the mechanisms through which PPARs affect CRBN expression level. Furthermore, our retrospective cohort study provided clinical evidence that the co-administration of PPAR agonists with IMiDs resulted in worse PFS and OS in patients with MM.
Our study is significant in several respects, providing the first evidence that PPARs agonists or antagonists may affect MM cell line sensitivity to lenalidomide by changing the methylation pattern of CpG islands in the CRBN promoter region. Moreover, PPAR agonists affect the outcomes of patients with MM when they are treated with IMiDs. This study further supports our previous research that PPAR agonists attenuate lenalidomide activity in MM and suggests that patients taking IMiDs avoid PPAR agonists. Our findings have important clinical relevance because over 1 in 4 myeloma patients has diabetes or dyslipidemia and many of these patients are taking PPAR agonists concurrently with IMiD treatment. Although additional and prospective studies are needed, our study suggested that for patients with myeloma and treated with IMiDs, the use of PPAR agonists should be avoided and other medications such as statins, metform, and insulin can be used for the management of dyslipidemia and/or type II diabetes.
A few studies are also consistent with our findings and support the association between CRBN methylation and IMiD resistance in MM. Haertle et al. identified a previously undescribed DNA hypermethylation in an active intronic CRBN enhancer [54]. Differential hypermethylation in this region was pronounced in IMiD-refractory MM. Methylation was significantly correlated with decreased CRBN expression levels. In vitro experiments with DNA methyltransferase inhibitors induced CRBN enhancer demethylation, and sensitizing effects of lenalidomide treatment were observed in two MM cell lines. However, other studies found no hypermethylation in the CRBN promoter in clinical samples [19]. Additionally, Dimopoulos et al. showed no involvement of promoter methylation in the regulation of CRBN expression in lenalidomide-resistant MM cell lines [55]. This inconsistency might be due to different treatment regimens used in the studies.
Apart from the prognostic value of CRBN level, CRBN genetic alterations such as gain/loss or mutation following lenalidomide treatment have been reported but the results were controversial. One study found mutations in CRBN in 6 (12%) out of 50 patients with relapsed MM [56], while another study found 3 patients with mutations in TP53 but no mutations in CRBN or IKZF1 out of 22 patients with relapsed MM [19]. Differences in treatment duration and combination with other agents could affect the incidence of mutations in CRBN following lenalidomide treatment. Further studies with larger patient sample size will be important to ascertain the incidence and role of mutations in CRBN and related genes in patients with relapsed MM.
Targeted protein degradation represents a rapidly growing area in drug discovery and development [57,58]. Moreover, small molecules or compounds that reduce targeted protein degradation, thereby increasing relative chemical sensitivity, also represent an important direction to overcome drug resistance. More recently, CRBN has gained popularity as an E3 ubiquitin ligase that can be hijacked by small-molecule degraders and used for targeted protein degradation [59,60]. The utility of CRBN in this context stems from the discovery that it binds IMiDs, resulting in the recruitment of IKZF1 and IKZF3 and their subsequent ubiquitination and proteasomal degradation [61]. Additionally, IMiDs can be incorporated into bifunctional degrader molecules (also called PROTACs), which can induce the degradation of target proteins beyond transcription factors, by bringing them into proximity to an E3 ligase [62][63][64]. These compounds are composed of an E3 ligaserecruiting element, such as IMiDs that recruit CRBN, and a binding motif for the target of interest. Bifunctional degraders function in the same manner as IMiDs, by facilitating the formation of ubiquitination-competent ternary complexes [65,66].
Drug-drug interactions between PPAR agonists and antitumor agents remain largely uncharacterized. Many anticancer treatment techniques interact with PPARγ ligands that synergistically enhance their efficiency. The combination of PPARγ ligands and tyrosine kinase inhibitors in chronic myelogenous leukemia is the most striking example [67]. Here, PPARγ activation renders leukemic stem cells that have previously been resistant to treatment susceptible to targeted therapy. In contrast, treatment with PPARγ ligands significantly alters the response and cell viability when co-treated with imatinib in chronic myeloid leukemia [67]. Therefore, given the possibility of interactions, a clinical evaluation of PPAR and the pharmacodynamic effects of co-administration with IMiDs will be useful to inform future clinical practices.
Aberrant lipid metabolism is recognized as a key feature of MM [68]. A previous study has reported a significant accumulation of lipids in MM cells after proteasome inhibition [69]. Our metabolic analysis, performed in a lenalidomide-resistant cell line, showed a similar trend after lenalidomide treatment in MM cell lines. This finding prompted us to hypothesize that MM cell survival depends on the maximal utilization of abnormally accumulated lipids. Whether lipid metabolism-modulating agents synergize with lenalidomide must be explored in the future. This requires additional research that is beyond the scope of this manuscript. With an in-depth understanding of tumor pathology and the evolution of new drug research and development technology, we believe that newer small-molecule compounds or partial agonists/antagonists will be developed in the near future.

Conclusions
To our knowledge, this is the first study to explore the response to IMiDs in patients who have MM with co-existing diabetes and/or dyslipidemia and treated with PPAR agonists. Between patients receiving IMiDs and patients receiving IMiDs and PPAR agonists, we found statistically significant differences in the overall response rate, PFS, and OS. More clinical samples should be used to detect the CRBN promoter methylation status. Furthermore, there were limitations to our study. It was a single center, retrospective analysis of data. Despite its limitations, the present study provides new information regarding drug-to-drug interaction that may be useful in guiding future studies.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cancers14215272/s1. Figure S1: PPARs agonist increase CRBN protein degradation; Figure S2: Uncropped blots and intensity ratio of Figure 2C. Figure S3: Uncropped blots and intensity ratio of Figure 3. Table S1: Clinical characteristics of patients treated with IMiD and with/without PPAR agonist; Table S2: Overall response rate in myeloma patients treated with IMiD alone or with concurrent PPAR agonists; Table S3: Multivariable cox regression analysis.
Author Contributions: Study conception and design undertaken, J.W. and Y.K.; methodology, J.W. and E.C.; statistical analysis and interpreted, J.W.; patient data extraction: B.P. and Y.K.; writing-original draft, J.W. and Y.K.; writing-review and editing, E.C. and Y.K.; funding acquisition, Y.K. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported by the Duke Cancer Institute Fund [to Y.K.] and the National Cancer Institute [grant numbers R44CA199767, R01CA197792, and R21CA234701 [to Y.K.]. The sponsors have no roles in study design, the collection, analysis and interpretation of data, the writing of the report or the decision to submit the article for publication.

Institutional Review Board Statement:
The retrospective study involving chart review of patient medical records was performed in accordance with the ethical standards of Duke University Institutional Review Board Committees on human experimentation (Pro00088486, approval date: 11 June 2017).

Data Availability Statement:
The data presented in this study are available in this article (and Supplementary Materials).