Author response: Targeting the fatty acid binding proteins disrupts multiple myeloma cell cycle progression and MYC signaling

Abstract

Article Figures and data Abstract Editor's evaluation eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Multiple myeloma is an incurable plasma cell malignancy with only a 53% 5-year survival rate. There is a critical need to find new multiple myeloma vulnerabilities and therapeutic avenues. Herein, we identified and explored a novel multiple myeloma target: the fatty acid binding protein (FABP) family. In our work, myeloma cells were treated with FABP inhibitors (BMS3094013 and SBFI-26) and examined in vivo and in vitro for cell cycle state, proliferation, apoptosis, mitochondrial membrane potential, cellular metabolism (oxygen consumption rates and fatty acid oxidation), and DNA methylation properties. Myeloma cell responses to BMS309403, SBFI-26, or both, were also assessed with RNA sequencing (RNA-Seq) and proteomic analysis, and confirmed with western blotting and qRT-PCR. Myeloma cell dependency on FABPs was assessed using the Cancer Dependency Map (DepMap). Finally, MM patient datasets (CoMMpass and GEO) were mined for FABP expression correlations with clinical outcomes. We found that myeloma cells treated with FABPi or with FABP5 knockout (generated via CRISPR/Cas9 editing) exhibited diminished proliferation, increased apoptosis, and metabolic changes in vitro. FABPi had mixed results in vivo, in two pre-clinical MM mouse models, suggesting optimization of in vivo delivery, dosing, or type of FABP inhibitors will be needed before clinical applicability. FABPi negatively impacted mitochondrial respiration and reduced expression of MYC and other key signaling pathways in MM cells in vitro. Clinical data demonstrated worse overall and progression-free survival in patients with high FABP5 expression in tumor cells. Overall, this study establishes the FABP family as a potentially new target in multiple myeloma. In MM cells, FABPs have a multitude of actions and cellular roles that result in the support of myeloma progression. Further research into the FABP family in MM is warrented, especially into the effective translation of targeting these in vivo. Editor's evaluation This paper has valuable findings that have practical implications within multiple myeloma and tumor microenvironment fields. It describes a family of genes regulating myeloma cell survival and proliferation. The approaches used are convincing, state-of-the-art, and rigorous. The data support the essential claims of the manuscript and have mechanistic depth. https://doi.org/10.7554/eLife.81184.sa0 Decision letter Reviews on Sciety eLife's review process eLife digest Multiple myeloma is a type of blood cancer for which only a few treatments are available. Currently, only about half the patients with multiple myeloma survive for five years after diagnosis. Because obesity is a risk factor for multiple myeloma, researchers have been studying how fat cells or fatty acids affect multiple myeloma tumor cells to identify new treatment targets. Fatty acid binding proteins (FABPs) are one promising target. The FABPs shuttle fatty acids and help cells communicate. Previous studies linked FABPs to some types of cancer, including another blood cancer called leukemia, and cancers of the prostate and breast. A recent study showed that patients with multiple myeloma, who have high levels of FABP5 in their tumors, have worse outcomes than patients with lower levels. But, so far, no one has studied the effects of inhibiting FABPs in multiple myeloma tumor cells or animals with multiple myeloma. Farrell et al. show that blocking or eliminating FABPs kills myeloma tumor cells and slows their growth in a dish (in vitro) and in some laboratory mice. In the experiments, the researchers treated myeloma cells with drugs that inhibit FABPs or genetically engineered myeloma cells to lack FABPs. They also show that blocking FABPs reduces the activity of a protein called MYC, which promotes tumor cell survival in many types of cancer. It also changed the metabolism of the tumor cell. Finally, the team examined data collected from several sets of patients with multiple myeloma and found that patients with high FABP levels have more aggressive cancer. The experiments lay the groundwork for more studies to determine if drugs or other therapies targeting FABPs could treat multiple myeloma. More research is needed to determine why inhibiting FABPs worked in some mice with multiple myeloma but not others, and whether FABP inhibitors might work better if combined with other cancer therapies. There were no signs that the drugs were toxic in mice, but more studies must prove they are safe and effective before testing the drugs in humans with multiple myeloma. Designing better or more potent FABP-blocking drugs may also lead to better animal study results. Introduction Fatty acid binding protein (FABP) family members are small (12–15 kDa) proteins that reversibly bind lipids (Hotamisligil and Bernlohr, 2015). The 10 human FABP isoforms are functionally and spatially diverse, consisting of ten anti-parallel beta sheets, which form a beta barrel that shuttles fatty acids across membranes of organelles including peroxisomes, mitochondria, nuclei, and the endoplasmic reticulum (Furuhashi and Hotamisligil, 2008). FABPs influence cell structure, intracellular and extracellular signaling, metabolic and inflammatory pathways (Hotamisligil and Bernlohr, 2015), and maintain mitochondrial function (Field et al., 2020). While most cell types express a single FABP isoform, some co-express multiple FABPs that can functionally compensate for each other if needed (Hotamisligil et al., 1996; Shaughnessy et al., 2000), suggesting that broad FABP targeting may be necessary. FABP insufficiencies in humans and mice induce health benefits (eg. protection from cardiovascular disease, atherosclerosis, and obesity-induced type 2 diabetes), suggesting these to be safe therapeutic targets (Cao et al., 2006; Maeda et al., 2005; Tuncman et al., 2006). Multiple myeloma (MM), a clonal expansion of malignant plasma cells, accounts for ~10% of hematological neoplasms (Rajkumar, 2020). Myeloma cell growth initiates in and spreads throughout the bone marrow, leading to aberrant cell proliferation and destruction of the bone (Fairfield et al., 2016). Treatments for myeloma patients have greatly improved within the past two decades (American cancer institute, 2022), but most patients eventually relapse, demonstrating the need to pursue more novel types of MM treatment. Few therapies are designed to specifically target molecules involved in the MM cell metabolism, despite recent findings that MM cells uptake fatty acids through fatty acid transport proteins, which can enhance their proliferation (Panaroni et al., 2022). Links between FABP4 and cancer have been demonstrated in prostate, breast, and ovarian cancer, and acute myeloid leukemia (AML; Al-Jameel et al., 2017; Carbonetti et al., 2019; Herroon et al., 2013; Lan et al., 2011; Mukherjee et al., 2020; Shafat et al., 2017; Yan et al., 2018; Zhou et al., 2019). FABP5 has been less widely studied in cancer, but is known to transport ligands to PPARD (Tan et al., 2002), which can intersect with many pro-tumor pathways that increase proliferation, survival (Adhikary et al., 2013; Di-Poï et al., 2002; Tan et al., 2001), and angiogenesis (Wang et al., 2006), and decrease tumor suppressor expression (Tan et al., 2001). Herein we explored the oncogenic function of the FABPs in MM by examining therapeutic targeting with FABP inhibitors (FABPi) in multiple cell lines in vitro, and using genetic knockout of FABP5, pre-clinical models, large cell line datasets, and multiple patient datasets. Our results suggest FABPs are a novel target in MM due to the plethora of important biological functions that FABPs modulate to control cellular processes at multiple levels. Results FABP5 is vital for MM cells and genetic knockout results in reduced cell number We first examined FABP gene expression in MM cell lines and found that FABP5 was the most highly-expressed FABP isoform in GFP+/Luc+MM.1S and RPMI-8226 cells (Supplementary file 1, Fairfield et al., 2021) and that some other FABPs were also expressed to a lesser extent (eg. FABP3, FABP4, and FABP6). FABP5 protein was also robustly expressed in these cells (Figure 1A, Figure 1—figure supplement 1A), and FABP5 consistently showed the expression in haematopoetic/lymphoid lineage lines within the Cancer Cell Line Encyclopedia (CCLE) at the gene level (Figure 1—figure supplement 1B) and protein level (Figure 1—figure supplement 1C, “DepMap 22Q2, 2022; Ghandi et al., 2019; Nusinow et al., 2020). In MM cell lines specifically, FABP5 was the most highly expressed at the gene level (Figure 1B) and FABP5 and FABP6 were the most highly expressed at the protein level (Figure 1—figure supplement 1D). In the Broad Institute’s Cancer Dependency Map (DepMap; Tsherniak et al., 2017), of all the FABPs, only FABP5 exhibited a negative CERES Score (–0.30) in all 20 MM cell lines, demonstrating a strong reliance on FABP5 for their survival (Figure 1—figure supplement 2A). Interestingly, all cancer types within the DepMap database had negative FABP5 CERES values (Figure 1—figure supplement 2B). Importantly, many fatty acid metabolism genes, including FABP5, had negative CERES scores (shown in blue) in MM cells (Figure 1—figure supplement 2C). Figure 1 with 8 supplements see all Download asset Open asset FABPi significantly impair MM cell growth and induces apoptosis. (A) Confocal overlay immunofluorescence images show FABP5 (red) expressed in cytoplasm of GFP+/Luc+ MM.1S cells. Nuclei identified with DAPI (blue), cells stained with secondary antibody alone (control) or primary plus secondary antibodies (FABP5 staining), scale bar = 200 µm. (B) Comparison of basal gene expression of FABP isoforms in 30 myeloma cell lines. Data extracted from the Cancer Cell Line Encyclopedia (CCLE; DepMap, Broad (2022): DepMap 22Q2 Public. figshare. dataset. https://doi.org/10.6084/m9.figshare.19700056.v2), filtered in excel, and graphs made in Graphpad PRISM (v7.04) using scatter dot plots (mean ± SEM). (C, D) MM cell numbers after being exposed to (C) BMS309403 and (D) SBFI-26 for 72 hr; 50 µM dose (~EC50) indicated by arrows. (E) GFP+/Luc+MM.1S cell numbers after treatment with inhibitors in combination (50 µM each). Vehicle vs BMS309403 (24 hr, *; 48 hr, ****; 72 hr, ****). Vehicle vs SBFI-26 (24 hr, *; 48 hr, ****; 72 hr, ****). Vehicle vs BMS309403 +SBFI-26 (24 hr, ***; 48 hr, ****; 72 hr, ****). BMS309403 vs BMS309403 +SBFI-26 (48 hr, **; 72 hr, ****). SBFI-26 vs BMS309403 +SBFI-26 (48 hr, **; 72 hr, ****). Two-way ANOVA analysis with Tukey’s multiple comparisons test analysis. (F) CellTiter-Glo analysis of human mesenchymal stem cells after treatment with BMS309403 or SBFI-26 for 72 hr. Data are mean ± SEM and represent averages or representative runs of at least three experimental repeats. One-way ANOVA with Dunnett’s multiple comparison test significance shown as *p<0.05. **p<0.01. ***p<0.001. ****p<0.0001. **** p<0.0001. Please see 8 supplements to Figure 1. Based on these initial findings, we next examined the effect of FABP5 knockout (KO) in MM cells. FABP5 KO (FABP5KO) MM.1R cells exhibited a 94% editing efficiency with a ~59% KO efficiency after expansion (Figure 1—figure supplement 3A and B). We observed an 84% reduction in FABP5 expression in the edited pool (Figure 1—figure supplement 3C), confirming functional FABP5 knockdown. FABP4 expression was not altered (Figure 1—figure supplement 3D), but FABP6 expression was increased in the edited cells (Figure 1—figure supplement 3E). FABP5 KO cells showed a slight reduction in cell numbers at 48, 72, and 96 hr, versus controls (Figure 1—figure supplement 3F). Pharmacological inhibition of FABPs reduces myeloma cell proliferation in vitro Having observed potential compensation among FABP family members in the FABP5KO cells, we next used two well-known FABP inhibitors (FABPi): BMS309403 and SBFI-26, which specifically and potently inhibit FABPs by binding their canonical ligand-binding pockets, or inducing conformational changes, for example by binding their substrate entry portal region (Hsu et al., 2017). Ligand-binding assays determined that BMS309403 has Ki values in solution of <2, 250, and 350 nM for FABP4, FABP3, and FABP5, and that SBFI-26 has Ki values of 900 and 400 nM for FABP5 and FABP7, respectively, as reported on the manufacturers’ datasheets (Hsu et al., 2017). BMS309403 and SBFI-26 consistently demonstrated dose-dependent decreases in myeloma cell numbers, in all 7 MM lines screened, at 72 hr (Figure 1C and D; Supplementary files 2 and 3) and earlier (Figure 1—figure supplement 4). BMS309403 (50 µM), SBFI-26 (50 µM), or the combination (50 µM BMS309403 +50 µM SBFI-26) reduced cell numbers at 24, 48, and 72 hr by 39%, 42%, and 83%, respectively in GFP+/Luc+MM.1S cells (Figure 1E), suggesting that targeting different FABPs, or using different FABP inhibitors, could be beneficial. Non-cancerous cells were much less sensitive to FABPi (Figure 1F), intimating the potential clinical translation of these or similar FABP inhibitors, as supported by prior literature showing the safety of FABP inhibitors (Al-Jameel et al., 2017; Mukherjee et al., 2020). No change in amount or localization of FABP5 protein after treatment with FABPi was observed by immunofluorescence (Figure 1—figure supplements 5 and 6) at 24 hr in GFP+/Luc+MM.1S or RPMI-8226 cells, or by western blotting at 24, 48, or 72 hr in GFP+/Luc+MM.1S cells (Figure 1—figure supplement 7A and B). Gene expression of FABP3, FABP4, FABP5, and FABP6 were also not consistently altered with treatments (Figure 1—figure supplement 7C) as assessed by qRT-PCR. These data suggest that FABP activity, but not protein expression, is decreased by these FABP inhibitors. Recombinant FABP4 and FABP5 did not affect MM.1S cell number (Figure 1—figure supplement 8A,B). FABPi induce gene expression changes in myeloma cells that affect a range of cellular processes and pathways linked to survival To identify transcriptional changes that may mediate the effects of FABP inhibition on cell number, we treated GFP+/Luc+MM.1S cells with a vehicle control, the single FABP inhibitors alone (50 µM), or the combination of FABPi (50 µM of each) for 24 hr in vitro, isolated total RNA, and performed RNA-Seq. Principal component analysis (PCA) demonstrated that the FABP inhibitor groups exhibited distinct gene expression profiles, and that the combination treatment differed the most from vehicle-treated cells (Figure 2A). Over 14,000 genes were analyzed, revealing 93 significant differentially expressed (DE) genes within all three treatment groups, compared to the vehicle control (FDR <0.2): 90 downregulated and 3 upregulated (Figure 2B; Supplementary file 4). Consistent with decreased levels of transcription, we also observed significantly lower levels of 5-hydroxymethylcytosine in cells treated with FABPi compared to vehicle-treated cells (Figure 2C), suggesting decreases in active chromatin. This finding is consistent with previous reports linking FABP depletion to DNA methylation signatures in other cancers (Mukherjee et al., 2020; Yan et al., 2018). Figure 2 with 3 supplements see all Download asset Open asset RNA sequencing analysis of MM1S cells treated with FABPi for reveals unique gene expression patterns. (A) Principal component analysis of cells after 24 hr treatments. (B) Venn diagram displays the overlapping and specific genes dysregulated with FABPi (FDR cutoff of 0.2). (C) Global hydroxymethylation DNA analysis of MM.1S cells after 24 hr of combination treatment. Data represent mean and +/- SEM using n=3 biological repeats, and * p<0.05 using an unpaired, two-tailed Student t-test. (D) Ingenuity pathway analysis of RNA-Seq results (p-value of overlap by Fisher’s exact test, significance threshold value of p<0.05(-log value of 1.3)). Stringdb (FDR cutoff of 0.2) of the combination therapy versus control showing (E) the 1 upregulated pathway and (F) 5 of the many downregulated pathways. MYC, a central node, is circled for emphasis. GFP+/Luc +MM.1 S cells were used for these experiments. Please see 3 supplements to Figure 2. To further understand the mechanisms of action of FABPi, we investigated which pathways were impacted in our RNA-Seq data using STRINGdb and IPA (Ingenuity Pathway Analysis). IPA was specifically used to investigate canonical pathways, while STRINGdb was used to examine connectivity of DE genes and enrichment for specific gene ontology terms, as well as molecules in Reactome and KEGG pathways. In total, 15 IPA canonical pathways were commonly dysregulated in all three treatment groups including Cell Cycle: G2/M DNA Damage Checkpoint Regulation, EIF2 Signaling, Sirtuin Signaling Pathway, and the NER pathway (Figure 2D; Supplementary file 5). The one upregulated pathway according to STRING was ‘cellular response to interferon gamma signaling’ in the combination group (Figure 2E; Supplementary file 6). The top downregulated pathways in the combination treatment by STRING analysis are in Supplementary file 7. Interestingly, both IPA and STRING databases revealed commonly downregulated pathways related to the unfolded protein response (UPR) or ER stress responses for BMS309403 (Figure 2—figure supplement 1A–C), SBFI-26 (Figure 2—figure supplement 3, and the combination Figure 2D and F). Three of the five downregulated Reactome pathways in the combination group were related to UPR or ER stress (Figure 2F), driven by molecular players such as XBP1, BIP (HSPA5), and IRE1 (ERN1) (Figure 2—figure supplement 3A). Downregulation of total XBP1 by the combination treatment was confirmed after 24 hr (Figure 2—figure supplement 3B) and heatmaps visually demonstrated the downregulation of genes involved in XBP1 signaling (Figure 2—figure supplement 3C) and the UPR (Figure 2—figure supplement 3D) as determined by IPA. Interestingly, MYC, a known oncogene, was a central node in STRING analysis (Figure 2F) and among the top 10 most downregulated genes in RNA-Seq from combination treatments (Supplementary file 8). FABPi induces protein changes in MM cells that affect a range of cellular processes and pathways linked to survival To identify protein changes resulting from FABPi, we treated GFP+/Luc+MM.1S cells with the single inhibitors (50 µM) or the combination (50 µM of each) for 48 hr, isolated total cell lysate proteins, and performed a mass spectrometry-based proteomic analysis. (Numbers of significant proteins, Supplementary file 9; protein names, Supplementary files 10-15). PCA analysis showed a tight grouping of samples based on treatments (Figure 3—figure supplement 1A); 15 proteins were commonly upregulated and 15 were commonly downregulated between all treatments (Figure 3—figure supplement 1B, C; Supplementary files 16 and 17). We then compared significant genes and proteins identified by both RNA-Seq and proteomics (Figure 3A and B). CCL3, a chemokine for monocytes, macrophages, and neutrophils, was upregulated by SBFI-26, BMS309403, and their combination in proteomics, and upregulated by the combination treatments in RNA-Seq. Ki67, a proliferation marker, and PTMA, a negative regulator of apoptosis, were both significantly downregulated in the combination treatment in RNA-Seq and proteomics, and in the single drug treatments in proteomics (Figure 3B), indicating cell death and cell cycle arrest likely result from FABPi. Figure 3 with 7 supplements see all Download asset Open asset Forty-eight hr proteomic analysis of MM1S cells treated with FABPi reveals a unique protein signature. MM.1S cells were assessed by proteomics after 48 hr treatments with BMS309403 (50 µM), SBFI-26 (50 µM) or the combination, and compared to results from RNA-Seq. N=3 biological replicates and three technical replicates Venn diagram comparison of (A) upregulated genes and (B) downregulated proteins in proteomics and RNA-Seq among BMS309403 and SBFI-26 treated cells compared to vehicle. (C–F) Pathway analysis of proteomic data of significantly upregulated or downregulated proteins in MM.1S cells treated with both FABPi (BMS309403 +SBFI-26). (C, D) String analysis of upregulated (C) or downregulated (D) pathways. (E) Top 10 significantly changed pathways with FABP inhibition. For IPA analysis, orange represents positive z-score, blue indicates a negative z-score, gray represents no activity pattern detected and white represents a z-score of 0. (F) Ingenuity pathway analysis of the Cell Death and Survival heatmap. Numbers in boxes represent: (1) Cell death of melanoma lines; (2) Cell death of carcinoma cell lines; (3) Cell death of neuroblastoma cell lines; (4) Cell death of breast cancer cell lines; (5) Cell death of connective tissue cells; (6) Cell death of fibroblast cell lines; (7) Cell viability of myeloma cell lines; (8) Apoptosis of tumor cell lines; (9) Apoptosis of carcinoma cell lines. GFP+/Luc +MM.1 S cells were used for these experiments. Please see 7 supplements to Figure 3. STRING analysis of proteomic data suggested many other systemic changes (eg, downregulation of DNA replication and other viability/proliferation processes and upregulation of lysosome, carboxylic acid catabolic process, and mitochondrial pathways) induced by the FABPi combination treatments (Figure 3C and D). STRING analysis also revealed interesting up- and downregulated pathways by BMS309403 or SBFI-26 treatments alone (Figure 3—figure supplement 2, Figure 3—figure supplement 3). IPA analysis revealed ‘EIF2 Signaling’ to have the highest negative Z-score for all FABPi treatments in proteomics (Figure 3E; Figure 3—figure supplement 4A, Figure 3—figure supplement 5A). IPA ‘Cell Death and Survival’ heatmap analysis showed increases in cell death and apoptosis pathways and decreases in cell viability pathways after FABPi combination treatment (Figure 3F; Figure 3—figure supplements 4B and 5B). Interestingly, MYC was the most significant predicted upstream regulator, found to be strongly inhibited in the BMS309403, SBFI-26, and combination treatments from IPA proteomic analysis (Supplementary files 18-20). Since MYC was found as a central node or commonly downregulated gene/pathway in our RNA-Seq and proteomic data analyses, we investigated MYC’s role in FABP signaling in myeloma cells. We confirmed decreased MYC expression in GFP+/Luc+MM.1S cells treated with the FABPi combination, and also saw a trend for this in 5TGM1-TK cells treated with SBFI-26 (Figure 3—figure supplement 6A, B). MYC protein level was also decreased in GFP+/Luc+MM.1S cells at 24, 48, and 72 hr with FABPi (Figure 4A and B), with similar trends observed in 5TGM1-TK myeloma cells (Figure 3—figure supplement 6C, D). The decrease in MYC-regulated genes with FABPi was also visualized in both the RNA-Seq (Figure 4C) and proteomic data (Figure 4D) by heatmap analysis. In RNA-Seq data, treatment with BMS309403 induced aberrant gene expression of 171 genes known to be regulated by MYC (Supplementary file 21), with 138 of those having expression patterns consistent with MYC inhibition. Similarly, co-treatment induced changes in 91 genes modulated by MYC (Figure 3—figure supplement 7; 68 consistent with MYC downregulation), while 29 MYC targets were aberrantly expressed with SBFI-26 treatment (Figure 2—figure supplement 2D; 18 consistent with MYC downregulation). Figure 4 with 1 supplement see all Download asset Open asset FABPi target MYC and the MYC pathway. (A) Representative western blot and (B) quantification of MYC protein and β-actin (housekeeping control) at 24, 48, and 72 hr after treatment with BMS309403 (50 µM), SBFI-26 (50 µM), or the combination. (C) RNA-seq and (D) Proteomic analysis of expression of genes/proteins involved in MYC signaling shown as heatmap visualizations. Curated lists are based on IPA MYC Pathway list, known MYC-regulated genes, and proteins present in proteomics. (E) 72 hr BMS309403 dose curve with and without Myc inhibitor 10058-F4 (37.5 µM) in MM.1S cells. (F) 72 hr SBFI-26 dose curve with and without 10058-F4 (37.5 µM) in MM.1S cells. Data represent mean ± SEM from n=3 biological repeats, analyzed with one-way ANOVA with significance shown as *p<0.05. **p<0.01. ****p<0.0001. GFP+/Luc +MM.1 S cells were used for these experiments. Please see 1 supplement to Figure 4. To test if MYC inhibition was a major cause of the FABPi effects on MM cells, we then pharmacologically inhibited MYC and tested a range of doses of FABPi. MYC inhibition alone dramatically reduced cell numbers at 72 hr, as expected, and FABP inhibition had less of an effect on MM cells when MYC was already inhibited (seen by a slope of ~0 for the black lines) (Figure 4E and F). This suggests that much of the effect of FABPi is through decreased MYC signaling, although the strong effect of the MYC inhibitor makes this difficult to determine unhesitantly. Similar results were seen at 24 and 48 hr (Figure 4—figure supplement 1). FABPi impair MM cell metabolism, mitochondrial function, and cell viability Having observed effects of the inhibitors on metabolic processes such as mitochondrial function and oxidative phosphorylation in the proteomic data, we next assessed mitochondrial function and metabolic changes using a Cell Mito Stress Test (Figure 5—figure supplement 1A). After 24 hr treatments, all FABPi treatments decreased basal mitochondrial oxygen consumption rates (OCR) and OCR dedicated to ATP production (Figure 5—figure supplement 1B). Maximal respiration and spare respiratory capacity were decreased with SBFI-26 and combination treatments, suggesting FABP inhibition reduces the ability of MM cells to meet their energetic demands. To determine the effects of FABPi on fatty acid oxidation (FAO) specifically, we treated tumor cells with etoxomir, an FAO inhibitor, with or without the combination FABPi treatment (Figure 5—figure supplement 2). The combination of FABPi alone again strongly reduced mitochondrial respiration in most of the parameters assessed. Interestingly, etoxomir treatment caused a slight, but significant reduction in OCR when it was administered, demonstrating some reliance of MM cells on FAO for mitochondrial respiration. However, the FABPi had a much greater effect on MM mitochondrial respiration than etoxomir alone, suggesting that FABPi treatment inhibited mitochondrial respiration through another mechanism. Also, since maximal respiration was decreased in the Etox +FABPi combination compared to FABPi alone, it appears that FABPi treatment does not completely block FAO when used alone. Overall, the data demonstrate that mitochondrial respiration is inhibited by FABPi. To assess whether metabolic dysfunction could be caused by damaged mitochondria, we utilized tetramethylrhodamine, ethyl ester (TMRE) staining and flow cytometric analysis to assess mitochondrial transmembrane potential. GFP+/Luc+MM.1S cells treated with BMS309403 or the combination (BMS309403 +SBFI-26) had decreased TMRE staining (Figure 5—figure supplement 3), suggesting that BMS309403 damages MM cell mitochondria. We next investigated if reactive oxygen species (ROS), a major byproduct of the electron transport chain, were changing in MM cells after FABPi treatment. CellROX staining showed that the combination FABPi treatment significantly increased total ROS at 24, 48, or 72 hr in MM.1S (ATCC), U266 and OPM2 cells (Figure 5A, Figure 5—figure supplements 4A and 5A, 6 A). We also found changes in superoxide, a ROS subspecies measured by MitoSOX, after FABP inhibition; in MM.1S (ATCC), BMS309403 and the FABPi combination increased superoxides over 72 hr (Figure 5B, Figure 5—figure supplement 4B). In U266, the FABPi combination increased superoxides at each time point, and BMS309403 increased superoxides at 48 and 72 hr (Figure 5—figure supplement 5B). In OPM2, all FABPi treaments increased superoxides at all timepoints (Figure 5—figure supplement 6B). Overall, FABP proteins are vital to MM cells for normal oxygen consumption, mitochondrial potential maintenance and ATP production, adaption to increased demands for energy, and control of ROS, including superoxides. Figure 5 with 10 supplements see all Download asset Open asset FABPi significantly induce reactive oxygen species, impair MM cell growth and induce apoptosis. (A) Reactive oxygen species measured by MFI (mean fluorescent intensity) with CellROX Green staining at 72 hr in MM.1S cells. TBHP is positive control. (B) Superoxide levels shown as MFI, determined with MitoSOX staining, at 72 hr in MM.1S cells. (C) MM.1S cell cycle states with the FABPi alone (50 µM) or in combination (50 µM of each). (D) Apoptosis in MM.1S cells with FABPi as in C. Data are mean ± SEM unless otherwise stated and represent averages or representative runs of at least three experimental repeats. One-way ANOVA with Dunnett’s multiple comparison test significance shown as *p<0.05. **p<0.01. ***p<0.001. ****p<0.0001. ATCC MM.1S cells were used for these experiments. Please see 10 supplements to Figure 5. We next investigated FABP inhibitor effects on MM cell cycle and apoptosis. In GFP+/Luc +MM.1 S, FABPi combination treatment increased the G0/G1 population at 24, 48, and 72 hr, and decreased G2/M at 48 and 72 hr, suggesting a G0/G1 arrest and a negative impact on cell cycle progression (Figure 5C, Figure 5—figure supplement 7). FABPi combination treatment also increased apoptosis in GFP+/Luc +MM.1 S cells at all three time points, and SBFI-26 did as well at 72 hr (Figure 5D). To determi