Targeting the MARCH5-MFN2 axis to enhance mitochondrial fusion and sensitize multiple myeloma cells to venetoclax

Abstract

Background: Accumulating evidence suggests that mitochondrial fission and fusion events are imbalanced in cancer due to defective activity of their key regulators. In this study, we investigated the functional role of the E3 ubiquitin ligase Membrane-Associated Ring-CH-Type Finger 5 (MARCH5) in regulating cell growth, metabolic reprogramming and drug resistance in multiple myeloma (MM) through the negative regulation of the mitochondrial fusion driver mitofusin 2 (MFN2).

Methods: Cell viability and apoptosis were evaluated in MM cell lines or in co-culture with stromal cells using the CellTiter-Glo® Cell Viability Assay and Annexin V/7-AAD staining, respectively. Clonogenic potential was assessed using methylcellulose-based colony formation assays. Protein stability was determined via cycloheximide chase experiments, while protein-protein interactions by co-immunoprecipitation. Mitochondrial ultrastructure was analyzed by transmission electron microscopy. Oxygen consumption was measured using high-resolution respirometry in live cells. Transcriptomic profiling was performed using the Illumina NGS platform, and mRNA and protein levels were quantified by quantitative RT-PCR and Western blot, respectively. In vivo anti-tumor efficacy was evaluated in NOD-SCID mice subcutaneously engrafted with MM cells, using an MFN2-inducible model or following intraperitoneal administration of leflunomide. Immunohistochemistry was used to analyze tumor xenografts and mouse organs.

Results: Knockdown of MARCH5 led to a pronounced elongation of mitochondria accompanied by increased expression of MFN2, likely resulting from reduced MARCH5-mediated ubiquitylation. Functionally, silencing MARCH5 impaired mitochondrial oxidative phosphorylation (OXPHOS) and reduced ATP production, ultimately leading to mitochondrial dysfunction and apoptosis in MM cells. Notably, similar phenotypic and functional effects were observed following either genetic overexpression or pharmacological activation of MFN2 using leflunomide, both in vitro and in vivo in a murine xenograft model of MM. Transcriptomic profiling of MARCH5-depleted cells revealed downregulation of gene sets associated with mitochondrial electron transport chain (ETC) and ATP synthesis, pathways implicated in the development of venetoclax resistance. Consistently, both MARCH5 knockdown and MFN2 upregulation enhanced the sensitivity of MM cells to venetoclax.

Conclusion: Shifting mitochondrial dynamics toward fusion by targeting the MARCH5-MFN2 axis impairs ETC and OXPHOS, thereby sensitizing MM cells to venetoclax. These findings provide preclinical evidence for the potential therapeutic use of MFN2 inducers to enhance venetoclax responsiveness of MM patients.

Keywords: MARCH5; Mitochondrial dynamics; Multiple myeloma; Venetoclax.

Fig. 1

MARCH5 is highly expressed in MM cell lines and modulates mitochondrial dynamics. A Western Blot analysis of MARCH5 protein expression in proteasome inhibitor (PI) sensitive and resistant isogenic MM cell lines; GAPDH served as a loading control. B TEM analysis of mitochondrial structure and morphology in AMO-BZB cells, after electroporation with 100 nM siMARCH5#1, siMARCH5#3 or negative control siRNA (NC). Mitochondrial length was calculated as the average of acquisitions from different samples. ****p < 0.0001. C Western blot analysis of DRP1, MFF and MFN2 expression in AMO-BZB cells after electroporation with 100 nM siMARCH5#1, siMARCH5#3 or negative control siRNA (NC); α-TUBULIN or GAPDH were used as loading control. D Western blot analysis of DRP1, MFF and MFN2 expression in AMO and H929 cells after lentiviral expression of MARCH5-FLAG cDNA (OE MARCH5), or the corresponding empty vector (EV); GAPDH was used as loading control

Fig. 2

MARCH5 negatively regulates mitochondrial fusion by promoting MFN2 ubiquitination and proteasomal degradation. A Western blot analysis of MFN2 and FLAG-tagged MARCH5 proteins in whole-cell lysates (input) and in anti-FLAG immunoprecipitation products from AMO cells following lentiviral expression of MARCH5-FLAG cDNA (OE MARCH5) or the corresponding empty vector (EV). B Fluorescence microscopy analysis of AMO-BZB cells stained for endogenous MFN2 (green) and MARCH5 (red); nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI). C Western blot analysis of MFN2 in EV- or MARCH5-FLAG-transduced AMO cells following treatment with 35 µg/mL cycloheximide for the indicated time points. Normalized MFN2 protein levels are shown in the accompanying graph. D Western blot analysis of ubiquitin in whole-cell lysates (input) and in anti-FLAG immunoprecipitation products from EV or MARCH5-FLAG-transduced AMO cells, 4 h after treatment with the proteasome inhibitor MG132 (20 µmol/L)

Fig. 3

MARCH5 regulates mitochondrial dysfunction and oxidative phosphorylation in MM cells. A Bubble chart for GO pathway enrichment analysis showing the top 10 down-regulated datasets. Bubble chart was generated in R (v4.3.2). Combined scores are shown by the circle area, while the circle color represents the range of the adjusted p-value. B Quantitative RT-PCR analysis of MARCH5, NDUFA2, NDUFS2, NDUFA8, SDHA, UQCR11 and COX5A mRNA expression levels in AMO cells, after electroporation with 100 nM siMARCH5#1 or negative control siRNA (NC). The results show the average ± SD of mRNA expression levels after normalization with β-actin and ΔΔCt calculations. *p < 0.05; **p < 0.001. C Quantitative RT-PCR analysis of MARCH5, NDUFA2, NDUFS2, NDUFA8, SDHA, UQCR11 and COX5A mRNA expression levels in AMO-BZB cells, after electroporation with 100 nM siMARCH5#1 or negative control siRNA (NC). The results show the average ± SD of mRNA expression levels after normalization with β-actin and ΔΔCt calculations. **p < 0.001 D Western blot analysis of NDUFS1, COX1, UQCRFS1, NDUFAB1 and COX4 proteins in AMO-BZB cells, 48 h after electroporation with 100 nM siMARCH5#1, siMARCH5#2 or negative control siRNA (NC); GAPDH was used as loading control. Histogram bars show the densitometric analysis expressed as fold change relative to the NC. E Histogram bars reporting real time Oxygen Consumption Rate (OCR) measurement using OROBOROS on AMO-BZB cells, 48 h after electroporation with 100 nM siMARCH5#1, or NC negative control siRNA. The graph reports basal respiration, spare capacity, maximal respiration, leak state, ATP production, and non-mitochondrial respiration. *p < 0.05; **p < 0.01; F Trypan blue cell count in AMO-BZB cells 48 h after electroporation with siMARCH5#1, siMARCH5#2, siMARCH5#3, or negative control (NC) siRNA. *p < 0.05; **p < 0.01. G Flow cytometry analysis of Annexin V/7-AAD-stained AMO-BZB cells, 48 h after electroporation with 200 nM siMARCH5#1, siMARCH5#2, siMARCH5#3, or negative control (NC) siRNA. Histogram bars represent the total Annexin V-positive population, including both Annexin V-positive only and Annexin V/7-AAD double-positive cells. *p < 0.05. H–I Western blot analysis of PARP, CASPASE 3, BAD, BIK, BAK, BID, PUMA, and NOXA in AMO-BZB cells, 48 h after electroporation with 100 nM siMARCH5#1, siMARCH5#2, siMARCH5#3, or negative control (NC) siRNA. GAPDH was used as a loading control

Fig. 4

Promotion of mitochondrial fusion via genetic MFN2 upregulation replicates MARCH5 loss and triggers anti-myeloma responses. A Western blot analysis of NDUFS1, COX1, COX4, and MFN2 proteins in independent AMO-BZB cell clones (CL.15 and CL.19) with doxycycline-inducible MFN2 expression. GAPDH was used as a loading control. B Histogram bars reporting real time Oxygen Consumption Rate (OCR) measurement using OROBOROS on AMO-BZB cell clones (CL.19) with doxycycline-inducible MFN2 expression, 72 h after treatment. The graph reports basal respiration, spare capacity, maximal respiration, leak state, ATP production, and non-mitochondrial respiration. *p < 0.05. C) Cell viability was assessed by Cell Titer Glo assay in AMO-BZB cell clones (CL.15 and CL.19) with doxycycline-inducible MFN2 expression, 6 and 9 days after treatment with doxycycline. *p < 0.05; **p < 0.01. D Flow cytometry analysis of Annexin V/7AAD-stained AMO-BZB cell clone (CL.19) with doxycycline-inducible MFN2 expression, 6 and 9 days after treatment with Doxycycline. E Western blot analysis of PUMA and NOXA expression in AMO-BZB cell clones (CL.15; CL.19); GAPDH was used as loading control. F In vivo tumor growth evaluation of subcutaneous xenografts with AMO-BZB (CL.19), receiving doxycycline or vehicle as control; administrations were performed via drinking water. Average ± SD of the tumor volume for each group is shown; p-values were obtained using two-tailed t-test. **p < 0.001. G Kaplan–Meier curves relative to AMO-BZB xenografts (CL.19) receiving doxycycline or vehicle as control (log-rank-test). Survival was evaluated from the first day of treatment until death or sacrifice. Percentage of mice alive is shown. H Western blot analysis of MFN2 protein levels in tumors retrieved from doxycycline or vehicle-treated mice; α-tubulin was used as loading control. Densitometric analysis of protein levels represents the mean ± SD from doxycycline-or vehicle-treated mice after normalization. I IHC analysis (20 × magnification) of Ki-67 expression in xenografts from AMO-BZB cells (CL.19) with doxycycline-inducible MFN2, retrieved from mice before sacrifice; representative images are reported. L Western blot analysis of BAD, BID, NOXA and PUMA protein levels in retrieved xenografts; GAPDH was used as loading control. Densitometric analysis of protein levels represents the mean ± SD from doxycycline-or vehicle-treated mice after normalization *p < 0.05; **p < 0.01

Fig. 5

Induction of MFN2 via leflunomide replicates MARCH5 loss and triggers anti-MM responses. A Western blot analysis of MFN2 was performed in AMO cells, 72 h after treatment with increasing doses of Leflunomide or DMSO as vehicle; GAPDH was used as loading control. B Heatmap shows cell viability assessed by Cell Titer Glo assay in MM cells treated with Leflunomide for 48 h. Viable cells are reported as percentage of vehicle. C Cell viability assessed by Cell Titer Glo assay in AMO and H929 cells treated with Leflunomide for 72 h, and then co-cultured with HS5 cells for 48 h. Viable cells were expressed as the percentage of vehicle. Data represent the average ± SD of three independent experiments. *p < 0.05. D Western blot analysis of NDUFS1, NDUFAB1, and COX4 expression in AMO cells, 72 h after treatment with Leflunomide; GAPDH was used as loading control. E Histogram bars represent real time Oxygen Consumption Rate (OCR) measurement using OROBOROS on AMO cells, 48 h after treatment with Leflunomide. The graph reports basal respiration, spare capacity, maximal respiration, leak state, ATP production, and non-mitochondrial respiration. *p < 0.05. F Western blot analysis of PUMA and NOXA expression in AMO cells, 72 h after treatment with Leflunomide; GAPDH was used as loading control. G In vivo tumor growth evaluation of subcutaneous AMO xenograft receiving i.p. Leflunomide (20 mg/kg), or vehicle as control; administrations were performed 5 days a week for a total of four weeks. Average ± SD of the tumor volume for each group is shown; p-values were obtained using two-tailed t-test. *p < 0.05. H Kaplan–Meier curves relative to i.p. Leflunomide-treated AMO xenografts compared to control xenografts (log-rank-test, *p = 0.0094). Survival was evaluated from the first day of treatment until death or sacrifice. Percentage of mice alive is shown. I Western blot analysis of MFN2 protein levels in tumors retrieved from Leflunomide- or vehicle-treated mice; α-tubulin was used as loading control. Densitometric analysis of protein expression represents the mean ± SD from Leflunomide-or vehicle-treated tumor xenografts after normalization. **p < 0.01. L IHC analysis (20 × magnification) of Ki-67 expression in AMO xenografts retrieved from mice, four weeks after Leflunomide treatment; representative images are reported. M Western blot analysis of BAK, BIK, BIM, BID, NOXA and PUMA protein levels in tumors retrieved from Leflunomide-or vehicle-treated mice; α-tubulin was used as loading control. Densitometric analysis of protein levels represents the mean ± SD from Leflunomide-or vehicle-treated mice after normalization

Fig. 6

MARCH5 influences Venetoclax response in MM cells. A Cell viability was assessed by CTG assay in AMO, AMO-BZB, H929, and H929-BZB cells 48 h after treatment with increasing doses of Venetoclax compared to vehicle (DMSO). B Cell viability was assessed by CTG assay in AMO-BZB cells after electroporation with 100 nM siMARCH5#1 or negative control (NC) siRNA, alone or in combination with Venetoclax. Viable cells are expressed as a percentage relative to vehicle-treated cells. *p < 0.05; **p < 0.001. C Heatmap showing combination indexes (CI), determined using SynergyFinder software, 48 h after combined treatment with Venetoclax and Leflunomide in AMO cells. D Cell viability was assessed by CTG assay in AMO cells after lentiviral expression of MARCH5 cDNA (OE MARCH5) or the corresponding empty vector (EV), 48 h after treatment with increasing doses of Venetoclax compared to vehicle (DMSO). **p < 0.01. E Histogram showing real-time Oxygen Consumption Rate (OCR) measurements using the OROBOROS system in AMO cells after lentiviral expression of MARCH5 cDNA (OE MARCH5) or EV, 48 h after Venetoclax treatment. The graph reports basal respiration, spare respiratory capacity, maximal respiration, leak state, ATP production, and non-mitochondrial respiration. *p < 0.05; **p < 0.01. F Flow cytometry analysis of mitochondrial ROS levels in AMO cells stained with MitoSOX Red after lentiviral expression of MARCH5 cDNA (OE MARCH5) or EV, followed by Venetoclax treatment. Histogram is representative of three independent experiments. Mean fluorescence intensity (MFI) values are reported. *p < 0.05; **p < 0.01. G Flow cytometry analysis of Annexin V/7-AAD-stained AMO cells after lentiviral expression of MARCH5 cDNA (OE MARCH5) or EV, treated alone or in combination with increasing doses of Venetoclax. H Western blot analysis of PARP1 and CASPASE 3 cleavage in AMO cells after lentiviral expression of MARCH5 cDNA (OE MARCH5) or EV, 48 h after treatment with Venetoclax or vehicle (DMSO). GAPDH was used as a loading control. I Flow cytometry analysis of mitochondrial membrane potential using TMRM staining, in AMO cells after lentiviral expression of MARCH5 cDNA (OE MARCH5) or EV, followed by Venetoclax treatment. Histogram is representative of three independent experiments. MFI values are reported. *p < 0.05; **p < 0.01