Peroxisome disruption alters lipid metabolism and potentiates antitumor response with MAPK-targeted therapy in melanoma

Melanomas reprogram their metabolism to rapidly adapt to therapy-induced stress conditions, allowing them to persist and ultimately develop resistance. We report that a subpopulation of melanoma cells tolerate MAPK pathway inhibitors (MAPKis) through a concerted metabolic reprogramming mediated by peroxisomes and UDP-glucose ceramide glycosyltransferase (UGCG). Compromising peroxisome biogenesis, by repressing PEX3 expression, potentiated the proapoptotic effects of MAPKis via an induction of ceramides, an effect limited by UGCG-mediated ceramide metabolism. Cotargeting PEX3 and UGCG selectively eliminated a subset of metabolically active, drug-tolerant CD36+ melanoma persister cells, thereby sensitizing melanoma to MAPKis and delaying resistance. Increased levels of peroxisomal genes and UGCG were found in patient-derived MAPKi-relapsed melanomas, and simultaneously inhibiting PEX3 and UGCG restored MAPKi sensitivity in multiple models of therapy resistance. Finally, combination therapy consisting of a newly identified inhibitor of the PEX3-PEX19 interaction, a UGCG inhibitor, and MAPKis demonstrated potent antitumor activity in preclinical melanoma models, thus representing a promising approach for melanoma treatment.

. ROS metabolism and mitochondrial dysfunction do not explain the increased sensitivity of PEX3-defficient melanoma cells to MAPK inhibition. (A) DCFDA relative mean fluorescence intensity (MFI) in human melanoma cells following PEX3 knockdown and treatment of indicated MAPK-targeted therapy agents (n=3). Equal volume of DMSO was added in the control groups. (B, C) Percent apoptosis (PI + /Annexin V + , PI -/Annexin V + ) detected in (B) siPEX3-or siCtrl-transfected A375M cells and (C) D4M.3a Cas9-Ctrl, 6D and 9G cells, following vemu or DMSO treatment with or without the presence of NAC. Corresponding H2O2-treated cells were cultured with or without NAC as positive controls (n=3). (A-C) Two-way ANOVA. (D) Schematic showing circularity (left) and aspect ratio (right) of a mitochondrion. (E, F) Representative electron microscope images of (E, left) siPEX3-or siCtrl-transfected A375M cells with or without vemu treatment, or (F, left) D4M.3a Cas9-Ctrl, 6D and 9G cells. The analyses of mitochondrial circularity and aspect ratio in each condition are presented (right). (E) Two-way ANOVA. (F) One-way ANOVA. (G) Seahorse analysis assessing mitochondrial oxygen consumption rate (OCR) in A375M cells following siPEX3 or siCtrl transfection and subsequent vemu or DMSO control treatment (n=3). All data are presented as mean ± SD. Figure S5. CD36 marks a distinct population of melanoma persister cells. (A) Violin plot of scRNA-seq data highlighting the distribution of CD36 and the cutoff line set to distinguish CD36 -(< 2.2) and CD36 + (≥ 2.2) cells in different phases of MAPKi treatments. (B) A total of 674 melanoma cells (Rambow 2018 dataset) were projected in a two-dimensional space by t-SNE, comparing the distribution of CD36 + cells (left) defined by high CD36 expression as a single marker (normalized gene expression ≥ 2.2, colored in red) versus SMCs (right) defined by AUCell analysis (SMC AUCell score ≥ 0.05, colored in purple). (C) Dynamics of the different melanoma cell states at the indicated time points. (D) Individual growth of A375M-derived melanomas treated with combined BRAF/MEK inhibitors (PLX4720+cobi), related to Figure 4E. (E) Percentage of CD36 + cells in total (CD45 -) tumor cells, isolated from A375M-derived melanomas following PLX4720+cobi treatment for 8 days. Two-sided unpaired t-test. Data represent mean ± SEM. (F, G) Percentage of CD36 + populations in A375M cells following (F) indicated MAPK inhibitors treatment for 48 hours, or (G) vemu treatment for indicated time. Representative flow cytometric images are presented below (n=3). One-way ANOVA. (H) Representative image of parental (unsorted), CD36 -and CD36 + persister A375M cells following vemu treatment for indicated days (representative of n=3). (I) Relative survival of CD36 + versus CD36 -persister A375M cells following C2-Cer treatment at indicated doses, measured by crystal violet staining. Representative images of crystal violet-stained cells are presented (n=3). (H, I) CD36 -and CD36 + persister A375M cells were sorted after 48-hour vemu treatment (see Figure 5A).   Table S1). Representative flow cytometric histograms are presented below (n=3). Two-sided unpaired t-test. Data are presented as Mean ± SD.  Table S1). Data represent mean ± SD (n=3). Two-way ANOVA. (C, D) Individual growth of D4M.3a Cas9-Ctrl-or 9G-derived melanomas in mice receiving indicated treatment, related to Figure 7C and 7E, respectively. Mice were treated with PPMP or vehicle (C) after tumors relapsed on PLX4720, annotated with red or grey triangles, or (D) after relapsed tumor reached a volume of 1,300 mm 3 . Median survival is indicated in each graph. All mice were kept on PLX4720 chow after PLX4720 treatment initiated when individual tumor first reached a volume of 200 mm 3 . Number of biological replicates (mice) is indicated in each graph.
Tumor initiation was determined once palpable tumors were formed. Tumors were then measured in length (L) and width (W). Tumor volumes (V) were calculated based on the formula V=3.1416/6*L*W 2 . For PLX4720 treatment, mice were switched to a special diet containing PLX4720 (AIN-76A rodent diet with 417mg PLX4720/kg, cohort, mice were kept on PLX4720 chow and were co-treated with PPMP by intraperitoneal injection at 5 mg/kg every 36 hours. Treatments were initiated when 1) the initial tumor volume of PLX4720-naïve mice reaches 200 mm 3 ( Figure S4G, S4J), or 2) tumors had relapsed after PLX4720 treatment (tumor volume increase ³ 10%, Figure   7B), or 3) the relapsed tumors had reached a size of 1300 mm 3 after PLX4720 treatment ( Figure 7D). For the NNC treatment cohort ( Figure S9A), when the tumor volume reached approximately 200 mm 3 , female NOD/SCID mice were switched to PLX4720 diet and simultaneously treated with NNC by intraperitoneal injection at 20 mg/kg, and/or with PPMP by intraperitoneal injection at 5 mg/kg every 48 hours. For the PLX4720-relapsed cohort ( Figure   S9C), mice were kept on PLX4720 diet until tumors relapse. NNC and PPMP treatments were then initiated when relapsed tumors reached a volume of approximately 400 mm 3 . For Figure 9A, PLX4720+cobi and/or PPMP+NNC treatments were initiated when A375M-dervied tumors reached an average size of 700 mm 3 and all mice were sacrificed on Day 10 after treatment initiation. For the 1205Lu-VCDR (BRAF/MEKi dual-resistant) cohort ( Figure   S9E), male NOD/SCID mice were used and PLX4720+cobi treatment were initiated once 1205Lu-VCDR-derived tumors were palpable. When the tumor volume reached approximately 250 mm 3 , PPMP+NNC treatments were initiated while mice were kept on PLX4720 diet and treated with cobi simultaneously. Unless otherwise indicated, mice bearing D4M.3a-derived melanomas were sacrificed when the tumor volume reached 1600 mm 3 ; mice bearing A375M-derived melanomas were sacrificed when the tumor volume reached 1000 mm 3 ; and mice bearing 1205Lu-VCDR-derived melanomas were sacrificed when the tumor volume reached 1500 mm 3 .

Quantitative real-time PCR
Cultured cells were pelleted, and RNA was prepared using the E.Z.N.A. total RNA isolation kit (OMEGA Bio- Tek). RNA concentrations were then quantified using a NanoDrop spectrophotometer (ThermoFisher Scientific) and cDNA was prepared from 1mg of total RNA using iScript cDNA Synthesis Kit (Bio-Rad). Target genes were quantified using the Applied Biosystems 7500 Fast Real-Time PCR System with SYBR Green real-time PCR master mix (Applied Biosystems). Two housekeeping genes were used for each assay. Primers used for qPCR are listed in Table S4.

Transmission electron microscopy (TEM)
A375M cells were allowed to grow for 48 hours following siRNA transfection to reach a confluence of 80%.
D4M.3a Cas9-Ctrl (Pex3 +/+ ), Pex3 +/-Clones 6D and 9G cells were cultured to 70-80% confluence. Cells were then trypsinized and resuspended in twice the volume of culture media. Cells were initially centrifuged in 15 mL conical centrifuge tubes, and approximately 3 x10 6 cells per cell line were washed in 1x PBS, then resuspended, and transferred to 1.5 mL microcentrifuge tubes. Cells were then centrifuged at 100g for 10 minutes, washed twice in 1 mL 1x PBS, followed by aspiration of PBS, retaining cell pellets. Using 2% glutaraldehyde fixation buffer (made by mixing 0.4ml glutaraldehyde 50%, 5ml 0.2M cacodylate buffer pH7.2 and 4.6mL distilled water), pellets were fixed for one hour at 4°C. Following fixation, pellets were centrifuged at 100g for 5 minutes, and washed twice with wash/storage buffer (made by mixing equal parts of 0.2M cacodylate buffer pH7.2 with distilled water). Pellets were then processed and imaged according to previously published methods (1).

Cellular and mitochondrial bioenergetics measurements
Bioenergetic profiling was performed according to Agilent Seahorse XF Cell Mito Stress Test Kit Guidebook.
Seahorse XF-96 cell culture microplates were first coated with a 50μg/ml working stock of poly-D-lysine and incubated at room temperature for 1-2 hours. 200μL of sterile water per well was added and aspirated from each well, then plates were left to air dry in an incubator without the plate lid. 20,000 A375M cells were transfected with siCtrl or siPEX3 and seeded in each well overnight. The next day, cells were changed to DMSO-or vemucontaining media, then transferred to a 37°C cell incubator with 20% O2 and 5% CO2 for approximately 18h. Cells were then washed twice in 1x PBS and incubated with Seahorse XF medium (120μL), and incubated for 1h in a non-CO2 37 °C incubator. XF96 sensor cartridges were placed on top of each well and 20μL, 22μL, and 25μL of oligomycin (2.5μM final), Carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone FCCP (2.0μM final), and rotenone/antimycin A (0.5μM final), were added to ports A, B, and C respectively. Plates were loaded and analyzed for cellular and mitochondrial bioenergetics according to the guidebook, alongside normalization to cell number.

Targeted Liquid chromatography tandem mass spectrometry (LC-MS-MS)
Following indicated treatments, cells were scrapped and washed twice with cold 1x PBS. Each sample was then separated into two parts: 90% cells were collected in a borosilicate tube for LC-MS-MS analysis to assess ceramide and HexCer concentrations, while 10% cells were collected in an Eppendorf tube for Bradford assay to assess total protein concentration for final normalization.
The samples were mixed by vortex and centrifuged at 3,000 RPM for 10 min at 4℃ to separate the aqueous (upper) and organic (lower) phases. The lower organic phase was carefully transferred into a separate test tube.
The remaining material underwent a second round of extraction with an additional 1mL of chloroform. Organic phases were combined and dried under ultra-pure nitrogen gas. The final material was solubilised with 125μL of reconstitution solution (chloroform:methanol 1:9). Ceramide and HexCer species were subsequently analyzed by LC-MS-MS using Acquity Premier UPLC I-Class with Xevo TQ-S micro-System (Waters, MA, USA). Data were acquired by MassLynx software and quantified by TargetLynx software. Lipid concentrations were normalized to protein quantity, and relative lipid concentrations were presented as nmol lipid per mg protein.

Access and re-analysis of the Banerjee 2021 yeast two-hybrid drug screen data
Raw data of the Banerjee 2021 LOPAC yeast two-hybrid drug screen (OD600 reading) were retrieved from Supplementary Table 1 of the original publication (2). Figure S8A was generated in the same manner as Figure 2 in the original publication using the mean OD600 data of two independent screens (2). Percent growth inhibition was then calculated as 100-100*(OD600: compound/OD600: DMSO). To assess the activity of each compound against human PEX3-PEX19 binding, the percent inhibition of the BD-PEX3/AD-PEX19-expressing yeast (%Inhibition PEX3-PEX19) was compared to the percent inhibition of the pGAD424/pGBT9-expressing yeast (%Inhibition general growth) grown in the presence of the same compound. The relative inhibition of human PEX3-PEX19 binding ( Figure S8B, Y axis) was calculated as %Inhibition PEX3-PEX19 -%Inhibition general growth.

PEX3 docking
Docking studies were performed with PEX3 and NNC using Autodock Vina (3) based on the crystal structures of enzymes as deposited in the RCSB Protein Data Bank (https://www.rcsb.org/). NNC was centered at the location of PEX19-binding site on PEX3, and genetic algorithm runs were performed for the ligand and receptor.

Colony formation assay
MAPKi-resistant melanoma cells were seeded into 6-well plates at 100,000 cells per well in the presence of indicated MAPK inhibitors. The next day, NNC (4μM), PPMP (see Table S1), or DMSO control were added to the