Targeting fatty acid oxidation enhances response to HER2-targeted therapy

Abstract

Metabolic reprogramming, a hallmark of tumorigenesis, involves alterations in glucose and fatty acid metabolism. Here, we investigate the role of Carnitine palmitoyl transferase 1a (Cpt1a), a key enzyme in long-chain fatty acid (LCFA) oxidation, in ErbB2-driven breast cancers. In ErbB2+ breast cancer models, ablation of Cpt1a delays tumor onset, growth, and metastasis. However, Cpt1a-deficient cells exhibit increased glucose dependency that enables survival and eventual tumor progression. Consequently, these cells exhibit heightened oxidative stress and upregulated nuclear factor erythroid 2-related factor 2 (Nrf2) activity. Inhibiting Nrf2 or silencing its expression reduces proliferation and glucose consumption in Cpt1a-deficient cells. Combining the ketogenic diet, composed of LCFAs, or an anti-ErbB2 monoclonal antibody (mAb) with Cpt1a deficiency significantly perturbs tumor growth, enhances apoptosis, and reduces lung metastasis. Using an immunocompetent model, we show that Cpt1a inhibition promotes an antitumor immune microenvironment, thereby enhancing the efficacy of anti-ErbB2 mAbs. Our findings underscore the importance of targeting fatty acid oxidation alongside HER2-targeted therapies to combat resistance in HER2+ breast cancer patients.

Fig. 1. Cpt1a ablation impairs ErbB2-driven mammary tumor progression and metastasis.

a Schematic of genetically engineered mouse model. b Lysates from end-stage NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L mice were immunoblotted with antibodies against Cpt1a and α-Tubulin. n = 5 mice per genotype in triplicate. c Kaplan–Meier analysis of mammary tumor onset in mice with wild-type Cpt1a alleles (NIC/Cpt1a^+/+, n = 20) and homozygous (NIC/Cpt1a^L/L, n = 20) conditional Cpt1a alleles. ****p < 0.0001; log rank test. d Total tumor volume for mice as in (a). n = 20 mice per genotype - ****p < 0.0001; unpaired, two-tailed Student’s t-test. e Left panel - Hematoxylin and eosin (H&E) staining of NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L tumors. Images are representative of ten independent tumors of each genotype. Scale bar represents 100 μm. Right Panel – Quantification of the tumor necrosis area as a percentage of the total tumor area. n = 15 mice per genotype - *p  <  0.05, ****p < 0.0001; unpaired, two-tailed Student’s t-test. Incidence (f) and average burden of lung metastasis (g). n = 15 mice per genotype - **p < 0.01; unpaired, two-tailed Student’s t-test. All error bars are expressed as mean values ± SD. Source data are provided as a Source Data file.

Fig. 2. Cpt1a-deficient cells exhibit metabolic dysfunction with reduced proliferation, OXPHOS, glycolysis and FAO.

a Proliferation was assayed in NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cells by Incucyte for 96 h. Data were normalized to confluency at t = 0. n = 3 cell lines in triplicate. ****p < 0.0001; unpaired, two-tailed Student’s t-test. b Cells in (a) were treated with Cpt1a inhibitor, Etomoxir (Etom), at the indicated concentrations or with dimethylsulfoxide (DMSO), and proliferation was assayed using Incucyte. n = 3 cell lines in triplicate - **p < 0.01 and ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. c Quantification of basal oxygen consumption rates (OCR) and extracellular acidification rates (ECAR) of NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cell lines. n = 4 cell lines per genotype in triplicate – ****p < 0.0001; unpaired, two-tailed Student’s t-test. d Lactate levels in conditioned media from NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cells (n = 4 cell lines per genotype in duplicate – ***p < 0.001 and ****p < 0.0001; unpaired, two-tailed Student’s t-test). e Quantification of maximal OCR in the presence of bovine serum albumin (BSA)-control or palmitate and in the presence or absence of Etomoxir (Etom). n = 4 cell lines in quadruplicate - ****p <  0.0001; one-way ANOVA with Tukey’s post-hoc test. f Schematic illustrating palmitic acid carbon tracing in the tricarboxylic acid (TCA) cycle. Shown is the pathway of U-¹³C-palmitate carbons (black) transferred among the molecules of the TCA cycle. Palmitic acid is degraded to Acetyl-CoA (M + 2) via fatty acid oxidation, which reacts with oxaloacetate to form citrate (M + 2). g U-¹³C-palmitate labeling of TCA cycle M + 2 isotopomers as a fraction of total metabolites. Cells were isolated 24 h after U-¹³C-palmitate tracer. n = 3 cell lines in triplicate. ****p < 0.0001; unpaired, two-tailed Student’s t-test. h NIC tumors were immunostained with the indicated antibodies and DAPI. Right panel - Images representative of ten independent mice per genotype. Scale bar represents 100 µm. Left panel - Staining was quantified and normalized to total cell number (DAPI). Minimum 10,000 cells per sample, ****p < 0.0001; unpaired, two-tailed Student’s t-test. All error bars are expressed as mean values ± SD. Source data are provided as a Source Data file.

Fig. 3. Deletion of Cpt1a alters glucose metabolism and impairs the reduction of electron carriers.

a Glucose consumption in conditioned media from NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cell lines (n = 4 cell lines per genotype in duplicate – ****p < 0.0001; unpaired, two-tailed Student’s t-test). b Glut1 (Slc2a1) transcript levels in NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cells normalized to Actb. n = 4 per genotype, analyzed in triplicate - ***p < 0.001; unpaired, two-tailed Student’s t-test. c NIC tumors were immunostained with the indicated antibodies and DAPI. Right panel - Images representative of ten independent mice of each genotype. Scale bar represents 100 µm. Left panel - Staining was quantified and normalized to total cell number (DAPI). Minimum 10,000 cells per sample, **p < 0.01; unpaired, two-tailed Student’s t-test. d Fractional ion abundance of glycolytic (Lactate, Alanine, Serine, and Pyruvate) and TCA Cycle (Citrate, a-KG, Succinate, Malate) intermediates following a pulse in U-¹³C-glucose of NIC/Cpt1a^L/L cells normalized to wild-type (NIC/Cpt1a^+/+) cells. n = 3 per genotype, analyzed in triplicate - ***p < 0.001, ****p < 0.0001; unpaired, two-tailed Student’s t-test. e Relative abundance of oxidized (nicotinamide adenine dinucleotide - NAD⁺, nicotinamide adenine dinucleotide phosphate - NADP⁺ and flavin adenine dinucleotide - FAD) and reduced (NADH and NADPH) forms of electron carriers in NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cells. n = 3 per genotype, analyzed in triplicate - ****p < 0.0001; unpaired, two-tailed Student’s t-test. f ATP Concentration in NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cells. n = 3 per genotype, analyzed in triplicate - ****p < 0.0001; unpaired, two-tailed Student's t-test. All error bars are expressed as mean values ± SD. Source data are provided as a Source Data file.

Fig. 4. Cpt1a-deficient cells undergo oxidative stress and upregulate Nrf2 expression and activity in vivo.

a Left panel - Representative fluorescence microscope image of NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cells with CellROX Orange reagent. Scale bar represents 200 µm. Right panel – Quantification of relative CellROX Fluorescence Intensity (absorgance units - A.U.) and percentage of CellROX labeled cells. n = 3 cell lines per genotype ***p < 0.001; unpaired, two-tailed Student’s t-test. b Unsupervised hierarchical clustering analysis of genes differentially up-regulated (red) and down-regulated (blue) in NIC/Cpt1a^L/L cells compared to NIC/Cpt1a^+/+ controls (n = 2 or 4 per genotype, analyzed in duplicate). c ENCORE and ChEA Consensus Transcription Factor analysis of up-regulated transcription factors in Cpt1a^L/L cells compared to wild-type NIC cells using Enrichr. d Left panel – End-stage tumors were stained with Nrf2 and ErbB2-specific antibodies and DAPI (nuclei). Scale bar represents 100 µm. Right panel – Nrf2 immunostaining was quantified and normalized to ErbB2-positive cells. n = 5 mice per genotype (minimum 10,000 total nuclei analyzed per mouse). ***p < 0.001; unpaired, two-tailed Student’s t-test. e Quantitative real-time polymerase chain reaction (QRT-PCR) analysis of Nrf2 target gene expression in NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cells. Nrf2 target expression was normalized to that of Actb. n = 3 cell lines per genotype in triplicate - ***p < 0.001, ****p < 0.0001; unpaired, two-tailed Student’s t-test. All error bars are expressed as mean values ± SD. Source data are provided as a Source Data file.

Fig. 5. Nrf2 modulates glucose uptake and metabolism in Cpt1a-deficient cells.

a NIC cells were treated with the Nrf2 inhibitors, Brusatol (100 nM), ML385 (2μM)), or vehicle (DMSO). Proliferation was assessed by Incucyte after 72 h. Data normalized to confluency at t = 0. n = 3 cell lines in triplicate - **p < 0.01, ****p <  0.0001; one-way ANOVA with Tukey’s post-hoc test. Glucose consumption (b) and QRT-PCR analysis of Glut1 gene expression (c) in cells treated as in (a). n = 3 cell lines per genotype in triplicate – *p < 0.05, ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. d Basal, maximal (Carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone – FCCP), ATP-synthesis coupled (Oligomycin A), and non-mitochondrial (rotenone/ antimycin A) oxygen consumption rates (OCRs) of NIC cells treated with DMSO or Nrf2 inhibitors. Representative of n = 3 cell lines per genotype per condition. e Quantification of basal OCR and extracellular acidification rates (ECAR) of cells as in (a–e). n = 3 cell lines per genotype in quadruplicate – *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. f NIC cells stably expressing short hairpin RNAs (shRNAs) against luciferase (control – shCon), Nrf2 or Keap1. Proliferation assessed using Incucyte after 96 h. n = 3 cell lines per genotype in triplicate - *p < 0.05, ***p < 0.001, ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. Glucose consumption (g) and QRT-PCR analysis of Glut1 expression (h) in cells as in (f) transduced with indicated shRNAs. n = 3 cell lines per genotype in triplicate – *p < 0.05, **p < 0.01, ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. i OCR analysis of NIC cell lines expressing control, Nrf2 or Keap1 shRNAs as in (d). Representative of n = 3 cell lines per genotype. j Quantification of basal OCR and ECAR of cells as in (e, f). n = 3 cell lines per genotype in triplicate –**p < 0.01, ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. All error bars are expressed as mean values ± SD. Source data are provided as a Source Data file.

Fig. 6. Proliferation of Cpt1a-deficient cells is compromised in high-lipid, low-glucose conditions in vitro or in vivo.

a Proliferation was assayed in NIC cells supplemented with palmitate under variable glucose concentrations for 96 h. Data were normalized to confluency at t = 0. n = 3 cell lines per genotype in triplicate – *p < 0.05, **p < 0.01, ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. b Schematic showing the experimental design of the ketogenic diet orthotopic xenotransplant study using two independent NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L cell lines. Created with BioRender.com c Plasma glucose and ketone levels were measured. n = 6 per group, ****p <  0.0001; one-way ANOVA with Tukey’s post-hoc test. d Tumor mass was assessed at endpoint. n = 6 mice per treatment group, **p < 0.01, ***p < 0.001; one-way ANOVA with Tukey’s post-hoc test. e Tumor burden was determined by weekly caliper measurements till control mice (wild-type NIC tumors on normal chow) reached endpoint (10 weeks). n = 6 mice per treatment group, ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. f Kaplan–Meier survival analysis. n = 6 mice per treatment group, *p < 0.05, **p < 0.01, ***p < 0.001; log rank test. g Quantification of Ki67, cleaved caspase 3 and ErbB2 immunostaining in endpoint tumors. **p < 0.01, ****p <  0.0001; one-way ANOVA with Tukey’s post hoc-test. All error bars are expressed as mean values ± SD. Source data are provided as a Source Data file.

Fig. 7. Cpt1a loss enhances ErbB2 monoclonal antibody response in tumor suppression.

a Kaplan–Meier survival curve of Breast Cancer (BRCA) (n = 1069) and HER2-positive BRCA (n = 666) patients with high or low CPT1A expression (log rank p = 0.0032 and p = 0.00059, respectively) segregated using median cut-off. b ROC-plotter analysis of CPT1A gene expression in Trastuzumab non-responder and responders (n = 120, ROC p-value = 1.2 e-02). Box plot with center line = median, box = 25th–75th quartile, whiskers = maxima/minima, outliers = open circle. c Proliferation assay in NIC cells treated with DMSO or Lapatinib supplemented with Palmitate (left panel) or Ketone (3-OHB, right panel) at varying glucose concentrations. Data normalized to confluency at t = 0. n = 2 cell lines per genotype in triplicate – *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. d Immunocompetent mice bearing orthotopic NIC/Cpt1a^+/+ and NIC/Cpt1a^L/L breast cancer allografts treated with IgG (control Ab) and 7.16.4 Ab (n = 10 per treatment group). Tumor weight (top panel) and representative images (bottom panel) at end-point. n = 10 mice per treatment group, *p < 0.05, **p < 0.01 and ***p < 0.001; one-way ANOVA with Tukey’s post-hoc test. e Top panel – Weekly tumor volume assessment during treatment. n = 10 mice per treatment group, **p < 0.01,***p < 0.001 and ****p < 0.0001; one-way ANOVA with Tukey’s post-hoc test. Bottom panel – immunoblot analysis of Cpt1a levels in endpoint tumors. f Quantification of Ki67 and cleaved caspase 3 by immunostaining endpoint tumors. **p < 0.01, ***p < 0.001, ****p < 0.0001; one-way ANOVA with Tukey’s post hoc-test. g Left panel - End-stage NIC tumors treated with IgG and 7.16.4 mAb immunostained with immune cell markers and DAPI. Images representative of 10 mice per treatment group. Scale bar: 100 μm. Right panel - Quantification of p-Stat1(Y701)⁺/ErbB2⁺, F4/80⁺/CD206⁺ and F4/80⁺/p-Stat1(Y701)⁺ by HALO Analysis. *p < 0.05, ***p < 0.001, ****p < 0.0001; one-way ANOVA with Tukey’s post hoc-test. All error bars are expressed as mean values ± SD. Source data are provided as a Source Data file.

Fig. 8. Inhibition of Cpt1a sensitizes ErbB2-driven tumor cells to HER2-targeted therapy.

Schematic illustrating the role of lipid and glucose metabolism in energy production and proliferation in HER2+ cells. Disruption of fatty acid oxidation (FAO), mediated by Cpt1a, diminishes the reduction of electron carriers, impairs oxidative phosphorylation (OXPHOS) and promotes oxidative stress. Cpt1a-deficient cells exhibit increased glucose uptake and reprogram glucose metabolism to reflect a reliance on the TCA cycle for NAD + /FAD reduction. Inhibiting Cpt1a can enhance the response to the ketogenic diet or HER2 monoclonal antibody therapy by promoting an antitumor immune microenvironment, thereby improving the clinical outcome of HER2-positive breast cancer patients. Created with BioRender.com.