Untargeted Metabolomics Reveals Acylcarnitines as Major Metabolic Targets of Resveratrol in Breast Cancer Cells

Abstract

Background/Objectives: Millions of new diagnoses of breast cancer are made each year, with many cases having poor prognoses and limited treatment options, particularly for some subtypes such as triple-negative breast cancer. Resveratrol, a naturally occurring polyphenol, has demonstrated many anticancer properties in breast cancer studies. However, the mechanism of action of this compound remains elusive, although prior evidence suggests that this compound may work through altering cancer cell metabolism. Our objective for the current study was to perform untargeted metabolomics analysis on resveratrol-treated breast cancer cells to identify key metabolic targets of this compound. Methods: MCF-7 and MDA-MB-231 breast cancer cells were treated with varying doses of resveratrol and extracted for mass spectrometry-based untargeted metabolomics. Data preprocessing and filtering of metabolomics data from MCF-7 samples yielded 4751 peaks, with 312 peaks matched to an in-house standards library and 3459 peaks matched to public databases. Results: Pathway analysis in MetaboAnalyst identified significant (p < 0.05) metabolic pathways affected by resveratrol treatment, particularly those involving steroid, fatty acid, amino acid, and nucleotide metabolism. Evaluation of standard-matched peaks revealed acylcarnitines as a major target of resveratrol treatment, with long-chain acylcarnitines exhibiting a 2-5-fold increase in MCF-7 cells and a 5-13-fold increase in MDA-MB-231 cells when comparing the 100 µM treated cells to vehicle-treated cells (p < 0.05, VIP > 1). Notably, doses below 10 µM showed an opposite effect, possibly indicating a biphasic effect of resveratrol due to a switch from anti-oxidant to pro-oxidant effects as dose levels increase. Conclusions: These findings suggest that resveratrol induces mitochondrial metabolic reprogramming in breast cancer cells in a dose-dependent manner. The biphasic response indicates a potential optimal dosage for therapeutic effectiveness. Further research is warranted to explore the mechanisms underlying these metabolic alterations and their implications for precision nutrition strategies in cancer treatment.

Keywords: breast cancer; metabolomics; mitochondria; resveratrol; triple-negative.

Figure 1

Multivariate analysis using all peaks from MCF-7 samples. (A) PCA plot of all treatments. (B) OPLS-DA plot of cells treated with 0 or 1 µM of resveratrol. (C) OPLS-DA plot of cells treated with 0 or 10 µM of resveratrol. (D) OPLS-DA plot of cells treated with 0 or 100 µM of resveratrol. All multivariate plots were scaled to unit variance. Green = 0 μM, blue = 1 μM, red = 10 μM, yellow = 100 μM.

Figure 2

Enriched metabolic pathways across all treatment groups in MCF-7 samples using the Functional Analysis module in MetaboAnalyst. All peaks for all four treatment groups were uploaded as a peak-intensity table in MetaboAnalyst, and ANOVA p-values were used to determine peaks significantly altered by treatment. The top five pathways by p-value are annotated. A full list of significantly enriched pathways is listed in Table 1. Darker red indicates a larger −log₁₀ (p-value).

Figure 3

Analysis of carnitine/acylcarnitine metabolites in MCF-7 samples. (A) Heatmap of 16 carnitine/acylcarnitine metabolites across all four treatment groups. Data are auto-scaled (mean-centered and divided by standard deviation) for each metabolite. (B) PCA plot of all four treatment groups using the 16 carnitine/acylcarnitine metabolites. Green = 0 μM, blue = 1 μM, red = 10 μM, yellow = 100 μM. (C) Biplot showing the loadings of each carnitine/acylcarnitine metabolite on top of the PCA from Figure 3B. Individual metabolites are indicated in purple. All multivariate plots were scaled to unit variance. Axis labels p(corr) [1],t(corr) [1] and p(corr) [2],t(corr) [2] refer to the correlation-scaled loadings (p) and scores (t) of the first and second principal components, respectively.

Figure 4

Box plots of peak-abundance values from long-chain acylcarnitines in MCF-7 samples across all four treatment groups. Red = 0 μM, green = 1 μM, dark blue = 10 μM, light blue = 100 μM.

Figure 5

Multivariate analysis using all peaks from MDA-MB-231 samples. (A) PCA plot of all treatments. (B) OPLS-DA plot of cells treated with 0 or 25 µM of resveratrol. (C) OPLS-DA plot of cells treated with 0 or 50 µM of resveratrol. (D) OPLS-DA plot of cells treated with 0 or 100 µM of resveratrol. All multivariate plots were scaled to unit variance. Green = 0 μM, blue = 25 μM, red = 50 μM, yellow = 100 μM.

Figure 6

Analysis of carnitine/acylcarnitine metabolites in MDA-MB-231 samples. (A) PCA plot of all four treatment groups using the 12 carnitine/acylcarnitine metabolites. Green = 0 μM, blue = 25 μM, red = 50 μM, yellow = 100 μM. (B) Biplot showing the loadings of each carnitine/acylcarnitine metabolite on top of the PCA from Figure 6A. Individual metabolites are indicated in purple. Axis labels p(corr) [1],t(corr) [1] and p(corr) [2],t(corr) [2] refer to the correlation-scaled loadings (p) and scores (t) of the first and second principal components, respectively. (C) Heatmap of 12 carnitine/acylcarnitine metabolites across all four treatment groups. All multivariate plots were scaled to unit variance. Data are auto-scaled (mean-centered and divided by standard deviation) for each metabolite.

Figure 7

Box plots of peak-abundance values from long-chain acylcarnitines in MDA-MB-231 samples across all four treatment groups. Red = 0 μM, light blue = 25 μM, green = 50 μM, dark blue = 100 μM.