The Role of Vitamin C in Cancer Therapy
Introduction
Today’s post was inspired by the following online blog post on the benefits of high intravenous vitamin C (aka ascorbate) doses in cancer therapy. This has been a rather controversial topic that attracted the attention of none other than two times Nobel prize winning scientist Linus Pauling, who ended up co-authoring a series of studies on the efficacy of vitamin C in treating cancer in the 70s.
His studies were widely criticised and even ridiculed due to poor design, while two subsequent studies published in the New England Journal of Medicine showed no beneficial effect (albeit not under the conditions tested by Pauling) and led to an end of any scientific interest on the potential of vitamin C as an anti-cancer drug in the following decades.
However, interest was revived in the past 15 years, as more knowledge on both tumors and Vitamin C pharmacokinetics became available. This resulted, among others, in a prominent study showing the efficacy of intravenous injections of high doses of vitamin C against KRAS/BRAF mutant cancer cells and a review discussing its possible modes of action of against tumors.
To look more in detail into how vitamin C can affect cells, we selected transcriptomic data from a study published in 2018 for re-analysis on our platform. In this study, the authors observed that vitamin C promoted apoptosis in breast cancer cells from patients and looked at whole transcriptome changes in a human breast cancer cell line.
Clustering Analysis
We started our re-analysis by showing that vitamin C has indeed a discernible effect on cancer cells, resulting in the formation of discrete clusters in a tSNE plot (Figure 1).
We then performed a Differential Gene Expression analysis using both the edgeR and DESeq2 methods and using an FDR threshold of q<0.05, as done in the article. WIth a minimum log fold change (LFC) of 0.5, we identified a total of 333 differentially expressed genes shared between the two methods (Figure 2). Relaxing the LFC to 0.2 resulted in 1,021 shared genes, more similar to the number obtained in the article, with 466 genes downregulated and 555 upregulated.
Among the most upregulated genes were TNFSF10 and CYP1B1, while the most downregulated genes included BNIP3, NDRG1 and ADM (Figure 3). TFRC, PGK1,PDK1, HK2 and BNIP3L, while not among the 12 most downregulated genes, were still significantly downregulated with a LFC>0.5, consistent with the results of the article. The upregulation of TNFSF10, in particular, is central to the article, as it encodes the TNF-related apoptosis-inducing ligand (TRAIL).
Pathway Analysis
To verify if the Gene expression profile induced by vitamin C exposure was consistent with signals induced by apoptosis, we performed a KEGG pathway analysis.
The most significantly upregulated KEGG pathway was “GRAFT_VERSUS_HOST_DISEASE” (q=0.003684), which is related to cell apoptosis (Figure 4). Furthermore, we observed the upregulation of glutathione metabolism (Figure 5), consistent with the theory that vitamin C can kill cancer cells by inducing oxidative stress.
There was also the upregulation of various pathways related to drug metabolism by cytochrome P450. Among the most down regulated pathways were the Ribosome pathway and oxidative phosphorylation. We also observed the downregulation of the glycolysis/glucogenesis pathway (Figure 6).
Although the downregulation was not statistically significant (though it still yielded a q<0.1), it could reflect the postulated ability of vitamin C to inhibit the increased glycolytic activity in cancer cells.
Please note that the figures above have not been produced using Omics Playground. The platform currently supports pathway analysis with WikiPathways, Reactome and GO graph.
Drug Connectivity Analysis
Finally, we took a look at the possible correlation between the profile generated in the breast cancer cell line selected in this study upon exposure to vitamin C and profiles stored in the Drug Connectivity Map database.
Since we were examining the profile generated by the equivalent of an anticancer drug, we looked at positively correlated drug profiles that might mimic its tumoricidal activity. The drug with the highest statistically significant (p<0.05) normalised enrichment score (NES) was torin-2, an mTOR inhibitor (Figure 7A).
Vitamin C was indeed shown to inhibit mTOR and lead to cell death in an ovarian cancer cell study. The second highest scoring hit was trametinib, a MEK-inhibitor (Figure 7B). Interestingly, a combination of a MEK-inhibitor and vitamin C has been used to maintain hypomethylation in mouse embryonic cells (here and here).
Furthermore, hypermethylation is the hallmark of many tumors and Vitamin C has been shown to induce DNA demethylation in cancer cells. It is intriguing to notice that Vitamin C combines the features of two different antitumor mechanisms that have been proposed as an effective combination therapy (here, here and here).
Conclusion
To conclude, the transcriptomic profile in this study supports the proposed mechanisms for the antitumoral properties of vitamin C. Especially when considering its safety and low cost, there are solid arguments in favour of its inclusion (as high dose intravenous injections) in cancer therapy trials. Something that should make us re-evaluate Linus Pauling’s legacy.
About the Author
Axel Martinelli’s academic background is in molecular biology and parasitology. He earned a Ph.D. on the genetics of strain-specific immunity against malaria infections and a master’s degree in bioinformatics with specialization in the analysis of omics data. During his postdoctoral career, he worked on genomics and transcriptomics studies and is currently the head of biology at Bigomics Analytics.