Research Spotlight: A New Pathway for Viral Mimicry Underlying the Anti-Cancer Mechanism of Mutant Isocitrate Dehydrogenase (MIDH1) Inhibition

Meng-Ju Wu, PhD, an instructor in investigation at the Krantz Center for Cancer Research in the Mass General Cancer Center and an instructor of Medicine at Harvard Medical School, is co-lead author of a new study in Science, Mutant IDHI Inhibition Induces dsDNA Sensing to Activate Tumor Immunity.

Hiroshi Kondo, PhD, a postdoctoral fellow at the Krantz Center for Cancer Research in the Mass General Cancer Center, is co-lead author of the study.

Ashwin Kammula, BS, a Computational Associate at the Broad Institute and the Krantz Center for Cancer Research in the Mass General Cancer Center, is co-lead author of the study.

Robert Manguso, PhD, an Assistant Professor at the Krantz Center for Cancer Research in the Mass General Cancer Center, an assistant professor of Medicine at Harvard Medical School and an Associate Member at the Broad Institute, is co-senior author of the study.

Nabeel Bardeesy, PhD, an Associate Professor at the Krantz Center for Cancer Research, an associate professor of Medicine at Harvard Medical School and an Associate Member at the Broad Institute, is co-senior author of the study.

Summary:

Mutations in Isocitrate Dehydrogenase 1 (IDH1) are present in a range of different cancer types, including difficult-to-treat liver and brain tumors. Mutant IDH1 (mIDH1) generates the “oncometabolite” 2-hydroxyglutarate (2HG), which interferes with DNA and histone demethylating enzymes, although specifically how 2HG causes tumor growth remains under investigation.

Moreover, while inhibitors of mIDH1 slow the growth of these tumor types, the specific mechanism underlying therapeutic efficacy has been elusive.

We report a new pathway underlying the anti-cancer mechanism of mIDH1 inhibitors, involving the process of “viral mimicry.”

Transposable elements (TEs) are relics of ancient viral infections scattered throughout mammalian genomes. TE reactivation and detection by cytosolic nucleic acid sensors acts as a cellular alarm, eliciting antiviral immunity (viral mimicry).

We show that the double-stranded (ds) DNA sensor, cGAS, is silenced by promoter DNA hypermethylation in mIDH1 liver and brain tumors.

Conversely, mIDH1 inhibition causes DNA hypomethylation and transcriptional activation of cGAS and of TEs encoding reverse transcriptase, the latter generating cytoplasmic dsDNA. Resulting cGAS signaling stimulates potent anti-tumor T-cell response.

Background

This project started years ago as an investigation into the mechanisms behind how mutations in the metabolic enzyme isocitrate dehydrogenase 1 (mutIDH1) drive tumorigenesis.

Mutations in this enzyme occur in many cancers but are prevalent in certain cancers such as gliomas, cholangiocarcinomas (a liver cancer of the bile duct), and myeloid leukemias.

The mutation causes IDH1, which normally converts isocitrate to a-ketoglutarate, to generate a different product called 2-hydroxyglutarate (2-HG).

This abnormal metabolite is a competitive inhibitor of several epigenetic enzymes that normally depend on a-ketoglutarate as a co-factor and are involved in demethylating DNA to promote gene expression.

High levels of 2-HG that build up in mutIDH1 tumor cells inhibit these enzymes and cause global DNA methylation and silencing.

There are several proposed mechanisms by which this process drives cancer, but the exact mechanisms in highly aggressive solid tumors such as cholangiocarcinoma and glioma are unknown.

Despite this lack of clear mechanistic insight, inhibitors of the mutant form of IDH1 have been approved for the treatment of cholangiocarcinoma and will likely soon be approved for glioma.

The drugs slow tumor growth but are not curative, suggesting that we need to know how they work so we can come up with combination approaches.

Years ago, the Bardeesy Lab made the first genetically engineered mouse model of IDH1 mutant cholangiocarcinoma and discovered that inhibition of mutIDH1 was effective in the models, but intriguingly found that they were only effective in mice with intact immune function.

The drugs didn’t work in immunodeficient mice, arguing a role for anti-tumor immunity.

This was when the Bardeesy and Manguso labs began working together to uncover the mechanism behind how the drugs work.

We have a previous paper on the topic, where we show that one of the mechanisms involve restoring tumor cell sensitivity to the T cell-derived cytokine interferon-gamma, which the mutIDH1 tumors lose responsiveness to because of DNA hypermethylation (Wu et al, Cancer Discovery 2022).

However, we had evidence that this didn’t explain the entire mechanism and have kept working on it since.

What’s New in this Study

In our new manuscript, recently accepted at Science, we’re revealing a completely unexpected and fascinating mechanism that explains the efficacy of the drug.

We have found that inhibition of mutIDH1 causes the re-expression of endogenous retroviruses and other transposable elements (TEs) in tumor cells. These relics of ancient viral infection are normally silenced, but the restoration of DNA de-methylation after inhibition of mutIDH1 causes their expression.

Some of these TEs contain intact viral genes and encode for viral proteins such as reverse transcriptase. We see that reverse transcriptase becomes upregulated in tumor cells and causes the accumulation of cytosolic double-stranded DNA which is then sensed by the cytosolic DNA sensor cGAS to activate a viral mimicry response in tumor cells, which then act like virally infected cells and produce cytokines called interferons to activate the immune system.

What is striking is that we also find that mutations in IDH1 cause the silencing of the cGAS gene in all solid tumors with the mutation, suggesting that the downregulation of this pathway to avoid immune activation is a hallmark of IDH1 mutant tumors.

We show that tumors lacking either the cGAS pathway or the endogenous retroviral reverse transcriptase activity have no response to mutant IDH1 inhibition.

Clinical Implications and Next Steps

It is the first example that we are aware of where the action of an FDA-approved medicine is dependent on the activity of factors like reverse transcriptase encoded by transposable elements in our genome.

While there are studies suggesting that the activity of transposable elements may contribute to causing cancer by inserting new copies of themselves throughout the genome and causing mutations, there are no studies suggesting that the efficacy of anti-cancer therapeutics would depend on the activity of these elements.

Developing strategies that potentiate viral mimicry and/or harness its immunostimulatory effects could increase the effectiveness of mIDH1 inhibitors.

Paper Cited:

Wu, M. J., Kondo, H., Kammula, A. V., Shi, L., Xiao, Y., Dhiab, S., Xu, Q., Slater, C. J., Avila, O. I., Merritt, J., Kato, H., Kattel, P., Sussman, J., Gritti, I., Eccleston, J., Sun, Y., Cho, H. M., Olander, K., Katsuda, T., Shi, D. D., … Bardeesy, N. (2024). Mutant IDH1 inhibition induces dsDNA sensing to activate tumor immunity. Science (New York, N.Y.), 385(6705), eadl6173. https://doi.org/10.1126/science.adl6173