LOS ANGELES – For decades, the drugs designed to fight small cell neuroendocrine cancers of the lung, prostate, and ovary have worked for a while and then stopped working. The tumors outrun chemotherapy. They develop resistance to immunotherapy. For patients diagnosed with these cancers, survival figures have barely moved in four decades. A study published this week in the Proceedings of the National Academy of Sciences suggests that the answer to this long failure may already be sitting on a pharmacy shelf, prescribed for something else entirely.
Leflunomide, which rheumatologists prescribe for rheumatoid arthritis, and teriflunomide, which neurologists use to slow multiple sclerosis, are both classified as autoimmune disease drugs. Neither was developed with cancer in mind. A team at the University of California, Los Angeles now reports that both drugs can halt the growth of small cell neuroendocrine cancer cells in the laboratory, exploiting a vulnerability that a genome-wide CRISPR screen detected after years of searching.
The vulnerability sits in a chain of molecular events that begins with the loss of a gene called RB. Small cell neuroendocrine cancers are defined, almost universally, by the deletion of RB, a tumor suppressor that ordinarily acts as a brake on cell division. When RB is gone, researchers had long assumed the cells simply proliferated without restraint. What the UCLA team found, using a screen that systematically disabled thousands of genes one by one across the full genome of cancer models to observe what happened, was more specific than that.
When RB is lost, the cancer cell shifts its survival dependence onto a different protein, E2F3. This is the phenomenon scientists call synthetic lethality: the cancer tolerates losing one gene, but it cannot survive losing both RB and E2F3. The effect, when E2F3 was disrupted in RB-deficient laboratory models, was a dramatic slowdown in tumor growth. Dr. Owen N. Witte, the study’s senior author and Presidential Chair in Developmental Immunology at UCLA, told UCLA Health that “losing one gene may not matter much, but losing both has a dramatic effect on tumor growth.”
E2F3 is not directly targetable with any approved drug. That is where a different enzyme enters the picture. The UCLA team found that DHODH, an enzyme responsible for producing pyrimidines, the molecular building blocks cells use to replicate DNA, controls the expression of E2F3. Inhibit DHODH, and E2F3 levels fall; when E2F3 falls, RB-deficient tumor cells lose the molecular support they depend on and their growth slows or stops. Blocking pyrimidine synthesis is how leflunomide and teriflunomide already work in their approved autoimmune indications, suppressing the runaway proliferation of immune cells that drives arthritis and multiple sclerosis.
The same molecular lever that makes those drugs effective against overactive immune cells applies, the UCLA team found, to the overactive cancer cells sharing the same RB deficiency. Leflunomide reached FDA approval for rheumatoid arthritis in 1998. Teriflunomide was cleared for relapsing forms of multiple sclerosis in 2012. Their toxicity profiles are well characterized over decades of clinical use, and their manufacturing infrastructure is already in place at commercial scale. That prior regulatory history is the study’s most concrete clinical implication.
“What’s exciting is that our findings open the door to applying existing drugs in a new way,” said Dr. Evan Abt, the study’s first author and an assistant professor of molecular and medical pharmacology at the David Geffen School of Medicine at UCLA. The research reflects more than a decade of work from the Witte Laboratory at UCLA’s Jonsson Comprehensive Cancer Center, building on earlier investigations into how prostate cancer transforms into its most lethal neuroendocrine form.
The practical logic of drug repurposing is well understood in oncology but rarely arrives with the mechanistic specificity the UCLA study provides. Developing a new cancer drug from early discovery to FDA approval typically takes more than a decade and costs several hundred million dollars, with clinical failure rates the pharmaceutical industry consistently places above 90 percent. A compound with an existing safety record and manufacturing infrastructure skips the longest and most expensive portion of that timeline. The question now is whether the laboratory efficacy translates to human tumors with sufficient consistency to justify a clinical trial design. Earlier this month, GLP-1 drugs approved for diabetes were found to cut overdose deaths by half in a study of 600,000 U.S. veterans, another instance of FDA-cleared compounds disclosing effects their original trials were never designed to detect.
The cancers this study addresses are among the most difficult in oncology. Small cell lung cancer accounts for roughly 15 percent of all lung cancer diagnoses and is distinguished by rapid growth and early spread to other organs; most patients are diagnosed at a stage when surgery is no longer an option. Small cell neuroendocrine prostate cancer typically emerges after standard hormone-based prostate cancer treatments have been applied, as though the tumor has reorganized itself to escape the therapy. Small cell carcinoma of the ovary strikes younger women and carries survival figures measured in months after diagnosis. All three share the near-universal loss of RB. All three have resisted every treatment approach that has succeeded against other cancer types, including the immunotherapy regimens that transformed outcomes for melanoma, lung adenocarcinoma, and certain bladder cancers.
The path from the PNAS result to an oncologist prescribing leflunomide for a patient with small cell lung cancer is not yet mapped. The UCLA experiments were conducted in laboratory cell lines and preclinical models; the study reports no clinical trial plans. Whether DHODH inhibition produces the same effect in the complex biology of an actual tumor, where immune cells, stromal tissue, and blood supply interact in ways no cell culture fully replicates, is the question the published data does not answer. The same CRISPR editing technology that made this vulnerability map possible is moving from research to clinical practice on other fronts: this month, as the FDA expanded approval of the CRISPR therapy Casgevy to children as young as two with sickle cell disease, the distance between what CRISPR can find in a laboratory and what it can do for a patient has been narrowing steadily.
According to UCLA Health’s announcement of the findings, the discovery points toward a class of cancers that has resisted treatment not because they are poorly understood biologically but because no one had found a dependency they could not escape. The RB deletion that defines small cell neuroendocrine cancers is so consistent across tumors of the lung, prostate, and ovary that the E2F3 dependency it creates is likely to hold across patients in ways that more variable tumor mutations do not. Whether leflunomide or teriflunomide reaches a patient with one of these cancers depends on decisions the PNAS paper does not make. The molecular argument for trying is now on record.
