Fedratinib Does Not Inhibit Thiamine Uptake or Induce Experimental Wernicke's Encephalopathy in Nonclinical Studies

Introduction:

Despite compelling efficacy data from Phase II and Phase III trials, clinical development of fedratinib was halted in 2013 after 8 of 670 treated patients presented with neurological symptoms suggestive of Wernicke's encephalopathy (WE). WE is a neurological disorder that develops due to thiamine deficiency that is treated or prevented simply by using thiamine supplementation. Autopsy studies indicate WE has a prevalence between 0.8-2.8% in the general population with a higher rate in MPN and cancer patients. Confirming a diagnosis of WE is complicated by comorbidities in patients with myeloproliferative neoplasms (MPN) including a high incidence of stroke, hepatic encephalopathy secondary to hepatic extramedullary hematopoiesis or sepsis related to marrow failure in an elderly population. Analysis of patients in fedratinib trials by a panel of neurology and neuroradiology experts indicates that between 1-5 patients of 670 treated experienced WE, less than what might be expected based on historical norms.

WE is commonly caused by either poor nutritional status or chronic alcohol use. The most common non-hematological adverse events for fedratinib are nausea, vomiting and diarrhea, suggesting they may be exacerbating factors for malnutrition in patients. Nonetheless, inhibition of thiamine transporter (THTR) has been proposed as a putative mechanism based on Caco-2 studies. Interpretation of data from these studies is complicated based on use of protein-free culture medias that don't replicate protein binding of drug in patients (96% protein bound) and for which the drug is sparingly soluble. No CNS lesions that would be predictive of WE were reported in the IND-enabling toxicology studies with fedratinib, but subtle neurological signs or histology signs may not be clear to non-experts.

To evaluate the potential of fedratinib to inhibit thiamine uptake, thiamine transporter studies were run using previously published methods with Caco-2 cells that overexpress thiamine transporters with the significant exception that they were performed using human FBS rather than cell culture media. This addresses concerns about protein binding, solubility and replicates conditions experienced by patients. Further, to evaluate the potential of fedratinib to induce WE non-clinically, rats were dosed orally with 40 mg/kg fedratinib (140% of 500 mg patient dose by allometric scaling) daily for 28 days with careful monitoring of neurological signs of disease followed by evaluation of brains for any signs of lesions consistent with WE.

Results:

Fedratinib had no effect on THTR-1 at concentrations up to 30 µM (6-12X higher than Cmax at highest dose of drug). Fedratinib IC₅₀ against THTR-2 was >30 µM. This indicates that fedratinib does not inhibit either thiamine transporter at plasma concentration achievable in patients. This is consistent with patient data in which plasma thiamine levels were normal for the 161 patients sampled at end of treatment in the fedratinib clinical studies.

In rats dosed orally with 40 mg/kg fedratinib for 28 days, there was no observed ataxia, opisthotonus, nystagmus, loss of righting reflex or any other overt neurological signs of thiamine deficiency. Examination of the brains of these animals found no lesions consistent with focal damage to vulnerable areas of brain, e.g. thalamus, inferior colliculus or brainstem. There was no difference in weight gain in these animals compared to control animals. This data indicates that at doses comparable or higher than the highest fedratinib dose used in phase III studies, there is no nonclinical evidence that fedratinib induces thiamine deficient disease.

Conclusion:

Transporter studies performed in the presence of human serum indicate that fedratinib does not inhibit thiamine transport at clinically achievable concentrations. This is consistent with plasma thiamine data available from patients. Rodent studies indicate that fedratinib does not induce thiamine deficient disease when administered orally at clinically relevant concentrations. Notably, rats are incapable of emesis, so this model doesn't account for effects of gastric distress. In total, these data suggest that fedratinib does not increase the risk of thiamine deficiency, beyond its potential to exacerbate malnutrition through poor management of preventable GI adverse events.