Are IDH Mutations a Predictor of Venetoclax Response?

Data from a small clinical trial of venetoclax in acute myeloid leukemia (AML) supported an FDA breakthrough therapy designation for use in combination with hypomethylating agents in treatment-naïve patients who are ineligible for high-dose induction chemotherapy (PMID: 31628431).  Approval of BCL-2 inhibitors raises the question of the mechanism of action in AML and patient selection for treatment. Unfortunately, no biomarkers or methods exist to predict venetoclax response in AML, making treatment selection challenging. However, 17p deletion is one of the key biomarkers suggested by NCCN for the selection of venetoclax in CLL patients. 

The Clinical Case

Nearly 20-25% of the AML population has IDH1/2 mutations, presenting a potential treatment option apart from chemotherapy. The first FDA IDH2 inhibitor, Enasidenib, was approved in 2017. Additionally, IDH1 inhibitors are currently undergoing clinical studies and the oral BCL-2 inhibitor venetoclax has shown efficacy in this population subset. Venetoclax (ABT-199, GDC-0199, Venclexta®; AbbVie Inc., North Chicago, IL, USA) is a highly selective oral agent that binds to the BH-3 binding domain of BCL-2, preventing BCL-2-mediated inactivation of pro-apoptotic proteins. 

In a phase 2 study of venetoclax 800 mg daily in patients with relapsed/refractory AML or those unable to receive intensive chemotherapy, IDH1/2 mutations were detected in 12 patients (PMID: 29882807). Additionally, a reported phase 1b dose escalation study of venetoclax in combination with 5-Aza or DAC enrolled 17 patients with IDH1/2 mutations. Among these patients, 10 (59%) demonstrated CR/CRi and three (18%) demonstrated a morphologic leukemia-free state.

A review article by Buege, MJ et al, 2018 (PMID: 29882807) focused on the clinical efficacy of venetoclax and identified the clinical benefits in the IDH mutant subgroup. However, the review did not include the scientific rationale or the molecular pathway involved. 

Cellworks Insight

We applied patient-specific genomics and computational biology methods to biosimulate a patient’s cancer and then determine the patient’s response to existing drugs. The Cellworks computational biology model (CBM) utilized in this study was previously outlined and published in studies of glioblastoma multiforme, multiple myeloma, ETP-ALL, CLL, MPN and MDS. 

The CBM is based on utilizing more than 25,000 published PubMed references and in-house experimental data. This model considers signaling pathway interactions important in cancer including growth factor signaling cascades, cytokines, chemokines, mTOR regulators, cell cycle regulators, oxidative and ER stress responses, cancer metabolism, autophagy and proteasomal degradation, DNA damage repair, apoptosis cascades and p53 signaling to predict a patient’s response to a single drug or a combination of drugs. This modeling system includes thousands of intracellular pathway elements capable of simulating functional interactions, including comprehensive coverage of the kinome, transcriptome, proteome, and metabolome.

The Cellworks network-based analysis is able to give more insight into the IDH1/2 mutation-based altered pathway and the impact of these mutations on various signaling pathways, metabolic pathways, epigenetics and complex machinery. Cellworks predictive model also helped to identify the biomarkers responsible for drug efficacy. We used this model and evaluated venetoclax efficacy on IDH mutated patient profiles and identified key molecular pathways involved in this process.

Scientific Rationale:

    
IDH mutation has been reported in Acute Myeloid Leukemias, Chondrosarcomas, MDS, AML, and Gliomas. Mutation of IDH1 leads to the abnormal accumulation of oncometabolite 2 -hydroxy glutarate, and loss of functional metabolite alpha-ketoglutarate, leading to a global reduction of hydroxylation processes in the cell. HIF1a is a prominent target of hydroxylation-mediated degradation. Reduced alpha keto-glutarate leads to the stabilization of HIF1a and an increase in its downstream transcriptional targeting of BCL-2 among other survival markers. Increased activity of the HIF1a transcription factor leads to an increased BCL2 expression and enhanced sensitivity to Venetoclax. 

We modeled and simulated two patients harboring IDH mutation. Our biosimulation predicted a response to venetoclax, which agrees with the clinical results. In patient 1820, PIK3CA-GOF, IBTK- LOF and PRKAA1 were also identified, a positive indicator of venetoclax response and IDH1 mutation. Similarly, in patient 2188, SUZ12- GOF, PDGFRA- GOF and ITGB1- GOF were identified along with IDH1 mutation contributing to venetoclax efficacy. Therefore, the IDH mutation status becomes a positive indicator of venetoclax response in patients.