Another Step Closer to Genomic-informed Therapeutic Dosing for Tuberculosis Drugs

Most antibiotics currently used to treat Mycobacterium tuberculosis (Mtb) were developed and approved more than 50 years ago. As such, many anti-tuberculosis (anti-TB) drugs have nonspecific mechanisms of action and a long list of potential adverse drug reactions (1). Therefore, treatment guidelines recommend weight-based dosing for several first-line anti-TB medications, including isoniazid and rifampin, to achieve optimal drug concentrations within narrow therapeutic windows (2). If drug concentrations become supratherapeutic, then patients are subjected to unnecessary toxicities and complications, including drug-induced hepatotoxicity (3, 4). Conversely, if drug concentrations remain subtherapeutic, then Mtb may continue replicating and/or developing drug-resistance mechanisms, both of which can result in treatment failure or worse (5, 6).

In 2001, the World Health Organization and the International Union Against Tuberculosis and Lung Disease recommended fixed-dose combination pills to simplify treatment regimens, reduce pill burden, and increase medication adherence for the treatment of TB disease (7). The implementation of fixed-dose combination pills improved the procurement and dispensing of TB medications but also minimized a clinician’s ability to personalize dose adjustments or exclude individual drugs based on tolerability and/or toxicity. The adoption of fixed-dose combination pills may have also shifted clinical attention away from measuring and understanding variations of drug metabolism and therapeutic drug concentrations among patients receiving TB therapy in TB-endemic settings.

Drug metabolism can be affected by several pathways, and the host genome may contain several polymorphisms that alter proteins within these metabolic pathways. For isoniazid, a pathway for activation to toxic intermediates was known in the 1970s (8), and the impact of varying acetylation rates on drug metabolism was described many decades ago (9, 10). In addition, genomic polymorphisms that impact the clearance for other prominent TB drugs, including rifampin, rifapentine, linezolid, and bedaquiline, have also been described (11–13). Thus far, only one published randomized clinical trial has shown that pharmacogenetics-based dosing of isoniazid, based on the NAT2 (N-acetyltransferase 2) genotype for acetylation activity, can reduce drug toxicities and early treatment failure for the treatment of pulmonary TB disease (14).

Despite knowing the genomic mutations that most significantly impact TB drug metabolism, pharmacogenomic testing for personalized TB drug dosing has not been endorsed for routine clinical practice. Part of this reluctance has been the lack of a practical, scalable, and cost-effective assay for genomic testing in TB-endemic settings. However, in recent years, rapid nucleic acid sequencing devices and tools to evaluate the presence of host genomic alleles have been developed. If such a pharmacogenetic testing approach could be highly accurate and affordable, then it could be implemented to achieve better, more personalized treatment for people initiating TB therapy.

In this issue of the Journal, Verma and colleagues (pp. 1486–1496) developed a customized sequencing panel on a rapid nanopore assay to detect key mutations in five human genes that are known to affect the metabolism of critical TB drugs (15). In this study, the team first validated their molecular sequence panel on 48 specimens from the 1000 Genomes Project. The experimental assay achieved 100% concordance when compared with whole-genome sequencing data generated from an Illumina platform. Thus, the team demonstrated the accuracy of their custom nanopore assay before testing blood specimens from individuals with TB disease.

Second, the team evaluated their rapid genomic sequence panel using whole blood specimens from 100 adults who were being treated for TB disease in South Africa. The team determined an individual’s acetylator status by interrogating the NAT2 gene, among other polymorphisms. Within this clinical cohort, the sequencing depth across eight amplicons was robust (having achieved >100-fold coverage for 99.8% of target amplicons), and participants were classified into slow, intermediate, or rapid drug-acetylator phenotypes. Then, the study team compared pharmacogenetic results with drug concentrations of isoniazid and rifampin, which were obtained several weeks after treatment initiation, in stored whole blood specimens. The primary results showed that isoniazid clearance was strongly associated with an individual’s acetylator status, as determined using their genomic sequencing panel. Similarly, rifampin clearance was also significantly different between people with and without a homozygous AASACrs1803155 allelic substitution. Overall, these findings are supportive of prior pharmacogenetic studies and provide much added value for conducting the genomic testing on a rapid nanopore assay.

The study provides important additional support to the growing literature that pharmacokinetic measurement of TB drugs, most importantly isoniazid, or screening for genomic alleles that influence drug metabolism may be valuable for clinical care and management of TB disease. However, additional studies are still needed to inform guidelines and future directions. First, epidemiologic studies may be helpful to enumerate the penetrance of the human genomic alleles that impact TB drug metabolism among key populations and diverse TB-endemic settings. Second, additional data from larger randomized controlled trials using pharmacogenetics-guided therapy will be needed to further quantify any reductions in adverse events and gains for improved treatment outcomes in various settings. Finally, cost-effectiveness analyses, particularly with new, clinic-based next-generation sequencing devices such as the robust, accurate nanopore instrument, can quantify the value gained with pharmacogenetic testing and personalized medication dosing.

Although this study was limited by conducting cross-sectional testing on stored blood specimens, the findings do suggest that more prospective studies, as well as randomized controlled trials, are warranted to better quantify the clinical outcomes and cost-effectiveness of pharmacogenetic testing at TB treatment initiation. Eventually, the first generation of anti-TB drugs, namely isoniazid and rifampin, should be replaced by anti-TB drugs that will be more potent for killing Mtb and less toxic to the human body. Fortunately, there is a robust pipeline of new drug candidates being evaluated in human studies. However, because the development and adoption of novel drugs has historically been slow, we should continue to optimize the delivery and dosing of the anti-TB drugs that are currently used around the world.