Protecting Pyrazinamide (PZA) and fluoroquinolones (FLQ) for successful multi drug resistant tuberculosis (MDR-TB) treatment

Using antibiotics selects resistance: although the consequence and the time before this resistance becomes so common it undermines the antibiotics effectiveness for routine use in the community varies enormously. Even when correctly managed, following international and national guidelines, due to the complexity of (tuberculosis) TB diagnostics a proportion of patients initially receive inappropriate first line therapy. In fact, inappropriate therapy may be given for a considerable period as there is often a delay of weeks to months before treatment failure or drug resistance is detected by bacterial culture, as a result of the slow growth rate of mycobacteria. In high (multi-drug resisitant Mycobacterium tuberculosis) MDR-TB settings the proportion of initially under treated patients can be significant. This means that these patients receive fewer active drugs than needed to prevent the emergence of resistance during the initial phase of their treatment. The precise contribution this under treatment makes to the generation of resistance is unknown. In this project we will take advantage of the clonal nature of tuberculosis (Eldholm and Balloux 2016) to study the development of resistance in tuberculosis patients and assess the contribution of inappropriate treatment. We will follow initially undetected MDR-TB patients who receive a period of suboptimal multi-drug treatment and compare the micro-evolution of their infecting bacteria with optimally treated patients using deep sequencing. We will pay particular attention to fluoroquinolone (FLQ) and the development of any pyrazinamide (PZA) resistance in early therapy, as recent information suggests FLQ (Streicher 2012) and PZA (Operario 2017) resistance acquisition is a common event. We have also recently proposed that PZA resistance is likely a genetic bottleneck for the bacteria (den Hertog 2015) which has significant implications for the design of future drug regimens (Anthony 2017, Fofana 2016). We will study a cohort from high two MDR-TB settings Moldova and Belarus (Jenkins 2013, Skrahina 2012). In these high MDR-TB burden settings recruiting patients that start directly on appropriate therapy and a matched group that initially receive some inactive drugs. Paired MTB cultures obtained at the time of diagnosis and at the moment MDR-TB infection is detected in inappropriately treated patients switching to second line therapy or adapted second/third line therapy will be subjected to deep sequencing to detect any accumulation of mutations, including minority variants. In parallel we will also use deep sequencing to monitor bacterial microevolution in a series of Dutch patients who receive individualised therapy. In this project clinical data will also be collected using a mobile phone android application developed for the project backed up by onsite review of the patient notes. In addition, we will collect blood spots to allow the measurement of circulating drug levels and confirm drug dosing as well as assess any influence of (sub optimal) dosing on the accumulation of mutations. Routine culture and molecular diagnostic will be performed at the study sites as part of the routine standard of care. Blood spots and DNA extracts will be shipped to the Netherlands for detailed analysis. WGS of isolates from inappropriately treated patients will be compared to isolates from patients directly receiving appropriate therapy. This study will not alter the treatment of the patients in any way. In the Netherlands we will apply next generation deep sequencing to DNA extracted from all cultures and screen known drug resistance associated mutational hotspots for mutations. The cohort of 10 -20 patients treated with fully individualized regimens will be recruited in the Netherlands. The use of next generation sequencing will allow us to detect genetic diversity in the patients and the emergence of resistance associated mutations that are present in only a small proportion of the bacterial population (<5%). The frequency of mutations and emerging genetic diversity associated with drug resistance will be compared in appropriately treated patients and inappropriately treated patients before and after a period of optimal or sub optimal multi-antibiotic therapy. Genetic data will be linked to detailed prescribing and serum drug level data from individual patients to understand how to better use available drugs. Mathermatical modelling will be perfromed to assess the potential impact and cost-effectivness of individualising therapy using WGS and theraputic drug monitoring. This study will use the newly developed deep sequencing methods in a systematic study with detained treatment information rather than a convenience sample which will provide results that can be used to directly inform and influence policy discussions at the national and world wide level.