Determining the persister populations in sputum during tuberculosis therapy. A supplementary study to the RIFASHORT trial.
Lead Research Organisation:
University of Leicester
Department Name: Infection Immunity and Inflammation
Abstract
Nearly ten million people developed TB in 2014 and 1.5 million died. Although drug resistance is a serious problem, the vast majority of infections are treatable with standard antibiotics effective against the TB bacterium, known as Mtb, which causes this devastating disease.
Treatment of TB currently requires patients to take antibiotics for 6 months. Even after this the disease returns in up to 5% of individuals, usually within a year after stopping therapy. Developing treatments with shorter duration and lower disease return (or relapse) rates is a major goal of global TB control efforts. Testing new treatments is very expensive, requires hundreds of patients and takes many years. Our project aims to reduce these barriers to testing and enable shortened treatment.
The need for prolonged therapy and the frequency of relapse are thought to be due to significant numbers of Mtb bacteria in an altered state that survive exposure to antibiotics; the altered bacteria are known as persisters. While growing cells of Mtb are rapidly killed in conventional lab tests, cells in the non-growing persister state are not.
Our team has been studying the persister phenomenon in TB for many years. In the past decade we discovered features of Mtb in the sputum (coughed up phlegm) of newly diagnosed patients that, surprisingly, indicated the presence of non-growing persisters amongst the bacteria that clinical labs normally isolate. These features are recognised in 3 ways, microscopy to detect fatty droplets inside the Mtb cells, lab culture in the presence of a growth stimulator known as Rpf and a genetic test to show which genes the bug is expressing (known as the transcriptome). Because all three indicate the present of non-growing Mtb in sputum and because the levels present varied from patient to patient we wondered whether these "Fat and Lazy" bacteria were related to the persisters. If this is so, then the numbers of Mtb with these properties in different patients might tell us how quickly they will respond to treatment and, once on therapy, how rapidly the persisters were being eliminated.
In preliminary studies we looked at the frequency of fatty Mtb cells in TB patients and related these to how patients responded to treatment. Although numbers were small, we found that patients with higher numbers of fatty Mtb responded more slowly. Then, with colleagues in Liverpool and Malawi, we asked whether counting fatty Mtb cells in sputum samples could tell us which patients were going to fail TB treatment in spite of being infected with an antibiotic sensitive bug. When we compared the levels of fatty Mtb between 9 treatment failures and 29 cured patients we found that the failures had, on average, double the frequency of these "persister-like" cells.
Here we aim to add value to a clinical trial testing a shortened treatment of TB by studying the ability of our tests to predict treatment failure. The RIFASHORT TB treatment trial provides an exceptional opportunity for us to test sufficient sputum samples for these properties to achieve 3 objectives: 1) re-testing the predictive capacity of counting fatty Mtb cells at one month; 2) determining the interrelationships between the fatty, Rpf and transcriptome features in sputum samples; and 3) determining of which test, applied within the first 2 months of treatment, is most strongly linked to treatment failure.
The results of this study will have clear potential to, 1) enable rapid assessment of new anti-TB treatments including new drugs effective against resistant Mtb, and 2) determine which patients might safely stop treatment after periods as short as 3 months. If we can reliably compare new treatments in 2 rather than 18 months, we will be able to test more alternatives and select the best with far less resource than currently required. If we can shorten TB treatment, many lives will be saved and much suffering alleviated.
Treatment of TB currently requires patients to take antibiotics for 6 months. Even after this the disease returns in up to 5% of individuals, usually within a year after stopping therapy. Developing treatments with shorter duration and lower disease return (or relapse) rates is a major goal of global TB control efforts. Testing new treatments is very expensive, requires hundreds of patients and takes many years. Our project aims to reduce these barriers to testing and enable shortened treatment.
The need for prolonged therapy and the frequency of relapse are thought to be due to significant numbers of Mtb bacteria in an altered state that survive exposure to antibiotics; the altered bacteria are known as persisters. While growing cells of Mtb are rapidly killed in conventional lab tests, cells in the non-growing persister state are not.
Our team has been studying the persister phenomenon in TB for many years. In the past decade we discovered features of Mtb in the sputum (coughed up phlegm) of newly diagnosed patients that, surprisingly, indicated the presence of non-growing persisters amongst the bacteria that clinical labs normally isolate. These features are recognised in 3 ways, microscopy to detect fatty droplets inside the Mtb cells, lab culture in the presence of a growth stimulator known as Rpf and a genetic test to show which genes the bug is expressing (known as the transcriptome). Because all three indicate the present of non-growing Mtb in sputum and because the levels present varied from patient to patient we wondered whether these "Fat and Lazy" bacteria were related to the persisters. If this is so, then the numbers of Mtb with these properties in different patients might tell us how quickly they will respond to treatment and, once on therapy, how rapidly the persisters were being eliminated.
In preliminary studies we looked at the frequency of fatty Mtb cells in TB patients and related these to how patients responded to treatment. Although numbers were small, we found that patients with higher numbers of fatty Mtb responded more slowly. Then, with colleagues in Liverpool and Malawi, we asked whether counting fatty Mtb cells in sputum samples could tell us which patients were going to fail TB treatment in spite of being infected with an antibiotic sensitive bug. When we compared the levels of fatty Mtb between 9 treatment failures and 29 cured patients we found that the failures had, on average, double the frequency of these "persister-like" cells.
Here we aim to add value to a clinical trial testing a shortened treatment of TB by studying the ability of our tests to predict treatment failure. The RIFASHORT TB treatment trial provides an exceptional opportunity for us to test sufficient sputum samples for these properties to achieve 3 objectives: 1) re-testing the predictive capacity of counting fatty Mtb cells at one month; 2) determining the interrelationships between the fatty, Rpf and transcriptome features in sputum samples; and 3) determining of which test, applied within the first 2 months of treatment, is most strongly linked to treatment failure.
The results of this study will have clear potential to, 1) enable rapid assessment of new anti-TB treatments including new drugs effective against resistant Mtb, and 2) determine which patients might safely stop treatment after periods as short as 3 months. If we can reliably compare new treatments in 2 rather than 18 months, we will be able to test more alternatives and select the best with far less resource than currently required. If we can shorten TB treatment, many lives will be saved and much suffering alleviated.
Technical Summary
Current standard therapy for drug-sensitive TB is associated with a relapse rate of ~5%. We urgently need tests that identify potential unsatisfactory treatment outcomes early in therapy with sufficient sensitivity and specificity to guide the implementation of clinical trials and the monitoring of individual patients. In the former case such tests could accelerate development and evaluation of new drugs and treatment regimens, in the latter the duration of therapy for individual patients could be individualised.
Here we build on our studies of mycobacterial lipid bodies (LBs) and Resuscitation promoting factor dependency (Rpf-D), both of which provide established links to antibiotic tolerant persister-like phenotypes and analyses of which have, respectively, been clearly linked to relapse in human and mouse infections. Our aim is to evaluate LB and Rpf assessments in sputum samples for their capacity to predict treatment responses.
By use of samples from and added to the RIFASHORT trial, we will obtain multiple readouts of the M. tuberculosis (Mtb) subpopulations in sputum that our data indicate are potentially responsible for relapse and slow responses to therapy. Critically, our recently published protocol for storage of sputum samples to enable deferred analysis of Rpf-D populations will enable assessment of this strong candidate persister readout on a scale sufficient to evaluate its potential predictive value.
Recent analyses of Mtb transcription in sputum and in vitro provide a further opportunity to link clinical assessment of treatment responses to Mtb physiology and our background understanding of LBs and Rpfs. This study will address the critical need for clinical biomarkers of treatment response in TB while simultaneously adding substantial value to the RIFASHORT trial. Our study will deliver new clinical knowledge in this key area and has significant potential to both accelerate development of new treatments and enhance individual patient management.
Here we build on our studies of mycobacterial lipid bodies (LBs) and Resuscitation promoting factor dependency (Rpf-D), both of which provide established links to antibiotic tolerant persister-like phenotypes and analyses of which have, respectively, been clearly linked to relapse in human and mouse infections. Our aim is to evaluate LB and Rpf assessments in sputum samples for their capacity to predict treatment responses.
By use of samples from and added to the RIFASHORT trial, we will obtain multiple readouts of the M. tuberculosis (Mtb) subpopulations in sputum that our data indicate are potentially responsible for relapse and slow responses to therapy. Critically, our recently published protocol for storage of sputum samples to enable deferred analysis of Rpf-D populations will enable assessment of this strong candidate persister readout on a scale sufficient to evaluate its potential predictive value.
Recent analyses of Mtb transcription in sputum and in vitro provide a further opportunity to link clinical assessment of treatment responses to Mtb physiology and our background understanding of LBs and Rpfs. This study will address the critical need for clinical biomarkers of treatment response in TB while simultaneously adding substantial value to the RIFASHORT trial. Our study will deliver new clinical knowledge in this key area and has significant potential to both accelerate development of new treatments and enhance individual patient management.
Planned Impact
We aim to deliver a step change in our ability to monitor the effectiveness of TB treatment and predict its outcome. Our work has great potential to accelerate the evaluation of new drugs and new treatment regimens. If we obtain further evidence that sputum analyses in the first 2 months of therapy predict the likelihood of cure in populations and individuals then the outcome will be transformative. This will have most impact on therapeutic trial design where alternate regimens could be rapidly compared and the most effective taken forward but also on individual case management where selected subpopulation analyses could be used determine the appropriate duration of therapy.
Our work will also produce new understanding of the interplay between Mtb subpopulations in human infection, their responses to drug exposure and the host environment.
WHO WILL BENEFIT AND HOW?
ACADEMICS, PRACTITIONERS AND AGENCIES DEVELOPING AND CONDUCTING CLINICAL TRIALS OF TB TREATMENT AND DEVELOPING NEW ANTIMICROBIALS (GOVERNMENT AND CHARITY RESEARCH AGENCIES AND PHARMACEUTICAL INDUSTRIES).
This work will deliver new insights into changes in Mtb populations in the first 2 months of therapy and the degree to which these map to the outcome of treatment. The study is designed to recognise the predictive capacity of separate and combined sputum analyses with respect to disease relapse/treatment failure separate from genetically determined bacterial antimicrobial resistance. Understanding which analyses are most discriminatory in this regard will allow their deployment in the future design of trials and to focus effort on further development of the selected analytical methods to make them more amenable for clinical deployment.
The ability to predict the outcome of therapy with defined levels of confidence within 2 months of commencement will enable a radical enhancement of trial design so that more options can be rapidly evaluated. While safety and governance issues will require extensive further evaluation, establishing clear potential here will incentivise rapid further development and evaluation.
Organisations with interests in the development and evaluation of new anti-TB therapies will benefit directly from our research outcomes. These include pharmaceutical companies in the UK, such as GSK and Astra Zeneca.
DEPARTMENT OF HEALTH; HEALTH PROTECTION ENGLAND; THE WORLD HEALTH ORGANISATION; GOVERNMENT; THE HEALTHCARE INDUSTRY. Primary beneficiaries will be organisations responsible for policies for the management of TB and the evaluation of new drug treatments. The long term goal is to shorten TB therapy and to facilitate evaluation of individuals' responses to treatment.
MEDICAL COMMUNITY AND PATIENTS AND THOSE AT RISK OF EXPOSURE TO MTB. Practitioners concerned with the treatment of TB will benefit from the tests evaluated and the understanding gained in this project. Patients will potentially benefit by shorter treatment regimens and by the possibility of individualising treatment duration in response to selected biomarker results.
Our work will also produce new understanding of the interplay between Mtb subpopulations in human infection, their responses to drug exposure and the host environment.
WHO WILL BENEFIT AND HOW?
ACADEMICS, PRACTITIONERS AND AGENCIES DEVELOPING AND CONDUCTING CLINICAL TRIALS OF TB TREATMENT AND DEVELOPING NEW ANTIMICROBIALS (GOVERNMENT AND CHARITY RESEARCH AGENCIES AND PHARMACEUTICAL INDUSTRIES).
This work will deliver new insights into changes in Mtb populations in the first 2 months of therapy and the degree to which these map to the outcome of treatment. The study is designed to recognise the predictive capacity of separate and combined sputum analyses with respect to disease relapse/treatment failure separate from genetically determined bacterial antimicrobial resistance. Understanding which analyses are most discriminatory in this regard will allow their deployment in the future design of trials and to focus effort on further development of the selected analytical methods to make them more amenable for clinical deployment.
The ability to predict the outcome of therapy with defined levels of confidence within 2 months of commencement will enable a radical enhancement of trial design so that more options can be rapidly evaluated. While safety and governance issues will require extensive further evaluation, establishing clear potential here will incentivise rapid further development and evaluation.
Organisations with interests in the development and evaluation of new anti-TB therapies will benefit directly from our research outcomes. These include pharmaceutical companies in the UK, such as GSK and Astra Zeneca.
DEPARTMENT OF HEALTH; HEALTH PROTECTION ENGLAND; THE WORLD HEALTH ORGANISATION; GOVERNMENT; THE HEALTHCARE INDUSTRY. Primary beneficiaries will be organisations responsible for policies for the management of TB and the evaluation of new drug treatments. The long term goal is to shorten TB therapy and to facilitate evaluation of individuals' responses to treatment.
MEDICAL COMMUNITY AND PATIENTS AND THOSE AT RISK OF EXPOSURE TO MTB. Practitioners concerned with the treatment of TB will benefit from the tests evaluated and the understanding gained in this project. Patients will potentially benefit by shorter treatment regimens and by the possibility of individualising treatment duration in response to selected biomarker results.
