Childhood tuberculosis: Integrating tools for improved diagnosis and vaccines

Tuberculosis still causes significant disability and death in many countries in the world, including half a million deaths in children each year. The diagnosis of TB is more difficult in children, as often it cannot be confirmed by "gold standard" methodology, which means showing presence of the TB bacilli in a patient, usually in their sputum. This is rarely possible in children, since they have fewer bacteria, often don't cough up sputum and still get very sick. They often don't receive treatment in time or not at all. Apart from looking for the TB bacilli, we hence use other more indirect tools to make a diagnosis, including observations of TB cases in the community and evidence that the host's immune system has been exposed to TB (patient) by doing a skin test. There are new technologies which can now be applied to blood samples of patients to see if their immune system expresses a characteristic "signature" of TB. However, these have not been used in children yet.
TB often occurs in households, which is where children are particularly likely to catch it from an infected adult. But not everyone gets it when exposed to a coughing "index case", which is the coughing person with active TB. It is likely that many factors like the organism in question, the conditions of the household and also the responses of the immune system of the individual child play a role. We have a large gap in our understanding of the transmission, susceptibility and what kind of immunity is needed to fend off the TB bacteria. This is why we have so far failed to make a better vaccine than the current vaccine, BCG, which is not fully protective. If we understand the mechanisms that underlie susceptibility and protection, this would be very useful for any trials of new vaccines, but also for the diagnosis, as we could look for specific markers associated with TB disease and protection. Such markers could also be used in vaccine or treatment trials, instead of purely relying on the bacteriologically confirmed cases, which is the current approach. This approach is very costly since it needs a very large number of children to participate in vaccine or treatment trials, as most of them will never show up with confirmation by bacteria and yet get sick with what we think is TB.
In our project, we will develop and test novel approaches by using new methods in the microbiology and immunology laboratory and have brought together a strong team of internal and external experts to be able to use all the state-of-the art available methodology.
In particular, we want to evaluate some of the new technologies to diagnose TB on the "signature" it leaves in the body and to use a statistical approach to bring together the individual pieces of the jigsaw that makes up TB diagnosis in children. We will also study children who do and do not get TB after exposure to the bacillus in their household, as the group who does not contract the disease or infection might hold the key to what is really protective against TB. This knowledge will be very important for making an improved vaccine, which can essentially mimic such protective mechanisms. With the help of more advanced statistical analysis of all the factors that contribute to a TB diagnosis we would like to come up with algorithms for the diagnosis that are just or nearly as good as the gold standard. We wish to conduct this project in the Gambia, where there is still a lot of TB. Thanks to the MRC Unit in The Gambia, we benefit from all the tools in the lab and field to collect samples and data for a very comprehensive assessment of every case.
Our long-term goal with this project is to enable the research community to develop a better diagnostic approach to childhood TB and understand how we should design the ideal vaccine against TB.

Tuberculosis causes significant morbidity and mortality in children worldwide. Bacteriological confirmation of TB in children is rare and lack of suitable alternative diagnostics is a major bottleneck to progress in identifying patients in need of treatment and in clinical trials of new vaccines and therapies. The MRC Unit, The Gambia can make a significant contribution in this field, since it has a track record in TB research, established data collection tools and state-of-the art laboratory facilities. Suitable cohorts of children with TB disease, infected and controls can be predicted. Our team of local and international investigators will further develop and evaluate existing and new tools based on both host immune response and microbiology and work with new external collaborators, who are highly experienced in statistical analysis. We will test samples from TB-affected children (exposed, infected or diseased) identified in household cohorts by using a combined approach of host and pathogen signatures and cellular and molecular methods, such as GeneXpert and biosignatures. We will characterise host responses associated with protection against infection in TB-exposed children who remain uninfected and who can serve as a model for protective immunity. We will expand our existing epidemiological database to include the epidemiological and microbiological context of household transmission and its impact on host responses. We will develop a novel statistical approach to design prediction algorithms for the diagnosis of childhood TB. Translation to Policy: Integration of measurable host responses into diagnostic algorithms will inform not only the diagnosis of the individual patient but also facilitate clinical trials of new vaccines and drugs. Understanding of protective immune mechanisms will guide vaccine design. Epidemiological datasets that link TB exposure, infection and disease can serve as a model for public health and TB control.

Tuberculosis remains a global emergency and affects and continues to kill adults and children in large numbers worldwide.
Our project proposal has been developed to address a critical bottleneck for patient management and clinical trials of new vaccines and drugs alike: a) the problem of accurate diagnostics for tuberculosis (TB) in children and b) the absence of markers of protection which can be used as endpoints for TB vaccine trials. This bottleneck is widely acknowledged by the scientific community and international TB program officers, and remains at the center of international research efforts. This was highlighted in the special supplement to the Journal of Infectious Disease, published on World TB Day 2012 and dedicated to basic research but also implementation obstacles for TB globally.
Without a better diagnostic approach, which doesn't exclusively rely on bacteriological confirmation, children are not likely to be included in any studies of improved treatment for TB, and any ongoing vaccine trials will continue to require large and expensive cohorts of patients to be followed prospectively over a long period of time. This is prohibitively expensive and will preclude the evaluation of promising vaccine candidates as there simply are not enough field sites available that can provide both sizeable paediatric cohorts and have the laboratory and clinical capacity to follow them up in clinical trials.
Our research program will impact on areas of discovery, development and delivery within childhood tuberculosis. All of these are required and have merit in driving forward the agenda to improve the diagnosis and treatment of individual cases of TB in children and ultimately also facilitate clinical trials for new treatments and vaccines.
Discovery: the identification of new biomarkers and host/pathogen signatures will drive the development of point-of care tests for TB by identifying new leads in either metabolomics, cytokine or gene signatures, which are amenable to further development with the help of industry partners and not-for profit organisations such as FIND/PATH. To understand mechanisms of protection against infection opens up a key avenue for vaccine design and evaluation, as ultimately, only a TB vaccine that prevents infection and disease will have the desired public health impact for TB control.
Development: we expect that the evaluation of recently identified biosignatures from adult populations will confirm preliminary results also in children and will then stimulate the development of more user-friendly versions of assays, including translation into nanotechnological and lateral flow-based assays, again partnering with industry in the future.
Delivery: The assessment of impact of epidemiological components for household transmission and the analysis of additional laboratory assays built into the diagnostic algorithms will lend support for the planning of interventions and mobilisation of resources for policy makers in TB and the national TB programs in a variety of settings, depending on resources available. The aim of this particular objective is to develop further policy links and share the algorithms and toolbox concept with our partners in NLTP in resource-poor and -rich settings. We anticipate that these algorithms could be tailored according to available resources and require to be evaluated in a variety of populations (+/- HIV). We will then liaise with experts in cost-effectiveness analysis (i.e. Imperial College Business school, HPA, WHO) to quantify the potential need for resources, depending on the algorithm chosen. This will allow the TB program decision makers to develop more appropriate predictions of impact depending on the settings that wish to adapt them.Through close links between our research and the National Leprosy and TB program (NLTP), the experience in The Gambia will inform international TB control efforts via communication channels such as the StopTB partnership and WHO.