Evaluation of host biosignatures for targeted screening for tuberculosis at the point of care, and monitoring of the response to TB treatment

Date
2020-03
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University, 2020
Abstract
ENGLISH ABSTRACT: Background The accurate and rapid diagnosis of TB disease remains one of the major challenges in the control of the disease. This is mainly due to limitations in the currently existing diagnostic tests. There is also an urgent need for the development of tools that may assist in monitoring the response of TB patients to treatment. This is important given the increasing burden of multi-drug resistant TB, the risk for treatment failure and of relapse after initial clinical cure and the need for early diagnosis of such poor outcomes. The work reported in this thesis investigated different biomarkers that may assist in addressing these challenges. Aims 1) To refine and validate different biomarkers selected from literature and evaluate novel biomarker combinations in a large Africa-wide study as biosignatures for the diagnosis of TB. 2) To evaluate combinations between different biomarkers selected from the literature and mycobacterial and clinical parameters (body mass index) as tools for monitoring the response to TB treatment and prediction of relapse. 3) To employ an unbiased proteomic approach in the search for new salivary biomarkers for TB. Methods Serum and saliva specimens collected through different multicentred projects including the African-European Tuberculosis Consortium (AE-TBC), the “Action TB” study and the “Tuberculosis Research Unit” (TBRU) study at different time points were evaluated. The concentrations of host inflammatory biomarkers in serum samples from all participants (aims 1 and 2 of the project) were evaluated using the Luminex multiplex platform. The investigation of new salivary biomarkers for TB was done following a discovery proteomics approach by mass spectrometry, using the Orbitrap QExacutive platform. Standard statistical analysis procedures as used in previously published studies were employed for data analysis. Main Findings 19 out of the 20 host inflammatory biomarkers that were selected from the literature on the basis of their performance in previous studies, and evaluated as diagnostic tools for TB in the AE-TBC cohort (n=1003 study participants, 277 of whom were finally diagnosed with TB and 726 of whom were diagnosed with other respiratory diseases [ORD]), showed potential in discriminating participants with TB and those with ORD individually, thereby confirming their potential as candidate TB diagnostic biomarkers. The diagnostic potential of modified versions of a previously published 7-marker signature (Apo A1, CFH, CRP, IFN-g, IP-10, SAA and transthyretin) was confirmed. However, we identified a novel 4 marker biosignature (I- 309, CRP, IP-10, and NCAM) and a novel smaller 3-protein biosignature (CRP, I-309 and NCAM) with strong potential as diagnostic biosignatures for active TB (AUCs of 0.91 and 0.90 respectively), regardless of HIV status. When the utility of host biomarkers selected from the literature was investigated for their potential to be used as biomarkers for monitoring of the response to TB treatment, it was observed that the concentrations of multiple biomarkers changed in the course of TB treatment. The levels of SAP, IP-10, and sIL-6R amongst other biomarkers increased with treatment, whereas those of Apo CIII, MMP-2 and Fas amongst others generally tended to decrease. Some of these proteins including IL-6, IP-10, IL-22 and C3 showed potential as individual biomarkers in predicting relapse both in the action tb and TBRU cohorts. However, the most important findings from this part of the work were the identification of a combination of TTP, BMI, sICAM, IL-22, IL-1β and C3 which predicted relapse at baseline, prior to the onset of TB treatment with 89% sensitivity and 94% specificity, and a combination of Apo CIII, IP-10 and sIL-6R which predicted relapse at the end of treatment with 71% sensitivity and 74% specificity. For the unbiased proteomic identification of salivary biomarkers part of the project, we identified a total of 948 unique proteins in saliva. After statistical analysis to identify the potentially most useful proteins, we identified 26 proteins with TB diagnostic potential. Importantly, two protein biosignatures including a five-protein signature comprising of P01011, Q8NCW5, P28072, Q99574 and A0A2Q2TTZ9 and a four-protein biosignature comprising of P30043, P35579, Q99574 and P31949. The five-protein signature ascertained TB with a sensitivity of 100% and specificity of 90.91% after leave-one-out cross-validation. Conclusion The usefulness of previously identified biomarkers was confirmed in large Africa-wide diagnostic study in the current thesis. Importantly, new smaller biosignatures comprising of as little as three proteins showed strong potential in TB diagnosis. Biosignatures comprising of combinations between immunological biomarkers, BMI and TTP were identified with strong potential for use as tools for monitoring of the response to TB treatment and prediction of TB relapse both prior to the initiation of treatment, early on after initiation of treatment and at the end of treatment. Results from this thesis show that the development of a rapid point-of-care diagnostic test comprising of 3 biomarkers is possible and that such a test will potentially be highly useful if used as a triage test for TB in peripheral level health centres all over Africa, which is where our study participants were recruited. We identified strong candidate biosignatures for prediction of relapse prior to the initiation of TB treatment, early on in treatment and at the end of treatment, and also novel salivary protein biomarkers with potential in the diagnosis of TB. However, the sample sizes were relatively small, and these novel findings require validation in larger studies, especially studies conducted in multiple geographical settings.
AFRIKAANS OPSOMMING: Agtergrond Die akkurate en vinnige diagnose van TB-siekte bly een van die belangrikste uitdagings in die beheer van die siekte. Dit is hoofsaaklik te wyte aan beperkings in die huidige diagnostiese toetse. Daar is ook 'n dringende behoefte aan die ontwikkeling van instrumente wat kan help met die monitering van die reaksie van TB-pasiënte op behandeling. Dit is belangrik in die lig van die toenemende las van weerstandige TB teen die medisyne, die risiko vir behandelingsmislukking en terugval na die aanvanklike kliniese genesing en die behoefte aan vroeë diagnose van sulke swak uitkomste. In die werk wat in hierdie proefskrif gerapporteer is, is verskillende biomerkers ondersoek wat hierdie uitdagings kan aanpak. Doelwitte 1) Om verskillende biomerkers wat uit literatuur geselekteer is, te verfyn en te bekragtig en nuwe biomerkmerkkombinasies te evalueer in 'n groot studie in Afrika as biosignature vir die diagnose van TB. 2) Om kombinasies te evalueer tussen verskillende biomerkers wat uit die literatuur gekies is en mykobakteriële en kliniese parameters (liggaamsmassa-indeks) as instrumente om die respons op TB-behandeling en voorspelling van terugval te monitor. 3) Om 'n onbevooroordeelde proteomiese benadering te gebruik in die soeke na nuwe speekselbiomerkers vir TB. Metodes Monsters van serum en speeksel wat deur verskillende multicentred-projekte ingesamel is, waaronder die Afrika-Europese Tuberkulose Consortium (AE-TBC), die 'Action TB' -studie en die 'Tuberculosis Research Unit' (TBRU) -studie op verskillende tydpunte, is geëvalueer. Die konsentrasies gasheer-inflammatoriese biomerkers in serummonsters van alle deelnemers (doelstellings 1 en 2 van die projek) is met behulp van die Luminex multiplex-platform beoordeel. Die ondersoek na nuwe speekselbiomerkers vir TB is gedoen volgens die ontdekking van proteomika-benaderings deur massaspektrometrie, met behulp van die Orbitrap QExacutive-platform. Standaard statistiese analise prosedures soos gebruik in voorheen gepubliseerde studies is gebruik vir data-analise. Uit 20 gasheer inflammatoriese biomerkers wat uit die literatuur geselekteer is op grond van hul prestasie in vorige studies, en geëvalueer is as diagnostiese instrumente vir TB in die AETBC- groep (n = 1003 deelnemers aan die studie, waarvan 277 uiteindelik met TB gediagnoseer is) en 726 van hulle is gediagnoseer met ander respiratoriese siektes (ORD), 19 het potensiaal getoon om deelnemers met TB en dié met ORD individueel te diskrimineer, en sodoende hul potensiaal as kandidaat-diagnostiese biomerkers vir TB te bevestig. merkerhandtekening (Apo A1, CFH, CRP, IFN-g, IP-10, SAA en NCAM) is bevestig, maar ons het 'n nuwe 4-merker bio-handtekening (I-309, CRP, IP-10 en NCAM) en 'n nuwe kleiner 3-proteïen biosignature (I- 309, NCAM en SAA) met 'n sterk potensiaal as diagnostiese biosignature vir aktiewe TB (AUC's van onderskeidelik 0,91 en 0,90), ongeag MIV-status. tydens die ondersoek na die potensiaal om gebruik te word as biomerkers vir die monitering van die respons op TB-behandeling, is daar waargeneem dat die konsentrasies van verskeie biomerkers gedurende die behandeling van TB verander het. Die vlakke van SAP, IP-10 en sIL-6R onder ander biomerkers het met behandeling gestyg, terwyl die vlak van onder andere Apo CIII, MMP-2 en Fas oor die algemeen geneig was om te daal. Sommige van hierdie proteïene as individuele biomerkers, insluitend IL-6, IP-10, IL-22 en C3, het potensiaal getoon soos om die terugval te voorspel, beide in die aksie tb- en TBRU-kohorte. Die belangrikste bevindings uit hierdie deel van die werk was egter die identifisering van 'n kombinasie van TTP, BMI, sICAM, IL-22, IL-1β en C3 wat terugval by die beginlyn voorspel het voor die aanvang van TB-behandeling met 89%. sensitiwiteit en 94% spesifisiteit, en 'n kombinasie van Apo CIII, IP-10 en sIL-6R wat terugval voorspel het aan die einde van die behandeling met 71% sensitiwiteit en 74% spesifisiteit. Vir die onbevooroordeelde proteomiese identifikasie van 'n speekselbiomerk-deel van die projek, het ons 'n totaal van 948 unieke proteïene in speeksel geïdentifiseer. Na statistiese ontleding om die mees bruikbaarste proteïene te identifiseer, het ons 26 proteïene met TBdiagnostiese potensiaal geïdentifiseer. Die belangrikste is dat twee proteïenbiohandtekeninge insluitend 'n vyfproteïne-handtekening bestaan uit P01011, Q8NCW5, P28072, Q99574 en A0A2Q2TTZ9 en 'n bio-handtekening met vier proteïene wat bestaan uit P30043, P35579, Q99574 en P31949. Die vyfproteïne-handtekening het TB vasgestel met 'n sensitiwiteit van 100% en spesifisiteit van 90,91% na kruisvalidering met verlof eenmalig. Afsluiting Die nut van voorheen geïdentifiseerde biomerkers is in die huidige tesis bevestig in 'n groot diagnostiese studie in Afrika. Die belangrikste is dat nuwe kleiner bio-handtekeninge wat bestaan uit so min as drie proteïene, sterk potensiaal het in die diagnose van TB. Biosignature wat bestaan uit kombinasies tussen immunologiese biomerkers, BMI en TTP, is geïdentifiseer met 'n sterk potensiaal om te gebruik as hulpmiddels vir die monitering van die respons op TB-behandeling en voorspelling van TB-terugval beide voor die aanvang van die behandeling, vroeg na die aanvang van die behandeling en aan die einde van die behandeling. Resultate uit hierdie proefskrif toon dat die ontwikkeling van 'n vinnige diagnostiese toets vir sorgsaamheid bestaande uit 3 biomerkers moontlik is en dat so 'n toets potensieel nuttig sal wees as dit gebruik word as 'n triage-toets vir TB in perifere vlak gesondheidsentrums oor die hele Afrika. , dit is waar ons studie-deelnemers gewerf is. Ons het sterk kandidaat-bio-handtekeninge geïdentifiseer vir voorspelling van terugval voor die aanvang van TB-behandeling, vroeg in die behandeling en aan die einde van die behandeling, en ook nuwe speekselproteïnebiomerkers wat potensiaal het vir die diagnose van TB. Die steekproefgroottes was egter relatief klein, en hierdie nuwe bevindings vereis validering in groter studies, veral studies wat in veelvuldige geografiese instellings uitgevoer is.
Description
Thesis (PhD)--Stellenbosch University, 2020.
Keywords
Tuberculosis -- Diagnosis, Treatment monitoring, Biochemical markers, Salivary proteins, Medical screening, Treatment response, UCTD
Citation