Thesis (PhD)--Stellenbosch University, 2008. / ENGLISH ABSTRACT: Setting Study conducted in Tygerberg, Cape Town in South Africa. Hypothesis Host biomarkers associated with the antimycobacterial immune response during active infection with M. tuberculosis and during anti-tuberculosis chemotherapy are indicative of bacterial killing in the host and can be used in models to predict eventual treatment outcome. Objectives 1. To investigate immune parameters that were selected in a biological context as biomarkers of the extent of disease and early response to anti-tuberculosis treatment. 2. To use selected immune parameters to characterise fast and slow responders to anti-tuberculosis therapy. Findings Evaluation of cytokine multiplex fluorescent bead-based immunoassays as a screening tool in the search for biomarkers The data showed that cytokine multiplex fluorescent bead-based immunoassays achieved acceptable recoveries to detect antigen-specific IFN- responses in whole blood supernatant making it attractive for biomarker screening. However, proper optimisation needs to be done and proper controls included when using these kits. Markers of extent of disease High levels of CRP at diagnosis were found to be associated with the presence of multiple cavities on chest X-rays. A high level of suPAR and sICAM-1 at diagnosis were associated with the extent of alveolar disease. Also significant were the associations between the level of granzyme B, LAG-3 at diagnosis and the size of the cavities. No significant associations were observed between sTNFRs or DR5 with the chest X-ray grading of tuberculosis disease. Early classification of fast and slow responders to anti-tuberculosis treatment After cross-validation classification, discriminant analysis (DA) and support vector machine (SVM) analysis of selected immune parameters (sICAM-1 CRP, granzyme B, suPAR, sTNFRs, LAG-3 and CD3dim/CD56+ (% of CD45+) resulted in a 75% to 100% correct classification of the fast responders and a 82% to 100% correct classification of the slow responders when using DA. For SVM, the correct classification of the fast responders ranged from 88% to 100%, and that for the slow responders ranged from 95% to 100%. Differential gene expression in fast and slow responders to treatment Direct comparison of fast and slow responders showed that IL-4 transcripts were significantly higher in the fast responders at week one after initiation of treatment when compared to slow responders. IL-42 was also differentially expressed. Although IL- was significantly up-regulated in both fast and slow responders after one week of treatment compared to diagnosis, IL- expression was more than two folds higher in slow responders than in fast responders. No significant differences between the fast and slow responders were observed in the expression of TGF-, TGF-RII, Foxp3 and GATA-3. Conclusion Predictive models for differential anti-tuberculous treatment responses combining host proteins are promising and should be included in larger prospective studies to find the optimal markers for inclusion into clinical trials of new drugs and for implementation into clinical practice. / AFRIKAANSE OPSOMMING: Ligging Studie onderneem in Tygerberg, Kaapstad, Suid-Afrika. Hipotese Gasheerbiomerkers wat verband hou met die antimikobakteriële immuunrespons tydens aktiewe infeksie deur M. tuberculosis en tydens teentuberkulose chemoterapie dui op bakteriële doding in die gasheer en kan in modelle gebruik word om die uiteindelike uitkoms van die behandeling te voorspel. Doelwitte 1. Om gekose immuunparameters in ’n biologiese konteks as biomerkers van die omvang van siekte en vroeë reaksie op behandeling te ondersoek. 2. Om gekose immuunparameters te gebruik om vinnige en stadige reageerders op teentuberkulosebehandeling te karakteriseer. Bevindings Evaluering van die sitokien veelvuldige fluoresseer-pêrelbaseerde immuuntoets (cytokine multiplex fluorescent bead-based immunoassays) as ’n siftingsinstrument in die soeke na biomerkers Die data het getoon dat die sitokien veelvuldige fluoresseer-pêrelgebaseerde immuuntoets in staat was om antigeenspesifieke IFN--respons te meet wat dit aanloklik maak vir biomerkersifting. Sorgvuldige optimering moet egter gedoen word en behoorlike beheer moet ingesluit word wanneer hierdie stelle gebruik word. Merkers van omvang van siekte Hoë vlakke van CRP by diagnose is getoon om verband te hou met die teenwoordigheid van veelvoudige holtes op die pasiënte se borskas x-strale. Hoë vlakke van suPAR en sICAM-1 by diagnose was assosieer met die omvang van alveolêre siekte. Die assosiasie tussen die vlakke van granzyme B, LAG-3 by diagnose en die grootte van die holtes was ook betekenisvol. Daar was geen betekenisvolle assosiasies toe sTNFRs of DR5 en die borskas x-straalgradering van tuberkulosesiekte nie. Vroeë klassifikasie van vinnige en stadige reageerders op teentuberkulosebehandeling Ná klassifikasie op grond van kruisstawing het diskriminant-analise (DA) en ondersteuningsvektormasjiene (SVM) van geselekteerde immuunparameters (sICAM-1 CRP, gransiem B, suPAR, sTNFRs, LAG-3 en CD3dim/CD56+ (% van CD45+)) gelei tot ’n 75% tot 100% korrekte klassifikasie van die vinnige reageerders met DA en ’n 82% tot 100% korrekte klassifikasie van stadige reageerders. Vir SVM het die korrekte klassifikasie van vinnige reageerders gewissel van 88% tot 100%, en vir stadige reageerders het dit gewissel van 95% tot 100%. Differensiële geenuitdrukking in vinnige en stadige reageerders op behandeling In vergelyking met die vlak by diagnose is die uitdrukkingsvlak van IL-4 in die vinnige reageerders betekenisvol opgereguleer met ’n faktor van 9.2 teen die eerste week ná die aanvang van behandeling, in kontras met die stadige reageerders. Daar was geen verskille tussen die vinnige en die stadige reageerders met betrekking tot die uitdrukking van TGF-, TGF-RII, Foxp3 en GATA-3 nie. Gevolgtrekking Voorspellende modelle vir differensiële tuberkulose behandelingsresponse wat gasheerproteïene kombineer, hou belofte in en behoort in groter prospektiewe studies ingesluit te word om die mees geskikte merkers te vind vir insluiting in kliniese proewe van nuwe middels en vir implementasie in kliniese praktyk.