Current tuberculosis diagnostic testing remains challenging despite years of development [3]. Poor sensitivity of tests based on conventional culture and molecular diagnostic methods highlights the need for novel diagnostics geared toward diagnosing active tuberculosis patients, especially those with paucibacterial infections [17]. In this study, we developed a new immunological diagnostic method that can differentiate between active TB and other pulmonary diseases based on cytokine release responses when patient PBMCs were exposed in vitro to MTB antigen. The sensitivity rate of the parallel-type combination assay for IFN-γ and IL-2 release was 87.9% as compared to the rate of 83.8% obtained using an assay based only on IFN-γ release. The greater sensitivity rate of the combination assay over that of the single release assay mainly stemmed from the addition of testing for IL-2 release as a parallel diagnostic marker. Importantly, this disparity also reflects the fact that approximately 16% of active TB cases had negative IFN-γ results, a quarter of which yielded positive IL-2-release results. Similarly, a recent systematic review provided an estimated pooled sensitivity rate for QFT-GIT results of 81% for active TB cases [8] and aligned with previous studies attributing low sensitivity and false-negative IGRA test results to immunosuppression [8]. By contrast, based on positive control results (PBMCs stimulated with phytohemagglutinin) as an indicator of a patient’s immune response, we did not identify immunosuppression as a risk factor for false-negative results obtained for active cases. Nevertheless, a potential explanation for negative IGRA results may relate to the spatial compartmentalization of T cells in the body during infection [18]. Considering that peripheral blood T cells are isolated from blood in order to measure IFN-γ levels via in vitro IGRA assays, such assays may miss TB antigen-specific T cells previously recruited to infection sites from the blood during the initial course of TB disease; this loss of blood antigen-specific T cells may thus explain the negative IGRA results we obtained for early TB cases.
Of note, we failed to detect IFN-γ release in a small number of active cases that harbored a small proportion of MTB-specific T cells that only secreted IL-2. In line with this observation, an experimental study on the dynamic relationship between IFN-γ and IL-2 profiles during the natural course of human tuberculosis demonstrated that newly detectable CD4 + T cells secreting only IL-2 were present during and after treatment [19]. This finding was potentially valuable as it could be used for differentiating between active TB and LTBI. In this work, we conducted a detailed analysis of MTB-specific cytokine excretion and confirmed that IL-2 release was more specifically associated with active TB cases than was IFN-γ release. During the course of chronic TB infection, MTB-specific IL-2-secreting T cells exhibiting high expression of PD-1 have been shown to emerge during the late phase of the disease that do not secrete IFN-γ [20]. The emergence of such cells may serve as a plausible explanation for the higher specificity observed for IL-2 versus IFN-γ release for assays used for detecting active TB. In such cases, series-type combination testing for both IFN-γ and IL-2 release provided increased specificity (96.0%) that was comparable to that obtained using molecular diagnostic testing of culture-positive cases. Thus, the series-type combination assay was more effective than other cytokine release tests evaluated here for diagnosing culture-negative cases.
In view of the diagnostic utility of combining IFN-γ and IL-2 release assays, we modeled false-positive and false-negative results obtained for active TB cases against the variable prevalence of active TB (Fig. 3) then we proposed two diagnostic algorithms suitable for different hospital settings. For general hospitals in China which are in charge of screening suspected tuberculosis cases, approximately 10% of patients with symptoms suggestive of TB are ultimately diagnosed with active TB [21]. In this type of clinical setting, use of parallel-type combination release assays with higher sensitivity would help clinicians identify more patients at high risk of active TB than would conventional smear microscopy. Meanwhile, in TB-specialized hospital settings the proportion of active TB cases of the total TB suspected cases approaches 50% [22], justifying use of the series-type combination assay to achieve earlier diagnosis of active TB patients and prevent administration of toxic treatments with limited clinical efficacy.
Another interesting finding of our study is the association between diagnostic accuracy and severity of disease in the population. The sensitivity rate of the series-type combination assay when used for testing of definite TB cases was significantly higher than the rate obtained for clinically diagnosed cases; this result aligns with results of a report by Chee and colleagues indicating that a quantitative T cell response is positively associated with mycobacterial burden and disease activity [23]. However, conflicting results reported for another experimental study ruled out any clear correlation between antigen burden and T cell response [24]. Nevertheless, although our combination assay results for active TB patients revealed enhanced in vitro IFN-γ/IL-2 release with increasing MTB-specific antigen dose, further research is needed to investigate the dynamics of MTB-specific IFN-γ/IL-2 secretion during anti-TB treatment.
This study had several obvious limitations. First, despite strict diagnostic criteria, some patients afflicted with respiratory illness may have been improperly diagnosed as active TB cases due to lack of laboratory support. Such a diagnostic error may have negatively impacted the reliability of our study conclusions. Second, although comorbidities such as diabetes and immunosuppression status are considered risk factors for negative immune responses, comorbidity data were not collected here and thus were not considered in our evaluation of the clinical performance of our immunological assays. Third, we only evaluated the diagnostic utility of combination assays for pulmonary TB cases and should therefore confirm whether these combination tests would be effective for diagnosing extrapulmonary TB as well.