In resource strapped high TB burden countries the need of the day is to have easily detectable biomarkers in accessible samples which distinguish latent infection from active tuberculosis at both pulmonary and extra-pulmonary sites, and for monitoring successful recovery post chemotherapy. Although samples from infective sites [1-5] may be more representative of biomarker changes associated with active disease, obtaining samples from the infected disease sites is invasive and in extra-pulmonary tuberculosis several sites are not accessible and not applicable at the field level. Peripheral whole blood therefore becomes the sample of choice for assessing biomarkers associated with active disease in tuberculosis [2]. Biomarkers assessed in whole blood range from antibodies [3] to gene expression profiling, [4] to cytokines, [5] with the markers of choice being cytokines associated with the pro- and down inflammatory arm of the immune response [6]. These cytokines upon in vivo activation are produced by several different cell types which are present in the blood. Therefore cytokine profiling rather than individual cytokines are now showing more promise.

Cytokines in whole blood have been analyzed in serum, plasma, supernatants of whole blood cultures after variable periods of incubation with or without stimulants. Although it has long been known that serum and Mycobacterium induced cytokine responses do not show parallel responses [2,7,8], the reasons for such differences are still not unclear [9]. In case of serum/plasma the inter-dilution effect can be considerable and the discrimination between patients with active tuberculosis and healthy BCG vaccinated endemic controls is different at different dilution (Figure S1; Hussain et al. manuscript in prep). It is therefore important that controls and test samples are compared at the same dilution or acceptable inter-dilution coefficient of variation is mentioned. However, the reports seldom mention this point. Because of this inter-dilution variability; the assay should also have a broad range of detection if several cytokine are being compared which can have log differences in concentrations. When analyzing cytokine concentrations in serum plasma or endogenously activated cells in the peripheral blood compartment [10] the issue of specificity comes up particularly if these readouts are to be used as point of care tests. Finally, because of low levels of cytokine secreting cells in this compartment very sensitive assays, with low limit of detection are required. Stimulation of whole blood cultures with specific mycobacterial antigens (ESAT-6, CFP-10) expands antigen specific T cells [11]. Again, there are two major limitations with this test. The boost with mycobacterial antigens is empirical and the concentrations of antigens used may have no relevance to the mycobacterial stimulation at the site of disease. Secondly, due to the non relevant concentrations of antigens, the cytokine profiles may alter during in vitro stimulation and may not reflect the events at the site of disease [12]. Finally, the choice of assay for readouts adds yet another layer of complexity to the issue of biomarker readout. It has been recognized that multiple biomarkers increase the specificity and sensitivity of assays. Multiplex technology has replaced individual biomarker assessment. This technology has several advantages over individual assessment. In addition to being more efficient and less time consuming, it can detect several cytokines in the same sample without intra-dilution and intrainter- assay variation (Human chemokine cytokine panel -29 Plex http://www.millipore.com/catalogue/item/MPXHCYTO-60K-29#; bdbiosciences.com). There are only limited reports that have evaluated multiplex technology with respect to different compartments [13,14]. Multiplex technology is quite expensive and therefore, can only be used for identification and selection of promising biomarkers. However, this technology does provide a single platform for assessing samples from varying sites and compartments and assessing concentrations over several logs to be tested simultaneously for discriminating cytokine profiles. There is a need to decide which compartment (Blood, serum, plasma, endogenously activated whole blood, T cells or stimulated WB) is appropriate and truly reflective of active tuberculosis and discriminatory among various tuberculosis states in an endemic area (exposed Vs non exposed, latently infected Vs non infected and disease severity) with high specificity and sensitivity and can be translated to a point of care test. Interferon g release assays do not discriminate between latent infection and active disease [15] and therefore multiplex biomarker profiling will be the test of choice. Finally, a user friendly strip test will have to be developed with stringent cut offs, if such cytokine profiling are to be made field applicable Point of Care Test. For this field to move forward a concerted effort at comparing different sample sources and cytokines profiles in well characterized patient and control groups using a multiplex methodology is an urgent need of the day for high TB burden countries.