Article Text


The science of tuberculosis: current concepts

Statistics from


1MPR Berry, 2D Chaussabel, 3OM Kon, 2J Banchereau, 1A O’Garra. 1Division of Immunoregulation, National Institute for Medical Research, London, UK, 2Baylor Institute for Immunology Research and Baylor Research Institute, Dallas, USA, 3St Mary’s Hospital, Imperial College Healthcare NHS Trust, London, UK

Introduction: Tuberculosis is a major and increasing cause of morbidity and mortality worldwide caused by infection with Mycobacterium tuberculosis (MTb). However, the majority of individuals infected with MTb remain asymptomatic, apparently retaining the infection in a quiescent form. It is thought that this latent infection is maintained by an active immune response. The immune response to MTb is complex and remains incompletely characterised. We hypothesised that obtaining whole genome expression profiles from the blood of patients with either latent or active tuberculosis could lead to improved insights into the protective immune response, as well as the development of potential biomarkers.

Methods: We recruited patients with active pulmonary tuberculosis and latent tuberculosis. All active tuberculosis patients included in the analysis were culture positive and all latent patients included were positive by both tuberculin skin test and interferon-gamma release assay (IGRA). Potentially immunosupressed individuals were excluded. We also recruited healthy controls, who tested negative by both skin test and IGRA. Whole blood was collected from all participants before treatment. RNA was extracted from the whole blood and microarray analysis performed using Illumina Sentrix Human-6 V2 BeadChip arrays (>48 000 probes). Statistical analysis and hierarchical clustering of samples was performed using Genespring, v7.1.3.

Results: After filtering out undetected transcripts and those with less than twofold change, 6269 transcripts were used for unsupervised clustering analyses by Pearson correlation. This revealed distinct expression profiles, or “disease signatures”, corresponding to the clinical diagnosis of latent or active tuberculosis. These disease signatures held true across the broad number of ethnicities represented in the study. To improve discrimination between clinical groups, statistical comparisons were performed yielding a list of 119 genes. Unsupervised hierarchical clustering using this list separated participants into three distinct groups of active tuberculosis, latent tuberculosis and healthy controls.

Conclusions: We identified distinct gene expression profiles in active and latent tuberculosis, which were also distinct from control individuals. Our findings will yield insights into the potential mechanisms of immune protection (latent tuberculosis) versus immune pathogenesis (active tuberculosis). They also have value for the development of biomarkers for use in diagnosis.


1TSC Hinks, 1DPS Dosanjh, 2JA Innes, 3G Pasvol, 2S Hackforth, 3H Varia, 1KA Millington, 1R Gunatheesan, 4JJ Deeks, 1A Lalvani. 1Tuberculosis Immunology Group, Imperial College London, London, UK, 2Department of Infection and Tropical Medicine, Birmingham Heartlands Hospital, Birmingham, UK, 3Department of Infection and Tropical Medicine, Northwick Park Hospital, Harrow, UK, 4Department of Public Health and Epidemiology, University of Birmingham, Birmingham, UK

Introduction and Objectives: The majority of individuals infected with Mycobacterium tuberculosis achieve lifelong containment of the bacillus, but what constitutes this effective host immune response is poorly understood. Antigen-specific CD4 IFN-γ-secreting T-cell responses to mycobacterial antigens are believed to be pivotal, but the relationship between the strength of these responses and the development of protective immunity or of pathology is unclear.

Methods: Using the ex vivo enzyme-linked (ELISpot) assay, we compared frequencies of IFN-γ secreting T cells among 205 prospectively recruited individuals with either non-recent latent infection, who have successfully achieved immune control, or with active disease, whose immune responses have failed to contain the bacillus.

Results: Among subjects with an IFN-γ ELISpot response to one or more RD1-encoded antigens, T cells from subjects with active disease recognised a greater number of peptide pools from those antigens than did subjects with latent infection (p = 0.002). There was a trend for any given region of difference-1 (RD1) antigen or peptide pool to be recognised by a greater proportion of subjects with active disease than with latent disease and these differences were significant for early secretory antigenic target 6 (ESAT-6) peptide pools, recombinant ESAT-6 antigen and recombinant culture filtrate protein 10 (CFP-10) antigen (fig A,B). T-cell frequencies in responding wells were greater in active than latent infection for summed pools of RD1 peptides (p⩽0.006), and for CFP-10 antigen (p = 0.029) (fig C,D). The hierarchy of immunodominance among antigens and peptides was highly correlated between active and latent disease. Responses to the DosR antigen α-crystallin were not associated with latency (p = 0.373). RD1-specific T-cell responses were greater among subgroups positive to tuberculin skin testing (TST). Proportions of individuals positive to TST were higher among active than latent cases (87.9% vs 44.1% p<0.001). Whereas RD1 antigen responses are higher in active infection, responses to PPD are strongly associated with TST (p⩽0.025), but not with clinical status (p>0.7).

Conclusions: Our results suggest that RD1-specific IFN-γ-secreting T-cell frequencies correlate with the presence of pathology rather than with protective immunity.


1A Dunleavy, 2CJ Smith, 2RAM Breen, 3HJ Stauss, 1MCI Lipman. 1Department of Thoracic Medicine, The Royal Free Hospital, London, UK, 2The Royal Free and University College Medical School, London, UK, 3Department of Immunology and Molecular Pathology, University College London, The Royal Free Hospital, London, UK

Introduction: Protective host immunity against Mycobacterium tuberculosis requires tissue T-cell responses in sites such as the lung. Most reported work has used blood samples—and this may not reflect accurately the local immune response. The recent appreciation of the heterogeneous nature of memory T cells; together with the possibility of a compartmentalised tissue response, led us to investigate whether we could demonstrate distinct differences in blood and lung T-cell memory populations between tuberculosis patients and relevant controls.

Methods: T-cell differentiation markers were studied in paired blood and lung samples (the latter obtained by sputum induction). Seventeen subjects with active tuberculosis and six healthy controls were recruited. Tuberculosis was pulmonary in 11 cases, lymph node in three and abdominal, cerebral and renal in one each. Two of 17 subjects were HIV co-infected. Fresh samples were stained with CD3 and CD4 and with one of the following combinations: CD45RA and CCR7; CD27 and CD62L; CD27 and CD28; CD57 and CD62L. All panels were analysed using four-colour cytometry.

Results: There were significant differences between blood and lung T-cell populations. At the latter site, tuberculosis patients had a significantly smaller proportion of central memory (CD45 RACCR7+, TCM) and a greater proportion of effector memory (CD45 RACCR7,TEM) T cells than healthy controls (fig). There was no significant difference in the same populations in blood (p = 0.62). In contrast, significant differences were demonstrated between CD27 CD62L, CD27 CD28 and CD57 CD62L CD4 T-cell populations (p = 0.001, p = 0.001 and p = 0.03, respectively) in blood but not sputum, when tuberculosis patients were compared with controls.

Conclusions: Memory T-cell populations differ markedly between blood and lung fluid and within specific populations at these sites. This highlights the limitations of blood-based assays when investigating tuberculosis pathogenesis. Furthermore, it suggests that simple phenotypic markers may lack the sensitivity required to discriminate those T-cell populations that are recruited and play an active role in immune regulation at tissue sites. The future use of antigen-specific assays may assist with this.

Abstract S147 Figure Distribution of CD3+CD4+ cells in induced sputurm based on CCR7 and CD45RA staining. *p = 0.01 **p = 0.003


1AB Hardy, 1R Varma, 2C Mullarkey, 1S Moffitt, 1T Collyns, 1JP Watson. 1Leeds Teaching Hospitals NHS Trust, Leeds, UK, 2Leeds Primary Care Trust, Leeds, UK

Background: NICE guidelines for new entrant tuberculosis screening recommend chest x ray (CXR) for immigrants from countries with tuberculosis incidence greater than 40/105, and tuberculin skin test (TST) for people with normal CXR from very high tuberculosis prevalence countries.1 The Leeds tuberculosis screening service serves a population of 750 000. Based on previous audit, we piloted a revised screening policy using first-line QuantiFERON-TB GoldTM (QFT) in high-risk immigrants from 1 January 2007. We have evaluated the cost effectiveness of this approach.

Methods: Initially, TST was offered to immigrants from countries with tuberculosis incidence 200–339/105, and QFT to those from countries with incidence greater than 340/105. When increased resources became available, all immigrants from countries with tuberculosis incidence greater than 200/105 had QFT. Those with positive QFT were invited for CXR. Costs were estimated for our protocol and for NICE protocol based on NICE estimates for cost of CXR £23.24, TST £13.69 and QFT £25.67.1 For the NICE protocol costings, the expected rate of positive TST for the group was estimated from the subgroup who had both tests, assuming that there would be no false negative TST.

Results: In 2007 we identified 2902 immigrants from countries with tuberculosis incidence greater than 200/105, 1336 were invited for screening (TST or QFT as described above) with a 32% attendance rate. 280 patients had QFT, of which 38% were positive, with less than 2% being indeterminate. Using the NICE approach the cost of screening these 280 immigrants would be £13 334 (£47.62 per immigrant) and would identify 83 cases of latent tuberculosis infection (LTBI). Using first line QFT followed by CXR the cost was £9652 (£34.47 per immigrant) and identified 106 LTBI. The cost to identify one case of LTBI following NICE guidelines would be £160.66 and using our protocol was £91.06.

Conclusions: Using QuantiFERON-TB GoldTM blood testing followed by CXR is more effective and more cost effective than NICE guideline for screening new entrants from high-risk countries.


1H Thuraisingam, 2J Lee, 2W Hoskyns, 3G Woltmann. 1Leicester City PCT, Leicester, UK, 2University Hospitals Leicester, Leicester, UK, 3Institute for Lung Health, Leicester, UK

Introduction: T-cell based interferon gamma release assays (TIGRA) are highly specific for the diagnosis of latent tuberculosis infection (LTBI) or tuberculosis disease. Given the relatively high incidence of tuberculosis in the city of Leicester, tuberculosis contacts were screened by Quantiferon (QFN) and/or chest x ray (CXR) in 2007.

Methods: All tuberculosis contacts below the age of 36 years and contacts of all ages of smear-positive pulmonary index cases were invited for TIGRA testing by QFN assay in 2007, 2–3 months after the notification date. Results are presented according to age and ethnicity of contacts as well as disease site, smear/culture positivity of the index case.

Results: 989 contacts of 269 index cases were identified and 464 contacts (47%) were QFN tested. 158/251 (63%) children (age <16 years), 220/383 (53%) of young adults (age 16–35 years), and 86/355 (24%) of older adults (age >35 years) were tested. QFN positivity was found in 11/158 (7%) children, 53/220 (24%) young adults and 37/86 (43%) adults above the age of 35 years. Indeterminate results were found in 11/58 (7%) children but in only 1/306 (0.3%) contacts above the age of 15 years. Positivity rates were similar for contacts of Indian subcontinent origin (80/399, 23.6%) and white caucasian contacts (13/58, 22.4%). Overall positivity rates were highly age dependent (table) but highly significant differences were seen between young contacts of respiratory and of non-respiratory index cases with the highest positivity rates in contacts of smear/culture-positive pulmonary index cases (age <10 years, 14.8%; age 10–19 years, 37%; age 20–29 years, 35.1%). Not a single positive QFN result was found in 34 contacts of non-respiratory index cases below the age of 20 years (age 0–19 years, 0%) but positivity rates increased above the age of 20 years even in contacts of non-respiratory index cases (age 20–35 years, 23.6%).

Conclusion: One-year results of a single step contact screening protocol involving TIGRA testing of all contacts below the age of 36 years are presented. Positivity is strongly correllated with smear/culture positivity and disease site of the index case. A probable cohort effect is seen in children and youngsters below the age of 20 years with no QFN positivity in those in recent contact with non-infectious index cases.

Abstract S149 Table Results summary of QFN testing in all contacts eligible for chemoprophylaxis by age groups (2007)
View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.