Detection of tuberculosis in HIV-infected and -uninfected African adults using whole blood RNA expression signatures: a case-control study

Abstract

Background: A major impediment to tuberculosis control in Africa is the difficulty in diagnosing active tuberculosis (TB), particularly in the context of HIV infection. We hypothesized that a unique host blood RNA transcriptional signature would distinguish TB from other diseases (OD) in HIV-infected and -uninfected patients, and that this could be the basis of a simple diagnostic test.

Methods and findings: Adult case-control cohorts were established in South Africa and Malawi of HIV-infected or -uninfected individuals consisting of 584 patients with either TB (confirmed by culture of Mycobacterium tuberculosis [M.TB] from sputum or tissue sample in a patient under investigation for TB), OD (i.e., TB was considered in the differential diagnosis but then excluded), or healthy individuals with latent TB infection (LTBI). Individuals were randomized into training (80%) and test (20%) cohorts. Blood transcriptional profiles were assessed and minimal sets of significantly differentially expressed transcripts distinguishing TB from LTBI and OD were identified in the training cohort. A 27 transcript signature distinguished TB from LTBI and a 44 transcript signature distinguished TB from OD. To evaluate our signatures, we used a novel computational method to calculate a disease risk score (DRS) for each patient. The classification based on this score was first evaluated in the test cohort, and then validated in an independent publically available dataset (GSE19491). In our test cohort, the DRS classified TB from LTBI (sensitivity 95%, 95% CI [87-100]; specificity 90%, 95% CI [80-97]) and TB from OD (sensitivity 93%, 95% CI [83-100]; specificity 88%, 95% CI [74-97]). In the independent validation cohort, TB patients were distinguished both from LTBI individuals (sensitivity 95%, 95% CI [85-100]; specificity 94%, 95% CI [84-100]) and OD patients (sensitivity 100%, 95% CI [100-100]; specificity 96%, 95% CI [93-100]). Limitations of our study include the use of only culture confirmed TB patients, and the potential that TB may have been misdiagnosed in a small proportion of OD patients despite the extensive clinical investigation used to assign each patient to their diagnostic group.

Conclusions: In our study, blood transcriptional signatures distinguished TB from other conditions prevalent in HIV-infected and -uninfected African adults. Our DRS, based on these signatures, could be developed as a test for TB suitable for use in HIV endemic countries. Further evaluation of the performance of the signatures and DRS in prospective populations of patients with symptoms consistent with TB will be needed to define their clinical value under operational conditions. Please see later in the article for the Editors' Summary.

Figure 1. Diagnostic process to identify TB cases, LTBI cases, and other diseases cases.

Figure 2. Study overview showing patient numbers and analysis pipeline.

HIV-, HIV-uninfected; HIV+, HIV-infected; TB, active tuberculosis (see Table 2).

Figure 3. Heatmaps showing clustering of training and test cohorts using transcriptional signatures.

Clustering of training (A/C) and test (B/D) cohorts using transcripts identified by elastic net for TB versus LTBI (A/B) and TB versus OD (C/D) (training: n _TB = 157 n _LTBI = 128/n _TB = 153 n _OD = 140, test: n _TB = 37 n _LTBI = 39/n _TB = 42 n _OD = 34). Rows are transcripts (transcripts shown in red are up-regulated, those in green are down-regulated) and columns are patients regardless of HIV status (purple, patients with TB; green, patients with LTBI; light blue, patients with OD).

Figure 4. Classification using the disease risk score on the test cohort and validation dataset.

Disease risk score and receiver operating characteristic curves based on the TB/LTBI 27 transcript signature (A/B) and the TB/OD 44 transcript signature (C/D) applied to the South African (SA)/Malawi HIV+/− test cohort (A/C) (n _TB = 37 n _LTBI = 39/n _TB = 42 n _OD = 34) and independent validation dataset comprising South African patients (B/D) (n _TB = 20 n _LTBI = 31 n _OD = 82). Sensitivity, specificity are reported in Table 3. HIV+, HIV-infected; HIV−, HIV-uninfected. Classification cut-offs: (A) 138.98; (B) 107.76; (C) 154.44; (D) 99.94.

Figure 5. Application of the transcript signatures to the South African and Malawi test cohorts by HIV status.

Disease risk score and receiver operating characteristic curves based on the TB/LTBI 27 transcript signature (A/B) and the TB/OD 44 transcript signature (C/D) applied to the HIV-uninfected (HIV−) (A/C) and HIV-infected (HIV+) (B/D) test cohort. Area under the curve, sensitivities, and specificities are reported in Table 3. Classification cut-offs: (A) 131.37; (B) 142.84; (C) 151.10; (D) 142.84.

Figure 6. Application of transcript signatures to the combined South Africa and Malawi cohorts.

Disease risk score and receiver operating characteristic curves based on transcript signatures of Berry et al. for TB versus LTBI (A/B/C) and TB versus OD (D/E/F) applied to the combined training and test cohorts in HIV-uninfected (HIV−) and HIV-infected (HIV+) (A/D), HIV− (B/E), and HIV+ (C/F) cohorts (Table 4 for sensitivities, specificities, and area under the curve). Classification cut-offs: (A) 1,847.73; (B) 1,777.65; (C) 1,898.97; (D) 172.12; (E) 170.30; (F) 173.70.