Neutrophil proteins as potential biomarkers for a sputum-based tuberculosis screening test

Mark Chambers¹Farina Karim¹Matilda Mazibuko¹Zoey Mhlane¹Lindiwe Madziwa¹Yunus Moosa²Sashen Moodley¹Monjurul Hoque²Emily Beth Wong^1,2,3,4Andriette Hiemstra⁵Stephanus Theron Malherbe⁵Belinda Kriel⁵Kim Stanley⁵Ilana Claudia van Rensburg⁵Ayanda Shabangu⁵Bronwyn Smith⁵Gerhard Walzl⁵Nelita Du Plessis⁵Timothy R Sterling⁶Mark Hatherill⁷Alasdair Leslie^1,2,8*

• ¹Africa Health Research Institute (AHRI), Durban, South Africa
• ²University of KwaZulu-Natal, Durban, South Africa
• ³Reproductive Health and HIV Institute, Hillbrow, South Africa
• ⁴Division of Infectious Diseases, Massachusetts General Hospital, Boston, Massachusetts, United States of America, Boston, United States
• ⁵SAMRC Centre for Tuberculosis Research, Tygerberg, South Africa
• ⁶Division of Infectious Diseases, Department of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee, USA, Nashville, United States
• ⁷South African Tuberculosis Vaccine Initiative, Institute of Infectious Disease & Molecular Medicine and Division of Immunology, University of Cape Town, Cape Town, South Africa, Cape Town, South Africa
• ⁸University College London, London, United Kingdom

The development of a rapid and affordable assay to screen participants for additional testing could streamline TB screening in resource-limited settings and for community-wide health screens. Sputum remains the primary testing sample, making it potentially ideal for a screening testing. Neutrophils are highly expanded in sputum from individuals with pulmonary TB with high specificity and have potential as a biomarker for TB. Three neutrophil associated proteins, neutrophil gelatinase associated-lipocalin (NGAL), the protein heterodimer S100A8/A9 and the protein death ligand-1 (PDL-1), were measured in presumptive TB cases from participants attending a primary healthcare clinic in Durban, South Africa, using commercially available ELISAs on a total of 79 participants from a 109-participant cohort. Participants with microbiologically confirmed TB were sampled after 1 month of treatment. Proteins were also measured in tongue swab samples in participants from this cohort at baseline. Baseline results were confirmed in a second TB cohort which recruited a total of 51 participants with presumptive TB from the Western Cape. Finally, we investigate sputum neutrophil protein levels in individuals with community-diagnosed asymptomatic TB. Significant increases in all proteins were detectable in sputum from clinic-diagnosed TB participants relative to symptomatic controls. Performance approached the WHO target product profile for a TB triage test, with ROC AUCs reaching 0.866 (with a 95% confidence interval of 0.7683 -0.9633) in the case of S100A8/A9. Sputum protein levels did not correlate with bacterial burden and did not consistently decrease following one month of drug therapy. Only PDL-1 was detectable in mouth swab samples. Sputum neutrophil proteins tended to be elevated in participants with asymptomatic community diagnosed TB, as compared to asymptomatic community controls within the Vukuzazi cohort using a sample size of 42 participants, although this was not significant. This study provides a proof of principle that neutrophil proteins can be easily measured in standard sputum samples and have potential as a screening test for TB. However, more work is needed to explore whether this approach, using these three neutrophil proteins, can meet the WHO target product profile for a triage test worth developing further.