Mycobacterial T-cell responses in BCG-vaccinated HIV-exposed infants and risk of Mycobacterium tuberculosis infection
AJ Warr1, C Anterasian2, JA Shah2,3, SM LaCourse2, SC DeRosa4, FK Nguyen2, E Maleche-Obimbo5, LM Cranmer6, D Matemo7, J Kinuthia7, GC John-Stewart2, TR Hawn2
1Baylor College of Medicine, Texas; 2U. of Washington, Washington; 3VA Puget Sound Healthcare System, Washington
4Fred Hutchinson Cancer Research Center, Washington
5U. of Nairobi, Kenya; 6Emory University and Children’s Healthcare of Atlanta, Georgia; 7Kenyatta National Hospital, Kenya
The correlates of risk of BCG induced immunity to Mycobacterium tuberculosis (Mtb) infection are unknown. HIV-exposed uninfected (HEU) infants are at increased risk of Mtb infection and tuberculosis disease. We hypothesized that mycobacterial antigen-specific T-cell responses following BCG vaccination are associated with risk of Mtb infection in HEU infants.
HEU infants were enrolled in Kisumu, Kenya and followed for 1 year in a randomized clinical trial of isoniazid to prevent Mtb infection and determine immunologic correlates of risk of Mtb infection (N=300). Cryopreserved PBMCs from 6-10 weeks of age were stimulated with Mtb whole cell lysate (TBWCL) or media control, and the percentage of CD4 cells expressing IFNγ, IL-2, and TNF were measured by flow cytometry. Infants underwent Quantiferon-Plus (QFT) and tuberculin skin testing (TST) at 14 months of age. TNF, IL-2, and IP-10 levels were also measured in QFT supernatants by Luminex. Infants were considered to have Mtb infection if they had a positive QFT, TST induration >=10mm, or QFT non-IFNγ cytokine expression >=90th percentile.
At 14 months of age, 1.2% of infants were QFT positive, 13.3% were TST positive (induration >=10mm), and an additional 6.1% were considered positive by non-IFNγ cytokine production in QFT supernatants (cytokine level >=90th percentile). A CD4 polyfunctional antigen-specific response to TBWCL was detected in 90.3% of infants (COMPASS posterior probability) at 6-10 weeks of age following BCG vaccination at birth. TBWCL CD4 responses did not differ between infants who subsequently developed Mtb infection (detected by QFT, TST, or cytokine) at 14 months of age: median CD4 polyfunctional score of 0.15 vs 0.13 (p=0.4), IFNγ+ percentage of CD4 cells 0.01 vs 0.01 (p=0.8), IL-2+ CD4 cells 0.03 vs 0.03 (p=0.9), or TNF+ CD4 cells 0.04 vs 0.05 (p=0.5). Maternal viral load and CD4:CD8 ratio were not associated with infant CD4 cytokine production.
Despite known differences in T-cell function and increased rates of Mtb infection and disease in HEU infants, Th1 T-cell responses measured at 6-10 weeks following BCG vaccination were not associated with increased risk of Mtb infection at 14 months as measured by QFT, TST, or QFT non-IFNγ cytokine production. Non-IFNγ cytokines were detected in certain infants with otherwise negative QFT results and may improve measurement of Mtb infection in children.