The Duffy-null state is associated with a survival advantage in leukopenic HIV-infected persons of African ancestry

Abstract

Persons of African ancestry, on average, have lower white blood cell (WBC) counts than those of European descent (ethnic leukopenia), but whether this impacts negatively on HIV-1 disease course remains unknown. Here, in a large natural history cohort of HIV-infected subjects, we show that, although leukopenia (< 4000 WBC/mm(3) during infection) was associated with an accelerated HIV disease course, this effect was more prominent in leukopenic subjects of European than African ancestry. The African-specific -46C/C genotype of Duffy Antigen Receptor for Chemokines (DARC) confers the malaria-resisting, Duffy-null phenotype, and we found that the recently described association of this genotype with ethnic leukopenia extends to HIV-infected African Americans (AAs). The association of Duffy-null status with HIV disease course differed according to WBC but not CD4(+) T-cell counts, such that leukopenic but not nonleukopenic HIV(+) AAs with DARC -46C/C had a survival advantage compared with all Duffy-positive subjects. This survival advantage became increasingly pronounced in those with progressively lower WBC counts. These data highlight that the interaction between DARC genotype and the cellular milieu defined by WBC counts may influence HIV disease course, and this may provide a partial explanation of why ethnic leukopenia remains benign in HIV-infected AAs, despite immunodeficiency.

Figure 1

Association of low WBC counts during disease with survival in HIV-infected subjects. Each panel shows Kaplan-Meier (KM) survival plots for subjects with an average WBC count during disease that was < (orange) or ≥ 4000 cells/mm³ (purple). The KM plots are for all subjects (left column), and those with baseline CD4⁺ T-cell counts of < 350 (middle column) ≥ 350 (right column) cells/mm³ in the entire cohort (top row), and subjects categorized as seroconverter (middle row) and seroprevalent (bottom row) component of the cohort. RH indicates relative hazards (reference RH = 1 represents those with WBC counts of ≥ 4000 cells/mm³); CI, confidence interval; P, significance value estimated using Cox proportional hazards modeling; and n, number of subjects.

Figure 2

WBC counts according to race/ethnicity and DARC −46C/C genotype. (A) Box-and-whisker plots depicting the distribution of the average of the total WBC counts during disease, in the major ethnic groups represented in the study population. AA indicates African American; EA indicates European American. OT indicates other ethnicities. Numbers at the top are P values obtained by Mann-Whiney tests for the indicated comparisons. N indicates number of subjects. (B) The proportion of subjects with leukopenia (initial WBC count < 4000 cells/mm³) in the indicated racial/ethnic groups. Vertically oriented numbers are significance values obtained by Pearson χ² test. (C) The proportion of AA (n = 397) possessing the indicated DARC T-46C genotype (122 subjects with T/C or T/T and 275 subjects with C/C) who were leukopenic at presentation. The significance value for the comparison was obtained using the Pearson χ² test. (D) The bar charts represent the mean (vertical bars) and 95% confidence intervals (error bars) for the average WBC counts during disease in the indicated racial/ethnic groups according to their DARC T-46C genotype. Numbers at the top are P values obtained by Mann-Whiney tests for the indicated comparisons. N indicates number of subjects. *Too few subjects (1 EA and 3 others), for whom the vertical bars are not shown.

Figure 3

Differential effects of low WBC counts on HIV disease course in HIV-infected EA and AA. KM plots depict survival curves computed according to race/ethnicity and average WBC count in the entire cohort (A), and HIV⁺ subjects categorized as seroconverter (B), and seroprevalent (C). RH indicates relative hazards (reference RH = 1 represents AAs with WBC counts ≥ 4000 cells/mm³); CI, confidence interval; P, significance values obtained by Cox proportional hazards modeling; and n, number of subjects.

Figure 4

Rate of HIV disease progression in HIV⁺ EAs, and HIV⁺ AAs who are Duffy-null (DARC −46C/C) or -positive (DARC −46C/T or T/T). Each panel shows KM survival plots for 3 groups of HIV-infected subjects: EA (blue), AAs possessing DARC −46 C/C (pink), and AAs possessing DARC −46 C/T or T/T (green). The KM plots are for all subjects (left column) and those with baseline CD4⁺ T-cell counts of < 4000 (middle column) or ≥ 4000 (right column) cells/mm³ in the entire cohort (top row), and those subjects categorized as seroconverter (middle row) and seroprevalent (bottom row). RH indicates relative hazards (reference RH = 1 represents AAs with DARC −46C/C); CI, confidence interval; P, significance value estimated using Cox proportional hazards modeling, n, number of subjects. Whether DARC −46C/T or T/T was associated with disease acceleration in leukopenic seroconverting AAs could not be evaluated as in this group; there were only 4 subjects who did not have −46C/C, and no death events had occurred (green survival curve in panel E).

Figure 5

Relationship between decrements in the average WBC counts and the hazard ratios for time to death associated with the DARC −46C/C genotype in HIV-infected AAs. The plot represents the hazard ratios (y-axis) for rate of progression to death estimated using Cox proportional hazards modeling in subjects with average WBC counts below the indicated cutpoints (x-axis). The hazard ratios were estimated over the range of WBC counts shown on the x-axis for every 50-cell count change. The reference group is those subjects who possess DARC −46T/T or T/C (RH = 1). Inset represents data showing an interactive effect of low total WBC count on the survival advantage associated with the DARC −46C/C genotype in HIV⁺ AAs. RH indicates relative hazards (reference RH = 1 represents AAs with DARC −46T/T or T/C); CI, confidence interval; and P, significance value estimated using Cox proportional hazards modeling with indicated parameters (a, b, and c).