Background: Asymptomatic CMV replication is frequent in HIV infected men, and is associated with increased immune activation, T cell proliferation and HIV disease progression. We hypothesized that persistent CMV replication influenced HIV DNA dynamics after initiation of antiretroviral therapy (ART) during early HIV infection.
Methods: We investigated 397 peripheral blood mononuclear cell (PBMC) samples collected from 96 CMV-seropositive, recently HIV-infected men. Participants started ART within a median of 5 months from estimated date of infection (EDI) (range: 0-8), achieved suppressed HIV RNA in blood within a median of 6 months on ART (range: 1-8) and were followed for a median of 17 months after ART start (range: 12-75). The median CD4 count at presentation was 485 cells/µl and the median peak HIV RNA was 4.8 log10 HIV RNA cp./ml. Levels of CD4-associated HIV DNA and CMV DNA were measured by droplet digital PCR for each time-point (mean 4 TP/participant). Using a general linear mixed-effect regression model, associations between HIV DNA decay, age, frequency of detectable CMV, nadir CD4 count, peak HIV RNA level, time from EDI to ART start, and time from ART start to virologic suppression were evaluated.
Results: Higher peak HIV RNA levels and higher frequency of detectable CMV in PBMC (>40% of sampled time-points) were associated with increased levels of HIV DNA during ART (p<0.01). Both factors were independently associated with higher HIV DNA in multivariable analysis (p<0.01). No other variable contributed significantly. The pattern of HIV DNA decay during ART differed significantly between participants with higher versus lower frequency of detectable CMV in PBMC (> vs < 40% of samples with detectable CMV). When considered separately, a linear model had a significantly better fit for HIV DNA decay in the low-frequency CMV group, while a quadratic model of HIV DNA decay had a better fit for participants with higher frequency of CMV DNA (p<0.01, see figure). In other words, individuals with more detectable CMV demonstrated a U-shaped pattern of HIV DNA decay while the low frequency CMV group had a linear pattern of decay.
Conclusions: Detectable CMV DNA in PBMC is associated longitudinally with higher HIV DNA levels, even among individuals who started ART early during HIV infection, suggesting that CMV replication may help maintain the stability of the HIV DNA reservoir. Future studies with anti-CMV therapeutics could help determine underlying mechanisms and if causal associations exist.
When considered separately, a linear model had a significantly better fit for HIV DNA decay in the low-frequency CMV group (<40% detectable CMV, left panel), while a quadratic model of HIV DNA decay had a better fit for participants with higher frequency of CMV DNA (>40% detectable CMV, right panel).