HIV Maintains an Evolving and Dispersed Population in Multiple Tissues during Suppressive Combined Antiretroviral Therapy in Individuals with Cancer

Abstract

While combined antiretroviral therapy (cART) can result in undetectable plasma viral loads, it does not eradicate HIV infection. Furthermore, HIV-infected individuals while on cART remain at an increased risk of developing serious comorbidities, such as cancer, neurological disease, and atherosclerosis, suggesting that during cART, tissue-based HIV may contribute to such pathologies. We obtained DNA and RNA env, nef, and pol sequences using single-genome sequencing from postmortem tissues of three HIV(+) cART-treated (cART(+)) individuals with undetectable viral load and metastatic cancer at death and performed time-scaled Bayesian evolutionary analyses. We used a sensitive in situ hybridization technique to visualize HIV gag-pol mRNA transcripts in cerebellum and lymph node tissues from one patient. Tissue-associated virus evolved at similar rates in cART(+) and cART-naive (cART(-)) patients. Phylogenetic trees were characterized by two distinct features: (i) branching patterns consistent with constant viral evolution and dispersal among tissues and (ii) very recently derived clades containing both DNA and RNA sequences from multiple tissues. Rapid expansion of virus near death corresponded to wide-spread metastasis. HIV RNA(+) cells clustered in cerebellum tissue but were dispersed in lymph node tissue, mirroring the evolutionary patterns observed for that patient. Activated, infiltrating macrophages were associated with HIV RNA. Our data provide evidence that tissues serve as a sanctuary for wild-type HIV during cART and suggest the importance of macrophages as an alternative reservoir and mechanism of virus spread.

Importance: Combined antiretroviral therapy (cART) reduces plasma HIV to undetectable levels; however, removal of cART results in plasma HIV rebound, thus highlighting its inability to entirely rid the body of infection. Additionally, HIV-infected individuals on cART remain at high risk of serious diseases, which suggests a contribution from residual HIV. In this study, we isolated and sequenced HIV from postmortem tissues from three HIV(+) cART(+) individuals who died with metastatic cancer and had no detectable plasma viral load. Using high-resolution evolutionary analyses, we found that tissue-based HIV continues to replicate, evolve, and migrate among tissues during cART. Furthermore, cancer onset and metastasis coincided with increased HIV expansion, suggesting a linked mechanism. HIV-expressing cells were associated with tissue macrophages, a target of HIV infection. Our results suggest the importance of tissues, and macrophages in particular, as a target for novel anti-HIV therapies.

FIG 1

Pairwise genetic distance in env (left) and nef (right) for cART-treated patients (in gray) and cART-naive patients (in white). Shown are the mean (thick black bar), interquartile range (white or gray box), and range (dotted lines). Outliers are shown as open circles.

FIG 2

Maximum-likelihood trees of the env (left) and nef (right) sequences for patient C02. Branches are drawn in substitutions/site according to the scale at the bottom. Circles are colored by the tissue of origin. Open circles, RNA; filled circles, DNA. Asterisks represent branches with support of >70.

FIG 3

Maximum-likelihood trees of the env (left) and nef (right) sequences for patient C04. Branches are drawn in substitutions/site according to the scale at the bottom. Circles are colored by the tissue of origin. Open circles, RNA; filled circles, DNA. Asterisks represent branches with support of >70.

FIG 4

Maximum-likelihood trees of the env (left) and nef (right) sequences for patient C05. Branches are drawn in substitutions/site according to the scale at the bottom. Circles are colored by the tissue of origin. Open circles, RNA; filled circles, DNA. Asterisks represent branches with support of >70.

FIG 5

BEAST estimated evolutionary rates. Mean (circles) and 95% confidence interval (vertical bars) evolutionary rates were estimated for the env and nef regions from the cART-infected and cART-naive patients in this study (C02 to KS3). Also shown are the rates reported in reference (env1); rates for progressors (env2) and nonprogressors (env3) from reference ; and pretherapy and posttherapy rates from reference . The y axis on log scale. Gray boxes indicate estimates from cART-treated patients.

FIG 6

RNAscope of cells expressing HIV gag-pol transcripts. Each punctum (pink dot) represents a single HIV gag-pol transcript in lymph node tumor (a and b) and cerebellum (c to f). Note that many puncta fuse together and appear as dense red staining of the entire cell. Cells were stained for markers of macrophages (CD68) and their activation/infiltration phenotype (CD163) (brown). Panel b is a ×630 magnification of the region outlined in panel a. Panels d and f are ×400 magnifications of the regions outlined in panels c and e, respectively. (Panels 6c and d are republished from reference with permission of the publisher.)