HIV-1 replication in the central nervous system occurs in two distinct cell types

Abstract

Human immunodeficiency virus type 1 (HIV-1) infection of the central nervous system (CNS) can lead to the development of HIV-1-associated dementia (HAD). We examined the virological characteristics of HIV-1 in the cerebrospinal fluid (CSF) of HAD subjects to explore the association between independent viral replication in the CNS and the development of overt dementia. We found that genetically compartmentalized CCR5-tropic (R5) T cell-tropic and macrophage-tropic HIV-1 populations were independently detected in the CSF of subjects diagnosed with HIV-1-associated dementia. Macrophage-tropic HIV-1 populations were genetically diverse, representing established CNS infections, while R5 T cell-tropic HIV-1 populations were clonally amplified and associated with pleocytosis. R5 T cell-tropic viruses required high levels of surface CD4 to enter cells, and their presence was correlated with rapid decay of virus in the CSF with therapy initiation (similar to virus in the blood that is replicating in activated T cells). Macrophage-tropic viruses could enter cells with low levels of CD4, and their presence was correlated with slow decay of virus in the CSF, demonstrating a separate long-lived cell as the source of the virus. These studies demonstrate two distinct virological states inferred from the CSF virus in subjects diagnosed with HAD. Finally, macrophage-tropic viruses were largely restricted to the CNS/CSF compartment and not the blood, and in one case we were able to identify the macrophage-tropic lineage as a minor variant nearly two years before its expansion in the CNS. These results suggest that HIV-1 variants in CSF can provide information about viral replication and evolution in the CNS, events that are likely to play an important role in HIV-associated neurocognitive disorders.

Figure 1. HIV-1 variants in the CSF of neurologically asymptomatic subjects can be R5 T cell-tropic.

(A) Maximum-likelihood phylogenetic trees. env sequences from the CSF are labeled with solid blue circles, and plasma sequences are labeled with solid red triangles. Bootstrap numbers ≥70 are indicated with an asterisk at the appropriate nodes, and bootstrap values are listed at critical nodes in the trees. The genetic distance scale bar indicates the number of nucleotide substitutions per site between env sequences. HIV-1 env sequences selected for phenotypic analyses are indicated with a black square, and X4-tropic envelopes are indicated with an underlined, superscript “X” following the envelope name; all other envelopes were R5-tropic. (B) Single-cycle infection of HIV-1 Env-pseudotyped reporter viruses on CD4^lowCCR5^high 293-Affinofile cells. Receptor expression was measured as follows: CD4^low = 1,214 receptors/cell, CD4^high = 97,003 receptors/cell, CCR5^high = 34,431 receptors/cell. Data are expressed as means of quadruplicate wells ± s.d., and results are representative of two independent experiments. (C) Single-cycle infection of HIV-1 Env-pseudotyped reporter viruses on primary human MDM. Data shown are means of duplicate wells ± s.d. for one donor, and were normalized to infection of the control macrophage-tropic HIV-1 Ba-L envelope.

Figure 2. Compartmentalized R5 T cell-tropic HIV-1 populations are found in the CSF of subjects diagnosed with HAD.

(A) Maximum-likelihood phylogenetic trees with compartmentalization and clonal amplification in the CSF. env sequences from the CSF are labeled with solid blue circles, and plasma sequences are labeled with solid red triangles. The phylogenetic tree characteristics are the same as those stated in Figure 1A. (B) Single-cycle infection of HIV-1 Env-pseudotyped reporter viruses on CD4^lowCCR5^high 293-Affinofile cells. Receptor expression was measured as follows: CD4^low = 1,214 receptors/cell, CD4^high = 97,003 receptors/cell, CCR5^high = 34,431 receptors/cell. Data are expressed as means of quadruplicate wells ± s.d., and results are representative of two independent experiments. (C) Single-cycle infection of HIV-1 Env-pseudotyped reporter viruses on primary human MDM. Data shown are means of duplicate wells ± s.d. for one donor, and were normalized to infection of the control macrophage-tropic HIV-1 Ba-L envelope.

Figure 3. Compartmentalized HIV-1 populations in the CSF are associated with HAD.

Maximum-likelihood phylogenetic trees are displayed for five HAD subjects with compartmentalization in the CSF. env sequences from the CSF are labeled with solid blue circles, and plasma sequences are labeled with solid red triangles. The phylogenetic tree characteristics are the same as those stated in Figure 1A.

Figure 4. Compartmentalized HIV-1 populations in the CSF of some HAD subjects are macrophage-tropic.

Phenotype data is displayed for HIV-1 env sequences from Figure 3. (A) Single-cycle infection of HIV-1 Env-pseudotyped reporter viruses on CD4^lowCCR5^high 293-Affinofile cells. Receptor expression was measured as follows: CD4^low = 1,214 receptors/cell, CD4^high = 97,003 receptors/cell, CCR5^high = 34,431 receptors/cell. Data are expressed as means of quadruplicate wells ± s.d., and results are representative of two independent experiments. (B) Single-cycle infection of HIV-1 Env-pseudotyped reporter viruses on primary human MDM. Data shown are means of duplicate wells ± s.d. for one donor, and were normalized to infection of the control macrophage-tropic HIV-1 Ba-L envelope.

Figure 5. Macrophage-tropic HIV-1 variants in the CSF/CNS can be detected prior to the development of severe dementia.

(A) Longitudinal phylogenetic analysis of env sequences from one subject with dementia. The genetic distance scale bar indicates the number of nucleotide substitutions per site between env sequences. HIV-1 env sequences selected for phenotypic analyses are indicated with a black square, and the node of divergence for the CSF M-tropic population is indicated with a blue open circle. Bootstrap numbers ≥70 are indicated with an asterisk at the appropriate nodes, and bootstrap values are listed at critical nodes in the tree. Months from HAD diagnosis correspond to the following sample dates: 7/8/2002 (−21), 12/3/2002 (−16), 4/8/2004 (0), 5/11/2004 (1), and 6/3/2004 (2). (B) Single-cycle infection of HIV-1 Env-pseudotyped reporter viruses on CD4^lowCCR5^high 293-Affinofile cells. Receptor expression was measured as follows: CD4^low = 1,214 receptors/cell, CD4^high = 97,003 receptors/cell, CCR5^high = 34,431 receptors/cell. Data are expressed as means of quadruplicate wells ± s.d., and results are representative of two independent experiments. (C) Single-cycle infection of HIV-1 Env-pseudotyped reporter viruses on primary human MDM. Data shown are means of duplicate wells ± s.d. for one donor, and were normalized to infection of the control macrophage-tropic HIV-1 Ba-L envelope.