A novel CCR5 mutation common in sooty mangabeys reveals SIVsmm infection of CCR5-null natural hosts and efficient alternative coreceptor use in vivo

Abstract

In contrast to HIV infection in humans and SIV in macaques, SIV infection of natural hosts including sooty mangabeys (SM) is non-pathogenic despite robust virus replication. We identified a novel SM CCR5 allele containing a two base pair deletion (Δ2) encoding a truncated molecule that is not expressed on the cell surface and does not support SIV entry in vitro. The allele was present at a 26% frequency in a large SM colony, along with 3% for a CCR5Δ24 deletion allele that also abrogates surface expression. Overall, 8% of animals were homozygous for defective CCR5 alleles and 41% were heterozygous. The mutant allele was also present in wild SM in West Africa. CD8+ and CD4+ T cells displayed a gradient of CCR5 expression across genotype groups, which was highly significant for CD8+ cells. Remarkably, the prevalence of natural SIVsmm infection was not significantly different in animals lacking functional CCR5 compared to heterozygous and homozygous wild-type animals. Furthermore, animals lacking functional CCR5 had robust plasma viral loads, which were only modestly lower than wild-type animals. SIVsmm primary isolates infected both homozygous mutant and wild-type PBMC in a CCR5-independent manner in vitro, and Envs from both CCR5-null and wild-type infected animals used CXCR6, GPR15 and GPR1 in addition to CCR5 in transfected cells. These data clearly indicate that SIVsmm relies on CCR5-independent entry pathways in SM that are homozygous for defective CCR5 alleles and, while the extent of alternative coreceptor use in SM with CCR5 wild type alleles is uncertain, strongly suggest that SIVsmm tropism and host cell targeting in vivo is defined by the distribution and use of alternative entry pathways in addition to CCR5. SIVsmm entry through alternative pathways in vivo raises the possibility of novel CCR5-negative target cells that may be more expendable than CCR5+ cells and enable the virus to replicate efficiently without causing disease in the face of extremely restricted CCR5 expression seen in SM and several other natural host species.

Figure 1. Sequence alignment of wild-type and mutant CCR5 genes.

(A) Partial nucleotide sequence alignment and (B) predicted protein sequence alignment of the wild-type smCCR5 and smCCR5Δ2 alleles (GenBank HM246693 and HM246694), the CCR5Δ24 alleles previously described in sooty and red capped mangabeys (GenBank AF07473 and AAC62474), and human wild-type CCR5 and CCR5Δ32 alleles (GenBank DQ217934 and U66285). Residues identical to the wild-type smCCR5 molecule are indicated by ‘.’ and gaps are shown as dashes. Location of the frameshift resulting from the smCCR5Δ2 deletion in TM4 is indicated by an inverted black triangle while the huCCR5Δ32 frameshift is indicated by an inverted white triangle. Transmembrane domains are indicated by shaded boxes and mis-sense amino acid sequences resulting from mutations are highlighted.

Figure 2. Surface expression of wild-type and mutant CCR5 variants in vitro.

293T cells were transfected with expression plasmids encoding smCCR5 wild-type (orange), smCCR5Δ2 (light blue), smCCR5Δ24 (magenta) and huCCR5-wt (green), along with empty expression plasmid (dark blue). Untransfected 293T cells (shaded red) serve as controls. Cells were stained with an anti-CCR5 mAb (clone 3A9; APC-conjugated), which recognizes the N-terminal region of both human and sooty mangabey CCR5 (right panel), or an isotype-matched control antibody (left panel).

Figure 3. CCR5 surface expression on sooty mangabey CD4+ and CD8+ T cells ex vivo.

(A) CCR5 staining on CD8+ T cells (left) and CD4+ T cells (right) of uninfected sooty mangabeys was carried out between 2004 and 2009, and analyzed according to genotype groups (mean ± SEM): CCR5 wild-type (W/W; n = 38), heterozygous (W/Δ; n = 34) and homozygous mutant (Δ/Δ; n = 7). Because both Δ2 and Δ24 alleles are functionally null, they were combined for this analysis. CCR5 expression by CD8+ T cells differs significantly between genotype groups (p<0.0001 by Kruskal-Wallis test), while the trend for CD4+ T cells does not reach statistical significance (p = ns). (B) FACS plots showing CCR5 staining on CD4+ and CD8+ T cells from representative homozygous wild-type, heterozygous and homozygous mutant animals.

Figure 4. SIV plasma viral load measurements in infected sooty mangabeys between genotype groups.

Plasma viral load measurements (VL log₁₀; means ± SEM) collected in surveys carried out between 2004 and 2009 of infected animals in the wild-type (n = 60), heterozygous (n = 49), and homozygous mutant (n = 7) genotype groups. The Δ2 and Δ24 alleles were combined for this analysis. The difference in VL between homozygous wild-type and CCR5-null genotype groups (W/W vs. Δ/Δ) is statistically significant (p<0.05; Dunn's multiple comparison test) whereas the differences for the heterozygous group (W/Δ vs. W/W; W/Δ vs. Δ/Δ) does not reach statistical significance (p = ns; Dunn's multiple comparison test).

Figure 5. Mutant smCCR5Δ2 does not support SIV infection in vitro.

293T cells were transfected with CD4 alone (yellow bars) or in combination with wild-type smCCR5 (blue bars), smCCR5Δ2 (green bars) or wild-type human CCR5 (red bars). Target cells were infected with luciferase-expressing pseudotype virions carrying Env glycoproteins that were cloned from plasma of two SIVsmm-infected sooty mangabeys (FFv: W/W SM; FNp: Δ2/Δ2 SM). Pseudotypes carrying the R5-tropic HIV-1 Env JRFL and virions lacking envelope glycoproteins served as controls. Infection was measured by relative light units (RLU) in cell lysates 3 days after infection (mean ± SD).

Figure 6. Alternative coreceptor utilization by SIVsmm Envs in vitro.

293T cells were transfected with CD4 alone or in combination with (A) the alternative coreceptors GPR1, GPR15 and CXCR6, or (B) CXCR4, CCR2b or CCR5. Target cells were then infected with luciferase-expressing pseudotype virions containing SIVsmm Envs from infected CCR5 wild-type (FFv: W/W SM) and CCR5-null (FNp: Δ2/Δ2 SM) animals. Envs from SIVmac239, HIV-1 JRFL and virions lacking Env served as controls. Infection was assayed by RLU (mean ± SD) measured 3 days after infection.

Figure 7. Effect of CCR5 blocking on SIVsmm use of CCR5 and entry into primary SM PBMC.

(A) 293T cells were transfected with CD4 in combination with wild-type smCCR5 or GPR15. Two days post-transfection, target cells were pretreated for one hour with or without the CCR5 antagonist, maraviroc (15uM), and then infected with pseudotype virions carrying SIVsmm Envs from a CCR5 wild-type animal (FFv) and a CCR5-null animal (FNp). Three days later, infection was measured based on RLU (mean ± SD) in cell lysates. (B) Growth curves from infection of primary SM PBMC (FAk: W/W SM; FAz: Δ2/Δ2 SM). Cells were stimulated for 3 days with PHA, then pretreated for one hour with or without maraviroc (15 uM), followed by infection with two different SIVsmm primary isolates (M923 and M951) in the continued presence or absence of maraviroc. SIV Gag p27 antigen levels in viral supernatants were measured by ELISA.