Gamma interferon-mediated inflammation is associated with lack of protection from intravaginal simian immunodeficiency virus SIVmac239 challenge in simian-human immunodeficiency virus 89.6-immunized rhesus macaques

Abstract

Although gamma interferon (IFN-gamma) is a key mediator of antiviral defenses, it is also a mediator of inflammation. As inflammation can drive lentiviral replication, we sought to determine the relationship between IFN-gamma-related host immune responses and challenge virus replication in lymphoid tissues of simian-human immunodeficiency virus 89.6 (SHIV89.6)-vaccinated and unvaccinated rhesus macaques 6 months after challenge with simian immunodeficiency virus SIVmac239. Vaccinated-protected monkeys had low tissue viral RNA (vRNA) levels, vaccinated-unprotected animals had moderate tissue vRNA levels, and unvaccinated animals had high tissue vRNA levels. The long-term challenge outcome in vaccinated monkeys was correlated with the relative balance between SIV-specific IFN-gamma T-cell responses and nonspecific IFN-gamma-driven inflammation. Vaccinated-protected monkeys had slightly increased tissue IFN-gamma mRNA levels and a high frequency of IFN-gamma-secreting T cells responding to in vitro SIVgag peptide stimulation; thus, it is likely that they could develop effective anti-SIV cytotoxic T lymphocytes in vivo. In contrast, both high tissue IFN-gamma mRNA levels and strong in vitro SIV-specific IFN-gamma T-cell responses were detected in lymphoid tissues of vaccinated-unprotected monkeys. Unvaccinated monkeys had increased tissue IFN-gamma mRNA levels but weak in vitro anti-SIV IFN-gamma T-cell responses. In addition, in lymphoid tissues of vaccinated-unprotected and unvaccinated monkeys, the increased IFN-gamma mRNA levels were associated with increased Mig/CXCL9, IP-10/CXCL10, and CXCR3 mRNA levels, suggesting that increased Mig/CXCL9 and IP-10/CXCL10 expression resulted in recruitment of CXCR3(+) activated T cells. Thus, IFN-gamma-driven inflammation promotes SIV replication in vaccinated-unprotected and unvaccinated monkeys. Unlike all unvaccinated monkeys, most monkeys vaccinated with SHIV89.6 did not develop IFN-gamma-driven inflammation, but they did develop effective antiviral CD8(+)-T-cell responses.

FIG. 1.

vRNA levels in lymphoid tissues of vaccinated and unvaccinated rhesus macaques 6 months after intravaginal challenge with SIVmac239. Tissue vRNA levels are expressed as log₁₀ vRNA copies per microgram of total tissue RNA. Each symbol represents an individual animal. The horizontal line shows the mean vRNA copy number for each experimental group. Note that some vaccinated-protected monkeys had undetectable tissue vRNA in some tissue samples; however, vRNA was detected in at least two tissue samples from each animal.

FIG. 2.

Relationship among vRNA levels, plasma vRNA levels, and number of virus-positive cells in lymph nodes of vaccinated and unvaccinated rhesus macaques 6 months p.c. (A) Linear correlations between lymph node and plasma vRNA levels in vaccinated (vaccinated-protected, n = 9; vaccinated-unprotected, n = 5) and unvaccinated (n = 6) monkeys. Note that the monkeys with undetectable plasma vRNA levels were assigned plasma vRNA levels of 500 copies/ml, the detection limit of the assay. (B) Correlation between tissue vRNA levels and virus-positive cells determined by ISH (of the same monkey tissues graphed in panel A) is shown. Similar results were obtained in the spleen (data not shown).

FIG. 3.

CD28 expression on CD8⁺ T cells in genital (left) and peripheral (right) lymph nodes relative to tissue vRNA levels. Note that tissues of unvaccinated monkeys with the highest vRNA levels had lower frequencies of CD8⁺ CD28⁺ T cells. Green circles, vaccinated-protected monkeys; blue triangles, vaccinated-unprotected monkeys; red diamonds, unvaccinated monkeys.

FIG. 4.

Relationship between IFN-γ mRNA levels or SIVgag-specific IFN-γ T-cell responses and vRNA levels in tissues of vaccinated and unvaccinated monkeys 6 months p.c. (A) IFN-γ mRNA levels relative to vRNA levels and challenge outcome. (B) Frequency of SIVgag-specific IFN-γ-secreting cells relative to vRNA levels in individual monkeys with different challenge outcome. Individual vaccinated-protected (green circles), vaccinated-unprotected (blue triangles), and unvaccinated (red diamonds) monkeys are represented.

FIG. 5.

Relationship between Mig/CXCL9 and IP-10/CXCL10 mRNA and tissue vRNA levels in lymphoid tissues of vaccinated and unvaccinated monkeys 6 months p.c. The symbols are defined in the legend to Fig. 4.

FIG. 6.

Frequency and anatomic distribution of SIV⁺, Mig/CXCL9⁺, and IP-10/CXCL10⁺ cells in the spleens of vaccinated and vaccine-naïve monkeys. (A, D, G, and J) Spleen from a representative vaccinated-protected animal (no. 28229). (B, E, H, and K) Spleen from a representative vaccinated-unprotected monkey (no. 30445). (C, F, I, and L) Spleen from a representative unvaccinated monkey (no. 28433). (A to C) SIV RNA-positive cells (large arrows) detected by ISH. Note that in unvaccinated monkeys, but not in vaccinated monkeys, vRNA is detectable as a diffuse fog of silver grains over follicular dendritic cells (small arrows). (D to F) ISH for IP-10/CXCL10 mRNA-positive cells. (G to I) ISH for Mig/CXCL9-positive cells. Fewer Mig/CXCL9-positive cells (arrows) were detected in the tissues of vaccinated monkeys than in those of unvaccinated animals. (J to L) Mig/CXCL9 protein-positive cells. Mig/CXCL9 protein-positive cells were rarely detectable in vaccinated-protected monkeys (J). The distribution of Mig/CXCL9-positive cells in vaccinated-unprotected monkeys (K) was predominantly in the white pulp. In contrast, Mig/CXCL9 protein-positive cells in unvaccinated animals (L) were detected in both the white and the red pulp (RP). Note the 100-μm scale bar in panel J.

FIG. 7.

Schematic representation of the proposed relationship among virus replication, IFN-γ-driven inflammation, and IFN-γ-induced effective anti-SIV CD8⁺-T-cell responses. Infection of rhesus macaques with SIV (step 1) results in the activation of CD8⁺ T cells (step 2), which secrete IFN-γ. IFN-γ drives the expansion and terminal differentiation of effective anti-SIV CTL (step 3), which are important in the control of virus replication. However, IFN-γ secretion also results in the induction of the proinflammatory chemokines Mig/CXCL9 and IP-10/CXCL9 (step 4). These chemokines attract activated CXCR3⁺ T cells (step 5) to the area of virus replication. These CXCR3⁺ T cells include activated SIV-specific CD8⁺ T cells (step 6) that can promote further IFN-γ production (positive feedback loop) and contribute to effective anti-SIV immunity. However, CXCR3⁺ CD4⁺ T cells are also recruited, and these activated T helper cells are substrates for virus replication (step 7). This cycle of inflammation occurs in essentially all unvaccinated animals after SIVmac239 infection. In SHIV89.6-vaccinated monkeys, IFN-γ-driven inflammation was present only in the lymphoid tissues of animals that were not protected from SIV challenge. Thus, IFN-γ-driven inflammation can directly contribute to viral replication and to the observed inability of prior lentiviral infection to protect some individuals from subsequent challenge with a pathogenic virus.