Loss of In Vivo Replication Fitness of HIV-1 Variants Resistant to the Tat Inhibitor, dCA

Abstract

HIV resistance to the Tat inhibitor didehydro-cortistatin A (dCA) in vitro correlates with higher levels of Tat-independent viral transcription and a seeming inability to enter latency, which rendered resistant isolates more susceptible to CTL-mediated immune clearance. Here, we investigated the ability of dCA-resistant viruses to replicate in vivo using a humanized mouse model of HIV infection. Animals were infected with WT or two dCA-resistant HIV-1 isolates in the absence of dCA and followed for 5 weeks. dCA-resistant viruses exhibited lower replication rates compared to WT. Viral replication was suppressed early after infection, with viral emergence at later time points. Multiplex analysis of cytokine and chemokines from plasma samples early after infection revealed no differences in expression levels between groups, suggesting that dCA-resistance viruses did not elicit potent innate immune responses capable of blocking the establishment of infection. Viral single genome sequencing results from plasma samples collected at euthanasia revealed that at least half of the total number of mutations in the LTR region of the HIV genome considered essential for dCA evasion reverted to WT. These results suggest that dCA-resistant viruses identified in vitro suffer a fitness cost in vivo, with mutations in LTR and Nef pressured to revert to wild type.

Keywords: HIV-1; Tat inhibitor; dCA; latency promoting agent; resistance; transcription.

Figure 1

Schematic representation of HIV NL4.3 and the MUT1 and MUT 2 mutations acquired to become resistant to dCA in vitro. (A) Mutations found across the genome. (B) Details of the mutations present in the HIV-1 5′ LTR. In green, mutations found in MUT1 dCA-resistant HIV-1 variant; in blue, mutations found in MUT2 dCA-resistant HIV-1 variant and in red, mutations found in both MUT1 and MUT2 dCA-resistant HIV-1 variants. The green rectangle depicts a 49-nucleotide insertion found in a MUT1 variant. Amino acid changes are indicated in parenthesis. Previously published in Mousseau et al., mBio 2019 [5].

Figure 2

Longitudinal analysis of plasma viral load and cell-associated HIV-1 DNA. (A) Plasma HIV-1 viral load was monitored longitudinally by real-time PCR. (B) Levels of cell-associated HIV-1 DNA were monitored longitudinally by real-time PCR. (C) Kaplan–Meier curve depicts the percent of animals with undetectable plasma viral load following exposure. Data in (A,B) are expressed as mean± SEM. Statistical significance was determined using a two-sided Kruskal–Wallis test in (A,B), and a Mantel–Cox test in (C). p values between WT and MUT1, WT and MUT2 in (C) are <0.0001 and 0.0033, respectively.

Figure 2

Longitudinal analysis of plasma viral load and cell-associated HIV-1 DNA. (A) Plasma HIV-1 viral load was monitored longitudinally by real-time PCR. (B) Levels of cell-associated HIV-1 DNA were monitored longitudinally by real-time PCR. (C) Kaplan–Meier curve depicts the percent of animals with undetectable plasma viral load following exposure. Data in (A,B) are expressed as mean± SEM. Statistical significance was determined using a two-sided Kruskal–Wallis test in (A,B), and a Mantel–Cox test in (C). p values between WT and MUT1, WT and MUT2 in (C) are <0.0001 and 0.0033, respectively.

Figure 3

HIV-1 plasma viral load monitored longitudinally by qPCR.

Figure 4

Analysis of human T cell levels in peripheral blood of HIV-1 infected humanized mice. Frequency of human CD4+ (A) and CD8+ (B) T cells in the peripheral blood was monitored longitudinally by flow cytometry. Data are expressed as mean ± SEM. Statistical significance was calculated using a two-sided Kruskal–Wallis test.

Figure 5

Frequency of activated T cells in peripheral blood of HIV-1 infected humanized mice. Activated CD4+ (A) and CD8+ (B) T cells in peripheral blood were identified by flow cytometry using anti-human CD38 and HLA-DR antibodies. Shown is a longitudinal analysis following HIV-1 exposure. Data are expressed as mean ± SEM. Statistical significance was calculated using a two-sided Kruskal–Wallis test.

Figure 6

Longitudinal analysis of cell associated HIV-1 RNA and proviral DNA in tissues from HIV-1 infected humanized mice. Levels of cell-associated RNA (A) and proviral DNA (B) per 10⁵ human CD4+ T cells in tissues of HIV-1-infected humanized mice were determined at necropsy. (C,D) depict the levels of cell-associated RNA and proviral DNA for all tissues combined, respectively. Data are expressed as mean ± SEM. Statistical significance was calculated using a two-sided Kruskal–Wallis test.

Figure 7

Frequency of CD4+ or CD8+ T cells in tissues of HIV-1-infected humanized mice. Percentage of human CD4+ (A) and CD8+ (B) T cells in tissues of HIV-1-infected humanized mice was determined by flow cytometry. Data are expressed as mean ± SEM. Statistical significance was calculated using a two-sided Kruskal–Wallis test.

Figure 8

Frequency of activated CD4+ and CD8+ T cells in tissues of HIV-1-infected humanized mice. Activated human CD4+ (A) and CD8+ (B) T cells in tissues of HIV-1-infected humanized mice were identified by the co-expression of hCD38 and HLA-DR. Data are expressed as mean ± SEM. Statistical significance was calculated using a two-sided Kruskal–Wallis test.

Figure 9

Levels of human IFNγ and IFNα in the plasma of HIV-1 infected humanized mice. Levels of plasma IFNγ (A) and plasma IFNα2 (B) in HIV-1-infected humanized mice at weeks 1, 2, and 3 post exposure were determined using a bead array kit. Data are expressed as mean ± SEM. Statistical significance was calculated using a two-sided Kruskal–Wallis test.