Elimination of HIV Reservoirs Harboring Intact Proviruses

Abstract

Human immunodeficiency virus (HIV) establishes persistent infection by integrating its proviral DNA into the host genome. While most integrated proviruses are defective, a small portion of proviruses are intact that represent a major obstacle to achieving an HIV cure. We have designed an approach for selective elimination of host cells harboring replication-competent HIV (SECH), through inhibition of autophagy and anti-apoptotic molecules during viral reactivation. However, the effects of SECH on the dynamics of intact and defective HIV provirus remain unclear. Here we show that SECH treatments reduce reservoirs harboring intact but not defective HIV proviruses by intact proviral DNA assays. Nested PCR and DNA sequencing analyses confirm that SECH treatments can delete full-length HIV-1 proviruses in humanized mice in vivo, and in patient samples ex vivo. Our data suggest that the SECH method is capable of effectively eliminating HIV reservoirs harboring intact HIV proviruses that can restart active viral replication.

Lay summary:

Inhibition of pro-survival mechanisms helps to eliminate the cells containing infectious HIV to clear the viral infection.

INTRODUCTION

HIV establishes chronic infection by integrating its proviral DNA into the host genome, leading to persistent infections that are difficult to be cleared (1). While targeting different stages of the HIV life cycle with combination antiretroviral therapy (ART) can inhibit active viral replication, continuous ART is necessary to prevent HIV rebound (2). Additionally, host factors that affect cell survival and immune responses may also affect the effectiveness for ART in viral suppression (3–6). Elimination of reservoirs containing intact HIV proviruses that are capable of producing new viruses will be ideal for viral clearance. However, the resistance of HIV reservoirs to depletion remains the principal barrier to achieving an HIV cure.

It has been shown that only a small portion of HIV-1-infected cells (2–3%) contain replication-competent proviruses, while most of the integrated proviruses are defective (7). To eliminate HIV infection, it will be important to delete all reservoir cells harboring intact HIV proviruses that are replication-competent. We have therefore tested an approach for HIV clearance by selective elimination of host cells harboring replication-competent HIV (SECH), to sensitize HIV-infected cells to cell death by targeting anti-apoptotic molecules and autophagy during HIV reactivation (8–10). The SECH regimen include ABT-263 (11), an inhibitor for Bcl-2 and Bcl-xL, and SAR405 (12), an autophagy inhibitor, in combination with latency reversal agents (LRAs), administrated orally once every two days for 30 to 35 times (8). An attachment inhibitor, Fostemsavir (13), and an integrase inhibitor, raltegravir (14), were included as ART to prevent new rounds of viral infection (8). We found that the SECH approach can clear HIV-1 infection in blood samples from ART-naive or ART-experienced people living with HIV-1 (PLWH) ex vivo, and in 40–77% of HIV-1-infected mice reconstituted with human hematopoietic stem cells (Hu-HSC) in vivo (8, 10).

The dynamics of integrated HIV-1 DNA in the host provides valuable insights into the maintenance and turnover of HIV-1 proviruses (15–18). To improve the efficacy for HIV clearance, it is important to understand how SECH treatments affect the turnover of HIV proviruses in the in vitro and in vivo settings. The intact proviral DNA assay (IPDA) can quantify intact and defective viral DNA through duplexed droplet digital PCR (ddPCR), by detecting the Packaging Signal (Ψ) near the 5’ end of HIV-1 genome and the Rev Responsive Element (RRE) in the Envelope (env) gene (7). However, IPDA fails to detect some infectious proviruses, or distinguish proviruses possessing both Ψ and RRE sequences but containing mutations in other regions (19–21). We therefore utilized both IPDA and near full-length provirus sequencing to determine the effects of SECH treatments on the intact and defective HIV-1 proviruses.

METHODS

HIV-1 infection and cure studies using humanized mice.

NSG-SGM3 mice were reconstituted with CD34⁺ human stem cells from individual donors (AllCells) to generate human CD34⁺ cell-reconstituted (Hu-HSC) mice as described (8). Three months after reconstitution, Hu-HSC mice were infected with HIV-1 AD8 (NIH AIDS Reagent Program, 1000 pfu/mouse, i.p.). Ten days after infection, Hu-HSC mice were given daily suppressive ART treatments with raltegravir (20 mg/kg b.w., Adooq Bioscience), Fostemsavir (BMS-663068; 20 mg/kg b.w., Adooq Bioscience) and Lamivudine (25 mg/kg b.w., Macleods Pharma.) orally. After 40 days of suppressive ART, the treatments were changed to SECH. For SECH treatments, IDB (ingenol-3,20-dibeozoate, 2 mg/kg b.w., ENZO Life Sciences), a protein kinase c activator, and JQ1 (35 mg/kg b.w., MedChemExpress), a BRD4 inhibitor (22, 23) were used as LRAs. ABT-263 (11), an inhibitor for Bcl-2 and Bcl-xL (40 mg/kg b.w., MedChemExpress), an autophagy inhibitor SAR405 (12) (40 mg/kg b.w., MedChemExpress), together with an integrase inhibitor, raltegravir (14) (20 mg/kg b.w.), an attachment inhibitor, Fostemsavir (24) (20 mg/kg b.w.), were formulated in the solvent containing 10% ethanol, 30% polyethylene glycol 400 (Sigma), and 60% Phosal 50 PG (Fisher Scientific), and administered by oral gavage once every 2 days. Raltegravir and Fostemsavir (20 mg/kg b.w.) alone were also administered on the alternate days. For the ART control group, raltegravir and Fostemsavir (20 mg/kg b.w.) were given daily. Tablets with non-steroid anti-inflammatory carprofen (2 mg in 5 g tablet, Bio-Serv) were supplied with regular diet to the mice treated by SECH or ART control to prevent potential inflammation according to our previous procedure (8). After 35 cycles of treatments by SECH or ART control, mice were kept for 2 months with no treatments to determine virus rebound as described (8) (Figure 1A). Experiments were performed with the approval of the Institutional Animal Care and Use Committee.

Figure 1. Preferential deletion of intact but not defective HIV-1 proviruses by SECH in humanized mice.

(A) The scheme of SECH regimen. Hu-HSC mice were infected with HIV-1 AD8 three months after reconstitution with CD34⁺ human hematopoietic stem cells for HIV cure studies. At day 10 post infection, the mice were treated with ART for 40 days, followed by 35 cycles of treatments with SECH regimens or ART control.

(B) HIV-1 viral titer in the blood was measured during the 35 cycles of treatments by SECH (left panel) or ART control (right panel).

(C) Assessment of HIV-1 rebound in Hu-HSC mice after withdrawal of ART or SECH treatment. The rates of HIV-1 rebound were significantly different in mice treated by SECH or ART (P =0.0294).

Quantification of HIV-1 by RT-PCR.

HIV-1 mRNA in the blood or cultured cells was quantified by RT-PCR according to our established protocols (8). LTR-GAG was amplified with forward primer (LTR-GAG-AF), 5’-GATCTCTCGACGCAGGACTC-3’ and reverse primer (LTR-GAG-AR), 5’-CGCTTAAT ACCGACGCTCTC-3’, and detected with the LTR-GAG probe, 5HEX/CCAGTCGCC/ZEN/GCCCCTCGCCTC/3IABkFQ. HIV-1 pol-1 was amplified with POL-1 forward primer (POL-1-AF), 5’-AGCAGGAAGATGGCCAGTAA-3’ and reverse primer (POL-1-AR), 5’-GGATTGTAGGGAATGCCAAA-3’, and detected with the pol-1 probe FAM/CCCACCAAC/ZEN/ARGCRGCCTTAACYG/3IABKFQ in iTaq Universal Probes Supermix (Bio-Rad).

Treatment of PBMCs from the people living with HIV (PLWH)

Experiments using de-identified samples from PLWH were performed with the approval of the Institutional Review Board of the Houston Methodist Research Institute. The cells were cultured in RPMI complete medium containing and 5 ng/ml IL-2 and 50 ng/ml M-CSF to maintain survival of CD4⁺ T cells, monocytes and monocyte-derived macrophages. Cells were cultured with SECH regimens containing 25 nM IDB, 20 nM ABT-263, 0.1 μM SAR405 and 0.25 μM JQ1, together with 0.2 μM BMS-626529 and 0.2 μM raltegravir. Only BMS-626529 and raltegravir were added in the ART control group. BMS-626529 is the active form of Fostemsavir (24), which was with raltegravir in vitro to prevent new round of virus infection. The cells were cultured for 2 days as one cycle of treatments, washed and cultured in the same medium for next cycle of culture. After 6 cycles (12 days) of treatments, RNA was extracted for RT-PCR analyses of HIV-1 mRNA and genomic DNA was for amplification of near full-length HIV-1 proviruses.

Intact provirus DNA assay and near full-length PCR.

The spleen cells from humanized mice were treated with ammonium chloride buffer (0.15 M NH4Cl, 10 mM KHCO3, 0.1 mM EDTA) to remove red blood cells, followed by different analyses. The brain from humanized mice were minced and incubated with liberase (0.65 U/ml, 5401020001, Sigma), papain (10 U/ml, LS003118, Worthington Biochemical Co.) and DNAse I (20 U/ml, M0303L, New England Biolabs) at 37 °C for 30 min. Brain leukocytes were isolated by Percoll-gradient separation as described (25). Brain cells or splenocytes (> 10 × 10⁶ of cells) were homogenized in lysis buffer (10 mM Tris, pH 8.0, 100 mM EDTA, 50 mM NaCl, 0.5% SDS and 200 μg/ml proteinase K), followed by extraction of genomic DNA with phenol and chloroform. DNA with the DNA shearing index (DSI) less than 0.38 was used for intact provirus DNA assay (IPDA) as described (26). IPDA was performed with digital PCR (dPCR) on a Qiagen QIAcuity One 5Plex digital PCR instrument as described (26). QIAcuity Probe PCR kit and the following primers and modified probes of Kinloch et al. (19) were used: HIV-1 Ψ Forward Primer, CAGGACTCGGCTTGCTGAAG; HIV-1 Ψ Reverse Primer-GCACCCATCTCTCTCCTTCTAGC; HIV-1 Ψ Probe, FAM-TTTTGGCGTACTCACCAGT-ZEN/Iowa BkFQ; HIV-1 env Forward Primer, AGTGGTGCAGAGAGAAAAAAGAGC, HIV-1 env Reverse Primer, GTCTGGCCTGTACCGTCAGC; HIV-1 env Probe, HEX-CCTTGGGTTCTTGGGA-ZEN/Iowa BkFQ. Primers and probes were synthesized by Integrated DNA Technologies.

CD4⁺ T cells from the spleen and CD11b⁺ cells from brain homogenate of Hu-HSC mice were purified using CD4 and CD11b MicroBeads (Miltenyi Biotec), respectively. The cells (10⁵) were lysed in lysis buffer (10 mM Tris, pH 8.0, 100 mM EDTA, 50 mM NaCl, 0.5% SDS and 200 μg/ml proteinase K), followed by extraction of genomic DNA with phenol and chloroform. HIV-1 provirus was amplified by PCR using the following primes: forward-AAATCTCTAGCAGTGGCGCCCGAACAG; reverse-TGAGGGATCTCTAGTTACCAGAGTC for 9054 bp proviral sequence. Nested PCR was performed to amplify the 8898 bp proviral sequence using HIV-Nest-F-1, ACAGGGACTTGAAAGCGAAAG and HIV-Nest-R-1, CTAGTTACCAGAGTCACACAACAGACG. The largest PCR product was purified and sequenced by Nanopore DNA sequencing (Quintara Bioscience, TX). Point mutations were identified by DNA sequencing were introduced into HIV-1 plasmid (pNL(AD8)) by site-specific mutagenesis. Wild type and mutant HIV-1 plasmids were transfected in 293T cell by Lipofectamine 2000 to produce wild type and mutant HIV-1 viruses. Virus in the supernatants were tittered by virus outgrowth assay using TZM-bl cells as described (8).

Statistical analyses.

GraphPad Prism (version 10.2.1) was used for statistical analyses of biological assays. Data were presented as the mean ± SD. P values were determined by two-tailed Student’s t-test or one-way ANOVA with unpaired two-tailed t test. Significant statistical differences (P< 0.05) are indicated. Fisher’s exact test was used for analyzing the rate of HIV-1 rebound in humanized mice.

Data availability.

All data supporting the findings of this study are available within the article and are available from the corresponding author.

RESULTS

Preferential reduction of intact HIV-1 proviruses in Hu-HSC by SECH treatments.

The SECH strategy can sensitize host cells to cell death induced by cytopathic HIV-1 genes during viral reactivation (8). Because only a small fraction of integrated HIV-1 proviruses are intact with replication potential, we asked whether SECH would be more effective in targeting intact proviruses capable of expressing viral genes that are cytopathic to the host cells. After establishing latent HIV-1 infection in humanized mice (Hu-HSC) by suppressive ART (Figure 1A and Supplementary Figure 1A), we treated HIV-1-infected Hu-HSC mice with SECH regimens that include ABT-263 (5) an inhibitor for anti-apoptotic molecules, and SAR405 (12), an inhibitor for autophagy, together with ingenol-3,20-dibeozoate (IDB) and JQ1 as LRAs once every 2 days as one cycle of treatments (Figure 1A and 1B). Fostemsavir and raltegravir were given daily as ART. The ART control group received Fostemsavir and raltegravir only. Similar to previous observations (8), Hu-HSC mice did not display significant loss of body weight or inflammatory responses in different organs. Similar to previous observations (8, 27), HIV-1 became undetectable in the peripheral blood after SECH treatment for 35 cycles (Figure 1B). The treatments were then stopped for 8 weeks. We observed that 69% of mice did not develop viremia after withdrawal of SECH treatments (Figure 1C). In contrast, the rest of SECH-treated mice and all mice treated by ART control showed viral rebound (Figure 1C). These results suggest that SECH treatments could clear HIV-1 infection in a portion of humanized mice.

We next analyzed HIV-1 proviruses in humanized mice after SECH treatments by IPDA. In the spleen of Hu-HSC mice treated by ART control (ART-HIV⁺), we could detect intact Ψ⁺Env⁺ HIV-1-proviruses, as well as defective proviruses containing only the 5’-(Ψ⁺) or 3’-portion (Env⁺) of the provirus (Figures 2A–C). Similarly, both intact and defective proviruses were found in the spleen of SECH-treated mice with HIV rebound (SECH-HIV⁺; Figures 2A–C). Interestingly, in SECH-treated mice without HIV rebound (SECH-HIV⁻), Ψ⁺Env⁺ intact HIV-1 proviruses were significantly reduced, while 5’- or 3’-portions of defective proviruses were still abundant (Figures 2A–C). These data suggest that SECH treatments can effectively reduce the intact but not defective HIV-1 proviruses.

Figure 2. Preferential deletion of intact but not defective HIV-1 proviruses by SECH in humanized mice.

(A) IPDA to detect 5’Ψ and 3’Env in HIV-1 proviruses in the spleen of humanized mice. IPDA for splenic genomic DNA from Hu-HSC mice treated by ART control or SECH with (HIV⁺) or without (HIV⁻) HIV rebound. Ψ⁺Env⁺ proviruses are indicated by circles among different treatments.

(B) Percentages of intact Ψ⁺Env⁺ proviruses, and defective proviruses with 5’ Ψ⁺ only or 3’ Env only are quantified in spleen cells (a). *P<0.05. NS, not statistically significant.

(C) Quantification of intact Ψ⁺Env⁺ proviruses (a) in spleen cells. *P<0.05. NS, not statistically significant.

(D) IPDA to detect 35’Ψ and 3’Env in HIV-1 proviruses in the brain of humanized mice. IPDA for brain genomic DNA from Hu-HSC mice treated by ART control or SECH with (HIV⁺) or without (HIV⁻) HIV rebound. Ψ⁺Env⁺ proviruses are indicated by circles among different treatments.

(E) Percentages of intact Ψ⁺Env⁺ proviruses, and defective proviruses with 5’ Ψ⁺ only or 3’ Env only are quantified in brain cells (c). *P<0.05. NS, not statistically significant.

(F) Quantification of intact Ψ⁺Env⁺ proviruses (c) in brain cells. *P<0.05. NS, not statistically significant.

HIV RNAs and proteins, as well as intact and defective proviruses are detectable in the brain of PLWH (28, 29). We therefore determined the effects of SECH treatments on HIV proviruses in the brain of Hu-HSC mice. In Hu-HSC mice without HIV rebound (SECH-HIV⁻), we also observed decreases in Ψ⁺Env⁺ intact but not 5’- (Ψ⁺) or 3’-portion (Env⁺) of defective proviruses by IPDA (Figures 2D–F). In contrast, in Hu-HSC mice treated by ART or SECH with rebound (ART-HIV⁺, SECH-HIV⁺), the intact Ψ⁺Env⁺ HIV-1 proviruses were not decreased in the brain (Figures 2D–F). These data suggest that SECH treatments preferentially reduce intact but not defective HIV proviruses in the brain of Hu-HSC mice not showing viral rebound.

Elimination of full-length HIV-1 proviruses by SECH treatments.

Because IPDA measures whether HIV-1 proviruses contain the regions corresponding to the Ψ sequence in the 5’ region and a part of Env in the 3’ region, mutations in other parts of the HIV-1 genome cannot be detected. We therefore examined whether the residual “intact” Ψ⁺Env⁺ signals detected by IPDA in SECH-HIV⁻ mice actually contain deletions or other mutations. We performed nested PCR to amplify near full-length HIV-1 proviruses (Figure 3A). Interestingly, in Hu-HSC mice with HIV rebound after ART or SECH treatments (ART-HIV⁺, SECH-HIV⁺), full-length HIV-1 proviruses (FL-HIV) were detected in the CD4⁺ T cells from the spleen (Figure 3B and Supplementary Figure 2A, highlighted with a box). However, in CD4⁺ T cells from the mice without HIV rebound (SECH-HIV⁻), full-length HIV-1 proviruses were not detected (Figure 3B). These results suggest that SECH treatments can delete T cell reservoirs harboring intact but not defective HIV-1 proviruses in Hu-HSC mice showing no viral rebound.

Figure 3. Clearance of full-length HIV-1 proviruses by SECH in humanized mice.

(A) The scheme of nested PCR for amplification of near full-length HIV-1 proviruses.

(B, C) Amplification of near full-length HIV-1 (FL HIV) in genomic DNA of CD4⁺ T cells from the splenocytes (B) or CD11b⁺ myeloid cells from the brain (C) of Hu-HSC mice treated by SECH with (HIV⁺) or without (HIV⁻) HIV rebound or treated by ART control.

Clearance of full-length HIV-1 proviruses in myeloid cells by SECH treatments.

Tissue-resident macrophages are an important cell type among HIV reservoirs (30–34). HIV-1 infections have been found in brain macrophages of humanized mice (35). We observed the reconstitution of human CD11b⁺ macrophage in the brain of Hu-HSC mice (Supplementary Figure 1B). In the brain of SECH-treated mice without HIV rebound, IPDA showed that intact HIV-1 proviruses was reduced (Figure 2D–F). We further examined intact HIV-1 proviruses in brain macrophages by near full-length PCR analysis. In human CD11b⁺ macrophages from the brain of mice treated by ART (ART-HIV-1⁺) or by SECH with HIV rebound (SECH-HIV-1⁺), we detected full-length HIV-1 proviruses by nested PCR (Figure 3C). In contrast, full-length HIV-1 proviruses were absent in brain macrophages from SECH-treated mice without HIV rebound (SECH-HIV⁻; Figure 3C). These provirus analyses suggest that SECH treatments can clear intact HIV reservoirs in tissue-resident macrophages of the brain, and potentially other organs or tissues of a portion of humanized mice.

Deletion of intact HIV-1 proviruses in PBMCs from PLWH by SECH treatments.

We have observed that SECH treatments can effectively clear HIV reservoirs in PBMCs from PLWH ex vivo (8). We therefore determined whether SECH treatments are indeed effective at reducing intact HIV-1 proviruses in the samples from PLWH (Figure 4A). Similar to previous observations (8), clearance of HIV-1 in PBMCs from PLWH as measured by RT-PCR was observed by six cycles of SECH treatments in vitro, but not treatments by ATR or ART plus IDB (2 days/cycle) (Figure 4B and Supplementary Figure 3). We next performed nested PCR to determine whether SECH treatments eliminated intact HIV-1 proviruses. In PLWH samples treated by ART control, we detected both full-length and shorter HIV-1 DNA (Figure 4C). In PLWH samples after in vitro SECH treatments, full-length HIV-1 DNA was undetectable (Figure 4C). In contrast, short proviral DNA products were still present (Figure 4C). Consistent to the previous observations (8), the cell viability was similar in PBMCs treated by ART or SECH (Figure 4D). Interestingly, an HIV-1 patient after treatment by stem cell transplantation is free of infectious virus but still contains detectable HIV-1 DNA (36). Our observations are consistent with this study that depletion of replication-competent proviruses can lead to an HIV cure without the clearance of defective viruses.

Figure 4. Clearance of full-length HIV-1 proviruses by SECH in PBMCs from people living with HIV (PLWH).

(A) ART-experienced PLWH were used for SECH treatment. B, black. C, Caucasian. ND, not detectable.

(B) Quantification of HIV-41 mRNA by RT-PCR in PLWH (P1-P5) PBMCs treated by ART or SECH.

(C) Amplification of near full-length HIV-1 proviruses in CD4⁺ T cells from PLWH PBMCs treated by ART or SECH.

(D) The percentage of cell survival and loss of live cells after one cycle of treatments by SECH or ART in PBMCs.

HIV-1 proviruses with deletions are retained after SECH treatments.

Because IPDA measures whether HIV-1 proviruses containing the regions corresponding to the Ψ sequence in the 5’ region and a part of Env in the 3’ region, mutations in other parts of the virus cannot be detected. We therefore examined whether the residual “intact” Ψ⁺Env⁺ proviruses detected by IPDA in SECH-HIV⁻ mice contain deletions or other mutations by DNA sequencing. Interestingly, in Hu-HSC mice with HIV rebound after treatments by ART or SECH (ART-HIV⁺, SECH-HIV⁺), full-length HIV-1 proviruses (FL-HIV) were detected in spleen cells (Figure 5). We also purified the largest provirus PCR products amplified from the mice without HIV rebound (SECH-HIV⁻) for DNA sequencing. We found that HIV-1 proviruses in mice from SECH-HIV⁻ mice contained large internal deletions, which are expected to disrupt the expression of multiple HIV-1 gene products (Figure 5). Full-length HIV-1 proviruses (FL-HIV) from majority of ART and SECH-treated HIV-1⁺ mice were found to be intact by DNA sequencing (Figure 5). Some point-mutations, including S346F and I982I in HIV POL and R95K in HIV VPR, were found in individual mice (Figure 5). Virus outgrow assays showed that HIV-1 carrying these point mutations could generate infectious viruses similar to the wild type control (Supplementary Figure 2B). This indicates that these mutations do not affect the life cycle and infectivity of HIV-1. In summary, only defective HIV-1 proviruses containing large deletions are detected in Hu-HSC after successful SECH treatments.

Figure 5. Sequencing analysis for HIV-1 proviruses.

Nested PCR of near full-length HIV-1 was performed using DNA of CD4⁺ T cells from Hu-HSC mice with (HIV⁺) or with (HIV⁻) HIV rebound after treatments by SECH or ART. The largest PCR product was purified and sequenced by Nanopore DNA sequencing. Dashed line: deleted sequences. Red bars indicate point mutations.

Because the amplification of HIV proviruses by near full-length PCR uses the primers within or close to the LTRs, only internal deletions, but not deletions in either 5’- or 3’-regions of the HIV genome would be detected by this assay. HIV reactivation by LRAs in the SECH regimen leads to the expression of HIV gene products that can trigger cell death (Figure 6). It has been shown that multiple HIV-1 gene products can independently activate different cell death pathways in host cells (37–44). Therefore, point mutations or small deletions affecting one gene may not be sufficient to destroy the cytopathic effects of HIV. Proviruses with large deletions are expected to disrupt multiple viral genes and fail to cause host cell death (Figure 6). Together, our data suggest that SECH treatments are effective in clearing reservoir cells harboring intact but not defective HIV-1 proviruses.

Figure 6. A diagram illustrating the deletion of intact HIV-1 proviruses by SECH.

SECH regimen includes the use of inhibitors for pro-survival autophagy and anti-apoptotic molecules together with LRAs to sensitize HIV reservoirs to cell death. LRAs in the SECH regimen induce the expression of cytopathic HIV genes from intact proviruses to trigger cell death. In contrast, the remaining defective HIV proviruses are not expected to express HIV genes to cause host cell death.

DISCUSSION

The “shock-and-kill” strategy uses LRAs to induce the expression of cytopathic HIV genes to induce cell death in the host cells. We have tested the SECH approach to further sensitize cell death in HIV reservoirs by targeting pro-survival autophagy and anti-apoptotic molecules during viral reactivation. We show that SECH treatments can eliminate reservoir cells harboring intact but not highly defective HIV proviruses in PBMCs from PLWH. Although HIV rebound was tested after SECH withdrawal for only 8 week, intact HIV proviruses were absent in humanized mice with no viral rebound, whereas defective proviruses still remain. This is in contrast to SECH-treated humanized mice that experience HIV rebound after withdrawal of the treatments in which both intact and defective proviruses are detectable. These data suggest that the elimination of cells harboring intact HIV proviruses is critical for the successful prevention of viral rebound and clearance of HIV infection by SECH.

For the clearance of HIV infection, it will be desirable to eliminate all reservoirs that harbor intact HIV proviruses that can express the virus. While IPDA was not performed for all mice due to limitations of cell numbers, near full-length PCR for all mice revealed that SECH treatments could clear intact but not defective HIV-1 proviruses in humanized mice in vivo, and in PLWH samples ex vivo. Mechanistically, LRAs in the SECH regimen induce the expression of cytopathic HIV genes from intact proviruses to trigger cell death. In contrast, the remaining defective HIV proviruses are not expected to express HIV genes to cause host cell death (Figure 6).

It has been shown that multiple HIV-1 gene products can independently mediate apoptosis in host cells (37–40, 45). However, some defective HIV-1 proviruses can still express viral genes (46–48). Point mutations or small deletions are unlikely to completely abolish HIV-1 gene expression and disrupt the killing of host cells by SECH. Indeed, DNA sequencing revealed large deletions in the proviruses from successfully treated mice. None of the mutants are predicted to express major HIV genes that cause cell death. Because defective HIV proviruses are not able to produce infectious viruses, their presence would pose no risks for viral rebound. However, certain defective HIV may cause alternative forms of cell death or trigger inflammation (49). It remains to be determined whether the defective HIV-1 proviruses that are not cleared by SECH can cause adverse responses after treatments.

Targeting cell death pathways may change the susceptibility of HIV reservoirs to killing by cytotoxic T cells (CTLs). Interestingly, a Bcl-1 inhibitor, ABT-119, can sensitize HIV reservoirs to elimination by CTLs (50). It is likely the inhibitors for anti-apoptotic molecules and autophagy in the SECH regimen can promote the killing of HIV-infected cells by CTLs. Although the Hu-HSC mouse model used in this study is not suitable for studying human T cell responses, it will be interesting to examine the potential synergistic effects of CTLs and SECH for HIV clearance by developing novel humanized mouse models with better human T cell development.

Inducing the expression of cytopathic HIV genes by LRAs is critical for triggering cell death in HIV-1-infected cells. However, reactivation of HIV-1 also induces the expression of anti-apoptotic molecules and autophagy in HIV-infected cells (8). Mechanistically, the promoters for anti-apoptotic and autophagy genes share the binding sites for transcription factors, such as NF-κB, NFAT, AP-1 and HIF-1α that bind to HIV LTR to drive viral gene expression (9). It is thus necessary to inhibit the pro-survival mechanisms that can counteract the cytopathic effects of induced HIV genes to sensitize HIV reservoirs to cell death. Targeting specific molecular pathways in HIV reservoirs refractory to cell death will be critical to improve the efficacy of SECH treatments. Our study suggests that the SECH approach can be developed as a therapy to eliminate lymphoid and myeloid HIV reservoirs of different organs and tissues.