Chim3 confers survival advantage to CD4⁺ T cells upon HIV-1 infection by preventing HIV-1 DNA integration and HIV-1–induced G₂ cell-cycle delay

A central issue of anti-HIV gene therapy is the ability of the therapeutic gene(s) to confer genetic resistance against viral infection to target cells. This leads to diminished virally induced cytophatic effects (CPE) and eventually to the extension of cell survival. However, not all antiviral genes studied so far possess these positive features.¹  In this regard, anti–HIV-1 transgenes have been grouped into 3 classes depending on the phase of the viral life cycle targeted.¹  Class I genes affect the early steps of the HIV-1 lifespan, ultimately preventing integration of proviral DNA, thus keeping the cells free of new infection; class II genes block viral gene expression, thereby reducing the overall production of toxic viral proteins and virions from chronically infected cells; and class III genes affect the assembly and release of new infectious viral particles.¹  According to this classification, mathematic models supported by experimental data predict that class I and class II genes protect from viral CPE and induce prolongation of cell lifespan, whereas class III genes do not. On this basis, the mode of action of an antiviral gene has to be carefully analyzed to achieve a powerful and lifelong therapeutic advantage. In this context, we developed preclinical gene therapy technology based on lentiviral vectors (LVs) expressing a natural mutant of the viral infectivity factor (Vif), named F12-Vif.²  Vif is an HIV-1 key protein whose major function is to block the antiviral activity of the cellular restriction factor human apolipoprotein B mRNA-editing enzyme catalytic polypeptide (APOBEC)–3G (hA3G).^3,4  However, the role of Vif is not limited to counteracting hA3G, but rather has been proved increasingly complex over time. Although Vif was implicated in the reverse transcription process for a long time,^5,6  recently it has been demonstrated to directly interact with both HIV-1 reverse transcriptase (RT) and genomic RNA.^7–11  In addition, Vif plays a direct role along with Vpr, though independent of it, in inducing viral cytopathycity and cell-cycle delay in G₂ phase by recruiting Cul5-ligase complex and degrading a yet-to-be-identified cell factor.^12–14  Altogether, Vif is rightly considered an important target for antiviral therapy because it is involved in many key viral and cellular processes linked to HIV-1 pathogenesis. We showed that F12-Vif inhibits HIV-1 replication²  despite degrading hA3G as efficiently as HIV-1 Vif does. As a consequence of a partial characterization of its mechanism of action, F12-Vif was labeled as a class III antiviral transgene.¹  In contrast, the F12-Vif derivative Chim3 is as lethal against HIV-1 as F12-Vif is, but is unable to degrade hA3G, and acting as a dominant-negative factor affecting the retrotranscripts accumulation and HIV DNA integration, it has to be categorized as a class I antiviral transgene.

Once established that Chim3 is a potent intracellular inhibitor of HIV-1 replication operating at the preintegration step of the HIV-1 life cycle, the goal of this study was 2-fold. First, we wanted to verify whether HIV-1–infected Chim3–genetically modified cells prolonged their survival in culture compared with control cells. Second, we wished to analyze the possible contribution of Chim3 on the HIV-mediated deregulation of the cell cycle. To this end, we set up an experimental model in which Chim3-expressing T cells were mixed with control cells at a 1:1 ratio and then infected with HIV-1. The endpoint analysis was the survival and possible enrichment of Chim3-expressing cells. Furthermore, we examined whether Chim3 played any role in regulating cell cycle either by itself or after HIV-1 infection. Interestingly, we established that Chim3-LV induces survival advantage to both CEM A3.01 cells and primary CD4⁺ T lymphocytes, and prevents the HIV-1 Vif-mediated cell-cycle delay in G₂ phase. Taken together, these results suggest that a therapeutic approach based upon the engineering of hematopoietic stem cells (HSCs) with Chim3 can give rise to HIV-1–resistant CD4⁺ T lineages, which through expansion and a long lifespan, ultimately repopulate the immune system of persons living with HIV-1/AIDS.

The HEK-293T cells were propagated in Iscove modified Dulbecco medium (IMDM) supplemented with 10% fetal calf serum (FCS; EuroClone Ltd) and a combination of penicillin-streptomycin and glutamine (PSG). The CEM A3.01 T cell line¹⁵  and ACH-2 cells¹⁶  were cultivated in RPMI 1640 supplemented with 10% FCS and PSG.

Human CD4⁺ T lymphocytes were purified from umbilical cord blood (UCB), cultivated, expanded by CD3/CD28 beads, and transduced as previously described.¹⁷ 

For survival advantage experiments, cells were analyzed by the Cedex AS20 automated cell culture analyzer (Innovatis AG) before mixing. For primary CD4⁺ T lymphocytes, the experiments were performed only when the different cell types were comparable under cell viability, density, size, morphology, and aggregation rate parameters.

The LVs empty-PΔN (empty-LV) and Chim3-PΔN (Chim3-LV) were described in Vallanti et al²  and Porcellini et al,¹⁷  respectively. Chim3 encodes the 45–amino acid region corresponding to amino acids 126 to 170 of the natural mutant F12-Vif inserted in a NL4-3 HIV Vif backbone together with the constitutive PGK-ΔLNGFR selection marker cassette.¹⁸  The green fluorescent protein (GFP)–LV was obtained by substituting the ΔLNGFR with the GFP gene using the XbaI and SacII sites in the empty-LV. VSV-G–pseudotyped LV stock production, cell transduction, and immune selection were performed as previously described.¹⁷  GFP-expressing cells were selected by the BD FACSVantage SE Cell Sorter (BD Biosciences).

The percentage of LV transduction, measured as ΔLNGFR or GFP expression, was evaluated by flow cytometric analysis (FACSCalibur; BD Biosciences). The NGFR level was measured by using either the anti-NGFR–PE antibody (Ab; C40-1457; BD Pharmingen) or the anti-NGFR–biotinylated Ab plus the SAv-PerCP reagent (BD Pharmingen), the latter only for the detection of ΔLNGFR in the GFP-LV–Chim3-LV mixture. The level of CD4 expression was assessed by either the mouse anti–human CD4 clone SK3 (BD Pharmingen) or RPA-T4 antibodies (eBioscience). Fluorescence-activated cell sorter (FACS) results were analyzed by the FlowJo software (TreeStar). The intracellular level of p24Gag was measured using either the Kc57-FITC or -PE monoclonal Abs (Beckman Coulter) and the BD Cytofix/Cytoperm-Fixation/Permeabilization kit following manufacturer's instructions (BD Biosciences). The T-cell receptor (TCR) Vβ repertoire was determined using the IOTest Beta Mark kit following the manufacturer's instructions (Beckman Coulter).

For cell-cycle analysis, cells were first synchronized at the G₁/S border by 0.75μM aphidicolin (Sigma-Aldrich) treatment for 16 hours. The drug was removed by several washes with phosphate-buffered saline (PBS) and, after 10 hours of release, cells were either mock-infected or infected with the VSV-G–pseudotyped Vif-proficient and Vif- and Vpr-deficient Δenv-HIV-1 at a multiplicity of infection (MOI) of 3 in the presence of aphidicolin. The double-deficient HIV-1 molecular clone was obtained by removing the SpeI-SalI HXB2 fragment containing 2 premature stop codons in the Vif and Vpr genes, respectively. The 2 viruses were produced as described in detail in Porcellini et al.¹⁷  After 24 hours of infection, aphidicolin was removed, and after 16 hours of release, cells were permeabilized and stained with p24Gag-FITC Ab. Then, 1 × 10⁶ FITC-labeled cells/mL were resuspended in PBS containing 100 μg/mL propidium iodide, 0.1% saponin, 5mM EDTA, and 50 μg/mL RNase A for 30 minutes at room temperature. Single cells were analyzed using a double discrimination function; relative cell-cycle distribution was calculated using the FlowJo software, and fitting curves were calculated with the Dean-Jett-Fox model.

Cells were acutely infected at the indicated MOI with either the X4 NL4-3 or the R5 AD8 HIV-1 molecular clones. Single-round infection experiments were performed using GFP-LV– and Chim3-LV–transduced cells infected with VSV-G–pseudotyped Vif-Vpr–proficient and Vif-Vpr–deficient HIV-1. After overnight infection, the virus was removed by several washes with PBS and fresh medium was added. After 72 hours, cells were analyzed for GFP, ΔLNGFR, and p24Gag expression by FACS as described in “FACS and cell-cycle analysis.” RT assay was performed as described in Vallanti et al.² 

The LV copy number and HIV DNA was measured by TaqMan PCR using an ABI PRISM 7900 instrument and analyzed by SDS 2.3 software (Applied Biosystems) as in Porcellini et al.¹⁷ 

HEK-293T cells were cotransfected with the Cul5-HA, WT-Vif-Myc, and Chim3 vectors at the indicated DNA ratio and after 36 hours were lysed in 150mM NaCl, 50mM Tris-HCl (pH 7.5), 0.5% NP-40, and a cocktail of protease inhibitors. A total of 500 μg of cell lysates per sample were immunoprecipitated with 5 μg of mouse anti-HA antibody (clone HA-7; Sigma-Aldrich) coupled to 20 μL of protein A–conjugated RMP Protein A Sepharose Fast Flow (GE Healthcare Bio-Sciences AB) for 3 hours rotating at 4°C. The beads were washed 4 times with the washing buffer (150mM NaCl, 20mM Tris-HCl [pH 7.5], 0.5% NP-40). After elution with 2× Laemmli buffer and boiling, the coimmunoprecitated complexes were resolved into SDS-PAGE followed by immunoblot using rabbit anti-HA Ab (HA-7; Sigma-Aldrich) at a 1:10 000 dilution or HIV-1 HXB2 Vif rabbit antiserum, obtained through the AIDS Research and Reference Reagent Program (Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health) from Dr D. Gabuzda, at a 1:1000 dilution.

Genomic DNA (gDNA), isolated by the QIAamp Mini kit according to the manufacturer's instructions (QIAGEN), was resolved on 0.8% agarose gels after overnight digestion with AflII, blotted onto nylon membranes (Hybond-N; Amersham), hybridized to 1 × 10⁶ disintegrations per minute (dpm) mL ³²P-labeled 1-kb ΔLNGFR fragment in PerfectHyb PLUS hybridization buffer (Sigma-Aldrich), and visualized by autography.

Statistical comparisons were determined by the 2-tailed Student t test. The level of significance chosen was P values less than .05. The data are presented as means plus or minus SEM.

Recently, we reported that Chim3 protects primary CD4⁺ T cells and HSC-derived macrophages from HIV-1 infection by exploiting a dual mechanism acting at the preintegration level, ultimately resulting in a drastic reduction of HIV-1 DNA integration.¹⁷  These observations prompted us to believe that Chim3-LV should bestow survival advantage on transduced cells with respect to control cells after HIV-1 infection. To test this hypothesis, we used the nonpermissive CEM A3.01 cell line as working model because these cells are highly sensitive to the CPE of HIV-1 and support efficient HIV-1 production.¹⁵  Initially, we verified whether Chim3 protected CEM A3.01 cells from HIV-1 infection in a manner similar to that observed in primary cells.¹⁷  Cells were engineered with either the empty-LV or the Chim3-LV as detailed in Vallanti et al²  and Porcellini et al.¹⁷  Both LVs contain the PGK-ΔLNGFR selection marker cassette, which allows detection and purification of transduced cells by an anti-NGFR Ab.¹⁸  Genetically modified cells were purified (> 95% purity), infected with the X4 HIV-1 NL4-3 molecular clone at an MOI of 0.1, and followed over 1 month for HIV-1 production. In agreement with our previous findings,¹⁷  Chim3-transduced cells, unlike empty-transduced cells, were protected from HIV-1 infection (Figure 1A). In empty-LV cells, the cytopathic/lethal effect of HIV-1 was detectable at day 21, 5 days after the peak of viral infection (day 16), whereas the viability of infected Chim3-expressing cells remained comparable with that of the uninfected cells (> 70%) throughout the entire course of infection (Figure 1B). Moreover, unlike empty-LV, Chim3-LV–transduced cells did not show down-modulation of CD4 HIV-1 receptor (Figure 1C), which is a typical feature of mainly X4 HIV-1–infected cells.^19,20  This finding excludes CD4 down-modulation as a possible additional cause of Chim3-mediated antiviral activity. Finally, we verified whether Chim3-expressing cells resist HIV-1 infection because 1 or few HIV-1–resistant cell clones emerge. To this aim, we carried out Southern blot analysis of gDNA extracted from empty-LV– and Chim3-LV–transduced cells after 28 days of HIV-1 infection using a ΔLNGFR-specific probe. Empty-transduced CD4⁻ cells showed a typical clonal genotype, which likely derives from the growth of the few cells that survived to HIV-1–induced cell death (Figure 1D; left panel). In contrast as a smear gDNA of Chim3-expressing cells appeared hybridized, typical of a transduced bulk cell population (Figure 1D; right panel).

Altogether, these findings indicate that Chim3-LV⁺ CEM A3.01 cells are protected from HIV-1 infection, express the CD4 receptor, and do not show signs of viral cytophaticity and cell clonality.

On this basis, we assumed that Chim3-expressing cells should prevail over unmodified cells when both cell types were infected within the same cell culture. To test this hypothesis, CEM A3.01 mock-transduced cells were mixed at a 1:1 ratio with Chim3-LV–transduced cells. The mixture was either left uninfected (duplicate culture) or infected (6-replicate culture) with NL4-3 HIV-1 at an MOI of 0.01. We measured HIV-1 production by RT activity assay and ΔLNGFR expression, as an estimate of Chim3-LV cells enrichment, by FACS using an anti-NGFR Ab. The mix population was productively infected, reaching the peak of virus release after 25 days (RT activity = 8248 cpm/μL over the value of 481 cpm/μL measured at day 7 after infection). Therefore, Chim3-expressing cells were exposed for a long time to the cytophatic/lethal effect of the virus derived from mock-transduced infected cells. Of note, despite this challenging condition, ΔLNGFR staining at 49 days of culture indicated that Chim3-LV cells were highly enriched in the infected mixture compared with the uninfected one (Figure 2A; 88% vs 41%), demonstrating that under strong HIV-1 selective pressure only Chim3-LV–transduced cells survive for longer than 3 months (Figure 2B). In agreement with the results of Figure 1C, unlike mock-transduced cells, Chim3-transduced cells did not down-modulate CD4 receptor after HIV-1 challenge (Figure 2C).

Finally, we wondered whether the HIV-1–resistant CD4⁺ Chim3-LV–enriched population maintained the HIV-resistant phenotype for a long time. To address this question, an aliquot of both the sestuplicate 4-month-infected culture and the corresponding duplicate uninfected culture were challenged with NL4-3 HIV-1 at an MOI of 0.1. Of note, Chim3⁺ cells were still protected from HIV-1 infection to an extent similar to that of the original 95% or greater Chim3-expressing cells, which were included as positive control (Figure 2D). Taken together, these results demonstrate that Chim3 is not silenced over time, and its antiviral activity is therefore permanent.

To exclude the contribution of any functional cis-acting elements present in the LV backbone itself, we mixed CEM A3.01 mock–Chim3-LV and mock–empty-LV cells at a 1:1 ratio and processed them following the same experimental conditions of Figure 2A and B. As expected, Chim3⁺ cells lived on and expanded over mock-transduced cells after HIV-1 infection. In contrast, the relative percentage of infected empty-LV cells as well as both uninfected cell mixtures remained substantially unchanged throughout the duration of the experiment (Figure 3A). This result indicates that there is no involvement of the LV backbone in inducing or reinforcing the therapeutic potential of Chim3-LV.

Next, to establish the level of HIV-1 infection in CEM A3.01 mock-Chim3 compared with mock-empty mixture, we measured the HIV-1 production by RT activity assay on a per-cell basis (Figure 3B) along with the integration of HIV-1 DNA on separated mock-transduced and transduced cell populations (Figure 3C). It is interesting to note that, although the initial contribution of mock cells was equal in both mixtures (approximately 60%), the level of HIV-1 production was overall lower in mock–Chim3-LV than in mock–empty-LV mixture (Figure 3B). Next, we separated the transduced cells from the corresponding mock cells of both mixtures by ΔLNGFR immune selection, and measured the HIV-1 DNA content in each cell type before the peak of HIV-1 production. Mock-transduced (100% HIV-1 DNA) and empty-LV cells derived from the mock–empty-LV mixture contained nonstatistically significant different levels of HIV-1 DNA (Figure 3C white bars, mock vs empty). In contrast, the overall DNA content of mock-Chim3 mixture was lower compared with the mock-empty mixture (Figure 3C black vs white bars). Of most relevance, mock cells separated from the mock-Chim3 mixture contained 70% less HIV-1 DNA than mock cells separated from the mock-empty control mixture (P < .01), indicating that less infectious and less numerous viral particles are spreading in the mock-Chim3 culture than in the control mix culture (Figure 3C black vs white mock bars). Furthermore, in agreement with our previous studies,¹⁷  Chim3-expressing cells carried the lowest content of HIV-1 DNA among the 4 subpopulations (Figure 3C black Chim3 bar).

Taken together, these findings demonstrate that the remarkable reduction of HIV-1 production observed in the mock–Chim3-LV cell mixture mostly depends upon the reduction of HIV-1 DNA integration in Chim3-expressing cells.

We carried out survival advantage experiments on mock–Chim3-LV and mock–empty-LV CD4⁺ T lymphocytes mixed at a 1:1 ratio and then infected with HIV-1 AD8 at an MOI of 1 (Figure 4A-B). A month after productive infection, uninfected and infected cells were stained for CD4 and ΔLNGFR expression and analyzed by FACS (Figure 4A-B). The percentage of ΔLNGFR⁺ cells was overall unchanged in uninfected cells of both mixtures compared with the initial 50:50 ratio (Figure 4A-B), whereas it was greatly increased (85.7%) only in the infected mock-Chim3 mix (Figure 4A right panel). We verified the possible clonality of the Chim3-enriched population by evaluating the TCR Vβ repertoire clonogram of the uninfected and infected cells after 40 days of culture (supplemental Figure 1, available on the Blood Web site; see the Supplemental Materials link at the top of the online article). The expansion of the Vβ2 clone, which is in general the most represented subtype in normal T lymphocytes,²¹  was clearly observed in infected cells (supplemental Figure 1). This result suggests that, as opposed to CEM A3.01, primary T cells are prone to clonal expansion after HIV-1 infection.

Then, we wanted to also estimate the level of HIV-1 proviral integration in CD4⁺ T lymphocytes to sustain our hypothesis that Chim3-expressing cells survive longer because they contain less HIV-1 DNA (Figure 4C). To this aim, we transduced CD4⁺ T lymphocytes either with the Chim3-LV or with the GFP-LV, instead of the empty-LV, as control. After cell sorting by 95% or more ΔLNGFR and GFP positivity, transduced cells were mixed at a 1:1 ratio. To avoid cell clonality, we performed short-term experiments. We infected the cell mix in a single round with a VSV-G–pseudotyped Δenv-HIV-1 at an MOI of 3; after 72 hours, we analyzed the intracellular p24Gag expression as a surrogate marker of HIV-1 DNA integration, as shown in Porcellini et al.¹⁷  We observed that Chim3⁺-infected cells expressed approximately 65% less p24Gag than control cells do (Figure 4C), providing support for the fact that these cells, harboring less HIV-1 DNA, are supposed to live longer than control cells.

In addition to the well-recognized role of HIV Vpr on cell-cycle disruption,¹⁴  it has been recently reported that Vif itself also induces a delay in G₂, which is independent of the action of hA3G and, interestingly, is exerted through the recruitment of the same E3 ubiquitin ligase complex that is involved in hA3G degradation.^12–14  Based on this notion, we decided to verify whether, similar to Vif, Chim3 can deregulate the cell cycle. To this aim, we transduced HEK-293T cells with empty-LV, WT-Vif-LV, and Chim3-LV at equal copy numbers (5 copies per cell). Transduced cells were transfected with Tat and Rev plasmids to trigger transgene expression. We then assessed both the induction of the Vif proteins by Western blot (Figure 5B) and the possible effect of Chim3 on cell-cycle perturbation by incubating the cells with propidium iodide (PI) to visualize DNA content as indicator of cell distribution into the cell-cycle phases (Figure 5A). To determine the amplitude of the delay in G₂, the G₂/G₁ ratio was calculated only on live-gated cells. We confirmed the results of Wang et al¹²  on Vif and, of interest, we established that, in contrast, Chim3 has no effect, suggesting that it is likely defective in recruiting the Cul5-E3 ligase complex and thereby unable to affect the cell cycle. This hypothesis was supported by our finding showing that Chim3 binds Zn poorly,²²  which is required for Cul5/Vif interaction.²³ 

To confirm this idea biochemically, we carried out coimmunoprecipitation analysis between Cul5 and Vif proteins (Figure 5C-D). HEK-293T cells were transfected with a fixed amount of Cul5-HA expression vector and increasing amount of either WT-Vif-Myc (Figure 5C) or Chim3 vectors (Figure 5D). Quantification of the band intensity, normalized by the level of the input extract used (WCE; supplemental Figure 2A), indicated that Chim3 interacts with Cul5 from 3.5- to 9.3-fold less than WT-Vif (supplemental Figure 2B).

Based on this finding, we wondered whether Chim3 could also interfere with the Cul5/WT-Vif interaction and therefore prevent the recruitment of the E3-ligase complex to destroy the yet-to-be-identified protein. To answer this question, we transfected fixed amounts of Cul5-HA and WT-Vif-Myc and increasing amounts of Chim3 vectors, coimmunoprecipitated the protein complexes with anti-HA Abs, and finally revealed them with specific anti-Vif and anti-HA Abs (Figure 5E). Interestingly, quantification of the band intensities, normalized by the level of the input extract used (WCE; supplemental Figure 2C), established that a 4-fold increase of Chim3 produced a decrement of WT-Vif-Myc interaction to Cul5 of approximately 50% and an increment in Chim3 interaction of approximately 50% (supplemental Figure 2D; column 13 vs column 8). This result indicates that large amounts of Chim3 compete with WT-Vif for Cul5 interaction of at least 50% and suggests the formation of possibly nonfunctional heterodimers or Chim3 homodimers interacting with Cul5.

Next, we wished to investigate whether the weak Chim3-Cul5 interaction and the interference of Chim3 on Cul5-Vif binding translated into a direct effect on cell-cycle progression in CD4⁺ T cells. As we showed that transduced T cells, unlike transduced HEK-293T cells, accumulated Vif proteins even in the absence of Tat and Rev,¹⁷  we first verified whether Chim3 itself alters the cell cycle of uninfected CD4⁺ T cells. Similarly to transduced HEK-293T, the G₂/G₁ ratio was comparable in uninfected empty-LV– and Chim3-LV–transduced cells (data not shown), indicating that Chim3 itself also does not affect cell cycle in the absence of HIV-1 infection in T cells. Next, we single-round–infected empty-LV and Chim3-LV CEM A3.01 and CD4⁺ T lymphocytes with VSV-G–pseudotyped Vif-Vpr–proficient and Vif-Vpr–deficient (ΔvifΔvpr) HIV-1 molecular clones at an MOI of 3 and then analyzed the cell-cycle profile in both p24Gag⁺ and p24Gag⁻ live-gated cells. To mimic the effect of HIV-1 particles emerging from empty-LV– and Chim3-LV–transduced cells after the first round of infection, both viruses were produced from HEK-293T cells expressing either the empty-LV or Chim3-LV in equal copy numbers. After infection with Vif-Vpr–proficient HIV-1 produced from empty-LV–transduced HEK-293T cells, the G₂/G₁ ratio was highest in empty target cells (Figure 6A), and slightly lower in Chim3-LV target CEM A3.01 cells (Figure 6C). In contrast, after infection with Vif-Vpr–proficient HIV-1 generated from Chim3-LV producer cells, Chim3-target cells showed the lowest ratio among the 4 conditions (Figure 6D). Interestingly, when the cells were infected with Vif-Vpr–deficient HIV-1, the G₂/G₁ ratios were similar in all conditions (Figure 6E-H), indicating a dominant-negative function of Chim3. Similarly, the p24Gag⁻ cells showed a cell-cycle profile similar to that of Δvif-Δvpr HIV-1–infected cells (supplemental Figure 3). Comparable results were obtained also with purified CD4⁺ T lymphocytes (supplemental Figures 4-5). Taken together, these results not only provide support for the previous findings on the involvement of Vif in affecting the cell cycle,^12–14  but also indicate that Chim3 prevents the G₂ delay induced by the HIV-1 Vif protein when it is expressed either in target or producer cells. A cumulative effect is appreciated when Chim3-expressing target cells are infected with HIV-1 generated from Chim3-producer cells (Figure 6D).

This study was conducted to investigate whether Chim3 extends the lifespan of transduced cells after HIV-1 infection. Because of the fact that Chim3 belongs to group I anti–HIV-1 transgenes for its mechanism of action,^1,17  we expected it was endowed with the ability to confer survival advantage. Here, we provide evidence that the weak, long-term expression of Chim3 prolongs the existence of Chim3-LV–transduced cells compared with mock- and empty-LV–transduced CEM A3.01 cells and primary CD4⁺ T lymphocytes after HIV-1 infection. We observed that primary T lymphocytes are more prone to cell clonality compared with the CEM A3.01 cell line in comparable experimental settings. This likely derives from the different culture conditions used for the 2 cell types. Primary cells are subjected to multiple CD3/CD28 stimulations that, in addition to the further restriction induced by HIV-1 infection, might explain the restriction of the Vβ repertoire. In this context, the largest sample of LV integration site analysis in CD4⁺ T cells isolated from patients undergoing a phase 1/2 clinical trial with the VIRxSYS anti-HIV vector VRX496 yielded no LV integration close to either proto-oncogenes or tumor suppressor genes known to contribute to lymphoid-specific cancers. Moreover, the patients, followed for 4 years after infusion with VRX496-modified autologous CD4⁺ T cells, developed no abnormal cell expansions. This finding is consistent with the observation that no T-cell malignancies have ever been associated with HIV-1 integration in millions of persons with HIV-1.²³ 

Interestingly, Chim3 expression affects neither the survival nor the cell-cycle progression of uninfected cells, suggesting that the antiviral and surviving functions of Chim3 strictly depend on the presence of HIV-1 Vif, which is a typical dominant-negative feature. In contrast to WT-Vif, Chim3 does not induce a delay in G₂ phase after infection with a Vif-proficient HIV-1, but not with a Vif-Vpr–deficient HIV-1. It is well documented that Vpr-induced cell-cycle arrest in G₂ promotes HIV LTR transcriptional activity, viral production, and the onset of apoptosis in infected cells.¹⁴  Similarly, Vif has been lately implicated in the cell-cycle delay in G₂, independent of Vpr,^12,13  by recruiting the same Cul5-E3-ligase complex responsible for the degradation of hA3G and then targeting a yet-to-be-identified cellular protein, different from hA3G, to the proteasome.²⁴  In this context, we established that Chim3 alone interacts with Cul5 from 3.5- to 9-fold less efficiently than WT-Vif. This result gives credit to our previous data demonstrating that Chim3 binds less efficiently than WT-Vif to a Zn affinity column,²²  and to the findings of Giri et al²⁵  showing that the presence of Zn enhances the strength of the interaction of Cul5 with a Vif HCCH peptide by 8-fold. Thus it is conceivable that the F12-Vif–specific 5–amino acid mutations of Chim3 located in the Zn-finger motif hinders Zn binding and consequently Cul5 interaction. Most importantly, we have shown that an excess of Chim3 halves the WT-Vif-Cul5 interaction and, at the same time, increases the binding of Chim3-Cul5, suggesting the formation of Chim3 homodimers and/or WT-Vif-Chim3 heterodimers interacting with Cul5. This is supported by the notion that the molar ratio between an HCCH peptide and Cul5 is 2:1, indicating that Vif binds to Cul5 as homodimers.²⁵  We propose that the inhibitory function of Chim3 on the deregulation of the cell cycle might derive indirectly from the inhibition of HIV-1 DNA integration,¹⁷  which ultimately results in a reduced total amount of newly synthesized HIV-1 Vif, but also in a direct effect of Chim3 on preventing the degradation of the still-unknown cellular protein involved in cell-cycle perturbation. We show here that upon HIV-1 infection, the level of proviral DNA is decreased in Chim3-expressing cells even when these cells are embedded in an environment much more challenging than that represented by the pure (≥ 95% Chim3⁺) CD4⁺ T lymphocyte and macrophage populations.¹⁷  Indeed, Chim3-expressing cells directly in contact with nonprotected HIV-1⁺ cells represent an environment closer to the in vivo physiologic one, where transduced cells are expected to be, at least initially, a minority. We have also recently reported that Vif acts as an auxiliary factor for HIV-1 RT, increasing its rate of association to RNA or DNA templates.²²  Thus collectively, a growing body of evidence sustains the concept that Chim3 acts at multiple levels, some of which, such as the block of HIV-1 DNA integration, the lower binding to a Zn-affinity column, and the prevention of G₂ delay of the cell cycle, are all independent of the presence of hA3G.

Based on the results we recently published^17,22  and those presented in this study, we draw the following model of action of Chim3 (Figure 7). During the first round of HIV-1 infection, the percentage of cells delayed in G₂ phase (Figure 6) and the number of retrotranscripts¹⁷  are lower in Chim3 cells than in control cells. Consistently, viral particles produced during the first round are identical to the input virus if budded out from mock cells, but less infectious and in lower number if budded out from Chim3-expressing cells because Chim3 preserves the antiviral action of hA3G, which is, in turn, accumulated into virions.¹⁷  Thus, during the second round of infection, the virus emerging from mock cells shows the same infectivity on target cells as during the first round, whereas the virus budded out from Chim3⁺ cells, being less infectious for carrying hA3G, integrates with lower efficiency in both mock- and Chim3-LV–transduced cells. During the first round of infection, Chim3 plays a role at the level of target cells, whereas during the second round of infection, it plays a role at the level of both target and producer cells. The amplitude of the effect observed in single-cycle infection experiments, which is reiterated during a long-term infection, cannot be anything but small. However, a small effect can eventually give rise to an exponentially amplified outcome after several rounds of infection and cell doublings.

Altogether, our preclinical findings are encouraging and warrant the investigation of this approach in an appropriate animal model,²⁶  paving the way toward clinical testing. Potentially, HSCs transduced with Chim3-LV once reinfused into patients might give rise to CD4⁺ T lymphocytes, which not only resist HIV-1 infection, but also survive and prevail upon the untransduced HIV-1 target cells. In the long term, these genetically modified, functionally active cells might overcome chronically infected cells.

We thank Anna Stornaiuolo for helpful suggestion on plasmid vector construction and Ermanna Rovida (CNR–Istituto di Tecnologie Biomediche, Milano) for helpful discussion.

This work was supported in part by TaKaRa Bio (Otsu, Japan).

Contribution: S.P. performed most of the experiments of the research and contributed to the conception and design of the experiments and to the interpretation of the results; F.G. and L.A. performed most of the experiments of the research and contributed to the conception and design of the experiments and to the interpretation of the results; B.M.P. performed Southern blot analyses; G.-P.R. supervised the conception and design of the experiments, the interpretation of the results, and the writing of the manuscript; and C.B. conceived and designed the experiments, supervised the crude data, interpreted the results, and wrote the manuscript.

Conflict-of-interest disclosure: The authors are employees of MolMed SpA. C.B. is inventor of the patent Pub. No. WO/2006/111866.

Correspondence: Chiara Bovolenta, MolMed SpA, via Olgettina 58, Milano, 20132 Italy; e-mail: chiara.bovolenta@molmed.com.