Inhibition of HIV-1 replication with stable RNAi-mediated knockdown of autophagy factors
© Eekels et al; licensee BioMed Central Ltd. 2012
Received: 10 January 2012
Accepted: 16 March 2012
Published: 16 March 2012
Skip to main content
© Eekels et al; licensee BioMed Central Ltd. 2012
Received: 10 January 2012
Accepted: 16 March 2012
Published: 16 March 2012
Autophagy is a cellular process leading to the degradation of cytoplasmic components such as organelles and intracellular pathogens. It has been shown that HIV-1 relies on several components of the autophagy pathway for its replication, but the virus also blocks late steps of autophagy to prevent its degradation. We generated stable knockdown T cell lines for 12 autophagy factors and analyzed the impact on HIV-1 replication. RNAi-mediated knockdown of 5 autophagy factors resulted in inhibition of HIV-1 replication. Autophagy analysis confirmed a specific defect in the autophagy pathway for 4 of these 5 factors. We also scored the impact on cell viability, but no gross effects were observed. Upon simultaneous knockdown of 2 autophagy factors (Atg16 and Atg5), an additive inhibitory effect was scored on HIV-1 replication. Stable knockdown of several autophagy factors inhibit HIV-1 replication without any apparent cytotoxicity. We therefore propose that targeting of the autophagy pathway can be a novel therapeutic approach against HIV-1
On the other hand, some viruses need autophagy to complete their replication cycle. Several positive-stranded RNA viruses such as poliovirus remodel intracellular membrane structures as scaffolds for their replication machinery . These membranous structures are thought to be autophagic vacuoles. For influenza A virus, two studies highlight two different aspects of the complex interaction between the invading virus and autophagy. One study reported that the intracellular concentration of autophagy marker protein LC3-II increased during influenza virus infection and pharmacological inhibition of autophagy reduced the viral titers, indicating that influenza requires autophagy . However, it has also been shown that influenza virus arrests autophagosome degradation, for which the viral M2 protein is solely responsible. This block of autophagy makes the infected cells more susceptible to apoptosis .
In case of human immunodeficiency virus type 1 (HIV-1), it is not clear to what extent autophagy influences the viral replication cycle, or whether the virus influences the autophagy pathway. It has been reported that the expression of HIV-1 Envelope protein (Env) on the surface of infected cells induces autophagy in bystander cells through gp41-mediated membrane fusion . The induction of autophagy subsequently leads to death of these uninfected cells [13, 14]. This mechanism has been used to explain the so-called "bystander-effect", which is the massive depletion of uninfected cells in HIV-1 infected individuals. Two studies also indicated that HIV-1 inhibits autophagy in the infected CD4+ T cell, shown by reduced expression of the two autophagy marker proteins LC3 and Beclin1 and analysis of infected cells by electron microscopy [15, 16]. Furthermore, the viral Nef protein prevents destruction of HIV-1 components in autolysosomes, thus blocking the antiviral role of autophagy in macrophages . Several autophagy factors were identified in a transient genome-wide RNAi screen for cellular co-factors of HIV-1 replication , suggesting that HIV-1 also needs autophagy or at least some of its components to complete the replication cycle. Indeed, stable knockdown of one of the identified autophagic co-factors, Atg16, resulted in long-term inhibition of HIV-1 replication . Also in primary cells it has been shown that several autophagy factors, e.g. Beclin1, Atg5 and Atg7 are required for HIV-1 replication [17, 20]. If HIV-1 indeed requires autophagy for its replication, inhibiting the pathway could be of therapeutic use. Since the virus relatively easily gains resistance against drugs targeting viral components, it has been hypothesized that targeting of cellular co-factors would make it more difficult for the virus to gain resistance [21, 22]. Therefore we sought to inhibit several autophagy factors (ATGs) via RNA interference (RNAi). Stable knockdown cell lines were generated, each cell line expressing a short hairpin RNA (shRNA) against a specific mRNA encoding an autophagy factor. Thirteen autophagy factors were tested, distributed along the autophagy pathway (Figure 1). Atg1/unc-51-like kinase (ULK1) is required for the initiation of autophagy. Autophagosome biogenesis is coordinated by the phosphatidylinositol 3-kinase (PIK3) complex containing the PIK3 p100 subunit (PIK3C3) and the PI3K p150 subunit (PIK3R4), which associate with Beclin1. WIPI1 regulates the transport of phosphatidylinositol-3-phosphate to the membranes. Two ubiquitin-like conjugation complexes are required; the first one forms a complex between Atg5 and Atg12, and Atg16 is non-covalently bound to this complex in a process that is catalyzed by Atg7 and Atg10. In the second conjugation system, LC3 is cleaved by Atg4 cystein proteases, essentially Atg4A and Atg4B, making it possible for Atg7 and Atg3 to generate the phosphatidylethanolamine (PE)-bound form of LC3: LC3-II.
We show that HIV-1 replication can be delayed in stable ATG knockdown cell lines. An additive inhibitory effect was observed when two ATGs were knocked down simultaneously, thus stressing the therapeutic potential of this strategy. Importantly, this HIV-1 replication delay was not accompanied by RNAi-induced cytotoxicity, suggesting that autophagy can be targeted in host cells without serious side effects.
For the selected shRNAs we first determined the ATG mRNA knockdown efficiency by RT-qPCR (Figure 3C). The results demonstrate a good knockdown (60-80%) for PIK3R4, Atg5, Atg10-3 and Atg16. A modest reduction in mRNA level (40-60%) was scored for Beclin1, Atg4A-1 and 3 and Atg10-5. For WIPI1 only a small reduction in mRNA expression (less than 20%) was measured. These results suggest that the knockdown level required for HIV-1 inhibition may vary per co-factor.
Autophagy activity upon ATG knockdown
Starvation-induced LC3 lipidation
To score for effects on cell proliferation, we performed the competitive cell growth or CCG assay . For that purpose the puromycin selection marker in the lentiviral constructs was replaced by the eGFP marker that allows the detection of transduced cells. Upon lentiviral-mediated transduction, the mixed culture of GFP+ and untransduced GFP- cells is followed longitudinally by FACS analysis. Based on the known doubling time of untransduced GFP-SupT1 cells, the decrease in GFP+/GFP- ratio over time was used to calculate the relative cell growth capacity, with the proliferation of control SHC1 cells set at 100% (Figure 3D). Knockdown of Atg5 has no effect on cell proliferation. A small cell growth reduction of less than 10% was observed for PIK3R4, Atg4A (both shRNAs) and Atg10 (both shRNAs). Three shRNAs (Beclin1, WIPI1 and Atg16) induce more significant cytotoxicity with a 10-20% reduced cell growth rate. Beclin1 and WIPI1 were abandoned after this point because the effect on HIV-1 replication may be caused by the cytotoxic effect of these shRNAs.
As an alternative means to inhibit the autophagy pathway, the autophagy inhibitor 3-methyladenine (3-MA) was used, which blocks the activity of the PIK3C3 and PIK3R4 kinases (Figure 1). First, 3-MA was tested in single cycle infection experiments on wild type SupT1 cells. Cells were pre-treated with 3-MA for 4 h, after which the drug was washed away before HIV-1 infection. Alternatively, cells were first infected with HIV-1 and then treated for 48 h. We also tested a combination of these treatments. All treated cell cultures were compared to the untreated control cells (Figure 4D). As reported earlier, we observed increased cell death in cultures that were treated with 3-MA for a prolonged period, including the samples that received 3-MA for 48 h post infection . Treating cells before infection did not cause significant changes in cell viability. The concentration of CA-p24 in the culture supernatant was dramatically reduced when cells were treated with 3-MA after infection (Figure 4D, right panel). However, more cells in the culture became CA-p24 positive and the mean production of CA-p24 per actively infected cell was slightly increased (Figure 4D, left and middle panel). Thus, cells treated with 3-MA post infection accumulate CA-p24 and yield a reduced CA-p24 concentration in the culture supernatant. These results indicate that a late replication step (from transcription to budding) is affected by 3-MA treatment after infection.
To test whether the simultaneous knockdown of two ATGs is tolerated by cells and whether the HIV-1 inhibition can be boosted, we generated double-knockdown cells expressing shRNAs against Atg16 and Atg5. Atg16 was chosen as its knockdown resulted in strong inhibition of HIV-1 replication with limited cytotoxicity . Atg5 knockdown resulted in good inhibition of HIV-1 replication without inducing cytotoxicity. Controls were untransduced SupT1 cells and the single-knockdown cell lines. The single knockdown cell lines were actually also transduced twice; the second transduction was performed with the SHC2 scrambled shRNA control. The first transduction used lentiviral vectors with the puromycin selection marker, whereas the second transduction was performed with vectors carrying a GFP-selection marker. To increase the change of scoring an additive effect, we purposely transduced cells at a relatively low multiplicity of infection (MOI) of 0.2. This will yield maximally 1 copy of each shRNA gene construct per doubly transduced cell, thus avoiding saturation of the RNAi mechanism.
Cells were incubated for 2 h in minimal medium to generate starvation signals in the absence and presence of protease inhibitors. The LC3-II level was detected by Western blot and normalized against the housekeeping protein GAPDH. Starvation-induced autophagy and autophagic flux were reduced in the single-knockdown cells and dramatically reduced in the double-knockdown cells (Figure 5E).
In this study we show that stable RNAi-mediated knockdown of autophagy factors leads to inhibition of HIV-1 replication. Multiple shRNAs against 13 autophagy factors (ATGs) were tested and inhibition of virus replication was scored for 7 ATGs. We confirmed knockdown of the targeted mRNAs and performed additional cell proliferation and autophagy tests, which allowed us to conclude that RNAi against PIK3R4, Atg4A, Atg5 and Atg16 results in HIV-1 inhibition due to a block of autophagy. For these 4 ATGs a specific reduction in the target mRNA expression level and a clear defect in the autophagy pathway were measured. For Atg10-1 and 3 no block in autophagy was observed, but a specific reduction in mRNA expression levels was measured and a minor effect on cell proliferation was observed. It is possible that the effect on HIV-1 replication is due to a cellular function of Atg10 other than its role in autophagy. Cell proliferation was not altered upon Atg5 knockdown, and only a small effect on cell growth was observed for PIK3R4 and Atg4A. Cell proliferation of Atg16 knockdown cells was 15% reduced compared to untransduced cells, but for this shRNA a specific reduction in Atg16 mRNA expression levels and a clear defect in autophagy were observed. For Beclin1 and WIPI1 a delay of HIV-1 replication was measured despite the absence of an effect on autophagy, and these shRNAs were therefore not selected for follow-up studies. The effect on HIV-1 replication could be due to the adverse effects on cell proliferation. An alternative explanation is that these specific shRNAs induce an Interferon alpha (IFNα) response that inhibits HIV-1 replication . An earlier genome-wide knockdown study has already identified ATG factors necessary for HIV-1 replication , although this was not confirmed in a second RNAi screen . However, none of the ATGs for which we scored inhibition of HIV-1 replication were identified in the knockdown study by Brass et al., except for Atg16. In primary cells it has been shown that knockdown of Beclin1 and Atg5  or Beclin1 and Atg7  results in inhibition of HIV-1 replication.
When expression of Atg5 and Atg16 was inhibited simultaneously, additive HIV-1 inhibition was measured. Knockdown of both ATGs did not have a greater impact on cell proliferation than that observed in the singly transduced cells. Atg16 knockdown seems to be the sole determinant of cell growth delay in the double knockdown cells, which may be surprising as both Atg5 and Atg16 are involved in the same step of the autophagy pathway (Figure 1). Such an additive effect was measured for autophagy activity, which indicates that the modest delay in cell growth measured for Atg16 knockdown is not directly related to the impact on the autophagy pathway.
When ATG knockdown cells were analyzed in single cycle infection experiments, we observed that less cells did produce intracellular CA-p24 and the concentration of CA-p24 in the culture supernatant was concomitantly reduced. This indicates that less cells became productively infected. The mean CA-p24 production per CA-p24 positive cells (mean fluorescence intensity or MFI) was not affected by ATG knockdown. Thus, fewer cells are productively infected, but these cells do synthesize as much CA-p24 as control cells. This result indicates that ATG knockdown leads to an early block of HIV-1 replication (e.g. entry or reverse transcription).
As an alternative to blocking autophagy with RNAi, we tested the inhibitor 3-methyladenine (3-MA). Surprisingly, we observed markedly different results in single cycle infection experiments with 3-MA compared to the shRNA-expressing cell lines. First of all, we observed that a higher percentage of the cells became positive for intracellular CA-p24, whereas a reduction was measured in the shRNA-expressing cell lines. As normally a certain percentage of HIV-1 infections become latent, this result could indicate that treatment with 3-MA results in less latently infected cells and more productively infected cells. A second observation was that cells treated with 3-MA after HIV-1 infection exhibited an increased MFI, meaning that cells produced more CA-p24 per cell than untreated cells. Less CA-p24 was measured in the culture supernatant. This is similar to what was described by Kyei et al., confirming their conclusion that 3-MA mediated blocking of autophagy leads to inhibition of virus budding into the culture supernatant. 3-MA has been used for several years as a specific inhibitor of autophagy, however, there is accumulating evidence that 3-MA can have pleiotropic effects, and the impact on autophagy should ideally be confirmed by alternative means such as shRNAs [27, 28].
Autophagy is a cellular pathway that is important in many viral infections, thus blocking the autophagy pathway could be of therapeutic value. In addition to viral infections, blocking autophagy has been proposed as a new therapeutic approach against cancer. Cancerous cells appear to exhibit increased autophagy activity that provides a survival mechanism when the cell is treated with chemotherapy . Blocking autophagy with 3-MA in combination with anti-cancer drugs has been used against several types of cancer, e.g. breast and colorectal cancer, and siRNAs to silence the ATGs Beclin1 and Atg5 have been tested against cervical cancer [30–32]. Therefore RNAi-mediated knockdown of autophagy factors could be of therapeutic value against viruses and other diseases. In this study, we used lentiviral vector-mediated delivery of shRNAs and this delivery system provides an attractive possibility to develop a durable therapy for HIV-1 patients. The ex vivo transduction of hematopoietic stem cells with lentiviral vectors expressing anti-viral and/or anti-co-factor shRNAs should guarantee the life-long generation of HIV-1 resistant cells, e.g. T cells and monocytes. This approach is the focus of further studies in our laboratory.
pLKO.1 DNA constructs expressing a specific shRNA were from the MISSION™ TRC-Hs 1.0 library . Constructs including the negative control constructs SHC001 and SHC002 (hereafter named SHC1 and SHC2) were obtained from Sigma-Aldrich as bacterial clones. Plasmid DNA was extracted using the Nucleobond Midiprep columns according to the manufacturer's instructions (Macherey-Nagel). Target sequences can be found on the website of Sigma-Aldrich [http://www.sigmaaldrich.com/life-science/functional-genomics-and-rnai/shrna/individual-genes.html]. The pLKO.1 constructs from the MISSION™ TRC-Hs 1.0 library contain a puromycin selection marker, which was replaced with the gene for enhanced eGFP (eGFP) as described earlier 
The T1249 peptide (WQEWEQKITALLEQAQIQQEKNEYELQKLDKWASLWEWF, Pepscan Therapeutics) was obtained as 10 000 × stock solution . The autophagy inhibitor 3-MA (Sigma-Aldrich) was diluted in 70% methanol and used at a final concentration of 10 mM. Protease inhibitors pepstatin A and E64d and the anti-LC3 antibody were purchased from Sigma-Aldrich. Raltegravir (MK-0518) was obtained from Bio-Connect services  and used at a final concentration of 1 nM. Lamivudine (3TC) was obtained from GlaxoWellcome and used at a final concentration of 33 pM.
The human embryonic kidney cell line HEK293T was grown in DMEM, supplemented with 10% FCS, 100 U/ml penicillin and 100 μg/ml streptomycin. The human T cell line SupT1 was cultured in Roswell Park Memorial Institute (RPMI) medium, supplemented with 10% FCS, 100 U/ml penicillin and 100 μg/ml streptomycin.
Lentiviral vectors were produced as described earlier . In short, HEK293T cells were co-transfected with pLKO.1-shRNA and the packaging plasmids (pVSV-G, pMDL and pRev-RRE) with Lipofectamin 2000 (Invitrogen). The medium was refreshed 1 day after transfection and the culture supernatant was harvested the next day. Aliquots of the culture supernatant with the lentiviral vectors were stored at -80°C. A sample was tested in CA-p24 ELISA.
SupT1 cells were seeded in 24-wells format at 1 × 105 cells/well and transduced with a fixed amount of lentiviral vector. Excess virus was washed away 1 day after transduction and selection of stably transduced cells was started by adding puromycin to the medium at a final concentration of 1 μg/ml. In the case of GFP-expressing cell lines, transduced cells were FACS sorted.
Culture supernatant was heat-inactivated at 56°C for 30 min in the presence of 0.05% Empigen-BB (Calbiochem). The CA-p24 concentration was determined by a twin-site ELISA with D7320 (Biochrom) as the capture antibody and the alkaline phosphatase-conjugated anti-CA-p24 monoclonal antibody EH12-AP (International Enzymes) as the detection antibody. Detection was performed with the Lumiphos plus system (Lumigen) in a LUMIstar Galaxy luminescence reader (BMG Labtechnologies). Recombinant CA-p24 produced in a baculovirus system was used as reference standard.
The HIV-1 molecular clone HIV-1LAI  was used to produce virus by transfection of HEK293T cells. HIV-1 production was measured by CA-p24 analysis in the culture supernatant. For HIV-1 replication studies, SupT1 cells were seeded in a 6-wells plate at 4 × 105 cells/well and infected with HIV-1 (0.2 ng CA-p24). HIV-1 replication was monitored by scoring for syncytia formation and longitudinal measurement of CA-p24 production in the culture supernatant.
For single cycle infection experiments SupT1 cells were incubated with HIV-1 for 4 hours. Excess virus was washed away and the cells were cultured in the presence of entry inhibitor T1249 (Pepscan) to block subsequent rounds of viral entry. Intracellular CA-p24 was analyzed by FACS and extracellular CA-p24 was measured by ELISA at 48 h post infection.
To assess the cytotoxicity induced by knockdown of autophagy factors, we used the competitive cell growth or CCG assay as described earlier . In brief, SupT1 cells were transduced with 0.1 or 1 μl lentiviral vector that expresses a shRNA and the GFP selection marker. The percentage of cells expressing GFP in the transduction mixture was analyzed longitudinally by FACS analysis. Twice weekly a sample of the culture was taken, cells were collected by centrifugation (4 min at 4,000 rpm, Eppendorf centrifuge) and resuspended in FACS solution (Phosphate buffered saline (PBS) + 2% FCS) and analyzed on FACScanto (BD Biosciences). The live cell population was determined with forward and side scatter. Fluorescence was normalized using unstained SupT1 cells. Based on the known doubling time of untransduced cells, the change in GFP+/GFP- ratio over the course of the experiment can be used to calculate the cell growth defect (%) of the GFP+ transduced cells .
For intracellular CA-p24 staining, cells were collected by centrifugation (4 min at 4,000 rpm, Eppendorf centrifuge) and fixed in 250 μl 4% formaldehyde for 5 min at room temperature. Cells were permeabilized with 500 μl BD Perm/Wash™ buffer (BD Pharmingen) and stained for at least 1 h at 4°C in 50 μl BD Perm/Wash™ buffer containing 5 μl 1:100 diluted antibody against CA-p24 conjugated with PE (monoclonal mouse, clone KC57, Coulter). Excess antibody was washed away by washing twice with 500 μl FACS solution, cells were resuspended in 250 μl FACS solution and analyzed on a FACScanto. Uninfected and unstained samples were used as negative controls.
mRNA knockdown levels for specific ATG targets were analyzed by RT-qPCR. RNA was isolated from 0.5 × 106 cells with the RNeasy kit (Qiagen) according to the manufacturer's protocol, including the DNase I treatment on the column. 1 μg RNA was used for reverse transcription (Thermoscript, Invitrogen) with Oligo-dT primers and cDNA synthesis was performed at 50°C. The cDNA was diluted 100 times and 5 μl of the diluted sample was used as template in a SYBR Green based RT-qPCR (SYBR Green FAST PCR, Qiagen) with an ABI Prism 7,000 detection system (Applied Biosciences). Primers for target genes and the internal control β-actin were from the Quantitect primer assays (Qiagen). The ΔΔCt method was used to calculate relative mRNA expression as described earlier .
Autophagy was induced by nutrient starvation (EBSS) for 2 h in presence or absence of the lysosome protease inhibitors E64d and pepstatin A (10 μg/ml each) to analyze the autophagy flux. To monitor the induction of autophagy, the relative amount of the PE-conjugated form (LC3-II) was determined by immunoblot analysis of whole-cell lysate using a rabbit polyclonal antibody against LC3. Cells were washed twice in PBS and lysed in buffer containing 50 mM Tris-HCl (pH 8), 1% Triton X-100, 100 mM NaCl, 1 mM MgCl2, 150 mM PMSF, and complete mini protease inhibitor cocktail (Roche Diagnostics). Cell lysates were electrophoresed in 12% SDS-PAGE and blotted to PVDF membranes. After a blocking step with PBS and 0.5% casein for 1 h at room temperature, blots were incubated overnight at 4°C with the anti-LC3 antibody in the blocking buffer. After 3 washes with PBS and 0.05% Tween, the blots were incubated for 1 h at room temperature with peroxidase-coupled antiserum diluted in blocking buffer. After further washes, the immune complexes were revealed by ECL (Millipore). The image capture was taken by the G:BOX camera system (Syngene) and intensity of the signals was analyzed with GeneTools software. The LC3-II signal was compared to that of the control housekeeping protein GAPDH.
We thank Stephan Heynen for performing CA-p24 ELISA experiments and Berend Hooibrink for FACS sorting. This research is financially supported by the Dutch AIDS fund, grant 2006006 and 2007028 and by SIDACTION. We also thank the Belgian Government for financial support through the Inter University Attraction Pole program, grant P6/41.