Strong In Vivo Inhibition of HIV-1 Replication by Nullbasic, a Tat Mutant

HIV-1 infection is effectively controlled by antiviral therapy that inhibits virus replication and reduces viral loads below detectable levels in patients. However, therapy interruption leads to viral rebound due to latently infected cells, which serve as a source of continued viral infection. Interest in strategies leading to a functional cure for HIV-1 infection by long-term or permanent viral suppression is growing. Here, we show that a mutant form of the HIV-1 Tat protein, referred to as Nullbasic, inhibits HIV-1 transcription in infected CD4+ cells in vivo. Analysis shows that stable expression of Nullbasic in CD4+ cells could lead to durable anti-HIV-1 activity. Nullbasic, as a gene therapy candidate, could be a part of a functional-cure strategy to suppress HIV-1 transcription and replication.

HIV-1-positive (HIV-1 ϩ ) subjects, M10-treated T lymphocytes survived preferentially in vivo, suggesting that M10 protected cells in infected individuals (8)(9)(10). More recently, clinical trials have advanced a zinc finger nuclease that targets CCR5 and renders treated CD4 ϩ T cells resistant to HIV-1 infection (4). Other viral entry inhibitors have fused a portion of the gp41 heptad repeat 2 region, a 34-amino-acid peptide, to the amino terminus of CCR5 or CXCR4 and were shown to protect human primary T cells (11) or, in combination with an siRNA targeted to CCR5, were protective in pigtailed macaques against simian-human immunodeficiency virus (SHIV) challenge (12). Finally, other recent approaches have used CRISPR-Cas9 editing of the cellular gene encoding the CXCR5 HIV-1 coreceptor and chimeric antigen receptor (CAR) T cell therapy for HIV-1 using engineered cells that can resist HIV-1 infection and train the immune system to eliminate HIV-1-infected cells (13)(14)(15). Clearly, application of therapeutic gene transfer approaches to HIV-1 therapy is progressing.
Our group has investigated a Tat transdominant negative mutant referred to as Nullbasic. Nullbasic is a 101-amino-acid Tat mutant in which the basic domain region residues 49 to 57 are replaced with the amino acids GGGGGAGGG (16). Unlike most transgene inhibitors of HIV-1 that have a single mechanism of action, Nullbasic disrupts three independent steps of virus replication: viral transcription by RNA polymerase II, Rev-mediated transport of viral mRNA, and reverse transcription (RT) of viral genomic RNA into double-stranded DNA (16,17). Nullbasic inhibits HIV-1 transcription by competing with Tat for binding to the positive transcription elongation factor P-TEFb (17), and further analysis revealed that transcriptional "shutdown" by Nullbasic is associated with chromatin modification within the long terminal repeat (LTR) promoter characteristic of inactive heterochromatin (18). Nullbasic also inhibits Rev function through a direct interaction with DDX1 and results in a redistribution of Rev localization in cells from the nucleolus and nucleoplasm to the nucleoplasm and cytoplasm (19). In vitro experiments indicate that Nullbasic inhibits HIV reverse transcription by binding to the viral reverse transcriptase in the virion to destabilize the viral core structure (20).
To follow up on these interesting outcomes from studies performed in vitro, human CD4 ϩ cell protection by Nullbasic was tested in a small-animal model of HIV-1 infection. An immunodeficient mouse strain, NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJl (NSG) (23,24), was used in two ways. First, CD4 ϩ cells were transduced with Nullbasic, sorted, infected with HIV-1, and then engrafted into mice (preinfection treatment). In a second approach, CD4 ϩ cells were infected with HIV-1 first, and Nullbasic was then delivered into the infected cells with a retroviral vector (postinfection treatment), sorted, and engrafted into mice.
Our results showed that preinfection treatment with Nullbasic strongly inhibited HIV-1 replication in vivo in the CD4 ϩ cell-engrafted animals (CD4 ϩ mice), with no detectable HIV-1 RNA in animal blood and a 2-to 3-log reduction in HIV-1 mRNA levels in organs compared to HIV-1-infected CD4 ϩ and mock-treated mice, respectively. In animals engrafted with postinfection-treated CD4 ϩ cells, Nullbasic delayed HIV-1 replication in blood and decreased HIV-1 replication in the organs. The difference in HIV-1 responses to Nullbasic in the two models is interesting and indicates that Nullbasic is a more effective inhibitor in cells prior to the expression of viral proteins. and ZSG was used as a control. The experimental procedure is summarized in Fig. 1A. The NB-ZSG-and ZSG-positive CD4 ϩ cell populations were enriched by fluorescenceactivated cell sorter (FACS) analysis, and nontransduced cells were also collected, which were used as a control in these experiments. The transduced and nontransduced cells were infected with HIV-1. Next, the HIV-1-infected cells were sampled at 16 h postinfection (p.i.) for Alu PCR analysis and at 3, 7, and 10 days postinfection (dpi) to analyze the HIV-1 mRNA levels. The cell samples collected at 10 dpi were assayed for viability and NB-ZSG or ZSG expression.
Genomic DNA samples were extracted from the cells collected at 16 h p.i. and analyzed by using HIV-1 Alu PCR assays. As shown in Fig. 1B, the relative levels of the integrated HIV-1 proviral DNA in the infected NB-ZSG-treated CD4 ϩ cells (NB-ZSG cells), ZSG-treated CD4 ϩ cells (ZSG cells), and CD4 ϩ cells were not significantly different.
At 2 dpi, a majority of NB-ZSG, ZSG, and CD4 ϩ cells were engrafted into animals, and the remaining cells were grown in vitro. Total cellular RNA was extracted from cell samples collected at 3, 7, and 10 dpi, and the levels of viral mRNA, relative to the levels of cellular glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA in the same sample, were measured. As shown in Fig. 1C, CD4 ϩ cells expressing NB-ZSG had significantly reduced levels of HIV-1 mRNA at all time points compared to ZSG and nontransduced CD4 ϩ cells. Overall, cells expressing NB-ZSG had 26-to 60-fold-reduced levels of viral mRNA compared to cells expressing ZSG or nontransduced CD4 ϩ cells. Cells collected at 10 dpi were also analyzed for live-cell populations by using a fluorometric cell assay. As shown in Fig. 1D, about 70% of NB-ZSG cells were alive, which was significantly higher than in HIV-1-infected ZSG cells and nontransduced CD4 ϩ cells, where ϳ54% cells were alive. There was no statistical difference in live-cell percentages between HIV-1-infected NB-ZSG cells and any of the uninfected CD4 ϩ cell groups. At 10 dpi, ZSG fluorescence-positive cell percentages were Ͼ89.8% for NB-ZSG cells and Ͼ93.7% for ZSG cells (Fig. 1E). NB-ZSG mRNA levels were measured by using RT-quantitative PCR (qPCR) analysis with oligonucleotide primers specific for NB and ZSG sequences. As shown in Fig. S1 in the supplemental material, levels of NB-ZSG mRNA in HIV-1-infected and uninfected cells were similar when measured by using oligonucleotide primers specific for either NB or ZSG sequences. The ZSG mRNA levels in ZSG cells were ϳ2.5-fold higher than the NB-ZSG mRNA levels in cells measured by using ZSG oligonucleotide primers. The results here support and extend our previous results showing that NB-ZSG inhibits HIV-1 replication in primary CD4 ϩ cells (21,22).
Preinfection treatment of human CD4 ؉ cells with NB-ZSG strongly inhibits HIV-1 replication in transplanted mice. Approximately 4 million HIV-1-infected or uninfected cells were injected into each NSG mouse. Groups of 6 animals were engrafted with NB-ZSG, ZSG, or nontransduced HIV-1-infected CD4 ϩ cells, and groups of 4 animals were engrafted with the corresponding cells that were uninfected. Blood samples were collected weekly from each animal for analysis of cell engraftment and viral RNA in plasma. Animals were sacrificed after 31 dpi, and the liver, lung, spleen, and kidney were collected from each animal to measure levels of human CD4 ϩ cells in tissue and viral mRNA levels in cells.
The levels of human CD4 ϩ cells in animal blood were measured by staining cell samples with an anti-human CD4-allophycocyanin (APC) antibody followed by flow cytometry analysis. Human CD4 ϩ cells were detectable in all blood samples at 10 dpi, and peak levels were reached at 17 dpi, when Ͼ85% of the cells measured by flow cytometry were human CD4 ϩ cells ( Fig. 2A). At 24 dpi, CD4 ϩ cell percentages in blood samples from HIV-1-infected ZSG and CD4 ϩ cell-engrafted mice declined, while the percentage of CD4 ϩ cells in HIV-1-infected NB-ZSG mice was significantly higher than that in HIV-1-infected ZSG mice ( Fig. 2A). At 31 dpi, the HIV-1-infected NB-ZSG mice had a significantly higher CD4 ϩ cell percentage in their blood samples than did the infected ZSG and CD4 ϩ mice ( Fig. 2A).
The ZSG fluorescence of NB-ZSG and ZSG cells was monitored by flow cytometry. As shown in Fig. 2B, there was no significant difference in the mean percentages of ZSG-positive cells in blood samples from HIV-1-infected compared to uninfected ZSG  Nullbasic Strongly Inhibits HIV-1 Replication In Vivo ® mice or from HIV-1-infected compared to uninfected NB-ZSG mice. The mean percentage of ZSG-positive cells was Ͼ94% in the ZSG mice for all time points. In NB-ZSG mice, the mean percentage of ZSG-positive cells was Ͼ80% up to 24 dpi and then decreased to 56% at the last time point. This reason for the decreased NB-ZSG expression at 31 dpi compared to the earlier time point is unknown.
The mouse plasma samples were analyzed for levels of HIV-1 RNA, and the results are shown in Fig. 2C. HIV-1 RNA was not detected by RT-qPCR at the first time point in all animals. At 17 dpi, HIV-1 RNA was measurable in 4 of the 6 mice from the HIV-1-infected ZSG and CD4 ϩ groups. In HIV-1 RNA-positive mice, the mean RNA levels were 1.57 ϫ 10 4 Ϯ 0.93 ϫ 10 4 copies/ml and 1.79 ϫ 10 4 Ϯ 1.22 ϫ 10 4 copies/ml plasma for CD4 ϩ mice and ZSG mice, respectively. At 24 dpi, the HIV-1 RNA levels in plasma samples from HIV-1-infected ZSG and CD4 ϩ mice were 1.14 ϫ 10 5 Ϯ 1.28 ϫ 10 5 copies/ml and 1.28 ϫ 10 5 Ϯ 0.91 ϫ 10 5 copies/ml plasma, respectively. At 31 dpi, the levels of HIV-1 RNA in plasma samples from HIV-1-infected ZSG and CD4 ϩ mice were 2.31 ϫ 10 5 Ϯ 2.73 ϫ 10 5 copies/ml and 1.56 ϫ 10 5 Ϯ 1.37 ϫ 10 5 copies/ml, respectively. There was no HIV-1 RNA detected in the HIV-1-infected NB-ZSG mice at any time point, and as expected, no HIV-1 RNA was detected in plasma samples from all uninfected mice (Fig. S2). The results indicate that NB-ZSG expression in CD4 ϩ cells prior to HIV-1 infection resulted in undetectable HIV-1 RNA in animal plasma for up to 31 dpi.
Low levels of HIV-1 mRNA were measured in cells from organs of HIV-1infected NB-ZSG mice. All animals were sacrificed at 31 dpi, and samples of the liver, lung, spleen, and kidney were collected. Total cells from organ samples were strained from the tissue and then stained with an anti-human CD4-APC antibody to identify human CD4 ϩ cells (Fig. 3A). No significant difference in the percentages of human CD4 ϩ cells in liver and lung in all the animal groups was observed, although a trend toward fewer human CD4 ϩ cells in these tissues was observed in the HIV-1-infected CD4 ϩ mice (designated CD4 ϩ _H) (Fig. 3A) and ZSG mice (designated ZSG_H) (Fig. 3A). In spleen samples, the levels of CD4 ϩ cells in ZSG_H and CD4 ϩ _H mice were significantly lower than those in uninfected ZSG mice (designated ZSG_C) and CD4 ϩ mice (designated CD4 ϩ _C), whereas there was no significant difference between the NB-ZSG_H and NB-ZSG_C mouse groups. However, there were no significant differences between the NB-ZSG_H mouse group and the ZSG_H and CD4 ϩ _H mouse groups. In kidney, NB-ZSG_H mice had a significantly higher CD4 ϩ cell level than that in ZSG_H mice. The ZSG_H mice had a significantly lower level of CD4 ϩ cells than that in ZSG_C mice. There was no significant difference between the NB-ZSG_H and NB-ZSG_C, ZSG_C, and CD4 ϩ _C mouse groups. The kidneys of NB-ZSG_H mice showed a trend toward higher levels of human CD4 ϩ cells than in CD4 ϩ _H mice, but the difference was not statistically significant. As a whole, the data show that levels of CD4 ϩ cells were similar in NB-ZSG mice, irrespective of HIV-1 infection, suggesting that Nullbasic strongly protected CD4 ϩ cell viability, which was evident at 31 dpi.
Flow cytometry was used to monitor the expression of ZSG and NB-ZSG by fluorescence. As shown in Fig. 3B, the mean percentage of ZSG-positive cells in organs was Ͼ96.0% for all ZSG mice irrespective of HIV-1 infection. For NB-ZSG mice, the mean percentage of NB-ZSG cells ranged from 58% to 78% in the uninfected and infected NB-ZSG mice, which was not statistically different from the levels observed in blood (Fig. 2).
The HIV-1 mRNA levels in the organ samples were analyzed as previously described (18) and normalized to human GAPDH mRNA levels in the same sample. The results showed that NB-ZSG mice had greatly reduced levels of HIV-1 mRNA in all organ samples. As shown in Fig. 3C, the mean HIV-1 mRNA levels in the organs of HIV-1infected NB-ZSG mice were lower by ϳ600 to ϳ2,800-fold than those in HIV-1-infected ZSG mice and CD4 ϩ mice. The level of inhibition by NB-ZSG observed here was much higher than the level of inhibition observed in vitro (Fig. 1A). Finally, and as expected, no human GAPDH mRNA signal was detected in mouse organ samples from nonengrafted mice (Fig. S3A), and HIV-1 mRNA was not detected in any uninfected mouse (Fig. S3B).
In summary, NSG mice engrafted with human CD4 ϩ cells transduced with NB-ZSG prior to HIV-1 infection had undetectable viral RNA in plasma, whereas high levels of viral RNA were detected in the plasma samples of control mice. In NB-ZSG mice, the levels of HIV-1 mRNA in CD4 ϩ cells isolated from the organs showed significant reductions compared to control animals. or ZSG virus-like particles (VLPs). After 48 h, the NB-ZSG-and ZSG-positive cells were enriched by FACS selection, and cells were collected to measure integrated HIV-1 DNA levels in the cells with an Alu PCR assay. There was no significant difference in the levels of integrated HIV-1 DNA among the three cell groups (Fig. 4B). Overall, the level of integrated DNA measured in this postinfection treatment model was ϳ500-fold higher than that in the preinfection treatment model. We measured NB-ZSG and ZSG mRNA levels in the different cell groups at 7 dpi using oligonucleotide primers specific for either the NB-ZSG or ZSG sequence as described above. As shown in Fig. S4, the relative levels of NB-ZSG mRNA measured were similar irrespective of the primers used in the RT-qPCR assay. Also, the levels of NB-ZSG mRNA measured here were similar to those in cells in the preinfection treatment scenario (Fig. S1A). Animals were engrafted with the cells at 4 dpi (2 days after FACS analysis). An aliquot of cells was grown in vitro and assayed at 7 and 10 dpi for viral mRNA. As shown in Fig. 4C, the level of HIV-1 mRNA at 7 dpi, normalized to the GAPDH mRNA level in the same sample, in NB-ZSG-treated cells was reduced by ϳ24to 25-fold compared to those in ZSG-treated and untreated CD4 ϩ cells and was reduced by Ͻ2-fold at 10 dpi.
Cell samples collected at 10 dpi were also assayed for the percentages of live cells and ZSG fluorescence-positive cells. At 10 dpi, the live-cell percentage of control cells, at ϳ45%, was significantly lower than that of NB-ZSG-treated cells, at ϳ66% (Fig. 4D), which was similar to the live-cell percentage observed in uninfected CD4 ϩ cells (Fig. 1D). The percentages of ZSG-positive cells were ϳ92% for NB-ZSG-treated cells and ϳ95% for ZSG-treated cells. The data indicate that NB-ZSG delayed HIV-1 replication and improved levels of live CD4 ϩ cells at 10 dpi compared to control cells.
NB-ZSG delays HIV-1 replication in NSG mice in the postinfection treatment model. Two days after the FACS analysis, 4 million NB-ZSG-and ZSG-treated or untreated cells were injected into groups of 6 animals for engraftment. Blood samples were taken at 14, 21, 28, and 39 dpi. As shown in Fig. 5A, human CD4 ϩ cells were detected at 14 dpi in blood samples, which reached a peak level at 21 dpi and decreased thereafter. However, in NB-ZSG-treated mice, the CD4 ϩ cell percentage in blood samples was significantly higher than that in untreated CD4 ϩ mice at 21 dpi and was significantly higher than those in both ZSG-treated mice and untreated mice at 28 dpi. At 39 dpi, the percentage of human CD4 ϩ cells in blood was below 6.8% in all animals, and no significant difference in the percentages of CD4 ϩ cells in the blood was observed.
The ZSG fluorescence in NB-ZSG and ZSG CD4 ϩ cells was analyzed by cell flow cytometry. The mean percentages of ZSG fluorescence-positive cells in NB-ZSG-treated animals were above 80% at 14, 21, and 28 dpi and decreased to ϳ78% at 39 dpi. The percentages of ZSG fluorescence-positive cells in ZSG-treated animals were above 90% at all time points, as shown in Fig. 5B. Overall, a trend toward lower levels of NB-ZSG cells than of ZSG cells was noted in the animal groups, but this difference was not statistically significant.
In summary, NB-ZSG mice demonstrated a delay in HIV-1 replication and increased levels of circulating CD4 ϩ cells at 21 dpi compared to CD4 ϩ -only mice (nontreated CD4 ϩ cell-engrafted mice) and at 28 dpi compared to both ZSG and CD4 ϩ -only mice.
There was no significant difference in viral RNA levels between animal groups after 14 dpi.
Trend toward reduced viral mRNA by NB-ZSG in human CD4 ؉ cells in organs. The animals engrafted with untreated CD4 ϩ or NB-ZSG-or ZSG-treated CD4 ϩ cells were sacrificed at 39 dpi. Organs were harvested to measure the percentage of human CD4 ϩ cells residing in tissue and the level of HIV-1 mRNA in the samples.
As shown in Fig. 6A, the level of human CD4 ϩ cells in all organ samples was below 20%. The mean CD4 ϩ cell level in liver of the NB-ZSG mice, at ϳ11%, was significantly higher than those in ZSG mice, at ϳ5%, and CD4-only mice, at ϳ6%. However, no significant differences in CD4 ϩ cell levels were observed in lung, spleen, and kidney samples for the animal groups. The mean levels of ZSG-fluorescent CD4 ϩ cells (Fig. 6B) in the organ samples from NB-ZSG mice were ϳ72% in the liver, ϳ64% in the lung, ϳ73% in the spleen, and ϳ53% in the kidney, and slightly higher mean levels of ZSG-fluorescent CD4 ϩ cells were observed in organ samples from ZSG mice, at ϳ83% in liver, ϳ76% in lung, ϳ79% in spleen, and ϳ75% in the kidney.
HIV-1 mRNA levels in the cells from organ samples were measured by RT-qPCR and normalized to human GAPDH mRNA levels in the same sample. The relative mean levels of HIV-1 mRNA in cells from liver, lung, spleen, and kidney of NB-ZSG mice trended toward reduced levels compared to those in samples from ZSG mice and CD4 ϩ -only mice. Of these, only the lung samples from NB-ZSG mice had a significant reduction of HIV-1 mRNA compared to the samples from ZSG mice and CD4 ϩ mice. However, no other significant differences in viral mRNA levels among liver, spleen, and kidney were observed. Liver samples from NB-ZSG and ZSG HIV-1-infected mice were also analyzed by RT-qPCR, as described above, for the expression of NB-ZSG and ZSG, respectively. The mean level of NB-ZSG mRNA was ϳ1.5 ϫ 10 5 copies (per 10 6 copies of GAPDH mRNA) (Fig. S5), which is comparable to levels of NB-ZSG mRNA measured in CD4 ϩ cells at 7 dpi (Fig. S4). The mean level of ZSG mRNA in CD4 ϩ cells was ϳ2to 3-fold higher than that of NB-ZSG, which was also observed in the preinfection treatment model (Fig. S1). No ZSG or NB-ZSG mRNA was detected in samples from HIV-1-infected CD4 ϩ -only mice (Fig. S5).
Overall, the combined data show that in the postinfection treatment model, NB-ZSG delayed virus replication and improved CD4 ϩ cell levels in blood and liver of HIV-1infected mice, but it did not significantly reduce HIV-1 mRNA levels in cells or virus levels in the blood after 14 dpi.

DISCUSSION
The Tat mutant variant called Nullbasic is a unique anti-HIV-1 agent because it can inhibit three different steps of viral replication: synthesis of viral mRNA by RNA polymerase II, reverse transcription of viral RNA into DNA, and Rev-mediated transport of viral mRNA (16). Our previous studies also showed that NB-ZSG inhibited HIV-1 replication in primary human CD4 ϩ cells (21,22), with no effects on CD4 ϩ cell viability with respect to levels of apoptosis, cell metabolic activity by a 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assay, or cell proliferation in vitro (21). Given these positive results, we further investigated Nullbasic as a possible gene therapeutic agent in vivo using NSG mice engrafted with human CD4 ϩ cells.
Here, retroviral vector-mediated gene transfer of Nullbasic in human CD4 ϩ cells engrafted in a mouse model demonstrated strong inhibition of virus replication in vivo in a preinfection treatment model. In a challenging postinfection treatment model, Nullbasic delayed virus replication and improved CD4 ϩ cell levels. There are several differences in the two models used in this study that may explain the observed differences in the inhibition of HIV-1 replication. The preinfection treatment model represents a typical gene transfer scenario where uninfected cells are treated via transduction and then challenged by virus. In this case, the inhibitor was delivered prior to infection and had time to express sufficient levels of the antiviral protein required for effective virus control (16,17,19,20,22). This is important with respect to Nullbasic because it inhibits HIV-1 replication in a dose-dependent manner (16,18). For example, the inhibitory effect of Nullbasic on HIV-1 can be reversed by siRNAs that downregulate NB-ZSG expression in HIV-1-infected Jurkat cells (18). Our experience is that robust expression of NB-ZSG posttransduction requires at least 2 days. This is evident in CD4 ϩ cells that expressed high levels of NB-ZSG prior to HIV-1 infection and were highly resistant to HIV-1 replication, as indicated by strongly reduced levels of HIV-1 mRNA in the cells. However, the postinfection treatment model allowed for HIV-1 replication in an activated CD4 ϩ T cell population for 48 h prior to FACS selection of NB-ZSG-positive CD4 ϩ cells, where the transduction rate of retroviral vector gene transfer was ϳ30 to 40% in the HIV-1-infected CD4 ϩ cells. Here, a number of cells would support productive virus replication. Also, Tat is made by infected cells during this time, which can be secreted by HIV-1-infected cells (26)(27)(28). It has been estimated that about two-thirds of all cellular Tat produced may be secreted by primary human CD4 ϩ T cells in vitro (29), which can affect other cells (28,30,31). The Alu PCR results (Fig. 4B) also confirmed that the level of integrated HIV-1 DNA in the postinfection treatment cells was 3 logs higher than that in the preinfection treatment cells (Fig. 1B). Nevertheless, in this challenging postinfection treatment model, NB-ZSG reduced virus mRNA levels in cells by ϳ25-fold in vitro at 7 dpi, delayed a detectable viral load by 14 days in vivo, and imparted a trend toward improved CD4 ϩ cell levels and lower viral mRNA levels both in vitro and in vivo, although in vivo, this was statistically significant only in the liver for increased CD4 ϩ cell numbers and in the lung for reduced viral mRNA. We noted higher levels of human cell engraftment in preinfection-treated cells than in postinfection-treated cells, which was evident in both blood and tissue samples. Several factors could have affected the levels of human cell engraftment in mice. For example, in vitro, the postinfection-treated CD4 ϩ cell population had 12-to 15-fold higher levels of HIV-1 mRNA at 7 dpi and increased levels of dead cells at 10 dpi compared to the preinfection-treated CD4 ϩ cell population. Animals that were engrafted with these postinfection treated cell populations showed increased viral loads in plasma samples compared to preinfection-treated CD4 ϩ cells, especially before 28 dpi. In addition, subtle differences in the way in which cells were handled, such as when the cells underwent FACS analysis, could have an effect on human cell engraftment levels in mice.
In this study, we did not observe that NB-ZSG had a significant adverse effect on human CD4 ϩ cell engraftment of NSG mice, as human CD4 ϩ cell levels were not significantly different among the uninfected animal groups. In both models, CD4 ϩ cells were detectable in the blood sample between 7 and 10 days and reached a peak between 14 and 17 days after the engraftment. The levels of human CD4 ϩ cell populations in NB-ZSG mice were generally the same as or higher than those in infected control animals. This is in line with our previous results of cytotoxic assays in vitro where no obvious negative effect of Nullbasic on host cells was observed (21,22).
The retention and expression of the transferred gene were confirmed by the detection of ZSG fluorescence and NB-ZSG mRNA in CD4 ϩ cells. In both models, the ZSG-positive cell levels were generally over 90% for the ZSG vector throughout the experiment period but varied between 48.5% and 98.5% for the NB-ZSG vector in blood cells during the experiment. The level of NB-ZSG-positive cells decreased from Ͼ90% to between 55 and 80% in CD4 ϩ cells detected in organ tissue. We previously observed similar decreasing expression of NB-ZSG in CD4 ϩ cells in vitro over time (21,22). The percentage of ZSG-fluorescent cells in NB-ZSG mice decreased over time compared to ZSG mice. Nevertheless, in the preinfection treatment model, stable expression of NB-ZSG was observed in Ͼ50% of the engrafted NB-ZSG CD4 ϩ cells for up to 31 dpi. Why the percentage of NB-ZSG-positive CD4 ϩ cells trended lower over time is unknown, but we suspect that stable expression of NB-ZSG is regulated by a combination of factors that may include the nature of the retroviral integration site in cellular DNA, the activation state of the cell, and effects of NB-ZSG on cellular factors. It is intriguing that in the postinfection treatment model, where viral loads were high, NB-ZSG expression was maintained in ϳ80% of the CD4 ϩ cells over 39 dpi.
Tat is essential for HIV-1 replication because it stimulates transcription from the viral promoter by interaction with P-TEFb and then binding to the trans-activation response element (TAR) RNA stem-loop structure in the nascent RNA transcript (32). The role of Tat in HIV-1 transcription can be inhibited directly by didehydro-cortistatin A (dCA), which binds to the basic domain in Tat and inhibits its interaction with HIV-1 TAR RNA (33), whereas Nullbasic is believed to inhibit HIV-1 transcription by binding to P-TEFb (19,34). In vitro, dCA was reported to inhibit HIV-1 replication in acutely infected human peripheral blood mononuclear cells (PBMCs), with a maximum inhibition plateau of 86%, and it inhibited virus production by CD4 ϩ T cells isolated from HIV-1-infected patients by ϳ25% (35). However, dCA is a very strong inhibitor of HIV-1 reactivation from latency both in vitro (36) and in vivo (37). Interestingly, viral rebound of HIV-1 in latently infected cells is suppressed even after dCA treatment is discontinued, suggesting that dCA causes prolonged transcriptional silencing of the viral promoter. While Nullbasic has demonstrated strong inhibition of HIV-1 reactivation in cell line models of viral latency, further work is required to determine if Nullbasic has a similar capacity to inhibit HIV-1 reactivation in primary cell models of HIV-1 latency in vitro and in vivo. We are currently testing the delivery of Nullbasic protein to human PBMCs using nanoparticle technologies, which should help to address these questions.
The gene therapeutic approach to cure HIV-1 remains an important area of research