In this study, we have shown that the expression and anti-HIV function of hA3G and hA3F vary with naïve CD4+ T helper cell differentiation to Th1 and Th2 effector cells. Cytokine polarization of naïve cells into Th1 and Th2 effectors had similar effects to transduction of naïve cells with Tbet or GATA3. In both cases, decreased expression of hA3G and hA3F was seen in Th2 cells relative to Th1 cells. These complementary methods demonstrate that the differences observed in the Tbet and GATA3 transduced cells were due to transcription regulated by those factors and not an artifact of over-expression. Such an opposing effect of differentiation on expression of hA3G and hA3F is consistent with earlier findings of opposing effects on the expression of several other genes in these two T helper subtypes 
. This hA3G and hA3F expression difference between Th1 and Th2 cells affected wild type, as well as vif
-deleted, HIV-1 infectivity.
Expression of Tbet in naïve helper cells has been shown to lead to production of IFN-γ[Σζαβo, 2000 #35]. In turn, that IFN-γ can act in an autocrine manner on Th1 cells 
. Extracellular neutralization of IFN-γ secreted by Tbet-transduced and control-transduced cells blocked basal and TCR-stimulated hA3G and hA3F expression. This is consistent with an autocrine effect of IFN-γ regulating hA3G and hA3F expression. GATA3 is known to inhibit the production of IFN-γ 
and no effect was observed with neutralizing anti-IFN-γ antibody or TCR stimulation of GATA3-transduced cells. This may be due to a GATA3-mediated block to production of IFN-γ, or a direct effect of GATA3 binding to the hA3G and hA3F promoters. These possibilities remain to be directly tested.
We verified that the difference in expression in cytokine-derived T helper cells led to a biological difference: infectivity of HIV-1 virions produced from Th1 and Th2 effectors varied inversely with their relative levels of cellular and virion hA3G and hA3F. Removal of vif resulted in reduced infectivity of virions produced from both cell types. The greater reduction of infectivity of virions produced from the Th1 cells is consistent with the relative greater APOBEC3 levels in those cells. Over-expression of hA3G in Th2 cells reversed the relative decrease in virion hA3G and the consequent relative increase in infectivity of virions produced from Th2 cells. The magnitude of the effect of the ectopically-expressed hA3G is likely underestimated here, as not every cell in this population is expressing the transduced hA3G. In addition, reduction of hA3G in Th1 cells also correlated with an increase in infectivity. We used neutralizing anti-IFN-γ antibody to decrease hA3G expression in Th1 cells because shRNA against hA3G or nucleofection (for introduction of siRNA against hA3G) proved toxic to in vitro
-derived Th1 cells, which are more prone to cell death than other cultured T cells 
.These results are consistent with the variation in virion infectivity being caused, at least in part, by the differences in cellular and therefore virion hA3G, rather than other effects of the cytokine derivation.
In this study, we observed reduction of infectivity associated with increased amounts of readily detectable virion hA3G without identification of any G-to-A hypermutation. Although hA3G and hA3F are cytidine deaminases, there is extensive evidence that hA3G also reduces HIV infectivity through other mechanisms that may be the major contributor to A3G's inhibition of reverse transcription 
. Previous studies that have observed hA3G-related hypermutation in vitro
differed from the short term virus replication allowed here, and instead used prolonged serial passage of HIV in transformed cell lines over-expressing hA3G 
. Therefore, it is likely that the difference in infectivity based on cell source of virus observed here is due to the other antiviral activities of hA3G that are not measured by hypermutation.
A major issue concerning the role of hA3G and hA3F in HIV-1 pathogenesis is the question of whether in vivo
variation in these cellular restriction factors affects replication of wild type (eg, vif
-competent) HIV-1. Although high level over-expression of hA3G does impair replication of wild type HIV-1 in cell lines 
, more recent studies have not conclusively determined if there is a correlation between the variation in cellular hA3G expression observed across HIV-infected individuals' peripheral blood mononuclear cells and the plasma viral load in these subjects 
. The present results clearly indicate that physiological variations in hA3G levels in primary cells are inversely correlated with hA3G content and infectivity of wild type virions. This more direct measure of biological relevance observed here supports the conclusions of earlier reports showing that greater A3G activity was associated with lower viral load set-point 
, and suggests that continued investigation of the effect of APOBEC3 restriction factors on vif
-competent HIV-1 pathogenesis in vivo
The present results are also consistent with earlier reports that HIV-1 spreads better through cultures of Th2 cells than Th1 cells 
. This effect was most apparent in the prior studies using CXCR4 (X4) tropic viruses 
, such as the viruses used here, and not explained by differences in expression of that co-receptor between Th1 and Th2 cells. The present results suggest, however, that virions produced from Th2 cells may be relatively more infectious than those produced from Th1 cells because of their relatively lower hA3G content. In an earlier study 
, CCR5-tropic HIV replicated equally well in Th1 and Th2 cells. Th1 cells express higher levels of CCR5 coreceptor than Th2 cells 
. Indeed, X4-tropic viruses were chosen for study here to minimize possible difficulty in interpretation of opposing effects of both increased CCR5 co-receptor expression and increased A3G expression in Th1 cell cultures, though further investigation into how co-receptor tropism affects infectivity is certainly warranted. Moreover, the wide inter-individual variation in hA3G and hA3F expression in our results (a 14 fold range in hA3G protein expression in Th1 cells and a four fold range in Th2 cells) suggests that there may be polymorphisms in the regulatory regions of the APOBEC3 promoters 
, or in factors that can modulate hA3G and hA3F expression or function. We hypothesize that this variation in hA3G and hA3F may contribute to the wide variation of progression time to AIDS among different patients. The Th1/Th2 cell balance may also vary across individuals based on several factors. Autoimmunity may lead to a Th1 cell skewing and parasitic infections may cause aTh2 cell predominance. Our findings suggest that a shift in this balance prior to, or during, HIV-1 infection may lead to compounded pathogenic effects. Decreased relative expression of hA3G and hA3F in Th2 cells may lead to a greater rate of decrease in that cellular pool, decreasing CD4+ help to B cells for antibody production. Also, an individual's variation in Th1/Th2 balance may lead to differences in HIV-1 genetic variation due to hA3G- and hA3F-mediated sub-lethal cytidine deamination of viral genomes over repeated cycles of infection 
The present study indicates that the regulation of expression of hA3G and hA3F, and their functional effect on HIV-1 infectivity, depends on the cytokine-regulated differentiation state of CD4+ T helper cells. Further molecular characterization of signals that modulate hA3G and hA3F expression will be needed. The current results provide compelling evidence that increasing hA3G in primary T cells impairs HIV-1 replication despite the presence of Vif. This validates inducing higher hA3G expression as a novel strategy for prevention of infection and/or treatment of the vif-positive viruses present in infected humans.