We assessed five in vitro Nef functions using clonal plasma HIV RNA sequences from 45 EC and 46 CP. We observed that EC Nef clones were generally functional, especially for Nef’s most characteristic activities, CD4 and HLA-I down-regulation. Nevertheless, median EC Nef activities were significantly lower for all five functions when compared to those from CP. Median CP Nef activities were consistent with that of HIV-1 strain SF2 used as a normalization control for all assays, indicating that our selection of chronic Nef clones is representative of chronic Nef isolates examined previously. The range in Nef activities observed here may help to resolve discrepancies between previous studies of HIV long-term non-progressors or controllers, which have reported relative preservation of CD4 and/or HLA-down-regulation function
[32, 33], inefficient Nef-mediated CD4 and/or HLA-down-regulation
[34–36] and reduced infectivity
 compared to CP. Our data suggest that there is in vivo pressure on Nef in EC to maintain CD4 and HLA-I down-regulation functions.
Relative functional impairments between EC and CP clones are not likely to be due to differences in Nef protein stability or expression levels, nor to recent descent from a defective common ancestor. Similarly, while enrichment of protective HLA alleles in EC may contribute to Nef sequence, it is not likely to be the only explanation for relative functional attenuation observed here, since differences between groups persisted after persons who expressed protective HLA alleles were excluded from analysis. Indeed, although significantly lower CD4 down-regulation activity was observed in B*57 compared to non-B*57 EC (Figure
5E), this was not true for other Nef functions, indicating that B*57 expression alone does not guarantee Nef attenuation in this group.
Rather, our results are consistent with functional variability of naturally occurring Nef sequences from EC, which may be attributable in part to non-canonical HLA-associated escape mutations selected in this rare group. Previously, in order to investigate the influence of HLA-associated viral polymorphisms on HIV-1 protein function in EC, we have made use of reference lists of common HLA-associated polymorphisms derived from population-level studies of chronically infected individuals
[20, 29]. However, such lists may not capture rare escape mutations that are unique to EC
[26, 27]. Therefore, we used our EC dataset to identify HLA-associated polymorphisms specific to this population, in a novel analysis of this type. It is important to note that this analysis was aimed at identifying HLA-B*57-associated polymorphisms in the EC dataset itself (and is therefore distinct from the comparative analyses described in
 that sought to identify polymorphisms enriched among EC compared to CP). Statistical power is a major limitation of this analysis, but HLA-associated polymorphisms can be identified in modestly sized datasets (e.g.
), especially if analyses are limited to specific alleles. At p < 0.05, we identified nine B*57-associated polymorphic sites specific to EC (Figure
6A), most of which differed from B*57-associated polymorphisms commonly identified in population-level analyses
[20, 29]. In contrast, B*57-associated polymorphisms identified in our CP dataset using the same approach were largely as expected
[20, 29, 30]. ELISpot reactivity to overlapping peptides spanning codons 28, 55, 85, 178 and 198 has been documented in B*57+ EC
; and V85L has been described to function as an escape mutation in the B*57/58-KF9 epitope
, further suggesting that these polymorphism may be due to CTL selection pressure in B*57 EC. In addition, paired Nef sequences from PBMC and plasma were available for two B*57 EC in the present study. The first exhibited identical amino acids in both compartments at all 7 of the residues putatively associated with HLA-B*57, while the second exhibited different amino acids at three of the seven B*57-associated sites (codons 85, 105 and 198). Although it is not possible to make conclusions based on only two patients, these data support evolution of these sites in vivo.
Notably, we observed dramatic inverse associations between the number of EC-specific B*57-associated polymorphisms and Nef-mediated replication, HLA-I down-regulation and CD74 up-regulation. CD4 down-regulation, the only function in which host expression of B*57 was in itself significantly associated with poorer function in EC, also displayed a modest, albeit not significant, negative relationship between the burden of B*57-associated escape mutations and function (Figure
6). These remarkable inverse relationships were particular to B*57 and not observed for other HLA alleles with similar frequency in our EC cohort. Taken together, results suggest that HLA-B*57-associated CTL pressures select for non-canonical polymorphisms in EC, which contribute additively to multiple functional impairments in EC Nef. Building upon previous studies of recombinant viruses encoding gag and pol sequences from the same EC cohort
[3, 4], our results support a complex relationship between B*57-associated immune pressures and Nef function.
In an exploratory analysis of Nef amino acid sequences, we identified 23 polymorphisms, located at 14 residues, associated with Nef function in EC. None overlapped with mutations previously identified in HIV non-progressors
 or with sites reported to affect HLA-I down-regulation activity in chronic infection
, a discrepancy that might be due to non-canonical polymorphisms observed in EC. Of interest, 8S was associated with Nef-mediated modulation of cell-surface CD4, HLA-I and CD74, an observation consistent with codon 8’s involvement in myristylation
. Also, 28D and 105X (in this case R), associated with modulation of CD74 and CD4, respectively, are EC-specific B*57-associated polymorphisms identified in the present study.
Viral genetic studies of EC feature numerous challenges and limitations. Although care was taken to choose a Nef clone that reflected each patient’s original bulk HIV RNA sequence, and to rule out proviral DNA contamination, potential biases associated with PCR amplification from extremely low copy-number templates must be acknowledged. On the other hand, use of a single sequence per patient alleviates potential biases associated with quasispecies approaches to compare samples with low vs. high genetic diversity. As our goal was to specifically investigate the function of Nef in EC, we employed recombinant virus (and single-protein expression) approaches to eliminate potential confounding effects of other HIV-1 proteins; however, such approaches may not reflect the characteristics of infectious molecular clones or whole-virus isolates recovered from PBMC, procedures that are rarely successful in EC
[21, 42]. Recombinant virus approaches are also inherently limited by potential incompatibilities between insert and backbone; our choice of a recombinant control strain (NL4.3-NefSF2) alleviates this to a minor extent. Although we assessed Nef activity using primary PBMCs and immortalized cell lines, Nef’s multiple functions
[16–18, 43] may vary in different cell types
[43–45]. Furthermore, the in vivo relevance of our observations - in particular, the extent to which these functional differences contribute to the viremia control in EC - remains unclear. Although our results are consistent with the transmission of partially attenuated Nef sequences in at least some EC, and/or further immune-mediated attenuation in others
[3, 4, 46], it is not possible to disentangle cause and effect in cross-sectional studies. Furthermore, Nef function may change over the infection course
, therefore longitudinal analysis of Nef function in controllers, beginning in the acute/early phase of infection, is warranted. Finally, although our results suggest that non-canonical polymorphisms in EC may contribute to attenuated Nef function, it will be important to validate these findings in larger EC cohorts. Despite these limitations, our study represents the largest linked analysis of multiple in vitro Nef functions in EC to date, and to our knowledge the only study assessing CD74 up-regulation and replication capacity in this group.