The SBBC consists of 8 individuals (subjects C98, C54, C49, C64, C18, C135, C83 and C124) who became infected with an attenuated strain of HIV via contaminated blood products from a common blood donor (subject D36) between 1981 and 1984 303238. The SBBC blood transfusion recipients have been referred to as recipients 7, 13, 12, 9, 10, 4, 8, and 5, respectively, in one previous study 30 and subjects A (C18), B (C64), C (C98), D (C54), E (C49) and F (C83) in an earlier publication 38. Viral attenuation has been attributed to gross deletions in the nef/long terminal repeat (LTR) region of the HIV genome 32. The clinical history and laboratory studies of the subjects from the first identification as SBBC members through 1998 has been described previously 30, and a detailed update of the clinical and laboratory data through 2006 has been described recently 39. Briefly, despite being infected from a single source, SBBC members now comprise slow progressors (SP) who have eventually experienced decline in CD4 T-cells after many years of asymptomatic infection (subjects D36, C98, C54), and "elite" long-term nonprogressors (LTNP) who have had stable CD4 T-cell counts and low or below detectable plasma HIV RNA levels for more than 20 years without antiretroviral therapy (ART) and remain asymptomatic (subjects C49, C64, C135) 283031. Five SBBC members have died of causes either unrelated to- or not directly related to HIV infection (C98, C54, C18, C83, C124) (Table 1). The SBBC therefore provides a unique opportunity to study the pathogenesis of, and immune responses to nef-deleted HIV infection in a naturally occurring human setting.

To determine whether changes in viral phenotype were occurring in SBBC members, HIV isolation was attempted from peripheral blood mononuclear cells (PBMC) collected longitudinally from all subjects except C124 and C83 2840, by selected PBMC coculture techniques 4041 (Table 2). For subjects with detectable but low HIV RNA levels (D36, C54, C98, C18), more than 10 HIV isolates were obtained from each of D36, C54 and C98 over a 5 to 6 year period between November 1994 and November 2000 40. Three HIV isolates were obtained from C18 over an 8 month period between July 1993 and March 1994. For subjects with consistently undetectable HIV RNA levels (C49, C64, C124, and C135), a single isolate was obtained from C64 from PBMC collected in February 1996. This was despite isolation attempts from 16 additional PBMC collections between November 1995 and March 2001 40. All isolates carried similar but distinct deletion mutations in the nef gene and LTR region 28293242, and were unable to synthesize Nef proteins detectable by Western blotting or immunofluorescence staining of infected cells (D. McPhee and A. Greenway, unpublished data). No isolates were obtained from longitudinal samples of PBMC collected from C49 or C135 over a 4 to 7 year period between February 1994 and October 2000, or from a single sample of PBMC obtained from C124 in March 1993 40. Thus, success of isolating nef-deleted HIV from SBBC members was strongly dependent on the presence of detectable plasma HIV RNA levels, with few exceptions.

When compared with wild type HIV isolates and isogenic controls with mutations in nef, replication capacity of SBBC isolates in PHA-activated PBMC was found to be consistent over time by viruses isolated from a particular subject, but heterogenous between subjects and fell into 3 distinct phenotypes 2840 (Table 2). Viruses isolated from C18 replicated rapidly to high levels similar to wild type HIV; viruses isolated from D36 and C98 replicated to lower levels; and viruses isolated from C54 and C64 were barely able to replicate to detectable levels. In contrast, all isolates replicated to equivalent levels in the Cf2-luc reporter cell line 414344 expressing CD4, CCR5 and CXCR4. Thus, SBBC isolates except those from C18 appear to have attenuated replication capacity in PHA-activated PBMC. Inhibiting the in vivo replication of HIV in D36 by HAART demonstrated a prolonged in vivo virion half life, with a first-phase slope of decline of HIV RNA 0.18/day 45 which is slower than that seen in all previously studied individuals infected with wild-type HIV after commencement of ART 46474849. Thus, the replication kinetics of D36 virus appears to be attenuated both in vitro and in vivo.

Analysis of coreceptor usage in transfected Cf2-Luc cells 41 showed that all isolates used CCR5 (R5) as the primary coreceptor for HIV entry, except viruses isolated from D36 prior to commencement of HAART which were dual tropic and used CCR5 and CXCR4 (R5X4) 2840 (Table 2). These results showed that nef-deleted HIV was capable of undergoing a coreceptor switch from R5 to R5X4 in vivo. An isolate obtained from D36 after commencement of HAART was CCR5-restricted and had features of an early archived HIV variant, but was genetically similar to HIV present in a CSF sample obtained from D36 during HIV-associated dementia (HIVD) 28. Thus, for D36, HIV present in CSF during HIVD was likely to be an early variant that underwent compartmentalized evolution in the CNS. Moreover, we showed for the first time that nef-deleted HIV is inherently capable of undergoing compartmentalized evolution in the CNS and causing neurologic disease in humans 28. Stepwise quasispecies diversity was observed in SBBC SP, whereas C49 displayed stable quasispecies diversity most similar to early variants in the SBBC (B. Herring et al., manuscript submitted). Extended analysis of alternative coreceptor usage showed that D36 and C54 isolates could use CCR2b, C18 and C54 isolates could use CCR3, and C18 isolates could use Gpr15 for HIV entry, albeit at low levels 40 (Table 2). Whether expanded usage of alternative HIV coreceptors by SBBC isolates contributes to HIV pathogenesis in these individuals is uncertain, but the unique signature of coreceptor usage for viruses isolated from different SBBC members suggests independent evolution for each virus after infection of each cohort member. This interpretation is consistent with results of Env heteroduplex tracking assays, Env heteroduplex mobility assays and Env V1V2 length polymorphism assays which also demonstrated independent evolution of HIV Env in each subject (50, and B. Herring et al., manuscript submitted).

To better understand changes in pathogenicity which may have contributed to HIV progression in D36, Jekle et al 51 used an ex vivo human lymphoid cell culture system to analyze the ability of 2 HIV viruses isolated from D36 to deplete CD4+ T-cells; one isolated in 1995 prior to the onset of AIDS (D36II) and another isolated in 1999 after the onset of disease progression (D36XI) (Table 2). Although both D36 isolates were less potent in depleting CD4+ T-cells than reference X4 and R5X4 isolates with intact nef genes, the 1999 isolate induced greater levels of CD4+ T-cell cytotoxicity than the 1995 isolate. Differences in CD4+ T-cell cytotoxicity between the 2 isolates were evident in CD4+/CCR5- cells, but not evident in CD4+/CCR5+ cells suggesting an increased ability to use CXCR4 by the 1999 isolate. Further studies with the CXCR4 inhibitor AMD3100 showed that, although both isolates were functionally R5X4 2840 (Table 2), the 1999 isolate had preferential use of CXCR4 whereas the 1995 isolate had preferential use of CCR5 for HIV entry. These studies showed evolution of R5X4 strains in D36 to a variant with higher cytopathic potential that was associated with increased use of CXCR4 in vitro and HIV progression in vivo.

Consistent with results of the study by Jekle et al 51, we showed alterations in HIV cytopathicity by sequential D36 isolates in cultures of monocyte-derived macrophages (MDM). Compared with the highly macrophage tropic R5 ADA and R5X4 89.6 isolates, all D36 viruses replicated in MDM to low levels and had delayed replication kinetics 52. There was no evidence of increased HIV replication in MDM by virus isolated from D36 after HIV progression. However, in support of the results obtained by Jekle et al 51, the 1999 isolate caused extensive cytopathicity in MDM similar to that present in cultures infected with ADA or 89.6, characterized by the presence of many syncytia 52. In contrast, earlier D36 isolates caused only few or occasional syncytia in MDM despite all D36 viruses replicating in MDM to similar levels. Thus, increased cytopathicity in MDM by the 1999 D36 isolate is most likely due to intrinsic pathogenic features of the Env that increase fusogenicity, similar to that which has been observed by particular neurotropic R5 and R5X4 viruses 535455. The increased Env fusogenicity may have contributed to greater cytopathicity by the 1999 D36 isolate and HIV progression in D36. Further studies to elucidate the molecular determinants of D36 Env that are associated with increased fusogenicity are in progress.

The effect of long-term infection with nef-deleted virus on CD4+ T cells was studied in detail for six SBBC members 56. Careful comparison with age- and transfusion-matched controls revealed the surprising result that SBBC members had an increased number of circulating CD45RO+ memory CD4+ T cells. This was unexpected, since these CD4+ T cells are widely believed to represent the main target of cytopathic HIV infection 57585960 (reviewed in 61), and loss of these cells ultimately leads to acquired immunodeficiency. Therefore, this result is consistent with the hypothesis that nef-deleted HIV has reduced pathogenicity in vivo.

Nevertheless, in the SBBC subjects studied with detectable plasma viral load, C54 and C98, there was concomitant elevation of CD8+ T cell activation, whereas the SBBC subjects with undetectable plasma viral load, C49, C64 and C135 had normal levels of CD8+ T cell activation 56. Therefore, within the SBBC, the situation was similar to the strong correlation seen between plasma viral load and CD8+ T-cell activation in subjects infected with wild type HIV 62. Furthermore, as described above, subjects D36, C98 and C54 exhibited a clear CD4+ T cell decline, albeit at a relatively slow rate 282930. This interesting finding argues that pathogenicity within the SBBC was more closely correlated with levels of viral replication (as assessed by plasma viral load) and CD8+ T-cell activation than with viral pathogenicity dictated by the presence or absence of nef. This finding likely represents the ability of host factors to modulate the pathogenicity of nef-deleted HIV-1 6364. CD8+ T cell activation may reflect lymphocyte turnover during HIV infection, which has been proposed to lead to disruption of normal homeostasis and eventual consumption of both memory and naïve CD4+ T cells 6566. However, we did not find evidence of dramatically increased CD4+ T cell turnover in these subjects, as determined by expression of Ki-67 as a marker of cell proliferation 56.

To determine whether an evolving nef/LTR sequence contributed to HIV progression in D36 and C98, we undertook a detailed longitudinal analysis of nef and LTR sequence changes occurring in D36, C98, C49, C54 and C64 over a 4 to 10 year period 29. Sequential analysis of nef/LTR demonstrated a gradual loss of nef sequence that differed in magnitude between subjects. A large deletion of 128 bp emerged in D36 effectively removing the entire nef gene with the exception of the region surrounding the nef start codon, the polypurine tract which contains terminal signals for HIV integration, and a 90 bp region of the nef/LTR overlap region surrounding the negative regulatory element. The pattern of nef/LTR sequence loss in C98 was remarkably similar to that of D36. The pattern of nef /LTR sequence loss in C54 was also similar, but less extensive than that observed in D36 and C98. However, the additional loss of nef/LTR sequence in C64 was comparatively minimal. These data are illustrated in Figure 2, where the nef/LTR sequences cloned from the earliest available and most recent blood samples of these subjects are compared. A more detailed longitudinal analysis of nef/LTR sequences in these subjects has been reported in Churchill et al 29. Thus, viruses harboured by D36, C54, C98 and C64 appeared to be evolving in a convergent fashion toward a highly deleted, minimal nef/LTR structure containing only sequence elements that are absolutely essential for HIV replication 29. The convergent nature of the nef/LTR sequence changes implies the presence of strong selection pressures that maintains the ability of defective HIV genomes to persist in vivo. The highly evolved nef/LTR sequences harboured by D36, C54 and C98 were strikingly similar to those that remained dominant in C49 for at least 10 years (Fig. 2) 29. Thus, the highly evolved nef/LTR structure appears to be stable, and in the case of C49 does not increase pathogenicity. However, taken together the results suggest the in vivo pathogenicity of nef-deleted HIV harboured by SBBC members is dictated by factors other than those that impose a unidirectional selection pressure on the nef/LTR region of the HIV genome. Due to the changes in the nef/LTR region, these other presumably host factors become more important in terms of disease outcome. This is exemplified by the marked variation from no disease progression with no detectable virus replication (C49 and C64), to no progression with a low viral load (C54), through to slow progression (D36 and C98).

Reversion to pathogenicity by nef-deleted SIV has been associated with restoration of a truncated Nef protein 26, acquisition of further deletions in the nef/LTR overlap region 67, and/or duplications of NF-κB binding sites in the LTR 67. In contrast to the SIV studies, the in vivo evolution of nef-deleted HIV in SBBC members was unidirectional toward a smaller nef/LTR sequence and the majority of the additional sequence loss was within the nef-alone region 29. Furthermore, none of the clones were capable of encoding Nef. Together, these results suggest improved viral replication by further deleting remnants of the nef gene. In addition, the presence of duplicated NF-κB binding sites in the LTR was not associated with clinical status of the SBBC subjects. Therefore, it is likely that viral factors that modulate the in vivo pathogenicity of nef-deleted HIV will be distinct from those in nef-deleted SIV. Interestingly, the unidirectional evolution toward the minimal nef/LTR sequence observed in SBBC members was strikingly similar to the pattern of evolution in a slow progressor infected with a nef/LTR-deleted variant of HIV circulating recombinant form 01_AE 34. The convergent pattern of nef/LTR evolution among viruses harboured by SBBC members is therefore unlikely to be due to a unique property of the infecting strain, but rather a positive selection that is common across clades.

Viruses harbored by SBBC members contain unique alterations of NF-κB and Sp-1 binding sites in the LTR that may affect transcriptional activity and thus, replication capacity 282932. Therefore, we examined the nucleotide sequence and transcriptional activity of nef/LTR clones obtained sequentially from D36 blood samples and from D36 CSF obtained during HIVD, to determine whether changes in LTR activity may contribute to neuropathogenesis of nef-deleted HIV-1 infection 68. We found that the transcriptional activity of CSF-derived nef/LTR clones was up to 4.5-fold higher than blood-derived clones isolated before and during HIVD when tested under basal, PMA- and Tat-activated conditions. The presence of duplicated NF-κB and Sp-1 binding sites or a truncated nef sequence in blood-derived nef/LTRs was not sufficient to mediate large increases in transcriptional activity. However, CSF-derived nef/LTRs had duplicated NF-κB and Sp-1 binding sites coupled with a truncated nef sequence, which formed a regulatory unit that significantly enhanced LTR transcription 68.

Previous studies showed that LTR variants with augmented transcriptional activity enhance replication of HIV 69. Therefore, to determine whether D36 nef/LTRs affect replication capacity of HIV in vitro, we produced and characterized full-length chimeric molecular clones of HIV NL4-3 carrying the nef/LTR nucleotide sequence of blood-derived D36 nef/LTRs or the CSF-derived D36 nef/LTR 68. We examined the capacity of chimeric NL4-3 viruses carrying D36 nef/LTRs to replicate in PBMC compared with wild type NL4-3 and the NL4-3ΔNef deletion mutant 70. Compared to wild type NL4-3, chimeric HIV containing the nef/LTR sequence of blood derived D36 viruses had attenuated replication kinetics, similar to NL4-3ΔNef. In contrast, chimeric HIV containing the nef/LTR of D36 CSF had enhanced replication capacity compared to wild type NL4-3. Thus, the nef/LTR derived from CSF of D36, which had augmented basal, PMA- and Tat-activated transcriptional activity compared to wild type and blood-derived D36 nef/LTRs, augmented replication of HIV in PBMC. Together, our results suggest unique features of the CSF-derived nef/LTR restore efficient replication capacity of nef-deleted HIV in PBMC by enhancing transcription. The results further suggest that nef and LTR mutations that augment transcription may contribute to neuropathogenesis of nef-deleted HIV.

In addition to nef and LTR, mutations or polymorphisms in other HIV genes including gag, rev, tat, vif, vpr, vpu and env have been detected in SP or LTNP 7172737475767778. A previous study of HIV rev alleles isolated from a subject with long-term nonprogressive HIV infection showed a persistent Leu to Ile change at position 78 in the Rev activation domain which attenuated Rev function and HIV replication capacity 73, providing evidence that defective rev alleles may contribute to long-term survival of HIV infection in some patients. A subsequent study of naturally occurring rev alleles with rare sequence variations in the activation domain showed variable reductions in Rev activity 79, although it was unclear from this study whether the observed reductions in Rev activity would be sufficient to attenuate HIV replication capacity. Of note, CTL escape mutations in the second coding exon of Tat have been shown to attenuate virus in vivo, suggesting that in vivo sequence changes in other regulatory HIV-1 genes may potentially affect HIV-1 pathogenesis 80. Since differences in HIV pathogenicity in SBBC members could not be fully explained by alterations in the nef/LTR region 29 or Env phenotype 4050, we characterized dominant HIV-1 rev alleles that persisted in SBBC LTNP (C18, C64) and SP (C98, D36) 81. We found that the ability of Rev derived from D36 and C64 to bind the Rev responsive element (RRE) in RNA binding assays was reduced by approximately 90% compared to Rev derived from NL4-3, C18 or C98. D36 Rev also had a 50–60% reduction in ability to express Rev-dependent reporter constructs in mammalian cells. In contrast, C64 Rev had only marginally decreased Rev function despite attenuated RRE binding. In D36 and C64, we found that attenuated RRE binding was associated with rare amino acid changes at 3 highly conserved residues; Gln to Pro at position 74 immediately N-terminal to the Rev activation domain, and Val to Leu and Ser to Pro at positions 104 and 106 at the Rev C-terminus, respectively. In D36, reduced Rev function and altered replication capacity was mapped to an unusual 13 amino acid extension at the Rev C-terminus. However, database analysis of rev sequence demonstrated that the presence of one or more of these rare amino acid changes was not able to discriminate between subjects with progressive or non-progressive HIV-1 infection. Moreover, none of these amino acid changes occurred in a previously identified LTNP with defective rev alleles 73. Thus, our studies suggest the contribution of any or all of these mutations to decreased RRE binding and/or attenuated Rev function by SBBC Revs, and possibly to slow or absent HIV progression, is likely to be context dependent.

It is presently unclear whether attenuated D36 Rev function in vitro equates to attenuated Rev function in vivo, and indeed whether attenuated Rev function contributed to slow progression of HIV infection in this subject. Extrapolation of these in vitro findings to an in vivo role for attenuated D36 rev alleles is difficult, since this subject and other SBBC members are infected with virus containing gross nef/LTR deletions which have been shown to cause significant viral attenuation in this cohort 282932. Nonetheless, our findings raise the possibility that attenuated Rev function may contribute, at least in part, to viral attenuation and slow HIV progression in D36.

To better understand the humoral immune response to nef-deleted HIV infection, we measured the total IgG responses in longitudinal plasma samples of SBBC members by Western blotting, and compared these with total IgG responses in a control group of LTNP with intact nef genes 40. We found a good correlation between total IgG responses in SBBC members and a detectable plasma VL, with plasma from C18, C54, D36 and C98 all being strongly reactive. Subjects C49 and C64, who consistently maintained undetectable HIV RNA copy numbers 2930, had significantly reduced total IgG responses. Furthermore, subject C135, who has also had consistently undetectable HIV RNA levels has not fully seroconverted after more than 20 years of infection 4042. These studies highlight the importance of adequate antigenic stimulation by nef-deleted HIV to drive antibody production. In contrast, we found that total IgG responses in the control group were uniformly potent, reflecting the fact that all these individuals had detectable VLs. Among SBBC members, the strongest antibody responses were observed in individuals with low but detectable VL set points, less than approximately 10,000 RNA copies/ml. This is consistent with a recent observation of an undetectable VL and weak, delayed antibody responses in an unrelated individual with nef-deleted HIV 34.

Recent studies have highlighted the change in HIV antibody responses with respect to antibody isotype switching after initial infection, in particular IgG₃ reactivity to p24 8283. Preliminary WB testing of p24 IgG₃ responses in the SBBC LTNP indicated there was reactivity for those individuals with the most potent total IgG responses (C18 and C54) (E. Verity, D. McPhee, K. Wilson, and D. Zotos, unpublished data). This hints at delayed isotype switching and hence delayed maturation of immune responses for at least some members of the SBBC.

In additional studies, we determined whether plasma from SBBC members had differences in ability to neutralize nef-deleted HIV strains. We did this by comparing the ability of longitudinally collected plasma from SBBC members or from control LTNP cohort members to neutralize the infectivity of HIV isolated from D36 and C18 40. We found that, for SBBC plasmas, neutralization of D36 or C18 viruses strongly correlated with VL, replication capacity of the isolated virus, and the strength of anti-HIV IgG responses. Plasma from SBBC members with undetectable VL was unable to neutralize the infectivity of these viruses. In contrast, plasmas from the control LTNP cohort were able to neutralize the infectivity of SBBC viruses with titres generally higher than that seen for SBBC members, but there was no correlation between neutralization and HIV RNA copy number or IgG responses in the control LTNP group.

A number of studies have suggested an increased frequency of LTNP and long term survivors (LTS) possess strong, cross-reactive neutralising antibody responses 848586. However, very few studies have investigated antibody responses in LTNP/LTS infected with nef-deleted HIV, despite a number of studies showing strong neutralising antibody responses in macaques infected with nef-deleted SIV 878889909192. However, studies by Greenough et al., 93 showed measurable neutralising antibody responses concurrent with a detectable VL for a single individual infected with nef-defective HIV. We therefore determined the breadth of neutralizing antibodies against HIV for SBBC members.

We found that plasma from C18, and to a lesser extent D36, C54 and C98 were able to potently neutralize the infectivity of a number of laboratory adapted and primary HIV strains, but little or no neutralizing activity was evident in plasma from C49, C64 and C135 40. Further analysis of cross reacting neutralizing antibodies in SBBC plasma demonstrated that plasma derived from C18 was able to potently neutralize the infectivity of HIV-1_ROK39 (clade A), HIV-1_SE364 (clade C), HIV-1_BCB93 (clade D), and HIV-1_92TH024 (CRF01_AE), while plasma derived from C98 was able to only weakly neutralize the infectivity of HIV-1_ROK39 and HIV-1_SE364 40.

Together, our studies on antibody responses in SBBC members demonstrated that infection with nef-deleted HIV can, in some individuals, induce antibody responses capable of potently neutralizing a broad range of isolates. It is possible that the breadth of antibody responses observed in SBBC members may be associated with unrestricted immunoglobulin class switching, which is inhibited by Nef 94. This would be counterbalanced by the very low viral replication in vivo (VL) and a detectable IgG₃ p24 response for C18 and C54. Hence, the broad antibody response may also reflect an early response to infection. The strongest antibody responses in SBBC members were observed in individuals with a detectable VL, supporting the model of a VL threshold which must be reached to provide adequate antigenic stimulation to drive antibody production 14. However, this higher level of virus replication places these individuals at greatest risk of disease progression. Signs of disease progression were observed for 2 SBBC members (D36, C98) 282930, demonstrating that the potent neutralizing antibody responses observed did not protect from disease progression. These results question the effectiveness of a broad neutralizing antibody response in individuals with established infection. However, this does not preclude an important role for neutralizing antibodies in preventing initial infection.

Analysis of cellular immune mechanisms suppressing nef-deleted virus in SBBC nonprogressors may provide insights relevant for HIV vaccine design. Several studies have demonstrated that the presence of a sustained Gag-specific CD8+ T-cell response is associated with protection against disease progression in cohorts of LTNP infected with nef-intact HIV 959697. In support, our studies have also shown that the distinguishing feature of LTNP harbouring nef-intact HIV is the predominance of Gag-specific CD8+ T-cell responses, which decline when these individuals start to progress (W.B. Dyer et al., manuscript in preparation). However, novel qualitative and quantitative differences in immune correlates of viral control may exist in LTNP infected with nef-deleted HIV. Longitudinal studies of T-cell responses in SBBC members have demonstrated a dominance of Pol-specific CD8+ T-cell responses rather than those against Gag 98 and W.B. Dyer et al., manuscript in preparation). This suggests that infection with nef-deleted HIV may give rise to a qualitatively different response, although a subset of SBBC members also had strong CD8+ T-cell responses to Gag 98. Nonetheless, CD8+ T-cell responses to Gag or Pol did not discriminate SBBC LTNP from SP.

In contrast, longitudinal analysis of HIV-specific CD4+ T-cells showed that all SBBC nonprogressors able to completely control plasma HIV RNA levels to below detectable levels (C49, C64 and C135) had persistent and strong T-helper proliferative responses to HIV p24 antigen, whereas these responses were absent in all progressors with persistent viremia (C54, C98 and D36) (Fig. 3). The notable exception was C18 who had detectable but low plasma HIV RNA levels without evidence of CD4+ T-cell loss, and had detectable p24 T-helper responses over a short period of approximately 12 months leading up to his non-HIV related death at 83 years of age. In this subject, the p24 T-helper responses coincided with an exponential increase in CTL responses against Pol antigens, measured by memory CTL precursor frequency assay 98, and IFN-gamma ELISPOT responses (W. Dyer, unpublished data). However, it is difficult to interpret the significance of the T-cell responses in C18 given the short window of analysis. Together, although derived from a small cohort of individuals, these results suggest that immune suppression of nef-deleted HIV-1 by SBBC LTNP may be dependent upon persistent T-helper responses irrespective of the CD8+ T-cell and neutralizing antibody response to viral antigens.