In this study, we have shown that three intrasubtype C superinfected individuals, in whom superinfection was detected within the first year of infection, have low to undetectable titers of autologous NAbs to their early/founder Env prior to superinfection and as late as 8-months post-seroconversion. This is in sharp contrast to ten matched non-superinfected controls similarly evaluated for neutralization of early/founder variants over the first year of infection, of which a majority mounted very potent neutralizing activities. This occurred as early as three-months post-seroconversion, when the median IC50 was 1896. Despite the small size of this study, the differences in autologous NAb titers were significantly different between the two groups (p = 0.039), and suggest that slower development of a humoral immune response increased susceptibility to intra-subtype superinfection in this cohort.
This result is consistent with a previous study of a subtype B MSM cohort, where low titers of autologous and heterologous NAbs were observed in the three superinfected individuals relative to matched non-superinfected controls . However in this same study, autologous pre-superinfection Envs were tested for neutralization only cross-sectionally against contemporaneous pre-superinfection and post-superinfection plasma, and heterologous breadth assays were performed against only two lab-adapted subtype B strains. Moreover, there was no evaluation of cross-neutralization of the superinfecting virus using plasma prior to superinfection . Nevertheless, there is a common observation that superinfection occurred during the first year of infection, and was associated with low autologous neutralizing antibody responses . These results are consistent with the hypothesis that higher susceptibility to superinfection during early infection may be, in part, due to diminished early humoral responses.
A different conclusion was reached from a study of superinfection in HIV-1 infected commercial sex workers in Mombasa, Kenya . There it was shown that while NAb breadth and potency were lower in superinfected individuals than in matched controls after approximately one year of infection, no difference in these parameters occurred immediately prior to superinfection (between 0.72-5 years post-infection) . In 4/6 cases identified in that study, superinfection occurred at or after two years of the initial infection, potentially allowing for development of stronger, yet still not protective, NAb responses . Thus in this multiple HIV-1 subtype sex worker cohort, NAb did not appear to provide any protection from superinfection. While the authors did not investigate autologous NAb responses to transmitted/founder Env glycoproteins in the study, responses to initial variants cloned from the time of superinfection detection and early Envs from within the first year of infection were evaluated .
To evaluate cross-neutralization breadth prior to superinfection, we evaluated the potential of pre-superinfection plasma to neutralize not only superinfecting variants, isolated at the time superinfection was detected, but also a subtype C reference panel of pseudoviruses. We found that pre-superinfection plasma was unable to neutralize superinfecting variants and had limited ability to cross-neutralize a panel of variants prior to superinfection, with a range of 0–7 (of 12) variants neutralized at very low IC50s (20–70) amongst all three superinfected cases. Heterologous breadth in non-superinfected control plasma samples was similarly limited, though some individuals did have greater breadth but not potency. These data are consistent with previous studies, which showed that early autologous NAbs in subtype C infection are monotypic with limited cross-neutralization potential [22, 23, 26, 31, 32]. Furthermore, it has been demonstrated that significant cross-neutralizing antibody breadth is unlikely to occur prior to chronic infection [33, 34].
Heterologous neutralizing antibody breadth did not necessarily correlate with strength or effectiveness of autologous NAb responses. Although some non-superinfected individuals clearly mount strong autologous responses, they may exhibit limited neutralizing breadth by primarily targeting single or nonconserved epitopes [22, 23, 25, 26, 31, 32, 35, 36]. In contrast, others with relatively low-titer autologous responses may in fact have wider breadth to multiple epitopes (or different epitopes), none of which confers a particularly effective neutralizing antibody response to the established infecting variant. Thus, this study suggests that, in the context of intrasubtype superinfections, either the ability to potently neutralize autologous virus or to target multiple epitopes could provide protection against superinfection. However, in the absence of both of these humoral responses, individuals may be predisposed towards superinfection.
Based on the data suggesting early deficits in NAb responses in superinfected individuals, but with little evidence for broadly neutralizing antibodies in the matched controls, we investigated whether levels of non-neutralizing antibodies also differed in the two groups prior to superinfection. We observed that superinfected individuals trended towards having lower levels of gp120-specific IgG antibodies prior to superinfection compared to controls, although this comparison did not achieve statistical significance (p = 0.115).
Similarly, we observed no reactivity to either consensus clade C or caseA2clB (clade B) V1V2-loop fusion proteins [8, 30] in plasma from superinfected individuals prior to superinfection. By contrast in 3/10 non-superinfected matched controls, we observed reactivity to both proteins during the first 6 months, and in 6/10 controls reactivity was seen against the consensus C protein during the first year of infection. Higher levels of IgG V1V2-loop binding antibodies have been correlated with protection from primary HIV-1 infection in vaccinees that remained uninfected in the RV144 trial [8, 9], and the data presented here are consistent with the concept that these antibodies may contribute toward protection in individuals that remained only singly-infected.
In the RV144 trial, levels of IgA antibodies capable of binding to gp120 were directly correlated with the risk of infection [8, 9]. It is of interest, therefore, that two of the three superinfected individuals showed the highest anti-gp120 plasma IgA levels amongst all study participants, while only two of the ten matched controls demonstrated positive IgA binding titers. One superinfected individual, ZM211F, showed no evidence of anti-gp120 IgA reactivity. However, this is consistent with the low overall HIV-1 specific humoral responses observed, including the lowest levels of V1V2-loop and gp120-specific IgG binding antibodies prior to superinfection. We have also found a statistically significant difference in anti-gp120 plasma IgA levels with respect to sexual exposure and potential HIV-1 acquisition risk, in that individuals either with superinfection (as a result of outside partnerships) or self-reported outside partnerships (in non-superinfected individuals) had significantly higher anti-gp120 plasma IgA responses (p = 0.005), as compared to non-superinfected controls without self-reported outside partnerships. This data corroborates those drawn from the RV144 trial that high plasma IgA levels may be a surrogate of HIV-1 exposure or a potential correlate of risk in the context of primary HIV-1 infection  and superinfection. We have yet to evaluate the mechanism by which these differences in plasma IgA levels may affect susceptibility to infection, however it has been suggested that high levels of IgA may interfere with other potentially protective antibody-mediated effector functions such as antibody-dependent cellular cytotoxicity (ADCC) . Non-neutralizing IgG antibodies could play a major role in increased mucosal barrier protection, sequestering the virus at the epithelial surface and in female genital tract mucus and contributing to Fc receptor-mediated antiviral activity [6, 37, 38]. Thus a diminished non-neutralizing IgG antibody response, compounded by potentially interfering IgA antibodies, could lead to reduced mucosal protection and higher susceptibility to superinfection. Future studies will elucidate whether non-neutralizing antibody-mediated antiviral activities contribute to protection from superinfection.