Assessing the genetic and functional characteristics of transmitted HIV-1 Env variants isolated from donors and recipients is important to identify any biological signatures associated with mucosal transmission that can be exploited in the design of prophylaxis strategies. To gain insights into the biologic properties of viruses capable of breaching the infant gastrointestinal mucosa during transmission through breastfeeding, we isolated and produced 30 Env variants from the breast milk of 8 HIV-1-infected women who did or did not transmit HIV-1 to their infants postnatally and investigated their genetic and phenotypic characteristics. In addition, T/F Envs from 7 infants infected through breastfeeding were also studied. This represents one of the largest phenotypic comparative analyses of postnatally-T/F variants and uniquely compares their virologic phenotype to nontransmitted breast milk Env variants.
Several studies have attempted to identify genetic differences between chronic and transmitted viruses, with disparate results. The majority have reported that acute viruses have shorter variable loops and fewer glycosylation sites than chronic viruses [9, 13, 14, 54]. However, while comparing sequences from clade B chronic (n=24) and T/F envs (n=17), Wilen and collaborators  found no genetic signatures associated with transmission. Interestingly, a larger sequence analysis of clade B env gene sequences allowed for the identification of a small number of signature sequences that could influence Env incorporation and viral entry . In this study, we found that based on the Env V2 sequence, all infant T/F variants and most maternal milk variants are likely to have been susceptible to the potentially-protective V2-specific antibody responses elicited by the RV144 vaccine. Moreover, we observed that a previously-identified unique sequence (isoleucine at position 6) in the signal sequence of the env gene of placentally-transmitted clade C variants  was common in Env variants of postnatal-transmitting women and postnatally-infected infants, although the observed high frequency of this mutation could be due to the relatively small number of mothers and infants included in the study. The location of this signature amino acid in the signal sequence of the Env gene is not likely to affect the structure or interactions of the Env protein, but could affect the efficiency of Env production. Association of an Env signature sequence with more than one MTCT route may indicate that this unique sequence provides a fitness advantage to the virus in the infant . Therefore, it would be important to analyze this putative genetic signature in the env signal sequence in a larger panel of both clade B and C MTCT T/F envs.
The ability of HIV-1 virions to perform key steps required to cross the infant gastrointestinal epithelial barrier could contribute to restricting the number of founder viruses. We tested this hypothesis by measuring the interaction of milk Env variants with intestinal epithelial cells. Further, we assessed the ability of gp120 from these Envs to bind to Galcer, the putative epithelial cell attachment factor for HIV-1 . Milk Env variants from transmitting and nontransmitting women had an equally low efficiency epithelial cell attachment and Galcer binding. These results are in accordance with previous reports of low level of transcytosis across epithelial barriers [55–57]. However, previous studies have reported that Galcer binds efficiently to both gp120 and gp41 [58–60]. The interaction of the HIV Env with Galcer may be dependent on conformational epitopes and multivalent interactions. Thus, future work should investigate the binding of monomeric and polymeric gp140 of transmitted variants to Galcer liposomes.
Previous studies have indicated that DCs are present in infant oral and intestinal sub-mucosa . The binding of HIV virions to these cells and their transfer to CD4+ cells may be an important step in the establishment of mucosal HIV-1 infection , especially in the setting of postnatal transmission, as carbohydrates in milk may compete for interaction with lectin-binding molecules on DCs [43, 62]. While there appeared to be lower DC-binding efficiency of postnatally-T/F Env variants compared to those in milk of nontransmitting women, the efficiency of DC-mediated virus transfer was not different between the transmitted and nontransmitted variants. The direct correlation between DC-mediated trans-infection efficiency and number of O-linked glycosylation sites suggests that this function may more dependent on glycan:DC interactions. In this study TZM-bl cells were used as target cells instead of primary CD4 + T cells, which may better represent in vivo function, in order to assure reproducibility and comparability across assays. Our results are similar to those of Wilen and collaborators  who reported no difference DC trans-infection efficiency between chronic and T/F clade B viruses from sexual transmission. Therefore, the ability of transmitted Env variants from other MTCT transmission routes and other virus subtypes should be investigated to determine if efficient DC to CD4+ T cell transfer is a particular feature of HIV-1 MTCT. Of note, the Env pseudoviruses and recombinant Env proteins used in these studies were primarily produced in 293 T cells, which may yield distinct glycosylation patterns compared to primary cells. Thus, these investigations mainly address functional differences in transmitted and nontransmitted Env variants at the amino acid level. However, a comparison of DC transfer efficiency of two breast milk full length, replication-competent Env variants revealed similar transfer efficiencies, suggesting that the Env variants produced in 293 T cells are representative of those produced in primary host cells in this assay.
Infant and adult T/F viruses poorly infect monocyte-derived macrophages (MDM) [18, 63] and primarily utilize CCR5 [18–20, 48]. In accordance with previous reports, all the postnatal T/F Env variants were CCR5-tropic viruses, and there was no difference in CCR5 utilization between transmitted and nontransmitted milk variants. Moreover, CD4 utilization and infectivity in TZM-bl cells were comparable between transmitted and nontransmitted milk variants. Thus, our results suggest no difference in viral entry mechanisms between infant postnatal T/F and milk nontransmitted Env variants.
Maternal antibodies are transferred to the fetus during pregnancy and, at birth, infants born to HIV-1 infected women have high levels of HIV-specific antibodies [30, 64]. Furthermore, HIV Env-specific functional antibodies are present in the milk of HIV-infected women [31, 65]. These maternal antibodies may provide some protection from HIV acquisition [32, 65–67]. However, studies have found conflicting results regarding associations between maternal and infant antibody responses and MTCT, probably related to differences in modes of transmission and virus clades being assessed [30, 68]. Maternal neutralizing antibodies may also exert selective pressure on viruses that could contribute to the transmission bottleneck. In support of this hypothesis, previous studies reported that viruses transmitted during breastfeeding are usually resistant to maternal autologous neutralization [27, 32]. However, other studies did not observe increased resistance of infant viruses to neutralization by maternal plasma [14, 63, 69, 70]. Furthermore, no difference in sensitivity to autologous neutralization was observed between clade C viruses from postnatally-infected infant and viruses from their mothers . Accordingly, we also found no difference in autologous milk or plasma neutralization of milk Env variants from women who did or did not transmitted HIV to their infants postnatally. In addition, most Env variants from women who transmitted HIV to their infants through breastfeeding and from postnatally infected infants were sensitive to neutralization by the new generation of highly-potent broadly neutralizing mAbs (PG9, PG16, VRC01, and CH31). Thus, our results, together with other recent reports, suggest that neutralization-resistance may not be a defining feature of HIV-1 clade C viruses transmitted during breastfeeding.
Our inability to discern major differences in the genotype and phenotype of postnatal-T/F variants could support that crossing the epithelial barrier is a stochastic event that does not depend on specific biologic properties of the Env. Fortunately, animal studies have established that infusion of broadly-neutralizing antibodies can protect the infant against postnatal HIV-1 acquisition . Thus, an infant vaccine to prevent HIV-1 transmission through breastfeeding would have to induce rapidly-elicited protective antibodies to ensure safe breastfeeding. Interestingly HIV vaccination of infants born to HIV infected women induces immune responses distinct from that of their mothers . As infant immune response could take weeks to be established, maternal vaccination to elicit antibody responses in breast milk and/or passive infusion of the infant with broadly-neutralizing, protective mAbs could assure undisrupted protection of infants throughout the breastfeeding period. Interestingly, passive immunotherapy with broadly-neutralizing mAbs against the CD4 binding site in infant rhesus monkeys was shown to accelerate the development of functional antibody responses following SIV infection . Thus, passive administration of specific mAbs prior to infant vaccination might also be beneficial for the induction of the rapid immune responses needed in the setting of breast milk transmission.