Immune-Based Approaches to the Prevention of Mother-to-child-Transmission of HIV-1: Active and Passive Immunization

Synopsis

Despite more than two decades of research, an effective vaccine that can prevent HIV-1 infection in populations exposed to the virus remains elusive. In the pursuit of an HIV-1 vaccine, does prevention of exposure to maternal HIV-1 in utero, at birth or in early life through breast-milk require special consideration? In this article we will review what is known about the immune mechanisms of susceptibility and resistance to mother-to-child transmission (MTCT) of HIV-1 and will summarise studies that have used passive or active immunisation strategies to interrupt -MTCT of HIV-1. We will also describe potentially modifiable infectious co-factors that may enhance transmission and/or disease progression (especially in the developing world). Ultimately an effective prophylactic vaccine against HIV-1 infection will need to be deployed as part of the Extended Programme of Immunisation (EPI) recommended by the World Health Organisation (WHO) for use in developing countries, so it is important to understand how the infant immune system responds to HIV-1 antigens, both in natural infection and presented by candidate vaccines.

Introduction

Mother to child transmission (MTCT) of HIV-1 infection remains an important cause of new HIV-1 infections worldwide, despite the increasing implementation of prevention strategies using anti-retroviral therapy (ART) across the developing world. In the year ending December 2008 an estimated 430, 000 children under the age of 15 years were newly infected with HIV-1 (UNAIDS Epidemic Update 2009), the great majority of whom acquired the infection from their mothers in low and middle-income countries. Despite clear evidence of significant progress, challenges remain for poor countries in providing comprehensive screening programmes for pregnant women and implementing the full range of prevention services for those identified as HIV-1-infected.

For those children who acquire perinatal infection, disease progression appears to be unusually rapid compared to that of adults, particularly in developing countries, where mortality rates as high as 20–52% have been reported in the first two years of life^{1 2}. A key factor contributing to the rapid disease progression observed in infants may be the persistently high levels of HIV-1 viremia observed throughout the first year of life, with the “set-point” viral load (VL) rarely falling more than 1 log below the peak VL^{3 4}, despite the early appearance of HIV-1-specific T-cell responses⁵. It is likely that co-infections acquired in babies in the developing world may play a role in rapid disease progression, such as human cytomegalovirus (CMV), which infects most West African children in the first year of life⁶ and is associated with high viral load in HIV-1 co-infected infants⁷. Co-infections, particularly malaria in pregnancy, may also contribute to the likelihood of transmission: therefore it is important to investigate and define the potential role of co-infections in MTCT so that these may be modified in preventive strategies.

Mother-to-child transmission of HIV-1: an immunological perspective

Vertical transmission of HIV-1 is not an inevitable consequence of exposure: in the absence of treatment 55–80% of HIV-1 exposed infants remain HIV-1 uninfected. This is striking when the large volumes of maternal blood containing HIV-1-infected cells circulating through the placenta throughout gestation are considered: in fact, HIV-1 can be detected with relative ease in the placentas of both transmitting and non-transmitting mothers⁸. Moreover, infants breastfed by untreated HIV-1 uninfected mothers will ingest hundreds of litres of HIV-1-contaminated milk but over 80% of them remain uninfected⁹. It is generally assumed that there are no serious long-term consequences of exposure to HIV-1 in utero or early life in children who escape infection. Nevertheless, uninfected babies born to HIV-1-infected mothers suffer from increased morbidity and mortality^{10, 11}. It has been difficult to distinguish the direct impact of HIV-1 exposure from the consequences of being born to a sick mother¹¹, but there are some suggestions that exposed uninfected babies may present with severe infections indicative of clinical immunodeficiency¹². With the widespread roll-out of PTMCT programmes in resource-poor settings, this represents an important area for future investigation.

When transmission does occur, in utero (transplacental) transmission accounts for an estimated 5–10% of infections, peripartum transmission (occurring during labor, delivery and early breastfeeding) contributes 10–15% of infections, whilst breastfeeding transmission accounts for 5–20% of MTCT (reviewed in¹³). In the developed world, the successful deployment of intervention strategies has reduced the overall transmission rate to substantially less than 5%. Although antiretroviral regimens and risk reduction counseling have been successfully used for pregnant women and their infants in many parts of the developing world, full implementation of these programs remains a challenge in many countries, especially where antenatal clinical attendance and HIV-1 screening is not widespread. In addition, the potential toxicities of and the development of drug resistance to anti-retroviral therapy (ART) in both mother and child are concerns. Therefore, the development of a safe effective immunoprophylaxis regimen begun at birth and continuing during breastfeeding, perhaps alongside neonatal chemoprophylaxis, remains an area of active research interest.

The major risk factors for MTCT include factors associated with maternal viral load (duration and composition of ARV regimen and viral load at delivery) and infant exposure to infected fluids (duration of rupture of membranes, type of delivery, presence of other maternal sexually-transmitted infections (STI), and breastfeeding duration)^{14 15}. Breastfeeding considered alone is associated with a near doubling of the risk of MTCT of HIV¹⁶. Avoidance of breastfeeding and reduction of other maternal risk factors for MTCT of HIV have led to remarkable declines in MTCT in settings with adequate resources. However, replacement feeding is only advisable when the “AFASS” criteria are met: namely, that formula feeding be culturally Acceptable, Feasible for the mother, Affordable, Sustainable and Safe - conditions rarely met in most areas of the world. Thus an ideal pediatric vaccine for PMTCT would combine the immediacy of passive immunization designed to protect the infant during the first vulnerable weeks of life with the durability of active immunization to protect against the repeated low dose homologous virus exposure delivered multiple times a day via breastfeeding.

Why does such a large proportion of HIV-1-exposed uninfected children resist infection? In utero transmission is thought to occur most commonly in the last trimester; it has been suggested that the placenta blocks the transmission of free HIV-1 virions¹⁷ (by mechanisms as yet unidentified) and infection depends either on breaks in the placental barrier or the transcytosis of infected cells¹⁸. There are some data that suggest that uninfected infants exposed to HIV in utero, at delivery or post-natally can develop HIV-1-specific T-cell responses^{19, 20 21 22}, which suggests that there has been sufficient exposure to replicating virus to prime such a response. Although the detection of an HIV-specific response in the absence of persistent infection does not necessarily imply that T-cell immunity contributes to protection, resistance to post-natal HIV-1 transmission was shown to correlate with the magnitude of the HIV-1-specific T-cell response in a prospective study in Nairobi, Kenya⁹. This finding provides some encouragement that enhancing the immune responses to HIV-1 through immunotherapeutic strategies in uninfected infants could confer protection against later infection.

Vaccine strategies to prevent mother-to-child-transmission of HIV-1

HIV-1 vaccine studies in adults

Three large-scale human efficacy studies have been completed to date. In the first of these, Vaxgen employed an envelope sub-unit protein construct with the aim of inducing a protective antibody response: however, a large phase III clinical trial showed no evidence of efficacy²³. The field moved towards testing T-cell-inducing vaccines and considerable hope was placed on a replication-defective adenovirus type 5 (Ad5)-based recombinant vaccine produced by Merck that appeared to be the most immunogenic available construct in human studies. This vaccine was tested in the phase 2b STEP trial (tested in high risk volunteers), which was terminated prematurely as study subjects were neither protected from HIV-1 infection nor experienced reduced viral load when infection occurred^{24 25}. This failure led to a great deal of introspection in the HIV vaccine field, as it called into question whether or not a vaccine designed to elicit a cytotoxic T-lymphocyte (CTL) response could indeed protect against HIV-1 infection, as well as casting doubt on the non-human primate models of HIV infection that did not accurately predict the outcome of the human trial. However, progressive improvements in the constructs used to elicit T-cell responses have provided some encouragement in the macaque model: for example, a heterologous rAd26 prime/rAd5 boost vaccine regimen expressing SIV Gag elicited broader and stronger cellular immune responses than had been seen with the homologous rAd5 regimen, which led to significantly lower setpoint viral loads as well as decreased AIDS-related mortality compared with control animals following challenge with a pathogenic SIV strain, SIVmac251²⁶. More recently, a novel approach using rhesus CMV constructs induced effector memory T-cells at mucosal sites that correlated with protection from infection in four of twelve macaques repeatedly challenged with the highly pathogenic SIVmac239 strain²⁷. Further encouragement to the HIV vaccine field has come from the results of the rv144 phase III efficacy trial of a combination of canary pox priming and HIV envelope protein boost tested in a low-risk population in Thailand²⁸. Although the incidence of new infections was low in this study, meaning that significant efficacy was only seen in one of three analyses (the modified intention-to-treat analysis), the vaccine appeared to confer 31% protection against infection, the first indication of efficacy in any human study. The mechanisms of protection are not yet known, but are unlikely to include either neutralizing antibodies or cytotoxic T lymphocytes (CTL), which are rarely induced by this vaccine approach.

Mucosal vaccines

Mucosal immunization has been demonstrated to induce mucosal responses and as the first sites of simian immunodeficiency virus (SIV) and recombinant simian-human immunodeficiency virus (SHIV) infection in neonatal rhesus macaques are mucosal surfaces, similar tissues are likely to be exposed in MTCT of HIV-1^{29 30}. HIV-1 specific immunity present at these sites may be most relevant to preventing transmission. Many cellular and cytokine properties of the neonatal oral and intestinal tissues are known to differ from the adult: these could either represent protective or susceptibility factors and include the levels of γδ T cells, NK cells, macrophages, dendritic cells, IgG, IgA and secretory IgA (SIgA) levels, and levels of cytokines such as TNF-α and IFN-γ³¹.

Mucosal vaccines developed for PMTCT of HIV-1 could exploit the common mucosa-associated lymphoid tissue (MALT) to explore oral or nasal delivery. Oral delivery, although attractive for neonates, has challenges such as induction of tolerance, limitations in the choice of safe effective adjuvants, requirement for large doses of antigen and the need for antigen stability in the gut. This last concern has been addressed through the development of lipid vesicles or polymeric nanoparticles that act as immunostimulants while preserving immunogens from intestinal enzymes. So far only a limited number of orally administered vaccines against HIV have been tested in humans. An attenuated canarypox vector (vCP 205) and Salmonella vaccine vector (CKS257) vaccine platforms were both reportedly well-tolerated in humans but with less than expected mucosal immunogenicity^{32 33}. As opposed to oral vaccine delivery, the main advantage of nasally administered vaccine is the requirement for smaller doses of antigen. Several vaccine formulations have been tested in adults including peptides, DNA, and live bacterial and viral vectors³¹. Mucosal vaccine development against HIV-1 appears to require the deployment of stronger adjuvants, the use of which may be associated with safety concerns in young children.

In summary, a pediatric vaccine to prevent HIV infection would rapidly induce both antibody and cellular responses detectable at mucosal surfaces that would remain at effective levels during the duration of breastfeeding. A greater understanding of the most important inductive and effector sites in the newborn would guide research to address such questions as to whether systemic immunization can induce sufficient mucosal immunity, if oral or nasal vaccines can be effective, and if appropriate adjuvants can be developed for an effective vaccine development strategy for use in HIV-1-exposed infants.

Neonatal Immunity

In conjunction with the challenges of designing an effective vaccine to prevent transmission across mucosal surfaces, pediatric vaccinologists have the additional difficulty of working with a neonatal immune system in a state of extreme change as the newborn adjusts to life outside its intrauterine environment^{34, 35}. Although neonates generally develop immune responses upon immunization as well as cell-mediated responses to several acute viral infections, several deficiencies in the immunoregulatory pathways of T cells and antigen presenting cells have been documented in human cord blood with important implications for the development of immune-based therapies to prevent MTCT of HIV-1. Cord blood T cells have lower basal expression of CD3 and adhesion molecules, defects in cytokine production and CD8+ T-cell activity³⁶, while monocytes and dendritic cells express lower levels of costimulatory molecules, have altered differentiation pathways and reduced cytokine/chemokine production^{37, 38}. Neonatal T cells have a bias against Th1 cell polarizing cytokines that leaves the newborn susceptible to microbial infection and may contribute to the impairment of neonatal immune responses^{39, 40}. An exception is bacilli Calmette-Guerin (BCG) vaccination: newborns are capable of mounting a Th1-type response of similar magnitude to that given later in life⁴¹. Moreover, BCG itself can act as an adjuvant for other vaccines⁴². Pre- and post-natal exposure to environmental microbial products that activate innate immunity might accelerate this maturation, diminishing the Th2 and/or enhancing Th1 cell polarization, and this could be incorporated into adjuvant design. Both natural and inducible CD4+CD25+ regulatory T-cell (Treg) numbers are increased in infancy and are speculated to be required for maintaining peripheral T cell tolerance through inhibition of Th1 cell immunity^43–45. It has been noted that exposure to foreign antigens, both allo-antigens⁴⁶ and antigens from P. falciparum present in placental malaria⁴⁷, promotes the development of regulatory T-cell populations at birth.

The B-cell compartment is also affected in newborns, with responses to vaccines characterized by lower antibody levels, restricted diversity of antibody repertoire and lower levels of IgG2 isotype as compared with responses induced in adults⁴⁸. The implications of these deficiencies in pediatric vaccine development may include a requirement for enhanced stimulation of antigen-presenting cells with increased avidity of CD3/TCR and co-stimulation molecule interactions to ensure immune activation.

Vaccine strategies for PMTCT of HIV infection

Animal models, especially the rhesus macaque-SIV/SHIV model, have been used to address areas of uncertainty in pediatric vaccine design, including the most appropriate vaccine, the timing of immunizations, and the duration of vaccine elicited responses. The SIV models for vertical transmission of HIV-1 and for neonatal vaccine development were validated in the 1990s and many important proof of concepts have been demonstrated including protection of newborn macaques from oral SIV/SHIV challenge through maternal vaccination and through administration of passive hyperimmune serum (SIVIG/HIVIG) or neutralizing monoclonal antibodies at birth^49–53. The mechanism of protection from infection could include antibody binding alone, classic viral neutralization, or NK cell-mediated antibody dependent cellular cytotoxicity (ADCC). But will passive immunization to prevent MTCT of HIV-1 work? There are many challenges to this approach, particularly the composition of the antibody cocktail, which would need to be effective against more than one HIV clade and circulating recombinant forms, and the scale-up delivery logistics are daunting. Alternatively, boosting of maternal antibody levels by vaccination represents an important strategy to augment passive immunity during the infant's early months of life, until the infant can be actively immunized: however, the role of maternal antibody in reducing the effectiveness of immune responses in the newborn, especially T cell responses, is a concern. A phase III randomized clinical trial to compare the standard single dose mother/infant nevirapine regimen for PMTCT with the addition of HIV immune globulin (HIVIGLOB) or a second arm of extended infant NVP dosing compared with the standard single dose NVP regimen alone without HIVIGLOB was completed in Uganda in 2006 (http://www.mujhu.org/hiviglob2.html). The study enrolled 722 mothers with 204 in the HIVIGLOB arm. The data from the HIVIGLOB arm were pooled with other data from Johns Hopkins trials in Ethiopia and India, and results are not yet available.

Live attenuated vaccines have the advantage of prolonged antigen delivery and stimulation of both innate and adaptive immunity. Modified vaccinia virus Ankara (MVA) and canarypox viral vectors (ALVAC) have been tested in the rhesus macaque/SIV-SHIV models in advance of human trials. MVA expressing SIV gag, pol, and env or expressing SIVmac1A11was used to immunize infant macaques at birth and at 3 weeks of age⁵⁴. The infants were challenged at 4 weeks of age with uncloned SIVmac251, using a multiple low dose challenge model to deliver virus three times a day over five days to mimic breast milk exposure. The immunization regimen was unable to prevent infection, however the immunized infants mounted antibody responses and had improved clinical outcome compared to controls. An attenuated recombinant canary pox vector expressing SIV gag, pol, and env (ALVAC-SIV) was used to immunize infant macaques at birth, 2 and 3 weeks of age, followed by repeated oral low-dose challenge⁵⁵. In this experiment, significantly fewer immunized infants were infected (6/16) compared to the unimmunized controls (14/16), demonstrating that neonatal immunization provided partial protection from infection. Lastly, a topical DNA vaccine containing HIV-1 gag and env (DermaVir) represents an immunization strategy that targets lymph node dendritic cells (DC). Rhesus macaques immunized with DermaVir generated HIV-1 specific Th1 and Th2 cytokines and antigen specific memory T cells, while serum antibody levels were boosted after p27/gp140 protein boosting. Following mucosal challenge, none of the animals were protected from infection, however 4/5 immunized monkeys had reduced peak and set point viremia⁵⁶.

Pediatric Clinical Trials of HIV vaccines

The most recently reported clinical trials of live attenuated vaccine constructs tested as PACTG 326 part 1 and part 2 include the ALVAC constructs vCP205 and ALVAC-HIV vCP1452. Immunization of neonates was well tolerated and induced lymphoproliferative and/or cytotoxic T cell responses in vaccinees: ~40% of infants immunized with ALVAC vCP205 and 75% of infants immunized with ALVAC vCP1452^{57, 58}, ^{McFarland, 2006 #1314} An MVA-vectored vaccine is also currently under evaluation in an open randomized phase I/II study evaluating safety and immunogenicity of a candidate HIV-1 vaccine, MVA-HIVA, administered to healthy infants born to HIV-1 infected mothers in Nairobi, Kenya. This active study is aiming to enrol 72 HIV-1 uninfected infants with 36 breastfeeding infants and 36 formula feeding infants by the end of 2010, with infants in follow-up for 18 months. Within each feeding group, infants will be randomized to receive MVA.HIVA or to remain unvaccinated. The design of the study will allow for multiple secondary aims for comparison of immunogenicity in the different feeding groups and the response to other national immunization program vaccines in the MVA.HIVA vaccinated or unvaccinated infants (Tomas Hanke, personal communication).

Co-infections have the potential to alter vertical HIV-1 transmission and vaccine efficacy

HSV-2

Genital co-infections play an important role in both vertical and horizontal HIV-1 transmission. A study of HIV-1 discordant couples in Uganda demonstrated that the presence of genital ulcers increased the probability of HIV-1 transmission per coital act 2.6-fold⁵⁹. Herpes simplex type-2 infection (HSV-2) is the primary cause of genital ulcer disease (GUD) in resource-poor settings. HSV-2 seroprevalence is high among HIV-1 infected populations globally, and is likely to have a significant population impact on HIV-1 transmission in endemic regions^{60, 61}; a recent meta-analysis of studies estimated HSV-2 infection confers a 3-fold increased risk of sexual HIV-1 acquisition⁶². Maternal HSV-2 seropositivity has also been associated with increased rates of peripartum HIV-1 transmission in some studies^{63, 64}, but not others^{65, 66}. Recurrent subclinical reactivation of HSV-2, characterized by HSV-2 shedding in the absence of ulcers, is common in HSV-2 infection⁶⁷. The presence of genital ulcers, or genital shedding of HSV-2 proximal to delivery has been associated with an increased risk of vertical transmission of HIV^{14, 66}. Several mechanisms may explain this association. First, ulcers may provide a physical breach in the mucosa that increases infant exposure to HIV-1 virions or infected cells beneath the epithelium. HSV-2 ulcers contain high levels of HIV-1 RNA⁶⁸, and genital HIV-1 replication is a strong risk factor for transmission¹⁴. The concentration of HIV-1 in ulcers may result from the homing of activated CD4 T cells^{69, 70}, which are potential targets for HIV-1 infection. Even in the absence of ulcers, HSV-2 shedding is associated with increased HIV-1 shedding in the genital tract^{71, 72}. HSV-2 may also affect vertical transmission by more generalized effects on maternal systemic viral load; plasma HIV-1 viral load is increased during sub-clinical HSV-2 reactivation and declines with antiviral suppression of HSV-2 replication^{67, 73}. A randomized controlled trial (RCT) of valacyclovir versus placebo for 12 weeks demonstrated a 0.53 log₁₀ copies/ml decrease in HIV-1 plasma viral load, and a 0.29 log₁₀ decrease in HIV-1 RNA in the genital tract of HIV-1/HSV-2 co-infected women⁷⁴. The effect of HSV-2 co-infection on systemic HIV-1 replication may result from increased immune activation or direct interactions between viruses (reviewed in⁷⁵). Additionally, acyclovir has been shown to have an inhibitory effect on HIV-1 replication in vitro^{76, 77}. Together, these data suggested that HSV-2 suppression may be an effective tool to prevent HIV-1 transmission. However, results from clinical trials of sexual transmission have been disappointing; HSV-2 suppression with acyclovir did not prevent either HIV-1 transmission or acquisition^78–80. Importantly, acyclovir does not appear to affect egress of CD4 cells from lesion sites following healing; suggesting HIV-1 target cells remain concentrated in the genital mucosa⁸¹, and this may explain in part the failure of acyclovir to prevent HIV-1 transmission despite reductions in viral load. However, failure in the sexual transmission model may not exclude HSV-2 suppression as a strategy to reduce vertical transmission. A pharmacokinetic study showed acyclovir is actively transported to the amniotic fluid and breast milk⁸²; if HSV-2 suppression reduces HIV-1 viral load in these compartments, there is the potential for this approach to affect in utero or breast milk HIV-1 transmission. A randomized trial is currently underway in Kenya to address this question (NCT00530777).

Malaria

There is significant overlap between the malaria and HIV-1 epidemics (reviewed in ⁸³). Malaria co-infection is associated with a transient, but significant increase (~0.25 log₁₀) in plasma HIV-1 RNA viral load^{84, 85}, but its role in transmission is unclear⁸⁶. Ayouba and colleagues reported higher rates of MTCT in Cameroon during the rainy season, and speculated that malaria may affect vertical HIV-1 transmission⁸⁷. A few studies have demonstrated a trend for increased vertical transmission in women with blood parasitemia during pregnancy, however this relationship disappears after adjusting for HIV-1 viral load^{88, 89} and other studies have shown no association^{90, 91}. Placental malaria is more commonly detected in HIV-1 infected women^{88, 89, 92, 93} and Malawian women with placental malaria were found to have a ~2.5 fold higher plasma HIV-1 RNA viral load⁹⁴. Some studies have demonstrated an association between placental malaria and increased rates of vertical transmission independent of HIV-1 viral load^{89, 92}; however other studies have shown no association⁹³. A study conducted in Western Kenya found HIV-1 transmission risk to be increased when placenta parasitemia was high (>10,000 parasites/ml) but reduced in cases of low parasitemia (<10,000 parasites/ml) compared to malaria-negative controls⁹⁰. The potential protective effect of placental malaria was confirmed in a second study conducted in Mozambique⁹¹. These conflicting findings may be attributable to different methods of detecting placental parasitemia, different regional epidemics (seasonal, holoendemic), unrecorded self-treatment of malaria, and including breast milk HIV-1 transmissions in the ascertainment of effect. Malaria may alter the risk of HIV-1 transmission in utero by causing inflammation in the placenta⁹⁵, increasing CC-chemokine production^{96, 97}, shifting cytokine production from Th2 to Th1-type responses⁹⁸ and increasing CCR5 expression on Hofbauer cells⁹⁹. Together these data suggest malaria has the potential to alter the risk of vertical HIV-1 transmission, but further studies are needed to understand the interactions between immune responses to these two pathogens.

Independent of HIV-1 infection, malaria during pregnancy is associated with obstetrical problems and adverse birth outcomes, making malaria prevention, diagnosis and treatment an important component of antenatal care for all women in malaria endemic areas.

Bacterial Vaginosis (BV) and other vaginal infections

In addition to HSV-2 as discussed above, other sexually transmitted infections (STI) are also associated with an increased risk of horizontal HIV-1 transmission and acquisition (reviewed in¹⁰⁰). Ulcerative and non-ulcerative STIs, as well as vaginitis and cervicitis of unknown etiologies are associated with increased HIV-1 shedding in the female genital tract^100–105, and treatment of genital infections is associated with declines in HIV-1 shedding^{104, 106}. Genital infections are likely to increase HIV-1 transmission via recruitment of activated HIV-1 infected cells to sites of inflammation. Disruption of healthy vaginal flora due to bacterial vaginosis (BV)^107–109, or vaginal washing has also been associated with an increased risk of HIV-1 acquisition¹¹⁰. A recent prospective study conducted in Kenya reported a 3-fold increased risk of in utero HIV-1 transmission from women diagnosed with BV at 32 weeks gestation compared to women with normal vaginal flora¹¹¹. However, a multi-site RCT of metronidazole versus placebo in three African countries found no difference in rates of vertical transmission compared to placebo, despite a 16% reduction in BV¹¹². Further studies are needed to determine whether restoration of normal vaginal flora can reduce vertical HIV-1 transmission.

Helminth infection is associated with reduced vaccine responses

The greatest need for an HIV-1 vaccine is in Africa and Asia where helminths may infect 25–76% of the healthy adults and children^113–116. The first indication that helminths may compromise vaccine responses was the observation of reduced Bacillus Calmette-Guérin (BCG) efficacy in resource-poor countries^{117, 118}. Though helminths comprise a diverse group of organisms with heterogeneous routes and targets of infection, the most commonly encountered helminths (digenean flukes, cestodes, and nematodes) have similar effects on the host immune system; co-infection suppresses IFN-γ production and induces a Th2-type CD4 response in their hosts (reviewed in¹¹⁹). Following BCG vaccination, newborns whose mothers had schistosomiasis or bancroftian filariasis infection had weaker IFN-γ responses to purified protein derivative (PPD) and increased production of IL-5¹²⁰. This trend was also found at 14 months of age; PPD-specific CD4 responses in children exposed in utero to helminths had highly Th2-skewed responses, indicating that helminth exposure during priming of CD4 responses in utero affected the profile of memory cells later in life¹²⁰. De-worming prior to revaccination with BCG improves PPD-specific IFN-γ production and T cell proliferation in adults¹²¹. Similarly, IFN-g and IL-2 responses to cholera vaccination were improved by de-worming of patients with Ascaris lumbricoides infection¹²².

Helminths may also affect HIV-1 disease progression and acquisition. Suppression of Th1-type responses may directly impair control over viral replication, and helminth-induced systemic immune activation^123–125 may accelerate HIV-1 disease progression. The effect of helminths on the control over HIV-1 replication has been studied elegantly in vivo using primate models; following infection with SHIV clade C, Schistosoma mansoni-infected rhesus macaques had higher SHIV viral load compared with non-parasitized controls, and maintained higher IL-4 and IL-10 production¹²⁶. Introduction of S. mansoni into chronically SHIV-infected animals who were aviremic resulted in reactivation of SHIV replication and decline in CD4+CD29+ cells¹²⁷. These experiments also demonstrated that helminth infection may affect HIV-1 acquisition; S. mansoni co-infected animals became infected at lower SHIV doses compared with non-parasitized control animals, and had higher peak viral loads and proviral loads in central memory CD4 T cells¹²⁸. One study in humans has found an increased risk of vertical HIV-1 transmission in the setting of maternal helminth infection⁸⁸. To date, two RCTs have evaluated treatment of helminthic infections as a strategy to reduce HIV-1 disease progression. Treatment with albendazole improved CD4 counts in HIV-1 infected individuals with Ascaris lumbricoides infection, but no effect was found in subjects infected with hookworm or Trichuris trichiura¹²⁹. Similarly, Kallestrup and colleagues found improved control over viral replication and increased CD4 counts in subjects treated for Schistosomiasis infection¹³⁰. In sub-Saharan Africa, periodic de-worming is now a recommended component of comprehensive HIV-1 care of women and children.

Taking into consideration the local helminth burden of target populations will enable the more strategic design and deployment of an HIV-1 vaccine. De-worming of vaccinees prior to or during immunization may improve the priming of a Th1-type response. In mouse models, elimination of Schistosoma mansoni with praziquantel prior to immunization with a HIV-clade C DNA vaccine resulted in much improved vaccine-specific IFN-γ responses¹³¹. Additionally, infection with vaccinia virus expressing HIV-1 gp¹⁶⁰ resulted in stronger CTL responses and more rapid viral clearance in nonparasitized mice compared to S. mansoni infected mice¹³². Though de-worming may improve immune responses to vaccination, re-infection rates are high, and this would be expected to reduce the benefit of de-worming in vaccines requiring multiple doses or boosting. Continual re-infection with helminths may similarly impair the efficacy of a non-sterilizing therapeutic HIV-1 vaccine. An alternate approach is the design of vaccines targeted to overcome the Th2 bias induced by helminths, by the strategic use of adjuvants^{133, 134}, cytokines¹³⁵, or Toll-like receptors (TLR) signalling^{136, 137}.

Concerns regarding infant vaccine development

Many key questions regarding the development of a successful adult HIV-1 vaccine are equally valid for a neonatal immune-based intervention for PMTCT. However, pediatric vaccine development also faces a series of unique concerns. These include, but are not limited to, regulatory/ethical issues applicable to vulnerable populations, physiologic constraints of blood volumes that may limit the degree of safety and immunogenicity testing, existing immunization schedules and potential vaccine interference, and simultaneous exposure to both vaccine and pathogen in the presence of maternal antibodies and the developing neonatal immune system.

There is general agreement that pediatric clinical trials are necessary for implementation of novel medical strategies for prevention or treatment of childhood illnesses. In the US, several regulatory agencies have committed to ensure ethical conduct in research involving children and development of pharmaceutical formulations specific for children (www.fda.gov/cder/pediatric/). Although government programs and assurances provide a level of support to including infants in clinical trials of vaccines designed for PMTCT, in countries with historic imbalances in political and medical power, community-based opinion affects the support of local testing. Infants and children are subject to cultural traditions as well, for example within the family, who is empowered to grant consent? Who makes medical decisions? What is the relative value of family members? Clearly there is need to include diverse populations in testing vaccines, and resources invested in community sensitization and education during all stages of the trial may help ensure continued involvement in subsequent trials.

A logistical concern for phase I/II studies in infants is the frequency and volume of blood collection, not only the physiological limit of body size, but the psychological limit of the caregiver's tolerance to blood draws from healthy infants. Many assays and procedures have been optimized for minimal blood volumes, including Elispot assays for the detection of antigen-specific responses¹³⁸, multicolor flow cytometry¹³⁹ and detection of HIV-1 DNA in dried blood spots for monitoring infection status¹⁴⁰. Continued development of assays that require minimal volumes of blood will benefit pediatric clinical trials.

Currently infants are immunized worldwide against an array of infections delivered over the first 2 years of life, including BCG, polio, hepatitis B, diphtheria, pertussis, tetanus, pneumococcus and Hemophilus influenza b (Hib). Investigations into the sequence of exposure to murine viruses have demonstrated that the magnitude and specificity of the immune response elicited by the most recent infection are modified by the host's history of previous infections¹⁴¹. In newborn mice and humans Mycobacterium bovis bacillus Calmette-Guerin (BCG) immunization induces a potent immune response and this response has been shown to alter immunity to unrelated vaccines¹⁴². Also, individual components of multivalent vaccines may induce responses that differ when given individually or in combination (reviewed in¹⁴³). Therefore, the timing of introduction of new vaccines into the existing Expanded Program of Immunizations has the potential to modify responses to both previous and subsequent vaccines.

Maternal antibodies have been clearly shown to interfere with the effectiveness of measles virus vaccination, although not affecting the immunogenicity of other vaccines administered in the presence of maternal antibodies¹⁴⁴. Passively transferred maternal antibodies may form antigen-antibody complexes with vaccine antigens thereby limiting vaccine exposure prior to the development of an infant specific immune response, although there is evidence for many vaccines of B-cell priming with development of protective antibody titers after boosting¹⁴⁵. Antigen-antibody complexes are efficiently taken up via Fc-mediated mechanism by professional antigen presenting cells, which may facilitate the development of antigen specific T- and B-cell responses. However in the setting of PMTCT of HIV, potential exists for co-circulation not only of maternal antibody and vaccine constructs aimed at priming immunity in the infant, but also maternal antibody and cell-free HIV-1 acquired from breastmilk. Thus, Fc-mediated uptake of antigen-antibody complexes by APCs could potentially lead to enhancement of HIV infection (reviewed in¹⁴⁶).

Vaccine constructs and adjuvants may also react differently in infants, although thus far, recombinant HIV-1 gp120 delivered either in alum or MF59, and recombinant canarypox vectored vaccines for HIV have proven safe in pediatric populations (PACTG 230, 326, and HPTN 027), whilst the testing of adjuvant CRM₁₉₇ for use in multivalent pediatric vaccines has shown improve immunogenicity of certain vaccines¹⁴⁷. Live attenuated SIV vaccines tested in neonatal macaques have also proved safe, with the rare exception of a multiply-deleted SIVmac239 that when administered to neonates, showed unexpected pathogenicity not initially observed in adults. Pathogenesis in adults was later documented in ~25% of vaccinated adults after a median of ~3 years of infection¹⁴⁸.

These concerns can be addressed readily through investment in the use of animal models, of continued testing in adults, and of basic research into underlying mechanisms of neonatal immune regulation, maternal antibody interference, and vaccine interference.

Conclusions and future directions

Although the HIV vaccine field still has some way to go before an effective vaccine to prevent infection becomes available, the special issues of mother-to-child transmission and infant immunisation deserve further study. We have highlighted the approaches tested to date, as well as highlighting the potentially modifiable infectious co-factors that can facilitate transmission of HIV-1 from mother to child in the developing world and commenting on some of the issues that will need to be considered in the development of a pediatric HIV vaccine. Even if deployment of strategies to prevent vertical transmission of HIV-1 becomes universal, a scenario that currently seems some distance away in resource-poor settings, it is very likely that a prophylactic HIV vaccine will ultimately need to be given as part of the Extended Programme of Immunisation (EPI). Therefore a better understanding of infant immune responses to candidate vaccine antigens and adjuvants is an important area for future investigation.