The role of infant immune responses and genetic factors in preventing HIV-1 acquisition and disease progression

INTRODUCTION

Mother-to-child HIV-1 transmission accounts for more than 700 000 new paediatric HIV-1 infections in developing countries each year [1]. This comprises less than one-third of the infants born to human immunodeficiency virus type 1 (HIV-1) infected mothers, the majority of whom remain uninfected despite recurrent risk for contact with the virus in utero, during delivery and through breastfeeding. A comprehensive approach to studying infant immunity against HIV-1 may provide insight into the determinants of HIV-1 acquisition, promote an understanding of resistance to infection in the setting of repeated exposure to the virus and contribute to the development of therapeutic interventions or vaccines against HIV-1 transmission.

Several unique features of mother-to-child HIV-1 transmission provide advantages in determining correlates of HIV-1 acquisition and viral immunity when compared to sexual HIV-1 transmission models. Both HIV-1 infected mothers and their exposed infants can be evaluated for viral and immunological factors associated with transmission. HIV-1 exposure can be characterized by quantifying maternal HIV-1 viral load in plasma, breast milk and genital tract secretions, and infant immune responses can be defined simultaneously or near the time of exposure. Timing of transmission can be estimated using HIV-1 polymerase chain reaction (PCR) at birth and at regular intervals after exposure during delivery and breastfeeding.

Vertical HIV-1 transmission risk is also higher than sexual transmission risk. Per sexual act, it is estimated that the risk of heterosexual transmission is approximately 0·1% in an antiretroviral naive population [2]. The risk of HIV-1 acquisition during delivery ranges from 10 to 20%, more than 100-fold higher than heterosexual transmission rates. Disparities between heterosexual and vertical HIV-1 transmission rates persist in the setting of antiretroviral therapy. This enables mother–child transmission studies to provide robust epidemiological data regarding specific immune mechanisms and combinations of immune responses that may constitute protective immunity against HIV-1.

This review examines the spectrum of innate, humoral and cellular immune responses and genetic factors that have been studied in infants who are HIV-1 infected or HIV-1 exposed and uninfected (Fig. 1).

Fig. 1.

Immune responses and genetic factors associated with mother-to-child HIV-1 transmission and paediatric HIV-1 disease progression. *May be maternally acquired in utero or via breast milk. †Alloimmunity is dependent on degree of maternal–infant HLA mismatch.

INNATE IMMUNITY

Innate immune responses are generated rapidly and are important in preventing and containing infections with a variety of viral pathogens. Broad innate immunity may also be capable of protecting against immune-escape viruses generated by more narrow adaptive immune responses. In HIV-1 transmission and disease progression, relevant innate mechanisms of immunity include the activity of natural killer (NK) cells and antiviral proteins such as the CC chemokines, CD8⁺ antiviral factor (CAF) and secretory leucocyte protease inhibitor (SLPI).

Natural killer (NK) cell activity

Natural killer (NK) cells induce inflammation and lyse infected cells without prior sensitization and in a non-HLA restricted manner. NK cells from HIV-1 infected individuals release the CC chemokines MIP-1α, MIP-1β and RANTES, three factors that have been shown to inhibit HIV-1 independently in vitro by blocking the CCR5 HIV-1 coreceptor [3,4]. NK cells also act by lysing HIV-1 infected cells via antibody-dependent cellular cytotoxicity (ADCC). This is initiated by binding of NK cell Fc receptors (CD16) to target cells coated with HIV-specific antibodies of the subclass IgG1 [5–7]. HIV-specific ADCC antibodies are directed against the viral envelope glycoproteins gp120 and gp41 and are distinct from virus-neutralizing antibodies [8].

There is conflicting evidence regarding the role of NK cells in containing HIV-1 in chronically infected children and in preventing vertical HIV-1 transmission. Several studies have evaluated HIV-specific ADCC antibody titres in sera of infants born to HIV-1 infected mothers and found that these antibodies are transferred efficiently across the placenta from mother to fetus [9,10]. However, there was no significant correlation between antibody titres at birth and either HIV-1 disease progression during 2 years of follow-up or mother-to-child HIV-1 transmission [9,10]. Active production of HIV-specific ADCC antibodies was observed in the majority of HIV-infected infants only after 12 months of age [10] and effector cells from HIV-1 infected children appear unable to generate NK cell-mediated cytotoxicity [11]. Thus, an immature immune system may account for the absence of ADCC-mediated NK protection against HIV-1 infection in neonates and young infants, despite adequate levels of passively transferred ADCC antibodies. This may contribute to rapid HIV-1 progression in children infected with HIV-1 early in life [10,11].

Non-cytotoxic T cell activity

In addition to mediating HLA-restricted cytolytic activity, CD8⁺ T lymphocytes can suppress HIV-1 in vitro by secreting a soluble factor or collection of factors. These non-entry inhibitors, known as CD8 antiviral factors (CAF), can block viral replication of both R5 and X4 viruses by inhibiting transcription regulation at the HIV-long-terminal repeat (LTR) [12,13]. CAF appears to be distinct from the CC chemokines but may be related to other known factors, such as the α-defensins-1, −2 and −3, and these may contribute in part to its anti-HIV-1 activity [14,15].

The presence of CAF in plasma has been associated with delayed HIV-1 disease progression in several adult cohorts [12,14,16,17]. The role of CAF in protecting against paediatric HIV-1 disease progression and mother-to-child HIV-1 transmission has been investigated ess extensively. Infants are capable of CAF production and this may be stimulated by exposure to HIV-1 in utero and during delivery. In one study, anti-HIV activity attributed to CAF was detected in 16 (52%) of 31 HIV-1 uninfected infants born to HIV-1 seropositive mothers and in none of the 12 control infants born to HIV-1 uninfected mothers [18]. Additional studies will be necessary to define the contribution of CAF to preventing HIV-1 disease progression in children and protecting against HIV-1 transmission in mother–infant cohorts. In both adults and children, CAF holds promise for new therapeutic and immune strategies that mimic its action or promote secretion of CAF factors.

Secretory leucocyte protease inhibitor (SLPI)

Endogenous proteins in saliva, genital secretions and breast milk may provide protection against mother-to-child HIV-1 transmission. Several soluble components of saliva have been demonstrated to have antiviral activity, including lysozyme, cystatins, lactoferrin and secretory leucocyte protease inhibitor (SLPI) [19,20]. Among these, only SLPI inhibits viral replication effectively at physiological concentrations. SLPI is a 12 kilodalton non-glycosylated protein that is secreted by acinar cells of submucosal glands and acts by targeting a host cell protein rather than by interacting with viral proteins (gp120, gp160), transcriptases or proteases [21–24]. One hypothesis is that SLPI stabilizes the host cell membrane after binding to a SLPI binding protein, thus inhibiting HIV fusion and preventing subsequent viral entry into host cells [25].

Three studies have evaluated the protective effect of maternal SLPI in preventing mother-to-child HIV-1 transmission [26–28]. SLPI levels in infant saliva were investigated in a mother–child cohort in Kenya and found to protect against HIV-1 exposure via breastfeeding [26]. In a second study in the Central African Republic, no differences were found when SLPI levels in colostrum and breast milk were compared for transmitting and non-transmitting mothers [27]. A third study, conducted in South Africa, evaluated SLPI levels in vaginal fluid at 28–32 weeks’ gestation and found a significant correlation between higher SLPI levels in vaginal fluids and decreased mother-to-child HIV-1 transmission [28]. The results of these studies are intriguing and suggest that SLPI in vaginal secretions and saliva is an important innate mechanism of defence against HIV-1 infection that may be used effectively for HIV-1 treatment or prevention.

Chemokines

HIV-1 replication is suppressed in vitro by the CC chemokines, MIP-1α, MIP-1β and RANTES, and the CXC chemokine SDF-1 [29–31] when these natural ligands bind to CCR5 and CXCR4 cell surface receptors and block or down-regulate coreceptors utilized by HIV-1 [32–35]. While there has been controversy regarding the role of chemokines in vivo, the majority of clinical studies among HIV-1 infected adults suggest that increased MIP-1α, MIP-1β and RANTES protect against progression of HIV-1 to clinical AIDS [36–40]. In paediatric HIV-1 infection, a positive correlation between CC chemokine levels and slow disease progression has also been observed [41].

These studies have encouraged investigators to explore the clinical application of chemokine-based therapies. These include the use of vaccines to increase production of chemokines and the development of antibodies or drugs that block HIV-1 entry or mimic the action of CCR5- and CXCR4-binding chemokines. Vaccines that induce chemokine expression result in down-regulation or blockade of important HIV-1 co-receptors and this may complement HIV-1 specific cellular and humoral protection [42]. Pharmacological or antibody-mediated blockade is another mechanism for down-regulating CCR5 and CXCR4 receptors. In vitro studies have shown that antibodies to these important HIV-1 co-receptors inhibit HIV-1 entry into cells, have a long half-life in vivo, and are able to cross the placenta [43,44]. In addition, inhibition due to chemokines has potential for being effective across subtypes because different HIV-1 strains use the same chemokine receptors to enter cells. This provides distinct advantages over other therapeutic modalities associated with viral mutations and resistance.

CC chemokines may also modulate mother-to-child HIV-1 transmission risk. Infants born to HIV-seropositive mothers who remain uninfected during follow-up have significantly higher levels of RANTES production from cord blood PBMCs than infants with perinatal HIV-1 infection [45]. CC chemokines may influence mother-to-child HIV-1 transmission risk by influencing HIV-1 replication in other maternal and infant compartments. Breast milk, cervicovaginal lavage samples, periodontal tissue and human placental tissue have been demonstrated to contain detectable levels of chemokines [46–49].

HUMORAL IMMUNITY

The humoral arm of the adaptive immune system plays an important role in preventing infection after viral exposures. However, a protective humoral immune response against HIV-1 has been difficult to characterize. It is generally accepted that the generation of high titre antibodies that neutralize HIV-1 is a desirable component of an effective HIV-1 vaccine. However, epitopes required to create an antibody response able to broadly neutralize virus are not characterized easily and HIV-1 mutates rapidly against most HIV-1 specific antibodies [50,51].

In addition to systemic humoral immunity, an effective humoral response at susceptible mucosal surfaces may protect against vertical HIV-1 infection. As infants pass through the birth canal and breastfeed, their oral, nasal, gastrointestinal and conjunctival mucosa are exposed to maternal genital secretions, blood and breast milk contaminated with HIV-1. Immunoglobulins, and in particular secretory IgA, may be critical for protection against viral infection at these sites and may play a similar role in preventing mother-to-child HIV-1 transmission.

HIV-1 specific IgG

Maternal IgG with specific activity against HIV-1 is acquired passively by the infant while in utero without selective antibody transfer [52]. Median time to loss of antibody is approximately 10 months and the majority of infants lose maternal IgG by 18 months of life [53]. In large cohort studies maternal HIV-specific IgG has not been associated with protection against mother-to-child HIV-1 transmission [54,55]. In the early 1990s, several studies reported that there was no difference in levels of maternal antibody to the third hypervariable region of gp120, one of the principal HIV-1 neutralization domains, between HIV-1 transmitting and non-transmitting mothers [56–58]. Later studies demonstrated that a high titre antigp160 response and high plasma virus load were independent risk factors for perinatal transmission of HIV-1 [55]. The increased risk of mother-to-child HIV-1 transmission with high anti-HIV antibody titres may be due to confounding, because women with high plasma viral load have high levels of anti-HIV antibodies [54].

Trials using hyperimmune serum containing virus-specific IgG have been conducted to determine whether passively acquired antiviral antibodies modulate virus transmission and disease progression. In the macaque model, simian immunodeficiency virus hyperimmune serum (SIVIG) given subcutaneously prior to oral SIV inoculation has been shown to protect newborns against infection [59], and when administered during early infection SIVIG has been associated with delayed disease progression in infant macaques [60]. These results suggest that passively acquired anti-HIV IgG may decrease perinatal HIV infection and may be an effective intervention.

The role of hyperimmune IgG has also been studied in a paediatric clinical trial [61]. The Pediatric AIDS Clinical Trials Group protocol 185 evaluated whether HIVIG infusions administered monthly during pregnancy and to the neonate at birth would significantly lower perinatal HIV transmission rates when added to zidovudine administered per the ACTG 076 regimen. This study did not demonstrate a protective effect for HIVIG; however, low rates of HIV-1 transmission in the setting of zidovudine prophylaxis limited the study's ability to address whether passive immunization can diminish perinatal transmission. Additional studies are planned and may show that there is a benefit to using HIVIG in developing countries where single-dose nevirapine or less aggressive zidovudine regimens are the standard of care and breastfeeding contributes many HIV-1 transmission events during the first 6 weeks.

Neutralizing antibodies

Protection against vertical HIV-1 transmission correlates with viral neutralization activity of HIV-1 specific antibodies [56,62–64]. Several studies have examined the ability of sera to neutralize its own virus (autologous neutralization) and virus from other mothers (heterologous neutralization) [62,63]. Non-transmitting mothers had neutralizing antibodies against autologous virus more frequently than transmitting mothers. In addition, all mothers with autologous neutralizing antibodies also neutralized at least two heterologous primary isolates. This provides evidence that broad neutralizing antibody responses contribute to reducing the risk of mother-to-child HIV-1 transmission.

There are several ways for an HIV-1 exposed fetus or neonate to benefit from neutralizing antibodies, including transplacental transfer and the administration of HIV-1 neutralizing antibodies via injection at the time of birth. Transplacental antibody transfer necessitates an effective neutralizing response in the pregnant woman, one that could be endogenous or vaccine-induced. Vaccination of HIV-1 infected pregnant women in the second and third trimesters has not been associated with changes in HIV binding and neutralization antibodies [65]. Passively administered monoclonal antibodies, on the other hand, hold promise for effective prevention of perinatal and breast milk HIV-1 transmission. Both pre- and postnatal treatment with a combination of three human neutralizing monoclonal antibodies have been shown to protect neonatal macaques from mucosal challenge with a simian-human immunodeficiency virus construct (SHIV)-vpu(+) [66]. In a subsequent study the same monoclonal antibody combination was used postnatally, thereby reducing significantly the quantity of antibodies necessary and rendering their potential use in humans more practical [67]. Two neonatal macaques treated with this regimen were protected against oral SHIV-vpu(+) challenge, while four untreated control animals became persistently infected [67].

To date, there have not been human trials to evaluate the safety or efficacy of passively administered human monoclonal antibodies. Monoclonal antibodies are costly and may be challenging to produce, but they are poised to play an important role in the prevention of mother-to-infant HIV-1 transmission. They may be particularly useful in settings where HIV-1 infected women are not identified until the time of delivery and therefore unable to benefit from antenatal interventions to decrease HIV-1 transmission.

Mucosal immunoglobulin A (IgA)

Infants are exposed to HIV-1 during delivery and through breastfeeding primarily via oropharyngeal and gastrointestinal mucosa. Secretory IgA may contribute to preventing HIV-1 infection at these mucosal surfaces by neutralizing virus. HIV-specific salivary IgA has been found in HIV-1-exposed, uninfected adults [68] and has been shown to neutralize different HIV-1 subtypes [69].

Studies have demonstrated that an HIV-specific IgA response can be elicited at oral and other mucosal surfaces with vaccines administered either via mucosal or systemic routes [70–72]. This HIV-specific IgA may be capable of neutralizing HIV-1 in vitro[71]. Research directed towards improved methods of induction of HIV-1 specific humoral immune responses at oral or genital mucosa is likely to have a substantial impact on future vaccine development because the vast majority of HIV-1 transmission occurs across mucosal surfaces [51,73].

CELLULAR IMMUNE RESPONSES

In concert with humoral immune responses, virus-specific cell-mediated immunity is responsible for containing and clearing viral infections. In HIV-1 infection, cellular immune responses contribute to controlling early viraemia and induction of these responses has been a target for vaccines designed to prevent the establishment of HIV-1 infection and disease progression. Cellular immunity against HIV-1 includes HIV-specific CD8⁺ cytotoxic T lymphocytes (CTLs), CD4⁺ T helper cells and natural killer (NK) cells. CD8⁺ and CD4⁺ T cell activity play an important role in acute and chronic HIV-1 infection and it is possible that these responses contribute significantly to protection against HIV-1 transmission.

CD8⁺ cytotoxic T lymphocytes (CTLs)

Acute HIV-1 infection in adults is marked by a rise in CD8⁺ cytotoxic T lymphocytes (CTLs) and control of viraemia before the onset of production of neutralizing antibodies [74,75]. CTL responses, although detected less consistently in acutely HIV-infected infants than in older children and adults [76–80], are considered to be important markers of early HIV-1 infection [77,78,81,82] (Table 1). Infant CTL responses may also delay HIV-1 disease progression [41,83,84] and in chronic HIV-1 infection, HIV-specific CTLs have been shown to correlate positively with CD4⁺ T cell numbers and inversely with HIV-1 RNA viral load, an association that was independent of the degree of immunosuppression [80,83]. These and several other studies that found no association between CTL activity and disease progression had relatively small sample sizes and included samples from children spanning a wide range of ages (Table 1).

Table 1.

Cellular immune responses in HIV-1 Infected infants and children: prevalence of responses and associations with HIV disease progression

Author, year	Age range	Prevalence of response		Protection against HIV-1 progression
HIV-1-specific CTL responses
Froebel, 1994 [79]	<1 month−8 years	6/9	(67%)	No
Luzuriaga, 1995 [78]	4–42 months	4/7	(57%)	No
Pikora, 1997 [77]	Cord blood:19 months	4/4	(100%)	Yes
Wasik, 1997 [90]	7–43 months	8/11	(73%)	Yes
Buseyne, 1998 [84]	4–21 months	5/10	(50%)	No
Wasik, 1999 [45]	Cord blood:14 months	4/7	(57%)	NA
Spiegel, 2000 [81]	1 month−13 years	NA/82	NA	Yes
Wasik, 2000 [41]	<1 month−5 years	NA/12	NA	No
Scott, 2001 [82]	<6 months	2/13	(15%)
	>6months	4/4	(100%)	No
Buseyne, 2002 [83]	5–16 years	env 21/47*	(45%)	Yes
		gag 29/47	(62%)	No
Buseyne, 2002 [134]	3–17 years	18/20	(90%)	Yes
Sandberg, 2003 [80]	1–16 years	55/65	(85%)	Yes
Summary	Cord blood: 17 years	139/197	(71%)	No: 6 studies
				Yes: 5 studies
HIV-1-specific T helper responses
Wasik, 1997 [90]	7–43 months	7/11	(64%)	Yes
Wasik, 1999 [45]	Cord blood-:14 months	4/7	(57%)	No
Wasik, 2000 [41]	<1 month−5 years	na/12	na/12	Yes
Kuhn, 2001 [102]	Cord blood	3/6	(50%)	No
	6 months	3/4	(75%)	No
Kuhn, 2001 [103]	3 months−11 years	13/49	(27%)	Yes
Summary	Cord blood: 11 years	30/77	(39%)	No: 2 studies
				Yes: 3 studies

NA = not available.

^*Only gag responses were included in summary to avoid repeated measures on the same children.

The role of infant HIV-1 specific CTL responses in preventing vertical HIV-1 transmission is less well established. The fetus is capable of making mature CTL responses against specific pathogens, such as cytomegolovirus [85]. The evidence in favour of a protective role for CTLs produced in utero and during early infancy is derived from investigations detecting HIV-1 specific responses in HIV-1 exposed infants who do not become infected [86–91] (Table 2). In these studies, CD8⁺ T cell responses specific to HIV-1 were found in peripheral blood and in cord blood obtained from HIV-1 exposed, uninfected infants. Overall, six (75%) of eight studies detected CTL responses in HIV-1 exposed, uninfected infants and 17 (25%) of the 69 infants studied had positive responses (Table 1). Small sample sizes and the use of a variety of assays with different levels of sensitivity may account in part for the wide range (0–100%) of positive HIV-1 specific CTL responses observed in different cohorts.

Table 2.

Cellular immune responses in HIV-1 uninfected infants and children with HIV-1 exposure

Author, year	Age range	Prevalence of response
HIV-1-specific CTL responses
Cheynier, 1992 [86]	2–35 months	3/3	(100%)
Buseyne, 1992 [87]	11–36 months	0/4	(0%)
Rowland-Jones, 1993 [88]	Cord blood: 13 months	1/1	(100%)
Aldhous, 1994 [135]	6–25 months	2/11	(18%)
De Maria, 1994 [136]	12–50 months	7/23	(30%)
McFarland, 1994 [89]	6–23 months	2/8	(25%)
Luzuriaga, 1995 [78]	1–18 months	0/10	(0%)
Wasik, 1999 [45]	Cord blood to 14 months	2/9	(22%)
Summary	Cord blood to 36 months	17/69	(25%)
HIV-1-specific T helper responses
Clerici, 1993 [104]	Cord blood	8/23	(35%)
Wasik, 1997 [90]	5 months−5 years	0/5	(0%)
Wasik, 1999 [45]	Cord blood to 14 months	2/9	(22%)
Kuhn, 2001 [102]	Cord blood*	30/80	(38%)
	6 months	23/41	(56%)
Summary	Cord blood to 5 years	40/117	(34%)

^*Only cord blood responses included in summary to avoid repeated measures on the same children.

Evidence that CTL responses protect against vertical HIV-1 transmission also comes from the finding that HIV-1 exposed infants are more likely to become infected with a viral strain that has escaped maternal CTL responses [92,93]. Studies evaluating CTLs in HIV-1 exposed, uninfected infants are ongoing to determine whether CTL responses protect against establishment of HIV-1 infection or serve merely as a marker for HIV-1 exposure. Additional research in this area will contribute data from a larger numbers of mother–infant pairs and may help to establish the clinical significance of these findings.

The majority of vaccines currently in clinical trials are designed to promote HIV-specific cellular immune responses [94,95]. Several of these candidate HIV-1 vaccines have elicited CTL responses successfully in HIV-1 uninfected adults [96,97], but few studies have looked at the immunogenicity of HIV-1 vaccines in infants or young children. In one study among HIV-1 infected infants with asymptomatic disease [98], infants immunized with a recombinant HIV-1 glycoprotein vaccine were significantly more likely to develop a CTL response than controls, supporting the immunogenicity of the vaccine in a paediatric population. It is not known whether this will translate into a protective cellular immune response for HIV-1 uninfected infants immunized with this vaccine and similar vaccine constructs.

CD4⁺ T cell responses

There is substantial evidence that HIV-specific CD4⁺ T helper cells contribute to control of viraemia in adults [99–101]. In HIV-1 infected infants, HIV-specific T helper responses have been detected and associated with non-progression (Table 1) [41,45,90,102]. Infants were found in one study to progress six times more rapidly to AIDS and death in the absence of production of HIV-specific interleukin-2 (IL-2), one of the primary cytokines secreted by CD4⁺ T cells [103]. CD4⁺ T cell activity has also been associated with early HIV-1 specific CTL responses in infants with delayed HIV disease progression [41].

Like CTL responses, HIV-specific CD4⁺ T helper responses have been reported in cord blood and peripheral blood from HIV-1 exposed, uninfected infants, suggesting that HIV-specific T helper cells contribute to preventing mother-to-child HIV-1 transmission. These infants did not acquire HIV-1 despite exposure to maternal virus in genital secretions, blood and breast milk [45,102,104]. In the largest of these studies, investigators found that a high proportion (approximately 40%) of HIV-1 exposed infants had CD4⁺ T cell responses in cord blood [102] that were associated with significant protection against HIV-1 acquisition during delivery and breastfeeding. These responses are reduced in mother–infant pairs receiving perinatal antiretrovirals for prevention of HIV-1 transmission [105]. More research into the immunological consequences of maternal antiretroviral therapy will help to clarify associations between this important intervention and perinatal and breast milk HIV-1 transmission.

It is unknown whether infants are capable of mounting CD4⁺ T helper responses similar to those demonstrated by adults after immunization with an HIV-1 vaccine [106]. Wasik and others have demonstrated a strong association between HIV-specific CTL activity and CD4⁺ T helper cell responses in infants [41,101]. The contribution of HIV-specific CD4⁺ T helper activity to eliciting and maintaining CTL responses may therefore be necessary to induce an effective response to a vaccine and requires further investigation.

GENETIC FACTORS

Genetic polymorphisms in HIV-1 co-receptors and human leucocyte antigens (HLA) influence susceptibility to HIV-1 infection and may determine the quality of the antiviral immune response irrespective of subtype or strain specificity. Greater understanding of these genetic factors may lead to the development of vaccines and pharmacological agents with cross-clade activity, and thus wider applicability.

HIV-1 co-receptor and chemokine polymorphisms

Some of the most compelling evidence that chemokines are linked to mother-to-child HIV-1 transmission and disease progression comes from associations of chemokine receptor polymorphisms with protection from HIV-1 acquisition. The most thoroughly investigated CCR5 receptor polymorphism is a natural 32-base pair deletion (CCR5Δ32) that results in a truncated protein not expressed on the cell surface [107]. Individuals who are homozygous for this mutation have been reported to be highly resistant to infection with R5 HIV-1 [108,109].

HIV-1 infected women who are heterozygous for CCR5Δ32 have decreased transmission to infants compared to HIV-1 infected mothers with wild-type CCR5 receptor genes [110]. Polymorphisms within the promoter region of CCR5 are found relatively frequently in African individuals and in a perinatal study among HIV-1 infected breastfeeding Kenyan women, heterozygosity for the 59356 C/T genotype was associated with increased mortality among women during postpartum follow-up [111]. However, there was no association with mother-to-child HIV-1 transmission or infant disease progression.

Much less well understood are the SDF-1 3′A and RANTES promoter polymorphisms [112–115]. The SDF-1 polymorphism, SDF-1 3′A, is a G to A change located in the 3′ untranslated region of an alternately spliced mRNA transcript that is common in African and other populations [112,113]. In a study in Kenya, the SDF-1 3′A allele was present in approximately 10% of pregnant HIV-1 seropositive study participants and heterozygosity for the SDF-1 3′A polymorphism was associated with increased late breast milk HIV-1 transmission [113]. Among 194 HIV-1 infected children in Italy, the SDF-1 3′A mutation was significantly correlated with accelerated disease progression [116]. In this study, 25% of long-term non-progressors and 42% of rapid progressors were heterozygous for SDF-1 3′A; none of the non-progressors were homozygous for the polymorphism [116].

The RANTES haplotypes G1 and G4 have two single nucleotide polymorphisms in the promoter region. In adults, these polymorphisms have been associated with slower HIV-1 disease progression and with an altered risk of HIV-1 acquisition, particularly in people who lack the CCR5Δ32 allele [114,115]. The effect of these polymorphisms on vertical HIV-1 transmission or HIV-1 infection in infants has not been defined.

Human leukocyte antigen (HLA)

Human leucocyte antigen (HLA) haplotypes can determine which HIV-1 epitopes bind and how effectively those epitopes are presented by class I and class II HLA to CD8⁺ T cells (CTLs) and CD4⁺ T cells (T helper), respectively. Thus, by presenting conserved immunogenic epitopes, specific HLA types may promote an immune response that protects against HIV-1 transmission or contains chronic HIV-1 infection [117].

Several class I and class II HLAs have been associated with an altered risk of HIV-1 disease progression in several different populations [118–128]. HLA homozygosity at class I loci has also been associated with more rapid disease progression in adults [118], presumably because homozygotes present a less diverse collection of epitopes to the immune system than heterozygotes. In children, more rapid HIV-1 disease progression has been associated with HLA A*23 and the class II allele DRB1*03, with slower progression associated with HLA DRB1*13 [125,129].

Associations between specific HLAs and HIV-1 mother-to-child transmission, discussed in a review by Goulder et al. [130] have been described predominantly in studies conducted in non-breastfeeding infants in Europe and North America [124–128]. HLAs DQB1*06 and DRB1*03 were associated with increased HIV-1 transmission, and several DQB1 alleles were associated with decreased transmission [124–128]. Limited research has been conducted on HLA associations and vertical HIV-1 transmission among breastfeeding populations from developing countries. In an African breastfeeding cohort [131,132], MacDonald and colleagues found that a group of functionally related HLA subtypes called the A2 supertype was associated with significantly decreased risk of infant HIV-1 infection during the first 6 months [132]. This study did not find protection against breast milk HIV-1 transmission with A2 supertype alleles, thus raising intriguing questions about how the nature and timing of exposure affects correlates of transmission.

In addition to examining the role of individual alleles, investigators have evaluated the role of HLA concordance between mother and child in vertical HIV-1 transmission [131]. These studies have determined that there is increased risk of transmission for infants who share a greater number of alleles with their mother. HLA concordance may permit transmission of escape mutants that have evaded a cellular immune responses shared by the mother and infant. Alloantibodies may also protect infants from HIV-1 infected cells from a mother with a different HLA type but this has not been confirmed [133].

Vaccine trials provide a unique opportunity to evaluate the contribution of HLA to cellular immunity against HIV-1. The distribution of HLA types in a population may also be relevant to the design of an effective preventive or therapeutic vaccine. Specific HLA haplotypes have been associated with responses to HIV-1 vaccines in adults [122]. Similar studies in paediatric populations may increase understanding of the role of HLA in modulating infant immune responses and mother-to-child HIV-1 transmission risk.

CONCLUSION

There are many potential correlates of infant immunity that are associated with altered risks for mother-to-child HIV-1 transmission and paediatric HIV-1 disease progression, making it likely that a combination of humoral, cellular and innate responses is responsible for prevention of HIV-1 acquisition in infants (Fig. 1). Neutralizing antibodies and several different components of the innate immune system are among the most promising candidates for a broad and effective immune response. Researchers have demonstrated that passive administration of monoclonal antibodies capable of neutralizing HIV-1 in vitro can prevent infection in animal models and these are moving toward trials in human paediatric populations. Data presented in this review also support a role for cellular immune responses in controlling viral replication and indicate that CD8⁺ or CD4⁺ T cell responses may provide protection against vertical HIV-1 transmission. Several innate immune responses, including CC chemokines and SLPI, have been associated with altered transmission risk and offer new targets for vaccines and therapeutic agents. Some studies have demonstrated synergy between innate and acquired immune responses. These findings are intriguing and suggest that future vaccines designed to induce innate immunity in conjunction with adaptive immunity may provide additional benefits.

Research using a mother-to-child transmission model may help to further define these immune responses and characterize synergistic interactions among them. This research will need to take into consideration the different types of transmission (in utero, intrapartum, breast milk) and their relative contribution to risk of HIV-1 infection and induction of protective immunity. For example, exposure producing effective immune responses in utero may not translate into protection against intrapartum or breast milk transmission. The age of the infant or child and the timing of infection are also important considerations since these may determine whether or not a child is able to generate an immune response. A more technical limitation of this model is the difficulty obtaining large volume blood samples from young infants for comprehensive immunological testing.

The challenges of conducting vertical HIV-1 transmission research are balanced by several factors that make this an extremely valuable model for determining protective immune correlates. These include relatively high HIV-1 transmission rates from mother-to-child, the ability to sample both mother and infant near the time of exposure, and relative precision regarding timing of infection. Thus, when conducted in parallel with investigations in adults, mother-to-child HIV-1 transmission may provide complementary information on correlates of protection, design of HIV vaccines, and development of other therapeutic agents that may benefit both adults and children.