|Home | About | Journals | Submit | Contact Us | Français|
Mother-to-child-transmission (MTCT) of human immunodeficiency virus (HIV) remains a significant cause of new HIV infections in many countries. To examine whether fetal immune activation as a consequence of prenatal exposure to parasitic antigens increases the risk of MTCT, cord blood mononuclear cells (CBMCs) from Kenyan and North American newborns were examined for relative susceptibility to HIV infection in vitro. Kenyan CBMCs were 3-fold more likely to be infected with HIV than were North American CBMCs (P = .03). Kenyan CBMCs with recall responses to malaria antigens demonstrated enhanced susceptibility to HIV when compared with Kenyan CBMCs lacking recall responses to malaria (P = .03). CD4+ T cells from malaria-sensitized newborns expressed higher levels of CD25 and human leukocyte antigen DR ex vivo, which is consistent with increased immune activation. CD4+ T cells were the primary reservoir of infection at day 4 after virus exposure. Thus, prenatal exposure and in utero priming to malaria may increase the risk of MTCT.
Malaria and human immunodeficiency virus (HIV) infections are leading causes of death among children and adults, respectively, in Africa [1, 2]. Morbidity attributable to these pathogens is particularly high during pregnancy and among neonates. Pregnant women infected with malaria or HIV have an increased risk of adverse events and outcomes, including anemia, spontaneous abortion, premature delivery, low neonatal birth weight, and increased infant mortality [3–8]. Maternal co-infection with both malaria and HIV may result in worse birth outcomes than either infection alone. Malaria-infected, HIV-positive pregnant women experience an increased prevalence of placental malaria, higher parasitemia, and clinical disease, compared with HIV-negative women [6, 9–14]. Conversely,malaria infection has been shown to increase HIV load and damage the placenta, causing microtransfusions from maternal to fetal circulation, and both of these are important risk factors in the vertical transmission of HIV from mother to child (MTCT) [15–19].
It is unclear whether malaria co-infection of HIV-positive women increases the risk for MTCT. Epidemiologic studies that have addressed this question have produced conflicting results . A study from Uganda found that placental malaria was associated with a ~3-fold greater risk for HIV vertical transmission . Similarly, a study from Cameroon found a positive association between MTCT and the peak rainy season when malaria transmission was high, suggesting a possible correlation between MTCT and maternal malaria infection, although malaria burden was not directly assessed . By contrast, no association between placental malaria and increased risk of MTCT was found on the Kenyan coast, which had lower levels of infection than did the studies from Uganda and Cameroon . Interestingly, a large study from western Kenya showed that high-density placental parasitemia was associated with an increased risk of MTCT, whereas low-density placental parasitemia was not , which suggest that the intensity or frequency of placental malaria may explain differences among studies.
These studies suggest that the relationship between placental malaria and MTCT is complex. To our knowledge, no study, however, has evaluated the potential impact of fetal factors on the risk of MTCT. For example, immune activation is vital for HIV infection; thus, in utero exposure to exogenous antigens and consequent fetal lymphocyte activation may increase susceptibility to HIV infection [24, 25]. Newborns of mothers infected with malaria and other intravascular parasites, such as schistosomiasis and lymphatic filariasis, are often primed to parasite antigens [26–30]. This activation of cord blood mononuclear cells (CBMCs) could increase the risk for HIV infection if exposed to virus in utero or perinatally. In support of this hypothesis, we previously demonstrated that helminth co-infections in HIV-positive women during pregnancy increase the risk for MTCT, especially among newborns who demonstrate in utero sensitization to helminth antigens . Here, we tested the hypothesis that CBMCs from Kenyan newborns of women infected with malaria exhibit greater susceptibility to HIV infection in vitro than do CBMCs collected from newborns of parasite-uninfected Kenyan or North American women. We also show that Kenyan CBMCs that demonstrate in utero priming to malaria blood stage antigens exhibit the greatest susceptibility to HIV infection in vitro.
Study population. Samples were obtained from newborns born at the Msambweni District Hospital, Coast Province, Kenya an area where malaria, schistosomiasis, lymphatic filariasis, and intestinal helminthes are endemic [26, 32]. The mothers were assessed for helminth and malaria infection during repeated assays performed on samples collected during antenatal clinical visits. All women were tested for HIV, and the offspring of HIV-positive women were excluded from the current study. Control CBMCs were prepared from healthy North American newborns delivered at University Hospitals of Cleveland (Cleveland, Ohio). Ethical approval was obtained from the Institutional Review Boards of the University Hospitals of Cleveland and the Kenya Medical Research Institute in Nairobi, and informed consent for the study was obtained from study subjects.
Determination of maternal parasite infection. Malaria infection was determined by blood smear and/or polymerase chain reaction performed on samples of maternal peripheral and intervillous blood; urinary schistosomiasis was assessed by the presence of ova in the urine; lymphatic filariasis was assessed by antigen detection in peripheral blood samples; and intestinal helminth infections were assessed by examination of stool samples, as described elsewhere .
CBMC sample collection and sensitization determination. CBMC isolation was performed by Ficoll-Hypaque density gradient separation. A subset of freshly collected CBMC sample was immediately cultured with media alone or with the following malaria antigens: recombinant malaria surface protein 1–42 (MSP1-42), alleles corresponding to 3D7 and FVO parasite lines (5 ug/mL each, kindly provided by Sanjay Singh, Carole Long, and David Narum of the Malaria Vaccine Development Unit, National Institutes of Health), pooled MSP1-42 peptides previously determined to be dominant CBMC epitopes in this population  (10 ug/mL for each peptide), Plasmodium falciparum ribosomal phosphoprotein peptides (PfP0; 10 ug/mL), and phytohemagglutinin (PHA; 1 ug/mL). Remaining CBMCs were cryopreserved.
Quantification of interleukin (IL)-2, IL-5, interferon γ, IL- 10, and IL-13 was performed on culture supernatants collected at 96 h using multiplex immunoassay (Millipore). Results were expressed in pg/mL by interpolation from standard curves. A positive response was scored when the following 2 criteria were fulfilled: (1) a net value for antigen-stimulated cells that was at least 2-fold greater than that of parallel cultures containing media alone and (2) responses to 2 malaria antigens. If cytokine production was not detectable in the negative control cultures, then >40 pg/mL cytokine was considered to be a positive response.
In vitro HIV infection. Cryopreserved CBMCs were thawed (all samples showed >85% viable cell recovery with preservation of immune function, as indicated by similar recall responses to malaria antigens between fresh and cryopreserved aliquots from the same cord blood sample) and rested overnight in Roswell Park Memorial Institute (RPMI) 1640 plus L-glutamine supplemented with penicillin, streptomycin (100 units/ mL), HEPES (10 mM), and 10% pooled human AB serum (cRPMI+10%). From this point forward, culture media consisted of cRPMI+10% supplemented with 1.0 ng/mL recombinant human IL-2 (rhIL-2; BD Biosciences). Infection with the R5 tropic virus HIVBaL (kindly provided by Dr. Eric Arts of Case Western Reserve/University Hospitals Cleveland Center for AIDS Research) was performed as follows: triplicate wells of 200,000 cells/well (1×106 cell/mL) in a 96-well tissue culture plate were cultured in the following conditions: (1) media alone, (2) immediate HIVBaL exposure with subsequent culture in media alone (virus plus media), (3) immediate HIVBaL exposure with subsequent culture in media containing 5.0 µg/mL PHA (Sigma-Aldrich; virus plus PHA), (4) 3 days of culture in media followed by HIVBaL exposure (media plus virus), (5) 3 days of culture in media containing pooled MSP1-42 peptides at 10 µg/mL each followed by HIVBaL exposure (malaria plus virus), and (6) 3 days of culture in media containing 5.0 mg/ mL PHA followed by HIVBaL exposure (PHA plus virus). Initially, a dose response curve was created to determine the optimal multiplicity of infection (MOI) for infection (see Results). Subsequently, cells were exposed to virus at an MOI of 0.001 for 4 h at 37°C after which cells were centrifuged, washed with fresh media, and resuspended according to the conditions outlined above. Every 3 days, one-half of the culture volume was removed, frozen, and replaced with fresh media containing appropriate supplements.
Reverse transcriptase (RT) assay. The removed culturemedia was stored at −80°C prior to the RT assay, which was performed according to standard protocols . In brief, 10 µL of culture supernatant was incubated with 25 µL of RT mix composed of Tris-HCl, KCl, dithiothretiol (DTT), 32P thymidine- triphosphate (dTTP), Poly r(A) template, poly d(T) primer, nonidet-P40, and α-32PdTTP for 2 h at 37°C. Ten µL of this reaction mixture was spotted onto Whatman filter paper, which was washed successively with 1 × saline-sodium citrate and 85% ethanol and allowed to dry. A known quantity of virus stock was included in all RT assays to verify consistency. Counts per min per microliter culture supernatant (cpm/µL) were quantified using a beta counter.
Flow analysis. One million cells were washed and stained with fluorochrome-labeled antibodies for 30 min in the dark at 4°C. For analysis of HIV infection, the following antibodies were used: CD3-PerCP (BD Biosciences); CD4-APC-Cy7, CD14-PE (eBioscience); and p24-FITC (Beckman Coulter). For analysis of CD4+ T cell activation markers the following antibodies were used: CD3-PerCP, Ki-67-PE (BD Biosciences); CD4-APC-Cy7, CD25-PE-Cy7, HLA-DR-FITC, CCR5-APC (eBioscience). Samples were processed (50,000 events) on a BD LSRII flow cytometer, and data was analyzed with FlowJo software (Tree Star). Gating was determined by fluorescence minus 1 staining, and compensation was calculated using CompBeads (BD Biosciences) and compensation platform in FlowJo. Integrated mean fluorescent indices were calculated as the product of percentage of positive cells and the geometric mean fluorescence intensity of the positive cells.
Statistical analysis. Fisher's exact test was used to make comparisons between groups for frequency of infection. Differences in peak viral replication between groups were determined by either Student's t test or analysis of variance (ANOVA). For expression of activation markers by flow cytometry, integrated mean fluorescent indices were log transformed and compared by ANOVA.
Determination of optimal MOI. To determine the optimal MOI for distinguishing potential differences in susceptibility to HIV in vitro, CBMCs from Kenyan (n = 6) and North American (n = 3) subjects were exposed to varying viral quantities (Figure 1). Maximal viral replication was observed at day 15 for all samples, as indicated by incorporation of α-32P-dTTP (data not shown). At day 15, an MOI of 0.05 infected all samples tested, whereas an MOI of 0.001 infected 4 of 6 Kenyan CBMCs and none of the North American CBMCs. This difference in infection at a lower MOI suggested heightened susceptibility to HIV in Kenyan CBMCs, compared with North American CBMCs.
Kenyan CBMCs are more susceptible to in vitro HIV infection than are North American CBMCs. To further examine this increased susceptibility of CBMCs from Kenyan newborns, compared with CBMCs from North American newborns, to HIV in vitro, additional cryopreserved CBMCs from Kenyan (n = 60) and North American newborns (n = 25) were examined. Samples were considered to be infected for a given culture if the measured cpm/µL was at least 5-fold greater than the no virus condition at 2 or more time points. In the absence of any exogenous stimulation (virus plus media), CBMCs from Kenyan newborns were significantly more likely to be infected with HIV, compared with North American CBMCs (21 of 60 vs 3 of 25 samples; P = .03; Figure 2A). This difference was further enhanced when viral exposure was followed by exogenous stimulation (virus plus PHA) (36 of 60 vs 6 of 25 samples; P = .004). PHA stimulation preceding viral exposure (PHA plus virus) resulted in HIV infection of 59 of 60 (98%) Kenyan and 25 of 25 (100%) North American CBMC samples.
Peak virus replication, as indicated by incorporation of α-32P-dTTP, was observed at day 15 for all samples and all culture conditions except PHA stimulation preceding viral exposure (PHA plus virus) when the peak was consistently observed at day 12. Among samples that were infected, peak viral production tended to be higher in Kenyan CBMC samples than in North American control samples at day 15 for virus plus media and virus plus PHA conditions, although differences were not statistically significant (Figure 2B).
To confirm the reproducibility of results, a second aliquot of Kenyan (n = 10) and North American (n = 10) CBMCs was thawed and infected as before. These replicates exactly match the results from earlier experiments in terms of classification of productive HIV infection, whereas peak levels of viral replication, as determined by RT assay, were also consistent (<20% difference in cpm/µL between replicate experiments; data not shown).
CBMCs of newborns sensitized to malaria antigens have increased HIV susceptibility. There was no clear association of either parity or maternal malaria, schistosomiasis, or geohelminth infection during pregnancy with increased CBMC susceptibility to HIV infection (Table 1). To examine the possibility that in utero immune priming to parasite antigens may increase susceptibility of Kenyan CBMCs to HIV infection in vitro, individual samples were stratified on the basis of prenatal immune priming to malaria blood stage antigens (Figure 2C and and2D).2D). One-half of Kenyan CBMC samples (30 [50%] of 60 samples) demonstrated recall responses to 2 malaria antigens and were defined as “malaria sensitized” (Figure 3). Samples classified as “malaria nonsensitized” generally did not produce any malaria antigen-driven cytokine. A few samples, however, demonstrated an isolated recall response to a single antigen, thereby failing to meet our criteria for immune priming.
Malaria-sensitized CBMCs were significantly more likely to be infected with HIV in vitro in the absence of exogenous stimulation (virus plus media) when compared with either malaria- nonsensitized (15 of 30 vs 6 of 30 samples; P = .03) or to North American CBMC samples (15 of 30 vs 3 of 25 samples; P = .004; Figure 2C). Although malaria-sensitized and malarianonsensitized CBMCs were similarly susceptible when virus exposure was followed by PHA stimulation (20 of 30 vs 16 of 30 samples; P = .43), malaria sensitized CBMC samples remained more likely to be infected, compared with North American CBMC samples (20 of 30 vs 6 of 25; P = .003). Among infected CBMC samples, peak viral replication was similar for malaria-sensitized versus malaria-nonsensitized CBMC samples (Figure 2D).
Stimulation with malaria antigens increases susceptibility of malaria-sensitized CBMCs to HIV infection. We further examined whether recall responses to malaria blood stage antigens increased susceptibility to HIV infection in vitro. Before virus exposure, a subset of CBMC samples from newborns described in Figure 2 (samples with sufficient numbers of frozen CBMCs) were exposed to pooled MSP1-42 peptides previously demonstrated to be dominant T cell epitopes in CBMCs from newborns on the coast of Kenya . In contrast to the addition of virus to CBMCs at the beginning of culture, preculture of CBMCs with media alone for 3 days prior to virus exposure failed to establish a productive infection in most samples from both Kenyan and North American subjects (4 of 20 vs 2 of 20 samples; Figure 4A). As expected, the addition of malaria peptides to North American CBMCs failed to enhance CBMC susceptibility to HIV infection (1 of 20 samples). In contrast, the addition of malaria peptides to CBMC samples classified as malaria sensitized showed a trend toward augmented HIV susceptibility, compared with parallel cultures of malaria-nonsensitized CBMCs (11 of 20 vs 6 of 20 samples; P = .2; Figure 4A). Malaria antigen- and PHA-driven Kenyan CBMC cultures infected with HIV produced similar levels of virus production, regardless of malaria sensitization status (Figure 4B).
CD4+ T cells are a primary reservoir for HIV at day 4 in vitro. Prenatal sensitization to malaria antigens has been shown to be predominantly attributable to CD4+ T cell responses . Therefore, we hypothesized that CD4+ T cells were the early target of HIV infection in this in vitro model. Flow cytometric analysis of p24 expression was performed for CD3+CD4+ T cells and CD14+ cells from malaria-sensitized samples known to be susceptible to HIV infection in the absence of exogenous stimulation (n = 5). Expression of p24 was undetectable in all samples at 24 h after virus exposure. However, at 96 h after virus exposure, 1.3% of CD3+ lymphocytes stained positively for p24 (Figure 5). Additional analysis indicated that all CD3+ cells that expressed p24 also expressed CD4 (data not shown). p24 was not detected in CD14+ cells (data not shown). Thus, productive viral infection is primarily detectable in CD3+CD4+ cells at 96 h after virus exposure.
Ex vivo flow staining of activation markers on CD4+ T cells. To determine whether malaria-sensitized newborns had evidence of increased immune activation, compared with malarianonsensitized or North American CBMC samples, ex vivo flow cytometric analysis of various CD4+ T cell activation markers was performed (n = 10 for each group). Little ex vivo proliferation, measured by Ki-67 expression, was observed, with no differences between groups (data not shown). Similarly, no differences were observed in the expression of CCR5 on CD4+ cells (data not shown). Malaria-sensitized CD3+CD4+ T cells demonstrated higher frequency and expression levels of CD25 (integrated mean fluorescent index, 6224) compared with malaria- nonsensitized (2049; P = .04) and North American CBMCs (1688; P = .04; Figure 6). Similarly, the frequency and expression of HLA-DR on CD4+ T cells was highest for malariasensitized CBMCs (integrated mean fluorescent index, 995), compared with malaria-nonsensitized (304;P = .05) andNorthAmerican CBMCs (328; P = .03). Thus malaria-sensitizedCBMCs contained CD4+ T cells that demonstrated increased ex vivo activation, which correlated with increased susceptibility to HIV infection in vitro.
Here, we show that CBMCs from Kenyan newborns in an area where multiple chronic parasitic infections are endemic are more susceptible to infection with an R5-tropic strain of HIV in vitro than are CBMCs from offspring born in North America. This increased susceptibility of Kenyan CBMCs to HIV infection was associated with in utero priming to malaria blood stage antigens and upregulation of CD25 and HLA-DR on CD4+ T cells. Stimulation of malaria-sensitized CBMCs further enhanced their susceptibility to HIV infection in vitro. HIV infected women are often coinfected with malaria and other chronic helminthic infections, which may lead to activation of fetal lymphocytes and, therefore, potentially increase the risk of MTCT. Thus prevention of malaria and chronic helminth infections (such as schistosomiasis, lymphatic filariasis, and geohelminths in pregnant women) may reduce MTCT, especially in conditions in which antiretroviral treatment may be difficult to reliably administer during pregnancy.
The failure to observe an association of maternal malaria or helminth infection with in vitro susceptibility to HIV is not surprising, because not all women expose their fetus to the parasite antigens necessary for in utero priming. The circumstances that might affect in utero sensitization are not clear, but it is likely affected by the duration, timing, and intensity of the parasite infection during gestation. Additional factors may include the integrity of the placenta and the pregnant woman's ability to generate appropriate antibodies to certain parasite antigens, because we have shown that immune complexes may be an important mechanism for transplacental transfer of malaria antigens .
Viral dose affected CBMC susceptibility to infection. At higher viral doses, all CBMCs tested became infected in vitro; however, at lower viral doses, more Kenyan samples thanNorth American samples remained susceptible. This increased susceptibility of Kenyan CBMCs occurred in the absence of exogenous stimulation, implicating in utero activation to exogenous antigens, particularly malaria and potentially other intravascular parasites, such as schistosomiasis and lymphatic filariasis, that are common in our population. In the event that the fetus or newborn should be exposed to virus, particularly at lower viral doses, this prenatal priming may increase the likelihood of MTCT. Importantly, when strong polyclonal activation preceded virus exposure in vitro, both Kenyan and North American samples were equally susceptible to infection.
One potential mechanism for fetal activation leading to increased viral susceptibility is by upregulation of CCR5, thereby facilitating viral invasion [37, 38]. Interestingly, such upregulation was not observed in the present study. This suggests that viral entry may not be the limiting step, which is an observation that is supported by similar levels of HIV minus strand strongstop DNA detected by real-time quantitative polymerase chain reaction 24 h after virus exposure among Kenyan and North American CBMCs (unpublished data). Rather, activated cells are likely to have greater pools of nucleotide precursors necessary for viral progression through reverse transcription, and/ or proviral gene transcription and translation is enhanced by host cell transcription factors, such as nuclear factor kB, activation protein 1, and/or nuclear factor of activated T cells . Although p24 was detected only in CD3+CD4+ T cells 4 days after virus exposure, the initial cell type infected in CBMCs remains unclear. Such studies are difficult because of the very low frequency of initially infected cells (unpublished data).
There are several limitations to the study. It is generally accepted that initial HIV infection occurs at mucosal surfaces [40–44], with dissemination to lymph nodes, where CD4+ cells act a reservoir [45–47]. Circulating CBMCs may not reflect those lymphocytes observed either at mucosal surfaces or in lymph nodes. Thus, our in vitro model of HIV infection may not fully represent in vivo conditions. In addition, although maternal infection with helminths during gestation was documented and failed to account for differences in observed in vitro HIV susceptibility, fetal sensitization to helminth antigens was not fully assessed. Because the prevalence of maternal helminth infection was high in the present cohort, and because we have previously shown that helminth-sensitized newborns born of HIV-positive mothers had an increased risk of MTCT in vivo , we cannot rule out the likelihood of fetal helminth sensitization contributing to the observed increased HIV susceptibility in vitro for Kenyan CBMCs.
In conclusion, this study demonstrates that CBMCs from Kenyan newborns primed in utero to parasite antigens aremore likely to become infected with HIV. Prevention of malaria and other chronic parasitic infections in HIV-positive pregnant women to block prenatal sensitization could reduce the risk of MTCT.
We thank medical superintendent Dr. Dawood Mwaura and nurses Victoria Saidi, Hashora Mwanguku, Zaituni Mwakileo, Fatuma Ngare, Ruth Notina, and Florence Wambua, for assisting with recruitment of women to the study, collection of cord blood samples, and care of the women and their newborns; Elton Mzungu, Kefar Wambua, Charles NgaNga, and Alex Osore, who performed many of the immunologic assays and parasitological examinations; Grace Methenge and Christine Lucas, for data entry and management; Drs. Zahra Toossi and Eric Arts, for assistance with the study concept and critical reading of the manuscript; and the women residing in the Msambweni location who participated in the study.
Potential conflicts of interest: none reported.
Financial support: National Institutes of Health (NIH) (grant AI064687), Merit Review Award from the Veterans Affairs Research Service, Case MSTP (NIH grant T32 GM07250 to K.S.), Case Geographic Medicine and Infectious Diseases Training Grant (NIH grant T32 A107024 to K.S.), and Case/University Hospitals of Cleveland Center for AIDS Research (NIH grant number: AI36219).
Presented in part: American Society for Tropical Medicine and Hygiene 58th Annual Meeting, November 2009, Washington, DC [abstract 368].