Submicroscopic infection of placenta by Plasmodium produces Th1/Th2 cytokine imbalance, inflammation and hypoxia in women from north-west Colombia

Women were recruited from January 2007 to April 2011, at the hospital obstetric facilities in the municipalities of Monteria (08°45′N, 75°53′W), and Puerto Libertador (07°53′35′ N, 75°40″16″ W) of the Uraba-Sinu-San Jorge-Bajo Cauca region of Colombia. This region has an estimated area of 43,506 km2, a population of 2.5 million at risk of malaria, with a mean annual parasite index of 35.8 cases/1,000 inhabitants [23]. The region has a low and stable malaria transmission intensity, with no marked fluctuations in the number of malaria cases during the year. The recorded prevalence of gestational malaria in the region was 9.1% by thick smear and 14.0% by PCR, while placental malaria was observed in 3.3% and 16.5%, respectively [24].

The study was part of a larger project aimed at exploring the epidemiology, clinical aspects and immunopathology of gestational and placental malaria in north-west Colombia, from which partial results have been reported elsewhere [21, 24]. A total 2,000 pregnant women were recruited in a sequential fashion for the main study. Clinical and epidemiological surveys were applied to all those women. Based on the availability and quality of material collected from peripheral blood and placental tissue, a subset of subjects was selected from the records of the 2,000 women recruited in the main project, to explore the effect of malaria infection on the pathophysiology of placental damage. Regardless of the time of the year or the year of collection, a total 50 subjects, non-matched, were studied: 25 with confirmed placental infection and 25 without Plasmodium spp. infection. The status of infection was defined by thick smear and real-time PCR (qPCR) of placental blood. Positive infected units were those with a negative diagnosis by microscopy and a positive qPCR result; negative units were negative by both tests.

Inclusion criteria were voluntary acceptance to participate in the study, signature of the informed consent, permanent residency (>1 year) in the malaria endemic region, negative history of pre-eclampsia and negative HIV and TORCH tests. The exclusion criterion was withdrawal of consent.

After enrolment, a questionnaire was completed which recorded data including age, number of pregnancies, number of malaria episodes during the ongoing pregnancy (based on antenatal records) and antimalarial treatment administered. In addition, labour/delivery and infant outcomes data, as well as maternal haemoglobin value, were obtained from the delivery chart. Placentas were processed immediately after delivery and maternal peripheral blood was obtained within 24 hours. After cleaning with saline, two ~1 cm³ sections of placenta were removed from the maternal side of the organ and blood was collected by aspiration into a tube containing EDTA. From mothers, 3 ml of peripheral blood was obtained and collected into EDTA containing vials. Thick smears were prepared from placental and mother’s blood and blood spots (approximately 100 μl of blood) were placed onto Whatman 3 mm filter paper for molecular diagnosis. One tissue fragment was preserved in RNA later® before processing for expression of markers of apoptosis, hypoxia and inflammation, and the remaining tissue fragment was fixed with 10% buffered formalin (pH 7.0-7.4), before embedding in paraffin, for the TUNEL assay.

Plasmodium infection was evaluated in mother’s peripheral and placental blood by microscopy of Field-stained thick films and qPCR. An experienced microscopist based at the field laboratory determined the presence of infection by counting the number of parasites per 200 leukocytes, based on a mean count of 8,000 leukocytes per microliter of blood. Samples were negative when no parasites were detected in 200 high power (100x) fields. For the qPCR, DNA was extracted from a 6 mm² fragment of placenta using the Chelex method described by Plowe et al.[25]. A qPCR was performed as described previously [26]. Briefly, samples were first tested for Plasmodium using a genus-specific set of primers and probe. The reaction was performed in a final volume of 25 μl containing 5 μl of DNA, 12.5 μl of TaqMan universal master mix (Applied Biosystems), 200 nM of each primer (Plasmo1 and Plasmo2), 50 nM of Plasprobe on the ABI 7500 FAST platform, under universal cycling conditions as published. Samples with a cycle threshold (Ct) value under 45 were tested in a duplex species-specific real-time PCR reaction for P. falciparum and P. vivax[26].

Cytokines were measured in mRNA isolated from maternal whole blood and placental tissue, while markers of tissue damage, hypoxia and inflammation were measured in mRNA isolated from placental tissue. Relative quantitation for expression analysis was performed using a reverse-transcription real time PCR assay (RT-PCR). Total RNA was extracted using QIAamp RNA Blood Mini® (QIAGEN) and cDNA was synthesized using First Strand cDNA Synthesis® (Fermentas). The reaction was set up in a Roche LightCycler® or ABI 7500. Table 1 details the primers and probes used. The following genes were analysed for expression: Fas, FasL; COX-1, COX-2, HIF, VEGF, and the cytokines IL- 2, IL-4, IL-10, IFN-γ and TNF. The efficiency of the PCR reactions was determined based on mRNA extracted from a stimulated BeWo cell culture or peripheral mononuclear cells from a donor. Complementary DNA was serially diluted and expression of β-actin was used to normalize the assays using the delta delta CT method (ΔΔCt) as described previously [27].

Apoptosis in placentas was assessed by a TUNEL assay as described earlier [28]. Placental sections were deparaffinized and placed on Fisherbrand Superfrost Plus® slides. Detection of cells undergoing apoptosis was performed using the kit DeadEnd Colorimetric System® Promega TUNEL, according to the manufacturer’s protocol. The apoptotic index was calculated based on the proportion of TUNEL stained cells observed in 10 fields (40x). The test was performed in a blinded fashion, for this samples were assigned a code which was revealed once all results were available.

Data were analysed using Excel and SPSS 19. Significance was set at p < 0.05. Non-parametric tests were used for data analysis and the U Mann–Whitney test and the Kruskal-Wallis test were used for correlations. According to the data distribution, the variable number of DNA copies/μl of template was grouped by convenience as follows: Group 1 (0 copies-no infection); Group 2 (0.01 to 1.99 copies/μl), Group 3 (2.00 to 100 copies/μl), and Group 4 (> 100 copies/μl) [16].

Pregnant women or guardians (in case of <18 years of age) signed a voluntary consent form. The study involved a minor risk and approval was granted by the Comite de Etica of Instituto de Investigaciones Medicas, Facultad de Medicina, Universidad de Antioquia (Approval Certificate number IIM 889ADV).

Maternal peripheral blood and placental tissue were evaluated for the expression (mRNA) of IFN-γ, TNF, IL-2, IL-4 and IL-10. In both compartments, expression of IFN-γ, TNF, and IL-10 was significantly higher in PM + when compared to PM-. Meanwhile, expression of IL-4 was high in placentas in PM+, and IL-2 was high in peripheral blood of the same group (Table 3). In peripheral blood of the PM + group, the cytokines TNF and IFN-γ were elevated and positively correlated (rho = 0.749, P = 0.000 with n = 50). In addition, for IL-4 and IL-10 a lower (rho = 0.431, P = 0.002 with n = 50), but nevertheless significant, correlation was observed in placenta, but not in peripheral blood (rho = 0.048, P = 0.740 with n = 50). Other positive correlations included IL-2 and TNF (rho = 0.676, p = 0.000 n = 50), IL-2 and IFN-γ (rho = 0.760, p = 0.000 n = 50), all in peripheral blood. In this same compartment, IL-4 did not correlate (p ≤ 0.406), but IL-10 positively correlated with IL-2, TNF and IFN-γ (p = 0.000).

In general, the differences observed in expression of cytokines in peripheral blood and placenta between PM + and PM-, persisted when analysis was performed taking into account the infecting species (Table 4).

According to TUNEL assay, the mean apoptotic index observed in the group PM + was 62.37% (range 55-79%) (Table 5), and this was significantly higher when compared with the PM- group (44.88%). Similarly, mean Fas expression was significantly higher in PM + than in PM- (6.52 vs. 3.42). However, FasL expression was higher in PM- than in PM + (6.95 vs. 5.60, p < 0.05) (Table 5).

The inflammation markers HIF and VEGF were observed high in the PM + group, but only VEGF was significantly different. As for the hypoxia markers, COX-1 and COX-2, both were significantly high in the PM + group. In general, all evaluated markers of tissue damage were up regulated in the placentas of the PM + group (Table 5).

Analysis of apoptosis, inflammation and hypoxia markers according to the Plasmodium species observed in placenta, confirmed similar results to those observed between the PM + group and non-infected placentas (Table 6).

No significant correlation was observed between IL-2 and any of the apoptosis/inflammation/hypoxia markers. Similarly, VEGF failed to correlate with the expression of any cytokine. On the other hand, the percentage of apoptotic cells and the expression of Fas, COX1, COX2 and HIF positively correlated with IFN-γ, TNF, IL-4 and IL-10. Meanwhile, FasL negatively correlated with the remaining variables (Table 7).

Placental parasite density was calculated using a qPCR assay as detailed in the methods section. In order to facilitate analysis, and based on the frequency distribution of data, results were categorized according to the copies of parasite DNA measured per PCR μl of total purified DNA: Group 1 (0 copies-no infection); Group 2 (0.01 to 1.99 copies/μl), Group 3 (2.00 to 100 copies/μl), and Group 4 (> 100 copies/μl) [16]. In general, significant differences were observed between Group 1 and the remaining groups as a whole. Copy number correlated positively and significantly with the percentage of apoptotic cells, and expression of Fas, COX1, COX2, HIF, IFN-γ, TNF, IL-4 and IL-10 (p = 0.000 and for IL4 p = 0.017). However, no significant correlation (p ≥ 0.078) was observed between expression of FasL, VEGF and IL-2 in relation to the copies of parasite DNA.