If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence to Alejandro G. Schijman, Ph.D., or Natalia A. Juiz, Ph.D., Instituto de Investigaciones en Ingeniería Genética y Biología Molecular “Dr Héctor Torres”, 2490, Vuelta de Obligado, C1428ADN, Buenos Aires, Argentina.
Molecular Biology of Chagas Disease Laboratory, Genetic Engineering and Molecular Biology Research Institute Dr. Héctor Torres (INGEBI-CONICET), Buenos Aires, Argentina
Molecular Biology of Chagas Disease Laboratory, Genetic Engineering and Molecular Biology Research Institute Dr. Héctor Torres (INGEBI-CONICET), Buenos Aires, Argentina
Molecular Biology of Chagas Disease Laboratory, Genetic Engineering and Molecular Biology Research Institute Dr. Héctor Torres (INGEBI-CONICET), Buenos Aires, Argentina
Address correspondence to Alejandro G. Schijman, Ph.D., or Natalia A. Juiz, Ph.D., Instituto de Investigaciones en Ingeniería Genética y Biología Molecular “Dr Héctor Torres”, 2490, Vuelta de Obligado, C1428ADN, Buenos Aires, Argentina.
Molecular Biology of Chagas Disease Laboratory, Genetic Engineering and Molecular Biology Research Institute Dr. Héctor Torres (INGEBI-CONICET), Buenos Aires, Argentina
Trypanosoma cruzi infection in women of reproductive age is associated with congenital transmission and adverse pregnancy outcomes. The placenta is a key barrier to infection. Gene expression profiles of term placental environment from T. cruzi–seropositive (SP) and -seronegative (SN) mothers were characterized by RNA-Seq. Nine pools of placental RNA paired samples were used: three from SN and six from SP tissues. Each pool consisted of female/male newborns and vaginal/cesarean delivery binomials. No newborn was congenitally infected. T. cruzi satellite DNA quantitative PCR in placental tissues and maternal and neonatal blood, and parasite 18S quantitative RT-PCR from placental RNA were negative, except in three SP women's bloodstream. To identify pathways associated with maternal T. cruzi infection, a gene-set association analysis was implemented: SP placental samples showed overexpression of inflammatory response and lymphocytic activation, whereas numerous biosynthetic processes were down-regulated. About 42 genes showed a significant fold-change between SP and SN groups. KISS1 and CGB5 were down-regulated, whereas KIF12, HLA-G, PRG2, TAC3, FN1, and ATXN3L were up-regulated. Several expressed genes in SP placentas encode proteins associated with preeclampsia and miscarriage. This first transcriptomics study in human term placental environment shows a placental response that may affect the fetus while protecting it from parasite infection; this host response could be responsible for the low rate of congenital transmission in chronic Chagas disease.
About 1.125 million women of fertile age are infected with Trypanosoma cruzi worldwide.
The incidence of congenital transmission from a chronically T. cruzi–infected pregnant woman to her fetus is relatively low, making the passage of the parasite across the placental barrier an infrequent event. Several mechanisms exerted by the placenta to protect the fetus from the infection have been reported, such as epithelial trophoblast turnover,
Different strategies for studying the mechanisms of placental resistance to T. cruzi infection have been implemented, namely, in vitro monolayer cell culture models, ex vivo infection of human chorionic villi explants, and mammal experimental models.
However, they all have disadvantages: cell cultures cannot reflect physiological conditions; in chorionic villi explants, the immune system, essential in this process, does not participate; and placentas of experimental models are dissimilar to human placentas.
In search of a better understanding of host–pathogen interaction in the pregnancy context, we aimed to characterize the gene expression profile of the term placental environment from T. cruzi–infected mothers in comparison with that of healthy noninfected mothers.
Materials and Methods
Ethics Statement
The research protocols followed the tenets of the Declaration of Helsinki and Guidelines according to Resolution N°1480/11 of the Ministerio de Salud from Argentina and were approved by the local medical ethics committees named Comité Provincial de Bioética de Jujuy and Comité de Bioética Dr. Vicente Federico del Giúdice – Hospital Nacional Prof. Alejandro Posadas. All mothers provided informed consent on their behalf before sample collection and after the purposes of the study were explained.
Subjects and Samples
Serodiagnosis of pregnant women was done by means of conventional serological methods and performed by the respective health centers based on routine assays. In maternal and umbilical cord blood samples, as well as placental tissues, presence of T. cruzi was tested using multiplex real-time PCR as previously described.
Maternal infection with other pathogens that produce congenital transmission and adverse pregnancy outcome, as well as missing data or incorrect sampling, were considered as exclusion criteria.
Fresh normal placentas were obtained after labor from vaginal or cesarean deliveries and placed within 24 hours at 4°C. Each placenta was dissected, and the middle section
at 2 cm distance from the umbilical cord was isolated and placed into RNAlater solution (Applied Biosystems, Foster City, CA). Total RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA) and stored at −80°C until used.
Transcriptomic Studies
A RNA-Seq experiment was performed in nine pools of two different placental RNA samples each: three pools (C1, C2, and C3) belonging to placentas from seronegative mothers (SN) and six pools (TC4 to TC9) from seropositive mothers (SP). Each pool consisted of a binomial of a female/male newborn and a vaginal/cesarean delivery. The cDNA libraries were prepared according to Illumina's TruSeq Stranded Total RNA with Ribo-Zero Gold for Human (Illumina, San Diego, CA), and a Hiseq 2.500 Illumina platform with 100-bp paired-end reads was used for sequencing. The RNA-Seq data have been submitted to the National Center for Biotechnology Information Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo; accession number GSE107376).
RNA-Seq Data Analysis
Raw reads were subjected to a quality test with FastQC software version 0.11.5 (Babraham Bioinformatics, Cambridge, UK; http://www.bioinformatics.babraham.ac.uk/projects/fastqc). Sequences were mapped implementing STAR software version 2.5,
and Picard software version 2.9.4 (Broad Institute, Cambridge, MA) tools were applied to calculate the gene expression level. Finally, the differentially expressed genes (DEGs) were identified using edgeR software version 3.6,
and principal component analysis was calculated in R software version 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria). Gene Ontology (GO) functional enrichment analysis was performed by the web-based application GOrilla (MULTI-KNOWLEDGE Project; http://cbl-gorilla.cs.technion.ac.il) and gene-set association analysis for RNA-Seq with gene permutation GSAASeqGP software version 2.0.
Gene interaction networks were analyzed by GeneMANIA software version 3.4.0 (University of Toronto, Toronto, ON, Canada; http://www.genemania.org) and visualized by the corresponding Cytoscape software version 3.0.2 plugins.
To validate the RNA-Seq results, total RNA pools used in transcriptomics and 12 independent samples (four from SN and eight from SP women) were analyzed. The cDNA was obtained implementing TaqMan Reverse Transcription Reagents (Applied Biosystems). Quantitative RT-PCR (RT-qPCR) of selected DEGs were performed using 1X FastStart Universal SYBR Green Master (Roche Diagnostics Corp., Indianapolis, IN) using the following cycling conditions: 95°C 10 minutes; 5 cycles of 95°C for 15 seconds and 64°C 1 minute, 40 cycles of 95°C for 15 seconds, and annealing temperature as listed in Table 1, for 1 minute. The expression levels were normalized to GAPDH with primers GAPDH 5′-GGTCTCCTCTGACTTCAACA-3′ (forward) and GAPDH 5′-GTGAGGGTCTCTCTCTTCCT-3′ (reverse).
Table 1Oligonucleotides Used in RNA-Seq Validation
Gene
Forward
Reverse
Ta, °C
Amplicon length, bp
HLA-G
5′-CTGGTTGTCCTTGCAGCTGTAG-3′
5′-CCTTTTCAATCTGAGCTCTTCTTTCT-3′
57
80
FN1
5′-CAAACACTAATGTTAATTGCCCA-3′
5′-GGTCTGCTAACATCACTCCA-3′
56
156
ATXN3L
5′-AAGATGAACATCTCCGCAGT-3′
5′-GGCCTTTCGTGTAGGTATGA-3′
56
214
TAC3
5′-GGATCATGCTGCTATTCACAG-3′
5′-ATGCATGTCACGTTTCTCGG-3′
56
245
CGB5
5′-CAGCATCCTACAACCTCCTG-3′
5′-CGCGGGTCATGGTGG-3′
67
277
KISS1
5′-GGACCTGCCTCTTCTCACCA-3′
5′-ATTCTAGCTGCTGGCCTGTG-3′
54
140
PRG2
5′-TGGATTGGAGGCAGGATCA-3′
5′-TGCCGTCAACCCACTGAAA-3′
54
61
Primer sequences of selected DEGs, annealing temperatures of RT-qPCR procedures (Ta) and amplicons length.
Primer sequences of selected DEGs are listed in Table 1. The relative expression of each sample was calculated using the 2−ΔΔCT method with the control group as calibrator. Statistical analyses were performed using GraphPad Prism software version 6.01 (GraphPad Software, La Jolla, CA).
Results
Mother/Newborn Binomials Characterization
Sixty-five pregnant women (age range 18 to 41 years) were consecutively enrolled at Posadas Hospital, Buenos Aires, from January 2014 to November 2015 and at Jujuy Maternal Hospital from January to September 2016 (Figure 1). Sixteen were SN and 49 SP, but 24 of these binomials were omitted because of exclusion criteria. Numbers of vaginal and cesarean deliveries and also numbers of female/male newborns in each case are detailed in Figure 1. All newborn umbilical cord and peripheral blood samples were collected during the first hour of life and had nondetectable T. cruzi satellite DNA by PCR, in agreement with microhematocrit findings. Placental tissues were also real-time quantitative PCR negative (Table 2).
Figure 1Flowchart of enrolled mother/newborn binomials in the prospective study.
Table 2Characteristics of Mother/Newborn Binomials under Study
Pool
Sample ID
Mother
Newborn
Placenta
Delivery
Gestation, weeks
Age, years
Serology
qPCR
qPCR
Sex
RT-qPCR and qPCR
C1
C3
26
−
ND
ND
M
ND
C
39
C52
29
−
ND
ND
F
ND
V
40
C2
C4
32
−
ND
ND
M
ND
C
38.4
C7
18
−
ND
ND
F
ND
V
41
C3
C6
33
−
ND
ND
M
ND
V
39.2
C8
28
−
ND
ND
F
ND
C
39.5
TC4
TC1
22
+
15p/mL
ND
M
ND
V
37
TC11
18
+
D/NQ
ND
F
ND
C
39.5
TC5
TC3
35
+
43p/mL
ND
M
ND
C
39.4
TC34
41
+
D/NQ
ND
F
ND
V
40.3
TC6
TC45
28
+
D/NQ
ND
F
ND
V
39
TC33
40
+
D/NQ
ND
M
ND
C
38
TC7
TC4
26
+
D/NQ
ND
F
ND
C
39.4
TC40
22
+
ND
ND
M
ND
V
38.5
TC8
TC42
20
+
ND
ND
M
ND
V
41
TC44
24
+
ND
ND
F
ND
C
37
TC9
TC5
31
+
ND
ND
F
ND
C
39.2
TC22
28
+
ND
ND
M
ND
V
40.5
Maternal age and serological diagnosis, newborn sex and results of T. cruzi satellite DNA qPCR expressed as parasites per milliliter (p/mL) in maternal and neonatal blood, as well as T. cruzi satellite DNA qPCR and 18S RNA–based RT-qPCR in placental tissues. It is also indicated delivery mode and time of gestation.
For RNA-Seq studies, only term pregnancies with times of gestation above 37 weeks (mean = 39.22 ± 1.15 weeks) were selected. Of 29 placental samples from SP, 12 were selected, grouped into six pools, and of 12 samples from SN, 6 were selected, grouped into three pools. The construction of pools took into account a balance (1:1) of newborn sex and mode of delivery variables, because they could become confounding factors (Table 2). These samples belonged to mothers with a mean age of 27.92 ± 7.55 and 27.67 ± 7.55 years, in the SP and SN groups, respectively (P > 0.05). No significant differences were found in time of gestation between the SP and SN groups (Table 2) (P = 0.597).
Of SP binomials, three pools were prepared with samples from women with positive bloodstream PCR results (TC4, TC5, and TC6) and another three with nondetectable parasite DNA in maternal peripheral blood after admission to maternity service (TC7, TC8, and TC9). Moreover, none of the placental tissues were positive for the determination of 18s RNA for T. cruzi (Table 2).
Relatedness among Gene Expression Patterns of Pools
First, principal component analysis was performed for a preliminary clustering of samples on the basis of complete gene expression data without prior statistical filtering (Figure 2). The first two principal components, which explain 86% (PC1 61% and PC2 25%) of the variance among samples, separate pools of SN (C1, C2, and C3) from those of SP (TC4 to TC9) mothers. However, the gene expression observed in the SP group could not differentiate between mothers with detectable or nondetectable T. cruzi DNA in peripheral blood close to delivery.
Figure 2Principal component analysis (PCA). PCA plot of human placental transcriptome from pooled samples of seronegative (SN) and seropositive (SP) mothers with detectable (PCR+) or nondetectable (PCR-) T. cruzi DNA in peripheral blood at the time of delivery. Circles cluster SN and SP samples. The x and y axes correspond to the first (PC1) and second (PC2) principal components, respectively, with their explained variances.
Only 42 genes were differentially expressed with a significance level <0.05 and a fold change |FC| >1.5 and 26 with |FC| >2, comparing SN and SP placentas (Table 3). The most represented transcript in SP group placentas was KIF12, a kinesin superfamily member with a role in microtubule cytoskeleton. Another interesting gene that turned out to be more expressed in the SP group was HLA-G. It belongs to the HLA class I that is expressed on fetal-derived placental cells and is related to immune tolerance in pregnancy. In addition, PRG2, a cytotoxin involved in antiparasitic defense mechanisms, MUC4, a mucin that provides protection to epithelial cells, and TAC3, the protein of which is expressed in the syncytiotrophoblast, had also a greater transcription level in the SP group.
Table 3List of Differentially Expressed Genes Detected by RNA-Seq Analysis
There was also a down-regulation of immunomodulatory activity genes in the SP group, such as IL1F10 and KISS1, inhibitors of chemotaxis and invasion, with a putative role in cell-matrix adhesion and that stimulate the secretion of gonadotropin. Related to the latter, the gene encoding the beta subunit of chorionic gonadotropin, CGB5, was also down-regulated, whereas fibronectine 1 (FN1), presents in cell surface and the extracellular matrix, was overexpressed in placentas from SP mothers. Finally, the most down-regulated gene was MIR6723, a microRNA that may be involved in post-transcriptional regulation.
Functional Pathway Analysis
To identify pathways significantly associated with the SP group in the placental environment, we first used the GOrilla tool for identifying enriched GO terms, considering genes with an adjusted P < 0.05 (Supplemental Table S1). Extracellular matrix organization emerged from positive FC genes, whereas pathways related to exocytosis, neutrophil degranulation, and glutamine and glutathione biosynthetic processes emerged from the list of negative FC genes.
A gene-set association analysis was also performed by GSAA-Seq, a cutoff-free strategy of genome-wide patterns in the measured gene expression. In this approach, choosing arbitrary cutoffs is avoided, and accumulative subtle changes in gene expression within the same gene set are identified. The resulting gene sets significantly associated with a false discovery rate q value < 25% and nominal P < 5% were represented in an enrichment map (Figure 3). Several pathways were overexpressed in the SP compared with SN placentas; some of them were related to immune response, cytoskeleton and calcium transport, cell cycle, DNA synthesis, and chromatin modifications. On the other hand, processes such as biosynthesis and respiratory chain, vesicle transport, and metabolism of lipids and hormones were down-regulated, as were networks associated with protein metabolism (synthesis and proteolysis by proteasome).
Figure 3Gene-set association analysis. Representation of biological pathways with false discovery rate q value < 25% and nominal P < 5%. Nodes correspond to gene sets, where red nodes correspond to up-regulated gene-sets and nodes in blue correspond to down-regulated ones, and links correspond to overlap of member genes. To improve layout, the networks were manually rearranged. Major clusters were manually labeled.
Within the top 10 GO pathways of the ranked list of over-represented genes from SP, which is based on the false discovery rate q values and on the Gene Set Normalized Association Score, which reflects how much a pathway is over- or under-represented, five pathways were related to calcium transport, and three to immune response (Supplemental Table S2). Interestingly, the fourth in this list was the GO Positive Regulation of Acute Inflammatory Response with 16 (57%) genes contributing to the enrichment score. In the eighth and tenth positions, there were T-cell Selection and Positive Regulation of Neutrophil Migration networks.
Regarding gene sets that were under-represented in placentas of SP mothers, several pathways related to respiratory chain were on the top of the list (Supplemental Table S2): Mitochondrial Respiratory Chain Complex I Biogenesis and Energy Derivation By Oxidation of Organic Compounds. Moreover, translation processes were also represented in the first positions: Translational Initiation and Translational Termination. Lastly, processes of vesicle transport were also affected by maternal infection, for example, Establishment of Protein Localization to Endoplasmic Reticulum (Supplemental Table S2).
RNA-Seq Validation
To validate the RNA-Seq data using RT-qPCR, seven DEGs were selected on the basis of their high FCs and low P values: FN1, ATXN3L, TAC3, CGB5 and HLA-G, PRG2, and KISS1. The expression patterns of these genes are shown in Figure 4. The results showed that the expression patterns as determined by RT-qPCR were consistent with those obtained by RNA-Seq (P > 0.05), confirming the accuracy of the RNA-Seq results in this study.
Figure 4RT-qPCR validation analysis of gene expression. Bars represent the mean gene expression levels in seropositive (SP) relative to seronegative (SN) samples, and vertical lines show SD. Data are expressed as means ± SD.
Maternal parasitic loads and strain placental tropism have been implicated in the likelihood of congenital transmission, which is an infrequent event in chronic Chagas disease pregnancy.
who reported the occurrence of congenital infection in both sisters sharing one placenta and no infection in the third one, indicative of the protective role of the placenta. Human genomic polymorphism in placental genes has been also associated to the occurrence of transmission.
Nevertheless, there is still a knowledge gap regarding the effect of the parasite infection in pregnancy outcome.
Placental Gene Expression Profile in SP Women
In the present study, even though a few genes (N = 42) had a significant FC between SP and SN groups, the deeper gene-set association analysis of the RNA-Seq data suggests the existence of a gene expression profile in the chorionic villi region of placentas from T. cruzi–seropositive mothers that differs from that of placentas from uninfected women, as revealed by principal component analysis.
This first study in human placental environment is consistent with our previous findings in placental gene expression from mice infected with T. cruzi compared with noninfected dams.
Indeed, the analysis showed several biosynthetic processes down-regulated in both murine and human placental studies, including lipid and protein metabolisms, whereas immunological gene sets were enriched in infected dams and mothers. In fact, in both studies, the inflammatory response and lymphocytic activation were overexpressed in placental tissues from infected groups. It has been already reported that newborns of T. cruzi–infected mothers are prone to produce higher levels of proinflammatory cytokines in comparison with those born to noninfected mothers, and there is increasing evidence suggesting that maternal infection during pregnancy affects the developing immune system of fetuses independent of the vertical transmission of pathogens.
Nevertheless, the placenta could reduce the damage produced by the inflammatory response by decreasing the exocytosis process, including neutrophil degranulation, because we have observed a significantly lower expression in related genes in the SP group. These results are also consistent with those from mice placenta, where genes associated with the secretory granule were down-regulated.
Moreover, vesicle transport processes also decreased in placentas from SP women and infected mice.
Because inflammatory activity leads to extracellular matrix damage, it is not surprising that we have found an activation of extracellular matrix organization–related genes. Previous studies had reported the importance of the extracellular matrix remodeling and that fibronectin was not affected at a protein level
Reorganization of extracellular matrix in placentas from women with asymptomatic Chagas disease: mechanism of parasite invasion or local placental defense?.
; however, our findings suggest that the placenta produces de novo synthesis of this protein, as well as collagen, during T. cruzi infection.
This differential gene expression in SP placentas was not associated with the presence of parasitic DNA or RNA in the tissues subjected to RNA-Seq analysis. Nevertheless, real-time quantitative PCR and RT-qPCR analysis in in-term placentas do not reflect the dynamics of placental parasite burden during pregnancy. Indeed, our transcriptomic analysis represents a placental terminal stage.
Differentially Expressed Genes in Placentas from SP Women
Infection with T. cruzi during pregnancy could result in low birth weight, still birth, spontaneous abortion, growth restriction, intrauterine fetal death, and fetal abnormalities.
There are some studies reporting similar or slightly higher prevalence of abortions and stillbirths in T. cruzi–seropositive mothers than in uninfected mothers, and there are also some contradictions about the consequences on low birth weight.
It is worth noting that in this study, SP human placentas showed a down-regulation of the KISS1 gene, which encodes kisspeptin, low serum levels of which have been related to preeclampsia and intrauterine growth restriction.
Interestingly, the gene CGB5 that encodes the beta 5 subunit of hCG was down-regulated in SP human placentas.
Moreover, the DEG TAC3 had a higher transcription activity in SP placentas. Its coding protein, neurokinin B (NKB), is expressed by syncytiotrophoblasts, and its overexpression has been also associated with preeclampsia.
Furthermore, PRG2 gene coding protein was also up-regulated in SP placentas. It encodes the preform of eosinophil major basic protein (proMBP), the predominant constituent of eosinophil granules, which have been associated with T. cruzi clearance through extracellular killing in the absence of antibodies.
so the higher transcription of its inhibitor observed in SP placentas might play a role in pregnancy outcome. Both PAPPA and hCG low levels have been proposed as biomarkers of preeclampsia.
Lower placental growth factor and higher free β-hCG and PAPP-A levels in the fetal circulation of near-term pregnancies complicated with severe preeclampsia.
Finally, GOrilla analysis allowed detection of altered glutamine and glutathione biosynthetic processes, consistently with a recent proteomics analysis of human placenta, which showed that, not only inflammatory response, but also glutathione metabolism dysfunction may contribute to preeclampsia.
Taken together, it is tempting to speculate that mothers with Chagas disease could be more prone to manifest preeclampsia than noninfected ones, which can be worth studying. This would be of relevance in chronic Chagas disease, because preeclampsia is associated with an increase of future heart failure and cardiovascular diseases.
Lack of association between Toxoplasma gondii infection and hypertensive disorders in pregnancy: a case–control study in a Northern Mexican population.
Another DEG was HLA-G with a positive FC, an immunomodulatory molecule involved in generating tolerance during pregnancy. A study of malaria showed that higher levels of HLA-G in infected mothers correlated with higher HLA-G levels in their newborns and infants; this phenomenon was related to a higher susceptibility of Plasmodium infection in children belonging to binomials with high HLA-G serum levels.
It would be of interest to study this phenomenon in the context of T. cruzi infection.
This is the first placental transcriptomics study in natural human infection, which may contribute to the understanding of placental mechanisms induced by maternal T. cruzi infection. Our results suggest that although most fetuses from infected mothers remain uninfected, they may be affected by the placental response that protects them from the parasite.
Acknowledgments
We thank Paula Beati (INGEBI-CONICET), Leandro Simonetti (INGEBI-CONICET), and Carlos David Bruque (ANLIS-MALBRAN) for their bioinformatics contributions.
Reorganization of extracellular matrix in placentas from women with asymptomatic Chagas disease: mechanism of parasite invasion or local placental defense?.
Lower placental growth factor and higher free β-hCG and PAPP-A levels in the fetal circulation of near-term pregnancies complicated with severe preeclampsia.
Lack of association between Toxoplasma gondii infection and hypertensive disorders in pregnancy: a case–control study in a Northern Mexican population.
Supported by the Argentinean Ministry of Science and Technology grants PICT 2014-1188 and PICT 2015-0074, and the European Union's Research and Innovation funding program grant ERANET-LAC HD 328 (A.G.S).
30018-X&id=gr1.jpg){kind=link}
30018-X&id=gr2.jpg){kind=link}
30018-X&id=gr3.jpg){kind=link}
30018-X&id=gr4.jpg){kind=link}