During pregnancy the maternal immune system has to adapt its response to accommodate the fetus. The objective of this study was to analyze the effects of smoking on the maternal immune system.
First-trimester decidual tissue and peripheral blood of smoking and nonsmoking women were analyzed by real time reverse transcription–polymerase chain reaction (RT-PCR) and flow cytometry. A mouse model was used to further analyze the effects of smoking. Murine tissue was analyzed by flow cytometry, real-time RT-PCR, and immunohistochemistry.
Smoking caused lower percentages of viable pups in mice and lower birthweights in humans. Smoking mothers, both mice and human, had more natural killer cells and inflammatory macrophages locally, whereas systemically they had lower percentages of regulatory T cells than nonsmoking controls.
Maternal smoke exposure during pregnancy influences local and systemic immune responses in both women and mice. Such changes may be involved in adverse pregnancy outcomes in smoking individuals.
Despite the well-known fact that smoking adversely affects pregnancy outcome, about 30% of pregnant women still smoke in The Netherlands. Fetal reactions to cigarette smoke were already described in 1935. After a mother started smoking a cigarette, an increase in fetal heart rate was found, and it was concluded that a toxic agent, probably nicotine, crossed the placenta to the fetus. Currently, it is clear that smoking during pregnancy causes lower birth weights, an increased incidence of spontaneous abortions and preterm birth, increased placental pathology, and increased stillbirth rates. Interestingly, it has also been shown that smoking women have a lower incidence of pregnancies complicated by preeclampsia. The mechanisms responsible for the toxic and protective effects of smoking on pregnancy and preeclampsia are not yet understood.
Tobacco smoke contains many substances and may affect the fetus directly, but it may also affect placental development and decidualization. Alternatively, smoking may achieve its effects by modulation of maternal immunological and endocrine parameters. During pregnancy the maternal immune system plays a key role in the success of pregnancy by adapting to accommodate the semiallogeneic fetus. Adaptations of the maternal immune system take place locally at the implantation site as well as in the peripheral circulation. Disturbances in this adaptation toward maternal tolerance have been associated with adverse pregnancy outcomes and other reproductive pathologies. Indeed, in nonpregnant women and animals, smoking alters immune responses. Therefore, adverse pregnancy outcomes in smoking women could be the result of immunological changes caused by smoking.
The exact influence of smoking on the maternal immune system during pregnancy is not known. Only 1 study investigated the effects of smoking on circulating maternal leukocytes during pregnancy, showing that smoking pregnant women had higher percentages of CD3+ cells, lower percentages of CD56+ cells, and a lower expression of CD54 (activation marker) on monocytes in peripheral blood in early pregnancy compared with nonsmoking pregnant women. It was concluded that smoking acts both as an inflammatory and antiinflammatory stimulus on the maternal immune system during pregnancy.
Important players in the adaptation of the maternal immune response are T cells (especially regulatory T cells), natural killer (NK) cells, and monocytes and macrophages. Regulatory T (Treg) cells are a subpopulation of T cells that regulate immune responses. They are important in the process of tolerance to self and foreign antigens, and they therefore play a role in autoimmune diseases, inflammatory diseases, transplantation, and pregnancy. NK cells are key players in the regulation of remodeling of spiral arteries and also have a role in regulation of invasion of trophoblasts. Functional and numerical changes of NK and Treg cells are both associated with adverse pregnancy outcomes.
Monocytes and macrophages are also involved in remodeling of spiral arteries and placental development. It has become clear that there are specific subsets of monocytes and macrophages, which are thought to have specific roles in inflammation, tissue remodeling, and vascularization, respectively, M1 (inflammatory) and M2 (remodeling/regulatory) macrophages. Macrophages in the placenta during healthy pregnancy appear to have a remodeling/regulatory role and are thus mainly of the M2 subtype.
The aim of this study was to analyze the effects of smoking on the maternal immune system during pregnancy locally in the uterus and systemically. In humans, smoking effects on the local and systemic maternal immune system were investigated in first-trimester decidual tissue and peripheral blood of smoking women. In this study, we used unique decidual material, collected during routine chorionic villous sampling between 10 and 12 weeks of pregnancy, which allowed us to study smoking effects in deciduas of ongoing pregnancies with known pregnancy outcomes. We furthermore used a mouse model to compare the results with human pregnancy and more specifically study changes in peripheral and local immune responses.
Materials and Methods
First-trimester decidual tissue with a known pregnancy outcome of 42 women, 21 smoking women, and 21 control women was selected from a tissue collection and analyzed by real-time reverse transcription–polymerase chain reaction (RT-PCR). In addition, data on peripheral blood T cell subsets of 6 smoking and 12 nonsmoking pregnant women were selected from a control cohort of a previously published study. To more specifically study changes in peripheral blood and local immune responses, we used a mouse model. Therefore, 6 pregnant mice were exposed to smoke and 6 control pregnant mice were exposed to fresh air.
To analyze the effects of smoking on the local maternal immune system, decidual tissue was analyzed by immunohistochemistry and real-time RT-PCR, and uterus-draining lymph nodes were analyzed by flow cytometry. For the analysis of the systemic immune response, systemic lymph nodes and spleens were analyzed by flow cytometry.
Collection of human decidual tissue
During routine chorionic villus sampling, performed vaginally between 10 and 12 weeks of gestation for maternal age, a previously chromosomal abnormal child, or serum screening-related risk for Down syndrome, surplus material, which was not needed for karyotyping, was obtained. Immediately after sampling, decidual tissue was mechanically separated from villi under a microscope by a qualified, experienced laboratory technician to minimize trophoblast contamination. The chorionic villi appeared as free-floating white structures with fluffy, filiforme branches, whereas decidual tissue had a more amorphous appearance and lacked distinct branches. Earlier studies showed that trophoblast cells were seldom found in decidual samples. Tissue was stored immediately at –20°C until further processing as described before.
Patients were informed that surplus material could be used for research. Follow-up of these pregnancies was available by a questionnaire returned by the patient postpartum. From this material, we selected decidual tissue from smoking women, and for each case we used a nonsmoking control woman matched for maternal age, parity, and gestational age at time of sampling. Information about smoking was obtained by medical history taking. For cases and controls, only pregnancies with a known pregnancy outcome longer than 30 weeks’ gestation were selected. In total, decidual tissue of 21 smoking and 21 nonsmoking women was selected for analysis ( Table 1 ).
|Characteristic||Nonsmoking (n = 18)||Smoking (n = 20)||<10 cigarettes per day (n = 11)||≥10 cigarettes per day (n = 9)|
|Age, y||37.6 ± 2.1||36.8 ± 4.3||37.5 ± 2.8||35.8 ± 5.7|
|Birthweight, g||3388.6 ± 591.6||3339.4 ± 600.2||3612.0 ± 574.8||3165.0 ± 548.5 a|
|Sex of child (% male)||28.6||35.0||36.4||33.3|
|Duration of pregnancy, d||273.1 ± 18.1||270.9 ± 31.1||268.6 ± 41.8||273.6 ± 14.3|
|Smoking average, (cigarettes per day)||0||9.2 ± 6.3 b||4.8 ± 2.6 b||15.0 ± 4.3 b|
To determine whether smoking had a dose-dependent effect, we divided the total smoking group into a heavy (≥10 cigarettes per day) and moderate smoking group (<10 cigarettes per day). The low percentage of male offspring is within normal variation and may have been influenced by the relatively high maternal age because there is evidence that a higher maternal age is associated with a shift from a male to a female majority in newborns.
PCR on human decidual tissue
Total ribonucleic acid (RNA) was isolated from decidual tissue using a RNA isolation Trizol kit (Invitrogen, Carlsbad, CA). To further purify RNA, a Turbo DNA free kit and column chromatography were used according to the manufacturer’s instructions (respectively, Applied Biosystems, Foster City, CA, and Bio-Rad Laboratories, Hercules, CA). Complementary deoxyribonucleic acid (cDNA) was reverse transcribed using a Superscript-II reverse transcriptase kit (Invitrogen).
As a housekeeping gene (HPRT), we used the forward primer 5′-GGCAGTATAATCCAAAGATGGTCAA-3′-GTCTGGCTTATATCCAACACTTCGT-3′ (Invitrogen), and probe: 6-FAM 5′-CAAGCTTGCTGGTGAAAAGGACCCC-3′ TAMRA (Eurogentic, Herstal, Belgium). For the other genes, interleukin (IL) 6, IL10, Tbet21 (Th1 response), Gata 3 (Th2 response), Rorγt (Th17 response), Foxp3 (Treg cell), CD56 (NK cell), CD68 (pan macrophage marker), NOS2 (iNOS, M1 macrophage), and CD206 (MRC-1, M2 macrophage), On-Demand gene expression assays were used (Applied Biosystems). PCR reactions were performed in triplicate in a volume of 10 μL consisting of 0.1 μL of MilliQ water, 5 μL PCR mix (Eurogentic), 1 μL of reverse and forward primer each, 0.4 μL probe, and 2.5 μL cDNA for the housekeeping gene.
For the other genes, the total volume of 10 μL contained 2 μL of MilliQ water, 5 μL PCR mix (Eurogentic), 0.5 μL assay mix, and 2.5 μL cDNA. Runs were performed by a 7900HT Fast real-time PCR system (Applied Biosystems), and RNA data were normalized to HPRT messenger RNA (mRNA) expression using 2 -ΔCt . After the run, it appeared that HPRT mRNA expression of 4 samples (3 controls, 1 case) was undetectable; these cases were excluded from this study. Undetectable cycle threshold (Ct) values of the gene of interest (more than 40) were interpreted as the maximum Ct value (40).
Human flow cytometry
To analyze the influence of smoking on the systemic maternal immune system in humans, data on T-cell subsets were selected from of a previously published control pregnant cohort and reanalyzed for effects of smoking. In short, peripheral blood samples were obtained from pregnant women well before the onset of labor, and T-cell subsets were analyzed using fluorescently labeled antibodies against CD4, CD25, and Foxp3 as described before. In total, 6 smoking pregnant women were selected and matched by gestational age, parity, and maternal age to 2 nonsmoking control pregnant women per case, leading to a control group of 12. Results obtained by flow cytometry were reanalyzed using Winlist software and cell subsets were calculated. Lymphocytes were selected based on forward and side scatter plots. Additional gates were used based on fluorescence emission wavelengths of the antibodies, and negative isotype controls were used to set the gates ( Supplemental Figure 1 ).
Female and male C57Bl6 mice, aged 8-10 weeks, were obtained from Harlan (Horst, The Netherlands). All mice were held under specific pathogen-free conditions on a 12:12 light-dark cycle and were administered food and water ad libitum. All animal protocols were approved by the local committee on animal experimentation (University of Groningen, Groningen, The Netherlands) and were performed under strict governmental and international guidelines on animal experimentation.
For experimental purposes, female mice were treated with 1.5 IU pregnant mare’s serum gonadotropin and 1.25 IU human chorionic gonadotrophin to induce simultaneous cycling. To induce pregnancy, 2 females were housed with 1 male. Mating was confirmed by vaginal plug detection. All female mice had been mated, 4 of the 6 smoking females (pregnancy rate 67%), and 5 of the 6 nonsmoking females (pregnancy rate 83%) ended up being pregnant ( Table 2 ). The morning after mating was called day 0.5 of gestation. Mated females were separated from males and were killed at day 11.5 of gestation. At day 11.5 paraaortic lymph nodes (uterus-draining lymph nodes), inguinal (systemic) lymph nodes, and spleens were removed. In addition, implantation sites were removed and fixed in 4% paraformaldehyde, zinc fixative or immediately snap frozen at –80°C for immunohistochemical and real-time RT-PCR analysis.
|Variable||Nonsmoking (n = 6)||Smoking (n = 6)|
|Pregnancy rate, %||83||67|
|Number of implantations||14.4 ± 2.1||14.0 ± 3.4|
|Number of resorptions||0.60 ± 0.6||3.4 ± 1.5 a|
|Viable pups, %||96||74 b|
|Maternal weight, g||21.5 ± 1.4||19.6 ± 0.8|
Cigarette smoke exposure
Cigarette smoke was generated using a TE-10 smoke exposure system of Teague Enterprises Smoke Exposure System (Woodland, CA). Female mice were exposed to fresh air or smoke in sessions of 7 hours continuously per day, from 7 days before mating until the day the animals were killed. Mice were exposed to 5 cigarettes the first day, 10 cigarettes the second day, 25 cigarettes the third day, 50 cigarettes the fourth, and 60 cigarettes on the fifth day and thereafter. Total particulate matter counts were at least 100. Kentucky 2R4F research-reference filtered cigarettes (The Tobacco Research Institute, University of Kentucky, Lexington, KY) were used.
Mouse decidual real-time RT-PCR
Frozen implantation sites were sectioned transversely at 4 μm and stained with hematoxylin. Frozen tissue of 1 control mouse was not available. Decidual and placental tissue was identified, selected, and cut from the slides using laser dissection microscopy (Leica, Rijswijk, The Netherlands). Decidual and placental tissue was collected separately. Total RNA was isolated from decidual and placental tissue using an mRNA easy kit (QIAGEN, Valencia, CA) according to the manufacturer’s instructions. cDNA was reverse transcribed using a Superscript-II reverse transcriptase kit (Invitrogen).
For the housekeeping gene (B2M) and for the other genes (IL6, IL10, Tbet21 [Th1 response], Gata 3 [Th2 response], Rorγt [Th17 response], Foxp3 [Treg cells], Nk1.1 [NK cells], CD68 [pan macrophage marker], NOS2 [M1- macrophages; iNOS], and CD206 [M2-macrophages; MRC-1]), On-Demand gene expression assays were used (Applied Biosystems). PCRs were performed in triplicate in a volume of 10 μL consisting of 2 μL of MilliQ water, 5 μL PCR mix (Eurogentic), 0.5 μL assay mix, and 2.5 μL cDNA. Runs were performed by a 7900HT Fast real-time PCR system (Applied Biosystems), and mRNA data were normalized to B2M mRNA expression using 2 -ΔCt . Undetectable Ct values (greater than 40) were interpreted as the maximum Ct value (40).
Mouse flow cytometry
Immediately after the animals were killed, paraaortic lymph nodes (uterus-draining lymph nodes), inguinal (systemic) lymph nodes, and spleens were removed and thoroughly dispersed into single cell suspensions in fluorescence-activated cell sorting (FACS) buffer (consisting of PBS with 2% fetal calf serum). Red blood cells in cell suspensions were lysed using NH 4 Cl solution.
For flow cytometric analysis, 1 million cells were stained according to standard methods. In short, cells were incubated for 5 minutes with 10% normal mouse serum to block a-specific binding. Cells from the lymph nodes were then incubated with a cocktail of fluorescently labeled antibodies for 30 minutes, including the following: Pacific Blue-labeled anti-CD3 (17A2; BioLegend, San Diego, CA), PerCP-labeled anti-CD4 (RM4-5; BD Pharmingen, San Diego, CA), Alexa 700-labeled anti-CD8 (53-6.7, Biolegend), allophycocyanin (APC)-labeled anti-CD25 (3C7; BD Pharmingen), and PE-Cy7-labeled anti-NK1.1 (PK 136; BioLegend). After surface staining, intracellular staining was performed according to the Foxp3-staining kit instructions (eBioscience, San Diego, CA) using fluorescein isothiocyanate (FITC)-labeled anti-Foxp3 (FJK-16s; eBioscience).
Cells from the spleen were not only incubated with the T-cell antibody cocktail described in the previous text but were also incubated with a monocyte antibody cocktail: PerCP-labeled anti-CD4 (RM4-5; BD Pharmingen), APC-labeled anti-CD11b (M1/70; BD Pharmingen), phycoerythrin (PE)-labeled anti-CD43 (1B11; Biolegend), PE-Cy7-labeled anti-Gr-1 (RB6-8C5; Biolegend), Pacific Blue-labeled anti-F4/80 (BM8; Biolegend), and FITC-labeled anti-MHC-II (2G9; BD Pharmingen).
Cells were analyzed on a LSRII (Beckton Dickinson, Lincoln Park, NJ) flow cytometer. Results were analyzed with Facs Diva software (Beckton Dickinson). Lymphocytes were selected based on forward and side scatter plots and subsequently staining for CD3, CD4, CD25, and Foxp3. To identify cells positive for the labeled antibodies, additional gates were set based on fluorescence emission wavelengths of the used antibodies and negative controls ( Supplemental Figure 1 ). For monocytes all viable cells were gated for CD11b and F4/80. CD11b-hi and F4/80-intermediate cells were considered monocytes, and these were gated for CD43 and GR1 to distinguish the GR1-hi and GR1-lo subset ( Supplemental Figure 2 ).
(Immuno)histochemical analysis mice uterine tissue
Implantation sites were fixed in paraformaldehyde and zinc fixative for 24 hours, transferred into 70% ethanol, and embedded in paraffin according to standard methods and sectioned transversely at 7 μm. Slides were dewaxed in xylene and rehydrated. For general histological examination, sections were stained with hematoxylin-eosin. First, myometrial, decidual, and placental sizes, and ratios of decidual spiral artery vessel to lumen area were measured. T cells and Foxp3+ cells were stained using primary antibodies against CD3 (ab16669; Abcam, Cambridge, MA), and Foxp3 (FJK16s; eBioscience). Antigen retrieval was performed by incubating sections (15 minutes at 400 W) in citrate buffer.
Sections were incubated with primary antibodies diluted 1:400 (CD3) or 1:50 (Foxp3) overnight at 4°C. Thereafter sections were incubated for 30 minutes with the appropriate secondary and tertiary peroxidated antibodies (CD3), or slides were incubated with a biotinylated secondary antibody and were incubated thereafter with a streptavidin-peroxidase complex (Foxp3). Positive staining was localized using diaminobenzidine tetrachloride (DAB). NK cells were stained using biotinylated lectin for dolichos biflorus (Sigma, St. Louis, MO) diluted 1:800 and were incubated overnight at 4°C. Sections were then incubated with a streptavidin-peroxidase complex for 30 minutes at room temperature, and positive staining was localized using 3-amino-9-ethyl-carbazole (AEC).
Subsets of macrophages were stained using primary antibodies against F4/80 (pan macrophage marker) (MCA497; Serotec, Raleigh, NC), IRF5 (M1 macrophage lineage marker) (10547-1-AP; Proteintech, Manchester, UK), and YM-1 (M2 macrophage lineage marker) (Mouse Chitinase 3-like; R&D Systems Europe Ltd, Abingdon, UK). Antigen retrieval was performed by incubating sections (15 minutes at 400W) in citrate buffer. Sections were incubated with diluted antibody solution, 1:50 (F4/80), 1:100 (IRF5), or 1:400 (YM-1) for 1 hour at room temperature (IRF5, YM-1) or were incubated overnight at 4°C (F4/80). Thereafter sections were incubated for 30 minutes with the appropriate secondary and tertiary peroxidated antibodies (IRF5, YM-1), or sections were incubated with a biotinylated secondary antibody and were incubated with ABC elite kit (Vector Laboratories, Burlingame, CA) (F4/80). Positive staining was localized using DAB (IRF5) or AEC (F4/80, YM-1). All sections were counterstained using hematoxylin and all incubation steps were followed by 3 washes in PBS for 5 minutes.
Except for the Foxp3 staining, which was performed in triplicate, every staining was performed in duplicate with at least 7 sections in between to avoid duplicate counting of cells. Sections were scanned using a NanoZoomer (Hamamatsu, Shizuoka, Japan). Morphometric analysis was performed by setting a staining threshold manually, based on negative control sections, and measuring the total stained area using ImageJ (National Institutes of Health, Bethesda, MD).
For statistical analysis, SPSS (SPSS, Inc., Chicago, IL) was used. All real-time RT-PCR data were log transformed before statistical analysis. Statistical comparisons between human smoking and nonsmoking groups were performed using 1-way analysis of variance and Dunnett post hoc testing. Statistical comparisons between smoke-exposed and control mice groups were performed using Student t tests. Pearson correlation coefficients were calculated to analyze the correlation between number of cigarettes and mRNA expression of several genes. P < .05 was considered to be significant, and P < .10 was considered a statistical trend.
Smoking influences the human maternal immune system locally and systemically
Smoking affects immune parameters in human decidual tissue
To investigate the effect of human maternal smoking on local immune parameters during pregnancy, first-trimester decidual tissue was analyzed for mRNA expression of different immune markers. When comparing the total smoking group with the control group, a significantly higher CD56 mRNA expression in smoking women was found as compared with nonsmoking women. No significant differences in expression of the other genes were found ( Figure 1 , A). However, when further dividing the smoking women into a moderate (<10 cigarettes per day) and a heavy smoking (≥10 cigarettes per day) subgroup, a significant increase was seen in IL6, Gata 3 (Th2 response), CD56 (NK cells), and iNOS (M1 macrophages) and a trend toward higher CD68 (pan macrophages) mRNA expression in heavy smoking women compared with nonsmoking women ( Figure 1 , B). Because iNOS (M1 macrophages) was increased in smoking women, we analyzed iNOS/CD68 and CD206/CD68 ratios to identify M1 and M2 macrophages as well as the iNOS/CD206 ratio as a measure for the ratio of M1/M2 macrophages ( Figure 1 , C). In moderate smokers a significant higher CD206/CD68 ratio was found, whereas in heavy smokers significantly increased iNOS/CD206 and iNOS/CD68 macrophage ratios were found ( Figure 1 , C). Interestingly, when calculating correlation coefficients (CC), significant positive correlations between the number of cigarettes and mRNA expressions were found for IL6 (CC 0.64 P < .001), CD56 (CC 0.58, P < .001), and CD68 (CC 0.55, P < .001) (data not shown).
Lower percentages of Treg cells in peripheral blood of smoking pregnant women
To analyze the influence of smoking on the systemic human maternal immune response, data of a previously published study were reanalyzed ( Table 3 ). Smoking pregnant women had significantly lower percentages of Treg cells (CD4+Foxp3+) than nonsmoking pregnant women ( P < .05) ( Table 4 ). No differences in percentages of effector T cells (Teff) (CD4+CD25+Foxp3–) cells were found between smoking and nonsmoking pregnant women ( Table 4 ).
|Characteristic||Nonsmoking (n = 12)||Smoking (n = 6)|
|Age, y||32.8 ± 4.6||30.0 ± 4.2|
|Birthweight, g||3798.6 ± 465.5||3273.0 ± 637.6|
|Sex of child (% male)||53.0||60.0|
|Duration of pregnancy, d||261.7 ± 22.5||265.3 ± 23.3|