Bacterial infections increase risk for pregnancy complications, such as preeclampsia and preterm birth. Unmethylated CpG DNA sequences are present in bacterial DNA and have immunostimulatory effects. Maternal exposure to CpG DNA induces fetal demise and craniofacial malformations; however, the effects of CpG DNA on maternal cardiovascular health have not been examined. We tested the hypothesis that exposure to synthetic CpG oligonucleotides (ODNs) during gestation would increase blood pressure and cause vascular dysfunction in pregnant rats. Pregnant and nonpregnant female rats were treated with CpG ODN (ODN 2395) or saline (Veh) starting on gestational day 14 or corresponding day for the nonpregnant groups. Exposure to CpG ODN increased systolic blood pressure in pregnant (Veh: 121 ± 2 mmHg vs. ODN 2395: 134 ± 2 mmHg, P < 0.05) but not in nonpregnant rats (Veh: 111 ± 2 mmHg vs. ODN 2395: 108 ± 5 mmHg, P > 0.05). Mesenteric resistance arteries from pregnant CpG ODN-treated rats had increased contractile responses to U46619 [thromboxane A2 (TxA2) mimetic] compared with arteries from vehicle-treated rats [Emax (%KCl), Veh: 87 ± 4 vs. ODN 2395: 104 ± 4, P < 0.05]. Nitric oxide synthase (NOS) inhibition increased contractile responses to U46619, and CpG ODN treatment abolished this effect in arteries from pregnant ODN 2395-treated rats. CpG ODN potentiated the involvement of cyclooxygenase (COX) to U46619-induced contractions. In conclusion, exposure to CpG ODN during gestation induces maternal hypertension, augments resistance artery contraction, increases the involvement of COX-dependent mechanisms and reduces the contribution of NOS-dependent mechanisms to TxA2-induced contractions in mesenteric resistance arteries.
- vascular function
- Toll-like receptor 9
NEW & NOTEWORTHY
Hypomethylated CpG sequences are frequent in bacterial, fetal, and mitochondrial DNA, all of which have been implicated in pregnancy complications. This is the first original research paper to provide evidence that exposure to stimulatory CpG oligodeoxynucleotide (ODN) during gestation induces maternal hypertension and vascular dysfunction.
infections and inflammation are associated with high risk of developing pregnancy complications, including preterm labor and preeclampsia (2, 19, 33, 38). Infectious agents express pathogen-associated molecular patterns (PAMPs), which are recognized by pattern-recognition receptors (PRRs) of the innate immune system (29). Toll-like receptors (TLR) are the most widely studied PRRs and their activation leads to signaling cascades that induce the production of inflammatory cytokines and interferon responses (16). In addition to PAMPs, TLRs mount an immune response when they recognize endogenous molecules released from dying cells and damaged tissue [damage-associated molecular patterns (DAMPs)] (16, 27, 28). TLRs are present in the maternal-fetal interface and play a significant role in the establishment of pregnancy (31). Excessive activation of TLRs, however, induces fetal demise and malformations (32) as well as preeclampsia-like symptoms in animal models (10, 40).
Unmethylated CpG motifs act as PAMPs that are exclusively recognized by TLR9 (17). CpG sequences are present at high frequency in bacterial but rare in mammalian DNA (17). Previous studies focusing on the effects of TLR9 activation on pregnancy outcomes reported that in vivo stimulation of TLR9 by CpG DNA led to fetal resorptions and aberrant craniofacial and distal limb development in mice (32). These studies, however, primarily focused on fetal development as the pregnancy outcome and did not address the effects of CpG DNA on maternal cardiovascular health. Poor maternal cardiovascular adaptations during pregnancy are underlying causes of maternal hypertension and are associated with increased maternal risk for future cardiovascular events (1, 13). Thus assessment of fetal growth and development in response to a perturbation during gestation provides a valuable yet incomplete evaluation of its effects on pregnancy outcomes.
Recently, we reported that in vivo treatment of male rats with synthetic oligodeoxynucleotides (ODNs) containing CpG motifs similar to those found in bacterial DNA, elicited hypertension, endothelial dysfunction, and oxidative stress (28). The effects of CpG DNA on the maternal vasculature during gestation, however, are currently unknown. Assessment of these effects is of high clinical significance since CpG ODNs are currently used as vaccine adjuvants for infections and immunotherapeutics against cancer (21). New vaccine technologies that include CpG ODNs as adjuvants may result in the exposure of the maternal and fetal tissues to immunostimulatory CpG sequences (32). Furthermore, hypomethylated CpG sequences are frequent in bacterial, fetal, and mitochondrial DNA, all of which have been affiliated with pregnancy complications (15, 34, 39). In this study, we tested the hypothesis that maternal exposure to CpG ODN would increase blood pressure and cause vascular dysfunction in pregnant Sprague-Dawley rats. We chose a low dose of CpG ODN and initiated the treatment after GD13 to avoid abnormal fetal and placental growth. This dose has been previously shown to activate the TLR9 signaling in isolated vascular tissues (28).
All protocols were approved by the Institutional Animal Care and Use Committee of Augusta University and all procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health.
Twelve-week-old virgin, female, Sprague-Dawley rats (Harlan Laboratories, Indianapolis, IN) were maintained in a temperature- and humidity-controlled environment under 12:12-h light-dark cycles and had free access to tap water and standard laboratory rodent chow. In-house acclimatization lasted 7 days before breeding and final experiments were performed when rats were 16–20 wk old. After a normal estrous cycle was determined using vaginal cytology, half of the female rats were randomly chosen to mate with a fertile male (15–20 wk old; Sprague-Dawley; Harlan Laboratories).
Experimental Design, Timeline, and Animal Treatment
Following overnight mating, vaginal smears were inspected for the presence of spermatozoa using light microscopy. The morning on which spermatozoa were found was considered day 1 of gestation (GD1). Pregnant rats were treated with a synthetic CpG ODN (ODN 2395; Invivogen, San Diego, CA) or saline (vehicle) via three intraperitoneal injections (100 μg/rat each ip) on GD14, GD17, and GD18. A similar, time-controlled, treatment regimen was used for nonpregnant rats. ODN 2395 has unmethylated CpG sequences (5′-tcgtcgttttcggcgc:gcgccg-3′, palindrome is underlined) and is designed with a phosphorothioate-modified backbone to provide nuclease resistance and increase half-life (22). We have previously shown that this treatment protocol elicits blood pressure increases in male rats (28). Furthermore, we have reported that this dose of ODN 2395 is sufficient to increase expression of proteins involved in TLR signaling in isolated mesenteric arteries and that cotreatment with the specific antagonist of TLR-9 (ODN 2088) abolished the effects of ODN 2395 (28). We have compared the effects of ODN 2395 to both nuclear DNA (has low frequency of CpG sequences and contains cytosines that are methylated) and saline in rats and found no differences between the two controls (28). Thus in the present study we employed only the saline control. Blood pressure was measured on GD19. Tissue harvest and vascular experiments were conducted on GD20.
Blood pressure measurements.
Systolic blood pressure (SBP) was measured noninvasively in nonanesthetized rats via the tail-cuff method using volumetric pressure recordings (Kent Scientific, Torrington, CT). The rats were acclimatized in restrainers that allowed unrestricted breathing before the commencement of the measurements. For each animal, we recorded 10 blood pressure values and an average was calculated. We chose the tail-cuff methodology because it does not require a surgical procedure that could lead to activation of the innate immune system due to tissue injury. Since the intervention took place in later pregnancy, there was no time for recovery from a surgical procedure.
Animal euthanization and tissue harvesting.
On GD20 (or the corresponding day for nonpregnant rats), blood was collected via cardiac puncture as rats were under a deep plane of anesthesia induced by isoflurane (5% for induction, 3% for maintenance, 100% oxygen). Rats were immediately euthanized with isoflurane overdose followed by cutting their diaphragm and removing their heart. The aorta, mesenteric arcade, and uterus with fetuses and placentas were excised. Fetuses and corresponding placentas were counted and weighed. Fetuses were then euthanized via decapitation. Arteries were placed in ice-cold physiological salt solution (PSS) of the following composition (in mM): 130 NaCl, 4.7 KCl, 14.9 NaHCO3, 5.5 dextrose, 1.18 KH2PO4, 1.17 MgSO4·7H2O, 1.6 CaCl2, and 0.026 EDTA (all Sigma-Aldrich, St. Louis, MO). Third order mesenteric arteries from the same rat were cleaned from perivascular connective and adipose tissue and were used for the following: 1) vascular reactivity experiments (immediately after the rat was euthanized), 2) immunoblotting (snap frozen in liquid nitrogen and stored at −80°C), and 3) reactive oxygen species (ROS) generation measurements [arteries were embedded in tissue freezing medium (Triangle Biomedical Sciences, Durham, NC) and frozen in liquid nitrogen]. Aortas were also prepared for vascular reactivity studies.
In vitro vascular reactivity.
Arteries were cut 2 mm in length and mounted in a wire myograph system (Danish Myo Technology, Aarhus, Denmark) using two 40-μm wires before resting tension was applied. Optimum resting tension was determined via a length-tension curve. Arterial segments were allowed to equilibrate for 45 min in a tissue bath filled with 5 ml PSS, continuously gassed with 95% O2-5% CO2 at 37°C. Vascular integrity was evaluated by contracting the arterial segments with potassium chloride (KCl; 120 mM). The viability of the vascular endothelium was examined by evaluating relaxation responses to acetylcholine (ACh; 3 × 10−6 M; Sigma-Aldrich) in arteries constricted with phenylephrine (PE; 3 × 10−6 M; Sigma-Aldrich). Concentration-response curves to α-adrenergic receptor stimulation (PE; 3 × 10−8 to 3 × 10−5 M) and thromboxane receptor (TP) activation (U46619, 10−9 to 10−5 M) were subsequently performed. Vascular smooth muscle endothelium-dependent and -independent dilatory responses were evaluated via concentration-response curves to ACh (10−9 to 10−5 M) and sodium nitroprusside (SNP; 10−10 to 10−6 M; Sigma-Aldrich), respectively. Some experiments were performed following a 30-min incubation of the vascular segments with the following inhibitors (or vehicle): NG-nitro-l-arginine [l-NNA; nonselective nitric oxide synthase (NOS) inhibitor; 10−4 M; Sigma-Aldrich], indomethacin [nonselective cyclooxygenase (COX) inhibitor; 10−5 M; Sigma-Aldrich], SC560 (COX-1 inhibitor; 10−6 M; EMD Millipore, Billerica, MA), and NS398 (COX-2 inhibitor; 10−6 M; EMD Millipore). None of the arterial segments were incubated with more than one inhibitor in consequent experiments to avoid carryover effects. All concentration-response curves to dilatory agonists were performed in arteries constricted with U46619 in a concentration that elicited isometric force corresponding to 70–80% of maximum response to KCl (120 mM).
Western blot analysis.
Arteries were homogenized in ice-cold lysis buffer containing T-Per tissue protein extraction solution (Thermo Fisher Scientific, Waltham, MA), 100 mM sodium orthovanadate (Na3VO4; Sigma-Aldrich), 100 mM phenylmethanesulfonyl fluoride (PMSF; Sigma-Aldrich), 1% proteinase inhibitor cocktail (Sigma-Aldrich) and phosphatase inhibitors (sodium fluoride and sodium pyrophosphate; Sigma-Aldrich). Equal amounts of protein (15–20 μg protein/lane) were resolved by electrophoresis on 10% SDS-PAGE gels and then transferred to nitrocellulose or polyvinylidene difluoride (PVDF) membranes. Membranes were incubated with primary antibodies overnight at 4°C. The immunostaining was detected using horseradish peroxidase-conjugated anti-rabbit immunoglobulin G (IgG; GE Healthcare, Buckinghamshire, UK) or anti-mouse IgG (GE Healthcare) for 1 h at room temperature. Results were normalized to the expression of a housekeeping protein (β-actin; Sigma-Aldrich). The following proteins were measured: COX-1 (1:2,000; Cell Signaling Technology, Danvers, MA), COX-2 (1:1,000; BD Biosciences, San Jose, CA), phosphorylated endothelial nitric oxide synthase Ser1177 (eNOSSer1177; 1:1,000; BD Biosciences), and total eNOS (1:1,000; BD Biosciences). Immunoreactive bands were visualized with an enhanced chemiluminescence detection system and quantified using UN-SCAN-IT gel analysis software (v. 6.1; Silk Scientific, Orem, UT).
ROS generation measurements.
ROS generation in mesenteric vessels was measured by staining with dihydroethidium (DHE) (28). Briefly, transverse cross-sections (10 μm) of frozen arteries were equilibrated for 10 min in phosphate buffer in a light protected humidified chamber at 37°C. Buffer containing hydroethidine (1 μmol/l; Sigma-Aldrich) was topically applied to each artery section and slides were then incubated at 37°C for 30 min. Slides were imaged using a Zeiss fluorescence microscope at ×20. Three different sections of each artery were analyzed for the presence of superoxide, as measured by DHE staining using ImageJ.
Sigmoidal curve fitting was performed on concentration-response curve data using GraphPad Prism software (v. 6.0; GraphPad Software, San Diego, CA). The maximal effect generated by the agonist (maximum constriction or dilation) and the EC50 (molar concentration of agonist producing 50% of the maximum response) were determined and presented as Emax and pEC50 (negative logarithm to base 10 of the EC50), respectively. Mean fetal and placental weights were calculated for each litter and the average for each group (vehicle-treated vs. ODN 2395-treated rats) was used for the statistical analysis as described below.
Values are presented as mean ± SE and n represents the number of animals used in each experiment. Group differences in Emax and pEC50 were determined using one-way ANOVA, followed by Tukey's post hoc test. Group differences in fetal and placental weights and protein expression were determined using an unpaired t-test. Three way ANOVA containing one repeated factor were used to determine: 1) the effects of ODN 2395 treatment in concentration-response curves [factor 1 = pregnancy (nonpregnant vs. pregnant), factor 2 = treatment (vehicle vs. ODN 2395), and factor 3 (repeated factor) = concentration]; and 2) the effects of COX and NOS inhibition on concentration response curves to U46619 within pregnant and nonpregnant groups [factor 1 = drug (drug alone vs. drug plus inhibitor), factor 2 = treatment (vehicle vs. ODN 2395), and factor 3 (repeated factor) = concentration]. Mauchy's criterion was used to assess the assumption of sphericity. When significant three-way interactions were found, simple main effects and two-way interactions were assessed. Bonferroni correction was used for multiple comparisons. The significance level of all tests was set at α = 0.05.
ODN 2395 Treatment Increased Blood Pressure in Pregnant but Not in Nonpregnant Rats
Treatment with ODN 2395 increased SBP (Fig. 1A) and spleen weights (Fig. 1B) in pregnant but not in nonpregnant rats. There were no differences in body weights between vehicle- and ODN 2395-treated nonpregnant [vehicle-treated (n = 13): 244 ± 2 g vs. ODN 2395-treated (n = 15): 242 ± 3 g, P > 0.05] and pregnant [vehicle-treated (n = 6): 335 ± 9 g vs. ODN 2395-treated (n = 8): 333 ± 8 g, P > 0.05] rats. Furthermore, treatment with ODN 2395 did not affect maternal heart weights [nonpregnant rats, vehicle-treated (n = 8): 0.68 ± 0.02 g vs. ODN 2395-treated (n = 11): 0.69 ± 0.01 g, P = 0.47; pregnant rats, vehicle-treated (n = 8): 0.77 ± 0.03 g vs. ODN 2395-treated (n = 9): 0.77 ± 0.03 g, P = 0.89] and kidney weights [nonpregnant rats, vehicle-treated (n = 7): 1.44 ± 0.05 g vs. ODN 2395-treated (n = 9): 1.47 ± 0.04 g, P = 0.62; pregnant rats, vehicle-treated (n = 8): 1.56 ± 0.04 g vs. ODN 2395-treated (n = 9): 1.62 ± 0.04 g, P = 0.27] in any group. Maternal treatment with ODN 2395 did not affect litter size [vehicle-treated (n = 6): 12.0 ± 2.0 vs. ODN 2395-treated (n = 8): 12.8 ± 0.5, P = 0.82], fetal weights [vehicle-treated (n = 6): 2.46 ± 0.213 g vs. ODN 2395-treated (n = 8): 2.18 ± 0.029 g, P = 0.24], and placental weights [vehicle-treated (n = 6): 0.52 ± 0.027 g vs. ODN 2395-treated (n = 8): 0.51 ± 0.017 g, P = 0.93] measured on GD20. Thus systemic treatment with ODN 2395 induced maternal hypertension and spleen enlargement but did not affect fetal and placental weights in late pregnancy.
Maternal Treatment with ODN 2395 Increases U46619-Induced Contractions in Mesenteric Resistance Arteries
There were no differences in maximum responses to KCl (120 mM) between groups (data not shown). Mesenteric resistance arteries from pregnant rats treated with ODN 2395 had greater contractile responses to U46619 compared with arteries from vehicle-treated dams, but this effect was not seen in the nonpregnant group (concentration × treatment × pregnancy interaction, P = 0.020; Fig. 2A and Table 1). ODN 2395 did not affect PE-induced contractions in mesenteric arteries from nonpregnant and pregnant groups (P > 0.05, Fig. 2B). In addition, treatment with ODN 2395 did not affect U46619-induced contractions in aortic segments from any group (P > 0.05; Fig. 2C).
Systemic treatment with ODN 2395 did not affect relaxation responses to ACh (pEC50, nonpregnant vehicle-treated: 7.69 ± 0.138 vs. ODN 2395-treated: 7.79 ± 0.138; pregnant vehicle-treated: 7.97 ± 0.143 vs. ODN 2395-treated: 7.83 ± 0.139, P = 0.62) and SNP (pEC50, nonpregnant vehicle-treated: 7.77 ± 0.182 vs. ODN 2395-treated: 7.81 ± 0.196; pregnant vehicle-treated: 8.13 ± 0.213 vs. ODN 2395-treated: 8.05 ± 0.137, P = 0.45) in mesenteric vessels from any group (Fig. 3, A and B). Mesenteric arteries from pregnant rats treated with vehicle had greater sensitivity to SNP at concentrations of 0.01 and 0.3 μM compared with arteries from nonpregnant vehicle-treated rats (Fig. 3B). In summary, in vivo treatment with ODN 2395 increased U46619-induced contractions in mesenteric resistance arteries from pregnant but not from nonpregnant rats and this effect was vascular bed and agonist specific.
ODN 2395 Modulates the Contribution of COX- and NOS-Dependent Mechanisms to U46619-Induced Contractions in Pregnant Mesenteric Arteries
Indomethacin (nonselective COX inhibitor) reduced U46619-induced contractions in mesenteric arteries from nonpregnant rats (Fig. 4A and Table 1). In pregnant rats, however, indomethacin reduced these contractions in ODN 2395-treated but not in vehicle-treated animals (concentration × treatment × inhibitor interaction, P = 0.006; Fig. 4B and Table 1). Selective inhibition of COX-1 by SC560 attenuated contractile responses to U46619 in mesenteric resistance arteries from ODN 2395-treated rats but not in vehicle-treated animals in both nonpregnant and pregnant groups (Fig. 5, A and B, and Table 1). Selective COX-2 inhibition by NS398 reduced contractile responses to U46619 in arteries from all groups (Fig. 5, C and D, and Table 1). COX inhibition abolished the differences in contractile responses to U46619 between arteries from vehicle-treated and ODN 2395-treated dams.
NOS inhibition increased U46619-induced contractions in arteries from nonpregnant and vehicle-treated pregnant rats. This effect was abolished in arteries from ODN 2395-treated rats (Figs. 6, A and B, Table 1).
ODN 2395 Treatment Increased COX-2 Expression in Mesenteric Resistance Arteries
Protein levels of COX-2 were increased in arteries from rats treated with ODN 2395 and this effect was seen in both nonpregnant and pregnant groups (Fig. 7, A and B). There were no changes in COX-1 expression in response to ODN 2395 treatment in any group (Fig. 7, C and D). In addition, eNOS expression and phosphorylation were not affected by in vivo treatment with low dose ODN 2395 (Fig. 7, E and F).
Treatment with ODN 2395 Increased ROS Generation in Mesenteric Resistance Arteries from Pregnant Rats
ROS generation was greater in arteries from pregnant ODN 2395-treated rats compared with arteries from vehicle-treated animals (Fig. 8A). Treatment with ODN 2395 did not affect ROS generation, as indicated by DHE fluorescence, in nonpregnant vessels (Fig. 8B).
In this study, we investigated the effects of maternal exposure to CpG ODN, a TLR9 agonist, on maternal blood pressure and vascular function during gestation. The main findings of our experiments demonstrated that in vivo treatment with CpG ODN: 1) induced maternal hypertension, 2) increased contractile responses in mesenteric resistance arteries, 3) abolished the anticontractile effects of NO in TxA2-induced contractions, and 4) increased vascular expression of COX-2 and potentiated the contribution of COX-derived mechanisms to TxA2-induced contractions in mesenteric resistance arteries.
We observed that treatment with synthetic CpG ODN induced maternal spleen enlargement and hypertension. The spleen plays an important role in the modulation of the immune system and spleen enlargement is often attributed to overactivity such as that seen during a bacterial infection (37). Accordingly, spleen weights may be used as markers of overactivity of the immune system (37). Unmethylated CpG motifs are present in bacterial DNA and have immunogenic properties due to their specificity for TLR9 (17, 34). Thus bacterial infections during pregnancy are likely to expose the maternal immune system to an increased load of unmethylated CpG DNA and excessive activation of TLR9 (39). Previous studies have shown that maternal treatment with a high dose of immunostimulatory synthetic ODNs in early murine pregnancy (GD6) led to fetal demise, high incidence of fetal craniofacial and limb malformations, placental necrosis and calcification, and increased circulating levels of Th1 cytokines (32). Subsequent studies confirmed these results and also demonstrated that the effects of CpG ODN on pregnancy outcomes were TLR9 dependent (39). These studies used a high single dose of CpG ODN (300–400 μg/dam, mice), they initiated maternal treatment in early gestation before placental development was complete, and most importantly, they did not assess maternal health. In our study, we chose a lower dose of CpG ODN (100 μg/dam, rats) and initiated treatment at mid-gestation to avoid fetal and placental defects that could have confounding effects on maternal vascular responses. In contrast to our study, Ito et al. (18) showed that CpG ODN treatment of pregnant mice improved maternal survival and prevented pathogen transmission to offspring following Listeria monocytogenes infection. It should be noted, however, that in our studies, the rats were “naïve,” while Ito et al. used CpG ODN to prime the immune system to exhibit a smaller response in a subsequent stimulus. This difference as well as other differences in the experimental design (i.e., drug dose, timing, duration of treatment) may explain the different outcomes. In addition, Ito et al. did not examine the maternal cardiovascular responses to treatment. In an attempt to report on maternal health, Ito et al. stated “ODN treatment had no discernible effect on the health of the mothers: all remained physically vigorous and gave birth at term” (18).
Interestingly, nonpregnant female rats had no hypertensive response to sustained treatment with CpG ODN, while our previous study demonstrated mild hypertension in male Sprague-Dawley rats treated with CpG ODN (same experimental design with this study) (28). These findings suggest that there may be sex differences in cardiovascular responses associated with innate immune system activation. Further investigation is required to address these observations.
Dams treated with CpG ODN had greater contractile responses in resistance arteries and this effect was pregnancy specific because CpG ODN treatment did not affect vascular function in nonpregnant animals. Others have reported that in vivo treatment with the viral mimetic poly I:C (ligand for TLR3) reduces endothelium-dependent relaxation in conduit arteries from pregnant but not from nonpregnant rats (40). These investigators, however, did not assess function of small vessels that contribute to peripheral vascular resistance and are relevant to blood pressure regulation. Furthermore, Tinsley et al. (40) did not examine contractile responses. We observed that arteries from pregnant rats treated with CpG ODN had greater contractile responses to a TxA2 mimetic compared with controls but there were no differences in contractile responses to α1-adrenergic receptor stimulation between groups. We did not observe a potentiation of U46619-induced contractions by CpG ODN in aortic preparations, suggesting that the effects of CpG ODN on vascular responses to TxA2 were vascular bed specific. TxA2 is a potent vasoconstrictor and plays a role in platelet aggregation. Women with hypertension during pregnancy have increased levels of TxA2 metabolites and isolated omental arteries from women with preeclampsia have increased contractile responses to a TxA2 mimetic (36). In addition, preexposure of isolated arteries to U46619 potentiated contractile responses to other contractile agonists. Our findings on the effects of CpG ODN on mesenteric resistance arteries and aortas do not exclude an effect of this treatment on other vascular beds that were not studied (i.e., renal circulation, uteroplacental, and coronary circulations).
We previously showed that although they were no differences in U46619-induced contractions between resistance arteries from pregnant and nonpregnant rats, pregnancy modified the signal transduction pathways following TP receptor stimulation (14). Thus mesenteric resistance arteries from pregnant rats achieved the same contractions with arteries from virgin animals but these contractions were developed via different mechanisms (14). Pregnancy induces an increase in endothelium-derived relaxing factors, which may desensitize TP receptors and inhibit their kinase-dependent downstream signaling. In this study, CpG ODN abolished the anticontractile effects of NO on U46619-induced contractions in pregnant but not in nonpregnant rats. Phosphorylation of eNOS, however, was not altered in arteries from pregnant ODN 2395-treated rats, suggesting that treatment with CpG ODN may have an effect on NO bioavailability and not on NO biosynthesis. In support of this, arteries from pregnant ODN 2395-treated rats had increased ROS production and this may account for reduced NO bioactivity.
Treatment with CpG ODN increased expression of COX-2 in mesenteric resistance arteries from both pregnant and nonpregnant rats but had no effect on COX-1 expression. It has been reported that activation of TLR9 signaling results in NF-κB translocation to the nucleus and induction of the COX-2 gene (among other inflammatory genes) (42). In vitro studies in isolated vessels showed that stimulation with ODN 2395 (at the same concentration used in our in vivo experiment) activated the TLR9 signaling (28). Thus it is possible that ODN 2395 increased expression of COX-2 in arteries from both groups via activation of TLR9. An alternative interpretation is that the increase in COX-2 expression was driven by ROS (26). Interestingly, resistance-sized vessels from women with preeclampsia have increased expression of COX-2 and this vascular inflammatory phenotype has been linked to neutrophil infiltration (35).
In our studies, COX-2 but not COX-1 contributed to U46619-induced contractions in mesenteric arteries from healthy female rats, regardless if they were pregnant or not. This may be specific to U46619, as previously it has been shown that U46619 induces COX in mesenteric arteries (3). We found that exposure to CpG ODN potentiated the involvement of COX-derived factors to U46619-induced contractions. When we used a specific inhibitor of COX-2, this effect was greater in pregnant compared with nonpregnant animals. Most importantly, inhibition of COX abolished the increased vascular contractile responses to U46619 seen after ODN 2395 treatment. It is noteworthy that increased contractions were seen only in pregnant rats treated with ODN 2395. Thus COX inhibition restored vascular responses to U46619 in pregnant rats. Potential efficacy of COX inhibition in hypertensive disorders of pregnancy is highlighted by the use of low-dose aspirin as a preventative intervention (24). Aspirin inhibits both COX-1 and COX-2. Recently, the American College of Obstetrics and Gynecology recommended the use of low-dose aspirin in women with high risk of developing preeclampsia (24). Yet, our knowledge with regards to the effects of aspirin on vascular function during pregnancy complications is limited.
Mesenteric arteries from pregnant rats treated with CpG ODN had increased ROS compared with nontreated rats and these effects were not seen in nonpregnant animals. Ding et al. (9) reported that ROS induced damage of mitochondrial DNA that led to activation of TLR9 signaling and initiation of an inflammatory process in human umbilical vein endothelial cells. Others have shown that TLR9 activation enhances ROS generation (23), suggesting that ROS is generated as a result of TLR activation. A relationship between COX and ROS has also been reported, with ROS inducing the generation of COX-2 (12). This relationship becomes reciprocal in hypertension (26). ROS quenches NO, reducing its availability and causing a reduction in vasodilatory mechanisms (4). These previous findings and our results suggest that there are complex interactions among TLR9, COX, and ROS that promote vascular dysfunction. Although COX and ROS were increased and NOS-dependent mechanisms were impaired in vessels from pregnant rats, ACh-induced relaxation was not affected by this treatment in our study. This is not surprising since ACh-induced relaxation in mesenteric arteries does not solely rely on NO as in other vessels (i.e., aorta) but other mechanisms such as prostacyclin and nonNO, nonprostacyclin factors (i.e., endothelium-derived hyperpolarization) significantly contribute to this relaxation (7). Thus a compensatory increase in these factors could potentially result in unchanged responses to ACh. This compensation is often seen at the initial stages of vascular disease (11).
Treatment with CpG ODN did not have any effect on maximum responses and EC50 values of the SNP concentration response curves, suggesting no effects on vascular smooth muscle responsiveness to NO. Interestingly, mesenteric arteries from pregnant rats treated with vehicle had greater sensitivity to SNP at concentrations of 0.01 and 0.3 μM compared with arteries from nonpregnant vehicle-treated rats. However, this effect of pregnancy was not seen in ODN 2395-treated animals. These data suggest that the expected increased responsiveness seen in the control animals during pregnancy failed to occur in the ODN 2395-treated rats.
Previous studies showed that TLR3 activation leads to the development of preeclampsia-like symptoms, including proteinuria, vascular dysfunction, and intrauterine growth restriction, in rodents (40). Other TLRs (TLR4, TLR7, and TLR8) have also been implicated in preeclampsia (5, 6, 19, 20, 41). In our study, a low dose of CpG ODN impaired maternal vascular function and increased maternal blood pressure without affecting fetal growth. Thus, and although we demonstrated that CpG ODNs have negative effects on maternal cardiovascular system, it still remains unclear whether CpG ODN/TLR9 interaction plays a causal role in the development of preeclampsia or other hypertensive disorders of pregnancy. Initial observations in our laboratory suggested that treatment of pregnant rats with CpG ODN reduced fetal weights measured at term (15). Substantial changes in rat fetal weights are seen at the end of pregnancy (i.e., more than a 1-gram increase from GD20 to GD21, greater than 40–50% increase). This is important because it is possible that any treatment during pregnancy may have differential effects on fetal weights depending on the gestational day they are recorded. Accordingly, fetal weights at different times of gestation as well as birth weights should be recorded in future studies for a comprehensive assessment of fetal development in response to CpG ODN.
Conclusions and Perspectives
In conclusion, we demonstrated that exposure to CpG ODN during gestation induces maternal hypertension, excess resistance artery constriction and oxidative stress (Fig. 9). Furthermore, CpG ODN altered the involvement of endothelium-derived factors (COX and NOS) to TxA2-induced contractions. To the best of our knowledge, this is the first original research report to examine maternal vascular function in response to CpG ODNs. These findings raise questions with regards to the safety of CpG ODNs containing vaccines during gestation and extend the findings of previous studies that showed negative effects of high doses of CpG ODNs on fetal development (32, 39). Hypomethylated CpG DNA is also present in fetal DNA and circulating cell-free fetal DNA is increased in complicated pregnancies. Maternal treatment with fetal DNA induced a high rate of fetal resorptions in wild type but not in TLR9 KO mice, suggesting that CpG containing fetal DNA induced poor pregnancy outcomes via TLR9 signaling (34). Finally, mitochondrial DNA contain unmethylated CpG sequences and has been shown to have immunogenic effects when released in the circulation from dying cells (30, 34, 43). Circulating and placental levels of mtDNA are actually greater in pregnancies with intrauterine growth restriction, a feature that is often seen in pregnancies with preeclampsia (8, 25). Taken together, these data demonstrate the clinical relevance of studying the effects of CpG DNA on maternal physiology during gestation. Our data suggest that maternal vascular function is a target of CpG ODN. The interactions among CpG DNA, its receptor, and its associated pathway may reveal novel therapeutic targets for hypertensive responses and vascular dysfunction during pregnancy.
This work was supported by American Heart Association (AHA) Grants 13SDG17050056, 13PRE14080019, and 15GNT25700451, the Society of Women's Health Research, and the National Council for Scientific and Technological Development (CNPq)-Brazil.
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: S.G. conception and design of research; S.G., C.F.W., C.G.M., and T.M. performed experiments; S.G. and C.F.W. analyzed data; S.G., C.G.M., T.M., and R.C.W. interpreted results of experiments; S.G. and C.F.W. prepared figures; S.G. drafted manuscript; S.G., C.F.W., C.G.M., T.M., and R.C.W. edited and revised manuscript; S.G., C.F.W., C.G.M., T.M., and R.C.W. approved final version of manuscript.
- Copyright © 2016 the American Physiological Society