Nanomaterial Inhalation During Pregnancy Alters Systemic Vascular Function in a Cyclooxygenase-Dependent Manner

Abstract

Pregnancy requires rapid adaptations in the uterine microcirculation to support fetal development. Nanomaterial inhalation is associated with cardiovascular dysfunction, which may impair gestation. We have shown that maternal nano-titanium dioxide (nano-TiO₂) inhalation impairs microvascular endothelial function in response to arachidonic acid and thromboxane (TXA₂) mimetics. However, the mechanisms underpinning this process are unknown. Therefore, we hypothesize that maternal nano-TiO₂ inhalation during gestation results in uterine microvascular prostacyclin (PGI₂) and TXA₂ dysfunction. Pregnant Sprague-Dawley rats were exposed from gestational day 10–19 to nano-TiO₂ aerosols (12.17  ± 1.67 mg/m³) or filtered air (sham-control). Dams were euthanized on gestational day 20, and serum, uterine radial arterioles, implantation sites, and lungs were collected. Serum was assessed for PGI₂ and TXA₂ metabolites. TXB₂, the stable TXA₂ metabolite, was significantly decreased in nano-TiO₂ exposed dams (597.3 ± 84.4 vs 667.6 ± 45.6 pg/ml), whereas no difference was observed for 6-keto-PGF_1α, the stable PGI₂ metabolite. Radial arteriole pressure myography revealed that nano-TiO₂ exposure caused increased vasoconstriction to the TXA₂ mimetic, U46619, compared with sham-controls (−41.3% ± 4.3% vs −16.8% ± 3.4%). Nano-TiO₂ exposure diminished endothelium-dependent vasodilation to carbaprostacyclin, a PGI₂ receptor agonist, compared with sham-controls (30.0% ± 9.0% vs 53.7% ± 6.0%). Maternal nano-TiO₂ inhalation during gestation decreased nano-TiO₂ female pup weight when compared with sham-control males (3.633 ± 0.064 vs 3.995 ± 0.124 g). Augmented TXA₂ vasoconstriction and decreased PGI₂ vasodilation may lead to decreased placental blood flow and compromise maternofetal exchange of waste and nutrients, which could ultimately impact fetal health outcomes.

Due to unique properties, such as increased strength, enhanced thermal/electric conductivity, and catalytic capacity, the incorporation of nanomaterials has increased in numerous products and processes (Matsoukas et al., 2015). Advanced materials exhibit novel or enhanced properties that are not dependent on nanoscale size (Kennedy et al., 2019). Nanomaterial exposures may occur throughout the life-cycle of advanced materials. This includes manufacturing, transport, use, and/or disposal (Giese et al., 2018). Products such as air/water filters, implantable devices, or contrast agents utilize nano-titanium dioxide (nano-TiO₂; Abukabda et al., 2017; Matsoukas et al., 2015). We previously demonstrated that inhaled nano-TiO₂ alters systemic arteriolar reactivity in the heart, spinotrapezius muscle, and mesentery (Abukabda et al., 2017; Knuckles et al., 2012; LeBlanc et al., 2010). The impact of nano-TiO₂ inhalation in the adult microcirculation is established; however, its influence on biological events where rapid vascular growth and adaptation are requisite, such as gestation, is understudied.

The Developmental Origins of Health and Disease hypothesis proposes that a hostile gestational environment compromises fetal health and predisposes individuals to adult disease (Barker, 1990; Barker and Martyn, 1992). This suggests that negative impacts of maternal toxicant inhalation exposure during gestation may influence offspring health. Pregnancy requires plasticity, adaptability, and rapid responses within the growing uterus to support fetal growth. As implantation and placentation progress, the human maternal vasculature must respond to trophoblast invasion by initiating spiral artery remodeling. The spiral arteries shift from high-resistance, coiled vessels to dilated, low-resistance vessels (Bibeau et al., 2016; Boeldt and Bird, 2017; Osol and Mandala, 2009), allowing for increased blood flow to the forming placenta, and facilitating nutrient/waste exchange with the fetus. In conjunction with functional adjustments, anatomic adaptation, and enlargement of the uterine vascular tree, including uterine arterioles, occurs through angiogenesis and increasing vessel length (Bibeau et al., 2016; Boeldt and Bird, 2017; Osol and Mandala, 2009). In humans, disruption of vessel development, vascular growth, or spiral artery invasion are associated with adverse gestational outcomes such as preeclampsia or fetal growth restriction (Bibeau et al., 2016; Boeldt and Bird, 2017; Giordano et al., 2010).

A potential pathway by which toxicant exposure may impact vascular reactivity is via arachidonic acid (AA) metabolites, which are vasoactive and essential to pregnancy. Phospholipase A₂ (PLA₂) hydrolyzes membrane phospholipids to release AA, while the rate limiting enzymes, cyclooxygenase (COX)-1, and COX-2, metabolize AA into prostaglandin (PGH₂) (Majed and Khalil, 2012). Specific PG synthases form prostacyclin (PGI₂), PGD₂, PGE₂, PGF_2α, or thromboxane A₂ (TXA₂; Majed and Khalil, 2012). Prostacyclin is a potent vasodilator, inhibitor of platelet aggregation (Moncada and Vane, 1978; Walsh, 2004), and is critical for trophoblast-mediated angiogenesis (Wang and Zhao, 2010). PGI₂ effects are counterbalanced by TXA₂, which functions as a potent vasoconstrictor and supports platelet aggregation (Boeldt and Bird, 2017; Moncada and Vane, 1978). Studies in humans and sheep have demonstrated that PGI₂ and TXA₂ are important for the regulation of vascular tone, blood pressure, and feto-placental blood flow (Darby et al., 2020; Mak et al., 1984; Sellers and Stallone, 2008). Rat studies on uteroplacental flow modifications in response to PGI₂ and TXA₂ have not been demonstrated; however, alterations to renal blood flow have been. Renal blood flow in pregnant rats was decreased with the addition of PGI₂, whereas TXA₂ had no effect on total peripheral resistance in pregnant rats, but increased mean arterial pressure (Jackson et al., 1982; Kriston et al., 1999). Disturbances in the balance between PGI₂ and TXA₂ have been associated with arterial thrombosis, vasotonic angina, pulmonary arterial hypertension, preeclampsia, and persistent pulmonary hypertension in the newborn (Majed and Khalil, 2012; Moncada and Vane, 1978; Walsh, 2004). In humans, preeclampsia has been associated with diminished PGI₂ concentrations and increased TXA₂ production (Walsh, 2004). Thus, an imbalance between PGI₂ and TXA₂ may contribute to the clinical symptoms observed in preeclampsia: hypertension, decreased uteroplacental blood flow, and platelet aggregation (Walsh, 2004).

We have reported (Knuckles et al., 2012) that in vivo spinotrapezius muscle arterioles had increased sensitivity to TX mimetic, U46619, and decreased sensitivity to the PGI₂ analog, iloprost after nano-TiO₂ inhalation exposure. Due to uterine microvascular reactivity changes after maternal nano-TiO₂ inhalation exposure (Bowdridge et al., 2019), and the importance of PGI₂ and TXA₂ during pregnancy, we hypothesized that maternal nano-TiO₂ inhalation during gestation alters uterine microvascular PGI₂ and TXA₂ reactivity. We tested this by examining the impacts of maternal nano-TiO₂ inhalation during gestation on microvascular reactivity adaptations in uterine arterioles, serum metabolite levels, and COX and phospholipase A₂ enzymatic activity.

MATERIALS AND METHODS

 

 

Animal model

Female Sprague-Dawley (SD) rats were purchased from Hilltop Laboratories (Scottdale, Pennsylvania) and single-housed in an American Association for Accreditation of Laboratory Animal Care (AAALAC) approved facility at West Virginia University (WVU). Rats were maintained under a regulated temperature of 20–26°C, relative humidity of 30–70%, and a 12:12-h light-dark cycle. Once acclimated for 48–72 h, rats were randomly assigned to either air (sham-control; N = 46) or nano-TiO₂ (N = 46) exposure groups before mating. Rat cages were lined with standard bedding (0.25-inch corncob), and had ad libitum access to standard chow (2918X; Envigo, Indianapolis, Indiana) and water throughout the acclimation and mating periods. Mating males (N = 12) ranged in age from 3 to 9 months, and females (N = 92) were 2–4 months of age at mating. Females were vaginally swabbed to assess for stage of estrous. When found in estrus, females were placed with single-housed male for up to 24 h, and then swabbed to check for sperm within the reproductive tract.

To increase the likelihood of a viable pregnancy and pups to study, pregnant rats were exposed after implantation during mid to late gestation (gestational day [GD] 10–19) for a total of 6 days. On GD20, rats were weighed and then anesthetized with isoflurane gas (5% induction, <5 min to effect, 2–3.5% maintenance for subsequent procedures). Animals were placed on a warm heating pad and maintained at a rectal temperature of 37°C. Blood, uterine artery, and uterine radial arteriole samples were collected, and rats were euthanized. All procedures were approved by the WVU Institutional Animal Care and Use Committee.

Nanomaterial

Nano-TiO₂ powder was obtained from Evonik (P25 Aeroxide TiO₂, Parsippany, New Jersey) and is composed of a mixture of anatase (80%) and rutile (20%) TiO₂. Particle characteristics have previously been determined, including primary particle size (21 nm), specific surface area (48.08 m²/g), and Zeta potential (−56.6 mV; Stapleton et al., 2018; Yi et al., 2013).

Inhalation exposure and aerosol characterization

Nano-TiO₂ aerosols were created with a high-pressure acoustical generator (IEStechno, Morgantown, West Virginia). The generator output was fed into a Venturi pump (JS-60M, Vaccon, Medway, Massachusetts) which further de-agglomerates particles, after which the nano-TiO₂ aerosol mix enters the whole-body exposure chamber. A personal DataRAM (pDR-1500; Thermo Environmental Instruments Inc, Franklin, Massachusetts) sampled air in the exposure chamber to determine aerosol mass concentration in real-time. Feedback loops within the software automatically adjust the acoustic energy needed to maintain a stable mass aerosol concentration throughout the exposure. Gravimetric aerosol sampling measurements were conducted with Teflon filters concurrently with the DataRAM measurements to obtain calibration factors. Gravimetric measurements were taken during each exposure to calculate the mass concentration measurement. Aerosol size distributions were measured in real-time in the exposure chamber at a target mass concentration of 12 mg/m³ via: (1) a high-resolution electrical low-pressure impactor (ELPI+; Dekati, Tampere, Finland); (2) a scanning particle mobility sizer (SMPS 3938; TSI Inc., St Paul, Minnesota); (3) an aerodynamic particle sizer (APS 3321; TSI Inc.); and (4) a micro-orifice uniform deposit impactor (MOUDI 115R, MSP Corp, Shoreview, Minnesota). Bedding material was soaked to maintain proper humidity (20–70%). Sham-control animals were exposed to high-efficiency particulate absorbance (HEPA)-filtered air only, with similar temperature and humidity conditions within an exposure chamber that was used only for sham-control animals.

Inhalation exposures were performed for 6 nonconsecutive days from GD 10 to 19 to maximize the chance of healthy gestations. Pregnant rats were exposed to a target concentration of 12 mg/m³, a concentration identical to that used previously for late gestation inhalation exposure studies (Stapleton and Nurkiewicz, 2014; Stapleton et al., 2013, 2018). To estimate lung deposition (dose) with nano-TiO₂ aerosols, the equation D = F•V•C•T was used (Nurkiewicz et al., 2008) where F is the deposition fraction (10%), V is the minute ventilation (208.3 cc), C is mass concentration (mg/m³), and T equals the exposure duration (min) (Yi et al., 2013). The exposure paradigm (12 mg/m³, 6 h/exposure, 6 days) produced a calculated cumulative lung deposition of 525 ± 16 µg with the last exposure occurring 24-h prior to the study on GD19. The calculations represent total lung deposition and do not account for lung clearance (MPPD Software v 2.11, Arlington, Virginia).

Pressure myography with isolated uterine microvessels

Once placentas and pups were removed, uteri were placed into a dissecting dish with physiological salt solution (PSS, in mmol/l: 129.8 NaCl, 5.4 KCl, 0.5 NaH₂PO₄, 0.83 MgSO₄, 19.0 NaHCO₃, 1.8 CaCl₂, and 5.5 glucose) and maintained at 4°C. A segment of the uterine radial arteriole was removed and transferred to a vessel chamber (Living Systems Instrumentation, Burlington, Vermont) containing oxygenated (21% O₂/5% CO₂) PSS, where it was cannulated with glass pipettes, and secured by nylon suture (11-0 ophthalmic, Alcon, UK). Arterioles were extended to their in-situ length, pressurized to 60 mm Hg, superfused with warmed (37°C) oxygenated PSS at a rate of 10 ml/min, and allowed to develop spontaneous tone. Internal and external arteriolar diameters were measured using video calipers (Colorado Video, Boulder, Colorado). Vascular tone ranges from 0% (maximum dilation) to 100% (maximal constriction), and only vessels with ≥20% spontaneous tone were included in this study. After equilibration, arteriolar vasoreactivity was analyzed.

Vasoreactivity assessments

Radial arterioles were treated abluminally with pharmacological agents to assess inner diameter changes via video calipers. Increasing concentrations of the endothelium-dependent vasodilator, acetylcholine (ACh: 10⁻⁹–10⁻⁴ M), endothelium-independent vasodilator, s-nitroso-N-acetyl-DL-penicillamine (SNAP: 10⁻⁹–10⁻⁴ M), and the COX metabolite vasodilator, carbaprostacyclin (10⁻²⁰–10⁻⁴ M) were used. Vasoconstrictive agents included the α₁-adrenergic vasoconstrictor, phenylephrine (PE: 10⁻⁹–10⁻⁴ M), and the COX metabolite vasoconstrictor, thromboxane (U46619, mimetic: 10⁻⁹–10⁻⁴ M). A selective, competitive antagonist for the prostacyclin receptor was also utilized (R01138452: 10⁻⁵ M) and the TXA₂ receptor was inhibited by the selective, competitive antagonist SQ29 548 (10⁻⁸ M). Inhibitor doses were based on a preliminary dose-response study (R01138452: 10⁻⁹–10⁻³ M; SQ29 548: 10⁻¹⁴–10⁻⁸ M) to determine the single most effective dose to prevent uterine vascular response from occurring in the presence of carbaprostacyclin or U46619, respectively. Steady-state diameter was recorded for at least 2 min after each dose, or 10 min after the addition of receptor antagonists. Once the dose-response curve was completed, the vessel chamber was washed with PSS to remove any excess drug and PSS was replaced with Ca²⁺-free PSS until maximum passive diameter was established. As mentioned above, arterioles with <20% spontaneous tone were not analyzed.

Calculations

Data are expressed as mean ± standard error of the mean (SEM). Spontaneous tone was calculated by the following equation:

$$\mathrm{Spontaneous}\mathrm{ }\mathrm{tone}\mathrm{ }\left(\mathrm{\%}\right)=\left\{\frac{\left(D_{\mathrm{m}}-D_{\mathrm{i}}\right)}{D_{\mathrm{i}}}\right\}\times 100$$

where D_m is the maximal diameter in response to calcium free and D_i is the initial steady-state diameter recorded prior to the experiments.

Active responses to pressure were normalized to the maximal diameter using the following formula:

$$\mathrm{Normalized}\mathrm{ }\mathrm{diameter}=D_{\mathrm{ss}}/D_{\mathrm{m}}$$

where D_ss is the steady-state diameter recorded during each pressure change.

The experimental response to ACh, SNAP, and PGI₂ are expressed as the following:

$$\mathrm{Diameter}\mathrm{ }\left(\mathrm{percent}\mathrm{ }\mathrm{maximal}\mathrm{ }\mathrm{dilation}\right)=\left\{\frac{\left(D_{\mathrm{ss}}-D_{\mathrm{con}}\right)}{(D_{\mathrm{m}}-D_{\mathrm{con}})}\right\}\times 100$$

where D_con is the control diameter recorded prior to the dose-response curve and D_ss is the steady-state diameter at each dose of the curve.

The experimental response to PE and U46619 is expressed using the following equation:

$$\mathrm{Diameter}\mathrm{ }\left(\mathrm{percent}\mathrm{ }\mathrm{maximal}\mathrm{ }\mathrm{constriction}\right)=-\left\{\frac{\left(D_{\mathrm{con}}-D_{\mathrm{ss}}\right)}{\left(D_{\mathrm{con}}\right)}\right\}\times 100$$

Wall thickness (WT) was calculated from the measurement of both inner (ID) and outer (OD) steady-state arteriolar diameters at the end of the Ca²⁺ free PSS wash using the following equation:

$$\mathrm{WT}=(\mathrm{OD}-\mathrm{ID})/2$$

Wall-to-lumen ratio (WLR) was calculated using the following equation:

$$\mathrm{WLR}=\frac{\mathrm{WT}}{\mathrm{ID}}$$

Although assessing vascular responses to the COX metabolites, distinct subpopulations of vessels became apparent that differed in their responsiveness to both PGI₂ and TXA₂. To determine if a vessel qualified as hyperresponsive, the following equation was used:

$$\mathrm{Thromboxane}\mathrm{ }\mathrm{Response}=-\left(\frac{D_{\max }-\mathrm{ }{D}_{\mathrm{con}}}{D_{\max }}\right)\times 100$$

$$\mathrm{Prostacyclin}\mathrm{ }\mathrm{Response}=\left(\frac{(D_{\max }-\mathrm{ }{D}_{\mathrm{con}})}{(D_{\mathrm{m}}-\mathrm{ }{D}_{\mathrm{con}})}\right)\times 100$$

where D_max is the diameter at the maximum dose response and D_con is the control diameter recorded prior to the dose-response curve. Once the responsiveness was calculated, vessels with percent maximal change >20% were considered hyperresponsive. Within the nano-TiO₂ exposed groups, these hyperresponsive vessels made up nearly 50% of the vessels tested, whereas within sham-controls, they constituted around 5% of vessels (n = 2).

Enzyme-linked immunosorbent assays

Blood samples (3 ml) were obtained from carotid cannulas placed during anesthesia, and serum was collected. Enzyme-linked immunosorbent assays (ELISAs) were performed on maternal serum for 6-keto-PGF_1α (stable PGI₂ metabolite) and TXB₂ (stable TXA₂ metabolite). All assays were performed according to manufacturer recommendations (Cayman Chemical Company, Ann Arbor, Michigan).

Enzymatic activity assay

Uterine arteries, implantation sites, and lungs were collected and individually homogenized in PBS (Thermo Fisher Scientific, Waltham, Massachusetts) containing a protease and phosphatase cocktail (Thermo Fisher Scientific). Tissue was homogenized with a bead mill homogenizer containing 1.6 mm stainless steel beads (Next Advance, Troy, New York) and spun in a refrigerated tabletop centrifuge (Hettich Lab Instrument AB, Stockholm, Sweden) at 14 000 × g for 15 min. Supernatant was collected, kept on ice, and total protein concentrations measured according to the Bradford method (Bradford, 1976). Collected protein samples were stored at −80°C until assayed for PLA₂ enzymatic activity (Thermo Fisher Scientific) and COX activity (Abcam, Cambridge, UK). To measure activity specific for cPLA₂, a sPLA₂ inhibitor (thioetheramide-PC) and an iPLA₂ inhibitor (bromoenol lactone) were utilized (Yano et al., 2011). All assays were performed according to manufacturer recommendations.

mRNA analysis by real-time PCR

Uterine arteries, implantation sites, lungs, and ovarian tissue were collected, flash frozen in liquid nitrogen, and stored at −80°C until processed for mRNA extraction. Total RNA was extracted via the RNeasy Kit (Qiagen, Hilden, Germany) based on manufacturer’s recommendations. Total RNA was then transcribed to cDNA via the High-Capacity RNA-to-cDNA kit (Thermo Fisher Scientific).

Using real-time PCR, we assessed genes involved in the production of prostacyclin, thromboxane, and/or their receptors: COX-1 (Ptgs1), prostacyclin synthase (Ptgis), thromboxane synthase (TXAS; Tbxas1), prostacyclin receptor (Ptgir), and thromboxane receptor (Tbxa2r). Primers were purchased from Integrated DNA Technologies (Coralville, Iowa); GenBank accession numbers, primer sequences, and references are listed in Table 1. Samples were analyzed in triplicate using iTAQ Universal SYBR Green Supermix (Bio-Rad Laboratories, Hercules, California). Relative fold changes in expression of candidate genes were obtained using the 2−ΔΔC_t method (Livak and Schmittgen, 2001). The C_t values were used to calculate ΔC_t values for genes of interest [C_t(test) – C_t(housekeeping)]. The housekeeping gene for normalization was β-actin (Actb). Statistical representation for each gene is based on fold change with respect to β-actin.

Table 1.

Real-Time PCR Primers, Their Sequence, Length, and References

GenBank	Forward Primer (5′-3′)	Reverse Primer (5′-3′)	Length (bp)	Reference
Actb NM_031144	GTAGCCATCCAGGCTGTGTTG	TGCCAGTGGTACGACCAGAG	1296	Zhang et al. (2017)
Ptgs1 NM_017043	CTCACAGTGCGGTCCAAC	CCAGCACCTGGTACTTAAG		Bernard et al. (2000)
Ptgis NM_031557	TGGTGTGGGATCTGCGTACA	CCTCCACTCCATACAGGGTCA	567	Numaguchi et al. (1999)
Ptgir NM_001077644	TGCTGGAACATCACCTACGT	GTTTCGAGCATAGGCCACAA	422	Numaguchi et al. (1999)
Tbxas1 D31798	GGGGCTTCTCAAGTTCGAAGT	CCCAACTTCCTCAGTCTTGAG	117	Tang and Vanhoutte, (2008)
Tbxa2r NM_017054.1	ATCTCCCATCTTGCCATAGTCC	CCGATGATCCTTGAGCCTAAAG	1880	Zhang et al. (2017)

Primers were used in lung, ovarian, uterine artery, and implantation site tissue.

Statistics

Dose-response curves were analyzed using two-way analysis of variance (ANOVA) with repeated measures and a Šídák’s multiple comparisons analysis for point-to-point differences when main effect significance was found. Uterine radial arteriolar characteristics were assessed via one-way ANOVA utilizing a Tukey post hoc analysis when statistical significance was found. Litter characteristics were assessed via two-way ANOVA utilizing a Tukey post hoc analysis when statistical significance was found. Dam characteristics, ELISA concentrations, enzymatic activities, and mRNA analyses were assessed via unpaired t test with Welch’s correction. All statistical analyses were performed with Graph Pad Prism 9 (San Diego, California). Significance was set at p ≤ .05, with N representing the number of dams per group and n being the number of vessels. The data in this study are reported as the mean ± SEM, unless stated otherwise.

RESULTS

 

Nanoparticle Aerosol Characteristics

The average real-time aerosol mass concentration throughout exposures was 12.17 mg/m³ with a SD of 1.67 (Figure 1A). The aerosol mobility diameter measured via SMPS and APS, had a count median diameter (CMD) of 116  ± 2.11 nm (Figure 1B). The aerosol aerodynamic diameter, assessed via ELPI, had a CMD of 169  ± 1.94 nm (Figure 1C). Aerosol size distribution, assessed by the MOUDI, had a mass median aerodynamic diameter of 1.03 µm with a geometric SD of 2.57 (Figure 1D). We have extensively characterized the morphology of agglomerates via electron microscopy (Abukabda et al., 2019; Bowdridge et al., 2019).

Figure 1.

Nano-TiO₂ aerosol real-time characterizations. Nano-TiO₂ aerosol characteristics were monitored and verified throughout exposure. Red lines in the figures represent size distribution curves based on a log normal fit of the size data, unless otherwise specified. A, Aerosol mass concentration was controlled via software over the 6-h exposure paradigm. Target average concentration (red line) real-time nano-TiO₂ aerosol mass concentration measurements (black line) were maintained at a target of 12 mg/m³. B, Aerosol agglomerate diameter was assessed by a SMPS (light gray) and an APS (dark gray). Based on results from the SMPS and APS the particle diameter, CMD = 116 nm, and had a GSD = 2.11. C, Electrical Low-Pressure Impactor (ELPI, light gray) assessed aerosol aerodynamic diameter at a CMD of 169 nm and a geometric SD of 1.94. D, Aerosol mass size distribution was evaluated by a Nano MOUDI; particles had an MMAD of 1.03 µm and a geometric SD of 2.57. Abbreviations: Nano-TiO₂, nano-titanium dioxide; SMPS, scanning mobility particle sizer; APS, aerodynamic particle sizer; CMD, count median diameter; GSD, geometric standard deviation; MMAD, mass median aerodynamic diameter; MOUDI, Micro-Orifice Uniform Deposit Impactor. A color version of this figure appears in the online version of this article.

Pregnant Rat and Litter Characteristics

Dams showed no difference in age or weight on GD20 between groups (Table 2). There were no significant differences in litter size or number of pups per horn between treatments (Table 2). Resorption sites (right horn) were significantly decreased in the nano-TiO₂ exposed group compared with sham-controls (Table 3). There was no difference between sham-control (N = 15) or nano-TiO₂ exposed (N = 20) litters for wet pup weight, wet placenta weight, dry pup weight, dry placenta weight, and placental efficiency (Table 3). The ratio of male to female pups did not significantly differ between treatment groups. Litter characteristics were split based on fetal sex to determine if nano-TiO₂ inhalation exposure during gestation negatively impacted fetal development in a sexually dimorphic manner. When the interaction of fetal sex and exposure were considered (Table 3), there was a significant decrease in nano-TiO₂ fetal female wet pup weight (3.633 ± 0.064 g) compared with nano-TiO₂ fetal males (3.998 ± 0.089 g) and sham-control fetal males (3.995 ± 0.124 g). The nano-TiO₂ fetal female dry pup weight was also significantly decreased compared with the nano-TiO₂ fetal male counterparts (0.4622 ± 0.014 vs 0.506 ± 0.013 g; Table 3). There was no significant difference for wet placenta, dry placenta, or placental efficiency when accounting for fetal sex. These data demonstrate that nano-TiO₂ inhalation exposure during gestation negatively impacts fetal growth in a sexually dimorphic manner.

Table 2.

Dam Physical Characteristics on GD20

Exposure	Dams (N)	Weight (g)	Age (Days)	Litter Size	Litter Distribution
					Left Horn	Right Horn
Sham-Control	46	349.5 ± 5.0	86.8 ± 3.8	11.2 ± 0.5	5.5 ± 0.4	5.6 ± 0.4
Nano-TiO2 exposed	46	347.6 ± 4.7	83.2 ± 3.6	10.7 ± 0.6	5.3 ± 0.4	5.4 ± 0.4

There were no significant differences between sham-control and nano-TiO₂ dams in weight, age, or litter size. Data are mean ± SEM.

Abbreviations: GD, gestational day; nano-TiO₂, nano-titanium dioxide; SEM, standard error of the mean.

Table 3.

Litter Characteristics Include Wet Pup and Placenta Weight, Dry Pup and Placenta Weight, and the Litters Sex Ratio Between Groups

Exposure	N	Wet Weight (g)		Dry Weight (g)		Placental Efficiency	Sex Ratio		Resorption Sites
		Pup	Placenta	Pup	Placenta		Male	Female	Left Horn	Right Horn
Sham-control	15	3.760 ± 0.074	0.758 ± 0.033	0.483 ± 0.009	0.107 ± 0.004	5.056 ± 0.216	0.482 ± 0.035	0.504 ± 0.034	0.500 ± 0.154	0.750 ± 0.228
Nano-TiO2 exposed	20	3.720 ± 0.060	0.716 ± 0.027	0.493 ± 0.011	0.113 ± 0.008	5.152 ± 0.304	0.482 ± 0.035	0.504 ± 0.034	0.286 ± 0.122	0.143 ± 0.078 *
Sham-control fetal males	15	3.995 ± 0.124	0.760 ± 0.032	0.499 ± 0.008	0.108 ± 0.005	5.168 ± 0.214	—	—	—	—
Sham-control fetal females	15	3.837 ± 0.107	0.757 ± 0.038	0.474 ± 0.009	0.109 ± 0.005	4.745 ± 0.184	—	—	—	—
Nano-TiO2 exposed fetal males	20	3.998 ± 0.089	0.709 ± 0.028	0.506 ± 0.013	0.104 ± 0.003	5.435 ± 0.288	—	—	—	—
Nano-TiO2 exposed fetal females	20	3.633 ± 0.064*,^	0.734 ± 0.29	0.462 ± 0.014^	0.106 ± 0.003	4.956 ± 0.297	—	—	—	—

The second portion is fetal characteristics based on the fetal sex for the litter average. N is the number of dams. Bold font is used to draw attention to the fact that the data points are significant. Data are mean ± SEM.

Abbreviations: GD, gestational day; nano-TiO₂, nano-titanium dioxide; SEM, standard error of the mean.

^* p ≤ .05 versus sham-control fetal males;

^{^} p ≤ .05 versus nano-TiO₂ fetal males.

When the uterine radial arteriole reactivity to PGI₂ and TXA₂ are considered, litter characteristics were split based on the corresponding uterine arteriole response to either PGI₂ or TXA₂. Based on the uterine arteriolar response, litter characteristics were split by augmented PGI₂ nano-TiO₂ exposed, nonaugmented PGI₂ nano-TiO₂ exposed, augmented TXA₂ nano-TiO₂ exposed, and nonaugmented PGI₂ nano-TiO₂ exposed. Litter characteristics based on corresponding uterine arteriolar response were not significantly different in any pup or placental measurements (data not included).

Arteriolar Characteristics

Resting characteristics of radial arterioles for sham-control and nano-TiO₂ exposed groups are shown in Table 4. There were no differences between groups for inner diameter, outer diameter, WT, or WLR. Based on results obtained from uterine radial arteriole reactivity, it was determined there were subpopulations of vessels that displayed augmented or nonaugmented responses to TXA₂ and PGI₂. Due to these unique responses, arteriolar characteristics were split based on radial arteriole responses to TXA₂ or PGI₂. There was a significant increase in tone for PGI₂ nonaugmented nano-TiO₂ exposed vessels compared with sham-control and total nano-TiO₂ exposed vessels. Passive inner and outer diameter were significantly decreased in the TXA₂ augmented nano-TiO₂ exposed vessels compared with both TXA₂ nonaugmented nano-TiO₂ exposed vessels and PGI₂ augmented nano-TiO₂ exposed vessels. These data provide evidence that maternal nano-TiO₂ exposure during gestation leads to anatomical changes within the uterine arteriole vessels.

Table 4.

Uterine Radial Arteriolar Characteristics on GD 20

Exposure	n	Inner Diameter (µm)	Outer Diameter (µm)	Tone %	Passive Diameter Inner (µm)	Passive Diameter Outer (µm)	Wall Thickness	Wall-to-Lumen Ratio
Sham-control	28	97.00 ± 5.362	131.3 ± 7.570	22.64 ± 0.569	128.8 ± 6.837	156.3 ± 8.340	12.84 ± 0.949	0.082 ± 0.003
Nano-TiO2 exposed	33	88.52 ± 4.238	117.9 ± 4.721	23.45 ± 0.704	118.7 ± 4.952	144.2 ± 6.127	12.60 ± 0.813	0.090 ± 0.004
Thromboxane nonaugmented nano-TiO2 exposed	18	100.1 ± 6.835	137.9 ± 7.904	22.12 ± 0.837	144.3 ± 9.019	175.0 ± 11.09	14.47 ± 1.186	0.106 ± 0.008
Thromboxane augmented nano-TiO2 exposed	17	83.94 ± 6.185	112.8 ± 7.317	25.53 ± 1.186	100.5 ± 3.396Φ , †	119.6 ± 4.071Φ , †	9.938 ± 0.649	0.087 ± 0.005
Prostacyclin nonaugmented nano-TiO2 exposed	15	84.27 ± 5.027	121.3 ± 7.525	30.80 ± 2.909* , ^	123.7 ± 7.481	148.4 ± 9.294	12.37 ± 1.173	0.099 ± 0.007
Prostacyclin augmented nano-TiO2 exposed	19	106.2 ± 7.755	143.8 ± 9.594	23.66 ± 0.952	139.2 ± 10.38Φ	167.0 ± 12.39Φ	12.91 ± 1.133	0.120 ± 0.008

n represents the number of vessels assessed. Bold font is used to draw attention to the fact that the data points are significant. Data are mean ± SEM.

Abbreviations: GD, gestational day; nano-TiO₂, nano-titanium dioxide; SEM, standard error of the mean; TXA₂, thromboxane A₂; PGI₂, prostacyclin I₂.

^* p ≤ .05 versus sham-control;

^{^} p ≤ .05 versus nano-TiO₂ exposed;

^Φ p ≤ .05 versus TXA₂ nonaugmented nano-TiO₂ exposed;

^† p ≤ .05 versus PGI₂ augmented nano-TiO₂.

Uterine Radial Arteriole Reactivity

Arteriolar responses to the endothelium-dependent vasodilator ACh (Figure 2A), endothelium-independent vasodilator SNAP (Figure 2B), and the adrenergic vasoconstrictor PE (Figure 2C), did not differ significantly between groups.

Figure 2.

Vascular reactivity of uterine radial arterioles. Dose-response curves were performed to assess vascular reactivity of uterine radial arterioles (n = 29 for sham-control and n = 18 for nano-TiO₂). A, Vascular reactivity following increasing doses of ACh, an endothelium-dependent vasodilator, (B) SNAP, an endothelium-independent vasodilator, and (C) PE, vasoconstrictor. *p ≤ .05 versus sham-control. Abbreviations: Nano-TiO₂, nano-titanium dioxide; Ach, acetylcholine; SNAP, s-nitroso-N-acetyl-DL-penicillamine; PE, phenylephrine.

Within the exposed vessels, a subpopulation of vessels that displayed an augmented vasoconstrictive response to TXA₂ was observed. We defined this as vessels that exceeded a 20% change in maximal response tone compared with their initial measurement before TXA₂ treatment. The maximal vasoconstrictive response to TXA₂ in this hyper-reactive subset of nano-TiO₂ exposed vessels was significantly greater compared with sham-control vessels (−41.3% ± 4.3% vs −16.8% ± 3.4%). These augmented vessels also exhibited a significant increase in cumulative vasoconstriction compared with nonaugmented nano-TiO₂ exposed vessels (−41.3% ± 4.3% vs −10.0% ± 2.0%). There was no difference between nonaugmented nano-TiO₂ exposed vessels and sham-control vessels. (Figs. 3A and 3B).

Figure 3.

Cyclooxygenase metabolites vascular radial reactivity. Subpopulations of nano-TiO₂ exposed uterine radial arterioles displayed hyperreactive responses to both COX metabolites. This subpopulation of vessels is shown herein as a separate group, and dose-response curves were assessed for vascular reactivity. A, Vascular responses following increasing doses of U46619, a TXA₂ mimetic (n = 29 for sham-control, n = 8 for nonaugmented nano-TiO₂, and n = 9 for augmented nano-TiO₂). B, Maximum response to U46619 across the dose-response curve. C, Vascular reactivity following increasing doses of carbaprostacyclin, a stable PGI₂ agonist (n = 29 for sham-control, n = 5 for nonaugmented nano-TiO₂, and n = 11 for augmented nano-TiO₂). D, Maximum response to carbaprostacyclin across the dose-response curve. *p ≤ .05 versus sham-control. ^p ≤ .05 versus nano-TiO₂. Abbreviations: nano-TiO₂, nano-titanium dioxide; COX, cyclooxygenase; TXA₂, thromboxane A₂; PGI₂, prostacyclin I₂.

A similar phenomenon occurred in nano-TiO₂ exposed vessels in the presence of PGI₂, with the augmented vessels presenting increased vasodilation (Figure 3C). At the maximal response dose, there was a significant difference between sham-controls and the augmented nano-TiO₂ exposed group (30.0% ± 9.0% vs 53.7% ± 6.0%; Figure 3D). There was also a significant difference between sham-control and nonaugmented nano-TiO₂ exposed vessels (30.0% ± 9.0% vs 1.0% ± 7.0%). Finally, the augmented nano-TiO₂ exposed vessels were significantly different when compared with nonaugmented nano-TiO₂ exposed vessels (53.7% ± 6.0% vs 1.0% ± 7.0%). It should be noted that 21% of all exposed vessels were hyperresponsive to PGI₂ and TXA₂, but the remainder only exhibited an augmented response to 1 agonist.

Arteriolar vasoconstriction to U46619 was tested in the presence of the thromboxane receptor antagonist, SQ29,548 for sham-control and nano-TiO₂-exposed groups. This inhibitor was utilized to test for the nonreceptor mediated effects of TXA₂. There was a significant difference between nonaugmented nano-TiO₂ alone and nonaugmented nano-TiO₂ with SQ29,548 vessels at the 10^{−6,−5,−4} doses (Figure 4A). At the maximal dose of U46619, there was a significant decrease in vasoconstriction of sham-control with SQ29,548 vessels compared with sham-control vessels alone (−5.4% ± 1.2% vs −15.4% ± 2.7%). At the maximal dose of U46619, there was also a significant decrease in vasoconstriction of nano-TiO₂ with SQ29,548 vessels versus sham-control arterioles (−4.3% ± 0.9% vs −15.4% ± 2.7%; Figure 4B). Figure 4 also highlights the uterine radial arteriole vasoconstriction to U46619 in sham-control, nano-TiO₂ nonaugmented, and augmented nano-TiO₂ vessels with or without SQ29,548. When comparing nano-TiO₂ exposed nonaugmented vessels to the augmented nano-TiO₂ vessels, there was a significant increase in vasoconstrictive capabilities of the augmented vessels across the higher doses (Figure 4C). The augmented nano-TiO₂ vessels also exhibited increased vasoconstriction compared with nano-TiO₂ vessels with SQ29,548 amongst the higher doses in the dose-response curve (Figure 4C). At the maximal dose (Figure 4D), there was a significant increase in vasoconstriction for augmented exposed vessels compared with nano-TiO₂ nonaugmented arterioles (−42.0% ± 7.0% vs −8.6% ± 0.9%). When compared with nano-TiO₂ nonaugmented with SQ29,548, there was a significant increase in vasoconstriction for the augmented vessels (−4.2% ± 0.9% vs −42.0% ± 7.0%) and for the augmented vessels with SQ29,548 (−4.2% ± 0.9% vs −32.6% ± 7.5%).

Figure 4.

TXA₂ vascular reactivity of uterine radial arterioles in the presence of antagonists. Uterine radial arterioles were exposed to a single dose of the thromboxane receptor antagonist, SQ29,548, and then a dose-response curve to U46619, a stable TXA₂ mimetic, was performed. A, Sham-control (n = 12) versus nano-TiO₂ (n = 11) vascular response following antagonist addition and increasing doses of U46619. B, Maximum response to U46619, in the presence of the receptor antagonist, across the dose-response curve. C, Nonaugmented nano-TiO₂ (n = 11) versus augmented nano-TiO₂ (n = 8) vascular response following antagonist addition and increasing doses of U46619. D, Maximum response to U46619 in the presence of the antagonist across the dose-response curve. *p ≤ .05 versus sham-control. ^p ≤ .05 versus nano-TiO₂. ^Φp ≤ .05 versus nano-TiO₂ with SQ29,548. Abbreviations: nano-TiO₂, nano-titanium dioxide; TXA₂, thromboxane A₂.

A significant decrease in uterine radial vasodilation was observed in sham-control versus all nano-TiO₂ vessels (42.0% ± 7.3% vs 15.9% ± 3.5%), as well as in sham-control vessels compared with all nano-TiO₂ with R01138452-treated arterioles (42.0% ± 7.3% vs. 7.3% ± 7.7%; Figure 5). Figure 5B highlights differences at the dose wherein sham-control vessels were most vasodilatory. When comparing augmented nano-TiO₂ exposed vessels to sham-control arterioles, there was a significant increase in vascular vasodilation for the augmented nano-TiO₂ vessels (Figure 5C). At the dose that elicited a maximal response in sham-control vessels, there was a significant increase in the vasodilation of augmented nano-TiO₂ vessels compared with nano-TiO₂ nonaugmented vessels (42.1% ± 4.7% vs 15.9% ± 3.3%). Figure 5D also highlights the significant increase in augmented radial arteriole vasodilation compared with nano-TiO₂ nonaugmented vessels treated with R01138452 (42.1% ± 4.7% vs 7.3% ± 7.7%) and when compared with the nano-TiO₂ augmented vessels treated with R01138452 (42.1% ± 4.7% vs 15.3% ± 4.4%). Overall, these data highlight how maternal nano-TiO₂ inhalation exposure during gestation results in uterine radial arteriolar dysfunction in response to TXA₂ and PGI₂.

Figure 5.

PGI₂ vascular reactivity of uterine radial arterioles in the presence of antagonists. Prostacyclin receptor antagonist, R01138452, was added to the bath as a single dose, and then a dose-response curve of carbaprostacyclin, a stable PGI₂ agonist, followed. A, Sham-control (n = 12) versus nonaugmented nano-TiO₂ (n = 8) vascular response following antagonist addition and then increasing doses of carbaprostacyclin. B, Maximum response to pretreated vessels with the prostacyclin receptor antagonist is represented in the bar graph. C, Nonaugmented nano-TiO₂ (n = 8) versus augmented nano-TiO₂ (n = 8) vascular response after R01138452 pretreatment to dose-response curve. D, Maximum response to carbaprostacyclin in the presence of the antagonist across the dose-response curve. *p ≤ .05 versus sham-control. ^p ≤ .05 versus nano-TiO₂. ^Φp ≤ .05 vs nano-TiO₂ with R01138,452. ^†p ≤ .05 versus augmented nano-TiO₂ with R01138. Abbreviations: PGI₂, prostacyclin I₂; nano-TiO₂, nano-titanium dioxide.

Serum 6-keto-PGF_1α and TXB₂

ELISA assays were utilized to measure 6-keto-PGF_1α, and TXB₂, the stable metabolites of PGI₂ and TXA₂, respectively. Circulating serum concentrations of 6-keto-PGF_1α were not different between sham-control and nano-TiO₂ exposed groups (597.3 ± 84.4 vs 667.6 ± 45.6 pg/ml; Figure 6A). Serum concentrations of TXB₂ were significantly decreased in nano-TiO₂ exposed dams (1483 ± 63.8 vs 1978 ± 313.2 pg/ml; Figure 6B). Therefore, maternal nano-TiO₂ inhalation exposure appears to influences the gestational endocrine environment by decreasing circulating levels of TXB₂.

Figure 6.

Circulating 6-keto-PGF_1α and TXB₂ serum concentration. Maternal sham-control (N = 9) versus nano-TiO₂ (N = 9) circulating levels of 6-keto-PGF_1α (A) and TXB₂ (B) in pg/ml. *p ≤ .05 versus sham-control. Abbreviations: 6-keto-PGF_1α, 6-keto-prostaglandin F_1α; nano-TiO₂, nano-titanium dioxide; TXB₂, thromboxane B₂.

Cytosolic Phospholipase A₂ Enzymatic Activity

Cytosolic PLA₂ (cPLA₂) was measured as an indicator of AA production and COX activity. Uterine artery levels of cPLA₂ were not different between sham-control and nano-TiO₂ exposed groups (13.4 ± 8.7 vs 18.6 ± 12.6 U/mg; Figure 7A). Implantation sites were also assessed, with cPLA₂ activity (Figure 7B) being significantly decreased in nano-TiO₂ exposed dams (54.3 ± 2.7) when compared with sham-controls (59.6 ± 3.2). The cPLA₂ activity changes seen in nano-TiO₂ exposed dams demonstrate that maternal inhalation exposure during gestation exerts an effect on AA production specifically within the uterus at the implantation site of the embryo.

Figure 7.

Cytosolic PLA₂ enzymatic activity. Cytosolic PLA₂ enzymatic activity (N = 5–8) in (A) uterine artery and (B) implantation sites in U/mg. *p ≤ .05 versus sham-control. Abbreviations: PLA₂, phospholipase A₂.

COX Enzymatic Activity

COX-2 activity within the lung was not significantly different between groups, but COX-2 activity tended (p = 0.067) to be lower in nano-TiO₂ exposed lung tissue when compared with sham-control (2.8 ± 3.7 vs 12.3 ± 5.4 µU/mg; Figure 8A). Uterine artery COX-2 activity did not differ significantly between sham-control or nano-TiO₂ exposed dams (49.1 ± 48.6 vs 19.8 ± 27.4 µU/mg; Figure 8B). COX-2 activity at implantation sites was not significantly different between sham-control and nano-TiO₂ exposed groups (2.3 ± 1.7 vs 1.8 ± 0.3 µU/mg; Figure 8C). COX-2 activity was not significantly different between groups, which may be due to the lag time between the last exposure and tissue collection.

Figure 8.

COX enzymatic activity. COX-2 enzymatic activity (N = 4–6) in (A) lung, (B) uterine artery, and (C) implantation sites in µU/mg. Abbreviations: COX, cyclooxygenase.

mRNA Expression of COX-1, PGIS, TXAS, PGR, and TXR

Expression of mRNA for COX-1, prostacyclin synthase, prostacyclin receptor, thromboxane synthase, and thromboxane receptor was determined in lung tissue from dams due to the direct exposure of that tissue to nano-TiO₂ (Table 5). There was a significant decrease in mRNA fold-expression for prostacyclin receptor in nano-TiO₂ exposed dams compared with sham-controls (1.0 ± 0.4 vs 0.3 ± 0.1). In contrast, prostacyclin synthase was significantly increased in lung tissue of nano-TiO₂ exposed dams compared with sham-control dams (1.0 ± 0.9 vs 2.5 ± 0.3). Expression of mRNA for COX-1, thromboxane synthase, or thromboxane receptor was not significant between groups. Assessment of implantation sites failed to detect differences between groups for thromboxane synthase or thromboxane receptor. Although there was not a significant difference between groups for COX-1 (1.0 ± 0.5 vs 5.6 ± 0.9), prostacyclin synthase (1.0 ± 0.2 vs 2.2 ± 0.4), and prostacyclin receptor (1.0 ± 0.2 vs 1.8 ± 0.9) all 3 had elevated gene expression in the nano-TiO₂ exposed tissue compared with sham-control. Assessment of placentas (sex not accounted for) was not significantly different between groups for COX-1, prostacyclin synthase, or thromboxane synthase. Although not significant, there was decreased prostacyclin receptor (0.46 ± 0.44 vs 1.00 ± 0.56) and increased thromboxane receptor (1.41 ± 0.38 vs 1.00 ± 0.37) gene expression in the nano-TiO₂ exposed tissue compared with sham-control. Maternal ovaries were also examined due to the importance of the ovaries for production of prostaglandins and their critical role in the rat’s ability to maintain a pregnancy. There were no significant differences noted in the ovary for any of the genes examined (data not shown). Examination of uterine arteries gene markers likewise failed to produce significant differences between nano-TiO₂ exposed and sham-control dams (data not shown). Maternal nano-TiO₂ inhalation exposure affected gene expression in lung tissue, which is likely due to the direct insult delivered by nanoparticle deposition.

Table 5.

RT-PCR results for lung implantation site, and placental tissue

	N	Lung		Implantation Sites		Placenta
		Sham-Control	Nano-TiO2	Sham-Control	Nano-TiO2	Sham-Control	Nano-TiO2
Cyclooxygenase-1	6	1.00 ± 0.61	0.73 ± 0.39	1.00 ± 0.52	5.57 ± 0.90	1.00 ± 0.42	1.09 ± 0.30
Prostacyclin synthase	6	1.00 ± 0.85	2.50 ± 0.28*	1.00 ± 0.22	2.24 ± 0.37	1.00 ± 0.14	1.32 ± 0.27
Prostacyclin receptor	6	1.00 ± 0.39	0.31 ± 0.13*	1.00 ± 0.23	1.78 ± 0.89	1.00 ± 0.56	0.46 ± 0.44
Thromboxane synthase	6	1.00 ± 0.28	0.48 ± 0.85	1.00 ± 0.51	0.56 ± 0.71	1.00 ± 0.14	0.78 ± 0.26
Thromboxane receptor	6	1.00 ± 0.49	0.98 ± 0.39	1.00 ± 0.08	0.84 ± 0.72	1.00 ± 0.37	1.41 ± 0.38

N represents the number dams used for tissue processing. Data are fold-change and represent mean ± SEM. Bold font is used to draw attention to the fact that the data points are significant. 

Abbreviations: SEM, standard error of the mean; nano-TiO₂, nano-titanium dioxide.

^* p ≤ .05 versus sham-control.

DISCUSSION

The primary aim of this study was to determine the effects of maternal nano-TiO₂ inhalation during gestation on uterine microvascular reactivity in response to the COX metabolites, PGI₂, and TXA₂. The exposure paradigm used in this study with nano-TiO₂ inhalation during gestation models and exposure pregnant women in an occupational setting would experience. This is mean to encompass the time periods in which these women may or may not be aware of their pregnancy and also continue to work until late gestation. We demonstrated that nano-TiO₂ inhalation exposure during gestation impaired endothelium-dependent vasodilation, augmented TXA₂-induced vasoconstriction in radial uterine arterioles, decreased fetal female pup mass, decreased resorption sites, increased lung prostacyclin synthase mRNA expression, decreased circulating TXB₂ concentration, and decreased implantation site cPLA₂. Decreased vasodilation as a result of inhalation exposure was an anticipated result, as it has been previously observed (Knuckles et al., 2012; LeBlanc et al., 2010). In a normal pregnancy, vascular tone decreases due to outward vascular remodeling, thus decreasing vascular resistance and increasing blood flow to the gravid uterus (Boeldt and Bird, 2017; Osol and Mandala, 2009). Vascular tone, which impacts vascular resistance and thus blood flow, is the product of a variety of factors including: myogenic activity, vasoconstrictor and vasodilatory influences, and flow or shear-stress (Hu and Zhang, 2021). It should be noted that arteriolar resistance is the result of two forms of tension regulatory elements: passive and active (Bohlen and Lash, 1994; Lash et al., 1991). Active tension is the dominant regulator of arteriolar diameter via physiologic signaling, such as vasoactive compounds, blood pressure, neurogenic activity, and metabolic state (Bohlen and Lash, 1994; Lash et al., 1991). Passive tension contributes broadly to arteriolar diameter and is associated with structural and anatomical characteristics of the arteriolar wall (Bohlen and Lash, 1994). Microcirculation heterogeneity is based on endothelial factors, smooth muscle cells, and the local microcirculation environment (Boegehold, 1998). Disruptions in vasoactive compound concentrations/activity (active tension), vascular remodeling (passive tension), and vascular resistance may ultimately disrupt blood flow and adversely impact fetal growth and development. Studies on rat uterine artery blood flow restriction found that decreased uterine blood flow via ligation, or due to ozone exposure, resulted in reduced fetal body weight (Miller et al., 2017; Wlodek et al., 2005). In this study, decreased vasodilation and increased vasoconstriction may have contributed to poor fetal health consequences as evidenced by decreased fetal female pup weights (Table 3).

The lung tissue collected from nano-TiO₂ exposed dams exhibited increased gene expression of PGIS and decreased prostacyclin receptor (Table 5). Additionally, serum from dams contained elevated circulating levels of PGI₂ (Figure 6). The increase in mRNA expression for PGIS in the lungs and elevated serum levels of PGI₂ seen in our study is likely in response to inflammatory signals arising from the lungs, as it is the initial site of deposition. We have previously shown that pulmonary nano-TiO₂ exposure is associated with increased interleukin (IL)-1β, 4, 5, 13, and tumor necrosis factor (TNF)-α in the plasma and bronchoalveolar lavage fluid, indicative of an innate inflammatory response (Abukabda et al., 2018). IL-1β has been reported to increase PGI₂ production in hypoxic environments (Camacho et al., 2011). Due to the interdependency of the pulmonary and cardiovascular systems, inflammatory responses that begin in the lung may potentially enter the circulation and result in systemic effects. This is part of the hypothesis by which inhaled air pollution impart negative cardiovascular effects (Robertson and Miller, 2018). Considered together during pregnancy, inflammatory signals arising from the lungs could ultimately have detrimental effects on uteroplacental blood flow.

Traditionally, a lung insult and the subsequent inflammatory response results in a swift increase of TXA₂ (Ishitsuka et al., 2004; Park and Christman, 2006; Wang et al., 2020). This is followed by increased PGI₂ acting as an antiinflammatory agent to increase blood flow to the inflammatory site (Davies et al., 1984; Park and Christman, 2006). Given previous work, we anticipated that TXB₂ would be increased after exposure, but this was not the case. However, thrombin and cytokines also stimulate production of PGI₂, likely in an effort to balance any vasoconstrictive responses to TXA₂ (Majed and Khalil, 2012). The increased PGI₂, along with elevated nitric oxide (NO) activity, is known to oppose TXA₂-induced vasoconstriction by desensitization of the thromboxane receptor (Goulopoulou, 2017). The increased TXA₂ would result in more robust vasoconstriction at the uteroplacental level, thus decreasing blood flow to the placenta and hindering the nutrient-waste exchange.

PGI₂ is mostly produced by endothelial cells but has an important role in the feedforward mechanism of parturition within the myometrium. TXA₂ is produced by endothelial cells, activated platelets, and macrophages. Circulating blood levels of rat 6-keto-PGF_1α is reported to be approximately 2.78 ± 0.05 ng/ml and rat TXB₂ is reported to be approximately 9.34 ± 1.19 ng/ml (Verlohren et al., 2008). Maternal nano-TiO₂ inhalation exposure decreased PGI₂-induced vasodilation. This is consistent with observations in preeclamptic patients (Boeldt and Bird, 2017; Hu and Zhang, 2021). The decrease we observed in endothelium-dependent vasodilation, similar to what we have previously reported (Abukabda et al., 2018; LeBlanc et al., 2010), is potentially the result of decreased NO bioavailability (Nurkiewicz et al., 2009), ultimately impairing vasodilatory capacity. It should be noted that the AA pathway is complex and requires multiple cells and receptors to complete its steps, and as such, is highly dependent on the microenvironment around these cells (Boegehold, 1998). The heterogeneity of the microcirculation is important to this point, in that the vessels being studied can produce different responses based on their regional location (tissue), segmental location, and circulating factors. We speculate this heterogeneity contributes to disparities between this study and previous studies. In isolated coronary arterioles, we observed that nano-TiO₂ inhalation had no significant effect on vasoconstriction to the TXA₂ mimetic, U46619 (LeBlanc et al., 2010). This is different from our study herein in which uterine radial arterioles were studied and had increased vasoconstrictor responses to TXA₂. Additionally, we found a subpopulation of vessels that were hyper-responsive to TXA₂ and/or PGI₂ (Figure 3). The increased TXA₂-induced vasoconstriction of radial uterine arterioles (Figure 3C) is seen in other gestational disease states such as preeclampsia, or intrauterine growth restriction (IUGR). In preeclamptic women, the endothelium reduces vasodilator production and fails to become insensitive to vasoconstrictors (Boeldt and Bird, 2017). Thus, these women present with increased uterine vascular resistance, when there would normally be low resistance to encourage nutrient-waste exchange in the placenta. Preeclamptic women also exhibit increased circulating levels of TXB₂, hypertension, proteinuria, and decreased fetal weight at term (Goulopoulou, 2017; Gunnarsdottir et al., 2018). Similarly, our model does show increased vasoconstriction (Figs. 3A and 4) and decreased fetal weight at term (Table 3). This may be evidence that uterine radial arterioles are compensating for decreased circulating TXA₂ by being hyper-responsive to TXA₂ stimulation. Increased vasoconstriction and decreased vasodilation, as seen in this study, have the potential to increase placental vascular resistance and ultimately lead to decreased blood flow to the placenta.

We observed a subpopulation of uterine radial arterioles that exhibited augmented vasodilation to PGI₂ (Figs. 3A and 5). Vessels hyper-responsive to PGI₂, unlike those that were hyper-responsive to TXA₂, presented minimal dilatory responses to the prostacyclin receptor antagonist (Figure 5D). The hyper-responsiveness of the radial uterine arterioles was surprising, mainly because we have previously demonstrated decreased vasodilatory responses after nano-TiO₂ inhalation exposures (Abukabda et al., 2018; LeBlanc et al., 2010). It should be noted that studies are lacking on these augmented responsive vessels to vasodilatory agents such as PGI₂. A similar augmented vasodilatory phenomenon can be found in some forms of IUGR and may be compensatory mechanisms aimed at maintaining fetal blood flow to near normal levels (Cohen et al., 2015; Fleiss et al., 2019). In a rat model where IUGR is induced by a low-Na⁺ diet to reduce maternal blood volume expansion and uteroplacental perfusion, it was found that radial uterine artery relaxation to carbachol, a parasympathomimetic that stimulates muscarinic and nicotinic receptors, was increased in IUGR rats (Bigonnesse et al., 2018). This study also found that relaxation in IUGR radial arterioles was increased in response to NO donors (Bigonnesse et al., 2018). These studies indicate that when vasoconstriction is augmented, such as in our study, the body attempts to maintain a homeostatic balance by increasing vasodilatory responses, suggesting our results could reflect a compensatory mechanism. In the absence of this compensatory mechanism, fetal outcomes may be more pronounced due to perturbed blood flow and nutrient-waste exchange between, the uteroplacental unit.

Although sex ratio was not significantly different, there appears to be a relationship between wet pup weight and sex in exposure groups that should be further studied in the future. Neither fetal sex nor location within the uterine horn was accounted for during tissue collection in this study, which may explain at least in part, the large variability seen in gene expression within the placenta (Table 5), implantation site tissue (Table 5), and enzymatic activity (Figs. 7B and 8C). It has been shown that fetal and placental size is sexually dimorphic, with male fetuses and their placentas being larger throughout gestation (Gonzalez et al., 2018). Maternal vascular function and effects of toxic exposure during gestation are also influenced by fetal sex (Gabory et al., 2013). In normal pregnancies with a male fetus, microvascular vasodilation induced by corticotrophin is enhanced compared with pregnancies with a female fetus (Clifton et al., 2012). Based on this information, it is plausible that some of the variability seen in AA metabolite gene expression within implantation sites and placental tissue may be due to sexually dimorphic responses after maternal nano-TiO₂ inhalation exposure during gestation. This would require further study to determine if toxicant exposure is causing sexually dimorphic changes to AA metabolites.

Maternal nano-TiO₂ inhalation exposure during gestation results in increased TXA₂-induced arteriolar vasoconstriction, decreased PGI₂-induced vasodilation, and decreased female fetal weight. Our exposure paradigm also provides evidence that maternal nano-TiO₂ inhalation exposure during gestation augments the microvascular response to thromboxane. Collectively, we speculate that alterations in vascular tone and reactivity compromise blood flow to the placenta thus decreasing placental nutrient and waste exchange. This environment is largely created by a balance of vasoactive elements, suggesting the modifications reported herein may be physiological adaptations in response to the insult of nano-TiO₂ inhalation during gestation. Future studies should investigate uterine inflammatory changes and modifications of placental blood flow in response to thromboxane and prostacyclin after maternal inhalation exposure during gestation. This would provide extended evidence and insight into placental hemodynamic modifications to maternal inhalation exposures during pregnancy.

AUTHOR CONTRIBUTIONS

Study design: J.A.G./E.C.B./T.R.N. Data Collection: J.A.G./K.L.G./E.C.B./E.D./K.S. Data analysis and interpretation: J.A.G./E.C.B./S.H./T.R.N. Animal Exposures: J.A.G./K.L.G./E.C.B./K.J.E./T.P.B./W.T.G./K.W. Article draft: J.A.G. Critical revisions and final decision to submit: all authors.