Renal redox response to normal pregnancy in the rat

Abstract

Normal pregnancy involves increased renal sodium reabsorption, metabolism, and oxygen consumption, which can cause increased oxidative stress (OS). OS can decrease nitric oxide (NO) bioavailability and cause pregnancy complications. In this study we examined the NO synthases (NOS) and redox state in the kidney cortex and aorta in early (E), mid (M), and late (L) pregnant (P) (days 3, 12, 20) and 2–4 days postpartum (PP) rats compared with virgin rats (V). Protein abundance of endothelial NOS (eNOS) was unchanged and neuronal NOS (nNOS)α fell at LP in the kidney cortex. Kidney cortex nNOSβ was elevated at MP, LP, and PP. No changes in aortic NOS isoforms were observed. Kidney cortex nitrotyrosine (NT) abundance decreased in EP, MP, and PP, whereas aortic NT increased in EP, MP, and PP. The NADPH oxidase subunit p22phox decreased in the kidney cortex at EP while aortic p22phox increased in EP and LP. No changes in kidney cortex NADPH-dependent superoxide production or hydrogen peroxide levels were noted. Kidney cortex cytosolic (CuZn) superoxide dismutase (SOD) was unchanged, while mitochondrial SOD decreased at EP and extracellular SOD decreased at MP and LP in the kidney cortex. Despite falls in abundance of kidney cortex SODs, total antioxidant capacity (TAC) was elevated in EP, MP, and PP in the kidney cortex. Aortic CuZn SOD deceased at PP, while the other aortic SODs and aortic TAC did not change. Data from this study suggest that the kidney cortex is protected from OS during normal rat pregnancy via an increase in antioxidant activity.

during pregnancy there is a widespread increase in metabolic activity that may lead to maternal systemic inflammation and oxidative stress even in normal pregnancy (6, 7, 14, 21, 22, 28). The kidney undergoes profound adaptations during pregnancy with a marked vasodilation and increase in glomerular filtration rate, as well as increased renal sodium reabsorption leading to positive sodium balance and maternal extracellular fluid volume expansion (3, 10, 12). This increase in sodium reabsorption requires an increase in metabolism that could lead to increased oxidative stress levels in the kidney. When reactive oxygen species are increased, nitric oxide (NO) bioavailability is reduced, and during pregnancy, this is associated with complications such as preeclampsia (7). In the case of the kidney, increased renal cortical NO production is required for the normal gestational renal vasodilation (2, 3, 9, 11, 13).

There is little information on the impact of pregnancy on the oxidant status of the normal kidney. In this study we examined the antioxidant capacity as well as markers of oxidative stress in the kidney cortex of virgin rats, during normal early, mid, and late pregnancy and 2–4 days postpartum in the Sprague-Dawley (SD) rat. We also repeated our earlier measurements of kidney cortex NO synthase (NOS) abundance (25), and for comparison, some of these enzymes and markers were also measured in the aorta.

METHODS

All rat experiments were approved by the committee on Animal Research of the University of San Francisco, California. Twenty-four female SD rats from Harlan (Indianapolis, IN) were mated with fertile male rats to generate pregnancy. Day 1 of pregnancy was determined by the presence of sperm in vaginal smears. Rats (n = 6 per group) were euthanized at early pregnancy (day 3), midpregnancy (day 12), late pregnancy (day 20), and postpartum. Pregnant and postpartum rats were age matched with virgin control female rats (n = 6). Kidney cortex and thoracic aorta/aortic arch were collected at the time of euthansia, flash frozen in liquid nitrogen, and stored at −80°C until further analysis.

Western blot.

Kidney cortex and aortic tissue abundance of proteins were detected using Western blotting, as previously described (23, 31). Tissue samples (200 μg of kidney cortex and aorta) were loaded on 7.5% or 12% polyacrylamide gels and separated by electrophoresis. Proteins were transferred onto a nitrocellulose membrane and stained with Ponceau red (Sigma-Aldrich, St. Louis, MO) to verify protein transfer and equal loading among samples. Membranes were then blocked, washed, and incubated overnight at 4°C with the primary antibody of interest. Blots were probed with the following primary antibodies: mouse anti-endothelial NOS (eNOS, 1:250 dilution) from BD transduction (San Jose, CA), mouse anti-neuronal NOS (nNOSα, 1:50 dilution) from Santa Cruz Biotechnology (Santa Cruz, CA), rabbit anti-nNOSβ (1:500 dilution) from Thermo Scientific Pierce (Rockford, IL), rabbit anti-extracellular SOD (ecSOD) (1:250 dilution) from Abcam (Cambridge, MA), rabbit anti-MnSOD (1:2,000 dilution) from Stressgen Bioreagents (Ann Arbor, MI), rabbit anti-Cu/ZnSOD (1:2,000 dilution) from Stressgen Bioreagents (Ann Arbor, MI), goat anti-p22phox (1:50 dilution) from Santa Cruz Biotechnology (Santa Cruz, CA), and mouse anti-nitrotyrosine (1:500 dilution) from EMD Millipore (Billerica, MA). Membranes were then incubated with the appropriate secondary antibody, goat anti-rabbit (1:3,000) from Bio-Rad (Hercules, CA), goat anti-mouse (1:2,000 dilution) from Bio-Rad, or donkey anti-goat (1:2,000 dilution) from Santa Cruz Biotechnology for 1 h at room temperature and then developed with enhanced chemiluminescent reagents (Thermo Scientific Pierce). For each protein probed, we ran two large gels with 25 lanes that consisted of a molecular ladder, an internal positive control (which is a tissue sample rich in abundance of our protein of interest), and tissue samples from virgin rats, early, mid, and late stages of pregnancy, along with postpartum rats. Samples were obtained from six rats at each stage of pregnancy and three samples were run on each gel. The bands were quantified by densitometry using the VersaDoc Imaging System and Software (Bio-Rad). Densitometry was normalized to Ponceau staining (total protein loaded) and an internal positive control, which allows for densitometric comparisons between samples among different membranes (23). Mean values for the virgin control group was set at 100% for comparisons between groups.

NADPH-dependent superoxide production.

NADPH-dependent superoxide production was measured by lucigenin (5 μmol/l, Sigma) chemiluminescence in the presence of NADPH (100 μmol/l, Sigma) as previously described (24). Briefly, kidney cortex samples were homogenized, and protein concentrations were measured by the Bradford assay (Bio-Rad). Thirty-five micrograms of kidney cortex homogenate diluted in physiological saline solution (PSS) were mixed with additional PSS or Tempol (10 mmol/l final concentration). Samples were placed in a 96-well plate and were incubated for 30 min at 37°C. Lucigenin and NADPH were added and after a 15-min dark period, plates were counted on an Orion microplate luminometer. NADPH-stimulated superoxide activity was expressed as relative luminescent unit (RLU) per microgram protein minus RLU per microgram protein in the presence of Tempol (i.e., RLU/μg protein − RLU/μg protein in Tempol).

Hydrogen peroxide.

Kidney cortex was homogenized in lysis buffer and then hydrogen peroxide levels (pmol/mg) were determined using the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (A22188, Invitrogen, Carlsbad, CA).

Total antioxidant capacity.

Total antioxidant capacity (TAC) was measured using the Antioxidant Assay Kit (709001, Cayman Chemical, Ann Harbor, MI), according to the manufacturer's instruction. This kit measures the ability of all aqueous and lipid-soluble antioxidants to inhibit the oxidation of 2,2′-azino-di-(3-ethylbenzthiazoline sulfonate) (ABTS) to ABTS⁺ by metmyoglobin, by reading absorbance at 750 nm. One microgram of protein from the kidney cortex and 10 μg of protein from the aorta were pipetted into 96-well plates. Sample antioxidant capacity is compared with Trolox (a water-soluble tocopherol analogue, 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid) standards at 0, 0.045, 0.09, 0.135, 0.18, 0.225, and 0.330 mM Trolox. The total antioxidant capacity is quantified as millimolar Trolox equivalents (mM Trolox/mg) at 750 nm wavelength.

Statistical analysis.

All data are expressed as means ± SE. Differences between virgin, early, mid, and late pregnant were analyzed using the one-way analysis of variance (ANOVA) multiple comparison test and the Bonferroni post hoc test. If analysis between groups failed the normality test, the Kruskal-Wallis one-way ANOVA multiple comparisons test was used with the Dunn post hoc test. In addition, the unpaired t-test was used to compare the virgin rat with specific stages of pregnancy and postpartum. Data were considered statistically different at P values <0.05.

RESULTS

NOS protein abundance.

Kidney cortex and aortic protein abundance of eNOS did not change during the time course of pregnancy or postpartum in the SD rats (Fig. 1, A and D). The nNOS splice variant nNOSα decreased in kidney cortex during late pregnancy compared with virgin rats (Fig. 1B) but showed no changes in the aorta (Fig. 1E). The abundance of nNOSβ increased during midpregnancy, late pregnancy, and postpartum compared with virgin rats in the kidney cortex (Fig. 1C) but remained unchanged in aorta (Fig. 1F).

Fig. 1.

Western blot analysis of kidney cortex endothelial nitric oxide synthase (eNOS) (A), neuronal NOS (nNOSα, B), and nNOSβ (C) protein abundance along with aortic eNOS (D), nNOSα (E), and nNOSβ (F) protein abundance during early (EP, day 3), mid (MP, day 12), and late (LP, day 20) pregnant Sprague-Dawley (SD) rats along with age-matched virgin (V) and postpartum (PP) rats. Representative blots with the molecular weight, positive control (+), and arrow pointing to the band of interest are provided for each protein of interest. Densitometry analysis is provided for each protein and was quantified as integrated optical density (IOD) normalized to Ponceau (PON) staining and the positive control (POS). Note that kidney cortex blots were cropped to exclude other stages of pregnancy not used in this study. All Western blot images shown here were done at the same time and are from the same membrane. *P < 0.05 vs. V and †P < 0.05 vs. EP.

Oxidative stress markers.

Abundance of p22phox and nitrotyrosine decreased in kidney cortex during early pregnancy compared with virgin rats (Fig. 2, A and B). Nitrotyrosine abundance also decreased in kidney cortex during late pregnancy and postpartum compared with virgin rats (Fig. 2B). In contrast, aortic p22phox abundance increased during early and late pregnancy compared with virgin rats (Fig. 2C), and there was a trend for an increase in the abundance of aortic p22phox during midpregnancy (P = 0.05). Nitrotyrosine abundance in aorta increased in early and midpregnancy and postpartum compared with virgin rats (Fig. 2D). Kidney cortex NADPH-dependent superoxide production and hydrogen peroxide content were unchanged during pregnancy and postpartum versus virgin rats (Fig. 3).

Fig. 2.

Western blot analysis of kidney cortex p22phox (A) and nitrotyrosine protein abundance (B) along with aortic p22phox (C) and nitrotyrosine protein abundance (D) during EP (day 3), MP (day 12), and LP (day 20) pregnant SD rats along with age-matched V and PP rats. Representative blots with the molecular weight, positive control (+), and arrow pointing to the band of interest are provided for each protein of interest. Densitometry analysis is provided for each protein and was quantified as IOD normalized to PON staining and the POS. Note that kidney cortex blots were cropped to exclude other stages of pregnancy not used in this study. All Western blot images shown here were done at the same time and are from the same membrane. *P < 0.05 vs. V, †P < 0.05 vs. EP, and ‡P < 0.05 vs. LP.

Fig. 3.

Kidney cortex NADPH-dependent superoxide production (A) and hydrogen peroxide (B) content during EP (day 3), MP (day 12), and LP (day 20) pregnant SD rats along with age-matched V and PP rats.

Antioxidants.

Protein abundance of ecSOD decreased in the kidney cortex in mid and late pregnancy versus virgin rats (Fig. 4A). Intracellular mitochondrial SOD (MnSOD) decreased in early pregnancy versus virgin rats, whereas intracellular cytosolic SOD (Cu/Zn SOD) protein abundance was similar in all groups in the kidney cortex (Fig. 4, B and C). In the aorta, ecSOD, MnSOD, and Cu/Zn SOD protein abundance was unchanged during pregnancy, although there was a decrease of Cu/Zn SOD in postpartum versus virgin rats (Fig. 4, E–G). Despite decreases in ecSOD and MnSOD in the kidney cortex during pregnancy, there was an increase in total antioxidant capacity in early and midpregnant rats versus virgin rats (Fig. 4D). The trend for increased total antioxidant capacity in postpartum versus virgin rats in the kidney cortex was close to statistical significance (P = 0.052) (Fig. 4D). No changes in aortic total antioxidant capacity were observed during pregnant or postpartum rats (Fig. 4H).

Fig. 4.

Western blot analysis of kidney cortex extracellular (ec) SOD (A), manganese (Mn) SOD (B), and copper/zinc (Cu/Zn) SOD protein abundance (C); and kidney cortex total antioxidant capacity (D) along with aortic ecSOD (E), manganese (Mn) SOD (F), and copper/zinc (Cu/Zn) SOD protein abundance (G); and aortic total antioxidant capacity (H) during EP (day 3), MP (day 12), and LP (day 20) pregnant SD rats along with age-matched V and PP rats. Representative blots with the molecular weight, positive control (+), and arrow pointing to the band of interest are provided for each protein of interest. Densitometry analysis is provided for each protein and was quantified as IOD normalized to PON staining and the POS. Note that kidney cortex blots were cropped to exclude other stages of pregnancy not used in this study. All Western blot images shown here were done at the same time and are from the same membrane. *P < 0.05 vs. V.

DISCUSSION

The main novel findings of this study are that despite increases in renal metabolic activity (due to increased sodium reabsorption) during normal pregnancy, there is no evidence of increased oxidative stress in the kidney cortex. The antioxidant capacity in the kidney cortex increases early during pregnancy and as a result alters the balance between pro- and antioxidant influences to promote an antioxidative state. The ability of the kidneys to lower oxidative stress during pregnancy may help to enhance NO bioavailability and thus facilitate maternal gestational renal vasodilation. In contrast, the thoracic aorta displays net evidence of increased oxidative stress.

Oxidative stress has been implicated in many pregnancy-associated pathologies such as preeclampsia (7), but the increased oxygen demand and cardiac output of normal pregnancy may also put the mother at risk of systemic and organ oxidative stress damage (3, 8). Indeed, normal pregnant women exhibit evidence of systemic inflammation and oxidative stress (7, 28) with a progressive increase in serum oxidized low-density lipoproteins (oxLDL) and a decrease in serum total antioxidant capacity from the first to the third trimester of pregnancy (6, 14, 28). Furthermore, during the third trimester of human pregnancy, increases in circulating leukocytes are associated with inflammation, and the generation of reactive oxygen species along with increases in plasma isoprostane levels and urinary 8-hydroxydeoxyguanosine excretion, a marker of DNA damage caused by oxidative stress (14, 21, 22). Our present studies in SD rats expand these observations and indicate that in the thoracic aorta, normal pregnancy is associated with increased p22phox (NADPH oxidase subunit) and increased nitrotyrosine (marker of oxidative stress) protein abundance. In the normal late pregnant SD rat, Li et al. (16) reported an exacerbated infarct size versus nonpregnant rats in response to acute cardiac ischemia-reperfusion (I/R) injury (16). In addition, superoxide production was much greater in the heart of the late pregnant rats versus the nonpregnant rats after the I/R injury, suggesting a preexisting elevation in oxidative stress and/or loss of antioxidant defenses in the heart in late pregnancy (16).

The primary focus of the present study was to examine the effects of normal pregnancy on the redox status of the kidney. The kidneys play a vital role in generating and maintaining plasma volume expansion in both human and rat pregnancy (3). The increase in plasma volume is mediated by renal sodium and water retention during pregnancy (3). A suboptimal volume expansion during pregnancy is associated with pregnancy complications and small for gestational-age babies (3). In addition to plasma volume expansion, a healthy pregnancy requires a robust renal vasodilation leading to increases in renal blood flow and glomerular filtration rate (2, 3, 8, 10, 12, 13).

An increase in renal NO production plays a critical role in mediating these renal hemodynamic changes during pregnancy (2, 11, 13). We recently reported that both message expression and protein abundance of a nNOS splice variant, nNOSβ, increases in the kidney cortex during pregnancy in the SD rat, whereas the “normal” nNOS isoform found in kidneys, nNOSα, and the eNOS tend to decrease (25). Our present observations support these earlier findings with an increase in nNOSβ abundance from midterm, no change in eNOS, and a decrease in nNOSα at late pregnancy in the kidney cortex. There are numerous splice variants of nNOS that are generated by different nNOS mRNA transcripts. In humans, these different splice variants are created by a complex mixture of cell-specific transcriptional regulation, alternative promoter regions of nNOS, and nucleotide inserts into the 5′ untranslated regions between exon 1 and 2 of nNOS that ultimately affect the fate of protein translation of the transcript (19, 29). Unfortunately, there is no information on the regulation of alternative splicing of nNOS in the rat, and this is an area that requires further study.

The regulation of NO production is complex, and there is considerable posttranslational control in which the redox state plays a major role in determining NO bioavailability. When levels of superoxide increase, NO is scavenged and converted to the highly reactive and toxic peroxynitrite. In addition, sustained high levels of superoxide can uncouple NOS and produce superoxide rather than NO (7, 27). Given the net renal sodium retention as well as the increase in filtration of sodium during normal pregnancy, there is a marked rise in renal sodium reabsorption. To increase sodium reabsoprtion during pregnancy, the kidney will increase its active sodium transport by increasing oxygen consumption and metabolism, which will generate increased oxidative stress (30). In the present study, however, we observed an early fall in p22phox abundance and no change in NADPH-dependent superoxide production or hydrogen peroxide production in the kidney cortex at any stage of pregnancy. In fact, the kidney cortex abundance of nitrotyrosine (a measure of peroxynitrite production) was decreased during pregnancy. On the antioxidant side, we observed unexpected decreases in kidney cortex MnSOD and ecSOD abundance, whereas Cu/Zn SOD abundance was maintained throughout pregnancy. Despite these changes in antioxidant protein abundances, there was an increased total antioxidant activity in early and midpregnancy. This may reflect the fact that the SODs are also subject to posttranslational modification so that enzyme activity and abundance do not necessarily change in parallel. Alternatively, there are several other agents that may have contributed to the increase in antioxidant activity (i.e., glutathione peroxidase, catalase, and other macro and small molecules) (18). Overall, it is clear that during normal pregnancy the kidney cortex experiences a net reduction in oxidative stress. This is even more marked in the pregnant spontaneously hypertensive rat (SHR) (15). In the nonpregnant SHR kidney there is evidence of inflammation and oxidative stress in the kidney that are dramatically reduced in late pregnancy (15). This protection may be linked to the beneficial shift in the angiotensin II receptor profile within the kidney so that the AT2 receptor predominates (15). There is also a fall in blood pressure in late gestation that may be associated with these intrarenal changes (15, 17). AT2 receptor activation facilitates renal vasodilation via an increase in NO production, which opposes the actions of the AT1 receptor (renal vasoconstriction, oxidative stress, and decreased NO bioavailability) by competing for ANG II. Given the benign effects of pregnancy on maternal blood pressure and kidney function, it is not surprising to note that multiple repetitive pregnancies have no long-term damaging effects with blood pressure, renal function, and protein excretion in the SHR (4).

In contrast to the normal rat kidney cortex, the aorta showed signs of increased oxidative stress (via increased nitrotyrosine and p22phox abundance) during pregnancy while the total antioxidant capacity was not altered. There was no change in the abundance of the aortic NOS isoforms in the present study. A previous study by Stennett et al. (26) also reported unchanged aortic eNOS abundance, although aortic phospho-eNOS abundance and NO production increased during late pregnancy. Future studies will need to determine aortic phospho-eNOS abundance, production of NO, NADPH-dependent superoxide production, and hydrogen peroxide levels during different stages of pregnancy to determine the net oxidative stress profile of the aorta during rat pregnancy.

In conclusion, there is an increase in the antioxidant capacity of the kidney cortex during normal pregnancy that protects the kidney from oxidative stress. The ability of the kidneys to maintain a normal or reduced level of oxidative stress during pregnancy is likely to facilitate renal gestational vasodilation, which requires increased NO bioavailability.

Perspectives and Significance

Pregnancy is considered a state of increased oxidative stress and inflammation due to increased metabolism and oxygen demand. During pregnancy the kidney will increase its oxygen consumption due to the increased metabolism (necessary to increase renal Na retention) that will lead to production of increased reactive oxygen species, which can compromise normal NO-mediated renal vasodilation during pregnancy. However, despite evidence of systemic oxidative stress in normal pregnancy, we report here that in the kidney cortex, nitrotyrosine levels fall and the antioxidant capacity increases. These findings indicate that the renal redox state is regulated differently from other organs (e.g., heart) and the systemic circulation. The net antioxidant response of the normal kidney to pregnancy permits increases in renal NO production that are required for the optimal gestational renal vasodilation. It should be noted also, that while excessive oxidative stress is damaging, there are situations in which reactive oxygen species perform a physiological role, such as during placentation (7). Thus moderate and targeted increases in reactive oxygen species may be required for an optimal pregnancy.

GRANTS

This work was supported by National Insititutes of Health Grants RO1DK-56843 and RO1HD-041571 (to C. Baylis). Funding for M. Cunningham is provided by T32DK-751825, C. West was funded by T32 HL-083810, and J. Sasser by was funding by American Heart Association Grant 11SDG6910000.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: M.W.C., J.M.S., and C.A.W. performed experiments; M.W.C., J.M.S., and C.A.W. analyzed data; M.W.C., J.M.S., and C.B. interpreted results of experiments; M.W.C. prepared figures; M.W.C. drafted manuscript; M.W.C., J.M.S., C.A.W., and C.B. edited and revised manuscript; M.W.C. and C.B. approved final version of manuscript; C.B. conception and design of research.