Ethics approval was granted by the Animal Ethics Committee of the Northern Sydney Local Health District (Resp/17/29), and work complied with the Australian code for the care and use of animals for scientific purposes (National Health And Medical Research Council, 2013). Animal experiments were performed at Kearns Animal Facility of Kolling Institute, Royal North Shore Hospital, Sydney, Australia. Our well-established mouse models of maternal obesity and diet-induced obesity were used in this study (Glastras et al., 2016b). Maternal obesity was induced by feeding 8 week old female C57BL/6 mice high fat diet (HFD 20 kJ/g, 43% of total energy from fat, Specialty Feeds, WA, Australia) for 6 weeks prior to mating and during the gestation and lactation periods. Control dams received standard rodent chow (11 kJ/g, Gordon’s Specialty Stockfeeds, NSW, Australia). Mating commenced at 14 weeks of age, and dams were allowed up to 72 h with males before being returned to their own cages. The presence of a vaginal plug was used to confirm pregnancy. During gestation, mice were injected subcutaneously with either normal saline (0.15 mL) or hydralazine (5 mg/kg) thrice weekly, as this dose is below the blood pressure-lowering dose of 50 mg/kg/day (Tampe et al., 2017). On postnatal day 1, the litter size was reduced to four pups, to avoid milk competition differences. On postnatal day 20, male offspring were weaned to either standard rodent chow or HFD. Only male offspring were studied as previous experimental models of CKD have generally shown that disease progression is more rapid in males than in females (Neugarten and Golestaneh, 2013), and the programming of offspring disease is more frequently observed in male offspring than in females (Cheong et al., 2016a,b). These effects are likely to be related to the differential effects of sex hormones (Whitaker et al., 2012). Offspring continued the same diet, and were weighed fortnightly until sacrifice at 32 weeks of age. There were eight offspring groups with an average of 12 animals per group. The offspring groups are abbreviated such that the first letter refers to the maternal diet (chow or HFD), and the second letter denotes the offspring diet. The groups were: control offspring (CC), lean offspring of obese mothers (HC), obese offspring of lean mothers (CH), obese offspring of obese mothers (HH), control offspring exposed to gestational hydralazine (CC + GH), lean offspring of obese mothers with gestational hydralazine (HC + GH), obese offspring of lean mothers with gestational hydralazine (CH + GH), and obese offspring of obese mothers with gestational hydralazine (HH + GH). A schematic representation of the animal model and experimental groups is shown in Figure 1.

All male offspring underwent intraperitoneal glucose tolerance testing (IPGTT) and blood pressure measurement using a non-invasive CODA^® tail vein cuff method (Kent Scientific, CT, United States) during the week prior to the endpoint. For IPGTT, animals were fasted for 5 h, then a glucometer (Accu-Chek^®, Roche Diagnostics) was used to measure the blood glucose level in tail blood. An intraperitoneal dose of glucose (2 g/kg, Phebra, Australia) was administered, and the blood glucose level was subsequently measured at 15, 30, 60, and 90 min. The trapezoidal rule was used to determine the area under the curve (AUC).

Offspring were harvested at 32 weeks after fasting for 5 h. Isoflurane (2–3%) was used for anesthesia, then cardiac puncture was performed to collect blood. Urine was collected via bladder puncture. The body was perfused with phosphate-buffered saline, prior to harvesting the kidneys, liver and fat. 10% formalin was used to fix tissues for histological analysis; alternatively, they were snapped frozen for extraction of protein, DNA and RNA.

Insulin was measured in mouse serum (n = 5) via the Ultra-Sensitive Mouse Insulin ELISA Kit (Crystal Chem, IL, United States), and insulin resistance index (HOMA-IR) was determined using the following equation: fasting insulin (mU/L) × fasting glucose (mmol/L)/22.5. Serum creatinine was determined (n = 8) using a colorimetric assay (Cayman Chemical, MI, United States). Serum was incubated with triacylglycerol reagent and measured colorimetrically to evaluate triglyceride concentration (n = 5–6). Non-esterified fatty acids (NEFA) were measured (n = 6–10) with a NEFA kit following the manufacturer’s instructions (WAKO, Osaka, Japan).

The spot urine sample was used to calculate the albumin:creatinine ratio (ACR) as a measure of albuminuria. Urine concentrations of albumin and creatinine were determined (n = 6–8) using a mouse albumin ELISA kit (Crystal Chem, IL, United States) and a colorimetric assay kit for creatinine (Cayman Chemical, MI, United States).

After extracting genomic DNA from renal tissue (n = 4–6) using the ISOLATE II Genomic DNA Kit (Bioline, London, United Kingdom), 5-methylcytosine (5mC) percentage of total DNA was calculated with the Methylated DNA Quantification Kit (Abcam, Cambridge, United Kingdom). An input of 150 ng of genomic DNA was used.

To assess glomerulosclerosis, formalin-fixed kidneys (n = 6) were sectioned at 2 μm, then underwent Period Acid Schiff (PAS) staining and were evaluated under a light microscope (Leica, Germany). The glomerulosclerosis index was determined for 20 glomeruli per slide by two independent, blinded investigators. Renal cortical collagen I and III deposition were assessed by performing picrosirius red staining on 2 μm kidney sections. Using a digital camera linked to a light microscope (Leica Application Suite, Leica, Germany), four consecutive images were captured and subsequently analyzed by Image J software (National Institutes of Health, United States). Kidneys stained by PAS and picrosirius red were then assessed blindly by two independent investigators for tubulointerstitial fibrosis, tubular vacuolation, tubular dilatation and casts. Representative images of these histological features are shown in Supplementary Figures 1, 2.

Formalin-fixed kidneys (n = 6) underwent sectioning at 4 μm and were attached to slides. These were deparaffinized, hydrated and heated in a water bath at 99°C for 20 min for antigen retrieval using 0.01 M citrate buffer pH 6. Hydrogen peroxide (Sigma-Aldrich, Dublin, Ireland) was used to quench endogenous peroxidase. Protein Block Serum Free (Dako, Glostrup, Denmark) was used to block slides, which were then incubated overnight at 4°C with primary antibodies against fibronectin (dilution 1:1,000, Abcam, Cambridge, United Kingdom), collagen IV (1:1,000, Abcam), collagen III (1:750, Abcam), collagen I (1:750, Abcam), 8-hydroxydeoxyguanosine (8-OHdG, 1:750, Bioss, MA, United States), and nitrotyrosine (1:500, Merck Millipore, Darmstadt, Germany). Secondary antibody incubation was achieved using the horseradish peroxidase anti-rabbit Envision system (Dako, Japan), and slides were stained with 3,3′-diaminobenzidine tetrahydrochloride (Dako) and Mayer’s hematoxylin. Four to six non-overlapping images per slide were photographed at 200× magnification (Leica Application Suite, Leica, Germany) and the stained area was quantitated by two independent investigators using Image J software.

RNA was extracted from kidney tissue (n = 6) with the RNeasy Plus Mini Kit (Qiagen, CA, United States), and used to generate cDNA with the iScript cDNA Synthesis Kit (Biorad, CA, United States). The QuantStudio 12K Flex Real-Time PCR System (Thermo Fisher Scientific) was used to perform RT-PCR. Mastermixes incorporated QuantiNova SYBR Green (Qiagen) and the primers listed in Table 1.

All results are presented as mean ± SEM. GraphPad Prism version 8.3.1 (San Diego, CA, United States) was used to perform statistical analysis. Analyses were performed using three-way analysis of variance (ANOVA) followed by Fisher’s LSD test. Differences in pre-conception maternal body weight (Figure 2A) were analyzed using an unpaired t-test. P-values of < 0.05 were considered statistically significant.

HFD-fed dams weighed significantly more than those receiving chow (p < 0.0001, Figure 2A), confirming the maternal obesity model. There were no pre-conception weight differences between obese dams who later received hydralazine or saline, or lean dams who were later treated with hydralazine or saline. At weaning, body weights of male offspring were similar, regardless of maternal diet or hydralazine treatment (Figure 2B).

At 32 weeks, offspring weaned to HFD weighed significantly more regardless of maternal diet or hydralazine administration (p < 0.0001). Body weight was similar between HC and CC (Table 2). While HH weighed significantly more than CC (p < 0.0001), their weights were similar to those of CH (Table 2). Gestational hydralazine did not alter offspring body weight at 32 weeks. Kidney weights were increased in CH compared to CC (p < 0.01), but were similar between CC, HC, and HH (Table 2). Gestational hydralazine did not affect kidney weights. Kidney:body weight ratio increased in HC compared to CC (p < 0.05), and decreased in HH compared to CC (p < 0.05, Table 2). CH and CC had similar kidney:body weight ratios. Liver:body weight ratio was similar between CC and HC, but increased above control levels in CH and HH (p < 0.0001 for both, Table 2). Gestational hydralazine did not affect offspring liver weight. Epididymal fat:body weight increased in HC, CH, and HH compared to CC (p < 0.01, p < 0.0001, and p < 0.0001, respectively, Table 2). CH and HH had increased retroperitoneal fat:body weight compared with CC (p < 0.0001 and p < 0.001, respectively, Table 2). Interestingly, retroperitoneal fat:body weight was higher in CH than in HH (p < 0.01). Gestational hydralazine did not affect offspring epididymal fat:body weight or retroperitoneal fat:body weight ratios. Mean blood pressure was similar in all offspring groups and remained unaffected by maternal diet, gestational hydralazine and offspring diet (Table 2).

As displayed in Table 3, glucose tolerance was similar between CC and HC. CH and HH both displayed impaired glucose tolerance compared with CC (p < 0.0001 and p < 0.05, respectively). CH displayed worse glucose tolerance than HH (p < 0.001). Gestational hydralazine did not significantly impact offspring glucose tolerance at 32 weeks. Independent of maternal hydralazine treatment, terminal blood glucose level increased in CH and HH compared with CC (p < 0.05 and p < 0.01, respectively). Terminal blood glucose level was similar between CH and HH. Serum insulin was elevated in CH and HH compared with CC (p < 0.05 for both). There were no differences in serum insulin between CC and HC. Gestational hydralazine did not affect serum insulin in offspring. CH and HH had significant insulin resistance compared to CC (p < 0.01 and p < 0.05, respectively). CC and HC had similar HOMA-IR scores. Gestational hydralazine did not affect insulin resistance in offspring.

Serum triglycerides were similar between CC, HC, and CH, but were increased above control levels in HH (p < 0.01, Table 3). Serum triglycerides were lower in HH + GH than in HH (p < 0.01, Table 3). Serum NEFA were similar in CC, HC, and CH, but increased in HH (p < 0.05, Table 3). Gestational hydralazine did not alter NEFA in offspring (Table 3).

Urinary ACR increased above CC in HC and CH (p < 0.05 for both, Table 3). However, there were no differences between CC + GH and HC + GH (p = 0.52), and CC + GH and CH + GH (p = 0.62). Urinary ACR also increased in HH compared to CC (p < 0.0001, Table 3). As expected, the highest degree of albuminuria was seen in the HH group. Interestingly, HH + GH had reduced urinary ACR compared to HH (p < 0.01). Urinary ACR was not improved by gestational hydralazine in CH + GH compared to CH, or in HC + GH compared to HC, and there were no differences between CC + GH, HC + GH, and CH + GH (Table 3).

Serum creatinine was similar between CC, HC, and CH, however it was higher in HH than in CC (p < 0.001, Table 3), in keeping with the increased albuminuria seen in this group. Serum creatinine was significantly reduced in HH + GH compared with HH (p < 0.0001), which also reflected the findings for albuminuria. No other offspring groups demonstrated any significant benefit on serum creatinine from gestational hydralazine. This suggests that protective effects of hydralazine on offspring serum creatinine are only seen in animals with severe renal impairment due to combined insults of obesity and maternal obesity.

Renal global DNA methylation was elevated in HC and CH compared to CC (p < 0.05 for both, Figure 3A). In HH, there was a trend toward increased global DNA methylation compared to CC, although this was not significant (p = 0.12). In all offspring groups who were exposed to gestational hydralazine, there was a reduction in global DNA methylation. Renal global DNA methylation was reduced in CC + GH compared to CC (p < 0.05), HC + GH compared to HC (p < 0.0001), CH + GH vs. CH (p < 0.0001), and HH + GH compared to HH (p < 0.0001, Figure 3A), confirming the efficacy of gestational hydralazine as a DNA demethylating agent.

DNA methyltransferases (DNMT) are key enzymes involved in DNA methylation (Arce et al., 2006), whereas ten-eleven translocation-3 (Tet3) is a DNA demethylating enzyme (Tampe et al., 2017). Renal DNMT1 and DNMT3a mRNA expression increased in HC compared to CC (p < 0.01 and p < 0.05, respectively), and there was a trend toward higher mRNA expression of Tet3 in HC than in CC (p = 0.11, Figures 3B–D). Compared to CC, CH had increased DNMT1 mRNA expression (p < 0.05) and trends toward higher mRNA expression of DNMT3a and Tet3 (p = 0.19 and p = 0.23, respectively, Figures 3B–D). HH had similar mRNA expression of DNMT1, DNMT3a, and Tet3 as controls. Gestational hydralazine reduced DNMT1, DNMT3a, and Tet3 in lean offspring of obese mothers (HC + GH vs. HC, p < 0.01, p < 0.05, and p < 0.01, respectively, Figures 3B–D). Similarly, in obese offspring of lean mothers, gestational hydralazine reduced DNMT1 and Tet3 (CH + GH vs. CH, p < 0.05 for both), and there was a trend toward decreased expression of DNMT3a (CH + GH vs. CH, p = 0.07, Figures 3B–D). mRNA expression of these enzymes were similar between HH + GH and HH.

Reflecting the pattern seen for albuminuria, the highest degree of glomerulosclerosis was observed in HH, where the glomerulosclerosis index (GSI) was approximately fourfold higher compared to CC (p < 0.0001, Figures 4A,B). HC and CH also had increased glomerulosclerosis compared to CC (p < 0.05 and p < 0.01, respectively), and in these groups, the GSI was approximately double that of controls. Gestational hydralazine decreased glomerulosclerosis in HH + GH compared to HH (p < 0.001, Figures 4A,B). GSI was not improved by gestational hydralazine in CH + GH compared to CH, or in HC + GH compared to HC and there were no differences between CC + GH, HC + GH, and CH + GH.

Picrosirius red staining was increased in the kidneys of HC, CH, and HH, compared with CC (p < 0.001, p < 0.001, and p < 0.05, respectively, Figures 5A,B). Gestational hydralazine reduced picrosirius red staining in HC + GH compared to HC (p < 0.05), but did not affect staining in other offspring groups (Figures 5A,B).

Increased tubulointerstitial fibrosis was observed in HC, CH, and HH compared to CC (p < 0.05, p < 0.01, and p < 0.001, respectively, Figure 5C). Tubular vacuolation increased above control levels in CH and HH (p < 0.0001 for both), but was similar between HC and CC (Figure 5D). Tubular dilatation was higher in CH and HH than in CC (p < 0.01 and p < 0.05, respectively), but was not significantly increased in HC (Figure 5E). Tubular casts were increased in CH compared to CC (p < 0.05), but were similar between CC, HC, and HH (Figure 5F). Gestational hydralazine did not alter tubulointerstitial fibrosis, tubular vacuolation, tubular dilatation, or casts in offspring (Figures 5C–F).

Assessment of fibronectin and collagens I, III and IV was performed to evaluate renal fibrosis. Fibronectin mRNA expression was elevated above control levels in CH and HH (p < 0.05 and p < 0.01, respectively, Figure 6A). Gestational hydralazine did not attenuate fibronectin mRNA expression in offspring. Renal fibronectin protein expression increased in HC, CH, and HH compared to CC (p < 0.05, p < 0.001, and p < 0.0001, respectively). However, fibronectin protein expression was not significantly upregulated in the HC + GH and CH + GH groups compared to CC + GH, which suggests that gestational hydralazine exerted a protective effect in offspring exposed to either maternal or dietary obesity (Figure 6B).

Collagen IV mRNA expression increased in CH and HH compared with CC (p < 0.001 and p < 0.05, respectively, Figure 7A). Renal collagen IV mRNA expression tended to increase above control levels in HC. There was a trend toward increased collagen IV mRNA expression in CH compared with HH (p = 0.06). CH + GH had reduced collagen IV mRNA expression in kidneys compared to CH (p < 0.01). However, no protective effects of hydralazine on collagen IV mRNA expression were observed in other groups (Figure 7A). Increased renal collagen IV expression was demonstrated in HC and CH compared to CC (p < 0.05 and p < 0.0001, respectively, Figure 7B). Collagen IV expression was similar between CC and HH. Interestingly, HH had reduced collagen IV expression compared with CH (p < 0.01, Figure 7B). Collagen IV expression was reduced in CH + GH compared to CH (p < 0.05), however, gestational hydralazine did not reduce collagen IV expression in any of the other offspring groups.

Renal collagen III mRNA expression was elevated above control levels in HC and CH (p < 0.001 and p < 0.01, respectively, Figure 8A), and there was a trend toward increased expression in HH compared to CC (p = 0.17). Gestational hydralazine attenuated collagen III mRNA expression in HC + GH compared to HC (p < 0.0001) and in CH + GH compared to CH (p < 0.01, Figure 8A). Collagen III protein expression was increased in CH compared to CC (p < 0.01, Figure 8B). There was a trend toward increased collagen III expression in HC compared to CC (p = 0.06), and expression was similar between CC and HH (Figure 8B). Renal collagen III expression was reduced in CH + GH compared to CH (p < 0.05, Figure 8B). In other groups, maternal hydralazine had no beneficial effect on offspring renal collagen III expression.

Renal collagen I mRNA expression was similar in all offspring groups and was unaffected by gestational hydralazine exposure (data not shown). Collagen I protein expression was also similar in all of the offspring groups (data not shown).

Obesity is a known pro-inflammatory state, and as expected, renal monocyte chemoattractant protein-1 (MCP-1) mRNA expression was increased in CH and HH compared to CC (p < 0.01 and p < 0.05, respectively, Figure 9A). Gestational hydralazine did not affect MCP-1 mRNA expression in offspring kidneys (Figure 9A). Exposure to obesity, maternal obesity or gestational hydralazine did not alter renal CD68 mRNA expression (Figure 9B).

Obesity and maternal obesity were observed to increase renal oxidative stress. Compared to CC, there was increased renal mRNA NADPH oxidase-2 (Nox2) expression in HC and CH (p < 0.05 and p < 0.0001, respectively, Figure 9C). There was a trend toward increased Nox2 mRNA expression in HH compared to CC (p = 0.07). CH had significantly increased Nox2 mRNA expression compared with HH (p < 0.001). Gestational hydralazine attenuated Nox2 mRNA expression in CH + GH compared to CH (p < 0.001). There was a near-significant reduction of Nox2 mRNA in HC + GH compared to HC (p = 0.05, Figure 9C), however, Nox2 mRNA expression was similar between HH + GH and HH.

Renal catalase mRNA expression was similar between CC, HC, CH, and HH (Figure 9D). Gestational hydralazine did not alter renal catalase mRNA expression (Figure 9D).

8-OHdG renal expression was substantially increased in CH and HH compared to CC (p < 0.0001 and p < 0.05, respectively), and in CH compared to HH (p < 0.01, Figure 9E). Gestational hydralazine attenuated 8-OHdG expression in CH + GH compared to CH (p < 0.05), but did not affect expression in other offspring groups (Figure 9E).

Renal immunohistochemistry demonstrated higher nitrotyrosine expression in HC and HH than in CC (p < 0.05 and p < 0.01, respectively, Figure 9F). Gestational hydralazine did not affect renal nitrotyrosine expression in offspring.