Introduction

Previous large-scale epidemiological studies have shown that low birth weight increases the risks for a variety of non-communicable diseases in adulthood such as hypertension, chronic kidney disease, diabetes, obesity, and metabolic syndrome [1]. Infants with malnutrition show decreases in kidney nephron counts and pancreatic beta cells and vascular endothelial cell dysfunction [2].

Stem and progenitor cells have a capacity for epigenetic memory [3,4,5]. Mechanical stress on stem cells is incorporated into stem cell function as memory [6]. As adaptation in utero is mediated by mechanisms based on epigenetic information to regulate gene expression, it is assumed that the phenotype suitable for the in utero environment will persist after birth, rather than becoming over-adapted to the favorable out-of-utero environment [7]. Thus, a new possibility for the theory of fetal onset of lifestyle-related diseases is that the fetal malnutrition causes abnormal epigenetic memory of tissue stem and progenitor cells that induces the inability to respond to repair demands at the time of tissue damage.

Mesenchymal stem cells (MSCs) are involved in the growth and development of mesenchymal tissues. Once an organ is injured, these MSCs leave the niche by downregulating cadherin expression and rapidly shorten their cell cycle as native stem cells, which repeatedly divide to repair the injured organ as in organogenesis [8]. Label-retaining cells (LRCs) were found to be progenitor cells in kidney that function as a source of proliferating cells during tubular regeneration and repair cells of renal injury in a rat model of acute renal failure with ischemia–reperfusion injury [9]. Endothelial progenitor cells (EPCs) repair damaged blood vessels in pathologies of oxidative stress such as hypertension and diabetes, and when EPC function is reduced, repair of endothelial damage is inadequate, which leads to vascular and renal injuries [10].

Taurine, which has an important role in the homeostasis of living organisms, is supplied to the body through dietary intake and biosynthesis in the body. Taurine in the fetus and newborn is mostly derived from the mother, and fetal taurine is supplied from maternal blood via the placenta [11]. Inadequate taurine intake during fetal life can suppress the formation of stem cells [12].

In this study, we hypothesized that mechanisms underlying hypertension and renal injury in adulthood caused by fetal malnutrition would be associated with abnormal epigenetic memory of stem and progenitor cells in newborns. We generated undernourished rats during gestation with or without taurine. We monitored blood pressure (BP) in offspring rats from the undernourished mothers and evaluated renal tubular repair cells, LRCs, and EPC function. Kidney-derived MSCs in the offspring rats were isolated in which we evaluated the generation of renin and Ang II and epigenetic characteristics in open chromatin regions.

Materials and methods

Ethics and animals

This investigation conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, 1996). The Ethics Committee of the Nihon University School of Medicine approved the research protocols involving the use of living animals in this study (Approval No. AP16M028-4).

Creation of gestationally malnourished rats

Seven-week-old female Wistar rats were purchased from Charles River Japan (Yokohama, Japan), and after 1 week of environmental acclimation, they were divided into a control group (Control) and low-protein diet group as the group with fetal malnutrition. Control was fed a normal MF diet (Oriental Yeast, Tokyo, Japan) with a protein content 23.1% and normal drinking water. The low-protein diet group was further divided into a low-protein group (LP) and taurine-supplemented group (LPT). LP was fed a low-protein diet (Oriental Yeast) with a protein content of 8% from 8 weeks age, and LPT was additionally fed 3% taurine water (FUJIFILM Wako Pure Chemicals, Tokyo, Japan). The diets and drinking water for each group are summarized in Supplementary Table 1. Nine-week-old male rats were cohabitated with 10-week-old females for 2 weeks. Males were fed a normal diet and normal drinking water for the entire time except while cohabiting with females. Only the male rat offspring were evaluated.

Experimental protocol

Offspring for Control were obtained from three mothers, whereas offspring for LP and LPT were obtained from six mothers each. The average number of days from the start of cohabitation with males to delivery was 27.3 days for Control, 25.3 days for the LP, and 27.5 days for LPT, not significantly different. The offspring in each group were weaned at 21 days of age and fed a normal diet and normal drinking water after weaning. Some offspring at 14 and 30 weeks of age were used in subsequent experiments.

Offspring observed up to 60 weeks of age were weighed once every 2 weeks (K-1 800 g automatic upper plate scale, Fuji Keiki, Nagoya, Japan), and BP was measured using the tail-cuff method (BP98A, Softron Corporation, Tokyo, Japan). The BP was measured five times consecutively in each rat, and the average BP of three measurements, excluding the highest and lowest BP values, was used.

After storage of their urine at 60 weeks of age, rats were anesthetized with isoflurane inhalation (induction 4–5%, maintenance 2–3%) to alleviate pain. After skin disinfection, the abdomen was opened through a median incision, and all blood samples were collected from the inferior vena cava.

Glomerular injury score (GIS) and tubular injury score (TIS)

At 30 and 60 weeks of age, rats in Control and LP were anesthetized, and kidneys were removed through a midline abdominal incision, fixed with 10% formalin, and blocked. Tissue specimens were stained with hematoxylin–eosin. To quantify the amount of matrix in the glomeruli, 50 glomeruli on each section were selected randomly. The percentage of each glomerulus occupied by mesangial matrix was estimated and assigned a score of 0–4 as follows: 0, normal; 1, involvement of up to 25% of the glomerulus; 2, involvement of 25–50% of the glomerulus; 3, involvement of 50–75% of the glomerulus; or 4, involvement of 75–100% of the glomerulus. Grading for GIS was performed as described previously [13]. To quantify the tubulointerstitial area, 20 areas in each renal cortex were selected randomly. The TIS graded basement membrane thickening, dilation, atrophy, interstitial inflammation, interstitial fibrosis, tubular necrosis, desquamation, and hydropic degeneration as follows: grade 0, none; grade 1, <10%; grade 2, 10–25%; grade 3, 26–50%; grade 4, 51–75%; and grade 5, >75% in an average of 20 fields per kidney coronal section.

LRCs in ischemia-reperfused kidney

LRCs were evaluated by a previously reported method [14]. ALZET osmotic pumps (ALZET Corp., Cupertino, CA, USA) were used for harvesting after 1 week of abdominal administration of bromodeoxyuridine (BrdU, 100 mg/kg/day) to each group of offspring at 11 weeks of age. Two weeks later at 14 weeks of age, isoflurane inhalation anesthesia (induction 4–5%, maintenance 2–3%) was performed, and after disinfection, the abdomen was incised and the right kidney was removed, ischemia was performed on the left kidney by clipping the left renal artery for 45 min. At 24 h later, the left kidney was removed and fixed with paraffin, and 4-μm-thick sections were subjected to fluorescent immunostaining using Monoclonal anti-BrdU (clone BU-1) (Sigma-Aldrich, St. Louis, MO, USA) and secondary antibody Alexa Fluor 488 (Goat anti-Mouse IgG) (Abcam, Cambridge, UK) to detect LRCs in BrDU-positive cells. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) stain, corrected using the formula BrdU-positive cell count/total cell count (DAPI-positive cell count) × 100, and compared in 15 views, n = 3 per group, each with 5 views.

EPC colony-forming assay

At 60 weeks of age, whole blood was collected from the inferior vena cava of rat offspring of each group, and mononuclear cells were isolated. Mononuclear cells were seeded at 5 × 106/5 mL/well in 6-well dishes and incubated in a 5% CO2 incubator for 24 h. At the same time, vitronectin and gelatin were placed in 24-well dishes, and cells were incubated in a CO2 incubator for 24 h. After incubation, non-adherent cells were collected, suspended in culture medium, counted, adjusted to 2 × 106/mL/well, reseeded, and cultured for 6 days. The number of colonies with a diameter of 0.13 mm or greater was counted in 4 wells/sample using a square eye micrometer, and the mean value of 1 well was calculated to evaluate EPC function.

Isolation of renal MSCs

Offspring in Control and LP at 14 weeks of age were anesthetized with isoflurane inhalation (induction 4–5%, maintenance 2–3%), and after disinfection, the abdomen was cut open and the kidneys were removed. Obtained kidneys were washed, crushed, filtered, and hemolyzed, and culture was started in 100 mm dish to obtain primary cultured cells. Cultures were grown in Dulbecco’s Modified Eagle’s Medium (DMEM; D6046, glucose 5.6 mM, Sigma-Aldrich) at a normal sugar concentration containing 10% fetal bovine serum in DMEM. Subconfluent cells in dishes were washed twice with phosphate-buffered saline (Sigma-Aldrich) except for the culture medium, and then incubated with TrypLE Express (12605010, Thermo Fisher Scientific, MA, USA). Cells were then detached and collected with TrypLE Express (12605010, Thermo Fisher Scientific) and cultured in passaging medium. When the cell count was increased to 1 × 107 cells/mL, CD44-positive cells were isolated using a Dynal CELLection Biotin Binder Kit (Thermo Fisher Scientific), and passaged culture was continued. Flow cytometry with a Gallios system (Beckman Coulter, CA, USA) was performed on these cells using CD34, CD44, CD45, and CD90 antibodies.

Differentiation of renal MSCs

Isolated kidney-derived MSCs were incubated with 5 ng/mL of transforming growth factor-β1 (TGF-β1) (R&D Systems, Minneapolis, MN, USA) for 0, 6, 12, and 18 days, and the protein was extracted from the cells for western blot analysis. Obtained cells were collected with 1 mL of PIPA buffer and solubilized (200 W, 10 s ON/20 s OFF/25 cycles) using a Bioruptor UCD-250 (Cosmo Bio, Tokyo, Japan) ultrasonic disruption machine, incubated on ice for 15 min, and centrifuged at 3000 rpm for 1 min at 4 °C, after which the supernatant was discarded.

The nuclear fraction of the precipitate was dispersed in 200 µL of PIPA buffer, and the cell nuclei were disrupted using a Bioruptor UCD-250 (250 W, 30 s ON/30 s OFF/20 cycles). Centrifugation was then performed at 15,000 rpm for 15 min to obtain the nuclear fraction of the supernatant. These extracts were quantified with a Pierce BCA protein assay kit, heat-denatured at 95 °C for 5 min, and subjected to 10% polyacrylamide gel SDS electrophoresis. For western blot analysis of cells, the separated proteins were electrophoretically transferred onto a polyvinylidene difluoride membrane. Then the membranes were incubated with primary antibodies including αSMA (1:100; Abcam, Cambridge, UK), h-caldesmon (1:200; Santa Cruz Biotechnology, Santa Cruz, CA), renin (1:200; Proteintech Group, Rosemont, IL), and LXRα (1:4000; Abcam) overnight at 4 °C, washed, and incubated with horseradish peroxidase-conjugated goat anti-rabbit (Jackson ImmunoResearch, West Grove, PA, USA) used for renin and LXRα or with anti-mouse IgG (Jackson ImmunoResearch) used for αSMA and h-caldesmon. Finally, the membranes were treated with chemiluminescent reagent, exposed to a cooled CCD camera (LAS-3000 Mini; Fujifilm, Tokyo, Japan) with which the images were acquired. Each protein was quantified and normalized by β-actin using Image J (NIH, Bethesda, ML, USA).

Measurement of renal renin and angiotensin II (Ang II) in renal medulla

At 60 weeks of age, five rats from each group were decapitated, and the right kidney was immediately removed. The left kidney was also used to extract protein from the renal medulla, and renin expression was evaluated by western blotting. The right kidney was easily removed and placed in an Eppendorf tube and stored in liquid nitrogen. Later, the stored material was thawed, and the peptides eluted using Sep-Pak C18 cartridges (Waters Associates, Milford, MA, USA). The eluate was dried and separated by centrifugation, further purified by a Shodex column (OPD-50, Showa Denko, Tokyo, Japan), and the intrarenal Ang II was measured using a radioimmunoassay as described previously [15].

RNA-sequence and ATAC-sequence analyses of renal MSCs

RNA sequencing analysis (RNA-seq, Rhelixa, Tokyo, Japan) was performed on 5 × 104 cells from each group of cultured MSCs from each litter of rats [16], and the variation in expression of each gene in renal MSCs from each group was quantitatively evaluated. Furthermore, assay for transposase-accessible chromatin (ATAC) sequencing analysis (ATAC-seq, Takara Bio, Shiga, Japan) was performed on cultured MSCs from rats in Control and LP using 5 × 104 cells in each group. ATAC-seq is an experimental method that can map the accessibility of chromatin by selectively detecting and sequencing open chromatin structures on a genome-wide basis [17] and can identify enhancers, or regulatory regions that activate transcription, within the extracted open chromatin regions. This method identifies regulatory regions that activate transcription, known as enhancers, in the extracted open chromatin regions. In other words, we identified regions where ATAC-seq signals were increased and activated. We identified the regions with a large difference in activation peaks between the two groups, and in particular, the regions where the distance between the transcriptional regulatory region and its target gene was close, i.e., genes that are considered to be more involved in the transcription process. By identifying genes with large differences between the two groups in the results and in the magnitude of expression variation determined by RNA-seq, we could identify genes that may be contributing to the differences in open chromatin regions between the two groups.

Statistical analysis

All results are expressed as mean ± standard error (SE). Significant differences between groups were analyzed with a t-test and Mann–Whitney U test by Stat Mate-III software (Atom, Tokyo, Japan), with P < 0.05 indicating statistically significant differences.

Results

Phenotype, body weight, and BP in each rat group

Figure 1a shows the phenotypes of offspring rats in each group at 3 weeks of age. LP showed sparse body hair and poor hairiness compared to the other two groups. The number of litters per fetus was significantly (P < 0.05) higher in Control than in LP and LPT (Fig. 1b). Body weight (BW) in LP and LPT was significantly (P < 0.05) lower than BW in Control after birth, and BW in LPT was significantly (P < 0.05) higher than BW in LP (Fig. 1c). BP in LP was significantly (P < 0.05) higher than BP in Control and LPT after 44 weeks of age. There was no difference in BP between Control and LPT (Fig. 1d).

Fig. 1
figure 1

Experimental results. a Appearance of offspring with fetal malnutrition at 3 weeks of age. b Number of rats birthed in each group. *P < 0.05 between two columns. Control rats are born from mothers fed a normal diet, and LP and LPT rats are born from mothers fed a low-protein diet without or with taurine, respectively. Data are means ± SE (n = 3 for Control, n = 6 each for LP and LPT). Changes in body weights (c) and systolic blood pressure (d) in each group. #P < 0.01 vs Control. *P < 0.01 vs Control. Data are means ± SE (n = 6). W weeks

Full size image

Blood and urine examinations

Figure 2 shows blood and urine examinations in each group at 60 weeks of age. There were no significant differences between the three groups in blood glucose, plasma renin activity, serum urea nitrogen, and serum creatinine (Fig. 2a–d) nor in urinary total protein (Fig. 2e). However, urinary albumin excretion was significantly (P < 0.05) higher in LP and LPT than that in Control (Fig. 2f).

Fig. 2
figure 2

Blood and urine examinations in offspring with fetal malnutrition at 60 weeks of age. Control rats are born from mothers fed a normal diet, and LP and LPT rats are born from mothers fed a low-protein diet without or with taurine, respectively. *P < 0.05 between two columns. Data are means ± SE (n = 6). PRA indicates plasma renin activity, BUN blood urea nitrogen, Cr creatinine, UTP urinary total protein excretion, UAlb urinary albumin expression. (a) Blood sugar, (b) PRA, (c) BUN, (d) Cr, (e) UTP/Cr, (f) UAlb/Cr

Full size image

Kidney morphological changes

Morphological changes in kidney were evaluated by GIS and TIS in Control and LP at 30 and 60 weeks of age. There was no significant difference in GIS between Control and LP at 30 weeks, but GIS in LP was significantly (P < 0.05) higher than GIS in Control at 60 weeks. There were no significant differences in TIS between Control and LP (Fig. 3).

Fig. 3
figure 3

Renal injury in offspring with fetal malnutrition. Glomerular injury score (GIS, a) and tubulointerstitial injury score (TIS, b) were evaluated by hematoxylin–eosin staining of kidney from offspring at 30 and 60 weeks of age. Control rats are born from mothers fed a normal diet, and LP rats are born from mothers fed a low-protein diet. *P < 0.05 between two columns. Data are means ± SE (n = 3). GIS glomerular injury score, TIS tubular injury score, W weeks

Full size image

Number of renal LRCs

Figure 4a shows staining with DAPI and BrdU in nephrotubules of kidney from Control and LP at 14 weeks. The number of BrdU-positive cells was significantly (P < 0.05) lower in LP and LPT than Control and was slightly increased in LPT compared to LP, but the difference was not significant (Fig. 4b).

Fig. 4
figure 4

Renal label-retaining cells in offspring with fetal malnutrition. a Immunostaining of bromodeoxyuridine (BrdU)-positive cells in ischemia-reperfused kidneys from offspring at 14 weeks of age. Magnification is 200×. Blue-stained areas are nuclei stained with 4ʹ,6-diamidino-2-phenylindole (DAPI), and green-stained areas are BrdU-positive cells. b Number of BrdU-positive cells. Data are means ± SE (n = 3). c Number of vascular endothelial progenitor cell (EPC) colonies from offspring with fetal malnutrition at 60 weeks of age. Number of colonies was counted in 4 wells/sample, and mean value of 1 well was calculated to evaluate EPC function. Data are means ± SE (n = 6). *P < 0.05, **P < 0.01 between two columns. Control rats are born from mothers fed a normal diet, and LP and LPT rats are born from mothers fed a low-protein diet without or with taurine, respectively

Full size image

EPC colony formation

EPC function was evaluated by the number of EPC colonies from peripheral blood in each group at 60 weeks of age. The number of EPC colonies in LP was significantly (P < 0.01) lower than that in Control, and the number of EPC colonies in LPT was significantly (P < 0.05) higher than that in LP (Fig. 4c).

Differentiation of renal MSCs

Supplementary Fig. 1 shows flow cytometric analysis for the isolated CD44-positive cells from kidney in each rat group. Renal MSCs were positive for CD44 and CD90 and negative for CD34 and CD45.

Differentiation of renal MSCs with TGF-β1 was evaluated in Control and LP by western blot analysis (Fig. 5). Renal MSCs from Control did not express h-caldesmon, whereas MSCs from LP abundantly expressed h-caldesmon before differentiation. The abundance of h-caldesmon in MSCs of LP at 0 and 12 days was significantly (P < 0.01) higher than that in Control. There was no difference in αSMA between MSCs from Control and LP. Renal MSCs from LP abundantly expressed LXR and renin before differentiation. The abundance of LXR and renin in MSCs of LP at 0 days was significantly (P < 0.01) higher than that in Control.

Fig. 5
figure 5

Differentiation of renal mesenchymal stem cells (MSCs) in offspring with fetal malnutrition. The kidney-derived MSCs from offspring were differentiated with TGF-β1 on days 0, 6, 12, and 18. Control rats are born from mothers fed a normal diet, and LP rats are born from mothers fed a low-protein diet. Expression of proteins was quantified by western blot analysis. a h-caldesmon heavy caldesmon, b αSMA alpha-smooth muscle actin, c LXRα liver X receptor-α, d renin. Data are means ± SE (n = 3). *P < 0.05, **P < 0.01 between two columns

Full size image

Renal Ang II and renin

Figure 6a shows renal Ang II levels in kidney in each group at 60 weeks of age. The renal Ang II level was significantly (P < 0.05) higher in LP than that in Control. Figure 6b shows expression of renin by western blot in kidney medulla in each group at 60 weeks of age. Expression of renin in renal medulla was significantly (P < 0.05) higher in LP than that in Control.

Fig. 6
figure 6

Renal-related test results. a Renal angiotensin II levels in offspring with fetal malnutrition at 60 weeks of age. b Western blot analysis of renin expression in renal medulla of offspring with fetal malnutrition at 60 weeks of age. Data are means ± SE (n = 3). *P < 0.05, between two columns. c Heatmap display of the hierarchical cluster analysis. Control rats are born from mothers fed a normal diet, and LP and LPT rats are born from mothers fed a low-protein diet without or with taurine, respectively

Full size image

Analysis of open chromatin regions in renal MSCs

Figure 6c shows a heatmap of RNA-seq for each of the three groups. RNA-seq, which was evaluated in three renal medulla from each group, shows obviously different heatmap patterns between the groups. ATAC-seq analysis was performed in renal MSCs from each group. In total, 9538 genes were extracted with a fixed cut-off and then selected by activated transcription regulatory regions showing high peak expression. Among them, genes were selected based on a large difference in peak activation between two groups and the close distance of regions between the transcriptional control region. Among 1421 genes showing peak expression of more than twice that in LP compared to Control, 203 genes were found to be within 500 bp of the transcription control region and the target gene. However, among 171 genes showing more than twice peak activation, 17 genes were selected that had a distance of closer than 500 bp between the transcriptional control region and the target genes.

In this study, we compared the expression of the aforementioned 17 genes in the Control and LP groups by RNA-seq. Among them, genes showing greater expression variation in LP than Control were extracted (Supplementary Table 2). The genes F2rl1, Chac1, and Tspan6 were considered to be the genes with differences in open chromatin regions between the two groups, and they may be genes possibly contributing to the effects of fetal malnutrition.

We evaluated the expression of F2rl1, Chac1, and Tspan6 in renal MSCs during differentiation to mesangial cells by real-time PCR analysis. Expression of F2rl1 mRNA was larger in LP than that in Control before and during differentiation (Supplementary Fig. 2). Expression of Chac1 mRNA was larger in LP than that in Control at 12 days of differentiation. There was no difference in the expression of Tspan6 mRNA between LP and Control (data not shown).

Discussion

Previous experiments showed that a low amount of maternal–fetal protein induced glucose intolerance and vascular endothelial dysfunction [18] and that taurine supplementation during the fetal period improved the BW in adulthood in a mouse model of fetal stunted growth caused by uterine artery ligation in the mother of the fetus [19]. The low-protein diet to pregnant mothers has been known to affect BP and lipid metabolism during the adult phase of offsprings [18]. On the other hand, low carbohydrate (calorie) diets were known to affect the nervous system during the adult phase of offspring [20]. According to these experiments, we created a model of fetal hyponutrition in which the mother rats were fed a low-protein diet with or without taurine.

In the present experiments, offspring with fetal malnutrition show lower BW than offspring with fetal normonutrition. However, BW was increased in offspring from the malnourished mothers with taurine supplementation. In addition, BP was increased in offspring with fetal malnutrition, but this effect was suppressed by taurine supplementation.

Offspring with fetal malnutrition showed increases in urinary albumin excretion and glomerular injury. We examined the dynamics of LRCs in the offspring, which were localized in the nuclei of distal and proximal tubules in each group. The number of LRCs was significantly decreased in offspring with fetal malnutrition compared to offspring with fetal normonutrition. It is possible that the kidney damage seen in offspring with fetal malnutrition is due to the inability to repair the kidney damage caused by the decreased number of LRCs.

It was recently reported that nutrients are also involved in the epigenetic regulation of EPCs [21]. Fetal-period progenitor cells are pivotal for fetal development. Oliveira et al. [22] reported impaired EPC function in the adult stages of a mouse model with fetal malnutrition, and Khorram et al. [23] reported that maternal malnutrition suppressed angiogenesis in fetal organs. In the present experiments, colony formation of EPCs from offspring with fetal malnutrition was markedly reduced compared to that in offspring with fetal normonutrition, which was fully restored by supplementing the undernourished mother with taurine. It has been established that impaired EPC function is associated with chronic kidney diseases [24]. Glomerular endothelium maintains glomerular function and podocyte function through podocyte–endothelial cell interaction [25]. It is possible that impaired EPC function may have contributed to the glomerular injury observed in offspring with maternal malnutrition in the present experiments. Moreover, taurine supplementation to malnourished mothers completely restored the reduced EPC function in their offspring. It is possible that the impaired vascular repair capacity of EPCs is involved in the mechanism of vascular injury in adulthood due to fetal malnutrition, especially taurine deficiency. The suppression of kidney LRCs from MSCs in offspring with maternal malnutrition was not restored by taurine supplementation to the malnourished mother. Taurine nutrient may be essential for the hematopoietic stem cell lineage but may not affect the MSC lineage.

Renal MSCs from offspring with maternal malnutrition expressed h-caldesmon and αSMA before incubation with TGF-β1, suggesting that they had already differentiated into mesenchymal cells in which the transcription factor LXR-α was also upregulated. LXR-α is known to bind to the promoter of the renin gene and promote renin transcription [26]. Matsushita et al. [27] found that when MSCs underwent differentiation into vascular smooth muscle cells by TGF-β1, they did not produce renin after being fully differentiated, but they did produce renin after full differentiation into vascular smooth muscle cells by LXR-α, and renin was also produced by LXR-α when the MSCs were more de-differentiated than fully differentiated. Chen et al. [28] recently showed that renin-producing cells existed during the differentiation of MSCs to vascular smooth muscle cells, and renin generation was observed in undifferentiated vascular smooth muscle cells but express high levels of renin via high expression of LXR-α. In the present experiments, MSCs from offspring with fetal malnutrition differentiated early into mesenchymal cells and produced high levels of renin. In addition, Ang II levels in kidneys from offspring with maternal malnutrition were higher than those of offspring with maternal normonutrition. These results suggested that the intrarenal renin–angiotensin system is enhanced in offspring with maternal malnutrition, which induces hypertension and renal injury in the adult stage of these offspring.

To evaluate epigenetic memory in stem cells, we analyzed the open chromatin regions by RNA-seq and ATAC-seq for differentiation defects in renal MSCs from offspring with maternal malnutrition. We identified three genes, F2rl1, Chac1, and Tspan6, whose expression varied widely in renal MSCs from offspring with maternal malnutrition. Expression of F2rl1 and Chac1 mRNAs was increased in MSCs from offspring with the maternal malnutrition. Thus, the F2rl1 and Chac1 genes may contribute to epigenetic differences in MSCs from offspring with maternal normonutrition and malnutrition. In particular, F2rl1, also called protease-activated receptor 2 (PAR2), is a type of G-protein-coupled seven-transmembrane receptor expressed in vascular endothelial cells and is involved in inflammation and tissue damage that contributes to changes in circulation and BP [29]. PAR2 expression is increased in the vascular endothelium of spontaneously hypertensive rats and is associated with oxidative stress [30, 31]. Moreover, PAR2 has been reported to induce renal tubular epithelial inflammation by inhibiting autophagy through the PI3K/Akt/mTOR signaling pathway [32], and the lack of Par2 attenuates renal tissue damage and genes related to inflammation, fibrosis, and oxidative stress [33]. From these reports, it is possible that F2rl1 contributed to the hypertension and renal injury observed in offspring with maternal malnutrition in the present study.

In the Graphical Abstract, we have summarized the mechanisms underlying the contributions to hypertension and renal injury by the abnormal epigenetic memory of stem and progenitor cells in offspring with fetal malnutrition. Fetal malnutrition causes abnormal epigenetic memory of stem and progenitor cells that induces the abnormal differentiation of MSCs, decreases in LRCs, and dysfunction of EPCs. The abnormal differentiation of MSCs activates renal renin–Ang II systems to induce hypertension and renal injury. Decreases in LRCs induce renal injury, and dysfunction of EPCs induces hypertension. Supplementation of taurine in the malnourished mother prevents the dysfunction of EPCs and subsequent hypertension that can occur in adulthood.