Neonatal- maternal separation primes zymogenic cells in the rat gastric mucosa through glucocorticoid receptor activity

Neonatal- Maternal Separation (NMS) deprives mammals from breastfeeding and maternal care, influencing growth during suckling- weaning transition. In the gastric mucosa, Mist1 (encoded by Bhlha15 gene) and moesin organize the secretory apparatus for pepsinogen C in zymogenic cells. Our current hypothesis was that NMS would change corticosterone activity through receptors (GR), which would modify molecules involved in zymogenic cell differentiation in rats. We found that NMS increased corticosterone levels from 18 days onwards, as GR decreased in the gastric mucosa. However, as nuclear GR was detected, we investigated receptor binding to responsive elements (GRE) and observed an augment in NMS groups. Next, we demonstrated that NMS increased zymogenic population (18 and and 30 days), and targeted Mist1 and moesin. Finally, we searched for evolutionarily conserved sequences that contained GRE in genes involved in pepsinogen C secretion, and found that the genomic regions of Bhlha15 and PgC contained sites highly likely to be responsive to glucocorticoids. We suggest that NMS triggers GR- GRE to enhance the expression and to prime genes that organize cellular architecture in zymogenic population for PgC function. As pepsinogen C- pepsin is essential for digestion, disturbance of parenting through NMS might alter functions of gastric mucosa in a permanent manner.


Results
NMS decreases body mass during postnatal development. As neonatal maternal separation is characterized by daily deprivations of feeding and maternal care (3 h/day until weaning), it might lead to changes in body mass gain, and so, we tracked it in suckling (10, 14 and 18 days), weanling (21 days) and young-adult rats (30 days) (Fig. 1a). We observed that body mass was reduced in NMS pups (10 to 21 days) when compared to control counterparts that suckled normally (Fig. 1b). After cessation of treatment at weaning (21 days), rats gained weight, and body mass was similar between groups in young-adults (30 days) (Fig. 1b).

NMS increases corticosterone levels and decreases GR in the gastric mucosa in pups and youngadults.
Corticosterone is essential for cellular and metabolic functions, and its levels are regulated by different mechanisms [27][28][29][30][31]42,43 . In rodents, the period from postnatal day 2 to 14 is considered hypo-responsive to stressful conditions, and so, the activity from hypothalamus-pituitary-adrenal (HPA) axis is reduced [44][45][46] . Because our hypothesis was based on the effects of NMS on corticosterone activity on gastric epithelial cells, we first measured its concentration in the plasma. We found that 10-and 14-day-old pups did not show changes in hormone levels ( Fig. 1c), but interestingly, after that, at 18, 21 and 30 days, the levels of plasma corticosterone increased when NMS and control rats were compared (Fig. 1c).
Next, we evaluated the effects of NMS on GR in the gastric mucosa at 18 and 30 days. During third postnatal week, the receptor is active in gastric epithelial cells 23,25,35 but levels and function might not be exclusively regulated by corticosterone. Therefore, we studied both distribution and function. At first, we observed that NMS decreased GR at gene expression (Nr3c1) and protein levels (Fig. 2a,b). Importantly, though treatment was stopped by weaning (21 days), the reduction was still detected at 30 days (Fig. 2a,b).
We also analyzed the tissue distribution of GR in terms of gland regions, and we found that immunolabeled cells are spread along the gastric gland at 18 days whereas at 30 days, they are more restricted to the neck-base transition (Fig. 2c). NMS did not alter this distribution. Cytoplasmic and nuclear immunostaining were observed, but not quantified. Upon determination of the GR labeling index, we verified that NMS decreased the distribution of GR-positive cells at 18 and 30 days. So, both in pups and young-adult rats, we demonstrated that NMS increased plasma corticosterone and decreased GR in the gastric mucosa.
Increased GR nuclear binding in the gastric mucosa after NMS. As demonstrated before 23,25,47 and confirmed above, GR is expressed in the rat gastric mucosa during postnatal development. In the cell, inactive GR is a cytoplasmic protein that is part of a multi-protein complex 33 . Once activated by glucocorticoids, GR rapidly translocates to the nucleus and can regulate gene expression through at least three different pathways: (a) directly binding to GRE 31,48,49 ; (b) tethering to other DNA-bound proteins that act as transcriptional factors, and (c) in a composite manner, when it binds to GRE and associates to other transcriptional elements 28 .
Different disorders affect glucocorticoid-GR-GRE cascade 49 and lead to changes in development and function 50 . As we detected high corticosterone concentration and reduced levels of gastric GR after NMS, we evaluated the effects of NMS on GR binding to GRE in the gastric mucosa at 18 and 30 days. By using a binding assay SCIentIfIC RePORtS | (2018) 8:9823 | DOI:10.1038/s41598-018-28223-1 ( Fig. 3a), we observed that NMS increased the response at both ages (Fig. 3b), and comparatively, the effect was more prominent at 30 days.
NMS targets zymogenic cells. The results above showed that NMS increased plasma corticosterone, and reduced GR distribution in the gastric mucosa, but effectively, receptor's activity augmented, as binding to GRE increased in pups and young-adult rats. Recently, we demonstrated that during early weaning, glucocorticoids take part of differentiation of zymogenic cell population 35 , which derives from mucous neck cells 38,40 . Because NMS is also a condition that disturbs feeding and behavior during development and the gastric mucosa is responsive to endogenous corticosterone 23,25,35 currently, we studied the molecules involved in the characterization of zymogenic cells. Also, to compare to response of this population to others components in the gastric mucosa, we evaluated the expression of genes involved in parietal and mucous neck cell activity ( Supplementary Fig. 1), and Sgk1 that is a target of GR in gastric mucosa. We observed that Muc6 was increased both at 18 and 30 days, whereas Atp4b was not changed. Sgk1 increased in pups, but not in young adults ( Supplementary Fig. 1). In zymogenic cells, we found that NMS increased the expression of pepsinogen C gene (PgC) at 18 days, but the effect was not maintained at 30 days (Fig. 4a). The expression of gastric intrinsic factor (Gif) (another product of ZC) was not altered in these cells ( Supplementary Fig. 1).
After protein synthesis, PgC is stored in granules and secreted to the lumen, where it is activated to pepsin. We observed that levels augmented only at 30 days (Fig. 4b). Because differences between gene expression and protein concentration could be related to product localization, whether cytoplasmic or luminal, we used immunohistochemistry to identify the cells and we estimated the labeling index. We found that at 18 and 30 days, NMS increased the distribution of zymogenic cells that were synthesizing PgC, as detected by granular cytoplasmic immunostaining (Fig. 4c).
Zymogenic cells derive from mucous neck cells and the differentiation process involves Mist1 transcriptional factor 38,51 and moesin protein, which is important in the organization of secretion apparatus, from mucous into serous pattern 41 . In order to verify whether NMS would influence cell reshaping, we evaluated the expression levels of Mist1 and moesin, and observed that treatment increased expression (Bhlha15 and moesin genes, respectively) at 18 and 30 days. The protein concentration of Mist1 was higher after NMS in pups and young-adult rats, whereas moesin increased only in pups (Fig. 5). (c) Total corticosterone plasma level was measured by RIA (ng/mL). Values were individually represented, and lines indicate means ± SD for control (blue) and NMS (red) groups. *P < 0.05 after Student t test. (n = 15/group/age for body mass and n = 4/group/age for total corticosterone levels).
After detecting that NMS increased GR nuclear binding and that it affected the differentiation of gastric zymogenic cell population, we searched for evolutionarily conserved genomic regions (ECR) in Bhlha15 and PgC genes that might function as GREs. We aligned human, rat and mouse sequences, identified those that were conserved and searched for GR binding sites in both coding and non-coding regions. We found several putative GREs for each gene (Supplementary Fig. S2 and Table S2).
The results presented above suggested that GR might be involved with zymogenic cell differentiation and function, and so we used sections from 30-day-old rats to compare the localization of GR immunostaining with the distribution of zymogenic and parietal cells, which identify the gland regions ( Supplementary Fig. 3). Different sections and animals were studied and GR was more concentrated at neck-base transition, where cells were immunolabeled for Mist1 and few parietal cells (DBA-stained cells) were detected.

Discussion
As neonatal-maternal separation impacts development, metabolism and behavior 10,12,[52][53][54] , and the growth of gastric mucosa might be affected, we currently investigated some important cellular and tissue aspects that could be targeted by corticosterone activity in rat pups and young adults. The first parameter to be checked was weight gain, and we found that up to weaning, NMS reduced body mass, when compared to counterparts that were suckling regularly, but after that period, as rats fed on chow and were not dependent on maternal care, weight gain was restored. Though we did not attempt to measure milk intake during NMS, which would be technically doubtful, we considered that as pups were weaned, the NMS group was able eat and to recover, especially because they were released from the disturbing condition induced by treatment.
Next, we evaluated corticosterone levels, and we observed that NMS did not alter the concentration at 10 and 14 days, corroborating studies that showed that such treatment does not influence hormone levels during hypo-responsive period 53 , which extends through the second postnatal week in rats 9,23,25,45,46 . Moreover, the lack of corticosterone response at 10 and 14 days also represents a protective mechanism to guarantee the ontogenesis of systems and organs, including the gastrointestinal tract. However, at 18, 21 and 30 days, we found that NMS increased corticosterone levels, suggesting that even though treatment ceased at 21 days, the effects were still detected in young adults, and the circulating glucocorticoid might interfere in cellular functions. Such high steroid concentration induced a negative feedback over GR in gastric epithelial cells, as both expression and distribution were reduced. When compared to other feeding and nutritional models, NMS differs from short-term fasting in pups, in which high corticosterone is concomitant with increased gastric GR and CBG binding, which renders a very low traffic into cellular nucleus, and so a reduced final activity 23 .
As the function of corticosterone-GR system has been shown to be important in the growth and differentiation of gastric mucosa, and results were suggestive of negative feedback during NMS, we tested whether nuclear receptors were binding to GRE, aiming to determine effectiveness. Of note, through morphology, we identified cytoplasmic and nuclear immunostaining, indicating different levels of activity in gastric cells. So, we compared GR-GRE interaction at 18 and 30 days, which represent important stages of growth in rats, and found that GR binding to GRE increased in both ages. We also registered high Sgk1 expression in NMS group in pups, indicating high GR activity on epithelial cells. Therefore, the practice of NMS during the first weeks of development induced alterations in corticosterone response that were noted during treatment (18 days) and in young adults as well, suggesting long-lasting effects. Similarly, early weaning (another disturbance of feeding and maternal-neonatal interaction) reprograms essential elements gastric growth 24,55 through a mechanism that involves the direct action of corticosterone on secretory cells 25,35 . We should mention that mucin secreting cells (both in the surface and neck of the gland) were investigated, and only Muc6 expression increased after NMS. The morphological parameters evaluated for cells distribution were not affected by the change of feeding pattern (data not shown). The conversion of pepsinogen C into pepsin is essential for digestive functions 56 and PgC transcription increases in parallel to gastric ontogenesis 57 during suckling-weaning period, in a way that by the end of third postnatal week, PgC is the main product of zymogenic cells 31,35,57,58 . Because NMS augmented GR-GRE binding  Supplementary Fig. 4). Densitometry was normalized to Beta-actin loading control. (c) Representative photomicrographs of the base region of gastric glands, showing PgC immunohistochemistry. Negative control is devoid of any reaction. Bars = 50 µm; high magnification at 30 days: bars = 25 µm. For both ages, the last panel shows X2 digital zoom to depict cellular details. Labeling index (LI) (%) was obtained as described for control (blue) and NMS (red) groups at 18 and 30 days. Values are individually represented as well as means ± SD. *P < 0.05 as compared to control group by Student t test.
in the nucleus of gastric epithelial cells and the zymogenic cell population might be a target for such response, we evaluated the expression, levels and distribution of PgC. We found that at 18 and 30 days, NMS increased PgC content, as a whole, though a differential regulation was observed on mRNA and protein expression in each age individually. As mentioned before, the reorganization of secretory apparatus from mucous neck cells (glycoproteins) into ZC during transition depends on the expression of scaling factors 38,40,41,51,59 , and among them, we studied Mist1 and moesin. We verified that NMS increased both factors at mRNA and protein levels, suggesting that Bhlha15 and moesin genes are kept activated by interference in maternal interaction during suckling period. Furthermore, as both Bhlha15 and PgC genomic regions contained multiple putative GRE, we considered that NMS might affect the differentiation of zymogenic cells as well as pepsinogen synthesis through corticosterone response. Also, the proximity of GR to ZC and parietal cells at the base is suggestive of an important signaling function in these cells. So, our current results are suggestive of a priming action of NMS through GR onto genes that regulate the secretory apparatus in zymogenic cells and thus, the function of PgC in the stomach.
In brief, we showed that neonatal-maternal separation affected body weight gain throughout postnatal development and increased the total corticosterone levels from 18 days onwards; although GR expression and protein levels were down-regulated, the receptor activation through binding to GREs increased; the expression of genes that encode the proteins essential for digestion was augmented by NMS, and levels remained high in young-adult rats. We concluded that the genes involved in the coordination of zymogenic cell differentiation and pepsinogen C synthesis might be reprogrammed by neonatal-maternal separation through corticosterone activity. Interestingly, these findings brought new data to studies conducted before on the important roles of corticosterone on pepsinogen levels 35,[60][61][62] , and more importantly, we were able to combine them to the advances of gastric gland architecture described recently and mentioned above, and to the delicate scenario of parenting and maternal-neonatal interactions. The clinical consequences of the changes currently shown should be explored in order to guide future directions in different fields.  Supplementary Fig. 4). Protein levels were determined after comparison to the loading control with Beta-actin. Values are individually represented as well as means ± SD for control (blue) and NMS (red) groups. *P < 0.05 as compared to control groups by Student t test.

Material and Methods
Rats and neonatal maternal separation. Wistar rats were obtained from the Animal Colony at the Department of Cell and Developmental Biology (Institute of Biomedical Sciences -ICB, University of São Paulo). Pregnant females were kept in individual cages at 22 °C under 12 h light and 12 h dark cycle (lights on at 06h00) with free access to food and water. Delivery was set as day 0 and litters were culled to 9 pups on the 2 nd postnatal day. All procedures were performed according to the guidelines of National Council of Ethics with Animals (CONCEA) and the protocols were approved by the Ethical Committee for Animal Use from ICB at University of Sao Paulo (17/2013). Experiments were conducted in order to minimize the number of animals and to allow best conditions during treatment.
On the 2 nd postnatal day, pups were randomly assigned into two groups, in which they either remained with the dam (control, C) or were separated from her (neonatal maternal separation, NMS). The control group suckled freely until weaning (21 days), whereas NMS pups were separated from the dams 3 h/day (10h00 to 13h00) (day 2 to 21) (Fig. 1a). For that purpose, pups were placed in warmed small plastic cages (stabilized to 32 °C). Weaning was set on day 21, and after that, all rats were fed ad libitum with chow (Nuvilab, CR-1 Quimtia SA, Paraná, Brazil). Body mass was recorded regularly throughout experimental period to control weight gain.
Corticosterone radioimmunoassay. After anesthesia (Isoflurane, Cristalia, Itapira, SP, Brazil), blood was collected from abdominal aorta (Fig. 1a) in heparinized tubes (Liquemine, 5000 IU, Roche, Basel, Switzerland), centrifuged (15000 g, 10 min), and the plasma was kept at −20 °C. Euthanasia was conducted at 17h00 to prevent circadian variation. Total corticosterone was measured by RIA using Rat 125 I corticosterone kit (ICN Biomedicals, Irvine, CA, USA). The amount of radioactivity was evaluated in a γ-scintillation counter (Wizard Automatic Gamma Counter, PerkinElmer, Waltham,, MA, USA). The intra-assay coefficient variation was 5.4% and the inter-assay coefficient variation was 9.3%. Stomach sample collection. Animals were euthanized for stomach collection (Fig. 1a). The organ was excised, opened along the lesser curvature, rinsed in 0.9% saline and submitted either to scraping mucosa (corpus region) or fixation in 10% formaldehyde. Samples that were isolated after scraping were used for extraction of RNA or proteins (total and nuclear contents). To that, the scraped mucosa was respectively collected in 100 µl  Fig. 1). Densitometry was performed with ImageJ (1.37 v Software, NIH Public Domain) and bands were cropped to represent control and NMS groups. Gels were run in replicates to confirm results. GR binding assay. Nuclear proteins were obtained from gastric fresh samples scraped and placed in pre-chilled glass pestles to be manually homogenized according to instructions (Active Motif). Lysates were obtained (1X hypotonic buffer, added of proteases inhibitors), incubated on ice (15 min), centrifuged (850 g, 10 min, 4 °C), and the supernatant was transferred into pre-chilled micro-tube for efficiency checking. After a sequence of pellet washings in buffers for separation of cellular fractions, the supernatant containing the nuclear fraction was collected, separated in aliquots and stored at −80 °C. Nuclear and cytoplasmic protein concentration were determined 64 .
The protein quality and the cross contamination were checked after Western blot, using 10 or 5 µg of cytoplasmic and nuclear aliquots, respectively. Proteins were separated into 12% SDS-PAGE and blots were conducted as described to be incubated with rabbit anti-laminB1 (5 µg/ml) (Abcam, Cambridge, UK) and monoclonal anti-αtubulin (5 µg/ml) (Sigma Aldrich) (overnight, 4 °C). Bands were detected with ECL kit, as described above. Nuclear protein lysates were used for GR binding assay with TransAM GR Kit (Active Motif) that contains the reagents and a 96-well plate on which an oligonucleotide carrying a GR consensus binding site (5′-GGTACA nnnTGTTCT-3′) was immobilized. To detect GR binding, we added to each well: complete binding buffer, wild-type oligonucleotides (provided as a competitor for GR binding) or mutant oligonucleotides (provided as non-effector to GR binding) and nuclear extract (Fig. 3a). As positive controls, we ran HeLa nuclear extract, and to blank wells, we added complete binding buffer. Each sample was tested in duplicates. The plate was incubated, washed, and the wells were exposed to anti-GR (1:1000 1 h, RT), and then to a secondary antibody (same dilution). The developing solution was added for colorimetric reaction, which was stopped and read at 450 nm (EspectraMax, Molecular Devices, Sunnyvale, CA, USA).
Immunohistochemistry. Non-serial sections were rehydrated and the peroxidase activity was blocked with 3% H 2 O 2 in methanol. For GR, antigen retrieval was performed with 10 mM citric acid (pH 6.0 in microwave). Non-specific reaction was blocked with 10% normal goat or rabbit serum. Sections were incubated (4 °C, overnight, Santa Cruz Biotecnology) with: polyclonal rabbit anti-GR (10 μg/ml) (M20), polyclonal goat anti-pepsinogen C (1 μg/ml) (I-19). After adding secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA, USA), the peroxidase activity was developed with H 2 O 2 and diaminobenzidine solution (Liquid DAB Kit, Dako, Carpinteria, CA, USA). Tissue sections were counterstained with Mayer's hematoxylin. Negative controls were performed with normal serum to replace the primary antibodies.
In order to evaluate GR and pepsinogen C labeling indices, only sections that were longitudinally gastric mucosa in the corpus region were considered. Representative images were acquired using light microscope (BX51, Olympus, Canada) and Image ProPlus v.5.2 (Media Cybernetics, Silver Spring, MD, USA). The labeling index (LI) for GR was determined by counting 2,500 epithelial cells that were reactive or not to GR immunohistochemistry along the gastric gland. GR LI was estimated as labeled cells/total epithelial cells X 100.
In order to study pepsinogen C distribution in the gastric mucosa, we considered the whole bottom of glands at 18 days, as cells were not completely mature, and the zymogenic cells, as morphologically identified at gland base in 30-day-old rats. So, at 18 days, we determined the labeling index (LI) by counting 1,000 epithelial cells that were PgC-positive or not. At 30 days, we counted 1,000 zymogenic cells as labeled or not. PgC LI was estimated as labeled cells/total cells X 100 (35).
Genomic analyses. The human genomic (hg19) regions for Bhlha15 and Pgc were aligned with the rat (rn4) and mouse (mm10) genomes in the ECR browser (ecrbrowser.dcode.org). The evolutionarily conserved regions that appeared after alignment in Pgc coding and surrounding non-coding regions of all three species were submitted to the publicly available TRANSFAC professional V10.2 through rVista 2.0 and analyzed for the presence of the glucocorticoid response element (GRE-C), its half site matrix (GR_Q6-01) and high affinity binding site (GR_01) in the ECR of all three species (Supplementary Fig. S2).
The Bhlha15 region is not annotated for the rat genome in the ECR browser, thus we first analyzed the evolutionarily conserved regions in the mouse and human sequences. The peaks that appeared in both species were submitted to TRANSFAC V10.2 through rVista 2.0, filtered for potential GR targets. Thereafter we used the mouse ECRs that presented potential GR targets to query the rat genome via BLAT in the USCS Genome browser, retrieving the homologous regions in the rat genome that corresponded to the mouse GR regulatory regions into the rat genome (rn4).

Statistical analyses.
The results were represented as means ± SD and analyzed by Student t test to evaluate differences between control and NMS groups. Statistical difference were considered when P < 0.05.