Offspring of obese mice display enhanced intake and sensitivity for palatable stimuli, with altered expression of taste signaling elements

Maternal body mass index and gestational weight gain predict future obesity status of the offspring. In studies of both rodents and non-human primates, maternal obesity also predicts a preference for palatable foods in the offspring. In this study, we used C57BL/6J mice to investigate whether an underlying cause for an increase in palatable food consumption in the offspring of obese mice was a change in taste function. Adult female mice were fed a normal chow (NC) or a high fat diet (HFD) for 5 weeks before mating, then also during the gestation (3 weeks) and lactation (3 weeks) periods, with offspring always maintained on a normal chow diet; thus the only experience offspring had with high fat food was via maternal exposure. Offspring exhibited similar weight, blood glucose levels and baseline water and chow intake in adulthood. Taste response was assessed after reaching maturity, using brief-access taste testing, with female offspring of obese dams showing an enhanced response to sucrose, and both sexes consuming more sucrose, sucralose and high fat diet if from obese mothers. Offspring also exhibited increased taste bud expression of mRNA for sweet receptor subunits T1R (Taste receptor type) 2 and 3, as well as other markers associated with taste signaling. Taste morphology in both groups appeared similar. Results indicate that obesity in the mother may lead to unhealthy feeding behavior in the offspring, correlating with altered expression of taste signaling elements, which likely drive increased avidity for palatable foods.

Perinatal flavor programming. In neonates, sweet, umami, and low concentrations of salty substances are preferred, whereas bitter and sour substances alone are typically rejected. Studies show that exposure to certain taste stimuli during infancy or early childhood can modify these seemingly innate tendencies, and alter dietary preferences in children [26][27][28][29][30][31] . For instance, studies by Mennella et al. 32 reveal that mothers fed carrot juice during their third trimester pass carrot flavor acceptance to their offspring. There is evidence linking the parental diet and nutritional status to multiple phenotypic traits in the offspring 33 , highlighting the perinatal period as a window for nutritional intervention that can have lifelong effects on health and dietary preference in the offspring. The present study was designed to examine the effects of maternal obesity on the taste bud, and on resultant taste behavior in the offspring of obese mice. We hypothesized that maternal obesity would increase behavioral responses to palatable foods via the modulation of taste transmission elements in the offspring, highlighting the taste bud as a locus for juvenile obesity.

Methods
Animals. All procedures were reviewed and approved by the Institutional Animal Care and Use Committee at Cornell University. Experiments were performed in accordance with relevant guidelines and regulations. In-house bred C57BL/6J female mice were assigned to one of two diets (see Supplemental Table 1) ad libitum at 8 weeks of age: normal chow (NC, 18% kcal from fat) or a high-fat diet (HFD, 58.4% kcal from fat). Females were maintained on their respective diets during the pre-mating period (8-13 weeks of age). After 5 weeks on respective diets, females were mated with healthy lean males maintained on NC. Day 1 of pregnancy was determined by detection of a copulatory plug. Pups were weaned at 3 weeks onto NC. At 7 weeks offspring were single housed and given a week to acclimatize before testing (Fig. 1). Additional groups of females were also placed on HFD only before mating (but not through pregnancy or lactation), or only after mating. Full composition of diets can be found at Ferramosca et al. 34 .
Baseline measurement. At 8 weeks of age offspring were weighed and baseline measurements of water and NC intake were recorded over 48 h. Before behavioral testing, mice were fasted for 4 h, and circulating blood glucose levels from tail blood was measured using a OneTouch UltraMini glucose meter (OneTouch, Systems, Sunnyvale, CA).
Brief access lickometry. Taste responses were measured using a brief-access Davis Lickometer (DiLog Instruments, Tallahassee, FL). Training and testing schedule was adapted from Glendinning et al. 35 . Before testing, mice were partially water and food restricted by providing 1 mL of water and 1 g of normal chow for 23.5 h. Once the mouse initiated licking on the presented bottle, the timer started for 5 s and then the shutter closed. Each test session lasted no more than 1 h in total, during which the mouse could initiate up to 5 blocks of 7 concentrations (i.e. 35 total presentations). Lick responses were normalized, fit to nonlinear variable slope concentration-response curves, and compared using extra sum-of-squares F test. Tastant concentration-lick ratio response curves were fitted to the mean data for each group using a classical four parameter logistic sigmoidal dose-response equation in the nonlinear regression suite of GraphPad Prism (v5.0, GraphPad Software, San Diego, CA).
Two bottle testing -sucrose and sucralose. Mice were tested with water and sucrose (0.001, 0.02, or 0.1 M) or sucralose (0.1, 0.3, or 1.0 mM). All solutions were prepared fresh, purchased from Sigma Aldrich, St Louis, MO, unless stated. Mice were given simultaneous access to the two bottles with consumption measured over 48 h. Bottle placement was random and switched after 24 h.
Diet preference and intake. Diet preference testing methods were adapted from Vucetic et al. 36 and Carlin et al. 37 . Mice were single housed and trained to consume NC from dual hoppers for 48 h. Following this, mice were presented with NC in one hopper and HFD in the other for 24 h as a training period, then for 2 test days, with positions swapped every 24 h, and intake of each averaged.
Taste bud isolation and RNA extraction. Circumvallate taste buds were isolated from mice after Lickometer testing and one week of wash out consuming NC and water ad libitum. Mouse tongues were freshly excised following euthanizing with CO 2 and cervical dislocation. Circumvallate taste buds were extracted as www.nature.com/scientificreports/ previously 38 . Additionally, a piece of the epithelium posterior to the circumvallate papillae, termed the "nontaste" area was collected as a non-chemosensory control epithelial tissue, and processed in parallel. Quantitative real-time RT-PCR using Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA) was run on a QuantStudio 6 Flex Real-Time PCR System (Thermo Fisher Sci, Waltham, MA). Relative quantification was performed in triplicates using QuantStudio PCR Software, based on the 2 −ΔΔCt method. Beta-Actin was used as the endogenous housekeeping gene for normalization of genes of interest (see Supplemental Table 2).
Postmortem tissue collection and staining. One hour prior to euthanizing, mice were injected with 5-HTP (5-Hydroxytryptophan, 2 mg/25 g body weight) to enhance immunofluorescence stain for serotonin positive taste cells. Mice were then euthanized with CO 2 and cervical dislocation. Tongues were excised and rinsed in PBS. The circumvallate papilla was carefully isolated with a sterile razor and fixed in 4% PFA at 4 °C for one hour, cryoprotected in 30% sucrose overnight, and then embedded in OCT. Perigonadal fat pads were dissected and weighed. Livers were collected and immediately fixed in 10% neutral buffered formalin fixative for staining with oil red O. Liver sections were imaged using an Aperio CS2 at 40 × magnification (Leica, Wetzlar, Germany). Tongue tissue was sectioned at 10 microns and stained with hematoxylin and eosin (H&E) or antibodies for immunofluorescence (see Supplemental Table 3). Images were taken using an Olympus IX-71 microscope (Olympus Corp, Tokyo, Japan) with a Hamamatsu Orca Flash 4.0 camera (Hamamatsu Photonics, Hamamatsu, Japan). In order to avoid double counting single cells from taste buds in the circumvallate, every 8th www.nature.com/scientificreports/ section was used for quantification. Random taste buds from the left and right trenches of the circumvallate were selected, with a minimum of 10 taste buds per mouse (n = 4 per sex for each treatment). Slides were mounted using DAPI Fluoromount-G (4′,6-diamidino-2-phenylindole, Southern Biotech, Birmingham, AL). The anterior two-thirds of the mouse tongue was excised and fixed in 4% PFA for 24 h before staining with 1% methylene blue for 1 min, with tissue then rinsed with PBS, and imaged with an Olympus dissection scope (Olympus Corp, Tokyo, JP) with a Lumenera Infinity HD scope-mounted camera (Lumenera, Ottawa, CA), counted using ImageJ from a region of interest representing a 1 × 1 mm 2 .
Statistical analysis. The effects of maternal HFD were analyzed using a two-way ANOVA with sex and maternal treatment groups as factors. When a significant interaction was identified, data were analyzed with post-hoc Tukey tests. Count data were analyzed with linear, or Poisson's loglinear regression, depending on distributions. All analyses were carried out in GraphPad Prism 5.0, and using IBM SPSS 24 (IBM Corp, Armonk, NY), where results with p < 0.05 were considered statistically significant.

Results
Offspring phenotype appears similar between treatments. At 9 weeks of age the body weights for both sexes between the maternal treatments was comparable ( Fig Sweet preference and intake over 48 h were measured for both caloric (sucrose; Fig. 3D,G) and non-caloric (sucralose; Fig. 3E,H) solutions, in a separate group of naïve mice (n NC = 20, n HFD = 20). No significant differences in preference to sucrose (Fig. 3D) or sucralose (Fig. 3E) versus water among the two groups was evident, however, total intake of sweet solutions was enhanced in the maternal HFD treated offspring, with offspring of HFD treated mice consuming more sucrose ( Fig. 3G; F 1,114 = 11.95, p < 0.001) and sucralose ( Fig. 3H; F 1,114 = 28.27, p < 0.001) than respective controls. Elevated sucrose and sucralose intake while preference ratios remained similar between groups may have been due to elevated water intake along with sweet solutions, as preference data would take both into account, however this remains speculative. Preference for, and intake of high-fat diet was also evaluated in offspring (Fig. 3F,I). Female offspring of HFD-fed dams consumed significantly more HFD relative to control females (F 1,29 = 5.684, p = 0.0239). Finally, HFD treatment applied either before conception but not while pregnant, or solely during gestation and lactation (thus in mice not obese before conception), did not elicit any change in sweet taste response in the offspring (Supplemental Fig. 2).

Expression of taste signaling elements influenced by maternal diet. Maternal diet influenced
expression of all sweet receptor and sweet signaling genes tested (Fig. 4, n = 4 M and 4F each group, samples tested in triplicate). 2-way ANOVAs tested both main effects of maternal diet and sex, as well as interaction effect. For clarity, only significance of main effect of diet is displayed in Fig. 4. Testing revealed a clear effect p sex = 0.061; p interaction = 0.080) was detected. All genes tested were taste-specific, and showed no modulation in non-taste tissue (Supplemental Fig. 3).

Offspring display comparable taste bud morphology between treatment groups. An alterna-
tive hypothesis leading to altered taste function or expression is a change in taste morphology. The fungiform papillae are easily accessible structures in the anterior tongue, housing one or more taste buds, and are often studied due to their easy access. No difference in fungiform density was detected between treatment groups (  www.nature.com/scientificreports/

Discussion
Effects of maternal HFD on body weight, blood glucose, baseline water and chow intake, fat pads, and hepatic lipids. By week 9, no differences in body weight between maternal treatment groups was evident in our study ( Fig. 2A), which corresponds to results from Vucetic et al. 36 . Additionally, there were no differences in perigonadal fat pad weights, blood glucose levels, or baseline water/chow intake between treatments for either sex (Fig. 2B-E). Other strains have shown contrasting results where offspring of HFD-fed dams had increased epididymal fat pad mass, and fatty livers, in Swiss dams 39 and increased epididymal fat pads and altered liver histology in C3H/HeJ dams 40 , however in each of these measures, mice in our study were similar.
Offspring of HFD-fed dams show increased avidity for palatable, high calorie stimuli. Female offspring of HFD-fed dams exhibited an enhanced licking response to sucrose compared to controls. This indicates that a lower concentration of sucrose induced the same level of licking response in the HFD group as would a higher concentration in controls. We also examined offspring of dams fed HFD solely during the pre-conception period, or solely during the gestation and lactation periods to determine if either time period could elicit the same increase in sweet responsiveness, but found no difference between groups ( Supplementary Fig. 2). This suggests that maternal obesity and also the action of consuming HFD while pregnant underlies results, and is in line with studies by Treesukosol et al. 20 in rats. Similarly, gestational exposure to sucralose in mice did not result in any change in sweet taste responsiveness, or taste receptor expression in the offspring 41 . Sex differences in the behavior of offspring from maternal feeding studies are common, with females noted in previous studies to be particularly prone to overconsumption 42,43 , possibly due to an alteration in VTA and nucleus accumbens muopioid receptor expression observed in female rats that seems absent in males 44 , or higher energy expenditure 45 or altered plasma leptin levels 46 in males. Maternal HFD treatment resulted in increased sucrose and sucralose intake in offspring of both sexes compared to controls, without any direct exposure to HFD, sucrose or sucralose. Interestingly in humans, a decrease in sweet taste response has been linked with a desire for greater concentrations of sucrose 47 , suggesting that an acute versus chronic change in taste may lead to contrasting outcomes. Although offspring had never been directly exposed to HFD, the female offspring of HFD fed dams consumed more HFD than controls (Fig. 3I), suggesting enhanced sweet taste may contribute to a propensity to overconsume palatable foods in general, other taste machinery may be altered as suggested by results in Fig. 4, or that the formulation of HFD presents as sweeter than regular chow; all likely outcomes. Other studies have also shown higher HFD intake in offspring of maternal HFD treatment [19][20][21] , and that a maternal "junk food" diet 42 , causes female rats to increase intake of fats, where males do not.

Maternal obesity alters gene expression in the taste buds of the offspring. Sweet taste usually
implies calories, and thus can induce an appetitive response to promote ingestion 48 . We examined gene expression of mRNA for sweet receptor subunits T1R2 and T1R3, as well as several additional taste signaling elements. In offspring of HFD fed mice, we found increased T1R2 and T1R3 sweet receptor mRNA expression, as well as for other components of the sweet signaling pathway (Fig. 4). Although the functional consequences of increased sweet receptor mRNA expression are unknown, genetic deletion of T1R3 in mice results in reduced consumption of carbohydrates and fats 49,50 . Additionally, rats fed HFD for 6 weeks showed decreased T1R3 expression, which was associated with lower intake and preference ratios for saccharin solutions 51 . While this may represent one mechanism for a change in taste responsiveness in offspring of HFD fed dams, several other explanations should not be discounted. Among these, an alteration in food reward, changes to second messenger cascades beyond those assayed in our experiments, or a varied pattern of taste innervation in the progeny of HFD fed vs control dams would all be rational. An additional mechanism that could potentially alter responsiveness to sweet and fat stimuli is an altered balance of cell-types within the taste buds of offspring of obese mice, with more PLCβ2 positive cells in our results. PLCβ2 marks type 2 taste cells, and thus the cells that are responsible for the detection of sweet, bitter and umami tastes. The free fatty acid G-protein coupled receptor GPR40 (Ffar1), a proposed fat taste receptor, is located at least primarily in Type 2 taste cells 52 , with another, CD36, colocalizing in taste buds with α-gustducin 53 , another well-accepted Type 2 marker. Nonetheless, the upregulated cell population did not seem to be sweet-sensitive, with T1R3 cells unchanged between groups, suggesting that the expanded population may have been primarily bitter-sensitive. The reduction in CD36 observed in female offspring of HFD fed mice, accompanied by an increase in GPR120 parallels the findings of Ozdener et al. in cultured mouse and human taste cells chronically exposed to linoleic acid 54 . Another potential mechanism for differences between offspring groups would be due to differing metabolite profiles arising from the breakdown of fatty acids in the diets (for more on fat taste signaling see reviews 55,56 ). The balance of fatty acids between control and high fat diets varied in our setup as well as simply the fat percentage 34 , with saturated and monounsaturated fatty acids in particular increased in the HFD group, which would have divergent effects on metabolism in the dam and likely also the offspring. Finally, another likely source of regulation in the offspring of HFD-fed dams would be central reward circuits governing intake. We would not rule out additional regulation of such circuits, as shown in the offspring of rats fed a junk-food diet by Ong and Muhlhausler 57 , occurring alongside taste receptor regulation, unbalancing feeding behavior in these animals.
This work provides evidence that the maternal environment can result in long-term programming effects on the taste system. Permanent alterations and stable long-term repression of some genes maintained through cell division, as suggested by our results, could be the result of DNA methylation or histone modification 58 , as a maternal HFD has been shown to influence epigenetic machinery 59 . Increased preference for fat and sucrose in offspring of maternal HFD fed animals is associated with global and gene specific decreases in DNA methylation in the offspring brain 36 . Thus, maternal obesity may mediate long-term changes in gene expression in the offspring, including expression of taste receptors, via epigenetic regulation. While the specific mechanism Scientific RepoRtS | (2020) 10:12776 | https://doi.org/10.1038/s41598-020-68216-7 www.nature.com/scientificreports/ remains speculative at this point, our results introduce taste to the growing list of metabolic alterations arising from fetal programming, and add to the growing evidence that the taste bud plays a role in the etiology of obesity.

Conclusions
Overall, this study is the first to demonstrate that maternal exposure to HFD during the perinatal period leads to long-term consequences for taste signaling components, as well as feeding behavior. Offspring of dams fed a high-fat diet exhibited enhanced licking responses for sucrose solutions, as well as increased intake of palatable stimuli (sucrose, sucralose and high fat diet), while remaining equally responsive to neutral stimuli (regular chow and water). From a public health standpoint, improving our knowledge of prenatal and early postnatal factors that program obesity in the offspring may provide insight into therapeutic targets to combat the obesity epidemic, a disease easier to prevent than to cure.