We previously generated a whole-body in vivo Krtcap3-KO on the WKY (WKY/NCrl; RGD_1358112) inbred rat strain (WKY-Krtcap3^em3Mcwi) and established a breeding colony at Wake Forest University School of Medicine (WFUSOM) in 2019 (Szalanczy et al., 2022). At Building A of WFUSOM, rats were housed in ventilated cages (46 cm × 24 cm × 20 cm) at 22°C in a 12 h light/dark cycle (dark from 18:00 to 6:00) at standard temperature and humidity conditions, and given ad libitum access to water. Bedding was 0.25 in corn cob with a paper puck for enrichment. Breeder rats and experimental rats prior to high-fat diet (HFD) start were given ad libitum standard chow diet (Lab Diet, Prolab RMH 3000, Catalog #5P00). At 6 weeks of age, experimental WT and KO rats were placed on experimental diet as described below (n = 8 female rats per genotype and stress condition).

To briefly describe the postnatal period of the experimental rats, litters were between eight to twelve rats and only rats from the first litter were used in the current study. Dam and sire were housed together during the postnatal period. Rats were not disturbed for the first 5 days after birth, and after were handled once per week by animal care staff during cage cleanings. In addition, pups were handled once by experimenters to collect ear punches for genotyping before weaning. Pups were away from their dam for no more than 10 min and tools were cleaned between each pup. This method of ear punch collection was also used in previous studies with the WKY-Krtcap3^em3Mcwi strain (Szalanczy et al., 2022; Szalanczy et al., 2023). Pups were exposed to the same lighting, noise, and traffic conditions of Building A as all other rats.

Experimental animals in the stress study were weaned at three weeks of age and housed as two rats of the same genotype per cage. Animals were housed at Building A in the same room as the breeding colony up until approximately 23 weeks of age, at which point they were transferred to the nearby Building B, which is better set up for rodent behavioral testing. At Building B, rats were housed in a male/female room in non-ventilated, open-air wire top cages (26.5 cm × 48 cm x 20 cm) with aspen shavings. Rats were maintained on the same 12 h light/dark cycle (dark from 18:00 to 6:00) and access to food and water remained ad libitum. Building B is an older building that does not allow temperature and humidity to be as tightly controlled as Building A, but conditions are appropriately managed with animal care health observation and veterinary oversight.

We also sought to evaluate Krtcap3 expression in multiple tissues between adolescent WT and KO rats from the colony without diet or stress exposure. Female WT and KO (n = 4 per genotype) rats were weaned at 3 weeks of age and placed into mixed-genotype cages with littermates, with two to four rats per cage. Rats were housed in Building A and maintained on the same chow diet as breeders. One week later, rats were euthanized, described below.

All experiments were performed using a protocol approved by the Institutional Animal Care and Use Committee at WFUSOM.

Experimental rats were genotyped at the Medical College of Wisconsin (MCW) as described elsewhere (Szalanczy et al., 2023).

Female experimental rats were weaned at 3 weeks of age, weighed, and placed two per cage in same-sex, same-genotype cages. Rats were randomly assigned to either a control or stress group, as described in detail below. Stress study rats were initially maintained on the same diet as breeders (see above), and body weight was recorded weekly starting at 4 weeks of age. To remain consistent with our previous studies (Szalanczy et al., 2022; Szalanczy et al., 2023), all rats began a HFD (60% kcal fat; ResearchDiet D12492) at 6 weeks of age. Rats were allowed access to diet ad libitum. Cage-wide food intake was recorded weekly from the time of HFD start until the end of UCMS.

From the time of weaning, the study lasted for 22 weeks (Figure 1). For rats assigned into the stress group, a mild stress exposure began at 22 days of age and lasted for approximately 12 weeks. Control rats underwent the same metabolic and behavioral phenotyping as stress rats, but were not exposed to the stress protocols. The first EchoMRI analysis was conducted 3 weeks after stress onset, at HFD start, while an acute restraint test took place 10 weeks after stress onset. The first two behavioral tests were the novelty suppressed feeding test (NSF) conducted 11.5 weeks after stress onset and the forced swim test (FST) conducted a week later. Another EchoMRI analysis was run 13 weeks after stress onset, leading immediately into a week-long period of individual housing to assess food intake. Following return to cage-mate, UCMS began 15 weeks after initial stress onset, and continued for 4 weeks, concluding with another EchoMRI analysis and an intraperitoneal glucose tolerance test (IPGTT). After rats were transferred to Building B they were given an open field test (OFT) 20.5 weeks after initial stress onset. 22 weeks after stress onset rats were euthanized (Sac). Blood was collected at three points to measure CORT responses: at the acute restraint test, immediately prior to UCMS, and at study termination.

Throughout the course of the study, body weight, and cage-wide food intake were measured weekly. EchoMRI analysis (EchoMRI LLC, Houston, TX) was conducted at three, 13, and 18.5 weeks after first stress exposure (Szalanczy et al., 2022). Immediately following the EchoMRI analysis 13 weeks after stress onset, rats began a seven day-long period of individual housing to measure food intake as previously described (Szalanczy et al., 2023). 19 weeks after stress onset, rats were fasted overnight before being administered an intraperitoneal glucose tolerance test (IPGTT) as performed elsewhere (Szalanczy et al., 2022; Szalanczy et al., 2023). We measured blood glucose (Contour Next EZ) at fasting and 15, 30, 60, 90, and 120 min after a 1 mg/kg glucose injection and calculated glucose area-under-the-curve (AUC) (Szalanczy et al., 2022; Szalanczy et al., 2023).

To measure adrenocorticotropic hormone (ACTH) in plasma collected at euthanasia, we used a sandwich ELISA kit (AbCam Ref #ab263880). Plasma samples were diluted 1:4 in 1X Cell Extraction Buffer PTR. The plate was incubated for 15 min in the TMB substrate, then Stop Solution was added, and the plate was analyzed at 450 nm against a 4-parameter standard curve.

We used a CORT competitive ELISA kit (ThermoFisher Ref # EIACORT) to analyze serum/plasma corticosterone collected throughout the study. As guided by the manufacturer’s instructions, samples were diluted at least 1:100, and analyzed at 450 nm against a 4-parameter standard curve.

RNA was extracted from fatty tissues using the RNeasy Lipid Tissue Mini Kit (Qiagen Cat # 74804). Non-fatty tissue, such as liver or intestine, were extracted by Trizol.

Pro-opiomelanocortin (Pomc) expression from the pituitary was measured between rats from the stress study to determine if there were genotype or stress exposure-driven differences on expression of the precursor to ACTH. We also examined expression of the glucocorticoid receptor (GR) nuclear receptor subfamily 3 group C member 1 (Nr3c1) in liver, colon, and pituitary between females of the main study to assess changes in CORT receptor expression by genotype and stress exposure. Colon and pituitary have high Krtcap3 expression (Supplementary Figure S1) while we previously investigated Nr3c1 expression in the liver (Szalanczy et al., 2023). We also investigated expression of hydroxysteroid 11-beta dehydrogenase 2 (Hsd11β2), responsible for catalyzing CORT deactivation, in the colon to determine if there were differences in CORT processing.

We then measured Krtcap3 expression between control and stress-exposed WT females in pituitary, adrenal, and colon to determine if expression had changed following exposure to the stress protocols. Results from the adolescent rats (described below) confirmed KO rats did not have Krtcap3 expression in these tissues, and they were excluded from this analysis.

We assessed Krtcap3 expression in adolescent females to confirm that Krtcap3 expression was significantly knocked-out in multiple tissues. We examined expression in tissues pertinent to metabolism or stress response: pituitary, adrenal, liver, whole hypothalamus, ileum, ovaries, and RetroFat.

To measure gene expression, either Gapdh or β-actin were used as housekeeping genes. Primers for all genes are found in Table 2. Fold change of the corresponding gene of interest transcript was calculated by the following equation: T r a n s c r i p t = 2 − ∆ ∆ C t Where ΔCt was the difference between the crossing threshold (Ct) of the gene of interest and the housekeeping gene, and ΔΔCt the difference between each sample ΔCt and the average control ΔCt. For analysis of the adolescent rats, the average control ΔCt was the mean ΔCt of WT rats. In gene expression analyses in the main study, the average control ΔCt was the mean ΔCt of control WT rats.

All data were analyzed in R (1.4.1103). Outliers were assessed either by Grubb’s test or by 1.5 * the interquartile range for phenotypes where n < 6 and were removed. Distribution was assessed by the Shapiro-Wilks test and data were transformed to reflect a normal distribution if necessary. Homogeneity of variance was assessed by Levene’s test.

For adiposity, behavioral, and gene expression results, data were first analyzed considering all four groups together. If ANOVA analysis indicated relevant interactions, the data were deconstructed appropriately. However, several phenotypes showed a visually stronger effect in one genotype relative to the other, even if ANOVA could not detect an interaction. In these cases, according to our a priori hypothesis that WT and KO rats would respond differently to stress, we split the data by genotype and assessed the effect of stress.

Single point adiposity and behavioral measurements were first assessed by a two-way ANOVA, where one factor was genotype (WT v KO) and the other was stress exposure (control v stress). Growth curves and cage-wide food intake were analyzed by a two-way repeated measures ANOVA, where the effect of genotype and stress exposure were examined over time. We measured cage-wide food intake from diet start until the end of UCMS (weeks 4–19 of stress). Food consumed during the week of individual housing was calculated as the average food consumed per day and the total food consumed during the week for each rat. If there was an interaction, or visually WT and KO rats responded differently, we split the data by genotype to evaluate the effect of stress or the effect of stress and time together.

Based on our previous study (Szalanczy et al., 2023) we anticipated there being three factors to consider when analyzing CORT and ACTH data from euthanasia: genotype, stress exposure, and euthanasia order (if the rat was euthanized first or second within a cage). We first examined CORT and ACTH from euthanasia using a two-way ANOVA, with only data from Rat 1 which we presumed to represent the basal condition, removing the consideration for the factor of order. The data were then analyzed by a three-way ANOVA and split according to significant interactions. We did not include rat order as a factor in analyses of CORT collected from the acute restraint test or prior to the UCMS protocol as both rats were handled at the same time. We used a mixed effects model to analyze CORT during the acute restraint, with the factors of genotype, stress exposure, and time restrained; Tukey’s multiple comparisons test compared CORT between the time points. Given the large effect of time, we then also assessed CORT at each time point separately with a two-way ANOVA. To evaluate CORT prior to UCMS exposure, we used a two-way ANOVA.

In the NSF, latency to feed and total time spent feeding could not be analyzed by a two-way ANOVA because there was a significant ceiling effect, as over one-quarter of the rats did not participate in the test in the given timeframe. Instead, we chose to re-conceptualize the latency to feed and time spent feeding phenotypes as participation in the test, and to then analyze by methods similar to those for survival analyses. Rats who participated in the test (that is, had a latency to feed less than 15 min) were considered “death events” while those who did not participate “survived”. We then analyzed this by the Kaplan-Meier estimator and log-rank test to examine the effect of stress exposure between each genotype.

Krtcap3 expression between WT and KO adolescents in multiple tissues was assessed by a t-test per tissue, and p-values for multiple comparisons were adjusted via the Holm-Sidak method. This same method was applied to comparing Krtcap3 expression between WT control and stress animals from the main study. Expression analyses of the other genes, which included KO rats, were first analyzed by a two-way ANOVA, and then appropriately deconstructed to examine effects of stress per genotype and effects of genotype per stress condition.

We verified the Krtcap3-KO in multiple tissues in female adolescent rats (Supplementary Figure S1). Compared to WT, KO rats had significantly lower Krtcap3 expression in all tissues that were assessed: pituitary (T_3.13 = 11.95, p = 6.17e-3), adrenal (T_4.77 = 11.03, p = 9.95e-4), liver (T_3.02 = 6.94, p = 0.024), hypothalamus (T_3.33 = 4.84, p = 0.026), ileum (T_3.05 = 8.34, p = 0.017), ovary (T_3.03 = 3.55, p = 0.038), and RetroFat (T_5.46 = 4.35, p = 0.024) tissues.

As expected, there were no differences in body weight at weaning (3 weeks of age) between WT or KO rats, and no differences in weight between animals assigned to the control or the stress groups (Supplementary Figure S2A). Surprisingly, despite exposure to noise stress for 3 weeks, there were no differences in body weight by genotype nor by stress exposure at 6 weeks of age at HFD start (Supplementary Figure S2B). There was, however, a modest increase in total fat mass in noise-exposed rats compared to control rats for both genotypes (F_{1, 26} = 5.38, p = 0.03; Supplementary Figure S2C).

Contrary to what we had hypothesized, the entirety of the stress protocol described here increased the body weight of the rats over time (F_{2.42, 67.89} = 3.13, p = 0.041) rather than decrease it. When WT and KO growth curves were assessed separately, however, stress exposure ultimately did not impact WT body weight over time (Figure 2A), while it did significantly increase KO body weight over time (F_{22, 308} = 2.03, p = 0.005; Figure 2A). Similarly, from the time of HFD start until the end of UCMS, stress led to an increase in cage-wide food intake over time (F_{1, 12} = 7.5, p = 0.018) as well as a stress by week interaction (F_{14, 168} = 1.88, p = 0.031) where food intake increased over time only in the stress-exposed rats. Although this effect appears to be driven mainly by the KO rats (Figure 2B), thereby explaining their increase in body weight over time, a larger n is needed to pick up an interaction and/or determine significant differences within each genotype separately (with two rats per cage the n is only 4 for each genotype/stress group).

EchoMRI analysis performed 13 weeks after mild stress initiation showed that stress-exposed rats of both genotypes had increased total fat mass relative to control counterparts (F_{1, 27} = 16.18, p = 4.17e-4; Figure 3A) as well as increased total lean mass (F_{1, 27} = 8.74, p = 6.4e-3; Figure 3B).

Within the single week of individual housing, 13 weeks after mild stress onset, there was a main effect of stress exposure on the sum of food consumed per rat (F_{1, 27} = 5.53, p = 0.026) but also a nearly significant interaction between genotype and stress exposure (F_{1, 27} = 3.44, p = 0.075). Between WT rats there were no differences in the total amount of food consumed during the week, but stress-exposed KO rats consumed a significantly larger quantity of food compared to control counterparts (T₁₃ = 3.3, p = 5.53e-3; Figure 3C), supporting the visual difference seen for cage-wide food intake (Figure 2).

Four weeks of UCMS exposure did not alter the patterns we had seen during the earlier mild stress protocol. From start to end of UCMS (weeks 15-19), stress-exposed WT and KO rats gained more weight than control counterparts (F_{1, 28} = 435.1, p = 1.61e-5). They also ate more over time (F4, 48 = 6.80, p = 0.02) than control rats (Figure 2). EchoMRI analysis conducted at the conclusion of UCMS exposure determined that stress exposure significantly increased total fat mass (F_{1, 28} = 5.58, p = 0.025), which was driven by differences in the KO rats (T₁₄ = 2.34, p = 0.035; Figure 4A). Stress exposure also increased total lean mass in both genotypes (F_{1, 27} = 4.34, p = 0.047; data not shown).

At the end of the study, stress-exposed rats had a trend toward increased body weight compared to control rats, for both genotypes (F_{1, 28} = 3.46, p = 0.074; Figure 4B). Similar to the post-UCMS findings, there was a main effect of stress for RetroFat mass (F_{1, 26} = 5.43, p = 0.028), that was driven by the KO rats (T₁₂ = 2.58, p = 0.024; Figure 4C) with no differences in WT rats. Stress had a modest impact on ParaFat (F_{1, 27} = 3.52, p = 0.071) and OmenFat (F_{1, 28} = 3.52, p = 0.071; data not shown). There were no differences in body length, tail length, or organ weight by genotype nor by stress exposure (data not shown). Control KO rats had a trend toward smaller adrenal glands compared to control WT rats (T₁₃ = 2.07, p = 0.059; Figure 4D), although there were no differences in adrenal gland weight between stress-exposed WT and KO rats. While there is a visual increase in adrenal gland weight between control and stress-exposed KO rats, due to large variation this was not statistically significant.

There were no differences in fasting glucose or glucose response to a glucose challenge, neither by genotype nor by stress exposure (data not shown).

There was a significant interaction between genotype and stress exposure regarding the overall movement of the rat during the NSF, as measured by the number of line crossings (F_{1, 28} = 13.22, p = 1.1e-3), where stress-exposed KO rats moved much less than their control counterparts (T₁₄ = 3.35, p = 4.8e-3; Figure 5A), with little difference in activity between WT groups (T₁₄ = 1.78, p = 0.096; Figure 5A). There were similar results in the number of center approaches, with a significant interaction between genotype and stress exposure (F_{1, 26} = 6.23, p = 0.019) where stress-exposed KO rats approached much less frequently than control KO rats (T₁₃ = 3.78, p = 2.3e-3; Figure 5B), with no significant differences between WT rats. When measuring the amount of time spent in the center there was a main effect of stress group, where control rats of both genotypes were more willing to spend time in the center of the box compared to stress-exposed rats (F_{1, 26} = 7.88, p = 9.4e-3; Figure 5C).

Because of ceiling and floor effects, we could not accurately measure statistical differences in latency to feed or the total time a rat spent feeding (Figure 5D). We instead compiled these phenotypes into one that examined participation in the test and used modified survival curves to analyze the likelihood of a rat participating given its genotype and stress exposure. There were no significant differences in WT participation in the test due to stress exposure, but stress-exposed KO rats had a significantly lower likelihood of participating in the NSF than controls (Χ² = 5.3, p = 0.02; Figure 5E).

We evaluated immobility and mobility in the FST on Day 2, plus the change in immobility from the first 5 min of Day 1 to Day 2, as measures of passive coping response to stress. There were main effects by stress exposure for both immobility and mobility on Day 2: as expected, stress exposure increased immobility in the test (F_{1, 27} = 10.21, p = 3.5e-3; Figure 6A) and decreased mobility (F_{1, 27} = 10, p = 3.8e-3; Figure 6B). In addition, there was a strong effect of stress exposure on the change in immobility between the days (F_{1, 26} = 24.04, p = 4.34e-5; Figure 6C), where stress-exposed rats had a smaller change between the days. Importantly, there was a main effect of genotype (F_{1, 26} = 4.25, p = 0.049) driven by differences in the stress-exposed rats (T₁₃ = 3.19, p = 7.2e-3; Figure 6C): KO rats had significantly greater immobility on Day 2 relative to Day 1, while WT rats had a slight decrease in immobility on Day 2 compared to Day 1. These differences support an increased response to FST stress in the KO rats exposed to mild stress relative to WT rats.

After UCMS exposure, we used an OFT to measure locomotor and anxiety-like behavior in the rats. Over the full course of the test, there was a main effect of both genotype (F_{1, 27} = 4.23, p = 0.049; Figure 7A) and stress (F_{1, 27} = 4.23, p = 0.049; Figure 7A), where KO rats spent more time moving than WT rats and stress exposed rats spent more time moving than controls. KO rats also reared significantly more relative to WT rats (F_{1, 28} = 4.49, p = 0.043), though this was primarily in the control rats (T₁₄ = 2.65, p = 0.019; Figure 7B) rather than the stress-exposed rats. KO rats also trended toward travelling a greater distance (F_{1, 27} = 4.09, p = 0.053; Figure 7C) than WT counterparts under both control and stress conditions.

While the two-way ANOVA did not show an interaction between genotype and stress exposure that impacted number of center entries, we previously saw in the NSF that stress affected willingness to approach the center in KO rats but not WT. When we examined each genotype separately in the OFT, we found that there were no differences by stress exposure in WT rats, but stress-exposed KO rats entered the center of the box significantly less frequently than control counterparts (T₁₃ = 3.78, p = 2.3e-3; Figure 7D). As we had previously seen in the NSF, stress exposure affected amount of time spent in the center, where stress-exposed rats of both genotypes spent significantly less time in the center than controls (F_{1, 27} = 4.41, p = 0.045; Figure 7E). Neither genotype nor stress exposure affected distance traveled in the center of the box, however (data not shown).

After 10 weeks of mild stress exposure, rats were administered an acute restraint test. There were no statistically significant differences in basal CORT by genotype nor by stress exposure (Figure 8A), although KO rats visually have a slightly lower basal CORT than WT. As restraint continued, CORT increased in all rats regardless of genotype or stress exposure (F_{2, 40} = 317.23, p = 2.9e-25), both after 10 (Q₂₅ = 20.07, p < 1e-4; Figure 8B) and 30 (Q₂₇ = 20.61, p < 1e-4; Figure 8B) min of restraint. When we examined the effect of genotype and stress exposure separately for each time point, there was a significant interaction after 10 min of restraint (F_{1, 26} = 4.73, p = 0.039) where stress-exposed KO rats had a lower CORT than control counterparts (T₁₃ = 2.29, p = 0.039; Figure 8C), with no differences in WT rats. Although stress-exposed KO rats maintain a visually lower CORT after 30 min of restraint compared to control KO rats, there was no longer a statistically significant difference.

Prior to UCMS, at week 15, we measured serum CORT in the rats and saw no differences by genotype nor by stress exposure (Supplementary Figure S3). This confirmed that the mild stress protocol did not increase basal CORT in WT rats and indicates that the changes in adiposity and behavior in KO rats occurred despite lack of changes in basal CORT.

We also measured CORT collected at euthanasia. Given prior results (Szalanczy et al., 2023), we expected that there would be an effect of euthanasia order on CORT, so we initially examined only the first rat of each cage, presumed to be the basal measurement. There was a significant interaction between genotype and stress exposure (F_{1, 12} = 9.93, p = 0.008), where stress-exposed KO rats had significantly elevated CORT relative to control counterparts (T_4.67 = 2.91, p = 0.036; Figure 9A) yet contrary to expectation stress-exposed WT rats had a trend toward lower CORT relative to control counterparts (T_5.98 = 2.11, p = 0.08). We do not fully understand why basal CORT was so high in control WT rats and these results may be confounded by an unknown stress at the time of sac. Similar to our previous work (Szalanczy et al., 2023), we also found that control KO rats have significantly lower serum CORT relative to control WT rats at the time of euthanasia (T_3.59 = 3.31, p = 0.035; Figure 9A), indicating KO rats have lower basal stress compared to WT rats, supporting the behavioral studies described above. We show serum CORT over time for both rats in the cage and just rat 1 (Supplementary Figures S3A, B). Note that because conditions of serum CORT collection differed at each time-point, time-points may not be directly comparable.

We next assessed the impact of euthanasia order in addition to genotype and stress exposure. Similar to our previous work (Szalanczy et al., 2023), there was a significant effect of order (F_{1, 24} = 21.01, p = 1.2e-4), with rats euthanized second showing higher serum CORT. There was still an interaction between genotype and stress exposure (F_{1, 24} = 8.65, p = 7e-3), but there was also a three-way interaction between genotype, stress exposure, and euthanasia order (F_{1, 24} = 6.6, p = 0.017). We split the data by genotype to confirm the impact of stress exposure and order on CORT, and found a significant main effects of stress (F_{1, 12} = 9.53, p = 9e-3; Figure 9B) and order (F_{1, 12} = 29.72, p = 1.47e-4; Figure 9B) in KO rats, demonstrating that chronic stress exposure increased CORT in the KO rats and that CORT rises in the second rat of the cage regardless of prior stress exposure. In WT rats, we found an interaction between stress exposure and order (F_{1, 12} = 6.28, p = 0.028) where order did not impact CORT in control WT rats but CORT was significantly elevated in the second rat of the cage in stress-exposed WT rats (T_5.25 = 3.91, p = 0.01; Figure 9B).

These findings were supported by additional analyses of the CORT ratio between the first and second rat of the cage for each genotype and stress condition. There was a nearly significant interaction between genotype and stress exposure (F_{1, 11} = 4.21, p = 0.065) where WT control rats had a much lower ratio relative to the other three groups (Figure 9C). These data indicate a stronger social stress response to removal of cage-mate in KO rats, which is seen only in WT rats when they have previously been exposed to chronic stress.

Given that there was a significant effect of euthanasia order on CORT, we initially examined ACTH in only the first rats of the cage. There were main effects of genotype (F_{1, 12} = 5.83, p = 0.033) and stress exposure (F_{1, 12} = 5.94, p = 0.031) where control WT rats had greater basal ACTH than control KO rats (T_4.98 = 3.02, p = 0.03; Figure 9D) and stress-exposed WT rats (T_5.99 = 2.46, p = 0.049; Figure 9D).

When we incorporated data from the second rats and the factor of euthanasia order into the analyses, there was an interaction between stress exposure and order (F_{1, 24} = 5.76, p = 0.025) separate from genotype. Specifically, in control rats there were no differences in ACTH between the two rats of the cage, whereas in stress-exposed rats the second rat of the cage had elevated ACTH compared to the first rat (T_{1, 14} = 8.71, p = 0.011; Figure 9E).

We investigated Nr3c1 expression in the pituitary, colon, and liver. We looked in the pituitary and colon because both tissues have high Krtcap3 expression (Szalanczy et al., 2023) and each are involved in stress response (Herman et al., 2016; Wiley et al., 2016), indicating that they may be the tissues of action for Krtcap3. Further, we also sought to verify previously reported changes in Nr3c1 expression in the liver in response to stress (Szalanczy et al., 2023). We found a very nearly significant main effect of stress-exposure on Nr3c1 expression in the pituitary (F_{1, 27} = 4.15, p = 0.052), where stress-exposed KO rats had greater Nr3c1 expression in the pituitary than control counterparts (T₁₄ = 2.33, p = 0.035; Figure 10A) with no differences in expression between WT rats. Similarly, there was a near interaction in colon Nr3c1 expression (F_{1, 25} = 4.2, p = 0.051) where there was no difference by stress exposure in WT rats, but stress exposure resulted in a trend toward decreased Nr3c1 expression in KO rats (T₁₃ = 2.06, p = 0.06; Figure 10B). In the control condition, KO rats also had significantly higher colon Nr3c1 expression than WT rats (T₁₃ = 2.2, p = 0.047; Figure 10B). These findings mimic the metabolic and behavioral findings, where differences are seen between control and stress-exposed KO rats but not between stress conditions in WT rats, further supporting an important role for Krtcap3 in the stress axis. As expected, Nr3c1 expression was also significantly increased in the liver tissue of stress-exposed rats (F_{1, 27} = 4.51, p = 0.043), but the difference here was driven by the WT rats (T₁₃ = 3.11, p = 8.2e-3; Figure 10C) with no changes in the KO rats. Similar to our findings in the colon, we also identified a difference in liver Nr3c1 expression between control WT and KO rats, where KO rats had significantly higher expression (T₁₃ = 2.26, p = 0.041; Figure 10C). On the other hand, there were no differences in pituitary Pomc expression or colon Hsd11β2 by genotype or by stress exposure (data not shown).

We measured Krtcap3 expression between control and stress-exposed WT rats in key tissues to evaluate if stress exposure altered expression. Ultimately, there were no differences in Krtcap3 expression between control and stress-exposed WT rats in the pituitary, the adrenal, or the colon (data not shown).

As opposed to the suspected decrease in food intake and adiposity in response to construction noise in WT rats from our previous study (Szalanczy et al., 2023), the chronic stress protocol here increased body weight over time only in the KO rats, with no effect in WT rats. Although stress exposure increased adiposity in both WT and KO rats early in the study, by the end the difference was only maintained in KO rats. Our findings here are consistent with the hypothesis that stress is necessary to increase eating and adiposity in rats with low Krtcap3 expression. The increased adiposity in response to stress in the KO rats from the current study more closely aligns with changes seen following psychosocial stress rather than physical stress (Patterson and Abizaid, 2013) and may result from the number of times cage-mates were separated during the study. In support of this, when all rats were socially isolated, the stress-exposed KO rats ate significantly more food compared to control KO rats. We posit that prior stress-exposure increased susceptibility to the psychosocial stress of social isolation in KO rats, who may have increased food intake to ameliorate the anxiety-inducing effects of the separation (Maniam and Morris, 2010; Kistenmacher et al., 2018; Hyldelund et al., 2022), while WT rats were unaffected. Overall, we show that KO rats responded differently to stress than WT rats and that these differences influenced adiposity. Fully elucidating the difference in response between WT and KO rats to different stressors and understanding the impact on adiposity is a key goal of future studies.

As we saw larger adiposity differences between control and stress-exposed KO rats, so too did we see larger behavioral changes in KO rats compared to WT, with greater increases in anxiety following stress exposure. Importantly, control KO rats exhibited lower anxiety measures in the OFT compared to WT rats, supportive of a lower basal stress state. In response to chronic stress, KO rats showed the expected increases in thigmotaxis in the OFT as well as hyponeophagia and significant decreases in exploratory activity in the NSF. WT rats, on the other hand, displayed only minimal differences in these measures after stress exposure, in contrast to findings after chronic restraint stress in WKYs (Jung et al., 2020). We also found that immobility in the FST increased in stress-exposed KO rats from Day 1 to Day 2 of the test, with minimal changes between days in stress-exposed WT rats. This suggests a greater passive coping response (Commons et al., 2017; Molendijk and de Kloet, 2019) or greater depressive-like behavior (Redei et al., 2022) in KO rats following early life stress. As the WKY rat is hyper-sensitive to stress (Redei et al., 2022), these findings support that WT rats may have been unable to distinguish between the stress of the study design (control rats) and the chronic stress procedures (stress-exposed rats): it is possible that both groups of rats were chronically stressed. The fact that KO rats show the expected changes in emotional behaviors in response to chronic stress suggests that low Krtcap3 expression may normalize the stress response in WKY rats, although more work is needed to test this hypothesis. If low Krtcap3 expression normalizes the stress response in WKY rats, then this may also explain why only KO rats consumed more HFD when stressed, as occurred during social isolation. Taken together, findings in the NSF, OFT, FST, and food intake under social isolation indicate that stress exposure leads to the expected increases in anxiety-like and depression-like behaviors when Krtcap3 expression is low.

Differences in plasma CORT and ACTH between WT and KO rats at euthanasia indicate that Krtcap3 expression likely plays a role in HPA axis function. As before (Szalanczy et al., 2023), we found that control KO rats had lower basal plasma CORT compared to control WT rats, supported by lower basal ACTH and slightly smaller adrenal glands. At the end of the study, stress-exposed KO rats had greater CORT than control counterparts with no differences in basal ACTH and a mild increase in adrenal gland weight. These data support that KO rats may have a more normal HPA axis response to chronic stress compared to WKY rats: increased basal CORT following stress may in turn lead to increased passive coping and anxiety-like behaviors, including increased food intake. Contrary to our initial expectations, stress-exposed WT rats did not have greater CORT than control WT rats and in fact had lower basal ACTH than control WT rats, which was surprising as prolonged stress is known to increase ACTH in WKYs (Pardon et al., 2003; Malkesman et al., 2006). Furthermore, there were no differences in adrenal gland weight between control and stress-exposed WT rats. These findings potentially support the hypothesis that control WT were chronically stressed due to the study design, although we are unable to distinguish between this hypothesis and the possibility that the early noise stress had a protective effect on later UCMS stress in these rats.

We also confirmed here that KO rats of either stress condition experienced a spike in CORT when their cage-mate is removed at euthanasia (Szalanczy et al., 2023), supporting that KO rats are more sensitive to psychosocial stress. Interestingly, we identified this same pattern in stress-exposed WT rats, but not control WT rats. It is possible that prior stress exposure sensitized the HPA axis to a novel acute stress (Franco et al., 2016) in WT rats, although the adiposity and behavioral data suggest this is not the case.

It is unclear why control WT rats had such a high basal CORT at euthanasia compared to the other groups. One explanation for this finding is that control WT rats had not fully adjusted to the room change for euthanasia, and the room change stress is responsible for the differences in basal CORT between control WT and KO rats at euthanasia. A second possibility is that the large change in housing environment to a different building coupled with a behavioral test in the final weeks of the study may have more greatly affected control WT rats than KO rats. This also supports that WT and KO rats respond to different stressors differently, as supported by work in other models (Singh et al., 1999; Kogler et al., 2015). Future studies to evaluate the neurocircuitry between WT and KO rats, before and after stress, are warranted to fully understand these differences in response (Bangasser et al., 2013).

The pattern of expression for the GR Nr3c1 may provide insight into the differences seen between WT and KO rats in the current study. We measured Nr3c1 expression in the pituitary, colon, and liver of WT and KO rats; while we only explored changes in gene expression here, the connection between Krtcap3 and Nr3c1 is also supported by evidence of protein-protein interactions elsewhere (Lievens et al., 2016).

The pituitary has high Krtcap3 expression in female WKY rats and is key to the HPA axis. The exact regulation of the GR under chronic stress conditions is unclear. Foundational in vitro work demonstrated that GR expression decreases following GC exposure in a mouse corticotrope or pituitary cell line (Sheppard et al., 1991; Williams et al., 1991) which aligns with findings in humans that early life stress is associated with methylation of NR3C1 in saliva or blood, leading to decreased gene expression and dysregulated HPA axis function (van der Knaap et al., 2015; Alexander et al., 2018; Holmes et al., 2019; Chatzittofis et al., 2021; Lewis et al., 2021; Chubar et al., 2023). However, in vivo work where animals are subjected to up to a week of stress has shown that expression of the GR increases in the pituitary (Sheppard et al., 1990; Nishimura et al., 2004; Noguchi et al., 2010). Similarly, in the current study we found that pituitary Nr3c1 expression increased when KO rats were exposed to stress, with no changes in WT rats. These findings further support that low Krtcap3 expression may “normalize” the HPA axis of the WKY rat. A change in pituitary Nr3c1 expression after stress exposure may be related to negative feedback regulation pathway of the HPA axis (Gjerstad et al., 2018) and could clarify the role of Krtcap3.

Krtcap3 expression is also high along the gastrointestinal tract of female WKY rats and can also be connected to the stress response (Wiley et al., 2016). Other work has shown that chronic stress decreases colonic Nr3c1 (Zheng et al., 2017; Muir et al., 2023), similar to our findings in the KO rats, lending further credence to the hypothesis that low Krtcap3 expression “normalizes” the HPA axis of WKY rats. In the colon we found that control KO rats have greater Nr3c1 expression than control WT rats and that stress exposure decreased Nr3c1 expression in KO rats to levels comparable to the WT rats. These data may indicate that control KO rats have improved intestinal integrity over WT rats due to the comparatively high Nr3c1 expression, but upon exposure to chronic stress and a HFD, Nr3c1 expression may decrease to minimize damage (Aranda et al., 2019; Shukla et al., 2022). The lack of change in Nr3c1 expression in WT rats after stress exposure supports that WT and KO rats responded to the study design differently. Krtcap3 may be acting in either the pituitary, the gastrointestinal tract, or both to influence GR expression and the effect of glucocorticoids on metabolism and behavior.

In the current study, we found that chronic stress increased liver Nr3c1 expression only in the WT rats, where previous work demonstrated increases in both WT and KO rats (Szalanczy et al., 2023). This highlights that the stress administered in this study did not recapitulate the stress in our previous work and that the role of Krtcap3 in stress response, and its relationship with Nr3c1 expression, is tissue-dependent (Costello et al., 2022). There remains a need to better understand how different types of stress regulate expression of the GR in different tissues and to identify downstream effects on health.

We did not see changes in Krtcap3 expression in the pituitary, adrenal, or colon between control and stress-exposed rats. Given that the control and stress-exposed WT rats had similar responses throughout the study, this finding is not surprising. Future work will therefore be needed to determine if stress alters Krtcap3 expression among these tissues.