Maternal Repetitive Hypoxia Prior to Mating Confers Epigenetic Resilience to Memory Impairment in Male Progeny

Abstract

We showed previously in a mouse model of vascular cognitive impairment and dementia involving chronic cerebral hypoperfusion (CCH) that repetitive hypoxic conditioning (RHC) of both parents results in the epigenetic, intergenerational transmission of resilience to recognition memory loss in adult progeny, as assessed by the novel object recognition test. The present study was undertaken in the same model to determine whether RHC treatment of one or both parents is required to confer dementia resilience intergenerationally. We found inherited resilience to 3 months of CCH in males is maternally mediated (p = .006). Statistically, we observed a strong trend for the paternal germline to contribute as well (p = .052). We also found that, in contrast to what is widely observed in males, females display intact recognition memory (p = .001) after 3 months of CCH, revealing a heretofore unidentified sexual dimorphism with respect to cognitive impact during disease progression. Overall, results of our study strongly implicate epigenetic changes in maternal germ cells, induced by our repetitive systemic hypoxic stimulus, contributing to a modified differentiation program capable of establishing a dementia-resilient phenotype in adult male first-generation progeny.

Epigenetic modifications to the genome, including imprinting, secondary to DNA methylation, posttranslational histone modifications, and the action of noncoding RNAs, alter gene expression and thus cellular or organismal phenotype. These epigenetically mediated phenotypes can be short- or long-lasting, depending on the nature of the stimulus. Some are also known to pass from parents to offspring in animals, in a phenomenon called intergenerational inheritance; the continued passage of phenotype from those offspring to subsequent generations is termed transgenerational inheritance (Perez & Lehner, 2019; Tuscher & Day, 2019; van Steenwyk et al., 2018). Findings from a number of studies indicate that epigenetic modifications in either the paternal or maternal germline are sufficient to reestablish the parental phenotype in their progeny; in some cases, such heritability also exhibits a sexual dimorphism, with only males or females inheriting the phenotype (Chan et al., 2018; Lempradl, 2020; Ratnu et al., 2017; Sharma, 2019; Wang et al., 2017). The vast majority of the heritable phenotypes in these documented cases are negative/maladaptive, induced as a result of some traumatic or adverse stress in parents and then manifested as disease, or disease vulnerability, in offspring (Jawaid et al., 2021; Yehuda & Lehrner, 2018).

We recently showed that positive/adaptive central nervous system phenotypes induced in somatic cells of parent (F0) mice by repetitive hypoxic conditioning (RHC) can also be transmitted intergenerationally, as a result of the additional epigenetic changes, this stimulus induces in their germ cells (Belmonte et al., 2022; Harman et al., 2020). Specifically, using a well-established mouse model of vascular cognitive impairment and dementia (VCID) secondary to 3 months of chronic cerebral hypoperfusion (CCH), we showed that adaptive epigenetic responses induced by RHC were recapitulated in the next generation. The responses were manifested functionally in hippocampal slices as a retention in long-term potentiation (LTP), an electrophysiological substrate of memory formation and synaptic plasticity, and in conscious mice as a resilience to recognition memory impairment (Belmonte et al., 2022). Herein, we undertook studies using the latter neurocognitive endpoint to determine whether this dementia resilience in first-generation (F1) mice results from RHC-induced changes in the germline cells of F0 males, F0 females, or both.

The aforementioned mouse model of VCID we employed herein and previously to document epigenetically mediated protection against CCH-induced dementia is well-established, but has relied almost exclusively on the use of male mice to characterize the underlying pathology and to assess the efficacy of potential therapeutics (Ben-Ari et al., 2019; Bhatia et al., 2022; Cao et al., 2022; Dominguez et al., 2018; Inaba et al., 2019; Khan et al., 2015; Ohtomo et al., 2020; Patel et al., 2017; Toyama et al., 2018; Tsai et al., 2015; Zheng et al., 2022). Thus, a secondary aim was to determine if impairments in recognition memory secondary to CCH, independent of treatment, exhibited a sexual dimorphism.

Method

We report herein our results of two distinct studies related to VCID protection in mouse model that used bilateral carotid artery stenosis (BCAS) with microcoils (Duncombe et al., 2017; Shibata et al., 2007) to establish a 3-month period of CCH. Nine-week-old male and female (25 each) C57BL/6J mice (Jackson Laboratory, Bar Harbor, ME, United States) were habituated to our vivarium for 1 week before being randomly assigned into experimental breeding groups that provided the F1 mice used in our two studies. All mice were housed under a 12-hr/12-hr light/dark cycle with food and water provided ad libitum. All surgical procedures and neurobehavioral tests were approved by our Institutional Animal Care and Use Committee and were performed in accordance with the APA Guidelines for Ethical Conduct in the Care and Use of Nonhuman Animals in Research and the NIH Guidelines for the Care and Use of Laboratory Animals.

F0 mice were treated with twenty-four 1-hr exposures of mild-to-moderate systemic hypoxia (RHC) every other day for 8 weeks, as described previously (Belmonte et al., 2022). Untreated controls consisted of age- and sex-matched cages of mice exposed to normal atmospheric oxygen (room air) of equivalent frequency and duration. Breeding pairs of F0 mice were established 5 days after the 24th hypoxia treatment to generate the following four F1 offspring groups: Those in which both mother and father received RHC, only the mother received RHC (the father served as a normoxic control), only the father received RHC (the mother served as a normoxic control), and both mother and father served as normoxic controls. Mice in each of the aforementioned F1 groups were derived from a minimum of four (and maximum of 9) distinct breeding pairs.

At 12–14 weeks of age, male and female F1 offspring born to normoxic control parents were randomized to receive either BCAS surgery or sham-BCAS surgery, to create untreated disease and untreated control (CTL) experimental groups, respectively. Three age-matched BCAS/disease mouse groups were established; specifically, they included mice born to RHC-treated mothers and fathers (Both-RHC), to RHC-treated mothers (Mo-RHC), and to RHC-treated fathers (Fa-RHC). None of the mice from these latter three groups were designated for sham-BCAS surgery because we previously confirmed in both sexes that, relative to CTL mice, those mice treated with RHC (without BCAS) exhibited normal corpus callosum myelination, hippocampal LTP responses, and recognition memory (Belmonte et al., 2022).

BCAS was performed aseptically, as described previously (Belmonte et al., 2022). In brief, mice were anesthetized with ketamine (87 mg/kg, i.p.) and xylazine (13 mg/kg, i.p.), and sustained-release buprenorphine (0.1 mg/kg) was administered subcutaneously, prior to the aseptic surgical procedure. Following a ventral midline incision on the neck, both carotid arteries were isolated from their vagus nerves, and gold-plated microcoils (0.18-mm internal diameter; 0.50-mm pitch; 2.5-mm total length; Motion Dynamics Corporation, Fruitport, MI, United States) were wrapped around them using a piece of 10-0 suture looped under the carotid to guide the artery around the coil. The incision was then closed with Vetbond tissue adhesive (3M, Saint Paul, MN, United States). Sham-BCAS mice were subjected to the same procedures, but without wrapping the coils around the arteries. For postoperative recovery, the mice were placed into individual new cages, in close proximity to a warming fan, and monitored until ambulatory. They were ultimately single-housed to prevent intracage fighting postsurvival surgery and provided Institutional Animal Care and Use Committee-required supplemental nesting.

After 3 months of CCH or normal cerebral perfusion, recognition memory was assessed in all F1 groups by the novel object recognition (NOR) test (Lueptow, 2017), as described previously (Belmonte et al., 2022). In brief, on the day before the test, mice were individually habituated for 15 min to an empty, dimly lit, black Plexiglass box with a white floor (20 × 20 × 20 cm). The next day, a testing session consisting of two trials was initiated. In the first trial (the familiar object exposure trial), mice were individually placed in the same Plexiglass box, now containing two identical objects—either two small glass bottles or two multicolored, 57-mL 5-hr Energy containers. Both objects were previously validated to be equally preferred by mice. Mice were allowed to explore the objects for 10 min before being returned to their home cages. Three hours later, the second trial (the novel object exposure trial) was initiated by returning the mice to the box that now contained one of the objects from the first trial, along with a novel object. Again, mice were allowed to freely explore the objects for 10 min. Which objects were used as familiar objects in the first trial, and which object was substituted in for the second trial as the novel object, was randomized among all of the mice tested. Ethanol (70%) was used to wipe down the entire box between each trial, including habituation, to eliminate olfactory cues.

Mouse behavior during the second trial was recorded by an overhead-mounted video camera. “Exploration” was quantified offline as the time spent sniffing or touching either object at a distance of <2 cm from the object. Time spent biting, climbing, and sitting on either object was not factored into this metric. Analysis of NOR behavior was conducted by video playback by an observer blinded to experimental conditions. Recognition memory was quantified by determining differences in total time spent exploring each object.

F0 and F1 animals were sacrificed by CO₂ inhalation after breeding or after NOR testing, respectively.

A priori data exclusion criteria employed included the following: (a) wounds experienced as a result of intracage fighting prior to CCH (the only period during which they were group-housed; n = 3); (b) concerns related to the success of the BCAS surgery and coil placement (n = 8); (c) a total exploration time in the second NOR trial that did not exceed 18 s (n = 54; Leger et al., 2013); or (d) urination during the second NOR trial, which was taken as an anxiety indicator and potential confounder of exploratory behaviors (n = 13). GraphPad Prism V9 (GraphPad Software, San Diego, United States) was used to analyze NOR data, and Grubbs’ test (GraphPad Outlier Calculator) was used to identify outliers. Differences in total exploration times of the familiar and novel objects were compared using the nonparametric Wilcoxon signed-rank test.

Results

Results of Study 1 are shown in Figure 1. We found that male F1 progeny from the CTL and Both-RHC groups exhibited significantly greater exploration times of the novel object, indicative of normal, unimpaired recognition memory. This was the primary expected outcome for the uninjured CTL group and, based on our previous study (Belmonte et al., 2022), for the Both-RHC group as well, confirming our finding that treating both parents with RHC prior to mating prevented deficits in recognition memory resulting from 3 months of BCAS-induced CCH. We now show that mice derived from matings in which only the mothers received RHC (Mo-RHC) exhibited significantly greater exploration times (p = .006) for the novel object after 3 months of CCH, whereas exploration times for the familiar and novel object in mice derived from matings in which only the fathers received RHC (Fa-RHC) were not significantly different, indicative of impaired recognition memory. However, statistically speaking, the difference was almost significant (p = .052), suggesting RHC-induced epigenetic changes in the paternal germline may contribute to the dementia-resilient phenotype in adult progeny as well.

Figure 1. The Exploration Times Scores for the Novel Object Recognition (NOR) Test of Recognition Memory in Male Mice, by Disease and Parental Treatment.

Note. Graph shows the exploration times of familiar and novel objects in adult F1 control (CTL) mice born to untreated F0 control parents and in four groups of adult F1 experimental mice derived from untreated F0 control parents (No-RHC), RHC-treated F0 fathers and mothers (Both-RHC), father-only RHC-treated F0 parents (Fa-RHC), and mother-only RHC-treated F0 parents, following 3 months of BCAS-induced chronic cerebral hypoperfusion (CCH). Significant differences in exploration time were found in the CTL, Both-RHC, and Mo-RHC groups (p = .004, p = .020, and p = .006), respectively. No differences in exploration time were found in the No-RHC and Fa-RHC groups (p = .129 and p = .052), respectively. All data are shown as mean ± SEM. A p value < .05 was considered significant. Sample size (“n,” representing the number of mice) is shown within each column. RHC = repetitive hypoxic conditioning; SEM = standard error of the mean.

* p < .05. ** p < .01.

Figure 2 shows the results of Study 2. Although male mice with 3 months of BCAS-induced CCH exhibited impairments in recognition memory, based on NOR testing, we found that, under identical experimental conditions, female mice showed no such deficits. Specifically, 3 months post-BCAS, female mice spent significantly more time exploring the novel object than the familiar object, as did both sexes of non-BCAS, normal controls. Of note, all of the mice in this study were “F1” mice with respect to being derived from in-house matings of untreated/control parents in conjunction with Study 1 above, but because neither of their F0 parents were treated with RHC, these findings are relevant to the VCID model used to generate them and hold no generation-specific dependencies.

Figure 2. The Exploration Times Scores for the Novel Object Recognition (NOR) Test of Recognition Memory, by Disease and Sex.

Note. Graph shows the exploration times of familiar and novel objects for control and BCAS-induced CCH mice of each sex. Significant differences in exploration times were found in the male CTL, female CTL, and female BCAS groups (p = .004, p = .023, and p = .001), respectively. No differences in exploration time were found in the male BCAS group (p = .129). All data are shown as mean ± SEM. A p value < .05 was considered significant. Sample size (“n,” representing the number of mice) is shown within each column. CTL = control; BCAS = bilateral carotid artery stenosis; SEM = standard error of the mean.

* p < .05. ** p < .01.

Discussion

Overall, results of our present studies provide novel insights into the parental transmission of an epigenetically induced resilience to dementia. Specifically, we show that maternal RHC treatment leads to a robust neurocognitive protective effect against memory impairment in their F1 progeny; that said, our findings also reveal some contribution to this phenotype from paternal RHC treatment. Secondarily, we identified a dimorphism by sex inherent in the pathology that defines this dementia model: Young adult female mice exhibit natural neuroprotection after 3 months of CCH, reflecting a unique, sex-dependent VCID clinical phenotype that warrants female-inclusive studies of both pathology and treatment.

To our knowledge, the adaptive phenotypes induced in mice by paternally mediated intergenerational, epigenetic modifications have been characterized by improvements in baseline metrics and not by responses that protect against disease. Specifically, paternal exercise enhances baseline cognitive abilities of offspring, as measured by a variation of the NOR test (McGreevy et al., 2019). Similarly, offspring born to fathers exposed to environmental enrichment exhibit enhanced cognition, as documented by the Morris water maze test of spatial reference memory and through recordings of LTP (Benito et al., 2018). However, neither of these studies concomitantly tested the same offspring outcome metrics following exercise or environmental enrichment of the mother. On the other hand, there are precedents for maternally mediated, intergenerationally inherited adaptive responses which do protect against disease or reverse mutant phenotypes, more closely resembling the effect of maternal RHC treatment found herein. In particular, maternal exercise prevented the development of detrimental metabolic phenotypes in adult progeny (Laker et al., 2014)—although whether a similar heritability results from paternal exercise were not examined. Similarly, maternal environmental enrichment restored LTP in mutant mice known to display impaired LTP responses at rest (Arai et al., 2009), whereas paternal environmental enrichment had no effect on this progeny phenotype. Altogether, these findings suggest stimuli experienced by both the father and mother can epigenetically modify their respective germline cells and result in offspring inheritance of an adaptive, and possibly protective, phenotype, but, as in our study, epigenetic modifications of the maternal germline may ultimately confer a greater likelihood of the given phenotype manifesting in F1 adult progeny. Accumulating evidence of sperm epigenetic heterogeneity (Laurentino et al., 2016) may also contribute to the interanimal variability we observed in phenotypic expression in the adult progeny of mice derived from father-only RHC-treated mating pairs.

Results of our second study, showing that adult female mice exhibit no significant loss of recognition memory after 3 months of CCH—as commonly reported for males (Ben-Ari et al., 2019; Bhatia et al., 2022; Cao et al., 2022; Dominguez et al., 2018; Inaba et al., 2019; Khan et al., 2015; Ohtomo et al., 2020; Patel et al., 2017; Toyama et al., 2018; Tsai et al., 2015; Zheng et al., 2022)—reveal a previously undescribed sexual dimorphism in this VCID model. This dimorphism is consistent with the finding that 4-month-old male mice treated with 1 month of oral estradiol post-BCAS are protected against BCAS-induced memory impairment in the NOR test (Dominguez et al., 2018). Along similar lines, 4-month-old adult female mice exhibit unimpaired recognition memory 4–6 weeks after BCAS but exhibited a strong trend toward impairment at 22 months of age (Wolf et al., 2017); no male mice were included in this study to potentially reveal the sexual dimorphism we identified in our 6-month-old mice at 3 months post-BCAS. Taken together, these findings begin to establish preclinical, mechanistic support for human epidemiological data implicating a relative advantage premenopausal women have over men in VCID (Conde et al., 2021; Gannon et al., 2019). Similar findings apply to Alzheimer’s (Merlo et al., 2017), with disease risk increasing postmenopause (Uddin et al., 2020). Future preclinical studies are warranted in young and aged mice of both sexes to interrogate the molecular mechanisms underlying what appears to be an age-dependent sexual dimorphism.

Our study was limited in several ways. Apparent maternal germline contributions to offspring phenotype can be confounded by several factors (Bohacek & Mansuy, 2017) that we did not experimentally refute, including the effects of mother-derived mitochondrial DNA (Stewart & Larsson, 2014), the intrauterine environment (Gaillard, 2015), and fostering behaviors. The first two caveats can be addressed using in vitro fertilization (Huypens et al., 2016), and cross-fostering can control for the latter. That said, it is difficult to envision how some supernormal level of maternal nurturing by RHC-treated mothers could lead to disease resilience in their offspring. It is also possible that the social isolation we imposed on all our mice following BCAS surgery as a prophylaxis against fighting could have resulted in interanimal variability in NOR-based assessments of recognition memory secondary to long-lasting changes in mood (depression, anxiety) and learning abilities (Mumtaz et al., 2018). This, as well as variability resulting from differential reductions in the magnitude and duration of CCH secondary to using 0.18-mm diameter microcoils for BCAS (Nishio et al., 2010; Zhou et al., 2022), plus germ cell-to-germ cell variations in epigenetic modifications induced by RHC, may have contributed to the variability in our outcome measurement that rendered advancing definitive conclusions about germline contributions to phenotype challenging. Going forward, studies with larger group sizes, and the inclusion of other measures of cognitive impairment—such as the Morris Water Maze, the 8-arm radial maze, the radial arm water maze, and so forth—would be of keen interest to more definitively differentiate maternally from paternally driven intergenerational epigenetic effects with respect to not only memory-associated outcomes but other neurobehavioral metrics such as anxiety, arousal, and fear, and potential neuromotor endpoints as well. Last, the translational relevance of future preclinical studies examining treatments for VCID, including RHC in mice, should routinely include both sexes, at longer durations of CCH (when females are also likely to exhibit cognitive impairment), and ideally should be conducted in older animals to more closely approximate human VCID.

In sum, we document for the first time in a mouse model of VCID that a beneficial, dementia-resilient phenotype can be induced epigenetically in the adult progeny of mothers treated with repetitive hypoxia prior to conception. Future investigations are needed to confirm our observed trend that the same protective effect on recognition memory may also be realized by exposing only fathers to this epigenetic stimulus. Using the same NOR neurocognitive assessment test of recognition memory, we also show that, after 3 months of CCH, female mice do not exhibit impairments in recognition memory like their male counterparts, unveiling a heretofore unrecognized sexually dimorphic disease phenotype whose clinical implications warrant additional study.