Prenatal Exposure to Diethylhexyl Phthalate Impairs the Recovery of Spatial Memory Post-traumatic Brain Injury in a Sex-speci c Manner

Min Kyoung Sun University of Georgia Amrita Kaimal University of Georgia Charles-Francois V Latchoumane University of Georgia Rameen Forghani University of Georgia Christopher E Lenear University of Georgia Phillip V Holmes University of Georgia Sheba M.J Mohankumar University of Georgia College of Veterinary Medicine Lohitash Karumbaiah (  lohitash@uga.edu ) University of Georgia https://orcid.org/0000-0001-7969-417X


Introduction
Motor and cognitive functions progressively decline after severe traumatic brain injury (sTBI) causing chronic disability 1,2 . Only 26% of patients who sustained moderate-to-severe TBIs report improved outcomes 5 years after the injury due to the complex cascade of pathophysiological events and comorbidities 3 . Neuroendocrine dysfunction is a common comorbidity among TBI patients as nearly two-thirds of the patients experience such a disorder 4 . Furthermore, neuroendocrine dysfunction could further contribute to stressinduced memory and executive function impairments 5 .
Di-2-ethylhexyl phthalate (DEHP) is an endocrine disrupting chemical (EDC) that has transgenerational effects on neuroendocrine function 6-8 . Since human exposure to phthalates is ubiquitous 9,10 , it is important to investigate if prenatal exposure of DEHP could contribute to neuroendocrine dysfunction, such as interference with the hypothalamic-pituitary-adrenal (HPA) axis, leading to impaired recovery from sTBI.
The HPA axis or the stress axis is activated in response to injury, in ammation and infection 11 . Activation of the HPA axis is essential for maintaining homeostasis and for timely recovery from TBI 12 . The HPA axis is made up of corticotrophin releasing hormone neurons located in the hypothalamus, corticotrophs in the anterior pituitary and the adrenal cortex that secretes glucocorticoids (corticosterone (CORT) in rodents and cortisol in humans) 13 . Glucocorticoids act through a negative feedback mechanism inhibiting further HPA axis activation 11 . The hippocampus contains glucocorticoid receptors (GR) and plays a critical role in this feedback inhibition by binding glucocorticoids 14 and directly in uencing neuronal excitability and restructuring neural connectivity 15 . In addition, uninhibited elevation in glucocorticoid levels impairs memory tasks associated with the hippocampus 16 .
Glucocorticoid levels post-sTBI can be elevated or decreased depending on the injury type, severity, time of the injury, and sex; however, there is a consensus that an inverted U-shaped relationship between functional performance and glucocorticoid secretion exists. Neither hyper-secretion nor hypo-secretion of glucocorticoids is bene cial for learning and memory 17 . Hence, stress axis disruption post-TBI can have a direct impact on learning and memory dysfunction.
Recent data from our laboratory suggests that prenatal exposure to 7.5mg/kg BW of DEHP can increase CORT levels in adult female offspring, but not male offspring (Mohankumar et al., Athens, 2021. Unpublished). These offspring also exhibit certain behavioral abnormalities. However, the impact of sTBI on animals prenatally exposed to DEHP has never been studied. Considering the evidence presented above, we hypothesized that prenatal exposure to DEHP could negatively impact spatial memory recovery post-TBI in a sex-speci c manner. To test this, rats prenatally exposed to DEHP were subjected to controlled cortical impact (CCI), and their spatial memory was tested for 4 weeks (Fig. 1). At the end of the observation period, CORT levels were measured in the serum, extent of tissue injury in the brain in terms of neuronal loss was determined by histological assessment, and neurotransmitter levels were measured in speci c brain areas.

Methods
Animals. All animal work and procedures were approved by the University of Georgia Institutional Animal Care and Use Committee (IACUC, A3437-01) and complied with the Public Health Service (PHS) Policy on Humane Care and Use of Laboratory Animals. All experiments were carried out in compliance with the ARRIVE guidelines. 7 pregnant Sprague Dawley dams (F0) were randomly assigned to be treated with either 10 µl of saline (control) or 10 µl of saline containing 7.5mg/kg bis(2-ethylhexyl) phthalate (DEHP) by oral gavage from days 6-21 of gestation. A total of 20 Sprague-Dawley rats (F1) offspring were obtained: 6 control males, 6 control females, 4 DEHP-exposed males, and 4 DEHP-exposed females. F1 rats were housed with their group mates. They were subjected to controlled cortical impact (CCI) at 6 months of age, and single housed post-CCI. All animal handlers were blinded to assigned groups post-CCI.
Controlled Cortical Impact Injury. A CCI injury to the motor cortex was induced using a custom-built pneumatic impactor that has been characterized by us previously 18 . Brie y, animals was anesthetized with 2-3% iso urane, and received a craniotomy to expose the left hind limb motor cortex region in the right hemisphere (bregma 0, 1.5mm lateral from the midline as the center of the impact) 19 . A 3mm diameter tip a xed to a pneumatically actuated piston was driven at an average velocity of 2.17m/s and an average cortical dwell time of 250 msec to create an injury to a depth of 2mm. The craniotomy was covered with 0.5% sterile seakem agarose (Lonza, NH) after cleaning out excess blood and debris surrounding the lesion. The incision was closed using nylon ber sutures (ACE, MA), and the animals were returned to their home cage for recovery. All animals were guillotined at study endpoint, and brain tissue and blood collected for further analyses. All females received CCIs and were sacri ced at diestrus. Behavioral Tests. The Morris Water Maze (MWM) task was used to measure spatial learning and memory following sTBI 20 . Brie y, training was performed in 3 phases each day: 1) in clear water with exposed platform (motor control), 2) repeated platform search in milk-opaci ed water and 3) free exploration without platform. All experiments were performed in a round pool (183 cm in diameter and 91 cm deep) that was lled with clear water (phase 1) or opaque water (containing powdered non-fat milk; phase 2 and 3). A square platform (10 cm x 10 cm) was placed in the south corner of the pool either visible (2 cm above water; phase 1), submerged 2 cm below water level (phase 2) or completely removed (phase 3). Visual cues were permanently a xed to walls on three sides of the pool. The platform was left exposed during all trials for the rst 2 days of training. Subsequently, the rat would perform one trial in phase 1 conditions, three in phase 2, and one in phase 3 per experiment day. The rat was released from either the north, east, or west corners in a random order. To evaluate spatial memory, the platform was removed, and the rat was released from the north corner facing the wall of the pool. Once released, the rat was allowed 2 minutes to search for the platform location.
Test sessions were recorded, and movement was tracked o ine using a custom Bonsai work ow 21 . Time spent in each quadrant was calculated using MATLAB® (Mathworks, inc.). At the end of the observation period, animals were sacri ced by rapid decapitation. Females were subjected to vaginal cytology and sacri ced when they were in the state of diestrus. Serum was separated from trunk blood and used for CORT radioimmunoassay. The brain was quickly harvested, frozen on dry ice and stored at -80ºC until further analysis.
CORT Measurement. Serum CORT was measured in duplicate using a double antibody radioimmunoassay as described previously 22 .
Brain sectioning and microdissection. A cryostat (Slee, London, UK) maintained at -10°C was used to section brain tissue at 300 µm thickness. Following this, the PVN, right (ipsilesional) cortex, contralesional cortex, ipsilesional hippocampus, and contralesional hippocampus were microdissected on a cold stage using the Palkovits' microdissection procedure with a stereotaxic brain atlas as a reference 23 . Microdissections were completed using a 500 µm diameter punch, and the brain punches were stored at -80°C until required for HPLC analyses.
Neurotransmitter analysis by HPLC-EC. HPLC-EC was used to analyze brain punches for norepinephrine (NE), dopamine (DA), and serotonin (5-HT). Brain punches were brie y homogenized in 0.05 M perchloric acid on ice and an aliquot was used for protein estimation (MicroBCA assay, Pierce, Rockford, IL). The remaining homogenate was centrifuged at 18,000 × g for 8 min at 4°C. The supernatant was injected with an internal standard (dihydroxybenzylamine, 0.05 M) into the autoinjector for HPLC analysis. The HPLC-EC system comprised of a 5-µm ODS reverse phase C-18 column (Phenomenex, Torrance, CA), a SIL-20AC autoinjector, a CTO-20AC column oven (Shimadzu, Columbia, MD) maintained at 37°C and a LC-4C detector (Bioanalytical Systems, West Lafayette, IN). The ow rate of the mobile phase was maintained at 1.8 ml/min using a LC-20AD pump (Shimadzu, Columbia, MD).
Chromatograms were analyzed for neurotransmitter concentrations using the Class VP software v 7.2 (Shimadzu, Columbia, MD).
Protein levels in tissue punches were measured using the micro bicinchoninic acid assay (Pierce, Rockford, IL). Samples were assayed in duplicate according to the manufacturer's protocol. Neurotransmitter concentrations in tissue samples were expressed as pg/µg of protein.
Immunohistochemistry. Non-consecutive coronal brain tissue sections were xed with 4% paraformaldehyde in phosphate-buffered saline (PBS) solution for ten minutes, rinsed with PBS three times and incubated in blocking buffer (PBS with 0.5% Triton-X100 containing 4% goat serum) for an hour. Sections were then incubated in blocking buffer containing primary antibody against NeuN (ABN91; 1:500) overnight at 4°C. The next day, sections were rinsed with washing buffer (PBS with 0.5% Triton-X100) three times followed by an hour of blocking. Sections were incubated in blocking buffer containing AlexaFluor 488 (Life Technologies, CA), for secondary binding. The sections were rinsed again with washing buffer, counterstained with NucBlue (Life Technologies, CA), and mounted with uoromount-G (SouthernBiotech, AL). Images were obtained using the Leica DM microscopy system (Leica Microsystems Inc., IL). Fiji (NIH) was used to process the raw images obtained for the presence of neurons 24 . Coronal sections between bregma +0.5mm and -0.5mm was used to evaluate the presence of neurons post-CCI.
Statistical Analysis. For all group comparisons, Kolmogorov-Smirnov or Shapiro-Wilk tests were performed to test for normality.
When sample size and normality were veri ed, student t-test was performed for two group comparisons, and Wilcoxon rank sum test was used for multiple pairwise comparisons. Two-way Analysis of Variance was used to compare two independent variables between groups. All statistical tests, except neurotransmitter data, were performed using Sigmaplot (Systat Software, Inc., CA) and R-studio (RStudio Inc.,MA). All neurotransmitter data were analyzed using unpaired t-tests to identify differences between the control group and DEHP group using GraphPad prism software 9.0.0. A p-value < 0.05 was considered to indicate a statistically signi cant difference. All values are reported as mean ± standard error of mean (s.e.m.).

Results
Spatial memory recovery was impaired in male DEHP-exposed rats. Spatial memory recovery was tested weekly on the Morris water maze (MWM) task over a period of 4 weeks post-sTBI (Fig. 2). Longer time spent in the target quadrant (South; initially trained with escape platform) indicated a stronger spatial memory 20 . Both control and DEHP-exposed male rats demonstrated decreased preference for the target quadrant, 1-week post-sTBI (Control male: 32.5±3.3%, DEHP male: 24.0±4%) in comparison to baseline (Control male: 41.3 ±3.3%, DEHP male: 34.7± 4%; p=0.04; Fig. 2a). Four weeks post-TBI, control rats were able to recover to baseline performance (41.7±3.3%), while DEHP rats showed a sustained de cit (29.2±4%, p=0.024). The preferred quadrant was once again con rmed by average tracing of male control and DEHP rats, with hyper-intense regions in the heatmap representing highly overlapping paths (Fig. 2b). Traces of male control and DEHP rat activity before TBI displayed hyper-intense regions localized in the target quadrant as marked with red lines. Locomotor activity was reduced 1-week post-TBI in both control and DEHP rats. 4 weeks post-TBI, the intensity of activity in the target quadrant was increased in male control rats but remained low in male DEHP rats. No signi cant differences in activity were detected between control and DEHP-exposed females (Fig. 2c). Both control female rats and DEHP female rats maintained a preference for the target quadrant 1 week and 4 weeks post TBI (Fig. 2c, d).
Effects of pre-natal exposure to DEHP on lesion area, and hippocampal neuronal number. To investigate the physiological damage caused by CCI, we quanti ed the cortical lesion area, as well as the number of neurons in the hippocampus (Fig. 3). No statistically signi cant differences in lesion area (male, p=0.411; female, p=0.461) or the number of hippocampal neurons (male, p=0.0774; female, p=0.293) between DEHP vs control animals were observed in both sexes.
Neurotransmitter changes in the cortex following sTBI in rats prenatally exposed to DEHP. There were no signi cant differences in neurotransmitter levels (pg/µg protein; Mean±SEM) in the ipsilesional or contralesional cortex between control and DEHP-exposed male rats 4 weeks post TBI (Table 1). In contrast, DEHP exposed female rats demonstrated signi cantly higher norepinephrine (NE), dopamine (DA) and serotonin (5-HT) levels (6.14±0.93, 1.7±1.09 and 1.64±0.95 respectively) compared to control (2.26±0.44, 0.322±0.14 and 0.39±0.1 respectively) in the ipsilesional cortex (Fig. 4). DA levels were moderately elevated only in females in the contralesional cortex in DEHP animals (0.3±0.04) compared to control (0.06±0.001; p<0.05). There were no signi cant differences between the two groups in the levels of NE or 5-HT in the contralesional cortex (Table 1).
Neurotransmitter changes in the hippocampus following sTBI in rats prenatally exposed to DEHP. Similar to observations recorded in the cortex, there were no signi cant changes in neurotransmitter levels (pg/µg protein; Mean±SEM) in the ipsilesional or contralesional hippocampus between control and DEHP-exposed male rats 4 weeks post-sTBI (Table 1). However, in female rats, we observed a marked increase in DA levels in DEHP exposed rats (1.13±0.09) compared to control (0.017±0.004; p<0.01). There were no other changes in NE or 5-HT levels in female rats (Table 1).
Changes in stress axis activity in male and female rats following sTBI. There were no signi cant changes in NE, DA or 5-HT levels in the paraventricular nucleus of the hypothalamus in both male and female rats post sTBI. Nor did we observe any changes in serum CORT levels between control and DEHP rats in both male and female rats ( Table 2).

Discussion
This study demonstrates that prenatal exposure to DEHP, a ubiquitous endocrine disrupting chemical, can impact spatial memory in a sex-speci c manner following sTBI. Prenatal DEHP-conditioned adult male rats displayed signi cantly reduced spatial memory, while control rats successfully recovered spatial memory de cits 4 weeks post-TBI (Fig. 5). In contrast to adult male subjects, prenatal-DEHP-conditioned adult females showed little spatial memory impairment following TBI. This was accompanied by increased monoamine levels in the ipsilesional cortex of DEHP-exposed female rats. DA levels were also elevated in the contralesional hippocampus of DEHP-exposed females. Although we suspected that changes in stress axis activity as measured by monoamine levels in the PVN and serum CORT would contribute to the sex-speci c recovery of spatial memory, we did not observe any changes in these parameters.
The sTBI injury in this experiment was con ned to the right cortex. We did not observe any signi cant differences in the lesion area or the number of NeuN staining neurons between the control and DEHP-exposed groups suggesting that prenatal DEHP exposure did not contribute to exaggerated post-sTBI degenerative responses (Fig. 3). However, it appeared to have impacted neurotransmission ( Table 1, 2). We have previously observed that prenatal DEHP exposure signi cantly decreases monoamine levels in the cortex of uninjured male rats but increases them in female animals (Mohankumar et al., Athens, 2021. Unpublished).
The monoamine levels observed after sTBI in the present study were about 4-fold lower than what we had observed in uninjured animals, but the pattern observed after prenatal DEHP exposure was preserved. This effect was apparent even 4 weeks after sTBI indicating that monoamine transmission was severely compromised by the injury in male rats and probably requires a longer recovery period to return to baseline. Interestingly in female rats, there was a signi cant increase in all three monoamines in the ipsilesional cortex suggesting a possible compensatory mechanism that contributes to their improved performance on the MWM. A similar compensatory increase in neurotransmission has been observed in the early phases of other degenerative disorders involving the cortex such as Alzheimer's disease 25 . This could be part of a wide spectrum of biological processes that contribute to brain function and homeostasis 26 . While some clinical reports indicate that female TBI patients tend to have a higher percentage of physiological abnormalities, leading to neurosurgical interventions and mortality [27][28][29][30] ; other studies suggest that females recover faster than males from TBI 31 . The inherent differences in white matter architecture, synaptic connections, differences in hormonal milieu and the related divergence in gene expression between males and females 32 could all be contributing factors and need further investigation.
The hippocampus is a major center for spatial memory in both rodents and humans 33,34 . Previous studies have demonstrated that an impact to the rodent cortex can result in spatial learning de cits despite an intact hippocampus 35,36 . Reduced synaptic integrity and apoptotic dentate gyrus granule cells are largely responsible for possible hippocampal memory impairments 37,38 . We examined the number of neurons present in the hippocampus to con rm that the injury was limited to the cortex and did not nd any differences between control and DEHP-exposed rats in both sexes (Fig. 3). However, we found increased levels of DA in the contralesional hippocampus in female DEHP-exposed rats that could be linked to their better performance on the MWM. DA in the hippocampus is important for various types of learning and memory 39 . Moreover, photostimulation of DA neurons in the midbrain has been shown to enhance memory persistence and improve hippocampal network dynamics 40 . Therefore, the increase in DA levels in DEHP-exposed females could have probably contributed to the better recall of the platform location on the MWM. A point to consider is that both control and DEHP-exposed females had the same performance on the MWM, yet DA levels in the hippocampus were only increased in DEHP-exposed females. Although the reason for this is unclear, there could be underlying processes that are taking place in DEHP-exposed animals that makes them over-compensate after sTBI. A recent study indicates that female rats are able to upregulate the lipid content in the hippocampus, which in return makes them resistant to DEHP, while males could not 41 . Moreover, intrinsic sex differences in hippocampal neurogenesis and function have long been addressed to emphasize how male and female brains react differently under the same injury/stress conditions 42,43 . Additional investigation is necessary to understand the mechanism in which the hippocampus is protected following sTBI despite pre-exposure to DEHP in female rats. Detailed examination of hippocampal integrity including synaptic plasticity in prenatal-DEHP-treated rats will facilitate better understanding of the effect of prenatal DEHP exposure on hippocampus-related learning and memory processes post-TBI.
The HPA axis is an essential regulator of stress and injury responses 11 . NE and 5-HT are believed to be stimulatory and DA inhibitory or without effect on the HPA axis 13 . Dysregulation of the HPA axis results in detrimental outcomes including poor memory function 44 . About 65% of moderate-to-severe TBI patients suffer from long-term cognitive impairments 45 , which could be exacerbated due to CORT imbalance. In this study, we did not observe any differences in NE, DA or 5-HT levels in the PVN between control and DEHP-exposed rats post TBI. Moreover, there were no differences in serum CORT between the two treatments or between the sexes. The levels of neurotransmitters in the PVN and CORT observed in this study are markedly lower than what has been observed in these rats without TBI (Mohankumar et al., Athens, 2021. Unpublished). Although we did not detect any signi cant differences in CORT or neurotransmitter levels, we cannot exclude the potential role HPA-axis plays in the prolonged memory impairment. Altered glucocorticoid receptor expression in the hippocampus post-TBI causes hippocampal neuron apoptosis and aggravates spatial memory impairment 46,47 . This could be a possible mechanism that is in play in male rats which hinders recovery post-TBI.

Conclusion
This study demonstrated that prenatal exposure to DEHP led to impaired recovery of higher cognitive function post-TBI in male, but not in female rats. These results also suggested that the sex and environmental exposures can signi cantly affect injury progression and recovery from TBI of individuals and accentuates the need for precision medicine.  Figure 1 Experiment Schedule. Pregnant Sprague Dawley rats were treated with saline (control) or 7.5 mg/kgBW of DEHP by oral gavage from day 6-21 of gestation. When male and female offspring were 6 months old, they were subjected to Morris water maze (MWM) training for 5 days and subsequently subjected to a controlled cortical impact (CCI)-induced severe traumatic brain injury (sTBI). Spatial memory was monitored at weekly intervals over 4 weeks post-TBI. Serum and brain were collected at the end of 4 weeks and processed for further analysis.   Monoamine neurotransmitters were measured in the ipsilesional cortex 4 weeks post-CCI. No difference was detected between control and DEHP-treated male rats in (a) norepinephrine, (b) dopamine or (c) serotonin levels. DEHP-treated female rats had higher levels of (a) norepinephrine, (b) dopamine and (c) serotonin compared to control rats 4 weeks post TBI. * indicates p<0.5. **indicates p<0.01.

Figure 5
Summary of ndings.Prenatal exposure to DEHP impaired spatial memory recovery in males following sTBI, while DEHPconditioned adult females increased monoamine levels in the ipsilesional cortex, attributing to minimal spatial memory de cit. In addition, increased dopamine level in the contralesional hippocampus was observed. Figure created with BioRender.com.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.