Introduction

In humans, long-term, excessive alcohol consumption results in a variety of somatic and central nervous system (CNS) insults [e.g., 1, 2]. In vivo brain structural magnetic resonance imaging (MRI) of uncomplicated alcoholism (that is, alcohol use disorder (AUD) absent clinically diagnosable consequences of chronic alcohol consumption such as Wernicke’s encephalopathy) commonly reveals tissue shrinkage in frontally distributed cortical regions [3,4,5] and subcortical regions including striatum [6,7,8] and thalamus [9, 10]. Factors that contribute to persistent or accelerated brain volume abnormalities are still being identified but likely include alcohol consumption variables of frequency and amount drunk, age at onset of alcohol misuse, and conditions resulting from years of abusive drinking, such as withdrawal signs and symptoms [11] and alcoholism-related nutritional deficiencies [12, 13]. Perhaps the most salient factor is age, which may render the older brain especially vulnerable to insult from other factors, including excessive alcohol consumption and attendant events [14]. Interaction of normal aging and alcoholism has been borne out in several cross-sectional [15,16,17] and longitudinal [18,19,20] studies in which accelerated volume declines appear by approximately age 45 years in alcohol-dependent groups. Assessing the association between age-alcoholism interactions and brain structure has new urgency considering current epidemiologic data indicating a major increase, measured throughout a decade, in the prevalence of AUD in individuals aged 65 years or older [21, 22].

Animal models of alcoholism offer the opportunity to investigate biological mechanisms of dependence, tolerance, and neuroadaptation, with the potential to test pharmacological and behavioral rehabilitation therapies. Given a number of previous studies that demonstrated similar, but non-enduring patterns of neuroimaging changes in response to high blood alcohol levels (BALs) [e.g., 23,24,25,26,27,28] we hypothesized that older animals may be necessary to observe persistent pathology [cf., 29]. Thus, the current set of experiments builds upon several previous studies that have used the modified Majchrowicz ethanol (EtOH)-intoxication model to achieve high BALs and longitudinally examine gross brain structure and metabolite responses in rats [24,25,26, 30]. Because human studies have demonstrated normalization of alcoholism-induced changes in metabolite levels with even short-term abstinence [e.g., 31, 32], cross-sectional, in vivo neurochemical differences examined at a single time point between AUD and healthy control individuals can be challenging to capture [33], but may provide insight into underlying mechanisms of pathology [e.g., 34] and contribute to CNS drug development [e.g., 35]. Multi-modal imaging of animal models, particularly MR Imaging (MRI) and Spectroscopy (MRS) studies—with the ability to measure changes pre- and post- exposure longitudinally—permit speculation regarding mechanism [e.g., 36, 37]. Structural and metabolic changes in response to high BALs achieved by binge (intragastric gavage) EtOH exposure have been consistent across our studies conducted in wild-type Wistar rats [24,25,26] and include ventricular enlargement [greater cerebrospinal fluid (CSF) volume], lower levels of N-acetyl aspartate [NAA, a marker of neuronal integrity, 38] and total creatine [tCr, energy buffer/shuttle, 38], and higher levels of choline containing compounds [Cho, involved in phospholipid synthesis and degradation, 38] together interpreted as reflecting altered brain energy utilization, disturbed osmolarity, or loss of white matter integrity [26]. While reliable, such changes are transient: following 1 week of recovery, ventricular volume and levels of NAA, tCr, and Cho return to baseline levels, no longer distinguishing control and EtOH-exposed animals. Thus, the first aim of the current study was to determine whether advanced age in rats, equivalent to ~45 human years, would produce more enduring brain structural and biochemical changes to alcohol exposure.

An additional source of variability contributing to brain abnormalities in AUD is sex. Although some literature has suggested that women are more vulnerable than men to the pathological effects of alcohol [39,40,41,42], other studies found that alcoholic women do not show disproportionately greater deficits than alcoholic men [43,44,45]. The rodent literature suggests that female relative to male rats of several strains drink more alcohol [46, 47, but see: 48,49,50]; but that male relative to female rats show greater evidence for withdrawal [51,52,53,54]. Thus, a second aim of the current study was to evaluate sex effects on brain structural and biochemical changes in response to alcohol exposure.

To accomplish these goals, 4-month old equivalent to ~16-year-old humans and 17-month old equivalent to ~45-year-old humans [55], male and female Fisher 344 rats were evaluated in a longitudinally controlled study using in vivo structural MRI and metabolic MRS. Specifically, these experiments tested the hypothesis that following equivalent alcohol exposure protocols, older relative to younger animals and male relative to female animals would exhibit more brain changes.

Materials and methods

Ethics statement

All experimental procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The Institutional Animal Care and Use Committees at SRI International and Stanford University approved all procedures.

Animals and treatment

Fisher 344 rats were acquired from the National Institute of Aging (https://ros.nia.nih.gov/). Experiments were conducted in two consecutive waves of 28 animals each comprising 14 young (~4 months, 7 female) and 14 old (~17 months, 7 female) animals for a total of 56 animals. As imaging data measures between cohorts were not different at any time point, data from the two waves were combined. Weight of animals at baseline was as follows: 14 old female (FO) 222 ± 15 g; 14 old male (MO) 428 ± 20 g; 14 young female (FY) 178 ± 11 g; 14 young male (MY) 312 ± 26 g. Animals had ad libitum access to standard laboratory chow and water.

After the baseline scans, eight animals from each of the four age + sex categories were exposed to EtOH. All animals were fasted overnight, and the EtOH group underwent a modified version of the Majchrowicz EtOH-intoxication model [56] as described in previous studies [24, 25, 30, 57]. Briefly, rats received an initial “loading” dose of 5 g/kg 20% w/v EtOH via oral gavage, then a maximum of 4 g/kg every 8 h (i.e., 2:00, 10:00, and 18:00) for 4 days. On each of the 4 days, animals were weighed, and tail vein blood samples (~300 μl) were collected at 8:00 each day to determine BALs in plasma based on direct reaction with the enzyme alcohol oxidase (Analox Instruments Ltd., UK). After the loading dose, EtOH was administered according to body weight, BALs, and behavioral evidence for intoxication [e.g., 56]. Control animals (n = 6 from each group) received consistent quantities of reverse osmosis water in volumes similar to those received by the experimental rats.

In total, five rats (three old males from wave 1 and two old females, one from each wave) did not survive binge EtOH exposure. Thus, results reported are for six controls (C) in each of the age + sex categories (i.e., FYC, FOC, MYC, MOC); eight in the young ethanol (E)-exposed groups (FYE, MYE), six in the old ethanol-exposed females (FOE); and five in the old ethanol-exposed males (MOE).

MR scanning procedures and data analysis

Schedule

All rats were scanned at baseline (scan 1), following 4 days of EtOH exposure and subsequent metabolic clearance of EtOH, ≥14 h (range: 13:30–22:30 h) after the last dose of EtOH (scan 2), and following 7 days of recovery (scan 3).

Anesthesia and monitoring

Isoflurane anesthesia was initiated at 2–3% and reduced to 1–2% at scan commencement. Rats were placed on an animal cradle equipped with built-in water circulation for body temperature control. Eye lubricant was used to anoint eyes for protection from dehydration. Silicon earplugs were affixed to protect ears from scanner gradient volume. Temperature and respiration were monitored throughout the ~2 h experiment. Animals received subcutaneous saline (10cc) for hydration at the end of the scan.

MRI acquisition

MR data were collected on a Bruker 70/16 US AVANCE III 7.0T system (Karlsruhe, Germany) with 380 mT/m gradient strength on each (X, Y, and Z) axis, slew rate of 3420 T/m/s, 16 cm bore size using a receive-only surface coil and ParaVision 6.1 software. A gradient-recalled echo (GRE) localizer scan was used to position the animals in the scanner and for graphical prescription of the subsequent scans. T2-weighted, high-resolution, TurboRare acquisition sequence: repetition time (TR) = 6774.8 ms; echo time (TE) = 33 ms; 0.2 mm isotropic voxels, matrix = 160 × 160; 2 averages; echo spacing = 11 ms; slice thickness = 0.5 mm; slice gap = 0; 64 interlaced slices; RARE factor = 16.

MRI processing

Preprocessing of each image included removal of noise [58] and inhomogeneity correction via ANTS 2.1.0 [59]. Each image was skull stripped by aligning a template to the scan via symmetric diffeomorphic registration [60] and the resulting deformation map was applied to the brain mask of the template. Image inhomogeneity correction was repeated on skull-stripped images. Bias-corrected, skull-stripped images were rigidly aligned to a template via ANTS 2.1.0 and used as input for further analysis. Structural images were segmented into CSF, gray matter, and white matter with finite mixture modeling (FMM) segmentation (ANTS atropos) producing a probability for each of the three tissue types for each voxel in the brain (Fig. 1). The final unit of measure for each tissue type was the integrated probability over the entire brain, yielding whole brain gray matter, white matter, and CSF volumes.

Fig. 1: Axial (top) and coronal (bottom) slices through a rat brain.
figure 1

From left to right are the T2-weighted structural images and the probability maps of CSF, gray matter, and white matter. The gray scale indicates the probability (0–100%) of the contents of each voxel by tissue type. The final unit of measure for each tissue type was the integrated probability over the entire brain, yielding whole brain gray matter, white matter, and CSF volumes.

Full size image

MRS acquisition

In vivo MRS scans were acquired in voxels placed in striatum or thalamus, each 3 × 3 × 3 mm with a point-resolved spin echo sequence (PRESS). The striatum was chosen because it is a node of the brain’s reward circuitry, is involved in motivated behavior, and is sensitive to volume loss in uncomplicated alcoholism [6, 7]. Further, profound sex differences have been observed in the striatum of male and female rats [61, 62]. The thalamus was similarly chosen because of demonstrated atrophy in alcoholism [63, 64] and also because our previous work using the Majchrowicz EtOH-intoxication model demonstrated decreases in tissue water transverse relaxation time and diffusivity measures selective to the thalamus [24]. For MRS, total acquisition time was 10:40 min (TR/TE = 2500/16 ms; 256 averages; 2 dummy scans; 4096 acquisition data points; spectral bandwidth 4000 Hz) with VAPOR (variable power and optimized relaxation delays) outer volume water suppression (bandwidth 200 Hz).

MRS metabolite quantification

LCModel analysis [Stephen Provencher, sp@lcmodel.CA] was run with the “water reference” option that provided reasonably meaningful absolute metabolite concentrations based on the unsuppressed water content of the voxel. The analysis window was 0.2–4.0 ppm. Data were pre-processed with zero-order phasing, referencing, and residual water line removal. Data were fitted to a linear combination of a number of metabolites in a simulated basis set designed for Bruker 7T acquisition at TE = 16 ms provided by Provencher containing alanine, creatine + phosphocreatine (tCr), glutamine (Gln), glutamate (Glu), glycerophosphorylcholine + phosphorylcholine (Cho), glutathione (GSH), inositol, lactate, N-acetylaspartate + N-acetylaspartylglutamate (NAA), scyllo-inositol, taurine, and several lipids and macromolecules. Only metabolite concentrations derived from fitted spectra consistently within average Cramér-Rao bounds <15% were considered.

Statistical analysis

Initial analysis of structural MRI data was conducted in R using a linear model (lm) analysis. All further analysis was conducted in JMP® Pro version 14.1.0 (SAS Institute Inc., Cary N.C., 1989–2019). For structural MRI, variables considered were whole brain, CSF, gray matter, and white matter volumes. For MRS, metabolites with average Cramer-Rao bounds below 15% were analyzed (i.e., NAA, tCr, Cho, Glu, Gln, GABA, GSH, Ins, taurine, and MM09). Baseline levels of each dependent variable were evaluated for effects of age, sex, and their interaction; for metabolites, region (i.e., striatum and thalamus) was additionally considered. Treatment effects on dependent variables were analyzed using omnibus repeated measures (time: baseline, binge, recovery) multivariate analysis of variance (MANOVA) considering age, sex, diagnosis, and their interaction with time. If the model was significant, effects of treatment (EtOH or control) were further considered. Significance required a p-value = 0.017 for structural MRI (i.e., Bonferroni corrected for 3 tissue types, 0.05/3 tissue types) and a p-value = 0.005 for MRS metabolite (i.e., Bonferroni corrected for 10 metabolites, 0.05/10 metabolites) derived models. Follow-up MANOVAs and t-tests within each age + sex (i.e., FO, FY, MO, MY) category considered treatment only. Finally, a data-driven approach entered all variables showing treatment effects into a JMP-based cluster analysis.

Results

Cumulative EtOH dose, BALs, and weights

EtOH animals were exposed to a cumulative 4-day total of EtOH as follows [mean ± standard deviation (SD)]: FOE 25.83 ± 1.94 g/kg; FYE 33.25 ± 1.46 g/kg; MOE 24.04 ± 3.32 g/kg; MYE 26.4 ± 1.49 g/kg (Supplementary Fig. 1a). Old male and old female EtOH groups were exposed to similar EtOH doses (p = 0.16); the young EtOH-exposed female rats tolerated a higher cumulative EtOH dose than the young EtOH-exposed  male rats (p < 0.0001).

Across the 4 days of exposure, peak BALs in the EtOH-exposed groups were as follows (mean ± SD): FOE 424.25 ± 141.69 mg/dL; FYE 333.25 ± 109.31 mg/dL; MOE 312.54 ± 88.05 mg/dL; MYE 512.48 ± 76.09 mg/dL (Supplementary Fig. 1b). Peak BALs were nominally higher in old female than old male rats (p = 0.09); peak BALs were lower in young female than young male rats (p = 0.002).

The model for weight was significant [F(8,92) = 20.9, p < 0.0001] and showed effects of age (p < 0.0001), sex (p < 0.0001), their interaction (p < 0.0001), and treatment (p < 0.0001). At all time points, old rats weighed more than young rats, and male rats weighed more than female rats; the age-by-sex interaction indicated that the weight difference between the younger and older male rats was greater than that between the younger and older female rats. At baseline, weights were not different between EtOH and control groups. Binge EtOH exposure caused significant weight loss in the male (young p = 0.02, old p = 0.01) but not female (young p = 0.32, old p = 0.46) EtOH-exposed rats; weight persisted in being lower at the recovery time point in the old male animals (young p = 0.08, old p = 0.01).

Age and sex differences at baseline

Table 1 summarizes baseline differences in volume measures and metabolite levels as a function of age and sex. A linear model (lm) analysis in R revealed that the total brain volume of the male rats was larger than that of the female rats [t(48) = 16.7, p < 0.0001] and that young rats were smaller than old rats [t(48) = −11.7, p < 0.0001] (Fig. 2). An age-by-sex interaction indicated that the volume difference between the younger and older male rats was greater than that between the female age groups [t(47) = −2.2, p = 0.03]. This age and sex pattern described each of the three constituent volumes (i.e., CSF, gray matter, white matter; Fig. 2 includes statistics for these results). The age-by-sex interaction was significant only for gray matter volume, showing a trend for white matter, and was not significant for CSF.

Table 1 Summary of baseline differences.
Full size table
Fig. 2
figure 2

MRI volumes at baseline. Baseline (a) total brain volume and (b) cerebral spinal fluid (CSF), gray matter, and white matter volume for each of the age+sex categories including results of lme in R showing effects of age (greater in older than younger Fisher 344 rats), sex (greater in male than female Fisher 344 rats), and an age-by-sex interaction wherein the total volume difference between the male young and old rats was greater than that between the female young and old rats. An age-by-sex interaction was significant in gray matter and showed a trend in white matter wherein volume difference between the male young and old rats was greater than that between the female young and old rats.

Full size image

NAA (NAA + NAAG, F(4,107) = 144.6, p < 0.0001) showed effects of age (p < 0.0001, old > young), sex (p < 0.0001, male > female), and region (p < 0.0001, thalamus > striatum). Glu (F(4,107)=19.8, p < 0.0001) showed effects of sex (p < 0.0001, male > female), region (p < 0.0001, thalamus > striatum), and an age-by-sex interaction (p = 0.02, young female < old female; young male > old male). GABA (F(4,107) = 29.2, p < 0.0001) and MM09 (F(4,107) = 5.2, p = 0.0007) showed sex (p < 0.007, male > female) and region (p < 0.002, thalamus > striatum) effects; Ins (F(4,107) = 84.1, p < 0.0001) showed age (p < 0.0001, old > young) and region effects (p < 0.0001, thalamus > striatum). Among the 10 metabolites, only GSH (F(4,107) = 4.0, p = 0.004) showed no region effect, and age and sex effects in the opposite direction: age (p = 0.05, young > old) and sex (p = 0.0008, female > male). tCr (Cr + PCr, F(4,107) = 47.0, p < 0.0001), Cho (GPC + PCh, F(4,107) = 17.7, p < 0.0001), Gln (F(4,107) = 6.8, p < 0.0001) showed only region effects (p < 0.0001, thalamus > striatum); taurine (F(4,107) = 423.1, p < 0.0001) also showed only region effects, but in the opposite direction (p < 0.0001, striatum > thalamus).

Binge EtOH exposure and recovery effects

Omnibus MANOVAs are presented in Table 2. Significant treatment effects are summarized in Table 3. MANOVAs by each age + sex category separately are presented in Supplementary Table 1. Supplementary Table 2 compares values [mean ± SD; MRI in cubic centimeters (cc), metabolites in IU; t-tests] in EtOH-exposed and control animals. Exemplary spectra are presented in Fig. 3. Graphs of metabolites by age and sex across three time points are presented in Fig. 4 and Supplementary Fig. 2 (GABA and MM09 not presented).

Table 2 Omnibus MANOVAs.
Full size table
Table 3 Summary of treatment effects.
Full size table
Fig. 3
figure 3

Voxel locations in striatum and thalamus shown in coronal, axial, and sagittal planes. Also pictured are spectra averaged across young (light green) or old (dark green), male (blue) or female (red) categories.

Full size image
Fig. 4: MRS-derived metabolites in thalamus and striatum.
figure 4

a NAA (NAA + NAAG), b tCr (Cr + PCr), c Ins, and d taurine levels in institutional units (IU) in the striatum and thalamus of young (light green) and old (dark green), male (blue) and female (red) groups across the three scan time points (1: baseline, 2: binge, 3: recovery; controls in gray). Note the shifted axis for tCr to permit representation of metabolite levels in both regions on a single graph. Also note taurine levels were higher in striatum than thalamus, while the opposite pattern pertained to the other three metabolites.

Full size image

The omnibus model for CSF volume was significant. All but the old females showed a treatment effect: binge EtOH exposure was associated with greater CSF volume in the EtOH-exposed relative to the control groups at the binge time point. Similarly, the omnibus model for gray matter volume was significant and showed treatment-by-time effects. All but the young females showed a treatment effect: binge EtOH exposure was associated with lower gray matter volume in the EtOH-exposed relative to the control groups at the binge time point. The model for white matter was not significant.

The overall model was not significant for NAA in either the striatum or thalamus. To permit comparison to previous studies, posthoc analysis showed treatment effects on striatal NAA in both male groups: binge EtOH treatment was associated with lower NAA levels in EtOH-exposed relative to control animals at the binge time point. The model was significant for tCr (i.e., Cr+PCr) in striatum and showed a treatment-by-time interaction: all but old females showed treatment effects with lower tCr levels in EtOH-exposed relative to control animals at the binge time point. The model was significant for Cho (i.e., GPC + PCh) in thalamus and showed a treatment-by-time interaction due to young female rats: EtOH-exposed relative to their controls had lower Cho levels at the binge time point, but this was not significant with Bonferroni correction.

The full model was significant for GSH in striatum and showed a treatment-by-time interaction significant only in young male animals: GSH levels were lower in EtOH-exposed than control rats at the binge time point. The model for Ins was significant in both regions. In the striatum, all but the young females showed significant treatment effects wherein binge EtOH treatment was associated with lower levels of Ins in the EtOH-treated relative to the control animals at the binge time point. In the thalamus, all but the old male rats showed significant treatment effects with the same pattern (i.e., lower Ins in the EtOH-exposed relative to control animals at the binge time point). The model was significant for striatal taurine and showed effects of treatment across all 4 age + sex categories: taurine was lower in the EtOH relative to the control groups at the binge time point. The full models were not significant for Glu, Gln, GABA, or MM09 in either the striatum or thalamus.

Cluster analysis

Relations between percent change (between baseline and binge) in weight, CSF and gray matter volumes, striatal (tCr, GSH, Ins, taurine) and thalamic (Ins) metabolites showing treatment effects, and cumulative EtOH dose and peak BALs (i.e., 10 variables) were explored using cluster analysis in EtOH-exposed animals only (i.e., n = 27). The resulting four clusters each explained 16–17% of the overall variance: cluster 1 (16.1%) included CSF and gray matter volume; cluster 2 (17.4%) striatal and thalamic Ins and striatal tCr; cluster 3 (17.3%) peak BAL and striatal GSH and taurine; cluster 4 (16.0%) cumulative dose and weight. For cluster 1, follow-up simple regressions demonstrated an inverse relationship between CSF and gray matter volume (r = −0.61, p = 0.0007); for cluster 2, striatal Ins positively correlated with striatal tCr (r = 0.51, p = 0.007), and thalamic Ins correlated with caudate Ins (r = 0.36, p = 0.06), but no other relationships reached or approached significance. For cluster 3, striatal taurine and GSH were correlated (r = 0.49, p = 0.009), but peak BALs showed only modest inverse relationships with each of the striatal metabolites (taurine r = −0.19, p = 0.15; GSH r = −0.29, p = 0.14). Finally, cumulative dose correlated with percent change in weight (r = 0.60, p = 0.0006).

Discussion

This longitudinal study of young and old, male and female Fisher 344 rats demonstrates fundamental differences based on age and sex in brain structure and neurochemical composition and further demonstrates effects of binge EtOH treatment. One caveat when comparing the current results with our earlier rodent studies of binge EtOH exposure is that previous experiments [24,25,26] used Wistar rats, whereas the present study used Fisher 344 rats as that is the only strain provided by the National Institute of Aging aged colony resource (https://ros.nia.nih.gov/).

Baseline differences based on age and sex

Total brain volume of male rats was larger than that of female rats and that of old rats was larger than that of young rats. This pattern was also present in the constituent tissue types (i.e., CSF, gray matter, white matter). The sex effects observed in these Fisher 344 rats are consistent with findings in humans because men as a group relative to women have larger total intracranial [65,66,67], cortical gray matter [68, 69], cortical white matter [70, 71], and CSF volumes [70, 72]. By contrast, larger total, gray matter, white matter, and CSF volumes in older than younger rats (i.e., continued brain growth with increasing age) diverges from human findings wherein total [e.g., 73] and gray matter volume [74,75,76,77] decreases steeply, white matter decreases but less precipitously [75, 78, 79], and CSF volume increases [72, 75] with age (generally after age 25) in both sexes [80, 81]. Evidence for continued brain growth with age in rodents comports with imaging studies conducted in alcohol-preferring male P rats tracked until 15 months of age [82] and in male Sprague-Dawley rats tracked until 13 months of age [83, also see 84,85,86].

With the exception of taurine, metabolite levels were higher in thalamus than striatum. Regional differences in metabolite levels are well-documented, in both human [e.g., 87, 88] and animal [e.g., 89] studies, but often, even those that include assessment of multiple regions evaluate treatment effects cross-sectionally rather than longitudinally [e.g., 90] precluding clear baseline estimates of regional differences. Previous studies in adult rats have demonstrated that taurine levels are high in striatum [relative to cortex and hippocampus, 91] and low in thalamus [relative to cerebellum and hippocampus, 92].

Metabolites that showed effects of older age included NAA, Ins, and GSH. MRS studies regarding aging effects on neurometabolite levels are equivocal [93,94,95,96,97] and likely related to the various methods used (ex vivo MRS studies may alter neurochemical levels during the tissue extraction and homogenization process) or to the different regions [e.g., 89, 92] or rat strains [98] evaluated. Higher NAA in the old relative to young animals comports with the observation of continued brain growth with increasing age; conversely, NAA is commonly reported to decline with aging in humans consistent with shrinking cortical volume [99]. Age-related increases in Ins [100, 101, also see 102]—a metabolite more highly concentrated in astrocytes than neurons [103]—maybe associated with age-related changes in astroglia [e.g., 104, 105]. Lower GSH was not observed in old (22-month) relative to young (3-month) Fisher 344 rats in MRS voxels placed in cortex or hippocampus [93], but postmortem assessment found significantly lower GSH in cortex, striatum, thalamus and hippocampus of middle aged (18-month) relative to young (3-month) Fisher 344 rats [106].

For nearly all metabolites, levels were higher in male than female rats. The only exception was GSH in the striatum, for which levels were higher in the female than male rats. Higher levels of neurometabolites in male than female rats have been reported in the few studies that have included both sexes [e.g., 102, 107,108,109] and may be related to the relatively bigger brains (i.e., greater total volume, gray and white matter volume) of male than female rats. Higher levels of the antioxidant GSH [e.g., 110] in the female than male rats may have contributed a protective effect against the more pronounced insult to CSF and gray matter volume observed in males caused by exposure to high EtOH levels.

Effects of binge EtOH treatment

Comporting with our previous studies performed in wild-type Wistar rats [24, 25, 30, 57], binge EtOH exposure caused transient ventricular enlargement (i.e., increase in CSF volume) that was significant in the young male rats but not present in the old female rats. Thus, the current results do not support the hypothesis that following equivalent alcohol exposure protocols, older animals exhibit more brain changes than young animals and indeed, demonstrate the opposite: young animals showed a greater change in CSF volume in response to binge EtOH treatment than their older counterparts. A caveat is that five of the old, EtOH-exposed animals died during EtOH exposure, thereby precluding brain-imaging evaluation at the binge time point.

Reduced ventricular expansion and gray matter volume loss in female than male rats in response to binge EtOH treatment could not be attributed to cumulative EtOH dose administered, because the old male and female EtOH groups were exposed to similar EtOH doses and the young female relative to the young male rats were exposed to a higher cumulative (relative to body weight) EtOH dose. Further, peak BALs, which were higher in the old female than male rats (though lower in the young female than male rats), cannot account for the sex differences in response to EtOH. That the young female rats had a higher cumulative dose of EtOH exposure (relative to body weight) but lower peak BALs appears to comport with the literature showing that female relative to male rats consume more EtOH [46, 47] possibly because relative to male rats, female rats have greater hepatic alcohol dehydrogenase (ADH) activity [111, 112].

Our previous studies of Wistar rats have consistently shown lower levels of NAA and tCr, and higher levels of Cho in response to binge EtOH treatment in voxels placed in dorsal hippocampus [25, 57] or whole brain [24]. The current study of Fisher 344 rats demonstrates that changes to tCr are reproducible across rat strains in the striatum. Although results of the model were not statistically significant, the NAA response in male Fisher 344 rats to high BALs was in the correct direction (lower in EtOH-treated animals relative to controls at the binge time point). Fisher 344 rats, however, did not demonstrate the predicted increase in Cho levels. The lack of an effect on Cho levels may be related to voxel location (i.e., thalamus and striatum in current study vs. hippocampus and whole brain in our previous studies) or to strain differences [e.g., 113,114,115] as cholinergic activity is constitutively higher in Fisher 344 relative to Wistar rats [see also 98, 116].

As this study was conducted on a higher strength magnet (7 T) than our previous spectroscopy studies (3 T) and used LCModel for analysis, additional metabolites were evaluated. Following binge-EtOH treatment, young male rats showed reduced GSH levels in striatum [117]; Ins levels were reduced in nearly all age + sex categories in both striatum and thalamus; and taurine was reduced in all age + sex categories in the striatum. Ins is reported as reduced in rat models of liver failure [118,119,120,121], after experimental cerebral ischemia [122], oxidative stress in astrocyte extracts from the rat brain [123], and following 8- [124] or 10- [125, 126] weeks of EtOH feeding. Ins is an organic osmolyte that participates in astrocytic volume regulation [127] and release of Ins from brain cells is thought to counteract cell swelling [128]. Thus, Ins depletion in response to binge EtOH treatment could be a response of astrocytes against cell swelling following exposure to high BALs. Taurine, which similarly contributes to the regulation of neuronal and astrocytic volume [129,130,131], is also reduced in models of liver failure [132, 133], in experimental cerebral ischemia [134], and oxidative stress in astrocyte extracts from the rat brain [123].

Limitations

Larger sample sizes would have been desirable, but the availability of young/old matched animals was limited as were the availability and cost of MR scanning. As with all MR spectroscopy studies, there is a frequency-dependent chemical-shift displacement between water and the various metabolites such that the voxel sampled is not precisely where it is prescribed anatomically, and the displacement varies depending on how far each metabolite is from water. Nonetheless, we used fairly large voxels on large structures, and the voxels as prescribed surely contained the mainstay of the target structure.

Conclusions

This study demonstrates baseline differences in brain structure and neurochemistry as a function of sex and age. Comporting with human studies, male relative to female rodent brain (total, gray, white, CSF) volumes are larger. Higher levels of NAA, Glu, and GABA in male than female rats may be associated with the larger brain size (i.e., more neurons). In contrast to the human condition, however, the rodent brain—as evidenced by total volume, gray and white matter volume, CSF volume, and NAA levels—continues to grow with older age. Ins levels higher with older age possibly reflect a compensatory mechanism, whereas GSH levels were lower with older age and may suggest loss of antioxidant activity.

Despite the use of a new strain, young male Fisher 344 rats demonstrated the pattern of ventricular enlargement and lower levels of NAA and tCr in response to binge EtOH exposure that we previously reported in young male Wistar rats. The cluster analysis showing a relation between CSF and gray matter volume changes may reveal CSF expansion at the cost of parenchymal shrinkage. The cluster analysis also supports the interpretation that neither cumulative dose nor weight directly influenced volume or metabolite levels. Instead, cumulative dose was directly related to change in weight as smaller (lower weight) animals tolerated the greatest cumulative dose. The cluster analysis showing a relationship between Ins and tCr may support a mechanism of reversible osmolarity changes due to intra- or inter- cellular fluid shifts [24] which might be due to decreased brain energy metabolism [124] in response to high BALs. Finally, higher GSH at baseline in female than male rats [135] and supports a mechanism whereby female rats may be protected against the more pronounced changes observed in male rats due to superior metabolic homeostasis mechanisms.