key: cord-0262860-hwltkplo
authors: Maphis, Nicole M.; Huffman, Tameryn Radcliff; Linsenbardt, David N.
title: The Development, but not Expression, of Alcohol Front-loading in C57BL/6J Mice Maintained on LabDiet 5001 is Abolished by Maintenance on Teklad 2920x Rodent Diet
date: 2022-02-25
journal: bioRxiv
DOI: 10.1101/2022.02.23.481358
sha: 5dbe2e66d9ce80e1555a72ff04e93dcb91e41254
doc_id: 262860
cord_uid: hwltkplo

Excessive alcohol (ethanol) consumption, such as binge-drinking, is extremely commonplace and represents a major health concern. Through modeling excessive drinking in rodents, we are beginning to uncover the neurobiological and neurobehavioral causes and consequences of this pattern of ethanol intake. One important factor for modeling binge drinking in mice is that subjects reliably drink to blood ethanol concentrations (BECs) of 80 mg/dl or higher. Drinking-in-the-dark (DID) is a commonly used mouse model of binge drinking, and we have shown these methods reliably result in robust ethanol front-loading and binge-level BECs in C57BL/6J (B6) mice as well as other ethanol-preferring mouse strains/lines. However, establishing the DID model in a new vivarium space forced us to consider the use of rodent diet formulations we had not previously used. The current set of experiments were designed to investigate the role of two standard rodent diet formulations on binge drinking and the development of ethanol front-loading using DID. We found that BECs in animals maintained on LabDiet 5001 (LD01) were double those found in mice maintained on Teklad 2920x (TL20). Interestingly, this effect was paralleled by differences in the degree of front-loading, such that LD01-fed mice consumed approximately twice as much ethanol in the first 15 minutes of the 2-hour DID sessions compared to TL20-fed mice. Surprisingly however, mice that developed front-loading during maintenance on the LD01 diet continued to display front-loading behavior after being switched to the TL20 diet. These data emphasize the importance of choosing and reporting diet formulations when conducting voluntary drinking studies and support the need for further investigation into the mechanisms behind diet-induced differences in binge drinking, particularly front-loading.

LD01-fed mice consumed approximately twice as much ethanol in the first 15 minutes of the 2-23 hour DID sessions compared to TL20-fed mice. Surprisingly however, mice that developed front-24 loading during maintenance on the LD01 diet continued to display front-loading behavior after 25 being switched to the TL20 diet. These data emphasize the importance of choosing and reporting 26 diet formulations when conducting voluntary drinking studies and support the need for further 27 investigation into the mechanisms behind diet-induced differences in binge drinking, particularly 28 front-loading. 29 30 Introduction: 31 A large percentage of mortality and disease, globally, can be attributed to alcohol (ethanol) 32 use. In the US alone, excessive ethanol drinking leads to 95,000 deaths and more than $250 33 billion in financial costs each year (numbers from 2010; CDC). Binge drinking, defined as a pattern 34 of drinking that leads to a blood ethanol concentration (BAC) at or over 0.08 g/dL (or 80 mg/dL; 35 NIAAA/NIH), is a particularly detrimental form of ethanol use. Unfortunately, binge drinking is also 36 extremely commonplace, and there are signs it may be increasing. For example, a survey 37 conducted during lockdowns associated with the Covid-19 pandemic observed that not only did 38 rates of ethanol consumption increase significantly in both males and females, but episodes of 39 binge drinking also increased (Pollard et al., 2020) , especially amongst previous binge drinkers 40 (Weerakoon et al., 2020) . Another study corroborated these findings and found that these 41 increases in general ethanol consumption and binge drinking during the Covid-19 pandemic also 42 occurred broadly across all sociodemographic sub-groups (Barbosa et al., 2021) . Thus, 43 identifying factors that contribute to drinking too much ethanol, too quickly, is critical for developing 44 strategies to improve public health. 45

Given ethical issues surrounding the imposition of binge ethanol exposure on human 46 subjects genetically vulnerable to developing an alcohol use disorder (AUD), frequent unreliability 47 of self-reporting in clinical studies, and limited access to human brain tissue, the use of reliable 48 animal models are paramount to uncovering the neurobiological underpinnings of binge drinking. 49

There are now many such rodent models, each of which has its own advantages (Belknap et al., importance. For example, greatly increasing the palatability of food has been shown to decrease 67 ethanol consumption (Dole et al., 1985) . However, more recently, seemingly subtle differences in 68 rodent diets have been found to lead to profound differences in ethanol consumption (Marshall et (JAX) at 8 weeks of age and allowed to acclimate to a reverse light/dark cycle (12 hours off/on) 77

and single-housing in standard 'Allentown' shoebox cages for at least two weeks prior to 78 experimentation. Animals had ad lib access to one of two rodent diets (detailed below) and water, 79 except during the 2-hour DID protocol, when (some) mice had access to a 20% ethanol solution, 80 instead of water. DID procedures were initiated when mice were at least 70 days of age. All Table 1 Blood samples were then centrifuged at 10,000 RPM for 10 minutes, and plasma was withdrawn 106 and stored at -20°C until enzymatic assessment of BECs using an Analox Ethanol Analyzer 107 (Analox Instruments, Lunenburg, MA). 108

A subset of mice were administered 2.0 g/kg ethanol (20% v/v) via intraperitoneal (i.p.) 110 injection on day 16, 24 hours following day 15 fluid access. Immediately following i.p. injection of 20% ethanol, mice were returned to their home cages for 1 or 2 hours, and then blood samples 112 were collected and processed as detailed above for determination of BECs. 113

Experiment 1 was the first experiment conducted in our new lab and was designed to test 115 our newly acquired and assembled volumetric sipper hardware/software using 12 male and 12 116 female mice. All mice were maintained on TL20 diet throughout and were given 15 consecutive 117 2-hour DID ethanol sessions. Experiment 2 used 40 female mice, maintained on either TL20 or 118 LD01 rodent diets, and half of which were assigned to either ethanol or water DID solutions. Mice 119 were given 15 consecutive 2-hour DID sessions before potential ethanol metabolism differences 120

were assessed on the 16 th day. Experiment 3 used 20 female mice, half of which received either 121 TL20 or LD01 rodent diets. Following 15 consecutive 2-hour DID ethanol sessions, mice in this 122 experiment were then provided the opposite rodent diet for a period of two weeks, before another 123 round of 15 consecutive 2-hour DID ethanol sessions were conducted. 124

125 Experment 1: Rodents maintained on Teklad 2920x (TL20) do not reliably binge-drink.

Repeated measures ANOVA did not detect any statistically significant differences in 127 ethanol consumption as a function of sex (F(1,22)=2.36; p = 0.1384) or day (F(14,308)=0.71; p = 128 0.3240; Figure 1A ). Assessment of day 15 ethanol intake (student's t-test, p=0.3083) and 129 consequent BEC (student's t-test, p=0.9957; Figure 1B ) also failed to detect any sex differences. 130

Furthermore, and most importantly, mean BECs were well below the 80 mg/dl binge threshold 131 (average of males and females = 39.43 ± 9 mg/dl), justifying experiments 2 and 3, which were 132 designed specifically to test the hypothesis that mice maintained on LD01 would consume 133 significantly more ethanol versus mice maintained on TL20. 139

The influence of TL20 and LD01 rodent diets on ethanol consumption, water consumption, 140 and BEC can be seen in Figure 2 with LD01 consuming mice quickly developing greater front-loading compared to TL20 consuming 155 mice ( Figure 2C ). An identical analysis for water consuming mice found no statistically significant 156 effects ( Figure 2D ). Unsurprisingly, differences in ethanol intake (student's t-test, p = 0.0374) were 157 paralleled by differences in BEC (student's t-test, p = 0.0062) ( Figure 2E Figure 1D) . 164

Experiment 3: Rodents maintained on LD01 reliably consume more ethanol, display 165 increased front-loading behavior, and achieve approximately two-fold higher BECs than 166 animals maintained on TL20.

The influence of TL20 and LD01 rodent diets on ethanol consumption, rate of ethanol 168 consumption, and BEC can be seen in (Figures 3B, 3D) . Differences in ethanol intake 175 (student's t-test, p = 0.0057) over the course of the first 15 days resulted in over two-fold difference 176

in BECs (student's t-test, p = 0.0007) ( Figure 3C) , with only animals in the LD01 group displaying 177 binge-levels BEC as a group (i.e. >80 mg/dl). However, even though animals switched from TL20 178 to LD01 consumed significantly more ethanol than on the previous diet ( Figure 3B, Figure 3D ), these differences were much smaller in magnitude (student's t-test, p = 0.3201) and did not lead 180 to statistically significant differences in BECs (student's t-test, p = 0.6333) ( Figure 3E) . 181

To determine the extent to which differences in ethanol consumption might be driven by 182 differences in ethanol drinking rate, i.e. front-loading, we next explored the time course of drinking 183 throughout the daily 2-hour sessions ( Figure 4A ). Differences in mean ethanol intake between the 184 two diet groups were initially mediated by a persistent drinking rate in the LD01 group ( Figure 4B) . 185 However, differences in mean intake quickly shifted to being attributable to differences in front-186 loading; more specifically, lack of front-loading in the TL20 diet group ( Figure 4C stability of differences in front-loading once they emerged ( Figure 4D ). We then explored the 2 nd 190 15-day block of drinking, and much to our surprise, did not observe any significant differences in 191 ethanol front-loading ( Figure 4E , Figure 4F ), principally because front-loading in the group that 192 had been established by mice previously maintained on LD01, continued front-loading once they 193 were switched to TL20 diet, even by the last day of the 2 nd 15-day block (Day 30, Figure 4G ). To 194 confirm these observations, we re-analyzed these data by generating a mean consumption value 195 for each subject within each block. Our evaluation of mean 2-hour ethanol consumption revealed 196 a significant block*diet interaction [F(1,17)=62.85; p<.0001], which post-hoc tests confirmed was 197 due to both within and between subjects differences at every level ( Figure 5A ). In contrast, 198 although our evaluation of the first 15-minute period during DID revealed a significant block*diet week block (D1-D15; Figure 5B ). Together, these findings suggest that something about the TL20 201 diet prevents animals from developing ethanol front-loading, but that once front-loading is 202 established under the LD01 diet, it will persist. These data support our long-standing hypothesis 203 that increases in the rate of ethanol consumption (i.e. front-loading) reflects an increase in 204 motivation to experience ethanol's post-absorptive effects and suggest that particular dietary 205 attributes may increase or decrease ones susceptibility to developing an AUD. 206 

236

Overall, these studies find that B6 mice maintained on LabDiet 5001 (LD01) reliably 237 consume more ethanol, consume ethanol more quickly, and achieve higher BECs than mice 238 maintained on Teklad 2920x (TL20). Furthermore, although differences in total fluid consumption 239 were non-specific to ethanol (they were also observed in water-consuming groups), ethanol front-240 loading behavior was uniquely impacted in ways that support it as a learned behavioral in their study, and although there were no significant differences in water consumption among the 261 6 diets, it was lowest in mice maintained on TL20. 262

The specific attributes of the 2 diets under study that led to differences we observed 263 remain unknown, but we considered several key factors. One we recognized early was that LD01 264 pellets are compressed whereas TL20 pellets are extruded. Extruded dietary components are 265 typically ground up finer resulting in a less dense pellet (Kurtz & Feeney, 2020), which we 266 observed also occurs in these two diets specifically, and is likely responsible for differences in 267 size, shape and color of fecal boli between the different diet groups (Supplementary Figure 1) . 268

These density differences could potentially lead to differences in fluid adsorption or other physical 269 barrier to rapid fluid ingestion. However, mice consuming TL20 were demonstrably capable of 270 front-loading, they just only did so after this phenotype was established during maintenance on 271 LD01 ( Figure 4F ). 272

We also carefully considered differences in ingredients. For example, the 273 carbohydrate/protein ratio of food has been shown to impact ethanol consumption; specifically, 274 that high carbohydrate to low protein ratio depresses ethanol consumption and vice versa 275 (Kampov-Polevoy et al., 1999). Our observations are in-line with these findings, as TL20 has a 276 higher carbohydrate/protein ratio than LD01, but these differences were extremely small and 277 therefore unlikely to be a major driver of our observations. Despite similar proportions of 278 macronutrients (Supplementary Table 1 and 2), the specific ingredients that were used in the 279 manufacture of the two diets did vary quite substantially. Primary differences were the sources of 280 protein between diets (Supplementary Table 1 and 2), with only LD01 containing soy-based 281 products that are known to contain phytoestrogens (PEs). PEs functionally and structurally mimic 282 mammalian estrogens and their active metabolites, which in turn can modulate estrogen-sensitive 283 pathways (Mäkelä et al., 1995) . In one study, animals fed a diet lacking PEs experienced deficits 284 in learning and memory, which was rescued when subjects were supplied exogenous equol (a 285 metabolite of estrogen; (Çalışkan et al., 2019) . In contrast, animals fed a diet high in PEs 286 experienced profound alterations in energy balance including reduced body weight, adiposity, and 287 increased lipid oxidation (Cederroth et al., 2007) . Our data potentially align with this report, as our 288 subjects gained slightly more weight over a two-week period on the non-phytoestrogen-containing 289 diet (TL20), as compared with the phytoestrogen high diet (LD01) (Supplementary Figure 2C) . A 290 recent mini-review on this subject was just published in which the authors suggested the existence 291 of a direct relationship between phytoestrogens and ethanol consumption (Eduardo and Abrahao, 292

Relatedly, different commercial rodent diets have been found to lead to profound 294 differences in the gut microbial community (Tuck et al., 2020) . These diet-induced differences in 295 microbial communities may impact the way food and other substances, like ethanol, are 296 metabolized. Notably, BEC is influenced by how quickly ethanol is emptied from the stomach and 297 the extent of metabolism that occurs after it passes through the stomach to the liver (Reviewed in 298 (Zakhari, 2006) . Since we did not thoroughly assess ethanol pharmacokinetics, diet-induced 299 alterations in ethanol pharmacokinetics remain a possible mechanism of the behavioral 300 differences. However, given we did not detect differences in ethanol metabolism, and we 301 observed a significant positive linear relationship between ethanol consumption and BEC 302 regardless of diet, we think this unlikely. 303

To our knowledge this is the first report that rodent diet can impact ethanol front-loading, 304

and that these diet-induced alterations in front-loading may be primary to decreases in total 305 ethanol consumption within a drinking session. Although additional work is necessary to describe 306 the mechanisms of these observations, these studies provide further support of our long-standing 307 hypothesis that increases in the rate of ethanol consumption (as indexed by front-loading) may 308 reflect an increase in motivation to experience ethanol's positive post-absorptive effects, and 309

suggest that particular dietary attributes may influence ones susceptibility to developing 310 deleterious patterns of ethanol consumption. 311

Show Reaction to Loss of Ethanol Reward Following Repeated Binge Drinking

Alcohol Consumption in Response to the COVID-19 317 Pandemic in the United States

Depletion of dietary 319 phytoestrogens reduces hippocampal plasticity and contextual fear memory stability in adult 320 male mouse

Rate of alcohol consumption in the 322 daily life of adolescents and emerging adults

A Phytoestrogen-Rich Diet Increases Energy 325 Expenditure and Decreases Adiposity in Mice

Toward an analogue of alcoholism in mice: criteria for 327 recognition of pharmacologically motivated drinking

Food composition can influence how much alcohol your 329 animal model drinks: A mini-review about the role of isoflavones

ASSOCIATION BETWEEN 331 PREFERENCE FOR SWEETS AND EXCESSIVE ALCOHOL INTAKE: A REVIEW OF 332 ANIMAL AND HUMAN STUDIES

Relative Fluid Novelty Differentially Alters the Time Course 334 of Limited-Access Ethanol and Water Intake in Selectively Bred High-Alcohol-Preferring 335

Alterations in the rate of binge ethanol consumption: 337 implications for preclinical studies in mice

Role of Novelty and Ethanol History in Locomotor 339 Stimulation Induced by Binge-Like Ethanol Intake

Agonism of the endocannabinoid system modulates binge-341 like alcohol intake in male C57BL/6J mice: involvement of the posterior ventral tegmental 342 area

Ataxic Effects and Alterations in Ethanol-Induced Locomotion Following Repeated Binge-345

Like Ethanol Intake Using the DID Model

Phytoestrogens are partial estrogen agonists 347 in the adult male mouse

Assessment of 349 the Effects of 6 Standard Rodent Diets on Binge-Like and Voluntary Ethanol Consumption

Changes in Adult Alcohol Use and Consequences 352 During the COVID-19 Pandemic in the US

Effect of different 354 standard rodent diets on ethanol intake and associated allodynia in male mice

Nutritional profile of rodent diets impacts 358 experimental reproducibility in microbiome preclinical research

Guidelines for the Care and Use of Mammals in Neuroscience and 360 Behavioral Research

Longer time spent at home during COVID-19 362 pandemic is associated with binge drinking among US adults

Overview: how is alcohol metabolized by the body?