The goal of the protocol presented here is to describe procedures to expose rats to moderate levels of alcohol during prenatal brain development and to quantify resulting alterations in social behavior during adulthood.
Alterations in social behavior are among the major negative consequences observed in children with Fetal Alcohol Spectrum Disorders (FASDs). Several independent laboratories have demonstrated robust alterations in the social behavior of rodents exposed to alcohol during brain development across a wide range of exposure durations, timing, doses, and ages at the time of behavioral quantification. Prior work from this laboratory has identified reliable alterations in specific forms of social interaction following moderate prenatal alcohol exposure (PAE) in the rat that persist well into adulthood, including increased wrestling and decreased investigation. These behavioral alterations have been useful in identifying neural circuits altered by moderate PAE1, and may hold importance for progressing toward a more complete understanding of the neural bases of PAE-related alterations in social behavior. This paper describes procedures for performing moderate PAE in which rat dams voluntarily consume ethanol or saccharin (control) throughout gestation, and measurement of social behaviors in adult offspring.
An estimated 1-5% of children are diagnosed with Fetal Alcohol Spectrum Disorders (FASDs)2, which include Fetal Alcohol Syndrome (FAS), partial FAS (pFAS), and Alcohol-Related Neurodevelopmental Disorders (ARNDs)3. Deficits in social behavior and cognition are among the most common adverse outcomes observed in children with FASDs4-7. Negative consequences are not limited to heavy prenatal alcohol exposure (PAE), as moderate PAE that does not lead to the conspicuous morphological, behavioral and cognitive deficits characteristic of FAS can cause comparatively subtle, but nonetheless persistent, deficits in humans with FASDs8-10 and non-human animals exposed to ethanol during brain development11. The importance of understanding the behavioral and corresponding neurobiological consequences of moderate PAE is underscored by current estimates indicating that the large majority of FASD cases fall within the less severe range of the spectrum12.
Several independent laboratories have reported alterations in rodent social behavior related to ethanol exposure during brain development, including decreased investigation and interaction1,13-15, altered play14,16,17, increased aggressive interactions17,18, alterations in responsiveness to social stimuli19-21, and deficits in socially acquired food preferences and social recognition memory22. Social behavior deficits have been observed following exposure to heavy (blood ethanol concentrations (BECs) ~300mg/dl)22,23 or more moderate levels of ethanol (BECs ~80mg/dl)1, and across a broad range of parameters for other significant factors including exposure timing, duration of exposure, and age at the time of behavioral measurement.
Previous research has demonstrated that alterations in specific aspects of social interaction in adulthood discriminate rats exposed to moderate levels of alcohol from control animals exposed to saccharin1,18. In particular, moderate PAE has consistently been associated with robust increases in wrestling, which suggests increases in aggressive behavior, and lower levels of social investigation (e.g., sniffing of the partner) in adulthood. Because alterations in social behavior are reliable consequences of PAE, the quantification of social behavior following PAE may hold importance for progressing toward a more complete understanding of the neural bases of PAE-related alterations in social behavior and the development of interventional approaches. The goal of this paper and the associated video is to provide instruction on the moderate PAE protocol and methods for quantification of social behavior in adult offspring that have reliably distinguished prenatal alcohol-exposed from non-exposed rat offspring.
All procedures described here and in the accompanying video have been approved by the Institutional Animal Care and Use Committees of the Health Sciences Center and the main campus of the University of New Mexico.
1. Prenatal Ethanol Exposure
2. Social Behavior
Over the course of many breeding rounds female rats in the ethanol condition consistently drink an average of about 2.1 g/kg of ethanol per 4 hr drinking session. Rat dams consume approximately one-half of the four hr total during the first 15 to 30 min after the introduction of the drinking tubes, resulting in a peak maternal serum ethanol concentration of about 60 mg/dl, measured at the 45 min time point. Over the remaining 3.5 hr of the drinking period, they continue to consume 5% ethanol at a lower, but relatively stable rate of 0.4 g/kg body weight/hour. This level and pattern of voluntary ethanol consumption has no significant effects on maternal weight gain, offspring birth weight, litter size, maternal care, placental wet weight, offspring weight at behavioral testing, or whole brain, hippocampal or cerebellar wet weights.
Representative means and SEMs from male saccharin- and ethanol-exposed rats for each behavioral measure are shown in Table 1. These data were pooled from prior experiments and include 16 males for each prenatal treatment condition. All animals were paired with partners from the same prenatal treatment condition. Robust alcohol-related alterations in the social behavior of female animals have not been observed in our studies using these procedures1, however, alcohol-related differences in female social behavior have been documented using other procedures15,23. Separate univariate analyses of variance (ANOVAs) performed in SPSS ver. 21 for Macintosh revealed that male ethanol-exposed rats had significantly higher duration [F(1, 30) = 19.12] and frequency [F(1, 30) = 6.80] of wrestling and decreased latency to the first instance of wrestling [F(1, 30) = 9.41]. Ethanol-exposed rats also spent less time engaged in anogenital sniffing [F(1, 30) = 5.17].
1a. Frequency | SAC | PAE |
Wrestling * | 2.00 (0.58) | 8.00 (2.23) |
Boxing | 1.81 (0.80) | 3.56 (1.47) |
Cross over/under | 1.06 (0.48) | 1.00 (0.29) |
Anogenital sniffing | 6.25 (1.20) | 3.75 (0.73) |
Body Sniffing | 19.75 (1.64) | 18.56 (2.06) |
Allogrooming | 2.31 (0.93) | 0.75 (0.27) |
Rearing | 56.50 (5.39) | 56.06 (5.40) |
Dig/Sniff Bedding | 32.06 (6.03) | 30.06 (5.27) |
1b. Duration (sec) | SAC | PAE |
Wrestling ** | 9.14 (2.31) | 39.81 (6.62) |
Boxing | 2.55 (1.43) | 3.81 (1.62) |
Cross over/under | 0.83 (0.39) | 1.03 (0.27) |
Anogenital sniffing* | 11.21 (2.10) | 5.69 (1.22) |
Body Sniffing | 27.21 (2.33) | 27.09 (3.73) |
Allogrooming | 13.50 (5.68) | 3.82 (1.79) |
Rearing | 120.31 (13.32) | 121.48 (12.13) |
Dig/Sniff Bedding | 119.59 (24.45) | 109.15 (21.41) |
1c. Latency (sec) | SAC | PAE |
Wrestling ** | 430.75 (50.51) | 209.98 (51.25) |
Boxing | 569.52 (48.14) | 525.63 (74.75) |
Cross over/under | 544.4 (65.21) | 429.01 (75.78) |
Anogenital sniffing | 107.68 (39.35) | 164.31 (44.09) |
Body Sniffing | 22.77 (6.14) | 16.80 (3.21) |
Allogrooming | 471.44 (70.82) | 588.52 (48.47) |
Rearing | 20.92 (7.65) | 11.94 (1.20) |
Dig/Sniff Bedding | 76.78 (25.78) | 117.66 (44.64) |
Table 1: Mean (+SEM) frequency (1a), duration (1b) and latency to first occurrence (1c) for each behavior quantified during the social interaction session for saccharin- (SAC) and prenatal alcohol-exposed (PAE) rats (n =1 6 per prenatal treatment group). [* p <0.05, ** p <0.005 ]
In addition to performing ANOVAs, performing a linear discriminant analysis to evaluate which variables best discriminate ethanol-exposed from saccharin-exposed animals is recommended18. For the present sample, the Box M’s test to test for equal variances could not be calculated because the number of independent variables was greater than the number of cases (a 5:1 ratio is typically recommended). The discriminant function revealed a significant association between groups and predictors, accounting for 74% of between group variability. An analysis of the structure matrix revealed that duration of wrestling (0.470), and latency to first occurrence of wrestling (-0.330) were significant predictors. Counts (frequency) of wrestling (0.280) and duration of anogenital sniffing (-0.244) were slightly weaker predictors. The cross-validated classification showed that overall, 71.9% of cases were correctly classified.
The prenatal alcohol exposure paradigm described here involves voluntary consumption of ethanol (5% v/v) by rat dams during pregnancy. There are a number of protocols for exposing non-human animals to ethanol during brain development represented in the literature, which differ with respect to the timing, dose, duration and route of ethanol administration as well as the species under investigation. Although a thorough treatment of the advantages of various exposure protocols is not provided here, several advantages of the voluntary drinking method for PAE described in this protocol are highlighted. Previously we utilized a liquid diet protocol, an approach commonly employed in this field of research, in which rat dams consume 5% ethanol as part of the primary food source24. Control conditions for this approach include a pair-fed group in which the caloric intake was yoked to that of ethanol-consuming dams and a group that has ad libitum access to chow. In the voluntary drinking paradigm described here rat dams in both groups (ethanol and saccharin) consume the same rat chow diet which reduces between group variability in nutrition and caloric intake, and minimizes the potential confound of stress associated with forced consumption of an unfamiliar food source25. This feature of the voluntary drinking paradigm also eliminates the need for a pair-fed control as with the liquid diet approach, which provides some practical and ethical benefits, including reductions in the number of experimental groups (from three to two), the number of animals used in the research, and the associated costs to perform the research. The 4 hr intermittent exposure pattern of voluntary ethanol consumption yields less variable drinking levels than observed with the 5% ethanol liquid diet protocol, which might reasonably be expected to similarly diminish variability in outcome measures observed in PAE offspring. Because blood ethanol concentrations achieved with any protocol are important to quantify and communicate, measurements of maternal serum ethanol concentrations should be conducted periodically (e.g., annually) to ensure that comparable BACs are being achieved across breeding rounds. Due to the potential for interactions between prenatal stress and ethanol exposure these measures should be performed in a separate cohort of females for which the offspring are not used in subsequent studies (see ref. 26). It is important to recognize that peak serum ethanol concentrations will occur approximately 45-60 min after the drinking tube is introduced. In lieu of BACs from the dams producing experimental offspring, the most useful measure to correlate with outcome measures is a rat dam’s daily ethanol consumption during pregnancy. However, restricting the range of drinking in rat dams to within one standard deviation of the mean for the group would likely constrain any meaningful correlation between ethanol consumed and a given outcome measure. Finally, evaluation of pre-pregnancy drinking is utilized to identify rats that drink at desired levels for subsequent phases of the drinking protocol. This aspect of the ethanol exposure paradigm also ensures that all female rats have experience drinking prior to pregnancy, which more accurately models human behavior in that drinking is unlikely to begin during pregnancy.
Several additional methodological issues and caveats related to the alcohol exposure paradigm should be considered. The voluntary drinking PAE paradigm described here occurs throughout gestational development, which in the rat corresponds roughly to the first two trimesters of gestational development in humans. Exposure during early postnatal development is utilized by many laboratories to model third trimester human equivalence (see e.g., ref. 27). Further, we note that the procedures presented here represent a chronic exposure protocol (rat dams drink every day). Importantly, precisely timed exposure to higher levels of alcohol (~287 mg/dl) limited to gestational day 15 has also been shown to alter social behavior15. The paradigm presented here involves voluntary alcohol consumption by rat dams, therefore, there is a limited range of blood ethanol concentrations that can be achieved with this approach. Achieving higher blood ethanol concentrations requires other methods of exposure (e.g., gavage, injection, vapor exposure). An additional point of importance concerns potential for alterations in maternal care that could complicate interpretation of behavioral effects associated with prenatal ethanol exposure. To address this, assessments of maternal care during pregnancy should be evaluated periodically. No effects of moderate drinking during pregnancy on maternal care have been detected using the procedures described here in the Long-Evans rat28. Evaluation of whether alterations in behavioral indicators of maternal care (see ref. 29), however, should be evaluated initially and periodically thereafter, particularly if deviations from the methods described here, including species and strain of animal, or higher concentrations of ethanol are utilized.
The behavioral procedures described here have yielded reliable alterations in specific aspects of social behavior (wrestling and investigation) in adult male rats exposed to moderate levels of alcohol during prenatal brain development1,18,30. The behaviors quantified here were selected based on a large body of extant literature31 to target partner-directed behaviors (e.g., wrestling, investigation) and other behaviors directed toward the environment (e.g., rearing, digging) that can be easily measured by way of video analysis. Discriminant analyses revealed that increases in wrestling provide the best discrimination between alcohol-exposed and saccharin-exposed rats among a broad range of social and non-social behavioral variables18. It is important to point out that effects of PAE on social behavior in female rats have not been observed using the methods described here1, although main effects of sex have been reported for several dependent measures including anogenital sniffing (female > male), body sniffing (male > female), wrestling (male > female), and boxing (male > female).
Although wrestling has been shown to be the primary aspect of social behavior altered by moderate PAE, the inclusion of other behaviors is important for establishing the selectivity of the behavioral effects and ruling out generalized behavioral deficits. The set of behaviors quantified here is by no means exhaustive. Behaviors of interest should be selected during preliminary work to capture the overall pattern of effects observed in a given set of data. This is particularly important if different alcohol exposure paradigms, or parameters, are utilized as variations in procedures could reasonably be expected to yield different behavioral outcomes. In addition to the behaviors described here, initially evaluating a broader set of behaviors including biting, scratching (self), full grooming sequences, truncated grooming sequences (indicative of anxiety), “lateral” display32,33, body shakes, chasing, and play behavior34-37 is recommended. Observation in 6-12 pairs of animals should be sufficient to identify behaviors that distinguish alcohol-exposed from non-exposed animals.
It is also important to consider that wrestling, depending upon the age at the time of measurement, could reflect genuine aggressive behavior or play behavior. Early lesions of the ventrolateral frontal cortex in rats, which has been linked to moderate PAE effects on social behavior1,18,30, result in increased play behavior38. In the rat, play behavior peaks during post-weaning development around postnatal days 30-40 33,34 and declines as animals approach adulthood. Alcohol exposure during brain development alters play behavior when measured prior to adulthood6,16,23 and could affect the rate at which play behavior decreases with age. Because the topographies of play and aggressive behaviors are similar a clear distinction can be difficult to achieve. In previous studies, conspicuous behavioral indicators of aggression, such as fighting or biting, have not been observed in moderate PAE rats. However, additional behavioral indicators can provide clues regarding the classification of these behaviors18. For example, attacks directed at the nape of the neck, the primary target of playful attacks, rarely occur in adult PAE rats18. In contrast, attacks directed toward the rump, a target of non-playful attacks39 were observed more frequently, suggesting that PAE-related increases in wrestling reflect aggression rather than play. Play behavior should be included in the analysis if social behavior is measured prior to adulthood or play behavior is conspicuous in the behavior of adult animals. Detailed methods for the analysis of play behavior are described by Himmler et al.35
Finally, it is important to note that social behavior using the methods described here can be influenced by the selection of the partner animal. In the representative results presented here animals were paired with familiar animals (cage-mates) from the same prenatal treatment condition1,18. The rationale for this selection was based largely on data demonstrating that housing control animals with ethanol-exposed rats alters social behavior in control rats14, as well as similar and reliable unpublished observations from our laboratory. These effects can potentially complicate identification of and interpretation of PAE-related alterations in social behavior. The methods described here for quantification of social behavior could, however, be applied to any variation of the source of the social partner, of which there are several possibilities including using a non-treated partner that comes from neither treatment condition1, using a partner animal from the same treatment condition (as described here), or a partner animal from a different treatment condition. Further, the familiarity of the partner animal can be manipulated to affect social interaction1. The selection of the partner condition and other variables related to social housing, sex, and exposure paradigm can be tailored to best meet the scientific goals of the particular laboratory while still utilizing the basic procedures for quantification of social behavior described here.
The authors have nothing to disclose.
Support provided by grant AA019462 to DAH and AA019884 to DDS.
Saccharin sodium salt hydrate | Sigma | S1002 | |
190 proof ethanol | Sigma | 493538 | |
Beaded glass drinking tubes | Fisher | 14-955K | |
Natural rubber white #4 stopper one hole | Plasticoid | LSG4M181 | |
1" bend tubes-ball point | Ancare | TD-199-3" | |
Paper rulers | N/A | N/A | www.vendian.org/mncharity/dir3/paper_rulers |
Apparatus for social interaction | Custom built | N/A | 95 cm X 47 cm X 43 cm |
Video cameras | N/A | N/A | Capable of recording low/no light conditions |
Infrared illuminators | Vitek | VT-IR1-12 | |
Teklad laboratory grade sani-chips | Harlan | 7090A | |
Brush and dustpan | N/A | N/A | |
Isopropyl alcohol | Sigma | W292907 | |
Chlorine Dioxide (1.5 mg Tablets) | Quiplabs | N/A | Prepare per manufacturer's recommendation |