We evaluated Gabra1 KO mice in a standardized phenotyping battery. This comprehensive phenotypic analysis also included a set of behavioral tests derived from the Irwin screen to determine the gross sensory and motor deficits (17). As part of this screen general physical health and empty cage spontaneous behaviors were observed by placing the mouse in an empty home cage for 60 s. Afterwards, the mouse was lifted by the tail and observed for limb splaying and forepaw and hindpaw clutching for 20 s.

Our KO mice were smaller in body weight and length, but viable and fertile. There were no differences between the KO and WT control mice in motor coordination, sensorimotor gating, depressive-like behavior, or learning and memory as assessed by inverted screen, pre-pulse inhibition, tail suspension, and trace fear conditioning assays, respectively (Table 4.2).

In the Irwin-derived initial behavioral screen, we observed tremor in Gabra1 KO mice as they moved freely about their cages, consistent with previous reports (9, 18). Despite the tremor, Gabra1 KO performed normally in an inverted screen assay that assesses motor strength and coordination (Table 4.2). In other reports, Gabra1 KO mice were reported as normal in swimming and balance beam tests (9). Variable results have been reported for performance on the rotarod (see Table 4.3 for comparison of phenotypes and corresponding references). Kralic et al. reported that Gabra1 KO mice fell off sooner in the accelerating rotarod test (18), while there was no effect of genotype when the speed was held constant within trials (with a stepwise increase in speed over trials) (9).

Several phenotypes were detected in assays with predictive validity for anxiolytic compounds, namely the marble burying (19–21), platform (22), and open field tests (23–26). One of the strongest phenotypes observed in the first cohort of mice was in the marble burying test. The marble burying (MB) test for anxiety-related and compulsive-like behaviors utilizes novelty-induced digging to assess the behavioral state of the subject and was performed as previously described in (27). Briefly, mice were individually placed into cages filled to a depth of 5 cm with clean bedding and 25 identical marbles evenly spaced across the bedding surface. After 30-min testing, each mouse was returned to its home cage and marbles covered by bedding 2/3 or more were scored as buried by an experimenter blind to genotype. There were no effects of sex or genotype by sex interaction; therefore, an unpaired t-test was used for all statistical analysis of MB. As seen in the Table 4.2, initial testing in MB revealed that Gabra1 KO mice buried less marbles when compared with WT cohort mates (t55  =  7.0; P  <  0.001, Fig. 4.2a). Follow-up testing on a new cohort of Gabra1 mice produced results similar to the initial screening (Fig. 4.2b); Gabra1 KO again buried significantly less marbles than WT mice (t37  =  4.3; P  <  0.001). The marble-burying assay is known to detect the activity of antidepressant and anxiolytic compounds, but does not always reveal a therapeutic direction (19–21, 28). Therefore, the significantly decreased digging behavior that we observed only provides an indication of altered anxiety behavior.

We evaluated locomotor activity of Gabra1 KO and WT mice in the open field (AccuScan Instruments Inc., Columbus, OH, USA; 42  ×  42  ×  30 cm³) as described in (27, 29). Center time and center-to-total distance ratio have been used as indicators of unconditioned (trait) anxiety (23–26). Published studies of Gabra1 KO mice have reported inconsistent effects using these measures (Table 4.3). In our initial characterization of the Gabra1 KO, we found decreases in locomotor activity, center measures (center time and center to total distance ratio) and exploratory behaviors (rearing and hole poke) in the open field (Table 4.2).

Having observed altered anxiety-like phenotypes in both MB and OF, we followed up with a more specific anxiety assay, the platform test. The platform test (an analog of the light/dark test) was performed as described in details in (22). In brief, a platform was positioned in the back right corner of the Digiscan open-field apparatus (AccuScan Instruments Inc., Columbus, OH, USA; 42  ×  42  ×  30 cm³). The animals were placed under the platform. Duration and distance in specific areas (under or outside the platform) were calculated using Versamap software (Accuscan Instruments). The platform test was conducted on a new cohort prior to the mice having any testing experience. The main measures of anxiety behavior in the platform test are time in the lighted area, i.e., outside of platform shelter, and the ratio of lighted to total area distance traveled (ld/td). Male Gabra1 KO mice spent less time in the lighted area (Fig. 4.3a) compared to male WT mice (t ₃₈  =  3.0; P  <  0.01). Male Gabra1 KO mice also exhibited a decreased ld/td ratio relative to male WT mice (t ₃₈  =  4.0; P  <  0.001, Fig. 4.3b). These findings of decreased time and distance spent in the lighted region in male Gabra1 KO mice are indicative of an anxiogenic phenotype. There was a nonsignificant trend in female Gabra1 KO mice toward reduced time spent on the lighted side when compared with female WT mice (t ₃₇  =  1.5; P  =  0.13, Fig. 4.3c). There was a nearly significant trend for decreased ld/td ratio in female Gabra1 KO mice (t ₃₇  =  1.96; P  =  0.05, Fig. 4.3d). Overall, Gabra1 deletion induced an anxiogenic-like phenotype in both male and female mutant mice in the platform test.

The platform, open field, and marble burying tests are measures of unconditioned anxiety. We also tested mice in fear-potentiated startle (FPS), a test of conditioned fear and anxiety that is sensitive to amygdala damage and GABA mediated drugs (30). In the FPS assay, a fear reaction to a discrete conditioned stimulus is measured rather than a response to an unconditioned fear-provoking unfamiliar environment as in exploration-based paradigms like the plus-maze or light-dark tests. Previously, we have reported a dose-dependent decrease in FPS with the GABA_A agonist chlordiazepoxide (31). An advantage of the FPS test is that it is not influenced by locomotor activity. The FPS test allowed us to examine if Gabra1 KO mice exhibited alterations in learned fear, often referred to as state anxiety, as opposed to unconditioned anxiety, commonly referred to as trait anxiety (32). Fear ­potentiated startle (FPS) was assessed using acoustic startle chambers (SR-LAB, San Diego Instruments, San Diego, CA, USA) for mice as described in (31). There were no significant differences due to genotype in FPS. Importantly, both WT and KO mice exhibited significantly higher startle in tone trials relative to no-tone trials, (F (1,144)  =  8.6, P  <  0.01, Fig. 4.4). This indicates that both WT and KO mice acquired conditioned fear to the tone stimulus and suggests that the conditioned fear is not affected by genetic disruption of Gabra1. Interestingly, Gabra1 KO mice exhibited significantly increased startle reactivity (F (1, 144)  =  10.5, P  <  0.01) in FPS. We also observed significantly increased startle reactivity in Gabra1 KO in PPI testing (Table 4.2). Additional evidence of Gabra1 involvement in startle behavior was reported by Homanics et al., who observed increased startle reactivity in Gabra1 homozygous knock-in mice (33). Startle reactivity has been used as a measure of unconditioned anxiety (34). Therefore, this result further substantiates the findings in marble-burying, platform, and open-field assays.

Our results show that Gabra1 KO mice are not impaired in FPS, an assay with a learning component, suggesting that this subunit is likely not critical for learned anxiety or classical fear conditioning. We also have not observed differences between genotypes in trace fear conditioning. Therefore, while it is unclear the specific conditions Gabra1 may modulate anxiety-mediated learning and memory, our study strongly suggests that Gabra1 is particularly crucial for trait anxiety and not important for conditioned anxiety. It is interesting to note that Dixon et al. (2008) report that genetic disruption of Gabra2 impairs ­conditioned anxiety (35).

Circadian influence on the GABA system has been well documented in mice and men (36, 37). We have followed this up using two additional cohorts. We tested mice in the open field at 10 am and 7 pm to determine the baseline differences between genotypes as well as circadian influences on activity and anxiety-related behaviors. When mice were tested during our normal testing hours (10  am), there was a significant effect of genotype on center time [F (1, 70)  =  30.5, P  <  0.001], and center to total distance ratio [F (1,70)  =  29.9, P  <  0.001] with decreases in these measures in KO mice (Fig. 4.5). The effect of genotype on total distance approached, but did not reach significance [F (1, 70)  =  3.9, P  =  0.051]. When tested at the beginning of the dark period (7  pm), there were significant effects of genotype on center time [F (1, 48)  =  20.6, P  <  0.001], center to total distance ratio [F (1, 48)  =  31.4, P  <  0.001], and total distance [F (1, 48)  =  43.3, P  <  0.001] with all measures decreased in the KO.