Methods

Animals

Asic1a_–/–_ mice, were generated as previously described (Wemmie et al. 2002). Mice were maintained on a congenic C57BL/6 J background. Mouse ages at the start of the fear-conditioning experiment ranged from P17 to P83. The developmental stages tested were pre-weaning (P17), early post-weaning (P23), juvenile and adolescent (P29, P42) and adult (P83). For pre-weaning fear-conditioning experiments, littermate controls (Asic1a_+/+_ and Asic1a_-/-) were generated by Asic1a+/-_ x Asic1a_+/-_ breeding. For post-weaning fear conditioning experiments mice were generated by Asic1a_–/–_ x Asic1_–/–_ and Asic1a_+/+_ x Asic1a_+/+_ breeding. Mice were harem-bred, and dams were separated before giving birth. Distinct sets of mice were used for each age cohort. All mice were included in all experimental phases of their respective age groups. Numbers of mice and litters used can be found in the legend of Fig. 1 and in Supplementary Table 1. Mice of both sexes were used for experiments, and as detailed in the statistical analysis section, we controlled for potential sex effects and tested for potential sex-by-genotype interactions. Mice were given ad libitum access to standard chow (7913, Teklad Standard Irradiated Diet, Inotiv Teklad Madison, WI, USA) and water. They were kept at a 12-hour light-dark cycle in a temperature- and humidity-controlled facility. Weaning took place between P21 and P25. Animal care met the standards set by the National Institutes of Health and all experiments were approved by the University of Iowa Animal Care and Use Committee.

Fig. 1: ASIC1A disruption impairs fear memory acquisition at all ages. Freezing at baseline and during fear memory acquisition in Asic1a knockout (-/-, blue) vs. wild-type (+/+, red) mice trained at: P17 (A); P23 (B); P29 (C); P42 (D); and P83 (E). N for P17: 20 wild-type (8 m, 12 f) and 21 knockout (13 m, 8 f) mice. N for P 23: 28 wild-type (14 m, 14 f) and 17 knockout (11 m, 6 f) mice. N for P29: 18 wild-type (10 m, 8 f) and 15 knockout (7 m, 8 f) mice. N for P42: 14 wild-type (7 m, 7 f) and 20 knockout (10 m, 10 f) mice. N for P83: 14 wild-type (7 m, 7 f) and 19 knockout (11 m, 8 f) mice. Baseline p-value in A-E) is based on square root transformation.*p < 0.05, **p < 0.005, ***p < 0.001.

Fear conditioning

Fear conditioning took place in Med Associates (Fairfax, VT) fear conditioning chambers. Motion thresholds of the VideoFreeze™ software (Med Associates) were adjusted to mouse age based on hand scoring of a representative group of mice (see Supplementary Table 2). We used automated scoring by the Video Freeze software (Med Associates) to assess freezing and maximum motion during the first foot shock (shock reactivity).

All experiments took place during the light cycle. Mice were given at least 30 min of habituation in their home cage after transfer to the experimental area. At postnatal day (P) 17, 23, 29, 42 and 83, we conditioned separate groups of mice. For visualization of the timeline, see Supplementary Fig. 1.

Mice were trained (context A) on day 0 (D0). 1% bleach was used as contextual odor. Pre-weaning mice were habituated to the chamber for 5 min in the darkness, followed by 3 min in the light and post-weaning mice were habituated in the light for 3 min. The additional habituation time in pre-weaning mice was based on initial experiments (data not shown) which demonstrated non-specific freezing without those additional 5 min. Freezing during the habituation period in the light is referred to as baseline freezing in the analysis and discussion. Following habituation, a series of five tones (20 s, 90 db, 3000 Hz), co-terminating with a 1-second foot shock (0.75 mA) were presented. The inter-trial interval between tones was 120 s. The last tone presentation was followed by 80 s without tone or shock. The period from the onset of the first tone through the end of the last 80 s post-shock interval (i.e., the one after the fifth tone-shock) is referred to as acquisition or fear memory acquisition in the analysis and discussion. Cued fear memory recall was probed on day 1 (D1, 24 h after training) in a novel context (context B) with altered lighting, floor texture and odor (peppermint). After habituation (5 min in the dark and 3 min with light for pre-weaning mice and 3 min with light for post-weaning mice), a 3-minute 90 db, 3000 Hz tone was played, followed by additional 4 min without tones or foot shocks. Freezing while the tone (cue) was presented is referred to as freezing during cue presentation or cue-evoked freezing and corresponds to cue-evoked fear memory. Cued fear memory maintenance (context B) was tested on day 15 (D15) with the same protocol as used on day 1. Contextual fear memory recall was tested on day 2 (D2, 48 h after training) by 5-minute exposure to the training context (context A), regardless of age at training, because initial experiments with pre-weaning mice (data not shown) suggested that the duration of context exposure did not affect the extent of context freezing. On day 16 (D16), contextual fear memory maintenance was tested with the same context (context A) and protocol used on day 2.

Statistical analysis

To probe for genotype-specific behavioral differences within the same age group and for developmental effects on behavior, we used mixed effects linear regression analysis. In all mixed-effects models, we controlled for potential sibling correlations by including a random effect for litter. We controlled for potential sex effects by including fixed covariate effects for sex and genotype-by-sex interactions. Data are presented as mean ± standard deviations of the raw data. The reported p values are provided by the regression model. In the case of highly skewed distributions – as determined by our statistician, Dr. Langbehn – group comparisons were performed using the square root of the scores. Follow-up models tested for significant genotype differences in behavioral trends over time. As above, mixed-effect models with random effects were used. The fixed effect predictors were genotype, test day, their interaction (the main parameter of interest), and sex.

Given that there was little numerical difference between baseline freezing at day zero (training day) and freezing upon re-exposure to the training context at 48 h after training we performed follow up-analysis to probe if mice of each age group, genotype and sex retained memory about the training context. The variance within groups tended to be smaller at baseline testing than on follow-up. This was especially notable for mouse ages P29 and beyond. To account for this as well as for within-mouse and within-litter associations, we fit linear models that combined random effects per mouse and within-litter variance and correlation that varied by testing day. The test-day-specific within-litter effects were estimated by iterative weighted least squares. The model fixed effects saturated the possible combinations of genotype, sex, and test day, and the learning comparisons of interest were estimated by the corresponding linear contrasts of those effects. Statistical analysis was performed using R, version 4, using the libraries lmerTest (v3.1), lme4 (v1.1) and emmeans (v1.1) and SAS (Version 9.4 with STAT version 15.3) Proc Mixed. For all statistical analysis a p-value < 0.050 was considered statistically significant and a p-value < 0.100 was regarded as statistical trend. Graphs represent raw data. The error bars in the graphs represent standard errors of the means of the raw data.