This study was carried out in accordance with the recommendations of the Institutional Animal Care and Use Committee of The George Washington University. A Cc2d1b null mouse line (1b-KO) was generated by the Knockout Mouse Project Repository (Project ID CDS 34981) at the University of California Davis, with the allele Cc2d1b^{tm1a(KOMP)Wtsi}. Cc2d1b null mice carry an engrailed 2 splice acceptor (En2SA) gene-trap allele with bicistronic expression of β-galactosidase as well as a neomycin resistance cassette, flanked by FRT (flippase recombinase target) recombination sites, in the genomic region between exons 2 and 3 of Cc2d1b (Figure 1A). Cc2d1a/1b double heterozygous (1a/1b-dHET) mice were generated by crossing Cc2d1a heterozygotes (1a-HET) with Cc2d1b heterozygotes (1b-HET). 1a-HET mice were bred from a Cc2d1a null mouse line (KO) generated by the Knockout Mouse Project Repository (Project Design ID 49663) at the University of California as was previously described by Oaks et al. (2017). All lines are maintained on a C57BL/6 background. For genotyping, polymerase chain reaction (PCR) amplifications were performed on 1 μL of proteinase K (New England Biolabs, Ipswich, MA, United States) digested tail DNA samples. PCR reactions (50 μL) consisted of GoTaq Flexi buffer (Promega, Madison, WI, United States), 100 μM dNTPs, 50 μM each of forward and reverse primers (sequence available upon request), 1 mM MgCl₂, and 1.25 U GoTaq Flexi DNA polymerase (Promega, Madison WI, United States), and were run with optimized reaction profiles determined for each genotype. A 25-μL aliquot from each reaction was analyzed by gel electrophoresis on a 1.0% agarose gel for the presence of the desired band.

To prepare tissue for histological analysis, deeply anesthetized mice were transcardially perfused with phosphate buffered saline (PBS) followed by 4% paraformaldehyde (PFA). Brains were removed and post-fixed in PFA. Cryosections from adult mouse brains were prepared by mounting in Neg-50 (Thermo Fisher Scientific, Waltham, MA, United States) and cut at 40 μm on a Cryostar NX50 cryostat (Thermo Fisher Scientific, Waltham, MA, United States), then stained with Hematoxylin and Eosin (H&E, VWR International, Radnor, PA, United States) to visualize tissue architecture. Imaging of H&E stained sections was performed on a Leica M165 FC stereo microscope (Leica Microsystems, Buffalo Grove, IL, United States).

A standardized battery of behavioral testing was applied to each cohort of animals, 1b-KO and 1a/1b-dHET male and female mice, at 3–4 months of age. As both 1b-KOs and 1a/1b-dHETs were generated from the same 1a/1b-dHET, breeding the wild-type (WT) controls were littermates shared by both cohorts and all behavioral tests were performed at the same time for WT, 1b-KOs, and 1a/1b-dHETs. Behavioral tests were performed in the Manzini lab behavioral suite in the George Washington University Animal Research Facility following a 60 min period of acclimatization. Initial characterization to analyze any neurological abnormalities including the analysis of basic motor and somatosensory function was performed on a subset of the behavioral cohort as described by Rogers et al. (1997): righting reflex, wire hang, gait analysis, tail pinch, and visual reach. Cognitive and social function and other behaviors were tested in the open field test, novel object recognition test (NORT) (Bevins and Besheer, 2006), Morris water maze (MWM) (Vorhees and Williams, 2006), and 3-chamber social interaction test (Nadler et al., 2004). Behavioral analysis was performed via automated animal tracking using ANY-maze (Stoelting, Wood Dale, IL, United States).

Loss of CC2D1A in humans causes a variable spectrum of ID, ASD, and seizures and the removal of Cc2d1a in the murine forebrain leads to several cognitive, social, and affective behavioral phenotypes (Manzini et al., 2014; Oaks et al., 2017). As no human mutations in CC2D1B have been identified to date, we asked whether loss of Cc2d1b in the mouse would lead to similar phenotypes as loss of Cc2d1a. A Cc2d1b-deficient line (1b-KO) had been generated from the Knockout Mouse Project (KOMP) as a gene-trap allele inserted in intron 2 of Cc2d1b (Figure 1A). We obtained heterozygous animals and bred them to homozygosity, finding that 1b-KO mice are born in Mendelian ratios (Figure 1B). Differently from Cc2d1a KO (1a-KO) pups, which die shortly after birth (Zhao et al., 2011; Al-Tawashi et al., 2012; Drusenheimer et al., 2015; Oaks et al., 2017), 1b-KO mice are viable, fertile, and indistinguishable from WT littermates (Figure 1C). Basic behavioral functions were tested in adult WT and 1b-KO males and females: coordination (righting reflex), strength (wire hang), locomotion (stride and gait), pain sensitivity (tail pinch), and vision (visual reflex). No differences were observed in basic sensory and motor function (Table 1). We confirmed via western blot analysis of cortical protein lysates that CC2D1B was completely absent in these animals and that CC2D1A was expressed at normal levels (Figure 1D). Cryosections generated from the adult brain of 1b-KO animals and stained using hematoxylin and eosin (H&E) showed no differences in brain size and organization from WT littermates (Figure 1E). In summary, loss of Cc2d1b does not affect respiratory function and deglutition in the infant as observed in 1a-KOs, and 1b-KO adult mice are indistinguishable from WT littermates.

CC2D1A and CC2D1B contain very similar protein domains and are thought to have redundant functions in endocytic traffic and gene transcription (Hadjighassem et al., 2009; Usami et al., 2012; Drusenheimer et al., 2015). Because CC2D1B LOF did not result in postnatal lethality, we wondered whether the two proteins would only be partially redundant. To test this hypothesis, we crossed 1b-KOs and 1a-KOs to generate Cc2d1a/Cc2d1b double heterozygous (1a/1b-dHET) and double KO mice (1a/1b-KO). As 1aKO pups die soon after birth (Zhao et al., 2011; Al-Tawashi et al., 2012; Oaks et al., 2017), we did not expect 1a/1b-KO animals to survive and we genotyped litters at postnatal day (P)0, collecting tissue from both live and dead pups. However, while dead 1a-KO and 1a-KO/1b-HET were found in the expected ratios, 1a/1b-dKO pups were never retrieved (Figure 2A), suggesting that double knockouts may die earlier during embryonic development. Examination of prenatal litters only identified 1a/1b-dKO tissue mid-gestation at E11.5, but the embryo was almost entirely absent, leaving only a hypomorphic and largely empty yolk sac (Figure 2B). These results indicate that removal of both CC2D1 proteins leads to early embryonic lethality.

1a/1b-dHETs were viable, fertile, and indistinguishable from WT littermates with normal gross brain anatomy (Figure 2C) and normal basic motor and sensory function (Table 1). We tested the expression levels of CC2D1A and CC2D1B in 1a/1b-dHET mice and found that as expected, only a half dose of each CC2D1 protein was present (Figure 2D). Thus, combined CC2D1 function is necessary for embryonic morphogenesis, but 1b-KO or 1a/1b-dHET animals develop normally, indicating that CC2D1A and CC2D1B have similar functions as it pertains to gross anatomical development and survival.

We have previously found that loss of Cc2d1a leads to a constellation of behavioral deficits: cognitive and social impairment, anxiety, hyperactivity, and repetitive behaviors (Oaks et al., 2017). We generated a cohort of 1b-KO and 1a/1b-HET male and female mice for behavioral analysis by crossing 1a/1b-HETs, so that we could compare behavioral performance in both lines at the same time. In the short-term memory version of the NORT (Bevins and Besheer, 2006) mice are placed in an arena with two identical objects that they are free to explore. After being removed back to their cages for 15 min, they are put in the arena where one of the now known objects has been substituted for a novel object (Figure 3A). In this test, WT male and female mice spend roughly four times longer exploring the novel object, while 1b-KOs and 1a/1b-dHETs show no difference (Figures 3B,C) (Males: WT, T2/T1 = 1.21 ± 0.32, T4/T3 = 3.90 ± 0.75, n = 10, p = 0.004∗∗; 1b-KO, T2/T1 = 1.05 ± 0.24, T4/T3 = 1.60 ± 0.46, n = 11, p = 0.309; 1a/1b-dHET, T2/T1 = 1.08 ± 0.23, T4/T3 = 1.62 ± 0.46, n = 12, p = 0.307. Females: WT, T2/T1 = 1.20 ± 0.25, T4/T3 = 4.39 ± 1.40, n = 10, p = 0.038 ∗; 1b-KO, T2/T1 = 0.84 ± 0.16, T4/T3 = 0.93 ± 0.24, n = 10, p = 0.757; 1a/1b-dHET, T2/T1 = 1.34 ± 0.48, T4/T3 = 1.46 ± 0.28, n = 10, p = 0.824). This deficit was not due to reduced interest in the objects, as animals spent similar amounts of time in exploratory behaviors, with 1a/1b-dHET males showing significantly more exploration (Figure 3D T1+T2 – Males: WT, t = 26.97 ± 5.75s, n = 10; 1b-KO, t = 23.17 ± 3.65s, n = 11, p = 0.999; 1a/1b-dHET, t = 65.27 ± 15.93s, n = 12, p = 0.167; Females: WT, t = 40.56 ± 5.19s, n = 10; 1b-KO, t = 71.67 ± 17.47s, n = 10, p = 0.423; 1a/1b-dHET, t = 54.30 ± 10.56s, n = 10, p = 0.960. Figure 3E T3+T4 – Males: WT, t = 21.93 ± 5.54s; 1b-KO, t = 17.91 ± 3.57s, p = 0.999; 1a/1b-dHET, t = 50.83 ± 16.0s, p = 0.640. Females: WT, t = 15.39 ± 2.12s; 1b-KO, t = 68.38 ± 26.04s, p = 0.090; 1a/1b-dHET, t = 31.86 ± 10.61s, p = 0.959. Figure 3F SUM T1,2,3,4 – Males: WT, t = 48.90 ± 9.35s; 1b-KO, t = 41.08 ± 6.20s, p = 0.942; 1a/1b-dHET, t = 116.1 ± 28.24s, p = 0.033 ∗; Females: WT, t = 55.95 ± 6.62s, 1b-KO; t = 140.1 ± 42.64s, p = 0.073; 1a/1b-dHET, t = 86.16 ± 20.52s, p = 0.660).

To further assess cognitive function, the 1b-KO mice were tested using the MWM paradigm which probes spatial memory acquisition, retention, and flexibility, by testing the ability of a mouse to learn, remember, and relearn the location of a platform hidden under opaque water (Morris, 1984). After the mice are trained using a visible platform to escape from the water, the platform is hidden under the surface in a different location and the animals undergo training on five consecutive days to learn the location of the platform. On the following day, memory retention is tested by removing the platform and measuring the amount of time the mouse spends in the area where the platform was previously located (probe trial). Finally, the position of the platform is changed and the animal must display flexibility by learning a new location (reversal). 1a-cKO animals show a delay in initial acquisition of the location of the HP, but after they learn, they can retain the memory in the probe trial, and learn a new location in the reversal (Oaks et al., 2017). Both 1b-KO and 1a/1b-dHET males and females presented deficits in this test (Figure 4). 1b-KO males and females and 1a/1b-dHET males were delayed in the HP acquisition showing significant differences in day 2 or 3 of the test (HP2 and HP3 in Figures 4B,F) (Males HP3: WT, t = 6.82 ± 0.69s, n = 11; 1b-KO, t = 10.97 ± 1.85s, n = 10, p = 0.042 ∗; 1a/1b-dHET, t = 11.99 ± 1.28s, n = 13, p = 0.0027 ∗∗. Females HP2: WT, t = 12.30 ± 1.32s, n = 13; 1b-KO, t = 19.62 ± 1.74s, n = 10, p = 0.0025 ∗∗; 1a/1b-dHET, t = 14.66 ± 1.64s, n = 11, p = 0.247). 1a/1bHET males and females were also affected in the probe trial where they spent less time in the platform quadrant during the first 15 s of the 60-s trial (Figures 4D,H) (Probe 15 s – Males: WT, t = 9.51 ± 0.83s, n = 11; 1b-KO, t = 6.13 ± 0.50s, n = 10, p = 0.0029 ∗∗; 1a/1b-dHET, t = 5.48 ± 0.80s, n = 13, p = 0.0021 ∗∗. Females: WT, t = 7.18 ± 0.80s, n = 13; 1b-KO, t = 5.77 ± 0.65s, n = 10, p = 0.203; 1a/1b-dHET, t = 4.36 ± 0.82s, n = 11, p = 0.022 ∗). Finally, 1b-KO males, but not females, were affected throughout the 60-s probe trial and spent less time exploring the correct quadrant in the probe trial testing memory retention (Figure 4D) (Probe 60 s – Males: WT, t = 25.40 ± 1.78s, n = 11; 1b-KO, t = 19.58 ± 1.30s, n = 10, p = 0.018 ∗; 1a/1b-dHET, t = 22.74 ± 2.63s, n = 13, p = 0.428. Females: WT, t = 21.19 ± 1.85s, n = 13; 1b-KO, t = 20.57 ± 1.54s, n = 10, p = 0.809; 1a/1b-dHET, t = 18.43 ± 2.62s, n = 11, p = 0.389). Animals heterozygous for loss of Cc2d1a or Cc2d1b alone showed normal behavioral performance (Supplementary Figures 1, 2 and Supplementary Table 1). In summary, loss of CC2D1B leads to cognitive deficits in both memory acquisition and retention. In general, males appear more severely affected than females in both 1bKO and 1a/1bHET lines, suggesting that CC2D1A and CC2D1B have overlapping roles in cognitive function.

1A-cKO animals showed increased mobility and reduced entry into the center of the open field arena, indicating hyperactivity and anxiety (Oaks et al., 2017). In addition, removal of Cc2d1a in the forebrain also leads to ulcerative dermatitis due to obsessive grooming and social interaction deficits (Oaks et al., 2017). 1b-KO males and females performed similarly to WT littermates in the open field test and showed no signs of hyperactivity or anxiety (Figure 5) (Distance – Males: WT, d = 25.16 ± 2.29m, n = 11; 1b-KO, d = 29.63 ± 1.96m, n = 11, p = 0.498; Females: WT, d = 34.65 ± 1.36m, n = 13; 1b-KO, d = 42.37 ± 3.28m, n = 11, p = 0.097. Time in center – Males: WT, t = 78.13 ± 5.23s, n = 11; 1b-KO, t = 83.17 ± 14.26s, n = 11, p = 0.988; Females: WT, t = 77.45 ± 11.78s, n = 10; 1b-KO, t = 87.75 ± 17.65s, n = 10, p = 0.969). Interestingly, 1a/1b-dHETs showed increased locomotion and avoidance of open spaces, as previously observed for the 1a-cKOs, but only in males, similar to the exploration in the NORT where increased exploratory behavior was only observed in 1a/1b-dHET males (Figures 5A,B) (Distance – Males: WT, d = 25.16 ± 2.29m, n = 11; 1a/1b-dHET, d = 35.85 ± 2.94m, n = 13, p = 0.0076 ∗∗; Females: WT, d = 34.65 ± 1.36m, n = 13; 1a/1b-dHET, d = 35.35 ± 2.51m, n = 11, p = 0.999. Time in center – Males: WT, t = 78.13 ± 5.23s, n = 11; 1a/1b-dHET, t = 41.32 ± 3.71s, n = 13, p = 0.0198 ∗; Females: WT, t = 77.45 ± 11.78s, n = 10; 1b-KO, t = 121.90 ± 15.19s, n = 11, p = 0.1225). No ulcerative dermatitis or obsessive grooming was observed in any of these mouse lines.

Finally, all mice were tested in the social approach version of the three-chambered test. In this test, the mouse is placed in an apparatus with three communicating chambers. In the left chamber, there is a novel mouse of the same sex under a wire cup, while in the right chamber there is an empty wire cup. Mice spend more time exploring and sniffing the stranger mouse than the object and this is considered a social action (Nadler et al., 2004; Kaidanovich-Beilin et al., 2011). The 1a-cKO showed no preference for the conspecific both as in the time spent around the mouse enclosure and the time spent sniffing the stranger mouse (Oaks et al., 2017). 1a/1b-dHET males and females and 1b-KO females behaved like WT mice in this test (Figures 5E–H). 1b-KO males were moderately affected showing non-significant difference between the empty cup and the stranger (Figure 5E) [Males: WT, time with mouse (tm) = 287.65 ± 26.81s, time with object (to) = 162.74 ± 18.15s, n = 11, p = 0.00098 ∗∗∗; 1b-KO, tm = 282.37 ± 34.83s, to = 187.98 ± 28.63s, n = 13, p = 0.082; 1a/1b-dHET, tm = 312.05 ± 39.03s, to = 159.39 ± 28.11s, n = 10, p = 0.0052 ∗∗. Females: WT, tm = 331.50 ± 19.14s, to = 152.70 ± 31.59s, n = 8, p = 0.00026 ∗∗∗; 1b-KO, tm = 362.93 ± 29.06s, to = 151.53 ± 29.39s, n = 8, p = 0.00016 ∗∗∗; 1a/1b-dHET, tm = 317.62 ± 20.89s, to = 172.47 ± 12.91s, n = 9, p = 0.00002 ∗∗∗]. The deficit in 1b-KO males was primarily due to a subset of animals showing preference for the object (Supplementary Figure 3). All genotypes showed significantly increased time spent sniffing the stranger mouse, indicating that once in the chamber the 1b-KO animals interact with the other animal (Figures 5F,H) [Males: WT, time sniffing mouse (tsm) = 66.47 ± 7.44s, time sniffing object (tso) = 36.61 ± 7.51s, n = 11, p = 0.0105 ∗; 1b-KO, tsm = 56.04 ± 10.78s, tso = 21.47 ± 5.78s, n = 13, p = 0.009 ∗∗; 1a/1b-dHET, tsm = 58.40 ± 8.65s, tso = 31.11 ± 8.95s, n = 10, p = 0.042 ∗. Females: WT, tsm = 60.11 ± 10.60s, tso = 31.15 ± 7.71s, n = 8, p = 0.044 ∗; 1b-KO, tsm = 96.68 ± 13.00s, tso = 29.93 ± 5.55s, n = 8, p = 0.00033 ∗∗∗; 1a/1b-dHET, tsm = 55.80 ± 5.66s, to = 18.26 ± 4.02s, n = 9, p = 0.00005 ∗∗∗].

In conclusion, 1b-KO and 1a/1b-dHET animals show only partially overlapping behavioral profiles in anxiety, hyperactivity, and sociability. 1b-KO mice of either sex do not appear anxious or hyperactive and only males show a mild sociability deficit in the three-chamber test. 1a/1b-dHET males are more similar to 1a-cKO mice, with increased locomotion and decreased time in the center of the open field. These results show that CC2D1A and CC2D1B only have partially redundant roles in cognitive and social function. Each of the Cc2d1 genes contributes to aspects of learning and memory and sociability, but Cc2d1a appears to be more critical for hyperactivity and anxiety. Interestingly, both lines display sexually dimorphic phenotypes with males being mildly more affected than females.