Global knockdown of glutamate decarboxylase 67 elicits emotional and auditory 1 abnormalities in mice 2

Reduced expression of glutamate decarboxylase 67 (GAD67), encoded by the Gad1 gene, is a consistent finding 33 in postmortem brains of patients with several psychiatric disorders, including schizophrenia, bipolar disorder and 34 major depressive disorder. The dysfunction of GAD67 in the brain is implicated in the pathophysiology of these 35 psychiatric disorders; however, the neurobiological consequences of GAD67 dysfunction in mature brains are not 36 fully understood because the homozygous Gad1 knockout is lethal in newborn mice. We hypothesized that the 37 tetracycline-controlled gene expression/suppression system could be applied to develop global GAD67 38 knockdown mice that would survive into adulthood. In addition, GAD67 knockdown mice would provide new 39 insights into the neurobiological impact of GAD67 dysfunction. Here, we developed Gad1 tTA/STOP-tetO biallelic 40 knock-in mice using Gad1 STOP-tetO and Gad1 tTA knock-in mice, and compared them with Gad1 +/+ mice. The 41 expression level of GAD67 protein in brains of Gad1 tTA/STOP-tetO mice treated with doxycycline (Dox) was 42 decreased by approximately 90%. The GABA content was also decreased in the brains of Dox-treated 43 Gad1 tTA/STOP-tetO mice. In the open-field test, Dox-treated Gad1 tTA/STOP-tetO mice exhibited hyper-locomotor activity 44 and decreased duration spent in the center region. In addition, acoustic startle responses were impaired in 45 Dox-treated Gad1 tTA/STOP-tetO mice. These results suggest that global reduction in GAD67 elicits emotional and 46 auditory abnormalities in mice. These GAD67 knockdown mice will be useful for elucidating the neurobiological 47 mechanisms of emotional abnormalities, such as anxiety symptoms associated with psychiatric disorders. can be used as GAD67 knockdown mice for studying the behavioral and neurochemical consequences of the global loss of GAD67 in mature brains. In this study, we successfully developed Gad1 STOP-tetO knock-in mice and 85 the subsequent Gad1 tTA/STOP-tetO mice. Herein, we report the behavioral abnormalities elicited by the global 86 knockdown of GAD67 in mice.


Immunoblot analysis 153
The mice were killed by decapitation. The brain hemispheres of neonates and the FCX, HIP and CER of 154 adult mice were quickly dissected on an ice-cold stainless plate. The tissues were immediately frozen in liquid 155 nitrogen and stored at -80°C until use. The frozen tissues were homogenized in ice-cold buffered sucrose (0.32 M) 156 solution containing 20 mM Tris-HCl (pH 7.5) and protease inhibitor cocktail (P8340, Sigma-Aldrich, Inc.). The 157 homogenates were centrifuged at 1,000 g for 10 min at 4°C, and the supernatants were collected as the protein 158 samples. The protein concentrations were determined using a TaKaRa BCA Protein Assay Kit (T9300A, Takara 159 Bio Inc., Japan). 160 The protein samples were diluted with electrophoresis sample buffer. Proteins (1.5 μg) were separated by 8% 161 SDS-polyacrylamide gels and transferred to a PVDF membrane. Blots were probed with respective antibodies to 162 GAD65/67 (1:1,000, rabbit polyclonal antibody) [29] and GAD67 (1:1,000, mouse monoclonal antibody, 163 Millipore, Code No. MAB5406). Immunoblots were developed using horseradish peroxidase-conjugated 164 secondary antibodies (GE Healthcare) and then detected with chemiluminescence reagents (ECL prime, GE 165 Healthcare) and visualized by the Light Capture AE-9672 (ATTO Co., Ltd.). After the detection of immunoblots, 166 the blotting membranes were washed with PBS several times and reprobed with a mouse monoclonal antibody to 167 β-actin (1:10,000, Medical & Biological Laboratories Co. Ltd., Code No. M177-3). The immunoblots of β-actin 168 11 were developed and visualized by the same protocol described above. The density of the bands was determined 169 using ImageJ software. The band densities of β-actin were used as the loading control. The

Motor coordination test 194
The performance of motor coordination in mice was tested by a rotarod apparatus (Ugo Basile, Comerio,195 Italy) according to a previous report [16]. Briefly, each mouse was placed in a separate lane of the apparatus on a 196 rotating cylinder (3 cm diameter) at 20 rounds per minute. The latency until the mouse fell from the cylinder (up 197 to 120 s) was recorded in three consecutive trials with 2 -3 min intervals, and the median latency was used for the 198 following analysis. If the mouse did not fall within 120 s, the latency to fall was recorded as 120 s. 199

Open-field test 201
Each mouse was placed in the center of an open-field apparatus (50 cm × 50 cm × 40 (H) cm) that was 202 13 illuminated by light-emitting diodes (30 lux at the center of the field) and allowed to move freely for 5 min. The 203 data were collected and analyzed using ImageJ OF4 (O'Hara & Co., Ltd., Tokyo, Japan), which is modified 204 software that is also based on the public domain ImageJ program. The procedure was performed according to our 205 previous report [ 110, and 120 dB) white noise stimuli (40 ms) were presented in quasi-random order and random intertrial 211 intervals (10 -20 s). In the PPI session, mice experienced five trial types: no stimulus; startle stimulus (120 dB, 40 212 ms) only; prepulse 70 dB (20 ms, lead time 100 ms) and pulse 120 dB; prepulse 75 dB (20 ms, lead time 100 ms) 213 and pulse 120 dB; and prepulse 80 dB (20 ms, lead time 100 ms) and pulse 120 dB. Each trial was repeated 10 214 times in quasi-random order and random intertrial intervals (10 -20 s). PPI was defined as the percent decline of 215 the startle response: 100 -[(startle amplitude after prepulse and pulse)/(startle amplitude after pulse only)] × 100. 216 The procedure was performed according to our previous report [19].

Development of GAD67 knockdown mice 262
We then crossed heterozygous male Gad1 tTA/+ mice and female Gad1 STOP-tetO/+ mice to generate 263 Gad1 tTA/STOP-tetO biallelic knock-in mice. Mice with four genotypes (Gad1 +/+ , Gad1 tTA/+ , Gad1 STOP-tetO/+ , 264 Gad1 tTA/STOP-tetO ) were born at the expected Mendelian frequency (Fig. 3 legend). Palate formation in mouse 265 neonates with four genotypes was observed, but none of them exhibited the cleft palate (Fig. 3a). Next, the 266 survival rates of these mice were examined in a small population. All Gad1 +/+ mice (total n = 15) survived until 8 267 weeks of age (Fig. 3b). Two Gad1 tTA/+ mice (total n = 15), one Gad1 STOP-tetO/+ mouse (total n = 17), and eight 268 Gad1 tTA/STOP-tetO mice (total n = 20) died within 8 weeks after birth (Fig. 3b). The proportion of genotypes in our 269 breeding colony (total n = 1,319) at the weaning period (P21 -P28) is shown in Fig. 3c (Figs. 4b -4d). Importantly, the expression level of GAD67 protein in the CER of Gad1 tTA/STOP-tetO mice was 282 markedly decreased in the presence of Dox compared with in the absence of Dox (Fig. 4d). By 283 immunofluorescence, GAD67 immunoreactivity in Gad1 +/+ mice was detected widely in the brain, particulary at 284 high levels in the olfactory bulb, globus pallidum, olfactory tubercle, substantia nigra, superior and inferior 285 colliculi, and deep cerebellar nuclei (Fig. 4e, upper panel). In brains of Gad1 tTA/STOP-tetO mice, the overall 286 immunoreactivity was reduced moderately without Dox treatment and severely with the treatment. Dox treatment 287 to Gad1 +/+ mice did not affect GAD67 immunoreactivity (data not shown). PV is expressed in a major subclass of 288 GAD67-positive inhibitory neurons [32]. No discernible changes in PV immunoreactivity were found between 289 Gad1 +/+ mice and Gad1 tTA/STOP-tetO mice with or without Dox treatment (Fig. 4e, lower panel). We assessed 3 290 independent mice in the respective groups and observed similar findings. These results suggest that Dox treatment 291 globally suppresses the expression of GAD67 in the brains of Gad1 tTA/STOP-tetO mice. 292 293

Behavioral abnormalities in GAD67 knockdown mice 294
We compared the behavioral phenotypes of Gad1 tTA/STOP-tetO mice with those of Gad1 +/+ mice in the presence 295 of Dox treatment. To avoid the effects of sex differences, male mice were only used in the following experiments. 296 We first investigated the body weights and the performance of motor coordination in Dox-treated 297 Gad1 tTA/STOP-tetO and Gad1 +/+ mice. No difference was observed in the body weights (t(17) = 1.066, p = 0.301, 298 Student's t-test, Fig. 5a) or the latency to fall from the cylinder in the rotarod test (t(17) = 0.772, p = 0.451, 299 Student's t-test, Fig. 5b) between the two genotypes. 300 We next conducted the open-field test, which is a well-accepted behavioral test to evaluate the anxiety-like 301 state of rodents [33]. The total distance, total duration of movement, moving speed, distance per movement, and 302 duration per movement were significantly increased in Dox-treated Gad1 tTA/STOP-tetO mice compared with 303 Dox-treated Gad1 +/+ mice (Table 1). In contrast, the total number of movement episodes was significantly 304 decreased in Dox-treated Gad1 tTA/STOP-tetO mice compared with Dox-treated Gad1 +/+ mice (Table 1). These 305 observations indicate that Gad1 tTA/STOP-tetO mice walk longer distances with less frequency. In addition, 306 Dox-treated Gad1 tTA/STOP-tetO mice walked a long time in the wall side and a short time in the center region 307 compared with Dox-treated Gad1 +/+ mice (Table 1 and Fig. 5c) (Fig. 5d). The PPI responses were significantly affected by the 315 effect of prepulse intensity (F(2,42) = 6.713, p = 0.003, two-way ANOVA) but not by the effect of genotype × 316 prepulse intensity interaction (F(2,42) = 0.577, p = 0.566, two-way ANOVA) or genotype (F(1,21) = 0.514, p = 317 0.481, two-way ANOVA) (Fig. 5e). 318 319 Brain GABA and glutamate content in GAD67 knockdown mice 320 The GABA content was significantly lower in Dox-treated Gad1 tTA/STOP-tetO mice than in Gad1 +/+ mice in the 321 FCX, HIP and CER (Table 2). On the other hand, the glutamate content in the respective brain regions was 322 20 comparable between these genotypes (Table 2). 323 324 Discussion 325 We first generated Gad1 STOP-tetO knock-in mice. All homozygous Gad1 STOP-tetO/STOP-tetO mice died on the day of 326 birth, and 57% of Gad1 STOP-tetO/STOP-tetO mice exhibited a cleft palate. The expression of GAD67 protein was 327 lacking in the brains of Gad1 STOP-tetO/STOP-tetO mice with or without the cleft palate. These phenotypes in 328 Gad1 STOP-tetO/STOP-tetO mice are consistent with those in Gad1 -/mice [15,31]. Therefore, the function of the Gad1 329 gene was eliminated by the insertion of the Neo-STOP-tetO cassette in the 5'-untranslated region of the Gad1 330 gene in mice. It has been reported that neonatal death in Gad1 -/mice is caused by respiratory failure rather than 331 impairment of suckling [15,35]. Therefore, neonatal death in Gad1 STOP-tetO/STOP-tetO mice may also be caused by 332 respiratory failure. 333 We next developed Gad1 tTA/STOP-tetO biallelic knock-in mice by crossing Gad1 tTA/+ and Gad1 STOP-tetO/+ parents. 334 Approximately 40% of Gad1 tTA/STOP-tetO mice died on the day of birth, and the number of Gad1 tTA/STOP-tetO mice at 335 P21-P28 in our breeding colony was smaller than the numbers of mice with the other genotypes. None of the 336 Gad1 tTA/STOP-tetO mice demonstrated a cleft palate. Unexpectedly, some adult Gad1 tTA/STOP-tetO mice died by 337 treatment with Dox. Therefore, GAD67 is important for survival not only in the neonatal period but also in 338 adulthood. However, the cause of death in Dox-treated Gad1 tTA/STOP-tetO mice is currently unknown. To resolve this 339 21 question, pathological examination is required in a future study. 340 Adult mice with Gad1 haplodeficiency demonstrated an approximately 40% reduction in GAD67 protein 341 levels in the whole brain compared with Gad1 +/+ mice [27]. Consistently, we observed that heterozygous Gad1 tTA/+ 342 and heterozygous Gad1 STOP-tetO/+ knock-in mice exhibited a 30 -50% reduction in GAD67 protein levels in the 343 FCX, HIP and CER compared with Gad1 +/+ mice. In the absence of Dox treatment, the expression level of 344 GAD67 protein in Gad1 tTA/STOP-tetO mice relative to Gad1 +/+ mice was dependent on the brain regions. In the 345 immunoblotting analysis, the expression of GAD67 protein in the CER was comparable between Gad1 tTA/STOP-tetO 346 mice and Gad1 +/+ mice. However, the expression of GAD67 protein in the FCX and HIP was significantly lower 347 in Gad1 tTA/STOP-tetO mice than Gad1 +/+ mice. Importantly, in the presence of Dox treatment, GAD67 expression was 348 reduced by approximately 90% in the brains of Gad1 tTA/STOP-tetO mice, compared with Dox-treated Gad1 +/+ mice. 349 The brain-wide reduction of GAD67 expression in Dox-treated Gad1 tTA/STOP-tetO mice was also observed in the 350 immunofluorescence analysis. These findings suggest that GAD67 expression is suppressed by treatment with 351 Dox in the brains of Gad1 tTA/STOP-tetO mice. 352 In adult mice with Gad1 haplodeficiency, the GABA content in the brain was reduced by 7 -20% from those 353 in wild-type control mice [15,32]. In this study, we found that the GABA contents in the FCX, HIP and CER of 354 Dox-treated Gad1 tTA/STOP-tetO mice were reduced by 27.0% -56.8% from those in Dox-treated Gad1 +/+ mice. 355 Therefore, the GABA reduction in the brains of Dox-treated Gad1 tTA/STOP-tetO mice was larger than that in Gad1 356 22 haplodeficient mice. Because approximately half of the brain GABA is produced by GAD65 in adulthood [36], 357 the remaining GABA in the brain of Dox-treated Gad1 tTA/STOP-tetO mice is mainly synthesized by GAD65. The 358 brain glutamate contents in Dox-treated Gad1 tTA/STOP-tetO mice were comparable to those in Dox-treated Gad1 +/+ 359 mice. Therefore, the glutamatergic system may be normal in Dox-treated Gad1 tTA/STOP-tetO mice. 360 GAD67 haplodeficient mice demonstrated several abnormal behaviors, such as hyper-locomotor activity, 361 reduced interactions with an unfamiliar mouse, and aggressive behavior. However, the emotional behaviors in 362