A tailored Cln3Q352X mouse model for testing therapeutic interventions in CLN3 Batten disease

CLN3 Batten disease (CLN3 disease) is a pediatric lysosomal storage disorder that presents with progressive blindness, motor and cognitive decline, seizures, and premature death. CLN3 disease results from mutations in CLN3 with the most prevalent mutation, a 966 bp deletion spanning exons 7–8, affecting ~ 75% of patients. Mouse models with complete Cln3 deletion or Cln3Δex7/8 mutation have been invaluable for learning about both the basic biology of CLN3 and the underlying pathological changes associated with CLN3 disease. These models, however, vary in their disease presentation and are limited in their utility for studying the role of nonsense mediated decay, and as a consequence, in testing nonsense suppression therapies and read-through compounds. In order to develop a model containing a disease-causing nonsense point mutation, here we describe a first-of-its-kind Cln3Q352X mouse model containing a c.1054C > T (p.Gln352Ter) point mutation. Similar to previously characterized Cln3 mutant mouse lines, this novel model shows pathological deficits throughout the CNS including accumulation of lysosomal storage material and glial activation, and has limited perturbation in behavioral measures. Thus, at the molecular and cellular level, this mouse line provides a valuable tool for testing nonsense suppression therapies or read through compounds in CLN3 disease in the future.


Results
Generation of Cln3 Q352X mice. Cln3 Q352X mice were generated by Applied StemCell Inc. using CRISPR technology to introduce a nonsense mutation in exon 16 (CAG > TAG), causing glutamine352 (Q) to be replaced with a premature stop codon (X). Two Cln3 guide RNAs were selected (Table 1), cloned into guide RNA/cas9 expression vectors, and transfected into mouse N2A cells for evaluation of non-homologous end joining efficiency (Fig. 1A,B). While mCLN3.g13 had greater efficiency (37% vs. 21%), mCLN3.g14 was chosen to generate donor DNA due to its optimal position to the point mutation.
A single stranded oligodeoxylnucleotide donor was made based on mCLN3.g14, with a silent mutation added to leucine351 (CTC > TTG) to prevent the repaired genome sequence from being re-targeted, and combined with mCLN3.g14 Cas-9 vectors and active guide RNA (Fig. 1C). The mCln3-Q352X CRISPR cocktail was injected into the cytoplasm of C57BL/6 embryos, and newborn mice screened for integration. Of the 100 embryos injected, 76 developed and were implanted into three surrogate mice, with 12 mice ultimately born. One cloned male was heterozygous for the Cln3 Q352X mutation and one male was homozygous, as confirmed by sequencing chromatograms (Fig. 1D, homozygous animal). These two animals were used as founders and mated to wildtype females to produce the F1 generation, which produced nine male and four female heterozygous Cln3 Q352X animals that were sequenced, shipped, and bred to generate homozygous Cln3 Q352X mice (Fig. 1E, heterozygous founder). Cln3 Q352X mice show reduced Cln3 transcript levels and variable TPP1 and PPT1 enzyme activities. mRNA was measured in various tissues from wildtype and Cln3 Q352X mice to determine Cln3 transcript levels. Significantly different Cln3 Q352X transcript levels ranged from 34 to 67% of wildtype and were seen in several tissues of both male and female Cln3 Q352X mice, including the cerebral cortex, thalamus, brainstem, lung, and kidney (Table 2). Cln3 Q352X males also had lower transcript levels in the spleen and liver, whereas Cln3 Q352X female mice displayed decreased transcript levels in the heart and eye.
Loss of CLN3 has been shown to affect lysosomal function [21][22][23] , and several recent articles have shown that various lysosomal enzymes, including PPT1 (protein product of CLN1) and TPP1 (protein product of CLN2), are altered in Cln3 Δ7/8 mice 24,25 . To examine whether this was the case in Cln3 Q352X mice, PPT1 and TPP1 enzyme activity assays were performed, showing decreased PPT1 activity in the liver of Cln3 Q352X mice and increased TPP1 activity in the thalamus and cerebellum (Table 2), consistent with previous reports 26,27 . Cln3 Q352X mice have typical pathological progression of Batten disease hallmarks. Wildtype and homozygous Cln3 Q352X mice were sacrificed at 6 and 21 months of age (early and late disease stage) to determine whether the mouse model displayed the typical pathological progression reported in CLN3 patients and other Cln3 models. In particular, astrocyte activation (GFAP + ), microgliosis (CD68 + ), and accumulation of www.nature.com/scientificreports/ lysosomal storage material (mitochondrial ATP synthase subunit C) were examined in the brain as hallmarks of CLN3 disease progression 28-32 . As regional pathogenesis starts in the somatosensory cortex and thalamus, and male and female patients have shown differences in disease progression, assays were analyzed across gender in the somatosensory barrel field (S1BF) cortex and the ventral posteromedial/ventral posterolateral (VPM/VPL) nuclei of the thalamus 4,33 . At 6 months, subunit C accumulation was already significantly present in both the S1BF and the VPM/ VPL of Cln3 Q352X mice while wildtype controls showed extremely minor amounts of positive immunoreactivity ( Fig. 2A). Astrocytosis was also prominent at 6 months with females showing increased values in the S1BF (Fig. 2B). Additionally, there were qualitative differences in astrocyte morphology between wildtype and Cln3 Q352X mice, with Cln3 Q352X astrocytes appearing slightly more hypertrophic and ramified. There was no significant difference, however, in microglial activation at this early 6 month timepoint (Fig. 2C). Pathology was widespread and significantly exemplified at 21 months, which is near end stage of a normal mouse lifespan. Subunit C accumulation was evident throughout the brain, increasing in both S1BF and VPM/VPL regions with no apparent sex differences (Fig. 3A). Astrocytosis followed a similar trend in both regions with obvious differences in morphology as Cln3 Q352X displayed extreme hypertrophy and ramification when compared to wildtype animals (Fig. 3B). Additionally, microglial activation was significantly elevated in Cln3 Q352X mice at 21 months in the VPM/VPL for both sexes while only males were significantly altered in the S1BF compared to their wild type counterparts (Fig. 3C).
Behavioral evaluation of Cln3 Q352X mice reveals sex-dependent motor coordination deficits. While other Cln3 mouse models are notorious for their limited and inconsistent behavior deficits, we  Table 1. Each oligo cassette contained 20 bp gRNA sequences with a guanosine at the 5′ end for optimal expression, and adherent ends for cloning at Bbs I sites. ( www.nature.com/scientificreports/ sought to thoroughly characterize any motor coordination deficits that may be present in Cln3 Q352X mice. When assessed on an accelerated rotarod test, Cln3 Q352X mice generally performed similarly to their wild type counterparts, and in some instances performed better compared to wild type mice (Fig. 4A). When assessed for motor coordination deficits using another assay, a modified vertical descent test, male Cln3 Q352X mice showed significant difficulties in climbing down and turning down the pole (Fig. 4B,C). These deficits were prominent from 3 months of age, with wild type males performing poorly at the test as they aged. Female Cln3 Q352X mice, however, showed inconsistent deficits in these assays. Lastly, in order to control for weight and anxiety as confounding factors in behavior results, mice were weighed at each time point and subjected to a light-dark test to assess their preference for dark areas. Cln3 Q352X male mice weighed less than their wildtype cohorts at 3, 6 and 18 months, while Cln3 Q352X female mice weighed less only at 6 months (Fig. 4D). Any weight disadvantages experienced by wildtype mice would be most prominent at these time points, which is not seen in the rotarod test, as there were no differences between wildtype and Cln3 Q352X mice at these time points. However, heavier wildtype mice may have been disadvantaged in the 18 month vertical descent test, as male wildtype mice took significantly longer to descend the pole when compared to the lighter Cln3 Q352X male mice. When assessed for anxiety and preference for dark areas, both groups spent the majority of their time in the dark side of the box with only one significant difference in female Cln3 Q352X mice at 1 month of age (Fig. 4E), indicating that anxiety likely did not contribute to the results of the other behavioral tests.

Discussion
Batten disease is a lysosomal storage disorder that, while individual subtypes are rare, represents the most common pediatric neurodegenerative disease worldwide. CLN3 Batten disease results from homozygous mutations in CLN3, which are present in ~ 80% of individuals inflicted with Batten disease. Although the function of the CLN3 protein is unknown, the pathogenic progression of the disease is well established in both human patients and animal models. Common symptoms of CLN3 Batten disease include progressive vision loss, neurocognitive decline, behavioral changes, and motor deficit with symptom onset beginning around ~ 4-8 years and death occurring at ~ 20 years. In mouse models of CLN3 Batten disease, disease progression results in whole brain atrophy, including region-specific atrophy in the cortex, hippocampus, thalamus, and cerebellum. Additionally, mice present with pathological hallmarks including reactive gliosis, neuroinflammation, and accumulation of Table 2. Cln3 Q352X mice show lowered Cln3 transcript abundance and variable PPT1 and TPP1 activity in multiple tissues. Cln3 transcript abundance of Cln3 Q352X mice was significantly lower in all tissues at 1-month of age. Fluorogenic PPT1 and TPP1 enzyme activity assays performed on tissue from 1 month old Cln3 Q352X and WT mice present variable PPT1 levels in several tissues of Cln3 Q352X mice, with a significant decrease in the liver. An increase in TPP1 enzyme activity was significantly observed in the thalamus and cerebellum in Cln3 Q352X mice. Unpaired t test, mean ± SEM, N = 2-4/sex/genotype. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. www.nature.com/scientificreports/ Figure 2. Cln3 Q352X mice exhibit common pathological hallmarks at 6-months of age. Immunohistochemistry on 6 month tissue showed significant increases in subunit C immunoreactivity (A) and GFAP immunoreactivity (B) in multiple brain regions for both male and female Cln3 Q352X mice. Microglial activation, as reflected by CD68 immunoreactivity (C), was decreased in the somatosensory cortex (S1BF) and thalamus (VPM/VPL) of male Cln3 Q352X mice as well as the S1BF of female Cln3 Q352X mice. A slight increase in microglial activation was observed in the VPM/VPL of female Cln3 Q352X mice. Two-way ANOVA, Fisher's LSD post-hoc. Mean ± SEM, ****p < 0.0001. All images are representative of male mice. Scale bar = 100 µm. www.nature.com/scientificreports/ www.nature.com/scientificreports/ ASM. However, while Cln3 mouse models recapitulate some motor and pathological signs of the disease, the visual and cognitive aspects of the disease in human patients do not develop. While the genetic knockout mouse models have been beneficial for studying the selective pathogenesis of Batten disease, recently, tailored mouse models have been generated to develop therapies to individual mutations. Generation of novel mouse models from patient mutations of Batten disease present with many of the same neuronal and visual phenotypes seen in patients 29,31,34-59 . Regional-specific vulnerability of cortical, hippocampal, thalamic, and cerebellar regions of the CNS have been identified in many of the murine models of Batten disease 29,31,37,38,44,60,61 . These models also help in longitudinal identification of pathological hallmarks and detection of glial activation before neuronal changes occur.
Cln3 Q352X displayed the typical pathological progression of disease hallmarks as other Cln3 mouse models. ASM was evident in the somatosensory cortex and the thalamus as early as 6 months ( Fig. 2A) and became widespread and extreme at 21 months of age (Fig. 3A). Similarly, reactive astrocytosis was prevalent at 6 months (Fig. 2B) which significantly worsened at 21 months with clear morphological differences between wild type and Cln3 Q352X mice (Fig. 3B). While microglial activation (CD68) at 6 months was not significant in either of the brain regions analyzed, it was significantly increased at 21 months in Cln3 Q352X mice, and it Cln3 Q352X mice appeared to have hypertrophic microglia at this stage in disease (Fig. 3C).
We have previously demonstrated that the neurological deficits in the Cln3 -/and Cln3 Δex7/8 mouse models of juvenile CLN3 disease strongly depend on the genetic background and gender 34 , and female CLN3 patients have reported more severe symptoms 33 . Therefore, the age-dependent disparate behavioral test results of male and female Cln3 Q352X mice are not surprising, though notably female Cln3 Q352X mice lacked a consistent, more severe phenotype. In the vertical pole test, Cln3 Q352X mice showed age dependent motor deficits compared to wild type mice. In the accelerating rotarod test, however, Cln3 Q352X mice performed better than wild type mice at certain ages. These surprising results are not unique to Cln3 Q352X mice, as we have previously shown that while Cln3 Δex7/8 female mice displayed motor deficits in the vertical pole test, they performed markedly better than their wild type and Cln3 −/− counterparts in the rotarod test 34 .
Although, decreased Cln3 Q352X transcript levels were seen in several organs and brain regions of both male and female Cln3 Q352X mice (Table 2), no statistically significant reduction of Cln3 Q352X mRNA was detected in the cerebellum, muscle, heart and eye of males, and the cerebellum, muscle, spleen and liver of females. It has been shown that inter-tissue and sex variations exist in the efficiency of nonsense-mediated mRNA decay in mice, and therefore, it is possible that these variations account for the differences seen in our model 62,63 . It has previously been shown that the activity of the soluble lysosomal enzyme, TPP1 (product of the CLN2 gene) is markedly increased in the brain of patients with juvenile CLN3 disease and also in the brain of Cln3 -/mice [24][25][26] . In Cln3 Q352X mice, however, TPP1 activity, was only significantly elevated in the thalamus and cerebellum, perhaps suggesting that the Cln3 Q352X mutation does not affect lysosomal function as severely as the most common disease-causing deletion (Cln3 Δex7/8 ) or the complete lack of Cln3 (Cln3 -/-). Importantly, the clinical presentation of the Cln3 Q352X patient mutation has not been reported, and it is therefore unclear whether the point mutation brings about less severe cellular manifestations and, as a consequence, disease course.
Taken together, we show pathological and behavioral characterization of a novel point mutant Cln3 Q352X mouse model, which shows progression of pathological symptoms similar to the common Cln3 Δex7/8 mouse model. Beneficially, Cln3 Q352X model enables the screening of read-through and nonsense suppression therapies in the future, that if tested in the common Cln3 Δex7/8 model would be confounded by the presence of novel amino acids. Therefore, this novel model shows its utility as a pathological model of CLN3 Batten disease, and its potential as a tool for screening tailored genetic therapies in the future.

Methods
Ethics statement. All animal research performed for this manuscript followed National Institute of Health (NIH) and Sanford Research Institutional Animal Care and Use Committee (IACUC) guidelines. The animal protocol for this study was reviewed and approved by the Sanford Research IACUC.
Mouse generation. Cln3 Q352X mice were generated by Applied StemCell Inc. (CA). Using CRISPR technology, a nonsense mutation was introduced in exon 16 (CAG > TAG) causing glutamine352 (Q) to be replaced with a premature stop codon. Addition of a silent mutation (CTC > TTG) directly before the nonsense mutation was necessary to generate the mouse line and codes for the same amino acid (leucine). Specific guide RNAs, oligo donors, and Cas-9 vectors are described in Fig. 1 and Table 1. gRNA non-homologous end joining efficiency was determined by Applied StemCell Inc. using a SURVEYOR assay on PCR products (Transgenomic Inc., #706020), and cloned and F1 generations were confirmed by sequencing chromatograms. formed three different tests to determine the impact of the Cln3 Q352X mutation on parameters of performance such as motor coordination, spatial awareness, and anxiety-like behavior. Mice were divided into two behavioral cohorts to minimize motor learning and repeated measures testing. Behavioral tests were performed on the same groups of male and female Cln3 Q352X mice and wild type mice in Cohort 1: 1, 3, and 6 months of age (n = 12 /sex/genotype) and Cohort 2: 9, 12, and 18 months of age (n = 11 of each sex/genotype). Before behavioral tests were performed, mice were allowed a 20 min acclimation period to their testing environment.

Rotarod.
Motor coordination was assessed using an accelerating rotarod test as previously described 64 . In brief, a Rota Rod Rotamex-5 (Columbus Instruments, OH) was utilized. Parameters for the rotarod test were as follows: start speed: 0 rotations per minute (rpm); end speed: 48 rpm, acceleration: 0.2 rpm per second, and total allotted test time: 240 s. Mice underwent three consecutive training runs on the rotarod before a 1.5 h rest period. After the rest period, mice were tested on the rotarod for three test trials comprising three consecutive testing runs with 15 min of rest between each trial. Latency to fall was analyzed as an average of the nine test runs for each mouse.
Modified vertical pole test. In order to test motor coordination and spatial awareness, a modified vertical pole test was performed as previously described 64 . Briefly, the vertical pole utilized a threaded metal pole attached to a padded base and consisted of two tests: (1) climb down: mice were placed at the top of the and timed until they climbed down to the base of the; and (2) turn downward: mice were placed at the top of the pole facing up and timed until they turn downward. For the climb down test, mice were given 60 s to climb down the pole, performing the test five consecutive times. During the turn downward test, mice were again given 60 s to turn around and face downward, performing the test four consecutive times. Mice that fall off the pole in either test were given a score of 60 s. The total sum of time in the trials for each test and for each mouse was calculated for analysis.
Light/dark box. Mice were placed individually into a box with a light side and a dark side (Stoelting Co., IL) and allowed to roam freely for 10 min. Mouse movement was recorded and tracked using ANYmaze equipment and software (Stoelting Co., IL). Software data was used to determine the amount of time spent in each area of the box. The time spent in the dark box is proportional to anxiety, as anxious mice spend little time in the light box.
Tissue collection and preparation. For enzyme analysis and transcript abundance, Cln3 Q352X mice and wild type mice (n = 4/sex/genotype) were euthanized using a carbon dioxide chamber at 1 month of age. Following euthanasia, mice were flushed with 10 ml of sterile phosphate buffered saline solution by cardiac perfusion. Collected tissues were flash frozen on dry ice, and included the cerebral cortex, cerebellum, striatum/thalamus, liver, and kidney. All flash frozen tissues were stored at − 80 °C until analysis. Tissues for enzyme analysis were homogenized, processed, and extracted using a Maxwell 16 LEV simplyRNA tissue kit (Promega), while tissues for enzyme analysis were homogenized in protein lysis buffer. Homogenization of tissues was performed in a Bertin Technologies Precellys 24 homogenizer using 2-ml tubes containing zirconium beads. For histological analysis, Cln3 Q352X mice and wild type mice (n = 6/sex/genotype) were euthanized at 6 and 21 months of age. Following euthanasia, animals were flushed with 10 ml of sterile phosphate buffered saline solution followed by a 10-ml perfusion of 4% paraformaldehyde in PBS both via cardiac perfusion. Extracted brains were placed into scintillation vials containing 4% paraformaldehyde in PBS. Brains were stored at 4 °C with PBS containing 0.02% sodium azide replacing the 4% paraformaldehyde solution after 24 h fixation. Brains were set in 5% low melting agarose and sliced coronally into 50 μm sections using a Leica VT1000S vibratome in a 1:6 serial collection.
Protein sample preparation. Protein sample preparation was performed as previously described 27 . Briefly, tissue were cut and placed into 2-ml tubes containing zirconium beads and 500 μl of protein lysis buffer and homogenized using a Bertin Technologies Precellys 24 homogenizer. Homogenized samples were centrifuged at 12,000g for 10 min with supernatant sample transfered to 1.5-ml microcentrifuge tubes and storage at − 80 °C. Total protein concentrations were determined using the Pierce 660 nm Protein Assay (Thermo Fisher Scientific). to fall at all time points on the rotarod, with significance at 9 and 12 months of age. Female Cln3 Q352X mice also displayed an increased latency to fall at all time points except 12 months, with significance at 3 and 6 months of age. (B) On the climb down test, an increased latency was observed for male Cln3 Q352X mice at 3, 9, and 12 months of age and a decreased latency at 18 months of age compared to male WT mice . Female Cln3 Q352X mice showed an increased latency at 9 months of age and a decreased latency at 18 months compared to female WT mice. (C) Time to turn downward test showed an increased latency in male Cln3 Q352X mice at 6 and 12 months of age compared to male WT mice. Female Cln3 Q352X mice only showed an increased latency at 6 months of age compared to female WT mice. (D) Significant differences in weight were observed at multiple time points. (E) WT and Cln3 Q352X mice spend the majority of their time in the dark side of the light/dark box with only one significant difference in female mice at 1 month of age. Two-way ANOVA, Fisher's LSD post-hoc. Mean ± SEM, *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. www.nature.com/scientificreports/ Batten-related lysosomal enzyme activity assays. Evaluation of TPP1 (tripeptidyl peptidase 1) and PPT1 (palmitoyl protein transferase 1) enzyme activities were completed using samples from Cln3 Q352X and wild type mice as previously described 14,27  RNA sample preparation and reverse transcription for cDNA synthesis. RNA and cDNA sample preparation were performed as previously described 27 . RNA sample concentrations were calculated and cDNA synthesis was executed on 1 μg of total RNA using GoScript Reverse Transcription System (Promega). RT-PCR Reactions were performed on an Applied Biosystems Veriti Thermal Cycler according to manufacturer's protocol. Samples were stored at − 20 °C.
Image acquisition and analysis. All microscope slides were scanned on a Leica DM6000B slide scanning microscope at 20X. Images were extracted in 2,400 × 2,400 pixel fields from appropriate regions and analyzed using an image threshold analysis using ImageJ (version 1.51). High-resolution images for figures were taken on a Nikon NiE microscope.
Statistical analysis. Statistical analyses were performed using GraphPad Prism (v6.04). Male and female Cln3 Q352X mice were compared to their same-sex cohort for all statistical analyses using two-way ANOVA with Fisher's LSD and outlier removal using the ROUT method, Q = 1%.