A huntingtin knock-in pig model recapitulates features of selective neurodegeneration in Huntington’s disease

SUMMARY

Huntington’s disease (HD) is characterized by the preferential loss of the medium spiny neurons in the striatum. Using CRISPR-Cas9 and somatic nuclear transfer technology, we established a knock-in (KI) pig model of HD that endogenously expresses full-length mutant huntingtin (HTT). By breeding this HD pig model, we have successfully obtained F1 and F2 generation KI pigs. Characterization of the founder and F1 KI pigs shows the consistent movement and behavioral abnormalities as well as early death that are germline transmittable. More importantly, HD KI pig brains display striking and selective degeneration of striatal medium spiny neurons. Thus, using a large animal model of HD, we demonstrate for the first time that overt and selective neurodegeneration seen in HD patients can be recapitulated by endogenously expressed mutant proteins in large mammals, which also underscores the importance of using large mammals to investigate the pathogenesis of neurodegenerative diseases and their therapeutics.

In Brief

A CRISPR-Cas9 knock-in pig model recapitulates the selective neurodegeneration observed in human Huntington’s disease patients

INTRODUCTION

A variety of neurodegenerative diseases, including Alzheimer’s (AD), Parkinson’s (PD), and Huntington’s (HD) as well as amyotrophic lateral sclerosis (ALS), share the common features of age-dependent neurological symptoms and selective neuronal degeneration (Wyss-Coray, 2016). Current research indicates all these diseases show the accumulation of misfolded proteins in the brain (Bates et al., 2015; Goedert et al., 2015; Balchin et al., 2016; Paulson et al., 2017). However, the cause behind the accumulation of misfolded proteins varies with the different diseases, and the mechanisms underlying particular neurodegeneration phenomenon remain underinvestigated. While genetic mutations contribute to a small fraction of AD, PD, and ALS; HD results from a monogenetic mutation, which is a CAG repeat expansion in the exon1 of the gene for huntingtin (HTT) (Bates et al., 2015). HTT is a multifaceted protein that is expressed ubiquitously and has numerous roles (Saudou and Humbert, 2016). The CAG repeat expansion (>36 CAGs) in the HTT gene leads to a polyglutamine (polyQ) expansion that causes HTT to misfold and aggregate in the brain. Similarly, the polyQ expansion also causes at least 8 other neurodegenerative diseases including various spinocerebellar ataxias as well as spinal-bulbar muscular atrophy (Orr and Zoghbi, 2007; Paulson et al., 2017).

The monogenic mutation feature makes HD an ideal model to investigate the pathogenesis of misfolded proteins common in neurodegenerative diseases. By expressing mutant HTT containing an expanded CAG repeat in different species, a variety of genetically modified animal models of HD have been established and characterized. Among these models, mouse models of HD have been widely used and provided valuable information regarding the pathogenesis and therapeutic development of HD. Although the HD mouse models show age-dependent accumulation of mutant huntingtin and its associated neurological symptoms, the HD knock-in mouse models, which express mutant HTT at the endogenous level, lack the overt and striking neurodegeneration, a typical pathological hallmark of HD (Levine et al., 2004; Crook and Housman, 2011). Similar to HD mouse models, other genetically modified mice that express different types of misfolded proteins including those for AD and PD also show the absence of overt and selective neurodegeneration (Ashe and Zahs, 2010; Dawson et al., 2010; Epis et al., 2010). There are considerable differences between rodents and primates. For example, the striatum, which is the most affected region in HD, consists of the caudate nucleus and putamen in the primate brains. However, the caudate nucleus and putamen are indistinguishable in rodents. The differences in neuropathology among the rodent and human brains indicate that species differences determine the nature of neuropathology and also highlight the demand for investigation of larger mammals that are closer to humans.

Pigs are genetically, anatomically, and physiologically closer to humans than small mammals. More importantly, the existing genetic manipulation tools enable the generation of a variety of pig models of human diseases (Prather et al., 2013; Holm et al., 2016). Somatic cell nuclear transfer (SCNT) in combination with CRISPR-Cas9 allows for genetic modifications of the endogenous pig genes (Zhou et al., 2015; Yang et al., 2016; Han et al., 2017). The SCNT leads to non-chimeric animals in the first generation that may recapitulate endogenous genetic mutation-associated phenotypes. In addition, the fast breeding period (5–6 months for sexual maturation) and large litter size (average 7–8 piglets) of pigs hold obvious advantages over non-human primates when considering the timeline to generate large animal models of human diseases.

Our previous studies showed that pigs overexpressing transgenic mutant HTT did not survive (Yang et al., 2010), which has prevented us from investigating neurodegeneration in adult animals. We used CRISPR-Cas9 to insert a large CAG repeat (150 CAGs) into the endogenous pig HTT gene in fibroblast cells and employed the SCNT to generate a HD knock-in (KI) pig model that expresses full-length mutant HTT at the endogenous level. This KI model allowed us to explore whether misfolded proteins at the endogenous level can cause neurodegeneration in large mammals. We found germline transmittable neurological phenotypes in different generations of this KI pig model. More importantly, we provide convincing evidence that mutant HTT causes striking and selective neurodegeneration that recapitulates the typical neurodegeneration feature in HD. Our findings support the idea that species differences are critical for the nature of neuropathology and strengthen the rationale for using large mammals to investigate the pathogenesis of neurodegenerative diseases and to identify their therapeutics.

RESULTS

Generation of HD knock-in pigs

CRISPR-Cas9 can break the double stranded DNAs to facilitate homologous recombination, a process that is required for genetic replacement of the targeted gene or genomic knock-in (Hsu et al., 2014; Sander and Joung, 2014). We designed two gRNAs to target the pig HTT intron after exon 1 to promote homologous recombination by replacing the pig HTT exon 1 with the human exon 1 containing a 150-CAG repeat. We transfected fetal pig fibroblast cells from a female Rongshui pig with the gRNAs and Cas9, as well as a donor vector that carries human HTT exon1 with 150-CAGs repeat flanked by two pig HTT DNA fragments (1 kb for each left and right arm) for homologous recombination (Figure 1A). We screened 2430 fetal pig fibroblast cells by PCR and identified 9 positive cell clones that contained the heterozygous expanded human HTT exon1 in the right locus of the pig HTT gene. We selected a cell clone for SCNT and obtained 2880 embryos, which were transferred into 16 surrogate pigs. Of these surrogate animals, 10 became pregnant, which yielded some miscarried fetuses and 7 naturally delivered piglets (Figure 1B). Genotyping by PCR analysis identified that 6 piglets carried the human HTT exon1 with expanded CAG repeats (Figure 1C). The female founder (F0) pigs (Rongshui) were used to mate with wild type male Bama miniature pigs (Liu et al., 2008) to generate F1 pigs (Rongshui/Bama), and the male F1 KI pigs were crossed with wild type female Bama pigs to yield F2 generation pigs. This breeding over the last two years gave rise to 15 F1 pigs and 10 F2 pigs, which are all positive for carrying the mutant HTT (Figure 1C–E). PCR and DNA sequencing (Figure S1) verified the human exon1 sequences and large CAG repeats in the targeted pig HTT allele in F1 and F2 KI pigs. Interestingly, genotyping revealed that the CAG repeats were unstable, ranging from 130 to 150 CAGs in the F0 founders, 113 to 206 CAGs in F1, and 118 to 230 CAGs in F2 generation (Figure 1F). The oldest F0 KI animal is 30 months of age, and F1 KI pigs are 12 months old, while F2 KI pigs are newborn piglets (Figure 1G, Figure S2). Whole genome sequencing revealed no off-targets in the F1 KI pig brain cortex (Figure S3). Thus, our characterization of the HD KI pigs was focused on F0 and F1 KI pigs to investigate their behavioral and pathological changes.

Figure 1. Generation of HD knock-in pigs.

(A) Schematic diagram of the strategy to generate HD KI pig via homologous recombination. Two gRNAs were used to target the pig HTT intron after exon 1 to promote DNA breaks and homologous recombination. The donor DNA consisting of human exon 1 HTT with 150 CAGs and homologous pig HTT sequences (left arm and right arm) was used to replace the endogenous pig exon 1 HTT in cultured pig fibroblast cells. Cells containing the knock-in (KI) allele were identified via PCR and selected for somatic nuclear transfer technology. (B) Reconstruct embryos (2880) were transferred to 16 pig surrogates, resulting in 62.5% pregnancy and 6 live birth HD KI pigs. (C–E) PCR analysis of targeted allele containing the expanded CAG repeats in the ear tissues of the HD KI founders (F0) (C), F1 (D), and F2 (E) generation pigs. Note that the sizes of PCR products vary because of different CAG repeat numbers. (F) The number of KI pigs in each generation and the range of the CAG repeat numbers in the HD KI pigs. (G) Representative photos of HD KI founders (F0–7, 14 months old), F1 (F1–5, 5 months old), and newborn F2 pigs (7 days old). Arrows indicate symptomatic KI pigs. Also see Figure S1, S3, S4, video 1, 2, 3 and 4.

Age-dependent and germline transmittable neurological symptoms

Given that the expanded CAG repeats are transmitted via germline cells, we wanted to know whether the HD KI pigs develop age-dependent neurological symptoms and whether these symptoms are transmittable via germline cells. First, we needed to verify the expression of full-length mutant HTT in the HD KI pigs. Thus, we isolated brain tissues from some F0 KI founders (F0–2, F0–3, and F0–6) at the age of 5 months and performed western blotting with 1C2 antibody that specifically reacts with the expanded polyQ repeat. Western blotting clearly showed the expression of full-length mutant HTT (arrow above 245 kD in Figure 2A) and multiple fragmented HTT products, which are absent in the brain tissues of wild type pigs. The expression profiling of the intact full-length mutant HTT and its degraded products is very similar to that in HTT KI mouse brains (Landles et al., 2010; Bhat et al., 2014; Wade et al., 2014). As with mutant HTT in rodent brains, full-length mutant HTT in the pig brains also undergoes a proteolytic process to generate multiple N-terminal HTT fragments that carry 1C2 labeled-polyQ repeats. Western blotting also suggests that mutant HTT is more abundant in the cortex than the striatum and cerebellum where the lowest level of HTT is seen. Such brain region-dependent expression levels were not found in the HTT KI mouse brains (Landles et al., 2010; Wade et al., 2014). Thus, different expression levels of mutant HTT in distinct brain regions are likely species-dependant.

Figure 2. Age-dependent phenotypes of HD KI pigs.

(A) 1C2 Western blots of the brain tissues from four HD KI pigs (F0–2, F0–3, and F0–6) aged at 5 months. (B) Body weight and survival curve of HD KI (F0) founder pigs (n=6 for each group). (C) Photo of a symptomatic HD KI pig (F0–5) and WT pig at the age of 5 months old. (D) The foot-printing assay revealing gait abnormalities in the KI pig (F0–5) compared with the WT control. (E) Quantification of stride lengths for front and rear footprints for WT and KI pigs (n= 3 per group). * P<0.05. (F) 1C2 Western blots of the brain tissues from HD F1 KI pigs (F1–14, F1–15) at 5 months of age. For western blots in (A) and (F), full-length mutant HTT is indicated by an arrow. Arrowheads indicate non-specific bands. Western blot analysis was repeated independently at least three times. (G) Body weight and survival curve of the HD F1 KI pigs (n = 15 for KI and n = 15 for WT). Data are analyzed by 2-way ANOVA with Bonferroni’s test and are presented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001.

HD KI pigs did not show obvious symptoms before the age of 4 months. By continuously monitoring the growth and body weights of HD KI founders, we saw that F0 KI pigs gained less body weight than the age- and sex-matched wild type pigs (Figure 2B). The old HD KI pigs often displayed wrinkled and sagging skin on their bodies (Figure 1G, Figure S2A, B), which is similar to transgenic TDP-43 pigs we generated previously (Wang et al., 2015). Moreover, some of F0 KI pigs died earlier between the ages of 5 to 10 months (Figure 2B). F0 KI pigs often showed walking abnormalities (Figure 2C, Supplemental video-1). Foot-print analysis demonstrated the abnormal walking pattern with asymmetric steps and reduced distance between front and hind paw of the HD KI pigs as compared with wild type pigs that showed normal alternate steps while walking (Figure 2D, E). Importantly, HD KI pigs displayed respiratory difficulty or irregular breathing patterns and abnormal movement (Supplemental video-1). The breath difficulty suggests that respiratory failure could contributes to the death of animals and is consistent with the observation that pulmonary dysfunction and aspiration pneumonia/suffocation are the major cause of the death in HD patients (Heemskerk and Roos, 2012; Reyes et al., 2014). We also examined F0 KI pigs for their motor function using treadmill, as we did previously with our transgenic ALS pigs (Yang et al., 2014). Video recording of a symptomatic HD KI pig at 5 months of age (F0–5) revealed that this KI pig was unable to run as compared with its age- and sex-matched control (Supplemental video-2). Other F0 KI pigs (F0–6 and F0–7) showed similar running difficulties (Figure S2A). In addition, this HD KI pig died two days after the treadmill test, suggesting that the HD KI pig was susceptible to exercise stress. This post-treadmill death did not allow us to use treadmill to continuously test more KI pigs for their limb movement impairment.

For F1 KI pigs (F1–14 and F1–15), western blotting also verified the expression of full-length mutant HTT and its fragments in the cortex, striatum, and cerebellum (Figure 2F), with a similar pattern to that in the F0 KI brains. Further, we also observed less body weight gain and the early death of F1 KI pigs (Figure 2G) and similar difficulties in body movement and breathing (Supplemental video-3, video-4), indicating that F1 KI pigs shared the similar phenotypes as F0 KI pigs (Supplemental table-1). For the dead KI founders (F0–3), gross examination revealed lung edema with hemorrhaging and extensively dilated aleveolar lumen (Figure S4A, B), further supporting the idea that respiratory failure was likely the cause of animal death.

Analyzing CAG repeats in different pig generations showed that the CAG repeat is unstable in KI pigs (Figure S4C). We further analyzed the CAG repeat numbers in different tissues in F0 (F0–5, F0–6) and F1 (F1–14, F1–15) pigs. The results showed that there are different CAG repeats in different tissues (Figure S4D, E). We also compared the repeat numbers in the ear tissues of KI pigs of different generations and found that the CAG repeat is unstable during both male and female germline transmission (Supplemental table-2). However, it remains to be determined how the repeat numbers change during germline transmission when more KI pigs are available. Although the instability of the CAG repeat has been reported previously in the rodents (Lloret et al., 2006), expansion of the CAG repeats appears to be more frequent in the pig genome than the mouse genome. Interestingly, a pig carrying 206 CAGs (F1–7) lived longer than other pigs carrying 113–138 CAGs (Figure S4C), suggesting that the CAG repeat length is not the sole determinant of the lifespan in these HD KI pigs and that environmental stress and other factors can also influence their lifespan. Indeed, we found that individually housing F1 KI pigs could reduce environmental stress and made them to live longer than F0 KI pigs that were kept in a large group. Also, the F1 KI pigs were generated by mating female Rongshui F0 pig with Bama male pigs so that F1 KI pigs carry mixed genetic backgrounds (Rongshui/Bama), which could also influence the age of onset. It also seems that the breathing difficulty is unique to the pig KI model since it is not found in mice and other animal models of HD.

Brain region-selective neuropathology in HD KI pigs

For brain pathology analysis, we examined two F0 and three F1 pigs (Supplemental table-3). We isolated the brains of the F1 KI pigs (F1–14 and F1–15) at 5 months of age and found that the brain size of these pigs was reduced when compared with the age-matched WT pigs. The thickness of the cortex and size of the striatum were smaller than those in the WT pig brain (Figure 3A, B). MRI analysis of some symptomatic HD KI pigs at the age of 5 months also showed an enlarged lateral ventricle and the reduced size of the striatum when compared with the age-matched WT control (Figure 3C). Quantification of the volumes of the striatum and lateral ventricle in MRI confirmed the reduced volumes in HD KI pig brains as compared with wild type pig brains (Figure 3D).

Figure 3. Gross morphology of HD KI pigs and their brain images.

(A) Gross morphology of the HD F1 KI (F1–14 and F1–15) pig brains at the age of 5 months showing the reduced thickness of their cortex (arrowheads) and decreased size of the caudate (c) and putamen (p) as compared with the age-matched WT control. (B) Brain weight and ratios of the caudate and putamen to the total brain size in the WT (n = 4) and HD-KI (n = 4) pig brain at the age of 5 months. HD-KI pigs showed the reduced brain weight and decreased volume of the caudate and putamen. * P<0.05. (C) Magnetic resonance imaging (MRI) analysis reveals T2-weighted coronal images of live WT (n=3) and HD KI (n=4) pigs at the age of approximate 5 months. The arrowheads indicate the enlargement of the frontal horns of the lateral ventricles in the HD KI pig brain compared with WT pig brain. (D) The volumes of the striatum and lateral ventricle in MRI were decreased in HD KI pigs. Data are analyzed by student’s T-test and presented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001.

To provide substantial evidence for the neurodegeneration in HD KI pig brains, we performed immunohistochemical experiments. Immunostaining of the HD KI pig brains with mouse anti-HTT (mEM48) revealed the distribution of mutant HTT and small aggregates in the neurites and bodies of neuronal cells, and 1C2 immunostaining, which selectively labels expanded CAG repeats, also revealed mutant HTT aggregates in the nuclei (Figure 4A). It seems that mutant HTT aggregates in the pig brain have different conformations that are recognizable by different antibodies. We further performed immunostaining of the KI pig brains (F0–5 and F0–6) with antibody to NeuN, a neuronal marker protein, and found a drastic reduction in the number of NeuN-positive cells in the striatum (Figure 4B). Immunofluorescent labeling with nuclear staining allowed us to estimate the relative numbers of NeuN-positive cells over the total number of cells. This assay also showed the greater decrease of NeuN-positive cells in the striatum than the cortex. However, NeuN-positive cells in the KI cerebellum remained unaltered as compared to the WT control (Figure S5).

Figure 4. The brain-regional dependent neuropathology of HD KI pigs.

(A) EM48 and 1C2 immunocytochemistry of the brain striatum of wild-type (WT) and HD F1 KI pig (F1–14) at the age of 5 months. Neuropil aggregates (middle) and nuclear aggregates (right) were evident by immunostaining with mouse antibodies EM48 and 1C2, respectively. Scale bars: 20 μm. (B) Anti-NeuN immunostaining of the cortex, striatum, and cerebellum of WT and HD KI pigs (F0–5 and F0–6) at the age of 5 months. NeuN immunostaining reveals the more extensive neuronal loss in the striatum than in the cortex of the HD KI pig as compared to the WT control at the age of 5 months. Scale bars: 50 μm. (C) Anti-GFAP immunostaining of the cortex, striatum, and cerebellum of WT and HD KI (F1–14) pigs at the age of 5 months. Scale bars: 50 μm. (D) Quantification of the numbers of NeuN-positive or GFAP-positive cells in WT and KI pigs (n = 6 fields/each brain section from 6 brain sections from 3 animals/per group). Data are analysised by student’s T-test and presented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001. See also Figure S5.

It is known that reactive gliosis with increased GFAP, an astrocytic protein, is the early pathological event in HD KI mouse brains (Lin et al., 2001; Yu et al., 2003; Palfi et al., 2007). Immunohistochemical staining showed the increased staining in GFAP in the F0 KI pig brain (Figure 4C). Also, immunostaining with the antibody to IBA1, a microglial cell marker, showed a marked increase in its labeling in the KI pig brain, which is also more abundant in the striatum than the cortex (Figure S6A). Quantification of the number of different types of cells revealed that the KI striatum had the most severe loss of NeuN-positive cells and the highest increase in glial cell numbers (Figure 4D; Figure S6B).

In HD F1 KI (FI–14) pig brains, we also saw the most severe loss of NeuN-positive cells in the striatum (Figure 5A). Because the neuronal loss in F1 pigs represents an important pathology that is germline transmissible, we performed unbiased stereology on 3 wild type and 3 symptomatic F1 KI pigs at the age of 4–5 months to quantify the neuropathology in F1 pig brains. Compared with WT controls, the density of NeuN positive cells is decreased to the greater extent in the caudate (WT 108858 ± 7449; KI 53569 ± 4908; ** P=0.0034) and putamen in KI pigs (WT 111398 ± 1565, KI 76936 ± 6404, ** P=0.0064) than the cortex (WT 143727 ± 4070, KI 113704 ± 9969, * P=0.0494). No significant difference is seen in the cerebellum in KI and WT pigs (WT 666820 ± 8989, KI 683743 ± 3633, P=0.8699). We also quantified GFAP-positive cells and found there are a greater number of GFAP-positive cells in the caudate (WT 36153 ± 2121, KI 71428± 3735, ** P=0.0012) and putamen (WT 34235 ± 4683, KI 72045 ± 6964, * P=0.0108) than the cortex (WT 31284 ± 2376, KI 52183 ± 3882, ** P=0.0101), thought the number of GFAP cells was similar in the WT (62506 ± 3219) and KI (62862 ± 1853) cerebellum. For reactive microglial cells, we quantified IBA1-positive cells. F1 KI pig brains also found increased IBA1 cells in the cortex (WT 37969 ± 6252 KI, 63715 ± 6091 * P=0.0136), putamen (WT 44311 ± 2362, KI 74596 ± 4364, ** P=0.0036), caudate (WT 44504 ± 3809, KI 86246 ± 3690, ** P=0.0014), but not in the cerebellum (WT 48997 ± 2299, KI 55061 ± 3808, P=0.2445) (Figure 5B).

Figure 5. The selective neuropathology of HD F1 KI pigs.

(A) NeuN immunostaining reveals the more extensive neuronal loss in the striatum than in the cortex of F1 KI (F1–14) pig as compared to the WT control at the age of 5 months. Scale bars: 20 μm. (B) Stereology analysis of NeuN-, GFAP-, and IBA1-positive cells in the dorsal caudate nucleus and putamen, prefrontal cortex, and the anterior lobe of the cerebellum in wild type (WT) and F1 KI pigs. Three F1 KI pigs and WT pigs at the age of 4–5 months were examined. * P<0.05; ** P<0.01. (C) Western blotting of the brain tissues of WT and F1 KI founders with 1C2 for mutant HTT and antibodies against NeuN or GFAP. Actin served as a loading control. (D) Quantitation of the ratios of GFAP or NeuN to β-actin on the western blots (n = 6 fields/each brain section from 6 brain sections from 2 animals/per group). Western blot analysis was repeated independently at least three times. Data are analyzed by student’s T-test and presented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001. See also Figure S5.

The selective loss of NeuN- positive cells was also confirmed by western blotting results, which showed a marked decrease in the NeuN level in the striatum as compared with the cortex and cerebellum (Figure 5C), and quantitative analysis of the relative level of NeuN (ratio to actin in Figure 5D). Also, increases in GFAP level, which reflect an early neurodegenerative event, occurred in the striatum and cortex of the HD F1 KI pig brain (Figure 5D).

Selective neurodegeneration in the striatum in a symptomatic HD KI pig

Although the striatum is the most affected brain region in HD, the neuronal loss in the striatum has a temporospatial distribution and occurs only with the medium spiny neurons while interneurons are primarily spared (Vonsattel et al., 1985). Examining 163 clinically diagnosed cases of HD revealed that the caudate nucleus was preferentially degenerated versus the putamen in early HD stages (Vonsattel et al., 1985). To investigate whether HD KI pig can recapitulate this pathological feature, we freshly isolated the brain of symptomatic HD F1 KI pigs (F1–14, F1–15) at the age of 5 months and immediately fixed the brain with paraformaldehyde. Immunostaining of the striatum with an antibody to DAPRR-32, which is selectively expressed in the medium spiny neurons, clearly showed the greater reduction of DAPRR-32-positive cells in the caudate nucleus than the putamen (Figure 6A). We also used anti-calbindin D28k that specifically labels the medium spiny neurons as well and confirmed the reduced number of medium spiny neurons in the striatum (Figure S6C, D). There are different types of interneurons that express pavalbumin, NPY, or choline acetyltransferase (ChAT) (Cicchetti et al., 2000). Despite the scarcity of these interneurons in the striatum, their density in the striatum is not different between WT and KI pig brains (Figure 6B). Quantitation of the relative numbers of DARRP-32 neurons and interneurons confirmed that only medium spiny neurons degenerated in the striatum as compared with different types of interneurons (Figure 6C).

Figure 6. Selective neurodegeneration in the caudate of the HD KI (F1–15) pig.

(A) DARRP-32 staining shows a fewer medium spiny neuron labeling in the caudate than the putamen in the HD KI (F1–15) pig at 5 months of age. (B) Labeling of interneurons by antibodies to parvalbumin, NPY, and ChAT in the striatum of wild-type and HD KI pig striatum. Scale bar in (A) and (B): 50 μm. (C) Quantitative analysis of the relative number of DARRP-32 positive medium spiny neurons and different types of interneurons in the striatum of the F1 KI and wild-type pigs. The data are presented as mean ± SEM (6 tissue sections from 3 pigs per group). * P<0.05; *** P<0.001 by Student’s t-test.

Ultrastructural alterations in the brains of the HD KI pig brain

We then performed electron microscopic examination (EM) to explore the nature of neuronal loss in the striatum of the F1 KI pig (F1–15). EM revealed dark neurons in the cortex and striatum in the KI pig brain, which are characterized by electron-dense cytoplasm and the absence of organelles and a nuclear membrane (Figure 7A). These dark neurons were frequently seen in the cortex than in the striatum, perhaps because more neurons in the striatum have been lost (Figure S7A). In the striatum, degenerated neurons displayed the electron-lucent cytoplasm showing degenerated organelles or swollen mitochondria (Figure 7A). Different types of neurodegeneration in the HD KI pig cortex and striatum suggest that nature of neuronal degeneration depends on cell-type. We also frequently observed reactive astrocytes and microglial cells near degenerated neurons. These reactive glial cells display increased density in chromatin on the nuclear membrane, highly clumped heterochromatin. Also, increased cytoplasmic region with variable sized cytoplasmic vacuoles as well as gaps in the nuclear membranes were evident (Figure 7B, Figure S7B).

Figure 7. Ultrastructural alterations in the HD KI pig brain.

(A) Electron microscopy showed dark neurons in the cortex and striatum of the F1–15 KI pig brain. In the cortex, most degenerated neurons show electron-dense cytoplasm with no clear organelles and no identifiable nuclear membrane. In the striatum, degenerated neurons show an electron-lucent cytoplasmic region containing degenerated or swollen mitochondria and organelles as well as irregular and disintegrated nuclear membranes. Astrocytes (AS) and microglial (M) cells are in the vicinity of neurons. (B) Reactive glial cells in the HD KI pig brain display dense chromatin on the nuclear membrane, variable sized cytoplasmic vacuoles, and swollen mitochondria. Microglial cells are identified by having electron-dense and highly clumped heterochromatin, accumulated distinctly under the nuclear envelope. Astrocytes are characterized by their oval or round nuclei with an irregular ring beneath the nuclear envelope and evenly dispersed accumulations in the interior of the nucleus. (C, D) Demyelinated axons and degenerated axons containing dark or disintegrated organelles are evident in the striatum (C) and globus pallidus (D) in the HD KI pig brain. In the globus pallidus, a reactive microglial cell (arrow) is among the demyelinated axons. An age-matched wild-type (WT) control served as a control. Scale bars: (A) and (B), 2 μm; (C), 0.5 μm; (D), 1 μm. See also Figure S7.

Axonal degeneration is the early neuropathological event in old HD KI mice (Li et al., 2001; 2003; Marangoni et al., 2014). In the F1 KI (F1–15) pig brain, we found substantial axonal degeneration in the striatum and its projection area, globus pallidus. Compared to the wild-type (WT) control, axons in the HD KI pig brain showed reduced myelination and contained degenerated organelles (Figure 7C, D). Altogether, these findings offer convincing evidence that the HD KI pig brains display severe neurodegeneration with a similar pattern to that in HD patient brains (Vonsattel et al., 1985; Bate et al., 2015).

DISCUSSION

By establishing a HD knock-in pig model, our findings demonstrate for the first time that large mammals can recapitulate overt and selective neurodegeneration as well as severe symptoms caused by the mutant protein that is expressed at the endogenous level. These findings also raise some important issues regarding the differences in the pathology and phenotypes between small and large mammalian animal models of neurodegenerative diseases.

The CAG repeat length inversely correlates with the age onset of HD symptoms in humans. The HD KI pigs express a large CAG repeat (mostly 140–150Q) that is similar to the repeats (110Q–175Q) in HD KI mice but show more severe phenotypes and selective neurodegeneration than HD KI mice (Levine et al., 1999; Menalled et al., 2003; Loh et al., 2013). The majority of HD patients carry the CAG repeats ranging from 37 to 48, and juvenile patients often bear >55 CAGs in the HD gene (Rosenblatt et al., 2006; Douglas et al., 2010). Very large repeats (214 and 265 CAGs) were found in either symptomatic infants or aborted fetuses (Milunsky et al., 2003; Seneca et al., 2004), suggesting that the larger CAG repeats are more deleterious. Various repeat numbers that exist after pig germline transmission indicate the instability of CAG repeats, which was also found in mice and humans. The differences in the phenotypes of HD KI pigs and mice; however, suggest that species differences determine the tolerance of animals to the expanded CAGs, and the lower mammals are more tolerant than large animals to the CAG expansion.

In addition to the typical HD movement phenotypes, HD KI pigs had severe respiratory difficulty. In HD patients, death occurs 15–30 years after the onset of symptoms and is usually due to pneumonia or respiratory failure (Heemskerk and Roos, 2012). The dysregulation of the respiratory system results in irregular breathing patterns and decreased pulmonary function in HD patients (Leopold and Kagel, 1985; Reyes et al., 2014). The cause of pulmonary dysfunction remains unstudied, as reports of HD patients show a clear indication that respiratory failure in HD is not merely a consequence of a neurodegenerative condition, but is integral to the disease process (Jones et al., 2016). It could be possible that neurodegeneration in the HD pig particularly affects the central nervous system or circuitry that controls respiratory function. Alternatively, mutant HTT may have a unique peripheral effect on the lung function in pigs. Furthermore, the respiratory difficulty phenotype has not been reported in transgenic rodent models of HD, indicating that HD KI pigs can recapitulate more symptoms seen in HD patients.

The most important finding in our study is the presence of robust and selective neurodegeneration in the HD KI pig brains, which mimics the severe and preferential neurodegeneration of the medium spiny neurons in HD patients. Mutant HTT preferentially targets the striatum in humans and mice (Vonsattel et al., 1985; Bates et al., 2015; Langfelder et al., 2016). Although mutant HTT can affect multiple types of cells and impaired cell-cell interactions contribute to HD pathology (Garden and La Spada, 2012), the medium spiny neurons are more vulnerable in the patient brains at the early stages of HD (Vonsattel et al., 1985). The severe neurodegeneration of medium spiny neurons in HD KI pig brains is evident by the loss of DARPP32- or calbindin-D28k- positive neurons, which are the majority of neurons (>90%) in the striatum. In contrast, the numbers of interneurons, which specifically express parvalbumin, NPY, or ChAT and are spared in HD, are not altered in HD KI pig brains compared with the WT control. Thus, this striatal degeneration in the KI pig brain remarkably recapitulates the selective degeneration of medium spiny neurons in HD. Also, the increased reactive gliosis, including increased GFAP and IBA1 staining, is evident for the glial response to neuronal damage. The transgenic pig model, which expresses the first 548 aa of HTT with 124 glutamines under the control of human HTT promoter, did not show distinct movement phenotypes and neurodegeneration (Baxa et al. 2013). The differences between HD KI and transgenic pigs suggest that the context and expression level of HTT are essential for neurodegeneration. Because the preferential loss of striatal medium spiny neurons in the HD KI pigs mimics the critical pathological feature in HD patients, this finding also points out the possibility of pig models for replicating selective neurodegeneration in other neurodegenerative diseases.

Ultrastructural alterations also confirm the severe loss of striatal neurons in HD pig brains revealed by EM, which identified dark neurons that were previously found in transgenic HD mice expressing small N-terminal HTT fragments (Turmaine et al., 2000; Yu et al., 2003). Also, HD KI pig brains show degenerated axons and demyelination, as reported in HD KI mice (Li et al., 2001; Menalled et al., 2003). However, these pathological changes were often found in old (>21 months) HD KI mice and these changes in HD KI mice were milder than those in HD KI pigs (Bayram-Weston et al., 2012). Considering the lifespan differences between rodent (average 2 years) and pig (average 15 years), neurodegeneration caused by large polyQ repeats in the HD KI pigs apparently occurs much earlier and to a more severe extent. It should be pointed out that the expression of a large polyQ repeat (150Q) in the KI pig brains is more likely to elicit the neuropathology resembling juvenile HD. However, stereological analysis showed that loss of medium spiny neurons in the sacrificed F1 KI pigs is not as robust as that in postmortem brains of HD patients. It is possible that the extent of neurodegeneration would become more severe if these F1 KI pigs were kept to live longer. The robust neurodegeneration in the HD KI pigs is undoubtedly an advantage of using the KI pig model to explore the mechanism underlying the selective neurodegeneration in HD and to develop effective therapeutics.

Why can HD KI pigs more faithfully recapitulate the phenotypes and neurodegeneration seen in HD patients? While addressing this issue would require substantial experiments to test a variety of hypotheses, several possible explanations exist. First, the species-dependent differences in lifespan, genomics, anatomy, and physiology play essential roles in determining the severity of neurodegeneration in different species. Indeed, the lack of distinguishable caudate nucleus and putamen structures in the rodent striatum accounts for the inability to mimic the preferential caudate degeneration in HD. Second, the development of the central nervous system is remarkably different in various species. The rapid development and maturation of the rodent brain may render neuronal cells resistant to toxic proteins. On the other hand, the toxic effect of misfolded proteins during the lengthy early brain development in large mammals may be required for the more severe neuropathology in adult brains after the differentiation and maturation of neuronal cells. In support of this idea, some studies show that toxic effects of misfolded proteins in the embryonic or postnatal stage can affect the development of neuropathology in the adult brains (Molero et al., 2016). Also, HTT in the brains of small and large mammals may associate with different partners and function differentially. The pig HTT exhibited 96% peptide sequence homology to human huntingtin (Matsuyama et al., 2000). Since transgenic YAC and BAC mice expressing human HTT with a large repeat show more severe neurodegenerative phenotypes than HD knock-in mice (Slow et al., 2003; Gray et al., 2008), it is also possible that the large glutamine repeat in pig HTT can lead to the molecular changes that are more similar to those caused by mutant human HTT. Addressing the above possibilities would help one to understand the pathogenesis of neurodegenerative diseases.

Although our findings indicate that the pig model can more faithfully recapitulate neurodegeneration seen in HD patients, there are some limitations in using large animal models because of the high animal cost, expensive facility, and more stringent regulations. Rodent models have provided us with valuable tools to investigate the pathogenesis of neurodegenerative diseases. However, the pig model serves as an important tool to validate essential findings and therapeutic targets. Stem cell therapy to replace degenerated neurons in neurodegenerative diseases is an attractive approach but requires the use of an animal model that replicates the neuronal loss seen in the patient brains. Lowering HTT by delivering antisense oligonucleotides (ASO) into the right lateral ventricle has been successfully used to alleviate neurological symptoms in HD mouse models (Kordasiewicz et al., 2012). The HD KI pig model showing striking neurodegeneration would serve as an ideal model for these therapeutic tests. Finally, the evidence for the neurodegeneration in HD KI pigs also paves an avenue for generating animal models to mimic selective neurodegeneration in other critical neurodegenerative diseases such as AD and PD and to develop effective therapeutic strategies.

STAR*METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal Anti-Polyglutamine-Expansion Diseases Marker(1C2)	Millipore	Cat# MAB1574
Mouse monoclonal anti-γtubulin	Sigma	Cat# T6557
Mouse monoclonal anti-NeuN	Millipore	Cat# MAB377
Rabbit polyclonal anti-GFAP	Millipore	Cat# AB5804
Rabbit polyclonal anti-Iba1	WAKO	Cat# 019-19741
Rabbit polyclonal anti-DARPP32	Chemicon	Cat# AB1656
Monoclonal Anti-Calbindin-D-28K	Sigma	Cat# C9848
Rabbit polyclonal anti-Neuropeptide (NPY)	Abcam	Cat# ab123951
Rabbit polyclonal anti-Parvalbumin	Abcam	Cat# ab11427
Rabbit polyclonal anti-ChAT	Chemicon	Cat# AB5042
Mouse monoclonal mEM48	Millpore	Cat# MAB5374
Donkey Anti-Rabbit	Jackson Immunolabs	Cat# 715-035-152
Donkey Anti-Mouse	Jackson Immunolabs	Cat# 715-035-151
Donkey Anti-Mouse IgG H&L (Alexa Fluor® 594)	Abcam	Cat# Ab150108
Bacterial and Virus Strains
Xl1-blue	Stratagene	Cat#200249
Chemicals, Peptides, and Recombinant Proteins
DAPI	Sigma	Cat# D9542
Collagenase IV	Sigma	Cat# C5138
PBS	Hyclone	SH30256
Trypsin-EDTA solution	Sigma	Cat# T4049
DMEM	Hyclone	SH30243.01
FBS	Hyclone	Cat# SH30084.03
Penicillin	Sigma	Cat# P3032
Streptomycin sulfate	Sigma	Cat# S6501
Dimethyl sulfoxide	Sigma	Cat# D8779
TCM-199	GibcoBRL	Cat# 31100-035
Cysteine	Sigma	Cat# C8152
Luteinizing hormone	Sigma	Cat# L5269
Follicle stimulating hormone	Sigma	Cat# F2293
Epidermal growth factor	Sigma	Cat# S4127
G418	Sigma	Cat# A1720
Critical Commercial Assays
Gene Pulser/MicroPulser	Bio-rad	Cat# 1652088
Gene Pulser Xcell	Bio-rad	Cat# 1652660
VECTASTAIN Elite ABC Kits	Vector	Cat# PK-2200
ECL™ prime western blotting detection Kit	Fisher scientific	Cat# 45-002-401
Deposited Data
Human gene definitions for HTT	NCBI	https://www.ncbi.nlm.nih.gov/gene/3064
Pig gene definitions for HTT	NCBI	https://www.ncbi.nlm.nih.gov/gene/397014
CAG repeats in HD KI pigs	This paper	GenBank: XXXX
Experimental Models: Cell Lines
Rongshui fetal fibroblasts 13-07	Lai’s lab	N/A
Experimental Models: Organisms/Strains
HD-KI F0 pigs	This paper	N/A
HD-KI F1 pigs	This paper	N/A
HD-KI F2 pigs	This paper	N/A
Rongshui mini-pig	Lai’s lab	N/A
Bama mini-pig	Lai’s lab	N/A
Oligonucleotides
sRNA-1	IGE, Guangzhou, China	GCACCGACCGTGAGTGCGGG
sRNA-2	IGE, Guangzhou, China	GCGGTGACGTCATGCCTCGG
LF primer	IGE, Guangzhou, China	GGAGAGCTGGGAGAGAATGCCAGTGTGACAGT
LR primer	IGE, Guangzhou, China	GCGGCTGAGGCAGCAGCGGCTGTGCCTG
RF primer	IGE, Guangzhou, China	GGCCTTCGAGTCCCTCAAGTCCTTCCAG
RR primer	IGE, Guangzhou, China	GCTTCTTGAAAGCCGTCCTCATGAAATGCCTTCCGT
Recombinant DNA
Plasmid: phCas9	Mali et al., 2013	Addgene Plasmid#41815
Plasmid: sgRNA with U6 promoter	Chen et al., 2013	Addgene Plasmid#48962
Plasmid: pU6-sgRNA1	sgRNA with U6 promoter	N/A
Plasmid: pU6-sgRNA2	sgRNA with U6 promoter	N/A
Plasmid: pBS-hHD	This paper	N/A
Software and Algorithms
GraphPad Prism 6	Graphpad Software	www.graphpad.com
SnapGene 3.0	GSL Biotech LLC	www.snapgene.com
Stereo Investigator 5.4.3	Micro Bright Field Bioscience	www.mbfbioscience.com
Image J	National Institutes of Health	https://imagej.nih.gov/ij/docs
Matlab R2015b	Math Works	https://cn.mathworks.com
Other
Axio Imager A2	Zeiss	Carl Zeiss, Germany
Axio Imager 2	Zeiss	Carl Zeiss, Germany
Zeiss LSM 800 Confocal Laser Scanning Microscope	Zeiss	Carl Zeiss, Germany
MRI scanner	Siemens	Erlangen, Germany
BTX Electro-Cell Manipulator 200	BT 600	Genetronics, USA

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Xiao-Jiang Li (xli2@emory.edu).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animals

Rongshui miniature pigs and Bama miniature pigs were breed at the animal facility of Guangzhou Institute of Biomedicine and Health (GIBH), Chinese Academy of Sciences. Both Rongshui and Bama pigs are local strains from Southern China. Animal use followed the NIH Guide for the Care and Use of Laboratory Animals. The Institutional Animal Care and Use Committees (IACUC) at Guangzhou Institute of Biomedicine and Health (GIBH), Chinese Academy of Sciences approved the animal use protocol. This study occurred in strict compliance with the “Guide for the Care and Use of Laboratory Animals (2011)” to ensure the safety of personnel and animal welfare. The founder (F0) KI pigs carry Rongshui genetic background, and F1 and F2 pigs were generated by crossing F0 pigs with Bama miniature pigs. F0 KI pigs of the Rongshui genetic background and F1 KI pigs of mixed Rongshui and Bama genetic backgrounds were used for pathological and phenotype examination. For pathological examination, male or female HD KI pigs at 113–331 days after birth were examined (also see Table S3). All F0 KI pigs (n=6) were female because they were generated from female donor fibroblast cells. F0 KI (6 females) and F1 KI pigs (9 males and 6 females) and gender- and sex- matched wild type pigs were examined for their body weight and survive from 2 to 30 months after birth for F0 KI pigs and from 2 to 12 months for F1 KI pigs. All HD KI pigs did not receive any drug treatment or other procedures to alleviate their symptoms. The pigs were maintained under in-door housing conditions at room temperature in the Animal Center of Guangzhou Institutes of Biomedicine and Health. Regular food and water were provided ad libitum.

Cell lines and tissue culture

Pig fetus fibroblast cells were grown in culture medium (Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, MEM non-essential amino acids (Life Technologies), 1 mM sodium pyruvate, at 37 °C in a humidified incubator containing 5% CO2. This cell is derived from a female 35-days Rongshui mini-pig fetus.

METHOD DETAILS

sgRNAs and donor vector

The CMV promoter-driven Cas9 plasmid was purchased from Addgene (#41815). The U6-sgRNA cloning vector was constructed by introducing 2 BbsI restriction sites to the downstream region of the U6 promoter of plasmid gRNA (#48962, Addgene). Target sgRNAs on the sequences of the intron after pig HTT exon 1 were designed following the 17 nucleotide truncated sgRNA rule that reduces off targets (Fu et al., 2014; Tsai et al., 2015). Two complementary oligo DNAs of sgRNAs were synthesized and then annealed to double-strand DNA, ligated to the BbsI sites of U6-sgRNA cloning vector to form sgRNA-expressing plasmid. SgRNA sequences are 5′-gcaccgaccgtgagtgcggg-3′ (sgRNA-1) and 5′-gcggtgacgtcatgcctcgg-3′ (sgRNA-2). The donor vector consists of the human HTT exon1 with 150 CAGs flanked by two homologous DNA arms (1 kb each) of pig HTT sequences required for homologous recombination. The pig exon1 was completely replaced by human exon1 with Nco1 and Apa1 restriction sites. This replacement resulted in two nucleotides differences (gc in lowercase) as TGAGTgcGGGCCC before Apa1 I site (underline) in pig HTT intron 1. These constructs were confirmed by sequence analysis (IGE, Guangzhou, China). The vector was linearized to release plasmid DNA before electroporation into the fibroblast cells.

Nuclear transfer and animal breeding

The constructed pCMV-Cas9, sgRNA-1, sgRNA-2, the HD donor vector fragment were co-transfected via electroporation (Gene Pulser Xcell, Bio-Rad, Hercules, CA, USA) into cultured pig fetal fibroblasts. Fibroblasts were isolated from a 35-day old fetus of a female Rongshui miniature pig. The transfected cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, HyClone, Logan, UT, USA) supplemented with 15 % fetal bovine serum (FBS, HyClone, Logan, UT, USA) at 39°C in an incubator with 5 % CO2. After the 24 h r ecovery, the cells were split into single cells and cultured for 10 days to form the colonies. The colonies of surviving and individual cells were passed into 48-well plates and analyzed via PCR to identify the colonies that replaced pig HTT exon1 with the human HTT exon1 containing the large CAG repeat. Using the same method described in our previous studies (Lai et al., 2002; Yang et al., 2010), we collected porcine oocytes for somatic cell nuclear transfers (SCNT). Cumulus-oocyte complexes (COCs) were cultured in maturation medium for 42–44 h at 39°C. Denuded mature oocytes were en ucleated, and the human HTT exon1 knock-in cells were used as donor cells for injection into the perivitelline space of oocytes. Two successive DC pulses at 1.2 kV/cm for 30 μs parameter via Electro Cell Manipulator 200 (Genetronics, San Diego, CA, USA) were used for cell fusion and in vitro fertilization. Reconstructed embryos were transferred to culture medium for incubation overnight at 39°C. Embryos (120–150) were then surgically transferred into the oviduct of a surrogate pig in estrus. Pregnancy was confirmed 30–35 post-transplantation days. The surrogate sows delivered the cloned piglets naturally. The F1 KI pigs were generated by crossing the female KI (F0) founders to wild-type male Bama miniature pigs, and male their offspring were crossed with female wild-type pigs to generate the F2 KI pigs.

PCR and genotyping

Cultured pig fibroblast cells were lysed in 15 μl of lysis buffer (0.45 % NP-40 plus 0.6 % Proteinase K) for 90 min at 56°C and then 10 min at 95°C to is olate DNAs for PCR. The primers were designed to amplify DNAs containing the homologous arm and the CAG repeats. The HD KI pigs were identified by PCR using primers for the left arm (Forward primer LF: 5′-GGA GAG CTG GGA GAG AAT GCC AGT GTG ACA GT -3′ and reverse primer LR: 5′-GCGGCTGAGGCAGCAGCGGCTGTGCCTG-3′) and right arm (Forward primer RF 5′-GGC CTT CGA GTC CCT CAA GTC CTT CCA G-3′ and reverse primer RR: 5′-GCTTCTTGAAAGCCGTCCTCATGAAATGCCTTCCGT-3′). The PCR conditions were 94°C for 5 min; 94°C for 30 s, 65°C for 30 s, 72°C for 1 min 3 0 s, for 35 cycles; 72°C for 5 min, and held at 12°C. The genomic DNA isolated from the blood of pi gs was used for genotyping and sequencing.

Western blot analysis, immunohistochemistry, and electron microscopy

Antibodies to expanded polyglutamine repeats (1C2, Millipore, MAB1574), γ–tubulin (Sigma, T6557), NeuN (Millipore, MAB377), GFAP (Millipore, AB5804), IBA1 (WAKO, 019-19741), DARPP32 (Chemicon, AB1656), Calbindin-D28k (Sigma, C9848), Neuropeptide (NPY) (Abcam, ab123951), Parvalbumin (Abcam, ab11427), and ChAT (Chemicon, AB5042) were used. Mouse anti-HTT (mEM48) was generated by us and used in the previous studies (Li et al., 2003; Yang et al., 2010). Secondary antibodies were all from Jackson Immunolabs. For western blotting, pig brain tissues were lysed in ice-cold RIPA buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1 mM EDTA pH 8.0, 1 mM EGTA pH 8.0, 0.1% SDS, 0.5% DOC, and 1% Triton X-100) containing Halt Protease Inhibitor cocktail (Thermo Scientific) and PMSF. The tissue lysates were incubated on ice for 10 min, centrifuged at 10,000 rpm for 10 min. The supernatants were loaded to SDS-PAGE, transferred to a PVDF membrane, and blocked with 5% milk/PBS for 1 h at room temperature. Primary antibodies were diluted in 3% BSA/TBST (50mM Tris-Hcl, pH 7.4 with 20μM Tween 20) and incubated with the blot membrane overnight at 4°C. The blotted membran e was washed in TBST 3 times (each time 5 min), followed by incubation with HRP-conjugated secondary antibodies in 5% milk/TBST for 1 h at room temperature. After three washes in TBST, ECL Prime (GE Healthcare) was used to detect immunoreactive signals on the membrane. For immunohistochemistry, isolated pig brain tissues were fixed 24 hours in 4% paraformaldehyde in 0.01 M PBS and then transferred into 30% sucrose to dehydrate at 4°C. The pig brain tissues were sec tioned at 20 μm using a freezing microtome. The slides were fixed for 10 min in 4% paraformaldehyde in 0.01 M PB, pre-blocked in 4% normal goat serum in 0.1% Triton X-100/PBS for 30 min. Slides were incubated with primary antibodies in 3% BSA/2%NGS/1×TBST overnight at 4°C. Secondary antibo dies were added after three washes with PBS. Fluorescence imaging was recorded after covering the slides with mounting buffer. The DAB staining was performed using the Avidin-Biotin Complex kit (Vector ABC Elite, Burlingame, CA, USA). Microscopic images were obtained using a Zeiss LSM 800 confocal microscope and Zeiss Axio Imager A2 inverted scope.

Quantification of neuronal cells and glial cells in F0 KI pig brains were obtained as described earlier (Yang et al., 2017). For a specific brain region in each animal, the relative numbers of NeuN, DARRP-32, calbindin-D28k, parvalbumin, NPY, ChAT, GFAP, or IBA1 positive cells per image were obtained. For GFAP staining, ImageJ software was also used to measure GFAP immunostaining intensity. The percentage of specific cell types of total cells was obtained by collecting more than six non-overlapping fields (20×) in each brain section, and at least 3 brain sections in each animal were examined. The average numbers of specific types of cells per image (20X) were used for statistical analysis.

The core electron microscopy facility at Emory University performed the EM examination as described in our previous studies (Li et al., 2003). Freshly isolated pig brain tissue blocks were fixed with 4% paraformaldehyde and 0.2% glutaraldehyde for 48 hours and sectioned using a vibratome. All sections used for electron microscopy were dehydrated in ascending concentrations of ethanol and propylene oxide/Eponate 12 (1:1) and embedded in Eponate 12 (Ted Pella Inc., Redding, CA). Ultrathin sections (60 nm) were cut using a Leica Ultracut S ultramicrotome. Thin sections were counterstained with 5% aqueous uranyl acetate for 5 min followed by Reynolds lead citrate for 5 min and examined using a Hitachi (Tokyo, Japan) H-7500 electron microscope.

Behavioral analysis

The movement capabilities of the HD KI pigs were examined as described in our previous study (Yang et al., 2014). For treadmill test, the pigs were placed on a treadmill in a closed cage to assess their running ability. A suitable cage was placed onto the treadmill to keep the pig and make it run on the conveyer belt. The animals were trained for 3 consecutive days before testing occurred. The speed of treadmill was kept at 3.0 km/h for 1 min when the pig ran on it. The better running time was recorded as the efficient running time. The treadmill running test was performed once a month after the pigs were grown. Because an HD KI pig died after the formal test of its motor function using treadmill test, we did not test all HD KI pigs for their treadmill performance. For gait test, we used a footprint tracking method following the same way to measure footprints of HD mice (Wang et al., 2008). HD KI pigs were trained to pass through the sandy pathway (80 cm wide and 4.5 m long), and their footprints during the movement process were recorded as pictures taken by a camera. The stride lengths for front and rear footprints on the pictures were measured. Wild type and HD KI pigs (n=3 each group) were examined to compare their stride lengths. Pig body weight was measured monthly, and animal mortality was recorded for survival time. For the symptomatic HD KI pigs, the video was used to record the dyskinesia and abnormal movements, and the gender- and age-matched wild type pigs served as controls for video recording.

Whole genome sequencing analysis

The Annoroad Company conducted whole genome sequencing of the pig genomic DNA and prepared the DNA library. The pig genome (Sus scrofa) was used to produce a custom-made index file, and PEMapper was used to map sequencing data. For each off-target locus, 50 bp of flanking region was added to each side of the locus for analysis. We next used the pileup file generated by PEMapper to retrieve the base-pair-level sequencing read coverage and reported the average. Top 20–25 off-target loci were selected by Blast analysis of the HTT gRNAs sequences in the pig genome for comparison with the non-targeted wild-type pig. Relative sequencing depth for most likely off-target loci by HTT gRNA-1 and gRNA-2 was calculated by normalizing the number of mapped reads in those loci to the genome-wide average of mapped reads.

QUANTIFICATION AND STATISTICAL ANALYSIS

MRI analyses

All animals were anesthetized with Zoletil TM 50 (Virbac S.A., France) before Magnetic Resonance Imaging (MRI) assessments. We examined three wild type pigs and four HD KI pigs at the age of 5 months. The images were acquired with a 3.0T clinical MRI scanner (Magnetom Symphony; Siemens, Erlangen, Germany). T2-weighted and T1-weighted MRI images of 1 mm slice thickness were obtained from each pig. The volumes of the striatum and lateral ventricle in MRI were calculated by Cavalieri method (Ekinci et al., 2008; Erbagci et al., 2012), which is one of the stereological techniques. Briefly, the converted JPEG image suitable for calculation was automatically placed over the image by the “Stereo Investigator Software”. A marker was selected and points that fell on the brain, striatum, and lateral ventricle were marked. After all data were entered to the system and the areas were marked, total brain, striatum, and lateral ventricle were calculated by the software program by multiplying the image voxel size with the number of voxels in it.

Stereology and quantification

For analyzing the relative numbers of neuronal and glial cells in specific brain regions in F1 KI pig brains, we performed unbiased stereology. Stereological cell counting and quantification were performed as described in our previous study (Xiang et al., 2014). At the age of 4–5 months, the brains of three HD KI pigs (F1–9, −14, −15) and three WT pigs were fixed with 4% paraformaldehyde in 0.01 M PBS (pH 7.4) immediately after death. The examined brain regions were cut in 40-μm serial sections, every tenth section was used for analysis. The selection includes levels throughout the dorsal caudate nucleus and putamen (on sections containing the external globus pallidus). For analyzing the cortex, the sections were obtained from the prefrontal cortex. For analyzing the cerebellum, section selection includes levels through the anterior lobe. The sampling of tissue sections followed a systematic, uniform, random sampling scheme to ensure an equal probability of being sampled. More than 8 sections in each brain region were obtained for stereological examination.

To assess the relative number of neuronal cells, astrocytes, and microglia cells in different brain regions of the same animals, we performed NeuN, GFAP, and IBA1 staining of the caudate nucleus, putamen, prefrontal cortex, and cerebellum of WT and HD KI pigs. To quantify NeuN-, GFAP-, and IBA1-positive cells, the optical-fractionator method was used, as implemented in the semiautomatic stereology system StereoInvestigator 5.4.3 (MicroBrightField, (Microbrightfield, Willston, VT, USA). The volume of the examined brain regions and the total number of NeuN-, GFAP-, and IBA1- positive cells in the examined brain regions were calculated by Stereo Investigator software. Quantification of cell number within the different brain regions was performed at X40 using a Zeiss AX10 microscope within a 200 μm X 200 μm grid size by two observers blind to experimental groups. The total volume of the striatum was measured by Cavalieri method (Gundersen et al. 1999). The same sections of the cell counts were used to estimate the volume of the examined brain region, which was calculated by multiplying the area by the average measured section thickness for each region and subject. The total number of NeuN-, GFAP-, and IBA1- positive cells was then divided by the volume to yield cell density and is presented as the number of positive cells per cubic mm.

Statistical analysis

When every two groups were compared, statistical significance was assessed using the two-tailed Student’s t-test. One-way ANOVA was used when analyzing multiple groups. For pigs that were repeatedly subjected to behavioral tests, we analyzed the data using two-way ANOVA. Data are presented as mean±SEM. Body weight and survival rate of HD KI pigs were obtained by monitoring 6 F0 KI pigs, 15 F1 KI pigs, and 21 age-matched wild type pigs. For pathological examination and western blots, at least 3 animals per group were used. Calculations were performed with GraphPad Prism software (GraphPad Software). A p-value of 0.05 was considered statistically significant.

DATA AND SOFTWARE AVAILABILITY

The CAG repeat sequences in HD KI pigs

The accession number for the raw sequencing reads of the CAG repeats (150, 226 CAGs) in HD KI pigs reported in this paper is GenBank XXXX (submission number 2087893).

Supplementary Material

1

Video-1 The abnormal walking and breathing as well as chorea movement of F0 KI (F0–5 and F0–6) pigs, related to Figure 1 and 2.

10. Figure S1. The genotyping of the HD KI pigs, related to Figure 1.

(A) PCR analysis of targeted allele in blood cells of the HD F0, F1, and F2 KI pigs. The PCR was performed using primers that could amplify the right arm of homologous DNA targeted in the endogenous locus of the pig HTT gene. (B) DNA sequences of CAG repeats-containing HTT gene in F2 KI pig cortex.

11. Figure S2. Behavioral analysis of HD KI pigs, related to Figure 2.

(A) Photos of HD KI (F0–7) founder (arrow) at the age of 15 months (left panel) and HD KI (F0–6) and WT pigs at the age of 5 months during treadmill running test (right panel). (B) Photos show HD F1 KI pigs at the age of 5 months. Arrows indicate symptomatic HD KI (F1–11 and F1–13) pigs. The age-matched wild-type control pig is next to the HD KI pig (right panel). (C) Photos show newborn HD F2 KI pigs at 7 days after birth.

12. Figure S3. Off-target assay and whole genome sequencing, related to Figures 1.

Whole genome sequencing analysis shows no off-target mutations in the cortex of F1 KI pig. Genomic DNAs from the cortical tissues of F1 KI (F0–6, F1–14, F1–15) pig were subjected to whole genome sequencing. Relative sequencing depth for most likely off-target loci by HTT gRNA-1 and gRNA-2 was calculated by normalizing the number of mapped reads in those loci to the genome-wide average of mapped reads. The on-target rate is not shown, as Cas9/sgRNA, generated nicks followed by homologous recombination that resulted in the same sequences as wild-type HTT for the on-target. Mismatched nucleotides are indicated in red.

2

Video-2 Treadmill test of F0 KI (F0–6) and WT pigs, related to Figure 1 and 2.

3

Video-2 Treadmill test of F0 KI (F0–6) and WT pigs, related to Figure 1 and 2.

4. Figure S4. The early death of HD KI pigs and instable CAG repeats, related to Figure 3.

(A) The pathology of the HD KI pig (F0–3) lung is evident by edema with hemorrhaging (arrows, upper panel) and extensively dilated alveolar lumen (red arrow) that is revealed by hematoxylin-eosin staining (lower panel). Scale bars: 50 μm. (B) Quantification of the area and length of alveolar lumen indicates the dilation of the alveolar lumen in the HD KI lung. *** p<0.001 by Student’s t-test. (C) The death and the age at death as well as the CAG repeat numbers of HD KI pigs in different generations. (D) Representative PCR analysis results showed different sizes of HTT DNAs containing an expanded CAG repeat in different tissues of F0–5 and F1–15 pigs. (E) DNA sequencing revealed different repeat numbers in different tissues in F0–5, F0–6, F1–14, and F1–15 pigs.

5. Figure S5. Reduced NeuN-positive cells in the HD KI pig brain, related to Figures 4 and 5.

(A) The brain cortex, striatum, and cerebellum from wild-type (WT) and HD F1 KI (F1–14) pigs at the age of 5 months were stained by immunofluorescent staining with anti-NeuN to label neuronal cells and DAPI to label the nuclei of cells. Note that the striatum of the HD KI pig brain shows the most significant reduction of NeuN-positive cells. (B) Quantitative analysis of NeuN-positive neuronal cells of the total cells per image (20X) in WT and HD KI pig brains. The average numbers of NeuN-positive cells were obtained by examining at least 8 randomly selected images per brain section, 6 sections from 3 animals per group, and data are presented as mean ± SEM, * P<0.05, *** P < 0.001 by Student’s-test. Scale bars: 100 μm.

6. Figure S6. Increased IBA1-positive microglial cells and selective degeneration of medium spiny neurons in the HD F1 KI (F1–15) pig brain, related to Figure 6.

(A) The brain striatum from wild-type (WT) and HD F1 KI pigs (F1–15) at 5 months of age were stained with antibodies to IBA1 for identifying microglial cells. Scale bars: 50 μm. (B) Quantitative analysis of IBA1-positive microglial cells per image (20X) in WT and HD KI pig brains. The average numbers of IBA1-positive cells were obtained by examining at least 6 randomly selected images per brain section, 6 sections from 3 animals per group, and data are presented as mean ± SEM, ***P < 0.001 by Student’s-test.

(C) Immunostaining of the F1 KI pig striatum (caudate and putamen) with anti-calbindin-D28k showing a fewer numbers of calbindin-D28k-positive cells in the caudate than the putamen. (D) Quantification of the Calbindin-D28k-positive neurons per image (20X) in the caudate and putamen of the KI pig. Scale bar: 20 μm.

7. Figure S7. Degenerated neurons and reactive glial cells in the F1 KI (F1–15) pig brain cortex, related to Figure 7.

(A) Electron microscopy revealing a dark neuron that is adjacent to an astrocyte (AS) in the middle photograph. Two degenerated neurons in the middle and right photographs show different appearances (electronic cytoplasmic dense or lucent) without clear nuclear membrane and structure. (B) Different types of reactive microglia (middle photograph) and astrocytes (right photograph) in the KI pig brain show the enlarged cytoplasm with various sizes of vacuoles and degenerated organelles as compared to normal microglia (M) and astrocyte (AS) in a wild type (WT) pig brain. Scale bars: (A), 2 μm; (B), 1 μm.

8

Video-3 The abnormal movement and breathing of F1 KI pigs, related to Figure 1 and 2.

9

Video-4 The eating behavior of F1 KI pig, related to Figure 1 and 2.

Highlights.

• CRISPR/Cas9 and somatic nuclear transfer enable the generation of HD knock-in pigs

• CAG repeat expansion and phenotypes in HD knock-in pigs are germline transmittable

• Mutant huntingtin causes selective neurodegeneration in specific brain regions

• HD knock-in pig brain shows the preferential degeneration of medium spiny neurons