Generation of HIP14 BAC mice & animal breeding strategy

A ∼190 kb long human Bacterial Artificial Chromosome (BAC) containing the entire HIP14 gene (RP11-463M12) was obtained from the Children's Hospital Oakland Research Institute (CHORI) (Figure S1a).

BAC DNA was prepared according to a standard protocol obtained from the Rockefeller University NINDS GENSAT BAC Transgenic project (http://www.gensat.org/GensatProtocols.pdf), microinjected into FVB/N fertilized embryos, and implanted in pseudopregnant female mice. Seven human-specific primers spanning the HIP14 gene and surrounding sequence were designed in order to characterize founders carrying the entire construct, and for subsequent routine PCR genotyping (Table 1).

Animal breeding strategy

BAC founders were identified using a standard genotyping protocol. The highest-expressing line (HB4) was chosen for further study, and BAC,Hip14+/−×HIP14+/− matings were set up to obtain mice for use in the study, namely wildtype (WT) (Hip14+/+), Hip14−/− (Hip14−/−), and the humanized BAC mice (BAC,Hip14−/−).

Mice were genotyped at wean using tail clips, with two PCR reactions: One for the human HIP14 BAC (Table 1, primer pair 5), and one for mouse Hip14 as described previously [23]. Genotypes of mice were re-confirmed at time of sacrifice.

Transgene copy number assessment by quantitative PCR

A primer pair specific for human HIP14 exon 1 (Table 1, primer pair 8) was designed using Primer Express 3.0 software (ABI). Relative transgene copy number was assessed on the ABI 7500 Fast system using Power SYBR Green PCR Mastermix (Applied Biosystems) and the relative quantitation settings. Signal was normalized to mouse actin (Table 1, primer pair 11).

Isolation of RNA & quantitative RT-PCR

Total RNA was isolated using the Qiagen RNeasy Plus Mini kit according to the kit instructions. Samples were subsequently treated with DNase I (Invitrogen). cDNA was generated using the SuperScript® III First-Strand Synthesis System (Invitrogen). qRT-PCR was performed on the ABI 7500 Fast Real Time PCR System (Applied Biosystems).

Primers specific for human HIP14 mRNA (Table 1, primer pair 9) were used in Figure 1a. For Figure 1b, primers specific for murine Hip14, annealing 3′ to the gene-trap site in intron 5 were used. Relative gene expression was normalized to mouse actin (as above).

Preparation of protein lysate and Western Blots

1 month mouse cortex or striatum tissue was homogenized using a Dounce homogenizer on ice in 1 volume of TEEN (50 mM Tris pH 7.5, 1 mM EDTA, 1 mM EGTA, 150 mM NaCl) +1% SDS buffer. After initial lysis, 4 volumes of TEEN +1% Triton X-100 was added. The sample was passed through a 25 gauge needle 5 times and sonicated for 5 seconds on 20% power, spun at 14,000 rpm at 4 degrees for 15 minutes, and the supernatant transferred to a new tube. Protein concentration was determined using the Bradford Assay (Biorad). The above buffers were supplemented with the following reagents: 1× Complete Protease Inhibitors (Roche), 1 mM sodium orthovanadate (Sigma), 800 mM PMSF (Sigma), 5 mM zVAD (Calbiochem). Equal amounts of protein were loaded and run on 4–12% Bis-Tris SDS-PAGE gels (Invitrogen). Immunoblots were obtained using an in-house polyclonal rabbit HIP14 antibody (PEP1) described previously [18]. Beta tubulin antibody T4026 was assessed as a loading control (Sigma).

Stereology & Neuropathological Assessments

Mice aged 1 month received tail injections of heparin and were terminally anesthetized by intraperitoneal injection of avertin, followed by intracardiac perfusion with fresh cold 4% paraformaldehyde in PBS, pH 7.0 for 10 minutes. Neuropathological assessments were performed as previously described [24]. All analysis was performed with the researcher blind to genotypes.

Quantification of DARRP-32, Enkephalin and Substance P Immunohistochemistry staining

Immunohistochemistry was performed on 1 month old mice of mixed gender.

Primary antibodies were: mouse anti-DARPP32, #C24-6a (from Dr. P. Greengard and Dr. H. Hemming), rabbit anti-enkephalin (Chemicon), rat anti-substance P (Accurate Chemical). Sections were incubated with horseradish peroxidase (HRP) conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) followed by DAB and 0.5% cresyl violet and visualized using a Zeiss Axioplan 2 microscope. Staining was quantified using MetaMorph (Universal Imaging Corporation). Relative levels of staining were calculated as the sum of the integrated optical density divided by the area selected, then multiplied by the sampling interval (×8) and section thickness (25 µm). The individual values were normalized to wild type.

Behavioral analysis

Mice were acclimatized to the behavioral testing holding room under reverse lighting (12–10 pm dark cycle) at least one week prior to behavioral testing at 3 months. Group-housed mice of both genders were assessed. Single-housed mice were removed from the analysis, and the experimenter was blind to genotype.

Palmitoylation assays

Protein palmitoylation was assessed using the Acyl-Biotin-Exchange with Immunoprecipitation (ABE/IP) method [36] as described previously [23]. For PSD-95, rabbit polyclonal antibody used for IP was generously provided by the late Dr. Alaa El-Husseini. PSD-95 mouse monoclonal antibody MAI-25629 was used for western blot (Affinity Bioreagents, Golden, CO).

For SNAP-25, a mouse monoclonal antibody SMI-81 (Covance, Emeryville, CA) was used for IP, and a rabbit polyclonal SNAP-25 antibody used for western blot (Synaptic Systems #111 002). Palmitoylation (the Biotin-BMCC label) was detected using a Streptavidin Alexa Fluor 680 conjugate antibody (Molecular Probes #S-32358). All palmitoylation assessments were done on whole brain of mice of combined gender aged 1–3 months. Each sample was split in two and processed as technical replicates, in order to reduce assay variability.

Body weight

Body weight of mice was recorded at 3 & 6 months of age prior to behavioral testing.

Statistical Analysis

All statistical analyses were performed using the Graphpad Prism software, version 5a. Parametric analysis of single timepoint data was performed using one-way ANOVA, with post-hoc Tukey test. Non-parametric analysis (for western blot and palmitoylation assays) was done using the one-way ANOVA Kruskal-Wallis test with post-hoc analysis using Dunn's Multiple comparison test. Behavioral data collected over multiple time points is assessed using repeated measures ANOVA with Bonferroni post-hoc analysis. Data are reported as mean ± SEM.

Generation of HIP14 BAC mice

A total of 494 FVB/N embyros were injected with a human HIP14 BAC containing the entire HIP14 gene, 378 of which survived and were implanted to pseudopregnant females. Forty-five live-born pups were produced, of which 11 were positive for the BAC by PCR genotyping. Nine of these 11 pups produced a PCR product for all primer sets (Figure S1b), indicating integration of an intact BAC construct. Two of eleven lines integrated a partial construct (Table 2).

Assessment of transgene copy number by qPCR in the founders indicated HB4 and HB6 as the two lines with the highest copy number (Figure S1c & d). These were selected for further study, and found to have comparable levels of HIP14 expression (data not shown). Line HB4 was crossed to the existing Hip14−/− line in order to asses for rescue in the present study.

mRNA levels in WT, Hip14−/−, and BAC mice

Intron-spanning primers specific for human HIP14 transcript confirmed the presence of human HIP14 expression in mice carrying the BAC transgene (BAC). An absence of amplification in the WT or Hip14−/− mice (Figure 1a) confirmed that this transcript arises from the human HIP14 transgene.

As mRNA expression had not been previously assessed in Hip14−/− mice, we investigated transcript levels in WT, Hip14−/−, and BAC mice. A murine Hip14−/− specific intron-spanning primer was designed to anneal downstream (exon 14–15) of the gene trap used to create the Hip14−/− mice, which lies in intron 5. Murine Hip14 transcripts were significantly reduced in Hip14−/− compared to WT littermates, but still detectable in Hip14−/− cortex. This was not significantly affected by the presence of the human HIP14 BAC in BAC mice (Figure 1b).

The human HIP14 transgene generates a modest level of human HIP14 protein expression

We confirmed protein expression of HIP14 from the transgene using a previously described HIP14 antibody [18], which detects both mouse and human HIP14; thus, the band detected represents total HIP14 protein. A faint band visible in Hip14−/− lanes was confirmed to be non-specific, as it is not eliminated upon peptide competition assay using a >500 molar excess of the peptide used to generate the antibody (Figure S2). The human HIP14 transgene is expressed at 36% and 35% of WT levels in striatum (Figure 1c) and cortex (Figure 1d) respectively. Protein expression was normalized to a beta tubulin loading control, and values are shown relative to wildtype levels. HIP14 levels were significantly increased in BAC mice relative to Hip14−/− in both striatum (p = 0.0091 normalized to WT, p = 0.0083 as raw values) and in cortex (p = 0.0016 normalized to WT, p = 0.019 as raw values).

Human HIP14 compensates for the neuropathological deficits in Hip14−/− mice

Mice lacking murine Hip14 demonstrate neuropathological deficits, including a 17% loss in striatal volume by embryonic day E17.5, and an accompanying reduction in striatal neuron count [23]. In order to assess whether the protein expressed from the human HIP14 BAC is functional, we assessed a series of similar neuropathological endpoints in Hip14−/− mice carrying the human HIP14 BAC transgene, as compared to wildtype and Hip14−/− littermates.

Brain weight is significantly decreased in Hip14−/− mice throughout their lifespan [23]. In the current study, whole brain weight and cerebellum are significantly decreased at 1 month in Hip14−/− mice relative to WT; BAC mice are not significantly different from WT in whole brain, cerebellum, or forebrain weight (Figures 2a–c). One-way ANOVA analysis shows significant rescue in cerebellum (Figure 2b) and a trend to restoration of WT values in whole brain (Figure 2a) and forebrain (Figure 2c). Similar observations are seen at 3 months (data not shown).

Similar to previous findings [23], Hip14−/− mice demonstrate a 15.7% reduction in striatal volume at 1 month. This is fully rescued to wildtype levels in Hip14−/− carrying the BAC transgene (Figure 2d). Striatal neuron count is reduced by 14.3% at 1 month in Hip14−/− mice and this is also significantly rescued to WT values in BAC (Figure 2e). Finally, an 11.9% reduction of cortical volume in Hip14−/− at 1 month is rescued in the presence of human HIP14 (Figure 2f).

Human HIP14 restores levels of DARPP32 and enkephalin in Hip14−/− MSNs

The striatal neuron populations affected in HD consist of GABAergic projection neurons and parvalbuminergic interneurons, while other neuron populations in the striatum remain relatively spared. The majority (95%) of striatal neurons consist of projection neurons; therefore, the striatal atrophy and cell loss observed in HD is largely due to loss of striatal GABAergic MSNs, cells that express high levels of DARPP-32 [37]. These MSNs are subdivided into two types: those expressing enkephalin and dopamine D2 receptors, and those expressing substance-P and dopamine D1 receptors. Levels of enkephalin and DARPP-32 are affected early in HD, whereas substance P remains largely unchanged until later in the disease [37]. As in HD patients, Hip14−/− mice demonstrate reduced levels of DARPP-32 and enkephalin in the striatum, while levels of substance-P are unchanged [23]. Similar to previous findings, striatal levels of DARPP-32 were reduced by ∼28% in Hip14−/− mice as compared to wildtype, and this was restored to wildtype levels in Hip14−/− expressing the human HIP14 BAC (Figure 3a). Enkephalin levels were reduced by ∼56% compared to wildtype, while the BAC mice were similar to wildtype (Figure 3b). Striatal levels of substance-P were similar for all genotypes (Figure 3c). This data demonstrates that BAC-derived human HIP14 compensates for the neurochemical deficits observed in the Hip14−/− mice.

Motor coordination impairments & altered locomotor activity in the Hip14−/− mouse is restored to normal by human HIP14

Hip14−/− mice demonstrate deficits in motor coordination as early as 3 months of age [23]. We therefore assessed mice for performance on an accelerating rotarod at 3 and 6 months of age. Repeated measures ANOVA reveals a significant effect of genotype (Figure 4a). BAC mice performed consistently better than Hip14−/− littermates, remaining on the rotarod apparatus for a longer time before falling. BAC performance was also superior to WT littermates, and this was more apparent at later trials. A non-significant trend of WT performance superior to Hip14−/− is apparent. The consistently superior performance in BAC mice suggests that human HIP14 BAC can completely compensate for motor coordination deficits seen in Hip14−/− mice.

Hip14−/− mice are hyperactive in various measures of dark-phase assessment of spontaneous activity, similar to the observations of hyperactivity in young YAC128 mice [24], [25]. Spontanous activity measures were assessed at 3 and 6 months of age (Figure 4b–j). The most robust changes were observed in stereotypic time (Figure 4c) and counts (Figure 4f), where repeated measures ANOVA revealed a significant effect of genotype. Post-hoc Bonferroni testing revealed an increase in both measures in Hip14−/− vs. WT, and a robust rescue in BAC mice. A significant effect of genotype was also observed in ambulatory time (Figure 4b) and counts (Figure 4e), distance traveled (Figure 4d), time resting (Figure 4g), and ambulatory episodes (Figure 4j), where Hip14−/− mice displayed hyperactivity and a normalization to wildtype values in BAC was observed. In summary, human HIP14 compensates for the behavioral deficits observed in mice lacking murine Hip14.

Palmitoylation deficits of HIP14 substrates in Hip14−/− mice are normalized to wildtype levels by human HIP14

HIP14 functions as a PAT for a number of critical neuronal proteins, including PSD-95 and SNAP-25 among others [19]. Palmitoylation of these neuronal proteins is decreased in the Hip14−/− mouse model, and the defect in palmitoylation may underlie defects in trafficking observed in HD [23]. We sought to confirm that the functional enzyme activity of BAC-derived human HIP14 is intact, and to assess whether the human protein product can compensate for the loss of murine HIP14. Similar to previous findings, we observed reduced palmitoylation of PSD-95 in the Hip14−/− mice (Figure 5a). Palmitoylation of SNAP-25 was similarly reduced in the Hip14−/− mice, and returned to wildtype levels in the BAC mouse (Figure 5b).

The reduction in body weight seen in Hip14−/− mice is only partially rescued by human HIP14

One of the features of HD in human patients is a progressive loss of weight [15], and Hip14−/− mice fail to gain weight [23]. We therefore assessed whether BAC-derived human HIP14 can compensate for this deficit. Hip14−/− mice aged 3 months demonstrate reduced body weight relative to WT, and this is partially rescued by human HIP14 in BAC mice in both genders (Figure 6). This pattern is more apparent by 6 months of age in both females (Hip14−/− 82.5% and BAC 87.6% of WT) and males (Hip14−/− 83.5% and BAC 90.8% of WT). Weekly measurements in a small subset of mice revealed that a similar pattern of partial rescue in body weight is present at early as 2 weeks of age (data not shown).

Conclusions

In summary, we have generated the first transgenic mouse model for HIP14, and demonstrate intact palmitoyl-transferease activity in a transgenic DHHC PAT model. Humanized mice for HIP14 were able to rescue defects in neuropathology, behavior, and HIP14 PAT activity, and partially restored body weight to wildtype levels. We report that human HIP14 compensates for the key features resulting from loss of the murine ortholog at an expression level equivalent to only ∼1/3 that of endogenous murine HIP14, which is sufficient for full rescue of most phenotypes.

This latter finding carries important therapeutic implications. We have previously shown that HIP14 is dysfunctional and displays impaired PAT activity in the brain of the YAC128 mouse model of HD despite normal levels of HIP14 protein [23]. The current study suggests that therapeutic interventions in human HD patients could be directed toward approaches to increase HIP14 activity, and that rescue of some phenotypes in YAC128 mice may be achieved with less than wildtype levels of functional HIP14.