Research ArticleAtherosclerosis

Familial Hypercholesterolemia and Atherosclerosis in Cloned Minipigs Created by DNA Transposition of a Human PCSK9 Gain-of-Function Mutant

See allHide authors and affiliations

Science Translational Medicine  02 Jan 2013:
Vol. 5, Issue 166, pp. 166ra1
DOI: 10.1126/scitranslmed.3004853


Lack of animal models with human-like size and pathology hampers translational research in atherosclerosis. Mouse models are missing central features of human atherosclerosis and are too small for intravascular procedures and imaging. Modeling the disease in minipigs may overcome these limitations, but it has proven difficult to induce rapid atherosclerosis in normal pigs by high-fat feeding alone, and genetically modified models similar to those created in mice are not available. D374Y gain-of-function mutations in the proprotein convertase subtilisin/kexin type 9 (PCSK9) gene cause severe autosomal dominant hypercholesterolemia and accelerates atherosclerosis in humans. Using Sleeping Beauty DNA transposition and cloning by somatic cell nuclear transfer, we created Yucatan minipigs with liver-specific expression of human D374Y-PCSK9. D374Y-PCSK9 transgenic pigs displayed reduced hepatic low-density lipoprotein (LDL) receptor levels, impaired LDL clearance, severe hypercholesterolemia, and spontaneous development of progressive atherosclerotic lesions that could be visualized by noninvasive imaging. This model should prove useful for several types of translational research in atherosclerosis.


Atherosclerosis is a low-grade inflammatory disease of the arterial wall caused primarily by cholesterol-containing low-density lipoprotein (LDL) particles in the blood (1). Considering the slow progress of the disease, there is ample time, in principle, to intervene against it with life-style changes, LDL cholesterol–lowering drugs, and other therapies before it leads to heart attack or ischemic stroke (2). Several challenges, however, limit progress in management of the disease. One is the lack of reliable biomarkers that can identify individuals at risk and that can be used to monitor the efficacy of investigational drugs in phase 2 trials (3, 4). Imaging of atherosclerotic plaques may achieve this, and several techniques are already established (57), but near–human-sized atherosclerosis models are needed to validate the methods and to facilitate further technological development (8). Second, drugs and interventional procedures that are effective in treating late-stage atherosclerosis are needed for those patients for whom primary prevention fails. Today, more than 20% of patients with an acute coronary syndrome will suffer another atherosclerosis-related event within 3 years despite intensive therapy (9).

Current mouse models are not supportive of these challenges. Hyperlipidemic mice with genetic defects in lipoprotein clearance, such as apolipoprotein E (apoE) knockout mice (10), have enabled many discoveries of molecular mechanisms in atherosclerosis, but the small size of the animals precludes research in intravascular devices and is a challenge to noninvasive imaging. Furthermore, many important features of human atherosclerosis, including adaptive intimal thickening, plaque angiogenesis, plaque rupture, and thrombosis, are rare or absent in mouse models (11).

Owing to their genomic, anatomical, and physiological resemblance to humans (12), pigs are increasingly used in preclinical studies (13, 14). For instance, to assess safety and efficacy of devices under development for intracoronary application, such as novel imaging and stent technologies, testing in porcine coronary arteries is considered the model of choice (15, 16). However, major limitations of the normal swine include a large, unmanageable size and a low propensity to atherosclerosis even with prolonged feeding of high-fat, high-cholesterol (HFHC) diets. Whereas genetic engineering solved this problem for mouse models 20 years ago (10), similar models have not been available in pigs or other large animals. In the present paper, we created such a genetic model by liver-specific overexpression of the D374Y gain-of-function mutant of the gene human proprotein convertase subtilisin/kexin type 9 (PCSK9) in Yucatan minipigs. This model developed severe hypercholesterolemia and human-like progressive atherosclerotic lesions on a HFHC diet and may be used for translational research in imaging technologies, intravascular devices, life-style factors, and drug therapy.


The PCSK9 gene is conserved between pig and man

PCSK9 binds hepatic LDL receptors (LDLRs), targeting them to lysosomal degradation, and is an important regulator of LDL cholesterol levels in humans and mice (17). Binding involves the catalytic domain of PCSK9 and the first epidermal growth factor–like repeat (EGF-A) domain of the LDLR (18, 19). The gain-of-function mutation D374Y increases affinity of this binding, which causes a severe form of autosomal dominant hypercholesterolemia in humans (20). To predict if human D374Y-PCSK9 would bind the porcine LDLR with high affinity, we searched genomic databases. The LDLR residues demonstrated to be important for binding of human D374Y-PCSK9 (18) were found to be conserved between pig and human (fig. S1). The porcine PCSK9 gene has not been described previously, but we identified a homologous sequence in the porcine CH242-60E11 bacterial artificial chromosome clone from the pig genome project (21). Using the Spidey (National Center for Biotechnology Information) tool to align the human PCSK9 mRNA sequence (NM_174936.3) to the genomic sequence of this clone, we predicted a porcine PCSK9 coding sequence with high homology to the human gene (fig. S2). Notably, the D374 residue was conserved.

Creation and functional testing of D374Y-PCSK9–expressing DNA transposons

Sleeping Beauty (SB) transposon–based vectors were engineered to express human PCSK9 or D374Y-PCSK9 under the control of the liver-specific human α1-antitrypsin promoter and a hepatocyte control region (HCR) from the gene that encodes apoE (Fig. 1A). This promoter provides efficient liver-specific expression in mice (22). The transposon also contained a puromycin resistance gene for selection. We used DNA transposition in HepG2 human hepatocellular carcinoma cells to verify the functional integrity of the transposon-delivered transgenes. As expected, PCSK9 and D374Y-PCSK9 expressed in stable cell lines were cleaved and led to markedly reduced LDLR levels (Fig. 1B).

Fig. 1

SB transposon–based human D374Y-PCSK9 construct. (A) Map of the plasmid. LIR and RIR, left and right inverted repeats, respectively; PGK, phosphoglycerate kinase promoter; Puro, puromycin resistance gene; pA, polyadenylation signal; HAAT, human α1-antitrypsin promoter. (B) The function of the D374Y-PCSK9–encoding transposon was confirmed in stable human HepG2 cell lines. Transposition of PCSK9 (P)– and D374Y-PCSK9 (D-P)–encoding transposons into HepG2 cells confirmed that the protein was expressed and cleaved, and caused efficient reductions in LDLR levels compared with control (C) HepG2 cells. Whole-cell lysates were analyzed.

Cloning of D374Y-PCSK9 transgenic minipigs

D374Y-PCSK9 transgenic primary Yucatan fibroblasts were created by transposition using hyperactive versions of SB transposases (SB100X and HSB3 for male and female cells, respectively). Both transposases were effective in creating puromycin-resistant cell clones compared with a catalytically inactive SB transposase. Resistant clones were pooled and used for somatic cell nuclear transfer by the handmade cloning technique (23), and the cloned embryos were transferred to recipient sows (24). We initially obtained three litters of pigs from seven embryo transfers. In two litters, eight female pigs (designated F1 to F8) were born alive, one (F8) of which died after few days. In another litter, nine male piglets (Fig. 2A) were born but failed to thrive, and despite intensive attempts to hand-rear them, all males were euthanized or died during the first 2 weeks. Autopsies revealed immature organs without specific malformations or signs of disease. From one of the piglets of this litter, fibroblasts were cultured and used for recloning of five live born male pigs. Two of these thrived (designated M1 and M2), one was euthanized for analysis (M3), and two died, probably from aspiration pneumonia.

Fig. 2

Genetic characterization of D374Y-PCSK9 transgenic founder pigs. (A) Newborn litter of male transgenic founders. (B) Southern blot to detect transgene integrations. The transgene was visualized using a probe recognizing a 1-kb sequence in the puromycin resistance gene cassette. The pigs were all transgenic with one or two insertions, except for F6, which harbored a high number of insertions. F5 is shown in a separate blot and was similar to F2, F3, and F7. (C) Genomic mapping of the two transgene insertions in the M1 founder by LDI-PCR. One insertion was mapped to chromosome Xq13 at the TA dinucleotide located at position 61349935 in the large intron 8 of the gene MAGT1 (magnesium transporter 1). The transgene is transcribed in the opposite direction to that of MAGT1. An Spe I site is located 4.6 kb downstream of the transgene integration site, which is consistent with the 6.5-kb fragment in the Southern blot in (B). The other insertion was mapped to chromosome 9q11 at the TA dinucleotide located at position 65836724, which is 10.8 kb upstream of the gene STEAP2 (six transmembrane epithelial antigen of the prostate 2). An Spe I site is located 4.1 kb downstream of the integration site, which is consistent with the 6.0-kb fragment in the Southern blot in (B). The M2 founder was genetically identical to M1.

Mapping of transgene integrations

Eight of the nine founder pigs (F1 to F5, F7, M1, and M2) had one or two transgenes inserted, whereas one (F6) harbored a large number of insertions (Fig. 2B). We mapped the position of a subset of the insertions in the nine pigs by long-distance inverse (LDI) polymerase chain reaction (PCR) and found all to be transposition events with the right inverted repeat flanked by the canonical TA site. One transgene was inserted into a large intron of the MAGT1 gene on chromosome X but transcribed in the opposite direction (Fig. 2C). The other transgenes were located in intergenic regions, one of which was on 9q11 (Fig. 2C and table S1). No integrations of the transposase-encoding plasmid were found.

Phenotype of founder pigs

All piglets expressed plasma D374Y-PCSK9 as determined by an enzyme-linked immunosorbent assay (ELISA) specific for human PCSK9. The range of expression at 14 weeks of age was between 460 and 26,500 ng/ml. Quantitative reverse transcription–PCR (RT-PCR) analysis in one founder pig (M3) showed abundant transgene expression in liver (Fig. 3A). Other tissues showed much lower expression levels, but transgene expression was detectable in heart and kidneys.

Fig. 3

Phenotype of D374Y-PCSK9 transgenic founder pigs. (A) Quantitative RT-PCR tissue analysis in a high-expressing founder, M3 (genetically identical to M1 and M2), euthanized at day 2. Gene expression was normalized to β-actin (ACTB). (B and C) Association between D374Y-PCSK9 expression and total cholesterol (B) and triglycerides (C) on low-fat diet. Data are fasting plasma measurements at 14 weeks of age in founders (n = 9) and age-matched wild-type (WT) pigs (n = 17). High-expressing founders are labeled F6, M1, and M2. (D) Plasma cholesterol in M1, M2, and WT males (n = 4) successively challenged with 4-week periods of HF and HFHC diets. (E) Plasma cholesterol in female founders (n = 7) and WT females (n = 7) challenged with HF and HFHC diets. Asterisk: At a later time point, this pig did not have higher plasma cholesterol than other WT pigs. (F and G) Plasma lipoproteins separated by density (F) or size (G) in M1, M2, and WT controls (n = 4) on a low-fat diet at 14 weeks of age, then on an HFHC diet at 22 weeks of age. (H) Plasma levels of apoB100 on HFHC diet at 22 weeks of age.

On a standard low-fat diet, there was an association between plasma D374Y-PCSK9 and plasma total cholesterol levels (Fig. 3B). Low-expressing D374Y-PCSK9 transgenic founders F1 to F5 and F7 had moderately increased plasma cholesterol (2.95 ± 0.50 mM, mean ± SEM) compared to wild-type Yucatan minipigs of the same age (2.15 ± 0.33 mM) (P < 0.001, t test). In high-expressing founders F6, M1, and M2, plasma cholesterol levels were about threefold higher than in wild-type pigs (Fig. 3B). Fasting plasma triglycerides were not increased by D374Y-PCSK9 expression on standard diet (Fig. 3C).

The D374Y-PCSK9 transgenic founders and age-matched wild-type controls were challenged with successive 4-week periods of feeding with high-fat (HF) diet containing 20% saturated fat and HFHC diet containing both 20% saturated fat and 2% cholesterol. Female founders were randomized to receive the diets in either order. On both diets, M1 and M2 reached much higher total cholesterol levels with mean levels of 10.2 mM on the HF (versus 2.7 mM in wild-type males) and 15.9 mM on the HFHC diet (versus 4.8 mM in wild-type males) (Fig. 3D). The high-expressing female founder (F6) also exhibited significant hypercholesterolemia upon dietary fat challenge, whereas the responses of the lower-expressing founders were less pronounced (Fig. 3E). Further analysis of the M1 and M2 founders showed that the hypercholesterolemia reflected increases in apoB100-containing lipoproteins with the density of LDL, but some of these lipoproteins had the size of very low density lipoprotein (VLDL) in gel filtration profiles (Fig. 3, F to H).

Reduced LDLR levels and LDL clearance in offspring

M1 and M2 founder pigs were bred with wild-type sows, and the inheritance of either of the segregating transgenes on chromosomes 9 and X (Fig. 2C) was found to increase plasma cholesterol in offspring on low-fat diet at 8 weeks of age (table S2). Note that because of the chromosome X–linked transgene, no wild-type females could be obtained in this cross.

The mechanism by which the D374Y-PCSK9 transgene caused hypercholesterolemia was investigated in female transgenic pigs with double transgene insertions. The hepatic LDLR level was reduced by ~90% compared to wild-type pigs (Fig. 4A). Furthermore, plasma clearance of injected human LDL was significantly retarded (Fig. 4B) with half-lives of 34.6 and 19.9 hours in transgenic and wild-type pigs, respectively (P = 0.005 by the extra sum-of-squares F test). Endogenous hepatic expression of porcine PCSK9 was unaltered by expression of the transgene (Fig. 4C), indicating that there was no compensatory regulation.

Fig. 4

Reduced LDLR levels in D374Y-PCSK9 transgenic pigs. (A) Immunoblots of hepatic membrane fractions, with and without an LDLR blocking peptide (BP), show levels of mature and precursor forms of the LDLR in transgenic (n = 4) and WT pigs (n = 4). The panel on the right compares the relative mean density (± SEM) of the mature LDLR band in the blot (n = 4 for each column). P value was determined with t test. (B) Clearance of injected human LDL, analyzed by a human-specific apoB ELISA. Data are mean plasma concentrations ± SEM (n = 4). Best-fit exponential curves with goodness-of-fit R2 are shown. *Comparison of half-lives: P = 0.005 by the extra sum-of-squares F test. (C) Expression of human and porcine PCSK9 mRNA in WT (n = 4) and transgenic pigs (n = 4). Data are means ± SEM.

Severe hypercholesterolemia during HFHC feeding

Wild-type males (n = 7), transgenic males with the chromosome 9 transgene (n = 5), transgenic females with the chromosome X transgene (n = 4), and transgenic females with double transgene insertions (n = 4) underwent prolonged feeding with HFHC diet to facilitate atherosclerosis development. Plasma D374Y-PCSK9 increased significantly on this diet in male and female pigs harboring the chromosome 9 transgene but remained stable in female pigs with the chromosome X transgene only (Fig. 5A). Plasma total and LDL cholesterol levels plateaued after about 20 weeks of feeding, with total cholesterol levels reaching ~20 mM in transgenic pigs (Fig. 5B). The levels of total and LDL cholesterol during the study, measured as area under the curve (AUC), were significantly higher in transgenic compared to wild-type pigs, but no differences in cholesterol levels were observed among groups of pigs with different transgene integrations (Fig. 5, B and C). High-density lipoprotein (HDL) cholesterol increased in all pigs during the first 4 weeks on the HFHC diet and remained elevated (Fig. 5D). Transgenic pigs had lower HDL cholesterol AUC and higher plasma triglyceride AUC compared to wild-type controls, whereas gender or the type of transgene integration did not exert statistically significant effects on HDL cholesterol or triglycerides (Fig. 5, D and E). Pigs had a uniform modest weight gain during the study, and final weight at 1 year of age was limited to 48.7 ± 1.3 kg (n = 20, mean ± SEM) with no differences among groups (Fig. 5F). The lipoprotein profile on low-fat diet was dominated by apoB100-containing LDL-sized lipoproteins (Fig. 5, G and H), but with continued HFHC feeding, apoB100-containing VLDL-sized lipoproteins also accumulated in both transgenic and wild-type pigs (Fig. 5, G and H). Plasma glucose, fructosamine, creatinine, and alanine aminotransferase levels were similar among all groups (fig. S3).

Fig. 5

Severe hypercholesterolemia in D374Y-PCSK9 transgenic offspring during HFHC feeding. (A) Plasma D374Y-PCSK9 levels were measured with a human-specific ELISA. (B to E) Total cholesterol (B), LDL cholesterol (C), HDL cholesterol (D), and triglyceride (E) levels were measured in fasting plasma during HFHC feeding. Data are means ± SEM. AUC is compared among groups with analysis of variance (ANOVA) followed by Bonferroni’s post test. (F) Body weight changes on the HFHC diet. Data are means ± SEM. (G and H) Gel filtration chromatography (G) and immunoblotting (H) show levels of apoB100-containing VLDL-sized lipoproteins on low-fat diet at 8 weeks of age, and after 8 and 28 weeks on HFHC diet. Analyses were performed on pooled plasma from WT (n = 7) and transgenic males (n = 5).

Accelerated development of progressive atherosclerotic lesions

Wild-type (n = 7) and transgenic (n = 13) pigs were euthanized after 46 (43 to 48) weeks of HFHC feeding to analyze the development of atherosclerosis. Because lipid profiles were similar among transgenic females with one or two transgene integrations, females were regarded as one group for the pathological analysis. We tested two hypotheses: (i) whether the transgene accelerated the development of atherosclerosis in male pigs and (ii) whether gender influenced atherogenesis in transgenic pigs. Atherosclerosis in the aorta, left anterior descending (LAD) artery, and iliofemoral artery was quantified, and plaque morphology was typed according to the human classification scheme by Virmani et al. (25) as non-atherosclerotic lesions (intimal thickening or xanthoma) or progressive atherosclerosis lesions (pathological intimal thickening or fibroatheroma). Atherosclerosis in the aorta was distributed in a characteristic pattern with raised lesions in the abdominal aorta and nonraised lesions in the short ascending aorta and aortic arch (Fig. 6A). Mean aortic surface area covered by atherosclerotic lesions was increased 2.1-fold and raised lesions 2.8-fold in transgenic compared with wild-type males (Fig. 6, B and C). Microscopic analysis of the largest raised lesion in the abdominal aorta showed progressive atherosclerotic lesions (pathological intimal thickening or fibroatheroma) in 100% of transgenic males, whereas only 28% of wild-type males had such lesions (Fig. 6C).

Fig. 6

Quantification and classification of atherosclerosis in aortic, coronary, and iliofemoral arteries. (A) Sudan IV–stained aorta (AA, abdominal aorta; TA, thoracic aorta) from a male D374Y-PCSK9 transgenic pig showing both raised (black arrows) and nonraised (white arrows) lesions. (B) Total surface area covered by atherosclerotic lesions in each animal. (C) Aortic surface area covered by raised lesions and plaque classification of the largest aortic lesion in each animal. (D) Average cross-sectional plaque size in the iliofemoral arteries and classification of the most advanced iliofemoral lesion in each animal. (E) Average cross-sectional area in the LAD and classification of the most advanced LAD lesion in each animal. Bars represent means ± SEM in (B) and (C) or median ± interquartile range in (D) and (E). P values for quantitative data were calculated with t test [left panels in (B) and (C)] or Mann-Whitney test [left panels in (D) and (E)]. P values for lesion type distribution were calculated by comparing the frequency of progressive atherosclerotic lesion types (pathological intimal thickening and fibroatheromas) between groups with Fisher’s exact test.

Consistently in iliofemoral arteries, all transgenic males developed progressive atherosclerotic lesions, with the average plaque area increased 7.6-fold (median fold change) compared with wild-type males (Fig. 6D). In the LAD, the cross-sectional plaque area was 1.8-fold (median fold change) higher in transgenic than in wild-type males, but this did not reach statistical significance. However, also in this vascular bed, lesions were significantly more advanced in transgenic males, which all had at least one progressive atherosclerotic lesion in the LAD, whereas this was only the case in 28% of wild-type males (Fig. 6E). No significant gender differences between transgenic males and transgenic females were observed in any of the vascular beds with respect to atherosclerosis quantity or plaque type. Four pigs (three transgenic females and one wild-type male) were outliers with respect to atherosclerosis in the iliofemoral artery and LAD. These pigs were all from the same litter, suggesting that a segregating genetic variant or alternatively intrauterine or postnatal factors may have been involved.

Human-like histological features

Lesion development in transgenic pigs modeled key pathophysiological aspects of human atherosclerosis. Similar to humans—but unlike mice—porcine lesions develop in preexisting intimal masses containing variable amounts of smooth muscle cells and extracellular matrix (26). Lesions resembling human xanthomas with infiltration of foam cells in the luminal intimal layer (Fig. 7A) and pathological intimal thickening with extracellular lipids located deep in the intima (Fig. 7B) were readily observed. Furthermore, fibroatheromas displayed the main components of their human counterpart including necrotic core formation, abundant fibrous tissue, calcification (Fig. 7C), plaque angiogenesis (Fig. 7D), and intraplaque hemorrhage (Fig. 7E). Additional examples of fibroatheromas from the abdominal aorta and iliac artery are shown in fig. S4.

Fig. 7

Histological features and noninvasive imaging. Representative images from the LAD. (A) Foam cell accumulation in the luminal part of an arterial intima containing smooth muscle cells and extracellular matrix. (B) Pathological intimal thickening with signs of extracellular lipids in a thickened arterial intima. (C) Coronary plaque displaying the main features of human fibroatheromas, including necrotic core formation (NC), fibrous tissue (F), and calcification (Ca). (D and E) Coronary plaque areas in (C) show angiogenesis (D) and intraplaque hemorrhage (E). Scale bars, 100 μm (A, B, D, and E); 500 μm (C). (F and G) FDG-PET/CT imaging of the pig in (C) showing FDG accumulation in the vessel wall of the aortic arch (AA) and in LAD. A sagittal (F) and oblique projection (G) is shown. The dashed line in (G) marks the level of the section shown in (C).

Imaging of atherosclerosis

Positron emission tomography (PET) with the tracer [18F]fluorodeoxyglucose (FDG) can detect inflammatory activity in atherosclerotic plaques and is increasingly used as a surrogate endpoint for human phase 2 studies addressing the safety and efficacy of anti-atherosclerotic drugs (5, 6). High reproducibility of the FDG-PET signal from the ascending aorta has been shown (7), making this the segment of choice for analysis. To show how D374Y-PCSK9 minipigs may be used to bridge between preclinical studies and human FDG-PET trials, we performed combined FDG-PET and computed tomography (CT) of the ascending aorta in the HFHC-fed D374Y-PCSK9 transgenic males (n = 5) and some of the transgenic females (n = 4) shortly before euthanization. Little uptake of FDG was detected in the myocardium, whereas clear signal from the aortic wall was obtained (Fig. 7, F and G). The measured target-to-background ratios (TBRs) in the ascending aorta of both males and females was in the upper range of TBRs measured in human studies (5, 6), as would be expected for a model of rapid atherosclerosis development.


Large-animal models that recapitulate human disease pathophysiology are needed for translational research in atherosclerosis, and recent advances in genetic engineering offer the potential to create such models by introducing genetic defects in well-characterized minipig strains. Gain-of-function mutations in PCSK9, of which D374Y is the most detrimental, cause a severe type of familial hypercholesterolemia with rapid development of atherosclerosis in humans and mice (20, 27). Here, we created a porcine model of accelerated atherosclerosis by liver-specific overexpression of human D374Y-PCSK9 in Yucatan minipigs. We and others have reported the feasibility of SB DNA transposition for transgenesis of cloned pigs (28, 29). Insertion of SB transposons is efficient and allows the integration of a clearly defined genetic segment. This is particularly appealing when a tissue-specific promoter drives the transgene, such as in the present study, making preliminary analysis of transgene expression in fibroblasts impossible. By balancing the level of transposition, we were able to produce several different founder animals harboring one to two transgene insertions that could be easily mapped in the pig genome. Active transgene expression was observed in all these animals, which lends support to the notion that the integrity of each transgene cassette is maintained and that analysis of individual cell clones before cloning is not required.

Plasma D374Y-PCSK9 levels in the D374Y-PCSK9 transgenic offspring were 5- to 30-fold higher than normal human plasma PCSK9 levels (median, ~400 ng/ml) (30), and the identical lipid profiles of transgenic pigs with different transgene integrations indicate that the maximal effect on LDL cholesterol levels was obtained with either of the segregating transgenes. Because the transgene in females with only the chromosome X insertion is silenced in ~50% of hepatocytes owing to random chromosome X inactivation, these data also support the predominant role of secreted, rather than intracellular, PCSK9 for LDLR regulation, which has previously been demonstrated in parabiotic mice (31).

LDL cholesterol, but not VLDL or HDL cholesterol, was increased on normal diet in transgenic minipigs, similar to the lipid profile of patients with D374Y-PCSK9 mutations (20). However, the increased triglyceride level reported for this patient group (20) and chow-fed D374Y-PCSK9 transgenic mice (27) was not observed in transgenic founders at 14 weeks of age. In offspring, the interpretation of basal triglyceride levels at 8 weeks of age was made difficult by the decline in levels that occurs in the months after weaning in Yucatan minipigs (32). During prolonged HFHC feeding, the level of VLDL-sized particles increased substantially in both wild-type and transgenic groups, but plasma triglyceride concentrations remained relatively low, although it was significantly increased in the transgenic pigs. This is consistent with our present (Fig. 3) and previous (33) observations that most of the VLDL-sized lipoproteins in pigs fed HFHC diets have the density of LDL. Such triglyceride-poor remnant particles are atherogenic in other models, including apoE knockout mice (10).

Similarities in cardiovascular anatomy with humans make the pig an attractive species in which to study atherosclerosis. Induction of rapid atherosclerosis in normal pigs is difficult, and most previous studies have relied on feeding with cholate or adjuvant atherosclerosis-accelerating procedures, such as balloon injury or streptozotocin-induced diabetes (14, 33, 34). Cholate is necessary to increase plasma cholesterol to atherogenic levels in normal pigs, but it is also a signaling molecule that increases insulin sensitivity (35) and may contribute to chronic inflammation (34, 36), and its use in atherosclerosis studies is therefore advised against (37). We and others have previously worked with a porcine model of familial hypercholesterolemia caused by natural recessive mutations in the gene encoding the LDLR (38). When fed an HFHC diet supplemented with cholate, castrated male pigs of this strain had hypercholesterolemia (~20 mM at 4 weeks declining to ~12 mM at 18 weeks) at a level comparable to that of D374Y-PCSK9 transgenic minipigs fed cholate-free HFHC diet (~12 mM at 4 weeks increasing to ~20 mM from 18 weeks onward) (33), and after balloon injury, they developed coronary atherosclerosis within and proximal to injured sites (33). The propensity of this strain to develop atherosclerosis, however, has not been studied in the absence of coronary instrumentation, castration, or dietary cholate. By contrast, the D374Y-PCSK9 transgenic Yucatan minipigs display consistent development of progressive atherosclerotic lesions with human-like features without the need for adjuvant procedures or cholate. Variability in aortic atherosclerosis was not greater than that commonly observed in inbred mouse models, making them useful for basic research or drug testing. Furthermore, the limited adult size of the Yucatan minipig strain means that clinical equipment and devices can be used without modification while the pigs are still small enough to facilitate easy handling and limit expenses for feeding and drug dosing.

The possibility to test imaging techniques, such as PET/CT, as part of the preclinical testing of a drug or device is an important feature of our model and may be used to guide decisions about such endpoints in subsequent clinical trials. Atherosclerosis imaging endpoints are becoming increasingly important for the drug development decision process (4, 5). The types of drugs that lend themselves to testing in minipigs, however, must be carefully considered. For some classes of drugs, such as lipoprotein-associated phospholipase A2 inhibitors (34), pigs are advantageous compared to other species, but this may not be the case for all types. Although we did not test statins directly, these cholesterol-lowering drugs have been reported to have low or even paradoxical treatment responses in pigs (39).

One limitation of our study is that we did not address whether the D374Y mutation in the PCSK9 transgene was functionally important in pigs. On the basis of sequence analysis, human D374Y-PCSK9 is likely to bind with increased affinity to the porcine LDLR, but at the high expression level obtained, this may not be important. Indeed, in human PCSK9 transgenic mice, the D374Y mutation augments hypercholesterolemia at physiological expression levels (27), but severe hypercholesterolemia can also be obtained by supraphysiological expression of normal human PCSK9 (31).

In conclusion, this report describes a model of familial hypercholesterolemia and accelerated atherosclerosis created by transgenesis in a large animal. The D374Y-PCSK9 transgenic Yucatan minipigs showed severe hypercholesterolemia on HFHC diets and development of progressive atherosclerotic lesions in several clinically important vascular beds. The model should be of general value for translational research involving imaging and intravascular devices and in cases where human-like pathogenesis of atherosclerosis is of critical importance.

Materials and Methods


The Danish Animal Experiments Inspectorate approved all animal procedures. Yucatan minipigs were purchased from Moran Farm Enterprises, and a breeding colony was maintained at our institution (Aarhus University). All pigs were housed in a specific pathogen–free stable facility. Diets were mixed by adding 20% (w/w) of lard with (HFHC) or without (HF) 2% cholesterol (Sigma-Aldrich) to a standard pig feed for growing pigs: 68.0% barley, 15.0% oat, 9.6% soy bean meal, 2.0% animal fat, 3.0% molasses, and 2.4% minerals and vitamins. Pigs on long-term HFHC diet were fed ad libitum until they reached a weight of 20 kg, after which feeding was reduced to 700 g divided into two daily portions.

DNA transposition

SB transposon donor plasmids derived from pSBT/PGK-puro (previously referred to as pTpuro) (40) were generated as described in the Supplementary Methods. Female fetal fibroblasts from the Yucatan minipig were cultured from fetuses at gestation day 44 (Supplementary Methods). Male neonatal fibroblasts were cultured from ear biopsies obtained 2 days after birth. For DNA transposition in male neonatal fibroblasts, cells were seeded at low densities (about 0.1, 0.25, and 0.5% confluency) in P10 dishes and, on the following day, transfected with 3 μg of pSBT/ApoEHCR-hAATp-D374Y-PCSK9 plasmid together with either 3 μg of pCMV-HSB3 [encoding the hyperactive HSB3 transposase (41)] or pCMV-mSB [encoding a catalytically inactive transposase (42)] in 18 μl of FuGENE6 (Roche). For transposition in female fetal fibroblasts, cells were seeded in six-well plates at a density of 2 × 105 cells per well and, on the next day, transfected with 1.0 μg of pSBT/D374Y-PCSK9 and 20 ng of either pCMV-SB100X [encoding the hyperactive SB100X transposase (43)] or pCMV-mSB. From 2 days after transfection, cells were selected with puromycin (1.0 μg/ml; Invitrogen) for 12 days.

Cloning and embryo transfer

The methods used to clone the piglets from fibroblasts have been published (23) and is described in further detail in the Supplementary Methods. Morulae and blastocysts cloned from D374Y-PCSK9 transgenic fibroblasts were collected on days 5 and 6 to be surgically transferred to Danish landrace sows on day 4 or 5 after heat, which was observed 5 days after weaning (24). Pregnancies were confirmed by ultrasonography on day 28, and piglets were delivered by caesarian section or vaginal birth 24 hours after induction with prostaglandin as described previously (24).


Pigs were sedated with an intramuscular injection of midazolam (1 mg/kg) and azaperone (8 mg/kg) and euthanized with a lethal dose of pentobarbital (4 ml/kg) after heparin (10,000 IU) was injected to prevent blood coagulation. The heart, aorta, and iliofemoral arteries were immersed in 4% phosphate-buffered formaldehyde for 6 hours followed by storage in cold phosphate-buffered saline. The aorta was immersed in Sudan IV solution (Sigma-Aldrich; 5 g/liter in 96% ethanol) for 5 min followed by 90-s washout in 96% ethanol. Scanned digital images were obtained, and a custom computer program, NPhase (SoftXCell), was compiled that allowed automatic quantification of sudanophillic lesions in addition to manual outlining of raised lesions that did not accumulate Sudan IV.

In each aorta, the largest raised plaque was identified by macroscopic inspection and excised for histological analysis. The proximal 3 cm of the LAD was sliced into 3-mm slices. The proximal 10 cm of the right common iliac artery and its extension into the external iliac artery and femoral artery were divided into 1-cm slices. Tissues were paraffin-embedded, sectioned (3-μm sections), and stained with hematoxylin and eosin using routine methods. Cross-sectional plaque area was measured with ImageJ (National Institutes of Health). All aortic, coronary, and iliofemoral sections were scored according to the classification of Virmani et al. (25), and the most advanced plaque type in each vessel was recorded for between-group comparisons. All microscopic analysis was performed blinded.

Statistical analysis

Statistical analyses were performed in Prism (GraphPad Software Inc.). Multiple group comparisons were performed with one-way ANOVA followed by Bonferroni’s post test. To test for differences in time series measurements, AUC was calculated and compared. Prespecified two-sample comparisons (for example, the effect of the transgene on atherosclerosis and the effect of gender) were done with Student’s t test or the Mann-Whitney test, as appropriate. Dichotomous data were ordered in contingency tables, and Fisher’s exact test was used to assess differences. Significance level was set at P < 0.05.

Supplementary Materials


Fig. S1. Alignment of the EGF-A domain of the human and porcine LDLR sequences.

Fig. S2. Alignment of the human and the predicted porcine PCSK9 protein sequences.

Fig. S3. Plasma glucose, fructosamine, alanine aminotransferase, and creatinine during HFHC feeding.

Fig. S4. Examples of histology from the abdominal aorta and iliac artery.

Table S1. Overview of mapped integration sites.

Table S2. Plasma cholesterol in D374Y-PCSK9 transgenic offspring.

Table S3. Primer sequences.

References and Notes

  1. Acknowledgments: We thank H. Kristiansen and all of the personnel at AU-Foulum’s stable facilities for neonatal care, blood sampling, and practical handling of pigs; K. D. Winther for veterinary consultations; and A. Pedersen, J. Adamsen, R. Kristensen, K. Villemoes, A. K. Nielsen, B. Synnestvedt, D. W. Jørgensen, L. M. Røge, Z. P. Nasr, L. D. Schröder, C. Berthelsen, and K. Rasmussen for providing excellent technical assistance. Funding: Danish Council for Independent Research|Medical Sciences, Lundbeck Foundation, Danish Heart Foundation, and Danish National Advanced Technology Foundation. Author contributions: R.H.A., E.F., and J.F.B. planned and conducted the phenotypic analysis of transgenic pigs. C.B.S., J.G.M., and J.F.B. designed and constructed the transgene. P.M.K., B.M., S.P., and J.F.B. planned and performed cell experiments. T.T. and M.B.M. contributed to the phenotypic analysis of pigs. R.H.A. and L.P.T. performed and analyzed PET/CT scans. P.M.K., Y.D., J.L., and Y.L. performed cloning supervised by G.V. and H.C., and M.S. performed embryo transfers. C.C., R.H.A., T.L., and L.B.N. performed analysis of plasma lipoproteins and other biochemical parameters. C.B.S., L.B., J.G.M., E.F., and J.F.B. planned the study and supervised experiments. R.H.A., E.F., J.G.M., and J.F.B. wrote the manuscript. All authors read and commented on the manuscript. Competing interests: Aarhus University has submitted a patent application involving the D374Y-PCSK9 transgenic pigs with J.F.B., C.B.S., P.M.K., L.B., J.G.M., and E.F. listed as inventors. L.B. is on the board of directors of Pixiegene A/S (Copenhagen, Denmark) as a scientific advisor. G.V. is partly employed by BGI Ark Biotechnology (Shenzhen, China). The other authors declare that they have no competing interests. Data and materials availability: All raw data are available upon request. D374Y-PCSK9 transgenic minipigs are available from Pixiegene A/S, which has acquired a license from Aarhus University to distribute the pigs. SB100X can be obtained under a materials transfer agreement with Z. Izsvak or Z. Ivics (Max Delbruck Center, Berlin).
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article