|Home | About | Journals | Submit | Contact Us | Français|
The pathologic hallmark of pseudoxanthoma elasticum (PXE) is ectopic mineralization of soft connective tissues. Recent studies have suggested that PXE is a metabolic disease, and perturbations in a number of circulatory factors have been postulated. One of them is fetuin-A, a 60-kDa glycoprotein synthesized in the liver and secreted into blood. Observations in targeted mutant mice (Ahsg−/−) and in cell culture model systems have demonstrated that fetuin-A is a powerful anti-mineralization factor in circulation, and the serum levels of fetuin-A both in patients with PXE as well as in a mouse model of PXE (Abcc6−/−) have been shown to be reduced by up to 30 percent. In this study, we tested the hypothesis that over-expression of fetuin-A in Abcc6−/− mice counteracts the ectopic mineralization. Delivery of a construct containing full-length mouse fetuin-A cDNA, linked to a His-tag, to the liver of these mice resulted in elevated serum levels of this protein. As a consequence, soft tissue mineralization, a characteristic of Abcc6−/− mice, was reduced by ~70 percent at 12 weeks of age. The results suggest that normalization of serum fetuin-A, either through gene therapy approaches or by direct protein delivery to the circulation, may offer novel strategies to treat PXE and perhaps other heritable disorders of soft tissue mineralization.
Pseudoxanthoma elasticum (PXE), a heritable multi-system disorder, is characterized by ectopic mineralization of soft connective tissues, with primary clinical manifestations in the skin, the eyes and the cardiovascular system (Neldner and Struk, 2002; Li et al., 2008). The classic forms of PXE are caused by mutations in the ABCC6 gene, which encodes ABCC6 protein, a putative transmembrane transporter expressed primarily in the liver and the kidneys (Pfendner et al., 2007, 2008; Li et al., 2008). The precise function of ABCC6, including its transport substrate(s), and the detailed mechanisms of the aberrant mineralization processes are currently unknown. It is clear, however, that the soft tissue mineralization plays a critical role in clinical manifestations, such as intermittent claudication, gastric bleeding, and early myocardial infarcts in the cardiovascular system. Furthermore, mineralization of Bruch’s membrane in the retina leads to angioid streaks and bleeding in the eye, leading to loss of visual acuity and blindness. Finally, histopathology reveals that the yellowish papules and inelastic plaques of the skin consist of accumulation of mineralized connective tissue, primarily pleiomorphic elastic fibers. Currently, there is no specific or effective treatment for PXE, but prevention of ectopic mineralization would be expected to ameliorate the clinical manifestations.
A number of gene/protein systems have been shown to participate in the control of the mineralization processes under physiologic conditions. For example, utilization of targeted mutant mouse (“knock-out”) models have demonstrated that proteins, such as fetuin-A, matrix gla protein, osteocalcin and ankylosis protein, apparently prevent premature mineralization under physiologic homeostasis of normal serum calcium and phosphate concentrations (Jahnen-Dechent et al., 1997; Luo et al., 1997; Ryan, 2001; Wang et al., 2005). In this study, we have focused on the potential of utilizing fetuin-A as an inhibitor of mineralization in a mouse model of PXE (Abcc6−/−), which we have generated by targeted ablation of the Abcc6 gene (Klement et al., 2005). These mice recapitulate features of human PXE, including extensive mineralization of soft connective tissues in the skin, the eyes, and the cardiovascular system. Similar to PXE, mineralization in these animals is late-onset (~5–6 weeks of age), and the calcium deposition is progressive throughout the life. A consistent finding in this mouse model is early-onset and progressive mineralization of the connective tissue capsule surrounding the vibrissae in the muzzle skin, a feature that we have previously suggested to serve as an early biomarker of the overall mineralization process (Jiang et al., 2007). The mineralization of the connective tissue capsule can be quantitated either by computerized morphometric analysis following histopathologic stains or by direct biochemical assay of calcium/phosphate content of the muzzle skin.
Fetuin-A (also known as the α2-Heremans-Schmid glycoprotein; the corresponding mouse gene symbol being Ahsg) is a systemic inhibitor of mineralization (Schäfer et al., 2003). It is highly expressed in the liver and secreted into circulation, and the plasma concentration of fetuin-A varies with age (Yang et al., 1992; Kazi et al., 1998). The importance of fetuin-A in preventing aberrant mineralization has been demonstrated by development of a mouse model in which the Ahsg gene was inactivated by targeted ablation. The Ahsg−/− mice developed severe calcification of various organs when the mice were placed on a diet rich in minerals and vitamin D, or on normal diet when the genetic deficiency was combined with a DBA/2 genetic mouse background with propensity for mineralization (Luo et al., 1997). It should be noted that the Ahsg−/− phenotype is not associated with changes in serum calcium or phosphate concentrations.
Two lines of evidence suggest that fetuin-A may contribute to the aberrant mineralization process in PXE. First, we have previously demonstrated that fetuin-A, which is expressed in the liver and subsequently transported in circulation, co-localizes with the mineral deposits in the connective tissue capsule of vibrissae in Abcc6−/− mice (Jiang et al., 2007). This was accompanied by ~20 percent reduction in fetuin-A levels in serum of Abcc6−/− mice at six months of age, as compared to age-matched wild-type controls. Secondly, immunologic measurements of fetuin-A levels in sera from 110 patients with PXE revealed that fetuin-A concentrations in these individuals’ sera were lower than those in 80 healthy, unrelated controls (0.55 ± 0.11 vs. 0.80 ± 0.23 g/l; mean ± SD, respectively; p<0.001) (Hendig et al., 2006). Thus, these studies suggest that fetuin-A, a major systemic inhibitor of mineralization, is deficient in PXE, and restoration of normal fetuin-A levels may counteract the progressive mineralization of connective tissues in PXE.
The purpose of this study was to over-express fetuin-A in a mouse model recapitulating features of PXE, including ectopic connective tissue mineralization. The hypothesis tested postulates that restoration of serum fetuin-A to normal levels, or exceeding them, counteracts the pathogenic mineralization process in this disease. For this purpose, an expression construct pLive-Fetuin-A consisting of an albumin promoter and full-length mouse fetuin-A cDNA linked to a His-tag was developed (see Materials and Methods). Several control experiments were performed to ensure the expression of this construct. First, MLE cells were transfected with this expression vector in culture, and the expression was followed by both RT-PCR as well as by immunofluorescence (Fig. 1). Before transfection these cells express essentially undetectable levels of fetuin-A mRNA (Fig. 1, lanes 4, 5), while after transfection, a specific band of 500 bp was readily observed by RT-PCR (lanes 1, 2). This band was similar to the band obtained using mouse liver mRNA as template (Fig. 1, lane 6). The possibility of DNA contamination was excluded by treatment of the same RNA samples with RNase, which abolished the signal (lane 7, 8 and 9). Immunofluorescence with anti-mouse fetuin-A antibody stained a select population of transfected MLE cells in culture, and the same cells were positive for staining with an anti-His antibody (Fig. 1B). Collectively, these observations indicate that the construct used for transfection was functional leading to expression of fetuin-A both at the mRNA and at protein levels.
Targeting of the expression constructs pLive-LacZ and pLive-Fetuin-A to the liver in vivo was tested by injecting them into the tail vein of Abcc6−/− mice using the hydrodynamic delivery method. This technique has been previously shown to deliver expression constructs to the liver with high efficiency (Zhang et al., 1999; Jiang et al., 2006). Injection of 30 μg of pLive-LacZ construct confirmed expression of this transgene in the liver by β-galactosidase staining of the entire liver lobes en block (Fig. 2A, upper panel) or by staining paraffin sections of the liver (Fig. 2A, lower panel). The expression was clearly detectable at one week after injection, appeared to be maximal at two weeks, and gradually declined so that little if any expression was noticeable at four weeks. The expression of the fetuin-A was examined by similar injection of the pLive-Fetuin-A expression vector to the tail vein of Abcc6−/− mice. Serum samples were obtained from the mice at 0, 2, 3 and 4 weeks after the injection, and the level of fetuin-A was determined by ELISA. The data were then compared to fetuin-A levels in Abcc6−/− mice simultaneously injected with the pLive-LacZ vector. A significant, close to 2-fold, increase in serum fetuin-A levels was noted after two weeks of injection (Fig. 2B). These results indicate that the system allows significant expression of the injected transgene to be expressed in the liver and this leads to a significant increase of fetuin-A levels in the serum of the corresponding mice.
The expression of fetuin-A, eight weeks after the injection, was also determined in the liver by quantitative RT-PCR. Small, but significant (p<0.01) increase, ~1.5-fold on the average, of the mRNA levels was noted in the liver of mice injected with the pLive-Fetuin-A construct, as compared to the pLive-LacZ vector (data not shown). Expression of recombinant fetuin-A in the liver of the mice injected with the pLive-Fetuin-A construct was also examined by immunofluorescence with the antibody recognizing the His-tag in the construct. Distinct cells with expression of the construct were noted in mice treated with pLive-Fetuin-A, while the immunofluorescence was negative in the control mice treated with the LacZ expression construct (Fig. 3A). Expression of the fetuin-A protein in the liver was also examined by Western analysis, which demonstrated enhanced expression in the mice treated with pLive-Fetuin-A construct (Fig. 3B).
The clinical correlate of this study revolves around the hypothesis that over-expression of fetuin-A can counteract the mineralization process in Abcc6−/− mice. To specifically test this hypothesis, we injected either the pLive-LacZ or pLive-Fetuin-A vector targeting the liver as above. The injections were repeated four weeks later, and the degree of connective tissue mineralization was determined at eight weeks of the original injections, i.e., at the age of 12 weeks of the mice. First, areas of muzzle skin containing vibrissae were biopsied and the total calcium content was determined by a chemical assay. The content of calcium, corrected for the tissue weight, was reduced in the biopsies obtained from Abcc6−/− mice treated with pLive-Fetuin-A, as compared to the pLive-LacZ controls (Fig. 4A). Secondly, reduced mineralization of the connective capsule in vibrissae of Abcc6−/− mice was also demonstrated by direct histopathologic evaluation upon staining with H&E or Alizarin Red (Fig. 4B). Computerized morphometric analysis of H&E stained sections revealed that the Abcc6−/− mice treated with the pLive-Fetuin-A construct had over 70 percent decrease in the degree of mineralization at eight weeks subsequent to gene delivery, as compared to the 12-week old mice treated with pLive-LacZ (p<0.05) (Table 1).
PXE is a multi-system, autosomal recessive disorder characterized by dystrophic mineralization of soft connective tissues in a number of organs. Clinically, the manifestations of classic PXE center on three major organ systems, the skin, the eyes, and the cardiovascular system (Li et al., 2008). The primary cutaneous lesions are small yellowish papules on the predilection sites at flexural areas, and the lesions progressively coalesce into larger plagues of inelastic, leathery skin. Histopathologic evaluation of skin reveals accumulation of basophilic elastotic material, as revealed by H&E stain, and characteristically these elastotic structures become mineralized in a progressive manner over the lifetime of the affected individuals. While PXE is associated with considerable morbidity and occasional mortality from cardiovascular complications, the phenotypic spectrum is highly variable with both inter- and intra-familial heterogeneity, and involvement of any given organ system may predominant in certain families. This variability has presented a diagnostic challenge, compounded by the fact that clinical manifestations are rarely present at birth and often do not become evident until the second or third decade of life.
The classic form of PXE is caused by mutations in the ABCC6 gene on chromosome 16p13.1, and to date over 300 distinct mutations have been described, representing over 1000 mutant alleles (Pfendner et al., 2007, 2008). The types of mutations include missense and nonsense mutations in 28 of the 31 exons of the gene, intronic mutations causing miss-splicing, small deletions and insertions, as well as large deletions spanning part or the entire coding region and sometimes including flanking genes as well. Identification of mutations in the ABCC6 gene can be used for confirmation of the clinical diagnosis, carrier detection, and pre-symptomatic identification of affected individuals with a family history of PXE. In spite of the impressive progress in molecular diagnostics in PXE, no clear explanation for the phenotypic heterogeneity is currently available, and careful comparisons of the mutation database with phenotypic details have been unyielding (Pfendner et al., 2007). It is evident, however, that the extensive and progressive connective tissue mineralization, the pathologic hallmark of PXE, is the cause of clinical findings in the affected tissues.
We have previously suggested that PXE is a metabolic disorder, and that in the absence of ABCC6 transporter activity in the liver, serum becomes deficient of a factor(s) necessary for prevention of aberrant, unwanted mineralization under normal homeostatic calcium and phosphate concentrations (Uitto et al., 2001; Jiang and Uitto, 2006). This hypothesis has been supported by our grafting models of Abcc6−/− and wild-type mice, as well as by in vitro mineralization models indicating that Abcc6−/− mouse serum has less ability to prevent the mineral deposition induced by inorganic phosphate in a cell culture system (Jiang et al., 2007, 2008). Thus, a number of serum factors, including fetuin-A, matrix gla protein, osteocalcin and ankylosis protein have been identified as potential candidate molecules for perturbations in PXE.
Fetuin-A is a 60-kD glycoprotein implicated as an anti-mineralization factor under physiologic conditions (Yang et al., 1992; Luo et al., 1997; Schäfer et al., 2003). Specifically, targeted ablation of the corresponding gene, Ahsg, results in severe calcification in mice, a process exacerbated by diet rich in minerals and vitamin D and amplified by genetic mouse background with propensity for mineralization. We have also demonstrated that in an in vitro aortic smooth muscle cell culture system, fetuin-A counteracts inorganic phosphate-induced mineralization in the presence of Abcc6−/− mouse serum (Jiang et al., 2007). In addition to these experimental data, the fetuin-A levels in sera from patients with PXE as well as in Abcc6−/− mice have been found to be reduced by about 20–30% in comparison to corresponding controls (Hendig et al., 2006; Jiang et al., 2007). The genetic evidence attesting to the role of fetuin-A as a systemic inhibitor of mineralization has been complemented by ultrastructural and biophysical data suggesting that fetuin-A forms colloidal spheres with calcium and phosphate which remain soluble (Heiss et al., 2003). Conceptually, these “calci-protein particles” may provide a mechanistic explanation as to how fetuin-A inhibits the mineralization process. Extensive mutation analyses have also demonstrated that the calcium phosphate precipitation inhibitory activity by fetuin-A resides in the amino-terminal cystatin-like domain which consists of a dense array of acidic residues on the extended β-sheet (Heiss et al., 2003).
The data presented in this study demonstrate that over-expression of fetuin-A can prevent ectopic mineralization of connective tissues in a mouse model of PXE developed by targeted ablation of the Abcc6 gene. These findings have been corroborated by in vitro findings in a cell culture system (Jiang et al., 2007), and collectively, these observations suggest that fetuin-A could potentially serve as a pharmacologic compound that could halt, and even perhaps reverse, the mineralization process which is the principal cause of clinical manifestations in PXE. Delivery of fetuin-A to the mice could utilize three potential strategies. First, delivery of the expression vectors to the liver, with subsequent integration to the genome, could potentially allow sustained expression of this anti-mineralization protein, analogous to the findings in our study. Secondly, ex vivo transfection of hepatocytes or stem-cell populations with potential for hepatocytic differentiation, with subsequent transplantation into the liver, would allow sustained expression of the transgene. Finally, direct administration of recombinant protein to the circulation could potentially elevate the fetuin-A levels and counteract the mineralization process. The latter strategy is obviously dependent on the stability and half-life of the exogenously delivered protein. These issues can be addressed experimentally in animal models, such as the Abcc6−/− mouse, and it is conceivable that with further optimization of protein production and delivery, direct protein replacement strategies can be developed towards clinical application of treatment of this currently intractable disease.
All animal studies were performed in accordance with the institutional guidelines of Thomas Jefferson University Animal Care Committee and National Institutes of Health. Mice were placed in cages in a temperature-controlled room with a 12-hr light-dark cycle and free access to food and water.
In all studies, male Abcc6−/− mice (4 weeks of age) were used and divided into two groups: (a) The control group was treated with the pLive-LacZ vector, and (b) the experimental group was treated with the pLive-Fetuin-A vector (see below). Each group consisted of four mice in each experiment.
The pLive™ Vector, which contains an albumin promoter, has been designed for prolonged, high-level expression of transgenes in mouse liver, and pLive-LacZ reporter vector were purchased from Mirus (Madison, WI). The pLive-Fetuin-A carrying the full-length cDNA ofthe mouse fetuin-A was constructed as follows: The mouse fetuin-A full length cDNA was first prepared by PCR amplification of mouse liver cDNA using a sense (5-′ttttcgcggccgctctggagcaa-3′) and an antisense primer (5′-ccaaaaccttgcggccgcgaatc-3′), both containing a Not I restriction site (underlined). The PCR product was separated by agarose gel electrophoresis and extracted from a gel slice (Qiagen, Valencia, CA). The purified fragment was digested with Not I, and subcloned into pcDNA3.1c/His vector (Invitrogen, Carlsbad, CA) to generate a full-length fetuin-A cDNA linked to a His-tag at the 3′ end. The cDNA fragment with His-tag was then cloned after resptriction digestion into the pLive™ vector backbone. The fidelityof the cDNA sequences was verified by automated DNA sequencing. Plasmid DNAs were purified with Endofree Plasmid Maxi Kit (Qiagen) according to the manufacturer’s protocol.
Mouse liver epithelial (MLE) cells were plated prior to transfection and grown to approximately 80% confluency. The cells were transfected with the pLive-Fetuin-A construct using FuGENE 6 transfection reagent according to the manufacturer’s instructions (Roche Diagnostic Co., Indianapolis, IN). Two days after transfection, the expression of mouse fetuin-A was investigated at mRNA level by RT-PCR and at protein level by immunofluorescence.
For RT-PCR, total RNA was extracted from transfected and non-transfected MLE cells with RNeasy Mini Kit (Qiagen, Valencia, CA) by applying the on-column DNase treatment, as recommended by the manufacturer’s instructions. The amount and quality of the RNA were verified by measuring the absorbance at 260/280 nm. The RNA samples were split into two sets. One set was treated with RNase before reverse transcription, while another set of RNA was used directly for reverse transcription. Random-primed reverse transcription of RNA was performed with the SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA), using 1 μg of RNA either treated with RNase or without such treatment. Two μl of RT reaction products were utilized for PCR to amply mouse fetuin-A cDNA. The primer sequences and the PCR conditions are available from the corresponding author upon request.
To investigate the expression of fetuin-A and His-tag fusion protein in the transfected cells, the cells were rinsed in 1x phosphate-buffered saline, then fixed in 50% methanol/acetone for 10 min, and blocked by 3% BSA for 1 hour at room temperature. Cells were then incubated for 1 hour with the primary antibodies, goat anti-mouse fetuin-A (1:200, R&D Systems) and mouse anti-His tag (Qiagen), respectively. The secondary antibody, either FITC conjugated anti-goat IgG or Texas Red conjugated anti-mouse (Jackson ImmunoResearch West Grove, PA), at 1:250 dilution, was applied for 1 hour.
Thirty μg pLive-Fetuin-A or pLive-LacZ construct was delivered into livers of 4-week old Abcc6−/− mice by hydrodynamic tail vein injection of 10% body volume of TransIT-QR hydrodynamic delivery solution (Mirus), as recommended by the manufacturer’s instructions, using a 27-gauge syringe needle. Two separate injections were performed at 4 weeks apart.
To analyze the liver biodistribution and expression of the pLive-LacZ vector, livers were isolated from control group at 1, 2, 3 and 4 weeks after injection and examined for β-galactosidase expression, using either paraffin-embedded sections or entire liver lobes. Six-micrometer sections or liver lobes fixed with 4% paraformaldehyde were stained for x-gal using the In Vivo Gen LacZ Detection kit (San Diego, CA). The sections were lightly counterstained with eosin before mounting.
Serum samples were collected from the mice at 1, 2, 3 and 4 weeks after injection. Fetuin-A levels in mouse serum were determined by an indirect ELISA. Ninety-six-well immunoplates (Nalge Nunc International, Rochester, NY) were coated overnight at 4°C with a monoclonal anti-mouse fetuin-A antibody (R&D Systems), 50 ng/well. The wells were washed three times with phosphate-buffered saline/0.05% Tween and then blocked with 5% BSA at room temperature for 2 hours. A standard curve was constructed by adding recombinant mouse fetuin-A in concentrations up to 50 ng/ml. For assay of mouse serum levels of fetuin-A, serum in 1:5,000 dilution was added. After appropriate incubations and washings, biotinylated goat anti-mouse fetuin-A antibody, 1:1,000 dilution, was added, followed by incubation for 1 hour at room temperature and washings. Horseradish peroxidase-avidin (Zymed, South San Francisco, CA) was then added, incubated for 1 hour, washed five times with phosphate-buffered saline/0.05% Tween, followed by color development using tetramethylbenzidine solution (Pierce, Rockford, IL) in 2 M H2SO4; the optical density was read at 450 nm.
Proteins were isolated from treated Abcc6−/− mouse livers at 8 weeks after the 1st injection and subjected to SDS/PAGE (7.5% gel) in the presence of a reducing agent. Protein was electrotransferred to polyvinylidene fluoride membrane, and nonspecific binding sites on the membrane were blocked by incubation in 5% milk for 1 hour at room temperature. The blot was incubated with anti-fetuin-A antibody at 1:1,000 dilution overnight at 4°C, and then incubated with horseradish peroxidase-labeled anti-goat secondary antibody at 1:80,000 dilution for 1 hour at room temperature. After three 10 min washings, the signal was determined with ECL Plus™ Western Blotting Reagent Pack (Amersham Biosciences, Piscataway, NJ).
Total RNA extraction and random-primed reverse transcription of RNA were performed as described above. SYBR Green PCR amplification of mouse fetuin-A was performed in a model 7000 sequence detector (Applied Biosystems, Foster City, CA, USA). The reactions were carried out in a 96-well plate in a 25 μl reaction volume containing 12.5 μl of 2x SYBR Green PCR Master Mix (Applied Biosystems), 0.3 nM concentration of forward and reverse primer, each, and 10 ng of cDNA. Liver cDNA samples in triplicate were used for each run with a standard protocol. The amount of specific mRNA in each RNA sample was quantified and normalized to Gapdh mRNA. The relative expression level of the target genes was calculated using the DDCt method (Real-time PCR software; model 7000 sequence detector; Applied Biosystems). Reaction specificity was determined by the dissociation curve immediately after the last reaction cycle, and visualized with the software Dissociation Curve 1.0 (7000 sequence detection system; Applied Biosystems).
Muzzle skin from both sides was harvested eight weeks after the 1st injection. One side of tissue was decalcified with 0.15 N HCl for 48 h at room temperature. The calcium content was measured colorimetrically by the o-cresolphthalein complexone method (CALCIUM (CPC) LIQUICOLOR, Stanbio Laboratory, Boerne, TX, USA), and the values for calcium were normalized to tissue weight.
The other side of tissue was embedded in paraffin and 6-μm sections were prepared for Hematoxylin-Eosin (H&E) or Alizarin Red staining. Computerized morphometric analysis of H&E stained sections of muzzle skin was performed, and specifically, the sections were examined with a Nikon model TE2000 microscope equipped with an AutoQuant Imaging system (Watervliet, New York, NY, USA). The number of vibrissae, with and without evidence of mineralization, was determined in several microscopic fields, and the degree of mineralization was expressed as both the percentage of the mineralized vibrissae and the percentage of area of mineralization per area of vibrissae (Jiang et al., 2007; LaRusso et al., 2008a,b).
Results are given as mean ± SE and/or median plus range. The data were analyzed by Student’s t-test, and differences were considered statistically significant at p<0.05.
The authors thank Carol Kelly for assistance. This study was supported by the NIH/NIAMS grants R01AR28450, R01AR52627 and R01AR55225. Dr. Jiang is the recipient of a Research Career Development Award from Dermatology Foundation.