|Home | About | Journals | Submit | Contact Us | Français|
Pseudoxanthoma elasticum (PXE) is an autosomal recessive multi-system disorder characterized by ectopic connective tissue mineralization, with clinical manifestations primarily in the skin, eyes and the cardiovascular system. There is considerable, both intra-and inter-familial variability in the spectrum of phenotypic presentation. Previous studies have suggested that mineral content of the diet may modify the severity of the clinical phenotype in PXE. In this study, we utilized a targeted mutant mouse (Abcc6−/−) as a model system for PXE. We examined the effects of changes in dietary phosphate and magnesium on the mineralization process using calcification of the connective tissue capsule surrounding the vibrissae as an early phenotypic biomarker. Mice placed on custom-designed diets either high or low in phosphate did not show changes in mineralization, which was similar to that noted in Abcc6−/− mice on control diet. However, mice placed on diet enriched in magnesium (5-fold) showed no evidence of connective tissue mineralization in this mouse model of PXE. The inhibitory capacity of magnesium was confirmed in a cell-based mineralization assay system in vitro. Collectively, our observations suggest that assessment of dietary magnesium in patients with PXE may be warranted.
Pseudoxanthoma elasticum (PXE; OMIM264800) is a heritable multi-system disorder caused by mutations in the ABCC6 gene (for review of PXE, see Li et al., 2008a). The pathologic hallmark of this disease is ectopic mineralization of connective tissues, primarily elastic structures, in a number of organs. Clinically, PXE manifests in three major organ systems, the skin, the eyes, and the cardiovascular system. In the skin, there is an accumulation of pleiomorphic elastotic structures in the mid-dermis, which become progressively mineralized. Clinically, these changes manifest as yellowish papules and plaques which tend to coalesce to form inelastic, leathery skin. The eye manifestations consist of angioid streaks resulting from mineralization and breaks in the Bruch’s membrane of the retina, and the ensuing neo-vascularization can lead to bleeding which causes primarily central vision loss. The cardiovascular pathology is due to mineralization of mid-layers of the arterial blood vessels, clinically manifesting with intermittent claudication, bleeding of the gastric blood vessels, and occasionally, early myocardial infarcts.
Timely and accurate diagnosis of PXE is often complicated by the late onset of manifestations and considerable, both intra- and inter-familial heterogeneity (Neldner and Struk, 2002; Li et al., 2008a). The reasons for this phenotypic variability are currently unknown, and attempts to establish phenotype/genotype correlations by comparing the types of mutations in the ABCC6 gene with phenotypic manifestations have been largely unyielding (Pfendner et al., 2007, 2008). It has been suggested, however, that a number of modifier genes may contribute to the clinical presentation, explaining, at least in part, the variability in PXE phenotype (Schön et al., 2006; Hendig et al., 2007; Li et al., 2008b; Vannakker et al., 2007; Zarbock et al., 2007). In addition, early retrospective, largely anecdotal, studies have suggested that high intake of dairy products (rich in calcium and phosphate) during childhood or adolescence correlates with the severity of PXE (Neldner 1988; Renie et al., 1984). Finally, recent preliminary studies suggested that treatment with an oral phosphate binder may halt, and even reverse, the progression of cutaneous and ocular findings in patients with PXE (Sherer et al., 2005). These studies collectively suggested that abnormalities in calcium/phosphate metabolism may contribute to the mineralization process, and therefore to the phenotypic presentation, in PXE.
We have developed a mouse model of PXE by targeted ablation of the Abcc6 gene in both alleles (Klement et al., 2005). These homozygous Abcc6−/− “knock-out” (KO) animals recapitulate the histopathologic and ultrastructural features of human PXE by manifesting with late-onset (~5 weeks) and progressive mineralization of connective tissues in the skin, eyes, and the arterial blood vessels. We have previously used this animal model to explore the effects of dietary modifications on the ectopic mineralization, and our preliminary studies suggested that the mineralization process in PXE may be exacerbated by changes in the mineral intake (LaRusso et al., 2007). In this study, we have specifically examined the roles of phosphate and magnesium by feeding these mice with custom designed diets enriched or depleted in these minerals. Our results indicate that increase in dietary magnesium prevents connective tissue mineralization in this mouse model of PXE.
In this study, we utilized Abcc6−/− mice as a model system for PXE which recapitulate genetic, histopathologic and ultrastructural features of the human disease. While these mice demonstrate ectopic mineralization in a number of tissues, we utilized calcification of the connective tissue capsule surrounding the vibrissae as a reliable biomarker of the mineralization process (Klement et al., 2005; Jiang et al., 2007). We have previously shown that mineralization of this capsule is one of the early signs of the overall mineralization process and is detectable around 5 weeks of age. The progressive, age-associated increase in the mineralization of the vibrissae-associated capsule can be demonstrated by histopathologic examination of sections stained by hematoxylin and eosin (H&E) stain (Fig. 1), or by special stains for calcium or calcium/phosphate complexes (Alizarin Red and von Kossa stains). The degree of mineralization can be quantitated by computerized morphometric analysis of the mineralized area in H&E stained sections or by determination of the calcium and phosphate content in biopsies from the muzzle skin containing the vibrissae (Jiang et al., 2007).
In the experiments performed here, Abcc6−/− mice at four weeks of age, i.e., prior to the onset of mineralization, were placed on specific diets, either high in phosphate, high in magnesium, low in phosphate or low in magnesium, as indicated in Table 1. Eight weeks later, at the age of 12 weeks, the mice were sacrificed, and the degree of mineralization of the vibrissae-associated connective tissue capsule was quantitated.
Examination of the vibrissae of wild-type (Abcc6+/+) mice on control diet did not show any signs of mineralization at 12 weeks of age, while Abcc6−/− mice fed the same control diet demonstrated clear foci of mineralization (Fig. 1, compare two left panels). This difference in mineralization was also demonstrated by quantitative assay of calcium and phosphate contents in the tissue, as well as by computerized morphometric analysis of the area of mineralization (Fig. 2).
As revealed by histopathology in Fig. 1, and quantitated by calcium/phosphate determination and by computerized morphometric analyses in Fig. 2, high or low phosphate content of the diet did not appreciably change the degree of mineralization, as compared to Abcc6−/− mice fed control diet. In contrast, the mineralization process in KO mice fed with high magnesium-containing diet (MgO) was completely abolished, and quantitatively these values were at the same low, essentially undetectable, levels as noted in wild-type Abcc6+/+ mice. While the KO mice fed a diet low in magnesium content showed somewhat elevated mean value of mineralization, as compared to the corresponding mice fed the control diet, this increase was not statistically significant, largely due to considerable variation in the values (Fig. 2).
We have recently adopted a cell-based in vitro system to identify factors that can alter the mineralization process (Jiang et al., 2007). In this assay, human aortic smooth muscle cells are cultured in medium containing either fetal calf serum (10%) or serum from WT (Abcc6+/+) or KO (Abcc6−/−) mice. After having reached ~80% of confluence, 2 mM inorganic phosphate (Pi) is added to the culture medium, eliciting a prompt mineralization of the cell layer. The mineralization can be visualized by phase contrast light microscopy and can be quantitated by the measurements of the calcium deposition in the cell layer by a chemical assay (Fig. 3). Addition of Pi into the culture medium results in ~60-fold increase in calcium deposition in the cell layer either in the presence of FBS or WT mouse serum (Fig. 3b). The degree of mineralization was even higher, about 90-fold, in cultures maintained in the presence of Abcc6−/− mouse serum, consistent with the notion that the KO mouse serum lacks a factor(s) required to prevent soft tissue mineralization under normal calcium and phosphate homeostatic concentrations. However, magnesium in concentrations varying from 0.1 to 1.0 mg/ml prevented the mineralization when added to the cultures maintained in the presence of either WT or KO mouse serum, in a concentration dependent manner (Fig. 3). These observations clearly suggest that the addition of magnesium can prevent mineralization in vitro and that the inhibition of mineralization by magnesium noted in the Abcc6−/− mice is not mediated by systemic factors, such as hormonal regulation.
In subsequent experiments, we allowed the mineralization process to take place in the presence of 2 mM Pi and KO mouse serum for 14 days, after which magnesium in the same concentrations was added to the culture medium (Fig. 4). Over the subsequent 14 days of follow-up, no change in the degree of mineralization was noted either by microscopic examination or by assay of calcium deposition in the cell culture layer. The latter observations suggest that the mineralization in this in vitro system, once it has occurred, can not be reversed by magnesium under these experimental conditions.
To examine the metabolic consequences of the experimental diets, the serum and urine concentrations of calcium, phosphorus and magnesium were determined at the end of the eight-week experimental diet (Tables 2 and and3).3). Relatively little variation was noted in the serum concentrations of these components, and the only statistically significant variation from the control range was reduction in phosphorus concentration in mice fed low phosphate diet (Group D) (Table 2). In contrast, significant changes were noted in calcium and phosphous concentrations in the urine at the end of the eight-week diet. Specifically, over 4-fold increase in calcium concentration and >80% reduction in the phosphorus concentration were noted in mice fed high magnesium diet (Group C) (Table 3). In contrast, the concentration of calcium was reduced by ~85% and the phosphorus concentration was increased by about 50% in mice fed low magnesium diet (Group E) (Table 3). As a result of these changes the Ca/P ratio in Group A (control), Group C (high magnesium), and Group E (low magnesium) were significantly different (p<0.01).
Since inhibition of ectopic mineralization by magnesium could involve changes in the parathyroid hormone (PTH) action (Mori et al., 1992), we also measured the serum concentrations of PTH by an ELISA in mice fed control diet (Group A) or diet high in magnesium (Group C) at 12 weeks of age. Although considerable individual variability in PTH values was noted, the average concentrations in Group A and Group C were not statistically different: 40.5 ± 13.6 and 56.9 ± 26.6 pg/ml, respectively (mean ± S.E.; n = 3–4; p>0.5).
Magnesium, a divalent cation, is an essential nutritional element that is crucial to a number of physiologic processes in humans. For example, magnesium plays a role in energy metabolism, muscle contraction, nerve impulse transmission, and bone mineralization, and it is estimated to be a required co-factor for some 300 enzymes (Baker and Worthley, 2002). Magnesium status is important for regulation of calcium balance through parathyroid hormone-mediated reactions, and secretion of parathyroid hormone and end-organ responsiveness to this hormone are dependent on the availability of magnesium (Mori et al., 1992). In this study, we demonstrate that elevated dietary magnesium prevents connective tissue mineralization in Abcc6−/− mice, a model of PXE. Specifically, while these KO mice reliably develop soft connective tissue mineralization noticeable as early as the fifth week of postnatal life when fed control diet, supplementation of the food with magnesium completely prevented the mineralization up to at least 12 weeks of age.
The mechanisms for the inhibition of ectopic mineralization by magnesium could be systemic interfering with the calcium metabolism mediated through the parathyroid hormone synthesis and/or secretion. In support of this suggestion are observations that magnesium deficiency is associated with insufficient parathyroid hormone action and can lead to reduced responses to calcitropic hormones (Mori et al., 1992). In fact, dietary magnesium reduction to 50% of the nutrient requirement has been shown to disrupt bone and mineral metabolism in rats (Rude et al., 2006). An alternate mechanism may involve direct interactions between magnesium and calcium ions in the mineralization process. Electron microscopic and X-ray microanalysis of the mineral deposits in PXE have shown direct correlation between the crystalline deposits and the presence of calcium, while the analysis showed these deposits being poor in magnesium (Calap et al., 1977). At the clinical level, treatment of patients for soft tissue mineralization with local application of magnesium sulfate in the calcified areas, together with oral administration of magnesium lactate for 4–6 months, has resulted in disappearance of mineral deposits, together with clinical improvement (Ditmar and Steidl, 1989; Steidl and Ditmar, 1990). These observations suggest that dietary magnesium in amounts sufficient to increase the total amount of magnesium in the body may be able to prevent the mineralization of soft connective tissues.
In general, mineralization develops in several contiguous stages: Amorphous calcium phosphate is first deposited at the early stages of the process, followed by gradual transformation into less soluble crystalline, apatite-like structures. In corroboration of our in vivo studies on PXE mice, we were able to prevent the calcium deposition by increasing magnesium in an in vitro assay system in a dose-dependent manner. However, addition of magnesium to the culture system 14 days after the mineralization had taken place was not able to reverse the mineral deposits. This finding could be explained by the continuous process of mineralization towards less soluble forms of crystalline apatite.
Our studies demonstrate that increased dietary magnesium can prevent ectopic mineralization of connective tissues in a mouse model of PXE, a finding that was corroborated by in vitro findings in a cell culture system. These observations suggest that dietary magnesium might also be helpful in patients with PXE, a notion that can be tested in carefully controlled clinical trials. However, additional, long-term preclinical studies utilizing the available animal models are required to test the potential consequences of dietary magnesium on bone metabolism and to monitor toxicity. Finally, it is appropriate to point out that mineral metabolism in the mice may differ from that in humans, and the clinical efficacy and potential side-effects in humans remain to be seen in clinical trials.
In this context, it should be noted that control of mineralization by dietary magnesium may have broader implications beyond PXE in situations such as arteriosclerosis. It is of interest that a study by investigators at the Centers for Disease Control and Prevention (CDC) in the United States, based on a National Health and Nutrition Examination Survey 1999–2000, concluded that substantial numbers of U.S. adults failed to consume adequate amounts of magnesium in their diets (Ford and Makldad, 2003). Collectively, these observations suggest that careful assessment of dietary magnesium, not only in patients with PXE but in populations at large as well, is warranted.
The PXE mouse model was developed by targeted ablation of the Abcc6 gene, as described previously (Klement et al., 2005). Heterozygous alleles were backcrossed for five generations into the C57BL/6J background and interbred to generate Abcc6−/− (KO) and Abcc6+/+ (WT) mice. The mice, housed in the Animal Facility of the Thomas Jefferson University, had free access to water and were maintained in a temperature- and humidity-controlled environment under 12-hour light/dark cycles. Female Abcc6−/− mice were divided into five groups (A, B, C, D, and E) at 4 weeks of age and placed on a special diet with or without mineral modifications in phosphate and magnesium concentration for eight weeks. All diets were specially ordered from Harlan Teklad, Madison, WI, USA. Group A was fed a control diet with the calcium, phosphorus and magnesium content found in a standard rodent diet (Table 1). Group B mice received the same diet but with a two-fold increase in phosphate, whereas Group C was fed the diet with a five-fold increase in magnesium. Groups D and E were fed the diet with reduced phosphate and magnesium contents, respectively. The percentage of phosphate was decreased by 50 percent and the magnesium content reduced by 80 percent. To monitor proper food intake and subsequent weight gain, food was checked daily and mice were weighed weekly.
The animal studies were approved by the Institutional Animal Care and Use Committee of Thomas Jefferson University.
Human aortic smooth muscle cells (Cascade Biologics™; Portland, OR) were cultured in DMEM growth medium. At ~80% confluence, the cells were switched to calcification-inducing medium (DMEM supplemented with 2 mM Pi) or to normal DMEM in the presence of either 10% fetal calf serum, or 10% mouse serum either from Abcc6+/+ or Abcc6−/− animals, and continued to be cultured for up to 2–4 weeks. The medium was changed every 2 days. The first day of culture in calcification-inducing medium was defined as day 0. For prevention experiments, magnesium (MgS04; Fisher Scientific, Fair Lawn, NJ) was added to some cultures at day 0 at the concentration of 0.1, 0.5 or 1.0 mg/ml, for 2 weeks. For reversal experiments, magnesium was added to some cultures at day 15 in calcification-inducing medium at the concentration of 0.1, 0.5 or 1.0 mg/ml, respectively, for additional 2 weeks. All experiments were performed in triplicate.
At day 14 or 28 of cell culture, the media were removed and the cell layers were decalcified with 0.6 N HCl for 24 hours at room temperature. The calcium content of the HCl supernatants was determined colorimetrically by the o-cresolphthalein complexone method (Calcium (CPC) Liquicolor®, Stanbio Laboratory, Boerne, TX, USA). After decalcification, the cells were rinsed three times with phosphate-buffered saline and solubilized with 0.1 N NaOH/0.1% SDS at room temperature. The protein content was measured with Protein Assay Kit (Bio-Rad, Hercules, CA), and the calcium content of the cell layer was normalized to protein content.
To quantify the calcium deposition in mouse vibrissae, the muzzle skin, which contains the vibrissae, was harvested and decalcified with 0.15 N HCl for 48 hours at room temperature. The calcium content was measured as above and the phosphate content was determined with Malachite Green Phosphate Assay kit (BioAssay Systems, Hayward, CA). The values for calcium and phosphate were normalized to tissue weight. Calcium and phosphorus in the serum and urine samples were quantitatively assayed as above.
The magnesium concentrations in the mouse serum and urine were measured using the QuantiChrom™ Magnesium Assay Kit (BioAssay Systems, Hayward, CA).
Mouse serum PTH concentrations were measured using Mouse Intact PTH Elisa Kit (Immutopics, Inc., San Clemente, CA).
For histopathological analysis of mineralization of vibrissae, muzzle skin was fixed in 10% phosphate-buffered formalin, embedded in paraffin, sectioned (5 µm), and stained with hematoxylin-eosin (H&E) using standard techniques.
Computerized morphometric quantification was used to examine H&E-stained sections of muzzle skin. The sections were examined with a Nikon model Te2000 microscope furnished with an Auto Quant Imaging system (Watervliet, New York, NY, USA). The number of vibrissae with and without perceptible mineralization was established in all sections, and the extent of mineralization was described as the percentage of area of mineralization per area of vibrissae.
The statistical differences between the means in groups of mice receiving various diets were calculated by the Student’s two-tailed t-test. Comparisons of continuous measures across all groups were completed using 2-sided Kruskal-Wallis nonparametric tests (Siegel and Castellan, 1988). The Kruskal-Wallis test is comparable to one-way analysis of variance, but without the parametric assumptions. Due to small sample size, yet high computational difficulty, exact tests based on Monte Carlo estimation are provided. For each of the paired group comparisons, an exact 2-sided Wilcoxon test was computed. All statistical computations were completed in StatXact v8.0.
Carol Kelly assisted in preparation of this manuscript. Dr. Terry Hyslop kindly advised in statistical analyses of the data. This study was supported by NIH/NIAMS grants R01 AR28450, R01 AR52627 and R01 AR55225.
CONFLICT OF INTEREST
The authors state no conflict of interest.
The Medical Ethical Committee of Thomas Jefferson University approved all described studies.