Topic Article
Pseudoxanthoma Elasticum Is a Metabolic Disease
Qiujie Jiang, Masayuki Endo, Florian Dibra, Krystle Wang and Jouni Uitto
Journal of Investigative Dermatology (2009) 129. 348-354; doi:10.1038/jid.2008.212

Pseudoxanthoma Elasticum: New Insights

Robb Marchione ¹, Nancy Kim ¹ and Robert S. Kirsner ¹

Journal of Investigative Dermatology (2009) 129, 258; doi:10.1038/jid.2008.407

Pseudoxanthoma elasticum (PXE) is an inherited disorder in which defects in the ABCC6 gene lead to calcification of connective tissue of the skin, eyes (Bruch’s membrane of the retina), and blood vessels (Neldner and Struk, 2002). Although the exact mechanism of how ABCC6 causes disease is not known, several hypotheses have been generated (Pfendner et al., 2008). The “metabolic hypothesis” suggests that gene defects lead to the absence of circulating factors normally needed to prevent mineralization in various tissues (Jiang et al., 2007), whereas the “PXE cell hypothesis” proposes that resident cells are altered by the genetic abnormality, leading to disease manifestations (Quaglino et al., 2000).

Although experimental and observational support exists for both hypotheses, a newly developed animal model clarifies the basis of the disease (Jiang et al., 2007). The PXE knockout mouse model demonstrates a delayed, slow onset of disease manifestations similar to those observed in patients and consistent with the metabolic hypothesis. Using this knockout model, in which mineralization of tissue begins about 5 weeks after birth, Jiang et al. (2009) were able to graft muzzle skin (with distinct vibrissae that typically mineralize in PXE) onto the back skin of ABCC6 discordant mice. Grafting experiments were performed in a bidirectional fashion between knockout and wild-type mice in two different time frames. Grafting experiments performed at 6 weeks enabled investigators to characterize disease progression and revealed that knockout skin grafted onto wild-type mice did not develop mineralization of the vibrissae, whereas wild-type skin grafted onto knockout mice did. Grafting experiments performed at 12 weeks enabled investigators to determine whether disease regression was possible because mineralization ordinarily occurs before 12 weeks of age. These experiments revealed that mineralized vibrissae from knockout mice did regress, to some extent, after being grafted onto wild-type mice. In summary, these experiments support the hypothesis that PXE has features of a metabolic disease.

Through the following questions, we examine this paper in greater detail.

QUESTIONS
1. How does PXE manifest clinically?
2. How is a knockout mouse created, and why and how was it used in this study?
3. Describe the method and techniques employed in this study.
4. What are the findings of this study?
5. What insights into PXE do the results imply?
6. What may be the clinical implications of this article?

ANSWERS
1. Pseudoxanthoma elasticum (PXE) is an inherited, clinically heterogeneous disorder, exhibiting variable manifestations, even among affected siblings. The condition primarily affects the skin, retina, and cardiovascular system. Typical skin lesions begin as yellowish papules that coalesce into plaques with a cobblestone appearance similar to that of the skin of a plucked chicken. Plaques first develop in the cervical area and then become more prominent in the flexural areas, manifesting as redundant axillary skin folds. The cervical lesions tends to arise between the ages of 8 and 12 years, with flexural plaques developing in the teenage years. Although skin biopsy is required for diagnosis (see below), Lebwohl et al. (2003) observed that the length of the horizontal chin crease is a valuable sign, with high diagnostic specificity in patients under 30 years of age. Rare cutaneous manifestations associated with PXE include acneiform papules on the neck and trunk, elastosis perforans serpiginosa, and a reticulated pigmented eruption.

The presence of elastic tissue in the thin membrane between the retinal pigment epithelium and the choriocapillaris, known as Bruch’s membrane, is responsible for PXE involvement of the eye. Bruch’s membrane becomes calcified and brittle, resulting in multiple cracks with resultant scarring as the patient ages. Ophthalmoscopically, one can observe grayish irregular lines, termed “angioid streaks” and resembling vasculature, radiating from the optic papilla. Although angioid streaks are usually not the direct cause of compromised vision, a complication of their formation is aberrant choroidal neovasculature with leaky, brittle vessel walls. These neovessels result in recurrent retinal hemorrhages manifesting as macular symptoms, peripapillary atrophy, disciform macular/foveal scarring, and central vision loss. Legal blindness is not uncommon among patients with PXE. Other ocular abnormalities include a mottling of the fundus (“peau d’orange” appearance), drusen, and comet-like streaks (Chassaing et al., 2005).

PXE-related cardiovascular disease results from the abundance of elastic fibers in arterial walls and manifests as occlusive arterial disease and/or mucosal bleeding. The elastin-rich internal elastic laminae of small and middle-sized vessels develop an atheroma histologically indistinguishable from atheroma due to other causes (e.g., tobacco use), resulting in segmental arterial narrowing with attendant limb arteritis, coronary artery disease, digestive angina, and cerebrovascular disease. Intermittent claudication of the lower limbs and fatigue of the upper limbs are the most common cardiovascular findings. Absence of peripheral pulses is also common and signals, along with precocious coronary artery disease, the need for a workup for PXE in younger patients. In terms of mucosal bleeding, gastrointestinal (GI) tract involvement is the most common complication. The exact mechanism of GI bleeding is unknown but may occur as a result of arterial involvement (Chassaing et al., 2005).

The histology of PXE is not pathognomonic, as it is indistinguishable from calciphylaxis, saltpeter intoxication, penicillamine intoxication, and calcific elastosis without perforations. Nevertheless, PXE skin lesions are characterized by an elastin anomaly in the mid-epidermis with relatively normal morphology in the papillary and deep dermal layers. Alcian blue and colloidal iron staining demonstrates an elastin band that undergoes swelling followed by granular degeneration, fragmentation, splitting, and curling of elastin fibers. von Kossa staining reveals prominent calcium depositions in the abnormal granular elastin matrix. Electron microscopy has demonstrated that the earliest histological sign of PXE is calcification of the elastic fibers, composed of hydroxyapatite and/or calcium phosphate. As mineralization accumulates, the characteristic fragmentation of elastic fibers occurs (Hu et al., 2003).

Around 2000, the gene responsible for most autosomal dominant and recessive forms of PXE, ABCC6, was mapped to 850 kb on chromosome 16p13.1. ABCC6 is in subfamily C of the ATP-binding cassette protein superfamily. Members of this subfamily are involved in signal transduction, drug and antibiotic resistance, and antigen presentation. ABCC6 is composed of 31 exons with a 4.5 kb open reading frame encoding the 1,503 amino acid protein, multidrug-resistance protein 6 (MRP6). There is relatively high expression of ABCC6 mRNA in liver and kidney cells, whereas, ironically, there is much lower expression in the affected tissues of PXE (see below). To date, approximately 50 mutations—particularly in exons 24 and 28–30 of ABCC6I —are reported to result in the PXE phenotype. The relationship between the molecules transported by MRP6 and PXE has yet to be elucidated, as various studies have demonstrated that MRP6 transports proteins ranging from glutathione conjugates to endothelin receptor antagonists (Hu et al., 2003).

2. Depending on the techniques applied, the knockout technique yields mice that completely lack a specified gene product, either in target tissues or in all cells. This technique is useful for studying the effects of losing gene expression in the host species under investigation. The generation of an ABCC6 knockout was described in detail by Kelment et al. (2005) and is summarized here. The development of this model was essential to elucidate the role of ABCC6 in the pathophysiology of PXE.

To create an ABCC6 knockout mouse, the investigators designed a targeting vector consisting of DNA with homology to portions of exons 15 through 18 of wild-type (WT) ABCC6, a neomycin/G418 resistance gene (NEO) with transcriptional orientation opposite to that of the ABCC6 genomic fragments, a herpes simplex virus-thymidine kinase (HSV-TK) gene, and a novel NotI restriction enzyme site to linearize the construct and allow for PCR analysis. Following standard techniques of targeting vector design, the homologous arms coinciding with DNA from exons 15 through 18 was used to ensure homologous recombination with the WT locus in the host genome. The homologous arms flank a neomycin resistance cassette that serves a dual purpose. First, when it integrates into the WT locus, it renders the coding sequence functionless. Second, it confers resistance to the antibiotic neomycin, which serves as a means of selecting out only the cells that have been successfully transfected. The thymidine kinase gene lies outside the homologous arms so that it will not be incorporated into the host genome during homologous recombination. This offers a means of negative selection whereby cells that randomly integrate this gene will incorporate a toxic nucleotide analogue into their DNA when the appropriate drug is added to the culture medium. The thymidine kinase gene thus helps to select against cells that have not engaged in homologous recombination at the desired host genome locus. The construct was then mass-produced using an Escherichia coli vector, and the resultant genetic material was used to transfect mice with a C57/J129 genetic background. Through homologous recombination, the targeting vector selectively entered the mouse genome at one of the ABCC6 loci and rendered it functionless by replacing WT DNA with the nonfunctional targeting construct.

Embryonic stem cells with one “knocked out” ABCC6 locus were then microinjected into blastocysts that were in turn implanted into pseudopregnant recipients. The resulting chimera, which contained one WT and one “knocked out” ABCC6 locus, were mated with C57BL/6J females. Heterozygotes containing one WT and one knockout gene were identified and mated successively until homozygous ABCC6 knockout mice could be identified (ABCC6 ^–/–).

These “knockout” mice were used to conduct the experiments carried out by Jiang et al. (2009) in the current study. As PXE is thought to result from mutations that render the ABCC6 gene product diminished or functionless, it follows that much could be learned about this condition by studying a model completely lacking expression of this gene product. Previous studies have demonstrated that this knockout mouse develops phenotypic and histologic features similar to those of patients with PXE (Jiang et al., 2007).

3. Once the knockout mice were created, the investigators sought to determine whether there was evidence for the proposed metabolic nature of PXE. Previous work has shown that the ectopic mineralization characteristic of PXE can be qualitatively measured by examining the histology of the connective tissue capsule of murine vibrissae. In particular, it has been observed that in ABCC6 ^–/– (knockout) mice—as demonstrated by hematoxylin/eosin, Alizarin Red, and von Kossa staining of serial chronologic sections—there is progressive mineralization of the connective-tissue capsule. This mineralization is not appreciable at 4 weeks but becomes increasingly apparent at 3 and 6 months. Furthermore, there is no appreciable mineralization of the capsule in ABCC6 ^+/+ (wild-type) mice at these time points.

Two major experimental designs were employed in this study: one focusing on potential prevention of ectopic mineralization, the other focusing on potential reversal of ectopic mineralization after it had occurred. In the prevention arm, Jiang et al. investigated whether the placement of muzzle skin from either knockout or wild-type mice onto the backs of mice with opposite ABCC6 genotypes would result in ectopic mineralization of the connective-tissue capsule of vibrissae. The overall goal of this study arm was to determine whether the engraftment of (–/–) muzzle skin onto a (+/+) host would result in the prevention of ectopic mineralization. Since neither ABCC6 ^/ or ^+/+ mice showed ectopic mineralization at 4 weeks, mice of this age were used in this arm of the study. Three groups were observed: one involving (+/+) donors and (–/–) hosts, the second involving (–/–) donors and (+/+) hosts, and a control involving (+/+) muzzle skin grafted onto (+/+) hosts to determine whether the process of transplantation played a role in mineralization. Transplanted tissue was analyzed with hematoxylin/eosin, Alizarin red, and von Kossa staining, along with transmission electron microscopy and computerized morphometric quantitation. von Kossa stains carbonates, phosphates, oxalates, sulfates, urates, chloride, and other anionic salts black and is sensitive but not specific for calcium deposition. Alizarin red is useful in staining for calcium phosphate deposition as it forms an orange-red aggregate with this mineral at a pH of approximately 4.2 and is a more specific stain for calcium than von Kossa. In addition to the control mentioned above, histology and PCR analysis were used to verify that the engrafted skin maintained the genotype of the donor and that neovascularization originated from the host tissue. Lastly, these transplant pairings were repeated in RAG1 mice to determine whether immune incompatibility played a role in the observed results . Experiments in RAG1 mice showed nearly identical results, disproving a role for immune incompatibility in the results.

In the reversal arm, Jiang et al. looked at whether the characteristic mineralization in ABCC6 ^–/– vibrissae, prominent at 12 weeks and beyond, could be reversed or undone if the affected tissue was engrafted onto an ABCC6 ^+/+ mouse. To accomplish this goal, three groups of 12-week-old mice were observed: one group of ABCC6 ^–/– donors engrafted onto ABCC6 ^+/+ recipients, another group of ABCC6 ^–/– donors engrafted onto ABCC6 ^–/– recipients, and a group of ABCC6 ^–/– not subjected to muzzle-skin transplant. As described above, the mineralization of the connective-tissue capsule was qualitatively and quantitatively observed with hematoxylin/eosin, Alizarin red, and von Kossa staining fortified with transmission electron microscopy and computerized morphometric quantitation.

4. In the prevention arm, at 8 weeks (i.e., 4 weeks after the initial transplant), biopsy specimens from the three groups were analyzed. A survey of vibrissae from the three groups was conducted. In the ABCC6 ^–/– mice who received (+/+) grafts, 16 of 56 vibrissae (28.6%) showed evidence of mineralization. Of the (+/+) mice who received (–/–) grafts, none showed evidence of capsule mineralization. Lastly, the control group of (+/+) mice with (+/+) grafts showed no capsule mineralization. Since (+/+) vibrissae grafted onto the skin of (–/–) mice showed evidence of calcification, the authors suggest that the local environment (as opposed to the cells themselves) at least contributes to this process. Conversely, and perhaps even more striking, the finding that (–/–) vibrissae engrafted onto a (+/+) background fail to exhibit mineralization suggests that even genetically deficient cells need the correct environment to develop disease and lends hope for an effective therapy that could prevent the ectopic mineralization characteristic of PXE.

In the reversal arm, at 6 months (i.e., 3 months after the initial transplant), biopsy specimens of the three groups were analyzed. Muzzle skin from 12-week-old ABCC6 ^–/– mice—i.e., mice with significant capsule mineralization—was transplanted onto the backs of either (+/+) or (–/–) mice. ABCC6 ^–/– muzzle skin at 3 months, the time of the grafting procedure for the experimental groups, was examined as a control and exhibited clear evidence of capsule mineralization. Kruskal-Wallis one-way analysis of variance demonstrated that there was a statistically significant degree of mineralization among the three groups in the reversal arm of the study. Furthermore, Dunn’s multiple-comparison testing showed a statistically significantly less mineralization in (–/–) grafts transplanted onto (+/+) mice as compared with (–/–) grafts transplanted onto (–/–) mice.

Other findings that lacked statistical significance but that are of some importance include the comparisons between (–/–) → (–/–) and (–/–) → (+/+) transplants compared with (–/–) muzzle skin. Of note, there was a 2.5 fold increase in mineralization of the (–/–) → (–/–) transplanted skin when compared with unadulterated (–/–) muzzle skin and 60% less mineralization in (–/–) → (+/+) transplants when compared with unadulterated (–/–) muzzle skin. These study populations were much smaller than the numbers involved in the prevention arm, and may have lacked adequate power to reach statistical significance.

Taken together, the data produced by the reversal arm suggest that the ectopic mineralization characteristic of PXE can be—at least partially—undone. Much in line with the prevention-arm data, these findings support the concept of PXE as more than a cellular process and promote the notion that systemic factors that manifest locally can contribute to ectopic mineralization. Furthermore, the reversal-arm data suggest that, even in subjects with established PXE lesions, the introduction of a therapeutic intervention may halt or even reverse progression of the disease.

5. Important insights can be drawn from these elegant experiments. It is relatively certain that if the PXE phenotype resulted only from localized cellular defects, grafted cells lacking ABCC6 would behave similarly regardless of the host genotype. Taken together, the data outlined in this study support the theory that PXE has at least a metabolic component with attendant cutaneous, ocular, and cardiovascular findings resulting from more than simply localized defects characteristic of resident cells in affected tissues.

Several hypotheses have been proposed to explain why the lack of ABCC6 results in the PXE phenotype. The notion that systemic metabolites play a role in PXE was elaborated, for example, by Le Saux et al. (2006), who showed that PXE fibroblasts cultured with normal human serum maintained structurally normal elastic fibers and that normal and PXE fibroblasts deposited abnormal aggregates of elastic fibers when maintained in the presence of serum from PXE patients. As mentioned above, ABCC6 is expressed in much higher quantities in the liver and kidneys, and one of its functions may be the transport of glutathione-conjugated anionic molecules (Belinsky et al., 2002). It has been proposed that a vitamin K precursor is secreted by ABCC6 from the liver as a glutathione (or glucuronide) conjugate and that this supplements the vitamin K that is relatively insufficient in peripheral tissues as a result of hepatic processing. Similar to the phenotype of γ-glutamyl carboxylase mutants that exhibit PXE-like lesions, calcification of elastic fibers occurs in tissues lacking adequate vitamin K, possibly because of the insufficient γ-carboxylation of glutamate residues on proteins known to counteract such ectopic calcification (Borst et al., 2008). There are myriad other possible metabolic explanations for this ectopic calcification, however, as several other serum proteins, including matrix Gla protein, fetuin-A, and ankylosis protein, have been shown experimentally to prevent it (Jiang et al., 2007).

One last insight that this study helps bring into relief is the possible bridge between the “metabolic” and “cellular” theories of PXE pathogenesis. ABCC6 mutants do exhibit phenotypic differences in protein expression, including that of matrix γ-carboxyglutamic acid protein (MGP), a substance known to inhibit soft-tissue calcification. As suggested by Gheduzzi et al. (2007), the disparity of γ-carboxylated versus noncarboxylated MGP in pathological PXE fibroblasts compared with controls indicates that a localized-cellular role for the ABCC6 mutation cannot be completely discounted.

6. The immediate clinical impact of this paper will be modest. However, this study and several of those mentioned above have signaled the importance of viewing PXE as more than a localized, cellular disease. This opens opportunities to rethink the disease in order to identify avenues for intervention. To date, there is little therapy for PXE other than vigorous monitoring for the numerous attendant morbidities associated with the disease. If PXE proves to be a metabolic process, perhaps medications can be designed to inhibit or replace the processes or deficiencies that contribute to it.

REFERENCES
Belinsky MG, Chen ZS, Shchaveleva I, Zeng H, Kruh GD (2002) Characterization of the drug resistance and transport properties of multi-drug resistance protein 6 (MRP6, ABCC6). Cancer Res 62:6172–7

Borst P, van de Wetering K, Schlingemann R (2008) Does the absence of ABCC6 (multidrug resistance protein 6) in patients with Pseudoxanthoma elasticum prevent the liver from providing sufficient vitamin K to the periphery? Cell Cycle 7:1575–9

Chassaing N, Martin L, Calvas P, Le Bert M, Hovnanian A (2005) Pseudoxanthoma elasticum: a clinical, pathophysiological and genetic update including 11 ABCC6 mutations. J Med Genet 42:881–92

Gheduzzi D, Boraldi F, Annovi G, DeVincienzi CP, Schurgers LJ, Vermeer C et al. (2007) Matrix Gla protein is involved in elastic fiber calcification in the dermis of pseudoxanthoma elasticum patients. Lab Invest 87:998–1008

Hu X, Plomp A, van Soest S, Wijnholds J, Paulus TVM, Bergen A (2003) Pseudoxanthoma elasticum: a clinical, histopathological, and molecular update. Surv Ophthalmol 48:424–38

Jiang Q, Li Q, Uitto J (2007) Aberrant mineralization of connective tissues in a mouse model of pseudoxanthoma elasticum: systemic and local regulatory factors. J Invest Dermatol 127:1392–402

Jiang Q, Endo M, Dibra F, Wang K, Uitto J (2009) Pseudoxanthoma elasticum is a metabolic disease. J Invest Dermatol 129:348–54

Klement JF, Matsuzaki Y, Jiang Q-J, Terlizzi J, Choi HY, Fujimoto N et al. (2005) Targeted ablation of the ABCC6 gene results in ectopic mineralization of connective tissues. Mol Cell Biol 25:8299–310

Lebwohl M, Lebwohl E, Bercovitch L (2003) Prominent mental (chin) crease: a new sign of pseudoxanthoma elasticum. J Am Acad Dermatol 48:620–2

Le Saux O, Bunda S, VanWart CM, Douet V, Got L, Martin L et al. (2006) Serum factors from pseudoxanthoma elasticum patients alter elastic fiber formation in vitro. J Invest Dermatol 126:1497–505

Neldner KH, Struk B (2002) Pseudoxanthoma elasticum. In: Royce PM, Steinmann B (eds) Connective Tissue and Its Heritable Disorders: Molecular, Genetic and Medical Aspects. Wiley–Liss: New York, 561–83

Pfendner E, Uitto J, Gerard GF, Terry SF (2008) Pseudoxanthoma elasticum: genetic diagnostic markers. Expert Opin Med Diagn 2:1–17

Quaglino D, Boraldi F, Barbieri D, Croce A, Tiozzo R, Pasquali-Roncherri I (2000) Abnormal phenotype of in vitro dermal fibroblasts from patients with pseudoxanthoma elasticum (PXE). Biochim Biophys Acta 1501:51–62

¹ Department of Dermatology and Cutaneous Surgery, University of Miami Miller School of Medicine, Miami, Florida, USA