It ANKH Necessarily So

Mineralization is a complicated process controlled by local (autocrine/paracrine) as well as systemic (endocrine) factors. It can take place in a highly organized physiological fashion, as occurs in bones and teeth, or in a less organized pathophysiological fashion, as is the case in the ectopic calcification of soft tissues such as the vasculature and joints. Both of these processes are distinct from the dystrophic calcification that occurs in association with cell death, for example. The cellular and molecular mechanisms that underlie mineralization are interrelated and involve coordination and balance of processes that either promote or inhibit mineralization. Whether mineralization initiates and proceeds at a given tissue is dependent upon mineral milieu and which components of the mineralization cascade that tissue expresses. Mineralization occurs when hydroxyapatite [HA; Ca₁₀(PO₄)₆(OH)₂), or more accurately, carbonate-rich apatite] is deposited in the extracellular matrix (1,2). A simplified but illustrative model that reflects our evolving understanding of this process is shown in Fig. 1.

Figure 1.

A model of mineralization emphasizing the role of the balance of local production and degradation of the potent inhibitor of mineralization, PPi. PPi inhibits physiological and pathological mineralization by poisoning initiation events and blocking efficient propagation and growth of HA, a biological crystal composed primarily of calcium (Ca) and Pi, with liberal substitution of Ca and Pi by other ions. PPi can be generated by metabolism of extracellular ATP by the phosphodiesterase ENPP1, also known as plasma cell membrane glycoprotein-1 (PC-1), and by the transport of intracellular PPi from cells by the transmembrane-spanning cell surface protein ANKH, the human homolog of the mouse progressive Ank. PPi levels are balanced by the action of the enzyme ALP, which hydrolyzes PPi to generate Pi.

Inorganic pyrophosphate (PPi) and inorganic phosphate (Pi) play central roles in this process. PPi functions as a potent negative regulator of the mineralization process by inhibiting initiation and/or propagation of HA deposition. Pi, most of which derives from the circulation, as opposed to that generated locally by PPi degradation, is a major component of the HA crystal and serves to promote mineralization via the formation and deposition of HA in a forming or propagating mineralization front.

One process that leads to the generation of PPi is the enzymatic hydrolysis of extracellular ATP to PPi by ectonucleotide pyrophosphatase/phosphodiesterase 1 (ENPP1), also referred to as plasma cell membrane glycoprotein-1 (3), which is expressed in mineralizing tissues such as bones and teeth. Another pathway for the generation of PPi is the secretion of PPi from cells by the transmembrane-spanning cell surface protein ANKH, the human homolog of the mouse progressive ankylosis protein (Ank) (2,4). Also central to the regulation of mineralization is the metabolism of PPi to Pi by the enzyme alkaline phosphatase (ALP), sometimes referred to as “tissue nonspecific” alkaline phosphatase (TNAP), because the isoform of the enzyme that participates in the mineralization process, as opposed to other isoforms with more limited distribution, is expressed in several tissues (5,6). ALP is a primarily membrane-bound ectophosphatase that hydrolyzes PPi to liberate Pi, which can participate with calcium in the generation of HA. As such, processes that generate PPi inhibit mineralization, and those that degrade PPi generate Pi and promote mineralization.

Human diseases and their associated animal models exist for defects of most steps of this process, and their description and study have contributed greatly to our current understanding of these processes. Loss-of-function mutations in TNAP cause the disease hypophosphatasia (5), which is associated with osteomalacia due to accumulation of PPi and inhibition of HA formation. Mutations in ENPP1 cause the disease generalized arterial calcification of infancy (7) due to the inability of vascular cells to form PPi. This suggests that vascular calcification may be a default process, kept in check by active inhibition of mineralization, and that vascular cell expression of ENPP1 is essential for the maintenance of normal vascular function and integrity. The fact that vascular cells mineralize in the absence of functional ENPP1 suggests that vascular cell ANKH-mediated PPi transport is insufficient to rescue the loss of ENPP1 activity. Recently, mutations in ENPP1 have also been reported as a second cause of autosomal recessive hypophosphatemic rickets (8,9), the first being mutations in dentin matrix protein 1 (DMP1) (10). Unlike rickets caused by TNAP mutations, wherein the effect is local, in both ENPP1- and DMP1-mediated rickets the disease is caused by systemically lower serum phosphorus secondary to elevations in blood levels of the phosphate-regulating hormone fibroblast growth factor 23 (FGF23). The mechanism by which ENPP1 or DMP1 mutations lead to excessive FGF23 production remains to be elucidated.

In terms of ANKH-related diseases, both gain-of-function and loss-of-function diseases appear to exist. ANKH’s role in mineralization regulation was first suggested when it was found that biallelic mutations in the mouse ortholog, Ank, were the cause of a spontaneously occurring progressive form of arthritis accompanied by mineral deposition, formation of bony outgrowths, and joint destruction (4). It is clear in the mouse model that loss of Ank function, which occurs in only the homozygous state, leads to PPi being trapped intracellularly and a relative lack of PPi extracellularly, thereby permitting ectopic mineralization to occur. Interestingly, Ank mice lack a vascular phenotype, suggesting that there is sufficient generation of PPi in the vasculature, possibly ENPP1-mediated, to prevent ectopic calcification. In humans, two different autosomal dominant diseases resulting from mutations in ANKH have been identified, craniometaphyseal dysplasia (CMD) (11,12) and chondrocalcinosis-2 (CCAL2) (13). CMD is characterized by overgrowth and sclerosis of the craniofacial bones and abnormal modeling of the metaphyses of the tubular bones. CCAL2 is characterized by arthritis caused by deposition of calcium pyrophosphate dihydrate crystals within articular cartilage. The elevated extracellular pyrophosphates seen in CCAL2 pathological tissue suggest that the underlying mechanism is gain-of-function mutations in ANKH, resulting in PPi transport into the extracellular matrix (14). It is not clear how the mutations that cause CMD account for the observed phenotype.

In this issue of the JCEM, Morava et al. (15) report the first autosomal recessive disease of ANKH. They describe a large consanguineous family of Turkish descent with mental retardation, deafness, ankylosis, and dental abnormalities in which all affected members are homozygous for mutations at the highly conserved leucine 244 residue (L244S). The authors went to great lengths to compare the findings in this family to the Ank mouse and conclude that there are “striking similarities between the human syndrome and the recessive ank mouse phenotype.” They assume a similar loss-of-function mechanism, but how this mutation leads to disease is not clear. Not only was cell surface ANKH expression preserved in cultures of skin fibroblasts from subjects and in HeLa cells transfected with L244S mutant ANKH, but also the plasma PPi levels were normal in homozygous subjects. Unfortunately, intracellular PPi levels, which have been demonstrated to be elevated in Ank mice, were not reported in subject fibroblasts (or transfected HeLa cells). The hearing loss in this family, which is due to stapes and incus/malleus fixation and stapes ankylosis, is similar to the ankylosis of the ossicles that leads to the hearing loss seen in CMD, a disease caused by heterozygous mutations in ANKH. Although a role for ANKH in the central nervous system has been previously suggested by both prominent ANKH expression in human tissue (16) and the fact that patients with CCAL have a lower seizure threshold (17), this is the first report of mental retardation in association with ANKH. This may represent an important insight into human cognitive function. However, given the consanguineous nature of the family, it remains possible that other genes may be responsible for this, and the findings should be interpreted with some degree of caution.

The authors report low levels of 25-hydroxyvitamin D, elevated PTH relative to heterozygotes, elevated 1,25-dihydroxyvitamin D, and mild hypophosphatemia, all of which they attribute to ANKH mutations. However, all of these findings are also consistent with vitamin D deficiency, and until the mineral metabolism is studied in vitamin D-replete subjects, these data should be considered preliminary. Furthermore, given the probable role of FGF23 in the hypophosphatemia of patients with ENPP1 mutations, it will be important to report the levels of FGF23 in these subjects. Bone density was assessed in just three homozygous subjects by dual x-ray absorptiometry, and minor abnormalities were reported (spine, −0.8 ± 0.9; femur, −1.8 ± 1.2). This was compared with a histomorphometric analysis of Ank mouse bones, which showed many dramatic differences, including greater than 50% reduction in bone volume, in comparison to wild-type mice. However, neither osteoid volume nor dynamic histomorphometry, which are important given the reported findings in the subjects, was assessed, nor were complete mineral metabolism data from the mice reported.

The identification of Ank mutations as the cause of the disease in the Ank mouse was a breakthrough, which, with the inclusion of the current work by Morava et al. (15), brings the number of diseases caused by mutations in ANKH to three: CMD, CCAL2, and now autosomal recessive mental retardation with deafness, ankylosis, and dental abnormalities. Although some of the findings by Morava et al. (15) are obviously important, the effort to draw similarities between the mouse and human data has the feel of trying to squeeze a round peg into a square hole. The ANKH story (and mineralization in general) is complex, and much remains unexplained. For progress in this area to take place, more analyses of the subjects is necessary, and differences between the animal models and human diseases need to be emphasized and explored rather than forced to fit. Until this takes place, to paraphrase Sportin’ Life from George and Ira Gershwin’s opera Porgy and Bess, when he spoke about what was written in the Bible, we too must advocate a cautionary “it ANKH necessarily so,” when interpreting some of what is written.