<<

Ann. N.Y. Acad. Sci. ISSN 0077-8923

ANNALS OF THE NEW YORK ACADEMY OF SCIENCES Issue: Skeletel Biology and Medicine Physiological role of explored in

1 2 Michael P. Whyte • 1 Center for Metabolic Disease and Molecular Research. Shriners Hospital for Children. St. Louis. Missouri, USA. 2Division of Bone and Mineral Diseases, Washington University School of Medicine at Barnes-Jewish Hospital, St. l.ouis, Missouri, USA Address for correspondence: Micnael P Whyte, M.D .. Shriners Hospital for Children, 2001 South Lindbergh Blvd. St. Louis. MO [email protected]

Hypophosphatasia (HPP) is the instructive or caused by loss-of-function mutation(s) within TNSALP, the gene that encodes the "tissue nonspecific" isoenzyme of alkaline phosphatase (TNSALP). HPP reveals a critical role for this enzyme in skeletal mineraliution. Increased extracellular levels of pyrido.xal 5' -phosphate and inorganic pyrophosphate (PP;) demonstrate that TNSALP is a phosphomonoester phosphohydrolase and a pyrophosphatase that hydrolyzes much lower concentrations of natural substrates than the artificial substrates of laboratory assays. Clearly, TNSALP acts at physiological pH and "alkaline phosphatase" is a misnomer. Aberrations of B6 metabolism in HPP revealed that TNSALP is an ectoenz.yme. PP; excesses cause chondrocalcinosis and sometimes . The skeletal disease is due to PP; inhibition ofhydroxyapatite crystal growth extracellularly so that crystals form within matrix vesicles but fail to enlarge after these structures rupture. Trials of alkaline phosphatase replacement therapy for HPP suggest that TNSALP functions at the level of skeletal tissues.

Keywords: ectoenzyme; inorganic pyrophosphate; osteomalacia; rickets;

Soon after Robison's report,1 his hypothesis was Introduction challenged. ALP was abundant in noncalcifying tis­ Alkaline phosphatase (ALP) was discovered in sues (e.g., liver, intestine, placenta), and extracellu­ 1923 by Robert Robison who suggested this en - lar fluid was supersaturated with P;. 3 Instead, other zyme functioned in skeletal mineralization by functions were proposed (see below).~ liberating inorganic phosphate (P;) f~r hydrox­ In the 1960s, electron microscopy rejuvenated yapatite (HA) crystal propagation (Fig. 1). 1 He Robison's hypothesis when the earliest HA crystals found considerable phosphatase activity in the were discovered within "matrix vesicles" (MVs).7 bone and cartilage of growing rats and hypothe­ MV s seem to be"buds of the plasma membrane sized that mineralization followed hydrolysis ofhex­ of and and are rich in osephosphoric esters. ln 1924, Robison and Soames ALP. During "primary" skeletal mineralization, HA demonstrated its peculiar alkaline pH optimum crystals appear and then grow within MVs. After in vitro. 1 Robison recognized, however, that this MVs rupture, "secondary" mineralization proceeds was not physiological and never referred to "alka­ as HA crystals enlarge and deposit into skeletal line phosphatase." He called his discovery "bone matrix. phosphatase." 1 By the 1990s, the postulated roles for ALP were Beginning in the 1930s, significant clinical in­ many,4~ including hydrolysis of Pi esters for the sight came from measuring ALP in serum because non -P; moiety, transferase action for synthesis of hyperphosphatas~mia usually indicates skeletal or P; esters, regulation of P; metabolism, mainte­ hepatobiliary disease. Quantitation of serum ALP nance of steady-state levels of phosphoryl metabo­ may still be the leading enzyme assay. 2 lites, and action as a phosphoprotein phosphatase.

doi 10.1111/j.17 49-6632.2010.05387.x 190 Ann. N.Y. Acad. Sci. 1192 (2010) 190-200

is, ALP hydrolyzes an inhibitor of calcification.3~-9 The principal candidate, inorganic pyrophosphate 'The SIGNIFICANCE ()f (PP;), impairs HA crystal growth and can be hy­ drolyzed by ALP.3•4 In fact (see below), plasma and PHOSPHORIC ESTERS urine levels of PPi are increased in hypophosphata­ sia (HPP).10 in METABOLISM Now, Robison's hypothesis is proven; the "tissue nonspecific" isoenzyme of ALP (TNSALP) is es­ By sential for skeletal mineralization. Verification came ROBERT ROBISON, P1t.O., O.Sc., F.R.S. ,,,1.. ..,,1....-,,v ..... ~,.-­ from the discovery and subsequent characterization •"""""""~_,...,...,,,..... ,.,_., ...., 4 __ _ of the inborn error of metabolism, HPP. -6 In 1988, loss-of-function mutation(s) in TNSALP were first documented in HPP. The pathogenesis of the de­ fective skeletal mineralization in HPP involves im­ 4 paired hydrolysis of PPi, ~ representing the second regulator ofmineralization anticipated by Robison. 1

I Ironically, nearly a century after the discovery of (t).,,. ALP, assay methods do not reflect Robison's ap­ 2 11 preciation of this enzyme. • fn both clinical and research laboratories, ALP is measured using non­ physiological alkalinity (e.g., pH 9.2-10.5) and high concentrations (millimolar) of artificial substrates (e.g., p-nitrophenylphosphate).2 Furthermore, bio­ THE NE.W YORK UNIVERSITY PRESS logic specimens are diluted into buffers without Pi, ,..,Ull111''1f()2' iQIIAltt un· • )IJ!iW YOltl: CITY although P; competitively inhibits ALP.2· 11 These M-•DOJII , ..t, ... ,. .. , .. lfl.LNU.O•o•,no 1,·IQVaUrt'lf' ,...... procedures igndre the pH optimum for ALPs being less alkaline for the lower concentrations of physio­ 2 6 logical substrates. • To understand HPP and what it teaches us about Figure 1. Robert Robison's 1932 summary of his re­ ALP, a brief review of ALP genomic structure and markable work concerning the importance of phospho­ protein chemistry is helpful. ric esters in metabolism (New York University Press, New York).1 Genomic structure and protein chemistry of alkaline phosphatase At plasma membranes, ALP perhaps conditioned transport of Pi> , fat, protein, carbohydrates, ALP (orthophosphoric-monoester phosphohydro­ and Na+ /K+. Sequence analyses suggested coupling lase, alkaline optimum, EC 3.1.3. 1) is found ubiq­ to other proteins, including collagen (see below). In uitously in plants and animals.2 In man, four the placenta, ALP bound the Fe receptor of [gG and ALP isoenzymes are encoded by four separate perhaps transcytosed this immunoglobulin. genes.5 Three are expressed in essentially a tissue­ Additional roles also emerged for ALP in skele­ specific distribution and form intestinal, placen­ tal mineralization, including a plasma membrane tal, and germ cell (placental-like) ALP. The fourth transporter for Pi, an extracellular calcium (Ca2+). is in all cells and designated tissue-nonspecific binding protein that stimulates calcium-Pi precipi­ ALP (TNSALP).5 Liver, bone, and kidney are es­ tation and orients mineral deposjtion into , pecially rich in TNSALP, and the gene mapping a Caz+ /Mgz+ -ATPase, or a phosphoprotein phos­ symbol is ALPL (ALP-liver), although the function phatase that wnditions matrix for . 8 Fur­ of TNSALP in bone, nut in liver, is known (st:t: thermore, certain domains seemed to bind ALP to below). The distinctive physicochemical properties types I, II, and X collagen in cartilage and bone. Nev­ ( e.g., heat stability, electrophoretic mobility) ofliver, ertheless, an early theory gained preeminence; that bone, and kidney ALP are lost with glycosidase

A1111. N.Y. AGa<.I St:i, 1192 (2010) 190 200 © 2010 Ntow Yv1k AGa<.l~111y of St:it!IIGl:!S. 191 Hypophosphatasia and alkaline phosphatase Whyte exposure, reflecting "secondary" isoem:ymes (iso­ phosphorylation-dephosphorylation of a serine forms) having an identical polypeptide sequence but residue. Dissociation of the covalently linked P; different posttranslationa( modifications. 5 seems to be the rate-limiting step. In fact, Pi is a 2 11 The TNSALP gene is at the tip of the short arm of potent competitive inhibitor of ALP. • However, chromosome 1 (lp36.l-p34); the other ALP genes P; may also stabilize the enzyme. 16 Catalytic activity are found at the end of the long arm ofchromosome requires Mg2+ as a cofactor.2 2 (2q34--q37).4 TNSALP seems to be the ancestral Uncertainties persist aboutALPbiosynthesis. The gene.4 human ALPs have a short signal sequence of 17-21 TNSALP exceeds 50 kb and contains 12 exons; 11 amino acid residues and a hydrophobic domain at are translated to form the 507-amin~ acid residue the carboxyterminus. 12 They link to plasma mem­ TNSALP. 12 TATA and Spl sequences may be regula­ brane surfaces tethered to the polar head group of tory elements, but basal expression seems to reflect a phosphatidylinositol-glycan moiety18 and can be "housekeeping» promoter effects, whereas differen­ released by phosphatidylinositol-specific phospho­ tial expression in tissues may use a posttranscrip­ lipase.14 However, their precise interactions with tional mechanism.s TNSALP has two promoters phosphatidylinositol may differ. Intracellular degra­ and two corresponding 51 noncoding exons, la and dation can involve proteosomes. 19 lb, resulting in two different mRNAs. Transcription Lipid-free ALP is found in the circulation. Yet the from the upstream promoter (la) is used in os­ mechanism for its release from cell surfaces is un­ teoblasts, whereas the downstream promoter (lb) known. The process could involve a C or D type is used in liver and kidney. 4 The tissue-specific ALP phosphatidase, detergent action, proteolysis, mem­ genes are smaller than TNSALP, primarily because brane fractionation, or lipolysis. of shorter introns. Amino add sequences deduced In healthy adults, nearly all serum ALP consists from their cDNAs suggest 87% positional identity of equal amounts of the bone and liver isoforms 4 5 between placental and intestinal ALP but only 50- of TNSALP. • Infants and children, and particu­ 60% identity with TNSALP. 5 larly newborns and adolescents, have. higher levels The cDNA sequence of TNSALP reveals of the bone isoform.2 Some people have intestinal five potential N-linked glycosylation sites. 12 N­ ALP in the circulation after ingesting a fatty meal, glycosylation is necessary for catalytic activity. 0- but usually this is just a small amount of ALP (max­ glycosylation characterizes the bone, but not liver, imum 20%).2 Placental ALP is typically expressed isoform.s only during the last trimester of .2 Clear­ In :2000, the crystal structure for placental ALP ance of circulating ALP probably involves the liver. was detailed. u The deduced active site of TNSALP is encoded by six exons comprised of 15 amino acid residues and features a nucleotide sequence con­ Hypophosphatasla served jn ALPs throughout nature. Subnormal extracellular concentrations of calcium 2 2 The ALPs are Zn + -metalloenzymes. Catalysis and P;, or P; alone, cause nearly all types of rickets or requires a multimeric configuration of identical osteomalacia. 20 HPP is an exception where the cir­ subunits with each monomer having one active culating levels are usually normal or elevated.4 With site and two zn2+ atoms that stabilize the tertiary the identification, beginning in 1988, of mutations structure. within TNSALP causing HPP, Robison's hypothesis In tissues, ALPs are tethered to cell surfaces, prob­ was confirmed. 1 Additionally, TNSALP was shown 14 ably as homotetramers. In the circulation, ALPs to be necessary for development of primary teeth. 2 are homodimeric. TNSALP in dimeric form has However, undisturbed organs/tissues in HPP, par­ etf3 topology for each subunit with a IO-stranded ticularly liver, led to the questioning of the signifi­ 5 13-sheet at its center. cance for TNSALP elsewhere.4•6 ALPs exhibit broad substrate specificities and pH optima in vitro. depending on the type and concentration of phosphocompound under­ History going catalysis.2 PPi as well as phosphoesters John C. Rathbun coined "hypophosphatasia" in 15 17 can be hydrolyzed. • The reaction involves 1948 when he reported an infant boy who died from

192 Ann. N.Y. Acad. Sci. 1192 (2010) 190-200 © 2010 New York Academy of Sciences. Whyte Hypophosphatasia and alkaline phosphatase

rickets and and whose ALP in serum, bone, Perinatal hypophosphatasia and other tissues was paradoxically subnormal.2 1 This most severe form of HPP is obvious in a Understanding the physiological role ofTNSALP neonate, manifests in utero with profound skele­ was advanced by the discoveries of elevated levels of tal hypomineralization, and typically causes death three phosphocompounds in patients. Jn 1955, ex­ at, or soon after, birth. cessive phosphoethanolamine (PEA) in urine pro­ Skeletal radiographs readily distinguish perinatal vided a second biochemical marker for HPP. In HPP from the most severe forms of osteogenesis 1965 and 1971, high levels of PPi in the urine and imperfecta or congenital dwarfism; the findings are plasma, 10 respectively, revealed the mechanism for diagnostic. In some stillborns, are nearly de­ 4 6 22 rickets or osteomalacia (see below). • • In 1985, void of mineral; others have severe rachitic changes. elevated plasma concentrations of pyridoxal 5/ - In the skull, membranous bones may calcify only phosphate (PLP), together with an understanding centrally, giving the illusion that cranial sutures are of vitamin B6 (B6) metabolism, disclosed an ec­ widely separated, although they may be functionally toenzyme function for TNSALP and explained how closed. the phosphocompounds accumulate extracellularly (see below). 17 Infantile hypophosphatasia Skeletal disease manifests after birth but before 6 Clinical features months of age4 when there is . HPP seems to affect all races. Then, rickets appears, leading to the diagnosis. B - Despite high levels ofTNSALP in bone, cartilage, 6 responsive epilepsy predicts a lethal outcome. 24 Rib liver, kidney, and adrenal tissue (and at least some fractures and chest deformity often lead to pneumo­ ubiquitously) in healthy individuals,2 HPP appears nia. Acquired hypercalcemia is common and may to disrupt only "hard tissues" directly.4 explain recurrences of vomiting as well as nephro­ Perhaps the most remarkable feature ofHPP is its cakinosis with renal compromise.4 extraordinarily wide-ranging expressivity, spanning The striking radiographic features are diagnostic. death in utero with nearly an unmineralized skele­ Cranial sutures may appear wide open but are the ton to difficulties with adult teeth without skeletal illusion from skull hypomineralization, and there disease.4 can be "functional" . Serial radio­ Many (approximately 200) mutations in graphs may reveal not only rickets but also gradual TNSALP, transmitted in various combinations by demineralization with fractures,24 heralding a lethal autosomal recessive inheritance or alone with dom­ outcome.25 Alternatively, there can be spontaneous inant/negative effects over several generations by unexplained improvement. autosomal dominant inheritance, occur in HPP.23 FmthnmnrP, othe-r e,:-netk and nongenetk factors condition HPP expressivity, as illustrated by variable Childhood hypophosphatasia 4 26 severity among siblings sharing TNSALP defects. This form has variable expressivity. • The diag­ Accordingly, the prevailing dassification scheme for nosis is made when is discovered af­ patients4 remains a clinical one from 1957. ter age 6 months. Premature loss of deciduous Five principal forms ofHPP are usually discussed. teeth (i.e., <5 years of age) is a major feature The age at diagnosis because of skeletal disease dis­ and results from hypoplasia of cementum.27 Con­ tinguishes the perinatal, infantile, childhood, and sequently, teeth slide out painlessly. Delayed walk­ adult types.4 Individuals without skeletal findings ing and waddling reflect the degree of skeletal dis­ but dental features only are said to have ''odonto­ ease. Patients can complain ofstiffness and and HPP."4 However, this nosology does not unambigu­ have appendicular muscle weakness consistent with ously classify all patients, and HPP is a wide-ranging a nonprogressive . 4 6 continuum. • Radiographs usually show characteristic The prognoses for these five major forms ofHPP "tongues" of lucency thot project from rachitic are determined by the skeletal complications. Typi­ growth plates into irregular and widened meta­ cally, the earlier the , the worse physes; these tongues are occasionally mistaken for the outcome.4 infection or leukemia (Fig. 2). Premature fusion

A11n. N.Y. Acad. Sci. 1192 (2010) 19G-200© 2010 New York Academyot Scienceg. 1&3 Hypophosphatasia and alkaline phosphatase Whyte

the cause is not understood. Calcium PPi dihy­ drate (CPPD) deposition, occasionally with attacks of pseudogout, is from increased PPi (see below). Some patients suffer degeneration of articular car­ tilage and "PP; arthropathy" and ossified ligaments resembling spinal .4

Odontohypophosphatasia This form is diagnosed when there are den­ tal manifestations but no evidence of rickets or osteomalacia.

Biochemical findings HPP can be diagnosed confidently from a consis­ tent medical history and physical findings, subnor­ mal serum ALP activity, and typical radiographic changes.4 Figure 2. Radiographic changes ofchildhood hypophos­ Hypophosphatasemia can occur in other con­ phatasia (HPP). (A) Metaphyscal irregularity, patchy os­ ditions.29 Rarely, severe teosderosis, and "tongues of radiolucency'' character­ and cleidocranial dysplasia manifest hypophos­ istic of childhood HPP are present in the right distal phatasemia due to little skeletal mass with impaired radius and ulna of this boy at age 16 months. (B) Ra­ cellular processing of bone ALP or hypo­ diolucendes in childhood HPP that are concerning for function, respectively. In general, HPP severity cor­ leukemia are seen here in a different patient. Before re­ relates inversely with age-appropriate serwn ALP ferral, he underwent extensive oncologic evaluation at activity. age 8 months because radiographs showed unusual "os­ In infantile HPP, hypercakemia and hyperphos­ teolytic" areas. Here, the metaphyses of the right distal phatemia are common.24·30 This seems largely from radius and ulna are expanded and osteopenic but with impaired calcium and P; uptake by a poorly growing ragged yet well-demarcated transition to irregularly hy­ and mineralizing skeleton. Low serum parathyroid pomincralized diaphyses. The same process, although less striking, appears at his elbow. These abnormalities were subsequently explained by HPP. of cranial sutures may raise intracranial pressure. Abnormalities may improve with growth plate closure after puberty (Fig. 3). Rarely, there is widespread bone marrow edema (Fig. 4).

Adult hypophosphatasia Presentation is typically after middle age. 4 Not in­ frequently, however, patients manifest rickets and premature tooth loss at a young age and then be­ come relatively healthy during their young adult life. Osteomalacia manifests as recurrent, poorly­ Figure 3. Spontaneous improvement in childhood HPP healing, metatarsal stress fractures. 28 With more ad­ at maturity. (A) The left knee ofa patient with childhood vanced disease, aching and tenderness in the thighs HPP shows considerable "popcorn calcification'' with is explained by femoral pseudofractures that will metaphyseal sclerosis as well as irregularity of his growth not heal unless they completely fracture. Early loss plates at age 17 years. (B) The abnormalities improved of secondary dentition is not uncommon, although spontaneously by age 20 years.

194 Ann. N.Y. Acad. Sci. 1192 (2010) 190- 200 © 2010 New York Academy of Sciences. Whyte Hypophosphatasia and alkaline phosphatase

HPP patients but can be unremarkable in mildly af­ fected subjects.9

Histopathological findings Histopathological disturbances in HPP seem to in­ volve only hard tissues directly. Except in odonto-HPP, bone reveals defective mineralization26 where ALP activity inversely re­ flects the degree ofosteoid accumulation. 26 Features of secondary are typically ab­ sent. 26·31 In growth plates and bone, the ceJlular sources of TNSA LP ( chondrocytes and osteoblasts) are present31 but TNSALP activity is deficient.26 Electron microscopy of perinatal HPP has re­ vealed a normal distribution of M Vs containing HA crystals, although they lack ALP.26 HA crystals fail to enlarge after MVs rupture.31 Accordingly, sec­ Figure 4. Bone marrow edema in infantile HPP. In ondary, but not primary, mineralization is compro­ two survivors of infantile HPP, marked bone marrow mised. edema (BME) affects the skeleton. (A) T2-weighted, fat­ Det1tition. In HPP, a paucity ofcementum despite suppressed magnetic resonance imaging of this 3-year­ cementoblasts explains the premature tooth loss.27 old boy shows BME in especially the proximal and distal Dentinogenesis also seems impaired, as revealed femurs. There is also evidence of BME in the ischium. by big pulp chambers. The excessive predentin Distal femoral physeaJ distortion is consistent with se­ width, increased interglobular dentin, and impaired vere HPP. (B) This 13-year-old girl has marked BME in cementum calcification parallel osteoidosis in bone. the left femur where a Looscr's zone is noted proximally Enamel is not impacted directly. (arrow).

Biochemical and genetic defects hormone (PTH) levels follow. In childhood HPP, TNSALP deficiency patients can have . Serum 25- Postmortem studies of lethal HPP revealed the hydroxyvitamin D, 1,25-dihydrol.."YVitam D, and TNSALP deficiency that defu1es this inborn error of PTH are typically normal unless affected by hyper­ metabolism. Profoundly Low ALP levels are in liver, cakemia or ren.11 compromise.26 hnnl', .an

Ann. N.Y. Acad. Sci. 1192 (2010) 190-200 © 2010 New York Acad0my of Sciences. 1115 Hypophosphatasia and alkaline phosphatase Whyte

in HPP patients worldwide (http://www.sesep. a bone-targeted human TNSALP from birth had

uvsq.fr/Database.html). Currently, approximately no skeletal and dental disease or B6 -responsive 200 TNSALP mutations are listed of which approx­ epilepsy. 35 imately 80% are missense. 32 Because HPP patients are often hyperphos­ phatemic,4 restriction and/or pharmacologic bind­ TNSALP structural defects ing of dietary Pi could theoretically be beneficial. Pi Many TNSALP mutations causing HPP change an competitively inhibits TNSALP2· 11 and suppresses amino add conserved in mammalian TNSALPs,23 TNSALP gene expression. including several in bacteria. Some disturb the cat­ alytic pocket or structurally important metal ligand binding sites; others probably compromise dimer Physiological role _of ALP explored in 4 23 formation. • Some impair intracellular movement hypophosphatasla 23 ofTNSALP. Discovery in 1988 that TNSALP mutation caused HPP proved Robison correct. 1 The dental manifes­ Prognosis tations showed that TNSALP is also necessary for Perinatal HPP is almost always fatal. Infantile HPP formation of primary teeth. often features clinical and radiographic deteriora­ Electron microscopy of postmortem HPP bone tion with approximately 50% of babies dying from and cartilage demonstrated that TNSALP enables 24 25 respiratory compromise. • Sometimes, consider­ skeletal mineralization. Secondary (extravesicular) able spontaneous improvement occurs, especially if not primary (vesicular) mineralization appears there is survival past infancy. Childhood HPP may compromised.31 improve when growth plates fuse. Unfortunately, re­ Defects in the cementum and dentin in HPP 28 currence of skeletal disease in adulthood is likely. teeth seem analogous to those in bone. A prominent Adult HPP causes recurrent and lingering orthope­ role for TNSALP during two critical phases of den­ dic difficulties. tal -initiation and completion­ has been proposed. In 2005, both cellular and acel­ Treatment lular cementum formation were impaired but not There is no established medical treatment, although mineralization of dentin.27 several approaches have been attempted, includ­ Although liver, kidneys, and adrenals are rich in 33 ing intravenous infusions of soluble ALP, marrow TNSALP,2 they seem to function normally in HPP 25 cell transplantation, and tcriparatide administra- ( see below), except that renal reclamation of filtered 28 . tion. Cortisone given to a few pediatric patients Pi is enhanced. This phenomenon does not seem to report~dly preceded normalization of serum ALP be always explained by suppressed circulating PTH activity and radiographic improvement, but not levels. consistently. Brief supplementation with Mg2+ It has been suggested that TNSALP deficiency or zn2+ has been unsuccessful. Bisphosphonates might impair the biosynthesis of phospholipids (derivatives of PPi) could be ineffective or pose fur­ causing pulmonary atelectasis in HPP; however, the 4 ther problems. respiratory problems are probably from rib cage Following documentation that plasma PLP and deformities, perhaps accounting for hypoplastic urine PEA and PP; fell as placental ALP corrected lungs. the hypophosphatasemia of pregnant carriers of HPP,34 i.v. infusions of purified placental ALP were used to correct hypophosphatasemia in a severely TNSALP substrates affected infant, but there was no clinical or radio­ Discoveries that PEA, PPi, and PLP accumulate en­ graphic improvement. This discouraging observa­ dogenously in HPP were essential for elucidating tion suggested even greater tissue requirements for the physiological role of TNSALP. Each phospho­ ALP, or perhaps ALP must be within the skele­ compound was inferred to be a natural substrate. ton and bound to plasma membranes for thera­ A preliminary study using 31 P magnetic resonance peutic efficacy.34 Now, recombinant TNSALP is be­ spectroscopy of HPP urine suggests at least several ing evaluated. In 2008, TNSALP-null mice given additional substrates.

196 Aon. N.Y. Acad. Sci. 1192 (2010) 190-200 © 2010 New York Academy of Sciences Whyte Hypophosphatasia and alkaline phosphatase

Phosphoethanolamine published observation). In TNSALP-deficient HPP In 1955, reports emerged that PEA is elevated fibroblasts, the various forms of B6 are unremark­ in HPP urine and plasma. In 1968, essentially able. Finally, autopsy tissues from perinatal HPP 1 no renal threshold was demonstrated for PEA contain normal levels of PLP and PL. K Accordingly, 17 excretion. TNSALP seemed to function as an ectoenzyme, 14· Although its origin is uncertain, PEA is not con­ as confirmed by studies of membrane attachment sidered a derivative of phosphatidylethanolamine, of porcine kidney ALP.36 that is, not from plasma membrane breakdown. Because TNSALP dephosphorylates PLP to PL The m.ijor source of PEA is reported to be liver, cxtrnceUularly, plasma levels of PL could be low in which metabolizes PEA to ammonia, acetaldehyde, HPP. However, only patients with extremely severe and Pi. Indeed, in one family :with adult HPP, urine H PP have low plasma PL concentrations---other PEA correlated inversely with serum liver (but not forms of HPP show normal or sometimes elevated bone) TNSALP. Now, PEA is understood to be part Levels. 17 of the phosphatidylinositol-glycan linkage appara­ B6 deficiency can cause renal lithiasis and epilepsy. tus36 and could be a degradation product from this Nephrocakinosis in infantile HPP is likely due to tether for cell-surface proteins (see below). hypercalciuria, but oxalate excess (a consequence of

B6 deficiency) has not been searched for in HPP. Pyridoxal 5' -phosphate Epilepsy in severe HPP has been associated with Discovery in 1985 that plasma PLP is elevated in cranial deformity, intracranial hemorrhage, and pe­ HPP advanced our understanding of ALP. 17 The riodic apnea; however, there may be another ex­ dietary forms of B6 are all converted to PLP in the planation. PEA caused seizures when given i.v. to liver. Organ ablation studies demonstrated liver as a severely affected infant during a study of PEA 4 6 the principal source of plasma PLP, where >95% of metabolism. • In two patients with perinatal HPP, PLP is coupled to albumin. Only a small amount epilepsy, and plasma PL levels below assay sen­ circulates freely. sitivity, pyridoxine did not correct seizures (per· Like other phosphocompounds, PLP cannot sonal observation). B6-responsive seizures herald cross plasma membranes but must be dephospho­ lethal HPP.24 In fact, tnsalp knockout mice manifest rylated to pyridoxal (PL). [n cells, PL is rephos­ epilepsy that requires pyridoxine administration to phorylated to PLP, which cofactors many enzymatic extend their lives (see below).38 reactions. Ultimately, PLP is degraded primarily in Observations concerning B6 metabolism in HPP the liver and excreted in urine. indicated an ectoenzyme role for TNSALP. 17 In In disorders that elevate serum ALP, plasma PLP 1980, discovery that porcine kidney ALP is bound isdecreased.37 Discovery of increased plasma PLP in to plasma membranes by phosphatidylinositol dis­ HPP disclosed the reciprocal relationship between closed a mechanism for attachment of TNSALP.36 PLP and TNSALP and the control of B6 metabolism In l 98~, cultivated osteosarcoma cells, and then der­ by ALP. 17 However, such effects may not be physi­ mal fibroblasts from patients with infantile HPP, ex­ ologically important because TNSALP controls ex­ posed to PLP and PEA confirmed this function (sec tracellular, not intraceJlular, PLP. Elevated plasma below).14 PLP in HPP reflect.~ a failure of extracellular hy­ drolysis of PLP. 17 Investigation of HPP suggested a Inorganic pyrophosphate generally inconsequential relationship because HPP Discovery in 1965 and 1971 that PPi levels are in­ 10 patients do not have symptoms ofB6 toxicity, such as creased in HPP urine and plasma explained the peripheral neuropathy. Similarly, there are no signs defective skeletal mineralization.9 PP; was known or symptoms of deficiency, such as stomatitis, der­ to potently inhibit mineralization22 by preventing matitis, peripheral neuritis, anemia, or depression, the growth of HA crystals. 9 in all but severely affected HPP babies with B6- Generation of PPi extracellularly, presumably 17 responsive seizures. Urine levels of degraded B6 from ATP, is not hindered in HPP because nucleo­ are unremarkable in childhood HPP, 18 where pa­ side triphosphate PP;-ase (NTP-PP;-ase, PC-1) ac­ tients respond normally to L-tryptophan loading, tivity is unremarkable in patient fibroblasts. Clear­ 32 a test for B6 deficiency (Whyte and Coburn, un- ance of PP; administered intravenously to adults

Ann. N.Y. Acad. Sci. 1192 (2010) 190-200

with HPP was markedly delayed (R.G.G. Russell, The ectophosphatase activity of TNSALP personal communication). hydrolyzed extracellular PEA and PLP under physi­ Consonant with in vitro effects of PPh minor ex­ ological conditions.39 Although some reports indi­ cesses of PP; may explain the precipitation of amor­ cated that ALP conditions cell growth and differen­ phous calcium phosphate and cakific periarthri­ tiation perhaps by influencing the phosphorylation tis of some adults with HPP. Furthermore, ALP of nucleotide pools, HPP fibroblasts proliferate nor­ can dissolve CPPD in vitro. •s This PP;-ase activ­ mally. TNSALP did not seem to be a phosphoprotein ity seems unrelated to hydrolysis of phosphoesters. phosphatase acting at the plasma membrane. I. 19 Thus, CPPD deposition leading to chondrocalci­ Phospholipid composition and rates of 32 P; accu­ nosis, pseudogout, and PP; arthropathy could be mulation were normal. due to failed hydrolysis of PPi. TNSALP knockout mice Circulating TNSALP Since 1995, tnsalp knockout mice have confirmed Several observations suggest that circulating (sol­ insight from HPP patients.4 The tnsalp knock­ uble) ALP is physiologically inactive. Infants with out mouse manifests the deranged B6 metabolism severe HPP who received plasma from patients of HPP, causing lethal seizures from deficient 'Y - with Paget bone disease or purified placental ALP aminobutyric acid in the brain. With pyridoxine demonstrated no significant clinical or radiographic treatment, the epilepsy is controlled and the an­ improvement despite ALP sometimes increased to imals survive long enough to develop skeletal and above normal levels. Such therapy failed to normal­ dental disease. Two models recapitulate the infantile 38 ize urinary PEA or PPi or plasma PLP. form of HPP remarkably well. In 2000, the min­ Deficiency of TNSALP within the HPP skele­ eralization defect was reproduced with osteoblasts ton seems to account for rickets and osteomala­ in culture. Defects in secondary skeletal mineraliza­ cia. In 1955, rachitic rat cartilage would calcify in tion were also confirmed by electron microscopy. HPP serum, yet HPP costochondral tissue would Subsequently, double knockout mice showed that not mineralize in a synthetic calcifying medium skeletal formation is essentially normal when both or in serum from healthy children.' Subsequently, tnsalp and the PPi-generating enzyme PC-1 (NTP­ transfection of ALP cDNA conferred both catalysis PP;-ase) are lacking, adding further support to ex­ against Pi esters and mineralization in a calcification perience with HPP patients. 5 Furthermore, tnsalp system. 16 There has been skepticism that such exper­ expression under control of the apo-e promoter in iments reflect biomineralization because increased the liver of tnsalp knockout mice improved skele­ 40 P; in metastable solutions precipitates calcium Pi, tal disease. It is uncertain if tnsalp liberated in Accordingly, it is probably important to augment especially high amounts into the circulation and ALP activity at the skeletal level to treat HPP. Mar­ reaching bone itself explains the beneficial effects. row cell transplantation for infantile HPP preceded clinical and radiographic improvement despite es­ Conflict of interest sentially unaltered biochemical abnormalities, sug­ gesting benefit from even small increments of ALP The author declares no conflicts of interest. activity within the skeleton. Transient transfection studies ofvarious TNSALPmutations also suggested References that small differences in ALP activity could account I. Robison, R. 1932. The Significance ofPhosphoric Esters in for lethal versus nonlethal outcomes.23 Metabolism. New York University Press. New York. 2. McComb, R.B., G. N. Bowers, Jr. & S. Posen. 1979. Alka­ Hypophosphatasia fibroblast studies line Phosphatase. Plenum. New York. Most. of our insights concerning ALP came from 3. Neuman, W. F. & M.W. Neuman. 1957. Emerging con­ investigations of patients with HPP. Before tnsalp cepts of the structure and metabolic functions of bone. knockout mice (see below), dermal fibroblasts in Am./. Med. 22: 123-131. culture also proved useful. 39 Such cells can be pro­ 4. Whyte, M.P. 2001. Hypophosphatasia. In The Metabolic foundly deficient ( <5% control) in ALP activity. and Molecular Bases offoherired Disease, 8th edn. Scriver,

198 Ann. N.Y Acad. Sci. 1192 (2010) 190-200 © 2010 New York Academy of Sciences. Whyte Hypophosphatasia and alkaline phosphatase

C.R., A.L. Beaudet, W.S. Sly & D. Valle, Eds.: 5313-5329. 17. Whyte, M.P., J.D. Mahuren, L.A. Vrabel & S.P. Coburn. McGraw-Hill. New York. 1985. Markedly increased circulating pyridoxal-5' - 5. Millan, J.L. 2006. Mammalian Alkaline Phosphatases: phosphate levels in hypophosphatasia: Alkaline phos­ From Biology to Applications in Medicine und Biotech­ phatase acts in vitamin B6 metabolism. J. Clin. Invest. nology. Wiley-VCH. Wdnheim, Germany. 76: 752-756. 6. Whyte, M.P. 2008. Hypophosphatasia: Nature's window 18. Whyte, M.1'., J.D. Mahuren, K.N. fedde, et al. 19118.

on alkaline phosphatase function in humans. In Princi­ Perinatal hypophosphatasia: Tissue levels of vitamin B6 ples ofBone Biology, 3rd edn. Bilezikian, J.P., L.G. Raisz & are unremarkable despite markedly increased circulat­ T.J. Martin, Eds.: 1573-1598. Academic Press. San Diego, ing concentrations of pyridoxal-5' -phosphate. Evidence California. for an ecloenzyme role for tissue-nonspecific alkaline 7. Anderson, H.C. 1969. Vesicles associated with calcifica­ phosphatase./. Clin. Invest. 81: 1234-1239. tion in the matrix ofepiphyseal cartilage.}. Cell. lJiol. 41: 19. Cai, G., 'J: Michigami, ·1: Yamamoto, et al. 1998. 59-72. Analysis of localization of mutated ti.m,c-nonspecific 8. Wuthier, R,E. & T.C. Register. 1985. Role of alkaline alkaline phosphatase proteins associated with neona­ phosphatase. a polyfunctional enzyme in mineralizing tal hypophosphalasia using green fluores,enl protein tissues. In The Chemistry and Biology ofMineralized 'fis­ chimeras./. Clin. Endocrinol. Metab. 83: 3936-3942. sues. W.T. Butler, Ed.: 113-124. EBSCO Media. Birming­ 20. Whyte, M.P. 2002. Rickets and osteomalacia (acquired ham. and heritable forms). In The Oxford Textbook of En­ 9. C.aswell, A.M., M.P. Whyte & R.G. RusseU. 1991. Hy­ docrinology and Diabetes. Wass, J.A.H. & S.M. Shalet, pophosphatasia and the extracellular metabolism of in­ Eds.: 697-715. Oxford University Press. New York, organic pyrophosphate: Clinical and laboratory aspects. NY. CRC Crit. Rev. Clin. Lab. Sci. 28: l 75-232. 21. Rathbun, J.C. 1948. Hypophosphata'sia, a new develop­ IO. Russell, R.G., S. Bisaz, A. Donath, et al. I 971. Inorganic mental anomaly. Am. J. Dis. Child. 75: 822-831. pyrophosphate in plasma in normal persons and in pa­ 22. Heinonen, J.K. 2001. Biological Role of Inorganic Py­ tients with hypophosphatasia, osteogenesis imperfecta, rophosphate. Kluwer Academic Publi.~hers. Norwell, MA. and other disorders ofbone. /. Clin. Tnvest. 50: 961-969. 23. Mornet, E., A. Taillandier, S. Pcyramaure, et al. 1998. 11. Coburn, S.P., J.D. Mahuren. M. Jain, et al. 1998. Alkaline Identification of fifteen novel mutations in the tissue­ phosphatase (EC 3.1.3.1) in serum is inhibited by phys­ nonspecific alkaline phosphatase (TNSALP) gene in Eu­ iological concentrations of inorganic phosphate.}. Clin. ropean patient-~ with severe hypophosphatasia. Eur. J. Endocrinol. Metab. 83: 3951-3957. Hum. Genet. 6: 308-845. 12. Weiss, M.J., K. Ray, P.S. Henthorn, eta/. 1988. Structure 24. Baumgartne1·-Sigl, S.B., E. Haberlandt, S. Mumm, et al. of the human liver/bone/kidney alkaline phosphatase 2007. Pyridoxine-responsive seizures as the first symp­ gene./. Biol. Chem. 263: 12002-12010. tom of infantile hypophosphatasia caused by t\vo novel 13. Le Due, H.M., T. Stigbrand, M.J. Taussig, iii al. lOOO. missense mutations (c.677T:;-C, p.M226T; c:.I 112C:;-T, Crystal structure of alkaline phosphatase from human p.T37 lI) of the tissue-nonspecific alkaline phosphatase placenta at 1.8 A resol11tion. J. Biological. Chem. 275: gene. Bone 40, 1655-1661. 9158-9165. 25. Cahill, R.A., D. Wenkert, S.A. Perlman, et al. 2007. Infan­ 14. Fedde, K.N., C.C. Lane & M.P. Whyte. 1988. Alkaline ti!e hypophosphatasia: Trial of transplantation therapy phosphatase in an ectoenr.yme that acts on micromolar using bone tragments and cultured osteoblasts. /. Clin. concentrations of natural substrates at physiologic pH Endocrinol. Metab. 92: 2923-2930. in human osteosarcoma (SAOS-2) cells. Arch. Biocl1em. 26. Fallon, M.D., S.L. Teitelbaum, R.S. Weinstein, et al. Biopl1ys. 264: 400-409. 1984. Hypophosphata_\ia: Clinicopathologic comparison 15. Xu, Y., T.F. Cruz & K.P. Pritzker. 1991. Alkaline phos­ ofthe infantile, childhood, and adult forms. Medicine 63: phatase dissol,ves calcium pyrophosphatase dihydrate 12-24. crystals. J. Rheumato/. 18: 1606-1610. 27. Van Den Bos, T., G. Handoko, A. Niehof, et al. 2005. 16. Farley, J.R. 1991. Phosphate regulates the stability of Cementum and dentin in hypophosphatasia. /. Dent. skeletal alkaline phosphatase activity in human osteosar­ Res. 84: 1021-1025. coma (SaOS-2) cells without equivalent effects on the 28. Whyte, M.P., S. Mumm & C. Deal. 2007. Adult hy­ level of skeletal alkaline phosphatase immuno-reactive pophosphatasia treated with . /. Clin. En­ protein. Calcif. Tiss_ue Int. 57: 371-378. docrinol. Metab. 92: 1203-1208.

Ann N.Y. Acad Sci. 1192 (2010) 190-200 © 2010 New York Academy of Sciences. 199 Hypophosphatasia and alkaline phosphatase Whyte

29. Macfarlane, J.D., J.H. Souverijn & F.C. Breedvdd. 1992. rophosphate, and pyridoxal 5' -phosphate (substrate ac­ Clinical significance ofa lowserum alkaline phosphatase. cumulation in carriers of hypopho~phatasia corrects Neth. J. Med. 40: 9-14. during pregnancy)./. Clin. Invest. 95: 1440,-1445. 30. Shohat, M., D.L. Rimoin, H.E. Gruber & R.S. Lachman. 35. Millan, J.L., S. Narisawa, I. Lemire, et al. 2008. Enzyme 1991. l'erinatal lethal hypophosphatasia: Clinical, radi­ replacement therapy for murine hypophosphatasia. }. ologic and morphologic findings. Pediatr. Radio/. 21: Bone Miner. Res. 23: 777-787. 421-427. 36. Low, M.G. & A.R. Saltiel. 1988. Structural and func­ 31. Anderson, H.C., H.H.T. Hsu, D.C. Morris, et al. 1997. tional roles of glycosyl-phosphatidylinositol in mem­ Matrix vesicles in osteomalacic hypophosphatasia bone branes. Science 239: 268-275. contain apatite-like mineral crystals. Am. J. Pathol. 151: 37. Anderson, B.B., H. O'Brien, G.E. Griffin & D.L. Mollin. 1555-1561. 1980. Hydrolysis of pyridoxal 5'-phosphate in plasma 32. Mornet, E. 2007. Tissue nonspecific alkaline phos­ in conditions with raised alkaline phosphate. Gut. 21: phatase gene mutations database Yvelines (France): 192-194. SESEP Laboratory at the University of Versailles­ 38. Fedde, K.N., L. Blair, J. Silverstein, et al. 1999. Alkaline Saint Quentin. c2004. (updated 2007 July 4; assessed phosphatase knock-out mice recapitulate the metabolic 2007 Dec 18). Available from: http://www.sesep.uvsq.fr/ and skeletal defects of infantile hypophosphatasia. /. Database.html. Bone Miner. Res. 14: 2015-2026. 33. Whyte, M.P., W.H. McAJister, L.S. Patton, et al. 1984. 39. Fedde, K.N., C.C. Lane & M.P. Whyte. 1990. Alkaline Enzyme replacement therapy for infantile hypophos­ phosphatase: ( tissue nonspecific isoenzyme) is a phos­ phatasia attempted by intravenous infusions of alkaline phoethanolamine and pyridaxal 5' -phosphate ectophos­ phosphatase-rich Paget plasma: Results in three addi­ phatase: Normal and hypophosphatasia fibroblast study. tional patients./. Pediair. 105: 926-933. Am.}. Hum. Genet. 47: 767-775. 34. Whyte, M.P., M. Landt, l.M. Ryan, et al. 1995. Alka­ 40. Murshed, M., D. Harmey, J.L. Millan, etal. 2005. Broadly line phosphatase: Placental and tissue-nonspecific isoen­ expressed genes accounts for the special restriction of zymes hydrolyze phosphoethanolamine, inorganic py- ECM mineralization to bone. Ge11es Dev. 19: 1093-1104.

200 Ann. N.Y. Acad. Sci. 1192 (2010) 190-200 © 2010 New York Academy of Sciences.