Systematic characterization of enzyme activity on ENPP1 deficiency disease phenotype

Abstract

ENPP1 codes for a type 2 transmembrane glycoprotein in the ectonucleotide pyrophosphatase/phosphodiesterase protein family. The enzyme’s catalytic activity hydrolyzes extracellular ATP to generate pyrophosphate and adenosine monophosphate (AMP), thereby contributing to downstream purinergic signaling pathways. Pathogenic variants in ENPP1 are associated with diseases such as Type 2 Diabetes, Generalized Arterial Calcification of Infancy (GACI), Ossification of the Posterior Longitudinal Ligament (OPLL), Autosomal Recessive Hypophosphatemic Rickets Type 2 (ARHR2), Cole Disease, and early onset osteoporosis. The human phenotypes associated with these diseases are varied, and include life threatening large arterial calcifications, phosphate wasting rickets due to FGF23 elevation, cutaneous hypopigmentation and punctate keratosis, ankylosis and ossification of spinal ligaments, premature osteoporosis, as well as insulin resistance and cardiovascular disease.To determine whether the above disease phenotypes correlate with ENPP1 enzyme activity we quantitated enzyme velocities of ENPP1 variants identified in 42 ENPP1‐deficient patients followed at the NIH, together with an additional 22 variants reported in the literature. We simultaneously evaluated catalytic velocity of all variants in triplicate, enabling a direct comparison of enzyme activity across disease phenotypes. Cole disease is induced by variants in the SMB domains of ENPP1, and variants associated with the Autosomal Dominant form reduced enzyme velocity to a greater extent than the Autosomal Recessive form – between 32% to 8% of wild type (WT) vs. 33% of WT levels, respectively. Further supporting the presence of a gene dose effect, monoallelic variants in ENPP1 associated with elevation of FGF23, phosphate wasting, and early onset osteoporosis, occurred in patients with variants that reduced enzyme velocity below 50% of WT, while the polymorphisms associated with insulin resistance and cardiovascular disease reduced enzyme velocity to 73% of WT levels. Additionally, variants associated with life‐threatening vascular calcifications in GACI, which only occur in autosomal recessive biallelic ENPP1 deficient patients, reduced ENPP1 enzyme activity over a surprisingly wide range, from 25% to 100% of WT levels. Finally, disease severity in GACI correlated with the more severely damaging ENPP1 variant rather than the average of both variant allele products, suggesting that the effects of biallelic ENPP1 variants on in vivo enzyme activity are dominated by the more damaging ENPP1 allele. In summary, comparison of enzyme velocities among various ENPP1 deficient disease phenotypes supports the notion that specific variants determine these variable phenotypes, and are sometimes, but not always, related to enzyme activity. ENPP1 gene dose and dominance of the severe isoform in biallelic disease are notable mechanisms for determination of phenotype in some forms of ENPP1 deficiency.