The role of membrane glycoprotein plasma cell antigen 1/ectonucleotide pyrophosphatase phosphodiesterase 1 in the pathogenesis of insulin resistance and related abnormalities

Abstract

Insulin resistance is a major feature of most patients with type 2 diabetes mellitus (T2D). A number of laboratories have observed that PC-1 (membrane [corrected] glycoprotein plasma cell antigen 1; also termed [corrected] ectonucleotide pyrophosphatase phosphodiesterase 1 or ENPP1) [corrected] is either overexpressed or overactive in muscle, adipose tissue, fibroblasts, and other tissues of insulin-resistant individuals, both nondiabetic and diabetic. Moreover, PC-1 (ENPP1) overexpression [corrected] in cultured cells in vitro and in transgenic mice in vivo, [corrected] impairs insulin stimulation of insulin receptor (IR) activation and downstream signaling. PC-1 binds to the connecting domain of the IR alpha-subunit that is located in residues 485-599. The connecting domain transmits insulin binding in the alpha-subunit to activation of tyrosine kinase activation in the beta-subunit. When PC-1 is overexpressed, it inhibits insulin [corrected]induced IR beta-subunit tyrosine kinase activity. In addition, a polymorphism of PC-1 (K121Q) in various ethnic populations is closely associated with insulin resistance, T2D, and cardio [corrected] and nephrovascular diseases. The product of this polymorphism has a 2- to 3-fold increased binding affinity for the IR and is more potent than the wild-type PC-1 protein (K121K) in inhibiting the IR. These data suggest therefore that PC-1 is a candidate protein that may play a role in human insulin resistance and T2D by its overexpression, its overactivity, or both.

Figure 1

Structure of PC-1. Major domains of the molecule are shown. EF-hand, Ca⁺² binding domain that maintains structure; Threonine 204, phosphodiesterase/pyrophosphatase site. The somatomedin-like domains are located near the plasma membrane and appear to be involved in interacting with the IR. The high cysteine region is involved in dimer formation.

Figure 2

PC-1 inhibition of IR β-subunit autophosphorylation in fibroblasts. A and B, Skin fibroblasts from the original patient with a 4-fold elevation of PC-1 were stimulated with either insulin (A) or IGF-I (B) for 5 min, and IR and IGF-IR autophosphorylation was measured by specific ELISAs. [Derived from Ref. .] C, 3T3 L1 mouse fibroblasts were transfected to overexpress human PC-1 4-fold. IR autophosphorylation in control 3T3 L1 and 3T3 L1 PC-1 fibroblasts was measured by specific IR ELISA.

Figure 3

Correlation between muscle and adipose PC-1 levels with in vivo insulin action. A, Correlation between PC-1 content in skeletal muscle of 28 nonobese, nondiabetic subjects and in vivo insulin sensitivity measured by the iv insulin tolerance test K_itt (51) (r = 0.51, P < 0.035) and positively correlated with plasma insulin levels. B, Correlation between PC-1 content in human adipose tissue of 19 nonobese, nondiabetic subjects and in vivo insulin sensitivity as indicated by K_itt values. PC-1 content negatively correlated with insulin sensitivity (r = −0.57, P = 0.011) and positively correlated with plasma insulin levels. [Panel A adapted from L. Frittitta et al.: Diabetologia 39:1190–1195, 1996 (51) with kind permission from Springer Science and Business Media. Panel B adapted from L. Frittitta et al.: Diabetologia 40:282–289, 1997 (62) with kind permission from Springer Science and Business Media.]

Figure 4

Fibroblast PC-1 levels. A, PC-1 content in cultured human skin fibroblasts from “normal” (healthy nonobese, nondiabetic) subjects with different insulin sensitivity (n = 12). Primary cultures of fibroblasts were established from 4-mm forearm skin biopsies. Horizontal bars indicate mean values. P = 0.01, sensitive vs. resistant subjects. B, Correlation between PC-1 content in cultured human skin fibroblasts from “normal” (healthy nonobese, nondiabetic) subjects and PC-1 content in skeletal muscle obtained from the same subjects. Over a wide range of PC-1 values, the content of PC-1 in skin fibroblasts closely reflected its content in muscle tissue (P = 0.01). [Panel A adapted from Frittitta et al.:Diabetes 47:1096–1100, 1998 (63) with permission from The American Diabetes Association. Panels A and B adapted from Goldfine et al.: Ann NY Acad Sci 892:204–222, 1999 (56).]

Figure 5

Studies in PC-1 transgenic mice. Insulin (A) and glucose (B) levels in 10 PC-1 transgenic (TG+) vs. 10 control mice (TG−). C, Glucose tolerance tests (five per group). D, Glucose metabolic index in tissues of transgenic animals overexpressing PC-1. Animals were given a constant insulin infusion, and the uptake of 2-[¹⁴C]DG was measured in various tissues (seven control and five transgenic). In transgenic animals, glucose uptake (R_g) was significantly decreased in diaphragm and soleus muscles and brain. [Adapted with permission from B. A. Maddux et al.: Am J Physiol Endocrinol Metab 290:E746–E749, 2006 (57).]

Figure 6

Genomic structure PC-1. The 5′ and 3′ UTRs are shown in dark gray. The 25 exons of the PC-1 gene are shown in light gray. The several single nucleotide polymorphisms so far reported to be associated with insulin resistance and related abnormalities are indicated.

Figure 7

Meta-analyses of the K121Q polymorphism. Meta-analyses on the association between the PC-1 K121Q polymorphism and T2D suggesting a significant diabetogenic role of this in modulating the risk of diabetes. Meta-analysis A, Ref. ; meta-analysis B, Ref. . For both, the odds ratios vs. K121K allele are calculated with a 95% confidence interval. In A, the KQ and QQ genes are shown separately. In both A and B, a significantly increased risk is given by each K121Q polymorphism. [Meta-analysis A adapted from N. Grarup et al.: Diabetologia 49:2097–2104, 2006 ( with kind permission of Springer Science and Business Media; B adapted from M. N. Weedon et al.: Diabetes 55:3175–3179, 2006 ( with permission from The American Diabetes Association.]

Figure 8

Traditional and STEM models of the IR. A, Traditional model of the IR with the major domains shown. The following regions are indicated: α-subunit domains: L1, long 1; CR, cysteine-rich; L2, long 2; α Fn 0, α fibronectin 0; α Fn 1, α Fibronectin 1; ID, insert-domain. β-subunit domains: ID, insert-domain; β Fn 1, β fibronectin 1; β Fn 2, β fibronectin 2; TM, transmembrane; TK, tyrosine-kinase; CT, C-terminal. B, Three-dimensional reconstruction (side view) from STEM. The arrow indicates the insulin-binding site at the L1 domains. The Fn 1, Fn 2, and TK domains of the β-subunit are shown. C, STEM data showing the effect of insulin binding to move the β-subunits. D, Model of proposed activation of the IR β-subunit by the CD. The CD is a connecting domain in the IR α-subunit that includes residues 485–599. In this model insulin binds to the α-subunit to activate the CDs which then move the β-subunits together to facilitate transphosphorylation. When PC-1 binds to the CD it blocks the movement of the β-subunits induced by insulin. [Derived from C. C. Yip and P. Ottensmeyer: Biol Chem 278:27329–27332, 2003 (126).]

Figure 9

PC-1 associates with the IR and blocks IR function: STEM model of PC-1 inhibition of the IR. A. Four types of cultured HTC cells transfected with the K121 of PC-1 were employed. [Derived from Ref. .] Control, Cells with low numbers of endogenous IRs; IR, cells transfected with large numbers of normal human IRs; Δ 485–599, cells transfected with large numbers of human IRs having a deletion of the IR at residues 485–599 in the α-subunit; and IGF-IR, cells transfected with large numbers of the human IGF-IR. PC-1 bound to either the IR or the IGF-IR was measured by ELISA. B, Comparison of binding of the K121 and Q121 allele of PC-1 to the IR. HTC cells, transfected with equal amounts of either the K or Q variants of PC-1, were also cotransfected with the IR. PC-1 bound to the IR was measured by ELISA. C, STEM data fitting PC-1 into the IR model to visualize its putative inhibition of IR function. Extracellular top view of the IR dimer is presented showing the calculated location of PC-1 binding to the Fn0 region interacting with the Fn2 regions. In this location, PC-1 would act to inhibit the Fn1/Fn2 domains from interacting with the associated intracellular TK domains and thus prevent them from rotating toward each other to carry out TK transphosphorylation (covered portion of double-headed arrows). The “hinge” region between Fn0 and Fn1 is indicated by a circle and labeled “tether”.