- I. Introduction
- II. Insulin Clearance
- A. Liver
- B. Kidney
- C. Other tissues
- D. Extracellular insulin degradation
- III. Cellular Insulin Uptake
- IV. Cellular Insulin Degradation
- A. Degradation products
- B. Assay for insulin degradation
- C. Insulin-degrading enzymes
- V. Biological Role of Insulin Degradation
- VI. Insulin-IDE-Proteasome Interactions and Control of Protein Degradation
- VII. Summary and Conclusions
THIS review is the third in a series of articles on insulin degradation in this journal (1, 2). As such, it will focus on work published since the last report, but older studies will be discussed when appropriate.
Interest in insulin degradation has long been confined to a limited audience. Recent information may stimulate more widespread attention since these data closely link insulin degradation and selected actions of the hormone. Insulin action is a complex process, which is not surprising considering the multiple cellular effects of insulin and the importance of the hormone for glucose, lipid, and protein turnover and for cell growth and differentiation. Insulin removal helps control the cellular response to the hormone by decreasing availability, but the degradative process may also be involved in mediating some aspects of insulin action (3). The insulin-degrading enzyme (IDE) has multiple cellular functions in addition to degradation, including binding and regulatory functions. IDE has regulatory functions for the activity of steroid receptors and proteasomes. Intracellular interactions of insulin with IDE may be involved in insulin control of cellular protein degradation and fat oxidation. Available data support an increased importance of insulin degradation and may open a new approach to understanding some actions of insulin.
II. Insulin Clearance
Insulin uptake and degradation is a feature of all insulin-sensitive tissues (4–6).
At physiological concentrations, uptake is mediated primarily by the
insulin receptor with a smaller contribution from nonspecific processes.
At higher concentrations, nonreceptor processes assume greater
importance. Insulin has a short plasma half-life (4–6 min), as would be
expected from the necessity to respond rapidly to changes in blood
glucose (2, 7).
The modeling of insulin kinetics is a technically difficult process,
and the mathematical ramifications of whole-body systems are beyond the
scope of this review. Earlier systems were described in previous reviews
in this series (1, 2). Recently, a five-compartment model was described (8) with an excellent discussion of alternatives. The modeling system has also been discussed in another recent publication (9).A. Liver
The liver is the primary site of insulin clearance (10, 11). Approximately 50% of portal insulin is removed during first-pass transit, but this percentage varies widely under different conditions. Hepatic uptake is not a static process, but rather is influenced by both physiological and pathophysiological factors. As discussed previously in detail (1, 2), hepatic uptake is incompletely understood and involves several different systems and controls. Since most uptake is a receptor-mediated process, very high concentrations of insulin (500–2000 μU/ml) result in a decrease in the fractional uptake, although total uptake is increased (12, 13). Prolonged increases in portal insulin levels also result in reduced clearance due to receptor down-regulation. Removal of insulin from the circulation does not imply immediate destruction of the hormone (14). A significant amount of receptor-bound insulin is released from the cell and reenters the circulation. Using a five-compartment model, Hovorka et al. (8) estimated that the mean residence time of endogenously secreted insulin was 71 min with 62 min spent bound to the liver receptor, 6 min bound to peripheral receptors, and 3 min in blood or interstitial fluid. With this model, 80% of the total insulin in the body was bound to liver receptors. Other tissues (e.g., muscle) also transiently bind and can release insulin back into the circulation (15).Nutrient intake alters insulin clearance (16, 17). In general, glucose ingestion increases hepatic insulin uptake, presumably due to signals from the gut since intraportal glucose infusion does not have this effect. The glucose-induced increase in insulin secretion may also decrease hepatic fractional extraction. Under normal physiological conditions, increasing doses of glucose (10 g, 25 g, and 100 g) result in increases in insulin secretion (1.8 U, 2.7 U, and 7.2 U) with decreasing hepatic extraction (67%, 53%, 42%). Fatty acids also alter hepatic and splanchnic insulin uptake and degradation (16, 18) and may be involved in the changes associated with type 2 diabetes (19).
Clearance rates are decreased in obesity and diabetes (14, 20–24) and with increases in other hormones such as catecholamines and GH (25). In Turner’s syndrome, insulin clearance is increased before treatment but decreased to normal with GH therapy (26). Several studies have suggested that the increase in circulating insulin in obesity and type 2 diabetes is due, at least in part, to a reduced hepatic clearance, although not all studies agree (27, 28). The type of obesity, i.e., abdominal vs. lower body, is important. Syndrome X, the genetic abnormality resulting in insulin resistance, hyperinsulinemia, dyslipidemia, hypertension, and other abnormalities, may include primary alterations in insulin clearance and degradation. If so, alterations in cellular degradative function may be an integral part of this syndrome (29). Krakover et al. (30) showed that estradiol and progesterone increased insulin binding and degradation, but that testosterone decreased degradation. They concluded that changes in sex hormone levels could contribute to altered insulin metabolism and peripheral hyperinsulinemia in androgenized women with abdominal obesity (28). Hypertension is also independently associated with altered insulin clearance (31).
Hepatic insulin removal and hepatic insulin effects are correlated (32, 33). This relationship suggests that simple hormone-receptor interaction is not sufficient to explain all of the actions of insulin. Okuda et al. (25) showed that GH administration increased glucose levels in the portal and hepatic veins, but the fractional hepatic extraction of insulin was unchanged before and after GH. These results are different from those in untreated animals and suggest that the insulin resistance due to GH is associated with a loss of the glucose-induced increase in hepatic insulin extraction.
Studies of sulfonylureas have also found an association between hepatic insulin effects and removal. Glyburide and glipizide have differential effects on the liver. Glyburide augments the hepatic effects of insulin whereas glipizide has limited hepatic effects but increases the peripheral delivery of insulin (34). Chen et al. (35) showed that glyburide administration increased hepatic insulin extraction whereas glipizide had no effect. Adrogue et al. (36) also showed an increase in hepatic insulin uptake after glyburide therapy. These results support a relationship between hepatic insulin uptake and action, as do studies showing that FFA inhibit insulin binding, degradation, and action in liver (18). Insulin uptake and metabolism are inseparable from full insulin action.
Given the importance of the liver in insulin clearance (11), it is not surprising that liver disease may result in a decrease in insulin clearance (37, 38), although not all studies agree (39). The decreased clearance is due both to reduced hepatic function and to portosystemic shunts, but not to decreased insulin-degrading activity as assessed by activity per mg of protein in biopsy specimens (37, 39). The reduced hepatic clearance is also associated with reduced insulin sensitivity (38), again supporting the relationship of insulin degradation to insulin action.
The primary cellular mechanism for hepatic uptake and degradation of insulin is a receptor-mediated process. Most hepatic uptake is due to hepatocytes, with Kupfer cells contributing about 15% to the total (2). Pinocytosis may be a significant factor in hepatocyte insulin uptake at high insulin concentrations (40). Pinocytosis, i.e., non-receptor-mediated insulin uptake, may also be involved in some insulin actions (41).
B. Kidney
The kidney is the major site of insulin clearance from the systemic circulation (42), removing approximately 50% of peripheral insulin. In addition, the kidney removes 50% of circulating proinsulin and 70% of c-peptide by glomerular filtration (43). Insulin analogs are also cleared by kidney (44).Insulin clearance by the kidney occurs by two mechanisms (43): glomerular filtration and proximal tubular reabsorption and degradation (42, 45). Glomerular clearance of insulin may occur both by nonspecific diffusion and by specific receptor-mediated transport. After entering the tubule lumen, more than 99% of the filtered insulin is reabsorbed by proximal tubule cells, primarily by endocytosis (46). Relatively little insulin is ultimately excreted in urine. The kidney also clears insulin from the postglomerular, peritubular circulation, also via receptor-mediated processes (47, 48). In man, about one-third of the total renal clearance is by this route.
Recent studies, primarily from Rabkin and collaborators, have clarified the cellular mechanisms of renal insulin metabolism. In general, insulin degradation by kidney cells is accomplished by the same processes as by liver (see Section III) (49). Insulin is internalized into endosomes where degradation is initiated (50). Some insulin is released from the cell by retroendocytosis (51). Intact insulin release is increased by bacitracin, an inhibitor of IDE (52) and a commonly used inhibitor of cellular insulin degradation (53). Since most preparations of bacitracin contain multiple inhibitors (54), conclusions drawn from studies using bacitracin must be considered tentative (55). Nevertheless, the effects of this agent support the idea of multiple pathways of cellular insulin processing. As with liver, isolated endosomes from kidney can degrade insulin, probably by IDE (47).
Unlike the liver, lysosomes play a greater and earlier role in kidney insulin degradation, with most of the endosomal insulin and partially degraded insulin fragments delivered directly to lysosomes where degradation is completed (49). Bacitracin also interferes with lysosomal delivery and degradation. Intracellular and endosomal products of insulin degradation in kidney are identical to hepatic products (46) and consistent with the action of IDE (see Section IV).
The kidney plays an even greater role in insulin clearance in insulin-treated patients with diabetes than in normal subjects (43). Since insulin administered by subcutaneous injection escapes first-pass removal by the liver, the kidney has increased importance in insulin removal in these patients. Renal failure may reduce insulin requirements dramatically and increase the potential for hypoglycemia in insulin-treated subjects. In patients with residual β-cell function, no exogenous insulin may be required for glucose control. Renal failure may result in hypoglycemia in insulin-secreting patients, at least partially, due to reduced clearance. Sulfonylurea therapy, especially chlorpropamide, is a major risk for hypoglycemia in patients with renal insufficiency due to reduced clearance of the drug and reduced removal of insulin. Metformin is also contraindicated in renal failure for other reasons.
Renal failure and uremia suppress cellular insulin metabolism by mechanisms not well established (43). Muscle and hepatic insulin clearance and degradation are decreased in uremic subjects. Circulating inhibitors of insulin degradation have been implicated. The effects of these inhibitors persist in isolated muscle and liver preparations (56). Hypertension is also associated with altered insulin metabolism in kidneys and muscle (31).
C. Other tissues
Insulin not cleared by liver and kidney is ultimately removed by other tissues. All insulin-sensitive cells remove and degrade the hormone. After liver and kidney, muscle plays the major role in insulin removal. The mechanism involves insulin binding to its receptor, internalization, and degradation as in other tissues (see Section III) (57).Insulin uptake and degradation occur in adipocytes (12, 58, 59), fibroblasts (12, 60, 61), monocytes (62), lymphocytes (63), gastrointestinal cells (64), and many other tissues (2). All cells that contain insulin receptors and internalization mechanisms can degrade insulin. As in liver, insulin action in adipocytes correlates better with degradation than with delivery (65).
D. Extracellular insulin degradation
Under normal conditions, almost all insulin is degraded intracellularly or at least by membrane processes. A recent study suggests that significant amounts of insulin may be cleared and degraded extracellularly in wounds (66). This degradation appears to be due primarily to IDE and may play a role in the wound-healing activity of insulin.
III. Cellular Insulin Uptake
Figure 1
shows a model for the cellular uptake of insulin by receptor-mediated
processes. The initial step is receptor binding. Receptor-bound insulin
serves as a reservoir that can return intact insulin to the circulation
or deliver the hormone to an intracellular site (67). Internalized insulin can be processed through multiple pathways (68, 69),
resulting in degradation or release from the cell intact (diacytosis or
retroendocytosis). Intracellular pathways of insulin processing include
degradation of insulin in endocytotic vesicles (50, 70)
or delivery of intact insulin and degradation products to other
subcellular sites. Internalized insulin can be found in cytosol,
nucleus, Golgi, or other locations (2, 3, 71–77).
Insulin may be degraded in these sites or delivered to lysosomes for
degradation. Ultimately, most internalized insulin, partially degraded
insulin, and insulin fragments localize to lysosomes for completion of
degradation (78).
Figure 1.
A
model for cellular handling and degradation of insulin. The following
are steps in insulin processing: 1. Insulin binding to its receptor; 1A.
Degradation on plasma membrane; 2. Release of intact insulin; 3.
Co-localization (“capping”) in plasma membrane; 4. Formation of coated
pits; 5. Endocytosis; 6. Initiation of degradation in endosomes; 6A.
Further intracellular processing of insulin and degradation products;
6B. Diacytosis of intact insulin and degradation products; 7.
Distribution of insulin or its degradation products; 7A. Cytoplasm,
nucleus, Golgi; 7B. Lysosomes; 8. Delivery of insulin to lysosomes from
organelles other than endosomes; 9. Recycling of the insulin receptor;
9A. Delivery to the plasma membrane by diacytosis; 9B. Return to
original function state. [Reproduced with permission from W. C.
Duckworth: Endocr Rev 9:319–345, 1988 (2 ). © The Endocrine Society.)
Most receptor-bound insulin is internalized into endosomes where degradation is initiated (82, 83). Internalization is a property of the insulin receptor. The receptor has a juxtamembrane NPEY region that is associated with the internalization property of other receptors. This sequence, however, is not necessary for insulin internalization. Berhanu (84) has shown that the internalization has multiple steps, both nonenergy and energy requiring. At this point, the requirements for receptor-mediated internalization are not established. The general consensus is that internalization requires receptor phosphorylation (85, 86), but not necessarily involvement of IRS-I or PI3 kinase (87). Calzi et al. (88) recently suggested that the two alternatively spliced variants of the receptor (one with and one without 12 amino acids encoded by exon 11) may endocytose insulin at different rates, although others disagree (89, 90). They also suggested that pp120 might increase internalization by the exon 11-absent isoform but not the exon 11-present isoform (88). While the concept that different forms of the insulin receptor have different functions and effects on different pathways is attractive and probably true, it seems unlikely that the complexity of insulin-cell interaction and insulin action will be so easily defined. It may well be that different mechanisms for insulin internalization result in different intracellular processing mechanisms. This is an area that clearly deserves further study.
After formation, endosomes rapidly acidify, which results in dissociation of insulin from its receptor. Endosomal degradation of insulin is initiated before acidification of the vesicles (91). This is also true for non-receptor-mediated endocytosis. The degradation products generated in non-acidified endosomes, in vitro, are identical with products extracted from intact cells (92). IDE is present in endosomes, and the degradation products are consistent with limited cleavage of insulin by this enzyme (91, 93, 94). After vesicle acidification and dissociation of insulin from the receptor in vitro, the degradation products increase in number and complexity and are less like the typical intracellular pattern (91). An acidic proteinase has been implicated in degradation in late, or acidified, endosomes (95).
Not all of the internalized insulin is degraded in endosomes. Endosomal degradation varies depending on insulin concentration, duration of exposure, and other factors (96). Even with controlled conditions, in vitro, the maximum amount typically degraded does not exceed 50%. The remainder of the insulin is delivered to other subcellular compartments including cytosol, nucleus, and lysosomes (76, 78, 97, 98). The mechanism whereby insulin reaches these compartments is unknown but, clearly, intracellular insulin processing involves multiple pathways with IDE involvement in these pathways (99). Internalized insulin can be cross-linked to cytosolic IDE (98), and inhibitors of IDE alter nuclear transport of insulin (71, 100). The intracellular association of insulin with IDE may have implications for some action of insulin (see Section V).
The role of lysosomes in the cellular degradation of insulin has long been debated. Autoradiographic studies show lysosomal localization, although the distinction between acidified endosomes and lysosomes is somewhat arbitrary in most studies (2). Nevertheless, labeled insulin or products can be found in lysosome-like structures after addition of radioactive insulin to intact cells. A number of careful biochemical studies have concluded that lysosomes play little or no role in cellular insulin metabolism (2). The most likely explanation for this discrepancy is that degradation occurring early after exposure of cells to insulin is nonlysosomal and mostly endosomal, whereas the hormone that escapes endosomal degradation, either intact or partially degraded, is ultimately delivered to lysosomes for complete metabolism. In this system nonlysosomal degradation is the initial step and lysosomal degradation is the final step.
Not all insulin is internalized by receptor-mediated processes. Insulin can also be internalized by pinocytosis. At high concentrations of insulin, this may be the predominant mechanism since insulin stimulates pinocytosis.
IV. Cellular Insulin Degradation
A. Degradation products
The products generated from insulin by IDE in vitro have been well established (67, 101–112). Figure 2 shows the major and minor cleavage sites in insulin produced by this enzyme. Examination of the susceptible bonds does not reveal specific amino acid requirements but, rather, conformational restrictions. All but two of the cleavages occur in a restricted portion of the molecule (Fig. 2). This has led to the conclusion that IDE recognizes the three-dimensional configuration of the substrate rather than specific bonds. This conclusion is supported by studies of other substrates of the enzyme (see Section IV.C) (113).
Figure 2.
..
Cleavage sites of insulin. Top, The three-dimensional structure of insulin is shown with arrows
indicating the bonds broken by IDE. Note that the majority of the
cleavages occur closely spaced in a “cleft” in the molecule toward the
back of this model and away from the receptor-binding region.
[Reproduced with permission from W. C. Duckworth et al.: Biochemistry 28:2471–2477, 1989 (103 ). © American Chemical Society.) Bottom, The primary structure of insulin is shown with arrows indicating the bonds broken by IDE.
B. Assay for insulin degradation
The standard assay for IDE is solubilization of[ 125I]iodoinsulin in trichloracetic acid (TCA) (114). For high-quality assays, a well characterized labeled material is essential. HPLC-purified 125I(A14) monoiodoinsulin is the best characterized and most like unlabeled insulin, although 125I(B26) monoiodoinsulin also has properties similar to native insulin (115). With partially purified enzyme preparations, these materials generally give comparable results, but more purified enzyme preparations have different activities toward A14- or B26-labeled insulin (116). This is due to selective preparation of multiple forms of IDE with variable specificity using different purification approaches. For valid results with TCA, enzyme activities must be kept within the linear portions of the assay (generally <20 a="" assay="" class="link link-ref link-reveal xref-bibr" data-open="R114" dilute="" diluting="" do="" is="" linearly="" many="" not="" partially="" preferable="" preparations="" purified="" sample="" shortening="" since="" soluble="" tca="" the="" time="" to="">114).In spite of its widespread use, and usefulness, the TCA assay greatly underestimates the proteolysis of insulin. HPLC is by far the most sensitive assay, able to detect a single peptide bond cleavage. Some IDE preparations may degrade 6–10 times as much insulin by HPLC assay as by TCA. In general, the discrepancy is reduced by purification but the HPLC assay remains 2–3 times more sensitive than TCA even with highly purified preparations. Since most enzyme characteristics have been determined by TCA assay, many properties [e.g., apparent Michaelis-Menton constant (Km)] must be taken as an approximation.
Other, generally more cumbersome, assays such as receptor binding or immunoprecipitation have been described and may have specialized uses (114). For most purposes the TCA assay remains most useful and HPLC most sensitive.
C. Insulin-degrading enzymes
1. IDE
a. Characteristics
A proteolytic activity, insulinase, which degrades insulin with a high degree of specificity was described by Mirsky and Broh-Kahn (117) more than 50 yr ago. Although insulinase was characterized extensively using homogenates and other relatively crude preparations, purification of this activity was generally unsuccessful, and the existence of a single, specific insulin-degrading enzyme became somewhat doubtful. This led to the suggestion that cellular insulin degradation was due to a sequential process with an initial reductive cleavage of the molecule and subsequent proteolysis of the separate chains. The disulfide bound cleavage was attributed to the enzyme glutathione insulin transhydrogenase — now called protein disulfide isomerase (PDI) — with A and B chain degradation due to cellular nonspecific proteinases (1). While it is now clear that this hypothesis is incorrect in detail, cellular insulin degradation does occur in a sequential fashion with several identified steps. The initial degradative step occurs in endosomes with two or more cleavages in the B chain. This is followed by reduction of the disulfide bonds by PDI, or a related enzyme, yielding an intact A chain and several B chain fragments (96). The insulin fragments are then further cleaved, probably by multiple proteolytic systems, including lysosomes.Most studies support IDE as the primary enzymatic mechanism for initiating cellular insulin processing and degradation. This protein is the major insulin-degrading activity in cell homogenates (99). Most activity is found in cytosol with small but significant amounts present in other subcellular fractions, including plasma membranes, endosomes, and peroxisomes (see Section IV.C.1.b). The characteristics of cellular insulin degradation, including the effects of modifiers (e.g., Ca++) and inhibitors [e.g., bacitracin and N-ethylmaleimide (NEM)], are consistent with the properties of IDE (118). Intracellular degradation is initiated in endosomes, which contain IDE (91, 92). Cellular and endosomal degradation respond to modifiers of IDE activity. Internalized insulin interacts with IDE (98, 119), and monoclonal antibodies to IDE microinjected into cells inhibit cellular insulin degradation (120). Overexpression of IDE increases the rate of insulin degradation by cells (121). All of these data support IDE as the major mechanism for cellular insulin degradation. This information does not exclude other cellular processes for insulin degradation. As with insulin action, insulin degradation is a complex and multicomponent process. Redundancy and alternative pathways are an integral part of the overall process, as is true for other hormones (122).
Although the insulinase preparation by Mirsky and Broh-Kahn was undoubtedly a mixture of both specific and various nonspecific proteases, the specific degrading enzyme has now been identified (123). The current preferred name is insulin-degrading enzyme (IDE; EC 3.4.24.56). Various laboratories have used different names, including insulin protease, insulin-specific protease, insulysin, insulin glucagon protease, and insulinase (1, 2, 124). Isolation of the cDNA by Roth’s laboratory (123) definitely established the existence of a single IDE, and collaborative work between Roth’s and our laboratory showed the identity of IDE and insulin protease (insulinase) (125).
IDE is a 110,000 mol wt Zn++ requiring metalloproteinase (126, 127) but with a distinct Zn++ binding site (128–130). Typical IDE is sensitive to chelators, although inactivation requires extensive treatment with EDTA. Under ordinary conditions, EGTA has relatively little effect, but 1,10-phenanthroline is a very effective inhibitor. Metal-deprived enzyme or enzyme mutated to remove the Zn++ binding site (131) retains substrate binding activity but is without proteolytic activity.
Other divalent cations may play a role in the activity of IDE. Chelator-inactivated enzyme can be reactivated by Zn++, Mn++, Co++, and Ca++. The Zn++ reactivation curve shows a biphasic pattern with low concentrations activating the enzyme and higher concentrations inhibiting. The other divalent cations that reactivate the enzyme show little inhibition at higher concentrations and, in fact, can increase activity above that of the starting material. In particular, Ca++ may play a role in the activity of the enzyme in cells. Ca++- depleted muscle has decreased insulin degradation and reduced IDE activity. Ca++ addition to muscle increases insulin degradation (118, 127, 132).
IDE is susceptible to inhibition by sulfhydryl-active agents such as p-chloromercuriphenyl-sulfonic acid (PCMBS) PCMBS and NEM, indicating a requirement for a sulfhydryl group. The sulfhydryl group may be necessary for conformation of the protein rather than involvement in catalytic activity. Recent evidence suggests the presence of a number of different conformations of IDE with the forms having different properties and activities (our unpublished observations)
Because of the multiple forms and other factors such as formation of complexes (see Section VI), the general characteristics of IDE have been the subject of controversy with different laboratories reporting different apparent molecular weights, Km values, pH optima, inhibitor susceptibility, and other properties. As an example, the apparent molecular weight by molecular sieve has varied from 60,000 to 300,000. In general, however, properties include a neutral, but fairly wide, pH optimum (6.0–8.5), a Km for insulin of 20 nm with less purified preparations tending to have a higher value (200 nm), and susceptibility to bacitracin, NEM, 1,10-phenanthroline, and PCMBS, but not to leupeptin, E64, phenylmethylsulfonyl fluoride, pepstatin, antipain, or thiorphan (1, 2, 124, 133).
Initially, the only homologous protein to IDE was an insulin-degrading Escherichia coli enzyme named protease III (134), which cleaves the B chain of insulin in the same bonds as IDE (135). The general characteristics of the human IDE are very similar to those of protease III (136). The bacterial enzyme lacks a sulfhydryl dependence, and the other major difference is in relative substrate affinity. Mammalian IDE binds and degrades insulin better than insulin-like growth factor-I (IGF-I) and IGF-II, whereas protease III binds IGF-II better than insulin and degrades the two peptides equally. The E. coli enzyme is 27% identical to the human enzyme with areas of more than 50% homology, suggesting that IDE is a highly conserved enzyme from an evolutionary standpoint. The bacterial enzyme is also a metalloproteinase, requiring Zn, as is IDE (137, 138).
Protease III has the same Zn++-binding motif as IDE (HXXEH) with the two histidines essential for Zn binding and a downstream glutamate (glu 169), the third Zn++-binding residue (130, 139). The Zn++-binding site is an inversion of the sequence found in classical metalloproteinases (HEXXH), supporting the concept that IDE is a member of a new superclass of metalloenzymes (137, 138). Mutation of the metal-binding site results in loss of catalytic, but not ligand, binding activity (131). His108 and Glu111 are necessary for catalytic activity (128), and His108, His112, and Glu189 are the zinc ligands of human IDE with Glu182 influencing binding (139). Other residues important for binding have also been identified (140).
Studies in Neurospora crassa have shown insulin-binding (141) and insulin-degrading activity (142) very similar to IDE. In addition, a periplasmic enzyme in Acinetobacter calcoaceticus has properties consistent with IDE, again showing that this enzyme is highly conserved from bacteria to man (143). It is of interest that the E. coli enzyme, protease III, is also periplasmic.
The human and mouse IDE genes have been mapped to chromosome 10 and chromosome 19, respectively (144). The IDE gene appears to be a single, complex gene. Transfection of the hIDE cDNA produces a recombinant protein indistinguishable from the human enzyme.
Several different transcripts of the IDE gene have been reported in different organisms and different tissues (145), suggesting that multiple proteins may be produced by this gene and it may be differentially expressed and developmentally regulated (121, 146, 147). In E. coli, two proteins, a 110-kDa and a 50-kDa protein, have been identified. The 110-kDa protein has insulin-degrading activity and is homologous with the mammalian enzyme. The structure and activity of the 50-kDa protein is undetermined.
In rat, a careful analysis by Baumeister et al. (148) showed two transcripts of 3.7 and 5.5 kb in various tissues, consistent with other studies and consistent with the production of at least two IDE-related proteins. Of particular interest was the finding of different transcripts in testis (3.7, 4.1, and 6.1 kb). The 4.1- and 6.1-kb transcripts correlated with testis-specific gene activation and sperm cell differentiation, supporting a role for IDE-related proteins in differentiation and development in specific tissues. The products of the IDE gene, and IDE itself, have multifunctional roles in the cell. These roles vary from tissue to tissue.
There have been several reports of endogenous inhibitors, but none of these putative inhibitors have been definitively identified and characterized. The best characterized and most purified of the reported inhibitors is the one described by Ogawa et al. (149), which was purified by a series of procedures to yield a 14 kDa protein band on SDS-PAGE. It inhibited insulin degradation by IDE in a competitive manner and also inhibited the cross-linking of[ 125I]insulin to IDE. The general characteristics of this material are similar to a less purified preparation described by McKenzie and Burghen (150). A much larger inhibitor (mol wt 67,000) was described in another preliminary study, but no additional information is available about this material (132). Inhibitors of IDE are also found circulating in blood, and a deficiency of these has been implicated in a rare case of insulin resistance (151). The nature of these inhibitors is unclear.
At least some of the variable properties described for IDE are due to the complexing of this enzyme to other cellular proteins (152). This enzyme can be isolated together with a number of other cytosolic proteins and can be shown to have functional relationships with other proteins and protein complexes. This property and other findings have raised the likelihood that IDE is a multifunctional cellular protein as discussed below.
b. Subcellular distribution
Most IDE (95%) is found in the soluble (100,000 × g supernatant), thus cytosolic, fraction. Small amounts of IDE have also been found on the plasma membrane (153–157), in endosomes (91), and in peroxisomes (158, 159). Peroxisomes have the highest relative concentration (i.e., enzyme per mg protein), which is not surprising since IDE has a peroxisomal targeting sequence. Even with this, these organelles contain only a tiny fraction (1–2%) of the total cellular IDE (our unpublished data). Cellular function of IDE does not require peroxisomal localization (160). The multiple cellular sties where IDE is found supports a multifunction role for this enzyme (see Sections V and VI).c. Other substrates
Although insulin has the highest affinity (based on apparent Km), a number of other peptides can be degraded by IDE. These include glucagon, IGF-II, atrial natriuretic peptide (ANP), and transforming growth factor-α (TGFα) (110, 161–166). The cleavage sites in TGFα have been identified (166) (Fig. 3). ANP has been shown to be a high-affinity substrate for IDE (165). IDE rapidly degrades ANP at the Ser25-Phe26 bond, with slower hydrolysis of three additional bonds (113). Several related peptides including proinsulin (167), EGF, and IGF-I, bind to the enzyme but are hydrolyzed at very slow rates, making them competitive inhibitors (164, 168). Proinsulin intermediates are degraded by IDE proportionally to their insulin-like activities (169). IDE is not a general peptidase since a wide variety of other peptides and proteins are not affected by the enzyme (reviewed in Ref. 2). The selectivity of the enzyme has led to the conclusion that substrate recognition depends on three-dimensional features of the peptide rather than proteolysis of specific peptide bonds (2). This property is consistent with one of the proposed functions of IDE, i.e., as an intracellular receptor for insulin and other growth factors. This enzyme was isolated as an intracellular receptor for insulin and related growth factors in Drosophila before its degradative function was appreciated (103, 170–173). Subsequent studies discussed below have emphasized the potential receptor role for this protein.
Figure 3.
Cleavage sites of TGFα. The three-dimensional structure of TGFα is shown with arrows
indicating the bonds broken by IDE. Note that the majority of the
cleavages occur in the same region of the molecule toward the back of
this model and away from the receptor-binding region. [Reproduced with
permission from F. G. Hamel et al.: Biochim Biophys Acta 1338:207–214, 1997 (166 ) with kind permission of Elsevier Science-NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands)
Based on apparent Km, insulin is the preferred substrate for IDE, but an increasing number of other substrates are being reported. Since the demonstration that glucagon is susceptible to IDE (161), IGF-II, TGFα, ANP, insulin B chain, β endorphin, amyloid, and oxidized hemoglobin have been added to the list (162, 163, 165, 166, 174–178). Some related proteins such as EGF, IGF-I, and proinsulin bind to the enzyme but are very slowly degraded (163). Proinsulin intermediates are degraded proportional to their similarity to insulin (169). Interestingly, the biological activity of the intermediates correlates well with their susceptibility to degradation (179).
Most studies indicate that the enzyme substrate specificity of IDE is determined by the three-dimensional structure of the substrate. Baumeister et al. (148) have proposed that a β-pleat in the substrate is required for substrate binding, but additional properties determine proteolytic susceptibility.
d. IDE in clinical diabetes
Alterations in insulin-degrading activity occur in a number of clinical conditions. The most extreme of these is marked insulin resistance to subcutaneous insulin injection but relatively normal response to intravenous insulin (180). Most cases of insulin resistance attributed to this syndrome have other explanations, (181) and an approach to diagnose subcutaneous insulin resistance has been described (182, 183). Other clinical conditions associated with excessive insulin degradation include pancreatitis (184) and cirrhosis (37).e. Other functions of IDE
Insulin degradation is unlikely to be the only cellular function of IDE. IDE is found in all cells, not only insulin-sensitive ones. The relative distribution of the enzyme in rat, as measured by IDE transcripts, is as follows: high levels in testis, tongue, brain, and brown adipose tissue; moderate levels in kidney, prostate, heart, muscle, liver, intestine, and skin; and low levels in spleen, lung, thymus, and uterus (145, 146). Some classically non-insulin-sensitive tissues, therefore, have higher amounts than insulin-sensitive organs. IDE is developmentally regulated (185). During rat development, from 6–7 days to adulthood, IDE mRNA levels increased in brain, testis, and tongue, decreased in muscle and skin, and were unchanged in other tissues (146). IDE has also been linked to various cellular functions including differentiation of muscle and other tissues (186).All insulin-sensitive cells contain IDE, but the enzyme is also present in non-insulin-sensitive cells, supporting a multifunctional role for this protein. The highest level IDE gene expression is found in germinal epithelium. IDE and IGF-I receptor mRNAs are colocalized in oocytes, while IDE colocalizes with the IGF-II receptor in spermatocytes. In general, IDE expression shows significant correlation with both insulin and IGF receptors. Since IDE degrades IGF-II as well as insulin, these data suggest a role for IDE in the degradation and perhaps action of both insulin and IGF-II (145).
The multiple roles of IDE are exemplified by studies from Delovitch and colleagues. In a series of elegant papers (187–190), these investigators have implicated processing of insulin by IDE as necessary for the recognition of insulin by T cells. Interestingly, further processing by reduction of the disulfide bands is required to process insulin into a T cell epitope. This is entirely analogous to the postulated multistep system present in insulin-sensitive cells and consistent with the cytosolic delivery of endocytosed antigens. Similarly, a portion of the B chain, B-(9–23), has been implicated in the epitope specificity of T cells from nonobese diabetic (NOD) mice and may be related to the etiology of the diabetes (191–193). Administration of this peptide protects the mice from development of diabetes (194). The B-(9–23) peptide is very similar to the potential products of IDE made by its cleavages at B-9/10 and B-24/25 (see Section IV.A), suggesting IDE may be involved in antigen processing in these animals.
Another homolog of IDE, which is involved in propheromone processing and bud site selection in yeast, was reported recently (195, 196). This yeast IDE is related to another member of the IDE family, arginine dibasic convertase, which functions as a prohormone-processing enzyme in rat (197). The yeast enzyme has the same essential Zn++ binding site as IDE. Relevant to other proposed nonproteolytic functions of IDE, the bud site selection property of the yeast enzyme appears to be independent of its catalytic function. Multiple functions of IDE are summarized in Table 1.
| .. | |
|---|---|
Table 1.
Reported functions of IDE and homologs
| Function | Reference |
|---|---|
| Binding and degradation: | |
| Insulin | 2 |
| Glucagon | 161 |
| ANP | 165 |
| TGFα | 162, 166 |
| IGF-II | 110, 163, 168 |
| β-Amyloid peptide | 176 |
| Thiolase-cleaved leader peptide | 158 |
| Binding, limited degradation: | |
| Proinsulin | 167 |
| IGF-I | 110, 163, 168 |
| EGF | 173 |
| Proteolytic processing: | |
| Yeast propheromone | 196 |
| β-Endorphin | 175 |
| Regulation: | |
| Myoblast differentiation | 186 |
| Yeast bud site selection | 195, 196 |
| Proteasome activity | 152, 233–235, 237 |
| Androgen, glucocorticoid receptors | 212, 213 |
| Presentation of insulin epitopes | 188, 190 |
2. PDI. Varandani and co-workers hypothesized that insulin degradation occurred in a sequential manner with disulfide cleavage and subsequent proteolysis (201). While the obligatory initial disulfide cleavage of this hypothesis has been disproven, the sequential nature of insulin degradation with multiple steps and multiple enzymes involved is well supported by the available data. Of great interest is the finding by Varandani that disulfide cleavage by PDI may be involved in some of the biological effects of insulin (202). While the mechanism for this is speculative, studies have shown that fragments of insulin, specifically portions of the B chain, have biological activity (2). Since the initial step in cellular insulin processing involves cleavage of the B chain of receptor-bound insulin by IDE, disulfide cleavage by PDI would produce intracellular fragments of insulin with potential biological activity. These data further support a role for insulin processing and degradation of insulin in some of the biological effects of insulin.
PDI has multiple roles, but the primary cellular function may be to catalyze protein folding in the endoplasmic reticulum (203–205). Noiva and Lennarz (206) reviewed PDI properties and functions recently and concluded that protein folding is likely to be a major activity of this enzyme but also concluded that its physiological function remains unclear. Given the abundance of this protein (0.4% of total cellular protein), multiple functions are likely, making the two primary enzymes involved in insulin metabolism, PDI and IDE, multifunctional cellular proteins.
3. Acidic proteinases. Lysosomal enzymes can degrade insulin in an acidic environment. At a pH of 6 or less, IDE has little proteolytic activity (207) and thus, in acidified vesicles, insulin degradation occurs by other enzymatic mechanisms. In most cells, with the possible exception of kidney, delivery of insulin to lysosomes is a later event (92, 208–210). Most lysosomal degradation consists of completion of degradation of partially degraded insulin molecules. The specific lysosomal enzymes that degrade insulin have not been completely defined, but it is likely that several different ones participate in the proteolysis of insulin and insulin fragments. Cathepsin D (2), an acidic proteinase that can degrade insulin, is a potential candidate.
An acidic proteinase that can degrade insulin has been reported in endosomes, but its overall role has not been established (95). The degradation products in acidified endosomes in vitro are different from those extracted from intact cells. The endosomal acidic enzyme may play a specialized function in insulin degradation. Relatively little is known about the characteristics of this enzyme other than a lack of inhibition by EDTA or NEM and a partial requirement for Mn++ (95).
V. Biological Role of Insulin Degradation
The biological function of insulin clearance and degradation is to remove and inactivate circulating insulin (81). Hormonal control of metabolism requires delivery and removal of the hormone for cellular regulation. From a conceptual standpoint, hormone degradation is as important as hormone secretion, albeit not nearly as well studied. For many years insulin degradation has been viewed as a passive “sink” and not considered to be involved in regulation of metabolism. It is now clear that insulin removal and degradation are regulated processes and that abnormalities in insulin clearance are integral to diseases such as type 2 diabetes. Furthermore, insulin degradation is inextricably linked to insulin action. All insulin-sensitive tissues degrade the hormone, and the first step in insulin degradation, as in insulin action, is binding to the receptor. This linkage has led to the suggestion that the degradation of receptor-bound insulin is part of the “off” signal for termination of insulin action. It seems highly likely that this is true for some actions of insulin (81, 211).Recently, there have been suggestions that insulin interaction with and degradation by IDE may play a more direct role in generating insulin effects. IDE complexes with, and regulates, certain cytosolic organelles, specifically proteasomes, androgen and glucocorticoid receptors, and possibly peroxisomes.
IDE activates proteasome and steroid receptors by increasing proteasome proteolytic activity and increasing steroid receptor binding to DNA (212, 213). Other associations between glucocorticoid activity and insulin degradation have been reported (214). IDE also associates with peroxisomes, but the effect on activity of these organelles is unknown. Insulin added directly to isolated proteasomes or peroxisomes decreases activity, resulting in inhibition of protein degradation and fatty acidβ -oxidation (215). Since inhibition of protein degradation, fatty acid oxidation, and steroid action are biological effects of insulin (3, 216–219), these findings raise the possibility that a direct cytosolic interaction of internalized insulin with IDE could be involved in these cellular effects (3).
Several earlier studies support the feasibility of this concept. Intracellular insulin has biological activity (220–223). Roth’s laboratory (98) showed that insulin added to intact cells could be cross-linked to cytosolic IDE similar to the phagosome-to-cytosol pathway for antigens (224). Draznin and Trowbridge (225) and our laboratory (226) showed that inhibitors of insulin processing prevented insulin effects on cellular protein degradation and amino acid transport but not insulin-stimulated glycogen synthesis.
Many other studies have shown that the various actions of insulin can be differentiated at the postreceptor level (227–231). Broadly defined, the actions of insulin can be divided into short-term (glucose uptake and metabolism), intermediate-term (protein and lipid turnover), and long-term (cell growth and mitogenesis) effects. Postreceptor signal transduction events for glucose uptake and DNA synthesis are clearly different. Significant progress has been made in understanding cellular signaling and regulatory events in the short- and long-term effects of insulin, but little attention has been given to the mechanism of insulin action on protein and lipid turnover (intermediate-term effects). Several excellent recent reviews on insulin action are available (217, 219, 232).
Over the past few years, our interest in insulin degradation and IDE led us to study the relationship of these to general cellular proteolysis. Based on these studies and studies from other laboratories, we propose that the effect of insulin on cellular protein turnover is due to a direct intracellular interaction of insulin with the cytosolic IDE-proteasome complex. IDE activates the proteasome; intracellular insulin binds to and is degraded by IDE, resulting in decreased proteasome activity. The same process may also be involved in insulin inhibition of glucocorticoid and androgen receptor activity and peroxisome activity (215).
VI. Insulin-IDE-Proteasome Interactions and Control of Protein Degradation
Isolation and purification of IDE result in coisolation and purification of other proteolytic activity (152). This nonspecific proteolytic activity does not degrade insulin but does degrade a wide variety of other peptides, proteins, and artificial substrates. On SDS-PAGE, most of this nonspecific proteolytic activity elutes as a series of bands in the 20,000–35,000 mol wt range (Fig. 4). This and other properties, such as activation by SDS, identifies it as the multicatalytic proteinase (MCP) a high mol wt (700,000) cytosolic proteinase with multiple (as many as five or more) distinct catalytic sites. IDE and MCP coisolate through numerous purification steps including insulin-affinity chromatography and immunoaffinity chromatography using a monoclonal antibody to IDE. The two activities can be separated on ion-exchange chromatography using a complex series of salt gradients and steps. Removal of IDE results in considerable loss of MCP activity.The addition of insulin to the crude or partially purified IDE-MCP complex, but not to the ion-exchange-purified MCP that does not contain IDE, results in a decrease in at least two of the catalytic sites of MCP (233). Both the chymotrypsin-like site and the trypsin-like site are inhibited by insulin in a dose-dependent, noncompetitive fashion (Fig. 5) (234). The effect of insulin on MCP can be restored by recombining the ion exchange column eluent, thereby adding IDE back to the complex.
Figure 5.
Insulin inhibition of proteasome activity. Dixon plots of the inhibition by insulin of the chymotrypsin-like (right) and trypsin-like (left)
activities of the proteasome. The pattern of the plots is consistent
with a noncompetitive type of inhibition. [Reproduced with permission
from R. G. Bennett et al.: Diabetes 46:197–203, 1997 (234 ). © American Diabetes Association.)
Other IDE substrates also decrease activity of the MCP complex (234). Materials that bind but are poorly degraded, such as proinsulin, are partial agonists but are also antagonists of the full effect of insulin. The relative effects of IDE substrates on MCP activity are proportional to their affinity for IDE.
While the in vitro studies suggest a potential direct effect of insulin on a cytosolic IDE proteinase complex, physiological significance requires that this occurs in intact cells. Although some studies have suggested insulin effects on lysosomal proteolysis, recent studies have shown that insulin alters cytosolic protein degradation, presumably proteasome activity. Barrett et al. (236) showed that the postprandial increase in insulin resulted in alteration of cytosolic degradative activity. We have recently shown that proteasomal activity can be assessed in intact cells using a membrane- permeable substrate. Insulin decreases the degradation of this material (237).
In follow-up studies we have shown that an inhibitory monoclonal antibody to IDE blocks the effect of insulin on cellular degradation by the proteasome. This suggests an insulin-IDE-proteasome interaction resulting in control of cellular proteolysis. Hepatocytes were loaded with the inhibitory antibody and then incubated with insulin. No effect of insulin on total protein degradation or on proteasome activity was seen in these cells, whereas nonspecific IgG did not block insulin action (235). In terms of overall cellular protein breakdown, the fraction controlled by insulin is relatively small, averaging 20% in short-term studies. Since cellular protein degradation is a complex, multicomponent process with numerous controls, the actual effect may vary in different conditions.
VII. Summary and Conclusions
Insulin degradation is a regulated process that plays a role in controlling insulin action by removing and inactivating the hormone. Abnormalities in insulin clearance and degradation are present in various pathological conditions including type 2 diabetes and obesity and may be important in producing clinical problems.The uptake, processing, and degradation of insulin by cells is a complex process with multiple intracellular pathways. Most evidence supports IDE as the primary degradative mechanism, but other systems (PDI, lysosomes, and other enzymes) undoubtedly contribute to insulin metabolism.
Recent studies support a multifunctional role for IDE, as an intracellular binding, regulatory, and degradative protein. IDE increases proteasome and steroid hormone receptor activity, and this activation is reversed by insulin. This raises the possibility of a direct intracellular interaction of insulin with IDE that could modulate protein and fat metabolism. The recent findings would place intracellular insulin-IDE interaction into the insulin signal transduction pathway for mediating the intermediate effects of insulin on fat and protein turnover.
Acknowledgment
The authors would like to thank Janet Corr for preparation of the manuscript.
1
Duckworth
WC
Kitabchi
AE
1981
Insulin metabolism and degradation.
Endocr Rev
2
:
210
–
233
Inga kommentarer:
Skicka en kommentar