The C-peptide of proinsulin is important for the biosynthesis of insulin but has for a long time been considered to be biologically inert. Data now indicate that C-peptide in the nanomolar concentration range binds specifically to cell surfaces, probably to a G protein-coupled surface receptor, with subsequent activation of Ca2+-dependent intracellular signaling pathways. The association rate constant, K ass, for C-peptide binding to endothelial cells, renal tubular cells, and fibroblasts is ∼3 ⋅ 109M− 1. The binding is stereospecific, and no cross-reaction is seen with insulin, proinsulin, insulin growth factors I and II, or neuropeptide Y. C-peptide stimulates Na+-K+-ATPase and endothelial nitric oxide synthase activities. Data also indicate that C-peptide administration is accompanied by augmented blood flow in skeletal muscle and skin, diminished glomerular hyperfiltration, reduced urinary albumin excretion, and improved nerve function, all in patients with type 1 diabetes who lack C-peptide, but not in healthy subjects. The possibility exists that C-peptide replacement, together with insulin administration, may prevent the development or retard the progression of long-term complications in type 1 diabetes.
- sodium-potassium-adenosine 5′-triphosphatase
- endothelial nitric oxide synthase
- renal function
- autonomic nerve function
- G protein
during the course of insulin synthesis, C-peptide is cleaved from proinsulin, stored in secretory granules, and eventually released into the bloodstream in amounts equimolar with those of insulin (45, 52, 53). C-peptide has an essential function in the synthesis of insulin in that it links the A and B chains in a manner that allows correct folding and interchain disulfide bond formation. When C-peptide is removed from proinsulin by proteolytic processing, the COOH-terminal part of insulin's B-chain becomes exposed and free to assume an appropriate conformation for effective interaction with the insulin receptor (51). A schematic representation of the proinsulin molecule and its components is given in Fig. 1.
After the discovery of the mode of insulin biosynthesis, several early studies addressed the question of possible physiological effects of C-peptide. Insulin-like effects on blood glucose levels and on glucose disposal after glucose loading were looked for but not found (14, 17). Rat C-peptide was, however, found to diminish glucose-stimulated insulin release in rats both in vivo and in vitro (6, 57, 58, 63, 64), whereas the corresponding findings in other animals were less conclusive (8, 32). C-peptide was also reported to inhibit arginine-stimulated glucagon release from the isolated perfused rat pancreas (63) and to inhibit fat-stimulated gastric inhibitory polypeptide secretion by intestinal cells (6). Despite these reported effects, it became generally accepted that C-peptide possesses little or no biological activity and has no other role than its participation in insulin synthesis (33), a role that is emphasized by its name, “connecting peptide.” Recently, new data have been presented demonstrating specific binding of C-peptide to cell surfaces in a manner that suggests the presence of G protein-coupled membrane receptors. C-peptide may thereby stimulate specific intracellular processes, influencing renal and nerve function in C-peptide-deficient type 1 diabetes patients. This contribution aims to review these new findings and to examine the possible physiological role of C-peptide.
The classic manner in which bioactive peptides exert their effects is via specific binding as ligands to receptors. In such binding, a limited region of the ligand serves as “active site,” effecting the binding to the receptor. This segment of the peptide is frequently well conserved across species borders. The binding process can usually be studied via binding of a labeled peptide in a radioligand assay. In the case of proinsulin C-peptide, the lack of the two basic concepts, a conserved active site and an established ligand assay, has long hampered the recognition of C-peptide as a bioactive hormone per se. Instead, nonreceptor membrane interactions have been suggested to explain some C-peptide effects (19). Recently, however, binding studies using new technology have established a typical receptor interaction for C-peptide (44). Consequently, it now appears that C-peptide can be recognized as a receptor ligand; it is just that some properties were difficult to define initially and that additional or multiple effects may exist. The molecular interactions that currently form the basis of a receptor concept are outlined below.
The first study describing interactions between C-peptide and cell membranes appeared in 1986. Binding of tyrosylated125I-labeled rat C-peptide 1 was examined by use of cultured rat islet tumor cells (10). Evidence for specific binding of C-peptide was reported, and a Scatchard plot of the data suggested a nonlinear course. The demonstration of C-peptide binding provided support for the earlier observations that C-peptide influences the function of the islet cells (6, 57, 63, 64). Interpretation of the results was complicated, however, by the ongoing secretion of C-peptide from the cells during the binding studies and by the existence of two different rat C-peptides (55). There is a lack of other studies describing C-peptide binding to cell membranes by use of the radioligand binding technique; only one report for skeletal muscle cell membranes, with negative result, is available (68). Relatively few binding sites per unit cell surface area (10) may have contributed to the difficulties in demonstrating cellular binding of C-peptide.
Binding evaluated via Na+-K+-ATPase stimulation.
Intracellular effects of C-peptide have been examined by using fresh preparations of proximal segments from the rat nephron, a well defined experimental model (1, 5). Addition of homologous C-peptide to the tubular segments was found to increase the intrinsic Na+-K+-ATPase activity in a concentration-dependent manner (42) (Fig.2). Moreover, pretreatment of the tubular segments with pertussis toxin completely blocked this effect. These and other observations indicate that a G protein interacting with a ligand-activated receptor may be involved in a C-peptide signal transduction pathway. Pertussis toxin treatment, known to affect the α-subunit of Gi proteins, may thus interfere with the interaction between the G protein and a loop region of a membrane-spanning receptor (15), which in turn could abolish the C-peptide effects. Although the definition of a receptor is lacking, the Na+-K+-ATPase studies (42, 43) have established that C-peptide is capable of eliciting molecular interactions in cellular systems and that these interactions can be measured, albeit indirectly. The above results support the notion that C-peptide effects are mediated via a receptor and that G proteins may be involved in the signal transduction pathway.
Further studies of the Na+-K+-ATPase activity of rat renal tubular segments involved a set of C-peptide fragments, essentially covering the entire length of C-peptide in different subsets and analogs (43). Two sets of rat C-peptide fragments were found to elicit stimulatory effects on Na+-K+-ATPase activity, suggesting the presence of two different types of interactive sites. One was localized to the COOH-terminal part of the molecule, with maximal activity from the COOH-terminal pentapeptide segment, and the other to the midsegment of C-peptide, with maximal activity from the segment corresponding to positions 11–19 (43).
The results for the COOH-terminal pentapeptide were characteristic of a receptor-ligand interaction. This pentapeptide gave full replacement of the entire C-peptide activity; the remaining part of the molecule, des-(27—31)-C-peptide, was without effect in this assay, and the effects were residue specific (43). Consequently, the pentapeptide data supported the notion of a specific C-peptide receptor that recognizes the COOH-terminal pentapeptide segment. Notably, the residues at positions 1 (Glu) and 5 (Gln) of the pentapeptide are conserved in most mammalian species (Table 1). The above results for C-peptide fragment stimulation of Na+-K+-ATPase were all obtained in studies involving rat tubular segments (43), but similar findings have also been made for rat pancreatic islets (60).
C-peptide fragments from the midsegment behaved in a different manner. Their replacement activity was incomplete (maximum 80% of that for the intact molecule), remaining parts of C-peptide did not lack activity in the assay, and their fragment activities were not residue specific (fragments with other or nonnatural d-amino acid residues were also partly active) (43). The data for the midsegment fragments suggested the existence of a second “site” with nonspecific interactions in a receptor-atypical manner, apparently less effective than the receptor-like interactions for the COOH-terminal pentapeptide (43).
In conclusion, the combined C-peptide fragment data support the concept of a traditional receptor binding of C-peptide, localized to its COOH-terminal segment, but they also indicate the possibility of secondary effects via other mechanisms and other segments of C-peptide.
In a report by Ido et al. (19), it was suggested that C-peptide effects may be elicited via a mechanism that defies the general rule that peptide hormones act by binding to stereospecific receptors. Human C-peptide was found to exert beneficial effects on vascular and neural dysfunction induced by experimental diabetes in rats. These effects were seen not only with the native C-peptide but also with itsd-enantiomer and with reverse-sequence C-peptide (19). The authors suggested that the effects of C-peptide may be independent of chirality and peptide bond direction and exerted via mechanisms similar to those of amphipathic antimicrobial peptides. The latter peptides elicit their effects by nonchiral interactions with membranes, resulting in the formation of ion channels and interference with phospholipase A activity (2). Ido et al. suggested that the glycine-rich segment in the central region of C-peptide (positions 13–17) was important for such biological activity (47). As discussed above, this conclusion agrees with the second site deduced from the fragment studies involving determinations of Na+-K+-ATPase activity (43). The central region of C-peptide is largely achiral and reasonably well conserved in mammalian species (Table 1). It should be noted, however, that the C-peptide molecule differs considerably from that of antimicrobial peptides, and that the negatively charged C-peptide may not easily interact with cell membranes to form cation-selective channels. Thus direct measurements under in vitro conditions have recently demonstrated that C-peptide fails to bind to lipid vesicles (16) and that its secondary structure is not altered in the presence of lipid membranes (16). Consequently, the nonreceptor effects, although observed in two different assays, testing vascular function (19) and Na+-K+-ATPase stimulation (43), are not only not ascribable to receptor interactions but also not explainable in terms of membrane interactions in general. Furthermore, in the study by Ido et al., heterologous C-peptide was used in a supraphysiological concentration (100 nM), and the effects required 2–3 days to become evident, which complicates the evaluation of those findings. It can be concluded that C-peptide midsegment effects may exist but do not appear to involve receptor or membrane interactions and do not interfere with conclusions regarding the receptor-like interactions of the pentapeptide COOH-terminal segment.
Fluorescence correlation spectroscopy.
New evidence that C-peptide binds to specific cell surface receptors has recently been reported (44). Thus fluorescence correlation spectroscopy (FCS) has been applied to the evaluation of C-peptide-membrane interactions. In FCS, the Brownian movements of a fluorophore-marked ligand are observed after excitation by a sharply focused laser beam (9, 38). The small volume element (<1 ⋅ 10− 15 l) in which the measurements are performed can be shifted from the incubation medium containing the labeled ligand to the membrane surface of cells growing in culture. From the autocorrelation function of the fluctuations in fluorescence intensity, it is possible to characterize the diffusion time of the labeled ligand when it is free in the incubation medium or bound to a cell membrane (9, 38). Compared with the radioligand binding method, the FCS technique has the advantage of higher detection sensitivity and an improved signal-to-noise ratio combined with submicron resolution (9, 38).
In the case of C-peptide, labeling for FCS was achieved by attachment of tetramethyl-rhodamine to the NH2-terminal end of the molecule. Specific binding of human C-peptide could be demonstrated for cultured human renal tubular cells, skin fibroblasts, and saphenous vein endothelial cells (44). The association rate constant (K ass) for C-peptide binding was ∼3 ⋅ 109M− 1, and the specificity of the binding was evidenced by the consistent displacement of bound C-peptide by excess unlabeled C-peptide. Addition of a peptide with the same amino acid composition as human C-peptide but with the residues arranged in random sequence (scrambled C-peptide), or of d-enantio C-peptide, failed to displace bound C-peptide, demonstrating the stereospecific nature of the binding. In contrast, addition of excess COOH-terminal pentapeptide competitively displaced bound C-peptide, indicating that the COOH-terminal segment is involved in the binding process as measured by FCS, in agreement with the ability of the pentapeptide to stimulate Na+-K+-ATPase activity (43). Proinsulin, which includes the pentapeptide segment, failed to elicit displacement of bound C-peptide, suggesting that the free COOH-terminal end of the segment is required for binding. Likewise, addition of insulin, insulin-like growth factor (IGF)-I, IGF-II, or neuropeptide Y (NPY) was not accompanied by displacement of bound C-peptide, indicating absence of cross-reactions with these hormones and their membrane receptors. Labeled insulin bound to cell membranes was not displaced by excess concentrations of C-peptide (44,68). In the case of proinsulin, the finding is in contrast to reports that C-peptide may bind with low affinity to a proinsulin receptor (22,23). Finally, preliminary FCS evidence indicates the presence of species specificity; at physiological concentrations, rat C-peptide fails to bind to human cells (unpublished observation).
Preincubation of cells with pertussis toxin was found to abolish FCS-measurable binding of C-peptide at physiological concentrations (44), in agreement with the findings from the Na+-K+-ATPase data (42, 43). The FCS binding results before and after treatment of the cells with pertussis toxin are consistent with a typical allosteric mechanism of signal transduction: before pertussis toxin treatment there was a high-affinity interaction between C-peptide and its membrane receptor, reflecting one configuration of the receptor. After pertussis toxin, the FCS data indicated only a small component of low-affinity interaction between C-peptide and the receptor, which then most likely has assumed a new configuration secondary to the effect of pertussis toxin on the α-subunit of the G protein (15).
It has been a consistent finding that effects of C-peptide cannot be demonstrated in normal animals or healthy subjects (14, 17,19, 21, 28, 64, 65); it is only in animals with experimental diabetes or patients with type 1 diabetes and, consequently, no or very low C-peptide plasma concentrations, that specific effects have been observed (11, 25-31, 35, 37, 50, 65). This may have to do with the binding characteristics for C-peptide (Fig.3). Half-saturation of C-peptide binding to renal tubular cells determined by FCS was found to occur already at a concentration of 0.3 nM, and full saturation was seen at ∼0.9 nM (44). Thus it is likely that, in healthy humans, receptor saturation is reached already at the ambient physiological C-peptide concentration (0.5–1.5 nM), so that no additional biological activity can be expected from a further increase in C-peptide concentration. This may explain why no C-peptide effects were seen in the early studies involving healthy humans and animals (14, 17, 64). For a further understanding of the detailed nature of the C-peptide binding and its physiological effects, the receptor structure will have to be determined. Moreover, the FCS results quantitate the binding affinity to that applicable to physiological C-peptide levels and provide a C-peptide binding assay.
The combined C-peptide binding data from the five different approaches outlined above provide new insights into C-peptide membrane interactions. There is evidence of stereospecific binding of C-peptide to a cell surface receptor, which most likely is G protein coupled. The binding occurs in the low nanomolar concentration range, and the COOH-terminal pentapeptide appears to mediate the effects. This interaction fits the classic ligand-receptor concept. In addition, there are indications of another type of nonspecific, nonchiral membrane interaction localized to the midsegment of C-peptide, which will require further evaluation for detailed understanding.
SIGNAL TRANSDUCTION AND CELLULAR EFFECTS
As mentioned above, rat C-peptide was found to elicit a concentration-dependent stimulation of Na+-K+-ATPase activity at nonsaturating Na+ concentrations (42) (Fig. 2). Rat C-peptides 1 and 2, which differ by two amino acid residues (55), were largely equipotent in stimulating Na+-K+-ATPase activity (43), and scrambled C-peptide had no detectable effect. Subsequent reports have demonstrated that C-peptide stimulates Na+-K+-ATPase activity also in rat sciatic nerve (19, 54) and granulation tissue (19), pancreatic islets (61), and red blood cells (4). C-peptide also ameliorates the impaired deformability of red blood cells from type 1 diabetes patients (35). This effect was abolished after pretreatment of the erythrocytes with ouabain, compatible with C-peptide effects being mediated via stimulation of Na+-K+-ATPase activity (35), known to be reduced in red blood cells from patients with type 1 diabetes (7).
The C-peptide effect on Na+-K+-ATPase activity of renal cells was inhibited by pretreatment of the cells with pertussis toxin (42), as outlined above. The results of Ohtomo et al. (42) also indicate that C-peptide activates Ca2+-dependent intracellular signaling pathways. Exposure of cultured proximal convoluted tubular cells to homologous C-peptide in the concentration range 10−11–10−9 M resulted in rapid and consistent increments of the intracellular Ca2+ concentration. It is noteworthy that the C-peptide concentration required for increasing the intracellular Ca2+ concentration (42) was less than that needed for stimulation of Na+-K+-ATPase activity in renal tubule segments (Fig. 2) (42). This is most likely related to the fact that the tubule segment, but not the cultured cells, underwent a preparation procedure involving freezing, thawing, and collagenase treatment, possibly resulting in interference with cell membrane structures (42). When the cultured renal tubule cells were maintained in a calcium-free medium, exposure to C-peptide failed to increase intracellular Ca2+ levels. Moreover, addition of FK506, a specific inhibitor of calcium/calmodulin-dependent protein phosphatase 2B (PP2B), resulted in complete inhibition of the stimulatory effect of C-peptide. PP2B is of major importance in the regulation of Na+-K+-ATPase activity in tubular cells because of its ability to convert the phosphorylated, inactive form of Na+-K+-ATPase to its dephosphorylated, active form (1, 18). The simplified, overall signal transduction pattern that emerges is thus that C-peptide activates a membrane receptor coupled to a pertussis toxin-sensitive G protein, with subsequent activation of Ca2+-dependent signaling pathways and stimulation of PP2B, resulting in increased Na+-K+-ATPase activity (Fig. 4).
The possibility exists that C-peptide may act in synergism with other hormones. This is suggested by the finding that the dose-response curve for C-peptide and Na+-K+-ATPase activity in renal tubular segments was shifted more than two orders of magnitude to the left, indicating increased C-peptide effectiveness, in the presence of subthreshold concentrations of NPY (42). NPY is released after activation of the sympathetic adrenergic system and is known to act synergistically with norepinephrine (59). Tissue levels of NPY are upregulated in animals with experimental diabetes and C-peptide deficiency (62). The results may suggest that C-peptide effects are dependent on sympathetic adrenergic activity. It is noted that, with regard to stimulation of Na+-K+-ATPase activity, no synergistic interaction was observed between C-peptide and insulin (42). In contrast, the smooth muscle-relaxing effect exerted by C-peptide (human) on rat muscle arterioles was potentiated by the presence of a low insulin concentration (24). Further experimental work will be required for a better understanding of these phenomena.
Administration of C-peptide to type 1 diabetes patients is accompanied by circulatory responses: it results in increased blood flow in skeletal muscle at rest (29) and during exercise (28), augmented capillary blood cell velocity and redistribution of skin microvascular blood flow (11), and increased renal blood flow (31). Again, no effects are observed in healthy subjects (28) or animals (21). C-peptide has also been found to increase blood flow, capillary filtration coefficients, and the permeability surface-area product in the isolated perfused rat hindquarter, primarily indicating recruitment of capillaries (36). The cellular mechanism underlying this vasodilator effect of C-peptide has not been fully established, but preliminary evidence suggests that C-peptide, via an increase in the intracellular Ca2+ concentration, stimulates endothelial nitric oxide synthase (eNOS) activity (30, 34, 37) (Fig. 4). Thus C-peptide (human) was found to increase NO release from bovine aortic endothelial cells in a concentration-dependent manner. The C-peptide-induced nitric oxide (NO) release was abolished by the addition of an NO synthase inhibitor (34). Further support for the notion that C-peptide administration may be accompanied by augmented NO formation is provided by the observation that forearm blood flow increments induced by C-peptide in type 1 diabetes patients can be inhibited by a NO synthase blocker (30). In agreement with the above findings, it has been reported that C-peptide induces a concentration-dependent dilatation of rat skeletal muscle arterioles and that this occurs via a NO-mediated mechanism (24).
C-PEPTIDE AND RENAL FUNCTION
Patients with type 1 diabetes frequently develop glomerular hyperfiltration early in the course of their disorder (39, 40). Adequate insulin therapy does not correct this phenomenon (46). In contrast, patients with type 2 diabetes, in whom insulin and C-peptide levels are within or above the normal range, do not show glomerular hyperfiltration or hypertrophy (12, 48). These considerations prompted studies of possible C-peptide effects on renal function in type 1 diabetes patients (31). A group of young patients without signs of late diabetic complications were studied under euglycemic conditions. C-peptide infusion for 3 h at a rate sufficient to raise its concentration to physiological levels (∼0.9 nM) decreased the glomerular filtration rate (GFR) by 7%, and renal plasma flow increased modestly. Infusion of C-peptide to reach higher concentrations (∼2.1 nM) was not accompanied by further changes in GFR or renal plasma flow. Even though the observed effect was modest, it did establish a significant renal effect of C-peptide (31).
The influence of C-peptide on glomerular hyperfiltration, functional reserve capacity, and renal protein leakage have been examined in streptozotocin diabetic rats (50). Administration of C-peptide (human) for 90 min was accompanied by diminished glomerular hyperfiltration (-20%), improved functional reserve as evidenced by augmented GFR after glycine loading, and a marked reduction (−70%) in protein leakage compared with diabetic control animals. The specificity of the C-peptide effects was evident from the observation that infusion of scrambled C-peptide had no effect. Further evaluation of renal C-peptide effects in patients with type 1 diabetes has been extended to include more prolonged administration. In a double-blind, randomized study, patients with type 1 diabetes received by subcutaneous pump infusion for 4 wk either insulin plus equimolar amounts of C-peptide or insulin alone (27). In the C-peptide-treated group, GFR decreased on average by 6%, whereas no changes in GFR were seen in the group receiving insulin only. Moreover, a significant reduction in the level of urinary albumin excretion was seen in the C-peptide group but not in the group given insulin only (27).
The above findings in patients have been further explored in a study involving C-peptide administration for 3 mo (25). The aim was to evaluate the possibility that C-peptide administration may reduce the level of microalbuminuria in patients with early signs of diabetic nephropathy. Patients were studied in a double-blind, randomized, crossover design and received C-peptide plus insulin for 3 mo and insulin only for 3 mo. All patients were normotensive and had urinary albumin excretion rates between 25 and 220 μg/min before the study. During the C-peptide study period, urinary albumin excretion decreased progressively to significantly lower values than those found during the control period (Fig. 5). The albumin excretion had decreased by ∼40% at the end of the 3-mo period, whereas no significant change occurred during the control period. A similar response was seen in the urinary albumin-to-creatinine ratio. The patients remained normotensive throughout the study, and glycemic control improved slightly but to the same extent in the two treatment groups. Thus the diminished albumin excretion during C-peptide administration could not be ascribed to a reduction in arterial blood pressure or an amelioration of blood glucose control (25). The mechanism underlying the beneficial effect of C-peptide on renal function in diabetes is not known. However, it is possible that C-peptide may have exerted a direct effect on the glomerular handling of albumin, as suggested by the studies of renal function in animals with experimental diabetes (50). As discussed above, C-peptide has the capacity to stimulate both renal Na+-K+-ATPase (42, 43) and eNOS (34), and it may be hypothesized that C-peptide can influence glomerular membrane permeability and transport, as well as regional blood flow of the kidney, possibly leading to improvements in renal function in the diabetic state.
In summary, the present evidence demonstrates that C-peptide has the capacity to diminish glomerular hyperfiltration and reduce urinary albumin excretion in both experimental and type 1 clinical diabetes. Studies involving C-peptide administration of longer duration will be required to determine whether C-peptide may have a role in the prevention and treatment of diabetic nephropathy.
C-PEPTIDE AND NERVE FUNCTION
Diabetic neuropathy is an important clinical feature of type 1 diabetes. Either the peripheral or the autonomic nerves, or both, may be involved. Autonomic nerve dysfunction is found in nearly 40% of the patients, even though just a few present with clear-cut symptoms. The etiology of diabetic neuropathy is not fully understood. In addition to the toxic effects of elevated blood glucose levels, the possible influence of vascular dysfunction involving the vasa nervorum has been implicated (56). Reduced levels of endoneurial Na+-K+-ATPase (19) and diminished nerve blood flow (3, 20) are reported for the diabetic state. The effect of C-peptide on nerve function in diabetes has been evaluated in animal studies (streptozotocin-diabetic rats). C-peptide (human) prevented the decreased caudal motor nerve conduction velocity (MNCV) in diabetic rats but did not affect MNCV in healthy control rats (19). In the same study, C-peptide administration prevented the diabetes-induced reduction of sciatic nerve Na+-K+-ATPase activity. Similar results have been reported for spontaneously diabetic insulin-deficient BB/W rats (54). Administration of homologous C-peptide for 2 mo resulted in significant improvements in MNCV and nerve Na+-K+-ATPase levels compared with untreated controls.
Data from nerve function studies in patients with type 1 diabetes are also available. Patients with symptoms of diabetic polyneuropathy were studied twice under euglycemic conditions and during a 3-h intravenous infusion of either human C-peptide or saline in a double-blind study (26). Plasma concentrations of C-peptide rose to levels within the physiological range during C-peptide infusion. Heart rate variability during deep breathing, an indicator of autonomic, primarily vagal nerve activity, rose markedly (+50%) during C-peptide infusion but did not change during saline administration (Fig.6). A significant improvement was also seen in the brake index during tilting in the patients who showed a reduced index before the study; no response was observed during saline infusion (26). Indexes of motor or sensory nerve function did not change significantly during C-peptide infusion. Nerve function has also been evaluated in a subgroup of the patients who received C-peptide for 3 mo in studies of renal function (25). The patients showed signs of autonomic nerve dysfunction before the study, and after 3 mo of C-peptide administration their heart rate variability during deep breathing had improved by 20%. In contrast, no change or a slight deterioration was seen in the same patients during the control period with insulin therapy only. Signs of sensory neuropathy were present before the study in six of these patients; improved sensory nerve function, as evidenced by significantly improved temperature threshold discrimination, was observed after the 3-mo C-peptide treatment but not after the control period (25). Metabolic control was similar during the two study periods (25).
In summary, the combined experimental and clinical data provide evidence that C-peptide administration may ameliorate nerve dysfunction in type 1 diabetes. The stimulatory effect of C-peptide on nerve Na+-K+-ATPase activity (19, 54) and eNOS (34) may provide a background to the findings. The possibility that C-peptide may be beneficial in the treatment of diabetic neuropathy warrants further studies involving more prolonged periods of C-peptide administration. All of the above results relate to experimental and clinical type 1 diabetes. Whether C-peptide might exert a similar influence on nerve function in patients with type 2 diabetes, a disorder characterized by elevated levels of both insulin and C-peptide during its first phase, remains to be determined. However, there is preliminary evidence to indicate that there may be resistance to the action of C-peptide in muscle tissue from type 2 diabetes patients (66).
C-PEPTIDE AND GLUCOSE UTILIZATION
Early studies of C-peptide effects demonstrated that supraphysiological concentrations of C-peptide increase and extend the hypoglycemic effect of insulin in alloxan-diabetic rats (64). At an early stage, a possible effect of C-peptide on blood glucose levels or the disposal of a glucose load was also investigated in healthy subjects and type 1 diabetes patients, but with negative results (14, 17). Direct examination of the influence of C-peptide on glucose transport in skeletal muscle under in vitro conditions has, however, shown that human C-peptide is capable of stimulating 3-O-methylglucose transport in incubated human muscle strips in a concentration-dependent manner (67). The effect was seen in muscle strips from both healthy subjects and type 1 diabetes patients and appears to occur via a mechanism that is independent of the insulin receptor and of receptor tyrosine kinase activation (68). The effects on glucose transport of supraphysiological concentrations of insulin and C-peptide were not additive (68), which raises the possibility that C-peptide may still stimulate muscle glucose transport via the insulin-stimulated rather than the exercise-mediated pathway.
The influence of C-peptide on glucose utilization in the in vivo situation has been studied using the clamp technique in streptozotocin diabetic rats (37, 65). Supraphysiological concentrations of human C-peptide were found to elicit marked increases in whole body glucose utilization, whereas scrambled C-peptide had no effect (65). Physiological concentrations of rat C-peptides 1 and 2 were found to be equally potent in stimulating whole body glucose utilization in the diabetic animals, but they had no effect in healthy, nondiabetic rats (37). A major proportion of the C-peptide-induced stimulation of glucose utilization could be blocked by treatment withN-monomethyl-l-arginine (l-NMMA), suggesting that the effect is elicited through a NO-mediated pathway. Although circulatory effects of l-NMMA need to be considered, it is noteworthy that the effect of C-peptide on glucose utilization also remained blocked when adenosine was co-administered with l-NMMA, in an attempt to overcomel-NMMA-induced reductions in blood flow (37).
Data on C-peptide and glucose utilization are also available from studies in humans. Type 1 diabetes patients have been examined under euglycemic conditions, by use of the clamp technique and C-peptide infusion, at two dose levels (31, 49). Whole body glucose turnover increased by 25% when C-peptide levels were raised to ∼0.8 nM, but no further increase was observed when C-peptide concentrations rose to ∼2.1 nM, in agreement with the concept that C-peptide binding to cell membranes becomes saturated already at ∼0.9 nM (44). Direct measurements of forearm muscle glucose uptake during C-peptide administration in type 1 diabetes patients (28, 29) have confirmed that the augmented whole body glucose utilization observed in the euglycemic clamp study is a consequence of increased muscle glucose uptake rather than inhibition of hepatic glucose production. The question can be raised as to whether the observed short-term (2-h) effects of C-peptide on glucose utilization in type 1 diabetes patients (31, 49) will translate into lower blood glucose levels and/or diminished insulin requirements during long-term C-peptide administration. This does not seem to be the case, because in patients receiving C-peptide plus insulin for 3 mo, blood glucose levels, indexes of glycemic control, and insulin doses were all unchanged compared with the same patients given insulin only (25). In summary, C-peptide's relatively marked stimulatory effect on glucose utilization in in vitro experiments and animal studies, which may be NO mediated, appears to be less pronounced in humans and detectable in short-term studies but not during prolonged administration of C-peptide.
SUMMARY: C-PEPTIDE IS A BIOLOGICALLY ACTIVE PEPTIDE HORMONE
The currently available information establishes that C-peptide is not as biologically inert as previously believed. Instead, it now emerges as an active peptide hormone with potentially important physiological effects. Even though C-peptide is formed from proinsulin and co-secreted with insulin, we should consider the possibility that C-peptide is a separate entity with biochemical and physiological characteristics that are different from those of insulin. New data now demonstrate the presence of significant C-peptide-cell membrane interactions, and there is direct evidence of stereospecific binding of C-peptide to a cell surface receptor, different from that of insulin and other related hormones (44). The COOH-terminal pentapeptide segment is essential for binding (44) and constitutes an active site (43) in similarity to other biologically active but unrelated peptides, such as gastrin (13) and cholecystokinin (41). The nature of the C-peptide receptor remains to be determined, but it is most likely G protein coupled (42, 44). The K ass for C-peptide binding is ∼3 ⋅ 109M− 1, and saturation of C-peptide binding occurs already at physiological concentrations (44), which explains why C-peptide effects have been difficult to observe in the past. These findings all agree with the classic concept of ligand-receptor interaction. There is also evidence of another type of interaction localized to the glycine-rich midsegment of the molecule (19, 43), the physiological importance of which remains to be established.
In addition to the binding data, the first outline of an intracellular signal transduction pattern for C-peptide emerges. It involves the activation of Ca2+-dependent signaling pathways, with subsequent stimulation of Na+-K+-ATPase (4, 19,42, 43) and eNOS activities (24, 30, 34). Both of these enzyme systems are known to show attenuated activities in type 1 diabetes, particularly in renal and nerve tissue. There is now evidence to indicate that replacement of C-peptide in type 1 diabetes is accompanied by improved renal function, as evidenced by correction of glomerular hyperfiltration (27, 31, 50) and diminished urinary albumin excretion (25, 27), and amelioration of nerve dysfunction (25, 26). C-peptide replacement together with insulin administration, a therapy possibly closer to nature's own way, may thus be beneficial in type 1 diabetes patients.
The work discussed in this review has been supported by grants from the Swedish Medical Research Council (no. 11201), the Swedish Natural Science Research Council (no. 34115), the Marianne and Marcus Wallenberg Foundation, the European Commission (Bio4 CT972123), and Schwarz Pharma (Monheim, Germany).
Address for reprint requests and other correspondence: J. Wahren, Section of Clinical Physiology A2:01, Karolinska Hospital, SE-171 76 Stockholm, Sweden (E-mail:).
- Copyright © 2000 the American Physiological Society