|Home | About | Journals | Submit | Contact Us | Français|
One of the most important unresolved issues in diabetes is the mechanism for the attenuated response to insulin, i.e. insulin resistance.
We hypothesize that the mechanism for the insulin resistance is due to uncontrolled protease activity in the plasma, on endothelial cells and in the tissue parenchyma. To examine this hypothesis we use of microzymographic techniques in the microcirculation, plasma zymography, and receptor labeling techniques with antibodies against an extracellular domain of the insulin receptor α.
The spontaneously hypertensive rat has an enhanced proteolytic activity and significant cleavage of the receptor with attenuated glucose transport. We present evidence for insulin receptor cleavage in a high fat diet and a transgenic model of diabetes.
These results suggest that cleavage of the extracellular domain of the insulin receptor, a situation that interferes with the ability for insulin to bind and provide an intracellular signal for glucose transport, may be involved in insulin resistance.
By definition, the metabolic syndrome X is associated with a multitude of organ and cell dysfunction. A large body of information is available about its clinical manifestations. But few proposals have been advanced that have put in place a conceptual framework to explain the diverse cell dysfunctions that accompany the syndrome, e.g. the insulin resistance, an elevated blood pressure, a prothrombotic or proinflammatory state. It is the goal of this report to propose a conceptual mechanism for a specific condition, the insulin resistance, as a candidate for a more general mechanism that may be responsible for the diverse dysfunction in the metabolic syndrome X and in other medical conditions.
Many hypertensives have telltale indicators for an inflammatory cascade [1,2]. Even though traditionally recognized for its elevated arterial blood pressure, the spontaneously hypertensive rat (SHR) also has a multitude of defects that are readily visible at the level of the microcirculation. While the SHR strain has no evidence for significant abdominal obesity, it exhibits a significant insulin resistance [3–9]. The insulin resistance is present in addition to endothelial apoptosis with loss of capillaries (capillary rarefaction) [10,11], a defect in leukocyte adhesion to the endothelium [12,13] with a chronically elevated circulating leukocyte count  (immune suppression), a prominent defect in the response to fluid shear stress [15–18] in addition to sleep abnormalities . The interesting aspect is that the insulin resistance appears in common with and yet distinct from other cell dysfunctions.
The insulin resistance in the SHR is qualitatively similar to the insulin resistance in other diabetic models, like the (Leprdb) mouse model  or the sand rat model . These models have in common a hyperglycemic state and an attenuated response to insulin .
In the past a large body of evidence in experimental models of hypertension and patients has been obtained which suggests that the enhanced organ injury in hypertension is accompanied by enhanced oxygen free radical formation (reviewed in [2,23,24]). Multiple forms of hypertension have enhanced free radical production [25–27]. In the SHR several classes of microvessels have an increased oxygen free radical formation, which if blocked leads to a reduction of blood pressure [25,28–31]. The free radical production in microvessels [32–34] is associated with enhancement of oxidase activity and reduction of enzyme activity that scavenges free radicals (e.g. superoxide dismutase, catalase) [35–37]. Reactive oxygen species modulate diverse functions  and exert secondary effects on microvessels, such as inhibition of endothelium-dependent vasodilation [39–42], reduced availability of nitric oxide and consequently enhanced arterial tone and blood pressure [31,43].
While oxygen free radical production has been linked to lipid peroxidation or deoxyribonucleic acid (DNA) damage, to date no conclusive mechanism has been advanced that links oxygen free radicals to specific cell dysfunctions, such as the insulin resistance. As alternative, we therefore focused on proteolytic pathway that may be associated with a complication such as insulin resistance.
The plasma protease activity in the SHR measured with a broad acting substrate (Enkzeck cleaved by metallo-, serine, acid and sulfhydryl proteases) is significantly elevated compared to the values in its normotensive control, the Wistar Kyoto (WKY) rat, or compared to the normotensive Wistar rat strain from which the SHR was generated (Figure 1). This overall protease activity in the SHR plasma detected by the broad reacting Enkzeck substrate can be reduced by a serine protease inhibitor (nafamostat mesilate; on average 20%) or by an matrix metalloproteinase (MMP) inhibitor (ethylenediaminetetraacetic acid (EDTA), 0.05 mM; on average 29%) suggesting the presence of both MMP and serine protease activity. The elevated MMP activity is also detected on microvascular endothelium by microzymography  with use of fluorescently quenched substrates specific for individual MMPs (Figure 2).
MMPs are expressed in a zymogen proform on endothelium and in the tissue parenchyma and under normal physiological conditions exhibit relatively low activity. Transformation of pro-MMP into an active enzyme requires cleavage or dislocation of the blocking peptide over the catalytic site of the enzyme. Among several mechanisms, a robust way to activate pro-enzymes in-vivo is by serine proteases (e.g. trypsin) . Another indication that serine protease activity in the SHR may be increased, is activation of protease-activated receptor-2 (PAR-2) . Treatment with angiotensin-converting enzyme (ACE) inhibitor (a serine protease inhibitor) blocks this process  (see below).
The protease activity in mice, homozygous for the diabetes spontaneous mutation (Leprdb) (on the C57BLKS background, ), or in the sand rat model  remains to be determined. Development of obesity in the db/db mouse and in diet-induced model of obesity induced mRNA levels for MMP-2, MMP-3, MMP-12, MMP-14, MMP-19, and tissue inhibitor of metalloproteinase – 1 (TIMP-1) (but not TIMP-3) in obese adipose but not lean adipose tissues. Pharmacological inhibition of the MMP activity prevents the expression of CCAAT/enhancer-binding protein β, a transcription factor that appears to play a role in the adipogenesis .
To directly examine possible cleavage of the extracellular domain of the insulin receptor we used a primary antibody against the extracellular domain of the insulin receptor alpha (Rα, Ñ sc-710 polyclonal antibody mapping to the N-terminus; 1:100) followed by a secondary antibody conjugated to the biotin/avidin with Vector NovaRED peroxidase enzyme substrate. In one set of experiments freshly harvested and heparinized cells of animals under general anesthesia (sodium pentobarbital, 50 mg/kg, i.p.) were directly labeled without centrifugation on a blood smear. Centrifugation was avoided since it may be associated with leukocyte degranulation, enzyme release, and receptor cleavage. The blood smear on a glass slide was air dried (37°C, 2 min), fixed in formalin solution, 10%, neutral buffered (1 min), and washed in deionized water (3×). The cell images were collected using a bright field microscope (40X and 60× objective) and digitally analyzed (NIH Image J) to determine light absorption as a measure for surface density of the extracellular domain of the insulin.
Male SHR (Harlan Sprague-Dawley) at age between 14 and 16 weeks were maintained on a normal rat chow without and with a chronic MMP inhibition (doxycycline, 5.4 mg/kg/day in drinking water for 24 weeks). These SHRs have on average a plasma glucose level of 118 mg/ml and 5.7% glycated hemoglobin (%A1C) as compared to 91 mg/ml and 4.1%A1C in the WKY rats and they have a reduced density of antibody binding sites on the extracellular domains of leukocytes and other cell types. The SHR has on average by a 5.5 % reduction of the extracellular domain density of the insulin receptor α on freshly collected and freely circulating leukocytes (Figure 3). The cleavage of the extracellular domain of the insulin receptor is significantly attenuated by chronic treatment with a broad acting MMP inhibitor (doxycycline, 5.4 mg/kg per day, 24 weeks), a process that also normalizes blood glucose and glycated hemoglobin values .
Since many of the circulating leukocytes in the SHR are in an activated state  and have high probability to be trapped in the microcirculation (instead of circulating), we also examined the insulin receptor density of naïve donor leukocytes from a Wistar strain after incubation with fresh heparinized SHR plasma (30 min exposure). This approach permits to directly examine whether the plasma of the SHR contains proteolytic activity with the ability to destroy the receptor and with minimal time for compensatory reactions (e.g. neogenesis, internalization). A period of 30 min exposure of naïve Wistar cells to SHR plasma reduces on average the extracellular binding site density of the insulin receptor α by 12% (as compared to 4% by WKY plasma and 0% by Wistar plasma). Such receptor cleavage by SHR plasma causes on average a 60% reduction of the transmembrane transport of fluorescent glucose (non-hydrolyzable glucose analog 6-NBD-deoxyglucose) into the cell cytoplasm (as compared to 13% cleavage by WKY plasma and 0% cleavage by Wistar plasma) .
This analysis showed that already after a 30 minute expose to SHR plasma (at room temperature), the extracellular domain receptor density on Wistar cells is significantly reduced by ~15% compared to the receptor density of the cells in control Wistar plasma. Even the WKY exhibits on average a ~4.3% reduction of the insulin receptor density, a situation that is in line with the fact that this (low pressure) control strain is not completely normal . The reduction of the extracellular domain insulin receptor density is accompanied by 61% reduction of glucose transport into the cell cytoplasm of naïve control cells of Wistar rats (detected by fluorescent analog transport  in SHR plasma (30 min incubation) and a 13% reduction in the WKY plasma as compared to the glucose transport of Wistar cells in Wistar plasma.
In order to test whether insulin receptor cleavage is detectable in other models with insulin resistance, we tested receptor cleavage on two alternative diabetes models, one that depends on calorie consumption (fat sand rat) and one with a genetic defect and spontaneous mutation of leptin receptor (Leprdb mice).
The study on the model that suffers from diabetes due to calorie overconsumption was carried out on the fat sand rat, an animal that naturally consumes low levels of calories and upon feeding normal Western diet will develop Type II diabetes. Male fat sand rats (Psammomys obesus, Harlan Sprague-Dawley)  at age between 5 and 6 weeks were transitioned after one week from a high-fiber low-calorie diet (Custom Rodent Diet, Sand Rat, 5L09; PMI Nutrition Int'l., LLC.) to a Western diet (LabDiet 5015) for a period of time between 24 and 28 weeks, while being kept on water ad libitum. The animals on such Western diet develop elevated plasma glucose levels (glucose levels raise from 73±9 mg/ml to 344±25 mg/ml, n= 3 rats, P<0.05) and a glycated hemoglobin level (HbA1C raise from 6.2±0.2% to >13%) One group of such diabetic rats was treated with an MMP inhibitor (doxycycline, 5.4 mg/kg/day in the drinking water) for a period of 24 weeks while the non-treated control group was kept on water. Treatment with the MMP inhibitor (doxycycline) served to raise the extracellular domain density of the insulin receptor by about 15% (Figure 4) in line with a significant reduction of the blood glucose (before treatment 94±41 mg/ml versus 87±7 mg/ml after treatment, n = 3 rats) and glycated hemoglobin level (Hb A1C before treatment 7.2±1.6%, after treatment 6.3±06%). The evidence supports the hypothesis that an environmental risk factor by calorie overconsumption in this model causes insulin receptor cleavage.
The study on the alternative genetic model of diabetes was with a leptin receptor deficiency. Heterozygote control mice (m Leprdb) (Background Strain: C57BLKS/J), homozygote Type 2 diabetic mice (Leprdb) (Background Strain: C57BLKS/J), and Leprdb null for TNFα (dbTNF−/dbTNF−) (Background Strain: C57BL/6J) were from Jackson Laboratory. The mice were maintained on a normal rodent chow diet. Male, 20–35 g nondiabetic m Leprdb mice, 40–60 g diabetic Leprdb mice and dbTNF−/dbTNF− mice were used in this study. The cross (dbTNF−/dbTNF−) of Leprdb with TNFα knockout mice is heterozygous for Leprdb and homozygous for TNFα knockout mice (TNF−/−). The dbTNF−/dbTNF− mice show phenotypes of hyperglycemia and obesity, the diabetic phenotype that is consistent with the penetrance of the leptin receptor mutation [47,51]. The mice were examined at an age between 14 and 16 weeks. The type 2 diabetic Leprdb mice had on average about a 19% reduction of the receptor labeling density, which was reduced to a non-significant 9% reduction in dbTNF−/dbTNF− (Figure 5). This evidence supports the hypothesis that genetic defects associated with insulin resistance may involve cleavage of the extracellular domain of the insulin receptor.
Thus, the underlying mechanism for the lack of insulin response in these two models and in the spontaneously hypertensive rat may be an uncontrolled degrading protease activity in the circulation and in the tissue parenchyma. Three widely different model of Type II diabetes have evidence for insulin receptor cleavage. The evidence is in line with the fact that MMP-1 (collagenase-1), MMP-2, and -9 (gelatinase A and B), MMP-7 (matrilysin), and MMP-14 (MT1-MMP) may cleave the insulin receptor alpha subunit at several positions in the extracellular domain (aa 27 to 760).
Many membrane receptors have cleavage sites for MMPs. Is it possible that other membrane receptors are subject to enzymatic cleavage of their extracellular domain? The SHR suffers from a defective adhesion mechanisms of its leukocytes to the postcapillary endothelium due to a reduced CD18 integrin expression [12,13] accompanied by a chronically elevated circulating leukocyte count . The extracellular domain density of CD18 on SHR leukocytes in the circulation is also reduced by SHR plasma to a similar degree as the insulin receptor. This CD18 receptor cleavage is abolished by chronic blockade of the MMP activity in the plasma, a process that also serves to return the circulating leukocyte count to normal values, likely since the circulating cell regain the ability to adhere to the vascular endothelium .
In a similar fashion the unchecked protease activity in the SHR leads to cleavage of the extracellular domain of the vascular endothelial growth factor receptor 2 (VEGFR2) . Such receptor destruction causes apoptosis of the endothelial cells. In the case of a capillary, which are composed of single narrow endothelial tube, endothelial cell apoptosis leads to loss of the entire capillary, i.e. to capillary rarefaction, a widespread phenomenon in the SHR. Thus the combined evidence suggests the first time that cleavage of membrane receptors may constitute a more general pathophysiological mechanism.
A common treatment for hypertension and diabetes is with ACE inhibitors. Could they possibly interfere with receptor cleavage? ACE inhibitors block MMPs in hypertensives [53,54]. Blockade of the MMP activity in SHR with ACE inhibitors lowers blood pressure [55–57] and glucose levels in diabetes. ACE inhibitors are effective against a variety of complications in the SHR [54,58]. Disruption of the blood brain barrier in the SHR involves MMPs [59,60]. In the brain of stroke prone SHR, the main forms of MMPs involved in the extracellular matrix alterations and basal lamina changes in cerebral microvessels are MMP-2 and MMP-9 whose activity is also significantly reduced by ACE inhibitors [54,61]. Early ACE inhibition affords a long lasting inhibition of cell death in the stroke prone SHR . In vitro studies have shown that ACE inhibitors are able to decrease the MMP activity directly, e.g. MMP-2 and MMP-9 are inhibited . ACE inhibitors may decrease MMP activity  besides lowering blood pressure. Most important, many of the complications that are associated with hypertension (e.g. atherosclerosis, cerebral and cardiac lesion formation and failure) are accompanied by enhanced MMP activity and reduced TIMP activity [63,65–67]. Thus, it is apparent that ACE inhibitors have a much broader enzyme blocking profile than just preventing the formation of angiotensin II. Their full range of activities will require a broader in-vivo analysis with inclusion of other reactions involving serine proteases, MMPs, and others.
The current evidence also suggests that other protease inhibitors, such a doxycycline, may serve to reduce the protease activity in diabetes and metabolic syndrome X. Their use in diabetes remains to be determined in controlled clinical trials. We noted that the doxycycline treatment needs to be prolonged for several weeks before a reduction of protease activity in association with blood pressure reduction and blood glucose levels could be observed.
There is an increasing body of evidence to suggest that diabetics have soluble fractions of receptors in their plasma. Soluble insulin receptor fragments include predominantly the extracellular alpha domain of the receptor, they are present in relatively low concentrations, elevated in both Type I and Type II diabetics, and correlate with the blood glucose levels . While in the past the significance of such receptor fragments in the plasma has been uncertain, their presence is in line with the current hypothesis that a proteolytic cleavage process generates them. These receptor fragments are an indicator of unchecked protease activity in the plasma and on endothelial cells of diabetics. The auto-digestion hypothesis is further supported by the evidence that soluble forms of receptors other than just the insulin receptor, e.g. the VEGFR2, are present in plasma of patients with metabolic syndrome and correlate with impaired angiogenesis . The concentrations of soluble VEGFR2 were found to correlate with those of the insulin receptor, indicating that the cleavage may not be limited to single receptor types . Proteolytic cleavage of surface receptor is observed under several pathophysiological circumstances [70–73] and includes soluble leukocyte adhesion receptors involved in inflammation .
The diversity of cell dysfunctions that accompany conditions, such as hypertension and the metabolic syndrome X, has escaped a rational explanation. The current report puts forward supporting evidence for a pathogenic mechanism that may put these mechanisms the first time under one conceptual roof that involves proteolytic degradation with direct destruction of the extracellular domain of receptors responsible for specific cell functions. Since many membrane receptors have extracellular binding domains that may be subject to destruction by degrading proteins, such as serine or cysteine proteinases, MMPs, and others, a variety of different receptors may be compromised by such a cleavage mechanism. The evidence in this report supports this notion in the case of the insulin receptor α in multiple forms of type II diabetes and in the specific case of the β2 integrins in the SHR . In other hypertensive and diabetic models and in patients the idea remains to be tested.
The cell dysfunction and end organ injury that is associated with unchecked degrading enzyme activity and direct membrane receptor damage may offer an opportunity for early intervention. Understanding the source and magnitude of these proteolytic enzymes, the reason why they are not completely blocked by native enzyme blockers (e.g. TIMPs, anti-α-trypsin), and development of novel interventions designed to block this enzyme activity in the circulation and in tissue parenchyma may open new opportunities for treatment. More sensitive techniques to measure protease activity in fresh unprocessed clinical samples need to be introduced into diagnosis . There is a need to identify fragments and cleavage sites generated by enzymatic destruction of membrane receptors, a process that may in part be hampered by peptide degradation in the presence of unchecked proteolytic activity in the plasma.
This research was supported by NIH grant HL 10881, by an unrestricted research gift of Leading Ventures (to GWSS), and by American Heart Association Scientist Development Grant (110350047A) and NIH grants (RO1-HL077566 and RO1-HL085119) (to CZ).
All animal protocols were reviewed and approved by the Animals Subjects Committees of the University of California San Diego and the University of Missouri.
Financial and competing interest disclosure
Dr. Schmid-Schönbein is a scientific consultant for Leading Ventures.
No writing assistance was utilized in the production of this manuscript.
* of interest
** of considerable interest