The current study is the first to investigate in humans in vivo whether T1DM affects the tolerance for IR and the protection by ischemic preconditioning. Patients with T1DM appear to be more resistant to IR in the forearm skeletal muscle. The efficacy of ischemic preconditioniong to limit IR-injury, however, is lower in patients with diabetes than in nondiabetic control subjects, and is even completely abolished during hyperglycemia.
Despite optimal reperfusion therapy, morbidity and mortality in patients suffering an acute myocardial infarction remain significant . Therefore, novel strategies to limit IR-injury are warranted. Based on the finding that ischemic preconditioning profoundly reduces myocardial infarct size, several novel treatment modalities have recently been developed and tested in humans, including remote preconditioniong , and pharmacological preconditioning . For patients with diabetes, this need for additional treatment is even more urgent, given the fact that the risk of cardiovascular events is higher, and the mortality rate following a cardiovascular event is increased, both in type 1 and in type 2 diabetic patients [2, 3]. In addition, the incidence of heart failure is higher in patients with T1DM . There are several potential explanations for this worse outcome following infarction in patients with diabetes. In diabetic patients suffering from a myocardial infarction, infarct size (estimated with nuclear imaging) is larger than in nondiabetic patients, both after thrombolysis and after primary percutaneous coronary intervention [20, 21]. It was suggested, however, that the small difference in infarct size could not completely account for the 4–6 fold increased mortality in the diabetic patients in this study . This was recently confirmed in an animal study, in which the mortality rate following a myocardial infarction was higher in diabetic rats than in control rats, despite similar infarct size. It was speculated that autonomic dysfunction in the diabetic rats contributes to this increased mortality .
Despite the clinical observations that suggest that patients with diabetes are more susceptible to myocardial IR-injury, animal studies have provided conflicting results. A vast amount of studies have explored the tolerance of myocardial tissue for IR in animal models of type 1 and type 2 diabetes mellitus [9, 10, 23–27]. Although some studies have reported that the myocardium is more resistant to IR-injury in animals with T1DM, other studies have found either no effect, or an increased susceptibility. Probably, this tolerance to IR is dependent on the duration of diabetes (an increased resistance has been reported in particular early after the onset of diabetes), the animal species, and the duration of IR . In addition, it has to be taken into account that the animal models of T1DM do not accurately reflect the human pathology of diabetes in all aspects (e.g. the toxins administered to destruct the pancreas might have alternative mechanisms of action, and do not reflect the auto-immune destruction that occurs in humans).
Experiments in animal models of myocardial infarction have demonstrated that comorbidities, including diabetes, myocardial hypertrophy, and hypercholesterolemia can limit the protective effect of (ischemic) preconditioning . Several animal studies have investigated whether the diabetic heart is still amenable to the cardioprotective effect of (ischemic) preconditioning. Most studies in animal models of T1DM (mainly diabetes induced by the administration of streptozotocin) have demonstrated that the cardioprotective effect of ischemic preconditioning is abolished [28, 29]. Also, acute hyperglycemia completely abolished the infarct size-limiting effect of ischemic precondioning . In Goto-Kakizaki diabetic rats, the cardioprotective effect of preconditioning could be restored, however, by increasing the intensity of the preconditioning stimulus, illustrating that diabetes increases the threshold for preconditioning . It has been suggested that the diminished potential for cardioprotection in diabetes is due to impaired function of the ATP-dependent potassium channel (KATP-channel), or due to decreased phosphorylation of important signalling kinases including Akt and glycogen synthase kinase (GSK)-3β [10, 26].
These findings of a reduced efficacy of ischemic preconditioning have been confirmed in experimental studies on IR-injury in human atrial trabeculae: in patients with diabetes, the cardioprotective effect of ischemic preconditioning was either abolished, or the threshold for cardioprotection was increased, possibly due to impaired opening of KATP-channels [32–34].
In our study, annexin A5 targeting after forearm IR was lower in patients with T1DM than in nondiabetic control subjects, indicating an increased resistance to IR. We postulate that this is, at least in part, due to the protective effect of insulin against IR-injury. The patients with diabetes and strict administration of insulin during the experiment only marginally differed from healthy controls with respect to the plasma glucose concentration while their plasma insulin concentration was significantly higher. These two groups did not differ at baseline in other respects. Therefore, we propose that the higher circulating insulin contributes to the observed tolerance to forearm IR in this group. This conclusion is supported by a large body of preclinical evidence indicating a protective effect of insulin against IR-induced cell death [35–37]. Based on this preclinical evidence, we expected an increased targeting of annexin A5 after ischemic exercise in the patients with T1DM who were studied during hyperglycemic conditions. This, however, was not observed. We postulate that this is because the protective effect of chronic insulin treatment was still present after skipping one dose of insulin and remained effective during the experiments.
Despite this reduced susceptibility for IR-injury, additional protection by ischemic preconditioning was much less effective in patients with T1DM than in healthy control subjects. In more detail, this protective effect was small, and only observed in patients during normoglycemic conditions. During hyperglycemia, ischemic preconditioning did not reduce annexin A5 targeting at all. In the patients who were studied during hyperglycemia, the plasma insulin concentration did not differ from that observed in healthy volunteers. Therefore, it is unlikely that insulin affected the protective effect of preconditioning in these patients, and it is more likely that the hyperglycemia abolished this protective effect. This is supported by previous preclinical observations in animals using myocardial infarct size as an endpoint [30, 38] and it is consistent with epidemiological evidence in humans . A possible explanation for this phenomenon is an acute impairment of mitochondrial KATP-channels in response to hyperglycemia . The patients who were studied during hyperglycemia were significantly older than the healthy volunteers and the patients who were studied during normoglycemia. Advanced age has been associated with a reduced efficacy of ischemic preconditioning to protect against ischemic cell death . However, the interaction between ischemic preconditioning and experimental condition remained when age was incorporated as a covariate in the analysis of variance, excluding that age is a significant confounder.
Taken together, these data provide human experimental evidence for aggressive normalization of plasma glucose in patients with T1DM who experience repeated bouts of ischemia, such as angina pectoris or transient ischemic attacks, in order to optimize benefit from endogenous protection by ischemic preconditioning. These observations fit well with clinical data suggesting benefit from insulin treatment in critically ill patients with a disturbed tissue perfusion .
There are several limitations of our study. Since targeting of annexin A5 after forearm IR was low in patients with diabetes, the potential window of protection by ischemic preconditioning was smaller. This could have affected the power of our study to detect an effect of ischemic preconditioning in these patients. Indeed, in the patients who were studied in normoglycemic conditions, ischemic preconditioning tended to reduce targeting: in 6 out of 8 volunteers targeting was reduced (p = 0.087). Our failure to detect a significant protection by ischemic preconditioning in this group of patients might therefore result from a lack of power due to reduced targeting at baseline. Despite this small window of opportunity to further reduce annexin A5 targeting, in the patients with type 1 diabetes the effect of ischemic preconditioning was significantly affected by the experimental condition (hyperglycemic/normoinsulinemic or normoglycemic/hyperinsulinemic), consistent with a direct effect of hyperglycemia on the efficacy of ischemic preconditioning to prevent targeting of annexin A5. Secondly, since we did not use a hyperinsulinemic clamp technique, serum glucose levels tended to increase in the normoglycemic group during the last hours of the reperfusion period. Furthermore, our experimental design resulted in differences in both serum glucose and insulin between the two groups of patients with diabetes, complicating the interpretation of the study. However, the current design was chosen to allow an optimal comparison between the healthy volunteers and the normoinsulinemic hyperglycemic patients with diabetes.