Low Circulating Protein C Levels Are Associated with Lower Leg Ulcers in Patients with Diabetes

Activated protein C (APC) promotes angiogenesis and reepithelialisation and accelerates healing of diabetic ulcers. The aim of this study was to determine the relationship between the incidence of lower leg ulcers and plasma levels of APC's precursor, protein C (PC), in diabetic patients. Patients with diabetes who had a lower leg ulcer(s) for >6 months (n = 36) were compared with age-, type of diabetes-, and sex-matched subjects with diabetes but without an ulcer (n = 36, controls). Total PC was assessed using a routine PC colorimetric assay. There was a significantly (P < 0.001) lower level of plasma PC in patients with ulcers (103.3 ± 22.7, mean ± SD) compared with control (127.1 ± 34.0) subjects, when corrected for age and matched for gender and type of diabetes. Ulcer type (neuropathic, ischaemic, or mixed) was not a significant covariate for plasma PC levels (P = 0.35). There was no correlation between PC levels and gender, type of diabetes, HbA1c, or C-reactive protein in either group. In summary, decreased circulating PC levels are associated with, and may predispose to, lower leg ulceration in patients with diabetes.


Introduction
Activated protein C (APC) is a plasma protease derived from its precursor, protein C (PC), which circulates in plasma at 3-5 μg/mL. APC was originally described as an anticoagulant but has recently been found to exert potent cytoprotective properties including the inhibition of in�ammation and apoptosis and maintenance of the endothelial and epithelial barriers [1][2][3][4]. APC exerts its cytoprotective effect through its receptor, endothelial protein C receptor (EPCR), which binds to both PC and APC with high affinity [5]. A soluble form of EPCR (sEPCR), circulating in normal human plasma [6], has similar affinity for binding PC as that of intact membranebound EPCR.
In humans, recombinant APC reduces the mortality rate in severe sepsis [7], and we have recently shown its potential application in the healing of chronic wounds [8,9]. Interestingly, biopsies taken immediately adjacent to chronic wounds in patients with diabetes exhibit very low total PC (PC plus APC) levels compared to normal skin [9]. In animal models, APC is neuroprotective aer stroke onset [10], protects diabetic nephropathy [11], signi�cantly inhibits the development of lung �brosis in bleomycin-induced lung injury [12], reduces intestinal injury in necrotizing enterocolitis [13], and accelerates healing in streptozotocin-induced diabetic rats [14]. In vitro, APC modulates keratinocyte and endothelial function towards a phenotype necessary to promote wound healing by enhancing reepithelialisation and angiogenesis [15][16][17]. Notably, total PC expression in skin surrounding lower leg ulcers in diabetic patients is lower than normal skin [9].
Taken together, these �ndings triggered our hypothesis that low total PC levels may predispose to lower leg ulcers in diabetes. e aim of the present study was to determine if an association exists between circulating levels of total PC and lower leg ulcers in patients with diabetes.

Methods
A total of 72 outpatients with diabetes mellitus participated. is study was approved by the Northern Sydney Health Human Research Ethics Committee, and written informed consent was obtained from each subject. e diagnosis of either type 1 or 2 diabetes mellitus was made according to the criteria of the American Diabetes Association [18]. irty-six patients had at least one lower leg ulcer, and these patients were matched for age, gender, and type of diabetes with 36 patients with diabetes with no history of previous or current lower leg ulcer. Ulcer types were classi�ed as neuropathic ( ), ischemic ( ), mixed neuropathic/ischemic ( ) or venous ( ). Peripheral ischemia was determined by absence of both dorsalis pedis and posterior tibial pulses on clinical palpation. Peripheral neuropathy was assessed by clinical insensitivity to a 10-gram mono�lament. All ulcers were located at or below the malleolus, except the venous ulcer which was located on the lower leg. Our control group consisted of matched patients with diabetes because patients with type 1 or type 2 diabetes have altered levels of circulating PC levels compared to normals [19,20]. Patients on warfarin or any anticoagulant therapy were excluded from the study.
Blood sampling was carried out in all subjects from an antecubital vein, and plasma was separated. All assays were performed in a routine diagnostic laboratory. Total PC was assessed using the Stachrom PC colorimetric assay aer activation of plasma PC with Agkistrodon contortrix venom (Diagnostica Stago, Asniers, France). Protein S and �brinogen were measured by an immunoturbidimetric and clot-based test, respectively, performed on a fully automated coagulation analyser (STAR by Diagnostica Stago). e prothrombin time (PT), activated partial thromboplastin time (APTT), and International Normalised Ratio (INR) were also performed on the STAR analyser. Soluble (s)EPCR was measured by ELISA (R & D Systems, Minneapolis, USA).
Data was tested for normality and, if not normally distributed, transformed to a normal distribution before modelling by linear regression using Stata 11.0. Age was included in the regression as a confounding covariate. Differences in proportions were by Fisher's Exact tests. To quantify any signi�cant associations within the data, pairwise Pearson correlation coefficients were calculated.

Results
Of the 72 patients with diabetes in this study, 36 had chronic (>6 month duration) lower leg ulcers, and 36 patients did not have any ulcers (control group). All dependent variables were normally distributed, except for sEPCR and HbA c which required log transformation. Results are shown in Table 1.
Between the 2 groups, there was no difference in age, sex, duration of diabetes, or HbA c . ere was no correlation between PC and gender, type of diabetes, HbA c , or CRP. ere was a negative correlation between PC levels and age ( , ) that remained when groups were analysed separately (Figure 1). e most striking difference was the signi�cantly lower levels of plasma total PC in patients with lower leg ulcers compared to control subjects ( ). Of the 36 patients with lower leg ulcers, 8 had total PC levels that were below the normal range (70%-180%), whereas only 1 of 36 control patients was lower than normal (Fisher's Exact ). Ulcer type was not a signi�cant covariate for plasma total PC ( ). Levels were ± 6% for neuropathic ( ) ulcers, ± 9% for ischaemic ulcers ( ), and 9 ± 9% for mixed neuropathic/ischaemic ( ). e protein C level of the only patient with a venous ulcer was 116%.
ere was no difference in protein S, APTT, or �brinogen levels between the 2 groups� however, there was signi�cantly higher INR, C reactive protein, and prothrombin time in patients with ulcers. INR was negatively correlated with PC in the ulcer group only ( , ). Plasma sEPCR levels did not statistically differ in patients with diabetes who had lower leg ulcers compared with matched controls.

Discussion
is is the �rst study to examine circulating PC levels in patients with diabetes who have lower leg ulcers compared to those without ulcers. Our results show that patients with diabetes who have lower leg ulcers have lower levels of plasma total PC than their counterparts without ulcers. ere was no difference in Protein S, APTT, �brinogen, or sEPCR between the two patient groups, suggesting that the low total PC levels in these patients are not a direct result of an altered coagulation pro�le. Blood glucose control appears to be unrelated to low PC levels in patients with lower leg ulcers, as there was no difference in HbA 1c between the two groups.
Test results were not available for all patients, and thus patient numbers are reduced for some tests, particularly CRP. Nonetheless, when corrected for age, there was a signi�cant increase in CRP in patients with lower leg ulcers compared to those without ulcers, suggesting increased in�ammation, which is a feature of diabetic skin ulceration. A further limitation to this study is that data on other diabetes complications, such as denervation or arteriopathy which may contribute to failed wound healing, were not available and may confound the observed relationships.
Liu et al. [21] have recently shown that high wound �uid concentrations of matrix metalloproteinase-(MMP-9) predict poor wound healing in diabetic foot ulcers. Considering that APC inhibits MMP-9 production by monocytes [17], it would be interesting to determine whether reduced APC in wound �uid results in increased MMP-9. Regardless of the mechanism, our results show an association between low total PC levels and lower leg ulcers in patients with diabetes. Whilst ∼20% of patients with diabetes and lower leg ulcers had total PC levels lower than the normal range, the mean total PC level (96%) of this group fell within the broad normal range (70%-180%). is demonstrates that it may not be necessary for total PC levels to be lower than the normal range for lower leg ulcers to occur. Further work is required to delineate what PC level constitutes "low" in terms of failed wound healing in diabetes.
Treatment of diabetic lower leg ulcers frequently presents a management challenge as they oen respond poorly to conventional wound management therapy, oen due to complications such as neuropathy and peripheral vascular disease. We found no signi�cant difference between these different ulcer types in this study. Previous evidence indicates that exogenous APC promotes healing of recalcitrant ulcers in patients with diabetes, acting via numerous different mechanisms including inhibition of in�ammation, stimulation of angiogenesis, and reepithelialisation [8,14,15]. APC is emerging as a potential therapeutic agent not only for lower leg ulcers but also for a number of other disorders including lung injury, spinal cord injury, and kidney injury [9]. Interestingly, plasma total PC levels are substantially decreased in patients who develop severe sepsis, and the level of total PC correlates inversely with morbidity and mortality [22]. Aer showing positive results in a clinical trial [7] and obtaining FDA approval for APC to treat sepsis in 10 years previously, Eli Lilly, the company who marketed the drug, controversially withdrew APC from the market in 2011.
Low circulating PC levels may either predict or be a consequence of lower leg ulcers. However, when combined with our previous �ndings that (i) APC treatment stimulates healing of lower leg ulcers [8,9] and (ii) total PC expression in skin surrounding lower leg ulcers is low [9], the current �ndings provide supportive evidence that low PC levels predispose to lower leg ulcers which do not heal in patients with diabetes. It is feasible that a blood test to measure PC may assist clinicians in the difficult judgement of whether a diabetic ulcer will heal or not. Further longitudinal clinical studies will help con�rm the value of such a test.