Sortilin levels are associated with peripheral arterial disease in type 2 diabetic subjects

Sortilin, a 95-kDa protein mainly expressed in hepatocytes, takes part in intracellular protein sorting between the trans-Golgi network and endosomes [1]. In a lesser quantity, it is also expressed on cell membrane and small quantities of sortilin have been found in circulating form. Sortilin is codified on chromosome 13 by the gene SORT1, whose polymorphisms have been linked with circulating levels of low-density lipoprotein (LDL) cholesterol and with the lifetime risk of developing clinically significant atherosclerosis [2‐12].

Sortilin seems to take part in apolipoproteins trafficking inside hepatocytes, binding to apoB100 and promoting its trafficking to endosomal system, thereby promoting its degradation and reducing circulating levels of very low-density lipoprotein (VLDL). Moreover, membrane sortilin acts as a receptor for circulating LDL, promoting their uptake by hepatocytes via an LDL-receptor (R)-independent mechanism [13‐17]. After the discovery of its link to atherosclerosis, there has been considerable interest in elucidating additional functions of sortilin in different cell types. Thus, sortilin has been discovered to play a role in intracellular cytokine traffic [18, 19] and to be expressed in platelets, from where it is released after activation [20]. Among those additional roles, the sortilin function as a membrane LDL-R has been implicated in LDL uptake by macrophages and in foam cell formation and directly in inflammatory mechanisms during atherosclerotic plaque formation and progression [21, 22]. Additionally, sortilin seems to be a fundamental player in hepatic and muscular response to insulin and may be a missing link between insulin resistance and hypercholesterolemia [23‐25].

Aside from its potential as a therapeutic target, sortilin has been found in increased quantities in patients with coronary artery disease (CAD) and in patients with CAD risk factors compared to control subjects [20, 26].

Peripheral artery disease (PAD) is a well-known manifestation of atherosclerosis, which is associated with considerable disability and mortality. Lower limb PAD (LL-PAD) is of particular concern in diabetic patients and in active or current smokers, where the incidence and prevalence of the disease are particularly high. Ankle-brachial index (ABI) is currently recommended as the main screening tool for PAD in diabetic patients and in those with multiple risk factors [27]. A reduction in ABI, however, represents an already clinically significant stage of the disease. A tool used to find patients with higher risk of developing significant LL-PAD would be useful in identifying patients at a particularly high risk of developing clinically significant PAD and in tailoring more aggressive preventive and diagnostic strategies.

Considering the already proven links between the sortilin and multiple aspects of atherosclerosis [28, 29], metabolic imbalance [30] and diabetes [31], it is possible to hypothesize a connection between sortilin and PAD in the diabetic scenario.

The aim of this study was to assess the possible role of sortilin as an easy-to-measure marker for the presence of LL-PAD in an in-patient population with type 2 diabetes mellitus (T2DM).

During the study period, 154 patients with a confirmed diagnosis of T2DM and not on statin therapy in the last 30 days, were included in the study. Eighty of the patients did not fulfill the criteria for a diagnosis of PAD, while the remaining 74 had a confirmed diagnosis of PAD or received a diagnosis of PAD during the index hospitalization. The demographic and clinical characteristics of the two groups are reported in Table 3. PAD patients had higher blood pressure values (P = 0.024), were more often smokers (P = 0.032) and were affected by CAD (defined as history of ischemic heart disease and/or previous coronary revascularization) more frequently (P = 0.031). There were no significant differences between groups regarding body mass index (BMI) (P = 0.52), median duration of diabetes (P = 0.97), fasting glucose (P = 0.75), glycated hemoglobin (P = 0.59), total cholesterol (TC) (P = 0.78), HDL-cholesterol (P = 0.67), LDL-cholesterol (P = 0.73) and triglyceride (P = 0.99). No statistical difference in terms of eGFR (P = 0.12), aspartate aminotransferase (P = 0.61), alanine aminotransferase (P = 0.45), total bilirubin (P = 0.33) and alkaline phosphatase (P = 0.56). Furthermore, there were no significant differences between groups regarding the complete blood count (Table 3). No statistical difference in terms of diabetic therapy was observed between the two patient groups. According to the Fontaine’s classification 36 patients were defined as stage II, 24 as stage III and 14 as stage IV.

Sortilin concentration was higher among patients with PAD than among those without PAD (1.61 ± 0.54 ng/mL versus 0.67 ± 0.30 ng/mL, P < 0.01), as highlighted in Fig. 1. Moreover, when evaluating sortilin concentration according to patients’ functional status, a trend was evident, with higher levels of circulating sortilin in patients with more advanced disease (Fig. 2).

The multivariate logistic regression analysis showed that, after adjustments for cardiovascular risk factors, male gender, smoking, hypertension, hyperlipidemia, total cholesterol, HDL-cholesterol, LDL-cholesterol and sortilin levels were independent determinants for PAD occurrence in patients with T2DM (Table 4).

The ability of the area under the ROC curve based on sortilin levels to predict the presence of PAD in diabetic patients was 0.9133 (Fig. 3).

The ABI is a very commonly used tool, easy to operate and it is able to identify PAD very effectively. When ABI is suggestive of PAD, however, we are often faced with a patient with an already advanced disease [39]. If we consider that PAD in diabetic patients can be even more aggressive and can manifest the symptoms later, the research for tools useful for an early diagnosis appears essential [40]. There are some potential biomarkers able to identify a subset of diabetic patients with PAD more prone to suffer major amputations; for instance, the tumor necrosis factor receptor 1 (TNFR1) has been associated with major amputation outcome [41]. Another possible biomarker for the presence of PAD in diabetic patients is represented by the high mobility group box 1 (HMGB1) which, together with the osteoprotegerin (OPG), has been associated with the presence of PAD in these patients [42]. Furthermore, also the fibroblast growth factor (FGF) 23 has been correlated with the presence of lower extremity atherosclerotic disease in diabetic patients [43]. In this scenario, sortilin may be of clinical interest as a simple, easy to measure biochemical marker for atherosclerotic disease. Numerous evidence suggests that sortilin is implicated in the pathogenesis of many inflammatory and metabolic diseases, including T2DM and atherosclerotic complications of diabetes [31]. In addition, sortilin has been shown to be an energetic regulator of lipid metabolism [44]. No definitive data about the role of sortilin in atherosclerosis and diabetes, however, are available. We have evidence that sortilin is decisive in the formation of atherosclerotic plaque and that therapeutic agents capable of reducing serum sortilin levels are able to improve atherosclerosis [30]. Sortilin-dependent uptake of LDL into macrophages represents an interesting mechanism of foam cell formation, the fundamental lesion of atherosclerotic disease [21]. There are data demonstrating that liver sortilin, however, could play a positive role in atherosclerosis, improving cholesterol metabolism and reducing LDL levels, which are fundamental players in the formation of atherosclerotic lesion [15]. Recent data have shown that a process dependent on autophagy degrades apoB that was diverted from the secretory pathway by sortilin and provides a mechanism contributing to the reduction of LDL cholesterol [45]. It is therefore possible that sortilin performs different functions depending on the district where it is produced, on the receptors which it binds and on the extracellular milieu where it is released. Oh and colleagues [26] have shown that this protein is associated with coronary artery disease and diabetes. To our knowledge, this is the first time that the sortilin serum concentration has been assessed as a potential biomarker for peripheral atherosclerotic artery disease of the lower limbs in a diabetic population. Our study shows a strong correlation between sortilin concentration and lower limb clinically significant atherosclerotic disease in T2DM patients. Previous experimental evidence has shown that sortilin is involved in the formation of atherosclerotic plaque. In particular, in a mouse model it has been shown that sortilin facilitates plaque formation by promoting the inflammatory setting [22]. It is therefore possible that in our population of diabetic patients, high sortilin values are responsible, at least in part, for atherosclerotic lesions of the lower limbs. An additional important data is that the levels of sortilin correlate with the PAD severity in diabetic patients, as if there were a dose-dependent relationship. Of particular interest, there is very little overlap between sortilin concentration in patients with PAD and in patients without PAD. The ROC curve also confirms that sortilin levels are able to predict the presence of PAD in our population of diabetic patients. If such result was confirmed, sortilin concentration could prove a biomarker with excellent sensitivity and specificity for peripheral artery disease in diabetic subjects.

A strength of this study is the fact that we enrolled only patients who were not on statin therapy. Given that statin therapy may affect sortilin concentration, our result, thus, are unlikely to be due to interference of pharmacological therapy. On the other hand, as many patients with diabetes or at high cardiovascular risk are more likely to be on statin therapy, our results need to be confirmed in a larger, more representative population. Statin therapy has been shown to reduce statin levels in patients with CAD [30]. Even circulating PCSK9 has been independently related to sortilin, and also their correlation is affected by the statin therapy [46]. Since our patient population was statin-free, the data we have documented are not affected by this aspect. At the same time, this finding could reinforce the hypothesis of using this biomarker to stratify patients, both from a diagnostic and a therapeutic point of view. It would be justified to assess PAD in diabetic patients with higher sortilin levels and it would be useful to start a more aggressive statin treatment in this subset of patients.

It is not possible to definitively clarify whether sortilin levels are a cause or effect of atherosclerosis, or even both, in a sort of vicious circle. The earlier studies on sortilin highlighted its causative role in atherosclerotic disease, mainly due to the action of intracellular sortilin. In fact, in the study by Musunuru and colleagues [15], protective SNP was associated with a reduced intracellular concentration of sortilin and it was this intracellular action that is thought as associated with atherosclerosis. It has been later shown, however, that sortilin is actively secreted by activated platelet and may play a role in foam-cell formation. Thus, circulating sortilin concentration may also be a consequence of atherosclerosis.

A limitation of our study is that its cross-sectional nature is not able to establish causal relationship between the findings. Even so, our result lend support to the hypothesis that sortilin actions in extra-hepatic tissue are important in the atherosclerosis physiopathology. A further limitation is the small number of patients included in the study, due to the particular condition that required the presence of diabetes and the absence of statin treatment. Another major limitation of our study is the lack of prospective data. We didn’t aim, however, to investigate sortilin as a maker for increased risk, we tested it as a marker for active disease. A further limitation of our study is in that we chose to define PAD as a clinically apparent disease. In fact, to overcome this limitation a prospective study would be needed. We believe that our present result may lay the foundation for such a subsequent prospective study.

We demonstrated that circulating sortilin levels are associated with PAD in diabetic patients, that sortilin levels correlate with disease severity and that, therefore, sortilin is a promising marker for PAD in diabetic patients. Furthermore, in addition to risk factor-based validated scores, sortilin may help clinicians to better stratify atherosclerotic risk in diabetic patients. Further studies to confirm this hypothesis are needed and warranted.