Neutrophil degranulation biomarkers characterize restrictive echocardiographic pattern with diastolic dysfunction in patients with diabetes

Abstract Objective To investigate the potential association between neutrophil degranulation and patterns of myocardial dysfunction in a cohort of patients with type 2 diabetes mellitus (T2DM). Background Two distinct phenotypes of diabetic cardiomyopathy have been described: a restrictive phenotype with diastolic dysfunction (restrictive/DD) and a dilative phenotype with systolic dysfunction (dilative/SD). However, the underlying determinants of these two patterns are not yet recognized. Methods In this single‐centre, observational, cross‐sectional study, 492 patients were recruited. Ultrasonographic measurements were performed by two experienced sonographers, blinded to the clinical data of the participants. Serum biomarkers of neutrophil degranulation were measured by enzyme‐linked immunosorbent sandwich assay (ELISA). Results After adjustment for confounders, resistin, myeloperoxidase, matrix metalloproteinase 8 and matrix metalloproteinase 9/tissue inhibitor of metalloproteinases 1 complex were positively associated with the restrictive/DD pattern compared with the normal pattern. Similarly, MPO was positively associated with the dilative/SD pattern compared with the normal pattern, and resistin was negatively associated with the dilative/SD pattern compared with the restrictive/DD pattern. Conclusions Neutrophil degranulation is associated with the restrictive/DD echocardiographic pattern in patients with T2DM, but not with the normal pattern and dilative/SD patterns. Neutrophils could have a pivotal role in the pathogenesis of myocardial dysfunction, and particularly diastolic dysfunction, in patients with T2DM.


| INTRODUCTION
Heart failure is a common comorbidity in patients with diabetes mellitus (DM), with heavy drawbacks on their quality of life and life expectancy, and consequent relevant social costs. 1 DM, and particularly type 2 DM (T2DM), is frequently associated with hypertension, obesity, lipid metabolism disorders and coronary artery disease; these factors could mediate the association between DM and heart failure. However, accumulating evidence highlights an additional disease-specific role of DM on myocardial remodelling and dysfunction, leading to the definition of diabetic cardiomyopathy as a primary myocardial disease. 2,3 Two distinct phenotypes of diabetic cardiomyopathy have been described: a restrictive phenotype with diastolic dysfunction (restrictive/DD) and a dilative phenotype with systolic dysfunction (dilative/SD). These two phenotypes-once considered consecutive stages of the same condition-are now rather considered as independent entities with different pathogeneses and divergent natural course. 4 The restrictive/DD pattern is characterized by a normal volume of the left ventricle, with a thickened and stiff wall. Enlarged cardiomyocytes with a preserved sarcomeric ultrastructure and increased deposition of extracellular matrix are the hallmark of this phenotype. 5 Conversely, the dilative/ SD pattern is characterized by an increased volume of the left ventricle. Cardiomyocytes are damaged with subversion of the sarcomeric structure and associated diffuse fibrosis. Both phenotypes are characterized by microvascular rarefaction and perivascular deposition of advanced glycation end products, more predominant in the dilative/SD phenotype. 4,6 Less is known about the role of innate immunity, and especially neutrophils, in the pathogenesis of diabetic cardiomyopathy and its subtypes. 7 The aim of this study was to investigate the potential association between neutrophil degranulation-evaluated through a panel of validated serum biomarkers-and echocardiographic patterns of myocardial dysfunction in a cohort of patients with T2DM.

| Enrolled subjects and study design
The present study has a single-centre, observational, crosssectional design enrolling Caucasian subjects participating in the CATAnzaro MEtabolic RIsk factors (CATAMERI) study. 8 Inclusion criteria were an established diagnosis of T2DM according to the American Diabetes Association diagnostic criteria 9 and age >18 years. Exclusion criteria were as follows: previous myocardial infarction or previous coronary revascularization; known coronary artery disease, valvular abnormality or hypertensive heart disease; and the presence of end-stage renal disease, chronic gastrointestinal diseases associated with malabsorption, chronic pancreatitis, history of any malignant disease, self-reporting alcohol consumption >20 g/d, positivity for antibodies to hepatitis C virus or hepatitis B surface antigen. The protocol (approval code: 2012.63) was approved by the Local Ethics Committee of Catanzaro (Comitato Etico Azienda Ospedaliera 'Mater Domini'), and written informed consent was obtained from each subject before commencing the study, in accordance with the principles of the Declaration of Helsinki.

| History, clinical parameters and anthropometric measures
Personal history, family history and current pharmacological treatments were collected using a standardized questionnaire. The anthropometric assessment was illustrated in the Supplementary Material.

| Endpoint adjudication and power study calculation
The primary endpoint of the study was the detection of whether surrogate biomarkers of neutrophil degranulation correlate with at least two echocardiographic patterns of myocardial dysfunction. The secondary endpoint was a significant correlation between the over-mentioned biomarkers and the ultrasonographic parameters of myocardial dysfunction. The adequacy of sample size was assessed assuming a 60% prevalence of subclinical myocardial dysfunction in patients with T2DM, with an equal prevalence of dilative/ SD and restrictive/DD pattern. 22 Based on previous studies in patients with heart failure, the minimum clinically relevant difference for resistin was calculated as 0.93 ng/mL, with a SD for difference of 1.64 ng/mL. 23 As a result, a sample size of 243 patients is considered sufficient to reach the desired power of 0.95.

| Statistical analysis
Continuous variables are expressed as median and interquartile range (IQR) or mean ± standard deviation (SD) according to their distribution assessed with the Kolmogorov-Smirnov test. Categorical variables are instead presented as number (%). A probability of type 1 error <0.05 is considered significant under the assumption of null hypothesis. Overall significance of differences between groups was assessed by the Kruskal-Wallis test and the chi-square test for continuous and categorical parameters, respectively. The significance of pairwise differences was assessed through the Mann-Whitney U test or univariate logistic regression, after log transformation of non-normally distributed variables. The correction for potential confounders was performed using multiple logistic regression analysis. The analysis was performed using the SPSS software package 23.0 (IBM Inc.).

| Clinical and biochemical variables and echocardiographic assessment of the study cohort
Demographic, physical and laboratory characteristics of the enrolled population are reported in Table 1. The enrolled population comprised 492 subjects, with a prevalence of male subjects (64.0%) in their late adulthood (median age: 61 years), who received a relatively recent diagnosis of T2DM (median duration of disease: 4 years). Most of the female subjects were in post-menopausal age (64%). Enrolled patients were overall obese (median BMI: 30.5 kg/m 2 ) with a good glycometabolic control (median HbA1c: 6.9%). As reported in Table S1, the majority of patients had associated hypertension (88.1%) and/or metabolic syndrome (77.7%). Three hundred and sixty subjects were on pharmacological treatment for hypertension, with inhibitors of renin-angiotensinaldosterone system being the most prescribed class of drugs (308 subjects). Most of the patients showed a low-grade inflammatory profile (median CRP level: 3.2 mg/L).
Complete echocardiographic parameters were available for 451 patients. Median values and IQR of parameters  analysed in the present study are reported in Table S2. LV diastolic dysfunction was diagnosed in 63.8% of subjects, and LV systolic dysfunction, in 30.6% ( Figure 1A); and LV hypertrophy was diagnosed in 52.3% of subjects, and LV dilation, in 35.7% ( Figure 1B). A restrictive pattern with diastolic dysfunction was observed in 177 subjects (35.8%), whereas a dilative pattern with systolic dysfunction was observed in 163 subjects (36.1%) ( Figure 1C).

| Neutrophil-related inflammatory biomarkers correlate with parameters of left ventricular dysfunction
As expected, inflammatory parameters belonging to neutrophil degranulation were positively correlated with neutrophils and white blood cell counts, with CRP levels and with each other (Table S3) It is noteworthy that CRP also positively correlated with body weight, the presence of metabolic syndrome, the markers of insulin resistance and LDL-cholesterol (Table S4).
As regards the parameters of LV dysfunction, serum levels of CRP, resistin, MPO, MMP-8 and MMP-9 and TIMP-1 positively correlated with the E/e' ratio; among them, resistin, MPO and TIMP-1 also positively correlated with the degree of diastolic dysfunction. MPO and TIMP-1 were also negatively correlated with the LV s' ( Table 2). As regards the parameters of LV remodelling, resistin positively correlated with interventricular septum (IVS) thickness and negatively correlated with the LV diameters and volumes. Similarly, MMP-9 positively correlated with the LV posterior wall (LVPW) thickness and negatively correlated with the LVEDVi (Table 2).

| Neutrophil-related inflammatory biomarkers are independently associated with patterns of myocardial dysfunction
Comparing the average values of neutrophil degranulation biomarkers among echocardiographic patterns (normal, restrictive with diastolic dysfunction and dilative with systolic dysfunction), a significant difference between normal and restrictive patterns was observed for resistin, MPO, MMP-8, MMP-9/TIMP-1 complex, TIMP-1 and TIMP-2 ( Figure 2). A significant difference in MPO values was also observed between normal and dilative patterns, whereas serum resistin levels significantly differed in restrictive and dilative patterns. Conversely, CRP did not differ across the different patterns of myocardial dysfunction. When compared to clinical, anthropometric and biochemical parameters, significant differences among the echocardiographic patterns were observed (Table S5). Those variables were then considered as potential confounders and included in logistic regression models ( Figure S1 and Table S6). After adjustments, resistin, MPO, MMP-8 and MMP-9/TIMP-1 complex confirmed their independent association with the restrictive pattern compared with the normal pattern (Figure 3 and Tables S6 and S7). Similarly, MPO was positively associated with the dilative pattern compared with the normal pattern, and resistin was negatively associated with the dilative pattern compared with the restrictive pattern (Figure 3 and Tables S8 and S9).

| DISCUSSION
The main finding of the present study is the association between neutrophil degranulation, evaluated through surrogate serum biomarkers, and the restrictive/DD pattern of myocardial dysfunction, evaluated through echocardiography, in patients with T2DM. This association is confirmed also after adjustment for potential confounders. Specifically, measured biomarkers are significantly higher in patients with restrictive/DD pattern than in those with normal pattern. Such difference has not been observed between patients with dilative/SD pattern and normal pattern, with the exception of MPO, showing significantly elevated values also in the dilative/SD pattern group. Notably, resistin values were significantly higher in the restrictive/DD pattern group than in the dilative/SD pattern group, whereas CRP failed to show any difference across the patterns of myocardial dysfunction. Overall, these findings might suggest some specific role for neutrophil degranulation in the pathophysiology of diabetic cardiomyopathy. This would not be surprising as the involvement of leucocyte in diabetic cardiomyopathy has been well established. 7 However, less is known about the role of neutrophils. 24,25 Actually, they are more prone to oxidative burst and secretion of extracellular traps, whereas phagocytic and chemotactic activity is reduced in patients with DM. [26][27][28] With regard to myocardial function, neutrophil degranulation has been associated with increased intraventricular diastolic pressure in patients with hypertensive heart disease 29 and severe chronic kidney disease, 30 but not yet in diabetic patients. Well established is instead the association between MMPs and LVDD in animal models. 31 Equally interesting is the evidence about the potential involvement of resistin in diabetic cardiomyopathy. Traditionally known as is cysteine-rich adipokine with downregulating effect of insulin signalling, 32 resistin is highly expressed in mononuclear cells and stored in the primary granules of neutrophils. 33 It has been supposed that resistin might sustain the chronic low-grade inflammation characterizing the metabolic disorders, 34 but clear evidence is lacking. 17,35 Indeed, only one single study has so far reported a positive association between resistin and myocardial diastolic dysfunction. 36 Instead, why MPO, MMP-8, TIMP-1, -2, and MMP-9/ TIMP-1 as well are significantly associated with restrictive/ DD, whereas only MPO is associated with dilative/SD still remains a controversial point. Although speculative, the development of restrictive/DD pattern can be assumed as a more complex process involving the deposition of extracellular matrix. This could be linked to a different neutrophil activation/degranulation pattern. In particular, MMP-8, and MMP-9-when balanced with TIMPs-exert pro-fibrotic effects. 37 Unfortunately, the study design of the primary cohort did not plan to collect neutrophils and we cannot perform any ex vivo experiment on neutrophil function. This should be considered a leading shortcoming of the present study. Beyond neutrophil degranulation biomarkers, we should also acknowledge that other determinants of echocardiographic patterns should be accounted for such as demographic, clinical and biochemical factors; namely, the presence of any abnormal pattern (both restrictive/DD and dilative/SD) was characterized by older age and the presence of hypertension. 1 The restrictive/ DD pattern was further characterized by the female sex, whereas the dilative/SD pattern was characterized by abdominal obesity, active smoking, higher severity of hypertension (estimated through the number of anti-hypertensive drugs), dyslipidaemia and worse glycometabolic control. Since the aim of the study was to investigate a correlation between serum biomarkers of neutrophil degranulation and myocardial function in patients with T2DM, subjects were not screened for the presence of coronary artery disease before enrolment, and neither were they retrospectively excluded on the basis of diagnostic tests performed afterwards. This could represent a limitation when attempting to implementing these results in the setting of diabetic cardiomyopathy. Furthermore, the definition of diastolic dysfunction was simplified in order to ease data collection and analysis, since the study does not aim at establishing an accurate diagnosis, rather detecting an association between a panel of biomarkers and an echocardiographic pattern. Notwithstanding this limitation, the prevalence of diastolic dysfunction reported in this paper is consistent with the one observed in previous studies. 22 As an additional limitation, we should acknowledge that this study is composed of patients with a relatively recent diagnosis of diabetes. This may somewhat limit the generalization of results, but it may also point out how neutrophil degranulation contributes earlier to the development of diabetic cardiomyopathy.
Finally, the prevalence of hypertension, obesity and metabolic syndrome was quite homogeneous among the three groups. Although model adjustments for these potential confounders were adopted, a relevant contribution of these comorbidities to myocardial dysfunction cannot be excluded. In particular, the adjustment for hypertension only concerns the presence of disease and the number of anti-hypertensive drugs, but not the duration of hypertension. The latter is often difficult to estimate as it often occurs asymptomatically but undoubtedly represents a missing confounder that should be acknowledged as a further study limitation.
In conclusion, the results of the present study suggest that neutrophil degranulation characterizes the restrictive/ DD echocardiographic pattern in patients with T2DM, compared with the normal pattern and the dilative/SD pattern as well. This result is consistent with previous observations in animal models but not in human beings. Future studies are needed to assess whether serum biomarkers could help in stratifying the risk of myocardial dysfunction in subjects with T2DM and normal myocardial function, as well as investigating the potential pharmacological strategies targeting neutrophil activity (eg inhibitors of MMPs) to prevent the progression of myocardial dysfunction to overt heart failure.