The aim of the present study was to investigate the role of GV (glycaemic variability) in diabetic vascular complications and to explore the molecular pathways modulated by glycaemic ‘swings’. We developed a murine model. A total of 30 diabetic mice received once daily basal insulin administration plus two oral boluses of glucose solution (GV group, named ‘V’) and 30 diabetic mice received once daily basal insulin plus two oral boluses of saline solution (stable hyperglycaemia group, named ‘S’) for a period of 30 days. Glycaemia was measured eight times daily to detect GV. Finally, postischaemic vascularization, induced by hindlimb ischaemia 30 days after diabetes onset, was evaluated. We found that GV was significantly different between S and V groups, whereas no significant difference in the mean glycaemic values was detected. Laser Doppler perfusion imaging and histological analyses revealed that the ischaemia-induced angiogenesis was significantly impaired in V mice compared with S group, after ischaemic injury. In addition, immunostaining and Western blot analyses revealed that impaired angiogenic response in V mice occurred in association with reduced VEGF (vascular endothelial growth factor) production and decreased eNOS (endothelial nitric oxide synthase) and Akt (also called protein kinase B) phosphorylation. In conclusion, we describe a murine model of GV. GV causes an impairment of ischaemia-induced angiogenesis in diabetes, likely to be independent of changes in average blood glucose levels, and this impaired collateral vessel formation is associated with an alteration of the VEGF pathway.
Cardiovascular complications are mainly responsible for the high morbidity and mortality in people with DM (diabetes mellitus). Indeed, diabetic subjects have a 3–5-fold higher risk of cardiac ischaemic disease death [1,2] and a 2–5-fold higher risk of stroke and peripheral arterial disease than non-diabetic subjects . Furthermore, long-term prognosis after a coronary event is significantly worse among people with DM than among those without DM . Usually, the risk factors for cardiovascular complications in DM are classified into glycaemic and non-glycaemic. Non-glycaemic risk factors include hypertension, dyslipidaemia, inflammation, microalbuminuria, abdominal obesity and smoking, whereas FPG (fasting plasma glucose), PPG (postprandial glucose), GV (glycaemic variability) and HbA1c (glycated haemoglobin; an index of glycaemic control of the last 3 or 4 months) belong to the first group. Recently, several epidemiological studies performed in both Type 2 and Type 1 diabetes [5,6] have suggested that FPG , PPG  and GV  are HbA1c-independent risk factors for vascular complications, leading to the concept of ‘the glucose tetrad’ (HbA1c, FPG, PPG and GV) . In particular, the role of GV in diabetic complications was first hypothesized by the authors of the DCCT (Diabetes Control and Complications Trial), who surmised that the mean HbA1c is not the most complete expression of the degree of glycaemia and that the risk of complications may be mostly dependent on other factors; indeed, even when HbA1c values were comparable between intensively and conventionally treated subjects, the latter group experienced a markedly higher risk of progression to retinopathy . A possible explanation is that glycaemic excursions were of greater frequency and magnitude among conventionally treated patients (who received fewer insulin injections), generating more ROS (reactive oxygen species) in ECs (endothelial cells). Subsequent in vitro and in vivo studies have confirmed this hypothesis [5,11–13]; however, to date, the cause–effect relationship between GV and in vivo diabetic vascular disease is still unclear.
Several of the long-term complications of diabetic vasculopathy are associated with aberrant angiogenesis. In fact, excessive angiogenesis plays a role in diabetic retinopathy, nephropathy and neuropathy, whereas inhibited angiogenesis contributes to impaired wound healing, and impaired coronary and peripheral artery diseases . After vascular damage, postischaemic angiogenesis is an important reparative mechanism and can ameliorate the outcome of diabetic vascular pathology. Impaired collateral vessel formation in response to ischaemic injury has been demonstrated in murine models of DM: hindlimb ischaemia created by ligation of the femoral artery has been associated with reduced formation of capillaries and a reduction in blood flow to the ischaemic limb in diabetic mice compared with non-diabetic mice . Nevertheless, previous studies have focused mostly on chronic hyperglycaemia consequences, and there is a lack of evidence on the importance of GV-related effects. Therefore the aim of the present study was to demonstrate a cause–effect relationship between GV and vascular complications, evaluated as impaired response to hindlimb ischaemia, and to analyse whether the severity of vascular dysfunction was higher with GV than with stable hyperglycaemia. Furthermore, we focused our attention on the VEGF (vascular endothelial growth factor)/Akt (also known as protein kinase B)/eNOS (endothelial nitric oxide synthase) pathway, which has been implicated in vascular alterations due to chronic hyperglycaemia .
MATERIALS AND METHODS
Mouse model of GV
The investigation was approved by A. Gemelli University Hospital's Institutional Animal Care and Use Committee. Male 8–12-week-old C57BL/6J mice were used for the experiments. All animals were allowed free access to the same standardized diet and water during the entire study. Diabetes was induced by administering streptozotocin (50 mg/kg body weight) (Sigma) in citrate buffer (pH 4.5) intraperitoneally during the fasting state and consecutively for 5 days, as described previously . Hyperglycaemia was verified on blood obtained from the tail vein by an Accu-Check Active glucometer (Roche) 2 days after streptozotocin injections. We considered mice to be diabetic when blood glucose was at least 16 mmol/l (normal, 5–8 mmol/l). Overall, 60 mice showed a blood glucose level of at least 16 mmol/l, both 1 and 2 weeks after streptozotocin injections, and were therefore included in the experimental group. Then, the diabetic animals were divided into two groups (30 mice each): the first group (GV group, named ‘V’) was treated with a subcutaneous injection of 0.5 unit of long-acting insulin glargine at 08.00 hours, plus two boluses of 0.7 ml of 33% glucose solution (at 10.00 hours and 14.00 hours) by gavage daily for 30 days; the second group (stable hyperglycaemia group, named ‘S’) received 0.5 unit of long-acting insulin glargine at 08.00 hours, plus two boluses of 0.7 ml of saline solution (at 10.00 hours and 14.00 hours) by gavage daily for 30 days (Figure 1). During the night, both groups received no intervention and lived in the same environmental conditions. For 30 days, the glycaemic values of each mouse were measured in tail blood by the glucometer at 08.00, 10.00, 12.00, 14.00, 16.00, 18.00, 20.00 and 22.00 hours (Figure 1). After the treatment and until the end of the study (subsequent 28 days; see below), both the V and S groups received a subcutaneous injection of 0.5 unit of glargine at 08.00 hours plus two boluses of 0.7 ml of saline solution (at 10.00 and 14.00 hours); the glycaemic profiles were measured as described previously. GV was measured by three well-known indices, namely S.D., MAGE (mean amplitude of glycaemic excursions; the mean of glycaemic excursions >1 S.D.) , and CONGA (continuous overlapping net glycaemic action; S.D. of differences between any glucose value and another one exactly n hours later) . Finally, 30 untreated normoglycaemic C57BL/6J mice (control group, named ‘C’) were also included in the model and glycaemic profiles were measured for 30 days as in V and S groups.
Schematic representation of the GV experimental protocol (a) and the mouse hindlimb ischaemia model protocol (b)
Mouse hindlimb ischaemia model
After 30 days, unilateral hindlimb ischaemia was induced in C, V and S mice (30 mice each), as described previously . Briefly, all animals were anaesthetized with an intraperitoneal injection of ketamine (60 mg/kg of body weight) and xylazine (8 mg/kg of body weight). The proximal and distal portions of the femoral artery and the distal portion of the saphenous artery were ligated. The arteries and all side branches were dissected free and excised. The skin was closed with 5–0 surgical sutures. A laser Doppler perfusion imager system (PeriScan PIM II; Perimed) was used to measure hindlimb blood perfusion before and immediately after surgery and then weekly, until the end of the study, for a total follow-up of 28 days after surgery. Before imaging, excess hair was removed from the limbs using depilatory cream and mice were placed on a heating plate at 40°C. To avoid the influence of ambient light and temperature, results were expressed as the ratio between perfusion in the right (ischaemic) to that in the left (non-ischaemic) limb.
At 1 and 4 weeks after surgery, mice were killed by intraperitoneal injection of an overdose of pentobarbital. The whole limbs were fixed in methanol overnight. The femora were carefully removed, and the ischaemic thigh muscles were embedded in paraffin. All the specimens were routinely fixed overnight in 4% buffered formalin and embedded in paraffin. Then, 4 μm sections of tissue samples were subjected to immunoperoxidase biotin–avidin reaction in the LSAB (labelled streptavidin–biotin) method to determine the CD31 and VEGF expression. CD45 was used as a marker for inflammatory infiltrate. Ki67 was used as a marker for proliferation. The cytoplasmic expression of Ki67, a nuclear protein expressed in the cell cycle, may be considered a functional phenomenon that is shared by normal tissues undergoing postnatal remodelling . The sections for immunohistochemistry were cut and mounted on 3-aminopropyltriethoxy-silane-coated (Sigma) slides, allowed to dry overnight at 37°C to ensure optimal adhesion, dewaxed, rehydrated and treated with 0.3% H2O2 in methanol for 10 min to block endogenous peroxidase. For antigen retrieval (not necessary for VEGF) the sections were microwave treated in 1 mM EDTA at pH 8 (for CD31) and pH 6 (for CD45 and Ki67) for 10 min and allowed to cool for 20 min. Endogenous biotin was saturated using a biotin blocking kit (Vector Laboratories). The sections were incubated at room temperature for 30 min with the following antibodies: purified rat anti-mouse CD31 [dilution 1:30; monoclonal (IgG2a); BD Bioscience], rabbit anti-mouse VEGF (dilution 1:100, polyclonal; Santa Cruz Biotechnology) and rabbit anti-mouse CD45 (dilution 1:50, polyclonal; AbCam), and incubated overnight with rabbit anti-mouse Ki67 [dilution 1:200; monoclonal (SP6); Novus Biological]. Binding was visualized using biotinylated secondary antibody (1 h of incubation) and the streptavidin–biotin peroxidase complex developed with diaminobenzidine. Finally, slides were counterstained with haematoxylin. Capillary density, proliferating cells and leucocyte infiltration were measured by counting six random high-power (magnification ×200) fields for a minimum of 200 fibres from each ischaemic and non-ischaemic limb on an inverted light microscope, and were expressed by the number of CD31+, Ki67+ or CD45+ cells/mm2. VEGF-positive cells were counted from four fields of view using a ×20 objective lens and the mean cell number/mm2 was established. Apoptosis was demonstrated in situ using a Mebstain Apoptosis kit II (Immunotech), and the apoptotic index was determined by dividing the total number of myocytes showing nuclear positivity by the total number of cells in the fields examined . Area was measured with an NIH image analysis system (ImageJ version 1.41). Two operators extracted the results independently.
Immunoblotting was performed on homogenates of muscle tissues. Protein concentration of samples was carefully determined by the protein assay (Bio-Rad Laboratories). Equal amounts of protein were subjected to SDS/PAGE using 4–12% gradient gels under reducing conditions (Bio-Rad Laboratories) and transferred to nitrocellulose membranes (GE Healthcare). To ensure equivalent protein loading and quantitative transfer efficiency of proteins, membranes were stained with Ponceau S before incubating with primary antibodies. Membranes were incubated with antibodies against VEGF (1:500; Santa Cruz Biotechnology), eNOS (1:1000; Santa Cruz Biotechnology), phospho-eNOS (Ser1177) (1:1000; Cell Signaling Technology), Akt (also called protein kinase B; 1:1000; Santa Cruz Biotechnology) and phospho-Akt (Ser473) (1:500; Cell Signaling Technology). Immunoreactive bands were detected by an ECL® kit (GE Healthcare). VEGF, eNOS, phospho-eNOS, Akt and phospho-Akt expressions were normalized using a mouse monoclonal anti-α-actin antibody. Densitometric analysis was performed and the results are expressed as a ratio compared with α-actin.
ELISA for VEGF
Ischaemic and control hindlimbs were harvested 7 days after surgery. VEGF levels were measured by ELISA (R&D Systems). The results were expressed as protein fold increase, calculated as the ratio between VEGF protein levels in ischaemic and control hindlimbs.
Statistical analysis was performed using STATA software (version 10.0). Results are expressed as means±S.E.M. Comparison of groups was made using ANOVA followed by Fisher's post hoc test. Repeated-measures ANOVA was performed to assess the improvement in perfusion over time within groups. Statistical significance was set at a probability value (P) of < 0.05.
A new model of GV
All the three considered indices of GV (MAGE, CONGA and S.D.) significantly differed between V, S and C groups (Figure 2 and Table 1); the difference in mgv (mean glucose value) between the V and S groups did not reach statistical significance (P=0.066), and there was no difference in mgv between S and C groups (P=0.396), whereas there was between V and C groups (P=0.001). Furthermore, there was no difference in glycaemic values at 20.00, 22.00, 08.00 and 10.00 hours among the three groups, whereas there was after the glucose gavage at 12.00, 14.00, 16.00 and 18.00 hours (results not shown); these results indicate that the difference in GV is a consequence of 12.00, 14.00, 16.00 and 18.00 hours glycaemic values, with a similar (not statistically different) glycaemic control between 20.00 and 10.00 hours (night profile). Overall, the results of our models clearly indicate that we can consider the V group as having a GV different from that of S and C groups, but similar mgv. This is of relevance because, to date, clinical studies on diabetes focused on the positive association between diabetic vascular complications and HbA1c, mathematically correlated to mgv [HbA1c=f(mgv)] . Therefore the glucose-related effects demonstrated in our model, in particular the angiogenic response to ischaemic injury between V and S groups, as well as S and C groups, are more likely to be a result of the effect of glucose oscillations. Furthermore, as insulin was administered in the V and S groups at the same units (0.5 unit), its effect on the angiogenic response to ischaemic injury is the same in the two groups. Of note, the 28 day analysis of the glycaemic data after surgery showed that there was no difference in GV indices and mgv between the V, S and C groups (results not shown).
Indices of GV and mean glycaemia derived from 7200 per-group glucose measurements (30 mice in each group)
|.||V compared .||S compared .||V compared .|
|Index .||with S .||with C .||with C .|
|.||V compared .||S compared .||V compared .|
|Index .||with S .||with C .||with C .|
Ischaemia-induced angiogenesis is significantly impaired in V mice compared with S mice
Laser Doppler perfusion imaging was performed before, immediately after and on days 7, 14, 21 and 28 after hindlimb ischaemia (Figure 3). Mean blood flow in untreated (C) mice reached 94% of the pre-ischaemic flow 28 days after hindlimb surgery. Perfusion recovery was significantly attenuated in S mice compared with C mice on postoperative days 7 (P=0.021), 14 (P=0.023), 21 (P=0.018) and 28 (P=0.032). Interestingly, the recovery was further impaired in V mice compared with S mice at 7 (P=0.023), 14 (P=0.026), 21 (P=0.008) and 28 (P=0.002) days after surgery. In addition, histological analysis revealed that the capillary density in ischaemic limb was significantly decreased in all diabetic mice compared with control mice and, importantly, there was a statistically significant difference between S and V mice (P=0.029; Figure 4).
Foot blood flow monitored
in vivo by laser Doppler perfusion imaging in C, S and V mice (ten mice in each group)
Representative photomicrographs of ischaemic muscle sections from C (a), S (b) and V (c) mice stained with an antibody directed against CD-31, 28 days after surgery (ten mice in each group)
To test whether the observed angiogenic responses were dependent on different proliferation, apoptosis or inflammation stimuli between the two diabetic groups or on an altered regulation of pro-angiogenic mechanisms, we analysed proliferation, apoptosis and leucocyte infiltration and noted that there was no difference between S and V mice according to all three aspects (Figure 5). Therefore it is possible to state that the observed difference in postischaemic vessel formation does not depend on a different proliferation, apoptotic or inflammatory response to hyperglycaemia between the two groups.
Representative photomicrographs of ischaemic muscle sections from S and V mice stained with an antibody directed against Ki67, with TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick-end labelling) and with an antibody directed against CD45, 7 days after surgery
Impaired angiogenic response in V mice occurs in association with reduced VEGF production and decreased eNOS and Akt phosphorylation
Immunostaining revealed that VEGF expression in the ischaemic tissue of S mice was reduced 7 days after surgery compared with control mice (Figures 6a and 6b) and that it was significantly impaired in V mice compared with S mice on postoperative day 7 (Figures 6b and 6c). Western blot analysis and VEGF-positive cells evaluation demonstrated that VEGF concentration in ischaemic tissue was also significantly higher in S mice than in V mice (Figures 6d and 6e), highlighting the crucial role of VEGF in impaired angiogenic response observed in V mice. To further investigate the mechanism by which the GV inhibits angiogenesis in diabetic mice, we evaluated eNOS and Akt phosphorylation in the ischaemic leg 7 days after surgery by Western blot analysis (Figure 7). GV reduced phosphorylation/activation of eNOS at Ser11777 and of Akt at Ser473. VEGF exerts many of its effects via the Akt pathway . The Akt pathway leads to downstream activation of eNOS and release of nitric oxide, which results in cGMP production. The phospho-Akt/total Akt ratio and the phospho-eNOS/total eNOS ratio (results not shown) were significantly lower in the V group compared with the S group, indicating that there is reduced downstream VEGF signalling.
Representative photomicrographs of ischaemic muscle sections from C (a), S (b) and V (c) mice stained with an antibody directed against VEGF, 7 days after surgery (ten mice each in group)
Representative Western blotting and relative attenuance evaluation of VEGF, eNOS, phospho-eNOS, Akt and phospho-Akt content in the ischaemic legs of C, S and V mice (ten mice in each group)
DM is a pathological condition of eminent epidemiological importance and several of the long-term diabetic complications are characterized by vasculopathy associated with aberrant angiogenesis. Experimental animal models of peripheral arterial disease in DM have shown attenuated perfusion recovery in response to ischaemia . A large number of hypotheses have been postulated to explain the impaired angiogenic postischaemic response in diabetes, such as the presence of vascular dysfunction characterized by both endothelial and vascular smooth muscle cell impairment , the decreased release or defective function of endothelial progenitor cells from the bone marrow , the exposure to chronic hyperglycaemia that leads to the non-enzymatic glycation of proteins and defective formation of new blood vessels , or the presence of abnormalities in growth factor signalling and/or expression, with maladaptive dysregulation of vascular growth factors pathways . But it is important to underline, once again, that previous studies on diabetic effects have focused principally on steady-state conditions of constant hyperglycaemia, and not on glycaemic control, glycaemic profiles and degree of GV. For this reason, we developed a mouse model of increased GV, to analyse how GV could contribute to ischaemia-induced vascularization in diabetic mice, and whose most relevant feature was the presence of a significant difference in GV comparing V group with the S group. In fact, in our model, an accurate (8 glycaemic values per day; for comparison, in the cornerstone study DCCT, GV was evaluated from 7 per day measurements ) and prolonged (58 days) evaluation of glycaemic profiles evidenced in V and S mice were different primarily between GV, and S and C. Therefore the differences found can be likely attributed primarily to GV. To our knowledge, this is the first animal model where GV is the primary difference and this represents the first innovative finding of this work. Furthermore, we used this model to analyse whether GV has more deleterious effects on vascular function than stable hyperglycaemia. In our experimental model, after hindlimb ischaemia, the presence of GV caused a significant impairment of postischaemic angiogenesis, compared with collateral vessels formation and blood flow recovery observed in untreated control mice and in stable diabetic mice. Moreover, V mice showed a reduced VEGF up-regulation and protein expression in ischaemic muscle after the ischaemic injury. The role of angiogenesis in diabetic CVDs (cardiovascular diseases) represents a major unresolved issue. Angiogenesis has attracted interest from opposite perspectives. Angiogenic cytokine therapy has been widely regarded as a fascinating approach both for treating ischaemic diseases and for enhancing atheroprotective functions of the endothelium ; conversely, several studies suggest that vascularization contributes to the development of atherosclerotic lesions and is a key factor for plaque destabilization, leading to plaque vulnerability and rupture . But it is evident that angiogenesis is an important defence mechanism from diabetic micro- and macro-vascular diseases that lead to the formation of new collateral vessels and consequent blood supply to peripheral ischaemic tissue. Previous studies have shown that large glycaemic swings exert deleterious effects in endothelial function and cardiovascular complication of DM  but until now there have been no data about the relationship between GV- and ischaemia-induced vascularizations.
To further investigate whether an alteration of the VEGF/Akt/eNOS pathway was involved in the postischaemic angiogenesis in our model, we evaluated eNOS and Akt activities, because Akt-dependent phosphorylation of eNOS at Ser11777 plays a key role in VEGF mediating vascular formation after ischaemic injury . Previously, a few studies have investigated modifications of the Akt and eNOS pathways in DM and, in fact, alterations in Akt or eNOS activity in diabetes were found in various cells and tissues, depending on experimental and clinical contexts. Defects in VEGF/Akt/eNOS activity have been reported in the endothelium of Type 2 DM models, possibly contributing to the development of endothelial dysfunction and loss of angiogenesis under these conditions . In contrast, Akt activity is increased in some tissues and vascular beds affected by complications of Type 1 DM . In the present study, we demonstrated that GV is associated with the phosphorylation/activation of eNOS and Akt in diabetic mice, indicating that there is impaired downstream VEGF signalling. Of note, as C mice did not receive insulin, the histological and functional differences between the C and S groups could be due to the administration of insulin; however, this hypothesis is not supported by evidence of the same differences between the S and V groups, both receiving the same type and dose of insulin (glargine, 0.5 unit).
These results are not in conflict with the existing evidence, as the studies that have analysed the angiogenic response to ischaemia in diabetic mice have been carried out without insulin treatment and in conditions of constant hyperglycaemia . These studies have documented a reduced ischaemia-induced vascularization but an increased up-regulation of VEGF, with a secondary functional alteration of this angiogenic factor. Our results underline that, when the DM is controlled with the administration of insulin, the postischaemic collateral vessel formation is reduced, with an impaired VEGF up-regulation and not significantly when compared with non-diabetic mice. In particular, our findings show that when significant blood glucose variability is induced in the same conditions of diabetes controlled by insulin treatment, the angiogenic response to ischaemic injury is further and significantly reduced with respect to non-diabetic and diabetic mice with chronic hyperglycaemia. Moreover, our data indicate that both VEGF up-regulation and eNOS and Akt phosphorylation/activation are also significantly reduced in these animals, suggesting a reduced downstream VEGF signalling in GV group mice.
Our experimental findings support the clinical evidence for the role of glycaemic control as a new potential risk factor for increasing CVD mortality and morbidity in DM . Indeed, because non-glycaemic risk factors are well recognized and treated in diabetic as well as non-diabetic subjects (often to lower targets in diabetics), something else should justify such a persistent excess of mortality due to diabetes. Obviously, the difference between diabetic and non-diabetic subjects is the degree of glycaemia, but, even after an intensive treatment of blood glucose (reduction of HbA1c/mean blood glucose), mortality remains high; hence, glycaemic risk factors other than HbA1c should be considered in the treatment of blood glucose disturbances in diabetes. Indeed, the glycaemic control (mgv) of the S group was not statistically significantly different from the C group. The mechanism through which GV induces loss of angiogenic properties after acute ischaemia could be related to an excess of oxidative stress in ECs . Oxidative stress is known to be the key mechanism in the pathogenesis of diabetes-related endothelial dysfunction. Oxidative stress is attributable to excessive production of ROS and the inactivation of nitric oxide by ROS is recognized to be a crucial factor in reducing nitric oxide bioavailability and the development of endothelial dysfunction . Interestingly, both chronic sustained hyperglycaemia and acute glycaemic fluctuations from peaks to nadirs, the two main determinants of GV, are conducive to the activation of oxidative stress with an overproduction of superoxide by the mitochondrial electron-transfer chain . Even if there are no data about the role of GV in the pathogenesis of diabetic ischaemic damage, our results suggest a possible correlation between the glucose fluctuations and angiogenic defects. This hypothesis deserves further evaluation and will represent the basis of our future studies.
In conclusion, this study designed and developed a murine model of GV. This will allow further studies to analyse the effect of GV on diabetic complications. Moreover, our experiments also showed that GV significantly reduces angiogenic reaction to peripheral ischaemic damage in diabetic mice, in the presence of similar average blood glucose levels. Finally, our data show that the reduced angiogenesis occurs in association with lowered VEGF up-regulation and with an impaired eNOS and Akt phosphorylation. These data provide support to the hypothesis that GV contributes to vascular complications of DM and control of GV plays a potentially important role in the management of diabetic patients.
This work was supported by the Catholic University School of Medicine.
We acknowledge the contribution of Dr Maria Emiliana Caristo, Director of Department of Animal House, Catholic University School of Medicine (Rome, Italy).
continuous overlapping net glycaemic action
endothelial nitric oxide synthase
fasting plasma glucose
mean amplitude of glycaemic excursions
mean glucose value
reactive oxygen species
vascular endothelial growth factor
Federico Biscetti, Dario Pitocco, Giuseppe Straface and Francesco Zaccardi participated in the design of the study, developed the hindlimb ischaemia model, performed data analysis and reviewed the paper prior to submission. Paola Rizzo, Raimondo De Cristofaro, Stefano Lancellotti and Tittania Musella carried out the immunoassays. Vincenzo Arena and Egidio Stigliano performed the immunohistochemical analysis. Giovanni Ghirlanda and Andrea Flex conceived the study, participated in its design and co-ordination, and helped in drafting the paper.
These authors contributed equally to the study.