Study participant characteristics

The median age of 32 study participants was 24 (range 18–34) years (59% male) with a mean ± SD duration of type 1 diabetes of 14.0 ± 6.3 years. Participants had a mean HbA1c of 65.3 ± 14.0 mmol/mol (8.1%) with a mean eGDR of 8.1 ± 2.1 mg kg−1 min−1. No participants had macrovascular complications and just two were on additional non-insulin therapies (both metformin) (Table 1).

Table 1 Summary of the baseline characteristics of all study individuals with type 1 diabetes

When stratifying participants according to their eGDR (Table 2), there were no statistically significant differences between groups in relation to sex, age, duration of diabetes or presence of retinopathy.

Table 2 Summary of the baseline characteristics of all study individuals with type 1 diabetes stratified according to eGDR Platelet activation in type 1 diabetes and healthy control participants

Under basal conditions, platelets from individuals with type 1 diabetes expressed significantly greater levels of CD62P compared with control participants (521 ± 246 vs 335 ± 67; p<0.001), with similar findings for PAC-1 (370 ± 165 vs 231 ± 88; p=0.011) and PS (869 ± 762 vs 294 ± 109; p=0.001; Fig. 1a–c, electronic supplementary material [ESM] Fig. 1). Treatment of blood with either SFLLRN (a thrombin mimetic) or CRP-XL (a collagen mimetic) alone or in combination led to increased expression of all activation markers in both groups. Examination of CD62P showed no clear difference in expression in the type 1 diabetes group and control participants. In contrast, PAC-1 binding was elevated in the type 1 diabetes cohort compared with control participants when stimulated with single agonists (for example, at 2 μmol/l SFLLRN MFI was 5583 ± 4960 in type 1 diabetes vs 2155 ± 487 in healthy control participants, p=0.001 and at 10 μg/ml CRP-XL MFI was 11,335 ± 5124 vs 8419 ± 1942, p=0.03) or a combination of agonists (11,375 ± 6689 vs 7861 ± 1458, p=0.01), as shown in Fig. 1b. We also observed a greater propensity for PS exposure in type 1 diabetes platelets compared with healthy control participants following stimulation, with CRP-XL alone (at 1 μg/ml CRP-XL 4356 ± 4719 vs 1165 ± 308, p=0.004 and at 10 μg/ml CRP-XL 10,561 ± 6181 vs 5259 ± 1125, p<0.001) or in combination with SFLLLRN (27,124 ± 10,105 vs 19,801 ± 1455, p=0.001), which are known to induce PS exposure. Consistent with published studies, we found stimulation of the thrombin activation pathway alone is insufficient to stimulate PS [26, 27, 32]. Taken together, these data demonstrate that type 1 diabetes platelets are partially activated under basal conditions and that these same platelets exhibit a greater sensitivity to activation when challenged with agonists.

Fig. 1

Multiparameter fluorescence flow cytometry to investigate platelet levels of (a) CD62P, (b) PAC-1 and (c) PS exposure in whole blood from study participants with type 1 diabetes (n=26–32) compared with healthy control participants (n=5–6). Expression is measured as MFI both at basal conditions and in response to stimulation with low and high dose single agonists (SFLLRN or CRP-XL) and high dose dual agonists (20 μmol/l SFLLRN and 10 μg/ml CRP-XL combined). *p≤0.05, **p≤0.01, ***p≤0.001. For comparison between two groups, unpaired t test or Mann–Whitney U tests have been used depending on distribution of data. Between multiple groups, ordinary ANOVA tests were carried out

Platelet activation in individuals with type 1 diabetes stratified by insulin resistance

Having observed platelet hyperactivity in individuals with type 1 diabetes compared with healthy control participants, we next examined the potential role of insulin resistance. The type 1 diabetes cohort was stratified according to eGDR, dividing into three groups: eGDR<6 mg kg−1 min−1, eGDR 6–8 mg kg−1 min−1 and eGDR>8 mg kg−1 min−1, to represent advanced, mild and normal insulin sensitivity, as per Nyström et al [18].

Under basal conditions, expression of all three activation markers was elevated in those with advanced insulin resistance compared with the other two groups (Fig. 2, ESM Fig. 2). When blood was treated with SFLLRN, CD62P was found to be elevated in individuals with advanced insulin resistance at 15,017 ± 5602 compared with mild insulin resistance (6304 ± 3478; p=0.01) or no insulin resistance (5226 ± 2565; p=0.007) (Fig. 2a). While we saw a similar pattern when examining PAC-1 binding, this only reached statistical significance at the higher concentration of SFLLRN, with advanced insulin resistance showing MFI of 19,339 ± 11,749 vs 7287 ± 2004 for mild insulin resistance (p=0.03; and 5187 ± 2872 for no insulin resistance [p=0.02], Fig. 2b). As expected, SFLLRN alone had no significant effect on PS (Fig. 2c).

Fig. 2

Multiparameter fluorescence flow cytometry to investigate platelet levels of (a) CD62P, (b) PAC-1 and (c) PS exposure in whole blood from study participants with type 1 diabetes (n=26–32) stratified according to eGDR threshold values derived from Nyström et al [18], eGDR<6 mg kg−1 min−1 (n=5–8), 6–8 mg kg−1 min−1 (n=6–9), >8 mg kg−1 min−1 (n=11–14). Expression is measured as MFI both at basal and in response to stimulation with low and high dose single agonists (SFLLRN or CRP-XL) and high dose dual agonists (20 μmol/l SFLLRN and 10 μg/ml CRP-XL combined). *p≤0.05, **p≤0.01, ***p≤0.001. For comparison between two groups, unpaired t test or Mann–Whitney U tests have been used depending on distribution of data. Between multiple groups, ordinary ANOVA tests were carried out

To determine whether platelet hyperactivity in the advanced insulin resistance group was agonist-specific, we next tested the effect of CRP-XL. We observed elevated CD62P expression in the advanced insulin resistance group compared with those with normal insulin sensitivity in response to the higher concentration of CRP-XL (29,167 ± 2177 vs 22,829 ± 2535, p<0.001), while differences with PAC-1 and PS failed to reach statistical significance (Fig. 2b, c).

Using dual stimulation, those with advanced insulin resistance demonstrated elevated levels of CD62P, PAC-1 and PS compared with the other type 1 diabetes groups (Fig. 2).

Platelet inhibition by PGI2 in individuals with type 1 diabetes and healthy control participants

Given previous work showing reduced platelet sensitivity to NO and PGI2 in type 2 diabetes [11], we speculated that platelet hyperactivity in type 1 diabetes may also be linked to disinhibition.

There was no difference in sensitivity to PGI2 between type 1 diabetes and healthy control participants for CD62P expression (Fig. 3a–d). However, PAC-1 inhibition by the higher dose PGI2 was diminished in the type 1 diabetes group compared with control participants following stimulation with 10 μg/ml CRP-XL (reduction of 79 ± 18% vs 94 ± 3%, p=0.01) and the SFLLRN/CRP-XL combination at both inhibitor doses (3 ± 14% vs 16 ± 7%, p=0.007 at 10 nmol/l PGI2 and 47 ± 29% vs 84 ± 8%, p<0.001 at 100 nmol/l PGI2; Fig. 3h). When examining PS exposure, we again observed hyposensitivity to PGI2 inhibition in the type 1 diabetes group compared with control participants when platelets were stimulated with CRP-XL or a combination of SFLLRN/CRP-XL, which was only observed with the higher concentration of the inhibitor (Fig. 3i). Taken together, these data suggest hyposensitivity of type 1 diabetes platelets to the antithrombotic actions of PGI2.

Fig. 3

Multiparameter fluorescence flow cytometry to investigate platelet expression of (a) CD62P and (e) PAC-1 and (i) PS exposure in whole blood from study participants with type 1 diabetes (n=26–32) compared with healthy control participants (n=5–6). Expression is measured as MFI both at basal and in response to stimulation with high dose single agonists (SFLLRN or CRP-XL) and high dose dual agonists (20 μmol/l SFLLRN and 10 μg/ml CRP-XL combined) as well as in response to inhibition with PGI2 at low (10 nmol/l) and high (100 nmol/l) doses. (b–d, f–h, j–l) Response has been measured as percentage reduction compared with expression following stimulation at each of the agonist doses. *p≤0.05, **p≤0.01, ***p≤0.001. For comparison between two groups, unpaired t test or Mann–Whitney U tests have been used depending on distribution of data. Between multiple groups, ordinary ANOVA tests were carried out

Platelet inhibition in individuals with type 1 diabetes stratified by insulin resistance

Insulin resistance influenced platelet inhibition by PGI2, with less inhibition of CD62P observed in those with advanced insulin resistance (Fig. 4a) and with larger differences detected with PAC-1 and PS exposure (Fig. 4d–i). With 1μg/ml CRP-XL alone, the inhibition of PAC-1 by 10 nmol/l PGI2 in those with advanced insulin resistance was significantly reduced compared with those with normal insulin sensitivity (28 ± 9% vs 52 ± 2% p<0.001), which was also evident with higher PGI2 concentration (Fig. 4e). Following dual agonist stimulation, the higher concentration of PGI2 (100 nmol/l) showed diminished inhibition in the advanced insulin resistance group compared with the normal insulin sensitivity group (Fig. 4f). Similarly, the inhibition of PS exposure by low and high PGI2 concentrations was diminished in those with advanced insulin resistance compared with the other two groups (Fig. 4g–i). These data suggest that insulin resistance in type 1 diabetes is associated with a reduction in platelet sensitivity to the key endogenous inhibitor PGI2.

Fig. 4

Multiparameter fluorescence flow cytometry to investigate platelet expression of (a) CD62P and (d) PAC-1 and (g) PS exposure in whole blood from study participants with type 1 diabetes (n=26–32), stratified according to eGDR threshold values derived from Nyström et al [18]. Expression is measured as MFI both at basal and in response to stimulation with single agonist (CRP-XL) and high dose dual agonists (20 μmol/l SFLLRN and 10 μg/ml CRP-XL combined) as well as in response to inhibition with PGI2 at low (10 nmol/l) and high (100 nmol/l) doses. (b, c, e, f, h, i) Response has been measured as percentage reduction compared with expression following stimulation at each of the agonist doses. *p≤0.05, **p≤0.01, ***p≤0.001. For comparison between two groups, unpaired t test or Mann–Whitney U tests have been used depending on distribution of data. Between multiple groups, ordinary ANOVA tests were carried out

Insulin resistance changes platelet subpopulation dynamics

The basis of functional platelet heterogeneity may lay in distinct receptor expression in response to physiological or pathophysiological mediators, along with size and sensitivity to activation [27]. Little is known regarding platelet subpopulations in individuals with type 1 diabetes. To address this, FAUST [31] was applied to flow cytometry data from unstimulated and dual agonist-stimulated platelets in the absence and presence of PGI2 from individuals with advanced insulin resistance or normal insulin sensitivity.

A total of eight platelet subpopulations (P1–P8) were detected in both cohorts, but critically with differing distribution (Fig. 5). These subpopulations were characterised by differential levels of CD62P, PAC-1 and PS, and they consisted of platelets with only activated αIIbβ3, CD62P−PAC-1+PS− (P1); platelets with α-granule secretion and activated αIIbβ3, CD62P+PAC-1+PS− (P2); platelets with only α-granule secretion, CD62P+PAC-1−PS− (P3); platelets with activated αIIbβ3 and PS exposure, CD62P−PAC-1+PS+ (P4); platelets with α-granule secretion, activated αIIbβ3 and PS exposure, CD62P+PAC-1+PS+ (P5); platelets with α-granule secretion and PS exposure, CD62P+PAC-1−PS+ (P6); platelets with only PS exposure, CD62P−PAC-1−PS+ (P7); and resting platelets, CD62P−PAC-1−PS− (P8).

Fig. 5

Platelet subpopulations. Whole blood was unstimulated (basal) or stimulated with SFLLRN and CRP-XL in the presence or absence of PGI2 (100 nmol/l) for 20 min prior to fixation. Samples were then analysed by flow cytometry where CD62P and PAC-1 expression and PS exposure were quantified. These data underwent platelet subpopulation analysis using FAUST. This discovered eight (P1–P8) platelet subpopulations present at basal, upon activation with SFLLRN and CRP-XL and in the presence of PGI2. These subpopulations were defined by differential CD62P, PAC-1 and AnnV binding. Platelet subpopulations are visualised on uniform manifold approximation and projection (UMAP) graphs from participants with (a) high eGDR (n=4) and (b) low eGDR (n=3). (c–e) Difference in platelet subpopulation abundance, presented in percentage (subpopulations, P1–P8), at basal (c), when dual agonist stimulated (d) and when dual agonist stimulated in presence of higher dose inhibition (100 nmol/l PGI2) (e). Data are expressed as mean ± SD. *p≤0.05, **p≤0.01, ***p≤0.001. For comparison between two groups, unpaired t test or Mann–Whitney U tests have been used depending on distribution of data. Between multiple groups, ordinary ANOVA tests were carried out

Under basal conditions, the most abundant subpopulation was P8 for both normal insulin sensitivity and advanced insulin resistance groups (94.6 ± 1.5 and 79.2 ± 1.7, respectively), showing that most platelets in the total population are quiescent. However, P8 abundance was significantly lower in those with advanced insulin resistance (p<0.001), accompanied by a significant increase in P3 abundance (p<0.001). These data suggest that individuals with type 1 diabetes and advanced insulin resistance have a specific subset of activated circulating platelets under basal conditions, expressing elevated CD62P. Dual stimulation of platelets led to a remodelling of platelet subsets, with platelets moving from P8 to P2 (CD62P+PAC-1+PS−), P5 (CD62P+PAC-1+PS+) and P6 (CD62P+PAC-1−PS+), but we found no significant differences between groups. After treatment with PGI2, participants with advanced insulin resistance demonstrated a different inhibitory profile. Participants with advanced insulin resistance had significantly higher levels of P2 platelets (p=0.03) and significantly lower levels of P3 (p=0.006) when compared with those with normal insulin sensitivity. This suggests that those with advanced insulin resistance have increased CD62P expression and αIIbβ3 activation in the presence of PGI2, implying that PGI2 is less effective at inhibiting the activation of these markers in individuals with advanced insulin resistance.