Fasting pancreatic polypeptide predicts incident microvascular and macrovascular complications of type 2 diabetes: An observational study

Abstract

Aims: Pancreatic polypeptide (PP) is elevated in people with vascular risk factors such as type 2 diabetes or increased visceral fat. We investigated potential relationships between PP and microvascular and macrovascular complications of diabetes.

Materials and methods: Animal study: Subcutaneous PP infusion for 4 weeks in high fat diet mouse model. Retinal mRNA submitted for Ingenuity Pathway Analysis. Human study: fasting PP measured in 1478 participants and vascular complications recorded over median 5.5 (IQR 4.9-5.8) years follow-up.

Results: Animal study: The retinal transcriptional response to PP was indicative of cellular stress and damage, and this footprint matched responses described in previously published studies of retinal disease. Of mechanistic importance the transcriptional landscape was consistent with upregulation of folliculin, a recently identified susceptibility gene for diabetic retinopathy. Human study: Adjusting for established risk factors, PP was associated with prevalent and incident clinically significant retinopathy (odds ratio (OR) 1.289 (1.107-1.501) p = 0.001; hazard ratio (HR) 1.259 (1.035-1.531) p = 0.0213), albuminuria (OR 1.277 (1.124-1.454), p = 0.0002; HR 1.608 (1.208-2.141) p = 0.0011), and macrovascular disease (OR 1.021 (1.006-1.037) p = 0.0068; HR 1.324 (1.089-1.61), p = 0.0049), in individuals with type 2 diabetes, and progression to diabetes in non-diabetic individuals (HR 1.402 (1.081-1.818), p = 0.0109).

Conclusions: Elevated fasting PP is independently associated with vascular complications of diabetes and affects retinal pathways potentially influencing retinal neuronal survival. Our results suggest possible new roles for PP-fold peptides in the pathophysiology of diabetes complications and vascular risk stratification.

Keywords: diabetes mellitus; diabetic retinopathy; microvascular disease; pancreatic polypeptide; visceral adiposity.

Figure 1

Ingenuity Pathway Analysis (IPA) of differential retinal gene expression between animals chronically treated with PP and controls (n = 10 in both groups following outlier elimination of 1st and 90th centiles). Panel (A) canonical pathways altered, (B) top eight upstream regulators up and down regulated, (C) Folliculin (FLCN) mechanistic network, (D) most and least enriched disease processes, (E) retinal degeneration-associated gene expression network with significantly enriched ophthalmic disease functions overlaid, (F) top matched human and mouse RNAseq analyses using IPA Analysis Match function, (G) comparison of effects on top canonical pathways between the present study and top matched analyses.

Figure 2

Flowchart illustrating the disposition of participants in the human study. Participants with an underlying diagnosis of Type 1, Monogenic, or Secondary Diabetes, or where the PP radioimmunoassay did not provide a result, were excluded from statistical analysis. NGT, normal glucose tolerance; PP, pancreatic polypeptide.

Figure 3. Summary of Cox regression analyses of logPP and established risk factors on incident vascular complications in participants with Type 2 Diabetes.

Forest plots of logPP and established risk factors for prediction of (A) incidence of the composite significant retinal disease endpoint, (B) incident maculopathy, (C) incident nephropathy, (D) incident ASCVD. Diabetes Duration refers to the recorded duration of diabetes at the time of blood sample collection. eGFR, estimated glomerular filtration rate (ml/min/1.73m2); LDL, low-density lipoprotein; logPP, log2 fasting pancreatic polypeptide; sBP, systolic blood pressure; TC:HDL, ratio of total cholesterol to HDL-C.

Figure 4. Cox proportional hazard regression of logPP, prediabetes, sex, age, and family history on progression to type 2 diabetes.

A two-fold increase in PP is associated with a 1.402 (1.081–1.818), p = 0.0109 fold increase in the momentary risk of progression to type 2 diabetes in 432 individuals with normal glucose tolerance or prediabetes states at recruitment. Panels (A) forest plot of logPP and established risk factors for prediction of incident Type 2 Diabetes in individuals with normal glucose tolerance or prediabetes, (B) cumulative hazard of Type 2 Diabetes onset by quartile of fasting logPP. BMI, body mass index; FHx, positive family history of diabetes.