Use of the NIDDK repository resources to - PowerPoint Presentation

1. Use of the NIDDK repository resources to uncover a new pathway for cardiovascular disease complications in type 1 diabetes Myra Lipes, MDJoslin Diabetes Center, Harvard Medical School

2. OverviewStudy collection: DCCT/EDIC (Diabetes Control and Complications Trial/ Epidemiology of Diabetes Interventions and Complications) study Resources: Serum, genomic DNA, clinical datasetsType of work: Mechanistic, pathogenic processesAccess pathway: R01 DK103609: “Cardiac Autoantibodies as Biomarkers of Heart Disease in Type 1 Diabetes”

3. Transgenic DQ8+mII-/- NOD miceHLA DQ8DQA1*0301DQB1*0302Autoimmune diabetes?Unexpectedly, DQ8+NOD mice died prematurely from heart failure mII-/- DQ8+mII-/- Autoimmune myocarditisOriginal goal: Create a humanized transgenic mouse model of T1DDQ8: Postulated to direct the autoimmune attack to pancreatic β-cellsBackground: What led to these studies?NormalheartDilatedheartDQ8 Insulin peptide Lee, Nat Imm 2001=NOD = Nonobese diabetic

4. Control B6 heartNOD heartCD4CD8B220CD4CD8B220InsulitisCD4CD8B220NOD pancreasGottumukkala et al. Science Translational Medicine 2012 “Autoimmune diabetes inflames the heart” Infarct expansionNormal infarctSimilar lymphocytic compositionExperimental myocardial infarction (MI) induces cardiac autoimmunity in NOD mice 250150100Post-MIPre-MIkD#1#2#1#2#3#4#5#6Pre-MIPost-MI#1#2#1#2#3#4#5#6Cardiac Autoantibodies Western blots of heart lysatesNOD seraControl B6 sera250150100kD

5. Created a panel of assays for cardiac autoantibody detection in humansCardiac antigen panel:MYH6 (encodes full-length α-myosin), the primary/initiating autoantigen in DQ8+NOD myocarditis (Lv, JCI, 2011)S1-MYH6 S2-MYH64. MYH7 (encodes full-length β-myosin) 5. cTnI (cardiac troponin I) Fluid-phase radiobinding assay format, the ‘gold-standard’ format for measuring islet autoantibodies in T1DGottumukkala et al. Sci Trans Med 201225,00050,00020,000S1-MYH6S2-MYH6MYH6~220 KdDiscovered a high prevalence of cardiac autoantibodies in ischemic cardiomyopathy patients with T1D, but not T2D

6. MYH6-S1 AAb index0.51.01.52.02.53.03.5P = 0.0030cTnI AAb index0.51.01.52.02.53.03.58.1-8.5 8.6-9.5 >9.5 HbA1c (%)(n=24) (n=34) (n=22)It-95*It-91*It-24*It-95*It-91*It-24*A surprising finding: 05101520067891011010203040HbA1c (%)41–45 years56–60 yearsAdjusted incidence of heart failure per 1000 patient-years 41–45 yearsCardiac AAb positivity in young adults with T1D and poor glycemic control Alfonso Galderisi, Lipes lab

7. “For patients with diabetes who had very poor glycemic control, the risks of death from any cause and of death from cardiovascular causes were 8 and 10 times as high, respectively, as those in the general population.”Could cardiac autoimmunity be a “missing link” between poor glycemic control and excess CVD risk in patients with T1D?Only ~50% of the effect of hyperglycemia on the risk of CVD in the DCCT/EDIC cohort was explained by traditional risk factors

8. Unique advantages of the DCCT cohort: The two treatment groups achieved markedly different mean A1C levels during the DCCT The Primary Prevention cohort was relatively healthy, young (mean, 27 y) with a short duration of T1D (mean 2.6 y) and devoid of diabetes complications at baseline Availability of longitudinal serum samples from the DCCT and rich clinical datasets with a 30-yr follow-up from the NIDDK CR How to best define a possible association between glycemic control and cardiac autoimmunity in T1D?Mean A1c=9.1%Mean A1c=7.2%

9. DCCT Primary Prevention Cohort n=726 83 cases 234 longitudinal samples83 controls266 longitudinal samplesExclusions:Diabetic kidney disease or CVD events during the DCCT711 eligible participants367 CONV treatment group/344 INT treatment groupCONV treatment group: mean HbA1c 9.0% during the DCCTINT treatment group: mean HbA1c 7.0% during the DCCT191 participants163 participantsSample availabilityand1:1 case/control matchingStudy Design Special thanks to the NIDDK CR staffContact the NIDDK CR support staff as early as possible to assess sample availability

10.     CharacteristicDCCT baseline (1983–1989)    P valueDCCT closeout (1993)   P valueHbA1c 7.0%(n=83)HbA1c 9.0%(n=83)HbA1c 7.0%(n=83)HbA1c 9.0%(n=83)Age, yr 28±626±70.0334±632±70.04Male sex (%)46 (55)46 (55)1.0046 (55)46 (55)1.00Diabetes duration, yr 2.5±1.22.5±1.40.867.8±1.97.9±2.30.63HbA1c, %8.0±1.29.1±1.4<0.0016.5±0.310.0±0.7<0.001Current smoker, (%)13 (16)14 (17)0.8310 (12)15 (18)0.26Body-mass index, kg/m223.1±2.524.0±2.80.0325.3±3.025.4±3.20.84Blood pressure, mm Hg Systolic114±12114±110.89114±11116±110.16 Diastolic72±873±90.4373±975±80.09 Hypertension, (%)*004 (5)1 (1)0.17Lipids HDL cholesterol, mg/dl52±1352±130.9152±1353±130.61 LDL cholesterol, mg/dl111±33108±280.47110±25114±280.36 Total cholesterol, mg/dl176±36173±330.58177±28184±330.16 Triglycerides, mg/dl68±2569±250.7475±3883±500.22 Hyperlipidemia, (%)†0016 (19)21 (26)0.33Microalbuminuria, AER (%)‡ 30 to <300 mg/24 hours001 (1)4 (5)0.17Apart from HbA1c levels, the two groups had similar CVD risk factors at the beginning and end of the DCCTSousa et al. Circulation 2019

11. Temporal dynamics of cardiac autoantibody expression during the DCCTDCCT yearMean HbA1c level (%)HbA1c 9.0% groupHbA1c 7.0% groupDCCT yearMean S1-MYH6 AAb levelsHbA1c 9.0% groupHbA1c 7.0% groupMean AAb IndexMean HbA1c levelsSousa et al. Circulation 2019

12. Percent positive ChagasCardiomyopathyn=51Controlsn=115HbA1c 7.0%n=83HbA1c 9.0%n=83HbA1c 7.0%n=140HbA1c 9.0%n=70Type 2 diabetesJoslin Heart Study2%1%46%22%11%1%51%20%10%4%8%7%4%Number of cardiac AAbType 1 diabetesDCCTPrevalence and profiles of cardiac autoantibodies Young adult DCCT participants with poor glycemic control showed similar AAb profiles to a heart failure cohort with Chagas cardiomyopathy No relationship between glycemic control and cardiac AAbs in patients with type 2 diabetes Sousa et al. Circulation 2019

13. Unexpectedly, the T1D DCCT participants with the highest titers and numbers of cardiac AAbs developed CVD events decades later Masked ID CVD eventYears from first positive AAb test1109CABG26714CVD death251319Acute MI25325CABG + HF20635CABGN/A*110911097147147141319325325Mean AAb IndexS1-MYH6P<0.001P<0.001P<0.001A1c7.0%A1c9.0%Sousa et al. Circulation 2019 S2-MYH6cTnI†Dotted lines show the 99th percentile cutoffs for AAb positivityA1c7.0%A1c9.0%A1c7.0%A1c9.0%The single fatal CVD event occurred in the only DCCTparticipant to test positive for all 5 cardiac AAbs

14. How could cardiac autoimmunity be linked to increased risk of atherosclerotic CVD outcomes? Cardiac autoimmunity↑ Coronary artery ...calcium ↑ CVD events Systemic inflammation:↑hsCRP? High-sensitivity C-reactive protein levels were measured during EDIC years 4−6 (NIDDK CR dataset: inflammatory_markers_musc.sas7bdat) ?

15. Serum hsCRP (mg/L)P=0.34A1c 7.0(n=67)A1c 9.0%(n=62)No difference in hsCRP levels between the two DCCT HbA1c groups Sousa et al. Circulation, 2019

16. Serum hsCRP (mg/L)P=0.34A1c 7.0(n=67)A1c 9.0%(n=62)2 AAbs(n=15)P=0.003Positivity for ≥2 cardiac AAbs was associated with markedly elevated hsCRP levels 1 AAb(n=114)median, 6.0 mg/Lmedian, 1.4 mg/L Serum hsCRP (mg/L)Suggests that subclinical inflammation may link cardiac AAbs to CVD outcomes Sousa et al. Circulation, 2019

17. FactorPostulated consequenceAt-risk individualsGeneral population Lack of thymic (‘central’) toleranceto α-myosin (MYH6) Peripheral blood enriched in CD4+ T cells specific to α-myosin AT1D and T2D with poor glycemic controlChronic hyperglycemia Subclinical myocardial injury BLeakage and exposure of heart muscle proteinsT1D with poor glycemic control Expansion and persistent activation of proinflammatory α-myosin-reactive CD4+ T cells Positivity for multiple cardiac autoantibodies particularly to MYH6 Increased “proinflammatory load” in the vasculature Accelerated atherosclerosis and increased risk of CVD events Dysregulated (‘overreactive’) adaptive immune responses to α-myosinCElevated hsCRP levelsProposed scheme of the relationship between hyperglycemia, cardiac autoimmunity, and increased CVD risk in T1D Sousa et al. Circulation, 2019

18. Impact of this NIDDK CR-based study included:NIH/NIDDK R01 DK125677: “Cardiac Autoimmunity as a Mediator of Cardiovascular Disease in Type 1 Diabetes”Stimulated other projects and further scientific discoveriesJDRF: “Validation of a New Approach for Cardiovascular Disease Risk Stratification and Therapeutic Targeting in T1D” → Scottish Diabetes Research Network Type 1 Bioresource (H. Colhoun) and CACTI cohort (J. Snell-Bergeon)Low Wang CC, Michels A. Expert Analysis: The missing link for cardiovascular disease in type 1 diabetes mellitus? Hyperglycemia-induced cardiac autoimmunity. Am Coll Cardiol. 2019 Jul 15.Featured in several editorials/commentaries ●“The studies by Sousa et al are elegant and represent a credible additional mechanism for accelerated CHD risk in type 1 diabetes…and require urgent validation”