Plasma apoM Levels and Progression to Kidney Dysfunction in Patients With Type 1 Diabetes

Apolipoprotein M (apoM), a 25-kDa lipocalin protein, has been associated with cardiovascular disease (1) and diabetes (2). apoM is expressed mainly in the liver, followed by the tubular epithelium of the kidney (3). apoM in circulation is primarily bound to HDL; however, apoM also binds to a minor extent with apoB-containing lipoproteins (1,4). Although 90–95% of plasma apoM is bound to HDL, only 5% of the HDL particles in normal subjects contain an apoM molecule (4), and plasma apoM can switch between lipoprotein subtypes depending on the relative concentration of each subtype (5).

apoM is one of the physiological carriers of the bioactive lipid sphingosine 1-phosphate (S1P) in circulation (6). Under normal physiological conditions, plasma apoM and S1P levels are positively associated with the cholesterol content of HDL and LDL (7,8); however, apoM, S1P, and the apoM-bound S1P (apoM/S1P complex) exert independent and interrelated effects on physiological and pathological functions, including endothelial integrity, inflammation, and lipid metabolism (9). Approximately 65% of plasma S1P is bound to apoM and 35% to albumin (10). S1P has a four times longer half-life when bound to apoM/HDL than to albumin (11), suggesting that binding of S1P to apoM/HDL may prevent S1P degradation and contribute to the vasoprotective (12) and antiatherogenic (13) properties of HDL.

The roles of apoM and S1P in diabetes, the effect of diabetes on the plasma apoM/S1P levels, and the potential role of apoM/S1P in the development of diabetes have been recently reviewed (14). Human studies showed that plasma apoM levels were lower in patients with maturity-onset diabetes of the young induced by HNF1A mutations than in patients with type 1 diabetes and relevant control subjects but could not distinguish patients with maturity-onset diabetes of the young from patients with type 2 diabetes (15).

In patients with type 2 diabetes, plasma apoM concentrations were found to be lower (16) or similar (17) compared with levels observed in control subjects. Patients with type 2 diabetes with diabetic nephropathy, defined by the presence of macroalbuminuria (MA), exhibited higher plasma apoM concentrations than patients without diabetic nephropathy, and the levels were also higher in patients with microalbuminuria compared with those with normal albumin excretion rate (AER) (18). In patients with type 1 diabetes, plasma apoM and S1P concentrations did not differ from those in control subjects; however, a shift of apoM and S1P to lipid-rich, light HDL2 particles was seen in the patients with type 1 diabetes (19). In the latter study, the diabetes complication status of the patients was not specified. In this study, we investigated the association of plasma apoM concentration and progression to diabetic kidney disease (DKD) in patients with well-characterized type 1 diabetes with different stages of DKD.

The Diabetes Control and Complications Trial (DCCT) was a randomized controlled trial of 1,441 patients with type 1 diabetes. At study entry, none of the participants had an estimated glomerular filtration rate (eGFR) <60 mL/1.73 m², systolic hypertension (≥140/90 mmHg), total cholesterol >200 mg/dL and/or LDL cholesterol (LDL-C) >160 mg/dL, and AER >200 mg/24 h. Each participant received a complete physical examination, including medical history, electrocardiogram, and routine laboratory analyses (20). Participants were randomized into two groups: receiving either intensive or conventional insulin therapy and followed for an average of 6.5 years. Approximately 95% of the DCCT participants were enrolled into an observational study, the Epidemiology of Diabetes Interventions and Complications (EDIC) study in 1994, to assess long-term effects of prior separation of glycemic levels on micro- and macrovascular outcomes in patients with type 1 diabetes (21). During EDIC, all patients were under the care of their personal physicians and encouraged to practice intensive insulin therapy. The current study was performed on a subgroup of 386 patients who did not participate in the studies performed as part of National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)/National Heart, Lung, and Blood Institute Program Project Grant at the Medical University of South Carolina (1996–2008) and were representative of the whole DCCT/EDIC cohort.

apoM was assayed in plasma samples collected at the time of their entry into the DCCT (baseline) and at the time of DCCT closeout (3–10 years later). Plasma samples were stored at −80°C prior to analysis. apoM levels (in nanograms per milliliter) were measured by a sandwich enzyme immunoassay using a biotin-labeled antibody (BioVendor, Asheville, NC).

The primary study end points for this analysis were progression to MA and/or chronic kidney disease (CKD). End points for each participant were individually assessed from the baseline DCCT visit (1983–1989) through year 18 of the EDIC study or end of study participation (4.8–28.5 years from baseline). Longitudinally measured eGFR was used to define progression to CKD, and AER defined progression to MA. Four-hour urine collections for measurement of AER were obtained annually during DCCT and alternating years during EDIC (20); eGFR was calculated based on annual serum creatinine levels using the Chronic Kidney Disease Epidemiology Collaboration equation (22). CKD was defined as any eGFR measure <60 mL/1.73 m². Time to CKD was defined as the time from DCCT baseline to the date of the first low eGFR value. Incident MA was defined as having an AER ≥300 mg/24 h at any time during the study. Time to MA was defined as the time from DCCT baseline to the date of the first elevated AER value. Although AER is generally associated with progression to CKD, with increases in AER preceding declines in measured eGFR, renal failure can occur in some patients with concurrent normal AER levels. Thus, each outcome was analyzed both together and independent of the progression of the other.

Demographic, clinical, and apoM measurements taken at DCCT baseline and closeout visits were tabulated for the entire cohort as well as by CKD and MA outcome groups. Standard descriptive statistics were used to summarize the general demographic and clinical data. The Kruskal-Wallis test was used to evaluate continuous baseline demographic and clinical measures across kidney dysfunction outcomes; the Pearson χ² test was used to assess the association for categorical variables.

Concentrations of apoM levels were positively skewed; therefore, natural logarithm transformations were applied prior to analysis. Following, apoM was standardized (z scores), and results represent the association between a difference of 1 unit in each biomarker z score and the risk of progression to kidney dysfunction. Repeated-measures logistic regression models were used to assess the association between apoM levels and progression to kidney dysfunction. The primary parameter of interest was the risk ratio (RR) for a 1-standard-unit increase in apoM levels (95% CI) for the progression to each outcome as compared with those that remain persistently normal. Models were adjusted for study design variables (baseline study cohort and randomized treatment assignment), baseline measures of eGFR/AER, and participant sex. Further models assessed the impact of lipids (LDL-C, HDL cholesterol [HDL-C], and triglycerides) on model associations. Modifying effects of DCCT treatment group assignment and baseline disease status (retinopathy and duration of diabetes) on the effects of the association between apoM and progression to kidney disease were examined for all models using appropriate interaction terms. All statistical analyses were performed using SAS system version 9.4 (SAS Institute, Cary, NC).

Access to the DCCT/EDIC data is available through the NIDDK Repository (https://repository.niddk.nih.gov/studies/edic/).

Baseline clinical and demographic characteristics of the subgroup studied are noted in Table 1. At the study baseline visit, participants averaged 27.2 (SD 7.3) years of age, were 51.0% female (n = 197), and had 6.0 (SD 4.3) years of diabetes duration. Although study participants were equally likely to be in either DCCT treatment arm (conventional, 49.7%, or intensive, 50.3%), participants who progress to DKD were less likely to be in the intensive treatment group. Participants who progress were also more likely to be female and have higher baseline HbA_1c.

Fifty-seven of the 386 participants (14.8%) progressed to either MA or CKD and 19 (4.9%) progressed to both MA and CKD during the study. Participants who progressed to MA had a median progression time from study baseline of 13.0 (interquartile range 10.7, 16.9) years, and those who progressed to CKD had a median progression time of 17.7 (interquartile range 14.0, 19.8) years.

At study baseline, participants who progressed to CKD had significantly higher apoM levels than those who remained normal (median [range] 3,111 [1,296, 6,164] vs. 2,795 ng/mL [1,206, 5,777]; P = 0.049). Further, at DCCT closeout, participants who progressed to MA, CKD, or both MA and CKD had significantly higher apoM levels as compared with those who remained normal (MA 2,962 [1,379, 5,407] vs. 2,764 ng/mL [976, 6,158], P = 0.027; CKD 3,257 [1,945, 5,407] vs. 2,764 ng/mL [976, 6,158], P < 0.001; MA and CKD 2,985 (1,945, 5,407) vs. 2,764 ng/mL [976, 6,158], P = 0.001). Overall increases in plasma apoM were associated with a significant increase in the risk of progression to CKD (RR 1.55 [95% CI 1.17, 2.06]) (Table 2) but not progression to MA (RR 1.18 [95% CI 0.97, 1.43]). Effects were primarily driven by a nominal increase in apoM level at DCCT closeout in those who progress to disease. Specifically, increases in apoM at DCCT closeout were associated with increased risk of progression to both MA (RR 1.30 [95% CI 1.01, 1.66]) and CKD (RR 1.69 [95% CI 1.18, 2.44]). Although the treatment group interactions with apoM on kidney outcomes were not statistically significant (P values >0.05), intensive treatment was associated with significant improvement in kidney outcomes overall, and stratified results are shown in Supplementary Table 1. Triglycerides were positively associated with CKD (10-unit increase; RR 1.04 [1.01, 1,07]; P = 0.008) but not MA. When included as a covariate in the modeling of CKD outcomes, parameter estimates are only slightly attenuated and remain significant (data not shown). LDL-C and HDL-C measured concurrently with apoM were not significantly associated with kidney outcomes in any of the models.

Our cohort of patients with type 1 diabetes shows normal levels of HDL-C both at DCCT baseline and closeout and no differences between patients who progress to kidney dysfunction and those who remained normal. Therefore, as recently postulated for cardiovascular disease (CVD) (23,24), the levels of HDL-C in patients with type 1 diabetes may not be a consistent predictor of nephropathy development regardless of the fact that decreased levels of HDL-C are well known predictors in the development of kidney disease and CVD in patients with type 2 diabetes (18).

The protective effect of HDL in the development of atherogenesis and kidney disease are related to the levels of paraoxonase, apoM, and S1P, among others, in HDL. apoM serves as a mediator of reverse cholesterol transport and plays a role in the formation of prebeta HDL and antioxidative properties of HDL (25). As the main carrier of S1P, apoM contributes to the vasculoprotective role of HDL through the engagement of S1P to the S1P receptors (6). In general, HDL3, the denser and cholesterol-poor HDL fraction, carries the highest levels of paraoxonase, apoM, and S1P. However, recent studies performed in type 1 diabetes examining the levels of apoM/S1P in normal subjects and subjects with type 1 diabetes found that there was a shift of the apoM/S1P complexes from dense to light HDL particles in patients with type 1 diabetes compared with control subjects. Interestingly, they also found that the apoM/S1P complexes present in the light HDL fraction were partially dysfunctional, being unable to activate the Akt signaling pathway and inhibiting less efficiently tumor necrosis factor-α–induced vascular cell adhesion molecule-1 expression in aortic endothelial cells (19). In the above study, possible relationships among the shift in the HDL subfraction carrying apoM and S1P and complications of diabetes were not examined. Their isolated HDL subfractions were, however, quite different in composition, and the HDL functionality was altered.

We have studied a cohort of patients with type 1 diabetes and were able to determine that apoM was higher in the patients who later progress to MA, CKD, or both and was associated with progression to MA in an early stage of the disease when AER and eGFR were still normal. That suggests that alterations in the composition and function of HDL in type 1 diabetes may be present very early in the disease process and that these alterations may indeed be associated with the development of nephropathy and likely CVD.

Whether these shifts in composition and functionality of HDL are present both in type 1 and type 2 diabetes and whether or not they can be reversed with treatment is worth exploring, as CVD and nephropathy are primary causes of death in both. The development of nephropathy and CVD is markedly intertwined in both, and finding a common cause (dysfunctional HDL) in the progression of these two complications of diabetes would represent a tremendous advance for the prevention of DKD and CVD in diabetes. As our data show, identification of these abnormalities can be made early in the disease process in type 1 diabetes.

Acknowledgments. The authors thank Waleed Twal, Department of Regenerative Medicine and Cell Biology, Medical University of South Carolina, and Yanchum Li, Division of Diabetes, Endocrinology, and Medical Genetics, Department of Medicine, Medical University of South Carolina, for their contributions to studies conducted to validate apoM measurement.

Funding. This work was supported by National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant R01 DK081352. The work was also supported by the Research Service of the Ralph H. Johnson Department of Veterans Affairs Medical Center. The DCCT/EDIC was sponsored through research contracts from the Division of Diabetes, Endocrinology and Metabolic Diseases of the NIDDK of the National Institutes of Health. Additional support was provided by the National Center for Research Resources through the General Clinical Research Center program and by Genentech through a Cooperative Research and Development Agreement with the NIDDK.

The contents of this manuscript do not represent the views of the Department of Veterans Affairs or the U.S. government.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. A.S., R.L.K., and M.F.L.-V. researched the data and obtained funding. M.F.L.-V. researched the data and wrote, reviewed, and edited the manuscript. N.L.B. and S.M.H. researched the data and wrote, reviewed, and edited the manuscript. N.L.B. and K.J.H. performed statistical analysis and wrote, reviewed, and edited the manuscript. W.T., and Y.L. contributed to the studies conducted to validate apoM measurement. N.L.B. and M.F.L.-V. are the guarantors of this work and, as such, had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.