Journal of Clinical Lipidology (2019) -, -–-

1 52
2 53
3 54
4 Original Article 55
5 56
6 57
7 Type 1 diabetes is associated with an increase 58
8 59
9 in cholesterol absorption markers but a 60
10 61
11 decrease in cholesterol synthesis markers in 62
12Q1 63
13 a young adult population 64
14 65
15 66
16
Q8
Ivana Semova, PhD1, Amy E. Levenson, MD1,2, Joanna Krawczyk, Kevin Bullock, PhD, 67
17 Kathryn A. Williams, MS, R. Paul Wadwa, MD, Amy S. Shah, MD, MS, 68
18 Philip R. Khoury, PhD, Thomas R. Kimball, MD, Elaine M. Urbina, MD, MS, 69
19 70
20 Sarah D. de Ferranti, MD, MPH, Franziska K. Bishop, MS, David M. Maahs, MD, PhD3, 71
21Q2 Lawrence M. Dolan, MD, Clary B. Clish, PhD, Sudha B. Biddinger, MD, PhD* 72
22 73
23 74
24 Division of Endocrinology, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA (Drs Semova, 75
25Q3 Levenson, Krawczyk, Williams, and Biddinger); Broad Institute of MIT and Harvard, Cambridge, MA, USA (Drs Bullock 76
26 and Clish); Biostatistics and Research Design Center, Boston Children’s Hospital, Boston, MA, USA (Dr Williams); 77
27 Barbara Davis Center for Childhood Diabetes, University of Colorado School of Medicine, Aurora, CO, USA (Drs 78
28 Wadwa, Bishop, and Maahs); Division of Endocrinology, Cincinnati Children’s Hospital Medical Center and University of 79
29 Cincinnati, Cincinnati, OH, USA (Drs Shah, Khoury, and Dolan); Division of Cardiology, Cincinnati Children’s Hospital 80
30 Medical Center and University of Cincinnati, Cincinnati, OH, USA (Drs Kimball and Urbina); and Department of 81
31 Cardiology, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA (Dr de Ferranti) 82
32 83
33 84
KEYWORDS: BACKGROUND: To optimize treatment and prevent cardiovascular disease in subjects with type 1
34 85
Cholesterol metabolism; diabetes, it is important to determine how cholesterol metabolism changes with type 1 diabetes.
35 86
Dyslipidemia; OBJECTIVE: The objective of the study was to compare plasma levels of campesterol and b-sitos-
36 terol, markers of cholesterol absorption, as well as lathosterol, a marker of cholesterol synthesis, in 87
Cholesterol-lowering
37 therapy; youth with and without type 1 diabetes. 88
38 Cardiovascular disease METHODS: Serum samples were obtained from adolescent subjects with type 1 diabetes (n 5 175, 89
39 risk; mean age 15.2 years, mean duration of diabetes 8.2 years) and without diabetes (n 5 74, mean age 90
40 Youth; 15.4 years). Campesterol, b-sitosterol, and lathosterol, were measured using targeted liquid chromatog- 91
41 Type 1 diabetes raphy tandem mass spectrometry, normalized to the control serum mean, and expressed in arbitrary 92
42 units. The markers were then compared between groups and correlated with the available cardiometa- 93
43 bolic variables. 94
44 RESULTS: Campesterol and b-sitosterol levels were 30% higher in subjects with type 1 diabetes and 95
positively correlated with hemoglobin A1c levels. In contrast, lathosterol levels were 20% lower in
45 96
46 97
47 Conflict of interest: none. 98
1 * Corresponding author. Children’s Hospital Boston, 3 Blackfan Circle,
These authors contributed equally to this work.2 Present address: Di-
48 vision of Pediatric Endocrinology, University of North Carolina School of
Boston, MA 02115, USA. 99
49 E-mail address: sudha.biddinger@childrens.harvard.edu 100
Medicine, Chapel Hill, NC.3 Present address: Division of Endocrinology,
Submitted June 3, 2019. Accepted for publication September 17, 2019.
50 Department of Pediatrics, Stanford School of Medicine, Stanford Diabetes 101
51 Research Center, Stanford University, Stanford, CA. 102

1933-2874/Ó 2019 National Lipid Association. All rights reserved.

https://doi.org/10.1016/j.jacl.2019.09.008

FLA 5.6.0 DTD  JACL1501_proof  16 October 2019  7:43 am

 2 Journal of Clinical Lipidology, Vol -, No -, - 2019

103 159
104 subjects with type 1 diabetes and positively correlated with triglycerides, body mass index, and systolic 160
105 blood pressure. 161
106 CONCLUSION: Plasma markers suggest that cholesterol absorption is increased, whereas choles- 162
terol synthesis is decreased in type 1 diabetes. Further studies to address the impact of these changes
107 163
on the relative efficacy of cholesterol absorption and synthesis inhibitors in subjects with type 1 dia-
108 betes are urgently needed.
164
109 Ó 2019 National Lipid Association. All rights reserved. 165
110 166
111 167
112 models of T1D show increased cholesterol absorption but 168
Introduction
113 decreased cholesterol synthesis.24–29 Consistent with this, 169
114 The risk of cardiovascular disease (CVD) in subjects intensive insulin treatment lowers cholesterol absorption in 170
115 with type 1 diabetes (T1D) remains exceedingly high, up to T1D subjects.30 Moreover, several excellent studies have 171
116 30-fold higher than in those without diabetes in some found increased cholesterol absorption and decreased choles- 172
117 populations,1 despite the numerous recent advancements in terol synthesis in T1D patients.30–33 However, limitations in 173
118 therapy.2 Furthermore, lifespan is compromised by 11– sample size have precluded their generalizability. 174
119 13 years in subjects with T1D compared with nondiabetic Here, we measured markers of cholesterol synthesis and 175
120 subjects, with CVD as the main cause of this discrepancy.3,4 absorption in a large cohort of adolescent subjects with 176
121 Therefore, an important goal of treatment for T1D subjects T1D and similar-aged controls.34 The use of an adolescent 177
122 is the prevention of CVD. cohort enabled us to avoid the confounding effects of lipid- 178
123 The Diabetes Control and Complications Trial/Epide- lowering drugs, as this population is largely drug na€ıve. 179
124 miology of Diabetes Interventions and Complications trial 180
125 showed conclusively that intensive insulin treatment lowers 181
126 cardiovascular risk5 and, indeed, subjects with good glyce- Methods 182
127 mic control show lipid profiles comparable to controls.6,7 183
128 However, most T1D subjects do not achieve their glycemic Study population 184
129 target,1 and poor control is associated with high rates of 185
130 dyslipidemia, including elevated total cholesterol (TC), As previously described,34 the study population consisted 186
131 low-density lipoprotein cholesterol (LDL-C), non–high- of individuals with T1D and individuals in the Denver area 187
132 density lipoprotein (HDL) cholesterol, and apolipoprotein with no chronic disease who participated in a study of cardio- 188
133 B levels.8-12 Thus, lipid-lowering drugs can become very vascular risk factors at the Barbara Davis Center for Child- 189
134 important. Indeed, up to 30% of the excess cardiovascular hood Diabetes. All subjects were 12–20 years of age. 190
135 death associated with poor control may be secondary to Individuals with T1D (diagnosed by the presence of islet 191
136 high cholesterol levels.13 cell antibodies or provider clinical diagnosis) had diabetes 192
137 Cholesterol synthesis inhibitors (statin drugs) have duration . 5 years. For the present study, all subjects 193
138 generally been found to be much more effective in lipid- (n 5 252) with plasma samples and data available for fasting 194
139 lowering than cholesterol absorption inhibitors (such as metabolic measurements and lipid parameters were 195
140 ezetimibe): thus, statins reduced LDL-C levels by 32%– included. Subjects who reported taking statin medications 196
141 38% in nondiabetic and type 2 diabetic subjects, whereas at the time of the study (n 5 3) were excluded from the ana- 197
142 ezetimibe only reduced LDL-C by 15%–19%.14–16 Simi- lyses because these medications have been shown to increase 198
143 larly, both adults and adolescents with hypercholesterole- cholesterol synthesis.35 The study was approved by the Col- 199
144 mia showed a 35%–50% reduction in LDL-C with statins, orado Multiple Institutional Review Board, and informed 200
145 but a 19%–42% reduction with ezetimibe.17–21 consent and assent (for subjects aged ,18 years) were ob- 201
146 Large studies comparing the effectiveness of statins vs tained in all subjects before participation in the study. 202
147 ezetimibe in T1D are lacking. T1D subjects appeared to 203
148 respond to statins, as they showed the same reduction in Anthropometric measures and laboratory assays 204
149 CVD for a given reduction in LDL as type 2 diabetic 205
150 subjects.22 However, a direct comparison of the two drugs in Anthropometric measures were obtained during study 206
151 a small cohort of patients showed ezetimibe to be more visits, including height to the nearest 0.1 cm, and weight to 207
152 effective than atorvastatin in lowering LDL-C in T1D sub- the nearest 0.1 kg using a Detecto scale (Detecto, Webb 208
153 jects (32% vs 19% reduction, respectively).23 City, Missouri). Subjects fasted for at least 8 hours before 209
154 These studies point to potential differences in efficacy of the study visit. Hemoglobin A1c (HbA1c) was measured on 210
155 different lipid-lowering drugs in T1D subjects. In this regard, the DCA Vantage (Siemens, Princeton, New Jersey) at the 211
156 it is important to consider the possibility that subjects with Children’s Hospital Colorado lab. Lipid parameters, 212
157 T1D may have unique changes in cholesterol metabolism that including TC, HDL-C, and triglycerides (TGs), were 213
158 alter the relative effectiveness of different drugs. Rodent obtained in the Clinical Translational Research Core lab 214

FLA 5.6.0 DTD  JACL1501_proof  16 October 2019  7:43 am

 Semova et al Cholesterol markers in type 1 diabetes 3

215 using a Beckman Coulter AU system (Beckman Coulter gas temperature; 11 L/min, sheath gas flow; 0.5 kV, nozzle 271
216 Inc, Brea, CA). LDL-C was calculated using the Friede- voltage; 150, high pressure RF; 90, low pressure RF. Peak 272
217 wald formula (no subjects showed TG . 400 mg/dL). integration was performed using Agilent MassHunter soft- 273
218 High-sensitivity C-reactive protein (Hs-CRP) was ware (Agilent, G3336AA); the area under the curve was 274
219 measured at the Children’s Hospital Colorado lab using a normalized to a standardized pooled serum sample and 275
220 multiplex assay platform Siemens BNII Nephelometer values were reported as arbitrary units. 276
221 (Siemens, Princeton, New Jersey). 277
222 Biostatistical analysis 278
223 Sample preparation and mass spectrometry 279
224 analysis Values are reported as mean 6 standard deviation (range) 280
225 or number (% from total). Sterol values were normalized to 281
226 Sterols were extracted from serum and derivatized to the control mean value. Bivariate Spearman correlations 282
227 picolinyl esters as previously described.36 Dried sterol deriv- were calculated between cholesterol absorption and synthe- 283
228 atives were reconstituted in acetonitrile and analyzed using a sis markers and potential covariates (HbA1c, fasting 284
229 targeted liquid chromatography tandem mass spectrometry glucose, C-peptide, body mass index (BMI), Hs-CRP, TG, 285
230 method operated on an Agilent 1290 U-HPLC coupled to TC, LDL-C, HDL-C, systolic blood pressure [SBP] and 286
231 Agilent 6495 triple quadrupole mass spectrometer. Briefly, diastolic blood pressure [DBP]). Significance was deter- 287
232 samples (10 mL) were injected onto a Hypersil GOLD col- mined using Student’s t-test, Fisher’s exact test, or two-way 288
233 umn (150 ! 2.1 mm, 3 mm, Thermo Electron) that was eluted ANOVA. P-value lower than 0.05 was considered significant 289
234 at a flow rate of 300 mL/min with acetonitrile/methanol/water for all analyses; n 5 249 total, 74 controls, 175 T1D for all 290
235 (40/40/20, v/v/v) with 0.1% acetic acid (mobile phase A) for measurements, except for HbA1c (n 5 246 total, 71 con- 291
236 0.5 minutes followed by a linear gradient to acetonitrile/ trols), C-peptide, and Hs-CRP (n 5 248 total, 73 controls). 292
237 methanol/water (45/45/10, v/v/v) with 0.1% acetic acid (mo- 293
238 bile phase B) over 19.5 minutes and held for 21 minutes. 294
239 Mass spectra were acquired using electrospray ionization Results 295
240 in the positive ion mode and using dynamic multiple reaction 296
241 monitoring scanning. Collision energies and precursor-to- Demographic, anthropometric, and 297
242 product ion transitions were determined using derivatized cardiovascular risk factors by study group 298
243 authentic reference standards. Mass spectrometer setting 299
244 were 3.5 kV, ionization voltage; 200 C, gas temperature; There were no statistically significant differences by age, 300
245 14 L/min, gas flow; 40 psi, nebulizer pressure; 325 C, sheath sex, BMI, Hs-CRP, and tobacco use between the groups. As 301
246 302
247 Q4
303
Table 1 Demographics and clinical parameters of subjects with T1D and nondiabetic controls
248 304
249 Control (n 5 74) T1D (n 5 175) P-value* 305
250 Age (y) 15.4 6 2.2 (12.0–20.0) 15.2 6 2.2 (12.0–19.4) .547 306
251 Sex (% female) 42 F/32 M (56.8%) 99 F/76 M (56.6%) 1.000 307
252 Tobacco use now 6 (8.1%) 12 (6.9%) .790 308
253 HbA1c (%) 5.3 6 0.3 (4.6–5.9) 9.0 6 1.6 (5.8–14.0) ,.001 309
254 Fasting glucose (mg/dL) 82 6 7 (64–99) 185 6 80 (40–451) ,.001 310
255 C-peptide (ng/mL) 1.7 6 0.6 (0.1–3.3) 0.1 6 0.1 (0.1–1.3) ,.001 311
256 BMI (kg/m2) 21.8 6 4.2 (13.8–33.5) 22.4 6 3.4 (15.3–35.3) .259 312
257 Hs-CRP (mg/dL) 0.9 6 1.4 (0.0–6.4) 1.3 6 2.3 (0.1–22.0) .084 313
TG (mg/dL) 83.5 6 41.5 (34.0–235.0) 86.8 6 50.5 (28.0–333.0) .586
258 314
TC (mg/dL) 147.7 6 28.0 (96.0–235.0) 159.4 6 34.8 (89.0–347.0) .006
259 315
LDL-C (mg/dL) 82.4 6 22.9 (43.0–168.0) 89.9 6 27.8 (42.0–250.0) .028
260 HDL-C (mg/dL) 48.6 6 9.3 (28.0–74.0) 52.1 6 10.6 (30.0–81.0) .010 316
261 SBP (mmHg) 109.0 6 8.5 (89.0–129.0) 113.1 6 8.5 (93.0–137.0) ,.001 317
262 DBP (mmHg) 64.3 6 6.1 (49.0–79.0) 69.0 6 6.5 (50.0–89.0) ,.001 318
263 Campesterol (AU) 1.00 6 0.35 (0.41–1.95) 1.31 6 0.41 (0.47–2.60) ,.001 319
264 b-Sitosterol (AU) 1.00 6 0.36 (0.39–2.25) 1.29 6 0.44 (0.50–2.72) ,.001 320
265 Lathosterol (AU) 1.00 6 0.51 (0.32–2.52) 0.79 6 0.38 (0.26–2.77) .001 321
266 T1D, type 1 diabetes; HbA1c, hemoglobin A1c; BMI, body mass index; Hs-CRP, high-sensitivity C-reactive protein; TG, triglycerides; TC, total 322
267 cholesterol; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; SBP, systolic blood pressure; DBP, diastolic blood 323
268 pressure; AU, arbitrary units. 324
Values are presented as mean 6 SD (range) or n (%). *P-value based on t-tests or Fisher’s exact test. Boldface type highlights statistical significance
269 325
with P , .05.
270 326

FLA 5.6.0 DTD  JACL1501_proof  16 October 2019  7:43 am

 4 Journal of Clinical Lipidology, Vol -, No -, - 2019

327 A B C 383
328 2 2 2 384
p<0.001 p<0.001 p=0.001
329 385
1.5 1.5 1.5
330 386

A.U.

A.U.

A.U.
331 1 1
387
1
332 388
333 0.5 0.5 0.5 389
334 390
335 0 0 0 391
Control T1D Control T1D Control T1D
336 392
337 Figure 1 Abundance of markers of cholesterol absorption (A, B) and synthesis (C) in T1D subjects and nondiabetic controls. Values are 393
338 presented as mean 6 SD; *P-value based on t-tests. AU, arbitrary units; T1D, type 1 diabetes. 394
339 395
340 expected, HbA1c (P , .001) and fasting plasma glucose In the complete cohort, serum levels of campesterol and 396
341 levels (P , .001) were significantly higher, and C-peptide b-sitosterol correlated positively with HbA1c and fasting 397
342 levels were significantly lower (P , .001) in T1D subjects. glucose and negatively with C-peptide. With regard to 398
343 TG levels were similar between controls and T1D subjects. plasma lipids, campesterol and b-sitosterol were positively 399
344 In contrast, TC (P , .01), LDL-C (P , .05), HDL-C associated with TC, LDL-C, and HDL-C. Both markers 400
345 (P , .05), SBP (P , .001) and DBP (P , .001) were signif- were negatively associated with BMI, whereas b-sitosterol 401
346 icantly higher in T1D subjects (Table 1). was negatively associated with SBP. When the T1D and 402
347 control subjects were analyzed separately, the association 403
348 Cholesterol absorption and synthesis markers with HbA1c remained significant in T1D subjects only. 404
349 However, the associations with TC, LDL-C, and HDL-C 405
350 The serum levels of campesterol and b-sitosterol remained significant in both groups. A significant negative 406
351 (cholesterol absorption markers) were 30% higher in T1D association with Hs-CRP emerged for both campesterol and 407
352 subjects compared with controls (P , .001, Table 1 and b-sitosterol in the control group only (Table 2). 408
353 Fig. 1). In contrast, the serum levels of lathosterol (choles- Serum levels of lathosterol correlated negatively with 409
354 terol synthesis marker) were 20% lower in T1D subjects HbA1c and positively with C-peptide. Lathosterol was 410
355 compared with controls. These markers also differed signif- positively correlated with Hs-CRP, particularly in the 411
356 icantly when comparing T1D to controls for male and for controls; SBP and BMI in all subjects; and DBP in T1D 412
357 female subjects separately (P , .05) but not when subjects. For the lipids, lathosterol was not correlated with 413
358 comparing male and female subjects within the T1D and HDL-C but was positively correlated with TG. On the other 414
359 control groups (P . .05) (Supplemental Fig. 1). hand, lathosterol, like campesterol and b-sitosterol, was 415
360 416
361 417
362 418
Table 2 Spearman correlation coefficients (P-value) for the markers of cholesterol absorption and clinical parameters in subjects with
363 419
T1D and nondiabetic controls
364 420
365 Campesterol b-Sitosterol 421
366 All (n 5 249) Control (n 5 74) T1D (n 5 175) All (n 5 249) Control (n 5 74) T1D (n 5 175) 422
367 423
HbA1c (%) 0.41 (,0.001) 0.11 (0.379) 0.25 (0.001)* 0.35 (,0.001) 0.06 (0.636) 0.20 (0.007)*
368 424
Fasting glucose 0.2 (0.002) 20.04 (0.742) 20.03 (0.668) 0.18 (0.005) 20.14 (0.233) 20.03 (0.680)
369 (mg/dL)
425
370 C-peptide (ng/mL) 20.35 (,0.001) 20.21 (0.080) 0.06 (0.414) 20.32 (,0.001) 20.24 (0.040)* 0.01 (0.889) 426
371 BMI (kg/m2) 20.18 (0.005) 20.36 (0.002) 20.15 (0.049) 20.22 (,0.001) 20.41 (,0.001) 20.18 (0.020) 427
372 Hs-CRP (mg/dL) 20.04 (0.539) 20.24 (0.044)* 20.04 (0.582) 20.05 (0.420) 20.26 (0.026)* 20.04 (0.628) 428
373 TG (mg/dL) 20.004 (0.951) 20.12 (0.294) 0.06 (0.467) 0.01 (0.831) 20.11 (0.329) 0.07 (0.363) 429
374 TC (mg/dL) 0.37 (,0.001) 0.33 (0.004) 0.33 (,0.001) 0.38 (,0.001) 0.31 (0.007) 0.36 (,0.001) 430
375 LDL-C (mg/dL) 0.33 (,0.001) 0.37 (0.001) 0.28 (,0.001) 0.33 (,0.001) 0.31 (0.007) 0.30 (,0.001) 431
376 HDL-C (mg/dL) 0.26 (,0.001) 0.26 (0.023) 0.2 (0.008) 0.28 (,0.001) 0.28 (0.014) 0.24 (0.002) 432
377 SBP (mmHg) 20.11 (0.085) 20.23 (0.050) 20.17 (0.022) 20.14 (0.022) 20.30 (0.009) 20.19 (0.014) 433
378 DBP (mmHg) 0.08 (0.184) 20.10 (0.404) 0.01 (0.847) 0.04 (0.549) 20.14 (0.232) 20.02 (0.760) 434
379 T1D, type 1 diabetes; HbA1c, hemoglobin A1c; BMI, body mass index; Hs-CRP, high-sensitivity C-reactive protein; TG, triglycerides; TC, total 435
380 cholesterol; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; SBP, systolic blood pressure; DBP, diastolic blood 436
pressure.
381 Boldface type highlights statistical significance with P , .05; asterisk indicates statistically significant difference between controls and T1D subjects.
437
382 438

FLA 5.6.0 DTD  JACL1501_proof  16 October 2019  7:43 am

 Semova et al Cholesterol markers in type 1 diabetes 5

439 495
Table 3 Spearman correlation coefficients (P-value) for the marker of cholesterol synthesis and clinical parameters in subjects with
440 T1D and nondiabetic controls 496
441 497
442 Lathosterol 498
443 All (n 5 249) Control (n 5 74) T1D (n 5 175) 499
444 500
HbA1c (%) 20.13 (0.046) 20.17 (0.149) 0.10 (0.188)
445 Fasting glucose (mg/ 20.04 (0.581) 0.16 (0.162) 0.11 (0.143) 501
446 dL) 502
447 C-peptide (ng/mL) 0.22 (,0.001) 0.34 (0.003)* 20.03 (0.666) 503
448 BMI (kg/m2) 0.36 (,0.001) 0.58 (,0.001) 0.31 (,0.001) 504
449 Hs-CRP (mg/dL) 0.18 (0.004) 0.61 (,0.001)* 0.05 (0.525) 505
450 TG (mg/dL) 0.39 (,0.001) 0.56 (,0.001) 0.35 (,0.001) 506
451 TC (mg/dL) 0.28 (,0.001) 0.46 (,0.001) 0.29 (,0.001) 507
452 LDL-C (mg/dL) 0.23 (,0.001) 0.30 (0.008) 0.27 (,0.001) 508
453 HDL-C (mg/dL) 20.03 (0.608) 0.08 (0.492) 20.04 (0.630) 509
SBP (mmHg) 0.25 (,0.001) 0.34 (0.003) 0.30 (,0.001)
454 510
DBP (mmHg) 0.12 (0.057) 0.21 (0.077) 0.21 (0.006)
455 511
456 T1D, type 1 diabetes; HbA1c, hemoglobin A1c; BMI, body mass index; Hs-CRP, high-sensitivity C-reactive protein; TG, triglycerides; TC, total 512
cholesterol; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; SBP, systolic blood pressure; DBP, diastolic blood
457 pressure.
513
458 Boldface type highlights statistical significance with P , .05; asterisk indicates statistically significant difference between controls and T1D subjects. 514
459 515
460 516
461 517
462 positively correlated with TC and LDL-C in all subjects with direct synthesis and absorption measurements,43–47 518
463 (Table 3). which were not practical in a cohort of this size. 519
464 These data raise the possibility that drugs that reduce 520
465 cholesterol absorption may be relatively more effective in 521
466 Discussion T1D subjects than nondiabetic and T2D subjects.14–16 522
467 Thus, additional, large-scale studies in T1D subjects, 523
468 In a cohort of 175 T1D and 74 control subjects, we find comparing the effectiveness of cholesterol absorption in- 524
469 that T1D is associated with higher campesterol and b- hibitors vs cholesterol synthesis inhibitors in lowering TC 525
470 sitosterol and lower lathosterol. Although campesterol, b- and LDL-C, as well as CVD risk, are urgently needed.48 526
471 sitosterol, and lathosterol are all associated with TC and 527
472 LDL-C levels in the complete cohort, campesterol and b- 528
473 sitosterol are positively correlated with HbA1c and fasting Acknowledgments 529
474 glucose levels, and lathosterol is positively correlated with 530
475 C-peptide, TG, BMI, SBP, and DBP. Moreover, Hs-CRP is This study was supported by the Juvenile Diabetes 531
476 negatively associated with campesterol and b-sitosterol in Research Foundation Ltd (grant 11-2007-694), National 532
477 controls but positively associated with lathosterol. Institute of Diabetes and Digestive and Kidney Diseases 533
478 These findings extend previous studies to a larger, (grant DK075360), NIH/NCRR Colorado CTSI grant 534
479 younger cohort,30 and confirm a positive correlation be- UL1 RR025780, American Diabetes Association grant 535
480 tween markers of cholesterol synthesis with TG, BMI, 9-18-CVD1-003 (I.S.), NIH training grant No. T32 536
481 SBP, and DBP.33,37,38 In addition, they document a positive DK007260 (J.K.), National Institute of Diabetes and 537
482 correlation between markers of cholesterol absorption and Digestive and Kidney Diseases grants K23 DK075360 538
483 HbA1c not observed in previous studies,39,40 perhaps and P30 DK116074 (D.M.M.), National Institute of 539
484 because of the inclusion of subjects treated with medica- Health National Heart, Lung, and Blood Institute grant 540
485 tions that could interfere with cholesterol absorption.41,42 R01-HL-109650 (S.B.B.), American Heart Association 541
486 An important limitation of the present study is the use of Established Investigator Award (S.B.B.), National Insti- Q5 542
487 serum markers of cholesterol synthesis and absorption tute of Diabetes and Digestive and Kidney Diseases grant 543
488 rather than direct measurements with isotope tracers. 5K12-DK-094721-04 (A.E.L.), and an SPARC grant from 544
489 Increases in dietary plant sterol content or reduced biliary the Broad Institute. 545
490 sterol secretion could increase plasma b-sitosterol and Authors’ contributions: I.S., A.E.L., R.P.W., F.K.B., 546
491 campesterol levels, independently of intestinal absorption. D.M.M., and S.B.B. designed the study; R.P.W., F.K.B., 547
492 In addition, b-sitosterol and campesterol are associated D.M.M. collected blood samples and anthropometric and 548
493 with HDL31; therefore, their increased abundance in T1D demographic data; I.S., A.E.L., J.K., and K.B. performed 549
494 subjects could be due, to some degree, to their increased LC-MS/MS; I.S., S.B.B., and C.B.C. analyzed the data; 550
HDL levels. Nonetheless, serum markers correlate well K.A.W. performed the biostatistical analysis; I.S. and

FLA 5.6.0 DTD  JACL1501_proof  16 October 2019  7:43 am

 6 Journal of Clinical Lipidology, Vol -, No -, - 2019

551 S.B.B. wrote the manuscript. All authors contributed to the 18. Clauss S, Wai KM, Kavey RE, Kuehl K. Ezetimibe treatment of pedi- 607
552 review and editing of the manuscript. atric patients with hypercholesterolemia. J Pediatr. 2009;154: 608
869–872.
553 19. Yeste D, et al. Ezetimibe as monotherapy in the treatment of hyper-
609
554 cholesterolemia in children and adolescents. J Pediatr Endocrinol 610
555 Supplementary data Metab. 2009;22:487–492. 611
556 20. Farnier M, et al. Effects of ezetimibe, simvastatin and ezetimibe/sim- 612
557 Supplementary data to this article can be found online at vastatin on correlations between apolipoprotein B, LDL cholesterol 613
558 https://doi.org/10.1016/j.jacl.2019.09.008. and non-HDL cholesterol in patients with primary hypercholesterole-
614
mia. Atherosclerosis. 2013;229:415–422.
559 21. Avis HJ, et al. Efficacy and safety of rosuvastatin therapy for children 615
560 with familial hypercholesterolemia. J Am Coll Cardiol. 2010;55: 616
561
References 1121–1126. 617
1. Miller RG, et al. A contemporary estimate of total mortality and car- 22. Cholesterol Treatment Trialists, C., et al. Efficacy of cholesterol-
562 diovascular disease risk in young adults with type 1 diabetes: the Pitts-
618
lowering therapy in 18,686 people with diabetes in 14 randomised tri-
563 burgh epidemiology of diabetes complications study. Diabetes Care. als of statins: a meta-analysis. Lancet. 2008;371:117–125.
619
564 Q6 2016;39:2296–2303. 23. Ciriacks K, Coly G, Krishnaswami S, Patel SB, Kidambi S. Effects of 620
565 2. Lind M, et al. Glycemic control and excess mortality in type 1 dia- simvastatin and ezetimibe in lowering low-density lipoprotein choles- 621
566 betes. N Engl J Med. 2014;371:1972–1982. terol in subjects with type 1 and type 2 diabetes mellitus. Metab Syndr 622
3. Livingstone SJ, et al. Estimated life expectancy in a Scottish cohort Relat Disord. 2015;13:84–90.
567 with type 1 diabetes, 2008-2010. JAMA. 2015;313:37–44.
623
24. Uchida K, et al. Altered bile acid metabolism in alloxan diabetic rats.
568 4. Huo L, Harding JL, Peeters A, Shaw JE, Magliano DJ. Life expec- Jpn J Pharmacol. 1979;29:553–562.
624
569 tancy of type 1 diabetic patients during 1997-2010: a national Austra- 25. Akiyoshi T, Uchida K, Takase H, Nomura Y, Takeuchi N. Cholesterol 625
570 lian registry-based cohort study. Diabetologia. 2016;59:1177–1185. gallstones in alloxan-diabetic mice. J Lipid Res. 1986;27:915–924. 626
571 5. Nathan DM, et al. Intensive diabetes treatment and cardiovascular dis- 26. Uchida K, Makino S, Akiyoshi T. Altered bile acid metabolism in non- 627
ease in patients with type 1 diabetes. N Engl J Med. 2005;353: obese, spontaneously diabetic (NOD) mice. Diabetes. 1985;34:79–83.
572 2643–2653.
628
27. Milliat F, et al. Short and long-term effects of streptozotocin on dietary
573 6. Lipid and lipoprotein levels in patients with IDDM diabetes control cholesterol absorption, plasma lipoproteins and liver lipoprotein re-
629
574 and complication. Trial experience. The DCCT Research Group. Dia- ceptors in RICO rats. Exp Clin Endocrinol Diabetes. 2000;108: 630
575 betes care. 1992;15:886–894. 436–446. 631
576 7. Sosenko JM, Breslow JL, Miettinen OS, Gabbay KH. Hyperglycemia 28. Biddinger SB, et al. Hepatic insulin resistance is sufficient to produce 632
and plasma lipid levels: a prospective study of young insulin- dyslipidemia and susceptibility to atherosclerosis. Cell Metab. 2008;7:
577 dependent diabetic patients. N Engl J Med. 1980;302:650–654.
633
125–134.
578 8. Guy J, et al. Lipid and lipoprotein profiles in youth with and without 29. Miao J, et al. Hepatic insulin receptor deficiency impairs the SREBP-2
634
579 type 1 diabetes: the SEARCH for Diabetes in Youth case-control response to feeding and statins. J Lipid Res. 2014;55:659–667. 635
580 study. Diabetes care. 2009;32:416–420. 30. Kojima H, et al. Effect of glycemic control on plasma plant sterol 636
581 9. Maahs DM, et al. Longitudinal lipid screening and use of lipid- levels and post-heparin diamine oxidase activity in type 1 diabetic pa- 637
lowering medications in pediatric type 1 diabetes. J Pediatr. 2007; tients. Atherosclerosis. 1999;145:389–397.
582 150:146–150.
638
31. Miettinen TA, Gylling H, Tuominen J, Simonen P, Koivisto V. Low
583 10. Petitti DB, et al. Serum lipids and glucose control: the SEARCH for synthesis and high absorption of cholesterol characterize type 1 dia-
639
584 Diabetes in Youth study. Arch Pediatr Adolesc Med. 2007;161: betes. Diabetes care. 2004;27:53–58. 640
585 159–165. 32. Gylling H, et al. Markers of absorption and synthesis of cholesterol in 641
586 11. Vaid S, Hanks L, Griffin R, Ashraf AP. Body mass index and glycemic men with type 1 diabetes. Diabetes Metab Res Rev. 2007;23:372–377. 642
control influence lipoproteins in children with type 1 diabetes. J Clin 33. Jarvisalo M, Raitakari O, Gylling H, Miettinen TA. Cholesterol ab-
587 Lipidol. 2016;10:1240–1247.
643
sorption and synthesis in children with type 1 diabetes. Diabetes
588 12. Schwab KO, et al. Spectrum and prevalence of atherogenic risk fac- care. 2006;29:2300–2304.
644
589 tors in 27,358 children, adolescents, and young adults with type 1 34. Brown TL, Maahs DM, Bishop FK, Snell-Bergeon JK, Wadwa RP. In- 645
590 diabetes: cross-sectional data from the German diabetes documenta- fluences of gender on cardiovascular disease risk factors in adolescents 646
591 tion and quality management system (DPV). Diabetes care. 2006;29: with and without type 1 diabetes. Int J Pediatr Endocrinol. 2016;2016: 647
218–225. 8.
592 13. Bebu I, et al. The relationship of blood glucose with cardiovascular
648
35. Kallien G, Lange K, Stange EF, Scheibner J. The pravastatin-induced
593 disease is mediated over time by traditional risk factors in type 1 dia- decrease of biliary cholesterol secretion is not directly related to an in-
649
594 betes: the DCCT/EDIC study. Diabetologia. 2017;60:2084–2091. hibition of cholesterol synthesis in humans. Hepatology. 1999;30: 650
595 14. Guyton JR, et al. Achievement of recommended lipid and lipoprotein 14–20. 651
596 levels with combined ezetimibe/statin therapy versus statin alone in 36. Honda A, et al. Highly sensitive analysis of sterol profiles in human 652
patients with and without diabetes. Diab Vasc Dis Res. 2011;8: serum by LC-ESI-MS/MS. J Lipid Res. 2008;49:2063–2073.
597 160–172.
653
37. Pihlajamaki J, Gylling H, Miettinen TA, Laakso M. Insulin resistance
598 15. Winkler K, Jacob S, Muller-Schewe T, Hoffmann MM, Konrad T. Eze- is associated with increased cholesterol synthesis and decreased
654
599 timibe alone and in combination lowers the concentration of small, cholesterol absorption in normoglycemic men. J Lipid Res. 2004;45: 655
600 dense low-density lipoproteins in type 2 diabetes mellitus. Atheroscle- 507–512. 656
601 rosis. 2012;220:189–193. 38. Lupattelli G, et al. Visceral fat positively correlates with cholesterol 657
16. Lakoski SG, et al. Indices of cholesterol metabolism and relative synthesis in dyslipidaemic patients. Eur J Clin Invest. 2012;42:
602 responsiveness to ezetimibe and simvastatin. J Clin Endocrinol Metab.
658
164–170.
603 2010;95:800–809. 39. Hallikainen M, Kurl S, Laakso M, Miettinen TA, Gylling H. Plant sta-
659
604 17. John CC, Regier MD, Lilly CL, Aly S. Long-term pharmacotherapy nol esters lower LDL cholesterol level in statin-treated subjects with 660
605 for elevated low density lipoprotein levels in children: A retrospective type 1 diabetes by interfering the absorption and synthesis of choles- 661
606 analysis. J Clin Lipidol. 2016;10:265–272. terol. Atherosclerosis. 2011;217:473–478. 662

FLA 5.6.0 DTD  JACL1501_proof  16 October 2019  7:43 am

 Semova et al Cholesterol markers in type 1 diabetes 7

663 40. Sittiwet C, et al. Cholesterol metabolism and non-cholesterol sterol volunteers of a randomly selected male population. Am J Epidemiol. 719
664 distribution in lipoproteins of type 1 diabetes: the effect of improved 1990;131:20–31. 720
glycemic control. Atherosclerosis. 2007;194:465–472. 45. Tilvis RS, Miettinen TA. Serum plant sterols and their relation to
665 41. Nakamura K, Matsui T, Adachi H, Yamagishi S. Involvement of cholesterol absorption. Am J Clin Nutr. 1986;43:92–97.
721
666 angiotensin II in intestinal cholesterol absorption. Pharmacol Res. 46. Simonen P, Gylling H, Miettinen TA. The validity of serum squalene 722
667 2010;61:460–465. and non-cholesterol sterols as surrogate markers of cholesterol synthesis 723
668 42. Inoue T, et al. Inhibition of intestinal cholesterol absorption might and absorption in type 2 diabetes. Atherosclerosis. 2008;197:883–888. 724
669 explain cholesterol-lowering effect of telmisartan. J Clin Pharm 47. Mashnafi S, Plat J, Mensink RP, Baumgartner S. Non-cholesterol ste- 725
Ther. 2011;36:103–110. rol concentrations as biomarkers for cholesterol absorption and syn-
670 43. Miettinen TA, Tilvis RS, Kesaniemi YA. Serum cholestanol and plant thesis in different metabolic disorders: a systematic review.
726
671 sterol levels in relation to cholesterol metabolism in middle-aged men. Nutrients. 2019;11. Q7 727
672 Metabolism. 1989;38:136–140. 48. Bishop FK, Wadwa RP, Ellis S, Rewers M, Maahs DM. Lessons 728
673 44. Miettinen TA, Tilvis RS, Kesaniemi YA. Serum plant sterols and learned from a lipid lowering trial in adolescents with type 1 diabetes. 729
674 cholesterol precursors reflect cholesterol absorption and synthesis in Int J Pediatr Endocrinol. 2012;2012:24. 730
675 731
676 732
677 733
678 734
679 735
680 736
681 737
682 738
683 739
684 740
685 741
686 742
687 743
688 744
689 745
690 746
691 747
692 748
693 749
694 750
695 751
696 752
697 753
698 754
699 755
700 756
701 757
702 758
703 759
704 760
705 761
706 762
707 763
708 764
709 765
710 766
711 767
712 768
713 769
714 770
715 771
716 772
717 773
718 774

FLA 5.6.0 DTD  JACL1501_proof  16 October 2019  7:43 am