Epidemiology/Health Services/Psychosocial Research ORIGINAL ARTICLE
Insulin and Amylin Release Are Both Diminished in First-Degree Relatives of Subjects With Type 2 Diabetes
NEGAR G. KNOWLES, MA WILFRED Y. FUJIMOTO, MD ease these changes occur. A number of MELINDA A. LANDCHILD, MA STEVEN E. KAHN, MB, CHB studies have been largely interpreted to suggest that whereas insulin resistance exists in individuals who are at high risk of developing diabetes and have normal glucose tolerance, -cell function is not OBJECTIVE — To determine whether first-degree relatives of individuals with type 2 diabe-   diminished (6–9), or if -cell dysfunction tes, who are at high risk of subsequently developing hyperglycemia, manifest alterations in -cell is present, this dysfunction is only mild function including an alteration in the co-release of insulin and amylin. and is present only when impaired glu- RESEARCH DESIGN AND METHODS — In 30 first-degree relatives and 24 matched cose tolerance exists (10,11). However, subjects with no family history of diabetes, -cell function was measured as the intravenous these assessments have not accounted for glucose–induced acute insulin response (AIRg) and acute amylin response (AARg). The insulin the fact that insulin sensitivity is an im- sensitivity index (SI) was quantified and used to account for the role of insulin sensitivity to portant modulator of the -cell response  ϫ modulate -cell function (SI -cell function). to secretagogues, and therefore reduced -cell function may easily be overlooked RESULTS — Fasting plasma glucose (5.3 Ϯ 0.1 vs. 5.1 Ϯ 0.1 mmol/l; means Ϯ SEM), Ϯ Ϯ (12,13). Thus, when the effect of insulin immunoreactive insulin (IRI) (68 7 vs. 57 6 pmol/l) and amylin-like immunoreactivity  (ALI) (5.5 Ϯ 0.6 vs. 4.7 Ϯ 0.7 pmol/l) were similar in relatives and control subjects, respectively. sensitivity on -cell function is accounted Ϯ Ϯ for, a comparable insulin response could Relatives were insulin resistant compared with control subjects (SI: 4.86 0.63 vs. 7.20 ϫ Ϫ5 Ϫ1 Ϫ1 Ϫ1 ϭ Ϯ Ϯ 0.78 10 min pmol l , P 0.01), but their AIRg (392 59 vs. 386 50 pmol/l) in fact be considered inappropriately low Ϯ Ϯ  and AARg (5.9 0.9 vs. 6.1 0.8 pmol/l) did not differ. When -cell function was determined in the face of insulin resistance. Using this ϫ Ϯ relative to insulin sensitivity, in the first-degree relatives, both AIRg (SI AIRg: 1.60 0.23 vs. approach, it is now being recognized that Ϯ ϫ Ϫ2 Ϫ1 Ͻ ϫ Ϯ Ϯ ϫ 2.44 0.31 10 min , P 0.05) and AARg (SI AARg: 2.39 0.35 vs. 4.06 0.56 -cell function is relatively decreased in 10Ϫ4 minϪ1, P Ͻ 0.05) were reduced. The molar proportion of ALI to IRI was not altered in Ϯ Ϯ some groups at high risk of developing high-risk subjects (1.75 0.16 vs. 1.71 0.15%). hyperglycemia (14–18). Amylin is a 37–amino acid peptide CONCLUSIONS — First-degree relatives of subjects with type 2 diabetes have diminished  -cell function at a time when they are not hyperglycemic, and this reduction affects insulin and that is produced by the -cell and is co- amylin responses proportionally. Thus, an altered amylin-to-insulin ratio is not likely to identify secreted with insulin in response to glucose individuals at high risk of developing type 2 diabetes. and nonglucose secretagogues adminis- tered orally or intravenously (4,5,19,20). Diabetes Care 25:292–297, 2002 It is the unique constituent of the islet amyloid deposits found in the vast major- ity of subjects with type 2 diabetes (21– he pathogenesis of type 2 diabetes -cell peptide known as amylin or islet 23). We and others have hypothesized includes reductions in insulin sensi- amyloid polypeptide can also be demon- that alterations in the handling of this T tivity and -cell function (1). The strated in subjects with type 2 diabetes peptide by the -cell may underlie the change in -cell function has been dem- (3–5). propensity of diabetic individuals to de- onstrated to include a reduction in insulin Although the changes in insulin sen- posit amyloid, resulting in a reduction in secretion in response to glucose (2). In sitivity and -cell function have been -cell mass and the loss of -cell secretory addition to this reduction in insulin re- demonstrated to exist once hyperglyce- capacity (24). In keeping with a reduction lease, it is apparent that a decrease in the mia is present, it has been debated as to in -cell secretory capacity, we have ob- release of the more recently described when during the development of the dis- served reductions in insulin and amylin responses in individuals with impaired ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● glucose tolerance (5). Others have found From the Division of Metabolism, Endocrinology and Nutrition, Department of Medicine, VA Puget Sound that the amylin response is increased un- Health Care System, and the University of Washington, Seattle, Washington. Address correspondence and reprint requests to Steven E. Kahn, MB, ChB, VA Puget Sound Health Care der conditions associated with insulin System (151), 1660 S. Columbian Way, Seattle, WA 98108. E-mail: [email protected]. resistance, such as obesity (4,25) and Received for publication 27 July 2001 and accepted in revised form 4 November 2001. pregnancy (26), including in women with Abbreviations: AARg, intravenous glucose–induced acute amylin response; AIRg, intravenous glucose– gestational diabetes who are at high risk induced acute insulin response; ALI, amylin-like immunoreactivity; GEZI, glucose effectiveness at zero for subsequently developing type 2 diabe- insulin; IRI, immunoreactive insulin; SG, glucose effectiveness; SI, insulin sensitivity index. A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion tes. Thus, changes in amylin have been factors for many substances. observed under scenarios that are associ-
292 DIABETES CARE, VOLUME 25, NUMBER 2, FEBRUARY 2002 Knowles and Associates ated with an increased risk of developing Table 1—Demographic characteristics and fasting plasma glucose, IRI, and ALI concentra- type 2 diabetes. tions in first-degree relatives of type 2 diabetic subjects and in control subjects First-degree relatives of subjects with type 2 diabetes are at increased risk of First-degree relatives Control subjects developing hyperglycemia. They there- fore provide an ideal group to study to n 30 24 determine the magnitude of changes in Age (years) 39.8 Ϯ 1.7 42.0 Ϯ 3.1 insulin sensitivity and -cell function that BMI (kg/m2) 29.2 Ϯ 1.2 26.7 Ϯ 0.8 may be present in such high-risk individ- Fasting glucose (mmol/l) 5.3 Ϯ 0.1 5.1 Ϯ 0.1 uals. Furthermore, as they are at increased Fasting IRI (pmol/l) 68 Ϯ 757Ϯ 6 risk of developing hyperglycemia, they Fasting ALI (pmol/l) 5.5 Ϯ 0.6 4.7 Ϯ 0.7 also represent an appropriate cohort for Data are means Ϯ SEM. examining whether amylin responses are altered in high-risk subjects. Therefore, were obtained 2, 3, 4, 5, 6, 8, 10, 12, 14, as the insulin sensitivity index (SI) and we have examined a group of first-degree 16, 19, 22, 23, 24, 25, 27, 30, 35, 40, 50, glucose effectiveness at basal insulin (SG) relatives of subjects with type 2 diabetes 60, 70, 80, 90, 100, 120, 140, 160, 180, (28). Glucose effectiveness at zero insulin ϭ Ϫ ϫ and a control group and quantified insu- 200, 220, and 240 min after glucose ad- [GEZI SG (SI fasting IRI)] was cal- lin sensitivity and both insulin and amylin ministration. The administration of tol- culated as a measure of insulin-indepen- responses as measures of -cell function butamide and the prolonged sampling dent glucose disposal (31). The in these individuals to determine whether schedule served to improve the ability to intravenous glucose–induced acute insu- these variables are altered in these high- identify the parameters (27) when the lin (AIRg) and acute amylin (AARg)re- risk subjects. In the process, we also ex- glucose and insulin data were analyzed sponses were calculated as the mean amined whether differential alterations in using the minimal model of glucose kinet- incremental response above basal from insulin and amylin responses may pro- ics developed by Bergman et al. (28). All the samples drawn during the first 10 min vide an additional useful marker for sub- samples were assayed for glucose and in- after intravenous glucose administration. jects at high risk of developing type 2 sulin, whereas only the basal samples and The percentile ranking for each individu- diabetes. those drawn up to 10 min after glucose al’s product of SI and AIRg (disposition injection were assayed for amylin. index) was determined using the formula RESEARCH DESIGN AND ␣ϭ ϫ ϩ METHODS Assays Z [ln(SI AIRg) 3.802]/0.5613 All blood samples were drawn into tubes Subjects containing EDTA, kept on ice before sep- which was originally determined by our The groups consisted of 30 (9 males/21 aration, and subsequently stored at Ϫ70°C group in a cohort of 93 apparently healthy females) individuals with at least one first- before being assayed. Plasma glucose was subjects (13). degree relative with a known diagnosis of measured by an automated glucose oxi- The glucose disappearance constant type 2 diabetes and 24 (14 males/10 fe- dase method. Plasma immunoreactive was calculated as the slope of the regres- males) subjects who had no known family insulin (IRI) was measured by a radioim- sion line relating the natural log of the history of type 2 diabetes. No subject was munoassay that has inter- and intra-assay glucose concentration versus time for taking medications that are known to af- coefficients of variation of 12 and 8%, samples drawn between 10 and 19 min fect glucose metabolism. All subjects gave respectively. The antibody used in this as- after glucose administration. written informed consent before partici- say cross-reacts fully with intact proinsu- Statistical analysis was performed us- pating in the study, which had been re- lin and its conversion intermediates. ing Statview SE ϩ Graphics (Abacus Con- viewed and approved by the Human Plasma amylin-like immunoreactivity cepts, Berkeley, CA). Data are presented Subjects Review Committee at the Uni- (ALI) was quantified using a two-site en- as means Ϯ SEM. Comparisons between versity of Washington. zyme-linked immunoassay system devel- groups were done using an unpaired Stu- oped by Amylin Pharmaceuticals (San dent’s t test, except when variables were Study methods Diego, CA) using antibodies F002 and non-normally distributed, in which case Weight and height were measured and F025 (29). This assay measures both gly- the Mann-Whitney U test was performed. used to calculate BMI as weight (kg)/ cosylated and nonglycosylated forms of Correlations were performed by linear re- height2 (m2). the peptide (29,30). It has inter- and in- gression. A P value of Ͻ0.05 was consid- After a 10-h overnight fast, a tolbut- tra-assay coefficients of variation of Ͻ15 ered significant. amide-modified frequently sampled in- and Ͻ10%, respectively, with a minimum travenous glucose tolerance test was detectable concentration of 1.6 pmol/l. RESULTS performed to quantify insulin sensitivity IRI was measured in duplicate, whereas and the first-phase insulin and amylin re- glucose and ALI were measured in tripli- Subject characteristics and fasting sponses. After basal sampling, glucose cate. measures (11.4 g/m2) was administered intrave- Age, BMI, fasting plasma glucose, IRI, and nously over 60 s, and 20 min later, tolbu- Calculations and statistical analysis ALI did not differ significantly between tamide (125 mg/m2) was injected Using the minimal model of glucose ki- the 30 first-degree relatives and 24 con- intravenously over 30 s. Blood samples netics, insulin sensitivity was quantified trol subjects (Table 1).
DIABETES CARE, VOLUME 25, NUMBER 2, FEBRUARY 2002 293 Insulin and amylin release in first-degree relatives
Insulin sensitivity, -cell function, and intravenous glucose tolerance SI, which was quantified using the fre- quently sampled intravenous glucose tol- erance test and minimal model of glucose kinetics, indicated that the first-degree relatives were insulin resistant. In the rel- Ϫ Ϫ atives, S was 4.86 Ϯ 0.63 ϫ 10 5 min 1 ϪI Ϫ pmol 1 l 1, whereas in the control sub- Ϫ Ϫ jects, it was 7.20 Ϯ 0.78 ϫ 10 5 min 1 Ϫ Ϫ pmol 1 l 1 (P Ͻ 0.05). The first-phase IRI (AIRg) and ALI (AARg) secretory re- sponses, determined over the first 10 min after intravenous glucose administration, did not differ in absolute terms between the two groups. In the first-degree rela- Ϯ tives, AIRg was 392 59 pmol/l and Ϯ AARg was 5.9 0.9 pmol/l, whereas in the control subjects, these measures were 386 Ϯ 50 and 6.1 Ϯ 0.8 pmol/l, respec- tively.  ϫ We determined whether the -cell re- Figure 1—Individual values for percentile scores for the disposition index (SI AIRg)in30 sponses were appropriate by calculating first-degree relatives of subjects with type 2 diabetes and 24 apparently healthy control subjects. Ϯ the product of SI and the acute insulin and The means SEM for the two groups are illustrated as is the 50th percentile. Among the relatives, ϫ ϫ 21 of the 30 (70%) individuals fall below the 50th percentile, whereas in the apparently healthy amylin responses (SI AIRg and SI control subjects, 12 out of 24 (50%) do so. The percentile score was determined based on the AARg). When the data were examined in this way, it was clear that the -cell re- formula derived in a separate healthy population (13). sponses in the first-degree relatives were not normal but were proportionally re- constant, tended to be reduced in the whole cohort and when they were subdi- duced. In these high-risk subjects, S ϫ first-degree relatives compared with the vided into the two groups. As illustrated Ϫ I Ϫ AIR was 1.60 Ϯ 0.23 ϫ 10 2 min 1 control subjects (1.64 Ϯ 0.12 vs. 1.84 Ϯ in Fig. 2B, there was a great degree of g Ϫ compared with 2.44 Ϯ 0.31 ϫ 10 2 0.12%/min; P ϭ 0.08). GEZI, a measure overlap in absolute AAR and AIR be- Ϫ g g min 1 in the control subjects (P Ͻ 0.05). of insulin-independent glucose disposal, tween individuals in both groups. For the Similarly, when -cell function was quan- did not differ significantly between the whole study group, these two responses ϫ Ϯ Ϫ1 ϭ tified as SI AARg, this measure was also two groups, being 0.016 0.001 min were strongly linearly related (r 0.89; significantly lower in the first-degree rel- in the first-degree relatives and 0.020 Ϯ P Ͻ 0.0001), and the results were not Ϫ Ϫ Ϫ Ϫ atives, being 2.39 Ϯ 0.35 ϫ 10 4 min 1 0.002 ϫ 10 2 min 1 in the control sub- different for the first-degree relatives (r ϭ Ϫ Ϫ versus 4.06 Ϯ 0.56 ϫ 10 4 min 1 in the jects (P Ͼ 0.1). 0.91; P Ͻ 0.0001) and control subjects control subjects (P Ͻ 0.05). (r ϭ 0.86; P Ͻ 0.0001). Again, the rela- To determine the distribution of the Relationship between amylin and tionship between these two parameters ϫ disposition index (SI AIRg) in the two insulin responses did not differ between the relatives and groups of subjects, we determined the Using linear regression, we assessed control subjects. Each individual’s AIRg percentile score for each subject based on whether the relationship between fasting and AARg were used to calculate the mo- ␣ the Z derived from SI and AIRg values in ALI and IRI levels was different between lar proportions of these responses for the a healthy population (13). As illustrated first-degree relatives and control subjects. two groups. The AIRg/AARg molar ratio in Fig. 1, although there was considerable As illustrated in Fig. 2A, these fasting val- did not differ between the first-degree rel- overlap in the percentile scores, the dis- ues were linearly correlated in the whole atives in whom it was 1.75 Ϯ 0.16% com- tribution of percentile scores differed be- population (r ϭ 0.42; P Ͻ 0.005). When pared with 1.71 Ϯ 0.15% in the control tween the two groups of subjects, with 21 the two groups were examined indepen- subjects. of the 30 relatives falling below the 50th dently, the relationship was not different percentile compared with 12 of the 24 between first-degree relatives (r ϭ 0.38; CONCLUSIONS — Many studies ex- control subjects. Thus, the mean percen- P Ͻ 0.05) and control subjects (r ϭ 0.49; amining the pathogenesis of type 2 diabe- tile score for the first-degree relatives was P Ͻ 0.05). The ratio of fasting ALI to fast- tes used absolute measures of insulin 29.6 Ϯ 6.0% compared with 47.7 Ϯ ing IRI did not differ between the two sensitivity and insulin responses and con- 6.7% in the control subjects (P Ͻ 0.05). groups, being 9.5 Ϯ 1.4% in the first- cluded that insulin resistance is the pri- As would be expected in a group of degree relatives versus 8.7 Ϯ 1.1% in the mary defect, especially when glucose individuals in whom both insulin sensi- control subjects. tolerance is still normal (6–11). However, tivity and insulin secretion are dimin- We also determined whether the based on findings in healthy control sub- ished, intravenous glucose tolerance, stimulated amylin (AARg) and insulin jects, the concept of a feedback loop be- determined as the glucose disappearance (AIRg) responses were related in the tween the insulin-sensitive tissues and the
294 DIABETES CARE, VOLUME 25, NUMBER 2, FEBRUARY 2002 Knowles and Associates
these high-risk subjects did not have dia- betes. The concept that these parameters may be reduced without a marked deteri- oration in glucose tolerance is apparent from studies of other groups at high risk for diabetes who have abnormalities in both insulin secretion and insulin sensi- tivity at a time when they have relatively normal glucose tolerance (14,17,18). These findings highlight the fact that more marked alterations in both insulin sensitivity and -cell function must exist for impaired glucose tolerance or diabetes to exist, and emphasize the importance of insulin-independent glucose disposal to glucose tolerance (34,35). Amylin is a secretory product of the -cell that is typically co-released with in- sulin (36). As is the case with insulin, in healthy subjects, the relationship be- tween insulin sensitivity and amylin levels is hyperbolic (37). Thus, by taking into account differences in insulin sensitivity, we were also able to demonstrate a reduc- tion in the amylin response in the first- degree relatives. It could be argued that it is not optimal to assess the amylin re- sponse in relationship to insulin sensitiv- ity because insulin sensitivity may not be a modulator of amylin release by the -cell. However, because insulin and amylin are packaged together in -cell se- cretory granules (38) and are typically co- released from the regulated secretory pathway in response to classic -cell secretagogues (36), because amylin levels have been shown to be increased in insu- Figure 2—Relationship between (A) the fasting values for IRI and ALI and (B) AIRg and AARg in first-degree relatives of subjects with type 2 diabetes (F) and apparently healthy control subjects lin-resistant states such as obesity ( ). The linear correlation for the whole cohort for both sets of variables is illustrated (fasting: (4,5,25), and because the relationship be- r ϭ 0.42; P Ͻ 0.005; stimulated: r ϭ 0.89; P Ͻ 0.0001). It is apparent that this relationship does tween amylin and insulin sensitivity is hy- not differ between groups. perbolic (37), it seems reasonable to assume that the effect of insulin sensitivity islet has emerged (12,13,17,32). Thus, it is applicable to populations studied to modulate the release of insulin- the relationship between insulin sensitiv- with different insulin assays. The utility of containing granules would apply to amy- ity and -cell function is hyperbolic, and this formula has also been demonstrated lin as well. the product of insulin sensitivity and by Elbein et al. (33), who demonstrated In addition, the molar proportion of -cell function, known as the disposition the heritability of the disposition index to amylin to insulin, both fasting and stim- index, is a constant. be 70% in subjects with normal glucose ulated, was not different between high- Using the disposition index as a mea- tolerance who have a sibling with type 2 risk and control subjects. The higher ratio sure of -cell function, first-degree rela- diabetes. These data raise the interesting during fasting is because amylin clearance tives had responses that ranged between possibility as to whether assessment of the is slower than that of insulin (25). Our the 1st and 88th percentiles, with 70% of percentile score for individuals at high findings differ from that of Kautzky- the first-degree relatives falling below the risk of developing type 2 diabetes could Willer et al. (26), who reported that preg- mean (50th percentile), whereas 12 con- possibly be used as a tool for predicting nancy was associated with an increase in trol subjects were above and 12 were be- subjects who are at risk. Such a possibility this ratio. It must be noted that our ap- low the 50th percentile. Because will require longitudinal studies of such proach does not allow us to discern small calculation of the percentile score with high-risk individuals. differences that may be identifiable if we this formula is independent of differences It is of interest that in the first-degree were sampling directly at the level of the in insulin assays, provided both high and relatives, impairments of both insulin pancreas. That such differences might oc- low insulin values are reliably measured, sensitivity and -cell function existed, yet cur is supported not only by the study of
DIABETES CARE, VOLUME 25, NUMBER 2, FEBRUARY 2002 295 Insulin and amylin release in first-degree relatives pregnant women (26) but also from mod- relatives of subjects with a history of type 4. Enoki S, Mitsukawa T, Takemura J, Naka- els in which more proximal sampling has 2 diabetes are both insulin resistant and zato M, Aburaya J, Toshimori H, Matsu- been performed in vitro and in situ (39– have impaired insulin and amylin re- kara S: Plasma islet amyloid polypeptide 41). However, based on our data, the cir- sponses to this insulin resistance. levels in obesity, impaired glucose toler- culating proportion of amylin and insulin In summary, we found that appar- ance and non-insulin-dependent diabetes mellitus. Diabetes Res Clin Pract 15:97– probably will not be useful as a clinical ently healthy first-degree relatives of sub-  102, 1992 marker of -cell dysfunction. jects with type 2 diabetes clearly have 5. Kahn SE, Verchere CB, Andrikopoulos S, The present finding of reductions in defects in insulin sensitivity and -cell Asberry PJ, Leonetti DL, Wahl PW, Boyko both insulin and amylin responses in function. The latter abnormality involves EJ, Schwartz RS, Newell-Morris L, Fuji- high-risk subjects suggests that the early reductions in first-phase insulin and amy- moto WY: Reduced amylin release is a changes in -cell function are fairly gen- lin release in response to glucose stimula- characteristic of impaired glucose toler- eralized, affecting the release of both these tion. However, these reductions are not ance and type 2 diabetes in Japanese peptides. Because amylin is the unique evident from the absolute responses and Americans. Diabetes 47:640–645, 1998 peptide constituent of the islet amyloid require a simultaneous assessment of in- 6. Haffner S, Stern M, Hazuda H, Mitchell B, deposits that are typically observed in in- sulin sensitivity. When the modulating Patterson J: Increased insulin concentra-  tions in nondiabetic offspring of diabetic dividuals with type 2 diabetes (21–23), it effect of insulin sensitivity on -cell func- parents. N Engl J Med 319:1297–1301, has been suggested that increased amylin tion is taken into account, the defect is 1988 release may play a role in the develop- obvious. Finally, we have found that de- 7. Lillioja S, Mott DM, Howard BV, Bennett ment of these deposits. The present find- spite the differences in insulin sensitivity PH, Yki-Jarvinen H, Freymond D, ings suggest the possibility that in high- and the reduction in -cell function, as Nyomba BL, Zurlo F, Swinburn B, Bogar- risk individuals, the absolute amylin determined by the insulin and amylin re- dus C: Impaired glucose tolerance as a responses and the proportion relative to sponses, the release of these two peptides disorder of insulin action: longitudinal insulin may not differ from those in low- is proportionate. Therefore, determina- and cross-sectional studies in Pima Indi- risk subjects. However, it is also clear tion of the relative proportions of amylin ans. N Engl J Med 318:1217–1225, 1988 from this study that, at the time we stud- and insulin is unlikely to be an additional 8. Warram JH, Martin BC, Krolewski AS,  Soeldner JS, Kahn CR: Slow glucose re- ied these subjects, the relative responses useful marker of -cell function in humans. moval rate and hyperinsulinemia precede were reduced. Thus, the question is still the development of type II diabetes in the open as to whether before the stage at offspring of diabetic parents. Ann Intern which we studied these subjects the abso- Acknowledgments— This work was sup- Med 113:909–915, 1990 lute output of amylin was increased and ported in part by National Institutes of Health 9. Martin BL, Warram JH, Krolewski AS, contributed to the initiation of amyloid Grants DK-02654, DK-17047, DK-50703, Bergman RN, Soeldner JS, Kahn CR: Role fibril formation. and RR-37; by the Office of Research and De- of glucose and insulin resistance in devel- velopment, Medical Research Service, Depart- The present study was performed in opment of type 2 diabetes mellitus: results ment of Veterans Affairs; and by the American of a 25-year follow-up study. Lancet 340: groups that were matched for BMI and Diabetes Association. N.G.K. and M.A.L. per- age, two parameters recognized to affect 925–929, 1992 formed this work as medical students sup- 10. Eriksson J, Franssila Kallunki A, Ekstrand insulin sensitivity (42,43). However, ported by the American Federation for Aging A, Saloranta C, Widen E, Schalin C, other factors such as body fat distribution Research. The amylin kits used in this study Groop L: Early metabolic defects in per- that may affect insulin sensitivity were not were a gift from Amylin Pharmaceuticals. sons at increased risk for non-insulin-de- assessed. Although such parameters may We are grateful to Valerie Larson for the care pendent diabetes mellitus. N Engl J Med have differed between the two groups, we of the subjects and Maggie Abrahamson, Dor- 321:337–343, 1989 do not believe these parameters would othy Winch, and Ruth Hollingsworth for tech- 11. Lillioja S, Mott DM, Spraul M, Ferraro R, have affected the outcome of these studies nical assistance. We wish to thank Daniel Foley JE, Ravussin E, Knowler WC, Ben- Porte Jr., Josep Vidal, and Miriam Cnop for because they affect insulin sensitivity, nett PH, Bogardus C: Insulin resistance useful discussions during the preparation of and insulin secretory dysfunction as pre- which we directly quantified, and related this manuscript. to parameters of -cell function. In our cursors of non-insulin-dependent diabe- tes mellitus: prospective studies of Pima assessments, we used an insulin assay that Indians. N Engl J Med 329:1988–1992, also recognizes proinsulin. Proinsulin References 1993 levels were shown to be disproportion- 1. Kahn SE, Porte D Jr: Pathophysiology of 12. Bergman RN, Phillips LS, Cobelli C: Phys- ately increased in subjects with type 2 di- type II diabetes mellitus. In Diabetes Mel- iologic evaluation of factors controlling abetes (44–47), and a small increase was litus. Porte D Jr, Sherwin RS, Eds. Stam- glucose tolerance in man: measurement of also demonstrated in individuals at high ford, CT, Appleton & Lange, 1996, p. insulin sensitivity and -cell glucose sen- risk of developing the disease (48,49). 487–512 sitivity from the response to intravenous  Whereas we have quantified the insulin 2. Porte D Jr: -Cells in type II diabetes mel- glucose. J Clin Invest 68:1456–1467, response shortly after granule exocytosis, litus. Diabetes 40:166–180, 1991 1981 any effect of the slower rate of clearance of 3. Sanke T, Hanabusa T, Nakano Y, Oki C, 13. Kahn SE, Prigeon RL, McCulloch DK, Okai K, Nishimura S, Kondo M, Nanjo K: Boyko EJ, Bergman RN, Schwartz MW, proinsulin would have made it more dif- Plasma islet amyloid polypeptide (amy- Neifing JL, Ward WK, Beard JC, Palmer ficult to demonstrate a difference between lin) levels and their responses to oral glu- JP, Porte D Jr: Quantification of the rela- first-degree relatives and control subjects. cose in type 2 (non-insulin-dependent) tionship between insulin sensitivity and Thus, we believe that the findings of the diabetic patients. Diabetologia 34:129– B-cell function in human subjects: evi- present study highlight that first-degree 132, 1991 dence for a hyperbolic function. Diabetes
296 DIABETES CARE, VOLUME 25, NUMBER 2, FEBRUARY 2002 Knowles and Associates
42:1663–1672, 1993 B, Nowotny P, Rabensteiner D, Wald- the simple explanation for inefficient pro- 14. Ryan EA, Imes S, Liu D, McManus R, hausl W, Pacini G, Prager R: Elevated islet insulin processing and islet amyloid for- Finegood DT, Polonsky KS, Sturis J: De- amyloid pancreatic polypeptide and pro- mation in type 2 diabetes. J Invest Med fects in insulin secretion and action in insulin in lean gestational diabetes. Diabe- 48:21A, 2000 women with a history of gestational dia- tes 46:607–614, 1997 38. Lukinius A, Wilander E, Westermark GT, betes. Diabetes 44:506–512, 1995 27. Prigeon RL, Kahn SE, Porte D Jr: Reliabil- Engstrom U, Westermark P: Co-localiza- 15. Larsson H, Ahren B: Failure to adequately ity of error estimates from the minimal tion of islet amyloid polypeptide and in- adapt reduced insulin sensitivity with in- model: implications for measurements in sulin in the B cell secretory granules of the creased insulin secretion in women with physiological studies. Am J Physiol 266: human pancreatic islets. Diabetologia 32: impaired glucose tolerance. Diabetologia E279–E286, 1994 240–244, 1989 39:1099–1107, 1996 28. Bergman RN, Ider YZ, Bowden CR, Co- 39. O’Brien TD, Westermark P, Johnson KH: 16. Cavaghan MK, Ehrmann DA, Byrne MM, belli C: Quantitative estimation of insulin Islet amyloid polypeptide and insulin se- Polonsky KS: Treatment with the oral an- sensitivity. Am J Physiol 236:E667–E677, cretion from isolated perfused pancreas tidiabetic agent troglitazone improves 1979 of fed, fasted, glucose-treated, and dexa- beta cell responses to glucose in subjects 29. Percy AJ, Trainor DA, Rittenhouse J, methasone-treated rats. Diabetes 40:1701– with impaired glucose tolerance. J Clin In- Phelps J, Koda JE: Development of sensi- 1706, 1991 vest 100:530–537, 1997 tive immunoassays to detect amylin and 40. Gedulin B, Cooper GJS, Young AA: Amy- 17. Weyer C, Bogardus C, Mott DM, Pratley amylin-like peptides in unextracted lin secretion from the perfused pancreas: RE: The natural history of insulin secre- plasma. Clin Chem 42:576–585, 1996 dissociation from insulin and abnormal tory dysfunction and insulin resistance in 30. Rittenhouse J, Chait BT, Bierle JR, Janes elevation in insulin-resistant diabetic rats. the pathogenesis of type 2 diabetes melli- SM, Park DR, Phelps JL, Fineman MS, Qin Biochem Biophys Res Commun 180:782– tus. J Clin Invest 104:787–794, 1999 J, Koda JE: Heterogeneity of naturally-oc- 789, 1991 18. Buchanan TA, Xiang AH, Kjos SL, Trigo E, curring human amylin due to glycosyla- 41. Kahn SE, Verchere CB, D’Alessio DA, Lee WP, Peters RK: Antepartum predic- tion. Diabetes 45 (Suppl. 2):235A, 1996 Cook DL, Fujimoto WY: Evidence for se- tors of the development of type 2 diabetes 31. Kahn SE, Klaff LJ, Schwartz MW, Beard lective release of rodent islet amyloid in Latino women 11–26 months after JC, Bergman RN, Taborsky GJ Jr, Porte D polypeptide through the constitutive se- pregnancies complicated by gestational Jr: Treatment with a somatostatin analog cretory pathway. Diabetologia 36:570–573, diabetes. Diabetes 48:2430–2436, 1999 decreases pancreatic B-cell and whole 1993 19. Butler PC, Chou J, Carter WB, Wang Y-N, body sensitivity to glucose. J Clin Endocri- 42. DeFronzo RA: Glucose intolerance of ag- Bu B-H, Chang D, Chang J-K, Rizza RA: nol Metab 71:994–1002, 1990 ing: evidence for tissue insensitivity to in- Effects of meal ingestion on plasma amy- 32. Clausen JO, Borch-Johnsen K, Ibsen H, sulin. Diabetes 28:1095–1101, 1979 lin concentration in NIDDM and nondia- Bergman RN, Hougaard P, Winther K, 43. Chen M, Bergman RN, Pacini G, Porte D betic humans. Diabetes 39:752–756, 1990 Pedersen O: Insulin sensitivity index, Jr: Pathogenesis of age-related glucose in- 20. Dechenes CJ, Verchere CB, Andrikopou- acute insulin response, and glucose effec- tolerance in man: insulin resistance and los S, Kahn SE: Human aging is associated tiveness in a population-based sample of decreased -cell function. J Clin Endocri- with parallel reductions in insulin and 380 young healthy Caucasians: analysis of nol Metab 60:13–20, 1985 amylin release. Am J Physiol 275:E785– the impact of gender, body fat, physical 44. Mako ME, Starr JI, Rubenstein AH: Circu- E791, 1998 fitness, and life-style factors. J Clin Invest lating proinsulin in patients with maturi- 21. Westermark P, Wernstedt C, Wilander E, 98:1195–1209, 1996 ty-onset diabetes. Am J Med 63:865–869, Hayden DW, O’Brien TD, Johnson KH: 33. Elbein SC, Hasstedt SJ, Wegner K, Kahn 1977 Amyloid fibrils in human insulinoma and SE: Heritability of pancreatic beta-cell 45. Temple RC, Carrington CA, Luzio SD, islets of Langerhans of the diabetic cat are function among nondiabetic members of Owens DR, Schneider AE, Sobey WJ, derived from a neuropeptide-like protein Caucasian familial type 2 diabetic kin- Hales CN: Insulin deficiency in non-insu- also present in normal islets. Proc Natl dreds. J Clin Endocrinol Metab 84:1398– lin-dependent diabetes. Lancet i:293– Acad SciUSA84:3881–3885, 1987 1403, 1999 295, 1989 22. Cooper GJS, Willis AC, Clark A, Turner 34. Kahn SE, Prigeon RL, McCulloch DK, 46. Saad MF, Kahn SE, Nelson RG, Pettitt DJ, RC, Sim RB, Reid KBM: Purification and Boyko EJ, Bergman RN, Schwartz MW, Knowler WC, Schwartz MW, Kowalyk S, characterization of a peptide from amy- Neifing JL, Ward WK, Beard JC, Palmer J, Bennett PH, Porte D Jr: Disproportion- loid-rich pancreases of type 2 diabetic pa- Porte D Jr: The contribution of insulin de- ately elevated proinsulin in Pima Indians tients. Proc Natl Acad SciUSA84:8628– pendent and insulin independent glucose with noninsulin-dependent diabetes mel- 8632, 1987 uptake to intravenous glucose tolerance litus. J Clin Endocrinol Metab 70:1247– 23. Clark A, Saad MF, Nezzer T, Uren C, in healthy human subjects. Diabetes 43: 1253, 1990 Knowler WC, Bennett PH, Turner RC: Is- 587–592, 1994 47. Kahn SE, Halban PA: Release of incom- let amyloid polypeptide in diabetic and 35. Best JD, Kahn SE, Ader M, Watanabe RM, pletely processed proinsulin is the cause non-diabetic Pima Indians. Diabetologia Ni TC, Bergman RN: Role of glucose ef- of the disproportionate proinsulinemia of 33:285–289, 1990 fectiveness in the determination of glu- NIDDM. Diabetes 46:1725–1732, 1997 24. Porte D Jr, Kahn SE: Hyperproinsuline- cose tolerance. Diabetes Care 19:1018– 48. Kahn SE, Leonetti DL, Prigeon RL, Berg- mia and amyloid in NIDDM: clues to eti- 1030, 1996 strom RW, Fujimoto WY: Proinsulin as a ology of islet -cell dysfunction? Diabetes 36. Kahn SE, D’Alessio DA, Schwartz MW, marker for the development of NIDDM in 38:1333–1336, 1989 Fujimoto WY, Ensinck JW, Taborsky GJ Japanese-American men. Diabetes 44: 25. Kautzky-Willer A, Thomaseth K, Pacini Jr, Porte D Jr: Evidence of cosecretion of 173–179, 1995 G, Clodi M, Ludvik B, Streli C, Waldhausl islet amyloid polypeptide and insulin by 49. Mykkanen L, Haffner SM, Kuusisto J, W, Prager R: Role of islet amyloid -cells. Diabetes 39:634–638, 1990 Pyorala K, Hales CN, Laakso M: Serum polypeptide secretion in insulin-resistant 37. Landchild MJ, Knowles NG, Yao Q, Wal- proinsulin levels are disproportionately humans. Diabetologia 37:188–194, 1994 den CE, Retzlaff BM, Knopp RH, Kahn SE: increased in elderly prediabetic subjects. 26. Kautzky-Willer A, Thomaseth K, Ludvik Increased -cell secretory demand is not Diabetologia 38:1176–1182, 1995
DIABETES CARE, VOLUME 25, NUMBER 2, FEBRUARY 2002 297