Insulin resistance is a key pathophysiologic feature of type by which chromium (Cr) supplementation affects insulin ac-
2 diabetes and is strongly associated with coexisting cardio- tion in vivo are currently unknown and an understanding of
vascular risk factors and accelerated atherosclerosis (1). As a these mechanisms would be required before firm recommen-
consequence, one of the most desirable goals of treatment for dations could be made regarding its routine use in the man-
patients with type 2 diabetes is increasing insulin sensitivity in agement of type 2 diabetes.
vivo. Although energy restriction and exercise greatly improve Cr use by the general public, and in diabetic patients in
insulin resistance, the long-term success of dietary interven- particular, has surpassed our ability as a scientific community
tion in humans is poor (2). Therefore, strategies to improve to provide evidence regarding its safety and efficacy. Part of
insulin resistance by pharmacologic means or nutritional sup- the problem stems from the lack of definitive, randomized
plementation represent a very attractive approach. Dietary trials because many of the earlier studies evaluating Cr use
supplementation with chromium picolinate (CrPic)2 has been were open-label studies and, therefore, generated substantial
proposed as one such nutritional intervention (3– 6). How- bias. Additional concerns are the lack of “gold standard”
ever, routine use of CrPic in subjects with diabetes is not techniques to assess glucose metabolism, the use of differing
currently recommended, and, indeed, the most recent clinical doses and formulations, and heterogeneous study populations.
practice recommendations from the American Diabetes Asso- As a result, a large body of conflicting data has been reported
ciation state that “chromium supplementation has no known that contributes greatly to the confusion among health care
benefit” (2). Further, the cellular and molecular mechanisms providers regarding use of Cr.
Several lines of evidence in both rodent and human studies,
however, suggest that Cr may modulate intracellular pathways
1
To whom correspondence should be addressed. E-mail: William.
of glucose metabolism and improve comorbidities associated
Cefalu@uvm.edu. with insulin resistance (3–9). Thus, the overall objective of
2
Abbreviations used: AUC, area(s) under the curve; CrPic, chromium picoli- this study was to evaluate the role of CrPic in improving the
nate; ECL, enhanced chemiluminescence; Glut, glucose transporter; HbA1c, clinical sequelae of the insulin resistance syndrome (e.g., dys-
hemoglobin A1c; ITT, insulin tolerance test; IPGTT, intraperitoneal glucose toler-
ance test; JCR, LA-cp; JCR, LA-corpulent; PMSF, phenylmethylsulfonyl fluoride; lipidemia, glucose intolerance, hyperinsulinemia) by use of a
TPN, total parenteral nutrition. rat model of insulin resistance.
1107
1108 CEFALU ET AL.
MATERIALS AND METHODS 120 min postglucose by tail cut (18). Insulin and glucose levels were
measured at each time point and the areas under the curve (AUC)
Study design were then determined.
The effect of CrPic was assessed in the JCR:LA-corpulent (JCR: Insulin tolerance test (ITT). An ITT was conducted before rats
LA-cp) rat, a strain incorporating the autosomal recessive cp gene were killed. After induction of anesthesia, a baseline tail cut was
that induces obesity (10,11). The JCR:LA-cp rat, when homozygous obtained, followed by intraperitoneal injection of regular insulin (5
U/kg) at time 0. Repeat tail cuts occurred at 5, 10, 15 and 30 min and
for the autosomal recessive cp gene (cp/cp), lacks membrane-bound
leptin receptors, leading to marked obesity (12). The cp/cp rats are then the rat was killed. The rate of glucose disappearance [mmol/
hyperphagic, become insulin resistant, hyperinsulinemic and hyper- (L 䡠 min)] was determined.
triglyceridemic, and develop advanced atherosclerotic disease as well For both the IPGTT and the ITT, rats inhaled 5% halothane in
as myocardial lesions consistent with an ischemic origin (13,14). The 100% oxygen via a facemask for 3– 4 min at a flow rate of 1.25 L/min,
hyperinsulinemia develops rapidly after 4 wk of age, with an age at then reduced to 2% in 100% oxygen. This method of anesthesia
half-maximum of 5.5 wk. Breeding is done using heterozygous rats allows the rats to recover completely between tail cuts and has been
(cp/⫹) and yields 25% obese rats (cp/cp) and 75% lean rats [a 2:1 mix shown to have minimal effects on insulin and glucose levels (19,20).
of cp/⫹ and ⫹/⫹ referred to as ⫹/?; for review see (15)]. Hyperten- Plasma was obtained at baseline, at wk 6 of treatment and at the
sion does not develop in this strain, thus providing a rat model of end of the study (12 wk) for determination of glucose and insulin
spontaneous cardiovascular disease that exhibits all of the aspects levels. A lipid profile was obtained at baseline and at wk 12 (end of
seen in obese, insulin-resistant humans, including vasculopathy, but study). Glucose was determined by an enzymatic method using the
without the confounding effects of hypertension. In addition to the Cobs Mira autoanalyzer (Roche Biomedical, Nutley, NJ). The CV for
Statistical analysis
Data were analyzed by repeated measures 2-way ANOVA using
the Scheffé F-test for post-hoc analysis. AUC for glucose tolerance
and insulin response were determined using the trapezoidal rule (24).
RESULTS
There was no effect of CrPic on daily food and water
intakes or body weights over the 90-d treatment period (Table
1). On the basis of the measured food and water intakes, the
control groups (both lean and obese) had daily elemental Cr
intakes ranging from 16 to 20 g/kg. The Cr-supplemented
groups had daily elemental Cr intakes ranging from 33 to 38
g/kg.
TABLE 1
Oral chromium picolinate (Cr-Pic) in JCR:LA corpulent rats that consumed water (control)
or Cr-Pic for 12 wk: weight, food, and water intake during study1
Lean Rats
Control 5 368.8 ⫾ 9.8 379.0 ⫾ 8.1 393.2 ⫾ 7.1 21.0 ⫾ 1.6 20.8 ⫾ 0.9 20.0 ⫾ 0.6 25.3 ⫾ 1.4 26.6 ⫾ 1.0 23.5 ⫾ 0.9
CrPic 5 374.2 ⫾ 9.5 392.0 ⫾ 10.1 413.6 ⫾ 6.8 21.5 ⫾ 1.3 21.1 ⫾ 0.6 20.0 ⫾ 0.3 26.1 ⫾ 0.5 27.7 ⫾ 1.0 26.0 ⫾ 0.6
Obese Rats
Control 5 644.4 ⫾ 21.9 695.0 ⫾ 23.1 737.5 ⫾ 24.2 28.6 ⫾ 1.6 29.6 ⫾ 1.0 29.8 ⫾ 1.1 35.6 ⫾ 1.6 37.6 ⫾ 1.8 35.7 ⫾ 0.7
CrPic 6 629.6 ⫾ 13.0 679.0 ⫾ 17.3 733.8 ⫾ 14.5 25.2 ⫾ 1.5 27.8 ⫾ 1.1 28.3 ⫾ 0.6 24.4 ⫾ 1.5 32.6 ⫾ 1.9 31.2 ⫾ 1.8
TABLE 2
Oral chromium picolinate (Cr-Pic) in JCR:LA corpulent rats
that consumed water (control) or Cr-Pic for 12 wk: lipids and
membrane-associated glucose transporter (Glut)-41
Membrane-
Cholesterol/HDL cholesterol associated
ratio Glut-42
Lean rats
Control 5 1.76 ⫾ 0.04 1.88 ⫾ 0.02 137.2 ⫾ 7.6
FIGURE 2 Insulin response of JCR:LA corpulent rats that con- CrPic 5 1.70 ⫾ 0.05 1.76 ⫾ 0.10 132.8 ⫾ 3.9
sumed water (control) or chromium picolinate (CrPic) for 12 wk to Obese rats
intraperitoneal glucose tolerance test (IPGTT) as assessed with area Control 5 2.51 ⫾ 0.18 3.19 ⫾ 0.35 93.8 ⫾ 6.9
under the curve (AUC). Values are means ⫾ SEM, n ⫽ 5– 6. *Different CrPic 6 2.56 ⫾ 0.22 1.86 ⫾ 0.10* 142.4 ⫾ 6.0**
from obese control, P ⬍ 0.001.
FIGURE 3 Plasma glucose response of obese JCR:LA corpulent FIGURE 4 Total plasma cholesterol levels in lean and obese
rats that consumed water (control) or chromium picolinate (CrPic) for 12 JCR:LA corpulent rats that consumed water (control) or chromium
wk to intraperitoneal glucose tolerance testing. Values are means picolinate (CrPic) for 12 wk for each treatment group over the course of
⫾ SEM, n ⫽ 5– 6. Plasma glucose was lower in treated rats at each time study. Values are means ⫾ SEM, n ⫽ 5– 6. *Different from obese control,
point , P ⬍ 0.0001. P ⬍ 0.01.
CHROMIUM AND INSULIN ACTION 1111
different hematological matrices in 53 subjects with type 2 metabolism; they used differing doses and formulations, eval-
diabetes compared with 50 controls; they reported significantly uated heterogeneous study populations and had widely varying
lower Cr levels in the plasma of the diabetic subjects compared periods of observation. Indeed, one study by Ravina et al. (42)
with the nondiabetic healthy controls. In contrast, Zima et al. suggested an effect that was observed after only 7 d of admin-
(40) suggested no alteration of Cr levels in type 2 diabetes; istration. Thus, the many confounders make these studies
however, only 11 subjects with type 2 diabetes were compared difficult to interpret when trying to suggest a consistent effect
with 19 healthy controls. If clinical states such as diabetes are of supplemental Cr on human carbohydrate metabolism. It
truly shown to be associated with diminished Cr levels, and if appears, however, that studies that specifically evaluated ⱕ200
supplementation generally leads to an increase in Cr concen- g of Cr as Cr chloride (CrCl) did not elicit a clinical response
tration, it is possible that diabetic patients may have inade- in subjects with type 2 diabetes (Table 3), whereas a more
quate dietary Cr intake. However, this area remains contro- consistent clinical response was observed with daily supple-
versial because studies demonstrating an inadequate Cr intake mentation of Cr ⬎200 g/d for a duration of at least 2 mo. In
in subjects with diabetes are not available. addition, other forms of Cr, especially CrPic, appeared to be
The controversy surrounding Cr as an adjunctive treatment more bioavailable and clinically more effective than CrCl in
in diabetes stems in large part from conflicting data reported in both human and rat studies (43).
previous studies of subjects with impaired glucose tolerance or Anderson et al. (4) provided evidence for a dose effect of
diabetes in which the diets were supplemented with Cr in an CrPic in a study of Chinese type 2 diabetic subjects. Short- (2
TABLE 3
Effect of chromium supplementation on carbohydrate metabolism in humans1
Results
Study Study
Reference type duration Dose (g) Subjects n Technique assessed Glucose Insulin HbA1c IS
1 Abbreviations used: DB, double blind; DM, Diabetes Mellitus; FBG, fasting plasma glucose; HbA1c, hemoglobin A1c; HOMA, homeostasis model
assessment; IGT, impaired glucose tolerant; IS, insulin sensitivity; IVGTT, intravenous glucose tolerance test; NA, not assessed; OGTT, oral glucose
tolerance test; OL, open label; 2, decreased; 1, increased; —, no change.
2 -cell sensitivity to glucose.
3 In hyperinsulinemic patients only.
1112 CEFALU ET AL.
form (pidolate vs. picolinate), duration of diabetes and status human trials with supplemental Cr (4,51–53), i.e., in studies of
of subjects (44). type 2 diabetic subjects, Anderson et al. (4) noted a significant
Our data suggest a role for supplemental CrPic in obese, drop in cholesterol levels, and Lee et al. (51) observed a 17%
insulin-resistant states because CrPic improved glucose toler- drop in triglyceride levels. Other reported human studies
ance and insulin levels in obese rats but not in lean controls. showed no significant effects on lipids with Cr nicotinic acid
It is not known whether the obese JCR:LA-cp rat is Cr (200 g) or CrPic (1000 g) supplementation (53,54). Ben-
deficient; thus, a limitation of this study is the lack of blood or eficial effects on lipids were also demonstrated in rats given a
urine Cr levels in the rats. A very important question would be synthetic, functional biomimetic Cr compound parenterally
whether blood Cr status could have explained the differences (52). Whether the improvement in lipid levels is secondary to
in the responses to Cr supplementation between lean and improvements in insulin levels per se or a direct effect of Cr on
obese rats; if so, it would suggest an abnormality in the insulin lipid metabolism, is currently unknown. However, there re-
signaling cascade in obesity that appears to be overcome with main interspecies differences in response to dietary changes
Cr supplementation. Whether hyperinsulinism, insulin resis- between rodent and human studies, and this effect on lipids,
tance and/or obesity, therefore, play a role in Cr metabolism before being extrapolated to human studies, will have to be
and/or excretion is an interesting question suggested by the evaluated specifically in human trials.
present studies. Such an observation may also explain in part Despite recognition of a specific Cr-deficient state, Cr re-
the reported discrepancies in response to CrPic in humans (see mains the only essential transition metal whose mechanism of
T. A. (1999) Chromium homeostasis in patients with type II (NIDDM) diabetes. 35. Nixon, D. E., Moyer, T. P., Squillace, D. P. & McCarthy, J. T. (1989)
J. Trace Elem. Med. Biol 13: 57– 61. Determination of serum nickel by graphite furnace atomic absorption spectrom-
6. Anderson, R. A. (1998) Chromium, glucose intolerance and diabetes. etry with Zeeman-effect background correction: values in a normal population
J. Am. Coll. Nutr. 17: 548 –555. and a population undergoing dialysis. Analyst 114: 1671–1674.
7. Anderson, R. A. (1997) Nutritional factors influencing the glucose/ 36. Moody, J. R. & Lindstrom, R. M. (1977) Selection and cleaning of
insulin system: chromium. J. Am. Coll. Nutr. 16: 404 – 410. plastic containers for storage of trace element samples. Anal. Chem. 49: 2264 –
8. Davis, C. M. & Vincent, J. B. (1997) Chromium oligopeptide activates 2267.
insulin receptor tyrosine kinase activity. Biochemistry 36: 4382– 4385. 37. Davies, S., McLaren, H. J., Hunnisett, A. & Howard, M. (1997) Age-
9. Davis, C. M., Sumrall, K. H. & Vincent, J. B. (1996) A biologically active related decreases in chromium levels in 51,665 hair, sweat, and serum samples
form of chromium may activate a membrane phosphotyrosine phosphatase from 40,872 patients—implications for the prevention of cardiovascular disease
(PTP). Biochemistry 35: 12963–12969. and type II diabetes mellitus. Metabolism 46: 469 – 473.
10. Russell, J. C., Ahuja, S. K., Manickavel, V., Rajotte, R. V. & Amy, R. M. 38. Ding, W., Chai, Z., Duan, P., Feng, W. & Qian, Q. (1998) Serum and
(1987) Insulin resistance and impaired glucose tolerance in the atherosclerosis- urine chromium concentrations in elderly diabetics. Biol. Trace Elem. Res. 63:
prone LA/N corpulent rat. Arteriosclerosis 7: 620 – 626. 231–237.
11. Mathe, D. (1995) Dyslipidemia and diabetes: animal models. Diabete 39. Ekmekcioglu, C., Prohaska, C., Pomazal, K., Steffan, I., Schernthaner, G.
Metab. 21: 106 –111. & Marktl, W. (2001) Concentrations of seven trace elements in different he-
12. Wu-Peng, X. S., Chua, S. C., Jr., Okada, N., Liu, S. M., Nicolson, M. & matological matrices in patients with type 2 diabetes as compared to healthy
Leibel, R. L. (1997) Phenotype of the obese Koletsky (f) rat due to Tyr763Stop controls. Biol. Trace Elem. Res. 79: 205–219.
mutation in the extracellular domain of the leptin receptor (Lepr): evidence for 40. Zima, T., Mestek, O., Tesar, V., Tesarova, P., Nemecek, K., Zak, A. &
deficient plasma-to-CSF transport of leptin in both the Zucker and Koletsky Zeman, M. (1998) Chromium levels in patients with internal diseases. Bio-
obese rat. Diabetes 46: 513–518. chem. Mol. Biol. Int. 46: 365–374.
13. Richardson, M., Schmidt, A. M., Graham, S. E., Achen, B., DeReske, M. 41. Rabinowitz, M. D., Gonick, H. D., Levine, S. R. & Davidson, M. B. (1983)
60. Potter, J. F., Levin, P., Anderson, R. A., Freiberg, J. M., Andres, R. & Elahi, 64. Evans, G. W. (1989) The effect of chromium picolinate on insulin
D. (1985) Glucose metabolism in glucose-intolerant older people during chro- controlled parameters in humans. Int. J. Biosoc. Med. Res. 11: 163–180.
mium supplementation. Metabolism 34: 199 –204. 65. Ravina, A., Slezak, L., Rubal, A. & Mirsky, N. (1995) Clinical use of the
61. Mossop, R. T. (1983) Effects of chromium III on fasting blood trace element chromium (III) in the treatment of diabetes mellitus. J. Trace Elem.
glucose, cholesterol and cholesterol HDL levels in diabetics. Cent. Afr. J. Med. Exp. Med. 8: 183–190.
29: 80 – 82. 66. Jovanovic, L., Gutierrez, M. & Peterson, M. (1999) Chromium supple-
62. Nath, R., Minoelia, J., Lyall, V., Sander, S., Kumar, V., Kapoor, S. & Dhar, mentation for women with gestational diabetes mellitus. J. Trace Elem. Exp. Med.
K. L. (1979) Assessment of chromium metabolism in maturity onset and 12: 91–97.
juvenile diabetes using chromium-51 and therapeutic response of chromium 67. Morris, B. W., Kouta, S., Robinson, R., MacNeil, S. & Heller, S. (2000)
administration on plasma lipids, glucose tolerance and insulin levels. In: Chro- Chromium supplementation improves insulin resistance in patients with Type 2
mium in Nutrition and Metabolism (Shapcott, D. & Hubert, J., eds.), pp. 213–222. diabetes mellitus. Diabet. Med. 17: 684 – 685.
Elsevier/North Holland, Amsterdam, The Netherlands. 68. Cheng, N., Zhu, X., Shi, H., Wu, W., Chi, J., Cheng, J. & Anderson, R.
63. Glinsman, W. H. & Mertz, W. (1966) Effect of trivalent chromium on (1999) Follow-up survey of people in China with Type 2 diabetes mellitus
glucose tolerance. Metabolism 15: 510 –520. consuming supplemental chromium. J. Trace Elem. Exp. Med. 12: 55– 60.