Lamson DS, Plaza SM
Kenmore, WA, USA.
The data on the standards for chromium requirements and the safety of various chromium compounds and doses are reviewed. The 350-fold difference between the acceptable daily intake and the calculated reference dose for humans of 70 mg per day seems without precedent with respect to other nutritional minerals. Previous claims of mutagenic effects of chromium are of questionable relevance. While studies have found DNA fragmentation (clastogenic effects) by chromium picolinate, anecdotal reports of high-dose chromium picolinate toxicity are few and ambiguous. The beneficial effects of chromium on serum glucose and lipids and insulin resistance occur even in the healthy. Serum glucose can be improved by chromium supplementation in both types 1 and 2 diabetes, and the effect appears dose dependent. Relative absorption of various chromium compounds is summarized and the mechanism of low molecular weight chromium binding substance (LMWCr) in up-regulating the insulin effect eight-fold is discussed. There is evidence of hormonal effects of supplemental chromium besides the effect on insulin. Chromium supplementation does result in tissue retention, especially in the kidney, although no pathogenic effect has been demonstrated despite considerable study.
Glucose Tolerance Factor and Low Molecular Weight Chromium Binding Substance
Although glucose tolerance factor was recognized almost 50 years ago, attempts to isolate the specific structure have eluded scientists to this day. In 1959, Mertz identified trivalent chromium from yeast as the active constituent of GTF, which, when given to brewer’s yeast-fed, chromium-deficient rats, corrected imbalances in carbohydrate metabolism.  Determination of the biologically active form of chromium focused on the isolation of GTF from brewer’s yeast. Acid hydrolysis with 5N HCL for a period of 18 hours was part of the isolation protocol,68 which would have destroyed most protein structures associated with the bioactive molecule. Yet this protocol led to the discovery of a low molecular weight molecule that was determined to consist of nicotinic acid, glycine, glutamate, and a sulfurcontaining amino acid.  Although some early studies used porcine kidney as a raw material in the acid hydrolysis isolation procedure, most used readily available yeast as the raw material for GTF studies.  Today, the term GTF is reserved for the organic chromium degradation product from yeast.  The question as to whether GTF is a biologically active substance or artifact centers on the ability to stimulate production of CO2 from glucose in rat adipocytes as a function of insulin concentration. GTF appears to function as a carrier of chromium to the chromium-deficient proteins in the cell. 
Analysis of mammalian tissue has resulted in the isolation of a lowmolecular-eight chromium binding substance (LMWCr) that in many ways is similar to yeast GTF.
Yamamoto et al  isolated two chromium-binding substances: a low molecular weight substance and a high molecular weight substance. The highmolecular- weight chromium binding substance (HMWCr) was isolated from both rabbit liver and mouse organ homogenate and has a molecular weight of 2600 and an ultraviolet absorbance of 260 nm. LMWCr has also been isolated in a number of mammalian organs, including rat lung,  rabbit,  mouse and canine livers,  and from bovine colostrum.  LMWCr has a molecular mass of 1500 and has an ultraviolet absorption maximum at 260 nm, which corresponds to the absorption maximum of GTF, also at 260 nm. [69,73] The LMWCr oligopeptide is composed of cysteine, glutamate, aspartate, and glycine. LMWCr differs from GTF in the combination of amino acids and does not contain nicotinic acid.  Yet acid hydrolysis of porcine LMWCr, similar to early protocols, yields GTFlike isolates.  LMWCr has been found to bind four chromium III ions in a multinuclear assembly much like that of calmodulin. Studies of Vincet et al57 have discovered that LMWCr is stored in the cytosol of insulin-sensitive cells in an apo (unbound form) that is activated by binding four chromium ions. This activation is the result of a series of steps stimulated by insulin signaling. LMWCr potentiates the action of insulin once insulin has bound to its receptor.  (Figure 3)
This insulin potentiating or autoamplification action stems from the ability of LMWCr to maintain stimulation of tyrosine kinase activity. [25,57] Once insulin is bound to its receptor, LMWCr binds to the activated receptor on the inner side of the cell membrane and increases the insulin-activated protein kinase activity by eightfold.  There is also evidence the autoamplification effect of LMWCr may be enhanced by the inhibition of phosphotyrosine phosphatase, which inactivates tyrosine kinase.  Further study is required to understand the inconsistent results of Davis et al  who found that LMWCr actually activates membrane-associated phosphotyrosine phosphatase in insulin sensitive cells. As insulin levels drop and receptor activity diminishes, LMWCr is transported from the cell to the blood and excreted in the urine.  When chromium is absorbed from the gut, it is carried by transferrin, which transfers chromium to the apo-LMWCr. Excess chromium is carried by albumin. It has been estimated that each millimeter of serum contains 2-3 mg of transferrin, of which only 30 percent is saturated with iron, leaving the remaining unsaturated sites able to bind trivalent chromium. 
In the last three decades evidence has been collected demonstrating that both exogenous and endogenous factors significantly alter absorption and ultimately bioavailability of chromium. A considerable variance with respect to absorption is reported in the literature. One study of men over age 60 found absorption of trivalent chromium from dietary consumption was approximately 1.8 percent.  Other sources cite absorption between 0.5 and 2.0 percent.  Variations in absorption of trivalent chromium can be traced to differences in the type of chromium ingested, competing minerals, and the effect of vitamins, proteins, drugs, and other nutritional factors used in combination (Table 1).
Dietary factors such as starch, ascorbic acid, minerals, oxalate, and amino acid intake can have a significant influence on chromium absorption. Carbohydrate intake has been shown to influence chromium urinary excretion and tissue concentration. Mice fed 51Cr-labeled chromium III chloride concomitantly with starch were found to have significantly higher concentrations of chromium in blood and tissue compared to those fed with chromium III chloride mixed with sucrose, fructose, or glucose.  Diets high in simple sugars have also been shown to increase urinary excretion of chromium by 10-300 percent, with no change in absorption rates.  Animals fed ascorbic acid with chromium supplementation demonstrated increased absorption.  A study of three women found that the ingestion of ascorbic acid (100 mg) in conjunction with chromium III chloride (1 mg) increased the absorption of chromium as measured in plasma levels. 
A number of minerals influence absorption. In rat studies, zinc supplementation reduced chromium absorption, while zinc deficiency had the opposite effect, elevating 51Cr levels.  A later study by Anderson et al  found no alteration in tissue levels of copper and zinc when mice were fed a diet with 5000 ng Cr III/g of feed. In in vitro rat studies, iron, manganese, and calcium have all been shown to depress intestinal transport of chromium at levels of only 100-fold that of chromium, while in the case of titanium, concentrations only 10 times that of chromium inhibited absorption.  In a study of rats fed 5000 ng/g of feed of a number of organic chromium compounds (chromium picolinate, nicotinate, acetate, glycinate, histidinate, or chloride), results showed that all compounds tested increased the iron content in the liver and spleen while decreasing iron levels in the heart. 
The interaction of iron and chromium is thought to be linked to the shared binding sites on transferrin. Sargent et al  first proposed the theory that increased iron stores due to hemochromatosis might result in the competitive inhibition of chromium binding, leading to diabetic symptoms. He found that patients with hemochromatosis did, in fact, have significantly less plasma chromium than iron-depleted patients. Chromium has been found to preferentially bind to the B site of transferrin. When saturation of transferrin with iron increases in hemochromatosis to over 50 percent, iron competes with chromium binding, affecting its transport.  This theory is further supported by studies of patients with hemochromatosis who were found to have significantly higher excretion of the unbound plasma chromium as well as a smaller blood pool of chromium due to the saturation of transferrin by iron. 
It has been found that substances forming chelates with chromium generally stimulate absorption and that EDTA (ethylenediaminetetracetic acid) or DL-penicillamine significantly increase absorption as measured by 51Cr levels.  However Chen et al94 found no significant difference in absorption when EDTA and 51Cr were administered to rats. Naturally occurring chelating agents, such as phytates and oxalates, have also been found to influence chromium absorption in both in vitro and in vivo rat studies. Rats fed chromium with oxalate were found to have higher 51Cr blood and tissue levels, while rats fed phytates with chromium had lower blood and tissue levels.
A number of amino acids have also been found to increase absorption of chromium from the intestine. It was found that a mixture of 20 amino acids nearly doubled the rate of absorption. Amino acids like histidine and glutamic acid that readily form complexes with chromium were also shown to increase absorption. 
Earlier studies found trivalent chromium had consistent absorption and excretion regardless of previous diet history (unlike the absorption of other elements).  In 1996 it was discovered that chromium analyses in biological samples prior to 1980 were inaccurate due to the state of early analytical instrumentation.  More recent, post-1980 studies, using more accurate instrumentation, now find dietary absorption to be inversely proportional to dietary chromium intake (as with other minerals). Humans consuming a self-selected diet with an intake of 10 mcg/day Cr III had an absorption of two percent, while an intake of 40 mcg/day provided absorption of only 0.5 percent. 
Different forms of trivalent chromium have distinct characteristics of absorption, with inorganic complexes of trace minerals known to have lower levels compared to organic complexes. Chromite ores, chromic oxide, and chromium III chloride have historically been shown to have the lowest levels of absorption. Ingestion of inorganic salts such as chromium III chloride have levels of absorption ranging between 0.4-1.3 percent, with a mean of 0.69 percent. [82,83,97]
Many authors cite absorption levels of 2- 3 percent of dietary chromium as organic complexes. [4,82] Chromium from brewer’s yeast was absorbed in the range of 5-10 percent,  although others were unable to duplicate these results.  Chromium picolinate was found to have absorption in humans estimated at 2.8 percent +/- 1.14 SD.  Studies on rats found that 3-8 times more chromium nicotinate was absorbed and retained than was chromium picolinate or chromium chloride. After 6-12 hours, tissues retained on the average 2-4 times more chromium nicotinate than chromium picolinate.  Similar results in rat studies using a number of different organic complexes of chromium found the relative absorption/retention as follows: Cr nicotinate > Cr picolinate > Cr chloride.  Concentrations of chromium picolinate in the liver and kidney were found to be 2-6 times higher than for chromium chloride- or chromium nicotinate-fed rats, with no detectable toxicity.