点击显示 收起
1 From the US Department of Agriculture, Human Nutrition Research Center on Aging, Tufts University, Boston.
2 The contents of this article do not necessarily reflect the views or policies of the US Department of Agriculture. 3 Supported by the US Department of Agriculture, Agricultural Research Service (contract 53-3-06-5-10). 4 Address reprint requests to RM Russell, USDA Human Nutrition Research Center on Aging, Tufts University, 711 Washington Street, Boston, MA 02111. E-mail: russell{at}hnrc.tufts.edu.
ABSTRACT
Dark adaptation has been used as a tool for identifying patients with subclinical vitamin A deficiency. With this functional test it was shown that tissue vitamin A deficiency occurs over a wide range of serum vitamin A concentrations. However, serum vitamin A concentrations >1.4 µmol/L predict normal dark adaptation 95% of the time. Other causes of abnormal dark adaptation include zinc and protein deficiencies. Stable isotopes of vitamin A and isotope-dilution techniques were used recently to evaluate body stores of vitamin A and the efficacy of vitamin A intervention programs in field settings and are being used to determine the vitamin A equivalences of dietary carotenoids. Vitamin A toxicity was described in patients taking large doses of vitamin A and in patients with type I hyperlipidemias and alcoholic liver disease. Conversely, tissue retinoic acid deficiency was described in alcoholic rats as a result of hepatic vitamin A mobilization, impaired oxidation of retinaldehyde, and increased destruction of retinoic acid by P450 enzymes. Abnormal oxidation products of carotenoids can cause toxicity in animal models and may have caused the increased incidence of lung cancer seen in 2 epidemiologic studies of the effects of high-dose ß-carotene supplementation. Major issues that remain to be studied include the efficiency of conversion of carotenoids in whole foods to vitamin A by using a variety of foods in various field settings and whether intraluminal factors (eg, parasitism) and vitamin A status affect this conversion. In addition, the biological activity of carotenoid metabolites should be better understood, particularly their effects on retinoid signaling.
Key Words: Vitamin A • retinoids • vitamin A deficiency • vitamin A toxicity • vitamin metabolism • stable isotopes • Robert H Herman Memorial Award in Clinical Nutrition
INTRODUCTION
This article covers some of the recent advances in the field of vitamin A deficiency and toxicity.
FUNCTIONAL TESTING
My involvement in the field of vitamin A began when Alex Krill, an ophthalmologist at the University of Chicago, approached me about his interest in developing a new visual function test for studying vitamin A deficiency. He wanted to study some patients with inflammatory bowel disease or celiac sprue at the gastroenterology clinics at the University of Chicago hospitals. I was eager to participate because I was interested in the prevalence of micronutrient deficiencies in these disease states. The first group of patients that we studied had relatively mild malabsorption due to chronic small-intestinal disease (1). In these patients, although fat malabsorption was relatively mild (mean fecal fat: 10 g/d with a 100-g-fat diet), the prevalence of reversible dark-adaptation abnormalities via vitamin A supplementation was in the range of 60%. Dark-adaptation curves for one of these patients, whose final threshold was grossly elevated at the beginning of study when his serum vitamin A concentration was 1.1 µmol/L, are shown in Figure 1. This patient's final dark-adapted threshold became normal after 30 d of treatment with 50000 IU (15 mg) vitamin A/d orally. The study pointed out a high frequency of vitamin A deficiency in patients with small-intestinal disease. The fact that many of these patients were taking routine vitamin supplements and that none of them was complaining of any kind of subjective symptom (eg, night blindness) suggested that this type of subclinical micronutrient deficiency was quite common in patients with Crohn disease and other chronic gastrointestinal diseases. Subsequent work showed that high prevalences of subclinical vitamin A deficiency also occur in clinic populations with alcoholic cirrhosis, primary biliary cirrhosis, a history of small-intestinal bypass surgery (which was once commonly performed for obesity), and pancreatic insufficiency (2–6).
FIGURE 1. . Dark-adaptation curves obtained from a vitamin A–deficient patient before (•) and after () treatment with 50000 IU (15 mg) vitamin A for 30 d. As time in the dark increased, luminance of smaller intensity was perceived. On treatment, the normal dark-adapted threshold decreased to the normal range, as indicated by the vertical bar, within 35–45 min. Adapted from reference 1.
In the meantime, Loerch et al (7) developed a new and unique way to diagnose vitamin A deficiency—the relative-dose-response test. This test requires 2 blood tests to be conducted 0 and 5 h after a physiologic dose of vitamin A. The test is based on the fact that in vitamin A–deprived states, resulting from an acute or chronic dietary deficiency, the plasma transport protein for vitamin A, retinol binding protein (RBP), accumulates in the liver (8). However, when vitamin A is made available from a dietary source, it becomes bound to the accumulated RBP and is promptly released into blood. Thus, in vitamin A–depleted organisms, the rise in serum retinol after a small dose of vitamin A is rapid, large, and sustained over a 5-h period. In contrast, in the vitamin A–sufficient state the rise in serum retinol reaches a lower and earlier apex, presumably because of a lower amount of accumulated apo-RBP and because the newly ingested dose goes into body storage pools rather than into the circulation. Thus, individuals with high relative dose responses are in a vitamin A–deficient state. The formula to calculate the relative dose response is as follows:
RETINOIC ACID DEFICIENCY
There has been recent interest in the effects of local tissue retinoic acid deficiency; thus, an alcoholic animal model was used to study this. Ethanol can compete with retinol for alcohol dehydrogenase, which catalyzes retinol oxidation to retinaldehyde, which then can be further oxidized to retinoic acid. It is possible that local tissue retinoic acid deficiency could be a molecular mechanism contributing to alcohol-induced liver injury (eg, proliferative activation of hepatocytes or hepatic fibrosis).
In work conducted by Wang et al (19), rats were fed a diet containing 36% of energy as alcohol or an isoenergetic diet containing maltose dextrin and no alcohol. Animals were pair fed for 4 wk; they were then killed and their tissues removed for analysis. Retinoic acid concentrations in rat liver and plasma after treatment with or without alcohol for 1 mo are shown in Table 2. It was found that treatment with a high dose of alcohol led to significant reductions in retinoic acid concentrations in both the liver and plasma. Because retinoic acid concentrations were significantly lower in the alcohol-fed animals, the authors hypothesized that alcohol ingestion can result in abnormal gene expression (19). Work from many groups showed that retinoic acid exerts profound effects on cellular growth and differentiation. Moreover, 2 families of nuclear retinoic acid receptors had been cloned (RAR and RXR) and were shown to be active in the receptor-mediated control of gene transcription. One of the proposed mechanisms for the antiproliferative effect of these retinoic acid receptors is through an interaction with the activator protein 1 (AP-1) complex made up of c-Fos and c-Jun. AP-1 mediates signals from several growth factors, inflammatory peptides, oncogenes, and tumor promoters, usually resulting in cell proliferation. AP-1–induced gene transcription can be inhibited by RAR and RXR when bound to retinoic acid.
View this table:
TABLE 2.. Concentration of retinoic acid in rat liver and plasma after treatment with or without (control) alcohol for 1 mo1
Thus, from this study, an intriguing hypothesis arises of how ethanol could result in proliferative activation of hepatic cells (Figure 3). Ethanol could inhibit retinoic acid synthesis by providing competition for alcohol dehydrogenase–catalyzed retinol oxidation. In addition, there might be increased destruction of the retinoic acid that is formed by increased P450 enzyme activity. Finally, ethanol might have an indirect effect in reducing retinoic acid concentrations in liver cells by increasing the mobilization of vitamin A to peripheral tissues (20). The decrease in tissue all-trans-retinoic acid concentrations could interfere with normal retinoid signal transduction by causing a functional down-regulation of RAR activity. As mentioned previously, RARs act as regulators of AP-1–responsive genes because RARs bound to retinoic acid can combine with the c-Fos–c-Jun complex and sequester it, thereby preventing it from binding to the AP-1 binding site. In the absence of retinoic acid, RARs can no longer bind to the c-Fos–c-Jun complex; thus, AP-1 can bind to DNA sequence motifs, resulting in the transactivation of target genes and cell proliferation. Such a mechanism could, in part, be responsible for alcohol-induced cell injury as well as malignant transformation. The tools of molecular biology are opening up new approaches for understanding the mechanisms whereby vitamin A deficiency and retinoic acid deficiency wreak havoc with tissues.
FIGURE 3. . Ethanol results in decreased liver retinoic acid concentrations by 1) increasing the mobilization of vitamin A from hepatic stores, 2) blocking the oxidation of retinol by inhibiting alcohol dehydrogenase (ADH), and 3) stimulating the hydrolysis of retinoic acid by cytochrome P450.
STABLE ISOTOPES
Relatively new technologies are being used to determine how best to combat vitamin A deficiency in a field setting: the use of stable isotopes and the analysis of samples with gas chromatography–mass spectrometry. De Pee et al (21) and Bulux et al (22) questioned the effectiveness of plant carotenoids in combating vitamin A deficiency. Each of these groups found no evidence of benefit on vitamin A nutritional status from the increased consumption of dark-green or yellow vegetables. Tang et al (23) reported the use of tetra- and octadeuterated retinyl acetate at different time points to assess changes in vitamin A status in Chinese children fed high-vegetable diets for 10 wk. To assess baseline vitamin A status, each child was fed octadeuterated retinyl acetate in corn oil on day 0; blood samples obtained on day 21 were used as the equilibration point. From days 22 to 92, carotenoid-rich, dark-green and yellow vegetables were fed to one kindergarten class, whereas light-colored vegetables were fed to children in a second kindergarten class. During this intervention period, conducted in the autumn, the 2 groups of children consumed 3 meals at school daily, 5 d/wk for 10 wk. After the intervention was complete on day 95, tetradeuterated retinyl acetate was administered to the children in each group to measure changes in body stores of vitamin A. The nutrient contents of both diets were equivalent except for the carotenoid content: the calculated retinol equivalents were 4 times higher in the dark-green and yellow-vegetable diet than in the light-colored-vegetable diet. After 95 d, serum retinol concentrations were not significantly different from baseline in the dark-green and yellow-vegetable group, but fell significantly (by 20%) in the light-colored-vegetable group. Mean serum vitamin A concentrations in these 2 groups of children were low, 1.0 µmol/L. Likewise, total liver reserves before and after the intervention did not change significantly in the dark-green and yellow-vegetable group, whereas they fell significantly (by 27 µmol) in the light-colored-vegetable group. Because the dark-green and yellow-vegetable group consumed 4.7 mg provitamin A carotenoids in their daily diet and the light-colored-vegetable group consumed only 0.7 mg provitamin A carotenoids in their daily diet, it was calculated that the provitamin A carotenoid intake by the dark-green and yellow-vegetable group prevented a loss of 7.4 mg retinol from the liver. With use of this estimate, it was calculated that ß-carotene from vegetable origin (under these study conditions) provided an estimated vitamin A equivalence of 25 to 1 by weight or a molar ratio of 13 to 1. Thus, in that study, vitamin A nutrition was sustained in Chinese kindergarten children who consumed dark-green and yellow vegetables with their meals. However, the vitamin A equivalence (by wt) of dietary carotenoids was less than the presently assumed ratios of 1 to 6 for ß-carotene and of 1 to 12 for other provitamin A carotenes. This issue is in need of far greater study.
VITAMIN A TOXICITY
Now I will turn to the opposite end of the vitamin A spectrum—vitamin A toxicity. My interest in vitamin A toxicity was sparked after treating a patient when I was a gastroenterology fellow at the University of Chicago. This patient had liver disease of obscure origin, but also had some odd symptoms: thinning of eyebrows; sparse, coarse hair; cheilosis; and bulging eyes (Figure 4). On careful questioning, she admitted to taking large doses of vitamin A, an average of 400000 IU (120 mg/d) for 8 y. Her liver biopsy showed hepatic congestion and fibrosis, particularly around the central vein. A second similar patient was identified within 2 mo. These patients were unique in that they showed a distinctive pattern of fibrosis and lipid disposition in their biopsy specimens. The report of these patients alerted the medical community that vitamin A toxicity may be more prevalent in clinic populations than recognized previously (24). Ellis et al (25) subsequently described other patients with vitamin A intoxication due to abnormal metabolism of vitamin A, specifically in patients with type I hyperlipidemia. Carpenter et al (26) described a familial clustering of vitamin A–intoxicated patients despite histories of only modest ingestion of the vitamin, thus implicating a possible genetic predisposition.
FIGURE 4. . A 58-y-old woman with vitamin A intoxication after taking on average 400000 IU (120 mg) vitamin A/d for 8 y. Note the dry skin, brittle hair, and cheilosis. This patient also had liver fibrosis.
Krasinski et al (27) were the first to point out the possible relation of age alone to a predisposition to vitamin A intoxication. These investigators showed that serum concentrations of vitamin A after a physiologic dose of vitamin A reached higher peaks in old people than in young people. As for the reason, it had been reported by Hollander and Morgan (28) that vitamin A was absorbed more readily in old rats than in young rats. However, Krasinski et al approached the problem somewhat differently. They fed humans high-fat, high–vitamin A meals and then conducted plasmapheresis several hours later. Chylomicrons and chylomicron remnants were laden with vitamin A esters and 24 h later the plasma was reinfused and the fall-off in serum retinyl esters was tracked in both old and young individuals. The fall-off in blood retinyl esters was significantly delayed 2-fold in older individuals than in younger individuals, which allowed for a transfer of vitamin A esters from chylomicrons into other lipoprotein particles such as LDLs. Once in LDL, potentially toxic retinyl esters are able to exist for 1 wk in the circulation as opposed to hours. In another study, Krasinski et al (29) showed that elderly people taking vitamin A supplements in amounts greater than the recommended dietary allowance tended to accumulate more retinyl esters in their fasting serum as the dose of vitamin A increased. Furthermore, they found that the longer the individuals took the vitamins containing vitamin A (ie, 5 y compared with <5 y), the greater the tendency for concentrations of potentially toxic retinyl esters to be high.
ß-CAROTENE TOXICITY
About the same time that these studies were published, ß-carotene toxicity was described by Leo et al (30) in the livers of alcohol-fed animals, which showed swollen mitochondria after ß-carotene feeding. Of interest is the possibility that retinoid metabolites of ß-carotene could also have biological and possibly toxic potential. Wang et al (31) showed that ß-carotene molecules in an in vitro system, in addition to splitting into retinal, could also be split at several double bonds, yielding apo carotenals and apo carotenoic acids. They showed that at low doses, these carotenoic acids could be converted directly to retinoic acid (32–34). That is, for retinoic acid to be formed, ß-carotene need not be converted to retinal first because in the presence of citral, which blocks the oxidation of retinal to retinoic acid, retinoic acid was still detected (35). Yeum et al (36) showed that this eccentric cleavage of ß-carotene could occur by a cooxidation mechanism in the cytosol. These investigations showed that when lipoxygenase was incubated with ß-carotene alone, very small amounts of eccentric cleavage products of ß-carotene appeared; however, when the substrate linoleic acid was added to the system, the cleavage metabolites of ß-carotene increased dramatically. Thus, it appears that eccentric cleavage can be initiated in tissues by a cooxidation mechanism and then possibly completed by either conversion to retinaldehyde to form retinoic acid or by a mitochondrial mechanism, as Wang et al (37) described, to form retinoic acid. However, the question arises as to what happens when these eccentric cleavage products accumulate in large amounts? Do they have biological activity of their own? Could these metabolites interfere with the action of retinoic acid? This may, in fact, partially explain the results from 2 carotene intervention trials in which the effects of high doses of ß-carotene supplements were studied in smokers and in asbestos-exposed workers (38, 39). These studies showed a higher incidence of lung cancer in smokers who consumed high doses of ß-carotene than in smokers who did not take ß-carotene supplements.
An animal model was used to try to mimic the results of these studies in humans (40). Ferrets were divided into 2 groups: ß-carotene supplemented and non-ß-carotene supplemented (control group). The dose of ß-carotene used was equivalent to 30 mg/d in the human intervention trials. The ß-carotene–supplemented and non-ß-carotene–supplemented groups were further divided into smoke-exposed and non-smoke-exposed groups. The smoke-exposed group was exposed to cigarette smoke within a chamber twice in the morning and twice in the afternoon for 30 min each time, providing an exposure equivalent to that from 1.5 packs of cigarettes/d in humans. The animals tolerated this exposure well; they experienced no decrease in appetite or weight and behaved no differently from non-smoke-exposed animals. Animals were treated for 6 mo and then killed. ß-Carotene concentrations in the plasma and lungs were greater in the ß-carotene–supplemented ferrets than in the nonsupplemented ferrets; however, ß-carotene concentrations in the lungs were significantly lower in the smoke-exposed ferrets than in the non-smoke-exposed ferrets in both the ß-carotene–supplemented and nonsupplemented control animals. Retinoic acid concentrations in the lung tissue were also significantly lower in all 3 treatment groups than in the control group (Table 3). The dramatic decreases in lung and blood ß-carotene concentrations as a result of smoke exposure correlated with the enhanced breakdown of ß-carotene into eccentric cleavage oxidation products.
View this table:
TABLE 3.. Concentrations of ß-carotene and retinoic acid in lung tissue of ferrets after 6 mo of treatment1
When the lung sections of the 4 groups of ferrets were examined, it was found that smoke exposure alone caused mild aggregation and proliferation of macrophages. However, localized proliferation of alveolar cells and alveolar macrophages and keratinized squamous epithelial cells were observed in the ferrets in the 2 ß-carotene–supplemented groups. The most severe proliferation of alveolar cells and squamous metaplasia was observed in the ß-carotene–supplemented, smoke-exposed ferrets. Keratinized squamous metaplasia was confirmed by immunohistochemical staining with anti-keratin antibody in the lung sections of all ferrets in the ß-carotene–supplemented, smoke-exposed and non-smoke-exposed groups. Retinoic acid concentrations were lower in the smoke-exposed ferrets than in the non-smoke-exposed ferrets, presumably because of increased oxidative breakdown. In turn, the expression of RAR ß (a subtype of RAR) activity was down-regulated in the lungs of the 3 treatment groups compared with that in the control group. RAR ß is known to play an important role in normal lung development, and primary lung tumors and lung cancer cell lines lack RAR ß expression (41–46). Thus, a role for RAR ß as a tumor suppressor gene in the lung has been proposed (47). Because lung carcinogenesis is also associated with an alteration in retinoid signaling involving the AP-1 complex, AP-1 transcriptional activity was studied in these ferrets (48). c-Fos and c-Jun expression were up-regulated in the ß-carotene–supplemented, smoke-exposed group. Additionally, AP-1 expression in this study was positively correlated with squamous metaplasia and inversely with RAR ß expression in these animals.
Thus, it appears that high doses of ß-carotene under highly oxidative conditions result in many eccentric cleavage oxidative breakdown products, which could have biological activity of their own. One possibility is that these products interfere with retinoic acid binding to retinoid receptors, but another likely possibility is that these metabolites induce local enzymes in the lung, such as P450 enzymes, which increase the catabolism of retinoic acid and thus diminish retinoic acid signaling. A local deficiency of retinoic acid can then result in squamous metaplasia. Salgo et al (49) reported that ß-carotene oxidation products promote the binding of benzo[a]pyrene (a smoke-borne carcinogen) to calf thymus DNA. Incubation of DNA with intact ß-carotene decreased such binding, whereas incubation with ß-carotene oxidation products (eg, 5,6-epioxide) for 1, 2, 3, and 4 h significantly increased the binding. These are all possible explanations for why toxicity occurs after high doses of ß-carotene and may explain the increased incidence of lung cancers observed in the 2 large intervention trials mentioned previously (38, 39).
What then are some major issues remaining concerning vitamin A deficiency and toxicity? With regard to deficiency, the efficiency of conversion of carotenoids in whole foods to vitamin A by using a variety of foods in various field settings and whether intraluminal factors (eg, parasitism) and vitamin A status affect this conversion should be studied. These issues are of utmost importance in developing countries. With regard to the toxicity of retinoids, the biological activity of carotenoid metabolites must be better understood in terms of their possible beneficial as well as harmful effects. Are these metabolites able to induce local tissue deficiencies of retinoic acid and thus diminish retinoid signaling? Do these metabolites have gene transcription activity of their own? These are both pertinent and important questions as we move into the 21st century and questions that I believe will be answered with use of available new technologies.
REFERENCES