2.2.1. When a Vitamin in not a Vitamin

Most animal species express a functional copy of L-gulonolactone oxidase, an enzyme with the synthesis of L-ascorbic acid as its only known function. In these animals, the regulation of GULO activity appears to depend primarily on substrate availability, namely the production of gulonolactone as a branch product from glucuronate synthesis [26] derived from UDP-glucose, a product of glycogen breakdown [52,53]. Thus, agents that stimulate glycogenolysis also stimulate ascorbate synthesis when an animal is in a fed state (*i.e.*, when glycogen is present); in contrast, prolonged fasting causes ascorbate synthesis to decline [54]. UDP-glucuronate is also needed for glucuronidation of xenobiotics, and there is a correlation between ascorbate synthesis activity and xenobiotic metabolism [26].

In each of the few mammalian species that do not synthesize ascorbic acid, such as guinea pigs, fruit bats, and primates, the loss of GULO has occurred at a genetic level. Although the mutations accumulated in this gene differ [55], the result is essentially the same: a loss of GULO activity. These species have adapted to the loss of *de novo* synthesis by consuming primarily plant sources of ascorbate. While all non-synthesizing animal species are at risk of developing scurvy and may die when vitamin C intake ceases for an extended period of time, this is a condition that does not occur normally in other animals, including most commonly used experimental animal models, such as mice and rats. Thus, in contrast to ascorbic acid-synthesizing species, in non-synthesizing species, including humans, vitamin C absorption is required to prevent deficiency and maintain health. This represents a fundamental shift from a perspective of diet-health interactions, as the absorption-derived *versus* glycogen-derived source of ascorbic acid may represent differences in ascorbate transport and carbohydrate metabolism that exist in humans and synthesizing animals.

Although evidence in the literature is limited, both rats and mice appear to poorly absorb ascorbate from the diet. Early studies on the small intestine in rats revealed that ascorbate uptake is a passive process, resulting mainly in intestinal mucosa accumulation but not transport to the circulation [56]. Several studies have demonstrated a profound difference in absorption between the rat and guinea pig small intestine [57–59], the latter displaying a robust transepithelial transport system that is sodium-dependent [59–61], similar to the human ileum [61,62]. A recent study in rats monitored the absorption of a single oral dose of ascorbic acid or dehydroascorbic acid given by gavage. The administration of 12 mg of dehydroascorbic acid led to a significant increase in plasma ascorbic acid concentration, but administration of 12 mg of ascorbic acid did not [63]. An inability to efficiently transport vitamin C is also seen in feeding studies, where mice fully capable of synthesizing ascorbate required at least 45 mg of ascorbate per day in their diet to show any significant increases in plasma ascorbate concentration [64]. Although it is difficult to extrapolate these doses to humans, allometric scaling based on calorie consumption suggests that a dose of 45 mg in a 20-g mouse is equivalent to about 3 g in a 70-kg person [65].

Studies on the bioavailability of different forms of ascorbic acid supplements also support the notion that ascorbate is poorly absorbed by rodent models. For example, Ester-C, a calcium ascorbate-threonate mixture, is reported to be more bioavailable in ODS rats than an equivalent dose of ascorbic acid [66]. However, the same comparison of supplements shows that Ester-C has a lower rate of absorption—and certainly no enhanced bioavailability—in human volunteers [67]. Furthermore, work in GULO knockout mice suggested an enhanced bioavailability of vitamin C contained in a kiwifruit puree compared to a pure ascorbic acid supplement gel [68]. However, more recent human data suggest that vitamin C from kiwifruit or a supplement is equally bioavailable [69].

Overall, these data support a species-specific route of ascorbate absorption. While indirect absorption of ascorbate may occur in rats and mice—likely mediated by ascorbic acid oxidation and transient formation and transport of dehydroascorbic acid—there is clear evidence of an active, sodium-dependent transport of vitamin C in guinea pigs and humans. Therefore, the use of rats or mice as a model of human vitamin C absorption and metabolism is ill advised. As indicated above, such studies with experimental animals should be avoided if they can be performed in humans, unless the purpose of the study is to better understand ascorbic acid absorption and metabolism in rodents. More importantly, no conclusions should and can be drawn from such studies for human vitamin C transport or metabolism.

Poor uptake of dietary ascorbate, or complete lack of it, is expected in animals that synthesize ascorbate, as they do not have a need for dietary ascorbate. In addition, high levels of ascorbate in the intestine would likely cause down-regulation of tissue ascorbate synthesis. However, the relationship between absorption and synthesis does not appear to be this simple. Genetically altered rat and mouse models lacking ascorbate synthesis have low tissue levels of ascorbate without supplementation [64,68,70,71]. Although these animals absorb dietary ascorbate, the levels needed to prevent scurvy or saturate tissues are relatively high compared to guinea pigs and humans on a body-weight basis (Table 1). These levels are especially high when contrasted to food sources of ascorbate. As an example, to saturate all tissues, GULO knockout mice need to be supplemented with 3.3 g/L of ascorbate in the drinking water, resulting in an intake of approximately 16.5 mg per day [68,72]. Based on allometric scaling, this dose in mice corresponds to about 1 g per day in a 70-kg person [65]. Although the metabolic rate of these animals likely contributes to high ascorbate requirements, this does not sufficiently explain data supporting an excessive inefficiency of intestinal absorption.
