I. INTRODUCTION

Top Previous Next References

Calcium is essential for life. It contributes to the electrical potential of cell membranes and is involved in many fundamental processes, including DNA synthesis, enzyme activity, photo- and chemosensory transduction, neurotransmitter release, membrane permeability, and intercellular communication. In vertebrates, calcium is also the major component of bone. Given these and many other functions, it is not surprising that disturbances in calcium metabolism have been implicated in most of the major chronic diseases, including osteoporosis, kidney disease, obesity, heart disease, and hypertension (for reviews, see Refs. 8, 20, 115, 116, 186, 187, 195, 231, 325). This has led to concern among nutritionists that the majority of the United States population does not eat as much calcium as they need (28, 191, 193, 196, 210, 212a, 310; see Refs. 118, 311 for contrasting views).

The pervasive contribution of calcium to physiology, coupled with recent advances in the ability to measure intracellular calcium levels, makes it a key area of study. For the past several years, there have been >1,000 peer-reviewed papers published each month that include "calcium" as a keyword (60a). There are several fine reviews of the homeostatic controls underlying the regulation of intracellular calcium, blood calcium concentrations, and bone (e.g., Refs. 159, 209). However, none considers the possibility that body calcium is regulated behaviorally as well as physiologically.

The purpose of this review is to readdress this imbalance. In fact, there is a moderately large body of work on calcium appetite, but it is widely scattered and badly splintered. In the avian literature, calcium appetite is a well-accepted phenomenon, although there is little connection made between studies on wild and domesticated birds. There have also been very few attempts to examine the physiological basis of the appetite and no studies of calcium gustation in birds. The mechanisms of calcium gustation have been studied most extensively in amphibians, but there are no accompanying behavioral or physiological studies for this class. Work on calcium appetite in mammals had a promising start in the 1930s (76, 242, 243) but foundered in the 1960s until recently. The lack of interest was due to work showing that rats could use arbitrary flavors to recognize sources of calcium (248, 263). From this, it was inferred that calcium appetite is solely a learned behavior, and thus uninteresting, or at least simply one of many learned appetites. This contrasted markedly with the appetite for sodium, which was considered to have innate properties and a clear anatomical and physiological substrate (248; see Refs. 67, 262 for review). The emphasis on sodium appetite as a unique behavior has overshadowed calcium appetite, and much of the information gleaned about calcium taste and intake comes from studies directed primarily at understanding sodium taste and intake.

Calcium appetite is the motivation to seek out or choose calcium-containing items. This implies that animals are capable of detecting calcium or some marker for it. The evidence for this is mustered in section II, which also deals with the spontaneous or "need-free" ingestion of calcium. Section III covers primarily behavioral observations that form the backbone of the phenomenology of calcium appetite. A major issue involves how calcium appetite is manifested. Does an animal deprived of calcium seek calcium ions, minerals, saltiness, bitterness, or something else? Section IV addresses this question of the appetite's specificity, or lack of it. Section V musters the evidence for and against various potential physiological mechanisms underlying calcium appetite. The final two sections raise the issue of whether or not humans have a calcium appetite and provide a brief commentary on data needs and potential future directions of this field.

This review is limited to vertebrates. Although several invertebrates, including American cockroaches (95), aquatic beetles (123), and various moth caterpillars (94, 96) reject high concentrations of calcium, nothing is known about the mechanisms involved.

	II. CALCIUM TASTE AND CALCIUM INTAKE BY CALCIUM-REPLETE ANIMALS

Top Previous Next References

Sources of calcium could potentially be detected by sight, smell, or taste. Of these, taste appears to be the most important, at least for mammals. There is evidence that birds use visual cues to locate calcium (145; see sect. IVA). Moreover, chickens offered separate sources of CaCO₃ or SrCO₃ were unable to distinguish between them unless a red color cue was present (131). But visual cues are not required for calcium detection because chickens respond appropriately when given a choice between diet containing CaCO₃ and calcium-deficient diet with white flour added to obscure color cues (131).

Humans and other mammals can detect calcium salts by smell (78, 295). However, at least in humans, the thresholds for detection of calcium odor are many times higher than those for taste (295), and wearing a nose clip does not influence attribution ratings of CaCl₂ solutions (121). There is no evidence that calcium salts specifically activate trigeminal afferent fibers, which would imply thermal, textural, or irritant qualities for calcium. Thus the emphasis here is on calcium taste.

Amphibians can detect submillimolar concentrations of calcium, which is consistent with levels found in their habitat (141, 228). One comprehensive investigation found changes in activity of the glossopharyngeal nerve of the frog when as little as 10 µM CaSO₄ or CaCl₂ was applied to the tongue (150).¹ Other investigators report slightly higher threshold concentrations (4, 40, 141, 151, 324), and 1 mM CaCl₂ is routinely used as a taste stimulus. High concentrations of CaCl₂ (e.g., 250 mM) produce less response than do lower ones (e.g., Ref. 324).

In contrast to the work with amphibians, which was designed primarily to uncover mechanisms of taste transduction (see sect. IIB), most of the early work on calcium sensitivity in mammals formed part of broad surveys of gustatory sensitivity to many compounds, with the focus being on NaCl (e.g., Refs. 17, 19, 87). The minimum concentration of orally applied CaCl₂ to evoke a response in the chorda tympani nerve of the rat is ~30-100 µM (132, 133, 135, 204). This agrees reasonably well with rat behavioral studies, in which concentration-preference functions for calcium salts were determined. The "thresholds" at which taste solution intake was significantly greater than water intake were 10-400 µM calcium (135, 294). The wide range can probably be attributed to differences in the calcium salt tested and rats' age, diet, and previous experience with calcium solutions.

Human calcium detection thresholds are similar to or slightly higher than those of the rat. Early investigators report detection thresholds ranging between 2 and 30 mM (median ~10 mM; Refs. 83, 96, 128, 309; see also Ref. 217). However, a recent study focused specifically on calcium found the mean detection threshold of 13 women for CaCl₂ was 0.16 mM (range 0.01-1 mM, Ref. 295). Detection thresholds for four other calcium salts (lactate, gluconate, hydroxide, phosphate) were similar to those for CaCl₂. The much lower values found in this compared with earlier reports are probably due to the use of more sensitive methods, the purity of the water used, the subject's recent calcium consumption, and the subject's sex. The calcium content of the water supply where this study was conducted (Philadelphia) averages ~0.9 mM (range 0.35-1.40 mM) calcium, depending on the season and water's source (218). Thus calcium must contribute variably to the taste of tap water (see also Ref. 326).

Calcium detection thresholds are fairly close to salivary calcium concentrations. Depending on the test conditions, average total saliva calcium levels range between 0.9 and 2.1 mM and ionized calcium from 0.5 to 1.0 mM (see Ref. 247 for an authoritative review). It is possible that the lower limit on calcium detection is determined by the level of ionized calcium in saliva.

Amphibian electrophysiology supports the idea of a specific calcium taste mechanism. Gustatory afferent fibers that are sensitive to CaCl₂ in the frog respond to calcium and strontium ions (150) but not undissociated CaCl₂, low concentrations of chloride, or water (141). The response to calcium is inhibited by NaCl, KCl, magnesium salts, and high (0.5 mM) concentrations of chloride (40, 141, 150, 324). This inhibition, the finding that the frog is more sensitive to calcium than it is to sodium (4), and other evidence (see Ref. 4) indicates that the mechanism of calcium transduction differs from that of sodium and other monovalent ions (141, 160; see Ref. 255 for a review).

In the mudpuppy, calcium-dependent ion channels are found in both the apical (chemoreceptive) and basolateral (synaptic) membrane of taste cells, but not in surrounding nontaste epithelial cells (see Ref. 251). Calcium is involved in both the active and passive components of the taste receptor membrane potential (149). The receptor potentials evoked by calcium are not simply due to it entering the cell through calcium channels (see also Refs. 150, 151). The critical transduction process involves modulation of resting potassium conductance (24). Both this finding and work with the frog suggests that transduction of calcium salts involves adsorption of the ion onto the taste cell membrane (24, 144, 150, 151; reviewed in Ref. 256).

This notion is particularly intriguing given the recent discovery of extracellular calcium receptors in mammals, including humans (37, 38). Perhaps such receptors mediate the adsorption and transduction of calcium taste. We recently have found positive staining for antibody to the extracellular calcium receptor (CaR) in taste buds of rat and one of two human fetus tongues examined (N. E. Rawson and M. G. Tordoff, unpublished results). There is also evidence that metabotropic glutamate receptors are found on the tongue and may be responsible for umami taste (46). Metabotropic glutamate receptors in the central nervous system (CNS) are sensitive to extracellular calcium (156, 257), but it is unknown whether this is true of the lingual receptors. Another mechanism underlying calcium transduction may involve ecto-calcium-dependent ATPase, which is found in taste buds of amphibians and mammals (hamsters) and is related to oral calcium concentrations (9). A role for epithelial Na⁺/Ca²⁺ exchange seems unlikely because blockade with 5-(N-4-chlorobenzyl)-2',4'-dimethylbenzamil had no effect on human subjects' perception of CaCl₂ taste (259).

Studies of the electrophysiological response to oral calcium in mammals are rudimentary. Early work compared calcium and other minerals with sodium. The integrated response of the chorda tympani nerve to continuous stimulation of the tongue with 100 mM CaCl₂ declined more rapidly than it did during stimulation with 100 mM NaCl (17), calcium salts evoked a longer discharge after a water rinse than did sodium salts (17), and adaptation of the tongue to 100 mM CaCl₂ did not influence the response to 100 mM NaCl (18). Both whole nerve and single-unit recordings from the chorda tympani nerve of the rat showed the response to orally applied 100 mM CaCl₂ was ~55% of the magnitude of the response to 100 mM NaCl (17, 19, 87; see Ref. 19 for similar studies in hamster, guinea pig, cat, and dog). In contrast, a more recent study found that CaCl₂ applied to the rat tongue produced a whole nerve chorda tympani response that was larger than that to monovalent chlorides (204). This is consistent with work showing that changes in receptor potential produced by application of 100 mM CaCl₂ to the rat tongue were "sometimes" larger than those produced by NaCl (147). Such results argue that the perception of calcium and sodium involve different transduction mechanisms, but are otherwise uninformative.

Individual gustatory afferent nerve fibers can be classified by the taste solution that causes the most pronounced (or "best") increase in action potential frequency. Recordings from both the whole chorda tympani nerve (132) and individual fibers (133) indicated that HCl- and NaCl-best fibers are stimulated about equally by oral application of 3, 10, and 31 mM CaCl₂ solutions. However, there was considerable variability. All HCl-best fibers responded to CaCl₂, but only some NaCl-best fibers were sensitive to it. There was little or no response of sucrose-best fibers to CaCl₂. Recordings from the geniculate ganglion of the rat produced different results (31). Application of 50 mM CaCl₂ to the tongue caused approximately six times more spikes in acid- than NaCl-sensitive units. Most units in the petrosal ganglion showed only small responses to CaCl₂. These units were more sensitive to quinine or atropine, suggesting that CaCl₂ may activate a class of alkaloid-sensitive units (30).

Smith et al. (273) performed a comprehensive test of gustatory evoked electrophysiological activity in the nucleus tractus solitarius and parabrachial pons of the hamster. In the nucleus tractus solitarius, HCl-best and NaCl-best but not sucrose-best units responded to CaCl₂. In the parabrachial pons, HCl-best, NaCl-best, and many sucrose-best units responded to CaCl₂. This paper is noteworthy in that the authors attempted to compare the similarity of tastes based on electrophysiological and behavioral responses, the latter being determined by generalization of conditioned taste aversions (see sect. IIB2). Except for quinine, the correspondence between electrophysiology and behavior was very close.

An interesting study of the chorda tympani response of SIc:ICR mice found that, in contrast to most other species tested, integrated responses to 100 mM MgCl₂ and CaCl₂ were considerably greater than those to monovalent chlorides (207). Cluster analysis of the responses of individual chorda tympani fibers found that ~23% were narrowly tuned to calcium and/or magnesium. As far as I know, this is the only direct evidence for specific calcium-magnesium taste fibers in mammals.

Various forms of multidimensional scaling have been used to compare the response profiles of neurons in the rat nucleus tractus solitarius (81, 82, 264). Responses evoked by orally applied CaCl₂ clustered with those evoked by MgCl₂, KCl, and several bitter substances (e.g., strychnine, nicotine, quinine). There was also a close clustering of CaCl₂ with sour tastes (acids) in one analysis (based on activity profiles across neurons) but not in another (based on activity profiles across time, Ref. 264). One study compared the response of nucleus tractus solitarius neurons to a variety of taste solutions in sodium-deprived and control rats (136). The main finding was that response profiles elicited by sugars and by sodium salts were more similar in sodium-deprived than control rats. However, the most striking difference in response profile occurred with 300 mM CaCl₂. In the replete animals, the response to 300 mM CaCl₂ approximated those of HCl, KCl, and citric acid (replicating, Ref. 136), but in the sodium-deprived animals, the response to CaCl₂ was closest to quinine. Given that the response profiles relate well to taste qualities described by humans, this study suggests that the rat finds CaCl₂ to be predominantly sour when sodium replete but predominantly bitter when sodium deprived.

The method of cross-adaptation has been used to characterize the electrophysiological response to oral calcium (272). The principle of this depends on the observation that during continuous or repeated exposure to a taste solution, electrophysiological and behavioral responses decrease, or adapt. If the tongue is exposed to one taste solution and then another, the response to the second solution is reduced in relation to its similarity to the first. In a comprehensive study involving cross-adaptation of the rat chorda tympani nerve response to CaCl₂, CaBr₂, and 17 other compounds, Smith and Frank (272) found that the proportion of the response eliminated by adaptation (i.e., the similarity) was the following for each cation: Na⁺ = 36%, Li⁺ = 28%, K⁺ = 76%, Mg²⁺ = 82%, and NH = 59%. For bitter, sour, and sweet, the corresponding values were ~18, 46, and 13%, respectively. Cross-adaptation of CaCl₂ and CaBr₂ was 102%, that is, almost perfect.

In summary, electrophysiological evidence from amphibians supports the existence of a specific calcium-sensitive transduction mechanism, and there is histological evidence from mammals that is consistent with this. Electrophysiological studies of gustatory responses in mammals have produced a wide range of not always consistent results. Calcium and sodium produce different patterns of integrated chorda tympani activity, and calcium stimulates fibers or neural units that are tuned to HCl, NaCl, sucrose, alkaloids, and calcium-magnesium. Multidimensional scaling and cross-adaptation methods indicate neural gustatory activity evoked by calcium is moderately similar to that evoked by potassium, magnesium, and ammonium, somewhat different from that evoked by acids, and very different from that evoked by sodium and sweeteners. The many differences between the results of various studies are probably due to the species being tested, the level of the nervous system where recordings are made, the other compounds used in comparison with calcium, and the methods used to determine which compounds produce the "best" firing.

It is noteworthy given the lack of specificity of calcium appetite (see sect. IV) that oral calcium probably modulates the transduction of other taste compounds. For example, K⁺ and Cl channels in mudpuppy are dependent on extracellular calcium (182), so tastes modulated by these mechanisms may be influenced by calcium availability. Consistent with this, if the frog tongue was adapted to 1-100 mM CaCl₂ for 2 min, subsequent responses to bitter substances were prolonged. This effect was unaltered by calcium channel blockade and did not occur if the tongue was adapted to NaCl or MgCl₂ (146). Studies examining the chorda tympani response of mixtures of CaCl₂ with other salts and sweeteners in hamsters produced complex results, including several examples of calcium influencing the response to other taste solutions (132, 133).

There has been very little attempt to determine how nonhuman animals perceive calcium. In a very clever study, Morrison (198) trained rats to press the appropriate bar after drinking quinine, HCl, or sucrose. He then presented them with 16 "test" salts including CaCl₂. When given 300 mM CaCl₂, the rats pressed levers associated with quinine and HCl about equally but did not press a lever associated with sucrose. The response pattern produced by CaCl₂ was "distinctly different" from all the other salts tested, except perhaps lead acetate. These findings were recently extended in a study involving tests of three concentrations of nine different taste compounds (100). Rats were trained to press one bar associated with NaCl and another associated with quinine, HCl, or NH₄Cl. After receiving 300 mM CaCl₂ the rats bar pressed in the ratio of 90% HCl-10% NaCl, 70% quinine-30% NaCl, and 80% NH₄Cl-20% NaCl. Thus they found 300 mM CaCl₂ to be about equally sour, bitter, and ammonium-like and not very salty. The rats responded more to NaCl and less to HCl and quinine after receiving 100 or 1,000 mM CaCl₂ than after 300 mM CaCl₂. Similar results were obtained in a marmoset tested with 100 mM CaCl₂. The authors could not easily associate these results with their prior gustatory electrophysiological results (e.g., Ref. 265).

An alternative method to examine the similarity of tastes involves taste aversion learning. Animals poisoned after drinking a taste solution later avoid solutions with similar tastes. This method has been used with hamsters that were poisoned after drinking 50 mM CaCl₂ and tested with nine other compounds (or vice versa). The hamsters strongly avoided MgSO₄, NH₄Cl, HCl, and quinine to roughly the same extent and weakly avoided NaCl, NaNO₃, DL-alanine, sucrose, and tartaric acid (273).

For studies of taste quality, humans have the advantage of being able to give oral reports about what they taste. There have been several attempts to match the intensity or taste quality of CaCl₂ with a reference taste solution, usually NaCl. Thus it has been reported that CaCl₂ is 81% as salty as the same concentration of NaCl [(309) cited in (217)], that 26 mM CaCl₂ and 26 mM CaBr₂ are the same intensity as 100 mM NaCl (183), and that 25 or 50 mM CaCl₂ are the same intensity as 200 mM NaCl (258).

Although early investigators considered calcium salts to be salty and bitter (83, 95, 142, 194), humans are remarkably inconsistent in describing calcium taste. At least one investigator reported that "subjects did not agree at all on the names to apply to the tastes of such substances as calcium chloride" (96). More recently, Schiffman and Erickson (258) asked subjects to rate the similarity of CaCl₂ to 18 other taste solutions and to describe them using 33 semantic differential scales. Multidimensional scaling suggested that CaCl₂ was approximately equally similar to salty and bitter tastes and relatively dissimilar from sweet tastes (Fig. 1). The semantic differential scales suggested that CaCl₂ is "generally a simple, minerally taste which is moderately bitter, smooth, warming, and soft. Only one subject (of 4) considered it to have a salty component... Two subjects found it quite soapy, but not obnoxious. Two others found it obnoxious, but not soapy. For two subjects, CaCl₂ was rated slow developing, flat, and nauseous. For the other two, it was rated fast developing and neither flat nor nauseous. It was uniformly considered nonfoodlike, but was rated poisonous by only one subject. " (258). Either different subjects perceive CaCl₂ quite differently or they use different language to describe its taste.

View larger version (13K): [in this window] [in a new window]   Fig. 1. Relationship of CaCl2 to other taste solutions and descriptors. Results of multidimensional scaling are based on 33 semantic differential scales. Note that CaCl2 (slightly left and above center) is not well described by either dimension and does not group with salty, bitter, or sour solutions. [From Schiffman and Erickson (258), with permission from Elsevier Science.]

The contribution of each of the four classic taste modalities to the taste of CaCl₂ and calcium lactate (CaLa) has been assessed (295). A 1 mM CaCl₂ solution was rated as 35% bitter, 32% sour, 29% sweet, and 4% salty. At higher concentrations, the sweet component diminished and the salty component increased so that 100 mM CaCl₂ was rated as 44% bitter, 20% sour, 1% sweet, and 35% salty. CaLa solutions were considered to be less bitter, less salty, and more sour than equimolar CaCl₂ solutions. In another study (121), subjects provided descriptors from a list of nine (sweet, sour, salty, bitter, soapy, sulfurous, metallic, other, none) that best described the taste of CaCl₂. The predominant rating at all three concentrations tested (10, 30, and 100 mM) was bitter; saltiness was not evident at 10 or 30 mM concentrations. The 10 mM CaCl₂ was also rated sweet/sour, soapy/metallic, other, and none; 30 mM CaCl₂ was rated salty, sweet/sour, sulfurous, and other; and 100 mM CaCl₂ was rated salty, sweet/sour, and soapy/metallic.

The method of cross-adaptation has been used to investigate the similarity of the tastes for calcium, sodium, potassium, and ammonium (183). Subjects were adapted to water, 100 mM NaCl, 110 mM KCl, 51 mM NH₄Cl, or 26 mM CaCl₂ by running the solution over the tongue. They then rated the intensity of the four chlorides and the equivalent bromides. Cross-adaptation between calcium and the other cations was minimal, suggesting different perceptual mechanisms were involved (see also Ref. 184).

In summary, the taste quality of calcium is not easily described, but it differs from the taste of sodium and other minerals. Calcium taste varies with both the form and concentration of salt tested, but nearly always includes sour and bitter components. Only low concentrations are sweet, and only high concentrations of CaCl₂ are salty. Although there is some convergence of electrophysiological and behavioral results (273), there are also many differences (265). One difference is that behavioral studies generally find a much larger bitter component to calcium than do most electrophysiological studies, although this may well be due to the mammalian electrophysiologist's preference for recording from the chorda tympani nerve, which carries little bitter-related information (e.g., Ref. 278).

The previous sections have dealt with the detection and quality of calcium taste, but there is also a "hedonic" property of calcium that does not relate easily to its psychophysical qualities. It is clear that high concentrations of calcium are unpleasant to humans (295) and are avoided by animals (e.g., Refs. 135, 294). However, there is also evidence that low concentrations are preferred over water and that animals frequently ingest more calcium than they need. Terminology in the literature is imprecise, but strictly speaking, this is a form of calcium intake that is unrelated to calcium appetite because it does not satisfy calcium homeostasis. The assumption is that under some circumstances calcium is consumed for no other reason than it tastes good.

The most common method to determine taste acceptability in nonhuman mammals is the two-bottle choice test. One bottle always contains water and the other contains a taste solution. Sprague-Dawley rats have been tested in this manner with a wide range of calcium solutions (0.2-300 mM, CaCl₂, phosphate, hydroxide, gluconate, and lactate; Ref. 294). Although there were slight differences in intake of the various salts, the general pattern of results was the same. Rats preferred calcium to water at calcium concentrations between ~0. 4 and 5 mM, showed indifference between 5 and 12 mM, and avoided higher concentrations. Slightly lower values were found in a follow-up study (294; Fig. 2). The inverted U-shaped preference curve is reminiscent of the curve shown by rats given NaCl solutions to drink (239).

View larger version (21K): [in this window] [in a new window]   Fig. 2. Inverted U-shaped concentration intake function for CaCl2 seen in Sprague-Dawley rats fed a calcium-replete (AIN-76A) diet. Humans also prefer low concentrations of calcium to water (326). [Adapted from Tordoff (294).]

This "spontaneous" or "need-free" intake is not confined to rodents. In a recent study (220), marmoset monkeys fed a calcium-replete diet drank significantly more 65 or 130 mM CaLa than water in two 7-h tests. Humans also follow the inverted U-shaped calcium concentration-liking function. Subjects reported 2 mM CaSO₄ to be preferred over water and other concentrations of CaSO₄. Similarly, 1.25 mM Ca(HCO₃)₂ tasted significantly better than 0.5 mM or 10 mM Ca(HCO₃)₂. There was no peak preference for CaCl₂, but the lowest concentration used (0.5 mM), which tasted about as good as water, may not have been low enough to reveal a preference (326). In other work (295), concentrations of CaCl₂ and CaLa of 3.16 mM or greater were disliked. High concentrations of CaCl₂ (31.6 and 100 mM) were rated as more intense and more disliked than equimolar CaLa solutions.

Unlike mammals, birds do not appear to find calcium solutions palatable, even after a period of dietary calcium deprivation (131, 145, 322). Nevertheless, hens fed diets containing sufficient calcium to meet requirements consume supplementary calcium grit when this is available (254; cited in Ref. 285). Under conditions of free access, calcium intakes can be highly variable and depend on the source of calcium (114).

Animals sometimes show "nutritional wisdom"; that is, they can regulate their diet choices to satisfy their physiological needs. One method to examine this with respect to calcium is to present calcium-replete animals with a continuous choice between a food without calcium and a separate source of calcium. This can be a "pure" powder or grit, a solution, or a second food incorporating calcium. This method differs from "need-free" calcium consumption (see sect. IIC) because in the need-free approach, the animal has its physiological requirement for calcium met by its maintenance diet but in the "diet choice" approach, it must consume the calcium source to avoid calcium deficiency. In fact, animals can recognize the nutritive value of calcium even while calcium replete (50). They most likely make appropriate diet choices before obvious signs of deficiency occur so the acquisition of calcium ingestion under these circumstances is probably not a response to deficiency.

Some of the earliest evidence for the existence of nutritional wisdom with respect to calcium comes from studies in which rats received a cafeteria choice of as many as 10 separate sources of nutrients. The rats consumed enough of each nutrient, including calcium, to match or exceed rates of growth produced by standard composite diets (e.g., Refs. 76, 78, 240-244). With the use of a much simpler method, rats given a choice between calcium-deficient diet and the same diet containing a calcium salt (CaLa, citrate, CaCO₃, CaHPO₄, or CaSO₄) chose between 20 and 34% of their calories from the calcium source, depending on the form and concentration of the calcium (263, 315). All the rats grew normally. Calcium intakes were higher in the rats given the high-calcium diets than those given the low-calcium diets, even though the high-calcium diets were preferred less.

Similar studies have been conducted with poultry. For example, hens given a choice between diets of high- or low-calcium content ate sufficient calcium to maintain normal egg production, although some groups were mildly hypocalcemic and had less dense eggs relative to birds given calcium-replete diets (124). In another study, hens given a choice between a high-energy, low-calcium diet and a low-energy, high-calcium diet had lower energy intakes and higher calcium intakes than did controls fed a composite diet. In this case, they produced eggs of superior shell quality (165).

A notable failure to find evidence for regulation of calcium intake involved cows fed diets containing various levels of calcium, and which also had access to solid CaHPO₄ for 1 h daily (58). In one experiment, CaHPO₄ intake increased during the first 10 wk of maintenance on low-calcium diet, but ~35% of the animals never approached the calcium source, and of those that did, there was great variability in intakes (0-1 kg CaHPO₄/day). In another experiment, heifers were fed diets containing 350, 3,100, or 4,150 mmol/day calcium and given continuous access to a choice of CaHPO₄ or defluorinated phosphate. Preference for the calcium source when fed the low-, medium-, and high-calcium diets was 77, 80, and 90%, respectively, which is the opposite result of what would be expected if calcium intake was regulated (see also sect. IIID).

The "diet choice" method has proven uniformly poor to determine calcium requirements, but this is not necessarily because the animal fails to show nutritional wisdom. It must balance the adverse, progressively developing effects of calcium deficiency against consumption of an unpleasant tasting calcium source. Sources of calcium differ substantially in acceptability, so intakes can be quite varied despite similar requirements (see discussion in Ref. 292).

	III. DEMONSTRATIONS OF CALCIUM APPETITE

Top Previous Next References

Calcium concentration in the sea is ~10 mM, although primieval levels may have been lower. Fresh water contains between 0.02 and 1.75 mM calcium (228), and calcium on land can be scarce. Thus the move from a marine environment to fresh water and land required the development of physiological mechanisms to help conserve calcium. It seems reasonable to assume that behavioral mechanisms to help locate calcium developed at the same time.

In some terrains, calcium is readily available as sedimentary rock (e.g., chalk and limestone). However, in many other areas, the ground does not provide sufficient calcium to support vertebrate life. Plants are generally an insufficient source of calcium (see Refs. 102, 307), so herbivores have a particularly difficult task. Many animals seek out and ingest salt licks if these are available. Although the emphasis has been on the behavior of animals obtaining sodium from salt licks, in many cases calcium is consumed as well (review in Ref. 67). However, the use of saltiness as a cue for calcium can be misleading because calcium and sodium do not always co-occur in rock outcrops (see Ref. 293).

An important source of calcium for many animals is the bones of their prey. Old bones are also chewed by herbivores, although this is most frequently associated with phosphorus deficiency (e.g., Refs. 26, 27, 67, 105, 112, 284). Calcium also comes from teeth, birds' eggs, and anthropogenic sources (e.g., mortar, putty), and in mammals, mother's milk.

There are three long-lived but disconnected literatures concerning the effects of calcium deficiency on subsequent calcium intake, which are summarized separately below.

"March 26, 1981 was a sunny day with light frost in the morning; in some places, the snow had completely melted. At ~6:00 A.M. on the wall of the house where I had spent the night, there were 50 red crossbills and the whole time they were eating mortar. Small flocks would fly to nearby spruce trees and then return to the wall. During the following two days, the number of crossbills increased, so that on the third day there were between 100 and 150" (283).

The above extract is typical of many published reports of calcium intake by wild birds. Most birds eat insects and/or plants that are relatively low in calcium, so these cannot fulfill calcium requirements (111; review in Ref. 108). For some species, restricted calcium availability limits geographical distribution (61, 101, 185, 318 and citations therein). However, many others supplement their diet by eating calcium-rich material such as bones, owl pellets, mortar, grit, and the shells of snails, crabs, marine mollusks, and other birds (33, 106, 109, 110, 129, 139, 155, 176, 252). Boreal chickadees consume ash rich in calcium (86); sandpipers eat grit, lemming bones, and teeth (176); sandwich terns eat shell fragments; and crossbills in coniferous forests eat bones, putty, and cement (206, 214, 283).

Graveland (106) provides a table of literature citations covering 28 species of birds that consume calcium-rich material. Observations of two species exemplify the rich repertoire of calcium-obtaining behaviors. First, banded-tail pigeons drink from mineral springs with high calcium content and pick up salts that have dried on pebbles on the seashore (178). In the summer in Hope, British Columbia, CaCl₂ was used to reduce dust on gravel roads. Immediately after CaCl₂-coated pebbles were scattered on the road, large flocks of pigeons were seen picking them up. The pigeons did not frequent this area at any other time. Second, the red-cockaded woodpecker caches pieces of bone by wedging them into the bark of trees (230). The purpose does not seem to be to store the bone for times of need. Instead, caching may afford the birds some advantage by reducing the time they spend obtaining calcium from raptor pellets on the ground, which is particularly hazardous for a woodpecker (230).

The demand for calcium is especially high during egg laying (see sect. VE). The importance of calcium appetite to reproductive success has been underscored by recent work aimed at understanding the decrease in bird populations in regions afflicted by acid rain. Acidity reduces the abundance of many invertebrates, such as snails, crayfish, and insects (see Ref. 279 for citations). Moreover, poor soil quality due to acid rain reduces the calcium content of leaves and needles (73). This, in turn, reduces the calcium content of caterpillars, which are an important food source for many species of tree birds (73). Reduced calcium availability leads to egg shell abnormalities and the consequent decline in bird populations (e.g., Refs. 73, 279; see sect. VE2).

For obvious commercial reasons, there has been considerable interest in dietary calcium levels and calcium appetite in chickens. Intuitively, the laying hen seems to be a good model for studying calcium appetite because of its substantial calcium requirement for egg production. However, egg production quickly ceases if a laying hen is fed a low-calcium ration (103). Moreover, bone stores of adult hens are substantial, and calcium is conserved very efficiently, making a long period of deprivation necessary (see Ref. 131). Thus most investigators have studied growing chickens or the combination of growing and egg-laying chickens.

Calcium requirements are ~100 mmol/day for the growing hen and 2.5-5.0 mmol/day for the adult (citations in Ref. 102). Typical diets contain 750-1,000 mmol Ca²⁺/kg, with the calcium provided as ground limestone. Gilbert (102) has written an outstanding review of the factors affecting calcium balance during egg production of hens.

The earliest reported experiment on calcium appetite (119), by Hellwald in 1931, involved chickens that were deprived of dietary calcium. Four hours before receiving free access to calcium, some of the birds were fed pulverized egg shell (calcium) inside macaroni, and the others were fed macaroni alone. The macaroni prevented the birds from tasting the calcium. When egg shells were made freely available, the group prefed calcium-containing macaroni ate 27 g, whereas the group prefed macaroni alone ate 91 g. This demonstrated not only that calcium-deprived chickens develop an appetite for calcium, but also that it can be assuaged without oral stimulation (see sect. VB).

Calcium-deprived chickens that received a choice between two diets initially preferred the alternative containing calcium (131, 322). Similarly, chickens given free access to CaCO₃ granules consumed the supplement in inverse relationship to the calcium content of the diet (285). This may have practical significance, because birds given low-calcium diets but free access to CaCO₃ granules consumed more calcium and laid more eggs than did controls fed a calcium-adequate diet but not given a supplement. Similar results were found with hens given cockleshell grit as a calcium supplement (286).

Whereas it has been relatively easy to see an effect of diet calcium availability on the intake of a solid calcium supplement, there have been two failures to find a preference of calcium-deprived chickens for calcium solutions. In one (322), calcium-deprived and control chickens drank similar amounts of 92 mM CaLa. This was attributed to the high sensitivity of birds to unpalatable fluids. In the other (131), calcium-deprived chickens preferred to drink water rather than 4% calcium borogluconate. The same birds tested with a choice between solutions of 4% calcium borogluconate and 1.4 µM quinine drank mostly quinine, but this was probably due to learning (see sect. VA). On the other hand, calcium-deprived birds preferred to drink considerably more of a 100 mM CaCO₃ suspension during a 10-day test than did birds that had never been calcium deprived (131).

Estimates of the amount of calcium required by rats for maximal growth and bone accretion range from ~0.55 to 1.2 mmol Ca²⁺/day (22, 44). The National Research Council recommends a minimum daily calcium intake of 1.5 mmol Ca²⁺/day (205), and most rodent diets provide considerably more calcium than is required [e.g., AIN diets contain ~125 mmol Ca²⁺/kg (1, 227), grain-based laboratory diets contain ~200 mmol Ca²⁺/kg].

Given that extracellular fluid (ECF) is ~20% of the rat's body weight, and calcium concentrations are ~2 mM, a 250-g rat has ~0.1 mmol calcium in its ECF. This is equivalent to a few bites (0.5 g) of Purina Chow or only 1 ml of 78 mM CaLa! It is a gross oversimplification to think that all ingested calcium contributes to ECF, but this calculation points out that intake of very small amounts of food or calcium solution have the potential to produce dramatic changes in the rats' extracellular environment.

The skeleton contributes 10.9% of the weight of the rat and is 47.4% water (271). We typically find calcium content of dry femur ash to be ~8 mmol/g, which implies that total skeletal calcium content of a 250-g rat is ~115 mmol. However, not all this calcium is accessible. After near-total calcium deprivation of growing rats for 3 wk, femur calcium concentration drops slightly (5-15%), and femur weight is considerably decreased (~50% relative to controls of the same age; unpublished results). These figures suggest that rats have available ~50-60 mmol of calcium in bone.

The experimental study of calcium appetite in rats began with a series of four papers by Richter and students (241-243, 245). In parallel with his work on adrenalectomy and NaCl appetite, Richter took advantage of the fact that rats without parathyroid glands do not maintain calcium balance. He found that rats with parathyroidectomy (PTX) maintained on a low-calcium diet increased intake of 65 or 78 mM CaLa solution (241, 243; see also Ref. 316; Fig. 3). Rats with PTX also drank more of other calcium salts (acetate, gluconate, or nitrate) and were able to select an adequate amount of calcium when given a choice of four different salts to drink (78 mM CaLa, 268 mM KCl, 357 mM NaLa, 180 mM CaCl₂, and 290 mM NaH₂PO₄; Ref. 242).

View larger version (26K): [in this window] [in a new window]   Fig. 3. One of the first demonstrations that parathyroidectomy of rats increases calcium lactate intake and that this can be reversed by reimplantation of the parathyroid glands. [From Richter and Eckert (242).]

Several early studies found evidence for a calcium appetite based on calcium-deprived rats' choice between a calcium-free and calcium-containing diet. For example, deprived rats ate ~67% of their food as diet containing 375 mM CaLa, whereas replete controls ate only 33% (315). Similar findings have been made using several concentrations of CaCO₃ and other calcium salts (citrate, phosphate, and sulfate; Refs. 248, 263, 315). More detailed discussion of these and related studies is deferred until section V.

Drummond (74) reported that pigs fed a ration deficient in lime when let into a small brick-floored courtyard "spent a large part of their time attempting to lick whitewash from the walls or to root out a fragment of cement or mortar from the brickwork." This was in contrast to pigs deprived of vitamin A that "search(ed) for the smallest blade of grass or a chance weed which might be growing in the cracks of the floor," or a nutritionally complete control group, which apparently did none of these behaviors.

There are many observations that cattle eat old bones (e.g., Refs. 26, 27, 67, 105, 112, 284), but this appears to be due to phosphorus deficiency. Phosphorus-deficient cows do not consume calcium salts in cafeteria experiments (67, 112). Lactating cows fed a low-calcium diet for 9 wk and then give 1 h/day access to calcium phosphate ate ~2.5 times more supplement after the low-calcium period than before it (57). Despite this increase, the authors concluded that the cows did not have a calcium appetite because they did not consume the calcium requirements computed by the National Research Council. However, during the low-calcium diet period, body weight and plasma calcium levels increased. Milk production decreased, suggesting the cows were very efficient at conserving calcium. Also, the use of calcium phosphate as the test compound was unfortunate because calcium-deficient rats avoid phosphorus (see Refs. 242, 245, 262).

	IV. SPECIFICITY OF CALCIUM APPETITE

Top Previous Next References

The literature described above shows that calcium-deficient birds and rats preferentially ingest calcium when given the opportunity. However, several studies suggest that they also ingest other substances, which raises the question of how specific is the appetite for calcium? The evidence is summarized below.

There are scattered allusions that calcium deficiency causes pica (intake of nonfood objects). However, these appear to be based on incidental observations in the wild (e.g., Refs. 212b, 257a) without empirical evidence that the animals were calcium deficient or that this was the only deficiency they suffered. An engaging review of the history of pica in humans mentions only infrequently the consumption of chalk or other calcium-containing items as a diagnostic symptom of pica, and the popular theory was that intake of such substances combatted gastric acidity (213). Moreover, pica appeared to be most often recognized in adolescent girls rather than subpopulations with high calcium demand (i.e., young children, pregnant and lactating women). A recent finding that calcium supplementation reduced premenstrual cravings for sweets and salts (288; see also Ref. 287) is consistent with a connection between calcium and indiscriminate ingestion, although the reduction in cravings may have been secondary to changes in appetite, pain, or other premenstrual symptoms. Perhaps more relevantly, a study of pregnant urban African American women, who consumed <60% of the recommended daily allowance for calcium, found ~8% reported pica, but this was mainly pagophagia (ice chewing) or amylophagia (ingestion of laundry starch). None ingested chalk or antacid tablets (77). Thus, if calcium-directed pica occurs in humans, it must be very rare.

Evidence for pica related to calcium deficiency in birds is much stronger. Early investigators reported that chickens given a low-calcium diet in a lime-free room became greatly excited when fed, pecking at the feeder's buttons and fingernails (119), and that calcium-deprived chickens were "much more active" than controls and five times more likely to peck at novel objects (buttons, chalk, cake balls, acorns; Ref. 322). Egg-laying female great tits maintained without calcium spent considerable time searching on the ground and frequently dug in the soil. The birds showed particular interest in white-colored objects (109). This was not simply a general activation of behavior because the birds increased their search efforts on the ground but not in trees or elsewhere in the aviary (109). Laying red-cockaded woodpeckers repeatedly investigated white pieces of paper on the ground (230). These behaviors could be considered cases of calcium deficiency-induced pica, but at least from the birds' viewpoint they may be appropriate because calcium-rich material is often found on the ground and is frequently light-colored (i.e., snails, bones, grit).

Some birds eat grit when calcium deprived. An early study tested growing pheasants given access to a grain-based (low-calcium) diet and various forms of grit (185). Intake of quartz grit increased progressively from basal levels of ~1 to ~7 g/day. When bone meal or CaCO₃ was added to the pheasants' diet, quartz grit intake dropped back to ~1 g/day. In another experiment, pheasants ate <2 g/day grit obtained from glacial gravel mix, which looks like quartz grit but contains calcium. They were then switched to quartz grit, and intakes rose progressively to >6 g/day. When returned to the glacial gravel mix, intakes returned to baseline within 2-3 days. These results with growing birds were replicated with adults, although, consistent with their lower demand for calcium, the effects were muted (185; see also Refs. 185, 252).

These studies provide solid evidence that grit intake depends on calcium status, although grit is also consumed to help grind food in the crop. It seems likely that the intake of quartz by calcium-deficient pheasants is a mistaken attempt to gain calcium. The finding that quartz intake increases even though this does not alleviate calcium deficiency is consistent with the possibility that some species of birds innately recognize calcium as small lightly colored particles.

It should be noted that the ingestion of non-calcium compounds by calcium-deficient animals is often considered "aberrant," but the behavior is generally forced by the experimenter because calcium is unavailable. Calcium-deficient animals given a choice between a non-calcium compound and calcium will always consume the calcium, unless the calcium source is extremely unpalatable (e.g., Refs. 167, 172).

Strontium (Sr) and calcium are in the same periodic group, and it thus might be expected that animals generalize from the taste of one compound to the other, in the same manner that sodium-deprived animals ingest LiCl and KCl (84, 203). Moreover, strontium can sometimes act as a physiological surrogate for calcium (see Refs. 166, 282).

Voluntary intake of 0.75 mM SrCl₂ or 0.56 mM strontium lactate solution by rats was increased by PTX (242) and could be reversed by reimplanting parathyroid glands. Dietary calcium deficiency increased the intake of 1 and 10 mM SrCl₂ solutions by rats in 24-h acceptance tests (48), but not 100 mM SrCl₂ in a 30-min test after water deprivation (49). Chickens could discriminate between diets containing 135 mmol/kg SrCO₃ and 200 mmol/kg CaCO₃ but not after flour had been added to hide the color (131). This strongly suggests that chickens cannot easily discriminate between the taste of strontium and calcium (see also sect. VA).

Like strontium and calcium, magnesium is a group IIa element, and thus it might be expected that animals confuse the three. However, magnesium has an intensely bitter taste. Calcium-replete rats showed a slight preference for very low concentrations of MgCl₂ (0.005-10 mM) over water in 24- or 48-h two bottle choice tests, and strongly rejected concentrations above ~50 mM (48, 294). Nevertheless, calcium-deprived rats drank more 10, 100, and 316 mM MgCl₂ than did replete controls (48), and rats with PTX increased intake of 61 mM MgCl₂ in one experiment (167) and 25 mM MgCl₂ (but not 0.69 mM magnesium lactate) in another (242). The increased intake of 25 mM MgCl₂ was reversed by reimplanting parathyroid glands (242; see also Ref. 168 and sect. VF2).

Phosphorus counteracts some of the physiological actions of calcium and consequently exacerbates the effects of calcium deficiency. It thus makes sense for calcium-deficient animals to avoid it. Consistent with this, rats fed a low-calcium diet showed lower intakes than did controls of several concentrations of sodium phosphate and potassium phosphate (48). Similarly, PTX reduced rats' intake of 37 mM dibasic sodium phosphate, and this could be reversed by administration of parathyroid extract or dihydrotachysterol (78, 242, 245, 316). The effects on phosphate intake were more immediate and greater than those seen in parallel experiments on calcium intake (75, 242, 243, 245, 316).

In contrast to the results with deprivation and PTX, during pregnancy and lactation phosphorus intake increases (36, 240). Perhaps this is the result of different physiological mechanisms at play in the different models of calcium intake. However, a simpler possibility is that the decrease in phosphate intake occurs only when calcium is scarce, which was not the case in the studies of ingestion during reproduction.

Snowdon (274) has argued that calcium deficiency is a cause of lead poisoning in children. This is based on findings that calcium-deprived rats (274, 275) and monkeys (137) ingested more lead acetate solution than did controls. Deprivation of other minerals (magnesium and zinc, but not iron) also increased lead acetate intake, but to a lesser extent (274).

Lead acetate is toxic so the question arises of why do calcium-deprived rats drink a solution with deleterious effects? Lead is easily incorporated into tissues of calcium-deficient animals (164), suggesting that in the absence of sufficient calcium, the animal may be able to use lead as a surrogate. Also, lead can elevate concentrations of 1,25-dihydroxyvitamin D [1,25(OH)₂D; Ref. 98], which increases calcium absorption and thus reduces the severity of calcium deficiency (see sect. VD). Consistent with this, calcium-deprived rats given access to lead acetate solutions gain more weight than do rats maintained on the same diet without lead to drink (274). Moreover, a flavored drink that was followed by an injection of lead acetate was avoided by calcium-replete but not calcium-deprived rats (274). Either the toxic effects of lead ingestion by calcium-deprived animals do not support a conditioned aversion, or the benefits of drinking lead to the calcium-deprived animal outweigh its deleterious effects.

Naive calcium-deprived rats licked more frequently than did replete controls for 20 mM lead acetate solution in a brief-exposure test (49). Such a rapid effect on intake argues that the response to lead acetate is driven by taste factors. Lead acetate has been called "lead sugar" and labeled as sweet by humans (see discussion in Ref. 274). However, calcium-replete rats avoid it in long-term tests (180) and do not generalize between lead acetate and sucrose (198). Even if lead acetate tasted sweet, this would be an unlikely cause for the increase in intake because calcium deprivation reduces intake of sweet solutions (see below). One explanation for the high intake of lead acetate is that it may taste like calcium. Morrison (198) trained rats to respond to "test" salts by pressing a lever associated with bitterness, sourness, or sweetness. CaCl₂ and lead acetate showed similar response profiles, which differed from the other 16 salts tested (198).

The relationship between calcium intake and sodium intake has been investigated in detail, in part because of the known link between calcium and sodium excretion (e.g., Refs. 34, 99, 148, 215), and in part because of the involvement of both calcium and sodium in hypertension (see reviews in Refs. 115, 186, 231, 276). NaCl intake is inversely related to both the amount of calcium in the diet and the duration of calcium deprivation (292, 305). Feeding a low-calcium diet to growing male rats progressively increased voluntary intake of 300 mM NaCl from control levels of ~10 to >60 ml/day after ~14 days. More severe calcium deprivation (>28 days) increased 300 mM NaCl intake to impressive levels (>125 ml/day; Ref. 305). Dietary calcium deprivation increased intake of a range of concentrations of NaCl (50-500 mM) (305) and several sodium salts including sodium acetate, bicarbonate, and glutamate (48, 290). Most observations were made with Sprague-Dawley rats, but the phenomenon has been demonstrated in other strains, including Fisher 344 rats, which are generally reluctant to drink NaCl (291), spontaneously hypertensive rats (SHR), and Wistar-Kyoto rats (WKY). The SHR appeared to be particularly susceptible to the effects of dietary calcium. Moreover, supplementing the calcium content of the SHR's diet decreased their already high NaCl intakes almost to levels of WKY controls (292).

The effects of PTX on NaCl intake are not as clear as are those of dietary calcium deprivation. Richter and Eckert (242) reported that with rats given PTX, "numerous single and multiple choice experiments made with sodium and potassium solutions... did not reveal cravings for these minerals." Similarly, rats with PTX did not increase NaCl intake when given a choice between 500 mM NaCl and 61 mM CaCl₂ solution (167). On the other hand, there are two reports, including one from Richter's laboratory, that rats with PTX increase NaCl intake in cafeteria choice experiments (76, 78), and a demonstration that rats with PTX pressed a bar to obtain NaCl in preference to CaCl₂ (172). One reason for the discrepancies is that animals with PTX given a choice between calcium and sodium solutions may ingest enough of the calcium solution to negate their calcium deficiency, and thus do not always need calcium.

Several studies have examined potential physiological mechanisms underlying calcium deficiency-induced NaCl intake. The renin-angiotensin-aldosterone system (RAAS) is the primary mediator of NaCl intake (67, 79, 88, 92), but indicators of RAAS activity, including blood pressure, urinary sodium excretion, and ECF volume, appear unperturbed in calcium-deficient rats (297, 303, 305). Moreover, treatment with an aldosterone antagonist or angiotensin II receptor inhibitor had no effect on the NaCl intake of calcium-deprived rats (303). Adrenalectomy, which generally increases NaCl intake, decreased the already high NaCl intakes of calcium-deprived rats, even during aldosterone replacement therapy (296). Thus calcium deficiency is arguably the clearest example of high NaCl intakes where a role for the RAAS has been effectively ruled out (see review in Ref. 297).

It seems more likely that the high intakes of NaCl produced by calcium deprivation are due to increased NaCl palatability (49, 189) and learning. Immediately after calcium-deprived rats drank NaCl they had increased plasma ionized calcium concentrations and reduced concentrations of PTH and 1,25(OH)₂D (298). In vitro studies indicate that physiological levels of sodium modulate the binding of calcium to plasma proteins. Small increases in plasma sodium concentrations release bound calcium into the ionized pool, which has the effect of temporarily reducing the severity of calcium deficiency, and thus reinforcing NaCl drinking behavior. Of course, the effect of drinking NaCl on plasma ionized calcium is only temporary, and the excess NaCl must be excreted, carrying calcium with it. Thus the rat is in a vicious cycle in which it exacerbates its deficiency because it attempts to benefit in the short term (see Refs. 297, 298).

Relative to replete controls, calcium-deprived rats have greater intakes of several chlorides (aluminum, ammonium, ferric, ferrous, potassium, zinc) but not nonsodium nonchlorides (ferrous sulfate, magnesium sulfate, potassium gluconate; Ref. 48).

Relative to replete controls, calcium-deprived rats have reduced intakes of several sweet solutions, including sucrose, saccharin, D-phenylalanine, aspartame, and cola beverage (48, 49, 190, 290, 293, 305; see also Ref. 39). They also avoid a sweet, solid carbohydrate when allowed to choose between separate sources of carbohydrate, fat, and protein (290). The reduced appetite for sweetness does not appear to be due to a general loss of appetite for carbohydrates or calories. Intake of Polycose, alcohol, and several fats was either increased or unaffected by calcium deprivation (290). The effects are reversible. The reduction in daily sucrose intake produced by feeding a calcium- or mineral-deficient diet could be eliminated by feeding bone meal or calcium gluconate (190).

The mechanism for these changes in sweet solution intake is unknown. One possibility is that they are mediated by changes in taste sensitivity. The sweeteners tested have diverse postingestive effects, and calcium-deprived rats ingest less saccharin or saccharin plus glucose mixture than do controls in brief-duration tests, which minimize postingestive effects (49, 298). Sweeteners increase intracellular calcium content of gerbil taste cells, and this depends on the presence of extracellular calcium (306). A complex relationship between extracellular calcium and sweetness perception has been observed in other species (157). It is thus conceivable that extracellular calcium modulates sweet taste transduction. It is also possible that changes in the concentrations of circulating hormones or other factors induced by calcium deficiency, rather than low levels of calcium per se, could be responsible for the changes in preference.

The possibility that a more general mechanism is involved is underscored by findings that deficiencies of sodium and zinc also influence sweet solution acceptance (14, 41, 42, 223, 237). This is not a function of general malnutrition because no changes in saccharin intake were observed in rats deprived of iron, magnesium, or phosphorus (293). One possibility is that deficiencies in calcium and other minerals interfere with CNS opioid activity, which governs intake of sweet compounds (see sect. VC1 and Ref. 290).

Calcium-deprived rats and controls had similar intakes of 0.368 mM sucrose octaacetate (bitter) and 2.5 mM citric acid (sour) in two experiments (293, 305). Calcium deprivation increased intake of 0.026 µM quinine sulfate but not higher concentrations (0.13-0.54 µM; Ref. 274, see also Ref. 48). Perhaps this extremely dilute quinine produces bitterness similar to the bitter component in the taste of calcium. Calcium deficiency had no effect on the intake of a number of other compounds, including citric, hydrochloric, and sulfuric acids (48).

	V. MECHANISMS UNDERLYING CALCIUM INTAKE

Top Previous Next References

There is strong evidence that the appetite for calcium involves an innate component. Most critically, calcium-deprived rats recognize and ingest novel calcium solutions within seconds of receiving them (49). They also sham-ingest CaCl₂ solutions even though this provides little or no postingestive benefit (189) and show latent learning for calcium solutions; that is, they associate an arbitrary taste with calcium consumption even though they are not calcium deficient and thus have no motivation to learn this association (50).

Nevertheless, it is clear that calcium-deprived animals can learn to prefer calcium-containing foods and drinks. In a series of studies using different calcium salts (CaCO₃, CaHPO₄ or CaSO₄), Scott et al. (263) fed rats diets with or without calcium and with or without anise flavor, according to a 2 × 2 design. All groups then received a choice between flavored diet and the unflavored diet of the opposite calcium content for 5 days, followed by 5 days with the initial diets. The results are difficult to extract from the paper, but it appears that the calcium-deprived but not replete rats' initial preferences were eliminated when the diet choice was first switched, and these were reinstated when the flavors were switched back again.

A simpler study was conducted by Rodgers (248), who tested rats maintained for 3 wk on one of two calcium-deficient diets. The rats then received a choice between one diet with 133 mmol/kg CaCl₂ added and the other without calcium. All the rats preferred the novel diet. This was surprising because one group ate novel calcium-deficient diet in preference to "familiar" maintenance diet with added calcium. This paper also included a demonstration that, in contrast to the results with calcium, sodium-deprived rats consumed a sodium-containing food, irrespective of its novelty. This disparity was the main impetus behind the conclusion that sodium appetite is purely innate, whereas calcium appetite is purely learned.

Given the repressive effect of this study on subsequent research into calcium appetite, it is worthwhile noting some of its shortcomings. First, no independent measures that the rats were calcium deficient were collected, and given that they consumed at least 1.325 mmol calcium in the first day of the 4-day test, it is unlikely that intakes during the test reflected calcium deficiency. Second, the design did not involve counterbalancing the complex diets used. It is impossible to know whether the low intake of calcium-containing diet is due to its bad taste or the rats' desire for novelty. Given that calcium-containing diets are unattractive to replete rats, it may be that the rats consumed as much calcium as they required over the first few hours of the 96-h test and then avoided the calcium source. Third, even if the experiment was methodologically sound, the conclusions do not follow from the results. A demonstration that learning can occur does not preclude innate mechanisms from also being involved.

The role of learning in determining calcium intake has been examined much more thoroughly in poultry. Wood-Gush and Kare (experiment 2 in Ref. 322) noticed that calcium-deprived chickens had more difficulty selecting a high-calcium diet on the day after food cup position was rotated than on the second day of each 2-day test. This led Hughes and Wood-Gush (131) to examine the contribution of learning to calcium appetite. Their most convincing demonstration involved calcium-replete and deprived chickens given a choice between diets containing 135 mmol/kg SrCO₃ and 200 mmol/kg CaCO₃. In one experiment, flour was added to the CaCO₃ source to hide its color. In another, a pink dye was added to the CaCO₃ source to make discrimination easier. With no color cues available, most birds showed positional preferences but not diet preferences. On the other hand, with a color cue present, 8 of 11 deprived birds but only 2 of 12 replete birds preferred the CaCO₃- to SrCO₃-containing diet. When, after a 10-day test period, the color was switched from the CaCO₃ diet to the SrCO₃ diet, the replete birds continued to eat ~50% from each cup, whereas the "calcium-deprived" group initially switched preference to the SrCO₃ diet (i.e., followed the color), but this preference dissipated to indifference within 3-4 days. The simplest explanation for this is that when initially given the choice, the calcium-deprived birds were able to 1) determine that the CaCO₃-containing diet was beneficial and 2) associate the diet's appearance with its beneficial effects.

These results were confirmed using slightly different methods (130). An extensive series of studies were conducted to characterize potential mechanisms that could account for the learned preference for calcium. Four hypotheses concerning the benefit that chickens could potentially derive from consuming calcium were tested. These included a general reduction in anxiety or arousal, the reversal of bone calcification, and thus reduced bone pain, a nonspecific general sense of well-being or a learned aversion to the deficient (maintenance) diet. There was no unequivocal evidence found for any of these mechanisms. Thus, although it is clear that chickens can associate the chemosensory aspects of calcium with its postingestive effects, the nature of the unconditioned stimulus remains obscure.

Learning is involved in two other areas of calcium appetite research. First, there is strong evidence that cues about calcium can be transmitted by social learning (see sect. VF4). Second, learning has been used to determine the strength of the rat's drive for calcium. Frumkin (97) gave intact rats and rats with PTX a series of daily 1-h tests with access to 81 mM CaLa, 2.4 mM saccharin, or both solutions simultaneously. The rats were poisoned with LiCl each time they drank one or other of the solutions. Intact rats poisoned after drinking CaLa stopped drinking this solution but rats with PTX continued to drink it. This was not because the calcium-deficient rats failed to learn to avoid a solution associated with illness because all animals poisoned after drinking saccharin stopped drinking it. Thus it appeared that the motivation to drink calcium of rats with PTX was greater than their learned aversion to the solution (see also Ref. 274). It is noteworthy that the toxic effects of lithium can be ameliorated by calcium supplementation (152).

Taste plays a dual role in the control of calcium intake. First, the taste of calcium and its vehicle (i.e., the rest of the diet) provides cues that can direct intake and be associated with the postingestive consequences of consuming calcium (see sects. II and VA). A good example of this is work showing that most rats with PTX given a cafeteria selection of nutrients could not select adequate amounts of calcium gluconate solution following lesions of ventrobasal thalamic gustatory regions (78). Second, the oral acceptability of calcium is mediated by calcium status. Whereas replete animals largely ignore concentrated calcium solutions, calcium-deprived ones ingest them avidly within seconds of receiving them (49), and sham ingest large volumes even though this provides little or no postingestive benefit (189).

Electrophysiological recordings from gustatory afferent nerves suggest that physiological calcium status influences calcium perception in the periphery (135, 158). In the frog, sensitivity to oral calcium depends on the concentration of calcium inside the tongue. The glossopharyngeal nerve response to water and a 1 mM CaCl₂ taste stimulus was positively related to the concentration of CaCl₂ infused through the lingual artery (158). The responses to quinine, HCl, and ethanol were either unaffected or decreased by lingual artery CaCl₂ infusion. In the rat, sensitivity to oral calcium is influenced by diet. The threshold concentration of oral calcium that elicited an electrophysiological response in the chorda tympani nerve was an order of magnitude lower in calcium-deprived than replete rats. The calcium-deprived animals also displayed greater integrated activity in response to low concentrations (<3 mM) of oral CaCl₂ and CaLa. In contrast, calcium-deprived rats were less sensitive than were replete controls to high (>32 mM) CaCl₂ concentrations. These results are consistent with the idea that the increased chorda tympani response at low calcium concentrations enhances the animal's ability to detect calcium, and the reduced response to high CaCl₂ concentrations lowers the perceived intensity of calcium and thus increases the acceptance of these normally unpalatable solutions (135). This link between calcium status and palatability is reinforced by recent data showing that calcium deprivation specifically increases the activity evoked by oral calcium of sucrose-best units in the nucleus tractus solitarius (188).

The mechanism by which calcium status influences gustatory sensitivity is unexplored. Richter and Eckert (242) believed that the craving for calcium depended on "chemical changes in the taste mechanisms in the oral cavity, making the calcium more desirable after parathyroidectomy than before," but this was apparently based solely on conclusions from his concurrent studies of sodium intake in rats with adrenalectomy (cf. Refs. 236, 238, 239). One obvious possibility is an effect of calcium on saliva composition. Diet calcium supplementation increases salivary calcium concentrations of children (23), although generally there is little or no relationship between calcium levels in plasma and saliva (72, 253). Alternatives include the direct modulation of transduction by extracellular calcium (see sect. IIB), or an effect of calciotropic hormones on taste mechanisms. There are parallels here to the literature on the modulation of sodium taste by sodium status (e.g.,