INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

POLYCYSTIC KIDNEY DISEASE (PKD) is the fourth most common cause of end-stage renal disease, with treatment costs exceeding a billion dollars per year in the United States (22). Human forms of this disease group have been linked to mutations of genes on chromosome 16 (25) and 4 (40) for dominant forms and 6 for the recessive form (53). The genes for the dominant forms are known to encode trans-membrane proteins that may interact with each other (42) and whose structure would be consistent with a role in cell-cell or cell-matrix signaling, but no definite function has yet been identified. Tubular dilatation is associated with increased rates of both epithelial proliferation (33) and apoptosis (51).

The Han:SPRD-cy rat is a model of PKD that shares autosomal dominant inheritance, progression through early adult life, and sexual dimorphism with human disease (14). The disease is characterized by progressive dilatation of nephrons in young animals, associated with marked interstitial inflammation and fibrosis, with associated nephron loss in older animals (5). Unlike human PKD, Han:SPRD-cy rat PKD has proved amenable to treatment. Modification of this model of PKD has been achieved with a variety of environmental manipulations, including dietary protein restriction (36), angiotensin-converting enzyme inhibition (37) or angiotensin receptor blockade (27), salt loading (27), methyl-prednisolone therapy (20), or hypocholesterolemic therapy (21).

The interstitial pathology of Han:SPRD-cy rat PKD bears many resemblances to other forms of chronic renal injury, including the five-sixths nephrectomy (1) and the spontaneous nephropathy of the Fisher 344 rat (30). These conditions also respond favorably to protein restriction and have also been ameliorated by alteration of the type of protein in the diet to a soy base (44, 49). This report describes studies undertaken to test the hypothesis that Han:SPRD-cy rat PKD would also be modified by the substitution of soy protein into the diet. As amelioration of Han:SPRD-cy rat PKD has been reported with alkali supplementation of the diet (46), urinary ammonium excretion was measured as an indicator of the impact of dietary changes in acid-base metabolism (10).

Previous studies of dietary intervention have provided few insights into biochemical alterations that might be related to altered pathology. Proton nuclear magnetic resonance (¹H-NMR) spectroscopy is powerful tool for the identification of compounds relevant to renal physiology and pathophysiology (3, 11, 16, 19, 28, 35, 47, 52). In a previous study using ¹H-NMR spectroscopy and imaging (38), we have shown that Han:SPRD-cy rat PKD is associated with an early alteration in renal tubular physiology that leads to urinary losses of a range of organic anions and amines, including some that may influence cell energy states. Similar although not identical changes have been reported in other forms of renal injury (3, 19, 28, 52). We used ¹H-NMR spectroscopy to explore the influence of a soy protein diet on renal biochemistry in the Han:SPRD-cy rat to determine whether altered disease expression could be associated with chemical change that might provide insight into mechanisms of renal injury in this model.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Han:SPRD-cy rats. Han:SPRD-cy rats were obtained from our own breeding colony that is derived from animals that were kindly provided to us by Dr. Benjamin Cowley (Univ. of Kansas Medical Center, Kansas City, KS). All animal procedures and care were examined by the University of Manitoba Committee on Animal Use and certified to be within the guidelines of the Canadian Council on Animal Care. Surviving male offspring of known Han:SPRD-cy heterozygotes were used in this study. Two-thirds of these animals would be expected to be heterozygous, as homozygotes in our colony rarely survive beyond weaning.

Rats in this study were placed on an experimental or control diet at weaning and fed ad libitum. The diets were obtained from ICN Biochemicals (Montreal, ON, Canada). The control diet was a standard casein-based diet (22.9% casein purified high nitrogen) with corn starch, sucrose, and corn oil (catalog no. 960260). The experimental diet substituted heated soy protein for casein. Both diets were supplemented with 0.3% DL-methionine. The animals were killed by pentobarbitone overdose after 8 wk on the diet, and blood and kidneys were collected for study.

Histology. Tissue from the left kidney was used for histological and immunohistochemical analysis. This tissue was fixed in 10% formalin for 60 min prior to embedding in paraffin and sectioning at 5 µm. Sections for measurement of cystic volume and qualitative study of renal histology were stained with hematoxylin and eosin. Sections for quantitative analysis of fibrosis were stained using methyl blue alone in adaptation of Masson's trichrome stain (13). Animals were classified as affected if a single longitudinal cross section of the kidney contained at least 10 areas of tubular dilatation with associated increase in extracellular matrix. High-power microscopic examination was used in cases of mild disease expression to ensure that lymphatic or empty vascular spaces were not being erroneously counted. Classification was confirmed by two experienced observers (N. Bankovic-Calic and M. R. Ogborn).

Cell proliferation was studied using immunohistochemical detection of proliferating cell nuclear antigen (PCNA). Antigen detection was enhanced by microwave preheating in citric acid buffer (33). Sections were incubated with a monoclonal anti-PCNA antibody (Dako M 0879; Dako, Glostrup, Denmark) at a dilution of 1:50 for 90 min. Secondary detection was achieved with a Vectastain Elite rat adsorbed anti-mouse immunoglobulin G kit (Vector Laboratories, Burlingame, CA) with peroxidase-diaminobenzidine color development.

Macrophages were identified using a monoclonal antibody against a 90- to 100-kDa protein that is expressed on lysosomal membranes and has many characteristics of the human CD68 antigen (Chemicon MAB1435; Chemicon International, Temecula, CA) at a dilution of 1:100 for 90 min. Secondary detection was achieved in the same way as for PCNA.

Apoptotic nuclei were identified using a cell death detection kit (Boehringer 1 684 817; Boehringer, Mannheim, Germany) with nickel-peroxidase color development using the manufacturer's procedures.

Image analysis. Image analysis procedures were performed with a system consisting of a Cohu high-resolution black and white camera connected to a computer via a PCVision-Plus video capture board. Images were captured using the Image Pro software package (Phoenix Biotechnology, Seattle, WA).

Renal volume was determined using the Cavalieri principle, as we have described previously (36). Parallel 2.5-mm sections of the left kidney were cut using a Lucite guide box constructed for this purpose. A video image of all sections, including a linear scale, was captured, and the area of the sections integrated using module 2500 of the Imagemeasure software package (Phoenix Biotechnology) and renal volume were calculated as the product of total section area and section thickness.

Measurement of relative tubular luminal area, i.e., the fraction of tissue section occupied by tubular lumen, was performed fluorometrically using the IM4100 module of the Image Measure software package (Phoenix Biotechnology), using low-power microscopic images captured via a Cohu high-resolution black-and-white video camera. Sections were viewed through a ×2 objective and Nikon television relay lens. The samples were illuminated with a wide-aperture condenser to ensure uniform lighting conditions. The profiling tool within the program was used to ensure uniform lighting of the captured video image. A 64 × 64-pixel rectangle was moved in an alternating horizontal and vertical path through the section from a random starting point until 50 measurements from each of four separate whole kidney tissue cross sections had been collected. The inner fluorometric threshold was adjusted for each slide to include all gray scale values equal to or greater than open tubular lumen. Some cysts contained eosinophilic debris, but this material had gray scale values less than that of solid renal tissues and did not influence results. The outer threshold was set to 1 (black) to include all stained areas of tissue, and the relative tubular luminal area was calculated according to the following formula where A is relative area ratio, and X_i and X_o are number of pixels within inner and outer threshold, respectively.

An average of 50 measurements from three to five different sections was used to determine the cyst area ratio. A measurement was accepted if the 95% confidence intervals of the mean were within 2% of the mean. Renal fibrous volume was measured in a similar way, using the proportion of section areas that had taken up methyl blue stain, as measured in densitometry mode of module 4100 of the Imagemeasure software package. In densitometry mode, the thresholds are reversed, with the outer threshold set to 255 (white) and the inner threshold set 1 gray scale unit over the actual measured value of an area of interstitial methyl blue staining. The product of the proportion and the reference renal volume, corrected to body weight, gives the final volume occupied by either renal cyst volume or renal fibrous tissue (48).

Macrophage numbers, PCNA-positive cells, and apoptotic cells were quantitated by counting, using module 2500 of the Imagemeasure software package. The macrophage counts were reported as a mean per high-power video field (×40 microscope objective) with at least 50 fields counted. Apoptotic cells and PCNA-positive cells were reported per millimeter of epithelium. The length of the apical surface of a randomly selected piece of epithelium was measured from a captured video image, using the same software module. Cells along the length of this piece that exhibited the appropriate nuclear staining characteristics were then counted. The score for the kidney was determined as the mean count divided by the mean length from 50 measurements.

Chemistry. Creatinine, serum albumin, urea, and cholesterol were determined in serum by spectrophotometric methods using Sigma kits (Sigma Chemical, St Louis, MO). Urinary ammonium was measured by an automated Vitros analyzer in the Clinical Chemistry Laboratories, Health Sciences Centre (Winnipeg, MB). Urine ammonium excretion is expressed as the absolute amount of ammonium excreted during the collection period, corrected to body weight.

¹H-NMR spectroscopy. Urine for ¹H-NMR spectroscopy was collected over a 6-h period commencing at 0800 h on the day before the animals were killed. Urine was collected under mineral oil to eliminate evaporation or change in pH due to formation and disassociation of carbonic acid. Urine volume was recorded, after which the samples were frozen at 70°C prior to NMR analysis. Samples were prepared for ¹H-NMR spectroscopy by the addition of 100 µl of ²H₂O containing 0.94 mM sodium-d4-(trimethylsilyl)propionate (TSP) as an internal reference to 500 µl of urine in a 5-mm NMR tube.

The right kidney was quickly removed when the animals were killed and then placed immediately in liquid nitrogen prior to storage at 70°C. Frozen whole kidney tissue was weighed, lyophilized, and then reweighed to determine water content. The dried tissue was pulverized under liquid nitrogen and then homogenized on ice in 0.3 M perchloric acid (10:1 vol/tissue weight) (39). Tissue debris was removed by centrifugation at 20,000 g for 10 min at 4°C to precipitate insoluble components. The supernatant from this procedure was adjusted to a pH of 7.2 with KOH and HCl. The sample was centrifuged at 20,000 g for 10 min at 4°C to precipitate KClO₄. The supernatant was frozen at 70°C and then lyophilized. The lyophilized sample was reconstituted in 1.5 ml of ²H₂O, containing known amounts of TSP as an internal standard of concentration and chemical shift. The pH was adjusted to 7.25 with appropriate deuterated compounds. Samples were refrigerated at 4°C for 1 h, then centrifuged at 4°C for 25 min at 16,000 g to remove additional salt. Samples were then transferred to 5-mm NMR tubes.

¹H-NMR spectra of urine or tissue extracts were obtained at 500 MHz (11.7 T), using a Bruker AMX500 spectrometer locked to the ²H₂O deuterium resonance and operating with a probe temperature of 310°K. Each urine spectrum was accumulated as the sum of 160 free induction decays acquired into 16,384 data points following an 80-degree pulse, using a sweep width of 7,042.25 Hz (digital resolution, 0.21 Hz/point), an acquisition time of 1.16 s, and a relaxation delay of 10 s for a total pulse recycle time of 11.16 s. The water signal was suppressed by presaturation during the latter 5 s of the relaxation delay. The extract spectra were obtained with similar parameters, except that a relaxation delay of 2 s was used and the presaturation time was 2 s. To compensate for the shorter relaxation delay, a 30-degree pulse was used. An exponential line broadening of 0.3 Hz was applied prior to Fourier transformation to yield the spectrum. Spectral peak positions were measured relative to the TSP peak (0 parts/million). Urine results were corrected to urine creatinine to minimize the impact of alterations in glomerular filtration rate and urine concentrating ability.

Peak assignments were based on previously published spectra (19, 28, 39, 47, 52) and by comparison with spectra of authentic compounds. For each metabolite of interest, an isolated peak or group of peaks was selected, and the intensity was measured relative to that of the TSP peak. The concentration of the metabolite in each sample was determined from this ratio, and the known concentration of TSP was used, after correcting for the relative number of protons contributing to the resonances and for the tissue weight or urine volume.

Statistical analysis. Statistical analysis was performed using one-way analysis of variance with the Bonferroni test for post hoc testing to compare the effect of dietary treatments in normal and affected animals, Student's t-test with the Welch correction or the Mann-Whitney test in samples with unequal variances using the Prism 2 software package (GraphPad Software, San Diego, CA).

	RESULTS

Top Abstract Introduction Methods Results Discussion References

A total of 15 affected animals were identified in the group receiving the control diet, and 19 affected animals were found in the soy diet group. Four normal animals received the casein diet, and five normal animals received the soy diet for the 8-wk period. Casein-fed animals demonstrated more rapid weight gain over the 8-wk feeding trial (P < 0.0001) (Table 1). This effect was seen to the same degree in unaffected animals.

Affected animals in the casein diet group had significantly higher serum levels of urea (P < .0001), creatinine (P = 0.002), and cholesterol (P = 0.028). There was a trend toward higher serum albumin in the soy protein-fed animals, but this did not reach statistical significance (Table 1). Soy-fed animals demonstrated significantly lower urine ammonium excretion, implying reduced renal acid excretion (Table 1).

The soy protein-fed animals demonstrated a significant reduction in renal cyst volume that was associated with reduced frequency of both apoptotic and proliferating tubular epithelial cells (Table 2). These changes were accompanied by a marked reduction in the amount of renal interstitial inflammation and fibrosis (Table 2). The casein-fed animals demonstrated diffuse tubular dilatation with surrounding areas of inflammation and fibrosis (Fig. 1A). The soy-fed animals demonstrated a more focal pathology, with areas of typical cystic change interspersed with regions in which normal renal architecture was preserved (Fig. 1, B and C). Renal water content corresponded to extent of cystic change with kidneys from casein and soy-fed normal animals containing means of 75.8 and 74.8% water, respectively, with kidneys from affected animals containing 86.9 and 78.3% water, respectively.

View larger version (112K): [in this window] [in a new window]   Fig. 1.   A: kidney (hematoxylin and eosin stained) from an 11-wk-old casein diet-fed animal, demonstrating diffuse tubular dilatation and extensive interstitial inflammation and fibrosis; magnification, ×20. B: kidney (hematoxylin and eosin stained) from 11-wk-old soy protein diet-fed animal, demonstrating focal tubular dilatation and corresponding localized areas of interstitial inflammation and fibrosis; magnification, ×20. C: higher-power (×100) detail of B, showing characteristic flattened epithelium and interstitial changes seen in a focal pattern in soy-fed animals.

Soy protein feeding was associated with significant increases in urinary excretion of citrate and dimethylglycine in normal animals. Representative ¹H-NMR spectra are shown in Fig. 2. A trend toward increased excretion was noted with other organic anions relevant to the citric acid cycle and methylamines, which did not reach significance because of the small number of normal animals studied. Soy protein feeding in affected animals was associated with significant increases in excretion of citrate, formate, -ketoglutarate, succinate, dimethylglycine, trimethylamine-N-oxide, and creatinine, compared with casein-fed affected animals (Table 3).

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Representative proton nuclear magnetic resonance (1H-NMR) spectra of urine from soy-fed (A) and casein-fed (B) Han:SPRD-cy rats. Peaks used for analysis are, from right, as follows: 1) alanine, 2) acetate, 3) succinate, 4) 2-oxoglutarate, 5) citrate, 6) dimethyl-amine, 7) dimethylglycine, 8) 2-oxoglutarate, 9) creatinine, 10) trimethlyamine-N-oxide, 11) taurine, 12) creatinine, 13) allantoin, 14) urea, 15) hippurate, and 16) formate.

Informative spectra of tissue perchloric acid extracts were not obtained from two normal animals receiving the casein diet. To improve statistical power, results of the remaining normal animals were pooled with those of normal animals fed the diet for a 4-wk period in a pilot phase of the study, as tissue profiles were identical at that time point. Results from four normal animals fed the casein diet and four fed the soy diet were included in this fashion. Analysis of perchloric acid extracts demonstrated that affected soy-fed animals maintained higher content per dry weight of tissue of succinate, despite increased urinary losses (Table 4). Representative spectra are shown in Fig. 3. Casein-fed animals retained higher tissue content of total cholines and taurine but had significantly lower levels of betaine, irrespective of disease status. Allantoin was significantly higher in affected casein animals than both normal casein animals and affected soy animals (P < 0.0001) but seemed to follow the spectrum of histological disease.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Representative 1H-NMR spectra of perchloric acid extracts of renal tissue from (A) soy-fed and (B) casein-fed Han:SPRD-cy rats. Peaks used for analysis are, from right, as follows: 1) lactate, 2) alanine, 3) acetate, 4) glutamate, 5) succinate, 6) cholines, 7) betaine, 8) taurine, 9) inositol, and 10) allantoin.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Dietary modification of renal disease remains an important experimental approach, despite recent disappointing results in short-duration human intervention studies (32). Most studies have considered the effects of modification of the amount of dietary protein. Some studies have shown, however, that alteration of the source of dietary protein may also have a profound influence on the expression of experimental renal disease. Williams and Walls (49) demonstrated improved survival and preserved renal function with soy protein substitution in a remnant kidney model of chronic renal failure. Iwasaki et al. (26) demonstrated similar benefit in the spontaneous nephropathy of the Fisher 344 rat. Shimokawa et al. (44) confirmed that this finding represented a benefit of soy, rather than a toxic effect of casein. The above studies did not extensively characterize pathological change and tended to emphasize glomerular pathology. Both of the above models, however, have marked tubulointerstitial pathology (1, 30). In general, tubulointerstitial changes correlate better with renal function and prognosis than glomerular change (43). The Han:SPRD-cy rat model of renal injury is clearly one in which tubulointerstitial change predominates (5). We have previously shown that reduction in dietary protein content modifies both tubular (36) and interstitial (6) architecture. Modification of the source of dietary protein content in this study has had a quantitatively larger effect than simple protein reduction. Our previous studies comparing 8 and 20% casein diets demonstrated persistence of a diffuse tubulointerstitial pathology, albeit to a lesser degree. Focal cystic and inflammatory lesions, with intervening tracts of normal renal parenchyma, appear specific to this modification. The impaired growth seen in soy-fed animals in this study raises concerns about nutritional adequacy. This raises the question as to whether the observed changes are simply due to reduced protein intake. The diet was prepared in a fashion for which we were cognizant of the issues of methionine deficiency and soy trypsin inhibitor content, which led to supplementation and heat treatment respectively (29). Animals were not pair fed, and a palatability issue may have contributed to the observed difference in weight gain. In addition to the above-mentioned qualitative differences in pathology produced by simple protein restriction, growth inhibition in our previous study was more severe. If the benefit observed in this study is simply due to protein restriction, we would expect to observe a lesser benefit proportional to the lesser degree of growth inhibition, whereas the reverse is true. The cystic change seen in the soy-fed animals is actually less on a milliliter per kilogram basis than seen in weanling animals in our colony (2.21 ± 0.32 ml/kg). It is difficult to determine, however, whether this represents regression or just a relative increase in size of other renal elements, with the hypertrophy that is a normal component of renal growth.

The mechanism by which different diets may be beneficial is unknown. Eddy (17) noted that dietary protein restriction produced improvement in the interstitial nephritis and fibrosis that follows puromycin aminonucleoside nephrosis. In those studies, this improvement was associated with decreased expression of transforming growth factor- (TGF-). In our study of protein restriction, TGF- expression was not significantly different between dietary groups but tended to be higher in the low protein group. The chemokine, monocyte chemoattractant protein-1, which is an important early signal in the initiation of inflammatory events, was significantly reduced by dietary protein restriction in that study (6). This protein may be expressed by a number of renal cell types and is downregulated by estrogenic compounds. Calorie restriction has been shown to influence pathology, independent of protein content in some models (31). The reduction in inflammation and fibrosis that was seen in our animals and other models of renal injury may be a major factor in the observed amelioration of PKD. Alteration of the interstitial collagen matrix is a potent modulator of tubular proliferation (50).

Speculation on the mechanism of benefit is limited by the fact that most soy protein-based diets, including that used in this study, do not use a pure protein. The products that are commonly used may vary in actual protein nitrogen content and are likely to contain a range of biologically active compounds (2), some of which are felt to contribute to other observed benefits of soy diets, including antineoplastic effects. Unheated soy protein may contain trypsin inhibitors that inhibit protein digestion, effectively limiting available protein to less than the stated dietary content. Soy protein feeding is known to enhance the conversion of polyunsaturated fatty acids to docosahexaenoic acid (45). Increased production of this complex lipid has been linked to benefits in a variety of inflammatory models and disorders, including renal disease (12). Soy protein diets are hypocholesterolemic, although this may in large part be due to increased stool losses of cholesterol as much as effects on synthesis (34). Gile et al. (21) have proposed that inhibition of cholesterol synthesis downregulates pathways critical to tubular proliferation in Han:SPRD-cy rat PKD, possibly through depletion of farnesyl pyrophosphate, a cofactor in G protein-dependent signaling. Reduction of serum cholesterol has been beneficial in other forms of renal injury (15). Eddy (18) has demonstrated that diet-induced hypercholesterolemia, combined with the modest remodeling stimulus of uninephrectomy, is a potent inducer of tubulointerstitial nephritis. Demonstration of the converse, i.e., lowering cholesterol will protect against interstitial nephritis, has not yet been described in that model. Soy protein concentrates, unless extracted with alcohol, are rich sources of genistein and other phytoestrogens (4). Genistein is a potent inhibitor of tyrosine kinase, the activity of which is critical to peptide growth factor-induced tubular hyperplasia in vitro (41). Soy protein diets are felt to be weakly estrogenic (4). Both the quantitative and qualitative changes that we have seen in this study are reminiscent of the differences seen between male and female Han:SPRD-cy rats and of the changes induced by castration or estrogen therapy (23). It is possible that the magnitude of the observed changes in pathology is a result of additive effects of a number of these dietary components.

Our finding of reduced ammonium excretion in soy-fed rats supports the concept that this diet had a net alkali effect compared with the casein diet, although impaired renal response to acidosis in the face of systemic acidosis would give similar urinary findings (10). Torres et al. (46) have shown amelioration of renal injury in male Han:SPRD-cy rats supplemented with alkali and acceleration of injury in female Han:SPRD-cy rats fed ammonium chloride. They speculate that interactions between complement and ammonia produced in response to an acid load, and possibly direct effects of ammonia on epithelial proliferation and matrix degradation are possible mechanisms for this observed benefit.

In a previous study (38), we demonstrated increased excretion of organic anions of the citric acid cycle in affected animals compared with normal when the animals were fed a standard chow diet. Bell et al. (8) have previously shown that feeding of a synthetic casein-based diet reduced excretion of these compounds in normal rats, which, they speculate, might reflect an effect on intracellular pH. They further demonstrated that excretion of some of these compounds decline with age. The concept of "dedifferentiation" that is popular among theories of PKD pathogenesis would thus fit well with the observations of our previous study. The less striking changes between normal and affected animals on these diets may also be a direct dietary effect on the excretion of the compounds of interest. The increased urinary losses of anions relevant to the citric acid cycle seen in soy-fed animals does not, however, correlate with expression of injury in this model, as has been reported in other types of renal injury (3, 11, 35). It is possible, however, that the successful retention of tissue succinate may be a marker of reduced tissue injury.

Our ¹H-NMR tissue extract studies do not strongly suggest a direct metabolic effect of soy protein on renal tubular injury. Our previous study reported larger differences in tissue content of organic anions and osmolytes based on wet weight determination (38). When corrected for the difference in water content, most of these differences diminish, but worse disease remained associated with significantly decreased tissue content of succinate and betaine when the comparisons are made between the two treatment groups. There was a trend to lower betaine and succinate in affected casein-fed animals, compared with normal casein-fed animals, as was observed in our previous study, but this did not quite reach significance, probably as a result of both small numbers and the correction to dry weight. Allantoin was the only compound identified the content of which varied with extent of disease alone. The studies indicated significant dietary effects on the entire profile of osmotically active molecules, although the sum of the principal osmolytes is similar in both dietary groups. It is therefore unlikely that amelioration of a disorder of cell volume is responsible for the observed clinical benefit. The marked accumulation of betaine in soy-fed animals may be a secondary response to reduced availability of taurine (9). The lower choline peak may reflect dietary effects or possibly enhanced betaine synthesis from choline, although the significance of this process within the kidney seems to be limited to the inner medulla (24). The function of betaine as a methyl donor has been protective against alcohol-induced liver injury (7), but no such metabolic role has yet been identified within the kidney.

Increased reliance on animal-based protein is a feature of increasing affluence in society. Our study provides further evidence that such dietary shifts may influence renal pathophysiology. The complex composition of soy protein preparations mandates future monocomponent dietary studies to identify active components for practical intervention in human disease.