The mechanism of insulin action is still incompletely understood and the subject of active investigation. After insulin binds to its cell surface receptor, many metabolic pathways are activated. Interest in these secondary signalling systems has resulted in the isolation of certain inositol phosphoglycan molecules which mimic some actions of insulin, especially those related to glucose oxidation and lipid metabolism(Reference Romero, Luttrell and Rogol1–Reference Kunjara, McLean and Greenbaum5). Interestingly, the inositol present in some of these putative insulin mediator preparations was predominantly or solely d-chiro-inositol (DCI) (Fig. 1(a)), an epimer of myo-inositol (MI) (Fig. 1(c)), which is the predominant mammalian inositol(Reference Mato, Kelly and Abler6, Reference Larner, Huang and Schwartz7). DCI differs from MI only in the orientation of a single hydroxyl group (labelled 6 in both DCI and MI). It is this hydroxyl group by which MI is attached to phosphatidic acid in phosphatidylinositol. The prefix ‘chiro’ refers to the fact that this inositol has optically active d and l enantiomers. chiro-Inositol also has been identified in preparations derived from the phosphatidylinositol anchors of membrane proteins(Reference Futerman, Low and Ackermann8–Reference Walter, Roberts and Rosenberry10). Both the chemical structure and the occurrence of DCI in insulin signalling molecules suggest that it may have a physiological role in mammals. Pinitol (1d-3-O-methyl-chiro-inositol) is structurally related to DCI, having a methoxyl group at position 3 (Fig. 1(b)). Pinitol was originally described in 1855 as a component of the sugar pine tree(Reference Berthelot11). It is also a prominent component of animal forage legumes(Reference Smith and Phillips12) and is found in beans including soyabeans(Reference Phillips, Dougherty and Smith13, Reference Streeter14).
Fig. 1 Structures of inositols: d-chiro-inositol (DCI) (a); pinitol (b); myo-inositol (c). DCI is numbered as recommended by the International Union of Pure and Applied Chemistry(34). Pinitol and myo-inositol are numbered according to the convention for DCI.
Diabetic patients (both type 1 and type 2) have increased urinary excretion of DCI, the magnitude of which correlates with plasma glucose and glycated Hb levels(Reference Ostlund, McGill and Herskowitz15). Pima Indians with type 2 diabetes have low levels of muscle chiro-inositol(Reference Kennington, Hill and Craig16). A similar abnormality in chiro-inositol excretion is found in diabetic db/db mice and streptozotocin diabetic rats(Reference Kawa, Przybylski and Taylor17). Although not all clinical trials have been positive(Reference Davis, Christiansen and Horowitz18), at least some insulin-resistant patients appear to have responded to treatment with DCI or pinitol with improved insulin sensitivity(Reference Iuorno, Jakubowicz and Baillargeon19, Reference Kim, Yoo and Kim20).
Studies of animal and human treatment with pinitol or DCI have been limited by lack of experimental control of dietary intake and incomplete knowledge of DCI metabolism. MI, a component of mixed diets(Reference Holub21), has been shown to be actively absorbed in the small intestine of hamsters(Reference Caspary and Crane22) and also is synthesised from glucose(Reference Daughaday, Larner and Hartnett23–Reference Hauser and Finelli25). However, there is no information about DCI absorption, and previous work on the possible bioconversion of dci from MI has not given consistent results. Inositol epimerase from bovine brain converted MI to neo- and scyllo-inositol but not to chiro-inositol(Reference Hipps, Ackermann and Sherman26). In cockroach fat body extracts, however, epimerisation of MI to chiro-inositol was observed(Reference Hipps, Sehgal and Holland27). Injection of 3H-labelled MI into rats showed conversion of MI to chiro-inositol as judged by counts co-migrating with chiro-inositol(Reference Pak, Huang and Lilley28).
In the present study we analysed dietary inositol contents of commercial animal diets and studied the absorption of pinitol and DCI and biosynthesis of DCI in rodents maintained on chemically defined diets.
Materials and methods
Materials
Diisopropylidene pinitol and l-chiro-inositol were gifts of Professor Laurens Anderson of the University of Wisconsin, Madison. Pinitol containing less than 1 % DCI was prepared from its isopropylidene derivative by hydrolysis in 0·1 m-HCl for 1 h at 110°C, followed by lyophilisation and recrystallisation from 90 % ethanol(Reference Anderson, MacDonald and Fischer29). DCI was purchased from Calbiochem. dl-1,2,3,4,5,6-[2H6]chiro-inositol(Reference Sasaki, Balza and Taylor30) was the gift of Dr Ken Sasaki (Connaught Center for Biotechnology Research, Toronto, ON, Canada). 1,2,3,4,5,6-[2H6]MI was purchased from MSD Isotopes (Montreal, QC, Canada). [2H3]pinitol labelled in the methyl radical was a gift of Dr Andrew Falshaw (Industrial Research Ltd, Lower Hutt, New Zealand). Ethanol-extracted casein (C3400) was purchased from Sigma (St Louis, MO, USA).
Diets
Purina Laboratory Rodent Diet 5001 (diet 5001) and Purina PicoLab Rodent Diet 20 5053 (diet 5053) were purchased from LabDiet (Richmond, IN, USA). Basal Diet 5755 (diet 5755; a chemically defined diet) was purchased from TestDiet (Richmond, IN, USA). Components of diet 5001 were kindly supplied by Dr Daniel Hopkins. The pinitol/DCI-deficient diet (Table 1) was based on diet 5755. The ingredients for both diets were dextrin, casein, sucrose, mineral mix, vitamin mix, cellulose, choline chloride, dl-methionine and fat. Maize oil (5 %) and lard (5 %) were used in diet 5755 whereas only soyabean oil (5 %) was used in our pinitol/DCI-deficient diet. Components of diet 5755 were analysed for total DCI and levels were generally very low. The original powdered cellulose contained 0·014 nmol total DCI/mg and was replaced by Avicel® cellulose (FMC Biopolymer, Philadelphia, PA, USA), which contained non-detectable total DCI.
Table 1 Pinitol/d-chiro-inositol-deficient diet formulation
ppm, Parts per million.
* Richmond, IN, USA.
† St Louis, MO, USA.
‡ Head Office: Bentonville, AR, USA.
§ Philadelphia, PA, USA.
∥ Final concentration of mineral mix components in the diet, as provided by the manufacturer: Ca, 0·60 %; P, 0·60 %; K, 0·40 %; Mg, 0·07 %; Na, 0·21 %; Cl, 0·24 %; F, 5·0 ppm; Fe, 63·0 ppm; Zn, 21·0 ppm; Mn, 65·0 ppm; Cu, 15·0 ppm; Co, 3·2 ppm; I, 0·57 ppm; Cr, 3·0 ppm; Mo, 0·82 ppm; Se, 0·23 ppm.
¶ Final concentration of vitamin mix in the diet, as provided by the manufacturer: vitamin A, 22·1 IU/g (12·1 μg/g); vitamin D3, 2·2 IU/g (0·055 μg/g); vitamin E, 50·1 IU/kg (50·1 mg/kg); vitamin K, as menadione, 10·4 ppm; thiamin hydrochloride, 20·6 ppm; riboflavin, 20·0 ppm; niacin, 90·0 ppm; pantothenic acid, 55·0 ppm; folic acid, 4·0 ppm; pyridoxine, 16·5 ppm; biotin, 0·4 ppm; vitamin B12, 20·0 μg/kg; choline chloride, 1400 ppm; ascorbic acid, 0·0 ppm.
Gas chromatography–mass spectrometry
For inositol analysis in plasma and urine, 2H-labelled internal standards for DCI, MI and pinitol were added and they were then processed by protein precipitation and anion exchange adsorption and dried as described previously(Reference Ostlund, McGill and Herskowitz15). Urine samples were further purified by solid-phase extraction. They were taken up in 0·5 ml water, applied to a 1·0 ml Supelclean™ LC-18 SPE column (Supelco, Bellefonte, PA, USA) previously equilibrated with 2 ml methanol, eluted with 2 ml water and lyophilised.
Both free and total inositols in feed are reported. To isolate free inositols, dried mouse stool homogenate or 50 mg samples of finely ground mouse chow received all three 2H-labelled internal standards and were extracted with 0·5 ml water at room temperature for 12 h. For total inositols, samples of ground mouse chow or dried stool homogenate received 0·5 ml 6 m-HCl with 2H-labelled internal standards for DCI and MI and were heated at 110°C for 24 h in tightly sealed tubes. Total DCI contents measured by acid hydrolysis include free and complex DCI such as inositol glycans or phosphates and free and complex pinitol, since acid hydrolysis converts pinitol to DCI and hydrolyses complex inositols to free inositols. The acid hydrolysate was separated from charred residue, dried under N2 at 85°C and taken up in 0·5 ml water. Samples were treated with anion exchange resins, passed over 6 ml LC-18 solid phase extraction columns (Supelco), and dried.
Lyophilised samples were derivatised for GC–MS by incubation overnight with 10 % pentafluoropropionyl imidazole in acetonitrile at 65°C and then separated on a 25 m × 0·25 mm internal diameter Chirasil-Val fused silica capillary column with 0·16 μm film thickness (Alltech Associates, Deerfield, IL, USA). The oven temperature was kept at 80°C for 0·5 min, then raised at a rate of 60°C/min to 100°C and held for 4 min, and again raised at a rate of 20°C/min to a final temperature of 185°C and held for 3·5 min. The effluent was analysed in an Agilent Technologies 5973 quadrupole mass spectrometer by negative-ion chemical ionisation MS using methane as the reagent gas (Hewlett Packard, Palo Alto, CA, USA). Selected ion monitoring was performed at m/z 573 and 576 (for natural and 2H-labelled pinitols, respectively) and at m/z 726 and 731 (for natural and 2H-labelled chiro-inositols and MI).
The enrichment of 2H-labelled water in urine was determined following alkaline exchange with acetone according to the published protocol, except that hexane rather than chloroform was used in the exchange procedure(Reference McCabe, Bederman and Croniger31). Exchanged acetone was separated isothermally at 60°C for 5 min on a 30 m × 0·25 mm internal diameter DB17-MS capillary column with 0·25 μm film thickness (J&W Scientific, Folsom, CA, USA). Selected ion monitoring was performed at m/z 58 and 59 using electron impact ionisation mode (70 eV). The incorporation of 2H from 2H-labelled water into MI was calculated by mass isotope distribution analysis considering both 2H and natural 13C enrichment.
Metabolic balance studies in rats fed Purina Rodent Diet 5001 or Basal Diet 5755
For metabolic balance studies, male Sprague–Dawley rats weighing 400–500 g were fed diet 5001 for at least 1 month, then housed individually in metabolism cages and fed powdered diets consisting of either diet 5001 or diet 5755 for 1 week. Dietary consumption was measured and urine and faeces were collected for 24 h. The total amount of DCI after HCl hydrolysis was measured by negative-ion chemical ionisation GC–MS, and DCI balance was computed as the difference between intake and output. Institutional and national guidelines for the care and use of animals were followed. All procedures involving animals were approved by the Washington University Animal Studies Committee.
Effect of stool bacteria on faecal total d-chiro-inositol levels
Equal amounts of stools from three rats fed diet 5001 were pooled and dispersed in 0·15 m-NaCl to a concentration of 0·2 g/ml and MI and glucose were added to final concentrations of 30 μm and 5·5 mm, respectively. Samples were incubated for 24 h at either 4°C (control for the measurements at 37°C) or 37°C in anaerobic jars, after which internal standards were added and the samples were hydrolysed and processed by negative-ion chemical ionisation GC–MS for total DCI content.
Absorption studies in rats fed Basal Diet 5755
Three rats weighing 427 (sem 27) g were housed individually in metabolism cages and fed diet 5755 for 3 d. Pinitol, 51·5 μmol (10 mg), was then given orally; urine and stools were collected for the following 2 d and analysed for total DCI. Then 25 nmol [2H6]DCI was given orally followed by an additional 2 d of urine and stool collection.
Inositol clearance over time in mice fed Purina PicoLab® Rodent Diet 20 5053 or the pinitol/d-chiro-inositol-deficient diet
Male C57BL/6J mice were fed diet 5053 for 1 week before randomly being assigned to diet 5053 (n 8) or the pinitol/DCI-deficient diet (n 8) for up to 5 weeks. Food intake was recorded twice per week and body weight once per week. Plasma, stools and urine were obtained before the experimental assignment (week 0), and at 1, 2 and 5 weeks after the diet assignment. For each time point, urine and stools were collected over a 24 h period in individual metabolism cages. Pinitol, DCI and MI levels in plasma, stools or urine were determined by GC–MS.
Measurement of inositol biosynthesis in mice fed the pinitol/d-chiro-inositol-deficient diet
To measure endogenous inositol (DCI and MI) synthesis, male C57BL/6J mice (n 3) were maintained on the pinitol/DCI-deficient diet for 10 weeks. Then, mice were provided with drinking water containing 20 % heavy water (2H-labelled water, catalogue number 151882; Sigma-Aldrich) for 1 week to label the ring hydrogens of inositols, followed by 24 h urine collection in individual metabolism cages.
At 5 weeks later, [2H6]MI (1 mg in water) was administered by intraperitoneal injection to each mouse (n 3, maintained on the pinitol/DCI-deficient diet for 15 weeks before intraperitoneal injection) to determine whether MI is converted into DCI in vivo. Urine was collected for 24 h each before and after injection. Urine samples were processed and GC–MS was performed as described above.
Statistical analyses
Group data are reported as mean values with their standard errors. Two-way repeated-measures ANOVA was used to analyse diet, time and interactions between diet and time with SAS Proc GLM (version 9.2; SAS Institute, Cary, NC, USA). Multiple comparisons were performed using the Tukey adjustment.
Results
Total acid-released d-chiro-inositol content of Purina rodent chows
Total DCI measured by GC–MS after hydrolysis with 6 m-HCl at 110°C for 24 h was a prominent component of laboratory animal chows (Table 2). Chows 5001 and 5053 each contained more than 4 nmol total DCI/mg. Diet 5001 contained 12·9 (sem 1·2) nmol total DCI/mg or 0·23 % by weight. DCI was present in diet 5001 in an amount approximately two-thirds that of MI, which was 20·2 (sem 0·7) nmol/mg or 0·36 % by weight following hydrolysis. Another rodent chow, chow 5053, currently used in our animal facility, contained comparable levels of total DCI at 10·94 (sem 0·45) nmol/mg (Table 2). Diet 5755 had only trace levels of total DCI and these levels were reduced even further in the pinitol/DCI-deficient diet.
Table 2 Nutrient composition of the diets*
Diet 5001, Purina Laboratory Rodent Diet 5001; diet 5053, Purina PicoLab® Rodent Diet 20 5053; diet 5755, TestDiet® 5755 Basal Diet; DCI, d-chiro-inositol; PFV, physiological fuel energy; NA, not analysed; ND, not detectable; MI, myo-inositol.
* The mouse chows and the pinitol/DCI-deficient diet (n 3, about 50 mg each) were weighed. Free (for pinitol, DCI and MI) and total (for DCI and MI) inositols were determined as described in Materials and methods. The rest of nutrient composition was from the manufacturers.
† Purchased from LabDiet (Richmond, IN, USA).
‡ Purchased from TestDiet (Richmond, IN, USA).
§ Sum of decimal fractions of protein, fat and carbohydrates multiplied by 16·8, 37·7 and 16·8 kJ/g, respectively.
∥ The limit of detection for DCI is 0·0003 nmol/mg.
Analyses of chow components showed that maize, wheat, meat, fish and dairy products contributed relatively little DCI (Table 3). However, total DCI was very high in the leguminous components of lucerne and soyabean meal in which it constituted 1·3 and 1·0 % of weight, respectively, or more than fifty times as much as in other chow components. l-chiro-Inositol, in contrast, was very low in all ingredients.
Table 3 Total chiro-inositol content of components of Purina Rodent Diet 5001
* Values for lucerne and soya for six replicates.
† l-chiro-Inositol could not be measured accurately in the presence of the large amount of d-chiro-inositol but it constituted less than 2 % of d-chiro-inositol.
The molecular form of DCI in diet 5001 was further investigated. Finely ground chow was suspended at room temperature in 1 m-HCl, then extracted and partitioned by the method of Bligh & Dyer(Reference Bligh and Dyer32) to yield aqueous and organic fractions. Less than 1 % of the total DCI was found in the organic fraction; in contrast, 83 % was recovered in the aqueous fraction with the remainder in the insoluble residue. Thus, substantial amounts of DCI did not appear to be present in phospholipids and DCI was present predominantly in the water-soluble form. Only 6 % of the total DCI was measurable without prior strong acid hydrolysis. The species containing DCI did not appear to be ionic (such as a phytate) because there was no change in the measured concentration when the sample was treated with an AG 501-X8(D) mixed bed ion exchange resin.
Metabolic balance studies in rats fed Purina Rodent Diet 5001 or Basal Diet 5755
Rats chronically fed diet 5001 consumed 921 μmol total DCI/kg body weight per d or 0·17 g total DCI/kg body weight per d (Table 4). Only 0·7 % of total DCI intake was found in the stool output, suggesting that gastrointestinal absorption was essentially complete. Urinary output was only 4·6 % of that consumed. The total excretion from the urine and stools represents 5·3 % of the total DCI intake, suggesting that the majority of the dietary inositols was metabolised. Thus, total DCI balance was strongly positive at 872·5 (sem 79·0) μmol/kg body weight per d.
Table 4 Metabolic balance of total d-chiro-inositol (DCI) in rats fed Purina Rodent Diet 5001*
(Mean values with their standard errors)
* Rats (n 3) were fed Purina Rodent Diet 5001 for at least 1 month and housed in individual metabolism cages for 1 week. Dietary consumption was measured and urine and faeces were collected for 24 h. The total amount of dci after HCl hydrolysis was measured by negative-ion chemical ionisation GC–MS, and dci balance was computed as the difference between intake and output.
Effect of stool bacteria on faecal total d-chiro-inositol levels
Since bacterial degradation or production of DCI in the gastrointestinal tract could have influenced these experiments, rat stool samples were fortified with MI and glucose (possible metabolic precursors of DCI) and then incubated at 37°C anaerobically for 24 h. A large amount of gas was produced indicating active bacterial metabolism, but little effect on measured total DCI levels was seen. There was a 9·5 % decrease in total DCI after 37°C incubation (10·54 (sem 0·27) nmol/ml at 4°C v. 9·85 (sem 0·19) nmol/ml at 37°C; P = 0·052), suggesting that a small amount of degradation might have occurred. Stool l-chiro-inositol was unchanged by incubation (1·11 (sem 0·36) nmol/ml at 4°C v. 1·32 (sem 0·11) nmol/ml at 37°C). Bacterial degradation or metabolism, therefore, could not account for the quantitative reduction of total DCI in stools compared with the diet.
Absorption studies in rats fed Basal Diet 5755
The gastrointestinal absorption of purified pinitol was studied next (Table 5). At 3 d before the experiment, rats were switched from diet 5001 to diet 5755 low in total DCI in order to reduce excretion of total DCI to approximately 0·5 μmol/kg body weight per d. Then 10 mg pinitol (containing 51·5 μmol total DCI) was given orally and the faecal and urinary excretion of total DCI was measured. Most of the material eliminated was excreted on day 1 when faecal loss was 0·9 % of intake and urinary loss was 21 %. These data suggest that pinitol is absorbed from the gastrointestinal tract and most is retained or metabolised. That somewhat more total DCI appeared in urine (21 %) compared with diet 5001 feeding (4·6 %; Table 4) may be due to the large single pinitol bolus administered. The day after pinitol feeding, total excretion in urine and faeces was similar to baseline.
Table 5 Gastrointestinal absorption of pinitol and [2H6]d-chiro-inositol (DCI) in rats fed Basal Diet 5755*
(Mean values with their standard errors)
ND, not detectable.
* Three rats weighing 427 ± 27 g were housed individually in metabolism cages and fed Purina Basal Diet 5755 for 3 d. Pinitol, 51·5 μmol (10 mg), was given orally with urine and stools collected for the following 2 d, which were analysed for total DCI. Then 25 nmol [2H6]DCI was given orally, followed by an additional 2 d of urine and stool collection. No [2H6]DCI was detectable before isotope administration.
Similar results were obtained when labelled free DCI tracer was given orally (Table 5). Absorption, estimated by recovery of [2H6]DCI in the stools for 2 d, was 98 %. The urinary output was only 14 % of that administered, consistent with incomplete urinary excretion.
Inositol clearance over time in mice fed Purina PicoLab® Rodent Diet 20 5053 or the pinitol/d-chiro-inositol-deficient diet
After a more defined diet was developed, we studied the clearance of inositols over a longer period of time. There were time and diet effects and interactions between time and diet effects on the food intake (Fig. 2(a)) and body weight (Fig. 2(b)). Food intakes (Fig. 2(a)) and body weights (Fig. 2(b)) were statistically different at 1, 2, 3, 4 and 5 weeks between the two diets.
Fig. 2 Food intakes (a) and body weights (b) of mice fed Purina PicoLab® Rodent Diet 20 5053 (diet 5053) (●) or a pinitol/d-chiro-inositol (DCI)-deficient diet (○). Male C57BL/6J mice (eight mice per diet) were fed diet 5053 or the pinitol/DCI-deficient diet for 5 weeks. Food intake was measured twice per week and the average represents weekly food intake (g/d). Body weight was determined at the beginning and once per week. Values are means, with standard errors represented by vertical bars. ** Mean value was significantly different from that of the mice fed diet 5053 (P < 0·001). Effects of diet (P = 0·0002), time (P < 0·0001) and interactions between time and diet (P < 0·0001) were statistically significant.
Both pinitol (Fig. 3(a), (c) and (e)) and DCI (Fig. 3(b), (d) and (f)) in plasma, stools or urine remained relatively stable over the 5-week period for the mice fed diet 5053. In contrast, both pinitol and DCI in these various compartments declined rapidly to a very low level over the 5-week period in mice fed the pinitol/DCI-deficient diet. Repeated-measures ANOVA showed that there was a diet effect on pinitol and DCI concentrations in plasma, stools or urine (Fig. 3). This indicates that the diet plays an important role in pinitol or DCI levels. A time effect was observed for all measurements except for stool pinitol, stool DCI or urine DCI. A time and diet interaction was seen for all parameters except for stool DCI. The differences at weeks 1, 2 and 5 between diet 5053 and the pinitol/DCI-deficient diet in all inositol measurements were statistically significant (P < 0·001).
Fig. 3 Plasma, stool and urine inositols over time in mice fed Purina PicoLab® Rodent Diet 20 5053 (diet 5053) (●) or a pinitol/d-chiro-inositol (DCI)-deficient diet (○): plasma pinitol (a); plasma DCI (b); stool free pinitol (c); stool free DCI (d); urine pinitol (e); urine DCI (f). Male C57BL/6J mice (eight mice per diet) were fed diet 5053 or the pinitol/DCI-deficient diet for 5 weeks. Blood, urine and stools were obtained at 0, 1, 2 and 5 weeks. Pinitol and DCI were analysed as described in Materials and methods. Values are means, with standard errors represented by vertical bars. ** Mean value was significantly different from that of the mice fed diet 5053 (P < 0·001).
Endogenous d-chiro-inositol/myo-inositol biosynthesis from 2H-labelled water in mice fed the pinitol/d-chiro-inositol-deficient diet
To determine whether the carbon ring hydrogen atoms of DCI were synthesised endogenously from water, male C57BL/6J mice (n 3) were fed the pinitol/DCI-deficient diet for 10 weeks and received drinking water containing 20 % (w/w) 2H-labelled water for 1 week followed by a 24 h urine collection while housed in metabolism cages. The resulting concentration of 2H-labelled water in urine was 8·70 (sem 0·52) % as determined by equilibration with acetone followed by GC–MS. Neither natural nor 2H-labelled DCI could be detected in the urine ( < 0·0067 nmol/ml). In contrast, 2H-labelled MI was present in the urine and it was calculated that 34·4 (sem 2·8) % of urinary MI was derived by biosynthesis from body water.
Conversion of myo-inositol to d-chiro-inositol in mice fed the pinitol/d-chiro-inositol-deficient diet
Male C57BL/6J mice (n 3) were maintained on the chemically defined pinitol/DCI-deficient diet for 15 weeks and then given 1·0 mg [2H6]MI by intraperitoneal injection. Following a 24 h isotope equilibration period, urine was collected in individual metabolism cages. [2H6]MI constituted 38·7 (sem 11·5) % of urinary MI after labelling, which indicated a large enrichment of the endogenous MI pool. Ionic masses corresponding to natural DCI (m/z 726) through to fully labelled DCI (m/z 731) were scanned, but no natural DCI or DCI with any amount of 2H labelling was detected ( < 0·0067 nmol/ml).
Discussion
Total DCI levels were unexpectedly high in common laboratory animal chows and constituted 0·23 % of the weight of diet 5001 (Table 2). The amounts consumed by animals were very large; for example, rats chronically ingested 166 mg total DCI/kg body weight per d while on diet 5001 (Table 4). Thus, a potential source of DCI is abundant in laboratory animal diets, principally in the legumes lucerne and soyabeans, which are ingredients of many commercial laboratory animal chows. However, meat and other animal products contained little total DCI. l-chiro-Inositol was present only in small amounts relative to DCI (Table 3). In lucerne and soyabeans, at least 98 % of the total chiro-inositol was the d-enantiomer. In diet 5053, only 1·5 % of the DCI was present in the free form, whereas 65 % was found in free pinitol. Another 33·5 % of DCI is in complex forms such as glycosides liberated by acid hydrolysis. In contrast to common rodent chows, diet 5755 and our pinitol/DCI-deficient diet had only trace amounts of total DCI (Table 2).
In the equilibrium state on diet 5001, there was a markedly positive total DCI balance of 872·5 μmol/kg body weight per d (Table 4). DCI must have been metabolised, since only about 5·3 % was recovered in the stools (0·7 %) and urine (4·6 %). The complete absorption (99·3 %) of total DCI suggests that both free and complex forms of pinitol or DCI were absorbed from the diet. Their complete absorption from the gastrointestinal tract was further confirmed by using purified pinitol and [2H6]DCI, since less than 2 % of the orally administered material was recovered in the stools (Table 5). More urinary output was observed on day 1 with feeding the natural pinitol than with the oral administration of [2H6]DCI, suggesting that their metabolic kinetics differed.
It has been reported previously that soil bacteria transform labelled MI into chiro-inositol(Reference Sasaki, Balza and Taylor30). Although the extent of conversion was small (4 % over 12 d), the possibility of bacterial synthesis or degradation of chiro-inositol in the large bowel was considered in our experiments under anaerobic incubation mimicking colonic fermentation. However, no net synthesis of DCI was found and only a small amount of DCI was degraded. Faecal bacteria do not appear, therefore, to be an important source of DCI in these animals.
Since pinitol is of plant origin(Reference Vernon and Bohnert33), it is expected to come exclusively from the diet. Thus pinitol in the plasma, stools or urine fell rapidly to a very low level over the 5-week period after switching from diet 5053 to the pinitol/DCI-deficient diet (Fig. 3(a), (c) and (e)). By contrast, pinitol levels in the same three compartments remained relatively stable when mice were fed diet 5053 (Fig. 3(a), (c) and (e)). Similarly, DCI fell rapidly to an extremely low level in the plasma, stools or urine of mice fed the pinitol/DCI-deficient diet, in contrast to relatively stable levels of these inositols in mice fed diet 5053 (Fig. 3(b), (d) and (f)). In fact, plasma, stools or urinary levels of pinitol or DCI were not detectable in mice fed the pinitol/DCI-deficient diet for 10 weeks. These results strongly suggest that DCI in body fluids derives from the diet and not from endogenous biosynthesis. Indeed, 2H-labelled water was not incorporated into 2H-labelled DCI, indicating no endogenous synthesis of DCI from body water hydrogen. In contrast, 34·4 % of urinary MI was synthesised from 2H-labelled water, consistent with endogenous synthesis of MI in the literature(Reference Daughaday, Larner and Hartnett23).
Bioconversion of MI to DCI appears not to occur in mice. 2H-labelled DCI was not detected in the urine of mice 24 h after intraperitoneal injection of [2H6]MI. The injected [2H6]MI constituted 38·7 (sem 11·5) % of urinary MI, indicating good enrichment of the putative precursor. Inconsistent results have been obtained in the past regarding conversion of MI to DCI. Inositol epimerase from bovine brain led to conversion of MI to neo- and scyllo-inositol but not chiro-inositol(Reference Hipps, Ackermann and Sherman26). In cockroach fat body extracts, however, epimerisation of MI to chiro-inositol was observed(Reference Hipps, Sehgal and Holland27). More recently, Pak et al. injected 37 MBq of [3H]MI intraperitoneally into rats over 3 d and found that it was converted to radioactive counts that co-migrated with chiro-inositol(Reference Pak, Huang and Lilley28). The percentage conversion was very large, 8 % in blood, 9 % in liver and 36 % in urine. However, DCI natural abundance of total inositols (sum of chiro-inositol and MI) is only 0·4 % in plasma and 2·3 % of urinary inositols in humans(Reference Ostlund, McGill and Herskowitz15). Furthermore, the ratio of basal (when all mice were on diet 5053) urinary DCI to urinary MI was barely 10·0 % in the present study. This raises the possibility that some of the counts co-migrating with chiro-inositol on paper chromatograms may not have been chiro-inositol. Furthermore, the method used did not distinguish d- from l-chiro-inositol(Reference Pak, Huang and Lilley28).
Our time-course study in mice comparing diet 5053 and the pinitol/DCI-deficient diet showed little change in body weight in either group over 5 weeks. However, mice in the pinitol/DCI-deficient diet weighed less (Fig. 2(b)) and had less food intake (Fig. 2(a)) when compared with those on diet 5053. We speculate that this could be due to differences in MI, DCI or other nutrients, or to reduced palatability of the pinitol/DCI-deficient diet. This experiment was performed to test whether or not DCI was synthesised endogenously rather than testing the function of DCI in the diet, which requires further research.
In conclusion, pinitol and DCI are present in large amounts in rodent chow diets and they are avidly absorbed from chow 5001. DCI was not synthesised in vivo or converted from [2H6]MI, suggesting that, like pinitol, it derives solely from the diet. The present study suggests that dietary control of inositols should be considered for future research in diabetes and insulin resistance. The chemically defined diet 5755 and the pinitol/DCI-deficient diet described here establish a deficient baseline for future control studies, and are suitable for the determination of the effects of DCI and pinitol on glucose and lipid metabolism.
Acknowledgements
The present study was supported by a National Institutes of Health (NIH) grant (R01 DK58698) and the Washington University Mass Spectrometry Resource (RR00954), Diabetes Research and Training Center (DK20579) and Clinical Nutrition Research Center (DK56341).
The contribution of each author was: R. E. O., design and writing; X. L., design, performance of experiments, statistical analysis of data, and writing; L. M., performance of experiments, GC–MS analysis of samples; C. G., performance of experiments and writing.
The authors declare no conflicts of interest.
