The enriched Ca isotope (⁴⁴Ca) as CaCO3 and the enriched Mg isotope (²⁵Mg) as MgO were obtained from Chemgas, (Boulogne, France). The atomic abundances of these enriched isotopes were as follows: ⁴⁰Ca = 3.41%, ⁴²Ca = 0.09%, ⁴³Ca = 0.03%, ⁴⁴Ca = 96.45% ⁴⁶Ca =< 0.01% ⁴⁸Ca = 0.02% and ²⁴Mg = 1.6%, ²⁵Mg = 97.8%, ²⁶Mg = 0.6%. HNO₃ (ultrapure), Mg and beryllium standard solutions (1 g/L) were obtained from Merck (Darmstadt, Germany). All other chemicals were of the highest quality available. Distilled water was used throughout. A Perkin-Elmer 6100DRC system (Perkin-Elmer Instruments, Courteboeuf, France) equipped with a Meinhard nebulizer was used for isotopic measurement, and a Perkin Elmer AA800 (Perkin Elmer Instruments, Courteboeuf, France) was used for total Mg measurement.

Eighty male Wistar rats aged 2, 5, 10 or 20 months were purchased from Janvier (Le Genest Saint Ile, France). They were fed a commercial pellet diet (Ssniff R/S-breeding – until 3 mo, then Ssniff R/S maintenance from 3 to 24 mo age). Two groups were formed for each age bracket to receive either a control diet or a semi-purified diet containing inulin until the end of the experiment. The composition of these two diets is given in Table 1. Tested inulin was purchased from Orafti, Tienen, Belgium (Raftaline^®). The target Ca and Mg levels in these diets were 5000 mg Ca/kg and 500 mg Mg/Kg diet. Powder diet (100 g) was made up with 100 ml of distilled water to form a kind of semi-liquid food prepared on-site each day. Chemical analysis of the diets offered confirmed the expected Ca and Mg contents in the experimental diets: 5107 mg Ca/kg and 5050 mg Ca/kg, and 495 mg Mg/kg and 514 mg Mg/kg in the control and inulin diets, respectively. Chemical analysis showed that the inulin contained approximately 40 mg Ca/kg and less than 1 mg Mg/kg. Dietary inulin level was maintained at 3.75% during the first 4 days and then 7.5% from day 5 until the end of the experiment. The 8 rat groups were given fresh food and water daily, made available ad libitum. Food consumption and body weight were recorded weekly. Throughout the experiment, the rats were housed two per cage (wire-bottomed to limit coprophagy) in a temperature-controlled room (22°C) with dark period from 08:00 pm to 08:00 am. Total experiment duration was 30 days. All procedures complied with the Institute's ethical guidelines on the care and use of laboratory animals.

215 mg of the ⁴⁴Ca (in carbonate form = 508 mg) and 255 mg of the ²⁵Mg (in oxide form = 412 mg) were first individually moistened with 2 ml of distilled water. One ml of HCl 12 N (ultrapure) was added to the ⁴⁴Ca suspension and two ml of HCl 12 N was added to the ²⁵Mg suspension to transform the carbonate and the oxide into soluble chlorides of Ca and Mg, respectively. Each solution was then diluted with 50 ml of distilled water, both solutions were then mixed, and pH was adjusted to between 3 and 6 with 1 N sodium hydroxide solution. The resulting study solution was then completed to 150 ml with distilled water and maintained for several days at +4°C until utilization. Total and isotopic Ca and Mg contents were checked before use.

The rats were transferred to metabolic cages and housed individually three days before the beginning of the isotopic balance study to allow them to adapt to their new environment. Animals received by gavage about 1.7 ml of isotopic solution. The urine and faeces of each rat were quantitatively collected for four consecutive days, and excreted isotopes in these two media and in the gavage solution were quantitatively determined by ICP/MS, as described below.

The rats were sacrificed just after the dark period (between 08:00 am and 10:00 am), i.e. at a time when cecal fermentation was still very active. After anesthesia (40 mg sodium pentobarbital/kg body weight), blood was withdrawn from the abdominal aorta, placed into tubes containing sodium heparin and centrifuged at 1,000 g for 10 minutes. Plasma samples were stored at 4°C for mineral analysis. The cecum, complete with contents, was removed and weighed (total cecal weight). The cecal wall was flushed clean with ice-cold saline, blotted on filter paper, and weighed (cecal wall weight). For each rat, duplicate samples of cecal contents were collected into 2 ml microfuge tubes and immediately frozen at -20°C until analysis. The pH of cecal content was determined on site using a Sentron pH-system 1001 portable pH-meter (Sentron Europe B.V. Ac Roden, The Netherlands). Supernatants of the digestive contents were obtained by centrifuging one of the two microfuge tubes at 20,000 g for 10 minutes at 4°C, and then frozen until analysis. One tibia was also sampled for Ca and Mg analysis.

Ca and Mg concentrations were determined in the plasma and urine after adequate dilution into 0.1% (w/v) lanthanum chloride. Diet aliquots, fecal materials and tibia were dry-ashed (10 hours at 500°C) and dissolved with concentrated HNO₃ and H₂O₂ on a heating plate until complete decoloration. The resulting mineral solutions were set at 10 ml with water and adequately diluted in 0.1% lanthanum chloride. Mineral concentrations were measured by atomic absorption spectrophotometry (on a Perkin-Elmer AA800) at wavelengths of 422 nm for Ca and 285 nm for Mg.

For isotopic ⁴⁴Ca and ²⁵Mg determination, samples were appropriately diluted before analysis using 1% HNO₃. Ca and Mg concentration and isotope ratios were determined by ICP-MS using Ca and Mg as external standard and beryllium as internal standard. The instrument operating conditions were set as follows after optimization with a solution of 1 μg indium/l: RF Power = 1050 W, Nebulizer Ar flow rate = 0.79 L/min, Auxiliary Ar flow rate = 1.2 L/min, Outer Ar flow rate = 15 L/min. Data acquisition parameters were set as follows: Sweeps/reading = 50, Readings/replicate = 1, Number of replicates = 3, Dwell time = 50 ms for ²⁴Mg, 75 ms for ⁹Be, ²⁵Mg, ²⁶Mg, and ⁴⁴Ca, 150 ms for ⁴²Ca and 300 ms for ⁴³Ca, Scanning mode = peak hopping. DRC operating conditions (for ⁴²Ca, ⁴³Ca and ⁴⁴Ca) were as follows: Cell Gas A Flow Rate = 0.5 L ammonia/min, RPa = 0, and RPq = 0.45.

Cecal SCFA concentrations, including acetic, propionic and butyric acid, were determined by gas-liquid chromatography on portions of supernatant fractions of cecal contents as previously described 14.

Ca and Mg each have different stable isotopes with the following natural abundances: ⁴⁰Ca = 96.941%, ⁴²Ca = 0.647%, ⁴³Ca = 0.135%, ⁴⁴Ca = 2.086% ⁴⁶Ca = 0.004% ⁴⁸Ca = 0.187% and ²⁴Mg = 78.99%, ²⁵Mg = 10.00% and ²⁶Mg = 11.01% 15. ⁴⁴Ca and ²⁵Mg isotopic enrichments were obtained, respectively, from the following equations: (⁴⁴Ca/⁴³Ca measured ratio - ⁴⁴Ca/⁴³Ca baseline ratio)/(⁴⁴Ca/⁴³Ca baseline ratio) and (²⁵Mg/²⁶Mg measured ratio - ²⁵Mg/²⁶Mg baseline ratio)/(²⁵Mg/²⁶Mg baseline ratio).

Non-absorbed ⁴⁴Ca and ²⁵Mg isotopes in the fecal or urine samples (coming only from the ⁴⁴Ca or ²⁵Mg isotope labels) were calculated as follows:

for ⁴⁴Ca (mg) = (total fecal or urine Ca (mg) × (natural abundance ⁴⁴Ca × enriched ⁴⁴Ca)/(1 + (natural abundance ⁴⁴Ca × enriched ⁴⁴Ca);

for ²⁵Mg (mg) = (total fecal or urine Mg (mg) × (natural abundance ²⁵Mg × enriched ²⁵Mg)/(1 + (natural abundance ²⁵Mg × enriched ²⁵Mg)).

Calculations were also made directly from ICP-MS data. The two modes of calculation give the same results when the ICP-MS quantitative procedure is used 16.

Intestinal absorption of ⁴⁴Ca and ²⁵Mg was then calculated as administered ⁴⁴Ca or ²⁵Mg - ⁴⁴Ca or ²⁵Mg excreted in the feces, and retention of ⁴⁴Ca and ²⁵Mg was calculated as administered ⁴⁴Ca or ²⁵Mg - ⁴⁴Ca or ²⁵Mg excreted in the feces and in the urine.

Total cecal SCFA content (μmol/cecum) was calculated as the supernatant SCFA concentration (μmol/ml) × cecal water (ml/cecum).

Soluble Ca and Mg levels in the cecal contents were determined on the supernatant concentration (μg/ml), and soluble Ca and Mg contents per cecum were calculated as (μg Ca/ml or μg Mg/ml) × cecal water (ml).

Values are given as means ± SD, and data were tested by 2-way ANOVA using the General Linear Models procedure of the Super ANOVA package (Abacus, Berkeley, CA). Post-hoc comparisons were performed using Fisher's least significant difference procedures. Differences of p < 0.05 were considered statistically significant. Simple linear correlation analysis was used to assess the relationships between intestinal absorption of Ca and Mg and other relevant parameters. Values of p < 0.05 were considered statistically significant.

Inulin intake at the dose of 7.5% showed only a tendency to decrease animal food intake in this study. The slight decrease in food intake in inulin-fed rats led to a non-significantly lower growth rate (p < 0.10) towards the end of the experiment in inulin-fed rats compared to controls. The lower calorific value of the inulin diets (-4%) compared to the control diets may also be responsible for this reduced weight gain. In addition, food intake decreased significantly with increasing age, as expected (data not shown).

As expected, inulin intake significantly increased cecal wall weight and cecal content and significantly decreased the pH of cecal content. These variables did not change with rat age. In addition, inulin intake considerably increased the individual and total pools of SCFA in the cecal contents (p < 0.0001). The effect of age on these SCFA pools was less clear. No significant age-related difference was observed amongst the control group rats, whereas in the inulin-fed group, the intestinal bacteria produced higher acetate, butyrate and total SCFA in the rats aged 10 mo than in the three other groups (p < 0.05).

The amount of gavaged ⁴⁴Ca was about 1.60 mg/rat, which led to a fecal ⁴⁴Ca enrichment of 10% to 20% in the 4-day feces pool. Fecal ⁴⁴Ca excretion expressed as mg/g of feces or as mg/day increased significantly with age. Consequently, net (mg) and relative (%) ⁴⁴Ca absorption were significantly lower in the aged rats than in the young adult or adult rats. In addition, urinary ⁴⁴Ca excretion (mg) increased significantly with age. Consequently, net (mg) and relative (%) ⁴⁴Ca retention were considerably lower in the aged rats than in the young adult or adult rats. Inulin intake significantly decreased fecal ⁴⁴Ca excretion, expressed as μg/g of feces or as μg, in all groups. Consequently, inulin intake significantly increased net (mg) and relative (%) ⁴⁴Ca absorption. Moreover, inulin intake increased urinary ⁴⁴Ca excretion (mg). Lastly, inulin intake significantly increased net (mg) and relative (%) ⁴⁴Ca retention in the four age-related groups compared to the control diet groups.

The amount of gavaged ²⁵Mg was about 2.50 mg/rat, which led to a fecal ²⁵Mg enrichment of 35% to 65% in the 4-day feces pool. Fecal ²⁵Mg excretion expressed as mg/g of feces or as mg increased significantly with age. Consequently, net (mg) and relative (%) ²⁵Mg absorption were significantly lower in the aged rats than in the young adult or adult rats. In addition, urinary ²⁵Mg excretion (mg) increased significantly with age. Consequently, net (mg) and relative (%) ²⁵Mg retention were significantly lower in the aged rats than in the young adult or adult rats. As expected, inulin intake significantly decreased fecal ²⁵Mg excretion, expressed as μg/g of feces or as μg, in all groups. Consequently, inulin intake significantly increased net (mg) and relative (%) ²⁵Mg absorption. Similarly, inulin intake increased urinary ²⁵Mg excretion (mg). However, inulin intake led to significantly higher net (mg) and relative (%) ²⁵Mg retention in all four groups compared to the control diet.

Mean plasma Ca varied from 95 to 102 mg/L, showing a tendency to increase with inulin intake (+2%, p = 0.0601) and to decrease with increasing age (-1%, p = 0.0619). Mean bone Ca varied from 202 to 228 mg/g dry weight, and was unaffected by inulin intake. However, mean bone Ca increased significantly with increasing age. Mean plasma Mg varied from 16.9 to 18.2 mg/L, showing a tendency to increase with inulin intake (+3%, p = 0.0570). However, mean plasma Mg was not modified by age. Plasma Mg increased in the inulin-fed aged rats (+6.7%), whereas there was no increase in the young and adult rats (-0.3%). Mean red blood cell Mg levels varied from 42.5 to 45.4 mg/L and remained unchanged when age increases or under inulin intake. Mean bone Mg levels varied from 3.72 to 3.92 mg/g dry weight, decreasing significantly with aging (p < 0.0001). However, mean bone Mg was unaffected by inulin intake.

Our results clearly showed that aged rats exhibited less efficient intestinal absorption and retention of Ca. ⁴⁴Ca absorption ranged from 48% without inulin to 66% under inulin intake in the young and adult rats and from 15% without inulin to 27% under inulin intake in the old and very old rats. This decline in Ca absorption with age is not new, and has already been reported in animal and human studies 252627 and is largely confirmed in this study. This decline is primarily due to an energy- and vitamin D-dependent Ca transport component in the elderly 28. Our results clearly showed that inulin intake increased the efficiency of Ca intestinal absorption and retention. The mean ⁴⁴Ca absorption in the four rat control groups was 26.3% compared to 40.2% in the four inulin-fed groups, with an overall increase in ⁴⁴Ca absorption of 53%. These results are in agreement with literature data showing that the effect of inulin on Ca absorption seems to be optimal in the early weeks, then decreasing gradually with experiment duration 202930. One possible explanation for this phenomenon is a down-regulation of the active pathway of intestinal Ca absorption after several weeks of feeding inulin, as previously reported 3132.

Our results showed that aged rats exhibited less efficient intestinal absorption and retention of Mg. ²⁵Mg absorption ranged from 56% without inulin to 86% under inulin intake in the young and adult rats and from 45% without inulin to 70% under inulin intake in the old and very old rats. This decline in Mg absorption with age is not well documented in the literature in either animal or human studies. Few, if any, incomplete studies have reported an age effect on Mg absorption 333435, and the results are inconsistent. Hence, to our knowledge, this is the first robust report to clearly show that Mg absorption decreases with age in the rat. Although Mg absorption is generally described as a passive phenomenon, one component of this absorption remains under hormonal control 3637, which may explain the observed results. Our results clearly showed that inulin intake considerably increased Mg intestinal absorption and retention efficiency. Mean ²⁵Mg absorption in the four rat control groups was 50.6%, compared to 78.0% in the four corresponding inulin-fed rat groups, with an overall increase in ²⁵Mg absorption of 54%. These results are in agreement with literature data showing that inulin intake considerably increases Mg absorption in animals and humans (see recent review 13).

Since Ca absorption is generally well controlled, the observed absorption increase under inulin intake may be down-regulated (known as a feed-back phenomenon) in adult rats. Thus, given that Ca absorption is low and the adaptative phenomenon less well controlled in aged rats, we hypothesized that inulin intake would lead to a much greater increase in Ca absorption in aged rats than in the young or adult rats. Conversely, since Mg absorption is only weakly controlled with a generally consistent increase under inulin intake in adult rats, we hypothesized that inulin intake would increase Mg absorption in both aged rats and young or adult rats to the same extent.

The relative increase in ⁴⁴Ca absorption under inulin intake was 41.5% and 84.5% in the younger (3 and 6 mo old) and older rats (11 and 21 mo old), respectively (figure 1). Although these increase percents are numerically more important in the older rats than in the younger rats, there was no statistically significant interaction between age and inuline. It is highly possible that the number of animals used in this experiment was not enough to reach significant level. Furthermore, the relative increase in ²⁵Mg absorption under inulin intake was 53.5% and 54.5% in the younger and older rats, respectively (figure 1). This indicates that the stimulatory effect of inulin on ²⁵Mg absorption was not age-dependent. It is possible that inulin intake may lead to a higher increase in ⁴⁴Ca absorption in the older rats than in the younger rats, whereas inulin intake leads to a similar increase in ²⁵Mg absorption in young, adult and aged rats, thus confirming the hypothesis we formulated for this study.

In conclusion, as expected, our results confirmed that short-term inulin intake stimulates the absorption of both Ca and Mg. Furthermore, not only these results confirmed that Ca absorption declines considerably with age but also showed for the first time that Mg absorption also declines with age in the rat. Moreover, these results confirmed our hypothesis of a greater stimulatory effect of inulin on Ca absorption in aged rats than in the young or adult rats, and a similar stimulatory effect of inulin on Mg absorption in aged rats and young and adult rats. Further studies are required to explore this effect on longer inulin intake periods and to validate these results on the stimulatory effect of inulin on Ca and Mg absorption in the elderly.