The experiment involved a 2 × 2 design: Rats were fed diets containing either 125 or 25 mmol/kg Ca2+, with the carbohydrate source derived from either a standard sucrose-cornstarch mixture or from fructose. Blood samples were collected at various times, as described below.
The protocol was approved by the Monell Chemical Senses Center Institutional Animal Care and Use Committee.
Diets
The composition of the four experimental diets is shown in Table
1. AIN-93G diet is a diet recommended by the American Institute of Nutrition (now the American Society for Nutrition) in 1993 for growing rats. It contains approximately 10 % sucrose and 52 % cornstarch by weight (Table
1). Fructose-based diets were identical to the AIN-93G-based diets except that virtually all the carbohydrate was provided as fructose (small amounts of sucrose were used as a vehicle for the mineral and salt mixes). These were provided in either “replete” (125 mmol/kg) or “low” (25 mmol/kg) calcium concentrations. The 125 mmol/kg calcium concentration is well above requirements for growth [
20,
21]. The 25 mmol/kg calcium concentration was chosen based on our previous work showing that it effectively reduced blood calcium concentrations and induced the motivation to consume calcium but it did not cause a failure to thrive (e.g., [
22,
23]). For convenience, we abbreviate the diets based on their carbohydrate source [i.e., similar to AIN-93G or fructose (FRU)] and the concentration of calcium (125 or 25 mmol/kg).
Procedures
After 3 days to acclimatize to the vivarium environment, the rats were assigned to four groups matched for body weight and the percentage of body weight that was fat, which was assessed using a Bruker Minispec nuclear magnetic resonance analyzer (see below). The experimental diets were introduced in two steps: On Day 0 (at age 24 days), half of the rats were switched to AIN25 diet; the other half remained on AIN125 diet. On Day 3, half of the rats in each AIN125 and AIN25 diet group were switched to fructose-based diets (FRU125 and FRU25). Thus, there were 18 rats in each of the four groups.
The body weight, food intake, and water intake of each rat was recorded daily. Body weight (±0.1 g) was measured using a top-loading balance. Food intake (±0.1 g) was based on the difference in weight between food given each day and food remaining the following day; it was corrected for spillage, which was collected under each cage. Body composition was assessed weekly. To accomplish this, each rat was carried to an adjacent room housing a Bruker Minispec LF110, placed into a plastic restraining tube, and then inserted into the core of the machine for ~90 sec while its body composition was assessed by magnetic resonance technology. The rat was then returned to its home cage.
To confirm the development of calcium deficiency, blood ionized calcium levels were measured on Days 10 and 17, at ~7 h after the start of the light period. Approximately 70 μl of blood was collected from the tip of the tail into a heparin-coated capillary tube (RAPIDLyte Multicap, Siemens) and within 15 sec analyzed using a Rapidlab 348 blood gas analyzer (Siemens, Germany) located in the vivarium. For this and subsequent tests, care was taken to collect blood samples from the rats in an order that controlled for disturbances and any delays between sampling and assaying. To do this, the samples were collected successively from rats in tetrads containing one representative of each of the four diet groups, and the order within each tetrad was counterbalanced.
Fasting blood glucose, insulin and triglyceride (TG) levels were measured in blood samples collected on Days 11, 18 and 42, at ~7 h after the start of the light period. Taking into account that young growing rats were relatively small at the beginning of the experiment we opted to use a morning fast instead of more severe overnight food deprivation: In small rodents prolonged overnight fasting induces a catabolic state, and along with standard housing conditions at subthermoneutral temperatures it can decrease metabolic rate [
24]. In our experiment, we aimed to collect measurements under less extreme physiological conditions. Rats were fasted for 5–6 h and then a blood sample was collected from the tip of the tail into an EDTA-treated tube. The tubes were centrifuged and plasma glucose and TG levels were analyzed using enzyme-based colorimetric assays (Cayman Chemical, MI). Insulin levels were measured using ELISA (Rat insulin kit # EZRMI-13 K, EMD Millipore, MA). Sometimes when the rats were young, insufficient blood was collected to conduct all the measurements. In these cases, insulin measurements were given priority.
An oral glucose tolerance test was given between Days 23–25, at ~7 h after the start of the light period. After an overnight fast, each rat was gavaged with 8 ml/kg of 25 % glucose solution (2 g/kg). Blood was collected at 0, 15, 30, 60 and 120 min. Plasma glucose and insulin levels were measured as described above.
An insulin tolerance test (ITT) was given on either Day 36 or 37, at ~7 h after the start of the light period. After a 5–6 h fast, each rat received insulin [0.75 IU/kg, 0.75 ml/kg, i.p. of 1 IU insulin (HumilinR, Eli Lilly) solution prepared in isotonic saline]. For this test, glucose levels were measured using an Accu-Check Aviva Plus blood glucose meter (Roche Diagnostics).
Data analysis
Body weights were analyzed using weekly measurements taken on the day of the body composition assessment. Daily food intakes were averaged into weekly blocks, based on the sum of daily measurements divided by the number of days; days when rats were fasted overnight were excluded from analysis.
Areas under the glucose and insulin curves during oral glucose tolerance tests were calculated using the trapezoidal rule [
25,
26].
Two-way ANOVAs with factors of Diet Carbohydrate and Diet Calcium were used to analyze the area under the glucose and insulin curves. For all other measures, three-way mixed-design ANOVAs were used, with the same two between-subject factors and Time as a within-subject factor. When appropriate, differences between pairs of means were assessed using Fisher LSD tests. The criterion for statistical significance was p ≤ 0.05.
One animal was excluded from the analyses of the oral glucose tolerance test due to unsuccessful gavage of the glucose load. The Week 3 body composition of one rat was lost due to a technical error.