This protocol describes the use of a highly palatable, western-style cafeteria diet to model overeating and obesity in rodents. Here, we provide a detailed outline of food selection, preparation and measurement, and explain methodological factors that assist in generating a robust and reproducible phenotype.
Obesity is rapidly increasing in incidence in developed and developing countries and is known to induce or exacerbate many diseases. The health burden of obesity and its comorbid conditions highlight the need for better understanding of its pathogenesis, yet ethical constraints limit studies in humans. To this end externally valid models of obesity in laboratory animals are essential for the understanding of being overweight and obesity. While many species have been used to model the range of changes that accompany obesity in humans, rodents are most commonly used. Our laboratory has developed a western-style cafeteria diet that consistently leads to considerable weight gain and markers of metabolic disease in rodents. The diet exposes rodents to a variety of highly palatable foods to induce hyperphagia, modeling the modern western food environment. This diet rapidly induces weight gain and body fat accumulation in rats allowing for the study of effects of overeating and obesity. While the cafeteria diet may not provide the same control over macronutrient and micronutrient profile as purified high-fat or high-fat, high-sugar diets, the cafeteria diet typically induces a more severe metabolic phenotype than that observed with purified diets and is more in line with metabolic disturbances observed in the overweight and obese human population.
Obesity and its related comorbidities make an enormous contribution to global health burden1 and account for 7% of disease burden in Australia2. A leading risk factor for obesity is consumption of unhealthy diets that are high in saturated fat and refined carbohydrates, and low in fiber and micronutrients3. Identifying targets for therapeutic intervention for obesity requires models that can systematically assess effects on multiple biochemical and physiological systems. Our understanding of the etiology of obesity has been advanced substantially by work using rodent models, where behavioral, metabolic and molecular effects can be studied across time under controlled conditions where environmental factors can be easily manipulated.
The cafeteria diet (CAF) model of diet-induced obesity consists of supplementing rodents' standard chow diet with a variety of palatable foods that are high in either saturated fat, refined carbohydrates, or both. Examples of these foods include cakes, sweet biscuits, and high-fat savory snacks (such as processed meats, cheese and chips). It reliably promotes hyperphagia and rapid weight gain in rodents. The key features of the model are the provision of a variety of highly palatable foods, designed to simulate the modern food environment. Access to variety increases food intake in rats over the short-term4 and in humans5 even when the foods are matched for palatability and vary only in flavor and olfactory cues4,6. However, one study showed that providing energy- and macronutrient-matched purified diets that varied in flavor and texture had no effect on long-term body weight gain in rats7, suggesting that nutrient composition and distinct post-oral effects of different foods may also contribute to overeating. Exposure to multiple tastes and textures overcomes sensory-specific satiety, which describes the decrease in desire to eat a recently eaten food relative to an alternative5. Across many cohorts in our laboratory, we have similarly observed that the use of highly palatable foods further amplifies overeating.
This CAF diet has been used for over 40 years, since Sclafani8 reported that female rats exposed to an assortment of ‘supermarket foods' (marshmallows, chocolate, peanut butter, cookies, salami and cheese among them) exhibited accelerated weight gain relative to controls. This and other early studies noted that CAF-style diets appeared to accelerate weight gain more effectively than pure high-fat or high-carbohydrate diets 8,9. Work in the 1980s characterized the macronutrient profiles10 and meal patterns11 of rats fed CAF diets, and showed profound changes to fat mass and insulin levels9,10 and thermogenesis12. Our group has used the CAF diet to model obesity for over two decades13,14 and during this time we have used several variants of the diet. Rats are presented with at least two sweet and two savory food items each day, in addition to regular chow and water. In recent years we have begun to supplement solid CAF foods with 10% sucrose solution. The ability to tailor the CAF diet to different experimental designs is a strength of the model.
CAF diets promote immediate hyperphagia (i.e., within the first 24 h) and steady gains in body weight and fat mass. However, a consequence of maximizing variety is that macronutrient and micronutrient intake is not controlled, a point some view as an insurmountable flaw15. Studies of diet-induced obesity more commonly use purified high-fat (HF) or combined high-fat, high-sugar (HFHS) diets, which offer precise control over nutritional content and are less labor-intensive than the CAF model, which requires daily monitoring and careful planning and execution of the schedule. The translational relevance of commercially available purified HF diets is a topic of ongoing debate, as their fatty acid profile and proportions of fat and sucrose may not align with human dietary intake16. While CAF diet does not offer the same degree of control over nutrient composition as purified diets, it aims to model the palatability and variety that characterizes food options in most modern societies.
The protocol described here has been optimized for use in rats. While we have used the CAF diet successfully in mice17,18, soft food grinding may introduce further error reducing the reliability of food intake measures19. This protocol is approved by the Animal Care and Ethics Committee at the University of New South Wales and complies with the Australian guidelines for the use and care of animals for scientific purposes (8th Edition) provided by the Australian National Health and Medical Research Council.
NOTE: Very few adverse effects have been observed in our short-term studies (i.e. <10 weeks ad libitum CAF access); there is no evidence of changes to general wellbeing, activity, sociability or anxiety-like behavior in rats on CAF diet20. After longer intervals (>16 weeks) very occasional cardiovascular incidents have been observed in CAF-fed rats.
1. Animal Acclimatization and Housing
2. Diet Selection and Setup
3. Cafeteria Diet Preparation
4. Food Intake Over 24 H
NOTE: Food intake measurements are conducted over a discrete 24 h period several times per week.
As shown in Figure 2A, CAF diet feeding produces a 2.5-fold increase in energy intake relative to chow controls, based on data from three cohorts of male Sprague Dawley rats, that is consistent over 6 weeks. Other studies have confirmed that this extent of hyperphagia is sustained over 1021 and 1622 week experiments. The weight curve (Figure 2B) indicates CAF diet feeding leads to a 20% difference in mean body weight compared with control after 3–4 weeks of diet, comparable to the body weight gain consistent with the onset of obesity in humans. At 6 weeks, mean difference in body weight gain between CAF and control groups is 67% (Figure 2C) and adiposity, determined by EchoMRI, is approximately doubled (Figure 2D). Cafeteria-fed rats typically eat 5–10% of their daily energy as chow (~5 g/rat/day).
Macronutrient intake profiles can be calculated on food intake measurement days using nutritional information from the product manufacturers. We observe consistent macronutrient intakes across sexes and ages, with CAF-fed rats consuming approximately 8% of energy as protein, 34% as fat and 58% as carbohydrate. Our maintenance chow provides 22% protein, 13% fat and 65% carbohydrate. Relative to estimates of macronutrient consumption in human populations (18.3% protein, 44.9% carbohydrate and 30.9% fat in Australia23; 15.7% protein, 48.7% carbohydrate and 33.7% fat in the United States of America24), our CAF-fed rats consume a lower proportion of energy as protein, a higher proportion as carbohydrate, and a comparable proportion from fat. However, when considering absolute intake, CAF-fed rats overeat all three macronutrients relative to controls (Figure 3A–C for males, Figure 3D–F for females), indicating that they are not protein deficient. Rather, the macronutrient composition results from the dramatic hyperphagia evoked, which is driven predominantly by excess carbohydrate and fat intake, not usually observed in human subjects who tend to become overweight and develop obesity more gradually.
A recent study we conducted suggests that female rats may be particularly vulnerable to the obesogenic effects of CAF diet. Energy intake was 3.8 times greater in CAF-fed females relative to controls, which was sustained over 6 weeks (Figure 4A). A 20% weight difference between groups was observed after only 2 weeks of CAF exposure (Figure 4B). After 6 weeks of diet exposure, body weight gain was 12 times greater in CAF rats (Figure 4C) and fat mass was doubled (Figure 4D) compared with healthy controls. The suggestion of greater susceptibility to diet-induced weight gain in females is supported by an earlier study by Sclafani and Gorman, which showed that a cafeteria diet induced significantly greater weight gain in females than in males25.
Providing too little of each CAF food may artificially constrain measurements of energy intake. This is most easily addressed by checking that there is some food remaining in the CAF cages 24 h after feeding (when the food is refreshed). Figure 5 shows how the use of multiple food sets for food intake can lead to substantial shifts in macronutrient intake despite comparable overall energy intake. In this instance, the variability was due primarily to the balance of fat and carbohydrate against a backdrop of consistent protein intake. Analyses of the food sets used indicated that high carbohydrate intakes were observed when a highly-preferred cake was provided; higher fat intakes were observed when this cake was absent.
Figure 1: Example of cafeteria diet across three days. CAF diet should consist of varied food each day to induce sustained hyperphagia. Preparing CAF foods can be streamlined by placing each cage’s food into a designated container. This allows for easy, well-timed delivery into each cage. Day 1 = chicken nuggets, beef-flavored dog food, chocolate cream biscuits, jam roll, high-fat purified diet. Day 2 = meat pie, chicken nuggets, scotch finger biscuits, caramel mud cake. Day 3 = dim sum, chicken-flavored dog food, custard cream biscuits, blueberry cheesecake. Please click here to view a larger version of this figure.
Figure 2: Representative results in multiple male cohorts. CAF diet produces a consistent increase in energy intake (A), and body weight (B) over 6 weeks in male adult rats. This is accompanied by significant increases in body weight gain (C) and fat mass (D) when assessed by EchoMRI at 4 weeks of diet. Data are represented as mean ± SEM; n = 48 for individual data; n = 12 for energy intake data (cage as the unit of analysis). Please click here to view a larger version of this figure.
Figure 3: Macronutrient intake in multiple male cohorts. CAF diet exposure increases total intake of carbohydrate (A), fat (B) and protein (C) for male rats. These increases are comparable in female rats for carbohydrate (D), fat (E) and protein (F) intakes. Data are represented as mean ± SEM; n = 12 or n = 4 cages (males and females respectively) averaged over 6 weeks. Please click here to view a larger version of this figure.
Figure 4: Representative results in a female cohort. Cafeteria diet produces a consistent increase in energy intake (A) and body weight (B) over 6 weeks in female adult rats. After 6 weeks of diet a substantial difference in body weight gain (C) and fat mass (D) is also observed. Data are represented as mean ± SEM; n = 12 for individual data; n = 4 for energy intake data. Please click here to view a larger version of this figure.
Figure 5: Macronutrient intake over time. Using multiple food sets to assess food intake can lead to differences in macronutrient intake over time given differences in individual preferences in a cohort of male rats. Data are represented as mean ± SEM; n = 4 cages. Note that overall macronutrient intake matches the pattern usually observed; CAF-fed rats consumed 60% energy as carbohydrates, 33% fat and 8% protein, as energy. Please click here to view a larger version of this figure.
Monday | Tuesday | Wednesday | Thursday | Friday | Saturday | Sunday | |
DAILY FOODS | Healthy chow and potable water | ||||||
CAKE | A | D | B | A | C | B | E |
PROTEIN 1 | A | C | D | A | C | D | A |
PROTEIN 2 | B | B | A | B | B | A | C |
COOKIE | B | A | C | B | D | C | A |
OPTIONAL ADDITIONAL DAILY FOODS | HFHS chow, 10% sucrose solution |
Table 1: Example weekly food plan for the cafeteria diet. The CAF diet promotes hyperphagia by providing a variety of palatable foods that are varied daily, as shown in this example weekly food plan. Letters A-E denote unique foods for that food group (for example, Cake A might denote chocolate mud cake and Cake B vanilla sponge). Food intake days, shaded in grey, should be positioned evenly across the week and ideally are kept as consistent as possible. While CAF diet always includes continuous access to healthy chow and water, optional additional daily foods can include HF or HFHS chows, and 10% sucrose solution.
PER 100g (derived from manufacturer) | ENERGY PER 1g | Recommended starting value (kJ/rat) | |||||||||||
Food | Energy (kJ) | Protein (g) | Total Fat (g) | Saturated Fat (g) | Total Carbohydrate (g) | Sugar (g) | Energy (kJ/g) | Protein (kJ) | Total Fat (kJ) | Saturated Fat (kJ) | Total Carbohydrate (kJ) | Sugar (kJ) | |
Protein A | 830.00 | 6.00 | 6.10 | 3.20 | 28.90 | 2.20 | 8.19 | 1.02 | 2.26 | 1.18 | 4.91 | 0.37 | 350.00 |
Protein B | 906.00 | 7.30 | 11.10 | 4.60 | 21.10 | 1.80 | 8.94 | 1.24 | 4.11 | 1.70 | 3.59 | 0.31 | 350.00 |
Cake A | 1470.00 | 4.60 | 13.30 | 3.70 | 52.40 | 33.10 | 14.61 | 0.78 | 4.92 | 1.37 | 8.91 | 5.63 | 200.00 |
Cake B | 1660.00 | 4.00 | 18.40 | 4.30 | 53.60 | 36.30 | 16.60 | 0.68 | 6.81 | 1.59 | 9.11 | 6.17 | 200.00 |
Cookie A | 1920.00 | 4.30 | 20.60 | 12.70 | 63.20 | 33.20 | 19.10 | 0.73 | 7.62 | 4.70 | 10.74 | 5.64 | 200.00 |
Cookie B | 2040.00 | 5.70 | 21.00 | 11.20 | 8.50 | 4.10 | 10.18 | 0.97 | 7.77 | 4.14 | 1.45 | 0.70 | 200.00 |
Table 2: Nutritional information for selected cafeteria food items. This table depicts the nutritional information obtained for several core items in the CAF diet. It is important to ensure that the daily options provide similar macronutrient availabilities, and that rats have access to adequate protein. For each daily set of foods used, it is helpful to calculate the overall energy density and the macronutrient content. The final column contains the recommended starting volume of each food (as energy per rat) for male Sprague Dawley rats at 200 g.
CAGE | WATER (g) | CHOW (g) | Cake A (g) | Protein A (g) | Protein B (g) | Cookie B (g) | ||||||
IN | OUT | IN | OUT | IN | OUT | IN | OUT | IN | OUT | IN | OUT | |
1 (CHOW) | ||||||||||||
2 (CAF) | ||||||||||||
3 (CHOW) | ||||||||||||
4 (CAF) |
Table 3: Example food intake sheet. The weights of each CAF food item (g, per cage) should be carefully recorded on a spreadsheet.
By exposing rats to a variety of highly palatable foods high in fat and sugar, the CAF diet protocol described here provides a reliable and robust model of the so-called ‘western diet’ eaten by many people. Hyperphagia—assessed as a significant increase in energy intake relative to controls—is observed within the first 24 h of exposure, with statistically significant body weight differences seen within weeks. Thus, CAF is an effective model of diet-induced obesity for rodents.
Several studies have reported that CAF-style diets produce a more exaggerated obesity phenotype than purified HF or HFHS diets. Sampey et al.26 showed that relative to rats fed a HF diet, CAF exposure led to greater liver and adipose tissue inflammation, poorer glucose tolerance and more insulin resistance. While that study used a lard-based HFD, two other experiments found that CAF diets increased adiposity relative to HFDs based on butter27 and coconut oil28. Similarly, Higa et al. found that while both HF and CAF diets increased visceral fat relative to control diet in mice, only CAF increased food intake and accelerated the onset of hyperglycemia, glucose intolerance and insulin resistance29. Another study in mice reported that in addition to more pronounced metabolic effects, CAF exposure increased the incidence of liver and heart pathologies (fibrosis, steatosis and apoptosis measures) relative to a purified HFD30. However, recent work has shown that a specifically formulated 'western diet' produced stronger effects on metabolic, adipose and inflammatory measures than a traditional CAF diet31. Further work identifying how different types of obesogenic diets affect metabolic outcomes is needed, as there are limited data comparing cafeteria diets with purified diets high in sugar, or high in both sugar and fat.
Most of our work with this model has been with outbred Sprague-Dawley rats. The CAF protocol has been optimized for investigating the metabolic effects32 of the modern food environment. We have used this model to study the microstructure of feeding across the day22, intermittent access models of ‘bingeing’33 and in experiments where energy intake is yoked to control levels34. More recent studies have examined dietary effects on cognition20,35 and the gut microbiome36,37. Versions of the CAF diet have also been used to study maternal obesity38 and to explore feeding responses to hypothalamic feeding peptides in obesity14.
It is important to note that while there is a large body of evidence indicating that the cafeteria diet induces hyperphagia and obesity in rats and mice, these studies have largely been conducted in outbred Sprague Dawley13,14,20,39,40,41,42 and Wistar6,26,31,43,44 rats, with relatively few studies performed in Balb/c17,18 and Swiss45,46 mice as well as other rodent strains. Therefore it is unknown whether the effects of cafeteria diet reported in the literature will be observed in other strains, especially as there are known strain differences in response to obesogenic diets in both rats47,48 and mice49. Additionally, starting age and weight are also important methodological factors that may modulate the effects of obesogenic diets on metabolic outcomes25. Most of our prior work has started rats on CAF diet in early adulthood, including the studies generating the representative data reported here.
Several local factors should also be considered. The food schedule proposed here may need to be modified when setting up CAF diet for use with a different strain or supplier. Local food supplies will determine the specific CAF foods to be used and consumption should be monitored carefully each time a new food is introduced. Nutritional information provided on food packages must be retained and checked over time to ensure that macronutrient calculations are accurate.
Successful application of the CAF model requires careful planning and daily monitoring of cages. Additional time is needed to purchase, thaw and prepare food items daily, and food intake measurement days are labor-intensive. These factors may pose logistical constraints and should be considered when evaluating the model for use. Researchers interested in adapting the CAF model should therefore consider that the reliable hyperphagia and obesity phenotype observed with CAF comes with reduced control over nutrient intake and increased preparation time.
Several limitations of the CAF model are important to consider. Since this model allows rodents to select foods, macronutrient intake cannot be determined for individual rats unless individual housing is employed. While the average macronutrient intake across cohorts is relatively stable, we observe variability in rats’ metabolic responses to CAF diet within cohorts, which may relate to differences in individual diet selection. Furthermore, the CAF diet items are not fortified, meaning micronutrient availability may be low. However, our rats always have access to healthy, nutritionally complete chow (which typically comprises 5–7% of energy intake) and are provided with nutritionally complete dog roll as a savory food on 3–4 days per week. It is also important to note that low micronutrient availability is observed in human western-style diets high in fat and sugar, and a high proportion of adults with obesity show micronutrient deficiency50.
There are also several important caveats to note regarding food intake measurement. As locating all fragments of food is impossible, it is important to ensure that an identical procedure is used for each cage. As food intake is measured on a per-cage basis, we analyze energy intake with cage as the unit of analysis, assuming equal consumption for all rats within. However, because the model is explicitly designed to maximize choice and variety, total energy intake, macronutrient and micronutrient profiles for individual CAF rats are likely to vary. Nonetheless, this provides an opportunity to study individual differences in consumption of, and metabolic response to, the CAF diet. Further studies comparing age-matched male and female rodents across the lifespan will be important to fully characterize sex differences in response to diet. Finally, rodent models cannot—and do not attempt to—recreate the complex range of economic, psychological and social factors that influence human eating behavior. However, given homologous neural circuitries underlying feeding behavior across mammals, and the similar physiological response to positive energy balance (i.e., in the deposition of fat and altered metabolic function), we believe this model holds value in understanding how poor diets and obesity alter body and brain function.
The authors have nothing to disclose.
The work was supported by NHMRC project grants (#568728, #150262, #1126929) to MJM.
2-5 L plastic bottle | For preparing 10% sucrose solution, if applicable | ||
Chopping board | Plastic is advised | ||
Freezer | For storing CAF foods | ||
Gordon's maintenance rodent chow | Gordon's Specialty Stockfeeds (Australia) | Maintenance diet used in our laboratory (14 kJ/g; 65% carb, 13% fat and 22% protein, as energy) | |
Large plastic storage boxes | All items above can be stored in containers for easy access | ||
Large spoon | For CAF diet preparation | ||
Microwave | For CAF diet thawing (when required) | ||
Non-serrated knife | For CAF diet preparation | ||
Paper towel | Important for cleaning work surfaces and the knife during CAF prep | ||
Plastic containers | These are for weighing CAF food items on measurement days | ||
Plastic funnel | For preparing 10% sucrose solution, if applicable | ||
Red light | As CAF diet should be refreshed near the onset of the dark phase each day, a red light will assist when working in the dark | ||
Tuna tins | For presenting 'wetter' CAF food items. Plastic containers may also be suitable | ||
Weigh container x 3 | Separate containers should be used to weigh rats, chow & bottles, and CAF foods | ||
Weighing scale | Sensitivity to 0.1g is recommended | ||
White sugar | For 10% sucrose solution, if applicable |