Dr. Michael Pellizzon of Research Diets, Inc.
, recently presented a webinar on the underappreciated impact of rodent diets on microbiome and gut health research
, where simple husbandry decisions can profoundly affect both animal model phenotypes and experimental outcomes.
Due to time constraints, many of the questions submitted during the webinar went unanswered. We present a full Q&A here.
Impact of Purified Diets
Q: Can purified diets be used for long-term experiments?
Dr. Michael Pellizzon (MP
): Yes. They have been used for very long experiments of one year or more in length. The animals are fine, and depending on the purified diet being fed, they may or may not live as long as those fed grain-based chow.
Q: Can you consider control animals on a purified diet to be healthy and does a purified diet support the normal development of animals?
: Purified diets have been used for a long time and have been considered to provide the right nutritional needs to support breeding and a long lifespan. There are some nutritional differences between chows and purified diets which can impact animals, for example in metabolic parameters and morphological changes in the cecum and colon. We think we can improve upon current purified diets by adding different types of fiber from what is currently used.
Q: Once you have switched animals to a new diet, how long until those animals are acclimated on the new diet?
: In general, mice can acclimate to a new diet pretty quickly. For metabolic disease, it may take a week or two on a purified diet prior to study initiation. The shrinkage of colon and cecum that is observed on some diets happens very quickly. The microbiome can change very quickly also, sometimes within a day.
Germ-Free Mice and Purified Diets
“With purified diets, we are able to understand how each ingredient affects biology in a very controlled way.”
–Dr. Michael Pellizzon, Research Diets, Inc.
: We have successfully made purified diets for use with germ-free mice
, though such diets must be irradiated at a higher level than we typically use. The irradiation facility Research Diets uses does not guarantee sterility of diet after irradiation and we recommend that the diets used for germ-free mice are tested for microbial load.
Q: How does the cecum of germ-free mice on purified diet with cellulose differ from that of germ-free mice fed chow?
: The colonic length and thickness of the mucosal lining of germ free mice were found to be similar for mice fed grain-based chow and a low fat purified diet with cellulose as fiber (Desai et al. 2016
, Figure 4C and F). I have heard from some of my collaborators that their germ-free mice fed a purified diet high in fat had smaller cecums than those fed a grain-based chow.
Q: At what level do you irradiate diets for germ-free mice and have you analyzed fat oxidation in the diet post-irradiation?
: We typically irradiate our diets at 10-20 kGy, but with diet that will be used with germ-free mice, that level may not be sufficient, so we irradiate twice to give a final dose of 20-40 kGy. We have measured fat oxidation; that does increase, depending on the irradiation dose and diet.
Q: Does double irradiated diet show any bacterial growth on culture?
: We have heard a couple of cases where our double irradiated diets had some contamination, but we are still investigating these cases. Because of this and just to be sure for ourselves that double irradiation will be enough to kill all bacteria, we are going to have some of our diets tested.
Our current test will be with non-irradiated diets compared to those that are irradiated once, twice, or 3 times at 10 — 20 kGy each run (which should be additive).
One bacterium that can survive irradiation is Deinococcus radiodurans. If it is present prior to irradiation, this should be killed when irradiating the diet 2 times, given there are data suggesting that 2.7 Mrad was enough to kill this bacteria
and our irradiation is 10 — 20 kGy each run (so 20 — 40 kGy if run twice. FYI, 1 Mrad = 10 kGy). Based on this study, other bacterial strains are much more sensitive and should be killed in the range of 10 - 20 kGy, so one irradiation run should kill these other bacteria.
We are aware that there will be dead bacteria, such as lactococcus, which is used to precipitate lactic acid casein, a common ingredient in purified diets. Irradiation at 10 kGy should kill these bacteria (if any survives pasteurization), but the DNA from this (and other) bacteria will be picked up by 16S rRNA sequencing in germ-free or anti-bacterial treated mice.
The company which irradiates our diets doesn't guarantee sterility, so we are unable to make the claim that the irradiated diets are 'sterile'. While we anticipate that double irradiation should be enough to kill all bacteria, we recommend that you do some microbial testing of the diet just be on the safe side if you plan to work with germ-free mice.
View the Taconic Biosciences' Webinar and Download the Presentation:
Q: Can purified diets be autoclaved for use with germ-free mice?
: Research Diets recommends that purified diets not be autoclaved. The high heat level may reduce certain vitamins and it may increase fat oxidation.
Irradiation is typically performed around 100°F, but autoclaving is done at much higher temperatures, so this can result in higher levels of fat oxidation and vitamin losses relative to irradiation. Autoclaving may add moisture to the diet. Also, advanced glycation end products may be produced by heating the diets at temperatures comparable to autoclaving (120 - 130°C for 20 - 30 min), which may alter certain parameters such as atherosclerosis and diabetes in certain mouse models (Lin et al., 2003
; Peppa et al., 2003
Particularly for a high fat diet, autoclaving will increase peroxidation of the diet, so we recommend against autoclaving.
Q: Can irradiation affect bioavailability of vitamins and minerals?
: Irradiation can reduce the thiamin and vitamin A levels in a diet. For diet that will be irradiated twice (as for use with germ-free animals), approximately 50% more vitamin mix is added in order to preserve the required levels in the diet after irradiation.
Q: Can germ-free mice become obese on a high fat purified diet?
: Some work has been done in this area. While conventional C57BL/6
are susceptible to obesity on a high fat purified diet, germ-free C57BL/6 mice are resistant to diet-induced obesity. Germ-free C3H
mice are sensitive to weight gain on a high fat purified diet (Bindels et al., 2017
In addition, it has been observed that the type of dietary fat can influence the potential for germ-free mice on a C57BL/6 background to gain weight (Kübeck et al., 2016
Q: Can autoclaved chows affect the immune system of germ-free mice? As they contain endotoxins, dead bacteria and viral particles could they still trigger TLR signaling?
: This is a very good question. It is possible that endotoxins in autoclaved chow may trigger the immune system of germ-free mice. A publication by Hrincir et al.
suggested that an LPS-rich sterile chow (ST1) can drive the expansion of B and T cells in Peyer's patches and mesenteric lymph nodes relative to a purified ingredient diet, AIN-93G in germ-free mice.
However, the conclusions may not be correct given there are many differences between the LPS-rich chow and the purified diet. Thus what is really driving these changes in the immune system between germ-free mice fed the chow compared to the purified diet is currently unknown.
Q: In my studies using purified diets, rats expressed high levels of LPS and pro-inflammatory cytokines. It is decreasing the gut health. Please address this.
: It is true that purified diets can certainly increase LPS and pro-inflammatory cytokines in mice, relative to those given a grain-based chow. I did have one slide that contrasted a purified high fat diet with a grain-based chow (Neyrinck et al., 2012
, Figure 4) and indeed demonstrated that LPS is much higher in mice fed the high fat diet versus those on the chow.
When adding in a soluble fiber such as wheat arabinoxylan oligosaccharides in the context of a purified high fat diet, LPS was reduced to a similar level as in the chow fed mice. Similarly, there was a reduction in pro-inflammatory cytokine IL-6 to levels similar to those of chow fed mice. This is very intriguing and really points to the importance of having soluble fiber in the diet to reduce LPS. Likely there is some influence by increasing tight junction proteins like Zonula Occuldens, which was suggested by this study, at least at the mRNA level. I find this work (and others like it) very interesting and suggests that it is quite easy to reverse the trend you are seeing if we add in some soluble fiber to purified diets, even when the diet is high in fat.
So, yes, I agree with you that purified diets can indeed increase LPS and pro-inflammatory cytokines relative to chow, but I think we can improve upon purified diets by simply adding in one or more soluble fibers. Ultimately, through this process of testing different fiber types (and other nutritional manipulations), we will learn a lot about the microbiome and gut health and how this affects overall health with purified diets.
Q: What are your thoughts on chow diets which come with extensive testing and certification such as accounting of micro- and macronutrients? How inferior are these to purified diets for experimental purposes?
: Chows are made with grains and animal by-products, each of which may contain multiple nutrients and non-nutrients. This can be a problem.
Understanding the diet used in an experiment is critical, particularly so for microbiome studies. There are a lot of factors in chow diets which may influence microbiota such as fiber and phytoestrogens. Ideally, diets for microbiome experiments should be open formula using purified ingredients, so you know what and how much you are feeding in terms of nutrient levels. This allows you to maintain consistency across batches.
Q: Within a given facility that purchases a single type of mouse chow over the course of years, can we consider data from two different time periods to be comparable? It seems you are saying that even without changing brand or type of chow, there may be considerable variability in the nutrient balance.
: Some nutrient (and non-nutrient) variability can be expected with grain-based chows, which may not be surprising given that each ingredient contains multiple nutrients and non-nutrients. There have been data suggesting that nutritional changes can occur from lot to lot of the same chow. In particular, this has been extensively reported for phytoestrogens (see Jensen and Hoitinga
, Brown and Setchell
, and Thigpen et al
.). Certain chows have been evaluated lot to lot and variability for nutrients and non-nutrients has been documented — see publication by Greenman et al
., Topham et al
. and Newberne et al
. These data demonstrate that chows can present significant variability.
Also, since most chow formulas are proprietary, and subject to change (e.g. to keep the level of a given nutrient constant from lot to lot, such as protein), such a change can also result in changes to the overall nutrient balance.
The other factor which I talked about that hasn't been studied for its lot to lot variability is fiber. Different chows have been found to have variable fiber levels and types (Wise & Gilburt, 1980
). Given the complexity of fiber in terms of the levels and different fiber types within each grain ingredient, I suspect that there will be differences in soluble and insoluble fiber levels from lot to lot of the same chow. I'm interested in testing whether this is the case from one lot of chow to the next, so hopefully at some point I will have more data to share.
: Taconic uses the NIH #31M
chow diet as its standard production diet. This is an open formula diet, and vendors providing this diet to Taconic provide a record of the amount of each ingredient used to manufacture every batch of feed and the results of a proximate analyses conducted on a representative feed sample collected from each shipment. Proximate analysis refers to a determination of the major components of a feed such as moisture level and crude protein to ensure they are consistent.
Q: Is there a database or other resource containing information on all the different fiber types which can be incorporated into a purified diet?
: There are many different types of fibers available. There does not appear to be a database which catalogs this information.
Q: What immune cell subsets that are changed by dietary fiber?
: There are data showing that fiber type can influence immune cells. Kelly-Quagliana et al. (Kelly-Quagliana, Nelson, & Buddington, 2003
) shows soluble fibers (inulin and oligofructose) actually influenced white blood cell counts in mice. I'm sure there are more data out there than this, but I think this is a pretty good example of how easy it is to modify immune function in mice by changing the fiber type.
Q: What is the recommended maximum amount of fiber in a diet? Can you mix a few different types of fiber to loosely mimic what is found in a chow diet?
: The level of fiber can be increased in a purified diet very easily. Matching chow can be challenging, since fiber in chow comes from plant cell walls within the chow ingredients. The soluble and insoluble fiber levels in a chow can be determined easily, and the levels of cellulose, lignin and pectin can be measured. It is possible to match levels of soluble and insoluble fiber in a chow when using a purified diet. Many different types of fibers can be added to a particular purified diet to loosely mimic the fiber composition of chow.
Q: Are there sex differences in the absorption of inulin and effects on the cecum?
: Inulin is not absorbed; it is fermented by the gut microbes. Adding inulin to the diet adds some energy because it is fermented to short chain fatty acids which provide energy once those are absorbed. I am not aware of a sex difference in how inulin is handled by males and females.
Q: Have you looked into diarrhea related to high doses of non-digestible fiber?
: It is pretty difficult for wild type mice and rats to get diarrhea on any diet, either with fiber or without, but there may be a change in fecal consistency.
With our own diet testing, there were no visible differences in feces when cellulose, inulin or fructooligosaccharides were fed to C57BL/6 mice over 2 weeks up to 200 g per 4084 kcals (close to 17% by wt). One older publication
suggested that there is some mild to marked diarrhea (i.e. loose stool) in rats fed oat bran. This is mentioned in the discussion (under laxation section). Some hamsters can get diarrhea if no fiber is present which can be corrected with increasing fiber levels (Hayes et al
DSS-induced colitis causes diarrhea, and this can be actually affected by the type of fiber present in a purified diet. Fructooligosaccharides added were found to increase diarrhea scores in this model fed a purified diet rapidly, and at the end of the 7 days, both purified diets had higher diarrhea scores than the chow fed group. Fructooligosaccharides had a tendency to increase this score when added to the chow (Goto et al
Q: What is the cost differential between chow and purified diets?
: Purified diets, regardless of their composition, have a higher cost compared to chow, usually several fold. This is because the ingredient costs are higher due to their increased level of refinement relative to chow ingredients. While their costs are higher, this increased cost comes with more confidence in the diet composition being consistent from lot to lot.
Some purified diets cost more than others, and this is mainly due to the length of time of manufacture, which can change dependent on several factors including the level and type of fat or other modifications to achieve a particular research goal.
Q: What is important with regards to the content of various fatty acids in chow vs purified diets?
: Chows can contain both animal and plant-based fat sources or in some cases just plant-based fat sources — Plant sources include ground wheat, ground corn, and soybean oil while animal based sources can include fish meal and porcine animal fat.
It's difficult to really know what the fatty acid profile would be from one lot to the next given most of these fat sources may vary in the concentration of fat or their fatty acid profile, which would be expected to alter the fatty acid profile in a given chow. In contrast, purified diets contain fat sources which are virtually all fat, from plant and/or animal based sources. Because fat sources in purified diets are from refined sources, the potential variation in the fatty acid profile will be less than that expected from sources in chows. Purified diets can contain different fat types, depending on the goal. Soybean oil and corn oil are commonly added as essential fatty acid sources.
Q: How does irradiation of a diet impact the gut microbiome?
: Based on literature searches, I haven't seen any studies where irradiated and non-irradiated diets were compared to see whether irradiation influences microbiome in rodent models.
The process of irradiation itself serves to reduce microbial load and any live bacteria, so irradiation can reduce any potential bacteria from the diet that may colonize in the gut. The DNA of any bacteria that was present in the diet will still be there and could be measured with modern measuring techniques such as 16s rRNA sequencing. Diet irradiation can cause some losses of certain vitamins including vitamin A and thiamin and can increase fat oxidation. However, I believe that such changes would not cause much effect on the gut microbiome in the cecum and colon given these nutrients would be absorbed in the small intestine.
Q: How does the form of diet (pelleted or liquid) affect the health/metabolism of mice?
: Typically, we prepare pelleted diets for mice, but we occasionally prepare powder versions which can be reconstituted in water (i.e. suspendible diets). We also prepare powder diets which would be fed as powders. I haven't come across data comparing health and metabolic parameters of mice fed a pelleted diet compared to a liquid diet.
However, I'm aware of some data contrasting diets that are fed as pellets vs. powder for metabolic measures and data have found that mice fed a powdered diet gain more weight (as body fat) compared to those fed pelleted diets and metabolism follows a similar pattern and affected metabolic parameters, such as increased plasma insulin and leptin (see Yan et al
Q: Because your literature searches have demonstrated that there has not been an improvement in the presentation of dietary information in Materials and Methods, are you planning on publishing a meta-analysis about what you've found so that we can use that to better inform investigators about how proper methods should be presented?
: A colleague and I at Research Diets are currently working on an opinion piece to submit to a journal that is focused on the widespread use of grain-based chows as a comparison diet to a purified high fat diet. We will use the data collected from my literature search as a means of pointing out that we've not been making progress.
Q: Discuss the proper formulation and use of the cafeteria-style diet.
: Cafeteria diets mimic products which are consumed by humans, such as cheese, salami, chocolate, and peanut butter, and there's definitely no such thing as a standard cafeteria diet. These are given to animals in varying types and levels, so it's not a definable way to feed an animal. I don't recommend the use of cafeteria-style diets. An alternative is to use a controlled, purified diet in order to affect metabolic parameters in desired fashion.
Q: Animals fed a high fat diet consume a higher amount of calories, but do not consume a higher volume of food. Can or should the high fat diet being supplemented with more vitamin and mineral mix?
: You're absolutely right that it is common to see that mice fed a high fat diet (HFD) consume less food, but perhaps more calories than those fed a low fat diet. For this reason, and because mice tend to eat more on a similar calorie basis than gram basis, we prepare the high fat purified diets with the same amount of vitamins and minerals as the matched low fat diet on a calorie basis. Therefore, mice that do tend to eat more calories on a high fat diet will also eat more vitamins and minerals. Animals eating the same number of calories (not grams) of each also eat the same amounts of vitamins and minerals.
Q: What is the translatability of this research to humans, who of course do not eat purified diets?
: That's a very good question. The same question could be asked of any animal experiment. Humans do eat the ingredients of purified diets. We eat starch, sugar and lard. Many of us consume some parts of purified diets. Purified diets use human-grade ingredients. To turn it around, do people eat diets that are similar to chow? Not really. With purified diets, we are able to understand how each ingredient affects biology in a very controlled way.
View the Taconic Biosciences' Webinar and Download the Presentation:
2. Bindels, L. B.; Munoz, R. R. S.; Gomes-Neto, J. C.; Mutemberezi, V.; Martínez, I.; Salazar, N.; Cody, E. A.; Quintero-Villegas, M. I.; Kittana, H.; Reyes-Gavilán, C. G. D. L.; Schmaltz, R. J.; Muccioli, G. G.; Walter, J.; Ramer-Tait, E. Microbiome 2017, 5 (1).
4. Desai, M. S.; Seekatz, M.; Koropatkin, N. M.; Kamada, N.; Hickey, C. A.; Wolter, M.; Pudlo, N. A.; Kitamoto, S.; Terrapon, N.; Muller, A.; Young, V. B.; Henrissat, B.; Wilmes, P.; Stappenbeck, T. S.; Núñez, G.; Martens, E. C. Cell 2016, 167 (5).
6. Goto, H.; Takemura, N.; Ogasawara, T.; Sasajima, N.; Watanabe, J.; Ito, H.; Morita, T.; Sonoyama, K. Journal of Nutrition 2010, 140 (12), 2121-2127.
9. Hrncir, T.; Stepankova, R.; Kozakova, H.; Hudcovic, T.; Tlaskalova-Hogenova, H. BMC Immunology 2008, 9 (1), 65.
12. Kübeck, R.; Bonet-Ripoll, C.; Hoffmann, C.; Walker, A.; Müller, V. M.; Schüppel, V. L.; Lagkouvardos, I.; Scholz, B.; Engel, K.-H.; Daniel, H.; Schmitt-Kopplin, P.; Haller, D.; Clavel, T.; Klingenspor, M. Molecular Metabolism 2016, 5 (12), 1162-1174.
13. Lin, R.-Y.; Choudhury, R. P.; Cai, W.; Lu, M.; Fallon, J. T.; Fisher, E. A.; Vlassara, H. Atherosclerosis 2003, 168 (2), 213-220.
16. Peppa, M.; He, C.; Hattori, M.; Mcevoy, R.; Zheng, F.; Vlassara, H. Diabetes 2003, 52 (6), 1441-1448.
17. Thigpen, J. E.; Setchell, K. D.; Padilla-Banks, E.; Haseman, J. K.; Saunders, H. E.; Caviness, G. F.; Kissling, G. E.; Grant, M. G.; Forsythe, D. B. Environmental Health Perspectives 2007, 115 (12), 1717-1726.