In this Interview Neuroscientist Steffen van Heijningen, shares with News-Medical the insights from the research on mice housing conditions.
Could you start by introducing yourself and saying a bit about your background?
My name is Steffen van Heijningen. I studied at the University of Groningen from 2010 to 2015, first as an undergraduate and then as a Master’s student. From 2015 onwards, I worked with Nutricia Research in the Netherlands on a Ph.D. project about metabolic programming. My thesis is entitled ‘The Driving Force of Metabolic Programming.’
In July, I started working at Noldus as a Marketing Communication Specialist, contributing to the neuroscience content that we communicate.
Please can you elaborate on your research area and your work to investigate the causes of obesity?
Obesity is a global epidemic, with much research and media coverage. Many researchers are working to investigate whether obesity is preventable, whether there is a reason that some people are predisposed to obesity and whether we can prevent this, for example, via early life nutritional interventions.
This research is based on environmental factors, the first of which take place in utero. Conditions in utero are influenced by the health and diet of the mother, so we need to provide the most optimal building blocks for healthy development and later life.
After birth, we see nutrient transfer through breastfeeding. Breastfeeding is also influenced by the diet of the mother, but we also see a nutritional difference in bottle feeding versus breastfeeding after birth. This nutritional environment in the first 1,000 days of life creates a window of opportunity for what is called ‘early life nutritional programming.’
This research is focused on using this window of opportunity to study the transition from early to adult life. The next stage of my research, in particular, involves searching for an appropriate model to predict this predisposition to obesity and then afterward prevent this with dietary intervention.
We already have a very clear obesity model in terms of the western-style diet, but we can also look at individual housing as an obesity model in comparison with a small litter rearing approach.
Why have mice historically been such a common research subject for this type of research?
Mice are widely used as a model organism in medical research because of their high similarity with the human genome. They are easy to take care of, you can breed them in large numbers and they are relatively cheap.
The most commonly used strain of mice - sometimes called the wild-type strain – is the C57 black 6 mouse, the very recognizable black mouse we see often.
In metabolic studies in general, this is a very established model. These mice are obesity-prone, allowing us to build upon earlier studies that were done in this context. During the first three weeks of the mice’ life, a postnatal leptin surge takes place that plays a role in the generation of hypothalamic feeding circuits and fibers.
Without this leptin surge, we see a lack of formation of these fibers. Interestingly, in humans, this does not occur. Instead, we see elevated leptin levels at the end of the prenatal period, which then declines quickly after birth.
Knowledge of this difference allows us to develop a postnatal model in mice to study this hypothalamic feeding circuitry, which relates to a prenatal period in humans. It creates an interesting window of opportunity for us to perform interventions and research into these feeding circuits.
What are the issues surrounding individual mouse housing versus social mouse housing for this type of research?
In metabolic studies, we often run into a problem with C57 black 6 mice in that we have to house them individually. We need to know their food intake, water intake and energy expenditure, and we want to collect feces for practical purposes. If we want to automate specific measurements, we need to be able to track these individuals in the cage.
Due to this, metabolic studies have often opted for the individual housing of mice. However, this approach can actually influence the outcome past the point of it being a relevant model.
When doing obesity research, I saw that the model itself was affecting the outcome in terms of intrinsic behavior and the metabolic and physiological parameters.
Social isolation is considered a stimulus-poor condition that can lead to boredom and increase reward sensitivity. This social isolation changes the mice’ energy use because individual animals do not have to expend energy on social behaviors such as interacting with cage mates, competing for food or generally engaging in feeding behavior.
Deprivation of social cues, for example, social facilitation, means there is less feeding competition, and the social status is entirely different.
Image Credit: Noldus
What impact did these factors have on the development of obesity models?
It is already known also that sustained social deprivation may provoke chronic emotional distress, affecting neuroendocrine and immune axes, which can also lead to alterations in body weight and fat deposition.
Our goal was to develop an obesity model with implications downstream, and it became evident that individually housing the mice was impacting a lot of factors that influence this metabolic profile, down to the behavioral level. These findings were led by Dr. Lidewij Schipper, my supervisor at Nutricia Research, with whom I worked very closely on this project.
We translated one of my thesis’s first chapters, which we published in PLOS ONE. It was titled ‘Individual Housing of Male Black 6 Mice After Weaning Impairs Growth and Predisposes for Obesity.’
As part of the study, at the point of weaning (postnatal day 21), the animals were transferred from their nest into either individual housing – an isolated state, or social housing - two mice in a cage. We saw that the social animals gained more weight towards postnatal day 42, which we used as a cutoff point for adulthood.
We transitioned the animals onto an adult diet at that stage and considered them as young adults. We observed a definite difference between the socially and individually housed mice in bodyweight accumulation.
During adulthood (postnatal day 42 to 126), we exposed the mice to a western-style diet, but irrespective of this diet, the individual animals gained more weight from this point on. We hypothesized that in earlier life and adolescence, individual housing impairs growth, but try to catch up later in life and add more weight than social animals.
We collated information on body composition, including lean body mass, fat mass, and femur length. Femur length is a robust parameter for overall development, lean mass and growth of the mouse.
In terms of fat mass, we saw that social animals weighed more than individual animals. Their fat mass however being significantly less. This led to the difference in body weight, but primarily a different body composition which leaned towards an obesogenic phenotype in the individually housed animals. This difference was still present in adulthood and possibly driven by femur length difference.
This finding proved that body composition is switched around in these two housing conditions. We hypothesized that thermogenesis had an effect here. UCP-1 (Uncoupling Protein 1) is responsible for heat generation in brown adipose tissue. This is considered a healthy fat, with babies having a lot of brown adipose tissue, for example.
It was possible that the individual animals were less capable of keeping each other warm through huddling behavior. Looking at adolescent versus adult mice, we saw higher activity in the brown adipose tissue in individually housed animals, indicating that this theory was robust.
Looking at the activity of UCP-1 in the white adipose tissue, also called ‘browning’ of the white adipose tissue, we saw that this effect was most significant in adulthood, where individual animals try to compensate for that activity due to the lack of social thermal regulation.
Overall, we found that individual housing impairs growth during adolescence but leads to increased body weight gain during adulthood regardless of diet, leading to an obesity-like phenotype.
Were these models also applicable to early life diets? How did you evaluate this?
Our next study saw us use the individual versus social model again, but we also exposed the mice to early life diets to look for validation of this phenotype in terms of obesity. This time we added behavior to the equation because we wanted to see if their behavior was also changing because of the housing conditions.
We measured distance moved in an open field because we know from the literature that individually housed animals become hyperactive in an open space. We also looked at an elevated plus maze where we studied anxiety-like behaviors, discovering that adolescent and adult individually housed animals had a higher overall ratio.
This was interesting because we had hypothesized that individually housed animals would be socially deprived, prompting them to show higher stress levels, but the elevated plus maze study indicated that they had reduced anxiety-like behavior.
We wondered why we were seeing this different pattern, so we explored this further by investigating how social housing affected specific behavioral outcomes. Using the elevated plus maze, we separated the socially housed animals into one group that was tested first and one group that was tested second.
We found that socially housed animals were influenced by the fact that they were separated from their cage mate. This highlighted that individual housing brings about behavioral consequences because of the deprivation of social interaction, but if we expose social animals used to being housed together to an environment where they were separated, this can also induce stress.
This was also affected by the mice’ hierarchal status within the cage, whether they were the dominant animal or the subordinate animal. We tested this using a tube test, and this was important in refining our methods and how we perform these behavioral tests in terms of habituation.
In terms of behavior, the most important thing we took away from this research is that we cannot compare individually housed mice versus socially housed mice due to these large variations in behavioral response. Testing order affects this as well as the hierarchy in social cages. It is important to keep this in mind when designing and performing a study.
Image Credit: unoL/Shutterstock
Where does this research fit within the wider literature? Have there been other studies into the impact of individual versus social housing on studies involving mice?
To my surprise, it was hard to find detailed articles looking into factors such as testing order and how this can obscure results, but it is nonetheless extremely important to consider the social lives of mice. Mice are a social species, and the literature states that they should be group-housed as long as the groups are stable.
This provides multiple benefits on an evolutionary level, including improved chances of survival of the individual, lowering predation risk and reducing energy costs via social thermoregulation.
In terms of welfare, people lean towards social housing, but this is where the dilemma arises. The Guide for the Care of Laboratory Animals from the National Research Council states that housing should be based on the natural living conditions of a species and that the home cage should be an environment that provides for the animal’s physiological and ethological needs.
The dilemma arises in terms of being able to effectively track the mice’s food and water intake and energy expenditure. This is explained in an article titled ‘To Group or Not to Group?’ This paper discusses how male mice live in groups, but they can become very territorial and aggression towards each other can be a serious problem that affects the welfare of these animals.
This can negatively impact animals in terms of stress, particularly in terms of social defeat, for example. Social defeat is an extremely stressful model which can happen in a home cage situation where multiple male mice are aggressing towards each other, and there is one mouse that keeps losing. This is the social defeat paradigm.
In this paper, they also state individual housing may not be an appropriate solution given the welfare implications associated with no social contact.
Given the various studies and findings on this topic, what would you say to anyone designing a study and looking for the optimal solution around housing mice?
The most optimal solution is based on a lot of factors that influence behavior. Handling, for example, tail or tube handling, can affect behavior. I tested both methods in my research to evaluate their impact on the mice’s stress levels.
Habituation is important, and we can consider separation from the cagemate before subjecting them to a test in which they have to perform on an individual level. The researcher itself can influence these factors by trying to control the variables and trying to create a level playing field as much as possible.
We know that social isolation induces anxiety-like behavior. In terms of individual housing, we wanted to see if these animals better coped with this isolation compared to socially housed animals, who were then separated before testing.
Another study used an isolated group and a control group to evaluate long-term isolation, finding that this could lead to increased anxiety-like behavior.
On a neuro level, we see social isolation induces anxiety-like behavior but also decreases BDNF levels, greatly affecting neuroplasticity and cognition in these animals.
The decision to house mice individually or socially is complex. Social housing is more naturally relevant, and individual housing does cause a depressive-like phenotype, leading to body weight and fat increase. Social housing creates some logistical challenges in terms of identification of the mouse, hierarchy and the need to account for testing order effects.
Social housing should be the housing method of choice. However, individual housing is inevitable in some cases. For example, in studies with cannulas, implants are problematic where there may be fighting behavior. This is also an issue in metabolic studies where we need to know the exact energy expenditure of one mouse instead of an entire cage.
My approach to this is to employ pragmatism. There is no perfect approach to housing mice, so this should be tailored to what you need for your research.
It is advisable to use concepts from previous studies and to learn from previous experience, taking those into account when developing your model. Another pragmatic point is to consider what you actually need from your data.
For example, I wanted to monitor the food intake of my animals constantly, but I had to decide if this was genuinely needed from individual mice or if I could just use the cages as a statistical unit. The latter is what I did in my previous study.
I was also privileged to be able to use a high N because of how I designed my study, and I was only working with dietary interventions so I could have a high number of animals and retain this statistical power when only using cages as a statistical unit.
Design a study by asking yourself, “What do I actually need? Can I acquire this data in a different form by analyzing or grouping this data to get the information needed to accurately test my hypothesis?”
Having done this, consider factors such as testing order, counterbalancing, habituating, testing during rest and active phases, and make sure you check everything beforehand to ensure no errors can be made.
How can Noldus help researchers in a similar situation or facing the same dilemma?
If I were to undertake this research project again, I would set up my research to include continuous monitoring with the Noldus PhenoTyper. This would have given me a full metabolic, behavioral and activity profile on the effects of social versus individual housing.
I would have been able to accurately monitor things like huddling behavior, social behavior, eating, drinking and activity patterns. Currently, I have small snippets of inner development, but these do not give me enough insight into how these metabolic profiles have been built up.
In terms of ‘snippets,’ I am referring to specific behavioral tests where I measured a few days’ energy expenditures, for example, using an indirect calorimetry setup. Using the Noldus PhenoTyper, I could have increased the timeframes in which I measured these variables to gather a large amount of information using a more powerful setup.
About Steffen van Heijningen
Born and raised on the Island of Curaçao, Steffen van Heijningen always aspired to go into science. At the age of 17 he moved to the Netherlands, where he studied Life Science & Technology, Biology and Behavioral Neuroscience at the University of Groningen. On the verge of submitting his doctoral thesis ‘The driving forces of Metabolic programming’, and with more than 6 years of hands-on experience with laboratory animals, Steffen is a true expert in his field and has faced many challenges during the design of his animal studies. Now stepping outside of academia, Steffen joined Noldus in the summer of 2021 to contribute to the Neuroscience pool of knowledge within the company, working remotely from his home town on Curaçao. Outside of being a scientist and working for Noldus, Steffen enjoys spending quality time with his two kids and wife, primarily being on the water and enjoying the ocean
About Noldus Information Technology
Noldus Information Technology was established in 1989 by Lucas Noldus, founder and CEO of the company. With a Ph.D. in animal behavior from Wageningen University, he developed the company’s first software tool during his research in entomology. Noldus has strived to advance behavioral research ever since, evolving into a company that provides integrated systems including software, hardware, and services.
We now offer a wide range of solutions for research in animal and human domains, including biology, psychology, marketing, human factors, and healthcare. We work with leading suppliers and develop innovative, state-of-the art products. We also offer excellent technical support and customer care. As a result, our systems have found their way into more than 10,200 universities, research institutes, and companies in almost 100 countries.
The success of our company is determined to a large extent by the enthusiasm and creativity of our employees. We encourage each other to think outside the box, which leads to unique products and services for our customers. And we are always on the lookout for new talent!