Introduction
Numerous reviews have been written to describe the specificities and advantages of the pig model in studies focusing on behavior, nutrition and neuroscience (Kornum and Knudsen, 2011, Roura et al., 2016, Val-Laillet, 2019). Besides the fact that the pig is an animal species of obvious interest in agriculture and human food production, it also has incomparable advantages when it comes to studying the consequences of the nutritional environments, diets or feed in itself on animal health, and allow its extrapolation to human health. Certain questions raised in animal production find an echo in the field of human nutrition and the problems inherent in loss of pleasure and appetite linked to changes in the environment or stressful situations, or in digestive and nutritional disorders induced by suboptimal/deleterious diets. The issue of excessive weight gain also exists, and although it affects pets more than production animals, it represents an undeniable parallel with the obesity pandemic described in human populations. This comes within the scope of the “One Health” concept.
This short paper is not a literature review but a quick panorama of the author’s teamwork in the context of neurobehavioral research in pigs. Here, instead of discussing about translational science, we will rather provide a quick overview on the exploratory methods that can be used in pigs to investigate the impact of food and diets on behavior and welfare, as well as the technologies that can be used to illustrate their brain correlates. These strategies can be combined with laboratory approaches aimed at investigating the gut microbiota, hormonal regulation, or immunity and inflammation processes to provide a picture of how the microbiota-gut-brain axis is impacted by nutrition. Of course, neurobehavioral approaches are often time-consuming and require complex paradigms and/or access to high-technology imaging platforms. Even though exploratory methods combining behavioural tests and in vivo brain imaging cannot be widely used in pig husbandry, they can be implemented in the scope of research and development strategies to provide valuable information about the effects of food and diet on animals’ welfare and health.
Investigating the effects of food and nutrition on pigs’ behavior
The main goal of nutrition is to provide calories and nutrients to feed the organism, maintain its basal metabolism, and ensure its growth and reproduction. But beyond these physiological and metabolic requirements necessary for survival, good health and perpetuation of species, food also contributes to the individuals’ hedonic needs and welfare. The nature of diet and food items available can modulate the animals’ behavior and activity, its cognition, the way it learns and memorizes things, the way it behaves with social partners and humans, the way it reacts and adapts to its environment. Since animal production environments are far from natural living conditions and mostly not optimal in terms of animal welfare and cognitive development, food can represent a vector to help the animals coping with their breeding conditions. On the contrary, when the nutritional environment is not adapted, it can have deleterious consequences on the animals’ development and welfare, contributing to cognitive, emotional, and health problems. To picture the impact of nutrition on pigs’ behavior, it is necessary to explore the different dimensions of animal behavior, and identify which factors deserve attention and possibly corrective interventions. Many behavioral tests have already been implemented in pigs to this purpose (Figure 1).
Eating behavior of course represents a major question of interest for animal productions. Feed must fulfill the animals’ nutritional requirements but also its hedonic expectations, which explains why so many feed producers work on feed palatability (the hedonic reward) to increase its appetence (its translation in terms of consumption). Eating behavior in itself depends on different cognitive components, including liking, wanting, and learning. In humans, it is quite easy to disentangle these dimensions using psychological tests and declarative questionnaires. In pigs and animals in general, this work can be achieved via specific experimental paradigms and apparatuses. Individual feed preferences (i.e. liking) can be explored via multiple-choice feed tests (e.g. using multi-compartment troughs, T- or Y-maze tests) and illustrated thanks to different criteria: e.g. visual/olfactory initial attraction conditioning the first choice, quantity consumed, time spent in contact with each feed, etc. We used these tests in juvenile pigs to compare different food flavors or feed additives (Clouard et al., 2012b), or to investigate the preferences for sweet and/or fatty foods in minipigs (Gautier et al., 2020), showing the existence of individual preferences towards specific flavours, tastes or textures. Meal microstructure (number and size of bites, interbites pauses, meal duration and intermeal periods, ingestion speed over time, etc.) can be investigated with computer-controlled automated feeders and provide additional data, notably to characterize satiation (short-term regulation signals) and satiety (longterm regulation signals) indicators (Val-Laillet et al., 2010, Boubaker et al., 2012, Clouard et al., 2014a), which is very interesting to evaluate the satiating effects of feed formulations. Food-rewarded operant conditioning is a potent strategy to investigate the “wanting” dimension of eating, since the quantity of work required to obtain a feed reward can be manipulated. In the progressive ratio task, the subject is required to make an increasing number of operant responses (e.g. to push a button or lever) for each successive reward, which is an indicator of food motivation (Ochoa et al., 2014). Different feed formulations can be compared using this approach to document the animals’ willingness to work for them.
In addition but also in relation to eating behavior, the emotional, cognitive and social dimensions of behavior can be explored via dedicated tests. To document the emotional reactivity of pigs, classical procedures such as the openfield test are widely used, where behavioral items are recorded in animals isolated (or not) in a novel environment, usually a square arena divided into virtual zones (Donald et al., 2011, Val-Laillet et al., 2013, Menneson et al., 2019). Locomotor activity, vocalisations, exploratory behavior, miction/defecation, escape attempts are part of the usual behavioral repertoire studied. Similar arenas can be used to assess the emotional reactivity to known or unknown humans, familiar or novel objects, or sudden stimuli (e.g. noise, light, falling object, etc.). All these tests are potent methodological tools to investigate the animals’ emotional reactivity to familiarity, novelty, suddenness, and impredictible environments. They are valuable to describe the individual behavioral traits (i.e. habitual behavioral patterns stable across situations and time and providing a foundation for predicting behavior). Conflicts of motivation are sometimes very interesting to look at in order to identify individual valuation of ressources and motivation priorities in pigs. This is the case for the novelty-suppressed feeding test (NSF): the animal is facing a trough containing a palatable feed but a novel and possibly frightening object stands alongside or above it (Menneson et al., 2019). The pig consequently has to fight against its fear to fulfill its motivation to eat. Contention tests also rely on such a conflict of motivation (i.e. “freeze, fight or flight”). In piglets, tonic immobility is correlated with emotional reactivity and stress/resignation susceptibility, similar to contention tests in the adult age (e.g. in a hammoc or Pavlov stand) (Menneson et al., 2019).
Finally, other kinds of tests can be designed to investigate the cognitive abilities of pigs, including social cognition (e.g. social discrimination and preferences using T- or Y-maze tests) (Val-Laillet et al., 2013), but also spatial learning and memorization for example (a paragon of hippocampal-dependent tasks). The holeboard test initially designed for rodents was successfully implemented in pig models, where the animal has to memorize and retrieve palatable food rewards hidden in a testing arena, independently from olfactory cues and on the basis of spatial cues only (Gieling et al., 2012, Gautier et al., 2018, Gautier et al., 2020). Using this paradigm, we demonstrated the deleterious effects of perinatal or adult exposure to high-sugar and high-fat diets, independently of weight gain. Every behavioural test has pros and cons as well as experimental constraints. In the case of the holeboard test, results are influenced by the cognitive abilities of the pig, i.e. its learning and memory capacities, but also by its motivation for the food rewards. When interpreting such tests, it is important to identify which behavioural dimensions are involved in the task, and how they can influence its outcomes. To disentangle these interconnections between different behavioral dimensions, it is possible to use different or complementary tests. For spatial learning and memorization for example, the holeboard task can be replaced by or completed with other different spatial tasks that are not reinforced by food reward, i.e. maze test where the success is rewarded by meeting with social partners or provision of toys instead of food (Kornum and Knudsen, 2011, Gautier et al., 2018).
Figure 1. Examples of behavioral tests implemented in pig models (conventional or miniature pigs) to investigate food preferences, motivation, and meal structure, emotional reactivity, responses to novelty or suddenness, spatial learning and memory, or social cognition.
Investigating the microbiota-gut-brain axis in pigs
Our research department at Inrae is renown for its expertise in functional brain imaging in pig models. For more than a decade now, we have implemented minimally invasive in vivo brain imaging for pigs in the context of nutrition research to investigate the impact of food and diet on brain metabolism as well as on the neurocognitive processes underlying eating behavior, with significant connections with welfare and health. This work was quite recently summarized in a review paper (Val-Laillet, 2019). Figure 2 illustrates the kind of food stimulation or treatments investigated in conventional or miniature pigs, as well as the brain areas for which differences in terms of metabolism or functional responses were observed in juvenile or adult pigs. Most of this research focused on the motivational and hedonic dimensions of eating, highlighting how the different organoleptic, nutritional and physical dimensions of food, and where it is perceived (e.g. oral vs. visceral), can influence liking, wanting, and learning.
Figure 2. A) Brain responses to food stimuli, diets, and nutrition-related interventions were explored in pig models using different brain imaging modalities such as nuclear imaging (single-photo computed tomography and positron emission tomography, SPECT and PET respectively) or functional magnetic resonance imaging (fMRI). Stimulations and treatments encompassed: a) visual stimuli, b) complex odors/flavors, c) basic tastes, d) duodenal sugar infusion, e) portal sugar infusion, f) gastric distension, g) butyrate supplementation, h) vagus nerve stimulation, but also chronic Western diets and surgical interventions such as Roux-en-Y gastric bypass. In the context of these studies, various neuronal networks were described including 1) visual occipital cortex, 2) brainstem, 3) olfactory bulb, 4) prepyriform cortex, 5) insular cortex, 6) hypothalamus, 7) prefrontal cortex, 8) striatum, 9) cingulate cortex, 10) amygdala, and 11) hippocampal areas. Excerpt from (Val-Laillet, 2019). B) Anaesthetized pig being equipped with nasal and lingual catheters to investigate its olfactogustatory perception during a brain imaging session. C) Example of sagittal (S.1 and S.2) and coronal (C.1 and C.2) brain images and 3D-reconstruction obtained in the pig model, with differences of brain metabolism observed in the nucleus accumbens (NAc), putamen (Put), and anterior prefrontal cortex (APFC). Excerpt from (Gautier et al., 2018).
Most of these studies were performed in the context of obesity and chronic exposure to Western diets in young or adult minipigs, in order to document the neurocognitive anomalies emerging from deleterious nutritional environments (Val-Laillet et al., 2011, Clouard et al., 2016a, Clouard et al., 2016b, Val-Laillet et al., 2017, Gautier et al., 2018, Gautier et al., 2020).
Comparisons between specific macronutrients in the diets were also performed, for example between fructose, glucose and starch (Ochoa et al., 2016a, Ochoa et al., 2016b), or between saturated fat, omega-6 and omega-3 fatty acids (Malbert et al., 2021), demonstrating that the quantity but also the quality of food is determinant for the behavioral and brain developments. Some nutrients or diets favor in the pig model as in humans the onset of nutritional pathologies, including obesity and metabolic diseases, but also eating and mood disorders. Another research thematic concerned the emergence of food preferences and aversions in juvenile commercial pigs, related to amino acids ratio, basic tastes (e.g. sweet), or to complex flavors as those used as/in feed additives (Gaultier et al., 2011, Clouard et al., 2012a, Gloaguen et al., 2013, Clouard et al., 2014a, Clouard et al., 2014b, Coquery et al., 2018). A third topic deals with the use of functional food ingredients, for which a focus will be provided in the last section of this paper. Such studies are valuable to identify the food properties that modulate food intake and eating behavior, and the use of functional brain imaging is a fantastic tool to understand the neural mechanisms underlying individual choices and motivation to eat.
Amongst the different brain imaging techniques available, two main approaches were used in pigs: nuclear brain imaging, including single photon computed tomography (SPECT) and positron emission tomography (PET), as well as functional magnetic resonance imaging (fMRI). Nuclear brain imaging techniques require the injection of a radiolabelled molecule of interest in the blood system of the subject. In our studies, we used SPECT to explore the blood brain perfusion and PET to explore the brain glucose metabolism. Both techniques are highly correlated but PET provides a better spatial resolution with sometimes a more interesting biological rationale, i.e. the consumption of glucose by the brain cells. They can be used to investigate basal brain metabolism or the regional brain responses to specific treatments or stimulations. But investigating different stimuli requires the repetition of imaging sessions and exposes the animals to radiation, which requires controlled housing and handling systems in respect of the nuclear authority regulations. The use of fMRI does not require ionizing radiation and relies on the BOLD (blood-oxygen-level-dependent) hemodynamic response, which changes according to the variations of oxyhemoglobin and deoxyhemoglobin in the brain tissues (i.e. oxygen consumption). Resting-state fMRI can be performed to investigate the resting brain functions and connectivity, while BOLD responses can be recorded in response to a wide variety of stimuli, which can be repeated many times along the same imaging session. All these techniques consequently have specific advantages and constraints, and the scientific question usually orientates the best choice of methodology.
When investigating the effects of food and diet on the individuals’ welfare and health, behavioral and brain measures are usually not sufficient, especially to understand the complex mechanisms underlying this relationship. Many of our research projects consequently combine these neurobehavioral explorations with measurements at the microbiota and gut levels to understand how the nutritional environment modulates the gut-brain communication and possibly the observed neurocognitive changes. The idea is usually to produce a systemic view of the metabolic, physiological, and neurocognitive effects of nutrition. Amongst the biological criteria that are recorded are gut hormones and neuropeptides involved in the regulation of homeostasis and food intake, but also biomarkers of immunity, inflammation, nervous communication, or intestinal barrier functions. The joint application of metabolomics and transcriptomics possesses the high efficiency of identifying key metabolic pathways and functional genes modulated by nutrition. It can be used also to investigate the metabolites produced by the gut microbiota, while metagenomics is used to represent its whole genome. Canonical correspondence analyses can be performed to delineate the complex relationships between sets of variables (e.g. environmental and nutritional factors, physiological and metabolic parameters from the host, gut microbiota composition and activity, etc.) by searching for latent (hidden) gradients that associate these sets of variables.
Such a systemic exploration strategy has already been implemented in adult minipigs to document the consequences of chronic exposure to Western diets (Val-Laillet et al., 2017, Gautier et al., 2018, Gautier et al., 2020), but also in commercial pigs to describe the microbiota-gut-brain axis and behavioral anomalies induced by chronic psychosocial stress for example (Menneson et al., 2019). In minipigs, we demonstrated that the chronic consumption of Western diet induced addiction-like behavioral and brain responses to sucrose, while obesity triggered anxiety-like, snacking-type behaviors, and negative metabolic outcomes. Overall, body weight gain and loss both modulated the corticostriatal neurocognitive responses to food. In commercial pigs, we showed that social isolation and stressing unpredictable environment induced systemic deleterious consequences at the microbiota-gut-brain and behavior levels (see next section). Further analyses will be performed to relate all these effects with specific gut microbiota populations and metabolites, with the aim to design potential interventional treatments targeting the gut microbiota. For example, in a study performed in juvenile commercial pigs, we demonstrated that supplementation with butyrate, which is a microbiota metabolite produced from the fermentation of dietary fibers, impacted brain activity in regions involved in cognition and pleasure (Val-Laillet et al., 2018b). A better understanding of the microbiome-host crosstalk via the microbiota-gut-brain axis cannot only shed light on healthy and efficient pig production but also promote our knowledge on this fascinating research topic (Wang et al., 2020).
Focus: The use of functional food ingredients to improve food pleasure and adaptation to stress
Functional food ingredients or sensory feed additives are increasingly used in animal nutrition to stimulate appetite, facilitate feed transitions, or improve animal welfare. Stressful environments or events are known to have a negative impact on well-being of course, but also on eating behavior, which sometimes leads to appetite loss and anorexia, which has been well described in pigs (Clouard et al., 2012a), or on the contrary to food cravings and hyperphagia. In a set of studies performed in the juvenile pig, we demonstrated positive effects of sensory feed additives or functional ingredients based on natural plant extracts and aromatic molecules on eating behavior, with increased feed preference and intake after a feed transition for example (Clouard et al., 2012b, Clouard and Val-Laillet, 2014). We also showed that sensory feed additives provided to sows during gestation/lactation and/or to piglets during the post-weaning period differentially impacted animals’ growth and feed intake, which highlighted the existence of nutritional programming and/or sensory conditioning during the perinatal period (Val-Laillet et al., 2018a).
To interpret these positive behavioral outcomes in terms of eating behavior, we explored the brain responses to one of these functional food ingredients in naïve or familiarized juvenile pigs.
We showed that the perception of the feed additive in familiarized individuals induced different brain responses in regions involved in reward anticipation and/or perception processes than the familiar control feed flavor in naive animals (Val-Laillet et al., 2016). Another study showed that these hedonic brain responses are dependent to the ingredient used and to its concentration (Coquery et al., 2019). In addition to positive effects on feed palatability, we hypothesized that some functional ingredients might help the animals to cope with degraded or stressful environments. To test this hypothesis, we conceived a chronic psychosocial stress model in the pig, combining social isolation with a poor and unpredictable environment (Menneson et al., 2019). Animals were characterized by a resignation behavior, altered hippocampal neuroplasticity, lower brain responses to novel food odors, higher insulin resistance and lower gut microbiota fermentation activity, revealing systemic stress-induced anomalies at the microbiota-gut-brain axis and behavioral levels (Menneson et al., 2019).
Using a spice functional food ingredient, we demonstrated increased expression of 5-HT1AR (serotonin receptor, a neurotransmitter involved in the reward system and mood regulation) and BDNF (a neuroplasticity marker) in the hippocampus and prefrontal cortex, respectively, as well as slight anxiolytic-like effects and significant modifications of the blood perfusion (via SPECT) in several brain areas involved in the regulation of emotions and cognition (Menneson et al., 2020b). In another study, we observed that the BOLD fMRI brain responses induced by a pharmacologically-induced acute stress were alleviated in animals previously supplemented with a food ingredient containing Citrus sinensis extracts (Menneson et al., 2020a). These results demonstrate that functional food ingredients can be used in pigs to modulate different behavioral dimensions and to correlate these effects with functional brain responses related to food reward, positive emotions and stress adaptation. This work opens the way to innovative nutritional strategies with the aim to improve the pigs’ welfare, cognitive development, and health.
Conclusions and perspectives
This overview provided a quick panorama of some exploratory approaches designed to investigate the effects of nutrition on the microbiota-gut-brain axis, behavior, welfare and health in pig models. Of course, most of these technologies and paradigms cannot be commonly implemented on farms because they are time-consuming, require costly machines and laboratory facilities, as well as dedicated staff. Though, all these approaches represent powerful tools in the context of research and development (R&D) strategies, to conceive better diets and innovative feed ingredients contributing to the “One Health” concept and objectives. Besides these systemic R&D approaches, complementary “easy-to-implement” methods must be developed and such aims are part of the advances in precision nutrition and modern phenotyping strategies in animal production. Recent review papers also illustrate the assets provided by mechanistic modeling and data-driven models for modern animal production systems (Ellis et al., 2020). Automated feeders, biological and activity sensors, environmental sensors, cameras and other image analysis tools can provide complex data streams from which it is possible to extract descriptive and predictive information in order to adapt nutrition and breeding conditions. Correlating these data with mechanistic information provided by the approaches described in this paper shall provide important interpretative keys to understand how animals perceive and adapt to their living conditions, and how it affects their behavior, welfare, and health.
Presented at the 2021 Animal Nutrition Conference of Canada. For information on the next edition, click here. 
