This presentation focuses on the research cost-benefit aspect of enrichment of housing conditions for laboratory rats and mice. The choice of this subject emerged because the session organizers requested a presentation on "laboratory animal housing enrichment" and included the following in their letter of charge:
The workshop will . . . focus . . . on identifying gaps in the current knowledge in order to encourage future research endeavors, assessing potential financial and outcome costs of unscientifically-based regulations on facilities and research, and determining possible negative impacts of arbitrary regulations on animal welfare.
The basic view put forth herein is that caution is warranted in the adoption of environmental enrichment procedures, because they may complicate interpretation of research results.
We begin by briefly discussing the history of what has come to be called "enriched housing." The first description of an effect of enhanced living conditions on behavior as an indication of altered brain function was the work of Hebb (1949), who compared rats that he reared as "pets" in his home with counterparts reared under normal laboratory conditions (an experiment unlikely to be repeated, given current regulations regarding research animal housing!). Hebb reported that the home-reared rats were superior to the laboratory rats on complex problem-solving tasks and that they continued to move ahead as they were tested on successive
178 SCIENCE-BASED GUIDELINES FOR LABORATORY ANIMAL CARE
tasks. Subsequently, students of Hebb or others inspired by him repeated the basic finding (in the laboratory), that a more stimulating rearing environment enhanced performance on complex learning tasks (e.g., Bingham and Griffiths 1952; Forgays and Read 1962).
Subsequently, Krech and colleagues (1960) reported effects of a similar rat housing environment, for which they adopted the term "enriched," on measures of the activities of enzymes involved in metabolism relating to cholinergic synaptic transmission. This program led to the discovery that some regions of the cerebral cortex were actually heavier and thicker in the "enriched condition" (EC) rats compared with "impoverished condition" (IC) rats kept in barren individual cages (Diamond and others 1966). Research stimulated by theirs triggered the first report of altered dendritic branching (Holloway 1966), although that paper used methods that were inadequate to quantify dendritic branching.
Research on the details of changes induced by such experiences, and that of a number of others, was inspired by the work of Rosenzweig and colleagues. For purposes of this illustration, the work of the Greenough laboratory is selectively emphasized here. An early replication of the Holloway (1966) study using quantitative methods (Volkmar and Greenough 1972) indicated that the dendritic branching of neurons in the rat visual cortex was altered in EC versus IC rats and that "social condition" rats housed in pairs in standard laboratory cages (SC) were intermediate, often differing statistically from both EC and IC rats. This latter result suggests that the more minimalist rodent enrichment procedures such as social housing, which are common in European laboratories and becoming more so in laboratories in the United States now, may actually bring about subtle but detectable changes in the brain. The enriched environment used by the Greenough laboratory, although likely falling short of Hebb's home is, by contrast, a very complex arrangement of objects for play and exploration as Figure 1 indicates.
The effects of these different environments are not restricted to the brain and to the behavior it enables. Significant peripheral somatic differences exist between rats housed in EC and those in IC, which could interact with various sorts of treatments or affect responses to edible reinforcements (Black and others 1989). These differences include, in rats in our laboratory, (1) greater body weight in IC than in EC rats, accompanied by (2) greater food consumption in the ICs, (3) more rapid maturation of the long bones in IC versus EC rats, (4) sometimes greater adrenal to body weight ratios in EC versus IC rats, (5) a higher kidney to body weight ratio in the EC group, and (6) a lower thymus to body weight ratio in the ECs (with no indication of diminished EC immune competence; Black and others 1989). The fact that the organ weight ratios differ in both directions (EC > IC and IC > EC) suggests that they do not reflect merely
ENVIRONMENTAL ENRICHMENT ISSUES 179
ENVIRONMENTAL ENRICHMENT ISSUES 179
the relatively lower body weights of the ECs. The goal here is not to try to explain why these differences occur, but rather to illustrate that a variety of experimental measurements could be affected by differences of this sort. Hence, on the basis of peripheral measures alone, utilizing rats exposed to enriched housing in other experiments could generate confounded results, and certainly switching from nonenriched to enriched animals could generate changes in experimental outcomes. It should also be noted that male and female rats can differ in their responses to enriched environments (e.g., Juraska 1991, 1998). Thus, basic somatic physiological processes are affected by rearing environment complexity, which can affect research outcomes if those processes are or affect variables of interest, and caution is warranted in introducing novel degrees of environment complexity or "enrichment" into ongoing research paradigms.
The brain effects of EC are even more profound. Neurons and their synapses, vasculature, and the two most prominent types of glial cells all are dramatically affected by exposure to an enriched environment. In visual cortex the number of synapses per neuron is 20 to 25% greater in EC rats compared with those in IC, with rats socially housed in cages typically little different from individually housed rats (Turner and
180 SCIENCE-BASED GUIDELINES FOR LABORATORY ANIMAL CARE
Greenough 1985). This effect is supported by equally substantial increases in the size of the dendritic fields of neurons (Volkmar and Greenough 1972). Synapse morphology and architecture are also different in EC versus IC rats (Jones and others 1997; West and Greenough 1972).
The volume of capillary per neuron is similarly selectively increased in EC rats (Black and others 1987), presumably in part to "power" the increased numbers of synapses, a substantial fraction of which are closely associated with mitochondria on the presynaptic side. Astrocytes, which serve to optimize many metabolic functions of neurons, and can be identified by the presence of their characteristic glial fibrillary acidic protein, are increased in both size and number in EC rats (Sirevaag and Greenough 1991). Moreover, synapses in EC rats are more completely covered by fine astrocytic processes than in IC rats (Jones and Greenough 1996). The other macroglial cell type, the oligodendrocyte that gives rise to the axonal myelination that enhances the speed of conduction of nerve impulses, is also affected by environment enrichment: EC rats have more myelinated axons in the corpus callosum than IC rats (Juraska and Kopcik 1988). All of the foregoing findings have been demonstrated in visual cortex (or connecting callosum), and many effects have also been demonstrated in other brain regions. Taken as a whole, these results indicate that the properties of most cell types and the ways in which they relate to each other in the brain may be altered by the housing environment.
Most of these effects also occur in rats put into enriched environments for the first time as adults. Most rats used in research are purchased from suppliers as young adults, typically shortly after they reach the point of sexual maturity, and are used as quickly as possible after they have become accommodated to their new surroundings, in an effort to minimize cost. If the rats were made to accommodate to an enriched laboratory environment, their bodies and brains might be in a state of relative physiological and structural turbulence at just the time they were expected to be ready to participate in experiments. Clearly research on or involving these variables will be affected, and research on other interacting variables might also be affected in unpredictable ways.
The mechanisms mediating these effects are largely unknown. Neuro-trophic factors such as "brain-derived neurotrophic factor" are known to be altered by environmental variations such as enrichment and exercise (Klintsova and others, submitted; Oliff and others 1998; A.Y. Klintsova, E. Dickson, R. Yoshida, and W.T. Greenough, manuscript in preparation), and these factors may well be a part of the process that generates the responses in brain physiology and structure in response to altered environmental conditions. This further complicates the stability of the background against which experimental effects are to be measured.
A novel finding that can be discussed only after it is accepted for
publication is Richard Smeyne's finding that the drug MPTP, which kills catecholamine neurons in the substantia nigra in humans and in conventionally housed rats, does not result in similar neural damage when rats are housed in an enriched environment. Although this finding, of course, suggests important therapeutic directions, it also illustrates the complications that might be induced in a well-developed paradigm by the sudden insertion of enriched housing procedures.
Numerous institutions, including my own, have recommended or even mandated enrichment procedures for laboratory rodents recently. In the case of my university, where group housing or the insertion of novel objects into the cage is mandated, a nonscientific poll of principal investigators using rodents found that only I was aware of this policy, despite the fact that those procedures were being applied to their animals at the time I asked them the question. Certainly they did not seem to have been asked whether they thought these procedures might interfere with their research, despite a clear policy guideline with regard to the following statement: "Investigators who must singly cage animals and feel that enrichment materials may confound their research objectives must provide justification." Enrichment appears to have been accepted as a "good thing," with little consideration of its possible effects on experimental outcomes. Taken literally, this University of Illinois policy might make it difficult to determine effects of enrichment that one did not know to exist. There is also tacit acceptance that group housing is superior to individual housing despite data that call into serious question whether this is true (Bartolomucci and others 2003).
Several presentations at the current meeting seemed similarly to espouse such a view. Perhaps most disconcerting is the arbitrary assumption that enrichment is better for the animals, with little data to support this assumption beyond the fact that the animals attend to enrichment objects and appear to play more vigorously when such objects are present. It appears in this case, as in the case of several other presentations at this workshop, that the animals' preferences are being allowed to drive, if not dictate, the issue of what constitutes enrichment. In this regard, it is of value to note that animals' preferences may not be the ideal guideline to what is of most value to them. In earlier research on addiction, which would probably not be permitted today, it was found that rats and monkeys given unrestricted or nearly unrestricted access to drugs of abuse (cocaine, amphetamine, methamphetamine, and alcohol) would self-administer these drugs within 1 month to the point of cessation of eating, refusal of hand-fed treats, and in many cases until dead, or near enough to death that researchers removed them from the experiment and provided life-saving measures to keep them alive (e.g., Johanson and others 1976; Pickens and Thompson 1971). This and similar findings in other
182 SCIENCE-BASED GUIDELINES FOR LABORATORY ANIMAL CARE
self-selection domains suggests that the animals' judgments are not always in synchrony with what appears optimal to their health (e.g., Galef and Beck 1990).
Thus we propose the following recommendation regarding the sudden and arbitrary insertion of environmental enrichment procedures into ongoing research: Caution is warranted. We should not assume that enrichment will not affect our research measurements or outcomes unless demonstrated otherwise. And we should not mandate enrichment of animals engaged (or to be engaged) in a research protocol unless the protocol has been explicitly shown not to be affected by the enrichment procedure to be used (or the effects are known and taken into account). Finally, just as we should use caution in generalizing from humans to mice about what we believe is best for a mouse or a human (see "To a mouse ..." by R. Burns [Douglas 1993]), we should also use caution when we generalize across more closely related species, until an experimental basis for doing so has been established.
Still thou art blest, compar'd wi' me The present only toucheth thee: But, Och! I backward cast my e'e. On prospects drear! An' forward, tho' I canna see, I guess an' fear!
And finally, The best-laid schemes o' mice an' men Gang aft agley
Was this article helpful?