Housing in the Animal Enclosure Module Spaceflight Hardware Increases Trabecular Bone Mass in Ground-Control Mice

The “weightless” environment encountered during spaceflight complicates the traditional concerns involved in animal housing, including nutrient delivery and waste management. Spaceflight experiments utilizing mice or rats require habitats that are specifically designed for this environment. During studies on the space shuttle, rodents are housed in a specially designed stainless steel meshed cage called the Animal Enclosure Module (AEM) (Brooks, 1981). Despite the retirement of the shuttle program in 2011, plans are in place to adapt the AEM for use on the International Space Station (ISS). AEMs are equipped with systems that provide food and water ad libitum (Zerath et al., 2002). In addition, these flight enclosures have a constant airflow mechanism designed to move free-floating animal waste towards an exhaust filter. AEMs allow for efficient containment of animals in the confines of the crew cabin, while also satisfying their homeostatic demands. The first test flight of the AEM was during shuttle mission STS-8 in 1983 (Smith et al., 1987). Since then, AEMs have been successfully utilized on over 20 missions involving both rats and mice.

Despite their demonstrated utility, AEM housing may have an effect on the physiology of the rodents contained within. Beyond consideration of the spaceflight environment, a non-standard cage environment like that found with AEMs may lead to unexpected changes in animal behavior that could complicate experimental results. These considerations are especially important in studies examining the skeletal effects of microgravity. In this case, even subtle changes in animal loading can distort experimental findings. For example, a study by Morey-Holton and colleagues revealed that group housing of rats in AEMs reduced their skeletal response to spaceflight by as much as 80% (Morey-Holton et al., 2000).

In the present study, we compared the effects of 13 days of AEM housing on various bone compositional, microarchitectural, and histological parameters. We documented significant effects of the AEM on the skeletal properties of mice, particularly in the trabecular bone compartment. This study was conducted as part of a ground-based parallel of an animal payload on space shuttle flight STS-108. Characterization of animal housing is important so that results from spaceflight experiments can be more reliably compared to ground-based controls and studies from other missions.

We have demonstrated that ground-based use of the AEM results in significant, albeit complicated, effects on murine skeletal physiology. With only modest effects on cortical bone, the AEM was found to produce its most dramatic effects in the trabecular bone compartment. At the proximal tibia and humerus, we observed significantly greater trabecular bone volume fraction and connectivity density in AEM-housed mice versus vivarium. Histological analysis at the proximal tibia suggests that increased accumulation of trabecular bone mass was likely due to a cessation of bone resorption, as there was a significantly lower osteoclast surface and eroded surface for AEM mice when compared to vivarium. Previous studies have determined that, beginning at approximately 6 weeks of age, C57BL/6J mice begin to continuously lose trabecular bone mass at multiple sites (Ferguson et al., 2003; Halloran et al., 2002). Given the 8-10 week age of these mice, the maintenance of trabecular bone structure at the proximal tibia and decline in trabecular bone structure at the proximal humerus of vivarium mice is not unexpected. Indeed, the greater trabecular bone mass and changes in microarchitectural parameters in AEM versus vivarium mice at multiple sites suggest that a significant hardware effect is present.

Histological analysis also revealed a decrease in osteoblast surface for AEM-housed animals versus vivarium at both the proximal tibia and the L5 vertebrae. Osteoid volume, surface, and thickness at the proximal tibia were all significantly lower for AEM-housed mice. There were also lower endpoint levels of serum osteocalcin and alkaline phosphatase for AEM mice. Taken together, these findings point to an overall decline in osteoblast activity and bone formation (Gundberg et al., 2002; Lumachi et al., 2009). This was not the case, however, for periosteal mineralization surface, which was used as an index of bone formation (Grimston et al., 2011), in femur cortical bone (Figure 2A). We also demonstrated a significant increase in mineralizing surface at the proximal tibia (Figure 4C). In considering the seeming incongruence of these histological and histomorphometric findings, it is also important to consider the methodology underlying the quantitative histomorphometric analyses. By administering a single Calcein label at day 0, one is actually visualizing mineralization sites present at any point over the course study, without being able to resolve fluctuations throughout. Given the results of the microCT and histological assays, it is likely that bone formation was relatively high in AEM-housed mice for the first half of the study, prior to adaptation of their activity and behavior to the new enclosure. This would have increased the relative mineralizing surface and contributed to the increased trabecular bone mass we documented via microCT. However, over the latter half of the study, the anomalous loading declined and bone formation decreased. This is evidenced by the aforementioned decrease in tibia osteoblast surface and decline in serum markers of bone formation. It is important to remember that all of these assays represent the state of bone physiology at the study endpoint only, which limits our interpretation. A single Calcein label was administered in order to limit variability in experimental design as mice in the parallel spaceflight experiment could only be administered a single label prior to launch. In retrospect, double labeling should have been utilized, as it would have provided dynamic indices. This will certainly be considered in future ground-based studies.

There were increased levels of serum phosphorus in AEM-housed animals versus vivarium. Parathyroid hormone (PTH) is one of the primary regulators of both serum calcium and phosphorus homeostasis (Bergwitz and Juppner, 2010), producing increased calcium release from bone breakdown and increased phosphorus excretion via the kidney. The relatively higher phosphorus levels suggest a lower PTH level. There was not, however, a difference in serum calcium. Given the proposed biphasic bone formation response, it may be that serum phosphorus levels were simply lagging behind calcium in returning to baseline following an early increase in bone formation. Additional insight could be gained from longitudinal measurement of serum PTH levels in future studies.

We must also consider that the observed increase in trabecular mineralizing surface, and potentially bone formation, occurred as a result of the reduced activity of bone-resorbing osteoclasts in AEM-housed animals. Normal (i.e., higher) levels of osteoclast-mediated bone resorption in vivarium-housed mice could have removed portions of the Calcein label that were not resorbed in the AEM-housed mice. The presence of these labels, and their measured size, are necessary for the quantification of histomorphometric parameters. These complications highlight one of the inherent limitations of single label histomorphometry (Arnett and Henderson, 1998). Future studies of the AEM should make use of double label histomorphometry over smaller time frames in order to definitively resolve early versus late changes in bone formation.

Differences in cortical bone parameters between AEM- and vivarium-housed mice were modest compared to findings in the trabecular compartment. There were no changes in cortical structural parameters, such as cortical bone volume or cortical bone thickness, for AEM versus vivarium. However, there was significantly lower femur stiffness for AEM-housed mice, accounted for by lower mineral content of the femur diaphysis. Compared to baseline, AEM-housed mice had greater maximum calculated resistance to torsion and cortical thickness – differences not seen in vivarium-housed mice. However, none of these changes were substantial enough to result in functional outcomes in terms of increased bone mass or strength. It is important to note, however, that cortical bone is generally less sensitive to loading and unloading compared to trabecular bone (Ducher et al., 2004; Warner et al., 2006).

Greater trabecular bone mass and trabecular mineralizing surface and reduced osteoclast surface are the most striking effects of AEM housing. To account for these findings, we propose that the structure of the AEM enclosure itself and the dynamics of group housing likely contribute to altered animal behavior and increased mechanical loading. The AEM has been designed in such a way that it will necessarily modify skeletal loading during normal activity. The wire mesh walls of the AEM module allow the animals to climb on the walls and roof, a behavior that is different from the standard “two dimensional” floor movement found in vivarium enclosures. This wire mesh is critical, however, for mechanisms that remove floating waste in microgravity (Brooks, 1981; Smith et al., 1987). In addition, the wire mesh provides an anchor point for mice during spaceflight, helping to reduce confounding effects of excessive stress on animal physiology (Sonnenfeld, 1999). It is also important to note that while the animals in the AEM are not crowded by NIH standards, they do have 60% less floor area per animal than the standard vivarium enclosures (130 vs. 81 cm²/mouse). The increased density of animals could have led to increased movement (running, jumping, fighting, climbing over other animals, etc.) that could contribute to increased loading and thus increased bone formation and accumulation of bone mass. In addition, it has been proposed that the interaction and companionship afforded by higher density group housing in the AEM may result in less stress for the animals when compared to standard vivarium housing (Morey-Holton et al., 2000). Although not measured here, the stress hormone cortisol and other glucocorticoids are potent activators of bone resorption in both humans and rodents (Gluer et al., 2007; Rehman and Lane, 2003). If glucocorticoid levels were lower in AEM versus vivarium, it may have contributed to the increase in bone mass that was observed. Measurement of stress hormone levels and adrenal mass would be worthwhile additions to future studies involving the AEM.

The record of spaceflight experiments utilizing animal models has yielded variable results with regards to the effects of housing environment on skeletal properties. Indeed, some studies demonstrated rather dramatic skeletal effects of housing type (Lafage-Proust et al., 1998; Morey and Baylink, 1978; Turner et al., 1979; Yagodovsky et al., 1976), while others saw relatively few changes (Bateman et al., 1998; Turner, 1995; Wronski et al., 1998). The results are so variable that one paper describes “arrested bone formation” (Spector et al., 1983), while another concludes there were “normal levels of [trabecular] bone mass and bone formation” (Wronski et al., 1998). Both the exaggerated and blunted effects of spaceflight on the rat skeleton may be an effect of the differing flight and ground control hardware housing conditions. The effects we observed are modest by comparison to the aforementioned studies, yet demonstrate a significant and important influence of the AEM housing environment that should be considered when planning future spaceflight experiments.

A recent study by Blotter and colleagues investigated the effects of ground-based housing of mice in the proposed “Mice in Space” (MIS) system (Blottner et al., 2009), which has not previously flown in space. In this study, investigators compared 25 days of housing in the MIS to individually ventilated cages. The investigators found no difference in bone structure or mineralization between the two groups. However, it is important to note the differences between the MIS and AEM enclosures, as well as variables related to experimental design. Importantly, the MIS system does not have the same wire mesh walls as the AEM. These walls provide an attachment point for the animals during spaceflight, helping to reduce animal stress in microgravity as well as during take off and landing. During ground-based use of the AEM, the mice are able to climbs the walls and roof of the AEM. As mentioned before, this increase in animal activity and loading may have contributed to the observed increase in bone mass. It is also important to note difference in animal age (3-4 months in MIS vs. approximately 2 months in the present study), study duration (25 vs. 13 days), and animal sex (male vs. female) that complicate a direct comparison. The AEM has proven utility on multiple spaceflight missions, although additional investigation of the AEM and novel enclosure systems would be warranted before making a final decision on future use aboard the ISS or elsewhere.

A broad analysis of current literature performed by Morey-Holton and colleagues indicates that flight hardware housing parameters determined the degree to which spaceflight caused skeletal degradation (Morey-Holton et al., 2000). Specifically, they compared skeletal parameters from rats following nine days of spaceflight. Rats were housed in either AEMs (group housed with six rats in each AEM) or in the Research Animal Housing Facility (RAHF; singly housed rats in each habitat chamber). While individual, RAHF-housed animals experienced substantial declines in bone formation and bone mass, no significant changes were seen in AEM-housed animals. A similar study by Wronski and colleagues found that group housing of rats in AEMs essentially abolishes the effects of spaceflight on the skeletal system (Wronski et al., 1998). Although they did not use the AEM, other studies have demonstrated the confounding effects of group housing of animals on skeletal properties (Vico et al., 1993; Vico et al., 1988). We note that space-flown group housed rats are in relatively crowded conditions and experience substantial loading during normal social activities. It is for this reason that we proposed modifying the AEM to fly mice on STS-108 (Dalton et al., 2003).

Aside from rodent housing in spaceflight hardware, there is also precedent for the physiological effects of housing environment found in research regarding the bone structure of laying hens (Whitehead, 2004). As with rodent housing, the space available to hens and their ability to experience static and activity-associated loading correlated with positive bone mass and strength parameters (Newman and Leeson, 1998). For example, hens housed in cages with perches or in low-level aviary systems had greater leg bone strength, but relatively less of a change in wing bone strength, whereas birds permitted to fly in high-level aviaries showed relatively greater improvements in wing bone strength (Fleming et al., 1994; Knowles and Broom, 1990; Silversides et al., 2012). Positive effects of housing environment on bone quality in these animals are likely due to increases in osteoblast-mediated bone formation, rather than a decrease in bone resorption (Newman and Leeson, 1998; Whitehead, 2004). Much like the present study, this work demonstrates the complicated and perhaps unpredicted consequences of animal housing on bone.

There are many variables to consider when designing a ground-based control for the spaceflight environment, including the spatial distribution of animals, their range of movement, as well as the lack of buoyancy-driven convection. Conventional thinking would suggest that AEM-housed animals on the space shuttle or ISS should be compared to AEM-housed animals on the ground. However, the unique three-dimensional distribution of animals within the AEM during spaceflight, compared to their two-dimensional distribution on the ground, inevitably results in an imperfect comparison; however, this will be the case for any gravity versus microgravity comparison. Despite the limitations of the AEM as a ground-based enclosure, one cannot discount the impressive wealth of information that has been gleaned from spaceflight experiments, including those rodent studies that utilize the AEM. However, considering the complicated effects of the AEM on skeletal parameters presented here and in previous studies, and the logistical hurdles necessary for a complete redesign of flight hardware, it is clear that some adjustments are needed. The best recommendation would be to give the mice a longer acclimatization period to the AEM, or at least the wire mesh cage configuration. Giving the animals two weeks prior to the start of the flight experiment may help to negate some of the early behavioral, activity, and stress changes that occur as the mice adjust to their new environment. In addition, minimizing animal crowding, as much as possible, would be ideal.

In summary, the present study was conducted as a parallel ground-based control of a space shuttle flight experiment aboard STS-108. A comparison of mice housed in AEMs with mice housed in standard vivarium cages revealed a substantial influence of flight hardware on skeletal parameters. The effect of AEM housing was primarily evident in the trabecular bone compartment, with greater bone mass seen at the proximal tibia and humerus. These effects were likely due to a suppression of osteoclast-mediated bone resorption combined with an early increase in osteoblast activity. Although the specific reasons for this effect are not clear, it is reasonable to hypothesize that altered skeletal loading, with mice living and climbing on the wire-meshed walls and interacting with each other, is a significant contributor. While the AEM will remain an integral part of animal experimentation on the International Space Station, it would be prudent to consider changes to both ground- and space-based utilization of these enclosures in order to minimize these effects.