INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

BECAUSE OF PRESENT-DAY EXTENDED stays on the International Space Station, potential Mars or Moon missions, and even a budding space tourism industry, it is becoming imperative that we understand the effects of spaceflight on human health. Studies in both rat models and humans have shown that the spaceflight environment can influence a number of physiological and immunological parameters. Changes in total body mass (1, 7, 8, 12, 21, 24, 42, 52, 54), thymus and spleen mass (12, 42, 45), circulating corticosterone (7, 32, 34, 54), mitogen-induced proliferation (7, 31-33, 36, 37, 47), cytokine production and reactivity (4, 13, 31, 32, 35, 37, 48, 49), and lymphocyte (Lymph) subpopulations (2, 23, 42, 46-48) have all been reported.

There are at least three aspects of the spaceflight environment that may be responsible for changes in immune parameters: mission-related psychological stress, low-dose/low-dose rate radiation, and changes in inertial condition (i.e., microgravity) (42, 50). Our laboratory has already reported results of studies with C57BL/6 mice characterizing the effects of low-dose -ray (18, 27, 41), proton (14, 15, 25-27, 40), iron, and silicon ion (19) radiation on immunity. Furthermore, our laboratory has collected preliminary data on the effects of hypogravity [Space Transport System (STS)-77 (42)] and hypergravity (17) on immune parameters, which suggest that the influences of these environmental factors are, in at least some respects, similar. At worst, they could interact synergistically and dramatically reduce immune efficacy (16, 50).

On December 5, 2001, the space shuttle Endeavor (STS-108) launched on a 12-day mission flying mice (C57BL/6J) for the first time for the duration of a shuttle mission. Pregnant mice flew on STS-90 but were euthanized early in the mission with only tissues from neonatal animals analyzed. The primary goal of this study was to examine spaceflight as a model for disuse-induced osteopenia in mice and to test the osteoclast inhibiting protein osteoprotegerin for the biotechnology company Amgen (Thousand Oaks, CA). Nonosseous tissues from the placebo-treated groups are being examined to explore the potential of space-flown mice as models for other biomedical disorders.

This paper is the first of two describing the results of the immunological assessment. Herein, we focus on population distributions in the spleen, bone marrow, and, to a limited extent, blood. Part II of this series covers body and lymphoid organ masses, cytokine expression, and activated T- and natural killer (NK)-cell populations (20). By providing this initial characterization, future investigations utilizing the wild-type C57BL/6 mouse model should be facilitated, allowing both the confirmation of the data presented here and the expansion of the current body of spaceflight research to include genetically modified strains (i.e., transgenics or knockouts), which typically have a mixed genetic background that includes the C57 strain. Furthermore, it is our hope that the data presented here make clear the necessity of characterizing the effect of spaceflight environment on the ability of the immune system to respond to an immunogenic challenge. Suboptimal response could result in overwhelming infection and possibly even death, as well as economic loss due to mission failure.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals, housing conditions, and specimen collection. Female C57BL/6J mice were ordered from Jackson Laboratory (Bar Harbor, ME) and delivered to the National Aeronautics and Space Administration (NASA) Life Sciences Support Facility (hangar L) at the Cape Canaveral Air Force Station, Florida, at 43 days of age. The mice were housed in vivarium cages at payload turnover until the age of 64 days, and standard rodent chow was replaced with food bars at 51 days of age. Food bars (American Institute of Baking Teklad no. TD 97071, Manhattan, KS) and water were provided ad libitum throughout the mission.

Splenocyte and bone marrow assessment using the ABC Vet Hematology Analyzer. Spleen and bone marrow were processed and evaluated as previously described by using the ABC Vet Hematology Analyzer (Heska, Waukesha, WI) (15, 18, 26, 41). Parameters characterized included granulocyte (Gran), eosinophil (Eos), monocyte/macrophage (Mono), and Lymph counts and percentages.

Hematology using the Advia 120 analyzer. After arrival at Jackson Laboratory, whole blood samples were diluted 1:4 in murine PBS (10 mM NaCl, 155 mM KCl, 10 mM glucose, 1 mM MgCl₂, 2.5 mM KHPO₄, pH 7.4), and complete blood counts were determined by using the Advia 120 multispecies whole blood analyzer (Bayer, Tarrytown, NY). Results reported here include counts and percentages of peripheral Gran, neutrophils (Neut), basophils (Baso), Eos, Lymph, and monocytes.

Flow cytometry analysis of bone marrow and spleen. Immunophenotyping was carried out by using a four-color FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA) (17, 40, 41). Monoclonal antibodies labeled with FITC, R-phycoerythrin, allophycocyanin, or peridinin chlorophyll protein were directed against specific markers on various cell populations. Data were collected on all events until 10,000 cells were counted in a CD45 (30-F11) vs. side-scatter Lymph gate.

Statistical analysis. The results were analyzed by using one-way ANOVA and Tukey's pairwise multiple-comparison test. A P value of <0.05 was selected to indicate significant differences among groups, whereas a P < 0.1 indicated a trend. These analyses were performed by using SigmaStat software, version 2.03 (SPSS, Chicago, IL).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Leukocyte subpopulations in the peripheral blood. There were no significant differences noted in the peripheral white blood cell (WBC) count among groups (Table 1). Assessment of five separate leukocyte subpopulations showed no significant effects on Lymph and Gran counts. However, a significant decrease in monocyte numbers compared with AEM, but not Viv, ground controls was observed.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Effect of flight condition on leukocyte subpopulations. A: lymphocytes; B: granulocytes; C: monocytes. Flt, flight group; AEM, animal enclosure module simulated flight control group; Viv, vivarium ground control group. Values are means ± SE. Post hoc Tukey's test: a P < 0.005 and b P < 0.001 vs. AEM; c P < 0.05 and d P < 0.005 vs. Viv.

Leukocyte subpopulations in the spleen. As shown in Fig. 1, there were highly significant effects of flight condition on proportions of splenic Lymph (P < 0.001) and Gran (P < 0.001). There was a flight-induced increase in Lymph percentages, with a corresponding decrease in Gran compared with both AEM and Viv controls. However, there were no changes noted in monocyte percentages. A similar lack of flight-induced change was noted in the Eos (data not shown for all groups); the percentages ranged from 1.05 ± 0.07% for AEM to 1.18 ± 0.07% for Flt.

Leukocyte subpopulations in the bone marrow. The flight effect was even more pronounced in the bone marrow differentials than in the spleen. Although there were no effects on cell counts (Table 1), there were highly significant effects of flight condition on Lymph (P < 0.005), Gran (P < 0.001), and monocyte (P < 0.005) proportions (Fig. 1). There were increases found in the Flt animals for both Lymph and monocyte proportions compared with AEM and Viv controls. There were corresponding decreases noted in Flt Gran percentages compared with AEM, but not Viv, ground controls. AEM controls were also significantly greater than Viv controls in this measure. Similar changes occurred in the Eos proportions (Flt = 3.66 ± 0.11%, AEM = 4.05 ± 0.10%, Viv = 3.62 ± 0.10%, P < 0.01 for effect of flight condition); the Flt and Viv groups were both significantly lower than the AEM mice, P < 0.05 for both.

Flow cytometric analysis of Lymph in the spleen. There were flight-dependent proportional differences noted in the major splenic Lymph subpopulations as measured by flow cytometry (Fig. 2). The proportions of CD3⁺ T and B220⁺ B cells were both significantly influenced by flight condition (both P < 0.05). T-cell proportions significantly decreased, whereas B cells increased, in the AEM, but not Viv, ground controls. There were no changes in NK1.1⁺ NK cell proportions.

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Effect of flight condition on T-, B-, and natural killer (NK)-cell subpopulations. A: CD3+ T cells. B: B220+ B cells. C: NK1.1+ NK cells. Values are means ± SE. Post hoc Tukey's test: a P < 0.05 and b P < 0.001 vs. AEM; c P < 0.001 vs. Viv.

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Effects of flight condition on T-cell subpopulations. A: CD3+ CD4+ T helper cells. B: CD3+ CD8+ T cytotoxic cells. C: CD4-to-CD8 ratios. Values are means ± SE. Post hoc Tukey's test: P < 0.001 vs. a AEM and vs. b Viv.

Flow cytometric analysis of the bone marrow. As was noted in the major leukocyte populations, the effects of flight condition were more pronounced on the Lymph populations in the bone marrow than in the spleen, with significant main effects found in T-, B-, and NK-cell percentages (P < 0.001, P < 0.001, and P < 0.05, respectively). However, the differences noted in this compartment were in the opposite direction from that of the spleen. CD3⁺ T-cell percentages were higher in Flt animals compared with both ground controls, with corresponding decreases in the B220⁺ B cells. There were also increases in the NK1.1⁺ NK-cell proportions in Flt animals compared with AEM, but not Viv, controls. With the exception of a significant decrease in B-cell counts in Flt mice compared with AEM controls, the changes in bone marrow percentages were not reproduced in their corresponding phenotype counts (Table 3).

Habitat temperatures. There was an inconsistency in the temperatures of the different habitats. Differences existed between Flt and AEM temperatures, despite very similar temperature profiles between the space shuttle middeck (23.8 ± 1.4°C, 28.6°C maximum, 18.4°C minimum) and the OES (23.9 ± 1.4°C, 28.7°C maximum, 18.3°C minimum). For the three Flt units, temperatures were consistently above 30°C (Flt 101: 31.8 ± 2.3°C, 35.7°C maximum, 22.4°C minimum; Flt 102: 33.6 ± 2.2°C, 37.3°C maximum, 22.7°C minimum; Flt 103: 32.7 ± 2.3°C, 37.3°C maximum, 22.7°C minimum). Maximum temperatures (>36°C) occurred briefly after launch (~4 h) and when the cabin was depressurized (14.7-10.2 psi for ~24 h) on flight day 5 for an extravehicular activity ("space walk") from the space shuttle middeck.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Due to large differences in launch vehicle, mission profile, flight hardware, ground controls, and animal models, data from flight experiments have sometimes been inconsistent and difficult to characterize. Despite these limitations, it is generally accepted that changes in inertial condition (i.e., hyper- or hypogravity), as well as other factors associated with spaceflight, can have a significant impact on a number of immune parameters.

The data show no significant differences in peripheral WBC counts after STS-108, although the values appeared to be slightly lower in Flt animals compared with ground controls. Because this is the first mouse study in which these parameters have been quantified, there are no other reports against which the present data can be directly compared. However, this is similar to results from some past flights using the rat model. Depending on the flight, peripheral WBC counts have either remained constant (28, 29) or decreased (2, 7, 23). Due to the constraints of this study, we were unable to obtain accurate splenic cell counts. However, others have reported decreases in total splenocyte counts after flight (22, 38). Furthermore, rat splenocytes labeled with the pan-leukocyte antibody OX-1 (different from the clone used here) either decreased (48) or remained constant (49) after flight.

Several studies have found that spaceflight causes neutrophilia. In rats, circulating Neut proportions consistently increased (7, 23, 28, 29), with a corresponding decrease in Lymph (2, 6, 7, 23, 28, 29). Eos counts have been shown to decrease (2), while proportions remain constant (28, 29). However, after STS-108, there were no significant effects of flight condition on the proportion or number of either the circulating Gran subpopulations (Neut, Baso, or Eos) or Lymph. The discrepancies are likely due to the choice in animal models. However, the lack of statistical significance reported here may also be due to the small sample size available for analysis. The decreases in peripheral monocyte counts in the Flt animals compared with AEM controls is consistent with most (2, 6, 7, 23), but not all (7, 28, 29), flights using other animal models.

In contrast to the blood, there were flight-induced increases in splenic Lymph proportions, with corresponding decreases in Gran. Furthermore, there were no significant changes in the splenic monocyte populations after STS-108. Although there appears to be nothing in the literature to confirm or contrast these findings, such a change would be consistent with a phenotypic shift between the splenic and blood compartments (42).

Flow cytometric analysis of splenocytes indicated that the spaceflight environment can influence distinct Lymph phenotypes. Specifically, the proportion of splenic T cells decreased, whereas B cells increased. Past flights characterizing splenic T-cell percentages in rat models are somewhat contradictory, with decreases (42), no changes (22, 38), or increases (23, 48, 49) all reported after spaceflight. Slightly less contradictory findings have been reported in B-cell percentages, which have been shown to either decrease (22) or remain unchanged (2, 23, 38) after spaceflight.

The inconsistencies in the splenic T- and B-cell results could be due to a number of factors, including flight profile, hardware, and species; all of which were variable across past flights. In addition to the differences in animal model, many of these flights were either Cosmos missions or space shuttle missions that included single- rather than group-housed animals. Handling, timing, and analysis techniques may also affect results. Most previous flight studies have utilized single-color labeling methods. The multicolor cytometric methods used after the STS-77 and STS-108 missions allowed us to characterize the various phenotypes with a greater degree of accuracy. However, despite these improvements in methodology, it is unlikely that a difference in technique would produce opposing results across flights. Because the flights that resulted in similar T- and B-cell shifts [STS-108, STS-77 (42), and STS-57 (22)] utilized nearly identical launch vehicles, flight hardware, and mission durations, discrepancies in Lymph shifts are more likely due to differences in these factors rather than in animal model or analysis technique.

Although we were unable to phenotype the peripheral blood, circulating T-cell percentages have remained constant (23), whereas B-cell percentages and counts consistently decreased (2, 23) after flight in the rat model. This suggests that there may have been a phenotypic switch between compartments in the rat, with B cells leaving the blood and entering the spleen. However, further study is needed to confirm this shift in the mouse model.

After STS-108, there was a shift in splenic T-cell subtypes. CD3⁺/CD4⁺ Th-cell proportions decreased, whereas CD3⁺/CD8⁺ Tc-cell proportions remained constant after flight. This led to a reliable decrease in the splenic CD4-to-CD8 ratio similar to at least one previous flight (42). A reduction in this ratio has frequently been associated with an immunocompromised state. However, without knowing what the circulating CD4⁺ and CD8⁺ counts are, it is difficult to make more firm conclusions.

As was seen in the general Lymph populations, there have been a number of flights that did not influence splenic T-cell subset percentages. The proportions of both Th (2) and Tc cells (2, 23, 38) have often remained constant. However, there have also been a number of flights that resulted in increases in Th- (22, 38, 49) and Tc-cell (22, 48, 49) proportions. In sharp contrast, there has only been one report, other than ours, in which splenic Th-cell percentages have decreased (23). Although CD4-to-CD8 ratios have not typically been reported in the past, we have found similar decreases across two similar flights with two different species (42). The discrepancies between the data presented here and that of past flights is somewhat disconcerting. There does not appear to be any hardware, launch vehicle, or flight duration-dependent rationale that adequately explains these inconsistencies. Furthermore, because we found similar results with two different species, this parameter does not appear to be a driving factor. Flow cytometric technique may have been of primary importance as the CD4⁺ and CD8⁺ T cells are relatively specific subpopulations. Previously used single-color labeling techniques may not have been exclusive enough (i.e., CD4 can be found on both Th cells and macrophages).

In any case, there appears to be a slight shift in immunological homeostasis away from the cell-mediated branch of adaptive immunity toward humoral mechanisms, as indicated by the phenotypic shift away from T cells toward B cells. Furthermore, as will be discussed in part II of this series (20), the changes noted in phenotype distributions are also consistent with the changes noted in IL-2 expression. Mitogen-induced IL-2 production was significantly lower in Flt mice than in both AEM and Viv ground controls. However, interferon-, tumor necrosis factor-, IL-2, and IL-4 expression did not appear to be influenced by spaceflight. Because the mitogen used specifically stimulated T cells (phytohemagglutinin), it is possible that the generalized T-to-B-cell population shift may be related to decreased capacity for IL-2 expression.

There were no significant effects of flight on NK cell proportions in the spleen. This is similar to what Sonnenfeld et al. (48, 49) found in the rat model. However, whereas there was an increase in the proportion of activated NK cells after the present flight (see part II of this series) (20), there have been reports of flight-induced decreases in splenic NK cytotoxicity (31, 44).

As we found no flight effects on the total bone marrow Lymph counts, there is little indication that hematopoietic mechanisms are responding to the changes noted in the spleen and blood. In contrast, although there were no effects of flight on monocyte, Lymph, or Gran counts, there were increases in the Lymph and monocyte proportions, with corresponding decreases in Gran. Because these changes are similar to those found in the spleen and peripheral blood, this is consistent with a shift in leukocyte homeostasis and/or their migration patterns. Rather than compensating for changes in peripheral phenotypes, the bone marrow appears to be inherently involved in these shifts.

Although not directly comparable, this fundamental shift in leukocyte homeostasis is not inconsistent with past experiments investigating hematopoiesis. Although there has been at least one report of an increase in hematopoietic measures (6), the majority of studies suggest that the response of bone marrow to various colony-stimulating factors has either decreased (1, 47-49) or remained constant (7, 23, 30, 52) after flight.

Flow cytometric analysis of bone marrow stem cell populations gives credence to a shift in leukocyte hematopoiesis. Although the results indicate that there were flight-induced decreases in the overall blast population, CD34⁺ parameters increased compared with AEM, but not Viv, controls. No significant differences, however, were seen in the Ly-6A/E⁺ population. This is of interest because CD34/Ly-6A/E⁺ cells are part of a self-renewing hematopoietic stem cell population (39), suggesting that, if there is indeed a change in hematopoietic mechanisms, it occurs while the cells are differentiating into more mature phenotypes.

The proportion of CD34/Ly-6A/E double-positive cells was also shown to be dependent on flight condition. Although post hoc analysis indicated that there were no significant differences across groups, this population is relatively small to begin with, and even a limited change could have large downstream consequences. Furthermore, as this population includes hematopoietic stem cells involved in the recovery of short-term multilineage bone marrow after lethal irradiation (39), changes could be of critical importance after a particularly large solar particle event.

In contrast to the general leukocyte distribution, flight-induced changes in bone marrow Lymph subpopulations were very different from, and in some cases opposed, those of the spleen. T-cell proportions increased with a corresponding decrease in B cells. Also, both Th- and Tc-cell percentages increased in the bone marrow after flight. As a result of this simultaneous increase, there were no significant differences noted in the CD4-to-CD8 ratio. There were also flight-induced increases in the NK-cell proportions. These phenotypic shifts in bone marrow Lymph are similar to previous reports (48). Together, these results suggest that, whereas hematopoietic mechanisms appear to be encouraging the leukocyte shift toward monocytes and Lymph, this cannot be said for all of the different Lymph phenotypes.

For some of these analyses, there were significant differences between AEM and Viv ground controls. Although this was the case for the proportion of circulating Lymph, most of the differences between ground controls were found in the analysis of the bone marrow (%Lymph, %Gran, %Eos, %blast, %CD34, number of CD34, and %Ly6/CD34). Similarly, Flt animals were different from one or the other of the ground control groups more often in the bone marrow (%Lymph, %Gran, %Eos, %blast, %CD34, and number of CD34 stem cells) than in blood (number of monocytes) or spleen (%T and %B cells). Most of these parameters were found to be significantly lower in Flt animals compared with AEM, but not Viv, controls. This suggests that animals in the ground AEMs may have been more stressed than even the Flt animals.

As stated previously, all mice in this study were treated with calcein before flight to enable the characterization of bone growth. A more detailed discussion of the effects of this label is included in part II of this series (20). However, calcein has a relatively short half-life (3) and is cleared from circulation within a few hours (43). In vivo studies strongly indicate that calcein has no effects on leukocyte population distribution or function (5, 10, 11, 53). Hence, we expected no significant effect of calcein on the parameters measured in this study.

A source of potential differences between flight and ground immune parameters may be the elevated temperature of the flight AEMs. NASA Ames Research Center personnel have run a test since the STS-108 mission and have identified the problem to be caused by recirculation of air within and between hardware units. As a result, some of the heat generated by the airflow exhaust fans in the AEMs increased the temperature of the intake air stream. The relative configuration and airflow characteristics of the ground AEM group were sufficiently different from those of the Flt group that waste heat was not recirculated and the inlet temperature more closely matched the room temperature. Similar elevated temperatures in flight have occurred previously with rats flown in the AEM habitats. For example, on STS-41, temperatures in flight AEMs peaked at 36°C (9, 51), and on STS-72 temperatures peaked at 35°C (unpublished data). Average temperature profiles were not published for these missions. Although temperature data have been published for very few flights, unusually high temperatures have not occurred on all flights (STS-77: 28.6 ± 1.0°C, 31.3°C maximum, 25.6°C minimum) (6).

There are several factors, different for each group, that impact the ability of mice to reject metabolic heat, the real physiological driver of temperature. These differences make it very difficult to simply rerun a set of Viv to mimic the high temperatures experienced by the Flt mice. First of all, in microgravity, buoyancy-driven convective heat flow does not exist (hot air rising because it is less dense than cooler air). Second, the forced-air convection (~0.25 ft./s to move waste toward the exhaust filter) of the hardware (Flt and AEM groups) causes convective rejection of metabolic heat to the degree that a slight "wind chill" effect (not characterized for mice) may effectively reduce the "temperature" slightly. Finally, the temperature within the vivarium cages (Viv group) is most likely much higher than the temperatures of the animal holding rooms because the air within the vivarium cages is stagnant due to the greatly restricted flow through the microisolator cage tops.

Therefore, the heat loads for each group (Flt, AEM, and Viv) are inevitably different. The degree to which forced air vs. buoyancy-driven convective forces results in heat rejection is not characterized for mice. However, forced-air convection is likely to remove more heat than buoyancy in this investigation, effectively reducing the heat load for the Flt and AEM groups and raising the heat load for the Viv group. Thus these effects will tend to reduce the differences that would otherwise be predicted based on temperature alone. Still, because of heat load differences, firm conclusions about the effects of the spaceflight environment on any physiological parameter, particularly those as sensitive as immunity, should be viewed with caution.

Another aspect of space experimentation that results in uncontrollable differences is the access to a three-dimensional environment in microgravity (Flt) vs. the two-dimensional floor space for the ground controls (AEM and Viv). Appropriately accounting for the translation from livable floor space to livable volume is not an issue that has been quantified but is inevitably going to impact social behavior and potentially stress. Because corticosterone measurements (a common indicator of stress levels) were not taken, this is purely speculative. We suggest that translating usable volume for the flight animals to more floor space for the ground controls may be important when the design of future spaceflight hardware for animals is considered. This, in addition to the convection discussion, is another example of how it is impossible to change only one variable when removing gravity. The "spaceflight environment" becomes the variable, and it is important to understand the complexity of this parameter.

Despite the described issues, these results suggest that the spaceflight environment can have a considerable influence on every level of immunity, from hematopoiesis to acquired immunity. With all of the changes in immune parameters reported here and elsewhere, regardless of the inconsistency of the data, it is surprising that the question of immune efficacy has not been directly addressed. Specifically, do these changes in immune homeostasis affect the ability to respond to an immunogenic challenge (e.g., bacteria, viruses, and fungi)? Ground-based models for various aspects of the spaceflight environment are currently headed in this direction. For example, we have already begun to characterize the effects of -, iron-, and silicon radiation-induced changes on the ability of the immune system to respond to, and develop memory for, a primary and secondary challenge (unpublished observations). Furthermore, plans for a continuation of this work using proton radiation (the most common form of space radiation), as well as hypergravity (as modeled by centrifugation), are in progress. The next logical step would be to expose animals to a relevant challenge while in the spaceflight environment. This type of information should help mission managers formulate a more accurate astronaut risk assessment for long-duration missions.