Trends in mucosal immunity in Antarctica during six Australian winter expeditions

Immunology and Cell Biology (2002) 80, 382–390 Research Article Trends in mucosal immunity in Antarctica during six Australian winter expeditions J L Y N N F R A N C I S , 1 M A R E E G L E E S O N , 1,2 D E S M O N D J L U G G , 3 R O B E R T L C L A N C Y , 1,2 J E F F M A Y T O N , 3 K E V I N D O N O V A N , 3 C H R I S T I N E A M C C O N N E L L , 3 TREVOR R TINGATE,3 BRUCE THORPE3 and ANNE WATSON3 1 Immunology & Microbiology School of Biomedical Sciences, Faculty of Health, University of Newcastle, Callaghan, Department of Immunology, Hunter Area Pathology Service, John Hunter Hospital, New Lambton, NSW and 3Polar Medicine Branch, Australian Antarctic Division, Kingston, Tasmania, Australia 2 Summary The mucosal immune status of Australian Antarctic personnel was monitored during six wintering expeditions at two Australian Antarctic Research Stations, Casey in 1992, 1993, 1994, and Mawson in 1992, 1995, 1996. Salivary immunoglobulin and albumin levels were examined for differences between stations and expeditions, and for monthly changes over the expedition year. Salivary IgA and IgM concentrations were on average higher for the 1993 Casey expeditioners, and all salivary protein levels were lower for 1996 Mawson expeditioners compared to levels of the other expeditions. The change in salivary IgA and IgM concentrations over the 1-year period revealed a consistent pattern between expeditions. Salivary IgA levels were lower in March, April and May compared to other months of the year (P = 0.0002). Salivary IgM levels were lowest in the first 4 months of the year, with peak levels in June and July (P < 0.0001). There were no changes in salivary IgG and albumin concentrations over the expedition year. Though the cause of the changes in salivary IgA and IgM levels over the year is unknown, the changes could reflect alterations in mucosal immunity in response to stressors associated with isolation. Key words: albumin, Antarctica, immunoglobulin A, immunoglobulin G, immunoglobulin M, mucosal immunity, saliva, stress. Introduction This study is the continuation of the initial investigation of mucosal immunity in Australian Antarctic expeditioners in 1992, which addressed the concern that Antarctic expeditioners may experience immunosuppression as a result of their wintering in Antarctica.1 Wintering Antarctic expeditioners are exposed to extremely harsh environmental and psychological conditions over a prolonged period of time. 2 It is important to understand the physiological and emotional changes that can result from such exposure in order to prevent, reduce or correct for deleterious effects.3 Several studies have demonstrated alterations in the humoral and cell-mediated immunity of Antarctic expeditioners in response to the stressors associated with prolonged Antarctic isolation.4–7 Depression of the cell mediated immune response (CMI) during Australian Antarctic expeditions has been shown to correlate with perceived anxiety. 5 This association is observed in both the shorter summer expeditions 8 and the year-long winter expeditions.5 Very little is known about the response of mucosal immune parameters in Antarctic expeditioners. 1 Many of the stressors present in the Antarctic environment are known to Correspondence: J. Lynn Francis, Immunology, Hunter Area Pathology Service, Locked Bag 1, Hunter Region Mail Centre, NSW 2310 Australia. Email: lynn.francis@newcastle.edu.au Received 7 January 2002; accepted 26 April 2002. impact on the mucosal immune system in other psychologically and physically stressful environments. 9–12 Changes in the levels of salivary immunoglobulins, in particular salivary IgA, are good indicators of mucosal immune status, and reflect the body’s response not only to infection, but also to acute and chronic stress of both a physical and emotional nature.10–16 Regardless of the initiating factor(s), suppression of the mucosal immune system increases the risk of mucosal infection. 17–20 The initial investigation of mucosal immune changes at Australian Antarctic stations was conducted at Casey, Davis and Mawson in 1992.1 Although some stations had statistically different mean levels of salivary IgA and IgG, the station differences during 1992 were not clinically relevant. The most interesting finding for the 1992 expeditions was the consistent pattern of variation in mean monthly salivary IgA and IgM over the course of the year at each station. The salivary IgA and IgM levels were lower in the first 4 months of the year, reached maximum values in July–August, and returned to yearly mean levels at the end of the year. The aim of the current study was to determine the patterns of change in mucosal immune parameters between different expedition years at two Antarctic stations, Casey and Mawson, and to examine the differences in mucosal immunity between the expedition years. The current study presents the changes in salivary immunoglobulins for expeditions to Casey in 1992, 1993 and 1994 and to Mawson in 1992, 1995 and 1996 from time of departure from Australia until isolation was broken the following year. Trends in mucosal immunity in Antarctica 383 Table 1 Demographic details for the winter expeditioners at the Australian Antarctic stations of Casey in 1992, 1993, 1994 and Mawson in 1992, 1995, 1996. Percentages of scheduled monthly samples collected and percentage of results available for statistical analysis after exclusion of unsuitable results Station and year Casey 1992 1993 1994 All Casey Mawson 1992 1995 1996 All Mawson All Expeditions Subjects Female:Male Median age (years) Percentage of monthly samples Percentage of results after exclusions* 16 16 17 49 2 : 14 2 : 14 3 : 14 7 : 42 30.6 (23.7–42.1) 35.1 (25.7–52.8) 34.3 (29.1–45.7) 33.2 (23.7–52.8) 95.2 (198/208) 76.4 (110/144) 94.1 (80/85) 88.6 (388/438) 94.4–97.0 99.1–100 86.3–98.8 94.3–98.2 27 8 17 52 101 1 : 26 1:7 2 : 15 4 : 48 11 : 90 31.7 (26.1–51.2) 31.7 (24.8–54.9) 36.4 (26.0–53.8) 33.5 (24.8–54.9) 33.5 (23.7–54.9) 95.4 (309/324) 86.3 (69/80) 90.8 (108/119) 92.9 (486/523) 90.0 (874/961) 88.3–90.0 85.5–95.7 96.3–98.1 91.1–93.0 92.6–94.3 *Percentage differed for each protein if insufficient sample was available for testing of all analytes. Lowest and highest percentages of results are indicated. Table 2 Geometric mean salivary IgA (mg/L) concentrations for each month of data collection for each year and station. The months with lower station levels (March–May) are bold Station and year Month* 0 1 2 3 4 5 6 7 8 9 10 11 12 All months Casey Mawson 1992 1993 1994 All years 1992 1995 1996 All years 54.60 45.60 64.72 52.98 42.10 61.56 43.82 66.69 63.43 66.02 59.74 58.56 47.94 55.15 – 86.63 72.97 82.27 90.02 79.84 108.85 – 72.97 77.48 102.51 – – 84.77 81.45 – 82.27 – – 51.94 – – 54.05 – 68.03 – – 66.02 71.52 60.95 72.97 64.07 56.83 62.80 68.03 66.69 62.18 70.11 71.52 58.56 7.94 64.72 62.80 70.11 69.41 54.05 52.98 71.52 76.71 64.07 90.92 78.26 64.07 79.04 – 68.72 92.76 113.30 69.41 69.41 54.05 54.05 47.47 47.94 51.42 62.18 – – – 61.56 – – – – – 46.53 48.91 56.26 68.03 74.44 62.80 – 79.04 60.34 69.41 75.19 69.41 56.83 52.98 58.56 60.34 58.56 75.94 75.19 63.43 79.04 79.04 65.37 Month 0 = Pre departure samples collected in Hobart in December of the previous year. Materials and Methods Subjects This study involved the Australian National Antarctic Research Expeditions (ANARE) that wintered at Casey in 1992, 1993 and 1994, and at Mawson in 1992, 1995 and 1996. Overall, data were available from 49 expeditioners from Casey, seven of whom were female, and 52 expeditioners from Mawson, four of whom were female. Four expeditioners from each station had insufficient data for assessment. The age distribution of expeditioners was consistent between stations and across years. Ages ranged from 24 to 55 years old, with a median age of 33 years (Table 1). Saliva collections Routine monthly saliva samples were collected by the station doctors throughout each expedition year and immediately stored at –70°C. Preliminary saliva samples were collected in Australia in late November or early December of the year preceding the study for the 1992 and 1994 expeditions and stored frozen at –70°C at the Australian Antarctic Headquarters in Hobart, Tasmania. Samples were collected 2 h postprandially by gently spitting into a collection tube. The saliva flow rate was not stimulated. Saliva samples were collected on 13 occasions at Casey in 1992 (Dec 1991–Dec 1992), nine in 1993 (Jan–Jun, Aug–Oct), and five in 1994 (Dec 1993, Feb, May, Aug and Oct 1994). At Mawson, saliva was collected on 12 occasions in 1992 (Dec 1991–Nov1992), 10 in 1995 (Dec 1994–Sep 1995) and seven in 1996 (May–Oct, Dec 1996). The summary of mean log IgA levels in Table 2 generally indicates the months of saliva collection encompassed by this study. Though the initial study plan was to collect saliva samples at every month for every expedition, some individual expedition study plans had to be revised due to expedition-specific constraints. The actual data collection rates were very high. Casey and Mawson collected over 95% of all scheduled samples in 1992, and 91% over the total study period. Samples were sent back to Australia frozen at the end of each study year. One technician assayed all samples from a given study year. Samples were excluded if they had deteriorated during storage, if the subject had been fasting at the time of collection, or if there was insufficient sample for analysis. In fasting samples, levels of IgA and IgM are uncharacteristically elevated due to alterations in flow rate, and deterioration of the samples is detected by exceptionally low levels of albumin indicative of protein degradation.9 Samples were identified as ‘fasting’ when all salivary protein values exceeded the Hunter Area Pathology Service 384 JL Francis et al. Figure 1 Distributions of the mean log salivary IgA, IgM, IgG and albumin concentrations for expeditioners at Casey in 1992, 1993 and 1994. The mean diamonds indicate expedition mean values (centre line), 95% confidence intervals (diamond height) and relative sample size (diamond width). The shorter lines within the diamonds are significance overlap lines. The Tukey-Kramer plots provide a visual representation of the degree of overlap of the expedition group means. (HAPS) adult non-fasting reference ranges17 and at least two of IgA, IgG and albumin were more than double the upper limit of the nonfasting reference range. Non-fasting ranges (mg/L) are: IgA (10–105), IgG (0–20), IgM (0–10), albumin (10–110). Fasting ranges (mg/L) are: IgA (20–500), IgG (0–95), IgM (0–15), albumin (20–370). A sample was considered deteriorated when IgA and albumin values were both less than 15 mg/L, and IgG and IgM were nil detected. Most expeditioners had valid samples for the majority of their collections, though a few had several of their samples rejected due to apparent fasting at the time of collection. Over 95% of all collected samples yielded usable results after excluding samples due to deterioration or fasting (Table 1). Salivary immunoglobulins and albumin ImmunoglobulinA, IgG, and IgM were measured in unstimulated whole mixed saliva by an in-house ELISA using commercially prepared unconjugated and biotin-conjugated antihuman antisera to detect IgA, IgG or IgM (Biosource International, Camarillo, CA, USA). Details of this method have been previously reported.1 Quality control was assessed against a range of three known positive controls (low, medium, high) and a known IgA negative control for all assays of the three classes of immunoglobulins. The average coefficient of variation (cv) for the entire study was 11% for IgA, 13% for IgG, and 16% for IgM. Salivary albumin was measured by rate nephelometry using a Beckman ARRAY analyser, standards and controls (BeckmanCoulter, Brea, CA. USA). The cv for albumin was 8%. Statistical analysis Comparisons of protein concentrations between the stations and between the expeditions were made by fitting one-way analysis of variance (ANOVA) models of individual’s mean yearly concentrations, and by fitting repeated measures ANOVA models with all observations. Monthly salivary protein concentrations were compared over the months of the year by fitting repeated measures ANOVA models and by assessing individuals’ changes in protein concentrations (person-centred values) in one-way ANOVA models. One-way ANOVA was used to compare individual mean concentrations of each protein between stations, and between expedition years at each station (Figs 1 and 2) using the statistical software package JMP.21 The resulting figures illustrated the similarities and differences in the salivary protein levels for each expedition. A mean salivary protein level was calculated for each individual from all of their concentrations over the year, and the distributions of these mean values were compared between the expedition years. The TukeyKramer test was used for the multiple pairwise comparisons of the years, to indicate the nature of the differences between expeditions. As the distributions of the salivary protein concentrations were skewed, the logged values were used in the ANOVA procedures. Resulting mean log values were antilogged to give the reported geometric means (gm) of the values in the original analytical scale of milligrams per litre (mg/L) (Tables 2 and 4). Monthly salivary protein concentrations were compared between expedition years and over the months of the year using repeated measures analysis of variance with the Huynh-Feldt (H-F) correction factor as implemented by the statistical package Stata.22 A repeated Trends in mucosal immunity in Antarctica 385 Figure 2 Distributions of the mean log salivary IgA, IgM, IgG and albumin concentrations for expeditioners at Mawson in 1992, 1995 and 1996. Mean diamonds and Tukey-Kramer plots are as for Fig. 1. measures procedure was used to properly account for the potential correlation of multiple measures on the same individual. Person-centred values were calculated to illustrate the patterns of change in salivary proteins for each expedition and for all expeditions over the year. Figure 3 represents the mean expedition change in salivary protein concentration for all expeditions at each month. For each expeditioner at each month, their yearly mean concentration was subtracted from their monthly concentration to give the amount that their monthly concentration was above or below their individual mean concentration at that month. For each expedition, the concentration differences were then averaged for each month. This indicated when protein concentrations tended to be higher or lower for expeditioners during the year, and was independent of mean expedition concentrations. One-way ANOVA analysis was used to compare the monthly mean concentration changes. Results of the repeated measures ANOVA indicated if concentrations were different between the months (e.g. geometric mean monthly concentrations for IgA in Table 2 and IgM in Table 4), but were not as useful for illustrating the specific nature of the differences. Results The median concentrations and ranges of the salivary proteins measured for all samples collected in this study are presented in Table 3. The salivary protein concentrations summarized by station and year of expedition illustrate the station and expedition differences. The median values in Table 3 correspond well to the ANOVA results presented as geometric mean values in the result sections. Differences in salivary proteins between the stations The average salivary IgA levels for all expeditioners at Casey in 1992, 1993 and 1994 (gm = 67.8 mg/L, 95% CI 63–73 mg/ L) were not significantly different to the average salivary IgA levels for all expeditioners at Mawson in 1992, 1995 and 1996 (gm = 65.6 mg/L, 95% CI 61–71 mg/L, P = 0.55). Overall, salivary IgM levels were higher at Casey (gm = 4.6 mg/L, 95% CI 3.7–5.6 mg/L) than at Mawson (gm = 3.1 mg/L, 95% CI 2.6–3.8 mg/L, P = 0.01), particularly in the Casey 1993 expedition (Fig. 1). The salivary IgG levels were not significantly different between Casey (gm = 12.7 mg/L, 95% CI 11–15 mg/L) and Mawson (gm = 10.6 mg/L, 95% CI 9–13 mg/L P = 0.15). Salivary albumin levels were on average higher at Casey (gm = 46.2 mg/L, 95% CI 39–55 mg/L) compared to Mawson (36.4 mg/L, 95% CI 31–43 mg/L, P = 0.04), due mainly to lower average levels recorded for expeditioners at Mawson in 1996 (Fig. 2). In summary, the mean salivary IgM and albumin levels were higher at Casey compared to Mawson, but the mean salivary IgA and IgG levels did not differ significantly. 386 JL Francis et al. Figure 3 Distributions of the mean monthly expedition changes in IgA, IgM, IgG and albumin concentrations for the expeditions at Casey in 1992, 1993, 1994 and at Mawson in 1992, 1995, 1996. Table 3 Median concentrations and ranges of salivary IgA, IgG, IgM and albumin in samples collected at Casey 1992, 1993, 1994 and Mawson 1992, 1995, 1996 Station and year Casey 1992 1993 1994 All Casey Mawson 1992 1995 1996 All Mawson Number of samples† IgA (mg/L) IgG (mg/L) IgM (mg/L) Albumin (mg/L) 192 110 79 381 57.0 (12–175) 83.4 (38–192) 67.7 (25–151) 66.4 (12–192) 15.1 (1–39) 17.0 (2–45) 8.8 (2–46) 15.1 (1–46) 3.9 (0–22) 6.3 (1–39) 3.9 (1–38) 4.6 (0–39) 47.6 (10–281) 45.8 (12–278) 36.5 (12–166) 43.3 (10–281) 278 66 106 450 67.5 (22–216) 65.6 (18–237) 59.6 (20–164) 65.8 (18–237) 17.0 (2–68) 15.9 (1–38) 5.0 (0–40) 13.8 (0–68) 3.2 (0–30) 3.5 (0–26) 2.0 (0–46) 3.0 (0–46) 41.2 (9–252) 66.3 (18–169) 20.0 (4–164) 38.7 (4–252) †This is the number of samples for IgA. Occasionally the number for other proteins is slightly less as indicated in the footnote to Table 1. Differences in salivary proteins between expeditions at the same station Immunoglobulin A Monthly mean log salivary IgA levels (Table 2) were significantly higher for expeditioners at Casey (Fig. 1) in 1993 than in 1992 or 1994 (1992 g = 55.5 mg/L, 1993 g = 85.2 mg/L, 1994 g = 65.9 mg/L, P < 0.0001). Mean log salivary IgA levels were more consistent in the Mawson expeditioners (Fig. 2) with no significant differences between 1992, 1995 and 1996 (1992 g = 70.3 mg/L, 1995 g = 62.2 mg/ L, 1996 g = 60.3 mg/L, P = 0.13). Immunoglobulin M There were no significant differences in mean log salivary IgM levels (Table 4) between the Trends in mucosal immunity in Antarctica 387 Table 4 Geometric mean salivary IgM (mg/L) concentrations for each month of data collection for each year and station. The months with lower station levels (January – April and predeparture) are bold Station and year Casey Mawson 1992 1993 1994 All years 1992 1995 1996 All years 2.44 2.46 3.35 2.69 2.23 4.18 3.03 5.64 4.31 5.21 4.62 5.47 3.67 3.67 – 5.26 5.16 6.89 8.00 7.39 9.03 – 5.26 4.71 8.25 – – 6.55 3.78 – 3.82 – – 4.10 – – 4.39 – 5.99 – – 4.31 3.29 3.49 3.97 4.01 3.71 5.00 5.16 5.64 4.57 5.00 5.87 5.47 3.67 4.48 2.46 3.60 2.86 2.34 2.83 4.53 5.42 4.39 3.46 2.77 3.42 3.74 – 3.35 4.81 5.21 3.94 3.22 3.00 3.03 5.64 3.46 3.60 3.46 – – – 3.86 – – – – – 4.31 2.75 2.27 3.10 2.56 3.42 – 2.41 2.86 2.92 3.82 3.00 2.48 2.86 4.22 4.26 3.39 3.35 2.80 3.42 3.74 2.41 3.32 ‡ Month 0 1 2 3 4 5 6 7 8 9 10 11 12 All months ‡ Month 0 = Pre-departure samples collected in Hobart in December of the previous year. expeditions at Casey (Fig. 1) (1992 g = 3.7 mg/L, 1993 g = 6.4 mg/L, 1994 g = 4.0 mg/L). Similar to salivary IgA levels, there was a trend for mean log salivary IgM levels to be higher in Casey expeditioners in 1993 (P = 0.06). No statistically significant differences in mean log salivary IgM levels were detected for expeditioners at Mawson (Fig. 2) during the three study years (1992 g = 3.4 mg/L, 1995 g = 3.9 mg/L, 1996 g = 2.6 mg/L, P = 0.33). Immunoglobulin G At Casey the mean log salivary IgG levels (Fig. 1) were on average lower (P < 0.002) for expeditioners in 1994 compared to those in 1992 and 1993 (1992 g = 14.3 mg/L, 1993 g = 16.1 mg/L, 1994 g = 9.0 mg/L). At Mawson (Fig. 2), the average salivary IgG levels were lower (P < 0.001) for expeditioners in 1996 compared to 1992 and 1995 (1992 g = 15.9 mg/L, 1995 g = 12.0 mg/L, 1996 g = 5.3 mg/L). Albumin The mean log salivary albumin levels were not significantly different across the three expedition years at Casey (Fig. 1). (P = 0.49) (1992 g = 51 mg/L, 1993 g = 47 mg/L, 1994 g = 41 mg/L). The average albumin levels at Mawson (Fig. 2) followed the same pattern as the IgG levels, with the average albumin level for expeditioners in 1996 significantly lower (P < 0.001) than those reported in previous expedition years (1992 g = 43 mg/L, 1995 g = 59 mg/L, 1996 g = 22 mg/L). In summary, there were differences in mean salivary IgA and IgG levels between the expeditions at Casey, and differences in mean salivary IgG and albumin levels between the expeditions at Mawson. Comparison of monthly salivary protein levels and changes over the expedition year Immunoglobulin A When each expedition was analysed separately, there were significant differences in mean monthly log IgA levels during the year (Table 2) for the expeditioners at Casey in 1994 (H-F corrected P = 0.001) and at Mawson in 1992 (H-F P = 0.003) and 1996 (H-F P = 0.03). The differences in monthly IgA levels during the year were not statistically significant after H-F correction in the other expeditions (Casey 1992 H-F P = 0.12, Casey 1993 H-F P = 0.15, Mawson 1995 H-F P = 0.12). The lowest IgA concentrations generally occurred between March and June though the patterns in mean monthly log IgA levels for each expedition varied. The person-centred monthly salivary IgA values were significantly different between the months for Casey 1993 (P = 0.016), Casey 1994 (P = 0.0009), Mawson 1992 (P = 0.0002), Mawson 1995 (P = 0.001) and Mawson 1996 (P = 0.002). The trend was the same for Casey 1992 though the salivary IgA values were not significantly different (P = 0.06). In addition to the tendency for the lowest mean IgA concentrations to occur between March and June (Fig. 3), the personcentred IgA values indicated that in the Mawson expeditions IgA levels tended to be higher in August and September. This pattern of higher IgA levels in later months was not consistent in the Casey expeditions. When the data from all expeditions were combined for analysis, the mean monthly log salivary IgA levels tended to be different over the year (H-F P = 0.058), as were the person-centred monthly IgA changes (P = 0.0005). Immunoglobulin A levels were lower in March, April and May compared to the other months of the year when assessed using both mean monthly log salivary IgA levels (Table 2) and person-centred monthly changes in salivary IgA concentration (Fig. 3). Immunoglobulin M The monthly log salivary IgM levels (Table 4) were statistically different across the year for expeditioners at Casey (H-F P = 0.006) and Mawson (H-F P = 0.0005) in the 1992 expeditions only. The differences in monthly IgM levels during the year were not statistically significant in the other expeditions (Casey 1993 H-F P = 0.14, Casey 1994 H-F P = 0.25, Mawson 1995 H-F P = 0.32, Mawson 1996 H-F P = 0.41). The pattern of lower monthly log IgM levels in the first 4 months of the year was observed for all expeditions. The person-centred salivary IgM values were significantly different across the months for expeditioners at Casey in 1992 (P < 0.0001) and 1993 (P = 0.03), 388 JL Francis et al. and expeditioners at Mawson in 1992 (P < 0.0001). The sparse data for Casey in 1994 and the small sample size for Mawson in 1995 reduced the power to detect differences in these years. The changes in monthly person-centred salivary IgM across the year for each expedition showed similar though not identical trends. Salivary IgM levels were generally lower in the first 4 months of the year, with peak values reached in May, June or July (Fig. 3). When the data from all of the expeditions were combined in analyses, the mean monthly log salivary IgM concentrations (H-F P = 0.025) and person-centred monthly changes in salivary IgM (P < 0.0001) demonstrated significant differences over the year. The pattern of lower salivary IgM levels early in the year (Table 4) was similar to that observed for salivary IgA levels, though salivary IgM levels tended to be low from the start of the expedition and peak salivary IgM levels were consistently observed mid-year (Fig. 3). Immunoglobulin G There were no significant differences in mean monthly log salivary IgG levels across the year (data not shown) during any of the Casey expeditions (Casey 1992 H-F P = 0.27, Casey 1993 H-F P = 0.34, Casey 1994 H-F P = 0.55). There were small but statistically significant differences in monthly log salivary IgG levels during the year at Mawson in 1992 (H-F P = 0.01) and 1996 (H-F P = 0.03). There were no significant differences in monthly log salivary IgG levels during the year for Mawson in 1995 (H-F P = 0.12). There were no significant differences between person-centred monthly salivary IgG values across the year for any of the Casey expeditions (Casey 1992 P = 0.092, Casey 1993 P = 0.53, Casey 1994 P = 0.47). At Mawson in 1992, the mean predeparture levels of IgG were significantly higher than those for the rest of the expedition year (P = 0.001). In the 1995 and 1996 Mawson expeditions, there were statistically significant differences between the 2 months with the largest differences in value only (Mawson 1995 P = 0.001, Mawson 1996 P = 0.035). There were no particular patterns to the monthly changes (Fig. 3). When data from all of the expeditions were combined for analyses, the mean monthly log salivary IgG levels were not significantly different over the year (H-F P = 0.13). The person-centred monthly IgG changes were significantly different (P = 0.02), due primarily to the dominance of the Mawson 1992 data with higher predeparture IgG levels. There was no trend over the rest of the year. Albumin The only statistically significant difference in mean monthly log salivary albumin levels across the year (data not shown) was for expeditioners at Mawson in 1996 (H-F P = 0.03). The mean log salivary albumin level at Mawson in December 1996 was higher than the levels for the previous 7 months. There were no significant differences for the other expeditions (Casey 1992 H-F P = 0.36, Casey 1993 H-F P = 0.44, Casey 1994 H-F P = 0.60, Mawson 1992 HP = 0.24, Mawson 1995 H-F P = 0.16). There were no statistically significant differences between months for personcentred monthly albumin changes for any of the expeditions (Fig. 3) (Casey 1992 p = 0.72, Casey 1993 P = 0.58, Casey 1994 P = 0.70, Mawson 1992 P = 0.15, Mawson 1995 P = 0.27, Mawson 1996 P = 0.14). When the data from all of the expeditions was combined for analyses, neither the mean monthly log salivary albumin levels (H-F P = 0.25) nor the person-centred monthly albumin changes (P = 0.42) were significantly different across months of the year, and there were no particular patterns to the mean monthly albumin levels or the mean monthly albumin changes over the year (Fig. 3). In summary, during the expedition year, mean monthly IgA concentrations were consistently lowest in March, April and May, and mean monthly IgM concentrations were lowest from January to April and highest in May, June or July There were no trends in mean monthly IgG or albumin levels. Discussion The most significant finding from this study was that the patterns of change in monthly salivary IgA and IgM concentrations observed for the 1992 Antarctic expeditioners were reinforced.1 The patterns of lower salivary IgA in March–May and lower IgM levels in the first 4 months of the year compared to subsequent months were generally consistent over the six expeditions studied. Peak salivary IgM levels invariably occurred during June and July in each expedition. Salivary IgA levels were back up to mean levels or reached peak levels in July to September. These patterns were observable for most expeditions, even though in separate expedition analyses they were not always statistically significant. When the data were summarized by year of expedition, giving each year equal weight, regardless of the number of expeditions or expeditioners, the patterns persisted, making it difficult to discount them as a manifestation of random variation. The patterns of change in salivary IgA concentrations are consistent with a response to the psychological stresses over the year in Antarctica.1 Initially at each station, new summer and winter expeditioners are present, and this has been reported to result in some discordance. Isolation of the winter expeditioners generally commences in February or March when the summer expeditioners depart, and after this departure the winter expedition group dynamics tend to stabilize (J. Wood pers.comm. 2000). A 4-week Norwegian isolation study of group interaction and dynamics similarly found that initially participants experienced the team as positive and cheerful, followed by a period of dissatisfaction with group management and organization, before individual mood and team function stabilized in a positive manner. 23 Studies of depressive symptoms in US Antarctic winter expeditioners indicate that even clinically normal individuals experience an increase in depressive symptoms in high-latitude environments during winter months and that these symptoms tend to peak mid-winter.24 Daily mood fluctuations have been associated with changes in salivary IgA, in particular, increased negative moods were associated with a concurrent reduction in salivary IgA.25 The characteristic psychological changes that occur in Antarctica are those that correspond to changes in salivary IgA levels. The relative stress within each Antarctic expedition varies. Each group of expeditioners has unique psychological characteristics and social dynamics. Some expeditions have more disharmony than others, either prior to the departure of the summer expeditioners or during the period of winter isolation. Individual psychological factors determine both the physical Trends in mucosal immunity in Antarctica and psychological response of that individual. 23 Not all stressors are equally influential on all individuals, and individuals cope differently.26 The scientific program to be undertaken during each expedition also results in different stressors, with varying degrees of physical and psychological stress imposed on the expeditioners. Some personnel have a high degree of exposure to extremes of temperature and demands of intense levels of physical exertion compared to others. The degree of difficulty in carrying out the winter’s programmed tasks is greatly influenced by these physical and climactic considerations. When evaluating the changes associated with response to isolation stressors, the changes within each individual are therefore most revealing. The importance of an individual’s change in mucosal immune status over time has been established in research with elite athletes in response to the stresses of competitive training.19,27 Markedly elevated levels of IgM at mucosal surfaces occur in individuals with salivary IgA deficiency, the extreme of impaired immune function.17,20 This is thought to be an adaptive effect against potential mucosal infection. The peak in salivary IgM levels observed in the Antarctic expeditioners following months of low salivary IgA levels may reflect a similar compensatory mechanism. This relationship was also apparent in a 7-month study of elite swimmers who experienced increased levels of salivary IgM coincident with suppressed salivary IgA following long-term training. 28 Salivary IgM concentrations normally only increase in response to antigenic stimulus such as mucosal infection. However, this explanation of the peak IgM levels observed in the Antarctic expeditioners in June or July each year is unlikely, as the levels of clinical illness during the Antarctic expeditions were very low (ANARE Health Register). It remains possible that subclinical events such as viral shedding may be associated with the changes in salivary IgM levels. Significant increases in EBV shedding and decreased cellular immunity have been observed at Casey,7 Davis and Mawson29 during the winter isolation with no associated clinical symptoms. The peak in IgM levels in the mid-winter solstice period may have an association with changes in mood state. Plasma levels of IgM have been found to be significantly increased for Norwegian summer expeditioners compared to pre- and post-expedition levels23 and a consistent correlation has been reported between plasma concentrations of IgM and psychological stress. 30 Expeditioners from different stations and study years had similar average salivary protein concentrations with only minor exceptions for IgA at Casey 1993 and IgM for Mawson 1996. Salivary IgG and albumin concentrations were similar on average between participants in all expeditions except for lower levels at Mawson in 1996. The observed differences between expeditions were small from a clinical perspective. These expedition differences may be due to specific expedition factors, but are more likely due to the different cohorts of individuals that comprised each expedition, as salivary immunoglobulin levels have a high variability between individuals compared to the variability of repeated measures within individuals.17 Statistically significant differences in concentrations between groups are possible while immunoglobulin levels remain within normal population reference limits. 23 The patterns of fluctuation in salivary IgA and IgM concentrations detected so strongly in the 1992 expeditions were more difficult to examine in the subsequent expedition 389 years due to constraints on the data collection. Assuming the patterns of change in salivary IgA and IgM levels are a reflection of psychological stress, the strength of the changes may also have been affected by the improvement in telecommunications between the Antarctic stations and the Australian mainland, reducing the overwhelming sense of isolation and associated stress. However, persistent alterations in cellular immunity subsequent to these improvements do suggest that this effect was minimal.7 Only a few investigations of immunity in the Antarctic have found evidence of variation in systemic immunity during expeditions. This may be due to inadequate frequency and timing of the measures, or inappropriate analysis.31 Repeated measures on individuals enables the direct study of the changes in those measures, given the appropriate design and statistical methods. 32 The fluctuations in mucosal immunity found in this study are likely to be associated with the isolation stressors common to all expeditioners over the year as opposed to individual expeditionerspecific stressors. Further investigation of the relationship of mucosal immune parameters with individual factors such as health and psychological status, and expedition-related factors of meteorological and ultraviolet radiation patterns may increase our understanding of the patterns of change in mucosal immune parameters during the Antarctic winter. Acknowledgements This work was supported by the Antarctic Science Advisory Committee (ASAC) Grants Scheme, The Australian Antarctic Division and the study was approved by The Australian Antarctic Division Ethics Committee (Human Experimentation). The authors are grateful to the ANARE members who participated in this study during the winters of 1992, 1993 and 1994 at Casey Station and 1992, 1995 and 1996 at Mawson Station. The authors would like to thank Ms Sharron Hall and Ms Deborah Capper of the Hunter Immunology Unit who performed the saliva assays, and Mr Adrian Flanagan (Hunter Immunology Unit) and Dr Peter Sullivan (Polar Medicine, Antarctic Division) for their assistance with collation of data. References 1 Gleeson M, Francis JL, Lugg DJ et al. A year in Antarctica: mucosal immunity at three Australian stations. Immunol. Cell Biol. 2000; 78: 616–22. 2 Lugg DJ. Current international human factors research in Antarctica. In: Harrison, AA, Clearwater, YA, McKay, CP (eds). From Antarctica to Outer Space: Life in Isolation and Confinement. New York: Springer-Verlag, 1990; 31–42. 3 Konstantinova IV, Fuchs BB. The Immune System in Space and Other Extreme Conditions. Chur, Switzerland: Harwood Academic Publishers, 1991. 4 Williams DL, Climie A, Muller HK, Lugg DJ. Cell-mediated immunity in healthy adults in Antarctica and the subantarctic. J. Clin. Laboratory Immunol. 1986; 20: 43–9. 5 Muller HK, Lugg DJ, Quinn D. Cell mediated immunity in Antarctic wintering personnel; 1984–92. Immunol. Cell Biol. 1995; 73: 316–20. 6 Pitson GA, Lugg DJ, Muller HK. Seasonal cutaneous immune responses in an Antarctic wintering group: No association with 390 7 8 9 10 11 12 13 14 15 16 17 18 JL Francis et al. testosterone, vitamin D metabolite or anxiety score. Arct. Med. Res. 1996; 55: 118–22. Tingate TR, Lugg DJ, Muller HK, Stowe RP, Pierson DL. Antarctic isolation: Immune and viral studies. Immunol. Cell Biol. 1997; 75: 275–83. Muller HK, Lugg DJ, Ursin H, Quinn D, Donovan K. Immune responses during an Antarctic summer. Pathology 1995; 27: 186–90. Gleeson M, Cripps AW, Clancy RL. Modifiers of the human mucosal immune system. [Review] [49 refs]. Immunol. Cell Biol. 1995; 73: 397–404. Ng V, Koh D, Chan G, Ong HY, Chia SE, Ong CN. Are salivary immunoglobulin A and lysozyme biomarkers of stress among nurses? J. Occup. Environ. Med. 1999; 41: 920–7. Gleeson M, Pyne DB. Exercise effects on mucosal immunity. Immunol. Cell Biol. 2000; 78: 536–44. Graham NMH, Bartholomeusz RCA, Taboonpong N, La Brooy JT. Does anxiety reduce the secretion rate of secretory IgA in saliva? Med. J. Aust. 1988; 148: 131–3. Gleeson M. Mucosal immune responses and risk of respiratory illness in elite athletes. Exercise Immunol. Rev. 2000; 6: 5–42. MacKinnon LT. Exercise and Humoral Immunity: Immunoglobulin, Antibody, and Mucosal Immunity. Advances in Exercise Immunology. Champaign, USA: Human Kinetics 1999; 159–200. van Rood YR, Bogaards M, Goulmy E, van Houwelingen HC. The effects of stress and relaxation on the in vitro immune response in man: a meta-analytic study. J. Behav. Med. 1993; 16: 163–81. Jemmott JBI, Borysenko JZ, Borysenko M et al. Academic stress, power motivation, and decrease in secretion rate of salivary secretory immunoglobulin A. The Lancet 1983; 25: 1400–2. Gleeson M, Cripps AW, Clancy RL, et al. The variability of immunoglobulins and albumin in saliva of normal and IgAdeficient adults. In: MacDonald, TT, Challacombe, SJ, Bland, PW, Stokes, CR, Heatley, RV, Mowat, AM (eds). Advances in Mucosal Immunology. Lancaster (UK): Kluwer Academic Press, 1990; 500–1. Gleeson M, Clancy RL, Cripps AW, Henry RL, Hensley MJ, Wlodarczyk JH. Acquired IgA deficiency. Pediatr. Allergy Immunol. 1994; 5: 157–61. 19 Gleeson M, McDonald WA, Pyne DB et al. Salivary IgA levels and infection risk in elite swimmers. Med. Sci. Sports Exercise. 1999; 31: 67–73. 20 Brantzaeg P, Farstad IN, Johansen F-E, Morton HC, Norderhaug IN, Yamanaka T. The B-cell system of human mucosae and exocrine glands. Immunol. Rev. 1999; 171: 45–87. 21 JMP Statistical Discovery Software (Computer program). Version 3 Gary, USA: SAS Institute Inc, 1999. 22 StataCorp. (Stata Statistical Software) (Computer program). Version 7.0 College Station, TX: Stata Corporation, 2001. 23 Ursin H. Psychobiological studies of individuals in small, isolated groups in the Antarctic and in space analogues. Environ. Behav. 1991; 23: 766–81. 24 Palinkas LA, Cravalho M, Browner D. Seasonal variation of depressive symptoms in Antarctica. Acta. Psychiatr. Scand. 1995; 91: 423–9. 25 Stone AA, Cox DS, Valdimarsdottir H, Jandorf L, Neale JM. Evidence that secretory IgA antibody is associated with daily mood. J. Pers. Soc. Psychol. 1987; 52: 988–93. 26 Wood J, Lugg DJ, Hysong SJ, Harm DL. Psychological changes in hundred-day remote Antarctic field groups. Environ. Behav. 1999; 31: 299–337. 27 Gleeson M, Ginn E, Francis JL. Salivary immunoglobulin monitoring in an elite kayaker. Clin. J. Sport Med. 2000; 10: 206–8. 28 Gleeson M, McDonald WA, Cripps AW, Pyne DB, Clancy RL, Fricker PA. The effect on immunity of long-term intensive training in elite swimmers. Clin. Exp. Immunol. 1995; 102: 210–16. 29 Mehta SK, Pierson DL, Cooley H, Dubow R, Lugg DJ. EpsteinBarr virus reactivation associated with diminished cell-mediatied immunity in Antarctic expeditioners. J. Med. Virol. 2000; 61: 235–40. 30 Endreson IM, Relling GB, Tonder O, Myking O, Walther BT, Ursin H. Brief uncontrollable stress and psychological parameters influence human plasma concentrations of IgM and complement component C3. J. Behav. Med. 1991–2; 17: 167–76. 31 Mishra SK, Segal E, Gunter E et al. Stress, immunity and mycotic diseases. J. Med. Vet. Mycol. 1994; 32: 379–406. 32 Diggle PJ, Liang K-V, Zeger SL. Analysis of Longitudinal Data. Oxford: Clarendon Press, 1994.