SEARCH

SEARCH BY CITATION

Keywords:

  • Canine;
  • Endurance;
  • Globulin;
  • IgG

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

Background: Serum immunoglobulin dynamics have not been studied in racing sled dogs, despite hypoglobulinemia having been reported during racing events.

Hypothesis/Objectives: Hypoglobulinemia in racing sled dogs is associated with decreases in serum IgA, IgE, IgG, and IgM concentrations during prolonged exercise.

Animals: One hundred and fifty-seven Alaskan sled dogs that successfully completed a 1,000 mile race.

Methods: Serum was obtained from 118 sled dogs within 1 month before the race and within 12 hours after completing the race. Serum also was obtained after 4 months of rest from 51 dogs that successfully completed the race, including 12 previously sampled dogs. Serum total protein ([TP]), albumin, and globulin ([Gl]) were measured, and serum IgA, IgE, IgG, and IgM were quantified by ELISA.

Results: The proportion of dogs with [Gl] ≤ 2.2 g/dL was significantly greater immediately after racing (38 of 118 dogs, 32.2%) than before racing (21 of 118 dogs, 17.8%, P= .005). Four months after racing, [Gl] was ≤ 2.2 g/dL in 23.5% (12 of 51) of dogs. [IgG] was significantly lower before (8.21 ± 4.95 mg/mL) and immediately after (7.97 ± 5.62) racing compared with 4 months after racing (18.88 ± 5.76). Serum [IgM] and [IgE] were higher and [IgA] was lower before racing compared with immediately after racing.

Conclusions and Clinical Importance: Sled dogs participating in long-distance racing have substantial decreases in [IgG] in addition to decreases in [IgM] and [IgE]. The pronounced hypogammaglobulinemia observed in a large proportion of racing sled dogs might predispose them to infectious disease.

Abbreviations:
[Gl]

serum total globulin concentration

[IgA], [IgE], [IgG], and [IgM]

serum immunoglobulin A, E, G, and M concentrations, respectively.

High intensity or long duration exercise can substantially impact immune function.1,2 In the last 2 decades, a large number of studies have documented the effects of training and exercise on immune function in human athletes.2 However, similar studies of animal athletes have been comparatively limited, and have pertained almost exclusively to horses.3–6 Despite recognition that distance running alters immune function in human athletes, few studies have specifically investigated the impact of exercise on immune variables in racing sled dogs.7–11 Elite sled dogs are at risk of changes in immune function because they participate in running events exceeding 1,000 miles in length, frequently preceded by several thousand miles of running training.12

Physical conditioning and athletic competition decrease serum total globulin ([Gl]) and immunoglobulins A, G, and M (IgA, IgG, and IgM) in human athletes.13–17 Mild to moderate hypoglobulinemia occurs in sled dogs training for and participating in long-distance races, and trained elite Alaskan sled dogs have been reported to have substantially lower [Gl] than sedentary dogs.18–20 Furthermore, in 1 study, participation of trained sled dogs in a simulated 500 mile race resulted in a decrease in [Gl] to concentrations comparable to those recorded in dogs with severe protein-losing gastrointestinal disease.20,21 The prevalence and clinical relevance of hypoglobulinemia in trained and racing sled dogs have not been clearly delineated, although dogs participating in racing events sometimes develop infectious diseases including pneumonia and diarrhea.22 More importantly, the impact of training and racing on specific immunoglobulin fractions in sled dogs has not been investigated. Determining these effects may assist the development of management interventions to decrease the potentially detrimental impact of training and competition on immune function in sled dogs. Therefore the purpose of the current study was to determine serum IgA, IgE, IgG, and IgM concentrations in trained Alaskan sled dogs before and within 12 hours after successfully completing the 2007 Iditarod Trail Race, and after 4 months of rest from training.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

One thousand three hundred and three dogs comprising 82 teams (maximum, 16 dogs per team) started the 2007 Iditarod Trail Race, which was approximately 1,000 miles in length. Before the start of the race, all dogs underwent physical examination, and CBC, blood biochemistry, and ECG examinations were performed. Blood samples were obtained by jugular venipuncture into 10 mL serum separator tubesa and were stored at − 20°C. Subsequently, a 2nd blood sample was similarly obtained from a convenience sample of 118 dogs within 12 hours after the dogs completed the race, and samples were retrospectively matched with each dog's prerace sample. This group of dogs consisted of 40% females and 60% males, with an age distribution (mean ± SD) of 4.2 ± 1.7 years. These dogs represented 13 different teams, and finished the race within the 1st 45 places in 9–12 days.

Four months after completion of the race, venous blood was similarly obtained from 51 dogs that had successfully completed the race and had been rested from training since the race. Because of geographic dispersion and logistical difficulties gaining access to previously sampled dogs, this group contained only 12 dogs from 2 teams within the original convenience sample group, and 39 dogs from an additional 4 teams that successfully completed the 2007 Iditarod, but that were not sampled before or immediately after the race. This group of dogs consisted of 47% females and 53% males with an age distribution of 4.4 ± 1.7 years. This group was derived from 6 different kennels, and the teams in this group finished the race within the 1st 50 places in 10–14 days.

Samples from dogs collected before, immediately after, and 4 months after the race were, respectively, labeled trained (TR; n= 118 dogs), raced (RA; 118), and rested (RE; 51). All samples were analyzed within 6 months of collection.

The study was approved by the Institutional Animal Care and Use Committee at Oregon State University. Participating mushers signed an informed consent form before initiation of the study.

Serum total protein ([TP]) and albumin concentrations were determined in all groups by chemical methods on a commercially available chemistry analyzer.b Serum total globulin concentration ([Gl]) was subsequently calculated for all dogs by subtracting each individual dog's measured albumin concentration from their measured [TP].

Serum IgA, IgE, IgG, and IgM concentrations were determined in 100 randomly selected dogs in the TR group, in the same 100 dogs in the RA group, and in 50 randomly selected dogs in the RE group, including 8 dogs contained within the TR and RA groups. Data pertaining to IgA and IgM was not available for 2 dogs in the TR and RA group, and for 1 dog for IgE and IgG in the TR and RA group (therefore n= 98 for IgA and IgM, and n= 99 for IgE and IgG). To determine correlation between serum and salivary IgA concentrations in racing dogs, cotton swabs were used to obtain saliva samples from 67 dogs in the RA group at the time that serum was collected. Swabs of saliva were only obtained from dogs that had not eaten in the previous 2 hours. Swabs of saliva were not obtained from dogs in the TR or RE groups.

Serum immunoglobulin concentrations were determined with commercially available canine specific ELISA kits according to the manufacturer's instructions.c,d Briefly, 96-well microtiter plates were coated overnight at 4°C with a 1 : 100 dilution of the appropriate goat anti-dog immunoglobulin isotype antibody. Plates were washed 3 times and treated with blocking solution at room temperature for 30 minutes, followed by 3 additional washes. Serum samples were diluted in sample dilution buffer at 1 : 1,000–1 : 20,000 for IgA, 1 : 100–1 : 1,000 for IgE, 1 : 10,000–1 : 1,000,000 for IgG, and 1 : 1,000–1 : 20,000 for IgM. Mucosal swabs for salivary IgA analysis were resuspended in 1.0 mL dilution buffer, vortexed for 10 seconds and then diluted 1 : 1,000–1 : 20,000 in dilution buffer. One-Hundred microliters of diluted sample was added to duplicate microtiter wells and incubated at room temperature for approximately 1 hour. A 7-point standard curve, in addition to blank negative and positive controls, was included in duplicate for each immunoglobulin isotype. After incubation, plates were washed 5 times and the appropriate horseradish peroxidase (HRP)-conjugated goat anti-dog immunoglobulin isotype diluted 1 : 20,000 in dilution buffer was added to each well. Plates were incubated for 1 hour at room temperature and washed an additional 5 times. After washing, HRP substrate was added to each well and plates were incubated until the highest standard was deep blue (approximately 5–10 minutes). The reaction was quenched with 1 M H2SO4. Plates were read at 450 nM with an ELISA plate readere,f and the concentration of the sample determined based on standard curve values and sample dilution. All samples were analyzed in duplicate. Samples were retested when coefficients of variation exceeded 15%.

Data were analyzed by ANOVA techniques in an unbalanced, incomplete block design to account for the unusual distribution of the groups.g Results are reported as least square means ± SD. Where appropriate, a McNemar's test for correlated proportions was applied to compare the TR and RA groups (matched samples). Variables were assessed for an effect of team by running a Covtest option in Proc Mixed. Additionally, because some of the dogs in the RE group were not identical animals to those in the TR and RA groups, data from dogs present in all 3 groups was analyzed by repeated measures ANOVA to determine if the results were representative of the main data set.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

Serum Total Protein, Albumin, and Globulin

Serum [TP] and albumin concentrations were significantly greater in the TR group compared with the RA and RE groups (P < .0001; Table 1). In the RA group, 28.6% (34 of 118 dogs) of dogs had serum [TP] < 5.5 g/dL as compared with only 2.5% (3 of 118) of dogs in the TR group (P < .0001) and 1.9% (1 of 51) of dogs in the RE group. Hypoalbuminemia (serum albumin concentration ≤ 2.9 g/dL) was observed in 6.8% (8 of 118 dogs) of dogs in the RA group, but in no dogs in the TR and RE groups. There was a significant effect of team on [TP] (P= .038) and albumin (P= .025) concentration. Repeated measures ANOVA of the data from the 12 dogs in the TR, RA, and RE groups revealed that serum [TP] was significantly greater at TR and RE (6.4 ± 0.4 and 6.2 ± 0.5 g/dL, respectively) compared with RA (5.8 ± 0.4 g/dL). Serum albumin concentration was significantly greater at TR (4.0 ± 0.3 g/dL) compared with RE (3.6 ± 0.2), and RA results were significantly lower than those of both the TR and RE groups (3.4 ± 0.2 g/dL).

Table 1.   Serum total protein, albumin and total globulin concentrations and albumin : globulin ratio in sled dogs participating in a long-distance race.
VariableTrainedRacedRestedReference Range
  1. Range and median displayed in parentheses.

  2. Trained, raced: n= 118; rested: n= 51.

  3. Different superscript letters indicate significantly (P < .05) different values within a row.

  4. Reference ranges provided by the laboratory performing the biochemical analysis. Derived from sedentary dogs of several breeds.

Total protein (g/dL)6.4 ± 0.5a5.8 ± 0.5c6.0 ± 0.4b5.1–7.1
(5.3–8.2, 6.4)(4.3–7.4, 5.8)(5.0–6.9, 6.1)
Albumin (g/dL)3.9 ± 0.3a3.3 ± 0.3c3.5 ± 0.2b2.9–4.2
(3.3–4.6, 3.8)(2.6–4.0, 3.4)(3.1–4.2, 3.6)
Globulin (g/dL)2.5 ± 0.4a2.4 ± 0.4b2.6 ± 0.3a2.2–2.9
(1.9–4.1, 2.5)(1.7–3.5, 2.4)(1.7–3.2, 2.5)
AG ratio1.5 ± 0.2a1.4 ± 0.2b1.4 ± 0.2a0.8–2.2

Serum [Gl] was significantly lower in the RA group compared with the TR and RE groups (P= .013; Table 1). Before racing, low [Gl] (≤ 2.2 g/dL) was recorded in 17.8% (21 of 118) of dogs. Immediately after racing, the proportion of dogs with [Gl] ≤ 2.2 g/dL increased to 32.2% (38 of 118), which was a significantly (P= .0052) greater proportion of dogs compared with the TR group. Four months after racing, [Gl] was ≤ 2.2 g/dL in 23.5% (12 of 51) of dogs. The albumin : globulin ratio was significantly (P < .0001) lower in RA compared with TR dogs (Table 1). For the 12 dogs present in all 3 groups, [Gl] also was lower in RA (2.4 ± 0.3 g/dL) than in TR and RE (2.5 ± 0.2 and 2.7 ± 0.4 g/dL, respectively), however differences were not significant (P= .07). There was no significant effect of team on [Gl] (P= .068) or albumin : globulin ratio (P= .073).

Serum Immunoglobulins and Salivary [IgA]

[IgG] was significantly (P < .0001) higher in RE dogs than in TR and RA dogs (Fig 1). [IgG] ranged from 0.8–21.3, 1.5–30.2, and 7.9–39.5 mg/mL in the TR, RA, and RE groups, respectively, and median values were 9.8, 9.4, and 17.4 mg/mL, respectively. [IgG] was extremely low (≤ 3 mg/mL) in 31% (31 of 99) of TR dogs and 38% (38 of 99) of RA dogs, but was not significantly different between the 2 groups (P= .07). [IgG] was > 11 mg/mL in all RE dogs tested except for 1 dog with a value of 7.9 mg/mL.

image

Figure 1.  Serum [IgG] in sled dogs participating in a long-distance race. Horizontal bar reflects mean. Trained, raced: n= 99; rested: n= 51. *Significant difference (P < .05) to other groups. Reference range 10–20 mg/mL.23,24

Download figure to PowerPoint

[IgM] and [IgE] were significantly (P < .0001) higher in TR versus RA dogs, but significantly lower than in RE dogs (Table 2). [IgA] was significantly (P= .0048) higher in the RA group compared with the TR group, but neither value was significantly different from the RE group. A significant effect of team was observed for [IgE] (P= .0126) and [IgG] (P= .0465), but not for [IgA] (P= .4659) or [IgM] (P= .0720). Weak positive correlation was documented between [Gl] and [IgG] immediately after racing (n= 97, r2= 0.056, P= .02) and in rested dogs (n= 50, r2= 0.099, P= .03). There was no significant correlation between [Gl] and [IgG] before racing, or between [Gl] and [IgM], [IgA], or [IgE].

Table 2.   Serum immunoglobulin concentrations in sled dogs participating in a long-distance race.
VariableTrainedRacedRestedReference Range
  1. Range and median are displayed in parentheses.

  2. Trained, raced: n= 98 (except for IgE where n= 99); rested: n= 50.

  3. Different superscript letters indicate significantly (P < .05) different values within a row.

  4. NA, not applicable.

IgA (mg/mL)1.23 ± 1.00b1.50 ± 1.07a1.43 ± 0.82ab0.4–1.624
(0.15–7.50, 0.96)(0.23–5.55, 1.25)(0.40–4.29, 1.19)
IgE (mg/mL)0.07 ± 0.07b0.05 ± 0.06c0.12 ± 0.10aNA
(0.01–0.48, 0.07)(0.01–0.36, 0.06)(0.01–0.48, 0.03)
IgM (mg/mL)0.97 ± 0.46b0.88 ± 0.44c1.21 ± 0.39a1.0–2.023,24
(0.21–2.45, 0.85)(0.25–2.07, 0.76)(0.75–2.58, 1.29)

Repeated measures ANOVA of the data from 8 dogs with immunoglobulin concentrations available in the TR, RA, and RE groups revealed that [IgG] (P= .0006) and [IgE] (P= .0006) were significantly lower in TR and RA than in RE ([IgG] 6.17 ± 5.45, 5.17 ± 4.55, and 19.45 ± 8.01 mg/mL, respectively; [IgE] 0.09 ± 0.07, 0.07 ± 0.06, and 0.14 ± 0.10 mg/mL, respectively). Serum [IgA] (P= .998) and [IgM] (P= .68) were not significantly different among groups for these dogs ([IgA] 0.88 ± 0.54, 0.87 ± 0.67, and 0.88 ± 0.48 mg/mL, respectively; [IgM] 1.18 ± 0.28, 1.09 ± 0.58, and 1.20 ± 0.31 mg/mL, respectively).

Salivary [IgA] in 67 dogs in the RA group was 0.26 ± 0.2 mg/mL (range, 0.022–0.75 mg/mL). There was no correlation between serum and salivary IgA concentrations in the sampled dogs (r2= 0.012; P= .38).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

The most striking finding of the current study was the observation of hypogammaglobulinemia in trained and racing sled dogs. Mean [IgG] before and soon after racing was approximately 8 mg/mL, below commonly reported reference values for healthy dogs.23,24 In contrast, rested sled dogs had a mean [IgG] of 18 mg/mL, comparable to values reported for healthy dogs of other breeds.24,25 Furthermore, [IgG] was ≤ 3 mg/mL, a concentration linked with immunodeficiency in other dog breeds, in 30% or more of the dogs studied before and immediately after racing.24–26 In comparison, all but 1 of the rested dogs had [IgG] of 11 mg/mL or greater, within reported reference ranges for healthy dogs of other breeds.24,25 Although a significant effect of team on [IgG] was observed, 11 of the 13 teams studied before and soon after the race had at least 1 dog with [IgG] ≤ 3 mg/mL, suggesting that factors in addition to team and race placing have an effect on [IgG]. Some dogs may be more susceptible to the effects of training and racing on serum [IgG] than others, because 54% (24 of 44 dogs) of the dogs that had a very low serum [IgG] recorded during the current study had very low [IgG] recorded at both the TR and RA time points. Mean and median results of [IgM] for the TR and RA groups also were below reference values reported for healthy dogs, but concentrations for the RE group were within reference ranges.23 Therefore, training for and participating in long-distance running events substantially decreases [IgG] and [IgM] in sled dogs, and in approximately 30% of dogs, hypogammaglobulinemia is substantial and possibly conducive to disease development.

Hypoglobulinemia was not a prominent finding in the current study, in contrast to some previous studies, although different reference ranges have been utilized in the studies.19,20 However, training and racing did cause mild decreases in serum [Gl], and specific individuals may be more substantially affected, particularly during racing. Decreases in serum [Gl] associated with long-distance running exercise might result from alterations to lymphocyte function or number, protein catabolism, plasma volume expansion, and other factors including application of unsuitable reference ranges.27 Wherever possible, breed-specific or activity-specific reference ranges for [Gl] should be applied. Reference ranges derived from truly sedentary sled dogs do not appear to be available, and specific results, such as [TP], albumin, and [Gl] concentrations are affected by training and racing.19,20 Greyhounds, another highly athletic dog breed, have lower [Gl] and α- and β-globulin fractions than sedentary breeds, even after retirement from athletic activities.28 A decrease in [Gl] has been previously linked to increasing fitness in training sled dogs and profound incremental decreases in [Gl] have been documented in sled dogs undergoing consecutive days of prolonged exercise.20,29 This suggests that both breed and activity may have substantial effects on [Gl] in dogs. However, the laboratory specific reference ranges applied in the current study were found to be virtually identical to those derived from a large number of trained sled dogs before participation in the Iditarod Trail Race,30 despite being derived from a population of healthy, sedentary nonsled dogs. Additionally, [IgG] in rested sled dogs in the current study was within reference ranges for nonathletic dog breeds, suggesting that training and racing have a substantial impact on [IgG] in dogs, which has not been previously documented.

The cause of the hypogammaglobulinemia in trained and racing sled dogs in the current study was not determined. In human subjects, even relatively brief periods of restricted energy intake during training negatively impact [IgM] and [IgG], and certainly some dogs may suffer appetite suppression during racing.17 Increased serum urea nitrogen concentrations and attenuation of muscle glycogenolysis have been reported in sled dogs participating in consecutive days of exercise, suggesting increased catabolism of protein for energy.31 Therefore, catabolism of endogenous protein sources may contribute to decreases in serum [Gl] and immunoglobulin concentrations in trained and racing sled dogs. Furthermore, sled dogs likely have some degree of gastrointestinal blood loss during racing that also could result in loss of serum immunoglobulins.12 The decrease in serum [TP] and albumin concentrations observed in the current study are consistent with previous findings in racing sled dogs.18,32 However, the disproportionate nature of the decreases in serum albumin and the various immunoglobulin fractions documented in the current study do not support blood loss as the primary reason for hypogammaglobulinemia.

The substantial increases in plasma volume that can occur with prolonged endurance exercise subsequently alter the concentration of various blood solutes and may result in relative hypoglobulinemia.27 Training of sedentary Beagles and Greyhounds was shown to increase plasma volume by 13.1 and 27.5%, respectively.33,34 However, dogs undergoing a single bout of prolonged exercise in a cool environment did not display changes in plasma volume.35 Furthermore, a 12-week training protocol induced negligible changes in plasma volume in Alaskan sled dogs, which was attributed to the existing fitness of the dogs at the commencement of the study.36 Increases in plasma volume because of endurance training are limited by initial fitness levels, duration of previous exposure and possibly genetic components.37 Although plasma volume alterations were not measured in the current study, the dogs' high degree of fitness may have attenuated major changes in plasma volume, at least from before to soon after the race. Additionally, the magnitude of the increase in [IgG] (exceeding 50%) in the rested dogs is unlikely to be explained by decreases in plasma volume alone. Furthermore, changes in [Gl] during the study were minimal, and serum [IgA] was highest when measured soon after racing; findings that are incongruous with expansion of plasma volume. This suggests that factors other than plasma volume expansion likely contribute to alterations in serum immunoglobulin concentrations in exercising sled dogs.

Although lymphocyte numbers and function were not examined in the present study, a significant decrease in lymphocyte number has been reported in sled dogs training for and participating in a long-distance race.12 In human subjects, blood lymphocyte concentrations are decreased by prolonged endurance exercise, and B lymphocyte function is transiently impaired after exercise.11,38 Exercising sled dogs also show evidence of adrenocortical stimulation, and alterations in serum cortisol concentrations may affect lymphocyte function or number.39

Perturbation of humoral immunity as a result of training and racing could increase susceptibility to infectious disease, which is of particular concern given the frequency with which commingling occurs in the sled dog population. During competition, sled dogs are afflicted with several conditions that are potentially infectious in nature, including respiratory disease and diarrhea.22 A study of 195 sled dogs participating in the 2006 Iditarod Trail Race indicated that approximately 15% of dogs had potentially nonprotective antibody titers to canine distemper virus before the race, despite mandatory requirements for vaccination to occur within 12 months of the race.40 However, antibody titers to canine parvovirus and canine distemper virus were documented to increase substantially from before to after the race in 31 and 17% of dogs, respectively, consistent with antigen exposure, although no dogs developed clinical signs of these diseases.40 Therefore, despite a substantial impact of training and racing on serum immunoglobulin concentrations, racing sled dogs appear to be capable of generating appropriate antibody responses to specific viral pathogens. Similarly, the ability of elite human athletes to mount an adequate antibody response to vaccine antigens is not impaired despite other perturbations of immune function.7,41,42 A weakness of the current study, however, was the inability to study the relationship between clinical disease and [IgG] during the race. Dogs to be studied were selected after successfully completing the race, and obtaining accurate information regarding clinical disease during the race was logistically difficult given the large number of dogs competing and limited access to dogs during the race. However, given the severity of the hypogammaglobulinemia in a substantial proportion of the dogs, further investigation of the pathophysiology and possible clinical impact of this phenomenon are warranted. Such a study should include investigation of dogs that fail to complete the race as a result of clinical disease, and investigation of potential cytokine alterations that may precipitate the observed changes in serum immunoglobulins.

In conclusion, training and racing results in significant decreases in [IgG] and [IgM] in racing sled dogs. The etiology and clinical implications of this phenomenon are currently unknown but warrant further investigation. Future studies may include establishing the correlation between infectious disease and serum immunoglobulin concentrations in racing sled dogs, investigation of plasma volume changes during training and racing, and the effects of nutritional and management interventions on specific aspects of immune function in training and racing sled dogs.

Footnotes

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

aCorvac serum separator tubes, Tyco Healthcare Group LP, Mansfield, MA

bHitachi 911, Roche-Boehringer Manheim, Indianapolis, IN

cDog Ig ELISA Quantitation Kits, Bethyl Laboratories, Montgomery, TX

dELISA Starter Accessory Package II, Bethyl Laboratories

eSPECTRAmax 340pc, Molecular Devices, Sunnyvale, CA

fSOFTmak PRO, Molecular Devices

gPC SAS Version 9.2, SAS Institute, Cary, NC

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

This study was supported by the American Kennel Club Canine Health Foundation and The Ohio State University. Appreciation is extended to Clint Warnke and Jana Fletcher for their assistance.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References