THE EFFECT OF DEXAMETHASONE ON HEMATOLOGIC PROFILES, HEMOSPORIDIAN INFECTION, AND SPLENIC HISTOLOGY IN HOUSE FINCHES (HAEMORHOUS MEXICANUS)

Corticosteroid production in birds, principally corticosterone, is regulated by the hypothalamo-pituitary-adrenal axis and, similarly to mammals, is involved in a wide range of physiologic functions including stress response, immune function, and inflammation (de Matos 2008). Of interest here is the effect of glucocorticoid-mediated immunosuppression on House Finches (Haemorhous mexicanus) as induced by dexamethasone. A tool used by other researchers aiming to mimic stress-induced immunosuppression in birds is the administration of corticosteroids such as dexamethasone (Applegate 1970; Jones et al. 1988; Akhtar et al. 2005; Singh et al. 2010; Chapman 2014; Berenjian et al. 2018; Schoenle et al. 2019; Romeo et al. 2020). Dexamethasone is a potent, long-acting, synthetic glucocorticoid that suppresses the hypothalamo-pituitary-adrenal axis to a greater degree, and for a longer period of time, in Common Pigeons (Columba livia) as compared to mammals (Westerhof et al. 1994; de Matos 2008). Dexamethasone-induced immunosuppression has been documented in Turkeys (Meleagris gallopavo) as increasing susceptibility to opportunistic bacteria as well as clostridial dermatitis (Huff et al. 1999, 2013). Chickens (Gallus gallus domesticus) have demonstrated an increased susceptibility to viral infections and dose-dependent lymphopenia in response to corticosterone administration (Gross et al. 1980; Lumeji 1994). Similarly, pigeons with intestinal coccidia revealed a dexamethasone dose-dependent increase in oocyst excretion (Haas 1987; Lumeji 1994).

Heterophils and lymphocytes comprise the majority of circulating avian leukocytes, and heterophil-leukocyte ratios are frequently used as a reliable indicator of immunosuppression or stress (Ots et al. 1998; Davis et al. 2008; Hõrak et al. 2013). This is principally manifested as a reduction in lymphocytes (B cells and natural killer [NK] cells) and an accompanied increase in heterophils (Dhabhar et al. 1995, 1996; Koutsos and Klansing 2014). As well as acute stress, an elevated heterophil-leukocyte ratio is considered an indicator of chronic stress (Davis et al. 2008; Müller et al. 2011). Consequently, this has been used as a parameter in welfare studies in commercial broilers and has been associated with increased mortality in field studies and handling stress in wild birds (Hocking et al. 2002; Sepp et al. 2010; Müller et al. 2011; Cirule et al. 2012; Hõrak et al. 2013; Krams et al. 2013). As expected, a single dose of corticosteroids has been shown to produce lymphopenia and heterophilia in both racing pigeons and chickens (Gröger and Grimm 1989; Lumeji 1994). In addition to alterations of the leukogram, administration of corticosterone in chickens has resulted in involution of the cloacal bursa, thymus, and spleen (Gross et al. 1980; Dohms and Metz 1991; Lumeji 1994; de Matos 2008). The overall effect of this is suppression of both humoral and cell-mediated immunity.

Our objective was to determine whether administration of dexamethasone would have an effect on the hematologic profile, natural hemoparasite infection, and splenic histology that would be suggestive of immunosuppression in House Finches. These changes might consist of perturbations in leukocyte populations and increases in parasite burden. Artificial induction of stress-induced immunosuppression would aid research that explores host responses to infectious diseases.

Eighteen juvenile house finches were captured between May and August 2017 in Ithaca, Tompkins County, New York, US (42°27′36″N, 76°27′0″W), under permit (New York State Fish and Wildlife License 39, Albany, New York, USA; US Geological Survey, Department of the Interi- or, Laurel, Maryland, USA, permit 22669). Birds were kept in individual wire bar cages (45×45×75 cm). The cages were placed inside large, semi-outdoor aviaries. In all cages, the arrangement of perches and water and food containers were identical. Water and food (Roudybush, Inc., Cameron Park, California, USA), mixed two thirds to one third with sunflower seeds), were offered ad libitum. Birds were part of a preceding experiment in which they had been infected with Mycoplasma gallisepticum (MG; strain Hofi MGCA2015) on 15 September 2017 (Reinoso-Pérez et al. 2020). At the time of our study (7 November 2017; d 0), all birds had various degrees of mild to moderate conjunctivitis, but were assumed to have cleared active infections based on results of real-time quantitative PCR tests designed to detect the presence of MG DNA in conjunctival swabs. The PCR tests were performed on all birds on d 0, 4, and 8 (Grodio et al. 2008). Except for one individual that tested positive for MG on d 0, all birds tested negative at all times. All experiments were approved by Cornell University's Institutional Animal Care and Use Committee under the 2017 renewal of protocol 2009-0034.

Birds were randomly assigned to an experimental or control group. Starting d 0, each bird in the experimental group (n=9) received an intramuscular injection of 25 µg (50 µL) of dexamethasone (manufacturer details were not recorded) daily for 8 d. The birds in the control group (n=9) received an injection of the same volume of saline solution. Injections were applied with a 25-ga needle on a 1-mL syringe and alternated between left and right sides of the pectoral muscle.

On d 0 (immediately before dexamethasone administration), 4, 8, and 9, a 25-ga needle was used to puncture the brachial vein to collect a small amount of blood onto heparinized capillary tubes; blood smears were made immediately after collection (Valkiūnas 2005). An automated stainer (Hematek 1000, model 4480, Miles Scientific, Division of Miles Laboratory, Inc., Naperville, Illinois, USA) was used to stain the smears with a modified Wright stain (Hematek® Stain Pack, Siemens Health Care Diagnostics Inc., Tarry-town, New York, USA). A leukocyte (WBC) differential and assessment for the presence of hemoparasites were performed. The WBC differential was performed independently by three observers (observers E.E.V.C., J.D.C.O., and T.P.) by counting 200 leukocytes at 100× oil immersion objective (eyepiece 10×; total magnification 1,000×) within the monolayer of the blood smear (Olympus BX40, Model BX40F4, Olympus Optical Co. Ltd., Tokyo, Japan). Two observers (observers J.D.C.O. and T.P.) independently recorded the presence and number of hemoparasites while performing the 200-cell WBC differential. As identification was restricted to visual characteristics, hemoparasites were categorized and counted as ‘Hemosporidian’ (Plasmodium and Hemoproteus spp.) or ‘Leucocytozoon sp.’ Parasitemia was additionally quantified for each smear based on 100 random fields, approximately 200 erythrocytes per field = 20,000 total erythrocytes, in which the number of infected cells was recorded by observer T (Godfrey et al. 1987). All blood smear observers were blinded to the experimental groups of the birds.

Birds were euthanized by carbon dioxide chamber on d 9. Necropsies were performed on all birds, with gross examination followed by 10% neutral buffered formalin fixation and histopathologic examination of major organs (brain, trachea, esophagus, lungs, heart, proventriculus, ventriculus, duodenum, pancreas, jejunum, cecum, colon, vent and bursa, liver, spleen, kidneys, gonads, adrenals). Immunohistochemical staining with mouse monoclonal antibodies to CD3 antigen was performed on the spleen using a streptavidin-horseradish peroxidase technique (Leica, Newcastle, UK) according to the manufacturer's instructions. Immunohistochemical staining with rabbit polyclonal antibodies to CD20 antigen was performed on a subset of spleen samples using a streptavidin-horseradish peroxidase technique (Thermo Scientific, Rockford, Illinois, USA) according to the manufacturer's instructions. Assessment of CD20 was not pursued further due to a lack of CD20-positive control immunoreactivity. Instead, B cell abundance was estimated using the presence or absence of lymphoid follicles.

Histologically, spleen sections were graded based on the presence and number of lymphoid follicles as a proxy for B cell abundance: absent follicles (grade 0), one follicle (grade 1), two to three follicles (grade 2), four or more follicles (grade 3), per whole spleen examined (Fig. S1). The proportion of CD3 immunoreactive cells (T cells) within the spleen sections was also graded 1–3: less than one third of parenchyma immunoreactive (grade 1); one third to two thirds of parenchyma immunoreactive (grade 2); greater than two thirds of parenchyma immunoreactive (grade 3; Fig. 1). Brain, trachea, esophagus, lungs, heart, proventriculus, ventriculus, duodenum, pancreas, jejunum, cecum, colon, vent and bursa, liver, spleen, kidneys, gonads, and adrenal glands were also histologically examined, and the presence or absence of small intestinal coccidiosis was specifically noted (Fig. 2). The observer of the histology was blinded to the experimental groups of the birds.

Hematologic results from observer counts (three observers for WBC differentials and two observers for hemoparasites) were averaged across observers, and comparisons among pretreatment (d 0) results and d 4, 8, and 9 were stratified by treatment group (dexamethasone vs. saline) and performed with Wilcoxon signed-rank test for matched-pairs. Spearman's rank correlation was used to describe the agreement among the observers. Histologic grades for lymphoid follicle presence and CD3 immunoreactivity in spleen were compared between treatment groups with a Wilcoxon rank-sum test. Proportions of birds with intestinal coccidiosis between treatment groups were assessed with Fisher's exact test. Statistical significance was set at P<0.05; all descriptive and statistical analyses were performed in Stata (StataCorp 2020).

Two birds from the experimental group and one bird from the control group died spontaneously on d 5, 6, and 8. In one dexamethasone-treated bird, the cause of death was likely intestinal coccidiosis (Fig. 2). In the remaining two birds, a definitive cause of death was not identified. Therefore, results for d 8 and d 9 include fewer than the 18 birds reported for d 0 and d 4.

White blood cells were identified as heterophils, basophils, eosinophils, lymphocytes (small and medium), and monocytes (Fig. 3). Averaged WBC differential counts comparing d 0 and d 4 blood smears in dexamethasone-treated group demonstrated an approximately threefold increase in relative heterophils and an approximately 55% relative decrease in lymphocytes (see Table 1 and Supplementary Material Table S1 for P values). The relative heterophilia was sustained at d 8 and d 9, although not statistically significant at the latter time point. A significant relative lymphocytosis was observed in the control group at d 8.

A relative eosinopenia was present in both the dexamethasone-treated and control groups at d 4 and d 8, when compared to d 0 counts. This was sustained at d 9 only in the treatment group. Within the treatment group, this represented a 38–80% decrease in relative eosinophils over the experiment's time course. Within the control group, this reduction ranged from 50% to 61% compared to d 0 values.

Monocytes were approximately doubled in the control group at d 8 and d 9 when compared to d 0 counts. No significant change in monocytes was observed in the treatment group and basophils did not exhibit any significant trends in either group.

In most blood smears, a few (not counted) immature erythrocytes characterized by a basophilic cytoplasm and clumped nuclear chromatic were observed, suggesting that a mild regenerative anemia had resulted from repeated blood collection.

Mean hemosporidian (Plasmodium and Hemoproteus spp., Fig. 3) counts comparing d 0 to d 4 and d 8 of the dexamethasone-treated group demonstrated a significant increase (see Table 2 and Supplementary Material Table S2 for P values). Although not a significant increase at d 9, the percentage of dexamethasone-treated birds infected ranged from 36% to 76% over the three time points compared to 0.9–3.6% in the control birds. Leucocytozoon sp. organisms (Fig. 3) were detected by one observer only. These were restricted to the dexamethasone-treated group and were found in different individual animals at different time points (two birds at d 0; one at d 4; one at d 8; one at d 9). Low numbers and lack of detection by a second observer precluded the performance of any statistical analysis.

All birds in the dexamethasone-treated group lacked histologically apparent splenic lymphoid follicles. This equated to a grade 0 (absent follicles) upon examination of the entire splenic section with statistical significance of P<0.001 (Table 3). In contrast, in the control group splenic histologic results were distributed across grade 1 to grade 3, with one to multiple lymphoid follicles present in each bird. No significant difference was observed in splenic CD3 immunoreactivity between groups. The prevalence of intestinal coccidiosis was greater in the treatment group, although the difference lacked statistical significance (P=0.077).

Spearman's rank correlation between the three observers (E.E.V.C., J.D.C.O., and T.P.) for WBC differential counts across all time points was good (>0.7) for heterophils (see Supplementary Material Table S3). Interobserver correlation for lymphocyte and basophil counts ranged between moderate (≥0.3 to <0.7) and good. Poor (<0.3) to good correlation was present for eosinophils, and poor to moderate correlation was noted for monocytes. Hemosporidian count correlations between the two observers (J.D.C.O. and T.P.) were poor to good across all time points (see Supplementary Material Table S3). A Spearman's signed-rank correlation of 0.724 was measured for hemosporidian counts performed by a single observer using a 200-cell WBC differential count and a full examination of the whole blood smear, regardless of treatment or sampling date (see Supplementary Material Table S4). Leucocytozoon parasites were only detected by one observer (J.D.C.O.), so no statistical analysis was performed.

Our findings confirmed changes in the hematologic profile, natural hemoparasite infection, and splenic histology of dexamethasone-treated House Finches that were compatible with immunosuppression and may effectively mimic a stress response. Additionally, interobserver agreement in both WBC differentials and parasite counts was documented and our observations may help to inform future research.

The relationship between increased glucocorticoids and heterophilia and lymphopenia is well established in birds (McFarlane and Curtis 1989); the leukocyte perturbations that we documented largely reflect this. Day 4 hematology results in the dexamethasone-treated group showed relative heterophilia and lymphopenia compatible with an acute stress response (McRee et al. 2018). Although the heterophilia was sustained to d 8, it was diminished at d 9 and the lymphopenia was absent at d 8 and d 9. This suggests that the initial dexamethasone-induced leukocytic changes waned despite continued daily dexamethasone administration. This is unusual, as prolonged dexamethasone, corticosterone, and heat stress exposure trials in Japanese quail and domestic fowl have each elicited sustained elevated heterophil-leukocyte ratios beyond a 9-d period (Jones et al. 1988; Berenjian et al. 2018; Nazar et al. 2018). It is plausible that species and seasonal variation in hemic responses to dexamethasone may account for the results seen, as may individual factors including age and comorbidities such as hemoparasitism, mycoplasmosis, and coccidiosis (Applegate 1970). Although previously unreported, perhaps an acclimation response to dexamethasone treatment occurred. Statistically significant eosinopenia was identified in treated birds at d 4, 8, and 9, but this was of equivocal significance to the treatment given eosinopenia in the control group at d 4 and d 8. Relative monocytosis in the control group was of unknown significance and may be related to sample factors, as only low numbers of monocytes were present in all counts. Although the control group exhibited a relative lymphocytosis at d 8, this was resolved at d 9 and represented very mild elevations that were unlikely to represent clinical significance.

The increased parasite burdens in dexamethasone-treated birds were suggestive of glucocorticoid-induced immunosuppression. Hemosporidian counts in dexamethasone-treated birds showed elevations that were significant at d 4 and d 8. Although not statistically significant (P=0.077), an analogous trend in intestinal coccidiosis incidence was noted in the treatment group histologically. Similarly, increased parasitemia in response to elevated glucocorticoids has been documented in both Red-winged Blackbirds (Agelaius phoeniceus) and House Sparrows (Passer domesticus) treated with corticosterone (Applegate 1970; Schoenle et al. 2019). In analyses of wild House Finches localized to areas of urbanization compared to rural regions, increased prevalence and severity of poxviral dermatitis and fecal coccidian oocysts were found (Giraudeau et al. 2014). The authors surmised that the urbanized environment may have contributed to physiologic stress and/or facilitated horizontal disease transmission.

Increasingly, complex interactions between parasite species and host immunomodulation have been recognized, particularly the frequent absence of an obvious linear relationship between parasite number to stress. Schoenle et al. (2019) showed that in Red-winged Blackbirds, a Plasmodium increase in response to glucocorticoid therapy only occurred in the absence of Hemoproteus infection. A similar interplay of glucocorticoid levels, different parasites, and season has been demonstrated in other bird species and mammals (Akhtar et al. 2005; Chapman 2014; Sild et al. 2014; Cizauskas et al. 2015; Romeo et al. 2020). In our experiment, differentiation between Plasmodium spp. and Haemoproteus spp. was not performed, as they cannot be reliably distinguished morphologically. Consequently, parasite count elevations may represent different hemosporidian genera and species, singular or in combination with each other. Leucocytozoon organism detection was rare and restricted to the dexamethasone-treated group. Any inferences about treatment effect on the latter hemoparasite were made difficult due to their scarcity. Additionally, both hematology and hemosporidian results may have been affected by variation in blood film quality. Given these findings, further work to better elucidate the relationship between dexamethasone therapy and different types of parasitism in house finches would be valuable.

Notably, histologic examination of the spleen from the dexamethasone-treated group in this study showed a uniform absence of lymphoid follicles, suggesting a decrease in the B cell population. Conversely, all control animals had at least one detectable follicular region. These results are in line with previous reports in chickens subjected to heat stress, where white pulp germinal centers exhibited lymphoid depletion (Lola et al. 2017). Loss of lymphocytes probably resulted from glucocorticoid-mediated cellular apoptosis, as has been demonstrated in mammals and fish (Power et al. 2002; Walsh et al. 2002; Salak-Johnson and McGlone 2007; Sandhu et al. 2012). Similarly, dexamethasone (30 µg/bird/day) given subcutaneously to Indian tropical birds, Jungle Bush Quail Perdicula asiatica, over a 21-d period resulted in decreased splenic mass and histologic density as well as increased circulating total leukocyte and lymphocyte counts (Singh et al. 2010). In the same study, splenocyte in vitro and in vivo interleukin-2 production was decreased, thereby diminishing T cell activation, and apoptosis was increased. A comparable effect is presumed to have occurred in our study, although splenic mass and interleukin-2 production were not measured. Immunohistochemistry for CD3 did not identify any trends in splenic T cell semiquantitative grading, despite glucocorticoid-induced immunosuppression being linked to apoptosis in the T cell progenitor population (Evans-Storm and Cidlowski 1995; Sapolsky et al. 2000). This might suggest that the effect of dexamethasone on T cells was variable in this study population, although the small sample size may have precluded detection of any subtle trends. It is also possible that our methods failed to detect a decrease in CD3-positive cells. Our semiquantitative method depended on the overall appearance of the entire splenic tissue and could have been affected by the obvious disappearance of lymphoid follicles in dexamethasone-treated birds. Regardless, sustained dexamethasone treatment profoundly impacted the histologic appearance of spleen in a manner consistent with previous investigations on avian stress and dexamethasone administration.

Significant to this discussion was the experimental infection with MG preceding our study and involving all study birds. For the following reasons we assert that this probably had no effect on our results. Firstly, all birds had been similarly infected, thus MG-infection could not explain the differences we found between dexamethasone-treated and control groups. Secondly, all birds were tested for the presence of MG by conjunctival swabs several times during the study and were, with one exception, consistently negative. The single exception was one bird that tested positive on d 0 but was negative at d 4 and d 8. Those negative results suggest that there was no active infection during our study, thus it holds little relevance to our findings.

Interobserver agreement for manual leukocyte differential counts varied dramatically between cell types in this study. This phenomenon has been examined in both human and veterinary applications, and its inherent variability is widely recognized (Kjelgaard-Hansen and Jensen 2006). In mammals, this limitation disproportionately affects nonsegmented leukocytes, with neutrophil counts typically demonstrating the highest reproducibility and lymphocytes, eosinophils, monocytes, and basophils associated with significant imprecision (Tvedten and Lilliehöök 2011). For birds, existing literature indicates high reproducibility of heterophil and lymphocyte counts and similar imprecision of other leukocytes (Post et al. 2003). Our findings were generally concordant with previous reports, with a good Spearman's rank correlation between results found by the three observers for heterophils across all time points. Greater variation was noted for lymphocytes and basophils, and even more variability was noted for eosinophils and monocytes. These results may be partially affected by the corresponding numbers of these cells across time points, with lymphocytes, heterophils, and basophils being the most numerous leukocytes observed. However, given that lymphocytes were typically the most abundant leukocyte, and that basophils often exceeded heterophils, other factors must be considered when explaining variation between observers. ‘Human factor’ variability contributes to both inter- and intraobserver agreement. This includes general considerations such as psychologic pressures, distraction, and fatigue, as well as the area of the slide holding the blood film that is examined for differential counting and any errors of cell identification, which may contribute to interobserver reproducibility (Fuentes-Arderiu et al. 2007). The ‘human factor’ appeared to disproportionately affect lymphocytes and basophils in this study. The cause of this was unclear, although lymphocytes are occasionally difficult to discern from immature erythrocytes and monocytes.

Correlations between the two observers for hemosporidian counts recording the presence and number of hemoparasites varied considerably across time points. The variation found in these counts was analogous to that found in laboratory professionals on malaria microscopy in malaria-endemic countries, where count agreement was markedly varied, but malaria detection was usually good (Ngasala and Bushukatale 2019). Good correlation was achieved across all time points between counts recorded by a single observer while performing a partial and full examination of the whole blood smear. The former is a nonstandardized method and is convenient because it can be performed concurrently with the WBC differential count. These findings suggest that this crude parasite count method has good correlation with the established whole blood smear method; however, further scrutiny of its use is warranted.

Overall, our findings indicate that dexamethasone has an effect on the hematologic profile of House Finches and suggest that it may be a useful method to induce immunosuppression in this species. These results may enable infectious disease researchers to further delve into host-pathogen interactions and the effects of stress in passerines. Our findings on interobserver agreement may better guide future experimental design.

The authors would like to extend their sincere gratitude to the histotechnicians at Cornell University's Animal Health Diagnostic Center who prepared slides for these cases.

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-21-00129.