Circulating γδ T Cells in Response to Salmonellaenterica Serovar Enteritidis Exposure in Chickens

ABSTRACT γδT cells are considered crucial to the outcome of various infectious diseases. The present study was undertaken to characterizeγδ (T-cell receptor 1+ [TCR1+]) T cells phenotypically and functionally in avian immune response. Day-old chicks were orally immunized with Salmonella enterica serovar Enteritidis live vaccine or S. enterica serovar Enteritidis wild-type strain and infected using the S. enterica serovar Enteritidis wild-type strain on day 44 of life. Between days 3 and 71, peripheral blood was examined flow cytometrically for the occurrence of γδ T-cell subpopulations differentiated by the expression of T-cell antigens. Three different TCR1+ cell populations were found to display considerable variation regarding CD8α antigen expression: (i) CD8α+high TCR1+ cells, (ii) CD8α+dim TCR1+ cells, and (iii) CD8α− TCR1+ cells. While most of the CD8α+high TCR1+ cells expressed the CD8αβ heterodimeric antigen, the majority of the CD8α+dim TCR1+ cells were found to express the CD8αα homodimeric form. After immunization, a significant increase of CD8αα+high γδ T cells was observed within the CD8α+high TCR1+ cell population. Quantitative reverse transcription-PCR revealed reduced interleukin-7 receptor α (IL-7Rα) and Bcl-x expression and elevated IL-2Rα mRNA expression of the CD8αα+highγδ T cells. Immunohistochemical analysis demonstrated a significant increase of CD8α+ and TCR1+ cells in the cecum and spleen and a decreased percentage of CD8β+ T cells in the spleen after Salmonella immunization. After infection of immunized animals, immune reactions were restricted to intestinal tissue. The study showed that Salmonella immunization of very young chicks is accompanied by an increase of CD8αα+high γδ T cells in peripheral blood, which are probably activated, and thus represent an important factor for the development of a protective immune response to Salmonella organisms in chickens.

Bacteria of the genus Salmonella are responsible for a variety of acute and chronic diseases in poultry. Moreover, paratyphoid Salmonella organisms are among the most important pathogens inducing food-borne zoonoses throughout the world. Poultry products represent a substantial source of salmonella infection in humans, with Salmonella enterica serovar Enteritidis and Salmonella enterica serovar Typhimurium being the serovars most frequently associated with human food poisoning. The manifestations and consequences of paratyphoid Salmonella infections may vary markedly depending on the age of the birds (5). While infections in young chicks are often characterized by considerable morbidity and mortality, birds aged some weeks and adults are far less susceptible to lethal effects of Salmonella exposure and may experience intestinal colonization as well as systemic dissemination without signs of illness.
Vaccination is widely regarded as the most realistic approach to efficiently control the invasive serovars of S. enterica serovar Enteritidis and S. enterica serovar Typhimurium in the poultry industry. In this context, comprehensive knowledge of immunological defense mechanisms of chickens against Salmonella exposure represents a crucial prerequisite for successful vaccine development. The immune-stimulating properties of various Salmonella vaccine preparations have already been examined for their potential to enhance antibody production, cell-mediated immunity, and induction of invasion of different cell populations in several organs or peripheral blood (3,8,9,13,23,24,39). Notably, T-cell reactions and cellular immune responses seem to be of central importance for defense against Salmonella infections in chickens (8,9,23,24).
T lymphocytes, which are characterized by their expression of special T cell receptors (TCR), have been classified into two subgroups: ␣␤ and ␥␦ T-cell receptor-bearing cells, with the former comprising the classical CD4 ϩ and CD8 ϩ cells. Avian ␣␤ T cells are additionally subdivided into V␤ 1 and V␤ 2 cells. Compared to the V␤ 1 population, the V␤ 2 ␣␤ T cells only constitute a small proportion of the total ␣␤ T-cell population of avian peripheral blood. The ␥␦ T-cell subgroup was first described in the middle of the 1980s (4,11). Meanwhile, a number of interesting distinctions between ␥␦ and ␣␤ T cells have been found, such as tissue distribution, repertoire restrictions, and nonclassical major histocompatibility complex (MHC) restriction (1,22). In addition, ␥␦ T cells were shown to possess antimicrobial activity in bacterial, parasitic, and viral infections (15,19,45) as well as to perform a wide range of functions, such as cytokine production, cytotoxic activity, immunomodulation, granuloma organization, and regulation of inflammation (10,12,26,35,41). Based on the expression of different V␥ or V␦ elements, tissue localization, TCR diversity, and the expression of some surface antigens, ␥␦ T cells could be divided into various functionally different subpopulations in mice and humans (14), cattle (25,46), and rats (42).
In mature chickens, T cells bearing the ␥␦ TCR comprise up to 50% of the peripheral T-cell pool (2). Recent studies in our lab revealed up-regulation of the percentage of CD8 ϩ -␥␦-receptor-positive T cells after S. enterica serovar Typhimurium infection of chicks in peripheral blood and tissues (8). Functional CD8 is a dimeric protein and exists either as a CD8␣␤ heterodimer or as a CD8␣␣ homodimer. These two forms are differentially expressed on a variety of lymphocytes and seem to serve distinct functions (21). Although the occurrence of a relatively large proportion of CD8 ϩ ␥␦ T cells expressing only the ␣-chain of the CD8 could be shown in spleens of embryos and young chicks (44), there is no knowledge of possible functions of these cells in chickens. In mice (28,29) and humans (27), CD8␣␣ is expressed on ␣␤ T-cell subsets displaying a memory phenotype. The development of memory CD8 ϩ T cells is a dynamic multistage process that includes an initial T-cell activation with adjacent T-cell expansion, a contraction phase (deletion, activation-induced cell death), and the final memory cell development (40). More recent work has indicated that cytokine receptors, such as interleukin-2 receptor (IL-2R), IL-7R, and IL-15R, all of which include the common ␥-chain as an integral signaling component, play an important role in the development, survival, and homeostatic proliferation of lymphocyte populations. This has been most clearly established for CD8 ϩ T cells. Moreover, the expression of IL-2, IL-7, and IL-15 cytokines and their receptors have been shown to be crucial for CD8 ϩ T-cell memory generation and maintenance (40). On the other hand, the accumulation of antiapoptotic factors, such as the Bcl-2 family members Bcl-2 and Bcl-x L , correlate with enhanced lymphocyte survival (40,36), and Fas-FasL interaction seems to be essential for peripheral T-cell deletion (36).
In the present study, peripheral blood CD8␣-positive and -negative ␥␦ T cells were characterized in detail with regard to their expression of several surface antigens (CD4, CD8␤, CD28, and MHC class II) and compared with the V␤1 and V␤2 ␣␤ T cells of avian peripheral blood. Additionally, morphological and functional aspects of the ␥␦ T-cell response in chickens after immunization and infection with S. enterica serovar Enteritidis were examined in blood and tissue. For this purpose, changes of the proportions of ␥␦ T-cell subsets characterized by CD8 and CD28 antigen expression, mRNA expression levels of IL-2R␣ and IL-7R␣, and apoptotic markers (Fas, Fas-L, and Bcl-x) were analyzed.

MATERIALS AND METHODS
Chickens. Specific-pathogen-free White Leghorn chickens were hatched at the facilities of the Friedrich-Loeffler-Institute, Jena, Germany, from eggs obtained from Charles River Deutschland GmbH, Extertal, Germany. Experimental and control groups were kept in separate rooms; commercial feed (in powder form without antibiotics or other additives) and drinking water were both available ad libitum. Cleaning and feeding regimens were organized which prevented crosscontamination effectively throughout the course of the experiment.
Experimental design and bacteriology. Salmonella vac E (TAD Pharmazeutisches Werk GmbH, Cuxhaven, Germany), a commercial live Salmonella enterica serovar Enteritidis (SE-LV) vaccine strain, and a nonattenuated Salmonella enterica serovar Enteritidis 147 (SE 147) wild-type strain (33) were used for oral immunization of the chicks on their first day of life. The viable count of the attenuated SE-LV was 1 ϫ 10 8 to 2 ϫ 10 8 CFU per bird, and that of the nonattenuated SE 147 was 1 ϫ 10 7 to 2 ϫ 10 7 CFU/bird. On day 44 of life, all immunized and nonimmunized chickens were infected orally using a nalidixic acid-resistant variant of the wild-type strain Salmonella enterica serovar Enteritidis 147 (SE 147 N) at a dose of 2 ϫ 10 8 CFU/bird. Oral administration of 0.1 ml bacterial suspension/bird was performed by instillation into the crop using a syringe with an attached flexible tube. Another control group was neither immunized nor infected. In summary, the following animal groups were investigated: group 1 (SE-LV/SE 147 N), immunization with SE-LV on day 1 of life and infection with SE 147 N on day 44 of life; group 2 (SE 147/SE 147 N), immunization with SE 147 on day 1 of life and infection with SE 147 N on day 44 of life; control group 3 (nothing/SE 147 N), no immunization, infection with SE 147 N on day 44 of life; control group 4 (nothing/nothing), no immunization, no infection.
All Salmonella strains used in this study were stored in the Microbank system (PRO-LAB Diagnostics, Ontario, Canada) at Ϫ20°C. The bacterial suspensions for immunization and infection were cultivated by shaking (18 h at 37°C) in nutrient broth (SIFIN, Berlin, Germany). Doses were estimated by measuring extinction at 600 nm against a calibration graph determined for the strains used and subsequently confirmed by plate counting on nutrient agar (SIFIN).
Bacterial counts in spleens were estimated after infection using a standard plating method as described previously (33). Briefly, homogenized organ samples were diluted in phosphate-buffered saline, plated on deoxycholate-citrate agar (SIFIN) supplemented with sodium nalidixate (50 g/ml) to detect the challenge organisms (SE 147 N), and incubated at 37°C for 18 to 24 h.
Peripheral blood leukocyte isolation. To study the dynamic of ␥␦ T-cell subsets after Salmonella exposure, peripheral blood leukocytes of five animals per day of examination (days 4,7,11,14,21,24,28,32,43,45,46,49,50,52,53,56,63, and 71 of life) and group were analyzed. For comparison of avian peripheral TCR1 ϩ with TCR2 ϩ and TCR3 ϩ T cells, peripheral blood of chickens was analyzed between days 49 and 56 of life (in total, 15 nontreated animals). For that, the animals were sacrificed and completely bled to death. Lymphocytes of peripheral blood were isolated as described previously (8). Briefly, heparinized blood was mixed with 3% hetastarch (Sigma Immuno Chemicals, St. Louis, Mo.) and centrifuged at 65 ϫ g for 10 min to allow erythrocytes to sediment. The leukocytes of the supernatant were washed twice in phosphate-buffered saline and subsequently used for flow cytometry.
Flow cytometry and cell sorting. For cell sorting, T-cell characterization, and analysis of the dynamic of ␥␦ T-cell subsets, fluorescein isothiocyanate-labeled mouse anti-avian TCR1 (␥␦), TCR2 (␣␤ V␤ 1 ), and TCR3 (␣␤ V␤ 2 ) antibodies, an R-phycoerythrin-labeled mouse anti-avian CD8␣ monoclonal antibody, and biotin-conjugated mouse anti-avian CD4, CD8␤, CD28 MHC class II antibodies (all from Southern Biotechnology Associates, Eching, Germany) were used. The biotin-conjugated antibodies were detected by Alexa Fluor 633-linked streptavidin (Molecular Probes, Leiden, The Netherlands). All antibody concentrations and dilutions were tested prior to starting the animal experiment. For every test, 2 ϫ 10 5 leukocytes were incubated in parallel with the appropriate monoclonal antibodies (30 min in the dark). Aliquots of 10,000 to 20,000 cells per sample were analyzed using a FACSCalibur (Becton Dickinson, Heidelberg, Germany) equipped with a 15-mW, 488-nm argon ion laser and a 633 diode laser.
Quantitative real-time RT-PCR. Total RNA was extracted from cell samples using the RNeasy mini kit (QIAGEN, Hilden, Germany) by following the manufacturer's instructions, eluted in 50 l RNase-free water per 50,000 cells, and stored at Ϫ80°C. For functional characterization of ␥␦ T-cell subsets, mRNA expression of avian cytokine receptor chains (IL-2R␣ and IL-7R␣) and apoptotic related proteins (Fas, Fas ligand, and Bcl-x) was determined by the QuantiTect SYBR green one-step reverse transcription (RT)-PCR kit (QIAGEN) according to the instructions of the manufacturer. Amplification and detection of specific products were performed using the Mx3000P real-time PCR system (Stratagene, La Jolla, CA) using the following cycle profile: one cycle of 50°C for 30 min and 96°C for 15 min, 45 cycles of 94°C for 30 s and appropriate annealing temperature for 30 s, followed by 72°C for 30 s. Primer sequences and annealing temperatures are given in Table 1. To avoid amplification of cellular DNA of each primer pair, at least one primer spans an intron-exon boundary. The threshold method was used for quantification of the mRNA level. ⌬C T (cycle threshold change) values were calculated on the basis of internal standard glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Results were expressed as 40- Immunohistochemistry. Samples from the cecum and spleen of animals (days 1 to 56 of life) were examined by immunohistochemistry to determine the invasion of different T-cell subsets. Cryostat sections of 7-m thickness were prepared, and detection of cells was done using unlabeled primary monoclonal 3968 BERNDT ET AL. INFECT. IMMUN.
antibodies against the TCR1, CD8␣, and CD8␤ antigens (Southern Biotechnology Associates, Eching, Germany) and a commercially available staining kit (PAP, ChemMate Detection kit, peroxidase antiperoxidase, rabbit/mouse; Dako-Cytomation, Hamburg, Germany). As a negative control, slides were incubated with preimmune mouse serum (dilution, 1:500) instead of the monoclonal antibodies. Sections were counterstained with hematoxylin and mounted with Canada balsam (Riedel de Haen AG, Seelze-Hannover, Germany). Image analysis. Immunohistological tissue preparations were examined by light-microscopic image analysis (analySIS 3.0; Soft Imaging System GmbH). All measurements were performed at a magnification of ϫ20. Using the computer software, the regions of interest were drawn interactively as a polygon (cecum) or rectangle (spleen) on the screen. At least 5 regions of interest of each tissue section and antibody were scanned, the percentage of antibody-stained areas was determined, and the mean values were calculated.
Statistical analysis. The data of flow cytometric analyses and image analyses were evaluated statistically. The Student t test for comparison of two independent samples was used for statistical evaluation of differences between the groups (each immunized group against the control group). P values of Յ0.05 were considered significant.
Viable counts of bacteria were converted into logarithmic form. For the purposes of statistical analysis, a viable log 10 count of Ͻ1.47 (limit for direct plate detection) obtained from a sample becoming positive only after enrichment was rated a log 10 value of 1.0. A sample which yielded no Salmonella growth after enrichment was rated a log 10 value of 0. Data were evaluated by multifactorial variance analysis. The factors considered were group and time. P values of Ͻ0.05 were regarded as statistically significant (software from Statgraphics Plus, Inc., Rockville, MD).

Clinical signs after Salmonella immunization and infection.
In the present study, morphological and functional aspects of the ␥␦ T-cell response in chickens after immunization and infection with S. enterica serovar Enteritidis were examined in blood and tissue.
The oral administration of attenuated live SE-LV did not result in morbidity or symptoms of intestinal inflammation. When using a dose of 1 ϫ 10 7 to 2 ϫ 10 7 CFU/bird of SE 147 in day-old chicks, strong colonization of the intestine and spread to liver was observed, but severe clinical signs were absent (34). Only a few birds showed signs of intestinal inflammation in the first week after administration. Infection with SE 147 N failed to generate any signs of morbidity or clinical symptoms in immunized and control birds.
Bacteriology. Oral immunization of chickens produced protective effects compared to nonimmunized controls ( Table 2). Immunization using the nonattenuated S. enterica serovar Enteritidis wild-type strain 147 led to the most efficient protection, and no challenge organisms were detected in the spleen in the course of the experiment.
Birds immunized using an attenuated live Salmonella enterica serovar Enteritidis vaccine (SE-LV) revealed a considerably diminished number of the challenge strain in the spleen (reduction of 1.0 to 1.5 log 10 units) compared to nonimmunized controls.
The majority of ␥␦ T cells were not found to express CD8␣ antigen on their surface. The CD8␣ ϩhigh TCR1 ϩ and CD8␣ ϩdim TCR1 ϩ cell populations occurred in similar proportions. In comparison, most of the TCR2 ϩ and TCR3 ϩ cells displayed diminished CD8␣ expression. A smaller amount of the TCR2 ϩ and TCR3 ϩ cells either expressed the CD8␣ antigen at a high concentration or showed no CD8␣ staining. Compared to TCR1 ϩ and TCR2 ϩ T cells, TCR3 ϩ cells were only found in small quantities in blood (Fig. 1A).
Approximately 40% of CD8␣ Ϫ TCR2 ϩ T cells also showed a distinct MHC class II expression. MHC class II expression of CD8 ϩ TCR3 ϩ cells was not investigated.
Dynamics of TCR1 ؉ cell subsets after Salmonella immunization in blood. To examine the possible induction of a ␥␦ T-cell response to Salmonella organisms, the percentage of TCR1 ϩ T cells characterized by their CD8␣ antigen expression was analyzed within the circulating lymphocyte population of immunized chicks compared to nonimmunized control animals. Additionally, the expression of CD8␤ and CD28 on the defined TCR1 ϩ cell subsets was investigated (Fig. 2).
CD8␣ ϩhigh TCR1 ϩ cells were generally found in very low numbers in the peripheral blood of immunized and nonimmunized chicks on day 4 of life. After immunization with SE-LV or SE 147 wild type, significantly (P Ͻ 0.05) higher numbers of these ␥␦ T cells were observed ( Fig. 2A). The peak in the number of CD8␣ ϩhigh TCR1 ϩ cells was reached on days 7 (SE-LV) and 11 (SE 147) of life. Subsequently, the percentages decreased to the level of nonimmunized animals. The percentages of CD8␣ ϩdim TCR1 ϩ cells in peripheral blood were significantly increased (P Ͻ 0.05) in animals immunized with SE 147 and/or SE-LV compared to nonimmunized controls (data not shown). A peak was seen on days 7 and 11 (SE 147) and on day 14 (SE-LV) of life. Afterwards, the number of these cells dropped to the level of the control birds until day 21 of life. From day 21 onward, a moderate but significant decrease (P Ͻ 0.05) of these cells was detected in immunized chicks compared to controls (data not shown).
CD8␣ Ϫ TCR1 ϩ cells were significantly reduced (P Ͻ 0.05) in immunized animals compared to control birds. This phenomenon was observed from day 21 onward and was most conspicuous in chickens immunized with SE 147. On day 43 of life, the percentages of the CD8␣ Ϫ TCR1 ϩ T cells within the lymphocyte population of the immunized birds were nearly the same as found in controls (Fig. 2B).
As a consequence of immunization of day-old chicks, the CD8␣ ϩhigh TCR1 ϩ -to-CD8␣ Ϫ TCR1 ϩ cell ratio within the ␥␦ T-cell population of immunized animals shifted significantly in favor of the CD8␣ ϩhigh TCR1 ϩ cell subset between days 7 and 14 of life (data not shown).
In immunized birds, the number of CD8␣ ϩhigh TCR1 ϩ cells expressing the ␤ chain of the CD8 antigen within the CD8␣ ϩhigh TCR1 ϩ cell population was significantly decreased between days 4 and 32 (Fig. 2C). This indicated an enormous increase of CD8␣␣ ϩhigh TCR1 ϩ cells in this ␥␦ T-cell subset.
Concerning the expression of the CD28 antigen on TCR1 ϩ cells, an increase of CD28 ϩ cells among the CD8␣ ϩhigh TCR1 ϩ T lymphocytes could be observed in immunized animals between day 4 and day 11 of age (Fig. 2D).
Dynamics of TCR1 ؉ cell subsets after Salmonella infection in blood. To evaluate a possible second or memory immune response to Salmonella infection on day 44 of life, the percentages of TCR1 ϩ T cells characterized by their CD8␣ antigen expression were analyzed in the previously immunized and nonimmunized chickens. Furthermore, the expression of CD8␤ and CD28 antigens on the defined TCR1 ϩ cell subsets was studied after infection between days 44 and 71 (Fig. 2).
In peripheral blood, the nonimmunized but infected group showed a significant (P Ͻ 0.05) increase of CD8␣ ϩhigh TCR1 ϩ cells 5 and 6 days after infection (days 49 and 50 of life) compared with birds neither immunized nor infected with SE 147 N. In animals immunized on the first day of life, the percentages of the CD8␣ ϩhigh TCR1 ϩ cells remained unchanged after Salmonella infection compared to the nonimmunized group. The percentages of the CD8␣ ϩdim TCR1 ϩ and CD8 Ϫ TCR1 ϩ cells were not significantly changed after infection.
Regarding CD28 antigen expression of the three TCR1 ϩ T cell subsets, no changes were observed after infection of chickens immunized on the first day of life.
Dynamics of TCR1 ؉ , CD8␣ ؉ , and CD8␤ ؉ cells after Salmonella infection in cecum and spleen. The local immune response after Salmonella immunization and infection has been examined by immunohistochemistry using monoclonal antibodies to CD8␣, CD8␤, and TCR1 antigens in the cecum and spleen.
In the cecum, significantly more CD8␣ ϩ , CD8␤ ϩ , and TCR1 ϩ cells were detected in SE-LV-and SE 147-immunized chickens than in the controls between days 4 and 21 of life (Fig. 3). In the ceca of previously immunized chickens, a second significant increase of CD8␣ ϩ , CD8␤ ϩ , and at lower degree, TCR1 ϩ cells were detected after challenge on day 44  and SE-LV-immunized birds but showed high standard deviations.
In spleens of SE-LV-and SE 147-immunized animals, the percentages of CD8␣ ϩ and TCR1 ϩ T cells were increased between days 8 and 21 of life. In contrast, the number of CD8␤ ϩ cells was significantly decreased between days 7 and 21 (Fig. 4). After challenge on day 44 of the already immunized animals, no changes concerning the number of the investigated T-cell populations were detectable. Only the nonimmunized but challenged control animals showed a decrease of CD8␤ ϩ cells, which was observed on days 45 and 46.

DISCUSSION
In the present study, the phenotype of avian T cells, particularly CD8␣ ϩ ␥␦ T cells, was specified to gain more insight into the role of single ␥␦ T-cell subsets in defense against Salmo- nella organisms of chickens. Based on our previous findings (8), the dynamics of CD8 ϩ and ␥␦ T-cell subsets was investigated after immunization and infection of chickens with attenuated and nonattenuated Salmonella enterica serovar Enteritidis strains in peripheral blood and organs. The phenotypic characterization of avian TCR1 ϩ , TCR2 ϩ , and TCR3 ϩ cells demonstrated the presence of completely separate cell populations not only concerning the TCR antigen but also because of the expression of further antigens. A closer relationship between the TCR2 ϩ and the TCR3 ϩ subpopulations than with TCR1 ϩ cells was shown by the relatively high frequency of CD4 ϩ antigen expression of both subpopulations. The detected different expression of the CD28 antigen on TCR2 ϩ and TCR3 ϩ cells indicates distinct activation levels of these T-cell populations. It has been already shown that TCR2 ϩ and TCR3 ϩ cells are separate sublineages of avian T cells by observation of a compensatory increase in TCR3 ϩ cell frequency when the development of the TCR2 ϩ subpopulation was inhibited (17). Other authors described the ability of TCR2 ϩ cells, but not of TCR3 ϩ cells, to migrate to the chicken intestine, where they provide help to mucosal B cells for immunoglobulin A production (16).
Each of the three TCR subsets could be further subdivided in three single subsets based on their intensive, weak, or absent CD8␣ expression. In contrast to CD8␣ ϩdim TCR1 ϩ cells, CD8␣ ϩdim TCR2 ϩ and CD8␣ ϩdim TCR3 ϩ cells of the White Leghorn chickens additionally expressed the CD4 antigen. These cells did not express the CD8␤ chain, indicating the expression of only the CD8␣␣ homodimeric form of the CD8 antigen. Accordingly, other studies demonstrated that, in some H.B15 chickens, a large subpopulation of peripheral blood CD4 ϩ T cells expresses only CD8␣ mRNA but not CD8␤ (30). The same authors have shown that these CD4 ϩ CD8␣␣ T cells can induce an in vivo alloreaction comparable to that of normal CD4 ϩ T cells, which represents direct evidence for peripheral blood CD4 ϩ CD8 ϩ T cells functioning as helper CD4 T cells. It seems that the expression of CD8␣ does not interfere with the function of peripheral blood CD4 ϩ T cells in chickens.
Based on CD8␣ antigen expression, we demonstrated for the first time the existence of three ␥␦ T-cell subsets in avian blood and showed that most of circulating CD8 ϩhigh TCR1 ϩ cells express the ␣␤ heterodimeric form of the CD8 antigen, whereas CD8 ϩdim TCR1 ϩ cells present the ␣␣ homodimeric form of the CD8 antigen. It seems that ␥␦ T cells are able to express both forms of the CD8 antigen depending on their localization or species investigated. Straube and Herrmann (42) considered rats to be the only species with frequent expression of the CD8␣␤ antigen by ␥␦ T cells. In spleens of avian embryos and young chicks, a relatively large proportion of CD8 ϩ ␥␦ T cells expressing only the ␣-chain of CD8 has been found (44).
The three ␥␦ T-cell subsets defined in dependence on their CD8␣ antigen expression showed different reactions after Salmonella immunization of day-old chicks. This fact indicates not only phenotypic but also functional differences between these subsets. A difference in the overall functional and/or activation status of human circulating CD8 Ϫ versus CD8 ϩ ␥␦ T cells has been postulated (25). A comparison of avian CD8␣ ϩhigh and CD8␣ ϩdim ␥␦ T cells in our experiment showed that the response of CD8␣ ϩhigh T cells was more evident. In swine, possible functional differences between CD8 ϩdim and CD8 ϩhigh ␣␤ T cells have been discussed in connection with different expression of a second ␣ or ␤ chain of the CD8 antigen (38,43).
The detected decrease of CD8␤ expression on CD8␣ ϩhigh TCR1 ϩ cells after Salmonella immunization indicates an increase in the number of CD8␣ ϩhigh ␥␦ T cells expressing the CD8␣␣ homodimer. Additionally, a decreased number of CD8␤ ϩ T cells was seen in the spleens of immunized chickens. Whether this was due to a down-regulation of the CD8␤ antigen on the single cell or an influx of CD8␣ ϩ homodimeric ␥␦ T cells has yet to be examined. However, this is the first demonstration of an increase of CD8␣␣ ϩhigh circulating ␥␦ T cells after Salmonella exposure of chicks. Similarly, rat splenic T cells showed persistent down-regulation of the expression of CD8␤ but not of CD8␣ on ␥␦ T cells after in vitro activation (42). CD8␣␤ is more effective than CD8␣␣ in facilitating recognition of the same peptide antigen by TCR in vitro (47). Thus, the CD8␤ down-regulation could presumably be a mechanism for limiting the generation of TCR-mediated signals after Salmonella infection in vivo. Moreover, a requirement of transient CD8␣␣ expression for memory T-cell generation has been proven by use of E8 1 -deficient mice that cannot generate CD8␣␣ homodimers (31). In mice (28,29) and humans (27), the CD8␣␣ antigen is expressed on distinct T-cell subsets that constitutively display a memory phenotype. That ␥␦ T-cell subsets are also able to differentiate into memory T cells has already been shown in humans (7,20). However, it is not known whether the rise of CD8␣␣ homodimeric ␥␦ T cells can be considered a suitable marker for memory T-cell generation or memory T cells themselves in chickens.
In this study, avian CD8␣␣ ϩhigh homodimeric, CD8␣ ϩhigh ␤ ϩ heterodimeric, and CD8␣ Ϫ ␥␦ T-cell subsets were characterized concerning their mRNA expression for interleukin receptor chains involved in the memory generation (IL-2R␣ and IL-7R␣) and the apoptosis-related proteins Fas, Fas-ligand, and Bcl-x after Salmonella stimulation for the first time. It was demonstrated that all three avian ␥␦ T-cell subsets were able to express mRNA of the investigated proteins. CD8␣␣ ϩhigh homodimeric ␥␦ T cells showed relatively reduced levels of IL-7R␣ and Bcl-x but an elevated level of IL-2R␣ mRNA expression compared to the CD8␣␤ heterodimeric and CD8␣ Ϫ ␥␦ T cells. Interestingly, the expression of the IL-7R␣ mRNA increased over the time of investigation from 6 to 13 days after Salmonella immunization. These results, together with the observed significant increase of the percentage of CD8␣␣ ϩhigh ␥␦ T cells and the increased CD28 antigen expression on the CD8␣ ϩhigh ␥␦ T cells after immunization, led to the assumption that the observed CD8␣␣ ϩhigh homodimeric ␥␦ T-cell subset represented activated ␥␦ T cells after clonal expansion that underwent the contraction phase of the primary immune response against Salmonella enterica serovar Enteritidis. Recent data showed that T-cell stimulation induces considerable modifications of IL-7R␣ chain expression (37,40). While naive T cells are IL-7R␣ positive, T-cell activation leads to downregulation of IL-7R␣ and large-scale apoptotic episodes thereafter, resulting in a substantial reduction in the number of antigen-specific T cells. Only a subpopulation of T cells escapes death and remains a stable population of memory T cells. Progression to memory is accompanied by the reexpression of the IL-7R␣ chain (37,40). A decrease in the percentage of peripheral CD8 Ϫ ␥␦ T cells in Salmonella enterica serovar Enteritidis-immunized birds was found. This might be due to (i) an up-regulation of CD8 molecule expression or (ii) a fast migration of CD8 Ϫ ␥␦ T cells to sites of infection, such as the intestine or spleen. It has been shown earlier, as well as in this study, that ␥␦ T cells occur in higher numbers in the gut and spleen after immunization of chickens with Salmonella enterica serovar Typhimurium (8) and Salmonella enterica serovar Enteritidis, respectively. The peripheral blood may be a major source for ␥␦ T cells of the intestine and vice versa. A proliferation and rapid recirculation of intestinal ␥␦ T cells into the blood was seen in swine (43). Additionally, it has been shown that bovine CD8 Ϫ ␥␦ T cells of newborn calves express E-selectin ligands, which are responsible for adhesion, and that they are able to migrate to acute and chronic sites of inflammation (18,32).
After infection, splenic Salmonella clearance was more rapid in immunized animals than in nonimmunized controls. Immunization using the nonattenuated S. enterica serovar Enteritidis wild-type strain 147 produced the strongest protection: no challenge organisms were detected in the spleen during the course of the experiment. The rapid clearance of the Salmonella organisms after infection of Salmonella-immunized birds did not correlate with changes of T-cell composition in peripheral blood or spleen after infection. In contrast, our data showed a restriction of the secondary cellular immune response to the intestine after Salmonella enterica serovar Enteritidis infection of already immunized birds. A strong increase of CD8 ϩ and TCR1 ϩ T cells in the intestinal lamina propria could be observed quickly after challenge especially of SE 147-immunized animals. Thus, the Salmonella clearance in the spleen correlated well with the observed secondary immune response in the cecum, as higher percentages of CD8 ϩ and TCR1 ϩ T cells in the cecum mean lower numbers of Salmonella organisms in the spleen after challenge. Similarly, Beal et al. (6) found a restriction of the antigenspecific humoral and cellular immune response to the intestine after Salmonella enterica serovar Typhimurium infection of immunized chickens.
Generally speaking, immunization of day-old chicks using live attenuated or nonattenuated Salmonella strains resulted in variable reactions of ␥␦ T-cell subsets. The emergence of ␥␦ T-cell populations in the blood indicates that they are responding to the pathogen and/or providing a protective immune response. The lower ␥␦ T-cell response after Salmonella live vaccine immunization than after Salmonella wild-type strain immunization of day-old chicks on the other hand seems to lead to a weaker protection after Salmonella infection. Distinct reactions of the defined circulating ␥␦ T-cell subsets after Salmonella immunization and infection of chickens indicate that phenotypically characterized ␥␦ T-cell subsets may serve different immunological purposes in chickens. The detected down-regulation of the CD8␤ antigen on avian circulating CD8␣ ϩhigh TCR1 ϩ cells and the low IL-7R␣ mRNA expression of CD8␣␣ ϩhigh ␥␦ T cells after Salmonella immunization of young chicks in this study might be a consequence of a transient activation or specialization of a single ␥␦ T-cell subset in vivo that eventually may lead to the generation of specific memory T cells.