CCR7 Is Mainly Expressed in Teleost Gills, Where It Defines an IgD⁺IgM⁻ B Lymphocyte Subset

This article is featured in In This Issue, p.835

Homeostatic chemokine receptors (i.e., CCR7, CXCR4, and CXCR5) regulate trafficking and homing of lymphocytes and Ag-presenting dendritic cells (DCs) to and within secondary lymphoid organs and through body cavities (1, 2). Specifically, CCR7 and its ligands, CCL19 and CCL21, participate in both the regulation of leukocyte trafficking and homing to the lymph nodes (LNs) (3–5), because CCL19 and CCL21 are constitutively produced on the high endothelial vessels of the LNs and by fibroblastic reticular cells and follicular DCs (6, 7). It is known that CCR7 is expressed by semimmature and mature DCs, thymocytes during defined stages of their development, naive B and T cells, regulatory T cells and a subpopulation of memory T cells known as central memory T cells (5); however, to date, most studies on CCR7-mediated migration have focused on its effects on DCs and T lymphocytes.

Naive B lymphocytes express low levels of CCR7, and the expression increases upon engagement of the BCR, thus facilitating T–B cell interactions within the LN (8). Interestingly, it has been described that CCR7 delivers signals that direct cells to the T cell areas, whereas CXCR5 guides cells to B cell follicles. Following activation, follicular B cells downregulate the CXCR5 and upregulate CCR7. This shift in chemokine receptor expression allows for the transient mobilization of follicular B cells toward the T cell zone, where activated B cells can receive help from CD4⁺ Th cells (5). Additionally, it has been demonstrated that CCR7 mediates egress of B lymphocytes from body cavities, because CCR7 deficiency results in a massive accumulation of CD4⁺ and CD8⁺ T cells and B-2 B cells in the peritoneal and pleural cavities (2).

No lymph nodes or Peyer's patches are present in teleost fish, and therefore, the spleen and the head kidney (which also functions as the main hematopoietic organ) are the main secondary lymphoid structures. It might be possible that less evolved lymphoid organizations in mucosal tissues such as the gills or the digestive tract are also capable of initiating an immune response locally, because B and T cells have been reported in most fish tissues (9, 10). In any case, there are many aspects of B cell biology still unknown in teleost fish. Three different Igs have been reported in teleosts, namely IgM (11), IgD (12), and IgT (13), designated as IgZ in zebrafish (Danio rerio) (14). Although IgM and IgD seem to be essential Igs present in all teleost species, IgZ/T are only present in some of them (15). In rainbow trout (Oncorhynchus mykiss), IgM, IgD, and IgT are all present, but only two subpopulations of B cells have been characterized to date according to the presence of membrane Igs (_memIgs): _memIgM⁺_memIgD⁺_memIgT⁻ (IgM⁺ cells) and _memIgM⁻_memIgD⁻_memIgT⁺ (IgT⁺ cells) (13, 16). In catfish (Ictalurus punctatus), together with _memIgD⁺_memIgM⁺, a second population _memIgD⁺_memIgM⁻ has also been described in PBLs (17); however, its role in defense is still unknown. In fact, many aspects of IgD regulation and function remain an enigma in teleost as well as in mammals (18).

A homolog to mammalian CCR7 has been recently reported in rainbow trout (19), and although no ligands for trout CCR7 are known yet for this receptor, its transcriptional regulation during pathogen encountering suggests a major role of the CCR7 in the mobilization of the lymphocytes to mucosal sites. In this study, we have characterized CCR7 at the protein level, showing that CCR7⁺ cells are mainly present in the gills. Furthermore, we have demonstrated that a major subpopulation of these CCR7⁺ cells in the gills constitute a novel _memIgD⁺_memIgM⁻ B cell population, not previously reported in trout. These _memIgD⁺_memIgM⁻ cells can also be detected at very low levels in other organs such as spleen, but not in PBLs. A possible homology to other subpopulations of _memIgD⁺_memIgM⁻ cells found in mammals is discussed. Altogether, our results suggest a novel role for CCR7 in this subpopulation of B cells, not previously identified in other fish species.

Female rainbow trout (Oncorhynchus mykiss) adults of ∼20–50 g were obtained from Centro de Acuicultura El Molino (Madrid, Spain) and maintained at the animal facilities of the Centro de Investigación en Sanidad Animal in a recirculating water system at 16°C, with 12:12-h light/dark photoperiod. Fish were fed twice a day with a commercial diet (Skretting). Prior to any experimental procedure, fish were acclimatized to laboratory conditions for at least 2 wk.

Rainbow trout eyed eggs at ∼360 degree-days (DD) postfertilization were also obtained from Centro de Acuicultura El Molino. Eggs were maintained in the same conditions as above and fed at 2 wk posthatching with a commercial diet. Individuals at different stages of the development were sampled for RNA extraction and immunohistochemistry as follows: eyed eggs (∼306, ∼354, and ∼402 DD), immediate posthatch fry (hatch, ∼450 DD), prefirst feeding fry (PFF; ∼562 DD), at the stage of full disappearance of the yolk sac (first feeding [FF]; ∼674 DD), and fry 3 wk following FF (fry, 786 DD).

All of the experiments described comply with the Guidelines of the European Union Council (2010/63/EU) for the use of laboratory animals and have been approved by the Instituto Nacional de Investigación Agraria y Alimentaria Ethics Committee.

A polyclonal Ab (pAb) against the rainbow trout CCR7 receptor was generated in rabbit by s.c. immunization with 200 μg purified CCR7 peptide (CTPYTAGSDQVR) corresponding to an extracellular domain of the receptor coupled to keyhole limpet hemocyanin and emulsified with CFA. Twenty-eight and 42 d after the first immunization, rabbits were boosted with 100 μg peptide–keyhole limpet hemocyanin. For the second and third immunizations, CFA was replaced by the incomplete form. Antiserum was harvested on day 62, and pAbs were purified using affinity chromatography (Abyntek). The concentration of the purified CCR7 pAb was measured by Bradford protein assay and the specificity and titer evaluated through ELISA.

The specificity of this anti-CCR7 pAb was verified in cells transfected with a plasmid coding for trout CCR7. For this, the complete coding sequence for trout CCR7 was cloned into pDisplay vector (Invitrogen Life Technologies) through restriction digestion. The CCR7 sequence was fused at the N terminus to the murine Ig κ-chain leader sequence, which directs the protein to the secretory pathway. Furthermore, recombinant proteins expressed from pDisplay contain the hemagglutinin A (HA) epitope for detection. Integrity, fidelity, and orientation of the final construct (pDisplay-CCR7) were confirmed by sequencing. HEK 293 cells were then transfected with the pDisplay-CCR7 construction or the corresponding empty plasmid using the X-tremeGENE HP DNA transfection reagent (Roche) and following the anufacturer’s instructions. Briefly, cells were seeded at 1 × 10⁵ in 24-well plates (Nunc) and grown for 24 h in DMEM (Lonza) supplemented with 2 mM l-glutamine, 100 IU/ml penicillin with 100 μg/ml streptomycin (P/S), and 10% FCS (Invitrogen). A 2:1 ratio of microliter X-tremeGENE HP DNA transfection reagent to microgram DNA was used. Transfection complex was prepared to a final concentration of 0.01 μg/ml plasmid DNA in serum-free DMEM without P/S for dilution. The transfection complex (50 μl) was added to each well, and cells were incubated for 48 h. Following the incubation, cells were collected, washed twice with PBS, and incubated with TM lysis buffer (Millipore) containing protease inhibitors (50× PI solution; Millipore). The cell lysates were clarified by centrifugation, and the clarified supernatant was precipitated overnight at 4°C by incubation with the anti-CCR7 pAb at 1 μg/ml. The complexes were then incubated with protein A-agarose immunoprecipitation reagent (Santa Cruz Biotechnology) for 3 h at 4°C and washed four times in PBS. The resultant immunoprecipitates were analyzed by native PAGE and Western blot using a mouse anti-HA mAb (Miltenyi Biotec). The samples were resolved on a 12% separation gel under nonreducing conditions and transferred to polyvinylidene difluoride membranes (0.45-μm pore size, Immobilon-P; Millipore) using a Mini protean system (Bio-Rad). Prestained SDS-PAGE protein standards (161-0318; BioRad) were included. After transference, membranes were blocked in PBS with 0.1% Tween 20 and 5% skimmed milk for 1 h at room temperature, then incubated overnight at 4°C with the anti-HA mAb in the same buffer, followed by incubation with a sheep ECL anti-mouse IgG HRP-linked whole Ab (GE Healthcare) for 1 h at room temperature. After three washes in PBS-Tween 20 buffer, ECL prime Western blotting detection reagent (GE Healthcare) was added to the membranes for chemiluminescent detection of the reactive bands using autoradiography film.

Additionally, transfected cells harvested using 0.05% trypsin solution and washed in cold PBS were stained with the anti-trout CCR7 pAb in staining buffer (PBS supplemented with 1% FCS and 0.02% sodium azide) for 30 min at 4°C. After two washes in staining buffer, cells were coincubated with an isotype-specific secondary Ab, Alexa Fluor 488 goat anti-rabbit IgG (H+L) (Invitrogen), and with a mouse anti–HA-PE mAb (Miltenyi Biotec) for 30 min at 4°C. Cells were washed three times and analyzed on an FACSCalibur flow cytometer (BD Biosciences). Nontransfected cells and cells transfected with the empty plasmid were used as controls.

A further verification of the Ab specificity was performed blocking the anti-CCR7 pAb with the specific synthetic peptide used for immunization. For this, a 30 min blocking preincubation of the pAb with the specific peptide at 1:1, 1:5, and 1:10 molar ratios was performed before cell staining for flow cytometry. A preincubation of the pAb with an irrelevant peptide (EATQAANSTQTDC) at a ratio 1:5 was also performed as further control.

The percentage of CCR7⁺ cells was analyzed in blood-depleted (buffer-perfused) naive fish as well as in PBLs. Fish were anesthetized with 30 mg/l benzocaine in water. Blood was extracted from the caudal vein with a heparinized needle and diluted 10 times with Leibovitz medium (L-15; Invitrogen) supplemented with P/S, 10 U/ml heparin, and 2% FCS. Subsequently, a transcardial perfusion was conducted to remove the circulating blood from the tissues. Heart was cannulated through the ventricle into the bulbus arteriosus for perfusion with 30 ml teleost Ringer solution (pH 7.4), with 0.1% procaine, using a peristaltic pump at a constant flow rate of ∼5 ml/min, whereas the atrium was cut to drain the blood out of the circulatory system. After perfusion, tissues (gill, thymus, spleen, head kidney, and midgut) sampled for leukocyte FACS staining were placed in L-15 with P/S, 10 U/ml heparin, and 2% FCS. All tissues with the exception of the midgut were pushed through a 100-μm nylon mesh, and the resulting cell suspensions were placed onto a 30/51% discontinuous Percoll gradient and centrifuged at 500 × g for 30 min at 4°C. Blood cell suspensions were placed onto a 51% Percoll density layer and centrifuged as above. The interface cells were collected, washed at 500 × g for 5 min in L-15 containing 0.1% FCS, and resuspended in L-15 with P/S and 2% FCS. Midgut was opened lengthwise, washed in PBS, and cut into small pieces. The midgut cell-extraction procedure started with one round agitation at 4°C for 30 min in L-15 medium with P/S and 5% FCS followed by an agitation in PBS with 1 mM EDTA and 1 mM DTT for 30 min. Finally, gut tissues were digested with 0.15 mg/ml collagenase (Sigma-Aldrich) in L-15 for 1.5 h at 20°C. All leukocyte fractions were collected and pooled, pushed through a 100-μm nylon mesh, and separated in a 30/51% discontinuous Percoll gradient as above.

Leukocytes from the different tissues obtained as described above were adjusted to a final concentration of 1 × 10⁶ cells/ml and incubated with anti-CCR7 pAbs (2 μg/ml) in staining buffer for 20 min at 4°C. The cells were washed twice with staining buffer and incubated with Alexa Fluor 488 goat anti-rabbit IgG (H+L) (Invitrogen) for 20 min. After three washes, cells were analyzed on an FACSCalibur flow cytometer (BD Biosciences). For double staining of cell-surface markers on gill CCR7⁺ cells, anti-IgM (mAb 1.14 mouse IgG, coupled to PE, 0.1 μg/ml) (20), anti-IgT (mAb mouse IgG2b, coupled to biotin, 1 μg/ml) (21), anti-IgD (mAb mouse IgG, 5 μg/ml) (22), anti-myeloid (mAb mouse IgG, coupled to PE, 0.2 μg/ml) (23), and anti-CD8α (mAb rat IgG, 7 μg/ml) (24) mAbs were used. The conjugates/secondary Abs for anti-IgT and anti-IgD were PE-Cy5 streptavidin (BD Biosciences) and allophycocyanin cross linked, F(ab′)₂ fragment goat anti-mouse (H+L) (Invitrogen), respectively, whereas the secondary Ab for anti-CD8α detection was an R-PE F(ab′)₂ fragment of goat anti-rat IgG (H+L) (Invitrogen). IgD mAb directly coupled to Alexa Fluor 647 (AF 647 Ab labeling kit; Life Technologies) was used for the _memIgM/_memIgD double labeling.

Gill leukocytes were sorted into either CCR7⁺ and CCR7⁻ cells or IgD⁺ and IgD⁻ cells using a BD FACSAria III (BD Biosciences) cell sorter. After sorting, 4 × 10⁴ cells from each population were used to obtain total RNA and perform subsequent cDNA generation and real-time PCR using Power SYBR Green Cells-to-Ct Kit (Life Technologies) according to the manufacturer’s instructions. Primers used for the transcriptional analysis of sorted populations had been previously designed and are listed in Table I.

Gill leukocytes were seeded in eight-well culture slides coated with human fibronectin (BD Biocoat; BD Biosciences) for adherence. After 30 min, cells were fixed in 10% neutral buffered formalin for 15 min, washed three times in PBS, and stained with the anti-CCR7 pAb and/or the anti-IgD mAb overnight at 4°C. Secondary Abs were then added (Alexa Fluor 488 donkey anti-rabbit IgG and Alexa Fluor 594 donkey anti-mouse IgG; Invitrogen) for 1 h at room temperature. Nuclei were stained with Hoechst 33342 (Pierce).

To analyze the levels of transcription of CCR7 in comparison with all known chemokine receptors at different stages of early trout development, total RNA was extracted from the samples using a combination of TRIzol (Invitrogen) and RNAeasy Mini kit (Qiagen). One milliliter TRIzol was added to the eyed eggs and hatch stages, and 2 ml was used for the stages of PFF, FF, and fry. Samples were then homogenized through mechanical disruption in TRIzol using a disruption pestle. A total of 200 μl/ml chloroform was then added and the suspension centrifuged at 12,000 × g for 15 min. The clear upper phase was recovered, mixed with an equal volume of 100% ethanol, and immediately transferred to RNAeasy Mini kit columns (Qiagen). The procedure was then continued following the manufacturer’s instructions, performing on-column DNase treatment. Finally, RNA pellets were eluted from the columns in RNase-free water and stored at −80°C until used.

Two micrograms RNA was used to obtain cDNA in each sample using the Bioscript reverse transcriptase (Bioline Reagents) and oligo(dT)_12–18 (0.5 μg/ml) following the manufacturer's instructions. The resulting cDNA was diluted in a 1:5 proportion with water and stored at −20°C.

Real-time PCR was performed with a LightCycler 480 System (Roche) using FastStart SYBR Green Master mix (Roche). Reaction mixtures containing 5 μl 2× SYBR Green Supermix, 300 mM gene-specific primers (Table I), and 2 μl cDNA template were incubated for 10 min at 95°C, followed by 40 amplification cycles (30 s at 95°C and 1 min at 60°C). A melting curve for each PCR was determined by reading fluorescence every degree between 60°C and 95°C to ensure only a single product had been amplified. Efficiency of the amplification was determined for each primer pair using serial 10-fold dilutions of pooled cDNA, performed in the same plate as the experimental samples. The efficiency was calculated as E = 10 ^(−1/s), where s is the slope generated from the serial dilutions, when log dilution is plotted against Δ threshold cycle number. Differences in the relative expression level of the genes among the different stages of development were determined using the Pfaffl method (25), comparing the mean expression of each group to the mean expression of the earliest stage (∼306 DD).

Excised organs from adult fish and whole fingerlings were fixed in Bouin’s solution for 24 h. Fish at hatch, PFF, and FF were fixed in Davidson’s fixative for 24 h. In all cases, after fixation, samples were embedded in paraffin (Paraplast Plus; Sherwood Medical) and sectioned at 5 μm. After dewaxing and rehydration, sections were subjected to an indirect immunocytochemical method to detect trout CCR7, IgM, or IgD. After a heat-induced epitope retrieval in Tris-EDTA buffer (pH 9) (800 W for 5 min and 450 W for 5 min in a microwave oven), the sections were preincubated in a blocking solution consisting of 2% BSA (Sigma-Aldrich) in Tris buffer with 0.2% Tween 20 (TBT) at room temperature for 10 min,and 10% normal goat serum in TBT for 10 min. Then sections were incubated with primary Ab solution overnight at 4°C. Rabbit anti-trout CCR7 pAb and mouse mAb anti-trout IgD were used at concentrations of 5 and 10 μg/ml, respectively. The anti-trout IgD mAb recognizes both the membrane and secreted forms of the IgD (22). A mouse anti-IgM mAb (kindly donated by Dr. Kurt Buchmann, University of Copenhagen, and Dr. Karsten Skjoedt, University of Southern Denmark) (26, 27) was used at a concentration of 10 μg/ml, also labeling the membrane and the secreted form of the IgM. Following this incubation, unbound primary Abs were washed off using TBT. The tissue was covered with Dako REAL detection System, alkaline phosphatase/RED, and rabbit/mouse (DakoCytomation) biotinylated secondary Ab and following the manufacturer’s instructions for staining. The specificity of the reactions was determined by omitting the primary Abs. Mayer’s hematoxylin (DakoCytomation) was used as nuclear counterstain, and mounting was conducted with Aquamount (Merck). Slides were examined with an Axiolab (Zeiss) light microscope.

Rainbow trout were challenged with viral hemorrhagic septicemia virus (VHSV) through bath infection to determine whether the infection induced changes in the number of CCR7⁺ cells in the gills. The VHSV challenge was performed as previously described (28). Briefly, 12 fish of ∼200 g were transferred to 4 l viral solution containing 5 × 10⁵ TCID₅₀/ml VHSV strain 0771. After 1 h of viral adsorption with strong aeration at 14°C, fish were transferred to their water tanks. A mock-infected group treated in the same way was included as control. At days 1 and 3 postinfection, six to seven trout from each group were sacrificed by overexposure to MS-222. Gills and kidney were sampled and processed for FACS analysis as described above. The spleen was also removed from these fish to evaluate viral transcription through real-time PCR using the same protocol as described for the evaluation of immune gene expression. Statistical differences between the number of CCR7⁺ cells in infected and mock-infected fish were determined using a Student t test (p < 0.05).

The pAb specific to rainbow trout CCR7 generated recognized the CCR7 extracellular peptide via ELISA (titer >1:512,000) and was also screened for its ability to recognize the full-length CCR7 molecule in transfected cells. For this, HEK cells were transfected with a pDisplay-CCR7 construct that expresses the rCCR7 in the cell membrane together with an HA tag. Forty-eight hours posttransfection, the binding ability of the anti-CCR7 pAb to pDisplay-CCR7–transfected cells was assessed in cell extracts from transfected cells immunoprecipitated with the anti-CCR7 pAb. Fig. 1A shows a band of the precipitated protein, present only in pDisplay-CCR7–transfected cells and not in cells transfected with the empty pDisplay plasmid carrying the HA tag. These results reveal that both the anti-HA and the anti-CCR7 recognize the same molecule, demonstrating the specificity of the anti-CCR7 pAb. Additionally, the specificity of anti-CCR7 pAb was assessed by flow cytometry. The anti-CCR7 pAb recognized ∼5% of the transfected cells, all of which were HA positive (data not shown). Finally, the binding specificity of the pAb to CCR7 was also assessed by blocking the Ag recognition site with the peptide used for the immunization. After the blockage of the Ab, only a residual population with nonspecific binding could be detected (Fig. 1B), indicating a specific recognition. Furthermore, this blockage was not observed when an irrelevant peptide was used. This control in which the pAb was preblocked with the specific peptide was included in all further flow cytometry experiments, and the values obtained were subtracted from the actual samples.

Once the specificity of the anti-CCR7 pAb was verified, we characterized the number of CCR7⁺ cells in different rainbow trout leukocyte populations, as well as their transcriptional responses (Table I). Perfused blood-depleted fish were always used to avoid contamination of blood cells in the tissue samples. We evaluated the number of CCR7⁺ cells in five to eight individual fish and consistently found that the highest population of CCR7⁺ cells was observed in the gills (Fig. 2A, Table II), with a mean percentage of 7.5% of cells within the leukocyte gate expressing CCR7 on their surface. A mean percentage of 3.7% CCR7⁺ cells were observed in the spleen, 1.8% in kidney, 1.7% in thymus, and the smallest proportion of CCR7⁺ leukocytes were obtained for midgut and PBLs (0.5%). The presence of CCR7⁺ cells in the gills, spleen, kidney, and thymus was also confirmed by immunohistochemistry, revealing a disperse distribution of CCR7⁺ cells inside the tissues (Fig. 2B).

Because CCR7 is an homeostatic receptor that participates not only in leukocyte trafficking during immune responses, but also in organ development and homeostasis, the transcription of CCR7 was assessed at key stages of the early trout development, comparing them to the levels of transcription of all other known chemokine receptors in rainbow trout. All chemokine receptors were transcribed as early as 306 DD postfertilization (Fig. 3A). CCR7 mRNA levels started to significantly increase after hatching, continuously increasing at each posterior step analyzed. An increased expression was observed after hatching for CCR6 and after FF for CCR9 and CXCR1, whereas no significant differences were observed among the different developmental stages for CCR9B, CCR13, or CXC receptor CXCR4.

To study whether this CCR7 expression at early life stages was also present in the gills, we conducted immunohistochemistry with samples obtained at hatch, PFF, and FF stages. Although some CCR7 staining was already visualized at hatching, the staining for CCR7 increased through the different developmental stages in the gills, in accordance with the results obtained in the transcriptional analysis of the complete individual (Fig. 3B). Our results reveal that CCR7⁺ cells are present in the gills from early developmental stages.

To determine which cells were expressing CCR7 in the gill in physiological conditions, we performed a double labeling of gill CCR7⁺ cells with specific leukocyte markers available for rainbow trout. Although _memIgM⁺ cells in the gills were the major leukocyte type (∼15.5%), only ∼18% of these _memIgM⁺ cells had CCR7 in the cell membrane (Fig. 4). A similar percentage was obtained for IgT cells that expressed CCR7 in the membrane. However, and despite the fact that _memIgD constituted only ∼8.8% of the gill leukocyte population, >64% of these cells were expressing CCR7. Interestingly, >75% of the cells expressing CCR7 were _memIgD cells. This experiment was repeated in five independent fish, and similar results were always observed, with some fish having all _memIgD cells expressing CCR7, and _memIgD cells accounting for most of the cells expressing CCR7 (80 ± 20% of _memIgD cells express membrane CCR7 [_memCCR7]). Myeloid cells were ∼14.3% of the gill leukocytes, but <15% of the cells expressed CCR7. Finally, only 4% of the CD8α cells in the gills expressed CCR7.

To confirm these results, we analyzed gene expression on the sorted CCR7⁺ population from the gills in comparison with the negative fraction. As expected, the transcripts for CCR7 were ∼10 times higher in sorted CCR7⁺ cells, but were also present in the negative fraction, suggesting that different cells types with no _memCCR7 may have the potential to express CCR7 in response to different stimuli (Fig. 5A). The sorted CCR7⁺ cell fraction contained mRNA for _memIgD as well as for secreted IgD, supporting the flow data. In contrast, although mRNA for total IgM was detected, the levels of transcription of the _memIgM remained undetectable. Additionally, transcripts for the MHC class II (MHC-II) and the DC marker CD83 were also detected in gill CCR7⁺. In correlation with the flow data, no CD3 or CD8α transcripts were found in the CCR7⁺ sorted population, strongly suggesting that T cells in the gills do not express _memCCR7 in physiological conditions.

As the majority of IgD⁺ cells were also CCR7⁺, we performed a transcriptional analysis of sorted IgD⁺ cells from the gills (Fig. 5A). We found that these cells transcribed CCR7 at levels ∼10 times higher than the negative fraction. These _memIgD cells transcribed both _memIgD as well as secreted IgD, suggesting that cells that still retain _memIgD also secrete IgD. In contrast, sorted IgD⁺ cells did not contain detectable levels of total or _memIgM transcripts, pointing again to the existence of the _memIgD⁺/_memIgM⁻ cell type in the gills. Interestingly, these cells also contained important mRNA levels of MHC-II and CD83.

To verify that a subpopulation of cells in the gills was coexpressing CCR7 and IgD, we also performed immunofluorescence in gill leukocytes (Fig. 5B). Multiple CCR7⁺ cells and IgD⁺ cells were found among gill leukocytes. Although some single-positive cells were detected for both markers, many double-positive CCR7⁺IgD⁺ cells were observed in concordance with the flow cytometry results.

Interestingly, when we compared the distribution of CCR7 in the gills to that of IgM and IgD by immunohistochemistry using serial sections from gill tissues, we observed that the distribution of IgM and IgD appeared very different (Supplemental Fig. 1). Apart from the strong IgM staining in the blood vessels of the secondary lamellae, IgM staining was visualized in the apical area of the primary lamellae as scattered cells. IgD staining was also observed in the apical area of the primary lamellae, but the distribution was more homogenous through the most exposed surface. Furthermore, positive IgD staining was also detected in and between secondary lamellae. When the distribution of CCR7 staining was compared, it resembled best that of IgD, being homogenously distributed through the primary lamellae as well as in and between secondary lamellae.

To determine that IgD⁺ cells bearing CCR7 constitute a new B lymphocyte subset with no IgM in the cell surface, we performed a triple IgM/IgD/CCR7 staining in gill leukocytes. We simultaneously performed the same analysis in splenocytes, because this population is characterized by low expression of CCR7. IgD/IgM/CCR7-labeled cells were analyzed by FACS, and leukocytes were identified and gated on the basis of forward light scatter (FSC)/side scatter (SSC). First, CCR7 expression levels were determined in leukocytes from gills (Fig. 6A) and spleen (Fig. 6C) and compared against the expression of _memIgD and _memIgM. As shown in Fig. 6A, most of the IgD⁺ cells showed surface CCR7, and there were virtually no IgM⁺ cells expressing CCR7 (0.49% CCR7⁺IgM⁺ against 10.7% CCR7⁺IgM⁻ cells). When _memIgD and _memIgM were plotted, we observed that almost no cells were double positive, but in contrast, _memIgD⁺_memIgM⁻ and _memIgD⁻_memIgM⁺ were the major cell types (Fig. 6B). CCR7 expression levels were much lower in the spleen. Although _memIgM⁺ splenocytes were very abundant in the spleen, just a very small fraction of this population expressed CCR7, which was rather expressed among the _memIgD⁺ splenocytes, because almost one third of the _memIgD⁺ cells expressed CCR7 (Fig. 6C). In this case, we could detect some _memIgD⁺_memIgM⁺ cells, although still ∼79% of the total number of _memIgD⁺ cells were _memIgM⁻ cells (Fig. 6D). Finally, to ascertain the contribution of the different B cell subsets to the expression pattern of CCR7 in these specific tissues, _memIgD⁺ and _memIgM⁺ populations in the gills (Fig. 6B) and spleen (Fig. 6D) were plotted to determine the CCR7 expression level in IgD⁺IgM⁻, IgD⁻IgM⁺, and IgD⁺IgM⁺ B cell subsets. In the gills, cells bearing only _memIgD showed the highest expression of CCR7, whereas cells bearing only _memIgM did not express the receptor. Moreover, _memIgD⁺_memIgM⁺ double-positive cells presented an intermediate level of CCR7 expression (Fig. 6B), suggesting that CCR7 expression was restricted to _memIgD⁺_memIgM⁻ or _memIgD⁺_memIgM⁺ cells and somehow excluded from _memIgD⁻_memIgM⁺ cells. Overall, a similar scenario was found in the spleen, although CCR7 expression levels were lower (Fig. 6D). Cells bearing only _memIgD contained a CCR7⁺ subpopulation, which presented the highest levels of CCR7 found in the spleen. _memIgM-bearing cells, not expressing _memIgD, did not express CCR7, whereas _memIgD⁺_memIgM⁺ double-positive B cells showed an intermediate CCR7 expression level, which was consistent with the results observed in the gills. Together, these data suggest that _memCCR7 expression in naive B cells is associated to the presence of IgD on the membrane of B cells, excluding the expression of the receptor from _memIgM⁺_memIgD⁻ B cells. The results also indicate that _memIgD⁺_memCCR7⁺ B cells are found preferentially in the gills.

To assess if an encounter with a pathogen can modify the distribution of the CCR7⁺ cell population in the gills, we performed a bath challenge with VHSV and then recorded the number of CCR7⁺ leukocytes at days 1 and 3 postinfection in both the gill and the kidney, a target organ for Ag presentation. We have previously established that at these days postinfection, VHSV induces the transcription of several chemokine genes in both gills and head kidney (28, 29). In some fish, the number of _memIgD cells was also assessed. First, we assessed viral transcription in the spleen of these infected fish to ensure that fish exposed to the virus became infected. Transcription of the N gene was detected in all infected fish at day 3 postinfection. As shown in Fig. 7A, despite the high variability in CCR7⁺ numbers observed among individual fish, the percentage of CCR7⁺ cells in the gills significantly decreased in response to VHSV in comparison with the levels observed in mock-infected controls at day 1 postinfection. On the contrary, the percentages of CCR7⁺ significantly increased in the head kidney of infected animals in comparison with mock-infected controls at both days 1 and 3 postinfection. When we analyzed the number of _memIgD⁺ cells in some of these samples, we observed that the level of _memIgD significantly correlated to the level of _memCCR7 in the gills (Fig. 7B). Based on these results and our previous findings that identified _memIgD⁺_memIgM⁻ cells as the major cell type expressing CCR7, we can speculate that these cells are playing a role in Ag sensing and presentation. In contrast, although the number of CCR7⁺ cells increase in the head kidney in response to the infection, _memIgD cells in these samples remained undetected. These results suggest two possible explanations for the increase in CCR7⁺ cells observed in head kidney in response to VHSV. Either activated CCR7⁺ cells from the gills are mobilized to the kidney while losing _memIgD in response to activation, or the viral infection itself triggers _memCCR7 expression in resident head kidney cells with no _memIgD.

The results from this study demonstrate that rainbow trout CCR7 is mainly expressed in the gills in physiological conditions. Within the gills, although we could detect a small percentage of IgM⁺ and IgT⁺ cells expressing CCR7, this receptor is mostly expressed in a subpopulation of B cells with _memIgD and no _memIgM that represents an important leukocyte subpopulation in gills (_memIgD⁺_memIgM⁻). In fact, the presence of CCR7 in the cell surface of trout B cells was associated to the expression of _memIgD, because B cells lacking IgD in the cell surface (_memIgM⁺_memIgD⁻) had no CCR7 expression.

Although IgD was mostly ignored in fish for some time after its discovery (12), recent evidence demonstrated the presence of secreted IgD, pointing to a conserved role of IgD in fish immunity (22). In mammals, IgM and IgD are coexpressed on the surface of naive B cells, which, upon Ag binding, downregulate IgD expression, accounting for the gradual disappearance of IgD from the cell surface of activated cells that goes along with somatic hypermutation and class-switch DNA recombination to diversify the Ig gene repertoire (30). These mechanisms still remain to be demonstrated in fish B cells, which do not undergo class switch. Naive B cells in trout blood and spleen have shown to express both IgM and IgD in the cell surface, with no single-positive B populations detected (22), although these experiments were performed in American rainbow trout, known to exhibit some genetic and immunological differences in comparison with European salmonid strains (31). On the contrary, some subpopulations of B cells expressing only IgD in the cell surface have been reported both in mammals (18, 32, 33) and fish (17). In mammals, two types of _memIgD⁺_memIgM⁻ cells have been described. A _memIgD⁺_memIgM⁻ population present mainly in the upper aerodigestive mucosa arises in humans after active IgM-to-IgD class switch. These plasmablast-like cells that retain IgD in the membrane secrete highly mutated mono- and polyreactive IgD, providing a layer of mucosal protection by interacting with pathogens and are either retained locally or circulated in the blood, where they can account for up to 0.5–1% of circulating B cells (18, 32). The second type of _memIgD⁺_memIgM⁻ make up to 2.5% of circulating B cells in humans. These naive B cells have Ab V region genes in an unmutated configuration, thus are fully mature cells that are autoreactive and functionally attenuated and therefore have been cataloged as a new type of anergic B cells (33). Clonal anergy is closely related to self-tolerance, and although different types of anergic cells exist, anergy is always associated with the absence of IgM (34). Although most B cells expressing surface Ig that bind autoantigens are eliminated throughout the development, some B cells escape these checkpoints and survive in the periphery as autoreactive B cells. However, their functionality is strongly reduced so they will no longer react to self-Ags (clonal anergy). If we take into account that clonal anergy is a way to inactivate B cells stimulated early in development when only autoantigens would be presented (35), it might make sense that equivalent cells are present in the gills of teleost fish continuously exposed to water-borne Ags to which fish should not react. In contrast, the presence of _memIgD⁺_memIgM⁻ in trout gills could be an equivalent population to the IgD-secreting plasmablasts found in upper respiratory tract of mammals (32), thus additional studies are needed to determine if these trout cells correspond to naive anergic cells or if they are class-switched cells with an Ag experience. Although in our studies trout _memIgD⁺_memIgM⁻ were mainly present in the gills and not in peripheral blood (data not shown), catfish PBLs contain a distinct _memIgD⁺_memIgM⁻ B cell population that can make up to 60–80% of peripheral blood B cells, depending upon the individual fish. These catfish _memIgD⁺_memIgM⁻ B cells resembled human activated IgM⁻IgD⁺ B cells in that they have plasmablast morphology, exhibit restricted IgL isotype usage, and produce a secreted form of IgD (17). In fact, all evidence from humans to fish show that IgD displays a considerable diversity in structure and abundance both within a single individual at different moments and between individuals and has been suggested to be the most evolutionarily dynamic Ig class among all vertebrate Igs (18). Because of this association of CCR7 with these IgD⁺IgM⁻ cells in fish, it would be interesting to study whether CCR7 is also expressed in all similar IgD-expressing mammalian B cells. Interestingly, IgD⁺IgM⁻CD38⁺ B cells that are selectively derived from human nasopharynx-associated lymphoid tissue, also cataloged as tolerogenic because of their expression of Ig V-gene repertoires that may allow considerable cross-reactivity and autoimmunity (36), are known to express CCR7 (37). Of course, whether trout _memIgD⁺_memIgM⁻ are in fact homologs of human anergic B cells remains to be demonstrated. The function of anergic cells in mammals can be rescued when sufficiently stimulated, normally through an interaction with both CD40L and IL-4 (33). This activation leads to proliferate and differentiate into plasmablast cells.

Accordingly to mammalian literature, it was expected that CCR7⁺ cells in the gills would be mainly T cells. In mammalian GALT, the desensitization of CCR7 (38), the genetic disruption of CCR7 (5), or natural mutations in CCL19 and CCL21 (38, 39) lead to a reduced homing of T cells into the Peyer's patches. Interestingly, B cell homing to these secondary lymphoid tissues has been shown to be less CCR7 dependent (5, 40). Moreover, T cells are known to be present in teleost gills, even though there is some controversy in relation to actual numbers because some authors report very high numbers of T cells in gills (41), whereas others have reported that only ∼4–9% of the gill leukocytes are in fact T cells (9). Due to the lack of specific markers for surface Ags that define specific subpopulations, we exclusively used an anti-CD8α Ab in combination with the anti-CCR7 pAb. But the fact that CD8⁺ cells did not express _memCCR7 and CD3 mRNA was not detected in CCR7⁺ sorted cells strongly suggest that naive T cells in the gills do not express _memCCR7 in physiological conditions.

Additionally, we have established that CCR7 is modulated in the gills in response to a viral infection, suggesting an important role for these _memCCR7⁺_memIgD⁺ cells in the early stages of the mucosal immune responses. Additional studies should be performed to determine whether the cell types that express _memCCR7 in response to the infection remain the same as those observed in physiological conditions in tissues different from the gills. In the gills, our results point to _memIgD⁺ as the major cell type also during infection because despite the decrease in the number of CCR7⁺ cells in gills in response to VHSV, the number of CCR7⁺ cells significantly correlates with the number of IgD⁺ cells in both control and infected fish. Because VHSV infection of gill leukocytes either in vitro or in vivo did not decrease the levels of transcription of CCR7 (data not shown), the reduction in the number of CCR7 cells in the gills seems to be a consequence of the mobilization of CCR7⁺ cells from the gills. One possible explanation is that these CCR7⁺ cells are mobilized to the head kidney, an organ in close relation to the gills, where we observed significant increases in the number of CCR7⁺ cells. However, in this scenario, _memIgD should be downregulated upon activation because we did not detect significant numbers of _memIgD cells in infected head kidney. An alternative hypothesis is that other cell types that did not express _memCCR7 in physiological conditions express _memCCR7 in response to activation. This might be possible because even if there is a low number of _memCCR7⁺ cells in the nonstimulated head kidney, moderate levels of CCR7 mRNA were detected in sorted head kidney T cells and IgM⁺ cells (19). Our results point to CCR7 as a major player in the early teleost immune response in mucosal tissues. This is also supported by previous results from our group that demonstrated that the infection with an intestinal parasite provokes a significant increase in CCR7 mRNA levels in IgM⁺ and IgT⁺ cells in the trout gut (19). The possible role of CCR7 in IgT⁺ cells remains to be investigated at the protein level, because we did detect a small percentage of IgT⁺CCR7⁺ in the gills, and IgT transcripts were observed in CCR7 sorted cells. Although _memIgD was not found in blood IgT⁺ B cells (21), it might be possible that _memIgD is found in IgT⁺ cells from mucosal tissues. In mammalian lungs, CCR7 regulates normal pulmonary leukocyte homeostasis because the lack of CCR7 induces pulmonary hypertension involving perivascular infiltration of B and T lymphocytes (42). The reason for this is that CCR7 not only efficiently infiltrates leukocytes to extralymphoid tissues and sites of inflammation but also mediates the exit of leukocytes from these tissues through afferent lymph vessels (43), although again, most of these studies were focused on T cells (44).

In summary, we have studied the physiological distribution of CCR7⁺ cells for the first time, to our knowledge, in teleost, revealing a major presence of CCR7⁺ cells in the gills from the very early developmental stages. Furthermore, most of these gill CCR7⁺ cells define a subpopulation of B cells with _memIgD and no _memIgM, not previously identified in rainbow trout. Finally, the number of cells with _memCCR7 is regulated in response to a viral infection both in gills and head kidney, revealing an important role of these CCR7⁺IgD⁺ cells at the initial phases of teleost mucosal immunity.

We thank Oriol Sunyer, Uwe Fischer, and Bernd Köllner for providing the mAbs against IgT, CD8, and myeloid cells, respectively; Kurt Buchmann and Karsten Skjoedt for providing the mAb against IgM used in immunohistochemistry; Pierre Boudinot for helpful scientific discussions; Antonia Gonzalez and Lourdes Peña for technical assistance with immunohistochemistry; and María Castro for the schematic fish drawing.

The authors have no financial conflicts of interest.