of February 15, 2018.This information is current as

Humoral and Cellular Immune ResponsesMucosal Inductive Site for Virus-Specific Nasal-Associated Lymphoid Tissue Is a

Thurnheer, Petra Fundova and John J. CebraAdrian W. Zuercher, Susan E. Coffin, M. Christine

http://www.jimmunol.org/content/168/4/1796doi: 10.4049/jimmunol.168.4.1796

2002; 168:1796-1803; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/168/4/1796.full#ref-list-1

, 21 of which you can access for free at: cites 41 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2002 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on February 15, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on February 15, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Nasal-Associated Lymphoid Tissue Is a Mucosal Inductive Sitefor Virus-Specific Humoral and Cellular Immune Responses1

Adrian W. Zuercher,* Susan E. Coffin,† M. Christine Thurnheer,* Petra Fundova,‡ andJohn J. Cebra2*

Peyer’s patches are known as mucosal inductive sites for humoral and cellular immune responses in the gastrointestinal tract. Incontrast, functionally equivalent structures in the respiratory tract remain elusive. It has been suggested that nasal-associatedlymphoid tissue (NALT) might serve as a mucosal inductive site in the upper respiratory tract. However, typical signs of mucosalinductive sites like development of germinal center reactions after Ag stimulation and isotype switching of naive B cells to IgAproduction have not been directly demonstrated. Moreover, it is not known whether CTL can be generated in NALT. To addressthese issues, NALT was structurally and functionally analyzed using a model of intranasal infection of C3H mice with reovirus.FACS and histological analyses revealed development of germinal centers in NALT in parallel with generation and expansion ofIgA� and IgG2a� B cells after intranasal reovirus infection. Reovirus-specific IgA was produced in both the upper respiratory andthe gastrointestinal tract, whereas production of reovirus-specific IgG2a was restricted to NALT, submandibular, and mesentericlymph nodes. Moreover, virus-specific CTL were detected in NALT. Limiting dilution analysis showed a 5- to 6-fold higherprecursor CTL frequency in NALT compared with a cervical lymph node. Together these data provide direct evidence that NALTis a mucosal inductive site for humoral and cellular immune responses in the upper respiratory tract. The Journal of Immu-nology, 2002, 168: 1796–1803.

T he mucosal immune system consists of two functionallydistinct types of tissue: 1) inductive sites, where naive Band T cells are clonally selected and expanded upon Ag

contact; and 2) effector sites, where activated B and T cells relocateafter Ag-priming in inductive sites to express their effector functions.This concept is best established for the components of the intestinalmucosal immune system. They are organized in the so-called gut-associated lymphoid tissue (GALT),3 and include Peyer’s patches(PP), mesenteric lymph nodes (LN), and dispersed lymphoid cells inthe epithelial layer and in the gut lamina propria (1). PP are inductivesites and have been described as the major location for Ag-specific Bcell activation and isotype switching to IgA (2) and generation ofIgA� memory B cells (3), as well as for the induction of Ag-specificCTL (4). Generally, these primed B and T cells emigrate from PP,undergo terminal differentiation, and eventually home to the laminapropria and the intra-epithelial lymphocyte compartment of the gut(5–7). In contrast, little is known about the anatomic location andfunctional potential of inductive and effector sites in the respiratory

tract. However, the known compartmentalization of mucosal immuneresponses depending on the route of Ag administration (8, 9) suggeststhat structures other than PP, e.g., in the respiratory tract, may functionas mucosal inductive sites.

In the murine upper respiratory tract, nasal-associated lymphoidtissue (NALT) is believed to be the equivalent of the Waldeyer’sring of humans (10). It consists of a paired lymphoid tissue locatedat the floor of the nasal cavity lined by ciliated respiratory epithe-lium, and has been postulated in different studies as a possiblefunctional equivalent in the upper respiratory tract to PP in the gut(reviewed in Ref. 11). Several findings support this hypothesis.First, the cellular composition of NALT is similar to PP (12–14);both tissues contain a major population of naive B cells as well asnaive (CD45RBhigh) T cells (12). Second, it has been shown in ratsthat both NALT and PP have overlaying epithelium containing Mcells (15, 16) and follicle-associated epithelium (17) that mayserve as entry-sites for different pathogens (18, 19). Third, uponstimulation with Ag, the major isotype of Ab produced by NALTB cells is IgA (12, 13, 20, 21). Salivary glands and tear glandswere suggested as possible effector sites for IgA production (22,23), which is reflected by the presence of IgA Abs in saliva andtear fluid. Despite these similarities, differences between PP andNALT, such as a markedly diverging expression and function ofhoming receptors (24), have been reported. Other hallmarks ofmucosal inductive sites, like formation of germinal centers as wellas isotype-switching and expansion of surface IgA� B cells uponAg-stimulation have not been directly demonstrated in NALT.Moreover, even though specific CTLs have been observed in cer-vical or mediastinal LN after intranasal (i.n.) immunization (25,26), it is unknown whether CTLs are induced in NALT.

Using i.n. infection with reovirus serotype 1/Lang, an estab-lished mucosal pathogen that infects and elicits mucosal immuneresponses in both the gastrointestinal (4, 27–29) and respiratorytract (30, 31), we have structurally and functionally analyzed mu-rine NALT. Germinal centers, as well as IgA� and IgG2a� B

Departments of *Biology and†Pediatrics, University of Pennsylvania School of Med-icine, Philadelphia, PA 19104; and‡Division of Immunology and Gnotobiology, In-stitute of Microbiology, Czech Academy of Science, Prague, Czech Republic

Received for publication June 27, 2001. Accepted for publication December 4, 2001.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby markedadvertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by Grant AI-23970 from the National Institutes of Health(to J.J.C.), and a fellowship from the Swiss Foundation for Medical-Biological Grants(Schweizerische Stiftung fuer medizinisch-biologische Studien; to A.W.Z.). The FlowCytometry Facility of the Cancer Center at the University of Pennsylvania is sup-ported by the Lucille P. Markey Trust.2 Address correspondence and reprint requests to: Dr. John J. Cebra, Department ofBiology, University of Pennsylvania, 415 South University Avenue, Philadelphia, PA,19104-6018. E-mail address: jcebra@sas.upenn.edu3 Abbreviations used in this paper: GALT, gut-associated lymphoid tissue; NALT,nasal-associated lymphoid tissue; i.n., intranasal(ly); PP, Peyer’s patch; LN, lymphnode; UEA, Ulex Europaeus agglutinin I; BALT, bronchus-associated lymphoidtissue.

Copyright © 2002 by The American Association of Immunologists 0022-1767/02/$02.00

by guest on February 15, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

cells, were induced and expanded in NALT, and potent reovirus-specific IgA responses were detected in the upper respiratory tract(NALT, palatine salivary glands, and submandibular LN) and gas-trointestinal tract (PP, small intestine, and mesenteric LN). Impor-tantly, virus-specific CTLs were detected in NALT as well as inmediastinal, submandibular, and cervical LN. Limiting dilutionanalysis revealed a 5- to 6-fold higher precursor-CTL frequency inNALT compared with cervical LN. Thus, NALT is a mucosalinductive site in the upper respiratory tract for specific humoraland cellular immune responses.

Materials and MethodsMice and media

Male C3HeB/FeJ (referred to as C3H), BALBc/ByJ, and DBA/2 mice werepurchased from The Jackson Laboratory (Bar Harbor, ME). C.B-17 (C.B-Igh1b/IcrTac) mice were from Taconic Farms (Germantown, NY). Allmice were used at the age of 8–14 wk. Stock mice were housed in theanimal facility of the Department of Biology, University of Pennsylvania(Philadelphia, PA). After infection with reovirus, mice were kept physi-cally separated from naive stock mice in a Trexler plastic isolator (StandardSafety, McHenry, IL).

L-929 fibroblasts were grown in M199 medium (Life Technologies,Grand Island, NY) containing 5% FCS (Life Technologies), 2 mM L-glutamine (Life Technologies), 1000 U/ml penicillin, and 0.1 mg/mlstreptomycin (Life Technologies). CTLs were grown in RPMI-1640(Life Technologies) containing 10% FCS, L-glutamine, penicillin,streptomycin, 50 �g/ml gentamicin (Life Technologies), and 50 �M2-ME (Sigma Aldrich, St. Louis, MO). For organ fragment cultures,Kennett’ s HY medium (Life Technologies) supplemented with 10%FBS, L-glutamine, penicillin, streptomycin, and gentamicin was used.

Histology, immunohistochemistry, and immunofluorescence

NALT was frozen in OCT compound (TissueTek; EMSCO, Philadelphia,PA), horizontal 5-�m cryosections were cut on a Cryocut 1800 cryotome(Leica; Dolbey-Jamison, Norristown, PA) and after air-drying fixed inice-cold acetone. Some slides were stained with hematoxylin (Mayer’shematoxylin; Sigma Aldrich) and eosin (Sigma Aldrich). For immunohis-tochemical staining, slides were incubated in 0.3% H2O2 for 30 min, thenblocked with Superblock (Pierce, Rockford, IL) for 30 min, and with bi-otin/avidin block (Vector Laboratories, Burlingame, CA) following themanufacturer’s instructions. Slides were incubated for 60 min with 50 �lof 2.5 �g/ml biotinylated anti-IgD (clone AMS9.1; BD PharMingen, SanDiego, CA), 2.5 �g/ml biotinylated anti-CD4 (clone GK1.5; BD PharM-ingen), 2.5 �g/ml biotinylated anti-CD8 (clone 53.6-72; BD PharMingen),or 4 �g/ml biotinylated Ulex Europaeus agglutinin I (UEA; Vector Lab-oratories), followed by an incubation with HRP-conjugated streptavidin(1/1000; BD PharMingen). All reagents were diluted in a solution of 10%Superblock in PBS. 3,3-Diaminobenzidine (Sigma Aldrich) was used assubstrate according to the manufacturer’s instructions, followed by coun-terstaining with hematoxylin (Gill’s hematoxylin; Fisher, Pittsburgh, PA),and overlaid with Permount (Fisher).

For immunofluorescence, slides were incubated for 30 min with 50 �l ofthe following FITC-conjugated reagents: PNA (Pierce, coupled to FITC inour laboratory as described in Ref. 32), 10 �g/ml anti-IgD (clone 11-26c.2a; BD PharMingen), 10 �g/ml anti-IgA (Southern Biotechnology As-sociates, Birmingham, AL), and 10 �g/ml anti-IgG2a (Southern Biotech-nology Associates). Sections were overlaid with Vectashield containing4�,6�-diamidino-2-phenylindole (Vector Laboratories).

Flow cytometry

Single cell suspensions (2 � 105–106/sample) of NALT or submandibular LNwere stained for 20 min at 4°C with FITC-conjugated anti-CD19 (clone 1D3;BD PharMingen), PE-conjugated anti-CD4 (GK1.5; BD PharMingen), FITC-CD8 (53-6.72; BD PharMingen), FITC-PNA, PE-� L chain (Southern Bio-technology Associates), FITC-IgA (Southern Biotechnology Associates), andFITC-IgG2a (Southern Biotechnology Associates). Cells were washed andfixed in 1% paraformaldehyde in PBS and analyzed on a FACScan flow cy-tometer (BD Biosciences, Mountain View, CA). WinMDI2.8 (The ScrippsResearch Institute, La Jolla, CA) software was used for evaluation.

Virus preparation and infection

Third passage stocks of reovirus type 1/Lang (33) were produced and pu-rified as described (34). For i.n. infection, mice were lightly anesthetized

with 0.1 mg/g body weight Avertin (2,2,2,-tribromoethanol; Aldrich, Mil-waukee, WI) and 1–2.5 � 107 PFU reovirus was applied to both nostrilswith a micropipette in a total volume of 25 �l saline/0.5% gelatin.

Virus titration

For determination of viral titers in various tissues, C3H mice (n � 3–4 pertime point) were sacrificed by CO2 asphyxiation and cervical dislocationon days 2, 5, 7, and 14 after i.n. infection. After perfusion through the rightventricle with 20 ml PBS, the upper right lobe of the lungs, the entiretrachea, one palatine salivary gland, and a 1-cm piece of terminal ilealsmall intestine were removed, washed, and weighed. The tissues were ho-mogenized in 3 ml saline/0.5% gelatin and serial dilutions incubated onmonolayers of L-929 fibroblasts in 6-well tissue culture plates (Costar,Cambridge, MA) for 45 min at 37°C, and thereafter overlaid with 3 ml of1% Agar in complete M199 medium and cultured at 34°C. Cultures wereoverlaid with more Agar/M199 after 3 and 6 days. Plaques were countedafter 7 days incubation.

Analysis of Ab production in organ fragment cultures

On days 0, 4, 7, 11, and 14 after i.n. infection, C3H mice (n � 3–4 per timepoint) were sacrificed and blood collected by heart puncture for serumisolation. After perfusion with 20 ml PBS, the entire small intestine andmesenteric LN were surgically removed. PP were visually detected andexcised from the small intestine. After decapitation, submandibular LNwere removed. NALT was isolated after removal of the mandible as de-scribed elsewhere (13). Palatine salivary glands were isolated after removalof the hard palate.

For organ fragment culture (35), tissues were sterilized by sequentialwashes as described in detail elsewhere (36). Individual PP or LN, 3 � 3mm pieces of small intestine from jejunum, individual palatine salivaryglands, and NALT from one mouse (a pair of NALT still attached to asmall piece of nasal epithelium overlaying the palate) were incubated inwells of 24-well tissue culture plates (Costar) in 1 ml complete Kennett’sHY medium for 7 days under a 90% O2/10% CO2 atmosphere at 37°C.Reovirus-specific IgM, IgA, and IgG2a Abs were measured by RIA. Forthis purpose, flexible polyvinyl plates (Serocluster; Costar) were coatedwith 2.5 � 109 particles of reovirus per well in 50 �l PBS overnight at 4°C.Plates were blocked with 1% BSA in PBS and incubated with organ frag-ment culture supernatant fluid overnight at 4°C. Thereafter, plates wereincubated for 6 h at room temperature with 125I-labeled anti-IgA, anti-IgG2a, or anti-IgM Abs (all from Southern Biotechnology Associates).Radioactivity of individual wells was measured using a 1272 Clinigammagamma counter (Wallac, Gaithersburg, MD). Total IgA Abs were mea-sured by RIA as described earlier (35), and a standard curve of purified,monoclonal IgA was used to convert cpm to nanograms per milliliter.

In vitro restimulation and analysis of cytotoxic lymphocytes

Seven days after i.n. infection of C3H mice with reovirus, single cell sus-pensions of NALT, mediastinal, cervical, and submandibular LN pooledfrom 16–20 mice were restimulated in vitro by incubation in 96-wellround-bottom microtiter plates (2 � 105/well; Costar) in the presence of5 � 104 virus-pulsed, irradiated, thioglycolate-elicited peritoneal exudatecells (4). After 24 h, ConA-conditioned medium (5% (v/v)) mixed withmethyl-a-D-mannopyranosid (100 �M; Sigma Aldrich) was added. Effectorcells were restimulated the same way after 7 days of culture, and after atotal of 13 days of culture, cellular cytotoxicity was assessed in a standard51Cr-release cytotoxicity assay (37). L-929 fibroblasts infected with reo-virus (or uninfected for control) were labeled with 100 �Ci 51Cr (NEN,Boston, MA) and then incubated with effector cells at different E:T ratios(3000 target cells/well). After 5 h of incubation in V-bottom microtiterplates (Costar), 100 �l supernatant fluids were collected, mixed with 1 mlscintillation fluid (Cytoscint; ICN Pharmaceuticals, Costa Mesa, CA), and�-emission measured on a LS6500 multipurpose scintillation counter(Beckman Coulter, Fullerton, CA).

Limiting dilution of precursor cytotoxic lymphocytes

Replicate cultures (n � 17–24) containing varying numbers of either NALT orcervical LN cells (4.7 � 102–1.2 � 105 for NALT, 3.9 � 103–5 � 105 forcervical LN) obtained from C3H mice 7 days after i.n. infection were restim-ulated in vitro for 7 days as described above. Thereafter, the contents of in-dividual microtiter wells were harvested and equally divided into two wells ofV-bottom microtiter plates containing either reovirus-infected or -uninfectedL-929 cells (3000 target cells/well, labeled with 51Cr) for a 5 h standard 51Cr-release cytotoxicity assay as described above. Cultures were considered todemonstrate cytotoxicity if the resulting 51Cr release was at least three timesmore than background of spontaneous release. Cytotoxicity was considered

1797The Journal of Immunology

by guest on February 15, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

virus-specific if the release from infected targets was at least 20% higher frominfected than from uninfected targets. The 95% confidence intervals of thelinear regression curves were calculated using the Sigmaplot 2.00 software(Jandel Corporation, SSPS, Chicago, IL).

ResultsStructure and cellular composition of naive murine NALT

We analyzed the structure of NALT by immunohistochemistry. Asshown in Fig. 1, NALT in the naive C3H mouse is organized indistinct B and T cell areas (Fig. 1). As in PP, B cells reside in follicularareas, whereas T cells, particularly CD8� T cells, predominantly oc-cupy the parafollicular spaces between the B cell follicles. NALT islined with an epithelium that binds the lectin UEA (Fig. 1). UEA isspecific for �-L-fucose residues which are typically present on Mcells and follicle-associated epithelium (16, 38). Staining with labeledUEA was restricted to the epithelium on the luminal side of NALT(Fig. 1). A detailed view of the luminal epithelium revealed stainingof individual cells within the epithelium. These data demonstrate that

NALT displays the typical structure and organization of a secondarylymphoid tissue and contains epithelial cells that potentially enable itto efficiently take up Ag from the nasal cavity.

As shown in Fig. 2, B cells were the most abundant cell populationin NALT comprising 47–78% of lymphocytes (Fig. 2, upper panel)depending on the strain of mice. The ratio of B:T cells varied from3.8:1 in C.B17 to 0.9:1 in C3H mice, whereas the CD4:CD8 ratiovaried from 3:1 in C.B17 to 1.6:1 in C3H mice (Fig. 2, lower panel).These data reveal that the cellular composition of NALT is similar toPP. As most studies using reovirus have been done in C3H mice, thisstrain was used for all further analyses.

Induction of germinal centers and generation of IgA� andIgG2a� B cells in NALT after infection with reovirus

To address the function of NALT, C3H mice were infected i.n.with 1–2.5 � 107 PFU of reovirus serotype 1/Lang in a volume of25 �l. Replicating virus was detected in the respiratory tract

FIGURE 1. Histological analysis of naive murine NALT. Serial horizontal sections of naive mouse NALT were stained with biotinylated anti-IgD andHRP-labeled streptavidin, biotinylated anti-CD4, and HRP-streptavidin, or biotinylated anti-CD8 and HRP-streptavidin. Original magnification �100.Sections were counterstained with Gill’s hematoxylin. Serial horizontal sections of paired NALT (N) were stained with H&E (�40) or biotinylated UEAand HRP-streptavidin or HRP-streptavidin alone (�200). Staining of the epithelium lining the luminal side of NALT is indicated by the arrows (�40).Staining of individual cells at higher magnifications is shown (�200 and �400). NC, nasal cavity. Sections were counterstained with Mayer’s hematoxylin.

1798 NALT IS A MUCOSAL INDUCTIVE SITE

by guest on February 15, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

(NALT, palatine salivary gland, lung, and trachea), and to a lesserextent in the small intestine, and was cleared within 7–14 daysafter infection (Fig. 3). These data demonstrate that the respiratorytract is the major site of viral infection after i.n. inoculation, butthat coincidental swallowing of virus leads to infection and pos-sible cross-priming of the gastrointestinal tract.

Next, we examined the effect of the virus infection on NALT.The average number of cells recovered from NALT increased 54%from 1.3 � 105 (n � 12 mice) before infection to 2.0 � 105 (n �91) at day 7 ( p � 0.003), and decreased to initial levels by day 14of infection (1.2 � 105; n � 14). In contrast, cell expansion wasmore dramatic and prolonged in submandibular LNs where cellnumbers increased 11-fold within the first 7 days after infection

(from 2.5 � 105/mouse in naive mice, n � 16, to 3.0 � 106/mouseat day 7, n � 19) and further to 1.8 � 106/mouse (n � 6) by day14 postinfection). FACS analysis showed that the percentage oftotal (CD19�) B cells remained stable during the course of infec-tion (data not shown). Similarly, the CD4:CD8 ratio did notchange during the course of infection. Induction of germinal centerB cells (PNA-binding, � L chainlow, Ref. 3) was observed inNALT on day 7 and increased through day 14, in parallel withexpansion of both surface IgAlow and IgG2alow germinal center Bcells, as well as surface IgAhigh and IgG2ahigh memory type B cells(Ref. 3; Fig. 4). Staining with anti-IgG2b Ab gave similar results(data not shown). A similar expansion of PNA�, IgA�, andIgG2a� B cells was observed in submandibular LNs, althoughwith slightly delayed kinetics compared with NALT (Fig. 4).

These findings were confirmed by immunofluorescence stainingof histological sections of NALT. NALT of naive mice was char-acterized by the presence of distinct areas of IgD� B cells (Fig. 5).No PNA-binding, IgA�, or IgG2a� cells were detected beforeinfection. Seven days after inoculation with reovirus, germinalcenters characterized by the presence of PNA-binding cells wereobserved preferentially at the periphery of NALT. These PNA-positive follicles colocalized with IgA-stained and IgG2a-stained

FIGURE 2. Lymphocyte composition of NALT. Single cell suspensionof NALT cells (2 � 105–106/sample) isolated from naive DBA/2, C.B-17,BALB/c, or C3H mice were stained with FITC-labeled anti-CD19, PE-anti-CD4, or FITC-anti-CD8. Numbers represent the percentage of positivecells among gated lymphocytes. One of two identical experiments isshown.

FIGURE 3. Clearance of reovirus after i.n. infection. Lightly anesthe-tized naive C3H mice were infected i.n. with 1–2.5 � 107 PFU reovirusserotype 1/Lang in a volume of 25 �l. Two, 5, 7, and 14 days postinfection,groups of mice (n � 3–4) were sacrificed and perfused with PBS. Palatinesalivary glands, one submandibular salivary gland, the upper left lobe oflung, trachea, and a 1-cm piece of terminal ileum were isolated, washed,weighed, and homogenized in PBS/0.5% gelatin. Serial dilutions were usedto infect monolayers of L-929 fibroblasts for a standard virus titration as-say. PFU were counted after 7 days incubation. Detection limit was �100PFU/g tissue. One of four similar experiments is shown.

FIGURE 4. Generation of germinal center (PNA-binding), sIgA�, andsIgG2a� B cells in NALT and submandibular LN after i.n. infection withreovirus: FACS analysis. Lightly anesthetized naive C3H mice were in-fected i.n. with 1–2.5 � 107 PFU reovirus serotype 1/Lang in a volume of25 �l. At 0, 5, 7, or 14 days postinfection, groups of mice (n � 3–4) weresacrificed and single cell suspensions of NALT cells (2 � 105–106/sample)were stained with PE-anti-mouse � L chain, FITC-PNA, FITC-anti-IgA, orFITC-anti-IgG2a. Numbers represent the percentage of positive cellsamong gated lymphocytes. One of four similar experiments is shown.

1799The Journal of Immunology

by guest on February 15, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

areas, whereas IgD� cells predominantly occupied the PNA-neg-ative regions. By day 14 of infection, PNA-binding cells wereconcentrated at the luminal side of NALT and strongly stainedindividual IgA� and IgG2a� cells were scattered throughoutNALT. These findings demonstrate that germinal center reactionswith generation and expansion of IgA- and IgG2a-switched B cellsare induced in NALT after i.n. infection with reovirus.

Production of reovirus-specific Abs in organ fragment cultures

Organ fragment cultures (35) of mucosal and lymphoid tissues ofthe respiratory and gastrointestinal tract were performed to addressproduction of total and virus-specific Abs after i.n. reovirus infec-tion. Marginal amounts of total IgA were produced by NALT andpalatine salivary glands of naive mice, whereas the components ofGALT produced substantial quantities of IgA before stimulationwith reovirus (Fig. 6A). After infection, total IgA production wasstimulated in NALT and palatine salivary glands, and simulta-neously a slight increase was observed in PP and small intestine.However, the total output of IgA was 4- to 150-fold higher (on day7 or 0, respectively) from PP compared with NALT. Similarly,production of total IgA was generally higher in small intestine andmesenteric LN compared with palatine salivary glands and sub-mandibular LN, respectively.

Maximal amounts of reovirus-specific IgA were produced byNALT 7 days after infection (Fig. 6B, top row). Whereas produc-tion rapidly declined from NALT, secretion of virus-specific IgAby palatine salivary glands was sustained for at least 2 wk. Onlymarginal amounts of virus-specific IgA were detected in other mu-cosal tissues of the oral cavity such as lip or tongue (data notshown). Large amounts of virus-specific IgA were detected in su-pernatant fluid of PP and small intestine fragment cultures indi-cating cross-priming during i.n. inoculation. Virus-specific IgA

was also produced by mesenteric and to a lesser extent by sub-mandibular LN which drain the gastrointestinal and upper respi-ratory tract, respectively. Compared with IgA, the production ofvirus-specific IgG2a was delayed in NALT, but by day 14 postin-fection, NALT was producing considerable amounts of virus-spe-cific IgG2a (Fig. 6B, middle row). Similarly, virus-specific IgG2aresponses were induced in submandibular and mesenteric LN. Incontrast, no appreciable production of virus-specific IgG2a oc-curred in PP, palatine salivary glands, or the small intestine. Virus-specific IgM was detected in PP, submandibular, and mesentericLN as early as 4 days postinfection, whereas no virus-specific IgMAbs were detected in supernatant fluids of NALT, palatine salivarygland, and small intestine fragment cultures (Fig. 6B, bottom row).The major isotype of virus-specific Ab induced in serum wasIgG2a concomitant with production of virus-specific IgM, but onlytrace amounts of virus-specific IgA (data not shown).

Generation of reovirus-specific CTL in NALT

It is not known to date whether cytotoxic cellular immune re-sponses can be induced in NALT. To address this, 7 days after i.n.infection with reovirus, cells from NALT, mediastinal, subman-dibular, and cervical LN from 16–20 mice were isolated andpooled. After in vitro restimulation for 13 days, the presence ofvirus-specific CTL was assessed in a standard 51Cr-release cyto-toxicity assay. Potent virus-specific CTL were obtained fromNALT, as well as from the LN draining the respiratory tract (Fig.7). Identical restimulation of NALT and LN cells isolated fromnaive mice did not result in the outgrowth of virus-specific CTL(Fig. 7).

However, limiting dilution analysis revealed a 5-fold higher fre-quency of precursor CTL in NALT compared with cervical LN(Fig. 8). This ratio is similar when the frequencies are adjusted to

FIGURE 5. Generation of germinal center (PNA-binding), IgA�, and IgG2a� B cells in NALT after i.n. infection with reovirus: histological analysis.Lightly anesthetized naive C3H mice were infected i.n. with 1–2.5 � 107 PFU reovirus serotype 1/Lang in a volume of 25 �l. At 0, 7, or 14 days afterinfection, NALT of individual mice was isolated and frozen. Serial horizontal sections were stained with FITC-labeled anti-IgD, FITC-PNA, FITC-anti-IgA,or FITC-anti-IgG2a. Luminal side of NALT is at the bottom of the picture. Original magnification �100.

1800 NALT IS A MUCOSAL INDUCTIVE SITE

by guest on February 15, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

total numbers of CD8� cells in the initial NALT or cervical LNpopulations (NALT: 23.5% CD8� cells, frequency � 962/106;cervical LN: 27.8% CD8� cells, frequency � 165/106; ratio �5.8:1).

DiscussionIn this study, we have analyzed the induction of reovirus-specific humoral and cellular immune responses in the upperrespiratory tract. We found a number of striking functionalsimilarities between the mucosa-associated lymphoid tissue ofthe upper respiratory and the gastrointestinal tract. Both sitescontain organized secondary lymphoid tissue lined by epithe-lium which possibly enables them to efficiently absorb Ag. At

both sites, germinal center reactions with expansion of specificIgA� and IgG2a� B cells ensue from infection with reovirus.Furthermore, both sites contain similar draining LN, i.e., mes-enteric LN in the gastrointestinal and submandibular LN in the

FIGURE 8. Limiting dilution analysis of precursor CTL in NALT (F)and cervical LN (E). Lightly anesthetized naive C3H mice were infectedi.n. with 1–2.5 � 107 PFU reovirus serotype 1/Lang in a volume of 25 �l.After 7 days, mice (n � 20) were sacrificed and perfused with PBS. Var-ious dilutions of single cell suspensions were cultured for 7 days as de-scribed above. Contents of individual wells were assessed for cytotoxicityagainst 51Cr-labeled reovirus-infected or uninfected L-929 fibroblasts in a51Cr-release assay. Results were evaluated as described in Materials andMethods. Frequencies (f) of precursor CTL among total cells are shown.

FIGURE 6. Production of total IgA and reovirus-specific Abs in theupper respiratory and the gastrointestinal tract after i.n. infection with reo-virus. Lightly anesthetized naive C3H mice were infected i.n. with 1–2.5 �107 PFU reovirus serotype 1/Lang in a volume of 25 �l. At day 0, 4, 7, 11and 14 days postinfection, groups of mice (n � 3–4) were sacrificed andperfused with PBS. NALT, PP, palatine salivary glands, a 1-cm piece ofjejunum, and submandibular and mesenteric LN were isolated and steril-ized by sequential washing. Organ fragment cultures of the tissues wereincubated for 7 days. A, Total IgA was measured by RIA using anti-Fabcoated polyvinyl plates. Serial dilutions of organ fragment culture super-natant fluid were incubated, and bound Abs detected with polyclonal 125I-labeled anti-IgA Ab. Cpm were converted to ng/ml using a standard curveof monoclonal-purified IgA. B, Supernatant fluids were assessed for reo-virus-specific IgA (top panels), IgG2a (middle panels), and IgM (bottompanels) by RIA. Polyvinyl plates coated with 2.5 � 109 particles of reo-virus per well were incubated with organ fragment culture supernatant fluidand bound Abs were detected with polyclonal 125I-labeled anti-IgA, anti-IgG2a, and anti-IgM Ab. Results from one of four similar experiments areshown.

FIGURE 7. CTL in NALT after i.n. infection with reovirus. Lightlyanesthetized naive C3H mice were infected i.n. with 1–2.5 � 107 PFUreovirus serotype 1/Lang in a volume of 25 �l. After 7 days, mice (n �16–20) were sacrificed and perfused with PBS. Single cell suspensionswere cultured in microcultures in the presence of virus-pulsed, irradiatedAPCs and 5% (v/v) ConA-conditioned medium. Cells were restimulatedafter 7 days of culture with fresh APCs and ConA-conditioned medium. Asa control, cells from naive mice were cultured under identical conditions(squares). After 13 days of culture, cells were used in a 5Cr-release cellularcytotoxicity assay with 51Cr-labeled, reovirus-infected (filled symbols) oruninfected (open symbols) L-929 fibroblasts as targets. After 5 h of incu-bation, �-emission of 100 �l supernatant fluids was measured on a scin-tillation-counter. Results show the percentage of specific lysis after sub-traction of spontaneous release. Spontaneous release was never above 18%of total release. One of three similar experiments is shown.

1801The Journal of Immunology

by guest on February 15, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

respiratory tract, that may serve to amplify the responsesgenerated in PP or NALT, respectively.

It is well accepted that mesenteric LN serve to amplify immuneresponses initiated in PP. Unlike this, it remains elusive which, andwhether at all, LN of the respiratory tract fulfill a similar functionfor NALT. However, some of our findings may implicate such arole for the submandibular LNs. Whereas only modest expansionof cells was observed in NALT after i.n. infection (�2-fold withinfirst 7 days postinfection), a more pronounced expansion of cellsoccurred in submandibular LN (11-fold) over the same time pe-riod. Furthermore, increased numbers of cells were still found insubmandibular LN 14 days postinfection, although by that timecell numbers in NALT were not significantly different from naivemice. In contrast, our analysis of precursor CTL frequencies inNALT vs cervical LN (and presuming similar numbers for sub-mandibular LNs) revealed a higher frequency of virus-specificCTL in NALT (Fig. 7). Moreover, kinetic analysis of generation ofPNA�, IgA�, and IgG2a� B cells showed a delay in the genera-tion of these cells in submandibular LN compared with NALT(Fig. 4), i.e., whereas percentages of PNA�, IgA�, and IgG2a�

cells peaked or reached a plateau after �7 days in NALT, therespective cells appeared later in submandibular LNs. Taken to-gether, these findings may suggest that submandibular LN can in-deed amplify the specific responses generated in NALT, and thatthese structures are functionally equivalent to PP and mesentericLN, respectively.

Despite these similarities, the essentially different physiologicalfunctions of the respiratory tract (gas exchange) vs the gastroin-testinal tract (uptake of food) result in distinct requirements for theimmune components present at these sites. GALT largely ignoresfood Ags to allow uptake of nutrients, and mounts only minimalimmune responses to the components of the resident bacterialflora, but at the same time serves to exclude potential pathogens. Incontrast, the immune components of the respiratory tract, in par-ticular the lower respiratory tract, are not permissive and attemptto keep the noncolonized lungs sterile. These functional differ-ences among a number of subtle differences might be best reflectedin the predominance of different isotypes at these sites. Clearly,IgA is the major isotype secreted in the gut, and its limited in-flammatory effector potential is the ideal “low-key” isotype for thissite. In contrast, there is increasing evidence that the lower respi-ratory tract is mainly controlled and protected by IgG Abs thathave broader effector potential than IgA. Studies in influenza-in-fected SCID mice after transfer of neutralizing mAbs of differentisotypes (39), and more recently in influenza-infected IgA knock-out mice (40) have provided elegant evidence for this notion. IgGAbs may originate from the systemic circulation and enter the lungby transudation (41), or be produced locally in bronchus-associ-ated lymphoid tissue (BALT). Indeed, BALT in rodents has beenshown to mount specific immune responses to Ags present in thelung lumen (reviewed in Ref. 42), as opposed to a more discretefunction in humans (43). It may be postulated that NALT mayserve an “intermediate” function between BALT and GALT; al-though it is in physical connection with the alimentary tract andnot sterile, NALT also serves as an important “gate-keeper” for thewell-protected lower respiratory tract. This intermediate statusmay be reflected in the mixed IgA/IgG2a isotype-pattern prevalentat this site. As shown in Fig. 6, NALT not only produces IgA, butalso considerable amounts of IgG2a Abs (Fig. 6B, middle row).Similarly, IgG2a was produced in larger quantities than IgA insubmandibular LN, whereas IgA predominated over IgG2a in mes-enteric LN. A more detailed comparison of the functions of BALTvs NALT in comparison to GALT will be particularly interesting.

We are currently investigating this in a model of reovirus infectionof rats.

Several other studies have described humoral immune responsesin murine NALT (12, 13, 20, 21). In these studies, Ab productionwas assessed using the ELISPOT technique. In general, the resultscorrelate well with our findings using the organ fragment culturetechnique. Both IgA- and IgG-producing cells were observed inNALT after i.n. infection with live influenza virus (21, 21, 44). Ourresults show local production of both IgA and IgG2a Abs in NALTafter i.n. infection with reovirus. This is in line with the reportedproduction of a mixed Th1/Th2-type cytokine profile after respi-ratory infection with reovirus (45). Others described IgA as theclearly predominating local isotype in NALT (12, 13). In thesestudies, cholera toxin subunit B was used either as Ag or as anadjuvant in combination with bacterial Ag, which may somehowmediate this skewing of the B cell response toward IgA. However,the use of the fragment culture technique may allow for a morequantitative analysis of Ab production than ELISPOT, as it inte-grates the number of specific cells with the actual secretion of Ab,yielding information about total Ab production capacity of an en-tire organ or tissue. For example, we show that 7 days postinfec-tion, NALT produces �4 times less total IgA than PP (Fig. 6A),which is probably due to the smaller number of total cells presentin NALT. Nevertheless, equal amounts of virus-specific IgA areproduced, which likely indicates a higher frequency of Ag-specificIgA-producing B cells in NALT than in PP.

In this study, we report the first direct demonstration of CTLgeneration in NALT. NALT CTL appear to kill infected targetsmore efficiently at lower E:T ratios than CTL from the drainingLNs of the same animals (Fig. 7). More strikingly, the 5- to 6-foldhigher precursor CTL frequency in NALT compared with cervicalLN (Fig. 8) clearly shows that NALT is a potent inductive site forspecific CTL responses upon i.n. infection. A similarly increasedprecursor CTL frequency in PP over peripheral LN was observedafter local gastrointestinal reovirus infection (4). It will be inter-esting to pinpoint effector sites of these CTL to establish whetherthey specifically emigrate to mucosal effector sites of the respira-tory tract through a homing-receptor mediated process, or whethera more random distribution to different nonlymphoid tissues oc-curs, as shown recently by Masopust et al. (46) in systemically andorally primed mice. Furthermore, we demonstrate the generationof germinal centers in NALT after i.n. infection along with expan-sion of IgA� and IgG2a� B cells, and local production of reovirus-specific IgA and IgG2a Abs. Altogether, these findings providedirect evidence that NALT is an inductive site for humoral muco-sal immune responses. Thus, tracking patterns of distribution ofCTL and specific B lymphoblasts after induction in local mucosalinductive sites (e.g., in NALT vs PP) may represent a promisingapproach to address cross-priming and communication betweendistant mucosal sites, and a way to experimentally assess questionsconcerning the regulation and functional characteristics of thecommon mucosal immune system.

AcknowledgmentsWe thank Haruka Hishiki for help with NALT isolation, and Alec McKayfor running the FACS and preparing radiolabeled Abs. We also thank Dr.Chris Cuff for his continuous support and helpful comments.

References1. Kato, T., and R. L. Owen. 1999. Structure and function of intestinal mucosal epi-

thelium. In Mucosal Immunology, 2nd Ed. P. L. Ogra, J. Mestecky, M. E. Lamm,W. Strober, J. Bienenstock, and J. R. McGhee, eds. Academic Press, San Diego,p. 115.

2. Craig, S. W., and J. J. Cebra. 1971. Peyer’s patches: an enriched source of pre-cursors for IgA-producing immunocytes in the rabbit. J. Exp. Med. 134:188.

1802 NALT IS A MUCOSAL INDUCTIVE SITE

by guest on February 15, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

3. Lebman, D. A., P. A. Griffin, and J. J. Cebra. 1987. Relationship between ex-pression of IgA by Peyer’s patch cells and functional IgA memory cells. J. Exp.Med. 166:1405.

4. London, S. D., D. H. Rubin, and J. J. Cebra. 1987. Gut mucosal immunizationwith reovirus serotype 1/L stimulates virus-specific cytotoxic T cell precursors aswell as IgA memory cells in Peyer’s patches. J. Exp. Med. 165:830.

5. Pierce, N. F., and J. L. Gowans. 1975. Cellular kinetics of the intestinal immuneresponse to cholera toxoid in rats. J. Exp. Med. 142:1550.

6. McDermott, M. R., and J. Bienenstock. 1979. Evidence for a common mucosalimmunologic system. I. Migration of B immunoblasts into intestinal, respiratory,and genital tissues. J. Immunol. 122:1892.

7. Cuff, C. F., C. K. Cebra, D. H. Rubin, and J. J. Cebra. 1993. Developmentalrelationship between cytotoxic �� T cell receptor-positive intraepithelial lym-phocytes and Peyer’s patch lymphocytes. Eur. J. Immunol. 23:1333.

8. Wu, H.-Y., and M. W. Russell. 1993. Induction of mucosal immunity by intra-nasal application of a streptococcal surface protein antigen with the cholera toxinB subunit. Infect. Immun. 61:314.

9. Haneberg, B., D. Kendall, H. M. Amerongen, F. M. Apter, J.-P. Kraehenbuhl, andM. Neutra. 1992. Induction of specific immunoglobulin A in the small intestine,colon-rectum, and vagina measured by a new method for collection of secretionsfrom mucosal surfaces. Infect. Immun. 62:15.

10. Kuper, C. F., P. J. Koornstrat, D. M. H. Hameleers, J. Biewenga, B. J. Spit,A. M. Dujivestjin, P. J. C. van Breda Vriesman, and T. Sminia. 1992. The roleof nasopharyngeal lymphoid tissue. Immunol. Today 13:219.

11. Sminia, T., and G. Kraal. 1999. Nasal-associated lymphoid tissue. In MucosalImmunology, 2nd Ed. P. L. Ogra, J. Mestecky, M. E. Lamm, W. Strober,J. Bienenstock, and J. R. McGhee, eds. Academic Press, San Diego, p. 357.

12. Wu, H.-Y., E. B. Nikolova, K. W. Beagley, and M. W. Russell. 1996. Inductionof antibody-secreting cells and T-helper and memory cells in murine nasal lym-phoid tissue. Immunology 88:493.

13. Heritage, P. L., B. J. Underdown, A. L. Arsenault, D. P. Snider, andM. R. McDermott. 1997. Comparison of murine nasal-associated lymphoid tissueand Peyer’s patches. Am. J. Respir. Crit. Care Med. 156:1256.

14. Asanuma, H., A. Hodson Thompson, T. Iwasaki, Y. Sato, Y. Inaba, C. Aizawa,T. Kurata, and S. Tamura. 1997. Isolation and characterization of mucosal nasal-associated lymphoid tissue. J. Immunol. Meth. 202:123.

15. van der Ven, I., and T. Sminia. 1993. The development and structure of mousenasal-associated lymphoid tissue: an immuno- and enzyme-histochemical study.Reg. Immunol. 5:69.

16. Takata, S., O. Ohtani, and Y. Watanabe. 2000. Lectin binding patterns in ratnasal-associated lymphoid tissue (NALT) and the influence of various types oflectin on particle uptake in NALT. Arch. Histol. Cytol. 63:305.

17. Jeong, K. I., K. Uetsuka, H. Nakayama, and K. Doi. 1999. Glycoconjugate ex-pression in follicle-associated epithelium (FAE) covering the nasal-associatedlymphoid tissue (NALT) in specific pathogen-free and conventional rats. J. Exp.Anim. Sci. 48:23.

18. Morin, M. J., A. Warner, and B. N. Fields. 1994. A pathway for entry of reovirusinto the host through M cells of the respiratory tract. J. Exp. Med. 180:1523.

19. Velin, D., G. Fotopopulos, F. Luthi, and J.-P. Kraehenbuhl. 1997. The nasal-associated lymphoid tissue of adult mice acts as an entry site for the mousemammary tumor retrovirus. J. Exp. Med. 185:1871.

20. Asanuma, H., C. Aizawa, T. Kurata, and S. Tamura. 1998. IgA antibody-formingcell responses in the nasal-associated lymphoid tissue of mice vaccinated byintranasal, intravenous and/or subcutaneous administration. Vaccine 16:1257.

21. Tamura, S., T. Iwasaki, A. Hodson Thompson, H. Asanuma, Z. Chen, Y. Suzuki,Z. Aizawa, and T. Kurata. 1998. Antibody-forming cells in the nasal-associatedlymphoid tissue during primary influenza virus infection. J. Gen. Virol. 79:291.

22. Wu, H.-Y., H. H. Nguyen, and M. W. Russell. 1997. Nasal lymphoid tissue(NALT) as a mucosal inductive site. Scand. J. Immunol. 46:506.

23. Carr, R. M., C. M. Lolachi, R. G. Albaran, D. M. Ridley, P. C. Montgomery, andN. L. O’Sullivan. 1996. Nasal-associated lymphoid tissue is an inductive site forrat tear IgA antibody responses. Immunol. Invest. 25:387.

24. Csencsits, K. L., M. A. Jutila, and D. W. Pascual. 1999. Nasal-associated lym-phoid tissue: phenotypic and functional evidence for the primary role of periph-eral node addressin in naive lymphocyte adhesion to high endothelial venules ina mucosal site. J. Immunol. 163:1382.

25. Porgador, A., H. F. Staats, B. Faiola, E. Giloba, and T. J. Palker. 1997. Intranasalimmunization with CTL epitope peptides of HIV-1 or ovalbumin with the mu-

cosal adjuvant cholera toxin induces specific CTL and protection against tumordevelopment in vivo. J. Immunol. 158:834.

26. Gallichan, W. S., D. C. Johnson, F. L. Graham, and K. L. Rosenthal. 1993.Mucosal immunity and protection after intranasal immunization with recombi-nant adenovirus expressing herpes simplex virus glycoprotein B. J. Infect. Dis.168:622.

27. Weinstein, P. D., and J. J. Cebra. 1991. The preference for switching to IgAexpression by Peyer’s patch germinal center B cells is likely due to the intrinsicinfluence of their microenvironment. J. Immunol. 147:4126.

28. Barkon, M. L., B. L. Haller, and H. W. Virgin VI. 1996. Circulating immuno-globulin G can play a critical role in clearance of intestinal reovirus infection.J. Virol. 70:1109.

29. Major, A. S., and C. F. Cuff. 1997. Enhanced mucosal and systemic immuneresponses to intestinal reovirus infection in �2-microglobulin-deficient mice.J. Virol. 71:5782.

30. Bellum, S. C., D. Hamadzic, A. Hodson Thompson, R. A. Harley, S. D. London,and L. London. 1996. Experimental reovirus serotype 1/strain Lang infection ofthe lung: a model for the study of the lung in context of mucosal immunity. LabInvest. 74:221.

31. Hodson Thompson, A., L. London, S. C. Bellum, D. Hamadzic, R. A. Harley,S. D. London. 1996. Respiratory-mucosal lymphocyte populations induced byreovirus serotype 1 infection. Cell. Immunol. 169:278.

32. Butcher, E. C., R. V. Rouse, R. L. Coffman, C. N. Nottenburg, R. R. Hardy, andI. L. Weissman. 1982. Surface phenotype of Peyer’s patch germinal center cells:implication for the role of germinal center B cell differentiation. J. Immunol.129:2698.

33. Tyler, K. L., and B. N. Fields. 1990. Reoviruses. In Virology, 2nd Ed.B. N. Fields, and D. M. Knipe, eds. Raven, New York, p. 1307.

34. Rubin, D. H., M. J. Kornstein, and A. O. Anderson. 1985. Reovirus serotype 1intestinal infection: a novel replicative cycle with ileal disease. J. Virol. 53:391.

35. Logan, A. C., K. P. Chow, A. George, P. D. Weinstein, and J. J. Cebra. 1991. Useof Peyer’s patch and lymph node fragment cultures to compare local immuneresponses to Morganella morganii. Infect. Immun. 59:1024.

36. Kushnir, N., N. A. Bos, A. W. Zuercher, S. E. Coffin, C. A. Moser, P. A. Offit,and J. J. Cebra. 2001. B2 but nor B1 cells can contribute to CD4� T cell-mediatedclearance of rotavirus in SCID mice. J. Virol. 75:5482.

37. Brunner, K. T., H. D. Engers, and J. C. Cerottini. 1976. The 51Cr release assayas used for quantitative measurement of cell-mediated cytolysis in vitro. InIn Vitro Methods in Cell-Mediated and Tumor Immunity. B. R. Bloom andJ. R. David, eds. Academic Press, New York, p. 423.

38. Clark, M. A., M. A. Jepson, N. L. Simmons, and B. H. Hirst. 1995. Selectivebinding and transcytosis of Ulex europaeus 1 lectin by mouse Peyer’s patchM-cells in vivo. Cell Tissue Res. 282:455.

39. Palladino, G., K. Mozdzanowska, G. Washko, and W. Gerhard. 1995. Virus-neutralizing antibodies of immunoglobulin G (IgG) but not of IgM or IgA isotypecan cure influenza virus pneumonia in SCID mice. J. Virol. 69:2075.

40. Mbawuike, I. N., S. Pacheco, C. L. Acuna, K. C. Switzer, Y. Zhang, andG. R. Harriman. 1999. Mucosal immunity to influenza without IgA: an IgAknockout mouse model. J. Immunol. 162:2530.

41. Scicchitano, R., R. F. Sheldrake, and A. J. Husband. 1986. Origin of immuno-globulins in respiratory tract secretions and saliva of sheep. Immunology 58:315.

42. Bienenstock, J., M. R. McDermott, and R. L. Clancy. 1999. Respiratory tractdefense: role of mucosal lymphoid tissue. In Mucosal Immunology, 2ndEd. P. L. Ogra, J. Mestecky, M. E. Lamm, W. Strober, J. Bienenstock, andJ. R. McGhee, eds. Academic Press, San Diego, p. 283.

43. Pabst, R. 1992. Is BALT a major component of the human lung immune system?Immunol. Today 13:119.

44. Matsuo, K., T. Iwasaki, H. Asanuma, T. Yoshikawa, Z. Chen, H. Tsujimoto,T. Kurata, and S. Tamura. 2000. Cytokine mRNAs in the nasal-associated lym-phoid tissue during influenza virus infection and nasal vaccination. Vaccine 18:1344.

45. Hodson Thompson, A., L. London, and S. D. London. 1999. Heterogeneouscytokine production by acutely stimulated bronchoalveolar T lymphocytes inreovirus 1/Lang-infected mice. Cell. Immunol. 198:120.

46. Masopust, D., V. Vezys, A. L. Marzo, and L. Lefrancois. 2001. Preferentiallocalization of effector memory cells in nonlymphoid tissue. Science 291:2323.

1803The Journal of Immunology

by guest on February 15, 2018http://w

ww

.jimm

unol.org/D

ownloaded from