Infection Mycobacterium Tuberculosis to Pulmonary Lymphocytes Are Associated with Immunity Integrin-Expressing T Β Expansion Of

Up-Regulation of VCAM-1 and Differential Expansion of β Integrin-Expressing T Lymphocytes Are Associated with Immunity to Pulmonary Mycobacterium tuberculosis This information is current as Infection of September 24, 2021. Carl G. Feng, Warwick J. Britton, Umaimainthan Palendira, Natalie L. Groat, Helen Briscoe and Andrew G. D. Bean J Immunol 2000; 164:4853-4860; ; doi: 10.4049/jimmunol.164.9.4853 Downloaded from http://www.jimmunol.org/content/164/9/4853

References This article cites 40 articles, 16 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/164/9/4853.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 24, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Up-Regulation of VCAM-1 and Differential Expansion of ␤ Integrin-Expressing T Lymphocytes Are Associated with Immunity to Pulmonary Mycobacterium tuberculosis Infection1

Carl G. Feng,* Warwick J. Britton,2*† Umaimainthan Palendira,* Natalie L. Groat,* Helen Briscoe,† and Andrew G. D. Bean3*

Immune responses rely on an intricate system of adhesion molecules to coordinate the homing and retention of lymphocytes in both secondary lymphoid tissues and at sites of infection. To deﬁne the events associated with pulmonary immune responses, the expression of endothelial addressins and integrins on T cells was analyzed during Mycobacterium tuberculosis infection. In infected lung, expression of endothelial VCAM-1, but not mucosal addressin cell adhesion molecule-1, was up-regulated from 4 wk ؉ postinfection and persisted to at least 12 wk. Subsequent analysis of the corresponding integrins expressed on lung CD4 and Downloaded from ؉ ␤ high ␤ ؊/low ␤ high CD8 T cells revealed an accumulation of 1 / 7 , and to a lesser extent 7 , integrin-expressing T cells during infection. ␣ high ␤ ؊/low ␤ ␣ Examination of integrin heterodimers showed that while 4 integrin was predominantly expressed on 1 / 7 cells, E ␤ high integrin was primarily associated with 7 . The majority of activated/memory T cells recruited during infection expressed high ␤ ␤ ␥ levels of 1 integrin and undetectable or low levels of 7 integrin. These T cells were capable of producing IFN- , a cytokine crucial high ␤ ؊ ␤ high ␤ for controlling M. tuberculosis infection. Rapid expansion of 1 , 7 , and 7 T cell populations in the lung upon secondary mycobacterial infection indicates the participation of these populations in the acquired immune response to the infection. Fur- http://www.jimmunol.org/ ␣ ␣ ␤ thermore, treatment of infected mice with mAb to 4 or 4 7 integrin led to a reduction in lymphocytes and increase in gran- ulocytes in the pulmonary inﬁltrate. These results reveal a crucial role for adhesion molecules in the generation of an effective pulmonary immune response to M. tuberculosis infection. The Journal of Immunology, 2000, 164: 4853–4860.

␣ ␤ he interactions between adhesion molecules on lympho- ficity of the integrin. 4 1 (VLA-4) integrin is mainly expressed cytes, known as homing receptors (HR),4 and their li- on leukocytes involved in binding to VCAM-1 and fibronectin (9). T gands on endothelium, vascular addressins, determine the Importantly, the VLA-4:VCAM-1 interaction facilitates recruit- specificity and magnitude of local immune responses (1–3). The ment of lymphocytes to inflamed extraintestinal mucosal tissues by guest on September 24, 2021 expression of addressins on activated endothelium directs the hom- (4, 10), exemplified by anti-VLA-4 and anti-VCAM-1 mAb treat- ing of lymphocytes from blood into sites of inflammation (4, 5) ment preventing the migration of lymphocytes into sites of inflam- Among these addressins, VCAM-1 is important in the attachment ␣ mation in vivo (11–13). Interestingly, 4 integrin can also associ- of lymphocytes to inflamed endothelium, following induction of ␤ ␣ ␤ ␣ ␤ ␣ ␤ ate with 7 integrin to form the 4 7 heterodimer. Like 4 1, 4 7 VCAM-1 expression by proinflammatory cytokines (6, 7). These integrin is able to bind to both VCAM-1 and fibronectin (14, 15), changes in the expression of vascular addressins result in an in- ␣ ␤ but a major ligand for 4 7 is the mucosal addressin cell adhesion creased efficiency of lymphocyte binding mediated by HR. molecule-1 (MadCAM-1). Because MadCAM-1 is only expressed One group of HR, the integrins, are heterodimic proteins con- on high endothelial venules of Peyer’s patches and mesenteric ␣ ␤ sisting of noncovalently associated - and -chains important in lymph nodes (16, 17), this interaction highlights the importance of regulating lymphocyte homing. They have been divided into sub- the ␣ ␤ heterodimer in the differential recirculation of lympho- ␤ ␤ ␤ ␤ 4 7 families according to their -chains (8), with the 1, 2, and 7 ␤ cytes to mucosal surfaces. Another member of the 7 integrin sub- integrin subfamilies important in the leukocyte-endothelial inter- family is ␣ ␤ . ␣ ␤ integrin retains lymphocytes within the ep- action (2). The ␣-chain of the heterodimer determines the speci- E 7 E 7 ithelium via binding to E-cadherin on epithelial cells (18, 19), rather than acting as a homing molecule for the mucosa (20–22). *Centenary Institute of Cancer Medicine and Cell Biology, Newtown, Australia; and Tuberculosis, primarily a lung disease, is the world’s leading †Department of Medicine, University of Sydney, Sydney, Australia cause of death from a single infectious agent, and represents 25% Received for publication May 3, 1999. Accepted for publication February 24, 2000. of all preventable deaths (23). The major protective immune re- The costs of publication of this article were defrayed in part by the payment of page sponse against M. tuberculosis infection is the activation of Ag- charges. This article must therefore be hereby marked advertisement in accordance specific CD4ϩ and CD8ϩ T cells. The secretion of IFN-␥ and with 18 U.S.C. Section 1734 solely to indicate this fact. other cytokines by these T cells leads to the containment of the 1 This work was supported by the National Health and Medical Research Council of Australia. C.G.F. is a recipient of the Australian Postgraduate Award. bacillus by activated macrophages (24, 25). The mechanisms that 2 Address correspondence and reprint requests to Dr. W. J. Britton, Centenary Insti- govern the recruitment of circulating lymphocytes into the M. tu- tute of Cancer Medicine and Cell Biology, Locked Bag No 6, Newtown, NSW, 2042 berculosis-infected lungs are not fully understood, particularly Australia. E-mail address: [email protected] with regard to the contribution of the different integrin families. 3 Current address: CSIRO, Division of Animal Health, Private Bag 24, Geelong, To explore the mechanisms that control lymphocyte migration Victoria, 3220 Australia. into the lung in normal and pathological situations, we used an 4 Abbreviations used in this paper: HR, homing receptor; BCG, bacille Calmette- Guerin; MadCAM-1, mucosal addressin cell adhesion molecule-1; VLA, very late aerosol model of M. tuberculosis infection in the mouse to inves- activation. tigate the changes in the expression of addressins on endothelium

and ␤ integrins on T cells in the lung during infection. In this color (clone CT-CD4), anti-CD8␣ Tri-color (clone CT-CD8a), and isotype study, we show that an up-regulation of endothelial expression of control Abs were purchased from Caltag (San Francisco, CA). Streptavidin VCAM-1 and differential expansion of ␤ integrin-expressing T conjugated with PE (Caltag) was used as a secondary reagent for biotin-labeled Abs. cells are important in the lymphocyte trafficking to the lungs following aerosol mycobacterial infection. These adhesion molecules Cell surface staining and flow cytometry are crucial to coordinate an appropriate pulmonary inflammatory The detailed procedures for the surface staining and FACS analysis of lung response. cells have been previously described (27). The staining of samples with isotype control Ab was used as reference to determine positive and nega- Materials and Methods tive populations. Mice Intracellular IFN-␥ staining C57BL/6 female mice were supplied by Animal Resources Centre (Perth, Single cell suspensions of lung cells were incubated in a six-well plate at WA, Australia), and were maintained in specific pathogen-free conditions 37°C for1htoremove adherent cells. Nonadherent cells (106/ml) were at the Centenary Institute animal facility until infection with then stimulated with plate-bound anti-CD3 mAb (PharMingen) (10 ␮g/ml) M. tuberculosis, when they were transferred to and maintained in a level 3 for 16 h in complete RPMI (RPMI supplemented with 10% FCS, 2 mM physical containment facility. Mice were used between 6 and 8 wk of age. L-glutamine, 10 mM HEPES, 0.5 ␮M 2-ME, 100 U/ml penicillin, and 100 ␮g/ml streptomycin). Brefeldin A (Sigma) (10 ␮g/ml) was added to the Bacteria and aerosol infection cultures for the final 4 h. Cells were washed and surface stained with rat anti-mouse CD4 or CD8 mAb (Caltag). The cells were fixed in 4% para- M. tuberculosis H37Rv (ATCC 27294) was grown in Proskauer and Beck formaldehyde for 20 min at room temperature, washed in permeabilization liquid medium (Difco, Detroit, MI), for 14 days at 37°C. Mycobacterium ␥ buffer (0.1% saponin in FACS buffer), then stained with anti-IFN- FITC Downloaded from bovis, bacille Calmette-Guerin (BCG; CSL strain), was prepared in sup- (clone AN18) in permeabilization buffer at 4°C for 30 min. Cells were then plemented Middlebrook 7H9 broth (Difco) for 14 days at 37°C. The bac- washed in permeabilization buffer, resuspended in FACS buffer, and ana- teria were washed, and enumerated on supplemented Middlebrook 7H11 lyzed on a FACScan flow cytometer (as described). Agar (Difco). Mice were exposed to M. tuberculosis H37Rv in a Middle- brook airborne infection apparatus (Glas-Col, Terre Haute, IN) at a pre- Immunohistology determined infective dose. Each mouse received ϳ102 viable bacilli per ␮ lung, as determined by culture of lung homogenates 24 h after infection. Air-dried frozen sections (4–6 m) were double stained by indirect im- munofluorescence with rabbit anti-cytokeratin polyclonal Ab (Dako, Car- Exposure to aerosol BCG occurred in an infection apparatus that delivered http://www.jimmunol.org/ an infectious dose of 103 viable BCG organisms to the lung. This resulted penteria, CA), to reveal epithelial cells, and either rat anti-mouse mAb in comparable kinetics in the bacterial load and T cell responses in the lung specific for E-cadherin (clone ECCD-2; Zymed, South San Francisco, CA), to that following aerosol M. tuberculosis infection (C. Feng and U. ICAM-1 (clone 3E2; PharMingen), MadCAM-1 (clone MECA-367; Palendira, unpublished data). For the secondary infection experiments, 8 PharMingen), or VCAM-1 (clone 429; PharMingen). This was followed by wk after primary infection with aerosol BCG, groups of mice were treated staining with either goat anti-rat or goat anti-hamster Ig conjugated to FITC with Isoniazid (Sigma, St. Louis, MO) (2% in drinking water) for 4 wk to (Caltag) and tetramethylrhodamine isothiocyanate-conjugated goat anti- reduce bacterial load. The mean (ϮSEM) CFU for three mice after anti- rabbit IgG (Southern Biotechnology Associates, Birmingham, AL). To in- biotic treatment and before secondary challenge was 227 (Ϯ32.15) per crease the sensitivity of staining, a single stain, three-step procedure for lung. Two weeks after the completion of drug treatment, half of these detection of MadCAM-1 was developed. Tissue sections were incubated infected and treated mice were exposed to a secondary aerosol infection with the mAb to MadCAM-1, followed by a rabbit anti-rat IgG (Dako). A with the same dose of the BCG as primary infection. The remaining mice FITC-conjugated goat anti-rabbit IgG (Caltag) was then used to visualize by guest on September 24, 2021 served as controls to illustrate the level of the integrin expression following the staining. Sections were stained with Abs, 20 min in a moist chamber at treated primary infection. room temperature, and after each step the sections were washed three times for 5 min each with PBS. Sections were examined on a Leitz DMR BE In vivo Ab treatment fluorescence microscope (Leica, St. Gallen, Switzerland). ␣ ␣ ␤ Protein G column-purified rat mAb to 4 (clone PS/2), 4 7 (clone Results DATK32), and an irrelevant control rat mAb (clone GL113) were used in Sustained up-regulation of VCAM-1 expression in the lungs of in vivo mAb treatment experiments. Mice were infected with aerosol M. tuberculosis and were injected (i.p.) with 500 ␮g of mAb on day 14, and M. tuberculosis-infected mice then on alternate days until day 28. Control mice received GL113 or were Recruitment of lymphocytes to nonlymphoid organs, such as the left untreated. This protocol was chosen because there was no significant increase in the numbers of lymphocytes in the lungs before day 14 postin- lung, is mediated by the interaction between addressins on endo- fection with aerosol M. tuberculosis (see Fig. 3B) (26). The concentration thelium and HR on lymphocytes. We compared the expression of of rat IgG (measured by ELISA) in the sera of mice was 4–18 ␮g/ml at ICAM-1, VCAM-1, MadCAM-1, and E-cadherin in normal lungs 24 h, and 2–5 ␮g/ml at 48 h after mAb treatment. At the end of the ex- and lungs at 2, 4, 8, and 12 wk following aerosol M. tuberculosis periments, one lung was homogenized and serially diluted for determina- infection. The expression of VCAM-1 was observed on the endo- tion of CFU of M. tuberculosis. The other lung was used for histological and FACS analysis. thelium only in infected lungs (Fig. 1). The up-regulated VCAM-1 expression was present from week 4 and maintained at least to Preparation of single cell suspensions from lung week 12 postinfection. By contrast, there was little change in the Animals were sacrificed by carbon dioxide narcosis at defined time points expression of ICAM-1 and E-cadherin (data not shown). Mad- after aerosol infection. The lungs were gently perfused with 20 U/ml hep- CAM-1 expression was not detected in either normal or infected arin (Fisons Pharmaceuticals, NSW, Australia) in PBS. Lung tissue was lung by either two- or three-step staining techniques, both of which minced and then incubated for 90 min at 37°C with shaking in RPMI revealed positive staining for MadCAM-1 expression in Peyer’s (1 lung/5 ml) supplemented with 50 U/ml collagenase I, type 4197 (Worth- ington, Freehold, NJ), and 13 ␮g/ml DNase I (Boehringer Mannheim, patches (data not shown). Mannheim, Germany). After incubation, a single cell suspension was pre- ␤ ␤ pared by removing large aggregates and debris by passage through a Differential expression of 1 and 7 integrin on lung T cells ␤ high ␤ Ϫ/low 100-␮m mesh. during M. tuberculosis infection: accumulation of 1 / 7 T cells in infected lungs Abs for flow cytometry Because VCAM-1 expression was up-regulated in the lung in- The following mAbs were used for flow cytometry: anti-CD44 FITC (clone fected with M. tuberculosis and both ␣ ␤ or ␣ ␤ integrins may ␣ 4 1 4 7 IM7; PharMingen, San Diego, CA), anti-CD49d ( 4 integrin) FITC (clone ␣ act as a counter-receptor for this addressin, we investigated the R1-2; PharMingen), anti-CD103 ( E integrin) FITC (clone 2E7; Phar- ␤ expression of ␤ and ␤ integrin on pulmonary T cells. In unin- Mingen), anti- 7 integrin-PE (clone M293; PharMingen), and biotin-con- 1 7 ␤ jugated anti-CD29 ( 1 integrin) (clone Ha2/5; PharMingen). Anti-CD4 Tri- fected lung, while the majority of lung T cells expressed low levels The Journal of Immunology 4855 Downloaded from http://www.jimmunol.org/

FIGURE 1. Sustained up-regulation of VCAM-1 expression in the lungs of M. tuberculosis-infected mice. Frozen sections from normal and infected lungs isolated at week 2, 4, 8, and 12 postinfection were stained by guest on September 24, 2021 with Ab against VCAM-1. The arrow indicates the positive staining of VCAM-1 (magniﬁcation ϫ200).

␤ ␤ of 1 integrin (Fig. 2A, upper panel), the expression of 7 integrin was heterogeneous (Fig. 2B, upper panel). M. tuberculosis infec- ␤ tion resulted in a shift in 1 integrin expression from predomi- ␤ low ␤ high nantly 1 to an accumulation of 1 T cells (Fig. 2A, lower ␤ panel). In contrast, changes in 7 integrin-expressing T cells displayed a reciprocal pattern, with a marked increase of T cells with ␤ undetectable or low levels of 7 integrin (Fig. 2B, lower panel). ␤ ␤ ϩ ϩ FIGURE 2. Differential expression of 1 and 7 integrin on T cells from Interestingly, a small population of CD4 and CD8 T cells ex- ␤ ␤ the lungs during M. tuberculosis infection. Histograms show 1 (A) and 7 ␤ ϩ ϩ pressing a higher level of 7 integrin also emerged during infection (B) integrin expression on lung CD4 and CD8 T cells from uninfected ␤ (Fig. 2B, lower panel). To relate the expression of 1 integrin to control mice (upper panels) or mice 8 wk postinfection (lower panels). The ␤ ␤ that of 7 integrin, the coexpression of these two -chains on dot plot (C) shows the costaining of the two ␤-chains. Proﬁles are repre- ϩ ϩ ␤ high CD4 and CD8 T cells was analyzed. Fig. 2C shows that 1 sentative of three experiments. The numbers are the mean percentages of ␤ ϩ ϩ T cells tended to express low or undetectable levels of 7 integrin, positive cells after gating on CD4 or CD8 T cells for three mice. and mycobacterial infection in lung resulted in an expansion of ␤ high/␤ Ϫ/low CD4ϩ and CD8ϩ T cells (Fig. 2C). 1 7 ␣ ␤ ␣ ␤ Development of 4 7 and E 7 T cell populations in lung ␣ ␤ high Expansion of 4 1 T cells in the lungs during M. during M. tuberculosis infection tuberculosis infection ␤ ␤ ␣ As for 1, 7 integrin can associate with 4 integrin (CD49d), but ␣ ␣ As the -chain of integrin heterodimers confers the speciﬁcity it may also pair with E integrin (CD103). Recent studies have ␣ ␤ high ␣ ␤ ␣ ␤ of binding, the expression of 4 1 (VLA-4) was analyzed. indicated that 4 7 and E 7 integrins are important in the homing M. tuberculosis infection led to an increase in numbers of both and retention of mucosal lymphocytes to mucosal surfaces (20, ϩ ϩ ␣ ␣ ␤ ␣ ␤ CD4 and CD8 T cells expressing 4 integrin. This increase was 21). We therefore compared the distribution of 4 7 and E 7 ␤ associated with the increased expression of 1 integrin (Fig. 3A). integrins on lung T cells in normal and infected lung by costaining ␤ ␣ ␣ Kinetic studies of the emergence of this population revealed that of 7 with either 4 or E integrin. Only a small percentage of T ␣ ␤ high ␣ ␤ the expansion of 4 1 T cells was apparent 4 wk after infec- cells expressed 4 7 integrin in normal lung (Fig. 4A, upper pan- Ͼ ϩ ␣ tion, and peaked at 8 wk with 100-fold increase for both CD4 el). M. tuberculosis infection, however, led to an increase in 4 ϩ ␣ ␤ high and CD8 4 1 T cells (Fig. 3B). integrin-expressing T cells (Fig. 4A, lower panel); the majority of 4856 CRUCIAL ROLE OF ADHESION MOLECULES IN PULMONARY TUBERCULOSIS Downloaded from http://www.jimmunol.org/

␣ ␤ ␣ ␤ FIGURE 4. Development of 4 7 and E 7 T cell populations during ␤ ␣ ␤ M. tuberculosis infection. The profiles show 7 vs 4 integrin (A) and 7 ␣ ϩ ϩ vs E integrin (B) expression on lung CD4 and CD8 T cells from normal (upper panels) and infected (lower panels) mice 8 wk postinfection. Pro- FIGURE 3. Accumulation of ␣ ␤ high T cells in the lungs during 4 1 files are representative of three experiments. The numbers are the mean ␤ ␣ M. tuberculosis infection. The profiles show 1 vs 4 integrin (A) expres- ϩ ϩ ϩ ϩ percentages of positive cells after gating on CD4 or CD8 T cells for sion on lung CD4 and CD8 T cells from normal (upper panel) and three mice.

infected (lower panel) mice 8 wk postinfection. Profiles are representative by guest on September 24, 2021 of three experiments. The numbers are the mean percentages of positive ϩ ϩ fore examined the expression of ␤ integrins on activated/memory cells after gating on CD4 or CD8 T cells for three mice. The kinetics of ␤ ␤ ␣ ␤ high T cells by costaining for CD44 with either 1 or 7 integrin. In expansion of 4 1 -expressing lung T cells over the course of infection is shown in B. The mean (ϮSEM) total number of lymphocytes for three agreement with our previous report (26), M. tuberculosis infection high mice per lung at weeks 2, 4, 8, and 12 were 0.9 ϫ 106 (Ϯ0.11), 9 ϫ 106 resulted in a significant enrichment of CD44 T cells (Fig. 5A), 6 6 ␤ (Ϯ1.50), 24.3 ϫ 10 (Ϯ2.47), and 11.3 ϫ 10 (Ϯ1.53), respectively. There and these T cells expressed high levels of 1 integrin (Fig. 5A). was no significant difference in number of total lymphocyte per lung be- However, a more complex pattern was displayed with regard to the ␤ ␤ Ϫ ␤ low tween normal and infected mice at 2 wk postinfection. expression of 7 integrin. Although subpopulations of 7 , 7 , ␤ high ␤ Ϫ/low and 7 activated/memory T cells were present, 7 T cells constituted the dominant activated/memory T cell population ␤ these T cells expressed low or undetectable levels of 7 integrin (Fig. 5B). ␣ (Fig. 4A, lower panel). Intriguingly, the expression patterns of E Ϫ ϩ ϩ ␥ ␤ high ␤ /low integrin on CD4 and CD8 T cells were different. In uninfected IFN- is predominantly produced by 1 / 7 T cells in ␣ ϩ the infected lungs lung, E was expressed on only a minority of CD4 T cells, ϩ ␣ ␥ whereas approximately one-half of CD8 T cells were E positive Because IFN- is an essential component of protective immunity (Fig. 4B, upper panel). After infection, there was a small, but against M. tuberculosis infection (28, 29), we examined the phe- ␣ ϩ ␥ ␤ definite, increase in the E CD4 T cells (from 2% to 10%) (Fig. notypes of IFN- -producing T cells by costaining for the inte- ␣ ϩ ␥ ␥ 4B, lower panel). The changes of E CD8 T cells were more grins and IFN- . As expected, the number of IFN- -producing T ␣ complex. There was a significant decrease in the proportion of E cells was greater in infected than that in normal lungs. The ma- CD8ϩ T cells (from 53% to 23%). However, as with the CD4ϩ T jority of the IFN-␥-producing CD4ϩ and CD8ϩ T cells expressed ϩ ␤ ␤ cells, a minor population of CD8 T cells expressing very high a high level of 1 integrin (Fig. 6A). With regard to 7 integrin ␣ ␤ ␥ levels of E 7 emerged in the infected lungs (from 1% to 6%). expression, the major IFN- -producing T cell population ex- ␣ ␤ Taken together, it appears that during infection, 4 integrin was pressed low or undetectable levels of 7 integrin, although IFN- ␤ Ϫ/low ␥ ␤ high ␣ primarily expressed on 7 T cells (Fig. 4A, lower panel), and -producing 7 , most likely in association with E integrin ␣ ␤ ␤ ϩ E 7 T cells tended to express high or very high levels of 7 (Fig. 4B), CD8 T cells were also evident (Fig. 6B). integrin (Fig. 4B, lower panel). Accelerated expansion of activated/memory T cell populations in ␤ ␤ Expression of 1 and 7 integrin on activated/memory T cells lung during secondary mycobacterial infection ␤ ␤ A preferential accumulation of some subsets of 1 and 7 integrin- Acquired immunity to mycobacterial infection is dependent on the expressing T cells in infected lung suggested that these T cells rapid expansion of memory T cells generated from primary infec- were of special relevance to the host immune response. We there- tion with CD4ϩ T cells as the predominant protective cells (24, The Journal of Immunology 4857 Downloaded from http://www.jimmunol.org/

␤ high ␤ Ϫ/low FIGURE 5. 1 / 7 activated/memory T cells emerge during infection. Lung cells isolated 8 wk after infection were analyzed for CD44 FIGURE 6. ␤ and ␤ IFN-␥-producing T cell subsets. After stimulating ␤ ␤ 1 7 and 1 (A)or 7 (B) integrin expression. Profiles are representative of three with plate-bound anti-CD3 mAb, lung cells from normal and infected mice experiments. The numbers are the mean percentages of positive cells after ␤ ␤ were surface stained for CD4, CD8, and 1 and 7 integrins, followed by gating on CD4ϩ or CD8ϩ T cells for three mice. ␥ ␥ ␤ intracellular staining for IFN- . The profiles show IFN- vs 1 integrin (A) ␥ ␤ and IFN- vs 7 integrin (B). Profiles are representative of three experiments. The numbers are the mean percentages of positive cells after gating ϩ ϩ on CD4 or CD8 T cells for three mice. by guest on September 24, 2021 ␤ ␤ 25). We therefore compared the expression of 1 and 7 integrin on CD4ϩ T cells during primary and secondary infection. After primary aerosol BCG infection was established, infected mice lungs. Histological examination of the infected lungs showed that were treated with antibiotics to reduce bacterial load before sec- granulomas in untreated infected (Fig. 8A) or control mAb-treated ondary aerosol challenge. The percentage of ␤ high CD4ϩ T cells 1 mice (Fig. 8B) were well defined and were comprised largely of in the antibiotic-treated mice (Fig. 7, group 2) was similar to that lymphocytes and macrophages (Fig. 8, E and F). In contrast, treat- in mice 2 wk after single aerosol infection (Fig. 7, group 3) and ment with integrin-specific mAb led to the formation of enlarged only slightly higher than in control uninfected mice (Fig. 7, group and disorganized lesions (Fig. 8, C and D), containing predomi- 1). Reexposure of the primary infected, antibiotic-treated mice to nantly neutrophils and necrotic debris (Fig. 8, G and H). Interest- secondary infection (Fig. 7, group 4) resulted in a rapid and sig- ingly, the treatment of mice with mAb to ␣ and ␣ ␤ integrin did nificant expansion of ␤ high CD4ϩ T cells (2.7-fold increase com- 4 4 7 1 not significantly alter the bacterial load in the lungs (Table I), pared with control group) by 2 wk postinfection (Fig. 7A). This despite the clinical deterioration of the mice. ␤ high ϩ increased percentage of 1 CD4 T cells corresponded with a ␤ Ϫ ϩ similar increase in the percentage of 7 CD4 T cells (2.5-fold ␤ high ϩ Discussion increase) (Fig. 7B). A significant increase in 7 CD4 T cells was also observed in the lungs after reexposure (3.2-fold increase) The control of pulmonary tuberculosis requires a sustained chronic (Fig. 7C). inflammatory response in the lung. Integrins have a role in lymphocyte-endothelial cell interaction and are important for the homing of activated/memory lymphocytes to inflamed lesions (4, 30– ␣ Treatment of M. tuberculosis-infected mice with mAb to 4 or 32). The relative contribution of the different integrins to ␣ ␤ 4 7 integrin leads to a detrimental pulmonary inflammatory recruitment of T cells to the lung, particularly during chronic my- response cobacterial infection, has not previously been investigated in de- ␣ ␣ ␤ To investigate further the contribution of 4 and 4 7 integrins to tail. To date, studies on mycobacterial infection have focused on the cell-mediated immune response to M. tuberculosis in the lungs, the LFA-1:ICAM-1 interaction. LFA-1high T cells are recruited to ␣ ␣ ␤ aerosol-infected mice were treated with mAb to 4 or 4 7 inte- the spleen after i.v. infection with M. tuberculosis (33, 34), and to grin from day 14 postinfection. Experiments were terminated on the lung following aerosol infection (26). Similarly, the expression ␣ ␣ ␤ day 28, as both 4 and 4 7 integrin mAb-treated mice were of ICAM-1 and VCAM-1 has been observed to increase in the skin ␣ hunched and wasting. Table I shows that treatment with mAb to 4 following BCG inoculation (35). ICAM-1-deficient mice, how- ␣ ␤ or 4 7 integrin resulted in a significant reduction in the numbers ever, did not have increased susceptibility to mycobacterial infec- of lymphocytes, and an increase in granulocyte numbers in the tion (36, 37), suggesting that other adhesion molecules may also be 4858 CRUCIAL ROLE OF ADHESION MOLECULES IN PULMONARY TUBERCULOSIS

to the lung following aerosol infection. In the lungs of uninfected mice, VCAM-1 was not detectable on pulmonary endothelium ␣ ␤ (Fig. 1), and 4 1 integrin was expressed at low levels on resident T cells (Fig. 3). After aerosol M. tuberculosis infection, the expression of VCAM-1 on endothelial cells occurred from week 4 and still persisted at week 12 in the infected lungs (Fig. 1). This up-regulated expression of VCAM-1 correlated with the accumu- ␣ ␤ lation of activated/memory T cells expressed high levels of 4 1 ␣ ␤ high ϩ integrin (Fig. 3). The peak of the influx of 4 1 CD4 and CD8ϩ T cells coincided with the maximum cellular infiltrate in the lungs, and these cells were the major IFN-␥-producing cells (Fig. ␤ high ␤ Ϫ ϩ 6A). Furthermore, the accelerated expansion of 1 / 7 CD4 T cells in the lung following secondary infection (Fig. 7, A and B) was associated with the rapid control of infection, and a reduced bacterial load than during primary infection (U. Palendira, manu- script in preparation). Taken together, these findings suggest that ␣ ␤ the 4 1:VCAM-1 interaction leads to the recruitment of activated T cells into the inflamed lungs and contributes to the control of mycobacterial infection in the lung. Analysis of addressins and Downloaded from integrin expression during Chlamydia infection also demonstrated ␤ high up-regulation of VCAM-1 expression and accumulation of 1 / ␤ Ϫ/low 7 -activated T cells in the genital mucosa (32), Thus, these ␣ ␤ findings underline the importance of the 4 1:VCAM-1 interaction in the homing of activated T cells to extraintestinal mucosa. Interestingly, other studies of genital Chlamydia infection have http://www.jimmunol.org/ ␣ ␤ revealed up-regulation of MadCAM-1 and recruitment of 4 7 integrin-expressing CD4ϩ T cells into genital tract (38), demon- strating that a complex pattern of addressin and integrin interaction may occur during inflammatory response to infection in extraintestinal mucosal tissues. ␣ ␤ ␣ ␤ Because VCAM-1 can bind to both 4 1 and 4 7 het- FIGURE 7. Accelerated changes of ␤ and ␤ integrin expression dur- 1 7 erodimers, we investigated the role of these two integrins in the ing secondary infection. Mice were infected with aerosol BCG; rested for cellular response to M. tuberculosis infection. Treatment of in- 8 wk, then treated with isoniazid for 4 wk; and rested for 2 wk and then ␣ by guest on September 24, 2021 reexposed to aerosol BCG. Lung cells were isolated from four groups of fected mice in vivo with mAb to 4 integrin, which binds to both ␣ ␤ ␣ ␤ mice (1, uninfected normal mice; 2, treated mice 14 wk after primary 4 1 and 4 7 integrins, led to a significant reduction in the num- infection; 3, mice 2 wk after single infection; and 4, treated mice 2 wk after ber of lymphocytes and to an influx of neutrophils into infected ␤ high ␤ Ϫ ␤ high ϩ ␣ secondary infection). The percentages of 1 , 7 , and 7 CD4 T lungs (Table I). It has been shown that anti- 4 mAb treatment does cells were compared. Results are representative of two experiments. Mean not block neutrophil recruitment (39). This dysregulation of the ␤ Ͻ fluorescence intensity (MFI) of 7 20 was defined as negative expression inflammatory response resulted in development of granulocyte- ␤ ␤ ␤ Ͼ ␤ high ␤ high for 7, whereas MFI of 1 and 7 200 was defined as 1 and 7 , predominant, disorganized infiltrates that progressed to necrosis Ϯ respectively. The mean ( SEM) percentage of the cell populations for (Fig. 8). Similar necrotic lesions have been reported in mice lack- three mice per group is shown. The significance of the differences between ing the critical protective cytokines, IFN-␥ and TNF (27, 28). This group 4 and groups 1, 2, and 3 was compared by unpaired Student’s t test p Ͻ 0.05). suggests that the influx of neutrophils is an attempt to compensate ,ء) for ineffective lymphocyte responses. Despite the altered inflammatory responses and the clinical deterioration in mice, there was required for cellular recruitment and so compensate for the defi- no significant difference in the bacterial load in the lungs between ciency of ICAM-1 during this infection. treated and untreated mice. This may be due to the relatively short The current study demonstrates that the interaction of VCAM-1 time course of the mAb treatment compared with slow rate of ␣ ␤ and 4 1 integrin has a role in the recruitment of activated T cells mycobacterial growth, or to other factors. Treatment with mAb to

␣ ␣ ␤ Table I. Treatment of M. tuberculosis-infected mice with mAb to 4 or 4 7 integrin leads to a detrimental lung response

mAb Treatmenta Lymphocyteb Granulocyteb Bacterial loadc n

Untreated 9.92 Ϯ 1.14 3.88 Ϯ 0.71 6.74 Ϯ 0.15 5 Control mAb 10.4 Ϯ 0.28 4.88 Ϯ 0.83 6.60 Ϯ 0.18 5 Ϯ ء Ϯ ءϮ d ␤ ␣ DATK32 (anti- 4 7 integrin) 6.48 0.94 11.0 2.60 7.02 0.15 5 Ϯ ء Ϯ ء Ϯ ␣ PS/2 (anti- 4 integrin) 6.10 0.57 10.5 2.07 7.22 0.25 4 a The mice were infected for 14 days with aerosol M. tuberculosis and then injected (i.p.) with rat mAb on day 14, then on alternate days until day 28 postaerosol infection. b Lung tissues were collected 28 days after aerosol infection of M. tuberculosis. Lymphocyte and granulocyte populations were deﬁned by FACS analysis. Results are expressed as mean cell number ϫ 106 Ϯ SEM/lung. c Ϯ Lung tissues were collected at 28 days after aerosol infection with M. tuberculosis. Results are expressed as mean log10 CFU SEM/lung. d The signiﬁcance of the differences in cell number and CFU between untreated and mAb-treated mice was compared using the Mann-Whitney test .(p Ͻ 0.05 ,ء) The Journal of Immunology 4859

FIGURE 8. Treatment of M. tuberculosis- ␣ ␣ ␤ infected mice with mAb to 4 or 4 7 integrin leads to a detrimental pulmonary inﬂam- matory response. The mice were infected for 14 days with aerosol M. tuberculosis and then

left untreated (A and E) or injected (i.p.) with Downloaded from ␣ ␤ control mAb GL113 (B and F), or anti- 4 7 ␣ mAb DATK32 (C and G), or anti- 4 mAb PS/2 (D and H) on day 14, then on alternate days until day 28 postaerosol infection. Treat- ␣ ␣ ␤ ment with mAb to 4 or 4 7 integrin led to large, disorganized cellular inﬁltrate with a predominance of neutrophils and increased http://www.jimmunol.org/ necrosis. Hematoxylin and eosin, magniﬁca- tion: A–D, ϫ50; E–H, ϫ200. by guest on September 24, 2021

␣ ␤ ϩ ϩ high ␤ high 4 7 integrin also reduced the number of lymphocytes in the lungs CD4 and CD8 T cells were CD44 (Fig. 5B), and the 7 ␣ ␤ ϩ ␥ by 40%. VCAM-1 may act as an alternative ligand for 4 7 inte- CD8 T cells were capable of producing IFN- (Fig. 6B), sug- grin (14, 15) with low levels of VCAM-1 preferentially recruiting gesting that these were activated/memory T cells. The expression ␣ ␤ high␤ Ϫ ␣ ␤ ␣ ␤ 4 1 7 cells, and high expression of VCAM-1 favoring 4 7 of high levels of E 7 integrin may favor the retention of these ␣ ␤ integrin binding (40). Therefore, the observed recruitment of 4 7- cells on epithelial surface via interaction with E-cadherin. Taken ␣ ␤ high expressing lymphocytes to the inflamed lungs could have occurred together, the activation and expansion of these E 7 T cells in ␣ ␤ high in the absence of detectable expression of MadCAM-1, presum- the infected lungs indicate that along with 4 1 T cells, this ably through the interaction with VCAM-1. small population may also contribute to the host pulmonary re- ␣ ␤ ␥ ␣ ␤ The expression of E 7 integrin on pulmonary lymphocytes is sponses through the production of IFN- . The precise role of E 7 less well understood. M. tuberculosis infection appeared to have integrin-expressing T cells in protective immunity against myco- ␣ ␤ ϩ differential effects on the expression of E 7 integrin on CD4 and bacterial infection in the lung needs to be further investigated. CD8ϩ T cells. In contrast to the CD4ϩ T cells, the proportion of In conclusion, up-regulated expression of VCAM-1 and the in- ␣ ϩ␤ high ϩ ␣ ␤ high E 7 CD8 T cells decreased significantly in infected lung creased numbers of 4 1 T cells during pulmonary tuberculosis ␣ ␤ ϩ ␣ ␤ owing to the expansion of 4 1-expressing CD8 T cells. This suggest that the 4 1:VCAM-1 interaction contributes to the re- ␣ ␤ observation supports the hypothesis that expression of E 7 on cruitment of activated T cells to the infected lung. There are also ϩ ␤ high CD8 T cells in lung is not related to inflammatory stimuli, but is significant increases in 7 integrin-expressing T cells that par- due to constitutive factors in the pulmonary microenvironment (31, ticipate in the prompt memory response to the reexposure to my- 41). Intriguingly, we also observed the emergence of a small pop- cobacterial infection. The complex pattern of integrin expression ϩ ␤ ulation of CD8 T cells with very high levels of 7 integrin as- and the impaired inflammatory response on treatment with mAb to ␣ ␤ high ␣ ␣ ␤ sociated with E integrin in infected lungs (Fig. 4B). These 7 4 and 4 7 integrins indicate that multiple adhesion molecules 4860 CRUCIAL ROLE OF ADHESION MOLECULES IN PULMONARY TUBERCULOSIS appear to contribute to the pulmonary inflammatory response to 22. Austrup, F., S. Rebstock, P. J. Kilshaw, and A. Hamann. 1995. Transforming ␤ ␣ aerosol M. tuberculosis. growth factor- 1-induced expression of the mucosa-related integrin E on lymphocytes is not associated with mucosa-specific homing. Eur. J. Immunol. 25: 1487. Acknowledgments 23. Snider, D. E., M. Raviglione, and A. Kochi. 1994. Global burden of tuberculosis. In Tuberculosis: Pathogenesis, Protection, and Control. B. R. Bloom, ed. ASM We thank Dr. Bernadette Saunders for helpful discussion and critical eval- Press, Washington, DC, p. 3. uation of tissue histology, and Joanne Spratt and Jason Compton for their 24. Andersen, P. 1997. Host responses and antigens involved in protective immunity to Mycobacterium tuberculosis. Scand. J. Immunol. 45:115. technical assistance. 25. Cooper, A. M., and J. L. Flynn. 1995. The protective immune response to My- cobacterium tuberculosis. Curr. Opin. Immunol. 7:512. References 26. Feng, C. G., A. G. D. Bean, H. Hooi, H. Briscoe, and W. J. Britton. 1999. Increase in IFN-␥ secreting CD8ϩ, as well as CD4ϩ, T cells in lungs following 1. Mackay, C. R. 1993. Homing of naive, memory and effector lymphocytes. Curr. aerosol infection with Mycobacterium tuberculosis. Infect. Immun. 67:3242. Opin. Immunol. 5:423. 27. Bean, A. G., D. R. Roach, H. Briscoe, M. P. France, H. Korner, J. D. Sedgwick, 2. Imhof, B. A., and D. Dunon. 1995. Leukocyte migration and adhesion. Adv. and W. J. Britton. 1999. Structural deficiencies in granuloma formation in TNF Immunol. 58:345. gene-targeted mice underlie the heightened susceptibility to aerosol Mycobacte- 3. Picker, L. J. 1994. Control of lymphocyte homing. Curr. Opin. Immunol. 6:394. rium tuberculosis infection, which is not compensated for by lymphotoxin. J. Im- 4. Salmi, M., and S. Jalkanen. 1997. How do lymphocytes know where to go: munol. 162:3504. current concepts and enigmas of lymphocyte homing. Adv. Immunol. 64:139. 28. Cooper, A. M., D. K. Dalton, T. A. Stewart, J. P. Griffin, D. G. Russell, and 5. Springer, T. A. 1995. Traffic signals on endothelium for lymphocyte recirculation I. M. Orme. 1993. Disseminated tuberculosis in IFN-␥ gene-disrupted mice. and leukocyte emigration. Annu. Rev. Physiol. 57:827. J. Exp. Med. 178:2243. 6. Osborn, L. 1990. Leukocyte adhesion to endothelium in inflammation. Cell 62:3. 29. Flynn, J. L., J. Chan, K. J. Triebold, D. K. Dalton, T. A. Stewart, and 7. Bevilacqua, M. P. 1993. Endothelial-leukocyte adhesion molecules. Annu. Rev. B. R. Bloom. 1993. An essential role for IFN-␥ in resistance to Mycobacterium Immunol. 11:767.

tuberculosis infection. J. Exp. Med. 178:2249. Downloaded from 8. Hynes, R. O. 1992. Integrins: versatility, modulation, and signaling in cell ad- 30. Engelhardt, B., F. K. Conley, P. J. Kilshaw, and E. C. Butcher. 1995. Lympho- hesion. Cell 69:11. cytes infiltrating the CNS during inflammation display a distinctive phenotype 9. Elices, M. J., L. Osborn, Y. Takada, C. Crouse, S. Luhowskyj, M. E. Hemler, and and bind to VCAM-1 but not to MAdCAM-1. Int. Immunol. 7:481. R. R. Lobb. 1990. VCAM-1 on activated endothelium interacts with the leuko- 31. Picker, L. J., R. J. Martin, A. Trumble, L. S. Newman, P. A. Collins, cyte integrin VLA-4 at a site distinct from the VLA-4/fibronectin binding site. P. R. Bergstresser, and D. Y. Leung. 1994. Differential expression of lymphocyte Cell 60:577. homing receptors by human memory/effector T cells in pulmonary versus cuta- 10. Butcher, E. C., and L. J. Picker. 1996. Lymphocyte homing and homeostasis. neous immune effector sites. Eur. J. Immunol. 24:1269. Science 272:60. 32. Perry, L. L., K. Feilzer, J. L. Portis, and H. D. Caldwell. 1998. Distinct homing 11. Nakajima, H., H. Sano, T. Nishimura, S. Yoshida, and I. Iwamoto. 1994. Role of pathways direct T lymphocytes to the genital and intestinal mucosae in Chlamy- http://www.jimmunol.org/ vascular cell adhesion molecule 1/very late activation antigen 4 and intercellular dia-infected mice. J. Immunol. 160:2905. adhesion molecule 1/lymphocyte function-associated antigen 1 interactions in 33. Andersen, P., A. B. Andersen, A. L. Sorensen, and S. Nagai. 1995. Recall of antigen-induced eosinophil and T cell recruitment into the tissue. J. Exp. Med. long-lived immunity to Mycobacterium tuberculosis infection in mice. J. Immu- 179:1145. nol. 154:3359. 12. Pretolani, M., C. Ruffie, J. R. Lapa e Silva, D. Joseph, R. R. Lobb, and 34. Orme, I. M., J. P. Griffin, A. D. Roberts, and D. N. Ernst. 1993. Evidence for a B. B. Vargaftig. 1994. Antibody to very late activation antigen 4 prevents anti- defective accumulation of protective T cells in old mice infected with Mycobac- gen-induced bronchial hyperreactivity and cellular infiltration in the guinea pig terium tuberculosis. Cell. Immunol. 147:222. airways. J. Exp. Med. 180:795. 13. Chisholm, P. L., C. A. Williams, and R. R. Lobb. 1993. Monoclonal antibodies 35. Abe, Y., K. Sugisaki, and A. M. Dannenberg Jr. 1996. Rabbit vascular endothe- ␣ lial adhesion molecules: ELAM-1 is most elevated in acute inflammation, to the integrin 4 subunit inhibit the murine contact hypersensitivity response. Eur. J. Immunol. 23:682. whereas VCAM-1 and ICAM-1 predominate in chronic inflammation. J. Leuko-

14. Chan, B. M., M. J. Elices, E. Murphy, and M. E. Hemler. 1992. Adhesion to cyte Biol. 60:692. by guest on September 24, 2021 ␣ ␤ 36. Johnson, C. M., A. M. Cooper, A. A. Frank, and I. M. Orme. 1998. Adequate vascular cell adhesion molecule 1 and fibronectin: comparison of 4 1 (VLA-4) ␣ ␤ expression of protective immunity in the absence of granuloma formation in and 4 7 on the human B cell line JY. J. Biol. Chem. 267:8366. 15. Ruegg, C., A. A. Postigo, E. E. Sikorski, E. C. Butcher, R. Pytela, and D. J. Erle. Mycobacterium tuberculosis-infected mice with a disruption in the intracellular ␣ ␤ ␣ ␤ adhesion molecule 1 gene. Infect. Immun. 66:1666. 1992. Role of integrin 4 7/ 4 P in lymphocyte adherence to fibronectin and VCAM-1 and in homotypic cell clustering. J. Cell Biol. 117:179. 37. Steinhoff, U., U. Klemm, M. Greiner, K. Bordasch, and S. H. Kaufmann. 1998. 16. Berlin, C., E. L. Berg, M. J. Briskin, D. P. Andrew, P. J. Kilshaw, B. Holzmann, Altered intestinal immune system but normal antibacterial resistance in the ab- ␣ ␤ sence of P-selectin and ICAM-1. J. Immunol. 160:1666. I. L. Weissman, A. Hamann, and E. C. Butcher. 1993. 4 7 integrin mediates lymphocyte binding to the mucosal vascular addressin MAdCAM-1. Cell 74:185. 38. Kelly, K. A., and R. G. Rank. 1997. Identification of homing receptors that 17. Streeter, P. R., E. L. Berg, B. T. Rouse, R. F. Bargatze, and E. C. Butcher. 1988. mediate the recruitment of CD4 T cells to the genital tract following intravaginal A tissue-specific endothelial cell molecule involved in lymphocyte homing. Na- infection with Chlamydia trachomatis. Infect. Immun. 65:5198. ␣ ture 331:41. 39. Lobb, R. R., and M. E. Hemler. 1994. The pathophysiologic role of 4 integrins 18. Cepek, K. L., S. K. Shaw, C. M. Parker, G. J. Russell, J. S. Morrow, D. L. Rimm, in vivo. J. Clin. Invest. 94:1722. and M. B. Brenner. 1994. Adhesion between epithelial cells and T lymphocytes 40. Rott, L. S., M. J. Briskin, D. P. Andrew, E. L. Berg, and E. C. Butcher. 1996. A ␣ ␤ mediated by E-cadherin and the E 7 integrin. Nature 372:190. fundamental subdivision of circulating lymphocytes defined by adhesion to mu- 19. Karecla, P. I., S. J. Bowden, S. J. Green, and P. J. Kilshaw. 1995. Recognition of cosal addressin cell adhesion molecule-1: comparison with vascular cell adhesion ␣ ␤ ␣ ␤ ␤ E-cadherin on epithelial cells by the mucosal T cell integrin M290 7 ( E 7). molecule-1 and correlation with 7 integrins and memory differentiation. J. Im- Eur. J. Immunol. 25:852. munol. 156:3727. ␤ 20. Shaw, S. K., and M. B. Brenner. 1995. The 7 integrins in mucosal homing and 41. Rihs, S., C. Walker, J. C. Virchow Jr., C. Boer, C. Kroegel, S. N. Giri, and ␣ ␤ retention. Semin. Immunol. 7:335. R. K. Braun. 1996. Differential expression of E 7 integrins on bronchoalveolar ␣ ␤ 21. Kilshaw, P. J., and P. Karecla. 1997. Structure and function of the mucosal T-cell lavage T lymphocyte subsets: regulation by 4 1-integrin cross-linking and ␣ ␤ ␤ integrin E 7. Biochem. Soc. Trans. 25:433. TGF- . Am. J. Respir. Cell Mol. Biol. 15:600.