A novel animal model of emphysema induced by anti-elastin autoimmunity

Abstract

Loss of immune tolerance to self-antigens can promote chronic inflammation and disrupt the normal function of multiple organs including the lungs. Degradation of elastin, a highly insoluble protein and a significant component of the lung structural matrix generates pro-inflammatory molecules. Elastin fragments (EFs), have been detected in the serum of smokers with emphysema, and elastin-specific T cells have also been detected in the peripheral blood of smokers with emphysema. However, an animal model that could recapitulate T cell specific autoimmune responses by initiating and sustaining inflammation in the lungs is lacking. Here we report the first animal model of autoimmune emphysema mediated by the loss of tolerance to elastin. Mice immunized with a combination of human EFs plus rat EFs, but not mouse EFs showed increased infiltration of innate and adaptive immune cells to the lungs and developed emphysema. We cloned and expanded mouse elastin-specific CD4⁺ T cells from the lung and spleen of immunized mice. Finally, we identified TCR sequences from the autoreactive T cell clones, suggesting possible pathogenic TCR that can cause loss of immune tolerance against elastin. This new autoimmune-model of emphysema provides a useful tool to examine the immunological factors that promote loss of immune tolerance to self.

Introduction

Elastin is a matrix protein, which comprises over 90% of assembled elastic fibers in the extracellular space, and provides the required tissue strength and elasticity necessary for multiple organs (1). Specifically, proper function of the lungs, vascular structures, and skin depend on their flexibility, as such they contain a much higher amount of elastin per dry weight than other organs (2). Under steady state, biogenesis of matrix molecules includes regular reorganization, however extracellular elastin matrix assembly is known as elastogenesis, primarily occurs during organ development and remain highly stable throughout life (3). As such, elastin degradation due to abnormal exposure to elastolytic enzymes expressed by innate immune cells, can result in organ dysfunction and life threatening diseases, of the lung (4–8), and vasculature (9–12).

Cigarette smoking causes a distinct pattern of lung parenchyma destruction characterized by loss of tissue elasticity and generation of elastin fragments (EFs) found in the serum (13, 14). We and others have shown that chronic exposure to cigarette smoke recruits innate and adaptive immune cells into the lung (5, 15–18). Activated innate immune cells (e.g., macrophage and neutrophils) release several elastin-degrading enzymes including neutrophil elastase, matrix metalloproteinase (MMP)9, MMP12, which can either directly cleave elastin, or block alpha one anti-trypsin, the absence of which is associated with severe emphysema (8, 19, 20). In addition to innate immune cells, activated adaptive immune cells (T and B lymphocytes) are recruited to the lungs of smokers, and adoptive transfer of CD8⁺ T cells have been shown to induce lung inflammation and emphysema (21–24).

We and others have shown that smokers who develop emphysema, harbor activated T helper 1 (Th1) and Th17 cells expressing interferon (IFN)-γ and interleukin (IL)-17A respectively in the lungs when compared to control subjects (25–27). Consistently CD4⁺ T cells isolated from the peripheral blood of smokers with emphysema show increased interferon IFN-γ and interleukin IL-17A expression in response to EFs which can be inhibited in the presence of MHC class II blocking antibodies (28, 29). The significance of adaptive immunity against elastin was shown in a longitudinal study whereby the magnitude of autoreactive immune responses to EFs, correlated with the severity of physiological decline over three years (30). Moreover, we have shown that auto-reactive T cell responses significantly correlate with emphysema severity, and lung function decline (28, 29). Collectively, human studies suggest that elastin-specific auto-reactive T cells persist in smokers with emphysema despite smoke cessation, which may contribute to progressive inflammation and result in the destruction of several elastin-rich organs.

Despite recent advances and a better understanding of the pathophysiological effects of chronic cigarette smoke-induced lung inflammation, little is known about the loss of immune tolerance to elastin. In this paper, we provide the methods that we utilized to develop a novel mouse model of emphysema that reproduces autoimmune inflammation against elastin that is found in smokers. Repeated immunization using non-self EFs (human and rat), but not mouse elastin, successfully broke tolerance against elastin in mice; the model recapitulated cigarette smoke-induced emphysema characterized by airspace enlargement and inflammatory cells infiltration in elastin rich organs. The precise contribution of EF reactive T cells to tissue damage is not fully known; however, w we cloned auto-reactive T cells and identified several potential pathogenic T cell receptors (TCRs) against mouse elastin. Here we describe the in vivo method for induction of EF specific T cell responses and isolation of both mouse and human TCR sequences for in vitro analysis of EF specific TCR function. This novel experimental mouse model provides a valuable tool that could be used to determine the underlying mechanisms involved in the loss of immune tolerance to EFs in elastin-rich organs. Furthermore, we anticipate that the techniques used to develop this model could be easily adapted to establish novel mouse models for other autoimmune diseases.

Materials and Methods

Mice

C57BL/6 mice (7 to 8 weeks old, females) were purchased from The Jackson Laboratory (Bar Harbor, ME); All experimental protocols were approved by the Institutional Animal Care and Use Committee of Baylor College of Medicine and followed the National Research Council Guide for the Care and Use of Laboratory Animals.

The human T cell clone

Peripheral blood mononuclear cells were isolated from smokers with emphysema and T cells reactive to hEFs were isolated and cloned as previously described (29). Studies were approved by the Institutional Review Board at Baylor College of Medicine and informed consents were obtained from all patients.

TCR identification of EFs-specific human CD4⁺ T clone

TC-378–4, a human elastin-specific CD4⁺ T cell clone isolated from peripheral blood mononuclear cells (PBMCs) from a smoker with emphysema (29), was used to generate human EFs-specific TCR expressing transfectant thymoma cell line.

Total RNA from TC-378–4 clone was extracted with TRIzol (ThermoFisher #15596026) following the manufacturer’s instructions. cDNA was synthesized with TCR-specific RT primers to increase specificity using a high-capacity cDNA reverse transcription kit (Applied Biosystems cat#4368814); 10× buffer, 25× dNTP mix, 50U RT enzyme, 40U RNase inhibitor and 325 nM final concentration of each TCR specific RT-PCR primer in a total 20μl volume. Primer sequences were as follows: 5’-AGCTGGACCACAGCCG-3’ (TRAC cDNA), 5’-GAAATCCTTTCTCTTGACCATG-3’ (TRBC1 cDNA), and 5’-GCCTCTGGAATCCTTTCTCT-3’ (TRBC2 cDNA). The reaction mix was incubated at 25°C for 10 minutes, 45°C for 120 minutes, and 85°C for 5 minutes.

Following RT, a multiplex-nested PCR was carried out in two rounds of PCR for TCRα- and TCRβ-chain amplification separately with reverse transcribed cDNA (50–200ng). In the multiplex PCR (1^st round PCR), the variable region of TCRα- or TCRβ-chain was amplified using a pool of multiplex variable region primers (Vα-pool or Vβ-pool) along with an external constant-region primer (TRAC_ext or TRBC_ext). Details related to the design of TCR primer sequences have been published (31). Briefly, Vα-pool and Vβ-pool were generated with 44 and 40 primers to detect each variable region of TCRα, and TCRβ respectively. A total 25μl PCR reaction was prepared with 5× Go Taq buffer, 80μM dNTPS, 3%DMSO, 2uM primer pool (final concentration of each primer was 46nM), 200nM constant primer, 1U Go Taq DNA polymerase (VWR #PAM8298), and 2.5ul RT reaction mixture. The PCR cycles were 2 minutes at 94°C followed by 34 cycles of 20 seconds at 94°C, 30 seconds at 54°C, and 1 minute at 72°C, followed by 7 minutes at 72°C.

For the nested PCR (2^nd round PCR), 2.5μl of the 1^st round PCR reaction was used as a template. Similarly, two separate PCR reactions were prepared for TCRα and TCRβ amplification in a final volume of 25μl. Reaction mix contained 5× Go Taq buffer containing loading dye, 80μM dNTPS, 3%DMSO, 200nM primer adaptor primer (TRAV_ada or TRBV_ada), 200nM internal constant primer (TRAC_int or TRBC_int), and 1U Go Taq DNA polymerase (VWR #PAM8298). The PCR cycles were 2 minutes at 94°C followed by 34 cycles of 20 seconds at 94°C, 30 seconds at 56°C, and 1 minute at 72°C, followed by 7 minutes at 72°C. Then, 5μl of the final reaction was run out 1% agarose gel to confirm successful PCR, and the 20μl remainder was purified for Sanger sequencing using a PCR purification kit (Zymo Research #11–305C).

For Sanger sequencing, the TRAV_ada and TRBV_ada primer were used for TCRα and TCRβ, respectively. TCR sequences were analyzed with the IMGT/V-Quest tool (http://www.imgt.org/).

Generation of human EFs-specific TCR retroviral vectors

The primers used in the multiplex-nested PCR reactions contain restriction sites for subsequent sub-cloning into the template retroviral vector pMSCVII-Ametrine (pMIA) to generate TCR expressing transfectant thymoma cell lines. The vector was modified to incorporate restriction enzyme cut sites for TCR cloning (32, 33).

Sub-cloning of the purified TCRα or TCRβ PCR products into pCR-Blunt II-TOPO vector (ThermoFisher #K280020) were performed using the nested reaction products. To generate blunt-ends, the PCR products were treated with DNA polymerase I Large-Klenow (Promega #M220C) for 30 minutes at 37°C followed by heat inactivation for 5 minutes at 70°C. Klenow-treated reaction products were ligated with pCR-Blunt II-TOPO vector at room temperature for 15 minutes as recommended by the manufacture. TOP10 competent cells (ThermoFisher #K280020) were used for transformations. The insertion of each TCRα and TCRβ chain was confirmed with specific restriction sites digestion: BstBI (NEB #R0519) and MfeI (NEB #R3589) for TCRα, and SnaBI (NEB #R0130) and SacII (NEB #R0157) for TCRβ. Following confirmation of the correct insertion sites, Sanger Sequencing was performed.

TCRβ containing pCR-Blunt II-TOPO vector was digested with BstBI and MfeI and DNAs were purified with gel-DNA purification kit (Zymo Research #11–300). In parallel, the template retroviral vector pMSCVII-Ametrine (pMIA) was digested with the same restriction enzymes and gel-purified using a gel DNA purification kit. Before ligation, the digested template vector was treated with 10U CIP enzyme (NEB #M0290) to prevent self-ligation for 1 hour at 37°C followed by heat inactivation for 10 minutes at 75°C and then was purified with a PCR purification kit (Zymo Research #11–305C). The ligation reaction was performed with the insert at 6 molar excess to the vector and about 150ng total DNA using DNA ligase for 30 minutes at room temperature. For transformations, DH5α competent cells (Invitrogen #18265017) were used. The successful insertion of the TCRβ chain was confirmed by test-digestion with the restriction enzymes and Sanger sequencing. For TCRα insertion into TCRβ containing retroviral vector, a similar process was carried out with different restriction enzymes, SnaBI and SacII.

Generation of human elastin-specific TCR expressing transfectant thymoma cell lines

A retroviral vector was generated using 4μg hTCR-pMIA (vector generated described above), 4μg pEQ-Pam3(-E) (packaging vector), and 2μg pVSVg (envelope vector) as provided in Transit 293T transfection reagent (Mirus #MIR2700) for HEK293T cells. After 24hrs, the media was replaced with 5% FCS complete DMEM, and 24hrs later, supernatant from hTCR transfected HEK293T cells was harvested and filtered through 0.45μm filters. Mouse 4G4 thymoma cells expressing human CD4, (1 × 10⁶) were suspended in 4ml supernatant from hTCR-transfected HEK293T cells mixed with polybrene at 6μg/mL, placed in a 6 well plate, and spun at 1800rpm for 90 minutes at 20°C. The second round of transduction was repeated the next day as described above and after additional 24hrs in culture, cells were placed in complete (C)-RPMI supplemented with 10% heat-inactivated fetal bovine serum (FBS). C-RPMI was prepared by adding 2-mercaptoethanol (5 × 10⁻⁵ M), L-glutamine (100×), sodium pyruvate (100×), MEM non-essential amino acids (100×), MEM vitamin (100×), penicillin/streptomycin (100×), and gentamicin (10 μg/ml). The hTCR transduced 4G4 cells were sorted by FACS for mCD3-BV510⁺, hTCR-Amet⁺, and hCD4-GFP^high cells representing stable and functional hTCR:mouse (m)CD3 complex on the cell surface (m:hTCR). The m:hTCR sorted cells were expanded for 3–5 days in a 6 well plate, and cell purity (>95%) was confirmed using flow cytometry.

Human leukocyte antigen (HLA) restriction identification and in vitro T cell assays

To identify HLA restriction, genomic DNA was isolated from 5×10⁵ TC-378–4 clone using a DNA isolation kit (Qiagen #69504) and used for high-resolution next generation sequencing (NGS), to identify HLA typing (Table 1). Based on the HLA typing, we used B-Lymphoblastoid cell line (IHW ID:09420) (purchased from International Histocompatibility Working Group; IHWG, Seattle, WA) as antigen presenting cells (APCs). m:hTCR transduced 4G4 cells (1 × 10^{4 )} were cultured with APCs (1 × 10⁵) in the presence of 30μg/ml of hEFs (EPC #QP45) for 4 days. For HLA blocking experiments, APCs were incubated with α-DQ, α-DR, or isotype control antibodies (2.5 μg/ml) for 1hour at 37°C before co-culture with m:hTCR cells. Mouse IL-2 production was measured in the supernatant using ELISA according to the manufacturer’s instructions (BD #555148).

Table 1.

HLA typing result

DRB1		DRB3	DRB4	DQB1		DQA1
07:01	13:03	01:01	01:03	02:02	03:01	02:01	05:05

Cigarette smoke-exposed emphysema mouse model

C57/BL6 mice (7 to 8 weeks old, females) purchased from The Jackson Laboratory (Bar Harbor, ME), were exposed to active smoke from commercial cigarettes (Marlboro 100 long, Marlboro) (15). In the cigarette smoke chamber exposure model, cigarettes are burned during 4–5 min/cigarette, that mimics the exposure to heated smoke. Smoke during each cycle is forced using 4 L/min air into the exposure chamber intermittently; each smoke cycle is designed to mimic the puffing of actual human smokers and to prevent asphyxiation of the mice: smoke cycles provide five seconds of heated cigarette smoke which is interrupted by twenty-five seconds of 4 L/min air using a timer controlled 2-way valve system (Humphrey, Kalamazoo, MI). The concentration of particulate matter and carbon monoxide during 30-minute smoke cycles of 1 cigarette in the mouse smoke chamber resulted in mean total particulate concentration of 550± 50 mg/m3 (±SD), and the average CO concentration in parts per million (ppm, v/v) during for one 30-minute smoke cycle experiment was 757 ppm (34). Mice were exposed to four cigarettes, five days a week for a total of six months. We used the number of cigarettes that approximated moderate to heavy smoking habits (e.g., greater than twenty pack year smoke exposure) in humans.

Emphysema mouse model induced by EFs immunization

C57BL/6 female mice (7 to 8 weeks old) were purchased from The Jackson Laboratory (Bar Harbor, ME). Human lung elastin fragments (hEFs; product ID QP45), rat lung alpha elastin fragments (rEFs; product ID RA50), and mouse lung elastin fragments (mEFs; product ID MLP54) were purchased form The Elastin Products Company (Owensville, MO). On day one, mice were immunized using subcutaneous (s.c.) injections with a mixture of 25 μg hEFs and 25 μg rEFs, or with 50 μg mEFs in Complete Freund’s Adjuvant (CFA) (BD, #263910) with 250 μg Mycobacterium tuberculosis (BD, #231141). After that, mice were immunized once a week by s.c. injections with a mixture of 25 μg hEFs and 25 μg rEFs, or with 50 μg mEFs in incomplete Freund’s Adjuvant (IFA) for a total of 7 weeks (see Schema in Figure 3).

Figure 3. Non-self EFs immunization-induced emphysema mouse model.

(A) Schematic diagram of EFs immunization. Mice were immunized with 25 μg human EFs (hEFs) and 25 μg rat EFs (rEFs) in CFA with 250 μg Mycobacterium tuberculosis by s.c. injection on day 1 (D1). Mice were then repeatedly immunized once a week (a total 7 times) with 25 μg human EFs and 25 μg rat EFs (h+rEFs) in IFA. Mice were analyzed one week after the last injection. (B) Representative three-dimensional images and (C) micro-CT quantification of lung volume from PBS or mixture of h+rEFs immunized mice. Average lung volume is shown below (B). (D) Representative images of H&E-stained lung sections; inset represents ×200 magnification. Scale bars, 100 μm. (E) MLI measurement from the indicated groups of mice. BALF analyses from the same group of mice showing total (F), macrophages (Mac) (F), lymphocytes (Lym) (G), and neutrophils (Neut) (H). Expression of MMP9 (I) and MMP12 (J) in BAL cells isolated from PBS or mixture of h+rEFs immunized mice. Data was combined with two different experiments. *P < 0.05, **P < 0.01 as determined by the Student’s t-test. Results are represented as mean ± SEM from three independent experiments with 4–5 mice in each group.

For analysis of experimental emphysema, immunized mice were euthanized and bronchoalveolar (BAL) lavage fluid, lungs, spleen, and aorta were collected. BAL fluid was collected by instilling and withdrawing 0.8 ml of sterile phosphate-buffered saline (PBS) twice through the trachea. Total and differential cell counts in the BAL fluid were determined with the standard hemocytometer and HEMA3 staining (Biochemical Sciences Inc, Swedesboro, NJ) using 150 μl of BAL fluid prepared with cytospin slides. Dissected lung tissues, spleen, and aorta were used to prepare single cell suspensions and/or histopathological studies using hematoxylin and eosin (H&E) staining as described previously (35).

Quantification of the experimental model of emphysema

To determine the severity of lung parenchymal destruction (emphysema), we used micro-computed tomography (microCT) and mean linear intercept (MLI) as previously described (34). The Animal Phenotyping Core at Baylor College of Medicine performed microCT studies on live anesthetized mice. Briefly, mice were placed in microCT (Gamma Medica, Salem, NH), and completed images of the chest were used to quantitate emphysema using Amira 3.1.1. software (FEI, Hillsboro, OR) (35).

Quantification of emphysema using MLI was done on blinded samples using ten fields that were randomly selected from the left lobe of lung. Paralleled lines were placed on the selected field and the number of intercepts was measured. MLI was calculated by dividing the length and the number of lines per field, multiplied by the number of intercepts.

Preparation of single cell suspensions from mouse thoracic aorta, lung, spleen, and lung draining lymph nodes (dLN)

Single cell suspensions from thoracic aorta were prepared by enzymatic digestion as previously described with minor modifications (36, 37). Briefly, thoracic aorta segments were dissected following vasculature perfusion with 2mM EDTA buffer(5ml), PBS (10ml), and FACS buffer (10ml) respectively. Thoracic aorta, segment above diaphragm, was cut into small pieces and digested in 3ml enzyme cocktail containing 400 U/ml collagenase type I (Millipore Sigma, #C0130), 120 U/ml collagenase type XI (Millipore Sigma, #C7657), 60 U/ml hyaluronidase type I (Millipore Sigma, #H3506), and 60 U/ml DNase I (Millipore Sigma, #11284932001) in RPMI supplemented with 10% heat-inactivated fetal bovine serum (hiFBS) for 50 minutes at 37°C. Digested products in suspensions were filtered through a 40 μm cell strainer (BD Falcon, San Jose, CA) followed by red blood cell (RBC) lysis (ACK lysis buffer, Millipore Sigma) for 3 min at room temperature.

Single cell suspensions from mouse lung, draining lymph nodes, or spleen were prepared by mechanically mincing collected organs though a 40 μm strainer followed by RBC lysis for 3 min at room temperature as described (38).

Mouse EFs specific CD4⁺ T cell cloning

Cloning of EFs specific CD4⁺ T cell was accomplished by serial dilution method as previously described with minor modification (39). Live CD3⁺CD4⁺CD25⁻ T cells were sorted using flow cytometry from single cell suspensions of the mouse lung, lung dLN, or spleen. The sorted cells were co-cultured with γ-irradiated [30 Gy (SI unite of absorbed dose of ionizing radiation)] CD3⁻ splenocytes isolated from CD45.1 congenic mouse as APCs in 1:2 ratio (T cells to APCs) in C-RPMI supplemented with 0.5% heat inactivated mouse serum (hiHMS, Equitech-Bio #SM-0100HI) in the presence of mEFs (30 μg/ml). After 3 days, the culture medium was replaced with C-RPMI supplemented with 10% hiFBS. After 2 days, the cells were placed in C-RPMI supplemented with 5% hiFBS. Ficoll-Pague Plus (GE Healthcare) gradient was used to remove dead cells by centrifugation at 500g for 20 minutes at room temperature. The collected live cells were placed in C-RPMI supplemented with 10% hiFBS and 10% T Cell Culture Supplement without ConA (BD Biosciences #354116). After 3 days, the culture medium was replaced with C-RPMI supplemented with 10% TCCS and incubated for 2 days. All steps from co-culture with APCs in the presence of mEFs were repeated for two additional rounds. After the third round of stimulation with mEF, T cells were diluted in C-RPMI supplemented with 10% hiFBS and 10% TCCS at serial concentrations in 96-well plate at concentrations of 1 cell per 100 μl and 10 cell per 100 μl. APCs (7 × 10⁵ cells per 100 μl) were added to each well in the presence of mEFs (final concentration 30 μg/ml). The medium was replaced every 3 days for 2 weeks, and at the same time T cells were re-stimulated by adding 7 × 10⁵ cells per 100 μl of APCs in the presence of mEF (30 μg/ml) until the growth of the cloned T cells becomes visible.

Identification of mouse elastin-specific TCRs

Mouse T cell clones reactive to mEFs were used to identify mouse elastin specific TCRs. RNA from T cell clone was extracted with TRIzol (ThermoFisher #15596026) following the manufacturer’s instructions. RNA from each T cell clone was reverse transcribed using a high-capacity cDNA reverse transcription kit (ThermoFisher cat#4368814) and mouse TCR specific RT primers to increase specificity; 10× buffer, 25× dNTP mix, 50U RT enzyme, 20U RNase inhibitor and 26 μM was used as the final concentration of each TCR specific RT-PCR primer in 20μl volume. Primer sequences were as follows: 5’- CTCAGCGTCATGAGCAGG-3’ (TRAC cDNA), 5’-CCATAGCCATCACCACCAG −3’ (TRBC1 cDNA), and 5’-CCATGGCCATCAGCACTAG −3’ (TRBC2 cDNA). The reaction mix was incubated at 25°C for 10 minutes, 45°C for 45 minutes, and 85°C for 5 minutes.

Following RT, we utilized the multiplex-nested PCR method designed with minor modification (40), in two rounds of amplification. The first round of PCR was performed with oligonucleotide mixture of 23 TRAV (each 363 nM final concentration), 19 TRBV forward (each 363 nM final concentration), single TRAC (3.3 μM final concentration) and TRBC (3.3 μM final concentration) reverse primers in total 25 μl containing 5× Go Taq buffer, 80μM dNTPS, 0.5U Go Taq DNA polymerase (VWR #PAM8298), and 2.5ul RT reaction product. The PCR cycles were 5 minutes at 94°C followed by 36 cycles of 30 seconds at 94°C, 30 seconds at 55°C, and 1 minute at 72°C, followed by 7 minutes at 72°C.

For the 2^nd round of amplification, 2.5μl of the 1^st PCR reaction product was used as a template for amplification of TCRα and TCRβ in two separate reactions using a corresponding internal sequence primer mix. The reaction components were similar to the 1^st round PCR using a mixture of 23 TRAV forward primers and single TRAC reverse for TCRα, and 19 TRBV forward and single TRBC reverse for TCRβ. The PCR cycles were 5 minutes at 94°C followed by 36 cycles of 30 seconds at 94°C, 30 seconds at 55°C, and 1 minute at 72°C, followed by 7 minutes at 72°C.

The PCR products were visualized on a 1% agarose gel, and purified using a PCR purification kit (ThermoFisher Scientific #K220001). The purified products were sequenced using TRAC or TRBC internal reverse primers for α and β PCR products, respectively. TCR sequences were analyzed with the IMGT/V-Quest tool (http://www.imgt.org/).

Auto-reactive T cell response against mEFs

CD4⁺ T cells from the lung or spleen of air- or cigarette smoke-exposed mice were isolated using mouse anti-CD4 conjugated MicroBeads (Miltenyi Biotec, #130–117-043) followed by AutoMACS (Miltenyi Biotec) positive selection. Single cells from the negative selection were used to isolate CD11c⁺ cells for APCs by positive selection with anti-CD11c conjugated MicroBeads (Miltenyi Biotec, 130–108-338). Isolated CD4⁺ T cells (0.5×10⁶) were co-cultured with γ-irradiated APCs (0.5×10⁵) in 10:1 ratio for 3 days with or without mEFs (30 μg/ml). After 3 days of co-culture, supernatants were collected and used to measure concentrations of cytokines: IFN-γ and IL-17A by Luminex.

T cell clone auto-reactivity was measured using co-culture with γ-irradiated CD3⁻ splenocyte APCs isolated using MicroBeads (Miltenyi Biotec, #130–094-973) with or without mEFs (30 μg/ml). T cell responses to mEFs were recorded as fold change; cytokine detected in the presence of mEFs divided by cytokine measured in the absence of mEFs(29).

Flow cytometry and antibodies

Flow cytometry was performed using BD LSR II (BD Biosciences) and data were analyzed with FlowJo software (Tree Star Inc., Ashland, OR). UV live/dead fixable dead cell stain (ThermoFisher) was used to exclude dead cells. For intracellular cytokine staining of IFN-γ and IL-17A, cells were stimulated in 10% FBS contained RPMI with phorbol 12-myristate 13-acetate (PMA, 10ng/ml; Sigma-Aldrich, St.Louis, MO), ionomycin (1μg/ml; Sigma-Aldrich), and brefeldin A (10μg/ml; Sigma-Aldrich) overnight. Cells were stained with following anti-mouse antibodies: Pacific blue-conjugated anti-CD3e (500A2), PE-Cy5-conjugated anti-CD4 (RM4–5), and allophycocyanin (APC)-Cy7-conjugated anti-CD8a (53–6.7) purchased from eBioscience (San Diego, CA). Cells were then fixed with FACS lysing solution (BD BioSciences, San Jose, CA), permeabilized, and stained with phycoerythrin (PE)-conjugated anti-IL-17A (Thermo Fisher, TC11–18H10), APC-conjugated anti-IFN-γ (Thermo Fisher, XMG1.2). For transcription factor analysis, cells were stained with PE-RORγt (Thermo Fisher, BD2), PE-Cy7-T-bet (Thermo Fisher, 4B10), Alexa Fluor 647-Foxp3 (Thermo Fisher, FJK-16s), and Brilliant Violet 421-Gata3 (Biolegend, 16E10A23) according to the manufacturer’s protocol (Thermo Fisher, Waltham, MA). For neutrophil and cDC analyses, following antibodies were used; eFluro450-B220 (Thermo Fisher, RA3–6B2), PE-CD11b (Biolegend, M1/70), APC-CD11c (Thermo Fisher, N418), BUV395-CD45 (BD Biosciences, BD30-F11), or BV510-Ly6G (BD Biosciences, 1A8).

mRNA isolation and quantitative PCR

Total RNA from thoracic aorta tissue treated with RNAlater RNA stabilization reagent (Qiagen, #1018087) was extracted using RNeasy fibrous tissue isolation kit (Qiagen, #74704) following the manufacturer’s instructions. Total RNA from mouse BAL cells was extracted with TRIzol (Invitrogen, Carlsbad, CA) following the manufacturer’s instructions. cDNA was synthesized with the High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Forster City, CA). Gene probes, MMP9 (Mm00600164_g1) and MMP12 (Mm00500554_m1), were purchased from Thermo Fisher. All data were normalized to 18S ribosomal RNA (Hs99999901_s1) expression.

Statistical analysis

Statistical analyses were performed using GraphPad Prism version 8 for Mac OS X (GraphPad Software, La Jolla, CA). All data shown in the figures are the mean ± standard error of mean (s.e.m.). We used one-way analysis of variance (ANOVA) with Bonferroni’s correction for multiple comparisons or the Student’s t-test using two-tailed parameters.

Results

Generation of HLA restricted human elastin-specific m:hTCR

We have previously identified and cloned several elastin specific-CD4⁺ T cells from PBMCs isolated from smokers with emphysema (29). Elastin-specific T cell clones and bulk T cells isolated from PBMC in smokers showed specificity for hEFs as determined by increased cytokine secretion, and a requirement for MHC class II recognition (27–29). To determine the optimal conditions that could be used to develop a novel elastin-induced model of emphysema, we sequenced TCRα- and TCRβ-chains of one the human T cell clones, TC-378–4, using multiplex-nested PCR and designed primers complementary to all human TCRα and TCRβ variable regions as previously described (31) (Fig. 1A–C). We next sub-cloned the purified human TCRα and TCRβ PCR products to a retroviral vector to generate human elastin-specific TCR by transducing murine surface TCR-deficient 4G4 thymoma cells with a hTCR retroviral vector expressing elastin-specific which we labeled as m:hTCR (Fig 1D).

Figure 1. Workflow schematic for human EFs-specific TCR identification and the results.

(A) Reverse transcribe (RT) mRNA from human EFs-specific T cell clone (TC-378–4) isolated from emphysema patient using TCR specific primers. (B) Detail of the multiplex-nested PCR (two rounds of PCR) to amplify hTCRα- and β- chain (Left). The amplified PCR products on 1% agarose gel (Right). (C) Sequencing result. (D) Construct map after hTCRα- and β- chain serial insertion into a retroviral vector to generate hTCR expressing 4G4 cells (m:hTCR cells) (Left). Representative flow cytometry is showing successful expression of the hTCR on the cell surface of a transduced 4G4 thymoma cell lines gated on Ametrine (Right). (E) HLA restriction assay with HLA blocking antibodies (DQ/DR) or isotype control antibody (2.5 μg/ml). HLA-matched antigen presenting cells (APCs) were pre-treated with blocking antibody for 1hour in 37°C and cultured with m:hTCR cells for 4 days in the presence of hEFs (30 μg/ml). ***P<0.001, **P<0.01, as determined by one-way ANOVA test with Bonferroni’s multiple comparison.

Next, we deep-sequenced the TC-378–4 clone to identify the endogenous MHC class II class molecules required for activation of m:hTCR in functional assays (Table 2). Co-culture of m:hTCR cells with endogenous MHC class II APCs, resulted in increased IL-2 production when compared to control (Fig. 1E). Further, blocking HLA DQ molecule with the anti-DQ antibody, but not anti-DR or isotype control, attenuated IL-2 production in m:hTCR transfectant cells stimulated with hEFs. These findings suggested that I-A molecule expressed in C57BL/6J strain, a mouse counterpart of human DQ (41), is required to present EFs by APCs in mice (Fig. 1E). Therefore, we utilized C57BL/6J strain to develop a novel antigen-induced model of emphysema.

Table 2.

Identified TCRs of mouse Elastin-specific T cell clone

Organ	Clone #	TCRα			TCRβ
		TRAV	TRAJ	CDR3α	TRBV	TRBJ	TRBD	CDR3β
Lung	A5	5–4 or 5D-4	52	CAASDTNTGANTGKLTF	19	2–3	1	CASSRGQGHAETLYF
	C5	14D-1	43	CAASWDNNAPRF	15	2–4	2	CASSVDWGNTLYF
	E1	17	58	CALEGQGTGXKLSF	13–2	2–5	1	CASGATGGADDTQ
	G1	12–2 or 12D-1	23	CALSDQNYNQGKLIF	5	2–5	2	CASSPDWDGDTQYF
	G2	14D-1	5	TVFWVKTQVVGQLTF	5	1–2	2	CASSQEMDSDYTF
	H4	10	47	CAARSKDYANKMIF	31	2–3	2	CAISSAETLYF
Spleen	A6	6–6 or 6D-6	34	CALLSSNTNKVVF	31	2–5	2	CAWSLGGAHQDTQYF
	B3	6–6 or 6D-6	34	CALLSSNTNKVVF	31	2–5	2	CAWSLGGAHQDTQYF
	B4	6–6 or 6D-6	34	CALLSSNTNKVVF	31	2–5	2	CAWSLGGAHQDTQYF
	D5	6–6 or 6D-6	34	CALLSSNTNKVVF	31	2–5	2	CAWSLGGAHQDTQYF
	E1	6–6 or 6D-6	34	CALLSSNTNKVVF	31	2–5	2	CAWSLGGAHQDTQYF
	F3	6–6 or 6D-6	34	CALLSSNTNKVVF	31	2–5	2	CAWSLGGAHQDTQYF
	H4	6–5 or 6D-5	16	CAXSEPSSGQKLVF	13–2	2–1	1	CASGDTGGYAEQFF

TCR V, D, and J nomenclature follow that of IMGT. Identical TCR identified in different clones is shown in bold.

Detection of mEF specific T cells in cigarette smoke-model of emphysema

We have previously shown that CD4⁺ T cells isolated from the PBMCs of smokers with emphysema differentiate to T helper type 1 (Th1) and Th17 cells in response to human lung elastin fragments, indicating a strong association between anti-elastin autoimmunity and emphysema in smokers (27). Although cigarette smoke-induced model of emphysema induces robust Th1 and Th17-mediated lung inflammation (15, 34), anti-elastin autoimmunity in this model had not been examined. To determine whether cigarette smoke induces elastin specific T cells, we examined mEF-specific auto-reactive T cells in C57BL/6J mice exposed to six months of cigarette smoke. Compared to air-exposed control mice, CD4⁺ T cells from emphysematous lungs exposed to cigarette smoke reacted to mEFs as determined by increased expression of IFN-γ (Fig 2A), indicating similar autoreactive immune responses as detected in the lungs of smokers with emphysema (28, 29). However, mEFs failed to induce IL-17A in CD4⁺ T cells under the same conditions (Fig 2B).

Figure 2. mEF specific auto-reactive CD4⁺ T cells from cigarette smoke-exposed mice.

Lung CD4⁺ T cells of air- or cigarette smoke-exposed mice were cultured with γ-irradiated CD11c⁺ cells with or without mouse EFs (30 μg/ml) for 3 days. After stimulation, the supernatants were harvested and used for the cytokine measurement. The concentration of IFN-γ (A) and IL-17A (B) were plotted as fold change over nil stimulation. Each dot represents a data point from an individual mouse. Data are from two independent experiments. *P < 0.05 as determined by the Student’s t-test.

Immunization with hEFs and rEFs induce emphysema in mice

Our findings so far indicate that cigarette smoke exposure induces Th1 specific loss of tolerance against elastin in C57BL/6J mice, but also highlight differences between animal models and human emphysema where the former occurs in the absence of respiratory infection, while the latter is often complicated with repeated bouts of microbial infections (42–44).

Therefore, to address the role of autoimmunity in the pathogenesis of emphysema, we immunized mice using elastin fragments. To accomplish this goal, we first chose C57BL/6J mice to immunize against mEFs isolated from the lungs of congenic mice. We found that repeated immunization with mEFs isolated from the lungs of C57BL/6J mice did not result in the induction of elastin specific T cells, lung inflammation or emphysema (Supplemental Fig. 1). We reasoned that immunization with a mixture of hEFs and rEFs, that have 72.6%, and 91% homology with mEFs (45), could induce autoimmunity in mice. Therefore, we next immunized with hEFs and rEFs (h+rEFs) using the same protocol (Fig 3A). We found that immunization with h+rEFs resulted in emphysema as determined by increased total lung volume quantified by micro-computed tomography (microCT) (Fig 3B–C), enlarged alveolar spaces detected by hematoxylin and eosin staining of lung sections (Fig 3D). Unbiased lung morphometry measurement (mean linear intercept; MLI) (Fig 3E) also showed emphysema in h+rEFs immunized mice when compared to control (PBS immunized) mice. Examination of lung inflammatory cells showed significantly increased numbers of macrophages lymphocytes and neutrophils (Fig 3F–H), that were present in the BAL fluid compared to control mice. Increased lung inflammation and emphysema was consistent with increased expression of matrix metalloproteinase 9 (MMP9) and MMP12 mRNA expression (Fig 3I–J).

Immunization with h+rEFs induces Th1 and Th17 lung inflammation

We next determined whether h+rEF immunization induces innate and adaptive inflammation in the lungs parenchyma. Single cell examination of lungs from h+rEF immunized mice showed increased relative abundance of lung CD11c⁺CD11b^High mDCs (Fig. 4A–B), and Ly6G⁺CD11b⁺ neutrophils (Fig. 4C–D), when compared to control (PBS immunized) mice. We next examined whether, in the same immunized mice, there is an induction of adaptive immune cells. Intracellular cytokine staining of CD4⁺ cells showed increased expression of IL-17A (Th17) and IFN-γ (Th1) producing cells in the lungs of h+rEF immunized mice, compared to controls (Fig. 4E–G). We also found an increased relative abundance of CD8⁺ (Tc17) and CD8⁺ (Tc1) T cells (Fig. 4H–I) in the same group. Notably, mice immunized with hEFs alone failed to develop lung inflammation or emphysema (Supplemental Fig. 2). Although immunization with rEFs alone resulted in mild increase in BAL cell numbers, enlarged alveolar space, and increased Th17 cell in the lung (Supplemental Fig. 3), the changes were less robust when compared to the combined h+rEF immunization protocol (Figs 3–4).

Figure 4. Non-self EFs immunization induces inflammation in the lung.

Representative (A) and cumulative (B) flow cytometry analysis of B220⁻CD11c⁺CD11b^High mDC in the lung of PBS or mixture of hEFs and rEFs (h+rEFs) immunized mice. Representative (C) and cumulative (D) flow cytometry analysis of CD45⁺Ly6G⁺CD11b⁺ neutrophils (Neut) in the lung of the same group of mice. Representative intracellular staining (E) and cumulative analysis of IL-17A (Th17) (F) and IFN-γ (Th1) (G) expressing CD4⁺ T cell subsets in the lung. Cumulative intracellular cytokine staining of IL-17A (Tc17) (H) and IFN-γ (Tc1) (I) expressing CD8⁺ T cell subsets in the lung. *P < 0.05, **P < 0.01 as determined by the Student’s t-test. Results are represented as mean ± SEM from three independent experiments with 5 mice in each group.

We next explored whether h+rEF immunization induced inflammation in other elastin rich organ (e.g., aorta). Consistent with Th1 and Th17 lung inflammation, we found increased relative abundance of RORγt⁺ T (Th17) and T-bet⁺ T (Th1) expressing T cells in the thoracic aorta (Fig 5A–B). Furthermore, we found increased expression of Mmp9 and Mmp12 in the same tissue indicating organ specific effect of inflammation on the affected organ (Fig 5C–D). h+rEF immunized mice also showed increased Th1 and Th17 cells in the spleen indicating the systemic impact of immunization in this model (Data not shown). Together, these findings strongly support that immunization with non-self EFs (e.g., hEFs and rEFs) but not mEF induce systemic and organ-specific inflammation in mice.

Figure 5. Non-self EFs immunization induces inflammation in the thoracic aorta.

(A) Representative staining of transcription factors (Foxp3 and RORγt) staining and cumulative analysis of RORγt expressing CD4⁺ T cells (Th17) in the thoracic aorta from PBS or mixture of hEFs and rEFs (h+rEFs) immunized mice. (B) Representative staining of transcription factors (T-bet and GATA3) staining and cumulative analysis of T-bet expressing CD4⁺ T cells (Th1) in the thoracic aorta from the same group of mice. Expression of MMP9 (C) and MMP12 (D) in the thoracic aorta isolated from the same group of mice. *P < 0.05 as determined by the Student’s t-test. Results are represented as mean ± SEM from two independent experiments with 5 mice in each group.

Isolation and identification of mEFs-specific CD4⁺ T cell clones

Our findings thus far provided evidence that immunization with r+hEF, but not mEFs, induces systemic innate and adaptive immune responses in mice. We next isolated CD4⁺ T cells from the lung and spleen of h+rEF immunized mice and co-cultured them with irradiated CD11c⁺ antigen presenting cells (APCs) isolated from the spleen of congenic mice (Supplemental Fig. 4A). After initial stimulation with mEFs, T cell cultures were subjected to 3 additional round of stimulus to enrich for mEF-specific T cell clones. Before each series of stimulation, dead cells were removed and live T cells quantified. As expected, T cells expanded in response to mEFs, suggesting that h+rEF immunized mice results in the induction of autoreactive T cells in mice (Fig. 6).

Figure 6. Proliferation of auto-reactive T cells in response to mouse (m)EFs.

Expansion of mEFs-specific CD4⁺ T cell from the lung (A) and spleen (B) of control (PBS) or h+rEF immunized was monitored using hemocytometer prior to serial dilution and clonal expansion. Cells were stimulated with mEFs (30 μg/ml) on Days 0, 14, and 28. Dead cells were removed on Days 5, 22, 34. Serial dilution and clonal expansion was done on Day 34.

Next, the antigen-enriched T cells were used for T cell cloning by serial dilution which resulted in multiple clones (Supplemental Fig. 4B). To identify the mEFs-specific TCR, each T cell clone was expanded and subjected to multiplex-nested PCR and TCR sequencing (Supplemental Fig. 4C–D). We identified 2 unique TCRs from a total of 7 T cells clones generated from splenocytes of immunized mice and 6 unique TCRs from a total of 6 T cell clones isolated from the lung cells (Table 2). Finally, we validated whether T cell clones were reactive against mEFs. Compared to CD4⁺ T cell isolated from control mice, T cell cloned from spleen and lung of r+hEF mice stimulated with mEFs exhibited increased IFN-γ secretion (Fig 7 A, C).

Figure 7. Responses of T cell clones from the lung and the spleen against mouse (m)EFs.

Each mouse EFs-specific T cell clone generated from the lung (total of 6 different T cell clones, A and B) and the spleen (a total of 7 different T cell clones, C and D) were cultured with γ-irradiated CD3⁻ splenocytes with or without mEFs (30 μg/ml) for 3 days. CD4⁺ T cells isolated from splenocytes were used as control (Ctrl, N=4). After co-culture, the supernatants were harvested and used for the measurement of cytokine concentration. The concentration of IFN-γ (A and C), and IL-17A (B and D) were plotted as fold change over nil stimulation. Each dot represents a data point from individual T cell clone. Results are mean ± SEM; *P<0.05 as determined by the Student’s t-test.

Discussion

In this report, we provide the first antigen specific model of emphysema as a novel tool that could be used to determine the pathophysiological role of adaptive immunity in cigarette smoke-induced lung disease. This model recapitulates smoke-induced lung disease in humans because we showed loss of immune tolerance to elastin, induction of autoreactive T cells from systemic and elastin rich organs, and pathophysiological changes in the lungs consistent with emphysema. Although instillation of elastase in the lungs has been used a non-smoke model of emphysema, to our knowledge, immunization with elastin fragments has not been shown to induce emphysema in mice.

Autoreactive T cells against elastin have been cloned from peripheral blood of smokers with emphysema and shown to secrete several proinflammatory cytokines in response to their cognate antigens (46). Although human association studies have provided strong support for cigarette smoke-mediated induction of autoimmunity as a potential mechanism for emphysema, whether loss of tolerance to elastin, independent of cigarette smoke, could cause emphysema has not been previously shown. This report provides the methods employed that successfully broke tolerance to elastin fragments and resulted in the generation of autoimmune inflammation in multiple elastin rich organs in C57BL/6J mice. We show here that immunization with mouse EFs failed to break tolerance to elastin; however, a combination of human and rat EFs was sufficient to induce autoimmunity in mice.

The elastin gene, ELN, contains a higher number of repeat elements than the rest of the genome, while alignment report of elastin DNA sequences in vertebrates shows little over 64% identity at the nucleotide and 72 % similarity at the amino acid level between human and mouse (45). Interestingly, the similarity between rat and mouse at the amino acid is reported at 91%, indicating substantial similarities between rodents that diverge from other mammals (45). Nonetheless, immunization of mice with mEFs failed to induce lung inflammation and emphysema. Prior studies have shown that immunization with unaltered (endogenous) proteins promote antigen-specific Tregs (47, 48); similarly, autoantigen-specific T cells against Apolipoprotein B detected in healthy cohorts were shown to be phenotypically biased toward Tregs (49). However, whether such regulatory cells are present in elastin-mediated autoimmunity remains unknown, and would be of importance to examine in the future studies.

Other autoimmune models have used human protein, (e.g., myelin oligodendrocyte glycoprotein) to induce T cell mediated autoimmunity in Experimental Autoimmune Encephalomyelitis (EAE) in mice, a model for neuronal degeneration that mimics encephalopathy seen in patients with multiple sclerosis (50). The mechanism by which combined h+rEFs can promote stronger autoimmune responses could be related to the strength and diversity of homologous antigens presented to T cells. Prior studies have established that sequence homology between self- and xeno-derived peptides could be the underlying mechanism resulting in cross-activation of autoreactive T cells (51). Specifically, the similarity between foreign and self-elastin peptides (human 71% and rat 91% homology with mouse elastin respectively), can expose T cells to multiple cryptic epitopes to promote epitope spreading and/or induce molecular mimicry to expand autoreactive T cells. Notably, immunization with xeno-molecules (hCollagen peptides, etc.) can promote tissue specific autoimmunity in mice (52) but to our knowledge development of immunity to mouse elastin induced by h+rEFs molecule has not been previously reported. The introduction of this breakthrough model opens a new paradigm and allows investigations into specific roles of TCR-specific pathogenic T & B lymphocytes that promote loss of immune tolerance in emphysema. A caveat of the current study is the time-course whereby the persistence of autoreactive T cells remains to be clear. However, several well-established mouse models of autoimmune diseases (e.g. EAE, arthritis, colitis etc.,) result in a transient loss of immune tolerance (53–55); despite their transient nature these models have provided invaluable new information to our understanding of the pathophysiology of autoimmunity. Future studies should include an exact time-course to examine whether and how autoimmunity can persist in this new model of emphysema.

Induction of autoimmune diseases requires multiple known and elusive etiologic factors that include environmental triggers, genetic predisposition, and threshold variability for autoimmunity in different organs (52). Consistently, development of emphysema in smokers is highly variable, with some estimates approximating 25% of all smokers (56–58), indicating a strong genetic susceptibility factor contributes to this disease. These findings highlight the need for an animal model that could be used to determine the stability of TCR/HLA/peptide tri-molecular complex to understand the pathobiology of emphysema.

Cigarette smoke represents one of the most critical environmental trigger associated with the induction of autoimmune inflammation in humans, including rheumatoid arthritis, multiple sclerosis, and as we have shown emphysema (15, 59–61). Multiple animal models of chronic exposure to cigarette smoke, have been developed to study the acute and chronic inflammatory recruitment of immune cells in the lung (62, 63). However, cigarette smoke-induced emphysema mouse model has some limitations to study anti-elastin autoimmunity, because cigarette contains over 5,000 chemicals, which can strongly activate innate immune cells producing elastolytic enzymes resulting in parenchymal destruction like emphysema even in the absence of adaptive immune cells (64). As such, the new model could provide the necessary tool to determine factors that allow the loss of immune tolerance to self.

A second critical factor in the development of autoimmunity in humans is the contribution of the HLA polymorphism. For instance, expression of DQB1*03:01 in rheumatoid arthritis patients has been found to be significantly associated with an increased incidence of chronic obstructive pulmonary disease (COPD) and emphysema, (65, 66). Using EFs-specific auto-reactive T cell clone generated from emphysema patient, we found HLA-DQ (DQA1*02:01 and 05:05, DQB1*02:02 and *03:01) corresponding to I-A molecule of MHC class II is required to present EFs in mice (Table 1 and Fig. 1).

Subcutaneous immunization of human and rat EFs induced organ specific inflammation (e.g. lungs and aorta) resembling smoke-induced inflammatory disease found in susceptible smokers (27, 28). Specifically, immunized mice developed inflammatory responses in the lung characterized by increased innate immune cells (mDCs and neutrophils), infiltration of adaptive immune cells (T cells), as well as lung parenchymal destruction. Moreover, for the first time, we generated EFs specific auto-reactive T cell clones from proliferating cells in response to mEFs, supporting the loss of tolerance against self-antigen.

Using this model, we have cloned multiple elastin-specific T cells with potential pathogenic autoimmune responses. TCR plays an essential role in the regulation of immunological tolerance. Specifically, TCR affinity induces the differentiation of T cell lineages such as Tregs and removes auto-reactive T cells during thymic selection. The highly variable CDR3 region of TCR provides antigenic specificities (67), a tool that could be used to characterize the dynamics of pathogenic T cell responses in autoimmunity in this disease model. Here we identified several CDR3 regions of TCR using autoreactive T cell clones against mEFs. Future studies will characterize pathogenic TCRs, that could play a critical role in the induction of inflammation in the lung and emphysema.

In conclusion, we provide the methods for development of the first mouse model anti-elastin autoimmunity that results in emphysema. This model could be used to examine the critical genetic, and environmental factors that reduce susceptibility to loss of immune tolerance to self-proteins. Further, the roadmap to generate m:hTCR expressing cell lines, could be utilized to develop retrogenic models to identify pathogenic autoreactive T cells in future studies. Similarly, future studies could determine the tolerogenic epitopes, which could induce and expand antigen-specific Treg to maintain the balance between inflammation and tolerance for immune homeostasis.

Key Points:

A novel model of autoimmunity against elastin resulted in emphysema. Elastin-specific TCR sequences identified autoreactive T cell clones in the lung.