Sarcoidosis is a granulomatous disease that primarily affects the lungs and is characterized by an accumulation of CD4+ T cells in the bronchoalveolar lavage (BAL). Previous work has indicated that HLA-DRB1*03:01+ (DR3+) patients diagnosed with the acute form of the disease, Löfgren’s syndrome (LS), have an accumulation of CD4+ T cells bearing TCRs using TRAV12-1 (formerly AV2S3). However, the importance of these α-chains in disease pathogenesis and the paired TCRβ-chain remains unknown. This study aimed to identify expanded αβTCR pairs expressed on CD4+ T cells derived from the BAL of DR3+ LS patients. Using a deep-sequencing approach, we determined TCRα- and TCRβ-chain usage, as well as αβTCR pairs expressed on BAL CD4+ T cells from LS patients. TRAV12-1 and TRBV2 (formerly BV22) were the most expanded V region gene segments in DR3+ LS patients relative to control subjects, and TRAV12-1 and TRBV2 CDR3 motifs were shared among multiple DR3+ LS patients. When assessing αβTCR pairing, TRAV12-1 preferentially paired with TRBV2, and these TRAV12-1/TRBV2 TCRs displayed CDR3 homology. These findings suggest that public CD4+ TCR repertoires exist among LS patients and that these T cells are recognizing the putative sarcoidosis-associated Ag(s) in the context of DR3.

Sarcoidosis is a multisystem granulomatous disorder of unknown etiology that predominantly affects the lungs (1). The clinical presentation and severity of sarcoidosis differ across individuals of different races and ethnic backgrounds. However, in individuals of primarily Scandinavian descent, an acute form of sarcoidosis, known as Löfgren’s syndrome (LS), occurs with a uniform clinical presentation of fever, bilateral hilar lymphadenopathy, erythema nodosum, and/or ankle arthritis (13). A consistent finding in sarcoidosis and LS is an immune response that is characterized by the accumulation of activated CD4+ T cells in the lungs.

The familial clustering of sarcoidosis supports a genetic contribution to disease development. Genetic susceptibility to LS has been strongly linked to the HLA-DRB1*03:01 (DR3) allele (4), and an association between the presence of DR3 and expansions of oligoclonal TRAV12-1 (previously designated as AV2S3)–expressing CD4+ T cells in the lung has been described (57). HLA-DRB3*01:01 (DRB3) is a similar MHC class II molecule that binds and presents nearly identical peptide epitopes as DR3 (8). Previous work has shown that TRAV12-1–expressing CD4+ T cells also accumulate in the lungs of DRB3+ LS patients (7, 9). Importantly, TRAV12-1–bearing CD4+ T cells disappear from the lungs upon disease resolution, indicating their involvement in disease pathogenesis and/or resolution (5, 10). Compared with the strong association between TRAV12-1 and LS, usage of TCRβ-chains has been less consistent, with expansions of CD4+ T cells using TRBV2, TRBV5, TRBV12, TRBV14, TRBV18, and TRBV24 having been previously demonstrated in differing sarcoidosis populations (6, 1115). In addition, no study has elucidated complete αβTCR pairs expressed on CD4+ T cells derived from the lungs of patients with active disease.

Previous work characterizing lung-accumulated CD4+ T cells in sarcoidosis patients has been hindered by a lack of suitable methodology for determining both α- and β-chain usage on single cells. In the current study, we overcame this limitation by using iRepertoire PCR to identify CD4+ TCR α and β gene usage and further used emulsion PCR (ePCR) and single-cell PCR (scPCR) to elucidate αβTCR pairing on CD4+ T cells derived from the lungs of DR3+ LS patients. In this article, we show that TRAV12-1 preferentially pairs with TRBV2 on CD4+ T cells. Furthermore, we identify shared and similar TRAV12-1 and TRBV2 CDR3 motifs across patients, suggesting the existence of public TCRs expressed in multiple DR3+ LS patients. Collectively, these findings suggest the accumulation of TRAV12-1/TRBV2–expressing CD4+ T cells in the lungs of LS patients in response to related Ags.

A total of 13 newly diagnosed sarcoidosis patients (9 LS and 4 non-LS) were included in the study (Table I). All patients were diagnosed with sarcoidosis according to criteria established by the World Association of Sarcoidosis and Other Granulomatous Disorders (16). These criteria included typical clinical and radiographic manifestations, an elevated bronchoalveolar lavage (BAL) CD4/CD8 ratio and, if required, biopsy evidence of granulomatous inflammation, as well as exclusion of other diagnoses. Informed consent was obtained from all subjects, and ethical approval was granted by the Stockholm County Regional Ethical Committee.

Genomic DNA was extracted from whole blood samples of all patients. HLA-DRB1 and DRB3 alleles were subsequently determined using the PCR sequence-specific primer technique (SSP-DR Low Resolution Kit; Olerup, Saltsjöbaden, Sweden), as previously described (17).

Bronchoscopy with BAL and PBMC isolation were performed as previously described (13, 17). Cells for iRepertoire, emulsion, and nonemulsion analyses were thawed and washed in RPMI 1640 supplemented with 10% FBS (both from HyClone, Logan, UT), 5% sodium pyruvate, 5% Penicillin/Streptomycin/Glutamine, and 5% HEPES (all from Life Technologies, Carlsbad, CA) (complete RPMI). BAL cells were incubated in 6 or 12 well plates for 20 min at 37°C to adhere alveolar macrophages. CD4+ T cells were obtained after positive selection cell sorting (Dynabeads; Life Technologies) and incubated overnight in c-RPMI supplemented with a low concentration (30 U/ml) of IL-2 to prevent cell death and to ensure cell viability before sorting. After 16–18 h, the cells were aliquoted equivalently into three 1.7 ml tubes for iRepertoire, emulsion, or nonemulsion RT-PCRs.

RNA was purified from CD4+ T cells according to the manufacturer’s instructions in the RNeasy Mini Kit (QIAGEN, Hilden, Germany). Two rounds of PCR using the iRepertoire human TCR α and β kits (iRepertoire, Huntsville, AL) were performed according to the manufacturer’s instructions.

Cells for emulsion and nonemulsion reactions were washed twice in MACS buffer (Dulbecco’s PBS with 0.5% FBS [both from HyClone] and 1% HEPES [Life Technologies]). Cells were centrifuged and resuspended in 50 μl of RT-PCR Master Mix (1× OneTaq One-Step RT-PCR buffer and enzyme [New England Biolabs, Ipswich, MA], 10 U RNase out [Invitrogen, Carlsbad, CA], 2.5 mg/ml acetylated BSA [Sigma-Aldrich, St. Louis, MO], 400 nM each of the C region primers and Step-Out primers, and 40 nM each of the V region primers, as described in Supplemental Table IA). The emulsion cells had 300 μl of the oil mixture added, according to the specifications in the Micellula DNA Emulsion and Purification Kit (Chimerx, Milwaukee, WI). The tubes were vortexed at 4°C for 2.5 min. Nonemulsion cells (in 50 μl of RT-PCR mix) were added to PCR tubes, and the vortexed emulsion cells (in 50 μl of RT-PCR mix plus 300 μl of oil) were divided into three PCR tubes. The cycling for PCR1 was as follows: 65°C for 2 min, 48°C for 40 min, 94°C for 2 min, (94°C for 30 s, 54°C for 60 s, and 68°C for 2 min) ×35 cycles, and 68°C for 10 min.

The separated triplicate emulsion samples were pooled together for purification using the Micellula DNA Emulsion and Purification Kit (Chimerx), according to the manufacturer’s instructions. The nonemulsion samples were purified using a GeneJET PCR Purification Kit (Thermo Scientific, Waltham, MA).

The first nested PCR (PCR2) used 2 μl of the nonemulsion PCR1 product or 20 μl of the emulsion PCR1 product in a 100-μl reaction. The master mix included 1× Standard Taq buffer and enzyme, 10 mM dNTPs, 100 nM each nested primer (Supplemental Table IB), and 1.6 μM blocking oligos (Supplemental Table IB). The cycling for PCR2 was as follows: 95°C for 30 s, (95°C for 30 s, 54°C for 30 s, and 68°C for 1 min) ×20 cycles, and 68°C for 10 min.

The second nested PCR (PCR3) used 2 μl of nonemulsion or emulsion PCR2 product in a 50-μl reaction. The master mix included 1× LongAmp Buffer and enzyme, 10 mM dNTPs, and 100 nM each nested primer (Supplemental Table IC). The cycling for PCR3 was as follows: 94°C for 30 s, (94°C for 30 s, 53°C for 30 s, and 65°C for 1 min) ×25 cycles, and 65°C for 10 min.

As a control, we performed nonemulsion reactions in which all primers and regents were present, but no oil mixture was used during PCR1. This setup allowed for any TCRα product to combine with any TCRβ product regardless of which CD4+ T cell each product originated.

BAL cells were thawed and placed in culture at 2.5 × 106 cells per milliliter at 37°C overnight in standard tissue culture medium supplemented with IL-2, as described above. Single-cell sorting of DAPI, CD3+, CD4+, TRAV12-1+ stained T cells was performed on a BD FACSAria (Becton Dickinson, San Jose, CA). Cells were directly sorted as 1 cell per well into reverse-transcription buffer in 96-well polypropylene plates. For cDNA synthesis and subsequent PCRs, cells were treated as previously described (18). Briefly, cells were lysed, and RNA was reverse transcribed using a combination of random hexamers and TCRAC/BC gene–specific primers. Superscript was added, and the sealed plates were incubated in an oven for 75 min at 55°C. For each T cell, TCRA and TCRB gene expression was ascertained separately in a series of three PCRs done in 96-well PCR plates, as previously described (18). The initial PCR was a multiplex reaction (40 TRAV and 33 TRBV primers plus C region primers, each at 0.5 μM) using Advantage 2 Polymerase Mix in a 50-μl reaction volume for 35 cycles. The first-round PCR products were diluted, and 2.5 μl of template was added to a 25-μl second-round PCR using nested TCRAC and BC primers on the 3′ end. The 5′ primer for the second-round PCR was a common sequence that was added to the 5′ end of each of the V region primers used in PCR round 1 (designated the Illumina short primer). After 35 cycles of amplification, the products were diluted and used as template in a third round of PCR. This final amplification of 18–20 cycles incorporated barcodes on both ends of each PCR product to enable identification of sequences run on the Illumina MiSeq platform. Thus, unique pairs of forward and reverse primers were added to each well of a 96-well plate that specified each T cell by its position on the plate.

PCR products were gel purified and quantified using a Qubit Fluorometer (Life Technologies). Samples were pooled separately for each application (iRepertoire, ePCR, or scPCR) and sequenced on a MiSeq instrument (Illumina) for paired-end 2 × 250 (ePCR and scPCR) or paired-end 2 × 150 (iRepertoire) reading. The TCR sequences presented in this article are listed according to International ImMunoGeneTics Information System nomenclature (http://www.imgt.org) and have been submitted to the Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo) under accession number GSE100378.

The open-source Galaxy Platform (https://usegalaxy.org) was used to process the raw sequencing files, as outlined below and as previously described (19). For each of the iRepertoire or scPCR samples, the raw read 1 and read 2 sequences were paired in Galaxy using the Pear function, and the samples were split into individual files using the Barcode Splitter function and used for MiTCR identification (see below). The emulsion and nonemulsion read 1 and read 2 sequences were paired at the 3′ end in Galaxy using the FASTQ Joiner function to generate 500-nt–length products with small portions of the C region at either end of the products. The products were trimmed using the Trim Sequences function in Galaxy to create an α read and a β read, which were exported and used for TCR identification and pairing. The TCR identification for all methods was performed using MiTCR (https://github.com/milaboratory/mitcr) and the command prompt instructions given by the developer. MiTCR is an open-source bioinformatics software that identifies TCR V, D, and J gene segment usage (International ImMunoGeneTics Information System nomenclature) and extracts the CDR3 sequence for each sequencing read. For the emulsion, nonemulsion, and scPCR samples, the α- and β-chains were paired based on identical cluster identifications using MiTCRCombine, as described previously (19). The results for all samples were analyzed further in Microsoft Excel. A cutoff of 10 reads was used for all samples (i.e., any sequence with a read count ≤ 9 was not included in any of the analyses).

iRweb (http://www.irepertoire.com) was used to calculate the number of CDR3s shared among patient groups and for determining D50 values. CDR3 sharing was calculated by the CDR3 Algebra function in iRweb. Sequences found in any other patient group were excluded. Thus, the numbers reflect only those CDR3 sequences that are shared exclusively in the patients indicated. CDR3 sequences were required to be identical (but with no restrictions on V or J usage) and must have been found in at least two patients in the given group. For D50, the number of unique CDR3s in a given sample was determined using the following formula: D50 = (number of unique CDR3s that make up 50% of the total reads * 100)/number of unique CDR3s.

CDR3 motifs were found among DR3+ LS patients by using a defined set of criteria. First, the motifs were required to be expanded (i.e., present at ≥10 reads in each patient as per the read cutoff defined above and encoded by at least two unique nucleotide variants). The sequences were then required to share an identical V region and have the same CDR3 length. Identical J usage was prioritized for determining whether two sequences were highly related, but sequences with different J usage were considered part of the same motif if the above stipulations were met. Finally, any sequences meeting all of the above criteria that were also found in control subjects were excluded.

Two-way ANOVA with the Dunnett multiple-comparisons test correction was performed on iRepertoire V usage data. Two-tailed unpaired t tests were used for D50 comparisons, PBMCs versus BAL cells, and chest x-ray stage comparisons. Pearson scores were determined for correlations between TCR V region usages and BAL CD4/CD8 ratios. The Kruskal–Wallis test with the Dunn posttest was performed for TRAV12-1/TRBV2 and TRAV26-1/TRBV20-1 preferential pairing testing. For analysis of preferential pairing between TRAV12-1/TRBV2 and TRAV26-1/TRBV20-1, the following was calculated: P(A) = percentage of pairs using TRAV12-1 or TRAV26-1 in a sample, P(B) = percentage of pairs using TRBV2 or TRBV20-1 in a sample, and P(AB) = percentage of pairs using TRAV12-1/TRBV2 or TRAV26-1/TRBV20-1 in a sample. If the chains were preferentially pairing, P(AB) > P(A) * P(B); if the chains were pairing by chance, P(AB) = P(A) * P(B). Therefore, to test for preferential pairing, a one-sample t test was used to calculate whether P(AB) − P(A) * P(B) was significantly different from 0 when all six DR3+ LS patients were included. GraphPad Prism v6.07 was used for all statistical analyses. A p value < 0.05 defined statistical significance.

One method for determining αβTCR usage in a semiquantitative manner is iRepertoire PCR, wherein multiplex PCRs separately amplify all Vα or Vβ regions of the TCR. For these analyses, BAL cells were collected from LS, non-LS, and control subjects (Table I), and iRepertoire PCR was performed on sorted CD4+ T cells. To determine the relative diversity of the sequences within each patient, we calculated D50 values for each patient sample. D50 measures the relative diversity of any given sample. Thus, the more diverse a sample, the closer the value will be to 50. For all PBMC samples (regardless of diagnosis), the TCR sequences expressed on CD4+ T cells were significantly more diverse compared with TCRs expressed on BAL CD4+ T cells of DR3+ LS patients (D50 range, 5.3–21.1 and 0.7–6.4, respectively, p = 0.005 for TCRα; D50 range, 3.4–24.4 and 0.8–5.4, respectively, p = 0.025 for TCRβ) (Fig. 1A). Similarly, sarcoidosis patients that were not DR3+ LS patients also had significantly less diverse sequences in their BAL than in the PBMC samples (1.9–11.8 and 1.5–5.1, TCRα and TCRβ, respectively, p < 0.05 for both versus PBMCs) (Fig. 1A). BAL from two fibrosis patients also had lower TCRα and TCRβ diversity compared with PBMCs (data not shown).

Table I.
Characteristics of sarcoidosis and control patients
LS Patients (n = 9)Non-LS Patients (n = 4)Control Patients (n = 3)
Gender (n; male/female) 9/0 3/1 2/1 
Age (y) 36 (33–42) 31 (29–44) 66 (46–68) 
Chest radiographic stage 0/I/II/III/IVa (n0/6/3/0/0 0/0/2/0/1b N/A 
Smoking status (n; nonsmoker/former/current) 3/4/2 2/1/0b 0/2/1 
Vital capacity (% of predicted) 98.0 (77.8–102.8) 79.0 (75.5–83.0) 94.0 (91.5–96.5)c 
DLCO (% of predicted) 102.0 (97.0–104.0) 76.0 (75.0–77.0) 87.0 (87.0–87.0)c 
FEV1 (% of predicted) 85.0 (82.0–92.3) 68.0 (56.0–75.5) 95.0 (93.0–98.5) 
BAL cell concentration (106 cells per liter) 278.1 (198.4–472.3) 284.2 (282.3–389.4) 316.7 (249.0–704.9) 
BAL macrophages (%) 79.3 (61.5–85.4) 66.6 (48.5–72.5) 88.3 (52.3–91.0) 
BAL lymphocytes (%) 16.4 (12.0–36.5) 29.3 (24.8–47.5) 10.6 (8.2–46.4) 
BAL neutrophils (%) 1.0 (0.8–2.2) 4.1 (2.6–4.1) 0.6 (0.3–0.9) 
BAL CD4/CD8 ratio 13.1 (6.3–16.8) 6.0 (5.5–9.7) 1.5 (1.5–4.1) 
HLA-DRB1*03+/DRB1*03DRB3*01+/DRB1*03DRB3*01 (n7/2/0 1/2/1 1/1/1 
LS Patients (n = 9)Non-LS Patients (n = 4)Control Patients (n = 3)
Gender (n; male/female) 9/0 3/1 2/1 
Age (y) 36 (33–42) 31 (29–44) 66 (46–68) 
Chest radiographic stage 0/I/II/III/IVa (n0/6/3/0/0 0/0/2/0/1b N/A 
Smoking status (n; nonsmoker/former/current) 3/4/2 2/1/0b 0/2/1 
Vital capacity (% of predicted) 98.0 (77.8–102.8) 79.0 (75.5–83.0) 94.0 (91.5–96.5)c 
DLCO (% of predicted) 102.0 (97.0–104.0) 76.0 (75.0–77.0) 87.0 (87.0–87.0)c 
FEV1 (% of predicted) 85.0 (82.0–92.3) 68.0 (56.0–75.5) 95.0 (93.0–98.5) 
BAL cell concentration (106 cells per liter) 278.1 (198.4–472.3) 284.2 (282.3–389.4) 316.7 (249.0–704.9) 
BAL macrophages (%) 79.3 (61.5–85.4) 66.6 (48.5–72.5) 88.3 (52.3–91.0) 
BAL lymphocytes (%) 16.4 (12.0–36.5) 29.3 (24.8–47.5) 10.6 (8.2–46.4) 
BAL neutrophils (%) 1.0 (0.8–2.2) 4.1 (2.6–4.1) 0.6 (0.3–0.9) 
BAL CD4/CD8 ratio 13.1 (6.3–16.8) 6.0 (5.5–9.7) 1.5 (1.5–4.1) 
HLA-DRB1*03+/DRB1*03DRB3*01+/DRB1*03DRB3*01 (n7/2/0 1/2/1 1/1/1 

All percentage values represent median (25th–75th percentile). Controls include one healthy and two (nonsarcoidosis) fibrotic lung disease patients.

a

Stage 0 = normal chest radiography, stage I = enlarged lymph nodes, stage II = enlarged lymph nodes with parenchymal infiltrates, stage III = parenchymal infiltrates without enlarged lymph nodes, and stage IV = signs of pulmonary fibrosis.

b

Unknown for one patient.

c

N/A for healthy control patient.

DLCO, CO diffusing capacity; FEV1, forced expiratory volume in 1 s.

FIGURE 1.

TCR sequence diversity in CD4+ T cells in the BAL and blood of LS, sarcoidosis, and control subjects, as determined by iRepertoire PCR. (A) D50 values, measuring sequence diversity (with numbers closer to 50 indicating more diverse samples), were calculated for each patient’s TCRα and TCRβ sequences derived from BAL CD4+ T cells. CD4+ T cells isolated from PBMCs were obtained from a subset of control and sarcoidosis/LS patients and used as a comparison. Horizontal lines indicate mean ± SD. The number of identical BAL CD4+ T cell CDR3s shared in at least two patients exclusively within each patient group is depicted in Venn diagrams for TCRα (B) and TCRβ (C) (left panels). The bar graphs (right panels) display the number of identical CDR3s shared in more than two, three, four, and five DR3+ LS or other sarcoidosis patients. **p < 0.01, *p < 0.05, two-tailed unpaired t test.

FIGURE 1.

TCR sequence diversity in CD4+ T cells in the BAL and blood of LS, sarcoidosis, and control subjects, as determined by iRepertoire PCR. (A) D50 values, measuring sequence diversity (with numbers closer to 50 indicating more diverse samples), were calculated for each patient’s TCRα and TCRβ sequences derived from BAL CD4+ T cells. CD4+ T cells isolated from PBMCs were obtained from a subset of control and sarcoidosis/LS patients and used as a comparison. Horizontal lines indicate mean ± SD. The number of identical BAL CD4+ T cell CDR3s shared in at least two patients exclusively within each patient group is depicted in Venn diagrams for TCRα (B) and TCRβ (C) (left panels). The bar graphs (right panels) display the number of identical CDR3s shared in more than two, three, four, and five DR3+ LS or other sarcoidosis patients. **p < 0.01, *p < 0.05, two-tailed unpaired t test.

Close modal

To analyze the extent of CDR3 homology within subject groups, we used a CDR3 sharing metric from iRweb. This tool does not take into account V or J usage but rather identifies identical CDR3 motifs in any given patient group. As seen in Fig. 1B, the CD4+ T cell compartment in the BAL of DR3+ LS and sarcoidosis patients had high numbers of CDR3α-chains exclusively present in each group when assessing sequences shared in at least two patients within the group (801 and 677 CDR3s, respectively). However, BAL CD4+ T cells from DR3+ LS patients shared more identical CDR3α-chains than sarcoidosis patients when the number of patients sharing the CDR3s was increased to at least three, at least four, or at least five patients (Fig. 1B, right panel). Similarly, a smaller number of CDR3β-chains was exclusively shared in DR3+ LS or sarcoidosis patients (277 and 227, respectively) (Fig. 1C). As was the case with CDR3α-chains, more identical CDR3β-chains were shared in at least three, at least four, and at least five DR3+ LS patients than sarcoidosis patients (Fig. 1C, right panel). In fact, no CDR3β-chains were shared in at least four sarcoidosis patients. Overall, these data suggest that BAL CD4+ T cells from DR3+ LS patients have a more restricted TCR repertoire compared with those cells in the circulation and that CD4+ T cells in the BAL of these patients contain a high proportion of shared CDR3α and CDR3β sequences.

Although CDR3 homology can indicate similar Ag specificity of TCRs (20, 21), we sought to determine V region usage on BAL CD4+ T cells in an unbiased fashion at the gene level to determine whether TCRs expressed on BAL CD4+ T cells of DR3+ LS patients also shared V region homology. TRAV12-1 expansions have previously been associated with DR3+ LS patients (57, 10); after determining overall TRAV usage in each patient group, significantly increased expression of TRAV12-1 was observed in DR3+ LS patients versus control patients (17.3 ± 2.5 versus 2.5 ± 0.6%, p < 0.001, Fig. 2A). Two DRB3+ LS patients also had an increased percentage of BAL CD4+ T cells expressing TRAV12-1. Non-LS sarcoidosis patients did not have significantly altered TRAV12-1 expression on BAL CD4+ T cells, and no other TRAV expansion was seen for any patient group versus control subjects (Fig. 2A, data not shown). However, in all patients (regardless of HLA genotype or diagnosis), >10% of the CD4+ T cells expressed TRAV26-1. Significantly increased expression of TRBV2 on BAL CD4+ T cells was seen exclusively in DR3+ LS patients relative to controls (12.8 ± 1.1 versus 2.9 ± 0.7%, p < 0.001, Fig. 2B), as has been previously reported (13, 15). DRB3+ LS patients did not have increased expression of TRBV2 on CD4+ T cells in the BAL, and increased expression of other TRBV segments was not seen in any patient group relative to control subjects (Fig. 2B, data not shown).

FIGURE 2.

TCR Vα and Vβ usage by BAL CD4+ T cells and correlations with BAL CD4/CD8 ratio and chest x-ray stage. Usage of TCR Vα (A) or Vβ (B) segments in four patient groups is shown, with LS patients divided by HLA genotype (DR3+ or DRB3+). Numbers in parentheses indicate the number of patients in each group. TRAV12-1 and TRBV2 usage was plotted versus BAL CD4/CD8 ratio, as determined by flow cytometry in DR3+ LS patients (C) and other sarcoidosis patients and controls (D). (E) Patients were separated based on chest x-ray stages, and the percentage of CD4+ T cells in the BAL using TRAV12-1 or TRBV2 in DR3+ LS patients was compared with control patient samples. Bars represent mean ± SD. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, two-way ANOVA versus the control patient samples with correction for multiple comparisons (A and B), Pearson scores (C and D), or two-tailed unpaired t tests (E).

FIGURE 2.

TCR Vα and Vβ usage by BAL CD4+ T cells and correlations with BAL CD4/CD8 ratio and chest x-ray stage. Usage of TCR Vα (A) or Vβ (B) segments in four patient groups is shown, with LS patients divided by HLA genotype (DR3+ or DRB3+). Numbers in parentheses indicate the number of patients in each group. TRAV12-1 and TRBV2 usage was plotted versus BAL CD4/CD8 ratio, as determined by flow cytometry in DR3+ LS patients (C) and other sarcoidosis patients and controls (D). (E) Patients were separated based on chest x-ray stages, and the percentage of CD4+ T cells in the BAL using TRAV12-1 or TRBV2 in DR3+ LS patients was compared with control patient samples. Bars represent mean ± SD. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, two-way ANOVA versus the control patient samples with correction for multiple comparisons (A and B), Pearson scores (C and D), or two-tailed unpaired t tests (E).

Close modal

PBMCs were available from a subset of sarcoidosis and control subjects, and CD4+ T cells were isolated and evaluated as described above for TRAV and TRBV usage (Supplemental Fig. 1). Because TRAV12-1 and TRBV2 expansions were seen in the BAL of DR3+ LS patients, we evaluated whether these expansions were significant over blood values. For these evaluations, we separated LS and non-LS patients, because previous work has demonstrated a correlation between LS and the expression of TRAV12-1 and TRBV2 in the BAL (57, 10, 13, 15). In DR3+/DRB3+ LS patients, TRAV12-1 expansions were ∼7.5-fold higher than in blood, and although DR3+ LS patients had significant expansion of TRBV2 in the BAL (Fig. 1), TRBV2 expression in the BAL of DR3+/DRB3+ LS patients was not significantly different compared with blood (Supplemental Fig. 2A, 2C). Additionally, non-LS sarcoidosis and patients with other lung diseases did not have expansion of TRAV12-1 or TRBV2 above blood levels (Supplemental Fig. 2B, 2C). These data support the hypothesis that CD4+ T cells using TRAV12-1 are actively recruited to the BAL of DR3+/DRB3+ LS patients. However, CD4+ T cells using TRBV2 were not significantly expanded in the BAL versus blood when DR3+/DRB3+ LS patients were combined.

Based on the selective expansion of TRAV12-1 and TRBV2 in the BAL of DR3+ LS patients, we focused on these particular α- and β-chains. To determine whether expansion of TRAV12-1 and TRBV2 CD4+ T cells in the BAL of DR3+ LS patients correlated with disease characteristics, expansions were compared with BAL CD4/CD8 ratio, age, and percentage of BAL lymphocytes. A significant correlation was found between TRAV12-1 usage and CD4/CD8 ratio but not between TRBV2 and CD4/CD8 ratio in DR3+ LS patients (r2 = +0.856 and +0.004, respectively, Fig. 2C). There was no significant correlation between TRAV12-1 or TRBV2 in non-LS and control patients and CD4/CD8 ratio (r2 = +0.001 and +0.026, respectively, Fig. 2D). No correlations were found for TRAV12-1 or TRBV2 with age or percentage of BAL lymphocytes (data not shown).

Disease severity of sarcoidosis is categorized by chest radiographic staging at the time of diagnosis, and the samples used for this study came from patients with different stages of disease (Table I). Although there was a significant difference in TRAV12-1 and TRBV2 usage for stage I and stage II DR3+ LS patients versus controls, there was no significant difference between the two stages among these patients (Fig. 2E). No differences were seen for non-LS patients with regard to TRAV12-1 or TRBV2 usage versus control patients for any stage or between stages (data not shown).

Despite the presence of shared CDR3s in each patient group (Fig. 1B, 1C), only DR3+ LS patients had skewing of the TCR repertoire with expansion of CD4+ T cells expressing TRAV12-1 and TRBV2. Therefore, we focused on TRAV12-1 and TRBV2 in these patients to identify disease-specific shared CDR3 motifs (as defined in 2Materials and Methods). Several similar and identical TRAV12-1 CDR3 motifs using TRAJ42 were expanded in DR3+ LS patients (Fig. 3A). For example, in Patients 1079 and 1088, 13 and 14 nucleotide combinations, respectively, were used to generate the amino acids proline (P) and arginine (R). Interestingly, with only one exception (a single sequence in Patient 1229, a healthy DR3+ individual), these TCRα sequences were not present in BAL CD4+ T cells isolated from DR3/DRB3 sarcoidosis patients or other control subjects. Fig. 4 shows a detailed view of the nucleotides used to generate the amino acids of three different TRAV12-1 CDR3 motifs, using the sequences from Patient 1088 as a reference. Similar results were found for each of the motifs shown in Fig. 3A (Fig. 4A–C, data not shown). For reference, the germline-encoded nucleotides are shown for TRAV12-1 and TRAJ42 (Fig. 4D). The sequence diversity occurred in the N region, N-terminal V region, and C-terminal J region, suggesting that these nucleotide differences seen in multiple LS patients were due to Ag selection and precluded the possibility of PCR contamination or artifact.

FIGURE 3.

Shared TRAV12-1 and TRBV2 CDR3 motifs in sarcoidosis patients. A subset of the TRAV12-1 (A) and TRBV2 (B) TCR CDR3 motifs that were shared and expanded in LS and non-LS patients versus control (Other and Healthy) patients. Orange (α) or blue (β) color indicates that the sequence was present in the patient sample, whereas gray indicates that the sequence was not found in the patient sample. Numbers indicate the unique nucleotide variants encoding the amino acid sequence in the specified patient, with darker colors representing higher numbers of nucleotide variants in a patient sample.

FIGURE 3.

Shared TRAV12-1 and TRBV2 CDR3 motifs in sarcoidosis patients. A subset of the TRAV12-1 (A) and TRBV2 (B) TCR CDR3 motifs that were shared and expanded in LS and non-LS patients versus control (Other and Healthy) patients. Orange (α) or blue (β) color indicates that the sequence was present in the patient sample, whereas gray indicates that the sequence was not found in the patient sample. Numbers indicate the unique nucleotide variants encoding the amino acid sequence in the specified patient, with darker colors representing higher numbers of nucleotide variants in a patient sample.

Close modal
FIGURE 4.

Oligoclonal expansion of TRAV12-1 sequences expressed in DR3+ LS patients. Shared and expanded CDR3 motifs in CD4+ T cells purified from the BAL of DR3+ LS patients were determined after iRepertoire PCR and deep sequencing. (AC) The three most expanded motifs that were shared among DR3+ LS patients are shown at the nucleotide level below the deduced amino acid sequences. Variants found in Patient 1088 are shown, and patients with identical sequences are listed in the far right column. Nucleotides highlighted in blue, red, and black are encoded by TRAV12-1, TRAJ42, and nontemplate bases, respectively. (D) Germline-encoded nucleotides are shown as a reference.

FIGURE 4.

Oligoclonal expansion of TRAV12-1 sequences expressed in DR3+ LS patients. Shared and expanded CDR3 motifs in CD4+ T cells purified from the BAL of DR3+ LS patients were determined after iRepertoire PCR and deep sequencing. (AC) The three most expanded motifs that were shared among DR3+ LS patients are shown at the nucleotide level below the deduced amino acid sequences. Variants found in Patient 1088 are shown, and patients with identical sequences are listed in the far right column. Nucleotides highlighted in blue, red, and black are encoded by TRAV12-1, TRAJ42, and nontemplate bases, respectively. (D) Germline-encoded nucleotides are shown as a reference.

Close modal

When searching for shared TRBV2 motifs within sarcoidosis patients, identical CDR3 motifs in multiple DR3+ LS patients were less common than seen for TRAV12-1 (Fig. 3B). Although the TRBV2 sequences of interest were not found in DR3/DRB3 sarcoidosis, control, and even non-LS patients, the expansions were expressed at lower frequencies at the nucleotide level relative to those seen in the TRAV12-1 motifs. However, there were several CDR3β sequences that were closely related and exclusively found in the DR3+ LS patients.

When examining all identical and highly related TRAV12-1 CDR3α and TRBV2 CDR3β motifs exclusively present in BAL CD4+ T cells of DR3+ LS patients, two consensus CDR3α motifs were identified using an online sequence alignment tool (22): CVV-PR-GGSQGNLIF and CVV-NRR-GGSQGNLIF (Fig. 5A). All sequences in the motif used TRAV12-1 (CVV) and TRAJ42 (GGSQGNLIF); however, all sequences had two or three n or p nucleotide additions and all deleted the germline asparagine (N) of TRAV12-1 and the tyrosine (Y) of TRAJ42. Additionally, two consensus CDR3β motifs were deduced from the shared TRBV2 sequences in DR3+ LS patients: CASS-EQGR-EEQFF and CASS-EQGGR-ETQYF (Fig. 5B). Sequences in the TCRβ motif all shared TRBV2 (CASSE) usage, but there were several sequences with deletion of the glutamic acid (E) of TRBV2 and utilization of a glycine (G) at that position. Moreover, most of the TRBV2 motif sequences shared TRBV2-5 (ETQYF) usage, but many of the sequences also had insertions and deletions in the J region, sometimes corresponding to differing J usage. All sequences making up the TRBV2 motifs had three or four n or p nucleotide additions between the V and J regions.

FIGURE 5.

Consensus BAL CD4+ T cell CDR3 motifs shared by DR3+ LS patients. Graphical representations of CDR3α (A) or CDR3β (B) motifs shared among DR3+ LS patients are shown after using a web-based sequence alignment tool (http://weblogo.berkeley.edu). Amino acids are color-coded based on properties, and black designates the conserved C-terminal CVV and CASS, as well as the N-terminal F, amino acids. Letter size reflects the frequency of appearance of a specific residue at a certain position.

FIGURE 5.

Consensus BAL CD4+ T cell CDR3 motifs shared by DR3+ LS patients. Graphical representations of CDR3α (A) or CDR3β (B) motifs shared among DR3+ LS patients are shown after using a web-based sequence alignment tool (http://weblogo.berkeley.edu). Amino acids are color-coded based on properties, and black designates the conserved C-terminal CVV and CASS, as well as the N-terminal F, amino acids. Letter size reflects the frequency of appearance of a specific residue at a certain position.

Close modal

The initial data generated by iRepertoire analyses suggested that the TRAV12-1 and TRBV2 chains might be paired together on lung-accumulated CD4+ T cells in DR3+ LS patients. Although one study showed that a small, but significant, percentage of CD4+ T cells from the BAL of LS patients coexpressed TRAV12-1 and TRBV2 by flow cytometry (15), no previous studies have sequenced and identified complete αβTCR pairs on BAL CD4+ T cells from these patients. Therefore, to address αβTCR pairing on BAL CD4+ T cells, ePCR was performed (19, 23).

Briefly, cells were resuspended in RT-PCR master mix with reagents and primers for amplifying TCRα and β-chain genes; following the addition of an oil mixture and vortexing, the contents were separated into a water-in-oil emulsion, capturing one cell per micelle (Fig. 6A). After three rounds of PCR, single bands with barcoded and joined αβTCR products were gel purified for each patient (Fig. 6B). Previous work using ePCR on hybridoma cells with known TCRs demonstrated that correct pairing of TCRα- and TCRβ-chains occurred with an efficiency of 85%, whereas the percentage of correctly matched pairs dropped to 5–10% when the emulsion step was not performed (19). Indeed, when we compared our emulsion samples with the nonemulsion samples, we saw high percentages of particular αβTCR pairs in the emulsion samples but observed no expansion in the nonemulsion samples (data not shown). However, as has been stated in previous work using this method, ePCR is not directly quantitative, and we likely observed overamplification of the top αβTCR pairs for most emulsion samples. For example, when assessing all samples, an average top value of 25.6% of the CD4+ cells in the BAL samples and 31.9% of the CD4+ cells in the PBMC samples were observed (data not shown) (19). Furthermore, the top αβTCR pair in the blood of one patient represented 76.23% of all TCRs expressed on CD4+ T cells (data not shown). Given those values, the output hierarchy is skewed toward the top αβTCR pairs. Thus, we used these data in a nonquantitative fashion and searched for αβTCR pairs of interest based on those expanded in the iRepertoire data.

FIGURE 6.

Pairing of TRAV12-1+ and TRBV2+ CD4+ T cells, as determined by ePCR. (A) PBMCs were labeled with CFSE, spun, and resuspended in RT-PCR master mix. Three hundred microliters of the oil mixture from the Micellula kit was added on top, and the contents were vortexed for 2.5 min. (B) CD4+ T cells were purified from BAL cells or PBMCs from one patient and used in emulsion (em) or nonemulsion (non) PCRs. A representative 2% agarose gel is shown displaying a DNA ladder (L) and the products obtained after PCR3. (C) Six DR3+ LS BAL samples were subjected to ePCR, and Vβ usage on TRAV12-1–utilizing BAL CD4+ T cells is shown. (D) Vα usage on TRBV2-utilizing BAL CD4+ T cells from six DR3+ LS patients. (E) Thirteen αβTCR pairs, as determined by ePCR. The fifth column (% α paired with β) indicates the percentage that the listed TCRα is paired with the listed TCRβ exclusively, as opposed to pairing with other TCRβ in a given patient sample. The sixth column (% β paired with α) indicates the percentage that the listed TCRβ is paired with the listed TCRα exclusively, as opposed to pairing with other TCRα in a given patient sample. Amino acids forming shared or similar CDR3 motifs are in color. (F) TRAV12-1 usage on TRBV2+ cells, TRBV2 usage on TRAV12-1+ cells, TRAV26-1 usage on TRBV20-1+ cells, and TRBV20-1 usage on TRAV26-1+ cells are shown for comparison among patient groups. Bars represent mean + SD. *p < 0.05, Kruskal–Wallis test with the Dunn posttest.

FIGURE 6.

Pairing of TRAV12-1+ and TRBV2+ CD4+ T cells, as determined by ePCR. (A) PBMCs were labeled with CFSE, spun, and resuspended in RT-PCR master mix. Three hundred microliters of the oil mixture from the Micellula kit was added on top, and the contents were vortexed for 2.5 min. (B) CD4+ T cells were purified from BAL cells or PBMCs from one patient and used in emulsion (em) or nonemulsion (non) PCRs. A representative 2% agarose gel is shown displaying a DNA ladder (L) and the products obtained after PCR3. (C) Six DR3+ LS BAL samples were subjected to ePCR, and Vβ usage on TRAV12-1–utilizing BAL CD4+ T cells is shown. (D) Vα usage on TRBV2-utilizing BAL CD4+ T cells from six DR3+ LS patients. (E) Thirteen αβTCR pairs, as determined by ePCR. The fifth column (% α paired with β) indicates the percentage that the listed TCRα is paired with the listed TCRβ exclusively, as opposed to pairing with other TCRβ in a given patient sample. The sixth column (% β paired with α) indicates the percentage that the listed TCRβ is paired with the listed TCRα exclusively, as opposed to pairing with other TCRα in a given patient sample. Amino acids forming shared or similar CDR3 motifs are in color. (F) TRAV12-1 usage on TRBV2+ cells, TRBV2 usage on TRAV12-1+ cells, TRAV26-1 usage on TRBV20-1+ cells, and TRBV20-1 usage on TRAV26-1+ cells are shown for comparison among patient groups. Bars represent mean + SD. *p < 0.05, Kruskal–Wallis test with the Dunn posttest.

Close modal

When analyzing TCRBV usage on BAL TRAV12-1+ CD4+ T cells from DR3+ LS patients, ∼30% of those cells coexpressed TRBV2 (Fig. 6C). When assessing BAL TRBV2 CD4+ T cells from these patients, ∼40% expressed TRAV12-1 (Fig. 6D). Using the shared TRAV12-1 and TRBV2 CDR3 motifs from Figs. 3 to 5 as references, we searched for αβTCR pairs containing identical or similar CDR3 regions. We identified 13 pairs with highly related CDR3α and CDR3β sequences that were found in multiple DR3+ LS patients (Fig. 6E). In many cases, the pairing was exclusive (i.e., a particular TCRα-chain paired with the listed TCRβ-chain at a high frequency relative to pairing with other TCRβ-chains, and vice versa), as shown in pairs 5, 6, and 13. However, several of the pairs were less exclusive on the TCRα side, TCRβ side, or both. For example, the αβTCR pair designated as clone 5 in Fig. 6E was exclusively found in the BAL of four DR3+ LS patients. Importantly, different nucleotides were used in these four patients to generate identical amino acid sequences (data not shown), suggesting the presence of a public TCR repertoire. Although other shared αβTCRs were found exclusively in DR3+ LS patients, the sequences did not use TRAV12-1/TRBV2 or did not share CDR3 homology in either chain (data not shown).

The higher percentage of TRAV12-1/TRBV2 pairing could possibly be due to random chance, because the relative abundance of each chain in these patients is high. Only two V regions were highly expressed in most BAL samples, regardless of diagnosis: TRAV26-1 and TRBV20-1 (Fig. 2A, 2B). Therefore, pairing between TRAV26-1 and TRBV20-1 was interrogated for all patient groups to determine whether highly expressed V regions preferentially pair together simply due to chance. TRAV26-1 and TRBV20-1 did not pair together at a high proportion in any patient group, as was seen between TRAV12-1 and TRBV2 exclusively in DR3+ LS patients (Fig. 6F). Using the preferential pairing equation described in 2Materials and Methods, the averaged value for TRAV12-1/TRBV2 pairing was significantly different from 0 (0.066 ± 0.059, p = 0.04), whereas the value for TRAV26-1/TRBV20-1 was not different from 0 (−0.0105 ± 0.0257, p = 0.36). These data suggest that TRAV12-1 and TRBV2 are preferentially paired on BAL CD4+ T cells in DR3+ LS patients, likely due to Ag selection.

To identify αβTCR pairs in a more quantitative fashion, as well as to validate the αβTCR pairs obtained by ePCR, we performed scPCR on DR3+ LS Patient 1244. BAL CD4+ T cells expressing TRAV12-1 were single-cell sorted into 96-well plates (Fig. 7A), cDNA was generated, and three rounds of PCR were performed (18). We obtained 86 complete αβTCR pairs. Similar to the ePCR-derived sequences in Fig. 6C, TRBV2 was the most prevalent TCRβ-chain paired with TRAV12-1 (Fig. 7B).

FIGURE 7.

αβTCR pairing on CD4+ T cells from the BAL of one DR3+ LS patient, as determined by single-cell PCR. (A) BAL CD4+ T cells were isolated from one DR3+ LS patient BAL sample and sorted by flow cytometry into 96-well plates. Cells were sequentially gated on lymphocytes (data not shown), CD3+DAPI, and CD4+TRAV12-1+, which were subsequently sorted into 96-well plates at one cell per well for scPCR. (B) After subjecting the cells to single-cell PCR and deep-sequencing analysis, the Vβ usage on TRAV12-1+ cells was determined for each well. (C) Cells expressing αβTCR pairs with shared or similar CDR3 motifs are shown as individual pairs, and the number of wells in which the pairs were found are shown; shared or similar CDR3 motifs are in color. (D) An online multiple sequence alignment tool (http://weblogo.berkeley.edu) was used to create a graphical representation of a consensus CDR3β motif found in multiple wells. Amino acids are color-coded based on chemical properties, and black designates the conserved C-terminal CASS, as well as the N-terminal F, amino acids. Letter size reflects the frequency of appearance of a specific residue at a certain position.

FIGURE 7.

αβTCR pairing on CD4+ T cells from the BAL of one DR3+ LS patient, as determined by single-cell PCR. (A) BAL CD4+ T cells were isolated from one DR3+ LS patient BAL sample and sorted by flow cytometry into 96-well plates. Cells were sequentially gated on lymphocytes (data not shown), CD3+DAPI, and CD4+TRAV12-1+, which were subsequently sorted into 96-well plates at one cell per well for scPCR. (B) After subjecting the cells to single-cell PCR and deep-sequencing analysis, the Vβ usage on TRAV12-1+ cells was determined for each well. (C) Cells expressing αβTCR pairs with shared or similar CDR3 motifs are shown as individual pairs, and the number of wells in which the pairs were found are shown; shared or similar CDR3 motifs are in color. (D) An online multiple sequence alignment tool (http://weblogo.berkeley.edu) was used to create a graphical representation of a consensus CDR3β motif found in multiple wells. Amino acids are color-coded based on chemical properties, and black designates the conserved C-terminal CASS, as well as the N-terminal F, amino acids. Letter size reflects the frequency of appearance of a specific residue at a certain position.

Close modal

We next assessed CDR3β homology among the TRBV2 sequences paired with TRAV12-1-expressing cells (Fig. 7C). Although pairs with similar CDR3 sequences in TCRα and TCRβ (e.g., pairs 2/3 and 7/8) were observed, the TRAV12-1 CDR3 regions were heterogeneous and used different TRAJ regions. However, the dominant TRBV2 CDR3β motif found in Patient 1244 (CASS-EGSR-GTAFF) (Fig. 7D) resembled the motif identified in the seven other DR3+ LS patients (CASS-EQGR-EEQFF) (summarized in Fig. 5B). Taken together, these results suggest a critical role for the TRAV12-1/TRBV2 TCRs in recognizing the DR3–peptide complex in LS patients.

CD4+ T cells accumulate in the lungs of DR3+ LS patients during active disease and consist of oligoclonal TCR expansion, suggesting that these T cells are recruited to the lung in response to conventional Ag stimulation (5, 12, 14, 24). In LS patients, the predominant TCRα-chain expressed on CD4+ T cells in BAL is TRAV12-1, and a strong correlation exists between DR3 and the expression of TRAV12-1 on BAL CD4+ T cells (57, 10). Additionally, expansion of several TCR β-chains in distinct sarcoidosis populations has been described (6, 1115). However, rarely have related TCRα- or β-chains been identified among multiple sarcoidosis patients, and no study has successfully characterized complete αβTCR pairs expressed on T cells derived from the lungs of multiple sarcoidosis patients. Using novel deep-sequencing approaches, we show that TRAV12-1 preferentially pairs with TRBV2 and identify public TRAV12-1/TRBV2 TCRs that likely play a critical role in Ag recognition in the lungs of LS patients.

CDR3α motifs have previously been identified in BAL CD4+ T cells from sarcoidosis patients, some of which resembled the CVV-RY-GGSQGNLIF and CVV-IGH-GGSQGNLIF sequences that we identified in multiple DR3+ LS patients, including CVV-DY-GGSQGN (24) and CVV-IGS-GGSQGNLIF (15). TCRβ sequences derived from skin biopsies of Kveim reactions (25) and from BAL CD4+ T cells after IL-2 culture (12) have not revealed shared or consensus CDR3β motifs. However, a previous study showed several TRBV2 motifs in the BAL of DR3+ LS patients (15); although the CDR3β motifs were fairly heterogeneous among patients, two individual sequences (CASS-EQGR-GETQYF and CASS-GPGGR-TEAFF) were found to be identical to or resemble the consensus TRBV2 sequences identified in multiple patients in the current study. Collectively, our data delineate several CDR3α and CDR3β motifs that are shared in the majority of the DR3+ LS patients in this study, including consensus CDR3α motifs (CVV-PR-GGSQGNLIF, CVV-NRR-GGSQGNLIF) and CDR3β motifs (CASS-EQGR-EEQFF, CASS-EQGGR-ETQYF). Although there were numerous motifs shared across patient groups, we specifically focused on motifs found only in LS DR3+ patients using TRAV12-1 and TRBV2. Although the pairs included in the determination of the CDR3 motifs were never found in control subjects or non-LS patients, one caveat to the study is that TCR motifs using other V regions may have been overlooked. Therefore, future studies are necessary to determine the specificity of the identified TCRs using TRAV12-1 and TRBV2.

Public T cells are characterized by the expression of identical or highly related TCR Vα and/or Vβ genes that are present in the majority of subjects, dominate the response to a specific epitope, and dictate disease severity, despite being restricted in nature (2633). Importantly, this type of TCR bias has been infrequently demonstrated in CD4+ T cells obtained from blood or the target organ of human subjects due, in large part, to unknown stimulatory Ags. We have previously characterized αβTCR pairs derived from the lungs of HLA-DP2–expressing chronic beryllium disease (CBD) patients that recognize an identical HLA-DP2–peptide/Be complex (21). In addition, the inverse relationship between expansion of CD4+ T cells expressing these public TCRs and disease severity suggests a pathogenic role for these T cells in CBD (21). However, increased expression of TRAV12-1 on CD4+ T cells in the BAL of LS patients has been associated with a better prognosis, with the disappearance of these public CD4+ T cells from the lung of LS patients with disease resolution (5, 10). Thus, the differences between public T cells in LS and CBD likely relate to Ag clearance from the lung. For example, in CBD, public T cells may drive loss of lung function and fibrosis as the result of the persistence of beryllium within the lung, whereas in LS, public T cells may be intimately involved in the clearance of an as-yet unknown Ag.

The public TCR chains expressing TRBV5-1 in CBD patients paired with different TRAV chains to generate beryllium-specific TCRs that recognized the same antigenic epitope (20, 34). Using site-specific mutagenesis along the CDRs of the TCRα and TCRβ genes, T cell recognition was dependent on several TCRβ residues in CDR1, CDR2, and CDR3, whereas only 1 aa of the CDR3α region was required for beryllium recognition. Given the nearly identical TCRβ-chains shared by LS patients, the data presented in this article might support a similar model to that observed in CBD; recognition of a sarcoidosis-associated Ag presented by DR3 could be dependent on the TCRβ CDR1, CDR2, and CDR3 regions. The role for the TCRα-chains on BAL CD4+ T cells with differing CDR3 sequences in Ag recognition might be minimal, and perhaps the CDR1 and CDR2 regions of the TCRα-chain are contacting the DR3 molecule more directly, driving the linkage between TRAV12-1–expressing CD4+ T cells and DR3. A recent publication depicted a three-dimensional computer-generated model of a theoretical TRAV12-1/TRBV2 TCR in complex with DR3–vimentin (15). The model predicted that specific residues on the DR3 molecule could directly interact with TCRα residues on the CDR1α and CDR2α loops and that TRBV2 residues could be important for DR3 recognition and for linkage to the TCRα-chain. Although the model predicted a binding topology that agrees with our current speculations, it is important to state that the Ags recognized by these TCRs and the mechanisms by which these TCRs bind to the DR3–peptide complex remain unknown.

We have known about the link between LS and the accumulation of CD4+ T cells expressing TRAV12-1 for >20 y (5). However, progress toward the delineation of the underlying mechanism(s) responsible for the recruitment of these cells to the lung has been hampered by an inability to generate large numbers of αβTCR-expressing T cell clones. Our ability to detect public TCRs in LS was aided by the use of deep-sequencing technology, including ePCR and scPCR. These approaches have several advantages, including their ability to determine αβTCR pairs on BAL CD4+ T cells directly ex vivo and without in vitro culture expansion. This is particularly important in a target organ in which the majority of T cells are terminally differentiated and, thus, incapable of vigorously undergoing repeated rounds of stimulation (35). Conversely, T cell cloning is an ineffective approach for obtaining large numbers of clones expressing unique αβTCRs, predominantly selecting those T cells that retain proliferative capacity and, thus, biasing the TCR repertoire. The current approach, in particular, scPCR, obviates the need for T cell cloning and can be used to delineate the most expanded αβTCRs expressed on T cells derived from any target organ. Disadvantages of this approach are expense and the requirement for a biased or skewed TCR repertoire. Thus, this deep-sequencing approach will likely not be applicable to the analysis of a diverse TCR repertoire, such as exists in blood.

We acknowledge that a weakness of our study is the small number of patient samples. This likely accounted for our inability to identify a correlation between CD4/CD8 ratio and the expression of TRBV2 on BAL CD4+ T cells in LS patients, as previously shown by Grunewald et al. (7). However, as we show in this article, large numbers of patients are not required when using deep-sequencing αβTCR repertoire methodology because of the depth of the analysis.

In conclusion, our findings identify public αβTCRs in the BAL of LS patients and strongly suggest that these public TCRs recognize putative sarcoidosis Ags and drive disease pathogenesis. Importantly, the delineation of complete αβTCR pairs with a distinct association with HLA and prognosis will enable future studies on Ag recognition.

We thank research nurses Gunnel de Forest, Margitha Dahl, and Heléne Blomqvist and biomedical analyst Benita Dahlberg (Respiratory Medicine Unit, Karolinska University Hospital, Solna, Sweden) for skillful assistance in sample preparation and processing.

This work was supported by National Heart, Lung, and Blood Institute/National Institutes of Health Grants HL062410 and HL102245 (to A.P.F.), the Swedish Heart Lung Foundation, the Swedish Research Council, the Mats Kleberg Foundation, the King Oscar II Jubilee Foundation, King Gustaf V’s and Queen Victoria’s Freemasons’ Foundation, the regional agreement on medical training and clinical research between Stockholm County Council and Karolinska Institutet, and the Karolinska Institutet (to J.G.).

The TCR sequences reported in this article have been submitted to the Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo) under accession number GSE100378.

The online version of this article contains supplemental material.

Abbreviations used in this article:

     
  • BAL

    bronchoalveolar lavage

  •  
  • CBD

    chronic beryllium disease

  •  
  • DR3

    HLA-DRB1*03:01

  •  
  • DRB3

    HLA-DRB3*01:01

  •  
  • ePCR

    emulsion PCR

  •  
  • LS

    Löfgren’s syndrome

  •  
  • scPCR

    single-cell PCR.

1
Newman
,
L. S.
,
C. S.
Rose
,
L. A.
Maier
.
1997
.
Sarcoidosis.
N. Engl. J. Med.
336
:
1224
1234
.
2
Grunewald
,
J.
2007
.
Clinical aspects and immune reactions in sarcoidosis.
Clin. Respir. J.
1
:
64
73
.
3
Judson
,
M. A.
2015
.
The clinical features of sarcoidosis: a comprehensive review.
Clin. Rev. Allergy Immunol.
49
:
63
78
.
4
Grunewald
,
J.
2012
.
HLA associations and Löfgren’s syndrome.
Expert Rev. Clin. Immunol.
8
:
55
62
.
5
Grunewald
,
J.
,
C. H.
Janson
,
A.
Eklund
,
M.
Ohrn
,
O.
Olerup
,
U.
Persson
,
H.
Wigzell
.
1992
.
Restricted V alpha 2.3 gene usage by CD4+ T lymphocytes in bronchoalveolar lavage fluid from sarcoidosis patients correlates with HLA-DR3.
Eur. J. Immunol.
22
:
129
135
.
6
Grunewald
,
J.
,
O.
Olerup
,
U.
Persson
,
M. B.
Ohrn
,
H.
Wigzell
,
A.
Eklund
.
1994
.
T-cell receptor variable region gene usage by CD4+ and CD8+ T cells in bronchoalveolar lavage fluid and peripheral blood of sarcoidosis patients.
Proc. Natl. Acad. Sci. USA
91
:
4965
4969
.
7
Grunewald
,
J.
,
M.
Berlin
,
O.
Olerup
,
A.
Eklund
.
2000
.
Lung T-helper cells expressing T-cell receptor AV2S3 associate with clinical features of pulmonary sarcoidosis.
Am. J. Respir. Crit. Care Med.
161
:
814
818
.
8
Nagvekar
,
N.
,
L.
Corlett
,
L. W.
Jacobson
,
H.
Matsuo
,
R.
Chalkley
,
P. C.
Driscoll
,
S.
Deshpande
,
E. G.
Spack
,
N.
Willcox
.
1999
.
Scanning a DRB3*0101 (DR52a)-restricted epitope cross-presented by DR3: overlapping natural and artificial determinants in the human acetylcholine receptor.
J. Immunol.
162
:
4079
4087
.
9
Grunewald
,
J.
,
J.
Wahlström
,
M.
Berlin
,
H.
Wigzell
,
A.
Eklund
,
O.
Olerup
.
2002
.
Lung restricted T cell receptor AV2S3+ CD4+ T cell expansions in sarcoidosis patients with a shared HLA-DRbeta chain conformation.
Thorax
57
:
348
352
.
10
Planck
,
A.
,
A.
Eklund
,
J.
Grunewald
.
2003
.
Markers of activity in clinically recovered human leukocyte antigen-DR17-positive sarcoidosis patients.
Eur. Respir. J.
21
:
52
57
.
11
Forman
,
J. D.
,
J. T.
Klein
,
R. F.
Silver
,
M. C.
Liu
,
B. M.
Greenlee
,
D. R.
Moller
.
1994
.
Selective activation and accumulation of oligoclonal V beta-specific T cells in active pulmonary sarcoidosis.
J. Clin. Invest.
94
:
1533
1542
.
12
Forrester
,
J. M.
,
Y.
Wang
,
N.
Ricalton
,
J. E.
Fitzgerald
,
J.
Loveless
,
L. S.
Newman
,
T. E.
King
,
B. L.
Kotzin
.
1994
.
TCR expression of activated T cell clones in the lungs of patients with pulmonary sarcoidosis.
J. Immunol.
153
:
4291
4302
.
13
Ahlgren
,
K. M.
,
T.
Ruckdeschel
,
A.
Eklund
,
J.
Wahlström
,
J.
Grunewald
.
2014
.
T cell receptor-Vβ repertoires in lung and blood CD4+ and CD8+ T cells of pulmonary sarcoidosis patients.
BMC Pulm. Med.
14
:
50
.
14
Moller
,
D. R.
,
K.
Konishi
,
M.
Kirby
,
B.
Balbi
,
R. G.
Crystal
.
1988
.
Bias toward use of a specific T cell receptor beta-chain variable region in a subgroup of individuals with sarcoidosis.
J. Clin. Invest.
82
:
1183
1191
.
15
Grunewald
,
J.
,
Y.
Kaiser
,
M.
Ostadkarampour
,
N. V.
Rivera
,
F.
Vezzi
,
B.
Lötstedt
,
R.A.
Olsen
,
L.
Sylwan
,
S.
Lundin
,
M.
Käller
, et al
.
2016
.
T-cell receptor-HLA-DRB1 associations suggest specific antigens in pulmonary sarcoidosis.
Eur. Respir. J.
47
:
898
909
.
16
Hunninghake
,
G. W.
,
U.
Costabel
,
M.
Ando
,
R.
Baughman
,
J. F.
Cordier
,
R.
du Bois
,
A.
Eklund
,
M.
Kitaichi
,
J.
Lynch
,
G.
Rizzato
, et al
.
1999
.
ATS/ERS/WASOG statement on sarcoidosis. American thoracic society/European respiratory society/world association of sarcoidosis and other granulomatous disorders.
Sarcoidosis, Vasc. Diffus. lung Dis.
16
:
149
173
.
17
Olsen
,
H. H.
,
J.
Grunewald
,
G.
Tornling
,
C. M.
Sköld
,
A.
Eklund
.
2012
.
Bronchoalveolar lavage results are independent of season, age, gender and collection site.
PLoS One
7
:
e43644
.
18
Michels
,
A. W.
,
L. G.
Landry
,
K. A.
McDaniel
,
L.
Yu
,
M.
Campbell-Thompson
,
W. W.
Kwok
,
K. L.
Jones
,
P. A.
Gottlieb
,
J. W.
Kappler
,
Q.
Tang
, et al
.
2017
.
Islet-derived CD4 T cells targeting proinsulin in human autoimmune diabetes.
Diabetes
66
:
722
734
.
19
Munson
,
D. J.
,
C. A.
Egelston
,
K. E.
Chiotti
,
Z. E.
Parra
,
T. C.
Bruno
,
B. L.
Moore
,
T. A.
Nakano
,
D. L.
Simons
,
G.
Jimenez
,
J. H.
Yim
, et al
.
2016
.
Identification of shared TCR sequences from T cells in human breast cancer using emulsion RT-PCR.
Proc. Natl. Acad. Sci. USA
113
:
8272
8277
.
20
Bowerman
,
N. A.
,
M. T.
Falta
,
D. G.
Mack
,
J. W.
Kappler
,
A. P.
Fontenot
.
2011
.
Mutagenesis of beryllium-specific TCRs suggests an unusual binding topology for antigen recognition.
J. Immunol.
187
:
3694
3703
.
21
Bowerman
,
N. A.
,
M. T.
Falta
,
D. G.
Mack
,
F.
Wehrmann
,
F.
Crawford
,
M. M.
Mroz
,
L. A.
Maier
,
J. W.
Kappler
,
A. P.
Fontenot
.
2014
.
Identification of multiple public TCR repertoires in chronic beryllium disease.
J. Immunol.
192
:
4571
4580
.
22
Crooks
,
G. E.
,
G.
Hon
,
J.M.
Chandonia
,
S. E.
Brenner
.
2004
.
WebLogo: a sequence logo generator.
Genome Res.
14
:
1188
1190
.
23
Turchaninova
,
M. A.
,
O. V.
Britanova
,
D. A.
Bolotin
,
M.
Shugay
,
E. V.
Putintseva
,
D. B.
Staroverov
,
G.
Sharonov
,
D.
Shcherbo
,
I. V.
Zvyagin
,
I. Z.
Mamedov
, et al
.
2013
.
Pairing of T-cell receptor chains via emulsion PCR.
Eur. J. Immunol.
43
:
2507
2515
.
24
Grunewald
,
J.
,
T.
Hultman
,
A.
Bucht
,
A.
Eklund
,
H.
Wigzell
.
1995
.
Restricted usage of T cell receptor V alpha/J alpha gene segments with different nucleotide but identical amino acid sequences in HLA-DR3+ sarcoidosis patients.
Mol. Med.
1
:
287
296
.
25
Klein
,
J. T.
,
T. D.
Horn
,
J. D.
Forman
,
R. F.
Silver
,
A. S.
Teirstein
,
D. R.
Moller
.
1995
.
Selection of oligoclonal V beta-specific T cells in the intradermal response to Kveim-Siltzbach reagent in individuals with sarcoidosis.
J. Immunol.
154
:
1450
1460
.
26
Venturi
,
V.
,
D. A.
Price
,
D. C.
Douek
,
M. P.
Davenport
.
2008
.
The molecular basis for public T-cell responses?
Nat. Rev. Immunol.
8
:
231
238
.
27
Turner
,
S. J.
,
P. C.
Doherty
,
J.
McCluskey
,
J.
Rossjohn
.
2006
.
Structural determinants of T-cell receptor bias in immunity.
Nat. Rev. Immunol.
6
:
883
894
.
28
Li
,
H.
,
C.
Ye
,
G.
Ji
,
J.
Han
.
2012
.
Determinants of public T cell responses.
Cell Res.
22
:
33
42
.
29
Wang
,
G. C.
,
P.
Dash
,
J. A.
McCullers
,
P. C.
Doherty
,
P. G.
Thomas
.
2012
.
T cell receptor αβ diversity inversely correlates with pathogen-specific antibody levels in human cytomegalovirus infection.
Sci. Transl. Med.
4
:
128ra42
.
30
Luo
,
W.
,
J.
Su
,
X.B.
Zhang
,
Z.
Yang
,
M.Q.
Zhou
,
Z.M.
Jiang
,
P.P.
Hao
,
S.D.
Liu
,
Q.
Wen
,
Q.
Jin
,
L.
Ma
.
2012
.
Limited T cell receptor repertoire diversity in tuberculosis patients correlates with clinical severity.
PLoS One
7
:
e48117
.
31
Frahm
,
N.
,
P.
Kiepiela
,
S.
Adams
,
C. H.
Linde
,
H. S.
Hewitt
,
K.
Sango
,
M. E.
Feeney
,
M. M.
Addo
,
M.
Lichterfeld
,
M. P.
Lahaie
, et al
.
2006
.
Control of human immunodeficiency virus replication by cytotoxic T lymphocytes targeting subdominant epitopes.
Nat. Immunol.
7
:
173
178
.
32
Ruckwardt
,
T. J.
,
C.
Luongo
,
A. M.
Malloy
,
J.
Liu
,
M.
Chen
,
P. L.
Collins
,
B. S.
Graham
.
2010
.
Responses against a subdominant CD8+ T cell epitope protect against immunopathology caused by a dominant epitope.
J. Immunol.
185
:
4673
4680
.
33
Billam
,
P.
,
K. L.
Bonaparte
,
J.
Liu
,
T. J.
Ruckwardt
,
M.
Chen
,
A. B.
Ryder
,
R.
Wang
,
P.
Dash
,
P. G.
Thomas
,
B. S.
Graham
.
2011
.
T cell receptor clonotype influences epitope hierarchy in the CD8+ T cell response to respiratory syncytial virus infection.
J. Biol. Chem.
286
:
4829
4841
.
34
Clayton
,
G. M.
,
Y.
Wang
,
F.
Crawford
,
A.
Novikov
,
B. T.
Wimberly
,
J. S.
Kieft
,
M. T.
Falta
,
N. A.
Bowerman
,
P.
Marrack
,
A. P.
Fontenot
, et al
.
2014
.
Structural basis of chronic beryllium disease: linking allergic hypersensitivity and autoimmunity.
Cell
158
:
132
142
.
35
Fontenot
,
A. P.
,
B. E.
Palmer
,
A. K.
Sullivan
,
F. G.
Joslin
,
C. C.
Wilson
,
L. A.
Maier
,
L. S.
Newman
,
B. L.
Kotzin
.
2005
.
Frequency of beryllium-specific, central memory CD4+ T cells in blood determines proliferative response.
J. Clin. Invest.
115
:
2886
2893
.

The authors have no financial conflicts of interest.

Supplementary data