Beryllium Skin Patch Testing to Analyze T Cell Stimulation and Granulomatous Inflammation in the Lung

This information is current as Andrew P. Fontenot, Lisa A. Maier, Scott J. Canavera, Tara of September 25, 2021. B. Hendry-Hofer, Mark Boguniewicz, Elizabeth A. Barker, Lee S. Newman and Brian L. Kotzin J Immunol 2002; 168:3627-3634; ; doi: 10.4049/jimmunol.168.7.3627

http://www.jimmunol.org/content/168/7/3627 Downloaded from

References This article cites 40 articles, 12 of which you can access for free at: http://www.jimmunol.org/content/168/7/3627.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 25, 2021

*average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2002 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Beryllium Skin Patch Testing to Analyze T Cell Stimulation and Granulomatous Inﬂammation in the Lung1

Andrew P. Fontenot,2*¤ Lisa A. Maier,*‡§ Scott J. Canavera,* Tara B. Hendry-Hofer,¤ Mark Boguniewicz,¶ Elizabeth A. Barker,¤ Lee S. Newman,*‡§ and Brian L. Kotzin*†§

Chronic beryllium disease (CBD) is characterized by granulomatous inflammation and the accumulation of CD4؉ T cells in the lung. Patch testing of CBD patients with beryllium sulfate results in granulomatous inflammation in the skin. We investigated whether the T cell clonal populations present in the lung of CBD patients would also be present in the involved skin of a positive beryllium patch test and thus mirror the granulomatous process in the lung. CBD patients with clonal TCR expansions in bronchoalveolar lavage (BAL) were selected for study. All three CBD patients studied had a positive response to beryllium sulfate application and a negative patch test to normal saline. Immunohistochemistry showed extensive infiltration with CD4؉ T cells and few, if any, CD8؉ T cells both at 3 days and at later times when granulomas were apparent. T cell infiltration early after skin Downloaded from testing appeared to be nonspecific with the TCR repertoire of infiltrating T cells being distinct from that present in BAL. At later times when granulomas were present, T cell clones in skin overlapped with those in BAL in all patients tested. Total TCR matches in skin and BAL were as high as 40% in selected V␤ T cell subsets. Studies of peripheral blood T cells before and after patch testing provided evidence for mobilization of large numbers of pathogenic beryllium-reactive T cells into the circulating pool. These studies using skin patch testing provide new insight into the dynamics of T cell influx and mobilization during granulomatous inflammation. The Journal of Immunology, 2002, 168: 3627Ð3634. http://www.jimmunol.org/

hronic beryllium disease (CBD)3 is a granulomatous dis- Considerable evidence suggests that CD4ϩ T cells are important order that develops in up to 16% of individuals exposed in the immune response to beryllium and in the immunopathogen- C to beryllium, depending on the nature of the exposure esis of CBD (11, 16–19). For example, the development of gran- and the genetic susceptibility of the individual (1–7). Exposure to ulomatous inflammation is associated with the accumulation of beryllium continues to be a major public health concern because of CD4ϩ T cells in the bronchoalveolar lavage (BAL). In a number of its continued use in ceramics, aerospace, electronics, and defense individuals, these T cells included oligoclonal T cell expansions industries (1). Although the lung is the predominant organ in- ϩ with particular TCRs specific for CBD (19). CD4 T cells in the by guest on September 25, 2021 volved, other organ systems such as the skin, liver, and lymphatic BAL of CBD patients have been shown to proliferate and release system may also be affected by the granulomatous inflammatory Th1 cytokines in vitro in the presence of APCs expressing class II response (8, 9). The current disease definition requires a history of MHC molecules (14, 15, 17). It is currently unknown whether the beryllium exposure, the presence of granulomatous and/or T cells present in BAL fluid reflect the same T cells that drive mononuclear cell inflammation in a biopsy specimen, and a granulomatous inflammation in the tissue. Lung tissue is relatively proliferative response of blood or lung T cells to beryllium salts inaccessible to repeated sampling or to large samples, as are many in vitro (10–12). Susceptibility to CBD has been associated organs affected by immune-mediated disease. In addition, the with particular alleles of the class II MHC molecule, HLA-DP CD4ϩ T cells in BAL have been studied only after many years of (5–7, 13), and the mechanism for this susceptibility has been disease, after granulomas are well established. Finally, CD4ϩ T tied to the ability of these HLA alleles to present beryllium to ϩ cell function has been assessed primarily by in vitro testing. An specific CD4 T cells (14, 15). assay system that analyzed the in vivo T cell-dependent process in the development of granulomatous diseases would likely add important new insight into the disease process. Departments of *Medicine, †Immunology, and ‡Preventive Medicine and Biometrics, University of Colorado Health Sciences Center, Denver, CO 80262; and Departments The immune basis of CBD was first suggested by the demon- of §Medicine and ¶Pediatrics, National Jewish Medical and Research Center, Denver, stration of a delayed-type hypersensitivity response in beryllium- CO 80206 exposed subjects following skin patch testing with beryllium salts Received for publication September 6, 2001. Accepted for publication January 30, 2002. (20, 21). However, the use of beryllium patch testing as a diagnostic tool decreased in popularity because of the potential risk of The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance inducing sensitization and the theoretical risk that the underlying with 18 U.S.C. Section 1734 solely to indicate this fact. disease would be exacerbated (22). Recently, patch testing has 1 This work was supported by National Institutes of Health Grants HL62410, been reinvestigated as a diagnostic tool (23–25). In 14 subjects ES06358, and HL03982, and General Clinical Research Center Grant M01-RR-0051 from the Division of Research Resources. with CBD, positive patch tests were associated with no detectable adverse effects (25). In the present study, we used beryllium patch 2 Address correspondence and reprint requests to Dr. Andrew P. Fontenot, Depart- ment of Medicine, Division of Clinical Immunology (B164), University of Colorado testing in CBD patients to compare the repertoire of T cells infil- Health Sciences Center, 4200 East Ninth Avenue, Denver, CO 80262. E-mail address: trating skin with BAL T cells in the same individual. We noted that [email protected] the CD4ϩ T cells infiltrating skin in response to beryllium varied 3 Abbreviations used in this paper: CBD, chronic beryllium disease; BeSO4, beryllium sulfate; BAL, bronchoalveolar lavage; CDR, complementarity-determining re- depending on the timing and pathology. Early T cell infiltrates in gion; TCRB, TCR ␤-chain gene. the skin, before granuloma formation, expressed a different set of

TCRs. In contrast, later in the course of granuloma formation, a Analysis of TCRBV gene expression in peripheral blood, BAL, significant fraction of skin T cells matched those present in BAL and skin biopsies and expressed beryllium-reactive TCRs previously identified as Total RNA was isolated using an acid guanidinium-phenol chloroform specific for CBD. These T cell specificities were also detected for method, and cDNA was synthesized using 2 ␮g RNA per 20-␮l reaction. the first time in the circulating pool. These studies demonstrate an cDNA synthesis reagents included reverse transcriptase (SuperScript RT; in vivo model of granuloma formation in humans with mobiliza- Life Technologies, Grand Island, NY), RNase inhibitor (Promega, Madi- tion and influx of pathogenic CD4ϩ T cells similar to that which son, WI), and dNTPs and random hexamers (both from Amersham Phar- macia Biotech). One microliter of the cDNA reaction was added per 50 ␮l occurs in the internal organ primarily affected by this disease. PCR mixture. PCR amplification (AmpliTaq; PerkinElmer, Branchburg, NJ) was performed for 35 cycles, and both 5Ј and 3Ј oligonucleotide prim- Materials and Methods ers were present at a concentration of 0.3 ␮M. Sequences of the 5Ј TCR Study population ␤-chain gene (TCRB)V3 primer, 5Ј TCRBV14 primer, 5Ј TCRBV2 primer, and the 3Ј CB primer are 5Ј-GTCTCTAGAGAGAAGAAGG The diagnosis of CBD was established using previously defined criteria AGCGC-3Ј,5Ј-GTCTCTCGAAAAGAGAAGAGGAAT-3Ј,5Ј-GCAACT (11, 26), including a history of beryllium exposure, the presence of gran- TCCAATGAGGGCTCC-3Ј, and 5Ј-TTCTGATGGCTCAAACAC-3Ј, re- ulomatous inflammation on lung biopsy, and a positive proliferative respectively. The PCR products were ligated into the pCR II TA cloning sponse of BAL T cells to beryllium sulfate (BeSO4) in vitro. We studied vector (Invitrogen, San Diego, CA), and the ligation products were trans- ϩ three CBD patients in whom CD4 T cell expansions were previously formed into Epicurian Coli XL-1 Blue supercompetent Escherichia coli identified in the BAL (two with TCR V␤3- and one with TCR V␤14- cells (Stratagene, La Jolla, CA). Colonies containing inserts were randomly expressing expansions) (18). CBD patients 1 and 3 were being treated with selected for nucleotide sequencing. Cycle sequencing was performed using corticosteroids at the time of enrollment in this study. Informed consent M13 reverse (5Ј-CAGGAAACAGCTATGAC-3Ј) and/or M13 forward was obtained from each CBD patient, and the protocol was approved by the (5Ј-CCCAGTCACGACGTTGTAAAACG-3Ј) sequencing primers and an Downloaded from Human Subject Institutional Review Boards at the University of Colorado automated ABI 377 sequencer (PE Applied Biosystems, Foster City, CA). Health Sciences Center and National Jewish Medical and Research Center (Denver, CO). Lymphocyte proliferation assay

5 BeSO4 patch testing of CBD patients Proliferation assays were performed using PBMCs (2.5 ϫ 10 cells/well) cultured in 96-well ﬂat-bottom microtiter plates in the presence of 1 ␮g/ml Patch testing was performed using a modiﬁed version of the methods of the ϫ Ϫ4 ϫ PHA for 48 h or BeSO4 at concentrations ranging from 1 10 to 1

Ϫ http://www.jimmunol.org/ North American Dermatitis Group (27). One percent aqueous BeSO4 in 10 6 M for 96 h. The wells were then pulsed with 1 ␮Ci of [3H]thymidine Finn chambers was applied to normal unabraded skin on the back for 48 h. for an additional 18 h, and incorporation of radioactivity was determined The patch test was interpreted at 96 h as follows: 1ϩ, weak (nonvesicular) ϩ ϩ by beta emission spectroscopy. Proliferation assays were performed in positive reaction; 2 , strong (vesicular) positive reaction; 3 , extreme triplicate. (bullous) positive reaction. A 3- or 4-mm punch biopsy was performed at abnormal patch test sites at 96 h and after ϳ2–5 wk. A patch test with normal saline was also performed at day 0 and biopsied at 96 h. A biopsy Results of normal skin was performed at day 0. Skin punch biopsy specimens were TCR expression in peripheral blood and BAL T cells of CBD placed in cell lysis solution and homogenized with a glass/Teflon homog- patients enizer (VWR Scientific Products, Aurora, CA) and stored at Ϫ70°C until processed for analysis. We have previously identified a subset of CBD patients with CD4ϩ T cell subset expansions in the BAL compared with blood by guest on September 25, 2021 Immunohistochemistry of BeSO skin patch tests 4 (18, 19). Three of these individuals were selected to undergo A second punch biopsy was collected from each subject at the designated BeSO patch testing in the current study and were restudied for ␮ 4 time points, frozen, and cut into 5- m sections. Abs directed against CD3, alterations in TCR V region expression. The percentage of CD4ϩ CD4 (both from BD Biosciences, San Jose, CA), and CD8 (monoclonal IgG1; DAKO, Carpinteria, CA) were used for immunohistochemistry. Avi- T cells in the BAL and blood expressing particular V regions for din-biotin complex peroxidase methods were used for staining based on patients 1–3 was determined by immunofluorescence staining and VectaStain Elite ABC kit (Vector Laboratories, Burlingame, CA). Endog- cytofluorographic analysis (Fig. 1). Patient 1 showed an increased enous peroxidase activity was quenched with 3% hydrogen peroxide, then percentage of V␤3ϩ and V␤13.1ϩ T cells in the BAL compared blocked with 1.5% normal horse serum for 30 min. Sections were incu- with the blood population (for V␤3, 12.3 vs 6.1%, respectively). In bated for 60 min with the primary Ab diluted in 1.5% normal horse serum, ϩ washed in PBS, and incubated for 30 min in biotinylated horse anti-mouse patient 2, 9.3% of the CD4 T cells in the BAL expressed V␤3as IgG Ab. Following washing, sections were incubated for 30 min with HRP- compared with 4.6% in the blood. Patient 3 demonstrated a V␤14 conjugated avidin-biotinylated complex, washed again with PBS, then de- expansion with 10.8% of the CD4ϩ T cells in the BAL expressing Ј veloped with 3 -diaminobenzidine (Vector Laboratories) and counter- V␤14 compared with 3% in the blood. stained with hematoxylin. All samples were also stained using H&E.

Isolation and analysis of TCR expression of peripheral blood Histopathology of BeSO4 patch test skin biopsies and BAL CD4ϩ T cells At 96 h after application of 1% BeSO4, all three individuals de- Mononuclear cells were isolated from heparinized blood by Ficoll- veloped a strongly positive (2ϩ) patch test reaction characterized Hypaque (Amersham Pharmacia Biotech, Piscataway, NJ) density gradient by erythema, induration, and vesicles. No individual developed a separation and from BAL as previously described (28). Peripheral blood positive reaction to normal saline. A separate study from this lab- cells from CBD patients were stained with FITC-labeled anti-CD4 mAb (BD Biosciences), and CD4ϩ T cells were sorted using a MoFlo cell sorter oratory showed that control subjects with contact dermatitis and no (Cytomation, Fort Collins, CO). Sorting of BAL cells from CBD patients history of beryllium exposure fail to develop a positive skin reac- ϩ Ͼ for CD4 T cells was not performed, due to the 80% predominance of tion following BeSO4 patch testing (25). Thus, patch testing of ϩ CD4 T cells in this compartment. control subjects was not repeated in this study. Mononuclear cells were analyzed by two-color immunofluorescence for Skin biopsies obtained at 96 h after patch testing showed mild to CD4 and TCR V region gene expression using mAbs directed to 11 different TCR V␤ receptors and one TCR V␣ receptor as previously de- moderate spongiosis involving the lower layers of the epidermis scribed (18, 29). Anti-TCR mAbs were biotinylated and streptavidin-PE and focal edema of the papillary dermis (Fig. 2A). In contrast to the (Fisher Biotech, Pittsburgh, PA) was used as a second-step reagent for skin of healthy individuals, which usually harbors only a small TCR staining. Cells were also double-stained with FITC-labeled CD3, number of lymphocytes, these biopsies showed a significant lym- CD4, and CD8 (all from BD Biosciences). The lymphocyte population was identified using forward and 90° light scatter patterns, and fluorescence phocytic infiltrate in a perivascular distribution. There was no ev- intensity was analyzed using a FACScan cytometer (BD Biosciences) as idence of granulomatous inflammation at this early time point. previously described. Immunohistochemistry of the biopsy showed that the infiltrating The Journal of Immunology 3629

BeSO4 patch testing. However, none of these apparently expanded clonal populations in skin were identified in BAL, blood, or later samples of skin (Fig. 3), and none of the TCRs expressed by BAL expansions were detected in these skin samples (see below). In addition, no TCR sequence similarities were present among these different skin samples, e.g., normal skin (day 0) vs after saline or beryllium patch testing. The presence of clonal populations in normal skin and skin following normal saline patch testing in the face of no inflammation and few infiltrating lymphocytes could have been due to the very small amounts of TCR mRNA initially present and subsequent PCR amplification. Clonal expansions (i.e., two or more repeated sequences from the same T cell clone) were completely absent from the peripheral blood V␤3ϩCD4ϩ subset of patient 1. In contrast to the early biopsies, T cells present in the skin later

in the course of the BeSO4 patch test reaction, at the time of granuloma formation, signiﬁcantly overlapped with clonal populations in the BAL of the same individual (Fig. 4). For example, in patient 1, one TCRBV3 sequence (BV3-QGD-BJ1S5) accounted for 6 of Downloaded from 64 BV3 sequences (9.4%) in BAL and 2 of 39 (5.1%) in the skin (Fig. 4). This and other TCRBV3 sequences present in skin and/or BAL were noted to have a TCR complementarity-determining region (CDR)3 motif characteristic of pathogenic beryllium-reactive cells in CBD, and some sequences were identical to those previ-

ously described (19). This CDR3 motif includes an invariant as- http://www.jimmunol.org/ partate (D) at position 96 of the ␤-chain and a length of 8 amino acids. One of the larger BAL clonal populations, BV3-PGGGLG- FIGURE 1. TCR V␤ repertoire in the BAL compared with blood in BJ2S4, accounted for ϳ16% of the BAL TCRBV3 repertoire and three CBD patients selected for patch testing. Data are expressed as the 5.1% of the skin TCRBV3 sequences. Overall, 21% of the skin ء ␤ ϩ mean percentage of CD4 T cells expressing a particular V . , Expanded TCRBV3 repertoire matched sequences present in BAL. Neither of ␤ Ͼ V subsets ( 2-fold) in the individual patients. these pathologic repertoires overlapped with sequences present in blood before patch test application. Similar to previous studies (19), peripheral blood T cells contained no clonal populations and lymphocytes expressed CD3 and CD4 (data not shown). Few no identifiable beryllium-reactive TCRs. by guest on September 25, 2021 ϩ CD8 T cells were present. In the later biopsy of patient 3, 33% of the TCRBV14 sequences Second skin biopsies of the same patch tests were done after in skin were also present in BAL (Fig. 4). For example, one ϳ2–5 wk. As shown in Fig. 2B, the spongiosis of the epidermis TCRBV14 sequence, BV14-KGGG-BJ1S5, accounted for 2 of 27 and the edema of the papillary dermis were resolved. Fig. 2C (7.4%) sequences in the skin and 3 of 37 (8.1%) sequences in the shows a poorly formed noncaseating granuloma. Staining of these BAL. In addition, this sequence was also present in the blood lesions with anti-CD68 showed clusters of histiocytes and con- CD4ϩ T cell population (1 of 35 sequences). The largest firmed their granulomatous nature (data not shown). Similar to the TCRBV14 clonal expansion in BAL was also present in the skin findings at 96 h, the lymphocytic infiltrate was primarily composed biopsy. In patient 3, we also examined the V␤3 subset to see ϩ of CD3 cells (Fig. 2D) that expressed CD4 (Fig. 2E). Few, if any, whether the similarity in skin and BAL repertoire after beryllium ϩ CD8 T cells were present. stimulation extended beyond the V regions expanded in BAL. The percentage of V␤3ϩCD4ϩ T cells in the BAL and blood of patient Analysis of expressed TCRB genes in CBD patients 3 were not different (4.1 and 5.2%, respectively) (Fig. 1). Similar We focused on the TCR V regions expressed by expanded CD4ϩ to the findings for the expanded T cell subsets, TCRBV3 sequence T cell subsets in the BAL. Selection for skin CD4ϩ T cells was not matches were present (Fig. 4). For example, one TCRBV3 se- performed due to the rarity of CD8ϩ T cells in the skin biopsies quence, BV3-MRGGG-BJ2S2, accounted for 9 of 38 BV3 se- and the difficulty of obtaining adequate numbers of lymphocytes in quences (24%) in the lung and 5 of 26 (19%) in the skin. For this a 3-mm punch biopsy specimen. BAL cells from CBD patients subset, nearly 31% of the skin sequences were present in the BAL were also not sorted for CD4ϩ T cells due to the great predomi- repertoire and, conversely, 73% of the BAL sequences were nance of CD4ϩ T cells in this compartment. However, we sorted present in the skin. CD4ϩ T cells from the blood of individual patients from which RNA and cDNA were prepared. TCRBV3 fragments from patients Mobilization of beryllium-reactive T cells into blood after patch 1 and 2 and TCRBV14 fragments from patient 3 were amplified, testing cloned in bacteria, and sequenced. Fig. 3 shows junctional region Previous studies have shown that beryllium-reactive T cells in the sequences from patient 1 that were found at least two times in BAL of patients with CBD are rarely present in the circulating normal skin and in skin 96 h after either normal saline or BeSO4 pool (18, 19). In the present work, we noted that this compartmen- patch testing. Each of these biopsies contained oligoclonal T cell talization was upset after patch testing. TCRBV14, BV3, and BV2 populations, with a few clonal expansions approaching 20–30% of sequences expressed by peripheral blood CD4ϩ T cells in patient ␤ ϩ the V 3 subset. We detected repeated sequences of individual T 3 before and after BeSO4 patch testing are shown in Fig. 5 (BAL cell clones in control samples (normal skin and skin following sequences are shown again for reference). Before patch testing, 1 normal saline patch testing) and in the 96-h skin biopsy following of 35 BV14, 1 of 18 BV3, and 0 of 28 BV2 sequences in blood 3630 SKIN PATCH TESTING TO BERYLLIUM

FIGURE 2. Histology of representative skin punch biopsies following BeSO4 patch testing. A, H&E staining of a skin biopsy obtained 96 h after 1% aqueous

BeSO4 patch test application. Mild to moderate spongiosis of the lower epidermis and edema in the papillary dermis are evident. A perivascular lymphocytic

infiltrate is also apparent (original magnification, Downloaded from ϫ10). B, H&E staining of a skin biopsy obtained 35 days after BeSO4 patch test application. A decrease in the amount of papillary dermal edema (compared with 96 h) is seen. Lymphocytic infiltration in the dermis is also present (original magnification, ϫ10). C, Higher power magnification of a selected area of B showing a poorly formed noncaseating granuloma (original mag- http://www.jimmunol.org/ nification, ϫ40). D, Expression of CD3 in an area of lymphocytic infiltration in a skin biopsy obtained 14 days after BeSO4 patch test application (original magnification, ϫ20). E, Expression of CD4 in the same biopsy as shown in C. by guest on September 25, 2021

matched those present in BAL. Remarkably, at day 35 following BV14-PKPTGVG-BJ2S6, was not detected in the blood CD4ϩ patch testing, 63% of the blood TCRBV14 sequences were present population initially, but following patch testing this sequence ac- in BAL and a similar percentage of BAL sequences were present counted for 22% of the TCRBV14 repertoire. For the BV3 subset, in blood (Fig. 5). The largest BAL TCRBV14 clonal expansion, each of the three large clones in BAL were detected in blood after

FIGURE 3. Expression of TCR V␤3 gene (TCRBV3) amino acid sequences in peripheral blood, BAL, and skin T cells of patient 1. The sequence for the junctional region encompassing CDR3 is shown. Skin samples were obtained on day 0 (D0) before patch testing and 96 h and 14 days (D14) after patch testing with either saline or BeSO4. TCRBV3 sequences found at least two times are shown. The number of identical sequences (deﬁned at the nucleotide level) is shown over the total number of sequences analyzed for a given anatomic site. The Journal of Immunology 3631

FIGURE 4. Expression of TCRBV amino acid sequences expressed in peripheral blood, BAL, and skin of three CBD patients. The TCR junctional region encompassing CDR3 is shown. BAL samples were obtained at various times before skin patch testing. Peripheral blood samples were obtained on day 0 (D0) just before patch testing. Skin samples were obtained 14 (patient 1), 12 (patient 2), and 35 days (patient 3) after patch testing with BeSO4. TCRBV sequences found at least two times or found in more than one anatomic site are shown. The number of identical sequences (deﬁned at the nucleotide Downloaded from level) is shown over the total number of sequences analyzed for a given anatomic site. The conserved cysteine (C) of the ␤-chain is designated as position 90. TCRBV14 and TCRBV3 sequences are included for CBD patient 3. http://www.jimmunol.org/

patch testing, but the extent of overlap between the blood and BAL BAL increased from 63 to 72% for BV14, from 9.4 to 53% for by guest on September 25, 2021 repertoires was less than that seen for the BV14 subset. For the BV3, and from undetectable to 46% for BV2. The in vitro expan- BV2 subset, none of the expanded BAL clones was detected in sion of several BV3 and BV2 clones, initially identiﬁed in BAL, blood after patch testing. was dramatic. In previous work, when BAL T cells were obtained from pa- The presence of these new sequences in the circulation sug- tients with CBD, in vitro stimulation with BeSO4 caused a further gested that functional responses of peripheral blood T cells to be- expansion of the clones already present (18). This is consistent ryllium might be altered after patch testing. To test this possibility, with the beryllium speciﬁcity of these clones. We observed the PBMCs from patients 1 and 2 before and after patch testing were same phenomenon when peripheral blood T cells from patient 3 cultured in the presence BeSO4 for 5 days (Fig. 6). Proliferative were obtained after patch testing and cultured for 5 days with responses of cells from patient 1 were not detectable (stimulation Ͻ BeSO4 (Fig. 5). The frequency of TCRs matching in blood and index 2) before patch testing. These cells proliferated normally

FIGURE 5. Expression of TCRBV14, BV3, and BV2 amino acid sequences expressed in peripheral blood, BAL, and skin of patient 3. The TCR junctional region encompassing CDR3 is shown. BAL samples were obtained at various times before skin patch testing and are shown again for comparison. Peripheral blood samples were obtained on day 0 (D0) just before and 35 days (D35)

after patch testing with BeSO4. TCRBV sequences found at least two times or found in more than one anatomic site are shown. The number of identical sequences (de- ﬁned at the nucleotide level) is shown over the total number of sequences analyzed for a given anatomic site. 3632 SKIN PATCH TESTING TO BERYLLIUM

uals with CBD demonstrated erythema, induration, and a vesicular reaction typical of a contact dermatitis. This inﬂammation was beryllium speciﬁc in that it did not occur when saline was used

instead of BeSO4, and, although not tested in this study, inflammation has not been seen after beryllium patch testing of control individuals who do not have a history of beryllium exposure or evidence of beryllium-reactive T cells (25). Although clonal T cell populations were present in skin early after patch testing, none of the TCRs expressed by these cells matched those present in BAL of the same patients or was recognizable as beryllium-specific when compared with past studies (19, 30). Still, it seems likely that a subset of these early infiltrating cells were beryllium specific. It FIGURE 6. Proliferation to BeSO4 in culture of peripheral blood lymphocytes obtained before and after patch testing. PBMCs were obtained is possible that the few Ag-specific cells were masked by a larger from patients 1 and 2 on day 0 (D0) just before patch testing and on day number of nonspecific T cells. In the setting of an inflammatory 14 (patient 1) or 12 (patient 2) after patch testing to beryllium. Proliferation response, the production of chemotactic factors directs migration was determined by incorporation of tritiated thymidine and expressed as of activated effector CD4ϩ T cells to sites of inflammation, re- the mean Ϯ SEM. The stimulation index for each condition and time point gardless of specificity (31, 32). As a related example, mice immu- is shown over each bar. Cells from patients 1 and 2 at both time points nized with myelin basic protein demonstrated minimal enrichment ␮ proliferated in a similar and normal manner in the presence of 1 g/ml of pathogenic T cell subsets in inflammatory brain lesions (33). Downloaded from PHA (data not shown). Instead, T cells at the site of pathology expressed heterogeneous TCRs, with the pathologic T cell subset being obscured by the influx of nonspecific T cells (33). It is possible that culturing of the in response to PHA (data not shown). In contrast, when blood was early skin T cells with IL-2 would have enriched for the subset of obtained from patient 1 on day 14 after patch testing, proliferation Ag-specific T cells as has been shown in studies of EAE (34) and

to each concentration of BeSO4 was apparent with stimulation in- sarcoidosis (35). http://www.jimmunol.org/ Ϫ dices as high as 14 at the previously determined optimal 1 ϫ 10 5 Our studies of early skin lesions also provide an important con-

M BeSO4 (19). The extent of proliferation of cells from patient 2 trol for T cells present later in the granulomatous process. Thus, also appeared to be significantly enhanced when obtained after the subsequent appearance of repertoires enriched for BAL-like patch testing, although the higher background proliferation re- TCRs cannot be attributed to nonspecific recruitment of previously sulted in very little change in stimulation indices. Peripheral blood activated cells to sites of inflammation. Considering the enormous

cells from patient 3 obtained at day 35 following the BeSO4 patch diversity of the T cell repertoire, finding T cells expressing iden- ϫ Ϫ5 test proliferated vigorously in response to 1 10 M BeSO4 tical TCRs in different target organs indicates selection by the (peak thymidine incorporation, 63,261 Ϯ 2,156; stimulation in- same Ag, in this case beryllium. We were also able to identify dex ϭ 221) (data not shown in Fig. 6). Unfortunately, blood from TCRs documented to be beryllium specific based on previous stud- by guest on September 25, 2021 day 0 was not available for comparison. ies of cloned V␤3ϩ T cells from BAL. These TCRs are characterized by a ␤-chain CDR3 motif with an invariant aspartate (D) at Discussion position 96 and a particular length (19, 30). Although not analyzed Considerable evidence indicates that CBD is dependent on the in this study, the paired TCR ␣-chain is also highly related among generation and accumulation of beryllium-specific CD4ϩ T cells the cells that express this beryllium specific ␤-chain. Although (17, 19). To study lung involvement, CD4ϩ T cells have been only one of the three patients in the current study had BAL T cells primarily obtained by washing the airspaces with saline and then with this TCR, previous studies have demonstrated similar (or collecting this fluid by aspiration, i.e., BAL. In patients with long- even identical) TCRs in almost all individuals with CBD caused by standing CBD, a population of beryllium-reactive CD4ϩ T cell exposure to beryllium in the ceramics industry (30). Patients 2 and clones are compartmentalized to this site (19). These cells recog- 3 in the current study were exposed in the nuclear weapons nize beryllium in the context of HLA-DP on APCs, and a subset is industry. marked by oligoclonal expansions and particular beryllium-reac- The presence and expansion of CD4ϩ T cells in skin that ex- tive TCRs (19). However, lung damage in CBD is secondary to the press TCRs identical to those in BAL provides new insight regard- granulomatous inflammation in lung tissue, and it is currently un- ing the pathogenic potential of these cells. Infiltration into tissue in clear how the beryllium-reactive cells in BAL participate in the association with the granulomatous process strongly suggests that tissue inflammation. This issue has been difficult to assess because these CD4ϩ T cells do the same in lung tissue. Our studies also of the small size and random nature of transbronchial lung biopsies show that these cells have the capability of recognizing beryllium as well as difficulty in repeated sampling. Also, it is unknown in other tissues, and it seems likely that granulomas in liver and whether the same T cells are involved in granulomatous inflam- lymphatic tissue in CBD involve the same subset of cells. It is mation at other non-lung sites and whether the compartmentaliza- currently not clear whether the pathogenic T cells arrive early in tion to lung can be altered by a stimulus elsewhere. In the present the skin and then expand over the next 2–5 wk or whether their study, we performed beryllium patch testing to address these is- arrival and expansion in the tissue are continuous during the gen- sues. The results suggest that patch testing creates an in vivo model eration of granulomas. The latter is supported by the high fre- of the pathogenic granulomatous process. The development of quency of these cells in the circulating pool later after patch granulomas in the skin was associated with mobilization and tissue testing. infiltration of the same T cell clones present in BAL. The granulomatous inflammation in CBD is identical to that in Few lymphocytes are present in the skin of healthy individuals. sarcoidosis, and these diseases share a number of clinical charac-

Within 96 h of applying BeSO4 onto the skin of individuals sen- teristics (36). The development of granulomatous inﬂammation in

sitized to beryllium, inﬂammation accompanied by inﬁltration with the skin following patch testing with BeSO4 is similar to the CD4ϩ T cells occurs. Similar to previous studies (25), all individ- Kveim-Siltzbach reaction seen in patients with sarcoidosis (37, The Journal of Immunology 3633

38). Following the intradermal injection of the Kveim-Siltzbach were able to observe the progression from influx of nonspecificT reagent (a suspension of sarcoid tissue extract), noncaseating gran- cells to the presence and expansion of beryllium-specific CD4ϩ T ulomatous inflammation develops after ϳ4 wk. Analysis of TCR cells at the time of granuloma formation. The kinetics of this pro- ␤-chain repertoire of these reactions showed oligoclonal T cell cess cannot be studied in the organs of individuals with established populations consistent with a conventional Ag response (38). disease and in vitro studies cannot provide the same insight shown Compared with BAL T cells from sarcoidosis patients in general, in this work. Whether patch testing can be used in non-granulo- similarities in the skin TCR CDR3 regions were noted to be sug- matous diseases thought to be mediated by CD4ϩ T cells remains gestive, but analysis of BAL TCR repertoire in the same individ- unknown. For example, patch testing may be useful to mobilize uals was not performed and identical TCRs were not found. Our and characterize pathogenic CD4ϩ T cells in diseases where these study is unique in this regard. cells are sequestered in target organs, such as the brain in multiple The majority of CBD patients will have blood cells that test sclerosis, pancreas in insulin-dependent diabetes mellitus, or even positive in a beryllium lymphocyte proliferation assay (1). Thus, joints in rheumatoid arthritis. Obviously, the safety of this proce- these individuals have beryllium-specific T cells in their circulat- dure in these diseases would need to be addressed in a rigorous ing pool. Still, we have noted that there is a separation of lung and fashion. peripheral blood CD4ϩ TCR repertoires in these patients (19). Precursors or progeny of clonally expanded BAL T cells are absent References or extremely rare in the circulating pool. The same was true in the 1. Newman, L. S., L. A. Maier, and B. Nemery. 1998. Interstitial lung disorders due to beryllium and cobalt. In Interstitial Lung Disease. M. I. Schwarz and T. E. present study, when blood T cells were analyzed before patch test- King, Jr., eds. B.C. Decker, Hamilton, Ontario, p. 367. ing. These results likely relate to the low frequency of blood T 2. Kreiss, K., S. Wasserman, M. M. Mroz, and L. S. Newman. 1993. Beryllium Downloaded from cells that proliferate in response to beryllium in culture, which is disease screening in the ceramics industry: blood test performance and exposure- disease relations. J. Occup. Med. 35:267. below the frequency detectable by TCR sequence analysis. How- 3. Kreiss, K., M. M. Mroz, B. Zhen, J. W. Martyny, and L. S. Newman. 1993. ever, it is also possible that blood and BAL beryllium-specificT Epidemiology of beryllium sensitization and disease in nuclear workers. Am. Rev. Respir. Dis. 148:985. cells are separate populations with separate TCRs. In the present 4. Kreiss, K., M. M. Mroz, L. S. Newman, J. Martyny, and B. Zhen. 1996. Ma- study, after 2–5 wk following patch testing, we noted a marked chining risk of beryllium disease and sensitization with median exposures below ␮ 3 increase in T cells with TCRs identical to those expanded in BAL. 2 g/m . Am. J. Ind. Med. 30:16. http://www.jimmunol.org/ 5. Richeldi, L., R. Sorrentino, and C. Saltini. 1993. HLA-DPB1 glutamate 69: a The beryllium-specific nature of these cells is further supported by genetic marker of beryllium disease. Science 262:242. the enhanced proliferation of peripheral blood lymphocytes and 6. Richeldi, L., K. Kreiss, M. M. Mroz, B. Zhen, P. Tartoni, and C. Saltini. 1997. the further expansion of particular T cell clones after in vitro stim- Interaction of genetic and exposure factors in the prevalence of berylliosis. Am. J. Ind. Med. 32:337. ulation with beryllium. Only considering the three TCRBV subsets 7. Wang, Z., P. S. White, M. Petrovic, O. L. Tatum, L. S. Newman, L. A. Maier, and (which account for ϳ20% of the repertoire), our results suggest B. L. Marrone. 1999. Differential susceptibilities to chronic beryllium disease ϩ contributed by different Glu69 HLA-DPB1 and -DPA1 alleles. J. Immunol. 163: that at least 2% of the peripheral blood CD4 T cells may be 1647. beryllium specific after patch testing. This frequency of Ag-spe- 8. Kriebel, D., J. D. Brain, N. L. Sprince, and H. Kazemi. 1988. The pulmonary cific CD4ϩ T cells has only been reported for some patients with toxicity of beryllium. Am. Rev. Respir. Dis. 137:464.

ϩ 9. Freiman, D. G., and H. L. Hardy. 1970. Beryllium disease: the relation of pul- by guest on September 25, 2021 chronic viral infections (39). In contrast, memory CD4 T cells monary pathology to the clinical course and prognosis based on a study of 130 specific for tetanus toxoid or purified protein derivative of Myco- cases from the U.S. beryllium case registry. Hum. Pathol. 1:25. bacterium tuberculosis in healthy individuals occur at a frequency 10. Epstein, P. E., J. H. Dauber, M. D. Rossman, and R. P. Daniele. 1982. Bron- choalveolar lavage in a patient with chronic berylliosis: evidence for hypersen- of 1 in 5,000 to 1 in 100,000 CD4 cells (40). Estimates of Th cell sitivity pneumonitis. Ann. Intern. Med. 97:213. precursor frequency in the blood following injection of a conven- 11. Rossman, M. D., J. A. Kern, J. A. Elais, M. R. Cullen, P. E. Epstein, O. P. Preuss, T. N. Markham, and R. P. Daniele. 1988. Proliferative response of bronchoal- tional Ag like keyhole limpet hemocyanin ranged from 1 in veolar lymphocytes to beryllium. Ann. Intern. Med. 108:687. 100,000 to 1 in 800,000 (41). 12. Mroz, M. M., K. Kreiss, D. C. Lezotte, P. A. Campbell, and L. S. Newman. 1991. Our results show that application of BeSO to the skin of indi- Reexamination of the blood lymphocyte transformation test in the diagnosis of 4 chronic beryllium disease. J. Allergy Clin. Immunol. 88:54. viduals with CBD results in a mobilization of beryllium-specific 13. Saltini, C., M. Amicosante, A. Franchi, G. Lombardi, and L. Richeldi. 1998. ϩ CD4 T cells into the circulating pool. Depending on the degree of Immunogenetic basis of environmental lung disease: lessons from the berylliosis systemic Ag exposure after patch testing, the lymphoid tissue model. Eur. Respir. J. 12:1463. 14. Fontenot, A. P., M. Torres, W. H. Marshall, L. S. Newman, and B. L. Kotzin. draining the patch test site or even the peri-lung lymphoid tissue 2000. Beryllium presentation to CD4ϩ T cells underlies disease susceptibility may be the sites contributing to the circulating cells. The results HLA-DP alleles in chronic beryllium disease. Proc. Natl. Acad. Sci. USA 97: 12717. prompted us to reevaluate the safety of patch testing in CBD pa- 15. Lombardi, G., C. Germain, J. Uren, M. T. Fiorillo, R. M. du Bois, tients. The safety of patch testing has been previously questioned W. Jones-Williams, C. Saltini, R. Sorrentino, and R. Lechler. 2001. HLA-DP (20, 21). More recent studies have found no evidence of clinical allele-specific T cell responses to beryllium account for DP-associated susceptibility to chronic beryllium disease. J. Immunol. 166:3549. deterioration or progression related to patch testing (25). Addi- 16. Newman, L. S., C. Bobka, B. Schumacher, E. Daniloff, B. Zhen, M. Mroz, and tional follow-up of the same patients indicated no evidence of T. E. King, Jr. 1994. Compartmentalized immune response reflects clinical se- disease exacerbation (L. S. Newman, unpublished observation). verity of beryllium disease. Am. J. Respir. Crit. Care Med. 150:135. 17. Saltini, C., K. Winestock, M. Kirby, P. Pinkston, and R. G. Crystal. 1989. Main- Furthermore, no clinical changes have been noted in the three in- tenance of alveolitis in patients with chronic beryllium disease by beryllium- dividuals patch tested in the current study, and follow-up is being specific helper T cells. N. Engl. J. Med. 320:1103. 18. Fontenot, A. P., B. L. Kotzin, C. Comment, and L. S. Newman. 1998. Expansions continued. It should be noted that we deliberately selected study of T-cell subsets expressing particular T cell receptor variable regions in chronic subjects for patch testing who no longer worked in the beryllium beryllium disease. Am. J. Respir. Cell Mol. Biol. 18:581. industry, as a further safeguard against the theoretical possibility of 19. Fontenot, A. P., M. T. Falta, B. M. Freed, L. S. Newman, and B. L. Kotzin. 1999. Identification of pathogenic T cells in patients with beryllium-induced lung dis- exacerbating the beryllium-induced immune response. However, ease. J. Immunol. 163:1019. the increased number of circulating beryllium-specific T cells after 20. Curtis, G. H. 1951. Cutaneous hypersensitivity due to beryllium: a study of thir- patch testing suggests the possibility that disease exacerbation may teen cases. AMA Archs. Dermatol. Syph. 64:470. 21. Curtis, G. H. 1959. The diagnosis of beryllium disease, with special reference to occur after patch testing, although none has been seen in the small the patch test. AMA Arch. Ind. Health 19:150. number of patients followed to date. 22. Waksman, B. H. 1959. The diagnosis of beryllium disease, with special reference to the patch test. AMA Arch. Ind. Health 19:154. Together, our results indicate that patch testing provides an in 23. Vilaplana, J., C. Romaguera, and F. Grimalt. 1992. Occupational and non-occu- vivo model of granulomatous disease in individuals with CBD. We pational allergic contact dermatitis from beryllium. Contact Dermatitis 26:295. 3634 SKIN PATCH TESTING TO BERYLLIUM

24. Haberman, A. L., M. Pratt, and F. J. Storrs. 1993. Contact dermatitis from be- 34. Kotzin, B. L., S. Karuturi, Y. K. Chou, J. Lafferty, J. M. Forrester, M. Better, ryllium in dental alloys. Contact Dermatitis 28:157. G. E. Nedwin, H. Offner, and A. A. Vandenbark. 1991. Preferential T cell re- 25. Bobka, C. A., L. A. Stewart, G. J. Engelken, L. E. Golitz, and L. S. Newman. ceptor V␤ gene usage in myelin basic protein reactive T cell clones from patients 1997. Comparison of in vivo and in vitro measures of beryllium sensitization. with multiple sclerosis. Proc. Natl. Acad. Sci. USA 88:9161. J. Occup. Environ. Med. 39:540. 35. Forrester, J. M., Y. Wang, N. Ricalton, J. E. Fitzgerald, J. Loveless, 26. Newman, L. S., K. Kreiss, T. E. King, Jr., S. Seay, and P. A. Campbell. 1989. L. S. Newman, T. E. King, and B. L. Kotzin. 1994. TCR expression of activated Pathologic and immunologic alterations in early stages of beryllium disease: re- T cell clones in the lungs of patients with pulmonary sarcoidosis. J. Immunol. examination of disease definition and natural history. Am. Rev. Respir. Dis. 139: 153:4291. 1479. 36. Newman, L. S., C. S. Rose, and L. A. Maier. 1997. Sarcoidosis. N. Engl. J. Med. 27. Adams, R. M. 1981. Patch testing: a recapitulation. J. Am. Acad. Dermatol. 336:1224. 5:629. 37. Siltzbach, L. E. 1961. The Kveim test in sarcoidosis: a study of 750 patients. 28. The BAL Cooperative Group Steering Committee. 1990. Bronchoalveolar lavage J. Am. Med. Assoc. 178:476. constituents in healthy individuals, idiopathic pulmonary fibrosis and selected 38. Klein, J. T., T. D. Horn, J. D. Forman, R. F. Silver, A. S. Teirstein, and comparison groups. Am. Rev. Respir. Dis. 141:S168. D. R. Moller. 1995. Selection of oligoclonal V␤-specific T cells in the intrader- 29. Fitzgerald, J. E., N. S. Ricalton, A. C. Meyer, S. G. West, H. Kaplan, C. Behrendt, mal response to Kveim-Siltzbach reagent in individuals with sarcoidosis. J. Im- and B. L. Kotzin. 1995. Analysis of clonal CD8ϩ T cell expansions in normal munol. 154:1450. individuals and patients with rheumatoid arthritis. J. Immunol. 154:3538. 39. Waldrop, S. L., C. J. Pitcher, D. M. Peterson, V. C. Maino, and L. J. Picker. 1997. 30. Fontenot, A. P., L. S. Newman, and B. L. Kotzin. 2001. Chronic beryllium dis- Determination of antigen-specific memory/effector CD4ϩ T cell frequencies by ease: T cell recognition of a metal presented by HLA-DP. Clin. Immunol. 100:4. flow cytometry: evidence for a novel, antigen-specific homeostatic mechanism in 31. Fontenot, A. P., and B. L. Kotzin. 2000. Immune recognition and responses. In HIV-associated immunodeficiency. J. Clin. Invest. 99:1739. Textbook of Respiratory Medicine, Vol. 1. J. F. Murray, J. A. Nadel, R. J. Mason, 40. Helms, T., B. O. Boehm, R. J. Asaad, R. P. Trezza, P. V. Lehmann, and and H. A. Boushey, Jr., eds. Saunders, Philadelphia, p. 413. M. Tary-Lehmann. 2000. Direct visualization of cytokine-producing recall anti- 32. von Andrian, U. H., and C. R. Mackay. 2000. T-cell function and migration: two gen-specific CD4 memory T cells in healthy individuals and HIV patients. J. Im- sides of the same coin. N. Engl. J. Med. 343:1020. munol. 164:3723. 33. Bell, R. B., J. W. Lindsey, R. Sobel, S. Hodgkinson, and L. Steinman. 1993. 41. Lane, H. C., G. Whalen, and A. S. Fauc. 1983. Antigen-induced human T cell Diverse T cell receptor V␤ gene usage in the central nervous system in experi- help: precursor frequency, radiation sensitivity, and allogeneic effects. J. Clin. Downloaded from mental allergic encephalomyelitis. J. Immunol. 150:4085. Invest. 72:636. http://www.jimmunol.org/ by guest on September 25, 2021