Associated Peptide Antigens of a Renal Carcinoma Cell Line by a Novel Mass Spectrometric Method1

[CANCER RESEARCH 58. 5803-5811. December 15. 1998] Direct Identification of Major Histocompatibility Complex Class I-bound Tumor- associated Peptide Antigens of a Renal Carcinoma Cell Line by a Novel Mass Spectrometric Method1

Thomas Flad, Bernhard Spengler, Hubert Kaibacher, Peter Brossart, Daniel Baier, Raimund Kaufmann, Peter Bold, Sabine Metzger, Martin Bliiggel, Helmut E. Meyer, Bernd Kurz, and Claudia A. MÃ¼ller2

Department of Hemalology. Oncology, and Immunology. University of TÃ¼bingen,72076 Tubingen, Germany Â¡T.F., P. Hr., D. B., B. K., C. A. M.I: Institute for Laser Medicine Â¡B.S., R. K., P. Bo., S. M.I and Center for Biological and Medical Research IB. S., R. K.I, University of DÃ¼sseldorf,40225 DÃ¼sseldorf,Germany: Medical and Natural Sciences Research Center, University of Tubingen. 72074 TÃ¼bingen.Germany Â¡H.K.Â¡:and Institute of Physiological Chemistry, University of Bochum, 44780 Bochum, Germtmv Â¡M.B., H. E.M.I

ABSTRACT tance in the rejection of solid tumors in animal models and humans (reviewed in Ref. 2 and 3). Therefore, recent research efforts strongly Melanoma and renal cell carcinoma (RCC) are thought to be the most concentrate on the identification of specific tumor antigens that acti immunogenic human tumors. Presently a series of tumor-specific peptides vate tumor-reactive T lymphocytes and thus promise to be effective in of melanoma is being tested in clinical trials with different immunother- apy protocols. In contrast, only one decameric peptide (SPSSNRIRNT) i/i vivo or in vitro immunization strategies. derived from one (ORF2) of three possible open reading frames (ORFs) of A very successful approach to characterize human tumor antigens a gene named RAGE (Renal tumor AntiGEn) was shown to be the target was pioneered by Boon and his group (4-7), who used peripheral for tumor-specific CTLs on renal carcinoma cells. One reason for the lack blood lymphocytes sensitized in vitro to tumor cells as screening of identification of tumor antigens on RCC compared with melanoma may instruments to identify DNA clones from genomic libraries of mela be the difficulty in generating tumor-specific CTLs as screening instru nomas. In this way, genes encoding tumor antigens were identified ments. Therefore, our approach was directly to isolate and identify peptides bound to HLA class I molecules of the HLA-A2 and -B8 homozygous after molecular expression cloning into tumor antigen loss variants. RCC line A-498. High performance liquid chromatography-fractionated For the first time, this approach led to the definition of a variety of peptides eluted with acid from immunoaffinity-purified HLA class I-pep- tumor antigens on human melanoma cells with either shared or tide complexes were sequenced and identified for the first time by the restricted distribution in different tumors (4-7). In several clinical novel and highly sensitive mass spectrometric method matrix-assisted trials, in vivo immunizations with the relevant T-cell epitopes of these laser desorption ionization-post source decay (MALDI-PSD) from minute tumor proteins are in progress to evaluate their potential in vivo amounts of 100 fmol to 1.5 pmol of the fractionated peptide samples. induction of immune responses and effectiveness in tumor destruction Fourteen peptide sequences first deduced from interpretations of the mass (8). spectra were also shown to fulfill other reliability criteria such as match ing the mass spectra of the respective synthetic peptides. Some peptides However, many human tumor cells are known to exert strong anergizing effects on T cells to escape immune recognition (9-11). were identified to be derived from genes preferentially activated in ma lignant tissues or resulted from a possibly mutated gene. The most prom These phenomena may account for the difficulties in routine expan ising candidate for a CTL epitope is a decameric peptide (PASKKTD- sion of tumor-reactive T cells from cancer tissue and their application PQK) derived from another possible ORF (ORF5) of the RAGE gene and as broader screening instruments for the definition of multiple tumor probably presented in association with HLA-B8. This peptide was synthe antigens that could be used as targets for specific immune recognition sized and used for the in vitro induction of CTLs that lysed the A-498 cells in different malignomas. On the basis of the observations that tumor- and another HLA-B8-positive RCC line significantly more strongly than either other AAGE-positive but HLA-B8-negative RCC lines or K562 cells. infiltrating cytotoxic T cells recognize peptide antigens in association Sensitive sequencing by MALDI-PSD thus may provide a powerful with HLA class I molecules, alternative strategies for characterization method of identifying potentially tumor-specific and HLA-restricted an of tumor antigens have been envisaged (12). Some of these ap tigens, even on native malignant cells and tissues. proaches involve direct release and sequence analysis of the MHC- associated peptides from the tumor cells. They use fractionation of INTRODUCTION peptide pools eluted with acid from MHC molecules by reversed phase-HPLC3 (4) before sequencing is performed. As pioneered by Since Gross (1) succeeded in protecting mice from carcinogen- Roetzschke et al. (13) and Van Bleek and Nathanson (14), conven induced tumors by prior s.c. immunization in 1943, tumor immunol- tional Edman sequencing has been successfully used for the identifi ogists aimed intensively at the in vivo induction and amplification of cation of alloantigenic and viral peptide epitopes that were dominant immune responses that effectively recognize and eradicate malignant species in such highly purified HPLC fractions. However, this ap cells. In the past two decades, numerous studies established that T proach only rarely led to the direct characterization of peptide anti cells as opposite to humoral immune responses are of major impor- gens on tumor cells (15, 16). Most extracted peptides representing tumor epitopes do not constitute the major peptide species in the Received 6/19/98: accepted 10/16/98. eluted HPLC fractions. Therefore, for direct sequencing of tumor 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 with peptide antigens associated with MHC molecules, more sensitive and 18 U.S.C. Section 1734 solely to indicate this fact. 1This work was partially supported by the "Fortune" Program 128.2093.1/2 of the differentiating techniques had to be applied. With the introduction of Medical Faculty of the University of Tuebingen. by the Medical Faculty of the University electrospray tandem mass spectrometry, high resolution separation of Duesseldorf. by the German Ministry for Education. Science. Research, and Technol and sequencing of single peptides from melanomas has been achieved ogy (BMBF, Contract 0310709). by the European Union (Contract MAT1-CT94-0034). by the German Ministry for Science and Research of Nordrhein-Westfalen, and by a grant from Thermo Bioanalysis, Lid.. Hemel Hempstead. United Kingdom. 3 The abbreviations used are: HPLC. high performance liquid chromatography: RCC, 2 To whom requests for reprints should be addressed, at Medical University Clinic. renal cell carcinoma: RAGE, renal tumor antigen: MALDI-PSD. matrix-assisted laser Department II. Section of Transplantation Immunology and Immunohematology, Otfried- desorption ionization-post source decay; NK, natural killer; mAb. monoclonal antibody; MUller-Strasse-10. D-72076 TÃ¼bingen.Germany. Fax: 49-7071-295755: E-mail: Claudia. TFA, trifluoroacetic acid; FACS, fluorescence-activated cell sorter; DC. dendritic cell; [email protected]. ORF. open reading frame: ESI-MS-MS. electrospray tandem mass spectrometry. 5803

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1998 American Association for Cancer Research. MALÃœI-PSD REVEALS TUMOR PEPTIDES OF A RCC LINE even in HPLC fractions still containing a mixture of different peptides separated from simultaneously detached, peptide-tree HLA class I molecules (17, 18). by ultrafiltration using a 10 kDa cutoff membrane (Sartorius. GÃ¶ttingen. For the first time, the following study presents the MALDI-PSD Germany). The final flow-through provided the acid-eluted peptide pools A mass spectrometry technique as an alternative sensitive method to (from the TH4 conjugated column) and B (from the W6/32.HL conjugated directly sequence tumor peptides eluted with acid from MHC class I column) that were lyophilized overnight and subsequently solubili/.ed in 1.5 ml molecules as exemplified on the human RCC line A-498. This RCC of water/acetonitrilenTA (60:40:0.1). The resulting solution was concentrated in a speed vac to 50 /Â¿I.Aliquots of the peptide-free MHC class I molecules line was used to identify a series of tumor-related peptides associated were analyzed by SDS-PAGE using the method of Laemmli and judged to be with HLA-A2 and HLA-B8. Renal carcinomas belong to the group of >95% pure, showing the expected bands of 44 kDa for HLA class I heavy tumors that are able to elicit antilumor immune responses. Recently. chains and 12 kDa for ÃŸ,-microglobulin. The final yields of HLA molecules RAGE-\ has been characterized as the first immunogenic tumor were quantified by the method of Bradford. antigen on RCCs (19) using the cDNA cloning method described Microbore Reversed Phase-HPLC. Peptides were separated by micro- above. We demonstrate that the sensitive MALDI-PSD technique also bore reversed phase-HPLC using a 125 x 2.1 mm C,â€žcolumn (Vydac. allowed the direct identification of another RACE-\ -related peptide as Hesperia. CA) connected with a HPLC intelligent pump model L 6200 (Merck- a further possible immunogenic endogenous antigen on RCCs. Hitachi. Darmstadt. Germany) and a diode array detector model 1000S (Ap plied Biosystems, Weiterstadt. Germany). An elution gradient consisting of 0-80% acetonitrile/0.05% TFA for 70 min with a flow rate of 150 /xl/min was MATERIALS AND METHODS applied. From the absorbance trace at 220 nm, each fraction containing absorbance peaks of the eluate was collected and stored at -80Â°C until Cell Lines. The RCC lines A-498 (HLA-A2/A2, B8/B8, Cw7/Cw7), sequence analysis. Caki-2 (HLA-A1/A11. BX/B51, Cw7/Cwl2), and 769-P (HLA-A3/A24. B7/ Peptide Analysis. Mass spectrometric analysis of the HPLC fractions was B7, Cw7/Cw7), as well as the proerythroblastic HLA antigen-negative cell line performed by MALDI-PSD instrumentation. The instrument used was a in- K562 used as a target cell tor NK cell activity, were purchased from the house-built laser reflectron time-of-flight mass spectrometer developed for American Type Culture Collection (Rockville, MD). The MZ-1851-RCC line optimized PSD analysis as a second-generation design derived from the (HLA-A1/A2. B44/B41. Cw7/Cw7) derived from the primary tumor of a original MALDI-PSD instrument (22, 23). The details of the instrumental patient with RCC was kindly provided by Alexander Knuth (Northwest Hos design will be described elsewhere. The mass spectrometer was equipped with pital, Frankfurt. Germany I. The peptide transporter-detective hybrid cell line a dual-stage gridded reflector that was coaxial with the ion beam path. The ion T2 (174 X CEM.T2: kindly provided by H-G. Rammensee. University of source was a two-stage acceleration system operated at 15 keV total acceler TÃ¼bingen;HLA-A2. B5) was used for peptide binding assays to HLA-A2. ation energy. For the studies described in this report, the instrument was run in All cell lines were maintained in vilro in RPMI 1640 (Life Technologies. delayed ion extraction mode (24). Inc., Grand Island. NY) supplemented with 2 mM L-glutamine, 25 mM From the HPLC-fractionated peptide peaks, samples were prepared for PSD HEPES, 10% heat-inactivated PCS, and 50 /xg/ml gentamicin at 37Â°Cand analysis using 2.5-dihydroxybenzoic acid as an embedding matrix according to 5% CO2. The RCC line A-498 was propagated in vilro in Triple-flasks standard procedures reported previously (22). In a first run of experiments, the (NUNC. Wiesbaden, Germany) to mass cultures and incubated with 1000 components of each sample containing a mixture of peptides were character units/ml of human recombinant IFN-y (PBH, Hannover, Germany) for 72 h ized by their molecular mass due to the production of stable molecular ions in to increase expression of HLA class I antigens before harvesting. After the MALDI mode setting of the instrumentation. In a second run. MALDI-PSD expansion. A-498 cells were detached by 10X trypsin/EDTA (PAA, Colbe, mass spectrometry was performed that allowed the analysis of the molecular Germany), washed once with PBS (Life Technologies. Inc.. Grand Island. masses of structurally specific product ions formed by decay of selected NY), quick-frozen as cell pellet, and stored at -70Â°C in 50 ml of BlueMax peptide precursor ions in the first field-free drift region of the spectrometer tubes (Falcon. Heidelberg. Germany I. after leaving the ion source. In the MALDI-PSD mode, a precursor selection Immunofluorescence Analysis. Expression of HLA class I molecules was device (ion gate) was used to obtain product ions only from (ideally) a single assessed by flow cytometry on a FACScan (Becton Dickinson. Mountain precursor peptide out of a mixture of components of different molecular View. CA) using the mouse mAbs W6/32.HL (anti-HLA-A.B.C common masses in samples of the HPLC-fractionated peptide peaks. The resolution determinant; Ref. 20) and TH4 (anti-HLA-B.C common determinant; pro M/AM of this selection step was in the range of 100-200. Thus, in many cases duced in our own laboratory) and W6/32.HK (20) as first labeling reagents and it was possible to acquire complete PSD product ion spectra from individual FITC-conjugated AffiniPure F(ab')2 fragments goat-anti-mouse IgG + IgM components of the HPLC fractions. (Dianova. Hamburg, Germany) as second labeling reagents of the analyzed cell Sequence analysis of unknown peptides was performed by computer-aided lines according to the procedure described previously (21). The fluorescence interpretation of the pattern of COOH-terminal, NH2-terminal, and internal intensity and scatter signals were analyzed by the WinMDI flow cytometry fragment ions observed for each peptide. Instrumental control and data acqui software (http://facs.scripps.edu/). sition and interpretation were supported by computer software developed in the Isolation of MHC Class I Molecules and Bound Endogenous Peptides. Institute of Laser Medicine in Duesseldorf. Different protein database search HLA-class I molecules of the RCC-linc A-498 were purified by means of programs'* were used as additional tools for structure analysis and for deter immunoaffinity chromatography. Briefly, 1 X 10"' IFN-y-stimulated A-498 mination of the protein source of the eluted peptides. The applied procedure for cells were incubated for 15 min in 30 ml of lysis buffer 110 mM Tris/HCI (pH reliable peptide sequence elucidation for database-listed or completely un 8.0), 0.2 mM phenylmethylsulfonyl fluoride (Boehringer Mannheim. Mann known peptides is described elsewhere (25). It was based on an approach of heim. Germany). 5 /XMleupeptin (Calbiochem, Bad Soden. Germany), and 10 combined use of mass calculations and combinatorics, recognition of complex /xM chymostatin (Boehringer Mannheim)] on ice and subsequently further ion signal patterns, as well as on knowledge of the fragmentation behavior of disrupted in 29Ã•NP40 (Serva, Heidelberg. Germany). The cell lysate was peptides. High reliability of a final sequence proposition after spectra inter cleared by ultracentrifugation (100.000 X g for 20 min at 4Â°C)and cycled pretation was assumed only if all observed signals of sufficient intensity had overnight through a precolumn (CNBr-activated Sepharose saturated with found a reasonable match and if all highly expected signals had indeed been Tris/glycine). Subsequently, it was passed over a series of immunoaffinity observed in the native spectrum. columns prepared from /V-hydroxysuccinimide-activated Sepharose (Pharma After interpretation of the MALDI-PSD spectra of the native sample, cia, Uppsala, Sweden) that were conjugated to: (a) the mAb TH4 for purifi additional methods for validation of the determined sequence were used. In cation of HLA-B and HLA-C antigens; and (Ã¨) the mAb W6/32.HL for several cases, unambiguous sequence determination was achieved by using purification of the remaining HLA class I antigens. The columns were first hydrogen-deuterium exchange of the investigated peptide to enhance structural washed with buffered detergent solution [50 mM sodium phosphate (pH 8.0), 100 mM NaCI, and 0.1% Zwittergent-12 (Calbiochem, La Jolla, CA)] and 4 http://www.mann.embl-heidelberg.de/Semces/PeptideSearch/FR_PeptidePattem thereafter with water to remove the detergents. Peptides were eluted from the Form.html. or http://falcon.ludwig.ucl.ac.uk/mstag.html, or http://prowl.rockefeller.edu/ bound HLA class I molecules with diluted TFA (pH 2.0) and subsequently PRO WL/pepfragch.html. or http://thompson.mbt.washington.edu/sequest. 5804

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1998 American Association for Cancer Research. MALDI.PSD REVEALS TUMOR PEPTIDES OF A RCC LINE information by acquisition of an additional MALDI spectrum and/or PSD period at 72Â°Cfor 22 s. The last step was followed by an extension at 72Â°Cfor spectrum of the fully deuterated peptide (26). For further confirmation of a 10 min. The amplification product was visualized by electrophoresis of 5 jiil of proposed sequence, synthetic peptides were compared in the mass spectro- the PCR reaction mixture on a 1.5% agarose gel and subsequently evaluated metrical analysis. Proposed peptide sequences were only accepted for final under UV light (312 nm) after staining with ethidium bromide. experimental approval if the majority of signals of reasonable intensity were Generation of DCs. DCs were generated from peripheral blood monocytes observed in spectra from both the native and the synthetic compound, whereby as described previously (28, 29). In brief, PBMCs were isolated from HLA-B8 absolute ion signal intensities were not taken as strong indicators for a matched positive buffy coat cells by Ficoll/Hypaque (Life Technologies. Inc., Grand Island. NY) density gradient centrifugation. Cells were seeded (1 X IO7 cells/3 comparison. Scores for the reliability of a sequence proposition were determined based ml per well) into six-well microtiter plates (Costar, Cambridge, MA) in RP10 on the ratio of ion abundances of explained signals in the spectrum of the medium (RPMI 1640 supplemented with 10% heat inactivated FCS. 2 mM native peptide and of the total ion abundance expressed in the spectrum L-glutamine. 50 Â¡J.M2-mercaptoethanol. and 50 /ig/ml gentamicin). After 2 h (SCORE_I) and based on the ratio of ion abundances of signals in the spectrum of incubation at 37Â°C,nonadherent cells were removed, and the adherent cells of the synthetic peptide that were not found in the spectrum of the native were cultured in RPK) medium supplemented with the following cytokines: sample, and of the total ion abundance expressed in the spectrum of the human recombinant granulocyte-macrophage colony-stimulating factor (Leu- synthetic peptide (SCORE_2). The overall score was determined as the product komax, Sandoz, NÃ¼rnberg,Germany. 100 ng/ml), interleukin 4 (Genzyme, of SCOREJ and SCOREJ2 and had to be above 0.3 for reliable definition of RÃ¼sselsheim,Germany: 1000 lU/ml), and tumor necrosis factor-a (Genzyme: a peptide sequence. 10 ng/ml). The phenotype of DCs was analyzed by flow cytometry on a FACScan (Becton Dickinson. Mountain View. CA) after 7 days of culture SCORE 1 = ('- using FITC- or phycoerythrin-conjugated mouse mAbs against CD86. CD40 (PharMingen, Hamburg. Germany), CD80, HLA-DR. CD54, CD 14 (Becton Dickinson. Heidelberg. Germany). HLA-A, HLA-B, HLA-C. CD83 (Coulter- SCOREJ1 = 0-*Ã®Â±f) Immunotech, Hamburg. Germany), and CDla (OKT6: Ormo Diagnostic Sys tems. Raritan, NJ). Induction of Antigen-specific CTLs. For CTL induction, 5 x IO5 DCs SCORE_OVERALL = SCORE_\ * SCORE_2 were pulsed with 50 /xg/ml synthetic RAGE peptide for 2 h, washed, and incubated with 2.5 x IO6 autologous PBMCs in RPK) medium. After 7 days In the above formula, / is the relative ion abundance determined from the peak of culture, cells were restimulated with autologous peptide-pulsed PBMCs, and areas in the spectra. In cases where the selectivity of the ion gate was not high enough to select 5 ID/ml human recombinant interleukin 2 (Genzyme, RÃ¼sselsheim,Germany) only one peptide precursor component, we tried to deduce sequence informa were added on days 1, 3, and 5. The cytolytic activity of induced CTLs was analyzed on day 5 after the last restimulation in a standard MCr-release assay. tion from the convoluted PSD spectrum applying also the above criteria for 5lCr-Release Assay. In a standard "Cr-release assay modified as de approval on the rationalized subtracted pattern of ion fragments for each scribed (30), target cells were labeled with [51Cr]-labeled sodium chromate proposed sequence. The resulting SCORE_1 in these cases was considerably (Amersham-Buchler, Braunschweig, Germany) in RPIO for 1 h at 37Â°C.IO4 lowered because only part of the ion signals of the convoluted spectrum could cells/well were transferred to round-bottomed, 96-well plates. Varying num be explained by the sequence of the target component. Synthesis and Purification of Synthetic Peptides. All completely identi bers of CTLs were added to give a final volume of 200 /j.1and incubated for 4 h at 37Â°C.At the end of the assay, supernatants (50 ^I/well) were harvested fied HLA-associated peptides were synthesized by the solid phase method using standard Fmoc chemistry on a MilliGen 9050 PepSynthesizer (Millipore. and counted in a /3-plate counter (Wallac, Turku, Finnland). The percentage of Bedford, MA), purified by reversed-phase HPLC and further validated for specific lysis was calculated as: purity by mass spectrometry. Experimental release - spontaneous release MHC Peptide Stabilization Assay. For an analysis of specific binding of % specific release = the eluted endogenous peptides to HLA-A2, T2 cells (2.5 X 10s) were Maximal release - spontaneous release X 100 incubated with 100 JAMpeptide for 24 h at 37Â°Cin MEM-a medium (Life Technologies, Inc., Grand Island, NY) as described previously (27). Changes Spontaneous and maximal release were determined in the presence of either in surface expression of HLA class I molecules of the T2 cells were evaluated medium or 1% Triton X-100, respectively. To increase the lysis of human by FACS analysis after staining with the mAbs W6/32.HL and TH4 and tumor cells, target cells were incubated for 24 h with 100 lU/ml IFN-y (PBH) expressed as X-fold increase in mean fluorescence intensity (Fl sample/Fl or pulsed with 50 p.g/ml synthetic peptide for 2 h. control) of the T2 cells exogenously loaded with peptide in comparison with the unmanipulated T2 cells. Reverse Transcription-PCR. Total RNA was prepared from different RESULTS RCC lines using the RNeasy-Kit (Qiagen, Hilden. Germany). After DNA Expression and Purification of MHC Class I Peptide Com digestion with DNase I (Boehringer Mannheim) for 20 min at 25Â°C,1-10 /xg plexes of the RCC Line A-498. The human RCC line A-498 con- of mRNA were reverse transcribed with 0.5 jug oligo(dT)|2_18, Moloney stitutively expresses significant amounts of HLA class I proteins that murine leukemia virus-reverse transcriptase and buffers from Life Technolo gies, Inc. (Eggenstein, Germany) at 37Â°Cfor 1 h. RAGE-l cDNA was ampli could be further enhanced by incubation with IFN-y as detected by fied in a 50-fjLÃ•finalPCR reaction mixture that consisted of 5 fil GeneAmp flow cytometry with the mAbs W6/32.HL and TH4 specific for all 10X PCR buffer [100 mM Tris-HCl (pH 8.3), 500 mM KC1, 15 mM MgCU, and HLA-A,B.C or -B.C antigens, respectively (data not shown). IFN-y- 0.01% (w/v) gelatin: Perkin-Elmer, NJ], 5 ^1 of DMSO (Sigma, MÃ¼nchen, stimulated A-498 cells were assumed to exhibit maximized presenta Germany), 0.7 Â¡j.1(0.175 ju.1of each base) deoxynucleotide triphosphate (25 tion of HLA class I-associated tumor peptides. mM each base; Pharmacia, Freiburg, Germany), 2 Â¡JL\ofcDNA solution from HLA class I-peptide complexes, first HLA-B.C antigens and sub reverse transcriptase (150 ng of cDNA), 1 ^1 (concentration, 100 ng//xl) of sequently the remaining HLA class I molecules, were isolated by forward primer A (nucleotide position 467-488 of the RAGE-\ cDNA se immunoaffmity chromatography from the lysate of 10'" A-498 cells. quence 5'-TTCGGGAGTGGTCAGACTGTCG-3'), as well as 1 /xl (concen Six hundred Â¿igof HLA-proteins were purified by the mAb W6/ tration, 100 ng//il) of reverse primer B (nucleotide position 722-743 of the RAGE-l cDNA sequence S'-TCTCAGCAGCAGATCACCCAGG-S'). 0.2 /j.1 32.HL, and 330 jug of HLA-proteins were obtained using the mAb of AmpliTaq DNA polymerase (5 units//xl: Perkin-Elmer), and 35.1 PL\of TH4. HLA-Peptide Separation and Analysis. The two acid-released bidistilled water. PCR amplification was performed in a GeneAmp PCR System 9600 (Perkin-Elmer Cetus Corp., Norwalk, CT) in 32 cycles. After a peptide pools A and B were separated by reversed-phase HPLC in 30 first denaturation step at 94Â°Cfor 1 min, each cycle consisted of a denaturation single peptide peak fractions ( 12 from the peptide pool A and 18 from step at 94Â°Cfor 20 s, an annealing phase at 60Â°Cfor 50 s. and an extension the peptide pool B: Fig. 1). MALDI-MS analysis of the HPLC peaks 5805

Â£ c C o o CM CM CM NM* C oC o

o V) JO

l l l l l i t i i i illllll

Time (min) Time (min) Fig. 1. Reversed-pha.se HPLC separation of the peptide pools A and B eluted from HLA-B.C or the remaining HLA class I molecules of the RCC line A-498. The chromatograms represent the peptide repertoire as detected by UV absorbance at 220 nm. A linear gradient was run for 0-70 min at 0-80% acetonitrile/0.05% TFA.

selection device (see "Materials and Methods"), the MALDI-PSD established that almost all fractions still contained up to 100 different peptides in the mass range from 700 to 1200 Da although frequently technique then provided complete sequence information on 18 pep- with a few dominating ones (a typical MALDI-MS spectrum of one tides from 17 different HPLC fractions of complex peptide mixtures HPLC fraction is shown in Fig. 2). With the use of the precursor ion using the sequence elucidation procedure described above. Three more peptides could be sequenced only for their first four to five amino acids due to insufficient structural information from the MALDI-PSD spectra of the native compounds. Hydrogen-deuterium Â¡Ã• exchange was not useful in these cases because of mass overlap of the deuterated compounds with signals of additional peptides in the HPLC peak fraction or because of too low an entry signal of the precursor ion. MALDI-PSD spectra of the respective synthetic pep tides were generated and compared with the mass "fingerprints" of the original MALDI-PSD of the 18 completely identified peptides. The MIO 1120 1130 sequence proposals of 14 native peptides (4 of peptide pool A and 10 of peptide pool B) could be confirmed by comparison with the synthetic peptides (Table 1). All of them reached an overall score of more than 0.3 with a score 2 not below 0.75. Four additional native peptide peaks had a good match with specific sequences of four source proteins as suggested by database search '"i algorithms. However, confirmatory MALDI-PSD analysis of the cor responding synthetic peptides clearly established that the respective ^Xyu^Jl^^ mass spectra differed considerably from the observed mass signals of the native peptide sequences in all four cases (Table 1). The failure of the analytical method in these cases indicated that peptide sequence Fig. 2. Typical MALDI mass spectrum recorded from a peptide peak collected from the reversed-phase HPLC separation of the peptide pool A at the retention time of 27.20 min. elucidation based solely on database-assisted interpretation of only This shows a dominaling mass peak at m/c = 1099.6 u, which is demonstrated by partially identified fragment ions ["sequence tag" approach (31)] is of comparative analysis of the deuterated sample to contain two compounds of identical precursor ion masses with differing numbers of exchangeable hydrogens (m/z = 1117.6 limited reliability for analysis of complex samples. In all four cases, u and 1121.6 u ). it was found that two precursor peptides were selected at the same 5806

Table I Evaluation of the identified peptides

Peptide"YLKVKGNVFILMEHIHKLFPRLKSKLFLGENISNFLALSDHHIYLYGMPRQILWVKEKVVALDLERKVESLQKDKVAEELQKDKVAELEDOKDHAVFHPKYKTELGLSEFTEYLYLLPAIVHIPASKKTDPQKRSERNNMMNIAQDRNIAIPEDFIKKADSADKLRFRALMW*1066.621132.64987.621152.581067.54976.521070.651087.591058.561058.56958.451014.551057.491037.631098.601150.491012.561148.571088.64SCORE_1C0.961.000.921.000.860.740.880.920.760.630.920.900.910.900.460.500.290.210.48SCORE_2'/1.000.781.000.920.950.781.000.990.810.300.870.921.000.990.820.290.600.650.49OVERALLSCORE'0.960.780.920.920.820.580.880.910.610.190.800.830.910.880.380.150.170.140.23Specialfeatures2nd componentLow signal/noiseSodium peptide2ndadduci in native component,signal/noise2nd low component2" component2"d component"Mutated" EL)OriginalLFA-3 sequence (COOH-terminus LE)SodiumLFA-3 sequence (COOH-terminus adducipeptide2nd in synthetic

component2nd component2nd component2nd component2n component2nd component " Bold sequences, confirmed peptides; thin sequences, disproved peptides. Monoisotopic molecular weight of the neutral molecules. ' Explanation of all peaks observed in the MALDI-PSD spectrum of the native peptide. ' All dominant peaks of synthetic peptide found in the MALDI-PSD spectrum of the native peptide. " Product of SCORE 1 and SCORE 2.

native peptide peptides except for one (LFA3, position 59-67) which showed only a rather low overall score with the database-listed sequence and a considerably higher score with a sequence where two amino acids were exchanged at the COOH terminus. The HLA class I-bound peptides were derived from two groups of precursor proteins. In the first category, six peptides corresponded to sequences from structural or housekeeping proteins present in many normal human cells, such as vimentin (32) or the ribosomal protein L19 (33, 34) or the surface protein LFA-3 (35), or from nuclear proteins involved in regulation of cell growth and proliferation such as the RNA helicase p68 (36) or tristetraproline (37). Three peptides of the L19 (ILMEHIHKL), the p68 helicase (YLLPAIVHI), and the tristetraproline (HPKYKTEL) proteins completely matched previ ously defined sequences bound to HLA-A2 of lymphoblasts (38, 39) or to transfected HLA-B8 molecules of a plasmacytoma cell line (40).

native peptide synthetic peptide Fig. 3. MALDI-PSD spectrum (upper part) recorded from the HPLC fractionated peptide peak (retention time. 28.59 min) of the peptide poo! B after precursor ion selection at mass m/i = 1088.6 u ([M + H]+ ion) in comparison to the spectrum (lower part) recorded from the corresponding synthetic peptide DLERKVESL. Observed ions resulting from regular backbone cleavages are labeled with their fragment type specification for each spectrum. Internal ions and less common fragment-type ions identified in the spectra are not labeled here but were used for spectra interpretation. The sequence identity of the two peptides was proved by this comparison, based on the determined overall score of 0.91. time by the precursor ion gate that were either of the same mass or had a difference of only one or two mass units. An example of a peptide b. 1S2 111 132 523 &U 1ST ess IflU 11.33 from the group of the 14 confirmed sequences of which the original mass spectrum matched the "mass fingerprint" of the corresponding 11S1 Ã•9S Ifii Hi ill US 3Â»S 3M 113 y.

synthetic peptide is shown in Fig. 3, whereas the mass spectrum of IOC ISO 200 ISO 300 350 Â«0 Â«50 500 550 600 650 TOO 750 800 WO 900 9SO 1000 1050 ' another peptide of which the proposed sequence was disproved by a different mass spectrum of the synthetic compound is presented in synthetic peptide Fig. 4. All peptide sequences mentioned in the following met the reliability Fig. 4. MALDI-PSD spectrum (upper part) recorded from the HPLC fractionated peptide peak (retention time, 28.56) of the peptide pool A after precursor ion selection at criteria as described above. mass m/z â€”1151.5 u ([M + H] + ion) in comparison with the spectrum (lower part) Autologous Peptides Bound to the HLA Class I Proteins of the recorded from the synthetic peptide RSERNNMMN corresponding to a sequence prop RCC Line. All of the 14 reliably identified peptide sequences were osition which was based on only partial sequence interpretation (RSERXXXXX), fol lowed by protein database search. Signals were labeled the same way as in Fig. 3. The derived from previously described cellular proteins (Table 2). Protein sequence identity of the two peptides was disproved by this comparison, resulting from a database searches showed 100% sequence similarity for all of these determined overall score of 0.15. 5807

Table 2 Aulologous peplides bound to Â¡heHLA-class I proteins of the RCC line A-498

no."Swiss-ProtSwiss-Pro!Swiss-ProtSwiss-ProtSwiss-ProtSwiss-ProtTR_NEW:Swiss-ProtSwiss-ProtSwiss-ProtSwiss-Prot: PeptideYLKVKGNVFILMEHIHKLQKDKVAEELDQKDHAVFHPKYKTELYLLPA1VHIDLERKVESLFPRLKSKLFLGENISNFLALSDHHIYLYGMPRQILWVKEKVVALGLSEFTEYLPASKKTDPQKPosition124-132137-14559-67244-251139-146148-156218-226129-136242-251115-123192-199175-183225-23344-53Protein60SdistributionWidespread; expressionConstitutive: 9LFA-3HDJ-2,ribosomal protein LI P14118: thatoverexpressbreast tumors overexpressedConstitutiveIndueible:3435463736. her-2/neuWidespreadWidespread

PI9256: DNAJ-2Tristetraproline P31689: rejectedtransplanted(?); human overexpressedInducibleConstitutiveConstitutiveConstitutiveConstitutiveConstitutiveInducibleConstitutiveConstitutive, kidneyWidespreadWidespreadWidespread (TTP)p68 P26651: helicase)VimentinApolipoproteinprotein (RNA PI7844: 5032414243454419 P08670G2425058; cells)PancreasMuscleSmooth(mesenchymal LFructose-bisphosphate-aldolase

A(Aldoa)SM22-alpha P04075:

homologNatural P37802: muscleActivated proteinprecursorkiller cells 4 P24001: TlymphocytesBrain,NK cells and (NK-4)Microtubule-associated IB(MAP protein P46821EMBL: cordRetina,spinal IB)RACE- 1.2.3 ORF-5Accession U46191 U46192 U46193Predominantdifferent tumorsCellular tumorspecificReferences33. " EMBL. European Molecular Biology Lab.

In the second group of peptides with tissue-restricted distribution, two by detailed analysis of the obtained convoluted fragment ion signal sequences first designated to be derived from unknown protein pre patterns of both components. The MALDI-PSD spectra of the native cursors were finally found to exhibit both a 100% match with the peptide mixture and of the synthesized peptide are shown in Fig. 5. It sequence of the only recently described protein apolipoprotein L (41), has to be noted that deconvolution and interpretation of the spectrum which has been claimed to be selectively expressed in the pancreas. of the mixed components could not be performed (or reconstructed) Four additional peptides were derived from antigens normally ex from the printed spectrum as shown in Fig. 5 but only from a thorough pressed in muscle [Aldoa (42), SM22-Alpha-Homologue (43)] or analysis of the mass spectrometrical raw data with computer assist nervous cells [MAP IB (44)] or from a protein [NK-4 (45)], which is ance. Although the amino acid sequence of the major component expressed in activated lymphocytes. Another HLA-A2 binding se could not be established in this study, the minor component was quence was identified as originating from the heat shock protein identified to be the &4GE-derived peptide PASKKTDPQK, the HDJ-2 (46). [M + H]+ ion of which indeed contains 22 exchangeable hydrogens. Analysis of the HLA Binding Specificity. Six of the confirmed Comparison of the MALDI-PSD spectra of the native sample with HLA-associated peptides (ILMEHIHKL, YLLPAIVHI, that of the synthesized peptide PASKKTDPQK demonstrated that DLERKVESL, FLGENISNFL, ALSDHHIYL, and GLSEFTEYL) most of the specific ion signals of the synthetic compound were showed the binding motif characteristic (main anchor positions are represented within the spectrum of the original peptide fraction ac shown in boldface) for association with HLA-A2 (47), another four cording to the intensity ratio of the two peptide components at peptides (YLKVKGNVF, HPKYKTEL, FPRLKSKL, and WVKEKVVAI), derived from TH4- or W6/32.HL-positive HLA antigens, carried characteristic amino acid anchors for binding to native peptide HLA-B8 (47, 48), and the remaining four peptides showed only partial binding motifs for HLA-A2 (QKDKVAEEL) or HLA-B8 (DQKDHAVF, YGMPRQIL, and PASKKTDPQK). All peptides containing the full or partial HLA-A2 but not the HLA-B8 binding motif were demonstrated to elevate HLA class I surface expression of the processing defective T2 cell line as shown by an increase of fluorescence labeling with the mAb W6/32.HL on FACS analysis (data not shown). Identification of an Endogenous Peptide of the Tumor Antigen RAGE Bound to HLA Class I Molecules. The endogenous decameric peptide PASKKTDPQK derived from the first described RCC tumor antigen RAGE (19) was isolated from W6/32.HL-positive HLA class I molecules of the cell line A-498. The respective peptide HPLC b. 91 1Ã•J2ifi 3Â«4512 fill221 SIS 133 1 ion moa sii MI iis sit ni 112 au IÂ«Ty. fraction at a retention time of 27.20 min (see Fig. 1) provided a prominent mass signal at m/z = 1099.6 in the first MALDI-MS analysis. A subsequent hydrogen-deuterium exchange revealed in a second MALDI run that this signal corresponded to at least two synthetic peptide different peptide masses split into the signals at 1117.6 u and 1121.6 u in the spectrum of the deuterated sample (Fig. 2). The MALDI-PSD Fig. 5. MALDI-PSD spectrum (upper pan) recorded from the RACE peplide and from fragment ion spectrum of the signal at [M + H]+ = 1099.6 u thus a second compound that were selected by setting the precursor ion gate to obtain product ions at the precursor ion mass m/- = 1099.6 u (see Fig. 3). Comparative spectrum (lower contained information from two peptide components, one major com pan) recorded from the corresponding synthetic peptide PASKKTDPQK. Signals were ponent containing 18 exchangeable hydrogens and one minor com labeled in the same way as in Fig. 3. The sequence identity of the minor compound of the native sample and of the synthetic peptide was proven by this comparison (overall score. ponent containing 22 exchangeable hydrogens. The sequence of the 0.38). even if deconvolution of the PSD spectra of the two components was critical due /MGÂ£-derived peptide was first established with computer assistance to the lower intensity of the target peptide. 5808

Induction of CTLs with the Synthetic RAGE Peptide. To test the immunogenicity of the identified endogenously presented peptide 60 - PASKKTDPQK of RAGE, PBMCs of a healthy individual carrying HLA-B8 were stimulated with autologous peptide-pulsed DCs in vitro. After two restimulations, the resulting effector cells exhibited strong cytolytic reactivity against the A-498 RCC line as well as against another HLA-B8-positive RCC line Caki-2 but only weak, probably NK cell-related, lytic activity against K562 as well as against the HLA-B8-negative RCC lines MZ1851-RCC and 769-P, which all were shown to express RAGE mRNA (position 467-743 of RAGE-l ; Fig. 6a). The cytolytic activity against the RCC line A-498 could be significantly increased by additional exogenous pulsing of the cell line with 50 jAg/ml RAGE peptide for 2 h (Fig. 6Â¿>).

DISCUSSION In this study, structural characterization of self-peptides bound to HLA class I molecules of an RCC line was first achieved by the MALDI-PSD technique. This technique was shown to provide a powerful and sensitive alternative tool to ESI-MS-MS for reliable identification of MHC-associated peptides, even if provided in com plex peptide mixtures. Among 14 unequivocally identified HLA-A2 or -B8 bound self-peptides, several sequences within the repertoire of 70 endogenous peptides were characterized that were derived from pro 60 teins with widespread expression in normal tissues. Some of these had been reported previously to be specifically overexpressed in diseased 50 renal tissue or tumors. In addition, a new antigenic sequence of the recently described first tumor-specific antigen RAGE of RCC was 40 shown to also belong to the endogenous HLA class I-associated 30 - peptide pool of the RCC line A-498 and is probably presented by HLA-B8. Both these groups of peptides, either derived from overex 20 pressed but otherwise ubiquitous antigens or from the tumor-specific antigen RAGE, may potentially serve as prime target structures of 10 - HLA-A2- or -B8-restricted therapeutic T cell autoreactivity in RCC. MALDI-PSD has been introduced recently as one of the potentially 10 15 20 25 30 most powerful methods of mass spectrometry for highly sensitive and E:T fast molecular structural analysis of peptides. In contrast to Edman Fig. 6. In vitro induction of CTLs in PBMCs of a healthy individual by stimulation degradation, separation of analyte molecules in a crystal lattice of with autologous dendritic cells pulsed with the synthetic RAGE peptide. a, a 4-5-fold dihydroxybenzoic acid as embedding matrix as well as the use of an increased cytotoxic reactivity of effector cells obtained after two restimulations is shown in a standard cytotoxicity assay against the RACE- and HLA-B8-positive RCC lines electronic precursor ion selection device allowed us to sequence A-498 (X) and Caki-2 (â€¢)versus the /MGE-positive. HLA-B8-negative RCC lines 769-P (O) and MZ1851-RCC (A), which revealed lysis similar to the HLA class I-negative cell line K562 (â€¢)used as a control for NK cell activity, b, the in vitro sensitized CTLs showed tumor cell lysis of the RCC line A-498 (â€¢),which was further increased by prior MELPKLKLSGWRLSSYSSPTLQSVLGSGTNGRVPVLRPLKCIPASKKTDPQKDLKP incubation of the target cells with 50 fig/ml RAGE â€¢ peptide for 2 h. APQQCRLPTIVRKGGR (this study) MSSAHPLRRSSPSSNRIRNTSTNNQFVPTMPLPPARNGGL (Gaugleretal.) [M + H]+ = 1099.6 u (Fig. 5). Because the identified peptide did not 1000 bind to HLA-A2 of the processing defective T2 cell line, presentation RAGE-1 â€”. 1118bp in association to HLA-B8 appeared to be most likely, although the peptide only partially matched characteristic anchor positions for this HLA-B alÃele(47, 48). RAGE-2 1168bp Analysis of the Expression of the Tumor Antigen RAGE. To further prove expression of RAGE, the RCC line A-498 was investi gated by RT-PCR for transcription of specific mRNA containing the RAGE-3 1235bp coding region of the identified peptide. Using primers flanking posi a Possibly translated ORF tion 467 to 743 of the published cDNA sequences of RAGE-l (19), <â€¢â€¢/.Probablynot translated ORF Sequences unrelated to the three fragments of 277, 367, and about 450 bp were amplified and RAGE-1 sequence shown by direct nonradioactive sequencing to correspond to RAGE- 1-3 cDNA sequences (19).5 Three other RCC lines tested, Caki-2, Fig. 7. Schematic representation of the previously identified RACE-\-i cDNA se quences. Closed black boxes indicate the position of the ORF5-dependent coding region 786-P, and MZ-1851-RCC, also expressed the same RAGE-1,2,3 for the antigenic peptide characterized in this study as well as of the previously defined mRNA species in our PCR assay (data not shown). ORF2-dependent transcribed antigenic peptide. The shadowed areas in RAGE-l-3 indi cate the most likely transcribed ORF2, ORF3, and ORF5 in each of the three reading frames. Sequences not related to the RAGE-l sequences as well as additional ORFs 5 B. Van den Eynde. unpublished results. probably not transcribed are indicated by differentially marked boxes. 5809

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1998 American Association for Cancer Research. MALDI-PSD REVEALS TUMOR PEPTIDES OF A RCC LINE single peptides in mixtures without further purification steps. Also, as an aberrantly present antigenic determinant in RCC. Like HLA- possibly minor contaminants of detergents in the peptide samples A2-restricted peptides of her-2/neu, such peptides could fulfill the from the HPLC separation appeared to be tolerated. The total amount requirements of surface expression above a certain threshold neces of peptide needed to be prepared on the target of the MALDI-PSD sary for T-cell recognition in malignant but not normal cells and might time-of-flight mass spectrometer is in general only in the range of be explored for in vitro priming of tumor-specific cytotoxic T cells 30-100 fmol. For all of the MHC-associated peptides elucidated in (52-54). All peptides identified by MALDI-PSD except the charac this study, the amount of prepared sample was estimated to be in the terized sequence from LFA-3 were found to completely match the range of 100 fmol to 1.5 pmol. Only a few percentages of the sample original sequences represented in the protein databases. Sequencing of were actually used up by acquiring a complete PSD spectrum. In the coding region of the latter peptide after specific amplification is contrast to ESI-MS-MS. the offline character of MALDI-PSD there warranted to further document this mutation of the original protein fore allowed us to investigate the prepared sample multiple times, e.g., LFA-3 as a tumor-specific antigen of the RCC line A-498. Point for reconfirming a sequence proposition by means of a more detailed mutations of proteins represented as antigenic MHC-bound peptides fragmentation study, or to perform on-target hydrogen-deuterium have been observed only in a few human tumors (55, 56). exchange or derivation procedures with the already analyzed sample In RCCs, only one antigen RAGE has been identified as a tumor- to achieve more structural information on a peptide. Although specific peptide antigen that is expressed in a small subpopulation of MALDI-PSD provided mass spectra based on N-, C- and high internal renal tumors and can be recognized by T cells. Four different mRNA ion fragmentation similarly to ESI-MS-MS, combined and computer species, RAGE-\-4 related to RAGE, have been identified and shown ized interpretations of the mass spectra seemed to be less complicated to contain up to five different ORFs. There is still uncertainty about because of the lack of additional signals from different and multiple- which of these ORFs is functional (19). The previously reported charged product particles of the same precursor ions. HLA-B7-restricted T-cell epitope has been suggested to be translated Nevertheless, routine comparison of the observed fragment ion from the ORF2 of the RAGE-l cDNA (19). However, the HLA class spectra of the native peptide sample with those of the deduced I-bound RAGE peptide of our study was found to originate from corresponding synthetic sequences was found to be a useful and ORF5 with the use of the same reading phase as ORF2. In contrast to necessary procedure to confirm first interpretations of mass spectra of ORF2, ORF5 has been shown also to be present in the RAGE-2,3 selected precursor ions. Four peptide sequences deduced solely on mRNA species (Fig. 7). Analysis of RAGE mRNA expression in three manually obtained incomplete sequence information and protein da further RCC lines indicated that the expression of RAGE-1,2,3 mRNA tabase searches of the recorded mass peaks had to be discarded when from ORF5 is less restricted in RCC than the expression of RAGE compared with the respective synthetic peptide. We found that in the containing ORF2.6 In vitro priming of peripheral blood lymphocytes case of the MHC-associated peptide samples, it was not sufficient to showed that the newly identified RAGE peptide can be applied to rely only on results of protein database searches based on partially induce cytotoxic T cells for specific lysis of HLA-B8-positive RCC (31) or completely noninterpreted mass spectra (49). A full interpre lines. Thus, this peptide may be used to stimulate lymphocytes from tation of the spectrum and a complete rationalization of the sequence healthy individuals to generate tumor-specific CTLs against RCC by proposition in a second step of peptide analysis was mandatory. primary in vitro immunization. Additional comparison with a synthetic peptide proved to be a very In summary, this study shows that direct sequencing of HLA- helpful confirmatory test for correct peptide sequencing. Without such associated peptides by MALDI-PSD allows the identification of nat further proof, the reliability of sequence propositions from database urally processed tumor antigens encoded by genes preferentially ac search programs was insufficient, particularly if two overlapping tivated in malignant tissue as well as resulting from mutations. The MALDI-PSD spectra of fragments from ion precursors of closely new directly defined HLA-bound peptides of the RCC line A-498 related molecular masses, which could not be separated by the ion extend the number of potential tumor-specific antigenic determinants gate, had to be deconvoluted. This observation questions the general of RCC for immunotherapeutic strategies and particularly establish reliability of solely database-proposed sequences of mass spectro- another sequence of RAGE as an endogenously processed and HLA metrical analysis, particularly if only short partial sequences can be class I-presented antigen of RCC lines. directly specified from the observed ion fragmentation and applied to database searches. This situation may be not unique to MALDI-PSD but of relevance for mass spectrometric peptide sequencing in general. ACKNOWLEDGMENTS Thus, MALDI-PSD like other mass spectrometrical approaches Our deepest appreciation is directed to one of the authors. Raimund Kauf offers the advantage of directly identifying naturally processed mann. Professor Dr. Raimund Kaufmann, born February 14. 1934, died Sep epitopes of proteins that are actually presented at the malignant cell tember 1, 1997, after a long illness. He gave considerable impetus to the surface and may be specifically recognized by cytotoxic T cells. By initializing steps of the project and contributed profoundly to the elucidation of using HPLC purification and sequencing by Edman degradation or MHC peptide sequences. We feel great sorrow at the loss of an innovative mass spectrometry, HLA-associated tumor peptides have been iden scientist who made a significant contribution to furthering the field of mass tified previously on melanomas (18) and on chronic myeloid leukemia spectrometry and laser medicine. blasts (16) in humans as well as on an UV light-induced murine tumors (50). In most previous analyses, as well as in our study of tumor cells, a large fraction of the identified HLA-associated peptides REFERENCES was found to be processed from high copy number proteins with 1. Gross. L. Intradermal immunization of C3H mice against a sarcoma that originated in an animal of the same line. Cancer Res.. 3: 326. 1943. widespread but low expression also in normal tissue. This particularly 2. Pardoll. D. M. Cancer vaccines: a road map for the next decade. Curr. Opin. applies for the ribosomal protein L19, which has been shown to be Immunol.. 8: 619-621. 1996. overexpressed in RCCs in association with increased levels of the 3. Roth. C.. Rochlitz. C., and Kourilsky, P. Immune response against tumors. Adv. Immunol.. 57: 281-350. 1994. her-2/neu proto-oncogene and was correlated with tumor metastasis 4. van der Bruggen. P.. Traversali. C.. ChÃ´mez. P.. Lurquin. C.. De Plaen. E.. Van den (51). Also, the identified naturally processed peptide of HDJ-2, a Eynde. B.. Knuth. A., and Boon, T. A gene encoding an antigen recognized by previously reported member of the heat shock protein-40 family with absent or very low expression in normal renal tissue (46), could serve 1Unpublished observation, manuscript in preparation. 5810

cylolytic T lymphocytes on a human melanoma. Science (Washington DC), 254: 31 Mann, M., and Wilm. M. Error-tolerant identification of peptides in sequence data 1643-1647, 1991. bases by peptide sequence tags. Anal. Chem., 66: 4390-4399, 1994. 5. Traversali, C., van der Bruggen, P., Luescher, 1. F., Lurquin, C, ChÃ´mez, P., Van Pel, 32 Sabo, E., Miselevich, 1., Bejar, J., Segenreich, M., Wald, M., Moskovitz, B., and A., De Plaen, E., Amar-Costesec. A., and Boon, T. A nonapeptide encoded by human Nativ, O. The role of vimentin expression in predicting the long-term outcome of gene MAGE-1 is recognized on HLA-A1 by cytolytic T lymphocytes directed against patients with localized renal cell carcinoma. Br. J. Urol., 80: 864-868, 1997. tumor antigen MZ2-E. J. Exp. Med., 176: 1453-1457, 1992. 33 Henry, J. L., Coggin, D. L., and King, C. R. High-level expression of the ribosomal 6. Van den Eynde, B., Peelers, O., De Backer, O., Gaugler, B., Lucas, S., and Boon, T. protein L19 in human breast tumors that overexpress erbB-2. Cancer Res., 53: A new family of genes coding for an antigen recognized by autologous cytolytic T 1403-1408, 1993. lymphocytes on a human melanoma. J. Exp. Med., 182: 689-698, 1995. 34 Davies, B., and Fried, M. The L19 ribosomal protein gene (RPL19): gene organization, 7. Traversali, C., van der Bruggen, P., Van den Eynde, B., Hainaut, P., Lemoine, C., chromosomal mapping, and novel promoter region. Genomics, 25: 372-380, 1995. Onta, N., Old, L., and Boon, T. Transfection and expression of a gene coding for a 35. Wallner, B. P., Frey, A. Z., Tizard, R., Mattaliano, R. J., Hession, C., Sanders, M. E., human melanoma antigen recognized by autologous cytolytic T lymphocytes. Immu- Dustin, M. L., and Springer. T. A. Primary structure of lymphocyte function- nogenetics, 35: 145-152, 1992. associated antigen 3 (LFA-3). The ligand of the T lymphocyte CD2 glycoprotein. J. 8. Maeurer, M. J., Storkus, W. J., Kirkwood, J. M., and Lotze, M. T. New treatment Exp. Med., 166: 923-932, 1987. options for patients with melanoma: review of melanoma-derived T-cell epitope- 36 Iggo, R. D., and Lane, D. P. Nuclear protein p68 is an RNA-dependent ATPase. based peptide vaccines. Melanoma Res.. 6: 11-24, 1996. EMBOJ., 8: 1827-1831, 1989. 9. Salvadori, S., and Zier, K. Molecular basis of T cell dysfunction in cancer is 37 Taylor, G. A., Carballo, E., Lee, D. M., Lai, W. S., Thompson, M. J., Patel, D. D., influenced by the paracrine secretion of tumor-derived IL-2. J. Immunol., 756: Schenkman, D. I., Gilkeson, G. S., Broxmeyer, H. E., Haynes. B. F., and Blackshear, 2927-2932, 1996. P. J. A pathogenetic role for TNF a in the syndrome of cachexia, arthritis, and 10. Kolenko, V., Wang, Q., Riedy, M. C., O'Shea, J., Ritz, J.. Cathcart, M. K., Rayman, autoimmunily resulting from tristetraproline (TTP) deficiency. Immunity. 4: 445- P., Tubbs, R., Edinger, M., Novick, A., Bukowski, R., and Finke, J. Tumor-induced 454, 1996. suppression of T lymphocyte proliferation coincides with inhibition of Jak3 exprÃ¨s- 38. Barouch, D., Friede, T., Stevanovic, S., Tussey. L., Smith, K., Rowland-Jones, S., sion and IL-2 receptor signaling: role of soluble products from human renal cell Braud, V., Mi-Michael. A., and Rammensee, H. G. HLA-A2 subtypes are functionally carcinomas. J. Immunol., 159: 3057-3067, 1997. distinct in peptide binding and presentation. J. Exp. Med., 182: 1847-1856, 1995. 11. Enk, A. H., Jonuleit, H., Saloga. J., and Knop, J. Dendritic cells as mediators of 39 Hunt, D. F., Henderson, R. A., Shabanowitz, J.. Sakaguchi, K., Michel, H.. Sevilir. N., tumor-induced tolerance in metastatic melanoma. Int. J. Cancer, 73: 309-316, 1997. Cox, A. L., Appella, E., and Engelhard, V. H. Characterization of peptides bound to 12. Engelhard, V. H. Direct identification of tumor associated peptide antigens. Springer the class I MHC molecule HLA-A2.1 by mass spectrometry. Science (Washington Semin. Immunopathol., 18: 171-183, 1996. DC), 255: 1261-1263, 1992. 13. Roetzschke, O., Falk, K., Wallny, H. J., Faath, S., and Rammensee, H. G. Charac- 40. DiBrino, M., Parker, K. C., Shiloach, J., Turner, R. V.. Tsuchida, T., Garfield, M., terization of naturally occurring minor histocompatibility peptides including H-4 and Biddison, W. E., and Coligan, J. E. Endogenous peptides with distinct amino acid anchor H-Y. Science (Washington DC), 249: 283-287, 1990. residue motifs bind to HLA-AI and HLA-B8. J. Immunol., /52: 620-631, 1994. 14. Van Bleek, G., and Nathanson, S. Isolation of an endogenously processed immu- 41. Duchateau, P. N., Pullinger, C. R., Orellana, R. E., Kunitake. S. T.. Naya-Vigne, J., nodominant viral peptide from the class I H-2Kb molecule. Nature (Lond.), 348: O'Connor, P. M., Malloy, M. J., and Kane, J. P. Apolipoprotein L, a new human high 213-216, 1990. density lipoprotein apolipoprotein expressed by the pancreas. Identification, cloning, 15. Mandelboim, O., Berke, G., Fridkin, M., Feldman, M., Eisenstein, M., and Eisenbach, characterization, and plasma distribution of apolipoprotein L. J. Biol. Chem., 272: L. CTL induction by a tumor-associated antigen octapeptide derived from a murine 25576-25582, 1997. lung carcinoma. Nature (Lond.), 369: 67-71, 1994. 42. Izzo, P., Costanzo, P., Lupo, A., Rippa, E., Paolella, G., and Salvatore, F. Human 16. Papadopoulos, K. P., Suicu-Foca, N., Hesdorffer, C. S., Tugulea, S., Maffei, A., and aldolase A gene. Structural organization and tissue-specific expression by multiple Harris, P. E. Naturally processed tissue- and differentiation stage-specific autologous promoters and alternate mRNA processing. Eur. J. Biochem., 774: 569-578, 1988. peptides bound by HLA class I and II molecules of chronic myeloid leukemia blasts. 43. Nagase, T., Miyajima, N., Tanaka, A., Sazuka, T., Seki, N., Sato, S., Tabata, S.. Blood, 90: 4938-4946, 1997. Ishikawa, K., Kawarabayasi, Y., and Kotani, H. Prediction of the coding sequences of 17. Henderson, R. A., Cox, A. L., Sakaguchi, K., Appella, E., Shabanowitz, J., Hunt, unidentified human genes. III. The coding sequences of 40 new genes (KIAA0081- D. F., and Engelhard, V. H. Direct identification of an endogenous peptide recognized KIAA0120) deduced by analysis of cDNA clones from human cell line KG-1. DNA by multiple HLA-A2.1 specific cytotoxic T cells. Proc. Nati. Acad. Sci. USA, 90: Res., 2: 37-43, 1995. 10275-10279, 1993. 44. Lien, L. L., Feener, C. A., Fischbach, N., and Kunkel. L. M. Cloning of human 18. Cox, A. L., Skipper, J., Chen, Y.. Henderson, R. A., Darrow, T. L., Shabanowitz, J., microtubule-associated protein 1B and the identification of a related gene on chro Engelhard, V. H., Hunt, D. F., and Slingluff, C. L., Jr. Identification of a peptide mosome 15. Genomics, 22: 273-280, 1994. recognized by five melanoma-specific human cytotoxic T cell lines. Science (Wash- 45. Dahl, C. A., Schall, R. P., He, H. L., and Cairns, J. S. Identification of a novel gene ington DC), 264: 716-719, 1994. expressed in activated natural killer cells and T cells. J. Immunol., 748: 597-603, 1992. 19. Gaugler, B., Brouwenstijn, N., Vantomme, V., Szikora, J. P., Van der Spek, C. W., 46. Alevy, Y. G., Brennan, D., Durriya, S., Howard, T., and Mohanakumar, T. Increased Patard, J. J., Boon, T., Schlier, P., and Van den Eynde, B. J. A new gene coding for expression of the HDJ-2 heat shock protein in biopsies of human rejected kidney. an antigen recognized by autologous cytolytic T lymphocytes on a human renal Transplantation (Baltimore), 61: 963-961. 1996. carcinoma. Immunogenetics, 44: 323-330, 1996. 47. Rammensee, H. G., Friede, T., and Stevanovic, S. MHC ligands and peptide motifs: 20. Bamstable, C. J., Bodmer, W. F., Brown, G.. GalfrÃ©,G.,Milstein, C., Williams, A. F., and first listing. Immunogenetics, 41: 178-228, 1995. Ziegler, A. Production of monoclonal antibodies to group A erythrocytes, HLA, and other 48. Sutton, J., Rowland-Jones, S., Rosenberg, W., Nixon. D., Gotch, F., Gao, X. M., Murray, human cell surface antigens -new tools for genetic analysis. Cell, 14: 9-20, 1978. N., Spoonas, A., Driscoll, P., Smith, M., Willis, A., and McMichael, A. A sequence 21. Reuss-Borst, M. A., BUhring, H. J., Schmidt, H., and MÃ¼ller,C. A. AML. Immuno- pattern for peptides presented to cytotoxic T lymphocytes by HLA B8 revealed by phenotypic heterogeneity and prognostic significance of c-kit expression. Leukemia analysis of epitopes and eluted peptides. Eur. J. Immunol., 23: 447-453, 1993. (Baltimore), 8: 258-263, 1994. 49. Yates, J. R., Eng, J. K., McCormack, A. L., and Schieltz, D. Method to correlate 22. Kaufmann, R., Spengler, B., and Liitzenkirchen, F. Mass spectrometric sequencing of tandem mass spectra of modified peptides to amino acid sequences in the protein linear peptides by product-ion analysis in a reflectron time-of-flight mass spectrom database. Anal. Chem., 67: 1426-1436, 1995. eter using matrix-assisted laser desorption ionization. Rapid Commun. Mass Spec- 50. Dubey, P., Hendrickson, R. C., Meredith, S. C., Siegel, C. T., Shabanowitz, J., trom., 7: 902-910, 1993. Skipper, J. C., Engelhard, V. H., Hunt, D. F., and Schreiber, H. The immunodominant 23. Kaufmann, R., Kirsch, D., and Spengler, B. Sequencing of peptides in a time-of-flight antigen of an ultraviolet-induced regressor tumor is generated by a somatic point mass spectrometer: evaluation of postsource decay following matrix-assisted laser mutation in the DEAD box helicase p68. J. Exp. Med., 785: 695-705, 1997. desorption ionization (MALDI). Int. J. Mass Spectrom. Ion Proc., 131: 355-385, 1994. 51. Stumm, G., Eberwein, S., Rostock-Wolf, S., Stein, H., Pomer, S., Schlegel, Jâ€žand 24. Kaufmann, R., Chaurand, P., Kirsch, D., and Spengler, B. Post-source decay and Waldherr, R. Concomitant overexpression of the EGFK and erbB-2 genes in renal cell delayed extraction in matrix-assisted laser desorption/ionization-reflectron time-of- carcinoma (RCC) is correlated with dedifferentiation and metastasis. Int. J. Cancer, flight mass spectrometry. Are there trade-offs? Rapid Commun. Mass Spectrom., 10: 69: 17-22, 19%. 1199-1208. 1996. 52. Peoples, G. E.. Goedegebuure, P. S., Smith, R., Linehan. D. C., Yoshino, I., and 25. Spengler. B. Post-source decay analysis in matrix-assisted laser desorption/ionization Eberlein, T. J. Breast and ovarian cancer-specific cytotoxic T lymphocytes recognize mass spectrometry of biomolecules. J. Mass Spectrom.. 32: 1019-1036, 1997. the same HER2/neu-derived peptide. Proc. Nati. Acad. Sci. USA, 92: 432-436. 1995. 26. Spengler, B.. Liitzenkirchen, F., and Kaufmann, R. On-target deuteralion for peptide 53. Fisk, B., Blevins, T. L., Wharton, J. T., and loannides, C. G. Identification of an sequencing by laser mass spectrometry. Org. Mass Spectrom., 28: 1482-1490. 1993. immunodominant peptide of her-2/neu protooncogene recognized by ovarian tumor- 27. Baas, E. J., van Samen, H. M.. Kleijmeer, M. J., Geuze, H. J., Peters, P. J., and specific cytotoxic T lymphocyte lines. J. Exp. Med., 787: 2109-2117, 1995. Ploegh, H. L. Peptide-induced stabilization and intracellular localization of empty 54. Brossart, P., Stuhler, G., Fiad. T., Stevanovic, S., Rammensee, H. G., Kanz, L., and HLA class I complexes. J. Exp. Med., 176: 147-156, 1992. Brugger. W. Her-2/neu-derived peptides are tumor-associated antigens expressed by 28. Sallusto, F., and Lanzavecchia, A. Efficient presentation of soluble antigen by human renal cell and colon carcinoma lines and are recognized by in vitro induced cultured human dendritic cells is maintained by granulocyte/macrophage colony specific cytotoxic T lymphocytes. Cancer Res., 58: 732-736, 1998. stimulating factor plus interleukin 4 and down regulated by tumor necrosis factor a. 55. Robbins, P. F., El-Gamil, M., Li, Y. F., Kawakami, Y., Loftus, D., Appella, E., and J. Exp. Med., 179: 1109-1118, 1994. Rosenberg. S. A. A mutated ÃŸ-cateningene encodes a melanoma-specific antigen 29. Zhou, L. J., and Tedder. T. F. CD14+ blood monocytes can differentiate into recognized by tumor infiltrating lymphocytes. J. Exp. Med., 183: 1185-1192, 1996. functionally mature CD83+ dendritic cells. Proc. Nati. Acad. Sci. USA, 93: 2588- 56. Wolfel, T., Hauer. M., Schneider, J., Serrano, M., Wolfel. C.. Klehmann-Hieb, E., De 2592, 1996. Plaen, E., Hankeln, T., Meyer zum Buschenfelde, K. H., and Beach, D. A pl61NK4a- 30. Brossart. P., and Bevan, M. J. Selective activation of Fas/FasL mediated cytotoxicity insensitive CDK4 mutant targeted by cytolytic T lymphocytes in a human melanoma. by a self-peptide. J. Exp. Med.. 183: 2449-2458, 1996. Science (Washington DC), 269: 1281-1284, 1995.

5811

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1998 American Association for Cancer Research. Direct Identification of Major Histocompatibility Complex Class I-bound Tumor-associated Peptide Antigens of a Renal Carcinoma Cell Line by a Novel Mass Spectrometric Method

Thomas Flad, Bernhard Spengler, Hubert Kalbacher, et al.

Cancer Res 1998;58:5803-5811.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/58/24/5803

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/58/24/5803. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.