- Email: [email protected]
Identification of a shared epitope recognized by melanoma-specific, HLA-A3-restricted cytotoxic T lymphocytes
Immunology Letters 90 (2003) 131–135
Identification of a shared epitope recognized by melanoma-specific, HLA-A3-restricted cytotoxic T lymphocytes Kevin T. Hogan a,b,∗ , Michael A. Coppola b,1 , Christine L. Gatlin b,2 , Lee W. Thompson a , Jeffrey Shabanowitz c , Donald F. Hunt c,d , Victor H. Engelhard e , Craig L. Slingluff Jr. a , Mark M. Ross b,3 a
Department of Surgery, University of Virginia, Box 801359, Charlottesville, VA 22908, USA b Argonex Inc., Charlottesville, VA 22903, USA c Department of Chemistry, University of Virginia, Charlottesville, VA 22908, USA d Department of Pathology, University of Virginia, Charlottesville, VA 22908, USA e Department of Microbiology, University of Virginia, Charlottesville, VA 22908, USA
Received 8 July 2003; received in revised form 26 August 2003; accepted 26 August 2003
Abstract We previously established a melanoma-reactive cytotoxic T lymphocyte (CTL) line that recognizes multiple epitopes in the context of HLA-A3. To increase the number of peptides available for use in a vaccine for the treatment of melanoma, we identified one of these epitopes, SQNFPGSQK, through a combination of epitope reconstitution experiments and mass spectrometry. The SQNFPGSQK peptide was also found to be associated with HLA-A3 on an additional melanoma tumor line, thus indicating that the peptide is not unique to the melanoma tumor line from which it was isolated and thus, unlikely to arise through a mutational event. Although the protein origin of SQNFPGSQK has yet to be established, the shared nature of this epitope and the fact that it elicits a natural immune response indicates that it warrants further study to determine its usefulness as a vaccine component for the treatment of melanoma. The peptide may also be useful as a research tool for evaluating spontaneous anti-tumor immune responses in patients with melanoma. © 2003 Elsevier B.V. All rights reserved. Keywords: Immunotherapy; Cancer; Antigen; Epitope; Cytotoxic T lymphocyte; Mass spectrometry
1. Introduction Significant progress has been made in identifying peptide epitopes that can stimulate a melanoma-specific cytotoxic T lymphocyte (CTL) response [1–3]. The tumor antigens giving rise to these peptides include cancer/testis antigens, differentiation antigens, and mutated gene products [1–3]. A number of these peptides have been tested in clinical trials, with some patients showing partial or complete tumor regressions [1,2]. These initial results indicate that a thera∗ Corresponding author. Tel.: +1-434-243-9861; fax: +1-434-924-8464. E-mail address: [email protected] (K.T. Hogan). 1 Present address: Department of Cell Biology, University of Virginia, Charlottesville, VA 22908, USA. 2 Present address: The Institute for Genomic Research, Rockville, MD 20850, USA. 3 Present address: MDS Proteomics, Charlottesville, VA 22911, USA.
0165-2478/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.imlet.2003.08.003
peutic melanoma vaccine may prove efficacious in the treatment of the disease, and thus further development of such a vaccine is warranted. One limitation of peptide-based vaccines is the fact that they are only useful for the treatment of individuals that express one or more of the class I major histocompatibility complex (MHC)-encoded molecules to which the peptides are capable of binding. Most of the melanoma peptides that have been identified are restricted to HLA-A2, and thus there is a need to identify peptides that bind to other class I MHC molecules so as to provide broad-based population coverage. Another limitation that must be overcome is that tumors can undergo immunoselection and lose the expression of a given antigen or class I MHC molecule, thus preventing recognition of the tumor by the corresponding CTL [3–6]. This limitation can be overcome by including in the vaccine peptides that are derived from multiple tumor antigens, as well as peptides that bind to different class I MHC
132
K.T. Hogan et al. / Immunology Letters 90 (2003) 131–135
molecules. Therefore, we have continued to pursue the identification of additional melanoma antigens. We previously described a melanoma-specific, HLA-A3restricted CTL line that recognizes a Pmel17/gp100-derived peptide [7]. This CTL line recognizes two additional epitopes, one of which has been recently identified.4 In the present report, we identify the third antigen which is seen by this CTL in the context of HLA-A3.
2. Materials and methods
absorbance detector. The column was made by packing a 27 cm bed of 10 m C18 particles in a section of 285 m o.d. × 75 m i.d. fused silica. Peptides in a selected second dimension fraction were loaded onto this column and eluted with a gradient of acetonitrile/0.67% triethylamine acetate (TEAA)/water in 0.1% triethylamine acetate/water, with the concentration of acetonitrile increasing from 0 to 60% in 40 min. The flow rate was approximately 300 nl/min, and fractions were collected into 25 l of 0.1% acetic acid every 30 s. In all RP-HPLC experiments, peptides were detected by monitoring UV absorbance at 214 nm.
2.1. Cell lines
2.5. CTL epitope reconstitution assay
The melanoma lines VMM12 and VMM18 were maintained in RPMI1640 supplemented with 5–10% FBS and 2 mM l-glutamine. C1R-A3 was maintained in the same media supplemented with 200 g/ml G418.
Aliquots of each RP-HPLC fraction were tested for the presence of peptides that could sensitize C1R-A3 targets for lysis by VMM18 CTL in a standard 4 h 51 Cr-release assay as previously described [9].
2.2. CTL line
2.6. Mass spectrometric analyses
VMM18-specific CTL have been described previously [7]. CTL were expanded in bulk culture using anti-CD3 antibody [8] and cryopreserved in aliquots of (1–5) × 107 cells for use in epitope reconstitution assays.
Active RP-HPLC fractions were screened by on-line RP-HPLC/electrospray ionization mass spectrometry (MS) using a homemade microcapillary column and a Finnigan-MAT TSQ 7000 triple quadrupole mass spectrometer (Finnigan, San Jose, CA). Approximately, 1% of the active RP-HPLC fraction was loaded onto a section of 185 m o.d. × 75 m i.d. fused silica packed with 10–12 cm of 10 m C18 particles. Peptides were eluted directly into the mass spectrometer using a 10-min 0–60% acetonitrile in 0.1 M acetic acid gradient. Ions were formed by electrospray ionization, and mass spectra were recorded by scanning between mass to charge ratios (m/z) 300 and 1400 every 1.5 s. Active second dimension RP-HPLC fractions were analyzed using an effluent splitter on the microcapillary HPLC column. The column (360 m o.d. × 100 m i.d. with a 25-cm C18 bed) was connected with a zero dead volume tee (Valco, Houston, TX) to two pieces of fused silica of different lengths (25 and 40 m i.d.). Peptides were eluted with a 34 min gradient of 0–60% acetonitrile in 0.1 M acetic acid. The 25 m capillary deposited one-fifth of the RP-HPLC effluent into the wells of a microtiter plate for use in a CTL epitope reconstitution assay, while the remaining four-fifths of the effluent was directed into the mass spectrometer, with mass spectra recorded as described above [9]. Peptide sequences were determined by collision-activated dissociation (CAD) tandem mass spectrometry using an LCQ (Finnigan) ion trap mass spectrometer and methods as described [10,11].
2.3. Isolation of HLA-A3 associated peptides HLA-A3 molecules were immunoaffinity purified from aliquots of (6–8) × 1010 VMM18 tumor cells as described [9] using HLA-A3-specific monoclonal antibody GAP-A3 bound to protein A-Sepharose. 2.4. Peptide fractionation Peptide extracts were fractionated by reverse phase-high performance liquid chromatography (RP-HPLC) using an Applied Biosystems model 140B system. The extracts were concentrated by vacuum centrifugation and injected onto a Higgins (Mountain View, CA) C18 HAISIL column (2.1 mm × 4 cm, 300 Å, 5 m). The peptides were eluted with a gradient of acetonitrile/0.085% trifluoroacetic acid (TFA) in 0.1% TFA/water, with the concentration of acetonitrile increasing from 0 to 9% (0 to 5 min), 9 to 36% (5 to 55 min), and 36 to 60% (55 to 62 min). Second dimension fractionations of selected first dimension (TFA) fractions were accomplished using the same gradient but with the substitution of heptafluorobutyric acid (HFBA) for TFA. A third dimension of RP-HPLC was achieved using an Eldex (Napa, CA) MicroPro pump, a homemade C18 microcapillary column and an Applied Biosystems model 785A UV 4
K.T. Hogan, M.A. Coppola, C.L. Gatlin, L.W. Thompson, J. Shabanowitz, D.F. Hunt, V. H. Engelhard, M.M. Ross, C.L. Slingluff, Jr. Identification of novel and widely expressed cancer/testis gene isoforms that elicit spontaneous CTL reactivity to melanoma, submitted for publication.
2.7. Peptide synthesis Peptides were synthesized using a Gilson (Madison, WI) AMS 422 multiple peptide synthesizer using conventional FMOC chemistry. Peptides were purified by RP-HPLC using a 4.6 mm i.d. × 100 mm long POROS (Perseptive
K.T. Hogan et al. / Immunology Letters 90 (2003) 131–135
133
Biosystems, Cambridge, MA) column and a 10-min 0–60% acetonitrile in 0.1% TFA gradient. 2.8. Pmel17/gp100 sequencing
2.9. Human subjects All research involving human subjects was approved by the University of Virginia Human Investigation Committee.
3. Results 3.1. VMM18 CTL recognize three distinct HLA-A3 restricted epitopes The peptides bound to HLA-A3 molecules on 8 × 1010 VMM18 tumor cells were purified as described in Section 2, and fractionated by RP-HPLC using TFA as the organic modifier. A CTL epitope reconstitution assay was performed using 2.5% of each RP-HPLC fraction (2 × 109 cell equivalents), and three peaks of activity were observed (Fig. 1). Peak B activity (fractions 26–28) corresponds to the previously described ALLAVGATK peptide from Pmel17/gp100 [7] and peak C activity (fraction 38) corresponds to a peptide from a newly identified cancer/testis antigen (see footnote 4).
Fig. 1. Reconstitution of VMM18 CTL epitopes by RP-HPLC-fractionated peptides eluted from immunoaffinity purified HLA-A3 molecules from VMM18 melanoma cells. C1R-A3 targets were incubated with HPLC fractions prior to standard 51 Cr-release assay at an E:T of 16:1. Three peaks of activity were observed.
peptide (Fig. 2). The peptides were further fractionated by a third round of RP-HPLC, using TEAA as the organic modifier. In CTL epitope reconstitution assays, the peak A antigenic peptide was found to elute in fractions 22–24 (Fig. 3). 3.3. The active peptide in peak A is SQNFPGSQK Mass spectrometric analysis of the active third dimension RP-HPLC fractions representing peak A indicated that the abundance of the m/z 497 ion strongly correlated with the CTL epitope reconstituting activity. An analysis of the fragment masses obtained from the CAD mass spectrum indicated that the m/z 497 peptide had the sequence SQNFPGSQK. This synthetic peptide was active in sensitizing C1R-A3 targets for lysis by VMM18 CTL at concentrations as low as 10 pM (Fig. 4). It was subsequently determined
80
% Specific Lysis
Total RNA was purified from VMM18 melanoma tumor cells by using RNAzolTM B (Tel-Test Inc., Friendswood, TX). First strand cDNA was synthesized by reverse transcribing 2 g of total RNA with oligo(dT) using the SuperScriptTM system (Gibco BRL, Gaithersburg, MD). The ExpandTM High Fidelity PCR system (Boehringer Mannheim, Indianapolis, IN) was used for polymerase chain reaction (PCR) amplification of 1 l of newly made first strand cDNA. The oligonucleotide primers for RT-PCR amplification of the human Pmel17/gp100 mRNA (accession #S73003) were PMEL1-5 (coding strand, position 11-31; 5 -GGA AGA ACA CAA TGG ATC TGG) and PMEL3-3 (non-coding strand, position 477–452; 5 -GAC ATA AAC AAA GCT TCT CTT CTG AG). To amplify a 467 bp PCR product the PCR components were heated to 94 ◦ C, after which 1 l of cDNA was added to each tube. Subsequent cycling conditions were 94 ◦ C for 15 s, 55 ◦ C for 30 s, and 72 ◦ C for 1 min for 30 cycles with a final extension of 72 ◦ C for 10 min. PCR products were analyzed on ethidium bromide stained agarose gels and sequenced.
60
40
20
0 20
3.2. Identification of the antigenic peptides Pooled active fractions 15–17 (peak A) were further fractionated by RP-HPLC using HFBA as the organic modifier. In CTL, epitope reconstitution assays, fractions 26 and 27 of this second fractionation of peak A contained the active
22
24
26
28
30 32
34
36
38
40
42
Fraction Number
Fig. 2. Reconstitution of CTL epitopes following a second round of RP-HPLC fractionation of pooled active first dimension fractions corresponding to peak A from the experiment shown in Fig. 1. The percentage of each HPLC fraction tested represents 3 × 109 cell equivalents. The E:T was 13:1.
134
K.T. Hogan et al. / Immunology Letters 90 (2003) 131–135
the peptide had an exact match in the Pmel17/gp100 sequence (residues 87–95). Nucleotide sequencing of the Pmel17/gp100 RT-PCR product from VMM18 cells yielded an exact match to the published sequence in this region, with no evidence of heterogeneity at these codons, suggesting that the sequence does not arise as the result of a mutation or rearrangement of the Pmel17/gp100 gene.
% Specific Lysis
80 60 40 20 0 10
14
18
22 26 30 Fraction Number
34
38
Fig. 3. Reconstitution of CTL epitopes following a third round of RP-HPLC fractionation of pooled active second dimension fractions corresponding to peak A from the experiment shown in Fig. 2. The percentage of each HPLC fraction tested represents 6 × 109 cell equivalents. The E:T was 16:1.
by RP-HPLC and mass spectrometry that the synthetic peptide SQNFPGSQK co-eluted with the unknown m/z 497 in the active second and third dimension fractions (data not shown), indicating that this sequence represents the naturally processed and presented epitope. 3.4. SQNFPGSQK is presented by the HLA-A3+ melanoma, VMM12 Mass spectrometric analysis of RP-HPLC fractionated peptides eluted from immunoaffinity purified HLA-A3 from the melanoma cell line VMM12 demonstrated that the SQNFPGSQK peptide was in the expected fractions (data not shown) and thus, represents a novel melanoma antigen that is presented, in association with HLA-A3 molecules, by two different melanomas. 3.5. The source gene for the SQNFPGSQK peptide is unknown Homology searches of SQNFPGSQK in the public non-redundant human protein database yielded no exact matches, although the seven N-terminal amino acids of
% specific lysis
80 60 40 20 0
1000 100 10
1
0.1 0.01 0.001
peptide concentration (nM)
Fig. 4. Epitope reconstitution with synthetic peptides. SQNFPGSQK (•–•), the peak A antigenic peptide, and ALNFPGSQK (•- - -•), a negative control, were used to sensitize C1R-A3 for lysis by VMM18 CTL at an E:T of 16:1.
4. Discussion VMM18 CTL were previously generated by the repeated stimulation of melanoma positive lymph node-derived lymphocytes with autologous VMM18 melanoma tumor [7]. Epitope reconstitution experiments demonstrated that the CTL recognize three HLA-A3-restricted antigens (see footnote 4). One of these antigens is the Pmel17/gp100-derived peptide ALLAVGATK [7], while the second antigen, RLSNRLLLR, is derived from a newly discovered cancer/testis antigen (see footnote 4). In the present report, we identify the third antigen as SQNFPGSQK. One criteria that needs to be fulfilled in order for an antigen to be included in a vaccine for the treatment of melanoma in a patient population, is that of being a shared antigen. SQNFPGSQK fulfills this criteria as it has been isolated from HLA-A3 molecules purified from both VMM18 and VMM12 melanoma tumor cells. A second important criteria is that the antigen should be capable of stimulating a CTL response. We have previously shown that the melanoma patient from which the VMM18 CTL were derived has spontaneous reactivity to the SQNFPGSQK peptide as determined by an ELISpot assay [12]. A third criteria by which to judge the suitability of an antigen for inclusion in a melanoma vaccine is that it is expressed in tumor cells but not in normal tissue. This is most often done by using PCR to assess the expression levels of the gene coding for the antigen in cDNA obtained from a large panel of normal tissues. We have been unable to make this assessment because we have not been able to identify the gene that codes for the protein from which the SQNFPGSQK peptide is derived. VMM18 CTL are, however, very specific for autologous tumor and are not reactive to normal cells, such as autologous EBV-transformed B cells. Thus, this criterion is at least partially satisfied. Homology searches of SQNFPGSQK in the public non-redundant human protein database yielded numerous hits with homology of six to seven amino acids, but no exact hit. Among the hits was the Pmel17/gp100 sequence NFPGSQK (residues 89–95), suggesting that there may be a mutation, rearrangement, or polymorphism of the Pmel17/gp100 gene in the VMM18 tumor. We are not aware of any reports of allelic variants of this gene, and the finding that the peptide is present on another melanoma would make the spontaneous mutation of a minimum of three bases in two adjacent codons an unlikely explanation for the origin of this epitope. Furthermore, nucleotide sequencing
K.T. Hogan et al. / Immunology Letters 90 (2003) 131–135
of the Pmel17/gp100 RT-PCR product from VMM18 cells yielded an exact match to the published sequence in this region, with no evidence of heterogeneity at these codons. A splicing event in which the 5 -end of the Pmel17/gp100 gene is replaced seems unlikely as a canonical splice site does not exist within the nucleotide sequence coding for the peptide. One possible explanation for the origin of the nucleotide sequencing coding the SQNFPGSQK peptide is that it is the product of a splicing event in which two exons of an unidentified gene are brought together, as is the case with the sequence coding the RLSNRLLLR peptide (see footnote 4). Although it does not explain the origin of the SQNFPGSQK peptide, a trivial explanation for the recognition of that peptide is that it is the result of cross-reactivity with the ALNFPGSQK peptide. We have, however, shown that VMM18 CTL do not recognize the ALNFPGSQK peptide [7]. Thus, the recognition of SQNFPGSQK cannot be explained by cross-reactivity with ALNFPGSQK. Because the SQNFPGSQK peptide represents a shared antigen that is capable of eliciting a spontaneous CTL response [12], we believe it warrants further investigation as to its origins.
Acknowledgements This work was supported by grants R01CA90815 (KTH), R01CA57653 (CLS), and F32CA72166 (LWT) from the National Cancer Institute, and by Argonex, Inc. The authors
135
would like to thank Diana Berry, Eyre Bigelow, Donna Deacon, Dominic Eisinger, Yanbin Li, Charles E. Lyons, Jr., Cara Miller, Karen Root, and Sandi Walton for their expert technical assistance. References [1] J. Weber, Cancer Invest. 20 (2002) 208. [2] G. Parmiani, C. Castelli, P. Dalerba, R. Mortarini, L. Rivoltini, F.M. Marincola, A. Anichini, J. Natl. Cancer Inst. 94 (2002) 805. [3] G.A. Ohnmacht, F.M. Marincola, J. Cell. Physiol. 182 (2000) 332. [4] C.L. Slingluff Jr., T.A. Colella, L. Thompson, D.D. Graham, J.C. Skipper, J. Caldwell, L. Brinckerhoff, D.J. Kittlesen, D.H. Deacon, C. Oei, N.L. Harthun, E.L. Huczko, D.F. Hunt, T.L. Darrow, V.H. Engelhard, Cancer Immunol. Immunotherapy 48 (2000) 661. [5] B. Seliger, M.J. Maeurer, S. Ferrone, Immunol. Today 21 (2000) 455. [6] B. Seliger, T. Cabrera, F. Garrido, S. Ferrone, Semin. Cancer Biol. 12 (2002) 3. [7] J.C. Skipper, D.J. Kittlesen, R.C. Hendrickson, D.D. Deacon, N.L. Harthun, S.N. Wagner, D.F. Hunt, V.H. Engelhard, C.L. Slingluff Jr., J. Immunol. 157 (1996) 5027. [8] P.D. Greenberg, M.A. Cheever, Surv. Immunol. Res. 4 (1985) 283. [9] K.T. Hogan, D.P. Eisinger, S.B. Cupp III, K.J. Lekstrom, D.D. Deacon, J. Shabanowitz, D.F. Hunt, V.H. Engelhard, C.L. Slingluff Jr., M.M. Ross, Cancer Res. 58 (1998) 5144. [10] A.L. Cox, J. Skipper, Y. Chen, R.A. Henderson, T.L. Darrow, J. Shabanowitz, V.H. Engelhard, D.F. Hunt, C.L. Slingluff Jr., Science 264 (1994) 716. [11] R.A. Henderson, A.L. Cox, K. Sakaguchi, E. Appella, J. Shabanowitz, D.F. Hunt, V.H. Engelhard, Proc. Natl. Acad. Sci. U.S.A. 90 (1993) 10275. [12] G. Yamshchikov, L. Thompson, W.G. Ross, H. Galavotti, W. Aquila, D. Deacon, J. Caldwell, J.W. Patterson, D.F. Hunt, C.L. Slingluff Jr., Clin. Cancer Res. 7 (2001) 909s.
Identification of a shared epitope recognized by melanoma-specific, HLA-A3-restricted cytotoxic T lymphocytes Kevin T. Hogan a,b,∗ , Michael A. Coppola b,1 , Christine L. Gatlin b,2 , Lee W. Thompson a , Jeffrey Shabanowitz c , Donald F. Hunt c,d , Victor H. Engelhard e , Craig L. Slingluff Jr. a , Mark M. Ross b,3 a
Department of Surgery, University of Virginia, Box 801359, Charlottesville, VA 22908, USA b Argonex Inc., Charlottesville, VA 22903, USA c Department of Chemistry, University of Virginia, Charlottesville, VA 22908, USA d Department of Pathology, University of Virginia, Charlottesville, VA 22908, USA e Department of Microbiology, University of Virginia, Charlottesville, VA 22908, USA
Received 8 July 2003; received in revised form 26 August 2003; accepted 26 August 2003
Abstract We previously established a melanoma-reactive cytotoxic T lymphocyte (CTL) line that recognizes multiple epitopes in the context of HLA-A3. To increase the number of peptides available for use in a vaccine for the treatment of melanoma, we identified one of these epitopes, SQNFPGSQK, through a combination of epitope reconstitution experiments and mass spectrometry. The SQNFPGSQK peptide was also found to be associated with HLA-A3 on an additional melanoma tumor line, thus indicating that the peptide is not unique to the melanoma tumor line from which it was isolated and thus, unlikely to arise through a mutational event. Although the protein origin of SQNFPGSQK has yet to be established, the shared nature of this epitope and the fact that it elicits a natural immune response indicates that it warrants further study to determine its usefulness as a vaccine component for the treatment of melanoma. The peptide may also be useful as a research tool for evaluating spontaneous anti-tumor immune responses in patients with melanoma. © 2003 Elsevier B.V. All rights reserved. Keywords: Immunotherapy; Cancer; Antigen; Epitope; Cytotoxic T lymphocyte; Mass spectrometry
1. Introduction Significant progress has been made in identifying peptide epitopes that can stimulate a melanoma-specific cytotoxic T lymphocyte (CTL) response [1–3]. The tumor antigens giving rise to these peptides include cancer/testis antigens, differentiation antigens, and mutated gene products [1–3]. A number of these peptides have been tested in clinical trials, with some patients showing partial or complete tumor regressions [1,2]. These initial results indicate that a thera∗ Corresponding author. Tel.: +1-434-243-9861; fax: +1-434-924-8464. E-mail address: [email protected] (K.T. Hogan). 1 Present address: Department of Cell Biology, University of Virginia, Charlottesville, VA 22908, USA. 2 Present address: The Institute for Genomic Research, Rockville, MD 20850, USA. 3 Present address: MDS Proteomics, Charlottesville, VA 22911, USA.
0165-2478/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.imlet.2003.08.003
peutic melanoma vaccine may prove efficacious in the treatment of the disease, and thus further development of such a vaccine is warranted. One limitation of peptide-based vaccines is the fact that they are only useful for the treatment of individuals that express one or more of the class I major histocompatibility complex (MHC)-encoded molecules to which the peptides are capable of binding. Most of the melanoma peptides that have been identified are restricted to HLA-A2, and thus there is a need to identify peptides that bind to other class I MHC molecules so as to provide broad-based population coverage. Another limitation that must be overcome is that tumors can undergo immunoselection and lose the expression of a given antigen or class I MHC molecule, thus preventing recognition of the tumor by the corresponding CTL [3–6]. This limitation can be overcome by including in the vaccine peptides that are derived from multiple tumor antigens, as well as peptides that bind to different class I MHC
132
K.T. Hogan et al. / Immunology Letters 90 (2003) 131–135
molecules. Therefore, we have continued to pursue the identification of additional melanoma antigens. We previously described a melanoma-specific, HLA-A3restricted CTL line that recognizes a Pmel17/gp100-derived peptide [7]. This CTL line recognizes two additional epitopes, one of which has been recently identified.4 In the present report, we identify the third antigen which is seen by this CTL in the context of HLA-A3.
2. Materials and methods
absorbance detector. The column was made by packing a 27 cm bed of 10 m C18 particles in a section of 285 m o.d. × 75 m i.d. fused silica. Peptides in a selected second dimension fraction were loaded onto this column and eluted with a gradient of acetonitrile/0.67% triethylamine acetate (TEAA)/water in 0.1% triethylamine acetate/water, with the concentration of acetonitrile increasing from 0 to 60% in 40 min. The flow rate was approximately 300 nl/min, and fractions were collected into 25 l of 0.1% acetic acid every 30 s. In all RP-HPLC experiments, peptides were detected by monitoring UV absorbance at 214 nm.
2.1. Cell lines
2.5. CTL epitope reconstitution assay
The melanoma lines VMM12 and VMM18 were maintained in RPMI1640 supplemented with 5–10% FBS and 2 mM l-glutamine. C1R-A3 was maintained in the same media supplemented with 200 g/ml G418.
Aliquots of each RP-HPLC fraction were tested for the presence of peptides that could sensitize C1R-A3 targets for lysis by VMM18 CTL in a standard 4 h 51 Cr-release assay as previously described [9].
2.2. CTL line
2.6. Mass spectrometric analyses
VMM18-specific CTL have been described previously [7]. CTL were expanded in bulk culture using anti-CD3 antibody [8] and cryopreserved in aliquots of (1–5) × 107 cells for use in epitope reconstitution assays.
Active RP-HPLC fractions were screened by on-line RP-HPLC/electrospray ionization mass spectrometry (MS) using a homemade microcapillary column and a Finnigan-MAT TSQ 7000 triple quadrupole mass spectrometer (Finnigan, San Jose, CA). Approximately, 1% of the active RP-HPLC fraction was loaded onto a section of 185 m o.d. × 75 m i.d. fused silica packed with 10–12 cm of 10 m C18 particles. Peptides were eluted directly into the mass spectrometer using a 10-min 0–60% acetonitrile in 0.1 M acetic acid gradient. Ions were formed by electrospray ionization, and mass spectra were recorded by scanning between mass to charge ratios (m/z) 300 and 1400 every 1.5 s. Active second dimension RP-HPLC fractions were analyzed using an effluent splitter on the microcapillary HPLC column. The column (360 m o.d. × 100 m i.d. with a 25-cm C18 bed) was connected with a zero dead volume tee (Valco, Houston, TX) to two pieces of fused silica of different lengths (25 and 40 m i.d.). Peptides were eluted with a 34 min gradient of 0–60% acetonitrile in 0.1 M acetic acid. The 25 m capillary deposited one-fifth of the RP-HPLC effluent into the wells of a microtiter plate for use in a CTL epitope reconstitution assay, while the remaining four-fifths of the effluent was directed into the mass spectrometer, with mass spectra recorded as described above [9]. Peptide sequences were determined by collision-activated dissociation (CAD) tandem mass spectrometry using an LCQ (Finnigan) ion trap mass spectrometer and methods as described [10,11].
2.3. Isolation of HLA-A3 associated peptides HLA-A3 molecules were immunoaffinity purified from aliquots of (6–8) × 1010 VMM18 tumor cells as described [9] using HLA-A3-specific monoclonal antibody GAP-A3 bound to protein A-Sepharose. 2.4. Peptide fractionation Peptide extracts were fractionated by reverse phase-high performance liquid chromatography (RP-HPLC) using an Applied Biosystems model 140B system. The extracts were concentrated by vacuum centrifugation and injected onto a Higgins (Mountain View, CA) C18 HAISIL column (2.1 mm × 4 cm, 300 Å, 5 m). The peptides were eluted with a gradient of acetonitrile/0.085% trifluoroacetic acid (TFA) in 0.1% TFA/water, with the concentration of acetonitrile increasing from 0 to 9% (0 to 5 min), 9 to 36% (5 to 55 min), and 36 to 60% (55 to 62 min). Second dimension fractionations of selected first dimension (TFA) fractions were accomplished using the same gradient but with the substitution of heptafluorobutyric acid (HFBA) for TFA. A third dimension of RP-HPLC was achieved using an Eldex (Napa, CA) MicroPro pump, a homemade C18 microcapillary column and an Applied Biosystems model 785A UV 4
K.T. Hogan, M.A. Coppola, C.L. Gatlin, L.W. Thompson, J. Shabanowitz, D.F. Hunt, V. H. Engelhard, M.M. Ross, C.L. Slingluff, Jr. Identification of novel and widely expressed cancer/testis gene isoforms that elicit spontaneous CTL reactivity to melanoma, submitted for publication.
2.7. Peptide synthesis Peptides were synthesized using a Gilson (Madison, WI) AMS 422 multiple peptide synthesizer using conventional FMOC chemistry. Peptides were purified by RP-HPLC using a 4.6 mm i.d. × 100 mm long POROS (Perseptive
K.T. Hogan et al. / Immunology Letters 90 (2003) 131–135
133
Biosystems, Cambridge, MA) column and a 10-min 0–60% acetonitrile in 0.1% TFA gradient. 2.8. Pmel17/gp100 sequencing
2.9. Human subjects All research involving human subjects was approved by the University of Virginia Human Investigation Committee.
3. Results 3.1. VMM18 CTL recognize three distinct HLA-A3 restricted epitopes The peptides bound to HLA-A3 molecules on 8 × 1010 VMM18 tumor cells were purified as described in Section 2, and fractionated by RP-HPLC using TFA as the organic modifier. A CTL epitope reconstitution assay was performed using 2.5% of each RP-HPLC fraction (2 × 109 cell equivalents), and three peaks of activity were observed (Fig. 1). Peak B activity (fractions 26–28) corresponds to the previously described ALLAVGATK peptide from Pmel17/gp100 [7] and peak C activity (fraction 38) corresponds to a peptide from a newly identified cancer/testis antigen (see footnote 4).
Fig. 1. Reconstitution of VMM18 CTL epitopes by RP-HPLC-fractionated peptides eluted from immunoaffinity purified HLA-A3 molecules from VMM18 melanoma cells. C1R-A3 targets were incubated with HPLC fractions prior to standard 51 Cr-release assay at an E:T of 16:1. Three peaks of activity were observed.
peptide (Fig. 2). The peptides were further fractionated by a third round of RP-HPLC, using TEAA as the organic modifier. In CTL epitope reconstitution assays, the peak A antigenic peptide was found to elute in fractions 22–24 (Fig. 3). 3.3. The active peptide in peak A is SQNFPGSQK Mass spectrometric analysis of the active third dimension RP-HPLC fractions representing peak A indicated that the abundance of the m/z 497 ion strongly correlated with the CTL epitope reconstituting activity. An analysis of the fragment masses obtained from the CAD mass spectrum indicated that the m/z 497 peptide had the sequence SQNFPGSQK. This synthetic peptide was active in sensitizing C1R-A3 targets for lysis by VMM18 CTL at concentrations as low as 10 pM (Fig. 4). It was subsequently determined
80
% Specific Lysis
Total RNA was purified from VMM18 melanoma tumor cells by using RNAzolTM B (Tel-Test Inc., Friendswood, TX). First strand cDNA was synthesized by reverse transcribing 2 g of total RNA with oligo(dT) using the SuperScriptTM system (Gibco BRL, Gaithersburg, MD). The ExpandTM High Fidelity PCR system (Boehringer Mannheim, Indianapolis, IN) was used for polymerase chain reaction (PCR) amplification of 1 l of newly made first strand cDNA. The oligonucleotide primers for RT-PCR amplification of the human Pmel17/gp100 mRNA (accession #S73003) were PMEL1-5 (coding strand, position 11-31; 5 -GGA AGA ACA CAA TGG ATC TGG) and PMEL3-3 (non-coding strand, position 477–452; 5 -GAC ATA AAC AAA GCT TCT CTT CTG AG). To amplify a 467 bp PCR product the PCR components were heated to 94 ◦ C, after which 1 l of cDNA was added to each tube. Subsequent cycling conditions were 94 ◦ C for 15 s, 55 ◦ C for 30 s, and 72 ◦ C for 1 min for 30 cycles with a final extension of 72 ◦ C for 10 min. PCR products were analyzed on ethidium bromide stained agarose gels and sequenced.
60
40
20
0 20
3.2. Identification of the antigenic peptides Pooled active fractions 15–17 (peak A) were further fractionated by RP-HPLC using HFBA as the organic modifier. In CTL, epitope reconstitution assays, fractions 26 and 27 of this second fractionation of peak A contained the active
22
24
26
28
30 32
34
36
38
40
42
Fraction Number
Fig. 2. Reconstitution of CTL epitopes following a second round of RP-HPLC fractionation of pooled active first dimension fractions corresponding to peak A from the experiment shown in Fig. 1. The percentage of each HPLC fraction tested represents 3 × 109 cell equivalents. The E:T was 13:1.
134
K.T. Hogan et al. / Immunology Letters 90 (2003) 131–135
the peptide had an exact match in the Pmel17/gp100 sequence (residues 87–95). Nucleotide sequencing of the Pmel17/gp100 RT-PCR product from VMM18 cells yielded an exact match to the published sequence in this region, with no evidence of heterogeneity at these codons, suggesting that the sequence does not arise as the result of a mutation or rearrangement of the Pmel17/gp100 gene.
% Specific Lysis
80 60 40 20 0 10
14
18
22 26 30 Fraction Number
34
38
Fig. 3. Reconstitution of CTL epitopes following a third round of RP-HPLC fractionation of pooled active second dimension fractions corresponding to peak A from the experiment shown in Fig. 2. The percentage of each HPLC fraction tested represents 6 × 109 cell equivalents. The E:T was 16:1.
by RP-HPLC and mass spectrometry that the synthetic peptide SQNFPGSQK co-eluted with the unknown m/z 497 in the active second and third dimension fractions (data not shown), indicating that this sequence represents the naturally processed and presented epitope. 3.4. SQNFPGSQK is presented by the HLA-A3+ melanoma, VMM12 Mass spectrometric analysis of RP-HPLC fractionated peptides eluted from immunoaffinity purified HLA-A3 from the melanoma cell line VMM12 demonstrated that the SQNFPGSQK peptide was in the expected fractions (data not shown) and thus, represents a novel melanoma antigen that is presented, in association with HLA-A3 molecules, by two different melanomas. 3.5. The source gene for the SQNFPGSQK peptide is unknown Homology searches of SQNFPGSQK in the public non-redundant human protein database yielded no exact matches, although the seven N-terminal amino acids of
% specific lysis
80 60 40 20 0
1000 100 10
1
0.1 0.01 0.001
peptide concentration (nM)
Fig. 4. Epitope reconstitution with synthetic peptides. SQNFPGSQK (•–•), the peak A antigenic peptide, and ALNFPGSQK (•- - -•), a negative control, were used to sensitize C1R-A3 for lysis by VMM18 CTL at an E:T of 16:1.
4. Discussion VMM18 CTL were previously generated by the repeated stimulation of melanoma positive lymph node-derived lymphocytes with autologous VMM18 melanoma tumor [7]. Epitope reconstitution experiments demonstrated that the CTL recognize three HLA-A3-restricted antigens (see footnote 4). One of these antigens is the Pmel17/gp100-derived peptide ALLAVGATK [7], while the second antigen, RLSNRLLLR, is derived from a newly discovered cancer/testis antigen (see footnote 4). In the present report, we identify the third antigen as SQNFPGSQK. One criteria that needs to be fulfilled in order for an antigen to be included in a vaccine for the treatment of melanoma in a patient population, is that of being a shared antigen. SQNFPGSQK fulfills this criteria as it has been isolated from HLA-A3 molecules purified from both VMM18 and VMM12 melanoma tumor cells. A second important criteria is that the antigen should be capable of stimulating a CTL response. We have previously shown that the melanoma patient from which the VMM18 CTL were derived has spontaneous reactivity to the SQNFPGSQK peptide as determined by an ELISpot assay [12]. A third criteria by which to judge the suitability of an antigen for inclusion in a melanoma vaccine is that it is expressed in tumor cells but not in normal tissue. This is most often done by using PCR to assess the expression levels of the gene coding for the antigen in cDNA obtained from a large panel of normal tissues. We have been unable to make this assessment because we have not been able to identify the gene that codes for the protein from which the SQNFPGSQK peptide is derived. VMM18 CTL are, however, very specific for autologous tumor and are not reactive to normal cells, such as autologous EBV-transformed B cells. Thus, this criterion is at least partially satisfied. Homology searches of SQNFPGSQK in the public non-redundant human protein database yielded numerous hits with homology of six to seven amino acids, but no exact hit. Among the hits was the Pmel17/gp100 sequence NFPGSQK (residues 89–95), suggesting that there may be a mutation, rearrangement, or polymorphism of the Pmel17/gp100 gene in the VMM18 tumor. We are not aware of any reports of allelic variants of this gene, and the finding that the peptide is present on another melanoma would make the spontaneous mutation of a minimum of three bases in two adjacent codons an unlikely explanation for the origin of this epitope. Furthermore, nucleotide sequencing
K.T. Hogan et al. / Immunology Letters 90 (2003) 131–135
of the Pmel17/gp100 RT-PCR product from VMM18 cells yielded an exact match to the published sequence in this region, with no evidence of heterogeneity at these codons. A splicing event in which the 5 -end of the Pmel17/gp100 gene is replaced seems unlikely as a canonical splice site does not exist within the nucleotide sequence coding for the peptide. One possible explanation for the origin of the nucleotide sequencing coding the SQNFPGSQK peptide is that it is the product of a splicing event in which two exons of an unidentified gene are brought together, as is the case with the sequence coding the RLSNRLLLR peptide (see footnote 4). Although it does not explain the origin of the SQNFPGSQK peptide, a trivial explanation for the recognition of that peptide is that it is the result of cross-reactivity with the ALNFPGSQK peptide. We have, however, shown that VMM18 CTL do not recognize the ALNFPGSQK peptide [7]. Thus, the recognition of SQNFPGSQK cannot be explained by cross-reactivity with ALNFPGSQK. Because the SQNFPGSQK peptide represents a shared antigen that is capable of eliciting a spontaneous CTL response [12], we believe it warrants further investigation as to its origins.
Acknowledgements This work was supported by grants R01CA90815 (KTH), R01CA57653 (CLS), and F32CA72166 (LWT) from the National Cancer Institute, and by Argonex, Inc. The authors
135
would like to thank Diana Berry, Eyre Bigelow, Donna Deacon, Dominic Eisinger, Yanbin Li, Charles E. Lyons, Jr., Cara Miller, Karen Root, and Sandi Walton for their expert technical assistance. References [1] J. Weber, Cancer Invest. 20 (2002) 208. [2] G. Parmiani, C. Castelli, P. Dalerba, R. Mortarini, L. Rivoltini, F.M. Marincola, A. Anichini, J. Natl. Cancer Inst. 94 (2002) 805. [3] G.A. Ohnmacht, F.M. Marincola, J. Cell. Physiol. 182 (2000) 332. [4] C.L. Slingluff Jr., T.A. Colella, L. Thompson, D.D. Graham, J.C. Skipper, J. Caldwell, L. Brinckerhoff, D.J. Kittlesen, D.H. Deacon, C. Oei, N.L. Harthun, E.L. Huczko, D.F. Hunt, T.L. Darrow, V.H. Engelhard, Cancer Immunol. Immunotherapy 48 (2000) 661. [5] B. Seliger, M.J. Maeurer, S. Ferrone, Immunol. Today 21 (2000) 455. [6] B. Seliger, T. Cabrera, F. Garrido, S. Ferrone, Semin. Cancer Biol. 12 (2002) 3. [7] J.C. Skipper, D.J. Kittlesen, R.C. Hendrickson, D.D. Deacon, N.L. Harthun, S.N. Wagner, D.F. Hunt, V.H. Engelhard, C.L. Slingluff Jr., J. Immunol. 157 (1996) 5027. [8] P.D. Greenberg, M.A. Cheever, Surv. Immunol. Res. 4 (1985) 283. [9] K.T. Hogan, D.P. Eisinger, S.B. Cupp III, K.J. Lekstrom, D.D. Deacon, J. Shabanowitz, D.F. Hunt, V.H. Engelhard, C.L. Slingluff Jr., M.M. Ross, Cancer Res. 58 (1998) 5144. [10] A.L. Cox, J. Skipper, Y. Chen, R.A. Henderson, T.L. Darrow, J. Shabanowitz, V.H. Engelhard, D.F. Hunt, C.L. Slingluff Jr., Science 264 (1994) 716. [11] R.A. Henderson, A.L. Cox, K. Sakaguchi, E. Appella, J. Shabanowitz, D.F. Hunt, V.H. Engelhard, Proc. Natl. Acad. Sci. U.S.A. 90 (1993) 10275. [12] G. Yamshchikov, L. Thompson, W.G. Ross, H. Galavotti, W. Aquila, D. Deacon, J. Caldwell, J.W. Patterson, D.F. Hunt, C.L. Slingluff Jr., Clin. Cancer Res. 7 (2001) 909s.