SIMULTANEOUS EXPRESSION OF T-CELL ACTIVATING ANTIGENS INRENAL CELL CARCINOMA: IMPLICATIONS FOR SPECIFIC

IMMUNOTHERAPY

MARK RINGHOFFER, CARLHEINZ R. MULLER, ARIANE SCHENK, HANSPETER KIRSCHE,MICHAEL SCHMITT, JOCHEN GREINER AND JURGEN E. GSCHWEND*

From the Departments of Internal Medicine III (MR, AS, HK, MS, JG) and Urology (JEG), University Hospital and German NationalBone Marrow Donor Registry (CRM), Ulm, Germany

ABSTRACT

Purpose: The activation of antigen specific T cells by tumor associated antigens (TAA) might bea promising treatment strategy for patients with renal cell carcinoma (RCC). We analyzed TAAexpression in patients with RCC as well as the prevalence of fitting HLA phenotypes andcalculated the percent of patients eligible for peptide vaccination trials.

Materials and Methods: A total of 41 RCC samples from primary tumors were analyzed forTAA expression by reverse transcriptase-polymerase chain reaction. Genes of interest wereMAGE-1, MAGE-3, G250 and PRAME since peptides derived from these genes have been shownto activate antigen specific cytotoxic T lymphocytes. Results were combined with data on the HLAgene and haplotype frequencies in the German population as an example of a white population.

Results: Tumor specific expression of at least 1 T-cell activating antigen was observed in allpatients. Of the patients 80% expressed 2 or more TAAs simultaneously. HLA molecules suitablefor presentation of the respective antigens were calculated to be expressed in 51% to 85% of whiteGerman patients. These results mirror with only minor variations most of the white populationsin Europe and North America.

Conclusions: We noted that T-cell activating tumor associated antigens are frequently ex-pressed in patients with RCC. Based on HLA expression analysis in a white population at least30% of patients with RCC are eligible for monovalent specific immunotherapy and 41% areeligible for polyvalent specific immunotherapy. These data are a rational basis for futureprospective vaccination trials in patients with RCC.

KEY WORDS: kidney; carcinoma, renal cell; antigens; T-lymphocytes; vaccination

During the last decade several tumor associated antigens(TAAs) have been identified, of which some are able to elicittumor specific immune responses. Screening experiments usingcytotoxic T lymphocytes (CTLs) or the serological screening ofrecombinant expression libraries (SEREX approach) led to theidentification of multiple T-cell epitopes. Furthermore, the useof prediction algorithms (that is the SYFPEITHI, BIMAS orPaProC algorithm) that integrate HLA related peptide motifsand proteasomal cleavage rules have been shown to be powerfultools for the prediction of the definite epitope that might bepresented by antigen presenting cells to T cells.1–3 Peptideantigens entered clinical vaccination trials, that is MAGE-3,tyrosinase and Melan A/MART1 in melanoma or prostate spe-cific antigen and prostate specific membrane antigen in pros-tate cancer. For all of these antigens the enhancement of spe-cific T-cell reactivity could be observed and in some casesclinical responses could be achieved.4–6

In renal cell carcinoma (RCC) the yield for truly immuno-genic and clinically applicable antigens has remained low.Although many identified antigens have been shown to elicitcellular immune responses, some have only limited therapeu-tic value due to low frequency of expression7 or their unfa-vorable expression pattern in tumor and normal tissues.

A basic problem accompanying peptide vaccination strate-gies, especially in heterogeneous solid tumors, is the selec-tion of tumor cells caused by antigen loss. In the melanomamodel it was demonstrated that metastases that developedduring successful peptide vaccination lost the immunogenicantigen.8 To overcome this problem it is desirable to establisha polyvalent vaccination strategy that combines several pep-tide antigens and, therefore, minimizes the risk of immuneescape.9 Compared with alternative strategies, such as vac-cination with tumor cell lysates10 or RNA11 as a source ofantigen, peptide vaccination offers the opportunity to assessthe T-cell response accurately with sophisticated techniques,such as ELISpot assays or tetramer staining.

Beside TAA expression the second prerequisite for the suc-cessful induction of a T-cell response is the presence of an HLAclass I molecule on the surface of antigen presenting cells. Theinteraction between antigenic peptides and the HLA moleculerelies on the peptide motif and the molecular structure of theantigen binding site. Since this interaction is rather specific, anantigenic peptide can only be present on specific HLA mole-cules.12 Therefore, an estimation of the eligibility of patientswith RCC for clinical peptide vaccination must consider TAAexpression as well as HLA allele distribution. The current studycontributes comprehensive data to estimate the number of pa-tients with RCC who would be eligible for an individualizedmonovalent or polyvalent peptide vaccination trial.

MATERIALS AND METHODS

Patients samples and cell lines. We prospectively analyzedtissue samples from 41 patients with RCC. Patients provided

Accepted for publication December 19, 2003.Study received local ethics committee approval.Supported by Institutional Grants P.731 and P.763 from the Uni-

versity of Ulm.* Correspondence: Department of Urology, University of Ulm,

Prittwitz-Strasse 43, 89075 Ulm, Germany (telephone: ��49–731-500–27808; FAX: ��49–731-500–33166; e-mail: juergen.gschwend@medizin.uni-ulm.de).

0022-5347/04/1716-2456/0 Vol. 171, 2456–2460, June 2004THE JOURNAL OF UROLOGY® Printed in U.S.A.Copyright © 2004 by AMERICAN UROLOGICAL ASSOCIATION DOI: 10.1097/01.ju.0000118383.86684.38

2456

informed consent. About 1 cm3 of a single tumor tissue sam-ple was prepared per patient and cryopreserved immediatelyafter radical nephrectomy. In all cases the diagnosis of RCCwas confirmed by pathological findings.

Reverse transcriptase-polymerase chain reaction (RT-PCR).Total RNA was isolated from snap frozen tumor samplesusing Trizol reagent (Gibco, Karlsruhe, Germany). Approxi-mately 2 �g of each RNA sample were as subjected to cDNAsynthesis using a SuperScript Kit (Gibco) containing oli-go(deoxythymidine) primer for reverse transcription. As acontrol, �-actin PCR was performed and monitored by aga-rose gel electrophoresis. Generally the reaction mixture wascomposed of a standard PCR buffer (10 mM tris-HCl, pH 8.3and 50 mM KCl) containing 1.5 mM magnesium chloride,deoxynucleoside triphosphates (250 nM each), primer pairs(10 pmol each) and 1.5 �l template cDNA, corresponding toabout 1 ng template cDNA, in a final volume of 50 �l 2.5 UAmpliTaq (Perkin Elmer, Foster City, California). The reac-tion was started with an initial denaturation step (95C for 5minutes), followed by the respective number of amplificationcycles and a final elongation step (72C for 10 minutes). AGene Amp PCR system 2400 device (Perkin Elmer) was usedfor the amplification procedure.

Primer sequences and PCR conditions. Table 1 lists primersequences and PCR conditions. All primer pairs were locatedin different exons to avoid the amplification of contaminatinggenomic DNA. The identity of PCR products was initiallyverified by direct sequencing with an ABI Prism 373 DNAsequencer (Applied Biosystems, Foster City, California). Dis-crimination among MAGE-3, MAGE-2 and MAGE-6 was fur-ther guaranteed by restriction analysis using the endonucle-ase Sca I (Boehringer Ingelheim, Mannheim, Germany),which digests MAGE-2 and MAGE-6 but not MAGE-3.

For G250 we established RT-PCR with a second primerpair since we found low sensitivity of the already publishedprimer pairs. All negative samples were evaluated againwith this second primer pair. Table 2 lists the results of these2 PCR reactions.

Calculation of the eligibility of patients with RCC for vac-cination trials. The fraction of patients with a fitting HLAallele was estimated based on the allele and haplotype fre-quencies of the German population according to Muller etal.13 It was compared with data on the United States Na-tional Marrow Donor Program (NMDP).14

The antigen frequency fA2 HLA-A2 can be obtained fromthe gene gA2 using formula 1, fA2 � 1 � (1 � gA2)2. Due to thelinkage disequilibrium between the loci HLA-A and HLA-Bthe formula for the frequency (fX) of individuals carrying atleast 1 of the antigens A1, A2, A24 and B35 is more complex.First, the gene frequencies gA1, gA2, gA24 and gB35, and the 2

locus haplotype frequencies gA1-B35, gA2-B35 and gA24-B35 areused to calculate the cumulative frequency (hC) of all HLA-A-B haplotypes carrying at least 1 gene for A1, A2, A24 andB35 using the formula 2, hc � gA1 � gA2 � gA24 � (gB35 �hA1-B35 � hA2-B35 � hA24-B35), where the term in parenthesescorresponds to the frequency of all B35 haplotypes withoutA1, A2 and A24. The desired frequency can then be calcu-lated in the same way as formula 1 using the equation 3,fc � 1 � (1 � hc)

2.To determine the percent of patients eligible for vaccina-

tion with a single peptide the appropriate values of HLAfrequency and the mRNA expression rate of a certain TAAhad to be multiplied. Thus, the eligibility rates eG, eM and ePfor the 3 peptides G250 (G), MAGE-3 (M) and PRAME (P),respectively, were calculated using formulas 4A to C, eG � fA2� fG, eM � fc � fM and eP � fA2 � fP, respectively, using theantigen frequency of HLA-A2 for G 250 and PRAME, thecombined frequencies of A1, A2, A24, B35 for MAGE-3, andthe respective expressions rates fG, fM and fP.

When calculating the eligibility rates for monovalent(emono), polyvalent (fpoly) and overall (ftotal) peptide vaccina-tion, one must consider the dependency of the HLA aspect ofthe peptide presentation. Overall eligibility can be calculatedas the sum of the eligibility rate contributed by individualswith the antigen HLA-A2, where the contributions of the 3peptide accumulate like independent variables and of indi-viduals with HLA-A1, HLA-A24 or HLA-B35 but withoutHLA-A2, which are able to present MAGE-3. The 2 groupscorrespond to the 2 terms in formula 4, etotal � (1 � (1 � fG) �(1 � fM) � (1 � fP)) � fA2 � fM � (fC � fA2) � (fG � fM � fP� fG � fM � fM � fP � fP � fG � fG � fM � fP) � fA2 � fM �(fC � fA2).

Here polyvalent vaccination is only feasible in individualswith the antigen HLA-A2. The eligibility frequency (epoly) canbe calculated by formula 5, epoly � (fG � fM � fM � fP � fP �fG � 2fG � fM � fP) � fA2. As a consequence, the difference isthe fraction of individuals eligible for a monovalent vaccina-tion, as calculated by formula 6, emono � etotal � epoly � (fG �fM � fP � 2 � fG � fM � 2 � fP � fG � 3 � fG � fM � fP) �fA2 � fM � (fC � fA2). These calculations are based on theobservation that TAA expressions are independent of eachother. In parallel the calculations were performed using theobserved fractions of expression of only 1 or more than 1TAA.

Statistical analysis. We used 2 � 2 contingency tables andthe 2-sided Fisher exact test to obtain R values and signifi-cance levels concerning the co-expression of the differentTAAs.15 The chi-square test was used to evaluate the rela-tionships of TAA expression to the features tumor size, grad-ing and histology.

TABLE 1. Primer sequences and RT-PCR conditions

Gene of Interest (sequence)Annealing

Temperature(°C)

Product Length(bp)

Denaturation/Annealing/Elongation (mins)

GenbankAccession No.

G 250 a:5� ACT GCT GCT TCT GAT GCC TGT 3� 68 494 2/2/1 NM_001216.15� AGT TCT GGG AGC GGC GGG A 3�

G 250 b:5� ACT TCA GCC GCT ACT TCC AA 3� 58 370 1/1/1 NM_001216.15� TCT CAT CTG CAC AAG GAA CG 3�

MAGE-1:5� CGG CCG AAG GAA CCT GAC CCA G 3� 64 421 1/1/1 NM_004988.25� GCT GGA ACC CTC ACT GGG TTG CC 3�

MAGE-3:5� ACC AAG GAG AAG ATC TGC CAG TGG GTC TC 3� 72 722 1/2/2 NM_005362.25� ACA GTC GCC CTC TTT TGC GAT TAT GG 3�

PRAME:5� GTC CTG AGG CCA GCC TAA GT 3� 64 822 1/1/1 NM_0061155� GGA GAG GAG GAG TCT ACG CA 3�

�-Actin:5� GCA TCG TGA TGG ACT CCG 3� 68 613 1/1/1 NM_001101.25� GCT GGA AGG TGG ACA GCG A 3�

RENAL CELL CANCER AND PEPTIDE VACCINATION TRIALS 2457

RESULTS

Patients and tumor characteristics. A total of 41 patientswith RCC, including 27 males and 14 females, were studied.Median age of our patients was 66 years (range 34 to 79).RCC histology and grading were classified according toThoenes et al.16 Table 2 shows an overview of the total data.Figure 1 shows the distribution of carcinomas among thesubgroups according to the 1997 TNM classification, gradingand histology.

Frequency of mRNA TAA expression in RCC samples. An-tigen mRNA expression was found in 25 of 41 patients (61%)for MAGE-3, in 25 of 41 (61%) for PRAME and in 38 of 41(93%) for G250 on RT-PCR. Of 30 patients with a clear cellhistology 29 (97%) had positive G250 mRNA. Although astrong signal was obtained in the myeloid cell line K562,which served as a positive control, none of the 18 RCC sam-ples expressed MAGE 1. Figure 2 shows an ethidium bromidestained agarose gel with the PCR product showing all inves-tigated antigens.

Each patient expressed at least 1 TAA. A total of 33 sam-ples (80%) showed a polyvalent antigen mRNA expressionpattern (fig. 3). We used the chi-square test to evaluate ifthere was any relationship between TAA mRNA expressionand tumor stage, grading or histology. We could not detectsignificant differences among these subgroups (data notshown) except for a correlation of G250 expression and achromophilic or clear cell phenotype (p � 0.02).

On further analysis we investigated if there was any cor-relation of the expression of the antigens under investiga-tions. The R value for the co-expression of each pair of the 3TAAs investigated was between –0.13 and 0.16 (2-sided

Fisher exact test p � 0.6). As a consequence, we assumed anindependent expression in formulas 4 to 6.

Calculating the number of patients eligible for specific im-munotherapy based on tumor antigen expression and HLAmolecule distribution. Specific immunotherapy requires apeptide consisting of an amino acid sequence with a distinctHLA anchor motif. Therefore, the number of patients suit-

TABLE 2. Patient characteristics and corresponding TAA expression pattern

CodePt

TNM Stage Grade �-Actin MAGE-1 MAGE-3 Prame G250 HistologySex — Age

UL-RCC-2 F — 54 pT1N0M0 G2 Pos Neg Neg Pos Pos Clear cellUL-RCC-3 M — 76 pT1N0M0 G2 Pos Neg Pos Pos Pos Clear cellUL-RCC-4 M — 60 pT1pN0M0 G2 Pos Neg Pos Pos Pos Clear cellUL-RCC-5 F — 77 pT3aNxM0 G1–2 Pos Neg Pos Neg Pos Clear cellUL-RCC-6 F — 66 pT2pN0M0 G1 Pos Neg Pos Neg Pos Clear cellUL-RCC-9 M — 74 pT3aNxM0 G1–2 Pos Neg Pos Neg Pos Clear cellUL-RCC-10 M — 61 pT3bpN0M0 G2—3 Pos Neg Neg Pos Pos Chromophilic/papillaryUL-RCC-11 M — 67 pT3bN0M0 G2–3 Pos Neg Pos Pos Pos Clear cellUL-RCC-12 M — 59 pT3bN0M0 G2 Pos Neg Pos Pos Pos Chromophilic/papillaryUL-RCC-13 M — 76 pT3aN0M0 G2 Pos Neg Neg Pos Neg ChromophobeUL-RCC-14 F — 62 pT1N0M0 G1 Pos Neg Neg Pos Neg ChromophobeUL-RCC-15 M — 69 pT1N0M0 G1 Pos Neg Pos Neg Pos Clear cellUL-RCC-16 F — 79 pT3bN0M0 G2 Pos Neg Pos Pos Pos Clear cellUL-RCC-17 M — 59 pT2pN0M0 G2 Pos Neg Pos Neg Neg Clear cellUL-RCC-18 M — 63 pT3bN0M0 G1 Pos Neg Pos Pos Pos Clear cellUL-RCC-19 M — 78 pT1pN0M0 G2–4 Pos Neg Pos Pos Pos Chromophilic/papillaryUL-RCC-20 M — 67 pT2pN0M0 G2 Pos Neg Pos Neg Pos Clear cellUL-RCC-21 F — 60 pT3pN0M0 G1 Pos Neg Pos Neg Pos Clear cellUL-RCC-22 F — 79 pT3bpNxM0 G2 Pos Not determined Pos Pos Pos Clear cellUL-RCC-23 M — 76 pT3pN0M0 G2 Pos Not determined Pos Neg Pos Clear cellUL-RCC-25 F — 74 pT3N0Mx G2–3 Pos Not determined Pos Pos Pos Clear cellUL-RCC-26 M — 72 pT3pN1Mx G1–3 Pos Not determined Pos Neg Pos Clear cellUL-RCC-27 F — 62 pT2NxMx G1 Pos Not determined Pos Pos Pos ChromophobeUL-RCC-28 F — 46 pT2NxM0 G1 Pos Not determined Pos Pos Pos Clear cellUL-RCC-29 M — 61 pT2NxMx G2 Pos Not determined Pos Pos Pos Clear cellUL-RCC-33 F — 79 pT3bpN0M0 G2 Pos Not determined Neg Neg Pos Chromophilic/papillaryUL-RCC-34 M — 63 pT3apN0M0 G2 Pos Not determined Neg Pos Pos Chromophilic/papillaryUL-RCC-35 M — 72 pT3apNxM0 G1–3 Pos Not determined Neg Neg Pos Chromophilic/papillaryUL-RCC-38 M — 69 pT1pN0M0 G2–4 Pos Not determined Neg Pos Pos Chromophilic/papillaryUL-RCC-39 F — 75 pT1N0M0 G3 Pos Not determined Pos Neg Pos Clear cellUL-RCC-40 M — 74 pT3bpN0M0 G2 Pos Not determined Neg Pos Pos Clear cellUL-RCC-41 M — 65 pT1pN0M0 G1 Pos Not determined Neg Pos Pos Clear cellUL-RCC-42 F — 78 pT1pN0M0 G2 Pos Not determined Neg Neg Pos Clear cellUL-RCC-43 M — 73 pT3apNxM0 G2 Pos Not determined Neg Pos Pos Clear cellUL-RCC-44 M — 57 pT3pN0M1 G2 Pos Not determined Neg Pos Pos Clear cellUL-RCC-45 M — 34 pT2pN0M0 G2 Pos Not determined Neg Neg Pos Clear cellUL-RCC-46 M — 50 pT1pNxM0 G2 Pos Not determined Pos Pos Pos Clear cellUL-RCC-47 M — 65 pT3pN0M0 G2 Pos Not determined Pos Pos Pos Chromophilic/papillaryUL-RCC-48 M — 66 pT1N0M0 G1 Pos Not determined Neg Pos Pos Clear cellUL-RCC-49 M — 65 pT1pN0M0 G1 Pos Not determined Neg Neg Pos Clear cellUL-RCC-53 F — 61 pT1pN0M0 G1 Pos Not determined Pos Neg Pos Clear cell

FIG. 1. Tumor characteristics of patient cohort

RENAL CELL CANCER AND PEPTIDE VACCINATION TRIALS2458

able for such a treatment strategy depends on the HLA genedistribution in a given population. To calculate the frequencyof individuals carrying at least 1 suitable HLA antigen (A1,A2, A24 and B35) we founded on the gene haplotype frequen-cies of a German reference population, as described. Compar-ing our German HLA data with a similar study of Americanpopulation data from the United States NMDP14 revealedextremely similar gene frequencies for their white population(HLA-A1 15.2% vs 15.7%, HLA-A2 28.7% vs 29.9%, HLA A 249.3% vs 9.2% and HLA-B35 9.7% vs 9.7%). Since the NMDPdatabase only gives the frequency of the haplotype HLA-A2-B35, we cannot give exact calculations for all 5 ethnic sub-groups of the United States. However, the eligibility for apeptide vaccination would be almost identical for whiteAmericans and white Germans, and it should be greater than60% for the Asian, Latin and native American populations.For black Americans a substantially lower eligibility rate ofabout 45% is expected due their lower frequency of HLA-A2.Tables 3 and 4 show the detailed calculation for the Germanpopulation.

DISCUSSION

In the current study we evaluated the expression patternof proven T cell activating TAAs that might be involved in theinduction of T-cell specific immunity. An important findingwas the frequent expression of MAGE-3 in RCC. This is ofparticular interest since MAGE-3 has been successfully in-volved in peptide vaccination trials in patients with melano-ma,17 demonstrating the induction of MAGE-3 specific CTLs.To date MAGE-3 derived CTL activating peptides have beenidentified for a wide range of HLA molecules, which rendersa vaccination feasible for a wide range of patients. Mean-while, several T-helper cell activating epitopes derived fromthe MAGE-3 gene have been described.18 This offers theopportunity of the simultaneous recruitment of CD4� andCD8� T cells through stimulation with HLA classes I and IIbinding peptides, which may result in a probably more effi-cient antitumor response. The intriguing importance ofMAGE-3 was most recently demonstrated in melanoma casessince only the induction of cytotoxic responses against thisantigen correlated with a clinical benefit.19 In the currentstudy the PRAME gene was expressed at a higher percent(61% vs 40%) than previously described.20 Since HLA-A2matched CTL activating peptides derived from the PRAMEgene have been described only recently,21 this is obviouslyanother interesting target structure. In accordance with thedata contributed by Neumann et al,20 we failed to detectMAGE-1 expression in our samples. The expression rate ofG250 was in the expected range in our series. For the G250antigen a CD4� helper T-cell stimulating epitope has alsobeen described.22

The importance of a polyvalent strategy for cellular immu-notherapy was shown in a recent study. A total of 18 patientswith stage IV melanoma were immunized with a polyvalentvaccine using dendritic cells pulsed with up to 4 melanomapeptide antigens. Regression of more than 1 tumor lesion wasobserved in 7 of 10 patients who showed enhanced T-cellresponses to more than 2 antigens. The overall immunity tomelanoma antigens was significantly associated with theclinical outcome.9

To evaluate the relevance of our findings for clinical pur-poses we calculated the number of patients suitable for sucha peptide vaccination trial based on data on HLA allele andhaplotype frequencies in Germany, and compared it toNMDP results.13, 14 We focused on the most frequent geno-types for which tumor associated antigens have been de-

FIG. 2. Ethidium bromide staining of 1.5% weight per volumeagarose gels with all investigated antigens (I). Samples or cell linesserving as positive control were MAGE-1 (K562) (a), MAGE-3 (K562)(b), PRAME (K562) (c), G250a (d) and G250b (UL-RCC-46) (e). G250PCRs and corresponding �-actin PCR were run on different gels. II,�-actin control for integrity and presence of amplifiable cDNA. M �marker.

FIG. 3. Simultaneous TAA expression in RCC samples. Y axisindicates actual number of TAA expressing samples.

TABLE 3. T-cell activating peptide motifs and calculation of eligible patients for vaccination trials

Gene of Interest(HLA type) Peptide Motif % HLA

Expression% TAA Expression

(No. pts)% Eligible forVaccination

G 250A2 HLSTAFARV 50.8 92.7 (38) 47.1MAGE-3: 85.2 61.0 (25) 52.0A1A2A24B35

EVDPIGHLYFLWGPRALVIMPKAGLLI, TFPDLESEFEVDPIGHLY

PRAME 50.8 61.0 (25) 31.0A2 VLDGLDVLL, SLYSFPEPEA, ALYVDSLFFL, SLLQHLIGL

Total of 41 patients.Peptides have T-cell immunogenicity and peptides are shown as single letter code for each amino acid.

TABLE 4. Subsequent estimated total eligibility rate

Observed/Calculated %

TAA Expression(No. pts)

CumulativeEligibility

Monovalent 19.5/17.5 (8) 29.9 /30.9Polyvalent 80.5/81.3 (33) 41.3 /40.9

Totals 100/98.8 (41) 71.2 /71.8Total of 41 patients.

RENAL CELL CANCER AND PEPTIDE VACCINATION TRIALS 2459

scribed. Our data show that about 70% of patients with RCCwould be eligible for a peptide vaccination trial. In more thanhalf of them a polyvalent approach seems to be feasible. Thisobservation is based on the assumption that there is nolinkage between the HLA phenotype and the incidence ofRCC or the expression of a certain TAA. Our calculationalgorithm may be easily adapted to further HLA genotypes,for example if affinity dependent binding assays23 may re-veal new relevant TAA/HLA molecule combinations.

We are aware that the current analysis has certain limita-tions. Clearly our analysis is a best case scenario that does notconsider that some patients have down-regulated antigen orHLA molecule expression. Also, we cannot state the antigenexpression on the protein/peptide level and how strong thosesignals must be to elicit a sufficient immune response. Accord-ing to differences in the efficacy of antigen presentation anddifferences in the T-cell receptor repertoire these factors mayconsiderably vary among patients. Finally, some patients withmetastatic disease undergo immune selection due to the heter-ogeneity of solid tumors, resulting in altered antigen expressionin the primary tumor, which we analyzed, and metastatic le-sions. However, some of these events would only appear duringtherapy and our results may hold true at least for the adjuvantsituation. The optimal approach to immunize these patientsremains to be determined but the prerequisite is always the pres-ence of immunogenic antigens on the surface of the tumor cell.

The current study focused on the eligibility of patients withRCC for clinical vaccination trials but is clearly not able topredict the outcome of such trials. For a clinical study in theemerging field of tumor immunization the therapeutic strategymust be efficient and at the same time safe and easy to monitor.Immunization with mere peptides like MAGE-3, PRAME andG250 or with peptide pulsed dendritic cells would most proba-bly fulfill these criteria because the acting antigens are knownand the immune response can be measured by ELISpot assay ortetramer staining. As we noted, the majority of patients withRCC in the adjuvant or palliative setting may be eligible for aclinical trial of peptide vaccination.

CONCLUSIONS

About 71% of patients with RCC are eligible for tumorspecific immunotherapy and in 41% of patients with RCCeven a polyvalent strategy is feasible. The estimation isbased on the expression frequencies of proven T-cell activat-ing TAA and the frequencies of the appropriate HLA class Imolecule(s) that are necessary for the presentation of theantigenic determinants toward CTLs. These data have adirect impact on the design of new phase I RCC studies.

S. Braun, K. Singer and A. Szmaragowska provided tech-nical assistance. Dr. Philipp Dahm, Duke University, NorthCarolina critically read the manuscript.

REFERENCES

1. Boon, T., Cerottini, J. C., Van den Eynde, B., van der Bruggen, P.and Van Pel, A.: Tumor antigens recognized by T lymphocytes.Annu Rev Immunol, 12: 337, 1994

2. Sahin, U., Tureci, O., Schmitt, H., Cochlovius, B., Johannes, T.,Schmits, R. et al: Human neoplasms elicit multiple specificimmune responses in the autologous host. Proc Natl Acad SciUSA, 92: 11810, 1995

3. Lu, J. and Celis, E.: Use of two predictive algorithms of the worldwide web for the identification of tumor-reactive T-cellepitopes. Cancer Res, 60: 5223, 2000

4. Jager, E., Ringhoffer, M., Dienes, H. P., Arand, M., Karbach, J.,Jager, D. et al: Granulocyte-macrophage-colony-stimulatingfactor enhances immune responses to melanoma-associatedpeptides in vivo. Int J Cancer, 67: 54, 1996

5. Tjoa, B. A. and Murphy, G. P.: Development of dendritic-cell

based prostate cancer vaccine. Immunol Lett, 74: 87, 20006. Nestle, F. O., Alijagic, S., Gilliet, M., Sun, Y., Grabbe, S., Dummer,

R. et al: Vaccination of melanoma patients with peptide- or tumorlysate-pulsed dendritic cells. Nat Med, 4: 328, 1998

7. Gaugler, B., Brouwenstijn, N., Vantomme, V., Szikora, J. P., Vander Spek, C. W., Patard, J. J. et al: A new gene coding for anantigen recognized by autologous cytolytic T lymphocytes on ahuman renal carcinoma. Immunogenetics, 44: 323, 1996

8. Jager, E., Ringhoffer, M., Karbach, J., Arand, M., Oesch, F. andKnuth, A.: Inverse relationship of melanocyte differentiationantigen expression in melanoma tissues and CD8� cytotoxic-T-cell responses: evidence for immunoselection of antigen-lossvariants in vivo. Int J Cancer, 66: 470, 1996

9. Banchereau, J., Palucka, A. K., Dhodapkar, M., Burkeholder, S.,Taquet, N., Rolland, A. et al: Immune and clinical responses inpatients with metastatic melanoma to CD34(�) progenitor-derived dendritic cell vaccine. Cancer Res, 61: 6451, 2001

10. Rieser, C., Ramoner, R., Holtl, L., Rogatsch, H., Papesh, C.,Stenzl, A. et al: Mature dendritic cells induce T-helper type-1-dominant immune responses in patients with metastaticrenal cell carcinoma. Urol Int, 63: 151, 1999

11. Heiser, A., Maurice, M. A., Yancey, D. R., Coleman, D. M.,Dahm, P. and Vieweg, J.: Human dendritic cells transfectedwith renal tumor RNA stimulate polyclonal T-cell responsesagainst antigens expressed by primary and metastatic tumors.Cancer Res, 61: 3388, 2001

12. Falk, K., Rotzschke, O., Stevanovic, S., Jung, G. and Rammensee,H. G.: Allele-specific motifs revealed by sequencing of self-peptides eluted from MHC molecules. Nature, 351: 290, 1991

13. Muller, C. R., Ehninger, G. and Goldmann, S. F.: Gene and haplo-type frequencies for the loci HLA-A, HLA-B, and HLA-DR basedon over 13,000 German blood donors. Hum Immunol, 64: 137,2003

14. Mori, M., Beatty, P. G., Graves, M., Boucher, K. M. and Milford,E. L.: HLA gene and haplotype frequencies in the North Amer-ican population: the National Marrow Donor Program DonorRegistry. Transplantation, 64: 1017, 1997

15. Agresti, A.: A survey of exact inference for contingency tables.Stat Sci, 7: 131, 1992

16. Thoenes, W., Storkel, S. and Rumpelt, H. J.: Histopathology and clas-sification of renal cell tumors (adenomas, oncocytomas and carcino-mas). The basic cytological and histopathological elements and theiruse for diagnostics. Pathol Res Pract, 181: 125, 1986

17. Thurner, B., Haendle, I., Roder, C., Dieckmann, D., Keikavoussi,P., Jonuleit, H. et al: Vaccination with mage-3A1 peptide-pulsedmature, monocyte-derived dendritic cells expands specific cyto-toxic T cells and induces regression of some metastases in ad-vanced stage IV melanoma. J Exp Med, 190: 1669, 1999

18. Chaux, P., Vantomme, V., Stroobant, V., Thielemans, K.,Corthals, J., Luiten, R. et al: Identification of MAGE-3epitopes presented by HLA-DR molecules to CD4(�) T lym-phocytes. J Exp Med, 189: 767, 1999

19. Reynolds, S. R., Zeleniuch-Jacquotte, A., Shapiro, R. L., Roses,D. F., Harris, M. N., Johnston, D. et al: Vaccine-induced CD8�T-cell responses to MAGE-3 correlate with clinical outcome inpatients with melanoma. Clin Cancer Res, 9: 657, 2003

20. Neumann, E., Engelsberg, A., Decker, J., Storkel, S., Jaeger, E.,Huber, C. et al: Heterogeneous expression of the tumor-associated antigens RAGE-1, PRAME, and glycoprotein 75 inhuman renal cell carcinoma: candidates for T-cell-based im-munotherapies? Cancer Res, 58: 4090, 1998

21. Kessler, J. H., Beekman, N. J., Bres-Vloemans, S. A., Verdijk, P., vanVeelen, P. A., Kloosterman-Joosten, A. M. et al: Efficient identifica-tion of novel HLA-A(*)0201-presented cytotoxic T lymphocyteepitopes in the widely expressed tumor antigen PRAME byproteasome-mediated digestion analysis. J Exp Med, 193: 73, 2001

22. Vissers, J. L., De Vries, I. J., Engelen, L. P., Scharenborg, N. M.,Molkenboer, J., Figdor, C. G. et al: Renal cell carcinoma-associated antigen G250 encodes a naturally processed epitopepresented by human leukocyte antigen-DR molecules toCD4(�) T lymphocytes. Int J Cancer, 100: 441, 2002

23. Kessler, J. H., Mommaas, B., Mutis, T., Huijbers, I., Vissers, D.,Benckhuijsen, W. E. et al: Competition-based cellular peptide bind-ing assays for 13 prevalent HLA class I alleles using fluorescein-labeled synthetic peptides. Hum Immunol, 64: 245, 2003

RENAL CELL CANCER AND PEPTIDE VACCINATION TRIALS2460