Serological analysis of human renal cell carcinoma

Authors

  • Gerard Devitt,

    1. Department of Tumor Progression and Immune Defence, German Cancer Research Center, 69120 Heidelberg, Germany
    2. Department of Applied Genetics, University of Karlsruhe, 76128 Karlsruhe, Germany
    Search for more papers by this author
  • Christiane Meyer,

    1. Department of Tumor Progression and Immune Defence, German Cancer Research Center, 69120 Heidelberg, Germany
    Search for more papers by this author
  • Nicole Wiedemann,

    1. Skin Cancer Unit, German Cancer Research Center, 69120 Heidelberg, Germany
    Search for more papers by this author
  • Stefan Eichmüller,

    1. Skin Cancer Unit, German Cancer Research Center, 69120 Heidelberg, Germany
    Search for more papers by this author
  • Annette Kopp-Schneider,

    1. Department of Statistics, German Cancer Research Center, 69120 Heidelberg, Germany
    Search for more papers by this author
  • Axel Haferkamp,

    1. Department of Urology, University Hospital, Ruperto Carola University, 69115 Heidelberg, Germany
    Search for more papers by this author
  • Richard Hautmann,

    1. Department of Urology, University Hospital, University of Ulm, 89075 Ulm, Germany
    Search for more papers by this author
  • Margot Zöller

    Corresponding author
    1. Department of Tumor Progression and Immune Defence, German Cancer Research Center, 69120 Heidelberg, Germany
    2. Department of Applied Genetics, University of Karlsruhe, 76128 Karlsruhe, Germany
    Current affiliation:
    1. Department of Infection and Cancer, German Cancer Research Center, 69120 Heidelberg, Germany
    • Department of Infection and Cancer, German Cancer Research Center, 69120 Heidelberg, Germany
    Search for more papers by this author
    • Fax: +6221-424760


Abstract

Serological analysis of cDNA expression libraries (SEREX) has proven to be a useful technique in the quest to elucidate the repertoire of immunogenic gene products in human cancer. We have applied the SEREX method to human renal cell carcinoma (RCC) in order to identify associated immunogenic gene products. cDNA expression libraries were prepared from a RCC tumor, a RCC cell line and human testis. The 3 libraries were screened with sera from 35 RCC patients and 15 healthy controls. Approximately 4.5 × 106 phage plaques were screened resulting in 234 positive clones, which corresponded to 74 different gene products. The seroreactivity toward 49 of these antigens was assessed. Seroreactivity to 21 (43%) of the antigens was similar in RCC patients and healthy controls, 9 antigens (18%) elicited antibodies more frequently and 19 antigens (39%) solely in RCC patients. In the reverse setting, reactivity of RCC patients' sera was tested against a panel of 44 previously identified “tumor-associated” antigens via the SADA (serum antibody detection array) method; 6 antigens reacted with RCC patients' and healthy donors' sera, 8 were reactive only with RCC patients' sera. From the 27 antigens identified by SEREX and SADA, which did not react with sera from healthy controls, 10 antigens reacted with a significant proportion of RCC patients' sera and 77% of RCC patients' sera reacted at least with one of these antigens. Sera from patients with non-malignant renal diseases or an autoimmune disease did not react with these 10 antigens. © 2005 Wiley-Liss, Inc.

Renal cell carcinoma accounts for up to 3% of all adult malignancies and is the most common neoplasm in the adult kidney. The incidence of RCC is increasing yearly.1 There is no internationally standardized treament for metastatic RCC and the only curative option for localized RCC is radical nephrectomy,2 as RCC is highly chemoresistant3 and radioresistant.4 However, RCC is thought to be immunogenic, as there are reports of spontaneous regression, albeit extremely rare.5 Identification of RCC-associated immunogenic molecules is therefore a priority in order to understand the underlying mechanisms of immunogenicity and to identify potential diagnostic, prognostic and therapeutic targets.

The repertoire of tumor antigens recognized by the immune system is referred to as the cancer immunome.6 The immunome comprises antigens defined by T cell epitope cloning,7, 8, 9 MHC peptide elution10, 11, 12 and serological methods such as SEREX (serological analysis of recombinant cDNA expression libraries)13, 14, 15 or SERPA (serological proteome analysis).16 All of these methods have been used successfully to identify antigens in a wide range of cancers. However, in comparison with other cancers such as melanoma, relatively few antigens have been identified to date in RCC.

One of the RCC-associated antigens identified by the T cell epitope cloning method is RAGE-1.17 The frequency of its expression in RCC ranges from 2 to 36% depending on which study it is read.17, 18, 19 The study by Neumann et al.18 also examined the expression of numerous other tumor-associated antigens. By RT-PCR they could detect expression of PRAME (preferentially expressed antigen in melanoma) in 40% of RCC samples and gp75 in 11%. Van Den Eynde et al.20 have identified another CTL epitope, which they named RU2AS. The epitope is derived from the protein product of the antisense mRNA of RU2S. The antigen was found not to be tumor-specific but to be a self-antigen with restricted tissue distribution.

The serological methods have had a little more success. Sahin et al. performed the first SEREX on RCC in 1995.13 They identified a gene (carbonic anhydrase XII) that was found to be overexpressed in 10% of RCC. A subsequent study identified a new member of the cancer/testis antigen family, namely SCP-1 (synaptonemal complex protein 1).21 It was found to be expressed in 3 of 36 RCC samples. Scanlan et al. analyzed 4 additional cases of RCC and found 169 clones representing 65 different gene products. Of these, 36 were coded for by known genes and 29 were novel gene products. All the genes were found to be expressed in a range of normal tissues. With regard to the immunogenicity, 12 of the antigens were found to have cancer-related seroreactivity, i.e., antibodies were detected in 5–25% of patients but not in healthy controls.22 Koreleva et al. used SEREX on 2 RCC specimens and identified 66 different genes. Of these, 18 genes were of unknown function. Four genes were found to have a cancer-related seroreactivity. However very little is known about these clones to date.23 The 2 proteomic studies of RCC have identified 12 antigens reacting with sera from RCC patients. Interestingly carbonic anhydrase I (CA I) was identified in both studies.16, 24 Two other carbonic anhydrase members, CA IX/G25025 and CA XII,26 have previously been shown to be RCC-associated antigens. The G250 antigen had been identified through a mouse monoclonal antibody specific for RCC and not normal tissue or other tumor tissues. It is expressed by more than 75% of primary and metastatic RCC and is the most promising candidate so far for immunotherapy of RCC.25

To broaden the panel of potentially diagnostic, prognostic or therapeutic targets for RCC as well as to define more completely the RCC immunome, we have performed an extensive serological analysis of this tumor type using the SEREX and SADA methodologies.

Abbreviations:

5-AZA-CdR, 5-aza-2′-deoxycytidine; CT, cancer-testis; CTCL, cutaneous T cell lymphoma; MM, malignant melanoma; RCC, renal cell carcinoma; SADA, serum antibody detection array; SEREX, serological analysis of cDNA expression libraries.

Material and methods

Cell culture, tissue samples and sera

The human renal cell carcinoma cell line KTCTL-26, a grade II clear cell RCC at stage pT2a, Nx, M1, was cultured and maintained in RPMI-medium supplemented with 10% fetal calf serum (FCS) and antibiotics/L-glutamin at 37°C, 5% CO2 in a humidified atmosphere. Treatment with 5-AZA-CdR (Sigma), was performed as described previously.27 Briefly, cells were seeded at a density of 3–4 × 105 cells/ml in a T175 tissue culture flask. When cells became firmly adherent, the medium was replaced with fresh medium containing 1 μM 5-AZA-CdR, every 12 hr for 2 days (4 pulses). Before RNA extraction, cells were cultured for an additional 48 hr in medium without 5-AZA-CdR.

Normal kidney and kidney tumor tissue samples from RCC patients were snap-frozen in liquid nitrogen immediately after surgery and stored at −80°C. Thirty-five pairs of RCC/normal kidney tissue have been used. Corresponding sera were stored at −20°C. The majority of the tumors were of the clear cell type (48%). There have been 6% of mixed type, 6% of the granular type and 6% of the oxyphil type. The histology of the remaining 12 samples is unknown. In addition, 10 sera each of patients with an autoimmune disease (psoriasis), 10 sera of patients with cutaneous T cell lymphoma (CTCL), 10 sera from patients with malignant melanoma and 10 sera of patients with non-malignant renal disease (2 sera from patients with chronic pyelonephritis and 8 sera from patients with renal calculus and accompanying nephritis) have been tested.

RNA extraction and construction of cDNA libraries

Total RNA was isolated from human tumor cell lines and human tissue samples using Tri Reagent™ (Sigma, Taufkirchen, Germany) as per the manufacturer's instructions. Poly A+ mRNA was isolated from total RNA samples with mini-oligo (dT) cellulose spin columns (Peqlab, Erlangen, Germany) as per the manufacturer's instructions.

cDNA was synthesized from the tumor of a 72-year-old woman with RCC of the mixed type and a staging of T1G1 and a clear cell RCC cell line treated with 5-AZA-CdR. The cDNA was ligated into the Uni-ZAP® XR lambda vector (Stratagene), followed by in vitro packaging. Both libraries contained ∼106 primary recombinants with an average insert size larger than 0.4 kb, which were amplified once prior to screening. The creation of the testis library has been described before.28 The library consisted of 106 primary recombinants with an insert size larger than 0.4 kb and was amplified to 1010 plaque-forming units.

Immunoscreening

To remove antibodies reacting with the vector system, sera (diluted 1:1) were preabsorbed twice on CNBr-activated Sepharose 4B (Pharmacia) columns coupled to lambda bacteriophage lysates of E. coli Xl-1 blue MRF. Sera were then diluted to 1:100 in 5% non-fat dry milk powder in TBS (Tris-buffered saline) containing Tween-20 (0.05%) (TBST) and preabsorbed 4 times (overnight at 4°C) on nitrocellulose membranes precoated with proteins derived from Escherichia coli and E. coli phage lysates.

Library screening was performed as described by Sahin et al.13 with the following modifications. Recombinant phages at a concentration of 2,000 pfu per plate were amplified for 6 hr and transferred to nitrocellulose membranes prewetted with 10 mM IPTG for an additional 15 hr at 37°C. Membranes were first washed 3 times with TBST then blocked with 5% NFDM in TBST. Membranes were then incubated in a 1:100 dilution of preabsorbed sera for 15 hr at 4°C. Following serum exposure, filters were washed 3 times in TBST and then incubated for 1 hr at room temperature, with an alkaline phosphatase-coupled secondary antibody (goat-anti-human IgG, Fc fragment; Dianova, Hamburg, Germany). Membranes were again washed 3 times with TBST and then processed with 4-nitro blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl-phosphate (NBT/BCIP). Serum-positive clones were subcloned and retested for serum reactivity as above. Positive phagemids were subcloned to monoclonality and submitted to in vivo excision of the pBluescript plasmid (protocol as described by the manufacturer; Stratagene). DNA was isolated with QIAprep spin miniprep following the manufacturer's protocol (Qiagen). The size of the insert was analyzed by XbaI/ApaI digestion and gel electrophoresis. For determining the reactivity of allogeneic sera with SEREX-derived clones, plates containing an equal number of serum-positive clones and negative control plaques were similarly processed (secondary SEREX).

DNA sequencing

Sequencing was performed commercially (MWG Biotech AG, Ebersberg, Germany) or using an automatic fluorescent sequencer (Model 377; PerkinElmer/Applied Biosystems, Forster System) and the dye terminator method according to the manufacturer's protocol (ABI PRISM Big Dye Ready Reaction Terminator Cycle Sequencing Kit, PerkinElmer). Primers were chemically synthesized (Operon Biotechnologies GmbH, Cologne, Germany).

RT-PCR and northern blotting

RT-PCR was performed on total RNA extracted from either normal kidney tissue, tumor tissue or cell lines. Reverse transcription was carried out with the ImProm-II™ reverse transcriptase (Promega) according to the manufacturers instructions.

Northern blots were prehybridized at 42°C for at least 1 hr with 15 ml prehybridization solution containing denatured salmon sperm DNA. Probe DNA (∼25 ng) was labelled with 50 μCi [α32P]dCTP using Rediprime™ II (Amersham). Before adding the denatured probe, 3 ml of 50% dextran sulphate was added to the prehybridization solution. Hybridization was performed overnight at 42°C. The next day the blots were washed for 15 min in 1× SSC/1% SDS at 42°C, 15 min in 0.2× SSC/1% SDS at 42°C and 15 min in 0.2× SSC/1% SDS at 55°C. The blots were sealed in plastic wrap and signals were detectable on X-ray film (Hyperfilm™ MP, Amersham-Pharmacia) after 1 hr to 14 days exposure at −80°C with intensifying screens. For reprobing, the blots were stripped in 40 mM Tris-HCL (pH 7.5) with 0.1× SSC and 1% SDS. For this, the solution was boiled and the blots were washed in this solution for 15 min in a boiling waterbath.

Secondary SEREX and SADA

For secondary SEREX, phages of interest were plated simultaneously at an equal density to negative control phages (i.e., without insert). Plaques were transferred to nitrocellulose membranes as before. Membranes were then cut into appropriate sized squares and screened with preabsorbed individual sera. The procedure was repeated for all clones in order to confirm results.

Phages corresponding to 44 previously identified antigens28, 29, 30 were screened with individual sera by the SADA method. Briefly, 300 μl of phages (at ∼2,000 pfu) were placed in duplicate wells of a 96-well microtitre plate. Different number of phages (24–26) were used per 96-well plate. The phages were then transferred to a rectangle shaped agar plate covered 3 hr previously with E. coli XL1-Blue MRF, top agar and 10 mM IPTG. The plates were incubated overnight at 37°C and were incubated the next day with nitrocellulose membranes for 3–4 hr. Membranes were either used immediately for screening with human serum or stored at −20°C.

Sequence analysis

Sequence analysis was performed using the HUSAR (Heidelberg Unix Sequence Analysis Resources) program from the Biocomputing Service group at the German Cancer Research Center, Heidelberg (http://genome.dkfz-heidelberg.de/). Database searches were performed with the blast programs for nucleic acids (BLASTN) as well as for protein sequences (BLASTP). Sequences were compared with those contained in the SEREX database of the Ludwig Institute for Cancer Research (http://www2.licr.org. CancerImmunome DB). Virtual northern blots were performed using the GeneFinder tool on the National Institute of Health, USA web page (http://cgap.nci.nih.gov/Genes/GeneFinder).

Statistics

Statistical analysis was done by the Fisher's Exact test.

Results

Screening of a RCC tumor and a RCC cell line library with pooled and individual sera

Phagemid libraries were constructed from RNA isolated from a RCC tumor and a RCC cell line treated with the demethylating agent 5-aza-2′-deoxycytidine (5-AZA-CdR), which has been shown to induce and/or upregulate the expression of cancer-testis (CT) antigens in neoplastic cells.27, 31

The RCC tumor library was screened initially with autologous serum. However, no positive clones could be identified. On screening with pooled sera, 5 clones (HD-RCC-1–5) were identified (Table I). Sequence analysis revealed 4 genes whose products are of unknown function: a DNA J domain containing protein (HD-RCC-1), actin associated filament protein (HD-RCC-3) and NADH Ubiquinone Oxidoreductase chain 5 (HD-RCC-4). Sequencing of HD-RCC-5 revealed the sequence at the 5′ end belonging to chromosome 19 and the 3′ end to chromosome 12. However, only the 3′ end, which corresponds to a region of chromosome 12p12 (AC006559), contains an open reading frame of 126 AA. No homologies could be found for the HD-RCC-2 clone. Screening of the cell line library with pooled and individual sera resulted in the identification of 10 clones (HD-RCC-6 to 15). All but one of the clones correspond to known genes (Table I) and four of the clones (HD-RCC-6, 7, 10 and 14) have been identified in previous SEREX studies.

Table I. Clones Identified Through The Screening of A RCC Tumor and A RCC Cell Line Library with Pooled and Individual Serum
HD-RCCcloneAccession no.IdentitySEREX homologExpected expression1Seroreactivity (%)
Normal (15)RCC (35)p-value
  • 1

    U: ubiquitously, R: restricted.

1AC008009DNAJ domain-containing protein U03NS
2HS1056L3  U03NS
3AC005383Actin-associated filament protein U03NS
4BQ614315NADH Ubiquinone Oxidoreductase chain 5 U03NS
5BC012607/AC006559Translocation Chr.19q13.2/Chr.12p12 R03NS
6NM_152233Sorting nexin 6NY-BR-48U03NS
7NM_002078Golgin subfamily a, 4 (GOLGA4)HOM-Br2-54, -MA1-12, othersU100100NS
8NM_003472DEK proto-oncogene U03NS
9BC013724Human ferritin, heavy chain U03NS
10NM_018003Uveal autoantigen (UACA)MO-BC-423, -440, othersU100100NS
11NM_152255Proteosome subunit, alpha type 7 U03NS
12NM_006793Peroxiredoxin 3 U711NS
13BC012423Superoxide dismutase 2 (MnSOD) U014NS
14NM_005751A kinase anchor protein 9MO-REN-1U2734NS
15BC012579KIAA1229, unknown protein U79NS

Only 1 out of 35 RCC patients' sera was reactive with 9 of these clones. No antibody responses could be identified against these clones in healthy controls. Five of the clones had elicited antibody responses at similar frequencies in healthy controls and RCC patients. Two of these clones (HD-RCC-7 and 10) are known autoantigens. While we detected seroreactivity in 100%, previous SEREX studies report seroreactivities between 0 and 100%. Only 1 clone (HD-RCC-13/MnSOD) demonstrated a higher level of seroreactivity in RCC patients (14%) but not in healthy controls (0%). A recent study also identified anti-MnSOD antibodies in 4 out of 6 RCC patients.24

Screening of a human testis library with pooled and individual sera

A human testis library was screened with individual sera from 35 RCC patients. Initially 2 × 104 plaques were screened. Three sera, which displayed the highest frequencies of reactivity, were used to screen 2 × 105 plaques. In total, 104 clones were identified corresponding to 59 different transcripts, 20 of which belong to unknown genes (Table II).

Table II. Clones Identified Through The Screening of Human Testis Library with Individual Serum
HD-TES cloneAccession no.IdentitySEREX homologExpected expression1Seroreactivity (%)
Normal (15)RCC (35)p-value
  • 1

    Ub: ubiquitously, Re: restricted, CT: cancer-testis antigen.

  • 2

    n.t.: not tested.

1NM_006267RAN-binding protein 2MO-BC-1083Ubn.t.2n.t. 
2BC002362Lactate dehydrogenase B (LDHB)NY-REN-46, NW-TWe 43, TC45Ub5360NS
3AC109815BAC clone  n.t.n.t. 
4NM_005054RAN-binding protein 2 like-1 (RANBP2L1)HOM-Gliom, GT 39, HOM-Ts-PMR1-11, othersCT4057NS
5AF093415Cell division protein Re2723NS
6BC036109SECIS-binding protein 2 (SBP2)MO-BC-204, NGO-Br-48 0230.08
7AC112246BAC clone  n.t.n.t. 
8AC007899RPII-53  n.t.n.t. 
9AC129915Chromosome 8 clone  011NS
10AP_000354_1Chromosome 22 clone  2746NS
11AF097485Transducin beta like-2 gene (TBL2) Re79NS
12BC008881Kinesin 2Hom-TSSemA-65, NGO-St-51, othersUb720NS
13BC047764Syntaxin-binding protein 3 Ub1311NS
14AF254756TSGA10 CT011NS
15NM_002078Golgin subfamily a, 4 (GOLGA4)HOM-Br2-54, HOM-MA1-12, othersUb100100 
16NM_016343Centromere protein FNGO-Pr-24, NGO-Br-7, othersUbn.t.n.t. 
17ALI 57387RP11-20F24  n.t.n.t. 
18BC012498Valosin-containing protein (VCIP) Ub017NS
19Z77885flow-sorted chromosome 6 TaqI fragment Ubn.t.n.t. 
20AF109135Archivillincl#11Ubn.t.n.t. 
21BC013835Beta-actinNY-SAR-52, KM-PA-3Ubn.t.n.t. 
22NM_005751A kinase anchor protein 9MO-REN-1Ub2734NS
23NM_030906Serine threonine kinase 33 (STK33) Re1317NS
24AK097129cDNA FLJ 39810  1329NS
25NM_006633RAS-GAP-related proteinTE2-35aUb714NS
26AY014284Nucleoporin NYD-SP7 Ubn.t.n.t. 
27HSA132440PLU-1 CT011NS
28BX161471DNA clone (CSODI069YD09) of placenta Ub717NS
29AF432211CLL-associated antigen KW11KW11 n.t.n.t. 
30BC059374Serine threonine kinase 31 (STK31) CT011NS
31NM_001726Bromodomain testis-specific protein (BRDT) CT2720NS
32AC114486RP11-1217.3  76NS
33AC008088.8Chr. 17 clone  n.t.n.t. 
34BC048287SEC63-like protein Ub731NS
35NM_025009Hypothetical protein FLJ13621 Ub2026NS
36NM_014648Zinc finger DAZ-interacting protein 3HOM-TSMa4-10, Lu22.2, MO-OVA-102Ubn.t.n.t. 
37AC097467RPII-27G13  76NS
38NM_002154Heat shock 70-kDa protein 4KM-PA-1, NGO-Br-43, othersUbn.t.n.t. 
39M18112Human poly(ADP-ribose) polymeraseNY-BR-59Ub2020NS
40BC012003KIAA0635NW-TK 190 n.t.n.t. 
41NM_001813Centromere protein E Ub720NS
42AL831917Thyroid hormone receptor interactor 8 Ub2729NS
43NM_003472DEK proto-oncogene Ubn.t.n.t. 
44AB014543KIAA0643MO-OVA-68, MO-CO-98Ub1317NS
45AC090768RPII-722P15  n.t.n.t. 
46NM_203292Retinoblastoma-interacting protein 8 (RBBP8) Ub13400.10
47NM_016061Yippee protein (Yippee) Ub7340.075
48NM_024581Chr. 6 ORF 60MO-TES-14, NGO-St-98Ub03NS
49BC058921Glutamyl prolyl tRNA synthetase Ubn.t.n.t. 
50AK001695Strawberry notch homolog 1 Ub2040NS
51AF458591Hypothetical protein Ubn.t.n.t. 
52BC012846Isocitrate dehydrogenase 1 Ubn.t.n.t. 
53BC001282High mobility group nuceosomal binding domain 4 Ubn.t.n.t. 
54NM_152233Sorting nexin 6NY-BR-48Ubn.t.n.t. 
55NM_003176SCP-1 meiosis specific proteinHom-TSRCC1-7, se2-1, othersCT09NS
56AY092062Breast cancer antigen U1717 Ubn.t.n.t. 
57BC015815Butyrophilin subfamiliy 3, member A3 Ub06NS
58AK025531cDNA FLJ 21878 fis Ub2031NS
59AC112246BAC clone  n.t.n.t. 

The PubMed database was examined for known clones. According to their predicted expression, the antigens belong to the following categories: CT, restricted expression, upregulated, normal/ubiquitous (autoantigens) and unknown antigens.

Six clones (HD-TES-4, 14, 27, 30, 31, 55) are predicted to be expressed only in testis tissue. These are RAN-binding protein 2 like-1 (RANBP2L1), TSGA10, PLU-1, serine/threonine kinase 31 (STK31), Bromodomain testis specific protein (BRDT) and SCP-1 meiosis specific protein (SCP-1). RANBP2L1 and SCP-1 have been previously identified by the SEREX method.21 PLU-1 and BRDT are putative CT antigens. PLU-1 has been shown to be upregulated in breast cancer.32 BRDT, also known as CT9, has been identified as a CT antigen in lung cancer.33 TSGA10 and STK31 have both been shown to be expressed in a testis-specific manner.34, 35 Their status as CT antigens, however needs further clarification.34, 35, 36

Restricted expression antigens are expressed in only a limited number of tissues, or have low level expression in many tissues but high expression in a selected few. Three clones, HD-TES-6/SECIS-binding protein 2 (SBP2), HD-TES-11/Transducin beta like-2 gene (TBL2) and HD-TES-23/serine/threonine kinase 33 (STK33) have restricted expression. SBP2 is overexpressed in testis37 and has been identified twice in SEREX studies. TBL2 is expressed as a 2.4 kb transcript predominantly in testis, skeletal muscle, heart and some endocrine tissues, with a larger (∼5 kb) transcript detected ubiquitously at lower levels.38 STK33 is expressed in many but not all tissues.39 In this study, STK33 has been isolated 3 times by 2 different sera.

Upregulated antigens are ubiquitously expressed but upregulated in cancer. Three genes, HD-TES-2, 18 and 46, fall into this category and correspond to lactate dehydrogenase B (LDHB), Valosin-containing protein (VCIP), and Retinoblastoma-interacting protein 8 (RBBP8). Serum antibodies against LDHB have been identified repeatedly by SEREX and overexpression has been found in breast cancer cell lines.40 Increased expression of VCIP has been found in pancreatic ductal, colorectal carcinoma and gastric carcinoma and correlates with lymph node metastasis and disease progression.41, 42, 43 RBBP8 is a binding partner for the tumor suppressor retinoblastoma (RB) protein. RBBP8 is expressed ubiquitously at low levels and is upregulated in many tumor cell lines.44 RBBP8 itself has not been identified previously by SEREX but two other family members (Rbbp6 and Rbbp7) have.

The greatest proportion of antigens identified in this screening are ubiquitously expressed antigens, which is also the case for most other SEREX studies.

The sequences of the 20 unknown antigens were virtually translated in all 3 frames to identify possible open reading frames (ORFs). Comparing the most likely ORFs to the protein databases using BlastP, only 8 antigens were found to have significant ORFs. Their homologies to other proteins are indicated in Table II. Expression of the clones was predicted by virtual Northern. No predictions could be made for 2 clones (HD-TES-9 and 10), 5 clones (HD-TES-28, 40, 44, 51, 58) appear to be expressed ubiquitously at low levels, HD-TES-35 is predicted to be expressed ubiquitously at high levels. Both HD-TES-40 and 58 are predicted to be upregulated in cancer. HD-TES-40 and 44 have previously been identified by SEREX.

The seroreactivity of 34 of the 59 clones was determined by secondary SEREX. Sixteen antigens (HD-TES-2, 5, 11, 13, 22, 23, 25, 28, 31, 32, 35, 37, 39, 42, 44, 58) elicit a similar frequency of antibody responses in both normal controls and RCC patients and can be considered as autoantigens. Five have been identified before by SEREX. Nine antigens (HD-TES-4, 10, 12, 24, 34, 41, 46, 47, 50) elicit antibody responses more frequently in RCC patients than in healthy controls, the difference in seroreactivity ranging from 1.4 (RANBP2L1) to 4.8 (Yippee protein) times higher in cancer patients. The remaining 9 antigens (HD-TES-6, 9, 14, 18, 27, 30, 48, 55, 57) elicit antibody responses only in cancer patients. Seroreactivity are found at frequencies of 3 to 23%. Interestingly, 4 (TSGA10, PLU-1, STK31, SCP-1) of the 9 antigens are CT antigens. VCIP (HD-TES-18) is known to be upregulated in some cancers. RCC patients' seroreactivity was found to differ significantly from the seroreactivity of controls for 2 ubiquitously expressed antigens, RBBP-8 and the Yippee protein and for SBP2, which is overexpressed in testis.

mRNA expression of SEREX clones

The mRNA expression of 38 of the 74 clones was examined by Northern blotting of 24 RCC patients (both normal and RCC tissue). Expression of the CT antigens were also examined by RT-PCR. For the CT antigens, no signals could be detected on Northern blots or by RT-PCR. This could be due to the low level expression of CT antigens. For all other antigens, signals could be detected on Northern blots. However, differential expression could be detected only in the case of HD-TES-46/RBBP8, one of the three clones with significantly different seroreactivity. Even for RBBP8, mRNA expression in RCC tissue was not significantly upregulated in all cases where seroreactivity has been observed (data not shown). Thus, exceptionally serum reactivity only could be verified by Northern blotting or RT-PCR, indicating that antibodies might be mostly either directed against posttranscriptional modifications or are mutation-specific, neither of which would be detected by Northern blot or RT-PCR.

SADA

SADA is based on the SEREX method and is used to determine seroreactivity against known antigens. We have analyzed the seroreactivity of RCC patients versus normal healthy controls for 44 different antigens, which have been identified in previous SEREX screenings of testis, melanoma and cutaneous T-cell lymphoma libraries.28, 29, 30 Table III lists the phages and summarizes the results; Figure 1 shows examples of healthy donors' and RCC patients' sera reactivity on the SADA membranes. Each antigen was spotted twice and each screening was repeated twice. Antigens were only considered positive if the duplicates were positive on both occassions. A spot was scored as positive if it was clearly darker than the spots corresponding to the negative phage. Reactivity against the tested antigens was divided into 3 categories: autoantigens, antigens to which RCC patients' sera reacted more frequently than healthy donors' sera and antigens to which exclusively RCC patients' sera were reactive.

Figure 1.

Seroreactivity of normal and RCC sera on SADA membranes: seroreactivity is shown for the serum of one healthy donor (N) and of one RCC patient (T). (a) membrane containing clones 1–22; (b) membrane containing clones 23–44. Positive clones are marked with a rectangle: (1) GOLGA4; (2) se57-1; (3) Lar-interacting protein 1a/b; (4) HEXIM-1.

Table III. SADA-Tested Clones (44)
SADA cloneAccession no.IdentityExpected expression1Seroreactivity (%)
Normal (15)RCC (35)p-value
  • 1

    U: ubiquitously, R: restricted, CT: cancer-testis antigen.

1U09284PinchU00 
2AJ243706Retinoblastoma-binding protein 2 homolog 1a (Rbbp21a)U03NS
3AF273056Retinal scDH reductase (RscDHr)U00 
4NM_022663cTAGE-5aCT03NS
5NM_019619Par-3 partitioning defective 3 homolog (Par-3ta)R0230.087
6 Negative phage (no insert)U00 
7NM_015224Retinoblastoma-associated protein 140 (RAP140)R03NS
8AJ132440PLU-1CT00 
9NM_014790KIAA0555Ub00 
10NM_022663cTAGE-1CT00 
11NM_003176Synaptonemal complex protein 1 (SCP-1)CT729NS
12AF328727Guanylate-binding protein5-taR00 
13AF273051CTCL tumor antigen se57-1 (se57-1)R017NS
14AJ000522Dynein heavy chainR00 
15NM_018204Cytoskeleton-associated protein 2 (CKAP2)R07NS
16NM_005365MAGE-A9CT00 
17NM_014635GRIP and coiled-coil domain-containing 2 (GCC-2)U2037NS
18AF430643Guanylate-binding protein-5aR00 
19AF181985Serine/threonine kinaseU00 
20NM_002078Golgin subfamily a, 4 (GOLG A4)U8794NS
21NM_007152Zinc finger protein 195U00 
22U19144GAGE-3CT03NS
23AC046174RP11_157D8U00 
24NM_002748Mitogen-activated protein kinase 6U00 
25NM_000462Ubiquitin protein ligase E3AU00 
26NM_014706Squamous cell carcinoma antigen recognized by T cells 3U00 
27X77567InsP3 phospataseU00 
28U14603Protein Tyrosine phospataseU00 
29BC015153Nuclear distribution gene C homologU00 
30NM_014497NP220 nuclear proteinU00 
31D14530Ribosomal proteinU00 
32U22815/6LAR-interacting protein 1a/bU10097NS
33AB021179Hexamethylene-bis-acetamide-inducible transcript (HEXIM-1)U011NS
34BC017297Hypothetical protein BM-009 00 
35NM_004226Serine/threonine kinase 17bU00 
36AY040871Cell division autoantigen 1 (CDA1)U726NS
37AC020978Unknown/chromosomal sequence 00 
38BC001887Zinc finger protein 133 (ZNF133)U7400.021
39NM_001006945RIF1 variant 2U00 
40BC007446Putative S1 RNA binding domain proteinU00 
41NM_005362MAGE-A3CT00 
42AC011726Chromosome 8 est sequence 00 
43BC019669Elongation factor 1 alphaU00 
44AY029066HumaninU00 

The GOLGA4 antigen (also identified in the SEREX screening; HD-RCC-7) and the Lar-interacting protein both fall into the category of autoantigens. GOLGA4 is already known as an autoantigen. The Lar-interacting protein has not been identified before as an autoantigen. While it was not the original intention, the presence of these autoantigens on the SADA membranes served as positive controls. A higher percent reactivity in RCC patients' than in healthy controls' sera was seen against 4 antigens (SCP-1, GCC-2, CDA1, ZNF 133), the difference varying between 29% versus 7% for SCP-1 and 40% versus 7% for ZNF 133. Seroreactivity against 8 antigens (Rbbp21a, cTAGE5a, Par-3ta, RAP 140, se57-1, CKAP2, GAGE3, HEXIM-1) was only found in RCC patients' sera, with reactivity frequencies varying from 1 to 8 patients' sera.

Selectivity of RCC-reactive sera

The SEREX and SADA screening revealed 27 antigens that were recognized by RCC patients', but not by healthy donors' sera. Antibodies against 10 of these antigens were detected in at least 11% of the RCC patients' sera. Thus, these antigens could be expressed preferentially in RCC or tumor tissue in general. To substantiate the hypothesis, sera of 10 patients with infectious kidney disease, 10 patients with an autoimmune disease (psoriasis), 10 patients with a lymphoid malignancy (cutaneous T-cell lymphoma, CTCL) and 10 patients with malignant melanoma (MM) were screened for reactivity against these antigens. This screening should provide evidence whether the antigens are kidney-, rather than RCC-specific, whether they may provoke an immune response based on an altered immune steady state (autoimmune disease) or whether they are generally upregulated/presented in an immunogenic form in different malignancies. All sera were also tested against the uveal autoantigen (UACA) as a positive control.

As summarized in Table IV, all sera reacted with UACA. Sera of two patients with CTCL reacted with Par-3ta and one patient with a CTCL reacted with se57-1. This is not surprising, as these two antigens were originally identified in a SEREX screening of a testis library with CTCL sera. None of the sera from patients with MM showed any reactivity. Importantly, too, none of the sera from patients with psoriasis and none of the sera from patients with non-malignant renal disease reacted with anyone of the 10 antigens. Thus, these 10 antigens clearly elicit antibody responses in a tumor-restricted manner and preferentially in RCC.

Table IV. Tumor/RCC Specificity of Selected Clones
CloneSeroreactivity of patients with
RCCNon-malignant renal diseasePsoriasisCTCLMMHealthy
UACA35/3510/1010/1010/1010/1015/15
Par-3ta8/350/100/102/100/100/15
SBP27/350/100/100/100/100/15
VCIP7/350/100/100/100/100/15
MnSOD6/350/100/100/100/100/15
se57-16/350/100/101/100/100/15
HD-TES-94/350/100/100/100/100/15
TSGA104/350/100/100/100/100/15
PLU-14/350/100/100/100/100/15
STK314/350/100/100/100/100/15
HEXIM-14/350/100/100/100/100/15

The overall reactivity profile of RCC patients' sera against these 10 antigens has been summarized in Figure 2. Antibodies were found in 11–23% of RCC patients and 77% of RCC patients had antibodies to at least one of these antigens. As far as it can be judged from the limited number of different histological types of RCC, reactivity against these antigens does not appear to be histological type-restricted.

Figure 2.

“RCC-specific” seroreactivities: seroreactivity against the panel of 10 antigens defined by SEREX and SADA, which were not recognized by sera from healthy donors, patients with non-malignant renal disease, psoriasis or MM, is shown. *Squamous cell carcinoma of renal pelvis

Discussion

Antibodies detected by SEREX are high-titered IgG, the generation of which requires T cell help. SEREX defined antigens are thus thought to reflect a T cell-mediated immune response and the SEREX as well as modifications of the SEREX method are, besides others, used to identify potential targets for cancer immunotherapy. In this study, 3 cDNA libraries (RCC tumor, RCC cell line and testis) and a panel of 44 previously identified antigens have been screened with sera from RCC patients in an attempt to elucidate, at least in part, the RCC immunome. RCC patients' sera recognized 74 antigens in the SEREX screening and 14 out of 44 known antigens screened by SADA. We want to briefly discuss the methods and the identified antigens.

The vast majority of clones identified by SEREX are ubiquitously expressed genes. In this respect our study compares well with other SEREX studies. The low number of clones identified in the RCC tumor and the RCC tumor line library, however, needs to be commented. In most SEREX studies, sera are diluted 1:100. We considered a higher dilution (1:3,000) as more appropriate to select for strong immunogenic entities. Though this consideration obviously was correct, we may have missed tumor-related antigens, as the majority of clones depicted were known autoantigens, like, e.g., potent GOLGA4, UACA and DEK.45, 46, 47 In the case of the tumor-derived library, it may also be that this patient's tumor is not very immunogenic. In fact, not one single clone was identified in the small-scale screening of the testis library using the serum corresponding to the patient from which the tumor library was derived. Because of the low efficiency of the screening with high serum dilution, we proceeded with the more common 1:100 dilution in the screening of the testis library and the SADA screening. Also, by performing an initial small screening of the testis library with each of the 35 RCC sera, the most reactive sera could be identified. This significantly enhanced the number of clones identified, which besides known ubiquitously expressed genes (27) included unknown genes (20) and a decent number of upregulated and CT antigens (9).

One drawback of the SADA method is the frequently observed high background, which may explain some of the discrepancies between equivalent clones screened by both secondary SEREX and SADA (e.g. SCP-1). We consider the inclusion of known autoantigens (in our study GOLGA4 and/or the UACA) as positive control to be quite helpful and recommend this as a general strategy. Still, the SADA method was less sensitive than secondary SEREX. Of 44 known antigens, only 14 were recognized by healthy donors' or RCC patients' sera. In comparison, all of the clones identified in the testis screen were recognized by at least one person's serum by secondary SEREX. However, the lack of sensitivity of SADA screening may also be an advantage for identifying cancer-related clones. Indeed, 57% of SADA, but only 26% of secondary SEREX-depicted clones react in our study exclusively with RCC sera.

We also want to comment on the failure, with one exception, to confirm upregulated gene expression by Northern blot or RT-PCR. We previously searched for RCC-associated gene expression by SSH and DD RT-PCR48, 49, 50 and expected to find upregulated gene expression for at least some of the gene products identified by SEREX and SADA. However, it is known, that the majority of antibodies recognize conformational epitopes and posttranslational modifications, which would not be reflected at the RNA level. The same accounts for point mutations, which again are a likely source of an antigenic entity.51 Although we are repeating the search for differential RNA expression, functional assays to screen the frequency and efficacy of T cell responses may be of greater relevance and we have started an immunogenicity analysis for two of the antigens.

The majority of the clones identified by SEREX and SADA represent autoantigens. The significance of autoantibody responses should however not be dismissed. Antibodies directed against normal antigens can be considered part of the natural autoantibody (na-Ab) repertoire. Na-abs were first suggested to have a “sewage” role whereby na-Ab are responsible for inactivating any biologically active molecules that are overproduced.52 Some na-Ab provide transportation of molecule-ligands as well as protection of ligands from proteolysis53 or possess enzymatic activity.54 The finding that some antibodies can modulate functions of intranuclear proteins in vivo, suggests that antibodies are also capable of translocation across cell membranes.55 The term Immunculus (immunological homunculus) has been proposed to describe the network of constitutively expressed na-Ab.56 Poletaev and Osipenko suggest that the Immunculus mirrors the organism's physiological state. In healthy people the Immunculus is relatively constant and is characterized by minimal individual quantitative variations. However, abnormal metabolic deviations, which precede or accompany different diseases, result in quantitative rather than qualtitative changes in this na-Ab network. Quantitative differences to the state of na-Ab in the healthy condition may be used for the early detection of potentially pathogenic metabolic changes including cancer.56 Thus, 15% of the clones elicited antibodies more frequently in RCC patients than in healthy controls. For some of these clones the difference in seroreactivities is relatively low. However, antibodies against 5 antigens have been detected significantly more frequent in RCC patients' sera (RBBP-8, Yippee, Par-3ta, ZNF-133, SBP2). Whether the increased detection of these antibodies in RCC patients reflects quantitative changes to the healthy steady state na-Ab network remains to be explored.

Twenty seven clones elicited antibody reactions only in RCC patients. Antibodies against 17 clones were found in only 1 or 2 of the 35 patients. These 17 clones may represent tumor-specific antigens and are thus not suitable for screening procedures. However, antibodies against 10 antigens were found in 11–23% of RCC patients and 77% of RCC patients have antibodies to at least one of these antigens. In addition, sera from patients with non-malignant renal diseases and from patients with psoriasis did not react with these antigens. Also, only 2 sera of patients with CTCL reacted with Par-3ta and 1 serum with se57-1, both antigens were originally identified by screening sera of CTCL patients. Notably, too, antibodies against these antigens were at least not obviously associated with the histological type of RCC, although—with the exception of clear cell renal cell carcinoma—far more samples would be required to confirm or disprove a statistically significant association between the histological type of RCC, the antigen expression profile and seroreactivity.

Without going into detail, the 10 antigens that induced antibodies preferentially in RCC patients may be briefly introduced. The mitochondrial antioxidant enzyme Mn superoxide dismutase (HD-RCC-13) has been reported to be both upregulated and downregulated in RCC and other cancer types.57, 58 SBP2 is a key player in the synthesis of selenocysteine and selenoproteins. Most of these enzymes are involved in redox reactions, which are frequently distorted in cancer patients.59 Some are involved in maturation, which is in line with the high expression of SBP2 in testis.37 TSGA10 was originally identified as a testis-specific gene. It is predicted to play a role in the sperm tail fibrous sheath.34 Overexpression was found in a variety of cancers.36 VCIP is a member of the “ATPases associated with various cellular activities” superfamily and plays a key role in the ubiquitin-dependent proteasome degradation pathway.60 VCIP is known to inhibit apoptosis via degradation of the inhibitor of NF1κB1α.60 Though upregulated in several cancers,41, 42, 43 VCIP antibodies have not been demonstrated for any cancer. Whether VCIP-specific antibodies are selectively found in RCC remains to be seen. PLU-1 is a nuclear protein proposed to function as a regulator of gene expression by interacting with the developmental transcription factors BF-1 and PAX9.61 In the adult, PLU-1 is expressed only in testis and ovary.32 The Par-3ta gene belongs to the family of partitioning-defective genes, which are involved in asymmetric cell division and polarized growth.62 In normal tissue, 3 splicing products of Par-3ta have been identified, whereas in tumor tissue only one intense band is seen (S.Eichmüller, unpublished observation). The clone used here (Par-3ta) is lacking the first of 3 PDZ domains and appears to be a tumor-specific splice variant of the Par-3ta gene. The se57-1 clone, identified in a SEREX screening of CTCL,28 is expressed in bone marrow, colon, small intestine, spleen, testis and trachea. Nothing else is known about this gene. The HEXIM-1 gene plays an inhibitory role in NF-κB-dependent gene expression in vascular smooth muscle cells.63 Two SEREX studies have identified seroreactivity against HEXIM-1 only in cancer patients.29, 64 STK31 has been identified recently in mouse spermatogonia but not in somatic tissues.35 Nothing else is known about this protein. A more detailed analysis of HD-TES-9 expression and function remains also to be done, because no homologies were found for the translated sequence of the 1 possible ORF and prediction of its expression by virtual northern was not possible.

We previously searched for RCC antigens by SSH and DD RT-PCR.48, 49, 50 As compared to the overall efficiency of these screenings, SEREX and SADA were of comparable efficacy. However, SEREX and SADA are without question, the methods of choice when searching for potential targets of an easy to perform serological screening. Besides CTL libraries and peptide elution, SEREX and SADA screenings are also well suited as a first step towards the identification of immunotherapeutic targets.

Our results indicate an enormous variability of the antigenic pattern of individual RCC, the reasons for which are unknown. We therefore suggest that a panel of RCC antigens should be used for diagnostic screening and for selecting the target antigen for immunotherapy. The combination of the 10 antigens identified here plus selected antigens from previous SEREX analysis of RCC may provide a starting basis.

Acknowledgements

We cordially thank S. Hummel for most valuable technical help.

Ancillary