• tumor immunology;
  • astrocytoma;
  • antigens


  1. Top of page
  2. Abstract
  6. Acknowledgements

The molecular characterization of antigens preferentially or exclusively expressed by astrocytomas and recognized by the autologous immune system are a prerequisite for the development of specific vaccines. To identify such antigens, we screened 5 cDNA expression libraries derived from astrocytomas and other gliomas for reactivity with high-titered IgG antibodies in the sera of astrocytoma patients using SEREX, the serologic identification of antigens by recombinant cDNA expression cloning. Autologous and allogeneic SEREX analysis of >5 × 106 clones with the sera of 18 astrocytoma patients revealed 10 antigens: the differentiation antigen glial fibrillary acidic protein (GFAP), Bax-inhibitor 1 (which was overexpressed in all glioma samples tested), 3 other molecules involved in the regulation of gene expression and proliferation (the nm23-H2-encoded nucleoside diphosphate kinase B, the Ran binding protein-2 and a DNA binding protein encoded by the son gene), SP40,40 (a complement inhibitory molecule), the chaperonin TCP-1, calnexin and 2 new gene products. No immune responses were detected against the “shared tumor” or “cancer testis antigens” that are regularly expressed in gliomas. Antibody responses in astrocytoma patients against antigens expressed by gliomas were rare and, with the exception of Bax-inhibitor 1 and the product of the son gene, were also found in apparently healthy controls. We conclude that although astrocytomas express a broad spectrum of antigens, they elicit antibody responses only rarely, most likely because of their intrinsic immunosuppressive effects. © 2001 Wiley-Liss, Inc.

A variety of immunotherapeutic and genetherapeutic strategies have been pursued in patients with malignant brain tumors to improve on results obtained with surgery, radiotherapy and chemotherapy.1, 2 A prerequisite for the success of tumor-specific immunotherapeutic strategies is the existence and identification of genes that are either exclusively or preferentially expressed in malignant compared with normal tissues.

We have recently shown3 that human gliomas express the so-called shared or cancer-testis antigens4 at a similar frequency as malignant melanomas, in which most of these antigens had been originally described. Due to their expression in a broad spectrum of tumors of different origins, but absence in normal tissues except for the testis, these antigens are ideal candidates for cancer vaccine development. However, the identification of additional tumor-specific antigens in human gliomas with a frequent expression in gliomas is warranted to allow for the development of widely applicable polyvalent glioma vaccines. To detect more of such antigens in astrocytomas, we analyzed the B-cell repertoire of 18 patients with astrocytomas and glioblastomas against antigens expressed by autologous and allogeneic astrocytomas and other gliomas using SEREX, the serologic analysis of antigens using recombinant cDNA expression cloning.5 Our results show that even though astrocytomas express a wide range of antigens, antibody responses against these antigens are rare in glioma patients.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Patients, sera and tissues

The study was approved by the local ethical review board (Ethikkommission der Ärztekammer des Saarlandes). Recombinant DNA work was done with the official permission and according to the rules of the State Government of Saarland. Sera and tissues were obtained during routine diagnostic or therapeutic procedures performed at Saarland University Medical School, Homburg, Germany and stored at −80°C until use. Normal tissues were collected from autopsies of tumor-free patients. Tumor samples used for SEREX and reverse transcriptase PCR analysis was checked microscopically for the presence of neoplastic tissue and the absence of contaminating normal brain tissue. The World Health Organization (WHO) brain tumor classification and the Daumas-Dupont/SAMS grading of astrocytomas were used for histologic diagnosis. Normal tissues were collected from autopsies of tumor-free patients.

Construction of cDNA expression libraries

Fresh glioma tissues were used to establish cDNA libraries, as described by us.6 Poly(A)+ RNA was prepared with an mRNA isolation kit (Stratagene, La Jolla, CA) out of total RNA isolated from fresh tumor biopsies.7 cDNA expression libraries were constructed by starting with 5–8 μg of poly(A) RNA. First-strand synthesis was performed using an oligo(dT) primer with an internal XhoI site and 5-methylCTP. cDNA was ligated to EcoRI adaptors and digested with XhoI. cDNA fragments were cloned in the sense direction with respect to the lacZ promotor into the bacteriophage expression vector λ-ZAP Express using a commercially available adaptor ligation system according to the manufacturer's instructions (Stratagene). After packaging into phage particles, these cDNA expression libraries were used to transfect Escherichia coli. The cDNA derived from a mixture of several different glioma tissues (see below) was constructed by starting with a mixture of 1 μg each of poly(A) RNA from each glioma sample.


The sera from patients with untreated gliomas that had been preabsorbed with transformed E. coli lysates were screened for IgG reactivity against transfectants derived from the different cDNA expression libraries using the phage assay described by us in detail previously.6 XLI MRF′ bacteria transfected with recombinant λ-ZAP Express phages were plated onto Luria-Bertani agar plates. Expression of recombinant proteins in lytic phage plaques on the bacterial lawn was induced with isopropyl-β-D-thiogalactopyraoside (IPTG). Plates were incubated at 37°C until plaques were visible and then blotted onto nitrocellulose membranes. The membranes were blocked with 5% low-fat milk in Tris-buffered saline and incubated with a 1:250 dilution of the patient's serum, which had been preabsorbed with transfected E. coli lysates. Serum antibodies binding to recombinant proteins expressed in lytic plaques were detected by incubation with alkaline phosphatase-conjugated goat anti-human IgG and visualization by staining with 5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetrazolium.

Sequence analysis of identified antigens

Positive clones were subcloned to monoclonality and submitted to in vivo excision of pBK-CMV phagemids. The nucleotide sequence of cDNA inserts was determined by using an Excel cycle sequencing kit (Epicentre Technologies, Madison, WI) on a LICOR automatic sequencer. Sequencing was performed according to the manufacturer's instructions starting with the vector-specific primers. Insert-specific primers were designed as the sequencing proceeded. Sequence alignments were performed with DNASIS (Pharmacia Biotech, Gaithersburg, MD) and BLAST software on EMBL, Genbank and PROSITE databases.

Detection of antibodies against defined antigens

Monoclonalized phages from clones that were reactive with the patient serum used for the immunoscreening were mixed with nonreactive phages of the cDNA library as internal negative controls at a ratio of 1:10 and used to transfect bacteria. IgG antibodies in the 1:100 diluted E. coli-preabsorbed sera from other patients and healthy controls were tested with the above-described phage assay to assess for antibody responses against the respective antigens.

Northern blot analysis

Northern blots were performed with RNA extracted from tumors and normal tissues as described before.8 Integrity of RNA was checked by electrophoresis in formalin/3-(N-morpholino)propane sulfonic acid (MOPS) gels. Gels containing 10 μg RNA/lane were blotted onto nylon membranes. After prehybridization the membranes were incubated with the specific 32P-labeled insert-specific cDNA probes overnight at 65°C in hybridization solution (50% formamide, 6 × standard saline citrate [SSC], 5 × Denhardt's solution, 0.2% SDS). The membranes were then washed at progressively higher stringencies, with the final wash in 1 × SSC and 0.2% SDS at 65°C. Autoradiography was conducted at −70°C for up to 7 days using Kodak X-OMAT-AR film and intensifying screen. After exposure the filters were stripped and rehybridized with a glyceraldehyde-3-phosphate dehydrogenase probe to prove RNA integrity.


  1. Top of page
  2. Abstract
  6. Acknowledgements

More than 5 × 106 recombinant clones represented in a variety of cDNA expression libraries derived from different astrocytomas and other gliomas were screened by the SEREX approach using the sera of 18 patients with gliomas. The patients from whom the sera had been taken had no obvious other disease apart from their glioma; in particular, the sera were negative for antibodies against common autoantigens such as antinuclear antibodies or rheumatoid factor. The results of this analysis are shown in Tables I and II. Fourty-eight clones encoded by 10 different genes reacted with IgG antibodies in the 1:250 diluted sera from the patients with astrocytomas.

Table I. Summary of SEREX Analysis of Astrocytomas and Other Gliomas Using Sera From Patients With Gliomas
SEREX analysisSource of cDNA1SerumClones positive/clones testedNo. of different antigens
  • 1

    Mixture of gliomas, 2 grade II astrocytomas, 5 glioblastomas, 2 oligodendrogliomas and 1 ependymoma.

1Astrocytoma grade IIAutologous astrocytoma grade II42/1.2 × 1066
2GlioblastomaAutologous glioblastoma1/1.2 × 1061
3Astrocytoma grade IIAllogeneic astrocytoma grade II2/1.0 × 1061
4GlioblastomaAllogeneic glioblastoma0/1.0 × 1060
5Mixture of gliomasAllogeneic astrocytoma grade II2/2 × 1041
6Mixture of gliomasAllogeneic astrocytoma grade II0/2 × 1040
7Mixture of gliomasAllogeneic astrocytoma grade II0/2 × 1040
8Mixture of gliomasAllogeneic astrocytoma grade II1/2 × 1041
9Mixture of gliomasAllogeneic astrocytoma grade III0/2 × 1040
10Mixture of gliomasAllogeneic glioblastoma0/2 × 1040
11Mixture of gliomasAllogeneic glioblastoma0/2 × 1040
12Mixture of gliomasAllogeneic glioblastoma0/2 × 1040
13Mixture of gliomasAllogeneic glioblastoma0/2 × 1040
14Mixture of gliomasAllogeneic glioblastoma0/2 × 1040
15Mixture of gliomasAllogeneic glioblastoma0/2 × 1040
16Mixture of gliomasAllogeneic glioblastoma0/2 × 1040
17Mixture of gliomasAllogeneic glioblastoma0/2 × 1040
18Mixture of gliomasAllogeneic glioblastoma0/2 × 1040
Table II. Antigens Detected by SEREX Analysis of Gliomas With Sera From Glioma Patients
AntigencDNA sourceSerumExpression specificity
SonAstrocytoma IIAutologousUbiquitous
SP-40,40Astrocytoma IIAutologousUbiquitous
BAX-inhibitor 1Astrocytoma IIAutologousOverexpressed in gliomas
HOM-GLIO-101Astrocytoma IIAutologousWidely expressed in tissues
2410004N11Rik homologAstrocytoma IIAutologousUbiquitous
Chaperonin TCP1Astrocytoma IIAutologousUbiquitous
nm23-H2Astrocytoma IIAllogeneic astrocytoma IIWidely expressed in tissues
GFAPGlioma mixAllogeneic astrocytoma IIAstrocytes
RanBP2Glioma mixAllogeneic astrocytoma IIUbiquitous

Detection of astrocytoma antigens by antibodies in autologous sera

Two patients, 1 with a grade II astrocytoma and 1 with a glioblastoma (formerly grade IV astrocytoma), were analyzed for antibodies against antigens expressed by their autologous tumor. The analysis of ∼1.2 × 106 expressed clones from each cDNA by the autologous sera yielded 42 clones that were detected by the pretreatment serum from patient 1, but only 1 clone was detected by antibodies in the serum of the second patient.

The sequence analysis revealed that 36 of the 42 clones detected in patient 1 were encoded by the son gene, which codes for a DNA binding protein; 1 clone each represented the product of the human SP-40,40 gene (a complement-inhibitory protein), the BAX- inhibitor 1 (BI-1), which is encoded by the formerly described tegt (testis-enhanced gene transcript) gene, and the chaperonin TCP 1. The insert of an additional clone is encoded by the human homolog of the recently described murine gene 2410004N11Rikm, which is ubiquitously expressed in benign and malignant tissues9 and 1 insert (HOM-GLIO-101) exhibited no homology with known genes listed in the gene bank. Its sequence has been submitted to the EMBO gene bank (accession number Aγ413270). The only clone detected by the autologous serum from patient 2 was identified as calnexin, a membrane protein of the endoplasmic reticulum.

Detection of astrocytoma antigens by antibodies in the sera of allogeneic astrocytoma patients

Clones (1 × 106) of an additional grade II astrocytoma and a glioblastoma, respectively, were screened using the sera from allogeneic patients with untreated glioma of the respective grade. One clone, representing nucleoside diphosphate kinase B, which is encoded by the nm23-H2 gene, was detected by antibodies in the patient with astrocytoma grade II, whereas no antigens were detected by antibodies in the serum of the glioblastoma patient.

Detection of antigens expressed by a variety of gliomas by antibodies in the sera of patients with astrocytomas of different grades

The results of the screening with individual sera and tumors made it evident that antibodies against antigens expressed by individual astrocytomas are rather infrequent in patients with astrocytomas. We therefore decided to construct a cDNA library from a mixture of 10 glioma biopsies of different histologic types; these included 2 astrocytoma grade II, 5 glioblastomas, 2 oligodendrogliomas and 1 ependymoma. Clones (>20,000) of the expression cDNA derived from this glioma mixture were tested with the sera from 14 glioma patients: 4 patients with astrocytoma grade II, 1 patient with astrocytoma grade III and 9 patients with glioblastoma. Only 2 antigens were identified, each detected by only 1 serum: 1 antigen represented glial fibrillary acidic protein (GFAP), a marker of astrocytic differentiation, and was represented 10 times among 20,000 clones tested. The second antigen, which was represented only once, was encoded by the RanBP2 gene, which codes for a protein containing binding sites for the GTPase Ran.

Northern blot analysis of selected antigens

To investigate whether the differential expression in different tissues might be responsible for the immunogenicity of the antigens expressed by the astrocytomas, we analyzed the expression of some of the antigens (for which this information is not available from the literature) by Northern blot analysis. Analysis of the antiapoptotic protein BI-1 revealed that its expression was between 5 and 10 times stronger in all 16 glioma samples tested compared with normal brain and other normal tissues (Table III, Fig. 1). In contrast to BI-1, the nucleoside diphosphate kinase B, which is encoded by the nm23-H2 gene, was not overexpressed in the tumor where it had been detected (data not shown); moreover, analysis of the nm23-H2 gene in normal and malignant tissues revealed that it was expressed in almost all normal tissues except for lung and rectum and in a wide spectrum of malignant tumors of different origins (Table IV). Similarly, the new gene HOM-GLIO-101 was expressed at a similar level in all benign and malignant tissues tested (Table V).

Table III. Northern Blot Analysis of BAX-Inhibitor 1 Expression
TissueNo. tested/no. positiveExpression level
Normal brain8/8Baseline
Normal stomach2/2Baseline
Normal liver1/1Baseline
Normal colon2/2Baseline
Normal kidney1/1Baseline
Peripheral blood lymphocytes3/3Baseline
Astrocytomas of different grades16/165–10 × baseline
Raji (Burkitt lymphoma cell line)1/12–5 × baseline
L540 (Hodgkin's-derived cell line)1/12–5 × baseline
Karpas (anaplastic large cell lymphoma)1/12–5 × baseline
thumbnail image

Figure 1. Northern blot analysis of the expression of the antiapoptotic protein Bax-inhibitor 1 (BI-1) in normal brain and in a glioblastoma. From left to right: cerebellum; frontal lobe; temporal lobe from normal brain; glioblastoma. Upper row: expression of BI-1; lower row: expression of GAPDH.

Download figure to PowerPoint

Table IV. Northern Blot Analysis of nm23-H2 Expression
TissuePresence or absenceNo. tested/no. positive
 Brain, frontal cortex+
 Brain, temporal cortex+
 Basal ganglia+
 Lung cancer1/2
 Hepatocellular carcinoma1/1
 Prostate cancer1/1
 Rectal carcinoma1/1
 Colon cancer1/1
 Breast cancer1/1
Table V. Northern Blot Analysis of HOM-GLIO-101 Expression
TissuePresence or absenceNo. tested/no. positive
 Normal brain+
 Adrenal glands+
 Blood lymphocytes+
 Breast cancer1/2
 Colon cancer1/1
 Lung cancer1/2

Specificity of antibody responses against glioma-associated antigens

To determine whether antibody responses against the antigens expressed by the astrocytomas are specific for glioma patients, we tested the sera of glioma patients and healthy controls against the 10 antigens detected in our study. As shown in Table VI, antibody responses against the glioma-associated antigens were rarely found in glioma patients; no antibodies against BI-1 and the son gene product were found in healthy controls; however, antibody responses against these antigens were also infrequent in astrocytoma patients (1/12 and 2/12, respectively). Antibody reactivities against GFAP occurred in 5/20 glioma patients, but also in 1/20 healthy controls, whereas antibodies against the other antigens occurred in the sera of glioma patients and healthy controls with similar frequencies.

Table VI. Antibody Reactivity Against Glioma-Associated Antigens in Glioma Patients and Healthy Controls
AntigenGlioma patients (no. positive/no. tested)Healthy controls (no. positive/no. tested)
BAX-inhibitor 11/120/17
2410004N11Rik homolog1/120/12
Chaperonin TCP 12/121/17
Ran binding protein 21/201/17


  1. Top of page
  2. Abstract
  6. Acknowledgements

Our analysis shows that astrocytomas express a wide spectrum of antigens. Although a classification of the antigens detected using the sera of 18 astrocytoma patients is difficult, it is intriguing that many of them are involved in gene expression and proliferation: the son gene encodes a conserved DNA binding protein mapping to human chromosome 21;10 the human nucleoside diphosphate kinase B, which is encoded by the gene nm23-H2, has the distinctive property of stimulating gene transcription;11 Ran-binding protein 2, a protein containing FG repeat motifs and 4 binding sites for the guanosine triphosphatase Ran, which is localized at the cytoplasmic periphery of the nuclear pore complex, is believed to play a critical role in nuclear protein import12 and in the nuclear export pathway;13 and finally, BI-1, which is a regulator of cell death pathways controlled by Bcl-2 and BAX.14

When overexpressed in mammalian cells, BI-1 suppresses apoptosis induced by Bax, etoposide, staurosporine and growth factor deprivation, but not by Fas (CD95). We now demonstrate for the first time that the antiapoptotic protein Bax-inhibitor 1 is overexpressed in gliomas. The overexpression of BI-1 in gliomas may confer a growth advantage on the respective neoplastic cells and might be the underlying cause for its immunogenicity in astrocytoma patients. However, whereas BI-1 was found to be overexpressed in all glioma samples analyzed, only 1 of 12 patients tested developed an antibody response against the product of this overexpressed gene.

Another interesting observation of our study was the demonstration that GFAP is immunogenic in humans, thus representing a differentiation antigen that is immunogenic in astrocytoma patients. The immunogenicity of GFAP in 2 patients with vascular dementia and a dementia associated with autoimmune disease has recently been reported.15 However, our study now shows that antibodies against GFAP can also be detected in healthy controls. This speaks against the hypothesis that underlying autoimmune, vascular or neoplastic disease of the central nervous system is a prerequisite for the development of such autoantibodies.

Our finding that the complement-inhibitory protein SP40,40 elicits immune responses in glioma patients is of interest in light of reports that SP-40,40 is expressed and deposited in brains with Alzheimer′s disease, where it is closely associated with amyloid deposition and seems to play a role in the pathogenesis of amyloid formation.16 It might thus be interesting to investigate patients with Alzheimer's disease for the presence of antibodies against this newly identified autoantigen.

The fact that a wide spectrum of normal structures of the cell that are neither exclusively expressed nor overexpressed in gliomas compared with normal tissue can become immunogenic in astrocytoma patients suggests that the context in which a protein is presented to the immune system (e.g., the context of “danger”) is more decisive for its immunogenicity (and breaking of tolerance) than its more or less restricted expression in a given tissue. According to this concept of contextual immunogenicity, the primary driving force of the immune system is not to distinguish between self and non-self, but rather the need to detect and protect against danger.17

Having demonstrated a comparatively frequent expression of the shared tumor or cancer testis antigens in human gliomas,3 we had expected frequent antibody responses against glioma-associated antigens to be detected by the SEREX analysis of astrocytoma patients. However, in contrast to our expectations, our study revealed that antibody responses against antigens expressed by astrocytomas are rare in patients with astrocytomas. Even though we screened >5 × 106 clones derived from 14 different gliomas with the sera from 18 astrocytoma patients, only 10 antigens were detected. This is far below the frequency of antigens detected that we and others have experienced with the SEREX analysis using the sera from patients with other tumors or autoimmune diseases.4, 18, 19, 19a The extensive analysis using the sera from 18 different patients with astrocytomas and expression libraries derived from different gliomas has met with limited success with respect to the identification of new astrocytoma antigens suitable for vaccine development. This also holds true for the antiapoptotic protein BI-1, which is strongly overexpressed in all glioma samples tested compared with normal brain; nevertheless, its significant baseline expression in normal brain would preclude its use as a target for immune responses induced by specific peptide or whole protein vaccines.

The reason for the paucity of (humoral) immune responses in astrocytoma patients against antigens expressed by their tumors is probably not due to the lack of expression of antigens by human gliomas, as shown by our recent analysis of the expression pattern of a whole panel of the so-called shared tumor antigens in human brain tumors.3 This analysis showed that human gliomas (oligodendrogliomas, oligoastrocytomas and astrocytomas) express cancer testis genes at a frequency comparable to that human melanomas and thus that they belong to the types of human tumors with the most frequent expression of cancer testis antigens. Similarly, it is unlikely that the blood-brain barrier is a major reason for the low frequency of antibody responses in glioma patients, because (even though the number of sera from astrocytoma patients analyzed in our study is limited), the results shown in Table I suggest that antibody responses appear to be more frequent in low-grade astrocytoma patients than in patients with high-grade astrocytomas, in whom the blood-brain barrier would be expected to be less stringent than in the low-grade patients. Instead, the impaired antibody response may be due to the known immunosuppressive effects of human astrocytomas, which are mainly operative at the tumor site. Although gliomas express MHC-I molecules regularly, the production of transforming growth factor-β and prostaglandin E2 probably impairs T-cell activation.

Whereas the expression of granulocyte and granulocyte/macrophage colony-stimulating factors has been shown in human gliomas,20 expression of proinflammatory genes such as B7-2 and interleukin-12 is lacking in these tumors.21 Finally, a factor(s) of a minimum weight of 49 kDa has recently been identified in the supernatant of glioblastoma cell cultures that not only suppresses T-cell responses but also alters the cytokine profiles of monocyte-derived antigen-presenting cells,22 which, in turn, might also inhibit T-cell help for B cells which is needed for the production of antibodies against antigens expressed by gliomas. Thus, in addition to the development of specific peptide or whole-protein vaccines, immunotherapeutic approaches in astrocytomas must design strategies to overcome the immunosuppressive effects of the malignant cells at the tumor site that impair both immune recognition and effective effector mechanisms of the immune system against these tumors.


  1. Top of page
  2. Abstract
  6. Acknowledgements

We thank U. Sahin helpful discussions and for providing some of the cDNAs.


  1. Top of page
  2. Abstract
  6. Acknowledgements