Immunoglobulins (Igs) are found thus far only to be produced by differentiated B lymphocytes. By immunohistochemistry analysis, in situ hybridization, and laser capture microdissection-assisted single-cell PCR, we demonstrate that human cancers of epithelial origin, including carcinomas of breast, colon, liver, lung, established epithelial cancer lines, as well as some normal lung tissues, also produce IgG in both cytoplasmic and secreted forms. Furthermore, blockade of tumor-derived IgG by either antisense DNA or antihuman IgG antibody increased programmed cell death and inhibited growth of cancer cells in vitro. More importantly, administration of antihuman IgG antibody also suppressed the growth of an IgG-secreting carcinoma line in immunodeficient nude mice. Our results support a role of tumor-derived IgG as growth factor for epithelial cancers. Prevalent expression of IgG in human carcinomas and its growth-promoting functions may have important implications in growth regulation and targeted therapy of human cancers.

Elevated level of serum IgG, IgA, or IgM antibodies is frequently observed in patients with cancers of epithelial origin, including carcinomas of breast, colon, and liver (1, 2). These tumor-reactive Igs3 have been interpreted as humoral responses of host to cancer growth (3). In fact, by Serex method, it has been firmly established that these tumor-reactive antibodies are capable of binding to normal and tumor-associated antigens, including those of cell surface and intracellular proteins (3, 4, 5). However, monoclonal gammopathy was observed not only in patients with neoplastic proliferation of B cell and/or plasma cells but also in those with malignant tumor of epithelial origin (6, 7, 8, 9). These observations suggest the possibility that the Ig is produced by the epithelial cancer cells. In this article, we present unexpected findings that IgG with unidentified specificity could be secreted directly from cancer cells, and these antibodies are involved in their survival and growth.

Tissue Samples.

Formalin-fixed, paraffin-embedded biopsy specimens of tumor and corresponding normal tissues were collected from Department of Pathology, Norman Bethune University of Medical Sciences and Department of Pathology, Taiping Road Hospital. Fresh biopsy samples were obtained from Department of Pathology, Beijing Railway Hospital. All procedures were performed with the approval of the Institutional Review Board and proper consent from patients (if any). For immunohistochemical staining, the biopsy tissues had been fixed immediately in 10% buffered formalin and, 24 h later, dehydrated in increasing concentrations of alcohol and embedded in paraffin. Sections of 5 μm were made. Fresh biopsy samples were used for the isolation of tumor cells.

Cell Lines.

Human breast cancer cell line (MCF-7) and colon cancer cell lines (HT-29, LOVO) were purchased from Beijing Institute for Tumor Prevention and Treatment; human cervical cancer cell line (HeLa MR) was a gift of Dr. Shouping Ji (Academy of Military Medical Science); cervical cancer cell line (HeLa S3), lung cancer cell line (A549), ovarian cancer cell line (CaOV3), and B lymphocytic leukemia cell lines (Raji and Daudi) were supplied by Beijing University Center for Human Disease Genomics. The MCF-7, HT-29, LOVO, HeLa MR, HeLa S3, A549, and CaOV3 were cultured in DMEM with 10% FCS, whereas Raji and Daudi cell lines were cultured in RPMI 1640 10% FCS at 37°C in a humidified atmosphere of 95% O2 and 5% CO2.

Immunohistochemistry.

The tissue sections were dewaxed, rehydrated through graded ethanol washes, placed in 10 mm citrate buffer (pH 6.0), and then heated twice in microwave oven for 5 min each. The slides were incubated with 0.3% hydrogen peroxide for 5 min, washed with PBS, and blocked in PBS plus 10% normal goat serum for 10 min. After excess blocking buffer was removed, the direct immunohistochemical staining was performed with goat antihuman IgG-HRP to detect IgG (1:100; Dako, Carpinteria, CA). Slides were incubated for 45 min with antibody at 37°C in a humidified chamber. After thorough washing, the bound antibody was detected using DAB; control sections were stained with goat antimouse IgG-HRP (1:100; Dako). Indirect immunohistochemical staining was performed to detect CD20 using mouse antihuman CD20 mAb (CD20cy, clone: L26, 1:100; Dako). To do so, the mAb was added to slides and incubated for 45 min with antibody at 37°C, after washing three times with PBS, the slides were incubated with goat antimouse IgG-HRP for 40 min at 37°C; after extensive washing, the bound antibody was detected using DAB.

Purification of Tumor Cells.

Tumor tissue suspension was prepared from fresh biopsy cancer tissues of the breast, colon, liver, and lung. After passing through a 90-mesh net and washing twice with Hanks solution, lymphocytes in the cell suspension were removed by gradient centrifugation in lymphocyte separation medium (Shanghai Chemical Reagent, Inc., Shanghai, China). The separated cancer cells were washed and resuspended in Hanks solution. To detect IgG in the isolated cancer cells, goat antihuman IgG-HRP (1:100) was used in direct immunocytochemical staining.

Extraction and Purification of Ig Molecules.

Cancer cells were treated with lysis buffer [50 mm Tris-HCl (pH 8.0), 1 mm EDTA (pH7.5), 25% sucrose, 0.8 m 2-mercaptoethanol, and 1 mm phenylmethylsulfonyl fluoride] on ice for 20 min. The lysate was centrifuged for 25 min at 15,000 rpm/min at 4°C. The pellet was discarded, and the supernatant containing the cytosol protein was added to CM-Sepharose (Pharmacia, Peapack, NJ), and the bound IgG was eluted by different concentrations (0.5, 0.8, 1.0, and 1.5 m) of NaCl. Fractions collected were analyzed by SDS-PAGE, Western blot, or capillary electrophoresis. For simplicity, sometimes the Ig in the cytosol protein was not separated by CM-Sepharose but directly analyzed by SDS-PAGE and Western blot after protein concentration was determined according to A280 nm.

SDS-PAGE and Western Blot.

The cytosol protein (crude protein at 40 μg/well; purified protein at 10 μg/well) was analyzed in 10% SDS-PAGE (under reducing and nonreducing conditions) and stained with Coomassie Brilliant Blue. For Western blot analysis, after electrophoresis, the separated proteins on the gel were transferred to polyvinylidene difluoride membrane. In direct staining, goat antihuman IgG-HRP (1:1000) was used. In indirect staining, mAbs against IgG heavy chain (1:1000; Sigma, St. Louis, MO), κ chain, and λ chain (Dako, Glostrup, Denmark; 1:1000) were used as primary antibodies, and goat antimouse IgG-HRP (Dako; 1:1000) was used as secondary antibody.

In Situ Hybridization.

Paraformaldehyde-fixed and paraffin-embedded tissue sections (5 μm) were deparaffinized, dehydrated, and incubated in 0.2 m HCl for 20 min. The sections were treated with proteinase K, fixed with paraformaldehyde, prehybridized at 42°C for 4 h, and hybridized overnight at 46°C with human IgG Fc cRNA as probe. After hybridization, the sections were washed in 2× SSC and 0.2× SSC (65°C) and then treated with RNase. The samples were incubated with antidigoxigenin antibody conjugated with alkaline phosphatase (dilution 1:500; Roche Diagnostics, Rotkreuz, Switzerland). 5-Bromo-4-chloro-3-indolyl phosphate and nitro-blue-tetrazolium (Sigma) were used for color reaction. In control experiments, the slides were incubated with RNase or with corresponding sense probes.

Northern Blot Analysis.

Total RNA was extracted from normal lung and lung carcinoma tissues using Trizol Reagent (Life Technologies, Inc.) followed by electrophoresis. The RNA was transferred onto nylon membranes and cross-linked by UV irradiation. The membranes were prehybridized using hybridization solution (Life Technologies, Inc.) for 6 h at 42°C and hybridized with 32P-labeled IgG Fc cDNA probe for 16 h at 68°C. After extensive washing, positive signals were displayed on a phosphor screen.

Single-Cell RT-PCR.

Frozen lung cancer samples were sectioned at 8 μm and mounted on slides, and slides containing frozen sections were quickly fixed in 70% ethanol for 5 min, stained with hematoxylin for 1 min, counterstained with eosin for 2 min, and then quickly rinsed in diethyl pyrocarbonate-treated water. Slides were dried in a fume hood for 5 min and then was used to LCM using the PALM Microlaser Systems (PALM Microlaser Technologies). The total RNA microdissected cancer cells was isolated following the standard RNeasy Mini Kit (Qiagen, Hilden, Germany) protocol.

Reverse transcription was carried out with Sensiscript RT Kit (Qiagen), following the standard protocol, with oligo(dT) primer. After reverse transcription, 1 μl of each reverse transcription reaction was used for Touchdown PCR with LA Taq Polymerase (TaKaRa Bio, Inc., Siga, Japan). The specific primers for Ig gene of variable region and CH1 were used to RT-PCR of IgG heavy chain (included V, D, and J region and CH1). Two Ig genes of variable region primers were: V3, 5′-GAGGTGCAGCTCGAGCAGTCAGG-3′ or V4f, 5′-CAGGTGCAGCTGCTCGAGTCGGG-3′; and IgG1 CH1 region primer, 5′-ACACCGTCACCGGTTCGG-3′. Touchdown PCR was performed for amplification. PCR product was also cloned into pGEM-T Easy Vector (Promega) and subjected to sequencing by ABI 3100 Genetic Analyzer (Applied Biosystem).

ELISA.

IgG in the culture supernatants of cancer lines was detected by double layer ELISA. A 96-well plate was coated with goat antihuman IgG (1:2000; Dako) in 50 μl PBS/well for 1 h at 37°C. The wells, after three successive washings with PBS, were incubated with 0.25% BSA and 0.05% Tween 20 in PBS for 30min. Cancer cell culture supernatants serially diluted (from 1:10) with PBS containing 0.25% BSA and 0.05% Tween 20 were added to the wells and incubated for 1 h at 37°C. After washing, 0.25% BSA and 0.05% Tween 20 in PBS were again added to each well and incubated for 10 min at 37°C. Rabbit antihuman IgG-HRP (1:5000) was added and the plate incubated for 1 h at 37°C. After final washing, the plate was developed with o-phenylenediamine in a buffer containing 0.012% hydrogen peroxide, 0.1 m citric acid and 0.1 m Na2HPO4.

FACS.

For detecting IgG in cancer cells, incubated cancer cell lines were harvested and washed with PBS, the cells were fixed in 70% ethanol for 24 h, washed with PBS and blocking with 10% FCS for 30min, cells were stained with rabbit antihuman IgG-FITC (Dako) for 30 min at 4°C, the cells after another washing with PBS, were analyzed by flow cytometry. To detect CD19 on cancer cells, cancer cell lines (A549, HT-29 and HeLaS3) and Raji cells were harvested and washed with PBS containing 1% FCS, they were stained with mAb against human CD19- PE (Becton Dickinson, Chicago) for 30min at 4°C, after washing with PBS, the cells were analyzed by flow cytometry.

RT-PCR.

Total RNA was extracted using Trizol reagent (Life Technologies, Inc.) and reverse transcription of total RNA was performed using oligo(dT) primer and SuperscriptII-RT (Life Technologies, Inc.) following manufacturer’s instruction. PCR primers used for amplifying the CDR3 region of the Ig heavy chain are shown as follows: the first PCR amplification with upstream primers V3 or V4f or V6, 5′-CAGGTACAGCTCGAGCAGTCAGG-3′, and the downstream primer LJH, 5′-TGAGGAGACGGTGACC-3′, was performed for 35 cycles under the following conditions: denaturing at 94°C for 1 min; annealing at 55°C for 1 min; and polymerization at 72°C for 1 min. Nest PCR amplification with primers FR3, 5′-ACACGGCYSTGTATTACTGT-3′, and VLJH, 5′-GTGACCAGGGTNCCTTGGCCCCAAG-3′, was performed for 30 cycles under the following conditions: denaturing at 94°C for 1 min; annealing at 57°C for 1 min; and polymerization at 72°C for 1 min. The PCR product, predicted as 60–100 bp in size, was separated on 2.5% agarose gel by electrophoresis. The identification of the PCR product was confirmed by DNA sequencing.

For amplifying RAG-1 and RAG-2, RQ1 RNase-free DNase (Promega, Madison, WI) was used to treat the RNA samples to eliminate any contaminating genomic DNA. Nest-PCR was performed as described previously (10).

Effect of ASODN.

ASODN 5′-ACACGGCTGTGTATT-3′ and RODN 5′-TGAGACTACCCGTCA-3′ were separately introduced into cultured cancer cells (10 μm ODN/106) by Electro square porator ECM830. The introduced cells were then incubated in DMEM plus 5% FCS at 37°C for 18, 28, and 42 h. IgG expression was determined using rabbit antihuman IgG-FITC. Cell apoptosis was assessed after annexin V and PI staining by flow cytometry.

In Vitro Effect of Anti-IgG Antibody.

Cancer cells were seeded in a 24-well plate at 6 × 104 cells/ml, incubated overnight at 37°C in DMEM with 2.5% FCS. The supernatant was discarded, and goat antihuman IgG antibody (25 μg/ml; Sigma) was added to the culture for 48 h additional incubation. Apoptosis of cancer cells was assessed as described above. Normal goat IgG (25 μg/ml) was used as control. In cell proliferation assay, cancer cells were seeded in 96-well plate at 1 × 103 cells/ml and incubated overnight in DMEM with 2.5% FCS. The supernatant was discarded, and the cancer cells were subsequently treated with goat antihuman IgG antibody (10 μg/ml; Sigma). Cell proliferation was determined daily by MTT incorporation.

Animal Studies.

Cultured HeLa MR cells (5 × 106) were s.c. inoculated into nude mice (BALB/c). When the tumor became palpable, polyclonal antibody of goat antihuman IgG [purified by protein A-Sepharose 4B column, a gift from PLA Logistic University (Changchun, China), with an antibody titer of 1:125 by double diffusion test] was injected 1.5 mg/tumor intratumorally for five times with a 3-day interval. The growth of the tumor was monitored every day. On day 21 after tumor cell inoculation, the tumors were excised and histologically examined. Nude mice with HeLa MR xenografts treated with normal goat IgG (PLA Logistic University, also purified by protein A-Sepharose 4B column), and PBS was used as controls.

IgG Expression in Human Cancer Tissues.

We first used human IgG-specific monoclonal or polyclonal antibodies in immunohistochemistry to determine the expression of IgG in malignant epithelial tumors. In all evaluated cancer tissues, including those from breast cancer (n = 10), liver cancer (n = 14), colon cancer (n = 6), and lung cancer (n = 12), positive staining was detected in the cytoplasm of all cancer tissues. In contrast, IgG was not detected in similar numbers of normal tissues of breast, colon, liver, and lung, except 4 of 10 normal lung tissues, some cells in the basal layer of epithelium and cells close to cancer tissues. In addition, immunoreactivity did not correlate with the amount of B lymphocyte (CD20-positive cells) and plasma cell found in cancer tissues; furthermore, tumor-infiltrating plasma cells were generally negative or weakly positive reaction by antihuman IgG staining (Table.1, Fig. 1 A).

To minimize contamination of infiltrating B cells in cancer tissues, we isolated tumor cells by gradient centrifugation from fresh biopsy tissues of breast cancer (n = 4), colon cancer (n = 4), liver cancer (n = 4), and lung cancer (n = 2). After removal of lymphocytes (including plasma cells) by gradient centrifugation, IgG in the separated breast, liver, and colon cancer cells were detected by immunocytochemical staining (Fig. 1,B). IgG was distinctly detected in the cancer cells by Western blot (Fig. 1,C). Protein extract of enriched cancer cells from a case of lung cancer was further purified by ion exchange chromatography on CM-Sepharose, and IgG in these tumor cells was confirmed in Western blot analysis using specific antibodies (Fig. 1 D) Finally, NH2-terminal sequence analysis of the purified protein detected the 14 amino acid residues that were identical to those of the published IgG sequences (N-VQLVESGGGLVHPG).

We also examined the expression of Ig mRNA by in situ hybridization using the cRNA of IgG Fc segment as a probe. Human IgG gene transcripts were demonstrated in all 6 breast cancer samples (Fig. 1,E), all 10 lung cancer samples, and a part of adjacent normal epithelial cells of the breast and lung. Furthermore, we also demonstrated IgG expression by Northern blot analysis using cDNA of IgG Fc segment in 5 lung cancer tissues and in 5 adjacent normal lung tissues. Finally, Ig γ chain mRNA was also detected in all 5 lung cancer tissues and 3 of 5 adjacent lung tissues (Fig. 1,F). Cancer cells were isolated by LCM technique from the same lung cancer sample (the protein purified from this sample was Ig by NH2-terminal sequence analysis as described above), and the IgG heavy chain (included V, D, J, and CH1 region) was amplified by single-cell RT-PCR. This result confirmed that Ig gene recombination, a hallmark of functional Ig, exists in the cancer cells. Moreover, three clonal sequences sequenced were completely identical (Fig. 1 H), suggesting limited diversity of cancer-associated IgG.

IgG Expression in Long-Term Cultured Human Cancer Lines.

To entirely eliminate contamination of B cells in cancer tissues, we examined the expression of IgG also in long-term cultured human cancer lines of epithelial origin. FACS analysis demonstrated IgG expression in breast cancer (MCF-7), colon cancer (HT-29, LOVO), liver cancer (BCL-7402), ovarian cancer (CaOV3), lung cancer (A549), cervical cancer (HeLaS3 and HeLa MR), and two B lymphoma cell lines (Raji and Daudi). IgG immunoreactivity located predominantly in cytoplasm (30∼90%) than in plasma membrane (15∼20%; Fig. 2,A). We have used antihuman CD19 (pan B cells) mAb to distinguish cancer cell lines from B-cell lines by flow cytometry. CD19 was found on B-cell line (Raji) but not on cancer cell line A549 (Fig. 2,B), HT-29, and HeLaS3 (data not shown). The culture supernatant of two cervical cancer cell lines was also positive for IgG as determined by ELISA, suggesting that these cancer lines secrete human IgG (Fig. 2,C). Consistent with this finding, Western blot analysis of lysate of cancer cell lines demonstrated immune staining at Mr 150,000, Mr 50,000, and Mr 25,000, which correspond to the molecular sizes for IgG, in several cancer cell lines using monoclonal antibodies against human Ig γ chain and κ chain (Fig. 2 D).

V-D-J recombination of Ig heavy chain was analyzed in these epithelial tumor lines. The third CDR (CDR3) of IgG heavy chain gene was amplified by RT-PCR using a set of primers. The CDR3 segments, which include the V-D-J sequence, were detected in HeLa S3, HeLa MR, A549 (GenBank No. AY035373) and HT-29 (GenBank No. AY035374; Fig. 2 E). The results demonstrated monoclonality of the V-D-J recombination in HeLa S3 and HT-29, although occasionally it contained single nucleotide deletion, insertion or mutation, whereas the Ig in HeLa MR and A549 did not. Interestingly, HT-29 line had identical sequence as HeLa MR. The CDR3 in HeLa S3 (GenBank No. AY05372) was also amplified, although the primer sequences used for FR3 and VLJH were not completely complementary.

The RAG-1 and RAG-2 are required for the formation of VDJ recombination. We next examined whether or not human cancers actively expressed these genes. Interestingly, RAG-1 and RAG-2 gene transcription in these cancer cell lines were detected by RT-PCR (Fig. 2 F), using the primers as described previously (10). Moreover, both the RAG-1 and RAG-2 segments from cancer cells were identical to the published sequences from B lymphocytes. Our results thus suggest that human epithelial cancer cells have ability to process VDJ recombination.

ASODN and Antihuman Ig Increases Apoptosis and Inhibits Growth of Cancer Cells.

To determine the functions of cancer-derived Ig, we first introduced a 15-mer S-oligos of ASODN into A549, HT-29, HeLa MR, and HeLa S3 cells to block the expression of human IgG. RODN was used for comparison. The ASODN was complementary to FR3 of the IgV heavy chain derived from A549, HT-29, and HeLa MR but not completely complementary to HeLa S3s. FACS analysis showed significant inhibition of IgG expression in A549 (inhibition rate: 50–70%), HT-29 (inhibition rate: 70–80%), and HeLa MR (inhibition rate: 30–50%) by ASODN. The inhibition of IgG expression induced by RODN in A549 and HT-29 was likely attributable to nonspecific effect, but it was significantly weaker than ASODN. As what we envisaged, IgG expression was not significantly affected in HeLa S3 by either ODN (Fig. 3 A). Our results support that ASODN is selective for the inhibition of Ig gene expression in tumor cells.

We next examined the effect of ASODN on survival of tumor cells. After ASODN was introduced into A549, HT-29, and HeLa MR cells accompanied with an inhibition of IgG expression, apoptosis of tumor cells increased significantly (Fig. 3,B). In contrast, no apoptosis occurred in HeLaS3 cells. Our results suggest that tumor IgG expression is required, at least, in part, for survival of tumor cells. The effects of tumor IgG on the growth and survival of tumor cells were additionally confirmed by a goat antibody specific for human IgG. Inclusion of this antibody-induced cell apoptosis and growth inhibition on A549, HT-29, and HeLa MR cells, whereas goat-derived normal IgG and PBS had not effect (Fig. 3,C). Accordingly, the antihuman IgG antibody significantly inhibited proliferation of these cancer cell lines in a MTT assay (Fig. 3 D). Goat-derived normal IgG or PBS did not induce cancer cell death and growth inhibition.

The effect of antihuman IgG in the growth of HeLa MR tumor in nude mice was evaluated. Inoculation s.c. of HeLa MR led to progressive growth of tumor. Tumor-bearing mice on day 7 were treated intratumorally on a 3-day interval using antihuman IgG. Tumor growth was significantly retarded as compared with those treated with normal goat IgG and PBS (Fig. 4,A), Histological examination demonstrated extensive necrosis and apoptosis in the mice treated with antihuman IgG but not control goat IgG or PBS (Fig. 4 B). Some of the anti-IgG antibody-treated tumors became liquefied (data not shown). Our results thus indicate that blockade of tumor-derived IgG inhibits growth of tumor and implicates a new method for cancer immunotherapy.

In this article, we describe an unexpected finding that human cancers of epithelial origin produce IgG. Moreover, antisense oligos, which block the expression of IgG, also inhibit the growth of tumor cells in vitro. Administration of anti-IgG antibody induced regression of xenografted human tumors that secrete IgG. Our results reveal a novel mechanism for growth of cancer cells and implicate a new target for the manipulation of progressive tumor growth.

Several methods were used to verify the expression of IgG from cancer cells. Initial immunohistochemical studies showed IgG immunoreactivity in many tissues derived from a variety of human cancer samples. In addition, tumor cells were isolated from cancer biopsy tissues, and additional analysis was performed. Western blot and NH2-terminal sequence analysis confirmed Ig molecular expression in the isolated cancer cells. Although it is possible that IgA and IgG are transferred into epithelial cells via poly-IgA/G receptor on plasma membrane (11), our further analysis by in situ hybridization and LCM-assisted single-cell RT-PCR essentially ruled out this possibility. Additional studies also detected IgG in many established tumor cell lines by FACS and Western blot analysis; furthermore, the FCS (5%) was used for culture of all cell lines and used in ELISA experiments to detect IgG in cancer culture supernatants. We did not observe significant increase in IgG level unless supernatants from cancer cell cultures were added. We also found that culture of tumor cells in serum-free medium (Chinese Hamster Ovary Cell Culture Media; Life Technologies, Inc.) up to 6 days did not eliminate the detection of IgG as shown by immunocytochemistry staining (12). Interestingly, we found that IgM was also expressed in these tumor cell lines, albeit in much lower level than IgG (data not shown). We have also found that Ig gene V-D-J rearrangement in conjunction with expression of RAG-1 and RAG-2 at mRNA level. Our results thus demonstrate the expression of IgG in human cancers of epithelial origin.

At this time, precise antigenic specificity of the cancer-derived Ig is unknown. High serum levels of autoantibodies were frequently found in patients with cancer (13, 14), and these autoantibodies often recognize intracellular proteins. There is yet no evidence that these autoantibodies are secreted by cancer cells. We found, however, that lung and breast cancer-derived IgG was capable of binding to smooth and striated muscles (data not shown). Moreover, the Ig V-D-J sequence of HT-29 was partially homologous with that of an autoantibody (GenBank no. Y17928). As the sequence of HeLa MR was identical to that of HT-29, it suggests that IgG from the two different cancer cell lines may have autoantibodies reacting to an identical antigen.

Tumor-derived IgG was found to involve, at least, in part, in the survival and growth of epithelial tumor cells because antihuman IgG and ASODN of Ig-induced apoptosis and growth inhibition of cancer cells in vitro and in vivo. It is likely that the cancer-derived Ig stimulates cell growth in an autocrine/paracrine fashion. As a matter of fact, Ig expression was not limited in malignant epithelial cells. Normal proliferating but not resting epithelial cells were also shown to express Ig, as seen in hyperplastic mammary glands, proliferating hepatocytes in cirrhotic liver (data not shown) and epithelium close to cancer tissue. Our results suggest that tumor-derived IgG contributes in cancer initiation in the precancerous stage when the epithelial cells are actively proliferating. Development of methods aiming at selective blockade of tumor-derived IgG thus may constitute a new approach for cancer therapy and prevention.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by Key Fundamental Research Grant 39830410 from the National Natural Sciences Foundation, China.

3

The abbreviations used are: Ig, immunoglobulin; HPR, horseradish peroxidase; DAB, 3,3′-diaminobenzidine; mAb, monoclonal antibody; CDR3, immunoglobulin heavy chain complementary-determining region 3; RT-PCR, reverse transcription-PCR; LCM, laser capture microdissection; FACS, fluorescence-activated cell sorting; FR3, immunoglobulin heavy chain framework 3 region; RAG, recombinase-activating gene; ASODN, antisense oligodeoxynucleotide; RODN, random oligodeoxynucleotide; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide.

Fig. 1.

IgG expression in cancer cells and in proliferating epithelial cells. A, IgG expression in epithelial cells in normal and cancerous tissues from breast, liver, colon, lung, and tumor-infiltrating plasma cells were examined by immunohistochemistry using the antibody against human IgG; CD20 expression on tumor-infiltrating B lymphocytes were examined immunohistochemically using the antibody against human CD20. B, IgG expression in purified cancer cells from breast and liver were examined immunocytochemically using the antibody against human IgG. C, SDS-PAGE and Western blot analysis were performed, and the cancer-derived IgG was observed in purified breast, liver, and colon cancer cells. D, IgG from purified lung cancer cell was extracted and purified by ion exchange chromatography on CM-Sepharose and detected by Western blotting. E, in situ hybridization of a cRNA segment of IgG in breast and lung cancer. F, Northern blot analysis of IgG mRNA in 5 lung cancer tissue samples and in 5 normal tissue samples adjacent to the tumor with a cDNA segment of IgG Fc as the probe. C, lung cancer tissue; N, normal lung tissue; P, peripheral blood mononuclear cells. G, single-cell RT-PCR analysis, after the frozen lung cancer sample slice was stained with H&E, single cancer cells were captured. H, the PCR product from single cancer cells was cloned into pGEM-T Easy Vector (Promega), and three clones were sequenced.

Fig. 1.

IgG expression in cancer cells and in proliferating epithelial cells. A, IgG expression in epithelial cells in normal and cancerous tissues from breast, liver, colon, lung, and tumor-infiltrating plasma cells were examined by immunohistochemistry using the antibody against human IgG; CD20 expression on tumor-infiltrating B lymphocytes were examined immunohistochemically using the antibody against human CD20. B, IgG expression in purified cancer cells from breast and liver were examined immunocytochemically using the antibody against human IgG. C, SDS-PAGE and Western blot analysis were performed, and the cancer-derived IgG was observed in purified breast, liver, and colon cancer cells. D, IgG from purified lung cancer cell was extracted and purified by ion exchange chromatography on CM-Sepharose and detected by Western blotting. E, in situ hybridization of a cRNA segment of IgG in breast and lung cancer. F, Northern blot analysis of IgG mRNA in 5 lung cancer tissue samples and in 5 normal tissue samples adjacent to the tumor with a cDNA segment of IgG Fc as the probe. C, lung cancer tissue; N, normal lung tissue; P, peripheral blood mononuclear cells. G, single-cell RT-PCR analysis, after the frozen lung cancer sample slice was stained with H&E, single cancer cells were captured. H, the PCR product from single cancer cells was cloned into pGEM-T Easy Vector (Promega), and three clones were sequenced.

Close modal
Fig. 2.

Ig immunoreactivity and Ig gene transcription in human cancer cell lines. A, IgG was localized in the cytoplasm or on the plasma membrane of cultured cancer cell lines by FACS analysis using an antibody to human IgG conjugated to FITC, and the number of IgG immunoreactive cells were measured. B, CD19 on cancer cell lines (A549, HT-29, and HeLaS3) and Raji cells was stained with mAb against human CD19-phycoerythrin. C, quantification of IgG level of culture supernatant of two HeLa cell lines was analyzed by ELISA using goat antihuman IgG-HRP; results are expressed as the value of A490 nm. D, forty μg of cell extract from HeLa MR and HT-29 were separated by 10% reducing SDS-PAGE or nonreducing SDS-PAGE, and proteins were transferred to polyvinylidene difluoride membrane; cancer-derived IgG was detected by Western blot with mAb against human IgG. E, the V-D-J rearrangement sequence and transcription of the Ig heavy chain were amplified by RT-nest-PCR. The primers for the first PCR amplification were segments of 5′-end of Ig genes (top) and complementary to 3′-end of J region (bottom); the primers for nest-PCR amplification were segments of FR3 and J region. The identification of the PCR product was confirmed by 2.5% agarose gel electrophoresis and DNA sequencing. F, RAG-1 and RAG-2 mRNA were also found in cancer cell lines by RT-nest-PCR. The fact that the amplified segments of RAG-1 and RAG-2 are on one exon of RAG-1 and RAG-2, respectively, excluded the effect of genomic DNA. The total RNA was treated with DNase before amplification. As a negative control, treated total RNA was also used as the template. 1R, A549 RNA as template; 1C, A549 cDNA as template; 2R, HT-29 RNA as template; 2C, HT-29 cDNA as template; 3R, HeLa MR RNA as template; 3C, HeLa MR cDNA as template.

Fig. 2.

Ig immunoreactivity and Ig gene transcription in human cancer cell lines. A, IgG was localized in the cytoplasm or on the plasma membrane of cultured cancer cell lines by FACS analysis using an antibody to human IgG conjugated to FITC, and the number of IgG immunoreactive cells were measured. B, CD19 on cancer cell lines (A549, HT-29, and HeLaS3) and Raji cells was stained with mAb against human CD19-phycoerythrin. C, quantification of IgG level of culture supernatant of two HeLa cell lines was analyzed by ELISA using goat antihuman IgG-HRP; results are expressed as the value of A490 nm. D, forty μg of cell extract from HeLa MR and HT-29 were separated by 10% reducing SDS-PAGE or nonreducing SDS-PAGE, and proteins were transferred to polyvinylidene difluoride membrane; cancer-derived IgG was detected by Western blot with mAb against human IgG. E, the V-D-J rearrangement sequence and transcription of the Ig heavy chain were amplified by RT-nest-PCR. The primers for the first PCR amplification were segments of 5′-end of Ig genes (top) and complementary to 3′-end of J region (bottom); the primers for nest-PCR amplification were segments of FR3 and J region. The identification of the PCR product was confirmed by 2.5% agarose gel electrophoresis and DNA sequencing. F, RAG-1 and RAG-2 mRNA were also found in cancer cell lines by RT-nest-PCR. The fact that the amplified segments of RAG-1 and RAG-2 are on one exon of RAG-1 and RAG-2, respectively, excluded the effect of genomic DNA. The total RNA was treated with DNase before amplification. As a negative control, treated total RNA was also used as the template. 1R, A549 RNA as template; 1C, A549 cDNA as template; 2R, HT-29 RNA as template; 2C, HT-29 cDNA as template; 3R, HeLa MR RNA as template; 3C, HeLa MR cDNA as template.

Close modal
Fig. 3.

Effect of antisense and antihuman IgG on cancer cell survival and proliferation in vitro. A, cancer cells (2 × 105) seeded in 6-well culture plate were separately treated with ASODN, RODN, or vehicle for 42 h. Cancer cells were harvested and fixed with ice-cold 70% ethanol and washed with PBS. They were immunostained with antihuman IgG-FITC antibody, and the number of positively stained cells were analyzed by FACS. B, A549, HT-29, and HeLa MR cells were treated with ASODN, RODN, or vehicle for 20, 28, and 42 h, respectively. After double staining with annexin V and propidium iodide, cell apoptosis was assessed by FACS. C, after treatment for 48 h with antihuman IgG, goat IgG, or PBS, cancer cells were harvested and fixed with ice-cold 70% ethanol and washed with PBS. Cell apoptosis was assessed by FACS as in B. D, cancer cells (1 × 103) seeded in 96-well culture plate were treated with goat antihuman IgG, goat IgG, or PBS. Cell proliferation was determined by MTT assay.

Fig. 3.

Effect of antisense and antihuman IgG on cancer cell survival and proliferation in vitro. A, cancer cells (2 × 105) seeded in 6-well culture plate were separately treated with ASODN, RODN, or vehicle for 42 h. Cancer cells were harvested and fixed with ice-cold 70% ethanol and washed with PBS. They were immunostained with antihuman IgG-FITC antibody, and the number of positively stained cells were analyzed by FACS. B, A549, HT-29, and HeLa MR cells were treated with ASODN, RODN, or vehicle for 20, 28, and 42 h, respectively. After double staining with annexin V and propidium iodide, cell apoptosis was assessed by FACS. C, after treatment for 48 h with antihuman IgG, goat IgG, or PBS, cancer cells were harvested and fixed with ice-cold 70% ethanol and washed with PBS. Cell apoptosis was assessed by FACS as in B. D, cancer cells (1 × 103) seeded in 96-well culture plate were treated with goat antihuman IgG, goat IgG, or PBS. Cell proliferation was determined by MTT assay.

Close modal
Fig. 4.

Growth inhibition of HeLa MR in nude mice treated with antihuman IgG. A, after tumor formation, antihuman IgG was injected intratumorally once every 3 days in 13 mice. Injection of normal goat IgG (n = 8) and PBS (n = 13) served as controls. B, on day 21 after treatment, the mice were sacrificed, and the tumors observed histologically (H&E staining).

Fig. 4.

Growth inhibition of HeLa MR in nude mice treated with antihuman IgG. A, after tumor formation, antihuman IgG was injected intratumorally once every 3 days in 13 mice. Injection of normal goat IgG (n = 8) and PBS (n = 13) served as controls. B, on day 21 after treatment, the mice were sacrificed, and the tumors observed histologically (H&E staining).

Close modal
Table 1

IgG expression in human cancers and normal tissues

CasePositive/total (%)aPositive degreeb
Malignant epithelial tumor Breast cancer 10/10 (100%) ++∼+++ 
 Liver cancer 14/14 (100%) ++∼+++ 
 Colon cancer 6/6 (100%) +∼++ 
 Lung cancer 8/12 (66%) ++ 
 Total 38/42 (90%) +∼+++ 
Normal epithelial tissue Normal breast tissue 0/10 (0) − 
 Normal liver tissue 0/14 (0) − 
 Normal colon tissue 0/6 (0) − 
 Normal lung tissue 4/10 (40%) 
 Total 4/40 (10%) 
CasePositive/total (%)aPositive degreeb
Malignant epithelial tumor Breast cancer 10/10 (100%) ++∼+++ 
 Liver cancer 14/14 (100%) ++∼+++ 
 Colon cancer 6/6 (100%) +∼++ 
 Lung cancer 8/12 (66%) ++ 
 Total 38/42 (90%) +∼+++ 
Normal epithelial tissue Normal breast tissue 0/10 (0) − 
 Normal liver tissue 0/14 (0) − 
 Normal colon tissue 0/6 (0) − 
 Normal lung tissue 4/10 (40%) 
 Total 4/40 (10%) 
a

Percentage of IgG positive cancer tissues and normal tissues.

b

Degree of positive IgG expression on tumor tissues and normal tissues is determined by immunohistochemical staining. −, positive cell under 10%; +, positive cells 30–50%; ++, positive cells 50–70%; +++, positive cells 70–100%.

We thank Drs. Dalong Ma (Peking University Center for Human Disease Genomics), Guizhen Yang (Norman Bethune University of Medical Sciences), and Sigano Akao (Institute of Microbiological Disease, Osaka University); Xinsheng Ye (The Natural Science Foundation of China) and Ning Mao (Institute of Basic Medical Science, Academy of Military Medical Science) for their comments and discussion; Dr. Michael M. Woolfson of University of York (United Kingdom) for his comments and suggestions; Drs. Baoyu Li (Department of Pathology, Beijing Railway Hospital); Xueyuan Xiao (Department of Biology, Beijing Normal University) and Jing Ma (Chinese PLA General Hospital) for their generous technical support; and the National Natural Science Foundation of China for financial support.

1
Olubuyide I. O., Salimonu L. S., Adeniran S. O. Soluble immune complexes and immunoglobulin (IgG, IgA and IgM) levels in Nigerians with primary liver cell carcinoma.
Afr. J. Med. Sci.
,
22
:
57
-62,  
1993
.
2
Manjula S., Aroor A. R., Raja A., Rao S. N., Rao A. Serum immunoglobulins in brain tumors.
Acta Neurochir.
,
115(3–4)
:
103
-105,  
1992
.
3
Taylor D. D., Gercel-Taylor C. Tumor-reactive immunoglobulins in ovarian cancer: diagnostic and therapeutic significance?.
Oncol. Rep.
,
5
:
1519
-1524,  
1998
.
4
Güre A. O., Altorki N. K., Stockert E., Scanlan M. J., Old L. J., Chen Y. T. Human lung cancer antigens recognized by autologous antibodies: definition of a novel cDNA derived from the tumor suppressor gene locus on chromosome 3p21.3.
Cancer Res.
,
58
:
1034
-1041,  
1998
.
5
Gure A. O., Stockert E., Scanlan M. J., Keresztes R. S., Jager D., Altorki N. K., Oid L. J., Chen Y. T. serological identification of embryonic neural proteins as highly immunogenic tumor antigens in small cell lung cancer.
Proc. Natl. Acad. Sci. USA
,
97
:
4198
-4203,  
2000
.
6
Tenenbaum N., Meignan S., Vincent J. P. Bi- and tri-clonal gammopathies and the elderly.
Ann. Med. Interne (Paris)
,
143
:
89
-93,  
1992
.
7
Maes B., Vanwalleghem J., Kuypers D., Van Damme B., Waer M., Vanrenterghem Y. IgA antiglomerular basement membrane disease associated with bronchial carcinoma and monoclonal gammopathy.
Am. J. Kidney Dis.
,
33
:
E3
1999
.
8
Hori H., Kihara Y., Nagashio Y., Hirohata Y., Abe S., Murata I., Otsuki M. Signet ring cell carcinoma of the colon with monoclonal gammopathy of undetermined significance.
Nippon Shokakibyo Gakkai Zasshi
,
97
:
1373
-1377,  
2000
.
9
Gregersen H., Mellemkjaer I., Ibsen J. S., Dahlerup J. F., Thomassen L., Sorensen H. T. The impact of M-component type and immunoglobulin concentration on the risk of malignant transformation in patients with monoclonal gammopathy of undetermined significance.
Haematologica
,
86
:
172
-179,  
2001
.
10
Lynch S., Kelleher D., McManus R., O’Farrelly C. RAG1 and RAG2 expression in human intestinal epithelium: evidence of extrathymic T cell differentiation.
Eur. J. Immunol.
,
25
:
1143
-1147,  
1995
.
11
Phillips J. O., Everson M. P., Moldoveanu Z., Lue C., Mestecky J. Synergistic effect of IL-4 and IFN-γ on the expression of polymeric Ig receptor (secretory component) and IgA binding by human epithelial cells.
J. Immunol.
,
145
:
1740
-1744,  
1990
.
12
Hou C., Qiu X., Sun J., Shen J., Tang P., Mao N. Detection of Ig-like protein in HeLa MR cells cultured in serum-free medium.
Bull. Acad. Mil. Med. Sci.
,
22
:
240
1998
.
13
Lorenz K. J., Susal C., Opelz G., Maier H. Relationship between progression of disease and immunoglobulin A-anti-Fab-/F(ab′)2 autoantibodies in patients with head and neck cancer.
Otolaryngol. Head Neck Surg.
,
118
:
130
-652,  
1998
.
14
Vernino S., Adamski J., Kryzer T. J., Fealey R. D., Lennon V. A. Neuronal nicotinic ACh receptor antibody in subacute autonomic neuropathy and cancer-related syndromes.
Neurology
,
50
:
1806
-1813,  
1998
.