Abstract
Laminin-5 (Ln-5), a heterotrimer composed of three different laminin chains [laminin-α3 (Ln-α3), laminin-β3 (Ln-β3), and laminin-γ2 (Ln-γ2)], is a major component of the basement membrane in most adult tissues. One of the chains, Ln-γ2, is a specific marker of invasive tumors because it is frequently expressed as a monomer in malignant tumors. However, there is no simple and direct method to detect the monomeric form of Ln-γ2 selectively in the presence of Ln-5 because all available antibodies recognize both monomeric and heterotrimeric forms of Ln-γ2. In this study, we developed a new monoclonal antibody (mAb) termed 1H3 that reacts specifically with human Ln-γ2 monomers during immunoprecipitation, ELISA, Western blotting, and immunostaining. Ln-5 was not recognized by mAb 1H3 after denaturation with detergents under nonreducing conditions, but reactivity was recovered when denaturation was done under reducing conditions. The epitope of the antibody was mapped to region on the coiled-coil structure formed between Ln-γ2 and its partner chains Ln-α3 and Ln-β3 in Ln-5, whose structure is further stabilized by disulfide bonds. In normal tissue samples, the basement membrane was stained with conventional antibody against Ln-γ2 but not by mAb 1H3. In contrast, tumor cells in tissue sections could be stained with mAb 1H3 as efficiently as with conventional antibody. Thus, mAb 1H3 holds promise as a powerful tracking tool for the specific detection of monomeric Ln-γ2 in vivo and in vitro and is potentially useful as a diagnostic tool for detecting tumors and as a vehicle for drug delivery to cancer tissues. [Cancer Res 2008;68(2):530–6]
Introduction
Laminins are assembled from three polypeptide chains designated α, β, and γ. There are several isoforms of each type of chain that associate in different combinations to constitute the burgeoning laminin protein family (1). For example, laminin-5 (Ln-5), the most abundant form of most of the basement membrane in adult tissues, contains laminin-α3 (Ln-α3), laminin-β3 (Ln-β3), and laminin-γ2 (Ln-γ2) chains as illustrated in Fig. 1A (2, 3). Among these polypeptide chains, Ln-γ2 has been found to be frequently expressed as a monomer in several types of malignant cancer cells based on experiments showing the lack of simultaneous expression of Ln-α3 and Ln-β3 chains (4, 5). Expression of monomeric Ln-γ2 is particularly predominant in the budding cells of tumor masses as observed in esophageal, stomach, colon, and cervical carcinomas, although it is not detectable in normal tissues (4, 6–9). These observations strongly emphasize the importance of monomeric Ln-γ2 as a specific marker for invasive carcinomas.
Detection of monomeric Ln-γ2 by sandwich ELISA using mAbs. A, schematic depiction of human Ln-5 and monomeric form of Ln-γ2 (mono). Left, Ln-5 composed of laminin α3, β3, and γ2 chains; right, monomeric Ln-γ2. Ln-γ2 has five domain structures (DI-DV). Arrows, frequently observed processing sites (ps) in Ln-γ2 (ps1 and ps2). B, specific reactivity of mAb 1H3 (top) and D4B5 (bottom) to monomeric Ln-γ2 and Ln-5 was analyzed by sandwich ELISA using 96-well plates. Each antibody was coated onto the 96-well plates and antigens were added to the well at the indicated doses. The captured antigens were measured as described in Materials and Methods.
Detection of monomeric Ln-γ2 by sandwich ELISA using mAbs. A, schematic depiction of human Ln-5 and monomeric form of Ln-γ2 (mono). Left, Ln-5 composed of laminin α3, β3, and γ2 chains; right, monomeric Ln-γ2. Ln-γ2 has five domain structures (DI-DV). Arrows, frequently observed processing sites (ps) in Ln-γ2 (ps1 and ps2). B, specific reactivity of mAb 1H3 (top) and D4B5 (bottom) to monomeric Ln-γ2 and Ln-5 was analyzed by sandwich ELISA using 96-well plates. Each antibody was coated onto the 96-well plates and antigens were added to the well at the indicated doses. The captured antigens were measured as described in Materials and Methods.
In addition, monomeric Ln-γ2 may also play an important role as a modulator of tumor cell behavior. Ln-γ2 has an epidermal growth factor (EGF)-like repeat in domain III (DIII) and this portion can be clipped out from Ln-γ2 by the action of matrix metalloproteinases (MMP), such as MMP-2 and membrane type 1-MMP (MT1-MMP; refs. 10, 11). The released DIII fragment can act as a ligand of the EGF receptor (EGFR) and elicit receptor-mediated intracellular signals (12). Interestingly, the monomeric form of Ln-γ2 shows 10-fold higher sensitivity to MT1-MMP than the γ2 chain in human Ln-5 (13). Thus, the monomeric form of Ln-γ2 expressed in human tumors may contribute to tumor cell behavior by activating EGFR. Indeed, aggressive melanoma Mum2B cells that express monomeric Ln-γ2 and MT1-MMP can form a tube-like network known as vascular mimicry when the cells are cultured in collagen gel. Knockdown of MT1-MMP expression by small interfering RNA or inhibition of tyrosine kinase activity of EGFR suppressed the tube formation (14). In spite of the importance of monomeric Ln-γ2 as a marker in tumors and as a tumor cell regulator, we do not have simple methods to detect it directly and sensitively because all the available antibodies against Ln-γ2 also react with Ln-5. Thus, the presence of monomeric Ln-γ2 can only be inferred by confirming the lack of Ln-α3 and Ln-β3.
In this study, we tried to develop a monoclonal antibody (mAb) that reacts selectively with the monomeric form of Ln-γ2. For this purpose, mice were immunized with purified monomeric human Ln-γ2 and antibodies produced by hybridoma clones were screened by ELISA using monomeric Ln-γ2 and Ln-5. Most of the hybridoma supernatants that recognized the monomer also recognized Ln-5, with the exception of one (mAb 1H3) that reacted preferentially with monomeric Ln-γ2. We evaluated this antibody as a practical tool to monitor monomeric Ln-γ2 in vitro and in vivo.
Materials and Methods
Preparation of human Ln-γ2 protein. Human Ln-5 and monomeric Ln-γ2 proteins were purified from culture medium of human gastric carcinoma STKM-1 and malignant melanoma Mum2B cells, respectively, as described previously (13).
Construction and expression of DIII, DIII/V, and DI/II. The cDNA of human Ln-γ2 was used as a template to generate DNA fragments encoding DI/II (nucleotides 1830–3580), DIII (nucleotides 1303–1737), and DIII/V (nucleotides 1–1820) by PCR. These PCR products were subcloned into the pcDNA3.1+ expression vector (Invitrogen), and their nucleotide sequences were confirmed. For transient expression of the encoded proteins, each expression plasmid (4 μg) was transfected into COS7 cells (100 mm dish) using Fugene 6 (Roche Diagnostics) and cells were cultured for 48 h in serum-free DMEM after 12 h of transfection. Conditioned medium was collected and concentrated by ammonium sulfate precipitation. Precipitates containing recombinant proteins were collected and dissolved in 20 mmol/L Tris-HCl (pH 7.5)/0.15 mol/L NaCl/0.005% Brij-35 buffer.
Immunization, hybridization, and hybridoma screening. BALB/c mice were injected i.p. with 50 μg of human monomeric Ln-γ2 protein emulsified in an equal volume of complete Freund's adjuvant (Difco Diagnostic Systems) and boosted twice at 2-week intervals with equal amount of monomeric Ln-γ2 protein with incomplete Freund's adjuvant (Difco Diagnostic Systems). Six weeks after the third immunization, the mice were injected i.v. with the same amount of Ln-γ2 protein without adjuvant. After 3 days of i.v. injection, the blood was collected and centrifuged at 3,000 rpm at room temperature and serum was collected as a positive control of antibody to monomeric Ln-γ2 protein during hybridoma screening.
Hybridization of the mouse myeloma line (P3U1) and spleen cells was performed by the methods at Kitayama Labes Co. Ltd. Hybridoma cells were cultured in hypoxanthine-aminopterin-thymidine selection medium for 10 days and the cells were reseeded into 96-well plates. Supernatants from each well having growing cells were examined for antibody production using purified monomeric Ln-γ2 and Ln-5 by a single ELISA assay. Limited dilution and cloning of cells were repeated twice and some hybridomas producing mAb reactive to monomeric Ln-γ2 were obtained.
Western blot analysis. Purified monomeric Ln-γ2 (1 μg) and Ln-5 (3 μg) proteins were separated on a 6% or 10% SDS-polyacrylamide gel and then transferred onto polyvinylidene difluoride (PVDF) membrane (0.45 μm; Millipore) for 1.5 h. The membranes were incubated in PBS and 0.05% Tween 20 containing 5% nonfat dry milk overnight at 4°C to block nonspecific binding and then probed with the antibodies. To detect the antibody bound to the antigen, the membranes were treated with secondary antibody conjugated to horseradish peroxidase (HRP). Peroxidase activity was detected using the enhanced chemiluminescence plus detection system (GE Healthcare; ref. 13).
Immunoprecipitation of Ln-γ2 with 1H3 antibody. Purified monomeric Ln-γ2 and Ln-5 proteins (1 μg each) were incubated overnight at 4°C with mAbs 1H3 and 2H2 in 20 mmol/L Tris-HCl (pH 7.5)/0.005% Briji-35/1% CHAPS (TBC) buffer (final volume, 200 μL). Suspended protein G-Sepharose (40 μL) was added and incubated at 4°C for 3 h with agitation. Protein G-Sepharose was washed thrice with TBC buffer and then centrifuged at 15,000 rpm to precipitate the antigen/antibody complexes. Finally, precipitated antigen-antibody complex was released from the protein G-Sepharose 4B (GE Healthcare) by addition of 2× SDS sample buffer/5% β-2-mercaptoethanol buffer (15).
Single and sandwich ELISA assays. Screening of antibodies produced by hybridoma cells was carried out by single ELISA. Ninety-six–well plates (Corning Costar) were directly coated with either purified monomeric Ln-γ2 (0.5 μg/mL) or Ln-5 (1 μg/mL) in PBS for 3 h at room temperature. The wells were blocked with 5% bovine serum albumin (BSA) for 1 h at room temperature. After washing with PBS thrice, serum-free culture media of hybridomas were added to the wells and allowed to react with each antigen for 1 h at room temperature. Then, each well was washed with PBS/0.05% Tween 20 thrice and bound antibodies were detected by using HRP-conjugated goat anti-mouse IgG (GE Healthcare). The absorbance of the tetramethyl benzidine–filled (Sigma) wells was determined at 450 nm with a Bio-Rad spectrophotometer.
For quantitative analysis, 96-well plates were coated with mAb D4B5 (1 μg/mL) or mAb 1H3 (10 μg/mL) in PBS overnight at 4°C. Preimmune mouse serum was also coated as a negative control. Blocking and washing of the wells were carried out similarly to the methods described above. Then, increasing doses of purified Ln-5 or monomeric Ln-γ2 (0–240 ng/mL) were added to the wells and allowed to react with each antibody for 1 h at room temperature. After washing, bound antigens were detected with polyclonal antibody (pAb) 2778 (0.5 μg/mL) and goat anti-mouse IgG conjugated to HRP. The absorbance (450 nm) of the tetramethyl benzidine–filled wells was measured with a Bio-Rad spectrophotometer.
Cells, cell culture, and preparation for cell lysate. Mum2B, a human melanoma cell line, was a gift from Professor Vito Quaranta (Vanderbilt University, Nashville, TN). STKM-1, a human gastric carcinoma cell line, was a gift from Dr. S. Yanoma (Kanagawa Cancer Center, Research Institute, Yokohama, Japan). Human gastric cancer TMK-1 cell line was a gift from Professor E. Tahara (Hiroshima University, Hiroshima, Japan). Human ovarian carcinoma SKOV-3 cell lines were purchased from the American Type Culture Collection, and human gastric carcinomas MKN7 and MKN28, skin carcinoma A431, breast carcinomas MDA-MB-231 and MCF-7, and lung carcinoma PC-8 cell lines were provided by Human Science Research Resources Bank. The cell lines were cultured at 37°C in a humidified atmosphere of 5% CO2/95% air. DMEM for Mum2B, MDA-MB-231, MCF-7, A431, and PC-8 cell lines and RPMI 1640 (Sigma) for SKOV-3, MKN7, MKN28, TMK-1, and STKM-1 cell lines were supplemented with 10 mmol/L HEPES, 1.2 mg/mL NaHCO3, and 2 mmol/L glutamate and used as basal medium and supplemented with 10% fetal bovine serum (Irvine Scientific). Cells were lysed with lysis buffer [20 mmol/L Tris-HCl (pH 7.5)/1% CHAPS/0.005% Brij-35/0.5 mol/L NaCl] and centrifuged at 15,000 rpm for 20 min at 4°C and the supernatant was collected as a clear lysate. The protein concentration was determined by a Bio-Rad Dye binding kit. BSA was used as the protein standard.
Human colon carcinoma tissues. Human colon adenocarcinoma and normal colon epithelial tissues were surgically obtained from Fukuoka University Hospital (Fukuoka, Japan). Frozen samples of these tissues were sliced to a thickness of 4 μm on glass slides. Proteins were then extracted by adding 75 μL of lysis buffer to the glass slide (total tissue lysate). The cell lysate was centrifuged at 15,000 rpm for 15 min at 4°C, and the supernatant was collected. The protein concentration was determined by a Bio-Rad Dye binding kit. BSA was used as the protein standard.
Immunohistochemistry. Frozen colon carcinoma and normal colon epithelial tissues were fixed with 4% paraformaldehyde/PBS for 10 min. Endogenous peroxidase was quenched with 3% H2O2 in methanol for 30 min. Tissue sections (4 μm) were treated mildly with protease XXIV for 1 min in 50 mmol/L Tris-HCl (pH 7.5), and the enzymatic reaction was stopped by washing thrice with PBS. Blocking was performed using the Histofine blocking kit (Nichirei Biosciences, Inc.) for 30 min. Samples were treated with either mAb 1H3 (10 μg/mL), D4B5 (5 μg/mL), or IgG (10 μg/mL) overnight at 4°C. After washing thrice with PBS, the sections were incubated with biotinylated rabbit anti-mouse immunoglobulin as a secondary antibody and then with HRP conjugated to streptavidin. Peroxidase activity was detected by using a solution of diaminobenzidine containing H2O2 in a Tris-HCl buffer, which develops a brown color in the presence of peroxidase activity.
Results
Specific detection of the monomeric form of Ln-γ2 using a mAb. The heterotrimeric structure of Ln-5 (left) and the domain structure of Ln-γ2 (right) are illustrated in Fig. 1A. Ln-γ2 is frequently processed by MMPs and major processed forms are also indicated in Fig. 1A (right). Human malignant melanoma Mum2B and human gastric carcinoma STKM-1 are representative cell lines that express Ln-γ2 as a monomer and as a heterotrimer (Ln-5), respectively. Thus, these cell lines were used as sources of monomeric Ln-γ2 and Ln-5, respectively (for purified proteins, see Fig. 3A).
Mice were immunized with monomeric Ln-γ2 purified from Mum2B cells and hybridoma clones were obtained. The mAbs obtained were screened by performing single ELISA assays using plates coated with either monomeric Ln-γ2 or Ln-5. Most of the mAbs that reacted with monomeric Ln-γ2 also recognized Ln-5 (data not shown). However, one clone (clone name: 1H3) produced an IgG1 class of mAb that reacted preferentially with Ln-γ2.
To confirm the reactivity and specificity of mAb 1H3 against monomeric Ln-γ2, the antibody was coated on assay plates and incubated with different concentrations of Ln-γ2 and Ln-5 as antigen. Antigens captured by the plates were detected by using anti-Ln-γ2 rabbit pAb (2778) and HRP-conjugated goat anti-mouse IgG. A negligible amount of Ln-5 was captured by mAb 1H3 even at the highest dose (240 ng/mL) used, whereas monomeric Ln-γ2 was captured in a dose-dependent manner (Fig. 1B,, top). Thus, mAb 1H3 preferentially recognizes the monomeric form of Ln-γ2. We also carried out similar experiments using a commonly used anti-human Ln-γ2 mAb, D4B5. Although mAb D4B5 captured the monomeric form of Ln-γ2 efficiently, it also reacted with Ln-5 (Fig. 1B , bottom) as reported previously.
We also performed immunoprecipitation assays using mAb 1H3 and mAb 2H2, an antibody that recognized both monomeric Ln-γ2 and Ln-5 during initial screening (data not shown). Purified monomeric Ln-γ2 and Ln-5 were immunoprecipitated with both antibodies and the immunoprecipitates were subjected to Western blotting using a pAb against Ln-γ2 (2778). Both monomeric Ln-γ2 and Ln-5 contained intact Ln-γ2 together with some processed forms of these proteins as detected by Western blotting (Fig. 2A, left; γ2, γ2', and DIII/V, respectively). Although the intact form of monomeric Ln-γ2 was precipitated by both antibodies (Fig. 2A,, middle), certain processed forms detected by the pAb 2778 were precipitated differentially by these two antibodies. This presumably reflects differences in the distribution of the recognition sites for these antibodies among the processed forms. In contrast, the Ln-γ2 chain in Ln-5 was precipitated by mAb 2H2 but not by mAb 1H3 (Fig. 2, right). Thus, the results are consistent with the ELISA assays and show that mAb 1H3 is selectively reactive with monomeric Ln-γ2.
Immunoprecipitation (IP) of Ln-γ2 by mAb 1H3. Purified monomeric Ln-γ2 or Ln-5 (1 μg) was incubated with immunoglobulin (IgG), mAb 1H3, or mAb 2H2 (1 μg/mL each). Antibody-antigen complexes were then precipitated and subjected to Western blotting (WB) using pAb 2778. In addition to the intact form of Ln-γ2 (140 kDa), NH2-terminal (70 kDa) and COOH-terminal (100 kDa) fragments processed at ps2 (Fig. 1A, right) are detected and indicated as γ2, DIII/IV, and γ2', respectively. Left, sample proteins (monomeric Ln-γ2 and Ln-5) used for the assay were directly analyzed by Western blotting using pAb 2778.
Immunoprecipitation (IP) of Ln-γ2 by mAb 1H3. Purified monomeric Ln-γ2 or Ln-5 (1 μg) was incubated with immunoglobulin (IgG), mAb 1H3, or mAb 2H2 (1 μg/mL each). Antibody-antigen complexes were then precipitated and subjected to Western blotting (WB) using pAb 2778. In addition to the intact form of Ln-γ2 (140 kDa), NH2-terminal (70 kDa) and COOH-terminal (100 kDa) fragments processed at ps2 (Fig. 1A, right) are detected and indicated as γ2, DIII/IV, and γ2', respectively. Left, sample proteins (monomeric Ln-γ2 and Ln-5) used for the assay were directly analyzed by Western blotting using pAb 2778.
Monomer-specific detection of LN-γ2 by Western blot analysis. We also tested whether mAb 1H3 displayed similar selectivity toward monomeric Ln-γ2 during Western blot analysis. Both pAb 2778 and mAb 1H3 detected equally well monomeric Ln-γ2 containing intact (γ2) and processed (γ2') forms as well as Ln-5 γ2 under conventional conditions of SDS-PAGE electrophoresis, which is done under reducing conditions (Fig. 3A). Conversely, mAb 1H3 only recognized monomeric LN-γ2 when electrophoresis was done under nonreducing conditions (Fig. 3B). Ln-5 γ2 was detected as a 450-kDa band by pAb 2778, reflecting the fact that this is a complex of three polypeptide chains held together by disulfide bonds (Fig. 1A,, left). In contrast, pAb 2778 detected monomeric Ln-γ2 as intact and processed forms as well as a high molecular weight band of 400 kDa, which probably corresponds to a Ln-γ2 homotrimer complex (Fig. 3B,, left). However, mAb 1H3 did not recognize Ln-γ2 in the 450-kDa Ln-5 complex, although it still reacted with monomeric Ln-γ2 (Fig. 3B,, right). Taken together, these results further confirm the specificity of mAb 1H3 for monomeric Ln-γ2 and show that this antibody can be used to specifically detect monomeric Ln-γ2 during Western blot analysis as long as electrophoresis is done under nonreducing conditions. The specificity and conditions of the antibodies used are summarized in Fig. 4D.
Use of mAb 1H3 in Western blot analysis. Purified monomeric Ln-γ2 (1 μg) and Ln-5 (3 μg) proteins were separated by 6% or 10% SDS-PAGE and blotted on PVDF membranes under reducing (A) or nonreducing (B) conditions, respectively, and then detected either by pAb 2778 or mAb 1H3. Arrows, Ln-γ2 homo-oligomer (γ2 oligo), monomer (γ2), its processed fragment (γ2'), and Ln-5.
Use of mAb 1H3 in Western blot analysis. Purified monomeric Ln-γ2 (1 μg) and Ln-5 (3 μg) proteins were separated by 6% or 10% SDS-PAGE and blotted on PVDF membranes under reducing (A) or nonreducing (B) conditions, respectively, and then detected either by pAb 2778 or mAb 1H3. Arrows, Ln-γ2 homo-oligomer (γ2 oligo), monomer (γ2), its processed fragment (γ2'), and Ln-5.
Epitope mapping of mAb 1H3. Recombinant DI/II, DIII, and DIV/V proteins of Ln-γ2 (250 ng/lane) were expressed in COS-1 cells and the conditioned medium was subjected to 12.5% SDS-PAGE followed by Western blot analysis using the indicated antibodies. A, pAb 2778. B, pAb 1963. C, mAb 1H3. pAb 1963 is an anti-rabbit pAb that is specific to DI/II of Ln-γ2 (10). The assay was carried out under reducing (Me+) and nonreducing (Me−) conditions. Arrows, DI/II, DIII, and DIII/V recombinant proteins. D, summary of the reactive sites of Ln-γ2 antibodies under nonreducing conditions.
Epitope mapping of mAb 1H3. Recombinant DI/II, DIII, and DIV/V proteins of Ln-γ2 (250 ng/lane) were expressed in COS-1 cells and the conditioned medium was subjected to 12.5% SDS-PAGE followed by Western blot analysis using the indicated antibodies. A, pAb 2778. B, pAb 1963. C, mAb 1H3. pAb 1963 is an anti-rabbit pAb that is specific to DI/II of Ln-γ2 (10). The assay was carried out under reducing (Me+) and nonreducing (Me−) conditions. Arrows, DI/II, DIII, and DIII/V recombinant proteins. D, summary of the reactive sites of Ln-γ2 antibodies under nonreducing conditions.
The cognate site of mAb 1H3 is located within the heterotrimeric coiled-coil region of Ln-5 γ2. The results of the Western blot analysis showed that mAb 1H3 specificity was retained after denaturation of the sample proteins with detergents as long as the assays were carried out under nonreducing conditions. However, mAb 1H3 reacted with Ln-γ2 released from the Ln-5 complex under reducing conditions. This suggested that the binding site of mAb 1H3 in Ln-γ2 was masked in Ln-5, whose structural stability depends on disulfide bond formation. To obtain an explanation for why mAb 1H3 differentially interacts with Ln-5 γ2 under reducing and nonreducing conditions, we performed epitope mapping. Ln-γ2 has five domains that are consecutively numbered from its NH2 terminus as illustrated in Fig. 1A. Three fragments containing different parts of Ln-γ2 (DI/II, DIII, and DIII/IV) were transiently expressed in COS7 cells as secreted proteins and partially purified. DIII/IV and DIII fragments were detected as 23- and 70-kDa bands, respectively, by Western blotting using pAb 2778 under reducing conditions (Fig. 4A, Me+). Under nonreducing conditions, both DIII and DIII/IV fragments were detected as monomers and homo-oligomers (Fig. 4A, Me−). The DI/II fragment was not recognized by pAb 2778 but was detected by pAb 1963 under both reducing and nonreducing conditions (Fig. 4B). mAb 1H3 recognized only the DI/II fragment under both reducing and nonreducing conditions (Fig. 4C). Thus, the epitope recognized by mAb 1H3 is in domain I or II, which forms the coiled-coil structure with the Ln-α3 and Ln-β3 chains and whose stability depends on disulfide bond formation. Unsurprisingly, this is the region of the ternary structure that is most severely occluded by the presence of other laminin chains (α3 and β3) and by disulfide bonds.
Use of mAb 1H3 for the detection of Ln-γ2 in tumor cell lines. Because mAb 1H3 selectively detects monomeric Ln-γ2 even by Western blotting, we used it to screen for the expression of Ln-γ2 in tumor cell lines. First, we tested Mum2B and STKM-1 cells as controls. Ln-γ2 was detected by mAb D4B5 in the extracts of both of these cells under nonreducing conditions (Fig. 5A, D4B5). Mum2B cells predominantly expressed monomeric Ln-γ2 (130-kDa band) with undetectable levels of heterotrimeric Ln-5 γ2 (450-kDa band) expression. On the other hand, STKM-1 cells expressed heterotrimeric Ln-5 γ2 with very little monomeric Ln-γ2 expression. Under the same assay conditions, mAb 1H3 detected the 130-kDa band in the Mum2B cell lysate but failed to detect the 450-kDa band in the STKM-1 cell extract (Fig. 5A, 1H3). Next, we extended the assay to eight other carcinoma cell lines derived from gastric, breast, skin, lung, and ovarian carcinomas. Both mAbs 1H3 and D4B5 detected monomeric Ln-γ2 as a 130-kDa protein in six carcinoma lines (TMK-1, MKN7, MDA-MB-231, MCF-7, A431, and PC-8) but not in MKN28 and SKOV-3 cell lines [Fig. 5B, D4B5 (top) and 1H3 (bottom)]. On the other hand, heterotrimeric Ln-5 γ2 (450 kDa) was detected only by mAb D4B5 in the MKN7 cell lysate (Fig. 5B, top). Thus, monomeric Ln-γ2 is expressed in a wide variety of carcinoma cell lines and mAb 1H3 can be efficiently used to distinguish its presence. These results also show for the first time that Western blotting under nonreducing conditions can be used to distinguish between monomeric and heterotrimeric forms of Ln-γ2 because they migrate to different positions on the SDS-PAGE gel under these conditions.
Screening for the expression of monomeric Ln-γ2 in human tumor cell lines. A, cell lysates prepared from Mum2B or STKM-1 cells (20 μg/lane each) were separated by 6% SDS-PAGE under nonreducing conditions and analyzed by Western blotting using mAbs 1H3 and D4B5. B, an additional eight human carcinoma cell lines were also analyzed similarly. Arrows, Ln-5 γ2 and monomeric Ln-γ2 chains. A and B, bottom, β-actin was detected as a loading control. C, summary of the expression of monomeric and heterotrimeric forms of Ln-γ2 in human cancer cell lines.
Screening for the expression of monomeric Ln-γ2 in human tumor cell lines. A, cell lysates prepared from Mum2B or STKM-1 cells (20 μg/lane each) were separated by 6% SDS-PAGE under nonreducing conditions and analyzed by Western blotting using mAbs 1H3 and D4B5. B, an additional eight human carcinoma cell lines were also analyzed similarly. Arrows, Ln-5 γ2 and monomeric Ln-γ2 chains. A and B, bottom, β-actin was detected as a loading control. C, summary of the expression of monomeric and heterotrimeric forms of Ln-γ2 in human cancer cell lines.
Application of mAb 1H3 for the detection of Ln-γ2 in human cancer tissues. To further evaluate the usefulness of mAb 1H3 in detecting monomeric Ln-γ2 in clinical samples, we carried out Western blot analysis on human colon cancer specimens. Tissue lysate was prepared from 4-μm-sliced frozen tissues from two colon carcinoma patients (patients #1 and #2) and subjected to Western blotting using mAbs 1H3 and D4B5 under nonreducing conditions. In the cancer tissue lysates, multiple Ln-γ2 bands migrating between 110 and 140 kDa were detected by both mAbs (Fig. 6A, 1H3 and D4B5). These bands presumably correspond to differently processed or glycosylated forms of Ln-γ2. In contrast, the 450-kDa band corresponding to heterotrimeric Ln-5 γ2 was only detected in the normal tissue lysate with mAb D4B5 but not with mAb 1H3 (Fig. 6A). This further confirms previous findings that monomeric Ln-γ2 is predominantly expressed in cancer tissues but not in normal epithelium.
Use of mAb 1H3 for the detection of monomeric Ln-γ2 in clinical samples. A, adenocarcinoma (T) tissues were obtained from two colon carcinoma patients and adjacent normal parts (N) were used as controls. Extracts were prepared from 4-μm-sliced frozen tissues on glass slides. The samples (15 μg/lane) were subjected to SDS-PAGE under nonreducing conditions and analyzed by Western blotting using mAbs 1H3 and D4B5 as indicated. Arrows, bands detected at 450, 140, 130, and 110 kDa. Right, human β-actin was used as a loading control. B, serial sections of normal colon (1–3) and carcinoma tissues (4–9) were examined by immunostaining with mAb D4B5 (3, 6, and 9) or 1H3 (2, 5, and 8). 1, 4, and 7, as a negative control, IgG was used for staining. The staining patterns were observed under a microscope (×200 or ×400 magnification) and representative fields are presented. 1 to 3, inset, enlargements (×1,200 magnification) of the squares surrounded by dots. Arrows, typical signals.
Use of mAb 1H3 for the detection of monomeric Ln-γ2 in clinical samples. A, adenocarcinoma (T) tissues were obtained from two colon carcinoma patients and adjacent normal parts (N) were used as controls. Extracts were prepared from 4-μm-sliced frozen tissues on glass slides. The samples (15 μg/lane) were subjected to SDS-PAGE under nonreducing conditions and analyzed by Western blotting using mAbs 1H3 and D4B5 as indicated. Arrows, bands detected at 450, 140, 130, and 110 kDa. Right, human β-actin was used as a loading control. B, serial sections of normal colon (1–3) and carcinoma tissues (4–9) were examined by immunostaining with mAb D4B5 (3, 6, and 9) or 1H3 (2, 5, and 8). 1, 4, and 7, as a negative control, IgG was used for staining. The staining patterns were observed under a microscope (×200 or ×400 magnification) and representative fields are presented. 1 to 3, inset, enlargements (×1,200 magnification) of the squares surrounded by dots. Arrows, typical signals.
Next, we used mAb 1H3 to immunostain frozen sections obtained from normal and carcinoma tissues from patient #1. The basement membranes of normal tissues were stained clearly with mAb D4B5 but not with mAb 1H3 (Fig. 6B, 1–3). In contrast, no basement membrane structure was observed in carcinoma tissue stained with mAb D4B5 but carcinoma cells invading the interstitial area were positively stained (Fig. 6B, 6 and 9). These carcinoma cells were also positive for staining with mAb 1H3 (Fig. 6B, 5 and 8). Thus, mAb 1H3 has strong potential as a tracking tool to detect the expression of monomeric Ln-γ2 directly in situ.
Discussion
We developed a new mAb (1H3) that preferentially recognizes monomeric Ln-γ2. The specificity of mAb 1H3 for monomeric Ln-γ2 is sufficient for the selective detection of monomeric Ln-γ2, even in the presence of Ln-5 γ2, by various immunologic detection methods, such as ELISA, immunoprecipitation, Western blotting, and immunohistochemistry.
The specificity of mAb 1H3 for monomeric Ln-γ2 most likely depends on the masking of its recognition site in Ln-5. Indeed, the epitope was mapped to domains I and II (Fig. 4), which are known to form a coiled-coil structure with the Ln-α3 and Ln-β3 chains. Thus, the epitope coincides with a region that, structurally speaking, is likely to be occluded in the presence of the other laminin chains of Ln-5. In addition, the stabilization of this region by multiple disulfide bonds may explain why mAb 1H3 does not react with Ln-5 even after it has been denatured and subjected to Western blot analysis under nonreducing conditions.
In parallel with the characterization of mAb 1H3, we also showed that expression of monomeric Ln-γ2 can be easily distinguished from that of Ln-5 γ2 by Western blotting under nonreducing conditions, even with commonly available antibodies. Under these conditions, Ln-5 γ2 can be detected as a 450-kDa protein cross-linked to other laminin chains via disulfide bonds, whereas non–cross-linked monomeric Ln-γ2 can be readily detected as a 130-kDa band. Thus, Western blotting under nonreducing conditions is the easiest way to detect monomeric Ln-γ2 with conventional antibodies, such as mAb D4B5. By using this method, we analyzed 10 carcinoma cell lines for Ln-γ2 expression (Fig. 5A and B). The expression patterns of these carcinoma cells are summarized in Fig. 5C. Eight of 10 cell lines expressed Ln-γ2. Six of the eight positive cell lines predominantly expressed Ln-γ2 in monomeric form (melanoma, lung, gastric, breast, skin, and ovarian cancers), whereas the other two expressed it predominantly as the Ln-5 heterotrimer (gastric carcinoma cell lines). These results corroborate previous reports claiming that malignant carcinoma cells frequently express Ln-γ2 in monomeric form, which were solely based on experiments showing the lack of Ln-α3 and Ln-β3 expression in carcinoma cells. Monomeric Ln-γ2 showed a single band of 130 kDa in the positive cell lines (Fig. 5A and B), whereas those detected in Fig. 4 under nonreducing conditions were 130 and 170 kDa in size. The Ln-γ2 detected in Fig. 5 presumably represents the proteins before secretion, whereas those detected in Fig. 3 correspond to secreted proteins purified from the conditioned medium. Thus, the differences in the molecular weight of Ln-γ2 between the two experiments are most likely due to differences in the glycosylation status and proteolytic processing of Ln-γ2.
Ln-γ2 was detected in two clinical samples by Western blotting under nonreducing conditions. Ln-5 γ2 was detected only in normal tissues where the expression level of the monomeric form is very low if not negligible (Fig. 6A,, D4B5). In contrast, most of the Ln-γ2 detected in colon carcinoma tissues was in the monomeric form (Fig. 6A,, 1H3). In addition, specificity of both antibodies (D4B5 and 1H3) was further confirmed even in immunostaining of the same sample in situ (Fig. 6B). The basement membrane of the normal tissue was stained with mAb D4B5 (Fig. 6B,, 3) but not with mAb 1H3 (Fig. 6B,, 2). This is because Ln-γ2 in the basement membrane is mostly part of the heterotrimer of 450 kDa as detected by Western blotting. In contrast, basement membrane structure was not evident in cancer tissue and the Ln-γ2 signal was detected by mAb D4B5 mostly within cancer cells (Fig. 6B,, 6 and 9). The Ln-γ2 in these carcinoma cells was also detectable by mAb 1H3, indicating that the Ln-γ2 signal detected by mAb D4B5 is in a monomeric state in these cells (Fig. 6B , 5 and 8).
In summary, we developed a new mouse mAb that selectively distinguishes between the monomeric form of Ln-γ2 and Ln-γ2 as part of the heterotrimeric Ln-5 complex. According to our knowledge, this is the first monomer-specific anti-Ln-γ2 antibody described in the literature. Because of this specificity, we believe it will be useful for the direct monitoring of monomeric Ln-γ2 expression in vivo and in vitro. Because the expression of the monomeric form of Ln-γ2 seems to be a specific marker of malignant tumor cells (4, 16, 17), the mAb 1H3 could have potential as a diagnostic tool for detecting invasive tumors in patients. In addition, it may also be possible to use this antibody as a vehicle to deliver antitumor agents to the vicinity of invasive tumors.
Acknowledgments
Grant support: Mitsui Life Science Welfare Foundation (N. Koshikawa), Special Coordination Fund for promoting Science, and Ministry of Education, Culture, Sports, Science and Technology of Japan Cancer Research grant-in-aid (N. Koshikawa and M. Seiki).
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.
We thank Professor Vito Quaranta for kindly providing us with Mum2B cell and anti-Ln-γ2 pAb and Drs. Andrew Sharabi (Baylor College of Medicine, Houston, TX) and Roy Zent (Vanderbilt University) for critical reading of this manuscript and for providing suggestions.