INTERNATIONAL JOURNAL OF ONCOLOGY 16: 549-553, 2000
`
`Analysis of glycoproteins in cancers and normal tissues
`reactive with monoclonal antibodies B3 and Bl
`
`ANGELA MARIANO1, ANGELINA DI CARLO2, IRA PASTAN3 and VINCENZO MACCHIA1
`
`Dipartimento di Biología e Patología Cellulate e Molecolare 'L. Califano', Università di Napoli 'Federico IF,
`Via Sergio Pansini 5, 80131 Napoli; Dipartimento di Patología Sperimentale, Biotecnologie Mediche,
`Infettivologia ed Epidemiología, Université di Pisa, Via Roma 55, 56126 Pisa, Italy;
`Laboratory of Molecular Biology, Division of Basic Sciences, National Cancer Institute,
`National Institutes of Health, 37 Convent Drive, Bethesda, 20892 MD, USA
`
`Received November 17, 1999; Accepted December 7, 1999
`
`Abstract. In this study we show by immunoblotting that
`Bl and B3, two newly isolated monoclonal antibodies,
`react with a variety of glycoproteins with different molecular
`weights expressed in stomach, pancreas, colorectal and breast
`cancers. The pattern of reactivity differed among cancers
`arising in different tissues, although no correlation has
`been observed with the histopathological characteristics of
`the lesion analysed. MAb B3 and MAb Bl, have a limited
`reactivity with peritumoral tissues, whereas react very
`strongly with metastatic lesion. Because of the limited
`reactivity of these antibodies with normal tissue, MAbs B3
`and Bl, armed with toxin in the form of recombinant
`immunotoxins, can be useful in treating certain kinds of
`cancer such as metastatic lesions. However, until current
`clinical trials are completed, wc will not know if they will be
`helpful in cancer treatment.
`
`Introduction
`
`Monoclonal antibodies that recognize antigens overexpressed
`in tumor cells are being used to develop new therapies for
`cancer (1). Many of these antibodies recognize normal or
`mutant proteins that are present on cancer cells. However,
`several of these antibodies recognize carbohydrate structures.
`Monoclonal antibodies that recognize the Lewisy antigen
`have been found to bind to many epithelial cancers, including
`cancers of the colon, ovary, breast, lung, esophagus, bladder
`and prostate. Besides reacting with epithelial cancers, mono-
`
`Correspondence to: Professor Vinccnzo Macchia, Dipartimento
`di Biología e Patología Cellularc e Molecolare 'L. Califano'. Via
`S. Pansini 5, 1-80131 Napoli. Italy
`
`Key words: monoclonal antibodies, LewisY carbohydrate antigens
`
`clonal antibodies Bl, B3 and BR96 also react with a limited
`number of normal human tissues (2,3). These include glands
`of the stomach, epithelium of the bladder and trachea and
`the superficial epithelium of the esophagus. Analysis of the
`reactivity of monoclonal antibodies Bl and B3 using a panel
`of carbohydrates has shown that they not only react with
`Lewisy but also react with some closely related carbohydrate
`structures and that the reactivities of MAbs Bl and B3
`with these related carbohydrates are different (2). Their
`differential reactivity with these carbohydrate structures
`indicate that they recognize different epitopes in Lewisy.
`Furthermore, a peptide mimitope has been isolated from a
`phage display library that binds to monoclonal antibody B3
`but not Bl (4). The BR96 antibody has also been found to
`react with carbohydrate structures related to Lcwisy (3). All
`three antibodies that bind to Lewisy have been shown to
`immunoprecipitate a variety of glycoproteins from carcinoma
`cell lines (2,5). Monoclonal antibodies Bl and B3 have been
`shown to react with glycoplipids extracted from carcinoma
`cells (our unpublished data). Despite extensive analysis of
`the reactivity of these antibodies with cultured cell lines,
`there is no information on the nature of the glycoproteins
`present in human tumors or normal human tissues that react
`with these antibodies.
`We have completed a phase I clinical trial utilizing an
`immunotoxin (LMB-1) in which MAb B3 is coupled to a
`mutant form of Pseudomonas exotoxin A (PE). In this trial,
`several clinical responses were observed (6). Because of the
`clinical activity of LMB-1, Phase II trials are being planned.
`Unfortunately, nothing is known about the nature of the
`antigens present in cancers that react with monoclonal
`antibody B3 or the related antibody Bl. This is particularly
`important because after binding, immunotoxins must be
`internalized in order to kill the target cell. To investigate
`the nature of the glycoproteins present in cancers and normal
`tissues, we obtained samples and nearby normal tissue removed
`at surgery. Membranes that react with these antibodies
`were then prepared and immunoblot analysis utilized to
`characterize the reactive glycoproteins. We find that cancers
`express a complex pattern of glycoproteins that reacts with
`MAbs B3 and Bl.
`
`IMMUNOGEN 2155, pg. 1
`Phigenic v. Immunogen
`IPR2014-00676
`
`

`

`550
`
`MARIANO etat MAb B3, M Ab HI AND CANCER
`
`Figure I. lmmunoblotling with MAb B3 of human gastrointestinal tumors.
`Molecular weight standards are shown on the left. In all samples, 15 (ig
`of protein was loaded on the gel. Lane 1. colon adenocarcinoma cell line
`(LS 174T); lane 2, colon adenocarcinoma; lane 3, caecum adenocarcinoma;
`lane 4. rectum adenocarcinoma; lane 5, rectum adenocarcinoma; lane 6,
`stomach adenocarcinoma; lane 7. stomach adenocarcinoma; lane 8. stomach
`adenocarcinoma; lane 9, stomach adenocarcinoma; lane 10. pancreatic
`adenocarcinoma.
`
`Figure 2. lmmunoblotling with MAb Bl of human gastrointestinal tumors.
`Molecular weight standards are shown on the left. In all samples. 15 fig
`of protein was loaded on the gel. Lane 1. colon adenocarcinoma cell line
`(LS 174T); lane 2, colon adenocarcinoma; lane 3, caecum adenocarcinoma;
`lane 4. rectum adenocarcinoma; lane 5. rectum adenocarcinoma; lane 6.
`stomach adenocarcinoma; lane 7. stomach adenocarcinoma; lane 8. stomach
`adenocarcinoma; lane 9, stomach adenocarcinoma; lane 10. pancreatic
`adenocarcinoma.
`
`Materials and methods
`
`Tissue and cells. Pathological human tissues were obtained
`in the operating room from patients undergoing surgery for
`neoplasms. Diagnosis of tumor was made by usual clinical
`and laboratory criteria and confirmed by histopathological
`findings. Peritumoral tissues, histologically normal, excised
`surgically along with large tumors were used as control
`tissue. Freshly collected tissues were cut in two parts: one
`was fixed in formalin for histological diagnosis and the
`other rapidly frozen and stored at -80°C to be analysed with
`immunoblotting. A431 (human epidermoid carcinoma cell
`line) and LS 174T (human colon adenocarcinoma cell line)
`were purchased from the American Type Culture Collection.
`Cell lines were cultured and maintained in Dulbecco's modified
`Eagle's Medium (DMEM) with pennicillin (50 |ag/ml),
`streptomycin (50 units/ml) and 5% fetal bovine serum.
`
`Materials. Aprotinin, PMSF, BSA. Hepes were purchased
`from Sigma Chemical Co. (St. Louis, MO). The BioRad
`protein assay reagent was from BioRad Laboratories. All
`other reagents were available from commercial sources.
`
`Antibodies. The monoclonal antibodies B3 and Bl were
`prepared in the Laboratory of Dr Ira Pastan (NCI, NIH,
`Bethesda, MD) purified from serum-free culture medium
`(B3) or ascites (Bl) by ammonium sulfate precipitation and
`chromatography on MonoQ and TSK-250 gel filtration
`columns. The purity of both antibodies was established by
`sodium dodecyl sulfate/polyacrylamide gel electrophoresis
`(SDS-PAGE) (7). Properties of MAbs B3 and Bl, both of the
`IgGlk isotype, have been previously described (2). Antibody
`against EGF-R (Ab 2913) was generated by immunizing a
`rabbit with an affinity purified EGF receptor preparation
`excised from an SDS-PAGE as previously described (8).
`
`Plasma membrane preparation. Approximately 500 mg of each
`tissue sample (wet weight) was washed, minced in 0.154 M
`NaCl on ice and subjected to five strokes of homogenization
`in a Teflon-glass homogenizer in 2 ml of buffer A (10 mM
`Tris HCl at pH 7.4, 50 mM NaCl, 1% aprotinin, 1% PMSF
`and 5 mM EDTA). The homogenate was centrifuged for
`10 min at 1,500 x g, the supernatant was further centrifuged
`for 30 min at 105,000 x g and pellets and supernatants
`were collected. The pellets (crude plasma membrane) were
`resuspended in ice cold buffer A using a glass homogenizer
`and pestle, divided in aliquots and stored at -S0°C. Protein
`was determined by the Bradford procedure (9). Cells
`harvested from culture dishes were resuspended in ice cold
`Buffer A and homogenized with 3 strokes in a Teflon-glass
`homogenizer. The homogenate was centrifuged for 30 min at
`105,000 x g and the crude membrane pellets were solubilized
`in IX Laemmli buffer.
`
`Immunoblotting of MAb B3. MAb Bl and anti-EGF-R Ab 2913
`to plasma membranes. Crude plasma membrane proteins
`from human peritumoral and tumoral tissues and from A431
`and LS 174T cells were resolved by SDS-PAGE performed
`with a Protean II cell system using 7.5% gels (BioRad). Prior
`to electrophoresis, the crude plasma membrane samples were
`resuspended in ice cold buffer A and diluted in loading
`sample buffer under reducing condition and heated at 95°C
`for 4 min as described by Laemmli (7). For Western blotting,
`samples after electrophoresis were transferred from gels to
`nitrocellulose sheets (Millipore) as described by Towbin et al
`(10). The nitrocellulose sheet was incubated with 5% non-fat
`dry milk in Tris-buffered saline at 20°C for 2 h and processed
`by adding the antibody B3 or the anti-EGF-R Ab 2913 for
`1 h at room temperature. Antibodies were detected by a
`secondary biotinylated Ab and ECL kit as recommended by the
`manufacturer (Amersham). All nitrocellulose sheets exposed to
`MAb B3 were stripped and reprobed with MAb Bl.
`
`IMMUNOGEN 2155, pg. 2
`Phigenic v. Immunogen
`IPR2014-00676
`
`

`

`INTERNATIONAL JOURNAL OF ONCOLOGY 16: 549-553. 2000
`
`551
`
`Figure 3. Immunoblotting with MAb 133 of human breast cancers and
`adjacent normal tissues. Molecular weight standards are shown on the
`left. In all samples. 15 U,g of protein was loaded on the gel. Lane 1, colon
`adenocarcinoma cell line (LS 174T); lane 2, breast cancer; lane 3, breast
`peritumoral tissue: lane 4, breast perilumoral tissue; lane 5. colon adeno­
`carcinoma cell line (LS 174T); lane 6. breast cancer.
`
`Figure 4. Immunoblotting with MAb Bl of human breast cancers and
`adjacent normal tissues. Molecular weight standards are shown on the
`left. In all samples, 15 |ig of protein was loaded on the gel. Lane 1, colon
`adenocarcinoma cell line (LS 174T); lane 2. breast cancer; lane 3, breast
`peritumoral tissue; lane 4. breast peritumoral tissue; lane 5, colon adeno­
`carcinoma cell line (LS 174T); lane 6, breast cancer.
`
`Results
`
`To analyse the nature of the glycoproteins present in cancers
`and normal tissues, several different types of cancers and
`adjacent normal tissue were analysed. Plasma membranes
`were prepared from these samples, the proteins were separated
`on 7.5% SDS-PAGE. and the gels stained with monoclonal
`antibodies B3 or Bl. For reference purposes a sample derived
`from a colon cancer cell line, LS 174T, was also examined.
`Furthermore, the gels after stripping procedure were probed
`with an antibody to tubulin as a positive control to ensure
`that protein loading was similar in each of the lanes (data not
`shown). Typical results with several cancers of gastrointestinal
`origin are shown in Figs. 1 and 2. It is evident that many
`glycoproteins of different sizes react with both antibodies.
`These range in size from over 200 kDa to under 44 kDa. A
`second point is that the pattern of reactivity varies when
`comparing tumors of the same type as well as comparing
`tumors of different types. Four colorectal cancer samples were
`analysed and they all showed different patterns. All four
`specimens reacted with MAb B3 although two of the four
`samples stained weakly showing a strong band with a
`molecular weight of about 200 kDa. One of the four specimens
`reacted strongly with MAb Bl showing a series of bands
`beginning with a high molecular weight band that was present
`at the top of the gel and extending down to the 44,000
`molecular weight standard; two of the samples showed
`moderate staining with the Bl antibody and one showed no
`reactivity with MAb Bl. The same amount of protein was
`loaded on the gel in all samples and equal loading was
`confirmed by the exposure of the stripped nitrocellulose
`sheet to tubulin antibody (data not shown). In regard to the
`stomach cancer samples, each of the four specimens analysed
`showed a different pattern. All four specimens reacted with
`both antibodies although one of the four samples stained
`
`weakly with both antibodies and another stained weakly
`with MAb B3 (Fig. 1, lane 7). This result is consistent with
`our previous observation using immunohistochemistry that
`indicated that about 75% of stomach cancers react with
`MAb B3 or MAb Bl (2). Furthermore, one pancreatic adeno­
`carcinoma has been analysed and it stained strongly with
`both antibodies showing a complex pattern of reactivity with
`a series of bands of different molecular weight.
`M Abs B3 and Bl react with many breast cancers.
`Therefore, two breast cancer samples were examined by
`SDS-PAGE and immunoblotting. The immunostained gels
`are shown in Figs. 3 and 4. In both examples, the cancer
`samples stained very strongly, whereas there was no detectable
`reactivity with the adjacent normal breast tissue. In both
`breast cancer samples, the pattern of reactivity was very
`similar. The strongest reactivity was with a broad band with
`a molecular weight of about 144 kDa. However, several
`other bands with lower and higher molecular weights were
`detected. The pattern in breast cancer of antibody reactivity
`was very different from the colon cancer cell line used as a
`standard.
`We also analysed one sample in which a colon cancer
`had metastasized to lung and several primary lung adeno­
`carcinomas. Figs. 5 and 6 show that the metastatic colon
`tumor sample reacted very strongly with MAbs B3 and Bl.
`There was very little reactivity with normal lung or with the
`primary lung cancer samples examined. The metastatic colon
`lesion showed a complex pattern of reactivity with both
`antibodies.
`The complex pattern shown in Figs. 1-6 could be the
`result of many different glycoproteins reacting with the
`antibody or the proteolytic degradation of a few proteins
`into smaller molecular weight species. To determine if
`proteolytic degradation had occurred, we stained a group of
`cancer samples with a polyclonal antibody to the human
`
`IMMUNOGEN 2155, pg. 3
`Phigenic v. Immunogen
`IPR2014-00676
`
`

`

`552
`
`MARIANO etal: MAbB3, MAbBl AND CANCER
`
`Figure 5. ImmunobloUing with MAb B3 of human lung adenocarcinomas.
`Molecular weight standards are shown on the left. In all samples, 15 u,g
`of protein was loaded on the gel. Lane I. colon adenocarcinoma cell line
`(LS 174T): lane 2. lung adenocarcinoma: lane 3, lung peritumoral tissue;
`lane 4, king peritumoral tissue; lane 5, lung metastases from primary colon
`adenocarcinoma; lane 6, lung adenocarcinoma.
`
`Figure 7. ImmunobloUing with Ab EGF-R of human primary tumors.
`Molecular weight standards are shown on the left. Protein concentration in the
`samples of crude plasma membrane fractions of primary human tumors and
`in the sample from A431 cells were about 150 \ig and 30 u.g respectively.
`Lane I, colon adenocarcinoma; lane 2. stomach adenocarcinoma; lane 3.
`epidermoid carcinoma cell line (A431); lane 4, breast cancer.
`
`Discussion
`
`The results shown here indicate that in stomach, pancreas,
`colorectal and breast cancers, and in metastatic cancers
`monoclonal antibodies B3 and Bl react with a variety of
`glycoproteins with different molecular weights. It is striking
`that the bands are very diffuse in contrast to the EGF receptor
`band which is sharp. This result indicates that the broad bands
`are composed of several different species of glycoproteins
`with different molecular weights. This is the expected result
`since there is no known mechanism whereby the Ley epitope
`would be specifically present on one or a few glycoproteins.
`In addition, the pattern of reactivity differed among cancers
`arising in different tissues, although no correlation has been
`observed with the histopathological characteristics of the
`lesion analysed. We have previously noted that MAbs B3
`and B1 react with a variety of glycoproteins present in MCF-7
`breast carcinoma cells (2). It is likely that the formation of Ley
`in the cancers is due to induction of a fucosyl transferase by the
`transformation process and that the addition of fucose occurs
`on many different glycoproteins (11). The internalization of
`the glycoproteins by endocytosis is able to bring the immuno-
`toxin into the cell so that intoxication and cell death can
`take place. The different pattern of glycoproteins present in
`different cancers is probably a reflection of the differences
`in the glycoprotein composition of these cancers due to
`differences in differentiation. It is currently not clear what
`selective advantage certain cancers gain by activating the Le-V
`pathway. Ley is present on normal stomach and continues to
`be expressed in many stomach cancers. However, there is
`very low Ley expression in normal colon whereas expression
`is very strong in the vast majority of colon cancers and of
`metastatic colon cancers. A clinical trial with LMB-1 has
`been completed in which shrinkage of colon and breast cancers
`has been observed (6). In this study different grades of
`shrinkage of the tumor masses has been observed. It is not clear
`why some tumors are more sensitive than others to LMB-1.
`
`Figure 6. Immunoblotting with MAb Bl of human lung adenocarcinomas.
`Molecular weight standards are shown on the left. In all samples, 15 [ig of
`protein was loaded on the gel. Lane I, colon adenocarcinoma cell line
`(LS 174T); lane 2, lung adenocarcinoma; lane 3, lung peritumoral tissue;
`lane 4, lung peritumoral tissue; lane 5, lung metastases from primary colon
`adenocarcinoma; lane 6, lung adenocarcinoma.
`
`EGF receptor which is known to be a single species. Fig. 7
`shows that a single strong band with a molecular weight of
`about 170 kDa is detected by the antibody in samples of
`colon, stomach and breast cancers. As a positive control,
`we used the A431 cell line which makes a large amount of
`the EGF-receptor. The major band detected by the antibody
`in this cell line is in the same position as the band in the
`cancer samples. This result indicates that the proteins in the
`membrane samples were not degraded to a significant degree.
`In addition, when the membrane proteins were stained with
`Coomassie blue, many sharp bands of different sizes were
`noted indicating very little degradation had occurred (data
`not shown).
`
`IMMUNOGEN 2155, pg. 4
`Phigenic v. Immunogen
`IPR2014-00676
`
`

`

`INTERNATIONAL JOURNAL OF ONCOLOGY 16: 549-553. 2000
`
`553
`
`There are several factors that could contribute to variable
`responses. One is that the maximum tolerated dose in humans
`is less than the dose that regularly causes tumor regression in
`animal models (12). If higher doses could be given to humans,
`more responses might be observed. Other possible explanations
`for variable responses are differences in tumor penetration
`due to variation in vascular permeability, differences in
`interstitial pressure in tumors (13), and differences among
`the cell surface glycoproteins to which Ley is attached. The
`latter is important because the immunotoxin must bind to
`glycoproteins that are internalized to kill cells.
`Several strategies are being described to increase the
`responses to immunotoxin. A recombinant
`immunotoxin
`that is smaller in size and can penetrate into tumors more
`effectively has been produced and a clinical trial with this
`smaller recombinant form of LMB-1 is ongoing at the
`National Institutes of Health. Therapy of solid tumors with
`immunotoxins is not expected to replace either surgery or
`radiation therapy. The therapeutic target for immunotoxins
`can be a useful addition to the standard procedures of surgery,
`radiation and chemotherapy. In practice, immunotherapy can
`be used to cure small tumors such as metastases. We have
`found that antibodies B3 and Bl react very strongly with
`metastatic lesions showing a complex pattern of reactivity.
`Specificity of both antibodies is determined by the
`distribution of the target antigens; we have also shown the
`limited reactivity of these antibodies with normal tissues.
`Therefore MAbs B3 and Bl, armed with toxin in the form
`of recombinant immunotoxins, can be useful in treating
`certain kinds of cancer such as metastatic lesions. However,
`until current clinical trials are completed we will not know if
`these new agents will produce substantial clinical responses
`and whether they will be useful drugs in cancer treatment.
`If the clinical trials are succesful we will try to correlate
`responses with the nature of the Ley-containing glycoproteins
`in each primary tumor or metastatic lesion to determine if
`certain patterns of antibody reactivity will predict a response
`to the treatment.
`
`References
`
`1. Junghans RP, Sgouros G and Schcinbcrg DA: Antibody-based
`immunotherapies for cancer. In: Chemotherapy and Biotherapy.
`2nd edition. Chabner BA and Longo DL (eds). Lippincott-
`Raven, Philadelphia, pp655-6S9, 1996.
`2. Pastan I, Lovelace ET, Gallo MG, Rutherford AV. Magnani JL
`and Willingham MC: Characterization of monoclonal antibodies
`Bl and B3 that react with mucinous adenocarcinomas. Cancer
`Res 51: 3781-3787, 1991.
`3. Hellstrom I, Garrigues J, Garrigues U and Hellstrom KE:
`Highly tumor-reactive, internalizing, mouse monoclonal
`antibodies to Ley-related cell surface antigens. Cancer Res 50:
`2183-2190,1990.
`4. Hocss R, Brinkmann U, Handel T and Pastan I: Identification of
`a peptide which binds to the carbohydrate specific monoclonal
`antibody B3. Gene 128: 43-49, 1993.
`5. Garrigues J, Anderson J, Hellstrom KE and Hcllstrom I: Anti­
`tumor antibody BR96 blocks cell migration and binds to a
`lysosomal membrane glycoprotein on cell surface microspikes
`and ruffled membranes. J Cell Biol 125: 129-142, 1994.
`6. Pai LH, Wittes R. Sctser A, Willingham MC and Pastan I:
`Treatment of advanced solid tumors with immunotoxin LMB-1:
`an anti-Lewisy antibody linked to recombinant Pseudomonas
`exotoxin, PE38. Nat Med 2: 350-353. 1996.
`7. Laemmli UK: Cleavage of structural proteins during the
`assembly of the head of bacteriophace T4. Nature 227: 680-685,
`1970.
`8. Bcguinot L, Werth D, Ito S, Richert N, Willingham MC and
`Pastan I: Functional studies on the EGF receptor with an
`antibody that recognizes the intracellular portion of the receptor.
`J Biol Chem 261: 1801-1807. 1986.
`9. Bradford MM: A rapid and sensitive method for quantitation
`of microgram quantities of protein utilizing the principle of
`protein-dye binding. Anal Biochem 72: 248-254, 1976.
`10. Towbin H, Staehelin T and Gordon J: Elcctrophoretic transfer
`of proteins from Polyacrylamide gels to nitrocellulose sheets:
`procedure and some applications. Proc Natl Acad Sei USA 76:
`4350-4354, 1979.
`11. Fukuda M: Possible roles of tumor-associated carbohydrate
`antigens. Cancer Res 56: 2237-2244, 1996.
`12. Pai LH and Pastan I: Immunotoxin therapy for cancer. JAMA
`78: 269-273, 1993.
`13. Jain RK and Baxter LT: Mechanisms of heterogeneous
`distribution of monoclonal antobodies and other macro-
`molecules in tumors: significance of elevated interstitial
`pressure. Cancer Res 48: 7022-7032, 1988.
`
`IMMUNOGEN 2155, pg. 5
`Phigenic v. Immunogen
`IPR2014-00676
`
`