[CANCER RESEARCH 52. 2310-2317, April 15, 1992] Second Generation Anti-MUCl Peptide Monoclonal Antibodies Pei-xiang Xing, Julie Prenzoska, Kaylene Quelch, and Ian F. C. McKenzie1 Austin Research Institute, Austin Hospital, Heidelberg 3084, Victoria, Australia ABSTRACT Second generation antibodies to mammary mucins were produced by immunizing mice with a peptide with a sequence deduced from that of the MUC1 complementary DNA sequence (PAHGVTSAPDTRPAPGS- TAP). Four monoclonal antibodies (BCP7-10) were produced which gave different reactions. BCP8 was similar in tissue reactivity (by immunoper- oxidase staining) to anti-breast cancer or anti-human milk fat globule membranes (HMFG) antibodies and reacted strongly with most breast cancers and a proportion of other adenocarcinomas, whether formalin fixed or fresh, and reacted less strongly with some normal tissues. The three other antibodies (BCP7, BCP9, BCP10) reacted only with fresh tissues or a single cell line (LS174T of colon cancer origin) and gave variable weak reactions. Like many anti-mucin antibodies BCP8 reacted with HMFG, but more strongly with deglycosylated HMFG; analysis with peptides by enzyme-linked immunosorbent assay indicated reactivity with an epitope contained in the amino acid motif PDTR and using the pepscan method, the minimum epitope was DTR. M Alis BCP7, BCP9, and BCP10 did not react with HMFG; substantial reactions were obtained with deglycosylated HMFG for BCP7 and weaker reactions with BCP9 and BCP10. The finding that BCP7 reacted with breast cancer tissues and deglycosylated HMFG suggested that the epitope recognized by BCP7 was masked in native form and exposed in cancer, indicating that BCP7 could be a useful agent for analyzing differences between normal and cancer mucins. The amino acid epitopes for these antibodies were VTSA (BCP7), GSTAP (BCP9), and RPAP (BCP10). For BCP8, amino acid substitution analysis of SAPDTR indicated that substitutions were poorly tolerated (except Q for T and L/Y for R), contrasting with the substitution analysis of anti-mucin antibody reactions where virtually any amino acid can be substituted for T, indicating that in the native state T (threonine) may be O-glycosylated. The use of synthetic peptides to produce antibodies similar to those produced using crude mucins or tumor extracts represents a major advance in the production of antitumor reagents. INTRODUCTION Mucins are produced in large amounts in breast cancer and many anti-breast cancer monoclonal antibodies react with these mucins (1-4). The analysis of such antibodies has advanced recently with the cloning of the cDNA2 encoding the protein core of the major mammary mucin (MUC1) (5-8). Within the cDNA sequence there are a VNTR where a 60-base pair nu- cleotide sequence is repeated 40-80 times (7-9). When this VNTR nucleotide sequence was translated to a 20-amino acid sequence, it was found that a number of antibodies to native mucins reacted with this peptide, in particular with the sequence APDTRPA in the VNTRs (10-12). We now report a series of second generation monoclonal antibodies made to synthetic peptides the sequence of which was derived from the VNTR sequence of the MUC1. These anti-peptide antibodies were found to react with breast cancer cells and appeared similar on tissue sections to the anti-breast cancer or anti-mucin antibod- Received 9/26/91; accepted 2/11/92. 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. 1To whom requests for reprints should be addressed. 2The abbreviations used are: cDNA. complementary DNA; HMFG, human milk fat globule membranes; mAb, monoclonal antibody; VNTR, variable number of tandem repeats. ies; they detected a range of epitopes, some not previously described. MATERIALS AND METHODS Peptide Synthesis. Peptides synthesized using an ABI peptide syn thesizer (Foster City, CA) were (Table 1): (a) those with sequences derived from the MUCJ gene, representing the hydrophilic region of the VNTR, regions 5' and 3' to the VNTR, and cytoplasmic tail peptides; (¿>) peptide MI29 derived from the MUC2 gene, and peptide SIB35, derived from the MUC3 gene (13, 14); (e) T4N1 representing the NH2 terminal of mouse CD4 and used as a negative control. In addition, peptides were synthesized on polyethylene pins (Cambridge Research Biochemicals, Cambridge, United Kingdom) (15): the first set contained twenty 6-mer overlapping peptides covering the twenty amino acids of the VNTR of MUC1 and were used to map the epitopes detected by mAbs (16); the second set contained 120 of the 6-mer peptides, where each amino acid in SAPDTR was substituted by the other nineteen amino acids; these were used to study the role of each amino acid in the epitope. MAb Production and Testing. The peptide C-p 13-32 (sequence shown in Table 1 and which contains an NH2-terminal cystine (C) so that dimers would form) was conjugated to keyhole limpet hemocyanin (11), and 100 ^g of C-pl3-32-keyhoIe limpet hemocyanin emulsified with complete Freund's adjuvant were injected i.p. into BALB/c mice. Second and third injections were given 4 and 6 weeks later without adjuvant, and a fusion performed 3 days later (17). Enzyme-linked immunosorbent assay and immunoperoxidase methods were used for examining supernatants of hybridomas (11, 17). The isotypes of mAbs were determined using a gel immunodiffusion kit (Serotec, England). HMFG and deglycosylated HMFG were prepared and tested as de scribed ( 17). The mAbs were tested by FACScan flow cytometry (Becton Dickinson, Mountain View, CA) using colon cancer cell lines HT29, HT29 SB, LS174T, and LOVO; breast cancer cell lines T47D, MCF7, MDA231; and pancreatic carcinoma cell line Pane 89 (18). The mAb E4.3 (IgG2a) to human CD46 and the mAb anti-Ly2.1 (IgG2a) to mouse antigen Ly2.1 were used as positive and negative controls, respectively. The mAbs BC2, HMPV, and 4B6 (IgM), produced by immunization with HMFG, and recognizing the peptide epitopes APDTR (BC2 and HMPV) and DTR (4B6), were used for comparison with BCP8 in reactions with HMFG and peptides (19). IgG antibodies were purified using a protein A-Sepharose 4B affinity column and IgM antibodies were precipitated by 33% saturated am monium sulfate (17). To study the relationship between the epitopes recognized by the mAbs, competitive binding radioimmunoassays were performed (17). Briefly, mAbs were labeled with 125I(Amersham Inter national, Amersham, England) using the chloramine-T method (20), and a constant amount of labeled mAb was mixed with un labeled mAb at increasing concentrations; the mixture was transferred to HMFG or C-pl3-32 coated plates and left at 4°Covernight. The binding of '"!- labeled antibody to the plates was measured using a gamma counter. The percentage of inhibition was then calculated by comparing the binding of mAb with and without unlabeled antibodies: % of inhibition = 11 —¿ Binding of mAb with inhibitor\ Binding without inhibitor / x 100 2310 Association for Cancer Research. by guest on August 25, 2020. Copyright 1992 American https://bloodcancerdiscov.aacrjournals.org Downloaded from
Transcript
[CANCER RESEARCH 52. 2310-2317, April 15, 1992]
Second Generation Anti-MUCl Peptide Monoclonal AntibodiesPei-xiang Xing, Julie Prenzoska, Kaylene Quelch, and Ian F. C. McKenzie1
Austin Research Institute, Austin Hospital, Heidelberg 3084, Victoria, Australia
ABSTRACT
Second generation antibodies to mammary mucins were produced byimmunizing mice with a peptide with a sequence deduced from that ofthe MUC1 complementary DNA sequence (PAHGVTSAPDTRPAPGS-TAP). Four monoclonal antibodies (BCP7-10) were produced which gavedifferent reactions. BCP8 was similar in tissue reactivity (by immunoper-oxidase staining) to anti-breast cancer or anti-human milk fat globulemembranes (HMFG) antibodies and reacted strongly with most breastcancers and a proportion of other adenocarcinomas, whether formalinfixed or fresh, and reacted less strongly with some normal tissues. Thethree other antibodies (BCP7, BCP9, BCP10) reacted only with freshtissues or a single cell line (LS174T of colon cancer origin) and gavevariable weak reactions. Like many anti-mucin antibodies BCP8 reactedwith HMFG, but more strongly with deglycosylated HMFG; analysiswith peptides by enzyme-linked immunosorbent assay indicated reactivitywith an epitope contained in the amino acid motif PDTR and using thepepscan method, the minimum epitope was DTR. MAlis BCP7, BCP9,and BCP10 did not react with HMFG; substantial reactions were obtainedwith deglycosylated HMFG for BCP7 and weaker reactions with BCP9and BCP10. The finding that BCP7 reacted with breast cancer tissuesand deglycosylated HMFG suggested that the epitope recognized byBCP7 was masked in native form and exposed in cancer, indicating thatBCP7 could be a useful agent for analyzing differences between normaland cancer mucins. The amino acid epitopes for these antibodies wereVTSA (BCP7), GSTAP (BCP9), and RPAP (BCP10). For BCP8, aminoacid substitution analysis of SAPDTR indicated that substitutions werepoorly tolerated (except Q for T and L/Y for R), contrasting with thesubstitution analysis of anti-mucin antibody reactions where virtually anyamino acid can be substituted for T, indicating that in the native state T(threonine) may be O-glycosylated. The use of synthetic peptides toproduce antibodies similar to those produced using crude mucins or tumorextracts represents a major advance in the production of antitumorreagents.
INTRODUCTION
Mucins are produced in large amounts in breast cancer andmany anti-breast cancer monoclonal antibodies react with thesemucins (1-4). The analysis of such antibodies has advancedrecently with the cloning of the cDNA2 encoding the proteincore of the major mammary mucin (MUC1) (5-8). Within thecDNA sequence there are a VNTR where a 60-base pair nu-cleotide sequence is repeated 40-80 times (7-9). When thisVNTR nucleotide sequence was translated to a 20-amino acidsequence, it was found that a number of antibodies to nativemucins reacted with this peptide, in particular with the sequenceAPDTRPA in the VNTRs (10-12). We now report a series ofsecond generation monoclonal antibodies made to syntheticpeptides the sequence of which was derived from the VNTRsequence of the MUC1. These anti-peptide antibodies werefound to react with breast cancer cells and appeared similar ontissue sections to the anti-breast cancer or anti-mucin antibod-
Received 9/26/91; accepted 2/11/92.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 inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1To whom requests for reprints should be addressed.2The abbreviations used are: cDNA. complementary DNA; HMFG, human
ies; they detected a range of epitopes, some not previouslydescribed.
MATERIALS AND METHODS
Peptide Synthesis. Peptides synthesized using an ABI peptide synthesizer (Foster City, CA) were (Table 1): (a) those with sequencesderived from the MUCJ gene, representing the hydrophilic region ofthe VNTR, regions 5' and 3' to the VNTR, and cytoplasmic tail
peptides; (¿>)peptide MI29 derived from the MUC2 gene, and peptideSIB35, derived from the MUC3 gene (13, 14); (e) T4N1 representingthe NH2 terminal of mouse CD4 and used as a negative control. Inaddition, peptides were synthesized on polyethylene pins (CambridgeResearch Biochemicals, Cambridge, United Kingdom) (15): the first setcontained twenty 6-mer overlapping peptides covering the twenty amino
acids of the VNTR of MUC1 and were used to map the epitopesdetected by mAbs (16); the second set contained 120 of the 6-mer
peptides, where each amino acid in SAPDTR was substituted by theother nineteen amino acids; these were used to study the role of eachamino acid in the epitope.
MAb Production and Testing. The peptide C-p 13-32 (sequence shownin Table 1 and which contains an NH2-terminal cystine (C) so that
dimers would form) was conjugated to keyhole limpet hemocyanin (11),and 100 ^g of C-pl3-32-keyhoIe limpet hemocyanin emulsified withcomplete Freund's adjuvant were injected i.p. into BALB/c mice.
Second and third injections were given 4 and 6 weeks later withoutadjuvant, and a fusion performed 3 days later (17). Enzyme-linked
immunosorbent assay and immunoperoxidase methods were used forexamining supernatants of hybridomas (11, 17). The isotypes of mAbswere determined using a gel immunodiffusion kit (Serotec, England).HMFG and deglycosylated HMFG were prepared and tested as described (17). The mAbs were tested by FACScan flow cytometry (BectonDickinson, Mountain View, CA) using colon cancer cell lines HT29,HT29 SB, LS174T, and LOVO; breast cancer cell lines T47D, MCF7,MDA231; and pancreatic carcinoma cell line Pane 89 (18). The mAbE4.3 (IgG2a) to human CD46 and the mAb anti-Ly2.1 (IgG2a) to
mouse antigen Ly2.1 were used as positive and negative controls,respectively. The mAbs BC2, HMPV, and 4B6 (IgM), produced byimmunization with HMFG, and recognizing the peptide epitopesAPDTR (BC2 and HMPV) and DTR (4B6), were used for comparisonwith BCP8 in reactions with HMFG and peptides (19).
IgG antibodies were purified using a protein A-Sepharose 4B affinity
column and IgM antibodies were precipitated by 33% saturated ammonium sulfate (17). To study the relationship between the epitopesrecognized by the mAbs, competitive binding radioimmunoassays wereperformed (17). Briefly, mAbs were labeled with 125I(Amersham Inter
national, Amersham, England) using the chloramine-T method (20),
and a constant amount of labeled mAb was mixed with un labeled mAbat increasing concentrations; the mixture was transferred to HMFG orC-pl3-32 coated plates and left at 4°Covernight. The binding of '"!-
labeled antibody to the plates was measured using a gamma counter.The percentage of inhibition was then calculated by comparing thebinding of mAb with and without unlabeled antibodies:
% of inhibition = 11 —¿�Binding of mAb with inhibitor\Binding without inhibitor /
x 100
2310
Association for Cancer Research. by guest on August 25, 2020. Copyright 1992 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
From two fusions, four antibodies, BCP7 (IgG2a), BCP8(IgG2b), BCP9 (IgGl), and BCP10 (IgM), were produced whichwere reactive with the immunizing synthetic peptide C-p 13-32,and were nonreactive with the control peptide T4N1 by enzyme-linked immunosorbent assay. These mAbs were further testedon other peptides, on whole and deglycosylated HMFG, andby the immunoperoxidase staining method on formalin-fixedand fresh tissues (Tables 1-4).
Reaction of mAbs with Synthetic Peptides, HMFG, andEpitope Mapping
To confirm that the mAbs reacted specifically with peptidesencoded by the VNTR region of MUC1, peptides representingdifferent parts of VNTR, carboxy and amino terminal toVNTR, cytoplasmic tail, and peptides encoded by the MUC2and MUC3 genes were separately examined (Table 1; Fig. 1).In addition, each antibody was tested on native and deglycosylated HMFG. Finally, epitope mapping was performed for eachantibody.
BCP7. In addition to the immunizing peptide (C-P13-32)this mAb reacted with peptides p 13-32 (monomer of C-p 13-32), pl-24, and p5-20 but did not react with pl-15 or A-pl-15(Fig. 1). By comparing the common sequences present in thesepeptides, it was clear that the amino acids AHGVTSA werelikely to contain the reactive epitope. To precisely define theepitope, the pepscan method was used (Fig. 2). In this methodoverlapping peptides were examined and indicated that 3 peptides, HGVTSA, GVTSAP, and VTSAPD, were clearly reactive with BCP7; these share the amino acids VTSA. This is anew epitope, not previously defined by antibodies made to
native mucins. Of interest was the reproducible activity of thepeptide GSTAPP, which has no overlapping amino acid withthe other 3 peptides (Fig. 2A); perhaps this attains a conformation similar to VTSA (ATS in GSTAPP would be verysimilar to VTS). It was also noted that BCP7 was nonreactivewith the native mucin HMFG but reacted strongly with deglycosylated HMFG (Fig. 3).
BCP8. This antibody gave reactions very similar to thosepublished elsewhere for anti-mucin antibodies which react withthe amino acids APDTR (11, 17). BCP8 reacted with theimmunizing peptide C-pl3-32 and its monomer pi3-32 andwith pl-24 (Fig. 1; Table 1). Of interest was the lack of reactivityof p5-20 and the differential reaction of A-pl-15 and pl-15indicating that the amino acids PDTR are weakly reactive, butAPDTR give the stronger reaction (Fig. 1). When overlappingpeptides were examined, the common amino acids were DTR,although, as will be further shown below, the epitope is PDTRwith some contribution by the preceding amino acid (A) (Figs.1, 2, and 4). This antibody reacted with both native and deglycosylated HMFG (Fig. 3).
BCP9. This antibody reacted with similar peptides to BCP8(Table 1), and was reactive with deglycosylated HMFG, but notwith native HMFG (Fig. 3; Table 1). Epitope mapping indicatedthe minimum epitope to be GSTAP (Fig. 2).
BCP10. This antibody reacted with the immunizing peptideand its monomer p 13-32, but only weakly with pl-24 (Table 1;Fig. 1). The reactive epitope appeared to be RPAP (Fig. 2).BCP10 also reacted with deglycosylated HMFG at high concentrations (>10 Mg/ml), but not with HMFG (Fig. 3).
Further Studies of the Epitope Detected by BCP8
mAb BCP8 resembled many antibodies made to breast cancercells or to HMFG in that it reacted with the epitope APDTR.
Table 1 Reaction of mAbs (BCP7-BCP10) with synthetic peptides tested by ELISA"
Peptide Amino acid sequence BCP7 BCP8 BCP9 BCP10
MUCl VNTRpl-24*p5-20p 13-32C-pl3-32cpl-15A-pl-15
5' region to VNTRp31-55p51-70
3' region to VNTRp344-364P408-423
Cytoplasmic tail of MUClp471-493P507-526
MUC2MI-29
MUC3SIB-35
Mouse CD4T4NI
HMFG"
Deglycosylated HMFG
PDTRPAPGSTAPPAHGVTSAPDTRPAPGSTAPPAHGVTSA
PAHGVTSAPDTRPAPGSTAP(C)PAHGVTSAPDTRPAPGSTAP
PDTRPAPGSTAPPAHAPDTRPAPGSTAPPAH
TGSGHASSTPGGEKETSATQRSSVPRSSVPSSTEKNAVSMTSSVL
NSSLEDPSTDWQELQRDISETGFNQYKTEAASRVNL
AVCQCRRKNYGQLDIFPARDTYH(C)YVPPSSTDRSPYEKVSAGNG
KYPTTTPISTTTMVTPTPTPTGTQTPTTT
CHSTPSFTSSITTTETTSHSTPSFTSSITTTETTS
KTLVLGKEQESAELPCECY
" —¿�,absorbance at 405 nm <0.25; +, 0.25-1.0; ++, absorbance 1-1.5; +++, absorbance >1.5 at 1/10 dilution of tissue culture supernatant of hybridomas.* In the peptides of VNTR pl-24 and p 13-32 the numbers refer to a 24-amino acid sequence of the VNTR (11). Other numbers (p31-55, p51-70, etc.) refer to the
amino acid position in the published sequence (7, 8).cC-pl3-32 was the immunizing peptide used to produce the BCP7-BCP10 mAbs.d For comparison HMFG and deglycosylated HMFG were also prepared and used to test in the enzyme-linked immunosorbent assay (17).
2311
Association for Cancer Research. by guest on August 25, 2020. Copyright 1992 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
Fig. 1. Enzyme-linked immunosorbenl assayreaction of anti-C-p!3-32 mAbs BCP7 (A).BCP8 (A), BCP9 (C). and BCP10 (D) usingculture supernatants in doubling dilution (II2048). Peptides used were: p5-20 (D); pI-24 (*);pl3-32 (Q); C-pl3-32 (O); pl4-24 •¿�pl-15(D); A-pl-15 (A); T4NI (A). The peptide T4N1was used as a negative control.
SECOND GENERATION mAbs TO BREAST CANCER
A
2.0-
1.5
1.0-
10-
10'" 10"3 10"2 10"' 10° 101
DILUTION OF OF BCP 7
10"-
0.5-
10"* 10"J IO"« 10"'
DILUTION OF BCP8
DILUTION OF BCP 9
10"' 10"J 1O"¿ 10" '
DILUTION OF BCP 10
10" 10'
2.0-
1.5-1.0-0.5-n
n-HC
GVTSAVTSAPVTSA=0GSTAPP.
PDTRPAPGSTAPPAHGVTSA
OVERLAPPING PEPTIDES
PDTRPAPGSTAPPAHGVTSA
OVERLAPPING PEPTIDES
B2.0-
1.5- AP DTP PP DTP PA
DTR PAP
IPDTRPAPGSTAPPAHGVTSA
OVERLAPPING PEPTIDES
lili... ..... JPDTRPAPGSTAPPAHGVTSA
OVERLAPPING PEPTIDES
Fig. 2. Reactivity of mAbs. BCP7 (A). BCP8(B), BCP9 (O, and BCP10 (D) with overlappingpeptides synthesized on polyethylene pins. Thesingle letter code on the abscissa represents thefirst amino acid of each overlapping 6-mer peptide.e.g.. P = PDTRPA. D = DTRPAP. etc.
It was therefore important to make a comparison of the reactionof the anti-peptide with anti-mucin antibodies in APDTR. Foranti-mucin antibodies such as 4B6, the epitope of which is 3-mer DTR (the same as BCP8), as defined by pepscan method,amino acid substitution studies using peptides on polyethylenepins show that P and R cannot be substituted by any otheramino acid (19), A and D can be substituted only by relatedamino acids, and T can be substituted by virtually any aminoacid (19) indicating it is likely to be O-glycosylated and essentially not part of the APDXR epitope (X = any amino acid).However, for BCP8 (the epitope of which is also DTR), whichwas produced by immunizing with a synthetic peptide C-pl3-
32 (where there is obviously no glycosylation), substitutionanalysis indicated that D and P could not be substituted, Rcould be substituted (by L and Y), but unlike 4B6, T could notbe substituted by amino acids other than Q. The difference inthreonine (T) substitution patterns of anti-mucin and anti-peptide antibody is clear (Figs. 4 and 5) and leads us to concludethat anti-mucin antibodies react with APDXR, and anti-peptideantibodies react with APDTR; i.e., T (threonine) is a criticalpart of the epitope for the anti-peptide antibodies but not forthe anti-mucin antibodies. The T is glycosylated and thereforehidden in native mucin although it is likely that the conformation of this epitope in native mucin and in a synthetic peptide
2312
Association for Cancer Research. by guest on August 25, 2020. Copyright 1992 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
Fig. 3. Enzyme-linked immunosorbent assay reaction of mAbs BCP7 (»),BCP8 (d),BCP9 (O), and BCP10 (•)with HMFG (A)and deglycosylated HMFG (B). Anti-HMFGmAbs BC2 (Q) and 4B6 (A) were positivecontrols. Anti-MUC2 mAb CCP58 (A) wasused as an negative control. Supernatants ofmAbs were used in tripling dilution.
Fig. 4. Reaction of mAb BCP8 with pep-tides synthesized on polyethylene pins. *,amino acids substituted by other amino acidsin the parent peptide SAPDTR shown on abscissa; percentage of mean parent response wascalculated by comparison of absorbance of substituted peptides with that of the parent pep-tide SAPDTR.
may differ. The finding that both HMFG and deglycosylatedHMFG react with the anti-peptide antibody BCP8 could indicate that not all threonine residuces in APDTR in HMFG areglycosylated,Competitive Inhibition of the mAbs Binding to HMFG and C-pl3-32
Antibodies known to react with slightly different epitopeswithin APDTR were examined with BCP8 in blocking studies.
SAPDT-
The antibodies used and epitopes detected by pepscan methodwere: BC2 (APDTR); HMPV (APDTR); 4B6 (DTR); andBCP8 (DTR). We examined the ability of these 4 antibodies toinhibit the binding of 125I-radiolabeled BCP8 (2 ¿ig/ml),4B6 (4
Mg/ml), or BC2 (4 /¿g/ml)to peptide C-p 13-32 (Fig. 6) or
HMFG (not shown). Other antibodies, BCP7 (VTSA), BCP9(GSTAP), BCP10 (RPAP), and CCP58 (to MUC2 peptide;epitope not mapped) (21), to different MUC1 or MUC2 epi-
2313
Association for Cancer Research. by guest on August 25, 2020. Copyright 1992 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
Fig. 5. Comparison of the reaction of BCP8 and 4B6 with substituted peptidesof SAPD'R (*, amino acid T which has been substituted by other amino acids).
topes were also examined. Clearly, all 4 antibodies (BC2, 4B6,HMPV, and BCP8) gave 100% inhibition; i.e., each was effectively able to totally block the binding of the radiolabeledantibody to either substrate (peptide or HMFG) (Fig. 6). Thedifferences noted in the curves are likely to be due to differencesin antibody binding; the important feature is that these antibodies all give ~100% inhibition; i.e., all antibodies react withAPDTR. It was also noted that antibodies (BCP7, BCP9,BCP10) to other epitopes gave little inhibition, which was ofinterest because the 3 other MUC1 epitopes all map within 15amino acids of each other, yet no inhibition occurred. However,reciprocal blocking studies were not performed with I25I-BCP7,-9, or -10 and it may be that these lower affinity antibodiescould easily be displaced by higher affinity anti-mucinantibodies.
Reaction of m.\ l>with Normal and Malignant Tissues byImmunoperoxidase Staining
The anti-peptide antibodies were tested on formalin-fixedand fresh tissue sections, by immunoperoxidase staining (Tables
2, 3, and 4). It was found that mAb BCP8 reacted with 39 of41 (95%) formalin-fixed breast cancer specimens, of which 34samples (82%) were stained strongly. Most of the staining wascytoplasmic and membranous, with some staining in the basement membrane of the glandular cells (Fig. 7), similar to thedistribution found with anti-mucin antibodies (17). BCP8 wasnot breast cancer specific, inasmuch as it reacted with cancersof colon, ovary, lung, kidney, pancreas, stomach, and prostatebut not with lymphoma and neuroblastoma (Table 2). BCP8also reacted with some normal epithelial tissues such as colon,salivary glands, ovary, lung, kidney, pancreas, and stomach butless strongly than with tumor tissues (Table 3). The other threemAbs had different reactivity patterns. BCP7 reacted weaklywith formalin fixed breast cancers (3 of 10), while it did notreact with any normal tissues. BCP9 and BCP10 did not stainformalin-fixed tissues. Both BCP7 and BCP8 reacted with freshbreast (4 of 4) and colon cancer tissues (4 of 4) but not withnormal colon tissue (Table 4). BCP9 and BCP10 reacted with1 of 4 of fresh breast cancer tissues but not with normal andmalignant colon tissue.
Reaction of mAbs with Human Cancer Cell Lines
The immunofluorescence technique was used to detect theepitopes recognized by mAbs in tumor cell lines (Table 5). MAbBCP8 reacted with the breast cancer cell lines T47D and MCF7and weakly with colon cancer cell line LS174T but was notreactive with other cell lines. MAbs BCP7, BCP9, and BCP10reacted only with LS174T (Table 5).
DISCUSSION
In the past monoclonal antibodies to mucins have beenproduced by immunizing with crude or purified mucins, particularly HMFG, or whole tumors or their extracts. In this way,many antibodies have been made which are reactive with breastcancer cells and on further analysis these have clearly falleninto several groups, wherein the antibody reacts with predomi-
Fig. 6. Percentage inhibition of mAbsBCP8 (2 jig/ml) (A). BC2 (4 Mg/ml) (B), and4B6 (4 fig/ml) (C), of binding to C-pl.V32 bymAbs BC2,4B6. BCP7. BCP8. BCP9, BCP10.HMPV. and anti-MUC2 antibody CCP58.
CONCENTRATION OF INHIBITORS (|ig/ml) CONCENTRATION OF INHIBITORS (jig/ml)
BC2
4B6
BCP7
BCP8
BCP9
BCP10
HMPV
CCP58
CONCENTRATION OF INHIBITORS (jig/ml)
2314
Association for Cancer Research. by guest on August 25, 2020. Copyright 1992 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
Fig. 7. Immunoperoxidase staining (arrows) of breast cancer tissue (A) , andnormal breast tissue (B) with mAb BCP8 at 1/10 dilution of supernatant, x 400.
nantly carbohydrate epitopes or predominantly peptide epitopes(9, 11, 22-26). With recent advances both in synthetic carbohydrate chemistry and in the cloning of the cDNA encodingthe protein core of the mucins comes the ability to make anti-
mucin antibodies by using synthetic moieties, be they carbohydrate or peptide, as described herein. This represents a substantial advance in being able to use clearly defined antigens ratherthan a mixture of materials obtained by extracts from tissuesor mucins. It remains to be seen in clinical studies whetherthese monoclonal antibodies are of greater diagnostic or therapeutic value than those made by immunizing with crude reagents; nonetheless important questions can be answered.
The four antibodies produced (BCP7, BCP8, BCP9, andBCP10) all reacted with the immunizing peptide and to a
varying extent with native and deglycosylated mucin and withtissue sections. One antibody, BCP8, most closely resemblesother antibodies made to mucins and tumors in that it reactswith formalin-fixed and fresh tissues; it has a differential reaction with tumors and normal tissue and, like many anti-mucinpeptide antibodies (12, 17,23), reacts with an epitope containedwithin amino acids APDTR. The other antibodies, BCP7,BCP9, and BCP10, show lesser reactions on formalin-fixedtissue but reacted with fresh tissues. This differentiation between formalin-fixed and fresh tissues is not unexpected, because formalin is known to cross-link amine groups of aminoacids and may destroy amino acid residues or mask the antibody-binding sites and would be expected to have a predominant effect on proteins rather than carbohydrate in tissues. Thefinding is of little relevance other than that it is more difficultto obtain fresh than formalin-fixed tissues for testing. Nonetheless, the antibodies, particularly BCP9 and 10, appear to havea restricted tissue distribution as seen from the limited results(Table 4) and this is further shown with reactivity on cell lineswhere only the colon cancer cell line LS174T was reactive,compared with BCP8 which reacted with LS174T and withseveral breast cancer cell lines (T47D, MCF7, and NH2A)(Table 5). However, antibodies BCP7 and BCP9 identified thenew peptide epitopes VTSA (BCP7) and GSTAP (BCP9),which have not been previously detected with monoclonal anti-mucin antibodies. It was interesting to find that BCP7 did notreact with native mucin (HMFG) but reacted strongly withdeglycosylated HMFG (Fig. 3) and breast cancer tissues. Thismay indicate that the amino acids VTSA recognized by BCP7are masked in native form and exposed in cancer mucin. Thereactivity with tumors could mean that T and S are not glyco-sylated in the reactive molecules.
A comparison of the activity of anti-APDTR antibodies madeagainst mucin (e.g., BC2) and against peptides (e.g., BCP8) inan amino acid substitution study indicated that the threonine(T) may be glycosylated in the native state (e.g., HMFG)although other explanations, such as conformational changes,are also possible. A recent consideration has been that as tumorsdevelop, the mucins have variations in their degrees of glyco-
sylation exposing more protein and that tumor mucins aretherefore different from normal mucins. The ability of anti-peptide antibodies (made against both mucins and syntheticmaterial) to react more strongly with deglycosylated materialwould support this. However, at this time there is no clearevidence that the glycosylation of threonine in tumors andnormal mucins are different.
The observations reported here are not new for antibody-antigen reactions, but are novel for antitumor antibodies. How-
Table 2 Reaction of mAb BCP8 with formalin-fixed malignant tissues tested by immunoperoxidase staining1
% of cellstainingCarcinomatissuesBreastColonLiverOvarianBladderLungKidneyStomachPancreaticProstateNo.
of positivestaining/totaltested39/413/81/11/2I/«1/32/24/41/22/20-525010200106-25100000000026-50420000020251-75811010110076-10026001011110Negative2501020010Staining
" The results of BCP7. BCP9. and BCP10 were not shown because BCP7 reacted weakly with formalin-fixed tissues, and BCP9 and BCP10 did not. BCP8 did not
react with other normal tissues, such as gallbladder (0 of 1), intestine (0 of 1), brain (0 of 1), muscle (0 of 5), liver (0 of 2). adrenal (0 of 2). parathyroid (0 of 1),prostate (0 of 2), spleen (0 of 1), and lymphocyte (0 of 3).
b Sebaceous gland staining.
Table 4 Reaction of mAb BCP7, BCP8, BCP9, and BCPIO with frozen normal and malignant tissues tested by immunoperoxidase staining
% of cellstainingTissueBCP7Breast
cancerColoncancerBCP8Breast
cancerColoncancerBCP9Breast
cancerBCPIOBreast
cancerNo.
of positivestaining/totaltested4/44/44/44/41/41/40-50000336-2500000126-5001041051-7521200076-100222000Negative000033Staining
" BCP7 and BCP8 did not react with normal colon tissue (0 of I). BCP9 and BCPIO did not react with normal colon (0 of 1) or with colon cancer tissue (0 of 4).
Table 5 Expression ofepitopes recognised by anti-peptide C-pl3-32 mAhs oncancer cell lines tested by flow cytometry"
Cancer andcelllineColon
HT29HT29 SB'
LS 174TLOVOBCP76.97.9
46.610.4BCP89.7
9.633.0
7.8BCP93.72.241.7
7.5BCPIO11.24.838.215.1E4.3*
Anti-Ly2.TBC2¿87.2
69.776.872.66.13.512.010.415.2
14.331.2
8.5
BreastMDA231T47DMCF7NH2A
PancreasPane 89
4.0 12.99.4 80.9
15.0 59.518.2 31.4
3.1 22.5
3.95.7
12.417.6
3.0
1.76.6
17.79.7
1.0
84.696.076.992.3
8.9
1.47.7
15.521.2
5.0
14.587.361.952.8
27.0" Values indicate the percentage of cells staining.* Positive control antibody to the human cell surface marker CD46.c Negative control antibody, anti-mouse T-cell surface antigen Ly2.1.* BC2 is an anti-HMFG mAb. which also reacts VNTR-encoded peptide.' HT29 SB is a stably differentiated mucus-secreting derivative of HT29 that
was isolated after treatment with 5 m.vi sodium butyrate for 23 days (27).
ever, the usefulness of anti-peptide antibodies for diagnosticserum tests, for immunoscintigraphy, or for therapeutic purposes remains to be evaluated. At this time, there is no clearevidence that peptide epitopes provide a preferential target forsuch antibodies in either diagnosis or therapy.
ACKNOWLEDGMENTS
We would like to thank Dr. J. Trapani and V. Apostolopoulos forhelpful discussions and T. Athanasiadis for secretarial assistance.
REFERENCES
1. Ceriani, R. L., Peterson, J. A.. Lee, J. Y., Moneada, R., and Blank, E. W.Characterisation of cell surface antigens of human mammary epithelial cellswith monoclonal antibodies prepared against human milk fat globule. Somat.Cell Genet., 9: 415-427, 1983.
2. Taylor-Papadimitriou, J., Peterson, J. A, Burchell, J., Arklie, J., Ceriani, R.L., and Bodmer, W. T. Monoclonal antibodies to epithelium specific components of the milk fat globule membrane: production and reactions withcells in culture. Int. J. Cancer, 28: 17-21, 1981.
3. Sekine, H., Ohno, T., and Kufe, D. W. Purification and characterazation ofa high molecular weight glycoprotein detectable in human milk and breastcarcinoma. J. Immunol., 135: 3610-3615, 1985.
4. McKenzie, I. F. C., and Xing, P-X. Mucins in breast cancer: recent immu-nological advances. Cancer Cells, 2: 75-78, 1990.
5. Genlder, S. J., Burchell, J. M., Duhig, T.. Lamport, D.. White, R., Parker,M., and Taylor-Papadimitriou, J. Cloning of partial cDNA coding differentiation and tumour-associated mucin glycoproteins expressed by humanmammary epithelium. Proc. Nati. Acad. Sci. USA, 84: 6060-6064, 1987.
6. Siddiqui, J., Abe, M., Hayes, D., Shani, E., Yunis, E., and Kufe. D. Isolationand sequencing of a cDNA coding for the human DF3 breast carcinomaassociated antigen. Proc. Nati. Acad. Sci. USA, 85: 2320-2323, 1988.
7. Wreschner, D. H.. Hareuveni. M., Tsarfaty, I., Smorodinsky. N., Horev, J.,Zaretsky, J., Kotkes, P., Weiss, M., Lathe, R., Dion, A., and Keydar, 1.Human epithelial tumor antigen cDNA sequences—differential splicing maygenerate multiple protein forms. Eur. J. Biochem., 189: 463-474, 1990.
8. Lightenberg. M. J. L., Vos, H. L., Gennissen. A. C. M., and Hilkens, J.Episialin. a carcinoma-associated mucin, is generated by a polymorphic geneencoding splice variants with alternative amino-termini. J. Biol. Chem., 265:5573-5578, 1990.
9. Gendler, S. J., Taylor-Papadimitrou, J., Duhig, T.. Rothbard, J., and Burchell, J. A highly immunogenic region of a human polymorphic epithelialmucin expressd by carcinomas is made up of tandem repeats. J. Biol. Chem.,263: 12820-12823, 1988.
10. Burchell. J., Taylor-Papadimitriou, J., Boshell, M., Gendler, S.. and Duhig,T. A short sequence, within the amino acid tandem repeat of a cancer-associated mucin. contains immunodominant epitopes. Int. J. Cancer, 44:691-696, 1989.
11. Xing, P-X. Reynolds. K., Tjandra. J. J., Tang, X. L., and McKenzie, I. F. C.Synthetic peptides reactive with anti-human milk fat globule membrane.Cancer Res., 50: 89-96. 1990.
12. Price, M. R., Hudecz, F., O'Sullivan, C., Baldwin, R. W., Edwards, P. M.,
2316
Association for Cancer Research. by guest on August 25, 2020. Copyright 1992 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
and Tendier, S. J. B. Immunological and structural features of the proteincore of human polymorphic epithelial mucin. Mol. Iminiino!.. 27: 795-802,1990.
13. Gum, J. R., Byrd, J. C, Hicks, J. W., Toribara, N. W., Lamport, D. T. A.,and Kim, Y. S. Molecular cloning of human intestinal mucin cDNAs:sequence analysis and evidence for genetic polymorphism. J. Biol. Chem.,264:6480-6487, 1989.
14. Gum. J. R., Hicks. J. W., Swallow, D. M., Lagace. R. L.. Byrd, J. E.,Lamport. D. T. A., Siddick. B.. and Kim, Y. S. Molecular cloning of cDNAsderived from a novel human intestinal mucin gene. Biochem. Biophy. Res.Commun., 171: 407-415, 1990.
15. Geyson, H. M., Rodda, S. J.. Mason, T. J., Tribbick, G.. and Schoofs. P. G..Strategies for epitope analysis using peptide synthesis. J. Immunol. Methods.702:259-274, 1987.
16. Xing. P-X, Reynolds, K., Pietersz, G., and McKenzie, I. F. C. Effect ofvariations in peptide sequence on anti-human milk fat globule membraneantibody reactions. Immunology. 72: 304-311, 1990.
17. Xing, P-X., Tjandra, J. J., Stacker. S. A., Teh, J. G.. Thompson, C. H.,McLaughlin. P. J., and McKenzie, I. F. C. Monoclonal antibodies reactivewith mucin expressed in breast cancer. Immunol. Cell. Biol., 67: 183-185,1989.
18. Johnstone, R. W., Andrew, S. M., Hogarth, P. M., Pietersz, G. A., andMcKenzie, I. F. C. The effect of temperature on the binding kenetics andequilibrium constants of monoclonal antibodies to cell surface antigens. Mol.Immunol., 27: 327-333, 1990.
19. Xing. P-X., Prenzoska, J., and McKenzie, I. F. C. Epitope mapping of anti-
breast and anti-ovarian mucin monoclonal antibodies. Mol. Immunol., inpress, 1992.
20. Harlow, E., and Lane, D. Antibodies: a Laboratory Manual, p. 328. ColdSpring Harbor, NY: Cold Spring Harbor Laboratory, 1988.
21. Xing, P-X., Prenzoska, J., Layton, G. T., Devine, L. P., and McKenzie, I. F.C. Second generation monoclonal antibodies to intestinal MUC2 peptidereactive with colon cancer. J. Nati. Cancer. Inst.. in press, 1992.
22. Hilkens, J., Brijs, F. F., and Ligtenberg, M. Complexity of MAM-6, anepithelial sialomucin associated with carcinomas. Cancer Res., 49: 786-793.1989.
23. Burchell. J.. Gendler. S.. Taylor-Papadimitriou. J.. Girling, A., Lewis, A.,Millis, R., and Lampost, D. Development and characterization of breastcancer reactive monoclonal antibodies directed to the core protein of thehuman milk mucin. Cancer Res.. 47: 5476-5482. 1987.
24. Stahli. C., Caravatti, M., Acschbacher, M.. Kocyba, C.. Takacs, B.. andCarmann. H. Mucin-like carcinoma-associated antigen defined by threemonoclonal antibodies against different epitopes. Cancer Res.. 48: 6799-6802. 1988.
25. Abe, M., and Kufe, D. Structural analysis of the DF3 human carcinoma-associated protein. Cancer Res., 49: 2834-2839, 1989.
26. Stacker, S. A., Tjandra, J. J.. Xing, P-X.. Walker, C. H., Thompson, C. H.,and McKenzie, I. F. C. Purification and biochemical characterization of anovel breast carcinoma associated mucin-like glycoprotein defined by antibody 3E1.2. Br. J. Cancer, 59: 544-553, 1989.
27. Angeron, C., and Laboisse, C. L. Emergence of permanently differentiatedcell clones in a human colonie cancer cell line in culture after treatment withsodium butyrate. Cancer Res., 44: 3961-3966, 1984.
2317
Association for Cancer Research. by guest on August 25, 2020. Copyright 1992 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
![Austin Tate [Ai Austin] AIAI, University of Edinburgh](https://static.documents.pub/doc/80x56/56812c3b550346895d90c479/austin-tate-ai-austin-aiai-university-of-edinburgh-56e603690dae6.jpg)
