isolation and identification of human ... - cancer research · a number of relevant reports dealing...

[CANCER RESEARCH 37, 1313-1322, May 1977]

Other authors reported the extraction of lung TM's solublein 50% ammonium sulfate,but also demonstrated theseantigens not to be cross-reactive with CEA (36, 37). Twoantigens extractable with 0.85% NaCI solution were isolatedby Mclntire and Sizaret (26). These antigens did not crossreact with CEA, AFP, a-2 H-globulin, or f3-fetoprotein. Watson et a!. (42, 43) utilized an immunoadsorption column towhich normal lung-soluble proteins were coupled to isolateantibodies to lung TM from rabbit antisera. One of theirlung TM's was cross-reactive with embryonic extractswhile the other was not. Both of these latter lung TM's werefound in extracts of several histopathological types of lungtumors. Recently, Hollinshead et a!. (15, 16) successfullyisolated and purified a lung TM and demonstrated its immunological specificity by means of skin testing. Thesedata, along with the detection of a circulating lung TM bycross-over immunoelectrophoresis by Viza et al. (41), support the existence of human lung TM. However, no attempthas been made to adequately characterize, identify, and/orcorrelate the above findings.

We report the isolation, characterization, and identification of 3 human lung TM's found in significant concentrations in several histopathological types of lung tumors.

MATERIALS AND METHODS

Tissue Extraction Protocol. Tumor tissue removed at surgery was collected into a sterile container and placed in anice bath. The tissue was taken directly to and examined bythe surgical pathologist who aseptically dispensed tissuesfor diagnostic and research purposes. Paired pathologyslides and pathology reports were requested at the time oftissue procurement.

Once in the laboratory, the tumor and nontumor siteswere isolated and processed separately. All work from thispoint on was performed under aseptic conditions. Onepiece of tissue from each specimen was removed for microbiological analysis and another piece was removed forpreparation of whole cellular vaccines and tissue culturestudies. The remainder of the tissue was first minced in BSSwithout glucose and phenol red, pH 6.6. These tissue sliceswere then homogenized with a Brinkmann/Polytron ModelPT1O-35 homogenizer (Brinkmann Instruments, Inc. , West

bury, N. Y.) using the PT10-ST generator. The latter was a 2-stage process, and the 1st 30-sec homogenization cycle

1313MAY 1977

Isolation and Identification of Human Lung Tumor-associatedAntigens1

Robert W. Veltri, Henry F. Mengoli, Peter E. Maxim, Shirley Westfall, Joseph M. Gopo, Chaog-Wei Huang,and Philip M. Sprinkle

Divisionof OtolaryngologyfR.W.V.,P. E. M., S. W.,P. M. S.)and Departmentof Microbiology(R. W.V.,H. F. M.J,WestVirginiaUniversityMedicalCenter,Morgantown,WestVirginia26506;Departmentof Biology,Universityof Zambia,Lusaka,Zambia(J. M. G.J;and Departmentof Biochemistry,UniversityofWashington,Seattle,Washington98195(C-W.H.J

SUMMARY

Three human lung tumor-associated antigens (TAA's)have been identified in soluble and membrane-solubilizedextracts of human squamous cell lung carcinoma with theuse of antisera raised in rabbits. The antigens were identified and partially characterized by means of an agaroseadsorption technique. These antigens, termed lung TAA's 1,2, and 3, are all soluble in 50% ammonium sulfate, areantigenically distinct, and do not cross-react with carcinoembryonic antigen or a-fetoprotein. Lung TAA's 1 and 2are oncofetal antigens demonstrable in soluble extractsfrom 24-week-old but not from 26-week-old fetal lungs.Rabbit antibodies to these lung TM's were not adsorbed bytypes A, B, and 0 human red blood cells, serum proteins, aswell as soluble or insoluble lung preparations. Of severalcommercial antisera to human proteins, none cross-reactedwith lung TAA 1, but anti-human liver ferritin cross-reactedwith lung TM 2, and anti-human lactoferrin cross-reactedwith lung TAA 3. Lung TAA 1 was partially adsorbed andcross-reacted with certain normal serum or plasma preparations used and appears to be a normal serum protein inCohn Fraction IV-4. Lung TAA 2 and 3 appear only in lungtumor-soluble extracts, whereas the lung TAA 1 was demonstrable in soluble extracts of breast, colon, cervical,and head and neck carcinoma. All may be tumor markers ofvalue in immunodiagnosis.

INTRODUCTION

The discovery of CEA2 in carcinoma of the colon (11) andthe AFP found in the serum of patients with hepatocellularcarcinoma (38) resulted in numerous publications on human TM's.

A number of relevant reports dealing with isolation, identification, and characterization of lung carcinoma (lungTM) need mention. Yachi et a!. (45) extracted 2 lung TM'ssoluble in 50% ammonium sulfate with molecular weightsexceeding that of 7S lgG and designated them X and Y.

1 This project was funded by a Cancer Immunodiagnosis Contract, NO1-

CB-43890, from the National Cancer Institute, USPHS.2 The abbreviations used are: CEA, carcinoembryonic antigen; AFP, a-

fetoprotein; TAA, tumor-associated antigen; BSS, Hanks' balanced salt solution; PBS, 0.01 M phosphate-buffered 0.14 M NaCI.

Received July 26, 1976; accepted January 28, 1977.

Research. on January 26, 2021. © 1977 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

A. W. Ve!trietal.

was performed in BSS with the speed setting number at 5on the rheostat(blade speed, 21,000 rpm). The homogenateat this point contains intact cells and debris when observedby light microscopy. This homogenate was centrifuged at8000 x g in a Sorvall RC-2B at 4Â°for 15 mm. The weight ofthe sediment was determined , and the necessary amount of0.05 MTris-0.002 M EDTA buffer, pH 7.2, required to effect a20% weight-volume suspension was determined. This wasused for another 30-sec homogenization cycle performed asabove, except the rheostat setting was number 8 (bladespeed was 55,000 rpm). The Tris-EDTA supernatant fluidobtained after centrifugation at 8000 x g for 15 mm described above was centrifuged at 100,000 x g for 60 mm at 4Â°in a Beckman Model LS-65 ultracentrifuge. The 2nd homogenization cycle sediment was retained for basic proteinstudies not reported at this time. The resulting supernatantfluid from the 100,000 x g step is henceforth referred to as100,000 x g soluble and the particulate sediment is termedthe membrane fraction.

Microbiological Analysis The tissue specimens for microbiology were homogenized as a 10% w/v suspension inBSS with the use of a TenBroeck homogenizer. One-tenthml of this solution was then inoculated onto a spectrum ofmicrobiological media including eosin methylene blue agar,5% sheep blood agar (1 incubated aerobic and 1 anaerobic),chocolate agar, potato-dextrose agar containing penicillin(500 units/mI) and streptomycin (200 @g/ml),and Hayflick'sMycoplasma agar plates (14). In addition, 0.1 ml was inoculated into each of 5 roller tubes of African green monkeykidney cell cultures, WI-38 lung fibroblasts, and Hep-2 cellcultures for virus isolation studies. The latter specimenswere â€œblindpassagedâ€•at least twice prior to discard . Routine methods for identification of any microorganisms isolated were used (1).

Tissue Procurement. In addition to tissues obtained atsurgery and autopsy at West Virginia University, 5 breastand 4 colon carcinoma samples were provided by Dr. HenryWeimer at the University of California at Los Angeles. Thesetissues were processed by the same protocol describedabove. Also, Dr. James McCoy of Litton Bionetics (Bethesda, Md .) provided 3 normal lungs. All these tissues werefresh frozen at surgery or autopsy, shipped under dry ice,received within 16 hr of shipping, and stored frozen untiluse.

Ammonium Sulfate Fractionation. The starting materialwas always 100,000 x g soluble, and solid ammonium sulfate was used to prepare 4 fractions (19). Sufficient ammonium sulfate was added to effect a 20% final concentrationand held at 4Â°for 1 hr with constant stirring. The latter wascentrifuged at 10,000 x g for 30 mm at 4Â°.The 20% solublesupernatant fluid was then adjusted to 50% saturation withsolid ammonium sulfate and treated as described above.The 50% soluble supernatant fluid was then adjusted to 80%saturation with ammonium sulfate and treated as above.The 20, 50, and 80% ammonium sulfate precipitates fromeach of the above procedures were dissolved in the smallestpossible volume of PBS, pH 7.2 and dialyzed against severalbuffer changes with the same PBS at 4Â°.The 80% ammonium sulfate supernatant fluid was dialyzed and saved aswell. Again, all extracts were placed in aliquots and storedat â€”80Â°.

3 M KCI Extraction. The procedure utilized was that reported by Meltzer et a!. (27) with minor modifications. The100,000 x g membrane pellets were used instead of intacttumor cells. The pellets were resuspended in 1 ml of DNase(P-L Biochemicals, Inc. , Milwaukee, Wis.) (1 mg/mI) with aDounce glass homogenizer and held at 4Â°for 30 mm. Subsequently, 10 ml of 3 M KCI were added per 0.1 g of 100,000x g membrane pellet and stirred at 4Â° for 16 to 24 hr.

The extract was centrifuged at 40,000 x g for 1 hr at 4Â°.Thesupernatant fluid was desalted by dialysis against severalchanges of PBS for 48 to 72 hr at 4Â°.All extracts were placedin aliquots and stored at â€”80Â°.

Triton X-100 Extraction Protocol. The protocol was modified from that of Razin (31). The 100,000 x g membranepellet was resuspended in 0.05 M Tris-0.1 M NaCI buffer, pH7.4, by glass Dounce homogenization with 100 ml buffer per9 of 100,000 x g membrane pellet. Subsequently, 16 mg ofTriton X-100 (Sigma Chemical Co. , St. Louis, Mo.) per ml ofmembrane suspension were added, and this suspensionwas held with stirring at 37Â°for30 mm. The suspension wascentrifuged at 100,000 x g for 60 mm at 4Â°.The supernatantfluid was dialyzed against PBS for 4 to 6 days and withnumerous changes. In no case was all Triton X-100 removed, even with alternate procedures utilizing single orbundle hollow fibers. However, this residual Triton X-100did not deleteriously affect the immunization or immunoassay protocols.

Lithium Diiodosalicylate Extraction Protocol. This procedure is modified from a protocol developed by Rosai et a!.(32). A 20% weight volume suspension of the 100,000 x gmembrane pellet was prepared by Dounce homogenizationin 0.05 M Tris-0.002 M EDTA-0.25 M sucrose buffer containing 0.1 M lithium chloride, pH 7.5. This suspension wasextracted, with 0.3 M lithium diiodosalicylate (Eastman Kodak, Rochester, N. Y.) in 0.05 M Tris-HCI, pH 7.5, at roomtemperature for 15 mm in a total volume of 2 ml. Subsequently 2 volumes of water were added and the suspensionwas mixed at 4Â°for 20 mm. The latter was centrifugedat 48,000 x g for 30 mm at 4Â°.The supernatant was dialyzedfor 72 hr against PBS, pH 7.2. Extracts were placed inaliquots and stored as described above.

Sonic Disruption. This protocol was modified from onedeveloped by Hollinshead et al. (16) and Davies (6). A Sonipen (Technic International Sonicator, Bergenfield, N. J.)with a 12.00- x 0.15-cm probe was used routinely. The100,000 x g membrane pellet was resuspended in 2 ml ofBSS with the use of a Dounce homogenizer. This suspension was treated sonically fo@I mm in an ice bath behind aprotective shield with the sonic oscillator set at power No. 4and tuning at one-half maximum. The suspension was centrifuged at 48,000 x g at 4Â°for 30 mm, and the supernatantfluid was saved (Sonic Extract 1). The pellet was resuspended as described above in BSS diluted 1:2 in distilledwater, and the protocol was repeated. The resulting supernatant fluid was termed Sonic Extract 2. The pellet from thisstep was resuspended as above, but in BSS diluted 1:4 indistilled water, and the protocol was repeated. The resultingsupernatant fluid was termed Sonic Extract 3. All supernatant extracts were placed in aliquots separately and storedat â€”80Â°.

Immunization Protocol. Four-kg New Zealand White rab

1314 CANCER RESEARCH VOL. 37



Human Lung TAA

bits obtained from the same vendor were used in groups of3 to 5 for immunization with tumor extracts. All rabbits wereprimed at Day 0 with 2.5 ml of Freund's incomplete adjuvantadministered into 4 injection sites: 2 i.m. into each hindthigh and 2 s.c. in the neck region. On Day 10 and at 2-week

intervals thereafter, each rabbit received 200 @gof proteinadsorbed onto 2 ml of the adjuvant Alhydrogel (GallardSchlesinger, Cane Place, N. Y.) for a total of 5 injections.The protein antigens were coupled to Alhydrogel by incubation at pH 6.5 for 1 hr at room temperature. Rabbits wereroutinelybled 5 to 6 days afterthe 5th immunization.

lmmunoassay Protocol. The method of DeCarvalho (7)was used to screen rabbit antisera to human lung tumorextracts. A 0.75% solution of Sea-Kem agarose (MarineColloids, Rockland, Maine) prepared in 1.0 M glycine,0.85% NaCI solution, and 0.1% sodium azide, pH 7.1, wasused to perform all immunodiffusion assays. The antiserumwells were 7 mm in diameter and had a capacity of 80 pi,while the antigen wells were 3 mm in diameter and had acapacity of 30 @l.The distance between the antiserum andantigen wells is 6 mm. This method was used routinely todetect lung TM by using normal serum, plasma, and/or100,000 x g soluble lung extracts as the adsorbent. Themethod was also used to evaluate a number of other solubleadsorbents. In all cases, the soluble adsorbent was used tofill the antiserum well and incubated for 2 hr at 37Â°, followed

by addition of the rabbit antiserum to human lung TM andincubation at 37Â°overnight.

Sephadex G-150 Gel Filtration. A 2.5- x 45-cm columnwas prepared with Sephadex G-150 and a running buffer of0.05 M Tris, pH 8.0, in 2 M NaCI. Ascending chromatographywas performed at room temperature, and the flow rate was3.6 mI/i sq cm/hr. Column eluates were monitored at 280nm, and fractions were collected and assayed for lung TMby the method of DeCarvalho (7).

DEAE-Cellulose Chromatography. Samples were equilibrated by dialysis against the starting buffer of 0.05 M Tris,pH 8.0-0.04 M NaCI. They were then applied to a 1.5- x 25-cm DEAE-cellulose column also equilibrated with the starting buffer. Proteins not bound with the starting conditionswere designated Peak 1. An elution was performed by increasing the NaCI concentration to 0.40 M and this yieldedPeak 2. Resulting peaks were concentrated with the use ofsingle hollow fibers (SHF-36; Biomed Instruments, Inc.,Chicago, III.) and assayed for lung TM by the DeCarvalhomethod (7).

Adsorption Analysis. All soluble antigens, except thosedescribed below, used to investigate specificity of rabbittrivalent antiserum to human lung TM were evaluated bythe DeCarvalho method described above.

Also, all soluble adsorptions were performed with bothunconcentrated and 2- to 10-fold concentrated preparationsto assure further the complete adsorption of possible antibodies to normal tissue components.

A few adsorptions were performed by alternate methods.Human A, B, and 0 erythrocytes were used for adsorbingsuch antisera by mixing equal volumes of washed, packederythrocytes and antisera, incubation at 37Â°for 1 hr, followed by 4Â°for1 hr with occasional mixing. Another methodused to rule out reactivity of our rabbit trivalent antiserumwith normal antigenic constituents of lung tissue utilized

adsorption with soluble and insoluble lung connective tissue prepared in the laboratories of Dr. Robert Burrell , WestVirginia University. One ml of rabbit antiserum was adsorbed with 10 mg, dry weight, of globulin-free, insolublelung connective tissue. This material was prepared according to a previously described method (13) and has beenshown to be composed of collagen and reticulin. A 2nd 1-mIaliquot of antiserum was adsorbed with an equal volume ofa soluble lung connective tissue extract (1.5 mg/mI). Thisantigen was prepared according to the method of Cate andBurrell (4). Although free of serum proteins, its chemicalnature has not been determined. Both adsorptions werecarriedout at37Â°for30 mm.

Adsorptions were also performed using human serumfractions prepared by the ethanol fractionation procedureof Cohn et al. (5) and purchased from Pentex Laboratories,Kankakee, III. These fractions were dissolved in PBS at aconcentration of 10 mg/mI and tested by the method ofDeCarvalho (7).

Identity Analysis. Several commercial antisera specific tohuman serum proteins were obtained from Behring Diagnostics (Sommerville, N. J.). Dr. Charles Todd, City of HopeHospital, Duarte, Calif. , provided purified CEA and highlyspecific precipitating antisera to CEA. Dr. Robert Mclntireofthe National Cancer Institute, Bethesda, Md., provided anAFP source and specific precipitating antisera to AFP.These antisera and our trivalent rabbit antisera as well asgoat monospecific antisera were cross-tested by immunodiffusion against 100,000 x g soluble lung tumor extractscontaining lung TM to rule out their possible cross-reactivity.

Definition of Immunological Reagents. Unless otherwiseindicated in the text, the polyvalent rabbit antiserum utilizedin these studies was raised to human lung TM present in a100,000 x g-soluble and/or 80% ammonium sulfate precipitate thereof. The extract was derived from a squamous cellcarcinoma of the right lung of a 60-year-old male Caucasian(Code Number LU-15). The description of the antiserum(No. 107) and the lung TM follows in â€œResultsâ€•

The monospecific goat antisera to human lung TM wasprepared according to a protocol reported to the AmericanAssociation of Cancer Research in Toronto, Canada, in1976 (25). Briefly, this protocol combines the techniques ofDEAE-cellulose chromatography, 4B Sepharose-dye affinitychromatography, Sephadex G-150 molecular sieve chromatography, and polyacrylamide gel disc electrophoresis. Thefinal purified products were isolated and identified in polyacrylamide disc gel slices and were used to immunize goatsaccording to the immunization schedule outlined above.The immunochemical specificity of these antisera is alsodescribed inâ€œResults.â€•

RESULTS

Identification of Lung TAA. Only microbiologically sterile100,000 x g soluble supernatant extracts of single orpooled normal lungs and/or normal serum and plasma wereused to remove antibodies produced against normal humanproteins.The proteinvaluesof solublenormal undilutedlung extracts ranged from 4.0 to ii .5 mg/mI, whereas

1315MAY 1977



R. W. Veltri et a!.

plasma and serum ranged from 70 to 110 mg/mI.None of the lung tumors utilized demonstrated any micro

bial contamination by test procedures mentioned above.Fig. 1 illustrates a typical immunodiffusion profile demonstratinglung TM. From leftto right,Well 1 demonstrateslung TM after adsorption with pooled normal human serumand 100,000 x g-soluble lung. Three immunoprecipitates,designatedlungTM, 1,2,and 3,going from the precipitatenearest the antiserum well toward the upper 100,000 x gtumor-soluble well, are demonstrable but none below, mdicating removal of normal reactivity. The center antiserumwell is a 0.8% NaCI solution adsorption control and the wellto the right is a tumor-soluble adsorption control. All lungTM reactivity was removed in the latter case. Normal lung100,000 x g soluble supernatants were tested as adsorbentsat protein concentrations of up to 115 mg/mI and gavesimilar results with this test system. Please note that theabove-described trivalent antiserum (No. 107) was used inexperiments which follow unless otherwise stated. Themonospecific quality of our goat antisera is illustrated inFig. 2. The outside wells contain the 3 goat antisera prepared against purified lung TM 1, 2, and 3. The centerantigen well contains Peak 2 from a DEAE-cellulose columnwhich has all 3 lung TM's present. Note the single bandsobtained with each antisera and that they are all immunologically nonidentical.

Column Chromatography of Lung TAA. In order to purifypartially and confirm effectively the immunological noni

dentity of the 3 lung TM's, we utilized a combination ofDEAE-cellulose and Sephadex G-150 chromatography.Thirty-two mg of human lung tumor-soluble protein, equilibrated against the starting buffer by dialysis, were appliedto a 1.5- x 25-cm DEAE-cellulose column. All 3 lung TM'seluted in Peak 2 with 0.4 M NaCI in a total of 9 mg of protein.The latter was concentrated to a volume of 4 ml and appliedto a 2.5- x 45-cm Sephadex G-150 column equilibrated in a0.05 M Tris buffer, pH 8.0, with 2 M NaCI. Chart 1 illustratesthe G-150 column elution profile. Region 1 contains onlylung TM 2, Region 2 contained only lung TM 1 reactivity,while the 3rd region contained some lung TM 1 but mostlylung TM 3 reactivity.

Antigenic Dissimilarity of Lung TAA. The 3 lung TM'spartially purified by DEAE-cellulose and Sephadex G-150chromatography were analyzed by immunodiffusion usingthe trivalent rabbit antiserum to confirm effectively theirimmunological nonidentity. Fig. 3 illustrates an immunodiffusion analysis of a lung tumor 100,000 x g soluble and thepartially purified lung TM. Region 1 (R,) contains only lungTM 2, Region 3 (R3) contains lung TM 1 and 3, whereas,

Region 2 (R2)contains only lung TM 1. Also note the 3 lungTM's in the 100,000 x g tumor soluble and no reactivitywith the normal human lung (NHL) or normal human serum(NHS) antigen-containing welIs@The immunological nonidentity of the 3 lung TM's is readily observed.

Distribution of Lung TAA in Different Types of LungTumor Extracts. Table 1 demonstrates the distribution of

T. sol

â€˜@â€”r-@ ffi4r:@:

@ .

C.NHL

IFig. 1. Detection of 3 lung TAA's by agarose adsorption immunodiffusion. All center wells contain a trivalent rabbit antiserum (No. 107) to tumor 100,000 x

g-soluble extract. Antisera wells were prerun with the adsorbents (left to right) normal human serum (NHS) and normal human lung 100,000 x g-soluble (NHL)extract, 0.85% NaCI solution, and lung tumor 100,000 x g-soluble extract. Upper antigen wells received tumor 100,000 x g-soluble extract, lower antigenwells received normal lung 100,000 x g-soluble extract. When the trivalent rabbit antiserum (No. 107) is adsorbed with NHS and NHL, 3 precipitin bandsremain to the tumor-soluble extract only. These are removed following adsorption with tumor-soluble extract; 0.85% NaCI solution adsorption yields thecomplex pattern shown in the center.

1316 CANCERRESEARCHVOL. 37

T.sol T.sol

NHS+NHL+107

.â€˜0.. NHLNHL



Lung TAA distributionin soluble andmembrane-solubilizedpreparationsPositive

immunodiffu

Totaloflungsion

reactionâ€•Lung

LungLungExtractionmodalityaextracts testedTAA 1 TAA2 TAA3Sonic

Extract 1â€•106 76SonicExtract282 32SonicExtract380 003MKCIr107

46LidV9130Triton

X-100'@96 5420%ASPâ€•1062450%ASP1074780%ASP109

1010100,000x 9 soluble1010 10 10

Human Lung TM

the 3 lung TM's in 9 different extracts of several squamouscell tumors by using the goat monospecific antisera. Theprotein concentration added to each antigen well was approximately 100 to 400 @g.With the exception of the Lidiprotocol, the sonic treatment, 3 M KCI and Triton X-100solubilization protocols gave comparable results, but allwere less reactive than the 100,000 x g-soluble, 50 and 80%ammonium sulfate precipitates. Note, however, the proteinvalues for membrane preparations were at least one-halfthose of soluble extracts, and considerably smaller volumeswere obtained.

Lung TAA Distribution in Several Lung Tumors andPaired Nontumor Sites. Table 2 demonstrates the mcidence of 1, 2, and 3 lung TM in 18 100,000 x g-soluble lungtumor extracts and 13 paired nontumor site 100,000 x gsoluble lung extracts. These 100,000 x g solubles wereanalyzed for the distribution of the 3 lung TM's using high

2

4

Elution Volume(ml)Chart I . Partial purification of lung TAA on Sephadex G-150. The 100,000

x g tumor-soluble extract was separated into 2 fractions on DEAE-cellulose.The 2nd fraction, containing all 3 lung TAA's, was run on Sephadex G-150and yielded the above profile. Regions R-1, R-2, and R-3 were tested byimmunodiffusion with the trivalent rabbit antiserum for lung TAA activity. Theantigens present are indicated at the base of the peaks.

Table 1

titered monospecific goat antisera prepared to purified lungTM (25). Note the uniform occurrence of lung TM 1, 2, and3 in almost all 100,000 x g soluble extracts analyzed. Alsonote the frequency of all 3 lung TM in matched nontumorsites.

Adsorption Analysis. Table 3 illustrates that several adsorbents were tested for ability to remove lung TM reactivity of a rabbit antiserum. The results indicate that lung TM1 and 2 are embryonic antigens, but they are probablyproduced at different times during embryogenesis. Alsolung TM 1 appeared at times to have reduced activitysubsequent to adsorption with normal human serum, lung,or plasma pools. Table 4 is a summary of adsorptions performed with Cohn fractions of human serum. The concentration of all fractions was adjusted to 10 mg/mI. OnlyFractions IV and lV-4 completely removed reactivity of goatantisera to lung TM 1.

Identity Analysis. These testswere performed as outlinedin â€œMaterialsand Methodsâ€•using the trivalent rabbit antisera to lung TM. Our anti-lung TM and antisera to humanserum proteins, CEA and AFP, obtained from commercialand private sources were cross-tested against 100,000 xsoluble extracts. Table 5 indicates neither lung TM 1, 2, or3 are CEA or AFP. However, lung TM 2 cross-reacted withantiserum to human liver ferritin and lung TM 3 crossreacted with antiserum to human lactoferrin. The latter resuits were confirmed with the monospecific goat antisera.

Specificity Analysis. Again, using the monospecific goatantisera and the method of DeCarvalho (7), 100,000 xsoluble tumor extracts from 6 breast, 4 colon, 5 cervical,and 3 head and neck carcinomas, and a Gardner's syndrome were investigated for lung TM 1, 2, and/or 3 reactivity (Table 6). Only lung TM 1 was demonstrable in theextracts of breast, colon, cervix, and head and neck carcinomas, whereas lung TM 2 and 3 appear to be lung specific, at least at the level of sensitivity of this test system.

DISCUSSION

There is convincing biochemical and immunochemicalevidence for the existence of one or more lung TM insoluble extracts of primary lung tumors of several histopathologically different types (10, 12, 20, 26, 36, 40, 42, 43).Further substantiation originates from in vivo (15, 16) and invitro (2, 21, 29, 35) cell-mediated immunological studieswith membrane-bound lung TM solubilized by severalmethods.

We report the isolation and partial characterization of 3serologically distinct soluble antigens designated lung TM1, 2, and 3 from human squamous cell carcinomas with theuse of a trivalent rabbit antiserum and goat monospecificantisera as immunochemical probes. All 3 antigens are sol

uble in SO% ammonium sulfate, as were previously described lung TM's (36, 37, 45). Furthermore, the lung TM'swere demonstrated most frequently in 100,000 x g and 80%ammonium sulfate precipitates thereof. The membrane-solubilized preparations from the sonic treatment, 3 M KCI,Triton X-100, but not the Lidi protocols also yielded the lungTM. The universal distribution of our lung TM in solubleand membrane extracts may relate them to antigens dem

a Method of extraction described in â€œMaterials and Methods.â€•

b Monospecific goat antisera were used.I. Protein concentration of membrane solubles was adjusted to

2.0 mg/mI, andthe ammoniumsulfateprecipitateswereadjustedto5.0 mg/mI.

dA5p ammonium sulfate precipitate.

1317MAY 1977



Lung TAA distribution in lung tumor and matched nontumorousextractsPositive

immunodiffusion reactionforâ€•Lung

TAA 1 Lung TAA 2 Lung TAA3CodeTumor

typeaTS@ NTS TS NTS TSNTSLU-iAdenocarcinoma++ + + ++LU-2Squamous++ + + ++LU-3Squamous+â€” + + ++LU-4Adenocarcinoma

squamous++ + + ++LU-SSquamous+

+ + + ++LU-6Squamous++ + + ++LU-7Squamous++ + + ++LU-8Squamous+â€” + â€” ++LU-9Carcinoidâ€”

NO â€” NO +NOLU-bSquamous+

+ + + ++LU-ilSquamous+NO + NO +NOLU-12Squamous++ + + ++LU-13Squamous++ + + ++LU-14Adenocarcinoma+â€” + â€” +â€”LU-i

5Small cell (oats)+ NO + NO +NOLU-16Squamous+NO + NO +NOLU-i7Squamous+NO + NO +NOLU-i8Adenocarcinoma++ + + + â€”

Adsorption analysisof lung TMantiseraResidual

LungTMActivityâ€•Proteinconcen

trationLungLungLungAdsorbent(mg/mI)TM 1TM 2TM3Human,ORBCNAb

+++Human,ARBCNA+++Human,BRBCNA+++Normal

soluble human lung2.74+++Poolednormal soluble human lung20.0+10++Humanembryonic lung 1 (24 wk)2.400+Humanembryonic lung 2 (26 wk)2.1+++Membrane

extract of normal lunge10.0+++Crudeconnectivetissueofnormallunge1.5+++Normal

human serum pool (Red Cross)Versatolâ€•7.5

+107.0 +10+ +++Acute

phaseserum (tuberculosis)10.0+++Acutephase serum (pyelonephritis)ii .0+++Pooled

normal human saliva0.45+++Poolednormal human plasma7.0 +++

R. W. Ve!tri et a!.

Table 2

a Histopathological diagnosis.

b Monospecific goat antisera were used.rTS, 100,000x g tumor soluble; NTS,100,000x g nontumor soluble, obtained from a remote

lung region not draining the tumor site; NO, no nontumorous site tissue obtained.

Table 3

a Trivalent rabbit antisera to lung TM used.

b NA, not applicable.C Prepared by Dr. Robert Burrell.

dA human serum standard obtained from General Diagnostics, Morris Plains, N.J.

onstrated by in vitro and in vivo cell-mediated techniques (2,15, 16, 21 , 29, 35). An important find in our studies was the

detection of all 3 lung TM's in matched nontumor sites oflungs removed at surgery. Several investigators have in thepast used such tissues as normal control materials and haveobtained somewhat conflicting results (2, 12, 15, 16, 21, 29,35). The positive value of such a find is that the soluble lungTM's 1, 2, and 3 are produced in quantities to be detected

in areas draining the tumor site and hence may be demonstrable in the patient's peripheral circulation.

All 3 lung TM's did not cross-react with CEA or AFP,although lung TM's 1 and 2 were shown to cross-react withan embryonicproteinin24-butnot26-week-oldfetallungextracts. Other researchers have described lung TM notidentifiable with CEA or AFP, but oncofetal in nature (37, 38,42, 43). Mclntire and Sizaret (26) described 2 lung TM's




Adsorption analysis of lung TMantiseraResidual

lung TMactivityâ€•Lung

TM Lung TAA LungTMAdsorbent'1 23Cohn

Fraction I + ++CohnFraction II + ++CohnFraction Ill + ++CohnFraction IV 0 ++CohnFractionV + ++CohnFraction VI + ++CohnFraction IV-i + ++CohnFraction IV-4 0 ++CohnFraction IV-5, 6 + ++@1

Pentex Laboratories, Kankakee,Ill.â€œTrivalent

rabbit antisera to lung TMused.Table

5Identity

analysisIdentity

reactionsâ€•Lung

LungLungAntiserumto TM 1 TM 2 TM3lgG,

IgA, 1gM 0 00al, antitrypsin 0 NDâ€•0a2,macroglobulin 0 00a2,acute phase protein 0 00Haptoglobin

0 00C-reactiveprotein 0 00Transferrin

0 00Ferritin0 +0Lactoferrin0 0+Hemopexin0 00Fibrinogen0 00CEA'0 00AFPâ€•0 0 0

Specificity of lungTMProtein

(mg/mi)Positive

immunodiffusionreactionforâ€•Lung

TM 1LungTM 2LungTM3BR'@i04

BR112BR112NBRil5BR119BR WVU-iCO71CO71NCO75CO81HNCHNCHNCCERWVU-lCERWVU-2CERWVU-3CER WVU-4CERWVU-5GAR2.65

3.302.402.963.256.803.301.601.251.642.402.403.103.762.884.406.205.205.85â€”

+â€”

â€”

+-

+â€”

+++++++++â€”

-â€”

â€”

â€”

â€”

â€”

â€”

â€”

â€”

â€”

â€”

Â±-

â€”

-

-

-

Â±-

-â€”

â€”

â€”

â€”

â€”

â€”

â€”

â€”

â€”

â€”

+-

--

â€”

-

â€”

â€”

-

-

Human Lung TM

Table 4 Table 6

a Monospecific goat antisera were used.

I' BR, breast carcinoma; CO, colon carcinoma; HNC, head and

neck carcinoma; CER, cervical carcinoma; GAR, Gardner's syndrome.

Lung TM 2 cross-reacts with monospecific antisera tohuman liver ferritin, but the antigen was found to possess amore acidic electrophoretic mobility on immunoelectrophoresis and to contain a lung-specific antigenic component byimmunodiffusion (work to be described elsewhere). Thesedifferences manifested between lung TM 2 and liver ferritinmay be similar to these of isoferritins described by Marcus(22) and Marcus and Zinberg (23). Ferritin levels have beenfound to be increased in a number of different neoplasmsand iron storage diseases (9, 17, 22). However, using ourantisera to lung TM 2 (lung ferritin), we were unable todetect the antigen in breast, colon, cervical, or head andneck carcinomas, which may further indicate that the lungTM 2 produced in lung cancer is antigenically distinct fromother normal human organ-specific ferritins. Alternatively,analysis of these samples by more sensitive radioimmunoassays and further immunochemical analysis with otherorgan specific ferritins (liver, spleen, heart) is needed toestablish actual specificity.

Lung TM 3 was shown to cross-react with commercialmonospecific antiserum to human lactoferrin. This is anunusual finding, since lactoferrin is an iron-binding proteinfound in a variety of secretions as well as leukocytes inhumans (24, 39). No evidence exists in the literature forlactoferrin as a tumor marker for any cancer. The value oflung TM 3 is also accentuated by the apparent specificitydemonstrated in lung versus breast, colon, cervical, andhead and neck carcinomas. However, as in the case of lungTM 2, this antigen (lactoferrin) awaits the application ofmore sensitive radioimmuno assay (24) and immunochemical analysis.

We have also prepared polyvalent antisera to 3 additionalsquamous cell lung tumors (LU 2, 7, and 12 from Table 2) in

â€˜IBoth trivalent rabbit and monospecific goat antisera were used

in these experiments.bND, not done.â€˜.Obtained from Dr. Charles Todd, City of Good Hope Hospital,

Duarte, Calif.d Obtained from Dr. Robert Mclntire, National Cancer Institute,

Bethesda, Md.

nonreactive with embryonic extracts, but these results andthose of others may be due to age of fetuses used as in ourstudies. Many of the above-described lung TM's were madequately characterized and lacked specificity. The partialpurification of our 3 lung TM's with DEAE-cellulose' ionexchange and ascending Sephadex G-150 chromatographyfacilitated demonstration of their immunochemical nonidentity and provided a basis for further purification andmolecular characterization (25).

Additional characterization of our lung TM has revealedthat lung TM 1 is a normal serum protein found in CohnFraction IV-4. These fractions are concentrates of poolednormal human serum prepared by sequential ethanol fractionation. Proteins that have been identified in Fraction lV-4include al , antitrypsin, haptoglobin, transferrin, hemopexin , Gc-globulins, thyroxine-binding globulin , cholinesterase, and hypertensinogen. Attempts to identify by immunodiffusion lung TM 1 as one of these proteins usingcommercial monospecific antisera to each proved negative.

1319MAY 1977



R. W. Veltri et al.

rabbits and demonstrated antibodies to lung TM 1, 2, and 3to be present. Furthermore, with such immunological reagents, we can reproducibly monitor purification of all 3lung TM's from tumor, matched nontumor site 100,000 xg-soluble extracts, as well as patient's serum (25).

Several normal serum proteins of nonIung neoplasms (3,8, 9, 18, 28, 33, 34, 44) and lung neoplasms (30) have shownpromise as tumor-associated antigens of possible diagnostic and prognostic significance but most lacking specificity.The numerous TM's reported for a variety of human neoplasms have not been adequately characterized to rule outthe possibility they are normal human serum proteins produced in excess during cancer. However, at least at thelevel of sensitivity of the present immunodiffusion assay,lung TM's 2 and 3 appear to be lung-related TM's, buttheir immunodiagnostic potential is as yet untested. Evidence for their potential immunodiagnostic value is suggested by our demonstration of the lung TM in large quantities in matched lung nontumor sites and by recent reportsof cell-mediated assays (2, 15, 16, 21, 29, 35) and detectionof a circulating antigen by means of cross-over immunoelectrophoresis (41). Current efforts are being directed toward the development of sensitive radioimmunoassays tobe applied to the detection of circulating lung TM in patient's serum or plasma.

ACKNOWLEDGMENTS

We wish to thank Dr. Robert G. Burrell, Dr. Billy E. Kirk, and Dr. David B.Yelton of the Department of Microbiology and the staff of the Departments ofPathology and Surgery for their excellent collaborative efforts. Also, weacknowledge the capable technical assistance rendered by John R. McKolanis, Joanne E. Wright, and Nancy R. Thomas.

REFERENCES

1. Blair, J. E., Lennette, E. H., and Truant, J. P. Manual of Clinical Microbiology. Bethesda, Md.: American Society of Microbiology, 1975.

2. Boddie, A. W., Holmes, E. C., Roth, J. A., and Morton, 0. L. Inhibition ofHuman Leukocyte Migration in Agarose by KCI Extracts of Carcinoma ofthe Lung. Intern. J. Cancer, 15: 823-829, 1975.

3. Carmil, A. Extreme Elevation of Serum Transcobalamin I in Patients withMetastatic Cancer. New EngI. J. Med., 292: 282-284, 1975.

4. Cate, C. C., and Burrell, R. G. Lung Antigen Induced Cell-MediatedImmune Injury in Chronic Respiratory Disease. Am. Rev. Respirat. Diseases, 109: 114â€”123,1974.

5. Cohn,E.J.,Strong,L.E.,Huges,W.L.,Mulford,D.J.,Ashworth,J.N.,Melin, H., and Taylor, H. L. Preparation and Properties of Serum andPlasma Proteins. IV. A System for the Separation into Fractions of theProteins and Lipoprotein Components of Biological Tissues and Fluids.J. Am. Chem. Soc., 68: 459-475, 1946.

6. Davies, D. A. L. Mouse Histocompatability Isoantigens Derived fromNormal and from Tumor Cells. Immunology, 11: 115â€”125,1966.

7. DeCarvalho, S. Detection of Neoantigens in the Serum of Patients withActive Neoplastic Diseases by the Absorption-Immunodiffusion Method.Oncology, 27: 193-234, 1973.

8. Eisen, H. N., Sakato, N. , and Hall, S. J. Myeloma Proteins as TumorSpecific Antigens. Transplant. Proc., 7: 209â€”214,1975.

9. Eshhar, Z., Order, S. E., and Katz, D. H. Ferritin, a Hodgkin's DiseaseAssociated Antigen. Proc. NatI. Acad. Sci. U. S., 71: 3956-3960, 1974.

10. Frost, M. J., Rogers, G. T., and Bagshawe, K. D. Extraction and Preliminary Characterization of Human Bronchogenic Carcinoma Antigen. Brit.J. Cancer, 31: 379-386, 1975.

I 1. Gold, P., and Freedman, S. 0. Specific Carcinoembryonic Antigens ofthe Human Digestive System. J. Exptl. Med., 122: 467-481 , 1965.

12. Granlund, D. J., and Rifts, R. E. Soluble Proteins of Human Bronchogenic Carcinomas. Mayo Clinic Proc., 51: 19-27, 1976.

13. Hagedorn, J. E., Burrell, R., and Andrews, C. E. Immunochemical Analy

sis of Lung-Reactive Antibodies in Human Serum. Am, Rev. Respirat.Diseases, 94: 751-755, 1966.

14. Hayflick, L. The Mycoplasmatales and The L-Phase of Bacteria. NewYork: Appleton-Century-Crofts, 1969.

15. Hollinshead, A. C., Sega, E., Stewart, T.H. M. Ricci, C., and Mineo,T. C.Comparison of Lung Cancer Antigens. Tumori, 61: 125-128, 1975.

16. Hollinshead, A. C., Stewart, T. H. M., and Herberman, R. B. DelayedHypersensitivity Reactions to Soluble Membrane Antigens of HumanMalignant Lung Cells. J. NatI. Cancer Inst., 52: 327-338, 1974.

17. Jacobs, A., and Worwood, M. Ferritin in Serum Clinical and BiochemicalImplications. New EngI. J. Med., 292: 951-956, 1975.

18. Kerbel, A. S., and Davies, A. J. S. The Possible Biological Significance ofFc Receptors on Mammalian Lymphocytes and Tumor Cells. Cell, 3: 105-112, 1974.

19. Korngold, L. The Distribution and Immunochemical Properties of HumanTissue and Tumor Antigens. Ann. N. V. Acad. Sci., 69: 681-697, 1957.

20. Louis, C. J., Blunck, J. M., and Richmond, L. M. Agarose-Gel Electrophoresis of Soluble Proteins from Bronchial Mucosa and BronchogenicCarcinoma. Oncology, 27: 324-335, 1973.

21. Marabella, P. C., Elias, L. L., Takita, H., and Mincwadce, J. MigrationInhibition Test in Lung Cancer Patients. J. Surg. Oncol., 7: 479â€”483,1975.

22. Marcus, D. M. Immunologic Aspects of the Breast. Am. J. Clin. Pathol.,64: 786-791, 1975.

23. Marcus, 0. M. , and Zinberg, N. Isolation of Ferritin from Human Mammary and Pancreactic Carcinomas by Means of Antibody Immunoadsorbents. Arch. Biochem. Biophys., 162: 493-501 , 1974.

24. Masson, P. L., Heremans, M. D., and Schonne, E. Lactoferrin, an IronBinding Protein in Neutrophilic Leukocytes. J. Exptl. Med., 130: 643-658, 1969.

25. Maxim, P. E., Wirtz. G. E., and Westfall, S. Purification of Human LungTumor Associated Antigens (LTAA). Proc. Am. Assoc. Cancer Res. 17:148, 1976.

26. Mclntire, K. A., and Sizaret, P. P. Detection of Antigens Associated withCarcinoma of the Lung. In: P. Bucalossi, U. Veronesi, and N. Cascinelli(eds.), Proceedings of the 1ith International Cancer Congress, Florence,Italy, October 20 to 26, 1974. Vol. 1, p. 172. Amsterdam: ExcerptaMedica, 1975.

27. Meltzer, M. S., Leonard, E. J., Rapp, H. J., and Borsos, T. TumorSpecific Antigen Solubilized by Hypertonic Potassium Chloride. J. NatI.Cancer Inst., 47: 703-709, 1971.

28. Naughton, M. A., Merill, D. A., McManus, L. M., Fink, L. M. Berman, E.,White, J. M., and Martinez-Hernandez, A. Localization of the B Chain ofHuman Chorionic Gonadotropin on Human Tumor Cells and PlacentalCells. Cancer Res., 35: 1887-1890, 1975.

29. Pierce, G. E., and DeVald, B. Microcytotoxicity Assays of Tumor Immunity in Patients with Bronchogenic Carcinoma CorreL ed with ClinicalStatus. Cancer Res. 35: 3577-3584, 1975.

30. Primack, A. The Production of Markers by Bronchogenic Carcinomas: AReview. Seminars Oncol., 1: 235-244, 1974.

31. Razin, S. Reconstitution of Biological Membranes. Biochim. Biophys.Acta, 265: 241-296, 1972.

32. Rosai, J., Tillack, T. W., and Marchesi, V. T. Membrane Antigens ofHuman Colonic Carcinoma and Non-Tumoral Colonic Mucosa: ResultsObtained with a New Isolation Method. Intern. J. Cancer, 10: 357-367,1972.

33. Rosen, S. W. Placental Proteins and Their Subunits as Tumor Markers.Ann. Internal Med. 82: 71-83, 1975.

34. Rosen, S. W., and Weintraub, B. D. Humours, Tumors and Caveats. Ann.Internal Med., 82: 274-276, 1975.

35. Roth, J. A., Holmes, E. C., Bodelie, A. W., and Morton, D. L. LymphocyteResponses of Lung Cancer Patients to Tumor Associated Antigen Measured by Leucine Incorporation. J. Thoracic Cardiovascular Surg., 70:613-618, 1975.

36. Schlipkoter, H. W., Idel, H., Barsoum, A. L., and Vollmer, U. J. AntigensCharacteristic for Tumors in Bronchial Carcinoma (Part 1). Zentr. Bakteriol. Parisitenk. Abt. I, 158: 109-124, 1973.

37. Sega, E., Natali, P. G., Ricci, C., Minneo, C. T., and Citro, G. LungCancer Tumor Associated Antigen Isolation by Gel Filtration and Characterization by Immunodiffusion. I. A. C. S., 2: 1278, 1974.

38. Tatannov, V. S. Detection of Embryospecific Globulin in the Blood Seraof Patientswith Primary LiverTumor. Vopr. Med. Khim., 10: 90-91 , 1964.

39. Tourville, D. R., Adler, R. H., Bienstock, J., and Tomasi, 1 . B. The HumanSecretory Immunoglobulin System : Immunohistological Localization ofy-A, Secretory â€œPieceâ€•and Lactoferrin in Normal Human Tissues. J.Exptl.Med.,129:411-429,1969.

40. Veltri, A. W., Sprinkle, P. M., and Burrell, A. B. Isolation and Identification of Lung Tumor-Associated Antigens (LTAA). Proc. Am. Assoc. CancerRes., 17: 131, 1b76.

41. Viza, D., Louvier, M., Phillips, J., Bouchiex, C. and Guerin, A. A. Solubilization of an Antigen Associated with Certain Bronchial Tumors. European J. Cancer, 11: 765-770, 1975.

1320 CANCERRESEARCHVOL. 37



Human Lung TM

42. Watson, A. D., Smith, A. G., and Levy, J. G. The Use of Immunoadsorbent Columns for the Isolation of Antibodies Specific for Antigens Associated with Human Bronchogenic Carcinoma. Brit. J. Cancer, 29: 183-188,1974.

43. Watson, A. D., Smith, A. G., and Levy, J. G. The Detection by Immunodiffusion of Tumor Associated Antigenic Components in Extracts of HumanBronchogenic Carcinoma. Brit. J. Cancer, 32: 300-309, 1975.

44. Waxman, S., and Gilbert, H. S. A Tumor Related Vitamin B,, BindingProtein in Adolescent Hepatoma. New EngI. J. Med., 289: 1053-1056,1973.

45. Yachi, A., Matsuura, Y., Carpenter, C. M., and Hyde, L. ImmunochemicalStudies on Human Lung Cancer Antigens Soluble in 50% AmmoniumSulfate. J. NatI. Cancer Inst., 40: 663-682, 1968.

1321MAY 1977



Fig. 2. The , , â€”â€˜@.. J are ; goat antiseraare contained in the large outer wells, while a DEAE-cellulose Peak 2 fraction containing lung TM's 1, 2, and 3 is contained in the center antigen well.

Fig. 3. Immunological nonidentity of lung TAAs 1, 2, and 3. A,, R@,and R3 refer to regions obtained from Sephadex G-150 fractionation of lung tumorsoluble extracts (see Chart 1). When each fraction was tested by immunbdiftusion with the trivalent rabbit antiserum, A, contained only lung TAA 2 activity, R3contained lung TM's 1 and 3, and R, contained only lung TAA 1 activity. All 3 antigens are present in the tumor-soluble extract, but no long TAA reactivityoccurs with normal human serum (NHS) and normal human lung (NHL).


A. W. Veltri et a!.



1977;37:1313-1322. Cancer Res Robert W. Veltri, Henry F. Mengoli, Peter E. Maxim, et al. AntigensIsolation and Identification of Human Lung Tumor-associated

Updated version

http://cancerres.aacrjournals.org/content/37/5/1313

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/37/5/1313To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



isolation and identification of human ... - cancer research · a number of relevant reports dealing...

Documents