Int. J. Cancer: 35, 707-7 14 (1985) 0 1985 Alan R. Liss, Inc.

PURIFICATION FROM A HUMAN LUNG CANCER CELL LINE OF A WATER SOLUBLE MOLECULE MEDIATING LEUKOCYTE ADHERENCE INHIBITION FOR PATIENTS WITH LUNG CANCER D.M.P. THOMSON’, Mary SUTHERLAND, Margaret DURKO, Tony DUBOIS, Rosemarie SCANZANO, Nabil LABATEYA and George SHENOUDA Montreal General Hospital, Montreal, Quebec, Canada,

Soluble lung tumor activity as determined by LAl’ was enriched by physiocochemical methods from chemically - defined spent medium of a lung cancer cell line (NCI-H69). To identify the polypeptide carry- ing the antigenic determinant, splenic lymphocytes of BALBk mice were immunized with the enriched iso- late and hybridized with mouse plasmacytoma cells. Eight hybrids were cloned successfully and produced MAbs that immunoprecipitated principally a single chain of M, 40,000 (p40) as well as minor chains of M, 25,000 @25) and M, 13,000 (p13) which were probably degradation products of p40. On ZD gels, p40 was composed of 7 spots with a p l of 6.3 to 7.6, which was not altered by neuraminidase digestion. Affinity chro- matography with MAb anti-p40 absorbed p40 and LA1 activity. The bound and recovered fraction was en- riched for p40 and LA1 activity. Affinity-purified p40 also contained the previously identified p25 and p13 as well as a M, 32,000 peptide (p32). MAb anti-p40 was directed to a common framework determinant on p40 since MAb anti-p40 bound to cancer cells from other organs. The comparatively lung cancer organ-specific determinant recognized by leukocytes from lung can- cer patients was not recognized by the MAb. Affinity- purified p40 triggered LA1 for leukocytes from pa- tients with lung cancer but not for leukocytes from control subjects or patients with colon cancer or ma- lignant melanoma in rigorous blind testing. Cross- reactivity was observed with leukocytes from patients with breast cancer. LA1 activity of affinity-purified p40 seems unlikely to result from an unidentified impur- ity. Thus a p40 molecule has been purified that is expressed on the membranes of lung cancer cells and triggers immunologically-mediated LAI.

Leukocyte adherence inhibition is a reliable in vitro assay (Halliday and Miller, 1972; Thomson et al., 1981a,b; Thomson et al . , 1979; Maluish and Halliday, 1979; Khosravi et al., 1983) that detects immunity to human organ-specific cancer neoantigens (OSNs) (MacFarlane et al . , 1982). Cellular recognition mech- anism of OSN can be mediated by antibody-dependent monocytes (Grosser et al., 1976; Marti et al . , 1976) and/or T cells of either T8 or T4 phenotypes (Shen- ouda et al., 1984; Shenouda and Thomson, 1984; Thomson and Halliday, 1984). Leukocyte nonadher- ence is an active cellular event (Thomson, 1984; Thomson et al. , 1983) and is triggered mostly in by- stander cells by oxidative metabolites of arachidonic acid released from the cells binding OSN (Thomson et al., 1982qb; Thomson, 1984; Thomson and Halliday, 1984).

To identify and purify the tumor cell product which can be recognized immunologically by the tumor host, the enriched OSN from the human lung cancer cell line NCI-H69 was used to produce mouse MAb. The resultant MAb was used in affinity chromatography to deplete and to enrich the OSN isolate of individual proteins, since it was reasoned that the polypeptide

responsible for LA1 activity could thus be identified. In this report, we describe a MAb to an M, 40,000 (p40) polypeptide, the principal protein of the isolate from the lung cancer cell line, that was used in affinity chromatography to enrich the bound fraction for p40 and for LA1 activity for leukocytes from lung cancer patients.

MATERIAL AND METHODS

Production and enrichment of lung cancer OSN The method for producing enriched OSN from the

spent medium of the NCLH69 lung cancer cell line grown in chemically-defined RPMI 1640 medium has been described in detail (Dubois et al . , 1985). OSN from about 80 1 of spent medium was concentrated and then separated by anion exchange, molecular sieve and Blue Sepharose CLdB chromatography (Pharmacia, Montreal, Que.) (Dubois et al., 1985). Also OSN in spent medium during the serum-free cycle from the human colon cancer cell line (HCT-15) was harvested and separated in an identical manner. Mouse immunization and production of hybridoma cell lines

Female BALB/c mice were given an i.p. injection of 100 pg of an enriched lung OSN emulsified in an equal volume of complete Freund’s adjuvant 6 weeks before fusion. Three days before fusion the chosen mice were given an i.p. injection of 50 pg of enriched OSN. Equal numbers of mouse splenocytes were fused with the mouse myeloma line, NS-1, with 50% (wiv) poly- ethylene glycol (M, 1,500) in Dulbecco’s modified Eagle’s medium (DMEM). The cells were grown in selection medium (Kohler and Milstein, 1975). Viable hybrid clones were tested for antibody production by ELISA using the OSN isolate and viable NCI-H69 cells, and for membrane staining using VP 107 plates (V and P, San Diego, CA) (Handley et al., 1982).

‘To whom reprint requests should be sent, at the Montreal General Hospital, 1650 Cedar Avenue, Montreal, Quebec H3G IA4, Canada.

’Abbreviations: BSA, bovine serum albumin; DMEM, Dul- becco’s modified Eagle’s medium; FBS, fetal bovine serum; HPLC, high-pressure liquid chromatography; HTf, human trans- ferrin; LAI, leukocyte adherence inhibition; MAb, monoclonal antibody; M,, relative molecular mass; NAI, nonadherence in- dex; OSN, organ-specific cancer neoantigen; PAGE, polyacry- lamide gel electrophoresis; PBS, phosphate-buffered saline, PH 7.3; SDS, sodium dodecyl sulphate; TPBS, Tween phosphate- buffered saline; i.p., intraperitoneal.

Received: December 3, 1984 and in revised form February 18, 1985.

708 THOMSON ET AL

Appropriate positive cultures were cloned by the lim- iting dilution method (McKearn, 1980) by plating in 96-well plates (Costar, Cambridge, MA). Three fu- sions were performed. Eight hybridoma lines were cloned successfully. All the antibodies discussed here were IgGl except F3-P5C4 which was IgG2,. Magni- fied hybridoma lines were maintained in hybridoma media at 37°C in a 5 % C02 buffered incubator. The hybridoma IgG was purified by means of a protein A- Sepharose column (Pharmacia) when it was derived from supernatants, and was isolated by passage through Blue Sepharose and anion exchange chromatography when it was obtained from ascites. AfJinity chromatography

The IgG enriched from ascites was dialysed against 40 vol of 0.1 M HEPES buffer PH 7.3. A sample containing 90 mg in 50 ml was added to 25 ml of washed Affigel 10 (Bio Rad, Richmond, CA) and ro- tated overnight at 4°C. The gel was washed with PBS ( 0 . 1 4 ~ NaCl, 0 . 1 ~ Na phosphate) resuspended in 2.5 ml IM ethanolamine-HC1 PH 8.0 in 50 ml PBS, incu- bated at 4°C for 1 hr with frequent agitation, decanted, transferred to a column and extensively washed with PBS followed by 0 . 0 5 ~ diethylamine PH 11.5 and returned to PBS with 0.02% sodium azide.

The enriched OSN isolate in 0.5 ml was dialysed against PBS, applied to the PBS-washed column and allowed to bind for I hr at 4°C. The unbound fraction was eluted with 50 ml PBS. The bound fraction was eluted with 100 ml of 0 . 0 5 ~ diethylamine PH 11.5 and neutralized with 30 ml of 1 M Tris-HC1 pH 7.8 as it came off the column. The column was then washed with PBS, and the first 30-ml wash was collected since it also contained bound material. The collected frac- tions were dialyzed, concentrated and tested by ELISA and LAI. A separate affinity column of MAb anti-p40 was prepared for the colon cancer OSN isolate. A control column of the same dimensions was prepared by linking MAb W6/32 purified from ascites. Hybrid- oma HB95 from American Type Culture Collection produced MAb W6/32 to a monomorphic determinant on class I HLA-A,B,C antigens. Membrane staining of live cells with MAb

Cells were harvested from tissue-culture flasks and mononuclear cells were isolated from 30 ml of periph- eral blood by Ficoll-Paque (Shenouda et al . , 1984). Cells (lo6) were distributed into tubes and incubated with hematall, with medium alone, with control super- natants from an IgG-producing hybridoma and with supernatants from the specific hybridomas. After 1 hr the cells were washed with cold hematall and then incubated with a fluorescein-labelled affinity-purified goat antibody to mouse IgG, diluted 1/10, for 1 hr on ice in the dark. The cells were then washed and left in 100 pl of 1% paraformaldehyde in hematall PH 7.2. The cells were analyzed by flow cytometry on a FACS I11 cell sorter for membrane staining. As a control, mononuclear cells were also stained directly with anti- Leu 1. Binding studies by ELISA

Supernatants from the wells containing the hybrido- mas were screened against the enriched OSN in 96- well polystyrene Immulon I1 microtitration plates (Dy- natech, Alexandria, VA), about 100 to 500 ng of pro- tein being used per well in 50 pl of solution in PBS p~

7.3. The protein was fixed to the plates by incubation overnight at 4°C. The MAb anti-p40 from FI-P2C3 hybridoma supernatants was used to monitor isolates from spent medium for p40. A sheep anti-mouse IgG peroxidase conjugate (Cappel, West Chester, PA) di- luted 1/1000 was added as the second antibody. The substrate was 1 mg/ml 0-phenylenediamine (Fisher, Montreal) and 0.012% (v/v) H202 in 0 . 1 ~ citrate buffer PH 4.5. The color change was read at 410 nm on a Microelisa Minireader I1 (Dynatech).

SDS and two-dimensional (20) PAGE SDS PAGE were run by the method of Laemmli

(1970) with a continuous 5 to 20% polyacrylamide gradient. 2D PAGE was run as described by O'Farrell (1975) with a PH gradient of 3.5 to 10 for separation in the first dimension. Radiolabelled samples of OSN were analyzed by 2D PAGE containing unlabelled standard proteins for isoelectric and mol. wt reference points. The gels were stained with Coomassie blue, dried, and then exposed to Kodak X-OMAT AR film between 2 intensifying screens (KYOKKO) at -40°C. Unlabelled samples were silver-stained with a Bio Rad kit according to the method of Merril et al. (1981). Imrnunoprecipitation

A goat anti-mouse IgG antibody (Cappel) was linked to a Kynar 301F (vinylidene fluoride) bead (provided by Roche Diagnostics) as described by Newman (1980). A '251-labelled sample of enriched OSN, con- taining 2 x lo6 cpm, in 200 p~ of TPBS-BSA was cleared with 8 ml of pelleted goat anti-mouse IgG Kynar by incubation overnight at 4°C. Aliquots of cleared antigen containing 120,000 cpm were mixed and incubated with 200 pl of the appropriate hybrid- oma supernatant (MAb) in 2.5-ml tubes overnight at 4°C. To each tube 1.0 ml of goat anti-mouse IgG Kynar was added. After thorough mixing, the tubes were incubated for 15 min at 25 "C and centrifuged at 200g for 5 min. The pellet was washed 5 times in TPBS-BSA and mixed with 30 pl of SDS PAGE sam- ple buffer (Laemmli, 1970) or 2D PAGE lysis buffer (O'Farrell, 1975). The mixtures were boiled and the supernatants recovered and run on SDS gels and 2D PAGE as described above. Monitoring of OSN activity in LAI

The antitumor reactivity of leukocytes was first es- tablished by testing with crude extracts of cancer in the LA1 assay as described previously by Grosser and Thomson (1975). With crude extracts of cancer, NAIs 2 3 0 are positive since NAIs of >95% of control subjects are less than 30 (Grosser and Thomson, 1975; MacFarlane et al . , 1982).

OSN activity of isolates was monitored with the same LA1 assay under slightly modified conditions (Dubois et a l . , 1985): crude control and unrelated cancer extract such as pancreatic cancer were added to both A and B tubes, and coded samples of either the isolate (10 pl) or a control buffer (10 p1) were added to A tubes. Medium 199 and leukocytes were then added to both tubes as in the standard LA1 assay. The final volumes in tubes A and B were 510 and 500 pl, respectively. In addition to leukocytes from lung can- cer patients, leukocytes from patients with colon, breast or bladder cancer or malignant melanoma and from control subjects were also tested. All test samples were

PURIFICATION OF p40 CANCER NEOANTIGEN 709

coded and tested in a random order. The results were expresed as a nonadherence index (NAI):

A - B NAI = - x 100

B

where A equals the number of cells nonadherent in a sample from tubes containing the unknown sample admixed with crude control cancer extract and B equals the number of cells nonadherent in a sample from tubes containing the crude control cancer extract. In this study with soluble antigen isolates, NAIs 230 were considered positive and in addition we determined the statistical difference in NAIs by Student’s independent t-test.

RESULTS

IdentiJication of monoclonal reactivities Viable hybridoma cells that were selected produced

high titers of antibody to give positive binding in ELISA with coupled OSN isolate and membranes of viable NCI-H69 cells, and to give negative binding in ELISA with coupled control proteins, albumin, car- bonic anhydrase and HTf. After 3 successive clonings to ensure pure hybridoma lines, the antigenic reactivi- ties were identified by immunoprecipitation using 1251- labelled OSN isolate.

All 8 hybridoma lines produced MAbs which precip- ated an M, 40,000 protein (p40): F1-P2C3, F1-P2D2, F z - P ~ C ~ , Fi-PiC4, F2-P3h9 F3-PiD4, F3-PiA3, F3- P5C4. Radioimmunoprecipitations of the MAbs, con- trol MAb IgG and MAb anti-class-I HLA antigens are shown in Figure 1. In addition to the heavy band with Mr of 40,000, faint bands with Mrs of 25,000 and of 13,000 were observed with the MAbs to OSN but not with the control MAb or with MAb to class-I HLA antigens (Fig. 1). On 2D gels only p40 was visible, and p40 had microheterogeneity with about 7 spots with a basic PI from 6.3 to 7.6 (Fig. 2). No difference in the radioimmunoprecipitation patterns of the mono- clonals was seen when either hybridoma cell superna- tant or Protein A-chromatography-purified MAbs were used. Cellular distribution of p4O

Viable tissue-cultured cells of lung cancer (NCI- H69), breast cancer (DU 4473, colon cancer (HCT

FIGURE 1 - SDS PAGE analysis of immunoprecipitates with MAb from 8 hybridomas produced to the enriched OSN iso- late and control MAb. Lane A; 1251-enriched lung cancer iso- late. Lanes B,C; control MAb supernatants containing mouse IgG. Lanes D-G; MAb supernatants to human class-I MHC determinants. Lanes H-0; ELISA-positive supernatants from 8 different cloned and expanded hybridomas.

FIGURE 2 - Two dimensional (2D gel electrophoresis of

isolate. Molecular weight and isoelectric points were esti- mated from the mobility of unlabelled standard proteins run with the sample.

MAb anti-p40 immunoprecipitated l2 z -labelled enriched OSN

15) and bladder cancer cells (T24) showed intense fluorescein membrane staining with MAb anti-p40 by FACS 111, whereas peripheral blood mononuclear cells and fibroblast cells were negative. Thus the MAb anti- p40 was directed to a molecule expressed on the mem- branes of parenchymal cancer cells. Distribution of p4O in isolates from spent medium

The distribution of p40 in the isolates by the four- step separation method was examined by ELISA (Ta- ble I). Most p40 did not bind to the Trisacryl ion- exchange column in the low-salt buffer. On the Se- phacryl S200 column, most p40 activity eluted slightly later than an ovalbumin standard (Dubois et al . , 1985). F44 with the highest p40 activity also had the highest LA1 activity, and F25 with the second highest concen- tration of p40 had the second highest LA1 activity. By cation-exchange HPLC as described previously (Du-

TABLE I - DISTRIBUTION OF p40 AND OSN ACTIVITY IN ISOLATES OF SPENT MEDIUM FROM NCI-H69

Fraction OSN

p401 activity’ by LAI

(vdassav) activity by

ELISA

Crude supernatant Trisacry 1 Unbound 0 . 0 1 6 ~ Na acetate 1 .OM Na acetate

Sephacryl S200 F150 F70 F44

F25

Blue Sepharose F44 Unbound

F25 Unbound Bound

Bound Cation exchange HPLC Time 0-7min

7-10 min 10-17 min 17-25 min

0.09-0.6 1

0.60-1.88 0.01-0.08 0.02-0.00

0.09-0.00 0.20-0.14 1.96-2.02 1.932 0.75-1.51 1.152

1.90- 1.96 0.01-0.46 1.72-1.25 0.08-0.03

0.03 0.91 2.09 0.20

10

2 > 10 > 10

2.0 0.5-1.0

0.75-1.5

0.25-0.5

0.5-1.0 Neg

Neg

> 1 1 .o 0.25-0.5 > 1

‘Wells coated with 500 ng proteidwell except where indicated. The upper and lower values from 3 separate preparations of 60-80 1 of spent medium.-*Tested at 125 ng/well.-3The amount of protein/assay required to give a positive LAI (NAI 2 30).

710 THOMSON ET AL.

bois et a/., 1985), most p40 eluted at about 10-17 min in the fraction with the highest LA1 activity (250 to 500 ng/assay).

MAb anti-p40 afinity chromatography MAb anti-p40 (F1-P2C3) affinity chromatography

was used to purify p40 after the third physicochemical separation step. After the first passage through the anti-p40 affinity column, LAI activity and P40 were somewhat depleted. A second passage through the af- finity column was required to remove all p40 detect- able by ELISA (Table II). The material that bound to the column on the first or second passage was re- covered by elution with 0 . 0 5 ~ diethylamine p~ 11.5 and was enriched for p40 (Table 11). The affinity - purified p40 gave a single sharp peak when subjected to molecular-sieve HPLC on a TSK-3000 column (Fig. 3).

TABLE I1 - RESULTS OF ISOLATING p50 BY MAb ANTI-p40 AFFINITY CHROMATOGRAPHY

prep. Protein ELISA4 of MAb

No. Mg (%) control 2:- Procedure ration

BlueSepharose 16 19.7 (100) 0.04 0.74 Unbound' 17 48.0 (100) 0.02 1.37

Affinity chromatography Unbound 16 5.6 (28) 0.02 0.02 (2 cycles) 17 16.1 (34) 0.03 0.026 Bound and 16 4.1 (21) 0.33 1.23 recovered 17 7.2 (15) 0.14: 1.59'

Mol. sieve 16 1.92 0.01 1.305 TSK-HPLC 17 1.03 0.09 1.79 'Sephacryl S200 fraction in the area where ovalbumin elutes; material

was pooled on the basis of ELISA results with MAb anti- 0 and then

mg and amount of protein under peak.-'Applied 1.66 mg and amount of protein under ~eak.-~100 ng proteiniwell unless otherwise indicated.-%O ng prote in /we~~ . -%~ ng protein/weII.

material was passed through a column of Blue Sepharose. - P Applied 2.85

Figure 4a,b,c shows the SDS PAGE analysis of the materials from the affinity column. Figure 4u (lane A) shows the IgG preparation which was subsequently linked to Affigel. The affinity-purified material shows principally p40 (Fig. 4u, Lane D, Fig. 4b, Lane C). In some affinity-purified preparations (Fig. 4a, Lane D), faint bands were present at 29kd and 67kd, pre- sumably representing mouse L chain and albumin which had bled from the material linked to the column. In other affinity-purified preparations, a faint band was observed at 32kd (Fig. 46, Lane C). Affinity-purified p40 was labelled with 1251 and run on SDS PAGE (Fig. 4c). The most intense band was p40. With prolonged overexposure of the film other radiolabelled compo- nents were also detected. Mouse L chain and albumin were detected as radiolabelled bands with mobilities compared to standards of 29 and 67kd. In addition, bands were observed at 32,25 and 13kd. Control MAb and W6132 immunoprecipitated trace amounts of 40kd and 13kd components and moderate amounts of 29kd, and the results can be explained by small amounts of F1-P2C3 MAb in the fraction because of bleeding from the affinity column (Fig. 44. MAb anti-p40 ( F I - P ~ C ~ ) , used to affinity purify the material, immunoprecipi- tated the bands at 40,32,29, and 25kd and the 2 bands at 13kd (Fig. 4c). Another MAb (F3-P1A3), which bound to another determinant of p40, as determined by competitive binding studies, did not precipitate p32 or the lower p13 component (Fig. 4 4 .

1 oc

s

4

7

(I)

3 w-

E C 0 cu (v

Q) 0 C a 0 (I)

9 n a

0

Alb 0 I II

I

L I I I I 0 10 20 30

Time (min) FIGURE 3 - Molecular-sieve profile on TSK-3000 SW (7.5

mm X 60 cm) HPLC of affinity - purified p40. The elution positions of 20 pg/50 pl injection volume of albumin and ovalbumin are indicated. 110 pg/100 pl injection volume of the affinity-purified p40 was applied and the sensitivity ad- justed to include the principal peak. A slight peak is seen in the area where IgG elutes. Column was equilibrated with 0 . 3 ~ NaCI, 0.05 Na phosphate, PH 7.0.

After 2 passages through the affinity column, the unbound fraction lacked the heavy p40 band and the remaining components had intensified. The purified p40 from the TSK-HPLC showed a single component on SDS PAGE after silver staining (Fig. 4b, Lane D), microheterogeneity on 2D gels with a PI of 6.3 to 7.3 (Fig. 5) and after radiolabelling showed $0, p32, p25 and p13 which gave an immunoprecipitation result identical to that shown in Figure 4c. Digestion with neuraminidase of '251-labelled p40 did not shift the position of the spots on 2D gels, suggesting that mi- croheterogeneity may not result from glycosylation of the protein (Fig. 6). On 2D gels, p32, p25 and p13 were not seen, presumably because of the small quantities. LA1 activity of p4O and p4&depleted fractions

Purified p40 and p40-depleted fractions were tested extensively in LAI. Leukocytes from patients with lung cancer showed LA1 to p40 with the maximum response observed at 100 ng (Table 111) although even 10 ngiassay gave a significantly positive LAI. Leuko- cytes from control subjects or patients with colon can-

PURIFICATION OF p40 CANCER NEOANTIGEN 711

FIGURE 4 - (a) SDS PAGE analysis of OSN materials from preparation 16 isolated by MAb anti-p40 affinity chromatog- raphy. Lane A; preparation of MAb IgG isolated from ascites which was linked to Affigel-10 to form the MAb affinity col- umn. Lane B; Applied OSN Fraction 44 unbound from Blue Sepharose. Lane C; Unbound material from MAb anti-p40 affinity column. Lane D; Bound and recovered material which shows trace contamination with mouse L chain and albumin. P40 is the principal isolate. Gel was silver-stained.

(b) SDS PAGE analysis of OSN materials from preparation 17 isolated by MAb anti-p40 affinity chromatography. Lane A; Applied OSN fraction 44 unbound from Blue Sepharose. Lane B; unbound material from MAb affinity column. Lane C; bound and eluted material from MAb affinity column; in addition to p40, the major component, a faint band at about 32kd is visible. Lane D; same material as in Lane C but sub- sequently separated by mol. sieve TSK 3000 SW (7.5 mm X 60 cm). Gel was silver-stained.

(c) SDS PAGE analysis of immunoprecipitates of affinity- purified, '251-labelled p40 (radiogram). Lane A; radiola- belled affinity-purified p40. Lane B; Immunoprecipitation with control MAb. Lane C; Immunoprecipitation with W6/32 (anti-class I HLA-A,B,C). Lane D; Immunoprecipitation with MAb anti-p40 (FI-P2C3), the antibody used for the affinity column. Lane E; Immunoprecipitation with MAb anti-p40 (F3-PIA3) which recognizes another common determinant on p40.

cer or malignant melanoma showed no LA1 with purified p40 (Table III). Leukocytes from patients with breast cancer often gave a positive LA1 to p40. In some instances, leukocytes from patients with bladder cancer gave positive responses to p40. After the first passage through the affinity column, the unbound frac- tion still retained strong LA1 activity at 500 nglassay. After a second passage through the affinity column,

the unbound fraction was LAI-positive at 1.0 pg, and negative at 500 ng/assay (Table 111). A third passage through the affinity column did not further decrease the LA1 activity of the unbound fraction and removed

FIGURE 6 - 2D PAGE of affinity-purified p40, radiola- belled and run without (A) and with (B) digestion with 2U neuraminadase (Sigma). Standard proteins were added to the sample in order to determine the PI and mol. wt of the ra- FIGURE 5 - 2D PAGE of affinity-purified p40 and silver-

stained. diolabelled p40.

7 12 THOMSON ET AL

TABLE U1- ANTIGENICITY BY LAI OF CODED BOUND AND UNBOUND FRACTIONS FROM ANTI-p4O AFFINITY CHROMATOGRAPHY

Leukocyte NAIto donors control/F44 control o.l Control

NAI to antiLp40 affinity chromatography fractions Unbound ( f ig) Bound (rg)

1.0 0.5 1.0 0.5 0.1 0.05 0.01 0.001

Lung cancer 89 & 7 - 3 f 2/ 2 f 3 34f 16f I + -3+3 83f 91f 109k 41f 27f 14f10 (n = 97)' 82 + 24***4 6*** 13 10 14** 1 4 % ~ 11*** lo** 8*

Control 2+4 2f61 - 6 f 7 10f 14+3 NT3 - 2 5 NT O f subjects 2+10 4 4 4 (n = 57)

Other cancers Colon 56f7 PT2 5f 8f -17 NT 1Of NT 7f

(n = 48) 3 12 f 6 4 3 Melanoma 57f PT -13f 7*4 -3f NT - 6 f 6 NT 8f

(n = 28) 13 5 10 4 Bladder 56 +4 - 1 f 51 -453 21f 9f13 NT - 5 f 4 NT 23f

(n = 33) 18f25 15 9* Breast 67k5 PT 34f 14f NT - 1 f 6 NT 36+

(n = 26) 16 16 7** In = Number tested.-*F"T = previously tested (Dubois er al., 1985).-3NT = not te~ted.-~Student's dependent t-test; *p < 0.05; **p < 0.01; ***p

< 0.001 comparing sample to control sample containing medium 199.

less than 10% of the applied protein. Although the unbound fraction was depleted of p40 by ELISA (Ta- ble II), the unbound fraction still had LA1 activity at 1.0 pg/assay with the same specificity as p40 (Table 111). Affinity-purified p40 was subsequently separated by TSK molecular sieve HPLC (Fig. 3) and showed LA1 activity similar to that of the material taken di- rectly from the affinity column (Table IV). When p40 was tested at 1.0 pg, there seemed to be less activity than at 0.1 pg, suggesting a bell-shaped dose-depen- dent LA1 response.

TABLE IV - ANTIGENICITY BY LAI OF AFFINITY-PURIFIED p40 FROM MOL. SIEVE HPLC

NAI to coded samples Leukocyte ",d$ Control Purified NO' ( f ig)

donor extracts 0.5 0.1 0.02 0.002

Lung 82+8 O f 58+ 60+ 21+ 13+ cancer 4 9*3 4* 9** 9

subject 4 5 7

(n = ~ 7 ) ~ Control -6+ -5+ -3+ 1f10 10+1

(n = 19) Other cancers Colon

(n = 28) Melanoma

(n = 14) Bladder

(n = 21) Breast

(n = 24)

78 + 12

78 + 17

55+8

82 + 13

3 f 5

-0 + 7 - 10 t5

- 8 5 7

10f5

NT

8 + 8

NT

13+9

10+3

6 + 5

34 f 8**

2 f 3

NT

NT

NT

'Bound and recovered material from MAb antiLp40 affinity chromatog- raphy was separated by TSK mol. sieve chromatography.-% = Number oftests.-'Student's dependent t-test; *p < 0.001; **p < 0.01.

The organ-specificities of the purified p40 from lung and colon cancer lines were correctly identified in rigorous, blind criss-cross testing with leukocytes from lung or colon cancer patients, respectively (Table V). The lung cancer p40 depleted (unbound) fraction tested at an identical protein concentration did not stimulate positive LAIs. The purified p4Os of lung and colon cancer were also coded so that neither investigator nor technicians were aware of their identity; the techni- cians then used the coded samples to test patients with

lung or colon cancer or benign diseases. After such testing for a week, the technicians correctly declared which sample was p40 lung and which was p40 colon (data not shown).

TABLE V - BLIND CRISS-CROSS TESTING BY LAI OF ANTIGENICITY OF SAMPLES FROM ANTI-p40 AFFINITY COLUMN

NAI to

Colon cancer 3 f 5*3 49 f 17* 9 f 6 - 8 f 10

Lung cancer 50 f 15** 16 + 9** -6 f 13 3 f 12

Malignant 5 f 2 11 + 2 4 f 3 6 f 4

(n = 14)*

(n = 24)

melanoma (n = 6)

'Sam les were tested at 500 ng protein.-% = Number of patients tested- s Student's independent t-test; *p < 0.001; **p < 0.01.

Speci3city of anti-p40 aflnity column It was possible that the LAI-active material was not

p40 but an impurity that had bound non-specifically to the MAb affinity column. LAI-active material was passed through an MAb anti-class-I HLA antigen col- umn (W6/32) composed of an identical amount of IgG (Table VI). The unbound fraction suffered no measur- able loss of LA1 activity. However, the bound fraction did have LA1 activity which was similar to that of the applied material, suggesting that some LAI-active and inactive materials bound equally and nonspecifically to the anti-HLA affinity column (Table VI). By compar- ison, from the anti-p40 affinity column, the unbound fraction lost LA1 activity, whereas the bound and re- covered fraction gained LA1 activity resulting in a 20 - fold enrichment in LA1 activity (Table VI).

DISCUSSION

Soluble lung tumor OSN activity as determined by LA1 was first enriched by physicochemical methods from the chemically-defined spent medium of a tissue- cultured lung cancer cell line, NCLH69. To identify the polypeptide carrying the antigenic determinant rec- ognized by human immune cells, MAb were raised to

PURIFICATION OF p40 CANCER NEOANTIGEN 713

TABLE VI - ENRICHMENT OF OSN ACTIVITY IN LAI BY IMMUNOAFFINITY CHROMATOGRAPHY WITH MONOCLONAL

ANTI-LAO ANTIBODY OR ANTI-CLASS-I HLA ANTIBODY

Total

me(%) Procedure protein Un,tS Sp. Act

Applied material Anti-HLA 14.5 0.25 58,000 4,000

Unbound 10.4 0.25 41,600 4,000

Boundand 0.9 0.5 1,800 2,000

Applied 19.7 0.25 78,800 4,000 material (100)

Anti-p4O Unbound 5.6 1 .0 5,600 1,OOO

(29) Bound and 4.1 0.05 52,000 20,000

( 100)

(72)

recovered (6)

recovered (21) 'The minimum amount of protein required per LA1 assay for a positive

result for leukocytes from patients with lung cancer.

the enriched fraction containing lung OSN activity by somatic cell hybridization (Kohler and Milstein, 1975). Eight hybridomas were cloned successfully, and all identified a protein of Mr 40,000 (p40). Affinity chro- matography with the MAb to the p40 (anti-p40) de- pleted the enriched tumor antigen isolate of both p40 and LA1 activity. The bound and recovered fraction was enriched for both p40 and LA1 activity. From these results, we conclude that p40 carries an organ- specific neoantigen determinant that is recognized by human leukocytes. The possibility that an impurity of the affinity-purified p40 accounted for LA1 activity seems unlikely. First, all proteins in the affinity-puri- fied fraction were precipitated by MAb to p40 except for trace amounts of mouse L chains and albumin. Second, affinity chromatography with MAb to anti- class-I HLA did not enrich the fraction for either p40 or LA1 activity.

Analysis of affinity-purified p40 indicated that other minor components were also isolated and probably represented proteolytic products of p40. Polypeptides of M, 32,000,25,000 and 13,000 were present on SDS gels and were immunoprecipitated by MAb anti-p40 but not by control MAb or MAb W6l32 to human class4 HLA antigens. Affinity-purified p40 showed a sharp single peak on molecular-sieve HPLC. On 2D PAGE, p40 showed microheterogeneity with 7 spots spreading over a PI of about 6.3 to 7.6. The position of the spots was not altered by digesting p40 with neuraminidase, suggesting that the microheterogeneity did not result from glycosylation. P40 was expressed on the membranes of lung, breast, bladder and colon cancer cells but not on peripheral blood mononuclear cells or a fibroblast cell line. Thus, MAb anti-p40 was

directed to a common framework determinant of the p40 molecule, whereas leukocytes from patients with lung cancer recognized an organ-specific determinant

The antigenic activity of affinity-purified p40 was tested extensively and blindly on human leukocytes. P40 triggered a dose-dependent LAI response for leu- kocytes from lung cancer patients but not for leuko- cytes from control subjects or from patients with colon cancer or malignant melanoma. Rigorous, blind criss- cross testing of affinity-purified p40 from lung and colon cancer showed that each expressed an organ- specific determinant, although each was purified by the same MAb anti-p40. Lung cancer p40 showed cross-reactivity with leukocytes from breast cancer patients which is identical to the result observed with crude extracts of cancer (Thomson et al., 1981a,b; MacFarlane et al., 1982; Thomson, 1982~).

Extracts of normal lung tissue do not trigger LA1 for leukocytes from lung cancer patients (Thomson et al., 1981a), suggesting that the OSN epitope of p40 is not expressed by normal lung tissues. Fetal lungs of 14 to 19 weeks express the OSN epitope, but fetal lungs of 20 weeks do not (Thomson, 1982b), suggesting that the OSN epitope of p40 is expressed during early fetal development and again during oncogenesis. It is not known whether p40 molecules without the OSN epi- tope are expressed by normal lung tissue. It is conceiv- able that normal lung tissues might express small amounts of the OSN epitope of p40; however, the amounts are inadequate to sensitize the normal host or to stimulate LA1 when normal lung tissue is used as an extract (Thomson et al., 1981~).

It is not known whether the OSN epitope is defined by a peptide, carbohydrate or lipid, or whether the OSN epitope can exist on proteins coded for by differ- ent genes. Moreover, it is not surprising that we failed to produce mouse MAb to the OSN determinant rec- ognized by human leukocytes inasmuch as mouse MAb to human class-I and -11 MHC antigens, for example, are usually directed to framework determinants, and only rare reports exist concerning the production of MAb to the same determinants as those recognized by specific human alloantisera from women with multiple pregnancies or from hypertransfused individuals. To obtain mouse MAb to the OSN determinant, it may be necessary to define the peptide carrying the OSN anti- genic determinant of p40 and chemically synthesize peptides to induce antibodies which will react with the same cognate sequence in the intact folded protein as do immune human leukocytes.

of p40.

ACKNOWLEDGEMENTS

Supported by a Grant from the National Cancer Institute of Canada. We thank Mrs. Mary Bergin for typing this manuscript.

REFERENCES

DUBOIS, A.E.J., THOMSON, D.M.P., SUTHERLAND, M.E., and GROSSER, N., and THOMSON, D.M.P., Cell-mediated antitumor SCANZANO, R., An isolated enriched organ-specific cancer immunity in breast cancer patients evaluated by antigen-induced neoantigen of human lung cancer for leukocyte adherence inhi- leukocyte adherence inhibition in test tubes. Cancer Res., 35,

GROSSER, N., MARTI, J.H., PROCTOR, J.W., and THOMSON. HALLIDAY. W.J., and MILLER, S., Leukocyte adherence inhibi- D.M.P., Tube leukocyte adherence inhibition assay for the detec- tion: a simple test for cell-mediated tumor immunity and serum tion of anti-tumour immunity. I. Monocyte is the reactive cell. blocking factors. Int. J . Cancer, 9,477-483 (1972). Int. J . Cancer, 18, 39-47 (1976).

bition assays. Cancer Res., (1985). In press. 2571-2579 (1975).

7 14 THOMSON ET AL

HANDLEY, H.H., GLASSY, M.C., CLEVELAND, P.H., and ROYS- TON. I., Development of a rapid microELISA assay for screening hybridoma supernatants for murine monoclonal antibodies. J. immunol. Meth., 54,291-296 (1982). KHOSRAVI, M., LIAO. S-K., THOMSON, D.M.P., and DENT, P.B., Relationship of melanoma-associated antigens to histocompati- bility antigen and p-2 microglobulin in material spontaneously shed by cultured human melanoma cells. Transplantation, 35,

KOHLER. G., and MILSTEIN, C., Continuous cultures of fused cells secreting antibody of predefined specificity. Nature (Lond.),

LAEMMLI, U.K., Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature ((Lond.), 227,680-685, (1970). MACFARLANE. J.K., THOMSON, D.M.P., PHELAN, K., SHENOUDA, G., and SCANZANO, R., Predictive value of tube leukocyte adher- ence inhibition (LAI) assay for breast, colorectal, stomach and pancreatic cancer. Cancer. 49, 1185-1193 (1982). MALUISH, A.E., and HALLIDAY, W.J., Hemocytometer leukocyte adherence inhibition technique. Cancer Res., 39,625-629 (1979). MARTI, J.H., GROSSER, N., and THOMSON. D.M.P., Tube leuko- cyte adherence inhibition assay for the detection of anti-tumour immunity. 11. Monocyte reacts with tumour antigen via cyto- philic anti-tumour antibody. Int. J. Cancer, 18,48-57 (1976). MCKEARN. T.J., Cloning of hybridoma cells by limiting dilution in fluid phase. In: R.H. Kennett, T.J. McKearn and K.B. Bechtol (ed.) Monoclonal antibodies, p. 374, Plenum Press, New York (1980). MERRIL. C.R., GOLDMAN, D., SEDMAN, S.A., and EBERT, M.H., Ultrasensitive stain for proteins in polyacrylamide gels shows regional variation in cerebrospinal fluid proteins. Science, 211, 1437-1438 (1981). NEWMAN. E.S., Preparation of Kynar resin particles as a dis- persed solid phase antibody. Amer. Ass. Cancer Res. 21, p.218 (Abstract 877), (1980). O’FARRELL. P.H., High resolution two-dimensional electropho- resis of proteins. J. biol. Chem., 250,4007-4021 (1975). SHENOUDA. G., THOMSON. D.M.P., and MACFARLANE, J.K. , Re- quirement for autologous cancer extracts and lipoxygenation of arachidonic acid for human T-cell responses in leukocyte adher- ence inhibition and transmembrane potential change assays. Can- cer Res., 44, 1238-1245 (1984). SHENOUDA. G., THOMSON, D.M.P., and MACFARLAND, J.K., Re- quirement for autologous cancer extracts and lipoxygenation of

258-266 (1983).

256,495-497 (1975).

arachidonic acid for human T-cell responses in leukocyte adher- ence inhibition and transmembrane potential change assays. Can- cer Res., 44, 1238-1245 (1984). THOMSON. D.M.P., Influence of tumor burden on host antitumor immunoreactivity. In: S. Haskill (ed.), Tumor immunify in prog- nosis: ihe role of mononuclear cell infiltration, Vol. 18, p. 113- 149, Marcel Dekker, New York (1982~). THOMSON. D.M.P., The LA1 response to human cancer: the biology of the LA1 response and features of the antigens. In: D.M.P. Thomson (ed.), Assessment of immune status by the leukocyte adherence inhibition test, pp. 127-171, Academic Press, New York (19826). THOMSON. D.M.P.. Various authentic chemoattractants media- ting leukocyte adherence inhibition. J. m t . Cancer Znst., 73, 595-605 (1984). THOMSON. D.M.P., AYENI, R.O., MACFARLANE. J.K., TATARYN, D.N., TERRIN, M., SCHRAUFNAGEL, D., WILSON, J., and MULDER, D.S., A coded study of antitumor immunity to human lung cancer assayed by tube leukocyte adherence inhibition. Ann. thor. Surg., 31, 314-321 (1981a). THOMSON, D.M.P., and HALLIDAY, W.J., Recognition of myelin basic protein by cancer patients’ T-lymphocytes in association with class-II MHC antigens on monocytes. Submitted (1984). THOMSON, D.M.P., NEVILLE, A.M., PHELAN, K., SCHWARTZ- SCANZANO, R., and VANDEVOORDE, J.P., Human cancers trans- planted in nude mice retain the expression of their organ-specific neoantigen. Europ. J. Cancer clin. Oncol., 17, 1191-1197 (1981b). THOMSON, D.M.P., PHELAN. K., MORTON, D.G., and BACH, M.K., Armed human monocytes challenged with a sensitizing cancer extract release substances pharmacologically similar to leukotrienes. Int. J . Cancer, 30,299-306 (1982~). THOMSON, D.M.P., PHELAN. K., and SCANZANO, R., Oxidative metabolism, cytoskeletal system, and calcium entry of leuko- cytes in the phenomenon of sensitizing cancer extract-induced leukocyte adherence inhibition. Cancer Res., 43, 1066-1073 (1983). THOMSON, D.M.P., PHELAN, K., SCANZANO, R., and FINK, A., Modulation of antigen-induced leukocyte adherence inhibition by metabolites of arachidonic acid and intracellular nucleotides. Inr. J. Cancer, 30, 311-319 (1982b). THOMSON, D.M.P., TATARYN. D.N., LOPEZ, M., SCHWARTZ, R., and MACFARLANE. J.K., Human tumor-specific immunity as- sayed by a computerized tube leukocyte adherence inhibition. Cancer Res., 39,638-643 (1979).