Immunocytochemical characterization of mouse monoclonal ACTH antibodies with a note on staining...

Histochemistry (1989) 9I : 115-122 Histochemi[stry �9 Springer-Verla~, 1989

Immunocytochemical characterization of mouse monoclonal ACTH antibodies with a note on staining conditions and control procedures

L.-I. Larsson t' 2, L. Scopsi 1 *, D. Medena 3, G. Racchetti 3, and Y.M. Galante 3 1 Unit of Histochemistry, University Institute of Pathological Anatomy, Copenhagen, Denmark 2 Department of Molecular Cell Biology, State Serum Institute, Building 81, Amager Boulevard 80, DK-2300 Copenhagen S, Denmark 3 Section of Biochemistry and Immunology, Recordati S.p.A., Milan, Italy

Accepted July 23, 1988

Summary. Mouse monoclonal antibodies (MAbs) have been produced against porcine ACTH and tested for their immunocytochemical utility. Ten out of 12 MAbs reacted with formaldehyde-fixed human ACTH[1-39] and fragments thereof. Cytochemical fragment testing revealed that 6 of the 10 MAbs recognized epitopes in the vicinity of the region where porcine ACTH differs from mouse ACTH (amino acids 26, 29 and 31). Both tissue and cytochemical model data indicate that many of the MAbs detected porcine ACTH with somewhat higher potency than human and rat ACTH (rat ACTH[1-39] is identical to mouse ACTH[1- 39]). MAbs Nos. 5, 8 and 12, in particular, revealed a highly satisfactory signal to noise ratio also in formalin-fixed, paraffin-embedded specimens. Most of the MAbs were potent in detecting both the high concentrations of ACTH congeners in corticotrophs and melanotrophs and the lower concentrations of such peptides in human antropyloric gastrin cells. Blocking of tissue endogenous peroxidase activity reduced reactivity towards the MAbs. This could be circum- vented by use of biotinylated primary antibodies combined with avidin/streptavidin-alkaline phosphatase detection. Availability of MAbs and of the corresponding synthetic antigen also made some quantitative comparisons and analyses of appropriate control procedures possible.

Introduction

Adrenocorticotropin (ACTH) is produced as part of a large precursor, proopiomelanocortin (POMC) (Eipper and Mains 1980). POMC is expressed in numerous cell types, including anterior pituitary corticotrophs, intermediate pituitary melanotrophs, central neurons, gastrointestinal gastrin cells, cells of the immune system and of the genitourin- ary tract (Chen et al. 1984; Endo et al. 1985; Feurle et al. 1980; Herbert et al. 1982; Krieger et al. 1977; Larsson 1977, 1979, 1981a; Liotta etal. 1979; Lolait et al. 1986; Pintar et al. 1984; Sanchez-Franco etal. 1981; Smith et al. 1986; Smyth 1983 ; Tanaka et al. 1982). The posttranslational pro-

Please send offprint requests to." Professor Lars-Ingc Larsson at the Department of Molecular Cell Biology

* Present address : Dr. Lucio Scopsi, Anatomia Patologica Istituto, Nazionale Tumori, Via G. Venezian, 1, 1-20133 Milano, Italy.

cessing of POMC varies with its site of production and results in i.a. ACTH[1-39] in anterior pituitary and in various modified (acetylated, amidated) ACTH[I-13] components (~MSH forms) and various modified (glycosylated, phosphorylated) ACTH[18-39]/ACTH[18-38] components (CLIP's) in the intermediate pituitary lobe and in most ex- trapituitary sites. Many tumours of both pituitary and ex- trapituitary origin express POMC and process its compo- nent peptides differently. Unfortunately, many currently used antisera are raised to purified ACTH and may therefore react a~so with contaminants present in such prepara- tions. Although it is possible to raise polyclonal antisera to synthetic ACTH peptide fragments, these are not available in sufficient quantity to be useful as standards for clinocopathological studies. Moreover, many of these antisera react poorly or not at all with conventionally fixed and embedded material.

We have produced monoclonal antibodies (MAbs) to porcine ACTH and have evaluated the utility of these for detecting formaldehyde-fixed ACTH in cytochemical models and in tissues from pig, man and rat. Also the region- specificities of the twelve different MAbs have been evaluated using synthetic peptides and a cytochemical model system. Finally, methods for examining the specificity of monoclonal antibody staining of tissue and the use of biotinylated monoclonal antibodies have been explored.

Materials and methods

Production and biotinylation of monoclonal antibodies. Porcine ACTH (Serva, Heidelberg, FRG) was conjugated to bovine serum albumin (BSA) with glutaraldehyde at at molar ratio of ACTH to BSA of 5.6:1, using the procedure of Reichlin (1980). I mg of ACTH, dissolved in 1.3 ml of 0.1 M sodium phosphate buffer (pH 7.0), was mixed with 1.3 ml of the same buffer containing 2.6 mg of BSA and t.3 ml of 21 mM re-distilled glutaraldehyde was added. The mixture was rotated at room temperature for 24 h and was then dialyzed against PBS at 4 ~ C for 48 h. Conjugation efficiency was determined using 125I-ACTH, followed by separa- tion of the products by Sephadex G 50 superfine gel filtration and checking them by sodium dodecylsulphate-polyacrylamide gel elec- trophoresis (SDS-PAGE) (Laemmli 1970), followed by autoradiog- raphy. Under the above conditions, about 60% of added ACTH was conjugated to BSA.

Female Balb/c mice (4-weeks-old) were immunized with BSA- conjugated ACTH, emulsified in an equal volume of Freund's corn-

116

plete adjuvant, following two different protocols. A) Three mice received on day zero 5 pg of ACTH-BSA conjugate in a volume of 50 gl in each hind footpad; on day 13 a second immunization was given using the same amount of conjugate in Freund's incom- plete adjuvant (Racchetti et al. 1987). After 3 days, popliteal lymph nodes were removed and pooled, a cell suspension was prepared and the lymphocytes were fused as described below. B) One mouse received 50 gg of emulsified ACTH-BSA conjugate in 400 pl given intraperitoneally (i.p.) on day zero, followed by booster injections according to the following schedule (Cianfriglia et al. 1983) : 50 gg in 400 pl i.p. on day 8, 5 ~g i.p. on day 14, 2.5 pg in 100 ~tl of PBS i.p. plus 2.5 Ilg in 50 ~1 of PBS intravenously on day 15 and day 16. On day ]7 the spleen was removed, the cells were prepared and fused as in protocol A.

Before fusion, the cells, either obtained from popliteal lymph nodes or from spleen, were forced twice through a 100 gm mesh nylon sheet with a sterile plunger and were washed three times in Hank's medium. The lymphocytes were fused with cells of P3 x 63 myeloma line, brought to log growth phase, following the procedure of Galfr~ et al. (1977), as described (Comitti et al. 1987 ; Rac- chetti et al. 1987).

Primary screening was carried out by enzyme-linked immuno- sorbent assay (ELISA) with porcine ACTH immobilized in the wells of Costar microtiter plates (by incubating 100 gl/well of a solution of 2 gg ACTH/ml in PBS at 4 ~ C overnight), saturated with 200 gl 1% horse serum in PBS. 50 pl of cell culture superna- tant were added, the plate was incubated for 2 h, 100 gl of a biotinylated rabbit anti-mouse IgG (H+ L) serum in dilution buffer (0.1% BSA in PBS+0.05% Tween 20+0.05% NAN3) were added, followed, after 1 h, by a 15 rain incubation with 100 gl of avidin- alkaline phosphatase conjugate in dilution buffer. The reaction was developed with 100 gl of a solution of 1 mg/ml ofp-nitrophenyl phosphate in diethanolamine buffer (pH 9.8) for 30 min and mea- sured at 405 nm in a Titertek Multiscan MC ELISA reader (Flow). After each step, the plate was washed 5 times with washing buffer (PBS containing 0.2% Tween 20 and 0.05% NAN3) in a Microplate automatic washer (Flow). All incubations were done at 37 ~ C.

Antibody haplotype was established by a second screening either with a goat anti-mouse IgG (Fc) or a goat anti-mouse IgM (mu), both biotin-labelled, followed by incubation with an avidin- alkaline phosphatase conjugate as above. Only hybridomas secreting immunoglobulins of G class were further considered. Positive hybrids were re-tested by ELISA to determine their specificity against whole human ACTH[1-39] and two of its fragments ([1-24] and [18-39]), immobilized on microtiter plate wells, as described above.

Three times cloned hybridomas were i.p. injected into pristane- primed Balb/c mice. After collection of the ascites fluid, the MAbs were purified by two precipitations at 50% ammonium sulfate satu- ration and extensively dialysed against PBS containing 0.05% sodium azide (PBS/NaNa). All manipulations were carried out at 4 ~ C.

The purity of each MAb was checked by SDS-PAGE in a 12.5% acrylamide slab gel according to Laemmli (1970). The purified MAbs were stored at - 8 0 ~ in 0.3 ml aliquots. They were biotinylated with SP-biotin (Jackson Immune Research Laborato- ries) at 1 mg/ml in borate buffer (pH 8.5) and 0.2 mg SP-biotin/ml for 4 h at room temperature, followed by extensive dialysis against PBS/NaN3. Reactivity of the biotinAabelled monoclonals was checked by solid phase ELISA on immobilized human ACTH in a direct assay with an avidin-alkaline phosphatase conjugate and by an indirect assay using a second antibody conjugated to horseradish peroxidase.

Paper eytochemical models. Different synthetic human ACTH fragments, purchased from Peninsula Laboratories or kindly donated from Gideon Richter Inc. were spotted on strips of Whatman No. 1 filter papers. Peptides used included ACTH[1-14], ACTH[I-24], ACTH[1-32], ACTH[18-39] and ACTH[34-39]. In addition, synthetic human ACTH[1-39] and a purified porcine ACTH preparation (corticotrophin peptide) was analyzed. For testing fragment reactivity of the MAbs, 2 gl droplets, containing 100 ng of the

peptides, dissolved in 5 m M HC1 or in redistilled water, were applied. Following drying the paper strips were fixed in paraformaldehyde vapours at 80 ~ C for 1 h as described (Larsson 1981b). The strips were then soaked for 30 min in 1% HSA in TBS and exposed to the different MAbs at 5 pg/ml, diluted in TBS, containing 0.25% BSA and 2 mg/ml of low molecular weight poly-L-lysine (PLL, Sigma) for 1 h at room temperature (Scopsi et al. 1986). After washing 3 x 10 rain in TBS (containing 1% triton X-100) the strips were exposed for 30 min at room temperature to peroxidase-Iabelled rabbit anit-mouse IgG (Dakopatts A/S, Copenhagen, Denmark), diluted 1 : 100 in TBS, containing 0.25% BSA and 2 nag/ ml PLL. Following three 10 minute washes in TBS-triton, the strips were developed in 3,Y-diaminobenzidine-H20~ medium (Graham and Karnovsky 1966).

Tissue material. Human foetal pituitary material was obtained at induced abortions at 12-28 weeks of gestation. Pig pituitaries were obtained at a local abattoir and rats were decapitated under light diethyl ether anaesthesia. In addition, samples from human antropyloric mucosa were obtained at surgery for gastric ulcer or cancer. Part of the material was immediately frozen in melting Freon-22, freeze-dried, fixed in paraformaldehyde vapours at 80 ~ C for 1 h and embedded in paraffin in vacuo (Bj6rklund et al. 1972; Larsson 1983), whereas part was fixed by immersion in 10% commercial formalin (corresponding to 4% paraformaldehyde) in 0.1 M sodium phosphate buffer pH 7.4 or in Bouin's fluid, containing 5% acetic acid and routinely processed for paraffin embedding.

Immunoeytochemical staining. Tissue sections were deparaffinized and hydrated and were occasionally exposed to methanol-H20 z (Larsson 1983) during the hydration sequence to block endogenous peroxidase activity. The hydrated sections were soaked in TBS for 5 min, exposed to 1% human serum albumin in TBS for 30 min at room temperature and then exposed to the various MAbs (Nos. 1-5, 7-13) diluted to contain 1-25 I-tg IgG per ml in TBS, containing 0.25% BSA. After incubation for either 1 h at room temperature or for 20 h at 4~ plus 1 h at room temperature, sections were rinsed in TBS-triton X-100 (3 x 15 rain) and exposed to peroxidase-conjugated rabbit antimouse IgG diluted 1:30 for 30 min at room temperature. Subsequently, sections were rinsed in TBS-triton (3 • 15 min) and submitted to 3,3'-diaminobenzidine- HzOz development for 8 rain, rinsed in distilled water, lightly coun- terstained with haematoxylin, dehydrated, cleared and mounted.

In one set of experiments, sections were stained with MAb No. 5, labelled with biotin, diluted to 5 or 15 ~tg IgG per ml as above and applied for 20 h at 4 ~ C + 1 h at room temperature. Subsequently, sections were washed in TBS (3 x 5 min), exposed to 0.1% triton X-100 in PBS (0.13 M NaC1 in 0.05 M sodium phosphate buffer pH 7.4), washed in PBS for 5 rain and exposed to an avidin-alkaline phosphatase conjugate (Dakopatts A/S, Co- penhagen, Denmark) or to a streptavidin-alkaline phosphatase complex (Detek-l-alk, Enzo Biochemicals, New York, NY) prepared and diluted as indicated by the manufacturers, using 0.5 m M EDTA, 0.5% triton X-100 and 0.1% BSA in PBS. After incubation at 37 ~ C for 30 or 60 rain, slides were rinsed as recommended by Enzo (10 m M KPO, pH 6.5, 0.5 M NaC1, 0.5% triton X-100, 1.0 m M EDTA, 0.1% BSA 4 x 5 rain), soaked in 100 m M Tris-HC1 buffer pH 8.8, containing 100 mM NaCI and 5 m M MgCI~ 2 x 2 rain and incubated in a developer containing 5-bromo-4- chloro-3-indolyl phosphate (BCIP) as substrate and nitroblue tetra- zolium (NBT) (Leary et al. 1983). Levamisole (Sigma) was added to the developer to inhibit endogenous (non-intestinal) alkaline phosphatase activity. Stock solutions of BCIP (50 mg/ml) in anhydrous dimethylformamide and of NBT (75 mg in 0.7 ml anhydrous dimethylformamide, to which 0.3 ml distilled water was added) were kept in the dark at 4 ~ C for, at most, 2 weeks. The developer was made up by adding 70 gl of the NBT and 10 gl of the BCIP stock solutions to 5 ml 100mM Tris-HC1 pH 8.8, containing 100 m M NaC1 and 5 m M MgClz. The development was controlled in the microscope and was terminated by rinsing in water after 10-15 min.

117

Control procedures. Controls used included conventional staining controls as well as absorptions of the antibodies with the appropriate synthetic antigen. In addition, substitution of MAbs with matched control IgG was performed. Three of the MAbs were kindly isotyped for us by Drs. T. Bech and O.D. Madsen (Hage- dorn Research Laboratories, Gentofte, Denmark), using an isotype kit (indirect immunoperoxidase ELISA) fi'om Zymed Laboratories. Isotyping revealed MAbs No. 5 and 11 to be of the IgGl subtype (kappa light chains) and No. 8 to be lgG2b (kappa light chains).

Table I. Specificity of the monoclonal antibodies towards synthetic ACTH fragments as tested on formaldehyde vapour-fixed paper cytochemical models

Monoclonal Peptide fragment antibody No.

ACTH ACTH ACTH ACTH ACTH [1-14] [1 241 [1-321 [18 39] [34-39]

1 High performance liquid chromatography (PHLC). The purity of 2 the synthetic human ACTH peptides and of commercial purified 3 porcine ACTH was controlled using a Waters high pressure liquid 4 chromatograph equipped with two model 510 pumps, a U6K injec- 5 tor, a model 441 UV absorption detecter (224 nm), a model 730 7 data module and a model 720 system controller. Radial-pak gBon- 8 dapak C~8 columns (8 mm I.D.) were used in a Z module for 9 reverse phase chromatography and were eluted with a linear gra- 10 dient of 15% 80% acetonitrile (Rathburn, Scotland) in 0.08% tri- 11 fluoroacetic acid (Merck) in glass-redistilled water over 45 min at 12 3 ml/min. ] 3

+ + -- _ + + -- _

- - + + - -

- - - - -1_ _

+ + -- _

_ _ + +

_ _ -1_ _ _

R e s u l t s a n d c o m m e n t s

Monoclonal antibodies

Forty-e ight hybr idomas secreting immunoglobul ins (of IgG haplotype) reacting with human ACTH[1-39] were obtained using immunizat ion pro tocol A, while only thirteen were obtained using pro tocol B. Fur the r screenings revealed that only 3 hybr idomas secreted antibodies reactive with ACTH[I -24] .

Cytochemieal models

Twelve M A b s were tested for reactivity against formaldehyde-fixed A C T H fragments in paper cytochemical models. Most of the M A b s were directed towards the middle region of the A C T H molecule and failed to react with either N- terminal (ACTH[I -14] or extreme C-terminal (ACTH[34- 39]) fragments (Table 1). One single M A b (No. 11), however, reacted strongly with synthetic ACTH[34-39]. Two M A b s (Nos. 1 and 13) completely failed to react with any of the A C T H fragments tested. These two MAbs also failed

to react with synthetic human ACTH[1 39] in concentrations up to i00 ng in the formaldehyde-f ixed models. As these M A b s were found to react with (unfixed) A C T H [ I - 39] in the E L I S A system, it is likely that the fixation des- t royed the epitopes recognized by them. No less than seven M A b s (Nos. 4, 5, 7, 8, 10, 1 l , 12) reacted with synthetic human ACTH[18-39] (CLIP). In addit ion, three M A b s (Nos. 2, 3 and 9) reacted with synthetic ACTH[1 24], but not with ACTH[18-39] or ACTH[ l -14] . The lat ter three M A b s also reacted with ACTH[ l -32 ] , which, in addit ion, was also detected by M A b No. 7. In Fig. 1, the different M A b s have been assigned a tentative "region-specif ic i ty" . As porcine A C T H had been used as antigen, we were cur- ious to see whether the M A b s were more potent in revealing porcine than human ACTH. As synthetic porcine A C T H [ I - 39] was not available, we used purified porcine ACTH. HPLC analysis of this p repara t ion revealed it to contain very numerous pept ide peaks, the p redominant of which eluted with the expected retention t ime o f ACTH[1-39] . Use of this prepara t ion at its nominal weight would, hence,

MAB If 11

" / f 4 , 5 , 8 , I 0 , 1 2

Fig. 1. Amino acid sequence of human ACTH[I-39] with the preliminarily assigned region-specificities of the monoclonal antibodies as based on the results shown in Table 1. At least four different epitopes, all localized in the C-terminal 25 amino acid residues, are detected. The epitope recognized by MAbs No. 2, 3 and 9 most likely includes the tryptic cleavage site from Lys 15 to Arg aS. This site is cleaved during processing of ACTH to ~MSH (ACTH[t-13] derivatives) and CLIP (ACTH[18-39] and ACTH[18-38] derivatives). This basic region of the peptide would also be expected to be much modified by formaldehyde. In pig ACTH, Ser 3~ is replaced by Leu ~1, while in mouse and rat ACTH, Gly 26 and Asp 29 are replaced with Va126 and Asn 29, respectively

118

Fig. 2a and b. Sections of rat pituitary (freeze-dried, paraformaldehyde-vapour fixation) stained with MAb No. 5 and peroxidase-labelled antirnouse IgG (a and h). Note strong staining of both anterior lobe corticotrophs and of intermediate lobe melanotrophs. Note also pseudoperoxidase activity in erythrocytes, a x 75, b x 230

underestimate the true levels of ACTH[1-39] contained in it. Nevertheless, when equal amounts (ng) of synthetic human ACTH[1-39] and purified porcine A C T H were spotted it was found that both MAbs No. 5 and 10 were more potent in detecting the porcine (detection limit in nominal weight: 12.5 ng for antibody No. 5, 25 ng for antibody

No. 10) than the human (detection limit 100 ng for both antibodies) peptide. As porcine and human ACTH[1-39] differ by only one amino acid (serine 31 in the human is replaced with a leucine residue in the porcine hormone), it is possible that this residue affects the binding of these two antibodies.

T a b l e 2. Reactivity of the monoclonal antibodies with anterior pituitary tissue from pig, rat and man and with human antropyloric gastrin cells

Monoclonal Tissue antibody No.

Pig Rat Human Human pituitary pituitary pituitary antropyloric

G ceils

1 + (+) (+) -

2 + + + + - 3 + + + + - 4 + + + + + + + + + + + + 5 + + + + + + + + + + + + 7 + + + + + 8 + + + + + + + + + + + + 9 + + + -

t0 + + + + + + + + 11 + + + + + + + + -

12 + + + + + + + + + + + + 13 . . . .

Tissue s taining e x p e r i m e n t s

Except for MAbs Nos. l and 13, all antibodies stained cells in human, pig and rat pituitaries (Fig. 2). Optimal staining results were obtained in freeze-dried, paraformaldehyde vapour-fixed and paraffin-embedded material as well as in formalin-fixed paraffin-embedded material. Bouin's fixative followed by paraffin-embedding produced much inferior results with most of the antibodies. Blocking of endogenous peroxidase activity with methanol-H202 resulted in losses of staining with most MAbs and was therefore abandoned. Several of the MAbs produced clearly stronger staining of porcine than of human pituitary cells. This may, as dis- cussed under cytochemical models, reflect a better reactivity of the MAbs with porcine ACTH but may, in tissue sections, equally well be ascribed to differences in the post mortem time and condition of the specimens, Nevertheless, several of the MAbs were useful for identifying human foetal ACTH cells (Table 2). MAbs Nos. 4, 5, 8 and 12 produced the strongest staining of human anterior lobe material, while Nos. 10 and 11 produced somewhat weaker, but still useful, reactions. The same general pattern was also observed in the porcine and rat pituitary specimens. No clear-cut differences were noted between the staining pat- terns of corticotrophs and melanotrophs. The optimal concentrations of MAbs were between 5 and 15 gg/ml and incubation at 4 ~ C overnight produced best results. At concentrations above 20 gg/ml background staining became troublesome, particularly with MAb No. 4.

As pituitary cells contain quite high concentrations of ACTH congeners we supplemented our studies by staining experimens of human antropyloric mucosa. In this location, gastrin cells have been shown to contain ACTH immunoreactants in lower concentrations than pituitary cells (Lars- son 1979, 1981 a). Staining experiments performed on paraffin sections of specimens either fixed in aqueous formalin or freeze-dried and paraformaldehyde vapour-fixed, revealed strong staining of gastrin cells with MAbs Nos. 4, 5, 8 and 12 and weaker, but distinct, staining with Nos. 7 and 10 (Fig. 3). Double staining experiments have shown the reactive cells to represent gastrin cells (Larsson: to be

119

published). Thus, most of the MAbs that produced strong staining of pituitary ACTH and MSH cells were also useful for revealing the lower concentrations of ACTH occurring in human gastrin cells. However, MAb No. 11 formed an exception in that it was unable to detect gastrin cells in any of the three human antra examined.

As blocking of endogenous peroxidase activity des- troyed much of the immunoreactivity recognized by the MAbs, we also tried to use biotinylated MAb No. 5 for avidin-alkaline phosphatase detection. Good results were obtained in pituitaries and antropyloric tissue using the biotinylated antibody at 5 gg/ml and the commercial avidin- alkaline phosphatase or streptavidin-alkaline phosphatase reagents, diluted twice that recommended by the manufacturers. Development time in the BCIP-NBT mixture was optimal at 5-10 rain. The BCIP and NBT stock solutions were renewed at least every second week. In the paraffin embedded tissue specimens, endogenous alkaline phosphatase activity never presented any problem with the short development used.

Controls

All conventional staining controls were negative and, hene, served to exclude the presence of interfering endogenous peroxidase (or alkaline phosphatase) in the cells stained by the monoclonal antibodies and also excluded unspecific binding of the second peroxidase-labelled anti-mouse IgG antibody to the tissues (Larsson 1983; Sternberger 1979). The remaining controls included absorption controls, using purifed or synthetic antigens, as well as substitutions of the primary antibodies (Nos. 5, 8 and 11) for isotype- matched MAbs (IgGl and IgG2b; kappa chain) at the same and 3- and 5-fold higher concentrations. Finally, tissue staining experiments were also performed with the MAbs preabsorbed with 2 mg/ml low molecular weight poly-L- lysine. Absorption controls, using purified porcine ACTH resulted in extremely high background staining of the material, probably reflecting binding of antibody-bound antigen to most tissue components. Therefore, absorptions were carried out with synthetic ACTH fragments. MAb No. 5 (at 5 gg/ml) was absorbed with 200 gg/ml of synthetic human ACTH[I 8 39]. Such preabsorbed antibody completely failed to react with antropyloric gastrin cells (Fig. 3), but still produced a weak staining of ACTH and MSH cells of the pituitary glands. In contrast, absorption of this MAb with 2 mg/ml poly-L-lysine did not change the staining pattern obtained, but evoked some decrease in the background staining. Finally, substitution with isotype-matched irrele- vant MAbs failed to produce any staining in either pituitary or antropyloric gland area.

D i s c u s s i o n

Our results show that several mouse MAbs directed against ACTH are useful for immunocytochemistry. Comparisons between human and porcine tissue specimens, indicate that, in general, the MAbs tended to produce stronger staining of porcine than of human material. As this result could be due to differences in ACTH content in the different species, differences in post mortem time and in handling of the material, we sought to corroborate our notion by quantitative comparisons in (formaldehyde-fixed) cytochemical models. As synthetic porcine ACTH[1-39] is not

120

Fig. 3a and b. Serial sections through human antropyloric mucosa stained with MAb No. 5 (a) or with CLIP-preabsorbed MAb No. 5 (b) and peroxidase-labelled antimouse IgG. Note strong reactivity in scattered epithelial endocrine cells and absence of staining in the absorption control. Pseudoperoxidase activity in erythrocytes is evident in both sections, x 307

available, we were forced to use a purified porcine ACTH preparation for this comparison. HPLC analyses docu- mented that the predominant peptide contained in this material eluted with the retention time of ACTH[1 39], but that it also contained several impurities. Therefore, the concentrations given represent over-estimates of the true ACTH[1-39] concentrations. Our comparisons, subject to these reservations, showed that MAb No. 5 was about 8-fold more potent in detecting porcine than human ACTH and that the corresponding figure for MAb No. 10 was 4-fold. The human and porcine ACTH[l-39] molecules differ in only one amino acid (position 31) and both MAbs were found to react with sites, close to or encompassing, this amino acid residue. In mouse, rat and human ACTH[1- 39], serine occurs as amino acid No. 31, whereas, in pig, a leucine replaces this amino acid. It is possible that the mice have preferentially recognized this as a foreign region

of the immunogen, Differences between mouse and pig ACTH[1-39] include, in addition, amino acids 26 and 29 and it is striking to note that the majority of the MAbs produced appear to be directed towards this region. Of all MAbs, only three (Nos. 2, 3 and 9), appear to react with the ACTH[1-24] region, which is invariable in mouse, rat, human, pig and cow.

The paper cytochemical models were fixed by formaldehyde vapours, which serve the dual purpose of immobilizing the peptides and to fix them in the same way as occurs in freeze-dried formaldehyde vapour fixed tissue specimens. As basic peptides may bind aggregated or labelled IgG un- specifically, models were stained after pretreating all im- munereagents with 2 mg/ml low molecular weight poly-L- lysine, which has been shown to eliminate this unspecific binding (Scopsi et al. 1986). By the fragment testings, a preliminary assignment of the region-specificity of the

121

MAbs towards formaldehyde-fixed ACTH could be derived. Only two of the MAbs (Nos. 1 and 13) failed to react with the formaldehyde-fixed antigen. The specificities of the remaining MAbs are illustrated in Fig. 1. It should be noted, however, that further fragment testings will be necessary to exactly pin-point the reactive amino acid residues.

Our staining experiments also indicated that many of the MAbs were potent in demonstrating ACTH immunoreactivity in both anterior pituitary corticotrophs, intermediate pituitary melanotrophs and antropyloric gastrin cells. The processing of the POMC precursor differs in these cells, resulting in the predominant production of ACTH[1-39] in the anterior lobe and of c~MSH and CLIP-type components in the intermediate lobe and in gastrin cells. The MAbs that produced strongest staining in the pituitary also produced strongest staining in the gastrin cells. These cells have been found to contain lower levels of ACTH components than the pituitary cells and were not detected by some of the more weakly reactive MAbs (Nos. 2, 3 and 9). It is likely that the explanation to this is that these MAbs are not sensitive enough to detect lower ACTH concentrations. However, only small amounts of ACTH[1-39] occur in antropyloric mucosa (about 2-3 ng/g wet weight), whereas much larger amounts of ~MSH- and CLIP-derived components occur (Larsson: to be published). Therefore, as MAbs Nos. 2, 3 and 9 recognize an epitope around amino acids 14-18, which is cleaved during formation of ~MSH and CLIP components, they may also be compromised in their reactivity to gastrin cells. Interestingly, MAb No. 11, which shows a medium to strong reactivity in pituitary material, completely failed to detect gastrin cells in any of the three human antra tested. This MAb is specific to the ACTH[34-39] region and, hence, detects ACTH[18-39]. Formation of underivafized or derivatized (phosphorylated and/or glycosylated) ACTH[I 8-38] and ACTH[I 8-39] have been described in rat neurointermediate lobe (Bennett etal. 1982). It is possible that MAb No. 11 depends upon the presence of amino acid No. 39 for reaction and further studies on the presence of ACTH[18-38] and ACTH[18-39] components in human antra are therefore needed.

In terms of general usefulness and signal to noise ratio MAbs Nos. 5, 8 and 12 gave the best results. Also No. 4 produced strong staining, but gave higher background. All of these MAbs produced good results on routinely fixed, paraffin-embedded specimens. This, together with the fact that they react both with human ACTH[1-39] and its processed forms (ACTH[18-39]/[18 38]), would make them useful for clinicopathological studies. It should be kept in mind, however, that these antibodies will detect both corti- coidogenically active ACTH[]-39] and the corticoidogeni- cally inactive C-terminal fragments.

One problem associated with these antibodies is that blocking of endogenous peroxidase activity by methanol- H202 pretreatment severely compromises their reactivity with tissue. One way of circumventing this is to employ alkaline phosphatase as a reporter molecule in lieu of peroxidase. In most of the paraffin embedded tissues examined with our biotin-avidin-alkaline phosphatase technique endogenous alkaline phosphatase activi~;y presented no problem. However, in cryostat-sectioned material addition of levamisole to the developer was desirable to block the endogenous non-intestinal alkaline phosphatase activity (un- published data).

An interesting aspect of monoclonal antibody immunocytochemistry relates to the use of controls. Standard stain- nig controls (i.e. the sequential deletion of the different antibody layers) constitute useful means for excluding unspecific staining due to e.g. endogenous enzyme activity or unwanted binding of secondary reagents (like avidin, second antibodies) to tissue components. However, in addition to these controls, it is necessary to establish that the primary antibodies react with tissue by their antigen-combining sites and not by other regions of the IgG molecule (e.g. the Fc region). This can be accomplished either by adding antigen to block the antigen-combining site of the monoclonal antibody (liquid-phase absorption) or by use of isotype- matched control IgG. Use of solid-phase antigen absorptions would be useless as these would remove all antibody protein from the staining solution. Use of liquid phase absorptions were unsuccesful when purified porcine ACTH was employed. Thus, the absorbed antisera produced a general unspecific staining of most of the tissue. With synthetic ACTH[18-39], such unspecific binding did not occur and with graded amounts of peptide the staining could be removed. The concentrations necessary for removing gastrin cell staining was 200 gg per ml of monoclonal antibody No. 5, diluted to 5 gg per ml. In the pituitary, such absorptions considerably reduced, but did not totally abolish, staining. This difference probably reflects the different concentrations of immunoreactants per cell in pituitary and antrum and, hence, with a constant avidity of the antibody, the concentrations of tissue-bound antigen and of antibody are the main variables determining staining intensity. In agreement with this, the same apparent staining intensities were reached in human pituitary corticotrophs at 5 gg/ml and in human gastrin cells at 15 gg/ml of monoclonal antibody No. 5. This probably reflects that, in sites of high antigen density, say perhaps only 10 per cent of available epitopes have to react with the monoclonal antibody, whereas, in sites of lower antigen density, a figure approach- ing 100% of available epitopes would have to react. Per- turbing this equilibrium will first abolish staining in the site containing low antigen concentrations.

From a practical point of view, absorption controls with the current monoctonal antibodies are economically prohib- itive. With a monoclonal antibody to insulin we have also found that very high concentrations of insulin (above 700 gg/ml) are needed even to diminish insulin cell staining. There are two possible explanations to this: firstly the avidi- ties of the monoclonal antibodies may be low, necessitating high concentrations of antigens. Secondly, polyclonal antisera, which usually can be absorbed out with very low antigen concentrations, may be able to form more stable and complex antigen-antibody aggregates upon absorption (cf. Sternberger 1979). Such complexes would form only with monoclonal antibodies if the antigen contained multiple identical epitopes. This could be achieved e.g. by cross- linking antigen molecules and, most likely, this is the situa- tion met with in tissue.

The succesful absorption control does not prove identity between the antigen detected in tissue and the antigen used for absorption. It merely shows that blocking of the antigen-combining site of the antibody prevents it from reacting with tissue. Proofs for identity must come from use of region-specific immunocytochemistry (Larsson and Rehfeld 1977) and from corroborative biochemical and histophysio- logical studies. Recently, a serious pitfall of the absorption

122

control was unravelled. Aggregated or labelled ant ibodies were found to react with basic and/or aromat ic amino acid- rich peptides, including A C T H and most of the opioid peptides. Absorp t ion of such ant ibodies against any basic peptide sequence (at high concentrations) abolished this staining (Scopsi et al. 1986). Accordingly, it has become necessary to perform also such controls. We employ poly-g-lysine (low molecular weight) as a convenient such reagent and our results show that absorpt ions against this reagent does not abolish staining of pi tui tary cor t icotrophs and melano- t rophs or of antropylor ic gastrin cells. Due to the above considerations alternative control procedures are highly desirable. A good alternative is the use of isotype-matched IgG (or of IgM) and has been successfully employed in the present study.

Acknowledgements. Grant support was from the Danish MRC, Cancer Society and Lundbeck foundation. Dr. Scopsi was sup- ported by a scholarship from the Danish MRC. Skilfulled technical assistance was given by Mrs. B. Traasdahl and K. Hogenhaven.

References

Bennett HPJ, Browne CA, Solomon S (1982) Characterization of eight forms of corticotropin-like intermediary lobe peptide from rat intermediary pituitary. J Biol Chem 257 : 10096-10102

Bj6rklund A, Falck B, Owman Ch (1972) Fluorescence microscopic and microspectrofluorometric techniques for the cellular localization and characterization of biogenic monoamines. In: Ber- son SA (ed) Methods of investigative and diagnostic endocrinology, vol 1, The Thyroid and biogenic amines. Rail, JE, Ko- pin, IJ (section editors). North-Holland, Amsterdam, pp 318- 368

Chen C-LC, Mather JP, Morris PL, Bardin CW (1984) Expression of proopiomelanocortin-like gene in the testis and epididymis. Proc Natl Acad Sci USA 81 : 5672-5675

Cianfriglia M, Armellini D, Massone A, Mariani M (1983) Simple immunization protocol for high frequency production of solu- ble antigen-specific hybridomas. Hybridoma 4:449-455

Comitti R, Racchetti G, Gnocchi P, Morandi E, Galante YM (1987) A monoclonal-based, two-site enzyme imunoassay of human insulin. J Immunol Methods 99:25-37

Eipper BA, Mains RE (1980) Structure and biosynthesis of pro- adrenocorticotropin/endorphin and related peptides. Endocrine Rev 1 : 1-27

Endo Y, Sakata T, Watanabe S (1985) Identification ofproopiome- lanocortin-producing cells in the rat pyloric antrum and duode- num by in situ mRNA-cDNA hybridization. Biomed Res 6: 253-256

Feurle GE, Weber U, Hehustaedter V (1980) Corticotropin-like substances in human gastric antrum and pancreas. Biochem Biophys Res Commun 95:1656-1662

Galfr~ G, Howe SC, Milstein C, Butcher GW, Howard JC (1977) Antibodies to major histocompatibility antigens produced by hybrid cell lines. Nature 266:550-552

Graham RC, Karnovsky MJ (1966) The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney. Ultrastructural cytochemistry by a new technique. J Histochem Cytochem 14:291-302

Herbert E, Birnberg N, Lissitsky J-L, Civelli O, Uhler M (1982) Regulation of expression of pro-opiomelanocortin and related genes in various tissues of the rat and mouse. In: Motta M,

Zanisi M, Piva F (eds) Pituitary hormones and related peptides. Academic Press, New York, pp 31-42

Krieger DT, Liotta A, Brownstein MJ (1977) Presence of corticotropin in brain of normal and hypophysectomized rats. Proc Natl Acad Sci USA 74: 648-652

Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227 : 680-685

Larsson L-I (1977) Corticotropin-like peptides in central nerves and in endocrine cells of gut an pancreas. Lancet II:1321-1323

Larsson L-I (1979) Radioimmunochemical characterization of ACTH-like peptides in antropyloric mucosa. Life Sci 25:1565 1579

Larsson L-I (1981a) Adrenocorticotropin-like and ~-melanotro- pin-like peptides in a subpopulation of human gastrin cell gran- ules: bioassay, immunoassay and immunocytochemical evi- dence. Proc Natl Acad Sci USA 78: 2990-2994

Larsson L-I (1981 b) A novel immunocytochemical model system for specificity and sensitivity screening of antisera against multiple antigens. J Histochem Cytochem 29:408-410

Larsson L-I (1983) Methods for immunocytochemistry of neurohormonal peptides. In: Bj6rklund A, H6kfelt T (eds) Hand- book of chemical neuroanatomy, vol 1. Methods in chemical neuroanatomy. Elsevier, Amsterdam, pp 147-209

Larsson L-I, Rehfeld JF (1977) Characterization of antral gastrin cells with region-specific antisera. J Histochem Cytochem 25:1317-1321

Leary J J, Brigati D J, Ward D (1983) Rapid and sensitive colorimet- ric method for visualizing biotin-labeled DNA probes hybrid- ized to DNA or RNA immobilized on nitrocellulose : Bio-blots. Proc Natl Acad Sci USA 80: 4045M049

Liotta AS, Gildersleeve D, Brownstein MJ, Krieger DT (1979) Biosynthesis in vitro of immunoreactive 31000 dalton corticotropin/flendorphin-like material by bovine hypothalamus. Proc Natl Acad Sci USA 76:1448-1452

Lolait S J, Clements JA, Markwick A J, Cheng C, McNally M, Smith AI, Funder JW (1986) Pro-opiomelanocortin messenger ribonucleic acid and posttranslational processing of beta-endorphin in spleen macrophages. J Clin Invest 77:1776-1779

Pintar JE, Schachter BS, Herman AB, Durgerian S, Krieger DT (1984) Characterization and localization of proopiomelanocortin messenger RNA in the adult rat testis. Science 225 : 632-634

Racchetti G, Fossati G, Comitti R, Putignano S, Galante YM (1987) Production of monoclonal antibodies to calcitonin and development of a two-site enzyme immunoassay. Mol Immunol 24:1169-1176

Reichlin M (1980) Use of glutaraldehyde as a coupling agent for proteins and peptides. Methods Enzymol 70:159-165

Sanchez-Franco F, Patel YC, Reichlin S (1981) Immunoreactive adrenocorticotropin in the gastrointestinal tract and pancreatic islets of the rat. Endocrinology 108:2235-2238

Scopsi L, Wang B-L, Larsson L-I (1986) Nonspecific immunocytochemical reactions with certain neurohormonal peptides and basic peptide sequences. J Histochem Cytochem 34:1469-1475

Smith EM, Morrill AC, Meyer III WJ, Blalock JE (1986) Cortico- tropin releasing factor induction of leucocytederived immunoreactive ACTH and endorphins. Nature 321:881-882

Smyth DG (1983) flEndorphin and related peptides in the pituitary, brain, pancreas and antrum. Br Med Bull 39:25 30

Sternberger LA (1979) Immunocytochemistry, 2nd edn. J. Wiley, New York

Tanaka I, Nakai Y, Nakao K, Oki S, Masaki N, Ohtsuki H, Imura H (1982) Presence of immunoreactive 7-melanocyte-stimulating

�9 hormone, adrenocorticotropin and flendorphin in human gastric antral mucosa. J Clin Endocrinol Metab 54: 392-396

Immunocytochemical characterization of mouse monoclonal ACTH antibodies with a note on staining...

Documents

Transcript of Immunocytochemical characterization of mouse monoclonal ACTH antibodies with a note on staining...