Histochem Cell Biol (1995) 104:309-316

9 Springer-Verlag 1995

Kunio Fujiwara 9Yukinobu Masuyama

Monoclonal antibody against the glutaraldehyde-conjugatedpolyamine,spermine

Accepted: 11 May 1995

A b s t r a c t We developed a mouse monoclonal antibody

(ASPM-29, mAb) against spermine (Spm) conjugated to human serum albumin (HSA) using glutaraldehyde-sodium borohydride, for applications in immunocytochemistry (ICC). The antibody specificity was evaluated by an enzyme-linked immunosorbent assay (ELISA) binding test, simulating the ICC of tissue sections. ASPM-29 showed an almost equal immunoreactivity to Spm and spermidine (Spd) but no reactivity to any of the other polyamine (PA)-related compounds tested. By use of this antibody, indirect immunoperoxidase staining was observed in different tissues fixed with glutaraldehyde in combination with borohydride reduction. In contrast, immunoreactivity was quite low in tissues fixed only with glutaraldehyde. Absorption controls indicated that the immunostaining could be completely inhibited by 50 gg/ml of Spm or Spd and partially inhibited by N-acetylspermine (Ac-Spm), Nl-acetylspermidine (N1-Ac-Spd), or NS-acetylspermidine (NS-Ac-Spd), but was hardly inhibited at all by other PA-related compounds or amino acids. The reactivity of the antibody with Spm conjugated on wells in an ELISA plate was inhibited by micromolar concentrations of Spm, Spd, Ac-Spm, N1-Ac-Spd, or NS-Ac-Spd, in decreasing order, but not by other small molecules. Dense ICC staining was observed in the paranuclear and basal cytoplasm of acinar cells of rat pancreas, submandibular gland and paratid gland, these results being in complete agreement with our recent ICC methods using other mAbs produced against N-(~'-maleimidobutyryloxy) succinimide-conjugated Spm.

Introduction Polyamines (PAs) are low molecular weight organic polycations present in all living cells and implicated in numerous biological processes including cell proliferation, nucleic acid and protein synthesis, and nucleic acid K. Fujiwara (~i~). y. Masuyama Faculty of Pharmaceutical Sciences, Nagasaki University, Bunkyomachi 1-14, Nagasaki 852, Japan

stabilization (for review see Cohen 1978; Algranati and Goldemberg 1977; Pegg 1986; Pegg and McCann 1988; Tabor and Tabor 1984). A better understanding of PA functions may be developed by immunocytochemical (ICC) studies revealing exact cellular and subcellular localizations of endogenous PAs. PA ICC was developed originally using polyclonal anti-spermine (Spin) serum (Fujiwara et al. 1983a; Hougaard et al. 1986, 1987) and quite recently using monoclonal antibodies (mAbs) against Spm conjugated with a carrier protein via the cross-linking agent N-(7-maleimidobutyryloxy) succinimide (GMBS) (Fuji-wara et al. 1994). In PA ICC model experiments in these studies it was suggested that PAs are affected in situ in cells by glutaraldehyde (GA) during the fixation process in PA ICC, resulting in PA derivatives that are immunoreactive with the antibody described above. In order to further elucidate the PA staining reaction mechanism, and to define the cellular pools of PAs precisely, a more direct ICC method using an antibody specific for PAs fixed in tissues is desired. In order to avoid PA redistribution, a strong fixative, GA, was inevitably needed in PA ICC (Hougaard et al. 1986, 1987; Fujiwara 1994; Fujiwara et al. 1994). Such fixation, however, may change the chemical structure of PAs at multiple amino groups in the molecule and, as a result, may change their antigenicity (Fujiwara et al. 1993, 1994; Fujiwara 1994). Also, it is known to be often the case that a hapten antibody recognizes not only the hapten molecule itself but also, in part, conjugation sites of a protein carrier (Campistron et al. 1986; Clements and Beitz 1985; Fujiwara et al. 1983b; Geffard et al. 1984, 1985; Nilsson etal. 1987; Onteniente etal. 1984; Peressini et al. 1984; Schipper and Tilders 1983), and thus that in ICC for small-sized molecules the best results can be expected when the epitope of the conjugate used for immunization closely resembles the fixed tissue antigen (Buijs et al. 1984; Campistron et al. 1986; Onteniente et al. 1984; Seguela et al. 1984). Therefore, we decided to use GA in a hapten-protein conjugate (antigen) preparation for immunization for specific antibodies, and for their location after tissue fixation in PA ICC.

310

In this report, we describe the production, specificity, and application of mAb against Spm using GA in conjugate production and tissue fixation. Immunostaining patterns of rat tissues obtained by the present PA ICC study were compared with those recently reported by us using other mAbs produced against GMBS-conjugated Spm (Fujiwara et al. 1994). Materials and methods

intraperitoneal cavity of an unimmunized BALB/c mouse but injected with 0.5 ml of Pristane (2,6,10,14-tetramethyl pentadecane, Sigma), had been plated. At days 12 and 20 post-fusion, wells were screened for reactivity using an ELISA method as described below. Two strongly reactive wells were identified, and cell lines from these wells were expanded, although cell lines from one well were lost during this process. Subcloning of the remaining cell line was done by limiting dilution. Ascites was raised in BALB/c mice after pretreatment with 0.5 ml Pristane intraperitoneally 1 week before intraperitoneal injection of 106 hybridoma cells. Sub-isotyping of mAb was done using the mouse monoclonal sub-isotyping kit 97-6550 from Zymed (San Francisco, Calif., USA).

Chemicals Spermine, spermidine (Spd), putrescine (Put), cadaverine (Cad), 1,3-diaminopropane (Dap), N-acetylspermine (Ac-Spm), Nl-ace tylspermidine (N~-Ac-Spd), NS-acetylspermidine (NS-Ac-Spd), Nacetylputrescine (Ac-Put), and o-phenylenediamine (OPD), in the form of hydrochloride salts, were purchased from Sigma Chemical St. Louis, Mo., USA. NI,N12-diacetylspermine.2HC1 (2Ac-Spm) and Nl,NS-diacetylspermidine.HCL (2Ac-Spd) were generously given by Dr. N. Seiler at Marrion Merrell Dow Research Institute, Strasbourg, France. Goat anti-mouse gammaglobulin and its fragment Fab' both labeled with horseradish peroxidase (HRP) were purchased from Cappel (West Chester, Pa. USA) and MBL (Nagoya, Japan), respectively. Synthesis of Spm-protein conjugate (antigen) Spm (Sigma, 14.5 mg) in 1 ml of a solution containing 1 M sodium acetate was incubated with 2.0 ml of 30 mM glutaraldehyde for 30 s with stirring. To this mixture was added 25 mg of carrier protein [human(HSA) or bovine serum albumin (BSA)] in 1.0 ml of sodium acetate and the entire mixture was incubated for 10 min with slow stirring at room temperature during which the pH of the solution was adjusted to 8. Then, 4 mg of NaBH 4 (Sigma) was added so that the double bonds were saturated. The coupling mixture was dialyzed against several changes of 10 mM sodium acetate overnight at 4 o C. Insoluble materials was removed by centrifugation at 10 000 g for 10 min. Immunization Four-week-old, female BALB/c mice were injected intraperitoneally with 100 btg of Spm conjugate emulsified in complete Freund's adjuvant (Difco). Subsequently, they received three injections of 50 btg of the conjugate alone at 2-week intervals. Following immunization, antisera were collected, and their antibody affinity was evaluated by enzyme-linked immunosorbent assay (ELISA) as described below. After the fourth immunization, one of the three mice showed a high antibody titer and affinity, and was selected for hybridization. One month after the last immunization injection, the mouse received a fifth fntraperitoneal booster injection and was killed 4 days later. Cell fusion Spleen cells (108) obtained from the immunized BALB/c mouse and 2x107 myeloma cells (P3/NS-1/1-Ag4-1), which were previously washed 3 times in serum-free RPMI 1640 medium, were incubated in the presence of 0.6 ml of 40% polyethylene glycol 1500 (Boehringer Mannheim) at 37 ~ C for 1 min. Ten milliliters of RPMI 1640 containing 10% fetal calf serum was added dropwise over 3 min with gentle stirring. Cells were centrifuged at 400 g for 5 rain and resuspended in HAT medium (100 gM hypoxanthine, 0.4 btM aminopterin, 16 btM thymidine) with 10% fetal calf serum, and plated out in 96-well tissue culture plates (Corning) at a density of 105 cells per well in which 105 feeder cells obtained from the

ELISA method In screening clones for the production of antibody against Spm, wells in microtiter plates (Nunc Immunoplate 1, Roskilde, Denmark) were coated with 100 btl of the Spm-HSA conjugate (10 btg/ml) in 10 mM TRIS-HC1 buffer, pH 8.5, containing 10 mM NaCI and 10 m M N a N 3 for 30 rain at 37 ~ C (Fujiwara et al. 1993). Wells were blocked with 100 gl of 1% skimmed milk in TRISbuffered saline (TBS) containing 0.1% NaN 3 for 1 h at room temperature. The wells were then incubated overnight at 4 ~ C with hybridoma culture supernatant or ascites fluids (diluted 1:100 000). After rinsing with TBS containing 0.05% Tween 20 (TBST), the wells were incubated for 1 h at room temperature with HRP-labeled goat anti-mouse IgG, or Fab', diluted 1:1000 in TBS. Bound enzyme activity was measured using 100 pl of 30 mM citrate buffer, pH 5.3 containing 0.5 mg/ml OPD and 0.012% H202, and was measured at 492 nm with an automatic ELISA analyzer (SLT-Lab Instruments, Salzburg, Austria).

ELISA binding test (ELISABT) Wells in a microtiter plate, coated with poly-L-lysine (30 btg/ml), were activated with 2.5% GA at pH 10.0 for t h according to our previous method (Fujiwara et al. 1993; Fujiwara 1994). Test compounds at various concentrations were added to the wells and incubated for 1 h at 25 ~ C. Excess aldehyde groups were blocked with 0.3% hydroxylamine for 30 rain and 0.5% NaBH 4 for 10 min at 25 ~ C. The wells were further blocked with 1% skimmed milk for 1 h for protein binding sites, and then incubated overnight at 40 C with mAb hybridoma culture supernatant (diluted 1:100). The wells were then incubated with HRP-labeled anti-mouse IgG (whole) (1:1000) for 1 h. The bound enzyme activity was then measured as described above. ELISA inhibition test (ELISAIT) Spin was conjugated to the wells of a microtiter plate in the same manner as previously described (Fujiwara et al. 1993; Fujiwara 1994). The wells were then incubated with a fixed concentration of ASPM-29 (1:100) and different compounds at various concentrations overnight at 4 ~ C, followed by incubation with goat antimouse IgG labeled with HRP (1:1000) for 1 h at room temperature. The bound HRP activity was measured using the ELISABT as described above. Tissue materials Male Wistar rats (250 g body weight) were anesthetized with pentobarbital intraperitoneally, perfused transcardially with 100 ml of saline for 1 rain, and subsequently fixed by perfusion with 300 ml of one of the following fixatives in 0.1 M phosphate buffer, pH 7.4 for 3 min at room temperature: (1) freshly made 2.5% GA (Nacalai Tesque); (2) a mixture of 2.5% GA and 4% paraformaldehyde; (3) 4% paraformaldehyde; (4) 5% carbodiimide [(1-ethyl-3-3-di-

311 methyl aminopropyl]carbodiimide, Sigma]. Specimens from various tissues were post-fixed in the same fixative overnight at 4 ~ C and subsequently routinely embedded in paraffin. In addition, human melanoma BD and neuroblastoma IMR 32 cell lines, cultured as detailed elsewhere (Fujiwara et al. 1993) were fixed in the same fixatives for 30 min at room temperature, and were then processed for PA ICC as described below. [mmunocytochemistry (ICC) This was performed essentially as previously described (Hougaard et al. 1987; Fujiwara et al. 1993), but with a slight modification. Deparaffinized sections were treated with 0.03% H202 in methanol for 30 min and then rehydrated in a graded ethanol series. Subsequently, they were treated with 0.2% NaBH4 for 10 rain, followed by 0.03% protease (from Bacillus arnyloliquefaciens, type V, Sigma, 8.9 U/rag) in TBS for 60 min at room temperature and exposed to the mAb ASPM-29 (0.03-0.5 btg/ml, the concentration of which was determined by the conventional 'sandwich ELISA' using chromatographically purified mouse IgG, Zymed, USA, as a standard) at 4 o C overnight.After brief rinsing in TBS, 1% Triton X-100, the site of the antigen-antibody reaction was revealed by the indirect immunoperoxidase method using HRP-labeled goat anti-mouse IgG/Fab' diluted 1:300 in TBS, and the HRP activity demonstrated with diaminobenzidine development (Graham and Karnovsky 1966). Controls included the use of conventional staining controls (second level controls) as well as absorption controls for the ASPM-29 mAb using Spin, Spd, Put, Cad, Dap, Ac-Spm, 2AcSpm, 2Ac-Spd, N,N'-diacetylputrescine (2Ac-Put) (Sigma), AcPut, NLAc-Spd, NS-Ac-Spd, lysine, arginine, ornithine, glycine and 7-aminobutyric acid (GABA) at a concentration of 50 btg/ml.

Antibody dilution Microtiter wells coated with a fixed amount of SpmBSA conjugate (10 gg/ml) were used to test for antibody binding using serial dilutions of ASPM-29 mAb (hybridoma culture supernatant). As shown in Fig. 1, with both assays using goat anti-mouse IgG (whole) or Fab' labeled with HRP as the second antibody marker, significant binding activity was observed at more than 4000 times dilution of the antibody. About 3 times higher immunoreactivity of ASPM-29 mAb was seen with the second antibody IgG (whole) than that with IgG fragment Fab', indicating that the second antibody titer used at the same dilution (1:1000) was higher with the whole IgG than with Fab' fragment. No antibody binding was seen

E

L--

0

=o 0

<

9

0

The immunogen Spm-HSA conjugate prepared via GA, elicited the production of anti-Spin antibody in each of the three mice. The antibody titers in the serum samples were examined by the ELISA method using Spm-BSA as a solid phase antigen, and they were found to peak 2 weeks after the third booster injection.

0

C~ O~

'

'

10

Results

Antibody response

0

10 1

10 2 10 3 Dilution (fold)

10 4

10 5

Fig. 1 Dilution curves for ASPM-29. Several dilutions of ASPM29 were incubated in microtiter wells coated with spermine (Spm)BSA conjugate overnight at 4~ C, followed by the procedure described for the ELISA method using horseradish peroxidose (HRP)-labeled goat anti-mouse IgG (whole) (@) or Fab' (9 each in a dilution of 1:1000 as the second antibody 0.4 E c

Generation and detection of mAbs to Spm From a fusion experiment 605 hybridoma lines were produced. Two hybridoma lines secreted antibodies that bound to the Spm-BSA conjugate but did not recognize BSA, as determined by the ELISA. Ten additional clones bound to both the Spm-BSA conjugate and BSA. In the ELISAs, we used immobilized Spm-BSA rather than Spm-HSA conjugates to reduce the number of unwanted positive clones that reacted with HSA determinants but not with Spm.

Sub-isotyping of subclones Both subclones of the hybridoma obtained by limiting dilution were found to produce only antibody of the IgG 1 sub-isotype.

g

0.3

g c

0.2

o 09 .Q

0.1 9

0.0

,

10 -3

l

10 -2

I

O I

I

10 -1 10 0 10 1 Concentration (gM)

i

I

i[

10 2

10 3

Fig. 2 Reactivity of ASPM-29 as measured by its immunoreactivity in the ELISA binding test (ELISABT). Activated wells as prepared for the ELISABT were incubated with various concentrations of Spm (O), spermidine (Spd) (e), or Spm (A) in which, however, a NaBH4 reduction step was omitted from the protocol of ELISABT. The wells were then reacted with ASPM-29 (1:100) followed with HRP-labeled goat anti-mouse IgG (whole) (1:1000). Note ASPM-29 mAb is immunoreactive with Spm and Spd only when they were reduced

312 0.8 E o~ 0.6 t-

0.4

o

J~

.o

<

0,2

0.010 -2

10 -1

10 0

10 1

10 2

10 3

Concentration (~tM)

Fig. 3 Changes in immunoreactivity of the bound Spm in the ELISABT after reaction with glutaraldehyde (GA). The wells which had been conjugated with Spin at various concentrations were reacted again with GA at a concentration of 0 ((2)), 0.025% (&), 0.25% (/~) or 2.5% (O) at pH 7.2, for 1 h at room temperature, and then subjected to the ELISABT using ASPM-29 mAb. Note enhanced immunoreactivity with GA concentrations

duced a dose-dependent saturation curve with Spm and Spd in the range between 1 btM and 400 btM. The extent of the immunoreaction was 100% with Spm and 90% with Spd at 10 btM. No immunoreaction occurred with Put, Cad, Dap, 2Ac-Spm, Ac-Spm, 2Ac-Spd, N1-Ac Spd, NS-Ac-Spd, 2Ac-Put, Ac-Put, lysine, arginine, ornithine, GABA or glycine. If the NaBH 4 reduction step is omitted from the protocol, no antibody binding occurred at concentrations ranging from 1 to 400 gM with any of the PAs or related compounds (Fig. 2). PAs possess plural primary and/or secondary amino groups and it is therefore possible that they might be fixed by the reaction of some of the groups with the fixative used in the ICC. In order to simulate the effect of GA on the immunoreactivity of PAs in situ with the antibody, wells on which various known amounts of PAs (Spm, Spd, Put, Cad, or Dap) had been conjugated were again incubated with various amounts of GA and were then subjected to the ELISABT. Among those about 2 times higher immunoreactivity was observed only with the treated Spm than with non-treated Spm (Fig. 3), and a very slight increase in immunoreactivity was noticed with the treated Spd (data not shown).

100

'~

ELISAIT

80

~- 40

lal 'ID e-

O m

20 0

........

10 -2

I

10 -1

. . . . . . . .

I

. . . . .

10 0 Inhibitor

~,,I

10 1

. . . . . . . .

I

. . . . .

10 2

~11

10 3

(btM)

Fig. 4 Reactivity of ASPM-29 as measured by its immunoreactivity in the ELISA inhibition test (ELISAIT). The curves show the amount (percentage) of bound enzyme activity (B) for various doses of Spin (0), Spd (0), N-acetylspermine (Ac-Spm; ~), Nl-acetylspermidine (N1-Ac-Spd, 9 I), NS-acetylspermidine (NS-Ac-Spd; &) or N1,N12-diacetylspermine (2Ac-Spm; [Z) as a ratio of that bound using the HRP-labeled second antibody alone (B~

with type-matched mAb (IgG 1) of anti-penicillin (Cosmobio). Thus, HRP-labeled goat anti-mouse IgG was used in the following experiments.

Evaluation of antibody specificity by ELISABT This ELISABT simulates the ICC of tissue sections, based on the principle of coupling the amino group of analytes to the wells of a microtiter plate activated with poly-L-lysine and GA, and incubating the wells by the indirect immunoperoxidase method (Fujiwara 1994; Fujiwara et al. 1993; 1994). As shown in Fig. 2, analysis of the relationship between the concentration of each of the PAs applied to the wells and the bound HRP activity pro-

This was achieved by the principle of competition between PAs or related analogs (free in solution) and a fixed amount of the bound antigen for a limited number of binding sites on the mAb. Calibration curves were plotted showing the relationship between concentrations of the analytes and the percentage of bound mAb, giving dose-dependent inhibition curves with Spm, Spd, AcSpm, N1-Ac-Spd, NS-Spd, and 2Ac-Spm (Fig. 4). The dose required for 50% inhibition of binding was used as an indication of the strength of inhibition. This dose was 32 btM with Spm or Spd, 50 btM with Ac-Spm and 110 gM with with NI-Ac-Spd or NS-Ac-Spd, and less than 300 btM with 2Ac-Spm. No inhibition occurred with Put, Cad, Dap, Ac-Put, 2Ac-Put, ornithine, lysine, arginine, or GABA.

Evaluation of tissue staining No significant staining was seen when tissues were fixed with a mixture of 2.5% GA plus 2% paraformaldehyde or 4% paraformaldehyde alone, the latter of which also resulted in poor tissue preservation. Fixation with 5%

Fig. 5a-d Different cell types stained for polyamines by immuno-

cytochemistry using ASPM-29 mAb (indirect method). Rat pancreas (a), submandibular gland (b), parotid gland (c), and adenohypophysis (d). Strong staining occurs in the paranuclear and basal cytoplasm of acinar cells of three types of exocrine tissues (a, b and c). Less strong staining was seen in the cytoplasm of the islets of Langerhans and adenohypophysis cells of two types of endocrine tissues (a and d, respectively), x350

~J

314 carbodiimide produced a diffuse pattern of immunostaining. Fixation with GA alone resulted in very weak staining of certain cells of rat tissues. However, GA fixation combined with reduction of tissue sections by NaBH 4 resulted in intense staining of such cells. Particularlly strong immunoreaction for PAs was observed in several cell systems active in peptide and protein synthesis (Fig. 5). Thus, strong staining was found in the paranuclear and basal cytoplasm of pancreatic acinar cells and of terminal portion cells (acinar cells) of submandibular glands and parotid glands, and the cytoplasm of human melanoma BD and neuroblastoma IMR 32 cell lines (data not shown). Less strong staining was seen in the cytoplasm of the islets of Langerhans, and the cytoplasm of cells in the adenohypophysis, in which, however, the degree of immunostaining varied from cell to cell in the tissue, although the cell types have not been precisely identified yet (Fig. 5). Only very little or no staining was found in the nuclei. Conventional immunocytochemical staining controls (second level controls) were all negative. The absorption controls for ASPM-29 mAb showed that an addition of Spm or Spd at a concentration of 50 Bg/ml into the antibody abolished all staining (Fig. 5) and that addition of Ac-Spm, N1-Ac-Spd, NS-Ac-Spd (50 gg/ml) weakened the immunostaining. No cross-reactivity was observed to PA-related compounds (Put, Cad, Dap, 2Ac-Spm, 2Ac-Spd, 2Ac-Put, Ac-Put) or amino acids (lysine, arginine, ornithine, glycine and GABA).

Discussion Previous ICC studies for PAs have been done using PA antibodies, both polyclonal (Fujiwara et al. 1983a, 1993; Hougaard et al. 1986, 1987) and monoclonal (Fujiwara et al. 1994), both of which were produced against PABSA (or HSA) conjugates prepared using a heterobifunctional cross-linking agent, GMBS. Such a conjugation method, however, led to complicated Spm-protein carrier complexes, because of the multiple reactive amino groups of PAs in the molecules, resulting in fact in the production of mAbs ASPM-1 and ASPM-2 with different antibody specificities (Fujiwara et al. 1994). The ASPM-2 mAb specific for acylated Spm or Spd, but not for Spm or Spd, as examined by ELISABT, proved to be very useful in PA ICC studies, staining a variety of cancer cell lines and mammalian tissues (Fujiwara et al. 1994). However, the exact staining mechanism for nonacylated free PAs (possibly Spm and Spd), occurring exclusively in situ in mammalian cells, has not fully been elucidated yet (Fujiwara et al. 1994). In order to examine the mechanism and to further define the cellular pools of PAs we have now prepared and characterized a mAb (ASPM-29) against Spm using GA in conjugate preparation. A large number of hybridomas were obtained, with 90% of wells yielding colonies, but only two had selective binding activity for Spm. One of the hybridomas, which showed a prolific growth rate, was cloned and

four of the resulting subclones continued to secrete the IgG 1 antibody. With conventional ELISA using a hapten-protein conjugate as a solid phase antigen, the mAb showed almost the same degree of immunoreactivity for Spin and Spd conjugates (see Fig. 1 in the case of Spm-GA-BSA conjugates) and no cross-reactivity for conjugates of structurally similar amines and amino acids, such as Put, Cad, Dap, ornithine, arginine and GABA. No antibody activity against the carrier HSA or GA-HSA was evident, demonstrating lack of recognition. An ELISABT indicates that ASPM-29 mAb had an equally strong immunoreaction with both Spm and Spd, which had been reduced with NaBH 4, but none at all with non-reduced Spm and Spd (Fig. 2). Furthermore, ASPM-29 mAb was demonstrated to be primarily specific for Spm or Spd by the ELISAIT, since in the test the highest immunoreactivity was found with Spm or Spd, and a significant cross-reactivity with Ac-PAs (Ac-Spm, N1-Ac-Spd, N sAc-Spd or 2Ac-Spm), such cross-reactivity, however, is of no significance in PA ICC without immunoreaction with the mAb when assayed in the ELISABT as a model of PA ICC (Figs. 1, 2, 4). Also of note in the ELISAIT is the fact that the degree of inhibition defined as an ECs0 value (32 gM with both the strongest inhibitors, Spm and Spd) seems to be much larger than expected, since the previously produced anti-Spm mAbs, ASPM-1 and ASPM-2, showed much stronger immunoreaction with Ac-Spm and 2Ac-Spm with EC50 values of 0.09 gM, and 0.32 gM, respectively (Fujiwara et al. 1994). Besides, the present ELISABT, where PAs and analogs conjugated on a solid phase were treated again with GA followed by NaBH 4, showed enhanced immunoreactivity only with S p m and to a lesser extent with Spd among the compounds tested (Fig. 3). This suggests the possibility that ASPM-29 mAb was generated recognizing SpmGA-HSA conjugates, in which Spm was reacted with GA at multiple amino groups in the molecule (possibly at both the primary amino groups), and then was reduced with NaBH 4, just as such GA conjugation with PAs occurs during tissue fixation of PA ICC (Fujiwara 1994; Fujiwara et al. 1994; Mattingly 1990; Stahl et al. 1978; Tabor et al. 1971). A previous PA ICC using polyclonal anti-Spm serum was reported by Hougaard et al. (1987; K. Fujiwara of the present team was involved) where it was pointed out that pretreatment of specimens with an appropriate protease concentration followed by dehydration in an ethanol series and air-drying overnight is essential for the ICC reaction. However, in the present ICC study using ASPM-29 mAb, the ethanol treatment and air-drying resulted in no significant immunostaining potency, though the optimal digestion conditions for most tissues were 0.03% protease for 1 h. Furthermore, in the ICC applications, the borohydride treatment, in combination with GA fixation, gave the most intense staining, being consistent with the results of ELISABT (Figs. 2, 5). This might indicate that the ELISABT system used may indeed predict how the given antibody will react with the

315 antigens fixed in the tissue. The ICC results obtained in this study indicate that PAs can be demonstrated specifically in the cytoplasmic compartments in all the exocrine and endocrine cell types and tumor cell lines tested, whereas the nuclei remain virtually unstained (Fig. 5). Also, of note is the finding that the immunostaining was much stronger with the exocrine cell types than with the endocrine cell types, suggesting a general difference in PA concentrations between the two cell types. Furthermore, the localization of PA-like immunoreactivity found in the paranuclear and basal cytoplasm rather than the apical cytoplasm of rat pancreas acinar cells and terminal portion cells of the submandibular gland and pararid gland is in complete agreement with our recent results obtained using other mAbs against GMBS-conjugated Spm (Fujiwara et al. 1994), suggesting that PAs exist in conjunction with endoplasmic reticular regions within the cells, which are closely associated with peptide or protein synthesis (Igarashi et al. 1978; Tabor and Tabor 1984; Pegg and McCann 1988). These results are, however, in contradiction with previous results of PA ICC using anti-Spin serum polyclonal (Hougaard et al. 1987), showing immunostaining in secretory granules of rat pancreas acinar cells and convoluted duct cells of the submandibular gland. Such immunostaining in the granules might occur due to unsuitable fixation conditions used for perfusion because of the characteristic PA diffusibility (Fujiwara et al. 1994). It was found in the present PA ICC study using ASPM-29 mAb that the higher the concentration of the strong fixative GA used under well-controlled conditions (1.0-5.0% GA in 300 ml total volume per rat weighing about 200 g), the more fixed and reproducible PA immunostaining patterns for rat tissues were obtained without any loss of PA antigenicity in situ (Fig. 5). Also demonstrated in this study was the fact that polyclonal anti-Spm serum (Fujiwara et al. 1983a) sometimes produced non-specific immunostaining especially in the cytoplasm and/or karyoplasm of the islet cells and duct cells of rat pancreas, but once an IgG fraction purified from the anti-Spm serum was used in PA ICC, the immunostaining was almost identical to the results of the present ICC using ASPM-29 mAb under the same conditions (see Materials and methods; Fig. 5). This suggests that such non-specific immunostaining might be due in part to IgM-class antibodies present in the anti-Spin serum, as has been often shown in other cases of ICC (Larsson 1988), because this was also true with PA ICC using three IgM-class mAbs specific to Spm previously produced by us (Fujiwara et al. 1994). In conclusion, as in the previous PA ICC using ASPM-2 mAb (Fujiwara et al. 1994), it is again proposed that during the fixation process of PA ICC, PAs were affected by GA at multiple amino groups in the molecule, converting to a variety of PA derivatives, certain of which may become strongly immunoreactive with ASPM-29 mAb after treatment with NaBH 4. ASPM-29, a mAb prepared newly in this study shows quite different immunochemical properties from the previous mAbs against GMBS-conjugated Spin in many respects, but

produced completely identical ICC results. Thus, ASPM29 mAb should have potential in PA ICC for the clarification of the biological role of PAs in a variety of cells. Acknowledgements I. Sakamoto, K. Hachiya and S. Ikutsuki are thanked for technical assistance, and E. O'Dowd for help in preparing the manuscript.

References Algranati ID, Goldemberg SH (1977) Polyamines and their role in protein synthesis. Trends Biochem Sci 2:272-274 Buijs RM, Geffard M, Pool CW, Hoorneman EMD (1984) The dopaminergic innervation of the supraotic and paraventricular nuclei. A light and electron microscopic study. Brain Res 323: 65-72 Campistron G, Buijs RM, Geffard M (1986) Specific antibodies against aspartate and their immunocytochemical applications in the rat brain. Brain Res 365:179-184 Clements JR, Beitz AJ (1985) The effects of different pretreatment conditions and fixation regimes on serotonin immunoreactivity: a quantitative light microscopic study. J Histochem Cytochem 33:778 Cohen SS (1978) What do the polyamines do? Nature 274: 209-210 Fujiwara K (1994) An evaluation of polyamine immunocytochemistry using immunocytochemical model systems. Histochemistry 101:287-294 Fujiwara K, Asada A, Kitagawa T, Yamamoto K, Ito T, Tsuchiya R, Sohda M, Nakamura N, Hara K, Tomonaga Y, Ichimaru M, Takahashi S (1983a) Preparation of polyamine antibody and its use in enzyme immunoassay of spermine and spermidine with fl-D-galactosidase as a label. J Immunol Methods 61: 217-226 Fujiwara K, Saikusa H, Kitagawa T, Takahashi S, Konishi Y (1983b) Specificity of antisera produced against mitomycin C. Cancer Treat Rep 67:1079-1084 Fujiwara K, Araki M, Kitagawa T, Inoue Y (1993) A new enzymelinked immunosorbent assay (ELISA) for studying immunocytochemical procedures using an antiserum produced against spermidine as a model. Histochemistry 99:477-483 Fujiwara K, Furukawa K, Nakayama E, Shiku H (1994) Preparation and characterization of monoclonal antibodies against polyamine, spermine: immunocytochemical localization in rat tissues. Histochemistry 102:397-404 Geffard M, Buijs RM, Seguela R Pool CW, Le Moal M (1984) First demonstration of highly specific and sensitive antibodies against dopamine. Brain Res 294:161-165 Geffard M, Dulluc J, Heinrich-Rock A-M (1985) Antisera against the indolealkylamines, tryptophan, 5-hydroxytryptophan, and 5-hydroxytryptamine, 5-methoxytryptophan, and 5-methoxytryptamine tested by an enzyme-linked immunosorbent assay method. J Neurochem 44:1221-1228 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 Hougaard DM, Fujiwara K, Larsson L-I (1986) Polyamine cytochemistry: comparisons between cytochemical, autoradiographic, immunocytochemical and chemical results in the prostate. Histochem J 18:321-328 Hougaard DM, Fujiwara K, Larsson L-I (1987) Immunocytochemical localization of polyamines in normal and neoplastic cells. Comparisons to the formaldehyde-fluorescamine and o-phtalaldehyde method. Histochem J 19:643-650 Igarashi K, Watanabe Y, Nakamura K, Kojima M, Fujiki Y, Hirose S (1978) Changes of substrate specificity by polyamines of ribonucleases which hydrolyze ribonucleic acid at linkages attached to pyrimidine nucleotides. Biochem Biophys Res Commun 67:1070-1077

316 Larsson L-I (1988) Immunocytochemistry, theory and practice. CRC Press, Boca Raton, Florida Mattingly PG (1990) Mono-protected diamines. N
Peressini S, Bmsco A, Pecci Saavedra J (1984) Basis for the specificity of anti-5-HT-like antisera in immunocytochemistry applied to the central nervous system. Histochemistry 80:597-602 Schipper J, Tilders FJH (1983) A new technique for studying specificity of immunocytochemical procedures: specificity of serotonin immunostaining. J Histochem Cytochem 31:12-18 Seguela R Geffard M, Buijs RM, Le Moal M (1984) Antibody against GABA: specificity studies and immunocytochemical results. Proc Natl Acad Sci USA 81:3888-3892 Stahl GL, Walter R, Smith CW (1978) General procedure for the synthesis of mono-N-acylated 1,6-diaminohexanes. J Org Chem 43:2285-2286 Tabor CW, Tabor H (1984) Polyamines. Annu Rev Biochem 53: 749-790 Tabor H, Tabor CW, De Meis L (1971) Chemical synthesis of Nacetyl-l,4-diaminobutane, Nl-acetylspermidine, and NS-acetylspermidine. Methods Enzymol 17B:829-833