Characterization of a monoclonal antibody that specifically binds to choline phospholipids and its use in immunocytochemistry.

0022-1554/92/$3.30

Vol. 40, NO.6, pp. 827-838. 1992 Printed in UXA.

The Joumal of Histochemistry and Cytochemistry

Copyright 0 1992 by The Histochemical Society, Inc.

Original Article

I

Characterization of a Monoclonal Antibody That Specifically Binds to Choline Phospholipids and Its Use in Immunocytochemistry' MANUEL P. MARK,2 TAKANORI TSUJI, JACQUES PORTOUKALIAN, ABDELHADI REBBAA, GHASSAN ZIDAN, and JEAN-VICTOR RUCH Instittrt de Biologie Me'dicale, INSERM-Universite' Louis PaJteut; Faculti de Me'decine, 67085 Strasbourg, France (MPM3nGZ,JVR), and Laboratoire d'lmmunologie et de Cance'mlogie &pirimentale, INSERM U 218, Centre L2on Birard, 69373 Lyon France (JeAR).

Received for publication August 1, 1991 and in revised form December 31, 1991; accepted January 16, 1992 (1A2418).

A monoclonal IgM (MC22-33F),raised in response to mouse embryonic dental papilla cells, was selected for further analysis on the basis of the unusual resistance of its epitope to detergent exuactiolls and protease treatmentsof cell cultures. Binding of MC22-33F to cultured cells was abolished after either pre-treatment of the cells with phospholypase C or pre-incubation of the hybridoma culture supernatant with multilam& phosphatidylcholine-containingvesicles. MC2233F reacted with phosphatidylcholine, with the phosphatidylcholine analogue dimethylphosphatidylethanolamine,and with sphingomyelin immobilized on polystyrene surfaces or in thin-layer chromatograms. Crossreaction with other phospholipids was not observed. The surface of cultured epithelial cells was labeled by MC22-33Fat sites of bleb formation. Combining immunostainingby MC22-33F and histochemical staining of cultured cells revealed codistribution of

Introduction Phospholipids are major structural components of biological membranes and can accumulate in the cytoplasm of certain cell types under various physiological or pathological conditions (for review see Wohan, 1964). Recently, fluorescent analogues of specific phospholipids have been developed to examine the transport of these molecules to different cell organelles (for review see Voelker, 1990). However, the use of such fluorescent tracers is restricted to living cells in monolayers. Localization of endogenous lipids in cells and tissues relies on histochemical methods of poorly defined specificities (Bonilla and Prelle, 1987; Fowler and Greenspan, 1985; Greenspan et al., 1985; for review see Pearse, 1968). The recently developed histoenzymatic technique employing phospholipase A2-gold

Supported by a grant from INSERM CJF 88-08 and by the Fondation pour la Recherche Mtdicale. Correspondence to: Dr. Manuel P. Mark, Institut de Biologie MCdicale, FacultC de MCdecine, UniversitC Louis Pasteur, 11, rue Humann, 67085 Strasbourg, France.

phospholipid-containing inclusions with either lysosomes or neutral fat droplets, and inhibition of lipid degradation by kanamycin resulted in a parallel accumulation of these inclusions and of neutral fats in the cytoplasm. Immunolabeling by MC22-33F of frozen mouse tissues was maximal in fat-storing and steroid-producingcells. Extracellular phospholipids present in d c m g cartilagesepta strongly reacted with MC22-33F. This monoclonal antibody o&rs an interesting alternative to histochemical lipid stains for investigating fatty metamorphosis and extracellular lipid deposition under physiological and pathological conditions. (jliistochem Cytochem 40:827-838, 1992) KEYWORDS:Monoclonal antibodies; Phospholipids; Immunocyto-

chemistry; Aminoglycosides; Matrix vesicles; Cartilage mineralization: Mouse.

complexes to detect phospholipids (Coulombe et al., 1988)appears promising. In the same context, the use of speclfic anti-phospholipid antibodies in immunocytochemistry is of potential interest. Monoclonal antibodies (MAb) have been obtained that recognize the stereospecific configuration of phosphatidylserine(Umeda et al., 1989) and phosphatidylcholine (PC) (Nam et al., 1990) or that distinguish between the lamellar and the hexagonal phase of phosphatidylethanolamine (PE) (Rauch et al., 1986). To our knowledge, such MAb have never been employed to immunolocalize phospholipids. In this study we report the characterizationof an MAb, MC2233F, to choline phospholipids and its use as a probe to examine their subcellular localization and tissue distribution.

Materials and Methods Cell. and Tirsues The rat osteosarcoma cell line ROS 17/23 (ROS cells) was a gift from Dr. Farach-Carson (The University of Texas at Houston, Dental Branch). Hu827

Downloaded from jhc.sagepub.com by guest on August 26, 2015

MARK,

828

man embryonicfibroblasts were donated by the Department of Cytogenetics, Hopital de Hautepierre, Strasbourg,and KB cells (transformedhuman oral epithelial cells) by the Department of Virology, Medical School, Strasbourg. Primary cultures of trypsin-isolated enamel organs from embryonic first mandibular molars (dental epithelial cells) were performed according to Lesot et al. (1984). Unless otherwise specified, the cells wcre cultured on glass coverslips in RPMI medium supplemented with 10% fetal calf serum and 0.1 mglml kanamycin. The cultures were grown in a humidified incubator under a 5% co2/95% air atmosphere. The medium was changed every 4 days. Unfixed frozen sections (8 pm thick) of newborn and adult Swiss mice tissues were collected on polylysine-coated glass slides.

Monoclonal Antibody The hybrid clone MC22-33F was derived from the spleen of a Fisher rat immunized intraperitoneally with mouse embryonic dental papilla cells (Zidan and Ruch, 1989).The spleen cells were fused with NS-1 mouse myeloma cells and selected in HAT medium. The wells were screened by immunofluorescence microscopy using unfixed frozen sections of 2-day-old mouse first lower molars. Positive hybridomas were cloned twice by limiting dilution. Immunoglobulinsubclassanalysis showed that the MAb MC2233F was an IgM (Zidan and Ruch, 1989).

Immunofluorescence Cells on coverslips were rinsed with phosphate buffered saline (PBS; 8.10 mM Na2HP04, 1.47 mM KH2P04, 137 mM NaCI, 2.70 mM KCI, 1.05 mM MgC12, and 0.90 mM CaCIz), fixed in 3% paraformaldehyde freshly prepared in 0.1 M sodium phosphate buffer, pH 7.4 (15 min at O'C), then washed with four changes of 0.1 M sodium phosphate buffer containing 50 mM N h C I for 30 min and eventually permeabilized by 0.2% Triton X-100 in PBS ( 5 min at OT). Frozen tissue sections were hydrated in PBS. The cells and the tissue sectionswere then incubated (30 min at 37°C) with undiluted hybridoma supernatant of MC22-33F (antibody concentration 15 pglml, as determined by ELISA using rat IgM as a standard).After washing in PBS (three times for 3 min), the specimens were incubated with fluorescein-conjugated affinipure F(ab)2 fragments of anti-rat IgM, mu chain specific (Jackson Immunoresearch Labs-Immunotech; Marseille, France), diluted 1:40 in PBS and were mounted in glycerol-PBS containing paraphenylenediamine. The sections were photographed with a Leitz Orthoplan photomicroscope. Non-immune rat IgM (Jackson; 40 pg/ml) served as a negative control of the immunostaining procedure In one set of experiments, aliquots of MC22-33F hybridoma supernatant were incubated with 0.5 mglml of a commercial preparation of egg phospholipids (phosphatidylcholine Type X-E; Sigma, St Louis, MO) or with 0.5 mglml of purified egg PC (Type V-E; Sigma), or with 1.0 mglml of phosphorylcholine chloride (Sigma) for 1 hr at 37°C before immunostaining. Both phospholipids and phosphorylcholine were added to the supernatant in the form of l o x suspensions or solutions in PBS.

Evaluation of the Fixation Conditions f o r Antigenic Preservation ROS cells were fixed in four different ways: (a) 3% paraformaldehyde (10 min at 0°C); (b) 95% ethanol (10min at 0°C); (c) 100% acetone (10 min at 0.C) after dehydration in 50% acetone (1min at 0°C);and (d) paraformaldehyde followed by 95% ethanol (10 min at 0°C each). The cells were then treated with 0.2% Triton X-100 in PBS (5 min at 0°C) and immunostained by MC22-33F as described above.

nm,PORTOUKALIAN, REBBAA, ZIDAN, RUCH

Effect of Enzyme Treatments and Pemeabilization Conditions on Antigen Preservation Unfixed ROS cells were first permeabilized with 0.20/0 Triton X-100 in PBS (5 min at O'C), then incubated for 15 min at 37'C with 20 Fglml PBS of either TPCK trypsin (Sigma) or pronase E (Sigma), or for 1 hr at 37'C with 1 Ulml PBS of phospholipase C (Sigma). The proteases were then inactivated by fetal calf serum and by extensive washing of the cell cultures with PBS. In another set of experiments, unfixed ROS cells were permeabilized for 10 min at 0°C with one of the following detergent solutions: (a) 0.2% Triton X-100 in 10 mM sodium phosphate buffer, pH 7.4, or in 50 mM sodium acetate buffer, pH 4.5, 50 mM Eis-HC1 buffer, pH 9, PBS plus 0.6 M NaCI, or PBS plus 0.6 M KCl; or (b) 1% Triton X-100, 0.5% deoxycholate in PBS with and without 0.6 M KCI. The cell residues still adhering to the coverslips after these treatments were then immunostained by MC22-33F.

CO-localization of MC22-33F Binding with Neutral Red Uptahe in Human Fibroblasts Cells on marked glass coverslips were incubated in a solution of neutral red (500 pglml PBS; C.I. 50040; Merck, Darmstadt, FRG) for 10 min at 37°C in COzlair. The dye solution was pre-warmed just before use. After incubation, the cells were mounted in PBS and photographed immediately using a Leitz Orthoplan microscope with a x 40, numerical aperture 0.75, Phaco 2 objective. The unfixed cultures were then permeabilized by 0.2% Triton X-100, immunostained by MC22-33F, mounted in PBS-glycerol containing paraphenylenediamine,and the same fields photographed previously with neutral red were identified and photographed by fluorescein epifluorescence.

Double Labeling of Human Fibroblasts with MC22-33F and Oil Red 0 Paraformaldehyde-fixed,permeabilized cell monolayerswere immunostained by MC22-33F, then washed in PBS and counterstained with oil red 0 (C.1. 26125; Merck) in isopropanollwater (Lillie, 1965) for 3 min.

Effect of Intracellular Accumulation of Aminoglycoside Antibiotics on MC22-33F Immunoreactivity in Cultured ROS Cells ROS cells were cultured in 2 % fetal calf serum in RPMI medium with or without 1 mglml kanamycin. After 3 or 5 days, the cultures were fixed with paraformaldehyde, permeabilized, immunostained by MC22-33F, and counterstained by oil red 0.

Enzyme-linked Immunosorbent Assays (ELISAs) Egg PC (Type V-E), egg PE (Type III), cardiolipin from bovine heart, and sphingomyelin(SM) from bovine brain were purchased from Sigma. Each phospholipid (1 mg) was deposited from organic solvent as a thin film on the bottom of a glass tube with a stream of nitrogen. When lipid mixtures were used, the components were mixed in 200 pl chloroform before evaporation. Residual solvent was removed by drying in high vacuum for 1 hr. One milliliter of PBS was added to the dry lipid and dispersion of the lipid was achieved by vortex mixing


A MONOCLONAL ANTIBODY

To CHOLINE PHOSPHOLIPIDS

ELISAs were performed essentially according to Hazeltine et al. (1988) with modifications.Polystyrene%-well culture plates (Nu&, Roskilde, Denmark) were coated with 50 pl/well of either PC, PE, cardiolipin, or SM or a mixture of PC and PE (1/1:w/w) or PC and cardiolipin (1/1:w/w). The lipids were suspended in PBS at a concentration of 50 pg/ml. The coating took place for 20 hr at 4'C. The coated plates were washed three times with PBS at 24'C, then saturated with PBS containing 15% fetal calf serum and 0.3% gelatin for 2 hr at 24°C to block any free binding sites. The MC22-33F hybridoma supernatant was then added (50 pl/well) and incubated for 3 hr at 24'C. The plates were washed with three changes of PBS containing 0.4% bovine serum albumin (PBS-BSA) at 24'C. then incubated with peroxidase-conjugatedanti-rat IgG + IgM (Jackson)diluted 1:lOOO in PBS-BSA for 20 hr at 4'C (75 pl/well). The unbound antibody was removed by washing in PBS-BSA and the bound peroxidase measured using 0-phenylenediamine as substrate (200 pl/well:lO mg of 0-phenylenediamine dissolved in 1 ml methanol, diluted to 100 ml with distilled H2O and 10 p1 of 30% H202 added before use). After appropriate color development (usually 45 min) the reaction was terminated by addition of 50 pl and the optical density measured at 492 nm. In one experiment we used a simplified detection procedure for immobilized phospholipids. PC or mixtures of PC and cardiolipin in PBS were incubated for 16 hr at 4'C in 35-mm plastic petri dishes (Nunc). After washing, the immobilized lipids were stained by indirect immunofluorescence using the protocol already described for cells and tissues. As a control of the unspecific binding of rat IgMs to phospholipids and polystyrene, we used a rat hybridoma cell supernatant containing a monoclonal rat anti-vimentin IgM, MC22-14A(Mark et al., unpublished study).

Immunodetection of Phosphoaminolzpids Afier Chromatography on Thin-layer Plates The PC analogues 1,2-dipalmitoyl-sn-glycero-3 phosphodimethylethanolamine (dimethyl-PE, DMPE) and 1,2-dipalmiroyl-sn-glycero-3 phosphomethylethanolamine(monomethyl-PE,MMPE) were purchased from Fluka (Buchs, Switzerland). PC, SM, DMPE, and MMPE were applied to silica gel 60 thin-layerplates (Merck).After migration, the lipids were immunostained using a 1:lOO dilution of the MC22-33F-containinghybridoma supernatant in PBS as described by Portoukalian and Bouchon (1986). A duplicate of the immunostained chromatogram was stained with the phosphorus-specificspray reagent of Dituner and Lester (Portoukalian et al., 1978).

Immunoelectron Microscopy Procedures Pre-embedding Method. Subconfluent ROS cell cultures were washed with PBS for 10 min at 37'C, then mechanicallyremoved from the culture surface with a rubber policeman. The cell suspension was centrifuged for 1 min in an Eppendorf tube containing a 10% gelatin solution in PBS. The cell pellet in gelatin-PBS was ice-cooled for 10 min, then cut into small pieces. These were immersed in 0.1 M sodium phosphate-buffered 4% paraformaldehyde and 0.25% glutaraldehyde solution (pH 7.4) for 6 hr at 4 ° C then washed in PBS (three times for 10 min), treated with 50 mM N&CI in PBS (16 hr, 4°C) and equilibrated in PBS containing 5 % , lo%, 15%, 20% sucmse for 2 hr, 2 hr, 6 hr, and 16 hr, respectively. Cell pellets were embedded in Tissue-Tek 4583 (Miles; Elkhart, IN) and snap-frozen in liquid nitrogen. Fifteen-micrometer cryosections were mounted on poly-L-lysine-coatedglass slides. After rehydration in PBS, sections were pre-treated with 0.5% metaperiodic acid PBS (10 min. 24'C) and freshly prepared 0.05% sodium borohydridein PBS (5 min. 24°C)(Wil-

829

lingham, 1983; Weber et al., 1978). After rinsing in PBS, sections were incubated with MCZZ-33F(12 hr, 4'C), then with the peroxidase-labeledsecondary antibody against rat IgM (diluted 1:40 in PBS) for 12 hr at 4'C. Sections were washed in PBS and refixed in 2 % glutaraldehyde in PBS for 20 min at 24°C. After rinsing, the sections were incubated for 20 min at 24% in a Hanker-Yates solution, freshly prepared by dissolving 75 mg of Hanker-Yates powder (Sigma) in 100 ml of 0.1 M Xis-HC1 buffer (pH 7.6) containing 1% dimethylsulfoxide (DMSO). The sections were developed (twice for 10 min) in the same solution without DMSO (to which was added 0.01% H202 before use). Sections were post-fixed for 1 hr at 24% with 1% osmium tetroxide in PBS, dehydrated with graded acetone, and flat-embedded in Epon 812 using upside-down hinged capsules (Mark et al., 1987).After polymerization of the resin (48 hr at 60%). proper portions were trimmed and reembedded in Epon 812 using a silicon rubber plate. Ultra-thin sections were cut with a diamond knife on a Reichert-Jung ultratome. Sections with slight uranyl acetate counterstaining were examined using a Siemens Elmiskop 102 electron microscope at an acceleratingvoltage of 60 kV. As negative control, non-immune rat IgM (Jackson)was substituted for the primary anti body.

Cell Surface Labeling. KB cell cultures on glass slides were fixed with 4% paraformaldehyde and 0.25% glutaraldehyde in PBS for 1 hr at 4'C. After washing with PBS containing NH4CI (1 hr, 24'C). cell monolayers were incubated with MC22-33F for 1 hr at 24°C. After rinsing, cells were incubated with biotinylatedanti-mouseIgM (140; Vector Labs, Burlingame, CA) for 1 hr, then labeled with 20-nm streptavidin-conjugated colloidal gold particles (diluted 1:30; Gibco BRL, Cergy Pontoise, France) for 1 hr at 24-C. As negative controls: (a) cell monolayers were pre-treated with phospholipase C (1 Ulml PBS) for 1 hr at 37°C or (b) non-immune rat IgM was substituted for the primary antibody. After washing away the excess of colloidal gold particles, sections were re-fixed in 2 % glutaraldehyde in PBS for 20 min. then post-fixed for 1 hr with 1% osmium tetroxide in PBS, dehydrated with graded acetone, and flat-embedded in Epon. After polymerization ofthe resin, proper portions were trimmed and re-embedded. Ultra-thin sections were obtained as described above.

Results The data are presented following the sequence of experiments that were undertaken with the aim of characterizing the MAb epitopes, then of gaining information on the significance of the immunocytochemical labeling.

Immunofluorescence Localization of the MC22-33F Epitope in Cultured Cells: Efects of Fktives, Permeabilization Conditions, and Enzyme Treatments on Antigen Preservation The immunostaining patterns in unfixed and paraformaldehydefived cells were identical. Without permeabilization, some dental epithelial cells and KB cells demonstrated punctate labeling in the plane of the plasma membrane (Figure 1). Unpermeabilized ROS cells and unpermeabilized human embryonic fibroblasts did not react with the MAb (not shown). In all cell types, the intracel-


830

MARK, TSUJI, PORTOUKALIAN. REBBAA, ZIDAN, RUCH

t

I I

Figure 1. Immunofluorescence staining by MC22-33F of mouse dental epithelial cells in primary culture. The cells were fixed with paraformaldehyde but not permeabilized. The staining of the cell surface is very heterogeneous and many cells were unlabeled (0). Original magnification x 400. Bar = 30 m. Figure 2. An overview of an ROS cell culture immunostained with (a,b) MC22-33F or (c) non-immune rat IgM. The cells were fixed with paraformaldehyde,then permeabilized with Triton X-100. (a) A group of cells demonstrating large intracellular immunofluorescent aggregates. (b) Corresponding phase-contrast micrograph. Original magnification x 400. Ear = 30 wm. Figure 3. Different aspects of the cytoplasmic immunofluorescence staining pattern with MC22-33F in ROS cells. Cells were fixed in 3% paraformaldehyde,then permeabilized. (a) Small discrete bodies are distributed throughout the cytoplasm, although most are usually clustered around the nucleus (n). (b) Spherical and doughnut-like (arrows) cytoplasmic structures. (c) Honeycomb-shaped aggregates in a large cell. The first aspect (a) is the most frequent. However, the three aspects might coexist in a single cell. Original magnification x 1200. Ear = 10 pm.


83 1

A MONOCLONAL ANTIBODY TO CHOLINE PHOSPHOLIPIDS

I I

Figure 4. Comparison of the intracellular immunostaining patterns with MC22-33F of ROS cells fixed either with acetone or with paraformaldehydeand ethanol post-fixation. Cells were plated at the same density and cultured for 3 days, fixed, then treated with Triton X-100 and immunostained in parallel. (a) After fixation with acetone the staining pattern appears reticular instead of punctate. (b) Cells fixed with ethanol demonstrate a diffuse pattern of staining and the intensity of immunofluorescence is considerably reduced. The two photographs were taken using exactly the same exposure conditions. Note that acetone fixation enhances the autofluorescence of nuclei. Original magnification x 1000. Bar = 10 pm. Figure 5. Immunofluorescencestaining by MC22-33F of subconfluent cultures of ROS cells. Before incubation with the MAb, unfixed cultures were treated with: (a) 1% Triton X-100-0.5% deoxycholate; (b) 1% Triton X-100-0.5% deoxycholate containing 0.6 M KCI; or 0.2% Triton X-100followed by (c) pronase or (d) phospholipase C digestion. lmmunolabeling of the cells is unaffected by high detergent concentrations, is increased by KCI and proteases, but is abolished by phospholipase C. Original magnifications: a-c = 400; d x 1000. Bars: a-c = 30 pm;d = 10 pm. Figure 6. Immunofluorescence staining by (a,b) MC2233F or (c) MC2214A of purified phospholipids immobilized on plastic petri dishes. (a&) PC 0.5 mglml. (b) PC 0.5 mg/ml plus cardiolipin 0.5 mglml. Original magnification x 1000. Bar = 10 pm.


832

M a r immunostaining observed after permeabilization was polymorphic, occurring mostly in punctate structures (distributed throughout the cytoplasm) but also in large elements u p to IO-pm (Figures 2a. 2b. and 3a-3c). Control cultures that were incubated with non-immune rat IgM instead of MC22-33F were negative (see Figure 2c). ROS cells were also futed with acetone or with ethanol, then treated with 0.2% Triton X-100 and immunostained with MC2233F. Acetone fiwtion caused considerable disturbance of the staining pattern but increased the staining intensity (Figure 4a). After ethanol fixation the cell labeling became extremely weak (not shown). Treatment of paraformaldehyde-fixed cells with ethanol caused the same dramatic decrease of antigenicity (Figure 4b). To determine suitable conditions for antigen extraction from the cell monolayer, unfixed ROS cells were treated with detergent solutions of differing concentrations, pH. and ionic strength before immunostaining. Treatment of the ROS cells with 0.6 M KCI during (or after) permeabilization caused an apparent increase in the staining intensity. by inducing the formation of large intracellular aggregates (compare Figure 5b with Figure 5a). With this exception. the detergent extraction procedures listed in Materials and Methods caused no significant change in the immunostaining pattern or intensity (see Figure 5a). The cell residues left on the glass coverslips after trypsin or pronase digestion of unfixed, permeabilized ROS cells were devoid of recognizable nuclei when examined under phase-contrast optics. However, they still reacted intensely with MC22-33F (Figure 5c). O n the other hand, digestion of unfixed, permeabilized cells with phospholipase C before incubation with the MAb abolished the immunostaining (Figure 5d).

Effect of Pre-incubation of MC22-33F with Choline Phospholipids on Its Ability to Bind to Cell Monolayers Binding of the MAb to cultured cells was abolished when the hybridoma supernatant was first incubated with an excess of choline phospholipids, but was unaffected by pre-incubation with phosphorylcholine.

Anti-phospholipid ELISAs and Immunochemical Detection of Phospholipids on Thin-layer Chmmatograms We have analyzed the binding of MC22-33F to lipid-coated microtiter plates. The results are summarized in Figures 6 and 7. Of the phosholipids tested in this assay, only PC and SM reacted positively, but, unexpectedly, mixtures of PC with cardiolipin were not recognized by the MAb. Under identical experimental conditions, Hazeltine et al. (1988) and Rauch and Janoff (1990) have reported efficient immobilitation to polystyrene of cardiolipin and of PE. However, binding of the different phospholipid species to plastic is difficult to quantitate. The reactivity of MC22-33F towards different phosphoaminolipids, including synthetic PC analogues with modified polar head groups, was further investigated on thin-layer chromatograms

a b C

d e f 9 h i

i 2.0

1.0

0

OD492 Figure 7. Reactivity of MC22-33Ftowards different phospholipids in an ELIassay. Microtiter plates were coated with 2 5 pg of (a) PC. (b)SM. (c) PE, and (d) cardiolipin (CL). (e$) The plates were coated with lipid mixtures consisring of PC (2.5 pg) plus PE (2.5 pg) or PC (25 pg) plus CL (25 pg). respectively. (9-1) The immobilized PC (2.5 wg) was digested with 50 pl of TPCK trypsin (20 ug/ml in PES).with pronase (20 pglml in PES)or with phospholipaseC (1 Ulml in PES), respectively, for 1 hr at 37% before incubation with the MAb. (I) The microtiter plate was coated with PC (2.5 pg) but MC22-33F was replaced by MC22-14A.

(Figures 8 and 9). The order of the reactivity in this assay was DMPE > SM > PC (Figure 9). MMPE (Figure 8) and PE (not shown) were not recognized by the MAb.

Combined Immunofluorescence and Histochemical Labelings Vital staining by neutral red allowed the co-localization of MC2233Fpositive structures with sites of neutral red uptake (compare Figure loa. neutral red uptake, with Figure lob, immunostaining by MC22-33F). Double staining by MC22-33F and oil red 0 revealed that the cytoplasmic areas containing large immunofluorescent vesicles also contained most of the neutral fat droplets (Figures l l a and llb). Furthermore. the centers of the immunofluorescent

- MMPE *

8a

._

-DMPE -PC

8b

Figure 8. Thin-layer chromatography of 1 of MMPE. DMPE. and PC in chloroform-methanol-water (60:35:8). followed by immunoperoxidasestaining by (a) MC22-33F or (b) staining by the reagent of D i m ” and Lester.


833


Figure 9. Thin-layer chromatography of a mixture of DMPE, PC, and SM (molecular ratios 1:l:l) in chloroform-methanol-water (55:45:10) followed by (a) immunoperoxidase staining by MC22-33F or (b) staining by the reagent of Dittmer and Lester. Bottom line: total amount of phospholipids (pg) deposited in the silica gel.

-DMPE PC

-

-SM

0.2

0.5 1.0 2-0 5.0

MC22-33F-labeled “rings” (Figures 3b and lla) were found to consist of neutral lipids (Figure llb). ROS cells cultured for 3 days in the presence of kanamycin did not differ significantly from control cells: both demonstrated an obvious decrease in the frequency of neutral fat bodies and of MC2233F-positivevesicles (not shown). The cells that survived a 5-day exposure to kanamycin were loaded with phospholipid inclusions (compare Figures 12a and 12b) and with large neutral fat droplets (Figures 12c and 12d).

Im munoelectron Microscopy KB cells exhibited many microvilli and blebs at their dorsal surface (Figure 13a). After colloidal gold surface labeling, the MC22-33F epitope was detected exclusively on the surface of some blebs (Figures 13b and 13c). This staining might correspond to the punctate labeling provided by indirect immunofluorescence (Figure 1). KB cells pre-incubated with phospholipase C were devoid of immunogold labeling (not shown). To identify the intracellular antigenic sites recognized by MC2233F, a pre-embedding technique was employed. In ROS cells, only membrane-like structures within lysosomes stained positively (Figures 14a and 14b).

Immunohistochemicul Loculizution of Phospholipids in Frozen Mouse Tissues with MC22-33F Many cell types and tissues were not stained by the MAb, e.g., epidermal cells, intestinal epithelial cells, dermal fibroblasts, spleen, myometrium, hyaline cartilage, and myelin sheets. The highest levels of immunoreactive phospholipids were detected in oil red 0-positive cell types, e.g., pre-adipocytes (Figures 15a-15d) multilocular adipocytes (not shown), cells of the ovarian stroma and of the theca interna (Figures 16a and 16b). granulosa cells (Figures I6a and 16b), and endometrial cells (Figures 17a and 17b). An exception was the adrenal cortex (not shown), which failed to react with the MAb. The cardiac muscle fibers, which in mouse contain many lipid droplets (see Dreibelbis, 1986), strongly reacted with MC22-33F (Figure 18). In calcifying cartilage, the MAb detected extracellular phospholipids in the septa surrounding hypertrophic chondrocytes (Figures 19a and 19b). Tissue sections incubated with non-immune rat IgM instead of MC22-33F were negative (see Figure 1%).

0.2

0.5 1.0 2.0 5.0

Discussion This study reports on the characterization of an MAb to choline phospholipids and on its use as a probe to detect these compounds in cell cultures as well as in tissues and organs. Several attempts have been made to raise cell type-specific or cell lineage-specific MAb using isolated cells as immunogens (Perry et al., 1990; Tsai et al., 1990; Zidan and Ruch, 1989). This approach gives rise to a large number of MAb of undefined epitope specificities. Our data exemplify the value of immunocytochemical technics to screen these MAb specificities (see also Mark et al., 1989; Willingham et al., 1987; Rapraeger et al., 1986). The binding of MC22-33F to cell monolayers has many unique features which, together, allowed identification of the epitope. This binding was unaffected by trypsin or pronase treatment or by prolonged exposure to high concentrationsof non-ionic detergents. Moreover, the MC2233F epitope was found to be stable over a wide range of pH and ionic strengths, properties that are uncommon for a peptide epitope in either soluble, polymerized, or membrane-incorporated forms. On the other hand, the binding was impaired by post-fixation with ethanol of paraformaldehyde-fixed cells, whereas acetone fixation of cell monolayers had the opposite effect. Since solubilization by ethanol and precipitation by acetone are properties of many phospholipids, this prompted us to orient the search towards a lipid epitope. Incubation of cell monolayers with phospholipase C resulted in inactivation of the antigen. Conversely, pre-incubation of the hybridoma supernatant with multilamellar vesicles prepared either from a crude extract of egg phospholipids or from highly purified egg PC abolished the binding of the MAb to the cell cultures. PC, SM, and DMPE, but not MMPE and PE, were recognized by the MAb in ELISA assays and/or on thin-layer chromatograms, suggesting that the methyl groups on the quaternary nitrogen of the choline moiety of the phospholipid were important for binding. Lipid mixtures composed of PC and cardiolipin did not react with MC22-33F. This indicates that the epitope might be masked on interaction with other phospholipids in vitro and perhaps also in vivo (see below). As already noted for other anti-PC IgMs (Nam et al., 1990), phosphorylcholine, the putative water-soluble hapten, did not impair the binding of MC22-33F to cell monolayers. It is well established that IgM antibodies are usually of low affinity. and it has been stated that this relatively low affinity mighc be compensated by multivalency (Goding, 1986). Therefore, it seems likely that MC22-33F might bind only to multiple, appropriately spaced phosphorylcholine residues in lipid bilayers.


834

MARK, TSUJI. PORTOUKALIAN, REBBAA. ZIDAN, RUCH

.. .

*

.-

12d Figure 10. Co-localizationof the MC22-33Fepitopewith neutral red uptake in human fibroblasts. (a) Cellswere incubated with neutralred in PBS and photographed under visible light. The cells were subsequently permeabilized, then incubated with MC22-33F. followed by fluorescein-conjugated anti-rat IgM. (b) The same cells were then rephotographedwith fluorescein epifluorescence. Arrows show examples of coincidence of neutral red staining and MC22-33Ffluorescent staining. Original magnification x 820. Bar = 10 vm. Figure 11. (a,b) Co-localization of MC22-33F-positivelarge cytoplasmic inclusionswith neutral fat droplets in a human fibroblast. Cells were immunostained by the MAb. washed in PBS, then counterstained with oil red 0. MC22-33F-positiveinclusion and neutral fat droplets are lying aside but are not superimposed. The arrows point to doughnut-shaped elements of which the cores consist of neutral fats. Original magnification x 1000.Bar = 10 pm. Figure 12. Effect of kanamycin on immunostainingby MC22-33F and neutral fat accumulation in ROS cell cultures. ROS cells were cultured for 5 days in 2% serum with (a.c,d)or (b) without the antibiotic. (a) Kanamycin induces accumulationof phospholipiddroplets in the cytoplasm of ROS cells, whereas (b) culturing the cells in 2% serum alone results in a considerable decrease of the immunostaining. (c.d) Higher magnification of examples of kanamycin-treated ROS Cells that have been double stained by MC22-33F and oil red 0. Original magnifications: a,b x 400; c,d x 1000. Bars: a.b = 30 pm; c.d = 10 vm.


835


Figure 13. lmmunodetectionof the MC2233F epitope at the surface of KB cells. (a) Scanning electron micrograph showing a typical aspect of a KB cell surface with many microvilli and blebs. (b) Transmission electron micrograph of KB cells; immunogold labelingwith MC2233F. Colloidal gold particles are localized on the surface of some blebs but are absent from the smooth cell surface and from the microvilli. (c) Negative control of the immunogold procedure. MC22-33F was replaced by non-immune rat IgM. Original magnification x 40,000. Bar = 0.25 pm. Figure 14. Transmission electron micrograph of ROS cells immunostained using the preembedding method. (a) Membranelike structures within lysosomes react with MC22-33F (arrows). (b) Similar structures (arrows) in a control employing non-immune rat IgM. Original magnification x 40,000. Bar = 0.25 pm.

F-‘

14a

Phospholipids are ubiquitous. The apparent selectivity of the immunostaining of cells in vitro and in situ is likely,to rely largely on binding characteristics of IgM antibodies and on phenomena of antigen “masking” rather than on differences in epitope affinities. DMPE is a metabolic intermediate of the methylation pathway of PC formation, but this pathway is of no quantitative significance in animal tissues (except in liver; for review see Kent, 1991); SM represents only a small percentage of the phospholipids in cells. PC, on the contrary, accounts for about half of the phospholipids in cells and thus represents the best of the three phospholipid candidates for binding to MC22-33F in cells. Synthetic multilamellar vesicules consisting of pure PC are labeled by indirect immunofluorescence employing MC22-33F. The possible dependence of MC2233F binding on fatty acid composition of PC, which has been reported for other anti-phospholipid antibodies (Nam et al., 1990; Kawagushi, 1987), might also explain part of the variable immunostaining. Plasma membrane phospholipids bound to the MAb only at sites of bleb formation, suggesting that these structures might display a molecular organization distinct from the rest of the plasma membrane (see Hale and Wurthier, 1987). Inside the cells, detection of phospholipids was restricted to vesicles of varying size and shape. At the electron microscopic level, these immunostained vesicles appeared as lysosomes containing membrane-like structures. Double labeling of the cells with MC22-33F and neutral red or oil red 0 demonstrated the frequent co-distribution of immunofluorescent structures with lysosomes and with neutral fat droplets. Aminoglycoside antibiotics, among which is kanamycin, are concentrated exclusivelyin lysosomes from cultured cells (Tulkens and Trouet, 1978) and inhibit lysosomal phospholipase C, sphingomyelinase, and acid lipases (Oshima et al., 1986; Hostetler and Hall,

-

A

14b

1982; Laurent et al., 1982;Tulkens and Van Hoof, 1980). This results in phospholipid accumulation and eventual formation of osmiophilic multilamellar bodies (Oshima et al., 1986). Accumulation proceeds slowly, and stable intracellular contents of aminoglycosides are obtained only after 4 days of culture (Tulkens and Trouet, 1978). Both rate and level of accumulation of kanamycin are increased on lowering the serum concentration of the culture medium (Tulkens and Trouet, 1978). Long-term exposure of ROS cells to kanamycin induced an intracytoplasmic accumulation of both phospholipid-containing and neutral fat-containing inclusions, indicating the existence of a functional link between these structures and glycerolipid catabolism. In newborn and adult mouse tissues, immunolabeled phospholipids were most abundant in cells known to store large amounts of triglycerides (e.g., adipocytes, cardiac muscle cells, endometrial cells) andlor cholesterol esters (e.g., steroid-producing cells). This seems to indicate that phospholipids are usually associated with cytoplasmic accumulation of neutral fats. Unexpectedly, immunoreactive phospholipids were also abundant in calcifying cartilage septa. Immunohistochemistry at the electron microscopic level will permit investigation of the link between these extracellular phospholipids and matrix vesicles. These latter structures, which are generated by plasma membrane blebbing of hypertrophic chondrocytes (Hale and Wurthier, 1987) might be involved in the initiation of mineralization (Boskey, 1985). In conclusion, the MAb MC22-33F, specific for choline phospholipids, appears as a new and interesting probe to investigate lipid accumulation in situ under physiological and pathological conditions.

Acknowledgments We thank Drs E Klein and G.Rebel for useful advice.


836

MARK, TSUJI, PORTOUKALIAN, REBBAA. ZIDAN, RUCH



Literature Cited Bonilla E. Prelle A (1987): Application of Nile blue and Nile red, two fluorescent probes, for detection of lipid droplets in human skeletal muscle. J Histochem Cytochem 35:619 Boskey AL (1981): Overview of cellular elements and macromoleculesimplicated in the initiation ofmineralization. In Butler WT, ed. The chemistry and biology of mineralized tissues. Birmingham, Ebsco Media, 331 Coulombe PA, Kan FWK, Bendayan M (1988): Introduction of a high resolution cytochemical method for studying the distribution of phospholipids in biological tissues. Eur J Cell Biol 46:564 Dreibelbis D, ed (1986): Bloom and Fawcett. A textbook ofhistology. 11th ed. Philadelphia, Saunders Fowler SD. Greenspan P (1981): Application of Nile red, a fluorescent hydrophobic probe, for the detection of neutral lipid deposits in tissue sections: comparison with oil red 0. J Histochem Cytochem 33:833 Goding JW (1986): Monoclonal antibodies: principles and practice. 2nd ed. London, Academic Press Greenspan P, Mayer EP, Fowler SD (1985): Nile red: a selective fluorescent stain for intracellular lipid droplets. J Cell Biol 100:965 Hale JE, Wuthier RE (1987): The mechanism of matrix vesicle formation. J Biol Chem 262:1916 Hazeltine M, Rauch J, Danoff D, EsdaileJM, Tannenbaum H (1988): Anriphospholipid antibodies in systemic lupus erythematosus: evidence for an association with positive Coombs and hypocomplementemia.J Rheumatol 15:80 Hostetler KY, Hall LB (1982): Inhibition of kidney lysosomal phospholipases A and C by aminoglycoside antibiotics: possible mechanism of aminoglycoside toxicity. Proc Natl Acad Sci USA 79:1663 Kawagushi S (1987): Phospholipid epitopes for mouse antibodies against bromelain-treated mouse erythrocytes. Immunology 62:ll Kent C (1991): Regulation of phosphatidylcholine biosynthesis. Prog Lipid Res 29237 Laurent G, Carlier MB, Rollman B, Van Hoof F, Tulkens P (1982): Mechanism of aminoglycoside-inducedlysosomal phospholipidosis: in vitro and in vivo studies with gentamicin and amikacin. Biochem Pharmacol31:3861 Lesot H, MeyerJM, Karcher-Djuricic V,Fabre M, RuchJV (1984): Behaviour of odontogenic epithelial cells in primary culture. J Craniofac Genet Dev Biol 4 2 2 1 Lillie RD (1965):Histopathologic technic and practical histochemistry. 3rd ed. New York, McGraw-Hill Mark MP. Butler WT, RuchJV (1989): Transient expression of a chondroi-

837

tin sulfate-related epitope during cartilage histomorphogenesisin the axial skeleton of fetal rats. Dev Biol 133:475 Mark MP, Prince CW, Gay S, Austin RL, Bhown M, Finkelman RD, Butler WT (1987): A comparative immunocytochemical study on the subcellular distributions of 44 kDa bone phosphoprotein and bone y-carboxyglutamic acid (G1a)-containing protein in osteoblasts. J Bone Miner Res 2:337 Nam KS, Igarashi K, Umeda M, Inoue K (1990): Production and characterization of monoclonal antibodies that specifically bind to phosphatidylcholine. Biochim Biophys Acta 1046239 Oshima M, Hashiguchi M, Shindo N, Shibata S (1986): Biochemical mechanisms of aminoglycosidecell toxicity. I. The uptake of gentamicin by cultured skin fibroblasts and the alteration of lysosomal enzyme activities.J Biochem 100:1575 Pearse AGE (1968): Histochemistry. Vol. 1. 3rd ed. London, Churchill Perry J, Gilligan M, Green E, Docherty J, Heath D (1990): Monoclonal antibodies to ROS 17/2.8 cell recognize antigens, some of which are restricted to osteoblasts and chondrocytes.J Bone Miner Res 5:187 PortoukalianJ, Bouchon B (1986): Hydrolysis of all gangliosidesincluding GM1 and GM2, on thin layer plates by Vibrio cholerae neuraminidase.J Chromatogr 380:386 PortoukalianJ, Meister R, Zwingelstein G (1978): Improved two dimensional solvent system for thin layer chromatographicanalysis of polar lipids on silica gel 60 precoated plates. J Chromatogr 152:169 Rapraeger A, Jalkanen M, Bernfield M (1986): Cell surface proteoglycan associateswith the cytoskeleton at the basolateral cell surface of mouse mammary epithelial cells. J Cell Biol 103:2683 Rauch J, Janoff AS (1990): Phospholipid in the hexagonal I1 phase is immunogenic: evidence for immunorecognition of nonbilayer lipid phases in vivo. Proc Natl Acad Sci USA 87412 RauchJ, Tannenbaum M, Tannenbaum H, Ramelson H. Cullis PR, Tilcock CPS, Hope MJ, Janoff AS (1986): Human hybridoma lupus anticoagulants distinguish between lamellar and hexagonal phase lipid systems.J Biol Chem 261:9672 %ai CC, McGuire MH, Melitt RJ, Ritter I11 RA, Xu J, Litwicki DJ, Rood-

man ST (1990): Monoclonal antibodies to human osteosarcoma: a novel Mr 26,000 protein recognized by murine hybridoma T-MMR2. Cancer Res 50:152 Tulkens P, Trouet A (1978): The uptake and intracellular accumulation of aminoglycoside antibiotics in lysosomes of cultured rat fibroblasts. Biochem Pharmacol 27:415 Tulkens P, Van Hoof F (1980): Comparative toxicity of aminoglycosideantibiotics towards the lysosomes in a cell culture model. Toxicology 17:195

4

Figure 15. Fat island from subcutaneous tissue of (a,b) a 1-day-oldand (c,d)a 2-day-old mouse, immunostained with MC22-33F (a,c). b and d are phase-contrast micrographs of a and c, respectively. The micrographs were taken from similar regions in the two animals. (a) High magnification of pre-adipocytes loaded with immunoreactive phospholipids. Note the absence of refringent lipid droplets in b. n, nuclei. (c.d) An overview of a fat island 24 hr later; immunostaining is confined to the peripheral cytoplasm of the differentiating adipocytes but absent from neutral fat droplets (I).c, capillaries. Original magnifications: a,b x 1000; c,d x 400. Bars: a,b = 10 pm; c,d = 30 pm. Figure 16. Mouse Graafian follicle. Double staining with (a) MC22-33F and (b) oil red 0. lmmunostaining is preponderant in cells of the theca interna (1) and of the cortical stroma (s)which accumulate neutral fats. The granulosa cells (9) are also immunostained. a, antrum. Original magnifications: a x 150; b x 100. Bars = 100 pm. Figure 17. Transverse section through the ampulla of the uterine tuba from an adult mouse. (a) Intense immunolabeling by MC22-33F is observed in the basal portion of the epithelial cells (e). The muscularis (m) is negative. (b) A similar section stained with oil red 0. I, lumen of the tube. Original magnifications: a x 320; b x 100. Bars: a = 30 pm; b = 100 pm. Figure 18. Myocardium from an adult mouse. Strong immunofluorescencestaining by MC22-33F is observed at the periphery of the muscle fibers. Original magnification x 320. Bar = 30 pm. Figure 19. Mandibular condyle from a I-day-old mouse. (a) Immunofluorescencestaining by MC22-33F, demonstrating phospholipid within the calcified cartilage septa (arrows). (b)Phase-contrast of view of a. (c) Section consecutive to a. MC22-33F was replaced by non-immune rat IgM in the immunostaining sequence. Original magnification x 400. Bar = 30 pm.


MARK, TSUJI, PORTOUKALIAN, REBBAA, ZIDAN, RUCH

838

Umeda M, Itarashi K, Nam KS, Inoue K (1989): Effective production of monoclonal antibodies against phosphatidylserine: recognition of phosphatidylserine by monoclonal antibody. J Immunol 143:2273 Voelker DR (1990): Lipid transport pathways in mammalian cells: Experientia 46:570 Weber K, Rathke PC, Osborn M (1978):Cytoplasmic microtubular images in glutaraldehyde-fixed tissue culture cells by electron microscopy and immunofluorescence microscopy. Proc Natl Acad Sci USA 75:1820

the GBS (glutaraldehyde-borohydride-saponin) procedure. J Histochem Cytochem 31:791 Willingham MC, Richert ND, Rutherford AV (1987): A novel fibrillar structure in cultured cells detected by a monoclonal antibody. Exp Cell Res 171:284 Wolman M (1964): Handbuch der Histochemie. Vol 5 . Stuttgart, Gustav Fisher Verlag. Zidan G, RuchJV (1989):Production of monoclonal antibodiesagainst mouse molar papilla cells. Int J Dev Biol 33:245

Willingham MC (1983): An alternative fixation processing method for preembedding ultrastructural immunochemistry of cytoplasmicantigens:


Characterization of an antiglycoprotein Ib monoclonal antibody that specifically inhibits platelet-thrombin interaction.

Isolation and Characterization of a Monobody with a Fibronectin Domain III Scaffold That Specifically Binds EphA2.

Isolation of an Aptamer that Binds Specifically to E. coli.

A cytotoxic monoclonal anti-leukemia antibody binds to histone H1.

Production and characterization of monoclonal antibodies that specifically bind to phosphatidylcholine.

Development of a monoclonal anti-ADAMTS-5 antibody that specifically blocks the interaction with LRP1.

A monoclonal antibody (8H3) that binds to rat T lineage cells and augments in vitro proliferative responses.

Characterization of [3H]acifluorfen binding to purified pea etioplasts, and evidence that protoporphyrinogen oxidase specifically binds acifluorfen.

Com, the phage Mu mom translational activator, is a zinc-binding protein that binds specifically to its cognate mRNA.

Monoclonal antibody Ki-67: its use in histopathology.

Generation and characterization of a human monoclonal antibody that neutralizes diverse HIV-1 isolates in vitro.

A monoclonal antibody to scopolamine and its use for competitive enzyme-linked immunosorbent assay.

Monoclonal antibodies to human endothelin-1: characterization and utilization in radioimmunoassay and immunocytochemistry.

Characterization of a neutralizing monoclonal antibody to Pasteurella haemolytica leukotoxin.

Generation and characterization of ixekizumab, a humanized monoclonal antibody that neutralizes interleukin-17A.

Cysteinylation of a monoclonal antibody leads to its inactivation.

Characterization of a monoclonal antibody that selectively recognizes a subset of Entamoeba histolytica isolates.

Drosophila transcriptional repressor protein that binds specifically to negative control elements in fat body enhancers.

A monoclonal antibody directed against the sporozoite stage of Plasmodium vivax binds to liver parenchymal cells.

A monoclonal anti-peptide antibody reacting with the insulin receptor beta-subunit. Characterization of the antibody and its epitope and use in immunoaffinity purification of intact receptors.

A monoclonal antibody binds to threonine 49 in the non-structural 1 protein of influenza A virus and interferes with its ability to modulate viral replication.

A protein which specifically binds to single stranded TTAGGGn repeats.

A conserved phosphoprotein that specifically binds nuclear localization sequences is involved in nuclear import.

Arabidopsis florigen FT binds to diurnally oscillating phospholipids that accelerate flowering.