Journal of Immunological Methods, 42 (1981) 355--365 © Elsevier/North-Holland Biomedical Press
355
S I M U L T A N E O U S D E T E C T I O N O F TWO D I F F E R E N T C E L L S U R F A C E A N T I G E N S BY E L E C T R O N M I C R O S C O P Y BY M E A N S O F MULTIVALENT HYBRID ANTIBODY WITH DOUBLE SPECIFICITY
E. MANDACHE, GABRIELA MO~?A, ELENA MOLDOVEANU, I. MORARU and V. GHETIE Babe~ Institute, Spl. lndependentei 99, R-76201 Bucharest, Rumania
(Received 24 September 1980, accepted 16 December 1980)
Multivalent hybrid antibody (MHA) complexes with dual specificity were prepared by combining two antibodies of different specificities, one against ferritin (Fer), the other against horseradish peroxidase (HRP), with protein A of Staphylococcus aureus (SPA). Electron microscopy of mouse spleen lymphocytes and thymocytes (previously coated with mouse IgG anti-Thy-1 antibody) treated with IgG anti-Fer/SpA/IgG anti-mouse Ig complex and Fer gave better resolution and higher accuracy than previously obtained with IgG anti-HRP/SpA/IgG anti-mouse Ig complex and HRP (Mandache et al., 1980). Surface Thy-1 alloantigen and Fc receptor (charged with human IgG) treated with a mixture of IgG anti-Fer/SpA/IgG anti-Thy-1 and IgG anti-HRP/SpA/IgG anti-human Fab could be simultaneously detected on the thymocyte surface by either light or electron microscopy using Fer and HRP. The concomitant visualization of Thy-1 alloantigen (with Fer) and FcR (with HRP) on mouse thymocyte clearly shows that their distribution is largely independent and that the amount of Thy-1 antigen is greater. These results show that electron microscopy with a mixture of MHA is a useful technique for simultaneous location of two antigenic markers on the cell surface.
INTRODUCTION
Multivalent h y b r i d a n t i b o d y (MHA) c o m p l e x e s with dual specificity were p r e p a r e d b y c o m b i n i n g t w o a n t i b o d i e s with d i f f e r e n t specificities with p r o t e i n A o f S t a p h y l o c o c c u s aureus (SPA) (Ghe~ie and Mota, 1980; MoSa and Ghe~ie, 1 9 8 1 ) . Such soluble c o m p l e x e s consist o f 4 m o l e c u l e s o f IgG and 2 m o l e c u l e s o f SpA (mol. wt. 6 8 4 , 0 0 0 daltons; m o l e c u l a r f o r m u l a : (IgG antiA / S p A / I g G anti-B)2) ( M o t a et al., 1978; Mihaescu e t al., 1979). E l e c t r o n m i c r o s c o p y o f m o u s e spleen l y m p h o c y t e s t r e a t e d with (antip e r o x i d a s e / S p A / a n t i - m o u s e Ig)2 c o m p l e x e s and p e r o x i d a s e s h o w e d strong specific staining o f l y m p h o c y t e m e m b r a n e s ( M a n d a c h e e t al., 1980). This result d e m o n s t r a t e d t h a t M H A c o m p l e x e s c o n t a i n i n g anti-peroxidase antib o d y are useful reagents f o r d e t e c t i n g various antigenic m a r k e r s o n cell m e m brane by electron microscopy. Results r e p o r t e d .here s h o w t h a t t h e M H A staining p r o c e d u r e m a y also be used with anti-ferritin antibodies. F u r t h e r m o r e , b y using a m i x t u r e o f t w o
356 MHA complexes consisting of anti-peroxidase and anti-ferritin IgG linked by SpA to different antibodies directed against two membrane antigens, we were able to locate the latter simultaneously on the cell surface. MATERIALS AND METHODS Pro te in s
Protein A of Staphylococcus aureus (SPA) (Pharmacia, Uppsala), peroxidase from horse radish (HRP) (BDH Chemicals Ltd.), ferritin from horse spleen (Fer) (Fluka AG Buchs SG), mouse Ig (mIg) (obtained by preparative electrophoresis), human IgG (hIgG) (Kabi, Stockholm) and its Fab fragments (prepared by papain digestion and chromatography on SpA-Sepharose 4B) were used. Antibodies
Rabbit anti-HRP serum was kindly supplied by Dr. Doina Onic~ (Babe~ Institute, Bucharest). Rabbit anti-mIg, anti-Fer and anti-Fab fragment were obtained by repeated injections of these antigens. Rabbit anti-Thy-1 serum was kindly supplied by Dr. A. Sulica (Babe~ Institute, Bucharest). IgG was isolated from these sera by affinity chromatography on SpA-Sepharose 4B (Hjelm et al., 1972). Purified antibodies were obtained from IgG anti-Fer, IgG anti-HRP and IgG anti-Fab by adsorption on glutaraldehyde insolubilized antigens (Avrameas, 1969) and elution in 0.2 M acetic acid. Cells
Spleen lymphocytes and thymocytes were isolated from CBA mice. The lymphocytes were subsequently purified by the Ficoll-sodium metrizoate technique (BSyum, 1968). Cells were washed and suspended in IC-65 medium (equivalent to TC-199 medium) (Cantacuzino Institute, Bucharest). Sheep red blood cells were coated with HRP (EHRP), Fer (EFer), Fab fragment (EFab) and SpA (ES) by the chromium chloride method (Gold and Fudenberg, 1969; Ghel;ie et al., 1975). Ferritin was also bound to chicken red blood cells by the same technique. Some EFer samples were labelled with fluorescein isothiocyanate (Serva, Heidelberg) as described by MSller (1974). Immunological analysis
Microagglutination of protein-coated sheep red blood cells and mouse thymocytes was performed in microtitre plates as previously described (Mih~escu et al., 1979). The titres were expressed as /ag MHA/ml of the highest dilution giving a positive reaction. Rosetting was carried out on cells treated with various amounts of MHA complexes, using as indicator cells a mixture of EHRP and EFer.
357
Preparation o f MHA with double specificity The preparation of hybrid anti-Fer/SpA/anti-mIg, anti-Fer/SpA/antiThy-1 and anti-HRP/SpA/anti-Fab complexes was performed as described by Ghe~ie and Mona (1980) and Mandache et al. (1980). Cell treatment Mouse spleen lymphocytes (107) were suspended in 0.15 ml IC-65 medium containing 0.015 M sodium azide and 0.05 ml hybrid antibody solution (5--20 pg) was added. After 30 min incubation at 4°C, the cells were washed thrice with cold IC-65 medium containing sodium azide and resuspended to 107 cells/0.1 ml. To this suspension, 0.025 ml of HRP (10--20 pg) and Fer (25 #g) in phosphate-buffered saline was added and the mixture was incubated for 30 min at 4°C. After repeated washings the cells were pelleted by centrifugation. Mouse thymocytes and spleen lymphocytes (10 ~ cells/ml) were treated with 10 pl hIgG (1 mg). After 30 min incubation at 4°C the cells were repeatedly washed with cold IC-65 medium containing sodium azide, resuspended at 107 cells/0.15 ml and then processed as described above. Control experiments with cell samples not treated either with hybrid antibody complex or with hIgG were also performed. Ultrastructural processing The cell pellet was fixed for 5 min with 1.5% glutaraldehyde in 0.067 M sodium cacodylate buffer, pH 7.4. For uniform fixation the cells were resuspended in the fixing medium and centrifuged. The cell samples treated with MHA containing anti-Fer antibody were washed 3 times in sodium cacodylate buffer and postfixed with 1% OsO4 in the same buffer, pH 7.4, then washed, pelleted again, incorporated in 1% Difco agar, sampled and processed for embedding in Epon 812. All the cell samples treated with hybrid mixtures containing anti-Fer and anti-HRP antibodies were washed in 0.05 M Tris-HC1 buffer, pH 7.4 and incubated for 20 min at room temperature in a medium containing 0.05% 3.3'-diaminobenzidine and 0.02% H202 in 0.05 M Tris-HC1 buffer, pH 7.4, according to Graham and Karnovsky (1966). The cells were washed, postfixed in 1% OsO4 in sodium cacodylate buffer, pH 7.4, and again washed and pelleted. The pellet was incorporated in 1% Difco agar at 65°C and divided into 1 mm 3 fragments. These were dehydrated and embedded for transmission electron microscopy. RESULTS The ability of anti-Fer/SpA/anti-Thy-1 and anti-HRP/SpA/anti-Fab complexes to react with the cognate antigens was demonstrated by agglutination of thymocytes and erythrocytes coated with HRP, Fer, and Fab (Table 1),
358 TABLE 1 Agglutinating activity of multivalent hybrid antibody. Titre ( pg ligand/ml) with :
Ligands
Anti-Fer/SpA/anti-Thy Anti-HRP/SpA/anti-Fab Anti-Fer/SpA/anti-mIg
Thymocytes
EFer
EHRP
EFab
4.5 NA NA b
0.2 NA 0.4
NA a 0.1 NA
NA 0.8 NA
= no agglutination. The hybrid agglutinates mIg-bearing spleen lymphocytes.
a NA
b
As Table 1 shows the anti-Fer/SpA/anti-Thy-1 complex was able to react with both EFer and t h y m o c y t e s but n o t with EHRP and EFab. Similarly, the anti-HRP/SpA/anti-Fab complex reacted only with the corresponding antigens, EHRP and EFab. T h y m o c y t e s coated with hIgG and treated simultaneously with both MHA complexes were able to form mixed rosettes with EHRP and EFer as revealed by either fluorescent EFer or Fer bound to chicken red blood cells (Fig. 1). A
a
b
f
c
d
Fig. 1. Mixed rosettes with chicken EFer and sheep EHRP formed by mouse lymphoid cells carrying bound hIgG and treated with a mixture of hybrid antibody complexes consisting of anti-Fer/SpA/anti-Thy-1 and anti-HRP/SpA/anti-Fab, a, b = spleen lymphocytes; c, d = thymocytes.
359
Fig. 2. Mouse spleen lymphocyte treated with anti-Fer/SpA/anti-mIg showing labelled areas with ferritin in monolayer (vertical arrows). The horizontal arrow shows a labelled plasma membrane fragment attached to another marked segment. Nucleus (N), mitochondria (m). x75,000.
Fig. 1 shows t h a t Thy-1 antigen and Fc receptor (carrying bound hIgG) were simultaneously detected on the cell surface by the a t t a c h m e n t of both EFer (with chicken red blood cells) and EHRP (with sheep red blood cells). Electron microscopy of spleen l y m p h o c y t e s treated with anti-Fer/SpA/ anti-mIg and ferritin, showed in transmission electron microscopy a specific labelling (Fig. 2). Control preparations showed no membrane labelling. The labelling pattern consisted of electron dense grains situated in a monolayer on the outer aspect of the plasma membranes and separated from the latter by a narrow space of approximately three times the thickness of a cell membrane. This pattern was seen on cell membranes in longitudinal thin sections. When a thin section containing a labelled cell membrane cut obliquely was examined, the Fer molecules no longer appeared as a monolayer but were superimposed. The distribution of Fer grains was in groups having approximately the same density of molecules with slight variation depending on section thickness (Fig. 2). The Fer molecules were found in clusters and never alone, and appeared on thin sections as segments of labelled plasma membrane separated by marker-free zones. These labelled segments had a very similar distribution to those obtained using MHA with anti-HRP specificity (Mandache et al., 1980). The only difference was that the plasma membrane segments labelled with Fer were shorter and more numerous than the segments labelled with HRP. Thus, we often found short Fer-labelled segments, some situated quite close to each other (Fig. 2). We also observed
360
Fig. 3. Surface o f a spleen l y m p h o c y t e treated with anti-Fer/SpA/anti-mIg showing distinct ferritin-labelled areas, a: longitudinal section showing ferritin monolayer, b: nonlongitudinal incidence showing indistinct plasma membrane and dispersed ferritin molecules, c: veil-like cytoplasmic extension in process o f engulfing a labelled area. X70,000.
some labelled plasma membranes featuring e n d o c y t o t i c aspects, viz. cytoplasmic veils (Fig. 3), pits and vesicles. Thy-1 alloantigen was detected on mouse t h y m o c y t e s with an indirect technique, namely treatment of cells with mouse IgG anti-Thy-1 followed by the anti-Fer/SpA/anti-mIg hybrid antibody. The labelling pattern had features similar to those previously obtained with MHA containing anti-HRP antibody (Mandache et al., 1980). The treatment of hIgG charged t h y m o c y t e s with a mixture of MHA containing anti-Fer/SpA/anti-Thy-1 and anti-HRP/SpA/anti-Fab enabled us to visualize simultaneously on the same cell both Fc receptor (FcR) and Thy-1 alloantigen. These two membrane antigens were identified by two distinct electron microscopic markers, namely ferritin for Thy-1 and peroxidase for FcR. All the cells investigated showed rounded shapes with more or less rough surfaces. We found no cells showing uropods or real microvilli (Fig. 4). The Fer labelling consisted of electron dense grains in monolayers on the outer aspect of the plasma membrane more or less equally spaced at approx. 30 nm. The Fer molecules were always organized in groups, set in rows and close to each other. These rows of Fer molecules varied in length, never exceeding 0.5 pm, and were spread all over the t h y m o c y t e plasma membrane (Fig. 4). The cells studied showed between 9 and 12 such Fer-labelled groups with no special orientation and separated b y large Fer-free segments (Figs. 5 and 6a). Compared with Fer, the HRP labelling pattern was less extensive in terms of number and dimensions of labelled segments as seen on ultrasections (Fig. 4). The typical pattern of peroxidase labelling consisted of electron dense black layers situated on the outer face of the t h y m o c y t e plasma membrane (Figs. 5 and 6a). The thickness of these layers was somewhat
361
Fig. 4. IgG-carrying thymocyte treated with anti-Fer/SpA/anti-Thy-1 and anti-HRP/SpA/ anti-Fab showing a round shape, a pit-like endocytotic feature (End) and several areas labelled either with Fer (thin arrows) or with HRP (thick arrows). N = nucleus, x14,000.
greater than the distance between the plasma membrane and the labelling ferritin molecules. No more than three or four peroxidase labelled zones were f o u n d on one cell but they were up to 1 pm in length. The lack of any peroxidase activity product in t h y m o c y t e s reacted with anti-Fer/SpA/anti-Thy-1 and incubated in a DAB-containing medium showed t h a t ferritin had no peroxidase-like activity. Besides the two distinct kinds of labelled areas -- rows of Fer grains and electron dense black layers with HRP -- some t h y m o c y t e s showed both fine grains and the electron dense black layer in apposition (Fig. 6b). This third kind of labelling occurred with a lower frequency, usually showing small clusters of ferritin spread through a larger peroxidase-labelled zone (Fig. 6b).
Fig. 5. Mouse thymocyte previously charged with IgG and treated with anti-Fer/SpA/antiThy-l and anti-HRP/SpA/anti-Fab showing four Fer-labelled patches (thin arrows) and two HRP-labelled areas (thick arrows). X40,000.
Fig. 6. a: IgG-charged thymocyte treated with anti-Fer/SpA/anti-Thy-l and anti-HRP/ SPA/anti-Fab showing ferritin-labelled segments (thin arrows), peroxidase-labelled segments (thick arrows), and mixed (ferritin and peroxidsse) labelled zone (arrow-heads). ~35,000. b: detail of mixed labelled zone from (I showing the apposition of ferritin molecules and peroxidase (arrow-heads). ~50,000.
363 DISCUSSION The presence of antibody molecules with two different specificities in a single hybrid complex was demonstrated by the agglutination of both EFer and thymocytes by anti-Fer/SpA/anti-Thy-1 and of both EHRP and EFab by anti-HRP/SpA/anti-Fab. The ability of a mixture of these two MHA to attach simultaneously and independently to the cell surface was clearly shown by formation of mixed rosettes between hIgG charged mouse thymocytes or spleen lymphocytes and HRP-coated sheep red blood cells and Fer-coated chicken red blood cells. The presence of FcR on mouse thymocytes was demonstrated by their ability to bind hIgG either in heat-aggregated (Santana and Turk, 1975) or in monomeric form (Ghe~ie et al., 1976). Since the SpA molecule in the hybrid complex is able to react by gel diffusion with hIgG (Mona et al., 1978), it might possibly allow binding of both anti-HRP/SpA/anti-Fab and anti-Fer/SpA/anti-Thy-1 complexes to the FcR-attached hIgG, thus giving misleading results. However we have shown that hybrid complexes, at concentrations up to 25 ~g/107 cells, do not bind to hIgG through SpA since anti-Fer/SpA/anti-mIg is not able to attach to hIgG charged thymocytes, as is demonstrated by their failure to form EFer rosettes. The similar ultrastructural pattern of the membrane antigens on spleen lymphocytes (mIg) and thymocytes (Thy-1) visualized either with Fer binding MHA or with HRP binding MHA substantiates the validity of our electron microscopy technique using MHA labelling procedures (Mandache et al., 1980). The fact that the Fer-labelled patches are shorter and more numerous than those obtained with HRP labelling can be satisfactorily explained by the better resolution and higher accuracy of Fer labelling which permits more precise localization of surface antigens. Thus two or many close clusters were visualized as one single patch when labelled with HRP binding MHA owing to the HRP activity product and its possible diffusibility. On the other hand, visualization of Fer molecules requires higher magnification compared with HRP labelling which although it works very well at lower magnification, itself requires an additional technical step (the DAB reaction). Spleen lymphocytes treated with anti-Fer/SpA/anti-mIg hybrid complex and Fer definitely showed a pattern of redistributed mIg although the surface mIg molecules are uniformly distributed on undisturbed mouse spleen lymphocyte membrane (De Petris, 1977). The patchy pattern recorded by us, which is similar to that reported by Raft and De Petris (1973) with divalent Fer-labelled rabbit anti-mIg antibody, is the result of cross-linking and aggregation of mIg by MHA. The patchy pattern of mIg in our experimen t was not followed by the later phase of capping and uropod formation because of the low temperature conditions and the presence of sodium azide (De Petris, 1977). However, the process of endocytosis was still present as already reported (De Petris, 1977).
364 When simultaneous detection of FcR and Thy-1 antigen on mouse thymocyte was performed with a mixture of MHA, both antigenic markers showed a patch-like pattern which might have resulted from redistribution induced by the multivalency of the ligands. However the FcR patches on the cell surface might represent the natural mode of distribution, as suggested by An (1979) in the case of FcR-bearing human lymphocytes. The normal distribution of Thy-1 antigen may be essentially a uniform one (De Petris, 1977) or a mosaic of discrete patches of variable size (Aoki et al., 1969). C o n c o m i t a n t visualization of FcR and Thy-1 antigen on the mouse t h y m o c y t e clearly shows that their distribution is largely independent and that there is a greater a m o u n t of Thy-1 antigen present. The apposition of Fer and HRP markers on some segments of plasma membrane is not attributable to the binding through SpA of both hybrid complexes to FcR-bound hIgG since, as shown above, the SpA in the complex is unable to interact with FcR-bound hIgG, at least at the MHA concentrations used by us. The occurrence of mixed labelled areas may be explained by assuming that after interaction with the corresponding ligands, both FcR and Thy-1 antigen are incompletely segregated in the redistribution process. This explanation implies that on the one hand the antigens are initially mixed together on the cell surface, at least one of them (Thy-1) having a uniform distribution; and on the other hand, t h a t patch formation of both membrane antigens is not hindered by the simultaneous binding of two multivalent ligands. Thus the mixed Fer- and HRP-labelled areas may result from the counter-current flow of Thy-1 antigen molecules through clusters of FcR molecules. However the disparity between the size of the hybrid complex and the size of antigenic sites on cell membranes introduces some uncertainty as to the actual position of and the redistribution pattern of such antigenic markers. Our results show that simultaneous detection of different membrane antigenic markers can be achieved by means of a mixture of two multivalent antibody complexes able to bind distinct electron microscopic markers (i.e. Fer and HRP). This technique may be particularly useful in locating functionally related receptors and in characterizing l y m p h o c y t e subpopulations bearing two distinct antigenic markers. ACKNOWLEDGEMENTS This work was supported by grants from the Rumanian Academy of Medical Sciences. The technical assistance of Mrs. Mariana Caralicea and Mrs. Cornelia Popescu is greatly appreciated. REFERENCES An, T., 1979, Immunology 36,859. Aoki, T., U. H~/mmerling,E. De Harven, E.A. Boyse and J. Old, 1969, J. Exp. Med. 130, 979.
365 Avrameas, S., 1969, Immunochemistry 6,825. BSyum, A., 1968, J. Clin. Lab. Invest. 21 (Suppl. 97). De Petris, S., 1977, in: Dynamic Aspects of Cell Surface Organization, eds. G. Poste and G.L. Nicholson (Elsevier]North-Holland Biomedical Press, Amsterdam) p. 643. Ghe~ie, V. and G. Mona, 1980, Mol. Immunol. 17, 395. Ghe~ie, V., G. St~lenheim and J. SjSquist, 1975, Scand. J. Immunol. 4,471. Ghe~ie, V., I. Moraru, A. Sulica, M. Gherman and J. Sjbquist, 1976, Rev. Roum. Biochim. 13, 263. Gold, E.R. and H.H. Fudenberg, 1967, J. Immunol. 99, 859. Graham, R.C. and M.J. Karnovsky, 1966, J. Histochem. Cytochem. 14, 291. Hjelm, H., K. Hjelm and J. Sj6quist, 1972, FEBS Lett. 28, 73. Mandache, E., E. Moldoveanu, G. Mona, I. Moraru and V. Ghe~ie, 1980, J. Immunol. Methods 35, 33. Mih~escu, S., A. Sulica, J. Sjbquist and V. Ghe~ie, 1979, Rev. Roum. Biochim. 16, 57. M611er, G., 1974, J. Exp. Med. 139,969. Mona, G. and V. Ghe~ie, 1981, Mol. Immunol. 18, 91. Mona, G., V. Ghe];ie and J. Sjbquist, 1978, Immunochemistry 15,625. Raff, M.C. and S. De Petris, 1973, Fed. Proc. 32, 48. Santana, V. and J.L. Turk, 1975, Immunology 28, 1173.