- Email: [email protected]
Atomic force microscopy to study direct neurite–mast cell (RBL) communication in vitro
Immunology Letters 74 (2000) 211 – 214
www.elsevier.com/locate/
Atomic force microscopy to study direct neurite–mast cell (RBL) communication in vitro Hiroyuki Ohshiro, Ryo Suzuki, Tadahide Furuno, Mamoru Nakanishi* Faculty of Pharmaceutical Sciences, Nagoya City Uni6ersity, Tanabe-dori, Mizuho-ku, Nagoya 467 -8603, Japan Received 27 April 2000; accepted 14 June 2000
Abstract Communication between nerves and mast cells is a prototypic demonstration of neuroimmune interaction. We used an in vitro co-culture approach comprising cultured murine superior cervical ganglia (SCG) and rat basophilic leukemia (RBL-2H3) cells. Atomic force microscopy (AFM) showed how neurites attached to a pseudopodium or a cell body of an RBL cell. After stimulation of SCG neurites with bradykinin or scorpion venom, RBL cells attached to neurites spread and flattened, and several discharged granules (0.5–1.0 mm in diameter) were found on the surface of the RBL cells. A neurokinin (NK)-1 receptor (i.e. substance P receptor) antagonist prevented the RBL degranulation. The results showed that activation of the SCG neurites with bradykinin or scorpion venom was able to elicit degranulation in RBL cells which were attached to neurites. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Neuroimmune interaction; Atomic force microscopy; Superior cervical ganglia; RBL-2H3 cells; Degranulation
1. Introduction During the last decade there has been an exponential increase in data illustrating that the nervous and immune systems are not disparate entities [1,2]. The nerve–mast cell relationship has served as a prototypic association and has provided substantial evidence for bidirectional communication between nerves and immune cells [3]. Early studies elegantly described the non-random spatial association of nerves and mast cells in a variety of tissues in which actual membrane –membrane contacts could be observed [4,5]. To understand these events we have recently studied direct neurite– mast cell communication using an in vitro co-culture approach and calcium imaging by confocal laser scanning microscopy (CLSM). Our results showed clearly that nerve–mast cell cross-talk can occur in the absence of an intermediary transducing cell and that the neuropeptide substance P, operating via NK-1 receptors, is
an important mediator of this communication [6]. Our findings have implications for the neuroimmune signaling cascades that are likely to occur during airways inflammation, intestinal hepersensitivity, and other conditions in which mast cells feature [6]. In the present paper we have studied by atomic force microscopy (AFM) the degranulation process in RBL cells (homologous to the mucosal mast cell), which were involved in neurite–RBL communication. AFM has advantages close to electron microscopy (EM) in spatial resolution, in addition, it is able to apply to specimens (cells) in nearly intact conditions [7–13]. Thus, it seemed that AFM was very useful to study direct neurite–mast cell communication in vitro. The results gave new information that direct neurite–mast cell cross-talk induced degranulation in RBL cells.
2. Materials and methods
2.1. Ner6e –RBL cell co-culture * Corresponding author. Tel.: +81-52-8363411; fax: + 81-528363414. E-mail address: [email protected] (M. Nakanishi).
Nerve–mast cell (RBL-2H3) co-culture was done in the same way in the previous paper [6]. Following a
0165-2478/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 2 4 7 8 ( 0 0 ) 0 0 2 1 1 - X
212
H. Ohshiro et al. / Immunology Letters 74 (2000) 211–214
published protocol [14,15], superior cervical ganglia (SCG) were dissected from newborn (0 – 48 h old) CBA mice (Japan SLC, Shizuoka) and rinsed in Hanks’ balanced salt solution containing 10 mM HEPES (pH 7.4). Details of the experimental procedure has been described in our previous paper [6]. For AFM measurements, a collagen (type I; Wako, Osaka)-coated polystyrene dish (Corning) was used instead of a matrigel-coated dish [6]. For co-culture experiments, RBL cells (5×104 cells per dish) were added to 48-h-old cultures of SCG neurites and incubated at 37°C for 72 h.
2.2. Cellular acti6ation Calcium mobilization was used as an index of cellular activation and was assessed by CLSM with Fluo-3 [6,16,17]. After 72 h of co-culture, cells were treated with culture medium containing Fluo-3-AM (2.5 mM for 30 min at 37°C; Molecular Probes, Eugene, OR) and then washed with HEPES buffer. Cells were observed with a CLSM (Zeiss, LSM-410; argon ion laser at 488 nm) and their images were captured and analyzed using IBM compatible computer software [6]. Neurite activation was evoked by either bradykinin (100 nM; Bachem, Bubendorf, Switzerland) or scorpion venom (Leiurus quinquestriatus herbaeus, 100 pg/ ml; Sigma). In the experiment of an NK-1 receptor antagonist, WIN51, 708 (50 mg/ml; Research Biochemicals International) was added to neurite –RBL co-cultures for 10 min before stimulation.
2.3. AFM measurements AFM was done using a NanoScope IIIa D-3000 (Digital Instruments). Samples for AFM measurements were prepared by spreading neurite – RBL coculture cells on the collagen-coated polystyrene dish and were fixed for 30 min with 2% glutaraldehyde in 200 mM So¨rensen’s phosphate buffered solution (pH7.4) at an appropriate time (20 min) after stimulation with bradykinin (100 nM) or scorpion venom (100 pg/ml) at 37°C. They were dried in air. A silicon cantilever tip (Nanosensors; NCH-10T, 33 N/m) was used for the experiments. AFM images were captured by recording feedback signals under constant tapping forces (1–10 nN). Scanning frequency was 0.8 Hz and acquisition points were 512×512. Here, AFM images were collected as 16-bit data and were presented with 256 pseudocolor steps [10 – 12].
3. Results The collagen-coated dish was used to measure a tapping mode AFM image of neurite – RBL communi-
cation in vitro [15]. We found that Ca2 + mobilization in neurite–RBL co-culture on the collagen-coated dish was induced as much as on the matrigel-coated dish. Addition of bradykinin, or scorpion venom, elicited neurite activation (Ca2 + mobilization) and, after a lag period, Ca2 + mobilization occurred in RBL cells. The experimental results correlated well with the previous data of neurite–RBL co-culture on the matrigelcoated dishes [6]. Then, we measured the cell surface structures of neurite–RBL association on the collagen-coated dishes by using AFM. We performed AFM measurements on dozens of neurite–RBL samples. A typical example of AFM images of the neurite–RBL communication is shown in Fig. 1, where SCG neurites extended on the surface of the dishes and attached to a cell body of an RBL cell with filopodium (an arrow head in Fig. 1A) [14]. A thin neurite with the growth cone extended over a fiber-like pseudopodium. An enlarged AFM image of the attached area between the growth cone and the pseudopodium is shown in Fig. 1B. Contact between the growth cone and the RBL cell occurred over about 7 mm. The result correlated well with the previous experimental results by EM [14]. The AFM image showed that the neurite–RBL association occurred in nearly intact conditions. A resting RBL cell in the neurite–RBL association had a spherical and nearly smooth cell body and a thin pseudopodium as shown in Fig. 1A. When SCG neurites were stimulated with bradykinin (100 nM), RBL cells spread and flattened, and several discharged granules existed on the cell surface of the RBL cell attached to neurites in Fig. 2. The percentage of RBLs responding to bradykinin stimulation on the collagen-coated dishes was 56 (n= 16) in degranulation. The diameter of the discharged granules was 0.5–1.0 mm. The size of the discharged granules in RBL cells correlated well with the previous reports for RBL cells stimulated with antigen [9,10]. However, the number of the discharged granules was much less than that stimulated with antigen [9,10]. This may be due to the local stimulation of RBL cells through neurites. Such kinds of morphological changes were not observed in RBL cells non-contacted with neurites (0%, n= 25). In the control experiments we stimulated RBL cells alone with bradykinin, however, there were no direct effects on RBL cells (0%, n=36). In addition, morphological changes and degranulation in RBL cells were mostly blocked in the presence of a NK-1 receptor (i.e. substance P receptor) antagonist (WIN51,708) (15%, n= 13). We further found that scorpion venom (100 pg/ml) also induced spreading and degranulation of RBL cells attached to SCG neurites and that the NK1 receptor antagonist blocked them.
H. Ohshiro et al. / Immunology Letters 74 (2000) 211–214
4. Discussion Considerable evidence exists for a consistent anatomical association between mast cells and nerves in tissues throughout the body [4]. The morphological juxtaposition of mast cells and nerves would, by itself, be of little interest if it were not for the evidence of physiological connectedness. Antigen activation of mast cells results in the release of a variety of effector molecules (e.g. arachidonic acid metabolites, 5-hydroxytryptamine, histamine, cytokines) that can modulate neuronal activity.
213
In the reciprocal communication pathway, in vitro studies have revealed that neuronal-derived messengers, such as neuropeptides, can evoke or regulate mast cell activation and/or degranulation in a tissue-, and species-dependent manner. Thus, while it was widely accepted that mast cell–nerve bidirectional communication occurs on a regular basis, it is still not clear if an intermediary cell is required to facilitate this functional cross-talk [6]. Using an in vitro model, we recently examined neurite–mast cell (RBL) units and have demonstrated that activation of a nerve fiber
Fig. 1. AFM images of SCG neurite–RBL interaction. AFM images were captured after neurite – RBL cells were fixed with 2% glutaraldehyde solution as described in Section 2. (A) An AFM image of neurites attached to a pseudopodium and a cell body of an RBL cell. A thin filopodium (an arrow head) of the neurite attached to a cell body of the RBL cell. A white bar is 10 mm. (B) An enlarged AFM image of a white square in Fig. 1A. A thin neurite with the growth cone extened over a fiber-like pseudopodium of an RBL cell. Association between the growth cone and the RBL cell occurred over 7 mm. Fig. 2. A typical example of AFM images of a bradykinin-stimulated neurite – RBL communication. The image was captured at 20 min after stimulation with bradykinin (100 nM). (A) An AFM image of a bradykinin-stimulated RBL cell. A white bar is 10 mm. (B) An enlarged AFM image of a white square in Fig. 2A. Several discharged granules were observed along the neurite.
214
H. Ohshiro et al. / Immunology Letters 74 (2000) 211–214
elicits an activation event (as indicated by Ca2 + mobilization) in an RBL in contact with the activated neurite. In the previous experiment we showed that both bradykinin and scorpion venom evoked neuronal activation, and that neither agent had any direct effect on RBL Ca2 + mobilization. Further, addition of either agent to SCG–RBL co-cultures resulted in RBL activation that was always preceded by neurite activation, as gauged by Ca2 + imaging. CLSM showed that increased RBL Ca2 + mobilization was invariably accompanied by membrane ruffling and an increase in cell diameter; morphological features associated with activation and/ or subsequent degranulation. However, CLSM images were not enough to describe details of the morphological changes of neurite – RBL association and the degranulation from the RBL cell after stimulation with bradykinin or scorpion venom. However, AFM images in the present experiment gave precise information for the morphological features of the interaction between neurites and RBL cells in nearly intact conditions. We are sure that AFM is more suitable than transmission electron microscopy and scanning electron microscopy to study the neurite– mast cell communication in vitro.
Acknowledgements We thank Professor Bienenstock and Dr McKay (both from McMaster University) for their valuable support and also for technical and intellectual input.
References [1] C.A. Ottaway, Gastroenterol. Clin. North Am. 20 (1991) 511– 529. [2] N. Cohen, J.A. Moynihan, R. Ader, Int. Arch. Allergy Immunol. 105 (1994) 101 – 106. [3] D.M. McKay, J. Bienenstock, Immunol. Today 15 (1994) 533– 538. [4] R.H. Stead, J. Bienenstock, in: M.M. Burger, B. Sordat, R.M. Zinkernagel (Eds.), Cell to Cell Interaction, Karger, Basel, 1990, pp. 170 – 187. [5] R.H. Stead, M. Tomioka, G. Quinonez, G.T. Simon, S.Y. Felten, J. Bienenstock, Proc. Natl. Acad. Sci. USA 84 (1987) 2975–2979. [6] R. Suzuki, T. Furuno, D.M. McKay, D. Wolvers, R. Teshima, M. Nakanishi, J. Bienenstock, J. Immunol. 163 (1999) 2410–2415. [7] G. Binnig, C.F. Quate, C. Gerber, Phys. Rev. Lett. 3 (1986) 930 – 933. [8] R. Guckenberger, M.H. Heim, G. Ceve, H.F. Knapp, W. Wiegrabe, A. Hidebrand, Science 266 (1994) 1538 – 1540. [9] A. Spudich, D. Braunstein, Proc. Natl. Acad. Sci. USA 92 (1995) 6976 – 6980. [10] R. Nakamura, T. Furuno, M. Nakanishi, Bioimages 6 (1998) 69 – 75. [11] R. Nakamura, M. Nakanishi, Immunol. Lett. 69 (1999) 307–310. [12] C. Kawaura, A. Noguchi, T. Furuno, M. Nakanishi, FEBS Lett. 421 (1998) 69 – 72. [13] A. Ikai, Surf. Sci. Rep. 26 (1996) 261 – 332. [14] M.G. Blennerhassett, M. Tomioka, J. Bienenstock, Cell Tissue Res. 265 (1991) 121 – 128. [15] M.G. Blennerhassett, J. Bienenstock, Neurosci. Lett. 120 (1990) 50 – 54. [16] T. Furuno, T. Hamano, M. Nakanishi, Biophys. J. 64 (1993) 665 – 669. [17] K. Nakato, T. Furuno, K. Inagaki, R. Teshima, T. Terao, M. Nakanishi, Eur. J. Biochem. 209 (1992) 745 – 749.
. .
www.elsevier.com/locate/
Atomic force microscopy to study direct neurite–mast cell (RBL) communication in vitro Hiroyuki Ohshiro, Ryo Suzuki, Tadahide Furuno, Mamoru Nakanishi* Faculty of Pharmaceutical Sciences, Nagoya City Uni6ersity, Tanabe-dori, Mizuho-ku, Nagoya 467 -8603, Japan Received 27 April 2000; accepted 14 June 2000
Abstract Communication between nerves and mast cells is a prototypic demonstration of neuroimmune interaction. We used an in vitro co-culture approach comprising cultured murine superior cervical ganglia (SCG) and rat basophilic leukemia (RBL-2H3) cells. Atomic force microscopy (AFM) showed how neurites attached to a pseudopodium or a cell body of an RBL cell. After stimulation of SCG neurites with bradykinin or scorpion venom, RBL cells attached to neurites spread and flattened, and several discharged granules (0.5–1.0 mm in diameter) were found on the surface of the RBL cells. A neurokinin (NK)-1 receptor (i.e. substance P receptor) antagonist prevented the RBL degranulation. The results showed that activation of the SCG neurites with bradykinin or scorpion venom was able to elicit degranulation in RBL cells which were attached to neurites. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Neuroimmune interaction; Atomic force microscopy; Superior cervical ganglia; RBL-2H3 cells; Degranulation
1. Introduction During the last decade there has been an exponential increase in data illustrating that the nervous and immune systems are not disparate entities [1,2]. The nerve–mast cell relationship has served as a prototypic association and has provided substantial evidence for bidirectional communication between nerves and immune cells [3]. Early studies elegantly described the non-random spatial association of nerves and mast cells in a variety of tissues in which actual membrane –membrane contacts could be observed [4,5]. To understand these events we have recently studied direct neurite– mast cell communication using an in vitro co-culture approach and calcium imaging by confocal laser scanning microscopy (CLSM). Our results showed clearly that nerve–mast cell cross-talk can occur in the absence of an intermediary transducing cell and that the neuropeptide substance P, operating via NK-1 receptors, is
an important mediator of this communication [6]. Our findings have implications for the neuroimmune signaling cascades that are likely to occur during airways inflammation, intestinal hepersensitivity, and other conditions in which mast cells feature [6]. In the present paper we have studied by atomic force microscopy (AFM) the degranulation process in RBL cells (homologous to the mucosal mast cell), which were involved in neurite–RBL communication. AFM has advantages close to electron microscopy (EM) in spatial resolution, in addition, it is able to apply to specimens (cells) in nearly intact conditions [7–13]. Thus, it seemed that AFM was very useful to study direct neurite–mast cell communication in vitro. The results gave new information that direct neurite–mast cell cross-talk induced degranulation in RBL cells.
2. Materials and methods
2.1. Ner6e –RBL cell co-culture * Corresponding author. Tel.: +81-52-8363411; fax: + 81-528363414. E-mail address: [email protected] (M. Nakanishi).
Nerve–mast cell (RBL-2H3) co-culture was done in the same way in the previous paper [6]. Following a
0165-2478/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 2 4 7 8 ( 0 0 ) 0 0 2 1 1 - X
212
H. Ohshiro et al. / Immunology Letters 74 (2000) 211–214
published protocol [14,15], superior cervical ganglia (SCG) were dissected from newborn (0 – 48 h old) CBA mice (Japan SLC, Shizuoka) and rinsed in Hanks’ balanced salt solution containing 10 mM HEPES (pH 7.4). Details of the experimental procedure has been described in our previous paper [6]. For AFM measurements, a collagen (type I; Wako, Osaka)-coated polystyrene dish (Corning) was used instead of a matrigel-coated dish [6]. For co-culture experiments, RBL cells (5×104 cells per dish) were added to 48-h-old cultures of SCG neurites and incubated at 37°C for 72 h.
2.2. Cellular acti6ation Calcium mobilization was used as an index of cellular activation and was assessed by CLSM with Fluo-3 [6,16,17]. After 72 h of co-culture, cells were treated with culture medium containing Fluo-3-AM (2.5 mM for 30 min at 37°C; Molecular Probes, Eugene, OR) and then washed with HEPES buffer. Cells were observed with a CLSM (Zeiss, LSM-410; argon ion laser at 488 nm) and their images were captured and analyzed using IBM compatible computer software [6]. Neurite activation was evoked by either bradykinin (100 nM; Bachem, Bubendorf, Switzerland) or scorpion venom (Leiurus quinquestriatus herbaeus, 100 pg/ ml; Sigma). In the experiment of an NK-1 receptor antagonist, WIN51, 708 (50 mg/ml; Research Biochemicals International) was added to neurite –RBL co-cultures for 10 min before stimulation.
2.3. AFM measurements AFM was done using a NanoScope IIIa D-3000 (Digital Instruments). Samples for AFM measurements were prepared by spreading neurite – RBL coculture cells on the collagen-coated polystyrene dish and were fixed for 30 min with 2% glutaraldehyde in 200 mM So¨rensen’s phosphate buffered solution (pH7.4) at an appropriate time (20 min) after stimulation with bradykinin (100 nM) or scorpion venom (100 pg/ml) at 37°C. They were dried in air. A silicon cantilever tip (Nanosensors; NCH-10T, 33 N/m) was used for the experiments. AFM images were captured by recording feedback signals under constant tapping forces (1–10 nN). Scanning frequency was 0.8 Hz and acquisition points were 512×512. Here, AFM images were collected as 16-bit data and were presented with 256 pseudocolor steps [10 – 12].
3. Results The collagen-coated dish was used to measure a tapping mode AFM image of neurite – RBL communi-
cation in vitro [15]. We found that Ca2 + mobilization in neurite–RBL co-culture on the collagen-coated dish was induced as much as on the matrigel-coated dish. Addition of bradykinin, or scorpion venom, elicited neurite activation (Ca2 + mobilization) and, after a lag period, Ca2 + mobilization occurred in RBL cells. The experimental results correlated well with the previous data of neurite–RBL co-culture on the matrigelcoated dishes [6]. Then, we measured the cell surface structures of neurite–RBL association on the collagen-coated dishes by using AFM. We performed AFM measurements on dozens of neurite–RBL samples. A typical example of AFM images of the neurite–RBL communication is shown in Fig. 1, where SCG neurites extended on the surface of the dishes and attached to a cell body of an RBL cell with filopodium (an arrow head in Fig. 1A) [14]. A thin neurite with the growth cone extended over a fiber-like pseudopodium. An enlarged AFM image of the attached area between the growth cone and the pseudopodium is shown in Fig. 1B. Contact between the growth cone and the RBL cell occurred over about 7 mm. The result correlated well with the previous experimental results by EM [14]. The AFM image showed that the neurite–RBL association occurred in nearly intact conditions. A resting RBL cell in the neurite–RBL association had a spherical and nearly smooth cell body and a thin pseudopodium as shown in Fig. 1A. When SCG neurites were stimulated with bradykinin (100 nM), RBL cells spread and flattened, and several discharged granules existed on the cell surface of the RBL cell attached to neurites in Fig. 2. The percentage of RBLs responding to bradykinin stimulation on the collagen-coated dishes was 56 (n= 16) in degranulation. The diameter of the discharged granules was 0.5–1.0 mm. The size of the discharged granules in RBL cells correlated well with the previous reports for RBL cells stimulated with antigen [9,10]. However, the number of the discharged granules was much less than that stimulated with antigen [9,10]. This may be due to the local stimulation of RBL cells through neurites. Such kinds of morphological changes were not observed in RBL cells non-contacted with neurites (0%, n= 25). In the control experiments we stimulated RBL cells alone with bradykinin, however, there were no direct effects on RBL cells (0%, n=36). In addition, morphological changes and degranulation in RBL cells were mostly blocked in the presence of a NK-1 receptor (i.e. substance P receptor) antagonist (WIN51,708) (15%, n= 13). We further found that scorpion venom (100 pg/ml) also induced spreading and degranulation of RBL cells attached to SCG neurites and that the NK1 receptor antagonist blocked them.
H. Ohshiro et al. / Immunology Letters 74 (2000) 211–214
4. Discussion Considerable evidence exists for a consistent anatomical association between mast cells and nerves in tissues throughout the body [4]. The morphological juxtaposition of mast cells and nerves would, by itself, be of little interest if it were not for the evidence of physiological connectedness. Antigen activation of mast cells results in the release of a variety of effector molecules (e.g. arachidonic acid metabolites, 5-hydroxytryptamine, histamine, cytokines) that can modulate neuronal activity.
213
In the reciprocal communication pathway, in vitro studies have revealed that neuronal-derived messengers, such as neuropeptides, can evoke or regulate mast cell activation and/or degranulation in a tissue-, and species-dependent manner. Thus, while it was widely accepted that mast cell–nerve bidirectional communication occurs on a regular basis, it is still not clear if an intermediary cell is required to facilitate this functional cross-talk [6]. Using an in vitro model, we recently examined neurite–mast cell (RBL) units and have demonstrated that activation of a nerve fiber
Fig. 1. AFM images of SCG neurite–RBL interaction. AFM images were captured after neurite – RBL cells were fixed with 2% glutaraldehyde solution as described in Section 2. (A) An AFM image of neurites attached to a pseudopodium and a cell body of an RBL cell. A thin filopodium (an arrow head) of the neurite attached to a cell body of the RBL cell. A white bar is 10 mm. (B) An enlarged AFM image of a white square in Fig. 1A. A thin neurite with the growth cone extened over a fiber-like pseudopodium of an RBL cell. Association between the growth cone and the RBL cell occurred over 7 mm. Fig. 2. A typical example of AFM images of a bradykinin-stimulated neurite – RBL communication. The image was captured at 20 min after stimulation with bradykinin (100 nM). (A) An AFM image of a bradykinin-stimulated RBL cell. A white bar is 10 mm. (B) An enlarged AFM image of a white square in Fig. 2A. Several discharged granules were observed along the neurite.
214
H. Ohshiro et al. / Immunology Letters 74 (2000) 211–214
elicits an activation event (as indicated by Ca2 + mobilization) in an RBL in contact with the activated neurite. In the previous experiment we showed that both bradykinin and scorpion venom evoked neuronal activation, and that neither agent had any direct effect on RBL Ca2 + mobilization. Further, addition of either agent to SCG–RBL co-cultures resulted in RBL activation that was always preceded by neurite activation, as gauged by Ca2 + imaging. CLSM showed that increased RBL Ca2 + mobilization was invariably accompanied by membrane ruffling and an increase in cell diameter; morphological features associated with activation and/ or subsequent degranulation. However, CLSM images were not enough to describe details of the morphological changes of neurite – RBL association and the degranulation from the RBL cell after stimulation with bradykinin or scorpion venom. However, AFM images in the present experiment gave precise information for the morphological features of the interaction between neurites and RBL cells in nearly intact conditions. We are sure that AFM is more suitable than transmission electron microscopy and scanning electron microscopy to study the neurite– mast cell communication in vitro.
Acknowledgements We thank Professor Bienenstock and Dr McKay (both from McMaster University) for their valuable support and also for technical and intellectual input.
References [1] C.A. Ottaway, Gastroenterol. Clin. North Am. 20 (1991) 511– 529. [2] N. Cohen, J.A. Moynihan, R. Ader, Int. Arch. Allergy Immunol. 105 (1994) 101 – 106. [3] D.M. McKay, J. Bienenstock, Immunol. Today 15 (1994) 533– 538. [4] R.H. Stead, J. Bienenstock, in: M.M. Burger, B. Sordat, R.M. Zinkernagel (Eds.), Cell to Cell Interaction, Karger, Basel, 1990, pp. 170 – 187. [5] R.H. Stead, M. Tomioka, G. Quinonez, G.T. Simon, S.Y. Felten, J. Bienenstock, Proc. Natl. Acad. Sci. USA 84 (1987) 2975–2979. [6] R. Suzuki, T. Furuno, D.M. McKay, D. Wolvers, R. Teshima, M. Nakanishi, J. Bienenstock, J. Immunol. 163 (1999) 2410–2415. [7] G. Binnig, C.F. Quate, C. Gerber, Phys. Rev. Lett. 3 (1986) 930 – 933. [8] R. Guckenberger, M.H. Heim, G. Ceve, H.F. Knapp, W. Wiegrabe, A. Hidebrand, Science 266 (1994) 1538 – 1540. [9] A. Spudich, D. Braunstein, Proc. Natl. Acad. Sci. USA 92 (1995) 6976 – 6980. [10] R. Nakamura, T. Furuno, M. Nakanishi, Bioimages 6 (1998) 69 – 75. [11] R. Nakamura, M. Nakanishi, Immunol. Lett. 69 (1999) 307–310. [12] C. Kawaura, A. Noguchi, T. Furuno, M. Nakanishi, FEBS Lett. 421 (1998) 69 – 72. [13] A. Ikai, Surf. Sci. Rep. 26 (1996) 261 – 332. [14] M.G. Blennerhassett, M. Tomioka, J. Bienenstock, Cell Tissue Res. 265 (1991) 121 – 128. [15] M.G. Blennerhassett, J. Bienenstock, Neurosci. Lett. 120 (1990) 50 – 54. [16] T. Furuno, T. Hamano, M. Nakanishi, Biophys. J. 64 (1993) 665 – 669. [17] K. Nakato, T. Furuno, K. Inagaki, R. Teshima, T. Terao, M. Nakanishi, Eur. J. Biochem. 209 (1992) 745 – 749.
. .