- Email: [email protected]
Effect of proinflammatory cytokines on cultured spiral ligament fibrocytes
Hearing Research 137 (1999) 155^159 www.elsevier.com/locate/heares
E¡ect of proin£ammatory cytokines on cultured spiral ligament ¢brocytes Kazuhide Yoshida, Issei Ichimiya, Masashi Suzuki, Goro Mogi * Department of Otolaryngology, Oita Medical University, 1-1 Idaigaoka, Hasama-cho, Oita 879-5593, Japan Received 10 December 1998; received in revised form 27 July 1999; accepted 3 August 1999
Abstract To clarify the effect of proinflammatory cytokines on spiral ligament (SL) fibrocytes, in vitro studies were performed using secondary cell cultures. Cultures from murine SL fibrocytes were stimulated by interleukin (IL)-1L or tumor necrosis factor (TNF)K, and secretion of various mediators was measured by enzyme-linked immunosorbent assay. After stimulation with the proinflammatory cytokines, IL-6, TNF-K, monocyte chemoattractant protein-1, KC, macrophage inflammatory protein-2, soluble intercellular adhesion molecule-1, and vascular endothelial growth factor levels were elevated. Secretion of these chemokines and other mediators could induce inflammatory cell movement, which would prolong the inflammatory response, leading to fibrocyte damage. Given that SL fibrocytes may play a role in cochlear fluid and ion homeostasis, such fibrocyte disruption could cause cochlear malfunction. ß 1999 Elsevier Science B.V. All rights reserved. Key words: Mouse; Inner ear; Interleukin-1L; Tumor necrosis factor-K; Chemokine; Soluble intercellular adhesion molecule-1
1. Introduction Recent morphological studies suggest that the spiral ligament (SL) may play a role in cochlear £uid and ion homeostasis. Localization of Na/K-ATPase (Ichimiya et al., 1994a ; Schulte and Adams, 1989) and gap junctions (Kikuchi et al., 1995) in the cochlea suggests that cochlear lateral wall cells, including the SL ¢brocytes, provide the anatomical substrate for K ion recirculation from hair cells back to the endolymphatic space. This hypothesis is consistent with electrophysiological changes in the cochlea of animals with experimentally induced endolymphatic hydrops where connexin 26 gap junction protein immunostaining is reduced without other noticeable histological changes (Ichimiya et al., 1994b). Recently, we also reported reduced connexin 26 staining in the SL after labyrinthitis following injection of the protein antigen keyhole limpet hemocyanin (KLH) into the scala tympani of systemically sensitized
* Corresponding author. Tel.: +81 (97) 586-5913; Fax: +81 (97) 549-0762.
guinea pigs (Ichimiya et al., 1998). The ¢ndings of this previous study led us to speculate that SL ¢brocytes may be responsible for the cochlear malfunction associated with some immunologically linked inner ear diseases. Furthermore, reduced connexin 26 staining in the SL in animals with experimental pneumococcal otitis media suggests that some noxious substances that pass through the round window membrane may also a¡ect the SL ¢brocytes (Ichimiya et al., 1999). To characterize the SL ¢brocyte features from an immunological point of view, we performed an in vitro study using cultured ¢brocytes. Because immunological mechanisms have been elucidated in mice, we cultured mouse ¢brocytes and investigated the e¡ect of proin£ammatory cytokines on these ¢brocytes. 2. Materials and methods 2.1. Cell culture Inner ear cell cultures were established as described by Gratton et al. (1996) with minor modi¢cations. CBA/JN Crj mice (male, 6^8 weeks of age) were anes-
0378-5955 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 9 9 ) 0 0 1 3 4 - 3
HEARES 3308 25-10-99
156
K. Yoshida et al. / Hearing Research 137 (1999) 155^159
thetized deeply with intraperitoneal pentobarbital and decapitated. The temporal bones were excised quickly, immersed in ice-cold minimum essential medium-K (MEM-K, Life Technologies, Rockville, MD, USA), and the basal turn of the cochlear lateral wall was dissected. The lateral wall fragments were placed gently on a type I collagen-coated Petri dish (Becton Dickinson Labware, Franklin Lakes, NJ, USA). To facilitate attachment, one drop of the culture medium consisting of MEM-K with 10% fetal calf serum (FCS), 1U Antimycotic-antibiotic (Sigma Chemical, St. Louis, MO, USA), and 1% ITS-G supplement (Life Technologies) were placed on the tissue fragments, followed by a cover glass. The Petri dish was placed in the incubator (37³C; 5% CO2 , 95% air) with maximum humidity for 24 h. Then the cover glass was removed gently, the culture medium was added slowly, and the explants were incubated further. The culture medium was changed every 3 days after rinsing with washing bu¡er, consisting of sterile phosphate-bu¡ered saline (sPBS) with a 1UAntimycotic-antibiotic solution. After about 2 weeks in primary culture, nearly con£uent monolayer cells were rinsed with the washing bu¡er, and trypsin/EDTA (0.05% :0.02%) were added to promote cell detachment. Culture medium was added after a 7-min incubation, and the detached cells were centrifuged at 1000 rpm for 10 min. The cell pellet was resuspended and distributed to new £asks at a subculture ratio of 1:4. Secondary cultures were maintained the same as the primary cultures. 2.2. Immunocytochemistry As described in a separate manuscript (unpublished data), immunostaining for caldesmon, Na/K-ATPase, S-100, or cytokeratin was performed to classify the cultures. Upon con£uence, the cells were subcultured into 2-well chamber slides (Becton Dickinson Labware), rinsed in PBS, ¢xed in 10% neutral-bu¡ered formalin for 10 min, and dehydrated. The slides were rinsed in PBS and exposed to a 5% solution of normal horse, rabbit, or goat serum; then incubated with mouse anti-caldesmon (diluted 1:300, Sigma Chemical), chicken anti-Na/K-ATPase (diluted 1:30 000, Cortex Biochem, San Leandro, CA, USA), rabbit anti-S-100 (diluted 1:1000, Sigma Chemical), or rabbit anticytokeratin (ready-to-use solution, Zymed Laboratories, San Francisco, CA, USA). Sections were then £ooded with biotin-conjugated horse anti-mouse immunoglobulin G (IgG) (diluted 1:200, Vector Laboratories, Burlingame, CA, USA), rabbit anti-chicken IgG (diluted 1:4000, Zymed Laboratories), or goat anti-rabbit IgG (diluted 1:200, Vector Laboratories) as appropriate. Bound primary antibody sites were visualized by incubating with ABC reagent (Vector Laboratories) fol-
lowed with 0.05% 3,3P-diaminobenzidine-0.01% H2 O2 substrate medium in 0.1 M phosphate bu¡er. After counterstaining with Veronal acetate-bu¡ered 1% methyl green solution, the slides were mounted for observation. Staining to con¢rm antibody speci¢city was done as described above, except that PBS was used in place of the primary antibodies. To identify the cell type of the culture, para¤n-embedded cochlea sections from normal mice were processed at the same step, and the staining patterns of the cochlear lateral wall were compared with the immunostaining pro¢le of the cultured cells. 2.3. Cytokine stimulation studies Secondary cultures were grown in six-well type I collagen-coated £at-bottom plates (5U104 cells/well, Becton Dickinson Labware) in the culture medium described above. The con£uent monolayer cells were incubated with 2 ml of culture medium for 24 h alone as controls or in the presence of the proin£ammatory cytokine, interleukin (IL)-1L or tumor necrosis factor (TNF)-K (each at 1 or 10 ng/ml; RpD Systems, Minneapolis, MN, USA). Conditioned medium was analyzed with the following enzyme-linked immunosorbent assay kits: IL-1L, IL-2, IL-4, IL-6, IL-10, IL-13, interferon (IFN)-Q, TNF-K, granulocyte-macrophage colony-stimulating factor (GM-CSF), macrophage in£ammatory protein (MIP)-1a, MIP-2, KC, vascular endothelial growth factor (VEGF) (RpD Systems), IL-3, IL-5, soluble intercellular adhesion molecule (sICAM)-1 (Endgen, Woburn, MA, USA), and chemoattractant protein (MCP)-1 (Cosmo Bio, Tokyo, Japan). All data are expressed as the mean þ S.D. (n = 3 assays of duplicate wells). Statistical analyses were performed using a one-factor ANOVA test. P 6 0.05 was considered signi¢cant. This study was performed in accordance with the Law Concerning the Protection and Control of Animals (Law No. 105, October 1, 1973), the Standards Relating to the Care and Custody of Laboratory Animals (Noti¢cation No. 9, March 27, 1980, Prime Minister's Of¢ce), and the Method for Sacri¢cing Laboratory Animals (Noti¢cation No. 40, July 4, 1995, Prime Minister's O¤ce). The animal use protocol was approved by the Committee on Animal Experiments (CAE) of Oita Medical University (Approval No. 9827004, May 22, 1998). 3. Results Structural characteristics of the secondary cultures were consistent with those of ¢brocytes. Spindle-shaped cells having large nuclei and scant cytoplasm with slen-
HEARES 3308 25-10-99
K. Yoshida et al. / Hearing Research 137 (1999) 155^159
157
Fig. 1. Murine SL ¢brocyte immunostaining shows a positive reaction for (A) caldesmon, (B) S-100, but a negative reaction for (C) Na/KATPase and (D) cytokeratin. Bar = 50 Wm.
der irregular processes were noted following subculture. These cells were immunopositive for caldesmon and S-100, but negative for Na/K-ATPase and cytokeratin (Fig. 1). The results match the immunostaining pro¢le of type I ¢brocytes of the SL in para¤n-embedded sections. We investigated the secretory pro¢le of the cultured cells after in vitro stimulation with the proin£ammatory cytokine, IL-1L or TNF-K. After the cells were incu-
bated with IL-1L, TNF-K was detected in the supernatant (Fig. 2). However, IL-1L was not detectable in the sample after stimulation with TNF-K (data not shown). Elevated IL-6 levels were observed in the supernatant after incubation with either proin£ammatory cytokine, especially with TNF-K (Fig. 2). However, IL-2, IL-3, IL-4, IL-5, IL-10, IL-13, IFN-Q, and GM-CSF were not detected in the samples incubated with IL-1L or TNF-K (data not shown).
Fig. 2. Cytokine secretion from murine SL ¢brocytes. TNF-K and IL-6 release was analyzed by enzyme-linked immunosorbent assay on supernatants collected from unstimulated ¢brocytes and ¢brocytes stimulated for 24 h with IL-1L or TNF-K. Measurements were performed in duplicate cultures; the data shown are the mean þ S.D. of two separate experiments. ND: not detected. *P 6 0.01 (by onefactor ANOVA).
Fig. 3. Chemokine secretion from murine SL ¢brocytes. MCP-1, KC, and MIP-2 release was analyzed on enzyme-linked immunosorbent assay on supernatants collected from unstimulated ¢brocytes and ¢brocytes stimulated for 24 h with IL-1L or TNF-K. Measurements were performed in duplicate cultures; the data shown are the mean þ S.D. of two separate experiments. ND: not detected. *P 6 0.01 (by one-factor ANOVA).
HEARES 3308 25-10-99
158
K. Yoshida et al. / Hearing Research 137 (1999) 155^159
Fig. 4. Secretion of sICAM-1 and VEGF from murine SL ¢brocytes. sICAM-1 and VEGF release was analyzed by enzyme-linked immunosorbent assay on supernatants collected from unstimulated ¢brocytes and ¢brocytes stimulated for 24 h with IL-1L or TNF-K. Measurements were performed in duplicate cultures; the data shown are the mean þ S.D. of two separate experiments. ND: not detected. *P 6 0.05; **P 6 0.01 (by one-factor ANOVA).
The chemokines MCP-1, KC, and MIP-2 were detected in the sample after incubation with TNF-K (Fig. 3). Stimulation with IL-1L also raised MCP-1 and KC levels although the increase was less impressive than with TNF-K (Fig. 3). MIP-2 was not detected after incubation with IL-1L (data not shown). MIP-1K was not detected in the supernatant incubated with either proin£ammatory cytokine (data not shown). The level of sICAM-1 was elevated after incubation with TNF-K (Fig. 4), while no change was detected after incubation with IL-1L (data not shown). Higher levels of VGEF were detected in the sample after stimulation with IL-1L or TNF-K (Fig. 4). All the measured levels described above except VGEF responded to stimulation in a dose-dependent manner. In the control samples, incubated with culture medium alone, the levels were signi¢cantly lower or undetectable. 4. Discussion Although inner ear cells have been cultured previously, most studies focused on the establishment of epithelial or endothelial cells. An exception is the recent work by Gratton et al. (1996) who successfully cultured SL ¢brocytes. They maintained long-term secondary ¢brocyte cultures from gerbils and morphologically identi¢ed the cells as type I ¢brocytes. In the present study, we applied their culture protocol to mice because mice are advantageous for immunological investigations. Structural culture characteristics were consistent with those of ¢brocytes, and the immunocytochemical results for caldesmon, S-100, Na/K-ATPase, and cytokeratin indicated that the cultures were composed of
type I ¢brocytes from the SL. While the culture protocol was similar between mice and gerbils, the immunocytochemical strategy to identify the ¢brocytes (Spicer and Schulte, 1991) was di¡erent and is explained in detail in a separate manuscript (unpublished data). Our in vitro study demonstrates the secreting pro¢les of the cytokines or other mediators from cultured inner ear cells. IL-1L and TNF-K, which were used as stimulants in this study, are representative cytokines involved in the change from acute to chronic infection or in£ammation. In this study, we demonstrated that IL-1L or TNF-K stimulates the SL ¢brocytes to produce in£ammatory response mediators, such as cytokines (IL-6, TNF-K), chemokines (MCP-1, KC, MIP-2), sICAM-1, and VEGF. The mediators were found to increase more following stimulation with TNF-K than with IL-1L although IL-1L was not detectable in the sample after stimulation with TNF-K. Since TNF-K increased after the cells were incubated with IL-1L, it is possible that IL-1L acts indirectly via the induction of TNF-K. IL-6 is a mediator of acute in£ammatory reactions and is known to act as a proliferation factor of B cells or plasma cells, or as a factor inducing antibody production. MCP-1, which belongs to the CC chemokine subfamily, is chemotactic for mononuclear cells, both monocytes and lymphocytes (Proost et al., 1996). KC and MIP-2 belong to the CXC chemokine subfamily (Cochran et al., 1983 ; Ryseck et al., 1989 ; TekampOlson et al., 1990). It has been suggested that these two are the functional homologues of human IL-8 because murine KC and MIP-2 induce a strong activation of chemotactic factor for neutrophils and because of their high a¤nity for IL-8 receptors (Driscoll, 1994). ICAM-1 is an adhesion molecule that appears in ¢broblasts when stimulated with IL-1L or TNF-K (Kishimoto et al., 1989). VEGF induces vascular neogenesis and promotes vascular permeability (Fong et al., 1995 ; Millauer et al., 1994; Shalaby et al., 1995). These mediators together might induce in£ammatory cell movement, which would prolong the in£ammatory response. Ultrastructural studies of the mesothelial cells bordering the SL have shown that these cells contain micropores allowing the perilymph free access to the SL (Lim, 1970). Furthermore, the culture medium composition is almost identical to that of perilymph. Thus, our present in vitro study provides evidence as to the probable e¡ects of proin£ammatory cytokines in perilymph on the ¢brocytes in the SL. As for in£ammatory cell transportation routes into the cochlea, two routes are assumed: one is from the general circulation through the spiral modiolar vein (SMV) (Harris et al., 1990), and the other is from the endolymphatic sac where in£ammatory cells are locally produced (Tomiyama and Harris, 1986). When the inner ear is exposed to
HEARES 3308 25-10-99
K. Yoshida et al. / Hearing Research 137 (1999) 155^159
bacteria, endotoxin, or other in£ammatory substances, monocytes and macrophages are transported into the perilymph, where they secrete IL-1L and TNF-K at the early stage. Also, in middle ear in£ammation, cytokines should appear in the perilymph because their molecular weight is low enough to permeate through the round window membrane. These proin£ammatory cytokines in the perilymph might act directly on vascular endothelial cells. However, it is also suspected that these cytokines act on the SL ¢brocytes to produce various in£ammatory mediators that could activate the in£ammatory cells and the neighboring vascular endothelial cells, and that these activated cells would participate in in£ammatory cell tra¤cking from blood circulation into the in£ammatory sites. One might also expect that such enhanced biological defense abilities could cause the SL ¢brocyte damage that is seen accompanied by the labyrinthitis (Ichimiya et al., 1998, 1999). Cytokines and adhesion molecules play a central role in the biological defense system and formation of in£ammatory conditions through intracellular signal transmission. Integrins, however, cannot bind to adhesion molecules at a steady state, but must be activated by intracellular signals transmitted through stimulation with various chemokines to become adhesive factors (Hynes, 1992). We speculate that MCP-1, KC, and MIP-2 produced in the SL ¢brocytes would be presented on the vascular endothelial cell surface through transcytosis and that they would consequently activate various in£ammatory cells. We also assumed that sICAM-1 and VGEF assist in£ammatory cell in¢ltration out of the blood vessels toward the in£ammatory site. The result of the present study is compatible with a previous immunocytochemical study that demonstrated a strong ICAM-1 expression in the SL and SMV at the early stage of labyrinthitis after KLH inoculation into the scala tympani of systemically sensitized rats (Suzuki and Harris, 1995). In some inner ear diseases whose etiology is attributed to immunologic mechanisms, we speculate that the SL is involved in acute and chronic in£ammation and that ¢brocyte disruption could cause malfunctioning of the cochlea. Moreover, it is assumed that there could be networks among in£ammatory cells, ¢brocytes, and vascular endothelial cells in the cochlea and that these networks are interconnected by chemokines, intercellular adhesion molecules, and other mediators. Thus, appropriate control of these networks might suppress in£ammation and inhibit tissue damage. Further investigations of ¢brocyte cultures will be required to clarify these issues. Even though one must consider that the results from in vitro studies are not always applicable in vivo, they will contribute toward providing the hypothesis that should be con¢rmed in vitro.
159
References Cochran, B.H., Re¡el, A.C., Stiles, C.D., 1983. Molecular cloning of gene sequences regulated by platelet-derived growth factor. Cell 33, 939^947. Driscoll, K., 1994. Macrophage in£ammatory proteins: biology and role in pulmonary in£ammation. Exp. Lung Res. 20, 473^490. Fong, G., Rossant, J., Gertsenstein, M., Breitman, M.L., 1995. Role of the Flt-1 receptor tyrosine kinase in regulating the assembly of vascular endothelium. Nature 376, 66^70. Gratton, M.A., Schulte, B.A., Hazen-Martin, D.J., 1996. Characterization and development of an inner ear type I ¢brocyte cell culture. Hear. Res. 99, 71^78. Harris, J.P., Fukuda, S., Keithley, E.M., 1990. Spiral modiolar vein: Its importance in inner ear in£ammation. Acta Otolaryngol. (Stockh.) 110, 357^365. Hynes, R.O., 1992. Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69, 11^25. Ichimiya, I., Adams, J.C., Kimura, R.S., 1994a. Immunolocalization of Na , K -ATPase, Ca -ATPase, calcium-binding proteins, and carbonic anhydrase in the guinea pig inner ear. Acta Otolaryngol. (Stockh.) 114, 167^176. Ichimiya, I., Adams, J.C., Kimura, R.S., 1994b. Changes in immunostaining of cochleas with experimentally induced endolymphatic hydrops. Ann. Otol. Rhinol. Laryngol. 103, 457^468. Ichimiya, I., Kurono, Y., Hirano, T., Mogi, G., 1998. Changes in immunostaining of inner ears after antigen challenge into the scala tympani. Laryngoscope 108, 585^591. Ichimiya, I., Suzuki, M., Hirano, T., Mogi, G., 1999. The in£uence of pneumococcal otitis media on the cochlear lateral wall. Hear. Res. 131, 128^134. Kikuchi, T., Kimura, R.S., Paul, D.L., Adams, J.C., 1995. Gap junctions in the rat cochlea: immunohistochemical and ultrastructural analysis. Anat. Embryol. 191, 101^118. Kishimoto, T.K., Larson, R.S., Corbi, A.L., Dustin, M.L., Staunton, D.E., Springer, T.A., 1989. The leukocyte integrins. Adv. Immunol. 46, 149^182. Lim, D.J., 1970. Surface ultrastructure of the cochlear perilymphatic space. J. Laryngol. Otol. 84, 413^428. Millauer, B., Shawver, L.K., Plate, K.H., Risau, W., Ullrich, A., 1994. Glioblastoma growth inhibited in vivo by a dominant-negative Flk-1 mutant. Nature 367, 576^578. Proost, P., Wuyts, A., Van Damme, J., 1996. The role of chemokines in in£ammation. Int. J. Clin. Lab. Res. 26, 211^223. Ryseck, R.P., MacDonald-Bravo, H., Mattei, M.G., Bravo, R., 1989. Cloning and sequence of a secretory protein induced by growth factors in mouse ¢broblasts. Exp. Cell Res. 180, 266^275. Schulte, B.A., Adams, J.C., 1989. Distribution of immunoreactive Na ,K -ATPase in gerbil cochlea. J. Histochem. Cytochem. 37, 127^134. Shalaby, F., Rossant, J., Yamaguchi, T.P., Gertsenstein, M., Wu, X., Breitman, M.L., Schuh, A.C., 1995. Failure of blood-island formation and vasculogenesis in Flk-1-de¢cient mice. Nature 376, 62^ 66. Spicer, S.S., Schulte, B.A., 1991. Di¡erentiation of inner ear ¢brocytes according to their ion transport related activity. Hear. Res. 56, 53^ 64. Suzuki, M., Harris, J.P., 1995. Expression of intercellular adhesion molecule-1 during inner ear in£ammation. Ann. Otol. Rhinol. Laryngol. 104, 69^75. Tekamp-Olson, P., Gallegos, C., Bauer, D., McClain, J., Sherry, B., Fabre, M., van Deventer, S., Cerami, A., 1990. Cloning and characterization of cDNAs for murine macrophage in£ammatory protein 2 and its human homologues. J. Exp. Med. 172, 911^919. Tomiyama, S., Harris, J.P., 1986. The endolymphatic sac: its importance in inner ear immune responses. Laryngoscope 96, 685^691.
HEARES 3308 25-10-99
E¡ect of proin£ammatory cytokines on cultured spiral ligament ¢brocytes Kazuhide Yoshida, Issei Ichimiya, Masashi Suzuki, Goro Mogi * Department of Otolaryngology, Oita Medical University, 1-1 Idaigaoka, Hasama-cho, Oita 879-5593, Japan Received 10 December 1998; received in revised form 27 July 1999; accepted 3 August 1999
Abstract To clarify the effect of proinflammatory cytokines on spiral ligament (SL) fibrocytes, in vitro studies were performed using secondary cell cultures. Cultures from murine SL fibrocytes were stimulated by interleukin (IL)-1L or tumor necrosis factor (TNF)K, and secretion of various mediators was measured by enzyme-linked immunosorbent assay. After stimulation with the proinflammatory cytokines, IL-6, TNF-K, monocyte chemoattractant protein-1, KC, macrophage inflammatory protein-2, soluble intercellular adhesion molecule-1, and vascular endothelial growth factor levels were elevated. Secretion of these chemokines and other mediators could induce inflammatory cell movement, which would prolong the inflammatory response, leading to fibrocyte damage. Given that SL fibrocytes may play a role in cochlear fluid and ion homeostasis, such fibrocyte disruption could cause cochlear malfunction. ß 1999 Elsevier Science B.V. All rights reserved. Key words: Mouse; Inner ear; Interleukin-1L; Tumor necrosis factor-K; Chemokine; Soluble intercellular adhesion molecule-1
1. Introduction Recent morphological studies suggest that the spiral ligament (SL) may play a role in cochlear £uid and ion homeostasis. Localization of Na/K-ATPase (Ichimiya et al., 1994a ; Schulte and Adams, 1989) and gap junctions (Kikuchi et al., 1995) in the cochlea suggests that cochlear lateral wall cells, including the SL ¢brocytes, provide the anatomical substrate for K ion recirculation from hair cells back to the endolymphatic space. This hypothesis is consistent with electrophysiological changes in the cochlea of animals with experimentally induced endolymphatic hydrops where connexin 26 gap junction protein immunostaining is reduced without other noticeable histological changes (Ichimiya et al., 1994b). Recently, we also reported reduced connexin 26 staining in the SL after labyrinthitis following injection of the protein antigen keyhole limpet hemocyanin (KLH) into the scala tympani of systemically sensitized
* Corresponding author. Tel.: +81 (97) 586-5913; Fax: +81 (97) 549-0762.
guinea pigs (Ichimiya et al., 1998). The ¢ndings of this previous study led us to speculate that SL ¢brocytes may be responsible for the cochlear malfunction associated with some immunologically linked inner ear diseases. Furthermore, reduced connexin 26 staining in the SL in animals with experimental pneumococcal otitis media suggests that some noxious substances that pass through the round window membrane may also a¡ect the SL ¢brocytes (Ichimiya et al., 1999). To characterize the SL ¢brocyte features from an immunological point of view, we performed an in vitro study using cultured ¢brocytes. Because immunological mechanisms have been elucidated in mice, we cultured mouse ¢brocytes and investigated the e¡ect of proin£ammatory cytokines on these ¢brocytes. 2. Materials and methods 2.1. Cell culture Inner ear cell cultures were established as described by Gratton et al. (1996) with minor modi¢cations. CBA/JN Crj mice (male, 6^8 weeks of age) were anes-
0378-5955 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 9 9 ) 0 0 1 3 4 - 3
HEARES 3308 25-10-99
156
K. Yoshida et al. / Hearing Research 137 (1999) 155^159
thetized deeply with intraperitoneal pentobarbital and decapitated. The temporal bones were excised quickly, immersed in ice-cold minimum essential medium-K (MEM-K, Life Technologies, Rockville, MD, USA), and the basal turn of the cochlear lateral wall was dissected. The lateral wall fragments were placed gently on a type I collagen-coated Petri dish (Becton Dickinson Labware, Franklin Lakes, NJ, USA). To facilitate attachment, one drop of the culture medium consisting of MEM-K with 10% fetal calf serum (FCS), 1U Antimycotic-antibiotic (Sigma Chemical, St. Louis, MO, USA), and 1% ITS-G supplement (Life Technologies) were placed on the tissue fragments, followed by a cover glass. The Petri dish was placed in the incubator (37³C; 5% CO2 , 95% air) with maximum humidity for 24 h. Then the cover glass was removed gently, the culture medium was added slowly, and the explants were incubated further. The culture medium was changed every 3 days after rinsing with washing bu¡er, consisting of sterile phosphate-bu¡ered saline (sPBS) with a 1UAntimycotic-antibiotic solution. After about 2 weeks in primary culture, nearly con£uent monolayer cells were rinsed with the washing bu¡er, and trypsin/EDTA (0.05% :0.02%) were added to promote cell detachment. Culture medium was added after a 7-min incubation, and the detached cells were centrifuged at 1000 rpm for 10 min. The cell pellet was resuspended and distributed to new £asks at a subculture ratio of 1:4. Secondary cultures were maintained the same as the primary cultures. 2.2. Immunocytochemistry As described in a separate manuscript (unpublished data), immunostaining for caldesmon, Na/K-ATPase, S-100, or cytokeratin was performed to classify the cultures. Upon con£uence, the cells were subcultured into 2-well chamber slides (Becton Dickinson Labware), rinsed in PBS, ¢xed in 10% neutral-bu¡ered formalin for 10 min, and dehydrated. The slides were rinsed in PBS and exposed to a 5% solution of normal horse, rabbit, or goat serum; then incubated with mouse anti-caldesmon (diluted 1:300, Sigma Chemical), chicken anti-Na/K-ATPase (diluted 1:30 000, Cortex Biochem, San Leandro, CA, USA), rabbit anti-S-100 (diluted 1:1000, Sigma Chemical), or rabbit anticytokeratin (ready-to-use solution, Zymed Laboratories, San Francisco, CA, USA). Sections were then £ooded with biotin-conjugated horse anti-mouse immunoglobulin G (IgG) (diluted 1:200, Vector Laboratories, Burlingame, CA, USA), rabbit anti-chicken IgG (diluted 1:4000, Zymed Laboratories), or goat anti-rabbit IgG (diluted 1:200, Vector Laboratories) as appropriate. Bound primary antibody sites were visualized by incubating with ABC reagent (Vector Laboratories) fol-
lowed with 0.05% 3,3P-diaminobenzidine-0.01% H2 O2 substrate medium in 0.1 M phosphate bu¡er. After counterstaining with Veronal acetate-bu¡ered 1% methyl green solution, the slides were mounted for observation. Staining to con¢rm antibody speci¢city was done as described above, except that PBS was used in place of the primary antibodies. To identify the cell type of the culture, para¤n-embedded cochlea sections from normal mice were processed at the same step, and the staining patterns of the cochlear lateral wall were compared with the immunostaining pro¢le of the cultured cells. 2.3. Cytokine stimulation studies Secondary cultures were grown in six-well type I collagen-coated £at-bottom plates (5U104 cells/well, Becton Dickinson Labware) in the culture medium described above. The con£uent monolayer cells were incubated with 2 ml of culture medium for 24 h alone as controls or in the presence of the proin£ammatory cytokine, interleukin (IL)-1L or tumor necrosis factor (TNF)-K (each at 1 or 10 ng/ml; RpD Systems, Minneapolis, MN, USA). Conditioned medium was analyzed with the following enzyme-linked immunosorbent assay kits: IL-1L, IL-2, IL-4, IL-6, IL-10, IL-13, interferon (IFN)-Q, TNF-K, granulocyte-macrophage colony-stimulating factor (GM-CSF), macrophage in£ammatory protein (MIP)-1a, MIP-2, KC, vascular endothelial growth factor (VEGF) (RpD Systems), IL-3, IL-5, soluble intercellular adhesion molecule (sICAM)-1 (Endgen, Woburn, MA, USA), and chemoattractant protein (MCP)-1 (Cosmo Bio, Tokyo, Japan). All data are expressed as the mean þ S.D. (n = 3 assays of duplicate wells). Statistical analyses were performed using a one-factor ANOVA test. P 6 0.05 was considered signi¢cant. This study was performed in accordance with the Law Concerning the Protection and Control of Animals (Law No. 105, October 1, 1973), the Standards Relating to the Care and Custody of Laboratory Animals (Noti¢cation No. 9, March 27, 1980, Prime Minister's Of¢ce), and the Method for Sacri¢cing Laboratory Animals (Noti¢cation No. 40, July 4, 1995, Prime Minister's O¤ce). The animal use protocol was approved by the Committee on Animal Experiments (CAE) of Oita Medical University (Approval No. 9827004, May 22, 1998). 3. Results Structural characteristics of the secondary cultures were consistent with those of ¢brocytes. Spindle-shaped cells having large nuclei and scant cytoplasm with slen-
HEARES 3308 25-10-99
K. Yoshida et al. / Hearing Research 137 (1999) 155^159
157
Fig. 1. Murine SL ¢brocyte immunostaining shows a positive reaction for (A) caldesmon, (B) S-100, but a negative reaction for (C) Na/KATPase and (D) cytokeratin. Bar = 50 Wm.
der irregular processes were noted following subculture. These cells were immunopositive for caldesmon and S-100, but negative for Na/K-ATPase and cytokeratin (Fig. 1). The results match the immunostaining pro¢le of type I ¢brocytes of the SL in para¤n-embedded sections. We investigated the secretory pro¢le of the cultured cells after in vitro stimulation with the proin£ammatory cytokine, IL-1L or TNF-K. After the cells were incu-
bated with IL-1L, TNF-K was detected in the supernatant (Fig. 2). However, IL-1L was not detectable in the sample after stimulation with TNF-K (data not shown). Elevated IL-6 levels were observed in the supernatant after incubation with either proin£ammatory cytokine, especially with TNF-K (Fig. 2). However, IL-2, IL-3, IL-4, IL-5, IL-10, IL-13, IFN-Q, and GM-CSF were not detected in the samples incubated with IL-1L or TNF-K (data not shown).
Fig. 2. Cytokine secretion from murine SL ¢brocytes. TNF-K and IL-6 release was analyzed by enzyme-linked immunosorbent assay on supernatants collected from unstimulated ¢brocytes and ¢brocytes stimulated for 24 h with IL-1L or TNF-K. Measurements were performed in duplicate cultures; the data shown are the mean þ S.D. of two separate experiments. ND: not detected. *P 6 0.01 (by onefactor ANOVA).
Fig. 3. Chemokine secretion from murine SL ¢brocytes. MCP-1, KC, and MIP-2 release was analyzed on enzyme-linked immunosorbent assay on supernatants collected from unstimulated ¢brocytes and ¢brocytes stimulated for 24 h with IL-1L or TNF-K. Measurements were performed in duplicate cultures; the data shown are the mean þ S.D. of two separate experiments. ND: not detected. *P 6 0.01 (by one-factor ANOVA).
HEARES 3308 25-10-99
158
K. Yoshida et al. / Hearing Research 137 (1999) 155^159
Fig. 4. Secretion of sICAM-1 and VEGF from murine SL ¢brocytes. sICAM-1 and VEGF release was analyzed by enzyme-linked immunosorbent assay on supernatants collected from unstimulated ¢brocytes and ¢brocytes stimulated for 24 h with IL-1L or TNF-K. Measurements were performed in duplicate cultures; the data shown are the mean þ S.D. of two separate experiments. ND: not detected. *P 6 0.05; **P 6 0.01 (by one-factor ANOVA).
The chemokines MCP-1, KC, and MIP-2 were detected in the sample after incubation with TNF-K (Fig. 3). Stimulation with IL-1L also raised MCP-1 and KC levels although the increase was less impressive than with TNF-K (Fig. 3). MIP-2 was not detected after incubation with IL-1L (data not shown). MIP-1K was not detected in the supernatant incubated with either proin£ammatory cytokine (data not shown). The level of sICAM-1 was elevated after incubation with TNF-K (Fig. 4), while no change was detected after incubation with IL-1L (data not shown). Higher levels of VGEF were detected in the sample after stimulation with IL-1L or TNF-K (Fig. 4). All the measured levels described above except VGEF responded to stimulation in a dose-dependent manner. In the control samples, incubated with culture medium alone, the levels were signi¢cantly lower or undetectable. 4. Discussion Although inner ear cells have been cultured previously, most studies focused on the establishment of epithelial or endothelial cells. An exception is the recent work by Gratton et al. (1996) who successfully cultured SL ¢brocytes. They maintained long-term secondary ¢brocyte cultures from gerbils and morphologically identi¢ed the cells as type I ¢brocytes. In the present study, we applied their culture protocol to mice because mice are advantageous for immunological investigations. Structural culture characteristics were consistent with those of ¢brocytes, and the immunocytochemical results for caldesmon, S-100, Na/K-ATPase, and cytokeratin indicated that the cultures were composed of
type I ¢brocytes from the SL. While the culture protocol was similar between mice and gerbils, the immunocytochemical strategy to identify the ¢brocytes (Spicer and Schulte, 1991) was di¡erent and is explained in detail in a separate manuscript (unpublished data). Our in vitro study demonstrates the secreting pro¢les of the cytokines or other mediators from cultured inner ear cells. IL-1L and TNF-K, which were used as stimulants in this study, are representative cytokines involved in the change from acute to chronic infection or in£ammation. In this study, we demonstrated that IL-1L or TNF-K stimulates the SL ¢brocytes to produce in£ammatory response mediators, such as cytokines (IL-6, TNF-K), chemokines (MCP-1, KC, MIP-2), sICAM-1, and VEGF. The mediators were found to increase more following stimulation with TNF-K than with IL-1L although IL-1L was not detectable in the sample after stimulation with TNF-K. Since TNF-K increased after the cells were incubated with IL-1L, it is possible that IL-1L acts indirectly via the induction of TNF-K. IL-6 is a mediator of acute in£ammatory reactions and is known to act as a proliferation factor of B cells or plasma cells, or as a factor inducing antibody production. MCP-1, which belongs to the CC chemokine subfamily, is chemotactic for mononuclear cells, both monocytes and lymphocytes (Proost et al., 1996). KC and MIP-2 belong to the CXC chemokine subfamily (Cochran et al., 1983 ; Ryseck et al., 1989 ; TekampOlson et al., 1990). It has been suggested that these two are the functional homologues of human IL-8 because murine KC and MIP-2 induce a strong activation of chemotactic factor for neutrophils and because of their high a¤nity for IL-8 receptors (Driscoll, 1994). ICAM-1 is an adhesion molecule that appears in ¢broblasts when stimulated with IL-1L or TNF-K (Kishimoto et al., 1989). VEGF induces vascular neogenesis and promotes vascular permeability (Fong et al., 1995 ; Millauer et al., 1994; Shalaby et al., 1995). These mediators together might induce in£ammatory cell movement, which would prolong the in£ammatory response. Ultrastructural studies of the mesothelial cells bordering the SL have shown that these cells contain micropores allowing the perilymph free access to the SL (Lim, 1970). Furthermore, the culture medium composition is almost identical to that of perilymph. Thus, our present in vitro study provides evidence as to the probable e¡ects of proin£ammatory cytokines in perilymph on the ¢brocytes in the SL. As for in£ammatory cell transportation routes into the cochlea, two routes are assumed: one is from the general circulation through the spiral modiolar vein (SMV) (Harris et al., 1990), and the other is from the endolymphatic sac where in£ammatory cells are locally produced (Tomiyama and Harris, 1986). When the inner ear is exposed to
HEARES 3308 25-10-99
K. Yoshida et al. / Hearing Research 137 (1999) 155^159
bacteria, endotoxin, or other in£ammatory substances, monocytes and macrophages are transported into the perilymph, where they secrete IL-1L and TNF-K at the early stage. Also, in middle ear in£ammation, cytokines should appear in the perilymph because their molecular weight is low enough to permeate through the round window membrane. These proin£ammatory cytokines in the perilymph might act directly on vascular endothelial cells. However, it is also suspected that these cytokines act on the SL ¢brocytes to produce various in£ammatory mediators that could activate the in£ammatory cells and the neighboring vascular endothelial cells, and that these activated cells would participate in in£ammatory cell tra¤cking from blood circulation into the in£ammatory sites. One might also expect that such enhanced biological defense abilities could cause the SL ¢brocyte damage that is seen accompanied by the labyrinthitis (Ichimiya et al., 1998, 1999). Cytokines and adhesion molecules play a central role in the biological defense system and formation of in£ammatory conditions through intracellular signal transmission. Integrins, however, cannot bind to adhesion molecules at a steady state, but must be activated by intracellular signals transmitted through stimulation with various chemokines to become adhesive factors (Hynes, 1992). We speculate that MCP-1, KC, and MIP-2 produced in the SL ¢brocytes would be presented on the vascular endothelial cell surface through transcytosis and that they would consequently activate various in£ammatory cells. We also assumed that sICAM-1 and VGEF assist in£ammatory cell in¢ltration out of the blood vessels toward the in£ammatory site. The result of the present study is compatible with a previous immunocytochemical study that demonstrated a strong ICAM-1 expression in the SL and SMV at the early stage of labyrinthitis after KLH inoculation into the scala tympani of systemically sensitized rats (Suzuki and Harris, 1995). In some inner ear diseases whose etiology is attributed to immunologic mechanisms, we speculate that the SL is involved in acute and chronic in£ammation and that ¢brocyte disruption could cause malfunctioning of the cochlea. Moreover, it is assumed that there could be networks among in£ammatory cells, ¢brocytes, and vascular endothelial cells in the cochlea and that these networks are interconnected by chemokines, intercellular adhesion molecules, and other mediators. Thus, appropriate control of these networks might suppress in£ammation and inhibit tissue damage. Further investigations of ¢brocyte cultures will be required to clarify these issues. Even though one must consider that the results from in vitro studies are not always applicable in vivo, they will contribute toward providing the hypothesis that should be con¢rmed in vitro.
159
References Cochran, B.H., Re¡el, A.C., Stiles, C.D., 1983. Molecular cloning of gene sequences regulated by platelet-derived growth factor. Cell 33, 939^947. Driscoll, K., 1994. Macrophage in£ammatory proteins: biology and role in pulmonary in£ammation. Exp. Lung Res. 20, 473^490. Fong, G., Rossant, J., Gertsenstein, M., Breitman, M.L., 1995. Role of the Flt-1 receptor tyrosine kinase in regulating the assembly of vascular endothelium. Nature 376, 66^70. Gratton, M.A., Schulte, B.A., Hazen-Martin, D.J., 1996. Characterization and development of an inner ear type I ¢brocyte cell culture. Hear. Res. 99, 71^78. Harris, J.P., Fukuda, S., Keithley, E.M., 1990. Spiral modiolar vein: Its importance in inner ear in£ammation. Acta Otolaryngol. (Stockh.) 110, 357^365. Hynes, R.O., 1992. Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69, 11^25. Ichimiya, I., Adams, J.C., Kimura, R.S., 1994a. Immunolocalization of Na , K -ATPase, Ca -ATPase, calcium-binding proteins, and carbonic anhydrase in the guinea pig inner ear. Acta Otolaryngol. (Stockh.) 114, 167^176. Ichimiya, I., Adams, J.C., Kimura, R.S., 1994b. Changes in immunostaining of cochleas with experimentally induced endolymphatic hydrops. Ann. Otol. Rhinol. Laryngol. 103, 457^468. Ichimiya, I., Kurono, Y., Hirano, T., Mogi, G., 1998. Changes in immunostaining of inner ears after antigen challenge into the scala tympani. Laryngoscope 108, 585^591. Ichimiya, I., Suzuki, M., Hirano, T., Mogi, G., 1999. The in£uence of pneumococcal otitis media on the cochlear lateral wall. Hear. Res. 131, 128^134. Kikuchi, T., Kimura, R.S., Paul, D.L., Adams, J.C., 1995. Gap junctions in the rat cochlea: immunohistochemical and ultrastructural analysis. Anat. Embryol. 191, 101^118. Kishimoto, T.K., Larson, R.S., Corbi, A.L., Dustin, M.L., Staunton, D.E., Springer, T.A., 1989. The leukocyte integrins. Adv. Immunol. 46, 149^182. Lim, D.J., 1970. Surface ultrastructure of the cochlear perilymphatic space. J. Laryngol. Otol. 84, 413^428. Millauer, B., Shawver, L.K., Plate, K.H., Risau, W., Ullrich, A., 1994. Glioblastoma growth inhibited in vivo by a dominant-negative Flk-1 mutant. Nature 367, 576^578. Proost, P., Wuyts, A., Van Damme, J., 1996. The role of chemokines in in£ammation. Int. J. Clin. Lab. Res. 26, 211^223. Ryseck, R.P., MacDonald-Bravo, H., Mattei, M.G., Bravo, R., 1989. Cloning and sequence of a secretory protein induced by growth factors in mouse ¢broblasts. Exp. Cell Res. 180, 266^275. Schulte, B.A., Adams, J.C., 1989. Distribution of immunoreactive Na ,K -ATPase in gerbil cochlea. J. Histochem. Cytochem. 37, 127^134. Shalaby, F., Rossant, J., Yamaguchi, T.P., Gertsenstein, M., Wu, X., Breitman, M.L., Schuh, A.C., 1995. Failure of blood-island formation and vasculogenesis in Flk-1-de¢cient mice. Nature 376, 62^ 66. Spicer, S.S., Schulte, B.A., 1991. Di¡erentiation of inner ear ¢brocytes according to their ion transport related activity. Hear. Res. 56, 53^ 64. Suzuki, M., Harris, J.P., 1995. Expression of intercellular adhesion molecule-1 during inner ear in£ammation. Ann. Otol. Rhinol. Laryngol. 104, 69^75. Tekamp-Olson, P., Gallegos, C., Bauer, D., McClain, J., Sherry, B., Fabre, M., van Deventer, S., Cerami, A., 1990. Cloning and characterization of cDNAs for murine macrophage in£ammatory protein 2 and its human homologues. J. Exp. Med. 172, 911^919. Tomiyama, S., Harris, J.P., 1986. The endolymphatic sac: its importance in inner ear immune responses. Laryngoscope 96, 685^691.
HEARES 3308 25-10-99