Ph.D. Program in Human Biology, School of Integrative and Global Majors, University of Tsukuba, Tsukuba, Ibaraki, JapanDepartment of Immunology, Faculty of Medicine, University of Tsukuba, Tsukuba, Ibaraki, JapanLife Science Center of Survival Dynamics, Tsukuba Advanced Research Alliance (TARA), University of Tsukuba, Tsukuba, Ibaraki, Japan
Allergy and Immunology Project Team, Center for Institutional Research and Medical Education, Nihon University School of Medicine, Itabashi-Ku, Tokyo, Japan
Department of Immunology, Faculty of Medicine, University of Tsukuba, Tsukuba, Ibaraki, JapanLife Science Center of Survival Dynamics, Tsukuba Advanced Research Alliance (TARA), University of Tsukuba, Tsukuba, Ibaraki, Japan
Although phosphatidylserine (PS) confined to the inner leaflet of plasma membrane is exposed on the cell surface when cells undergo apoptosis, viable cells, including mast cells (MCs), also externalize PS in certain cellular states.
However, the pathophysiological significance of PS exposure on viable cells remains elusive. To address the role of PS externalization on the cell surface of viable MCs, we monitored PS surface exposure on bone marrow–derived cultured MCs (BMMCs) by confocal microscopy after stimulation with trinitrophenyl (TNP)-specific IgE and TNP-ovalbumin (OVA) in the presence of PSVue 643, a fluorescent dye with rapid binding capacity for PS. The dye began to accumulate on the cell surface of live BMMCs within 600 seconds after FcεRI stimulation, whereas the nonstimulated BMMCs remained negative for the staining (Fig 1, A; see Video E1 in this article's Online Repository at www.jacionline.org), indicating that PS is externalized within 10 minutes after activation.
Fig 1Imaging analyses of PS and CD300a on BMMCs during degranulation. A, Time-lapse montage of BMMC degranulation in the presence of PSVue643 by confocal microscopy. Dead cells were used as a PSVue-643 staining positive control. Scale bars, 10 μm. See also Video E1. B, Imaging flow cytometry of degranulated BMMCs stained with anti-CD300a and Annexin V. Representative gating (left); single-cell images (center); and colocalization of CD300a and PS analyzed using bright detail similarity R3 based on images of single cells (see this article's Methods section in the Online Repository at www.jacionline.org) in the indicated gate (right). C, FRET analysis between CD300a and PS. Representative confocal pictures (left) and FRET+ cell quantification (right) (see this article's Methods section). Scale bars, 5 μm. Data are representative of more than 3 independent experiments. Error bars indicate SD. APC, Allophycocyanin; Em, emission; Ex, excitation; NBD-PS, 1-Oleoyl-2-{6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl}-sn-glycero-3-phosphoserine; PE, phycoerythrin. **P < .01.
By imaging flow cytometry analyses, we found that both PS and CD300a showed a polarization and colocalization on the cell surface of mouse BMMCs (Annexin V–stained) and cultured human synovial MCs (hMCs) 15 minutes after FcεRI stimulation (Fig 1, B; see Fig E1, A and B, in this article's Online Repository at www.jacionline.org). Fluorescence resonance energy transfer (FRET) analyses using 1-oleoyl-2-{6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl}-sn-glycero-3-phosphoserine (NBD-PS) as a donor and a non-neutralizing anti-CD300a antibody (TX10) conjugated with Alexa546 as an acceptor demonstrated increased FRET efficiency calculated by sensitized emission between CD300a and PS during degranulation (Fig 1, C). Treatment with a neutralizing anti-CD300a antibody (EX42) that interferes with the binding of CD300a with PS dramatically decreased the FRET+ cell number during degranulation (see Fig E1, C and D), demonstrating the direct cis-interaction between CD300a and PS in the colocalized region of degranulating MCs. PS also showed a colocalization with CD107a on the cell surface of degranulating (ie, CD107a+ Annexin V+) BMMCs (see Fig E1, E), suggesting that PS exposure on the cell surface of MCs is associated with degranulation. Indeed, treatment of BMMCs with the MC- degranulating stimuli ATP and ionomycin induced the exposure of PS as well as CD107a on the cell surface. In contrast, neither CD107a nor PS was observed on the cell surface of BMMCs after stimulation with either MC activator LPS or IL-33
(see Fig E1, F). Unlike FcεRI stimulation, however, stimulation with neither ATP nor ionomycin induced the polarization of CD300a and PS on the cell surface of MCs (see Fig E1, G). Together, these data demonstrate that PS is promptly exposed together with CD107a during degranulation of MCs and colocalized with CD300a.
We next examined the degranulation of wild-type (WT) and Cd300a−/− BMMCs following stimulation of FcεRI by addition of TNP-specific IgE and TNP-OVA. Flow cytometry analyses showed that the size of the CD107a+ population was significantly larger in Cd300a−/− BMMCs than in WT BMMCs from 8 minutes (480 seconds) to 30 minutes (1800 seconds) after stimulation (Fig 2, A and B). To confirm the cell-intrinsic effect of CD300a on degranulation, equal numbers of WT and Cd300a−/− BMMCs were mixed and stimulated with TNP-specific IgE and TNP-OVA. The CD107a+ population was again observed to be larger in Cd300a−/− BMMCs than in WT BMMCs (see Fig E2, A and B, in this article's Online Repository at www.jacionline.org). Moreover, Cd300a−/− BMMCs released a larger amount of β-hexosaminidase than did WT BMMCs when they were analyzed 30 minutes after stimulation (Fig 2, C). We found that the population of propidium iodide (PI)+ dead cells was comparable between the cultures of WT and Cd300a−/− BMMCs before and after stimulation of FcεRI (see Fig E2, C). Moreover, even in the low concentrations of BMMCs in the culture, in which trans-interactions of CD300a with PS were unlikely, Cd300a−/− BMMCs still showed increased degranulation compared with WT BMMCs (see Fig E2, D and E). Treatment with a neutralizing anti-CD300a antibody (EX42) increased CD107a expression on WT BMMCs to a level comparable to that of Cd300a−/− BMMCs (Fig 2, D). Similarly, a neutralizing anti-human CD300a antibody increased the CD107a expression during FcεRI-mediated degranulation of hMCs (Fig 2, E). Together, these results further support the involvement of cis-interaction, rather than trans-interaction, of CD300a with PS in the suppression of MC degranulation. In contrast, CD107a expression was comparable between WT and Cd300a−/− BMMCs stimulated with either ATP or ionomycin (see Fig E2, F). Interestingly, FcεRI stimulation also induced polarization of FcεRI and colocalization of FcεRI with CD300a and PS (see Fig E2, G and H). Thus, the same spatiotemporal localization of CD300a, PS, and FcεRI on MCs might cause CD300a-mediated suppression of FcεRI-mediated, but not ATP- or ionomycin-mediated, CD107a expression during degranulation. Together, these imaging and functional analyses suggest that the cis-interaction of CD300a with PS specifically suppresses FcεRI-mediated signaling for degranulation of MCs. Indeed, Syk phosphorylation was higher in Cd300a−/− BMMCs than in WT BMMCs after antigen challenge at 10 minutes (600 seconds) after FcεRI stimulation, but was comparable between WT and Cd300a−/− BMMCs at 2 minutes (120 seconds) after stimulation (see Fig E3 in this article's Online Repository at www.jacionline.org), consistent with our observation that degranulation was significantly higher in Cd300a−/− than in WT BMMCs at 8 minutes (480 seconds), but not 2 minutes (120 seconds), after FcεRI stimulation (Fig 2, B).
Fig 2Functional interaction between CD300a and PS externalized during degranulation of MCs. A, Representative plots showing CD107a expression of antigen-stimulated WT or Cd300a−/− BMMCs gated on the PI− c-Kit+ cells. B, Kinetics of CD107a expression after antigen stimulation. C, β-Hexosaminidase release in the culture after degranulation. D and E, Effect of anti-mouse CD300a or anti-human CD300a neutralizing antibodies on degranulation of WT and Cd300a−/− BMMCs (Fig 2, D) and cultured human synovial MCs (Fig 2, E) gated on the PI− c-Kit+ cells. F and G, Change in intrarectal temperature in WT (n = 6) and Cd300a−/− (n = 6) mice (Fig 2, F) or WT mice that had been injected with control (n = 8) or anti-CD300a (n = 9) antibody (Fig 2, G) after intravenous sensitization with TNP-specific IgE, followed by intravenous challenge with TNP-OVA. Data are representative of 3 independent experiments (Fig 2, A-E). Data are pooled from 2 to 3 experiments and error bars indicated SD (Fig 2, F and G). APC, Allophycocyanin. *P < .05, **P < .01, and ***P < .001.
We further analyzed the role of CD300a in the pathogenesis of passive systemic anaphylaxis (PSA). WT and Cd300a−/− mice were intravenously injected with TNP-specific IgE, followed by intravenous challenge with TNP-OVA. Although rectal temperature of WT and Cd300a−/− mice decreased to a similar level by 20 minutes (1200 seconds) after the challenge, the recovery of the rectal temperature after this period was slower in Cd300a−/− mice than in WT mice (Fig 2, F). Interestingly, this seems to correlate to the time course of the self-regulation of MCs degranulation in vitro by CD300a. Similar results were also observed when mice were intraperitoneally injected with a neutralizing anti-CD300a antibody (EX42) (Fig 2, G), suggesting that the CD300a-PS interaction suppressed PSA. Moreover, polarized PS was detected only in MCs, rather than at regions surrounding MCs, in the ear tissue of mice after injection with Flag-tagged MFG-E8-D89E (milk fat globule-EGF factor 8 protein- D89E) together with TNP-OVA (see Fig E4 in this article's Online Repository at www.jacionline.org). These results suggest that the cis-interaction regulated MCs degranulation in vivo as well as in vitro.
In conclusion, by combining imaging and functional analyses of MC degranulation, we revealed the physical and functional associations of externalized PS with CD300a via cis-interaction on viable MCs during degranulation (see Fig E5 in this article's Online Repository at www.jacionline.org). Given that MCs are widely distributed in many tissues in a scattered manner without contact with each other
such self-regulation of MC degranulation represents a novel strategy that MCs evolved to control their own activation, adding another layer of regulation in allergic responses.
We thank S. Tochihara and W. Saito and the member of the Shibuya Laboratory for secretarial assistance and comments and discussions, respectively.
Methods
Cells
BMMCs were generated by culturing WT and Cd300a−/− mouse bone marrow cells in the presence of 10 ng/mL stem cell factor (455-MC/CF, R&D Systems, Minneapolis, Minn) and 4 ng/mL IL-3 (403-ML, R&D Systems) as previously described.
Briefly, weekly passages were performed by seeding 2 × 106 cells in 10 mL medium. Cells were cultured for 5 to 8 weeks before use. Cd300a−/− mice were described previously.
Activation of human synovial mast cells from rheumatoid arthritis or osteoarthritis patients in response to aggregated IgG through Fcγ receptor I and Fcγ receptor II.
All mice experiments were conducted in accordance with the guidelines of the Animal Ethics Committee of the University of Tsukuba Animal Research Center. The human mast cell study was approved by the Ethics Committee of the Nihon University School of Medicine (RK-160112-2), and all the subjects provided written informed consent in accordance with the Helsinki Declaration of the World Medical Association.
Antibodies, other reagents, and flow cytometry
Anti-mouse CD107a (1D4B), anti-mouse c-Kit (2B8), anti-mouse IgE (RME-1), mouse IgG1 (MOPC-21), anti-human CD107a (H4A3), and anti-Flag (L5) antibodies and fluorescein isothiocyanate (FITC)-Avidin were purchased from Biolegend (San Diego, Calif). Human IgE Myeloma (401152) was purchased from CALBIOCHEM. Anti-human IgE (Dε2) (E124.2.8) was purchased from Beckman Coulter (Brea, Calif). Annexin V, trinitrophenyl (TNP)-specific mouse IgE (C38-2) was purchased from BD Bioscience (Franklin Lakes, NJ). Anti-Syk (#2712) and anti-pSyk (#2711) antibodies were purchased from Cell Signaling Technology (Danvers, Mass). TNP-OVA, MFG-E8-D89E, MFG-E8-EPT, neutralizing and non-neutralizing anti-mouse CD300a antibodies (EX42 and TX10, respectively), CD300a-human Fc protein and antihuman CD300a (mouse IgG1, TX49) were made in our laboratory, as previously described.
1-Oleoyl-2-{6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl}-sn-glycero-3-phosphoserine (NBD-PS) was from Avanti. LPS (L2880) derived from Escherichia coli O55:B5, ATP (A26209), and ionomycin (I9657) were purchased from Sigma-Aldrich (St Louis, Mo). PSVue480 and PSVue643 were purchased from Molecular Targeting Technologies (West Chester, Pa). Recombinant mouse IL-33 was purchased from R&D Systems (3626-ML).
CD300a-Fc–blocking experiment was done with either apoptotic cell induced by dexamethasone or degranulated BMMCs. Briefly, 105 cells were incubated with CD300a-Fc together with 50 ng anti-CD300a antibody (TX10 or EX42) or 100 ng MFG-E8-D89E on ice for 15 minutes, and further stained by anti-CD107a (for BMMCs only) and anti-human IgG Fc-PE (M1310G05, Biolegend) and phycoerythrin (PI).
Flow cytometry analyses were performed using an LSRFortessa system (BD Bioscience), and data were analyzed by using FlowJo software (BD Bioscience).
Degranulation and other stimulations
For mouse in vitro stimulations, MCs were sensitized with 1 μg/mL TNP-specific IgE overnight, washed twice with Tyrode's buffer, preincubated with reagents as indicated for 30 minutes, and then challenged with 1 ng/mL TNP-OVA for 30 minutes or indicated time point. β-Hexosaminidase activity was measured as previously reported.
Briefly, 50 μL of the culture supernatant was mixed with 50 μL of 4-nitorophenyl-N-acetyl-β-d-glucosaminide (1.3 mg/mL; Sigma, N9376) in substrate buffer (0.4 M citric acid, 0.2 M NaH2PO4, pH 4.5) and incubated at 37°C for 3 hours. The reaction was stopped by adding 100 μL of glycine (0.2 M, pH 10.7), the solution was thoroughly mixed, and then absorbance of 415 nm was determined. CD107a was measured by flow cytometry after staining with anti-cKit and anti-CD107a antibodies and PI. Apoptotic thymocytes were induced in 1 μM dexamethasone with 107 cell/mL RPMI medium for 12 hours. Cell density pictures were acquired under 20× objective lens under bright field and converted into binary by ImageJ.
For human MC stimulation, cultured human synovial MCs were sensitized with 500 ng/mL human IgE, stained with 50 μg/mL anti-human CD300a or isotype control antibody, challenged with 300 ng/mL anti-human IgE antibody for 30 minutes, followed by staining with anti-human CD107a and PI, and analyzed by flow cytometry.
For measuring PS exposure of BMMCs after stimulations by LPS (1 μg/mL), IL-33 (150 ng/mL), TNP-OVA (10 ng/mL), ATP (0.5 mM), or ionomycin (2500 ng/mL), the cells were incubated with indicated reagents for 20 minutes after IgE sensitization and then stained with antibody against CD107a, annexin V, and PI. For measuring degranulation after ATP (0.5 mM) or ionomycin (500 ng/mL) stimulation, the stimulation time was 30 minutes.
Live imaging
For time-lapse imaging, BMMCs were sensitized as mentioned above and then incubated in 500 μL of Tyrode's buffer containing 1 mM PSVue-643 fluorescent probe (Polysciences, Inc, Warrington, Pa) in a glass-bottom dish (CELLview, Greiner Bio-One, Kremsmünster, Austria) for 30 minutes before gently adding 10 μL of TNP-OVA (100 ng/mL). Cells were monitored under a laser scanning confocal microscope (Olympus FV10i) at 10-second intervals under 60× optical magnification. Data were analyzed and exported by FV10-ASW (Olympus, Tokyo, Japan). The video and montage were generated by ImageJ software (NIH, Bethesda, Md) with 10 frames/s. The confocal images of single mast cells were also collected and analyzed on the same platform with PS stained by PSVue480 and CD300a by Alexa647-conjugated anti-CD300a.
For imaging flow cytometry, the Image Stream Mark II system (Amnis) was used to observe single cells after 15-minute degranulation and staining. Data were acquired and analyzed by using the Inspire and Ideas software packages (Merck, Kenilworth, NJ), respectively. The Bright Detail Similarity R3 Feature (based on Pearson correlation coefficient) was adopted as the localization measurement according to the manufacturer's instructions. Briefly, the Bright Detail Similarity R3 Feature value was calculated with a customized imaging mask to identify aggregation of the molecule of interest (CD107a or PS) on the cell surface. To detect colocalization of CD107a and PS, the mask was set on the CD107a channel with the following parameters: Threshold (M05, Ch05, 73) and Peak (M05, Ch05, Bright 10). To detect colocalization of CD300a and PS, the mask was set on the PS channel with following parameters: Threshold (M05, Ch05, 60) and Peak (M05, Ch05, Bright 4). The data were exported as a Flow Cytometry Standard file and analyzed by FlowJo (BD Bioscience).
FRET analysis was used as direct evidence of cis-interaction between other receptors.
To analyze the cis-interaction between CD300a and PS, TNP-specific IgE-sensitized BMMCs were first stained by nonblocking anti-CD300a antibody TX10 (Alexa546 conjugated) for 15 minutes and then labeled by 500 nM NBD-PS in HBSS with 1 mM CaCl2 for 8 minutes in room temperature. The cells were immediately washed by 5 mg/mL fatty acid-free BSA in HBSS with 1 mM CaCl2. The stained cells were challenged by 10 ng/mL TNP-OVA and observed under live imaging conditions with laser set 471 nm/559 nm and filter set 490 to 540 nm and 570 to 620 nm.
FRET efficiency was calculated by sensitized emission according to the FRET package instructions in FV10-ASW (Olympus).
where
are the collection efficiency in donor and acceptor channel, respectively;
are quantum yield of the donor and acceptor, respectively.
f : is acceptor with donor excitation.
e is donor with donor excitation;
g : is acceptor with acceptor excitation;
and e, f, and g were obtained using donor and acceptor double-stained samples (FRET samples).
Donor spectral bleedthrough (DSBT) and acceptor spectral bleedthrough (ASBT) were calculated by images from the donor/acceptor single- stained sample excited by designated laser according to the package instruction to obtain the value a, b, c, and d.
FRET+ cell percentage was calculated by counting the FRET+ cells in each field.
Passive systemic anaphylaxis
WT and Cd300a−/− mice (age-matched 8- to 14-week-old females) were sensitized by intravenous (IV) administration of 5 μg of TNP-specific IgE (BD Biosciences, C38-2) for 24 hours and then IV challenged with 40 μg TNP-OVA. Body temperature was measured intrarectally at the indicated time points. For antibody blocking, a neutralizing anti-CD300a or isotype (400 μg/mice) was injected intraperitoneally 5 hours before antigen challenge.
For staining ear tissue sections, 50 μg/mice MFG-E8-D89E (Flag-taged) were IV injected together with or without 40 μg/mice antigen. Ear tissue was harvested 10 minutes after injection and fixed with formalin. Paraffin sections were deparaffinized, antigen retrieval was done by using AR6 buffer (PerkinElmer, Waltham, Mass), staining for MCs by using FITC-Avidin (Biolegend), and PS by using PE-anti-Flag (L5, Biolegend).
Western blot analysis
One hundred thousand BMMCs were degranulated as described above and immediately washed with ice-cold PBS at the indicated time points. Cells were lysed with 1% (wt/vol) NP40 in the presence of 1mM sodium orthovanadate. The lysates were immunoblotted with antibodies against Syk or phosphorylated Syk.
Statistical analysis
Statistical analyses were performed using GraphPad Prism software (GraphPad Software). For comparing between 2 groups, statistical significance was determined by 2-tailed unpaired Student t test with or without correction by Holm-Sidak method. For comparing more than 2 groups, statistical significance was determined by 2-way ANOVA multiple comparisons with Bonferroni or Sidak test. Error bars indicate SEM unless otherwise mentioned.
Fig E1PS exposure of degranulating MCs. A and B, Confocal images of degranulated BMMCs (Fig E1, A) and human synovial MCs (Fig E1, B) stained with anti-mouse and anti-human CD300a, respectively, and analyzed in the presence of PSVue480. C, FRET analysis of the CD300a-PS interaction on degranulated BMMCs pretreated with either isotype or anti-CD300a neutralizing antibody. Representative merged image (left) and FRET+ cell quantification (right). White arrows indicate FRET+ cells. D, Flow cytometry analysis of apoptotic mouse thymocytes or degranulating BMMCs incubated with MFG-E8-D89E, neutralizing anti-mouse CD300a (EX42), or nonneutralizing anti-mouse CD300a (TX10) together with mouse CD300a-Fc in the presence of 2 mM CaCl2, followed by a PE-conjugated antibody against human IgG and PI. E, Imaging flow cytometry analysis of degranulated BMMCs stained by anti-CD107a and annexin V. Gating (left), images (middle), and colocalization of CD107a and PS (right) of single cells from indicated gates (see theMethodssection). F, Flow cytometry analysis of BMMCs after indicated stimulations and staining with anti-CD107a, annexin V, and PI. G, Imaging flow cytometry analysis of ATP or ionomycin-stimulated BMMCs stained with anti-CD300a antibody and PSVue643. Data are representative of 2 (Fig E1, C, D, F, and G) to 3 (Fig E1, A, B, and E) independent experiments. Scale: A and B, 5 μm; C, 50 μm. APC, Allophycocyanin; NBD-PS, 1-oleoyl-2-{6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl}-sn-glycero-3-phosphoserine; PE, phycoerythrin.
Fig E2A and B, WT and Cd300a−/− BMMCs were equally mixed and degranulated by IgE-antigen complex and analyzed by flow cytometry. Data are representative plots (Fig E2, A) and the mean CD107a expression (Fig E2, B) in each genotype. C, PI+ cell before and after degranulation in WT and Cd300a−/− BMMCs. D and E, WT and Cd300a−/− BMMCs were diluted into different density (scale 50 μm) (Fig E2, D), and then CD107a+ cells were measured after degranulation in PI- gated cells (Fig E2, E). F, WT and Cd300a−/− BMMCs were stimulated with 0.5 mM ATP or 500 ng/mL ionomycin and analyzed by flow cytometry. G and H, Imaging flow cytometry of naive or antigen-stimulated BMMCs stained with anti-IgE and anti-CD300a antibodies (Fig E2, E), or PSVue643 plus either anti-IgE or anti-cKit (Fig E2, H). Data are representative of 2 (Fig E2, D-H) to 3 (Fig E2, A-C) independent experiments. APC, Allophycocyanin; FSC-A, forward scatter-area; PE, phycoerythrin; PE-Cy7, phycoerythrin-cyanin7; SSC-A, side scatter area.
Fig E3Western blot analysis of Syk phosphorylation in whole-cell lysates of degranulated WT and Cd300a−/− BMMCs at indicated time points. Data are representative of 2 independent experiments.
Fig E4Immunohistochemistry analysis of PS exposure in MCs during PSA. WT mice were sensitized with TNP-specific IgE 24 hours before injection of 50 μg MFG-E8-D89E (Flag-taged) together with or without 40 μg TNP-OVA antigen. Ten minutes after the injection, ear tissue sections were stained for MCs by FITC-conjugated Avidin and for PS by PE-conjugated anti-Flag antibody. Scale bar: 50 μm (tissue view) and 5 μm (enlarged cell view). FITC, Fluorescein isothiocyanate.
Time-lapse video of Fig 1, A. BMMCs were sensitized by anti-TNP IgE, seeded in Tyrode's buffer containing PSVue 643 with gentle addition of TNP-OVA under monitoring by confocal microscopy (the images at 10-second intervals). The speed of the video is 10 frames/s. Scale: 10 μm.
References
Elliott J.
Surprenant A.
Marelli-Berg F.
Cooper J.
Cassady-Cain R.
Wooding C.
et al.
Membrane phosphatidylserine distribution as a non-apoptotic signalling mechanism in lymphocytes.
Activation of human synovial mast cells from rheumatoid arthritis or osteoarthritis patients in response to aggregated IgG through Fcγ receptor I and Fcγ receptor II.
This study was supported in part by grants provided by the Japan Society for the Promotion of Science (KAKENHI) (grant nos. 16H06387 and 18H05022 to A.S. and grant nos. 16H05350 and 17H05495 to C.N.-O.).
Disclosure of potential conflict of interest: The authors declare that they have no relevant conflicts of interest.
30384-7&id=gr1.jpg){kind=link}
30384-7&id=gr2.jpg){kind=link}
30384-7&id=fx1.jpg){kind=link}
30384-7&id=fx2.jpg){kind=link}
30384-7&id=fx3.jpg){kind=link}
30384-7&id=fx4.jpg){kind=link}
30384-7&id=fx5.jpg){kind=link}