Migration of nerve growth cones requires detergent-resistant membranes in a spatially defined and substrate-dependent manner

Abstract

Motility of nerve growth cones (GCs) is regulated by region-specific activities of cell adhesion molecules (CAMs). CAM activities could be modified by their localization to detergent-resistant membranes (DRMs), specialized microdomains enriched in signaling molecules. This paper deals with a question of whether DRMs are involved in GC migration stimulated by three CAMs; L1, N-cadherin (Ncad), and beta1 integrin. We demonstrate that L1 and Ncad are present in DRMs, whereas beta1 integrin is exclusively detected in non-DRMs of neurons and that localization of L1 and Ncad to DRMs is developmentally regulated. GC migration mediated by L1 and Ncad but not by beta1 integrin is inhibited after DRM disruption by micro-scale chromophore-assisted laser inactivation (micro-CALI) of GM1 gangliosides or by pharmacological treatments that deplete cellular cholesterol or sphingolipids, essential components for DRMs. Characteristic morphology of GCs induced by L1 and Ncad is also affected by micro-CALI-mediated DRM disruption. Micro-CALI within the peripheral domain of GCs, or even within smaller areas such as the filopodia and the lamellipodia, is sufficient to impair their migration. However, micro-CALI within the central domain does not affect GC migration. These results demonstrate the region-specific involvement of DRMs in CAM-dependent GC behavior.

Figure 1.

CAM localization to DRMs in cerebellar granule cells. Cultured cells from P8 mice were processed by detergent extraction and sucrose density gradient centrifugation. The gradient was divided into 10 fractions from the top (fraction 1) to the bottom (fraction 10), and each fraction was tested for expression of GM1 and TfR by Western blotting (A). The GM1-positive insoluble fractions (DRM) or the TfR-positive soluble fractions (non-DRM) were combined and processed for Western blot analysis of L1 (B), Ncad (C), and β1 integrin (D). In some experiments (bottom panels in B and C), the cells were treated with 4 μM lovastatin before detergent extraction. (E) The lovastatin dose–response curves for disruption of DRMs as assessed by detergent insolubility of L1 and Ncad (B and C). The value in ordinate was defined as D/(D+N) × 100 (%), where D and N were the intensity of blots in the DRM and non-DRM fractions, respectively. Each bar represents four independent experiments. *, P < 0.05; ***, P < 0.001; one-way ANOVA followed by Dunnett's post-test compared with neurons with no lovastatin treatment.

Figure 2.

Developmental changes of L1 and Ncad localization to DRMs in the cerebellum. The cerebellums from P3 (A), P8 (B), P15 (C), P28 (D), and P56 (E) mice were processed by detergent extraction, sucrose density gradient centrifugation, and fractionation. Each fraction was tested for expression of GM1, TfR, L1, Ncad, and β1 integrin by Western blotting. The similar localization pattern was reproducibly obtained in three independent experiments.

Figure 3.

Neurite growth on L1 and Ncad is affected by depletion of cellular cholesterol or sphingolipids. (A–D) Differential interference contrast images of cerebellar granule cells cultured for 36 h on L1 (A and B) or Ncad (C and D) in the absence (A and C) or presence (B and D; 4 μM) of lovastatin. Bar, 50 μm. (E–H) Cerebellar granule cells (E and F) or DRG neurons (G and H) in culture were treated with indicated concentrations of drugs, and the length of their neurites was measured (n = 100 for each bar). *, P < 0.05; ***, P < 0.001; one-way ANOVA followed by Dunnett's post-test compared with neurons with no drug treatment (control). (I) DRG neurons were cultured in the presence of indicated concentrations of fumonisin B₁ or NB-DNJ for 8 h. Expression of GM1 and actin in the cells was detected by Western blotting.

Figure 5.

Micro-CALI of GM1 disrupts DRM integrity as assessed by Thy-1 insolubility in cold detergent. NIH-3T3 cells (A–D) or DRG GCs (E–I) incubated in the presence of FITC-CTxB (A–C and E–G), FITC-BSA (D and H), or the FITC-RGD peptide (I) were irradiated with a laser at the areas outlined in white. After detergent extraction, Thy-1 was labeled by immunocytochemistry (B, D, F, H, and I). Differential interference contrast images of Thy-1–labeled cells (B and F) are also shown (A and E, respectively). As a control, cells that had been subjected to micro-CALI but not to detergent extraction were processed for Thy-1 immunocytochemistry (C and G). Fluorescent images of FITC-CTxB (J) or the FITC-RGD peptide (K) bound to the DRG GC surface are also shown. Bars: (A–D) 20 μm; (E–I) 10 μm; (J and K) 10 μm.

Figure 4.

Codistribution of L1 and Ncad with cross-linked DRM patches in GCs. (A–Y) DRG GCs that had been treated to form visible CTxB patches were double labeled for β1 integrin (A–E), L1 (F–O), or Ncad (P–Y). GCs cultured in the presence of 4 μM lovastatin were used in some experiments (K–O and U–Y). Images showing CTxB patches (A, F, K, P, and U) were processed for intensity thresholding, and a red mask was generated on pixels whose intensity values fall within the upper 10% of the total integrated intensity in each GC (B, G, L, Q, and V). Similarly, green masks were generated (D, I, N, S, and X) on unprocessed images showing CAM distribution in the GCs (C, H, M, R, and W). Red and green masks were superimposed to reveal regions of colocalization appearing yellow (E, J, O, T, and Y). Bars, 5 μm. (Z) Quantitative analysis of CAM localization with CTxB patches using superimposed images (for example, E, J, O, T, and Y). The ordinate indicates the ratio of colocalized (yellow) mask area to the green mask area in GC. Each set of experiments involved 9–17 GCs. ***, P < 0.001; one-way ANOVA followed by Tukey's post-test compared with untreated GCs.

Figure 6.

Micro-CALI–mediated DRM disruption affects GC behavior on L1 and Ncad. Time-lapse image sequences of DRG GCs migrating on L1 (A and B), Ncad (C and D), or laminin (E). The areas outlined in black were irradiated with a laser for 30 s (from −0.5 to 0 min) in the presence of either FITC-CTxB (A, C, and E) or FITC-BSA (B and D). Bar, 5 μm.

Figure 7.

Quantitative analyses of changes in GC behavior induced by micro-CALI of GM1. (A–C) DRG GCs migrating on L1, Ncad, or laminin was irradiated with a laser in the presence of FITC-CTxB, FITC-BSA, or the FITC-RGD peptide, as shown in Fig. 6. Each set of experiments involved 7–31 GCs. Their migration rates (μm/10 min) were measured immediately before and after the laser irradiation (A and B). The average length of filopodia of each GC was quantified immediately before and 10 min after the laser irradiation (C). For this set of experiments (C), lamellipodia-dominated GCs were intentionally selected on a laminin substrate. ***, P < 0.001; paired t test compared with GCs before laser irradiation. (D) Migration rates of DRG GCs on L1 (n = 8) and Ncad (n = 5) were measured every 10 min. The entire area of the GC was subjected to micro-CALI of GM1 at 0 min.

Figure 8.

GC migration on L1 and Ncad is impaired by micro-CALI–mediated DRM disruption in the P-domain but not in the C-domain. (A–C) Time-lapse image sequences of DRG GCs migrating on Ncad (A and B) or L1 (C). Either the P-domain or the C-domain was subjected to micro-CALI of GM1 at the areas outlined in black at 0 min (A–C) and 20 min (C). Bars: (A and B) 5 μm; (C) 10 μm. (D) Migration rates of DRG GCs immediately before and after laser irradiation within the P-domain or the C-domain in the presence of either FITC-CTxB or FITC-BSA. Each set of experiments involved 13–25 GCs. ***, P < 0.001; paired t test compared with migration rates before laser irradiation.

Figure 9.

Micro-CALI–mediated DRM disruption in either the lamellipodia or the filopodia is sufficient to impair GC migration on L1 and Ncad. (A and B) Time-lapse image sequences of DRG GCs migrating on L1. Either the lamellipodia (A) or the filopodia (B) was subjected to micro-CALI of GM1 at the areas outlined in black. All the filopodia were targeted by four times of laser irradiation (B). Bars, 5 μm. (C) Migration rates of DRG GCs immediately before and after micro-CALI of GM1 in the lamellipodia or the filopodia. Each set of experiments involved 10–12 GCs. ***, P < 0.001; paired t test compared with migration rates before laser irradiation.