Utilization of commercial collagens for preparing well-differentiated human beta cells for confocal microscopy

Abstract

Introduction: With technical advances, confocal and super-resolution microscopy have become powerful tools to dissect cellular pathophysiology. Cell attachment to glass surfaces compatible with advanced imaging is critical prerequisite but remains a considerable challenge for human beta cells. Recently, Phelps et al. reported that human beta cells plated on type IV collagen (Col IV) and cultured in neuronal medium preserve beta cell characteristics.

Methods: We examined human islet cells plated on two commercial sources of Col IV (C6745 and C5533) and type V collagen (Col V) for differences in cell morphology by confocal microscopy and secretory function by glucose-stimulated insulin secretion (GSIS). Collagens were authenticated by mass spectrometry and fluorescent collagen-binding adhesion protein CNA35.

Results: All three preparations allowed attachment of beta cells with high nuclear localization of NKX6.1, indicating a well-differentiated status. All collagen preparations supported robust GSIS. However, the morphology of islet cells differed between the 3 preparations. C5533 showed preferable features as an imaging platform with the greatest cell spread and limited stacking of cells followed by Col V and C6745. A significant difference in attachment behavior of C6745 was attributed to the low collagen contents of this preparation indicating importance of authentication of coating material. Human islet cells plated on C5533 showed dynamic changes in mitochondria and lipid droplets (LDs) in response to an uncoupling agent 2-[2-[4-(trifluoromethoxy)phenyl]hydrazinylidene]-propanedinitrile (FCCP) or high glucose + oleic acid.

Discussion: An authenticated preparation of Col IV provides a simple platform to apply advanced imaging for studies of human islet cell function and morphology.

Keywords: CNA35; lipid droplets; mitochondria; type IV collagen; type V collagen.

Figure 1

Single-cell suspension of human islet cells was seeded on glass surfaces coated with type IV collagen 5533 (C5533), type IV collagen 6745 (C6745), and type V collagen (Col V) and cultured for 5 days. (A) The number of nuclei per an islet cell cluster and (B) the proportion of INS⁺ and GCG⁺ cells were counted after staining for insulin (INS, green), glucagon (GCG, white), and DAPI (blue). (C) Representative images. (D) Human islet cells prepared as in (A) and stained for INS (green), NKX6.1 (red), and DAPI (blue) and % of INS+ cells with nuclear NKX6.1 was measured. (E) representative images. Data are mean ± SEM. n = 4 donors. Scale bar, 10 μm. Statistics by RM one-way ANOVA. n.s.; not significant, *p < 0.05.

Figure 2

Dispersed human islet cells seeded and cultured on type IV collagen 5533 (C5533), type IV- collagen 6745- (C6745), and type V collagen- (Col V) coated glass surfaces were visualized by anti-insulin antibody (green) and DAPI (blue). (A) Representative 20× objective epifluorescence microscope image showing DAPI (scale bar, 100 μm). (B) The number of clusters and (C) cell cluster area in 0.44 mm² field were counted and expressed as an average for C5533 in each of five donors as 100%. n = 3 fields per donor × 5 donors total 15. (D-F) Z-stack images of islet cell clusters on each collagen surface was captured via a confocal microscope and processed by Imaris software to count the number of nuclei along the Z-axis for each cluster (D) average number of nuclei for each donor (n = 4 donors). (E) data expressed taking average of C5533 as 100% for each donor (n = 8 clusters per donor × 4 donors = 32). (F) 3D representative images (axis in μm, see Supplementary Video 1 ). (G, H) Beta-cell border was visualized by a mixture of anti-NTPDase3 and syntaxin 1 antibodies (green). Beta-cell identity was confirmed by nuclear expression of NKX6.1 (white). Beta-cell area was measured as in methods in three donors. n = 6 to 29 cells. Each dot represents one cell. Data are mean ± SEM. Statistics by RM one-way ANOVA. n.s, not significant; *p < 0.05.

Figure 3

(A) 1.5 μg/lane of type IV collagen C5533 (C5533), type IV collagen C6745 (C6745), and type V collagen (Col V) preparations were run on SDS-PAGE under reducing condition and visualized by silver staining. The 50-kDa band of C6745 (red rectangle) was analyzed by mass spectrometry ( Supplementary Figure 2A ). (B) Two lots of C6745 (Lot#1 and #2) were separated by SDS-PAGE under non-reducing (-DTT) and reducing (+DTT) conditions and visualized by silver staining (left) or by immunoblot for rabbit IgG (right) after transfer to a nitrocellulose membrane. (C) 1.5 μg of C5533, C6745 (Lot#1 and Lot#2), and Col V were applied to a nitrocellulose membrane and incubated with CNA35, as described in the methods. (D) 5.7 μg/lane for each collagen preparation was separated by SDS-PAGE and incubated with CNA35, as described in the methods.

Figure 4

(A) Insulin contents and (B) insulin secreted at indicated glucose concentrations were obtained from human islet cell clusters cultured for 5 days on type IV collagen C5533- (C5533), type IV collagen C6745- (C6745), and type V collagen- (Col V) coated glass surfaces as described in methods. (C) Ratio of insulin secreted at high and low glucose. n = 7 donors (A) and 8~10 donors (B, C). (D, E) TIRF image was captured for insulin-positive granules on glass surface coated by three coating preparations after incubation in 2.8 or 16.8 mM glucose for 30 min. (D) The number of granules per cell for each donor (n = 4). (E) Data are expressed taking the average number of granules per cell for cells on C5533 at 2.8 mM glucose in each donor as 100% and data from four donors are combined (n = 8~21). Data are mean ± SEM except for (B) and (E) that indicate medium. Statistics by RM one-way ANOVA (A) or mixed-effect models (B, C). n.s, not significant; *p < 0.05.

Figure 5

The morphology of mitochondria was assessed in beta and alpha cells plated on C5533 and culture in neuronal medium only or with addition of FCCP or glucose + OA. (A) Representative images showing mitochondria-GFP (green), INS or GCG (white), and DAPI (blue). Scale bar, 10 μm. (B) Mitochondrial length and (C) form factors in beta and alpha cells were obtained in three donors. Each dot represents the mean value of data obtained from 10 to 20 images for each donor (see Supplementary Figures 3C–F from which mean values were obtained). Data are mean ± SEM. RM one-way ANOVA with Dunnett’s multiple-comparison test. n.s; not significant, *p < 0.05.

Figure 6

Human islets from three donors aged 16 (donor 1), 32 (donor 2), and 60 (donor 3) year-old (yo) were plated on C5533 and cultured in neuronal medium only or with addition of FCCP or glucose + OA as in methods. LDs were visualized by overnight incubation with Bodipy C12 (red). (A) Representative images from donor 2. DAPI in blue. Beta cells are marked by a white line. Scale bar, 10 μm. (B) Number of LDs corrected by beta-cell area. n = 8 to 20 images. (D) Size of LDs expressed as area (μm²). Each dot represents single LD and data combine LDs from 8 to 20 images for each condition. Data are mean ± SEM. Statistics by one-way ANOVA with Dunnett’s multiple-comparison test. n.s; not significant, and *p < 0.05.