Fig 1.
Schematic depiction of the FCCS approach.
An array of antibodies against cell-surface antigens is printed on a hydrogel coated glass slide. Live cells in suspension are captured on specific antibody spots by interaction between their surface antigens and the printed antibodies. The captured cells are fixed and labeled for one or more intracellular markers of functional relevance (e.g. insulin, glucagon, somatostatin for pancreatic beta, alpha and delta cells, respectively). Populated spots enriched by cells with the desired label designate markers that are preferentially associated with the respective cell functionality. Following validation of the marker, it is used to isolate cells by flow cytometry. The isolated cells are then used as an input sample for the next iteration of this procedure. The iterative application of the procedure allows identification of additional markers that further refine the isolation.
Fig 2.
Identification and validation of markers for initial enrichment of specific cell types within human islets of Langerhans.
(a) Representative images of populated antibody spots with different enrichments of insulin- (red), glucagon- (green) and somatostatin-positive cells (blue). On the right are spots mostly populated by non-endocrine cells, and on the left are spots populated by different proportions of alpha, beta and delta cells. Corresponding phase contrast images are shown on the right of each image (10x magnification). (b) Real-time qPCR analysis of several cell-type specific genes in different marker-isolated populations (insulin for beta cells, glucagon for alpha cells, somatostatin for delta cells and trypsin for acinar cells). CD9+ refers to top 10% expressing cells (CD9high). Shown is mean expression relative to unsorted (bulk) cells +/- SE (n = 3; biological replicates correspond to different donors; * p < 0.05, ** p < 0.01).
Table 1.
Cell-surface markers expressed by CD9+ cells as detected by the second iteration of the FCCS.
Fig 3.
Second iteration of FCCS identifies marker combinations for improving the isolation of insulin producing cells.
(a) Representative images of spots containing top 10% CD9-expressing cells (CD9high), immunostained for insulin and somatostatin. Spots with high, medium and low beta cell enrichment are presented from top to bottom. Respective phase contrast images (10x magnification) are shown to the right. (b) qPCR analysis of cell-type specific genes in islet cells fractionated based on CD9/CD56 combinations (CD9high/CD56+, CD9high/CD56-, CD9-/CD56+ and CD9-/CD56-). Shown is mean expression relative to unsorted (bulk) cells +/- SE (n = 3; biological replicates correspond to different donors; * p < 0.05, ** p < 0.01).
Fig 4.
Improved purity and yield following a second iteration of the FCCS.
(a) Left: Flow cytometry analysis of CD9, CD56 and intracellular insulin expression in islet samples. Co-localization of insulin+ cells (blue overlay) with CD9high/CD56+ expressing cells (determination of gated cells is described in S2 File and Figure D of S1 File). Insulin- cells are shown in red. Right: histogram of all cells (red) vs. the insulin+ population (blue) and the top 20% of CD9-expressing cells within the CD56+ population (green). Dashed line indicates the gate for top 10% of the population by CD9 expression. This demonstrates the potential of the CD9/CD56 combination to identify more relevant cells as compared to only CD9. (b) Flow cytometry analysis of CD9, CD56 and intracellular glucagon (left) or somatostatin (right) in islets samples. Staining for glucagon and somatostatin is overlaid in blue on the CD9/CD56 panel, demonstrating co-localization of glucagon with CD9-/CD56+ expressing cells (left) and co-localization of somatostatin with CD9high/CD56+ (right). Glucagon and somatostatin negative cells are shown in red. (c) Left: comparison of purity (% of insulin+ cells in the isolated fraction) obtained by CD9/CD56-based isolation and CD9-based isolation at different choices of gating. Right: Comparison of CD9/CD56-based and CD9-based isolation with respect to beta cell yield (fraction of insulin+ cells out of total beta cells) and purity. The yield is calculated based on the same gates as the purity. Changes in CD9/CD56 gates were made by modifying the threshold of isolation equally along both axes. (d) Tabulation of the data points in (c).