Stromal cells support the survival of human primary chronic lymphocytic leukemia (CLL) cells through Lyn-driven extracellular vesicles

Abstract

Introduction: In chronic lymphocytic leukemia (CLL), the tumor cells receive survival support from stromal cells through direct cell contact, soluble factors and extracellular vesicles (EVs). The protein tyrosine kinase Lyn is aberrantly expressed in the malignant and stromal cells in CLL tissue. We studied the role of Lyn in the EV-based communication and tumor support.

Methods: We compared the Lyn-dependent EV release, uptake and functionality using Lyn-proficient (wild-type) and -deficient stromal cells and primary CLL cells.

Results: Lyn-proficient cells caused a significantly higher EV release and EV uptake as compared to Lyn-deficient cells and also conferred stronger support of primary CLL cells. Proteomic comparison of the EVs from Lyn-proficient and -deficient stromal cells revealed 70 significantly differentially expressed proteins. Gene ontology studies categorized many of which to organization of the extracellular matrix, such as collagen, fibronectin, fibrillin, Lysyl oxidase like 2, integrins and endosialin (CD248). In terms of function, a knockdown of CD248 in Lyn⁺ HS-5 cells resulted in a diminished B-CLL cell feeding capacity compared to wildtype or scrambled control cells. CD248 is a marker of certain tumors and cancer-associated fibroblast (CAF) and crosslinks fibronectin and collagen in a membrane-associated context.

Conclusion: Our data provide preclinical evidence that the tyrosine kinase Lyn crucially influences the EV-based communication between stromal and primary B-CLL cells by raising EV release and altering the concentration of functional molecules of the extracellular matrix.

Keywords: CD248; Lyn kinase; chronic lymphocytic leukemia; extracellular matrix; extracellular vesicle; filopodia.

FIGURE 1

Influence of tyrosine kinase inhibitor Dasatinib on the release of extracellular vesicles (EVs) in stromal cells. HS-5 (A) and WI-38 (B) cells (2 × 10⁶/ml) were cultivated at 37°C in serum-free RPMI-1640 with dilution series of Dasatinib as indicated. DMSO (0.1% in PBS) served as a control. After 24 h, cells and the supernatant were separated by centrifugation (10 min, 300 × g). The supernatant was cleared by centrifugation at 3,000 × g and afterwards subjected to ultracentrifugation (90 min, 110,000 × g). After ultracentrifugation, the EV pellet was suspended in PBS and the count was determined by nanoparticle tracking analysis. The cell pellet was used for cell viability determination. The viability of the DMSO-treated aliquots served as controls (100%). The statistic evaluation was performed by the Kruskal–Wallis test.

FIGURE 2

Influence of Lyn on the extracellular vesicles (EV) release in stromal cells. HS-5 cells (wt) and Lyn-deficient counterparts (KO) were compared. (A) The expression of Lyn was determined in both cell types by western blot. (B) The growth conditions were tested in a kinetic study by determination of the cell count at time points as indicated. (C) The EVs in the cell-free supernatants were isolated by sedimentation for 90 min at 110,000 × g and visualized by cryo-transmission electron microscopy (TEM). (D) To compare the EV concentration both HS-5 cell types were cultivated for 24 h in medium with EV-depleted serum. After precentrifugation steps at 300 × g, 3,000 × g, and 10,000 × g, the EVs were sedimented at 110,000 × g, suspended in 500 μL PBS and purified by size exclusion chromatography (SEC). Then, the EV count and diameter of the suspension was determined by nanoparticle tracking analysis. (E) To plot which EV size is predominantly influenced by Lyn in HS-5 cells, the EV diameter and count are shown in a histogram. A dashed line separates particles ≤ 200 nm (s-EVs) and ≥ 200 nm (l-EVs). The percentage of particles < 200 nm and > 200 nm is calculated and depicted in a bar chart. The influence of Lyn on the EV release was also tested in the supernatant of wt and Lyn-defective StromaNKtert cells, another stromal cell line. (F) The expression of Lyn was determined in both StromaNKtert types by western blot. (G) The particle count and diameter were determined by nanoparticle tracking analysis. This time, the EVs were not purified by SEC but aliquots were filtered through a 200 nm filter. (H) To plot which EV size is predominantly influenced by Lyn in HS-5 cells, the EV diameter and count are shown in a histogram. A dashed line separates particles ≤ 200 nm (s-EVs) and ≥ 200 nm (l-EVs). The percentage of particles < 200 nm and > 200 nm is calculated and depicted in a bar chart. The results were statistically evaluated by a two-tailed, non-parametric t-test (Mann–Whitney) (ns = not significant, *P < 0.05, ^***P < 0.001).

FIGURE 3

Influence of Lyn from stromal cells on the extracellular vesicles (EV) uptake. (A) The uptake/binding of EVs from MEC-1 cells (chronic lymphocytic leukemia, CLL) was tested in Lyn-proficient and Lyn-deficient HS-5 cells. EVs from MEC-1 cells were isolated by ultracentrifugation for 90 min at 110,000 × g, stained with DiD and washed with PBS. Both HS-5 cell types were cultivated for 24 h in the presence of a dilution series (1–10 μg/ml) of these EVs. Then, the cells were washed and analyzed by flow cytometry. (B) Confocal microscopy of Lyn-proficient and Lyn-deficient HS-5 cells with CSFE-labeled MEC-1 EVs. Cells were incubated with labeled EVs and the cell membrane was subsequently stained with CellMask deep red. (C) The internalization process was further analyzed by imaging flow cytometry investigation of 10,000 wt and Lyn-deficient HS-5 cells. Representative images are shown of Lyn-proficient (upper left) and Lyn-deficient HS-5 cells (lower left), incubated with CSFE-labeled EVs for 60 min at 37 °C. The loading controls of three independent experiments with their geo means are shown in the middle panel and the internalized intensity (internalization erode) of the green fluorescence with the geo means is depicted (right panel). The results were statistically evaluated by a two-tailed, parametric t-test (Mann–Whitney) (^**P < 0.01, ns, not significant).

FIGURE 4

Influence of Lyn on filopodia in stromal cells. (A) wt HS-5 cells were transfected with CD63-eGFP and grown to confluency. Then, cells were treated with Alexa Fluor Phalloidin (594) to stain F-actin. Images with single stains and a fused image are depicted. Arrow heads show protrusion-associated CD63⁺ extracellular vesicles (EVs). (B) Filopodia were identified in Lyn-proficient and -deficient HS-5 cells after F-actin staining. Filopodia were measured and calculated in fixed cells using the FiloQuant plugin of Fiji software (19). (C) Filopodia count per cell and filopodia length were determined. The results were statistically evaluated by a two-tailed, unpaired, non-parametric t-test (Mann–Whitney) (^****P < 0.0001).

FIGURE 5

Lyn contributes to generate chronic lymphocytic leukemia (CLL)-supportive extracellular vesicles (EVs). (A) Primary CLL cells (0.3 × 10^6/well in a 96-well plate) of 11 different donors were cultivated in RPMI-1640 medium with 10% FBS and a dilution series of purified EVs from HS-5 cells as indicated. After 24 h the cell viability was determined with CellTiter-Glo 2.0. Statistics: Brown–Forsythe ANOVA test. (B) The differential effect of EVs (4 μg/ml) from Lyn-proficient and -deficient HS-5 cells on the viability of primary CLL cells was compared with untreated CLL cells in a kinetic study (0–96 h). At the end of the indicated incubation period, the cell viability was determined with CellTiter-Glo 2.0. Statistics: two-tailed, unpaired t-test with Welch’s correction at indicated time points (^***P < 0.001, ^****P < 0.0001).

FIGURE 6

Proteomic study of extracellular vesicles (EV) proteins from Lyn-proficient and deficient HS-5 cells. Differentially expressed proteins from EVs of Lyn-proficient and -deficient HS-5 cells were identified under the following condition (p-adjusted ≤ 0.05; 1 ≤ log₂ fold change ≤ –1) (Supplementary Table 1). Gene Ontology (GO) studies are depicted, i.e., molecular function (A), cellular component (B), and the Reactome (C). Enrichment of proteins with a relation to the extracellular matrix are shown (D).

FIGURE 7

CD248 contributes to the support of chronic lymphocytic leukemia (CLL) cells. (A) CD248-defective HS-5 cells were generated. The defect was confirmed by western blot in cell lysates of HS-5 wt, CD248 knockout (CD248 KO), and the scrambled control (scr3). (B) Primary CLL cells (1.5 × 10⁶/ml) were cocultivated with HS-5 cell variants as indicated (1 × 10⁴/ml) or without HS-5 cells. After indicated time points, the non-adherent CLL cells were suspended and transferred to another well for the cell viability measurement using CellTiter-Glo2 as indicated. The results show the means ± SEM of nine different patients. Statistics is shown between CD248 KO and the scr3 control at indicated time points. (C) To test the influence of CD248 on the release of particles, wt and CD248-defective HS-5 cells were cultivated for 24 h in medium with extracellular vesicles (EV)-depleted serum. Then, the particles count was determined in the supernatants of seven aliquots by nanoparticle tracking analysis. (D,E) Filopodia were identified in fixed and F-actin-stained confocal images of wt and CD248 KO HS-5 cells. The filopodia count per cell and filopodia length were measured and calculated using the FiloQuant plugin of Fiji software (19). The non-parametric, two-tailed t-test (Mann–Whitney) was performed when indicated (*P < 0.05, **P < 0.01, ns, not significant).