Blocking CLEC14A-MMRN2 binding inhibits sprouting angiogenesis and tumour growth

Abstract

We previously identified CLEC14A as a tumour endothelial marker. Here we show that CLEC14A is a regulator of sprouting angiogenesis in vitro and in vivo. Using a human umbilical vein endothelial cell spheroid-sprouting assay, we found CLEC14A to be a regulator of sprout initiation. Analysis of endothelial sprouting in aortic ring and in vivo subcutaneous sponge assays from clec14a(+/+) and clec14a(-/-) mice revealed defects in sprouting angiogenesis in CLEC14A-deficient animals. Tumour growth was retarded and vascularity reduced in clec14a(-/-) mice. Pull-down and co-immunoprecipitation experiments confirmed that MMRN2 binds to the extracellular region of CLEC14A. The CLEC14A-MMRN2 interaction was interrogated using mouse monoclonal antibodies. Monoclonal antibodies were screened for their ability to block this interaction. Clone C4, but not C2, blocked CLEC14A-MMRN2 binding. C4 antibody perturbed tube formation and endothelial sprouting in vitro and in vivo, with a similar phenotype to loss of CLEC14A. Significantly, tumour growth was impaired in C4-treated animals and vascular density was also reduced in the C4-treated group. We conclude that CLEC14A-MMRN2 binding has a role in inducing sprouting angiogenesis during tumour growth, which has the potential to be manipulated in future antiangiogenic therapy design.

Figure 1

SiRNA knockdown of CLEC14A reveals a role for CLEC14A in endothelial sprouting. [A] SiRNA duplexes targeting CLEC14A can efficiently knockdown CLEC14A mRNA expression in HUVEC, as determined by qPCR. Relative expression was determined by normalising expression to flotillin2. [B] Knockdown of CLEC14A at the protein level was determined for both siRNAs by Western blot analysis. Tubulin was used as a loading control. [C] Representative images of sprout outgrowth after 16 hours for control or clec14a targeted siRNA treated HUVEC. Scale bars are equal to 100 μm. [D] Quantitation of sprouts for 27 spheroids (9 spheroids from 3 cords) for control and CLEC14A knockdown HUVEC; Kruskal-Wallis statistical test p<0.001. [E] Representative images of sprout outgrowth after 24 hours for mixed control (green) and clec14a targeted siRNA treated HUVEC (red). Scale bars are equal to 100 μm. [F] Quantitation of the percentage of tip and stalk cells derived from control (CON) and CLEC14A knockdown (KD) HUVEC; two-way ANOVA statistical test with Bonferroni post-tests *** = p<0.001, ns = not significant.

Figure 2

Loss of CLEC14A inhibits sprouting in vitro and in vivo. [A] Schematic diagram of clec14a gene in C57BL/6 (clec14a^+/+) or C57BL/6^{(Clec14atm1(KOMP)Vlcg)} (clec14a^−/−) mice. [B] Quantitative PCR analysis of cDNA generated from three clec14a^+/+ mice (white bars) and three clec14a^−/− mice (black bars) for the 5′ untranslated region (UTR), coding sequence (CDS) and 3′ UTR of clec14a. Relative expression was determined by normalising expression to flotillin2. [C] Western blot analysis of CLEC14A protein expression in lung lysates from clec14a^+/+ and clec14a^−/− mice using polyclonal antisera against murine CLEC14A. Tubulin was used as a loading control. [D] Representative images of the aortic ring sprouting assay from clec14a^+/+ and clec14a^−/− mice. Scale bars are equal to 200 μm. Quantitation of tubes formed per ring [E], and quantitation of the longest distance migrated away from the aortic ring by an endothelial tube per aortic ring [F], data from 48 rings per genotype, 6 mice for each genotype; Mann-Whitney statistical test p<0.001. [G] Representative images of haematoxylin and eosin stained sections of sponge implant from clec14a^+/+ and clec14a^−/− mice, sections at the centre of the sponge were analysed. Black and white images represent the masks generated during the threshold analysis for quantitation. [H] Quantitation of cellular invasion, by threshold analysis of haematoxylin and eosin stained cellular material within the sponge implants shown in G; Mann-Whitney statistical test p<0.05. [I] Quantitation of vessel density; Mann-Whitney statistical test p<0.001. [J] Sections of liver and sponge tissue stained with x-gal from clec14a^−/− mice, counterstained with haematoxylin and eosin. Arrows indicate vessels stained with x-gal and the increased intensity in the sponge tissue compared to the liver. Scale bars are equal to 100 μm.

Figure 3

Loss of CLEC14A inhibits tumour growth. [A] Lewis lung carcinoma (LLC) tumour growth in clec14a^+/+ (black line with dots) and clec14a^−/− (black line with squares) mice; two-way ANOVA statistical analysis, * = p<0.05, ** = p<0.01, *** = p<0.001. [B] Representative images of LLC tumours. [C] Endpoint tumour weight for 7 clec14a^+/+ (dots) and 7 clec14a^−/− (squares) mice; Mann-Whitney statistical test p<0.001. [D] Representative images of immunofluorescent staining of LLC tumour sections stained for murine CD31. Scale bars are equal to 100 μm. Quantitation of vessel density [E] and percentage endothelial coverage [F] from clec14a^+/+ and clec14a^−/− mice; Mann-Whitney statistical test p<0.0001. [G] Sections of liver and LLC tumour tissue from clec14a^−/− mice stained with x-gal, counterstained with haematoxylin and eosin. Scale bars are equal to 100 μm.

Figure 4

MMRN2 binds to CLEC14A. [A] 20 μg CLEC14A-ECD-Fc or Fc was used to precipitate interacting partners. Precipitates and HUVEC lysates were separated on an SDS-PAG and blotted for MMRN2 (top panel) or CLEC14A-ECD-Fc (bottom panel). [B] CLEC14A was immunoprecipitated from HUVEC lysates using polyclonal antisera against CLEC14A. Immunoprecipitates were analysed by Western blot for MMRN2 (top panel) and CLEC14A (bottom panel).

Figure 5

CLEC14A monoclonal antibodies block MMRN2-CLEC14A interaction. [A] HEK293T cells expressing HA-CLEC14A or an empty vector (CON) were stained with an HA tag antibody (column 1), or monoclonal antibodies against CLEC14A C2 (column 2) or C4 (column 3). Cells were analysed by flow cytometry and displayed as histograms of increasing fluorescence (x-axis) versus counts (y-axis). [B] HUVECs transfected with negative control siRNA duplexes or siRNA duplexes targeting CLEC14A were probed with C2 or C4 antibodies and analysed as in A. [C] HUVECs were pre-treated with blocking buffer (−), 100 μg C2 antibody or 100 μg C4 antibody, prior to C2-FITC staining. Cells were analysed by flow cytometry. Geometric means were normalised to staining for the cells pre-treated with blocking buffer. [D] as for C, except stained with C4-FITC. [E] CLEC14A-Fc (5 μg) protein G agarose bead complexes were blocked with 7.5, 15, 30, μg mIgG or C2, prior to MMRN2 pulldown from HEK293T lysates. Precipitates were separated by SDS-PAGE and blotted for MMRN2 (upper panel) and CLEC14A-Fc (lower panel). [F] as in E, except C4 was used to block instead of C2.

Figure 6

MMRN2-CLEC14A interaction blocking antibody inhibits endothelial tube formation and sprouting in vitro and in vivo. HUVECs were plated onto Matrigel and grown in the presence of 20 μg/ml mIgG1 (CON), C2, or C4 antibodies for 16 hours. [A] Representative images. Scale bars are equal to 100 μm. Tube formation was analysed for number of meshes [B], number of branches [C] and the branch length [D]. Representative data from 1 of 3 independent experiments; Kruskal-Wallis statistical test, * = p<0.05, ** = p<0.01, ns = not significant. HUVEC spheroids embedded in a collagen gel were stimulated to sprout with VEGF supplemented with 20 μg/ml mIgG1 (CON), C2 or C4. [E] Representative images. Scale bars are equal to 100 μm. [F] Quantitation of sprouts per spheroid for 27 spheroids from 3 independent experiments; Kruskal-Wallis statistical test *** = p<0.001, ns = not significant. Aortic rings from C57BL/6 mice were cultured in the presence of 20 μg/ml mIgG1 (CON), C2 or C4. [G] Representative images. Scale bars are equal to 200 μm. [H] Quantitation of tubes formed per ring, data from 30 rings, at least 6 mice were used for each condition; Kruskal-Wallis statistical test *** = p<0.001, ns = not significant. [I] Mmrn2 expression can be efficiently knocked down by siRNA duplexes in HUVEC, as determined by qPCR. Relative expression was determined by normalising expression to flotillin2. [J] Endothelial sprout outgrowth after 16 hours VEGF treatment of control or mmrn2 targeted siRNA treated HUVEC spheroids. Quantitation of sprouts for 27 spheroids (9 spheroids from 3 cords); Kruskal-Wallis statistical test p<0.001. [K] Representative images of sponge implants injected with bFGF and mIgG1 (CON) or C4 antibody. [L] Quantitation of cellular invasion into these sponge implants by threshold analysis of haematoxylin and eosin stained cellular material; Mann-Whitney statistical test p<0.01. [M] Quantitation of vessel density from K; Mann-Whitney statistical test p<0.001.

Figure 7

MMRN2-CLEC14A interaction blocking antibody inhibits tumour growth. [A] Mice injected with LLC were treated with 100 μg injections of mIgG1 (black line with dots; n=7) or C4 antibody (black line with squares; n=7); two-way ANOVA statistical analysis, ** = p<0.01, *** = p<0.001. [B] Representative images of LLC tumours. [C] Endpoint tumour weight for 7 mIgG1 treated mice (dots) and 7 C4 antibody treated mice (squares); Mann-Whitney statistical test p<0.001. [D] Representative images of immunofluorescent staining of LLC tumour sections stained for murine CD31. Scale bars are equal to 100 μm. Quantitation of vessel density [E] and percentage endothelial coverage [F] from mice treated with mIgG1 or C4 antibody; Mann-Whitney statistical test p<0.001.

Figure 8

CLEC14A-MMRN2 binding is an important component of endothelial sprout formation and a regulator of tumour growth. ECM = extracellular matrix, PM = plasma membrane.