VEGFR-1 expressed by malignant melanoma-initiating cells is required for tumor growth

Abstract

Melanoma growth is driven by malignant melanoma-initiating cells (MMIC) identified by expression of the ATP-binding cassette (ABC) member ABCB5. ABCB5(+) melanoma subpopulations have been shown to overexpress the vasculogenic differentiation markers CD144 (VE-cadherin) and TIE1 and are associated with CD31(-) vasculogenic mimicry (VM), an established biomarker associated with increased patient mortality. Here we identify a critical role for VEGFR-1 signaling in ABCB5(+) MMIC-dependent VM and tumor growth. Global gene expression analyses, validated by mRNA and protein determinations, revealed preferential expression of VEGFR-1 on ABCB5(+) tumor cells purified from clinical melanomas and established melanoma lines. In vitro, VEGF induced the expression of CD144 in ABCB5(+) subpopulations that constitutively expressed VEGFR-1 but not in ABCB5(-) bulk populations that were predominantly VEGFR-1(-). In vivo, melanoma-specific shRNA-mediated knockdown of VEGFR-1 blocked the development of ABCB5(+) VM morphology and inhibited ABCB5(+) VM-associated production of the secreted melanoma mitogen laminin. Moreover, melanoma-specific VEGFR-1 knockdown markedly inhibited tumor growth (by > 90%). Our results show that VEGFR-1 function in MMIC regulates VM and associated laminin production and show that this function represents one mechanism through which MMICs promote tumor growth.

Figure 1

Vasculogenic/angiogenic pathways in human melanoma. A, tumorigenicity of ABCB5⁺ versus ABCB5⁻ melanoma cells in human to NSG mouse xenotransplantation experiments. B, representative immunofluorescence staining of ABCB5 (green) and CD271 (red) expression in clinical melanoma specimens; nuclei (blue). C, pathway activation across ABCB5⁺ MMIC. Genes represented by red nodes (circles) are overexpressed in ABCB5⁺ relative to ABCB5⁻ human melanoma cells; those represented by black nodes are expressed at lower levels. Black lines show known gene interactions, and gene functions in vasculogenesis/angiogenesis or as drug targets are annotated (red lines). Gene relationships are based on Ingenuity Pathway Analysis. D, VEGFR-1 mRNA expression determined by real-time PCR in ABCB5⁺ versus ABCB5⁻ human melanoma cells. E, representative flow cytometric plots of VEGFR-1 protein expression on ABCB5⁺ MMIC (top) and ABCB5⁻ melanoma cells (bottom). Aggregate mean percentages are shown on the right. F, representative immunofluorescence double staining of ABCB5 (red) and VEGFR-1 (green) expression in melanoma specimens, with nuclei counterstained in blue. Arrows indicate zones of membrane coexpression (yellow).

Figure 2

VEGF/VEGFR-1 signaling in human melanoma cells. A, representative immunofluorescence staining for CD144 (top) and VWF (bottom) expression (red) by purified ABCB5⁺ or ABCB5⁻ melanoma cells before and after VEGF treatment; nuclei (blue). Aggregate analysis of 6 distinct melanoma specimens is shown on the right. B, representative immunofluorescence staining for CD144 expression (red) by melanoma cells treated with VEGF as above but in the presence or absence of anti-VEGFR-1 blocking mAb or isotype control mAb; nuclei (blue). Aggregate analysis of 6 distinct melanoma specimens is shown on the right. C, tube formation of melanoma cells treated with VEGF in the presence or absence of anti-VEGFR-1 blocking mAb or isotype control mAb. Aggregate analyses of numbers of tubes per microscopy field and tube lengths (means ± SE, n = 3 replicate experiments) are shown on the right. NS, nonsignificant.

Figure 3

In vivo expression of the VM-associated markers ABCB5 and laminin. A–D, immunohistochemistry for ABCB5 protein (left), laminin (middle), and ABCB5 (red)/laminin (green) immunofluorescence double staining (right) in (A) clinical melanoma, (B) clinical melanoma xenografts, and (C) melanoma line xenografts, detecting identical patterns of VM-associated reticular channel-like reactivity (also detected in D) by PAS staining. Zones of close spatial association between ABCB5 and human laminin are indicated by arrows (A–C, right). E, ISH for ABCB5 mRNA (inset is sense control).

Figure 4

Detection of communicating VM patterned networks. A, immunohistochemistry of tumor xenografts after intravenous administration of anti-ABCB5 antibody (left) or control Ig (right) to melanoma xenograft-bearing mice. B, conventional histology discloses channels to be associated with linear lamella of PAS-positive extracellular matrix (inset) intimately associated with tumor cells (arrows). H&E, hematoxylin and eosin. C, transmission electron microscopy of a melanoma specimen depicting conventional tumor angiogenesis involving formation of spaces lined by flattened endothelial cells, containing erythrocytes, and surrounded by tumor cells (left) and other channels (middle and right), also containing erythrocytes, which are lined by extracellular matrix consistent with basement membrane (middle, arrows; right at high magnification, asterisks) and surrounded by tumor cells.

Figure 5

In vivo requirement for VEGFR-1 for efficient tumor growth. A, TV (mean ± SE) and (B) representative immunofluorescence double staining of ABCB5 (red) and laminin (green) expression with nuclear counterstaining (blue) of VEGFR-1⁺ versus VEGFR-1⁻ melanoma cell–derived xenografts. C, VEGFR-1 mRNA expression (top) and protein expression (bottom, molecular size unit: kD) in VEGFR-1 shRNA–transfected melanoma cells compared with controls; untr. ctrl., untreated control. D, representative immunohistochemistry (human ABCB5, human laminin, murine CD31) and immunofluorescence double staining of human ABCB5 and laminin (middle), revealing in the case of ABCB5 and laminin the extent of VM in melanomas that developed from control versus VEGFR-1 shRNA knockdown tumor xenografts or, in the case of CD31, the extent of the physiologic angiogenic response; mag, magnification. E, quantitative image analysis of laminin VM immunoreactivity for melanomas derived from control or VEGFR-1 shRNA–transfected melanoma xenografts (n = 6 recipient mice per experimental group). Y-axis, percentage of pixelated area with reactivity (mean ± SE). F, TVs (mean ± SE) 3 weeks following xenotransplantation of control or VEGFR-1 shRNA–transfected human melanoma cells (left). Typical macroscopic appearance of tumors dissected 3 weeks following transplantation of control shRNA- or VEGFR-1 shRNA–transfected melanoma cells (right).