Integrin αvβ3 acting as membrane receptor for thyroid hormones mediates angiogenesis in malignant T cells

Abstract

The interaction of lymphoid tumor cells with components of the extracellular matrix via integrin αvβ3 allows tumor survival and growth. This integrin was demonstrated to be the membrane receptor for thyroid hormones (THs) in several tissues. We found that THs, acting as soluble integrin αvβ3 ligands, activated growth-related signaling pathways in T-cell lymphomas (TCLs). Specifically, TH-activated αvβ3 integrin signaling promoted TCL proliferation and angiogenesis, in part, via the upregulation of vascular endothelial growth factor (VEGF). Consequently, genetic or pharmacologic inhibition of integrin αvβ3 decreased VEGF production and induced TCL cell death in vitro and in human xenograft models. In sum, we show that integrin αvβ3 transduces prosurvival signals into TCL nuclei, suggesting a novel mechanism for the endocrine modulation of TCL pathophysiology. Targeting this mechanism could constitute an effective and potentially low-toxicity chemotherapy-free treatment of TCL patients.

Figure 1

THs mediate proliferative effects in TCL cell lines. (A) mRNA levels of ITGAV, ITGB3, and THRA in a panel of TCL cell lines in comparison with mRNA levels in normal T cells. Results shown are the mean ± SEM of n = 3 independent experiments. (B) Protein levels of integrin αvβ3 and TRαβ obtained by flow cytometry in CUTLL1, SUDHL1, OCI-Ly12, and HuT 78 cell lines. Results are representative of n = 3 experiments. (C) CUTLL1 cells treated for 24 hours with free or agarose-bound T3 (1 nM) and T4 (100 nM) or a combination of the same concentrations of both free (TH) or AG-coupled (TH-AG) hormones were evaluated by Cell Titer Blue assay. Results shown represent the percent of proliferative cells respective to CT. (D) Cell proliferation after 24-hour treatment with TH or TH-AG compared with CT (dashed line) in the complete TCL cell line panel. (E) Evaluation of DNA synthesis by [³H]TdR incorporation of CUTLL1, HuT 78, and OCI-Ly12 cells after 24 hours of hormone treatment vs CT. (F) Cyclin mRNA expression levels in CUTLL1 cells by quantitative reverse-transcription polymerase chain reaction (qRT-PCR) after 2-, 6-, 12-, and 18-hour treatment with TH-AG (right) and free TH (left) compared with CT. (G) Cyclin D1 mRNA expression levels in HuT 78 and OCI-Ly12 cells by qRT-PCR after a 2-, 6-, and 12-hour treatment with TH-AG and free TH compared with CT. (H) PCNA protein levels by flow cytometry in CUTLL1 cells (right) and mRNA (left) levels in CUTLL1, HuT 78, and OCI-Ly12 cells after 24-hour treatment with free TH and AG-bound TH compared with CT. qRT-PCR results shown are the mean ± SEM of at least n = 3 independent experiments. For PCNA protein levels, a representative of 4 experiments is displayed. CT, control.

Figure 2

Integrin αvβ3 is the membrane receptor for THs in human TCL. (A) Blockade of the TH-mediated proliferative effect by 1 nM RGD peptide, added 10 minutes before 24-hour hormone treatment, was analyzed by Cell Titer Blue assay. (B) CUTLL1 cells were transfected by electroporation with siRNA against ITGAV (si-ITGAV), ITGB3 (si-ITGB3), or noncoding siRNA as control (si-CT). mRNA levels of target genes were analyzed by qRT-PCR 24 and 48 hours posttransfection. (C) Effect of ITGAV and ITGB3 knockdown on CUTLL1 cell growth by Trypan Blue staining. (D-F) Knockdown of ITGAV and/or ITGB3 abrogates the TH proliferative effect as measured by [³H]TdR incorporation (D) and cell proliferation by Cell Titer Blue assay (E) after 24 hours of TH treatment in CUTLL1 cells. Similar results were found in mature TCL, HuT 78, and OCI.Ly12 cells (F). (G) Effect of 250 ng/mL vitronectin ligand and/ or THs on cell proliferation in CUTLL1, HuT 78, and OCI-Ly12 cells measured by [³H]TdR incorporation after 24-hour treatment. (H) Growth of OCI-Ly12 cells cultured in an artificial ECM system that offers an RGD ligand, in the presence or absence of physiological levels of TH and in the presence or absence of the inhibitor of integrin αvβ3 cilengitide. Cell proliferation and clustering was measured by microscopical examination. Representative photographs (scale bar, 100 μm) of 4 replicate experiments are shown.

Figure 3

THs acting at the membrane receptor initiate a transcriptional program. (A) A graphical heatmap representation of changes (log₂ fold change) in gene expression of CUTLL1 cells after treatment with TH-AG vs CT (AG alone); n = 3 independent samples for each condition. Selected Gene Ontology categories for upregulated genes are shown on the right. (B) The top network identified by Ingenuity Pathway Analysis on mobilized genes by AG-bound hormones treated cells vs CT. (C) Validation of a select number of genes from the network shown in panel B, 24 hours posttreatment with TH-AG in immature CUTLL1 cells and mature TCL cell lines, HuT 78, and OCI-Ly12. (D) Regulation of IL-4 and VEGFB mRNA levels after 2 hours of treatment with free TH in the presence and absence of integrin αvβ3. (E) NF-κB localization by confocal microscopy of CUTLL1 cells treated or not (Control) with THs for 15 minutes. Representative photographs (scale bar, 20 μm). (F) Effect of 1 µM of NFκB inhibitor, BAY 11-7082, on TH-mediated regulation of IL-4 and VEGFB genes in CUTLL1, HuT 78, and OCI-Ly12 cells.

Figure 4

THs contribute to TCL malignant phenotype through angiogenesis induction. (A) Pearson correlation between ITGAV and ITGB3 with VEGFB and VEGFA levels in TCL patients. (B) mRNA levels of VEGFA and VEGFB in the panel of immature and mature TCLs after 24-hour treatment with TH-AG compared with CT. (C) Transcript abundance of VEGFA (right) and VEGFB (left) was evaluated in CUTLL1, HuT78, and OCI-Ly12 cell lines transfected for 48 hours with siRNA against ITGAV (si-ITGAV), ITGB3 (si-ITGB3), or noncoding sequence (si-CT) and then treated for 24 hours with TH-AG or agarose alone (as CT). (D) Transcript abundance of VEGFB (right) and VEGFA (left) was evaluated in CUTLL1, HuT78, and OCI-Ly12 cell lines after 24-hour treatment with 250 ng/mL vitronectin (VN). (E) Induction of HMEC1 cell migration by conditioned medium from CUTLL1 cells treated or not with TH-AG. Representative photographs (scale bar, 60 μm) of the endothelial cells that migrate through the chamber membrane in the presence of CT or conditioned medium. (F) Quantitation of HMEC1 cell migration by conditioned medium from CUTLL, HuT 78, and OCI-Ly12 cells treated or not with TH-AG and preincubated with the anti-VEGF bevacizumab (10 μg/mL) vs vehicle. (G) Cell migration quantitation of HMEC1 cells in the presence or absence of conditioned medium from si-RNA–transfected CUTLL1 cells treated with TH-AG vs CT for 24 hours. Mean ± SEM of at least 3 independent experiments are shown.

Figure 5

Autocrine effect of TH-induced VEGF in TCL proliferation. (A) TCL proliferation after 24-hour treatment with recombinant VEGF vs CT. (B) DNA synthesis measured by [³H]TdR incorporation in CUTLL1, HuT 78, and OCI-Ly12 cells exposed to TH and treated with the anti-VEGF bevacizumab or vehicle and the VEGFR inhibitor axitinib or vehicle. Results shown are the mean ± SEM of independent triplicates. (C) CUTLL1 cells were transfected with si-ITGAV, si-ITGB3, or si-CT and injected into SCID mice (n = 4 for each treatment). Tumor growth from day 4 to 15 after implantation was measured by the AUC. (D) Levels of VEGFA and active caspase-3 by immunohistochemistry staining in tissue sections of the tumors obtained from panel A. Representative photographs are shown (scale bar, 50 μm). (E) Representative photographs of CD31 staining of tissue sections from the CUTLL1 tumors. Quantification of the blood vessel area (lumen) from the CUTLL1 tumors (right).

Figure 6

Pharmacologic inhibition of integrin shows anti-lymphoma effect in PTCL-NOS and ALCL mouse models. (A) Tumor volume (measured by AUC, from day 1 to day 20) in OCI-Ly12 xenografted mice treated with vehicle vs cilengitide. (B) DNA-binding capacity of NF-κB family members assessed by an enzyme-linked immunosorbent assay–based assay in extracted nuclear fractions of disaggregated lymphoma tissues from the tumors obtained in panel A. (C) Microvascular lumen (by intravascular isolectin) and apoptotic bodies (by terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling assay) in representative tissue sections of the tumors from vehicle- and cilengitide-treated OCI-Ly12 mice. Scale bar, 25 μm (upper) and 100 μm (lower). The white lines show representative maximum lumen size measurements. (D-E) AUC of tumor growth of ALCL ALK-negative (D) and ALCL ALK-positive (E) patient-derived lymphomas in NSG mice treated with vehicle vs cilengitide. (F) Cartoon representation of TH regulation of TCL prosurvival pathways via the activation of αvβ3 integrins.