Fig 1.
Proteomic analysis of receptor tyrosine kinases in mouse KSHV-induced KS-like tumors shows activation of PDGF receptor-alpha.
(A) Mouse Phospho-Receptor Tyrosine Kinase (RTK) Array Kit used to quantify levels of phosphorylation of 39 RTKs in mECK36 tumors pointing to major activation spot corresponding to PDGF receptor alpha chain. (B) Bar graph and pie chart from densitometry for the higher-exposure blot are equally color coded for the most prominent signals. (C) PDGFRA and phospho-PDGFRA (left panel) or c-KIT and phospho-cKIT (right panel) determined in 3 different samples of Mouse Normal Skin and mECK36 tumors from 3 different mice by immunoblotting.
Fig 2.
Phosphorylation of PDGFRA in mECK36 and KS tumors is associated with the presence of its PDGF ligands and localize to areas of KSHV infection.
(A) PDGFA, PDGFB and GAPDH determined in 3 different samples of Normal Skin and mECK36 tumors from 3 different mice by immunoblotting. (B) Immunohistochemical staining of mouse KS-like mECK36 tumors for PDGFA, PDGFB, LANA, and phospho-PDGFRA. Representative image of the FFPEs obtained from the mECK36 tumorigenesis experiment N = 5.(C) Immunohistochemical staining of human KS biopsies for PDGFA, PDGFB, LANA, and phospho-PDGFRA. (D) IFA of mECK36 tumors for KSHV LANA (red) and phospho-PDGFRA (green). Cell nuclei were counterstained with DAPI (blue). White arrows indicate cells that co-stain for LANA and p-PDGFRA.
Fig 3.
KSHV-mediated PDGF upregulation and PDGFRA activation in mECK36 cells and tumors.
(A) Fold-changes in KSHV gene expression and PDGFRA ligands between mECK36 cells and mECK36 tumors determined by RT-qPCR in triplicate and are presented as means ± SD. *P < 0.05. (B) Total and phospho-PDGFRA together with its ligands PDGFA and PDGFB determined by immunoblotting in mECK36 cells (duplicate) and three mECK36 tumors from 3 different mice.(C) Total and phospho-STAT3 together with Total and phospho-AKT were determined by immunoblotting in mECK36 cells (duplicate) and three mECK36 tumors from 3 different mice.(D) Fold-changes in PDGFRA ligands and KSHV gene expression between doxyxcyclin induced and un-induced mECK36 cells stably transfected with a Tet-inducible RTA were measured by RT-qPCR after 24 hours of induction. Data were from three independent experiments carried out in triplicate and are presented as means ± SD. *P < 0.05.
Fig 4.
KSHV vGPCR can activate PDGFRA by upregulation of its ligands PDGFA/B in mECK36 cells.
(A) mRNA levels determined by RT-qPCR of angiogenesis factors, cytokines and REDOX genes in Tetracycline-inducible vGPCR (TET-vGPCR) and control mECK36 stimulated with doxycycline for 24 hours. Data were from three independent experiments carried out in triplicate and are presented as means ± SD. *P < 0.05. (B) and (C) PDGFB-Luc promoter activity in 293T cells co-transfected with empty vector (control), vGPCR, and either a constitutively activated Rac1 (RacQL) construct (B) or a dominant negative Rac1 (RacN17) (C). Data were from three independent experiments carried out in triplicate and are presented as means ± SD. *P < 0.05. (D) Western blot analysis for PDGFB of TET-vGPCR mECK36 cells induced with doxycycline for 24 hrs in the presence of Rac1 inhibitor EHT1864, the NOX inhibitor DPI, or the ROS scavenger NAC. (E) Phosphorylated and total PDGFRA levels of mECK36 cells stimulated with conditioned media from TET-vGPCR mECK36 cells induced with DOX in the presence or the absence of Rac1, NOX and ROS inhibitors. (F) PDGFB mRNA and vGPCR levels of cells in (D) determined by RT-qPCR. Data were from three independent experiments carried out in triplicate and are presented as means ± SD. *P < 0.05.
Fig 5.
PDGFR signaling drives proliferation and angiogenesis in mECK36 cells and tumors via a Rac1-ROS-NOX dependent pathway.
(A) Rac1 activation levels determined by a GTP-bound Rac1 pull-down assay/ Rac1 immunoblotting of mECK36 cells treated with PDGF-BB in the presence or in the absence of Rac1 inhibitor EHT1864. (B) ROS production (superoxide) of mECK36 cells stimulated with PDGFBB. Rac1 inhibitor (EHT1864) or NOX inhibitor (DPI) were added before PDGF stimulation (C) mRNA expression levels of c-Myc, VEGFA and KSHV LANA determined by RT-qPCR of mECK36 cells stimulated with PDGF-BB in the presence of absence of Rac1 inhibitor EHT1864, ROS scavenger NAC, or NOX inhibitor DPI. Data were from three independent experiments carried out in triplicate and are presented as means ± SD. *P < 0.05. (D) mRNA expression levels of KSHV v-Cyclin, v-FLIP, v-IL6, vGPCR and ORF8 determined by RT-qPCR of mECK36 cells stimulated with PDGF-BB. Data were from three independent experiments carried out in triplicate and are presented as means ± SD. *P < 0.05. (E) and (F) Proliferation (E) or VEGF secretion measured by ELISA (F) of mECK36 cells stimulated with PDGF-BB in the presence or absence of NAC, DPI or EHT1864. Data were from three independent experiments carried out in triplicate and are presented as means ± SD. *P < 0.05. (G) Phosphorylated and total PDGFR levels in NAC-treated and control mECK36 tumors were determined by immunoblotting. (H) mRNA levels of PDGFs and PDGFRs in NAC treated and control mECK36 tumors were determined by RT-qPCR. Data were from three tumors per treatment and are presented as means ± SD. *P < 0.05.
Fig 6.
Multi tyrosine kinase inhibitors that can target PDGFRA block KSHV-mediated tumorigenesis.
(A) Tumor growth curve from mice with established subcutaneous mECK36 tumors treated with vehicle (PBS) or Imatinib (150 mg/Kg twice daily) by oral administration. Data indicate mean tumor size ± SD (n = 10). (B) mRNA levels of VEGFA in Imatinib treated and control mECK36 tumors determined by RT-qPCR. (C) Tumor growth curve from mice with established subcutaneous mECK36 tumors treated with vehicle (PBS) or Sunitinib (80 mg/Kg/day) by oral administration. Data indicate mean tumor size ± SD (n = 10). (D) Phosphorylated and total PDGFR levels from Imatinib-treated, Sunitinib-treated and control mECK36 tumors determined by immunoblotting. (E) IFA of frozen sections of Imatinib-treated, Sunitinib-treated and control tumors stained with pan-endothelial marker CD31 (red) or lymphatic microvessel marker VEGFR3 (red). Nuclei were counterstained with DAPI (blue). Bar graphs show the relative intensity signals of CD31 and VEGFR3 from 10 different fields. *P < 0.05.
Fig 7.
Maintenance of tumorigenesis in KSHV-negative mECK36 tumors through PDGFRA activating mutations.
(A) Phosphorylated PDGFRA, total PDGFRA, PDGFA and PDGFB levels from KSHV+ve mECK36 and KSHV-ve mECK36 tumors determined by immunoblotting. (B) mRNA levels of PDGFs and PDGFRs in KSHV+ve mECK36 and KSHV-ve mECK36 tumors determined by RT-qPCR. Data are from three tumors carried out in triplicate and are presented as means ± SD. *P < 0.05. (C) ELISA of Platelet-Derived Growth Factor AA (PDGF-AA) and Platelet-derived growth factor subunit BB (PDGF-BB) in KSHV+ve mECK36 and KSHV-ve mECK36 tumor tissues. Data are from three tumors and are presented as means ± SD. *P < 0.05. (D) Immunohistochemistry staining of KSHV+ve mECK36 and KSHV-ve mECK36 tumor tissues using antibodies against PDGFA, PDGFB, LANA, and phospho-PDGFRA. (E) Mouse Growth Factor Antibody Array used to detect 30 Mouse Growth Factors in KSHV+ve and KSHV-ve tumors. Data is presented as fold change expression between KSHV+ve mECK36 and KSHV-ve mECK36 tumor tissue. (F) Sequence alignment of the hot spot region for oncogenic mutations in the activation domain of TKs. The D842V mutation in PDGFRA was only found in the cDNA of tumorigenic KSHV-negative mECK36 cells. (G) Tumor growth curve from mice with established subcutaneous KSHV-negative mECK36 tumors treated with Imatinib (150 mg/Kg twice daily) or Sunitinib (80 mg/Kg/day) by oral administration. Data indicate mean tumor size ± SD (n = 10).
Fig 8.
Tyrosine-kinase truncated dominant-negative mutants of PDGFRA block KSHV tumorigenesis in mice.
(A) Structure of the truncated dominant-negative forms of PDGFRA and PDGFRB. (B) Phosphorylated PDGFR and total PDGFR determined by immunoblotting in mECK36 cells transfected with dominant-negative forms of PDGFRA, PDGFRB, or empty vector control stimulated with either PDGF-AA (80 ng/mL) or PDGF-BB (20ng/ml) for 10 min. (C) Tumor growth curve from mice following subcutaneous injection of mECK36 cells transfected with dominant-negative forms of PDGFRA, PDGFRB, or empty vector control. Data indicate mean tumor size ± SD (n = 10).
Fig 9.
PDGDRA phosphorylation is consistently found in spindle-cells of AIDS-Kaposi’s sarcoma lesions and localizes to areas of KSHV infection.
(A) Staining AIDS-KS biopsies from a ACSR tissue microarray (TMA) showing that phospho-PDGFRA localizes to areas of LANA staining in two characteristic samples of the 59 out of 66 skin KS tumors which were strongly phospho-PDGFRA+ve/ LANA+ve. (B) Example of one of the 7 phospho-PDGFRA-ve (LANA+ve) AIDS-KS tumors of the TMA. (C) Normal control tissue from the TMA (skin). (D) Example of a KS tumor with strong phospho-PDGFRA staining and a low percentage of LANA+ve cells. (E) LANA and phospho-PDGFRA were scored from 0 to 3 depending on the signal strength of the antibody staining (bottom table) and total number of PDGFRA+ve/LANA+ve (59) and PDGFRA-ve/LANA+ve (7) biopsies are shown over the 66 skin AIDS-KS biopsies analyzed (top table).
Fig 10.
KSHV-dependent and independent mechanisms of PDGFRA-driven sarcomagenesis.
The upper panel describes one possible PDGFRA activation mechanism supported by data from this paper: lytically or abortive-lytically infected KS cells expressing vGPCR can drive sarcomagenesis by Rac1-NOX-ROS mediated upregulation of PDGF leading to PDGFRA activation in latently infected cells that promotes proliferation and VEGF angiogenesis (paracrine oncogenesis). This mechanism would be sensitive to Imatinib, which inhibits PDGFRA oncogenic signaling. The lower panels show in the left panel; the scenario for non-viral sarcomas driven by ligand-mediated activation of PDGFRA (soft tissue sarcomas) which are sensitive to Imatinib and to anti-PDGFRA antibody therapy [56]. The right panel shows the scenario for our KSHV-ve KS-like sarcomas driven by the PDGFRA D842V mutations—the most frequent PDGFRA mutation in GIST—which make these tumors Imatinib-resistant.