The structure and function of tumor blood vessels profoundly affects the tumor microenvironment. Signals mediated through the lysophosphatidic acid receptor 4 (LPA4) promote vascular network formation to restore normal vascular barrier function in subcutaneous tumors and thus improve drug delivery. However, the characteristics of the vasculature vary by organ and tumor types, and how drug delivery and leukocyte trafficking are affected by modification of vascular function by LPA in different cancers is unclear. Here, we show that LPA4 activation promotes the formation of fine vascular structures in brain tumors. RhoA/ROCK signaling contributed to LPA-induced endothelial cell–cell adhesion, and RhoA/ROCK activity following LPA4 stimulation regulated expression of VCAM-1. This resulted in increased lymphocyte infiltration into the tumor. LPA improved delivery of exogenous IgG into brain tumors and enhanced the anticancer effect of anti–programmed cell death-1 antibody therapy. These results indicate the effects of LPA on vascular structure and function apply not only to chemotherapy but also to immunotherapy.

Significance:

These findings demonstrate that lysophosphatidic acid, a lipid mediator, promotes development of a fine capillary network in brain tumors by inducing tightening of endothelial cell-to-cell adhesion, facilitating improved drug delivery, and lymphocyte penetration.

Glioblastoma multiforme (GBM) is the most malignant of all brain tumors. Standard treatment is maximal surgical resection followed by radiotherapy and concomitant and adjuvant chemotherapy with temozolomide (TMZ; ref. 1). However, median overall survival with this standard care is less than 15 months. Hence, a new strategy for treating GBM is urgently required.

Antiangiogenic therapy was proposed as a suitable treatment strategy because highly proliferating microvasculature is a key pathologic feature of GBM due to tumor cell secretion of large amounts of VEGF. This factor is predominantly involved in angiogenesis, supporting tumor growth and invasion (2). Bevacizumab, a neutralizing antibody against VEGF, is an antiangiogenic drug approved for malignant gliomas. High-dose bevacizumab reduces the degree of vasculature and mediates antitumor effects in a mouse glioma model (3). However, in the clinical setting, the overall survival of newly-diagnosed GBM patients was not improved by the addition of bevacizumab to chemoradiotherapy with TMZ (4, 5).

The formation of structurally abnormal blood vessels in tumors is a major clinical problem, due to their nonfunctionality which results in decreased drug delivery and decreased immune cell infiltration (6). Hence, tumor vascular normalization is now a widely accepted concept that would not rely on inhibiting angiogenesis or pruning immature vessels in tumors but rather would promote the formation of a functional vascular network to improve drug delivery, reverse intratumoral hypoxia, and encourage immune cell infiltration (7). Bevacizumab was the first drug for controlling tumor angiogenesis utilized in the clinic. However, high-dose bevacizumab does not alter the intratumoral environment appropriately because it prunes immature and microvessels, only partially improving vascular structure (3, 8). Combinations of appropriate doses of antiangiogenic agents together with anti–PD-L1 therapy show synergistic effects in brain tumor models (9). In contrast to antiangiogenics, as outlined above, the promotion of tumor vascularization is a new approach based on enhancement of the vascular network to improve drug delivery and immune cell infiltration (10, 11).

Lysophospholipids such as lysophosphatidic acid (LPA) and sphingosine-1-phosphate (S1P) are newly recognized vessel-regulatory factors which have also been documented as active in vascular formation and integrity maintenance (12, 13). Here, we focused on LPA which is mainly synthesized from lysophosphatidylcholine by the enzyme autotaxin (ATX; ref. 14). Angiogenesis does not develop normally in ATX-deficient mice, resulting in embryonic lethality (15). In a previous study, we reported that LPA induces fine vascular network growth through LPA receptor 4 (LPA4) signaling and improves drug delivery in a subcutaneous tumor model in mice. LPA4 activation induces cortical actin formation under the endothelial cell (EC) membrane and stabilizes the localization of VE-cadherin at the cell membrane, which enhances endothelial intercellular junction integrity (16). Here, we have continued to analyze the function of LPA4 for vascular formation in an orthotopic tumor implantation model for brain tumors. The characteristics of the vasculature are different in different organs, and vascular formation in tumors also varies depending on the microenvironment. In the present study, we investigated whether LPA promotes vascular network formation in brain tumors using GL261, a murine glioma cell line, as well as primary tumor cells from GBM patients. For comparison, we also examined a Lewis lung carcinoma (LLC) model. In addition, we investigated whether vascular modification by LPA in the brain tumor model alters immune cell infiltration and drug delivery into tumors, as well as the effects of immune-checkpoint inhibitor treatment.

Reagents

1-oleoyl-LPA and VPC31144(S) were purchased from Avanti Polar Lipids. LPA and VPC31144(S) were dissolved at 10 mmol/L in 50% ethanol. On the day of administration, LPA was diluted with PBS to the appropriate concentration. Y27632 (WAKO) was dissolved in water at 10 mmol/L. PD98059 (WAKO), U0126 (WAKO), SP600125 (WAKO), and SQ22536 (WAKO) were dissolved in DMSO at 50 mmol/L and SB202190 (WAKO) at 10 mmol/L. VX702 (Selleck) was also dissolved in DMSO at 1 mmol/L. LY294002 (WAKO) was dissolved in ethanol at 50 mmol/L and wortmannin (WAKO) at 1 mmol/L. 2′, 5′-Dideoxyadenosine (2′5′DDA; WAKO) was dissolved in water at 100 μmol/L. NF023 (Abcam) was dissolved in water at 50 mmol/L. All of the above stocks were stored at −20°C. Pyrrolidine dithiocarbamate (PDTC; Sigma-Aldrich) was dissolved in water at 1 mol/L and YM-254890 (WAKO) in DMSO at 10 mmol/L; both were stored at 4°C.

Mice

All experimental procedures in this study were approved by the Institutional Animal Care and Use Committee of Osaka University and Osaka University Committee for Recombinant DNA Experiments. C57BL/6 and SCID mice were purchased from SLC. LPA4 KO mice on the C57BL/6 background were generated as described previously (17). Mice 8 to 12 weeks of age were used for these experiments.

Cell culture

GL261 (mouse glioblastoma) was kindly provided from Dr. Shingo Takano (Department of Neurosurgery, Faculty of Medicine, University of Tsukuba, Tsukuba, Ibaraki, Japan) without further authentication (18). LLC was purchased from RIKEN Cell Bank. Both cell lines were maintained in DMEM (Sigma-Aldrich) supplemented with 10% FCS (Sigma-Aldrich) and 1% penicillin/streptomycin (P/S; Life Technologies). GDC40 was isolated from human GBM tissue as previously described (19) and cultured using the neurosphere culture technique in DMEM/F-12 (1:1; Sigma-Aldrich) with epidermal growth factor (20 ng/mL; PeproTech Inc.), fibroblast growth factor 2 (20 ng/mL; PeproTech), leukemia inhibitory factor (10 ng/mL; Millipore), B27 supplement (final 2%; Thermo Fisher Scientific), and heparin (5 mg/mL; Sigma-Aldrich; ref. 20). MS-1 (mouse pancreatic islet EC line; ATCC) was maintained in DMEM supplemented with 5% FCS and 1% P/S. OP9 (mouse fibroblast; RIKEN) was maintained in MEM (Sigma-Aldrich) supplemented with 10% FCS and 1% P/S. LLC and GL261 cells were stably transfected with the expression vector pEGFP N-1 (Clontech) using Lipofectamine 2000 (Invitrogen). The transfected cells were cultured in selection medium containing 200 μg/mL G418 (Geneticin; Gibco). All cell lines passaged for less than 3 months after thawing. Because of health appearance, no Mycoplasma test was done.

Ethics statement

For using human GBM tumor samples and human GBM–derived xenografts in the GDC40 model, written informed consent was obtained from patients, the studies were conducted in accordance with recognized ethical guidelines (Declaration of Helsinki), and the study was approved by Osaka University Institutional Review Board.

Tumor model

The orthotopic brain tumor model using GL261 was established as previously described (21). Briefly, 8- to 10-week-old female C57BL/6 mice were anesthetized with medetomidine, midazolam, and butorphanol, and 2 × 105 cells in 1 μL PBS were stereotactically injected into the right basal ganglia using a Hamilton syringe. The transplantation locus was determined as 2 mm to the right of and 2 mm caudal from the bregma, and 2 mm deep into the brain. For quantitative evaluation of vascular density, length, perfusion, hypoxia, and drug delivery, LPA (1 mg/kg) was intravenously injected into mice bearing GL261 brain tumors on days 7, 9, 11, and 14 after inoculation, and tumors were harvested on day 15. The LLC brain metastasis model and GDC40 xenograft model were established in the same way as GL261. Establishment of subcutaneous xenografts of LLC and measurement of tumor size were determined as previously reported (16). Briefly, 7 days after inoculation, mice were continuously treated with i.p. administration of vehicle or LPA (3 mg/kg) for 15 days. For quantitative measurement of vascular density and vessel length, we used a semiautomated computational tool (AngioTool; available in the public domain at http://angiotool.nci.nih.gov, Image J: available at https://imagej.nih.gov/ij/).

In vivo survival analysis

TMZ was purchased from MSD. Weekly TMZ (20 mg/kg) was administered i.p. from day 11. Anti–programmed cell death (PD)-1 antibody was purchased from BioXcell (clone RMP1-14). For LLC subcutaneous tumor, anti–PD-1 antibody (100 μg/mouse) was injected i.p. from day 7 three times a week for 2 weeks. For the GL261 brain tumor model, anti–PD-1 antibody (200 μg/mouse) was injected from day 7. Injections were given 3 times a week in the first week and twice a week in the following weeks. Equivalent amounts of isotype control IgG (clone 2A3, BioXcell) were injected into control mice. Mice were observed daily and euthanized when they showed 20% weight loss or neurological deficits. Fifty days after tumor inoculation, brains of surviving mice were harvested and examined by hematoxylin and eosin staining for remnant tumor.

Immunofluorescence staining

The procedure for tissue preparation and staining was as previously reported (22). Briefly, fixed specimens were embedded in optimal cutting temperature compound (Sakura Finetek) and sectioned at 20 or 40 μm. After postfixation with 4% paraformaldehyde (PFA), sections were blocked. For immunohistochemistry, anti-mouse CD31 mAb (Millipore), anti-human CD31 (eBioscience), anti-P2Y9 (LPA4) polyclonal antibody (Bioss), anti-STEM121 mAb (Takara Bio, Inc.), anti-mouse VE-cadherin mAb (BD Biosciences), anti-mouse claudin 5 mAb (Thermo Fisher Scientific), anti-mouse occludin mAb (BD Biosciences), anti-human ZO1 mAb (BD Biosciences), PE-conjugated anti-mouse CD4 (eBioscience), anti-mouse CD8a mAb (BioLegend), anti-mouse CD106 (VCAM-1) mAb (BioLegend), anti-mouse CD54 (ICAM-1) mAb, Alexa Fluor 488–conjugated human/mouse MECA-79 mAb (eBioscience), anti–NF-κB mAb (Cell Signaling Technology Inc.), and anti-mouse PD-L1 (BioLegend) were used as primary antibodies. As secondary antibody, anti-hamster Alexa Fluor 488 (Jackson ImmunoResearch Laboratories), anti-hamster Cy3 (Jackson ImmunoResearch Laboratories), anti-hamster Alexa Fluor 647 (Jackson ImmunoResearch Laboratories), anti-rabbit Alexa Fluor 488 (Jackson ImmunoResearch Laboratories), anti-rat Alexa Fluor 488 (Jackson ImmunoResearch Laboratories), or anti-rat Alexa Fluor 546 (Jackson ImmunoResearch Laboratories) were used as appropriate for the primary antibodies. Biotinylated antibodies were developed using ABC kits (Vector Laboratories). Nuclear staining was performed using Hoechst 33342 (Sigma-Aldrich) or TOPRO3 (Invitrogen).

Assessment of blood vessel function

Tumor vascular perfusion was evaluated as described previously (16). In short, fluorescein Lycopersicon Esculentum (Tomato) lectin (0.05 mg/mouse; VECTOR) was injected intravenously into mice harboring GL261 brain tumors. Tumors were dissected 10 minutes after lectin injection. Blood flow was interpreted by the lectin positivity in ECs. To measure hypoxia in tumor tissues, Hypoxyprobe-1 (60 mg/kg, i.p.; Hypoxyprobe) was injected 2 hours before tissues were harvested. Tumor sections were stained using the anti-Hypoxyprobe antibody, following the manufacturer's instructions. For evaluation of drug delivery, 1% Evans Blue (EB; Sigma-Aldrich) in PBS was injected i.p. into mice bearing GL261 brain tumors receiving 150 mg/kg. Thirty minutes after injection, mice were perfused with 4% PFA in PBS, and tumors were resected. For evaluation of vascular permeability, three types of FITC-conjugated dextran (MW 3–5 kDa, 70 kDa, and 2 mDa, 125 mg/kg, Sigma-Aldrich) were injected intravenously into mice bearing GL261 brain tumors. One hour after injection, tumors were resected as described above. To evaluate retention of macromolecules in brain tumors, 70 kDa FITC dextran (125 mg/kg) was injected intravenously. Twenty-four hours after injection, tumors were resected after intracardiac perfusion of 4% PFA.

Cell preparation

Tissue dissection procedures and preparation of single-cell suspensions were as previously reported (23). Briefly, GL261 brain tumor–bearing mice were perfused with PBS containing 4% FCS, and tumors were then dissected. Single-cell suspensions were stained with FITC-conjugated anti-CD45 mAb and APC-conjugated anti-CD31 mAb (BD Biosciences). CD31+CD45ECs were analyzed and sorted using a SORP FACSAria (BD Biosciences). For cell staining of lymph nodes, anti-mouse CD16/CD32 (Mouse BD Fc Block), BV421-conjugated anti-mouse CD3e mAb (BD Biosciences), APC-conjugated anti-mouse CD8a mAb (BD Biosciences), PE-conjugated mouse CD4 mAb (eBioscience), PE-conjugated anti-mouse CD11b mAb (eBioscience), BV421-conjugated anti-mouse/human B220 mAb (Biolegend), and APC-conjugated anti-mouse CD11c mAb (Biolegend) were used. For cell staining of GL261-EGFP, APC-conjugated anti-mouse CD274(PD-L1) mAb (Biolegend) was used. Stained cells were analyzed on a BD LSRFortessa X-20 Cell analyzer (BD Biosciences).

Real-time RT-PCR analysis

RNA was extracted using RNeasy Mini Kits (Qiagen), and cDNA was generated using reverse transcriptase with the ExScript RT reagent Kit (Takara Bio). Real-time RT-PCR was performed using a Stratagene Mx3000P (Stratagene). The PCR was performed on cDNA using specific primers with the following sequences: 5′-TGG CAA AGT GGA GAT TGT TGC C-3′ and 5′-AAG ATG GTG ATG GGC TTC CCG-3′ for GAPDH, 5′-CCG CTT CCA TTT CCC TAT TT-3′ and 5′-AAA ACC GTG ATG TGC CTC TC-3′ for LPA1, 5′-CCA TCA AAG GCT GGT TCC T-3′ and 5′-TCC AAG TCA CAG AGG CAG TG-3′ for LPA2, 5′-TTC CAC TTT CCC TTC TAC TAC CTG-3′ and 5′-TCC ACA GCA ATA ACC AGC AA-3′ for LPA3, 5′-GCC CTC TCT GAT TTG CTT TT-3′ and 5′-TCC TCC TGG TCC TGA TGGTA-3′ for LPA4, 5′-AGC GAT GAA CTG TGG AAG G-3′ and 5′-GCA GGA AGA TGA TGA GAT TGG-3′ for LPA5, 5′-TGT GCC CTA CAA CAT CAA CC-3′ and 5′-TCA CTT CTT CTA ACC GAC CAG-3′ for LPA6, 5′-TAA GGC ACG GGT AGC ACT CA-3′ and 5′-AAC TCG CGG ACA ACG ATG TT-3′ for claudin 5, 5′-CTGGATCTATGTACGGCTCACA-3′ and 5′-GGGATCAACCACACAGTAGTG-3′ for occludin, 5′-GCC AGA GAA AAG TTG GCA AG-3′ and 5′-ATG GTC AGT CCCAGC ATC TC-3′ for ZO-1, 5′-CCC GTC TTT ACT CAA TCC ACA-3′ and 5′-ATC TGG GTC CAC AAC AGT CAG-3′ for VE-cadherin, 5′-TGG TCA TCG CCA CCT TAA TAG-3′ and 5′-CAG AAA TCT TCT CGC TGT TGG-3′ for PECAM, 5′-AGC CTC AAC GGT ACT TTG GAT A-3′ and 5′-TCT CCA TGA GAA CAA TGG TGA C-3′ for VCAM-1, 5′-GTG ATG GAG GGG GTC AGG A-3′ and 5′-GGG ATG GGA CAG CCT AAA CT-3′ for Ccl21, 5′-GGC CAC GGT ATT CTG GAA GC-3′ and 5′-GGG CGT AAC TTG AAT CCG ATC TA-3′ for Cxcl13, 5′-CTG GTT CTC TGG ACC TTC CC-3′ and 5′-GGT GCA CAG AGC TGA TAG CC-3′ for Ccl19, 5′-TCT TGC TGC CAC CTC TCT TG-3′ and 5′-CTC ACT GGT GTA GCT GGT GG-3′ for Glycam1. Expression levels of the target genes were normalized to the GAPDH level in each sample.

Western blotting

Confluent MS-1 cells were cultured in 6-well plates and serum-starved in DMEM supplemented with 0.5% BSA after which they were stimulated with 100 μmol/L LPA or VPC31144(S) for 24 hours. After washing with ice-cold PBS, cells were extracted in RIPA Buffer (Thermo Fisher Scientific) containing Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific). After centrifugation at 15,000 rpm for 5 minutes at 4°C, the supernatants were heated at 95°C for 5 minutes with 6 × SDS sample buffer. Analysis by SDS-PAGE was as previously described (24). Anti-phospho p38 mAb (Cell Signaling Technology), anti-p38 (Cell Signaling Technology), anti-mouse VCAM-1 mAb (Cell Signaling Technology), anti-mouse pErk1/2 (Cell Signaling Technology), anti-mouse Erk (Cell Signaling Technology), and anti-mouse GAPDH mAb (Millipore) were used as the primary antibodies.

siRNA knockdown

RNA knockdown experiments using siRNAs were conducted as previously described (16). Briefly, siRNA for LPA4, LPA6, and GNA13 and negative control siRNA were purchased from Thermo Fisher and were transfected into MS-1 cells using Lipofectamine RNAiMAX (Thermo Fisher) according to the manufacturer's instructions. Knockdown efficiency was verified using quantitative real-time PCR in the previous study (16).

RhoA activity assay

MS-1 cells were grown to confluence in 6-well plates. After serum starvation in DMEM supplemented with 0.5% BSA for 3 hours, MS-1 cells were stimulated with LPA (10 μmol/L), VPC31144(S) (10 μmol/L), or vehicle for 3 minutes. Activation of RhoA was measured by G-LISA RhoA activation assay biochem kit (Cytoskeleton, Inc.), following the manufacturer's instructions.

MS-1 stimulation and immunostaining

MS-1 cells were grown to confluence on glass-bottom plates, serum starved in DMEM supplemented with 0.5% BSA for 3 hours, and stimulated with or without LPA (100 μmol/L). The cells were then fixed with 4% PFA and permeabilized using 0.1% Triton X-100. After blocking, cells were labeled with anti-mouse VCAM-1 (BioLegend) and visualized with Alexa Fluor 488–conjugated goat anti–rat IgG followed by nuclear staining with TOPRO3.

Microscopy

For immunofluorescence staining, samples were analyzed using a Leica TCS/SP5 confocal microscope (Leica Microsystems) or Leica DM5500B microscope (Leica Microsystems). Images were processed with the Leica application suite (Leica Microsystems) and Adobe Photoshop CS6 software (Adobe Systems). All images shown are representative of more than five independent experiments.

Statistical analysis

All data are presented as mean ± SEM. Comparisons between multiple treatments were made using one-way ANOVA, followed by the Mann–Whitney U test. Pair-wise comparisons between treatments were made using the Student t test. A P value of <0.05 was considered to be significant. For survival analysis of brain tumors, the Kaplan–Meier method was used, and the group differences were assessed by the Log-rank test.

LPA promotes fine vascular formation in brain tumors

We transplanted GL261 cells stereotactically into the right basal ganglia of B16BL/6 mice. LPA was administered intravenously on days 7, 9, 11, and 14 after inoculation of cancer cells. Tumors were harvested at day 16. In the control tumor, blood vessels were not uniform in size and were scarce in the center of the tumor. Moreover, enlarged abnormal vessels were detected in the control tumor (Fig. 1A). In contrast, after LPA administration, the number of vessels increased and most of them were fine capillaries (Fig. 1B–D). Coverage by pericytes and the basal membrane was not significantly altered by LPA (Supplementary Fig. S1A–S1D).

Figure 1.

LPA stimulates LPA4 and promotes vascular network formation in brain tumors. A–D, GL261 brain tumor sections stained with anti-CD31 mAb. Representative images of mice treated with vehicle (A) or LPA (B) three times a week from day 7 to day 14. The dashed lines indicate the tumor edge. Scale bar, 500 μm. Quantification of total vessel length (C) and vascular density according to the vessel caliber (D) at the tumor center (n = 9 tumors per group). E, Quantitative real-time PCR analysis for LPA receptors 1–6 in ECs from GL261 brain tumors (n = 4 tumors per group). F and G, GL261 brain tumor–bearing wild-type (top) and LPA4KO mice (bottom) were treated with LPA or VPC31144(S) (VPC). F, Representative images of sections stained with anti-CD31 mAb (green) and TOPRO3 (blue) are shown. G, Quantification of vascular density (n = 5 tumors per group). H and I, SCID mice inoculated with human GBM cells intracranially were treated with weekly injection of vehicle, LPA, or VPC31144(S). Tumors were harvested 12 weeks after inoculation. H, Representative image of sections stained with anti-CD31 mAb (green) and anti-STEM121 mAb (red). The dashed lines indicate the tumor edge. I, Quantification of vascular density (n = 5 per group). All experiments were repeated at least twice. Error bars, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S1.

Figure 1.

LPA stimulates LPA4 and promotes vascular network formation in brain tumors. A–D, GL261 brain tumor sections stained with anti-CD31 mAb. Representative images of mice treated with vehicle (A) or LPA (B) three times a week from day 7 to day 14. The dashed lines indicate the tumor edge. Scale bar, 500 μm. Quantification of total vessel length (C) and vascular density according to the vessel caliber (D) at the tumor center (n = 9 tumors per group). E, Quantitative real-time PCR analysis for LPA receptors 1–6 in ECs from GL261 brain tumors (n = 4 tumors per group). F and G, GL261 brain tumor–bearing wild-type (top) and LPA4KO mice (bottom) were treated with LPA or VPC31144(S) (VPC). F, Representative images of sections stained with anti-CD31 mAb (green) and TOPRO3 (blue) are shown. G, Quantification of vascular density (n = 5 tumors per group). H and I, SCID mice inoculated with human GBM cells intracranially were treated with weekly injection of vehicle, LPA, or VPC31144(S). Tumors were harvested 12 weeks after inoculation. H, Representative image of sections stained with anti-CD31 mAb (green) and anti-STEM121 mAb (red). The dashed lines indicate the tumor edge. I, Quantification of vascular density (n = 5 per group). All experiments were repeated at least twice. Error bars, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S1.

Close modal

It has been reported that LPA has 6 specific receptors (LPA1-LPA6), and LPA4 ligation showed vascular-promoting effects in a subcutaneous tumor model (16, 25). Here, we investigated whether LPA4 is also involved in vascular-promoting effects in the GL261 brain tumor model. Two weeks after inoculation of GL261 cells, ECs (identified as CD45CD31+ cells) were isolated from tumor tissues by FACS. Quantitative RT-PCR analysis suggested that LPA4 and LPA6 are expressed on ECs in GL261 brain tumors (Fig. 1E). Next, we stained GL261 brain tumor sections with anti-LPA4 antibody and confirmed that CD31+ECs express LPA4 at the protein level (Supplementary Fig. S1E). We confirmed the specificity of the anti-LPA4 antibody by lack of LPA4 in ECs of GL261 tumors generated in LPA4 KO mice. We also stained human GBM tissues and found that ECs therein did express LPA4 (Supplementary Fig. S1F). In order to assess whether LPA4 is involved in the vascular-promoting effect mediated by LPA in brain tumors, we inoculated GL261 cells into the brains of LPA4 KO mice. We then injected LPA intravenously into these LPA4 KO mice bearing GL261 brain tumors in the same way as in wild-type mice. As expected, no promotion of vascularization was seen in the LPA4 KO mice. VPC31144(S), an analog of LPA which preferentially binds LPA4 rather than LPA6, had the same effect on the tumor vasculature. Consistent with this, the vascular-promoting effect of VPC31144(S) was absent in LPA4 KO mice (Fig. 1F and G).

Next, we established a brain metastasis model by injecting LLC tumor cells into C57BL/6 mice and subsequently treated them with LPA or VPC31144(S). LPA4 expression was confirmed by immunostaining in this model (Supplementary Fig. S1G). Results indicate that vascular density increased in LPA- or VPC31144(S)-treated tumors generated in wild-type mice but not LPA4 KO mice (Supplementary Fig. S1H and S1I).

Finally, we tested the effect of LPA in a xenograft model using primary human GMB cells (GDC40; ref. 18). We transplanted 2 × 105 GDC40 cells into brains of NOD/SCID mice. LPA or VPC31144(S) was intravenously injected weekly from day 7. Brain tumors were harvested 12 weeks after inoculation. Tumor sections were stained with anti-mouse CD31 and anti-STEM121 antibody reacting specifically with a cytoplasmic protein of human cells. Results showed that vascular density was increased in both LPA- and VPC31144(S)-treated tumors (Fig. 1H and I). Taking these results together, we conclude that LPA4 activation by LPA promotes fine vascular network formation in brain tumors.

LPA induces functional vasculature in brain tumors

Relative to normal blood vessels, the tumor vascular is structurally and functionally abnormal. Abnormal blood vessel formation in tumors is a clinical problem, because it decreases drug delivery and immune cell infiltration into the tumor (7). We assessed vascular function in tumors treated with LPA in the brain tumor models. First, we injected FITC-conjugated lectin intravenously into mice bearing GL261 brain tumors and evaluated blood perfusion into the tumors. Blood perfusion was improved when the animals were treated with LPA (Fig. 2A and B). Next, we analyzed intratumoral hypoxia and found that LPA treatment reduced the hypoxic area (Fig. 2C and D), suggesting that functional blood vessels are induced by LPA. Drug delivery into the brain is controlled by the blood–brain barrier (BBB). In malignant brain tumors, the BBB is damaged and macromolecules can enter into interstitial spaces. This leads to high interstitial pressure and worsens blood flow and drug delivery into tumors; however, the extent of BBB dysfunction depends on the tumor model. Here, we used EB (MW 960.8), which is widely employed to assess the integrity of the BBB in different GBM models (26), to confirm that the function of the BBB was impaired in our GL261 brain tumor model. We found that extravasation of EB was observed only at the edge and not in the center of the tumor (Supplementary Fig. S2A), suggesting that drug delivery is hindered especially in the center. In order to compare drug delivery status, we evaluated EB leakage after only 30 minutes, a very early time point after injection, and found that more of the dye was detectable in the tumor center of LPA-treated tumors (Fig. 2E). This effect was no longer seen when LPA4 KO mice were used in this tumor model. Next, we examined the distribution of 3 different sizes of FITC-conjugated dextran (MW 3–5 kDa, 70 kDa, and 2 mDa) into mice with GL261 brain tumors treated with LPA or vehicle. In the control tumor, macromolecules such as 70 kDa and 2 mDa dextran extravasated; however, 3–5 kDa dextran were barely detected in the tumor center. In contrast, in the LPA-treated tumors, larger amounts of 3 to 5 kDa dextran exited blood vessels, but extravasation of 70 kDa and 2 mDa dextran was restricted (Fig. 2F). These results indicate that LPA induced the normalization of blood vessel function and reduced the retention of large-size molecules. To confirm this, we quantified the retention of 70 kDa dextran injected 24 hours before tumor resection. Higher amounts of dextran were observed in the control tumor than in LPA-treated tumors. These differences were not seen when LPA4 KO mice were treated with LPA (Fig. 2E–H).

Figure 2.

LPA induces functional vascular network formation in GL261 brain tumors. A and B, GL261 brain tumor–bearing mice were treated with LPA, and vessel perfusion in tumors was visualized by FITC-conjugated lection. A, Representative images stained with anti-CD31 mAb (red). Scale bar, 200 μm. B, Quantification of perfused blood vessels of the tumor center by enumerating CD31+lectin+ vessels among total CD31+ vessels (n = 4 tumors per group). C and D, Analysis of hypoxia in brain tumors treated as in A. C, Tumor sections were stained with Hypoxyprobe (pimonidazole; green), anti-CD31 mAb (red), and Hoechst 33342 (blue). Scale bar, 1 mm. D, Quantification of hypoxic area (n = 8 tumor per group). E and F, Analysis of drug delivery in GL261 brain tumors generated in wild-type or LPA4 KO mice. Mice were treated with vehicle or LPA. E, Representative image of EB (150 mg/kg) injected 30 minutes before euthanization. Sections were stained with anti-CD31 mAb (green). Dashed lines in the left-side panels indicate the tumor edge. Dashed boxes in the left panels are magnified in the right panels. Scale bar, 200 μm (left panels) and 40 μm (right panels). F, Representative images of parenchymal penetration of FITC-conjugated dextran (green) with different molecular sizes. Tumor-bearing mice were intravenously injected with 3–5 kDa, 70 kDa, or 2 mDa dextran 60 minutes before euthanization. Sections were stained with anti-CD31 mAb (red) and TOPRO3 (blue). Scale bar, 40 μm. G and H, Analysis of retention of FITC-conjugated 70 kDa dextran (green) 24 hours after intravenous injection. Sections were stained with TOPRO3 (blue). Dashed lines indicate the tumor edge. Scale bar, 200 μm. H, Quantification of the FITC-positive area (n = 4 tumor per group). I and J, Survival of wild-type (I) and PA4 KO (J) mice bearing GL261 brain tumors. Mice were treated with LPA (1 mg/kg) three times a week from day 7 and/or weekly temozolomide (20 mg/kg) from day 11. K and L, Tumor vasculature after LPA and/or TMZ treatment. K, Representative images of tumor sections stained with anti-CD31 mAb (green) and TOPRO3 (blue). L, Quantification of vascular density at tumor center (n = 3 tumors per group). Scale bar, 500 μm. All experiments were repeated at least twice. Error bars, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. n.s., nonsignificant. See also Supplementary Fig. S2.

Figure 2.

LPA induces functional vascular network formation in GL261 brain tumors. A and B, GL261 brain tumor–bearing mice were treated with LPA, and vessel perfusion in tumors was visualized by FITC-conjugated lection. A, Representative images stained with anti-CD31 mAb (red). Scale bar, 200 μm. B, Quantification of perfused blood vessels of the tumor center by enumerating CD31+lectin+ vessels among total CD31+ vessels (n = 4 tumors per group). C and D, Analysis of hypoxia in brain tumors treated as in A. C, Tumor sections were stained with Hypoxyprobe (pimonidazole; green), anti-CD31 mAb (red), and Hoechst 33342 (blue). Scale bar, 1 mm. D, Quantification of hypoxic area (n = 8 tumor per group). E and F, Analysis of drug delivery in GL261 brain tumors generated in wild-type or LPA4 KO mice. Mice were treated with vehicle or LPA. E, Representative image of EB (150 mg/kg) injected 30 minutes before euthanization. Sections were stained with anti-CD31 mAb (green). Dashed lines in the left-side panels indicate the tumor edge. Dashed boxes in the left panels are magnified in the right panels. Scale bar, 200 μm (left panels) and 40 μm (right panels). F, Representative images of parenchymal penetration of FITC-conjugated dextran (green) with different molecular sizes. Tumor-bearing mice were intravenously injected with 3–5 kDa, 70 kDa, or 2 mDa dextran 60 minutes before euthanization. Sections were stained with anti-CD31 mAb (red) and TOPRO3 (blue). Scale bar, 40 μm. G and H, Analysis of retention of FITC-conjugated 70 kDa dextran (green) 24 hours after intravenous injection. Sections were stained with TOPRO3 (blue). Dashed lines indicate the tumor edge. Scale bar, 200 μm. H, Quantification of the FITC-positive area (n = 4 tumor per group). I and J, Survival of wild-type (I) and PA4 KO (J) mice bearing GL261 brain tumors. Mice were treated with LPA (1 mg/kg) three times a week from day 7 and/or weekly temozolomide (20 mg/kg) from day 11. K and L, Tumor vasculature after LPA and/or TMZ treatment. K, Representative images of tumor sections stained with anti-CD31 mAb (green) and TOPRO3 (blue). L, Quantification of vascular density at tumor center (n = 3 tumors per group). Scale bar, 500 μm. All experiments were repeated at least twice. Error bars, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. n.s., nonsignificant. See also Supplementary Fig. S2.

Close modal

Next, we assessed whether LPA improves drug delivery in the brain tumor model using TMZ, an anticancer drug widely used to treat GBM patients (1). Intravenous administration of LPA was started on day 7 after cancer cell inoculation, and LPA was injected every other day (3 times/week; on day 1, then on the second day and finally 2 days later). TMZ was i.p. injected weekly from day 11. LPA alone did not affect survival, but TMZ improved it. On the other hand, a combination of LPA with TMZ increased the anticancer effect of the latter (P = 0.0028, Log-rank test; Fig. 2I). However, the same experiment in LPA4 KO mice did not result in enhanced effects of TMZ, suggesting that LPA4 is indeed involved in mediating improved drug delivery and increased survival duration (Fig. 2J). Parallel to the survival benefit, combination therapy induced more uniform and finer vessel formation than LPA alone (Fig. 2K and L).

We assessed the improvement of drug delivery facilitated by LPA in a brain metastasis model using intracranial inoculation of LLC and with 5-FU as the anticancer drug. As expected, wild-type but not LPA4 KO mice treated with LPA and 5-FU survived longer than those given 5-FU alone (P = 0.0048, Log-rank test; Supplementary Fig. S2B and S2C). In addition, treatment with LPA and 5-FU enhanced fine capillary network formation (Supplementary Fig. S2D and S2E). Taking these results together, we conclude that LPA4 activation by LPA promotes functional vascular network formation in brain tumors and enhances the effects of anticancer drugs.

LPA activates RhoA/ROCK and induces cortical actin fiber formation in ECs

LPA induces cortical actin fiber formation under the EC membrane and stabilizes the membrane localization of VE-cadherin through LPA4 signaling in the subcutaneous tumor model (16). Here, we confirmed that LPA induced membrane localization of VE-cadherin in the GL261 brain tumor model (Supplementary Fig. S3A–S3C). Thus, tightening of adherent junctions is induced in both the brain tumor and the subcutaneous tumor cell inoculation model. Molecules composing tight junctions such as claudin 5, occludin, and ZO1 were visualized in tumor blood vessel ECs. We found that ECs in control tumors seemed not to be adhering tightly as reflected in junctional protein gaps between ECs (Supplementary Fig. S3D–S3F). This weak form of tight junction was not altered by LPA treatment. We also confirmed that expression of mRNA for molecules composing tight junctions was not affected by LPA treatment (Supplementary Fig. S3G).

The RhoA/ROCK pathway is known to be relevant for actin fiber formation (27). We hypothesized that the RhoA/ROCK pathway is involved in LPA-induced cortical actin fiber formation of ECs. MS-1 cells from a commercially-available EC line express LPA4 and LPA6 as observed in tumor ECs (Fig. 1E; ref. 16). We found that RhoA was activated following stimulation of confluent MS-1 cells with LPA or VPC31144(S) (Fig. 3A). Next, we evaluated the amounts of cortical actin fiber and VE-cadherin at cell–cell junctions in confluent MS-1 cells stimulated by LPA or VPC31144(S) in the presence or absence of Y27632 (a ROCK inhibitor; ref. 28). The increases in the amounts of cortical actin fiber and VE-cadherin induced by LPA or VPC31144(S) were suppressed in the presence of Y27632 (Fig. 3B–D). These results indicate that stimulation of LPA4 promotes cortical actin formation and strengthens interendothelial adherent junctions through the RhoA/ROCK pathway.

Figure 3.

LPA4 signaling activates the RhoA/ROCK pathway and induces cortical actin fiber formation under the endothelial cell membrane. A, Relative activation of RhoA. Confluent MS-1 cells were stimulated with vehicle or 10 μmol/L LPA or VPC31144(S) (n = 3 experiments per group). B–D, Localization of VE-cadherin (green) and F-actin (red) in MS-1 cells. Monolayers of MS-1 cells were pretreated with Y27632 (10 μmol/L) or vehicle for 60 minutes and subsequently stimulated with 10 μmol/L LPA or VPC31144(S). The area enclosed by the dashed box is magnified in the bottom three plots (B). Scale bar, 10 μm. Quantification of VE-cadherin (C) and F-actin (D) at cell–cell junctions relative to control (n = 16 points per group). All experiments were repeated at least twice. Error bars, mean ± SEM. **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S3.

Figure 3.

LPA4 signaling activates the RhoA/ROCK pathway and induces cortical actin fiber formation under the endothelial cell membrane. A, Relative activation of RhoA. Confluent MS-1 cells were stimulated with vehicle or 10 μmol/L LPA or VPC31144(S) (n = 3 experiments per group). B–D, Localization of VE-cadherin (green) and F-actin (red) in MS-1 cells. Monolayers of MS-1 cells were pretreated with Y27632 (10 μmol/L) or vehicle for 60 minutes and subsequently stimulated with 10 μmol/L LPA or VPC31144(S). The area enclosed by the dashed box is magnified in the bottom three plots (B). Scale bar, 10 μm. Quantification of VE-cadherin (C) and F-actin (D) at cell–cell junctions relative to control (n = 16 points per group). All experiments were repeated at least twice. Error bars, mean ± SEM. **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S3.

Close modal

LPA increases tumor-infiltrating lymphocytes in brain tumors

Increases in numbers of functional vessels in tumors and improvement of the tumor microenvironment are believed to be related to immune system activity against the tumor (6). Here, we tested whether LPA affects lymphocyte infiltration into GL261 brain tumors. We injected LPA into GL261 brain tumor–bearing mice following the same schedule as described in Fig. 1B. We then evaluated tumor sections and found that numbers of CD4+ and CD8+ T cells increased with LPA administration in tumors generated in wild-type but not in LPA4 KO mice (Fig. 4A–C). LPA injection into normal mice had no effect on lymphocyte counts in the lymph nodes (Supplementary Fig. S4A). Moreover, we also examined infiltration of myeloid cells into tumors and found no significant differences (Supplementary Fig. S4B and S4C). To investigate the mechanism responsible for this increase of tumor-infiltrating lymphocytes (TIL), we focused on VCAM-1 and ICAM-1, which is a leukocyte adhesion molecule expressed by ECs (29). We found that the number of VCAM-1–expressing vessels increased in LPA-treated wild-type but not LPA4 KO mice (Fig. 4D and E). Furthermore, we found that the number of ICAM-1–expressing vessels increased in LPA-treated wild-type mice (Supplementary Fig. S4D and S4E). High endothelial venules (HEV) are other candidates for lymphocyte recruitment (30), and we hypothesized that LPA changes immature tumor vessels into HEVs. However, LPA administration did not alter the expression of MECA79, an HEV marker protein, by GL261 tumor ECs (Supplementary Fig. S4F and S4G). It has been suggested that HEVs located in the tumor induce infiltration by lymphocytes (9, 31–33). Therefore, we tested whether LPA treatment also alters the expression of HEV-related molecules (Ccl19, Ccl21, Cxcl13, and Glycam1). However, partly because of the large individual variability, we could not detect any significant differences of their expression. These data suggest that LPA induces VCAM-1 expression on ECs via LPA4 and thus increases the number of TILs.

Figure 4.

LPA induces leukocyte adhesion molecule expression on endothelial cells and increases the number of tumor-infiltrating lymphocytes. A–C, Lymphocyte infiltration into GL261 brain tumors generated in wild-type or LPA4 KO mice. Mice were treated with vehicle or LPA. A, Representative images of tumor sections stained with anti-CD4 mAb (red), anti-CD8 mAb (green), and anti-CD31 mAb (blue). Scale bar, 100 μm. Quantification of the number of tumor-infiltrating CD4-positive (B) and CD8-positive (C) lymphocytes (n = 12 tumors per group). D and E, VCAM-1 expression on ECs in GL261 brain tumors generated in wild-type or LPA4 KO. Mice were treated with vehicle or LPA. D, Representative images of tumor sections stained with anti–VCAM-1 mAb (green) and anti-CD31 mAb (red). Scale bar, 100 μm. E, Quantification of the percentage of VCAM-1–expressing vessels among total blood vessels (n = 12 tumors per group). F–K, Analysis of VCAM-1 expression on MS-1 cells under various conditions. MS-1 cells were stimulated with different doses of LPA or VPC31144(S). Cells were stained with anti–VCAM-1 mAb (green), and nuclei were labeled with TOPRO3 (blue). Scale bar, 100 μm. F, Time course of VCAM-1 expression was evaluated. Quantification is shown in G (n = 4 experiments per group). H, VCAM-1 expression on MS-1 cells 24 hours after VPC31144(S) stimulation. Quantification is shown in I (n = 4 experiments per group). J, MS-1 cells were transfected with negative control siRNA (siCTR), LPA4 siRNA (siLPA4), or LPA6 siRNA (siLPA6) and then stimulated with 100 μmol/L LPA or vehicle. Quantification is shown in K (n = 4 experiments per group). All experiments were repeated at least twice. Error bars, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. n.s., nonsignificant. See also Supplementary Fig. S4.

Figure 4.

LPA induces leukocyte adhesion molecule expression on endothelial cells and increases the number of tumor-infiltrating lymphocytes. A–C, Lymphocyte infiltration into GL261 brain tumors generated in wild-type or LPA4 KO mice. Mice were treated with vehicle or LPA. A, Representative images of tumor sections stained with anti-CD4 mAb (red), anti-CD8 mAb (green), and anti-CD31 mAb (blue). Scale bar, 100 μm. Quantification of the number of tumor-infiltrating CD4-positive (B) and CD8-positive (C) lymphocytes (n = 12 tumors per group). D and E, VCAM-1 expression on ECs in GL261 brain tumors generated in wild-type or LPA4 KO. Mice were treated with vehicle or LPA. D, Representative images of tumor sections stained with anti–VCAM-1 mAb (green) and anti-CD31 mAb (red). Scale bar, 100 μm. E, Quantification of the percentage of VCAM-1–expressing vessels among total blood vessels (n = 12 tumors per group). F–K, Analysis of VCAM-1 expression on MS-1 cells under various conditions. MS-1 cells were stimulated with different doses of LPA or VPC31144(S). Cells were stained with anti–VCAM-1 mAb (green), and nuclei were labeled with TOPRO3 (blue). Scale bar, 100 μm. F, Time course of VCAM-1 expression was evaluated. Quantification is shown in G (n = 4 experiments per group). H, VCAM-1 expression on MS-1 cells 24 hours after VPC31144(S) stimulation. Quantification is shown in I (n = 4 experiments per group). J, MS-1 cells were transfected with negative control siRNA (siCTR), LPA4 siRNA (siLPA4), or LPA6 siRNA (siLPA6) and then stimulated with 100 μmol/L LPA or vehicle. Quantification is shown in K (n = 4 experiments per group). All experiments were repeated at least twice. Error bars, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. n.s., nonsignificant. See also Supplementary Fig. S4.

Close modal

It has been reported that human umbilical vein endothelial cells express leukocyte adhesion molecules induced by inflammatory cytokines (34). We investigated whether LPA induces VCAM-1 expression on LPA4-expressing MS-1 cells. Under control conditions, VCAM-1+ cells were rare; however, LPA induced VCAM-1 expression in a dose- and time-dependent manner (Fig. 4F and G). In order to determine which receptor is involved in inducing VCAM-1 expression, we stimulated MS-1 cells with VPC31144(S) and confirmed that it did induce VCAM-1 expression (Fig. 4H and I). Next, we decreased the expression of either LPA4 or LPA6 by siRNAs (siLPA4 or siLPA6). Knockdown efficacy confirmed by quantitative real-time PCR was almost the same as previously reported (LPA4 4.2%, LPA6 12.1%; ref. 16). The expression of VCAM-1 induced by LPA was perturbed by LPA4 but not by LPA6 knockdown (Fig. 4J and K). Taking these results together, we conclude that LPA stimulates LPA4 and induces VCAM-1 expression on ECs.

LPA4 signaling induces nuclear translocation of NF-κB and promotes VCAM-1 transcription

LPA receptors belong to the family of G-protein–coupled receptors (35). In order to determine which G-protein mediates LPA signaling resulting in VCAM-1 expression, we tested 2′5′DDA and SQ22536 which inhibit cAMP (Gs inhibitors), YM254890 (a Gq inhibitor), and NF023 (a Gi inhibitor). None of these inhibitors prevented VCAM-1 expression on MS-1 cells after stimulation with LPA (Supplementary Fig. S5A). Next, we reduced the expression of Gα12/13 by silencing GNA13 (siGNA13) and found that VCAM-1 expression was now inhibited (Fig. 5A and B). It is known that RhoA/ROCK is involved in downstream signaling of Gα12/13, and that RhoA/ROCK induces VCAM-1 expression on ECs (36, 37). Therefore, we inhibited ROCK by exposing MS-1 cells stimulated with LPA to Y27632. This resulted in the suppression of LPA-induced VCAM-1 expression (Fig. 5C and D).

Figure 5.

VCAM-1 expression of MS-1 cells induced by LPA is controlled by RhoA, p38, and NF-κB. A–D, Analysis of VCAM-1 expression on MS-1 cells. MS-1 cells were stained with anti–VCAM-1 mAb (green), and nuclei were labeled with TOPRO3 (blue; A and C). Scale bar, 100 μm. MS-1 cells were transfected with negative control siRNA (siCTR) or siRNA for GNA13 (siGNA13) and then stimulated with 100 μmol/L LPA or vehicle. Representative images are shown in A, and quantification is shown in B (n = 4 experiments per group). MS-1 cells were pretreated with Y27632 or DMSO (control) for 60 minutes and stimulated with 100 μmol/L LPA. Representative images are shown in C, and quantification is shown in D (n = 4 experiments per group). E, Western blotting for detecting phosphorylated p38 and VCAM-1. Confluent MS-1 cells were stimulated with 100 μmol/L LPA. F, Quantification of VCAM-1–positive MS-1 cells pretreated with PD-98059 (20 μmol/L, 60 minutes), U0126 (10 μmol/L, 60 minutes), SB202190 (5 μmol/L, 60 minutes), VX702 (1 μmol/L, 60 minutes), SP600125 (20 μmol/L, 60 minutes), LY294002 (50 μmol/L, 30 minutes), or wortmannin (100 nmol/L, 30 minutes). MS-1 cells were subsequently stimulated with 100 μmol/L LPA or vehicle for 24 hours (n = 4 experiments per group). G and H, Time course of nuclear translocation of NF-κB in MS-1 cells stimulated with 100 μmol/L LPA. G, Representative images of MS-1 cells stained with anti–NF-κB mAb (green) and anti–VE-cadherin mAb (red), and nuclei were labeled with TOPRO3 (blue). Scale bar, 50 μm. H, Quantification of internalized NF-κB (n = 4 experiments per group). I, Quantification of VCAM-1–positive MS-1 cells pretreated with PDTC for 60 minutes. Cells were subsequently stimulated with 100 μmol/L LPA (n = 4 experiments per group). J, Quantitative real-time PCR of VCAM-1 mRNA expression in MS-1 cells pretreated with PDTC for 60 minutes and subsequently stimulated with 100 μmol/L LPA for 12 hours (n = 4 experiments per group). All experiments were repeated at least twice. Error bars, mean ± SEM. **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S5.

Figure 5.

VCAM-1 expression of MS-1 cells induced by LPA is controlled by RhoA, p38, and NF-κB. A–D, Analysis of VCAM-1 expression on MS-1 cells. MS-1 cells were stained with anti–VCAM-1 mAb (green), and nuclei were labeled with TOPRO3 (blue; A and C). Scale bar, 100 μm. MS-1 cells were transfected with negative control siRNA (siCTR) or siRNA for GNA13 (siGNA13) and then stimulated with 100 μmol/L LPA or vehicle. Representative images are shown in A, and quantification is shown in B (n = 4 experiments per group). MS-1 cells were pretreated with Y27632 or DMSO (control) for 60 minutes and stimulated with 100 μmol/L LPA. Representative images are shown in C, and quantification is shown in D (n = 4 experiments per group). E, Western blotting for detecting phosphorylated p38 and VCAM-1. Confluent MS-1 cells were stimulated with 100 μmol/L LPA. F, Quantification of VCAM-1–positive MS-1 cells pretreated with PD-98059 (20 μmol/L, 60 minutes), U0126 (10 μmol/L, 60 minutes), SB202190 (5 μmol/L, 60 minutes), VX702 (1 μmol/L, 60 minutes), SP600125 (20 μmol/L, 60 minutes), LY294002 (50 μmol/L, 30 minutes), or wortmannin (100 nmol/L, 30 minutes). MS-1 cells were subsequently stimulated with 100 μmol/L LPA or vehicle for 24 hours (n = 4 experiments per group). G and H, Time course of nuclear translocation of NF-κB in MS-1 cells stimulated with 100 μmol/L LPA. G, Representative images of MS-1 cells stained with anti–NF-κB mAb (green) and anti–VE-cadherin mAb (red), and nuclei were labeled with TOPRO3 (blue). Scale bar, 50 μm. H, Quantification of internalized NF-κB (n = 4 experiments per group). I, Quantification of VCAM-1–positive MS-1 cells pretreated with PDTC for 60 minutes. Cells were subsequently stimulated with 100 μmol/L LPA (n = 4 experiments per group). J, Quantitative real-time PCR of VCAM-1 mRNA expression in MS-1 cells pretreated with PDTC for 60 minutes and subsequently stimulated with 100 μmol/L LPA for 12 hours (n = 4 experiments per group). All experiments were repeated at least twice. Error bars, mean ± SEM. **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S5.

Close modal

It has been reported that LPA4 signaling activates several different mediators (38); accordingly, we found that LPA or VPC31144(S) induces phosphorylation of Akt, Erk, and p38 in MS-1 cells (Fig. 5E; Supplementary S5B). To investigate which signal cascade mediates VCAM-1 expression, we used PD-98059 and U0126 (MEK inhibitors), SB202190 and VX702 (p38 inhibitors), SP600125 (JNK inhibitor), and LY294002 and wartmannin (Akt/PI3K inhibitors). Of these agents, only the p38 inhibitors prevented VCAM-1 expression (Fig. 5F; Supplementary S5C). Subsequently, we searched for transcription factors involved in VCAM-1 upregulation. It was reported that NF-κB plays an important role in VCAM-1 transcription in ECs (39). We therefore investigated nuclear translocation of NF-κB in MS-1 cells after stimulation for 1 hour with LPA (Fig. 5G and H). We hypothesized that NF-κB controls VCAM-1 expression on MS-1 cells stimulated by LPA. To test this, we used the NF-κB inhibitor PDTC, which was found to suppress Vcam-1 expression by MS-1 cells at both the mRNA and protein levels (Fig. 5I and J; Supplementary S5D). Taking all the data together, we conclude that LPA4 couples with Gα12/13, and its signaling induces nuclear translocation of NF-κB through Rho/ROCK and p38, leading to VCAM-1 transcription.

LPA enhances the anticancer effect of anti–PD-1 antibody therapy

Although LPA treatment increased the number of TILs in GL 261 brain tumors, it did not lead to a survival benefit when used as a single agent (Fig. 2I). One possible reason for this is that tumor cells express immune-checkpoint proteins and escape from immune destruction. To test this, we transduced the EGFP gene into GL261 (GL261-EGFP) and LLC (LLC-EGFP) cells and generated GL261 brain tumors and LLC subcutaneous tumors which could be distinguished from recipient mouse cells by green fluorescence. We confirmed that both EGFP-expressing tumor cells expressed PD-L1 (Supplementary Fig. S6A and S6B). It has been reported that the expression of PD-L1 is enhanced by antiangiogenic therapy (9). However, we found that PD-L1 expression by GL261 brain tumor cells was not affected by LPA administration (Supplementary Fig. S6C). Next, we investigated whether LPA improves the delivery of IgG into the tumor parenchyma by intravenously injecting IgG into mice harboring GL261 brain tumors. The injected IgG distributed mainly into the outer part of the tumor but did not penetrate into the center of the tumor in controls. In contrast, in the LPA-treated mice, the injected IgG penetrated with a radial pattern from the tumor center and entered deep into the parenchyma (Fig. 6A and B). Next, we determined the number of lymphocytes binding anti–PD-1 antibody after its i.p. injection into GL261 brain tumor–bearing mice. The percentage of total lymphocytes binding anti–PD-1 antibody increased in the tumors of LPA-treated mice (Fig. 6C and D). Finally, we determined the effect of combining LPA with anti–PD-1 antibody treatment in both tumor models. In the GL261 brain tumor model, anti–PD-1 therapy alone extended median survival by 5 days relative to the control group (P = 0.0005, Log-rank test). However, a combination of LPA and anti–PD-1 antibody extended median survival by 11 days relative to the single-agent anti–PD-1 antibody (P = 0.0011, Log-rank test; Fig. 6E and F). In this experiment, 2 mice survived more than 50 days; we harvested the brains of these mice and stained the sections with hematoxylin & eosin to confirm the presence of resident tumor cells. This revealed that tumor cells had been almost completely eliminated such that these mice showed complete responses to this combination treatment (Supplementary Fig. S6D–S6F). In the LLC subcutaneous model, single-agent anti–PD-1 therapy failed to mediate significant antitumor effects. However, combined anti–PD-1 antibody and LPA was able to inhibit tumor growth (Fig. 6G and H). Taking these data together, we conclude that LPA improves the delivery of exogenous IgG and enhances infiltration of lymphocytes into tumors, thus facilitating anti–PD-1 antibody therapy.

Figure 6.

LPA enhances the antitumor effect of anti–PD-1 therapy. A and B, Tumor parenchymal delivery of IgG intravenously injected into GL261 brain tumor–bearing mice. Mice were treated with vehicle or LPA. Tumors were harvested 1 hour after IgG injection. A, Representative images stained with anti-CD31 mAb (red). Injected IgG was labeled with Alexa Fluor 488–conjugated anti-rat IgG. The dashed lines indicate tumor edge, and the dashed white boxes in the top panels are magnified in the bottom panels. Scale bar, 200 μm (top panels) and 40 μm (bottom panels). B, Quantification of IgG-positive areas (n = 3 tumor per group). C and D, Infiltration of anti–PD-1 antibody–conjugated lymphocytes. GL261 brain tumor–bearing mice were treated with vehicle or LPA. Anti–PD-1 antibody was administered 24 hours before tumor resection. C, Representative images stained with anti-CD3 mAb (red) and anti-vW F pAb (blue). Injected anti–PD-1 antibody was labeled with Alexa Fluor 488–conjugated anti-rat IgG. Scale bar, 40 μm. D, Quantification of the percentage of CD3+ lymphocytes conjugated with anti–PD-1 antibody within total CD3+ lymphocytes (n = 3 per group). E and F, Survival of GL261 brain tumor–bearing mice treated with LPA and/or anti–PD-1 antibody. Administration schedule is shown in E. G and H, LLC subcutaneous tumor growth treated with LPA and/or anti–PD-1 antibody. Administration schedule is shown in G. All experiments were repeated at least twice. Error bars, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. n.s., nonsignificant. See also Supplementary Fig. S6.

Figure 6.

LPA enhances the antitumor effect of anti–PD-1 therapy. A and B, Tumor parenchymal delivery of IgG intravenously injected into GL261 brain tumor–bearing mice. Mice were treated with vehicle or LPA. Tumors were harvested 1 hour after IgG injection. A, Representative images stained with anti-CD31 mAb (red). Injected IgG was labeled with Alexa Fluor 488–conjugated anti-rat IgG. The dashed lines indicate tumor edge, and the dashed white boxes in the top panels are magnified in the bottom panels. Scale bar, 200 μm (top panels) and 40 μm (bottom panels). B, Quantification of IgG-positive areas (n = 3 tumor per group). C and D, Infiltration of anti–PD-1 antibody–conjugated lymphocytes. GL261 brain tumor–bearing mice were treated with vehicle or LPA. Anti–PD-1 antibody was administered 24 hours before tumor resection. C, Representative images stained with anti-CD3 mAb (red) and anti-vW F pAb (blue). Injected anti–PD-1 antibody was labeled with Alexa Fluor 488–conjugated anti-rat IgG. Scale bar, 40 μm. D, Quantification of the percentage of CD3+ lymphocytes conjugated with anti–PD-1 antibody within total CD3+ lymphocytes (n = 3 per group). E and F, Survival of GL261 brain tumor–bearing mice treated with LPA and/or anti–PD-1 antibody. Administration schedule is shown in E. G and H, LLC subcutaneous tumor growth treated with LPA and/or anti–PD-1 antibody. Administration schedule is shown in G. All experiments were repeated at least twice. Error bars, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. n.s., nonsignificant. See also Supplementary Fig. S6.

Close modal

Here, we have shown that LPA promotes functional vascular network formation in brain tumors and improves drug delivery. LPA activates the RhoA/ROCK pathway and induces cortical actin fiber formation under the EC cell membrane, which leads to membrane localization of VE-cadherin and tightened EC–cell adhesion. It is widely accepted that abnormal and immature tumor vessels leak macromolecules into the interstitial space and that their retention increases the interstitial pressure. This decreases blood perfusion, a phenomenon designated “enhanced permeability and retention (EPR)” (40). We conclude that LPA normalizes vascularization allowing small-molecule delivery and counteracts macromolecule delivery due to the EPR. Accordingly, LPA administration permitted penetration of TMZ to inhibit tumor growth.

It has been reported that ATX, an enzyme involved in LPA synthesis, is overexpressed in GBM and contributes to cell motility and glioma invasion (41, 42). However, these reports did not focus on the tumor ECs. Moreover, we showed that LPA alone did not affect the survival of mice harboring GL261 brain tumors. Other reports described that inhibition of ATX or LPA receptors increases the radiosensitivity of GBM cells in hind limb models (43, 44). Reports that inhibition of LPA induces antiangiogenic effects are similar to the results presented here on LPA-promoting angiogenesis, but contrasting results that LPA is involved in malignancy in previous papers may derive from the use of different tumor models which would have different tumor microenvironments. Current standard therapy for GBM includes surgery and radiotherapy plus concomitant and adjuvant TMZ chemotherapy (1). Improvement of the intratumoral microenvironment is reported to enhance the anticancer effect even of radiotherapy (45, 46). Therefore, treatment with LPA may be incorporated into current chemoradiotherapy regimes for GBM patients.

We show that LPA strengthens EC–cell junction adhesion and reduces the leakage of large molecules. Therefore, we presumed that LPA reduced the extravasation of IgG, the molecular weight of which is approximately 160 kDa. However, exogenously-injected IgG was found to be widely distributed in the tumor following LPA treatment. It has been reported that delivery of IgG is mediated by transcytosis dependent on the neonatal Fc-receptor expressed on ECs (47). This therefore implies that extravasation of IgG would not be affected by the tightness of EC–cell junctions. Further precise analysis is required to understand these processes. Nonetheless, it is suggested that LPA treatment affects the BBB function in tumor ECs to enhance IgG delivery.

Enhanced delivery of IgG induced by LPA treatment will also benefit immunotherapy using anti–PD-L1 antibody because antibody binding to tumor cells and their anticancer effects depend on the delivery of IgG into the tumor parenchyma (48). Recently, adoptive T-cell transfer approaches using T-cell receptor– and chimeric antigen receptor–modified T cells have been attracting a great deal of attention among immunotherapists (49). LPA increases the number of TILs, and therefore such treatment might amplify the anticancer effect of these therapies as well.

Inflammatory cytokines stimulate enhanced VCAM-1 transcription in ECs via activation of NF-κB (39). Other reports indicate that activation of RhoA leads to transcription of VCAM-1 (37). We showed that LPA activates NF-κB through RhoA/ROCK and p38, which induce expression of VCAM-1. The NF-κB pathway in ECs is known to induce vascular inflammation and to enhance vascular permeability (50, 51). However, we also demonstrated that LPA strengthened EC–cell junctions. This is because LPA-mediated RhoA/ROCK activation triggers other signaling pathways which lead to cortical actin formation and membrane localization of VE-cadherin. Therefore, lymphocyte infiltration and tightened interendothelial junctions are two characteristics induced by LPA. Nevertheless, we need to further clarify these results in terms of the mechanism underlying these two contradictory features depending on EC phenotypes such as venules and capillaries and the functional heterogeneity of ECs in tumors.

In our study, we showed that LPA treatment results in increases of TILs. It has been reported that HEVs located in the tumor increase TILs (9, 31–33). However, LPA does not induce HEV-related molecules. Thus, we focused on the adhesion molecules. Leukocyte infiltration begins with rolling on ECs, whereby leukocytes and ECs contact each other, but this weak adhesion mediated by carbohydrate chains on leukocytes and E-selection on EC does not result in arresting the movement of the leukocytes. The next step required for leukocyte infiltration is firm adhesion between leukocytes and ECs. ICAM-1 and VCAM-1 on ECs contribute to this tight adhesion. In vivo, LPA was found to increase expression of ICAM-1 and VCAM-1, but not E-selectin at least in mRNA level. This therefore documents that LPA controls leukocyte adhesion to EC without the involvement of E-selection. It has been reported that the tumor microenvironment contributes to the induction of lymphocyte anergy (52), and that altering the microenvironment can restore immune responses against tumors and delay cancer progression (53). We determined the percentage of Treg (FOXP3+ CD25+) within CD3+ lymphocytes; however, LPA treatment did not alter Treg infiltration. Further precise analysis is required to identify the mechanism whereby Treg infiltration is not altered even though the tumor microenvironment is improved by LPA in terms of the infiltration of other T cells.

A single treatment modality is rarely effective in GBM patients, and combinations of multimodal treatments, such as surgical resection, chemotherapy, radiotherapy, and immunotherapy, are essential. Increasing the extent of resection is associated with improved survival of malignant glioma patients, and surgical resection remains of great importance for GBM treatment. However, extensive resection may be associated with postoperative dysfunction. TMZ is the chemotherapeutic agent of choice for GBM treatment, but median survival is less than 15 months even when combined with radiotherapy. In addition, TMZ is immunosuppressive and reduces the antitumor effect of anti–PD-1 therapy (54). Immunotherapy may have fewer side effects than other types of cancer therapy, but its anticancer efficacy may not be sufficient and may be too delayed to control aggressive disease. In order to develop less invasive and more effective treatments, control of the tumor vasculature may emerge as an important element together with these modalities. In this context, LPA is a promising candidate for clinical application.

No potential conflicts of interest were disclosed.

Conception and design: D. Eino, N. Takakura

Development of methodology: D. Eino, Y. Tsukada

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D. Eino, Y. Tsukada, H. Naito, Y. Kanemura, T. Iba, T. Wakabayashi, H. Kidoya

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D. Eino, Y. Tsukada, K. Takara

Writing, review, and/or revision of the manuscript: D. Eino, Y. Tsukada, N. Takakura

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Kanemura, F. Muramatsu, H. Arita, N. Kagawa, Y. Fujimoto

Study supervision: H. Kishima, N. Takakura

We thank Dr. S. Ishii and D. Yasuda (Department of Immunology, Akita University Graduate School of Medicine) for providing us LPA4 KO mice. We thank N. Fujimoto, M. Ishida, and Y. Mori for technical assistance. This work was partly supported by the Japan Agency for Medical Research and Development (AMED) under grant numbers 18cm0106508h0002, 18gm5010002s110, Japan Society for the Promotion of Science (JSPS) Grants-in-Aid for Scientific Research (A) (16H02470), Grant-in-Aid for Young Scientists (B) (15K19968), Grant-in-Aid for JSPS Fellows (17J04925), Project MEET, Osaka University Graduate School of Medicine, Grant from Mitsubishi Tanabe Pharma Corporation, and Integrated Frontier Research for Medical Science Division, Institute for Open and Transdisciplinary Research Initiatives, Osaka University.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Stupp
R
,
Mason
WP
,
van den Bent
MJ
,
Weller
M
,
Fisher
B
,
Taphoorn
MJ
, et al
Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma
.
N Engl J Med
2005
;
352
:
987
96
.
2.
Takano
S
,
Yoshii
Y
,
Kondo
S
,
Suzuki
H
,
Maruno
T
,
Shirai
S
, et al
Concentration of vascular endothelial growth factor in the serum and tumor tissue of brain tumor patients
.
Cancer Res
1996
;
56
:
2185
90
.
3.
von Baumgarten
L
,
Brucker
D
,
Tirniceru
A
,
Kienast
Y
,
Grau
S
,
Burgold
S
, et al
Bevacizumab has differential and dose-dependent effects on glioma blood vessels and tumor cells
.
Clin Cancer Res
2011
;
17
:
6192
205
.
4.
Gilbert
MR
,
Dignam
JJ
,
Armstrong
TS
,
Wefel
JS
,
Blumenthal
DT
,
Vogelbaum
MA
, et al
A randomized trial of bevacizumab for newly diagnosed glioblastoma
.
N Engl J Med
2014
;
370
:
699
708
.
5.
Chinot
OL
,
Wick
W
,
Mason
W
,
Henriksson
R
,
Saran
F
,
Nishikawa
R
, et al
Bevacizumab plus radiotherapy-temozolomide for newly diagnosed glioblastoma
.
N Engl J Med
2014
;
370
:
709
22
.
6.
Jain
RK
. 
Antiangiogenesis strategies revisited: from starving tumors to alleviating hypoxia
.
Cancer Cell
2014
;
26
:
605
22
.
7.
Jain
RK
. 
Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy
.
Science
2005
;
307
:
58
62
.
8.
Naito
H
,
Takara
K
,
Wakabayashi
T
,
Kawahara
H
,
Kidoya
H
,
Takakura
N
. 
Changes in blood vessel maturation in the fibrous cap of the tumor rim
.
Cancer Sci
2012
;
103
:
433
8
.
9.
Allen
E
,
Jabouille
A
,
Rivera
LB
,
Lodewijckx
I
,
Missiaen
R
,
Steri
V
, et al
Combined antiangiogenic and anti-PD-L1 therapy stimulates tumor immunity through HEV formation
.
Sci Transl Med
2017
;
9
.
10.
Huang
Y
,
Goel
S
,
Duda
DG
,
Fukumura
D
,
Jain
RK
. 
Vascular normalization as an emerging strategy to enhance cancer immunotherapy
.
Cancer Res
2013
;
73
:
2943
8
.
11.
Wong
PP
,
Demircioglu
F
,
Ghazaly
E
,
Alrawashdeh
W
,
Stratford
MR
,
Scudamore
CL
, et al
Dual-action combination therapy enhances angiogenesis while reducing tumor growth and spread
.
Cancer Cell
2015
;
27
:
123
37
.
12.
Kono
M
,
Mi
Y
,
Liu
Y
,
Sasaki
T
,
Allende
ML
,
Wu
YP
, et al
The sphingosine-1-phosphate receptors S1P1, S1P2, and S1P3 function coordinately during embryonic angiogenesis
.
J Biol Chem
2004
;
279
:
29367
73
.
13.
Yukiura
H
,
Hama
K
,
Nakanaga
K
,
Tanaka
M
,
Asaoka
Y
,
Okudaira
S
, et al
Autotaxin regulates vascular development via multiple lysophosphatidic acid (LPA) receptors in zebrafish
.
J Biol Chem
2011
;
286
:
43972
83
.
14.
Aoki
J
. 
Mechanisms of lysophosphatidic acid production
.
Semin Cell Dev Biol
2004
;
15
:
477
89
.
15.
Tanaka
M
,
Okudaira
S
,
Kishi
Y
,
Ohkawa
R
,
Iseki
S
,
Ota
M
, et al
Autotaxin stabilizes blood vessels and is required for embryonic vasculature by producing lysophosphatidic acid
.
J Biol Chem
2006
;
281
:
25822
30
.
16.
Takara
K
,
Eino
D
,
Ando
K
,
Yasuda
D
,
Naito
H
,
Tsukada
Y
, et al
Lysophosphatidic acid receptor 4 activation augments drug delivery in tumors by tightening endothelial cell-cell contact
.
Cell Rep
2017
;
20
:
2072
86
.
17.
Sumida
H
,
Noguchi
K
,
Kihara
Y
,
Abe
M
,
Yanagida
K
,
Hamano
F
, et al
LPA4 regulates blood and lymphatic vessel formation during mouse embryogenesis
.
Blood
2010
;
116
:
5060
70
.
18.
Sakamoto
N
,
Uemae
Y
,
Ishikawa
E
,
Takano
S
,
Nakai
K
,
Yamamoto
T
, et al
Glioma immunotherapy with combined autologous tumor cell and endothelial cell vaccine in vivo
.
Neurol Med Chir (Tokyo)
2012
;
52
:
194
201
.
19.
Bamba
Y
,
Shofuda
T
,
Kanematsu
D
,
Nonaka
M
,
Yamasaki
M
,
Okano
H
, et al
Differentiation, polarization, and migration of human induced pluripotent stem cell-derived neural progenitor cells co-cultured with a human glial cell line with radial glial-like characteristics
.
Biochem Biophys Res Commun
2014
;
447
:
683
8
.
20.
Kanemura
Y
,
Mori
H
,
Kobayashi
S
,
Islam
O
,
Kodama
E
,
Yamamoto
A
, et al
Evaluation of in vitro proliferative activity of human fetal neural stem/progenitor cells using indirect measurements of viable cells based on cellular metabolic activity
.
J Neurosci Res
2002
;
69
:
869
79
.
21.
Tu
J
,
Hu
Z
,
Chen
Z
. 
A combination of radiosurgery and soluble tissue factor enhances vascular targeting for experimental glioblastoma
.
BioMed Res Int
2013
;
2013
:
390714
.
22.
Naito
H
,
Kidoya
H
,
Sakimoto
S
,
Wakabayashi
T
,
Takakura
N
. 
Identification and characterization of a resident vascular stem/progenitor cell population in preexisting blood vessels
.
EMBO J
2012
;
31
:
842
55
.
23.
Naito
H
,
Wakabayashi
T
,
Kidoya
H
,
Muramatsu
F
,
Takara
K
,
Eino
D
, et al
Endothelial side population cells contribute to tumor angiogenesis and antiangiogenic drug resistance
.
Cancer Res
2016
;
76
:
3200
10
.
24.
Kidoya
H
,
Ueno
M
,
Yamada
Y
,
Mochizuki
N
,
Nakata
M
,
Yano
T
, et al
Spatial and temporal role of the apelin/APJ system in the caliber size regulation of blood vessels during angiogenesis
.
EMBO J
2008
;
27
:
522
34
.
25.
Kihara
Y
,
Maceyka
M
,
Spiegel
S
,
Chun
J
. 
Lysophospholipid receptor nomenclature review: IUPHAR Review 8
.
Br J Pharmacol
2014
;
171
:
3575
94
.
26.
Leten
C
,
Struys
T
,
Dresselaers
T
,
Himmelreich
U
. 
In vivo and ex vivo assessment of the blood brain barrier integrity in different glioblastoma animal models
.
J Neurooncol
2014
;
119
:
297
306
.
27.
Hall
A
. 
Rho GTPases and the actin cytoskeleton
.
Science
1998
;
279
:
509
14
.
28.
Ando
K
,
Fukuhara
S
,
Moriya
T
,
Obara
Y
,
Nakahata
N
,
Mochizuki
N
. 
Rap1 potentiates endothelial cell junctions by spatially controlling myosin II activity and actin organization
.
J Cell Biol
2013
;
202
:
901
16
.
29.
Barreiro
O
,
Yáñez-Mó
M
,
Serrador
JM
,
Montoya
MC
,
Vicente-Manzanares
M
,
Tejedor
R
, et al
Dynamic interaction of VCAM-1 and ICAM-1 with moesin and ezrin in a novel endothelial docking structure for adherent leukocytes
.
J Cell Biol
2002
;
157
:
1233
45
.
30.
Martinet
L
,
Garrido
I
,
Filleron
T
,
Le Guellec
S
,
Bellard
E
,
Fournie
JJ
, et al
Human solid tumors contain high endothelial venules: association with T- and B-lymphocyte infiltration and favorable prognosis in breast cancer
.
Cancer Res
2011
;
71
:
5678
87
.
31.
Schmittnaegel
M
,
Rigamonti
N
,
Kadioglu
E
,
Cassara
A
,
Wyser Rmili
C
,
Kiialainen
A
, et al
Dual angiopoietin-2 and VEGFA inhibition elicits antitumor immunity that is enhanced by PD-1 checkpoint blockade
.
Sci Transl Med
2017
;
9
.
32.
Tian
L
,
Goldstein
A
,
Wang
H
,
Ching Lo
H
,
Sun Kim
I
,
Welte
T
, et al
Mutual regulation of tumour vessel normalization and immunostimulatory reprogramming
.
Nature
2017
;
544
:
250
4
.
33.
Uldry
E
,
Faes
S
,
Demartines
N
,
Dormond
O
. 
Fine-tuning tumor endothelial cells to selectively kill cancer
.
Int J Mol Sci
2017
;
18
.
34.
Zhou
Z
,
Connell
MC
,
MacEwan
DJ
. 
TNFR1-induced NF-kappaB, but not ERK, p38MAPK or JNK activation, mediates TNF-induced ICAM-1 and VCAM-1 expression on endothelial cells
.
Cell Signal
2007
;
19
:
1238
48
.
35.
Anliker
B
,
Chun
J
. 
Lysophospholipid G protein-coupled receptors
.
J Biol Chem
2004
;
279
:
20555
8
.
36.
Bian
D
,
Mahanivong
C
,
Yu
J
,
Frisch
SM
,
Pan
ZK
,
Ye
RD
, et al
The G12/13-RhoA signaling pathway contributes to efficient lysophosphatidic acid-stimulated cell migration
.
Oncogene
2006
;
25
:
2234
44
.
37.
Cernuda-Morollon
E
,
Ridley
AJ
. 
Rho GTPases and leukocyte adhesion receptor expression and function in endothelial cells
.
Circ Res
2006
;
98
:
757
67
.
38.
Lee
Z
,
Cheng
CT
,
Zhang
H
,
Subler
MA
,
Wu
J
,
Mukherjee
A
, et al
Role of LPA4/p2y9/GPR23 in negative regulation of cell motility
.
Mol Biol Cell
2008
;
19
:
5435
45
.
39.
Kuldo
JM
,
Westra
J
,
Asgeirsdottir
SA
,
Kok
RJ
,
Oosterhuis
K
,
Rots
MG
, et al
Differential effects of NF-κB and p38 MAPK inhibitors and combinations thereof on TNF-α- and IL-1β-induced proinflammatory status of endothelial cells in vitro
.
Am J Physiol Cell Physiol
2005
;
289
:
C1229
39
.
40.
Nichols
JW
,
Bae
YH
. 
EPR: Evidence and fallacy
.
J Control Release
2014
;
190
:
451
64
.
41.
Kishi
Y
,
Okudaira
S
,
Tanaka
M
,
Hama
K
,
Shida
D
,
Kitayama
J
, et al
Autotaxin is overexpressed in glioblastoma multiforme and contributes to cell motility of glioblastoma by converting lysophosphatidylcholine to lysophosphatidic acid
.
J Biol Chem
2006
;
281
:
17492
500
.
42.
Hoelzinger
DB
,
Nakada
M
,
Demuth
T
,
Rosensteel
T
,
Reavie
LB
,
Berens
ME
. 
Autotaxin: a secreted autocrine/paracrine factor that promotes glioma invasion
.
J Neurooncol
2008
;
86
:
297
309
.
43.
Schleicher
SM
,
Thotala
DK
,
Linkous
AG
,
Hu
R
,
Leahy
KM
,
Yazlovitskaya
EM
, et al
Autotaxin and LPA receptors represent potential molecular targets for the radiosensitization of murine glioma through effects on tumor vasculature
.
PLoS One
2011
;
6
:
e22182
.
44.
Bhave
SR
,
Dadey
DY
,
Karvas
RM
,
Ferraro
DJ
,
Kotipatruni
RP
,
Jaboin
JJ
, et al
Autotaxin inhibition with PF-8380 enhances the radiosensitivity of human and murine glioblastoma cell lines
.
Front Oncol
2013
;
3
:
236
.
45.
Brizel
DM
,
Dodge
RK
,
Clough
RW
,
Dewhirst
MW
. 
Oxygenation of head and neck cancer: changes during radiotherapy and impact on treatment outcome
.
Radiother Oncol
1999
;
53
:
113
7
.
46.
Nordsmark
M
,
Bentzen
SM
,
Rudat
V
,
Brizel
D
,
Lartigau
E
,
Stadler
P
, et al
Prognostic value of tumor oxygenation in 397 head and neck tumors after primary radiation therapy. An international multi-center study
.
Radiother Oncol
2005
;
77
:
18
24
.
47.
Cooper
PR
,
Ciambrone
GJ
,
Kliwinski
CM
,
Maze
E
,
Johnson
L
,
Li
Q
, et al
Efflux of monoclonal antibodies from rat brain by neonatal Fc receptor, FcRn
.
Brain Res
2013
;
1534
:
13
21
.
48.
Iwai
Y
,
Ishida
M
,
Tanaka
Y
,
Okazaki
T
,
Honjo
T
,
Minato
N
. 
Involvement of PD-L1 on tumor cells in the escape from host immune system and tumor immunotherapy by PD-L1 blockade
.
Proc Natl Acad Sci U S A
2002
;
99
:
12293
7
.
49.
Hinrichs
CS
. 
Molecular pathways: breaking the epithelial cancer barrier for chimeric antigen receptor and T-cell receptor gene therapy
.
Clin Cancer Res
2016
;
22
:
1559
64
.
50.
Hughes
DP
,
Marron
MB
,
Brindle
NP
. 
The antiinflammatory endothelial tyrosine kinase Tie2 interacts with a novel nuclear factor-kappaB inhibitor ABIN-2
.
Circ Res
2003
;
92
:
630
6
.
51.
DebRoy
A
,
Vogel
SM
,
Soni
D
,
Sundivakkam
PC
,
Malik
AB
,
Tiruppathi
C
. 
Cooperative signaling via transcription factors NF-kappaB and AP1/c-Fos mediates endothelial cell STIM1 expression and hyperpermeability in response to endotoxin
.
J Biol Chem
2014
;
289
:
24188
201
.
52.
Corbet C, Feron O. 
Tumour acidosis: from the passenger to the driver's seat
.
Nature reviews Cancer
2017
;
17
:
577
93
.
53.
Calcinotto
A
,
Filipazzi
P
,
Grioni
M
,
Iero
M
,
De Milito
A
,
Ricupito
A
, et al
Modulation of microenvironment acidity reverses anergy in human and murine tumor-infiltrating T lymphocytes
.
Cancer Res
2012
;
72
:
2746
56
.
54.
Mathios
D
,
Kim
JE
,
Mangraviti
A
,
Phallen
J
,
Park
CK
,
Jackson
CM
, et al
Anti-PD-1 antitumor immunity is enhanced by local and abrogated by systemic chemotherapy in GBM
.
Sci Transl Med
2016
;
8
:
370ra180
.