Plasminogen Kringle 5 Induces Apoptosis of Brain Microvessel Endothelial Cells: Sensitization by Radiation and Requirement for GRP78 and LRP1

Recombinant plasminogen kringle 5 (rK5) has been shown to induce apoptosis of dermal microvessel endothelial cells (MvEC) in a manner that requires glucose-regulated protein 78 (GRP78). As we are interested in antiangiogenic therapy for glioblastoma tumors, and the effectiveness of antiangiogenic therapy can be enhanced when combined with radiation, we investigated the proapoptotic effects of rK5 combined with radiation on brain MvEC. We found that rK5 treatment of brain MvEC induced apoptosis in a dose- and time-dependent manner and that prior irradiation significantly sensitized (500-fold) the cells to rK5-induced apoptosis. The rK5-induced apoptosis of both unirradiated and irradiated MvEC required expression of GRP78 and the low-density lipoprotein receptor-related protein 1 (LRP1), a scavenger receptor, based on down-regulation studies with small interfering RNA, and blocking studies with either a GRP78 antibody or a competitive inhibitor of ligand binding to LRP1. Furthermore, p38 mitogen-activated protein kinase was found to be a necessary downstream effector for rK5-induced apoptosis. These data suggest that irradiation sensitizes brain MvEC to the rK5-induced apoptosis and that this signal requires LRP1 internalization of GRP78 and the activation of p38 mitogen-activated protein kinase. Our findings suggest that prior irradiation would have a dose-sparing effect on rK5 antiangiogenic therapy for brain tumors and further suggest that the effects of rK5 would be tumor specific, as the expression of GRP78 protein is up-regulated on the brain MvEC in glioblastoma tumor biopsies compared with the normal brain. [Cancer Res 2009;69(13):5537–45]

Glioblastoma tumors have a dismal prognosis with a median survival of 12 to 15 months. As the tumors typically exhibit angiogenesis (1–3), antiangiogenic therapy may represent an effective therapeutic strategy. In this study, we investigated the proapoptotic effect of the recombinant form of the fifth kringle domain of plasminogen (rK5) on human brain microvessel endothelial cells (MvEC). Irradiation is known to promote the ability of other antiangiogenic agents to inhibit tumor growth (4–6) and irradiation is a standard initial therapy for patients with glioblastoma tumors (2); therefore, we also investigated the potential promotion of rK5-induced apoptosis of brain MvEC by irradiation.

It has been shown previously that rK5 induces changes indicative of apoptosis in nonbrain MvEC (7, 8) and inhibits proliferation of basic fibroblast growth factor-stimulated calf pulmonary arterial endothelial cells and bovine adrenal capillary endothelial cells (8, 9). The ability of rK5 to inhibit neovascularization has been shown directly in a rat model of hyperoxia-induced retinal neovascularization (10). Moreover, stable expression of K5 in U-87MG glioblastoma cells before their subcutaneous propagation in nude mice resulted in inhibition of angiogenesis and tumor growth (11) and inhibition of angiogenesis was seen when colorectal carcinoma cells stably expressing K5 were propagated subcutaneously in athymic nude mice (12).

The cell surface binding protein for rK5 on dermal MvEC is glucose-regulated protein 78 (GRP78; ref. 7). GRP78 is a member of the heat shock protein (HSP) 70 family, and its up-regulation is part of the general cellular defense mechanism of stressed cells called the unfolded protein response (reviewed in refs. 13, 14). The increased expression of members of the unfolded protein response in tumors suggests that they may be promising therapeutic targets (reviewed in ref. 15). GRP78 associates with the scavenger receptor low-density lipoprotein receptor-related protein 1 (LRP1) on the cell surface (16). LRP1 is known to signal upon associating with and internalizing its ligands (17–20). The biological consequences of its internalization are varied and depend on 1 of ∼30 different ligands it binds (reviewed in refs. 21–23) as well as its association with other cell surface proteins, the cell type, and the experimental conditions (16, 20, 24, 25). As the rK5 inhibition of retinal capillary endothelial cell proliferation was unaffected by RGD-containing peptides (10), it likely occurs in an integrin receptor-independent manner.

We found that rK5 induces apoptosis of brain MvEC and that prior irradiation significantly sensitizes the cells to rK5-induced apoptosis resulting in a dose-sparing effect. In both unirradiated and irradiated brain MvEC, the rK5-induced apoptosis requires expression of GRP78 and LRP1 as well as the activation of p38 mitogen-activated protein kinase (MAPK).

Materials. rK5 (ABT-828) expressed in yeast was provided by Abbott Laboratories; its native folded structure was verified by nuclear magnetic resonance as being comparable with K5 naturally derived from human plasminogen by elastase cleavage. PD98059 was purchased from Calbiochem, and SB202190 and SP600125 were from A.G. Scientific. Recombinant receptor-associated protein was purchased from Maine Biotechnology Services and dialyzed before use.

Cell culture. Primary human brain MvEC were purchased from Cell Systems, used at passages 2 to 8, and propagated as recommended in 70% CSC medium and 30% M199 medium, with 10% low-endotoxin fetal bovine serum (FBS; BD Biosciences). Unless otherwise indicated, cells were split the day before the experiment and then replated onto collagen-coated wells in M199 medium with 10% low-endotoxin FBS, 5 ng/mL basic fibroblast growth factor, and 10 ng/mL vascular endothelial growth factor (VEGF) 4 h before treatment with rK5. Cells were subconfluent at the time of treatment.

Analysis of cell surface expression. For analysis of the cell surface expression of GRP78 and LRP1, cells were labeled with biotin (26), lysed in 1% NP-40 lysis buffer [20 mmol/L Tris, 137 mmol/L NaCl, and 2 mmol/L EDTA (pH 8.0) with 10% glycerol and with protease inhibitors], immunoprecipitated with monoclonal antibody anti-LRP1 or goat anti-GRP78 IgG, the immunoprecipitates harvested by centrifugation, washed, subjected to SDS-PAGE, transferred to an Immobilon-P membrane, reacted with horseradish peroxidase-conjugated streptavidin, and developed using chemiluminescence (Amersham). Alternatively, cells were incubated with 10 μg/mL primary antibody (30 min, 4°C), washed, incubated with 10 μg/mL Alexa 488-conjugated secondary antibody (30 min, 4°C), washed, fixed, and subjected to fluorescence-activated cell sorting analysis using a FACScan instrument as described previously (27). Goat anti-GRP78 (directed toward the NH₂ and COOH termini) was purchased from Santa Cruz Biotechnology and monoclonal antibody anti-LRP1 was from Calbiochem.

Immunoblot analysis. Cells and tissue samples were lysed in radioimmunoprecipitation assay lysis buffer [0.05 mol/L HEPES (pH 7.4), 0.15 mol/L NaCl, 1% deoxycholate, 1% Triton X-100, and 0.1% SDS] with protease inhibitors as described previously (28). Equivalent amounts of lysate (typically 100-130 μg) were separated on SDS-PAGE, transferred to an Immobilon-P membrane, probed with primary antibody (4°C, overnight), washed, reacted with a horseradish peroxidase-conjugated secondary antibody (Bio-Rad), and developed using enhanced chemiluminescence (Amersham; ref. 28). Antibodies were purchased as follows: rabbit anti-p38 MAPK, anti-phospho-p38 MAPK (Y182), anti-HSP70, anti-c-Jun NH₂-terminal kinase (JNK), and anti-calreticulin (Santa Cruz Biotechnology); rabbit anti-cleaved caspase-3 or -7 (Calbiochem); rabbit anti-von Willebrand factor IgG (Chemicon); monoclonal antibody anti-actin (Sigma); mouse anti-caspase-3 (Cell Signaling); and monoclonal antibody anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH; Research Diagnostics).

Small interfering RNA down-regulation. SMARTpool small interfering RNA (siRNA) consisting of a pool of four SMARTselection-designed siRNA duplexes directed toward GRP78, HSP70, calreticulin, p38 MAPK, and extracellular signal-regulated kinase (ERK) 1 were purchased from Dharmacon, and siRNA directed toward LRP1 was purchased from Santa Cruz Biotechnology. HP-validated siRNA directed toward JNK2 was purchased from Qiagen. Two nonsense mutations were introduced into the JNK2 siRNA sequence as a custom control purchased from Qiagen. Transient transfections were carried out using the HiPerFect Transfection Reagent (Qiagen) according to the manufacturer's guidelines. The addition of liposome without siRNA was administered in control conditions.

Apoptosis assays. Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assays were carried out using the ApopTag peroxidase in situ apoptosis detection kit (Chemicon) as per the manufacturer's instructions and as described previously (28). Tumor necrosis factor-α (TNF-α) induction of apoptosis was used as a positive control (29). For determination of caspase-3/7 activity, a caspase-3/7 luminescent activity assay was done using the Caspase-Glo-3/7 Assay kit (Promega).

Analysis of human tissue samples. Tissue samples were obtained from the Cooperative Human Tissue Network of the National Cancer Institute and the University of Alabama at Birmingham Brain Tumor Bank in accordance with University Human Tissue Committee policies. Tumors were histologically graded according to the WHO classification of brain tumors (1). Frozen normal adult brain (cortex and white matter) and glioblastoma tumor samples obtained at autopsy within 18 h of death were homogenized in radioimmunoprecipitation assay lysis buffer with protease inhibitors as described (30) for Western blot analysis. For immunohistochemical analysis, frozen sections, as well as formalin-fixed and paraffin-embedded, normal brain and glioblastoma tumor samples were prepared from surgical biopsy samples and treated with blocking buffer to inhibit endogenous peroxidases and prevent nonspecific protein binding, reacted with the primary antibody in 5% bovine serum albumin/PBS/0.01% Tween 20 (4°C, 20 h), washed, reacted with a horseradish peroxidase-conjugated secondary antibody (22°C, 1 h) followed by the 3,3′-diaminobenzidine substrate (ScyTek; ref. 30), and counterstained with hematoxylin.

Statistics. After determining the data were normally distributed, a two-sample t test was used for data analysis and P < 0.05 was considered significant.

rK5 induces apoptosis of primary human brain MvEC. rK5 treatment has been shown to induce apoptosis of dermal MvEC (7). To determine the potential of rK5 treatment to induce apoptosis of brain MvEC, we treated primary human brain MvEC plated onto collagen type IV and grown in complete medium (10% FBS) in the presence of VEGF (10 ng/mL) and basic fibroblast growth factor (5 ng/mL) with rK5 at concentrations of 1 to 10,000 ng/mL for 17 h. We found that 5,000 ng/mL rK5 was necessary to induce a significant increase in the numbers of TUNEL-positive human brain MvEC (≈8% positive) at 17 h (Fig. 1A). Immunoblotting analysis as well as caspase-3/7 activity analysis confirmed that treatment of the cells with 5,000 ng/mL rK5 for 17 h resulted in a significant increase in cleaved caspase-7 and that treatment for periods greater than 17 h (25 and 40 h) did not result in a further increase in the amounts of cleaved caspase-7 (data not shown).

Irradiation sensitizes primary human brain MvEC to rK5-induced apoptosis. To evaluate the effect of prior irradiation on rK5-induced apoptosis of brain MvEC, we irradiated (2 or 5 Gy) primary human brain MvEC, allowed the cells to recover for 20 h, and replated them onto collagen-coated wells for 4 h before treatment with rK5 for 17 h. In the absence of treatment with rK5, only a low percentage of brain MvEC that were irradiated exhibited TUNEL positivity (Fig. 1B). Notably, the percentage of TUNEL-positive cells in the brain MvEC that had been irradiated before treatment with 10 ng/mL rK5 was similar to the percentage of TUNEL-positive cells in unirradiated MvEC that were treated with 5,000 ng/mL rK5 (Fig. 1B). Thus, the prior irradiation appeared to increase the sensitivity of the cells to rK5, the irradiation had a dose-sparing effect, but did not seem to increase the numbers of cells that were susceptible to rK5. Similarly, irradiation of the brain MvEC (2 or 5 Gy) followed by rK5 treatment (3-5,000 ng/mL, 17 h) resulted in a significant increase in the cleavage of caspase-3 or -7 that was maximal at 10 ng/mL rK5 (Fig. 1C; data not shown), confirming that irradiation plus rK5 induces apoptosis. Consistent with the results of the TUNEL assays, irradiation alone (5 Gy) did not induce a significant increase in cleaved caspase-3 (Supplementary Fig. S1A). The time course of rK5-induced caspase-7 cleavage in the irradiated brain MvEC was maximal at 17 h and detectable as early as 3 h post-rK5 treatment (Fig. 1D). Collectively, these data establish that prior irradiation of the brain MvEC resulted in a significant reduction in the dose of rK5 required for optimal induction of apoptosis.

GRP78 is necessary for rK5-induced apoptosis of unirradiated and irradiated primary human brain MvEC. Analysis of the expression of GRP78 on the cell surface of the MvEC used in the analyses of rK5-induced apoptosis was determined by cell surface biotinylation followed by detergent lysis, immunoprecipitation with anti-GRP78 antibody, and SDS-PAGE analysis. The data indicated that GRP78 was expressed on both unirradiated and irradiated human brain MvEC (Fig. 2A). Although irradiation has been reported to up-regulate GRP78 (31), we found that the surface expression was equivalent in the irradiated and unirradiated brain MvEC (Fig. 2A), suggesting that an up-regulation of GRP78 does not contribute to the enhanced sensitivity of the irradiated brain MvEC to low doses of rK5 in our experiments.

After further confirming the cell surface expression of GRP78 on the brain MvEC by fluorescence-activated cell sorting analysis (data not shown), we determined the requirement for GRP78 in rK5-induced apoptosis by down-regulating GRP78 with specific duplex siRNA. As a control, we down-regulated the endoplasmic reticulum chaperone protein calreticulin, which is expressed, in part, on the cell surface (19), or we down-regulated HSP70. Down-regulation of GRP78 and calreticulin was confirmed by Western blotting (Supplementary Fig. S1B) and we have established previously that HSP70 protein is down-regulated by >70% with 50 nmol/L siHSP70 treatment in the unirradiated human brain MvEC (data not shown). No adverse effects of the siRNAs on cell viability or morphology were detected over the time course of these experiments (data not shown). The down-regulation of GRP78 significantly blocked rK5-induced apoptosis (5,000 ng/mL; 17 h) of the unirradiated brain MvEC as determined using a TUNEL assay and blotting for cleaved caspase-3 (Fig. 2B), whereas the down-regulation of calreticulin or HSP70 had no effect.

Similar results were obtained on down-regulation of GRP78 in irradiated cells. Down-regulation of GRP78 (>70%) was achieved on treatment of the irradiated brain MvEC with 50 nmol/L siGRP78 (Supplementary Fig. S1C). Down-regulation of GRP78 in the irradiated brain MvEC significantly blocked rK5-induced apoptosis (10 ng/mL rK5; 17 h) as detected using the TUNEL assay and blotting for cleaved caspase-7 (Fig. 2C). The down-regulation of HSP70 had no effect. Further support for a role for GRP78 in rK5-induced apoptosis of the irradiated brain MvEC was obtained by the results of treatment with an anti-GRP78 antibody (10 μg/mL) that is directed toward the NH₂ terminus. This inhibited rK5-induced apoptosis (data not shown) as described previously for unirradiated dermal MvEC (7). These data indicate that GRP78 is necessary for rK5-induced apoptosis of both unirradiated and irradiated primary human brain MvEC.

LRP1 is necessary for rK5-induced apoptosis of unirradiated and irradiated human brain MvEC. GRP78 is known to associate with LRP1 on the cell surface (16); therefore, we examined the potential role of LRP1 in rK5-induced apoptosis of MvEC. The expression of LRP1 on the cell surface was verified using cell surface biotinylation followed by detergent lysis and immunoprecipitation with a monoclonal antibody anti-LRP1, which recognizes the 85-kDa light chain. SDS-PAGE analysis of the immunoprecipitates indicated that LRP1 is expressed on the surface of both unirradiated and irradiated MvEC and further indicated the levels of expression of LRP1 on the cell surface were not altered by irradiation (Supplementary Fig. S2A). To determine whether LRP1 is required for the rK5-induced apoptosis of the human brain MvEC, we pretreated the cells with receptor-associated protein, a competitive inhibitor of ligand binding to LRP1 (32). Prior addition of 60 nmol/L recombinant receptor-associated protein blocked rK5-induced apoptosis in the irradiated and unirradiated cells as detected using the TUNEL assay and cleavage of caspase-3 (Fig. 3A and B, respectively). These results were confirmed by down-regulation of LRP1 with pooled specific duplex siRNA. Down-regulation was confirmed by Western blotting of the irradiated and unirradiated cells (Supplementary Fig. S2B and C, respectively) and no effect of siLRP1 administration on cell viability or morphology was detected over the time course of these experiments. Down-regulation of LRP1 post-irradiation blocked rK5-induced apoptosis (10 ng/mL), whereas the down-regulation of HSP70 as a control had no effect (Fig. 3C). In the unirradiated cells, down-regulation of LRP1 significantly blocked the rK5-induced apoptosis (5,000 ng/mL), whereas the down-regulation of calreticulin had no effect (Fig. 3D). These data suggest that LRP1 internalization of GRP78 is likely necessary for rK5-induced apoptosis in both irradiated and the unirradiated brain MvEC.

p38 MAPK is necessary for rK5-induced apoptosis of the irradiated human brain MvEC. Stress is known to induce the activation of p38 MAPK (33) and p38 MAPK can promote a proapoptotic signal (34–36). We found a time-dependent increase in phosphorylated p38 MAPK in the irradiated cells treated with rK5 (Fig. 4A). Treatment with a small-molecule inhibitor of p38 MAPK (SB202190) at 2.5-fold the IC₅₀ (1.5 μmol/L) significantly blocked the proapoptotic effect of rK5 in the irradiated brain MvEC (Fig. 4B). In contrast, the proapoptotic effects of rK5 were not blocked by treatment with inhibitors of other kinases at 2.5-fold the IC₅₀: 5 μmol/L PD98059 (MEK inhibitor), SP600125 (JNK inhibitor), and FR180204 (ERK inhibitor; Fig. 4B). To further evaluate the requirement for p38 MAPK in the irradiated MvEC, we down-regulated p38 MAPK with pooled specific duplex siRNA and down-regulated ERK1 as a control (Supplementary Fig. S3A). TUNEL analysis indicated that the down-regulation of p38 MAPK in the irradiated brain MvEC blocked the proapoptotic effect of rK5 (Fig. 4C), whereas the down-regulation of ERK1 had no effect.

On analysis of the unirradiated cells, we also found that the p38 MAPK inhibitor SB202190 at 2.5-fold the IC₅₀ significantly blocked the proapoptotic effect of rK5 (Fig. 5A,, lanes 2 and 3), whereas the ERK inhibitor FR180204 did not (Fig. 5A,, lanes 2 and 4). Similarly, down-regulation of p38 MAPK in the unirradiated cells (Supplementary Fig. S3B) blocked the proapoptotic effect of rK5 as detected by TUNEL assay, whereas the down-regulation of ERK1 had no significant effect (Fig. 5B). These data suggest that p38 MAPK is a necessary downstream signaling effector in the proapoptotic effect of rK5 in irradiated and unirradiated human brain MvEC.

Although JNK has been reported to be involved in signaling that promotes apoptosis (34), we found that the JNK inhibitor, SP600125, at 2.5-fold the IC₅₀ in the irradiated cells (Fig. 4B) or the unirradiated cells (Fig. 5A) did not block the proapoptotic effect of rK5. However, a small increase (30%) in phosphorylation of the 54-kDa JNK2 isoform was detected in a time course analysis of JNK phosphorylation post-rK5 treatment of the unirradiated cells (data not shown). Down-regulation of JNK2 with 200 nmol/L specific duplex siJNK2 RNA (70% down-regulation of the 54-kDa isoform) followed by TUNEL assay confirmed that down-regulation of JNK2 in the unirradiated cells did not block the proapoptotic effect of rK5 (5,000 ng/mL; data not shown).

GRP78 expression is higher on brain MvEC in glioblastoma tumor samples compared with the normal brain. As our ultimate goal is to develop a novel therapeutic strategy for glioblastoma tumors, we evaluated the expression of GRP78 in frozen and paraffin sections of 14 glioblastoma tumor biopsy samples by immunohistochemistry. We found moderate (2+) expression of GRP78 in an estimated 30% of MvEC in the tumor portion of all glioblastoma samples (n = 14) and moderate (2+) expression in 30% of tumor cells (Fig. 6A). The staining localized to the cell membrane and the cytoplasm. The staining was specific for GRP78, as it was largely competed out by preincubation of the anti-GRP78 antibody with a 50-fold molar excess of GRP78 peptide (data not shown). Staining with rabbit IgG was used as a negative control and staining with anti-von Willebrand factor antibody was used as a positive control. Using this technique, GRP78 expression was below the limit of detection in endothelial and glial cells in the frozen and paraffin sections of normal brain (n = 13), although low (1+) expression of GRP78 was detected in neurons in the normal brain (data not shown). To confirm the up-regulation of GRP78 in the tumors compared with the normal brain, we immunoblotted lysates from nine normal brains and nine glioblastoma tumor samples with anti-GRP78 antibody followed by stripping and reprobing with antibodies directed toward GAPDH and HSP70. The immunoblotting data were quantitated by densitometric analysis and GRP78 expression was normalized to HSP70 and to GAPDH and plotted as bar graphs (Fig. 6B). We found on average a 2.0- to 2.5-fold increase in GRP78 expression relative to HSP70 or when normalized to GAPDH in the nine tumor samples compared with the nine normal brain samples (GRP78/HSP70 densitometry ratio: normal brain mean = 0.38 ± 0.06 and glioblastoma multiforme mean = 0.92 ± 0.2, P < 0.05; GRP78/GAPDH densitometry ratio: normal brain mean = 0.36 ± 0.05 and glioblastoma multiforme mean = 1.15 ± 0.3, P < 0.05). These data support our immunohistochemistry results, indicating that GRP78 expression is up-regulated in glioblastoma tumors in vivo.

In this report, we show that rK5 can induce apoptosis of brain MvEC and that irradiation significantly sensitizes primary human brain MvEC to the proapoptotic effect of rK5. We found that rK5 treatment of brain MvEC induces apoptosis, as measured by several different assays, when administered at 5,000 ng/mL (≈500 nmol/L). This is similar to the dosage of rK5 required to induce apoptosis of dermal MvEC (7). Importantly, we found that prior irradiation significantly sensitized the brain MvEC to rK5-induced apoptosis, as a 500-fold lower dose of rK5 (10 ng/mL) induced a similar percentage of TUNEL-positive cells and cleavage of caspase-3 or -7.

Few studies have focused on irradiation as a potential sensitizing agent for antiangiogenic therapy. Our studies are consistent with the recent report of Jin and colleagues (37) that rK5 combined with irradiation enhanced the antiangiogenic effect of rK5 in a mouse tumor model of Lewis lung carcinoma cells propagated subcutaneously. Under our experimental conditions (10% FBS with 10 ng/mL VEGF and 5 ng/mL basic fibroblast growth factor), irradiation alone did not induce significant cell death or caspase-3/7 cleavage. Other investigators have reported a proapoptotic effect of irradiation (2 or 5 Gy) alone on dermal MvEC that were propagated and radiated in reduced serum (2% FBS), and this was attributed to reduced levels of the antiapoptotic protein Bcl2 (38, 39). We did not find an altered level of Bcl2 protein post-irradiation of the MvEC at two time points in our experimental conditions.⁵

This may be due to the higher levels of FBS and added VEGF, as the level of Bcl2 expression is known to be regulated by VEGF (39) or it may reflect the use of brain MvEC rather than dermal MvEC. rK5 has been reported to induce autophagy of nonbrain endothelial cells in different experimental conditions (40); therefore, it is possible that rK5-induced autophagy contributes to the apoptosis we have observed in the brain MvEC.

Our finding that GRP78 is required for rK5-induced apoptosis of human brain MvEC is consistent with the report of the requirement for GRP78 in the proapoptotic effect of rK5 on dermal MvEC, and of a direct interaction of rK5 with recombinant GRP78 (7). GRP78 is an endoplasmic reticulum chaperone protein that is expressed on the cell surface in stress conditions and in tumors (14, 15); however, we found no change in the cell surface expression post-irradiation in the time course and conditions of our experiments. This suggests that an increase in GRP78 expression post-irradiation is not the mechanism by which irradiation sensitizes the cells to rK5-induced apoptosis. The ligand binding to cell surface GRP78 appears to determine the signal generated. We and others (7) detect a proapoptotic signal when rK5 binds to cell surface GRP78 on MvEC, whereas α₂-macroglobulin binding to cell surface GRP78 on human prostate cancer cells and macrophages initiates a pro-proliferation signal (41). The voltage-dependent ion channel also has been reported to be a receptor for rK5 (42).

We found that LRP1 is necessary for the rK5-generated proapoptotic signal. This suggests that LRP1 internalization of GRP78 is likely necessary for the proapoptotic effect of rK5 in the brain MvEC. The signal generated upon LRP1 binding and internalizing its ligand is dependent on the cell and the environmental context (21). For example, a promigratory signal is generated when LRP1 binds and internalizes the complex of cell surface calreticulin and thrombospondin-1 in coronary artery endothelial cells (19, 20). Differential LRP1 signaling is thought to be due to different adaptor molecules that bind to the phosphorylated tyrosine residue(s) in the LRP1 cytoplasmic tail and thereby activate different downstream effectors (21, 23, 43).

We found that p38 MAPK was activated with rK5 treatment post-irradiation and that p38 MAPK was necessary for rK5-induced apoptosis of the irradiated and unirradiated brain MvEC. In contrast, neither ERK nor JNK were necessary for the rK5-induced apoptosis of the human brain MvEC. Notably, the LRP1 family member LRP8, expressed on platelets, is known to activate p38 MAPK upon binding its ligand, the low-density lipoprotein, suggesting that the LRP family of proteins is capable of signaling to p38 MAPK (44).

In support of the potential clinical relevance of our findings, we show that the expression of GRP78 is increased on the MvEC in glioblastoma tumor samples compared with the normal brain. GRP78 expression was also increased on the tumor cells in the glioblastoma biopsy samples compared with glial cells in the normal brain. Our results are consistent with those in the literature, indicating that GRP78 expression is up-regulated in tumors (reviewed in refs. 15, 45, 46). Two conditions frequently found in malignant tumors, hypoxia and hypoglycemia, are known to up-regulate GRP78 expression (14, 47). Recently, GRP78 expression has been associated with chemoresistance, making it a promising target for cancer therapy (reviewed in refs. 15, 45, 48).

In summary, our data indicate that irradiation sensitizes primary human brain MvEC to the apoptosis-inducing effect of rK5 and that this proapoptotic signal requires LRP1 internalization of GRP78 and p38 MAPK activity. As the cell surface binding partner for rK5 (GRP78) and its scavenger receptor partner (LRP1; ref. 49) are both expressed on MvEC in glioblastoma tumor biopsy samples, these data suggest that rK5 treatment post-irradiation should be considered in the design of new therapies for patients with glioblastoma tumors.

D.J. Davidson and J. Henkin are employees of Abbott Laboratories and own stock in Abbott Laboratories. The other authors disclosed no potential conflicts of interest.

Grant support: NIH, National Cancer Institute grants CA109748 and CA127620 (C.L. Gladson) and Abbott Laboratories.

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.

We thank Rhonda Carr for assistance in preparing this article and Dr. Fiona Hunter for critical review.