Targetable T-type Calcium Channels Drive Glioblastoma

Glioblastoma (GBM) is the most common primary malignant brain tumor. GBM is associated with a dismal prognosis and an average life expectancy of approximately 15 months despite optimal therapy consisting of surgery, chemotherapy, and radiation (1). Lack of therapeutic success is attributed to various factors, including rapid tumor cell infiltration of the brain, inter- and intratumoral heterogeneity, limited diffusion of therapeutic drugs across the blood–brain barrier, and brain/tumor parenchyma as well as to the presence within the tumor of GBM stem cells (GSC) that are resistant to radio- and chemotherapy and that are capable of tumor generation and unlimited self-renewal (2–5). Overcoming GSC resistance to existing therapies or induction of GSC differentiation could greatly enhance therapeutic outcome.

Calcium signaling plays a ubiquitous role in many cellular regulatory processes (6–9), including proliferation, apoptosis, and gene transcription (10, 11). In addition, alteration of free cytosolic calcium results in the transcription of immediate early genes (12, 13), such as c-fos, c-jun, the cyclic AMP response element, and the serum response element. Expression of these genes triggers entry into the cell cycle through expression of cyclins and cyclin-dependent kinases (14–17). Calcium influx is mediated by voltage-gated Ca++ (Cav) channels (18), of which there are five types: L-type, P-type, N-type, R-type, and T-type. Aberrant expression and activity of T-type calcium channels (Cav3.2) have been implicated in cancer (19) through their role in the regulation of cell-cycle progression. In line with this, an increase in intracellular calcium, regulated by Cav3.2 expression, has been shown to regulate GBM cell proliferation (20, 21). Furthermore, calcium entry participates in the complex network responsible for mouse embryonic stem cell self-renewal (22). A very recent study, published while this manuscript was in revision, showed that membrane-depolarizing channel blockers induce selective glioma cell death by impairing nutrient transport and unfolded protein/amino acid responses (23).

This work aimed at studying the expression, functions, mechanisms of action, and therapeutic targeting of calcium channels with mibefradil and shRNA in combination with cytotoxic therapies. Mibefradil (C29H38FN3O3; Supplementary Fig. S1) is a T-type FDA-approved calcium channel blocker previously marketed by Roche as Posicor for the treatment of hypertension. Our data demonstrate that Cav3.2 is highly expressed in human GBM and GSCs and that expression correlates with patient survival. Inhibition of Cav3.2 suppressed both GSC growth and stemness and in vivo xenograft growth and sensitized GSCs to chemotherapy. Mechanistically, mibefradil altered multiple cancer-regulatory pathways as well as the expression of several oncogenes and tumor suppressors in GSCs. Mibefradil inhibited HIF1α/HIF2 expression under hypoxic conditions.

Human GBM cell lines U87, A172, U373, T98G, and SNB19 were obtained from ATCC within the past 4 years and used at less than 20 passages. U1242, U251, and SF767 were kind gifts from Drs. Isa Hussaini and Dr. Benjamin Purow (University of Virginia, Charlottesville, VA), and Dr. Russel Pieper (University of California, San Francisco, San Francisco, CA), respectively. They were obtained within the past 10 years and used at less than 35 passages. Primary GBM cells (GBM-6 and GMB-10), a gift from Dr. Jann Sarkaria (Mayo Clinic, Rochester, MN), were isolated from patients who underwent surgery at the Mayo Clinic and were used at less than 10 passages (24). GSCs XO-1, XO-2, XO-3, XO-4, XO-8, and XO-9, kindly offered by Dr. Deric Park, were isolated from GBM specimens obtained from patients undergoing surgery at the University of Pittsburgh (Pittsburgh, PA) and were used at less than 8 passages. GSCs 206, 827, and 578 (used at less than 12 passages) were isolated from patient surgical specimens. GSCs were characterized for in vivo tumorigenesis, pluripotency, self-renewal, stem cell markers, and neurosphere formation (25). Matched CD133-positive and CD133-negative cells isolated from GBM surgical specimens (used at one passage) were kindly provided by Dr. Rainer Glass (University Clinics Munich, Germany). All cell lines underwent testing for species and mycoplasma infection using Mycoplasma Detection Kit-QuickTest (Biotool). GBM surgical specimens were obtained from the University of Virginia Brain Tumor Bank according to procedures that were approved by the Review Board of the University of Virginia. All tumors were characterized by an experienced neuropathologist.

Immunoblotting was performed as described previously (26). Antibodies used were Cav3.2, Nestin, Tuj-1, and HIF1α (Santa Cruz Biotechnology), p27, GFAP, BMI1, MAP2, Bax, Sox2, mTOR, HIF2, and PARP (Cell Signaling Technology), β-actin and GAPDH (Santa Cruz Biotechnology).

Total RNA was extracted from GSCs using RNeasy Extraction Kit (Qiagen). cDNA was synthesized using the iScript cDNA Synthesis Kit (Bio-Rad), and quantitative PCR analysis was performed using the CFX Connect Real-Time System (Bio-Rad). GAPDH was used as an endogenous control. Primer sequences are listed in the Supplementary Methods.

Lentiviral control vectors (sh-control) or vectors encoding Cav3.2 shRNA (sh-Cav3.2; pooled; Santa Cruz Biotechnology) or two additional sh-Cav3.2 (Applied Biological Materials) were generated. GSCs were seeded in poly-l-ornithine precoated plates and infected with the lentiviruses.

Cell death was assessed by Trypan blue assay. Cell proliferation was assessed by cell counting for 5 days as described previously (27). The Alamar Blue assay was utilized according to the manufacturer's instructions. All experiments were performed three times.

Differential expression of Cav3.2 was analyzed in GBM (n = 607) and normal unmatched brain samples (n = 11) from The Cancer Genome Atlas (TCGA) data. The effect of Cav3.2 expression on patient survival was assessed using the cBioportal website (www.cBioportal.org). Differential expression of Cav3.2 and its correlation with survival was also assessed in the REMBRANDT database, which used an Affymetrix HG U133 v2.0 Plus platform to analyze 178 GBM samples.

GSCs were incubated in a hypoxic incubator (Thermo Fisher Scientific) with a 94:5:1 mixture of N₂/CO₂/O₂. The cells were treated with mibefradil (5 μmol/L) or vehicle (water) in normoxia (5% O₂) or hypoxia (1% O₂) for 24 to 48 hours. The cells were subsequently subjected to immunoblotting for hypoxia-inducible factor (HIF) proteins.

Proteomic screening was performed by reverse-phase protein array (RPPA) as described previously (28–30). Protein analytes were chosen for analysis based on their previously described involvement in key aspects of tumor biology. All antibodies were validated for single-band specificity and for ligand induction (phospho-specific antibodies) by immunoblotting prior to use on the arrays as described previously (28–30). Additional details are in the Supplementary Methods.

GSCs were treated with mibefradil (5 μmol/L) or vehicle control for 24 hours at 37°C prior to RNA extraction. RNA sequencing (RNA-seq) was performed by Hudson Alpha as described in the Supplementary Methods.

Functional rescue experiments were performed to determine whether molecules regulated by mibefradil from proteomic and transcriptomic screenings mediate the effects of mibefradil on GSCs. The experimental details are described in the Supplementary Methods.

The therapeutic effects of inhibiting Cav3.2 with mibefradil were assessed using an orthotopic GSC-based xenograft mouse model. GSC 827 cells (3 × 10⁵) were stereotactically implanted into the corpus striatum of immunodeficient mice (n = 10/treatment group). Six days after implantation, the animals were treated with control (H₂O) or mibefradil (24 mg/kg) by oral gavage every 6 hours and/or temozolomide (100 mg/kg body weight) intraperitoneally once a day for 4 days. Treatment stopped for 7 days. Mice were treated for two cycles. Tumor volumes were visualized and quantified by MRI, and animal survival was determined.

IHC was used to assess the expressions of Ki67, cleaved caspase-3, SOX2, and GFAP in brain tumor xenograft sections. IHC was performed as described previously (31) using antibodies against Ki67, cleaved caspase-3, SOX2, or GFAP (Cell Signaling Technology), and secondary antibodies were conjugated with the fluorescent dyes Alexa 488 or Alexa 555 (Thermo Fisher Scientific).

To evaluate the statistical significance of in vivo animal experiments, we used both two-sample t test and nonparametric Wilcoxon rank sum test. The continuous variable RPPA data generated were subjected to both unsupervised and supervised statistical analyses. Statistical analyses were performed on final microarray intensity values obtained using R version 2.9.2 software (The R Foundation for Statistical Computing). If the distribution of variables for the analyzed groups was normal, a two-sample t test was performed. If the variances of two groups were equal, two-sample t test with a pooled variance procedure was used to compare the means of intensity between two groups. Otherwise, two-sample t test without a pooled variance procedure was adopted. For nonnormally distributed variables, the Wilcoxon rank sum test was used. Significance levels were set at P < 0.05.

For RNA-seq, raw FASTQ sequencing reads were chastity filtered to remove clusters having outlying intensity corresponding to bases other than the called base. Filtered reads were assessed for quality using FastQC. Reads were splice-aware aligned to the Ensembl GRCh38 genome using STAR (32), and reads overlapping GRCh38.82 gene regions were counted using featureCounts (33). The DESeq2 Bioconductor package (34) in the R statistical computing environment was used for normalizing count data, performing exploratory data analysis, estimating dispersion, and fitting a negative binomial model for each gene comparing the expression for each cell line comparison. After obtaining a list of differentially expressed genes, log fold changes, and P values, Benjamini–Hochberg FDR was used to correct P values for multiple testing.

To determine whether Cav3.2 is deregulated in GSCs and human tumor specimens, we measured Cav3.2 protein expression by immunoblotting and compared it with the expression in commonly used established cell lines and normal brain. The data show that Cav 3.2 is highly expressed in all GSCs and in some GBM cell lines (Fig. 1A). To determine whether the expression of Cav3.2 is enriched in the stem cell fraction, we assessed the level of Cav3.2 in CD133-positive and CD133-negative cells derived from the same tumors (elevated SOX2 expression and reduced GFAP expression in CD133-positive cells compared with CD133-negative cells confirmed the stem cell identity of these cells). Cav3.2 expression was several fold higher expressed in the CD133-positive fraction than in the corresponding CD133-negative fraction. This suggests a link between elevated Cav3.2 expression and the stem cell state (Fig. 1B). The majority of GBM tumor samples expressed high levels of Cav3.2, while expression in normal brain was uniformly low (Fig. 1C).

We also analyzed Cav3.2 expression in the TCGA and REMBRANDT databases. The TCGA analysis showed that 11% of GBM patients display either amplification, mutations or mRNA upregulation of Cav3.2. Patients with an alteration of Cav3.2 demonstrated a trend toward worse survival than patients with normal Cav3.2 [Cav3.2 mRNA expression z-scores (RNA-Seq V2 RSEM), with a z-score threshold 1.0; Fig. 1D, top]. The REMBRANDT analysis also showed worse survival with high expression of Cav3.2 at thresholds of 25 percentile (Fig. 1D, bottom).

The above data demonstrate that Cav3.2 is highly expressed in a subset of GBM tumors and GSCs and that high expression may correlate with poor prognosis.

To assess the effects of Cav3.2 inhibition on GSC function, we first tested the effects of mibefradil on cell growth, proliferation, and death, also in combination with temozolomide. GSCs (827, 206, and 578) were treated with mibefradil (2.5–5 μmol/L) and/or with temozolomide (400 nmol/L) and analyzed for growth by Alamar blue assay. The results showed that mibefradil significantly inhibited cell growth and enhanced the inhibition of GSC growth by temozolomide (Fig. 2A). To determine the effects of mibefradil on GSC proliferation, cells were treated with mibefradil (5 μmol/L) and/or temozolomide (400 nmol/L) and counted for 5 days. Mibefradil significantly inhibited cell proliferation and enhanced the inhibition of GSC proliferation by temozolomide (Fig. 2B). To determine whether mibefradil affects cell death, we treated cells as above and performed a Trypan blue assay. The results showed that mibefradil significantly enhances cell death also in combination with temozolomide (Fig. 2C). The above data show that mibefradil exhibits anticancer stem cell effects also in combination with temozolomide.

Cav3.2 has been associated with differentiation of mouse embryonic stem cells (35). On the basis of that, we hypothesized that Cav3.2 might also regulate GSCs. To test this hypothesis, we determined the effects of mibefradil on the in situ expression of stemness and differentiation markers in GSCs using immunocytochemistry for the stemness and differentiation markers Sox2, GFAP, and Tuj1. Mibefradil treatment increased the differentiation markers, GFAP (astrocytes) and Tuj1 (neurons; Fig. 2D) and inhibited the expression of stem cell marker Sox2 (Fig. 2D). We confirmed the downregulation of stemness markers Nestin, Bmi1, and Sox2 as well as the upregulation of differentiation markers GFAP and MAP2 through immunoblotting (Fig. 2E). We also found, through RPPA, that mibefradil treatment decreased the protein level of CD133 while simultaneously increasing the level of GFAP (Fig. 4D). The above data indicate that mibefradil induces GSC differentiation.

To confirm that the cell death and growth suppression induced by mibefradil was attributed to inhibition of Cav3.2, we silenced Cav3.2 with shRNA and analyzed the effect on cell proliferation and death. Cells were infected with lentivirus encoding either sh-control or sh-Cav3.2 for 48 hours. Cav3.2 knockdown by sh-Cav3.2 was verified by immunoblotting (Fig. 3A). The cells were counted at 3 and 5 days after lentivirus infection. Cav3.2 silencing significantly inhibited cell proliferation of GSCs (Fig. 3B). To assess whether Cav3.2 silencing affected cell death, cells were infected as above and a Trypan blue assay was performed. The data showed that silencing of Cav3.2 significantly enhanced cell death (Fig. 3C). To exclude off-target effects of sh-Cav3.2, we used two additional sh-Cav3.2 and found similar effects on cell proliferation and death (Supplementary Fig. S3). The above data show that silencing of Cav3.2 expression leads to comparable anticancer effects on GSCs as mibefradil treatment.

Hypoxia and HIF levels have been associated with GSC maintenance and resistance to therapy (36, 37). We hypothesized that Cav3.2 inhibition might affect hypoxia-induced HIF1 and HIF2 in GSCs. Thus, we determined Cav3.2 expression in a hypoxic environment and the effect of mibefradil on HIF1α and HIF2 expression. GSCs were grown in 1% oxygen for 24 hours, and the level of Cav3.2 was determined by immunoblotting. Hypoxia induced Cav3.2 expression (Supplementary Fig. S2A). HIF1α and HIF2 were increased in GSCs under hypoxic conditions compared with normoxic conditions (Supplementary Fig. S2B and S2C). HIF1α and HIF2 expressions were significantly suppressed by mibefradil treatment (Supplementary Fig. S2B and S2C), suggesting that mibefradil might overcome GBM resistance to chemotherapy partially through inhibition of HIF1α and HIF2 in GSCs.

To investigate the mechanism of action of Cav3.2 blockade with mibefradil in GSCs, we performed proteomic screenings using RPPAs in GSCs treated with or without mibefradil for 1 or 24 hours. The arrays measured the expressions and activation of >300 proteins associated with cancer. Mibefradil induced dramatic and significant changes in GSC signaling including inhibition of the AKT/mTOR prosurvival pathway as evidenced by decreased activation of AKT, mTOR, and 4EBP1 and upregulation of LKB1 and Tuberin/TSC2 phosphorylation (Fig. 4A). Mibefradil also induced signaling changes associated with the induction of cell-cycle arrest and apoptosis, including inhibition of survivin and activation of BAX, caspase-9, PARP, p27, and Rb (Fig. 4B and C). Mibefradil also affected molecules associated with DNA damage repair and autophagy (activation of ATM and LC3B; Fig. 4C). In addition, mibefradil inhibited the expression of the stemness marker CD133 and increased the expression of the astrocytic marker GFAP (Fig. 4D). Select random RPPA results (changes in p27, BAX, and cleaved PARP) were verified by immunoblotting (Fig. 4E). The above data show that mibefradil-mediated Cav3.2 inhibition inhibits GSC malignancy by acting on multiple key cancer-regulatory pathways.

To determine whether BAX and p27 mediate the proapoptotic effect of mibefradil, we assessed the effects of silencing p27 and BAX on mibefradil-induced cell death. We transfected GSC cells with either si-control, si-BAX, or si-p27 before treating them with mibefradil for 48 hours. Inhibition of either BAX or p27 abrogated mibefradil-induced cell death (Fig. 5A). Silencing of BAX and p27 was confirmed by immunoblotting (Fig. 5B). To deduce the role of mTOR in mibefradil-mediated suppression of proliferation and cell death, we overexpressed mTOR prior to treating GSCs with mibefradil for 48 hours. Overexpression of mTOR partially mitigated mibefradil-induced cell death and growth suppression (Fig. 5C). Overexpression of mTOR was confirmed by immunoblotting (Fig. 5D). The above data show that the anticancer effects of Cav3.2 inhibition are partly mediated by the induction of apoptosis and inhibition of cell-cycle progression via regulation of mTOR, BAX, and p27.

As calcium is required for gene transcription (38, 39), we hypothesized that mibefradil might alter gene expression in GSCs. To test this hypothesis, we performed RNA deep sequencing (RNA-seq) on GSCs treated with vehicle control or mibefradil for 24 hours. Mibefradil significantly altered the expression of many genes (Fig. 6A and B). The RNA-seq data were partially verified by qRT-PCR (Fig. 6C). Notably, mibefradil induced the expression of several tumors suppressors, such as TNFRSF14 and HSD17B14, while simultaneously downregulating the expression of several oncogenes such as PDGFA, PDGFB, TGFB1, METTL7B, EGR3, and TNFRSF12A in GSCs. Interestingly, mibefradil-induced inhibition of METTL7B, a novel oncogene found overexpressed in primary stem-like GBM (40), and EGR3 (41), both of which may partially account for mibefradil-induced differentiation as well as decreased GSC proliferation and survival. RNA-seq data were deposited at the GEO database (accession # GSE95106).

We tested the effects of mibefradil alone and in combination with temozolomide on the growth of established GSC xenografts and animal survival. GSC 827 cells (3 × 10⁵) were stereotactically implanted in the striata of immunodeficient mice (n = 10). Six days after implantation, mibefradil (24 mg/kg bodyweight) was administered per oral gavage every 6 hours for 4 days. Temozolomide (100 mg/kg bodyweight) was concurrently administered intraperitoneally once a day for 4 days. The above treatment plan was repeated 7 days later. MRI scans were performed 21 days after surgery, and tumor sizes were measured. Animal survival was also assessed. The data show that mibefradil alone significantly inhibited tumor growth. This inhibition was of similar magnitude as temozolomide alone. Combined treatment with mibefradil and temozolomide inhibited tumor growth in an additive fashion (Fig. 7A and B). Similarly, both mibefradil and temozolomide alone significantly prolonged animal survival with the combination of both drugs further prolonging animal survival (Fig. 7C). IHC analysis of the xenografts revealed downregulation of the proliferation marker Ki67 and upregulation of the apoptosis regulator cleaved caspase-3 in mibefradil-treated xenografts. The expression of stem cell marker SOX2 was decreased, whereas the astrocyte marker GFAP was elevated in mibefradil-treated tumors (Fig. 7D). These data suggest that mibefradil is a promising drug for GBM therapy.

Changes in intracellular Ca²⁺ levels can modulate signaling pathways and gene transcription that control a broad range of cellular events, including those important to tumorigenesis and cancer progression (42). Calcium channels are differentially expressed in malignant cells and alterations in their activity highlight the possible use of calcium channels as targets for new therapies (43, 44). Mibefradil is an orally bioavailable blocker of T- and L-type calcium channels, marketed by Roche as Posicor, for the treatment of hypertension. Although blood pressure–lowering effects are mediated by L channel blockade, mibefradil has end organ-protective effects that are mediated by T-type calcium channel inhibition, resulting in reduced proliferation (45). Mibefradil was previously tested for effects upon GBM cell growth (46), but its mechanism of action in GBM and GSCs is not well understood. In this study, we investigated the expression, function, mechanisms of action, and therapeutic value of targeting of T-type calcium channels in GBM and GSCs.

We showed that expression of Cav3.2 is increased in GBM tumors and GSCs. Increased expression of Cav3.2 in GBM correlated with poor prognosis, suggesting the possibility of targeting Cav3.2 for GBM therapy and improved patient survival. We established that mibefradil sensitized GSCs to temozolomide treatment, a key chemotherapeutic agent used in the treatment of GBM. GSCs partly mediate resistance to chemo- and radiotherapy and have become vital targets for reversal of chemoresistance. Notably, after surgical intervention and chemotherapy, resistant GSCs survive and are able to initiate and maintain malignant growth of GBM (4). We showed that mibefradil induced the differentiation of GSCs, as evidenced by the downregulation of GSC stemness markers, CD133, Nestin, Bmi1, and Sox2, and upregulation of differentiation markers GFAP, Tuj1, and MAP2. Although Cav3.2 inhibition strongly impairs GSC malignant parameters, it likely also affects differentiated bulk GBM cells as has been shown for the U87 cell line (21).

To establish whether the anticancer properties of mibefradil are attributed to inhibition of Cav3.2, we used shRNA to specifically silence Cav3.2. Silencing of Cav3.2 displayed similar anticancer effects as mibefradil treatment. This indicates that inhibition of Cav3.2 is the primary mechanism through which mibefradil exerts its known anticancer effects. We also discovered that Cav3.2 is upregulated in GSCs under hypoxic conditions and, importantly, mibefradil downregulated HIF1α and HIF2. Hypoxia strongly correlates with poor patient survival, therapeutic resistance, and an aggressive tumor phenotype. It was also previously demonstrated that GSCs are critically dependent on the HIFs for survival, self-renewal, and growth (47, 48). This suggests that mibefradil can inhibit GSC malignant parameters by reducing hypoxic pressure and inhibiting HIFs.

We investigated the mechanism of action of mibefradil on GSCs. Using proteomic screening, immunoblot validation, and functional rescue experiments, we found that cell cycle and apoptosis signaling pathways are significantly activated by mibefradil in GSCs, including decreased phosphorylation of AKT, mTOR, and 4EBP1 and upregulation of p27 KIP1, BAX, FOXO1, and cleaved PARP, leading to decreased tumor cell proliferation through downregulation of key survival pathways and cell-cycle arrest, respectively. Furthermore, silencing of p27 or Bax, both of which are upregulated in response to mibefradil treatment, abrogated drug-induced cell death, suggesting a role for p27 and BAX in mediating mibefradil toxicity. Also, overexpression of mTOR, which is downregulated in response to mibefradil treatment, inhibited drug-induced cell death.

To understand the effect of mibefradil on the transcriptome, we performed RNA-seq. We found that mibefradil treatment resulted in a decrease in expression of several oncogenes, such as PDGFA, PDGFB, TGFB1, METTL7B, EGR3, and TNFRSF12A, and an increase in the expression of numerous tumor suppressors, including TNFRSF14 and HSD17B14, confirmed by qRT-PCR. These data provide insights into the molecular basis of responsiveness to Cav3.2 inhibition in cancer cells, suggesting an important role for mibefradil in the regulation of genes implicated in cancer. In addition to the changes in protein function and gene transcription described above, blocking Cav3.2 might also influence GSC function by affecting resting membrane potential. In fact, recently published research has shown that brain tumor stem cells have elevated resting membrane potential that is regulated by connexins and EAG2 potassium channels to influence these cells' malignant functions (49, 50).

Mibefradil significantly reduced in vivo tumor growth and prolonged animal survival. In addition, combination of mibefradil with temozolomide further enhanced therapeutic effect and survival. These data have promising translational potential. Mibefradil, an FDA-approved low-toxicity therapeutic, could easily translate into clinical trials. Cavion LLC undertook a trial to assess the safety of mibefradil in 30 healthy patients, followed by a second trial (NCT01480050), in conjunction with NCI (Rockville, MD), to assess the efficiency and optimal dosage of mibefradil sequentially administered in combination with temozolomide in patients with recurrent GBM. A third trial (NCT02202993) is currently enrolling recurrent GBM patients to study the effect and safety of mibefradil combined with hypofractionated radiotherapy. This emphasizes the relevance of our study in establishing the molecular and functional basis of Cav3.2 targeting for GBM therapy. Our data also suggest that mibefradil is likely to achieve greater clinical benefits in patients in combination with temozolomide.

Altogether, our data show novel roles and mechanisms of action of T-type calcium channel Cav3.2 in GSCs and GBM (Supplementary Fig. S4) and support the use of mibefradil in combination with temozolomide for GBM therapy.

D. Schiff is a consultant/advisory board member for Cavion. E.F. Petricoin is the chief science officer/consultant at Perthera, Inc., is a consultant at Ceres Nanosciences, Inc., has ownership interest (including patents) in Ceres Nanosciences, Inc. and Perthera, Inc., is a consultant/advisory board member for Ceres Nanosciences, Inc. and Perthera, Inc. L.S. Gray has ownership interest (including patents) in Cavion, LLC. No potential conflicts of interest were disclosed by the other authors.

Conception and design: Y. Zhang, D. Schiff, L.S. Gray, R. Abounader

Development of methodology: Y. Zhang, N. Cruickshanks, F. Yuan, L.S. Gray, R. Abounader

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Zhang, N. Cruickshanks, F. Yuan, M. Pahuski, I. Gallagher, S. Hatef, J. Lee, E.F. Petricoin

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Zhang, N. Cruickshanks, F. Yuan, M. Pahuski, A.F. Koeppel, S.D. Turner, E.F. Petricoin

Writing, review, and/or revision of the manuscript: Y. Zhang, N. Cruickshanks, F. Yuan, J. Wulfkuhle, E.E. Bar, D. Schiff, S.D. Turner, R. Abounader

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Zhang, N. Cruickshanks, B. Wang, C. Papanicolas

Study supervision: Y. Zhang, R. Abounader

Other (consultant): E.E. Bar

This work was supported by grants from the Commonwealth Research Commercialization Fund of Virginia (CRCF), the Virginia Biosciences Health Research Corporation (VBHRC), and NIH R01 grants NS045209 to R. Abounader and NIHR01 CA134843 to R. Abounader.

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.