Survivin Monoclonal Antibodies Detect Survivin Cell Surface Expression and Inhibit Tumor Growth In Vivo

Survivin (BIRC5, API4) is a protein that belongs to a family of apoptosis inhibitors and acts in concert with the mitotic spindle apparatus to regulate cell division and chromosomal segregation (1). It is expressed during the G₂–M phase of the cell cycle in association with the spindle microtubule organizing center (2). Survivin functions in critical roles at other cellular loci to regulate the cell cycle and to inhibit apoptotic cell death. It is frequently expressed by cancer cells of many different types, but its expression is uncommon in normal adult fully differentiated tissues (3).

Expression of survivin in gliomas and other cancers is associated with a poor prognosis (4–7) and refractoriness to chemotherapy (8). Epitopes of survivin are presented by MHC I complexes on the surface of tumor cells making them immunologically targetable by specific cytotoxic CD8⁺ T cells. Patients with cancer have anti-survivin antibodies (9) and survivin-specific T cells (10) in their peripheral blood. Therefore, survivin is immunogenic and its immunogenicity may be enhanced with the aid of survivin mimics (11, 12). Several studies have looked at active, specific vaccination against survivin to treat various cancers (13–15), including malignant gliomas (16).

Although it was initially defined as an intracellular molecule, more recently, survivin has been identified in association with exosomes produced by cervical and prostatic carcinoma cells (17, 18). While the intracellular functions of survivin have been characterized extensively, less well appreciated is its presence and function on the outer membrane of tumor cells where it is potentially targetable using antibody-mediated immunotherapeutic strategies. Here we identify survivin expression on the surface of a wide range of cancer cell types using complementary techniques and demonstrate that mAbs derived from immunization with an engineered survivin peptide have antitumor properties in different mouse tumor models.

GL261, U87, HT-29, HeLa, A2780, A1207, B16f1, Jurkat, Raji, MCF7, PC-3, and HEK293T cells were grown in complete DMEM containing 10% FCS, 5,000 units of penicillin/streptomycin, 50 μmol/L 2-mercaptoethanol, 25 nmol/L HEPES, and 1× nonessential amino acids at 37°C in 5% CO₂ with media changes two to three times per week. For animal studies, GL261 and B16f1 were obtained from the Division of Cancer Treatment and Diagnosis (DCTD) repository at the National Cancer Institute, NIH (Bethesda, MD). Additional cell lines were obtained through ATCC. No further authentication studies were done.

Anti-c-Myc Tag (9E10) was purchased from Thermo Fisher Scientific. Anti-Caveolin-1 and anti-CD71 (D7G9X) were purchased from Cell Signaling Technology. Anti-Flotillin-1 (18/Flotillin-1) was purchased from BD Transduction Laboratories. Anti-GST was purchased from ABM labs. Anti-survivin (60.11) was purchased from Novus Biologicals. Anti-Gα q/11/14 (G7), anti-EGFR (R-1) and anti-GAPDH (FL-335) were purchased from Santa Cruz Biotechnology.

Survivin peptide synthesis was performed using Fmoc chemistry and solid support resin and followed by conjugation to keyhole limpet hemocyanin (KLH) for immunization (Genscript USA). Additional survivin peptides, including N-terminal biotinylated survivin, were synthesized by CS Bio. Hybridomas were generated under contract with Genscript USA. Ten C57BL/6 mice were prebled at −4 days and followed with immunization subcutaneously with SVN53-67/M57-KLH on day 0. Mice were boosted on days 14, 28, and 56 and test bled at 21 and 35 days. Cells were fused to create hybridomas on day 60. Immune responses were tested by ELISA and Western blot analysis with target and irrelevant tagged protein 7 days after each boost immunization. On the basis of the test bleed result, two animals were selected for electrofusion. A fusion efficiency of one hybridoma per 2,500 splenocytes was expected. An average of 1 × 10⁸ B cells from each spleen of immunized mouse provides an anticipated recovery of 2 × 10⁴ hybridoma clones. All fused cells from each cell fusion were plated into 15 × 96-well plates. Supernatants were screened by indirect ELISA with antigen peptide. Twenty positive clones were seeded into 24-well plates, and five positive primary clones were subcloned by limiting dilution. The clones were carried for a maximum of three generations. Subcloned cells were again screened by indirect ELISA with both wild-type (SVN53-67) and engineered (SVN53-67/M57) peptides. Two stable subclonal cell lines of each primary clone were chosen for cryopreservation based on confirmed antigen-recognition and normal doubling time. Isotype identification for all the subcloned cell lines was performed. Subclone cultures were scaled up for production of purified IgG.

To generate hSurvivin-Myc/FLAG, human survivin fused to Myc and FLAG tags at its C-terminus was first excised from a commercial cDNA clone (OriGene). The survivin-Myc/FLAG insert was then used as the template in a PCR reaction with primers hSVN-EcoRI-F (5′-GCGATCGAATTCGTCGACTGG-3′) and hSVN-XbaI-R (5′-ACTCCTCTAGAAAACCTTATCGTCGTC-3′). Digested insert was then ligated into pcDNA3.1 (Thermo Fisher Scientific), and the construct was validated by sequencing. To generate the GST-hSurvivin plasmid used for protein purification, the survivin ORF (but not Myc or FLAG tag sequence) was excised from a commercial cDNA clone (OriGene) using EcoRI and NotI, and ligated into pGEX-5X-2 (GE Healthcare Life Sciences).

Overnight cultures of BL21-CodonPlus (DE3)-RIL (Agilent) containing GST-hSurvivin or the empty pGEX-5X-2 vector (500 μL) were used to inoculate 10 mL of Terrific Broth containing ampicillin and 80 μmol/L ZnCl₂, which was then grown to mid-log phase at 37°C with shaking. Expression was induced with 0.2 nmol/L IPTG for 5 hours at 30°C, after which bacteria were pelleted, resuspended in buffer containing 50 nmol/L Tris-Cl, pH 7.4, 150 nmol/L NaCl, 5 nmol/L MgCl₂, 80 μmol/L ZnCl₂, 1 nmol/L DTT plus protease inhibitors, and lysed by freeze/thaw and sonication. Supernatants were incubated with Glutathione Agarose (Thermo Fisher Scientific) for 4 hours at 4°C, and then beads were collected and washed. GST and GST-Survivin were eluted in buffer containing 50 nmol/L Tris-Cl, pH 8.0, 20 nmol/L reduced l-Glutathione and 0.1% Triton-X by incubation for 10 minutes at room temperature, then elutions were concentrated using 10 KD MWCO concentrator columns (Thermo Fisher Scientific). Protein concentrations were measured by bicinchoninic acid (BCA) assay and adjusted by adding buffer without glutathione or Triton-X.

Antibody binding to survivin peptides was measured using the AlphaScreen reagent Mouse IgG Detection Kit (Perkin-Elmer). Reactions were carried out in volumes of 50 μL in wells of an opaque, half-area 96-well plate, in buffer consisting of PBS pH 7.4, 0.1% BSA, and 0.01% Tween-20. Each well contained a final concentration of 20 μg/mL Acceptor Beads, 20 μg/mL Donor Beads, 0.5 nmol/L antibody (2C2 or 30H3) and 0.02–20 nmol/L biotinylated survivin peptide. Briefly, antibody was incubated with anti-mouse IgG Acceptor Beads plus dilutions of biotinylated survivin peptide for 30 minutes at room temperature, after which Streptavidin Donor Beads were added. The reactions were incubated another 30 minutes in the dark at room temperature and read on an EnVision Excite Multilabel Reader (Perkin-Elmer). Data was fit using GraphPad Prism 7, and K_d values were extrapolated from saturation curves.

Cultured cells were seeded on sterile glass coverslips in complete media overnight. Cells were washed with PBS, fixed for 15 minutes at room temperature in 4% paraformaldehyde and (if noted) permeabilized for 10 minutes using 0.05% Triton-X. Coverslips were incubated with the indicated survivin antibodies for 2 hours at room temperature, washed with PBS, incubated with secondary antibodies for 1 hour at room temperature, and then washed in PBS overnight at 4°C before mounting on glass slides and analyzing by fluorescence microscopy.

Formalin-fixed paraffin-embedded sections from a multi-tumor tissue microarray and one multi-organ normal tissue array (FDA999n; Biomax; 0.6 mm per core) were cut at 4 μm, placed on charged slides, and dried at 60°C for 1 hour. Slides were deparaffinized at room temperature through a series of xylene and graded ethanol baths. Slides were pretreated in citrate buffer (BioCare Medical CB910) for 60 minutes in a steamer and then cooled for 20 minutes. Slides were quenched in 0.3% H₂O₂ for 10 minutes, followed by protein block (X909, Dako) for 5 minutes. Slides were incubated with 2C2 antibody (1:100; 2 μg/mL) in a humidity chamber for 16 hours. Slides were loaded onto the Dako Autostainer PLUS and mouse Envision HRP was applied for 30 minutes (Dako K4007). Diaminobenzidine (Dako) was applied for 10 minutes. Slides were counterstained with hematoxylin for 2 minutes, dehydrated, cleared, and coverslipped.

Aliquots of anti-survivin 2C2 (0.2 μg each) were preincubated with 100 μg of immunizing survivin peptide or scrambled peptide for 30 minutes at room temperature. Cells were trypsinized, washed, and suspended in staining buffer. Preincubated antibody was added, and cells were incubated at room temperature for 1 hour. Cells were washed with PBS and stained with anti-mouse-FITC antibody for 30 minutes at room temperature, washed again, and data acquired with a BD Fortessa flow cytometer running FACSDiva software with final analysis using FCS Express v4.0.

Cultured cells were collected either by gently dislodging with PBS or by limited trypsinization, then washed with PBS and resuspended in DMEM to a concentration of 10⁶/mL. Cells were treated with either trypsin (1 mg/mL) or proteinase K (0.5 mg/mL) for 20 minutes at 37°C, then proteases were inactivated by the addition of either 1 mg/mL soybean trypsin inhibitor (for trypsin) or protease inhibitor diluted to 1× and 1 mmol/L phenylmethylsulfonylfluoride. Cells were washed twice with PBS, then resuspended in Cell Staining Buffer (Biolegend) or Cell Staining Buffer plus Fixation Buffer (Biolegend) plus 0.1% Triton-X. Cells were then stained with 0.2 μg of 2C2 for 1 hour at room temperature, washed once with PBS, incubated with FITC goat anti-mouse antibody, and analyzed by flow cytometry.

Lipid rafts were isolated following the detergent-free method of Macdonald and Pike, 2005 (19). Briefly, cells were scraped from three to four 15-cm culture plates into PBS, collected, and resuspended in 1 mL of Base Buffer containing 20 nmol/L Tris-Cl, pH 7.8, 250 nmol/L sucrose, 1 nmol/L CaCl₂, and 1 nmol/L MgCl₂, plus protease inhibitors. In the case of HEK293T, cells in 15-cm dishes were transfected with the hSVN-Myc/FLAG construct using Lipofectamine 2000 (Thermo Fisher Scientific) 24 hours prior to collection. After 15 minutes on ice, cells in Base Buffer were homogenized through a 22-gauge needle 20 times and centrifuged at 1,000 × g for 10 minutes. The post-nuclear supernatant was collected on ice, and the nuclear pellet was subjected to homogenization and extraction again in 1 mL Base Buffer. Both post-nuclear supernatants were pooled and stored at −20°C. Upon thawing, the post-nuclear supernatant was mixed with 2 mL of 50% Optiprep (Stemcell Technologies) and dispensed into a 12-mL tube. Two milliliters each of 20%, 15%, and 10% Optiprep solutions were then layered above the post-nuclear supernatant, followed by 2-mL base buffer. Gradients were centrifuged at 52,000 × g for either 2 (U87) or 3 hours (A1207 and HEK293T) in a Beckman Coulter L9-55M Ultracentrifuge equipped with SW28 rotor. After centrifugation, fractions were collected as 0.67 mL, and 25 μL of each was loaded onto 12% SDS-PAGE for Western blot analysis.

Lipid rafts in GL261 cells were labeled using the Vybrant Lipid Raft Labeling Kit (Molecular Probes) as recommended by the manufacturer. Briefly, cells were detached from culture dishes and resuspended in cold complete media containing 1 μg/mL Alexa Fluor 594 dye-labeled CT-B, incubated for 10 minutes at 4°C, washed in cold PBS, and then resuspended in anti-CT-B antibody solution (anti-CT-B diluted 1:200 in compete media) for 15 minutes at 4°C. Cells were washed twice with PBS, then resuspended in Cell Staining Buffer (Biolegend) and stained with 0.2 μg of 2C2 for 45 minutes at room temperature, washed with PBS, incubated with FITC goat anti-mouse antibody, and analyzed by imaging flow cytometry on an ImageStreamX Mark II Imaging Flow Cytometer (AMNIS/Millipore).

Cells were lysed in RIPA buffer and protein concentration was measured using Pierce BCA Assay Reagent. Aliquots of 20 μg total protein were mixed with Laemmli buffer, boiled, and separated on 12% SDS-PAGE gels, and transferred to polyvinylidene difluoride membranes, and then blocked with 5% nonfat milk in TBST for 1 hour. Antibodies were applied in blocker overnight at dilutions of 1:1,000 (60.11) or 1:100 (2C2 and 30H3), after which the membranes were washed, incubated with HRP-conjugated secondary antibody, washed again, incubated with ECL reagent, and exposed to X-ray film. For Western blotting of purified protein, 0.2 μg of GST or GST-Survivin were loaded in triplicate onto a 12% SDS-PAGE gel, transferred, and blots were processed as above using anti-GST at 1:1,000 and 2C2/30H3 at 1:500. ECL Chemiluminescence was captured using the Bio-Rad ChemiDoc Touch Imaging System.

GL261 cells were plated on sterile 96-well plates (2 × 10³ cells per well). Cells were treated with 2C2 and 30H3 antibodies in DMEM at 10 μg/mL in quadruplicate. Daily, cells were washed with PBS and fixed with 100% methanol at 4°C for 10 minutes. Fixed cells were stained with crystal violet for 10 minutes, washed, and allowed to dry. Crystal violet was solubilized in 10% acetic acid and absorbance measured on a Synergy (TM) HTX multi-mode microplate reader at 595 nm. Data were analyzed with GraphPad PRISM.

C57BL6 and nude (nu/nu) mice were implanted with murine GL261 glioma cells or B16f1 melanoma cells. For flank tumors, a suspension of 2 × 10⁷ Gl261 glioma cells or 1 × 10⁶ B16f1 melanoma cells in 100 μL of PBS was injected subcutaneously in male C57Bl/6 (immunocompetent) mice (Charles River) and in nude (immunocompromised) NCr-nu/nu mice (Charles River). After 3 days, mice were randomly assigned to groups and were given intraperitoneal (i.p.) injections of 10 μg of anti-survivin antibody 2C2 or 10 μg of IgG control antibody. Antibodies were administered every 5 days for a total of four doses. Tumor growth was measured daily using calipers and volumes were calculated according to the formula V = {a × b²}/2, where V is volume and a and b are perpendicular diameters of the tumor. Subcutaneous tumor volume significance was measured using Wilcoxon matched pairs signed rank test. All mouse flank tumor model experiments were approved by the Roswell Park Institutional Animal Care and Use Committee (IACUC).

Male C57BL/6 mice (Charles River) were anesthetized with isoflurane and fixed in a stereotactic head frame (David Kopf Instruments). A midline scalp incision was made and the bregma was identified. Stereotactic coordinates were measured (2.0 mm lateral and 1.2 mm anterior to the bregma). A burr hole was drilled at this point and 1 × 10⁵ GL261 cells suspended in 2.5 μL of DMEM was injected through a Hamilton syringe with a fixed, 25-gauge needle at a depth of 3.0 mm relative to the dura mater. Injections were performed at 1 μL/minute. After 3 days, mice were randomized into groups and injected with 10 μg anti-survivin antibody or 10 μg IgG for control. Antibodies were administered subcutaneously every 5 days for a total of five doses. SurVaxM (100 μg) was administered in Montanide ISA 51 with 100 ng mGM-CSF. Mice were followed for signs of neurologic deficits as an indicator of tumor growth and were sacrificed according to established criteria. Significance was assessed by Kaplan–Meier methodology and Mantel–Cox test for all groups. All mouse intracranial implantation experiments were approved by the Roswell Park IACUC.

Antibody-dependent cell-mediated cytotoxicity (ADCC) was assessed using a mFcγRIV ADCC Reporter Bioassay Kit (Promega). For experiments using prebound target cells, U87 cells were trypsinized, washed with complete media, resuspended at a concentration of 10⁶ cells/mL, and recovered at 37°C for 2 hours. Cells were collected and resuspended in assay buffer and incubated with nonspecific mouse IgG or mAb 2C2 for 30 minutes at 37°C. Cells were washed to remove unbound antibody, diluted, and used as directed for the assay. Effector cells were diluted in assay buffer according to the manufacturer's instructions. In 96-well opaque culture plates, 25 μL effectors were added to either 50 μL of prebound U87 cells, 50 μL of assay buffer containing increasing concentrations of 2C2, or to 25 μL of peptides at increasing concentrations, after which 25 μL 2C2 was added to a final concentration of 4 μg/mL. Plates were covered and incubated at 37°C in a humidified, 5% CO₂ incubator for 5 to 6 hours, after which luminescence was measured using Bio-Glo Reagent (Promega).

A panel of 19 murine hybridoma cell lines was developed against the engineered survivin peptide (SVN53-67/M57). Supernatants were screened, and high-affinity IgG mAbs were produced. Purified antibodies were characterized for target recognition and specificity. Cross-reactivity of individual hybridoma culture supernatants to wild-type survivin peptide SVN53-67 from 13 representative cell lines are shown in Fig. 1A. We chose to focus on two of these clones (2C2 and 30H3) with different affinities and isotype. Antibodies were compared with a commercially available survivin antibody (60.11; Novus Biologicals) in Western blots of total protein derived from human embryonic kidney 293T cells (Fig. 1B). All three antibodies detected a 16.5 kDa survivin species with limited minor bands (Fig. 1C). To measure antibody affinity, we utilized the fluorescence proximity-based AlphaLISA assay in which binding K_d was measured from the saturation of antibodies by biotinylated wild-type survivin peptide SVN53-67, (Fig. 1F). Anti-survivin antibody 2C2 (IgG2a) displayed approximately 4-fold greater binding affinity (K_d = 0.56 nmol/L) than mAb 30H3 (IgG1; K_d = 2.3 nmol/L) for wild-type survivin (Fig. 1F).

Both antibodies 2C2 and 30H3 detected filter-bound survivin peptides SVN53-67 and SVN 53-67/M57 with little reactivity to scrambled peptide or KLH carrier protein (Fig. 1D). The addition of wild-type survivin peptide to the antibodies during immunoblotting effectively blocked their detection of survivin peptide, demonstrating high antibody specificity (Fig. 1E). Antibody 2C2 detected strong survivin expression in a wide range of human cancers, including glioblastoma, invasive ductal breast carcinoma, lobular breast carcinoma, carcinoid tumor, clear cell renal carcinoma, hepatocellular carcinoma, medullary thyroid carcinoma and pheochromocytoma in a tissue microarray (Fig. 4). As determined by the pathology coinvestigator (J. Qiu), patterns of survivin expression included nuclear, cytoplasmic, apical, and membranous. Normal human brain, breast, and liver revealed no specific survivin expression, although modest nuclear staining was seen in renal tubular epithelium (Fig. 4H).

Analysis of nonpermeabilized U87 and A1207 human glioma cells by flow cytometry confirmed that cell-surface binding was present and was competed out by wild-type survivin peptide, but not scrambled peptide (Fig. 2A). Survivin was not detected on the surface of nonpermeabilized cells that had been treated with either trypsin or proteinase K (Fig. 2B). Treatment with proteases did not impair the ability of antibody to detect intracellular survivin following subsequent permeabilization. Flow cytometry analysis using anti-survivin 2C2 identified survivin expression on the surface of a diverse set of human and murine cancer cell lines, including those derived from gliomas (U87, A1207, and GL261), ovarian (A2780), breast (MCF7), prostate (PC3), colon (HT29), skin (B16 melanoma) and cervical (HeLa) cancers, leukemia (Raji), and lymphoma (Jurkat; Fig. 2A and 2C). Survivin was also detected in permeabilized tumor cell lines with 2C2 and 30H3 by immunofluorescence microscopy (Fig. 2D), and similar results were obtained with a commercially available survivin antibody (60.11) to an undefined survivin sequence. Nonpermeabilized cells showed significant surface binding of survivin antibodies with focal concentrations of fluorescence on the surface of the cell membranes of both murine and human cancer cell lines.

To test whether survivin is contained within lipid rafts of the cell membrane, we performed cell fractionation. Western blotting of fractionated lysates from U87 (Fig. 3B) and A1207 (Fig. 3C) human glioblastoma cells revealed the presence of survivin in the low-density lipid raft fraction (Fig. 3A) along with known lipid raft markers flotillin-1, EGFR, and caveolin-1. Cells transfected with an expression vector encoding a c-Myc/FLAG-tagged survivin construct also revealed detectable recombinant myc-survivin within the lipid raft fractions (Fig. 3D). Similarly, imaging flow cytometry showed colocalization of survivin and bound cholera toxin B (CT-B) (a lipid raft marker) in some, but not all, lipid rafts (Fig. 3E).

Both 2C2 and 30H3 anti-survivin antibodies inhibited growth of GL261 gliomas and B16 melanomas in flank tumor models compared with irrelevant IgG control antibody (Fig. 5A–D). There was no significant difference between 2C2 and 30H3 with respect to tumor inhibition, despite their difference in affinity for the target antigen. Although the inhibition of tumor growth by 2C2 and 30H3 was more effective in immunocompetent mice (Fig. 5A and B), both survivin antibodies inhibited GL261 and B16 tumor growth in immunodeficient (nude) mice as well (Fig. 5C and D). Vaccination of immunocompetent (C57BL/6) mice with SVN53-67/M57-KLH (SurVaxM) produced effective suppression of tumor growth (Fig. 5F). In comparison, treatment of tumor-bearing mice with pooled serum from vaccinated mice was also effective at suppressing tumor growth in naïve GL261 tumor-bearing mice, although less so than vaccine (Fig. 5F; P < 0.05). Survivin antibody (2C2) also significantly prolonged the survival of C57BL/6 mice with orthotopic (intracerebral) GL261 gliomas compared with either nonspecific PBS or IgG control (Fig. 5E; P = 0.004). There was no significant difference between nonspecific IgG and PBS controls (P = 0.089). In these experiments, treatments with vaccine and antibodies began 4 days after tumor cell implantation.

To assess one potential mechanism by which anti-survivin antibodies could mediate antitumor effects, we measured for ADCC in vitro. Preliminary experiments using splenocytes did not demonstrate measurable ADCC. To improve sensitivity, ADCC was assessed using a reporter assay that measures bioluminescent signal from engineered Jurkat effector cells expressing a FcγRIV-luciferase construct. We also found that Jurkat cells express survivin on their cell surface (Fig. 2C). Consequently, 2C2 antibody triggers FcγRIV-mediated luciferase expression, through binding Jurkat effector cells even in the absence of target cells (Fig. 6A). In addition, competition using free wild-type survivin peptide disrupts the 2C2-mediated, autoactivation of effector cells; whereas, scrambled peptide does not (Fig. 6B). Because of antibody recognition of survivin by effector (Jurkat) cells, subsequent assays were performed using U87 target cells that had been preincubated with 2C2 antibody, followed by washing to remove unbound antibody (Fig. 6C). Target U87 cells preincubated with normal mouse IgG did not trigger effector cell luminescence, whereas target cells preincubated with 2C2 stimulated mFcγRIV engagement and luciferase expression. In contrast to the results with 2C2, mAb 30H3 did not increase luciferase activity due to the fact that mouse IgG1 does not bind mFcγRIV (20, 21). To rule out any direct nonimmunologic effect upon cell viability, we assayed cell growth in the presence of antibody. Incubation of GL261 cells with 2C2 and 30H3 antibodies had no detectable effect upon in vitro cell viability (Fig. 6D).

Survivin is expressed by many different types of cancer cells. Patients with cancer have been shown to have anti-survivin antibodies in their circulation, even without active, specific vaccination (9). The presence of anti-survivin antibodies and survivin-specific T cells (9, 10) in patients with cancer indicates that survivin is released systemically by tumor cells and that it is immunogenic to some degree. Survivin peptide sequences, including those derived from the wild-type survivin molecule, and proteins encoded by various survivin splice variants, have been used to develop different cancer vaccine strategies (16, 22, 23–25). In these studies, the assumption that survivin is largely an intracellular protein target has led to a focus on T-cell–mediated antitumor mechanisms in these vaccine-based approaches. As survivin has been regarded as an intracellular molecule, its presentation as an immunologic target is postulated to be via processed survivin epitopes displayed by MHC-I molecules for TCR recognition (10, 11, 15).

In this study, we found that the survivin protein is present on the outer surface of a wide range of human and murine tumor cell lines. In addition, we developed mouse mAbs that inhibited tumor growth in several murine tumor models, perhaps via targeting of survivin associated with the outer surface of the tumor cell membrane. Our work shows, for the first time, that antibodies generated through survivin vaccination can also have antitumor effects. Thus, the finding of surface expression of survivin on tumor cells changes the paradigm for survivin-targeted immunotherapy to include antibody-mediated approaches as well.

Although the origin of circulating anti-survivin antibodies in patients with cancer is unclear, survivin protein has been identified in patient serum (26–30) and in association with tumor-derived exosomes (18). We have recently reported that the survivin protein is detectable by imaging flow cytometry on the outer surface of glioma-derived exosomes present in the circulation of patients with malignant gliomas, but not in normal noncancer controls (31). Moreover, vaccination of glioma patients using the SVN53-67/M57-KLH peptide vaccine may affect levels of survivin-bearing exosomes in the bloodstream (31). This study suggests that antibodies could arise from an immunologic response to tumor membrane survivin expression as well. It is also possible that exosomal survivin is derived from lipid raft survivin during the process of exosome generation and release (31).

Treatment of mice with purified anti-survivin antibodies or with pooled antisera from survivin-vaccinated mice inhibited tumor growth. Vaccination was more effective in preventing tumor progression than injection of antibody alone, and tumor growth inhibition was greater in syngeneic, immunocompetent, C57BL/6 mice than in immunodeficient (nu/nu) mice, suggesting that cellular immunologic elements may be necessary for antibody-mediated tumor inhibition. While some ADCC activity is observed in highly sensitive luciferase reporter systems, it is unclear whether ADCC is a major component of the antitumor effect of survivin antibodies in vivo. Nevertheless, both antibodies were similarly effective suggesting that the biological actions of anti-survivin antibodies are more complex in vivo than in cell culture and that ADCC probably does not account solely for the antitumor effects of anti-survivin antibodies.

For this study, we specifically selected antibodies generated in response to vaccination that were cross-reactive to the wild-type survivin sequence so that high affinity for the target antigen on tumor cells would be assured. In addition to activity in immunocompetent mice, these antibodies have efficacy against GL261 and B16 tumors in immunocompromised mice. In fully immunocompetent mice, vaccine was more effective than mAbs in inhibiting tumor growth, consistent with the fact that cytotoxic T cells and helper support are important components of vaccine action against tumors (11). As nude mice retain B cells, NK cells, macrophages and dendritic cells, which can all potentially mediate ADCC and other immunologic processes, additional studies will be required to define the mechanism underlying antibody-mediated antitumor effects. For example, we cannot exclude the possibility that antitumor effects of survivin antibodies in vivo are either fully, or in part, the result of opsonization, cell-mediated tumor killing, or interference with survivin-containing exosomes in the tumor microenvironment (29).

As previously shown with SurVaxM vaccine (11) and replicated here, antibody 2C2 prolongs the survival of immunocompetent mice with orthotopic (cerebral) GL261 gliomas indicating that the blood–brain barrier does not fully prevent survivin-specific antibodies and activated effector cells from gaining access to cerebral tumors. Moreover, it has been shown that the GL261 glioma model produces some disruption of the BBB as measured by high molecular weight tracers such as Evans Blue (32). Thus, cellular and antibody-mediated, survivin immunotherapies may have utility in treating cerebral tumors.

Although our work highlights the potential clinical importance of membrane-bound survivin, the cellular mechanics of surface expression and the biological role of survivin on the cell surface both remain to be defined. In one study, the protein encoded by solute carrier gene family 5a, member 8 (SLC5A8) was shown to translocate survivin to the plasma membrane in cancer cells by way of direct protein–protein interaction (33). Beyond that, survivin lacks a transmembrane domain and it is not known to interact with specific cell membrane components; however, it does possess amphipathic alpha helical structure (34). The focal pattern of expression observed by immunofluorescence microscopy suggests the likelihood that cell-surface survivin may be specifically associated with membranous lipid rafts. Packaging of exosomal survivin has previously been reported to be dependent on lipid rafts and to be associated with Hsp70, Hsp90, and PRDX1 proteins (35). If, as several studies suggest, exosomal survivin has tumor-promoting properties in the microenvironment, antibodies to survivin could also interfere with the action of these nanobodies, thus inhibiting tumor growth.

There are at least six survivin isoforms that appear to be self-regulating and may have distinct functions (36, 37). Survivin splice variants are generated through combinations of exons 1–4 with two alternate exons 2B and 3B (36–38). Survivin-2B is formed by the addition of an alternate exon 2, whereas, survivin-ΔEx3 contains a deletion of exon 3 that results in a truncated protein due to a frame shifted stop codon (36). A high ratio of survivin-ΔEx3 to wild-type survivin has been identified in aggressive cancers. In contrast, when survivin-2B is predominant over wild-type survivin, a reduced antiapoptotic potential is observed (39–46). High expression of survivin-ΔEx3 is also associated with high proliferative activity and tumor recurrence, while survivin-2B is associated with a lower rate of tumor recurrence (47–49).

Survivin variants can differentially traffic within the cell and can have variable subcellular localization (50, 51). For example, survivin-2B has been reported to preferentially localize in the cytoplasm, whereas, survivin-ΔEx3 preferentially accumulates in the nucleus (50, 52). Wild-type survivin and survivin-ΔEx3 have also been observed to interact within the mitochondria (40). Such observations demonstrate the multifaceted nature of survivin biology and raise the possibility that the unique survivin epitopes present in these isoforms may be differentially targetable via specific antibodies. It is possible that a surface-survivin isoform or conformational variant of survivin on the cell membrane is being detected by 2C2 and 30H3. In our study, the survivin epitope from which 2C2 and 30H3 antibodies are derived is conserved in all survivin isoforms described to date. Therefore, we are unable to make firm conclusions about which specific isoforms are present on the outer surface of tumor cells. Such information could help guide the development of highly specific antibody-mediated surviving-targeting tumor therapies.

While cancer vaccines provide one method for anti-survivin cancer immunotherapy, there are several potential advantages to antibody mediated approaches. Unlike with peptide vaccines, the effectiveness of mAbs is not HLA-dependent. Antibody therapy could potentially also have more immediate action against cancer cells than vaccines, which may take weeks to generate active immune responses and which may not work well in patients who are immunocompromised by their tumors or other treatments. Finally, with the presence of survivin protein on the surface of cancer cells, surface-survivin antibody-armed, CAR T-cells could have broad applicability for treating survivin-expressing cancers.

R.A. Fenstermaker is an employee of and has ownership interests (including patents) at MimiVax, LLC. J. Qiu reports receiving speakers bureau honoraria from and is a consultant/advisory board member for Merck. M.J. Ciesielski is an employee of and has ownership interests (including patents) at and reports receiving commercial research grants from MimiVax, LLC. No potential conflicts of interest were disclosed by the other authors.

Conception and design: R.A. Fenstermaker, S.A. Figel, M.J. Ciesielski

Development of methodology: R.A. Fenstermaker, S.A. Figel, J.-X. Qiu, P.M. Galbo, M.J. Ciesielski

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.A. Figel, J.-X. Qiu, T.A. Barone, S.S. Dharma, E.K. Winograd, P.M. Galbo, L.M. Wiltsie, M.J. Ciesielski

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R.A. Fenstermaker, S.A. Figel, J.-X. Qiu, E.K. Winograd, P.M. Galbo, M.J. Ciesielski

Writing, review, and/or revision of the manuscript: R.A. Fenstermaker, S.A. Figel, J.-X. Qiu, T.A. Barone, E.K. Winograd, L.M. Wiltsie, M.J. Ciesielski

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R.A. Fenstermaker, S.A. Figel, M.J. Ciesielski

Study supervision: M.J. Ciesielski

This project was funded by research grants from American Cancer SocietyRSG-11-153-01-LIB (to M.J. Ciesielski), the Roswell Park Alliance Foundation, Mr. Philip H. Hubbell, and the Linda Scime Endowment. This work was supported by Roswell Park Comprehensive Cancer Center and National Cancer Institute (NCI) grant P30CA016056.

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.