Cancer stem cells (CSC) drive tumorigenesis and contribute to genotoxic therapy resistance, diffuse infiltrative invasion, and immunosuppression, which are key factors for the incurability of glioblastoma multiforme (GBM). The AC133 epitope of CD133 is an important CSC marker for GBM and other tumor entities. Here, we report the development and preclinical evaluation of a recombinant AC133×CD3 bispecific antibody (bsAb) that redirects human polyclonal T cells to AC133+ GBM stem cells (GBM-SC), inducing their strong targeted lysis. This novel bsAb prevented the outgrowth of AC133-positive subcutaneous GBM xenografts. Moreover, upon intracerebral infusion along with the local application of human CD8+ T cells, it exhibited potent activity in prophylactic and treatment models of orthotopic GBM-SC–derived invasive brain tumors. In contrast, normal hematopoietic stem cells, some of which are AC133-positive, were virtually unaffected at bsAb concentrations effective against GBM-SCs and retained their colony-forming abilities. In conclusion, our data demonstrate the high activity of this new bsAb against patient-derived AC133-positive GBM-SCs in models of local therapy of highly invasive GBM. Cancer Res; 75(11); 2166–76. ©2015 AACR.

Many tumor types are organized as differentiation hierarchies with undifferentiated, stem-like cells at the apex. These cells are commonly known as cancer stem cells (CSC); they express typical stem cell and progenitor markers and are capable of persistent self-renewal and proliferation (1–4). Using patient-derived CSCs, we and others have shown that only these cells are tumorigenic upon xenotransplantation into immunodeficient mice whereas forced differentiation can completely abrogate their tumorigenicity (3, 5–8). CSCs have repeatedly been reported to preferentially locate to the invasive tumor front (9–11). This has also been documented for CD133+ pancreatic CSCs and CD133+ glioblastoma multiforme stem cells (GBM-SC; refs. 5, 12) and emphasizes the crucial role of CSCs in tumor progression. In addition, CSCs, especially in GBM, have been reported to be relatively resistant to conventional chemo- and radiotherapy and to be immunosuppressive (13, 14). Taken together, it is likely that therapeutic targeting of CSCs, for example, by immunotherapeutics, could improve the outcomes of tumor patients, including patients with GBM.

Bispecific antibodies (bsAb) exhibiting specificity for a tumor cell surface antigen and for the invariant CD3 signaling complex of the T-cell receptor can recruit any T cell, independent of its antigen specificity, to tumor cells. Upon binding, the T cells become activated, leading to a fast and efficient lysis of bound target cells. The power of this approach has not only been demonstrated in preclinical tumor models but also already in clinical trials investigating CD19×CD3 bsAbs in patients with leukemia and EpCAM×CD3 bsAbs in patients with EpCAM+ carcinoma-associated ascites or peritoneal carcinomatosis, where the bsAb was locally administered into the peritoneum (15–17).

Currently known cell surface markers of CSCs are receptors of developmental self-renewal pathways and other receptors found on lineage-negative cells, such as AC133, a stem cell–specific, N-glycosylation–dependent epitope of CD133 (18–20). The epitope is a marker not only for CSCs of several types of carcinomas but also for tumor stem cells of sarcomas, melanoma and highly aggressive brain tumors, including GBM (12, 21–24). GBM (WHO grade IV) is the most prevalent and most aggressive primary brain tumor in adults. The median survival time of patients with GBM is 14.6 months, with no curative treatments at present. The highly invasive growth pattern (which makes complete surgical resection impossible) and the high chemo- and radioresistance are mainly responsible for the incurability of GBM (25).

Here, we report on the development of a recombinant single-chain AC133×CD3 bsAb that redirects polyclonal CD3+ T cells to AC133+ GBM-SCs, inducing their efficient lysis. In the presence of human CD8+ T cells, this novel bsAb prevented the outgrowth of GBM-SC-derived subcutaneous (s.c.) tumors and, upon local administration, that of orthotopic brain tumors. Local application together with human CD8+ T cells also exerted strong antitumoral effects on established, orthotopically growing GBM-SC–derived brain tumors. In contrast, normal hematopoietic stem cells (HSC), a subset of which is AC133-positive, were more resistant to AC133×CD3 bsAb–mediated lysis, retaining their colony-forming abilities. Our data demonstrate potent preclinical activity of this new AC133×CD3 bsAb against patient-derived AC133+ CSCs and CSC-derived tumors in models of local therapy of highly invasive GBM.

Cell lines and cell culture

The human glioma cell line U251 was obtained from ATCC and transduced with an L1-puro-CD133 lentivirus to overexpress the CD133 antigen. Wild-type (WT) U251 cells and CD133-overexpressing U251 cells were cultured as previously described (26). NCH421k is a short-term patient-derived GBM-SC line and was cultured in serum-free medium supplemented with EGF and basic FGF (bFGF) as previously described (8). All three cell lines were transduced with a lentivirus coding for a fusion protein consisting of the latest generation of firefly luciferase and a neomycin resistance cassette, L1-FFneo-IRESneo, constructed in our laboratory according to standard procedures.

Differentiation of GBM-SCs

NCH421k neurospheres were transferred into DMEM containing 10% FCS and 10 nmol/L retinoic acid in the absence of GBM-SC growth factors for 3 to 4 weeks.

Fluorescence microscopic analysis of GBM-SC differentiation

NCH421k neurospheres were dispersed and centrifuged onto Cyto-Spin slides; NCH421k cells exposed to prodifferentiation culture conditions were grown on coverslips. For fluorescence microscopic analyses, the cells were stained with either phycoerythrin (PE)-labeled CD133/2 (293C3) monoclonal antibody (mAb; Miltenyi) or polyclonal rabbit anti-glial fibrillary acidic protein (GFAP) antibody (DAKO) followed by Alexa Fluor 546–labeled goat anti-rabbit F(ab′)2 fragments.

Construction of the recombinant AC133×CD3 bsAb

The cDNA sequences corresponding to the variable regions of the light and heavy chains of the anti-AC133.1 mAb (ATCC HB-12346; ref. 20) were obtained by phage display according to published methods (27), and those for the humanized anti-human CD3 mAb (clone UCHT1) were obtained from published sequences (GenBank: AJ853735.1). The cDNA sequences of the heavy and light chain polypeptides of the AC133 mAb were confirmed by nanoLC tandem MS analysis on an LTQ-Orbitrap XL+ETD mass spectrometer using trypsin (Promega), GluC (Roche), and LysN (U-Protein Express) proteolytic digestion followed by Mascot database searching of an in-house curated murine IPI database harboring a compendium of publically available mouse IgG cDNA sequences.

For bacterial expression, the bispecific single-chain construct was gene-synthesized (Genscript) and cloned into the pOPE101 vector (Progen) between the Nco1/BamH1 sites. As a specificity control, we produced a PSMA×CD3 bsAb from the VL and VH regions of the anti-PSMA mAb 3/A12 (27) and the humanized anti-human CD3 mAb (clone UCHT1).

Protein expression and purification

The bsAb cloned into the pOPE101 vector was transformed into E. coli XL1 Blue competent cells (Stratagene) and expressed and purified as described (27). The purified eluates were analyzed by SDS-PAGE and stained with Coomassie Brilliant Blue.

Flow cytometric binding analyses

The target cells (2.5 × 105) were incubated with increasing concentrations of the AC133×CD3 bsAb for 45 minutes at 4°C, followed by incubation with 1 μg/mL of mouse anti-c-myc mAb (Calbiochem) for 45 minutes. The cells were then incubated with 1.5 μg of anti-mouse PE-conjugated F(ab′)2 fragment (Dianova) for 30 minutes and analyzed on a Cytomix FC-500 flow cytometer (Beckman Coulter).

Preparation of effector T cells

Human peripheral blood mononuclear cells (PBMC) were isolated from fresh buffy coats by Pancoll (Pan Biotech) density gradient centrifugation using standard procedures. The CD8+ and CD4+ T cells were then isolated from the PBMCs using a human isolation kit (Miltenyi Biotec) or by FACS, respectively.

Assessment of T cell proliferation

Non-preactivated CD8+ T cells were labeled with 1 μmol/L carboxyfluorescein diacetate succinimidyl ester (CFSE) and incubated with CD133-overexpressing U251 cells or U251 WT cells in the presence or absence of 5 nmol/L of the bsAb at an effector-to-target cell (E:T) ratio of 10:1. CFSE dilution was analyzed by flow cytometry after 24 hours.

Quantitative in vitro cytotoxicity assay

The efficiency of the bsAb to redirect T cells and thereby kill antigen-expressing cells was determined using a FACS-based cytotoxicity assay. The target cells were labeled with PKH-26 (Sigma), transferred into a 48-well plate and allowed to adhere. Increasing concentrations of the AC133×CD3 bsAb were then added to the wells, along with nonactivated CD8+ T cells at an E:T ratio of 10:1. After incubation for 24 hours at 37°C in a CO2 incubator, the cells were collected and analyzed after adding 7-AAD (BD Pharmingen).

Subcutaneous tumor models

All animal experiments were performed in accordance with the German Animal License Regulations and were approved by the animal care committee of the Regierungspräsidium Freiburg (registration numbers: G-10/64, G-13/12, and G-13/101). NOD/SCID mice 6 to 8 weeks old (Charles River Laboratories) were s.c. injected into the flank with 5 × 106 CD133-overexpressing U251 cells mixed with 1 × 106 nonactivated human CD8+ T cells. Thereafter, the mice received daily intravenous (i.v.) injections of 10 μg of the AC133×CD3 bsAb for 14 days. Freshly isolated CD8+ T cells were injected again i.v. on days 6 and 12. Control groups consisted of mice receiving daily i.v. injections of PBS instead of the AC133×CD3 bsAb with or without co-injection of human CD8+ T cells. To assess the tumorigenicity of the patient-derived NCH421k GBM-SCs, 1 × 106 NCH421k cells were treated in vitro with 5 × 106 CD8+ T cells and 5 nmol/L of either the AC133×CD3 or the PSMA×CD3 bsAb for 96 hours. Posttreatment, the surviving cells were cultured without the bsAb for 5 days, collected thereafter and injected into the right or the left flanks of NOD/SCID mice, respectively. Growth of the xenografts was monitored by caliper measurement 3 times per week.

Orthotopic tumor models

Tumor cells were transduced with firefly luciferase for noninvasive monitoring of brain tumor growth by bioluminescence imaging (BLI). In this model, 1 × 105 CD133-overexpressing U251 cells or NCH421k GBM-SCs were implanted into the brain of 6- to 8-week-old NMRI nude mice (Harlan). Osmotic pumps (model 1007D; Alzet) were used to deliver the bsAb intracerebrally (i.c.). The pumps were filled with either the AC133×CD3 or the PSMA×CD3 bsAb at a concentration of 250 μg/mL in PBS. For implantation, the tumor cells were resuspended in 4 μL PBS and implanted into the brain with a Hamilton syringe (coordinates: 2 mm anterior and 3 mm to the right of the bregma at a depth of 3 mm). After the injection, the surface was cleaned and the brain infusion cannula (Brain Infusion Kit 3, 1–3 mm; Alzet) was placed in the injection site and fixed to the skull using cyanoacrylate adhesive (Alzet). The filled pumps were connected to the brain infusion cannula according to the manufacturer's instructions and placed s.c., delivering 3 μg of the bsAb per day for a period of 7 days. Tumor growth was monitored noninvasively using BLI acquired on an IVIS spectrum imaging system (Perkin Elmer) 3 times per week. In the co-implantation model, 2 × 105 nonactivated human CD8+ T cells were co-implanted together with the tumor cells. In the established tumor model, CD8+ T cells were injected on days 7 or 14 after tumor cell implantation, followed by the installation of the osmotic pumps and the brain infusion cannula.

Histopathology

Brains fixed with 4% paraformaldehyde were cut in a vibratome (VT-100S; Leica) in horizontal sections 60 μm apart. The brain sections were mounted and stained with hematoxylin.

CT imaging, image reconstruction, and analysis

For the determination of orthotopic tumor mass, contrast-enhanced microCT scans were conducted using Imeron350 (Bracco Imaging) as contrast agent. After inhalation narcosis with isoflurane, mice were placed and fixed on a heated bed and anesthesia was continued. Immediately before the start of the CT scan, the mice received a single injection of 120 μL Imeron350 i.v. The head region of each mouse was scanned in one bed position for 90 seconds using a 360° rotation step, a tube voltage of 65 keV, and a tube current of 300 μA. Images of the CT scans were reconstructed with a voxel size of 0.12 × 0.12 × 0.12 mm and a T30 kernel size, using the software provided by the manufacturer. Analysis of the CT images was performed using the AMIDE software. The thresholds of all images were scaled to show best contrast between both kinds of soft tissue, brain, and contrast-enhanced tumor. The contrast-enhanced visible tumor mass in each slice was marked by placing voxels of interest (VOI) to generate a 3D VOI tumor shape. The volume was then displayed by the software in mm3.

Human colony-forming cell assays

Human CD34+ HSCs were obtained from human leukapheresis products derived from granulocyte colony-stimulating factor (G-CSF)-mobilized donors using standard procedures. Briefly, donors were treated with human G-CSF, blood cells were collected by apheresis, and CD34+ cells were isolated by immunomagnetic separation using the CliniMACS affinity-based technology (CliniMACS CD34 MicroBeads; Miltenyi Biotec). For colony-forming assays, 5,000 to 10,000 CD34+ cells/mL were seeded in triplicate in 3-mL methylcellulose (MethoCult; Stem Cell Technologies). The dishes were incubated for 10 to 14 days at 37°C in 5% CO2. Thereafter, colonies were counted under an inverted microscope (Olympus IMT2).

Flow cytometric detection of AC133 expression

After incubation with FcR blocking agent, tumor cells or CD34+ HSCs were stained with anti-AC133-PE (Miltenyi).

Statistical analysis

Results are presented as means ± SD. Data were compared using the unpaired 2-tailed Student t test. P < 0.05 was considered significant. Analyses were performed using GraphPad Prism software version 6.0 (GraphPad Software Inc.).

Construct design, expression, and purification of the AC133×CD3 bsAb

For construction of the T-cell–recruiting bsAb, the sequences coding for the variable light chain (VL) and variable heavy chain (VH) regions of the anti-AC133 and the humanized UCHT1 anti-human CD3 mAb (Fig. 1A) were arranged in the domain order depicted in Fig. 1B, with standard glycine-serine linkers. The gene-synthesized construct was ligated into the pOPE101 expression vector, which contains the pelB leader sequence that allows secretion of the bsAb to the periplasm, a c-myc tag for protein detection, and a hexa-His tag for purification. The bsAb was purified from the periplasmic extract by immobilized metal affinity chromatography with high purity (Fig. 1C). The yield of the bsAb was about 250 to 300 μg/L of bacterial culture.

Figure 1.

Construction scheme, expression, and purification of the AC133×CD3 bsAb. A, construction scheme of the recombinant bsAb derived from the VL and VH regions of AC133 and CD3 antibodies. B, scheme of the vector for bacterial expression of the recombinant bsAb. C, SDS-PAGE of the Ni-affinity–purified bsAb stained with Coomassie Brilliant Blue, demonstrating high purity of the 55-kDa recombinant protein. M, prestained molecular weight marker. D, CSC features of the NCH421k cells. Representative examples of fluorescence microscopic images of NCH421k cells kept under stem cell culture conditions or prodifferentiation conditions. Note the dramatic changes in cell size and morphology as well as the differences in expression of the GBM-SC–specific CD133/2 epitope and the differentiation marker GFAP. Nuclei were stained with DAPI. A light microscopy picture was included to show the morphology and size/shape of anti-CD133/2–stained differentiated cells. DAPI, 4′,6-diamidino-2-phenylindole.

Figure 1.

Construction scheme, expression, and purification of the AC133×CD3 bsAb. A, construction scheme of the recombinant bsAb derived from the VL and VH regions of AC133 and CD3 antibodies. B, scheme of the vector for bacterial expression of the recombinant bsAb. C, SDS-PAGE of the Ni-affinity–purified bsAb stained with Coomassie Brilliant Blue, demonstrating high purity of the 55-kDa recombinant protein. M, prestained molecular weight marker. D, CSC features of the NCH421k cells. Representative examples of fluorescence microscopic images of NCH421k cells kept under stem cell culture conditions or prodifferentiation conditions. Note the dramatic changes in cell size and morphology as well as the differences in expression of the GBM-SC–specific CD133/2 epitope and the differentiation marker GFAP. Nuclei were stained with DAPI. A light microscopy picture was included to show the morphology and size/shape of anti-CD133/2–stained differentiated cells. DAPI, 4′,6-diamidino-2-phenylindole.

Close modal

CSC features of NCH421k cells

Under stem cell culture conditions, NCH421k cells are small and round and show high surface expression of the stem cell–specific CD133 epitopes AC133 (CD133/1) and CD133/2, usually coexpressed with each other (18, 26), but no expression of the astroglial differentiation marker GFAP (Figs. 1D and 2E); undifferentiated NCH421k cells are highly tumorigenic (5). Withdrawal of stem cell mitogens and exposure to FBS and retinoic acid initiate differentiation, whereupon the cells attach to the cultivation flask, lose expression of the stem cell–specific CD133 epitopes and adopt a morphology typical of differentiated brain cells, gaining expression of GFAP (Fig. 1D); differentiation is accompanied by loss of tumorigenicity (5).

Figure 2.

In vitro characterization of the functionality of the AC133×CD3 bsAb. A, binding specificity of the AC133×CD3 bsAb. FACS analysis of the binding to CD133-overexpressing U251 glioma cells (left) and to CD3-positive Jurkat cells (right) after incubation with various concentrations of the bsAb. B, T-cell activation by the bsAb. CD8+ T cells were labeled with CFSE and proliferation-induced CFSE dilution was monitored by FACS after a 24-hour incubation with CD133-overexpressing or WT U251 cells (E:T ratio 10:1) in the presence or absence of 5 nmol/L bsAb. C, microscopic visualization of T-cell activation and lysis of AC133+ tumor cells. CD133-overexpressing or WT U251 cells were incubated with CD8+ T cells in the presence of the bsAb as described in B. Activation-induced T-cell proliferation and tumor cell lysis were observed only in the presence of AC133+ tumor cells. CD8+ T cells cultured with the T-cell mitogen PHA and IL2 were used as positive control to visualize activation-induced T-cell proliferation. D, quantification of tumor cell lysis by FACS-based PKH assay. Tumor cells (PKH-labeled) and CD8+ T cells (E:T ratio, 10:1) were incubated at various concentrations of the AC133×CD3 bsAb. Lysis of CD133-overexpressing and WT U251 cells was analyzed after 24 or 48 hours for NCH421k GBM-SCs, respectively. E, FACS analysis of AC133 expression levels on CD133-overexpressing or WT U251 cells compared with NCH421k GBM-SCs. MFI, mean fluorescence intensity; OE, overexpressing; WT, wild-type. F, comparison of CD8+ and CD4+ T-cell–mediated cytotoxicity. PKH-labeled NCH421k cells and effector cells (E:T ratio, 5:1) were incubated at either 1 or 5 nmol/L of AC133×CD3 bsAb, and tumor cell lysis was analyzed by flow cytometry after 72 hours. Data (mean and SD of technical replicates) are representative of two independent experiments with similar results.

Figure 2.

In vitro characterization of the functionality of the AC133×CD3 bsAb. A, binding specificity of the AC133×CD3 bsAb. FACS analysis of the binding to CD133-overexpressing U251 glioma cells (left) and to CD3-positive Jurkat cells (right) after incubation with various concentrations of the bsAb. B, T-cell activation by the bsAb. CD8+ T cells were labeled with CFSE and proliferation-induced CFSE dilution was monitored by FACS after a 24-hour incubation with CD133-overexpressing or WT U251 cells (E:T ratio 10:1) in the presence or absence of 5 nmol/L bsAb. C, microscopic visualization of T-cell activation and lysis of AC133+ tumor cells. CD133-overexpressing or WT U251 cells were incubated with CD8+ T cells in the presence of the bsAb as described in B. Activation-induced T-cell proliferation and tumor cell lysis were observed only in the presence of AC133+ tumor cells. CD8+ T cells cultured with the T-cell mitogen PHA and IL2 were used as positive control to visualize activation-induced T-cell proliferation. D, quantification of tumor cell lysis by FACS-based PKH assay. Tumor cells (PKH-labeled) and CD8+ T cells (E:T ratio, 10:1) were incubated at various concentrations of the AC133×CD3 bsAb. Lysis of CD133-overexpressing and WT U251 cells was analyzed after 24 or 48 hours for NCH421k GBM-SCs, respectively. E, FACS analysis of AC133 expression levels on CD133-overexpressing or WT U251 cells compared with NCH421k GBM-SCs. MFI, mean fluorescence intensity; OE, overexpressing; WT, wild-type. F, comparison of CD8+ and CD4+ T-cell–mediated cytotoxicity. PKH-labeled NCH421k cells and effector cells (E:T ratio, 5:1) were incubated at either 1 or 5 nmol/L of AC133×CD3 bsAb, and tumor cell lysis was analyzed by flow cytometry after 72 hours. Data (mean and SD of technical replicates) are representative of two independent experiments with similar results.

Close modal

AC133×CD3 bsAb binds to AC133+ tumor cells and CD3+ T cells

Antigen binding of the AC133×CD3 bsAb to CD133-overexpressing U251 glioma cells, which are AC133-positive (Fig. 2A, left), and to CD3-expressing Jurkat T lymphoma cells (Fig. 2A, right) was demonstrated by flow cytometry. The equilibrium dissociation constants (Kd values) were 6.4 and 2.8 nmol/L, respectively. The bsAb did not show any binding to antigen-negative U251 WT cells.

AC133×CD3 bsAb activates resting T cells in the presence of AC133+ tumor cells

T-cell activation is usually accompanied by proliferative expansion. To track T-cell divisions induced by the AC133×CD3 bsAb, we labeled naïve human CD8+ T cells with the dye CFSE before incubation with tumor cells and the AC133×CD3 bsAb. Dividing T cells could only be noticed in the presence of AC133+ tumor cells, but neither in the presence of AC133-negative WT tumor cells nor in the absence of the bsAb (Fig. 2B). Likewise, clusters of dividing T cells, similar to those observed in cultures of phytohemagglutinin (PHA)-stimulated peripheral blood lymphocytes, could be microscopically detected only in the presence of AC133+ tumor cells but not in the presence of AC133-negative tumor cells (Fig. 2C). The costimulatory molecules CD80 and CD86 were not expressed on the tumor cells (data not shown).

AC133×CD3 bsAb redirects T cells to potently kill AC133+ tumor cells

Light microscopy also visualized the agent's potent and specific cytolytic potential. While virtually all CD133-overexpressing U251 glioma cells were lysed after a 24-hour incubation with the AC133×CD3 bsAb and CD8+ T cells at an E:T ratio of 10:1, antigen-negative U251 WT cells remained intact, forming a dense layer of adherent tumor cells (Fig. 2C). Specific cytotoxicity was quantified using a standard PKH-26 assay (Fig. 2D). The cytotoxic effect of the AC133×CD3 bsAb appeared to depend on the AC133 expression level of the target cells, which is about 10 to 15 times higher on CD133-overexpressing U251 glioma cells than on NCH421k patient–derived GBM-SCs (Fig. 2E); the EC50 values for the killing of the two different cell lines were 0.06 and 0.77 nmol/L, respectively. CD4+ T cells also executed bsAb-induced target cell killing. However, lysis was less efficient than with CD8+ T cells and only observed after 72 hours of coincubation with NCH421k cells and AC133×CD3 bsAb (Fig. 2F), whereas the CD8 T-cell kill was already observed after 48 hours (Fig. 2D).

Antitumor effects of the AC133×CD3 bsAb in subcutaneous and orthotopic glioma models

To assess the in vivo efficacy, we first studied whether i.v. injected AC133×CD3 bsAb could prevent the outgrowth of s.c. tumors when CD133-overexpressing U251 glioma cells were coimplanted with polyclonal human CD8+ T cells into immunodeficient mice. As shown in Fig. 3A, daily i.v. injections of the AC133×CD3 bsAb for 14 days after the coimplantation efficiently prevented the outgrowth of the CD133-overexpressing tumors. In contrast, when PBS was i.v. injected instead or when human T cells were not coimplanted, the outgrowth of the CD133-overexpressing tumors was not prevented.

Figure 3.

Antitumor activity of the AC133×CD3 bsAb in s.c. xenograft models. A, AC133×CD3 bsAb prevents the outgrowth of s.c. CD133-overexpressing U251 gliomas. CD133-overexpressing U251 cells and CD8+ T cells were coinjected s.c. into NOD/SCID mice. On days 6 and 12, the animals received an additional injection of CD8+ T cells i.v. Tumor outgrowth was prevented in mice treated with the bsAb (daily i.v. for 14 days as indicated by arrows) but not in control animals, which received PBS (n = 4 mice per group). The tumor growth rates were similar in the presence or absence of human CD8+ T cells. B, loss of tumorigenicity of patient-derived GBM-SCs following in vitro pretreatment with AC133×CD3 bsAb. NCH421k GBM-SCs were treated in vitro with CD8+ T cells and either the AC133×CD3 or the PSMA×CD3 bsAb for 96 hours. The surviving cells after treatment were injected into the right or the left flank of NOD/SCID mice, respectively (n = 4). The tumor volumes in each group are shown as means ± SD. *, P < 0.05; **, P < 0.01.

Figure 3.

Antitumor activity of the AC133×CD3 bsAb in s.c. xenograft models. A, AC133×CD3 bsAb prevents the outgrowth of s.c. CD133-overexpressing U251 gliomas. CD133-overexpressing U251 cells and CD8+ T cells were coinjected s.c. into NOD/SCID mice. On days 6 and 12, the animals received an additional injection of CD8+ T cells i.v. Tumor outgrowth was prevented in mice treated with the bsAb (daily i.v. for 14 days as indicated by arrows) but not in control animals, which received PBS (n = 4 mice per group). The tumor growth rates were similar in the presence or absence of human CD8+ T cells. B, loss of tumorigenicity of patient-derived GBM-SCs following in vitro pretreatment with AC133×CD3 bsAb. NCH421k GBM-SCs were treated in vitro with CD8+ T cells and either the AC133×CD3 or the PSMA×CD3 bsAb for 96 hours. The surviving cells after treatment were injected into the right or the left flank of NOD/SCID mice, respectively (n = 4). The tumor volumes in each group are shown as means ± SD. *, P < 0.05; **, P < 0.01.

Close modal

Patient-derived CSCs are increasingly used in xenograft models to validate therapeutic targets and to develop anti-CSC therapies. Therefore, we investigated whether the AC133×CD3 bsAb could prevent the outgrowth of s.c. tumors initiated from patient-derived AC133+ GBM-SCs. To this end, NCH421k GBM-SCs were first incubated with naïve human CD8+ T cells and either the AC133×CD3 bsAb or a prostate-specific membrane antigen (PSMA)×CD3 bsAb (produced as a control) for 96 hours in vitro. After 5 more days under CSC culture conditions, the remaining cells were injected s.c. into the right or the left flank of immunodeficient mice, respectively. As shown in Fig. 3B, preincubation with the AC133×CD3 bsAb led to a complete loss of in vivo tumorigenicity, whereas on the other hand, preincubation with the PSMA×CD3 bsAb and human CD8+ T cells did not prevent the outgrowth of tumors.

We also evaluated the therapeutic potential of the AC133×CD3 bsAb in orthotopic glioma models. For this purpose, we locally infused either the AC133×CD3 or the PSMA×CD3 bsAb continuously for 7 days with osmotic pumps and brain infusion kits. In a first set of experiments, luciferase-transduced CD133-overexpressing U251 glioma cells were coimplanted i.c. with human CD8+ T cells. Tumor growth monitoring by BLI (Fig. 4A) and tumor size determination by contrast-enhanced CT (Fig. 4B) or immunohistochemistry of brain slices (Fig. 4C) at the end of the experiments revealed that the AC133×CD3 bsAb efficiently prevented the outgrowth of CD133-overexpressing U251 gliomas. The same experiment was then performed using patient-derived AC133+ NCH421k GBM-SCs. As shown in Fig. 5A–C, despite the considerably lower AC133 expression compared with CD133-overexpressing cells, continuous local infusion of the AC133×CD3 bsAb, but not that of the PSMA×CD3 control bsAb, also efficiently prevented the outgrowth of GBM xenografts from NCH421k GBM-SCs coimplanted with human CD8+ T cells.

Figure 4.

Antitumor activity of the AC133×CD3 bsAb against orthotopic U251 gliomas overexpressing CD133. CD133-overexpressing U251 cells were coimplanted with CD8+ T cells into the brain of nude mice, followed by local infusion of either the AC133×CD3 or the PSMA×CD3 bsAb using 7-day Alzet osmotic pumps (n = 6–7 mice per group). A, tumor growth was monitored by BLI. B, differences in tumor volume at the end of the experiment (6.37 ± 1.02 mm3 for the AC133×CD3 bsAb-treated group and 52.40 ± 6.45 mm3 for the PSMA×CD3 bsAb-treated control group) were recorded by contrast-enhanced CT. Coronal and sagittal sections are shown; the tumor area is highlighted in green. C, in addition, the tumors were analyzed postmortem by staining brain slices with hematoxylin and scanning them with a whole-slide scanner. Representative sections are shown. ***, P < 0.001; values represent means ± SD.

Figure 4.

Antitumor activity of the AC133×CD3 bsAb against orthotopic U251 gliomas overexpressing CD133. CD133-overexpressing U251 cells were coimplanted with CD8+ T cells into the brain of nude mice, followed by local infusion of either the AC133×CD3 or the PSMA×CD3 bsAb using 7-day Alzet osmotic pumps (n = 6–7 mice per group). A, tumor growth was monitored by BLI. B, differences in tumor volume at the end of the experiment (6.37 ± 1.02 mm3 for the AC133×CD3 bsAb-treated group and 52.40 ± 6.45 mm3 for the PSMA×CD3 bsAb-treated control group) were recorded by contrast-enhanced CT. Coronal and sagittal sections are shown; the tumor area is highlighted in green. C, in addition, the tumors were analyzed postmortem by staining brain slices with hematoxylin and scanning them with a whole-slide scanner. Representative sections are shown. ***, P < 0.001; values represent means ± SD.

Close modal
Figure 5.

Antitumor activity of the AC133×CD3 bsAb against orthotopic xenografts initiated from patient-derived GBM-SCs. NCH421k GBM-SCs were coimplanted with CD8+ T cells into the brain of nude mice, followed by local infusion of either the AC133×CD3 or the PSMA×CD3 bsAb using 7-day Alzet osmotic pumps (n = 6–7 mice per group). A, tumor growth was monitored by BLI. B, differences in tumor volume at the end of the experiment (5.66 ± 1.11 mm3 for the AC133×CD3 bsAb-treated group and 45.85 ± 13.98 mm3 for the PSMA×CD3 bsAb-treated control group) were recorded by contrast-enhanced CT. Coronal and sagittal sections are shown; the tumor area is highlighted in green. C, in addition, the tumors were analyzed postmortem by staining brain slices with hematoxylin and scanning them with a whole-slide scanner. Representative sections are shown (*, P < 0.05; values represent means ± SD). D, treatment of established NCH421k tumors. CD8+ T cells were locally injected 7 days after tumor cell implantation, followed by osmotic pump–mediated microinfusion of the AC133×CD3 or the PSMA×CD3 bsAb from day 7 to day 14. Tumor growth was monitored by BLI. Dashed lines represent each individual mouse (n = 5 mice per group). E, delayed treatment of NCH421k tumors from day 14 to day 21 after tumor cell implantation (n = 6 mice per group).

Figure 5.

Antitumor activity of the AC133×CD3 bsAb against orthotopic xenografts initiated from patient-derived GBM-SCs. NCH421k GBM-SCs were coimplanted with CD8+ T cells into the brain of nude mice, followed by local infusion of either the AC133×CD3 or the PSMA×CD3 bsAb using 7-day Alzet osmotic pumps (n = 6–7 mice per group). A, tumor growth was monitored by BLI. B, differences in tumor volume at the end of the experiment (5.66 ± 1.11 mm3 for the AC133×CD3 bsAb-treated group and 45.85 ± 13.98 mm3 for the PSMA×CD3 bsAb-treated control group) were recorded by contrast-enhanced CT. Coronal and sagittal sections are shown; the tumor area is highlighted in green. C, in addition, the tumors were analyzed postmortem by staining brain slices with hematoxylin and scanning them with a whole-slide scanner. Representative sections are shown (*, P < 0.05; values represent means ± SD). D, treatment of established NCH421k tumors. CD8+ T cells were locally injected 7 days after tumor cell implantation, followed by osmotic pump–mediated microinfusion of the AC133×CD3 or the PSMA×CD3 bsAb from day 7 to day 14. Tumor growth was monitored by BLI. Dashed lines represent each individual mouse (n = 5 mice per group). E, delayed treatment of NCH421k tumors from day 14 to day 21 after tumor cell implantation (n = 6 mice per group).

Close modal

Effects on established orthotopic GBM-SC–derived glioma xenografts

In addition, we investigated the effect of the AC133×CD3 bsAb on established tumors using NCH421k GBM-SCs, which form highly invasive tumors in the brain of immunodeficient mice (5, 8). Luciferase-transduced NCH421k GBM-SCs were implanted i.c., and the local implantation of human CD8+ T cells and osmotic pumps filled with either the AC133×CD3 bsAb or the PSMA×CD3 bsAb was delayed until day 7 when the BLI signal of the NCH421k tumors had reached between 1 × 105 and 1 × 106 photons/s. As shown in Fig. 5D, in 4 of 5 mice treated with the AC133×CD3 bsAb, a strong inhibition of tumor growth was observed. In contrast, all the mice in the control group treated with the PSMA×CD3 bsAb showed progressive tumor growth. To confirm activity of the AC133×CD3 bsAb in later stage, established brain tumors, treatment was started only after 14 days and conducted until day 21. As shown in Fig. 5E, the tumors progressed between days 20 and 35 in PSMA×CD3 bsAb-treated mice, whereas tumor growth was again strongly inhibited in animals treated with the AC133×CD3 bsAb. These data demonstrate that the AC133×CD3 bsAb has antitumor activity against established tumors in orthotopic xenograft models with patient-derived CSCs.

AC133×CD3 bsAb does not affect HSCs at concentrations effective against tumor cells

Although the AC133×CD3 bsAb could be administered locally into brain tumors as described above, a proportion of the locally administered bsAb may escape into the blood circulation and affect extracerebral AC133+ cells. About 40% to 70% of human lineage-negative CD34+ HSCs are AC133-positive (28). We therefore studied the effect of the AC133×CD3 bsAb on the survival and function of human CD34+ HSCs. We first compared the AC133 expression on patient-derived NCH421k GBM-SCs and CD34+ HSCs by flow cytometry. As shown in Fig. 6A, the antigen-positive proportion of the CD34+ HSCs showed 5- to 7-fold lower expression than NCH421k GBM-SCs, implying that CD34+ HSCs might be less prone to AC133×CD3 bsAb-mediated lysis than the patient-derived CSCs. At several different concentrations ranging from 0.1 to 5 nmol/L of the AC133×CD3 bsAb and an E:T ratio of 5:1, the CD34+ HSCs indeed remained viable, whereas the NCH421k GBM-SCs were eliminated (Fig. 6B). Finally, we assessed the functionality of the HSCs by analyzing their colony-forming potential at two different concentrations of the AC133×CD3 bsAb. Except for a minor reduction in burst-forming unit-erythroid colony formation at a bsAb concentration of 1 nmol/L, no considerable effect of the bsAb on erythroid and myeloid colony formation was observed at these bsAb concentrations and an E:T ratio of 5:1 (Fig. 6C). Taken together, these data strongly suggest that there is a therapeutic window not only upon i.c. but also upon systemic administration of the AC133×CD3 bsAb, where AC133+ CSCs are killed but HSCs are spared.

Figure 6.

The multilineage reconstitution potential of human HSCs is not affected at AC133×CD3 bsAb concentrations effective against GBM-SCs. A, FACS analysis showing considerably lower AC133 expression of CD34+ HSCs compared with NCH421k GBM-SCs. Results shown are representative of four different experiments. B, FACS-based cytotoxicity assay showing unaffected CD34+ HSCs at various concentrations of the AC133×CD3 bsAb (in the range between 0.1 and 5 nmol/L) that mediate T-cell–induced killing of NCH421k GBM-SCs. Target cells were labeled with PKH and incubated with CD8+ T cells at an E:T ratio of 5:1 for 48 hours. C, hematopoietic colony-forming unit assays testing the functionality of CD34+ HSCs. Following in vitro incubation of CD34+ HSCs with CD8+ T cells at an E:T ratio of 5:1 and 0.5 or 1 nmol/L AC133×CD3 bsAb for 48 hours, the CD8+ T cells were removed and different colony-forming assays were performed. Except for a minor reduction of burst-forming unit-erythroid colony formation at 1 nmol/L bsAb, colony formation was not negatively affected. Data (mean and SD of technical replicates) are representative of two (B) or three (C) independent experiments with similar results.

Figure 6.

The multilineage reconstitution potential of human HSCs is not affected at AC133×CD3 bsAb concentrations effective against GBM-SCs. A, FACS analysis showing considerably lower AC133 expression of CD34+ HSCs compared with NCH421k GBM-SCs. Results shown are representative of four different experiments. B, FACS-based cytotoxicity assay showing unaffected CD34+ HSCs at various concentrations of the AC133×CD3 bsAb (in the range between 0.1 and 5 nmol/L) that mediate T-cell–induced killing of NCH421k GBM-SCs. Target cells were labeled with PKH and incubated with CD8+ T cells at an E:T ratio of 5:1 for 48 hours. C, hematopoietic colony-forming unit assays testing the functionality of CD34+ HSCs. Following in vitro incubation of CD34+ HSCs with CD8+ T cells at an E:T ratio of 5:1 and 0.5 or 1 nmol/L AC133×CD3 bsAb for 48 hours, the CD8+ T cells were removed and different colony-forming assays were performed. Except for a minor reduction of burst-forming unit-erythroid colony formation at 1 nmol/L bsAb, colony formation was not negatively affected. Data (mean and SD of technical replicates) are representative of two (B) or three (C) independent experiments with similar results.

Close modal

bsAbs recruiting polyclonal T cells to tumor cells are capable of inducing very potent tumor cell lysis (see Fig. 2C, D, and F). An important prerequisite for the efficacy of T-cell–recruiting bsAbs in the treatment of solid tumors is the presence of T cells in the tumor, particularly CD8+ cytotoxic T cells. Although some solid tumors are T-cell–inflamed, it is likely that the T-cell frequencies in advanced tumors are often not sufficient for complete elimination of tumor cells by bsAb-redirected T cells (29). However, the approach may be more efficient in minimal residual disease, for example, after incomplete surgical resection of invasive brain tumors. Autologous polyclonal T cells and T-cell–recruiting bsAbs locally applied may destroy residual tumor cells that are present in the resection cavity wall or have invaded the surrounding tissue. Earlier reports on local postsurgical treatment of high-grade gliomas with T-cell–recruiting bsAbs have shown clinical antitumor activity. Tumor regressions and prolonged survival in the majority of patients were reported in a study where lymphokine-activated killer cells preloaded with chemically cross-linked bsAbs with anti-glioma activity (but unknown specificity) were locally applied with Ommaya reservoirs (30). Also, resting T cells preloaded with chemically prepared EGFR-specific T-cell–recruiting antibody fragments were applied in a similar way and evidence for clinical activity was observed in some patients (31). We have observed that T-cell–recruiting bsAbs are rapidly internalized after binding to T cells, which would adversely affect the efficacy of T cells preloaded with bsAbs. We therefore did not use preloaded T cells but continuously infused the bsAb locally for several days after i.c. injection of CD8+ T cells at the tumor implantation site. Moreover, we targeted the AC133 epitope, an intensely studied CSC marker expressed on CSCs of many tumor entities, including high-grade gliomas. CSCs appear to be ideal targets particularly for postsurgical treatments because they preferentially locate to the invasive front of highly aggressive tumors, thereby often escaping surgical resection (5, 9, 10, 12). Moreover, they are responsible for tumor recurrences due to their high resistance to genotoxic therapies and their unique ability of tumor initiation (1–3, 5, 32). We indeed observed strong antitumoral effects of the AC133×CD3 bsAb on orthotopic GBM-SC–derived tumors (see Fig. 5), which are highly invasive (5, 8).

CSCs as targets for T-cell–recruiting antibodies in solid tumors have already been addressed by others. In these studies, EpCAM×CD3 bsAbs were effective against CSCs, including CSCs that coexpress EpCAM and CD133 (33–35). However, EpCAM is not expressed on all CSCs of EpCAM+ carcinomas (36, 37), and nonepithelial malignancies, such as primary brain tumors, leukemias, and sarcomas, are EpCAM-negative. Another promising target for T-cell–recruiting bsAbs is the highly tumor-specific EGFR variant EGFRvIII, which is expressed in 30% of GBM tumors and some other malignancies. EGFRvIII-specific T-cell–recruiting antibodies have shown activity against orthotopic EGFRvIII-overexpressing U87 gliomas in immunodeficient mice (38). It has recently been shown that EGFRvIII is associated with GBM-SCs and coexpressed with CD133 as well as other stem cell and progenitor markers (39). EGFRvIII-specific T-cell–recruiting Abs may therefore also induce T-cell–mediated lysis of CSCs, including GBM-SCs, in EGFRvIII+ patients. CSCs may also be lysed by cytokine-induced killer cells coated with chemically cross-linked anti-CD133/CD3 bsAbs, which have recently been shown to lyse conventional CD133high tumor cell lines better than uncoated cytokine-induced killer cells (40).

The AC133/CD133+ fraction among primary GBMs was reported to range from ≤1% to 50% (21, 23). In the latter study, a considerable proportion of tumors contained >10% or even >25% of AC133+ cells. Furthermore, the frequencies of CD133+ glioma cells and their proliferation seem to increase considerably at GBM recurrence (41). AC133+ tumor stem cells could perhaps also be targeted in brain tumors other than GBM. For medulloblastomas, 6% to 21% AC133+ cells have been reported (23). Particularly high frequencies (1%–36% at diagnosis and 30%–70% at relapse) have been found in teratoid/rhabdoid brain tumors (42). AC133+ tumor stem cells have also been detected in ependymomas (23).

T-cell–recruiting bsAbs may cause off-tumor effects by recruiting T cells to antigen-positive normal tissue, particularly stem cells. However, while fetal human neural stem cells are AC133-positive, postnatally, neurogenic astrocytes in the stem cell regions of the brain do not seem to express AC133 (43), which would be a good prerequisite for local AC133-targeted therapies of brain tumors. Moreover, local application into the tumor or the resection cavity may be capable of minimizing potential adverse effects. An advantage of the simultaneous, but separate, local application of bsAb and T cells could be that the application of the bsAb, which has a short half-life and is rapidly internalized into the T cells, can be stopped any time. Local application of T cells and antibody therapeutics can be more efficient than systemic treatment of locally advanced or incompletely resected intra- and extracerebral tumors, and feasibility and efficacy in patients will likely further increase with new emerging technologies (44–47).

Outside the brain, AC133 is expressed on subsets of normal stem cells in adult tissues (19). We found that human CD34+ HSCs express considerably less surface AC133 than NCH421k GBM-SCs. Nevertheless, further preclinical studies are required to find out whether AC133+ HSCs could be spared upon in vivo administration (local or systemic) of the AC133×CD3 bsAb or to what extent hematopoiesis would be affected in case of depletion of the AC133+ HSC subset. In case of tolerable normal-tissue toxicity, treatment of extracerebral malignancies, particularly such with relatively high frequencies of AC133+ CSCs, could be considered, for example, colon cancer (containing 2%–19% AC133+ cells; ref. 24) or ovarian cancer (containing 0.3%–35% AC133+ cells; refs. 48, 49). Pancreatic cancers and lung cancers usually contain smaller populations of AC133+ tumor cells (mean: 1.8% and 5%, respectively; refs. 12, 50). Because of their preferred localization at the invasive tumor front and their high relevance for tumor invasion and metastasis (5, 12), elimination of residual AC133+ CSCs after surgery may nevertheless enhance treatment outcomes even in tumors with low frequencies of functionally highly relevant CSCs.

In conclusion, our preclinical data show high activity of the AC133×CD3 bsAb against patient-derived AC133+ GBM-SCs in vitro as well as in vivo in s.c. and orthotopic models of brain tumors, including established, invasive, GBM-SC–driven brain tumors. Locally applied together with autologous CD8+ T cells as performed in this study, this new CSC-specific T-cell–recruiting antibody may be beneficial for the treatment of highly aggressive CSC-driven tumors, particularly when administered into surgically created resection cavities of brain tumors containing AC133+ CSCs. The efficacy may further be enhanced by simultaneous targeting of CSCs with other surface markers, if present, or by combination with approaches inducing long-lasting antitumor immunity against tumor cells. Patients eligible for therapy could perhaps be selected and monitored by AC133 mAb-mediated imaging (5). Together with immunohistochemical analyses, noninvasive imaging could also help acquire as much information as possible on normal-tissue expression of AC133, an indispensable prerequisite for potential clinical trials of AC133-targeted therapies.

S. Prasad, S. Gaedicke, M. Hettich, E. Firat, and G. Niedermann have ownership interests (including patents). J. Schueler Employment (other than primary affiliation; e.g., consulting) is Oncotest GmbH Department Head at Oncotest GmbH. No potential conflicts of interest were disclosed by the other authors.

Conception and design: S. Prasad, G. Niedermann

Development of methodology: S. Prasad, M. Machein, K. Klingner, J. Schüler, C. Herold-Mende, U. Elsässer-Beile, G. Niedermann

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Prasad, M. Machein, G. Mittler, F. Braun, E. Firat, J. Schüler, D. Wider, R.M. Wäsch

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Prasad, G. Mittler, F. Braun, M. Hettich, D. Wider, R.M. Wäsch, G. Niedermann

Writing, review, and/or revision of the manuscript: S. Prasad, M. Machein, F. Braun, M. Hettich, R.M. Wäsch, C. Herold-Mende, G. Niedermann

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Gaedicke, M. Hettich, E. Firat, J. Schüler, R.M. Wäsch, C. Herold-Mende, U. Elsässer-Beile

Study supervision: G. Niedermann

The authors thank Prof. AL. Grosu for her continuous support.

This work was supported by a grant from the Clotten Foundation to G. Niedermann.

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

1.
Nguyen
LV
,
Vanner
R
,
Dirks
P
,
Eaves
CJ
. 
Cancer stem cells: an evolving concept
.
Nat Rev Cancer
2012
;
12
:
133
43
.
2.
Yan
K
,
Yang
K
,
Rich
JN
. 
The evolving landscape of glioblastoma stem cells
.
Curr Opin Neurol
2013
;
26
:
701
7
.
3.
Kreso
A
,
Dick
JE
. 
Evolution of the cancer stem cell model
.
Cell Stem Cell
2014
;
14
:
275
91
.
4.
Visvader
JE
,
Lindeman
GJ
. 
Cancer stem cells in solid tumours: accumulating evidence and unresolved questions
.
Nat Rev Cancer
2008
;
8
:
755
68
.
5.
Gaedicke
S
,
Braun
F
,
Prasad
S
,
Machein
M
,
Firat
E
,
Hettich
M
, et al
Noninvasive positron emission tomography and fluorescence imaging of CD133+ tumor stem cells
.
Proc Natl Acad Sci U S A
2014
;
111
:
E692
701
.
6.
Maugeri-Sacca
M
,
Di Martino
S
,
De Maria
R
. 
Biological and clinical implications of cancer stem cells in primary brain tumors
.
Front Oncol
2013
;
3
:
6
.
7.
Piccirillo
SG
,
Reynolds
BA
,
Zanetti
N
,
Lamorte
G
,
Binda
E
,
Broggi
G
, et al
Bone morphogenetic proteins inhibit the tumorigenic potential of human brain tumour-initiating cells
.
Nature
2006
;
444
:
761
5
.
8.
Campos
B
,
Wan
F
,
Farhadi
M
,
Ernst
A
,
Zeppernick
F
,
Tagscherer
KE
, et al
Differentiation therapy exerts antitumor effects on stem-like glioma cells
.
Clin Cancer Res
2010
;
16
:
2715
28
.
9.
Siebzehnrubl
FA
,
Silver
DJ
,
Tugertimur
B
,
Deleyrolle
LP
,
Siebzehnrubl
D
,
Sarkisian
MR
, et al
The ZEB1 pathway links glioblastoma initiation, invasion and chemoresistance
.
EMBO Mol Med
2013
;
5
:
1196
212
.
10.
Kahlert
UD
,
Maciaczyk
D
,
Doostkam
S
,
Orr
BA
,
Simons
B
,
Bogiel
T
, et al
Activation of canonical WNT/beta-catenin signaling enhances in vitro motility of glioblastoma cells by activation of ZEB1 and other activators of epithelial-to-mesenchymal transition
.
Cancer Lett
2012
;
325
:
42
53
.
11.
Liu
S
,
Cong
Y
,
Wang
D
,
Sun
Y
,
Deng
L
,
Liu
Y
, et al
Breast cancer stem cells transition between epithelial and mesenchymal states reflective of their normal counterparts
.
Stem Cell Rep
2014
;
2
:
78
91
.
12.
Hermann
PC
,
Huber
SL
,
Herrler
T
,
Aicher
A
,
Ellwart
JW
,
Guba
M
, et al
Distinct populations of cancer stem cells determine tumor growth and metastatic activity in human pancreatic cancer
.
Cell Stem Cell
2007
;
1
:
313
23
.
13.
Bao
S
,
Wu
Q
,
McLendon
RE
,
Hao
Y
,
Shi
Q
,
Hjelmeland
AB
, et al
Glioma stem cells promote radioresistance by preferential activation of the DNA damage response
.
Nature
2006
;
444
:
756
60
.
14.
Di Tomaso
T
,
Mazzoleni
S
,
Wang
E
,
Sovena
G
,
Clavenna
D
,
Franzin
A
, et al
Immunobiological characterization of cancer stem cells isolated from glioblastoma patients
.
Clin Cancer Res
2010
;
16
:
800
13
.
15.
Bargou
R
,
Leo
E
,
Zugmaier
G
,
Klinger
M
,
Goebeler
M
,
Knop
S
, et al
Tumor regression in cancer patients by very low doses of a T cell-engaging antibody
.
Science
2008
;
321
:
974
7
.
16.
Heiss
MM
,
Strohlein
MA
,
Bokemeyer
C
,
Arnold
D
,
Parsons
SL
,
Seimetz
D
, et al
The role of relative lymphocyte count as a biomarker for the effect of catumaxomab on survival in malignant ascites patients: results from a phase II/III study
.
Clin Cancer Res
2014
;
20
:
3348
57
.
17.
Frankel
SR
,
Baeuerle
PA
. 
Targeting T cells to tumor cells using bispecific antibodies
.
Curr Opin Chem Biol
2013
;
17
:
385
92
.
18.
Kemper
K
,
Sprick
MR
,
de Bree
M
,
Scopelliti
A
,
Vermeulen
L
,
Hoek
M
, et al
The AC133 epitope, but not the CD133 protein, is lost upon cancer stem cell differentiation
.
Cancer Res
2010
;
70
:
719
29
.
19.
Grosse-Gehling
P
,
Fargeas
CA
,
Dittfeld
C
,
Garbe
Y
,
Alison
MR
,
Corbeil
D
, et al
CD133 as a biomarker for putative cancer stem cells in solid tumours: limitations, problems and challenges
.
J Pathol
2013
;
229
:
355
78
.
20.
Yin
AH
,
Miraglia
S
,
Zanjani
ED
,
Almeida-Porada
G
,
Ogawa
M
,
Leary
AG
, et al
AC133, a novel marker for human hematopoietic stem and progenitor cells
.
Blood
1997
;
90
:
5002
12
.
21.
Zeppernick
F
,
Ahmadi
R
,
Campos
B
,
Dictus
C
,
Helmke
BM
,
Becker
N
, et al
Stem cell marker CD133 affects clinical outcome in glioma patients
.
Clin Cancer Res
2008
;
14
:
123
9
.
22.
Chen
R
,
Nishimura
MC
,
Bumbaca
SM
,
Kharbanda
S
,
Forrest
WF
,
Kasman
IM
, et al
A hierarchy of self-renewing tumor-initiating cell types in glioblastoma
.
Cancer Cell
2010
;
17
:
362
75
.
23.
Singh
SK
,
Hawkins
C
,
Clarke
ID
,
Squire
JA
,
Bayani
J
,
Hide
T
, et al
Identification of human brain tumour initiating cells
.
Nature
2004
;
432
:
396
401
.
24.
O'Brien
CA
,
Pollett
A
,
Gallinger
S
,
Dick
JE
. 
A human colon cancer cell capable of initiating tumour growth in immunodeficient mice
.
Nature
2007
;
445
:
106
10
.
25.
Stupp
R
,
Hegi
ME
,
Mason
WP
,
van den Bent
MJ
,
Taphoorn
MJ
,
Janzer
RC
, et al
Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial
.
Lancet Oncol
2009
;
10
:
459
66
.
26.
Tsurumi
C
,
Esser
N
,
Firat
E
,
Gaedicke
S
,
Follo
M
,
Behe
M
, et al
Non-invasive in vitro imaging of tumor-associated CD133/prominin
.
PLoS One
2010
;
5
:
e15605
.
27.
Elsasser-Beile
U
,
Wolf
P
,
Gierschner
D
,
Buhler
P
,
Schultze-Seemann
W
,
Wetterauer
U
. 
A new generation of monoclonal and recombinant antibodies against cell-adherent prostate specific membrane antigen for diagnostic and therapeutic targeting of prostate cancer
.
Prostate
2006
;
66
:
1359
70
.
28.
Ebener
U
,
Brinkmann
A
,
Zotova
V
,
Niegemann
E
,
Wehner
S
. 
[Expression of AC133 vs. CD34 in acute childhood leukemias]
.
Klin Padiatr
2000
;
212
:
90
8
.
29.
Gajewski
TF
,
Schreiber
H
,
Fu
YX
. 
Innate and adaptive immune cells in the tumor microenvironment
.
Nat Immunol
2013
;
14
:
1014
22
.
30.
Nitta
T
,
Sato
K
,
Yagita
H
,
Okumura
K
,
Ishii
S
. 
Preliminary trial of specific targeting therapy against malignant glioma
.
Lancet
1990
;
335
:
368
71
.
31.
Jung
G
,
Brandl
M
,
Eisner
W
,
Fraunberger
P
,
Reifenberger
G
,
Schlegel
U
, et al
Local immunotherapy of glioma patients with a combination of 2 bispecific antibody fragments and resting autologous lymphocytes: evidence for in situ t-cell activation and therapeutic efficacy
.
Int J Cancer
2001
;
91
:
225
30
.
32.
Firat
E
,
Weyerbrock
A
,
Gaedicke
S
,
Grosu
AL
,
Niedermann
G
. 
Chloroquine or chloroquine-PI3K/Akt pathway inhibitor combinations strongly promote gamma-irradiation-induced cell death in primary stem-like glioma cells
.
PLoS One
2012
;
7
:
e47357
.
33.
Cioffi
M
,
Dorado
J
,
Baeuerle
PA
,
Heeschen
C
. 
EpCAM/CD3-Bispecific T-cell engaging antibody MT110 eliminates primary human pancreatic cancer stem cells
.
Clin Cancer Res
2012
;
18
:
465
74
.
34.
Jager
M
,
Schoberth
A
,
Ruf
P
,
Hess
J
,
Hennig
M
,
Schmalfeldt
B
, et al
Immunomonitoring results of a phase II/III study of malignant ascites patients treated with the trifunctional antibody catumaxomab (anti-EpCAM x anti-CD3)
.
Cancer Res
2012
;
72
:
24
32
.
35.
Herrmann
I
,
Baeuerle
PA
,
Friedrich
M
,
Murr
A
,
Filusch
S
,
Ruttinger
D
, et al
Highly efficient elimination of colorectal tumor-initiating cells by an EpCAM/CD3-bispecific antibody engaging human T cells
.
PLoS One
2010
;
5
:
e13474
.
36.
Yamashita
T
,
Honda
M
,
Nakamoto
Y
,
Baba
M
,
Nio
K
,
Hara
Y
, et al
Discrete nature of EpCAM+ and CD90+ cancer stem cells in human hepatocellular carcinoma
.
Hepatology
2013
;
57
:
1484
97
.
37.
Imrich
S
,
Hachmeister
M
,
Gires
O
. 
EpCAM and its potential role in tumor-initiating cells
.
Cell Adh Migr
2012
;
6
:
30
8
.
38.
Choi
BD
,
Kuan
CT
,
Cai
M
,
Archer
GE
,
Mitchell
DA
,
Gedeon
PC
, et al
Systemic administration of a bispecific antibody targeting EGFRvIII successfully treats intracerebral glioma
.
Proc Natl Acad Sci U S A
2013
;
110
:
270
5
.
39.
Emlet
DR
,
Gupta
P
,
Holgado-Madruga
M
,
Del Vecchio
CA
,
Mitra
SS
,
Han
SY
, et al
Targeting a glioblastoma cancer stem-cell population defined by EGF receptor variant III
.
Cancer Res
2014
;
74
:
1238
49
.
40.
Huang
J
,
Li
C
,
Wang
Y
,
Lv
H
,
Guo
Y
,
Dai
H
, et al
Cytokine-induced killer (CIK) cells bound with anti-CD3/anti-CD133 bispecific antibodies target CD133(high) cancer stem cells in vitro and in vitro
.
Clin Immunol
2013
;
149
:
156
68
.
41.
Tamura
K
,
Aoyagi
M
,
Ando
N
,
Ogishima
T
,
Wakimoto
H
,
Yamamoto
M
, et al
Expansion of CD133-positive glioma cells in recurrent de novo glioblastomas after radiotherapy and chemotherapy
.
J Neurosurg
2013
;
119
:
1145
55
.
42.
Chiou
SH
,
Kao
CL
,
Chen
YW
,
Chien
CS
,
Hung
SC
,
Lo
JF
, et al
Identification of CD133-positive radioresistant cells in atypical teratoid/rhabdoid tumor
.
PLoS One
2008
;
3
:
e2090
.
43.
Pfenninger
CV
,
Roschupkina
T
,
Hertwig
F
,
Kottwitz
D
,
Englund
E
,
Bengzon
J
, et al
CD133 is not present on neurogenic astrocytes in the adult subventricular zone, but on embryonic neural stem cells, ependymal cells, and glioblastoma cells
.
Cancer Res
2007
;
67
:
5727
36
.
44.
Bousquet
G
,
Darrouzain
F
,
de Bazelaire
C
,
Ternant
D
,
Barranger
E
,
Winterman
S
, et al
Intrathecal trastuzumab halts progression of CNS metastases in breast cancer
.
J Clin Oncol
2014
;
32
:
1
5
.
45.
Stephan
SB
,
Taber
AM
,
Jileaeva
I
,
Pegues
EP
,
Sentman
CL
,
Stephan
MT
. 
Biopolymer implants enhance the efficacy of adoptive T-cell therapy
.
Nat Biotechnol
2015
;
33
:
97
101
.
46.
Adusumilli
PS
,
Cherkassky
L
,
Villena-Vargas
J
,
Colovos
C
,
Servais
E
,
Plotkin
J
, et al
Regional delivery of mesothelin-targeted CAR T cell therapy generates potent and long-lasting CD4-dependent tumor immunity
.
Sci Transl Med
2014
;
6
:
261ra151
.
47.
Potts
MB
,
Silvestrini
MT
,
Lim
DA
. 
Devices for cell transplantation into the central nervous system: Design considerations and emerging technologies
.
Surg Neurol Int
2013
;
4
Suppl 1
:
S22
30
.
48.
Curley
MD
,
Therrien
VA
,
Cummings
CL
,
Sergent
PA
,
Koulouris
CR
,
Friel
AM
, et al
CD133 expression defines a tumor initiating cell population in primary human ovarian cancer
.
Stem Cells
2009
;
27
:
2875
83
.
49.
Skubitz
AP
,
Taras
EP
,
Boylan
KL
,
Waldron
NN
,
Oh
S
,
Panoskaltsis-Mortari
A
, et al
Targeting CD133 in an in vitro ovarian cancer model reduces ovarian cancer progression
.
Gynecol Oncol
2013
;
130
:
579
87
.
50.
Bertolini
G
,
Roz
L
,
Perego
P
,
Tortoreto
M
,
Fontanella
E
,
Gatti
L
, et al
Highly tumorigenic lung cancer CD133+ cells display stem-like features and are spared by cisplatin treatment
.
Proc Natl Acad Sci U S A
2009
;
106
:
16281
6
.