Novel PSCA targeting scFv-fusion proteins for diagnosis and immunotherapy of prostate cancer

Abstract

Purpose

Despite great progress in the diagnosis and treatment of localized prostate cancer (PCa), there remains a need for new diagnostic markers that can accurately distinguish indolent and aggressive variants. One promising approach is the antibody-based targeting of prostate stem cell antigen (PSCA), which is frequently overexpressed in PCa. Here, we show the construction of a molecular imaging probe comprising a humanized scFv fragment recognizing PSCA genetically fused to an engineered version of the human DNA repair enzyme O6-alkylguanine-DNA alkyltransferase (AGT), the SNAP-tag, enabling specific covalent coupling to various fluorophores for diagnosis of PCa. Furthermore, the recombinant immunotoxin (IT) PSCA(scFv)-ETA′ comprising the PSCA(scFv) and a truncated version of Pseudomonas exotoxin A (PE, ETA′) was generated.

Methods

We analyzed the specific binding and internalization behavior of the molecular imaging probe PSCA(scFv)-SNAP in vitro by flow cytometry and live cell imaging, compared to the corresponding IT PSCA(scFv)-ETA′. The cytotoxic activity of PSCA(scFv)-ETA′ was tested using cell viability assays. Specific binding was confirmed on formalin-fixed paraffin-embedded tissue specimen of early and advanced PCa.

Results

Alexa Fluor^® 647 labeling of PSCA(scFv)-SNAP confirmed selective binding to PSCA, leading to rapid internalization into the target cells. The recombinant IT PSCA(scFv)-ETA′ showed selective binding leading to internalization and efficient elimination of target cells.

Conclusions

Our data demonstrate, for the first time, the specific binding, internalization, and cytotoxicity of a scFv-based fusion protein targeting PSCA. Immunohistochemical staining confirmed the specific ex vivo binding to primary PCa material.

Introduction

Prostate cancer (PCa) is the second most common cancer in men worldwide and the sixth most common cause of cancer-related death in men worldwide (1.1 million cases estimated in 2012). It is the leading cause of cancer mortality in economically-developed countries (903,500 cases estimated for the year 2012).

Current diagnosis of PCa relies on digital rectal examination (DRE), detection of increased serum concentrations of prostate-specific antigen (PSA), magnetic resonance imaging (MRI), and transrectal ultrasound (TRUS)-guided biopsy (Heidenreich et al. 2014a). Although routine PSA screening seems to contribute to a reduced PCa mortality by detecting cancer at an earlier curable stage (Bokhorst et al. 2016), PSA screening alone is insufficient for accurate diagnosis and classification of PCa. The abundance of the soluble antigen also increases in non-malignant disorders of the prostate (Farwell et al. 2007; Madu and Lu 2010). Only for advanced metastatic PCa, molecular imaging agents targeting prostate-specific membrane antigen (PSMA) for radioimmunoscintigraphy like the monoclonal antibody (mAb) conjugate indium-111 capromab pendetide (ProstaScint, Cytogen, Princeton, NJ, USA) (Wu et al. 2014) and positron emission tomography (PET)/computer tomography (CT) [gallium-labeled PSMA ligand (68 Ga-PSMA)] is available (Afshar-Oromieh et al. 2013, 2015). New diagnostic markers that can accurately discriminate between indolent and aggressive variants of PCa are, therefore, required to increase the accuracy of diagnosis.

While watchful waiting/active surveillance is an option especially for older patients with slowly growing organ confined tumors, current curative treatments for patients with localized PCa include radical prostatectomy, radiation therapy, and cryosurgery (Heidenreich et al. 2014a; Valerio et al. 2014). Progressive PCa is treated with chemotherapy or hormone therapy like androgen ablation therapy, which leads to apoptosis of the androgen-dependent tumor cells. However, after an initial response, this therapy fails in a high percentage of patients resulting in recurrent androgen-independent PCa (AIPC). At present, there is no effective therapy for AIPC available. Chemotherapy has mainly palliative effects for patients with hormone refractory PCa, but also shows some survival effects (Chen et al. 2014; Tannock et al. 2004). Therefore, new therapies and therapeutic targets on PCa, especially for late-stage disease and aggressive tumors, are desperately needed (Heidenreich et al. 2014b; Long et al. 2005). Furthermore, specificity of treatment needs to be improved, since chemo- and to a lesser extent radiotherapy are associated with severe side effects due to their unspecific mode of action (DeVita and Chu 2008; Schrama et al. 2006).

Nowadays, 17 tumor-reactive antibodies are FDA approved and widely used in clinical practice and many more are being evaluated in clinical trials (Vacchelli et al. 2015). MAbs in clinical studies for PCa therapy include those approved for established targets on other solid tumors, like anti-epidermal growth factor receptor (EGFR) mAb cetuximab (Erbitux^®, Merck Serono) (Cathomas et al. 2012), and the fully human mAb panitumumab (Vectibix^®, Amgen) (Carey et al. 2012), anti-human epidermal growth factor receptor-2 (HER2) mAb trastuzumab (Herceptin, Roche) (Ziada et al. 2004) as well as the antivascular endothelial growth factor (VEGF) mAb bevacizumab (Avastin^®, Roche) (Small and Oh 2012). A mAb directed against a PCa specific target is the anti-PSMA mAb J591 (Akhtar et al. 2012; Tagawa et al. 2013). Recently, a phase II study with the 177 lutetium radiolabeled J591 in patients with progressive metastatic CRPC resulted in significant PSA declines in 60% of the patients (Vallabhajosula et al. 2016).

Another prostate-specific surface marker that is highly suitable for antibody-based therapy is the prostate stem cell antigen (PSCA), a cysteine-rich 123-aa glycosylphosphatidylinositol-(GPI) anchored membrane glycoprotein of the Thy-1/Ly-6 family, originally identified through an analysis of genes upregulated in the human CaP LAPC-4 xenograft model (Reiter et al. 1998). Because of its 30% homology to stem cell antigen type 2 (SCA-2), it was named PSCA, although it is neither a stem cell marker nor is it uniquely expressed in the prostate (Saeki et al. 2010).

PSCA is overexpressed in more than 80% of local PCa and in 100% of bone metastatic lesions (Gu et al. 2000; Reiter et al. 1998). It is also up-regulated in other cancers including those of bladder and pancreas (Wente et al. 2005). PSCA is a promising marker for late-stage PCa, as PSCA over-expression correlates with advancing tumor stage, grade, and progression to androgen independence (Gu et al. 2000; Han et al. 2004; Lam et al. 2005; Zhigang and Wenlu 2007).

In normal human tissues, PSCA mRNA expression is found in the basal cells and secretory luminal cells of the prostate, with lower expression in placenta and very low expression in kidney and small intestine (Cunha et al. 2003; Reiter et al. 1998; Ross et al. 2002). The limited expression in normal tissues with high expression in a large proportion of human prostate tumors, including metastatic and hormone refractory, makes PSCA an attractive target also for immunotherapy of advanced PCa. As levels of PSCA are increased in the prostate intraepithelial neoplasia (PIN) lesions that subsequently progress to cancer compared to those that do not (Zhigang and Wenlu 2007), it may also be a relevant prognostic maker concerning the aggressiveness.

So far, there are no αPSCA antibodies in clinical evaluation, although αPSCA monoclonal antibodies including hu1G8 have been reported to inhibit tumor growth and prolong the survival of mice bearing human PCa xenografts (Cunha et al. 2003; Reiter et al. 1998; Ross et al. 2002).

Although several mAbs and a few antibody–drug conjugates (ADCs) have been approved for therapeutic use in humans, and are an integral part of routine cancer treatment, their size (150 kDa) as well as the Fc domain can limit tumor penetration. Because of their reduced size, single-chain fragment variable (scFvs) antibody derivatives have favorable characteristics regarding tumor penetration. Moreover, the absence of the Fc region reduces off-target effects to Fc-receptor positive cells (Colcher et al. 1998; Kampmeier et al. 2010; Pavlinkova et al. 1999; Yokota et al. 1992). Recombinant immunotoxins (ITs), generated by fusion of scFvs to bacterial or plant derived toxins, are promising tools for targeted cancer treatment. After binding to tumor-associated cell surface receptors, they are internalized and subsequently induce apoptosis (Antignani and Fitzgerald 2013; Pastan et al. 2007). Another advantage to ADCs is that recombinant expression of these fusion proteins in bacterial and mammalian cells in large quantities without need for further modification reduces costs of production (Ahmad et al. 2012).

To date, there are no PSCA-specific ITs described. Nevertheless, several PSCA(scFv)-based therapeutics have shown encouraging results in preclinical studies. A bispecific scFv with specificities for PSCA and CD8 was able to retarget T cells to the cancer cells finally leading to cell death in vitro (Michalk et al. 2014). A chimeric antigen receptor (CAR) consisting of a PSCA(scFv) and the CD3 ζ chain signaling domain combined with CD28 and OX40 derived costimulatory signaling domains activated PSCA-CAR engineered T cells in response to PSCA-expressing tumor cells leading to delayed tumor growth in preclinical studies (Hillerdal et al. 2014). In addition, recently, the novel chemotherapeutic molecule JDF-12 encapsulated in nanoparticles was successfully targeted to PCa cells by surface functionalization with a scFv that recognizes the PSCA extracellular domain. Targeted delivery resulted in increased anti-tumor effects and reduced systemic toxicity (Fang et al. 2015).

Next to development of specific therapies, development of specific diagnostics is one strategy to further improve survival of PCa patients. Molecular whole body imaging of PCa has been shown successfully with anti-PSCA mAb hu1G8-based affinity-matured minibodies applied in PET of murine PCa xenografts (Knowles et al. 2014; Lepin et al. 2010). Such antibody derivatives can be used for the generation of novel theranostics by application of the SNAP-tag technology. As an engineered version of the human O(6)-alkylguanine-DNA alkyltransferase (hAGT), the SNAP-tag reacts specifically and rapidly with O(6)-benzylguanine (BG)-modified molecules in a defined 1:1-stoichiometry (Kampmeier et al. 2009; Kolberg et al. 2013). Fused to scFv molecules, it mediates efficient protein labeling without affecting the ligand’s binding activity (Hussain et al. 2011, 2013; Kampmeier et al. 2010). Previously, the generation of theranostics by conjugation of IRDye700dx to SNAP-tag fusion proteins (von Felbert et al. 2016) as well as recombinant ADCs by conjugation with BG-modified auristatin has been reported (Woitok et al. 2016), demonstrating the suitability of SNAP-tag technology not only for diagnostic approaches but also for the generation of novel ADCs.

Here, we report the successful derivation of a scFv from the mAb hu1G8-derived affinity-matured minibody variant A2 [based on published sequence (Wu et al. 2010)] and its genetic fusion to the truncated version of Pseudomonas exotoxin A (ETA′), yielding the IT PSCA(scFv)-ETA′, which was characterized in vitro and on primary PCa material. Since PSCA is expressed in nearly all prostate tumors, but is downregulated in cultured PCa cell lines in vitro (Taylor et al. 2012), the PSCA transgenic HEK-PSCA cell line was generated for analysis of PSCA(scFv) fusion proteins. We observed target-specific binding and cytotoxic activity of the IT against the HEK-PSCA cells with IC₅₀ values in the pM range. The scFv was also combined with the SNAP-tag to generate imaging probes for validation of its diagnostic potential in vitro and for potential diagnostic use in vivo (Amoury et al. 2013; Kampmeier et al. 2009, 2010). Binding and internalization of this fusion protein could be demonstrated on PSCA⁺ cells and the functionality was further confirmed by immunohistochemical analysis on primary tumor material. Since the novel IT showed efficient elimination of PSCA-expressing cells, it should be characterized in more detail for its application in anticancer therapy.

Materials and methods

Bacterial strains

Escherichia coli DH5α was used for the amplification of plasmids and E.coli BL21 Star DE3 for the expression of recombinant ITs.

Plasmids, oligonucleotides, and cloning

Plasmids were prepared by alkaline lysis and purified using the NucleoSpin^® Plasmid kit (Macherey–Nagel, Düren, Germany). Restriction fragments and PCR products were separated by agarose gel electrophoresis and extracted using the NucleoSpin^® Gel and PCR Clean-up kit (Macherey–Nagel). Synthetic oligonucleotides were synthesized by MWG (Martinsried, Germany).

We used a modified version of the pSecTAg/HygroB vector named pMS (Stocker et al. 2003) for expression in mammalian cells and a modified version of the pET27b vector (Novagen, Wisconsin, USA) named pMT (Barth 2002; Matthey et al. 1999; Tur et al. 2003) for bacterial expression.

Cloning of mammalian and bacterial expression vectors

The PSCA(scFv) sequence used for the generation of PSCA(scFv) fusion proteins was synthesized by GeneArt (Life Technologies Inc., Regensburg, Germany). This sequence is derived from an affinity-matured minibody (A2) (Lepin et al. 2010; Leyton et al. 2008) generated from the humanized mAb (hu1G8) (Olafsen et al. 2007). PSCA(scFv)-SNAP was constructed by cloning of the scFv sequence via Sfi I/Not I restriction into the pMS vector system containing the N-terminal SNAP-tag sequence (Kampmeier et al. 2009, 2010; Stocker et al. 2003). For mammalian expression, the vector also contains a cytomegalovirus promoter, a zeocin resistance gene used for the selection of transfected HEK-293T cells, an Igκ-leader (Igκ-L) for protein secretion, a tandem Myc-His₆-tag for detection and purification, and an internal ribosomal entry site, which enables the co-translation of the enhanced green fluorescent protein (EGFP) (Kampmeier et al. 2009, 2010; Stocker et al. 2003). Successful cloning was verified by sequencing and control restriction digest.

For bacterial expression, the PSCA(scFv) was inserted into the bacterial expression vector pMT-ETA′ using restriction sites Sfi I and Not I. This vector contains an isopropyl-β-d-1-thiogalactopyranoside (IPTG)-inducible lac operon, a pelB signal peptide, an enterokinase cleavable His₆-tag, and a modified ETA′ sequence (Bruell et al. 2003).

For generation of PSCA transgenic cells, PSCA template DNA [cDNA clone (IRAUp969A02102D)] was obtained from Imagenes (Berlin, Germany). The DNA was amplified with gene-specific primers containing appropriate restriction sites for cloning of PSCA into pMS without Igκ-L (Jorissen et al. 2009). The primers were 5′-PSCA-NheI (5′-ATGAACGCTAGCATGAAGGCTGTGCTGCT-3′), 3′-PSCA-NotI (5′-GATGTATGATTGGCGGCCGCCTATAGCTGGCCGGGTCCCCAGAGCA-3′). Restriction sites are underlined.

Culture and transfection of HEK-293T cells

HEK-293T cells (ATCC: CRL-11268) were cultured in RPMI 1640 (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal calf serum (Biochrom AG, Berlin, Germany), 100 U/ml penicillin, and 100 µg/ml streptomycin (Gibco) in a humidified atmosphere at 37 °C and 5% CO₂.

HEK-293T cells were transfected with 2 μg of plasmid DNA and 3 μL Roti-Fect (Carl Roth GmbH, Karlsruhe, Germany) according to the manufacturer’s instructions. Transfected cells were cultured under selection pressure with 100 μg/ml zeocin (Invivogen, Toulouse, France). For the production of larger amounts of protein, cells were cultured in TripleFlasks™ (VWR, Radnor, Pennsylvania USA). For surface expression of PSCA, HEK293T cells were transfected and cultured according to generate the transgenic HEK-PSCA cells. Single-cell clones of PSCA-tranfected cells were isolated by fluorescence-activated cell sorting (FACS) on an Influx™ cellsorter (Becton & Dickinson, Heidelberg, Germany) using the (PE)-labeled murine αPSCA mAb (αPSCA^PE) (clone 7F5; Santa Cruz, USA) (Morgenroth et al. 2007).

Purification of the recombinant PSCA(scFv)-SNAP fusion protein

The recombinant SNAP-tag fusion protein was purified from cell-free supernatants via the C-terminal His-tag using a 5 mL Ni–NTA Superflow cartridge (Qiagen, Hilden, Germany) on an ÄKTA protein purification system (GE Healthcare Europe GmbH, Freiburg, Germany). Briefly, after equilibration of the cartridge with wash buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10 mM imidazole, pH 8), supernatant containing the His-tagged protein was applied with a low flow rate followed by thorough washing with buffer (10, 30, and 40 mM imidazole). Bound His-tagged protein was eluted using elution buffer (50 mM NaH₂PO₄, 300 mM NaCl, 250 mM imidazole, pH 8) and the elution factions were dialyzed at 4 °C overnight against PBS. After concentration to 0.5–1 mg/mL with a 10-kDa molecular weight cutoff (MWCO) VivaSpin column (Sigma-Aldrich, Taufkirchen, Germany), protein was stored in aliquots at −20 °C.

Periplasmic expression and purification of the IT PSCA(scFv)-ETA′

After heat-shock transformation, the recombinant IT PSCA(scFv)-ETA′ was expressed in the periplasm of E. coli BL21 Star (DE3) (Novagen) under osmotic stress in the presence of compatible solutes as previously described (Barth et al. 2000a; Barth et al. 2000b). The periplasmic fraction was recovered and EDTA was removed by overnight dialysis against phosphate-buffered saline (PBS, pH 7.4). The recombinant IT was purified from the periplasmic fraction by IMAC purification as described above. Preparative SEC was carried out as described elsewhere (Hristodorov et al. 2014), elution fractions containing the IT were combined, concentrated to 0.5–1 mg/mL with a 30-kDa molecular weight cutoff (MWCO) VivaSpin column (Sigma-Aldrich), and stored in aliquots at −20 °C.

SDS-PAGE and western blotting

To confirm protein integrity, proteins were visualized on Coomassie Brilliant Blue-stained sodium dodecyl-polyacrylamide gels (SDS-PAGE) and detected on corresponding western blots by anti-His₆ mAb in case of PSCA(scFv)-SNAP, whereas PSCA(scFv)-ETA′ was detected using αETA′, 1:5000, 230 ng/μl diluted in PBS (Barth et al. 1998). Bound antibody was detected with alkaline phosphatase (AP)-conjugated monoclonal goat anti-mouse-IgG (GaM^AP) (Dianova, Hamburg, Germany) diluted 1:5000-fold. Bound detection antibody was detected using nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP) substrate (Life Technologies, Thermo Fisher Scientific, Waltham, MA, USA), and the protein size was compared to a pre-stained broad-range protein marker (New England BioLabs (NEB), Schwalbach, Germany). The protein concentration was determined by densitometry after Coomassie Brilliant Blue gel staining compared to bovine serum albumin (BSA) standards using AIDA-Advanced Image Data Analyzer Version 4.27.039 (Raytest Isotypenmessgeräte GmbH, Straubenhardt, Germany).

Protein labeling with BG derivatives of organic fluorophores

Purified SNAP-tag fusion proteins were conjugated with BG-modified fluorophores (NEB) by incubation with a 1.5-fold molar excess of fluorophore for 1 h at room temperature or overnight at 4 °C in the dark. Residual fluorophore was removed by gel filtration chromatography using PD10 columns (GE Healthcare) or dialysis. Alexa Fluor^® 647 (NEB) conjugated protein was visualized after SDS-PAGE with the Maestro™ in vivo fluorescence imaging system (Cambridge Research & Instrumentation Inc. (CRi), Woburn, MA, USA) using the appropriate filter set.

Flow cytometry to determine binding, affinity and internalization of fusion proteins

The cell-binding activity of PSCA(scFv) fusion proteins was evaluated using the PSCA⁺ cell line HEK-PSCA and HEK-293T cells as a control. Approximately 4 × 10⁵ cells were incubated with 1 μg of fusion protein in 50 μL of MACS buffer (PBS + 0.5% (w/v) BSA and 2 mM EDTA) for 30 min on ice. Proteins were detected with His₆ mAb diluted 1:1000 followed by PE-labeled goat anti-mouse IgG (H + L) F(ab′)₂ (GaM^PE) (Dianova, Hamburg, Germany) diluted 1:200 in 50 µl MACS buffer. As a positive control, cells were incubated with the αPSCA^PE antibody described above. In between staining steps, cells were washed twice in 1.8 mL PBS in a conventional cell washer (Baxter, USA). Fluorescence intensities were measured using a BD FACSCalibur flow cytometer (Becton & Dickinson) and evaluated using the Cyflogic software (CyFlo Ltd, Turku, Finland).

Protein internalization was evaluated by flow cytometry as described by Cizeau et al. (2009). Briefly, 4 × 10⁵ HEK-PSCA cells were incubated with 1 μg of PSCA(scFv)-ETA′ for 30 min on ice. Subsequently, cells were washed with PBS to remove unbound protein and then incubated in culture medium for 30, 60, 120, or 180 min at 37 °C. The non-internalization control sample was kept at 4 °C and set to 100% fluorescence intensity in the final evaluation. All samples were washed with PBS as described above and the remaining surface-bound protein was detected with mAb His₆ and GαM^PE in flow cytometry.

The number of PSCA molecules expressed on the cell surface was determined using the QIFIKIT^® (Dako, Hamburg, Germany). Flow cytometry analysis was carried out according to the manufacturer’s protocol using the αPSCA^PE antibody described above.

For the determination of the affinity constant K_d, different concentrations (up to 830 nM) of PSCA(scFv) SNAP were incubated with target cells as described above. Mean fluorescence intensities (MFIs) were evaluated via Cyflogic and compared by non-linear-regression determinations in GraphPad Prism 4.0 (La Jolla, CA, USA).

Confocal microscopy of PSCA(scFv)-SNAP internalization

To visualize the internalization of the scFv-SNAP probes, these were coupled to BG-Alexa Fluor^® 647 (NEB). To improve attachment of the cells, 96-well black µClear^®cell culture microplates (Greiner-Bio-One GmbH, Frickenhausen, Germany) were pre-coated with 5 µg/cm² fibronectin (Sigma-Aldrich) in 25 µl/well medium for 2 h at 37 °C. Wells were washed once with medium, 7.5 × 10³ HEK-293T or HEK-PSCA cells seeded per well and incubated over night at 37 °C and 5% CO₂. Before staining with the BG-Alexa Fluor^® 647-labeled fusion proteins, cells were starved for 3 h in 50 µl RPMI with 0.5% FCS (v/v) to reduce unspecific effects. Cells were then incubated with 1 µg of PSCA(scFv)-SNAP fusion protein for 1 h at 37 °C in 50 µl RPMI with 0.5% FCS (v/v). The cells were washed with RPMI and further incubated for 30 or 150 min, respectively. Binding was analyzed after 30 min at 4 °C. The nuclei were counterstained with Bisbenzimide Hoechst 33342 (Sigma-Aldrich) and confocal microscopy images were taken with an Opera System (Perkin Elmer, USA).

Colorimetric XTT cell viability assay

The cytotoxic activity of the IT against HEK-PSCA cells was measured using the XTT-based cell proliferation kit II (Roche, Mannheim Germany) according to the manufacturer’s instructions. Briefly, serially diluted PSCA(scFv)-ETA′ in 96-well plates was mixed with 2 × 10⁴ HEK-PSCA or HEK293T control cells in 100 μl final volume. Untreated cells, PSCA(scFv)-SNAP, H22-ETA′, and camptothecin were used as controls. XTT labeling reagent and N-methyl electron coupling reagent were mixed in 1:0.02 molar ratio. Thereafter, 50 µL of the XTT detection solution was added to the cells and incubated for 4 h at the standard cell culture conditions. Cell viability was determined by colorimetric analysis of the reduction of XTT to formazan at 450 nm absorbance wavelength and 630 nm reference wavelength using an Epoch Microplate Spectrophotometer (BioTek Instruments GmbH, Bad Friedrichshall, Germany). The required IT concentration to achieve a 50% reduction of cell viability (IC₅₀) relative to the untreated control cells was calculated using GraphPad prism v5 (GraphPad Software, La Jolla, CA, USA). All experiments were carried out at least three times in triplicates.

Immunohistochemistry

Paraffin-embedded PCa sections were kindly provided by Prof. Gattenlöhner (Pathology, Justus-Liebig University Hospital, Gießen, Germany) and were obtained during routine clinical practice in accordance with the principles of the Declaration of Helsinki. PCa biopsy specimen of four patients with early-to-intermediate stage (Gleason scores 6–7) is moderately differentiated, whereas prostatic chips obtained by transurethral resection of the prostate (TURP) of one patient (Gleason score 9) represent poorly differentiated advanced PCa. The slides were rehydrated and dewaxed as described elsewhere (Niesen et al. 2015a). Sections were circled using a Dako-pen (Dako). The slides were incubated overnight at 4 °C with either the commercial rabbit αPSCA antibody (Acris Antibodies GmbH, Herford, Germany) diluted 1:75, or recombinant PSCA(scFv) SNAP or PSCA(scFv)-ETA′ fusion proteins. The slides were then washed three times for 5 min with PBST. For the SNAP fusion proteins, the αSNAP mAb⁶⁹ diluted 1:400 and for the IT the αETA′ antibody diluted 1:30 in PBS were added and incubated for 8 h at 15 °C in 1% (v/v) goat serum in PBS. Afterwards, the slides were washed three times for 5 min with PBST and incubated overnight with GαM^AP (Dianova) or AP-conjugated goat anti-rabbit (GαR^AP) (Dianova) diluted 1:400 in 1% goat serum in PBS. After washing twice in PBS/Tween and once in Tris–HCl (0.1 M, pH 8.5), AP activity was detected using New Fuchsin staining. Subsequent counterstaining with hematoxylin and eosin (H&E) was performed as described elsewhere (Niesen et al. 2015a). Tissue sections were analyzed using an Olympus BX 41 light microscope (Olympus, Hamburg, Germany) and ProgRes^® CapturePro Version 2.8.8 software (Jenoptik Optical Systems, Jena).

Results

Generation of PSCA⁺ HEK293T cells

Since PSCA expression is downregulated in cell culture of PCa cell lines (Taylor et al. 2012), HEK-293T cells were transfected with pMS-PSCA. Cell surface expression of PSCA was confirmed by analyzing the binding activity of the αPSCA^PE mAb. Flow cytometric analysis with the QIFIKIT^® (Dako) revealed that PSCA was expressed at high levels on the surface of HEK-PSCA cells (16285 ± 2400 receptors/cell), whereas HEK-293T cells showed almost no expression (142 ± 138 receptors/cell) (Table 1).

Table 1.

IC₅₀ values (nM) of PSCA(scFv)-ETA′ on HEK-PSCA clones with different PSCA expression levels

	HEK293T	HEK-PSCA clone 1	HEK-PSCA clone 2
PSCA(scFv)-ETA′	109 ± 17.7 (n = 3)	0.026 ± 0.04 (n = 4)	0.36 ± 0.17 (n = 5)
PSCA(scFv)-SNAP	ND	ND	ND
PSCA expression level	142 ± 138 (n = 3)	16,285 ± 2400 (n = 3)	12,124 (n = 2)

The IC₅₀ values as means ± SD indicate the concentrations of the ITs required to achieve a 50% reduction in protein synthesis relative to untreated control cells, as shown in Fig. 4. The PSCA(scFv)-SNAP control did not show any cytotoxicity. Only a marginal effect of the IT could be measured (>100 nM) on the HEK-293T cell line, which is used as PSCA⁻ control cell line. PSCA expression levels are presented as mean number of surface antigen ± SD

ND not determined

Expression and purification of the IT and SNAP‑tag fusion proteins

We successfully cloned and periplasmically expressed the ETA′-fusion protein PSCA(scFv)-ETA′ in E. coli BL21 Star (DE3) cells. The imaging probe PSCA(scFv)-SNAP was expressed in HEK-293T cells (Fig. 1a). All proteins were successfully purified under native conditions by IMAC using the His₆-tag. In case of the IT, purification was performed on Ni–NTA Sepharose gravity flow columns at 4 °C. Remaining impurities and degradation products of PSCA(scFv)-ETA′ were removed by SEC yielding approx. 10 mg protein/L bacterial culture. Western blot with αETA′ mAb showed an electrophoretic mobility corresponding to the mass of 70 kDa predicted from the amino acid sequence (Fig. 1c). The yield of PSCA(scFv)-SNAP was 2.2 mg/L culture supernatant with 94% purity. Successful BG-Alexa Fluor^®647 labeling of PSCA(scFv)-SNAP was confirmed by Maestro™ in vivo fluorescence imaging system after separation by SDS-PAGE (Fig. 1b). Binding analysis showed binding of PSCA(scFv)-SNAP-647 on HEK293T-PSCA cells but only minimal background staining on HEK293T cells (data not shown).

Fig. 1.

Expression and purification of αPSCA fusion proteins. a SDS–polyacrylamide gel stained with Coomassie Brilliant Blue (lane 1) and corresponding western blot (lane 2) showing PSCA(scFv)-SNAP purified by FPLC-IMAC and detected with His6 mAb and a goat anti-mouse AP-labeled secondary antibody (GαMAP). b Confirmation of protein labeling with BG-Alexa Fluor^® 647 by visualization of the fluorescence signal after SDS-PAGE using the deep red filter set in CRi Maestro (lane 2) and afterwards staining with Coomassie Brilliant Blue (lane 1). c SDS–polyacrylamide gels stained as above showing purified PSCA(scFv)-ETA′ before (lane 1) and after (lane 2) size exclusion and the corresponding western blot (lanes 3 and 4) detected with the αETA′ and GαMAP secondary antibody. The anticipated protein sizes are ~46 kDa for PSCA(scFv)-SNAP and ~70 kDa for PSCA(scFv)-ETAʹ

In vitro binding analysis of the PSCA(scFv) fusion proteins

As mentioned above, PSCA expression on PCa cell lines is described to be downregulated in culture (Taylor et al. 2012) and in line with that none of the four different PCa cell lines available in our lab (PC-3, LNCaP, LNCaP-C4-2 and 22Rv1) showed significant PSCA expression in flow cytometry [QIFIKIT^®, Dako (data not shown)]. However, upregulation of PSCA expression in vivo could be shown for LNCaP-C4-2 cells in murine xenografts (Taylor et al. 2012). Flow cytometry analysis of the generated PSCA⁺ HEK-PSCA sublines confirmed the specific binding activity of the PSCA(scFv) fusion proteins, whereas no binding to the PSCA⁻ parental cell line HEK293T was detectable (Fig. 2a). The relative MFI values were ~75 to 240 times higher on HEK-PSCA cells compared to the control cell line HEK-293T. A medium binding affinity of 80 nM for PSCA(scFv)-SNAP was determined flow cytometrically (Fig. 2b).

Fig. 2.

Cell-binding activity and affinity of PSCA(scFv) fusion proteins. a Specific cell binding activity of PSCA(scFv)-SNAP (1, 4) and of the recombinant IT PSCA(scFv)-ETA′ (2, 5) was analyzed by flow cytometry compared to αPSCAPE antibody 7F5 (3, 6). Binding was demonstrated on PSCA + HEK-PSCA cells (1, 2, 3), and PSCA-HEK-293T cells (4, 5, 6) were used as negative control. Bound proteins were detected using His6 mAb and PE-labeled goat anti-mouse IgG (H + L) F(ab′)2 (GαMPE), FL-2 fluorescence channel/PE. The filled gray curve shows the background control, and binding of the fusion proteins is shown as solid black curves. b For the determination of the affinity constant Kd, different concentrations (up to 830 nM) of PSCA(scFv)-SNAP were incubated with HEK-PSCA cells. Geometric mean fluorescence intensities (MFI) (less the background control) were evaluated via Cyflogic and the Kd value was estimated by non-linear-regression determinations in GraphPad Prism (exemplarily shown for one experiment in triplicates, data are shown as mean ± SD)

Internalization of PSCA(scFv) IT and PSCA(scFv)-SNAP

Internalization of PSCA(scFv)-ETA′ and PSCA(scFv)-SNAP by PSCA⁺ HEK-PSCA cells was analyzed using flow cytometry-based internalization assays (Fig. 3d). We found that 30% of the fusion proteins were internalized within 10 min. After incubation for 180 min, approximately 70% of both fusion proteins were internalized.

Fig. 3.

Internalization of PSCA(scFv) IT and PSCA(scFv)-SNAP. The internalization of PSCA(scFv)-SNAP labeled with BG-Alexa Fluor^® 647 visualized in PSCA + HEK-PSCA cells (a, b) by confocal microscopy and Opera^® high-content screening. Nuclei were counterstained with Hoechst 33342. The internalization was measured after incubation for 90 min at 37 °C (a). Binding is shown after incubation for 90 min at 4 °C (b, non-internalization control). PSCA-HEK-293T cells (c) served as negative control. Each experiment was carried out at least three times. Scale bar 20 μm. d Efficiency of internalization for PSCA(scFv)-ETA′ and PSCA(scFv)-SNAP was determined by flow cytometry. The geometric mean fluorescence intensity (percentage MFI) is shown at different timepoints. Cells incubated at 4 °C are set to 100% in the evaluation. Surface bound fusion proteins were detected with His6 mAb and a GαMPE secondary antibody

Confocal microscopy of Alexa Fluor^®647-labeled PSCA(scFv)-SNAP bound to HEK-PSCA cells verified that a significant amount of protein was internalized at 37 °C within 30–60 min, leading to intracellular accumulation of the signal (Fig. 3a). As expected, no internalization was observed at 4 °C (Fig. 3b). No detectable protein was taken up by the PSCA⁻ cell line HEK-293T (Fig. 3c).

Elimination of PSCA⁺ cells by PSCA(scFv)-ETA′

The colorimetric XTT viability assay confirmed that incubation of PSCA⁺ target cells with increasing concentrations of the IT for 72 h resulted in decreased viability but not in the control cell line HEK-293T (Fig. 4). The IC₅₀ values are in the range of 26 ± 4 pM, as summarized in Table 1. Unspecific cytotoxicity of PSCA(scFv)-SNAP against the PSCA⁺ cell lines was not observed.

Fig. 4.

Cytotoxicity of PSCA(scFv)-ETA′. Cytotoxicity of PSCA(scFv)-ETA′ was assessed using an XTT assay. The cell lines HEK-PSCA (a, c) and HEK-293T (b) were incubated with serial dilutions of the sterile IT (a, b) or PSCA(scFv)-SNAP (c) as negative control. The concentration required to achieve a 50% reduction in protein synthesis (IC50) relative to untreated control cells was calculated using the GraphPad Prism software. Data are shown as mean ± SDs from three independent experiments carried out in triplicate

Immunohistochemical staining of primary tumor tissue with PSCA(scFv) fusion proteins

To confirm binding to primary tumor tissue, the PSCA(scFv) fusion proteins were used for immunohistochemical staining of formalin-fixed and paraffin-embedded (FFPE) PCa biopsy sections of early-to-intermediate PCa and TURP prostatic chips of advanced PCa (exemplarily shown in Fig. 5). PSCA(scFv)-SNAP was analyzed on three and PSCA(scFv)-ETA′ on two different early-to-intermediate PCa specimen. All analyses were also performed on advanced PCa specimen of one patient. Hematoxylin and eosin staining was used for histological survey staining of the tissue. Using a commercial αPSCA antibody (Acris antibodies) in combination with New Fuchsin staining resulted in specific staining of basal as well as luminal PCa cells in all tissue sections analyzed (Fig. 5a, e). The PSCA-specific binding of the IT (Fig. 5g) and PSCA-SNAP (Fig. 5b, f) fusion protein was also confirmed by New Fuchsin staining. No signal was detected in the negative controls incubated with the detection antibodies in the absence of the fusion proteins (Fig. 5c, d, h). The PSCA(scFv) fusion proteins were essentially staining the same cells as the commercial αPSCA antibody. In agreement with the specific binding seen in flow cytometry, both PSCA(scFv) fusion proteins were able to specifically recognize primary cells from all PCa biopsies tested. Weaker signals were detected for PSCA(scFv)-SNAP on PCa specimen of an intermediate stage, whereas stronger signals were detected on TURP prostatic chips representing advanced PCa specimen.

Fig. 5.

Immunohistochemical staining of primary tumor tissue with PSCA(scFv) fusion proteins. Representative images of the different FFPE tumor sections stained with H&E. Early (a–d) and advanced (e–h) PCa specimen of tumor tissue biopsies are shown. Specific binding of commercial rabbit αPSCA antibody (a, e) detected with GαRAP as well as specific binding of αPSCA-SNAP (b, f) and of αPSCA-ETA′ (g) followed by mAbs αSNAP or αETA, respectively, detected with GαMAP secondary Ab stained with red New Fuchsin substrate. The detection antibodies alone were used as a control [GaRAP (c), αSNAP + GαMAP (d), TC-1 + GαMAP (h)]. Objective lens × 20, scale bar 50 μm

Discussion

The specific and direct elimination of tumor cells has become increasingly important for targeted cancer therapy. More than a dozen mAbs are currently approved by the FDA for different oncologic indications with many more undergoing clinical development (Reichert 2014; Scott et al. 2012; Vacchelli et al. 2015). However, the administration of a single therapeutic mAb to patients with solid tumors often results in low efficacy, whereas combination of mAbs with the standard chemotherapy has a more beneficial impact (Sheng et al. 2015). ADCs have been developed to combine the targeting specificity of mAbs with the cell-killing activity of a cytotoxic payload (Panowksi et al. 2014). In addition, although only two ADCs have been approved by the FDA, many more are currently being evaluated with encouraging results in clinical studies (Feld et al. 2013). Another alternative are recombinant ITs composed of a tumor-specific antibody genetically fused to an apoptosis inducing enzyme eliciting direct killing of tumor cells in very low concentrations following internalization (Becker and Benhar 2012; Pastan et al. 2007). Compared to mAbs, scFv-based ITs show improved penetration of solid tumors due to their smaller size (Colcher et al. 1998; Yokota et al. 1992). This in combination with their straightforward production process makes ITs promising candidates for targeted cancer therapy (Becker and Benhar 2012; Pastan et al. 2007).

For treatment of PCa, no mAbs or ADCs have been approved so far. Nevertheless, various forms of PCa specific immunotherapies are being investigated in clinical trials including mAbs (Cathomas et al. 2012; Small and Oh 2012; Ziada et al. 2004), dendritic cell-based vaccines, immune checkpoint inhibitors, viral-based vectors, and whole cell-based vaccines (Tse et al. 2014). Sipuleucel-T (APC8015), an immunotherapeutic agent consisting of antigen-presenting cells, loaded ex vivo with a recombinant fusion protein of prostatic acid phosphatase linked to granulocyte–macrophage colony-stimulating factor, demonstrated an improvement in overall survival in phase III, placebo-controlled trials (Pieczonka et al. 2015; Sheikh et al. 2013). Other candidates for immunotherapy of PCa are PSMA (Kuroda et al. 2010; Wolf et al. 2010) and CD3-PSMA (Elsasser-Beile et al. 2006) as well as EGFR-specific ITs (Niesen et al. 2015b). Wolf et al. demonstrated high and specific toxicity of the αPSMA scFv immunotoxin D7-PE40 for PSMA-expressing PCa cells in vitro and in vivo (Wolf et al. 2010).

The limited expression of PSCA in normal tissues with high expression of a large proportion of human prostate tumors, including metastatic and hormone refractory, makes PSCA another attractive target for diagnosis and immunotherapy of advanced PCa (Gu et al. 2000; Reiter et al. 1998; Zhigang and Wenlu 2007).

We, therefore, developed a PSCA-specific SNAP-tag fusion protein for diagnostic purposes and a recombinant IT for targeted elimination of PSCA⁺ tumor cells and evaluated their activity on PSCA⁺ cells. Since expression of PSCA is downregulated in cultured PCa cell lines (Taylor et al. 2012), the PSCA⁺ transgenic cell line HEK-PSCA was generated and homogenously expressing subclones were successfully established by single-cell sorting for characterization of the PSCA(scFv) fusion proteins.

Flow cytometric analysis of PSCA(scFv)-SNAP on HEK-PSCA cells revealed an affinity constant of 80 nM for PSCA(scFv)-SNAP (Fig. 2b) indicating specific, medium-affinity binding to the receptor, whereas the original hu1G8 antibody binds to PSCA with a higher affinity (K _d = 1 nM) ³² in comparison with the affinity-matured minibody variant A2 (K _d = 11 nM) (Lepin et al. 2010). Confocal microscopy of PSCA(scFv)-SNAP coupled to BG-Alexa Fluor^® 647 further confirmed binding and subsequent efficient internalization (Fig. 3a).

The suitability of scFv-SNAP-tag fusions proteins for binding and internalization studies in vitro and tumor targeting studies in vivo has been demonstrated in several previous studies in our working group (Amoury et al. 2013; Kampmeier et al. 2009, 2010). In agreement with their specific binding characteristics on PSCA⁺ cells in vitro, PSCA(scFv)-ETA′ and PSCA(scFv)-SNAP were able to bind specifically to primary tumor material from PCa patients, both in early and advanced tumors. These data confirm that PSCA might be a promising target for PCa diagnosis. Besides immunhistochemical staining of patient biopsies during routine laboratory examinations, PSCA could be a potential target for molecular imaging approaches. Preclinical studies with radiolabeled hu1G8 parental antibody and corresponding minibody demonstrated their suitability for tumor detection by PET. Due to its smaller size and the resulting rapid blood clearance, the minibody showed a favorable pharmacokinetic profile for tumor imaging (Leyton et al. 2008). Other studies with different scFv-based diagnostics confirmed this data, showing that high tissue penetration and fast blood clearance result in an optimal tumor to background fluorescence after shorter circulation time and lower exposure of healthy tissue compared to full size mAbs (Robinson et al. 2005; Sundaresan et al. 2003). PSCA(scFv)-SNAP coupled to near-infrared (NIR) -dyes is a promising candidate for in vivo optical imaging of PCa or could be a tool in imaging-guided surgical approaches (Gioux et al. 2010; Sonn et al. 2016).

To generate the IT PSCA(scFv)-ETA′, we fused the well-characterized truncated bacterial toxin ETAʹ to a scFv derived from an affinity-matured minibody (A2) (Lepin et al. 2010; Leyton et al. 2008), generated from the humanized mAb (hu1G8) (Olafsen et al. 2007). Bacterial expression and purification by IMAC and SEC yielded 10 mg IT per L bacterial culture (Fig. 1c), which is comparable to or higher than yields reported for other recombinant scFv-ETA proteins (Singh et al. 2007, 2008; Tur et al. 2003). Specific binding to the PSCA⁺ cell line was confirmed (Fig. 2a). Flow cytometric analysis of protein internalization revealed that PSCA⁺ HEK-PSCA cells internalized 70% of surface-bound PSCA(scFv)-ETA′ within 180 min (Fig. 3b). Similar internalization times were described for another ETA-based IT (Cizeau et al. 2009). Next to high specificity, this fast and efficient uptake by target cells is one important prerequisite for activity of the IT.

We found that PSCA(scFv)-ETA′ showed specific and potent cytotoxicity against the two genetically modified PSCA⁺ cell clones with IC₅₀ values of approximately 26 pM (Table 1). Lower IC₅₀ values were achieved against the subline expressing the highest level of PSCA (Table 1). These IC₅₀ values′ are comparable to or even lower than those of similar ITs that have been characterized by others (Bang et al. 2005; Barth et al. 2000b; Bruell et al. 2003; Schwemmlein et al. 2006; Tur et al. 2003). Binding to primary tumor material from PCa specimen together with its high cytotoxicity makes PSCA(scFv)-ETA′ a promising candidate IT for PCa immunotherapy. Several recombinant scFv-ETA-based ITs, some of them targeting members of the EGFR family, have been evaluated in phase I trials with varying results (Becker and Benhar 2012; Hassan et al. 2014; Kreitman et al. 2012; Schrama et al. 2006).

These usually were equipped with the truncated versions PE40, lacking the natural binding domain Ia of the toxin, or PE38, additionally lacking a large part of domain Ib. While promising outcomes were primarily seen for hematologic malignancies, treatment of solid tumors remains challenging. In addition, vascular leak syndrome, hepatotoxicity, and kidney toxicity were observed in some cases (Weidle et al. 2014). However, immunogenicity of the bacterial protein is a limiting factor for its clinical application (Kreitman et al. 2012; Weidle et al. 2012). To overcome this problem, ETA-based ITs with drastically reduced immunogenicity were generated by removal of B- and T-cell epitopes. Immunogenicity was further reduced by removing a large part of PE38 domain II resulting in an IT, where the scFv is connected to domain III (PE24) by a furin cleavage site (FCS). These PE24-based ITs showed reduced off-target effects some of them also a longer serum half-life (Kaplan et al. 2016).

Nevertheless, ITs based on the truncated ETA versions are still immunogenic. A fully human cytolytic fusion protein (hCFP) based on human effector molecules could, therefore, further increase safety of targeted cancer therapy. Promising results using hCFPs have already been published (Cao et al. 2013; Weidle et al. 2012). We have previously shown that hCFPs based on granzyme B and angiogenin or microtubule-associated tau protein can efficiently eliminate human tumor cells in vitro and in vivo (Cremer et al. 2015; Hehmann-Titt et al. 2013; Stahnke et al. 2008). By use of the well characterized and highly effective bacterial toxic effector domain ETA′, we have generated and characterized the first PSCA-specific IT and demonstrated its functionality. Thus, we could demonstrate the suitability of PSCA as a target for recombinant ITs and created a basis for continuous development of improved ITs and hCFPs. The corresponding SNAP-tag imaging probes confirmed efficient binding and internalization of the hu1G8 derived scFv and can be applied in development of analogous imaging constructs for further in vivo-imaging approaches. Thus, both PSCA(scFv) fusion proteins proved their potential for future development of diagnostics and targeted therapy for PCa and other PSCA-positive malignancies.

Compliance with ethical standards

Conflict of interest

The authors declare no conflicts of interest.

Ethical standard

In accordance with the Helsinki Declaration of 1975, primary tissue samples were obtained during routine clinical practice at the University Hospital Giessen approved by the appropriate ethics committee.