ANTICANCER RESEARCH 30: 1573-1578 (2010)

Generation of an Antibody against the Inhibitor KEPI and Characterization of the Epitope

KATJANA DASKALOW1,2*, PRISCA BOISGUERIN3, BURKHARD JANDRIG2, FRANK K.H. VAN LANDEGHEM5, RUDOLF VOLKMER3, BURKHARD MICHEEL1 and JÖRG A. SCHENK1,4

1Institute of Biochemistry and Biology, University of Potsdam, Golm, Germany; 2Max-Delbrück-Center for Molecular Medicine (MDC), Berlin, Germany; 3Institute of Medical Immunology, Charité - Universitätsmedizin, Berlin, Germany; 4UP Transfer GmbH, Hybrotec, Potsdam, Germany; 5Institute of Neuropathology, University of Bonn Medical Center, Bonn, Germany

Abstract. A monoclonal antibody against the potential heart and skeletal muscle (1). The of human KEPI tumor suppressor kinase-enhanced protein phosphatase 1 consists of four exons and is located on 6q24- (PP1) inhibitor KEPI (PPP1R14C) was generated and q25. KEPI expression is regulated by morphine, and characterized. Human KEPI was expressed in Escherichia phosphorylated KEPI inhibits PP1, which modulates the coli and used to immunize Balb/c mice. Using hybridoma physiological activities of several other proteins (1). technology, one clone, G18AF8, was isolated producing Recently, a deletion in this region in a patient was found to antibodies which bound specifically to the KEPI protein in cause growth failure, cardiac septal defect, thin upper lip and ELISA, immunoblotting and flow cytometry. The antibody asymmetric dysmorphic ears (2). A frequent loss of was also successfully applied to stain KEPI protein in heterozygosity (LOH) in 6q23-q25 in breast and other types paraffin sections of human brain. The epitope was mapped of cancer suggested the presence of involved in tumor using peptide array technology and confirmed as development. Microcell-mediated chromosome transfer GARVFFQSPR. This corresponds to the N-terminal region studies showed that the introduction of a normal human of KEPI. Amino acid substitution analysis revealed that two fragment encompassing the region 6q23.3- residues, F and Q, are essential for binding. Affinity of q25 can suppress the neoplastic phenotype of breast cancer binding was determined by competitive ELISA as 1 μM. In cell lines (3). In microarray studies on melanoma cell lines, Western blot assays testing G18AF8 antibody on brain the KEPI gene was found to be highly methylated and samples of several species, reactivity with hamster, rat and therefore down-regulated compared to normal cells (4). chicken samples was found, suggesting a broad homology of Expression studies of KEPI in breast cancer cell lines, breast this KEPI epitope in vertebrates. This antibody could be used tumors, and metastases by reverse transcriptase-polymerase in expression studies at the protein level e.g. in tumor tissues. chain reaction (RT-PCR) showed a reduced or complete loss of KEPI expression (5). Therefore, expression studies on the The human kinase-enhanced protein phosphatase 1 (PP1) protein level would help to identify KEPI as a tumor inhibitor KEPI, also known as protein phosphatase 1 suppressor protein. We describe here the generation of a regulatory subunit 14C (PPP1R14C) is expressed in brain, KEPI-specific monoclonal antibody, the characterization of the epitope bound by this antibody and the use of the antibody to identify the KEPI protein in tissue samples.

*Present address: AJ Innuscreen GmbH, Berlin, Germany. Materials and Methods

Correspondence to: Jörg A. Schenk, UP Transfer GmbH, Hybrotec, Cloning of KEPI cDNA, protein production and purification. The c/o Fraunhofer Institute for Biomedical Engineering, Am cDNA of kinase C-enhanced PP1 inhibitor (KEPI) (GenBank Mühlenberg 13, D-14476 Potsdam, OT Golm, Germany. Tel: +49 AL096708) was cloned in pET-30 Xa/LIC and pET-32 Xa/LIC 33158187231, Fax: +49 3319771143, e-mail: joerg.schenk@up- vectors (Novagen, Schwalbach, Germany) according to the transfer.de manufacturer’s protocol. Escherichia coli BL21-Gold(DE3)pLysS (Stratagene, Amsterdam, the Netherlands) were transformed with Key Words: Hybridoma, monoclonal, KEPI, peptide array, the generated plasmids by heat shock. For induction of KEPI fusion substitutional analysis. protein, an overnight culture was grown (37˚C, 250 rpm) in LB

0250-7005/2010 $2.00+.40 1573 ANTICANCER RESEARCH 30: 1573-1578 (2010) media containing appropriate antibiotic to an optical density according to standard protocols. Stable cell lines were generated by OD680nm of 0.4-0.6 before induction with isopropyl-thio-beta-D- selection in 600 μg/ml G418 containing medium for approximately 2 galactopyranoside (IPTG) to a final concentration of 100 μM. weeks. For protein detection in Western blot, the cells were washed Maximum concentration of pET30 and pET32 fusion protein was twice in phosphate-buffered saline (PBS) and lysed in 300 mM NaCl, obtained after 16 h and 4 h, respectively. The osmotic shock fraction 50 mM Tris-HCl pH 7.6, 0.5% Triton X-100 and 1 mM of whole cell extract (OS) was isolated from pelleted bacteria by phenylmethylsulfonylfluoride (PMSF). FACS analysis was performed adding 5 mM MgSO4 (1/5 volume of the culture) and extracting the on a FACScalibur (Becton-Dickinson, Heidelberg, Germany). Cells clarified supernatant by centrifugation for at least 30 min. OS of were incubated with 4% formaldehyde for 10 min and intracellular non-transformed E. coli were used as control. The KEPI protein was staining was performed with 10 μg/ml G18AF8 or murine anti-GFP purified using TALONspin columns (BD Biosciences, Heidelberg, (1 μg/ml, Roche Diagnostics, Penzberg, Germany), followed by Germany) according to the manufacturer’s instructions. For biotinylated goat anti-mouse-Ig (Dianova, Hamburg, Germany) and mammalian expression of KEPI and KEPI-GFP fusion proteins, streptavidin-phycoerythrin (Pharmingen Biosciences, San Diego, CA, plasmids pEGFP-C2 (Clontech, Heidelberg, Germany) or pcDNA3.1 USA). All reagents were diluted in 5% (w/v) saponin. (Invitrogen, Heidelberg, Germany) were used. Immunohistochemical and immunofluorescence examinations. The Generation and characterization of monoclonal antibodies. study was performed with approval of the Institutional Review Board Hybridoma technology was applied for the generation of murine of the Charité, in accordance with Berlin law. Tissue samples were monoclonal KEPI-specific antibodies. Female Balb/c mice were obtained from ten female patients with an intracranial metastasis of immunized three times with unpurified heat-denatured KEPI pET30 a known ductal mamma carcinoma. They were fixed in buffered 4% fusion protein. The booster immunization was performed with formaldehyde overnight and paraffin embedded. Four-μm paraffin purified KEPI pET32 fusion protein. Five days later, electrofusion of sections were used for hematoxylin and eosin staining, alcian blue- spleen cells with myeloma cells (Sp2/0, ATCC CRL-1581) in the PAS reaction and immunohistochemical staining with antibodies presence of polyethylene glycol 8000 was performed as described against cytokeratins (Lu5; DAKO, Hamburg, Germany) and Ki-67 elsewhere (6). Selected hybrids were cultivated in RPMI-1640 (MIB-1; DAKO) to identify the tumor region. According to a medium (containing 10% (v/v) fetal calf serum (FCS), 2 mM standard protocol for immunohistochemistry, sections were dewaxed glutamine, 50 μM β-mercaptoethanol) and subcloned by limiting and subjected to microwave pretreatment (3×4 min in 10 mM citrate dilution on mouse peritoneal feeder cells. Culture supernatants of buffer, pH 6.5, 600 W; Bosch, Berlin, Germany). Endogenous clones and subclones were tested in an enzyme immunoassay peroxidase activity was blocked by incubating sections with 0.6% (ELISA) for antigen binding with KEPI pET32 fusion protein bound hydrogen peroxide for 15 min at room temperature. Subsequently, to the solid phase. Class and subclass of monoclonal antibodies were sections were incubated with normal goat serum diluted 1 to 20 in determined as described elsewhere (7). The antibodies were purified PBS for 20 min and then incubated with the antibody against KEPI from cell culture supernatant by affinity chromatography using overnight (diluted 1 to 100, in 10% FCS in PBS). Goat anti-mouse Protein A columns (GE Healthcare, Munich, Germany). IgG labelled with horseradish peroxidase (diluted 1 to 100, 90 min; DAKO) was applied and after detection with Vectastain ABC Elite SDS-PAGE and Western blot. Samples were separated in a gradient kit (Vector Laboratories, Wertheim, Germany) immunopositive cells (12.5 to 20%) SDS-polyacrylamide gel according to Laemmli (8). were visualized with 3,3’-diaminobenzidine (Sigma, Germany). Proteins were stained in 0.25% Coomassie Brilliant Blue R-250 Nuclei were counterstained with hematoxylin. For negative control, (Serva, Heidelberg, Germany) and destained. For Western blot, the the primary antibody was omitted. separated proteins were blotted on nitrocellulose membranes (PROTRAN NC BA3; Schleicher & Schuell, Dassel, Germany) Peptide array synthesis and binding studies. Cellulose membrane- using a semi-dry blotting apparatus. After staining with Ponceau S, bound peptides were automatically prepared according to standard the membrane was blocked in Western blocking reagent (Roche, SPOT synthesis protocols using a MultiPep synthesizer (Intavis AG, Mannheim, Germany) (diluted in PBS, pH 7.4, 0.1% Tween 20), Cologne, Germany) as described in detail previously (11). For incubated for 1 h in 0.5 μg/ml purified antibody G18AF8 and generation of the sequence files, the software LISA 1.78 (in-house incubated with secondary peroxidase conjugated goat anti-mouse software) was used. The generated peptide arrays were synthesized IgG(H+L) antibody (Southern Biotech, Birmingham, AL, USA) on an N-modified cellulose-amino-hydroxypropyl ether membrane (1:20,000 in blocking reagent). Detection was performed with the (N-CAPE) (12). The cellulose-bound peptides were pre-washed ECL™-substrate (Lumigen, Southfield, MI, USA) according to the once with ethanol (1×10 min), with Tris-buffered saline (TBS), pH manufacturer’s instructions. To detect the reactivity with KEPI 8.0, (3×10 min) and then blocked for 3 h with blocking buffer protein of other species, the Mega Western Protein Array: Multiple (Sigma-Genosys, Cambridge, MA, USA) in TBS pH 8.0, containing Species (normal) system (BioChain, Hayward, CA, USA) was used 5% sucrose. The membranes were incubated with antibody G18AF8 according to the manufacturer’s instructions with the same reagents (1 μg/ml in blocking buffer) overnight at 4˚C and thereafter washed as described above. with TBS pH 8.0 (3×10 min). All antibodies were protein A- purified. The interaction of the antibodies with the membrane-bound Transfection of HEK293, MCF-10 cells and intracellular FACS peptides were detected by a peroxidase labelled goat anti-mouse staining. Human embryonal kidney cell line HEK293 (ATCC: CRL- IgG (Sigma-Genosys) (1:500 in blocking buffer) for 2.5 h at room 1573) and human breast cancer cell line MCF-7 (ATCC: HTB-22) temperature. To remove excess of antibody, the membrane was were cultivated as described elsewhere (9, 10) and transfected with washed with TBS, pH 8.0, (3×10 min). An Uptilight HRP blot 10 μg of plasmids pEGFP-C2-KEPI, pEGFP-C2, pcDNA3.1-KEPI or chemiluminescent substrate (Uptima-Interchim, Montluçon, France) pcDNA3.1, respectively, using calcium phosphate co-precipitation was applied for detection using a Lumi-Imager (Boehringer

1574 Daskalow et al: Monoclonal Antibody against KEPI

Figure 1. Flow cytometric evaluation of GFP (A) and G18AF8 staining (B) of HEK293 cells transfected with GFP-KEPI (open curve) or untransfected HEK293 cells (shaded curve). An unrelated protein GFP-profilin (C, D) was used. Binding of primary antibody was detected after incubation with biotinylated secondary antibody and enhancement with streptavidin-phycoerythrin.

Mannheim GmbH, Mannheim, Germany). The signal intensities were recorded as Boehringer light units (BLU) using the LumiAnalyst™ software. The binding affinity (Kd) was measured by the BIACORE × system (Uppsala, Sweden) according to standard protocols (12).

Results

The cDNA of kinase C-enhanced PP1 inhibitor (KEPI) was successfully cloned into pET30 and pET32 vectors, which was proven by sequencing the cloning sites. After transformation of E. coli, both KEPI pET30 and KEPI pET32 fusion proteins were expressed after induction with IPTG. Analysis of the recombinant proteins in SDS-PAGE and Coomassie staining showed an estimated size of about 31 kDa (for KEPI pET30) and 35 kDa (for KEPI pET32). For antibody generation mice were immunized with KEPI-pET30 Figure 2. Immunochemistry of a human brain metastasis showing and KEPI–pET32 fusion proteins. Out of several hybridoma groups of carcinoma cells expressing KEPI (brown) within the cytoplasm, some of them also within the nucleus (blue; arrow). The cell lines, one clone was established whose antibody showed staining intensity was weak to moderate. Other parts of this metastasis a specific binding to both KEPI-fusion proteins. This of ductal mamma carcinoma show no expression of KEPI. KEPI antibody, G18AF8 belonged to the IgG2b subclass. In immunohistochemistry (brown) with hematoxylin counterstaining (blue), Western blot analysis with lysates from KEPI-transfected magnification ×400. MCF-7 breast cancer cell line, monoclonal antibody G18AF8 recognized a specific band of about 18 kDa which corresponds to the molecular weight of KEPI. Mock- KEPI) were also stained by G18AF8. The bands correspond transfected MCF-7 cell lysates did not react (5). In previous to about 48 kDa (GFP-KEPI) and several degradation experiments it was shown that most breast cancer cell lines products (32 kDa, 20 kDa) (data not shown). In flow do not express KEPI mRNA (5). Cell lysates from HEK293 cytometric experiments, it was shown that antibody G18AF8 cells transfected with GFP-KEPI fusion plasmid (pEGFP-C1- recognized HEK293 cells transfected with GFP-KEPI, but not

1575 ANTICANCER RESEARCH 30: 1573-1578 (2010)

GFP-profilin (as a control fusion gene) transfected or non- transfected HEK293 cells. Transfection was validated by flow cytometry using GFP fluorescence (Figure 1). An unrelated murine monoclonal antibody was not able to stain these cells in FACS analysis (data not shown). To determine the applicability of antibody G18AF8 to paraffin-embedded tissue, we stained intracranial metastases of ductal mamma carcinomas. The immunohistochemical examination using paraffin sections revealed a moderate to faint nuclear and cytoplasmic staining of solitary tumor cells and smaller groups of cells in six patients. Four metastases did not show expression, neither nuclear nor cytoplasmic (Figure 2). Our results clearly indicate the capacity of the KEPI antibody to work on formaldehyde-fixed, paraffin-embedded tumor tissue. As the antibody G18AF8 was able to detect KEPI in Western blot and flow cytometry, the detected epitope should be linear. Therefore, it was further characterized using a pepscan approach for antibody epitope determination (data not shown). In a first approach, 6mer peptides from the KEPI sequence with an overlap of 4 amino acids were used. The antibody gave the strongest signal with the peptide sequence, RVFFQS, and weaker signals with overlapping peptides GARVFF and FFQSPR. The main epitope was therefore concluded to be GARVFFQSPR, corresponding to N-terminal amino acids 18-27 in the human KEPI protein. Kd measurements of the 10mer epitope revealed an affinity in the low micromolar range (1 μM ± 0.5 μM). By downsizing the epitope to a 6mer (RVFFQS), a clearly loss of Figure 3. Substitutional analysis of the epitope. A: Each residue of the affinity (22 μM ± 0.1 μM) was found. To further analyze the key KEPI epitope GARVFFQSPR was replaced with 20 naturally occuring positions of the KEPI epitope, a substitutional analysis was L-amino acids. Black spots denote interactions of the peptide analogs performed. Every amino acid from the epitope GARVFFQSPR with antibody G18AF8. All spots in the left hand columns are identical was substituted with 20 of the naturally occuring L-amino acids and represent the wild-type (wt) peptide. All other spots are single substitution analogs, the rows defining the sequence position replaced (Figure 3A). Variance analysis revealed that two amino acids (F and the columns defining the amino acid used as a replacement. B: To and Q) are essential for antibody binding and three others have determine objectively the number of countable spots for substitutional a variance from 40-50% (Figure 3B). The exchange of a single analysis, the signal intensities for each spot in the substitutional amino acid at any other position did not influence the antibody analysis were measured as BLU values for antibody G18AF8. The binding. Therefore, a minimal antibody epitope with the percentage of replacement variability (V) of each sequence position was calculated (V=BLU/20X100) and plotted against the sequence position. sequence GAxxxFQSxx could be assumed. Knowing the bound As a result, we observed three classes: a class of low variability epitope, we searched for homologies of human KEPI in different (V<20%), a class of medium variability (50%>V>20%), and a class of species. As KEPI should be expressed in brain (1, 13), we high variability (V>50%). stained a commercially available brain tissue blot from different animal species with our monoclonal antibody. We were able to detect the protein in human brain tissue, as well as in hamster, rabbit, chicken and rat. There were very low signals in dog, pig detected by the antibody was mapped using cellulose bound and monkey (Figure 4). According to the Swiss Prot database, peptide array (pepscan) and the binding motif is further the only known KEPI sequences are human, murine and rat, depicted using amino acid substitution analysis and Kd- where the epitope GARVFFQSPR is identical. measurements. The affinity of G18AF8 to its epitope peptide was in a micromolar range, which should be sufficient for Discussion diagnostic purposes. A minimal antibody epitope with the sequence GAxxxFQSxx was discovered. Only a few other We were able to generate a monoclonal antibody specifically publications have studied KEPI expression so far, mostly at binding to the human PP1 inhibitor KEPI. This antibody the mRNA level or using polyclonal antisera (1, 13, 14). binds its antigen in ELISA, immunoblotting, flow cytometry, Recently, KEPI knock out mice were generated to investigate and immunohistochemistry on paraffin sections. The epitope the influence of KEPI on the morphine receptors in the brain

1576 Daskalow et al: Monoclonal Antibody against KEPI

Figure 4. Detection of KEPI protein in brain tissue of different species. Western blot was performed with purified antibody G18AF8 on a MegaWestern protein array (Biochain Institute, Inc.) representing brain tissue of 15 different species (see Materials and Methods).

(15). The authors suggest a role for KEPI action in adaptive References responses to repeated administration of morphine that include analgesic tolerance and drug reward (15). As the 1 Liu QR, Zhang PW, Zhen Q, Walther D, Wang XB and Uhl GR: biological role of KEPI, especially in cancer and metastasis, KEPI, a PKC-dependent protein phosphatase 1 inhibitor is not yet absolutely clear, antibody G18AF8 should be a regulated by morphine. J Biol Chem 277: 13312-13320, 2002. powerful reagent for examining KEPI expression in different 2 Caselli R, Mencarelli MA, Papa FT, Uliana V, Schiavone S, Strambi M, Pescucci C, Ariani F, Rossi V, Longo I, Meloni I, cell types and tissues, and might also be a diagnostic tool. Renieri A and Mari F: A 2.6 Mb deletion of 6q24.3-25.1 in a Possibly, the expression of KEPI might correlate with patient with growth failure, cardiac septal defect, thin upper lip and cancer stage or prognosis. As the epitope is identical in asymmetric dysmorphic ears. Eur J Med Genet 50: 315-321, 2007. mouse, rat and hamster, this antibody could also be valuable 3 Theile M, Seitz S, Arnold W, Jandrig B, Frege R, Schlag PM, for studies in these and in other species. The antibody was Haensch W, Guski H, Winzer KJ, Barrett JC and Scherneck S: A also binding chicken and rabbit KEPI. Surprisingly, the defined chromosome 6q fragment (at D6S310) harbors a putative antibody is not able to detect murine KEPI, suggesting that tumor suppressor gene for breast cancer. Oncogene 13: 677-685, 1996. either KEPI was not expressed in the murine brain or that 4 Bonazzi VF, Irwin D and Hayward NK: Identification of post-translational modifications hide the peptide epitope. candidate tumor suppressor genes inactivated by promoter Although the KEPI sequence of other species is not known, methylation in melanoma. Genes Cancer 48: 10- we can state that hamster KEPI should be very homologous 21, 2009. or identical in amino acids 18-27. Even the avian KEPI 5 Wenzel K, Daskalow K, Herse F, Seitz S, Zacharias U, Schenk homolog should be similar in this region. In summary, a JA, Schulz H, Hubner N, Micheel B, Schlag PM, Osterziel KJ, murine monoclonal antibody specific for the PP1 inhibitor Ozcelik C, Scherneck S and Jandrig B: Expression of the protein phosphatase 1 inhibitor KEPI is down-regulated in breast cancer KEPI was generated and characterized. It is the first cell lines and tissues and involved in the regulation of the tumor monoclonal antibody binding this antigen described so far suppressor EGR1 via the MEK-ERK pathway. Biol Chem 388: and might help understanding the role of KEPI in 489-495, 2007. cardiovascular and brain function (16) as well as its role as 6 Daskalow K, Boisguerin P, Jandrig B, Volkmer R, Micheel B and a potential tumor suppressor (5). Schenk JA: Epitope mapping of antibodies against S-tagged fusion proteins and molecular weight markers. Biosci Biotechnol Acknowledgements Biochem 72: 346-351, 2008. 7 Schenk JA, Matyssek F and Micheel B: Interleukin 4 increases We are grateful to B. Vogt and A. Menning for help with FACS the antibody response against rubisco in mice. In Vivo 18: 649- analysis and K. Messerschmidt for help in the cell fusion. This work 652, 2004. was supported by: Deutsche Krebshilfe (grant number: 10-1249). 8 Laemmli UK: Cleavage of structural proteins during the P.B. is supported by the Deutsche Forschungsgemeinschaft (DFG assembly of the head of bacteriophage T4. Nature 227: 680-685, Grant VO885/3-2). 1970.

1577 ANTICANCER RESEARCH 30: 1573-1578 (2010)

9 Luebbe L, Schenk JA, Naundorf H, Karsten U and Wunderlich 14 Erdodi F, Kiss E, Walsh MP, Stefansson B, Deng JT, Eto M, V: Reverse transformation of human mammary carcinoma cells. Brautigan DL and Hartshorne DJ: Phosporylation of protein Anticancer Res 19: 5119-5125, 1999. phosphatase type-1 inhibitory proteins by integrin-linked kinase 10 Sheriff A, Vogt B, Baumgart M, Montag C, Hollenbach B, and cyclic nucleotide-dependent protein kinases. Biochem Schenk JA, Ulrich J, Elías F and Micheel B: Intracellular capture Biophys Res Commun 306: 382-387, 2003. of B7 in antigen-presenting cells reduces costimulatory activity. 15 Drgonova J, Zimonjic DB, Hall FS and Uhl GR: Effect of KEPI Biochem Biophys Res Commun 301: 873-878, 2003. (Ppp1r14c) deletion on morphine analgesia and tolerance in mice 11 Frank R: The SPOT-synthesis technique. Synthetic peptide of different genetic backgrounds: when a knockout is near a arrays on membrane supports – principles and applications. relevant quantitative trait locus. Neuroscience 165: 882-895, J Immunol Methods 267: 13-26, 2002. 2010. 12 Boisguerin P, Leben R, Ay B, Radziwill G, Moelling K, Dong L 16 Eto M: Regulation of cellular protein phosphatase-1 (PP1) by and Volkmer-Engert R: An improved method for the synthesis of phosphorylation of the CPI-17 family, C-kinase-activated PP1 cellulose membrane-bound peptides with free C termini is useful inhibitors. J Biol Chem 284: 35273-35277, 2009. for PDZ domain-binding studies. Chem Biol 11: 449-459, 2004. 13 Gong JP, Liu QR, Zhang PW, Wang Y and Uhl GR: Mouse brain localization of the protein kinase C-enhanced phosphatase 1 Received October 12, 2009 inhibitor KEPI (kinase C-enhanced PP1 inhibitor). Neuroscience Revised March 16, 2010 132: 713-727, 2005. Accepted March 19, 2010

1578