RESEARCH ARTICLE

Screening of a Novel Peptide Targeting the ProteoglycanLike Region of Human Carbonic Anhydrase IX Shoaib Rana, Felix Nissen, Thomas Lindner, Annette Altmann, Walter Mier, Juergen Debus, Uwe Haberkorn, and Vasileios Askoxylakis

Abstract The extracellular domain of human carbonic anhydrase IX (CA IX) is extended by a proteoglycan-like region (PGLR). The aim of the present study was the development of novel molecules with specificity for PGLR, which may be used for tumor targeting and imaging. PGLR was chemically synthesized, and phage display biopanning was performed. The identified ligand PGLR-P1 was labeled with 125I and characterized for target binding and metabolic stability. In vitro characterization included kinetic, competition, and internalization studies on CA IX–positive renal cell carcinoma SKRC 52 cells. The CA IX–negative cell lines HEK293 wt and BxPC3 were used as negative controls. In vitro binding experiments revealed an increasing affinity of 125I-PGLR-P1 to SKRC 52 cells but not to negative control HEK293 wt and BxPC3 cells. Internalization studies indicated an exclusive cell membrane binding. Biodistribution analysis demonstrated a higher accumulation in SKRC 52 tumors than in most normal tissues after perfusion. In vivo blocking led to a significant decrease in tumor uptake. Our findings indicate that PGLR-P1 is a promising lead structure for the development of new peptide-based ligands targeting the PGLR of CA IX and reveal challenges that need to be considered for peptide-related molecular imaging.

ANCER BELONGS TO the leading causes of death worldwide and is characterized through an increasing burden.1 Treatment of the disease is usually multimodal, with radiotherapy and chemotherapy known to be powerful therapeutic approaches. However, treatment outcome is influenced by the tumor microenvironment. Hypoxia is a major factor for chemo- and radiotherapy failure. Subphysiologic oxygen levels occur in many solid tumors, mainly as a result of abnormalities in the microvessel structure and limited oxygen diffusion.2–4 At the molecular level, the transcription factor hypoxia-inducible factor 1a (HIF-1a) accumulates under hypoxic conditions, resulting in the activation of several target genes, which promote tumor growth and metastasis, enhance therapy resistance, and lead to reduced therapeutic efficacy.5

C

From the Clinical Cooperation Unit Nuclear Medicine, German Cancer Research Center, INF 280, Heidelberg, Germany; Department of Radiation Oncology, University of Heidelberg, INF 400, Heidelberg, Germany; and Department of Nuclear Medicine, University of Heidelberg, INF 400, Heidelberg, Germany. Address reprint requests to: Vasileios Askoxylakis, MD, Department of Radiation Oncology, University of Heidelberg, Im Neuenheimer Feld 400, 69120, Heidelberg, Germany; e-mail: [email protected].

DOI 10.2310/7290.2013.00066 #

2013 Decker Publishing

Human carbonic anhydrase IX (CA IX) is a prominent target gene of HIF-1a.6–8 CA IX, a membrane-associated member of the CA zinc metalloenzyme family, is overexpressed in many cancer entities,9,10 whereas it shows only a limited expression in normal tissues.9,11 The features of CA IX make the protein an attractive candidate for the development of tumor hypoxia–targeting approaches, which might find application within therapeutic or molecular imaging strategies. Over the past few years, several scientific projects have focused on the identification of CA IX–specific molecules, leading to the development of CA IX–specific sulfonamides and monoclonal antibodies. A prominent example is the chimeric monoclonal antibody G250, which is currently being investigated in phase III clinical trials.12,13 More recent efforts have focused on the identification of peptides with affinity for CA IX. Based on their lower molecular weight, peptides possess attractive properties for targeting and imaging approaches, such as improved tumor penetration and faster blood clearance. In addition, peptides are less immunogenic,14,15 although the improvements in peptide chemistry enable their effective and low-priced synthesis. A novel peptide ligand (CaIX-P1) with affinity for the extracellular domain of human CA IX was recently identified in our group using the high-throughput technology of phage display.16 After characterization of the properties of CaIX-P1, studies focused on the

Molecular Imaging, 2013: pp 1–12

1

2

Rana et al

improvement of its targeting and metabolic characteristics.17,18 The results of these studies led to the hypothesis that CaIX-P1 might be a candidate for a novel molecule targeting human CA IX. However, the fact that the exact binding site of CaIX-P1 on the extracellular domain of the target protein is not characterized yet, and the fact that the extracellular catalytic domain of CA IX shows a strong homology to the catalytic domains of further members of the CA family, such as CA XII and CA II, may limit the potential of the peptide for specific CA IX targeting. The aim of the present study was the identification and the development of further novel peptides with affinity for regions of the extracellular domain of CA IX with no homology to other members of the CA family. To achieve this goal, sequence alignments of individual CAs were carried out, revealing that CA IX exhibits a proteoglycan like-region (PGLR), which is unique for the target (Figure S1, online version only). As the PGLR is not posttranslationally modified, the 75–amino acid sequence was chemically synthesized and used as a target for phage display selection. After four biopanning rounds with a commercially available Ph.D.12 library (New England Biolabs, Frankfurt, Germany),19 the linear dodecapeptide PGLR-P1 (NMPKDVTTRMSS) was identified. The novel ligand was chemically synthesized and evaluated in vitro and in vivo. Specificity in vitro was confirmed using the target as an isolated protein and expressed by cells. Organ distribution studies in mice bearing CA IX– positive tumors revealed increased tumor to organ ratios after animal perfusion, whereas the in vivo specificity was confirmed by in vivo blocking experiments.

Materials and Methods Cell Lines Cell lines were cultivated at 37uC in a 5% CO2 incubator. The human renal cell carcinoma cell line SKRC 52 was obtained from O. Boerman (University of Nijmegen, Nijmegen, the Netherlands). SKRC 52 and HEK293 wt were cultured in RPMI-1640 with GlutaMAX (Invitrogen, Karlsruhe, Germany) containing 10% (v/v) fetal calf serum (Invitrogen). Human pancreatic carcinoma BxPC3 cells were cultured in RPMI-1640 with additional D-glucose (4.5 g/L) (Invitrogen) containing 10% (v/v) fetal calf serum. Phage Display A linear 12–amino acid peptide library (Ph.D.12) was used for biopanning. Panning was performed in solution phase on streptavidin beads coated with the biotinylated,

chemically synthesized PGLR (75 amino acids, N-terminally biotinylated, 80 nM for each round). Uncoated streptavidin beads were used as the control. The individual biopanning rounds were carried out according to the manufacturer’s protocol (New England Biolabs, E8110S). After four selection rounds, clones were picked and phage single-stranded deoxyribonucleic acid (DNA) isolation was performed (QIAprep Spin M13 Kit, Qiagen, Hilden, Germany). DNA sequencing was carried out through GATC Biotech (Konstanz, Germany). Peptide Synthesis, High-Performance Liquid Chromatography (HPLC) Analysis, and HPLC–Mass Spectrometry Analysis Peptides were synthesized by standard Fmoc-based solidphase peptide synthesis using an Applied Biosystems (Darmstadt, Germany) 433A synthesizer as previously described.18 Analytical reversed-phase high-performance liquid chromatography (HPLC) was performed on an Agilent 1100 HPLC system equipped with a Chromolith Performance RP-18e column (100 3 3 mm; 13 nm [130 A˚]; Merck KGaA, Darmstadt, Germany). All peptide synthesis products were characterized by liquid chromatography–mass spectrometry (LC/MS) using an Orbitrap Mass Spectrometer (Exactive, Thermo Fisher Scientific, Bremen, Germany) coupled to an Agilent 1200 HPLC system equipped with a Hypersil Gold C18 column (2.1 3 200 mm, 1.9 mm; Thermo Fisher Scientific). Radiolabeling of Peptides and Radio-HPLC Radiolabeling with iodine 125 or iodine 131 was conducted on an added tyrosine residue using the chloramineT method. Radio-HPLC was performed on an Agilent 1100 HPLC system equipped with a radioactivity detector (GABI Star, Raytest GmbH, Straubenhardt, Germany) using a Chromolith Performance RP-18e column (100 3 3 mm; 13 nm [130 A˚]; Merck KGaA). Binding Experiments on Immobilized Protein Binding of 125I-labeled PGLR-P1 was performed on immobilized PGLR, CA XII, and CA II and on the immobilized recombinant extracellular domains of CA IX and fibroblast growth factor receptor (FGFR). For immobilization, the target proteins were incubated at a concentration of 200 nM in 96-well MaxiSorp plates (Affymetrix, Frankfurt, Germany) for 24 hours, followed

Screening of a Peptide Targeting the PGLR of Human CA IX

by three washing steps with 500 mL phosphate-buffered saline (PBS). Incubation with 125I-PGLR-P1 was carried out in 500 mL PBS for 60 minutes at 4uC. After incubation, the plates were washed three times with 500 mL PBS. The proteins were denatured and solubilized by the addition of 500 mL NaOH (0.3 M), and radioactivity was measured using a gamma-counter. Bound radioactivity was calculated as the percentage of the applied dose. To evaluate the specificity of radioligand binding, competition experiments with the unlabeled PGLR-P1 and the PGLR-P1-all-D peptide at a concentration of 1025 M were carried out. In Vitro Binding Experiments and Internalization Studies A total of 500,000 SKRC 52, HEK293 wt, or BxPC3 cells were seeded into six-well plates and cultivated in 3 mL of incubation medium at 37uC for 24 hours. Binding and internalization studies were performed as previously described.18 To determine specific versus nonspecific binding, the cells were incubated with unlabeled competitors at concentrations varying from 10–3 to 10–10 M. PGLR-P1-all-D, octreotate (TATE), and PDGFR-P1 were used as control competitors. CA IX–negative HEK293 wt and BxPC3 cells were used as negative control cell lines. Organ Distribution Studies Organ distribution experiments were performed in 9-weekold female BALB/c nu/nu mice carrying subcutaneously transplanted SKRC 52 tumors. Animals were obtained from Charles River WIGA (Sulzfeld, Germany). For tumor transplantation, a cell suspension of 5 3 106 SKRC 52 cells was injected subcutaneously into the mouse upper hind limb, and the tumors were grown to a size of 1.0 cm3. 131Ilabeled PGLR-P1 was administered intravenously (approximately 1 MBq in 100 mL PBS), and 15, 60, and 120 minutes postinjection, the animals were sacrificed. Tumor, blood, and selected tissues (heart, spleen, liver, kidney, muscle, intestine, and brain) were removed, drained of blood, and weighed, and the radioactivity was measured using a gamma-counter (LB 951G; Berthold Technologies, Wildbad, Germany). The organ uptake was calculated as percent injected dose per gram tissue (% ID/g). For perfusion experiments, a catheter was put in the ascending aorta and perfusion was performed with 25 mL of 0.9% NaCl through a cut in the liver. After perfusion, samples of tumor and organs were removed and measured as described. To prove the in vivo specificity, blocking experiments were carried out. For in vivo blocking, the

3

radiolabeled PGLR-P1 peptide was coinjected with 100 mL of unlabeled peptide at a concentration of 10–3 M, and biodistribution was investigated 15 minutes after intravenous application. All animal experiments were carried out in conformity with the German laws for protection of animals and were in compliance with European laws. Study approval was received by the Regierungspra¨sidium Karlsruhe, Abteilung 3, Baden-Wu¨rttemberg, Germany (file reference: 35-9185.81/G-132/04). Fluorescence Flow Cytometry Flow cytometry was performed using a DAKO Galaxy Flow-Cytometry System (DAKO, Glostrup, Denmark) equipped with Partec analysis software (Mu¨nster, Germany). Cells (1 3 106) were trypsinated and incubated in RPMI-1640 (containing 10% [v/v] fetal calf serum) for 30 minutes at room temperature to allow recovery from the trypsin stress. The cells were washed twice with 1 mL PBS, blocked with 1 mL PBS 5% bovine serum albumin for 30 minutes, and washed again twice with 1 mL PBS. Cells were then incubated for 60 minutes with mouse anti–CA IX (2.5 mg/106 cells, MAB2188, R&D Systems, Wiesbaden, Germany) on ice, followed by washing twice with 1 mL PBS. The secondary antibody Alexa Fluor 488 goat antimouse (10 mg/mL, A21131, Invitrogen) was added and incubated for 30 minutes on ice. After incubation, cells were washed twice with 1 mL PBS and resuspended in 1 mL cold PBS, and 5 3 103 particles were counted from each sample. Autofluorescence was measured after cellular treatment without the secondary antibody and used to adjust the forward scatter and side scatter detector settings. Real-Time Quantitative Polymerase Chain Reaction and Western Blot Analysis For real-time quantitative polymerase chain reaction (PCR), total cellular ribonucleic acid was isolated from confluent SKRC 52, HEK293 wt, and BxPC3 cells in 75 cm2 cell culture flasks using the TRIzol method (TRIzol Reagent, Invitrogen) as previously described.18 For Western blot analysis, SKRC 52, HEK293 wt, and BxPC3 cells were grown to 80% confluency and Western blot was performed as previously described.18 Rabbit IgG monoclonal anti–human CA IX antibody (1:1,000 dilution, ab108351, Abcam, Cambridge, UK) was used as the primary antibody and antirabbit IgG with conjugated horseradish peroxidase (1:1,000 dilution, HAF008, R&D Systems) as the secondary antibody. For the loading control experiments, goat anti–glyceraldehyde 3-phosphate

4

Rana et al

dehydrogenase (GAPDH) antibody (1:75,000 dilution, PAB6637, Abnova, Heidelberg, Germany) was used as the primary antibody and rabbit antigoat–horseradish peroxidase (1:2,000 dilution, PK-AB718-9320, PromoKine, Heidelberg, Germany) as the secondary antibody. Serum Stability Assay The metabolic stability of 125I-labeled PGLR-P1 was investigated in both human (H4522, Sigma-Aldrich, Steinheim, Germany) and mouse (Invitrogen) serum at 37uC. At selected time points, aliquots were taken, the serum proteins were precipitated, and the supernatant was analyzed by radio-HPLC. Statistics The Student t-test was used to determine statistical significance (SIGMASTAT Jandel Scientific, Erkrath, Germany). Differences were considered significant at the 5% level.

Results Radiolabeling of Peptides All peptides were radioiodinated (125I and 131I) using the choramine-T method.20 Unlabeled peptides were separated

from the radiolabeled derivatives by preparative HPLC. The specific activity after separation was 50 6 5 GBq/mmol. Synthesis of the PGLR PGLR consisting of 75 amino acids was chemically synthesized by solid-phase peptide synthesis. HPLC-MS analysis revealed a high yield and purity (. 95%) of the synthesized protein (Figure 1). Phage display biopanning Four selection rounds were performed on the synthesized PGLR with a linear 12–amino acid peptide library. After the final selection round, single-stranded DNAs of 22 clones were sequenced and analyzed. The sequence NMPKDVTTRMSS (PGLR-P1) was identified in 46% of all sequenced clones. PGLR-P1 Binding Studies to Various Proteins The target specificity of the linear dodecapeptide NMPKDVTTRMSSy (PGLR-P1) was initially investigated on the following proteins: synthesized PGLR, entire CA II and CA XII proteins, recombinant extracellular domains of CA IX, and FGFR. Binding studies on immobilized proteins indicated a radioligand accumulation of about

Figure 1. HPLC-MS analysis after chemical synthesis and preparative HPLC of the biotinylated proteoglycan-like region (PGLR). [M + H]+ 5 8392.7586 (calculated); [M + H]+ 5 8392.7510 (found). Final purity . 95%.

5

Screening of a Peptide Targeting the PGLR of Human CA IX

10% on PGLR and 1.8% on the recombinant extracellular domain of CA IX. No binding was observed on the negative control proteins CA II, CA XII, and FGFR (, 0.1%) These results were highly statistically significant, with p , .01 (Figure 2). CA IX Expression in SKRC 52, HEK293 wt, and BxPC3 Prior to cellular binding experiments, renal cell carcinoma SKRC 52, human embryonic kidney HEK293 wt, and human pancreatic carcinoma BxPC3 cells were tested for CA IX expression. Real-time PCR (Figure 3A), Western blot (Figure 3B), and flow cytometry (Figure 4) revealed exclusive expression of CA IX in SKRC 52 cells. CA IX expression in BxPC3 and HEK293 wt was not observed. Determination of Affinity and Specificity of PGLR-P1 in Cell Binding Experiments The binding kinetics of radiolabeled PGLR-P1 was investigated in vitro on SKRC 52, HEK 293 wt, and BxPC3 cells. An increasing accumulation over time was shown for CA IX–positive SKRC 52 cells, with a maximal binding of about 4.9% applied dose/106 cells after 2 hours of incubation. No binding of the radiolabeled peptide was noticed on the negative control target cell lines HEK293 wt and BxPC3 (p , .01) (Figure 5A). To test whether the binding of PGLR-P1 peptide was not solely based on its charge, the all-D enantiomer, consisting of the same amino acid sequence in D form, was

synthesized, labeled, and investigated for cellular binding. Kinetic experiments on CA IX–positive SKRC 52 cells revealed no binding for the radiolabeled all-D peptide (p , .01) (see Figure 5A). In addition to kinetics, internalization experiments of radiolabeled PGLR-P1 were carried out. Internalization studies revealed an increase in the membrane-bound activity over time. Only a small fraction (0.8%) of the applied activity was taken up by the cells after 2 hours of incubation (Figure 5B). Competition Experiments Competition experiments were carried out on both immobilized protein and CA IX–expressing SKRC 52 cells. Studies on the immobilized extracellular domain of CA IX using the unlabeled PGLR-P1 peptide as a competitor showed a strong inhibition of the radioligand binding (p , .05). When the PGLR-P1–all-D peptide was used as the negative control competitor, no inhibition of the radioligand accumulation was noticed (Figure 6A). Competition experiments on CA IX–positive SKRC 52 cells using the unlabeled PGLR-P1 peptide as a competitor at different concentrations demonstrated an increasing inhibition of the radioligand binding with increasing competitor concentration (Figure 6B). The half-maximal inhibitory concentration (IC50) value was calculated to be about 4 mM. The respective dissociation constant (Kd) value was calculated to be 4.46 mM (Figure 6C). Using

Figure 2. Binding of 125I-PGLR-P1 on chemically synthesized PGLR, immobilized entire CA II and CA XII protein, and recombinant extracellular domains of CA IX and fibroblast growth factor receptor (FGFR). Radioligand incubation was performed for 60 minutes at 4uC (mean 6 SD; n 5 3). *p , .01.

6

Rana et al

Figure 3. (A) Real-time polymerase chain reaction and (B) Western blot analysis of CA IX expression in SKRC 52, HEK293 wt, and BxPC3 cells. CA IX is expressed in the form of 54 and 58 kDa protein bands. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the loading control.

Figure 4. Flow cytometry analysis of CA IX expression in SKRC 52, HEK293 wt, and BxPC3 cells. Autofluorescence was determined for all cell lines (A 5 SKRC 52; B 5 BxPC3; C 5 HEK293 wt). The bottom half of the figure illustrates the fluorescence signal after incubation with an antihCA9 antibody (D 5 SKRC 52; E 5 BxPC3; F 5 HEK293 wt).

Screening of a Peptide Targeting the PGLR of Human CA IX

7

Figure 5. A, In vitro kinetics of 125I-PGLR-P1 and 125I-PGLR-P1-all-D on SKRC 52, HEK293 wt, and BxPC3 cells. Incubation was performed for time periods varying between 10 and 120 minutes. B, Binding and internalization experiments of 125I-PGLR-P1 on SKRC 52 cells. Cells were incubated with the radioligand for 10, 30, 60, and 120 minutes at 37uC (mean 6 SD; n 5 3). *p , .01.

PGLR-P1–all-D, TATE, and PDGFR-P121 as control competitors at the same concentration, a significant reduction in the competition of 125I-PGLR-P1 was demonstrated (p , .05) (Figure 6D). Serum Stability Studies Characterization of the peptide’s stability in human serum revealed a time-dependent proteolytic degradation with a serum half-life of approximately 20 minutes (Figure 7A). A rapid peptide degradation was revealed when PGLR-P1 was incubated in mouse serum (Figure 7B). In Vivo Pharmacokinetics of PGLR-P1 Organ distribution analysis showed a tumor accumulation of about 2% ID/g tissue after 15 minutes of circulation in the bloodstream. Higher activities were measured for blood (< 3%), liver (< 3%), and kidney (< 19%). In vivo kinetics revealed a time-dependent decrease in both tumor and selected tissues (Table 1). The rate of clearance from most normal tissues was higher than for the tumor, resulting in an increase in the tumor to organ ratios. To reduce the blood background, organ distribution studies were also performed after animal perfusion (Figure 7C). Perfusion experiments demonstrated a strong radioactivity reduction in well-perfused organs, including the heart, lung, and liver, leading to a significant increase in the tumor to organ ratios (Table 2). To prove the in vivo specificity of PGLR-P1, blocking experiments with coinjection of unlabeled peptide and

radioligand were carried out and the ratios of binding with competitor to binding without competitor were calculated. In vivo blocking studies demonstrated a binding decrease only in the tumor. In particular, at 15 minutes after intravenous application, about 60% of the tumor uptake without blocking was measured (Figure 7D). Binding in all normal tissues remained constant or slightly increased, leading to a significant decrease in the tumor to organ ratios (see Table 2).

Discussion CA IX is a transmembrane zinc metalloenzyme that is overexpressed in various tumors. The fact that CA IX is a target gene of the transcription factor HIF-1a and is overexpressed under conditions of subphysiologic oxygen levels makes the molecule an attractive target for the development of molecular diagnostic and therapeutic strategies focusing on tumor hypoxia. The aim of this study was to identify a novel peptide with specific binding properties for the PGLR of human CA IX. The rationale for the project is based on the fact that PGLR is part of the extracellular domain of CA IX but not of the extracellular domain of other known members of the CA family. The strategy chosen to achieve this goal included the chemical synthesis of the PGLR of human CA IX, followed by biotinylation, immobilization, and peptide screening using the high-throughput technology of phage display. Since the PGLR of CA IX does not contain posttranslational

8

Rana et al

Figure 6. A, Competition experiments of 125I-PGLR-P1 on the recombinant extracellular domain of CA IX. Target protein was coated at a concentration of 200 nM overnight on MaxiSorp plates. B, Competition of 125I-PGLR-P1 using the unlabeled peptide at various concentrations. Incubation was performed for 60 minutes on SKRC 52 cells at 37uC. C, Scatchard plot analysis of the cellular competition studies. D, Competition of 125I-PGLR-P1 on SKRC 52 cells using different peptides as a competitor at the same concentration (10–5 M). Incubation was performed for 60 mintues at 37uC (mean 6 SD; n 5 3). *p , .05.

modifications, the chemical synthesis of the target was chosen because it offers important advantages, such as lower cost and time-intensive target production and easier modifications for target immobilization, including biotinylation. After four rounds of peptide screening on the chemically synthesized PGLR of human CA IX using phage display technology, the dodecapeptide PGLR-P1 (NMPKDVTTRMSS) was isolated. A D-tyrosine was added on the C-terminus of the peptide for radioiodination, and labeling was performed. Binding experiments on immobilized proteins revealed higher affinity and binding capacity to PGLR and the extracellular domain of human CA IX compared to negative control targets, such as CA II, CA XII, and FGFR. The binding capacity on PGLR was higher compared to CA IX, which is probably explained by the

greater accessibility of the peptide epitopes and/or by structural and conformational differences between the two targets. Furthermore, binding of radiolabeled PGLR-P1 could be strongly inhibited by the unlabeled ligand, not by negative controls, confirming target specificity. The hypothesis of a specific target binding is further supported by the results of in vitro experiments on various well-established tumor cell lines, with differential CA IX expression, investigated by quantitative real-time PCR, Western blot, and flow cytometry analysis. Kinetic experiments of radiolabeled PGLR-P1 revealed an increasing accumulation on CA IX–positive SKRC 52 cells but not on the CA IX–negative control cell lines, indicating a correlation between target expression and peptide binding. In addition, competition experiments revealed a concentration-dependent inhibition of radioligand binding

Screening of a Peptide Targeting the PGLR of Human CA IX

9

Figure 7. High-performance liquid chromatograms of the metabolic stability of 125I-PGLR-P1 in (A) human serum and (B) mouse serum. C, Organ distribution of 131I-labeled PGLR-P1 in female BALB/c nu/nu mice carrying SKRC 52 tumors with and without perfusion. Activity concentration (%ID/g tissue) in tumor and control organs. Incubation was performed for 15 minutes (three animals per experiment). D, In vivo blocking after coinjection of 131I-labeled PGLR-P1 with unlabeled PGLR-P1 at a concentration of 10–3 M. Ratios of binding with competitor to binding without competitor after 15 minutes (three animals per experiment).

by the unlabeled PGLR-P1 but not by other randomly chosen peptides, providing further strong indications for target specificity. Furthermore, the results of studies using the D-isomer of PGLR-P1 as a competitor provide Table 1. Organ Distribution of 131I-Labeled PGLR-P1 in Female BALB/c nu/nu Mice Carrying SKRC 52 Tumors %ID/g Postinjection (n 5 3 animals) Tissue Blood Heart Lung Spleen Liver Kidney Muscle Intestinum Brain Tumor

15 min 3.013 6 0.195 1.319 6 0.181 2.797 6 0.163 1.239 6 0.157 4.244 6 0.823 18.291 6 3.895 0.652 6 0.128 2.075 6 0.331 0.104 6 0.018 1.903 6 0.259

PGLR 5 proteoglycan-like region.

60 min 0.457 0.220 0.563 0.351 0.259 1.323 0.131 0.330 0.036 0.475

6 6 6 6 6 6 6 6 6 6

0.118 0.041 0.229 0.072 0.084 0.167 0.031 0.027 0.042 0.178

120 min 0.151 0.140 0.156 0.110 0.079 0.495 0.050 0.124 0.013 0.112

6 6 6 6 6 6 6 6 6 6

0.035 0.112 0.018 0.022 0.021 0.100 0.023 0.093 0.016 0.019

evidence that cellular binding of the novel ligand is not mediated by its molecular charge, strengthening the hypothesis of specific accumulation. Moreover, internalization studies demonstrated the majority of the radioactivity to be bound on the cellular membrane of SKRC 52 cells, which is in concert with the extracellular localization of the PGLR. The use of a ligand for targeting or imaging purposes requires advantageous in vivo pharmacokinetics, characterized by a higher tumor accumulation compared to normal organs. Although organ distribution studies indicated a slightly higher uptake in SKRC 52 tumors than in most normal tissues, the radioactivity values in kidneys, liver, and blood were higher, which is very disadvantageous for clinical applications because it would result in enhanced unspecific background. Furthermore, our biodistribution experiments showed a reduction in the peptide uptake with time progression, not only in normal organs but also in the tumor, which is also disadvantageous for in vivo targeting. However, the fact that the radioactivity reduction was lower for the tumor than for

10

Rana et al

Table 2. Tumor/Organ Ratios of 131I-Labeled PGLR-P1 in Female BALB/c nu/nu Mice Carrying SKRC 52 Tumors 15 min after Intravenous Application (3 animals per experiment) Tumor to Organ Ratios Heart Lung Spleen Liver Kidney Muscle Intestine Brain

Without Perfusion

After Perfusion

1.716 6 0.130 0.778 6 0.096 1.668 6 0.065 0.629 6 0.060 0.097 6 0.038 2.906 6 0.454 1.191 6 0.233 17.294 6 2.625

2.924 6 0.114* 3.014 6 0.419* 1.825 6 0.346 1.066 6 0.486* 0.127 6 0.026 3.603 6 1.187 1.310 6 0.219 20.499 6 7.313

In Vivo Blocking 0.855 0.444 0.815 0.300 0.086 1.367 0.558 7.568

6 6 6 6 6 6 6 6

0.307* 0.162* 0.178* 0.127* 0.002* 0.382* 0.170* 1.989*

PGLR 5 proteoglycan-like region. For in vivo blocking, 100 mL of unlabeled PGLR-P1 (10–3 M) was coinjected with the radiolabeled peptide. *p , .05.

the normal organs, resulting in a slight increase in the tumor to organ ratios for up to 60 minutes, as well as the fact that perfusion experiments demonstrated a significant decrease in peptide accumulation in well-perfused organs but not in the tumor, allows the hypothesis that PGLR-P1 might be a promising candidate for a lead structure targeting human CA IX. The rationale for perfusion studies was to investigate the radioligand uptake in the tumor and the normal tissues without the blood background. However, such studies have two major drawbacks: (1) they are not feasible in patients and (2) their results might be influenced by tissue perfusion. Well-perfused organs are characterized by increased washout compared to poorly perfused tissues. Still, the in vivo perfusion experiments provide important indication for in vivo specificity of the novel ligand. This is further supported by the results of in vivo blocking experiments. Coinjection of unlabeled peptide resulted in a significant reduction of the radioligand binding only in the tumor but not in the normal tissues and in a significant decrease in the tumor to organ ratios. These results confirm the in vivo specificity of the tumor uptake and reveal the potential of the new ligand as an imaging probe or therapy delivery agent. Major ligand improvement is, however, necessary prior to its in vivo application. Efforts to improve the ligand’s characteristics include the optimization of both binding affinity and metabolic stability. In regard to binding affinity, our in vitro data revealed Kd and IC50 values in the micromolar range. Since affinities in the single-digit nanomolar range are required for clinical applications,22 affinity maturation is absolutely essential. In this respect, different strategies might be applied, including targeted modifications after identification of the binding site in the ligands sequence23 or peptide multimerization.21

A further major aspect refers to the high blood values resulting in an increased radioactivity background. Besides possible interactions of the radiolabeled peptide with serum proteins, the enhanced blood levels are explained by the serum instability of PGLR-P1. Stability studies revealed rapid peptide degradation by both human and mouse serum proteases, resulting in circulation of radiolabeled fragments. Therefore, a major goal of further investigations is the metabolic stabilization of PGLR-P1. To achieve this goal, a cyclic derivative of PGLR-P1 (cPGLR-P1) was chemically synthesized after conjugation of two cysteines at the N- and C-terminals of the peptide. However, investigation of the targeting properties of cPGLR-P1 demonstrated a significantly reduced target binding (Figure S2, online version only). The fact that structural and conformational changes in the sequence of PGLR-P1 reduce its binding affinity further supports the hypothesis of specific targeting. However, it also reveals challenges associated with the optimization of linear oligopeptides. Further approaches to achieve this goal include the identification of the degradation site in the binding sequence of the peptide, followed by an exchange of single amino acids by unnatural ones, which cannot be recognized by serum proteases, the methylation24 or acetylation25 of the binding sequence, or the grafting of the binding motif into a stable scaffold structure.26–28 In regard to the unfavorable biodistribution, a further factor that might limit the safety of conclusions based on the pharmacokinetic results refers to the used animal model. Previous analyses have shown that mouse CA IX also contains a PGLR, which has a homology to the PGLR of human CA IX with a 57.4% identity.29 However, normal mouse tissues are known to possess different expression patterns of CA IX, characterized by increased protein

Screening of a Peptide Targeting the PGLR of Human CA IX

expression levels in the stomach, pancreas, and large intestine and lower but detectable signals for other tissues, such as the kidneys and the liver.30 An analysis of the accumulation of PGLR-P1 in mouse organs with high CA IX expression revealed a decreased peptide binding. In particular, peptide binding in the stomach, pancreas, and colon was reduced to 1.2% ID/g, 1.3% ID/g, and about 1.0% ID/g, respectively (data not shown). This result is similar to the biodistribution of well-established antihuman CA IX anitbodies, such as M75 and cG250, which also reveal decreased uptake values for mouse stomach, pancreas, or colon31,32 and indicates that our new peptide ligand does not have an increased affinity for mouse CA IX.

Conclusion The results of our studies indicate that the identified peptide PGLR-P1 selectively binds to the PGLR of CA IX. Despite the in vivo limitations, our data allow the hypothesis that PGLR-P1 might be a promising lead structure for the development of novel peptide-based ligands that can find application within imaging and targeting strategies. The future of the novel ligand will depend on how effectively issues of major importance such as serum stability, rapid blood clearance, binding affinity, and high uptake in normal tissues, such as kidney, will be resolved.

Acknowledgments We thank Ursula Schierbaum and Karin Leotta for their help in performing the animal experiments and Ulrike Bauder-Wu¨st for support with the flow cytometry analyzes. We also thank Matthias Strieker for his help with the HPLC analysis. Financial disclosure of authors: S.R. received financial support from the Tumorzentrum Heidelberg/Mannheim. V.A. received financial support from the Medical Faculty of the University of Heidelberg. The funders had no role in the design of the study, data collection and analysis, or preparation of the manuscript. Financial disclosure of reviewers: None reported.

References 1. Jemal A, Bray F, Center MM, et al. Global cancer statistics. CA Cancer J Clin 2011;61:69–90, doi:10.3322/caac.20107. 2. Wilson WR, Hay MP. Targeting hypoxia in cancer therapy. Nat Rev Cancer 2011;11:393–410, doi:10.1038/nrc3064. 3. Dewhirst MW, Cao Y, Moeller B. Cycling hypoxia and free radicals regulate angiogenesis and radiotherapy response. Nat Rev Cancer 2008;8:425–37, doi:10.1038/nrc2397.

11

4. Chitneni SK, Palmer GM, Zalutsky MR, et al. Molecular imaging of hypoxia. J Nucl Med 2011;52:165–8, doi:10.2967/jnumed.110. 075663. 5. Vaupel P. Hypoxia and aggressive tumor phenotype: implications for therapy and prognosis. Oncologist 2008;13 Suppl 3:21–6, doi:10.1634/theoncologist.13-S3-21. 6. Grabmaier K, A de Weijert MC, Verhaegh GW, et al. Strict regulation of CAIX(G250/MN) by HIF-1alpha in clear cell renal cell carcinoma. Oncogene 2004;23:5624–31, doi:10.1038/sj.onc. 1207764. 7. Semenza GL. Targeting HIF-1 for cancer therapy. Nat Rev Cancer 2003;3:721–32, doi:10.1038/nrc1187. 8. Askoxylakis V, Millonig G, Wirkner U, et al. Investigation of tumor hypoxia using a two-enzyme system for in vitro generation of oxygen deficiency. Radiat Oncol 2011;6:35, doi:10.1186/1748717X-6-35. 9. Wykoff CC, Beasley NJ, Watson PH, et al. Hypoxia-inducible expression of tumor-associated carbonic anhydrases. Cancer Res 2000;60:7075–83. 10. Alterio V, Hilvo M, Di Fiore A, et al. Crystal structure of the catalytic domain of the tumor-associated human carbonic anhydrase IX. Proc Natl Acad Sci U S A 2009;106:16233–8, doi:10.1073/pnas.0908301106. 11. Saarnio J, Parkkila S, Parkkila AK, et al. Immunohistochemistry of carbonic anhydrase isozyme IX (MN/CA IX) in human gut reveals polarized expression in the epithelial cells with the highest proliferative capacity. J Histochem Cytochem 1998;46:497–504, doi:10.1177/002215549804600409. 12. Uemura H, Nakagawa Y, Yoshida K, et al. MN/CA IX/G250 as a potential target for immunotherapy of renal cell carcinomas. Br J Cancer 1999;81:741–6, doi:10.1038/sj.bjc.6690757. 13. Bleumer I, Knuth A, Oosterwijk E, et al. A phase II trial of chimeric monoclonal antibody G250 for advanced renal cell carcinoma patients. Br J Cancer 2004;90:985–90, doi:10.1038/sj.bjc.660 1617. 14. Laverman P, Sosabowski JK, Boerman OC, et al. Radiolabelled peptides for oncological diagnosis. Eur J Nucl Med Mol Imaging 2012;39 Suppl 1:S78–92, doi:10.1007/s00259-011-2014-7. 15. Marr A, Markert A, Altmann A, et al. Biotechnology techniques for the development of new tumor specific peptides. Methods 2011;55: 215–22, doi:10.1016/j.ymeth.2011.05.002. 16. Askoxylakis V, Garcia-Boy R, Rana S, et al. A new peptide ligand for targeting human carbonic anhydrase IX, identified through the phage display technology. PLoS One 2010;5:e15962, doi:10.1371/ journal.pone.0015962. 17. Askoxylakis V, Ehemann V, Rana S, et al. Binding of the phage display derived peptide CaIX-P1 on human colorectal carcinoma cells correlates with the expression of carbonic anhydrase IX. Int J Mol Sci 2012;13:13030–48, doi:10.3390/ijms131013030. 18. Rana S, Nissen F, Marr A, et al. Optimization of a novel peptide ligand targeting human carbonic anhydrase IX. PLoS One 2012;7: e38279, doi:10.1371/journal.pone.0038279. 19. Smith GP. Filamentous fusion phage: novel expression vectors that display cloned antigens on the virion surface. Science 1985;228: 1315–7, doi:10.1126/science.4001944. 20. Hunter WM, Greenwood FC. Preparation of iodine-131 labelled human growth hormone of high specific activity. Nature 1962;194: 495–6, doi:10.1038/194495a0.

12

Rana et al

21. Askoxylakis V, Marr A, Altmann A, et al. Peptide-based targeting of the platelet-derived growth factor receptor beta. Mol Imaging Biol 2013;15:212–21, doi:10.1007/s11307-012-0578-7. 22. Merlo A, Hausmann O, Wasner M, et al. Locoregional regulatory peptide receptor targeting with the diffusible somatostatin analogue 90Y-labeled DOTA0-D-Phe1-Tyr3-octreotide (DOTATOC): a pilot study in human gliomas. Clin Cancer Res 1999;5: 1025–33. 23. Marr A, Nissen F, Maisch D, et al. Peptide arrays for development of PDGFRbeta affine molecules. Mol Imaging Biol 2013;15:391– 400, doi:10.1007/s11307-013-0616-0. 24. Fischer PM. The design, synthesis and application of stereochemical and directional peptide isomers: a critical review. Curr Protein Pept Sci 2003;4:339–56, doi:10.2174/138920303348 7054. 25. John H, Maronde E, Forssmann WG, et al. N-terminal acetylation protects glucagon-like peptide GLP-1-(7-34)-amide from DPP-IVmediated degradation retaining cAMP- and insulin-releasing capacity. Eur J Med Res 2008;13:73–8. 26. Boy RG, Mier W, Nothelfer EM, et al. Sunflower trypsin inhibitor 1 derivatives as molecular scaffolds for the development of novel

27.

28.

29. 30. 31.

32.

peptidic radiopharmaceuticals. Mol Imaging Biol 2010;12:377–85, doi:10.1007/s11307-009-0287-z. Zoller F, Schwaebel T, Markert A, et al. Engineering and functionalization of the disulfide-constrained miniprotein min-23 as a scaffold for diagnostic application. ChemMedChem 2012;7: 237–47, doi:10.1002/cmdc.201100497. Zoller F, Markert A, Barthe P, et al. Combination of phage display and molecular grafting generates highly specific tumor-targeting miniproteins. Angew Chem Int Ed Engl 2012;51:13136–9, doi:10.1002/anie.201203857. Available at: http://www.uniprot.org/uniprot/Q8VHB5. Hilvo M, Rafajova M, Pastorekova S, et al. Expression of carbonic anhydrase IX in mouse tissues. J Histochem Cytochem 2004;52:1313–22. Chrastina A, Zavada J, Parkkila S, et al. Biodistribution and pharmacokinetics of 125I-labeled monoclonal antibody M75 specific for carbonic anhydrase IX, an intrinsic marker of hypoxia, in nude mice xenografted with human colorectal carcinoma. Int J Cancer 2003;105:873–81, doi:10.1002/ijc.11142. Carlin S, Khan N, Ku T, et al. Molecular targeting of carbonic anhydrase IX in mice with hypoxic HT29 colorectal tumor xenografts. PLoS One 2010;5:e10857, doi:10.1371/journal.pone.0010857.

Screening of a novel peptide targeting the proteoglycan-like region of human carbonic anhydrase IX.

The extracellular domain of human carbonic anhydrase IX (CA IX) is extended by a proteoglycan-like region (PGLR). The aim of the present study was the...
1MB Sizes 2 Downloads 0 Views