European Journal of Pharmaceutical Sciences 52 (2014) 69–76

Contents lists available at ScienceDirect

European Journal of Pharmaceutical Sciences journal homepage: www.elsevier.com/locate/ejps

The gastrin/cholecystokinin-B receptor on prostate cells – A novel target for bifunctional prostate cancer imaging Alexander Sturzu a,b,⇑, Uwe Klose a, Sumbla Sheikh b, Hartmut Echner b, Hubert Kalbacher b, Martin Deeg c, Thomas Nägele a, Christian Schwentner d, Ulrike Ernemann a, Stefan Heckl a a

Department of Neuroradiology, University of Tübingen, Germany Peptide Synthesis Laboratory, Interfaculty Institute of Biochemistry, University of Tübingen, Germany Mass Spectrometry Laboratory, Medicinal Research Centre, University of Tübingen, Germany d Department of Urology, University of Tübingen, Germany b c

a r t i c l e

i n f o

Article history: Received 25 March 2013 Received in revised form 22 October 2013 Accepted 22 October 2013 Available online 6 November 2013 Keywords: Prostate cancer Gastrin receptor Magnetic resonance imaging Bimodal contrast agent

a b s t r a c t The means of identifying prostate carcinoma and its metastases are limited. The contrast agents used in magnetic resonance imaging clinical diagnostics are not taken up into the tumor cells, but only accumulate in the interstitial space of the highly vasculated tumor. We examined the gastrin/cholecystokinin-B receptor as a possible target for prostate-specific detection using the C-terminal seven amino acid sequence of the gastrin peptide hormone. The correct sequence and a scrambled control sequence were coupled to the fluorescent dye rhodamine and the magnetic resonance imaging contrast agent gadolinium (Gd)-1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA). Expression analysis of the gastrin receptor mRNA was performed by reverse transcriptase polymerase chain reaction on PC3 prostate carcinoma cells, U373 glioma, U2OS osteosarcoma and Colo205 colon carcinoma cells. After having confirmed elevated expression of gastrin receptor in PC3 cells and very low expression of the receptor in Colo205 cells, these two cell lines were used to create tumor xenografts on nude mice for in vivo experiments. Confocal lasers scanning microscopy and magnetic resonance imaging showed a high specificity of the correct conjugate for the PC3 xenografts. Staining of the PC3 xenografts was much weaker with the scrambled conjugate while the Colo205 xenografts showed no marked staining with any of the conjugates. In vitro experiments comparing the correct and scrambled conjugates on PC3 cells by magnetic resonance relaxometry and fluorescence-activated cell sorting confirmed markedly higher specificity of the correct conjugate. The investigations show that the gastrin receptor is a promising tumor cell surface target for future prostate-cancer-specific imaging applications. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction For prostate cancer patients’ prognosis and further treatment the detection of the prostate cancer’s lymph node metastases is decisive (Messing et al., 1999; Walsh, 2002). To this means, in magnetic resonance imaging (MRI), intravenously injected iron particles have been applied (Harisinghani et al., 2006; Harisinghani and Weissleder, 2004). The reticuloendothelial system (RES) takes up these iron particles. In T2-weighted magnetic resonance imaging, the iron particles lower the signal intensity of the RES, while

⇑ Corresponding author. Address: Department of Neuroradiology, University of Tübingen, Hoppe-Seyler-Str. 3, 72076 Tübingen, Germany. Tel.: +49 1739501785. E-mail address: [email protected] (A. Sturzu). 0928-0987/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ejps.2013.10.013

disruption of the RES, for example by lymph node metastases, does not show lowered signal intensity (Harisinghani et al., 2006; Harisinghani and Weissleder, 2004; Deserno et al., 2004). However, with this procedure, lymph node metastases cannot be differentiated from inflammation, fat or fibrosis disrupting the RES. As it is not the cancer cells themselves that are detected, but the surrounding RES, metastases also cannot be clearly designated to the organ they originated from (e.g. prostate, bladder). We picked the peptide hormone gastrin and its receptor as a candidate for a tissue-specific MRI marker. Gastrin binds to the G-coupled – gastrin-/cholecystokinin-B receptor found in the brain and in the intestinal tract (Dufresne et al., 2006). It has been shown to influence cell proliferation and migration in human glioblastomas (Camby et al., 1996, 1994; De

70

A. Sturzu et al. / European Journal of Pharmaceutical Sciences 52 (2014) 69–76

Hauwer et al., 1998), and specifically induce over-expression of genes involved in human U373 glioblastoma cell migration (Kucharczak et al., 2001). Autocrine stimulation by glycine-extended gastrin has been reported for colon cells (Hollande et al., 1997) and activation of gastrin and its receptor have been found to be an early event in the adenoma–carcinoma sequence (Smith and Watson, 2000). To confirm the gastrin receptor as suitable target for our approach we examined the gastrin-/cholecystokinin-B receptor in four different human tumor cell lines (PC3 prostate carcinoma, Colo205 colon carcinoma, U373 glioma and U2OS osteosarcoma) via mRNA expression analysis. We then used the seven C-terminal amino acids of the gastrin peptide sequence (AFWGDMY) to target the gastrin receptor. To this peptide’s N-terminus we coupled a lysine carrying the fluorescent dye rhodamine on its e-amino group chain and the magnetic resonance imaging (MRI) contrast agent gadolinium (Gd)-1,4,7, 10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) on its a-amino group. As a negative control we also synthesized a conjugate in which the seven amino acids of the gastrin sequence were randomly rearranged (AYGWMDF). This control conjugate also contained the N-terminal lysine carrying both the MRI and fluorescence markers. While a lot of gastrin conjugates have already been studied, the advantage of the present approach is that the peptide

component is relatively small but still carries two imaging agents for different imaging methods. The cellular binding, cytotoxic effects and diagnostic value of the correct conjugate and the scrambled control conjugate were evaluated in vitro by fluorescence assisted cell sorting and magnetic resonance relaxometry of carcinoma cell lines and in vivo on mice bearing carcinoma cell xenografts.

2. Materials and methods 2.1. Synthesis of the correct and scrambled seven-amino-acid gastrin conjugates (Fig. 1)

C1: GdDOTA-Lys(RITC)-Ala-Tyr-Gly-Trp-Met-Asp-Phe C2: GdDOTA-Lys(RITC)-Ala-Phe-Trp-Gly-Asp-Met-Tyr Conjugate synthesis was performed on 0.1 mmol scale by solid phase peptide synthesis on a Eppendorf ECOSYN P peptide synthesizer (Eppendorf-Biotronik, Hamburg, Germany) using the fmoc (fluorenylmethyloxycarbonyl) strategy. Tentagel S rink amid resin (Rapp-Polymere, Tübingen, Germany) was used as carrier which

Correct conjugate: GdDOTA-Lys(RITC)-Ala-Tyr-Gly-Trp-Met-Asp-Phe

Scrambled conjugate: GdDOTA-Lys(RITC)-Ala-Phe-Trp-Gly-Asp-Met-Tyr

Fig. 1. Structures of the correct and scrambled gastrin conjugates.

A. Sturzu et al. / European Journal of Pharmaceutical Sciences 52 (2014) 69–76

resulted in the final peptide conjugates being amidated at their Cterminus. O-(Benzotriazol-1-yl)-N,N,N0 ,N0 -tetramethyluronium tetrafluoroborate (TBTU) was used as coupling agent and diisopropyl-ethylamine (DIPEA) as adjuvant base for amino acid coupling in sequential peptide elongation. The fmoc protective group was cleaved off with 25% piperidine in dimethylformamide. The N-terminal lysine carrying rhodamine on its e-amino group and GdDOTA on the a-amino group was introduced as fmoc-lysine with Dde (1-(4,4-dimethyl-2,6-dioxocyclohex-1-ylidene)-ethyl) protecting the amino group of the side chain. For introduction of DOTA the fmoc protective group was cleaved off as in the previous elongation steps. Coupling was carried out with 3 equivalents of 1,4,7,10-tetraazacyclododecane1,4,7-tris(tert.butylester)-10-acetic acid (Macrocyclics, Dallas, USA), 3 equivalents of TBTU and 6 equivalents of DIPEA during 1.5 h at room temperature. Then the Dde side chain protective group was cleaved off by repeatedly treating the resin with a solution of 2.5% hydrazinhydrate in DMF over the course of one hour. After several washing steps with DMF, rhodamine was coupled to the peptide conjugate using 0.5 mM rhodamine-isothiocyanate with equal amount of DIPEA in DMSO (dimethylsulfoxide) at room temperature over night. The peptide sequences synthesized were analyzed and separated by RP-HPLC (Merck Hitachi, L-4000 A UV detector) using a C8 column(150 * 10 mm; Reprosil 100; Dr. Maisch GmbH, Germany) with the following solvent systems: (A) 0.055% v/v trifluoroacetic acid in water. (B) 0.05% v/v trifluoroacetic acid in 80% v/v acetonitrile in water. Elution was performed using a linear gradient from 20% to 100% B within 40 min at the flow rate of 2.5 ml/min and at 214 nm absorbance. The purified conjugate (>97% purity) was then lyophilized. The resulting substance was stirred for 5 h at 50 °C with 1 equivalent of Gd(III)chloride hexahydrate in water at pH 5.6 (adjusted with NaOH). After acidification with diluted acetic acid the substance was relyophilized. The products were verified by MALDI-TOF (matrix assisted laser desorption/ionisation – time of flight) mass spectrometry on a Bruker Daltonics reflex IV (Bruker Daltonics Inc., Billerica, MA, USA). C1: theoretical mass = 2057.0, measured mass = 2057.5; C2: theoretical mass = 2057.0, measured mass = 2058.1. 1H NMR spectroscopy on a Bruker Avance II spectrometer (Bruker Biospin, Billerica, MA, USA) confirmed that the conjugates of identical mass had different sequences.

2.2. mRNA expression analysis 1.5 ml cups containing cell pellets of approximately 1  106 cells of PC3 prostate carcinoma, U373 glioma, Colo205 colon carcinoma and U2OS osteosarcoma (grown in DMEM-GlutaMAX containing 4.5 g/l glucose and pyruvate (Gibco, Invitrogen, Darmstadt, Germany) with 10% FBS Gold (PAA laboratories, Pasching, Austria)) were prepared using the cell culture methods and conditions described in the previous sections. The cell pellets were also deep frozen in liquid nitrogen. Total RNA of the cell line and tissue samples was extracted using the Nucleospin RNA II kit (Macherey–Nagel, Düren, Germany) according to the manufacturer’s instructions. RNA (1 lg per transcription) was transcribed to cDNA using the iScript kit (BioRad, Munich, Germany) according to the manufacturer’s instruction. For Real-Time PCR, MicroAmp™ Fast Optical Reaction Plates, Power SYBR™ Green PCR Master Mix, MicroAmp™ Optical Adhesive Film and the 7500 Fast Real Time PCR System (Applied Biosystems, Darmstadt, Germany) were used in a reaction scaled to 10 ll reaction volume.

71

The exon spanning primers for the gastrin-/cholecystokinin-B receptor (GastrinR) were designed from the genomic sequence – accession number D21219.2 (for: GCCTCGTGTCTGCAGTG; rev: CTTGGTGGCGGACCCTGCT). The constitutively expressed glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as reference (for: TGGATCTGACGTGCCGC; rev: TGCCTGCTTCACCACCTTC). For each primer pair a threshold point was chosen in a region where all amplification curves showed exponential growth. From the cycle number at which this threshold was crossed for the samples with GastrinR primers we could calculate the relative GastrinR mRNA content of each sample. By putting these GastrinR mRNA values into relation with the same samples’ constitutively expressed GAPDH mRNA values, we obtained the relative GastrinR mRNA expression level of the tissue the samples were taken from. 2.3. MR-relaxometry For MR relaxometry, PC3 human prostate carcinoma cells were grown to 70–80% confluency in 75 cm2 culture flasks (Corning Costar, Bodenheim Germany), then detached with Accutase™ (PAA laboratories, Pasching, Austria), harvested and subsequently aliquoted into Eppendorf tubes (6  106 cells per tube). The detached cells in the tubes were incubated with the both the correct and scrambled Gd-DOTA-rhodamine-gastrin conjugates (25 lM). After a one-hour incubation period at 37 °C in an atmosphere of 5% CO2, the cells were washed three times in PBS and centrifuged at 800 rpm for 5 min. Concentration was chosen after previous experience with a different, bigger gastrin conjugate (Sturzu et al., 2012). Native control samples were only incubated with medium and washing and centrifugation were performed in line with the other MRI samples. In vitro imaging was performed with a 3 T whole body MRI-system (Trio, circular polarised wrist coil, Siemens, Erlangen, Germany). Sagittal T1-weighted MR images were obtained using the following spin echo sequence: TR (repetition time): 200 ms, TE (echo time): 7.4 ms, flip angle 90°, averages: 1, concatenations: 2, measurements: 1, number of slices: 19, distance factor: 30%, slice thickness: 3 mm, field of view read: 180 mm, field of view phase: 100%, base resolution: 256, phase resolution: 100%, voxel size: 0.7  0.7  3.0 mm, scan time: 1:48 min. T1 relaxation times were evaluated from signal intensities obtained by multiple spin echo measurements: TR: 20–8000 ms (50 different TR values), TE: 6.4 ms, flip angle 90°, averages: 1, measurements: 1, number of slices: 1, slice thickness: 1 mm, field of view read: 120 mm, field of view phase: 87.5, base resolution 128, phase resolution: 100%, voxel size: 0.9  0.9  1 mm. Analyses and calculations were performed using a Matlab program (Math Works, Natick, MA, USA). T1 values were approximated by a three-parameter fit procedure. The relaxation rate value obtained as the reciprocal value of the T1 relaxation time correlates directly to the presence of gadolinium conjugate. All signal curves were examined and found to be monoexponential. The investigations were performed in triplicate. 2.4. Flow cytometry For fluorescence activated cell sorting (FACS), PC3 human prostate carcinoma cells were grown in 75 cm2 culture flasks (Corning Costar, Bodenheim Germany) (70–80% confluency) as described in the confocal laser scanning microscopy and MR relaxometry sections. Accutase™ (PAA laboratories, Pasching, Austria) was added to achieve detachment of the cells, which were then harvested and aliquoted into Eppendorf tubes (Eppendorf, Hamburg,

72

A. Sturzu et al. / European Journal of Pharmaceutical Sciences 52 (2014) 69–76

Germany) (1  106 cells per tube). The cells were incubated in detached state for one hour with the correct and scrambled GdDOTA-rhodamine-gastrin conjugates (25 lM). Then the cells were washed once with PBS and centrifuged at 800 rpm for 5 min. Then 300 ll FACS buffer (D-PBS containing 1% paraformaldehyde) was added. The samples were measured immediately. Approximately 20.000 events were recorded per sample. Fluorescence excitation was achieved by an argon laser (488 nm). Rhodamine fluorescence was detected using a 580–610 nm bandpass filter. All investigations were performed in triplicate. Mean rhodamine fluorescence values of the samples were acquired using the WinMDI software (Joseph Trotter, Scripps Research Institute, San Diego, CA, USA) and then statistically evaluated.

1.5 mg (ca.730 nmol) of conjugate was predissolved in 10 ll pure ethanol and then diluted with 300 ll of PBS. This was done to keep the injection volume as small as possible. Three mice each were injected with solutions of either the correct or scrambled conjugate. After 30 min of circulation the mice were sacrificed and the tumors and organs subsequently removed, embedded in Jung tissue-freezing medium (Leica Instruments, Nussloch, Germany), and frozen in liquid nitrogen. Frozen sections were cut at 5 lm (LC3000 cryotome, Leica Biosystems GmbH, Nussloch, Germany) and rhodamine staining of the frozen sections prepared from the kidney as primary excretory organ for the conjugates and both tumor xenografts were examined by CLSM. Confocal laser scanning microscopy was performed on an inverted LSM510 laser scanning microscope (Carl Zeiss, Jena, Germany) (objective: LD Achroplan 40  0.6). For rhodamine fluorescence excitation, the 543 nm line of a helium–neon laser with appropriate beam splitters and barrier filters were used.

2.5. Mouse tumor xenografts Animal experiments were approved by the Committee for Animal Experiments of the Regional Council (Regierungspräsidium) of Tübingen. PC3 human prostate carcinoma and Colo205 human colon carcinoma were grown to 80% confluency at 37 °C, 5% CO2 (vol/vol) in RPMI-1640 Ready Mix Medium containing L-glutamine and 10% fetal bovine serum (FBS)-Gold (PAA laboratories, Pasching, Austria), in 75 cm2 culture flasks (Corning Costar, Bodenheim Germany). Cells were detached with Accutase™ (PAA laboratories, Pasching, Austria), harvested, washed with phosphate buffered saline (PBS, PAA laboratories, Pasching, Austria) and then drawn up into a 30G Micro-Fine™ insulin syringe (Becton Dickinson, Franklin Lakes, NJ, USA). Approximately 5  105 cells were injected subcutaneously into the lower flanks (PC3: right flank, Colo205: left flank) of male CD1 Nu/Nu mice (Charles River WIGA, Sulzfeld, Germany). Due to relatively low number of cells injected and no use of collagen gel, tumor growth was slower than expected. Experiments were performed after 10–14 days on mice bearing tumors of 3–5 mm diameter.

2.6. Confocal laser scanning microscopy (CLSM)

relative GastrinR mRNA expression [a.u.]

In preparation for CLSM mice bearing PC3 and Colo205 tumor xenografts were treated with the conjugates (three mice per conjugate).

2.7. In vivo magnetic resonance imaging (MRI) Mice bearing PC3 and Colo205 xenografts were anesthesized with a 3-component narcosis (Fentanyl, Midazolam, Medetomidin). Native T1 weighted axial magnetic resonance images were taken in the same 3 T scanner used in the in vitro experiments with a circular polarised wrist coil using the following settings: TR: 800 ms, TE: 9.1 ms, flip angle 90°, averages: 2, measurements: 1, number of slices: 25, slice thickness: 2 mm, field of view read: 60 mm, field of view phase: 52, base resolution 256, phase resolution: 100%, voxel size: 0.2  0.2  2 mm, time: 6:03 min. Then 1.5 mg (ca.730 nmol) of the scrambled conjugate was dissolved as described in the CLSM section and injected intraperitoneally. Post-contrast MR images were taken continuously for 2 h, using the exact same settings as for the native images. Then the correct Gd-DOTA-rhodamine-gastrin conjugate (1.5 mg, ca. 730 nmol) was dissolved and injected intraperitoneally. Again, magnetic resonance images were taken continuously for 2 h. Finally 0.5 ml (250 lmol) of Gd-DOTA (DOTAREM; Guerbet, Paris, France) solution was given intraperitoneally to and images were taken for another 25 min. Then the narcosis was stopped with a 3-component antidote (Naloxon, Flumazenil, Atipamezol) and the mice were returned to their cages. The scrambled and correct MRI contrast agent conjugates as well as the commercial contrast agent were used sequentially on

1.20 1.00 0.80 0.60 0.40 0.20 0.00 PC3

U373

U2OS

Colo205

Fig. 2. mRNA levels analyzed by RealTime-PCR. Mean GastrinR mRNA expression of human PC3 prostate carcinoma, U373 glioma, U2OS osteosarcoma and Colo205 colon carcinoma cell lines was depicted as the quotient of the relative GastrinR mRNA value and the relative mRNA value of the constitutively expressed GAPDH. PC3 cells showed approximately 8800-fold higher expression than the Colo205 cells. U373 and U2OS samples showed much lower expression at only 100-fold and respectively 280-fold higher expression compared to the Colo205 cells. GastrinR expression on Colo205 cells was lowest.

73

A. Sturzu et al. / European Journal of Pharmaceutical Sciences 52 (2014) 69–76

the same mouse. This was done to exclude any possible differences in MRI signal between the correct and scrambled conjugates based on heterogeneity of tumor xenograft growth and development in different test specimen. To obtain the signal-to-noise-ratio the mean signal intensity in a circle of 0.3 mm diameter placed on each of the tumor xenografts and the background were obtained for all four of the different treatments (native, scrambled conjugate, correct conjugate, saturation with DOTAREM). Two mice in which the xenografts were of the same approximate size and localized on a coronal plane orthogonal to the spine were examined in this way.

A

TR=40

TR=80

TR=140

TR=450

Correct Scrambled

B 2.557

MR-relaxometry: 25µM Gd-DOTA conjugates on PC3 cells

3. Results 3.1. Expression analysis for the gastrin receptor (GastrinR) on cell lines The GastrinR mRNA expression level was highest on the PC3 prostate carcinoma cell line (8800-fold higher than the lowest sample Colo205). The gap to the next cell lines U2OS (280-fold higher than Colo205) and U373 (100-fold higher than Colo205) was considerable. The clearly lowest GastrinR expression was exhibited by the Colo205 cell line (Fig. 2).

Relaxation rate [1/s]

2.357 2.157 1.957 1.757 1.557 1.357 1.157 0.957 0.757 0.557 Correct conjugate

3.2. MR relaxometry of the correct and scrambled conjugate on PC3 human prostate carcinoma cells Stronger cellular binding or uptake of the correct conjugate compared to the scrambled conjugate was already visible in the axial images at different repetition times (TR). At a TR of 40 ms the presence of the gadolinium in the sample incubated with the correct conjugate decreased the T1 relaxation by enough to result in an MRI signal while the scrambled conjugate sample was not yet visible. At higher TR values the scrambled conjugate samples became visible but the correct conjugate sample consistently showed stronger signal which meant more gadolinium in the sample. Quantitative evaluation in which the value for the native PC3 cells was used as background resulted in approximately double conjugate binding or uptake of the correct conjugate compared to the scrambled conjugate into the PC3 cells (Fig. 3A and B).

3.3. Fluorescence Assisted Cell Sorting (FACS) Experiments with both the correct and scrambled conjugate on PC3 human prostate carcinoma cells showed that rhodamine staining intensity after incubation with the correct conjugate is twice as high as after incubation with the scrambled conjugate (Fig. 3C).

3.4. Confocal laser scanning microscopy of mouse organ sections after intraperitoneal conjugate application Since the mice were not anesthesized for the intraperitoneal application of the conjugates we could observe that both the correct and scrambled conjugates did not cause any obvious adverse effects during the 30 min circulation time. The organ sections of the kidneys as excretory organs showed similar moderate staining for the correct and the scrambled conjugates. Staining of the PC3 tumor cell xenografts was clearly stronger with the correct conjugate compared to the scrambled conjugate and also to the kidney sections. The Colo205 tumor xenografts showed only very little to no staining at all (Fig. 4).

mean rhodamine fluorescence [a.u.]

C

40 35 30 25 20 15 10 5 0

Scrambled conjugate

native

FACS: 25µM Gd-DOTA conjugates on PC3 cells

Correct conjugate

Scrambled conjugate

Fig. 3. Magnetic resonance imaging and relaxometry. (A) Axial T1 weighted magnetic resonance (MR) images of PC3 prostate carcinoma cells (ca. 1  106 cells in 0.5 ml cups) taken at different repetition times (TR). The three samples incubated with the correct Gd-DOTA-rhodamine-gastrin conjugate (left side) showed first MR signals at a lower TR value (40 ms) and at higher TR values (80 ms, 140 ms) their MR signal was stronger than that of the three samples incubated with the scrambled conjugate (right side). At higher TR value (450 ms) the samples with both the correct and scrambled conjugate showed high MR signal strength. (B) The relaxation rate value (reciprocal value of the T1 relaxation time) correlates directly to the uptake of gadolinium conjugate. The relaxation rate of the PC3 prostate carcinoma cell samples incubated with the correct Gd-DOTA-rhodamine-gastrin conjugate was higher (2.18) than that of the scrambled conjugate (1.26) which in turn was higher still than the value of the native PC3 cell samples (0.56). The value of the native cells was used as background in this diagram. (C) Fluorescence Assisted Cell Sorting (FACS) PC3 human prostate carcinoma cells after incubation with 25 lM Gd-DOTA-rhodamine-gastrin conjugates (correct and scrambled). The mean rhodamine fluorescence of the samples with the correct conjugate was almost double as high as that of the sample with the scrambled conjugate.

3.5. In vivo magnetic resonance imaging In the MRI images taken after application of the scrambled conjugate no signal intensity increase was observed in the tumor xenografts compared to the native images (Fig. 5). Injection of the correct conjugate led to a slight signal intensity increase in the Colo205 xenograft and to a strong increase in the PC3 xenograft. This increase happened during the first 20 min after application and was then constant for the next 100 min. Finally, intraperitoneal application of a high dose of DOTAREM (Guerbet, Paris, France) immediately caused a high MRI signal intensity increase in both tumor xenografts. Other tissues like the leg and back muscles also showed increased MRI signal (Fig. 5). The obtained signal-to-noise-ration (SNR) values clearly showed that the correct

74

A. Sturzu et al. / European Journal of Pharmaceutical Sciences 52 (2014) 69–76

Rhodamine

Superposition Transmission/ Rhodamine

Scrambled conjugate

PC3 Tumor

Correct conjugate

Scrambled conjugate

Colo205 Tumor

Correct conjugate

Fig. 4. Confocal laser scanning microscopy (CLSM). Fusion images (transmission + rhodamine) and rhodamine fluorescence images of 5 lm thick frozen sections of mouse tissue (PC3 tumor xenograft, Colo205 tumor xenograft) were taken after 30 min of circulation with either the scrambled or the correct Gd-DOTA-rhodamine-gastrin conjugate. No staining with either of the conjugates was found in the Colo205 tumor xenografts. The PC3 tumor xenograft was strongly stained by the correct conjugate while the scrambled conjugate did also cause staining but with much lower intensity.

conjugate had a much stronger effect than the scrambled conjugate and also demonstrated a clear PC3 preference (Fig. 5). The signal enhancement achieved with the scrambled conjugate compared to the native image was under 10% in both the PC3 and Colo205 tumor xenografts. With the correct conjugate the enhancement was almost 60% for the PC3 xenograft and just over half that value at 32% for the Colo205 xenograft (Fig. 5). The values for the second mouse not shown in the figure were under 5% enhancement in both xenografts with the scrambled conjugate and a slightly lower enhancement along with a slightly smaller difference between PC3 and Colo205 with enhancement values of 55% and 36% respectively with the correct conjugate.

Signal enhancement values calculated for the images taken after two hours were very similar (within two percent points of the 30 min values). The mice recovered well from the anesthesia and no adverse effects of the conjugates were observed in the days and weeks after treatment. 4. Discussion Until now, the focus in the field of gastrin-related research was more on the gastrin-releasing-peptide-/bombesin-receptor which has been found in prostate carcinoma. Among numerous other studies, in this context, a dual-labelled receptor-targeted peptide

75

A. Sturzu et al. / European Journal of Pharmaceutical Sciences 52 (2014) 69–76

native PC3 Tumor

Xenograft

PC3

Colo205

SNR

3.56

3.08

Colo205 Tumor

Scrambled conjugate 30 minutes

Xenograft

PC3

Colo205

SNR

3.84

3.28

Enhancement [%]

7.86

6.51

Correct conjugate 30 minutes

Xenograft

PC3

Colo205

SNR

5.66

4.06

Enhancement [%]

58.84

32.13

DOTAREM 15 minutes

Xenograft

PC3

Colo205

SNR

7.4

7.4

Enhancement [%]

107.72

140.5

Fig. 5. Magnetic resonance imaging (MRI). Axial T1 weighted MRI images of a CD1 nude mouse carrying PC3 prostate carcinoma cell and Colo205 colon carcinoma cell xenografts in its flanks. In the native image (top) the xenografts showed signal intensities comparable with the surrounding tissue. The lower abdomen showed patches of fat with very high signal intensity. In the image taken 30 min after the intraperitoneal application of the scrambled Gd-DOTA-rhodamine-gastrin conjugate (second from the top) the signal intensity in the xenografts remained virtually unchanged while the lower abdomen with the exception of the test is showed maximal MRI signal intensity due to the intraperitoneal application of the Gd-containing conjugate. The image taken 30 min after intraperitoneal application of the correct Gd-DOTA-rhodamine-gastrin conjugate (third from the top) showed strongly increased signal intensity in the PC3 xenograft and a slight signal intensity increase in the Colo205 xenograft. The complete lower abdomen showed maximal MRI signal intensity resulting from fat and intraperitoneal contrast agent. In the image taken 15 min after the saturation of the mouse with 0.5 ml DOTAREM (i.p.), both the PC3 and Colo205 xenografts showed very high MRI signal intensity. An increase in MRI signal intensity could also be observed in the other tissues (leg and back muscles). Signal-to-noise-ratio (SNR) values and enhancement percentages compared to the native image were calculated and depicted in a table next to the respective image.

conjugate has already been used to target the gastrin-releasingpeptide receptors in prostate carcinoma xenografts in mice by magnetic resonance and optical imaging (Lin et al., 2011). Another G-protein coupled receptor, the gastrin/cholecystokinin-B receptor (GastrinR) has only been used targeting medullary thyroid carcinoma with radiolabelled gastrin peptides (Behr et al., 2001; Breeman et al., 2008). Using GastrinR mRNA expression analysis on four different carcinoma cell lines we now found that GastrinR expression predominates in PC3 prostate carcinoma cells (Fig. 2). The gastrin peptide hormone is naturally active in variations of different length (34, 17, 14 amino acids) with a conserved C-terminus (Rehfeld, 1981). Three different C-terminal gastrin heptapeptide conjugates, each labelled with a different fluorescent dye, have already been examined (Czerwinski et al., 1994). In contrast

to the previous fluorescence-labelled conjugates, we inserted another lysine to enable easy dual-labelling at its a- and e-amino group respectively. In vitro magnetic resonance relaxometry (MRI) and flow cytometry on PC3 cells showed that our conjugate was receptor specific. In both methods our conjugate with the correct gastrin sequence showed approximately double signal intensity compared to the conjugate with the scrambled gastrin sequence (Fig. 3B, C). In vitro MRI experiments were performed directly on the cells. Since the gadolinium that causes the increase in MRI signal is consistently bound to the e-amino group of an N-terminal lysine in both conjugates we were not expecting any difference in T1 relaxivity of the conjugates themselves. Confocal laser scanning microscopy (CLSM) examinations of the organ and tumor sections from mice treated with our conjugates

76

A. Sturzu et al. / European Journal of Pharmaceutical Sciences 52 (2014) 69–76

showed that both the correct and the scrambled conjugate stained the PC3 cell xenografts, albeit at different rates, while the Colo205 cell xenografts showed only traces of staining (Fig. 4). As already observed in in vitro experiments, staining of the PC3 cell xenografts with the correct conjugate was clearly higher than with the scrambled conjugate. Finally the in vivo MRI experiments on mice bearing both PC3 cell and Colo205 cell xenografts could reproduce the in vitro results. No marked increase of MRI signal in the PC3 xenografts was observed after application of the scrambled conjugate. Subsequent application of the correct conjugate led to strong increase of MRI signal in the PC3 xenografts. The MRI signal in the Colo205 xenografts remained at a low level with the correct and mutant conjugate (Fig. 5). The fact that both the PC3 and Colo205 xenografts showed strong MRI signals after application of the DOTAREM contrast agent confirmed that the difference observed with the correct gastrin conjugate was tissue-specific. One point of discussion here is our decision to apply the conjugates intraperitoneally and not intravenously. Intravenous injection would be preferable to achieve higher blood concentrations and avoid the ‘pass first’ effect which means all of the i.p. conjugate is absorbed into the portal vein and has to pass through the liver where it can be metabolized before it can reach its target, the tumor. Because we wanted to have a direct comparison between the correct and scrambled conjugate on the same animal and on the same xenografts we opted for sequential injection. Both repeated i.v. injection over the course of the 4–5 h of this experiment as well as a venous access were impractical because of the diminutive size of the test animals. This left intraperitoneal application as the next best method. With 1.5 mg of conjugate we used as high a dose as possible while staying at 300 lL of injection volume. The effect of this i.p. application was a delayed and reduced delivery into the blood stream compared to i.v. application. Taking previous examinations that were comparing the intraperitoneal and intravenous delivery of the peptide hormone Insulin into account (Schade et al., 1979), we were expecting maximal blood levels 30 min to 2 h after i.p. application at about 10% of the concentration we would have achieved by i.v. injection. We were not concerned about the conjugates being metabolized in the liver during the first pass after absorption into the portal vein. This effect surely has occurred but with the high dose of conjugate (25 lmol per kg body weight) was deemed negligible. By injecting a high dose of conjugate i.p. we created a depot of substance and the combination of intraperitoneal resorption and renal and hepatic elimination resulted in an equilibrium concentration in the blood stream. This effect could be deduced from the very similar enhancement values of the 30 min and 2 h images. Therefore all the images represented a summation of the effects of all the contrast agents that had been applied up to that point. This presented the problem that after sequential application of the conjugates the still present scrambled conjugate could either interfere with the binding of the correct conjugate or that just by the presence of a double dose of gadolinium the MRI signal could be increased in an unspecific manner. While we expected a slight unspecific increase, the observed increase in MRI signal intensity after injection of the correct conjugate was considerable and must be attributed to specific binding. This was also confirmed by the signal-to-noise-ratio values. Another point that remains unclear in this study is whether the association of the gastrin conjugates with the cells and tumor tissues is receptor-bound on the membranes or whether the conjugates have been internalized. In a previous study with a rhodamine-labelled 17 amino acid gastrin conjugate cytoplasmic dot-like staining was observed in confocal laser scanning images after 4 h of incubation (Sturzu et al., 2012). Since the incubation times were all shorter in the present study, one must assume that

the observed staining is still membrane-bound but an eventual internalization of the conjugates is likely. This internalization could lead to the effect that the marking of GastrinR positive tissues persists even after the conjugate has been cleared from the blood stream. Since GastrinR is also present in non-malignant tissue like the brain and the intestinal tract (Dufresne et al., 2006), further studies have to be performed to elucidate the correct conjugate’s specificity in a same-species environment. However in a first step we could demonstrate that our dual-labelled conjugate could differentiate between human prostate carcinoma cell xenografts on one side and human colon carcinoma xenografts as well as normal mouse tissue on the other side in the mouse model. This suggests that the conjugate could be a candidate for further investigations. If the prostate preference could be confirmed, such gastrin conjugates would be of value in the detection and demarcation of prostate tumor lymph node metastases. Acknowledgement This study was supported by the Wilhelm-Sander Foundation and by the Interdisciplinary Center for Clinical Research, University of Tübingen. References Behr, T.M., Gotthardt, M., Barth, A., Béhé, M., 2001. Imaging tumors with peptidebased radioligands. Quart. J. Nucl. Med. 45, 189–200. Breeman, W.A., Fröberg, A.C., de Blois, E., et al., 2008. Optimised labeling, preclinical and initial clinical aspects of CCK-2 receptor-targeting with 3 radiolabeled peptides. Nucl. Med. Biol. 35, 839–849. Camby, I., Salmon, I., Rorive, S., et al., 1994. Characterization of the influence of antihormone and/or anti-growth factor neutralizing antibodies on cell clone architecture and the growth of human neoplastic astrocytic cell lines. J. Neurooncol. 20, 67–80. Camby, I., Salmon, I., Danguy, A., et al., 1996. Influence of gastrin on human astrocytic tumor cell proliferation. J. Natl. Cancer Inst. 88, 594–600. Czerwinski, G., Wank, S.A., Tarasova, N.I., et al., 1994. Synthesis and properties of three fluorescent derivatives of gastrin. Lett. Peptide Sci. 1, 235–242. De Hauwer, C., Camby, I., Darro, F., et al., 1998. Gastrin inhibits motility, decreases cell death levels and increases proliferation in human glioblastoma cell lines. J. Neurobiol. 37, 373–382. Deserno, W.M., Harisinghani, M.G., Taupitz, M., et al., 2004. Urinary bladder cancer: preoperative nodal staging with ferumoxtran-10-enhanced MR imaging. Radiology 233, 449–456. Dufresne, M., Seva, C., Fourmy, D., 2006. Cholecystokinin and gastrin receptors. Physiol. Rev. 86, 805–847. Harisinghani, M.G., Weissleder, R., 2004. Sensitive, noninvasive detection of lymph node metastases. PLoS Med. 1, e66. Harisinghani, M.G., Saksena, M.A., Hahn, P.F., et al., 2006. Ferumoxtran-10enhanced MR lymphangiography: does contrast-enhanced imaging alone suffice for accurate lymph node characterization? Am. J. Roentgenol. 186, 144–148. Hollande, F., Imdahl, A., Mantamadiotis, T., et al., 1997. Glycine-extended gastrin acts as an autocrine growth factor in a nontransformed colon cell line. Gastroenterology 113, 1576–1588. Kucharczak, J., Pannequin, J., Camby, I., et al., 2001. Gastrin induces over-expression of genes involved in human U373 glioblastoma cell migration. Oncogene 20, 7021–7028. Lin, Y.H., Dayananda, K., Chen, C.Y., et al., 2011. In vivo MR/optical imaging for gastrin releasing peptide receptor of prostate cancer tumor using Gd-TTDA-NPBN-Cy5.5. Bioorg. Med. Chem. 19 (3), 1085–1096. Messing, E.M., Manola, J., Sarosdy, M., et al., 1999. Immediate hormonal therapy compared with observation after radical prostatectomy and pelvic lymphadenectomy in men with node-positive prostate cancer. New Engl. J. Med. 341, 1781–1788. Rehfeld, J.F., 1981. Four basic characteristics of the gastrin–cholecystokinin system. Am. J. Physiol. 240, G255–66. Schade, D.S., Eaton, R.P., Friedman, N., Spencer, W., 1979. The intravenous, intraperitoneal, and subcutaneous routes of insulin delivery in diabetic man. Diabetes 28, 1069–1072. Smith, A.M., Watson, S.A., 2000. Gastrin and gastrin receptor activation: an early event in the adenoma–carcinoma sequence. Gut 47, 820–824. Sturzu, A., Sheikh, S., Klose, U., et al., 2012. Novel gastrin receptor-directed contrast agents – potential in brain tumor magnetic resonance imaging. Med. Chem. 8, 133–137. Walsh, P.C., 2002. Surgery and the reduction of mortality from prostate cancer. New Engl. J. Med. 347, 839–840.