Feasibility of experimental BT4C glioma models for somatostatin receptor 2-targeted therapies.

Abstract Somatostatin receptor subtype 2 (sstr2) is regarded as a potential target in malignant gliomas for new therapeutic approaches. Therefore, visualizing and quantifying tumor sstr2 expression in vivo would be highly relevant for the future development of sstr2-targeted therapies. The purpose of this study was to evaluate sstr2 status in experimental BT4C malignant gliomas. Methods. Rat BT4C malignant glioma cells were injected into BDIX rat brain or subcutaneously into nude mice. Tumor uptake of [68Ga]DOTA-(Tyr3)-Octreotide ([68Ga]DOTATOC), a somatostatin analog binding to sstr2, was studied by positron emission tomography/computed tomography (PET/CT). Additionally, subcutaneous tumor-bearing mice underwent PET imaging with 5-deoxy-5-[18F]fluororibose-NOC ([18F]FDR-NOC), a novel glycosylated peptide tracer also targeting sstr2. Ex vivo tissue radioactivity measurements, autoradiography and immunohistochemistry were performed to study sstr2 expression. Results. Increased tumor uptake of [68Ga]DOTATOC was detected at autoradiography with mean tumor-to-brain ratio of 68 ± 30 and tumor-to-muscle ratio of 9.2 ± 3.8 for rat glioma. High tumor-to-muscle ratios were also observed in subcutaneous tumor-bearing mice after injection with [68Ga]DOTATOC and [18F]FDR-NOC with both autoradiography (6.7 ± 1.5 and 4.3 ± 0.8, respectively) and tissue radioactivity measurements (6.5 ± 0.8 and 4.8 ± 0.6, respectively). Furthermore, sstr2 immunohistochemistry showed positive staining in both tumor models. However, surprisingly low tumor signal compromised PET imaging. Mean SUVmax for rat gliomas was 0.64 ± 0.28 from 30 to 60 min after [68Ga]DOTATOC injection. The majority of subcutaneous tumors were not visualized by [68Ga]DOTATOC or [18F]FDR-NOC PET. Conclusions. Experimental BT4C gliomas show high expression of sstr2. Weak signal in PET imaging, however, suggests only limited benefit of [68Ga]DOTATOC or [18F]FDR-NOC PET/CT in this tumor model for in vivo imaging of sstr2 status.

Prognosis of glioblastoma, the most malignant glioma, remains poor [1]. somatostatin receptor subtype 2 (sstr 2 ) has been studied as a promising target in malignant gliomas for new treatment strategies, such as transfer of genes to human glioblastoma cells via targeting adenoviral vectors to sstr 2 [2], or managing experimental gliomas with somatostatin analog that carries chemotherapeutic agent [3]. Preliminary clinical studies have also investigated sstr 2 -targeted radionuclide therapy with yttrium-90-labeled 1,4,7,10-tetraazacyclododecane-1,4,7, 10-tetraacetic acid-(Tyrosine 3 )-octreotide ([ 90 Y] doTAToC) in human gliomas [4,5]. The therapeutic mechanism in all these new treatment strategies depends on sstr 2 expression, which is a prerequisite for a favorable effect. Therefore, a suitable experimental glioma model with established sstr 2 expression is warranted. Furthermore, recognizing the sstr 2 expression of the tumor in vivo before the potential therapy would be beneficial in order to optimize the treatment protocols [6]. in vivo PET imaging of gliomas with radioligands binding to sstr 2 , such as Gallium-68-labeled doTA-(Tyr 3 )-octreotide ([ 68 Ga]doTAToC), however, has not been previously studied systematically for either experimental nor human gliomas.
[ 68 Ga]doTAToC PET has been extensively studied with neuroendocrine tumors and more recently with meningeomas both of which express sstr 2 [7,8]. [ 68 Ga]doTAToC demonstrates favorable pharmacokinetic properties with rapid binding to its G-protein coupled receptor as well as rapid clearance via the kidneys [7]. However, introduction of hydrophilic sugar moieties to the peptide tracer has shown to improve various properties of the tracer, such as enhanced penetration of the bloodbrain barrier and increased tumor accumulation due to higher affinity [9]. Therefore, a new sstr 2 targeting glycosylated peptide tracer 5-deoxy-5-[Fluorine -18] fluororibose-1-nai 3 -octreotide ([ 18 F]FdR-noC) was explored for the first time in preclinical practice [10].
The aim of this study was to evaluate experimental bT4C glioma models (orthotopic glioma in bdiX rats and subcutaneous tumors in nude mice) for expression of sstr 2 by means of immunohistochemistry, Western blot, uptake of [ 68 Ga]doTAToC and [ 18 F]FdR-noC by PET/CT in vivo, and by autoradiography and measurements of tissue radioactivity ex vivo. Additionally, [ 18 F]FdG PET/CT was performed to mice with subcutaneous tumors. To our knowledge, sstr 2 status and [ 68 Ga]doTAToC uptake has not been previously studied in experimental glioma models. Furthermore, a novel glycosylated peptide tracer [ 18 F]FdR-noC targeting sstr 2 is introduced.

Material and methods
Animal models bT4C rat glioma cell line was originally derived from fetal bdiX rat brain cells, which were transferred to long-term culture after in vivo exposure to n-ethylnitrosurea [11]. Cells were grown in dulbecco's modified Eagle medium (dmEm) supplemented with 10% fetal calf serum (Gibco, Paisley, scotland), 2 mm glutamine, 2 mm sodium pyruvate and 50 mg/ml streptomycin at 37°C in the presence of 5% Co 2 . male bdiX rats (Charles River, France) weighing 263-354 g (n  9) were obtained for the stereotactic injections of bT4C glioma cells into the rat brain as described earlier [12]. briefly, rats were anesthetized intraperitoneally (i.p.) with ketamine and medetomidine and placed into stereotactic apparatus (Kopf instruments, Tujunga, CA, usA). A total of 10 000 bT4C cells in 5 ml of optimem medium were slowly injected with a 25 ml Hamilton syringe 27 G needle (Hamilton, bonaduz, switzerland) into the right corpus callosum (coordinates: 1.0 mm caudal to bregma, 2.0 mm right to sutura sagittalis and depth of 2.5 mm). needle was left in place for 5 min before removal to avoid backflow. The presence of rat bT4C orthotopic gliomas was verified by magnetic resonance imaging (mRi) 13 days after bT4C cell injection. Animals were kept under isoflurane anesthesia and mRi was performed using the 4.7 T magnet (magnex, Abington, uK) interfaced to varian unity inova (Palo Alto, CA, usA) console. T2-weighted spin-echo sequence (TR/TE  2.5 s/0.08 s) was used. image reconstruction was performed using matLab 7.1. software (The mathWorks inc., natick, mA, usA).
To establish subcutaneous tumor model, male athymic nude mice (Harlan) were kept under isoflurane anesthesia (induction 3-4% and maintenance 2-3%) while 250 000 bT4C cells in 100 ml of matrigel (matrigel matrix, bd biosciences) were injected subcutaneously into the neck and left limb of each mouse (n  16).
All animal care and experiments were approved by the national Animal Experiment board in Finland (21.12.2011 and carried out in compliance with Finnish laws relating to the conduct of animal experimentation. PET data acquired in list mode were iteratively reconstructed with a two-dimensional (2d) ordered subsets expectation maximization algorithm (osEm2d) with 4 iterations and 16 subsets into 12  5 s, 12  30 s, 8  60 s, 3  300 s and 3  600 s frames. A reconstructed image had 38 frames and 128  128 159 matrix size with a voxel size of 0.776  0.776  0.796 mm.
Quantitative analysis was performed by defining spherical volumes of interest (voi) on the tumor, normal brain, heart, liver and kidney. brain tumor was localized next to the cranial drill hole. Timeactivity curves (TACs) were extracted from the corresponding dynamic images (Carimas 2.8; Turku PET Centre, Turku, Finland; http://www.turkupetcentre.fi/ carimas). The average radioactivity concentration in the voi (kbq/mL) was used for further analyses. The uptake was reported as standardized uptake value (suv), which was calculated as a ratio of radioactivity of the voi and injected radioactivity (kbq) divided by body weight (g). The radioactivity remaining in the tail was compensated.
Tumor uptake in [ 18 F]FdG PET/CT was reported as scaled suv that takes into consideration the blood glucose levels measured before and after the scanning. scaled suv is calculated as the mean suv multiplied by blood glucose level of individual animal divided by mean blood glucose level of all the animals.
dynamic PET data from bdiX rats was used for [ 68 Ga]doTAToC tracer kinetic modeling. Logan plots for rat bT4C orthotopic gliomas with muscle as a reference tissue were applied to attain distribution volume ratio (dvR). We used 10 min as the starting point for the linear regression. binding potential (bP) corresponds to the density of available receptors and is calculated as bP  dvR -1. muscle was selected as a reference tissue due to its low sstr 2 expression.

Autoradiography analysis
Tumor, brain and muscle uptake of [ 68 Ga]doTA-ToC and [ 18 F]FdR-noC were studied with digital autoradiography. Tissues were dissected immediately after PET/CT imaging, frozen in dry ice and sectioned with cryomicrotome into serial 8-and 20-mm sections at -15°C. Tissue sections were then thaw-mounted onto microscope slides and opposed to an imaging plate (Fuji imaging Plate bAs-TR 2025, Fuji Photo Film Co., Ltd., Tokyo, japan). After an exposure time of 2.5 h for [ 68 Ga]doTAToC and 4 h for [ 18 F]FdR-noC, the imaging plates were scanned with the Fuji Analyzer bAs-5000 (Fuji Photo Film Co., Ltd., Tokyo, japan; internal resolution of 25 mm).
images of tumor, brain and muscle sections were analyzed for count densities (photostimulated luminescence units [PsL]/mm 2 ) with an image-analysis program (Tina 2.1, Raytest isotopenmessgeräte GmbH, straubenhardt, Germany). Circular Rois were defined for the tumor in accordance with hematoxylin and eosin staining from areas of highest tumor cellularity. The background count densities were subtracted from the image data. several tissue sections were analyzed for each animal and results are expressed as mean values ( sd).

Sstr 2 immunohistochemistry and Western blot analysis
Hematoxylin and eosin (H&E) staining and sstr 2 immunohistochemistry were performed on 8 mm slides adjacent to the 20 mm slides used for autoradiography analysis. The sstr 2 monoclonal antibody umb-1 (Abcam, Cambridge, uK) has previously been extensively characterized [14,15]. Endothelial cells known to express sstr 2 functioned as positive internal control and brain white matter and muscle as negative control [16]. Additional immunohistochemical stainings (Cd34 and Cd68) and supplementary Table i (available online at: http:// informahealthcare.com/doi/abs/10.3109/0284186X. 2014.925577)with details of the conditions used for the immunohistochemistry are presented in the Appendix. sstr 2 expression in bT4C cell line was further studied in vitro using Western blot analysis (detailed protocol in the Appendix). Tissue samples from rat adrenal gland, kidney, and spleen known to express sstr 2 served as positive controls [17].

Statistical analyses
data are presented as mean sd. normality was tested with shapiro-Wilk test. statistical dependence between bP and suv was calculated with spearman's rank correlation. independent t-test was used for parametric variables and independent-samples mann-Whitney u-test was applied for non-parametric variables to compare the uptake between the tumor and reference tissues, and the uptake between [ 68 Ga] doTAToC and [ 18 F]FdR-noC in biodistribution data. p-values were calculated using two-tailed test and less than 0.05 was considered statistically significant. statistical analyses were conducted using sPss 17 for mac (sPss, inc., Chicago, iL, usA).

Animal model and BT4C glioma histology
bT4C glioma cell injection induced intracranial tumor in seven bdiX rats of nine as confirmed by mRi 13 days after injection. The H&E staining demonstrated tumors with high cellularity, small necrotic areas and an infiltrative border. neoplastic cells displayed marked pleomorphism and nuclear atypia. mitotic figures were frequent and neovascularization was also detected. Histologically, rat orthotopic bT4C glioma was highly similar to human glioblastoma. mean tumor dimensions were 4.2  3.0 mm. subcutaneous bT4C tumors were induced successfully in all nude mice reaching an average size of 5.4  4.7 mm. Histologically they resembled orthotopic bT4C glioma, however, with less neovascularization and necrotic areas. in addition, subcutaneous bT4C tumors presented diffuse leukocytosis and areas of edematous subcutaneous fat tissue. The calculated bP for [ 68 Ga]doTAToC in the rat orthotopic bT4C gliomas using muscle as a reference tissue was 0.76  0.39 (n  5). Logan plot from one rat was excluded from the analysis due to low R-squared value of 47%. For other plots, mean R-squared value was 99.8% meaning good fitting of the data to the kinetic model used. no correlation, however, was found between bP and suv max values obtained from PET data from 30 to 60 min after injection.

Tissue radioactivity measurements and stability of [ 68 Ga]DOTATOC and [ 18 F]FDR-NOC
Ex vivo biodistribution data from tissue radioactivity measurements at 60 min after injection is presented in Table i. subcutaneous bT4C tumors in mice showed high uptake of both [ 68 Ga]doTAToC and [ 18 F]FdR-noC when compared to the reference tissue muscle (p  0.001 for both tracers). between the two tracers, tumor uptake was higher with [ 18 F] FdR-noC (p  0.021). However, due to increased radioactivity concentration also in reference tissues, the tumor-to-muscle and tumor-to-brain ratios were actually diminished with [ 18 F]FdR-noC compared to [ 68 Ga]doTAToC. Rat orthotopic bT4C gliomas were not analyzed with a gamma counter in order to obtain intact brain tumor for autoradiography analysis.
Tissue radioactivity measurements after [ 68 Ga] doTAToC injection showed high values in the kidneys and urine and low values in other non-target organs (Table i). Rapid clearance from blood and renal excretion was also confirmed by [ 68 Ga]doTA-ToC uptake data obtained from 60 min dynamic PET imaging in mice (Figure 2). [ 18 F]FdR-noC, however, in addition to elimination via kidneys was also eliminated via hepatobiliary pathway as indicated by significantly increased radioactivity in the liver and small intestine compared to [ 68 Ga] doTAToC (p  0.001).
in vivo stability of [ 68 Ga]doTAToC in plasma of bdiX rats and nude mice at 60 min after injection was excellent; the percentage of intact Receptors were homogenously distributed in rat glioma with positive staining in more than 95% of the neoplastic cells. subcutaneous bT4C tumors in nude mice showed more heterogeneous sstr 2 immunostaining compared to orthotopic gliomas with positive staining in 40-80% of the neoplastic cells. The staining pattern was mainly intracellular but also membranous staining was detected. Results of Cd34 and Cd68 immunostaining are presented in the Appendix.
Discussion suitable experimental glioma model expressing sstr 2 is warranted as the implications for sstr 2 -targeted therapies in gliomas continue to expand. in addition, detecting sstr 2 in vivo by PET imaging would be highly beneficial for the planning of therapies targeting sstr 2 and in the follow-up of these strategies. in this study, we validated the sstr 2 status of two bT4C glioma models with PET/CT imaging, autoradiography, tissue radioactivity measurements, [ 68 Ga]doTAToC was 99-100%. Also [ 18 F]FdR-noC showed high in vivo stability with 94-96% of intact tracer in plasma at 60 min after injection. major radioactive metabolite in plasma was present at retention time 4.85 min. intact [ 18 F]FdR-noC detected in urine 60 min after injection varied from 10% to 81%.

Autoradiography
The autoradiography analysis (Table ii) showed a significant uptake of [ 68 Ga]doTAToC in the rat orthotopic bT4C glioma when compared to contralateral normal brain (p  0.002) or muscle (p  0.002) with mean tumor-to-brain ratio of 68.1  30.4 and tumor-to-muscle ratio of 9.2  3.8. Also subcutaneous bT4C tumors presented in autoradiography high uptake of both [ 68 Ga]doTAToC and [ 18 F] FdR-noC compared to muscle (p  0.001 for both) with tumor-to-muscle ratio of 6.7  1.5 and 4.3  0.8, respectively. uptake of [ 68 Ga]doTAToC within the orthotopic bT4C glioma was mainly uniform as demonstrated in Figure 3. slightly higher uptake was observed in subcutaneous bT4C tumors in areas of edematous fat tissue. These areas, however, were disregarded from analysis.

Immunohistochemistry and Western blot analysis
immunohistochemical staining of sstr 2 was distinctly positive in rat orthotopic bT4C glioma our autoradiography analysis demonstrated a 68-fold higher uptake of [ 68 Ga]doTAToC in rat orthotopic bT4C glioma compared to contralateral normal brain. doTAToC binds predominantly to sstr 2 thus verifying 'functional' sstr 2 in bT4C tumors. in line with this, sstr 2 expression was also immunohistochemically observed in majority (95%) of rat bT4C glioma cells. The binding potential for [ 68 Ga] doTAToC in rat gliomas (0.76  0.39) was positive though remained at a low level. it must thus be speculated that [ 68 Ga]doTAToC tumor accumulation, in addition to active receptor binding, may also be passive due to a defective blood-brain barrier. disruption of blood-brain barrier is manifested also in rat bT4C glioma as indicated by mRi contrast enhancement [18]. The relationship between bloodbrain barrier integrity and uptake of [ 68 Ga]doTA-ToC is important for further validation of sstr 2 -targeted therapy in glioma and is in the focus of our ongoing study. Rat bT4C orthotopic gliomas were all visualized in PET/CT with mean suv max 0.64  0.28 from 30 to 60 min from injection with [ 68 Ga]doTAToC. However, most subcutaneous bT4C tumors could not be visualized with [ 68 Ga]doTAToC PET/CT possibly relating to less intensive immunohistochemical sstr 2 staining in subcutaneous tumors. Consequently, we wanted to study whether glycosylation of the tracer peptide would improve tumor uptake and induce sufficient signal intensity for PET imaging. it has been well documented that sugar conjugation is a powerful method to improve tracer pharmacokinetics [9,19]. schottelius et al. studied the pharmacokinetics of a sstr 2 agonist conjugated with different carbohydrates, such as glucose and maltose, and found elevated tumor uptake with excellent tumor-to-organ ratios [9]. using [ 18 F]FdR as a prosthetic group for labeling peptides is a novel method for glycosylating peptides for PET [10,13]. indeed, significant increase in [ 18 F]FdR-noC uptake was detected in subcutaneous bT4C tumors compared to [ 68 Ga]doTAToC with both methods of autoradiography and tissue radioactivity measurements. visualization of the subcutaneous tumors with [ 18 F]FdR-noC PET/CT, however, was not   Table i. Time-activity curves from dynamic PET data show that liver uptake of [ 18 F] FdR-noC was elevated already during the first minutes after injection and remained high for the whole 60 minutes suggesting rapid excretion via hepatobiliary pathway (Figure 2). in addition, [ 18 F] FdR-noC was eliminated through the kidneys with an extent similar to [ 68 Ga]doTAToC. This did not correspond to earlier studies where glycosylation of the tracer peptides improved pharmacokinetics due to increased hydrophilicity of the tracer leading to favored excretion pathway via kidneys and reduction in liver uptake [9,19]. our opposite finding of increased liver uptake with glycosylated [ 18  [ 18 F]FdR-noC uptake was significantly higher in almost all non-target organs, including brain, when compared to [ 68 Ga]doTAToC uptake in mice bearing subcutaneous tumors (Table i). This may be due to improved biological properties of the glycosylated tracer, such as bioavailability or penetration of the bbb as discussed earlier. secondly, tracer peptides ToC and noC have different affinity profiles for somatostatin receptor subtypes. While both peptides have high affinity for sstr 2 , noC demonstrates stronger binding also to somatostatin receptor subtypes 3 and 5 (sstr 3 and sstr 5 ) [22]. sstr 2 is known to be the most abundant subtype in most tumor types. However, the effect of sstr 3 and sstr 5 expression to [ 18 F]FdR-noC uptake cannot be excluded since their expression was not analyzed in this study. sstr 2 immunostaining with umb-1 antibody detected sstr 2 protein expression throughout the rat orthotopic bT4C glioma in almost all tumor cells. staining pattern in subcutaneous bT4C tumors was more heterogeneous with areas of negative staining as well. umb-1 is a monoclonal antibody that has been thoroughly studied in recent years and is considered to be the best available tool for semiquantitative characterization of tumor sstr 2 expression [15]. in addition, reliability of umb-1 staining was warranted in our study with both positive and negative internal controls. The subcellular localization of sstr 2 immunoreactivity was mainly cytoplasmic in both orthotopic and subcutaneous bT4C tumors. This corresponds well to a previous in vivo study where internalization of sstr 2 in rat AR42j tumor cells was seen already 2.5 min after i.v. injection of sstr 2 agonist and was still detectable at 6 h after injection [14]. sstr 2 receptor internalization and subsequent accumulation into the tumor cells is important especially considering the efficacy of targeted radiotherapy with somatostatin analogs, such as [ 90 Y]doTAToC.
[ 68 Ga]doTAToC PET/CT is in clinical use for staging and follow-up of neuroendocrine tumors. it is also used to evaluate the suitability of the tumor for in vivo sstr 2 targeting radiotherapy since neuroendocrine tumors have shown good correlation between sstr 2 immunostaining and increased uptake of [ 68 Ga] doTAToC in PET/CT imaging [23]. Körner et al., studied immunohistochemistry of neuroendocrine tumors with umb-1, the same antibody used in our study, and correlated it to the actual receptor levels detected by autoradiography [15]. They found that intensive umb-1 staining in more than 10% of tumor cells predicted high autoradiographic receptor levels, which is a prerequisite for therapeutic sstr 2 targeting. in our study, especially rat orthotopic bT4C gliomas showed very intensive positive staining with umb-1 antibody in more than 95% of the glioma cells suggesting a potential role for this experimental model in sstr 2 -targeted therapies.
We have validated sstr 2 status of two bT4C glioma models and conclude that these tumors express sstr 2 as demonstrated with autoradiography, tissue radioactivity measurements, immunohistochemistry, and Western blotting. Especially rat orthotopic bT4C glioma is warranted to be further evaluated in treatments attempting to target sstr 2 in malignant glioma. unfortunately, follow-up using [ 68 Ga]doTAToC or [ 18 F]FdR-noC PET/CT imaging was suboptimal due to low tumor signal.

Declaration of interest:
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
This study was conducted within the Finnish Centre of Excellence in Cardiovascular and metabolic diseases supported by the Academy of Finland, university of Turku, Turku university Hospital, and Åbo Akademi university. in addition, the study was financially supported by a grant from the Foundation for the Finnish Cancer organisations. Aida Kiviniemi is a Ph.d. student supported by the national Graduate school of Clinical investigation.