Preclinical Evaluation of Radiolabeled Peptides for PET Imaging of Glioblastoma Multiforme

In this study, we have compared four 68Ga-labeled peptides (three Arg-Gly-Asp (RGD) peptides and substance-P) with two 18F-tracers clinically approved for tumor imaging. We have studied in vitro and in vivo characteristics of selected radiolabeled tracers in a glioblastoma multiforme tumor model. The in vitro part of the study was mainly focused on the evaluation of radiotracers stability under various conditions. We have also determined in vivo stability of studied 68Ga-radiotracers by analysis of murine urine collected at various time points after injection. The in vivo behavior of tested 68Ga-peptides was evaluated through ex vivo biodistribution studies and PET/CT imaging. The obtained data were compared with clinically used 18F-tracers. 68Ga-RGD peptides showed better imaging properties compared to 18F-tracers, i.e., higher tumor/background ratios and no accumulation in non-target organs except for excretory organs.


Introduction
Glioblastoma multiforme (GBM) is an aggressive primary brain tumor and, despite the current intensive therapy regimen, median overall survival period is only 16-21 months from the date of diagnosis [1]. Therefore, early and precise diagnosis of GBM is a key aspect for appropriate patient care approach [2]. Imaging techniques play an important role in the diagnosis of GBM and subsequent disease monitoring. Although magnetic resonance imaging (MRI) is the first-line technique for GBM imagining, positron emission tomography (PET) is increasingly used in gliomas diagnosis as well [3,4]. The diagnostic information acquired with 18 F-fluorodeoxyglucose ( 18 F-FDG), the gold standard in clinical PET tumor imaging, is relatively variable; however, amino acid PET tracers, such as 11 C-methionine ( 11 C-MET), 18 F-fluoro-L-DOPA ( 18 F-FDOPA) or 18 F-fluoroethyltyrosine ( 18 F-FET) allow a more accurate glioma characterization [5]. The better tumor characterization could also be the case of more recent radiotracers developed for PET imaging of α v β 3 integrins [4,6]. α v β 3 integrin is highly expressed in glioma cells and generally in neovascular endothelial cells, in conjunction with tumor-related angiogenic processes [7]. The expression rate of α v β 3 integrin increases with the malignancy grade of the glioma and is particularly important in high-grade gliomas, where this integrin may facilitate tumor progression [8]. The radiotracers that bind α v β 3 integrin contain one or more tripeptide Arg-Gly-Asp (RGD) amino acid sequences providing high affinity towards α v β 3 integrins [9][10][11][12]. Although some of these RGD peptides were labeled with fluorine-18, this approach appears not to be convenient for practical use in terms of demanding radiochemical synthesis despite its great diagnostic potential [10,13].
The preparation of RGD peptides, labeled with radiometals like gallium-68 via a metal complexing approach, is much more straightforward in preclinical as well as in clinical settings [6,12]. Recently, RGD peptide coupled with a suitable macrocyclic chelator (e.g., dimeric 68 Ga-NOTA-PRGD 2 peptide) was successfully tested in human gliomas and showed better properties than 18 F-FDG in differentiating high and low-grade gliomas [6]. Apart from RGD peptides and 18 F-FDG, radiolabeled substance P that targets NK-1 receptors is another promising approach for imaging GMB [14]. In addition to the aforementioned tracers, there were also attempts to visualize glioblastoma with 18 F-FLT [15].
Nevertheless, a comparison of 18 F-FDG has never been performed to date with NODAGA-c(RGDyK) 2 , a dimeric RGD peptide employing NODAGA as the chelating moiety, with RGD monomer DOTA-c(RGDfK) and DOTA-substance P in preclinical GBM imaging. Despite the fact that the affinity for α v β 3 integrins is lower for monomeric and dimeric than multimeric RGD peptides, monomeric and dimeric RGD peptides usually have more favorable pharmacokinetic and imaging properties in vivo. The smaller size of mono-and dimeric RGD peptides result in faster pharmacokinetics and are often a more suitable biodistribution profile for tumor imaging [11,14]. In this study, we have focused on the preclinical evaluation of 68 Ga-labeled peptides (DOTA-c(RGDfK), NODAGA-c(RGDyK), NODAGA-c(RGDyK) 2 , and DOTA-substance P) and comparison of their in vivo behavior with two clinically established 18 F-based tumor radiotracers ( 18 F-FDG and 18 F-FLT) in a mouse xenograft model of human glioblastoma. Our glioblastoma model is based on U-87 MG cell line, a well-established tumor model expressing α v β 3 integrins and NK-1 receptors [16,17].

Radiolabeling of Studied Peptides
The radiochemical purity of 68 Ga-labeled peptides determined by HPLC was 95%. The lowest radiochemical purity of 96.3% was reported for 68 Ga-DOTA-substance P, while the highest radiochemical purity of 99.8% was observed for 68 Ga-DOTA-c(RGDfK)

In Vitro Characterization
The results of in vitro evaluation of 68 Ga-labeled peptides are presented in Table 1. All four studied peptides appeared as hydrophilic entities having partition coefficient values around -3. The three studied RGD peptides showed low plasma protein binding values ranging from 1% to 3%, while substance P displayed slightly different results with approximately 29% fraction bound to the plasma proteins. Another tested in vitro parameter was the stability of 68 Ga-peptides in the presence of an excess of Fe 3+ (0.1 M FeCl 3 ). RGD peptides showed high stability of 95-99% under these conditions. However, the stability of 68 Ga-DOTA-substance-P in the presence of a competing metal was 79% after 120 min of incubation. In the presence of an excess concentration of another competing chelator (6 mM DTPA), RGD peptides displayed stability of 99%, while the stability of substance P decreased to 56% after 120 min of incubation. The results of stability test in neutral pH and human plasma were in the range of 98-99% for all four tested 68 Ga-peptides (Table 1). Table 1. In vitro characterization of tested peptides. Partition coefficient (log P), human plasma protein binding and in vitro stability of 68 Ga-peptides under various conditions presented as the mean of radiochemical purity ± standard deviation (n = 3 and n = 1 for 68 Ga-DOTA-substance P) at different time points.

In Vitro Uptake
Uptake studies using U-87 MG cell line showed the accumulation of approximately 0.4-0.8% of the applied dose in cells, with the highest accumulation observed for 68 Ga-NODAGA-c(RGDyK) 2 and the lowest for 68 Ga-DOTA-c(RGDfK) (Figure 1). The presence of an excess of a cold ligand resulted in the decrease of the uptake of radiolabeled ligand in all tested 68 Ga-RGD peptides except 68 Ga-DOTA-substance P, which had an uptake of 0.3% of the applied dose under both normal and blocked variants of the experiment. For details, see Figure 1.

In Vivo Stability of 68 Ga-Peptides
The urine of animals injected with 68 Ga-peptides was collected 30 and 90 min after the administration for HPLC analysis, and the respective radiochromatograms of 68 Ga-peptides and urine are summarized in Figure 2. The HPLC analysis of 68 Ga-RGD peptides compared to appropriate urine samples showed no major shift in the retention time and the shape of the chromatographic peaks is almost similar to the peaks of observed with original peptides and urine samples. The percentage of radiochemical impurities is practically negligible. However, the urine samples of animals injected with 68 Ga-DOTA-substance P show a clear shift in the retention time of main peak (approximately 1 min) and appearance of new peaks suggesting the formation of more hydrophilic metabolites.

Ex Vivo Biodistribution
The biodistribution results were similar for all four tested 68 Ga-peptides showing very low accumulation of the tracers in non-targeted tissues ( Figure 3). The highest radioactive accumulation was registered in kidneys and ranged between 6-11% ID/g for 68 Ga-RGD peptides and 15-17% for 68 Ga-DOTA-substance P. In summary, all tested 68 Ga-labeled peptides showed rapid pharmacokinetics with the main renal excretion route and very low accumulation in non-targeted tissues. Ex vivo biodistribution data for 18 F-FDG displayed relatively low accumulation in non-targeted tissues (1-8% ID/g) with the exception of heart, where the accumulation reached 40% ID/g. 18 F-fluorothymidine ( 18 F-FLT) showed uniform accumulation and retention of 5-8% ID/g in all examined tissues for 30 and 90 min post injection (p.i.). Both 18 F-tracers had relatively higher accumulation in non-targeted tissues and slower elimination from blood compared to 68 Ga-compounds. Comparative tumor to blood/kidney/muscle ratios were calculated from the data obtained from ex vivo biodistribution studies ( Table 2). Tumor to blood ratios 30 min p.i. were in the range of 1-2.3 for all tested 68 Ga-peptides. Ninety minutes p.i., the tumor to blood ratio values ranged from 14-23 for 68 Ga-RGD peptides and 2.8 for 68 Ga-substance P. Similarly, the tumor to blood ratio 30 min and 90 min p.i. for 18 F-FDG was 2.0 for both time points and 1.1 and 2.4, respectively, for 18 F-FLT. Table 2. Tumor to blood, tumor to muscle and tumor to kidney ratios calculated from ex vivo biodistribution data for four studied 68 Ga-peptides and two 18   2.6. µPET/CT Imaging PET/CT imaging of 68 Ga-labeled peptides clearly showed the highest accumulation of tracer in the kidneys and bladder U-87 MG tumor xenografts were visualized with all three 68 Ga-RGD peptides, with the highest accumulation of radioactivity noted for 68 Ga-NODAGA-c(RGDyK) 2 . 68 Ga-substance P did not visualize the tumor in any of the studied time intervals. PET/CT of U-87 MG tumor xenografts with 18 F-FDG tracer allowed clear imaging of the tumor; however, high radioactivity signal was also observed in non-target tissues such as spleen, heart, brain, and brown adipose tissue. 18 F-FLT PET/CT imaging of U-87 MG xenografted mice 90 min p.i. showed the accumulation of the tracer mainly in the gastrointestinal tract, spleen, and bladder. The tumor on the right flank of the animal could also be imaged with 18 F-FLT. The acquired three-dimensional PET/CT images are shown in Figure 4. Using acquired PET data we have determined standard uptake values (SUV) for the tumor and brain regions and calculated tumor-to-brain ratios listed in Table 3. These ratios have ranged from 0.64 up to 4.70.

Discussion
MRI imaging and tumor biopsy still remain as the diagnostic gold standard for GBM. Nevertheless, various studies now confirm the diagnostic potential and advantages of PET imaging for GBM patients due to better tumor delineation or monitoring tumor progression [18,19]. Recently, radiolabeled RGD peptides were successfully used to image other human carcinomas with PET technology [20] and RGD-sequence peptides remain emerging candidates for glioblastoma imaging [21][22][23].
In the past 20 years, RGD peptides underwent relatively extensive preclinical development, beginning with 18 F-based 18 F-galacto-RGD peptide [12]. This promising tracer was substituted with 68 Ga-labeled variants due to its better targeting properties and more simple synthesis. The cyclization of the structure and its multimerization was also part of the development, which indicated that dimerization of the structure and cyclic form show the best in vivo imaging results. The multimeric (four and more moieties) RGD structure has the disadvantage of slower pharmacokinetics and higher image background. Among the most studied RGD-based tracers, RGDyK and RGDfK sequences, the published data are more favorable towards RGyK due to its faster renal elimination [24]. To our knowledge, there are very few complex reports focusing on the comparison of traditional 18 F-tracers with 68 Ga-labeled RGD peptides for glioblastoma imaging. In this comparative study, we reveal the potential of 68 Ga-RGD peptides under preclinical settings and show their advantages over clinically used 18 F-FDG and 18 F-FLT. We cover here the whole issue ranging from radiolabeling, in vitro characterization, in vivo biodistribution studies to PET/CT imaging of U-87 MG xenografted tumors in mice.
The radiolabeling of studied peptide tracers with 68 Ga resulted in excellent radiochemical purity (97-99%) for all RGD peptides and slightly lower value for 68 Ga-substance P (96%). DOTA modified tracers required elevated reaction temperatures (95 • C), while NODAGA-modified tracers provided excellent labeling results at room temperature. Log P values of all peptide tracers indicated that these molecules are highly hydrophilic, which was subsequently confirmed by animal experiments.
The plasma protein binding values of all three tested RGD peptides were very low (1-3%), except 68 Ga-DOTA-substance P that showed a higher (about 29%) protein binding value, what is in accordance with previously published data [12]. Nevertheless, we did not register much higher blood levels of 68 Ga-DOTA-substance P in animals compared to 68 Ga-RGD peptides. In the presence of an excess of competing chelator (DTPA), all tested 68 Ga-RGD peptides displayed excellent stability (99.06-99.65%), while 68 Ga-DOTA-substance P had an average stability. The presence of an excessive amount of Fe 3+ ions had lower effect on radiochemical purity, resulting in the lowest stability again for 68 Ga-DOTA-substance P. These findings confirmed previously published data, indicating that NODAGA has better chelating properties for 68 Ga 3+ compared to DOTA [12]. Additionally, all 68 Ga-peptides showed excellent stability under physiological pH and in human plasma, with a negligible decrease in the radiochemical purity. These findings are in general in accordance with previous published data [12].
The specificity of tracer binding to U-87 cells was tested by incubating the cells with an excess concentration of a cold tracer. The highest binding was detected for 68 Ga-NODAGA-c(RGDyK) 2 , while 68 Ga-DOTA-c(RGDfk) and 68 Ga-DOTA-substance P showed the lowest uptake of about 50% compared to 68 Ga-NODAGA-c(RGDyK) 2 . This is in accordance with previous studies that showed the benefits of multimerization and the superiority of RGDyK sequence over RGDfK [11,25]. The RGD dimer possesses higher cell binding property compared to monomer as shown by Li et al. [11]. Blocking the in vitro uptake of 68 Ga-tracers by cold ligand was successful for all RGD peptides except 68 Ga-DOTA-substance P, despite the fact that U-87 MG cells express neurokinin (NK1) receptor, the target of substance P [26].
Determination of in vivo stability of studied tracers in mice urine showed that RGD peptides are excreted in unchanged form, while 68 Ga-DOTA-substance P is rapidly metabolized and excreted in different forms compared to the parent molecule. This is clearly visible in radiochromatograms (Figure 2), where the 68 Ga-substance P peak is shifted and branched in urine samples. HPLC analysis of urine from 68 Ga-RGD peptides-injected animals showed a similar radiochromatogram as radiolabeled peptide. These findings indicate that 68 Ga-DOTA-substance P have a poor in vivo stability compared to other tested 68 Ga-RGD peptides.
Ex vivo biodistribution studies revealed the highest uptake of 68 Ga-tracers in kidneys as the principal organ of their excretion, where 68 Ga-substance P had the highest uptake of 15% ID/g. The GBM tumors had relatively low uptake values of about 2% ID/g, except for 68 Ga-NODAGA-c(RGDyK) 2 with 6% ID/g for 30 min p.i. These findings are in accordance with previously published data by other research groups, where tumor accumulation for RGD peptides ranges from 2.5% to 4% ID/g [11,27]. Ex vivo biodistribution data for 68 Ga-DOTA-substance P showed relatively low uptake across all studied organs with exception of kidneys, the tumor uptake 30 min p.i. was 2.14% ID/g; compared with the work of Mozaffari et al., we obtained very similar results. However, the tumor uptake in the article was 3.36% ID/g, i.e., higher when compared to our data, in what could be explained by the structural difference of studied compounds [28]. The accumulation of RGD peptides in tumors appears to be quite a fast process. This is indicated by the maximum tumor uptake of 68 Ga-NODAGA-c(RGDyK) 2 in 30 min p.i. Tumor to blood and tumor to muscle ratios are the best ex vivo parameters for prediction of an appropriate tumor to background contrast, which can result in suitable images. In this respect, 68 Ga-NODAGA-c(RGDyK) showed the best tumor to blood ratio of 23.1 in 90 min p.i. All three tested 68 Ga-RGD peptides revealed excellent tumor to blood ratios (14.7-23.1), in contrast to 68 Ga-DOTA-substance P (2.8). The best results for tumor to muscle (background) ratio was observed for 68 Ga-DOTA-c(RGDfK). 18 F-FDG biodistribution displayed typical high uptake of the tracer in the heart (almost 42% ID/g) and also in the spleen, what is in line with published data [29]. Finally, 18 F-FLT exhibited relatively uniform accumulation across all studied organs (4-6% ID/g). Nevertheless, this was expected according to the literature [30]. The decrease of tracer uptake between 30 and 90 min post injection was apparent, but it was not the case for spleen, intestine, bone and tumor, On the contrary to 68 Ga-tracers, 18 F-tracers had tumor to blood ratio of 2.0. Tumor to kidney ratio was low for all studied 68 Ga-labeled compounds and all time points, where the lowest values were for 68 Ga-DOTA-substance P. 18 F-tracers had lower tumor to muscle ratios and higher tumor to kidney ratio in comparison to 68 Ga-tracers. These results are fully in accordance with the main excretion route of all 68 Ga-peptides.
In accordance with our ex vivo data, PET/CT imaging with 68 Ga-tracers showed kidneys and bladder as the organs with the highest accumulation of radioactivity. However, the tumors were clearly visualized using the three studied 68 Ga-RGD peptides. The best tumor images were obtained with 68 Ga-NODAGA-c(RGDyK) 2 , whereas the other two 68 Ga-RGD peptides displayed lower radioactivity signal in the tumor. These results are in concordance with PET RGD imaging data published by Li et al. and Provost et al. [11,23]. We did not observe any substantial signal in tumors with 68 Ga-substance P. 18 F-FDG PET scan was also in the line with our ex vivo data and previously published imaging data, showing predominant accumulation of radioactivity in the heart, spleen, tumor and interscapular brown adipose tissue [23]. Similarly, data from 18 F-FLT PET imaging were in congruence with ex vivo results with the highest activity seen in the spleen, gastrointestinal tract, tumor, and largest bone tissue sites, such as knees, what confirms the findings published by Chan et al. [30]. The data gained from PET acquisitions were also used for the uptake quantification of the tracers in tumor and brain regions by calculation of standardized uptake values (SUV). Using the SUVs, we have calculated tumor to brain ratios in order to reveal potential tumor to background ratio for the situation of orthotopic glioblastoma. Interestingly, all the ratios ranged between 4 and 5 with the exception of 18 F-FDG (0.64). These results suggest that 18 F-FDG is not an appropriate tracer for intracranial tumor imaging because of the high background of the tissue. Surprisingly, 68 Ga-DOTA-substance P had comparable tumor to brain ratio as 68 Ga-RGD peptides and 18 F-FLT; however, the SUVs were very low in both tumor and brain regions in the case of 68 Ga-substance P. Among 68 Ga-RGD peptides and 18 F-FLT, all the tumor to brain ratios were very similar not preferencing any of the studied tracers. Nevertheless, the highest value (4.70) of tumor to brain ratio was registered for 68 Ga-DOTA-c(RGDfK). In summary, 68 Ga-RGD peptides showed much favorable kinetics and biodistribution compared to 18 F-FDG and 18 F-FLT. Despite relatively low tumor to blood ratio, tumors could be imaged on PET scans with 18 F-tracers presumably due to the high tumor uptake.
The accumulation of radiolabeled RGD peptides as ligands of α v β 3 integrins reflect the expression rate of these receptors and the tracer could monitor tumor progression and its metastases [31]. The integrin receptors are saturable as was shown in our in vitro cell study. Therefore, it is important to inject the lowest possible dose (low molar amount of RGD peptide) to avoid saturation of target receptors with the cold fraction of the peptide and consecutive decrease in the detected signal in tumor tissues. In contrast to 68 Ga-RGD peptides, 18 F-tracers do not image the rate of integrin expression but only the metabolic rate of compounds (glucose or thymidine). The metabolic rate is also a very useful characteristic of a tumor and could be used to predict various biological characteristics [32].
In conclusion, our study confirmed α v β 3 integrins as a good target for GBM imaging using 68 Ga-RGD peptides that have a favorably biodistribution profiles and a rapid pharmacokinetics.
RGD peptides are very versatile and are not exclusive for GBM as they can image various other malignancies, such as breast cancer, rectal cancer, NSCLC. [33]. The possibility of radiolabeling 68 Ga-labeled RGD peptides in laboratories without the need an on-site cyclotron facility give them an undisputed advantage over 18 F-tracers.

Chemicals
All reagents were purchased from commercial sources as analytical grade and used without further purification. Two clinically used radiopharmaceuticals were employed in this study, 18 F-fluorotymidine (RadioMedic, Rez, Czech Republic) and 18 F-fluorodeoxyglucose (UJV, Rez, Czech Republic), for comparison with non-RGD peptide DOTA-substance P and RGD peptides coupled with chelators, namely DOTA-c(RGDfK), NODAGA-c(RGDyK) and NODAGA-c(RGDyK) 2 . and. All "cold" peptides were purchased from ABX-Advanced biochemical compounds (Radeberg, Germany) with >95% purity. The structures of these four peptides are shown in Figure 5. 68 GaCl 3 was eluted from a 68 Ge/ 68 Ga generator (Eckert & Ziegler Eurotope GmbH, Berlin, Germany) with 0.1 N HCl using a fractionated elution approach.

Radiolabeling and in Vitro Characterization of Studied Peptides
Gallium-68 was obtained from 68 Ge/ 68 Ga generator (Eckert & Ziegler, Berlin, Germany) by elution with 0.1 M HCl (Sigma Aldrich, St. Louis, MO, USA). The optimal 68 Ga-labeling conditions for all peptides were evaluated as follows: 5 µg of peptide (1 µg/µL) dissolved in LC-MS ultrapure water (Sigma Aldrich, St. Louis, MO, USA) mixed with 30 µL of 1.14 M sodium acetate trihydrate (Merck, Darmstadt, Germany) and with 300 µL of 68 Ga eluate. The mixture was incubated for 10 min at RT for RGD peptides with NODAGA and at 90 • C for DOTA modified tracer. For substance P, the labeling conditions were slightly different: 10 µg of substance P, temperature of 95 • C, incubation time of 15 min. The radiochemical purity of 68 Ga-peptides was determined by HPLC analysis employing gradient elution (water to acetonitrile). The HPLC method was described in detail previously [12]. for HPLC analysis to determine the radiochemical purity were taken at selected time points (30,60, and 120 min) and directly injected on HPLC.

Stability in Physiological pH
The peptides were labeled as described above. Subsequently, 100 µL of 1.14 M sodium acetate was added to adjust the pH to 7.0. The mixture was incubated at 37 • C and the samples for HPLC analysis to determine the radiochemical purity were taken at selected time points (30,60, and 120 min) and directly injected on HPLC.

Plasma Stability
100 µL of the labeled peptide was mixed with 400 µL of human plasma (Sigma Aldrich, St. Louis, MO, USA). This mixture was incubated at 37 • C and the samples for HPLC analysis to determine the radiochemical purity were taken at selected time points (30,60, and 120 min). The individual samples were adjusted before the injection on HPLC-every sample was mixed with acetonitrile (Roth, Karlsruhe, Germany) and placed on the shaker (450 rpm) for 60 s and then centrifuged (20,000 RCF, 3 min), followed by removing the supernatant. Finally, the supernatant was analyzed using HPLC.

Determination of Partition Coefficient (Log P)
Labeled peptide was diluted with 600 µL of phosphate-buffered saline (PBS), and 50 µL of the solution was mixed in the test tube with 450 µL of PBS and 500 µL of octanol. The sample was placed on the shaker (1000 RPM) for 20 min and then centrifuged at 15,000 RPM for 1 min. Subsequently, 50 µL from both octanol and water layer were collected and measured on an automatic gamma counter (PerkinElmer, Waltham, Massachusetts, USA). Partition coefficient was calculated as the decadic logarithm of the ratio of octanol/water activity.

Plasma Protein Binding
10 µL of the labeled peptide was mixed with 190 µL of human plasma (Sigma Aldrich, St. Louis, MO, USA), and incubated at 37 • C. Twenty-five µL of sample were collected at 30, 60, and 120 min time intervals for the determination of protein binding. The samples were separated by size-exclusion chromatography (illustra MicroSpin G-50 Column, GE Healthcare, UK) by centrifugation (2000 × g, 1 min). The eluate (protein-bound) and appropriate column (non-protein-bound) were measured on an automatic gamma counter. The percentage of plasma protein binding (PPB) was calculated as the ratio of the activity in the eluate and the activity of the whole sample (column plus the eluate). The blank determination of protein binding, i.e., incubation of the radiotracer in PBS buffer followed by separation of the sample on gel column was performed as the control step.

Determination of Metabolic Stability in Vivo
The experimental animals were intravenously injected with tested radiolabeled peptides (5 MBq corresponding to 1 µg of the peptide per animal) and urine was collected 30 and 90 min post-injection. The urine samples were directly analyzed on HPLC using the same method as for other stability assays. The obtained radiochromatograms were compared with the radiochromatograms of studied 68 Ga-peptides. The goal of this experiment was to determine whether the peptides are metabolized in vivo or are excreted in unchanged form. The cells were incubated with 68 Ga-labeled RGD peptides (20 µM) alone and also with excess (2 mM) of cold RGD peptide for 90 min (37 • C) in the buffer described above. The cells were rinsed with PBS, lysed in 0.1 M NaOH and measured for radioactivity in a gamma counter (PerkinElmer, Waltham, Massachusetts, USA). The uptake was calculated as the mean of the percentage of the applied dose ± standard deviation.

Ex Vivo Biodistribution
For ex vivo biodistribution studies, radiolabeled tracers were diluted with saline. The prepared solutions were applied retro-orbitally (r.o.) to the mice [35] at a dose of 1-2 MBq per mouse corresponding to 0.2-0.4 µg of the peptide. The mice were sacrificed by cervical dislocation 30 min and 90 min post-injection and organs of interest (blood, spleen, pancreas, stomach, intestine, kidneys, liver, heart, lungs, muscle, bone, and tumor) were collected. The samples were weighed and the radioactivity was measured by an automatic gamma counter. The accumulation of radiotracer in studied organs was expressed as a percentage of injected dose per gram of organ.

µPET/CT Imaging
Experimental animals were imaged using microPET/SPECT/CT system Albira (Bruker, Billerica, Massachusetts, USA). Mice were r.o. injected with the radiolabeled tracer at a dose of 5-10 MBq (corresponding to 1-2 µg of peptide for animals injected with 68 Ga-peptides) per animal. Anesthetized animals were placed in a prone position in the Albira system before the start of imaging. Static PET/CT images were acquired over 30 min interval starting 30 and 90 minutes after injection. A 10-min PET scan (axial FOV 148 mm) was performed, followed by a double CT scan (axial FOV 65 mm, 45 kVp, 400 µA, at 400 projections). Scans were reconstructed with the Albira software (Bruker Biospin Corporation, Woodbridge, CT, USA) using the maximum likelihood expectation maximization (MLEM) and filtered back-projection (FBP) algorithms. The PMOD software (PMOD Technologies, Zurich, Switzerland) was used for the image analysis and tracer uptake quantification. Three-dimensional volume rendered images were created in VolView software (Kitware, Clifton Park, NY, USA). The data obtained by PET scan were quantified in PMOD via manually selected volumes of interest (VOIs) and expressed as standardized uptake values (SUV). Using SUVs for the tumor and brain regions the tumor to brain ratios were calculated.

Conclusions
Studied 68 Ga-RGD peptides displayed excellent in vitro characteristics as well as in vivo behavior; on the contrary, 68 Ga-DOTA-substance P failed in the in vivo stability test and was not able to image the tumor. In vivo comparison of 68 Ga-RGD peptides with 18 F-FDG and 18 F-FLT in glioblastoma tumor model in mice revealed better imaging properties for 68 Ga-RGD peptides. 68 Ga-RGD peptides showed higher tumor to background ratio and a low accumulation in non-target organs except the excretory organs. These properties make 68 Ga-RGD peptides, in particular, 68 Ga-NODAGA-c(RGDyK) 2 , more suitable for preclinical imaging of GBM than clinically used 18 F-tracers.