18F-Fluorothymidine-Pet Imaging of Glioblastoma Multiforme: Effects of Radiation Therapy on Radiotracer Uptake and Molecular Biomarker Patterns

Introduction. PET imaging is a useful clinical tool for studying tumor progression and treatment effects. Conventional 18F-FDG-PET imaging is of limited usefulness for imaging Glioblastoma Multiforme (GBM) due to high levels of glucose uptake by normal brain and the resultant signal-to-noise intensity. 18F-Fluorothymidine (FLT) in contrast has shown promise for imaging GBM, as thymidine is taken up preferentially by proliferating cells. These studies were undertaken to investigate the effectiveness of 18F-FLT-PET in a GBM mouse model, especially after radiation therapy (RT), and its correlation with useful biomarkers, including proliferation and DNA damage. Methods. Nude/athymic mice with human GBM orthografts were assessed by microPET imaging with 18F-FDG and 18F-FLT. Patterns of tumor PET imaging were then compared to immunohistochemistry and immunofluorescence for markers of proliferation (Ki-67), DNA damage and repair (γH2AX), hypoxia (HIF-1α), and angiogenesis (VEGF). Results. We confirmed that 18F-FLT-PET uptake is limited in healthy mice but enhanced in the intracranial tumors. Our data further demonstrate that 18F-FLT-PET imaging usefully reflects the inhibition of tumor by RT and correlates with changes in biomarker expression. Conclusions. 18F-FLT-PET imaging is a promising tumor imaging modality for GBM, including assessing RT effects and biologically relevant biomarkers.


Introduction
Glioblastoma Multiforme (GBM) is a highly aggressive cancer that experiences a high rate of recurrence despite combination therapy [1]. There is significant need for tumor imaging modalities or biomarkers capable of identifying early stage tumors, posttreatment changes, and signs of tumor progression and recurrence of GBM [2]. Positron emission tomography (PET) is an imaging modality widely used in oncology for clinical staging, monitoring of treatment efficacy, and followup for disease recurrence [3,4]. 18 F-Fluorodeoxyglucose ( 18 F-FDG) PET, which measures cellular glucose metabolism as a function of the enzyme hexokinase, is the most common clinically utilized PET modality. However, for imaging of high-grade brain tumors like GBM, 18 F-FDG-PET imaging has limited utility. The high baseline level of glucose uptake within the brain results in a low signal-tonoise ratio and obscured tumor imaging [5,6]. Consequently, several recent preclinical and clinical studies have investigated the use of alternative PET radiotracers, including 18 F-Fluorothymidine ( 18 F-FLT) and 11 C-Methionine ( 11 C-MET), for imaging of GBM [6][7][8]. 18 F-FLT, a nucleoside radiotracer, has emerged as a safe and promising tool to be used in conjunction with 18 F-FDG-PET for studying aggressive and proliferative tumors like GBM [9,10]. Thymidine uptake and metabolism require 2 The Scientific World Journal Thymidine Kinase-1 (TK-1) activity, which is elevated in the normal cellular S-phase and neoplastic tissues [11]. Additionally, unlike carbon-based thymidine, 18 F-FLT has demonstrated higher specificity for TK-1 and reduced incorporation in DNA [12]. 18 F-FLT-PET imaging has been shown in several clinical studies to correlate with cellular proliferation and tumor progression in high-grade glioma, as determined by the Ki-67 proliferation index [13][14][15]. Furthermore, Chen et al. demonstrated that increased radiotracer uptake in highgrade glioma was associated with reduced patient survival [13]. 18 F-FLT-PET imaging has also been shown to correlate with treatment response in clinical studies of breast cancer and high-grade glioma [8,16]. Similar studies have also been performed in other cancer types including lymphoma and soft tissue sarcoma [17][18][19]. We wished to study the effects of radiation on 18 F-FLT-PET imaging of GBM tumors and secondarily effects on markers relevant to hypoxia, angiogenesis, and DNA damage and repair. Such data could provide additional critical information about treatment response and disease progression.  [20]. After the procedure, the mice were returned to their cages and observed until the anesthesia had cleared and subsequently monitored on a daily basis for health, normal behavior, and visual tumor growth.

Orthotopic Intracranial Tumors.
Six-week-old nude/ athymic mice were injected intracranially with 200,000 U251-GFP-LUC cells. First, mice were anesthetized by IP injection with ketamine (120 mg/kg) and xylazine (10 mg/kg). Additionally, mice received meloxicam (5 mg/kg) through IP injection in the preoperative and postoperative period. Once confirmed to be unconscious by toe-pinch, each mouse individually was placed in a stereotactic head frame (Stoelting, Wood Dale, IL, USA). Once placed in the stereotactic frame, the operative area was decontaminated using betadine solution with a cotton tip applicator. All surgical procedures were performed using autoclaved or sterile instruments in a clean environment. First, a 1 cm midline scalp incision was made and the bregma was visualized. Next, a small burr hole was made through the skull using a Dremel drill 0.5 mm anterior and 2 mm lateral to the bregma as performed by Baumann et al. [21]. 200,000 cells in 10 L of 1xPBS were drawn into a Hamilton syringe. Then, the syringe was inserted into the arm of the stereotactic system with the 30G needle placed directly over the newly created burr hole. The syringe was inserted through the burr hole 3.3 mm from the skull surface. Cells were injected at a low rate (approximately 0.5 L/min) in order to minimize the pressure created at the injection site. Upon completion of the injection, the needle was removed, the burr hole was covered with bone wax, and the incision was closed using cyanoacrylate. After operation, the mice were returned to their cages and observed until the anesthesia had cleared and subsequently monitored on a daily basis for health, normal behavior, and tumor growth.

Bioluminescence Imaging (BLI).
The mice with tumors underwent weekly BLI to observe tumor growth and quantify tumor size [22]. BLI was conducted starting 1 week after tumor injection and continued on a weekly basis in a centrally shared IVIS Lumina II In Vivo Imaging System (Caliper Life Sciences; Hopkinton, MA, USA) according to the UPENN SAIF guidelines. First, the mice were anesthetized as described above (see Section 2.3.1) and confirmed by negative toe-pinch test. This was followed by dorsal, subcutaneous injection of 60 L of USP grade Dluciferin (D-Luciferin potassium salt; Gold Biotechnology; St. Louis, MO, USA). Imaging was conducted 5 min after D-luciferin injection over a period of 30 min to effectively determine the maximum luminescence intensity. Region of Interest (ROI)determination and analysis were conducted with the Living Image v4.0 software (Caliper Life Sciences; Hopkinton, MA, USA). Software-generated surface radiance measurements (photons/sec/cm 2 /steradian) were used to calculate Flux max values (photons/second) in order to determine tumor size. Tumors were allowed to grow without intervention until Flux max values approached 1×10 9 (1 +10).

Radiation Therapy.
Mice tumors with Flux max values on the order of 1 × 10 10 photons/sec (∼4-6 weeks after injection) were anesthetized via IP injection with ketamine (140 mg/kg) and xylazine (10 mg/kg). They then were individually irradiated with single-dose unilateral 16 Gray (Gy) radiation to the right side tumor in a departmentowned unidirectional anterior-to-posterior beam X-Ray unit (Schneeman Electronics (Grants Pass, OR), model A-9002-100) with full body lead shielding except at the implanted tumor site and were then sacrificed 4-6 weeks later.

2.4.3.
MicroPET/CT Imaging. Experimental and control mice were subsequently followed by serial 18 F-FLT-PET imaging. Both 18 F-FDG and 18 F-FLT microPET and microCT imaging were conducted at the UPENN SAIF laboratory. For flank tumors, 18 F-FLT-PET microPET/CT was conducted prior to radiation therapy (RT) and 5 and 13 days (∼1 week and 2 weeks) after RT. Mice were maintained on a 1-3% isoflurane and oxygen anesthesia system (VetEquip Inc; Pleasanton, CA) flowing at a rate of 1.0 liter/min. Radiotracer was delivered via a lateral tail vein injection with a maximum of 200 Ci-500 Ci and a 60 min uptake time. Animal temperature was maintained using a warm air source and subsequently monitored (Vetronics; West Lafayette, IN, USA). MicroPET imaging was performed on a Philips Medical System (Cleveland, OH, USA) small animal imaging machine (A-PET). Small animal microCT was subsequently performed on an ImTek scanner (CTI Molecular Imaging Inc. (now Siemens; Malvern, PA)) with a spatial resolution <50 mm. After imaging, mice were housed in a SAIFdesignated area for handling radiotracers for 24 hours and monitored for radiotracer clearance before returning to the normal housing facility. MicroPET/CT images were retrieved in Analyze (.hdr) format. Radiotracer uptake was quantified by calculation of the percent of injected dose (%ID) of radiotracer located within an ellipsoid ROI. Total injected

Intracranial MicroPET/CT Imaging.
To compare the patterns of uptake of 18 F-FDG or 18 F-FLT in healthy nude/ athymic mice, the mice underwent microPET imaging 60 min following tail vein injection with either radiotracer. Intracranial radiotracer uptake was quantified based on % injected dose (%ID). 18 F-FDG uptake into the normal brain was >4x higher than 18 F-FLT uptake (%ID 1.43 versus 0.33, respectively) (Figure 1(a)). 18 F-FDG uptake was uniform throughout the intracranial space when viewed in the coronal, axial, and sagittal planes, whereas 18 F-FLT uptake was visually undetectable (Figure 1(b)). These results indicate that there is a high background signal with 18 F-FDG that is not present for 18 F-FLT. Concentration of both radiotracers within the bladder was noted to be comparable (data not shown). Such imaging validates that 18 F-FLT uptake within the healthy brain is minimal compared to conventional 18 F-FDG and suggests that intracranial 18 F-FLT uptake would increase with tumor presence. We thus compared uptake patterns of 18 F-FDG and 18 F-FLT in intracranial tumors. Healthy nude/athymic mice ( = 5) were intracranially injected with U251-GFP-LUC glioma cells and underwent weekly BLI with subsequent 18 F-FLT-PET imaging (Figure 2(a)). Intracranial tumor growth across all mice demonstrated an exponential growth pattern as measured by changes in mean BLI Flux max, with increased proliferation beyond 4 weeks (Figure 2(b)). 18 F-FLT-PET imaging was performed in select mice ( = 2), and intracranial 18 F-FLT uptake was quantified by %ID at weeks 1 and 7, which correlated with a BLI Flux max, of 1 × 10 6 and 1 × 10 9 , respectively. Radiotracer uptake was ∼4.7x greater at 7 weeks versus 1 week (%ID 1.58 versus 0.33, resp.) (Figure 2(c)). Visually, radiotracer uptake was most appreciated in the larger intracranial tumors (7 weeks after The Scientific World Journal injection). Additionally, radiotracer uptake at 7 weeks was localized to the tumor site, with minimal uptake in the surrounding normal brain tissue (Figure 2(d)). These data demonstrate that 18 F-FLT concentrates preferentially within tumor and suggest that radiotracer uptake correlates with tumor size.

Flank Tumor
MicroPET/CT Imaging. We wished to investigate the effect of radiation therapy (RT) on radiotracer uptake. However, the limited size of intracranial tumors in the mice did not permit sufficient imaging resolution. Additionally, mice with intracranial tumors with Flux max values greater than 1 × 10 9 were quickly terminal (data not shown) and therefore unlikely to survive sufficiently long after irradiation for extended followup. Therefore, we resorted to flank xenografts for subsequent studies. Experimental mice ( = 3) with bilateral U251-GFP-LUC flank tumors were exposed to a single unilateral fraction of 16 Gy RT to the right side flank tumor and tumor growth of the irradiated and control nonirradiated tumors was subsequently followed serially with BLI (Figure 3(a)). Flux max in control tumors progressively increased in the weeks following radiation, while RT tumors demonstrated reduction in BLI starting after 1 week (Figure 3(b)).
In addition to BLI, 18 F-FLT-PET imaging of select experimental mice ( = 2) was performed before RT and 1 and 2 weeks following single dose 16 Gy RT of the right side flank tumor. Decreases in 18 F-FLT radiotracer uptake in RT tumors were noted by 1 week after RT and sustained until 2 weeks after RT. Decreases in %ID in RT tumors at 2 weeks ranged from 23% to 64%. (Supplemental Figure 1, see Supplementary Material available online at http://dx.doi.org/10.1155/2013/796029) When comparing changes in %ID in RT versus control tumors, RT tumors exhibited significant reduction in radiotracer uptake by 2 weeks (Figure 3(c)).
Visually, 18 F-FLT-PET imaging on both coronal and axial imaging suggests equal radiotracer enhancement between control and treatment tumors prior to RT (Figure 4(a)). Uptake patterns following radiotherapy, however, were different between treated and untreated tumors (Figure 4(b)). Left side control tumors were found to exhibit peripheral enhancement with notable central pallor, suggesting hypoxia induced central necrosis with the formation of pseudopalisades, a well-characterized phenomenon often seen in GBM [23,24]. Quite strikingly, irradiated tumors were diminished in size, and the smaller tumors showed uniform radiotracer uptake throughout the tumor mass but with decreased intensity in comparison to control tumors.
For these pilot studies, a single dose RT regimen was utilized to investigate the effects of RT on tumor growth. Such a technique is similar to hypofractionated stereotactic radiosurgery which is used clinically for patients with cancer [25,26]. These results in our mouse model encourage clinical follow-on studies with fractionated RT regimens. for histologic examination. Hematoxylin and Eosin (H&E) staining appeared comparable between control and irradiated tumors ( Figure 5(a)). In contrast, irradiated tumors showed markedly decreased Ki-67 staining ( Figure 5(b)). These results suggest that RT led to decreased tumor proliferation, thus reflected by diminished Ki-67 activity. Similar findings of diminished Ki-67 after radiation have also been noted in the clinic in both astrocytoma and oral squamous cell carcinoma [27,28].

Effects of Radiation Therapy on Cellular
We also sought to evaluate if decreased tumor proliferation secondary to irradiation also corroborated with initiation of cell death. H&E staining was compared to Gamma-H2AX ( H2AX) expression using immunofluorescence (IF). H2AX is a DNA histone that undergoes phosphorylation on the Ser139 residue in the presence of double-stranded DNA (dsDNA) breaks, such as that induced by RT. Increased H2AX signaling is suggestive of successful RT delivery [29][30][31]. Greater attention was given to expression levels within the tumor core, given that GBM is known to express phenotypic and genotypic heterogeneity in tumor markers at or within the tumor core [32,33]. Control tumors demonstrated increased eosin staining within the tumor core (Figure 6(a)). These areas also correspond with reduced H2AX foci (red) and DAPI signaling (blue). Peripheral to the tumor core, multiple H2AX foci with corresponding DAPI staining are noted. Such findings suggest a lack of viable cells within the tumor core, likely due to tumor necrosis. Similarly, such necrotic foci have been demonstrated to be a hallmark of GBM on both pathology and radiology [34].
In irradiated tumors, H&E staining also revealed large, centrally located areas that are composed of a milieu of mixed hematoxylin and eosin localization (Figure 6(b)). In comparison to control tumors, IF studies for H2AX (red) expression in irradiated tumors demonstrated increased biomarker expression with positive DAPI (blue) staining of nuclei in within the central milieu. This is best visualized at higher resolution (20x). Positive H2AX expression and positive DAPI staining in these central areas are suggestive of post-irradiation cell recovery and repair in the setting of radiation induced apoptosis [35][36][37]. 18 F-FLT-PET Imaging, and Markers of Angiogenesis, and Hypoxia. The reduced proliferation of tumor and tumor necrosis may be due at least in part to RT effects on the tumor vasculature. There is evidence that GBM undergoes vascular changes relating to hypoxia and angiogenesis, and these often impact tumor  growth and treatment resistance. Such findings have been visualized clinically with PET imaging [38]. We therefore sought to compare 18 F-FLT-PET imaging to IF staining for markers for angiogenesis and hypoxia. Hypoxia is known to be involved in tumor recurrence, chemo-and radioresistance, invasion, and decreased patient survival [24,39]. We chose to examine expression of vascular endothelial growth factor (VEGF) and hypoxia induction factor-1 (HIF-1 ), as both have been studied extensively and demonstrated to be active in GBM [24,[40][41][42][43]. 18 F-FLT-PET imaging of control tumors 2 weeks after unilateral RT demonstrated a noticeable reduction in radiotracer uptake within the tumor core as shown in both axial and coronal planes, potentially due to vascular insult ( Figure 7(a)). IHC staining patterns for VEGF and HIF-1 (red) correlate with regions of increased eosin staining, as previously shown. Expression of both biomarkers may be mutually related, as increased hypoxia secondary to proliferation can induce VEGF production as a means to stimulate angiogenesis and reduce hypoxia [23,34,44]. IF staining reveals similarly absent DAPI (blue) staining with increased fluorescence for VEGF and HIF-1 . This superimposed central presence of VEGF and HIF-1 in nonirradiated tumor cores supports our prior hypothesis of central necrosis [24].

Impact of Radiation Therapy on
In irradiated tumors, 18 F-FLT radiotracer uptake was also diminished within the tumor core, similar to control tumors (Figure 7(b)). IHC staining for VEGF and HIF-1 (red) was visually identical. IF staining of tumor cores was positive for increased expression of both biomarkers with concomitant positive DAPI (blue) signaling. Similar increases in VEGF expression after RT in GBM have been demonstrated in vitro [45]. Similarly, RT is known to induce hypoxia and to increase HIF-1 expression [46,47]. Additionally, in C6 rat gliomas, HIF-1 upregulation was noted within 48 hours and up to 8 days following 8 Gy RT [48]. These data suggest that increased biomarker expression following high dose 16 Gy RT is secondary to ongoing tumor repair.

Conclusions
18 F-FLT-PET imaging provides an exciting alternative to 18 F-FDG-PET in the study of GBM. In our intracranial mouse model, we demonstrated successful 18 F-FLT imaging of intracranial GBM orthotopic tumors with radiotracer sparing of the intact mouse brain. Additionally, 18 F-FLT uptake within tumors correlated with tumor size.
In heterotopic flank xenografts, 18 F-FLT radiotracer uptake patterns suggest irradiated tumors demonstrate detectable but globally diminished radiotracer uptake as seen by decrease in %ID. This is most likely due to tumor cell death and reduced postirradiation proliferation (Ki-67). Irradiated tumors also showed changes for biomarkers of dsDNA breaks ( H2AX), hypoxia (HIF-1 ), and vascular remodeling (VEGF).
These data suggest that 18 F-FLT radiotracer uptake varies due to changes in the GBM tumor microenvironment, and that PET imaging may readily identify these changes. Such data encourage further in vivo and clinical studies with novel PET radiotracers such as 18 F-FLT, with hopes of ultimately improving the imaging and care of patients with GBM.