PET imaging of medulloblastoma with an 18F-labeled tryptophan analogue in a transgenic mouse model

In vivo positron emission tomography (PET) imaging is a key modality to evaluate disease status of brain tumors. In recent years, tremendous efforts have been made in developing PET imaging methods for pediatric brain tumors. Carbon-11 labelled tryptophan derivatives are feasible as PET imaging probes in brain tumor patients with activation of the kynurenine pathway, but the short half-life of carbon-11 limits its application. Using a transgenic mouse model for the sonic hedgehog (Shh) subgroup of medulloblastoma, here we evaluated the potential of the newly developed 1-(2-[18F]fluoroethyl)-L-tryptophan (1-L-[18F]FETrp) as a PET imaging probe for this common malignant pediatric brain tumor. 1-L-[18F]FETrp was synthesized on a PETCHEM automatic synthesizer with good chemical and radiochemical purities and enantiomeric excess values. Imaging was performed in tumor-bearing Smo/Smo medulloblastoma mice with constitutive actvation of the Smoothened (Smo) receptor using a PerkinElmer G4 PET-X-Ray scanner. Medulloblastoma showed significant and specific accumulation of 1-L-[18F]FETrp. 1-L-[18F]FETrp also showed significantly higher tumor uptake than its D-enantiomer, 1-D-[18F]FETrp. The uptake of 1-L-[18F]FETrp in the normal brain tissue was low, suggesting that 1-L-[18F]FETrp may prove a valuable PET imaging probe for the Shh subgroup of medulloblastoma and possibly other pediatric and adult brain tumors.

Multimodal brain imaging integrating MRI, computer tomography (CT) and positron emission tomography (PET), is a pivotal aid in the diagnosis of brain tumors 6,7 . Complementing each other, MRI and CT scans provide structural information, while PET allows for an important quantitative and highly sensitive metabolic assessment enabling early diagnosis of tumors and prediction of response to therapy 8 . In the clinic, [ 18 F]fluorodeoxyglucose (FDG) PET has become a routine procedure for evaluating the glucose metabolism of tumors. However, the application of FDG in the diagnosis of brain tumors, and especially of low grade tumors, is limited due to the high uptake of FDG in normal grey matter. As such, a number of amino acid-based radiotracers have been developed and used for the evaluation of amino acid-involved biological process in tumors 8 . Among them, tryptophan-based PET tracers show great promise for brain tumor imaging [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25] .
While PET imaging with 1-L-[ 18 F]FETrp shows promise for detecting brain tumors 32 , previous studies have been limited to mouse xenograft models using established cell lines and primary tumor cells from adult patients, not taking into account limitations imposed by the blood brain barrier (BBB). In medulloblastoma, while changes in the BBB in the WNT form allow passage of chemotherapeutic agents into the brain and have high cure rates, the BBB is intact in the chemoresistant Shh form 37 . In the present study, we evaluated 1-L-[ 18 F]FETrp for PET imaging of the most common malignant pediatric brain cancer medulloblastoma in an immunecompetent, transgenic mouse model for the Shh subgroup in which tumors occur spontaneously 38 Fig. 2A,B. The tracer highly accumulated in the cerebellar area of the tumor-bearing mouse ( Fig. 2A), while in the cerebellar area of the control mouse the uptake was negligible (Fig. 2B). Mouse brains processed for histological hematoxylin and eosin (H&E) staining after PET imaging confirmed the presence of cerebellar tumors in the symptomatic mice (Fig. 2C), suggesting that 1-L-[ 18 F]FETrp could serve as a PET imaging probe to detect medulloblastoma.

1-L-[ 18 F]FETrp uptake kinetics in Smo/Smo medulloblastoma tumors.
To further investigate tracer kinetics, dynamic small animal PET studies were conducted in tumor-bearing mice (n = 3). As shown in Fig. 3A  www.nature.com/scientificreports www.nature.com/scientificreports/ post injection (2.26 ± 0.34) and plateaued until around 120 min post injection. Thereafter, the tracer washed out slowly from the brain tumor (SUV at 135 min: 2.14 ± 0.47; SUV at 155 min: 2.11 ± 0.41; SUV at 175 min: 2.04 ± 0.44). The activities in tumor-distant contralateral cerebellum and the cerebrum of the same tumor mice maintained a low level (SUV <0.77) over 180 min (n = 3, p < 0.05). Based on these time-activity curves, the image contrast-time (tumor-to-contralateral cerebellum ratio) profile was generated (Fig. 3C). The target-to-nontarget (tumor-to-normal cerebellum) ratio increased over time and peaked at between 40 and 90 min (standard uptake  www.nature.com/scientificreports www.nature.com/scientificreports/ value ratio (SUVR) = 3.5-3.7) resulting in a favorable contrast within this time frame. In addition, as shown in Fig. 3B, the tracer accumulated in the kidneys and bladder over time, suggesting renal excretion of this tracer. It should be noted that the skull and bone uptake maintained a low level over time (Fig. 3B), indicating that the tracer was stable with regard to in vivo defluorination.

Specificity of the 1-L-[ 18 F]FETrp uptake in medulloblastoma tumors.
It is common that the L-enantiomer of an amino acid PET tracer has higher tumor uptake than its D-counterpart 39,40 . Consistent with this, in breast cancer, it has previously been shown that L-and D-enantiomers of 1-[ 18 F]FETrp have different uptake profiles 33 . In these studies, the in vitro cell uptake in MDA-MB-231 cells and the in vivo tumor uptake in MDA-MB-231 mouse xenografts suggested that the uptake of 1-L-[ 18 F]FETrp is much higher than the D-counterpart. To confirm whether the enantiomers also displayed different imaging properties in medulloblastoma, 60 min dynamic PET scans were performed with 1-D-[ 18 F]FETrp in tumor-bearing Smo/Smo mice. As expected, the uptake of the D-enantiomer in nonlesioned cerebellum and cerebrum was very low in the tumor mice ( Fig. 4) with an SUV <0.5 during the entire 60 min scan. In addition, the tracer uptake in tumors was also low albeit slightly higher than that in the normal brain regions (contralateral cerebellum and cerebrum). These results suggest that like breast cancer xenografts, medulloblastoma tumors preferably take up the 1-L-[ 18

Discussion
For pediatric brain cancers, non-invasive PET imaging methods that can detect tumors with high sensitivity and specificity have lagged behind adult applications. In this study, we provide evidence that the tryptophan amino acid derivative 1-L-[ 18 F]FETrp may prove suitable as a PET imaging probe for the most common malignant pediatric brain cancer medulloblastoma. Using the Smo/Smo transgenic medulloblastoma mouse model, we show that 1-L-[ 18 F]FETrp can delineate cerebellar tumors with high contrast and specificity. 1-L-[ 18 F]FETrp accumulation likely occurred due to active uptake by the tumor rather than simple leakiness of the BBB as the uptake of the D-enantiomer was highly reduced compared to the 1-L-[ 18 F]FETrp counterpart. Further, the negligible skull uptake of 1-L-[ 18 F]FETrp indicated its good in vivo stability towards defluorination, a desirable feature for brain tumor imaging. Defluorination would result in activity accumulation in bone, spreading radioactivity from the skull to nearby brain tissue through a partial volume effect and confounding imaging quantification. Thus, 1-L-[ 18 F]FETrp may be particularly well-suited for imaging tumors in the brain.
Previous studies in breast and brain tumor xenografts demonstrated that 1-L-[ 18 F]FETrp can detect abnormal tryptophan metabolism in these tumors 32,33 . However, by using xenografts, these studies did not evaluate the ability of 1-L-[ 18 F]FETrp to cross the BBB and detect tumors directly in the brain, a prerequisite for brain tumor imaging in the clinic. With the use of the transgenic Smo/Smo mice in which medulloblastoma tumors occur spontaneously, we were able to use a relevant preclinical mouse model to demonstrate the ability of 1-L-[ 18 F]FETrp to directly detect brain tumors with a high contrast. In addition, this well-established animal model enabled us to circumvent additional limitations of patient-derived xenograft (PDX) models, in which the interpretation of the imaging data may be compromised due to the use of severely immunocompromised mice. Furthermore, the inevitable local brain injury in orthotopic xenografts occurs during the injection of the tumor cells into the brain, which may cause neuroinflammation. This is particularly important in the case of imaging probes such as the 1-L-[ 18 F]FETrp, as abnormal tryptophan metabolism in tumors has been linked to the kynurenine inflammatory pathway 28,41,42 . Based on our studies in an immunocompetent, genetically engineered preclinical medulloblastoma model with spontaneous tumor occurrence and intact blood-brain barrier, we now suggest that 1-L-[ 18 F]FETrp, in combination with other commonly used imaging modalities, may have clinical relevance for the diagnosis, metabolic characterization and evaluation of response to treatment of medulloblastoma.
With regard to tumor uptake, L-amino acid radiotracers in general show remarkable advantages over their corresponding D-isomer due to the preference of amino acid transporters and related enzymes for the L-isomer. As such, most amino acid PET radiotracers currently used in the clinic are L-isomers 39,40 . However, at times, D-amino acid PET radiotracers display a similar tumor uptake as their L-counterparts and in some cases, have superior contrast performance to L-isomers due to their fast clearance from the blood pool [43][44][45] . Furthermore, the tryptophan analog, 1-D-methyl-tryptophan (1-D-MT) has proven to be superior to its L-counterpart and has been widely used in the clinic. Nevertheless, similar to previous findings in a breast cancer model 33,46 , our studies in Smo/Smo mice showed that the brain uptake of 1-D-[ 18 F]FETrp was very low as compared to that of its L-counterpart, possibly due to preference differences in the amino acid transporter. However, it should be noted that although the tumor uptake of 1-D-[ 18 F]FETrp in Smo/Smo mice was low, it was still slightly higher than that in the normal brain regions (contralateral cerebellum and cerebrum). This could be due to increased vascularization of the tumor tissue or to a generally increased tryptophan metabolism in tumor cells as both Dand L-tryptophan can be catabolized by the KP rate-limiting enzyme IDO 47 . In medulloblastoma, IDO1 and/or TDO2 mRNA expression was reported to be increased in human tumor samples across all four subgroups 31 , but additional pathways known to affect tryptophan metabolism may induce 1-L-[ 18 F]FETrp uptake in medulloblastoma as well. This includes mechanistic target of rapamycin(mTOR) signaling, a metabolic pathway that has been shown to crosstalk with the IDO1 pathway in medulloblastoma 48 . mTOR promotes medulloblastoma tumor progression 49,50 and targeting mTOR may be a promising strategy to treat medulloblastoma 51  www.nature.com/scientificreports www.nature.com/scientificreports/ final product of 1-L-[ 18 F]FETrp for animal studies. 1-D-[ 18 F]FETrp was synthesized using the same procedure but starting from the D-precursor and was purified by the same chiral HPLC column. The quality control of the synthesized radiotracer was performed on an Agilent Infinity 1260 HPLC system equipped with an Astec CHIROBIOTIC ™ T Chiral column (5 μm, 250 × 4.6 mm), with a mobile phase of ethanol in water (7/3, v/v) at a flow rate of 1 mL/min through a UV detector (λ = 230 nm) followed by a radioactivity detector. The decay corrected radiochemical yields were in the range of 20-30% (n = 6). The molar activity was in the range of 15-170 GBq/µmol. The total synthesis time was about 75 min. Both the radiochemical purity and enantiomeric excess of 1-L-[ 18 F]FETrp were more than 95%. The same procedure was used for the synthesis of 1-D-[ 18 F]FETrp and the tracer was obtained with good radiochemical purity (>95%) and optical purity (>90%). The detailed radiolabeling and chromatographic information are included in the Supporting Information (Supplementary Figs. S1-4).

Micro-PET imaging.
For all animal studies, applicable institutional and/or national guidelines for the care and use of animals were followed and the study was approved by the Nemours Institutional Animal Care and Use Committee (IACUC). Smo/Smo mice developed in the laboratory of Dr. James Olson (Fred Hutchinson Cancer Research Center) were described previously 38  Micro-PET imaging was performed in Smo/Smo mice symptomatic for brain tumors (n = 9). Symptoms of brain tumor development included head tilt, decrease in activity, ataxia and bulging posterior skull. Asymptomatic littermates that had not developed brain tumors were used as control (n = 5). Imaging was performed with a PerkinElmer G4 PET-X-Ray scanner. Five min prior to imaging, the animals were anesthetized using 2-3% isoflurane in oxygen and then the mice were injected with 1.5-2.2 MBq of 1-L-[ 18 F]FETrp or 1-D-[ 18 F]FETrp in saline (total volume less than 100 µL) via the tail vein. The mice were placed onto the temperature-controlled imaging bed under isoflurane anesthesia for the duration of the imaging. Dynamic imaging was performed for one hour or three hours immediately after injection.
Image reconstruction and data analysis. Images were reconstructed on the scanner with default parameters. The duration of each frame was ten min. For the PET quantification, regions of interest (ROIs) within the tumor, the cerebellum, and the cerebrum were drawn manually on the frame 50-60 min post injection and copied to all other frames. The ROI depicting the tumor covered the tumor region with high uptake over multiple slices, while the ROI of the cerebellum was drawn on the contralateral side with a similar volume. The ROI of the cerebrum was drawn in the middle of the cerebrum with a volume of 3.65 × 3.65 × 2.28 mm 3 . ROIs were drawn in ITK-SNAP (version 3.2). Data analyses were performed in MATLAB. The resulting quantitative data were expressed in SUV, which were calculated as the ratio of regional averaged radioactivity in Becquerel per cubic centimeter to injected radioactivity in Becquerel per gram of body weight.
Histology. After imaging, mice were euthanized, and brains were dissected, fixed with formalin, and embedded in paraffin using standard procedures. Entire brains were sectioned, and one of every ten sections from 5 µm serial sections was deparaffinized, rehydrated, and stained with H&E using standard protocols. All sections were analyzed for the presence of tumor and images were taken with a Nikon Eclipse 80i microscope.
Statistical analysis. Data are presented as mean ± standard deviation (SD). A two-tailed paired t test was used for the comparison of the tracer uptake in the brain. A P value less than 0.05 was considered statistically significant.