[18F]FLT and [18F]FDG PET for Non-invasive Treatment Monitoring of the Nicotinamide Phosphoribosyltransferase Inhibitor APO866 in Human Xenografts

Introduction APO866 is a new anti-tumor compound inhibiting nicotinamide phosphoribosyltransferase (NAMPT). APO866 has an anti-tumor effect in several pre-clinical tumor models and is currently in several clinical phase II studies. 3′-deoxy-3′-[18F]fluorothymidine ([18F]FLT) is a tracer used to assess cell proliferation in vivo. The aim of this study was non-invasively to study effect of APO866 treatment on [18F]FLT and 2-deoxy-2-[18F]fluoro-D-glucose ([18F]FDG) uptake. Methods In vivo uptake of [18F]FLT and [18F]FDG in human ovary cancer xenografts in mice (A2780) was studied at various time points after APO866 treatment. Baseline [18F]FLT or [18F]FDG scans were made before treatment and repeated after 24 hours, 48 hours and 7 days. Tumor volume was followed with computed tomography (CT). Tracer uptake was quantified using small animal PET/CT. One hour after iv injection of tracer, static PET scans were performed. Imaging results were compared with Ki67 immunohistochemistry. Results Tumors treated with APO866 had volumes that were 114% (24 h), 128% (48 h) and 130% (Day 7) relative to baseline volumes at Day 0. In the control group tumor volumes were 118% (24 h), 145% (48 h) and 339% (Day 7) relative to baseline volumes Day 0. Tumor volume between the treatment and control group was significantly different at Day 7 (P = 0.001). Compared to baseline, [18F]FLT SUVmax was significantly different at 24 h (P<0.001), 48 h (P<0.001) and Day 7 (P<0.001) in the APO866 group. Compared to baseline, [18F]FDG SUVmax was significantly different at Day 7 (P = 0.005) in the APO866 group. Conclusions APO866 treatment caused a significant decrease in [18F]FLT uptake 24 and 48 hours after treatment initiation. The early reductions in tumor cell proliferation preceded decrease in tumor volume. The results show the possibility to use [18F]FLT and [18F]FDG to image treatment effect early following treatment with APO866 in future clinical studies.


Introduction
During development of new anti-cancer drugs, methods to assess early treatment effect and discriminate between responders and non-responders are widely needed. The use of positron emission tomography (PET) to discriminate between responders and non-responders early following treatment initiation is desirable. The PET technique allows for non-invasive characterization of biological functionality in tumor tissue. 39-deoxy-39-[18F]fluorothymidine ([18F]FLT) is used to assess cell proliferation in vivo by PET, by measuring the activity of thymidine kinase 1 (TK1) which is up-regulated in the S-phase of cell cycle [1][2][3][4][5][6]. Uptake of [18F]FLT has in several pre-clinical models been shown to decrease following treatment with various chemotherapeutics much earlier than a response in tumor volume is evident [7][8][9][10][11][12][13][14][15][16]. However, changes in [18F]FLT uptake following treatment are variable and dependent upon mode of action for the different chemotherapeutic drugs [7,[9][10][11]13].
APO866 is a new anti-tumor compound which inhibits nicotinamide phosphoribosyltransferase (NAMPT), an enzyme involved in the biosynthesis of NAD. APO866 inhibits biosynthesis of cellular NAD from niacinamide by inhibiting NAMPT. Reduction in cellular NAD levels leads to ATP depletion and apoptosis [17,18]. Cancer cells have a high NAD turn over compared with normal cells. NAD is an important cofactor for poly(ADP-ribose) polymerase 1 (PARP1) which is involved in DNA repair and therefore cancer cells are more sensitive to NAMPT inhibition than normal cells. APO866 has anti-tumor effect in many tumor cells in vitro and several pre-clinical human xenograft tumor models [19,20]. A clinical phase I study with APO866 has already been conducted [21], and APO866 is currently in several clinical phase II studies.
APO866 treatment decreases cellular ATP content [17,19] and ATP is a cofactor for TK1 [22]. We therefore hypothesized that [18F]FLT uptake would be influenced by APO866 treatment due to the decrease in ATP and possibly a change in tracer uptake could be used to predict treatment response at an early stage.

Tumor Model
Animal care and all experimental procedures were performed under the approval of the Danish Animal Welfare Council (2006/ 561-1124). Female NMRI (Naval Medical Research Institute) nude mice (6-8 weeks old) were acquired from Taconic Europe (Lille Skensved, Denmark) and allowed to acclimatize for one week in the animal facility before any intervention was initiated. The human ovarian carcinoma cell line A2780 [23] (a gift from R. Ozols, Fox Chase Cancer Center Philadelphia, PA, USA) was used. A total of 10 7 cells in 100 mL medium mixed with 100 mL Matrixgel TM Basement Membrane Matrix (BD Biosciences, San Jose, CA, USA) were injected subcutaneously into the left and right flank respectively during anesthesia with 1:1 v/v mixture of HypnormH (Janssen Pharmaceutica, Beerse, Belgium) and Dor-micumH (Roche, Basel, Switzerland). The cell line has been tested free of mycoplasma. Cells were cultured in RPMI (Roswell Park Memorial Institute) medium 1640+ GlutaMAX (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal calf serum (Biological Industries, Israel) and 1% penicillin-streptomycin (Invitrogen) in 5% CO 2 at 37uC.

Experimental Design
Mice were divided into groups (n = 6-12 tumors per group) receiving either APO866 or vehicle treatment and [18F]FLT or [18F]FDG scans. Treatment was started at day 14-16 after implantation of tumor cells, where tumor volumes were on average 265 mm 3 . Mice in the treatment groups received APO866 (figure 1) in saline/PBS with 2% HP-b-CD and 9% propylene glycol at a dose of 15 mg/kg IP twice daily for 7 days. This is the maximum tolerated dose of APO866 and inhibits growth of the A2780 xenograft tumor model [20]. The control groups received vehicle twice daily for 7 days. The mice were scanned at baseline before treatment start with [18F]FLT or [18F]FDG and subsequently [18F]FLT or [18F]FDG scans were repeated at 24 h (Day 1), 48 h (Day 2) and 168 h (Day 7) hours after start of treatment with either APO866 or vehicle. Tumor volume was followed by microCT during the experiments [24][25][26]. In addition to the microCT measurement of tumor volume, tumor sizes were followed by external caliper measurements where tumor volume was calculated using the modified ellipsoidal formula K(length*width 2 ). At Day 7 tumors were excised and transferred to 280uC for Ki67 immunohistochemistry analysis. All tumors were weighted and tumor weight was correlated with both microCT and caliper determined tumor volume.

Synthesis of [18F]FLT and [18F]FDG
[18F]FLT was synthesized using 3-N-Boc-1-[5-O-(4,49-dimethoxytrityl)-3-O-nosyl-2-deoxy-b-D-lyxofuranosyl]thymine as precursor on a GE TracerLab MX Synthesizer as described earlier [7]. All reagents and [18F]FLT cassettes were purchased from ABX (Radeberg, Germany  [27]. One hour after tracer injection mice were anaesthetized with 3% sevoflurane (Abbott Scandinavia AB, Solna, Sweden) mixed with 35% O 2 in N 2 and fixed on a bed in presence of three fiducial markers allowing fusion of PET and CT pictures. A 10 min PET scan was acquired using a MicroPET Focus 120 (Siemens Medical Solutions, Malvern, PA, USA). After data acquisition, PET data were arranged into sinograms and subsequently reconstructed with the maximum a posteriori (MAP) reconstruction algorithm. The pixel size was 0.86660.86660.796 mm and in the center field of view the resolution was 1.4 mm full-width-at-half-maximum. The images were not corrected for attenuation or scatter.
Following the microPET scan, a microCT scan was acquired with a MicroCATH II system (Siemens Medical Solutions). A 7 minute and 10 seconds CT scan was performed with parameter settings: 360 rotation steps, tube voltage 60 kV, tube current 500 mA, binning 4 and exposure time 310 ms. The pixel size was 0.09160.09160.091 mm.
PET and microCT images were fused in the Inveon software (Siemens Medical Solutions). Before fusion region of interests (ROIs) were drawn on the CT pictures manually by qualitative assessment covering the whole tumors and subsequently tumor volume and tracer uptake, assessed by standard uptake values (SUV) mean and maximum, were generated by summation of voxels within the tomographic planes. SUV was defined as (C T *W)/D inj , where C T is radioactivity in the tissue, W is weight of the animal and D inj is injected dose. SUVmean gives information of the mean concentration of tracer and SUVmax is a measure of the voxel within the ROI which have the highest tracer concentration. For tumor volume measurements ROIs were drawn manually by delineation of the tumor boundary by qualitative assessment in several of the coronal planes on the microCT images. Summation of voxels from all the planes generated the tumor volume.

Ki67 Immunohistochemistry
After freezing at 280uC tissue samples were fixed in 4% neutral buffered formalin and subsequently embedded in paraffin. The anti-human Ki-67 clone MIB-1 (Dako cat.no F7268/FITC conjugated mouse antibody) was used. The reactions were detected using standard polymer technique with few modifications. Negative controls were performed by omission of primary antibody. Sections were dewaxed in xylene, rehydrated and pretreated before immunostaining. The slides were incubated with MIB-1 1:100 diluted in Power Block solution (BioGenex,cat. no. HK085-5K) for 30 min at room temperature. The sections were rinsed in TBST (365 minutes) and incubated with rabbit anti-FITC (Invitrogen, cat. no. A6413 1:250) for 20 min at RT. The reactions were detected with Super Sensitive polymer-HRP reagent (Biogenex, cat. no. QD410D-XCX) and visualised with DAB ImmPACT Substrate chromogen solution (Vector laboratories cat. co. SK-4105) following the manufactures instructions. Finally, slides were counterstained with Mayers haematoxylin.
The level of Ki67 present in the samples was measured as mean of three hot spots on one slide from each tumor. Hotspots were defined as areas with the highest Ki67 intensity and were identified by scanning the entire tumor slide in 106magnification. Ki67 positive cells were defined as cells with positive nucleus staining. The Ki67 staining index was defined as the number of Ki67 positive cells divided by the total number of tumor cells in an area on the 606magnification. Positive and negative controls from human tissue were placed on each slide for validation of the immunohistochemical procedure.

Statistical Analysis
Comparison of tumor volume between the treatment and control group were calculated using an unpaired student's t-test. For intra-group comparisons of tracer uptake ANOVA with Bonferroni post-hoc test was used. All data were tested to be normal distributed by means of Kolmogorov-Smirnov test. Correlations between SUVmax ratios and tumor growth were calculated using linear regression. Calculations were made in PASW 17.0 (IBM, Armonk, NY, USA). Data are reported as mean6SEM (unless stated otherwise) and P,0.05 was considered statistically significant.

Ki67 Immunohistochemistry
Expression of Ki67 protein measured by immunohistochemistry showed a significant correlation with [18F]FLT SUVmean (r 2 = 0.75; P,0.001). Differences in expression of Ki67 protein between the treatment (28.3%) and control group (52.5%) were significant (P = 0.022) ( figure 7). Some of the treated tumors had 0% Ki67 protein expression. Treatment with APO866 for 7 days resulted in no observable Ki67 positive cells in the slides for these tumors.

Discussion
In this study we investigated the effect of APO866 treatment on tumor uptake of [18F]FLT and [18F]FDG by PET in the A2780 ovarian cancer tumor model. By groups, uptake of [18F]FLT decreased after initiation of treatment and was significantly different from baseline already at 24 hours after treatment initiation. By groups, uptake of [18F]FDG did not change on  A measure of the maximal tracer uptake is often used when assessing treatment responses rather than the mean tracer uptake [28]. With SUVmax the most active metabolic region of the tumor is measured, and one advantage of this is that some of the volume delineation problems often found in SUVmean values is overcome. Much discussion has centered about if inhibiting the most active metabolic or proliferative regions with the anti-cancer treatment is likely to give the best treatment effect. Maximal uptake values are in several studies shown to be a prognostic factor for outcome [28][29][30][31][32]. In the present study we found that changes in SUVmax following treatment correlated better with later tumor volume effect than changes in SUVmean. Furthermore, we found a much greater decrease in [18F]FLT SUVmax values at Day 7 compared to SUVmean values.
Lower uptake of tracer at Day 7 compared to baseline for both [18F]FLT and [18F]FDG in the treatment groups was most likely not due to partial volume effect because of a decrease in tumor volume as the tumors did not shrink during the treatment. The changes in tracer uptake were therefore assumed to be due to physiologic changes in the tumor tissue which were further confirmed by the difference in the quantity of Ki67 positive cells between the treatment and control groups. Positive correlation between [18F]FLT uptake and the amount of Ki67 positive cells have earlier been reported by others [33][34][35][36].
APO866 is a relatively slow acting anti-cancer drug without any immediate cytotoxicity leading to ATP depletion and apoptosis [17,19]. Since ATP is a co-factor for TK1, a decrease in ATP level will possibly decrease uptake of [18F]FLT, and a change in ATP tumor level is likely to account for some of the [18F]FLT changes we found in this study. This may also explain why we observed earlier changes in [18F]FLT compared to [18F]FDG. Tumor ATP level is decreased 72 hours after treatment start with APO866 [19]. However, already at 24 hours after treatment start a decrease in NAD content was observed. The early change in NAD levels may contribute to the change in [18F]FLT tracer uptake we found already 24 hours after treatment start in this study. No studies have yet analyzed the level of ATP in mouse tumor models following treatment with APO866 and more experiments are accordingly needed in order to elucidate the exact mechanism behind the decreased [18F]FLT and [18F]FDG uptake following treatment with APO866.
Anti-cancer drugs being cytostatic or having low cytotoxicity often require long treatment time before exerting their effect on tumor volume. The advantage of measuring physiological changes as we do in this study with [18F]FLT and [18F]FDG PET compared to the tumor volume criteria is that we are able to measure a treatment effect earlier than we can measure a change in tumor volume.
When the tumors were studied individually, we found a significant positive correlation between the changes in SUVmax for both [18F]FLT and [18F]FDG between baseline and 48 hours and the tumor growth. Accordingly, those tumors which had the greatest decrease in SUVmax values for both [18F]FDG and [18F]FLT 48 hours after treatment initiation were those which had the lowest increase in tumor volume at the end of the experiment. It indicates that we, in addition to finding a decrease in the whole treatment groups, on the individual tumor level already at 48 hours after treatment initiation were able to identify the tumors that responded best to the treatment as it were those that had the greatest change in SUVmax values.
In conclusion, we found a decrease in tumor [18F]FLT uptake following treatment with the nicotinamide phosphoribosyltransferase inhibitor APO866. After 7 days of treatment tumor uptake of [18F]FDG was half of baseline uptake. Additionally, we found that the tumors which responded best to the treatment were those that had the greatest change in both [18F]FLT and [18F]FDG at 48 hours after treatment initiation. These results show the possibility to use [18F]FLT and [18F]FDG to image treatment effect early following treatment with APO866 in future clinical studies.