Visual Abstract
Abstract
Aggressive cancers such as triple-negative breast cancer (TNBC) avidly metabolize glutamine as a feature of their malignant phenotype. The conversion of glutamine to glutamate by the glutaminase enzyme represents the first and rate-limiting step of this pathway and a target for drug development. Indeed, a novel glutaminase inhibitor (GLSi) has been developed and tested in clinical trials but with limited success, suggesting the potential for a biomarker to select patients who could benefit from this novel therapy. Here, we studied a nonmetabolized amino acid analog, 18F-fluciclovine, as a PET imaging biomarker for detecting the pharmacodynamic response to GLSi. Methods: Uptake of 18F-fluciclovine into human breast cancer cells was studied in the presence and absence of inhibitors of glutamine transporters and GLSi. To allow 18F-fluciclovine PET to be performed on mice, citrate in the tracer formulation is replaced by phosphate-buffered saline. Mice bearing triple-negative breast cancer (TNBC) xenografts (HCC38, HCC1806, and MBA-MD-231) and estrogen receptor–positive breast cancer xenografts (MCF-7) were imaged with dynamic PET at baseline and after a 2-d treatment of GLSi (CB839) or vehicle. Kinetic analysis suggested reversible uptake of the tracer, and the distribution volume (VD) of 18F-fluciclovine was estimated by Logan plot analysis. Results: Our data showed that cellular uptake of 18F-fluciclovine is mediated by glutamine transporters. A significant increase in VD was observed after CB839 treatment in TNBC models exhibiting high glutaminase activity (HCC38 and HCC1806) but not in TNBC or MCF-7 exhibiting low glutaminase. Changes in VD were corroborated with changes in GLS activity measured in tumors treated with CB839 versus vehicle, as well as with changes in VD of 18F-(2S,R4)-fluoroglutamine, which we previously validated as a measure of cellular glutamine pool size. A moderate, albeit significant, decrease in 18F-FDG PET signal was observed in HCC1806 tumors after CB839 treatment. Conclusion: 18F-fluciclovine PET has potential to serve as a clinically translatable pharmacodynamic biomarker of GLSi.
Reprogramming of energy metabolism has been recognized as a hallmark of cancer (1). Dysregulation of cellular metabolic pathways enables cancer cells to meet their energetic and biosynthetic needs but also provides opportunities to selectively target cancer cells while sparing normal tissues. Targeting such a cancer-specific metabolic signature could be useful for treating triple-negative breast cancer (TNBC), which lacks a subtype-specific treatment because of absence of the estrogen receptor, the progesterone receptor, and the human epidermal growth factor receptor 2. Glutaminolysis, the metabolic pathway of glutamine, is used by many aggressive cancers, including TNBC (2,3), and hence might be exploited for therapy. Conversion of glutamine to glutamate by the mitochondrial enzyme glutaminase represents the first and rate-limiting step in the pathway. In cancers, it is the kidney type rather than the liver type of glutaminase that is predominantly expressed. Highly potent and specific inhibitors of glutaminase (GLSi) have been developed. Several of these agents have shown promise in preclinical studies (4–7), and one agent, CB839 (Telaglenastat developed by Calithera Bioscience), has advanced into clinical trials. Although CB839 has been well tolerated in clinical trials, efficacy has been variable (8–10). These trials, though, did not include biomarkers to assess target impact and guide patient selection, potentially confounding the overall study results. Although enhanced glutamine metabolism is frequently observed in TNBC—more so than in other subtypes of breast cancer—individual TNBCs are highly heterogeneous and exhibit a wide range of glutaminase activity (4). As such, a method to evaluate the pharmacodynamic response of TNBC to GLSi would be highly valuable for patient selection in the clinical setting when glutaminase activity cannot be readily assessed.
Previous studies have elucidated an inverse relationship between glutaminase activity and glutamine pool size, providing a paradigm for imaging glutamine metabolism with amino acids analogs. Tumors with a high glutaminase activity have a low cellular glutamine concentration (pool size) accompanied by a high glutamate (the product of the glutaminase) level (4). Upon glutaminase inhibition, the cellular glutamine pool increases because of blockade of glutamine conversion to glutamate (4,11). It follows that such an increase in tumor cellular glutamine pool size can serve as a specific pharmacodynamic marker for GLSi.
Our approach toward measuring the change in tumor glutamine pool size relies on tracers that mimic bidirectional glutamine transport into and out of the cancer cell but are not metabolized by glutaminase. Under these conditions, and assuming stable kinetics for glutamine transport and metabolism over the course of the imaging study (∼1 h), the PET signal will fit a single reversible tracer kinetic model (12) in which the distribution volume (VD) provides a direct measure of intracellular glutamine concentration. We and others have previously shown that cellular uptake of glutamine analog 18F-(2S,4R)4-fluoroglutamine (18F-4F-glutamine) is mediated by the native glutamine transporters (13) but is a poor substrate for kidney-type glutaminase expressed in the tumor (14,15). Furthermore, in human breast cancer xenograft models, the VD of 18F-4F-glutamine PET or the tumor-to-blood ratio (T/B) 30–45 min after injection can serve as a pharmacodynamic marker of GLSi (15,16).
18F-4F-glutamine, however, is an investigational agent and has been reported to undergo defluorination (17,18), which might confound the PET signal and potentially mask bone disease, a common site of breast cancer metastasis. To overcome these limitations, we considered 18F-fluciclovine (Axumin; Blue Earth Diagnostics), a metabolically inert analog of l-leucine (16), with uptake into cancer cells mediated by glutamine transporters (19,20). Hence, 18F-fluciclovine may serve as an alternative tracer for measuring glutamine pool size and the impact of GLSi. Although 18F-fluciclovine has been studied under investigational settings for breast cancer (21,22), its utility as a metabolic marker has not been established. Here, we performed a proof-of-principle study of 18F-fluciclovine as a pharmacodynamic marker for GLSi, comparing with 18F-4F-glutamine by employing the same tumor models and analysis methods (15,16). We were able to show that the uptake of 18F-fluciclovine is mediated by glutamine transporters; 18F-fluciclovine has reversible kinetics that reflect cellular glutamine pool size, and short-term exposure to GLSi leads to significant increases in 18F-fluciclovine VD similar to those in 18F-4F-glutamine.
MATERIALS AND METHODS
Materials
18F-fluciclovine was provided by Blue Earth Diagnostics. A vehicle solution of 25% (w/v) hydroxypropyl-β-cyclodextrin in 10 mM citrate (pH 2) (15) and CB839 formulated in the vehicle or in dimethyl sulfoxide were provided by Calithera Biosciences. 18F-(2S,4R)4-fluoroglutamine (15) and 18F-FDG were produced at the PET Center, University of Pennsylvania.
The following reagents, assay kits, and cell lines were purchased: ion-exchange resin columns (catalog no. 731621) from Bio-Rad; a citrate assay kit (catalog no. MAK057), l-glutamine, GPNA (l-γ-glutamyl-p-nitroanilide), and BCH (2-aminobicyclo-(1,2)-heptane-2-carboxylic acid) from Sigma; syringe filters (catalog no. 09-720-3) and 96-well strips plate (catalog no. 07-200-97) from Fisher Scientific; human TNBC (HCC1806, HCC38, and MDA-MB-231) and estrogen receptor–positive (MCF-7) breast cancer cell lines from American Type Culture Collection. The cell lines were authenticated using the short-tandem-repeat DNA profiling method and were used within 6 mo from the date of purchase. Cells were maintained in complete culture medium consisting of RPMI-1640 (catalog no. 25-506; GenClone) supplemented with 10% HyClone fetal bovine serum without antibiotics.
Removal of Citrate from the Formulation of 18F-Fluciclovine
After the deionized water was released from the column resin, the column was loaded with the tracer solution (∼0.9 mL) provided by the vendor. An equal volume (0.9 mL) of phosphate-buffered saline (PBS) was added to the column, and the eluted solution was collected (∼0.9 mL in ∼1–2 min) in a sterile test tube. This was repeated by adding PBS (0.9 mL) and collecting the elution in another sterile test tube. The 18F activity in the eluted solution was estimated by a dose calibrator (CRC-7; Capintec), whereas the pH and residual citrate in the solution were assessed by pH test strips and citrate assay, respectively.
In Vitro Cell Uptake of 18F-Fluciclovine
Breast cancer lines HCC38, HCC1806, and MCF-7 were included. Cells (30,000, 25,000, and 30,000 cells per well for HCC38, HCC1806, and MCF-7, respectively) were seeded in a 96-well strip plate and incubated overnight in complete culture medium. To study tracer uptake, the cells were incubated in PBS containing 5 mM glucose and 100 μM glutamine to which 18F-fluciclovine (∼300,000 cpm/well) was added at time zero. To test whether 18F-fluciclovine uptake is mediated by glutamine transporters, glutamine (5 mM), GPNA (1 mM), BCH (10 mM), or dimethyl sulfoxide control (0.05%) in PBS containing 5 mM glucose was added at time zero. After a specified incubation time (5, 15, 30, 60, and 120 min), supernatant was aspirated, and the wells were washed twice with cold PBS. After washing, activity in each well was counted on a γ-counter (2470 Wizard2; Perkin Elmer), and the amount of protein was estimated by the Lowry method (15). To inhibit glutaminase activity, the cells were incubated in culture medium containing 1 μM CB839 or 0.05% dimethyl sulfoxide as a control for 24 h before cell uptake studies because cellular metabolite concentration remains unchanged during 24 h of incubation with CB839 (4); the drug was present during incubation with the radiotracer.
Human Breast Cancer Xenograft Models and Treatment with GLSi (CB839)
All animal procedures were approved by the institutional animal care and use committee of the University of Pennsylvania. NCR athymic nu/nu mice (female, 7 wk old) were purchased from Charles River. To establish the human breast cancer xenografts, 1 × 106 HCC1806 cells in 100 μL of PBS and 5 × 106 MCF-7 or HCC38 cells in 150 μL of Matrigel (Corning) solution (1:1 mixed with PBS) were subcutaneously inoculated into the right flank of the mice. Tumor size was measured by a reader who was blinded to the PET image analysis using a caliper in 2 orthogonal directions (a and b, with b being the shorter dimension) using the formula volume = (1/6) × π × a × b2. CB839 formulated in the vehicle solution (15) was administered via oral gavage (200 mg/kg twice daily) for 2 d; control mice received the same volume of vehicle solution (∼0.25 mL).
Quantification of Glutaminase Activity in Cells and Tumor Tissues
To process cells for glutaminase activity, a pellet of 10 million cells was suspended in 1 mL of homogenization buffer (50 mM tris-acetate, pH 8.6; 15 mM K2HPO4; 150 mM KCl; 0.25 mM ethylenediaminetetraacetic acid; and 1 mM dithiothreitol including 1x complete protease inhibitor) in a 2-mL tube containing ceramic beads (Bertin Instruments). Tumor tissue was clamp-frozen in liquid nitrogen upon euthanasia of the mouse and kept in a −80°C freezer. The homogenization buffer (as above) was added to the weighed tissue (v/w = 10/1) in a 2-mL tube containing ceramic beads.
Cell or tissue samples were homogenized at 4°C using a Precellys Evolution homogenizer equipped with Cryolys Evolution (Bertin Instruments). The supernatant was removed and passed through a gel filtration spin column (Zeba spin desalting column; Fisher Scientific). The filtered supernatant and substrates which include glutamate dehydrogenase, nicotinamide adenine dinucleotide phosphate, and glutamine were mixed to start the enzymatic reaction generating reduced nicotinamide adenine dinucleotide phosphate, which was detected by fluorescence (excitation wavelength, 340 nm/emission wavelength, 460 nm) and recorded every minute for 15 min at 25°C in a microplate reader (SpectraMax M5; Molecular Devices). Specific activity was calculated from the background-corrected velocity (nmol reduced nicotinamide adenine dinucleotide phosphate/min) divided by the amount of protein. A standard curve was produced using pure glutaminase (catalog no. G8880; Sigma).
In Vivo PET/CT Imaging and Analysis of Dynamic PET Images
In vivo PET/CT was performed on the Molecubes modular system (Molecubes Corp.) before and after 4 doses of CB839 or vehicle treatment. Approximately 7,400–9,250 kBq (200–250 μCi) of 18F-fluciclovine in 0.2 mL was injected into the tail vein catheter, and a 45-min dynamic PET scan was started immediately and followed by a 2-min CT scan. The dynamic PET data were reconstructed with a temporal resolution of 10 s/frame × 6 frames, 1 min/frame × 9 frames, and 5 min/frame × 7 frames. 18F-FDG PET imaging was applied to a separate cohort of mice subjected to the same treatment. After 4 h of fasting, the mice were injected with 3,700 kBq (100 μCi) of 18F FDG and allowed to recover from the anesthesia. Two hours after 18F FDG injection, the mice were anesthetized and underwent a 15-min static PET scan.
PET data were analyzed using PMOD software, version 3.711. A spheric region of interest (ROI) equal to an eighth the volume of the tumor (measured by caliper before imaging) was placed over the hottest region of the tumor while avoiding activities from nearby bones by referencing the CT image. The time–activity curve of the tumor was then obtained from the ROI. To measure the activity in the blood, a 2 × 2 × 2 mm cubic region of interest was placed over the left ventricle of the heart of the mouse by referencing the CT image. The image analysis methods used in this study match those of 18F-4F-glutamine (15,16), including the tumor ROI size. Time–activity curves for the blood and tumor were imported to the general kinetic modeling tool (PKIN) of PMOD for analyses. The preliminary analyses showed that the kinetic data fit both a 1- and a 2-compartment model, but the 2-compartment model fit returned a small k3 value (0–0.02/min), consistent with a reversible uptake of the tracer as reported earlier for 18F-4F-glutamine in mice (16) and patients (23). Henceforward, given reversible uptake, the Logan plot (24), a simplified graphical method based on compartment model analysis, was used to estimate the tracer VD of the tumor.
Statistical Analysis
Data are presented as mean ± SD (represented by error bars). Statistical tests were performed using GraphPad Prism, version 6. A 2-sided Student t test was used to evaluate difference between 2 groups, with α set at 0.05.
RESULTS
Citrate assay results demonstrated that the column filtration protocol almost completely removed citrate and that more than 80% of 18F radioactivity was recovered in eluted PBS with an activity concentration of 170–185 MBq (4.6–5 mCi)/mL, suitable for injection into mice (Supplemental Figs. 1A and 1B; supplemental materials are available at http://jnm.snmjournals.org). The citrate-replaced tracer solution appeared safe, without causing weight loss over 7 d (Supplemental Fig. 1C).
To assess whether 18F-fluciclovine uptake is mediated by glutamine transporters, tracer uptake was measured in the presence and absence of cold glutamine and pharmacologic inhibitors of alanine-serine-cysteine transporter 2 (GPNA) and large neutral amino acid transporter 1 (BCH) as shown in Fig. 1. 18F-fluciclovine uptake was blocked approximately 100%, 82%, and 56% by cold glutamine, BCH, and GPNA, respectively, at 30 min compared with the control (P < 0.00001 for all).
Cellular uptake of 18F-fluciclovine is mediated by glutamine transporters. Uptake into HCC1806 cells was blocked by excess of glutamine (A), BCH (B), or GPNA (C). Glutamine (5 mM), BCH (10 mM), or GPNA (1 mM) was added at time zero to incubation medium made of PBS containing 5 mM glucose (final concentration of these reagents are specified in parentheses). Each point in graph represents mean ± SD from 4 replicates. Gln = glutamine.
We then evaluated the impact of GLSi on 18F-fluciclovine uptake in human TNBC (HCC38 and HCC1806) and in estrogen receptor–positive breast cancer (MCF-7) cells, which exhibit distinct glutaminase activity (HCC38 > HCC1806 ≫ MCF-7). Consistent with a reversible uptake pattern, the uptake peaked at around 30 min, followed by washout as shown in Figures 2A–2C: at 30 min, tracer uptake was significantly higher in CB839-treated HCC38 and HCC1806 cells than in vehicle controls, but the difference was minimal in MCF-7 cells. Corroborated by glutaminase activities in the presence and absence of GLSi (Fig. 2D), these data suggest that inhibition of active glutaminase results in a large increase in glutamine pool size that is reflected in significant increases in 18F-fluciclovine uptake in these cells.
Cellular uptake of 18F-fluciclovine and glutaminase activity in TNBC and estrogen receptor–positive breast cancer cells. Dynamic uptake profile of HCC38 (A), HCC1806 (B), and MCF-7 (C) cells is shown, as well as glutaminase activities after CB839 or vehicle treatment (D). Error bars are SD. P values compare CB839 vs. vehicle group in D: HCC38/4.58 × 10−5, HCC1806/7.12 × 10−5, MCF-7/0.22. In A–C, each point in graph represents mean ± SD from 4 replicates; in D, each bar represents mean ± SD from 3 mice. VEH = vehicle.
Logan plots of PET data acquired before treatment revealed a distinct slope (VD) for the 3 models (HCC38 < HCC1806 < MCF-7, Supplemental Fig. 2A), inverse to their glutaminase activity. In vivo static PET images (last frame) revealed an increase in 18F-fluciclovine PET signal in HCC38 and HCC1806 tumor after GLSi treatment (Supplemental Fig. 2B). Changes in VD were then examined: in the HCC1806 model, a significant increase in VD was observed after a short course (2 d) of CB839 treatment (pretreatment vs. posttreatment VD: 1.24 ± 0.26 vs. 2.05 ± 0.26, P = 0.0055, paired t test, n = 6, Fig. 3A); for comparison, the VD of the vehicle control group showed a nonsignificant decrease (pretreatment vs. posttreatment VD: 1.33 ± 0.24 vs. 1.05 ± 0.20, P = 0.24, paired t test, n = 7). Changes in T/B shown in Supplemental Figure 3 are consistent with changes in VD. Two additional TNBC models were examined in fewer mice: in HCC38 model (high glutaminase), VD was increased by an average of 23% after CB839 (n = 3), compared with a 37% decrease after vehicle treatment (n = 2); in MDA-MB-231 model (low glutaminase activity), VD was decreased by 5% and 7% after CB839 treatment (n = 1) and vehicle treatment (n = 2), respectively.
18F-fluciclovine VD before and after GLSi treatment and change in VD induced by GLSi or vehicle treatment and tumor glutaminase activities in TNBC and estrogen receptor–positive BC. (A) Pre- and posttreatment VD of individual mice bearing HCC1806 tumor enrolled in vehicle-treated group (n = 7) or CB839-treated group (n = 6). (B) Percentage change in VD [100* (Posttreatment VD-Pretreatment VD)/Pretreatment VD] of vehicle-treated and CB839-treated groups for HCC1806 tumors (n = 7 and 6, respectively) and for MCF-7 tumors (n = 5 and 4, respectively). (C) Glutaminase activities of vehicle- and CB839-treated tumors (HCC1806 after CB839: 1.46 ± 0.13 [n = 4]; HCC1806 after vehicle: 4.56 ± 0.44 [n = 3]; MCF-7 after CB839: 1.07 ± 0.17 [n = 5]; MCF-7 after vehicle: 0.99 ± 0.12 [n = 3].
We further compared the VD of 18F-fluciclovine versus that of 18F-4F-glutamine PET (Fig. 4): in the HCC1806 model, a large increase in VD (slope of Logan plot) was observed in both 18F-fluciclovine and 18F-4F-glutamine after CB839 treatment compared with baseline (Figs. 4A–4B); in MCF-7, only a modest change in the slope was detected (Figs. 4C–4D). In some mice, an overall increase in 18F PET signal was observed after CB839 treatment, likely due to an increase in plasma glutamine level after CB839 exposure (15), leading to decreased clearance of 18F-4F-glutamine or 18F-fluciclovine. Groupwise, CB839 treatment led to an increase in VD (89% ± 39%), compared with a decrease (−16% ± 11%) after vehicle treatment in the HCC1806 model (Fig. 3B, P < 0.05); modest changes were observed in MCF-7 tumors after CB839 or vehicle treatment (P = 0.6). These values compare favorably with those of 18F-4F-glutamine PET: T/B estimated directly from PET images acquired from 30 to 45 min is a proxy of VD and was increased by 34% ± 14% in HCC1806 tumors after CB839, compared with −11% ± 24% after vehicle treatment (15), and T/B and VD were highly correlated (r2 = 0.92) (16). Glutaminase enzymatic activity measured from tumor tissues revealed a large, 3-fold, reduction in HCC1806 tumors after CB839 treatment compared with vehicle controls, whereas there was no significant change in MCF-7 tumors (Fig. 3C).
Comparison of 18F-fluciclovine vs. 18F-4F-glutamine PET for detecting pharmacodynamic responses to GLSi. Shown are Logan plots and PET images of 18F-Fluciclovine from representative mice bearing HCC1806 (A) and MCF-7 tumor (B) at baseline and post CB839 treatment. Logan plots and PET images of 18F-4F-Gln from HCC1806 (C) and MCF-7 (D) bearing mice are shown for the same treatment. Sample sizes for 18F-4F-glutamine study were reported earlier (15), whereas those for 18F-fluciclovine are described in Figure 3. Same color scale (based on percentage of maximum signal intensity) is applied to both pretreatment and posttreatment PET images.
Finally, we tested whether 18F-FDG PET can detect responses to CB839 treatment (Supplemental Figs. 4A–4B). A moderate albeit significant decrease in 18F-FDG PET signal (percent of injected dose per gram, %ID/g) was observed in HCC1806 tumors after CB839 treatment (pretreatment vs. posttreatment %ID/g: 2.94 ± 0.24 vs. 2.29 ± 0.21, n = 11, P < 0.05); a decrease was also observed in the vehicle group, but it was not significant (pretreatment vs. posttreatment %ID/g: 3.47 ± 0.59 vs. 2.47 ± 0.016, n = 5, P = 0.31).
DISCUSSION
Blockade of 18F-fluciclovine cell uptake by glutamine, as well as by pharmacologic inhibitors of large neutral amino acid transporter 1 (BCH) and alanine-serine-cysteine transporter 2 (GPNA), confirms that 18F-fluciclovine enters the cells via glutamine transporters. Indeed, prior studies have shown that alanine-serine-cysteine transporter 2 mediates glutamine uptake in many cancer cells (13,25), including TNBC (26). In vivo PET studies demonstrated a large increase in VD in TNBC models exhibiting high glutaminase activity (HCC1806 and HCC38) after a short course of GLSi treatment. However, no increase was detected after vehicle (control) treatment or in models exhibiting low glutaminase activity (MD-MBA-231 and MCF-7). These results are corroborated by glutaminase activities in tumor specimens and support that increased 18F-fluciclovine uptake reflects inhibition of active glutaminase. The ability of 18F-fluciclovine to report a pharmacodynamic response to GLSi enables early detection of target engagement and prediction of GLSi futility if the tumor fails to demonstrate an alteration in tracer uptake kinetics in response to the drug. This approach using a target-specific drug with short-term exposure as a perturbagen, and a mechanistically matched tracer, has been tested in other scenarios, including human studies of drug efflux inhibitors (27), estrogen receptor agonists (28), and estrogen receptor blockade (29). It is possible, and likely, that longer exposure may lead to differences in results, as was seen in studies that tested residual 18F-fluciclovine as a marker of resistant, viable tumor (22). The time course of the increase in tumor glutamine pool size and matched alteration in 18F-fluciclovine kinetics will need to be studied in future preclinical and clinical studies testing 18F-fluciclovine as a marker to guide glutaminase-targeted therapy.
Our data also revealed that 18F-FDG PET is not a sensitive marker for GLSi, likely because glucose uptake or metabolism is not directly impacted by a short exposure to GLSi (4). 18F-FDG can likely serve as a marker of cell viability or death over extended periods, as it does for a variety of cancer cell types and treatments (30).
The approach of using amino acid analogs that are not metabolized or substrates for biosynthesis differs fundamentally from that of metabolized tracers such as 11C-glutamine (31), which participates in biosynthesis and is metabolized through glutaminolysis to produce 11C-CO2 and radiolabeled metabolites (e.g., 11C-glutamate), leading to a complex distribution of the radiolabel that must be accounted for. Although 11C-glutamine has recently been studied in human subjects (32), the complexity behind the 11C PET signal would be a challenge for proper interpretation of PET signal changes and widespread clinical adoption. In contrast, complexity is mitigated by nonmetabolized tracers, such as 18F-fluciclovine, whereby enzymatic activity and drug effect are inferred from changes in pool size (33,34). The favorable comparison of the VD of 18F-fluciclovine versus 18F-4F-glutamine (Fig. 4) strongly suggests that clinical translation could be facilitated by 18F-fluciclovine because of its superior chemical stability, widespread consensus regarding lack of metabolism (35), and commercial availability.
Our study had a few limitations. First, whereas changes in VD observed in 2 other TNBC models (HCC38 and MDA-MB-231) were consistent with the expectation, their sample sizes were relatively small. However, a primary goal of this study was to compare the VD of 18F-fluciclovine with that of 18F-4F-glutamine previously estimated in HCC1806 and MCF-7 models (15,16). Second, tumor glutamine pool sizes were not directly quantified; in a previous study, however, glutamine concentration was measured independently by proton nuclear magnetic resonance technique in HCC1806 and MCF-7 tumors after the same treatment regimen, and a strong correlation was found between tumor glutamine pool size and T/B, a proxy of VD (15).
CONCLUSION
The findings of the current study suggest that 18F-fluciclovine PET provides a tool for assessing tumor glutamine pool size and can serve as a pharmacodynamic marker for GLSi (such as CB839). The favorable biochemical properties and established availability of this tracer make it an attractive candidate for future human translation.
DISCLOSURE
This study was supported by R21CA198563, R01CA211337, and Komen SAC130060. 18F-fluciclovine and CB839 were generously provided by Blue Earth Diagnostics and Calithera, respectively. No other potential conflict of interest relevant to this article was reported.
KEY POINTS
QUESTION: Can 18F-fluciclovine PET report the pharmacodynamic effect of GLSi in glutaminase-active breast cancers?
PERTINENT FINDINGS: 18F-fluciclovine is taken up in breast cancer cells via glutamine transporters and is not metabolized; thus, 18F-fluciclovine is capable of tracking glutamine pool size. In human breast cancer models of differential glutaminase activity, a significant increase in 18F-fluciclovine VD was observed after GLSi treatment in tumor models exhibiting high glutaminase activity but not in those with low activity.
IMPLICATIONS FOR PATIENT CARE: A 18F-fluciclovine PET marker might provide information that can help select patients for treatment targeting glutamine metabolism such as GLSi.
ACKNOWLEDGMENTS
We are grateful to Dr. Varsha Viswanath and Daniel Kranseler for assistance in data processing and to Drs. Karen Linder and Hamed Hossein of Blue Earth Diagnostics for insightful discussions.
Footnotes
Published online Jun. 30, 2022.
- © 2023 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- Received for publication March 23, 2022.
- Revision received June 21, 2022.