L-Glutamine (Gln) has many important functions in the normal physiology, but its role in cancer biology has created a demand for glutamine PET radioligands (Zhu et al., 2017), such as (2S,4R)-4-[18F]fluoroglutamine ([18F]FGln) (Lieberman et al., 2011; Venneti et al., 2015). In contrast to L-glutamine, 18F-labelled glutamine analogues are not metabolized in the TCA cycle, but can be incorporated into proteins and peptides (Yang et al., 2017), and can be used assess transporter (ASCT2) activity or protein synthesis rate.
Human trials with [18F]FGln PET have provided promising results, with high uptake in aggressive tumours in several cancer types (Venneti et al., 2015; Dunphy et al., 2018). Fasting prior to PET study is not required, because tumour uptake was not decreased in non-fasting patients (Dunphy et al., 2018); there even seems to be a trend of increased tumour uptake in non-fasting patients compared to the fasting patients, possibly due to competition at transport or enzymatic processes. [18F]FGln uptake in bone marrow is relatively high, and reduced uptake could be used as an early indicator myelosuppression during chemotherapy (Zhu et al., 2019).
[18F]FGln is mainly transported by ASCT2, as shown by reduction of its uptake caused by ASCT2 inhibitor GPNA (Lieberman et al., 2011), and correlation of tumour uptake and ASCT2 expression (Hassanein et al., 2016). Impaired BBB or neuroinflammation do not lead to increased [18F]FGln uptake in brain models (Zhu et al., 2017). Cytoplasmic glutaminase can metabolize [18F]FGln into [18F]fluoroglutamate, which can further be transformed into α-ketoglutarate by alanine aminotransferase (AAT), with the loss of [18F]F- (Ploesll et al., 2012). Like L-glutamine, [18F]FGln can be transported from cytoplasm into mitochondria; mitochondrial enzymes transform it there again into [18F]fluoroglutamate, but the next step produces [18F]fluoro-α-ketoglutarate and does not lead to release of [18F]F- (Cooper et al., 2012). The fraction of [18F]FGln metabolites trapped in tumour cells is relatively low after one hour in animal model (Zhou et al., 2017). [18F]FGln seems not to be a good substrate for glutaminase, and its uptake represents mostly the glutamine transport and intracellular glutamine pool size (Zhu et al., 2017). MYC oncogenes increase glutamine metabolism in tumours, and [18F]FGln uptake is higher in MYCN-amplified than non-MYCN-amplified neuroblastoma cell lines and animal xenografts (Li et al., 2019). Mitochondrial glutaminase is the rate-limiting step in tumour cells ability to use glutamate in the TCA cycle, and small-molecule inhibitors for this enzyme have been developed. Inhibiting the enzyme leads to increased glutamate pool in the cytoplasm, which leads to increased competition for efflux, and can be seen as increased [18F]FGln uptake in the tumours of mice (Zhou et al., 2017).
The radioactivity uptake one hour after [18F]FGln injection is highest in the bone marrow and pancreas, and high also in the liver, heart muscle, and small intestine; (Dunphy et al., 2018). Excretion into urine is fast, and probably prevents the usage in imaging tumours close to urinary bladder. Uptake in the kidneys is very high, but decreases by time (Lieberman et al., 2011).
After bolus injection in human studies, 18F is cleared from the blood and plasma in biphasic manner (Dunphy et al., 2018).
Blood-to-plasma ratio, which may be needed for conversion of blood to plasma concentrations or vice versa, was 0.74±0.2 at 5 min and 0.81±0.3 at 60 min after tracer injection in humans (Dunphy et al., 2018). The ratio at 5 min is higher than would be obtained if 18F were in the plasma alone; that, and the the slow increase of the ratio is probably due to the free [18F]F-, which is known to be transported across the red blood cell membrane. The increase and the high variability in the ratios suggests that population-based blood-to-plasma transformation is not feasible.
In rats, plasma-to-blood ratio follows a sigmoidal pattern, with all radioactivity initially in the plasma, and reaching the ratio reaches level ∼1.3 at 60 min (unpublished data from Turku PET Centre).
Parent tracer fractions in human plasma, as measured with HPLC, were 86±19% at 1 min, 88±9% at 5 min, 81±8% at 15 min, and 65±21% at 30 min after injection; practically all the rest radioactivity in the plasma was due to free [18F]F- (Dunphy et al., 2018). The percentage of radioactivity trapped in the plasma pellet increased from ∼10% at 1 min to ∼26% at 30 min, suggesting that it is due to labelled polypeptides (Dunphy et al., 2018). In the study by Grkovski et al (2020) the parent fractions were initially lower and then higher, and 73±8% at 65 min and 69±9% at 158 min.
The fast defluorination causes marked uptake in the bones; at 60 min, the bone uptake was about half of the uptake seen in [18F]NaF PET studies (Dunphy et al., 2018). Blocking of the 2-amino group of [18F]FGln by formation of dipeptide can greatly improve the in vitro stability of the tracer, while after venous administration [18F]FGln is rapidly released from the dipeptide (Zha et al., 2018), but this will probably not help to solve the problem of in vivo defluorination.
In rat studies, the fraction of free (not bound to proteins) and unchanged [18F]FGln can be fitted with power function (Watabe et al., 2000); fractions are ∼85% at 5 min, ∼65% at 30 min, and ∼60% at 60 min after administration (unpublished data from Turku PET Centre). The plasma kinetics seem to differ between female and male mice (Miner et al., 2020).
Data analysis methods
Detailed modelling of [18F]FGln PET data may not be feasible, because the injected dose may not be pure (2S,4R)-4-[18F]fluoroglutamine, but may contain marked proportion of its stereoisomer, (2R,4R)-4-[18F]fluoroglutamine, which as an analogue of D-glutamine has very different pharmacokinetics and slower tissue uptake. Linear Logan plot and small k3 in irreversible two-tissue compartmental model (Zhou et al., 2017; Viswanath et al., 2021) applied to mice PET data suggest that tumour uptake is reversible, which further supports the use of [18F]FGln as a probe of glutamine transport, and not for measurement of protein synthesis rate. Since labelled glutamate and fluoride have tissue uptake, Zhou et al (2017) used whole blood TAC as input function in the analysis. In mouse glioma model, with metabolite and plasma protein binding corrected plasma input, irreversible tissue uptake component was observed, but still dominated by the reversible uptake component (Miner et al., 2020 and 2021).
High bone uptake may hinder the analysis of tumours that are close to bone, especially in the cortex of the brain.
In human cancer studies SUV peak method and tumour-to-plasma ratio have been used (Dunphy et al., 2018). Optimal scan time may depend on the tumour (Xu et al., 2020). In oncological brain studies, where healthy brain area is available, it has been used as reference region in calculation of SUV ratio (Venneti et al., 2015). Based on AIC, reversible two-tissue compartmental model (R2TCM) fitted best the [18F]FGln data from most lesions in the brain and thoracic and abdominal area (Grkovski et al., 2020), although the difference between irreversible and reversible models was not large. K1 from the compartmental model correlated best with SUV; while k3 was low in ∼half of the lesions, it was found to be reduced by treatment with glutaminase inhibitor (Grkovski et al., 2020). Although reversible one-tissue model (R1TCM) performed worst according to AIC, the VT based on it correlated better than VT from R2TCM with the results of Logan plot (Grkovski et al., 2020).
Adapted SUV peak method has been used in animal tumour models (Zhou et al., 2017). Tumour-to-blood ratio correlates well with VT, and can be used to detect the increased tumour cell glutamate pool after glutaminase inhibition (Zhou et al., 2017; Viswanath et al., 2021).
Cooper AJL, Krasnikov BF, Pinto JT, Kung HF, Li J, Ploessl K. Comparative enzymology of (2S,4R)-4-fluoroglutamine and (2S,4R)-4-fluoroglutamate. Comp Biochem Physiol B Biochem Mol Biol. 2012; 163(1): 108-120. doi: 10.1016/j.cbpb.2012.05.010.
Jeitner TM, Kristoferson E, Azcona JA, Pinto JT, Stalnecker C, Erickson JW, Kung HF, Li J, Ploessl K, Cooper AJL. Fluorination at the 4 position alters the substrate behavior of L-glutamine and L-glutamate: Implications for positron emission tomography of neoplasias. J Fluor Chem. 2016; 192(A): 58-67. doi: 10.1016/j.jfluchem.2016.10.008.
Lieberman BP, Ploessl K, Wang L, Qu W, Zha Z, Wise DR, Chodosh LA, Belka G, Thompson CB, Kung HF. PET imaging of glutaminolysis in tumors by 18F-(2S,4R)4-fluoroglutamine. J Nucl Med. 2011; 52(12): 1947-1955. doi: 10.2967/jnumed.111.093815.
Ploessl K, Wang L, Lieberman BP, Qu W, Kung HF. Comparative evaluation of 18F-labeled glutamic acid and glutamine as tumor metabolic imaging agents. J Nucl Med. 2012; 53(10): 1616-1624. doi: 10.2967/jnumed.111.101279.
Venneti S, Dunphy MP, Zhang H, Pitter KL, Zanzonico P, Campos C, Carlin SD, La Rocca G, Lyashchenko S, Ploessl K, Rohle D, Omuro AM, Cross JR, Brennan CW, Weber WA, Holland EC, Mellinghoff IK, Kung HF, Lewis JS, Thompson CB. Glutamine-based PET imaging facilitates enhanced metabolic evaluation of gliomas in vivo. Sci Transl Med. 2015; 7(274): 274ra17. doi: 10.1126/scitranslmed.aaa1009.
Yang L, Venneti S, Nagrath D. Glutaminolysis: a hallmark of cancer metabolism. Annu Rev Biomed Eng. 2017; 19: 163-194. doi: 10.1146/annurev-bioeng-071516044546.
Zhou R, Pantel AR, Li S, Lieberman BP, Ploessl K, Choi H, Blankemeyer E, Lee H, Kung HF, Mach RH, Mankoff DA. [18F](2S,4R)4-fluoroglutamine PET detects glutamine pool size changes in triple-negative breast cancer in response to glutaminase inhibition. Cancer Res. 2017; 77(6): 1476-1484. doi: 10.1158/0008-5472.CAN-16-1945.
Zhu L, Ploessl K, Zhou R, Mankoff D, Kung HF. Metabolic imaging of glutamine in cancer. J Nucl Med. 2017; 58: 533-537. doi: 10.2967/jnumed.116.182345.
Updated at: 2023-02-14
Created at: 2018-02-28
Written by: Vesa Oikonen, Anne Roivainen