PET imaging of hepatocellular carcinoma with anti-1-amino-3-[¹⁸F]fluorocyclobutanecarboxylic acid in comparison with l-[S-methyl-¹¹C]methionine

Currently, there is no “good” radiotracer for PET imaging to supplement the standard radiological imaging of primary liver cancers such as hepatocellular carcinoma (HCC), the third leading cause of cancer-related death world-wide [1]. Natural amino acids such as methionine (Met) was radiolabeled with ¹¹C for cancer imaging based on upregulated amino acid transport and protein synthesis in different tumor types, which resulted in a high uptake of Met seen with PET imaging [2]. The most notable results so far came from brain tumor studies where lower background uptake of Met in the brain led to more favorable tumor-to-background uptake ratios than could be achieved with a commonly used glucose analog, 2-[¹⁸F]fluoro-2-deoxy-d-glucose (FDG) [3]. Uptake of FDG reflects the increased cellular glucose metabolism, which has dramatically improved patient management in a large variety of cancers. Unfortunately, FDG has severe limitations for imaging primary liver cancers such as HCC due to a high false negative rate as many HCCs have relatively low FDG uptake [4]. A different radiotracer is needed for effective PET imaging of liver cancer.

l-[S-methyl-¹¹C]methionine ([¹¹C]Met, Fig. 1a) was tested for HCC imaging [5, 6]. Although the initial clinical study showed a high uptake of [¹¹C] Met in HCC [6, 7], the 20-min short half-life of ¹¹C decay limited further clinical development of [¹¹C]Met for liver cancer applications. The non-natural amino acids provided an alternative. The advantages of these synthetic amino acid analogs include their ability to incorporate longer-lived radionuclides such as ¹⁸F, which makes routine clinical applications feasible through commercial batch production for regional distribution, and the lack of radiolabeled metabolite formation, which simplifies kinetic analysis and avoids possible confounding accumulation of activity in non-target tissues.

Synthetic amino acid analog anti-1-amino-3-[¹⁸F]fluoro-cyclobutyl-1-carboxylic acid (anti-[¹⁸F] FACBC, short for [¹⁸F] FACBC Fig. 1b) was developed and FDA-approved as Fluciclovine (¹⁸F) or Axumin® for PET imaging of recurrent prostate cancer. This study was designed to investigate [¹⁸F]FACBC when “repurposed” for PET imaging of primary liver cancer such as HCC in comparison with [¹¹C]Met. The mechanism of [¹¹C]Met uptake in HCC was elucidated through our previous investigations [8, 9]. Contrariwise, [¹⁸F]FACBC is not metabolized by design, and its uptake is mediated mainly by amino acid transporters, which include l-type amino acid transporter 1 (LAT1), LAT2, and also the alanine, serine, and cysteine preferring transport system 2 (ASCT2), all of which are highly conserved among the species [10]. The affinity of these amino acid transporters for [¹⁸F]FACBC is well characterized [11] in the order of high to low (Michaelis-Menten kinetics, K_m, in micromolar) as ASCT2 (92.0 ± 32.0) >> SNAT2 (222.0 ± 29.3) ≥ LAT1 (230.0 ± 84.5) >> LAT2 (738.5 ± 87.6), where SNAT2 is the Na⁺-coupled neutral amino acid transporter 2 (also known as solute carrier family 38 member 2, or SLC38A2) while ASCT2, LAT1, and LAT2 are likewise known as SLC1A5, SLC7A5, and SLC7A8, respectively. Our analyses (Fig. 2) of the human RNA-seq data from The Cancer Genome Atlas (TCGA) showed a higher level of expression of ACST2 in HCC than in the surrounding liver tissues and LAT1, which revealed a similar difference. In comparison, SNAT2 and LAT2 showed no differential expression between liver tumor and tissue. ASCT2 seemed to be the main active transporter that is responsible for [¹⁸F]FACBC uptake in HCC. We thus tested [¹⁸F]FACBC for PET imaging of HCC, in comparison with [¹¹C]Met, by using a unique animal model of HCC with clinical relevance.

As a clinically relevant animal model, the eastern woodchuck (Marmota monax) developed HCC after chronic viral hepatitis infection when it harbors a DNA virus—the woodchuck hepatitis virus (WHV), a member of the family Hepadnaviridae, of which human hepatitis B virus (HBV) is the prototype. Like HBV, WHV infects woodchuck liver to cause acute and chronic hepatitis, which leads to the development of HCC within 2–4 years of life. The HCC in woodchucks is considered as recapitulating the human HCC with similar pathology and natural history [12‐14]. Our bioinformatics analysis (see the “Results” section) showed that ASCT2 and LAT1 have high expression in HCC than in the surrounding liver tissues in the woodchuck models. In particular, the base mean of ASCT2 expression is higher than that of LAT1 in HCC, which might have led to different uptake patterns between [¹¹C]Met and [¹⁸F]FACBC during PET imaging.

Radiolabeled amino acids were developed to image the increased level of amino acid transport that occurs in many tumor cells in contrast to the surrounding tissues [20]. Naturally occurring amino acids are usually labeled with ¹¹C such as [¹¹C]Met, which limits their use in routine clinical scans. The advantages of these non-natural, non-metabolized amino acid analogs include the ability to incorporate longer-lived radionuclides such as ¹⁸F and the lack of radio-metabolite formation.

Synthetic amino acid analogs with the ¹⁸F label are preferable for clinical cancer imaging if they achieve a comparable tumor-to-background uptake ratio to that of natural amino acids. (4S)-4-(3-¹⁸F-fluoropropyl)-l-glutamine ([¹⁸F]FSPG) is an l-glutamate derivative that is a substrate to the sodium-dependent cystine-glutamate exchange-transporter ×(C)(−) [21]. ¹⁸F-FSPG was tested in a human trial and showed contrast uptake in moderately and poorly differentiated HCCs, but remained challenging with well-differentiated HCC, similar to FDG [22]. There is a radiolabeled glutamine, l-[5-¹¹C] glutamine, and analog ¹⁸F-(2S,4R)4-fluoroglutamine, that were designed for imaging glutaminolysis in tumors [23, 24]. However, both displayed a high liver background uptake, therefore are not suitable for liver cancer imaging. The same is true for l-[2-¹⁸F]fluorotyrosine [25], which was initially developed for imaging increased activity of amino acid transporters in brain tumors.

There are other small molecule radio-ligands tested for PET imaging of HCC. Among them, [¹¹C]acetate and [¹¹C]choline (and [¹⁸F]fluoridated choline) have shown uptake in HCC [26‐29], but their utility needs to be clinically tested. Recently, Ga-68-labeled short peptide ligand against prostate-specific membrane antigen (PSMA), developed originally for PET imaging of prostate cancer, was tested for PET imaging of HCC [30, 31]. Since the PSMA targets seemed to be on tumor-associated vascular endothelial cells, not on liver cancer cells [32], further investigations are needed.

In this study, [¹⁸F]FACBC was tested in comparison with [¹¹C]Met for PET imaging of HCC using a clinically relevant animal model of spontaneous HCC. [¹⁸F]FACBC was developed and approved for PET imaging of prostate cancer. For prostate cancer imaging, Na⁺-coupled neutral amino acid transporter SNAT2 and another neutral amino acid transporter ASCT2 as well as Na⁺-independent l-amino acid transporter LAT1 are all upregulated. A recent independent clinical PET imaging study [33] discovered [¹⁸F]FACBC uptake in human HCC by accident. While ASCT2 is the main amino acid transporter that is responsible for [¹⁸F]FACBC uptake [34], LAT1 is the main transporter responsible for [¹¹C]Met uptake [35].

ASCT2 and LAT1 are upregulated in HCC based on bioinformatics analysis and validated by PCR (Figs. 2 and 3), and both [¹¹C]Met and [¹⁸F]FACBC displayed uptake in liver cancer. For the woodchuck models of HCC, which is a spontaneously developed HCC after chronic hepatitis infection, no human viruses or tissues are involved. The pathological assessment revealed the liver tumor as moderately differentiated similar to the human HCC that is proliferative in nature (Fig. 7). The woodchuck indigenous amino acid transporters exhibit a high degree of homology in amino acid sequence to the human orthologs (Fig. 4 and supplementary). This homology most likely allowed the radiolabeled amino acids, [¹¹C]Met and [¹⁸F]FACBC, developed as substrates for human transporters to be transported also by the woodchuck’s indigenous transporters during PET imaging.

Comparison between [¹¹C]Met and [¹⁸F]FACBC for their uptake and retention in HCC revealed some differences between the two radiolabeled amino acids. The difference in uptake (Fig. 5) can be explained by the possibly not entirely overlaying expression or distribution pattern of the two amino acid transporters, ASCT2 and LAT1, as each is responsible for a preferred radiotracer, [¹⁸F]FACBC and [¹¹C]Met, respectively. Future work will focus on validation by co-IHC staining the two transporters to show the differential localization. The difference in retention/clearance, mainly in the liver as reflected by the T/L ratio (Fig. 6), can be explained by differential metabolism of the two radiotracers. Hepatic metabolism of Met was studied previously [8, 9]. Briefly, after transport by the high affinity and low capacity LAT1, the low affinity but high-capacity LAT2, and also ASCT2, Met is converted primarily into amino-acyl-tRNA by the enzyme amino-acyl RNA synthetase. This process is the first step in protein synthesis, which is enhanced in liver cancer as well as in tumors. [¹¹C]Met can also be converted into S-adenosylmethionine via methionine adenosyltransferase and through PE-methylation pathway into phospholipid incorporation in the liver and liver cancer. This explains the apparent plateau (almost steady T/L ratio) in [¹¹C]Met uptake in both HCC and liver parenchyma as tracer metabolism was on-going during PET imaging and the radio-metabolites were retained.

On the other hand, [¹⁸F]FACBC is not metabolized by design and would not be retained after transported. Thus, efflux of a non-metabolized could potentially explain the decline of the initial uptake of [¹⁸F]FACBC in HCC. Figure 6 indicates that while the [¹⁸F]FACBC signal in both liver and HCC decreased over time, the liver background signal fell faster, which resulted in an increase in T/L ratio, indicating increased intracellular to extracellular transport. Another l-amino acid transporter that could be responsible for this observation could be SLC43A1 since it is involved in the efflux of branched chain amino acids from the liver to blood [36, 37]. Shown in Fig. 1, [¹⁸F]FACBC is an analog of branched-chain amino acid l-leucine and a likely substrate for SLC43A1. TCGA data showed a high level of SLC43A1 expression in both liver and liver cancer with a much tighter range of expression in the liver than in HCC. Future work would also need to include co-IHC staining of SCL43A1 for liver and HCC in addition to those for ASCT2 and LAT1, as well as conducting efflux assays of [¹⁸F]FACBC with liver tissues. The low extracellular levels of competing amino acids may also contribute to the efflux of [¹⁸F]FACBC.

[¹⁸F]FACBC is a radiopharmaceutical approved by FDA. “Repurposing” it for use other than prostate cancer will be straightforward. The neutral amino acid transporter (ASCT2) mainly responsible for [¹⁸F]FACBC uptake in the prostate cancer seemed also to be the primary transporter that is upregulated in HCC. Both [¹⁸F]FACBC and [¹¹C]Met depicted HCCs as discovered through the use of a clinically relevant animal model of spontaneously developed HCC in the woodchucks. The uptakes of [¹⁸F]FACBC and [¹¹C]Met rely on different principal amino acid transporters, ASCT2 mainly for [¹⁸F]FACBC and LAT1 for [¹¹C]Met, which could lead to different uptake patterns of the two radiotracers for the same HCCs. Once further developed and validated, PET imaging with [¹⁸F]FACBC will potentially help a number of clinical management decisions.