Synthesis and Preclinical Evaluation of Two Novel 68Ga-labeled Bispecific PSMA/FAP-targeted tracers with 2-Nal-containing PSMA-targeted Pharmacophore and Pyridine-based FAP-targeted Pharmacophore

1. Introduction

Prostate cancer is the second most common cancer and the fifth leading cause of cancer death in men worldwide, with an estimation of 1,410,000 new cases and 375,304 deaths in 2020 [1]. Based on the recent cancer statistics by Siegel et al. [2], 5-year relative survival rates of localized and regional prostate cancer are > 99%, however it decreased significantly to 32% for metastatic prostate cancer patients. Therefore, early and accurate detection is important to further improve the survival rates of prostate cancer patients. Recently, the application of radiopharmaceuticals for non-invasive detection and treatment of prostate cancer has shown positive outcomes, such as improving positive-lesion detection rate and prolonging imaging-based progression-free survival [3,4]. One of the promising radiotracers for metastatic prostate cancer diagnosis, which has been approved by the US FDA, is [⁶⁸Ga]Ga-PSMA-11, that targets prostate-specific membrane antigen (PSMA) [5,6]. PSMA is a type II transmembrane glycoprotein, also known as folate hydrolase 1 and glutamate carboxypeptidase II [7,8], which has been shown to be overexpressed in prostate tumor and the neovasculature of other types of cancer, such as oral squamous cell, breast, and renal cancers [9,10,11]. Therefore, many attempts to develop PSMA-targeted radioligands for prostate cancer imaging and radioligand therapy have been made.

The clinical data from the PSMA-targeted radioligand therapeutic agent, [¹⁷⁷Lu]Lu-PSMA-617, were shown to be effective in treating metastatic castration-resistant prostate cancer (mCRPC) patients [3]. However, patients with low to no PSMA-expressing lesions are ineligible to benefit from this treatment [12,13]. This intra-patient heterogeneous PSMA expression in prostate cancer lesions can lead to worse overall survival due to PSMA-negative lesions being misdiagnosed as false-negative [14,15]. One strategy to improve lesion detection in this patient cohort is by using bispecific radioligands to target PSMA and other overexpressed proteins in prostate cancer lesions, such as fibroblast activation protein (FAP). FAP is a type II transmembrane serine protease [16] that is highly expressed by cancer-associated fibroblasts (CAFs) in the tumor microenvironment (TME) [17]. It is overexpressed in 90% of epithelial tumors, including prostate cancer [18,19], and its expression is associated with worse prognosis [20], making it a promising target for cancer diagnosis and therapy. In fact, many new promising FAP-targeted radioligands are being developed and some are being validated clinically, such as [⁶⁸Ga]Ga-FAPI-04 [21], [⁶⁸Ga]Ga-FAPI-46 [22], and [¹⁷⁷Lu]Lu-FAP-2286 [23]. Despite the potential of FAP-targeted radioligands shown in these studies, their tumor retention and efficacy have not been promising in the clinic. Hence, by targeting PSMA which is overexpressed in tumor and FAP which is overexpressed in the tumor stroma, we hypothesized that it would potentially increase the lesion detection rate and tumor retention in prostate cancer patients.

Previous studies have demonstrated the potential of utilizing bispecific radioligands to increase detection sensitivity of prostate cancer imaging, such as PSMA/gastrin-releasing peptide receptor (PSMA/GRPR) bispecific radioligands that have higher tumor uptake in PSMA- and GRPR-expressing mouse tumor models compared to the monospecific radiotracers [24]. Recently, several PSMA/FAP bispecific radioligands have been reported: ⁶⁴Cu-, ¹⁸F, and ⁶⁸Ga-labeled PSMA/FAP bispecific radiotracers by Boinapally et al. [25], Hu et al. [26], and Wang et al. [27], respectively. Their studies confirmed the potential usage of PSMA/FAP bispecific ligands to improve tumor uptake in PSMA- and FAP-expressing tumor models in comparison to the monospecific counterparts.

Recently, our group also developed three ⁶⁸Ga-labeled bispecific PSMA/FAP radiotracers which incorporate quinoline-based FAP pharmacophore and anthracene-containing PSMA pharmacophore [28]. We observed a high blood uptake (5-12 %ID/g at 1 h post-injection) in the mouse model with decreased tumor uptake in comparison to the ⁶⁸Ga-labeled monospecific counterparts, PSMA-targeted [⁶⁸Ga]Ga-HTK03041 and FAP-targeted [⁶⁸Ga]Ga-FAPI-04. We suspected that the longer blood retention could be caused by the increased lipophilicity of the ligands and potential interaction, such as π–π stacking between the quinoline ring and anthracene ring which hindered the binding of tracers to FAP and PSMA, leading to decreased tumor uptake. Our goal in this study is to solve these issues and improve the binding affinity and tumor uptake by replacing the pharmacophores with less lipophilic motifs, such as replacing Ala(9-Anth) in the PSMA-targeted pharmacophore with 2-Nal and replacing the quinoline motif in the FAP-targeted pharmacophore with a more hydrophilic pyridine.

Many potent pyridine-based FAP inhibitors (FAPIs) have been reported [29,30], however there are still few reports on the development of pyridine-based FAP-targeted tracers. Previously we synthesized and evaluated two novel ⁶⁸Ga-labeled pyridine-based FAP-targeted tracers, [⁶⁸Ga]Ga-AV02053 and [⁶⁸Ga]Ga-AV02070, and compared their binding affinity and tumor uptake to that of [⁶⁸Ga]Ga-FAPI-04 [31]. We discovered that although [⁶⁸Ga]Ga-FAPI-04 has higher binding affinity towards FAP and higher tumor uptake, both of our pyridine-based tracers have lower uptake in blood, muscle and bone, leading to much higher tumor-to-background contrast ratios. Our results suggest that pyridine-based FAPI is more hydrophilic than quinoline-based FAPI, and has potential to help reduce blood uptake. Furthermore, we found that Ga-AV02070 (Figure 1A), which has a carbonyl group at the para position to the pyridine nitrogen, has a better binding affinity to FAP and a higher tumor uptake compared to Ga-AV02053, which has the carbonyl group at the meta position to the pyridine nitrogen. Hence, the pharmacophore of AV02070 is a promising candidate for the design of FAP-targeted tracers.

In this paper, we report the design, synthesis, and evaluation of two bispecific PSMA/FAP-targeted radiotracers, [⁶⁸Ga]Ga-AV01084 and [⁶⁸Ga]Ga-AV01088 (Figure 1B). The PSMA binding motif of AV01084 and AV01088 was based on the 2-Nal-containing PSMA-targeted tracer, [⁶⁸Ga]Ga-PSMA-617 (Figure 1A) [32], and their FAP-targeted motif was derived from our pyridine-based FAP-targeted radiotracer, [⁶⁸Ga]Ga-AV02070 (Figure 1A). The difference between the two tracers is the position of DOTA chelator linked to the lysine, which is the ε-amino group in AV01084 and the α-amino group in AV01088. Their potential for prostate cancer imaging was evaluated by in vitro competition binding assay, PET imaging, and ex vivo biodistribution studies in preclinical PSMA-expressing LNCaP and FAP-expressing HEK293T:hFAP tumor models. The results were then compared with those of corresponding monospecific tracers, [⁶⁸Ga]Ga-PSMA-617 and [⁶⁸Ga]Ga-AV02070.

2. Results

2.1. Synthesis of PSMA/FAP Bispecific Ligand

AV01084 and AV01088 (Figure 1B) were synthesized on solid phase. Briefly, Lys(Lys(ivDde)-Gly-tranexamic acid-2-Nal)-urea-Glu(OtBu)-OtBu was first constructed on solid phase, followed by amide coupling of the FAP-targeted motif, compound 5, and DOTA chelator. To synthesize compound 5 (Scheme 1), compound 1 was first coupled with 2,3,5,6-tetrafluorophenol (TFP) to obtain activated ester 2 in 71% yield. Compound 3 which was synthesized following literature procedures [31] was first Boc-deprotected using trifluoroacetic acid (TFA), followed by coupling with compound 2 to obtain compound 4 in 45% yield. t-Butyl protecting group of compound 4 was removed using TFA and compound 5 was obtained quantitatively.

To synthesize AV01084, compound 5 was coupled to Lys(Lys(ivDde)-Gly-tranexamic acid-2-Nal)-urea-Glu(OtBu)-OtBu, followed by removal of the ivDde group at the Lys side chain and subsequent coupling with the DOTA chelator. To synthesize AV01088, DOTA chelator was first coupled to Lys(Lys(ivDde)-Gly-tranexamic acid-2-Nal)-urea-Glu(OtBu)-OtBu, followed by deprotection of amino group at the Lys side chain and coupling with compound 5. The DOTA-conjugated ligands were then cleaved off from resin and purified by HPLC (Table S1). AV01084 and AV01088 were obtained in 12% and 7.2% yields, respectively.

Detailed syntheses and characterizations of nonradioactive Ga- and ⁶⁸Ga-complexed AV01084 and AV01088 are provided in the Supplementary Materials (Tables S2-S3). Nonradioactive Ga-complexed AV01084 and AV01088 were obtained in 60% and 21% yields, respectively, and their ⁶⁸Ga-labeled analogs were obtained in 33 – 64% decay-corrected radiochemical yields with ≥ 44 GBq/µmol molar activity and > 95% radiochemical purity.

2.2. Binding Affinity and Lipophilicity

The binding affinities of Ga-AV01084, Ga-AV01088, and Ga-AV02070 to PSMA were measured by a cell-based binding assay using PSMA-expressing LNCaP prostate cancer cells and were compared to the previously published Ga-PSMA-617 (K_i = 1.23 ± 0.08 nM) [33]. These ligands inhibited the binding of [¹⁸F]DCFPyL to LNCaP cells in a dose-dependent manner (Figure 2A) and the calculated K_i(PSMA) values for Ga-AV01084, Ga-AV01088, and Ga-AV02070 were 11.6 ± 3.25, 28.7 ± 6.05, and > 1,000 nM, respectively (n = 3).

The binding affinities of Ga-AV01084, Ga-AV01088, and Ga-PSMA-617 to human FAP were measured by an enzyme inhibition assay using Suc-Gly-Pro-AMC as the FAP substrate and were compared to the previously published Ga-AV02070 (IC₅₀ = 17.1 ± 4.6 nM) [31]. The human FAP enzymatic activity on the substrate was inhibited by these ligands in a dose dependent-manner (Figure 2B). The calculated IC₅₀ values for Ga-AV01084, Ga-AV01088, and Ga-PSMA-617 were 10.9 ± 0.67, 16.7 ± 1.53 and > 1,000 nM respectively (n = 3).

The lipophilicity of the ⁶⁸Ga-labeled AV01084 and AV01088 were calculated using LogD_7.4 measurement. The LogD_7.4 values of [⁶⁸Ga]Ga-AV01084 and [⁶⁸Ga]Ga-AV01088 were -3.61 ± 0.07 and -3.66 ± 0.25, respectively. The values indicate that both tracers are highly hydrophilic.

2.3. PET Imaging, ex vivo Biodistribution, and Blocking Studies

Representative PET images acquired at 1 h post-injection using [⁶⁸Ga]Ga-AV01084, [⁶⁸Ga]Ga-AV01088, [⁶⁸Ga]Ga-PSMA-617 and [⁶⁸Ga]Ga-AV02070 are provided in Figure 3. All the radiotracers were excreted primarily through the renal pathway. LNCaP tumor xenografts were clearly visualized by [⁶⁸Ga]Ga-AV01084 and [⁶⁸Ga]Ga-AV01088, but not visualized by [⁶⁸Ga]Ga-AV02070 (Figure 3A). HEK293T:hFAP tumor xenografts were barely visualized by the bispecific tracers ([⁶⁸Ga]Ga-AV01084 and [⁶⁸Ga]Ga-AV01088), but not visualized by [⁶⁸Ga]Ga-PSMA-617 (Figure 3B). The bispecific tracers have bone and joint uptake, which is commonly observed for FAP-targeted tracers. There was a high kidney uptake in mice injected with the bispecific tracers and [⁶⁸Ga]Ga-PSMA-617, but not in mice injected with [⁶⁸Ga]Ga-AV02070. Co-injection of [⁶⁸Ga]Ga-AV01088 with the PSMA inhibitor 2-PMPA (500 µg) reduced the uptake of [⁶⁸Ga]Ga-AV01088 in LNCaP tumors to almost background level (Figure 3A), confirming the uptake is PSMA-mediated. Similarly, co-injection of [⁶⁸Ga]Ga-AV01088 with FAPI-04 (250 µg) reduced the uptake of [⁶⁸Ga]Ga-AV01088 in HEK293T:hFAP tumors to almost background level (Figure 3B), confirming the uptake is FAP-mediated.

Biodistribution studies were conducted at 1 h post-injection with ⁶⁸Ga-labeled AV01084, AV01088, and AV02070 in LNCaP tumor-bearing mice (Figure 4, Table S4), and the previously reported data from [⁶⁸Ga]Ga-PSMA-617 obtained using the same tumor model are included for comparison [33]. The results were consistent with the observation from their PET images. Tumor uptake values of [⁶⁸Ga]Ga-AV01084, [⁶⁸Ga]Ga-AV01088, [⁶⁸Ga]Ga-PSMA-617 and [⁶⁸Ga]Ga-AV02070 were 9.05 ± 1.54, 8.85 ± 1.25, 16.7 ± 2.30, and 0.74 ± 0.21 %ID/g, respectively. There was a very low tumor uptake in mice injected with [⁶⁸Ga]Ga-AV02070, showing a very minimal FAP expression level in this tumor model. The uptake levels of these tracers in major organs and tissues are consistent with the trends observed in their PET images of LNCaP tumor-bearing mice. The bispecific tracers ([⁶⁸Ga]Ga-AV01084 and [⁶⁸Ga]Ga-AV01088) have significantly higher blood and bone uptake values than the monospecific tracers, [⁶⁸Ga]Ga-PSMA-617 and [⁶⁸Ga]Ga-AV02070 (blood uptake: 1.56-2.26 vs 0.28-0.75 %ID/g, p < 0.05; bone uptake: 0.88-2.47 vs 0.09-0.56 %ID/g, P < 0.05). This results in significantly lower tumor-to-blood ratio (4.09 ± 0.77 for [⁶⁸Ga]Ga-AV01084, 4.10 ± 0.88 for [⁶⁸Ga]Ga-AV01088 and 27.7 ± 6.28 for [⁶⁸Ga]Ga-PSMA-617) and lower tumor-to-bone ratio (3.70 ± 0.83 for [⁶⁸Ga]Ga-AV01084, 3.75 ± 1.28 for [⁶⁸Ga]Ga-AV01088, and 96.5 ± 47.6 for [⁶⁸Ga]Ga-PSMA-617). [⁶⁸Ga]Ga-AV01084 has higher heart, kidney and adrenal glands uptake than [⁶⁸Ga]Ga-AV01088 (Figure 4, P < 0.05).

Biodistribution studies were also conducted at 1 h post-injection with [⁶⁸Ga]Ga-AV01084, [⁶⁸Ga]Ga-AV01088 and [⁶⁸Ga]Ga-PSMA-617 in HEK293T:hFAP tumor-bearing mice (Figure 5 and Table S5), and the previously reported data from [⁶⁸Ga]Ga-AV02070 [31] using the same tumor model are included for comparison. Tumor uptake values of [⁶⁸Ga]Ga-AV01084, [⁶⁸Ga]Ga-AV01088, [⁶⁸Ga]Ga-PSMA-617 and [⁶⁸Ga]Ga-AV02070 were 1.90 ± 0.41, 1.20 ± 0.25, 0.26 ± 0.01, and 7.93 ± 1.88 %ID/g, respectively. There was a very low tumor uptake in mice injected with [⁶⁸Ga]Ga-PSMA-617, showing very minimal PSMA expression in this tumor model. The uptake levels of these tracers in the major organs and tissues are consistent with the trends observed in their PET images of HEK293T:hFAP tumor-bearing mice (Figure 4B and Table S5)

Co-injection of 2-PMPA reduced the average uptake of [⁶⁸Ga]Ga-AV01088 in LNCaP tumor xenografts by 67% (8.85 ± 1.25 %ID/g down to 2.95 ± 1.13 %ID/g at 1 h post-injection), confirming the specific uptake of [⁶⁸Ga]Ga-AV01088 in LNCaP tumor xenografts (Figure 6). Co-injection of FAPI-04 reduced the average uptake of [⁶⁸Ga]Ga-AV01088 in HEK293T:hFAP tumor xenografts by 43% (1.2 ± 0.25 %ID/g down to 0.51 ± 0.14 %ID/g at 1 h post-injection), confirming the specific uptake of [⁶⁸Ga]Ga-AV01088 in HEK293T:hFAP tumor xenografts as well.

3. Discussion

Previously, we synthesized three ⁶⁸Ga-labeled bispecific PSMA/FAP radiotracers, [⁶⁸Ga]Ga-AV01017, [⁶⁸Ga]Ga-AV01030, and [⁶⁸Ga]Ga-AV01038 [28], which incorporate a quinoline-based FAP-targeted pharmacophore and an anthracene-based PSMA-targeted pharmacophore of [⁶⁸Ga]Ga-HTK03041. The PSMA and FAP in vitro competition binding assays indicated that our bispecific tracers have lower binding affinity to PSMA compared to the monospecific PSMA-targeted ligand, Ga-HTK03041, but maintained a similar binding affinity to FAP when compared to the monospecific FAP-targeted ligand, Ga-FAPI-04. However, the bispecific tracers have higher uptake in major organs (blood, muscle, bone, and heart) and decreased tumor uptake in the mouse models in comparison to the ⁶⁸Ga-labeled monospecific counterparts, [⁶⁸Ga]Ga-HTK03041 and [⁶⁸Ga]Ga-FAPI-04.

Therefore, in this report we selected the pharmacophores of [⁶⁸Ga]Ga-PSMA-617 (in brown, Figure 1) and [⁶⁸Ga]Ga-AV02070 (in blue, Figure 1) for the design of our PSMA/FAP bispecific tracers as they are more hydrophilic and have high affinity for PSMA and FAP, respectively. These two pharmacophores are separated by a Lys-Gly linker (Figure 1B). The PSMA-targeted pharmacophore (Lys(tranexamic acid-2-Nal)-urea-Glu and the linker (Lys-Gly) was constructed directly on solid phase using the commercially available amino acids. For AV01084, the pyridine-based FAP-targeted ligand and the DOTA chelator were coupled to the α-amino group and side-chain of Lys, respectively. While for AV01088, the DOTA chelator and the pyridine-based FAP-targeted ligand were coupled to the α-amino group and side-chain of Lys, respectively. This allows us to investigate the effect of the position of the DOTA chelator on binding affinity and biodistribution of the bispecific tracers.

The enzymatic assay (Figure 2B) confirmed that the FAP binding affinities of Ga-AV01084 (IC₅₀ = 10.9 ± 0.67 nM) and Ga-AV01088 (IC₅₀ = 16.7 ± 1.53 nM) were comparable or even slightly better than that of the previously reported Ga-AV02070 (IC₅₀ = 17.1 ± 4.60 nM) [31]. To investigate if the PSMA-targeted pharmacophore has any effect on the overall FAP binding of our bispecific ligands, we also measured the FAP binding affinity of Ga-PSMA-617. The very weak binding affinity of Ga-PSMA-617 (IC₅₀ > 1,000 nM) suggests that the potent FAP binding affinity of our bispecific ligands is contributed mainly by the AV02070 pharmacophore.

The PSMA binding affinities of the bispecific ligands (K_i = 11.6 – 28.7 nM) were inferior to that of Ga-PSMA-617 (K_i = 1.23 ± 0.08) [33]. However, when compared to the bispecific tracers from our previous report (K_i = 20.1–54.4 nM), changing Ala(9-Anth) to 2-Nal, and changing quinoline to pyridine significantly increased the binding affinity of PSMA/FAP bispecific tracers to PSMA. This supports our hypothesis that there might be an interaction between the pharmacophores that interferes with the overall binding of the bispecific ligands to PSMA. Moreover, Ga-AV02070 has very minimal binding affinity to PSMA (K_i > 1,000 nM), indicating that the PSMA binding affinity of our bispecific ligands is contributed mainly by the PSMA-617 pharmacophore.

Despite having a lower uptake in the LNCaP tumor compared to [⁶⁸Ga]Ga-PSMA-617 (16.7 ± 2.30 %ID/g), [⁶⁸Ga]Ga-AV01084 and [⁶⁸Ga]Ga-AV01088 (9.05 ± 1.54 and 8.85 ± 1.25 %ID/g, respectively) had better tumor uptake than our previously reported three bispecific PSMA/FAP tracers ([⁶⁸Ga]Ga-AV01017, [⁶⁸Ga]Ga-AV01030 and [⁶⁸Ga]AV01038) (4.25 – 5.17 %ID/g) [28]. Moreover, co-injection of [⁶⁸Ga]Ga-AV01088 with 2-PMPA decreased the tumor uptake by 67%, indicating that the tumor uptake is specific to PSMA. We also observed a significant bone uptake (1.97 ± 1.15 %ID/g) when [⁶⁸Ga]Ga-AV01088 was co-injected with 2-PMPA, whereas there was a decreased bone uptake to background level (0.20 ± 0.06 %ID/g) when [⁶⁸Ga]Ga-AV01088 was co-injected with FAPI-04. This indicates that the bone uptake of [⁶⁸Ga]Ga-AV01088 might be due to specific binding of the FAP-targeted pharmacophore. Uptake of [⁶⁸Ga]Ga-AV01088 to HEK293T:hFAP tumor is also FAP-mediated, as demonstrated by the significantly decreased tumor uptake obtained when co-injected with FAPI-04.

Unlike the improved uptake in LNCaP tumor xenografts, the uptake of [⁶⁸Ga]Ga-AV01084 and [⁶⁸Ga]Ga-AV01088 (1.20-1.90 %ID/g) in HEK293T:hFAP tumor xenografts remains significantly lower than that of the monospecific counterpart, [⁶⁸Ga]Ga-AV02070 (7.93 ± 1.88 %ID/g, P < 0.05). One possible reason is the addition of the succinic acid linker, which might have interfered with the binding of our tracers to FAP, thus decreasing the tumor uptake. Piperazine-based linkers have been shown to be important for maintaining good tumor uptake of FAP-targeted tracers [34]. Therefore, incorporating a piperazine-based linker can also be considered for future modifications to improve uptake of PSMA/FAP bispecific tracers to the FAP-expressing tumors.

Compared to the monospecific tracers, the bispecific tracers have higher blood, bone, and muscle uptake. However, the current modified tracers have significantly decreased blood retention (0.79 – 2.26 %ID/g) when compared to the PSMA/FAP bispecific tracers in our previous report (5.75 – 11.94 %ID/g) [28]. This leads to better LNCaP tumor-to-blood contrast ratios for [⁶⁸Ga]Ga-AV01084 (4.09 ± 0.77) and [⁶⁸Ga]Ga-AV01088 (4.10 ± 0.88) (Table S4) than our previously reported PSMA/FAP bispecific tracers (0.48 – 0.89) [28]. The decreased blood retention might be contributed by increased hydrophilicity of the new bispecific tracers shown by their low LogD_7.4 values (< –3.60). This is consistent with previous findings that more hydrophilic radiotracers will have faster pharmacokinetics and clearance, hence, have better tumor-to-background contrast ratio [35]. In addition, there are no significant differences in the tumor uptake or the tumor-to-background contrast ratios between [⁶⁸Ga]Ga-AV01084 and [⁶⁸Ga]Ga-AV01088, which indicates that the position of the DOTA chelator to the lysine linker does not have a crucial effect on the pharmacokinetics of the tracers.

Previously, Wang et al. [27] reported a ⁶⁸Ga-labeled PSMA/FAP bispecific tracer, [⁶⁸Ga]Ga-FAPI-PSMA (Figure 7), consisting of a PSMA-targeted pharmacophore from PSMA-617 (in brown, Figure 7) and an FAP-targeted pharmacophore from FAPI-04 (in blue). It would be difficult to directly compare the performance of our bispecific radiotracers with [⁶⁸Ga]Ga-FAPI-PSMA as different tumor models were used for evaluation: PSMA-expressing LNCaP tumors and FAP-expressing HEK293T:hFAP tumors used in this report; PSMA-expressing 22Rv1 tumors and FAP-expressing U87 MG tumors used by Wang et al [27]. [⁶⁸Ga]Ga-FAPI-PSMA was shown to have better tumor uptake (SUVmax = 1.32 and 1.67 for 22Rv1 and U87 MG tumors, respectively) compared to the monospecific tracers, [⁶⁸Ga]Ga-PSMA-617 (SUVmax = 0.25 for 22Rv1 tumors) and [⁶⁸Ga]Ga-FAPI-04 (SUVmax = 0.45 for U87 MG tumors). This demonstrates the potential of utilizing PSMA/FAP bispecific tracers to improve the tumor uptake. However, similar to our results, [⁶⁸Ga]Ga-FAPI-PSMA also had higher background uptake compared to the monospecific tracers.

Replacing the quinoline-based FAP-targeted pharmacophore with a pyridine-based FAP-targeted pharmacophore helps decrease the background uptake (blood, bone, and muscle) of the bispecific tracers, however it also results in a significantly lower uptake in HEK293T:hFAP tumors. Therefore, future attempts on the use of a pyridine-based FAP-targeted pharmacophore for the design of PSMA/FAP bispecific tracers need to include optimization of linkers such as the use of a piperazine-based linker to further improve FAP binding affinity and tumor uptake.

4. Materials and Methods

4.1. Synthesis of Bispecific PSMA/FAP-targeted Ligands

Detailed information for the synthesis, purification, and characterizations of AV01084 and AV01088, and their nonradioactive Ga-complexed standards and ⁶⁸Ga-labeled analogs are provided in the Supplementary Materials (Tables S1-S3).

4.2. Cell Culture

The LNCaP cells obtained from ATCC (via Cedarlane, Burlington, Canada) were cultured in RPMI 1640 medium supplemented with 10% FBS, penicillin (100 U/mL) and streptomycin (100 μg/mL) at 37 °C in a Panasonic Healthcare (Tokyo, Japan) MCO-19AIC humidified incubator containing 5% CO₂. The HEK293T:hFAP cells [28] were cultured in DMEM GlutaMAX™ medium supplemented with 10% FBS, penicillin (100 U/mL) and streptomycin (100 μg/mL). Cells were grown until 80-90% confluence and washed with sterile phosphate-buffered saline (PBS, pH 7.4) and collected after 1 min trypsinization. The cell concentration was counted in triplicate using a hemocytometer and a manual laboratory counter.

4.3. LogD_7.4 Measurement

The lipophilicity characteristics of the ⁶⁸Ga-labeled tracers were determined by calculating the logarithm of the distribution coefficient (logD) in n-octanol/phosphate-buffered solution (PBS) of pH 7.40. Purified ⁶⁸Ga-labeled tracer (50 µL) was added into a test tube containing 3 mL n-octanol and 3 mL PBS. The mixture was vortexed for 1 min, followed by centrifugation for 15 min at 3,000 rpm. The two layers (1 mL each) were then collected separately and the radioactivity was counted using a Perkin Elmer (Waltham, MA, USA) Wizard2 2480 automatic gamma counter. After adjusting the counts to background and calculating the ratio of the activity of the organic to that of the aqueous phase, the logD_7.4 values were calculated.

4.4. In vitro PSMA Competition Binding Assay

The PSMA binding assays were conducted following previously published procedures using LNCaP cells and [¹⁸F]DCFPyL as the radioligand [28,33,36]. Data analyses of K_i were performed using the nonlinear regression algorithm of GraphPad Prism 7 (San Diego, CA) software.

4.5. In vitro FAP Fluorescence Assay

The half-maximal inhibitory concentration (IC₅₀) values of the tested ligands for FAP were measured by in vitro enzymatic assay following previously published procedures [28,31]. The recombinant human FAP (Biolegend; 0.2 µg/mL, 50 µL) was added into costar 96-well plate. PBS and varied concentrations (0.2 pM to 2 µM) of tested nonradioactive Ga-complexed standards were added to each wells (in duplicate) containing the recombinant human FAP. After being incubated for 30 min at 37 °C, 50 µL of Suc-Gly-Pro-AMC (2 µM, Bachem) was added to each well. The fluorescent signals were acquired at 15, 30, 45, and 60 min using FlexStation 3 Multi-Mode Microplate Reader with excitation at 380 nm and emission at 460 nm. The IC₅₀(FAP) was calculated using “nonlinear fit model” built-in model in GraphPad Prism 7.02 software.

4.6. Ex vivo Biodistribution and PET/CT Imaging Studies

Imaging and biodistribution studies were performed using male NOD.Cg-Rag1tm1Mom Il2rgtm1Wjl/SzJ (NRG) mice following previously published procedures [37,38,39]. The experiments were conducted according to the guidelines established by the Canadian Council on Animal Care and approved by Animal Ethics Committee of the University of British Columbia. The mice were briefly sedated by inhalation of 2.5% isoflurane in oxygen, and 100 µL LNCaP (2×10⁵ cells) or HEK293T:hFAP (8.5×10⁶ cells) cells were inoculated subcutaneously behind the left shoulder. When the tumor grew to 5 – 8 mm in diameter over 3 – 4 weeks and 4 – 5 weeks for HEK293T:hFAP and LNCaP tumors, respectively, the mice were used for PET/CT imaging and biodistribution studies.

PET/CT imaging experiments were carried out using a Siemens (Knoxville, TN, USA) Inveon micro PET/CT scanner. Each tumor-bearing mouse was injected with ~4 – 6 MBq of ⁶⁸Ga-labeled tracer through a lateral caudal tail vein under 2.5% isoflurane in oxygen anesthesia, followed by recovery and roaming freely in its cage during the uptake period. At 50 min post-injection, a 10-min CT scan was conducted first for localization and attenuation correction after segmentation for reconstructing the PET images, followed by a 10-min static PET imaging acquisition.

For biodistribution studies, the mice were injected with the radiotracer (~2 – 4 MBq) as described above. For blocking studies, PSMA tumor-bearing mice were co-injected with 2-PMPA (500 μg/mouse) and HEK293T:hFAP tumor-bearing mice were co-injected with FAPI-04 (250 μg/mouse). At 1 h post-injection, the mice were euthanized by CO₂ inhalation. Blood was withdrawn by cardiac puncture, and organs/tissues of interest were collected, weighed and counted using a Perkin Elmer Wizard2 2480 automatic gamma counter.

4.7. Statistical Analysis

Data were analyzed with the GraphPad Prism, version 7.02 and Microsoft (Redmond, WA, USA) Excel software. One way ANOVA and multiple t tests were performed for all organs in the biodistribution studies of [⁶⁸Ga]Ga-AV01084 and [⁶⁸Ga]Ga-AV01088 in LNCaP and HEK293T:hFAP tumor models. A statistically significant difference was considered present when the adjusted P value was less than 0.05 using the Holm–Sidak method.

5. Conclusions

Two novel ⁶⁸Ga-labeled PSMA/FAP bispecific tracers were designed, synthesized, and confirmed to have the ability to bind both PSMA and FAP in vitro and in vivo. Compared with the monospecific tracers, the bispecific tracers have decreased binding affinities towards PSMA, but retain comparable binding affinities towards FAP. Compared with three of our previously reported PSMA/FAP bispecific tracers ([⁶⁸Ga]Ga-AV01017, [⁶⁸Ga]Ga-AV01030, and [⁶⁸Ga]Ga-AV01038), both [⁶⁸Ga]Ga-AV01084 and Ga-AV01088 have better PSMA binding affinity, and improved tumor uptake in PSMA-expressing xenografts and tumor-to-background (blood, muscle, and bone) contrast ratios. Further optimization on the selection of linkers should be explored to improve the binding affinity, pharmacokinetics, and tumor uptake of the PSMA/FAP bispecific tracers.