Improving Tumor-to-Background Contrast Through Hydrophilic Tetrazines: The Construction of ¹⁸F Labeled PET Agents Targeting Non-Small Cell Lung Carcinoma

Abstract

Bioorthogonal reactions have been widely used in biomedical field. ¹⁸F-TCO/Tetrazine ligation is the most reactive radiolabelled inverse electron demand Diels-Alder reactions, but its application had been limited due to modest contrast ratios of the resulting conjugates. Here we describe the use of hydrophilic tetrazines to improve tumor to background contrast of neurotensin receptor targeted PET agents. PET agents were constructed using a rapid Diels-Alder reaction of the radiolabeled trans-cyclooctene (¹⁸F-sTCO) with neurotensin (NT) conjugates of a 3,6-diaryltetrazine, 3-methyl-6-aryltetrazine, and a derivative of 3,6-di-(2-hydroxyethyl)tetrazine. Although cell binding assay demonstrated all agents have comparable binding affinity, the conjugate derived from 3,6-di-(2-hydroxyethyl)tetrazine demonstrated the highest tumor to muscle contrast, followed by conjugates of the 3-methyl-6-aryltetrazine and 3,6-diaryltetrazine.

Graphical Abstract

Hydrophilic tetrazines lead to improved tumor to background contrast in PET probe construction.

Due to the high selectivity and ability to proceed rapidly in biological context, bioorthogonal reactions have played a unique role in biomedical research including nuclear medicine.^[1] In particular, the inverse electron demand Diels-Alder (IEDDA) cycloadditions between tetrazines and strained dienophiles represents the fastest class of bioorthogonal reaction reported so far,^{[1b, 2]} and radiolabeled trans-cyclooctenes have emerged as particularly important dienophiles for constructing positron emission tomography (PET) tracers at low concentrations.^{[2f, 3]} In 2010, tetrazine-TCO ligation was first used for pretargeted radiochemistry applications by Robillard and co-workers.^[4] Our groups described the first ¹⁸F-labeled TCO,^[5] and Weissleder and co-workers described pretargeted ¹⁸F-imaging using tetrazine-TCO chemistry.^[6] Previously, we developed the conformationally strained trans-cyclooctenes sTCO and the more hydrophilic dTCO as dienophiles with especially fast kinetics in Diels-Alder reactions with tetrazines.^{[3b, e]} Thereafter, Kuntner and Mikula and coworkers designed an ¹⁸F-labeled tetrazine that rapidly combined with sTCO and other trans-cyclooctene and could be applied to conjugations in live animals.^[7] For applications in pretargeted imaging, Airaksinen and coworkers described a method for radiolabeling of tetrazines via 5-[18F]fluoro-5-deoxyribose,^[8] and an Al[¹⁸F]NOTA-labeled tetrazine radioligand was described by Lewis and co-workers^[9]. To retain the advantages associated with initially attaching the tetrazine component to the biomolecule,^[10] an ¹⁸F-labeled analog of sTCO was developed and shown to display rapid kinetics towards a PEGylated diphenyl-s-tetrazine analog (DiPhTz), which could be successfully used in PET probe construction (Scheme 1A).^[11] Although the resulting probe showed high tumor uptake, the background signal was high, resulting in modest imaging contrast. Subsequently, ¹⁸F-labeled probes based on dTCO were described by Bormans and used for pretargeted PET imaging.^[12] Despite great progress, there is still a need to develop new methods for probe construction that combine fast kinetics with improved tumor-to-background contrast.

Scheme 1.

(A) Tetrazine/trans-cylooctene based bioorthogonal reactions have been used for ¹⁸F-radiolabeling applications. (B) Hydrophilic tetrazines used in conjunction with ¹⁸F-sTCO rapidly provides probes with improved tumor-to-background in PET imaging applications.

More recently, we demonstrated that a more hydrophilic ¹⁸F-labeled trans-5-oxocene (¹⁸F-oTCO) in tetrazine ligation can be used to construct PET agents with improved tumor-to-background ratios in neurotensin receptor (NTR) imaging.^[13] Use of ¹⁸F-oTCO led to improvements over analogous probes constructed from either ¹⁸F-sTCO or ¹⁸F-dTCO. However, sTCO retains the advantages of a simpler synthesis and is ~30-fold more reactive than oTCO.^[14] Accordingly, we sought to explore the feasibility of improving tumor contrast by increasing the hydrophilicity of tetrazines while still taking advantage of the rapid kinetics enabled by sTCO. Here, we describe the use of a new hydrophilic tetrazine that can be prepared simply and constructed to ¹⁸F-based probes for in vivo imaging (Scheme 1B). Comparison is drawn to PET probes derived from 3-methyl-6-aryltetrazine and a 3,6-diphenyltetrazine derivative. We choose neurotensin/neurotensin receptor-1 system (NT/NTR1) as the model to evaluate in vivo imaging results of the new hydrophilic tetrazine derived tracers in NSCLC tumor model H1299 (Scheme 2).

Scheme 2.

Synthesis of a hydrophilic tetrazines 1 and 2 and NT-conjugates 3a-c. Construction of 4a, 4b and 4c.

Hydrophilic 3,6-di-(2-hydroxyethyl)tetrazine 1 was prepared in 33% yield from 2-cyanoethanol as displayed in Scheme 2.^[15] To improve the safety of the oxidation step, we developed a procedure based on the Knorr pyrazole synthesis for quenching hydrazine with 1,3-diacetylacetone prior to addition of PIDA oxidant. The diol was desymmetrized with p-nitrophenylchloroformate to give 2 in 48% yield. Conjugation with a PEG₁₁ amino acid, NHS ester formation and conjugation with NT-peptide gave tetrazine 3a. ¹⁸F-sTCO was synthesized according to our previously reported procedure,^[11] and conjugated with 3a to give ¹⁸F-NT conjugate 4a. Stopped-flow kinetics were used to evaluate the rate of tetrazine 1 with sTCO derivatives in PBS at 25 °C, with measured bimolecular rate constants k₂ 11,500 M^–1s^–1. The high reactivity of the new tetrazine towards sTCO retained its advantage of using it for the construction of PET-imaging probes.

Analogous procedures were used to prepare neurotensin conjugates of 3-methyl-6-aryltetrazine (3b) and diphenyltetrazine (3c) as well as their corresponding ¹⁸F-sTCO conjugates 4b and 4c. ¹⁸F-labeled neurotensin conjugates 4a, 4b and 4c were obtained in 41.1 ± 9.3%, 51.8 ± 5.9% and 42.7 ± 6.2% yield respectively, with > 98% radiochemical purity after HPLC separation (Fig 1). The final product identity was confirmed by comparing the radiolabeled agents with corresponding ¹⁹F standards (Fig S6). The logP values for 4a, 4b and 4c were −0.98 ± 0.08, −0.87 ± 0.09, and −0.69 ± 0.03, respectively. The HPLC retention time of 4a-4c correlated with these logP values, with 4a eluting most rapidly and 4c being most retained.

Figure 1.

Radio HPLC profile of purified of 4a (14.47 min), 4b (15.34 min) and 4c (16.78 min).

The receptor-binding affinity of the three NT analogues was compared using a competitive cell-binding assay (Fig 2).^[16] All 3 conjugates inhibited the binding of ¹²⁵I-NT (8–13) to H1299 cells in a dose-dependent manner. The IC₅₀ values of 4a, 4b, 4c and parent NT peptide were 78.6 ± 1.7 nM, 62.3 ± 3.4 nM, 73.0 ±1.5 nM and 16.7 ± 2.1 nM, respectively. These results demonstrate that the modification of the NTR-ligand slightly reduced the binding affinity to NTR1 but the binding affinities of the three conjugates were similar.

Figure 2.

Competitive cell binding assay of parent NT peptide, 4a, 4b and 4c.

Before testing these PET tracers in vivo, we first evaluated the NTR1 expression level in the NSCLC model H1299 by immunofluorescence both in cells and tumor tissues. As shown in Fig S7, H1299 cells incubated with fluorescent probe NT-FITC showed strong green fluorescence indicating the binding between NT-FITC and H1299 cells. An NT blocking group confirmed the target specificity. This result demonstrated that the NTR1 was positively expressed in H1299 cells. Intense immunofluorescence staining can also be seen in H1299 tumor tissues using 1:100 diluted anti-NTR1 primary antibodies. These data further confirmed that the NTR1 was highly expressed in H1299 tumor tissues and correlated well with the references.^[17] In fact, previous studies have suggested that NTR is upregulated in 59.7% lung cancer patients but not in normal lung tissues, making it a promising target for lung cancer management.^{[17b, 18]}

After confirming the positive expression of NTR1 in H1299, we further evaluated the in vivo behavior of the three PET tracers with static PET scans. As shown in Figure 3, both 4a and 4b showed prominent tumor uptake at 1 and 4 h post injection (p. i.). Diphenyltetrazine derived 4c showed a relatively high background signal but also visible tumor uptake. The quantitative analysis on PET images showed that the highest tumor uptake was observed in 4c with 2.51 ± 0.21 % ID/g and 2.24 ± 0.13 % ID/g at 1 and 4 hour p.i. The tumor uptake of more hydrophilic 4b was 2.42 ± 0.50 % ID/g and 1.76 ± 0.09 % ID/g at 1 and 4 hour p.i. while 4a showed tumor uptake of 1.90 ± 0.04 % ID/g and 1.45 ± 0.15 % ID/g at 1 and 4 h p.i., respectively. Although the more hydrophilic PET tracers 4a and 4b showed a modest reduction in tumor uptake, they exhibited higher tumor-to-background contrast with tumor to muscle ratio of 7.86 ± 2.45, 7.90 ± 0.79 and 6.11 ± 0.99, 6.35 ± 0.34 at 1h and 4h p. i., respectively (Table 1). On the contrary, the tumor to muscle ratio of 4c was 1.56 ± 0.40 and 2.87 ± 0.14 at 1 and 4 h p.i. (Table 1). Compound 4c showed long blood retention even at 4h p.i., which may be the cause of the higher tumor uptake and relatively high background signal. On the other hand, the more hydrophilic tracers 4a and 4b showed faster clearance rates from the blood thus leading to lower tumor uptake and better contrast. The highest uptake for all three tracers was observed in kidneys indicating a renal clearance pathway. The liver uptake was higher than tumor uptake for 4c but lower than tumor uptake for the more hydrophilic tracers 4a and 4b. We point out that probe stability could also play a role in tumor uptake and distribution. Because 4a-c bear the same targeting ligand and dihydropyridazine core, comparable stability is expected for these three agents. Nonetheless, more detailed metabolic profiles of these tracers need to be investigated in further studies. We also compared the uptake difference between 4a and 4b. The PET images showed higher tumor uptake of 4b than 4a at both time points, while the tumor to muscle ratio of 4a is 1.27-fold and 1.24-fold higher than that of 4b at 1 and 4 hours p.i, respectively. Thus, 4a demonstrated slightly decreased uptake but increased contrast compared with 4b.

Figure 3.

Representative PET/CT images of (a) 4a, (b) 4b and (c) 4c in H1299 tumor bearing mice at 1 and 4h post-injection (p.i.). The tumors are circled in orange. Tumor to liver and tumor to muscle ratio of (d) 4a, (e) 4b and (f) 4c at 1 and 4h p.i..

Table 1.

Tumor-to-muscle uptake ratio of 4a, 4b, and 4c in H1299 xenografts at 1 and 4h p.i.

Tumor/muscle	1h	4h
4a	7.86 ± 2.45	7.90 ± 0.79
4b	6.11 ± 0.99	6.35 ± 0.35
4c	1.56 ± 0.40	2.87 ± 0.14

In order to confirm the targeting specificity, a blocking study was performed in which 100 μg of unlabeled NT peptide was coinjected with the ¹⁸F-tracer 4b (Fig 4). The tumor uptake was reduced to 1.33 ± 0.05%ID/g at 1 h p.i. in the blocking study. The significant decrease of tumor uptake demonstrated the specificity of the tracer.

Figure 4.

Representative PET/CT images of H1299 tumor bearing mice (n=3) at 1h post-injection of 4b (a) without blocking dose and (b) with blocking dose. (c) Quantitative uptake of major organs derived from PET images.

In summary, NTR1-targeted PET tracers were prepared from three NTR1-conjugates of tetrazines of varied hydrophilicity. All three PET probes showed prominent tumor targeting properties with the more hydrophilic tetrazine derivatives giving the highest tumor-to-background contrast and best imaging quality. The study demonstrates that the tumor to background contrast of the tetrazine/TCO ligated probes by fine-tuning the hydrophilicity of tetrazine. The resulting PET tracers might be considered as promising candidates for NTR1 imaging in non-small cell lung carcinoma.

Experimental Section

Experimental details, synthetic procedures and analytical data are available in the supporting information.