A 99mTc-Labelled Tetrazine for Bioorthogonal Chemistry. Synthesis and Biodistribution Studies with Small Molecule trans-Cyclooctene Derivatives

A convenient strategy to radiolabel a hydrazinonicotonic acid (HYNIC)-derived tetrazine with 99mTc was developed, and its utility for creating probes to image bone metabolism and bacterial infection using both active and pretargeting strategies was demonstrated. The 99mTc-labelled HYNIC-tetrazine was synthesized in 75% yield and exhibited high stability in vitro and in vivo. A trans-cyclooctene (TCO)-labelled bisphosphonate (TCO-BP) that binds to regions of active calcium metabolism was used to evaluate the utility of the labelled tetrazine for bioorthogonal chemistry. The pretargeting approach, with 99mTc-HYNIC-tetrazine administered to mice one hour after TCO-BP, showed significant uptake of radioactivity in regions of active bone metabolism (knees and shoulders) at 6 hours post-injection. For comparison, TCO-BP was reacted with 99mTc-HYNIC-tetrazine before injection and this active targeting also showed high specific uptake in the knees and shoulders, whereas control 99mTc-HYNIC-tetrazine alone did not. A TCO-vancomycin derivative was similarly employed for targeting Staphylococcus aureus infection in vitro and in vivo. Pretargeting and active targeting strategies showed 2.5- and 3-fold uptake, respectively, at the sites of a calf-muscle infection in a murine model, compared to the contralateral control muscle. These results demonstrate the utility of the 99mTc-HYNIC-tetrazine for preparing new technetium radiopharmaceuticals, including those based on small molecule targeting constructs containing TCO, using either active or pretargeting strategies.


Introduction
Recent developments in bioorthogonal coupling chemistry offer a new way to create nuclear molecular imaging and therapy agents by allowing the targeting vector and its radiolabelled coupling partner to be administered separately, in contrast to the traditional method of administering a single molecule containing both the vector and the radiolabel [1]. Pretargeting allows time for the vector to accumulate at the site of interest and clear from non-target organs before injection of a radiolabelled coupling partner that reacts selectively with the targeting vector [2,3]. This approach creates the opportunity to use short half-life isotopes with macromolecular targeting vectors such as antibodies, which typically circulate for long periods of time and clear slowly [4][5][6]. The approach is also useful for radiometals and small molecule targeting constructs, where direct labelling is often not possible due to a lack of suitable metal-binding functional groups or where the addition of large radiometal-chelate complexes is detrimental to the ability of the conjugate to bind the target of interest.
The challenge of labelling small molecule targeting vectors with radiometals has been a particularly significant issue for targeted technetium radiopharmaceuticals. 99m Tc is the most widely used radionuclide in diagnostic medicine; however, the majority of agents used clinically are perfusion type radiopharmaceuticals [7][8][9]. There are a limited number of examples of targeted Tc agents, largely because of the challenge in mitigating the influence of Tc-chelate complexes on targeting constructs [10]. Pretargeting and bioorthogonal chemistry may offer a means of overcoming these barriers.
One particularly effective pretargeting method involves the inverse electron demand Diels-Alder (IEDDA) reaction between tetrazines and trans-cyclooctene (TCO) [11][12][13]. This system offers a number of advantages including biocompatibility, modest prosthetic group size, and rapid coupling kinetics. Herein we report the use of the IEDDA reaction with a 99m Tc-labelled tetrazine and two TCO-derived small molecules; one that targets bone, and a second that is designed to target sites of bacterial infection. These studies demonstrate the utility of the IEDDA reaction for developing new technetium radiopharmaceuticals with small molecule targeting constructs using both active targeting and pretargeting strategies.

General Procedures
Unless otherwise stated, all chemical reagents were purchased and used as received from Sigma-Aldrich, without further purification. Solvents were purchased from Caledon. 1 H and 13 C NMR spectra were recorded on Bruker AV 200 or AV 600 spectrometers. 1 H NMR chemical shifts are reported in ppm relative to the residual proton signal of the NMR solvents. Coupling constants (J) are reported in Hertz (Hz). 13 C NMR chemical shifts are reported in ppm relative to the carbon-13 signal of the NMR solvents. High resolution mass spectra (HRMS) were obtained on a Waters QToF Ultima Global spectrometer. 99m-Pertechnetate [ 99m TcO 4 ]was obtained in saline solution from a 99 Mo/ 99m Tc generator supplied by Lantheus Medical Imaging.
Reverse phase analytical HPLC was performed using a Waters 2489 HPLC equipped with a Waters 2489 UV/Vis (λ = 254 nm or 350 nm) and a Bioscan flow-count gamma detector (model 106). Spectra were recorded and processed on Empower 2 software (Waters). Semipreparative HPLC was performed using a C18, reverse-phase Phenomonex Gemini column (250 × 10 mm, 5 μm) with HPLC grade water (solvent A) and acetonitrile (solvent B) used as eluents. The flow rate was 4 mL/min using a gradient of 95% to 5%; solvent A over 30 min. Analytical HPLC was performed using a C18 reverse-phase Phenomonex column (25 × 4.60 mm, 5 micron) with HPLC grade water containing 0.1% trifluoroacetic acid (TFA) (Solvent A) and acetonitrile containing 0.1% TFA (Solvent B) as eluents. The flow rate was 1 mL/min using a gradient of 95% to 5% solvent A over 30 min. Kinetics studies were conducted on a Cary 50 UV/Vis spectrometer. Radioactivity for biodistribution studies was measured using a Perkin Elmer Wizard 1470 gamma counter.
Animal studies were approved by the Animal Research Ethics Board at McMaster University in accordance with Canadian Council on Animal Care (CCAC) guidelines. Balb/c and CD1 female mice (5 to 6 weeks old) were purchased from Charles River Laboratories and were maintained under Clean/Biohazard Level 2 conditions with 12 h light/dark cycles and given food and water ad libitum. Animals were housed for 1 week after arrival to acclimatize prior to initiation of any studies. No animals died during any of the biodistribution studies, prior to sacrifice at the endpoints. Mice were monitored at 4 h and 20 h post injection of bacteria. There was no indication of pain or distress at those times, although a few animals indicated discomfort in the infected leg at the time of sacrifice.

Reaction kinetics
Kinetic measurements were carried out in triplicate under pseudo-first order conditions using a fixed amount of tetrazine (10 −4 M) in methanol. The disappearance of the tetrazine absorption at 534 nm was measured at room temperature immediately upon the addition of (E)cyclooct-4-enol (TCO-OH). Studies were repeated using different concentrations of TCO-OH (10 mM, 15 mM and 20 mM) and the second order rate constant determined (S1 File).

Radiolabelling HYNIC-Tetrazine 3 with 99m Tc
Compound 2 (15 mg, 0.036 mmol) was dissolved in 50% v/v TFA:DCM and the mixture stirred over ice for 2 h before being concentrated to dryness. The resulting solid 3 was dissolved in 2:1 v/v DCM:MeOH and transferred to eppendorf tubes prior to evaporation under a stream of nitrogen. A sample of 3 (100 μg, 310 nmol) was subsequently dissolved in water (100 μL) prior to the addition of tricine (500 μL, 100 mg / mL in water), 500 μL 99m TcO 4 -(740 MBq) and 10 μL of stannous chloride dihydrate (3 mg / mL in EtOH). The solution was vortexed for 2 min and heated at 60˚C for 30 min, at which time the reaction was allowed to cool to RT and the desired product 4 isolated by semi-preparative HPLC. HPLC R t = 13 min; Radiochemical yield = 75%.

In vitro stability
Stability studies were performed following isolation of 4 by semi-preparative HPLC, where the product was first dried using a Biotage V10 evaporation system and the residue reconstituted in 0.9% saline (5.55 MBq / mL saline). Samples were incubated at 37˚C and then evaluated by analytical HPLC at various time points over 4 h (S2 File).

Biodistribution of 4
Biodistribution studies were performed using 4-8 week old female CD1 mice (Charles River Laboratories, Montreal, PQ). Mice were administered approximately 0.54 MBq of 4 (100 μL in 0.9% NaCl) by tail vein injection and sacrificed at t = 0.5, 1, 2 and 6 h post injection (n = 3 per time point). Decay correction was used to normalize activity measurements to time of dose preparation for data calculations. Following isolation of tissues and fluids, the mean percent injected dose per gram (%ID/g) and SEM for each sample were calculated.

Preparation of 5
Following isolation by HPLC and evaporation to dryness using a Biotage evaporator, compound 4 was formulated in 0.9% saline (0.37 MBq / 100 μL) containing TCO-BP (20 mg / kg) and incubated at room temperature for 15 min prior to administration.

Biodistribution of 4 and TCO-BP
For pretargeting, a 5 mg/mL solution of TCO-BP in saline (100 μL) was administered (20 mg / kg) by tail-vein injection to female Balb/c mice 1 h prior to injection of 0.6 MBq of 4 (100 μL). For studies involving the active targeting complex 5, a saline solution of the product also administered by tail vein injection. At 6 h post-injection of the labelled compounds, animals were anesthetized with 3% isoflurane and euthanized by cervical dislocation. Fluids, bone (knee, shoulder), and select tissues were collected, weighed, and counted in gamma counter. Decay correction was used to normalize activity measurements to time of dose preparation for data calculations. Data is expressed as %ID/g.

Synthesis of TCO-vancomycin 6
Preparation of the TCO-vancomycin derivative was performed following the procedure reported by Weissleder and coworkers [14].

Preparation of active targeting agent 7
Following isolation by HPLC and evaporation to dryness using a Biotage evaporator, compound 4 (0.37 MBq / 100 μL saline) was added to 1 μM 6 in phosphate-buffered saline (PBS) containing 1% fetal bovine serum (FBS). The mixture was left at room temperature for 15 min prior to initiation of biodistribution studies.

In vitro bacteria binding assay
An aliquot of Staphylococcus aureus (S. aureus, ATCC 25923) culture was added to 30 mL of lysogeny broth (LB) in a sterile 250 mL baffled flask. The culture was grown in a shaker incubator at 37˚C and 300 rpm for 16 to 24 h until bacteria reached log phase of growth. Bacterial suspensions were aliquoted into 1.5 mL microtubes and centrifuged for 2 min at 11,400 × g. The pellet was washed twice with LB and resuspended in LB to an OD 600 of 2.0. Each 1.5 mL sample was transferred to new 1.5 mL microtubes and centrifuged for 2 min at 11,400 × g. The supernatant was discarded and pellets resuspended in 1.5 mL LB containing either 20 μM 6 or 20 μM 6 + 200 μM vancomycin (each done in triplicate). Each sample was transferred into a separate 5 mL round-bottom tube and placed in a shaking incubator for 30 min, at 37˚C. Following pre-incubation with vancomycin-TCO, samples were washed twice with LB and resuspended in 1.5 mL of LB containing either 4 (7.4 Bq / μL) alone or the same amount of 4 in the presence of 200 μM vancomycin. Each sample was transferred to a new 5 mL round-bottom incubation tube and cultures were placed in a shaker incubator at 37˚C and 300 rpm. At 1 and 6 h following addition of 4, a 500 μL aliquot from each sample was taken and cells were pelleted by centrifuging for 2 min at 11,400 × g. A 480 μL aliquot of each supernatant was transferred to a separate plastic RIA test tube (Perkin Elmer, 1270-401). Pellets were washed once with PBS (500 μL each), and centrifuged as described above. Supernatant (500 μL) was collected and combined with supernatant previously collected for that pellet. Each pellet was resuspended in 400 μL of 5% SDS in water and incubated for 30 to 45 min at RT. A 300 μL aliquot of each pellet lysate was collected in a separate RIA test tube. Pellets and supernatants were then counted using a gamma counter. Data were analyzed and expressed as percent of 4 bound to S. aureus compared to total activity added. For all studies, mice were anesthetized with 3% isoflurane, euthanized by cervical dislocation, and tissue and fluid samples were collected at 1 h and 6 h post injection of the labeled compound (n = 3 per time point). The left (control) and right (infected) calf muscles and all other tissues and fluids were collected, weighed and counted using a gamma counter. Decay correction was used to normalize activity measurements to time of dose preparation and data reported as %ID/g.

Synthesis of the tetrazine ligand 2
There is only one report of a 99m Tc labelled tetrazine that was prepared using a hydrazinonicotinic acid (HYNIC) ligand [15]. We similarly employed a HYNIC ligand, but in place of a methyl substituted tetrazine we opted to use the more reactive and commercially available reagent (4-(1,2,4,5-tetrazin-3-yl)phenyl)methanamine (Fig 1). This particular tetrazine was selected because of its high reactivity towards TCO and proven effectiveness for in vivo IEDDA reactions [16]. The target ligand was prepared by coupling the commercial tetrazine with the Boc-protected HYNIC precursor 1 [17] in the presence of pyBOP in DMF. Isolation of the pink product 2 was achieved by silica gel chromatography in 65% yield.

Kinetic studies
The kinetics of the reaction between 2 and (E)-cyclooct-4-enol (TCO-OH) were followed spectrophotometrically by monitoring the decrease in absorbance at 534 nm as a function of time. The reaction was complete in seconds (S1 File), with a second order rate constant of 92 ± 6 M -1 s -1 in methanol. This was similar to that of the parent tetrazine (82 ± 2 M -1 s -1 in methanol) measured in head-to-head studies.

Radiolabelling with 99m Tc
Prior to labelling, the Boc groups were removed from 2 by treatment with TFA in DCM ( Fig  1). Labelling of 3 with 99m Tc involved treatment with 99m TcO 4 and SnCl 2 in the presence of a co-ligand (Fig 2) [18]. The co-ligand serves to occupy vacant coordination sites and can be easily varied, providing a convenient means of tuning the pharmacokinetics of the metal complex [19]. Tricine was selected here as the coligand as it has been shown to be highly effective at stabilizing Tc-HYNIC complexes and minimizing non-specific binding in vivo [20]. Optimizing the labelling conditions involved varying the reaction temperature, time, pH and the amount of 3. The reaction yield was particular sensitive to the amount of tetrazine ligand used where 100 μg was optimal. Ultimately, the desired product 4 was obtained in 75% radiochemical yield and 98% radiochemical purity (Fig 2) when the reaction was run at 60˚C for 30 min. Because the synthesis was performed directly from 99m TcO 4 -, which is the chemical form of technetium that is eluted from 99 Mo/ 99m Tc generators, the product can be produced easily, making it readily accessible and translatable for use in clinical studies.

In vitro stability and TCO reactivity
The stability of 4 in 0.9% saline was assessed over 4 h by HPLC and showed minimal decomposition (S2 File). The reactivity of the labeled tetrazine with TCO-OH was tested subsequently. γ-HPLC analysis showed rapid disappearance of the peak associated with 4 and appearance of multiple peaks (S3 File) associated with the formation of the tautomeric and diastereomeric isomers of dihydropyridazine and pyridazine [21].

Biodistribution of 4
The biodistribution of the 99m Tc-HYNIC-tetrazine 4 alone was evaluated in healthy CD1 mice at multiple time points. Compound 4 showed modest levels in the blood at 0.5 h (4.8 ± 0.1% injected dose (ID)/g) which ultimately decreased to 1.7 ± 0.3%ID/g at 6 h (Fig 3). Activity was observed in the stomach and thyroid, indicating the presence of pertechnetate. It is worth noting that the levels in these organs did not increase over 6 h, suggesting that the compound was not continually degraded during the timeframe of the study. The product showed significant uptake in the gall bladder that decreased significantly by 6 h post-injection. In addition, 4 was detected in multiple organs and tissues, which suggests that the labelled tetrazine should be available to react with TCO-conjugates in vivo. These results are consistent with the biodistribution of other HYNIC-tricine derivatives [22][23][24][25] and follows a similar trend to that recently reported for the analogous complex derived from a methylated tetrazine derivative [15,26].

Targeting bone using a TCO-bisphosphonate (TCO-BP) and 4
The next step was to test the ability of 4 to react with TCO derivatives in vivo. To this end a TCO-bisphosphonate derivative (TCO-BP) was employed. Bisphosphonates bind to regions of calcium accretion in bone and have been used as radiometal binding ligands and in routine clinical imaging of skeletal injuries, when labelled with 99m Tc [27][28][29]. Since it can target turnover in normal bone in mice, which is typically found in shoulder and knee joints, and is easily prepared in a single step, TCO-BP is a convenient reagent to evaluate new radiolabelled tetrazines in vivo. To this end, a solution of TCO-BP in saline was administered i.v. to Balb/c mice (20 mg/kg), 1 h prior to injection of 4. For comparison, an actively targeted construct was prepared by combining 4 with TCO-BP to produce the 99m Tc-HYNIC-tetrazine-TCO-BP complex 5, prior to administration (Fig 4). For both biodistribution studies, mice were euthanized and tissues collected 6 h post-injection. The radioactivity in different tissues and fluids was determined and % ID/g calculated.   For pretargeting (Fig 5), there was significant uptake in regions of active bone metabolism, mainly the knee (9.6 ± 0.7%ID/g) and shoulder (7.0 ± 1%ID/g). This was significantly higher than that for the control in which 4 was given alone (S4 File), where uptake was < 1%ID/g in bone at all time points. The actively targeted construct 5 also showed significant uptake in the knee (13.3 ± 0.4%ID/g) and shoulder (8.1 ± 0.5%ID/g). Compound 5 exhibited lower radioactivity concentrations in the blood and liver, which is likely due to the increased polarity of the bisphosphonate conjugate compared to the more hydrophobic parent tetrazine. The data clearly demonstrates that compound 4 can react with TCO derivatives in vivo, and that the IEDDA reaction is a convenient way to prepare actively targeted constructs.

Targeting infection using TCO-vancomycin
Having demonstrated the ability to localize 4 to bone using pretargeting and active targeting with TCO-BP, we next tested a TCO-vancomycin derivative as means of localizing 4 to sites of infection. There has been a longstanding search for Tc-radiopharmaceuticals for detecting bacterial infection [30,31]. One approach has been to radiolabel antibiotics directly with 99m Tc [30]. This met with limited success largely due to lack of stability of the metal complex in vivo and the detrimental influence coordination of the antibiotic to 99m Tc had on delivery and  binding to the infection. One approach to overcoming these issues is to use TCO-derivatized antibiotics, pretargeting and a 99m Tc-labelled tetrazine. As was mentioned, the same chemistry can be used to generate actively targeted radiopharmaceuticals. To these ends, TCO-vancomycin (6), which has been used to image bacteria in vitro [14], was employed as a model targeting molecule along with 4 to create new radiopharmaceuticals for imaging infection using both active and pretargeting strategies (Fig 6).
Initial in vitro studies were performed with S. aureus that was treated with 6 for 30 min in the presence or absence of vancomycin as a blocking agent. The amount of 6 used and incubation time were below that which would is known to have a bactericidal effect [14]. Following a washing step to remove unbound 6, compound 4 was added and the samples incubated at 37˚C for 1 and 6 h prior to washing. At both time points there was significantly more activity in the samples containing 6 alone compared to those containing 6 and excess vancomycin. At 6 h, the amount of 4 bound to S. aureus was 2.5-fold higher than the control containing the blocking agent (Fig 7). These results are consistent with prior reports, which showed that 6 binds specifically to vancomycin binding sites on bacteria, and that it maintains reactivity towards tetrazines when bound to bacteria [14]. In vivo testing of 4 to target sites of bacterial infection.
Compounds 4 and 6 were evaluated for their ability to target sites of infection in vivo using active and pretargeting approaches. Infections were created by administering S. aureus to the right calf muscle in Balb/c mice, with the non-injected left calf muscle as the contralateral control. Similar infection models has been used extensively to evaluate bacterial infection imaging agents [32][33][34][35][36][37]. For the active targeting study, 4 was combined with 6 ( Fig 7) and injected i.v. (tail vein), within 15 min. Tissues and fluids were collected at 1 and 6 h post-injection to determine the %ID/g (Fig 8). The infected calf muscle showed a 3-fold increase in radioactivity at 1 h compared to the control muscle. The radioactivity in the infected calf remained constant over 6 h post-injection (2.2 ± 0.4%ID/g at 1 h and 2.1 ± 0.6%ID/g at 6 h). The spleen also showed some accumulation of radioactivity that remained over the duration of the study (2.6 ± 0.4%ID/g at 1 h and 2.7 ± 0.8%ID/g at 6 h).
Results from pretargeting studies (Fig 9) showed a 2-fold higher radioactivity concentration localized to infected versus non-infected control muscle at 1 h (2.0 ± 0.1%ID/g versus 0.9 ± 0.1% ID/g, respectively), increasing to 2.5-fold higher localization at 6 h post injection (1.7 ± 0.1%ID/g and 0.7 ± 0.1%ID/g, respectively). As with the active targeting study, accumulation within the spleen occurred and remained consistent over the duration of the study (4.8 ± 0.4%ID/g at 1 h and 4.8 ± 0.7%ID/g at 6 h). The pretargeting studies show that the 99m Tc-labelled HYNIC derivative can localize preferentially to sites of infection.

Conclusions
A new Tc-labelled tetrazine was prepared and shown to react in vivo via the IEDDA reaction with two TCO-derived small molecules; one targeting regions of bone metabolism, and the second S. aureus infections. While pretargeting has the disadvantage of requiring administration of two compounds, it does offer a new approach for labelling small molecules with the most widely used radionuclide in diagnostic medicine. Furthermore, with access to 4, the high yielding and selective IEDDA reaction can also be used to prepare actively targeted Tc radiopharmaceuticals from small molecules that do not have the appropriate donors or arrangement of heteroatoms to bind the radiometal directly.