Coordination Chemistry of [Y(pypa)]⁻ and Comparison Immuno-PET Imaging of [⁴⁴Sc]Sc- and [⁸⁶Y]Y-pypa-phenyl-TRC105

Abstract

Both scandium-44 and yttrium-86 are popular PET isotopes with appropriate half-lives for immuno-positron emission tomography (immuno-PET) imaging. Herein, a new bifunctional H₄pypa ligand, H₄pypa-phenyl-NCS, is synthesized, conjugated to a monoclonal antibody, TRC105, and labeled with both radionuclides to investigate the long-term in vivo stability of each complex. While the ⁴⁴Sc-labeled radiotracer exhibited promising pharmacokinetics and stability in 4T1-xenograft mice (n=3) even upon prolonged interactions with blood serum proteins, the progressive bone uptake of the ⁸⁶Y-counterpart indicated in vivo demetallation, obviating H₄pypa as a suitable chelator for Y³⁺ ion in vivo. The solution chemistry of [^natY(pypa)]⁻ was studied in detail and the complex found to be thermodynamically stable in solution with a pM value 22.0, ≥3 units higher than those of the analogous DOTA- and CHX-A”-DTPA-complexes; the ⁸⁶Y-result in vivo was therefore most unexpected. To explore further this in vivo lability, Density Functional Theory (DFT) calculation was performed to predict the geometry of [Y(pypa)]⁻ and the results were compared with those for the analogous Sc- and Lu-complexes; all three adopted the same coordination geometry (i.e. distorted capped square antiprism), but the metal-ligand bonds were much longer in [Y(pypa)]⁻ than in [Lu(pypa)]⁻ and [Sc(pypa)]⁻, which could indicate that the size of the binding cavity is too small for the Y³⁺ ion, but suitable for both the Lu³⁺ and Sc³⁺ ions. Considered along with results from [⁸⁶Y][Y(pypa-phenyl-TRC105)], it is noted that when matching chelators with radionuclides, chemical data such as the thermodynamic stability and in vitro inertness, albeit useful and necessary, do not always translate to in vivo inertness, especially with the prolonged blood circulation of the radiotracer bound to a monoclonal antibody. Although H₄pypa is a nonadentate chelator, which theoretically matches the coordination number of the Y³⁺ ion, we show herein that its binding cavity, in fact, favors smaller metal ions such as Sc³⁺ and Lu³⁺ and further exploitation of the Sc-pypa combination is desired.

Graphical Abstract

H4pypa was conjugated to an antibody via a newly synthesized H4pypa-phenyl-NCS; promising immuno-PET imaging with 44Sc was demonstrated

Introduction

Medical imaging is a vital part of drug discovery in oncology. Advancing at a rapid pace, Single-Photon Computerized Tomography (SPECT) and Positron-Emission Tomography (PET) provide information complementary to the anatomical images generated by X-ray Computed Tomography (CT), thereby allowing non-invasive imaging of the tumor microenvironment with excellent sensitivity.¹ Both rely on the detection of the photons emitted (SPECT, 100–250 keV) or indirectly generated (PET, 511 keV) during the radioactive decay,^2,3 but the latter generally has higher sensitivity and spatial resolution (< 2–3 mm vs. 6–8 mm).^4,5,6

Positron-emitting radionuclides are either non-metallic or metallic; the former are incorporated into the carrier via a covalent linkage,⁷ while the latter are coordinated to a bifunctional multidentate chelating ligand, or chelator, conjugated to a targeting molecule (i.e. antibody, antibody fragment, peptide, etc).^2,3,8 The combination of a radionuclide and a bio-vector requires an appropriate match of the radioactive and biological half-lives. The monoclonal antibody is a promising targeting vector with high binding selectivity and recognition of the cellular target, and thus, it is ideal for immuno-PET imaging when combined with an appropriately long-lived positron-emitting radionuclides.^9,10 Two attractive PET isotopes are scandium-44 (t_1/2 = 3.97 h) and yttrium-86 (t_1/2 = 14.7 h).^11,12 Scandium-44 has a very high positron-branching ratio (I_β+ = 94%, Eβ⁺_avg = 632 keV)^11,13 and can be produced by either a titanium-44/scandium-44 generator or cyclotron irradiation of a calcium-44 target (⁴⁴Ca(p,n)⁴⁴Sc).^14,15 The latter allows production in large quantity to address clinical needs. Furthermore, both preclinical and clinical studies have proposed scandium-44 as a better imaging surrogate for lutetium-177 (t_1/2 ~ 6.64 d).^13,16,17 Although lutetium-177 emits an imageable γ-ray, using it in low-dose dosimetry study can be unreliable as high therapeutic dose can follow different pharmacokinetics.¹⁸ Most importantly, the therapeutic isotopologue, scandium-47 (t_1/2 = 3.35 d, Eβ⁻_avg=162 keV), possesses highly favorable decay characteristics for therapy.¹⁹ These two constitute a chemically equivalent theranostic pair that permits direct translation of the imaging data to the radiation doses of the therapeutic version.^20,21,22 As for yttrium-86, the high energy β⁺-particles and γ-rays are the major concerns since it impacts the image resolution and complicates the logistics.²³ Also, the emission of a plethora of γ-rays increases the gamma-coincidences, and subsequently the quantitative errors, which are further amplified in tissues with high attenuation, such as bone, posing difficulties in detection of bone metastases.^20,24 Nonetheless, yttrium-86 is still useful in pre-therapy “scout” imaging for the extensively used therapeutic yttrium-90 (t_1/2 = 2.67 d, β⁻).^8,25,26

Free scandium-44 and yttrium-86 are rapidly taken up in bone and liver; therefore, a carefully tailored coordination chemistry approach for a chelator is required for specific delivery of the radiation dose to the tumor site.^27,28 An appropriate ligand should not only bind the radiometal ion with great thermodynamic stability, but also should result from a synthetically friendly scheme with high functional versatility because the chemical properties of the linker can significantly alter the properties of the whole radiotracer, and consequently the biodistribution. When a temperature- and pH-sensitive monoclonal antibody is involved, the radiolabeling conditions are limited to ~RT and pH = 6–7.²⁹ The tetracarboxylato-macrocyclic DOTA (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid, Chart 1) is incompatible with these conditions, albeit dominating the chelation for a myriad of metal ions, including those of scandium-44 and yttrium-86.^3,4,8 In this regard, chelators that combine fast radiolabeling kinetics at room temperature and high stability are desired. The recently reported nonadentate H₄pypa demonstrated promising coordination chemistry with scandium-44, with an advantage of RT-radiolabeling.³⁰ As for yttrium-86, although CHX-A”-DTPA (Chart 1) often replaces DOTA in radioimmunoconjugates due to more efficient radiolabeling at ambient temperature,^28,31,32 it was shown that the ⁸⁸Y-labeled p-nitrobenzyl derivatives of CHX-A”-DTPA steadily dissociated in serum over time, while yttrium-88 gradually accumulated in the cortical bone of femurs after the labeled antibody-conjugates was injected into mice.³³ Another human serum stability study with ⁹⁰Y-labeled CHX-A”-DTPA-Trastuzumab also proved transchelation of radioactivity with time.³¹ Others such as H₄octapa, the NETA- and the pyclen-based chelators (Chart 1) have each been investigated to a different extent.^{34,31,35,36,37,38,39,40} Despite a few promising findings, some of the chelators were derived from rather complicated synthetic routes with limited functional variations, hindering kg production and clinical translation.³⁷ Aside from that, it will be advantageous to employ the same chelator-conjugate for both imaging and therapeutic radionuclides since the resulting radiotracers are more probable to have similar chemical properties.

Chart 1.

Chemical structures of subject ligands.

Considering all these factors, the goal of this report is to explore and evaluate the possibility of employing H₄pypa in immuno-PET imaging using scandium-44 and/or yttrium-86. H₄pypa is a nonadentate picolinate-carboxylate chelator built around a central pyridine unit that connects the pendant arms, increases the overall complex rigidity, and provides a convenient locus for bifunctionalization. Its nine-coordinating binding cavity has demonstrated high affinity for a range of small- to medium-sized trivalent metal ions (e.g. Sc³⁺, In³⁺, Lu³⁺) at room temperature,^30,41 a property that provides versatility in switching treatment modality and complexing both diagnostic and therapeutic radionuclides such as scandium-44 and lutetium-177. Since H₄pypa possesses high affinity for Lu³⁺ at room temperature,⁴¹ while Y³⁺ and Lu³⁺ ions have rather similar physical properties and chelator selectivity,²⁸ it was anticipated that H₄pypa would be a suitable chelating agent for yttrium-86 as well; however, detailed investigations are still required. For radiopharmaceutical applications, complex properties such as the thermodynamic stability, stability in the presence of blood serum proteins, radiolabeling efficiency under mild conditions, as well as the biological stability and properties are important considerations. In order to conjugate H₄pypa to an antibody (TRC105, an anti-CD105 monoclonal antibody that inhibits angiogenesis and tumor growth),^42,43 a N-isothiocyanate (NCS)-bifunctionalized H₄pypa (H₄pypa-phenyl-NCS) was developed via a facile synthetic route and labeled with both scandium-44 and yttrium-86 after conjugation. Biodistribution and PET/CT imaging studies were performed to compare their pharmacokinetics. Prior to the animal studies, the basic coordination chemistry of the [Y(pypa)]⁻ complex anion was investigated, including the thermodynamic stability via a series of potentiometric and spectrophotometric titrations. Furthermore, density-functional theory (DFT) calculation predicted the complex geometry, while the radiolabeling efficiency and the stability in mouse serum were demonstrated with yttrium-86.

Results and Discussion

Synthesis and Characterization

A N-hydroxysuccinimide (NHS)-functionalized pypa (^tBu₄pypa-alkyl-NHS) has previously been reported for peptide conjugation on solid-phase with high coupling efficiency.⁴¹ However, because tert-butyl carboxylate groups are supposed to be hydrolyzed after bioconjugation using trifluoroacetic acid (TFA), it is not compatible with the planned antibody conjugation. Fortunately, the para-hydroxyl (p-OH) group in the central pyridine in compound 7 renders H₄pypafunctionally versatile. Theoretically, any spacer with a leaving group can be integrated through nucleophilic substitution with the p-OH moiety, engendering a much simpler linker-switching procedure because the bifunctional precursor (compound 7) can be produced in bulk and repeatedly used for attachment of different linkers, helpful for preliminary screening. Herein, with an intention to incorporate a reactive N-isothiocyanate (NCS) group into H₄pypa for antibody-coupling, commercially available N-​boc-​2-​(4-​aminophenyl)​ethanol was selected as the starting material of the spacer (Scheme 1), and tosylated with p-tosyl chloride in tetrahydrofuran (THF) and 6 M sodium hydroxide (NaOH) aqueous solution at room temperature (compound 8, 71%). The p-tosyl leaving group allowed the linker to be coupled to compound 7, reproduced following published protocol.⁴¹ The linker was used in 0.1–0.2 equivalent in excess, stirred with compound 7 and anhydrous potassium carbonate (K₂CO₃) (4 equiv) at room temperature for 24–48 h to ensure complete alkylation. Due to the instability of compound 9 on silica columns, isolation was not performed, and indeed, not required, if all compound 7 was alkylated because the excess linkers would not interfere the following reactions and could be removed during the HPLC purification in the later step. An important key for this reaction is that the chelator should be stirred with the base vigorously for about an hour prior to the addition of the linker. After that, the tert-butyl groups and the tert-butyloxycarbonyl (boc) group in compound 9 were simultaneously removed with TFA in dichloromethane (DCM) (1:1) at room temperature overnight to give H₄pypa-phenyl-NH₂ (compound 10, 50%), which was isolated using the reverse-phase high-performance liquid chromatography (HPLC) as a single peak at t_R = 17.8 min (A: H₂O/0.1% TFA B: acetonitrile (ACN)/0.1% TFA, 5–60% B over 40 min, 10 mL/min). The purified aniline compound was finally activated with thiophosgene (CSCl₂) in a vigorously stirred biphasic solution of hydrochloric acid (5%, aq)/glacial acetic acid (4:1) and chloroform (CHCl₃) at room temperature overnight. The resulting product, H₄pypa-phenyl-NCS (compound 11), was isolated with HPLC using the gradient above (t_R = 35.0 min), and then lyophilized to a fluffy white solid in 30% yield.

Scheme 1.

Reagents and conditions. i) SOCl₂, MeOH, RT-60 °C, 26 h, >99%; ii) BnBr, ACN, K₂CO₃, 60 °C, overnight, 64%; iii) NaBH₄, dry MeOH, RT, overnight, 82%; iv) PBr₃, dry CHCl₃/dry ACN, 60 °C, overnight, 70%; v) K₂CO₃, dry ACN, RT, 24 h, 73%; vi) Pd/C (10% w/w), H₂ (g), MeOH, RT, overnight; vii) K₂CO₃, dry THF, RT, 24–48 h, 90%; viii) p-TsCl, THF, 6 M NaOH, 71%; ix) TFA/DCM, overnight, 50%; x) CSCl₂, 1 M HCl/ glacial AcOH/CHCl₃, RT, 24 h, 30%

Complexation and Characterization

Non-bifunctional H₄pypa was synthesized following the published protocol,⁴¹ and then complexed with yttrium(III) perchlorate hexahydrate (Y(ClO₄)₃·6 H₂O, 1.1 equiv) in water (pH ~ 7) at room temperature for 1 h to form [Y(pypa)]⁻. The complex was characterized by different nuclear magnetic resonance (NMR) spectroscopic techniques at ambient temperature (Figures 1 and S19–S22) and high-resolution electrospray-ionization mass spectrometry (HR-ESI-MS) (Figure S23). Characterizations of the [Sc(pypa)]⁻ complex were reported previously and two isomeric species were identified in a temperature-dependent conformational equilibrium.³⁰ As the temperature increased from 25˚C to 85˚C, the equilibrium position further shifted towards the major isomer with sharper ¹H NMR peaks.³⁰ This geometric isomerism, however, was not observed with [Y(pypa)]⁻ which formed a single species in neutral aqueous solution without observable fluxionality on the NMR timescale, indicated by the sharp and well-defined ¹H NMR signals (Figure S19). Furthermore, the [Y(pypa)]⁻ complex appeared to have 2-fold rotational symmetry about the central pyridine, evidenced by five chemically distinct aromatic ¹H signals in a pattern concordant with the symmetric ligand (two triplets: δ 8.23 (H_h) and δ 7.87 (H_f) ppm; three doublets: δ 8.08 (H_g), 7.80 (H_e), and 7.45 (H_d) ppm). Also, there were no additional ¹³C NMR peaks compared to the spectrum of H₄pypa (Figure S20). The sharp diastereotopic coupling peaks in the aliphatic region confirmed the metal coordination (Figure S19) and the correlations between the diastereotopic ¹H were realized by COSY experiments (Figure S22), showing ²J – couplings between H_a and H_a’, H_b and H_b’, as well as between the two peaks merged in H_c. H_f (t, 1H) represented the para-H atom in the central pyridyl ring and underwent ³J – coupling with the ortho-H_d. Similarly, the para-H_h atoms (t, 2H) in two chemically equivalent picolinate arms showed the expected ³J – correlations with both neighboring ortho-H_g and -H_e atoms. H_b and H_c signals were assigned to the methylene-H atoms on the pyridyl backbone and both picolinate arms, respectively, confirmed by the intense ⁴J – coupling with the neighboring ortho-H in the pyridines. The most upfield methylene-H_a(a’) signals were assigned to the acetic arms.

Figure 1.

The chemical structure of [Y(pypa)]⁻ with ¹H NMR assignments - reference to Figure 2.

Comparing the ¹H NMR spectra of [M(pypa)]⁻ M = Sc, Lu and Y (Figure 2), the coordination environment of the Y³⁺-complex was more similar to that of Lu³⁺; both exhibited 2-fold rotational symmetry about the central pyridine (Figure 2).⁴¹ The major difference between the two was the diastereotopic behavior of the ¹H signals at δ ~ 4.66 ppm which was significantly weaker in [Lu(pypa)]⁻ than in [Y(pypa)]⁻. Nonetheless, both Lu³⁺- and Y³⁺-pypa complexes were vastly different from the Sc³⁺ analogue, which existed in solution as two asymmetric isomers with broader ¹H NMR signals at room temperature. Even at 85˚C, the [Sc(pypa)]⁻ complex was not isomerically pure. These observations could be due to the size similarity between the Lu³⁺ and Y³⁺ ions, while both are larger than the Sc³⁺ ion (Sc³⁺ (0.87 Å) < Lu³⁺ (0.98 Å) < Y³⁺ (1.02 Å), 8-coordinated).⁴⁴ Regarding the isomerism of the [Sc(pypa)]⁻ complex, it is noteworthy that the previous biodistribution studies using the isomeric mixture of the ⁴⁴Sc-labeled H₄pypa-peptidomimetic-conjugate targeting prostate-specific membrane antigen demonstrated high overall in vivo stability (the isomers could not be separated).³⁰ However, since the biological properties are also heavily dependent on the bioconjugation method and the conjugated targeting vector, it is essential to evaluate the pharmacokinetics of each bioconjugate, even if the same radiometal-chelate is used.

Figure 2.

¹H NMR spectra (400 MHz, D₂O) of H₄pypa (RT), [Y(pypa)]⁻ (RT), [Lu(pypa)]⁻ (RT) and [Sc(pypa)]⁻ (RT & 85˚C).

DFT Calculations

The geometry of the [Y(pypa)]⁻ anion (Figure 3A) was calculated by DFT and compared to those of [Sc(pypa)]⁻ and [Lu(pypa)]⁻ (Figure S25). The calculated bond lengths suggested nine-coordination for all three complexes, consistent with the ¹H NMR spectra where diastereotopic splitting of the methylene-H atoms were observed (Figure 2). All three pypa-complexes shared the same stable geometry of distorted capped square antiprism (Figure 3B). As reported previously, the smaller Sc³⁺ ion also complexed with H₄pypa in another conformation ~22.4 kJ/mol less stable than the distorted capped square antiprism.³⁰ The small energy difference allowed both [Sc(pypa)]⁻ isomers to co-exist in solution and interconvert in a temperature-dependent manner, as confirmed in the ¹H NMR spectra (Figure 2).³⁰ The average bond lengths of the less stable [Sc(pypa)]⁻ geometry were expectedly longer than the more stable one.³⁰ In order to directly compare the structural information of [M(pypa)]⁻ (M = Sc, Lu and Y), the less stable [Sc(pypa)]⁻ conformer was not considered due to its completely different geometry, and only the bond lengths (Table 1) and the X-M-Y bond angles (Table 2) of the shared geometry were concerned. As shown in Table 1, the bond distances between the metal ion and the donor atoms increased in tandem with the ionic radii (i.e. Sc³⁺ < Lu³⁺ < Y³⁺), indicating weaker coordination bonds in the [Y(pypa)]⁻ complex among all,⁴⁴ while the X-M-Y bond angles were smaller for the bigger metal ions (Table 2). The observations were consistent with the trend reported by Bazargan et al. based on the Cambridge Structural Database.⁴⁵ The flexibility and the relationship between the bond length and the bond angle allow the binding cavity of the ligand to match itself with different metal ions.⁴⁵

Figure 3. (A).

DFT calculated geometry for the [Y(pypa)]⁻ anion. (Geometries calculated for the [Sc(pypa)]⁻ and [Lu(pypa)]⁻ anions are in Figure S25). (B) [Y(pypa)]⁻ coordination environment showing only Y³⁺ and the donor atoms (A perspective looking through the central pyridine).

Table 1.

DFT-calculated metal-donor bond lengths for the [M(pypa)]⁻ (M = Sc, Lu, Y) anions.

Donor atom		Bond length (Å)
		[Sc(pypa)]−	[Lu(pypa)]−	[Y(pypa)]−
Backbone pyr-N	N1	2.4960	2.5150	2.5405
Backbone-N	N2	2.5661	2.6058	2.6354
Backbone-N	N3	2.5647	2.6057	2.6353
Picolinate-N	N4	2.3495	2.4495	2.4803
Picolinate-N	N5	2.3500	2.4501	2.4793
Acetate-COO	O1	2.2115	2.3514	2.3723
Acetate-COO	O2	2.2149	2.3494	2.3753
Picolinate-COO	O3	2.1872	2.3341	2.3647
Picolinate-COO	O4	2.1925	2.3375	2.3618

Table 2.

DFT-calculated X-M-Y bond angles for the [M(pypa)]⁻ (M = Sc, Lu, Y) anions.

X	Y	M
		Sc3+	Lu3+	Y3+
Backbone pyr-N	Backbone-N	65.89˚	65.49˚	64.76˚
Backbone pyr-N	Backbone-N	65.93˚	65.36˚	64.93˚
Picolinate-N	Backbone-N	67.74˚	66.69˚	66.02˚
Picolinate-N	Backbone-N	67.67˚	66.59˚	66.19˚
Backbone-N	Acetate-COO	70.83˚	68.99˚	68.28˚
Backbone-N	Acetate-COO	70.75˚	69.03˚	68.17˚
Picolinate-N	Picolinate-COO	70.38˚	67.55˚	66.73˚
Picolinate-N	Picolinate-COO	70.30˚	67.42˚	66.82˚

Based on that, the ideal M-N bond length is ~2.5 Å and the ideal N-M-N bond angle is ~69˚ in a five-membered NCCNM chelate ring, which is preferred by the larger metal ions (i.e. ionic radius ~1 Å),^46,47,48 the coordination bonds between the metal ion and the N atoms of two tertiary amines are slightly more favorable in [Sc(pypa)]⁻ (~2.56 Å) compared to [Y(pypa)]⁻ (~2.64 Å) (Table 1), so as the N-M-N bond angles (Table 2). Despite the differences, all three metal ions demonstrated strong preferences for the oxygen donor atoms (i.e. significantly shorter M-O bonds) due to their hard nature.

Solution Thermodynamics

The stability constant of a metal complex is an important thermodynamic parameter, despite the lack of predictability of the kinetic inertness. Because the metal complexation reaction always occurs in competition with the protonation equilibria of the ligand, knowledge of the protonation constants of the ligand is mandatory to determine the stability constant of its metal complexes. H₄pypa was previously determined to possess nine protonation sites, the most basic of which deprotonated with a pKa = 7.78 (protonated tertiary amines in the pyridyl backbone), suggesting that at physiological pH (~7.4), almost all the donor atoms are free for chelation. Complex formation equilibria with more basic chelators, such as DOTA (Chart 1, highest pKa = 12.60⁴⁹) and CHX-A”-DTPA (Chart 1, 12.30⁵⁰) will occur with higher competition between the protons and the basic sites on the ligand. In this light, a more comprehensive parameter that allows for a better comparison of the affinity of different chelators for a metal ion is pM. The pM value is widely adopted for this purpose and is defined under physiologically relevant conditions (−log [M]_free at [L] = 10 μM, [M] = 1 μM and pH = 7.4). It accounts not only for the stability of the metal complexes, but also for the ligand basicity (pKa values) and denticity, and can provide a more accurate prediction of the in vivo thermodynamic stability and speciation of metal-scavenging at physiological pH = 7.4.⁵¹

To investigate the complexation equilibria of the Y³⁺-H₄pypa system, a series of titration studies were conducted. Since the complexation between the Y³⁺ ion and H₄pypa was complete at pH ≤ 2, below the electrode threshold (Figure S26), direct determination of the thermodynamic stability constant of [Y(pypa)]⁻ (log K_Y(pypa)) was not feasible. Instead, a ligand-ligand competition method with EDTA as competitor (Chart 1) and acidic in-batch UV spectrophotometric titrations (Figure S27) was applied to determine the formation constant of the protonated species ([Y(Hpypa)]) prior to those of [Y(pypa)]⁻ and [Y₂(OH)(pypa)₂]^3- using direct potentiometric titration. Potentiometric and spectrophotometric experimental data were refined using the Hyperquad2013⁵² and HypSpec2014⁵³ programs. The thermodynamic stability constant of [Y(pypa)]⁻ (log K_[Y(pypa)]⁻) was finally calculated to be 21.60(1) (0.16 M NaCl, 25˚C) (Table 3). Although it is 2.49 and 3.10 units lower than those of the DOTA- and CHX-A”-DTPA- complexes, respectively,^34,50 the pY value (22.0), a more accurate assessment, as described above, was ≥3 units higher.^34,50 The difference was partially a result of the drastically lower basicity of H₄pypa. When comparing to other M³⁺-H₄pypa systems (M³⁺ = Sc³⁺, In³⁺, Lu³⁺ and La³⁺),^30,41 pY value was consistent with the inverse correlation between the size of the metal ion and the metal sequestering ability of H₄pypa (except the Sc³⁺-complex).^30,41 With similar ionic radii (Y³⁺ = 1.075 Å vs. Lu³⁺ = 1.032 Å, 9-coordinated),⁴⁴ pY is also very close to pLu, implying similar thermodynamic stability in both systems (Figure 4). Another interesting observation is that with larger metal ions such as Y³⁺ and La³⁺, H₄pypa tends to form a 2:2 metal:ligand-hydroxide species under basic conditions, rather than the 1:1 complex as seen in the systems with the Sc³⁺, In³⁺ and Lu³⁺ ions (Table 3).

Table 3.

Stepwise stability constants (log K) of H₄pypa with Sc³⁺, Lu³⁺, Y³⁺ and La³⁺ ions.

Equilibrium reaction	Sc3+	Lu3+	Y3+	La3+
M3+ + L ⇆ ML	26.98(1)30	22.20(2)41	21.60(1) (a)	19.74(3)41
ML + H+ ⇆ MHL	3.55(2)30	3.60(6)41	2.14(5)(b) 2.29(c)	3.24(5)41
M(OH)L + H+ ⇆ ML	10.34(3)30	10.77(8)41	-	-
M2L2(OH) + H+ ⇆ M2L2	-	-	37.23(a)	34.40(6)41
pM(d)	27.130	22.641	22.0	19.941

^a)from direct UV-potentiometric titration method at 25 °C, I = 0.16 M NaCl;

^b)from EDTA potentiometric competition titration method at 25 °C, I = 0.16 M NaCl;

^c)from direct UV-acidic competition titration method at 25 °C, I ≠ 0.16 M NaCl;

^d)pM is defined as −log [M]_free at [L] = 10 μM, [M] = 1 μM and pH = 7.4.

Figure 4.

pM vs. ionic radius⁴⁴ for M³⁺ and ligands of interest.

Radiolabeling of [⁸⁶Y][Y(pypa)]⁻ and Mouse Serum Challenge Experiments

Because the radiotracer is usually formulated with very low chelator concentrations (<10⁻⁵ M), it is important to ascertain the affinity of a chelator for a radiometal ion under dilution. In this regard, one common useful study is the concentration-dependent radiolabeling experiment in which a constant amount of radioactivity is added to a series of diluted chelator solutions and then incubated for a desired period. With an intention to use H₄pypa in immuno-PET imaging, the labeling experiment of H₄pypa with yttrium-86 was performed under antibody-compatible conditions (pH = 7 and ambient temperature) using 0.2 M HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) buffer and ~4.1 MBq of yttrium-86. The mixtures were agitated at room temperature for 7, 15, 30 and 60 min, and then the radiochemical yield percentage (RCY%) of [⁸⁶Y][Y(pypa)]⁻ was determined with silica aluminum-backed thin layer chromatography (TLC) plates, developed in diethylenetriamine-pentaacetic acid (DTPA) buffer (0.1 M, pH = 5.5). Attempts to identify the complex by the radioactive high-performance liquid chromatography (HPLC) failed as [⁸⁶Y][Y(pypa)]⁻ decomplexed in the HPLC column, probably due to the fast acid-assisted dissociation kinetics at low concentrations of the metal complex in the acidic mobile phases (ACN/H₂O/0.1% TFA solution, pH ~ 2) (Figure S28). Regarding the radio-metalation, the efficiency was concentration-dependent (96±1% at 10⁻⁵ M vs. 88±3 at 10⁻⁶ M after 15 min) (Figure 5A and Table S1) and that at the lower concentrations (i.e. 10⁻⁶ M and 10⁻⁷ M) gradually improved with time (83±2%, 7 min vs. 93±1, 60 min at 10⁻⁶ M; 15±2%, 7 min vs. 26±2, 60 min at 10⁻⁷ M) (Table S1). Furthermore, the apparent molar activity of the [⁸⁶Y][Y(pypa)]⁻ complex (the radioactivity was corrected to the end-of-bombardment (EoB)) significantly increased with time, from 79 GBq/μmol at 7 min to 105 GBq/μmol at 60 min, meaning that more radioactivity could be used to achieve quantitative RCY if the complexation time is longer. The efficient radiolabeling at room temperature is an advantage of non-macrocyclic/acyclic chelator and was also seen with CHX-A”-DTPA. A radiolabeling study of yttrium-90 and CHX-A”-DTPA-DUPA-Pep (a peptide conjugate) conducted by Benjamin et al. obtained ~98% RCY in 5 min at room temperature (25 μg of peptide, pH = 5.5 and 3.7–3.8 MBq).⁵⁴ On the other hand, quantitative radiolabeling with DOTATOC in 10 min required heating at 100˚C (pH = 4),⁵⁵ which is not compatible with the monoclonal antibody, but an expected requirement for a constrained macrocycle. The radiolabeling results for [⁴⁴Sc][Sc(pypa)]⁻ were previously reported,³⁰ and similar to [⁸⁶Y][Y(pypa)]⁻, the radiochemical yields were both concentration- and time-dependent (95±0% at 10⁻⁵ M vs. 90±2 at 10⁻⁶ M after 5 min; 9±3%, 5 min vs. 21±1, 60 min at 10⁻⁷ M, pH = 5.5, 2.9 MBq). In spite of the similar radiolabeling behavior, only [⁴⁴Sc][Sc(pypa)]⁻ could be isolated in the acidic HPLC system as a single sharp peak at 13.4 min (same stationary and mobile phases as [⁸⁶Y][Y(pypa)]⁻ were used), implying much slower acid-assisted dissociation kinetics at pH ~ 2 than for the ⁸⁶Y-analog.³⁰ Nevertheless, the efficient radiolabeling of H₄pypa at room temperature with both radionuclides renders it a promising alternative to the “gold standard” chelator, DOTA, in building immuno-constructs.^55,56

Figure 5. (A).

Concentration-dependent radiolabeling studies of H₄pypa with ⁴⁴Sc in NH₄OAc buffer (0.5 M, pH=5.5) and ⁸⁶Y in HEPES buffer (0.5 M, pH=7) at room temperature over 15 min. (B) Mouse serum stability of [⁸⁶Y][Y(pypa)]⁻over 48 h at 37˚C.

After the successful radiolabeling results, it was deemed prudent to test the stability of the [⁸⁶Y][Y(pypa)]⁻ complex in mouse serum. In the study, the labeled complex was incubated in mouse serum at 37˚C over 48 h, and the intact percentage of the complex was determined by silica aluminum-backed TLC plates, developed in DTPA buffer (0.1 M, pH = 5.5). To validate the experiment, a control sample containing only yttrium-86 in the labeling buffer and mouse serum was prepared, showing only one radioactivity spot at the solvent front on the developed TLC plate, which confirmed that both the free and the transchelated yttrium-86 were moved by DTPA. The study showed that ~97% of [⁸⁶Y][Y(pypa)]⁻ remained intact after 2 days which is encouraging (Figure 5B and Table S2).

PET/CT Imaging and Biodistribution Studies

Following injection of ⁴⁴Sc- or ⁸⁶Y-labeled H₄pypa-phenyl-TRC105 (~18 and 11 MBq/mouse, respectively) into 4T1-xenograft-bearing mice (n = 3), serial PET/CT images (maximum intensity projection, MIP) were acquired at 30 min, 4 h, 8 h and 18 h post-injection (p.i.) for the ⁴⁴Sc-labeled tracer, as well as 30 min, 4 h, 24 h and 48 h p.i. for the ⁸⁶Y-counterpart (Figures 6–8). A higher dose of the ⁴⁴Sc-labeled radiotracer was injected in order to acquire the late-timepoint images. Radioactivity uptakes of selected organs were analyzed with quantitative regions-of-interest (ROI) imaging and plotted in Figure 7 in % injected-dose-per-gram (% ID/g). The ex vivo biodistribution results at the last time point are shown in Figure 8 and Table S3 which confirmed the accuracy of the image-based ROI quantification. Both radioimmunoconjugates showed expected uptakes in the blood, tumor, liver, kidney and spleen due to the antibody circulation, metabolism and excretion. High uptake in lung was seen in other reported TRC105-constructs, perhaps due to the prolonged residence in the blood pool.^57,58,59 The tumor was delineated shortly after 4 h post-injection (p.i.) with 10.2±1.6% ID/g and 8.1±0.6% ID/g for ⁴⁴Sc- and ⁸⁶Y-constructs, respectively. In the case of ⁴⁴Sc, the tumor uptake increased by ~6% ID/g from 4 to 18 h p.i. (16.1±1.9% ID/g), but only ~1% ID/g for [⁸⁶Y][Y(pypa-phenyl-TRC105)] between 4 and 24 h p.i, and then stabilized over the following 24 h (9.1±0.9% ID/g at 48 h p.i.). Additionally, for the ⁴⁴Sc-tracer, the radioactivity in blood decreased in tandem with that in the liver without significant change in the bone uptake throughout the course of study, indicating the long-term in vivo stability of the [⁴⁴Sc][Sc(pypa)]⁻ complex. Unfortunately, for the ⁸⁶Y-counterpart, despite the clearance in the blood pool, the bone uptake jumped by nearly 3-fold from 30 min (2.1±0.3% ID/g) to 48 h p.i. (5.7±0.5% ID/g), almost double compared to that of the [⁸⁶Y][Y(DTPA-TRC105)] (~3% ID/g at 48 h p.i.).⁵⁸ The liver activity also increased by ~2.9% ID/g within the first 4 h. Taken together, it suggests that [⁸⁶Y][Y(pypa-phenyl-TRC105)] is unstable in vivo. The in vivo transchelation of the ⁸⁶Y-construct could also explain the significantly lower tumor accumulation compared to that of the ⁴⁴Sc-counterpart. This outcome was not expected as [⁸⁶Y][Y(pypa)]⁻ appeared to be stable in mouse serum, with high thermodynamic stability comparable to the Lu³⁺-H₄pypa system which was stable in vivo.⁴¹ A possible explanation could be that the conjugation changed the coordination environment of the [⁸⁶Y][Y(pypa)]⁻ complex. Furthermore, since the radiopharmaceutical is usually injected in a very low concentration and is further diluted as it is distributed in the blood volume, the rate of dissociation of the complex, rather than the thermodynamic equilibrium, will dictate the in vivo stability.^3,60,61 This is further enhanced when there are orders-of-magnitude more concentrated endogenous ligands (e.g. transferrin) competing with the chelator-conjugate in binding the radiometal ions.³ Therefore, the disappointing pharmacokinetics of [⁸⁶Y][Y(pypa-phenyl-TRC105)] are likely a result of the insufficient kinetic inertness and metabolic stability of the complex. The result with [⁴⁴Sc][Sc(pypa-phenyl-TRC105)] was deemed a great success. The tumor was visualized at the highest contrast at 18 h p.i. (Figure 6A). The biodistribution pattern of [⁴⁴Sc][Sc(pypa-phenyl-TRC105)] closely resembled that of [⁸⁶Y][Y(DTPA-TRC105)] with uptake in blood, lung, liver, kidney, spleen and tumor, and no abnormal bone and liver retentions which manifest signs of in vivo stability (Figure 7A),⁵⁸ evidencing the aptness of H₄pypa in scavenging ⁴⁴Sc.

Figure 6.

PET/CT MIP Images of (A) [⁴⁴Sc][Sc(pypa-phenyl-TRC105)] (B) [⁸⁶Y][Y(pypa-phenyl-TRC105)] at different post-injection time points.

Figure 8.

Ex vivo biodistribution data of [⁴⁴Sc][Sc(pypa-phenyl-TRC105)] (18 h p.i.) and [⁸⁶Y][Y(pypa-phenyl-TRC105)] (48 h p.i.).

Figure 7.

Quantitative ROI analysis of the in vivo PET imaging data of (A) [⁴⁴Sc][Sc(pypa-phenyl-TRC105)] (B) [⁸⁶Y][Y(pypa-phenyl-TRC105)] at different post-injection time points.

Conclusion

The hypothesis of this work was to explore H₄pypa in immuno-PET imaging with scandium-44 and yttrium-86, which eventually could be useful for theranostic applications with long-lived therapeutic radionuclides such as lutetium-177. H₄pypa previously demonstrated favorable radiolabeling properties and complex stability with scandium-44 and lutetium-177.^30,41 Due to the similar chemical properties between Lu³⁺ and Y³⁺ ions,²⁸ H₄pypa was anticipated to hold promise for the Y³⁺ ion as well. The [Y(pypa)]⁻ complex presented a single symmetric species with the DFT-calculated bond lengths longer than both [Sc(pypa)]⁻ and [Lu(pypa)]⁻ complexes. Although the Y³⁺-H₄pypa system was thermodynamically favorable (pY value of 22.0, more than 3 units higher than those of the DOTA- and CHX-A”-DTPA-complexes^34,50), and the [⁸⁶Y][Y(pypa)]⁻ complex appeared to be stable in mouse serum as well, when [⁸⁶Y][Y(pypa-phenyl-TRC105)] was injected into a cohort of 4T1-xenograft mice (n = 3), progressive bone accumulation was observed due to the undesirable release of free yttrium-86 in vivo. On the other hand, the pharmacokinetics of [⁴⁴Sc][Sc(pypa-phenyl-TRC105)] were highly favorable with long-term in vivo stability. Although the previously reported [⁴⁴Sc][Sc(pypa-C7-PSMA617)] showed promising biological stability as well,³⁰ this finding is important because, firstly, H₄pypa-phenyl-NCS was a new bifunctional chelator which could lead to different metal coordination properties. In addition, antibody-conjugates possess much longer blood half-life (1–3 weeks) which prolonged the interactions between the radio-complex and the blood serum proteins, while the PSMA-conjugates tend to be excreted very quickly;⁶² therefore, the in vivo stability, particularly at late timepoints, would be more accurate when examined with an antibody-construct. The long-term in vivo stability of [⁴⁴Sc][Sc(pypa)]⁻ encourages future immuno-PET applications using a ⁴⁴Sc-labeled pypa-phenyl-antibody, especially in the dosimetry study for the long-lived therapeutic radionuclides such as scandium-47 and lutetium-177, which can be conveniently chelated with the same bioconjugate to attain more comparable biodistributions.

Experimental Section

Materials and Methods

All solvents and reagents were purchased from commercial suppliers (TCI America, Alfa Aesar, AK Scientific, Sigma-Aldrich, Fisher Scientific, Fluka) and were used as received. Deionized water was filtered through the PURELAB Ultra Mk2 system. ¹H, ¹³C{¹H}, ¹H-¹³C{¹H} HSQC and COSY NMR spectra were recorded at ambient temperature on Bruker AV400 instruments, as specified; the NMR spectra are expressed on the δ scale and were referenced to residual solvent peaks. Low-resolution (LR) mass spectrometry was performed using a Waters ZG spectrometer with an ESCI electrospray/chemical-ionization source, and high-resolution electrospray-ionization mass spectrometry (HR-ESI-MS) was performed on a Micromass LCT time-of-flight instrument at the Department of Chemistry, University of British Columbia. Analyses of radiolabeled compounds were performed with both thin layer chromatography (TLC) (i.e. silica aluminum-backed TLC plates and iTLC-impregnated with silica gel (iTLC-SG) strips) purchased from Agilent Technologies. The radio-TLC scanner model was the Cyclone Storage Phosphor System (Cyclone Plus) from Perkin Elmer and the HPLC system was Dionex Ultimate 3000. DIONEX Acclaim C18 5 μm 120 Å column (250 mm × 4.60 mm) was used for separation of free radioactivity and radio-complex. ⁴⁴Sc and ⁸⁶Y were provided by the UW-Madison Cyclotron Lab as 0.1 M HCl solutions.

Dimethyl 4-hydroxypyridine-2,6-dicarboxylate (1)

Thionyl chloride (SOCl₂) (9.50 mL, 0.130 mol, 5 equiv) was added slowly using a syringe to a stirred suspension of chelidamic acid monohydrate (5.28 g, 26.2 mmol, 1 equiv) in methanol (60 mL) in a two-neck round-bottom flask at 0°C. The mixture was stirred at room temperature for 24 h and then refluxed for an additional 2 h. The solvent was removed under reduced pressure gently at room temperature and then deionized water was added at 0˚C. The mixture was neutralized with aqueous 1 M K₂CO₃ solution and the precipitate was filtered by vacuum filtration, and then washed with 50% MeOH in water (~10 mL). The white precipitate was dried under reduced pressure to give a white solid (5.54 g, >99%). ¹H NMR (400 MHz, 298 K, (CD₃)₂SO): δ 6.74 (s, 2H), 3.72 (s, 6H). ¹³C NMR (100 MHz, 298 K, (CD₃)₂SO): δ 165.7, 149.2, 116.6, 52.7. LR-ESI-MS: calcd. for [C₉H₉NO₅ + Na]⁺ 234.0; found [M + Na]⁺ 234.2

Dimethyl 4-(benzyloxy)pyridine-2,6-dicarboxylate (2)

To a round-bottom flask with a stirred solution of compound 1 (1.65 g, 7.82 mmol, 1 equiv) in dry acetonitrile (ACN) was added anhydrous K₂CO₃ (2.19 g, 15.8 mmol, 2.02 equiv) and benzyl bromide (1.02 mL, 8.60 mmol, 1.1 equiv). The reaction mixture was refluxed overnight at 60˚C. K₂CO₃ was filtered out by vacuum filtration and then washed with DCM. The filtrate was concentrated in vacuo and then purified through a silica column (CombiFlash R_f automated column system, 24 g gold silica column, A: DCM B: MeOH, 0–5% B). The product fractions were rotary-evaporated to give a white powder (1.51 g, 64 %). ¹H NMR (400 MHz, 298 K, CDCl₃): δ 7.90 (s, 2H), 7.44–7.38 (m, 5H), 5.23 (s, 2H), 4.01 (s, 6H). ¹³C NMR (100 MHz, 298 K, CDCl₃): δ 150.0, 129.0, 128.9, 127.9, 115.0, 71.0, 53.4. LR-ESI-MS: calcd. for [C₁₆H₁₅NO₅ + Na]⁺ 324.1; found [M + Na]⁺ 324.1

(4-(Benzyloxy)pyridine-2,6-diyl)dimethanol (3)

To a round-bottom flask with a stirred solution of compound 2 (8.74 g, 29.0 mmol, 1 equiv) in dry MeOH (90 mL) was added sodium borohydride (NaBH₄) (3.29 g, 87.1 mmol, 3 equiv) in three portions over 30 min at 0˚C. The reaction mixture was stirred at room temperature. After 24 h, the mixture was diluted with CHCl₃ (50 mL) and then quenched with saturated aqueous NaHCO₃ solution (50 mL). The organic phase was separated and the bulk of MeOH in the aqueous layer was removed in vacuo to give an aqueous solution which was extracted with CHCl₃ (50 mL × 4). The combined organic phases were dried over anhydrous sodium sulfate (Na₂SO₄), and then clarified by filtration. The filtrate was rotary-evaporated to give a white solid (5.86 g, 82%).¹H NMR (400 MHz, 298 K, CDCl₃): δ 7.42–7.35 (m, 5H), 6.79 (s, 2H), 5.12 (s, 2H), 4.70 (s, 4H). ¹³C NMR (100 MHz, 298 K, CDCl₃): δ 184.4, 166.5, 162.7, 160.6, 149.6, 135.6, 128.9, 128.6, 127.6, 117.2, 111.8, 107.7, 106.5, 106.1, 105.2, 70.2, 64.5. LR-ESI-MS: calcd. for [C₁₄H₁₅NO₃ + Na]⁺ 268.1; found [M + Na]⁺ 268.2

4-(Benzyloxy)-2,6-bis(bromomethyl)pyridine (4)

Compound 3 (1.76 g, 12.6 mmol, 1 equiv) was suspended in dry ACN/CHCl₃ (40 mL, 50:50 v/v) in a three-neck round-bottom flask. Phosphorus tribromide (PBr₃) (3.60 mL, 37.9 mmol, 3 equiv) in CHCl₃ (5 mL) was added dropwise using a dropping funnel to the stirred solution of compound 3 at 0˚C over 15 min. The reaction mixture was stirred at 60˚C for 18 h and then saturated aqueous Na₂CO₃ solution was added slowly to quench the reaction at 0˚C. The aqueous phase was extracted with CHCl₃ (50 mL × 3). The combined organic phases were dried over anhydrous Na₂SO₄, and then clarified by filtration. The filtrate was rotary-evaporated to yield a colorless oil which later solidified to a white solid (3.28 g, 70%). ¹H NMR (400 MHz, 298 K, CDCl₃): δ 7.43 (m, 5H), 7.36 (s, 2H), 5.37 (s, 2H), 4.95 (s, 4H). ¹³C NMR (100 MHz, 298 K, CDCl₃): δ 170.9, 154.5, 133.2, 129.5, 129.3, 128.3, 113.2, 73.0, 25.3. LR-ESI-MS: calcd. for [C₁₄H₁₃⁷⁹Br₂NO + H]⁺ 369.9; found [M(⁷⁹Br) + H]⁺ 369.9

Tert-butyl 6-(((2-(tert-butoxy)-2-oxoethyl)amino)methyl)picolinate (5).

To a round-bottom flask with a stirred solution of tert-butyl 6-formylpicolinate⁴¹ (0.500 g, 2.40 mmol, 1 equiv) in dry MeOH (20 mL) was added tert-butyl glycinate (0.320 g, 2.40 mmol, 1 equiv). The mixture was stirred for 1 h at room temperature and then sodium cyanoborohydride (NaBH₃CN) (0.310 g, 4.87 mmol, 2 equiv) was added. The reduction reaction was continued for 3 h at room temperature before quenching with saturated NaHCO₃ in water (10 mL) and then extraction with DCM (20 mL × 3). The combined organic phases were dried over anhydrous Na₂SO₄, and then clarified by filtration. The filtrate was concentrated in vacuo and the residue was purified through a silica column (CombiFlash R_f automated column system, 12 g gold silica column, A: DCM B: MeOH, 0–5% B). The product fractions were combined and rotary-evaporated to give a pale yellow oil (0.550 g, 70%).¹H NMR (400 MHz, 298 K, CDCl₃): δ 7.76 (d, J = 7.7 Hz, 1H), 7.65 (t, J = 7.7 Hz, 1H), 7.43 (d, J = 7.6 Hz, 1H), 3.91 (s, 2H), 3.26 (s, 2H), 1.48 (s, 9H), 1.32 (s, 9H).¹³C NMR (100 MHz, 298 K, CDCl₃): δ 171.1, 163.9, 159.7, 148.6, 137.2, 124.9, 123.0, 81.9, 81.0, 54.2, 51.0, 27.9. LR-ESI-MS: calcd. for [C₁₇H₂₆N₂O₄ + H]⁺ 323.2; found [M + H]⁺ 323.1

Di-tert-butyl-6,6’-((((4-(benzyloxy)pyridine-2,6-diyl)bis(methylene))bis((2-(tert-butoxy)-2-oxoethyl)azanediyl))bis(methylene))dipicolinate (6)

Compound 4 (0.400 g, 1.30 mmol, 1 equiv), K₂CO₃ (595 mg, 4.31 mmol, 3.3 equiv) and KI (434 mg, 2.61 mmol, 2 equiv) were added sequentially to a stirred solution of compound 5 (0.837 g, 2.60 mmol, 2 equiv) in dry ACN (15 mL) in a round-bottom flask. The mixture was stirred at 30˚C for 24 h. K₂CO₃ was removed by centrifugation and then washed with DCM (10 mL × 3). The combined supernatants were concentrated in vacuo and then purified with a silica column (CombiFlash R_f automated column system, 12 g gold silica column, A: DCM B: MeOH, 0–5% B). The product fractions were rotary-evaporated to give a pale-yellow oil (0.670 g, 73 %). ¹H NMR (400 MHz, 298 K, CDCl₃): δ 7.92–7.61 (m, 6H), 7.52–7.30 (m, 5H), 7.12 (s, 2H), 5.11 (s, 2H), 4.03 (s, 4H), 3.86 (s, 4H), 3.33 (s, 4H), 1.57 (s, 18H), 1.43 (s, 18H). ¹³C NMR (100 MHz, 298 K, CDCl₃): δ 170.5, 166.2, 164.0, 160.2, 148.6, 137.2, 136.1, 128.5, 128.1, 127.7, 125.6, 123.3, 123.0, 107.8, 81.8, 80.9, 69.7, 64.4, 59.8, 56.1, 53.4, 28.0. LR-ESI-MS: calcd. for [C₄₈H₆₃N₅O₉ + Na]⁺ 876.5; found [M + Na]⁺ 876.6

Di-tert-butyl-6,6’-((((4-hydroxypyridine-2,6-diyl)bis(methylene))bis((2-(tert-butoxy)-2-oxoethyl)azanediyl))bis(methylene))dipicolinate (7)

Compound 6 (0.170 g, 0.200 mmol) was dissolved in dry MeOH (20 mL) in a three-neck round-bottom flask and saturated with N₂(g). Pd/C (10% w/w, 0.1 equiv) was added under a stream of N₂(g). The flask was purged with N₂(g), followed by H₂(g) from a balloon. The mixture was stirred vigorously at room temperature overnight under H₂ atmosphere, and then Pd/C was filtered off through a pre-wet (MeOH) Celite bed, washed with MeOH (10 mL × 5). The filtrate was rotary-evaporated to a pale-yellow oil (0.150 g) and used without further purification. LR-ESI-MS: calcd. for [C₄₁H₅₇N₅O₉ + H]⁺ 764.4; found [M + H]⁺ 764.6

4-((Tert-butoxycarbonyl)amino)phenethyl 4-methylbenzenesulfonate (8)

N-​boc-​2-​(4-​aminophenyl)​ethanol (1.97 g, 8.28 mmol, 1 equiv) was dissolved in THF (12 mL) and cooled to 0°C with an ice-water bath. NaOH aqueous solution (6 M, 11.9 mL) was added, followed by dropwise addition of para-tosyl chloride (3.16 g, 0.0169 mol, 2 equiv) in THF (24 mL) under N₂(g). After stirring at 0˚C for 1 h, the reaction mixture was warmed to room temperature and further stirred overnight. The mixture was extracted with DCM (30 mL × 3). The combined organic phases were washed with 1 M NaOH aqueous solution (40 mL × 2) and deionized water (40 mL × 2), and then dried over magnesium sulfate (MgSO₄). The mixture was clarified with filtration, evaporated in vacuo and then purified through a silica column (CombiFlash R_f automated column system, 24 g gold silica column, A: DCM B: MeOH, 0–5% B). The product fractions were rotary-evaporated to give a white solid (2.30 g, 71%). ¹H NMR (400 MHz, 298 K, CDCl₃): δ 7.68 (d, J = 8.3 Hz, 2H), 7.25 (dd, J = 13.4, 8.7 Hz, 4H), 7.01 (d, J = 8.4 Hz, 2H), 6.45 (s, 1H), 4.16 (t, J = 7.0 Hz, 2H), 2.89 (t, J = 7.0 Hz, 2H), 2.43 (s, 3H), 1.51 (s, 9H).¹³C NMR (100 MHz, 298 K, CDCl₃): δ 152.8, 144.8, 137.3, 133.1, 130.7, 129.9, 129.6, 128.0, 118.8, 80.7, 70.8, 34.8, 28.5, 21.7. LR-ESI-MS: calcd. for [C₂₀H₂₅NO₅S + H]⁺ 392.1; found [M + H]⁺ 392.1

Tetramethyl 6,6’,6”,6”’-((((4-(4-((tert-butoxycarbonyl)amino)phenethoxy)pyridine-2,6-diyl)bis(methylene))bis(azanetriyl))tetrakis(methylene))tetrapicolinate (9)

To a round-bottom flask with a stirred solution of compound 7 (72.1 mg, 0.0944 mmol, 1 equiv) in dry THF (1 mL) was added anhydrous K₂CO₃ (52.1 mg, 0.378 mmol, 4 equiv). The mixture was stirred vigorously for 1 h at 25˚C before the addition of compound 8 (42.1 mg, 0.108 mmol, 1.14 equiv). The mixture was stirred for 48 h at 25˚C when compound 7 was completely consumed. The solvent was evaporated in vacuo, and the residue was resuspended in DCM (6 mL). K₂CO₃ was removed by centrifugation and washed with DCM twice (~5 mL each). The combined organic phases were washed with saturated NaHCO₃ in water (10 mL × 2), water (10 mL × 2) and brine (10 mL × 2), and then dried over anhydrous MgSO₄. The drying agent was filtered off and the filtrate was concentrated in vacuo to a yellow oil. The product was confirmed by MS and then used without further purification in the next step. LR-ESI-MS: calcd. for [C₅₄H₇₄N₆O₁₁ + Na]⁺ 1005.5; found [M + K]⁺ 1005.9

6,6’,6”,6”’-((((4-(4-Aminophenethoxy)pyridine-2,6-diyl)bis(methylene))bis(azanetriyl))tetra-kis(methylene))tetrapicolinic acid (10)

Compound 9 (166 mg, 0.169 mmol, 1 equiv) was dissolved in TFA/DCM (1:1) (4 mL). The mixture was stirred overnight vigorously at room temperature, and then concentrated to dryness in vacuo. The crude product was re-dissolved in deionized water, and then purified through reverse phase HPLC (A: ACN/0.1% TFA, B: H₂O/0.1% TFA, 5–60% A over 40 min, 10 mL/min, t_R = 17.8 min). The combined product fractions were dried in vacuo to give a yellow oil (55.7 mg, 50%).¹H NMR (400 MHz, 298 K, D₂O): δ 8.23 (t, J = 7.9 Hz, 2H), 7.97 (d, J = 7.8 Hz, 2H), 7.91 (d, J = 7.7 Hz, 2H), 7.51 (d, J = 8.4 Hz, 2H), 7.43 (d, J = 8.5 Hz, 2H), 6.91 (s, 2H), 4.49 (s, 4H), 4.34 (t, J = 5.8 Hz, 2H), 4.28 (s, 4H), 3.82 (s, 4H), 3.17 (t, J = 5.8 Hz, 2H). ¹³C NMR (100 MHz, 298 K, D₂O): δ 13C NMR (101 MHz, D2O) δ 174.1, 169.6, 164.9, 154.1, 152.6, 146.5, 144.3, 139.5, 131.0, 128.7, 128.4, 125.2, 123.3, 110.7, 70.6, 57.8, 57.2, 33.9.. LR-ESI-MS: calcd. for [C₃₃H₃₄N₆O₉ + H]⁺ 659.7; found [M + H]⁺ 659.4.

H₄pypa-phenyl-NCS (11).

Compound 10 (68.0 mg, 0.103 mmol, 1 equiv) was dissolved in 1 M HCl/ glacial acetic acid (2 mL, 4:1 v/v) in a round-bottom flask; thiophosgene (CSCl₂) (119 μL, 1.55 mmol, 15 equiv) in CHCl₃ (2 mL) was then added dropwise using a Pasteur pipette to the stirred mixture which was stirred vigorously at room temperature overnight. After the reaction completed, CHCl₃ was removed with a Pasteur pipette. The aqueous phase was further washed with CHCl₃ (1 mL) which was removed by a Pasteur pipette. The process was repeated 4 times and then the residue was injected onto reverse phase HPLC (A: ACN/0.1% TFA, B: H₂O/0.1% TFA, 5–60% A over 40 min, 10 mL/min, t_R = 35 min). The product fractions were combined and lyophilized to give a fluffy white solid (21.7 mg, 30 %). ¹H NMR (400 MHz, 298 K, CD₃CN:D₂O 1:1): δ 8.62 – 8.50 (m, 4H), 8.27 (d, J = 7.3 Hz, 2H), 7.90 (d, J = 8.4 Hz, 2H), 7.84 (d, J = 8.4 Hz, 2H), 7.50 (s, 2H), 4.90 (s, 6H), 4.31 (s, 4H), 3.65 (t, J = 6.3 Hz, 2H). ¹³C NMR (100 MHz, 298 K, CD₃CN:D₂O 1:1): δ 173.5, 165.9, 155.0, 142.2, 131.2, 131.2, 128.7, 126.5, 126.5, 125.4, 111.2, 58.7, 56.6, 56.6, 56.2, 56.2. HR-ESI-MS: calcd. for [C₃₄H₃₂N₆O₉S+H]⁺ 701.2030; found [M+H]⁺ 701.2028

Na[^natY(pypa)]

H₄pypa · 2 TFA· 2 H₂O (21.8 mg, 4.16 × 10⁻⁵ mol, 1 equiv) was dissolved in D₂O (1 mL) in a scintillation vial and 0.1 M NaOD solution (aq) was added to adjust the pH to 7; Y(ClO₄)₃ · 6 H₂O (45.4 μL, 4.58 × 10⁻⁵ mol, 1.1 equiv) was then added. The mixture was stirred at room temperature for 1 h and used to confirm the complexation by LR-ESI-MS and NMR. ¹H NMR (400 MHz, 298 K, D₂O): δ 8.23 (t, J = 7.8 Hz, 2H), 8.08 (d, J = 7.7 Hz, 2H), 7.87 (t, J = 7.8 Hz, 1H), 7.80 (d, J = 7.8 Hz, 2H), 7.45 (d, J = 7.8 Hz, 2H), 4.71 – 4.58 (m, 4H), 4.40 (d, J = 14.7 Hz, 2H), 4.05 (d, J = 14.8 Hz, 2H), 3.97 (m, J = 17.0 Hz, 2H), 3.47 (d, J = 17.0 Hz, 2H).. ¹³C NMR (100 MHz, 298 K, D₂O): δ 178.7, 172.5, 156.7, 152.9, 151.1, 141.8, 140.2, 125.6, 124.1, 123.0, 64.6, 63.2. HR-ESI-MS: calcd. for [C₂₅H₂₁N₅O₈⁸⁹Y + 2Na]⁺ 654.0244; found [M + 2Na]⁺ 654.0251.

DFT Calculation

All DFT simulations were performed as implemented in the Gaussian 09 revision D.01 suite of ab initio quantum chemistry programs (Gaussian Inc., Wallingford, CT). B3LYP functional,^63,64 and the effective core potentials LanL2DZ basis sets for scandium^65,66 and yttrium^63,67,68 were applied to optimize the structural geometry in the presence of water solvent (IEF PCM as implemented in G09) without the use of symmetry constraints. Normal self-consistent field (SCF) and geometry convergence criteria were conducted for all the calculations. The calculated structures were visualized using Mercury 4.1

Solution Thermodynamics

All potentiometric titrations were carried out with a Metrohm Titrando 809 and a Metrohm Dosino 800 with a Ross combined electrode. A 20 mL and 25˚C thermostated glass cell with an inlet-outlet tube for nitrogen gas (purified through a 10% NaOH solution to exclude any CO₂ prior to and during the course of the titration) was used as a titration cell. The electrode hydrogen ion concentration was calibrated daily by direct titration of HCl (aq) with freshly prepared NaOH aqueous solution and the results were analyzed with the Gran procedure⁶⁹ in order to obtain the standard potential E° and the ionic product of water pK_w at 25˚C and 0.16 M NaCl (aq) used as a supporting electrolyte. Solutions were titrated with carbonate-free NaOH (aq, 0.16 M) that was standardized against freshly recrystallized potassium hydrogen phthalate aqueous solution. In the study of complex formation equilibria, the determination of the stability constants of the [Y(Hpypa)] species was carried out by two different methods. The first method used UV-vis spectrophotometric measurements on a set of solutions containing 1:1 metal to ligand molar ratio ([H₄pypa] = [Y]³⁺ = 1.33 × 10⁻⁴ M) and different amounts of standardized HCl (aq) and NaCl (aq) to set the ionic strength constant at 0.16 M when possible. The equilibrium H⁺ concentration in this UV in-batch titration procedure at low pH solutions (2 ≥ pH ≤ 0) was calculated from solution stoichiometry, not measured with a glass electrode. For the solutions of high acidity, the correct acidity scale H⁰ was used.⁷⁰ The spectral range was 200–400 nm at 25˚C and 1 cm path length. The molar absorptivities of all the protonated species of H₄pypa calculated with HypSpec2014⁵³ from the protonation constant experiments⁴¹ were included in the calculations. The second method used competition pH-potentiometric titrations with EDTA as a ligand competitor and the composition of the solutions was [Y]³⁺ = Y³⁺ ~ 1.51 × 10⁻³ M, [H₄pypa] ~ 6.88 × 10⁻⁴ M and [EDTA] ~ 1.51 × 10⁻³ M at 25˚C and I = 0.16 M NaCl (aq). The stability constants for the complexes formed by EDTA and Y³⁺ were taken from the literature.⁷¹ Direct pH-potentiometric titrations of the Y³⁺-H₄pypa systems were also carried out. The Y³⁺ metal ion solution was prepared by adding the atomic absorption (AA) standard solution to a H₄pypa solution of known concentration in the 1:1 metal to ligand molar ratio. Ligand and metal concentrations were in the range of 8.24 × 10⁻⁴ M. The exact amount of acid present in the AA standard solution was determined by Gran’s method⁶⁹ titrating equimolar solutions of Y(III) and Na₂H₂-EDTA. Each titration consisted of 100–150 equilibrium points in the pH range 1.6–11.5; equilibration time for titrations was up to 5 min for metal complex titrations. Three replicates of each titration were performed. Relying on the stability constants for the species [Y(Hpypa)] obtained by the two different methods, the fitting of the direct potentiometric titrations was possible and yielding the stability constants in Table 3. All the potentiometric measurements were processed using the Hyperquad2013⁵² software while the obtained spectrophotometric data were processed with the HypSpec2014⁵³ program. Proton dissociation constants corresponding to hydrolysis of Y(III) aqueous ions included in the calculations were taken from Baes and Mesmer.⁷² The overall equilibrium (formation) constants log β referred to the overall equilibria: pM + qH + rL ⇆ M_pH_qL_r (the charges are omitted), where p might also be 0 in the case of protonation equilibria and q can be negative for metal-hydroxo species. Stepwise equilibrium constants log K correspond to the difference in log units between the overall constants of sequentially protonated (or hydroxide) species. The parameter used to calculate the metal scavenging ability of a ligand towards a metal ion, pM, is defined as −log [Mⁿ⁺]_free at [ligand] = 10 mM and [Mⁿ⁺] = 1 μM at pH = 7.4.⁷³

Production and Radiochemical Separation of Yttrium-86

Yttrium-86 (t_1/2 = 14.7 h, 34% β⁺, E_β+max = 3.2 MeV) was produced in a 16 MeV GE PETtrace biomedical cyclotron using enriched [⁸⁶Sr][SrCO₃] targets of pressed powder. Following irradiation, the radiochemical isolation of the yttrium-86 was performed by single column extraction chromatography, as previously described.⁷⁴ Briefly, the target material was dissolved in 5 mL of 9 M HCl (aq), and loaded onto a column filled with a resin functionalized with N,N,N′,N′-tetrakis-2-ethylhexyldiglycolamide (branched DGA, Eichrom). Subsequent washes were carried out using 9 M HCl (aq, 15 mL) and 0.5 M HNO₃ (aq, 15 mL) to remove bulk strontium and other trace metal contaminants. No-carrier-added yttrium-86 was eluted with 0.1 M HCl (aq, 4 × 0.3 mL). The activity was assayed in a Capintec CRC-15R dose calibrator (setting #850/2).

Production and Radiochemical Separation of Scandium-44

Scandium-44 was cyclotron-produced using ^natCa[p,n]^4xSc nuclear reactions on pressed targets of metallic calcium (300–350 mg). Target preparation was performed in air, and rapidly mounted in the cyclotron to reduce calcium oxidation. Irradiations were performed at 20 μA for 1 h with direct water cooling, and a 12.7 μm Nb foil was used to degrade the beam energy from the nominal 16 MeV to 14.1 MeV. Under these conditions, a scandium-44 production yield of 0.4 mCi/μAh was obtained through the reaction ^natCa(p,n)⁴⁴Sc. Isolation of the produced scandium-44 was carried out by single column extraction chromatography using a N,N,N’,N’-tetrakis-2-ethylhexyldiglycolamide functionalized resin (DGA-branched, Eichrom).⁶⁵ The target was dissolved in 9 M HCl (aq, 10 mL) and passed through a 1 mL fritted solid phase extraction (SPE) tube filled with the DGA resin (~120 mg), loading the scandium-44 and eluting bulk Ca²⁺. Remaining Ca²⁺ was removed by rinsing the column with 4 M HCl (aq, 20 mL). Next, a 12 mL wash with 1 M HNO₃ (aq) was performed to elute possible trace metal contaminants such as Zn, Fe and Cu. Finally, the scandium-44 was eluted in a small volume using 0.1 M HCl (aq, 4 × 500 μL fractions). The radionuclidic and chemical purities were confirmed by high purity germanium (HPGe) gamma spectrometry and microwave plasma atomic emission spectroscopy (MP-AES), respectively.

Radiolabeling Studies

An aliquot of ligand solution (H₄pypa) (10 μL) was mixed with HEPES solution (0.5 M, pH = 7, 87 μL), followed by [⁸⁶Y][YCl₃] in HCl (aq) (4.14 MBq, 3 μL). The reaction mixtures were incubated at ambient temperature over the desired time period; 3 μL was then spotted on a silica aluminum-backed TLC plate, and developed in DTPA buffer (0.1 M, pH = 5.5). The TLC plate was read by a TLC reader, showing the free metal ion migrated to the solvent front while the complex stayed at the baseline. The areas of both peaks were used to calculate RCY%.

In vitro Mouse Serum Challenge

To the radiolabeled sample (100 μL, 1.4 MBq), an equal volume (100 μL) of the mouse serum was added. The mixture was incubated at 37˚C and 5 μL aliquots were collected at desired time points (0.5 h, 4 h, 24 h, 48 h). Each aliquot was spotted onto a silica aluminum-backed TLC plate next to the control spot (yttrium-86 in buffer with serum), and then developed in DTPA buffer (0.1 M, pH = 5.5). The TLC plate was read by the TLC reader. The free metal migrated to the solvent front while the complex stayed at the baseline. The areas of both peaks were used to calculate % intact.

Bioconjugation of H4pypa-phenyl-TRC105 and Radiolabeling with Yttrium-86 and Scandium-44

TRC105 (TRACON Pharmaceuticals) was prepared for radiolabeling through conjugation with H₄pypa-phenyl-NCS. The pH of the TRC105 solution was adjusted to ~9.0 with 0.1 M sodium carbonate buffer and H₄pypa-phenyl-NCS chelator was added in a 10:1 (chelator:antibody) molar ratio. After reacting for 2 h at room temperature, H₄pypa-phenyl-TRC105 was purified by size exclusion chromatography (PD-10, GE-Healthcare) using phosphate-buffered saline (PBS) as the mobile phase. For animal studies, 2.7 μg of H₄pypa-phenyl-TRC105 per MBq of yttrium-86 (or scandium-44) was added to the activity solution in sodium acetate buffer (0.1 M, pH = 5.5) and incubated at 37˚C for 30 min. After labeling, [^AE][E(pypa-phenyl-TRC105)] (^AE = ⁸⁶Y, ⁴⁴Sc) was purified by size exclusion chromatography using PBS as the mobile phase. Radiochemical yields (RCY) were measured via iTLC-SG plates developed in DTPA buffer (0.1 M, pH = 5.5) for the ⁸⁶Y-conjugate and sodium citrate buffer (0.4 M, pH = 4.5) for the ⁴⁴Sc-conjugate. The radiochemical yields were consistently above 90% for both radiotracers.

PET/CT Imaging and Biodistribution Studies

Animal experiments were conducted with the approval of the University of Wisconsin Institutional Animal Care and Use Committee (IACUC) and in strict accordance with the NIH guidelines for the care and use of laboratory animals (NIH Publication No. 85–23 Rev. 1985) and the relevant guidelines and regulations (protocol ID: B00000657). 4T1 cells were cultured in RPMI 1640 growth medium (Invitrogen) with a 10% fetal bovine serum (FBS) supplement. During culturing, cells were incubated at 37˚C with 5% CO₂. Tumors were grafted in four-week-old female Balb/c mice by subcutaneous injection of 2–3 × 10⁶ cells, suspended in 100 μL of 1:1 mixture of RPMI 1640 and Matrigel (BD Biosciences).

For imaging studies, mice bearing 4T1 murine breast cancer tumors were injected via the tail vein with 11–18 MBq of either [⁸⁶Y][Y(pypa-phenyl-TRC105)] or [⁴⁴Sc][Sc(pypa-phenyl-TRC105)]. PET scans of 20 million coincidence events per mouse were obtained using an Inveon PET/CT scanner (Siemens) at different post-injection timepoints. Measurements of regions-of-interest (ROI) in PET images were performed using the Inveon Research Workspace (Siemens).

Following the final imaging timepoint, mice were euthanized via CO₂ asphyxiation, and major organs were excised, wet-weighed, and radioactive content was measured using a gamma counter (PerkinElmer). Results of PET ROI analysis and biodistribution studies are presented as %ID/g.