Evaluation of macrocyclic hydroxyisophthalamide ligands as chelators for zirconium-89

The development of bifunctional chelators (BFCs) for zirconium-89 immuno-PET applications is an area of active research. Herein we report the synthesis and evaluation of octadentate hydroxyisophthalamide ligands (1 and 2) as zirconium-89 chelators. While both radiometal complexes could be prepared quantitatively and with excellent specific activity, preparation of 89Zr-1 required elevated temperature and an increased reaction time. 89Zr-1 was more stable than 89Zr-2 when challenged in vitro by excess DTPA or serum proteins and in vivo during acute biodistribution studies. Differences in radiometal complex stability arise from structural changes between the two ligand systems, and suggest further ligand optimization is necessary to enhance 89Zr chelation.


Introduction
The development of zirconium-89 ( 89 Zr: t 1/2 = 78.4h; β + ; 23%, 909 keV)-based PET radiopharmaceuticals has intensified since comprehensive procedures for producing 89 Zr in high specific activity and attaching it to monoclonal antibodies (mAbs) were first described [1][2][3].The half-life and positron emission properties of 89 Zr make it ideal for the labeling of mAbs, which require extended circulation time for effective in vivo antigen targeting.Several 89 Zr-labeled mAbs are currently in clinical trials [4][5][6].
In this report, we describe the development of BFCs containing four 2-hydroxyisophthalamide coordinating units for 89 Zr chelation (Fig 1).We became interested in hydroxyisophthalamide (IAM) based-ligands, since these coordination motifs are similar to those used by bacteria for metal ion sequestration and have been used successfully as chelators for a variety of lanthanide metal cations [25][26][27][28][29][30][31][32].Additionally, they offer potential advantages for 89 Zr BFC design [33].These aromatic systems bind through a combination of phenolic and carbonyl oxygen atoms, which should aid coordination to hard radiometals such as zirconium.Furthermore, the IAM unit can be easily derivatized to quickly expand the diversity of ligands that can be synthesized and evaluated.

Materials and methods
Zirconium-89 ( 89 Zr: t ½ = 78.4h, β + : 22.8%, E β+max = 901 keV; EC: 77%, E γ = 909 keV) was purchased from Washington University School of Medicine (St. Louis, MO) or Zevacor, Inc. (Dulles, VA).Unless otherwise noted, all other chemicals were purchased from Sigma-Aldrich Chemical Co.(St.Louis, MO), and solutions were prepared using ultrapure water (18 MO-cm resistivity).Radiochemistry reaction progress and purity were analyzed using a Waters analytical HPLC (Milford, MA), which runs Empower software and is configured with a 1525 binary pump, 2707 autosampler, 2998 photodiode array detector, 2475 multichannel fluorescence detector, 1500 column heater, fraction collector, Grace Vydac 218MS C18 column (5 μm, 4.6 × 250 mm, Grace Davidson, DeerField, IL) and a Carrol Ramsey 105-s radioactivity detector (Berkeley, CA).All ligands (1 and 2) and associated Nat Zr-complexes were monitored at 220 nm using a mobile phase consisting of 0.01% TFA/H 2 O (solvent A) and 0.01% TFA/acetonitrile (solvent B), and a gradient consisting of 0% B to 70% B in 20 min at a flow rate of 1.2 mL/min.In addition, radio-TLC was conducted on a Bioscan AR 2000 radio-TLC scanner equipped with a 10% methane:argon gas supply and a PC interface running Winscan v.3 analysis software (Eckert & Ziegler, Berlin, DE).Varian ITLC-SG strips were employed using a 50 mM DTPA (pH 7) solution as eluent, and the complex 89 Zr (ox) 2 as a standard control.Radioactive samples were counted using a Perkin Elmer 2480 Wizard 1 gamma counter (Waltham, MA).Preparation of ligand 1 has been reported [27].Starting compound 3, which was used for the synthesis of 2, was synthesized as described previously [32].

Density functional theory calculations
Ground state density functional theory calculations were performed at the Molecular Graphics and Computational Facility, College of Chemistry, University of California, Berkeley using Gaussian 09 [34].The ground state geometries of Zr-1 and Zr-2 (S3 Fig and S1 Table ) were optimized using the B3LYP functional, treating the light atoms (H through O) with the 6-31G (d,p) basis set [35][36][37][38].The Zr atom was treated with the effective core potential MWB28 [39,40].No solvent, symmetry constraints, or counterions were included in the calculations.Crystal structures of the related H22 linked 2-hydroxyisophthalamide (IAM) ligands bound to Tb III were used as starting points for these Zr IV complexes [27].In addition to geometry optimizations, frequency calculations were performed, and all structures are presented in energy minimized conformations.

Radiolabeling of macrocyclic 2-hydroxyisophthalamide ligands (1, 2) with 89 Zr
The complexation of 89 Zr with macrocyclic 2-hydroxyisophthalamide ligands (1, 2) was achieved by reacting 5-10 μg (5-10 μL,1.0 mg/mL in water) of each ligand with an aliquot of 89 Zr(ox) 2 (0.6 mCi, 22.2 MBq) diluted in 100 μL of water and pH adjusted to 7-7.5 using 1 M Na 2 CO 3 .The 89 Zr-1 reaction was incubated at 95˚C for 2 h; the 89 Zr-2 reaction was incubated at 50˚C for 1 h.Formation of 89 Zr-1 and 89 Zr-2 was monitored by radio-TLC using Varian ITLC-SG strips and 50 mM DTPA (pH 7) as the mobile phase.In this system, unchelated 89 Zr forms a complex with DTPA and elutes near the solvent front (denoted as the red line at 130 mm in radio-TLC chromatograms); the 89 Zr-ligand complex remains at the origin (denoted as the red line at 30 mm in radio-TLC chromatograms) (S4 Fig) .The identity of each radioactive complex was further confirmed by comparing its radio-HPLC elution profile to the UV-HPLC spectrum of its nonradioactive Zr-complex (S5 and S6 Figs).

Determination of partition coefficients (log P)
The partition coefficient (log P) for each complex was determined by adding 5 μL of each 89 Zrlabeled complex (5 μCi; 0.19 MBq) to a mixture of 500 μL of octanol and 500 μL of water [41].The resulting solutions (n = 8) were vigorously vortexed for 5 min at room temperature and then centrifuged for 5 min to ensure complete layer separation.From each sample, a 50 μL aliquot was removed from each phase.Radioactivity in each phase was counted separately in a gamma counter.Each organic phase was washed with water before gamma counting.The partition coefficient was calculated as a ratio of counts in the octanol fraction to counts in the water fraction.The log P values were reported as an average of four measurements (S2 Table ).

Animal model
All studies were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Animals of the National Institutes of Health, and approved by the Wake Forest University Health Sciences Institutional Animal Care and Use Committee (Protocol # A14-081).Normal mice used in the experiments were received from The Jackson Laboratory, USA and were accommodated in the housing room operated according to the Animal Welfare Act, the guidelines established by the NIH and all other Federal and State statutes and regulations pertaining to laboratory animal care.All animals were acclimated to a 12-hour light/dark cycle and received food and water ad libitum.All animals were checked by veterinary staff daily to ensure well-being, proper environment and monitor for signs of distress.Clinical signs of distress in laboratory rodents include decreased activity, pilo-erection, un-groomed appearance, excessive licking and scratching, self-mutilation, abnormal stance, hunched appearance, rapid or shallow respiration, grunting, dilated pupils, aggressiveness towards handler, high pitch vocalizations, change in feeding activity, and attempts to separate from group.In the event of clinical distress, animals were euthanized after consulting the on-call veterinarian.To minimize suffering during the tail vain injection, animals were kept under deep isoflurane anesthesia.At the end point of the study, animals were anesthetized with isoflurane and were euthanized by cervical dislocation, which is consistent with the recommendations of the American Veterinary Medical Association guidelines on Euthanasia.Animals were monitored every 30 minutes during the experiment, but during the experiment, none of the animals died or were found ill before experimental end point.

Biodistribution studies
Biodistribution studies were based upon a previously published procedure [42].Briefly, female NIH Swiss mice (6-8 wk old, n = 6) were injected with each 89 Zr-labeled complex (0.55 MBq (15 μCi)/mouse) via the tail vein, and sacrificed at 2, 4, 24, 48 and 72 h post-injection.Organs and tissues of interest were excised and weighed, and radioactivity was counted on a gamma counter.The percent injected dose per gram (%ID/g) was calculated by comparison to a weighed, counted standard for each group (S5 and S6 Tables).

Statistical methods
All of the data are presented as mean±SD or mean (95% confidence intervals).For statistical classification a Student's t test (two-tailed, unpaired) or one-way anova (with Tukey's multiple comparison post-test) was performed using GraphPad Prism software (San Diego, CA).p<0.05 was considered significant.

Results and discussion
Ligand 1 was synthesized according to published procedures [27].Preparation of an IAM bimacrocyclic ligand 2 (Scheme 1) began with 2-benzyloxy-1,3-phenylenebis((2-thioxothiazolidin-3-yl)methanone) 3, which was condensed with tetrakis-(2-aminoethyl) ethylene diamine 4 under pseudo-first order conditions to provide the activated tetra-amide 5.This was reacted with tris-(2-aminoethyl)amine 6 under high dilution conditions to form the bi-macrocycle 7. The remaining activated amide in 7 was reacted with amine 8 to provide bi-macrocycle 9. Protective groups were removed using a solution of concentrated hydrochloric acid in acetic acid to provide bi-macrocycle 2. Both ligand 1 and ligand 2 were completely characterized and observed using NMR spectroscopy, high resolution mass spectrometry and elemental analysis.The nonradioactive Zr-1 and Zr-2 complexes were prepared by reacting ligands (1 and 2, 1 equiv.each) with Zr(acac) 4 (0.75 equiv.) in methanol.Characterization of both complexes by NMR revealed spectra containing a complex pattern of multiplets that could not be resolved further.However, high resolution mass spectrometry (S1 and S2 Figs) and elemental analysis confirm the identity of each complex.Interestingly, several peaks were observed in the UV-HPLC chromatograms (S5 and S6 Figs), but based upon mass spectrometry and elemental analysis, it is believed that these peaks are structural isomers that form upon the coordination of the Zr 4+ ion in solution.It is not implausible to imagine the formation of several structural isomers, which are dependent upon the number of phenolic and carbonyl oxygen atoms involved in metal ion complexation.Attempts were made to isolate each isomer by HPLC for NMR analysis, but isomeric inter-conversion was observed during this process, which made our attempts at further characterization unsuccessful.While attempts to elucidate the molecular structure of Zr-IAM complexes using single crystal x-ray diffraction analysis are ongoing, we performed density functional theory calculations on both Zr-1 and Zr-2 to model ground state coordination geometries (S1 Table ).If we neglect the alkyl amine linker arm, the structure of Zr-1 (S3 Fig) has twofold rotational symmetry, where the symmetry axis passes through the midpoint of the central ethylene units and through the metal center.Modeling revealed the energy minimized structure of Zr-2 could exist with two distinct binding modes (A or B; S3 Fig), although both are quite similar.In structure A, the amine sidearm is connected to the amide not involved in Zr 4+ ion coordination.In structure B, the amine sidearm is connected to the amide directly involved in Zr 4+ ion coordination.Despite these differences, structure A was found to be only 0.42 kcal/mol lower in energy than structure B, which is a very small difference given the size and flexibility of the ligands.Thus it is likely that both A and B are present at significant quantities in solution, which further suggests that the multiple peaks observed in the HPLC chromatograms result from coordination isomers in solution.
Our initial radiochemistry studies with 1 revealed that elevated reaction time and temperature were necessary for quantitative radiolabeling (S4 Fig) .Although quantitative radiolabeling was achieved, these conditions would be too harsh for radiolabeling of 1-mAb conjugates.Accordingly, we reduced the structural rigidity of 1 by partially uncoupling one of the isophthalamide coordinating units to yield 2. Using this strategy, the time and temperature needed to quantitatively radiolabel 2 with 89 Zr (S4 Fig) decreased markedly.Furthermore, an A s of 0.7 GBq/μmol was observed, which is comparable to the A s of similar 89 Zr-complexes reported in the literature [14,19,23].
Lipophilicity (log P) plays an essential role in estimating the biodistribution of 89 Zr-complexes in vivo [41].Using the water-octanol partition method, the log P (pH 7) of 89 Zr-1 and 89 Zr-2 was determined to be -2.97±0.02and -1.45±0.06,respectively, compared to -2.83±0.04 for 89 Zr-DFO (S2 Table ).These negative values suggest that the complexes possess hydrophilic character, which is expected to result from the abundant hydrogen bonding motifs displayed by the radiometal complexes while in solution.These values are similar to recently published values for 89 Zr-terepthalamide and 89 Zr-hydroxypyridinone complexes [14,19,23].Complexes of 1 and 2 also contain 4 and 3 tertiary amines in the scaffold, respectively, in addition to the primary amines, some of which might be protonated at neutral pH.
Initial stability studies were also conducted by incubating 89 Zr-1 or 89 Zr-2 in the presence of a 50 mM DTPA solution over 7 days (S3 Table ). 89Zr-1 was more resistant to DTPA challenge than 89 Zr-2.Only 15% and 28% of 89 Zr-1 underwent transchelation after 1 and 7 days, respectively, while 37% and 74% of 89 Zr-2 underwent transchelation during the same study period. 89Zr-DFO displayed stability similar to 89 Zr-2 under these conditions, with 45% and 59% transchelation after 1 and 7 days, respectively.Likewise, incubation in human serum revealed 89 Zr-1 was more resistant to serum protein challenge than 89 Zr-2 (S4 Table ).While 75% of 89 Zr-1 remained intact after 7 days, only 17% of 89 Zr-2 was observed at the same time point. 89Zr-DFO remained intact upon incubation with serum.The resistance of 89 Zr-1 vs. 89 Zr-DFO to DTPA challenge may arise from the enhanced binding that occurs through the four IAM units, which can accommodate the 89 Zr 4+ ion to form an eight coordinate complex.On the other hand, reduced stability of 89 Zr-2, observed as 89 Zr 4+ ion transchelation to DTPA or serum proteins, may be caused by the reduced structural rigidity of ligand 2, which results in dissociation of an IAM coordinating unit and subsequent loss of the coordinated 89 Zr 4+ ion.  8Zr-2 and 89 Zr-DFO from the blood, liver, kidney, and bone; and complete biodistribution details can be found in SI (S5-S7 Tables).Animals injected with 89 Zr-1 and 89 Zr-2 demonstrated elevated levels of radioactivity in the blood at early time points when compared to animals injected with 89 Zr-DFO [19].However by 72 h p.i., blood retention was comparable in animals receiving 89 Zr-DFO and 89 Zr-1, but not 89 Zr-2.( 89 Zr-DFO vs. 89 Zr-1 vs. 89 Zr-2: (mean%ID/g±SD; [one-way ANOVA]: 0.00033±0.0008vs. DFO is a hexadentate acyclic ligand that forms a +1 charged complex.Conversely, 1 is a trimacrocycle comprised of 24 and 30-atom rings, whereas 2 forms a bi-macrocyclic complex of 24 and 27 membered rings; both ligands are octadentate (Fig 1).The altered charge of 89 Zr-1 and 89 Zr-2 when compared to 89 Zr-DFO might enhance their interactions with serum proteins, which retard clearance from the blood stream.These observations correlate well with the in vitro serum stability data obtained for both 89 Zr-IAM complexes.
Animals receiving 89 Zr-DFO had significantly less radioactivity in kidney tissue when compared to the 89 Zr-IAM complexes at every time point.By 72 h p.i., animals receiving 89 Zr-DFO, 89 Zr-1 or 89 Zr-2 had excreted 49%, 71%, and 54% of the activity recorded at 2 h ( 89 Zr-DFO vs 89 Zr-1 vs. 89 Zr-2: (mean%ID/g±SD; [one-way ANOVA value]: 0.69±0.098vs. 4.71 ±0.84 vs. 3.60±0.79;[F(2,15) = 82.11,p < 0.0001]).As the 89 Zr-IAM complexes diffuse from the blood pool into the kidney, pH may decrease, which can cause changes in the charge or structure of these radiometal complexes.These changes may induce aggregation or transchelation, which contribute to the retention of radioactivity in kidneys.Initially, less radioactivity was observed in the kidneys of mice injected with 89 Zr-2; however, the amounts for both IAM complexes become similar within 48 hours, consistent with the slower plasma clearance of 89 Zr-2.Further study is needed to ascertain whether increased radioactivity in the kidney relative to DFO would occur using antibody-chelator conjugates.
Since bone marrow is often considered a dose-limiting organ, radioactivity retention in bone was also examined [11].Animals injected with either 89 Zr-IAM complex demonstrated significantly more radioactivity in their bone tissue when compared to animals injected with 89 Zr-DFO at 72 h p.i. ( 89 Zr-DFO vs 89 Zr-1 vs. 89 Zr-2: (mean%ID/g±SD; [one-way ANOVA value]: 0.079±0.014vs. 0.11±0.006vs. 0.68±0.33;[F(2,15) = 17.57, p = 0.0001]).Interestingly, the biodistribution trends for 89 Zr-1 and 89 Zr-2 are very different.The amount of radioactivity in the bone of animals that received 89 Zr-1 diminished over time, and may reflect perfusion and clearance kinetics rather than demetallation and incorporation into the bone matrix.A similar phenomenon has been observed with a recently described hydroxypyridinone ligand [14].By contrast, radioactivity in the bones of mice receiving 89 Zr-2 continually increased during this experiment, and most likely reflects radiometal chelate instability, transchelation, and metabolic sequestration of the 89 Zr within the phosphate-rich bone matrix.

Conclusions
Herein, we describe the synthesis and evaluation of hydroxyisophthalamide ligands as bifunctional chelators for 89 Zr using log P, DTPA, and serum challenge studies.Additionally, the biodistribution of 89 Zr-1 and 89 Zr-2 was evaluated in normal mice.While the preparation of 89 Zr-1 required higher temperatures and longer reaction times compared to 89 Zr-2, it was more resistant to exogenous ligand challenge.Animals injected with 89 Zr-1 retained less radioactivity in the blood, liver, and bone than those receiving 89 Zr-2.These differences are believed to be caused by increased flexibility engineered into ligand 2. These data suggest that IAM ligands, which were previously found useful for the sequestration of a variety of lanthanide metal cations, can also function as 89 Zr chelators.Additional ligands that synergize the facile radiochemistry of 2 with the enhanced stability demonstrated by 89 Zr-1 are currently being evaluated in 89 Zr-immuno-PET applications, and will be reported in due course.

89Zr- 1
and 89 Zr-2 were evaluated in vivo through biodistribution studies in normal mice.Fig 2 displays the clearance properties of 89 Zr-1,