Novel Arginine-containing Macrocyclic MMP Inhibitors: Synthesis, 99mTc-labeling, and Evaluation

Matrix metalloproteinases (MMPs) are involved in tissue remodeling. Accordingly, MMP inhibitors and related radiolabeled analogs are important tools for MMP-targeted imaging and therapy in a number of diseases. Herein, we report design, synthesis, and evaluation of a new Arginine-containing macrocyclic hydroxamate analog, RYM, its hydrazinonicotinamide conjugate, RYM1 and 99mTc-labeled analog 99mTc-RYM1 for molecular imaging. RYM exhibited potent inhibition against a panel of recombinant human (rh) MMPs in vitro. RYM1 was efficiently labeled with 99mTcO4− to give 99mTc-RYM1 in a high radiochemical yield and high radiochemical purity. RYM1 and its decayed labeling product displayed similar inhibition potencies against rhMMP-12. Furthermore, 99mTc-RYM1 exhibited specific binding with lung tissue from lung-specific interleukin-13 transgenic mice, in which MMP activity is increased in conjunction with tissue remodeling and inflammation. The results support further development of such new water-soluble Arginine-containing macrocyclic hydroxamate MMP inhibitors for targeted imaging and therapy.

rupture risk. However, RP805 has a relatively prolonged blood circulation which may not be optimal for early time imaging of vascular diseases. In addition, the poor water solubility of RP805 precursor has impeded blocking studies in vivo at suitable concentrations. Therefore, we sought to pursue new macrocyclic hydroxamate analogs and related imaging agents with improved physicochemical properties for in vivo MMP-targeted imaging and therapy, with the ultimate goal of future clinical translation.
Previous studies performed on anti-succinate-based hydroxamic acids have elucidated the structural requirements of this class of inhibitors for anti-MMP activity, which include a hydroxamate as a zinc-binding site, as well as α, P1′, P2′ and P3′ substituents 1 . The preferred conformation of this type of inhibitors displays α and P2′ substituents on the same side with close proximity spatially, which can be cyclized as shown in RP805 and its precursor. Wide variations of P3′ substituent can be tolerated. Based on this structure-activity relationship (Fig. 2), we designed a new macrocyclic hydroxamate-based analog, RYM, containing an Arginine (Arg) residue at P3′ position for improving the hydrophilicity. Furthermore, we designed an analog, RYM1, containing a hydrazinonicotinamide (HYNIC) as a precursor for 99m Tc-labeling to form 99m Tc-RYM1 for pilot studies. Herein, we report the related synthesis, 99m Tc labeling, MMP inhibition, and other evaluations in vitro and ex vivo. In vivo evaluation of 99m Tc-RYM1, including biodistribution, imaging and binding specificity in murine models of carotid and aortic aneurysm, is reported elsewhere 23 . Fig. 3, the synthetic strategy includes an initial synthesis of a protected anti-succinate-based macrocyclic acid analog (I-7) 24,25 . The synthesis was started from an anti-succinic acid derivative, i.e. (2R,3S)-3-(tert-butoxycarbonyl)-2-isobutylhex-5-enoic acid which was first converted into its benzyl ester (I-1) in the presence of 1,8-diazabicyclo [5.4.0]undec-7-ene (DBU) and benzyl bromide in toluene. Hydroboration of I-1 by 9-borabicyclo[3.3.1]nonane (9-BBN) in tetrahydrofuran (THF), followed by oxidation with H 2 O 2 , gave the hydroxyl derivative (I-2). The conversion of I-2 into its bromide derivative (I-3) was achieved using CBr 4 and triphenylphosphine (Ph 3 P) in dichloromethane (DCM). The benzyl ester group of I-3 was removed by catalytic hydrogenation to give its acid analog (I-4). Coupling of I-4 with L-tyrosine benzyl ester in the presence of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDCI) and hydroxybenzotriazole (HOBT) in N,Ndimethylformamide (DMF) gave the anti-succinate-tyrosine benzyl ester (I-5). Cyclization of I-5 in the presence of Cs 2 CO 3 in anhydrous acetonitrile gave the protected anti-succinate macrocyclic derivative (I-6). Finally, the macrocyclic acid analog (I-7) was obtained by removing the benzyl group of I- 6 in the presence of 10% palladium on carbon (Pd/C) and HCOONH 4

in methanol.
As depicted in Fig. 4, we used Fmoc-Arg(Mtr)-OH as the starting material because the protecting group Mtr is stable in trifluoroacetic acid (TFA) solution compared with another commonly used protecting group Pbf in Fmoc-Arg(Pbf)-OH 26 . Therefore, Fmoc-Arg(Mtr)-OH was first reacted with CH 3 NH 2 in the presence   (I-9). The reaction between I-7 and I-9 in the presence of EDCI/HOBT afforded the Arg(Mtr)-containing macrocycle derivative (I-10), followed by de-protection of I-10 with TFA to give the carboxylic acid analog (I-11). The pre-activation of I- 11 in the presence of HOAT, HATU, and DIEA in DMF, followed by reaction with O-(tert-butyldimethylsilyl)hydroxylamine (TBDMS-ONH 2 ) gave I-12. Finally, the target compound RYM was obtained by de-protection of I-12 in a solution of TFA, water, TIS, and phenol.
We used a similar strategy for synthesis of the HYNIC conjugate (2). Therefore, Fmoc-Arg(Mtr)-OH was first reacted with NH 2 CH 2 CH 2 NH-Boc in the presence of EDCI/HOBT in DMF to give Fmoc-Arg(Mtr)-NH 2 CH 2 CH 2 NH-Boc (I-13) as depicted in Fig. 5. De-protection of I-13 with 5% piperidine in DCM gave H-Arg(Mtr)-NH 2 CH 2 CH 2 NH-Boc (I-14). The reaction between I-7 and I-14 in the presence of EDCI/HOBT afforded the Arg(Mtr)-containing macrocycle derivative (I-15), followed by de-protection of I-15 with TFA to give the carboxylic acid analog (I-16). I-16 was reacted with (Boc) 2 O in aqueous solution of Na 2 CO 3 to form the Boc-protected analog I-17. The pre-activation of I-17 in the presence of HOAT, HATU, and DIEA in DMF, followed by reaction with TBDMS-ONH 2 gave I-18. I-18 was de-protected in a solution of TFA, water, TIS, and phenol to give I-19. Boc-HYNIC was prepared in two steps 27 and activated in the presence of HATU, HOAT, DIEA, and DIEA in DMF. The activated Boc-HYNIC-OAT was reacted with I-19 to give I-20 which was de-protected in a solution of TFA, TIS, phenol, and water to give the desired HYNIC conjugate, RYM1.
The two crude products (RYM and RYM1) were purified by semi-preparative HPLC using aqueous acetonitrile as mobile phase to give the desired fractions as identified by LC-MS. The final products were obtained as white powder from lyophilization, which should be their corresponding TFA salts of the guanidine group. Both were further identified by analytical HPLC, high resolution mass spectroscopy (HR-MS, Supplemental Figure) and 1 H/ 13 C NMR analysis. MMP inhibition assays. MMP inhibition assays were carried out based on the kinetic effects of an inhibitor on the MMP-mediated catalytic cleavage of a fluorogenic substrate, namely Mca-Lys-Pro-Leu-Gly-Leu-Dpa-Ala-Arg-NH 2 , as reported by others previously 9,23,28 . In the first set of studies, the inhibitory effect of RYM was evaluated against 5 activated recombinant human (rh)MMPs, including rhMMP-2, rhMMP-7, rhMMP-9, rhMMP-12, and rhMMP-13. GM-6001, a broad-spectrum MMP inhibitor, was also tested for comparison. As expected, GM-6001 showed potent inhibitory activities against the five rhMMPs without significant selectivity and with the K i values close to those reported in the literature (Table 1). RYM also had potent MMP inhibition with K i values at nM concentrations. Unlike GM-6001, RYM showed modest selectivity against MMP-12 among the 5 tested rhMMPs.  Tumor necrosis factor α converting enzyme (TACE) structurally belongs to the ADAM (short for a disintegrin and metalloproteinase) family of proteases, but its catalytic site is similar with that of MMPs 25 . Therefore, we evaluated RYM and GM-6001 for inhibition against TACE using rhTACE and the fluorogenic peptide substrate, Mca-Pro-Leu-Ala-Gln-Ala-Val-Dpa-Arg-Ser-Ser-Ser-Arg-NH 2 . RYM had strong TACE inhibition with a K i value of 4.9 nM, which is close to the K i value of rhMMP-2 inhibition. Interestingly, GM-6001 showed much weaker inhibition against rhTACE (K i 113.1 nM) than rhMMPs. Next, we evaluated RYM1 for its inhibition against rhMMP-12 and TACE. This showed K i values of 2.2 ± 0.5 nM (rhMMP-12) 23 and 8.6 ± 2.2 (TACE). As such, both RYM and RYM1 had similar inhibitory potency against rhMMP-12 and TACE.
To further confirm the radiolabeling product, RYM1 was also labeled with tricine and TPPTS in the presence of SnCl 2 (Fig. 6). The 99m Tc-RYM1 samples obtained from both methods had identical retention time in radio-HPLC spectra. In addition, a mixture of 99m Tc and the succinate solution without RYM1 was also heated at 90 °C for 10 min. Radio-HPLC analysis showed only one major peak of [ 99m Tc]-colloid or [ 99m Tc]-co-ligand   product at 3.3 min but there was no radioactive peak at 13.8 min. Combined, these results confirmed the successful labeling to give 99m Tc-RYM1. One advantage of HYNIC ligand is to allow the use of different co-ligands such as TPPTS, tricine, and EDDA to give various 99m Tc labeling analogs of different hydrophilicity [30][31][32] . Therefore, we did 99m Tc-labeling of RYM1 in the presence of SnCl 2 and tricine to give a 99m Tc-labeled product which was expected to be 99m Tc-RYM1 analog with two tricine co-ligands, i.e. 99m Tc-RYM1 (tricine) 2 (Fig. 6). To this end, a solution of 5 µL/1 mM in DMSO, 99m Tc, and tricine-Sn solution was heated at 90 °C for 20 min to give the 99m Tc-labeling analog. Radio-HPLC showed the desired radioactive peak at 15.5 min in high radiochemical purity and yield (98%).
Stability studies. The resulting 99m Tc-RYM1 was monitored by the radio-HPLC analysis for stability over 24 h post-labeling. As shown in Fig. 7, both 99m Tc-RYM1 and RP805 showed good radiochemical stability in saline buffer at room temperature over 24 h. Both also exhibited good stability in murine blood samples incubated at 37 °C for 2 h (Fig. 7). However, the 99m Tc-labeled analog containing two tricine co-ligands, i.e., 99m Tc-RYM1(tricine) 2 , showed low stability in saline. The desired peak at 15.5 min was decreased to ~90% after keeping at room temperature for 8 h (Fig. 7f). A new peak at 5.6 min was increased, which might be related with free 99m Tc release from 99m Tc-RYM1(tricine) 2 , as suggested by radio-HPLC analysis of the 99m TcO 4 − solution under the same condition. Based on the ratio of the two radioactive peaks, 90% of 99m Tc-RYM1(tricine) 2 was decomposed when kept at room temperature for 24 h. LogP measurements. Compounds RYM and RYM1 have cLogP values of −0.23 and 0.50, respectively.
Accordingly, RYM1 has increased hydrophilicity compared with RP805 precursor (cLogP 3.65). The two 99m Tc labeled compounds, 99m Tc-RYM1 and RP805 were evaluated for their partitions between n-octanol and water or Tris buffer (pH 7.4) ( Table 2) 33 . The negative Log values indicated that both 99m Tc-labeled compounds were water soluble. Compared with RP805, the lower partition coefficient and shorter retention time in radio-HPLC indicated the improved hydrophilicity of 99m Tc-RYM1, as expected.

Effects of 99m
Tc-labeling on MMP binding. Next, we evaluated the effect of 99m Tc-labeling on RYM MMP inhibition, using rhMMP-12 as a representative enzyme. A solution of 99m Tc-RYM1 prepared from 2 µg of RYM1 was decayed at −80 °C for two days, lyophilized, and re-dissolved in saline for MMP inhibition assays. The decayed solution (which contains a mixture of the unlabeled precursor (RYM1), the added co-ligands, and the decayed products 99m Tc-RYM1) showed similar MMP inhibition potencies with RYM1 based on the fluorescence change kinetics, i.e. relative fluorescence units per min (RFU/min), suggesting that the labeling procedure had no significant effect on MMP binding (Fig. 8).
Tissue binding ex vivo. We have shown that the lungs of lung-specific interleukin (IL)-13 transgenic (tg) [Club cell 10-kDa protein (CC10)-IL-13 Tg] mice, a pre-clinical model of pulmonary pathology, express elevated MMP activity in conjunction with tissue remodeling and inflammation 20 . Therefore, as a prelude to in vivo imaging studies 23 , we evaluated binding of 99m Tc-RYM1 to lung tissue homogenates of CC10-IL-13Tg mice (n = 2). Consistent with the MMP-binding property of 99m Tc-RYM1, there was considerable retention of this tracer, which was reduced by 87% upon co-incubation with an excess of RYM, demonstrating its binding specificity in the lung tissue ( Fig. 9).

Discussion
We have designed, synthesized, and evaluated a new zwitterionic anti-succinate-based macrocyclic hydroxamate analog RYM, its HYNIC conjugate RYM1 and related 99m Tc labeled analog 99m Tc-RYM1. The compounds, which have good water solubility and nanomolar inhibition in vitro against a panel of rhMMPs and TACE, represent a new category of dual MMP/TACE inhibitors. Tc-RYM1 and RP805; (b) stability of 99m Tc-RYM1 in saline; (c) stability of RP805 in saline; (d) stability of 99m Tc-RYM1 in blood; (e) stability of RP805 in blood; (f) stability of 99m Tc-RYM1(tricine) 2 in saline. A.U.: arbitrary units.

LogP
Octanol/Tris* Octanol/water Table 2. LogP values of 99m Tc-RYM1 and RP805. The data represent the mean ± SD of 2 independent experiments, each performed with triplicate determinations. *pH 7.4. Hydroxamate-based MMP inhibitors such as GM-6001, batimastat (BB-94) and marimastat (BB-2516) bind to the zinc ion of the catalytic domain of activated MMPs for potent broad-spectrum inhibition 2,3,34 . The clinical use of these inhibitors is hampered by toxicity and side effects, which may stem from lack of targeting specificity and limited selectivity. Accordingly, other non-hydroxamate and hydroxamate MMP inhibitors with improved MMP selectivity have been developed, and are undergoing preclinical and clinical evaluation 9,28,35-38 . Despite the limitations of hydroxamate inhibitors as therapeutic agents, these compounds are of value in molecular imaging, which is governed by a different set of criteria. Accordingly, we have used RP805, a 99m Tc-labeled hydroxamate tracer, to image MMP activation by SPECT imaging in pre-clinical models of cardiovascular and pulmonary pathology without any evidence of toxicity 17,[19][20][21][22][39][40][41] . Prior to this work, RP805 was the only 99m Tc-labeled imaging agent that has shown promise in imaging MMP activation in cardiovascular pathology. However, RP805 has a relatively long circulation time, which could be a barrier to vascular imaging. Indeed, given the physical proximity of the blood pool and vessel wall as well as the small size of the arterial wall, imaging in atherosclerosis, aneurysm and other vascular pathologies is critically dependent on the contrast between their signals. A long clearance time would require delayed imaging which can adversely affect the contrast and signal to background ratio. In addition, the limited water-solubility of RP805 precursor is a barrier for establishing the specificity of the MMP signal in blocking studies. This is a key issue, as most cardiovascular pathologies are associated with inflammation and enhanced tissue permeability that promote non-specific uptake of the tracers. Therefore, it is critical to demonstrate signal specificity for further development and translation. We hypothesized that novel hydroxamate-based MMP inhibitors with distinct physicochemical properties, favorable pharmacokinetics and better in vivo targeting could be attractive candidates for development of alternative tracers with improved imaging characteristics. Given the paucity of promising MMP-targeting imaging agents and aforementioned limitations of RP805, the development of such rationally-designed, improved MMP tracers would have a major impact on cardiovascular medicine with paradigm-shifting potential.
Water solubility or hydrophilicity is a determinant of a drug governing its formulation, pharmacokinetics and in vivo performance. Many MMP inhibitors such as GM-6001 and batimastat (BB-94) have very poor water solubility, and it is reported that glycine incorporation at P3′ is a critical structural component to achieve both good in vitro and oral activity. To develop novel water soluble MMP inhibitors, we first calculated the cLogP values GM-6001 and several related compounds using ChemDraw Professional 16.0 (Fig. 10). Compared with GM-6001 (cLogP: 0.89), the RP805 parent macrocyclic hydroxamate amide analog (cLogP: 1.42) and its glycine-containing analog (cLogP: 0.85), incorporation of an Arg side chain was expected to considerably increase the hydrophilicity of the resulting novel analogs. Indeed, both RYM (cLogP: −0.23) and RYM1 (cLogP: 0.50) had relatively low cLogP values. In addition, the positively charged guanidine of Arg and negatively charged hydroxamate groups displayed around the anti-succinate-based macrocycle scaffold show the structural characteristics of a zwitterionic molecule, a feature which may improve in vivo distribution and specific targeting [42][43][44] . Therefore, as a novel approach to improving water-solubility and biodistribution, we synthesized several new Arg-containing macrocyclic hydroxamate analogs, including RYM, its HYNIC conjugate (RYM1) and related 99m Tc-labeled analog, 99m Tc-RYM1. Macrocyclic hydroxamate MMP inhibitors can be synthesized starting from anti-succinate derivatives, as reported previously 24,25 . To synthesize new Arg-containing analogs, we used alternative multiple protection and de-protection strategies in 13   into hydroxamic acid, and (4) HYNIC conjugation. As expected, the Arg-containing macrocyclic hydroxamate MMP inhibitors, RYM and RYM1, showed good water solubility (10 µg/µL). Furthermore, similar to RP805 precursor 16,20 , the compounds displayed nanomolar inhibition in vitro against a panel of rhMMPs including rhMMP-2, rhMMP-7, rhMMP-9, rhMMP-12, and rhMMP-13, as well as TACE. The results confirmed our hypothesis that the Arg at P3′ can improve the water solubility greatly but does not perturb the binding of hydroxamate group to the catalytic Zn 2+ of rhMMPs. K i values of RYM against 5 rhMMPs showed the relative inhibition potency in the following order: rhMMP-12 > rhMMP-2 > rhMMP-7~rhMMP-13 > rhMMP-9. Albeit not high, the rhMMP-12 selectivity of RYM over the other rhMMPs is remarkable compared with GM-6001, one of non-selective potent hydroxamate MMP inhibitors. We deduce that the selectivity may be ascribed to the macrocyclization between α substituent and P2′ substituent which may favor for rhMMP-12 especially 45 . Both RYM and RYM1 had similar K i values against rhMMP-12, indicating the HYNIC motif and related linker might not impact the MMP binding significantly.
Because of the structural similarity between the ADAM and MMP families of proteases, we evaluated the binding of RYM to TACE (ADAM17), as a representative ADAM 25 . TACE is often upregulated and activated (and implicated) in inflammatory disorders, where MMP activation also plays a key role 46 . Accordingly, it is reasonable to speculate that by targeting both TACE and MMP activation, RYM-based tracers may prove to be more effective in tracking inflammation in vivo. Direct comparison of these tracers (and selectivity toward other ADAMs) would require dedicated sets of studies that are beyond the scope of this report.
HYNIC has been widely used as a chelator for 99m Tc-labeling. It needs two co-ligands such as TPPTS and tricine or two tricine [30][31][32] . As described for RP805, RYM1 was conveniently labeled with 99m Tc in the presence of both TPPTS and tricine in a formulation to form the stable 99m Tc-labeling product, 99m Tc-RYM1, in high radiochemical purity and yield. The labeling can also be achieved in the presence of SnCl 2 to give 99m Tc-RYM1 by an alternative method. We also confirmed two co-ligands i.e. TPPTS and tricine led to higher radiochemical stability than two tricine co-ligands. Importantly, compounds RYM1 and the decayed solution of 99m Tc-RYM1 displayed a similar inhibition profile against rhMMP-12 enzymatic activity, indicating the 99m Tc-labeling procedure does not have a significant effect on key macrocyclic hydroxamate motif, its MMP binding and inhibition. Furthermore, 99m Tc-RYM1 exhibited specific binding with lung tissue from IL-13 transgenic mice ex vivo, which express high levels of MMP activity 20 .
We expected the structural differences between RP805 and 99m Tc-RYM1would lead to different biodistribution, pharmacokinetics, and targeting localization. This was empirically demonstrated in murine models, where 99m Tc-RYM1 showed faster blood clearance with residual blood activity reduced by more than 50% at 1 h post-injection, and significantly higher specific uptake in several organs relative to RP805 23 . Importantly, in these studies the MMP signal in aneurysm in vivo correlated with MMP activity, as determined by zymography ex vivo. Similar to aneurysm, MMPs play a key role in the pathogenesis of several pulmonary diseases. We confirmed tissue binding of RYM1 and its specificity in IL-13 Tg lungs, where there is considerable tissue remodeling and inflammation 20 . This is associated with upregulation of several members of MMP family, including MMP-12 and -13, and enhanced MMP-2 and MMP-12 activity. We have shown the feasibility of MMP-targeted imaging with 99m Tc-RP85 in this model and demonstrated that the MMP signal correlates with tissue inflammation 20 . A future study comparing 99m Tc-RYM1 and 99m Tc-RP805 imaging in this model is of interest. However, it is impossible to determine to what extent each specific MMP (or ADAM) contributes to 99m Tc-RYM1 (or 99m Tc-RP805) binding, as this would require highly selective inhibitors that are currently unavailable. Similarly, MMP gene deletion (even if the animals were available) could alter the lung biology and confound the interpretation of the studies.
Combined, our data suggest that 99m Tc-RYM1 is a suitable imaging agent for molecular imaging of MMP activation in vivo, especially in vascular pathologies, where rapid blood clearance is critical. While beyond the scope of this manuscript, based on these promising data we are currently working on securing permission for first in-human trial of this tracer. The Arg-containing macrocyclic hydroxamate MMP inhibitors can also serve as lead compounds for structural modification and further development to provide analogs with different linkers and metal chelators for various imaging and therapeutic applications.

Summary.
To address the limitations of existing MMP imaging agents, we have designed, synthesized, and evaluated a new Arg-containing anti-succinate-based macrocyclic hydroxamate analog (RYM), its HYNIC conjugate (RYM1) and related 99m Tc labeling analog ( 99m Tc-RYM1). Both RYM and RYM1 displayed nanomolar potent inhibition in vitro against a panel of rhMMPs and TACE. RYM1 allows for facile efficient 99m Tc-labeling to give its 99m Tc-labeled analog, 99m Tc-RYM1 in a high radiochemical yield. As a new generation MMP-targeted tracer, 99m Tc-RYM1 exhibits specific binding with inflamed lung tissue ex vivo. These results, along with favorable in vivo imaging studies reported elsewhere 23 , support further development of such water soluble Arg-containing macrocyclic hydroxamate MMP inhibitors with improved physicochemical properties and favorable pharmacokinetics for molecular imaging and potentially, therapeutic applications.

Materials and Methods
All methods were carried out in accordance with relevant guidelines and regulations.

Intermediate 2 (I-2).
To a stirred solution of I-1 (9.0 g, 24.9 mmol) in 30 mL anhydrous THF, at 0 °C, was added dropwise 9-BBN in THF (200 mL, 100 mmol) over a period of 30 min and allowed to stir at room temperature overnight. 10 mL water was added dropwise after the solution was cooled in an ice bath. A solution of 9.9 g NaOAc in water (30 mL) was added, followed by the addition of 30% H 2 O 2 (30 mL) dropwise. The mixture was stirred at room temperature for 60 min and concentrated under vacuum. The resulting aqueous solution was extracted with ethyl acetate (20 mL × 4). The combined ethyl acetate solution (80 mL) was washed with 1 N HCl (20 mL × 2), water (20 mL × 2), and brine (20 mL × 2). After it was dried over anhydrous MgSO 4 , the solution was filtered and concentrated under vacuum. The resulting residue was purified by silica gel chromatography (silica gel, 20% ethyl acetate/hexane) to give the title compound

Intermediate 3 (I-3).
To a stirred solution of I-2 (7.0 g, 19.0 mmol) and CBr 4 (12.6 g, 38 mmol) in anhydrous DCM (30 mL), was added Ph 3 P (10.4 g, 40 mmol) in small portions for 30 min. The resulting mixture was stirred at room temperature for another 2 h, followed by adding 30 mL hexanes. The resulting mixture was transferred to a short column of silica gel for quick elution using DCM and hexane (1:1). The desired fractions were combined and concentrated to get a crude product which was further purified by column chromatography (silica gel, 2% ethyl acetate/hexane).

Intermediate 4 (I-4).
A mixture of I-3 (5.0 g, 11.3 mmol) and Pd/C (2.0 g, 10%, wet) in 30 mL methanol was stirred under hydrogen gas at room temperature for 30 min. The mixture was filtered, followed by washing the Pd/C with methanol (5 mL × 4). The combined methanol filtrate was concentrated and the product (3.

Intermediate 5 (I-5).
To a stirred solution of I-4 (3.0 g, 8.5 mmol), HOBT (1.7 g, 12.7 mmol), and Tyr-OBz (3.4 g, 12.7 mmol) in 20 mL anhydrous DMF cooled in an ice bath, was added EDCI (2.4 g, 12.7 mmol) and stirred at room temperature for 2.5 h. The resulting mixture was concentrated under high vacuum, re-dissolved with ethyl acetate (100 mL), and washed with water, 1 N HCl, water, 1 N Na 2 CO 3 solution, water, and brine. The ethyl acetate solution was dried over anhydrous MgSO 4 , filtered, and concentrated under vacuum. The resulting residue was purified by chromatography (silica gel, 20% ethyl acetate/hexane). 3 (I-6). To a stirred solution of Cs 2 CO 3 (3.2 g, 23.4 mmol) in 500 mL anhydrous acetonitrile at 60 °C, was added dropwise a solution of I-5 (3.0 g, 5.3 mmol) in 50 mL acetonitrile, over a period of 1 h. The resulting mixture was stirred at 60 °C for another 3 h and concentrated under vacuum. The product was re-dissolved with ethyl acetate and filtered, followed by washing the solid with ethyl acetate for 5 times (10 mL × 5). The combined ethyl acetate filtrate was washed with 1 N HCl solution, water, and brine. The ethyl acetate solution was dried over anhydrous MgSO 4 , filtered, and concentrated. The resulting residue was purified by silica gel chromatography using (silica gel, 20% ethyl acetate/hexane) to give the product (1.6 g, 60%

Intermediate 7 (I-7).
A mixture of I-6 (5.0 g, 9.5 mmol), 10% Pd/C (2.0 g), and HCOONH 4 (5 g) in 15 mL CH 3 OH was stirred at room temperature for 2-3 h until the hydrogen gas evolved was observed. The mixture was further stirred for another 20 min and filtered, followed by washing the Pd/C with CH 3 OH (5 mL × 4). The methanol filtrate was concentrated and acidified with 1 N HCl solution to get a residue which was extracted with ethyl acetate. The ethyl acetate solution was washed with water (10 mL × 2) and brine (10 mL × 2), dried over anhydrous MgSO 4 , filtered, and concentrated. The product (3.7 g, 90%) was obtained and used for the next step without further purification. LC-MS: Observed [MH] + 434.2 and [MNa] + 456.2.

Intermediate 8 (I-8).
A mixture of Fmoc-Arg(Mtr)-OH (1.2 g, 1.97 mmol), HOBT (0.27 g, 2.0 mmol), CH 3 NH 2 2 M in THF(2 mL, 4.00 mmol) was dissolved in 5 mL anhydrous DMF and cooled at 0~5 °C in an ice bath, followed by adding EDCI (600 mg, 3.1 mmol). The mixture was stirred at room temperature overnight and concentrated under high vacuum. The residue was triturated with 1 N HCl, the solid obtained was filtered, and further washed with 1 N HCl, 5 N Na 2 CO 3 , and H 2 O. The solid product was dried to give 1. Intermediate 12 (I-12). A mixture of I-11 (200 mg, 0.26 mmol), HOAT (86.0 mg, 0.6 mmol), HATU (236.0 mg, 0.6 mmol), and DIEA (16.0 mg, 0.9 mmol) was dissolved in 4 mL anhydrous DMF. The mixture was stirred at room temperature for 20 min, followed by adding TBDMS-ONH 2 (137.0 mg, 0.93 mmol). The mixture was further stirred at room temperature overnight and concentrated. The residue was triturated with 1 N HCl, the solid obtained was filtered, and washed with 1 N Na 2 CO 3 , water and brine. The product was further purified by flash column chromatography to give 100.0 mg (50%) of the title compound.
Target compound, RYM. I-12 (50 mg, 0.065 mmol) was dissolved in a mixture of 9.0 mL TFA, 0.5 mL H 2 O, 0.25 mL TIS, and 0.25 g phenol. The mixture was stirred at room temperature overnight and concentrated under vacuum. The residue was added into 10 mL cooled diethyl ether. The precipitated product was collected by centrifugation and purified by semi-preparative HPLC using an eluent of aqueous acetonitrile. The desired fractions were collected and lyophilized to give the title compound   (20 mL), and water (20 mL) was refluxed at 95~100 °C for 4 h. The mixture was concentrated under vacuum and the residue was dissolved in 40 mL water. The resulting solution was acidified with 1 N HCl to reach a pH 5~5.5. The solution was kept in a refrigerator overnight. The precipitated solid was collected by filtration, washed with cold water (10 mL × 2) and ether (10 mL × 2). The solid was dried to give 8.5 g (87%) of the title compound. ES-MS: Observed [MH] + 154.1.

Boc-HYNIC.
A mixture of 4.0 g HYNIC and 8.0 g Na 2 CO 3 in 60 mL of 1,4-dioxane and 60 mL of water was cooled in an ice bath. 8.2 g of Boc 2 O was added, and stirred at room temperature for 5 h. The mixture was concentrated and the residue was acidified with 1 N HCl solution, the precipitated solid was collected by filtration and dried to give the title compound.
LogP measurement. An aliquot (10.0 µL/~100 µCi) of RP-805 or 99m Tc-RYM1 was added into a mixture of 500.0 µL octanol and 500.0 µL water or tris solution (1 M, pH 7.4) in a 1.5 mL Eppendorf vial. The vial was mixed by vortex violently for 3 min and then centrifuged at 13,000 rpm for 5 min. The two phases i.e. octanol and aqueous layers separated were obtained carefully using a micropipette. The two solutions were centrifuged at 13,000 rpm for 5 min. The samples (10.0 µL) taken from the octanol layer and the aqueous layer were measured for radioactivity in triplicates by r-counter (PerkinElmer, Waltham, MA). The logP octanol/water and logP octanol/tris values were calculated by the following equation: logP = log (decay-corrected radioactivity in octanol sample/ decay-corrected radioactivity in aqueous sample).
Tissue binding. Ex vivo binding of 99m Tc-RYM1 to biological tissue was evaluated in lung tissue homogenate of CC10-IL-13Tg mice (n = 2) showing elevated basal MMP activity 20 . Lung tissue (247 ± 85 µg, in duplicates was incubated with 3 nM of 99m Tc-RYM1 for 60 min at 37 °C with or without the preincubation of 20 µM of RYM. After tracer incubation, tissue homogenate was spin down and washed 3 times with PBS, before resuspension in protein lysis buffer, gamma-counting (Wizard2, PerkinElmer) and protein concentration was measured (Pierce BCA Protein Assay Kit, Thermo Fisher Scientific and Multiskan Ascent Microplate Photometer, Thermo Labsystems) in order to determine the amount of tissue used. All experimental protocols were approved by Yale University and VA Connecticut Institutional Animal Care and Use Committees.
Data availability. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.