Radiometal-Containing Aryl Diazonium Salts for Chemoselective Bioconjugation of Tyrosine Residues.

Tyrosine is an attractive target for chemo- and site-selective protein modification. The particular chemical nature of tyrosine residues allows bioconjugation chemistry with reactive aryl diazonium salts via electrophilic aromatic substitution to produce diazo compounds. In this work, we describe the preparation of 64Cu- and 68Ga-labeled 1,4,7-triazacyclononane-1,4,7-triacetic acid (NOTA)-diazonium salts as building blocks for azo coupling chemistry with tyrosine and tyrosine-containing peptides and proteins under mild conditions. 2-S-(4-aminobenzyl)-1,4,7-triazacyclononane-1,4,7-triacetic acid (p-NH2-Bn-NOTA) was used to form the corresponding 64Cu- and 68Ga-labeled complexes, followed by diazotization with NaNO2 in the presence of HCl. 64Cu- and 68Ga-labeled NOTA complexes were prepared in high radiochemical yields >80% starting from 20 μg of p-NH2-Bn-NOTA. Conversion of p-NH2-Bn-NOTA complexes into diazonium salts followed by azo coupling with l-tyrosine afforded 64Cu- and 68Ga-labeled tyrosine in radiochemical yields of 80 and 56%, respectively. Azo coupling with tyrosine-containing hexapeptide neurotensin NT(8-13) afforded 64Cu- and 68Ga-labeled NT(8-13) in radiochemical yields of 45 and 11%, respectively. Azo coupling of 64Cu-labeled NOTA-diazonium salt with human serum albumin (HSA) gave 64Cu-labeled HSA in radiochemical yields of 20%. The described azo coupling chemistry represents an innovative and versatile bioconjugation strategy for selective targeting of tyrosine residues in peptides and proteins.

Introduction

The modification of peptides and proteins is frequently accomplished by site-directed bioconjugation chemistry with canonical amino acid residues with a strong dominance of lysine and cysteine. Over the last decade, various chemoselective bioconjugation strategies, including biorthogonal click chemistry, have gained much attention as versatile bioconjugation tools for the synthesis of radiolabeled peptides and proteins for positron emission tomography (PET) and single-photon emission computed tomography (SPECT).[1−3]

Tyrosine is a particularly interesting amino acid for the chemoselective modification of peptides and proteins because of its low abundance, comprising only 3% of the total amino acid content of proteins.[4] Initial strategies for site-directed tyrosine modification involved three-component Mannich-type reactions,[5] ene-type reactions with cyclic diazocarboxamides,[6] Pd-mediated allylations,[7] cerium-promoted one-electron oxidative couplings,[8] and single-electron oxidative couplings.[9] Another highly versatile bioconjugation strategy for the site-directed modification of tyrosine residues is azo coupling chemistry with aryl diazonium salts (Figure ).[4,10]

Figure 1

Azo coupling chemistry of tyrosine residues with aryl diazonium salts.

Azo coupling chemistry with radionuclides was first described for the labeling of proteins with 111In for in vivo scintillation studies using acyclic chelators such as ethylenediaminetetraacetic acid (EDTA).[11] More recently, azo coupling was used for the incorporation of 111In into tyrosinamide-containing polymers for subsequent pharmacokinetic studies.[12] Incorporation of 111In was performed using a postconjugation labeling strategy. Bifunctional chelator 2-S-(4-aminobenzyl)-diethylenetriamine pentaacetic acid (p-NH2-Bn-DTPA) was converted into the respective diazonium salt followed by azo coupling to tyrosinamide-containing polymers prior to the labeling with 111In.[12]

In this work, we describe the preparation of 64Cu- and 68Ga-labeled 1,4,7-triazacyclononane-1,4,7-triacetic acid (NOTA)-diazonium salts as versatile building blocks for azo coupling chemistry with tyrosine residues under mild conditions. The first description of a tyrosine targeting strategy in a virus with diazonium salts was reported in 2004 by Francis et al.[13] The group described the chemoselective modification of tyrosine residues on the interior surface of bacteriophage MS2 using aryl diazonium salts in an electrophilic aromatic substitution reaction. Inspired by this first report of targeting tyrosine residues in a virus with diazonium salts, our group recently used this bioconjugation strategy for the synthesis of 64Cu-labeled tobacco mosaic virus as the nanoparticle platform for cancer imaging with PET.[14] In the present work, we have demonstrated the versatility of the azo coupling strategy with 64Cu- and 68Ga-labeled NOTA diazonium salts for the radiolabeling of l-tyrosine, tyrosine-containing peptide neurotensin NT(8–13), and protein human serum albumin (HSA).

Results

Chemistry and Radiochemistry

Many examples of radiometal-labeled peptides and proteins use the postconjugation labeling approach with peptides and proteins decorated with a chelator such as NOTA or 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid. However, numerous attempts to decorate l-tyrosine or NT(8–13) with NOTA using NOTA-diazonium salts according to the postconjugation labeling approach were not successful (Table S1). No product formation was observed based on liquid chromatography–mass spectrometry (LC–MS) analysis of the crude reaction mixture and LC–MS analysis of peaks collected after high-performance LC (HPLC) separation. However, using 3-fluoroaniline as a model compound instead of p-NH2-Bn-NOTA, desired products were observed suggesting that the chemical nature of macrocycle NOTA prevented the formation of the respective diazonium salt or subsequent azo coupling reaction with l-tyrosine (Table S2).

However, when natCu- and natGa-labeled p-NH2-Bn-NOTA were used in the reaction with l-tyrosine according to a preconjugation labeling approach, expected diazo compounds were formed (Table S3). A similar trend was observed with reactions involving tyrosine-containing peptide NT(8–13) (Table S4).

Radiosynthesis of 64Cu- and 68Ga-labeled NOTA diazonium salts 4 and 5 was performed starting from [64Cu]CuCl2 and [68Ga]GaCl3 in 0.25 M NH4OAc buffer (pH 5.5) in the presence of 20 μg of p-NH2-Bn-NOTA 1, followed by in situ diazotization with NaNO2 with HCl at 4 °C (Figure ).

Figure 2

Radiosynthesis of 68Ga- and 64Cu-labeled aryl diazonium salts 4 and 5.

Chelation of p-NH2-Bn-NOTA with 64Cu and 68Ga gave complexes 2 and 3 in high radiochemical purity >94% based on radio-thin-layer chromatography (TLC) analysis (Figure S15). In situ formed 64Cu- and 68Ga-labeled NOTA diazonium salts 4 and 5 were used without purification for subsequent azo coupling chemistry.

Bioconjugation of radiolabeled aryl diazonium salts 4 and 5 with tyrosine residues was tested with l-tyrosine, hexapeptide neurotensin NT(8–13), and protein HSA. In the first set of reactions, conjugation of diazonium salts 4 and 5 with l-tyrosine in borate buffer at pH 8–9 gave diazo compounds in 6 and 7 in 56 and 80% radiochemical yields, respectively (Figure ).

Figure 3

Azo coupling of radiometal diazonium salts 4 and 5 with l-tyrosine.

Next, 64Cu- and 68Ga-labeled diazonium salts were coupled with tyrosine-containing hexapeptide neurotensin NT(8–13). The reaction of diazonium salts 4 and 5 with 200 μg of NT(8–13) gave respective radiometal-labeled NT(8–13) derivatives 8 and 9 in 45 and 20% radiochemical yields, respectively, after HPLC purification (Figure ).

Figure 4

Azo coupling of diazonium salts 4 and 5 with neurotensin NT(8–13).

The mild reaction conditions for the azo coupling reaction are particularly suitable for radiolabeling experiments with proteins. We selected HSA as a model protein to test the azo coupling bioconjugation with radiometal-containing diazonium salts.

HSA contains 18 tyrosine residues, which makes the protein a valid candidate for azo coupling chemistry. Because of the long biological half-life of HSA compared to tyrosine and NT(8–13), we performed azo coupling with 64Cu-labeled diazonium salt 5. The reaction of diazonium salt 5 with 100 μg of HSA in borate buffer at pH 8–9 at 4 °C for 15 min gave 64Cu-labeled HSA 10 in radiochemical yields of 20% after purification with size-exclusion chromatography (SEC) (Figure ).

Figure 5

Radiolabeling of HSA.

The collected fractions containing 64Cu-labeled HSA 10 were analyzed with radio-TLC and radio-sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), confirming a radiochemical purity greater than 95%. Coomassie staining and phosphorimages clearly showed the expected 66 kDa band of HSA (Figure S17).

Challenge Experiments with Histidine

The chemoselectivity of azo coupling chemistry with radiometal aryl diazonium salts toward tyrosine residues was assessed with histidine challenge experiments. The competitive azo coupling chemistry of 64Cu- and 68Ga-labeled diazonium salts 4 and 5 in the presence of equimolar amounts of l-tyrosine and l-histidine in borate buffer was carried out at different pH values (7, 8, and 9). Ratios of labeled tyrosine versus histidine were determined by radio-HPLC. 64Cu- and 68Ga-labeled histidine were predominantly observed at pH 7, whereas a mixture of 64Cu- and 68Ga-labeled tyrosine and histidine was found at pH 8 and 9 (43% labeled tyrosine versus 17% labeled histidine). Histidine challenge experiments at pH 9 were also carried out with the natCu-labeled diazonium salt resulting in the formation of 64% azo-coupled tyrosine versus 36% azo-coupled histidine (Figure S18). These results confirmed better chemoselectivity of azo coupling reaction with tyrosine at pH 9, whereas azo-coupled histidine preferentially formed at a lower pH of 7.

Determination of IC50 Values

Determination of IC50 values was carried out with a radiometric competitive binding assay with different concentrations of natCu- and natGa-containing NT(8–13) in neurotensin receptor 1 expressing HT-29 cells (n = 3).[15,16] [125I]Tyr3-neurotensin at a molar activity of 81.4 GBq/μmol (PerkinElmer) was used as the radiotracer. When compared to unmodified neurotensin NT(8–13), natGa- and natCu-containing NT(8–13) displayed subnanomolar IC50 values of 0.04 and 0.16 nM, respectively. Parent peptide NT(8–13) gave an IC50 of 0.03 nM in the assay (Table ).

Table 1

IC50 Values of Neurotensin NT(8–13), natCu-Labeled Neurotensin NT(8–13), and natGa-Labeled Neurotensin NT(8–13)

peptide	IC50 value (nM)
NT(8–13)	0.03
Cu-NOTA-NT(8–13)	0.16
Ga-NOTA-NT(8–13)	0.04

PET Imaging

The biodistribution of 64Cu-labeled HSA 10 was studied in normal BALB/c mice using PET imaging. Radiotracer [64Cu]Cu-HSA (5 MBq) in 200 μL of phosphate-buffered saline (PBS) was injected into the tail vein of BALB/c mice. Mice were imaged using dynamic data acquisition for 2 h followed by a static scan at 24 h. The PET image showed high radiotracer uptake in the heart and slow blood pool clearance over the first 2 h p.i. Radioactivity accumulation in the liver and spleen was visible at 24 h p.i. (Figure ).

Figure 6

Coronal PET images of 64Cu-labeled HSA in normal BALB/c mice.

Discussion

Innovative and versatile bioconjugation strategies are on high demand for the chemoselective incorporation of radionuclides into peptides and proteins for molecular imaging and radiotherapy. Radiometals are predominantly incorporated into peptides and proteins through the postconjugation labeling concept. Popular strategies involve radiolabeling of chelator-decorated peptides and proteins, including antibodies and other immunoconjugates, with radiometals for PET and SPECT imaging such as 68Ga, 64Cu, and 111In or those for radiotherapy such as 177Lu and 225Ac. Preconjugation labeling strategies are especially popular in 18F chemistry as demonstrated with numerous prosthetic groups used for the labeling of peptides with 18F. Chemoselective incorporation of radionuclides via metal chelators or prosthetic groups into peptides and proteins mainly exploits bioconjugation chemistry with nucleophilic groups such as amine and thiol as present in lysine and cysteine residues. A rather underutilized amino acid for peptide and protein conjugation is tyrosine. Aryl diazonium salts are known to be mild electrophiles which undergo electrophilic substitution reactions with aromatic compounds, such as the phenol side chain of tyrosine, which are activated by electron-donating substituents such as hydroxyl and amino groups.

The present work describes a systematic study on the scope and limitations of using 68Ga- and 64Cu-labeled NOTA diazonium salts for preconjugation labeling of peptides and proteins. Attempts to decorate tyrosine and NT(8–13) with NOTA via diazo coupling with NOTA diazonium salts for postconjugation labeling with 68Ga and 64Cu failed. The reason for the unsuccessful azo coupling with NOTA diazonium salts without the coordinating metal remains unclear.

However, the chemical nature of macrocycle NOTA containing three free carboxylic acid residues seems to prevent successful azo couplings as the reaction proceeded smoothly with 3-fluoro-aniline as a model compound and natCu- and natGa-coordinated NOTA.

64Cu- and 68Ga-labeled NOTA diazonium salts were prepared in situ through the conversion of the respective NOTA–amine complexes. Azo coupling chemistry with l-tyrosine, NT(8–13), and HSA proceeded rapidly under mild aqueous conditions in good radiochemical yields, making the labeling strategy particularly compatible with the sensitive nature of delicate biomolecules such as proteins. The chemistry is highly versatile and flexible as demonstrated with the labeling of amino acid tyrosine, hexapeptide NT(8–13), and protein HSA.

The chemoselectivity of azo coupling reactions with diazonium salts toward tyrosine and histidine residues was recently studied.[17−19] Results are varying greatly, but several studies that described a high degree of tyrosine site-specificity in azo coupling reactions were reported even in the presence of competing amino acid targets such as histidine.[19] Other studies demonstrated more selectivity for tyrosine at pH 8.5 but increased selectivity for histidine at pH 9.1 and higher.[20]

In contrast, a more recent study reported more selectivity for tyrosine residues over histidine at pH 4–10, with a significant increase in selectivity at pH 7.[21] In the present work using radiometal-containing aryl diazonium salts, challenging experiments with tyrosine and histidine suggested a trend for histidine labeling occurring predominantly at pH 7, whereas tyrosine labeling predominantly occurred at higher pH in the range of 8–9. This trend may be attributed to the difference in pKa values of the ionizable side chains of these amino acid residues. The pKa of histidine (6.0) is lower than the pKa of tyrosine (10.5); therefore, there would be a greater proportion of deprotonated histidine in solution at pH 7 to react in the azo coupling reaction. When both amino acids are deprotonated, tyrosine preferentially reacts in the azo coupling reaction at pH 9, and there would likely be a sufficient proportion of deprotonated tyrosine in solution to compete with histidine in the azo coupling reaction. Based on these results and supporting reports in the literature, coupling of radiometal-containing aryl diazonium salts with peptides and proteins should be carried out at pH ≈ 9 to achieve a favorable tyrosine selectivity profile.

Modification of peptides and proteins can affect binding affinity to respective target proteins and receptors, which is a special challenge for the design and synthesis of radiolabeled peptides and proteins for targeted molecular imaging and therapy. In this line, we tested the effect of radiolabeled diazo conjugates of NT(8–13) on the binding to neurotensin receptor NTR1. Recently, we have measured several multimeric neurotensin(8–13) analogues for their binding to the NTR1 to identify suitable candidates for radiolabeling with 18F.[15,16] All tested NT(8–13) were modified at the N-terminal end of the peptide. Most peptides displayed rather low inhibitory potency compared to parent compound NT(8–13). In the present study, NT(8–13) was modified at the tyrosine residue. The radiometric competitive binding assay gave IC50 values in the subnanomolar range (0.04 and 0.16 nM) close to the IC50 value of parent compound NT(8–13) (IC50 = 0.03 nM). These data reveal that the introduction of the bulky diazo motif at the tyrosine residue had only little effect on the receptor binding. When binding affinities were compared to other neurotensin NT(8–13) analogues in the literature, which were modified at the N-terminal end of the peptide to facilitate radiolabeling with 64Cu, 68Ga, or 177Lu, IC50 and Kd values were in the nanomolar range.[22−24]

Given that IC50 values of the tyrosine-modified peptide we have described were in the picomolar range, the N-terminal modified peptides described in the literature displayed lower inhibitory potency than the tyrosine-modified peptide. The tolerance of the tyrosine residue in NT(8–13) for the introduction of large substituents, while retaining favorable receptor binding was demonstrated with various 125I-labeled neurotensin derivatives.[25]

Experiments with 64Cu-labeled HSA in normal mice further confirmed the suitability of the preconjugation labeling method with radiometal aryl diazonium salts for the preparation of radiolabeled proteins for PET imaging. The obtained biodistribution profile of 64Cu-labeled HSA is characteristic for a blood pool marker. A significant heart uptake and slow blood clearance, as well as radioactivity accumulation in the liver over time, are typical for proteins such as HSA (Figure S21). The results are comparable to the recent literature describing PET imaging with 18F-labeled rat serum albumin (RSA) and 64Cu-labeled HSA.[26,27] In a 2015 study conducted by Jagoda et al., PET imaging with 18F-labeled RSA showed uptake in the heart and peripheral vasculature, with retention up to 2.5 h postinjection.[26] These findings are comparable to those of the present study, where PET imaging with 64Cu-labeled HSA also showed uptake in the heart and arteries, with retention up to 2 h postinjection. In a 2016 study conducted by Choe et al., PET imaging with 64Cu-labeled HSA showed liver metabolism from 16 to 45 h postinjection.[27] These findings are also comparable to those of the present study, where PET imaging with 64Cu-labeled HSA also showed liver metabolism 24 h postinjection. The favorable in vivo profile of 64Cu-labeled HSA, prepared using the described diazonium-based tyrosine labeling method, further demonstrates the feasibility of this strategy for the introduction of radiometals into proteins.

Conclusions

Azo coupling is a versatile bioconjugation technique that can be applied to PET radiochemistry for the chemoselective incorporation of radionuclides onto tyrosine residues under mild conditions. Preparation of 64Cu- and 68Ga-labeled diazonium salts and their subsequent bioconjugation with tyrosine residues was carried out on the amino acid, peptide, and protein level to afford the expected azo conjugates in good radiochemical yields. Determination of IC50 values with natCu- and natGa-labeled neurotensin NT(8–13) confirmed tolerance of the tyrosine residue in NT(8–13) to accommodate bulky substituents. PET imaging studies with 64Cu-labeled HSA prepared by azo coupling chemistry showed the expected biodistribution profile, with significant uptake in the heart and slow clearance from the blood. The described preconjugation labeling strategy proceeds rapidly under mild conditions with various tyrosine-containing biomolecules making azo coupling chemistry a highly efficient and versatile labeling tool for the incorporation of radiometals into tyrosine residues of peptides and proteins.

Experimental Section

General

All chemicals were obtained from Sigma-Aldrich (St. Louis, MO, USA), with the exception of 2-S-(4-aminobenzyl)-1,4,7-triazacyclononane-1,4,7-triacetic acid (p-NH2-Bn-NOTA 1) which was obtained from Macrocyclics (Dallas, TX, USA), neurotensin(8–13) from Bachem (Torrance, CA, USA), and HSA from Millipore (Billerica, MA, USA). LC–MS data were recorded on an Agilent 6130 mass spectrometer with Agilent 1260 HPLC. 64Cu (350–400 MBq) was produced via the 64Ni(p,n)64Cu nuclear reaction in high specific activity on a biomedical cyclotron from Washington University, St. Louis, USA. It is typically delivered in 12 μL of 0.01 N HCl ([64Cu]CuCl2). A 50 mCi (1850 MBq) iThemba Laboratories 68Ge/68Ga-generator from isoSolutions Inc. (Vancouver, B.C., Canada) was used as the 68Ga source. Buffers for radiolabeling with 64Cu and 68Ga were of trace-metal grade. Radio-TLC was performed using either silica plates (TLC silica gel 60 F254, EMD Millipore, Billerica, MA, USA) or reverse phase (RP-18) plates (ALUGRAM RP18W/UV254 by Macherey-Nagel, Dueren, Germany). SEC was performed using a Bio-Rad Econo-Pac 10 DG (Bio-Rad, Mississauga, ON, Canada) or GE PD MidiTrap G-25 (GE Healthcare, Mississauga, ON, Canada). Semipreparative HPLC of radiolabeled neurotensin NT(8–13) was performed on a Gilson system (Mandel Scientific, Guelph, ON, Canada) with a 321 pump, a photodiode array detector, and a HERM Bertolt radiodetector installed with the HPLC column Phenomenex Luna 10u C18(2) 100A, 250 × 10 mm. UV absorbance was monitored at 210 and 254 nm wavelength. The mobile phase consisted of water with 0.2% trifluoroacetyl as solvent A and acetonitrile as solvent B. The phosphor imager used for autoradiography was a Typhoon 9400.

Chemical Synthesis

The synthesis of nonradioactive reference compounds was carried out according to the following general procedures:

Synthesis of natCu- or natGa-Labeled p-NH2-Bn-NOTA Complexes

natCu-Labeled p-NH2-Bn-NOTA 2

An excess of copper(II) chloride dihydrate and EDTA in 1 M ammonium acetate buffer pH 6 was incubated at room temperature for 30 min before it was added to a solution of p-NH2-Bn-NOTA 1 in 1 M ammonium acetate buffer pH 6. The reaction was incubated at 60 °C for 1 h before being incubated at room temperature for 24 h. The reaction was purified by HPLC (5% CH3CN 0–15 min, 5–30% CH3CN 15–25 min, 30–80% CH3CN 25–30 min, and 80% CH3CN 30–40 min) and characterized by LC–MS. LC–MS, ESI-positive: 469.1 [M + H]+.

natGa-Labeled p-NH2-Bn-NOTA 3

An excess of gallium trichloride and EDTA in 1 M sodium acetate buffer pH 4.5 was shaken at room temperature for 5 min before it was added to a solution of p-NH2-Bn-NOTA 1 in 1 M sodium acetate buffer pH 4.5. The reaction was incubated at room temperature for 24 h. The reaction was purified by HPLC (5% CH3CN 0–15 min, 5–30% CH3CN 15–25 min, 30–80% CH3CN 25–30 min, and 80% CH3CN 30–40 min) and characterized by LC–MS. LC–MS, ESI-positive: 475.1 [M + H]+.

General Procedure for Diazotization and Bioconjugation with Tyrosine Residues

To a solution of natCu- or natGa-labeled p-NH2-Bn-NOTA 2 and 3 in 1 M hydrochloric acid, 1 equiv of sodium nitrite was added. The reaction was periodically shaken at 4 °C for 2 min. The product was prepared in situ and used for the next step without purification.

To 1 equiv of l-tyrosine or neurotensin NT(8–13) in 1× PBS pH 7.2, the diazotized solution was added. The pH of the solution was adjusted to 9 by the addition of 4 M NaOH, and the reaction was incubated at 4 °C for 30 min. Products were purified by HPLC (5% CH3CN 0–15 min, 5–30% CH3CN 15–25 min, 30–80% CH3CN 25–30 min, and 80% CH3CN 30–40 min) and characterized by LC–MS. natCu- and natGa-labeled l-tyrosine 6 and 7 were isolated in 28 and 7% yields, respectively. LC–MS, ESI-positive: 662.1 [M + H]+. LC–MS: 667.2 [M + H]+. The natCu- and natGa-labeled neurotensin NT(8–13) 8 and 9 were isolated in 76 and 66% yields, respectively. LC–MS, ESI-positive: 1297.4 [M + H]+. LC–MS: 1302.3 [M + H]+.

Radiosynthesis

[64Cu]CuCl2 was obtained from Washington University (St. Louis, MO, U.S.A). [68Ga]GaCl3 was eluted from a 68Ge/68Ga generator using 0.6 N HCl.

General Procedure for Radiosynthesis of 64Cu- or 68Ga-Labeled Diazonium Salts 4 and 5

To a solution of p-NH2-Bn-NOTA 1 in 0.25 M ammonium acetate buffer pH 5.5, 5–10 MBq of 64Cu or 68sGa was added. The reactions were incubated at 37 °C for 15 min and monitored by radio-TLC using a 10% 1 M ammonium acetate in the methanol solvent system. The product had an Rf value of 0.3 and was used in the next step without purification. The radiochemical purity was >94%. Next, 100 μL of 1 N hydrochloric acid and 10 μL of a 2 M sodium nitrite solution were added. The reaction was shaken at 4 °C for 10 min. The radiolabeled diazonium salt product was used, in situ, for the next step without purification.

General Procedure for Bioconjugation with Tyrosine Residues Using 64Cu- or 68Ga-Labeled Diazonium Salts 4 and 5

To 100 μg of peptide NT(8–13) or protein (HSA) in 0.1 M borate buffer pH 8.8, the 64Cu- or 68Ga-labeled diazonium salt was added. The pH was adjusted to 9 with the addition of 4 M NaOH, and the reaction was shaken at 4 °C for 30 min. Peptide products were purified by HPLC (5% CH3CN 0–15 min, 5–30% CH3CN 15–25 min, 30–80% CH3CN 25–30 min, and 80% CH3CN 30–40 min) and characterized by radio-TLC and radio-HPLC, based on the retention time of the nonradioactive reference compound. Protein products were purified by SEC and characterized by radio-TLC and SDS-PAGE with Coomassie staining and radioactive scan. The 64Cu- and 68Ga-labeled l-tyrosine 6 and 7 were afforded in 56 and 80% radiochemical yields, respectively; 68Ga- and 64Cu-labeled neurotensin NT(8–13) derivatives 8 and 9 were isolated in 12 and 45% radiochemical yields, respectively; 64Cu-labeled HSA 10 was obtained in 20% radiochemical yield.

IC50 Determination

IC50 values were determined using a radiometric competitive binding assay.[15,16] [125I]Tyr3-neurotensin, with a molar activity of 81.4 GBq/μmol from PerkinElmer, was used as a radiotracer. HT-29 cells plated on tissue culture T-25/75 flasks were incubated in culture media in humidified 5% CO/95% air at 37 °C for three to five days. Cells were washed three times with Tris buffer, suspended in 8 mL of Tris buffer, and homogenized. Cell densities ranged from 4 × 106 to 7 × 106 cells/mL.

A 100 μM peptide stock solution was prepared using water, and serial dilutions were prepared using Tris buffer (50 mM Tris-HCl, 1 mM EDTA, 0.1% bovine serum albumin, and 0.5 mM o-phenanthroline, cOmplete, EDTA-free protease inhibitor cocktail, pH 7.4), with concentrations ranging from 10–5 to 10–12 M. A 100 μL aliquot of peptide solution was combined with 200 μL of cell homogenate, 500 μL of Tris buffer, and 200 μL of the radiotracer (n = 3). The solution was incubated at room temperature for 30 min. The free and bound ligands were separated by filtration, using Whatman GF/B glass fiber filters and a Brandel cell harvester (Gaithersburg, MD, USA). Filters were washed three times with Tris buffer and quantified using a gamma counter (Wizard-2, PerkinElmer). IC50 values were calculated by nonlinear regression, using sigmoidal dose-response curves from GraphPad Prism 7 software (Figure S20).

SDS-PAGE

SDS-PAGE was used to characterize 64Cu-labeled HSA. Radiolabeled proteins were visualized with a combination of Coomassie staining and a radioactive scan of the gel. To a 30 μL aliquot of each sample, 6 μL of 5× Laemmli stain with dithiothreitol was added before incubation at 95 °C for 5 min. Samples were run, along with a protein ladder, on a Bio-Rad Mini-PROTEAN TGX precast gel (10% Tris buffer) in a Bio-Rad Mini-PROTEAN Tetra Cell with 1× running buffer at 200 V (35 mA) for 40 min. The gel was exposed to a Fujifilm BAS-MS 2025 imaging plate for 45 min, and the imaging plate was scanned using a Typhoon 9400 phosphor imager. The gel was then incubated in Coomassie Brilliant Blue R-250 (Bio-Rad) for 20 min at room temperature before destaining with destaining solution.

PET Imaging Studies

All animal experiments were carried out in accordance with guidelines of the Canadian Council on Animal Care (CCAC) and approved by the local Animal Care Committee of the Cross Cancer Institute. PET experiments using a normal BALB/c mouse were carried out to determine the biodistribution profile of 64Cu-labeled HSA 10. Isoflurane in 100% oxygen (gas flow, 1 L/min) was used as a general anesthetic. Body temperature was kept constant at 37 °C. Following anesthetization, the mouse was immobilized in the prone position in the center field of view of an Inveon preclinical PET scanner (Siemens Preclinical Solutions, Knoxville, TN, USA). Radioactivity of the injection solution in a 0.5 mL syringe was measured using a dose calibrator (AtomlabTM 300, Biodex Medical Systems, New York, U.S.A.) prior to injection. The emission scan of a 120 min dynamic PET acquisition was initiated. Following a 15 s delay, 5 MBq of radiochemically pure 64Cu-labeled HSA 10 in 200 μL of PBS was injected into the tail vein. Data acquisition continued for 120 min in the 3D list mode, after the experiment list mode data were sorted into sinograms with 61 time frames (10 × 2, 8 × 5, 6 × 10, 6 × 20, 8 × 60, 10 × 120, and 9 × 300 s). In addition to the dynamic 120 min scan, another static scan was measured after 24 h p.i. with a scan duration time of 60 min. Image files were reconstructed using the maximum a posteriori reconstruction mode. Correction for partial volume effects was not performed. Image files were further processed using ROVER v2.0.21 software (ABX GmbH, Radeberg, Germany). Masks defining 3D regions of interest (ROI) were set, and the ROIs were defined by 50% thresholding. Mean standardized uptake values [SUVmean = (activity/mL tissue)/(injected activity/body weight), mL/g] were calculated for each ROI, and time-activity curves were generated using GraphPad Prism 5.0 (GraphPad Software Inc., La Jolla, CA, U.S.A.).