In Vivo Evaluation of Indium-111–Labeled 800CW as a Necrosis-Avid Contrast Agent

Reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA) unless stated otherwise. Solvents were purchased from Honeywell Riedel-de-Haën™ (Seelze, Germany). 800CW-N-hydroxysuccinimide ester (2) was purchased from Westburg BV (Leusden, The Netherlands). DOTA-PEG₄-NH₂ (3) was purchased from Chematech (Dijon, France). [¹¹¹In]InCl₃ from Mallinckrodt BV (Petten, The Netherlands). 4T1-Luc2 cells were purchased from PerkinElmer (Boston, MA, USA). Cell culture media was obtained from Sigma-Aldrich (St. Louis, MO, USA) or Gibco Life Technologies (Waltham, MA, USA). ESI-MS analysis was performed on a TSQ Quantum Ultra system equipped with a Surveyor Autosampler Plus and MS Pump Plus from Thermo Fisher Scientific (San Jose, CA, USA). HPLC analysis was performed on a system from Waters (Milford, MA, USA) with a quaternary pump, a 2998 PDA detector, a radio detector consisting of a NaI crystal detector with a Canberra Osprey-DTB, and dedicated software. Solvent A was water + 0.1 % v/v TFA, and solvent B was ACN + 0.1 % v/v TFA. HPLC method A is Gemini C₁₈ semi-preparative column (10 × 250 mm, 5 μm, Waters) and a gradient profile of 0–1 min 20 % B in A, 1–25 min to 30 % B at a flow rate of 3.0 ml/min. HPLC method B is Symmetry C₁₈ analytical column (4.6 × 250 mm, 5 μm, Waters) and a gradient profile of 0–1 min 10 % B, 1–20 min towards 60 % B at a flow rate of 1.0 ml/min. HPLC method C is Symmetry C₁₈ analytical column (4.6 × 250 mm, 5 μm, Waters) and a gradient profile of 0–1 min 22 % B, 1–20 min towards 51 % B (solvent A was 0.2 M Tris-HCl buffer pH 8.5 + 10 % MeOH and solvent B was MeOH) at a flow rate of 1.0 ml/min. Autoradiography was performed using super resolution phosphor screens and a Cyclone® Plus system (Perkin Elmer, Waltham, MA, USA). NIRF imaging was performed on an Odyssey flatbed scanner system (800 nm channel, laser intensity 5.0; Li-Cor, Lincoln, NE, USA). Bioluminescence imaging (BLI) was performed in an IVIS-Spectrum imager (Perkin Elmer). SPECT/CT images were obtained with a VECTor-5, equipped with a high-sensitivity 3.0-mm pinhole collimator (MILabs, Utrecht, The Netherlands). The mages were analyzed with Pi-Mod (version 3.901) and visualized in VivoQuant (version 2.50, patch 3). Activity was accurately quantified on a Wizard 3″, 1480 γ-counter (Perkin Elmer). Histochemical dead cell staining was performed using the DeadEnd™ Colorimetric TUNEL System (Promega, Madison, WI, USA) and subsequent imaging on a NanoZoomer 2.0HT digital slide scanner (Hamamatsu, Hamamatsu City, Japan).

Compound 2 (5.0 mg, 4.4 μmol) was dissolved in DMSO (50 μl) and added to a mixture of DOTA-PEG₄-NH₂ (3, 3.1 mg, 3.8 μmol) and DiPEA (1.78 μl, 10.1 μmol) in DMSO (200 μl). The resulting solution was stirred overnight at room temperature. Subsequently, the mixture was directly purified over semi-preparative HPLC, method A (t_R 14 min), yielding a green solid (4.1 mg, 55 %). The purity was > 99 % as determined by HPLC, method B (t_R 10.1 min, Supplementary data 1). ESI-MS: m/z calculated for C₇₂H₁₀₃N₈O₂₅S₄⁺ was 1607.59. m/z found was 802.67 [M-3H]²⁻; 534.75 [M-4H]³⁻.

Compound 1 (430 μg, 250 nmol) was added into a conical vial, and quenchers were added to mimic reaction conditions of radiolabeling (gentisic acid and sodium ascorbate with a final concentration of 3.5 mM). The reaction was buffered to pH 4–4.5 with sodium acetate (53.6 μl, 2.5 M) and diluted with Milli-Q water to a final reaction volume of 135 μl. Then, InCl₃ (375 nmol, native indium ICP standard in 0.05 M HCl) was added, and the sealed vial was heated to 70 °C for 30 min. The vial was allowed to cool to room temperature and DTPA (4 mM, 5 μl) was added to complex remaining indium. The conversion yield was > 99 % as determined by HPLC, method B (t_R 9.5 min, Supplementary data 2). ESI-MS: m/z calculated for C₇₂H₁₀₀N₈O₂₅S₄In⁺ was 1719.47. m/z found was 860.19 [M+H]²⁺.

Into a conical vial, gentisic acid and sodium ascorbate (both 50 mM in 10 μl) were added to 1 (1.4/14/70 μg; 0.8/8.0/40 nmol; for low, medium, or high dose, respectively) to prevent radiolysis [20]. The reaction was buffered to pH 4–4.5 with sodium acetate (1 μl, 2.5 M) and diluted with Milli-Q water for a final reaction volume of 135 μl. Then, [¹¹¹In]InCl₃ (80 MBq) was added and the sealed vial was heated to 70 °C for 30 min. The vial was cooled to room temperature for 5 min, and DTPA (4 mM, 5 μl) was added. Radiochemical purity (RCP) of [¹¹¹In]1 was > 95 % as determined by HPLC, method C (t_R 14.1 min). Radiochemical yield was determined by an instant thin-layer chromatography (iTLC-SG paper, Aligant, eluent: 10 % w/v NH₄Ac to MeOH 9:1, [¹¹¹In]In-DTPA eluted to top of paper and [¹¹¹In]1 remained at the baseline).

4T1-Luc2 cells were cultured in RPMI-1640 medium containing 10 % fetal bovine serum and 1 % penicillin at 37 °C under a humidified atmosphere with 5 % CO₂. The cells were seeded in a black 96-well plate with transparent bottom (typically 1.0 × 10⁴ cells per well) and grown until 70–80 % confluency. Then, from half of the wells, the medium was removed, the cells were killed with EtOH (4 μl, 70 %) and washed once with PBS. The live cells in the other half of the wells were washed with PBS, and subsequently, all cells were incubated for 15 min with or without 800CW or it conjugates (100 nM in culture medium) as indicated. The cells were washed thrice with PBS, fixated with 4 % formalin, and the whole plate was imaged on the Odyssey. This experiment was performed in triplicate.

4T1-Luc2 cells were seeded in two 12-well plates (1.0 × 10⁵ cells per well) and grown until 70–80 % confluency. After incubation, the medium was removed from half of the wells, after which the cells were killed with EtOH (50 μl, 70 %) and washed once with PBS. The live cells were washed with PBS, and subsequently, all cells were incubated for 1 h at 37 °C with or without the radiotracers (3.3 MBq/nmol, 100 nM in 400 μl culture medium). Then, the cells of one plate were washed twice with PBS, detached with 1.0 M NaOH, and collected in tubes for γ-counting. The cells of the other plate were washed twice with PBS; the whole plate was imaged on a phosphor screen (2 h) and the Odyssey. This experiment was performed in triplicate.

All animal experiments were approved by the Animal Welfare Committee and conducted in accordance with accepted guidelines. Female BALB/cAnNRj-nude mice (6 to 8 weeks old) were housed in ventilated cages in groups of four to six mice and were provided standard laboratory animal food pellets and water ad libitum. A week after arrival, 1.0 × 10⁴ 4T1-Luc2 cells suspended in 15 μl matrigel: PBS (1:1) were injected bilaterally on the shoulders. Tumor growth was monitored every 3 days with a caliper. After 2 weeks, mice were injected with d-luciferin (150 mg/kg, i.p. injection). Ten minutes later, BLI was performed for 30 s with an open filter under isoflurane anesthesia (4 % induction, 1.5 to 2 % maintenance in 100 % O₂₎. All groups consisted of three mice with bilateral tumors.

Tumor-bearing mice received an intravenous injection of [¹¹¹In]1 (20 MBq, labeled to 0.2, 2, or 10 nmol, 200 μl PBS) in the tail vein. At 6 and 24 h post injection (h.p.i.), SPECT/CT imaging was performed under isoflurane anesthesia, while the body temperature was maintained constant. Static images of the 5-cm axial field of view were obtained over a total scan time of 28 min and 20 s, followed by a 2-min full body CT. Dynamic scans were obtained over a total duration of 1 h with 30 timeframes, directly after injection of [¹¹¹In]1.

Acquired images were reconstructed using SR-OSEM with 9 iterations and 128 subsets on a 36 × 36 × 35 mm matrix with 0.80 × 0.80 mm isotropic voxels. The images were further analyzed in Pi-Mod and VivoQuant. Regions of interest were manually drawn around the tumors, heart, and muscle. Subsequently, the percentage of injected dose (%ID) and the tumor to background ratio (TBR) were determined.

After imaging, the blood, tumors, skin, pancreas, liver, spleen, small intestines, colon, ovaries, kidneys, lungs, heart, muscle, bone, lymph nodes, and brain were collected and weighted, and activity was accurately quantified in a γ-counter. After counting, the tumors were frozen in liquid nitrogen and adjacent 10-μm cryosections of tumor centers were prepared for autoradiography, NIRF imaging, and TUNEL staining.

Precursor 1 was successfully synthesized via an amide coupling reaction between 2 and 3 (Fig. 1). After purification, 4.1 mg (55 % yield) of precursor was isolated as a green solid, with a purity of > 99 % (HPLC), and was stored in titrated aliquots of 15.6 nmol at − 20 °C.

The RCP and RCY of [¹¹¹In]1 were > 95 % with molar activities achieved up to 200 MBq/nmol. The antioxidant gentisic acid and ascorbic acid in the reaction mixture stabilized radiochemical purity. After 2 days at room temperature (10 MBq/nmol/200 μl), RCP was > 95 % (Supplementary data 3). All in vitro and in vivo experiments were executed within 3 h after radiolabeling.

We further examined the effect of the metal chelation and indium labeling on the necrosis avidity of 800CW in dead or alive 4T1-Luc2 cells. Based on fluorescent signal from the cells after washing, the uptake of 800CW and its conjugates was determined (Fig. 2a).

The following conditions were tested: In-DOTA-PEG₄-800CW (^Nat.In-1), labeled with non-radioactive indium; DOTA-PEG₄-800CW which was heated at 70 °C for 30 min (1^a) [20], untreated 1, 800CW-carboxylate (4), DOTA-PEG₄-NH₂ (3) as a negative control, and medium for background subtraction. The experiment was performed in triplicate (Fig. 2a). Based on the fluorescent signal, we found that neither the procedures required for radiolabeling nor the indium complexation of 800CW had a significant effect on uptake by dead cells (Fig. 2b).

Later, we determined the [¹¹¹In]1 uptake by dead cells and used [¹¹¹In]In-DOTA-PEG₄-NH₂ ([¹¹¹In]3) as a negative control. The wells were washed, and the whole plate was imaged on a phosphor screen for visualization of the radioactivity and on the Odyssey (Fig. 2c) for quantification of the fluorescence. The fluorescent signal from the dead cells was a 67-fold stronger than that from the live cells (Fig. 2d). As expected, no fluorescence signal was observed from the [¹¹¹In]3-treated wells. Simultaneously, a second plate was prepared, from which the cells were collected from the wells after incubation and washing. The radioactivity was measured in a γ-counter for accurate quantification of the radioactivity uptake (Fig. 2e). There was no uptake of [¹¹¹In]3 by dead or alive cells, however a 23-fold higher uptake of [¹¹¹In]1 by dead cells as compared to live cells.

[¹¹¹In]1 was administrated in female nude mice inoculated with 4T1-Luc2 breast cancer cells, which develop rapidly growing tumors with spontaneous necrotic cores. The necrotic core was first confirmed by BLI since 4T1-Luc2 cells express luciferase. As a result, necrotic tumors showed a ring-shaped BLI signal from their viable outer rim and a much lower signal from the center (Supplementary data 4) [9].

In the first in vivo experiment, we evaluated the dose range in which the tumors could be imaged. Therefore, three doses of 1 (i.e., 0.2 nmol, 2.0 nmol, and 10 nmol; from here referred to as low, medium, and high dose) with a constant amount of radioactivity (i.e., 20 MBq) per injection were tested. At 6 and 24 h.p.i., SPECT/CT imaging was performed. Then, the animals were euthanized and the organs and tumors were collected for biodistribution and histological analysis.

The uptake in the tumors and kidneys were clearly visible on SPECT images at 6 and 24 h.p.i (Fig. 3 a and b). The SPECT signal from the tumors was quantified and showed no significant difference in the total counts between high and medium dose at 6 h.p.i. On the other hand, the total counts were significantly lower in the low-dose group (44 kBq/ml) compared to the high-dose group (61 kBq/ml, P = 0.03, Fig. 3c). However, no major differences were observed in %ID between the three doses at 6 h.p.i. (Fig. 3d). At 24 h.p.i, no significant differences in either total counts or %ID was seen between the three doses. Similarly, the biodistribution studies also showed no significant differences in %ID/g between the three dose groups at 24 h.p.i. (Fig. 4a), confirming these results. Even though the overall uptake in the tumors was low, the tumors were clearly visible on SPECT images due to high TBRs (Fig. 3e). The TBR for the low-dose group was the highest (10:3), followed by medium-dose group (7:4), and high-dose group (4:2, P = 0.087). At 24 h.p.i., however, there were no significant differences between the three doses (Fig. 3c–e). The uptake of the NACA into the necrotic regions of the extracted tumor was confirmed by autoradiography, NIRF imaging, and TUNEL staining of the cryosections (Fig. 3f).

Next, we evaluated the optimal time-point for imaging. The NACA uptake in %ID/g was determined at 3, 6, 24, 48, and 72 h.p.i. At each time-point, mice were euthanized and the organs were collected to determine the biodistribution of the NACA (Fig. 4). The tumor uptake did not increase over time, and no significant differences were observed between time-points. However, it can be noted that the NACA is rapidly excreted from the body with 3%ID/g left in the kidneys after 3 h. As a result, the tumor to blood ratios increased over time, up to 29:7 at 72 h.p.i. (Supplementary data 5).

We performed dynamic scans of the first hour after injection of the NACA, in which we observed renal clearance of most of the NACA in the first recorded frame recorded 0–5 min post injection (Supplementary data 6). At 1 h.p.i., urine was collected and analyzed by radio-HPLC (method C, 0.2 MBq injected, recovery > 95 %). We observed that > 98 % of the activity eluted at 14 min indicating that the NACA was excreted intact (Fig. 5).