Synthesis and in vitro evaluation of [¹⁸F]BMS-754807: A potential PET ligand for IGF-1R

Abstract

Radiosynthesis and in vitro evaluation of [¹⁸F](S)-1-(4-((5-cyclopropyl-1H-pyrazol-3-yl)amino)pyrrolo[ 2,1-f][1,2,4]triazin-2-yl)-N-(6-fluoropyridin-3-yl)-2-methylpyrrolidine-2-carboxamide ([¹⁸F]BMS-754807 or [¹⁸F]1) a specific IGF-1R inhibitor was performed. [¹⁸F]1 demonstrated specific binding in vitro to human cancer tissues. Synthesis of reference standard 1 and corresponding bromo derivative (1a), the precursor for radiolabeling were achieved from 2,4-dichloropyrrolo[2,1-f][1,2,4]triazine (4) in three steps with 50% overall yield. The radioproduct was obtained in 8% yield by reacting 1a with [¹⁸F]TBAF in DMSO at 170 °C at high radiochemical purity and specific activity (1–2 Ci/μmol, N = 10). The proof of concept of IGF-IR imaging with [¹⁸F]1 was demonstrated by in vitro autoradiography studies using pathologically identified surgically removed grade IV glioblastoma, breast cancer and pancreatic tumor tissues. These studies indicate that [¹⁸F]1 can be a potential PET tracer for monitoring IGF-1R.

Insulin-like growth factors (IGFs or IGF-I and IGF-II) are growth hormones that have high sequence homology to insulin and function as regulators of cellular proliferation, apoptosis, energy metabolism and various organ-specific functions.^1,2 The functions of IGFs are mediated through two tyrosine kinase receptors IGF-1R, IGF-2R and IGF binding proteins (IGF-BP).^3,4 Overexpression of IGF-1R has been found in many cancers affecting multiple aspects of malignancy and metastases. These include mesenchymal, epithelial and hematopoietic classes of a wide variety of tumors/ cancers.^5–12 In addition, a role in brain development and normal brain functions, alterations of IGF-1R are also found in Alzheimer’s Disease, traumatic brain injury, Amyotrophic Lateral Sclerosis, Fredreich Ataxia and aging.^13–18 IGF-1R overexpression has also been reported in brain tumors such as glioblastoma, neuroblastoma, astrocytomas, meningiomas and medulloblastoma.^19–22 The biological roles of IGF-1R in malignancy have been demonstrated by preclinical, epidermological and clinical studies.^6–8,23,24 IGF-IR is a potential therapeutic target and there are several active clinical trials evaluating anti IGF-1R effects in various diseases.^6–8 These agents include antibodies for IGF-I, monoclonal antibodies for IGF-1R, small molecule targeting tyrosine kinase receptor inhibitors (TKRIs) and ligand binding antibodies and antisense oligonucleotides. Although clinical trials with at least six monoclonal antibody based IGF-1R therapy are still under preclinical and phases I–III clinical evaluation, many clinical trials based on antibody based IGF-1R therapy reported negative results and safety concerns.^25–30 There is a growing interest in the development of small molecule TKRIs inhibitors of IGF-1R.^10,24 Several small molecules based on TKRIs are in various phases of clinical investigations (phases I–III).^7–11 The identification of noninvasive biomarkers would enable detection and quantification of IGF-1R in vivo to monitor target occupancy and treatment responses and accelerate development of medications that target IGF.

To date, five major classes of ligands have been considered for IGF-1R imaging including proteins, antibodies, peptides, affibodies and small molecule TKRIs.²³ SPECT ligands based on I¹²⁵ analogues of IGF-I were not successful in in vivo imaging IGF-I due to high protein binding and in vivo deiodination.²³ In vivo imaging of IGF-1R with ¹¹¹In-IGF1(E3R) in a mouse breast cancer xenograft tumor model shows good tumor contrast, a strong linear correlation with in vitro and in vivo IGF-1R expression level and tracer uptake. ²³ R1507, a monoclonal antibody in human was radiolabeled with PET and SPECT isotopes (¹¹¹In, ⁸⁹Zr and ¹²⁵I) and tested in tumor model.³¹ Higher tumor uptake of ¹¹¹In-R1507 and ⁸⁹Zr-R1507 was observed in comparison to that of ¹²⁵I-R1507.^{32 111}In-R1507 SPECT has also been utilized to predict the response to anti-IGF-1R therapy in human bone sarcoma xenografts.³³ Although antibody based imaging is successful in mice models, its translation to human study is not practical as the imaging probe requires days for equilibration, and offer poor target/nontarget contrast due to the expression of IGF-1R in normal tissues. There have been unsuccessful attempts to image IGF-1R antibodies with quantum dots using fluorescent techniques.^23,34,35 IGF-1R based PNA has been recently radiolabeled with ⁶⁴Cu and ⁹⁹mTc and tested in IGF-1R overexpressing breast cancer xenografts in mice.^36,37 A few radioligands based on IGF-1R selected antibody have been labeled with ¹¹¹In and show modest tumor uptake and tumor to blood ratio. ³⁸ Recently, [¹⁸F]FDG and [¹⁸F]FLT have been employed to access response to IGF-1R inhibitors in preclinical and human subjects.^39–41

Some of the ligands listed above show promise for imaging IGF-1R, but do not cross the blood brain barrier (BBB) and hence are limited to use in imaging studies outside the brain. To quantify binding to IGF-1R in brain requires development of selective non-peptide PET ligands. High affinity, lipophilic TKRIs of IGF-1R are potential candidates for developing such imaging agents. We have screened a large number of TKRIs for this purpose and identified BMS-754807 as a small molecule dual inhibitor of IGF-1R/insulin receptor (IR), that has been clinical evaluation, as a lead ligand for imaging using PET.^42–44 BMS-754807 (1) is an orally bioavailable, potent and reversible small molecule inhibitor of the IGF-1R/IR family kinases (K_i <2 nmol/L), which is currently in phase II clinical trial for the treatment of a variety of human cancers.^42–46 The selectivity of compound 1 over other kinases, presence of metabolically stable fluorine in the 2-substituted pyridine ring, which is amenable for radiolabeling using nucleophilic displacement with [¹⁸F]fluoride⁴⁷ and a calculated lipophilicity (clogP) 3.5, prompted us to choose it as a candidate radiotracer for imaging IGF-1R with PET.

BMS-754807 (1) and the corresponding bromo precursor (1a) for radiosynthesis, were prepared based on modification of a published procedure.⁴⁵ Preferential displacement of the C-4 chloride in compound 4 with 5-cyclopropyl-1H-pyrazol-3-amine (5) followed by displacement at C-2 position with (S)-2-methylpyrrolidine-2-carboxylic acid gave the corresponding carboxylic acid 7 in two steps with 65% yield. The carboxylic acid was then condensed with 6-bromopyridin-3-amine or 6-fluoropyridin-3-amine to yield the precursor 1a or the standard 1 in 78–70% yield (Scheme 1).⁴⁸ Radiosynthesis of [¹⁸F]BMS-754807 was initially optimized by exchange reaction with cold BMS-754807 under microwave conditions using [¹⁸F]KF/K222. However, the bromo precursor 1a did not undergo radiofluorination using [¹⁸F]KF/K222 under thermal or microwave conditions (Scheme 1). However, the radioproduct was obtained in 8% yield by treating bromo precursor 1a with [¹⁸F]TBAF in DMSO at 170 °C in high radiochemical purity (>95%) and specific activity (1–2 Ci/μmol).⁴⁹ Nucleophilic aromatic radiofluorination via the displacement of 2-bromo pyridyl system without using catalyst is widely reported in literature.⁴⁷ [¹⁸F]TBAF was in situ generated from freshly prepared tetrabutylammonium bicarbonate by azeotropic distillation with [¹⁸F]fluoride. Tetrabutylammonium bicarbonate in turn was freshly prepared by bubbling carbon dioxide from dry ice into tetrabutylammonium hydroxide (4 mL, 13% by weight in water, 0.5 M) for 4 h at room temperature. The radioproduct was stable in 10% ethanol–saline solution for over 6 h. The experimental partition coefficient (logP) of [¹⁸F]1 was measured as 2.8, indicating that the radioligand has adequate lipophilicity for brain imaging.⁵⁰

Scheme 1.

Synthesis and radiosynthesis of BMS-754807 and [¹⁸F]BMS-754807.

After optimizing the synthesize of [¹⁸F]1, the proof of concept of IGF-1R imaging with the radiotracer was performed in surgically removed and pathologically identified grade IV-glioblastoma, breast cancer and pancreatic tumor using phosphor imager autoradiography (Fig. 1).^51–53 Pathologically identified frozen sections of glioblastoma IV, pancreatic tumor and breast cancer were used for phosphor image studies. The slide mounted tumor sections were brought to room temperature and incubated with [¹⁸F]1 (0.1 nM) for 1 h.⁵⁴ The adjacent sections were incubated with 1 μM of GSK1838705A, a specific IGF-1R ligand to determine nonspecific binding. After incubation, sections were washed, dried and exposed to phosphor imaging screen. As evident from Figure 1, there is specific binding of [¹⁸F]1 in the three tumor/cancer tissues tested. Grade IV gliobalstoma shows higher tracer uptake and specific binding in comparison to pancreatic tumor and breast cancer tissues. The binding rations of total verses non-specific binding obtained are 5.25 (N = 9), 1.92 (N = 4) and 1.7 (N = 4) for grade IV glioblastoma, pancreatic tumor and breast cancer, respectively.

Figure 1.

Phosphor image of [¹⁸F]1 in surgically removed cancer tissues. Column 1: total binding of [¹⁸F]1 in glioblastoma IV sections; column 2: nonspecific binding of [¹⁸F]1 in glioblastoma IV sections; column 3: total binding of [¹⁸F]1 in pancreatic tumor sections; column 4: nonspecific binding of [¹⁸F]1 in pancreatic tumor sections; column 5: total binding of [¹⁸F]1 in breast cancer tissues; column 6: nonspecific binding of [¹⁸F]1 in breast cancer tissues.

In summary, we successfully synthesized [¹⁸F]1, a potential imaging agent for IFG-1R. The total time required for radiosynthesis was 60 min from EOS in 8% yield with excellent chemical and radiochemical purities and specific activity. Autoradiography studies by phosphor imaging indicate that [¹⁸F]1 binds to IGF-1R of surgically removed postmortem human glioblastoma grade IV, breast cancer and pancreatic tumor tissues. [¹⁸F]BMS-754807 is the first small molecule radioligand that showed promise for imaging IGF-1R in tumor/cancer. The results indicate that [¹⁸F]BMS-754807 can be a potential PET imaging agent for in vivo monitoring IGF-1R.