Synthesis and preliminary PET imaging of 11C and 18F isotopologues of the ROS1/ALK inhibitor lorlatinib

Thomas Collier, Marc D. Normandin, Nickeisha A. Stephenson, Eli Livni, Steven H. Liang, Dustin Wooten, Shadi A. Esfahani, Michael G. Stabin, Umar Mahmood, Jianqing Chen, Wei Wang, Kevin Maresca, Rikki N. Waterhouse, Georges El Fakhri, Paul Richardson, Neil VasdevNature Communications 2017 · DOI

Abstract

Lorlatinib (PF-06463922) is a next-generation small-molecule inhibitor of the orphan receptor tyrosine kinase c-ros oncogene 1 (ROS1) and anaplastic lymphoma kinase (ALK). It is currently under investigation for treating non-small cell lung cancers (NSCLC), particularly those with central nervous system (CNS) metastases. A critical requirement for therapeutic success is the drug's ability to penetrate the blood-brain barrier (BBB). In this study, we developed $^{11}$C- and $^{18}$F-labeled isotopologues of lorlatinib to evaluate its biodistribution and brain permeability using positron emission tomography (PET). By employing non-traditional radiolabeling strategies, including an automated "Loop Method" for $^{11}$C and iodonium ylide-based radiofluorination for $^{18}$F, we achieved reliable tracer production. PET imaging in non-human primates confirmed high brain permeability, validating the molecular design of lorlatinib for CNS-targeted therapy.

Introduction: The Challenge of Brain Metastases

Many patients with NSCLC driven by ALK or ROS1 mutations eventually develop brain metastases. While targeted therapies exist, the blood-brain barrier often acts as a physical barrier, preventing drugs from reaching effective concentrations in the brain. Lorlatinib was specifically engineered to be highly brain-permeable to address this "sanctuary site" for tumor growth. However, non-invasively quantifying drug delivery to brain lesions in vivo remains a challenge. This research focuses on creating radioactive versions (isotopologues) of lorlatinib that are chemically identical to the parent drug, enabling real-time visualization of its pharmacokinetics via PET imaging.

Radiolabeling Methodology

The structural complexity of lorlatinib presented significant challenges for traditional radiochemistry. The team developed two primary pathways:

Carbon-11 ($^{11}$C) Labeling

Initial attempts at direct one-step methylation of the precursor (R)-2 were unsuccessful due to the high nucleophilicity of the pyridine nitrogen, which competed with the target amide nitrogen. To overcome this, the researchers utilized a two-step "Loop Method." This approach employed a diBoc-protected precursor, (R)-3, which effectively shielded the reactive pyridine site, allowing for selective methylation and subsequent deprotection to yield the desired tracer, $^{11}$C-1.

Figure 6: The successful two-step "Loop Method" for Carbon-11 labeling using the diBoc-protected precursor (R)-3.

Fluorine-18 ($^{18}$F) Labeling

Traditional nucleophilic fluorodenitration led to molecular decomposition. Instead, the team applied a spirocyclic iodonium ylide-based radiofluorination strategy. This milder technique is better suited for the "mildly activated" aromatic rings found in lorlatinib and provided significantly higher radiochemical yields (14% uncorrected) compared to traditional methods (<1%).

Key Insights: Overcoming Chemical Reactivity

A major hurdle in the synthesis was the unintended reactivity of the pyridine ring. In direct methylation attempts, the radioactive methyl group preferred the pyridine nitrogen, forming the quaternary ammonium species $^{11}$C-7 instead of the desired amide $^{11}$C-1.

Figure 5: Reaction scheme illustrating the competing methylation pathways and the formation of the undesired pyridine-labeled byproduct.

The success of the diBoc-protection strategy highlights the importance of tailored protecting group chemistry in the development of clinical PET tracers. Furthermore, the use of iodonium ylides proves to be a robust alternative for labeling complex molecules that cannot withstand harsh fluorination conditions.

In Vivo Validation and Imaging Results

The tracers were evaluated in both rodent and non-human primate (NHP) models to assess tumor targeting and biodistribution:

  • Tumor Targeting: In mice bearing human H3122 (EML4-ALK-positive) NSCLC xenografts, the tracer showed specific uptake (2.2–2.37% ID/g). When co-injected with non-radioactive lorlatinib, tumor uptake dropped to <0.4% ID/g, confirming the tracer's specificity for the ALK/ROS1 targets.
  • Brain Permeability: PET imaging in rhesus macaques confirmed rapid and high brain uptake, with a Standardized Uptake Value (SUV) of approximately 2 in the cerebellum within 10 minutes. Approximately 1.4% of the total injected dose reached the brain, validating the molecular engineering goals of the drug's original developers.
  • Dosimetry: The effective dose was calculated at 0.0035 mSv/MBq, suggesting the tracer is safe for future human clinical trials.

Conclusion and Future Directions

This study successfully demonstrates the production of $^{11}$C and $^{18}$F isotopologues of lorlatinib using innovative radiochemical techniques. The confirmation of high brain permeability in primates is a significant milestone, supporting the drug's clinical use for CNS metastases. While the $^{11}$C tracer is limited by its 20-minute half-life, the $^{18}$F version (110-minute half-life) offers greater potential for widespread clinical distribution. Future work will focus on using these tracers in human patients to measure drug concentrations directly within brain tumor lesions, optimizing dosing strategies for improved therapeutic outcomes.

Summary updated

Abstract

Lorlatinib (PF-06463922) is a next-generation small-molecule inhibitor of the orphan receptor tyrosine kinase c-ros oncogene 1 (ROS1), which has a kinase domain that is physiologically related to anaplastic lymphoma kinase (ALK), and is undergoing Phase I/II clinical trial investigations for non-small cell lung cancers. An early goal is to measure the concentrations of this drug in brain tumour lesions of lung cancer patients, as penetration of the blood-brain barrier is important for optimal therapeutic outcomes. Here we prepare both <sup>11</sup>C- and <sup>18</sup>F-isotopologues of lorlatinib to determine the biodistribution and whole-body dosimetry assessments by positron emission tomography (PET). Non-traditional radiolabelling strategies are employed to enable an automated multistep <sup>11</sup>C-labelling process and an iodonium ylide-based radiofluorination. Carbon-11-labelled lorlatinib is routinely prepared with good radiochemical yields and shows reasonable tumour uptake in rodents. PET imaging in non-human primates confirms that this radiotracer has high brain permeability.
BibTeX
@inproceedings{collier2017synthesis,
  author = {Thomas Collier and Marc D. Normandin and Nickeisha A. Stephenson and Eli Livni and Steven H. Liang and Dustin Wooten and Shadi A. Esfahani and Michael G. Stabin and Umar Mahmood and Jianqing Chen and Wei Wang and Kevin Maresca and Rikki N. Waterhouse and Georges El Fakhri and Paul Richardson and Neil Vasdev},
  title = {{Synthesis and preliminary PET imaging of 11C and 18F isotopologues of the ROS1/ALK inhibitor lorlatinib}},
  booktitle = {Nature Communications},
  year = {2017},
  doi = {10.1038/ncomms15761},
}
Figure