The present study revealed the following main results, (i) Trk inhibitor [18F]TRACK is taken up into TrkA expressing human KM12 colon cancer cells and this uptake can be inhibited by pan-Trk inhibitor entrectinib; (ii) [18F]TRACK is also taken up into KM12 tumor tissue as well as brown adipose tissue (BAT) in vivo which can also be blocked by entrectinib; and (iii) amitriptyline acts as a functional Trk agonist and increases uptake of [18F]TRACK in peripheral target tissue.
Trk expression and function has been mainly studied in the neuronal system as well as for neurological disorders and neurodegenerative diseases [15–19]. In addition, Trk fusion protein expression and activation promoting cell proliferation has been found in many cancers [1]. However, based on their rarity, variety and clinical sample detection limits to date, functional effects and importance in cancers are not fully understood yet [45]. In 2018 and 2019, first generation Trk inhibitors larotrectinib and entrectinib were introduced into the clinic for targeted therapy effectively blocking the kinase domain activity [46]. Next generation Trk inhibitors selitrectinib and repotrectinib are currently undergoing clinical development to overcome therapy resistance [47]. Clinical data have shown that specific therapy with first generation Trk inhibitors achieves high response rates independent of tumor histology, age or NTRK fusion type [48]. Patient selection for Trk therapy may become a key step for personalized medicine and treatment decisions. As current clinical testing for NTRKs is still based on invasive biopsy sampling or prohibitive costs for novel bioassays such as NGS [45], access to non-invasive imaging methods such as PET would tremendously facilitate therapy decision making processes in the clinical setting.
After numerous imaging probe developments [37], [18F]TRACK has been developed as an optimized 18F-labeled Trk targeting PET radiotracer [43, 44]. So far, [18F]TRACK has been mainly evaluated for its binding to Trk in the brain. While it has shown substantial brain uptake in preclinical studies in murine, monkey and human brain [43, 44], in vivo proof of target specificity remains challenging.
In the present study, inhibition of [18F]TRACK cellular uptake into KM12 cells using pan-Trk inhibitor entrectinib was successful. KM12 represents a human colon cancer cell line expressing TrkA and more specifically tropomyosin 3-TrkA (TPM3-NTRK1) fusion protein [14, 49]. Since non-radiolabeled TRACK and entrectinib resulted in similar IC50 values in the nanomolar range when competing with [18F]TRACK binding in KM12 cells. These results serve as a strong functional proof-of-target in this colon cancer model in vitro. In enzymatic assays, entrectinib has shown IC50 values of 1–5 nM against isolated TrkA/B/C [50]. TRACK resulted in an IC50 of 0.3 to 4.2 nM IC50 for human TrkA/B/C measured in an ATP-based enzymatic assay [43]. However, uptake into KM12 tumors in vivo was similar to muscle tissue (SUV60min 0.43 versus 0.46). This could be attributed to the muscle expression of Trk specifically TrkB as known from the literature [51, 52]. Interestingly, uptake profiles over time resulted in a later slight washout from muscle tissue while uptake into tumor tissue continued to rise even after 2 h. resulting in a T/M ratio of 1.1.
PET imaging experiments also revealed uptake into brown adipose tissue (BAT) in mice reaching significantly higher uptake levels (SUV60min > 1) versus KM12 tumor tissue. In the literature, adipose tissue has been shown to express TrkB mRNA [53] and TrkB protein in bovine adipose tissue of lactating cows [54]. The neuronal BDNF/TrkB axis and its function in adipose tissue plays an important role for metabolic regulation and its influence on food intake and obesity. Therefore, it seems not surprising to detect uptake of [18F]TRACK in murine BAT.
TrkB protein expression in brown and white adipose tissue as well as KM12 tumor samples was confirmed using Western blotting and immunohistochemistry and compared to TrkB expression in the whole mouse brain. Experimental analysis revealed that TrkB expression in the brain is high and ubiquitous. Immunohistochemistry also showed somewhat higher expression of TrkA over TrkB in the mouse brain. In contrast, expression of TrkB was stronger than TrkA on the membrane surface of fat cells analyzed from murine BAT samples confirming previous observations in lactating cows [54]. However, present analysis of KM12 tissue with immunohistochemistry did not reveal substantial expression of both TrkA and TrkB. TrkA was only found in small inclusions in the tumor tissue confirming only a somewhat weak expression although KM12 cells are known to strongly express TrKA [55] although as the fusion protein TPM3-NTRK1 [56]. It is possible that the antibody used for western blot and immunohistochemistry experiments in the present study was not able to detect the fusion protein.
The present functional in vivo analysis using radiotracer [18F]TRACK supports the fact of lower protein expression and functional availability in KM12 tumors versus higher expression in BAT. To confirm specific targeting of [18F]TRACK to Trk (mainly TrkB) in BAT, in vivo blocking studies were performed with non-radioactive TRACK reference compound for self-blocking and pan-Trk inhibitor entrectinib. With both compounds similar blocking effects on the radiotracer uptake were observed at 30 min p.i. (minus 23% with TRACK and 18% with entrectinib). However, this effect was not persistent and at 60 min p.i. no blocking was detectible indicating reversible effects over time. Based on the reversibility of the observed effects it is difficult to conclude that these in vivo experiments using standard Trk inhibitors and therefore antagonists binding to the intracellular kinase domain would serve as proof-of-target in vivo although they show clear blocking effects in KM12 cells in vitro.
Consequently, an alternative approach using an agonist on Trk receptors might provide a way to prove specific targeting of TrkB via radiotracer [18F]TRACK in vivo. A basic analysis from 2009 revealed that the classic tricyclic antidepressant amitriptyline directly binds to the extracellular domain of both TrkA and TrkB triggering their dimerization and activation [57]. It was found that amitriptyline is in fact acting as an agonist on both TrkA and B in neuronal cells and that inhibition of TrkA eliminated its neuroprotective effect without impairing its antidepressant activity. Subsequent follow up studies found that amitriptyline induces neurite outgrowth in rat primary cortical neurons through Trk receptor and MAP kinase activation [58] and that similarly to neurotrophins amitriptyline also acts as a Trk receptor agonist on the regeneration of afferent cochlear synapses in the inner ear [59]. Using amitriptyline in the present study, overall uptake levels of [18F]TRACK in KM12 tumor tissue and in BAT could be significantly increased by 21 and 23% indicating that ligand activation of TrkA and TrkB in both tissues resulted in increased binding of [18F]TRACK at the intracellular kinase binding site as well. For comparison, radiotracer uptake in the whole brain was increased by only 13%, which did not reach statistical significance, though the trend confirmed the observations in the peripheral tumor and BAT tissue. Interestingly, [18F]TRACK uptake into muscle tissue was also elevated by 15%, however, only after 30min and diminished after 60min which is different from the effect in tumor and BAT tissue attributing maybe to different intracellular ATP levels to compete for binding with [18F]TRACK or different dimerization forms of the oncogene Trk versus the muscular Trk. It is known, that muscle tissue expresses Trk [51] and specifically TrkB and that BDNF mimetics such as 7,8-di-hydroxyflavone [52] or other antidepressant drugs such as fluoexitine can activate muscular TrkB by direct binding to the receptor [60]). Taken together, tricyclic antidepressant amitriptyline also acts as a Trk agonist in peripheral target tissue by activating Trk receptors resulting in an uptake increase of Trk targeting radiotracer [18F]TRACK. In addition, this in vivo experiment also serves as a pharmacodynamic proof-of-target experiment.
Finally, experiments combining agonistic activation of Trk with amitriptyline (binding at the extracellular domain) and Trk inhibition with entrectinib (binding the intracellular kinase domain) revealed that there was a small but also systematic blocking effect detectable at 60 min post injection of radiotracer [18F]TRACK in both KM12 tumor tissue (Δ8%) and BAT (Δ14%). This effect was persistent in the equilibrium phase of the radiotracer’s in vivo distribution which was in contrast to the observed reversible effects detected without amitriptyline. This combination experiment may serve as the best functional proof-of-target for selective target binding of [18F]TRACK in vivo and in combination with the in vitro results.