Preclinical molecular imaging in oncology – From chemical synthesis to clinical translatable applications

Background: Molecular imaging (MI) is fundamental in clinical settings for diagnosis, treatment selection, intraoperative guidance and treatment efficacy evaluation. This field has progressed rapidly in recent years, driven by the need for diagnostic tools and as a tool to monitor the efficacy of new cancer treatments, including gene-based therapies. MI in combination with reporter genes allows in vivo monitoring of enzyme expression and follow-up of the success of gene-directed enzyme prodrug therapies (GDEPT), where the suicide gene also acts as reporter gene. In this thesis, nitroreductase NfsB (NTR) from E. coli has been concurrently employed as a reporter gene (Papers I and III) and suicide gene (Paper III). The work of this thesis has also focused on the contrast agents of two main MI methodologies, namely optical imaging (mainly fluorescence imaging (FLI) for longitudinal imaging (Papers I – III) and fluorescence image-guided surgery (FIGS) (Paper II)) and nuclear-based PET/CT imaging (Paper III). Aims: To highlight and resolve the discrepancies present in the literature in regard to the near-infrared (NIR) dye CytoCy5S. For this purpose, four NIR dyes were synthesised and analysed. Based on their physicochemical and optical properties and their in vivo behaviour, we aimed to define the best candidate for preclinical optical imaging of NTR (Paper I). To compare how common NIR dyes affect the in vivo behaviour of conjugates and to identify the optimal dye that could be used to improve the further development of fluorescent conjugates for FIGS (Paper II). To repurpose the hypoxia radiopharmaceutical 18F-FMISO for PET/CT imaging of NTR expression in an in vivo GDEPT setting (Paper III). Materials and methods: We synthesised, purified and characterised four NTR substrates (1 – 4), including alternately described structures currently referred by the name CytoCy5S. A comparative NTR enzymatic assay was performed to assess the spectroscopic characteristics of the different reductively activated dyes. The NTR expressing triple-negative breast carcinoma cell line, MDA-MB-231 NTR+, was employed to compare, both in vitro and in vivo, the suitability of these fluorophores as reporters of NTR activity. Comparison of the reporting properties was achieved by flow cytometry, fluorescence microscopy and optical imaging, both in vivo and ex vivo (Paper I). The NHS esters of ZW800-1, ZW800-1 Forte, IRDye® 800CW, ICG, and Cy7 derivative were conjugated to folate through an ethylenediamine linker to give conjugates 1 – 5, respectively. The optical properties of conjugates 1 – 5 were determined by spectroscopy, their specificity assessed in vitro by flow cytometry and FLI, and their biodistribution studied in vivo and ex vivo using FLI in a subcutaneous Skov-3 ovarian cancer model (Paper II). The efficacy of 18F-FMISO to report NfsB NTR activity in vivo was investigated using the MDA-MB-231 model. For validation, subcutaneous xenografts of cells constitutively expressing NTR were imaged using 18F-FMISO PET/CT and FLI with CytoCy5S. Further, examination of the non-invasive functionality of 18F-FMISO PET/CT in reporting NfsB NTR activity in vivo was assessed in metastatic orthotopic NfsB NTR expressing xenografts and metastasis confirmed by bioluminescence imaging. 18F-FMISO biodistribution was acquired ex vivo by an automatic gamma counter measuring radiotracer retention to confirm in vivo results. To assess the functional imaging of NTR-based GDEPT with 18F-FMISO, PET/CT was performed to assess both gene transduction and cytotoxicity effects of prodrug therapy (CB1954) in subcutaneous models (Paper III). Results: The introduction of the methoxy group in 3 and the sulfonate and cyanine scaffold in 4 resulted in a modest bathochromic shift compared to 1 and 2. All compounds were confirmed as NTR substrates with delayed kinetics observed for 2 and low emission intensities for 3 and 4. In vitro, 3 was found to display the highest NTR+/NTR- ratio, followed by 1. 4 performed poorly in vitro and in vivo. 1 and 2 provided similar suitability for NTR interrogation, being 1 slightly more sensitive at smaller tumour volumes and presenting faster kinetics and a functional group for conjugation to biomolecules. We determined that substrate 1 is the ideal candidate for further studies (Paper I). We demonstrated time- and receptor-dependent binding of folate conjugates in vitro and in vivo. Healthy tissue clearance characteristics and tumour-specific signal varied between conjugates 1 - 5. 2 revealed the highest contrast in folate receptor alpha (FRα) positive xenografts and was the only conjugate that reported significant differences between Skov-3 and A549 (p < 0.05). Whilst conjugates 1, 2 and 3 are renally cleared, hepatobiliary excretion and no or very low accumulation in tumours was observed for 4 and 5 (Paper II). 18F-FMISO retention was detected in NTR+ subcutaneous xenografts, displaying significantly higher PET contrast than NTR- xenografts (p < 0.0001). Substantial 18F-FMISO retention was evident in metastases of orthotopic xenografts (p < 0.05). Accordingly, higher 18F-FMISO biodistribution was prevalent ex vivo in NTR+ xenografts. 18F-FMISO NfsB NTR PET/CT imaging proved useful for monitoring in vivo NTR transduction and the cytotoxic effect of prodrug therapy (Paper III). Conclusions: We have fully characterised the chemical and biological properties of four NIR dyes and clearly defined which is the best for imaging of NTR (Paper I). Our work highlights the high impact the physicochemical properties of the dye have on the tumour accumulation and biodistribution of the conjugate and demonstrates the potential of alternative fluorophores to IRDye® 800CW for targeted approaches in FIGS (Paper II). We have demonstrated that 18F-FMISO is a readily implementable PET probe to be employed as companion diagnostic test for NTR-based GDEPT systems (Paper III).Doktorgradsavhandlin