Computational analysis of the structural features of the family of FABP proteins and their role in Prostate Cancer: a comparative study between Homology Modeling and AI powered predictions aimed to establish the FABP12 protein structure

Prostate cancer (PCa) is the second most common cancer in men after lung cancer and is one of the leading causes of cancer-associated death in men. With cancer risk strongly increasing with age, PCa incidence exhibits a direct correlation with the development index of any given region. Highly developed regions, with higher life expectancy, such as USA, UK, EU(Bray et al., 2018), show higher incidence in comparison to less developed countries. These regions, however, exhibit a higher rate of annual increase of incidence and higher PCa mortality rates(Rebello et al., 2021). Localized PCa has excellent prognosis, with 5-year survivability rates of 60-99%, with highly effective options of treatment, such as Androgen Deprivation Therapy (ADT), which inhibits one of the major drivers in the disease advancement, the androgen receptor. These values inevitably decrease drastically as the cancer metastasize, degenerating into metastatic castration-sensitive prostate cancer (mCSPC), and eventually develops resistance to treatment (metastatic castration-resistant prostate cancer (mCRPC)). Fatty acids (FAs) and their metabolism play a crucial role in tissues with high rate of growth as an energy source and as metabolic intermediates for membrane biosynthesis, energy storage and the generation of signalling molecules. Due to their hydrophobic nature, FAs transportation has to be achieved either through membrane diffusion or via transportation by specific proteins. Fatty acid binding proteins (FABPs) are cytosolic proteins regulating all functions of cell lipid transportation and storage, and their overexpression has been found to be a marker for tumor advancement and worsening of prognosis. In particular, the most recently discovered member of the FABP family, FABP12, plays an important role in PCa degeneration from local to metastatic. FABP12 inhibition with general FABP inhibitors has however been proven effective in slowing down the process. Despite having identified the primary structure, FABP12 tertiary structure has yet to be elucidated trough experimental methodologies, such as X-ray diffraction or Nuclear Magnetic Resonance (NMR) analysis. To bridge this gap, in order to further the research for more specific inhibitors for FABP12 that might prove effective in arresting the development of metastasis in PCa, this study proposes the use of in silico predictions with different softwares, Molecular Operating Environment (MOE) being the first and employed to perform Homology modelling, the gold standard for in silico protein structure prediction, and the more recent, AI powered software, AlphaFold. First, the sequences of the other FABPs have been compared among themselves to establish reciprocal similarity degree and which structure to use as template for each FABP in the first testing phase, in order to adjust the software’s prediction parameters. Then all sequences are compared with FABP12 sequence to establish the best template, that has also been used as control. Structures obtained through Homology modelling with MOE and through Alphafold predictions have been compared, using different metrics, with experimentally known structure of other FABPs in order to assess the better methodology. Then the possible FABP12 structures obtained with these two softwares have been once again compared and ranked in order to find the most suitable candidate for ligand interaction simulations, which represents the next step in research for more specific FABP12 inhibitors.