Table 1 Examples of data mining approaches for the identification of drug candidates for new diseases starting from libraries of approved pharmaceuticals

Case 1: Repurposing drugs for Fragile X Syndrome

Healx, in partnership with FRAXA Research Foundation, employs machine learning algorithms and computational biology as the basis of its in silico Disease-Gene Expression Matching (DGEM) pipeline – with subsequent pharmacological expert reviews - to identify drug-disease connections. This approach results in several candidates, of which three were tested in in vivo mice studies, and the most promising chosen to progress through Phase IIa trials. It took only 15 months from project initiation to readiness for clinical trial [50, 51].

Case 2: Tolcapone repurposed to treat transthyretin amyloidosis (ATTR)

Following the use of a proprietary virtual screening platform, SOM Biotech identified tolcapone (SOM0226, or CRX-1008) as potential treatment for ATTR, a rare genetic degenerative disease where abnormal build-up of amyloid takes place and deposited in different organs and tissues, notably the nervous system and myocardium. Clinical validations with Phase II trials led to the granting of orphan drug designation by the FDA for all types of ATTR. The knowledge gained was leveraged against the execution of a licensing agreement for the clinical development and commercialization of this repurposed drug, an oral medication used as adjunct in the treatment of Parkinson’s disease [52, 53].

Case 3: TEE886 in pseudo-adrenoleukodystrophy

APTEEUS, through its ID2STOP Orphan (Individualized Drug Selection Technology for Orphan Patients) program, gathered over 1500 marketed drugs to build a pharmacopeia which can be systematically screened against a patient’s cells which bear the causative effect of the disease. Once potential drug candidates were identified, functional assays were carried out to determine efficacy and potency in rescuing the disease phenotype. In a case of a pseudo-adrenoleukodystrophy patient, incubation of his skin fibroblasts with TEE886 shows restoration of the profile of very long chain fatty acids which would otherwise accumulate in all cells in his body (Conference communication, [54]).

Case 4: Fluspirilene as candidate anti-cancer drug

The cyclin-dependent kinase 2 (CDK2) is an attractive anti-cancer drug target given its roles in controlling cell proliferation. A group of academic researchers created a free, open-source protein-ligand docking software to conduct in silico screening of FDA-approved small molecular drugs against CDK2. Nine compounds were subsequently tested in vitro of which the anti-psychotic drug fluspirilene was identified as a potential CDK2 inhibitor. Further in vivo mice studies show the potential of the repurposing of fluspirilene as anti-hepatocellular carcinoma [55].