Tinnitus as Side Effect: a Drug-Target Network Analysis

ELGOYHEN AB; LANGGUTH B; NOVAK W; SCHECKLMANN M; DE RIDDER D; VANNESTE S

Congreso; 7th INTERNATIONAL TRI CONFERENCE ON TINNITUS; 2013

Background A wide variety of approaches have been used in an attempt to delineate the neural correlates of tinnitus. These range from experiments in animal models to brain imaging studies in humans. However, the exact mechanisms underlying this phantom perception are still unknown. A new approach can be based on the availability of large drug databases which include tinnitus as a side effect. Methods In order to explore the tinnitus target space, we built and analyzed a drug-target network based on compounds that elicit tinnitus as side effect. A total of 300 drugs with side effect tinnitus were extracted from the  SIDER 2 database. We subsequently looked for all reported molecular targets of these drugs in a manually curated database that was generated with information downloaded from the DrugBank and PDSP Ki databases. Using Cytoscape 3 for network data integration and analysis, we generated a bipartite ‘drug–target network’. Results Of the 300 total drugs, 181 formed a giant component, the largest connected component of the network. Within it, two distinct modules were relevant: one that included central nervous system acting drugs and the other cyclooxygenase inhibitors. A third isolated component included cardiovascular acting drugs, mainly angiotensin converting enzyme inhibitors. The target network also displayed a main giant interconnected component which included molecules from different gene families. Only cyclooxygenase 2, cyclooxygenase 1, ACE (angiotensine converting enzyme), 5HT1A (serotonine type 1A receptor), SLC6A4 (serotonin transporter), SCN5A (sodium channel protein type 5 subunit alpha) were significant above random. These findings recapitulate the fact that tinnitus might be both of peripheral and central origin, being the cyclooxygenases most likely linked to a cochlear site of genesis and 5HT1A, SLC6A4 and SCN5A to a central mechanism. In order to further validate these results we generated drug-target networks with drugs that produce either hearing loss or hyperacousis as side effects, two symptoms that sometimes accompany tinnitus. Similar to that observed in the tinnitus network, cyclooxygenases were enhanced above random in the hearing loss network, suggesting a peripheral site of action of cyclooxygenase inhibitors. On the contrary, whereas the cyclooxygenases were not present in the hyperacousis network they shared with tinnitus 5HTA1, SLC6A4, SCN5, recapitulating the proposed central origin of hyperacousis. Conclusions These results indicate that high-throughput electronic-biology approaches based on in silico data mining of existing databases and integration of this information through network analysis can further help understand tinnitus and related disorders. In addition, it might aid towards the design of better pharmacotherapies.