A method for identification of anti-hiv human mirna mimics

Medicilon’s Chemistry department has more than 100 chemists, who are experienced in the cooperation with major domestic and international pharmaceutical and biotech companies. Our services cover a variety of research interests in novel drug research, including target validation, hits evaluation, lead optimization, candidate nomination, preclinical development and IND filing. Email:marketing@medicilon.com.cn web:www.medicilon.com

The present invention relates to methods for the identification of anti-HIV miRNAs and anti-HIV pharmaceutical compounds using high-throughput screening methods, comprising: transfecting reporter cells with a panel of miRNAs, infecting the reporter cells with HIV, screening the cells to identify miRNAs that modulate HIV infection and identifying the specific pathways, nucleic acids and/or polypeptides that are targeted by the miRNAs. The invention further provides for the identification and screening of anti-HIV pharmaceutical compounds having known activity against the specific pathways, nucleic acids and/or polypeptides that are targeted by the miRNAs for efficacy in the treatment of HIV. The invention also provides for the use of miRNA mimics, miRNA inhibitors and pharmaceutical compounds (including oncology drugs and kinase inhibitors) in the treatment and/or prevention of HIV infection.

Four fields of view (100 um apart) were acquired from the central region of each well imaged. Images were acquired in the 405 and 488 fluorescent channels for each field of view. Wells were imaged using the 20x ewld objective on the Image Xpress Ultra automated confocal microscope system (Molecular Devices). The imaging parameters were kept constant for all replicates of both miRNA and targeted drug/compound screens.

Image analysis

HCS analyses rely on the discrimination of multiple biological descriptors between images derived from experimental and control wells for HIT identification. HCS analyses thus require the extraction quantitative, subcellular measurements (features) from the acquired images in order to identify subsets of these features that can accurately and reproducibly classify specific biological phenotypes within experimental wells.

Feature extraction algorithms were utilized in order to extract 16 subcellular measurements from the acquired images. These 16 factors are derived from 15 descriptors previously described by Genovesio et al. 2011 to be relevant indicators of the infective state of host cells, with addition of 'infection efficiency' as a descriptor of the percentage of infected cells per treatment well.

A bespoke support vector machine-learning (SVML) pattern recognition algorithm was trained to identify phenotypes of interest using the relevant control wells. Transfection of GHOST(3) with an siRNA targeting the human CD4 receptor (25 nM) 48 hours prior to exposure to HIV was shown to significantly suppress HIV infection and produce a visually distinct phenotype representative of the inhibition of HIV replication. Knockdown of the CD4 receptor required for HIV entry in GHOST(3) cells 48 hours prior to exposure to HIV resulted in suppressed GFP reporter signal and when compared to natural infection controls (high GFP signal) creates contrasting visual phenotypes required for subsequent image analysis. siCD4 wells were included on every screening plate in the miRNA screens as positive controls for suppressed HIV infection. Likewise the addition of the HIV-1 integrase inhibitor, Raltegravir (10 uM) to GHOST(3) culture medium 24 hour prior to exposure to HIV was also shown to produce a distinct phenotype representative of suppressed HIV replication. The transfection of siCD4 and Raltegravir treatment were thus utilised in the miRNA and drug/compound screens respectively as positive controls for the inhibition of HIV-1 replication.

SVML algorithms were trained using these controls and were thus able to generate a decision boundary (HIT boundary) that defines a region of 16 dimensional space which is representative of the Raltegravir or siCD4 control wells and which is also distinct from the controls representative of a natural infection phenotype. Experimental wells from the miRNA and drug screens, which were classified within the HIT boundary closer to the siCD4 and Raltegravir controls, were identified as positive HITs for the inhibition of HIV replication. miRNA mimics or inhibitors which were able to enhance activation of the HIV LTR promoter were identified by evaluating the average GFP intensity in experimental wells, relative to the natural infection controls. miRNA molecules which produced a greater than 2-fold increase in average GFP intensity relative to control wells were classified as HIV enhancer miRNA molecules. Only miRNA molecules and drugs which were shown to elicit concordant phenotypes across replicate screens were included in the final list of HITs. Additionally, using the values obtained for cell counts per well (as per Hoechst staining) we were also able to evaluate the cytoxicity of specific drugs and compounds using the "drug only" screen format as a control. Drugs/compounds which were shown to be cytotoxic specifically in response to HIV exposure in either the phenocopy and clearance screen formats but not in the drug only screens were also identified as drugs/compounds of interest.