High-content image-based drug screen identifies a clinical compound against cell transmission of adenovirus

Human adenoviruses (HAdVs) are fatal to immuno-suppressed people, but no effective anti-HAdV therapy is available. Here, we present a novel image-based high-throughput screening (HTS) platform, which scores the full viral replication cycle from virus entry to dissemination of progeny. We analysed 1,280 small molecular weight compounds of the Prestwick Chemical Library (PCL) for interference with HAdV-C2 infection in a quadruplicate, blinded format, and included robust image analyses, and hit filtering. We present the entire set of the screening data including all the images, image analyses and data processing pipelines. The data are made available at the Image Data Repository (IDR) 1, accession number idr0081. Our screen identified Nelfinavir mesylate as an inhibitor of HAdV-C2 multi-round plaque formation, but not single round infection. Nelfinavir has been FDA-approved for anti-retroviral therapy in humans. Our results underscore the power of image-based full cycle infection assays in identifying viral inhibitors with clinical potential.

. Proteolytically matured HAdV progeny is released upon rupture of the nuclear envelope and 52 the plasma membrane [25][26][27] . 53 54 Currently, there is no effective therapy available against HAdV disease. The standard of care is 55 the nucleoside analogue Cidofovir, with poor clinical efficacy 7,28 . The problem is exacerbated by 56 the shortage of a suitable small animal model for HAdV disease, although Syrian Hamsters are 57 susceptible to HAdV-C infection and give rise to viral progeny 29 . Here, we developed an image-58 based procedure to identify novel inhibitors of HAdV infection in cell culture. We used the 59 commercially available Prestwick Chemical Library (PCL) comprising 1,280 off-patent mostly 60 FDA-approved small molecules (listed in Supplementary Table 1). The PCL comprises com-61 pounds against diseases including infection and cancer [30][31][32] . 62 63 Here, we performed a phenotypic screen against HAdV-C2 infection employing automated 64 fluorescence microscopy and image-based scoring of the progression of multi-round infections 65 using the Plaque2.0 software 33 (Figure 1a and b). The screen was performed in 384-well plates 66 (for representative images, see Figure 1c). It features robust imaging, image analysis and data 67 processing, as concluded from two parallel procedures carried out at independent institutions, the 68 Department of Molecular Life Sciences at University of Zurich (UZH), and the Biomolecular 69 Screening Facility at Ecole Polytechnique Fédérale de Lausanne (EPFL).

71
Five phenotypic features were used to score the effects of the compounds on HAdV-C2-dE3B-72 GFP infected human lung cancer epithelial A549 cellsthe number of infected and uninfected 73 cell nuclei, the infection index (infected nuclei per total nuclei), the number of plaques (areas of 74 infection foci originating from a single infected cell) and the integrated signal of the infection 75 marker green fluorescence protein (GFP) encoded in the reporter virus genome. All data are 76 available at the Image Data repository (IDR) 1 , IDR accession number idr0081, and can be 77 accessed via the IDR web client. Raw and scored infection phenotypes are provided for UZH and 78 EPFL analyses. Rigorous assay development ensured a high assay quality, as indicated by mean 79 Z'-factors of 0.52 for the plaque numbers. The screening was performed in four biological 80 replicates at high reproducibility, and compounds that gave significant toxicity in uninfected cells 81 were excluded during hit filtering. Imaging, image analysis and scoring by the two independent 82 teams yielded well correlated scores and a congruent list of top hits, provided in Table 1. We 83 confirmed the antiviral efficacy of Nelfinavir in a follow-up study (Georgi et al., in preparation). 84 85 86

88
Virus 89 HAdV-C2-dE3B-GFP was produced as described 25  The PCL compound arrangement as dispensed by EPFL in four subset plates A -D comprising 115 each screening set replicate 1 -4 was blinded and replaced by UZH with internal identifier 116 (Supplementary Tables 2 and 3, compoundIdentifier 1 to 1280). The identity of the compounds 117 was only disclosed after the screening process and hit filtering (Supplementary Tables 2 and 3  118 and Table 1, PCL_ID Prestw-1 to Prestw-1804 and compoundName). 119 120

Compounds 121
The PCL was obtained from Prestwick Chemical (Illkirch, France The Z'-factor was computed using R version 3.3.2 35 according to Equation (1) 197 where σ+ is the SD of the positive control, σ-is the SD of the negative control, μ+ the mean of the 199 positive control and μthe mean of the negative control.

201
Screening data processing 202 Plaque2.0 results were further processed and filtered. Data structure and repository 219 The screening data comprise the information collected during assay development, including 220 stability, quality and screening of the PCL itself. The latter two were imaged on two different 221 microscopes. We provide the parameters used for Plaque2.0 image analysis, and the code for 222 the subsequent hit filtering in R. The data structure as available for download at the IDR 1 , 223 accession number idr0081, outlined in Figure 2a. Moreover, the data can be viewed conveniently 224 on the IDR web client, where it is structured as outlined in Figure 2b.

226
Data sets and file types 227 The data provided for download consists of three data sets 1 to 3 (see Figure 2a).

241
The data provided for browsing via the IDR web client consists of five screens A to E (see Figure  242 2b).

243
-idr0081-study.txt summarizes the overall study and the five screens that were performed.

244
-screenA contains the assay stability test plates performed at UZH prior to Z'-factor plates (preZ) 245 and the screen (preScreen). idr0081-screenA-library.txt provides thorough information on the 246 tested compounds including PubChem identifiers and their plate layout. idr0081-screenA-247 processed.txt presents the results of the Plaque2.0-based image analysis. idr0081-screenA-248 mean.txt summarises the infection scores per pre plate.

249
-screenB contains the assay quality test plates (Z'-factor plates a and b) performed at UZH. 250 idr0081-screenB-library.txt provides thorough information on the tested compounds including 251 PubChem identifiers and their plate layout. idr0081-screenB-processed.txt presents the results of 252 the Plaque2.0-based image analysis. idr0081-screenB-mean.txt summarises the infection scores 253 per Z'-factor plate.

254
-screenC contains the assay quality test plates (Z'-factor plates a and b) performed at EPFL. 255 idr0081-screenC-library.txt provides thorough information on the tested compounds including 256 PubChem identifiers and their plate layout. idr0081-screenC-processed.txt presents the results of 257 the Plaque2.0-based image analysis. idr0081-screenC-mean.txt summarises the infection scores 258 per Z'-factor plate.

273
Assay stability 274 The wet-lab screening pipeline was optimized regarding liquid handling, cell seeding, virus 275 inoculum, positive and negative controls, infection time, as well as imaging and image analysis.

276
This ensured a high assay stability and reproducibility. Furthermore, all compounds, especially 277 media and supplements, the BSA for tubing saturation, PFA-and Hoechst-supplemented fixative 278 were prepared as large batch from a single lot and stored as single-use aliquots. Prior to every 279 experiment, assay stability with respect to cell and infection phenotype was tested on pre-plates 280 according to the established wet-lab, imaging and image analysis pipeline. Since the solvent 281 control had already been spotted in 10 µl PBS, no further PBS was added prior to cell seeding. 282 Periodically, the virus stock dilution was tested and adjusted for experiments if necessary.

284
Assay quality determination: Z'-factor 285 The accuracy of the wet-lab, imaging and image analysis pipeline was assessed by two 286 independently imaged and analyzed Z'-factor plates (Table 2 and Figure 3). 3σ Z'-factors of 287 numberOfInfectedNuclei, infectionIndex and numberOfPlaques were in the range of 0.30 to 0.57, 288 scoring good to excellent. totalVirusIntensity (Z'-factors between -0.07 to 0.08) were not suitable 289 to identify HAdV infection inhibitors, while numberOfNuclei (Z'-factors between -1.11 to -8.10) 290 was not a useable readout either. Additionally, the Z'-factors were determined for each of the 16 291 screening plates (Table 3 and Figure 4). 3σ Z'-factors of numberOfInfectedNuclei, infectionIndex 292 and numberOfPlaques were in the range of 0.27 to 0.57, scoring good to excellent.

294
Independent analysis and filtering 295 Imaging, image analysis and screening data processing were performed by two independent 296 research teams at UZH and EPFL, as depicted in Figure 1. Raw and scored infection phenotypes 297 are shown for UZH and EPFL analyses (Supplementary Tables 2, 3 and Supplementary Tables  298 4, 5, respectively). Both dry-lab pipelines confirmed the high assay quality (Tables 2 and 3). 299 During hit filtering, PCL compounds that gave significant toxicity in uninfected cells were excluded 300 during hit filtering ( Figure 5, Table 4). As summarized in Figure 6

449
Of the 1,278 PCL compounds tested, 126 PCL compounds are found to be toxic, as shown in red, and 450 listed in Table 4. A549 cells were treated with PCL compounds in duplicates according to the screening

482
The quality of the screening platform was assessed prior to screening of the PCL by two independent Z'-