1. Introduction
JC polyomavirus (JCPyV), a small non-enveloped double-stranded DNA virus, infects 50–80% of the population [
1,
2]. The virus is shed in urine and is transmitted via the fecal–oral route. Infection commonly occurs during early childhood through adulthood and results in a lifelong infection in the kidneys [
3,
4]. In healthy individuals, the initial infection is maintained at low levels and presents asymptomatically. However, during severe prolonged immunosuppression, JCPyV can spread from the kidneys into the central nervous system, where it results in a lytic infection of the glial cells in the brain, specifically astrocytes and oligodendrocytes [
5,
6]. JCPyV infection of the brain leads to the disease progressive multifocal leukoencephalopathy (PML), which is characterized by rapid neural demyelination and lesion formation. Symptoms of PML include impaired sensory function, paralysis, and the deterioration of cognitive abilities [
7]. Early detection and effective treatments of the underlying immunosuppression have increased the survival rate of PML, but despite these efforts, the disease outcome is often fatal: sometimes as rapid as within a year of symptom onset [
5,
8]. Viral spread and the development of PML is most common in individuals with uncontrolled human immunodeficiency virus (HIV) infection or those undergoing prolonged immunomodulatory treatments for cancer, organ transplant or immune-mediated diseases such as multiple sclerosis (MS) [
9].
Currently, there are no approved targeted treatments for JCPyV infection or PML, and clinicians are limited to treating the underlying immunosuppression by ceasing the use of immunomodulatory drugs or implementing highly active antiretroviral therapy (HAART) for HIV management [
5,
10]. A significant risk of immune system rescue is the development of PML immune reconstitution inflammatory syndrome (IRIS), which is a disease with substantial morbidity that sometimes leads to worsening of the neurological deficits caused by PML [
11,
12]. Direct antiviral therapies most commonly targeting viral entry or replication have been evaluated with limited success. Nucleoside analogues, cytarabine and cidofovir, topoisomerase inhibitor topotecan and serotonin receptor antagonist mirtazapine have shown promising reductions of infection in vitro but lacked clinical benefit or relevance in clinical trials [
5,
13,
14,
15,
16]. The absence of approved and effective treatments for PML emphasizes the critical need for additional research exploring the infectious cycle of JCPyV to identify novel therapeutic targets and antiviral therapies.
JCPyV infection can occur through both receptor-independent entry through extracellular vesicles and receptor-mediated entry [
17]. To initiate infection through receptor-mediated entry, the external component of the JCPyV capsid, viral protein (VP) 1, interacts with alpha2,6-sialic acid containing lactoseries tetrasaccharide (LST-c) and glycosaminoglycans (GAGs) on the host cell membrane [
18,
19]. Following attachment, the virus enters the cell through clathrin-mediated endocytosis using the G protein-coupled receptor (GPCR) type 2 serotonin receptor, 5-hydroxytryptamine receptor (5-HT
2R) subtypes A, B and C [
20,
21,
22]. Super-resolution microscopy studies have shown that JCPyV localizes with 5-HT
2Rs during viral entry and that viral particles induce clustering of the 5-HT
2Rs at times consistent with entry [
23]. For proper viral endocytosis to occur, the 5-HT
2Rs must be bound by scaffolding protein β-arrestin. The G protein-coupled receptor kinase 2 (GRK2) is also required for infection, presumably by initiating β-arrestin recruitment to the receptor [
24,
25,
26,
27]. Activation of the 5-HT
2R also leads to the recruitment of additional endocytic proteins clathrin and adaptor protein (AP2), resulting in viral entry into clathrin-coated pits [
24]. Following clathrin-mediated endocytosis, the virus traffics in endosomes to the ER and lastly to the nucleus, where the viral genome is transcribed and replicated [
28]. β-arrestin recruitment by the 5-HT
2Rs has also been suggested to trigger the host mitogen-activated protein kinase (MAPK) cascade, which is a signaling pathway necessary for JCPyV to complete the infectious cycle [
29,
30]. The MAPK cascade terminates in the phosphorylation of extracellular signal-regulated kinase (ERK), which recruits and activates host transcription factors to drive productive viral infection [
30,
31,
32]. Inhibitors of 5-HT
2Rs, clathrin-mediated endocytosis, and the MAPK cascade significantly reduce JCPyV infection, suggesting that viral entry and GPCR signaling are promising therapeutic targets [
21,
22,
23,
30,
33].
Additional support for viral entry as a promising therapeutic target was discovered in a large-scale high-throughput drug screen of the National Institute of Health Clinical Collection performed by Bond et al. [
34]. In this study, the largest portion of drugs identified as hits that reduced JCPyV infection in vitro (36%) were characterized as receptor agonists/antagonists, particularly of GPCRs, making this group of considerable interest due to the known involvement of GPCRs and GPCR signaling in the JCPyV infectious cycle [
22,
30,
31,
34]. Further, GPCR signaling is a conserved mechanism of infection across viral families, as GPCR antagonists have been shown to reduce the infection of flaviviruses, influenza virus, and SARS-CoV-2, among others [
26].
This study characterized potential antiviral treatments against JCPyV by validating seven receptor-specific inhibitors identified as hits in the high-throughput drug screen published by Bond et al. [
34]. Hits that were approved by the Food and Drug Administration (FDA) with an average target receptor gene count higher than zero in SVGA cells [
35] were evaluated, and drugs reducing infection were further assessed for their mechanistic action. Two GPCR-associated inhibitors, cetirizine and paroxetine, significantly reduce JCPyV infection across immortalized and primary cell types at early time points in the infectious cycle correlating with viral entry and GPCR signaling. The results described herein suggest viral entry and GPCR signaling as possible therapeutic targets to inhibit JCPyV infection and the potential of 5-HT
2R- and GPCR-related antagonists to be repurposed and used as antiviral therapeutics.
2. Materials and Methods
2.1. Cells and Viruses
SVGA cells were maintained in complete minimum essential medium (MEM) (Corning, Corning, NY, USA) containing 10% fetal bovine serum (FBS) (Bio-techne, Minneapolis, MN, USA) 1% penicillin–streptomycin (P/S) (Corning) and 0.2% Plasmocin (InvivoGen, San Diego, CA, USA). Cells were passaged 2–3× weekly upon 90–100% confluency. Human embryonic kidney cells (HEK293A) stably expressing 5-HT
2CR tagged with YFP (HEK-2C-YFP) [
22] or Dendra2 (HEK-2C-Dendra2) [
23] were maintained in Dulbecco’s modified Eagle’s medium containing (DMEM) (Corning) 10% FBS, 1% P/S, and 0.2% Plasmocin. To sustain receptor expression, cells were maintained in 1% G418 (MP Biomedicals, Irvine, CA, USA). Cells were passaged 2–3× weekly upon 90–100% confluency, and media was replaced every 48 h (h). Normal human astrocytes (NHAs) were maintained in complete astrocyte medium (CAM) with 2% FBS, 1% P/S, and 1% astrocyte growth supplement (ScienCell, Carlsbad, CA, USA). Cells were passaged upon 90% confluency in accordance with ScienCell distributer protocol, and media was replaced every 48–72 h. Renal proximal tubule epithelial cells (RPTECs) were maintained in complete renal epithelial growth medium (REGM) (ATCC, Gaithersburg, MD, USA) with an added epithelial cell growth kit (ATCC) and 1% P/S. Cells were passaged upon 90% confluency in accordance with ATCC protocol, and media was replaced every 48 h. All cell lines were propagated in a humidified incubator at 37 °C with 5% CO
2 and tested for mycoplasma every 6 months. SVGA cells and HEK-2C cells were generously provided by the Atwood Laboratory (Brown University, Providence, RI, USA), while RPTECs and NHAs were obtained from ATCC and ScienCell, respectively.
JCPyV strain Mad-1/SVEΔ was generated and propagated as described in [
28,
36]. All experiments, except when purified labeled virus is indicated, were performed using a crude supernatant virus stock titered by focus-forming unit (FFU) infectivity assay in SVGA cells. JCPyV was purified as described in [
28], and pure JCPyV was labeled with Alexa Flour 647 as described in [
18].
2.2. Agonists/Antagonists and Antibodies
All inhibitors used had a purity of at least 98% and include telmisartan (Selleckchem, Houston, TX, USA), paroxetine (Abcam, Cambridge, MA, USA), cetirizine (Selleckchem), pergolide (Medchem Express), pilocarpine (Selleckchem), pindolol (Sigma Aldrich, St. Luois, MO, USA), and rizatriptan (Selleckchem). All drugs were resuspended in DMSO (Tocris Bioscience, Minneapolis, MN, USA), and DMSO was also used as a volume-specific vehicle control. Figures or figure legends list the concentrations used for each inhibitor.
The primary antibodies used in FFU and ICW assays include PAB962, a monoclonal antibody (mAb) derived from a hybridoma supernatant for the detection of the JCPyV large TAg protein graciously provided by the Tevethia Laboratory (Penn State University, Hershey, PA, USA) [
37], and an antibody against phosphorylated ERK (pERK) (Cell Signaling Technology, Danvers, MA, USA). Secondary antibodies used include polyclonal goat anti-mouse and goat anti-rabbit Alexa Fluor 488 and 594 antibodies (Thermo Fisher, Waltham, MA, USA) and LI-COR 800 goat anti-mouse and anti-rabbit antibodies (LI-COR, Lincoln, NE, USA). DAPI (Thermo Fisher) was used to stain cell nuclei in FFU assays, and CellTag 700 (LI-COR) was used as a cell count normalization stain for In-Cell Western assays. Antibodies used for staining after pulldown assay and Western blot include primary antibodies against β-arrestin and GFP (Cell Signaling Technology) as well as a LI-COR 800 anti-rabbit secondary antibody (LI-COR). Entry assays quantified using confocal microscopy were stained using a primary antibody against pan-cadherin (Abcam) and a polyclonal goat anti-mouse Alexa Fluor 488 secondary antibody (Thermo Fisher).
2.3. Cell Viability Assay
The MTS assay (G3581) (Promega, Madison, WI, USA) was used according to the manufacturer’s instruction to determine the cell viability for each cell type after the treatment of specified inhibitor concentrations. To mimic infectivity assay protocols, cells were plated to 70% confluency and pre-treated with drug or DMSO volume control at 37 °C for 1 h. Cells were then mock infected with appropriate media alone and incubated with an inhibitor or DMSO control for 48 h (SVGA and NHAs) or 72 h (RPTECs). MTS reagent was added for 1 h at 37 °C before absorbance was measured at 490 nm using an Agilent BioTek Cytation 5 Imaging Reader (Agilent Technologies, Santa Clara, CA, USA). Drug concentrations maintaining >80% cell viability in comparison to the appropriate DMSO control were deemed to not induce significant toxicity and were considered usable concentrations. Experiments were performed in triplicate wells for at least 3 replicates.
2.4. JCPyV Infection
SVGA, HEK-2C, RPTEC, and NHA cells were plated to 70% confluency in 96-well plates in MEM, DMEM, REGM, CAM or MEM, respectively. The inhibitors and DMSO volume control were diluted in appropriate media to concentrations indicated in figures and incubated at 37 °C for 1 h. Cells were infected with JCPyV at a multiplicity of infection (MOI) of 1 FFU/cell (SVGAs and HEK-2Cs), 3 FFU/cell (RPTECs), or 2 or 3 FFU/cell (NHAs) in the absence of drug at 37 °C for 1 or 2 h. Infections were fed with 100 uL/well of drug- or DMSO-containing media at appropriate concentrations and incubated at 37 °C for 48 h (SVGAs, HEK-2Cs, and NHAs) or 72 h (RPTECs). Cells were then fixed in 4% PFA at RT for 10 min directly in media, media was removed, and 4% PFA was added for another 10 min without media prior to processing via FFU or In Cell Western (ICW) assays.
2.5. FFU Infectivity Assay
Fixed cells were washed in 0.1% PBS-Tween (PBS-T) 3× for 5 min each before staining. Cells were permeabilized with 1% Triton X-100 in PBS at room temp (RT) for 15 min and blocked with 10% goat serum in PBS at RT for 1 h with gentle rocking. JCPyV TAg-specific primary antibody PAB962 (1:5) in PBS was added to the cells at RT for 1 h. Cells were washed 3× for 5 min each with PBS-T and then incubated with a polyclonal goat anti-mouse Alexa Flour 594 or 488 secondary antibody (1:1000) in PBS at RT for 1 h. Again, cells were washed 3× with PBS-T for 5 min each, and cell nuclei were stained with DAPI (1:1000) in PBS at RT for 5 min. PBS-T was used to wash 2×, and PBS was added for storage at 4 °C. All infections were performed in triplicate wells for at least 3 biological replicates.
Infected and stained cells were visualized and quantified using epifluorescence microscopy on a Nikon Eclipse Ti microscope (Micro Video Instruments Inc., Avon, MA, USA). The percent infection was quantified by dividing TAg-positive cells by the number of total cells visualized by DAPI and multiplying by 100. Quantification was performed at 10× or 20× magnification for 5 fields of view (FOV) per well. DAPI-positive cells were counted using an automated binary algorithm in the Nikon NIS Elements Basic Research software (versions 4.50.00 and 6.02.01) separating cells by intensity, diameter, and circularity, while TAg-positive cells were counted manually [
25,
34].
2.6. In-Cell-Western Assay
Fixed cells were washed with PBS-T 3× for 5 min each and permeabilized with TX-100 in PBS at RT for 15 min. Cells were then blocked with TBS Odyssey Blocking Buffer (LI-COR) at RT for 1 h before being incubated with primary antibody specific for JCPyV TAg (PAB962) (1:5) or pERK (1:750) in TBS Odyssey Blocking Buffer at 4 °C overnight (O/N) with gentle rocking. Cells were washed 3× in PBS-T 5 min each and then incubated with LI-COR 800 secondary anti-mouse or anti-rabbit antibody (1:10,000) and CellTag 700 (1:500) at RT for 1 h. Lastly, cells were washed 3× with PBS-T.
Cells were washed with PBS, and liquid was removed before scanning at 700 and 800 nm channel intensities using a LI-COR CLx Infrared Imaging system. The scanner was set to 42 μm resolution, medium quality, and 3.0 mm focus offset [
32]. Image studio software was used to align channels after scanning, and the ICW module was used to outline each well. The intensity values inside the wells for the 700 and 800 channels were recorded, and infection was quantified by dividing the 800-channel intensity value by the 700-channel intensity value and multiplying by 100. Finally, values were normalized to the appropriate control [
34].
2.7. Time of Addition Assay
SVGA cells were plated to 70% confluency in 96-well plates using complete MEM. Treatment with inhibitors was initiated either 1 h before infection (pre-treatment), during infection (0 hpi), or at indicated time points after infection (2, 4, 6, 12 or 24 hpi). For all treatment conditions, drugs were diluted in 10% MEM and incubated at 37 °C. Infection with JCPyV was performed in the presence of drug at 37 °C unless otherwise indicated, and infection was fed with 10% MEM containing drug or 10% MEM alone. At 48 hpi, cells were fixed with 4% PFA at RT for 10 min directly in media, media was removed, and 4% PFA (without media) was added for another 10 min. Cells were stained for JCPyV TAg (1:5) and analyzed via ICW assay. Experiments were performed in triplicate for 3 biological replicates.
2.8. pERK Signaling Assay
SVGA cells were plated to 90% confluence in 96-well plates using complete MEM. Inhibitors were diluted to concentrations indicated in figures in 10% MEM and incubated at 37 °C for 1 h. Cells were then pre-chilled at 4 °C for 15 min before being mock infected with MEM or infected with JCPyV in the absence of drugs at 4 °C for 1 h. Infections were fed with 10% MEM containing appropriate drug concentrations at 37 °C for 15 min and then fixed in 4% PFA at RT for 10 min directly in media, media was removed, and 4% PFA (without media) was added for another 10 min. Cells were stained for pERK (1:750) and CellTag (1:500) and analyzed via ICW assay. Experiments were performed in triplicate for 3 biological replicates.
2.9. Attachment Assay by Flow Cytometry
SVGA or HEK-2C cells were plated to 100% confluency in 12-well plates and treated with DMSO or drug at 37 °C for 1 h. Cells were washed with PBS and then incubated with Cellstripper (Corning) at 37 °C for 15 min to detach cells from the plate. Detached cells were pelleted at 452× g at 4 °C for 5.5 min and washed with PBS. JCPyV labeled with Alexa Fluor 647 (JCPyV-647) in phenol red free MEM was added to cells and incubated on ice for 1 h with agitation every 15 min. Cells were washed with PBS, pelleted by centrifugation, then fixed in 4% PFA on ice for 10 min. Each sample was resuspended in PBS and analyzed for viral attachment by flow cytometry using an LSRII system (BD Biosciences) with a 650 nm AP-C laser line. At least 10,000 events were recorded per sample, and data analyses were performed using BD FACSDiva (version 6.2) and FlowJo software (version 10.10). Gates were generated to exclude dead and complex cells.
2.10. Entry Assay by Confocal Microscopy
SVGA or HEK-2C-YFP cells were plated to 70% confluence in 96-well glass-bottom plates (CellVis number 1.5, CellVis, Sunnvale, CA, USA). Cells were pre-treated with drug or DMSO control in 10% MEM (SVGA) or 10% DMEM (HEK-2C-YFP cells) at 37 °C for 1 h and then pre-chilled at 4 °C for 45 min. JCPyV labeled with Alexa Fluor 647 [
17] (MOI = 4 FFU/cell) was added at 4 °C for 1 h, and cells were then fed with pre-warmed media containing drug or DMSO at 37 °C for 2 h. Fixation was completed by adding 4% PFA at RT for 10 min with media present, removing the media, and adding 4% PFA (without media) for another 10 min. After fixation, SVGA cells were washed with PBS 3× and incubated in block buffer (2% goat serum, 0.2% TX-100 and 0.1% BSA) at RT for 1 h while rocking. A primary antibody against pan-cadherin (1:75) in block buffer was added to the cells and incubated at 4 °C O/N while rocking. Cells were washed 3× with PBS and incubated with anti-mouse 488 secondary antibody (1:1000) at RT for 1 h while rocking. Another 3 washes with PBS were performed, and DAPI (1:1000) was added at RT for 10 min while rocking. After fixation, HEK-2C-YFP cells were washed 3× with PBS, stained with DAPI (1:1000) at RT for 10 min while rocking, and then washed 3× with PBS. Plates were stored in PBS until visualization. Samples were visualized at 63× magnification (oil immersion) using a Zeiss LSM 980 Confocal Microscope with Airyscan 2 and the Zen Blue 3.8 software. Diodes of 405 nm, 488 nm and 639 nm were used to acquire images, and cross-sections of at least 30 cells per sample were analyzed. Images were analyzed in ImageJ for viral internalization. HEK-2C-YFP cell images were subjected to a median filter in ImageJ to remove background noise. Region of interest (ROIs) were defined using the polygon selection tool to exclude the plasma membrane in either the 488 channel or differential interference contrast (DIC) images [
23]. Viral internalization was measured by comparing relative fluorescence units per cell for background corrected samples. Experiments were performed for 3 biological replicates, and graphs represent at least 90 cells per treatment.
2.11. Serotonin Receptor Pulldown Assay
HEK-2C cells expressing 5-HT2CR tagged with YFP were plated to 90% in 6-well plates and pre-treated with paroxetine or DMSO at 37 °C for 1 h. Plates were then pre-chilled at 4 °C for 45 min and incubated with either JCPyV (MOI = 3 FFU/cell) or 5-HT [200 μM] in 10% DMEM at 4 °C for 1 h. Pre-warmed paroxetine/DMSO-containing DMEM was added to each well at 37 °C for 15 min before removal by scraping. Detached cells were pelleted at 414× g at 4 °C for 10 min and resuspended in lysis buffer (10 mM Tris HCl, 150 mM NaCl, 0.5 mM EDTA, and 0.5% Triton X-100 [pH 7.5]) supplemented with phosphatase (Sigma-Aldrich) and protease (Sigma-Aldrich) inhibitors on ice for 30 min with vigorous mixing every 10 min. Cells were pelleted at 21,130× g at 4 °C for 10 min, while GFP-Trap magnetic agarose beads were washed 3× with wash buffer (10 mM Tris HCl, 150 mM NaCl, 0.5 mM EDTA, and 0.5% Triton X-100 [pH 7.5]) and separated using a magnetic tube rack. The sample supernatant was mixed with the beads and incubated at RT for 2 h while rotating. Beads were washed 3× before being transferred to a new tube in 50 μL buffer. Samples were mixed 1:1 with Laemmli sample buffer (BioRad, Hercules, CA, USA) and boiled at 95 °C for 10 min. Samples were processed by SDS page using a 10% TGX mini gel (BioRad), and proteins were transferred to a nitrocellulose membrane (BioRad) at 2.5 amps (25 V) for 3 min using a BioRad Semidry Transblot unit. Membranes were equilibrated in TBS for 5 min followed by blocking using Odyssey blocking buffer (LI-COR) at RT for 1 h while rocking. TBS-T was used to wash membranes 3× before incubating with a β-arrestin 1/2 primary monoclonal antibody (1:500) in block buffer at 4 °C O/N. Membranes were washed 3× and incubated with LI-COR 800 anti-rabbit secondary antibody (1:10,000) in block buffer at RT for 1 h. Another TBS wash was completed before membranes were imaged on a LI-COR Odyssey CLx system using the 800 channel. Membranes were stripped of antibodies using NewBlot stripping buffer (LI-COR) according to manufacturer’s instructions and then re-probed with an anti-GFP primary monoclonal antibody (1:1000) to detect 5-HT2R-YFP at 4 °C O/N. Blots were then washed, incubated with LI-COR 800 anti-rabbit secondary antibody for the detection of 5-HT2R-YFP, and membranes were washed and imaged as previously described. ImageJ was used to determine percentage of bound β-arrestin by comparing the background subtracted mean signal of each β-arrestin band to the corresponding receptor protein band. Within each sample, the β-arrestin band was normalized to 5-HT2 expression, and results were compared across paroxetine- or DMSO-treated samples challenged with either 5-HT or JCPyV.
2.12. Receptor Cluster Analysis by Super-Resolution Microscopy
HEK-2C-Dendra2 cells were plated to 50% confluency in 96-well glass-bottom plates (CellVis) using phenol-free DMEM supplemented with 4 mM glucose, 10% FBS, and 1% G418 (Corning). Cells were treated with paroxetine or DMSO (vehicle) at 37 °C for 1 h and then pre-chilled at 4 °C for 1 h. Cells were incubated with JCPyV-647 at 4 °C for 1 h and then fed with pre-warmed media containing drug or DMSO for 5 or 15 min. Cells were fixed with 4% PFA for 10 min and washed 2× with PBS. Samples were stored in PBS until imaging [
23].
Imaging was completed utilizing a custom fluorescence photoactivation localization (FPALM) microscope [
38]; an Olympus IX71 inverted microscope with a 60X objective, 2× telescoping lens in the detection path, and an Andor iXon + electron-multiplying charge-coupled device for detection. Samples were imaged in GLOX buffer (10%
w/
v glucose, 10 ug/mL Catalase, 50 ug/mL Glucose OXidase, 1% beta-mercaptoethanol, and 10 mM Tris pH 7.4) to induce Alexafluor-647 blinking for directed stochastic optical reconstruction microscopy (dSTORM) [
39,
40]. Dendra2 activation and readout was achieved utilizing 405 nm and 558 nm lasers, respectively. Alexa Fluor-647 readout was achieved utilizing a 638 nm laser. Prior to beam splitting, two 405/488/561/635 notch filters (Semrock, Rochester, NY, USA) and a 561-notch filter (Semrock) located in the microscope turret were utilized to filter out the laser background. A DiO 405/488/561/635 multiband dichroic mirror (Semrock) also in the microscope turret was used to reflect lasers into the back aperture of the objective. Fluorescence collected by the objective passes through a tube lens and into a 2× telescope. After the telescope, the fluorescence reaches a FF580-FDi01 dichroic mirror (Semrock) that splits the light into the transmitted (red) and reflected (green) channels. The red channel has a 664 nm RazorEdge LP filter (Semrock), while the green channel has a 580/40 BrightLine filter (Semrock). A total of 10,000 images per cell were captured at a frame rate of 30 Hz and an electron multiplication gain of 200 [
23].
Images were localized, concatenated, and analyzed (and rendered) utilizing custom MATLAB scripts [
23,
41]. Briefly, images were background subtracted utilizing a temporal median filter, and thresholds for blinking were set by sample. An α-histogram of relative fluorescence in both the red and green channels was utilized to select for Dendra2 and Alexaflour-647 localizations [
41]. Dendra2 was identified as α < 0.7, and AlexaFluor-647 was identified as α > 0.83. Images from single cells were then concatenated, drift corrected, and duplicate pixels were removed [
42]. Clusters were identified using a single-linkage cluster analysis (SCLA) algorithm [
43]. Clusters were defined as imaged areas with a density of at least three times the cell-averaged density (units of localizations per unit area), a radius of at least 50 nm, and a minimum number of 10 localizations. Cluster properties were then analyzed for area, density, and perimeter, or radial distribution function (RDF) analysis was performed on each cluster, averaged over all clusters within a given cell, and then further analyzed or averaged [
44]. Cluster density is represented by the average cluster density per cell. RDF functions show the average density at distance X from the cluster center. RDF functions are averages of all clusters found, and they are not differentiated by cell or replicate.
2.13. Statistical Analysis
Two-sample Student’s t-tests were performed using Microsoft Excel (Version 16.89.1) to determine statistical significance. Mean values of triplicate samples were compared assuming unequal variance. Variation in a given population was determined by calculating the standard error of the mean (SEM) by dividing the standard deviation of a population by the square root of the sample size.
Statistical differences in cluster densities were determined by a Wilcoxon signed-rank test, assuming unequal variance in population distributions due to the bimodal structure of population densities. RDF statistical analysis compared densities as a function of radius r measured from the cluster center (r = 0) out to r = 500 nm with a spacing of 10 nm. Statistical significance was quantified using a Wilcoxon signed-rank test with a Bonferroni–Holm correction.
4. Discussion
The lack of therapeutic options for treating JCPyV infection and PML highlights an opportunity for research to increase our understanding of viral infection and consequently identify novel therapeutic targets. To bridge existing knowledge gaps and discover potential antiviral therapies, this study validated seven FDA-approved, GPCR-associated hits identified in a high-throughput drug screen for reduction in JCPyV infection (
Table 1) [
34]. Upon further investigation two therapeutics, cetirizine and paroxetine were found to significantly reduce JCPyV infectivity in immortalized SVGA cells and were selected for further characterization of their antiviral properties (
Figure 1). The majority of drugs that did not reliably reduce JCPyV infection were non-5-HT
2R-related agonists or antagonists. This supports previously established knowledge that JCPyV utilizes 5-HT
2Rs to mediate viral entry through β-arrestin-driven endocytosis [
20,
22]. Moreover, this study also suggests that other GPCRs, such as the histamine H1 receptor inhibited by cetirizine, may play a role in the JCPyV infectious cycle.
Cetirizine, a histamine H1 receptor antagonist currently used to relieve allergies [
46], and paroxetine, a serotonin reuptake transport inhibitor and GRK2 inhibitor prescribed for depression [
47], were further evaluated for their capacity to reduce JCPyV infection in primary cells. Compared to immortalized cells, primary cell lines provide a more clinically accurate human cell model void of variances in signaling pathways caused by immortalization. The treatment of NHA cells, a primary human astrocyte cell type, with cetirizine and paroxetine resulted in a significant reduction in JCPyV infection using both therapeutics (
Figure 2A,C), indicating a conserved drug mechanism across immortalized and primary cell lines for both paroxetine and cetirizine. In RPTECs, a primary human kidney cell line, cetirizine and paroxetine also resulted in a significant reduction in viral infection (
Figure 2B,D), suggesting that the mechanism of infection targeted by these drugs is conserved across both major cell types targeted by JCPyV in the host. Notably, the reduction in JCPyV infection in NHAs and RPTECs with paroxetine or cetirizine treatments yielded only minimal dose-dependent effects, which may be due to the narrow range of drugs tested and that higher doses were not included due to drug toxicity at higher concentrations.
Interestingly, cetirizine has also been shown to reduce the infectious ability of lentivirus pseudotyped with the SARS-CoV-2 spike protein [
48] and was deemed a promising therapeutic approach in the management of COVID-19 during the recent pandemic [
49]. Antihistamines are mainly developed for their use as allergy medications, but recent studies have revealed an increased applicability as anti-infective agents against viruses, bacteria, and parasites. Histamine receptor antagonists have been identified to reduce infection of several viruses such as SARS-CoV-2, filoviruses, and influenza virus through both in vitro and in vivo studies. The most common mechanism of action for these drugs involves targeting viral entry either by blocking receptor internalization or fusion events [
50].
To characterize how cetirizine reduces JCPyV infection, the drug was applied at different time points throughout the infectious cycle, and a significant reduction in infection was observed at all evaluated time points. However, the greatest reduction occurred up to 6 h post-infection during the early stages of the infectious cycle (
Figure 3A) [
25,
28]. Given the known function of cetirizine as a receptor antagonist, the impact of the drug on viral attachment and entry was explored using flow cytometry and confocal microscopy, respectively. Surprisingly, neither JCPyV attachment nor entry was reduced by treatment with cetirizine, suggesting that cetirizine is impacting a post-entry step in the infectious cycle (
Figure 3B,C). Based on these findings, it is likely that cetirizine affects viral trafficking, signaling, or possibly viral transcription, yet the early timing of drug addition may have led to more significant impacts on infection in the time course experiment due to the length of treatment time (
Figure 3A). Moreover, it is possible that the mechanism by which cetirizine reduces JCPyV infection is independent of its currently known function as a histamine receptor antagonist [
46]. Additional exploration is necessary to further characterize how cetirizine inhibits viral infection.
Paroxetine is a serotonin reuptake transporter antagonist, and it is also an inhibitor of GRK2, which is an essential protein that mediates β-arrestin binding to the 5-HT
2Rs [
26,
51,
52]. Prior research from our lab has shown that GRK2 is required for JCPyV internalization and infection, and the knockdown of GRK2 using siRNA reduces β-arrestin coupling to the 5-HT
2Rs [
24]. β-arrestin and GRK involvement are required for the productive infection of many other viruses belonging to vastly diverse virus families [
26]. Filoviruses, flaviviruses, and SARS-CoV-2 have all evolved to rely on signaling emanating from these two proteins, making them promising therapeutic targets [
26].
To define the antiviral mechanism of paroxetine, a time-of-addition assay was performed, and paroxetine was found to reduce infectivity early in the infectious cycle (0–12 hpi) (
Figure 4A), mainly correlating to viral attachment and entry. Additional investigation by flow cytometry demonstrated that viral attachment was not affected by paroxetine treatment in either glial (SVGA) cells or kidney (HEK-2C) cells (
Figure 4 and
Figure 5). However, an analysis of viral entry by confocal microscopy revealed that paroxetine significantly impaired viral internalization in both cell types (
Figure 4 and
Figure 5), suggesting that reduced viral entry could be due to the altered receptor properties induced by paroxetine.
To further explore whether paroxetine treatment affects 5-HT
2R properties, receptor analysis upon infection was performed using super-resolution microscopy FPALM. Previous research from our lab using FPALM has shown that JCPyV induces serotonin receptor (5-HT
2R) clustering at 5 mpi, which is consistent with clathrin-mediated endocytosis of the virus [
23,
24]. In this study, FPALM analysis demonstrated that 5-HT
2cR clusters induced by JCPyV infection were decreased in cluster density and had altered cluster shapes in the presence of paroxetine at 5 mpi compared to the DMSO control (
Figure 5E,F). Additionally, a bimodal distribution of cluster densities was observed in the paroxetine-treated cells at 5 mpi (
Figure 5E), suggesting paroxetine may not fully inhibit receptor desensitization but rather reduce the number and size of clusters that could facilitate entry, reducing the chance of JCPyV internalization. This is further supported by the RDF analysis showing paroxetine reduces receptor density in clusters at 5 mpi up to 200 nm from the cluster center but does not entirely eliminate receptor clustering. However, this reduction occurs in the most dense clusters based on the thresholding used to identify clusters, suggesting the effects of paroxetine are particularly potent in reducing potential sites of viral internalization. Moreover, at 15 mpi, the cluster density in DMSO-treated cells is reduced in comparison to DMSO-treated cells at 5 mpi, suggesting that while the receptor is still clustering, the clusters are less dense at later times. This suggests that 5-HT
2CR clusters at 5 mpi in response to JCPyV infection, and paroxetine treatment reduces the density of clusters, resulting in a decreased internalization of viral particles. Taken together, these data are consistent with our previous work and demonstrate that early receptor clustering may play a crucial role in viral endocytosis, which is reduced by paroxetine treatment. As a GRK2 inhibitor, it was hypothesized that treatment with paroxetine would prevent the recruitment of β-arrestin to the receptor, which is critical for JCPyV-receptor mediated endocytosis and infection.
To determine whether paroxetine inhibits β-arrestin recruitment to the receptor, 5-HT
2R pulldown assays were performed. The results demonstrated a sizable reduction in receptor-bound β-arrestin upon treatment with paroxetine (
Figure 6B), supporting the idea that paroxetine impacts viral entry by inhibiting β-arrestin binding to the 5-HT
2R likely by inhibiting GRK2. Utilizing GRK2 as a drug target has also shown to be an effective method against influenza virus infection. Paroxetine treatment reduced influenza virus infection during viral entry, as this virus relies on GRK, β-arrestin, and MAPK signaling for internalization and uncoating [
26,
51]. Similar to influenza virus, β-arrestin recruitment during JCPyV infection activates the MAPK signaling pathway, resulting in the phosphorylation of ERK, which is an event required for productive infection [
30,
31,
53].
The impact of paroxetine treatment on pERK activation was characterized, and it was confirmed that the drug significantly reduced MAPK signaling during infection (
Figure 6C). Decreased pERK expression likely results as a downstream effect of the lowered β-arrestin binding during viral entry caused by paroxetine treatment. However, it is also possible that the mere presence of JCPyV inside the host cell could activate the MAPK pathway independently of GRK2 and β-arrestin, and thus a decreased level of internalized virus in the presence of paroxetine could result in a subsequent reduction in pERK expression.
Although viral entry seems to be an important step affected by paroxetine, the reduction in internalization is considerably lower than the total amount of reduction observed during the initial validation assay. One explanation may be that paroxetine also inhibits other steps in the JCPyV infectious cycle, resulting in an additive effect. Alternatively, this could be due to JCPyV virions entering through alternative mechanisms such as a non-receptor-dependent process [
54]. For this to be established, however, additional research of other potential pathways is warranted. Interestingly, paroxetine was identified in a NIH-CC drug screen as an inhibitor of coronavirus [
55] mouse hepatitis virus (MHV) and was demonstrated to reduce MHV entry. In addition to paroxetine, several other antidepressants have demonstrated antiviral activity against Ebola virus and SARS-CoV-2 through mechanisms involving viral replication and interaction with viral glycoproteins or by exhibiting direct virucidal effects [
56,
57]. In a study where paroxetine reduced the viral RNA copy number of SARS-CoV-2 in vitro, a combination of inhibitory effects was observed, supporting the proposal that paroxetine can inhibit viral infection through multiple modes of action [
56]. Antidepressants have also been revealed to have antiviral benefits in clinical studies where a significant association between the intake of these drugs and a decreased risk of both mortality and hospitalization was observed in patients with ongoing SARS-CoV-2 infection [
56,
58].
Altogether, the results from this study characterize receptor-associated inhibitors paroxetine and cetirizine as potential therapeutics to treat JCPyV infection and the resulting neurological disease PML. It is expected that the GPCR inhibitors in this study would interfere with receptor-dependent endocytosis, but not receptor-independent endocytosis [
17], yet cetirizine did not reduce viral entry. Previous reports have identified the viral entry of JCPyV as a promising target for drug discovery, but despite reassuring results in vitro using drugs like mirtazapine, clinical trials presented with mixed results and limited benefit [
16,
33,
59,
60]. Both drugs validated in this report are currently FDA approved and could possibly be repurposed into antiviral therapies with limited side effects and toxicity. This process is a well-demonstrated approach and was recently utilized for the discovery of treatment options during the COVID-19 pandemic [
61]. Paroxetine is known to cross the blood–brain barrier and could therefore be used to reduce infection both in the kidneys and the brain, whereas cetirizine has a diminished ability to cross the blood–brain barrier and would be limited to treatment in the kidneys [
62,
63]. Further, it would be of interest to determine the effect of combining paroxetine with cetirizine or perhaps combining these inhibitors with other FDA-approved antiviral agents that block a post-receptor entry step in the viral lifecycle. For example, brincidofovir, an FDA-approved drug for poxvirus, inhibits mouse polyomavirus (MuPyV) infection in vitro by blocking a post-transcriptional step in MuPyV infection and reduces viremia and infection in the kidney in vivo [
64]. It is possible that combining brincidofovir with paroxetine or cetirizine might result in a more profound inhibition of JCPyV infection. Thus, further exploration of the clinical implications of cetirizine and paroxetine and possible combinations with antivirals is warranted.