key: cord-0778343-aexlwq79
authors: Kaptein, Suzanne JF; Jacobs, Sofie; Langendries, Lana; Seldeslachts, Laura; ter Horst, Sebastiaan; Liesenborghs, Laurens; Hens, Bart; Vergote, Valentijn; Heylen, Elisabeth; Maas, Elke; De Keyzer, Carolien; Bervoets, Lindsey; Rymenants, Jasper; Van Buyten, Tina; Thibaut, Hendrik Jan; Dallmeier, Kai; Boudewijns, Robbert; Wouters, Jens; Augustijns, Patrick; Verougstraete, Nick; Cawthorne, Christopher; Weynand, Birgit; Annaert, Pieter; Spriet, Isabel; Velde, Greetje Vande; Neyts, Johan; Rocha-Pereira, Joana; Delang, Leen
title: Antiviral treatment of SARS-CoV-2-infected hamsters reveals a weak effect of favipiravir and a complete lack of effect for hydroxychloroquine
date: 2020-06-19
journal: bioRxiv
DOI: 10.1101/2020.06.19.159053
sha: 4a4c39c386ca1c7d91d03cb65b2ed64572452af7
doc_id: 778343
cord_uid: aexlwq79

SARS-CoV-2 rapidly spread around the globe after its emergence in Wuhan in December 2019. With no specific therapeutic and prophylactic options available, the virus was able to infect millions of people. To date, close to half a million patients succumbed to the viral disease, COVID-19. The high need for treatment options, together with the lack of small animal models of infection has led to clinical trials with repurposed drugs before any preclinical in vivo evidence attesting their efficacy was available. We used Syrian hamsters to establish a model to evaluate antiviral activity of small molecules in both an infection and a transmission setting. Upon intranasal infection, the animals developed high titers of SARS-CoV-2 in the lungs and pathology similar to that observed in mild COVID-19 patients. Treatment of SARS-CoV-2-infected hamsters with favipiravir or hydroxychloroquine (with and without azithromycin) resulted in respectively a mild or no reduction in viral RNA and infectious virus. Micro-CT scan analysis of the lungs showed no improvement compared to non-treated animals, which was confirmed by histopathology. In addition, both compounds did not prevent virus transmission through direct contact and thus failed as prophylactic treatments. By modelling the PK profile of hydroxychloroquine based on the trough plasma concentrations, we show that the total lung exposure to the drug was not the limiting factor. In conclusion, we here characterized a hamster infection and transmission model to be a robust model for studying in vivo efficacy of antiviral compounds. The information acquired using hydroxychloroquine and favipiravir in this model is of critical value to those designing (current and) future clinical trials. At this point, the data here presented on hydroxychloroquine either alone or combined with azithromycin (together with previously reported in vivo data in macaques and ferrets) provide no scientific basis for further use of the drug in humans.

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) first emerged in Wuhan, 51

China in December 2019 1 . From there, the virus rapidly spread around the globe, infecting 52 more than 8 million people so far (June 18) [https://covid19.who.int/]. SARS-CoV-2 is the 53 causative agent of coronavirus disease 2019 . Common clinical manifestations of 54 COVID-19 are fever, dry cough, paired in a minority of patients with difficult breathing, muscle 55 and/or joint pain, headache/dizziness, decreased sense of taste and smell, diarrhea, and 56 nausea 2 . A small subset of patients will develop to acute respiratory distress syndrome 57 (ARDS), characterized by difficult breathing and low blood oxygen levels, which may directly 58 result into respiratory failure 2 . In addition, an overreaction of the host's immune and 59 inflammatory responses can result in a vast release of cytokines ('cytokine storm'), inducing 60 sepsis and multi-organ damage, which may lead to organ failure 3 . To date, more than 440,000 61 patients worldwide succumbed to COVID-19. Hence, in response to the ongoing pandemic 62 there is a desperate need for therapeutic and prophylactic options. 63

At present, no specific antiviral drugs have been developed and approved to treat infections 64 with human coronaviruses. Nonetheless, antiviral drugs could fulfill an important role in the 65 treatment of COVID-19 patients. Slowing down the replication of SARS-CoV-2 by antiviral 66 treatment could be beneficial and prevent or alleviate symptoms. In addition, antiviral drugs 67 could be used as prophylaxis to protect health care workers and high-risk groups. However, a 68 specific, highly potent antiviral drug for SARS-CoV-2 will take years to develop and evaluate 69 in clinical studies. Therefore, the main focus for COVID-19 treatment on the short term is on 70 hamsters were treated daily for 5 consecutive days with either hydroxychloroquine or 158 favipiravir, starting 1 day prior to exposure to the index hamster. 159

To study the contribution of the fecal-oral route to the overall transmission of SARS-CoV-2, 160 index hamsters were inoculated as described earlier. On day 1 or 3 pi, the index hamsters 161 were sacrificed after which sentinel hamsters were placed in the dirty cages of the index 162 hamsters. Food grids and water bottles were replaced by clean ones to minimize virus 163 transmission via food or water. At day 4 post exposure, the sentinels were sacrificed. Tissues 164 (lung, ileum and stool) were collected from index and sentinel hamsters and processed for 165 detection of viral RNA and infectious virus. 166

Hydroxychloroquine (HCQ) and its active metabolite desethylhydroxychloroquine (DHCQ) 168

were quantified in EDTA-plasma samples. A total of (i) 50 µL sample and (ii) 10 µL of internal 169 standard (IS) solution (hydroxychloroquine-d4 1500 ng/mL in water) were added to a tube and 170 mixed. After addition of 50 µL 5% perchloric acid, samples were shaken for 5 min and 171 centrifuged for 5 min at 16,162 g. Five µL of the supernatant was injected onto the HPLC-172 column. 173 HPLC analysis was performed using a Shimadzu Prominence system (Shimadzu, Kioto, 174 Japan) equipped with a Kinetex C18 column (100mm length x 2.1mm i.d., 2.6 µm particle size) 175 1/x weighting) and DHCQ (quadratic 1/x² weighting) were between 10 and 2250 ng/mL. Between-run imprecision over all QC levels (10, 25, 400, 2000 ng/mL) ranged from 2.84 to 184 11.4% for HCQ and from 5.19 to 10.2% for DHCQ. 185

Starting from the measured total trough plasma concentrations measured at sacrifice after 4 187 or 5 days of HCQ treatment, total lung cytosolic concentrations of HCQ were calculated. First, 188 the mean trough total plasma concentration of HCQ was used as a starting point to estimate 189 the whole blood concentrations considering a blood to plasma ratio of 7.2, as reported by Tett 190 and co-workers 19 Subsequently, as the HCQ efficacy target is intracellular, the cytosolic / total HCQ 201 concentration ratio was estimated, based on (i) relative lysosomal lung tissue volume, as well 202 as the contributions of interstitial and intracellular volumes to total lung volume and (ii) the pH 203 partition theory applying a pKa value of HCQ of 9.67. Based on these calculations (data not 204 shown), lung cytosolic HCQ concentrations are corresponding to 6% of the total lung tissue 205

Total cytosolic lung tissue concentration=total lung tissue concentration ×0.06 207

The calculated total cytosolic lung concentration was compared with EC50 concentrations 209 previously reported in literature, ranging from 0.72 µM to 17.3 µM 21-23 . 210

Hamster tissues were collected after sacrifice and were homogenized using bead disruption 212 (Precellys) in 350 µL RLT buffer (RNeasy Mini kit, Qiagen) and centrifuged (10,000 rpm, 5 min) 213 to pellet the cell debris. RNA was extracted according to the manufacturer's instructions. To 214 extract RNA from serum, the NucleoSpin kit (Macherey-Nagel) was used. Of 50 μL eluate, 4 215 μL was used as a template in RT-qPCR reactions. RT-qPCR was performed on a 216

LightCycler96 platform (Roche) using the iTaq Universal Probes One-Step RT-qPCR kit 217 (BioRad) with N2 primers and probes targeting the nucleocapsid 16 . Standards of SARS-CoV-218 2 cDNA (IDT) were used to express viral genome copies per mg tissue or per mL serum. 219

Lung tissues were homogenized using bead disruption (Precellys) in 350 µL minimal essential 221 medium and centrifuged (10,000 rpm, 5min, 4°C) to pellet the cell debris. To quantify infectious 222 SARS-CoV-2 particles, endpoint titrations were performed on confluent Vero E6 cells in 96-223 well plates. Viral titers were calculated by the Reed and Muench method using the Lindenbach 224 calculator 24 and were expressed as 50% tissue culture infectious dose (TCID50) per mg tissue. 225

For histological examination, the lungs were fixed overnight in 4% formaldehyde and 227 embedded in paraffin. Tissue sections (5 μm) were analyzed after staining with hematoxylin 228 and eosin and scored blindly for lung damage by an expert pathologist. The scored 229 parameters, to which a cumulative score of 1 to 3 was attributed, were the following: 230 congestion, intral-alveolar hemorrhagic, apoptotic bodies in bronchus wall, necrotizing 231 bronchiolitis, perivascular edema, bronchopneumonia, perivascular inflammation, 232 peribronchial inflammation and vasculitis.

Micro-CT data of hamster lungs were acquired in vivo using dedicated small animal micro-CT 235 scanners, either using the X-cube (Molecubes, Ghent, Belgium) or the Skyscan 1278 (Bruker 236

Belgium, Kontich, Belgium). In brief, hamsters were anaesthetized using isoflurane (2-3% in 237 oxygen) and installed in prone position into the X-cube scanner using a dedicated imaging 238 bed. A scout view was acquired and the lung was selected for a non-gated, helical CT 239 acquisition using the High-Resolution CT protocol, with the following parameters: 50 kVp, 960 Visualization and quantification of reconstructed micro-CT data were performed with 249

DataViewer and CTan software (Bruker Belgium). As primary outcome measure, a semi-250 quantitative scoring of micro-CT data was performed as previously described 25 . Visual 251 observations were blindly scored (from 0 -2 depending on severity, both for parenchymal and 252 airway disease) on 5 different, predefined transversal tomographic sections throughout the 253 entire lung image for both lung and airway disease by two independent observers and 254 averaged. Scores for the 5 sections were summed up to obtain a score from 0 to 10 reflecting 255 severity of lung and airway abnormalities compared to scans of healthy, WT control hamsters. 256

As secondary measures, imaging-derived biomarkers (non-aerated lung volume, aerated lung 257 volume, total lung volume and respective densities within these volumes) were quantified as 258 previously 16,26,27 or a manually delineated volume of interest covering the lung, avoiding the heart and main blood vessels. The threshold used to distinguish aerated from non-aerated 260 lung volume was manually defined and kept constant for all data sets 26, 27 . 261

GraphPad Prism (GraphPad Software, Inc.). was used to perform statistical analysis. 263

Statistical significance was determined using the non-parametric Mann Whitney U-test. P 264 values of ≤0.05 were considered significant. 265

We further characterized SARS-CoV-2 infection and readouts of disease in hamsters to be 268 replicates in the lungs of the hamsters, with viral RNA being detected in the lungs from day 1 273 pi and reaching a maximum level of ~7 log10 RNA copies/mg tissue at 4 days pi (Fig 1A) . A 274 similar kinetic profile was found in the ileum and stool samples, albeit at lower levels of 2-3 275 log10 RNA copies/mg of tissue. Titrations of homogenized lung tissue contained infectious 276 particles from 1 day pi and reached levels of ~5 log10 TCID50/mg tissue from day 2 pi onwards 277 (Fig 1B) , which is in line with the viral RNA levels. Infected animals displayed a slight weight 278 loss of about 5% by day 2 pi, which was completely resolved by day 4 pi (Fig 1C) . infiltrates and lung consolidation on day 3 pi (Fig 1D) . Analysis by H&E staining of lungs of 286 infected hamsters at day 4 pi showed signs of bronchopneumonia and peribronchial 287 inflammation, which were not present at the day of inoculation (Fig 1E) . 288

Next, we treated hamsters with antiviral molecules for four consecutive days starting one hour 290 before intranasal infection with SARS-CoV-2. At day 4 pi, a micro-CT scan was performed, 291 after which the animals were sacrificed and organs were collected for quantification of viral 292 RNA, infectious virus titers and lung histopathology (Fig 2A) . Twice-daily treatment with 293 favipiravir was done orally with a loading dose of 600 mg/kg/day at day 0 pi and 300 mg/kg/day 294 from day 1 pi onwards. Favipiravir-treated hamsters presented a decrease of 0.9 log10 RNA 295 copies/mg lung tissue, compared to untreated infected hamsters ( Fig 2B) ; a lesser effect was 296 observed in the ileum and stool of treated animals (Fig 2B) . A modest reduction in infectious 297 titers of 0.5 log10 TCID50/mg was observed in the lungs of favipiravir-treated animals (Fig. 2C) . 298 Treatment with favipiravir caused over 5% weight loss at day 3 and 4 pi, which is slightly more 299 than that of the untreated animals (Fig. 2D ). This could be due to the effect of administering a 300 relatively high volume of compound per os (which was at the limit of 10 mL/kg/ day) or due to 301 some toxicity of the molecule. Despite the very modest reduction in viral load, no obvious 302 change (improvement or worsening) of the rather subtle radiological and histological lung 303 pathology could be observed in favipiravir-treated hamsters (Fig. 2E-G) . Quantification of 304 micro-CT-derived biomarkers support these observations and quantify a relatively small 305 burden of radiological lung consolidation upon infection that does not change with favipiravir 306 treatment (Fig. 2F) . 307 HCQ sulphate was tested alone or in combination with azithromycin at a dose of 50 mg/kg/day 308 (equivalent to 39 mg/kg HCQ base) administered intraperitoneally once daily. When in 309 combination, azithromycin was given orally once daily at a dose of 10 mg/kg/day. Treatment 310 with HCQ alone resulted in a very modest reduction of 0.3 log10 viral RNA copies/mg lung, and 311 no reduction in viral RNA load in the ileum or stool compared to untreated infected hamsters 312 ( Fig 2B) . When combined with azithromycin, no additional reduction of viral RNA was observed 313 in the organs of infected animals (Fig 2B) . Virus titrations of the lungs also revealed no 314 significant reduction after treatment with HCQ alone or in combination with azithromycin (Fig  315   2C) . The weight loss of the animals treated with HCQ follows along the lines of the untreated 316 animals with < 5% weight loss during the whole experiment, while the combination treatment 317 with azithromycin caused a slightly greater weight loss at day 1 and 2 pi, from which the animals could partially recover (Fig 2D) . Similarly, no radiological improvement was observed 319 between non-treated animals and animals treated with HCQ or HCQ in combination with 320 azithromycin, which was confirmed by quantification of micro-CT-derived biomarkers of lung 321 pathology (Fig 2E-G) . Fig. 1 ). This indicates that 335 the fecal-oral route only marginally contributes to the transmission SARS-Cov-2 between 336 hamsters, thereby confirming the results of a previous study 18 . We therefore continued by 337 focusing on transmission of the virus via direct contact only. 338

Using the transmission model, we investigated the prophylactic potential of HCQ and favipiravir 339 against SARS-CoV-2. Sentinel hamsters received a daily dosage for 5 consecutive days with 340 either HCQ or favipiravir, starting 24 hours prior to exposure. Each individual sentinel hamster 341 was co-housed with an index hamster that had been intranasally inoculated with SARS-CoV-342 2 the day before (Fig 3A) . Index hamsters were sacrificed 4 days pi and sentinels 4 days post 343 exposure, after which the viral loads in lung, ileum and stool were determined. Index hamsters 344 had ~7 log10 viral RNA copies/mg in the lungs, whereas untreated sentinel hamsters had ~4 log10 viral RNA copies/mg in the lungs (Fig 3B) . Even though the variability between 346 individual hamsters in the sentinel groups was more pronounced than in the index groups, no 347 reduction in viral RNA was observed in either favipiravir-or HCQ-treated sentinel hamsters. 348

Also in ileum and stool, the viral RNA levels were not reduced by compound treatment. The 349 infectious viral loads in the lungs were also not reduced by treatment with either compound 350 (Fig 3C) , which is in line with the viral RNA data. In contrast to index hamsters, sentinel 351 hamsters did not lose weight, but gained around 8% of body weight by day 4 pi. Sentinels that 352 received HCQ or favipiravir treatment gained less body weight than the untreated sentinels 353 (5% and 2%, respectively) (Fig. 3D) . Pathology scores derived from micro-CT scans of 354 hamsters revealed multifocal pulmonary infiltrates and consolidations in some but not in all 355 hamsters (Fig. 3E, 3F) . Also, micro-CT-derived biomarkers showed no difference in lung 356 pathology between untreated and treated sentinel hamsters (Fig. 3G) , further confirming that 357 hydroxychloroquine and favipiravir failed to prevent SARS-CoV-2 infection in a transmission 358 model. 359

Based on the measurement of trough concentrations of HCQ at sacrifice (n=14), a mean ± SD 361 trough plasma concentration of 84 + 65 ng/mL (0.251 ± 0.19 µM) was found (Fig 4A) . This is 362 comparable to the plasma trough concentrations that were detected in cynomolgus macaques 363 (treated with a dosing regimen of 90 mg/kg on day 1 pi (loading dose) followed by a daily 364 maintenance dose of 45 mg/kg) 14 and in patients (3-5 days after starting treatment with 200 365 mg three times daily) 14 . The peak viral load in the lungs was not significantly associated with 366 plasma HCQ concentrations in individual hamsters (Fig 4B) , suggesting that a higher HCQ 367 exposure did not result in a more pronounced reduction of viral infection. 368

According to Equation 1, a whole blood concentration of 1.804 ± 1.39 µM was calculated (Fig  369   4C ). Subsequently, applying Equation 2, this resulted in a total lung concentration of 90.18 ± 370 69.42 µM, indicating that the lung tissues achieved HCQ concentrations above the reported in 371

vitro EC50 values, ranging from 0.72 to 17.31 µM, with a median value of 4.51 µM and an interquartile range of 5.44 (25-75%) 29 . To estimate 90% of inhibition of viral replication (EC90), 373 the EC90 was equated to 3 times the EC50, resulting in a target lung concentration of 13.53 ± 374 16.31 µM. In this case, the efficacy target at trough would be reached when applying this 375 dosing regimen (i.e., 50 mg HCQ sulphate/kg/day). However, it is important to note that the 376 total lung tissue concentrations described above consist of both intracellular and interstitial 377 HCQ concentrations. As the in vivo antiviral mechanism(s) of action of HCQ against SARS-378

CoV-2 has not been clarified yet and might not be exclusively by inhibition of endosome 379 acidification 30 , HCQ concentrations were calculated in cytosolic lung tissue, in the endosomal-380 lysosomal compartment of cells and in the interstitial compartment. Assuming that cytosolic 381 HCQ concentrations are only 6% of total tissue concentrations, a total cytosolic lung tissue 382 concentration of 5.41 ± 4.17 µM was calculated. This value was in line with the median in vitro 383 EC50 value, but is well below the estimated EC90 value. Also the interstitial concentration was 384 calculated to be 5.41 µM. In contrast, the endosomal/lysosomal HCQ concentration was 385 calculated to be 1.9 mM, which is much higher than the estimated EC90.

In a previous study, we showed that wild-type Syrian hamsters are highly susceptible to SARS-388

CoV-2 infections 16 . Here, we further characterized the hamster infection model to allow the use 389 of this model for antiviral drug evaluation. In agreement with previous studies, upon intranasal 390 inoculation, we observed that the virus replicates efficiently to peak levels (~6 log10 TCID50/mg) 391 in the lungs on day 4 pi., which is supported by radiological and pathological evidence. 392

Although the virus was also present in the ileum and stool of infected hamsters, levels were 393 significantly lower (~2.5 log10 copies/mg). Besides serving as efficient replication reservoirs of 394 SARS-CoV-2, the hamsters also efficiently transmit the virus to co-housed sentinels 15, 18 . Here, 395

we demonstrated that the virus is mainly transmitted via direct contact and only to a limited 396 extent via the fecal-oral route. The variability observed in the virus titers in the lungs of the 397 sentinels is probably due to differences in the infection stage of the animals. 398

Besides hamsters, a variety of other animals have been tested for their permissiveness to 399 SARS-CoV-2, of which ferrets and non-human primates were the most sensitive ones 31-35 . In 400 ferrets, infectious SARS-CoV-2 was only detected in the nasal turbinate and to a lesser extent 401 in the soft palate and tonsils, but not in the lungs 35 . Although, in a different study infectious 402 virus in the lungs of ferrets was detected, levels remained close to the limit of detection 33 . This 403

indicates that ferrets support SARS-CoV-2 replication, albeit to a lesser extent than hamsters. 404

In SARS-CoV-2-infected macaques (both rhesus and cynomolgus) virus levels were the 405 highest in nasal swabs and the lungs 32,34 . SARS-CoV-2 infection resulted in moderate transient 406 disease in rhesus macaques, whereas cynomolgus macaques remained asymptomatic, but 407 did develop lung pathology as seen in COVID-19 34 . Although aged macaque models may 408 represent the best models for studying more severe COVID-19 disease 36 , both the high costs 

A novel coronavirus from patients with pneumonia in 502 China

The trinity of COVID-19: immunity, 504 inflammation and intervention

Clinical characteristics of 82 death cases with COVID-506 19

Medical treatment options for COVID-19

Identification of antiviral drug candidates against SARS-510

CoV-2 from FDA-approved drugs

FDA approved drugs with 512 broad anti-coronaviral activity inhibit SARS-CoV-2 in vitro

In vitro inhibition of severe acute 515 respiratory syndrome coronavirus by chloroquine

Screening of an FDA-approved 518 compound library identifies four small-molecule inhibitors of Middle East respiratory 519 syndrome coronavirus replication in cell culture

Favipiravir as a potential countermeasure against 522 neglected and emerging RNA viruses

Remdesivir and chloroquine effectively inhibit the 524 recently emerged novel coronavirus (2019-nCoV) in vitro

Remdesivir, lopinavir, emetine, and 526 homoharringtonine inhibit SARS-CoV-2 replication in vitro

Favipiravir strikes the SARS-CoV-2 at its 529 Achilles heel, the RNA polymerase

Hydroxychloroquine or chloroquine with 531 or without a macrolide for treatment of COVID-19: a multinational registry analysis. 532 Lancet

Hydroxychloroquine in the treatment and 534 prophylaxis of SARS-CoV-2 infection in non-human primates

Simulation of the clinical and pathological 536 manifestations of Coronavirus Disease 2019 (COVID-19) in golden Syrian hamster 537 model: implications for disease pathogenesis and transmissibility

STAT2 signaling as double-edged 539 sword restricting viral dissemination but driving severe pneumonia in SARS-CoV-2 540 infected hamsters

First cases of coronavirus disease 2019 542 (COVID-19) in the WHO European Region

Pathogenesis and transmission of SARS-CoV-2 in 545 golden hamsters

A dose-ranging study of the pharmacokinetics of 547 hydroxy-chloroquine following intravenous administration to healthy volunteers

Stereoselective disposition of 550 hydroxychloroquine and its metabolites in rats

Dosing Design of Hydroxychloroquine for the Treatment of Severe Acute Respiratory 553 Syndrome Coronavirus 2 (SARS-CoV-2)

In vitro screening of a FDA approved chemical 555 library reveals potential inhibitors of SARS-CoV-2 replication

Hydroxychloroquine, a less toxic derivative of chloroquine, is 558 effective in inhibiting SARS-CoV-2 infection in vitro

A simple method of estimating fifty per cent endpoints

Iterative CT Reconstruction Using 562 Shearlet-Based Regularization

Radiosafe micro-computed tomography for 564 longitudinal evaluation of murine disease models

Longitudinal micro-CT provides 566 biomarkers of lung disease that can be used to assess the effect of therapy in 567 preclinical mouse models, and reveal compensatory changes in lung volume

Longitudinal, in vivo assessment of invasive 570 pulmonary aspergillosis in mice by computed tomography and magnetic resonance 571 imaging

Optimizing Hydroxychloroquine 573 Dosing for Patients With COVID-19: An Integrative Modeling Approach for Effective 574 Drug Repurposing

The possible mechanisms of action 576 of 4-aminoquinolines (chloroquine/hydroxychloroquine) against Sars-Cov-2 infection 577 (COVID-19): A role for iron homeostasis?

Infection and Rapid Transmission of SARS-CoV-2 in 579 Ferrets

Respiratory disease in rhesus 581 macaques inoculated with SARS-CoV-2

Antiviral Efficacies of FDA-Approved Drugs against 583 SARS-CoV-2 Infection in Ferrets

Comparative pathogenesis of COVID-19, MERS, 585 and SARS in a nonhuman primate model

Susceptibility of ferrets, cats, dogs, and other 587 domesticated animals to SARS-coronavirus 2

Animal models of mechanisms of SARS-589

CoV-2 infection and COVID-19 pathology

Structural Basis for Potent Neutralization of 591 Betacoronaviruses by Single-Domain Camelid Antibodies

A Randomized Trial of 593 Hydroxychloroquine as Postexposure Prophylaxis for Covid-19

Press release: No clinical benefit from use of 596 hydroxychloroquine in hospitalised patients with

Lung tissue distribution of drugs as a key factor for COVID-19 600 treatment

Favipiravir: Pharmacokinetics and Concerns About Clinical Trials for 602 2019-nCoV Infection

Favipiravir versus Arbidol for COVID-19: A 604 Randomized Clinical Trial

Pharmacokinetics of Favipiravir in Critically Ill 606 Patients with COVID-19

Dose rationale for favipiravir use in patients infected 608 with SARS-CoV-2

Favipiravir pharmacokinetics in Ebola-610 Infected patients of the JIKI trial reveals concentrations lower than targeted

A review of the safety of favipiravir -a potential 613 treatment in the COVID-19 pandemic?

demonstrated by a single dose with a single-domain antibody from a llama immunized with 415 prefusion-stabilized coronavirus spikes 16, 37 , thereby validating the use of hamsters to evaluate 416 treatment options against SARS-CoV-2. In addition, our data also indicate that hamsters are 417 highly amenable for studying the potential antiviral effect of small molecules on virus 418 transmissibility in a pre-and post-exposure setting. 419In an effort to contribute to the debate on the efficacy of (hydroxy)chloroquine and favipiravir 420 in COVID-19 patients, we evaluated both re-purposed drugs in our hamster infection and hamsters. In SARS-CoV-2 infected ferrets, HCQ treatment was also not able to significantly 426 reduce in vivo virus titers 33 . In addition, a recent study in SARS-CoV-2-infected cynomolgus 427 macaques showed that HCQ alone or in combination with azithromycin did not result in a 428 significant decrease in viral loads, both in a therapeutic and in a prophylactic setting 14 . On the 429 other hand, clinical trials with HCQ for the treatment of COVID-19 patients have resulted in 430 conflicting results and controversy. This is especially the case with clinical studies conducted 431 in the early stage of the pandemic, which were mostly small anecdotal studies. Results of large, 432 placebo-controlled, randomized trials are now becoming available. A randomized trial of HCQ 433 as post-exposure prophylaxis after high-to-moderate-risk exposure to COVID-19 showed that 434 high doses of HCQ did not prevent SARS-CoV-2 infection or disease similar to COVID-19 38 . 435In the RECOVERY trial, a large UK-based clinical study to evaluate potential therapies, HCQ 436 treatment did not result in a beneficial effect in mortality or hospital stay duration in patients 437 hospitalized with COVID-19 39 . These data are in agreement with our results in the hamster 438 model and clearly underline the importance of preclinical studies in animal models in the drug 439 development/repurposing process. 440The lack of effect observed for HCQ in this study and potentially also in other studies may be 441 explained by a pharmacokinetic failure. High lung concentrations of HCQ are caused by 442 massive accumulation ('ion trapping') of the compound in acidic lysosomes, which is driven by 443 a pH gradient between cytosol (pH 7.2) and lysosomes (pH 5). However, taking into account 444 the pH partition theory and considering the relative volumes of lung cellular and interstitial 445 compartments, only 6% of total HCQ concentrations in lung tissue is present in the cytosol of 446 lung cells. The other 94% of HCQ is present in the interstitial compartment and intracellularly 447 in lysosomes/endosomes or other subcellular fractions, or bound to proteins. Starting from the 448 measured trough concentrations from treated hamsters at day 4 or 5, the calculated HCQ 449 concentration in the endosomal compartment was 1.9 mM, which would be well above the 450 EC90 target. In contrast, cytosolic concentrations in the lung were only slightly higher than the 451 EC50 values reported in the literature, and far below the EC90 target. Although alkalization of 452 endosomes has been proposed as one of the key mechanisms of the broad-spectrum antiviral 453 effect of HCQ, the mechanism of action against SARS-CoV-2 has not been completely 454 unraveled 30 . Therefore, the very low cytosolic concentrations of HCQ in the lung may explain 455 the absence of an antiviral effect of HCQ against SARS-CoV-2 in vivo. Increasing the HCQ 456 dose to reach the EC90 might not be feasible in terms of safety, as it may lead to an increased 457 risk of QTc prolongation and fatal arrhythmia. In future studies, lung tissue distribution of (re-458 purposed) antiviral drugs should be taken into account, along with specification of the 459 subcellular target site, as recommended by Wang and Chen 40 . 460In contrast to HCQ, favipiravir was able to inhibit virus replication in intranasally infected already showed that in COVID-19 patients with mild symptoms (fever and respiratory 468 symptoms without difficulties in breathing) the clinical recovery rate at day 7 was higher in the 469 favipiravir-treated group compared to the control group, which received treatment with 470 arbidol 42 . However, for COVID-19 patients with hypertension and/or diabetes as well as 471 critically ill patients, the clinical recovery rate was not significantly different between groups, 472suggesting that favipiravir might be useful for patients with mild symptoms, but not for severely 473 ill patients. One concern with favipiravir is that it has been reported that the trough (FWO -12R2119N) . 630

The authors declare no competing interests. 632