key: cord-0851687-gpofwyux
authors: Savarino, Andrea; Tarek, Mohammad
title: Pharmacokinetic bases of the hydroxychloroquine response in COVID-19: implications for therapy and prevention
date: 2020-04-28
journal: nan
DOI: 10.1101/2020.04.23.20076471
sha: 8883f2335e3b638bd3cc9d0b8a5376cbad54d95c
doc_id: 851687
cord_uid: gpofwyux

Chloroquine/hydroxychloroquine has recently been the subject of intense debate in regard to its potential antiviral activity against SARS-Cov-2, the etiological agent of COVID-19. Some report possible curative effects, others do not. In order to shed some light on this rather controversial topic, we used mathematical modelling to simulate possible scenarios of response to hydroxychloroquine in COVID-19 patients. Our computer-aided simulations show that hydroxychloroquine may have an impact on the amplitude of the viral load peak but that viral clearance is not significantly accelerated if the drug is not administered early enough (i.e. when viral loads range from 1 to 1,000 copies/mL). Although some authors had used the trough plasma concentrations or the theoretical drug distribution in the lung to model the effect of chloroquine/hydroxychloroquine on COVID-19, the theoretical drug response based on the trough whole blood concentrations of the drug agreed well with the results of the clinical trials so far reported. Moreover, the effects of chloroquine/hydroxychloroquine could be fully explained when taking into account also the capacity of this drug to raise cell-mediated responses against the productively SARS-Cov-2-infected cells. On the whole, the present study suggests that chloroquine/hydroxychloroquine has a narrow therapeutic window, which overlaps with the highest tolerated doses. These considerations may have implications for development of anti-COVID-19 combination therapies and prevention strategies.

Chloroquine, and its hydroxy-analog hydroxychloroquine, was first proposed in 2003 for treatment of the severe acute respiratory syndrome (SARS) [1] and, due to its promising in-vitro effects [2] , was then repurposed as a drug for treatment of COVID-19. As clinical trials in COVID-19 have so far produced mixed results, further study on the interaction of this drug with the virus/host dynamics in vivo is necessary in order to provide insight into the optimal timing of administration and association with other interventions. However, after many decades of clinical use of hydroxychloroquine, part of its pharmacokinetics is still unknown, and a number of studies still rely on the drug plasma concentrations [3] , which are not an optimal indicator of the biodistribution of this drug, which is a weak base that becomes readily entrapped in intracellular acidic organelles, due to the Henderson-Hasselbach law [1] .

A recent advancement likely to shed more light in the interplay of SARS-CoV-2 with antiviral drugs has been the availability of a mathematical model able to simulate the virus/host interplay and matching clinical observations [4] . Similar models developed by some of the authors of that study had proven able to simulate the viral dynamics in HIV/AIDS, providing important therapeutic hints [5, 6] . In the present paper, we use this model to simulate possible scenarios of response to hydroxychloroquine in COVID-19 patients.

In the present paper we used the system of differential equations built by Gonçalves et al. [4] , as follows.

These are ordinary differential equations describing the viral dynamics in a model of target cell population with an eclipse phase accounting for antiviral effectiveness of chloroquine against viral reproduction and target cell populations.

T represents the target cells assumed to be 4×10 8 [7] based on an assumption of 30 mL nasopharyngeal volume [8] that maintain a cellular concentration of 1.33 × 10 7 cells/mL [4] which become infected with a constant parameter β which changes according to a constant parameter (1-ε) where ε reflects the antiviral effectiveness of chloroquine simulated at {0%, 33% ,70%,90%} based on an assumption of reduction in the viral basic reproductive number R 0 by the antiviral [4] , I1 represents the infected cells in the eclipse phase, I2 represents the productively infected cells which starts the productivity phase after average time 1/k where k values considered were {1,3,5} d -1 to allow prediction of intervention timing effect on viral propagation where releasing virions is controlled by constant parameter p which changes according to a constant parameter (1-ε) and assumed by default to be 10 d -1 , infected cells are later cleared with a constant parameter rate c which was assumed to be 5 d -1 following previous models of viral infection [9] , while Infected cells I2 are cleared later from circulation at a constant rate δ assumed by default to be 0.55 d -1 . V reflects the viral population that was modelled by changing V 0 at {1 1 , 10 3 , 10 5 } copies/mL ( Mathematical modelling was implemented using Python programming language v 3.7 (http://www.python.org), ODEINT function of SciPy package was used for solving ordinary differential equations [10] , numerical solutions were applied through NumPy package [11] and simulations were plotted within a 10 days' time frame using Matplotlib package [12] . 

CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 28, 2020. 1, р = 1 advanced stages of COVID-19. The simulation of δ = 1.56 with a starting viral load of 10 5 copies/mL is not provided as considered to be unrealistic.

Using the system of differential equations developed by Gonçalves et al. [4] , we showed that hydroxychloroquine may have an impact on the amplitude of the viral load peak. The effect on viral clearance is only evident in our simulations of mild viral loads, being less pronounced when viral loads are 10 5 viral RNA copies/mL, i.e. a viral load commonly found in throat swabs during symptomatic disease [13] .

Using the parameters adopted by the authors, the model shows that there is an early peak in viral replication followed by a gradual and slow abatement of viral load accompanying consumption of target cells (Figure 1 and Supplemental figures S1 and S2, first row). In their treatment simulations, Gonçalves et al. [4] concluded that hydroxychloroquine had no sufficient effect on the viral load dynamics because the pharmacokinetic parameter that they used to conduct the simulations was the plasma drug concentration, which, as mentioned above, is unlikely to reflect the biodistribution of this drug. In light of the plasma concentrations, the in-vivo drug efficacy that they calculated was 33%. The second row of Figure 1 and Supplemental figures S1 and S2 shows our simulations under those conditions. When we used the whole blood concentration (including the drug compartment entrapped in lysosomes) [14] , we calculated a drug efficacy of 0.7, based on the concentration-response curves published by Yao et al. [15] . In this case, a more pronounced decrease in the viral peak was observable ( Figure 1 and Supplemental figures S1 and S2, third row).

Another aspect that needs attention is the double mechanism of the antiviral effects of hydroxychloroquine. As derived from studies on a huge number of viruses, lysosomal entrapment of hydroxychloroquine is highly likely to affect both viral entry (simulated by Gonçalves et al. [4] , and post-entry stages in cellular organelles such as viral budding and viral particle glycosylation [1, 2] (not considered in the aforementioned study). According to the data of Wang et al. [2] , chloroquine displayed approximately half of its antiviral activity when it was added after virus adsorption onto cells, confirming that part of its activity occurs at a post-entry level. Therefore, as we also wanted to estimate the impact of hydroxychloroquine-induced impairment of viral production at times beyond the 48 h period considered to estimate hydroxychloroquine efficacy in vitro [15] , we decreased the viral production rate by 35%, i.e. half of the value calculated on the basis of the drug efficacy on the whole viral replicative cycle. We obtained a further decrease of viral load at peak and steady-state (Figure 1, fourth row) .

The data that we used to estimate the response of SARS-CoV-2 to hydroxychloroquine were based on the trough blood concentrations provided by Morita et al. [14] , which were based on a drug regimen of 400 mg/day. Gautret et al. [3] recently published the trough serum concentrations of hydroxychloroquine obtained from patients with COVID-19 and treated with a higher dosage (i.e. 600 mg/day) of hydroxychloroquine. Unfortunately, no blood concentrations were reported by Gautret et al., but these could be proportionally derived from the ratios provided by Morita et al. [14] , by maintaining a similar plasma/blood concentration ratio. This calculation led to a prediction of drug levels in whole blood of 6.16 μM. By plotting this value to the dose-response curves published by Yao et al. [15] , we extrapolated an efficacy of 90%: we adjusted the viral infection and production accordingly and obtained a further decrease in the viral load peak (Figure 1 and Supplemental figures S1 and S2, fifth row). This simulation may also account for hydroxychloroquine administered at 400 mg/day in combination with another antiviral agent with moderate potency.

Among the immune-modulating effects of hydroxychloroquine is its capacity to induce antigen cross-presentation and increase the immune cytotoxic response [16, 17] . Enhanced cytotoxic responses result in increased death of the infected cells. By increasing the death rate of productively infected cells (an effect of cytotoxic immunity) by only 3 fold (data derived from Accapezzato et al. [16] ), we could observe a further decrease in the peak amplitude and a shorter time to viral load exhaustion (Figure 1 and Supplemental figures S1 and S2, sixth row).

The simulations that we provide confirm current hypotheses that the virologic response to hydroxychloroquine in COVID-19 patients has a pharmacokinetic basis, and that the drug dosages with an acceptable toxicity profile have a narrow window of overlap with the antiviral effective concentrations. These results are in line with previous analyses of HIV clinical trials, showing that the dosages adopted safely in the clinic are in the lowest range of the therapeutic window, with significant, though yet partial, effects observable only at the highest doses administered [18] .

These observations are in line with the clinical data reported so far. A small trial in France suggested that hydroxychloroquine helps treat COVID-19 [3] and a small study from China [19] reported opposite results. Both studies have their own limitations and very limited statistical power, due to the fact that they were prepared during an emergency. The discrepancy of the results obtained may reside, according to the results of the present study, in dosage of hydroxychloroquine adopted, with the French study reporting positive results using 600 mg/day of hydroxychloroquine and the Chinese study being unable to show any efficacy following adoption of a lower dosage (400 mg/day). Furthermore, the reported response to hydroxychloroquine in the French study is incomplete, and the authors had to add another experimental drug repositioned as an antiviral (azithromycin) in order to increase efficacy. These considerations are in line with the results obtained in randomized clinical trials using chloroquine and hydroxychloroquine at an equivalent dosage superior to that reported in the two aforementioned studies [20] [21] [22] . As the safety margin of chloroquine/hydroxychloroquine is narrow [23] , administration of higher dosages is unfortunately hampered by toxicity, as shown by Borba et al. [22] , although recent data show that the toxicity observed by Borba et al. could in fact be ascribed by the concomitant administration of azithromycin [24] . Our results agree in part with the simulations conducted by Arnold and Buckner [25] , showing that only the 600 mg/day regimen might have a significant impact on viral replication according to all parameters that they adopted, and those of Garcia-Cremades et al. [26] , showing that a dosage higher than or equal to 600 mg/day is necessary to obtain a sustained remission. The approach that we have followed is more conservative than that adopted by Arnold and Buckner [25] , in that we used the whole blood concentrations as a measure of prediction of tissue accumulation rather than the predicted accumulation in the lung. According to the model adopted by Arnold and Buckner [25] , the drug lung concentrations are hundreds of times higher than those in plasma and are predicted to reach a concentration able to inhibit 100 percent of SARS-CoV-2 replication. We decided not to follow this prediction because 1) tissue accumulation of hydroxychloroquine in lung is mainly driven by tissue macrophages, which are not a main target for the virus, 2) according to the data published by Yao et al. [15] , 100 per cent of inhibition of coronavirus replication is not reachable experimentally using hydroxychloroquine. Also Smit et al. reached similar conclusions, based on pharmacokinetic considerations [23] .

Based on our simulations, we suggest that chloroquine may not be a "silver bullet" eradicating COVID-19 infection but, on the other side, it might display some beneficial effects in vivo . Our simulations go beyond the other models published, indicating that earlier virologic negativization should not be expected when hydroxychloroquine is administered at dosages of 400 mg/day but only at dosages of 600 mg/day, i.e.within the narrow therapeutic window. Instead, we believe that the theoretical capacity of hydroxychloroquine to decrease the viral load peak when viral replication is most active deserves attention and further investigation. The results of our simulations are in line with those of a recent meta analysis of the so far available clinical data, showing that hydroxychloroquine may decrease disease severity but not viral clearance when administered to symptomatic COVID-19 patients [27] .

One limitation of the present study resides in the fact that we only modeled the viral dynamics but not the effects of the molecule on immune mediated mechanisms such as the cytokine storm [1] . Be that as it may, an impact on the immune hyperactivation can in this case be inferred from the estimated potential of the drug to decrease the viral load peak, the amplitude of which is associated with deleterious immune hyperactivation in coronavirus infections [1] . This hypothesis is in line with the ability of SARS-CoV-2 to infect CD4+ T-cells [28] , the target cells of HIV, which is characterized by immune hyperactivation, as well [18] .

Finally, we show that the effect of chloroquine/hydroxychloroquine on antigen cross-presentation should be considered. Cross-presentation is a phenomenon in which a dendritic cell presents the antigen to CD8+ T-cells, improving their priming. Activation of CD8+ T-cells after priming through antigen recognition can induce the selective killing of the infected cells [16] . Taking into account this parameter it is possible to predict that hydroxychloroquine may accelerate viral clearance but only when administered early. Accordingly, we did not model a scenario of increased cell-mediated responses when viral loads are higher, because immune dysregulation is in this case likely to have taken place. Enhanced cell-mediated responses may not only induce selective killing of the infected cells but also enhance immunity for better long-term control of the infection.

Hydroxychloroquine is currently being investigated also as a prophylactic agent [29] . Given the great interest that this strategy has so far evoked in the context of COVID-19 [30, 31] , future vaccines against this disease should be considered in co-administration with chloroquine/hydroxychloroquine. In case chloroquine/hydroxychloroquine should prove ineffective in protecting from SARS-CoV-2 acquisition, it is plausible that it may mitigate disease severity by attenuating viral replication.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 28, 2020. . 

Effects of chloroquine on viral infections: an old drug against today's diseases?

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

Hydroxychloroquine and azithromycin as a treatment of COVID-19: results of an open-label non-randomized clinical trial

Timing of antiviral treatment initiation is critical to reduce SARS-Cov-2 viral load. medRxiv

Modeling latently infected cell activation: viral and latent reservoir persistence, and viral blips in HIV-infected patients on potent therapy

A cure for AIDS: a matter of timing?

Kinetics of influenza A virus infection in humans

Combination Antiviral Therapy for Influenza: Predictions From Modeling of Human Infections

Zika plasma viral dynamics in nonhuman primates provides insights into early infection and antiviral strategies

SciPy 1.0: fundamental algorithms for scientific computing in Python

The NumPy array: a structure for efficient numerical computation

Matplotlib: A 2D graphics environment

SARS-CoV-2 viral load in upper respiratory specimens of infected patients

Population pharmacokinetics of hydroxychloroquine in Japanese patients with cutaneous or systemic lupus erythematosus

In Vitro Antiviral Activity and Projection of Optimized Dosing Design of Hydroxychloroquine for the Treatment of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2)

Chloroquine enhances human CD8+ T cell responses against soluble antigens in vivo

Enhancement of T cell-mediated immune responses to whole inactivated influenza virus by chloroquine treatment in vivo

Chloroquine and beyond: exploring anti-rheumatic drugs to reduce immune hyperactivation in HIV/AIDS

A pilot study of hydroxychloroquine in treatment of patients with common coronavirus disease-19 (COVID-19)

Treating COVID-19 with Chloroquine

Breakthrough: Chloroquine phosphate has shown apparent efficacy in treatment of COVID-19 associated pneumonia in clinical studies

Chloroquine diphosphate in two different dosages as adjunctive therapy of hospitalized patients with severe respiratory syndrome in the context of coronavirus (SARS-CoV-2) infection: Preliminary safety results of a randomized, double-blinded, phase IIb clinical trial (CloroCovid-19 Study). medRxiv

Chloroquine for SARS-CoV-2: Implications of Its Unique Pharmacokinetic and Safety Properties

Safety signals for QT prolongation or Torsades de Pointes associated with azithromycin with or without chloroquine or hydroxychloroquine

Hydroxychloroquine for treatment of SARS-CoV-2 infection? Improving our confidence in a model-based approach to dose selection

Optimizing hydroxychloroquine dosing for patients with COVID-19: An integrative modeling approach for effective drug repurposing

Virological and Clinical Cure in Covid-19 Patients Treated with Hydroxychloroquine: A Systematic Review and Meta-Analysis

SARS-CoV-2 infects T lymphocytes through its spike protein-mediated membrane fusion

Can post-exposure prophylaxis for COVID-19 be considered as one of outbreak response strategies in long-term care hospitals?

Is Hydroxychloroquine a possible post-exposure prophylaxis drug to limit the transmission to health care workers exposed to COVID19?

Chloroquine or hydroxychloroquine for prophylaxis of COVID-19