Cytotoxicity evaluation of chloroquine and hydroxychloroquine in multiple cell lines and tissues by dynamic imaging system and PBPK model

Chloroquine (CQ) and hydroxychloroquine (HCQ) have been used in treating COVID-19 patients recently. However, both drugs have some contradictions and rare but severe side effects, such as hypoglycemia, retina and cardiac toxicity. To further uncover the toxicity profile of CQ and HCQ in different tissues, we evaluated the cytotoxicity of them in 8 cell lines, and further adopted the physiologically-based pharmacokinetic models (PBPK) to predict the tissue risk respectively. Retina, myocardium, lung, liver, kidney, vascular endothelium and intestinal epithelium originated cells were included in the toxicity evaluation of CQ and HCQ respectively. The proliferation pattern was monitored in 0-72 hours by IncuCyte S3, which could perform long-term continuous image and video of cells upon CQ or HCQ treatment. CC50 and the ratio of tissue trough concentrations to CC50 (RTTCC) were brought into predicted toxicity profiles. The CC50 at 24 h, 48 h, 72 h of CQ and HCQ decreased in the time-dependent manner, which indicates the accumulative cytotoxic effect. HCQ was found to be less toxic in 7 cell types except cardiomyocytes H9C2 cells (CC50-48 h=29.55 μM; CC50-72 h=15.26 μM). In addition, RTTCC is significant higher in CQ treatment group compared to HCQ group, which indicates that relative safety of HCQ. Both CQ and HCQ have certain cytotoxicity in time dependent manner which indicates the necessity of short period administration clinically. HCQ has the less impact in 7 cell lines proliferation and less toxicity compared to CQ in heart, liver, kidney and lung.

The Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), was 84 first emerged in China and has spread globally due to its high transmissibility and 85 infectivity, resulting in an unprecedented global public health challenge (1, 2). As of 86 April 20, 2020, more than 2,400,000 cases have been confirmed around the world, 87 according to data supplied by Johns Hopkins University, and at least 58,000 people 88 have died from the disease (2). Judging from current status, most patients have a good 89 prognosis, nevertheless approximately 20% of the patients with COVID-19 90 experienced critical complications, including arrhythmia, acute kidney injury, 91 pulmonary edema, septic shock, and acute respiratory distress syndrome (ARDS)

92
(3-6). Apart from primarily inflammation in the lungs, it is also suggested that other 93 vital organs like kidneys, heart, gut, as well as liver, were also suffered severe damage 94 according to the autopsies, suggesting that individuals or older with chronic 95 underlying diseases appear to have a higher risk for developing severe outcomes. 96 Such huge numbers of infected people call for an urgent demand of effective and 97 available drugs to manage the pandemic. Unfortunately, at present, there are still no 98 specific antiviral drugs for prevention or treatment of COVID-19 patients. Recent 99 publications have demonstrated that chloroquine (CQ) and hydroxychloroquine (HCQ) 100 efficiently inhibited SARS-CoV-2 infection in vitro assay (7-9). CQ, together with its 101 derivate HCQ, has been commercialized as antimalarial drugs in the clinic for several 102 decades. HCQ has also been broadly used in autoimmune diseases treatment, such as 103 systemic lupus erythematosus (SLE) and rheumatoid arthtitis (10-13). Several clinical 104 trials have confirmed that both CQ and HCQ were superior to the control group in 105 inhibiting the exacerbation of pneumonia, improving lung imaging findings, as well 106 as promoting the virus negative conversion and shorten the disease course. 11 107 Moreover, the U.S. Food and Drug Administration (FDA) also approved CQ and 108 HCQ for emergency use to treat hospitalized patients for COVID-19. Although 109 exhibiting apparent efficacy and acceptable safety profile for COVID-19 treatment, 110 CQ and HCQ still have some potential concerns with prolonged usage, including 111 heart rhythm disturbances, gastrointestinal upset, retinal toxicity, in particular for retinopathy (11,(14)(15)(16)(17). Additionally, Risambaf et al. found that CQ/HCQ may 113 increase the risk of liver and renal impairment when it used to treat .

114
Toxicity tolerability in key tissues about drug effectiveness and side effect were 115 critical to understand their mechanism and to optimize dosing regimen by integrating calculated at the given target organ, respectively. The data suggest that HCQ was 127 demonstrated to be much less toxic than CQ, at least at certain key tissues (heart, liver, 128 kidney, and lung). Taken together, this study provides the information regarding 129 cytotoxicity in a wide spectrum and will be beneficial for both pharmacologists and

133
The effect of CQ on cell proliferation 134 To gain the more comprehensive cytotoxic information upon CQ and HCQ treatment,   Both CQ and HCQ show strong and immediate toxicity on all 8 cell lines upon 168 treatment more than 300 μM of CQ or HCQ. As shown in Figure 1 and 2, when the 169 concentration of CQ or HCQ is higher than 300 μM, the proliferation shows a sudden 170 decline or brake compared with lower dosing regimens. H9C2 (heart) 、 HEK293( kidney), and IEC-6 (intestine), are the more sensitive cells to CQ compared 172 with 5 other cell lines, as their CC50 value at 72 h are less than 20 μM (17. with that of 48h in Vero, which may be due to special drug metabolism or stability in 183 it. As the selection index (SI) is the safe range to evaluation the drug effect.  Table 1) (9). Therefore, we can preliminarily conclude that the selectivity 188 index (SI) of HCQ is higher than that of CQ in most cell types. Using our PBPK models, we simulated the tissues concentrations of HCQ (600 mg 192 BID for 1 day, 200 mg BID for day 2 to 5) and CQ (500 mg BID for 7 days) (19, 20).

193
The Cmax of tissue concentrations were summarized in Table 2. Results of simulated 194 tissue concentration showed that tissue trough concentration of CQ in liver and lung 195 reached the highest level of drug accumulation (227.545 μg/ml), which is 3 times 196 more than that in heart (60.598 μg /ml). However, the tissue trough concentration of 197 HCQ in lung is the highest level (25.633 μg/ml) compared with liver, kidney and heart 198 (Table 2 and Figure 4).

199
In order to better predict the toxicity risk of CQ and HCQ in different tissues, we used 200 the ratio of simulated tissue trough concentration to CC50 (R TTCC ) to predict the risk 201 of tissue toxicity for the safety profile of these two drugs in the given tissues. As 202 shown in Figure 3, we systematically compared the toxicity between CQ and HCQ, 203 the R TTCC value of CQ is 6-87 times more than that of HCQ in lung, heart, kidney and 204 liver, which suggests that the toxicity risk of HCQ in the above tissues is much lower 205 than that of CQ. were obtained as previously reported. The lung to blood concentration ratio for CQ 233 and HCQ (obtained from animal studies) was used to predict the drug concentration in 234 the lungs, heart, liver, and kidney. To better investigate the potential toxicity in vivo 235 and in vitro, we proposed R TTCC (ratio of tissue concentration and CC 50 ) derived from 236 PBPK model to predict the risk of toxic profiles in different tissues. We compared the 237 R TTCC data collected from heart, liver, kidney, lung, and revealed HCQ has shown 238 significantly safe profiles than that of upon CQ treatment (9). However, recent 239 publication reported that CQ was safer than HCQ according to SI (7,9). We speculate 240 that the safety difference might be due to their complex pharmacokinetic 241 characteristics in vivo, which possessed specific distribution and long half-life of 242 around days. In short, based on our just published study, we further developed the 243 novel parameters to predict the potential toxicity besides the traditional selectivity 244 index (SI), (the ratio of the CC 50 to EC 50 ), which is a commonly accepted to measure 245 the window between cytotoxicity and antiviral capacity (9). As a result, our data 246 shows that kidney, lung and heart are prone to the toxicity of CQ, otherwise lung and 247 kidney are relative vulnerable upon HCQ treatment ( Figure 5). In the meantime, 248 considering the un-negligible effect on cardiocytes and retina cells, of which the most 249 patients with the severe symptoms are more likely suffered the dysfunction in heart 250 and eye sight with aging simultaneously. Therefore, ECG monitoring should be 251 necessary during clinical usage, even for the patients only infected with COVID-19 252 but without the underlying diseases. In addition, the more attention should be paid to 253 the patients in the changes of their eye sight when using HCQ.

254
In this study, we perform dynamic imaging system to accurately and precisely 255 monitor the whole proliferation process other than conventional CCK8 assay.