Dimethylarsenic acid (DMA) accumulation positively correlates with realgar-induced subchronic toxicity in rats

The toxicity of realgar depends largely on different arsenic species accumulation and distribution in the body. Here, after continuous oral administration of different doses of realgar for 90 days and subsequent 60-day withdrawal period, clinical observations, food consumption, body weights, blood biochemistry, hematology, and histomorphological examination of rats were performed. Realgar 40mg·kg−1·d−1 and 170 mg·kg−1·d−1 of realgar (which is equivalent to 40-fold and 100-fold the maximum clinical dose, respectively) can cause toxicity in rats, including degreased body weight, peripheral blood neutrality abnormal ratio of granulocytes and lymphocytes, hypercoagulability of the blood, liver and kidney tissue damage, liver and kidney may be the main toxic target organs of realgar. The no observed adverse effect level (NOAEL) dose is 10 mg·kg−1. At the same time, the content and distribution of arsenic species in tissues were determined. The content of total arsenic (tAs) and Dimethylarsenic acid (DMA) in the tissues of the realgar group was significantly higher than those of the control group. After 60 days of discontinuation, the DMA content in the realgar group decreased, but it was still higher than that in the control group, and liver and kidney damage occurred during the administration period basically returned to normal. Therefore, the authors speculated that when the DMA content in the tissue exceeds a certain range, liver and kidney toxicity will be induced. However, when the DMA content is lower than the above threshold after drug withdrawal, the liver and kidney lesions can return to normal.


34
Realgar is mineral medicine containing arsenic, the main component of which is As4S4 35 and its content is more than 90%. As a traditional Chinese medicine, realgar has a 36 widely usage in clinic for thousands of years. Currently, more than 30 realgar 37 preparations are still in use in China [1]. Realgar is effective in the treatment of various 38 diseases such as infectious diseases (malaria, syphilis, parasitic infections), 39 gastrointestinal diseases (ulcers), nervous system diseases (convulsions), rheumatic 40 diseases and skin diseases (psoriasis) [2][3][4]. Modern researches also show that realgar 41 has a significant anticancer effect in the clinic, especially on hematological cancer. 42 Since the treatment of leukemia in the 1970s, more and more studies have shown that 43 realgar has a potential therapeutic effect on various cancers. There is a growing 44 understanding of the mechanism of action of realgar as an anti-tumor agent [5]. For 45 example, realgar and realgar-containing formulations are effective and low toxicity in 46 inducing apoptosis or differentiation of acute promyelocytic leukemia (APL) cells. 47 However, realgar has been known as poisons since ancient times because arsenic is a 48 well-known toxic substance. It was found that arsenic can mediate oxidative stress, 49 induce hepatic apoptosis, cause chromosomal abnormality, alter growth factors 50 expression, inhibit DNA repairing, increase the secretion of inflammatory factors and 51 up regulate apoptosis related proteins, so that exposure to arsenic may induce liver 52 disease, cardiovascular disease, nephrotoxicity, neurotoxicity and carcinogenicity [6-53 8]. 54 The poor solubility of realgar is a barrier to its bioavailability, and long-term high-dose 55 administration of realgar is necessary for clinical complete remission of cancer, but its 56 potential toxicity limits its use. The long-term accumulation of toxic arsenic compounds 57 in the body can cause abnormal tissue structure and function of the body [9-11]. Due to 58 the known toxicity characteristics of arsenic [12][13][14], the long-term toxicity of realgar 59 has always been a concern. However, it is unscientific to treat the toxicity of realgar 60 and arsenic equally. Lu et al [15] compared the long-term renal toxicity of realgar, 61 sodium arsenite (1/100 of the arsenic level) and sodium arsenate (1/50 of the arsenic 62 level) in mouse. The results showed that the renal pathological lesions caused by 63 sodium arsenite and sodium arsenite were more severe than that caused by realgar. 64 Sodium arsenate and sodium arsenite induced the expression of heme oxygenase-65 related gene heme oxygenase-1, causing kidney damage, and realgar has no effects on 66 this. Thus, it suggests that the toxicity of arsenic is closely related to its chemical form.

67
There are studies that have demonstrated that if long-term ingestion of realgar, the total 68 arsenic content in rat tissues will increase significantly and will cause damage to the 69 liver and kidney of rats [16][17]. However, there are still few studies on its metabolic 70 characteristics and exact toxic substances and toxic effects. 71 It is known that the metabolism of realgar is different from that of other substances 72 containing arsenic, and the accumulation of As species may also be different. After 73 administration of realgar, the arsenic metabolism is converted to various arsenic 74 compounds. The toxicity of arsenic depends largely on its chemical structure, oxidation 75 state and metabolic pathways. Various arsenic metabolites exhibit differing toxicities. 76 For example, IC50 values of the main arsenic species is: AsⅢ < AsⅤ< MMA≤DMA < 77 AsC< AsB [18][19][20]. Thus, the quantity of each of these forms accumulated in tissues 78 could be an important biomarker for evaluating arsenic toxicity. A large number of 79 experiments have been conducted to study the changes in the total arsenic content of 80 realgars entering the body, but there are few studies on their metabolism and toxicity 81 profiles in the body. The association between the accumulation and distribution of 82 arsenic species and the long-term toxicity of realgar is still not fully explored. The 83 speciation information will contribute to improve our understanding of realgar 84 metabolic transformation and toxicity mechanism. In this study, high performance 85 liquid chromatography-inductively coupled plasma mass spectrometry (HPLC-ICP-86 MS) [21] was used to analyze arsenic species and the relationship between arsenic 87 metabolism and realgar toxicity after long-term oral administration of realgar was 88 evaluated. With the results we can figure out an underlying toxic substance basis and 89 the main toxicity target organ of realgar, so that to provide scientific basis for further 90 study the mechanism of the toxicity of realgar, clinically rational application of male-91 containing drugs, and their in-depth development. 92 realgar is 1.7 mg·kg -1 /d. In this study, the four doses set were 0 mg·kg -1 ·d -1 , 10 mg·kg -143 1 ·d -1 , 40 mg·kg -1 ·d -1 and 170 mg·kg -1 ·d -1 were set, which is equivalent 6-fold, 24-fold 144

Reagents and Instruments
and 100-fold of 1.7 mg·kg -1 , respectively. Thus, the rats were randomly delivered as 4 145 groups, 18 male rats and 18 female rats in each group. Rats in the 0 mg·kg -1 ·d -1 group 146 were orally administrated with 0.3% Carboxymethyl Cellulose Sodium (CMC-Na) 147 solution each day, and realgar was dissolved with 0.3% CMC-Na ultrapure water. Rats 148 administration volume is 5 ml/kg, and the other three realgar groups' rats were given 2 149 mg·ml -1 , 8 mg·ml -1 , 34 mg·ml -1 realgar-0.3% CMC-Na solution respectively. After 60 150 and 90 days of administration, each group of five male rats and female rats were 151 sacrificed. Finally, each of the remaining groups of rats ceased drug treatment and were 152 fed for 60 days and then were sacrificed on the 30th and 60th days of discontinuation, 153 respectively. During the experiment, the general condition of the rats was recorded daily, 154 including gait, general activity, hair, stool, urine, and so on. Body weight of all rats and 155 food consumption per cage were measured weekly. 156 Before each sacrifice, rats were individually placed into metabolic cages to collect 24 157 h-urine. Urine was qualitatively analyzed and indicators included urine glucose (GLU), 158 ketone body (KET), bilirubin (BIL), pH, specific gravity (SG), occult blood (BLO), 159 protein (PRO), urobilinogen (URO), leukocytes (Leu). Thereafter, rats were 160 anesthetized with phenobarbital sodium 30 mg·kg -1 i.p, then the blood of each rat was 161 taken from abdominal aorta to detect the hematology (including Red blood cell count

Rat biological sample collection and pretreatment
192 At each sacrifice, rats' blood, urine, feces, liver, kidney and brain of rats were obtained 193 to test the arsenic species therein. Determination of arsenic species in blood is usually 194 done with 2 ml of 10% EDTA-K2 anticoagulant. 5 ml of urine and all feces in the 195 metabolic cage were required to detect the exact arsenic levels in urine and feces. At 196 least 1g of organs (liver, kidney and brain) should be kept to measure the level of arsenic 197 in it. All samples were kept at -80℃ before the detection of arsenic species. Before 198 testing, the liver, kidney and brain were homogenized and the excrement was dried and 199 crushed. We respectively weighed the above samples, urine and plasma 0.5g each, then 200 put them into the microwave digestion tank, add 4mL nitric acid and 1mL hydrochloric 201 acid for digestion. Then the samples were taken out for cooling, transferred to a 50mL 202 polytetrafluoroethylene volumetric flask while the digestion tank is washed several 203 times with deionized water and the wash liquid were poured into the measuring bottle.

204
After that, we added deionized water until the liquid level reached 50ml mark and mix 205 the liquid evenly to get the final samples. 206 hydrochloric acid (HCL) and then added to an ion exchange resin chromatography 211 column, which are then eluted with to give an eluate. Since only arsenic species such 212

Analysis of arsenic species accumulation and distribution in
as MMA and DMA can adsorb to the column, AsⅢ and AsⅤ can be separated from 213 total arsenic through the column. The column is then separated by adding ultrapure 214 water and MMA and DMA are obtained by collecting the dissociation liquid. Next, 215 dilute HCL eluate containg AsⅢ and AsⅤ is passed through an anion strong basic 216 exchange column and then subsequently AsⅤ is absorbed into the column, thus 217 obtaining AsⅢ by collecting the eluate. In order to get AsⅤ, we add HCl to the anion 218 exchange column to dissolve AsⅤ to obtained AsⅤ HCl solution. Finally, all AsⅢ、 219 AsⅤ、MMA and DMA samples were detected by ICP-MS except that the fecal samples 220 were determinate by inductively coupled plasma atomic emission spectrometry (ICP- hematology and blood biochemistry values of male and female rats were analyzed 232 separately. The arsenic content data of male and female rats were combined and 233 analyzed. These data were performed using parametric one-way analysis of variance 234 (ANOVA) followed by LSD (equal variances assumed) or Tamhane's T2 (equal 235 variances not assumed) multiple comparison test. The difference between each realgar-236 treated group and the control group were compared. Statistical differences in 237 quantitative data were obtained by analysis of variance (ANOVA), qualitative urinary 238 data and histopathological changes were statistically analyzed with Kruskal-Wallis H-239 test and Mann-Whitney U-test. Statistical analysis using SPSS 17.0 software. A p-value 240 of <0.05 was considered to be statistically significant. 241

Oral realgar caused a degrease in body weight in rats 243
After 90 days of administration, except the rats in 170 mg·kg -1 ·d -1 realgar group had 244 depilated, while other animals were generally in good condition. In addition, from the 245 60th day to the 90th day, compared with the 0 mg·kg -1 ·d -1 control group, the body 246 weight of female rats in the 40 mg·kg -1 ·d -1 realgar group was significantly decreased (p 247 < 0.05). After discontinuation of treatment, the body weight of male rats in the realgar 248 group showed a downward trend compared with the control, but there was no statistical 249 difference (Fig 1). There was no difference in the average food consumption between 250 the four groups during treatment and after discontinuation (data not shown). Since the 251 weight-reduced rats returned to normal after drug withdrawal, we believe that this 252 adverse reaction may be mainly due to the larger volume and the higher concentration 253 of the drug, which affects the rats' feeding and digestive functions, rather than the toxic 254 effects of realgar.  After treatment with realgar for 90 days, compared with the control rats, in male and 261 female rats of the realgar medium-dose group and high-dose group, lymphocyte ratio 262 (Ly%) showed an upward trend, and neutrophilic granulocyte ratio (Gr%) showed a 263 decreasing trend (p<0.05, p<0.01) ( Table 1). We retrospectively reviewed the results of 264 the 60-day study, with no apparent abnormalities in Ly% and Gr% (data omitted). After 265 discontinuation of treatment, the values of Ly% and Gr% in the high-dose and medium-266 dose group of the realgar gradually returned to normal, and there was no difference 267 from the control group 60 days after drug withdrawal (Table 2). Other than that, there 268 were no obvious abnormal changes in other hematological parameters during realgar 269 administration, and after 60 days of discontinuation of the drug, there were no delayed 270 toxicities in the realgar hematological parameters. Therefore, long-term (90-day) use of 271 large doses of realgar (100 times the maximum clinical dose) may have an effect on the 272 immune function of rats. However, the specific impact on immune function needs to be 273 further studied. After the realgar is no longer used, peripheral blood neutrophils and 274 lymphocytes ratios of the rats returned to normal. 275

291
Realgar may cause hypercoagulability in rats 292 During the treatment period, compared with the rats in 0 mg·kg -1 ·d -1 control group, PT 293 and APTT of animals (♂, ♀) in low, middle and high dose realgar groups significantly 294 decreased (p <0.01), so long-term use of realgar can cause blood hypercoagulation. The 295 hypercoagulable state of blood has the risk of thrombosis, which in turn causes damage 296 to various systems of the body, in particular blood vessels and the heart. 60 days after 297 drug withdrawal, the APTT and PT values (♂, ♀) of the realgar group gradually 298 returned to normal ( only initially suspect that realgar will cause clinical hypercoagulability. Since the DMA 307 in the blood of the realgar rats was significantly higher than that of the control group, 308 whether or not the hypercoagulability state is related to the high level of DMA and its 309 specific mechanism，all these requires further study. 310 Realgar may damage kidney function and may also cause 311 hypercholesterolemia and abnormal electrolyte levels 312 After 60 days of administration, urinary protein (PRO) and white blood cell count 313 (WBC) in realgar groups were significantly higher than those in the control group (p 314 <0.05). However, after 90 days of dosing, this condition did not worsen, and they 315 gradually returned to normal after discontinuation. There are no other significant 316 differences in urinalysis results (data not shown). 90 days after the administration of 317 realgar, the blood urea nitrogen (BUN) values of rats in high-dose group (♂) and low-318 dose group (♀) increased significantly (p < 0.05) ( Table 3). While after the 60 days of 319 drug withdrawal, there was no obvious abnormality in BUN value in the realgar group. 320 In addition, cholesterol (CHO) levels were significantly increased in the realgar group 321 (♂) 60 days after administration (p < 0.05), and all the realgar group rats CHO until 90 322 days after dosing and 30 days after discontinuation values (♂, ♀) are still higher than 323 control rats (p < 0.05). Serum BUN and urinary protein were significantly increased in 324 rats, suggesting that renal function may be impaired. Compared with 0 mg·kg -1 ·d -1 325 control group, there was no significant difference in serum biochemical indexes in 326 realgar groups' rats after 60 days of withdrawal (Table 4). The result of blood electrolyte 327 analysis showed that after 60 days of realgar administration, the K+ value of some 328 realgar groups (♂, ♀) decreased (p < 0.05). Furthermore, after 90 days' administration 329 and 60 days after drug withdrawal, the K+ values of male rats in some realgar groups 330 were also significantly lower than that of the control rats (p < 0.05). However, after 90 331 days of realgar treatment, the K+ value of female rats in high-dose realgar group was 332 significantly higher than that of the control group (p < 0.01) (Table3-4). In short, long-333 term high-dose use of realgar has certain effects on rat cholesterol metabolism and 334 blood electrolyte (K + ) levels, but its specific causes and the relationship with arsenic 335 levels in the body is still unclear. 336    were observed after administration of realgar for 90 days (Table 5). However, after 60 356 days of discontinuation, in male rats of the realgar group, the liver ROW was lower than the control rats, and the kidney ROW was higher than that of the control rats (Table  358 6). All animals were observed during dissection. The organs were normal in texture. No 359 obvious swelling, bleeding, adhesions or ulcers were observed in all organs, thorax, 360 abdomen and pelvis. 361 Results are shown as the mean ± SEM for 3 rats per sex in each group. Significantly different from 368 0 mg·kg -1 ·d -1 controlgroup by ANOVA: a P < 0.05, b P < 0.01, c P < 0.001.

369
The results showed that at a dose of 10 mg·kg -1 , which is equivalent to 6 times the 370 maximum clinical dose, realgar has no obvious toxicity to rats' livers after continuous 371 administration for 90 days. With the increase of the realgar dosage, the liver damage 372 caused by it is aggravated. After 60 days of administration, the hepatocytes of rats in 373 middle-and high-dose realgar group showed mild swelling, degeneration, necrosis and 374 focal interstitial infiltration of inflammatory cells. 90 days after administration, the 375 pathological changes of the liver in the middle and high doses of realgar were 376 significantly aggravated, and there was a statistically significant difference compared 377 with the control group (p<0.05). In particular, a few animals exhibited degeneration of 378 flake hepatocytes, eosinophilic degeneration, steatosis, and even bridging necrosis. 379 After 30 days of discontinuation, the high-dose realgar group still had mild swelling 380 and a small amount of inflammatory cell infiltration, and the lesions were statistically 381 different from the control group (p <0.05). After 60 days of treatment discontinuation, 382 there was still a small amount of inflammatory cell infiltration and mild hepatocyte 383 swelling in the liver of high-dose realgar group rats, but there was no statistical 384 difference between the realgar group and liver histological group (Fig 2). 385 The pathological changes of liver tissue were obvious, but there was no obvious 386 abnormality in serum biochemical indicators of reactive liver function. This indicated 387 that the changes of serum biochemical indicators after realgar induced liver injury were 388 later than pathological changes. Thus, when taking realgar for a long period of time, the 389 biochemical indicators commonly used to detect liver function may not reflect early 390 liver injury. Therefore, new sensitive and specific biomarkers reflecting early liver 391 injury should be supplemented. In recent years, the study of drug hepatotoxicity 392 biomarkers has also achieved great development. Several new metabolites, proteins, 393 and nucleic acid microRNAs [26] have been studied as potential biomarkers of liver 394 injury, but still in the experimental research stage. It has been reported that oxidative 395 stress, apoptosis, and up regulation of apoptosis related proteins are the main 396 mechanisms for arsenic induced hepatotoxicity [27-28]. In the large-dose realgar group 397 rats, there was still obvious hepatic injury 30 days after drug withdrawal. Until 60 days 398 after drug withdrawal, the above-mentioned liver tissue pathology basically returned to 399 normal. 400 After continuous administration of realgar for 60 days in rats, a small number of 401 kidney lesions began to appear. Kidney proximal tubule lesion and mild glomerular 402 lesions were the main lesion, manifested as unclear renal tubular border, luminal 403 narrowing into a star or occlusion and other lesions. After 90 days of continuous 404 administration, the rats in the realgar group showed renal tubular epithelial cells 405 swelling, cytoplasm loosening, vacuolar degeneration, partial proximal epithelial cells 406 necrosis and shedding, the formation of cell tube near the lumen, mild glomerular 407 swelling, congestion and a small amount of inflammatory cell infiltration etc. And as 408 the realgar dose increases, the lesions worsen (Fig 2). These pathological changes were 409 statistically significant compared with the renal tissue of the control group (p <0.05, p 410 <0.01). After 30 days of withdrawal, the renal lesions in the three doses of realgar group 411 were alleviated, but the renal lesions in the high-dose group were still more severe than 412 those in the control group (p <0.05). After 60 days of discontinuation, although one-413 third of the rats in the high-dose group had slight tubular epithelial swelling and 414 degeneration, there was no significant difference in renal lesions compared with the 415 control group. In conclusion, when the dose of realgar was >10 mg·kg -1 and the rats 416 were orally administered for 60 days, pathological damage occurred in the kidney, and 417 pathological damage was exacerbated with increasing dose and/or time of 418 administration. Kidney proximal tubules may be the main target site for realgar kidney 419 toxicity. There is a positive correlation between renal injury and realgar dosage and 420 administration time. Some studies have reported that arsenic can cause oxidative 421 damage to the kidneys [29-31]. The renal damage caused by long-term oral 422 administration of realgar is reversible, with the prolonged withdrawal time, arsenic in 423 the kidney gradually decreased, and kidney damage gradually reduced to recovery.

424
In addition, the histological and cellular structures of the brain, lungs, bronchus, adrenal 425 gland, pancreas, bladder, testis, ovary, uterus, and other organs of realgar groups had 426 no obvious pathological changes compared with the control group during 427 administration and withdrawal. Some studies have shown that early exposure to arsenic 428 can cause brain cholinergic deficits. Arsenic accumulates in brain astrocytes and affects 429 their viability and glutathione metabolism [32][33][34]. Due to the limitations of this 430 experimental method, neurotoxicity cannot be directly reflected. Therefore, in this study, 431 there was no direct and concrete evidence that realgar would damage the brain or

448
Realgar led to a significant increase in DMA content in tissues 449 At the end of 60 days' administration, the total arsenic content (tAs) in rats of realgar 450 groups was significantly higher than that of 0 mg·kg -1 realgar control group (p <0.01). 451 With the increase of the realgar dosage, the content of tAS in rats also increased 452 significantly. The content of tAs in different biological samples of rats, arranged in 453 descending order of urine> feces> plasm> kidney> liver> brain. The content of tAs in 454 the above tissues of the realgar group was significantly higher than that of the control 455 group. In plasma, arsenic species content from high to low were 456 DMA>MMA>AsⅢ>AsⅤ. Compared with 60 days after administration, plasma DMA 457 and tAs levels were increased in each of the realgar group after 90 days of 458 administration, but there was no significant difference. After 90 days' treatment, the 459 accumulation of plasm DMA in the low-, medium-and high dose realgar groups were 460 11.11, 19.93 and 23.65 times that of the control group, respectively. After stopping the 461 drug, the arsenic content in the rats of realgar group was significantly decreased, and 462 the arsenic content decreased more significantly with the prolonged realgar withdrawal 463 time. However, tAs and DMA content remained significantly higher than the control 464 group even after stopping for 60 days (p <0.001). Except for DMA, there was no 465 significant difference in the other arsenic levels in plasma among the three realgar 466 groups compared with the control group (Fig 3).

473
The content of tAs in rats' liver of realgar groups was less than one-tenth the content of 474 the plasma. The accumulation of arsenic in the liver was positively correlated with the 475 realgar dose. The content of arsenic species in the liver of realgar group rats were 476 DMA>AsⅢ>AsⅤ>MMA>AsB>AsC. After 90 days of oral administration of realgar, 477 the accumulation of DMA in the low, middle and high dose realgar groups were 12.58, 478 17.75, 28.79 times that of the control group, respectively, which was significantly 479 higher than that of the control group (p <0.001), so we conjectured that DMA may play 480 a role in arsenic-induced liver toxicity. There was no significant increase in liver DMA 481 and tAs levels after 90 days of realgar dosing compared to 60 days after dosing. After 482 discontinuation of treatment, all arsenic species levels in the realgar group decreased, 483 while the DMA content was still higher than that of the control group (p <0.001) (Fig  484  4).

491
DMA, MMA and tAs levels in the kidney of realgar rats were significantly higher than 492 those of normal rats. (p <0.01, p<0.05, p <0.01), and the levels of DMA, MMA and tAs 493 in the kidneys of rats increased significantly with the prolongation of the administration 494 time. Compared to 60 days after the administration, although the increase of DMA and 495 MMA levels in the kidneys of the realgar-treated rats was not significant at 90 days 496 after the administration, the pathological lesions were significantly aggravated. After 497 discontinuation of treatment, the content of tAs and DMA in the kidneys of the realgar 498 groups gradually decreased, but they were still higher than those of the control group 499 (p<0.05). In addition, the renal MMA content also decreased rapidly too, and there was 500 A B no significant difference compared with the control group (Fig 5). 60 days after drug 501 withdrawal, renal function and renal tissue damage almost completely returned to 502 normal. In view of the above results, we speculate that DMA and MMA in the kidney 503 may be related to arsenic-induced oxidative stress

514
Despite the existence of blood-brain barrier, the content of tAs and DMA in the rats' 515 brain after realgar administration for 90 days was significantly higher than those in the 516 control group (p<0.001, p<0.001). This may illustrate DMA can penetrate the blood-517 brain barrier. Compared with 60 days after realgar administration, the content of tAs in 518 the brain of rats increased 90 days after administration, but the DMA content did not 519 increase significantly. After discontinuation of treatment, tAs and DMA levels in the 520 realgar group began to decrease but remained higher than the control group (p <0.05). 521 There was no difference in other arsenic species levels in the brains of realgar and 522 control rats throughout the experiment (Fig 6). Since arsenic is a toxin associated with 523 neurological disease and impaired cognitive function, in order to further understand the 524 relationship between arsenic (DMA) accumulation and brain damage, we believe that 525 a series of in vitro and in vivo tests are needed for research. 526

534
The content of tAs in feces and urine of realgar treated group was significantly higher 535 than that of the control group (p <0.05). As the dosage of realgar was increased, the tAs 536 in feces and urine of the corresponding rats also increased. In urine, DMA and MMA 537 were the two main arsenic species, and the amount of DMA were obviously higher than 538 MMA. Some reports had shown that high DMA% (and low iAs% and MMA%) in urine 539 represent enhanced methylation capacity which thought to decrease the susceptibility 540 to arsenic-related toxicity [36][37]. Compared with the 60th day after administration, 541 the arsenic content in rat urine increased and the arsenic content in feces decreased 90 542 A B days after administration. After realgar stopped for 30 days, the rats in the treatment 543 group were the same as those in the control group, and almost no obvious arsenic was 544 detected in the feces. 60 days after drug withdrawal, the urinary DMA and tAs levels in 545 the realgar group were significantly decreased, but still higher than the control group (p 546 <0.05) (Figs 7-8).

560
It is known that most of the arsenic species (MMA, DMA, AsⅢ, AsⅤ, AsB and AsC) 561 are excreted from rat urine and feces after oral administration of realgar. Once 562 discontinuation, the arsenic content in urine and feces of the realgar group dropped 563 drastically, and the decrease was greater than the decrease of arsenic in blood and 564 tissues. After 60 days of discontinued realgar, total arsenic and DMA levels in the blood, 565 liver, brain and kidneys of the realgar rats were still significantly higher than those of 566 the control rats. That is to say, after oral administration of realgar for long time, most 567 of the arsenic species are excreted, but the elimination of arsenic species that 568 accumulated in tissues is relatively slow. This may be one of the reasons that liver and 569 A B kidney tissues of rats still show pathological damage after drug withdrawal. 570

571
It can be seen from the results of this study that when the dose of realgar is≤10 mg·kg -572 1 , continuous oral administration for 90 days does not cause obvious toxicity in rats. 573 While 40mg·kg -1 ·d -1 and 170 mg·kg -1 ·d -1 of realgar (corresponding to 40-fold and 100-574 fold the maximum clinical dose, respectively) can cause toxicity in rats, liver and 575 kidney may be the main toxic target organs of realgar. Therefore, the no observed 576 adverse effect level (NOAEL) dose in this study is considered to be 10 mg·kg -1 .

577
NOAEL is known to be an important reference when setting safety limits for human 578 body equivalent dose (HED). International regulatory agencies recognize a 100-fold 579 safety factor for non-carcinogens. In order to ensure safety, the safety factor for 580 carcinogenic and teratogenic substances is formulated to be more prudent, and there are 581 different regulations depending on different situations. According to the long-term 582 toxicity test guidelines of ICH M3, the clinical safety of drugs with a drug cycle of 2 to 583 4 weeks can be referred to the 3-month repeated-dose toxicity test of rodents. Therefore, 584 based on the NOVEL dose in this study, the authors preliminarily concluded that the 585 safe use of realgar was approximately 0.1 mg·kg -1 · d -1 for 2-4 weeks of clinical use. 586 In addition, the kidney and liver are the main organs for arsenic accumulation. The 587 degree of liver and kidney injury and the accumulation of arsenic were positively 588 correlated with the dose and time of administration of realgar. DMA is the main arsenic 589 species after metabolism of realgar, so it may be the main material basis for realgar 590 exerting efficacy and toxicity. Modern studies have shown that DMA exert toxic and 591 anti-cancer effects by inducing cell differentiation and apoptosis. The authors 592 speculated that the content of DMA in tissue has a certain safety range, beyond this 593 range, liver and kidney toxicity will be induced. With the increase of DMA 594 accumulation time or the accumulation of accumulation, the toxic effect will increase. 595 However, the specific quantitative range of DMA-induced toxicity requires more 596 intuitive and accurate tests to determine. Although large doses or long-term use of 597 realgar have the risk of inducing liver and kidney toxicity, this toxic effect is reversible, 598 that is, it can be restored after drug withdrawal. Therefore, it is recommended that liver 599 and kidney function, blood coagulation, fat metabolism and electrolyte levels should be 600 routinely monitored when the realgar preparation is used for more than 2 weeks. Once 601 a toxic reaction occurs, medication should be stopped immediately to restore organ 602 dysfunction. 603