In humans, chronic liver disease (CLD) is a large group of clinical conditions characterized by impairment of liver function and morphology lasting over 6 months. The most common causes of CLD are alcohol consumption and viral infection. Less common etiological factors of CLD include nonalcoholic steatohepatitis, biliary obstructive disorders, right sided heart failure, Wilson’s disease, hemochromatosis, and chronic use of specific medications [1]. Regardless of the etiology, untreated CLD leads to cirrhosis, which is defined as pathological remodeling of the liver parenchyma with fibrotic scar formation and presence of regenerative nodules. The pathogenesis of this disease involves the destruction of parenchymal cells by the causative agent. Apoptosis or necrosis of hepatocytes induces an inflammatory reaction that activates hepatic stellate cells (HSCs) via multiple cytokines such as IGF-1, TGF-$\alpha$, and EGF. Activated HSCs differentiate into myofibroblasts, which secrete large amount of collagen and extracellular matrix (ECM) components into the space of Disse [2]. CLD often shows no signs or symptoms. At a later stage of the disease, signs and symptoms result from decreased parenchymal cell volume (e.g., hypoalbuminemia, coagulopathy, encephalopathy, jaundice) or portal hypertension (e.g., ascites, esophageal varices) caused by collagen deposition in perisinusoidal spaces. The primary treatment for CLD is withdrawal of the causative agent (alcohol, medications) or treatment of the underlying disease (viral infection, biliary obstructive disorders). At an early stage of cirrhosis, treatment involves dietary and electrolyte therapy, blood component transfusions, antibiotic therapy, and administration of drugs reducing portal hypertension ($\beta$-blockers, simvastatin). In cases of complications such as bleeding from esophageal varices, additional life-saving surgical procedures are needed [3]. When signs and symptoms of CLD are severe and the patient meets the appropriate Child-Pugh score criteria, the only effective therapeutic option is liver transplantation. However, organ transplantation has important drawbacks including high cost, low organ availability, the need for lifelong immunosuppression therapy, and risk of multiple complications [4, 5, 6]. Therefore, it is imperative to develop an effective therapy that will be able to slow down or even stop fibrosis progression to avoid the need for liver transplantation. Stem cell-based therapies could potentially slow down fibrosis by inhibiting the inflammatory reaction [7, 8].

Animal models of hepatotoxicity induced by hepatotoxins, including carbon tetrachloride (CCl${}_4$) and D-galactosamine (D-GalN), can be used in preclinical assessment of CLD therapies. Such models allow for long-term observation that cannot be performed in patients for ethical and medical reasons, particularly observation of the progression of liver diseases [9, 10] characterized by fibrosis, as well as assessment of the efficacy and safety of an advanced therapy. Unfortunately, the various mechanisms of xenobiotic toxicity and the different sensitivity of animal species and strains to intoxication result in different intoxication effects and, consequently, some difficulties in interpreting the data and extrapolating them to humans [10]. To induce fibrosis in mice and rats, the CCl${}_4$ dose should be 10–200 $\mu$L per 100 g bw, administered two to three times a week. Fibrosis incubation time varies by protocol and is usually between 4 and 12 weeks. To induce cirrhosis, incubation time should be extended to 12–20 weeks. The ability of D-GalN to induce cirrhosis is questionable. Compared to CCl${}_4$, D-GalN causes less severe inflammatory reaction, and thus D-GalN has lower potency to induce fibrosis. To induce fibrosis in mice and rats, the D-GalN dose should be 50–150 mg/100 g bw and 25–50 mg/100 g bw, respectively, and the period of exposure to this hepatotoxin should be at least 12 weeks. Despite the above-mentioned drawbacks, D-GalN intoxication better histopathologically mimics human viral infection compared to CCl${}_4$.

In the present study, we compared the development of histopathological changes induced in Sprague Dawley rat and BALB/c mouse experimental models by repeated injections of CCl${}_4$ and D-GalN at doses and frequencies described as effective [11, 12, 13, 14]. We analyzed the usefulness of these models with a view to their future introduction into preclinical experiments. In particular, based on histopathological and molecular changes it allowed us to identify potential experimental time cut-off points when pharmaceutical or stem cell therapeutic intervention might be effective. So far, the effectiveness of antifibrotic cell therapy has been proven in animal models using primary hepatocytes, mesenchymal stem cells (MSCs) [15, 16, 17], and embryonic stem cells (ESCs) [18]. Recently, a novel cell therapy using immunomodulatory human amniotic cells (hAC) has been proposed [7, 19, 20, 21, 22, 23].

In this study, we assessed the progression of histopathological changes in the livers of experimental animals, in the chronic liver injury model, to select the best time for potential pharmaceutical or stem cell therapy. During long-term intoxication, in response to parenchymal inflammatory reaction, stellate cells produce collagen and other extracellular matrix components. Collagen fibers enter the perisinusoidal space of Disse and, as it were, seal the permeable barrier between the sinusoids and the liver parenchyma, effectively inhibiting the flow of substances and migration of administered cells that could leave the vascular bed. It has been observed that in people without liver fibrosis, this organ is a frequent site of metastasis of malignant tumors from other organs due to the favorable hemodynamic conditions of flowing blood, permeability of the sinusoidal wall, and the small diameter of sinusoids, which favors the formation of microemboli from circulating cells. The resulting emboli retain cells within the vascular network of the organ, promoting their migration into the liver parenchyma, which would be impossible if the cells were suspended in the flowing bloodstream [38]. In contrast, metastases from distant organs are found very rarely in patients with developed cirrhosis.

Nevertheless, cirrhosis itself requires treatment which, in addition to eliminating etiological factors, involves slowing down the progression of the disease, including fibrosis, and supporting the functioning of the liver. Failures of CLD cell therapies were most likely due to unfavorable biodistribution of the administered cells caused by advanced fibrosis [39, 40]. So far, no detailed model studies have been conducted to analyze how the duration of the intoxication, and thus the degree of fibrosis, affects the biodistribution of stem cells administered for therapeutic purposes [11, 14, 41, 42, 43, 44].

Experimental models used to induce liver injury include intoxication with CCl${}_4$ or D-GalN. The main mechanism of liver injury by CCl${}_4$ involves its cytochrome P450 (CYP)-dependent biotransformation resulting in free radical generation, cell membrane oxidation, cytoplasmic ion imbalance, DNA damage, necrotic cell death, and inflammation [45]. In the groups where mice and rats were intoxicated with repeated injections of CCl${}_4$, we observed a decrease in CYP2E1 expression at some time points, which may be related to advanced histopathological changes and secondary hepatocyte damage via a free radical mechanism, and is consistent with the observations that CCl${}_4$ is a potent CYP2E1 inhibitor [46, 47]. On the other hand, the slight increase in CYP2E1 expression observed by us in mice and rats intoxicated with D-GalN at certain time points literature data confirmed a small effect of D-GalN-intoxication on CYP2E1 expression in rodents [48, 49, 50].

D-GalN is a highly hepatospecific compound. Unlike other hepatotoxins, D-GalN does not directly damage other organs and does not cause irritation when injected. In hepatocytes, D-GalN is eliminated by the galactose metabolic pathway, including D-GalN phosphorylation to galactosamine-1-phosphate (GalN-1-P), followed by its conversion to UDP-galactosamine, which has a higher affinity for uridine diphosphate (UDP) than for galactose. The trapping effect leads to uridine deficiency, inhibition of RNA and protein synthesis, and apoptotic cell death [51].

The primary lasting effect of our intoxication procedures was the progressive liver fibrosis. The mechanism of fibrosis in models using CCl${}_4$ and D-GalN is based on an inflammatory reaction initiated by damage to the liver parenchyma. In response to inflammation, hepatocytes and Kupffer cells produce, among others, TGF-$\beta$ and PDGF, which are involved in the activation of stellate cells responsible for the process of synthesize extracellular matrix components and fibrosis through the deposition of a large amount of collagen I and moderate amounts of collagens III and IV in the perisinusoidal space [44, 52, 53]. We found that the groups of CCl${}_4$-intoxicated rats and mice, where the dynamics of fibrosis were fastest, had the greatest increase in the percentage area occupied by $\alpha$-SMA${}^{+}$ stellate cells.

The data obtained by us and other research teams suggest that the onset of fibrosis in the models using CCl${}_4$ occurs very early [3, 37, 38], and the rapid dynamics of its progression make it possible to distinguish different stages of organ fibrosis. In rats and mice chronically intoxicated with CCl${}_4$, we observed early fibrosis in the form of fibrous expansion of collagen fibers from some pericentral areas at week 2 of intoxication. Both groups developed established fibrosis after 6 weeks of CCl${}_4$ intoxication. In other studies, both in mouse and rat models using CCl${}_4$, the development of established fibrosis was also observed between weeks 8 and 12 of intoxication [54, 55, 56, 57]. Bubnov et al. [12], who used the same dose and regimen of intoxication as in our experiment, based on ultrasound examination and standard staining of liver sections, suggest the development of cirrhosis in rats at week 8 of intoxication. Nevertheless, in rats at week 12 we observed the presence of regenerative nodules indicative of the development of incomplete cirrhosis. Overall, the data we obtained, in particular the morphometric analysis of the percentage area occupied by collagen fibers in the liver parenchyma and the Ishak score, suggest the onset of early cirrhosis in rats as late as at week 12 of intoxication. Studies in C57B mice using the same xenobiotic dose and intoxication regime as in our study (12.5 $\mu$L/100 g CCl${}_4$ bw every fifth day) showed the development of established fibrosis after 4 weeks and of incomplete cirrhosis after 8 weeks [11], indicating that the BALB/c mouse strain is less prone to develop fibrosis than the C57B strain [58]. In order to obtain a picture of cirrhosis produced in mice, it would be necessary to extend the duration of intoxication with CCl${}_4$, e.g., to 20 weeks and/or to increase its dose, e.g., to 30 $\mu$L/100 g [59]. In CCl${}_4$-intoxicated rats, where fibrosis was most evident, we noted an upward trend in the percentage of thick collagen fibers with a simultaneous decrease in the percentage of thin fibers over time. The process of pathological fibrosis is accompanied by an extracellular growth of collagen fibers in thickness. Our observation is consistent with literature data suggesting a relationship between an increase in the percentage of larger-diameter fibers and the processes of pathological fibrosis, aging of the body, and would healing [23, 48, 49].

At the molecular level, the process of fibrosis is influenced by a number of factors, and in particular depends on the expression of the genes TGF$\beta$, COL1A1, and COL3A1. Increased expression of these genes results in the deposition of collagen at necrotic sites and formation of connective tissue scars. Increased amounts of type III collagen are found in many fibrotic conditions, such as liver and kidney fibrosis. However, these parameters are not used in routine clinical practice during the diagnosis and assessment of prognosis of CLD as their expression level does not correlate with the severity of the disease [60]. Transforming growth factor-$\beta$ (TGF$\beta$) is an inflammatory cytokine that stimulates stellate cells to produce collagen and plays an important role during the development of chronic liver injury in both experimental animals and humans [14]. TGF-$\beta$ promotes fibrogenesis in three different ways: it inhibits extracellular matrix degradation by inhibiting matrix metalloproteinases; it induces myofibroblast formation; and it induces matrix production through SMAD protein-dependent or -independent mechanisms [61]. In general, animal experiments indicate that the expression of genes responsible for fibrosis fluctuates over time [62]. As a result, despite the intensification of the extracellular processes of thickening of collagen fibers seen in microscopic images, we found a very low expression of the COL1A1 gene in the test animals, stable expression of COL3A1 in rats, and increased expression of COL3A1 in CCl${}_4$-intoxicated mice (at all studied time points).

We observed morphological features of liver injury accompanying the fibrosis resulted from chronic liver injury [12], where the CCl${}_4$ dose in intoxicated rats was gradually reduced during the experiment. These changes were such as ballooning degeneration, inflammatory infiltration, hepatocyte edema, signs of organ steatosis in the form of lipid droplets in the cytoplasm of hepatocytes, and necrosis, mainly within zone 3 of the hepatic acinus, i.e., around the central vein. In mice intoxicated with repeated injections of 12.5 $\mu$L/100g bw CCl${}_4$, we observed the same features and localization of liver injury as in rats. However, in mice we did not observe liver steatosis. Thus, histopathological changes in mice and rats were localized similarly to those in chronic viral hepatitis or alcohol-induced liver fibrosis [63].

An additional effect of liver injury in the CCl${}_4$ models was an increase in proliferative activity and/or in apoptotic activity of hepatocytes in rats and in mice. The coincidence of changes in apoptotic and proliferative activities in these models may indicate the simultaneous action of injury and repair processes of variable intensity associated with alternating injury and restoration of the structure and function of the liver [64]. Similar dynamics of injury also occur in humans in viral hepatitis and chronic alcohol consumption, where damage progresses during disease exacerbations, while during remissions, functional and morphological regeneration of the organ occurs [65]. It was shown in acute liver diseases that increased expression of HGF and its receptor c-Met induces hepatocyte proliferation and organ regeneration. Decreased HGF expression in chronic diseases is explained by a partial loss of the ability of the liver parenchyma to regenerate [41, 66].

In the liver fibrosis models using repeated injections of D-GalN (25 mg/100 g bw in rats and 75 mg/100 g in mice), we observed the slow dynamics of fibrosis progression over time. Microscopic changes were accompanied by a transient increase in COL3A1 gene expression. Similar changes, especially in mice, have been previously noted by other researchers. In the rat models using repeated injections of 25 mg/100 g D-GalN bw, the presence of early or established fibrosis was demonstrated at week 12 of intoxication [14, 62]. In order to induce irreversible cirrhosis in the rat model, the duration of the intoxication should be extended to up to 6 months [67]. In the mouse model, an attempt to induce cirrhosis with a relatively high dose of 150 mg/100 g D-GalN bw administered regularly once a week for 13 weeks failed [68]. Overall, our results and literature data suggest that the rat and mouse models of CLD using repeated D-GalN injections offer significantly less opportunity to achieve established liver fibrosis and cirrhosis within a predictable and defined period of frame.

Fibrosis observed in D-GalN-intoxicated rats was accompanied by a relatively small inflammatory reaction and weakly pronounced visible histopathological changes, such as hepatocyte edema and ballooning degeneration. In the mouse model of chronic D-GalN administration, we did not observe any significant histopathological changes. In both rat and mouse models of D-GalN-induced CLD, other researchers observed little or no damage to the liver parenchyma [14]. Also, the increase in apoptotic activity only at certain time points after intoxication in rats, lack of apoptotic activity in mice, and temporary tendency to increase in proliferative activity observed in both rats and mice indicate relatively low potential of the D-GalN doses administered chronically to induce CLD. The scant literature data indicate a small effect of repeated low-dose injections of D-GalN on the induction of apoptosis [14, 62]. Increased apoptotic activity in the liver parenchyma is much more pronounced in the acute or fulminant injury model, where a single injection of high doses of D-GalN is used [69, 70]. Nevertheless, the decrease in the expression of the HGF gene and its receptor c-Met, found in animal groups studied by us, mostly those treated with D-GalN, coincides partially with literature data, which describe an increase in HGF expression in the initial stage of chronic hepatitis and a decrease in the expression of this gene in the advanced stage of the disease [66].

It should be noted that the histopathological changes recorded in the rat and moue livers in both intoxication models were not clearly reflected in the results of tests carried out on the blood of the test animals. In the models using CCl${}_4$ and D-GalN, we did not observe a clear temporal trend of changes in the values of the tested markers (ALT, AST), which may be the result of the doses used, the regenerative capacity of the liver, and the 72-hour interval between the collection of the material for testing and last xenobiotic injection. The blood picture in the studied groups also did not show a specific trend of change. In humans, the increase in transaminase activity in CLD has only auxiliary diagnostic value. In the course of chronic liver injury, aminotransferase values are often normal and their significant increase suggests exacerbation of the disease [42]. Our previous studies show a strong upward trend in the above-mentioned parameters during the first 48 hours after CCl${}_4$ injection in rats and mice and D-GalN injection in rats (but not in mice) in the models of acute liver injury [43].

To sum up, the used doses of hepatotoxins allowed to show subtle differences in their effects over time, while avoiding too abrupt and advanced changes without the possibility of tracing intermediate stages. We assessed the severity of reversible and irreversible structural and functional changes in the liver in individual CLD models to determine the potentially optimal time of therapeutic intervention. Such CLD models might apply to different types of pharmaceutical treatments and to stem cell therapy. However, it should be noted that stem cell therapy has much more variables that might completely render the “timing” a less important limiting factor. For instance, the type of transplanted cells, their number, the pre-treatment/pre-genetic engineering of injected cells, the route of injection, the co-transplantation of more than one cell type, and many others are very powerful limiting factors that might influence its effectiveness. Based on this preclinical study, we can expect that in experimental rodents with established fibrosis large amount of collagen fibers which seal the sinusoids barrier, might prevent implanted cells homing process.

Summarizing the results obtained in the mouse and rat models using CCl${}_4$, we can conclude that the dynamics of liver fibrosis in CCl${}_4$ treated animals were greater than in the D-GalN groups.

In mice, and rats CCl${}_4$ model we indicated two appropriate times for therapeutic intervention, which to varying degrees reflect the real clinical situation and may potentially differ in the obtained results: early intervention before week 4 of intoxication (during early fibrosis stage) and late intervention after week 8 of intoxication (when signs of established fibrosis are present).

We do not recommend rodent models of D-GalN-induced liver fibrosis due to the long incubation period, poor effect, and high costs.

CCl${}_4$, Carbon tetrachloride; D-GalN, D-galactosamine; CLD, chronic liver disease; i.p., intraperitoneal; ALT, alanine transaminase; AST, aspartate transaminase; ALP, alkaline phosphatase; TP, total protein; COL1A1, Type I collagen; COL3A1, Type III collagen; TGF-$\beta$, Transforming growth factor beta; c-Met, Tyrosine-protein kinase Met; HGF, Hepatocyte growth factor; CYP2E1, Cytochrome P450 2E1; PPAR-$\alpha$, Peroxisome proliferator-activated receptor alpha.

All data generated or analyzed during this study are included in this published article.

PC and MK designed the research study. PC supported the studies financially (grands), provided help and advice. MK, EK, ŁL, AP, AS-S, EB, BS, MH, MM, AG, JP performed the research. PC, MK, ŁL, AS-S, BS, AP analyzed the data. PC and MK wrote the manuscript. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.

Animal experiments were approved by the Animal Experiments Ethical Committee of Medical University of Silesia, Katowice, Poland (decision no. 18/2018).

We thank the Silesian Analytical Laboratory (Katowice; Poland) for performing serum biochemistry analysis and blood morphology assessment.

The studies were supported by institutional grants (SUM Katowice) no: KNW-1-103/N/8/0 and KNW-1-100/K/9/0.

The authors declare no conflict of interest.