A Comparative Analysis of Human Hepatic Lesions in Dengue, Yellow Fever, and Chikungunya: Revisiting Histopathological Changes in the Light of Modern Knowledge of Cell Pathology

1. Introduction

Arboviruses (arthropod-borne viruses) develop part of their replication cycle in insects and can be transmitted to humans and other animals by the bite of hematophagous arthropods. These viruses have been identified in almost all continents, and tend to have a geographic distribution as part of a special ecological subsystem represented by viruses, vectors, amplifying hosts, and reservoirs [1,2,3].

In recent years, the importance of increased climate and environmental changes due to anthropogenic factors such as deforestation favor the transposition of the barrier between species [4]. In addition, the increase in activities, such as trade and tourism, helps spread these viruses from endemic regions to different parts of the world [5].

Favor the increase in arbovirus infections. Many countries have experienced the periodic emergence of arboviruses [6]. The establishment of arboviruses in certain places is based on the complex interaction of biotic and abiotic factors, which favor the occurrence of epidemics and epizootics in different continents [7].

Arboviruses such as yellow fever virus (YFV), dengue virus (DENV), and chikungunya virus (CHIKV), belonging to the Flavivirus genus and Flaviviridae family (YFV and DENV) and Alphavirus genus and Togaviridae family (CHIKV), are related to seasonal circulation in tropical and subtropical areas [8,9,10]. According to the World Health Organization, 47 countries are endemic or have endemic regions of yellow fever (YF), of which approximately 72% are located in Africa and 28% in Central and South America [11].

In South America, countries with a concentration of confirmatory diagnoses for the disease are Peru, Bolivia, and Brazil [12]. DENV, on the other hand, has a more extensive area of circulation, with cases of dengue fever (DF) being common in 100 countries worldwide [13,14]. In the Americas, approximately 500 million people are at risk of infection. Brazil and Colombia are the South American countries that led the ranking of cases of the disease [15].

Unlike YFV and DENV, which are more widely circulated viruses, CHIKV has recently demonstrated an impact on a global scale. Autochthonous transmission has been reported in 40 countries worldwide. However, 115 countries have reported cases of chikungunya fever (CF), with approximately 41% of the reported cases being in the Americas, and when observing South America, as well as in DF, chikungunya cases are concentrated in Brazil and Colombia [15].

These arboviruses develop transmission cycles in wild and urban environments and are transmitted by the bites of mosquitoes belonging to the Culicidae family [16,17]. Currently, in Brazil and South America, DF and CF cases are predominantly urban, while in YF, the prevalence is in the sylvatic cycle of the disease, determined by epizootics in different regions of the country, mainly in the southeast, northeast, and north [18]. Brazil is considered the main source of the circulation and dispersion of these arboviruses in other Latin American countries [19,20,21,22].

YFV, DENV, and CHIKV are responsible for causing different clinical manifestations in infected individuals, several factors influence the aggravation of the disease. They are associated with both the characteristics of the host and the characteristics of the viruses, including the patient's age, viral load, pattern of immune response, secondary infection, and interaction of viral proteins and host cell receptors. This host-virus relationship, plays an important role in the clinical outcome of the disease, as it determines not only the viral tissue tropism but also a multitude of intracellular mechanisms that can alter the overall outcome of the infection [23,24,25].

Approximately 15% to 60% of individuals infected with YFV develop symptoms of the disease, with evolution to death in 20% to 50% of cases [26,27]. In DF cases, lethality is also associated with the infecting serotype, which has four serotypes (DENV-1 to DENV-4), which influence the patient's clinical outcome [28]. In CF cases, most infected individuals develop symptoms. Severe cases tend to have atypical characteristics of the disease (hepatitis, myocarditis, and neuropathies) [29,30].

These arboviruses have been observed to replicate and cause significant damage to various organs of the host, such as the heart, lung, thymus, spleen, kidney, and liver; therefore, they are considered to be multi organic pathogens [30,31,32,33,34]. The liver is considered a target organ in YFV and DENV infections and is characterized by CHIKV infection [35,36]. The pathogenesis of liver injury involves several mechanisms, including direct viral cytopathic effects and immune-mediated injury [37].

The diagnosis of patients infected with these viruses can be performed using several techniques, including serology, viral isolation, immunohistochemistry (IHC), molecular biology, and histopathology [38,39,40,41,42,43,44,45]. Previous studies in animal models and histopathological results in humans have allowed the characterization of clinical and pathological changes that develop during viral pathogenesis [46].

The hepatic histopathological patterns observed in the pathogenic profiles of these arboviruses show different degrees of degenerative lesions, with the presence of inflammatory infiltrate, macro-and microvesicular steatosis, necrosis, alterations in hepatic sinusoidal endothelial cells (HSEC), and Kupffer cells [12,47,48,49,50,51,52,53,54,55,56].

Taking into account the viral pathogenesis of YFV, DENV, and CHIKV that are associated with hepatic impairment, this study aimed to characterize, quantify, and compare the patterns of histopathological alterations in human liver samples from patients with YF, DF, and CF, contributing to the improvement of the criteria for histopathological diagnosis of these diseases through systematic and comparative characterization, which can also help expand clinical-epidemiological criteria and expand knowledge regarding the harmful processes developed in human liver tissue in YFV, DENV, and CHIKV infections.

2. Materials and Methods

This was an analytical cross-sectional study with retrospective samples of patients with confirmed laboratory diagnoses who died due to infection and complications by YFV, DENV, and CHIKV. Patients' ages were 15 and 63 years (mean age = 37 years) and 60% were male. The samples came from the states of Tocantins, Goiás, Distrito Federal and Paraíba, where most cases of infections came from the state of Goiás, between 2000 and 2016.

Samples collected from patients with a confirmed diagnosis of YFV, DENV, and CHIKV infections by clinical/epidemiological, histopathological, IHC, and molecular criteria were included.

Samples that did not have confirmatory histopathological, IHC, or molecular reports were excluded. Samples that did not have enough material to make slides, that is, the amount of material was not enough for making slides, and its use could compromise the storage of a part for the return to the laboratory block file. In addition, the samples that were identified were impaired.

Seventy liver tissue samples from individuals with YF, DF, and CF were included in the study (36 samples of YF, 17 of DF, and 17 of CF) from the biobank of the Pathology Section of Institute Evandro Chagas (SAPAT/IEC), located in the Municipality of Ananindeua, arising from the epidemiological surveillance of the period from 2000 to 2017. These cases were diagnosed based on clinical/epidemiological criteria, histopathological examination, detection of viral antigens by IHC, or viral genome by polymerase chain reaction with reverse transcription-polymerase chain reaction in blood or frozen liver samples for diagnosis (RT-PCR).

Liver samples obtained during routine autopsies were included as negative controls. Twenty liver samples from patients with a negative molecular diagnosis for the main hepatotropic viruses, with no morphological changes, and from the Death Verification Service of the Metropolitan Region of the city of Belém, State do Pará, Brazil (Supplementary Figure).

Samples that did not have confirmatory histopathological, IHC, or molecular reports were excluded. Samples that did not have enough material to make slides, that is, the amount of material was not enough for making slides, and its use could compromise the storage of a part for the return to the laboratory block file. In addition, the samples that were identified were impaired.

All procedures performed during this study were approved by the Research Ethics Committee of the Institute Evandro Chagas, following the recommendations by the resolution of National Health Council No. 466/12. It was approved with perish number 2.364.226. Free and informed consent terms were waived, as the samples collected were obtained post-mortem, and all data will be handled and analyzed anonymously, without the nominal identification of individuals.

4. Discussion

Arboviruses, such as YF, DF, and CF, are considered diseases that pose risks to global health, as they behave epidemiologically in the form of epidemics and epizootics, with outbreaks of a seasonal pattern of greater or lesser impact on public health. In the tropical regions of Africa and South America, biotic and abiotic factors favor the establishment and circulation of these arboviruses, such as rainfall, climate, temperature, presence of dispersing vectors, and amplification of hosts and reservoirs [8,59].

Brazil is considered the main source of the circulation and dispersion of these arboviruses in other Latin American countries [19]. In South America, Brazil, Bolivia, and Peru are the countries that lead the ranking of confirmed cases of YFV, DENV, and CHIKV infections. In Brazil, cases of DF and CF are predominantly urban, while in YF, the prevalence is in the sylvatic cycle of the disease, determined by epizootics in different regions of the country, mainly in the southeast, northeast, and north [18]. In northern Brazil, the Amazon region predominates these arboviruses and is classified as an endemic area for these diseases [20,21,22].

Such diseases have several clinical outcomes, framed as asymptomatic, mild, and fatal cases; the latter is characterized by multiple organ dysfunction, hemorrhage, severe jaundice, hepatitis, myocarditis, and neuropathies [26,27,60]. These clinical manifestations depend on several factors, such as viral load, tissue tropism, genetic factors, and immune evasion mechanisms, referring to both the virus and host. In addition, the interaction of viral proteins and host cell receptors may be relevant, as their interaction dictates not only viral tissue tropism but also a multitude of intracellular mechanisms that can alter the overall outcome of the infection [24,25].

Although several studies have addressed the clinical manifestations and hepatic histological changes of YF and DF, little is known about the hepatic involvement in CHIKV infection or hepatic histopathological differences in fatal cases of YF, DF, and CF. It is known that these three arboviruses are correlated with the development of liver tissue damage mechanisms, and worsening of the infection is associated with a greater probability of a lethal outcome. What are the histopathological similarities and differences in liver involvement among these three arboviruses?

Previous studies have reported several causes of the pathogenesis of these viruses, which are inherent to the host and virus. The viral proteins nsP2 and nsP3 of the alphavirus CHIKV and NS1, NS4A/B, and NS5, which belong to the flaviviruses YFV and DENV, play roles in the development of viral cytopathic effects. In addition to being associated with persistence in host cells, it promotes the attenuation of antiviral responses by inhibiting key factors of the immune response, resulting in viral dissemination in the host organism [25,61,62,63,64,65,66].

Considering the multi organic character of these arboviruses, related to infection and proliferation in different organs of the host [52,67]. the hepatic histopathological findings referring to YF, DF, and CF associate YFV infection, DENV, and CHIKV to different degrees of degenerative lesions, with the presence of inflammatory infiltrate, macro and microvesicular steatosis, necrosis, and changes in HSEC and Kupffer cells [30,31,47,48,51,52,68].

In our study on the semiquantitative histopathology of 70 liver samples from patients with YF, DF, and fatal CF, the quantification of histopathological changes showed abnormalities associated with cell swelling, macro and microvesicular steatosis, lytic and coagulative necrosis, apoptosis, hyperplasia, hypertrophy of Kupffer cells, endothelial changes, and inflammatory infiltrate. Cellular swelling in DF and CF cases showed homogeneous intensity, with no differences between the analyzed acinus zones; however, for CF cases, the swelling was intense, and in cases of moderate DF, in YF, an intensity variation was observed between the zones, in which in Z1 the swelling was intense, and moderate in Z2 and Z3.

Although there are no data on swelling in the liver tissue in CF cases, several studies have shown that YFV and DENV infections are associated with the development of hepatocyte swelling [69,70,71,72,73,74]. According to a study by Quaresma et al., the hepatocytes in Z1 and Z3 showed a greater intensity of this alteration. Cell swelling is attributed to changes in the mechanisms controlling water and ion concentrations due to direct viral action on the cell membrane or the immune response generated by the host [75,76].

As for macro-and microvesicular steatosis, the results showed that there was no significant difference between arboviruses with uniform intensity between them in macrovesicular steatosis. However, a moderate difference in intensity was observed between the acinar zones, and the cases of YF, DF, and CF showed a moderate increase in macrovesicular steatosis in Z2. In microvesicular steatosis, this alteration was not identified in CF cases in any of the three areas evaluated, whereas the intensity observed in YF and DF cases was equivalent, with moderate microvesicular steatosis in Z1 and Z3 and intense in Z2. Similar to macrovesicular steatosis, a preferential concentration of small lipid vesicles was observed in the cytoplasm of hepatocytes located in the midzonal area. A predominance of microvesicular steatosis was also observed in cases of YF and DF, unlike the results for macrovesicular steatosis, in which there was no difference between the arboviruses studied (Figure 9).

In YF, macro-and microvesicular steatosis is a change commonly observed in infected patients, present in the three zones of the hepatic acinus, but it is predominant in Z2 [57,77,78]. DF is associated with fatty liver degeneration [79,80]. Recently, Win et al. reported findings in human liver samples from patients who died of dengue hemorrhagic fever (DHF), which showed extensive cell damage and microvesicular steatosis [37]. Most studies relate YFV and DENV infections to the detection of microgoticular steatosis, the presence of which is related to changes in lipid metabolism, one of the mechanisms involved in hepatomegaly, a clinical manifestation associated with hepatitis observed in YF and DF [31,57,70].

Although data on hepatic fatty degeneration in CF are scarce, flavivirus and alphavirus infections have lipid metabolic implications, influencing the immune response against infection and the maintenance of viral pathogenic mechanisms [62]. In addition, studies have reported that the synthesis of lipid mediators during viral infection is related to the signaling, control, and maintenance of the immune response and viral pathogenesis [82,83,84,85,86].

During infection, the virus can control the lipid metabolism of the host cell in order to meet the demands required in the viral replication process [87,88]. Since the mobilization and recruitment of lipid droplets is an essential mechanism in the bioenergetic demand for viral proliferation, which occurs through the formation of replication complexes in regions of lipid concentration [89,90,91,92]. Thus, lipid metabolism is directly linked to the development of YFV, DENV, and CHIKV pathogenesis [93,94,95].

According to a study by Quaresma et al., the intensity of steatotic hepatocytes coincides with the intensity of apoptotic processes. Two studies on chronic hepatitis correlated steatosis with unfavorable prognosis [96,97]. Pre-existing non-alcoholic steatohepatitis is also an aggravating clinical factor that may potentiate liver failure during the disease [98].

Regarding the necrotic processes analyzed in liver samples from patients infected with YFV, DENV, and CHIKV, the degree of injury ranged from mild to moderate. There was no significant difference between arboviruses in lytic and coagulative necrosis, with equivalent intensity of injury. However, in lytic necrosis, there was a preferential concentration of lesions in the midzonal area of the YF, DF, and CF samples, with a moderate degree of intensity. In evaluating coagulative necrosis, all cases in the three zones of the hepatic acinus corresponded to a mild degree of this alteration, with no significant difference between the arboviruses or zones (Figure 10).

According to the histopathological findings of other studies, severe forms of DF are determined by the presence of hepatocellular necrosis, usually observed in Z1 and Z2 of the hepatic acinus [36,99]. In addition, several studies have associated the frequency of distinct histopathological findings in YF and DF with the Z2 of the hepatic acinus, but this pattern has also been observed in infections by other arboviruses [51,57,76,100,101,102].

In our findings, Z2 was the area with the highest concentration of damage, with the degree of injury varying from moderate to very intense, particularly in the case of YF, which presented the highest levels of harmful processes in the hepatic parenchyma. According to Pan et al., the development of injurious mechanisms in Z2 of the hepatic acinus is characteristic of the pathogenesis of flavivirus infections.

This pattern of injury in the midzonal area is attributed to the intrinsic characteristics of hepatocytes in this area, which have a lower presence of oxygen and nutrients when compared to Z1 [70,99,104]. In a study by Olímpio et al., lytic and coagulative necrosis were recurrent findings in the three hepatic zones of patients with YF, but the highest frequency was in Z2. The triggering of these cell destruction processes may be related to the activation of a varied set of defense cells, generating a strong immune response characterized by a cytokine storm that causes tissue damage [47].

Hepatic histopathological findings from human cases of co-infection with DENV-CHIKV showed that patients had coagulative necrosis, especially around the hepatic vein [30,33,68]. In animal models of CHIKV infection, infected mice presented on the third day after infection, foci of hepatocytic necrosis dispersed in the liver parenchyma without a specific zone [105,106,107,108].

Regarding apoptosis, we observed significant differences between arboviruses; CF cases showed the absence of this alteration in all analyzed areas. In DF cases, the results were homogeneous between the zones; that is, apoptotic hepatocytes showed a homogeneous distribution in the liver parenchyma, with moderate intensity in the evaluated areas. The cases of YF were related to the highest values of apoptotic processes of the arboviruses studied, emphasizing the midzonal area, presenting a very intense lesion. The periportal and centrilobular areas of YF samples showed intense and moderate apoptosis, respectively. From these data, it was possible to observe a clear predominance of the alteration in the cases of YF, especially in Z2 (Figure 6).

Apoptosis is a prominent histopathological feature of YFV infection, seen in hepatocytes, Kupffer cells, and endothelial cells. In YF, liver injury mechanisms are complex and involve the cytopathic effects inherent to viral replication and the development of the host's immune response to neutralize viral replication by activating, producing, and releasing cytokines, chemokines, and reactive oxygen species (ROS) and immune cells (Figure 11) [48,57,70,74,76,109,110].

The intense apoptotic event in hepatocytes in YF is also associated with the development of coagulopathy, a clinical manifestation of severe YF. Since hepatocytes are the main producers of most circulating clotting factors, the loss of these factors via hepatocyte destruction or some other mechanism can also result in an imbalance in the coagulation cascade, promoting hemorrhagic manifestation, a classic clinical feature in the fatal outcome of the disease [111]. The presence of apoptosis, lytic necrosis, swelling, and steatosis of hepatocytes is related to liver failure, a clinical manifestation observed in severe YF [112].

In the liver, two pathways of apoptosis were observed: extrinsic and intrinsic, which are dependent and independent of death receptors. Death receptors are the main mediators of apoptosis in the liver [113]. Enveloped viruses such as flavivirus and alphavirus are present in the outer layer of the host cell membrane derived from the endoplasmic reticulum (ER) and are mainly rich in phosphatidylserine (PS) and phosphatidylethanolamine (PE) (Figure 11) [114,115,116,117,118].

PS and PE are relevant components for interaction with the host's entry receptors and are important for the processes of tropism, pathogenicity, and viral infectivity. PS is mainly found in the outer membrane of apoptotic cells and is used as a “signal” to tissue macrophages, this mechanism triggered by PS by viruses as binding molecules trigger processes of “apoptotic mimicry” [119].

Regarding the hyperplasia and hypertrophy of Kupffer cells related to YF, DF, and CF cases. Data regarding hyperplasia showed a clear difference between arboviruses in terms of alteration. DF cases had a higher frequency of hyperplastic Kupffer cells, and an increase in the intensity of hyperplasia in the midzonal area of YF, DF, and CF cases was also identified. As for hypertrophy, Z1 and Z2 in the cases of YF, DF, and CF showed corresponding values with mild and moderate hypertrophy, respectively. In Z3, there was a difference between arboviruses; CF cases showed moderate cellular hypertrophy, while YF and DF cases showed mild hypertrophy. Unlike hyperplasia, in hypertrophy, CF cases showed a higher frequency of moderate intensity than YF and DF. Regarding the evaluated zones, the midzonal area was more affected, but in the case of CF, the midzonal and centrilobular areas were equally affected.

Kupffer cell hyperplasia and hypertrophy in YF have been widely documented in the scientific literature by several authors [7,100,120]. Hyperplasia and Kupffer cell destruction have been identified in most fatal DF cases [37,80,121,122]. Liver injury observed in hemorrhagic fevers, such as DF and YF, revealed that hepatocytes and Kupffer cells might be the target cells for viral replication [69,123]. Viral antigens were detected in Kupffer cells and hepatocytes from patients with DHF near damaged areas, suggesting an association between viral replication and liver damage [124]. Viral antigens were also detected in Kupffer cells in liver samples from fatal CF cases [34].

Three axes referring to hepatic endothelial alterations in the cases of YF, DF, and CF were evaluated. In the comparative analysis, SEA was predominant in cases of YF and CF, with moderate intensity in the three different zones of the hepatic acinus. In relation to DF, the midzonal area of the cases was more affected, with a moderate degree of alteration, while in Z1 and FD Z3, the change was slight. In sinusoidal congestion, CF cases did not show any change in any of the analyzed areas; therefore, the prevalence of congestion was in YF and DF cases, emphasizing the Z2 of YF cases. In contrast to congestion, in sinusoidal dilatation, we observed that the cases of YF and DF did not present any alteration in any of the analyzed zones; however, in the cases of CF, sinusoidal dilatation was moderate in all the zones, that is, there was no specific zone of involvement.

Viral antigens have been detected in HSEC and Kupffer cells in arbovirus infections [125,126,127]. In a study by Zellweger et al., a DENV-infected mouse model showed an intense HSEC infection. Local HSEC injury may result in changes in portal microcirculation, resulting in secondary ischemic injury and portal hypertension that may be more pronounced in some patients [37]. Reversal of portal flow has also been demonstrated in flavivirus infection, which, together with circulatory dysfunction of the splanchnic vascular bed, may explain several features of liver dysfunction [129].

Sinusoidal congestion has been characterized in YF infections [77]. Sinusoidal capillaries show intense circulation of macrophages/lymphocytes, and immune processes generate hepatocyte injury and necrosis of sinusoid capillaries due to the release of cytokines and activation of the complement system, as well as endothelial dysfunction [130]. Two other studies related this change to fatal DF cases, pointing to hemorrhagic congestion in the hepatic sinusoids [121,122]. Most patients have moderate-to-severe hepatic sinusoidal congestion [36,121,131].

In a study by Sharp et al., liver samples from fatal CF cases showed the presence of the viral antigen in vessels, Kupffer cells, HSEC, and portal connective tissue; leukocyte concentration was also observed within the vascular channels of the liver. The most frequently identified histopathological features were chronic portal hepatitis (mild), steatosis (mild), rare hepatocyte necrosis/apoptosis, and sinusoidal congestion with leukocytosis. A study by Agarwal et al. involving hepatic histopathological analysis of mice infected with CHIKV demonstrated the presence of dilatation of the hepatic sinusoids of these animals as a result of viral infection.

Of the alterations identified in the study referring to PT, the main alterations analyzed were moderate, with edema being equivalent to the three arboviruses studied (Figure 11), and in the alteration of the portal vein, the cases of DF were more intense. Congestion and alteration of the biliary canaliculus were mild and observed only in cases of YF and DF, and absent in cases of CF. However, the alteration of the portal artery was mild for the three arboviruses studied. In the study by Duarte-neto et al., microscopic liver analysis of YFV-infected individuals demonstrated the presence of edema in the PT and the space of Disse. None of the cases had bile duct injury.

Although the results obtained in the research related to histopathological alterations in DF, in general, were classified as mild to moderate, an explanation for these results would be the infectious serotype of the analyzed samples. Given that the study did not aim at the molecular characterization of the serotypes responsible for the infection, it is unknown whether the serotype of the samples examined was related to a different manifestation of DF, since the different clinical manifestations of the disease are also correlated with the DENV serotype. Patients infected with DENV-2 have a higher frequency of severe forms of the disease than those infected with other serotypes [132,133]. These patients manifest persistent vomiting, epigastric pain, plasma leakage, and shock [134].

According to Halsey et al., patients infected with DENV-3 showed a prevalence of musculoskeletal and gastrointestinal manifestations, whereas DENV-4 was associated with respiratory and cutaneous manifestations. Among the serotypes, DENV-2 was the most frequently isolated in cases of DHF and dengue shock syndrome, followed by DENV-3, DENV-1, and DENV-4 [28,136]. Another possibility is the individual aspects of each case, considering that the intrinsic characteristics of the host also influence the outcome of the disease.

As for the acinar and portal inflammatory infiltrates, our results were characterized by mild to moderate intensity, with the main cell types identified as lymphocytes and plasma cells. In YF and DF cases, the concentration of acinar inflammatory infiltrate was more evident in Z2 than in other zones of the hepatic acinus. Given the intense apoptotic process observed in the YF samples, the values referring to inflammation were disproportionate to the intensity of the liver injury.

Our data corroborate the previous findings of YF, in which the involvement of the liver parenchyma is disproportionate to the inflammatory findings. Furthermore, as previously mentioned, the pathogenic profile of YFV is correlated with the promotion of extensive apoptosis, a process linked to the non-induction of significant inflammatory response, thus explaining the scarcity of inflammatory infiltrates [47,48,57,137].

In cases of DF, as the processes of necrosis and apoptosis were generally associated with mild to moderate intensity, the limitation of inflammatory cells in the samples is predictable. Several studies have identified that the inflammatory infiltrates in YFV and DENV infection cases consist mainly of lymphocytes [26,69,76,138]. YF cases are characterized by the predominance of mild-to-moderate lymphocytes, both in the acinar area and in the PT [77,100,101,120].

Data from the two human studies collectively propose an immunometabolic basis for symptomatic YFV infection. The anabolic demands of infection can exacerbate cellular stress and trigger ROS responses, cell death, and inflammation [139,140]. According to Chan et al., immunometabolic responses support symptomatic YF and possibly other flavivirus infections.

Although CF is characterized by extensive inflammatory responses related to musculoskeletal injuries, hepatic inflammatory involvement was not intense in our study. Studies analyzing liver tissue from human cases and animal models infected with CHIKV have shown foci of mild inflammatory infiltrates composed of mixed inflammatory cells [68,108].

Infection promoted by flaviviruses and alphaviruses is based on the manipulation of the host's metabolism through the control of the survival of infected cells, thus affecting cell death pathways. Cell death pathways can be divided into two opposing processes: accidental cell death (ACD) and regulated cell death (RCD). While ACD is a consequence of severe and rapid injury (osmotic forces, pH variations, and lytic viral replication), RCD is based on regulated molecular machinery, implying modulation by pharmacological, genetic, or infectious components [142,143].

The different cell death processes have different morphological characteristics, with three main types: apoptosis, autophagy-dependent cell death, and necrosis. Apoptosis is characterized by chromatin condensation, nuclear fragmentation, cytoplasmic shrinkage, and the formation of “apoptotic bodies”, which are subsequently destroyed by regional phagocytes [143]. Necrosis is mainly characterized by swelling, rupture of the plasma membrane, and efflux of cytoplasmic content into the extracellular environment without evident phagocytosis or lysosomal degradation by neighboring cells [142].

Autophagy-dependent cell death is defined by the distinctive features of extensive vacuolization of the cytoplasm, leading to phagocytosis and degradation by lysosomes. The triggering of this process usually takes longer, allowing a longer period of viral proliferation before cell destruction. However, in a broader context, autophagy is associated with inhibition rather than induction of cell death and is considered a mechanism for maintaining cell homeostasis after stress signals (hypoxia, ROS, PRR activation, and others) [144,145].

The activation of other described mechanisms of non-apoptotic RCD is directly linked to viral load during infection. Necroptosis, which has morphological characteristics similar to necrosis, is an independent process of caspases [146]. Pyroptosis is a highly inflammatory form of lytic RCD that is dependent on the formation of pores in the plasma membrane by members of the gasdermin (GSDM), protein family which generally correlates with the activation of inflammatory caspases (CASP1, 4, or 5) [142,147]. These processes have been described in infections by intracellular pathogens, and are related to the induction of an antimicrobial response and involvement in viral pathogenesis [148,149].

Viral cytopathic mechanisms develop through the interaction between viral molecular components and cellular machinery. Flaviviruses and alphaviruses induce, activate, and inhibit several antiviral response mechanisms, including the development of different cell death processes, influencing the different morphological characteristics presented by cases of YFV, DENV, and CHIKV infection.

Finally, our results indicated differences in the patterns of histopathological alterations in human liver samples from fatal infections by YFV, DENV, and CHIKV, related to hepatic acinus, portal tract, and inflammatory infiltrate, with intensity variation among the studied arboviruses and predominance of lesions in certain areas of the liver parenchyma, resulting in tissue damage and liver dysfunction (Figure 12).

5. Conclusions

The cases of YFV, DENV, and CHIKV infections promoted different intensities of hepatic lesions in the analyzed hepatic acinus zones (Z1, Z2, and Z3). There was a predominance of alterations in the hepatic acinus among the analyzed areas, referring to the midzonal area (Z2) in YF, DF, and CF. Among the arboviruses studied, hepatic involvement in cases of YFV infection was more intense, characterizing the classic hepatic profile of the disease.

In the semiquantitative evaluation, high-intensity cell swelling, microvesicular steatosis, and apoptosis were observed, classified as intense to very intense. Regarding other histopathological changes evaluated in the hepatic acinus, macrogoticular steatosis, lytic/coagulative necrosis, and Kupffer cell hyperplasia/hypertrophy ranged from mild to moderate.

From the endothelial alterations in the hepatic acinus of the cases of YF, DF, and CF, there was a variation from mild to moderate, with the highlights of the cases of YF and CF presenting a higher frequency of moderate intensity. In PT, there was also a variation in intensity from mild to moderate, emphasizing edema and alteration of the portal vein, with all arboviruses presenting a moderate degree of edema. DF cases had relatively high values in portal vein alteration compared to YF and CF.

Infection by YFV, DENV, and CHIKV was accompanied by a mild-to-moderate inflammatory infiltrate consisting predominantly of lymphocytes and plasma cells, especially in cases of YF and DF, with a degree of intensity disproportionate to the involvement of the hepatic parenchyma. These data substantially contribute to understanding the pathogenesis of these arboviruses and the process of hepatotropic injury developed during infections, contributing to the improvement of diagnostic criteria for diseases and expanding our knowledge about the pathogenesis of YFV, DENV, and CHIKV in human liver samples.