The role of liver macrophages in viral liver pathogenesis

Abstract

Liver macrophages play important roles in the pathophysiology of liver fibrosis and hepatocellular carcinoma. However, liver macrophages are a heterogenous population and have differing roles in maintenance of liver function and response in disease. In a healthy liver, macrophages play a critical role in antigen processing, maintaining tolerance to the high levels of gut-derived bacterial products, and regulating inflammation through cytokine response. However, macrophages also play a critical role in liver pathology, specifically in the context of viral infection. The liver is targeted by multiple viruses, including human immunodeficiency virus, hepatitis B virus, and hepatitis C virus, which dysregulate macrophage functions to affect liver disease. Infection with any of these viruses is associated with increased risk of developing hepatocellular carcinoma, and coinfection further accelerates the progression to liver disease and cancer. However, the exact mechanisms by which liver macrophages contribute to disease in the context of viral infections are not well defined. This is a particularly acute issue in human immunodeficiency virus–infected populations, which have high incidence of hepatitis B virus and hepatitis C virus coinfection. To address this knowledge gap, this review describes the populations of macrophages in the liver, outlines the current models and limitations associated with the study of liver macrophages, discusses the function and role of liver macrophages in the context of viral infection, and describes the mechanisms by which these cells contribute to hepatocellular carcinoma and fibrosis. We then use this information to propose focus areas for the liver macrophage field to better address and resolve viral liver disease.

Infection of the liver by human immunodeficiency virus, hepatitis B virus, and/or hepatitis C virus impact the bidirectional molecular mechanisms that mediate the effects of macrophage and liver dysfunction in disease.

Key Concepts

• Many blood-borne viruses have profound effects on liver function and pathogenesis.

• Viral infections (human immunodeficiency virus, hepatitis B virus, hepatitis C virus) alter macrophage function, increasing risk of developing liver fibrosis and hepatocellular carcinoma.

• Both infected and virally exposed macrophages contribute to liver disease.

• Liver macrophages have diverse roles in liver function and disease pathogenesis.

• Liver macrophage diversity is poorly understood, as markers for distinct liver macrophage subpopulations and their response to disease are limited.

Open Questions

• What are the best markers and/or functions to more fully define liver macrophage populations?

• How do discrete liver macrophage populations contribute to viral liver disease?

• What are the bidirectional molecular mechanisms that mediate the effects of macrophage and liver dysfunction in disease?

• How can new models of liver dysfunction address the role of various macrophage populations?

1. Introduction

The liver is a functionally diverse organ that performs a variety of critical functions. One of the primary functions of the liver is the metabolism and removal of toxins and drugs in the blood received from the small intestine via the portal vein.¹ Other important liver functions include detoxification, aiding digestion through bile production, acute-phase protein synthesis, production of glucose, lipid metabolism, glycogen storage, and iron homeostasis.^1–3 During these processes, the liver is constantly exposed to gut-derived bacterial products with inflammatory potential, such as lipopolysaccharides (LPS).^1,2,4,5 As such, the liver must be highly tolerant to insult, while simultaneously ready to respond to dangerous toxins and chemicals for its role in detoxification. To fulfill these tasks, the liver has a number of cells that work in conjunction to maintain homeostasis, all of which can be altered in a disease state.³

The most common cell in the liver is the hepatocyte, the parenchymal cell of the liver. Hepatocytes comprise ∼80% of the liver volume and drive functions including metabolism and detoxification. These cells also play a key role in activating the innate immune system to maintain liver health.⁶ Hepatocytes produce bile from cholesterol to facilitate efficient digestion and absorption of dietary fats. In addition to regulating metabolism, bile acids can also function as ligands for nuclear receptors and cell surface G protein-coupled receptors to regulate the immune system.^3,6 Hepatocytes are the host cells for hepatitis B and C viruses (HBV and HCV, respectively), which can cause virus-mediated injury to the hepatocyte that is often caused by activation of an inflammatory response. Persistent inflammation can lead to the development of cirrhosis and liver disease.^7,8

In addition to the hepatocytes, there are a number of nonparenchymal cells (NPCs) in the liver. One of the most important of these is the hepatic stellate cells (HSCs). HSCs reside in the space of Disse and play a vital role in liver physiology. In the healthy liver, HSCs are the largest reservoir of vitamin A in the body.⁹ When the liver is injured, HSCs differentiate into fibroblast-like cells and produce extracellular matrix (ECM) to protect the liver from further damage. In chronic liver disease, overactive HSCs remodel the surrounding ECM, leading to reduced liver function due to increased stiffness of the liver microenvironment, eventually progressing to liver fibrosis.⁹ HSCs are susceptible to infection with human immunodeficiency virus type 1 (HIV-1) and can be activated in response to viral proteins from HIV-1, HBV, and HCV.^10–13 These cells are the main contributors to fibrosis in the liver through the production of deposition of collagen, resulting in liver stiffness and fibrosis—key steps in the development of hepatocellular carcinoma (HCC). Their role in HIV-1 driven liver disease will be discussed briefly later in the review. Kupffer cells (KCs), the resident tissue macrophages of the liver, are cells of the innate immune system that phagocytose bacteria have roles in initiation of inflammation as well as wound healing and are the main topic of this review.¹⁴ In addition to HSC and KC, other NPC include cholangiocytes and liver sinusoidal endothelial cells (LSECs). Cholangiocytes are epithelial cells that line the bile ducts of the biliary tree whose main function is to modify and transport hepatocyte-derived bile.¹⁵ These cells are the second most abundant type of cell in the liver. LSECs are highly fenestrated cells that comprise the microvasculature of the liver and separate the hepatocytes and HSCs from circulating blood cells. LSECs also have a role in maintaining HSC quiescence.

Although not often considered an immune organ, the liver contains a significant portion of the host macrophage population; over 80% of tissue resident macrophages are found in the liver, and hepatic macrophages comprise 20% to 35% of total liver NPCs.^14,16 At homeostasis, liver macrophages (LMs) are involved in the metabolism of iron and bilirubin, scavenging bacteria and damaged RBCs, monitoring tissue integrity, modulating the immune response, and are likely important in the pathogenesis of liver disease.^17–20 This review will describe how LMs contribute to the development of liver disease in the context of viral disease through the upregulation of proinflammatory and profibrotic mediators, as well as the activation of HSCs and contribution to the disruption of normal hepatocyte function. Critically, the different populations of LMs and how they contribute to cancer and fibrosis are difficult to define, and in this review, we attempt to differentiate the populations of these cells into specific groups: KCs, monocyte-derived macrophages (MDMs), and subgroups of macrophages from other sources—as the literature currently defines these as LMs, with little discussion on whether they should be distinguished by function. Common models used to study LMs will be discussed, and the importance of LMs in virally mediated liver disease, including the development of fibrosis and cancer, will be described. Due to the current state of the field, the remainder of this review will continue to define all macrophages found in the liver as LMs, especially in the context of human models, as the majority of the literature does not distinguish between the groups outlined above.

2. LMs at homeostasis

The majority of macrophages found in the liver are located in the periportal region of the liver sinusoids and are components of the vascular wall and therefore do not directly contact hepatocytes.^14,20–23 Instead, some LMs, especially those thought to be yolk-sac derived, canonically defined as KCs, are found on the layer of LSECs that separate hepatocytes from the blood cells via the space of Disse.^20,23

The focus of our review will largely be the immunological functions of LMs, but they also play a key role in many other functions associated with maintenance of liver function and homeostasis. Although hepatocytes play the primary role of iron storage and cholesterol synthesis in the liver,^6,15 they are not able to manage these products through phagocytosis—thus this function must be mediated by LMs. LMs play a critical role in removing aged and damaged RBCs, and when LMs are depleted, a significant reduction in serum iron levels occurs.^24,25 Iron homeostasis maintenance may also require recruitment of MDMs to the liver, indicating interplay between multiple populations of macrophages in the liver—including CLEC4F⁺ LMs that phagocytose desialylated platelets and are discussed in more depth later in the review, which is an example of how different populations of LMs may execute different functions.^26,27 Additionally, LMs phagocytose foreign particles, endotoxins, and cells undergoing apoptosis.^28–30 Thus, it can be summarized that some of the key roles involving LMs that are not necessarily associated specifically with immune function are related to phagocytosis and the presence of scavenger receptors, which will become important later when we seek to differentiate populations of LMs.

LMs are perhaps best known for their role in immune function, including antigen presentation and the activation and recruitment of CD8+ T cells, monocytes, and other immune cells.^21,31,32 Exposure to bacterial LPS results in increased NLRP3, tumor necrosis factor (TNF)-α, and other inflammatory markers in LMs, highlighting the importance of this cell type in immune surveillance.^33–38 The exposure of LMs to the outflow of the portal vein means that these cells are constantly exposed to microbial products, including LPS, released by gut microbiota. This causes the LMs to actively work to maintain a phenotypic balance between immune activation and tolerance.²³ LMs are typically thought of as having a more anti-inflammatory phenotype, involving secretion of interleukin (IL)-10 and transforming growth factor (TGF)-β and upregulation of scavenger receptors including MARCO and the aforementioned CLEC4F.^27,39,40

It is important to note that use of the binary proinflammatory and anti-inflammatory phenotypes does not fully reflect the breadth of macrophage function or the nuance associated the elaboration of distinct cytokines. While these phenotypes are often shortened to M1 (proinflammatory) or M2 (anti-inflammatory),⁴¹ more recent data have shown overlaps in these descriptors, and they are now often considered an oversimplification of series of more complex responses and phenotypes.^42–44 Indeed, LMs specifically have been shown to express both pro- and anti-inflammatory responses to distinct stimuli.^45–47 Despite this complexity, the binary pro-/anti-inflammatory simplification of macrophage activity remains in regular use because it is a useful way to describe the broad role for macrophages in pathogenic conditions and compare complex responses between diseases. For example, in cancer biology, which drove much of the initial definition of macrophages as M1 and M2, these definitions are valuable for associating clinical outcomes or tumor immunological activity with overall macrophage phenotype, while acknowledging the phenotype is more complex than binary comparison.^48,49

LMs, especially in mouse models, are often associate with a more tolerogenic and anti-inflammatory role of LMs, but this is not fully defined. Further, an “anti-inflammatory” phenotype in macrophages has also been linked to fibrosis through multiple mechanisms, and the cytokines and factors associated with an anti-inflammatory phenotype do not drive those outcomes. For example, secretion of the “anti-inflammatory” cytokine TGF-β both mediates tissue fibrosis and modulates the deposition of ECM.⁵⁰ And beyond its role in fibrosis and ECM regulation, TGF-β can promote Th17 cell development in the presence of IL-6,⁵¹ which can drive tissue inflammation. Similarly, the cytokine IL-10 is generally categorized as an anti-inflammatory cytokine that can suppress T-cell responses and release of IL-2, interferon (IFN)-γ, and TNF-α.^52,53 But IL-10 can also have immunostimulatory effects including roles in the development of memory T cells,⁵⁴ expansion of CD8+ T cells,^55,56 as well as some effects that are associated with antitumor immunity.⁵⁷ Thus, phenotypes defined as anti-inflammatory could also be profibrotic and drive liver disease, exemplifying the potential issues with the reductive nature of binary phenotypes. Despite these complexities, the bulk of the literature does define IL-10 and TGF-β as exhibiting primarily an anti-inflammatory function, as well as the promotion of a tolerogenic liver microenvironment, and this is true that the vast majority of the studies cited here use this definition. In the context of what we have sought to emphasize in this review, this role of promoting tolerance is especially important in the persistence of chronic viruses in the liver, including HBV, HCV, and HIV.^18,58–64 Thus, for the purposes of this manuscript, we opted to maintain the definition of these and other cytokines consistent with the literature being cited, but to add additional descriptors to better describe the role the cytokines were playing in that environment.

3. Macrophages and cancer

Macrophages play complex and often dual roles in cancer development. This may largely depend on their functional phenotype, which can shift between a proinflammatory (generally considered antitumor)⁶⁵ and anti-inflammatory (generally considered protumor)⁶⁶ state.⁶⁷ In many cancers, including those in the liver, macrophages adopt a tumor-associated phenotype, known as tumor-associated macrophages (TAMs). TAMs often exhibit characteristics commonly associated with macrophages that have anti-inflammatory function and promote tumor progression. This includes the production of mediators such as TGF-β, which promotes wound healing and fibrosis, along with the activation of other cells that can contribute to the development of disease such as HSCs in the context of liver cancer.^67–72 TAMs are also associated with markers such as CD163 and CD206, which generally are found on macrophages that produce anti-inflammatory cytokines, and are thought to promote cancer development.⁶⁷ Macrophages that are more tumor promoting” that secrete CD163 are also associated with significant fibrosis, especially in patients with viral hepatitis.⁷³ Further, macrophages expressing CCL18 and CD206 were positively correlated with increased tumor size, as well as showing infiltration in patients with advanced HCC.⁷⁴ TAMs also support angiogenesis by secreting growth factors like vascular endothelial growth factor (VEGF), which helps in the formation of new blood vessels, supplying the tumor with necessary oxygen and nutrients. These factors also protect the tumor by inhibiting cytotoxic T cells and natural killer cells, allowing cancer cells to evade immune detection and destruction, and facilitate metastasis by secreting enzymes such as matrix metalloproteinases (MMPs), which degrade the ECM and allow tumor cells to invade surrounding tissues.^66,75–79 TAMs can also promote an immunosuppressive microenvironment, specifically in the context of HBV-associated HCC.⁸⁰

Macrophages mediated chronic inflammation can also promote cancer development. Persistent inflammatory signals from macrophages can induce DNA damage through the release of reactive oxygen species (ROS) and nitrogen species, increasing the likelihood of mutations that drive tumorigenesis.^75,81–83 TAMs also play a role in helping cancer cells evade apoptosis by activating survival pathways such as NF-κB and STAT3 signaling. Moreover, macrophages can contribute to chemoresistance, enabling cancer cells to withstand treatment and repopulate after chemotherapy.^84–87 Specifically in the context of the lung and the liver, it appears that macrophages may drive the activation of cell types that promote tissue stiffness^88,89—which is considered a precursor to cancer development and will be elaborated on further down in this review.

The balance between macrophages that exhibit pro- and anti-inflammatory functions and the regulation of acute vs. chronic inflammation is central to the regulation of fibrosis and the development of HCC.^90–92 Fibrosis is a pathological response to wound healing resulting from the deposition of excessive amounts of ECM proteins, such as collagen. In the liver, this leads to a disruption in vascularity and an increase in stiffness that results in loss of function.⁶⁹ This effect is largely driven by HSCs, a major source of ECM proteins that can proliferate into myofibroblasts when activated.⁶⁹ HSCs can be activated by a number of processes, but many of them are initiated by LMs.⁶⁹ This response is generally broken up into stages—an initial inflammatory response that activates the HSCs, followed by a wound-healing fibrogenic response.⁶⁹

This process has been shown in a number of studies, which demonstrate the key role for LMs in fibrosis. In response to liver injury, macrophages can produce higher levels of IL-17A resulting in increased TNF-α, IL-6, and IL-1β. This increase in proinflammatory cytokine production was correlated with an increase in collagen 1 in cocultured HSCs, and the resulting fibrosis was abrogated in the absence of LMs.⁹³ In a murine model of induced hepatic fibrosis, LMs produced proinflammatory TNF-α and activated HSCs that produced profibrotic factors such as α-SMA and collagen 1, leading to significant liver fibrosis and hepatic injury.⁹⁴ The induction of TNF-α in LMs is also important because TNF-α exposure enhances macrophage expression of TIMP-1, a marker that is commonly correlated with increased liver fibrosis when expressed on both macrophages and hepatic myofibroblasts.^95,96 Coculture models also show that HSCs cultured with media from LMs incubated with leptin, which has profibrotic effects on the liver, show increased expression of profibrotic proteins collagen I, TIMP1, TGF-β1, and α-SMA, as well as increased motility and proliferation.⁹⁷ Together, these findings highlight the delicate interplay between proinflammatory and anti-inflammatory roles of LMs, as well as how these cells may affect the surrounding environment, particularly in regard to HSCs in order to promote fibrosis.

Given the central role of fibrosis in the development of HCC, it is not surprising that LMs also play a key role in the development and regulation of this disease. LMs can promote tumor development through expression of NADPH oxidase 1 and MMP-2, as well as the mechanisms that lead to fibrosis.^98,99 Furthermore, the increased presence of macrophages exhibiting anti-inflammatory function has been shown to be associated with poor prognosis in HCC patients.^100,101 This may be the result of a few different mechanisms, including increased recruitment of immune cells, dysregulation of angiogenesis by macrophages, and the promotion of HCC proliferation and metastasis.^{100,102–104} Despite the current literature showing strong evidence for the role of LMs in the development of fibrosis and HCC, the understanding of how viral disease may play into this is still not well understood and warrants further study. In the next section, the review will focus on the viruses most commonly associated with the development of HCC. These are HBV and HCV, which are the primary cause of HCC and result in over 1 million deaths per year, globally.¹⁰⁵ Additionally, these infections are commonly comorbid with HIV, which on its own can result in an increased risk of developing liver disease.¹⁰⁶ Research on how these viruses affect liver pathogenesis largely focuses on the effect of specific proteins, which is summarized in Table 1, and the overall effects are summarized in Table 2 and further elaborated upon in the section below.

Table 1.

Specific effects of viral proteins on liver cell function, including hepatocytes, HSCs, and LMs.

Virus	Protein	Function	Effect	Cell type	Citations
HIV	GP120	Surface protein on HIV	Apoptosis	Hepatocytes	Vlahakis et al. 2003107
			Apoptosis	Hepatocytes	Rizza et al. 2011108
			HSC migration	HSCs	Del Corno et al. 201610
			Cytokine dysregulation (CCL2, CXCL8)	HSCs	Del Corno et al. 201610
			ERK 1/2 phosphorylation	HSCs	Hong et al. 201211
			cytokine dysregulation (a-SMA, collagen 1)	HSCs	Hong et al. 201211
			Cytokine dysregulation (monocyte chemotactic protein-1 [MCP-1], IL-6, ferroportin [MTP-1])	HSCs	Bruno et al. 2010109
			HSC migration	HSCs	Bruno et al. 2010109
			Cytokine dysregulation (TNF-a)	LM	Joshi et al. 2011110
			Cytokine dysregulation (CCL2, CCL4, CXCL8, IL-1B)	LM	Del Corno et al. 201610
			Interaction with TLR4	LM	Del Corno et al. 201610
			Cytokine dysregulation (MIP-1a, RANTES)	LM	Bautista et al. 2002111
			Activation of AKT and ERK signaling (hypoxia indusible factor 1a [HIF-1a], TGF-B1)	LMs, hepatocytes, HSCs	Xu et al. 2024112
	p24	Core protein of HIV	Reduced phagocytic activity	Macrophage	Kedzierska et al. 2003113
	Tat	Transactivator	Cytokine dysregulation (CXCL10)	LM	Qu et al. 2012114
			Cytokine dysregulation (TNF-a)	LM	Joshi et al. 2011110
			Inhibition of phagocytosis	Macrophage	Debaisieux et al. 2015115
			Cytokine dysregulation (IL-1B, IL-6, TNF-a, MCP-1, ICAM-1 [CD54], vascular cell adhesion molecule 1 [VCAM-1], CCR2, CCR5)	Macrophage	Meng et al. 2022116
	Nef	Regulatory protein	Impair macrophage activity (oxidative burst, phagocytosis)	Macrophage	Olivetta et al. 2014117
			Cytokine dysregulation (TNF-a)	Macrophage	Olivetta et al. 2014117
			CD36 downregulation	Macrophage	Olivetta et al. 2014117
			Increased tissue infiltration by macropahges	Macrophage	Verollet et al. 2015118
			Resistance to apoptosis	Macrophage	Abbas et al. 2014119
	Vpu		Cytokine dysregulation (alpha-smooth muscle actin [aSMA-1] COL-1, VEGF)	HSCs	Patel et al. 201412
			Cytokine dysregulation (TGF-B)	LM	Patel et al. 201412
HBV	HBx	Multifunctional regulatory protein	Macrophage polarization	Macrophage	Lan et al. 2022120
			Suppression of apoptosis	Hepatocytes	Li et al. 2003121
			Increased proliferation and migration	HSCs	Chen et al. 2014122
			Cytokine dysregulation (TGF-B, aSMA, Col1)	HSCs	Chen et al. 2014122
			Cytokine dysregulation (TGF-B, aSMA, Col1, TIMP-1)	HSCs	Gong et al. 2016123
			Cytokine dysregulation (TGF-B, connective growth tissue factor [CTGF])	HSCs	Guo et al. 2009124
	HBsAg	Surface protein of HBV	Macrophage polarization (M2)	Macrophage	Li et al. 2022125–129
			Cytokine dysregulation (IL-6, TNF-a, IL-10)	LM	Boltjes et al. 2015130,131
			Induction of NK cells	LM	Boltjes et al. 2015130,131, Hou et al. 2017132
			Cytokine dysregulation (IL-12)	LM	Hou et al. 2017132
			Cytokine dysregulation (IL-10, TGF-B)	Macrophage	Xiang et al. 2022133
	HBeAg	Core protein of HBV	Cytokine dysregulation (IL-10, TGF-B)	Macrophage	Xiang et al. 2022133
			Inhibition of NF-kB signaling, ROS	Liver macropahge	Yu et al. 2017134
			Cytokine dysregulation (reduction in IL-1B production)	Liver macropahge	Yu et al. 2017134
			Proliferation, motility, contraction	HSCs	Xie et al. 2021135
			Cytokine dysregulation (CCL2, CCL5, CXCL10, TNF-a)	HSCs	Xie et al. 2021135
			Cytokine dysregulation (TNF-a, IL-6, IFN-B)	Macrophage	Li et al. 2022125–129
			Cytokine dysregulation (IL-6, TNF-a)	Macrophage	Tian et al. 2024136
HCV	Core		Cytokine dysregulation (TGF-B)	Hepatocytes	Benzoubir et al. 2013137
			Resistance to ROS-mediated apoptosis	Hepatocytes	Severi et al. 2007138
			Cytokine dysregulation (CCL2, CXCL10)	Macrophage	Song et al. 202174
			Inhibition of TLR2 activation	Macrophage	Song et al. 202174
			Cytokine dysregulation (TGF-B, CTGF, aSMA, COl1)	Hepatocytes	Shin et al. 2005139
	NS3/4A protease	Cleaves capsid to form mature viral proteins	Cytokine dysregulation (IL-28)	Hepatocytes	Ding et al. 2012140
			Resistance to ROS mediated apoptosis	Hepatocytes	Rios-Ocampo et al. 2019141
			Cytokine dysregulation early in infection (CCL2, TNF-a)	Macrophage	Bansal et al. 2015142
	NS5A		Cytokine dysregulation (IFN-B, TNF-a, IL-18)	Macrophage	Sepulveda-Crespo et al. 2021143
			Stimulation of monocytes	Macrophage	Li et al. 2005144
			Induction of ROS	Hepatocytes	Sene et al. 2010145
	E1/ E2		Cytokine dysregulation (IL-10, CD163)	Macrophages	Vijayamahantesh et al. 2022146
			Induction of anti-inflammatory macrophage polarization (CCL13, CXCL12, IL-10)	Macrophage	Kwon et al. 2019147
	NS2		Cytokine dysregulation (TGF-B, aSMA, Col1, TIMP-1)	HSCs	Li et al. 202330

The literature was reviewed for papers that used isolated proteins from HIV, HCV, and HBV in order to determine the effects of these proteins on liver cells. These effects are described, along with the cell type in which they were observed and arranged by virus and protein.

Table 2.

The overall effects of viral monoinfection and coinfection on hepatocytes, HSCs, and LMs.

Virus	Cell type	Effect	Citations
HBV	Hepatocytes	Transformation	Rawat et al. 2015148
HBV	HSCs	Cytokine dysregulation (CXCR3)	Yuan et al. 2021149
HBV	HSCs	Suppresion of apoptosis	Yuan et al. 2021149
HBV	HSCs	Increase motility, growth	Yuan et al. 2021149
HBV	Macrophage	Increased ROS	Xie et al. 2021135
HBV	HSCs	Cytokine dysregulation (aSMA, col1A)	Huang et al. 2015150
HBV	Hepatocytes	Cytokine dysregulation (TGFB, PDGF)	You et al. 2023151
HBV	HSCs	Cytokine dysregulation (Col1A1)	Li et al. 2022125–129
HBV	HSCs	Cytokine dysregulation (aSMA, PDGF)	Bai et al. 2012152
HBV	Macrophage	Suppression of macrophage function	Bei et al. 2023153
HBV	Macrophage	Cytokine dysregulation (TGF-B, IL-10, TNF-a, IL-1B)	Bei et al. 2023153
HCV	Hepatocytes	Cytokine dysregulation (CXCL10)	Brownell et al. 2013154
HCV	Macrophage	Cytokine dysregulation (CD206, CD163, iNOS, hARG1)—M2	Bility et al. 2016155
HCV	Macrophage	Cytokine dysregulation (TIMP1, COL1A)	Bility et al. 2016155
HCV	Macrophage	Cytokine dysreglation (decreased IL-12, IL-10)	Bility et al. 2016155
HCV	Macrophage	Cytokine dysregulation (CCL5)	Sasaki et al. 2017156
HCV	HSCs	Cytokine dysregulation (IL1B, IL-6, CCL5, TGFV1, Col4a, MMP, aSMA)	Sasaki et al. 2017156
HCV	HSCs	Cytokine dysregulation (aSMA)	Cheng et al. 201761
HCV	HSCs	Differentiation of HSCs	Kim et al. 2019157
HCV	HSCs	Cytokine dysregulation (TGF-B)	Kim et al. 2019157
HCV	Hepatocyotes	Apoptosis (Fas, BIK, p53)	Silberstein et al. 2016158
HCV	Hepatocytes	Cytokine dysregulation (TGF-B)	Hall et al. 2010159
HCV	HSCs	Cytokine dysregulation (TGF-B, IL-6, TNFa)	Cevallos et al. 2024160
HCV	Macrophage	Inflammasome activation	Chattergoon et al. 2014161
HCV	Macrophage	TREM-1 upregulation	Hyun et al. 2019162
HCV	Macrophage	Increased macrophage population	Tapia-Abellán et al. 2012163
HCV	Macrophage	Cytokine dysregulation (increased IL-10, decreased IL-6, IL-12)	Tapia-Abellán et al. 2012163
HCV	Hepatocytes	Increased epithelial mesenchymal transition	Bose et al. 2012164
HIV	Monocytes	Cytokine dysregulation (IL-2, TGF-B, CD25, aSMA, collagen)	Cevallos et al. 2024160
HIV	Macrophage	Inflammasome activation	Chattergoon et al. 2014161
HIV	Hepatocytes	Hepatocyte depletion	Dagur et al. 2018165
HIV	HSCs	Cytokine dysregulation (Col1A, TIMP1)	Lin et al. 2011166
HIV	HSCs	ROS production	Lin et al. 2011166
HIV	HSCs	ROS production	New-Aaron et al. 2022167
HIV	HSCs	Cytokine dysregulation (IL-6)	New-Aaron et al. 2022167
HIV	Hepatocytes	Upregulation of TRAIL R2, reduction in p53 function	Babu et al. 2009168
HIV	Hepatocytes	Apoptosis: mitochondrial dysfunction, ROS	Gruevska et al. 2021169
HIV	Macrophage	TREM-1 upregulation	Hyun et al. 2019162
HIV + HBV	Hepatocytes	Cytokine dysregulation (CXCL10, CXCR3)	Singh et al. 2020170
HIV + HBV	Hepatocytes and HSCs	Cytokine dysregulation (TGF-B)	Xu et al. 2024112
HIV + HBV	Hepatocytes	Persistence of HIV DNA in hepatocytoes of HBV+ individuals on ART	Zerbato et al. 2023171
HIV + HCV	HSCs	Increased proliferation	Akil et al. 2019172
HIV + HCV	HSCs	Cytokine dysregulation (CCL2, IL1A, IL1B, IL13RA2, MMP1)	Akil et al. 2019172
HIV + HCV	Coculture HSCs, Monocytes	Cytokine dysregulation (collagen)	Cevallos et al. 2024160
HIV + HCV	HSCs	Cytokine dysregulation (Col1A, TIMP1)	Lin et al. 2011166
HIV + HCV	HSCs	ROS production	Lin et al. 2011166
HIV + HCV	Hepatocytes	HCV upregulates HIV LTR activation and gene transcription	Sengupta et al. 2013173
HIV + HCV	Hepatocytes	Infection with HIV leads to increased HCV gene expression	Kong et al. 2014174
HIV + HCV	Macrophage	TREM-1 upregulation	Hyun et al. 2019162
HIV + HCV	HSCs	Increased proliferation and invasion	Shi et al. 2016175
HIV + HCV	HSCs	Cytokine dysregulation (TIMP1)	Shi et al. 2016175(#1367)

This summary table provides an overview of how viral monoinfection and coinfection affect hepatocytes, HSCs, and LMs, highlighting how coinfection generally worsens the impact on cell function and overall liver health.

4. LMs and viruses associated with HCC

4.1. Hepatitis B virus

HBV infects more than 2 billion people worldwide and is one of the primary causes of liver cancer. This virus is a member of the Hepadnavirus family and is a hepatotropic virus that primarily infects human hepatocytes,⁷ although HBV has also been shown to exist in extrahepatic reservoirs, such as lymphocytes and peripheral blood mononuclear cells. Thus, the primary interaction between macrophages and HBV is through exposure to HBV proteins secreted from hepatocytes, which is the focus of the remainder of this section.^176,177

In patients with a chronic HBV infection (CHB), as well as mouse models of CHB, macrophages are largely polarized toward an anti-inflammatory phenotype with a consequent reduction in proinflammatory cytokine secretion. LMs isolated from patients with CHB display upregulated expression and secretion of anti-inflammatory cytokines such as IL-10, CD163, and TGF-β, while expression and secretion of proinflammatory cytokines such as IL-1β, IL-6, IL-12, and TNF-α were decreased.^{125,134,178–180} Supporting the anti-inflammatory impact of HBV, when MDMs from healthy donors were differentiated to an proinflammatory phenotype and exposed to CHB patient-derived HBV antigens, expression of IL-1β, IL-6, and TNF-α was significantly decreased.^{126,134,178,181} As IL-1β has been shown to be sufficient to inhibit HBV replication, the inhibition of its expression may be a mechanism by which HBV modulates the immune system to promote chronicity.^126,178 Indeed, proinflammatory MDM prevented HBV infection in primary human hepatocytes when cocultured. However, when MDMs displaying a proinflammatory phenotype were pretreated with supernatant from HBV-expressing cells during differentiation, they expressed less IL-1β and IL-6 and were unable to prevent infection of primary human hepatocytes by HBV.¹⁷⁸ There are several different HBV proteins associated with the modulation of macrophage inflammation and numerous focus on the impact of specific HBV antigens.

The HBe antigen (HBeAg) is thought to modulate the host immune system to prevent a robust immune response and aid in the chronicity of an HBV infection.^182,183 HBeAg may accomplish this through the inhibition of the NLRP3 inflammasome and expression of IL-1β via downregulation of nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signaling.^125,126,181 In both samples collected from participants and in in vitro assays, the loss of HBeAg expression reduces, but does not eliminate, the inhibition of NLRP3 activation and IL-1β during an HBV infection, indicating that HBeAg is not the sole factor regulating this facet of HBV biology.¹³⁴ However, several studies found that treatment with recombinant HBeAg activated NF-κB signaling, increased activation and increased production of TNF-α and IL-6 in macrophages,^135,136 as well as possibly inducing apoptosis in these cells.¹⁸⁴ Taken together, data suggest that HBeAg can both induce an immunotolerant phenotype in macrophages, specifically in the modulation of inflammasomes, as well as drive proinfalmmatory activity.

The HBcAg is the capsid protein of HBV and is largely considered immunogenic during an HBV infection.^183,185,186 The immunogenicity of HBcAg is tied to its arginine-rich C-terminal domain, which upregulated IL-6, IL-12p40, and TNF-α expression in a Toll-like receptor (TLR)-2- and Heparan Sulfate ProteoGlycan-dependent manner in macrophages.¹⁸⁵ THP-1 cells, a monocytic cell line, exposed to HBcAg also displayed upregulated IL-1β expression resulting in the inhibition of HBV replication when cocultured with an HBV-expressing cell type.¹²⁶ Together this indicates that HBcAg has an overall activating effect on macrophage function.

While the role of HBsAg in regulating host immunity is not well understood, patients who had undergone HBsAg seroconversion, ie had undetectable levels of HBsAg in the blood, also had significantly lower levels of secreted CD163 (sCD163) compared with patients who had not achieved seroconversion. High levels of sCD163 correlated with more severe inflammation and fibrosis.¹⁸⁷ However, because seroconversion of HBsAg is a rare event, indicating successful immune control of a CHB, and HBsAg⁻ patients would also be HBeAg⁻, the implications of this study for the role of HBsAg in immune regulation are somewhat limited. LMs cultured in vitro and exposed to HBsAg show an increased expression of IL-1β, IL-6, TNF-α, and CXCL8 and increased staining for CD40⁺/CD80⁺/CD86⁻. Additionally, HBsAg-treated LMs can increase the percentage of CD56⁺/CD69⁺ NK cells and the expression of IFN-γ by NK cells.^126,130 In vitro, MDM activated by LPS exposure and treated with HBsAg showed decreased expression of IL-6 and IL-1β.¹⁸¹ Additionally, both an in vivo mouse model and coculture of macrophages and HBV-expressing HepG2.2.15 cells showed that macrophages displaying proinflammatory markers had decreased expression of iNOS and TNF-α, while macrophages displaying anti-inflammatory markers showed increased expression of IL-10 and Arg1.¹²⁷ Overall, much like the role of HBeAg in regulating host immune response, it is likely that HBsAg acts in a complimentary manner with other HBV antigens to promote immunotolerance and chronicity during an HBV infection.

Modulation of macrophage phenotype by HBV would have functional consequences for the activation of additional immune system components such as B and T cells during an HBV infection. Said et al.¹⁸² found that there was a significant increase in the number of CD68⁺/CD86⁺ macrophages compared with CD68⁺/CD80⁺ or CD68⁺/PDL1⁺ macrophages in liver samples from patients with CHB. An increased ratio of CD86⁺/CD80⁺ macrophages would be expected to skew the activation of T cells to a T helper-type 2 (Th2) response, which promotes immunotolerance. During CHB, aberrant activation of macrophages associated with anti-inflammatory function and the Th2 response could lead to the upregulated expression and secretion of profibrotic cytokines such as TGF-β, platelet-derived growth factor (PDGF), VEGF, insulin-like growth factor-1, and Galactin-3, thereby contributing to the development of liver fibrosis and potentially HCC.^188–191

4.2. Hepatitis C virus

HCV is another hepatotropic virus that infects more than 58 million people globally and is another leading cause of liver cancer. This Flavivirus primarily infects hepatocytes but unlike HBV, it has also been reported to infect macrophages, microglia, and lymphocytes.^192–195 Macrophage infection is hypothesized to occur primarily via uptake of infectious viral particles and interaction with scavenger receptors.^{143,196–199} Although the rate of infection of nonhepatic cell types by HCV is low during chronic HCV infection, the interaction between HCV and the host immune system is critical to maintaining the chronicity of an HCV infection and promoting pathogenesis.^{131,191,194,200–202} The ability of HCV to infect immune cells means that both secreted and nonsecreted proteins that remain intracellular, specifically the nonstructural proteins NS2, NS3, NS4A, NS4B, NS5A, and NS5B, have the potential to modulate the host immune system during an HCV infection.^203–205 Despite this possibility, much of the immunomodulatory activity connected to HCV infection is associated secreted proteins including Core, envelope (E1 and E2), and NS1 and involve TLR signaling. The HCV core antigen (HCcAg) can induce TLR2 heterodimerization with TLR1 or TLR6 and upregulate the expression of the inflammatory modulators IL-8, IL-10, IL-1β, PD-L1, and TNF-α.^{143,202,206–209} HCcAg can also induce tolerance to inhibit subsequent TLR2 activation and induce cross-tolerance to additional agonists for TLR2 and TLR4.²¹⁰ During chronic HCV infection, induction of cross-tolerance to TLR2/4 by HCcAg also impaired IL-6 and IL-17 expression as a result of NF-κB inhibition.²¹⁰ HCcAg can also to inhibit the IFN response during a chronic HCV infection via TLR3 binding, inhibiting the response to double-stranded RNA.²⁰⁶ HCcAg-mediated inhibition of TLR3-mediated antiviral activity may also enable persistence of an HCV infection in LMs by preventing NLRP3 inflammasome activation, which would otherwise trigger upon recognition of double-stranded HCV RNA.²⁰¹ In addition to the HCcAg, the HCV E2 protein can also drive inflammation by upregulating IL-1β, IL-6, IL-10, and TNF-α via binding to TLR2/4.

The HCV nonstructural proteins may also alter immune signaling via both intracellular and extracellular interaction. Although its secretion has not been confirmed, extracellular NS3 displays a TGF-β-mimetic activity and is able to bind TGF-β type 1 receptor (TβR1),²⁰¹ driving fibrosis²¹¹. This interaction is enhanced by TNF-α, which is upregulated during HCV infection. Similar to HCcAg, extracellular NS3 is also capable of binding to TLR2 to induce heterodimerization with TLR1/6 and upregulate the expression of IL-8, IL-10, and TNF-α.^143,208 By upregulating TβR1-induced signaling cascades and inhibiting the TLR3-mediated antiviral response NS3 is able to promote immunotolerance and an early anti-inflammatory effect with the long-term consequence of promoting fibrosis as seen in chronic HCV infections.^202,212,213 The inhibitory effects of NS3 on expression of IFN-α/β/γ and the chemokines CCL-5, CXCL-8, and CXCL-10 in macrophages can be enhanced by the NS4A protein.²¹⁴ The NS3/4A complex can also inhibit innate immunity by blocking the antiviral response generated by RIG1 upon detection of dsRNA,²¹⁵ and in vitro studies show that NS3/4A expression ablated RIG1-mediated IRF3 and NF-κB activation to inhibit IFN-β secretion.^216,217

The NS5A protein can bind to TLRs, in this case TLR4, to upregulate IFN-β, TNF-α, and IL-18.¹⁴³ Upregulation of TLR4-mediated signaling has been observed in HCV-infected patients, indicating that NS5A may have a proinflammatory/profibrotic effect during an HCV infection. NS5A stimulates monocytes by binding to TLR4 to initiate p38- and PI3K-dependent IL-10 production and inhibit IL-12 production,¹⁴⁴ driving monocyte secretion of TGF-β and inhibiting NKG2D expression on circulating NK cells to prevent IFN-γ production and destruction of HCV-infected cells.¹⁴⁵ NS5A can also upregulate NF-κB and STAT-3 activation via induction of ROS and upregulation of Ca²⁺ signaling, increasing inflammation,²¹⁸ as well as inhibiting protein kinase R (PKR) dimerization and subsequent PKR-mediated eukaryotic initiation factor-2a phosphorylation in HCV-infected cells, blocking downstream IFN signaling.²⁰⁵ Overall, the potential for an NS5A-mediated proinflammatory/profibrotic effect during an HCV infection is likely offset by the immunoregulatory activity of NS5A that generates a tolerogenic response. Taken together, these studies demonstrate that both secreted and nonsecreted HCV proteins modulate a variety of innate immune functions to sustain and advance viral infection and associated liver pathology.

4.3. Human immunodeficiency virus

Globally, more than 40 million people are living with human immunodeficiency type 1 (HIV-1), and about 1.3 million people are newly infected annually.²¹⁹ HIV-1 is a lentivirus within the family retroviridae. The viral genome consists of 2 ssRNA molecules enclosed within a virus particle comprised of a number of different proteins, many of which can be secreted and have systemic effects in the context of chronic infection, including the nonstructural proteins Tat, Rev, Nef, Vif, Vpr, and envelope protein gp120.²²⁰ HIV-1 infects cells that express the CD4 receptor and also either a CCR5 or CXCR4 coreceptor, which includes monocytes and macrophages, dendritic cells, and CD4+ T cells. HIV pathogenesis generally proceeds through destruction of CD4+ T cells and the development of opportunistic infections, tissue damage mediated by HIV-1, and/or systemic damage as a result of immune dysfunction or a combination of these effects.²¹⁹ While there are safe and effective antiviral therapies (ARTs) for HIV-1, these are noncurative and despite treatment, people with HIV (PWH) have an increased risk of developing liver-related diseases, including HCC.^221–224

Compared with the general population, PWH have a 2.79-fold greater risk for HCC, develop HCC at a younger age,²²⁵ have lower median survival times, more advanced liver disease at the time of identification, and higher lymphocyte infiltration in the tumors.²²¹ In the current era, liver-related disease is a leading cause of death for PWH, but the exact mechanisms by which HIV drives this are unclear.⁴ Studies have hypothesized a role for chronic inflammation, cytokine dysregulation, and/or continued exposure to viral proteins from HIV-infected cells both within and outside the liver, as current therapies do not block the release of viral proteins.^226,227 In untreated PWH, both HIV RNA and p24Gag, the HIV-1 capsid protein, are present in LMs, monocytes recruited to the liver, HSCs, and in a few cases, hepatocytes.^228–230 HIV DNA can also be detected via digital droplet polymerase chain reaction (PCR) in more than 50% of sampled liver tissues in ART-compliant PWH.^231,232 HIV RNA and viral proteins have also been detected in the liver and the LMs of simian immunodeficiency virus (SIV)-infected rhesus monkeys.^233–237 And there are also conflicting data as to whether LM can be productively infected with HIV-1. Some groups report no replication competent virus is released postinfection, while others reporting that this does occur in primary human cells and SIV models.^{163,233,235,237–245} Further, despite the detectable presence of HIV-1 DNA and RNA in macrophages, the mechanisms by which LMs contribute to HIV-1-associated disease in liver remain unclear.

One possible mechanism is chronic immune activation in response to persistent infection. During HIV-1 infection, tissue-resident macrophages in the lung, spleen, brain, and other cells of the myeloid lineage are thought act as viral reservoirs of HIV-1.²⁴³ If LM are indeed productively infected with HIV-1, they may also act as reservoirs, although this remains undetermined and the effects of HIV-1 infection on LM populations are still being defined. Some studies report that human LMs are depleted during infection,²⁴¹ while others observe a transient increase in the number of LMs in nonhuman primate (NHP) livers.²³³ This change in LM population is usually associated with the presence of inflammatory mediators such as TNF-α, CCL3, and profibrotic factor TGF-β.²³⁵ Chronic inflammation and production of inflammatory mediators is a hallmark of HIV infection and may be a result of immune dysregulation, exposure to viral proteins²⁴⁶ or virally mediated changes in macrophage phenotype or function. Indeed, HIV-1-infected LMs showed increased expression of TLR4 and CD14, indicating that HIV-1 infection may predispose LMs to a more robust inflammatory response.²⁴⁴ Under homeostatic conditions, LM can tolerate high levels of gut LPS, but during HIV-1 infection, gut microbial translocation is increased and infected LMs are more sensitized to LPS stimulation, releasing more proinflammatory cytokines such as IL-6 and TNF-α due to constant exposure to LPS and other bacterial products from the gut.^244,247

During homeostasis, LMs tolerate repeated interactions with LPS in an NF-κB-dependent manner.^60,248 However, exposure to HIV-1 proteins increases NF-κB activity on innate immune cells, specifically macrophages,²⁴⁹ potentially through the upregulation of triggering receptor expression of myeloid cells (TREM1). Increased expression of TREM1 on LMs can amplify inflammatory responses involved in the development of HCC and promote liver injury and fibrosis, and TREM1 levels on LMs are increased in response to infection with HIV-1.^162,250 The HIV-1 proteins Pro and Rev upregulate the expression of TREM1 mRNA through the NF-κB signaling pathway, and inhibition of TREM1 resulted in an abrogated inflammatory immune response after HIV stimulation, although the mechanism through which this occurs is not exactly defined.¹⁶² Additionally, TREM1 induction in the context of infection with HIV-1, and specifically with exposure to Vpr, Tat, or gp120, has been shown to prolong macrophage survival.²⁵¹

This constant immune activation, along with the presence of ROS produced by activated LMs, results in the activation of HSC,⁹ possibly contributing to liver damage and fibrosis and leading to the development of HCC. In a transgenic rat model of HIV-1 in which HIV-1 rats were genetically modified to contain the entire HIV-1 genome, HIV-1 rats showed increased hepatic steatosis and liver inflammation as well as increased macrophage infiltration in the liver.²⁵² Furthermore, treatment with HIV-1 envelope protein gp120 seems to effect LMs in both mouse and human primary cell models, upregulating the chemokines CXCL1 and CCL5,¹¹¹ with a TLR4-dependent increase in CCL2, CCL4, CXCL8 and IL-1 β.¹⁰ These studies suggest that LMs are influenced by HIV-1 proteins, especially through TREM1, but the link between these effects and an increased risk of developing HCC requires further study.

While many of the effects of HIV on liver disease are driven directly through the impact of the virus on LM, the indirect impact of HIV-LM interactions are also critical to the disease process. For example, HSCs can be activated by LMs during infection with HIV-1.^10,253 These cells can actively contribute to liver fibrosis and the development of HCC through the production of ECM and profibrotic cytokines.^1,9,175,254 HSCs produce these profibrotic factors, such as TGF-β, because of activation, which occurs in response to oxidative stress, innate immune signaling through TLR4, and proinflammatory cytokines such as IL-20.²⁵⁵ Indeed, fibrotic gene expression, contract, proliferation and other morphological changes associated with fibrogenesis have been observed in the human LX-2 hepatic cell line in the presence of HIV-1 infection.¹¹ Infection of LX-2 cells with an X4 tropic strain of HIV-1 increased expression of proinflammatory cytokines, as well as collagen and the HIV-1 coreceptor CXCR4,²⁵⁶ and although infection with a R5 tropic strain did not induce expression of TGF-β, collagen, or α-SMA, it did result in an increase in IL-6 and a loss in lipid droplets, which have been shown to drive fibrosis.²⁵⁷ Further coculture studies also showed that LX-2 cells incubated with HIV-infected Jurkat cells resulted in an increase in ECM proteins and profibrotic cytokines.¹⁶⁰ Collagen expression also increased in response to the HIV-1 envelope glycoprotein gp120 and matrix protein p17.^11,13,257 Expression of TGF-β is also increased in LX-2 coculture with U1 monocytic cells stably expressing the HIV-1 accessory protein Vpu.¹² In summary, these studies suggest a role for HSCs in developing HIV-associated liver disease. Still, the exact role of these cells in HIV/HBV coinfection-mediated liver injury requires further investigation, especially in the context of how LMs may activate HSCs to become more fibrotic and proliferative; encouraging a microenvironment that may be more friendly to the development of HCC.

4.4. Viral liver pathogenesis

The pathogenic impact of viral infection on the liver, as outlined above and in Tables 1 and 2, most likely drives the development of HCC and liver disease (Fig. 1). These effects are most likely mediated via direct effects of infection and viral proteins on LMs, hepatocytes, and HSCs, as well as through intercellular communication between these cells that is dysregulated by viral infection. LMs are a heterogeneous population that includes both canonical KCs, the resident tissue macrophages in the liver, and MDMs, which traffick into the liver from the blood. These 2 populations play distinct roles within the liver, although they both likely play roles in the regulation of inflammation and fibrosis. KCs are primarily localized in the sinusoids and serve as the first line of defense against pathogens, whereas MDMs trafficked into the liver upon injury.²⁵⁸ The involvement of both populations is critical to understanding how viral infections dysregulate the liver microenvironment, and it is critical to note that the majority of studies done on macrophage function in the context of viral liver disease do not differentiate between these 2 populations. Most of the LMs studied are CD68 positive—as both KC and MDM are—but there is little further categorization of LM populations. This is likely due to the lack of markers for and research into this populations, something that we discuss in detail later in the review.

Fig. 1.

Model of effects of viral proteins and infection on LMs. Coinfection is represented on the left and monoinfection on the right. Effects of HBV proteins on LMs are represented in yellow, HIV in red, and HCV in blue. Effects of LMs on cells are represented in green boxes. Proposed interactions in coinfection are represented with a dotted line. Created in BioRender. Gaskill, P. (2025) https://BioRender.com/m8p32su.

In HBV infection, LMs produce proinflammatory cytokines, including TNF-α and IL-1β,^131,135,259 which can disrupt hepatocyte function and activate HSCs, promoting fibrosis and oncogenic pathways (Fig. 1). Anti-inflammatory cytokines like IL-10 and TGF-β^63,153,178 are also produced, likely contributing to immune evasion and fibrotic remodeling. Notably, HBV proteins such as HBsAg and HBeAg drive these responses^{130,135,181,260} in LMs, potentially exacerbating parenchymal inflammation through their proximity to hepatocytes in the space of Disse. This elevated inflammation could damage hepatocytes and drive changes in the liver microenvironment. But HBV also directly impacts hepatocytes, in which it has been shown to drive transformation,¹⁴⁸ and dysregulate cytokines associated with anti-inflammatory/profibrotic effects like TGF-β and PDGF.¹⁵¹ This transformation effect is largely driven by HBx^7,261,262 (Fig. 1).

And, HSCs respond to HBV as well as HCV and HIV primarily through cytokine dysregulation, specifically an increase in profibrotic cytokines such as α-SMA, TGF-β, and Col1A.^{128,151,152,156,157,160,166,263} Viral infection also drives increased motility and proliferation in these cells, key steps in the development of fibrosis.^149,157 Most of these effects have been associated with specific viral proteins. The HBV proteins HBx, HBsAg, and HBeAg increase HSC production of profibrotic cytokines such as TGF-β and α-SMA,^{122,128,135,254} while HBeAg and HBx have been shown to drive proliferation, motility, and contraction.^122,135 In HCV infection, the NS2 protein seems to drive profibrotic cytokine production, increasing HSC TGF-β, α-SMA, Col1, and TIMP-1 production.^{128,146,214,264} In HIV infection, both gp120 and Vpu induce similar profibrotic effects^10,12,109 (Fig. 1).

A mix of pro- and anti-inflammatory profiles is also seen in LMs during HCV infection, suggesting that HCV proteins can drive phone phenotypes. In HCV-infected individuals, LMs show increased inflammasome activation¹⁶¹ and production of both the proinflammatory cytokine TNF-α and the anti-inflammatory IL-10,¹⁶³ but decreased secretion of proinflammatory cytokines IL-6 and IL-12.^155,163,188 In addition, LMs from HCV infected individuals show markers associated with profibrosis/anti-inflammatory macrophage function like CD206 and CD163.^155,188 The HCV proteins, including core protein, NS5a, and E1/E2, have all been shown to stimulate macrophage TNF-α production,^142,146,265 while the NS2 promotes macrophage polarization toward what may be a more anti-inflammatory and profibrotic phenotype characterized by an upregulation in CD206, CD163, hARG1 and profibrotic markers such as TIMP1 and Col1A.^155,188,266 During HCV infection, these responses impact both hepatocytes, through increased oxidative stress and apoptosis, and HSCs, driving their activation and collagen deposition. The presence of macrophages within the sinusoids and parenchyma likely amplifies their effects on nearby hepatocytes and HSCs, disrupting homeostasis across the liver. Indeed, HCV upregulates markers of apoptosis such as Fas, Blk, and p53 in hepatocytes,¹⁵⁸ while also inducing increased epithelial mesenchymal transition¹⁶⁴ and upregulating CXCL10 production,²⁶⁷ These effects are driven by a number of HCV proteins including Core,^137,138 and highlight the balance between virally drive apoptosis and survival (Fig. 1).

The specific effects of HIV on LMs are less well defined, with both prosurvival¹⁶² and profibrotic cytokine production being observed.¹⁶⁰ These effects may be driven by multiple HIV proteins, as gp120, Tat, and Nef can drive proinflammatory cytokine production in LMs,^10,114–117 while Vpu can drive profibrotic cytokine production.¹² As with HBV and HCV, these changes could affect both hepatocytes and HSCs by altering the microenvironment, fostering inflammation, and promoting fibrogenesis. Studies have shown that hepatocytes may be depleted upon HIV infection¹⁶⁵ and that HIV may reduce in p53 production—thereby increasing survival¹⁶⁸—while also driving apoptosis through mitochondrial dysfunction and an increase in ROS.¹⁶⁹ Other studies support this showing increased hepatocytes apoptosis in response to HIV protein gp120,^107,108 although little work has been done examining the effects of other HIV proteins on this cell type. Taken together, these data suggest that much of the impact of viral infection, or viral protein exposure, is mediated by LMs, and how these dysregulated LM affect liver homeostasis is similar among infections.

Overall, the presence of HBV, HCV, and HIV, both KCs and MDMs produce a variety of inflammatory and fibrogenic cytokines but also interact directly with hepatocytes and HSCs, driving fibrosis, inflammation, and cancer pathways (Fig. 1). Specifically, proinflammatory effects may be driven by HIV gp120 and Nef, HCV Core and NS5a proteins, and HBV HBsAg and HBeAg. For all 3 viruses, the upregulation of proinflammatory cytokines likely occurs through an NF-κB mediated pathway, as the most commonly upregulated cytokines are TNF-α, IL-6, and IL-1β. Anti-inflammatory effects were seen in response to exposure to including HBx, HBeAg (HBV), NS3/4a, Core, NS2 (HCV), and Vpu (HIV) (Fig. 1). The reason for these contradictory effects is not clear, nor is it apparent whether these the pro- and anti-inflammatory processes are occurring in the same or distinct LM subsets. It may be that infiltrating MDM exposed to, or producing, some of the proinflammatory viral proteins are the primary inflammatory agents, while resident KCs, responding to observed pathogenic stimuli, ie the viral proteins, promote an environment associated with the production of anti-inflammatory cytokines that aid in the promotion of fibrosis. However, further research is needed to be able to parse these differential effects. Similarly, it is possible that many of the changes in hepatocytes and HSCs are driven by LMs, but again, the specific subpopulation of LMs is not clear. Specifically in the context of HIV-1 infection, it appears that macrophages residing in the liver are depleted following initial infection.^233,241 This loss may result in the recruitment of MDMs to fill this gap, with these newly infiltrating MDM displaying a proinflammatory phenotype that biases the liver away from tolerance and toward chronic inflammation. LMs are also key regulators of HSC activation, and aberrant LM function may increase the production of profibrotic cytokines, which can drive liver stiffness, fibrosis, and the development of HCC.²⁶⁸

Importantly, most of these data result from studies of monocultures or monoinfection, despite the fact that coinfection is common (summarized in Table 2), as we discuss below. Another missing feature of these data is the understanding and accommodation for the fact that in vivo infections always impact mutliple cell types within the liver sinusoid. The balance between the effects of all of these viruses on liver pathology centers around immune tolerance vs immune response, of which macrophages are key regulators. This underscores the need for more research on how viruses broadly impact LM functions. Rather than focusing solely on specific cytokine production or effects on individual cell types, studies should aim to understand the overarching mechanisms. New models are needed to address these issues, particularly models of the different subpopulations of LMs and their markers and functions, particularly in understanding roles of distinct LM populations in the development of fibrosis. Additionally, there is an urgent need for coculture systems that enable more nuanced interrogation of the complex relationship between different LM, HSC, and hepatocytes in the context of viral infections. Understanding these processes, and the cell type–specific interactions driving them, will require a deeper focus on the specific roles and markers of these macrophage subsets, as their localization within the liver uniquely positions them to mediate systemic effects on the organ.

4.5. Viral coinfection

As noted above, there is a relative dearth of information about the molecular effects of liver virus coinfection, as the majority of studies addressing coinfection with HBV/HIV or HCV/HIV are largely epidemiological. Approximately 10% of PWH are coinfected with HBV,²⁶⁹ and that risk of HCV infection is 6 times higher in PWH than the uninfected population.²⁷⁰ Further, coinfection substantively worsens disease, as PWH coinfected with HBV experience an increased risk of developing liver disease, more aggressive disease, and increased risk of liver-related mortality.²⁷¹ These individuals also have higher levels of HBV viremia, a less effective antibody response to the HBV vaccine, and progress to chronic hepatitis at an accelerated rate compared with people infected with only HBV or HIV.^272,273 Although antivirals for HIV and HBV can stop viral replication and prevent infection of new cells, they do not (i) remove integrated HIV DNA or episomally maintained HBV cccDNA in already infected cells,²⁷⁴ (ii) stop production of viral proteins from these sources,^{222,227,275,276} and/or (iii) stop chronic inflammation.^4,5,221,277 Recently published work also indicates that liver cells including hepatocytes, HSCs, and LMs, incubated with HBV+ and HIV+ supernatants and treated with recombinant cytokines and HIV gp120, resulted in increased profibrotic gene expression and liver fibrogenesis through the activation of protein kinase B (AKT) and extracellular-regulated signal kinase (ERK) signaling.¹¹² Similarly, in people with HCV/HIV coinfection, there is a lower rate of clearance of HCV and an increased risk for a secondary episode of HCV viremia, allowing the virus to spread to more cells.^278,279 Furthermore, coculture experiments with HIV and HCV showed increased fibrotic cytokine TGF-β and an increase in ROS that lead to increased expression of collagen and TIMP1,²⁸⁰ and increases fibrogenic gene expression in HSCs.¹⁷² Thus, coinfection of HIV with either HBV or HCV would likely increase the risk of developing liver disease, although the mechanisms by which this may occur are unclear (Fig. 1).

It is possible that the HBV infection could reduce macrophages secretion of proinflammatory cytokines and increase production of TGF-β, enhancing fibrosis and promoting cancer growth (Fig. 1). However, this effect is not consistent across all HBV proteins, as HBeAg inhibits NF-κB signaling and increases anti-inflammatory cytokine production, HBcAg increases IL-6, and HBsAg increases TNF-α and IL-6, but can also drive macrophage polarization to a more anti-inflammatory phenotype. These differences may be the result of the different levels of HBV proteins that are expressed during the stages of chronic infection.²⁸¹ In regard to HCV, structural proteins can upregulate expression of inflammatory modulators, and nonstructural proteins can disrupt this response, so it is not clear which effect would predominate and why. This once again emphasizes the balance between infection and response that plays a role in chronic LM activation and the development of HCC that is associated with these infections. Notably, the upregulation of TREM-1 that occurs in response to HIV-1 infection may bias LMs toward a prosurvival phenotype, allowing the effects of HBV and HCV proteins to continue to affect these cells for longer periods of time, and resulting in the persistence of an altered liver microenvironment. These hypotheses require further inquiry, which is difficult as the challenges involved in culture and experimentation with these viruses, particularly HCV and HBV, and especially in liver cells, remain profound and many of the methods needed to study these mechanisms have yet to be developed.

4.6. Defining populations of LMs

As noted above, LMs are a heterogenous population of myeloid cells that exist as multiple subpopulations. Current data define these subpopulations using a variety of distinct surface markers, which are shown in Table 3. While the use of these markers is often used to define specific subpopulations and phenotypes in individual studies, the use of these markers between experiments, models, and species is inconsistent. The majority of studies that distinguish KCs from other macrophages in the liver, as well as delineating functions of these cells, are done in mouse models due to availability and flexibility in the types of studies that can be performed including depletion and lineage tracking (Table 3). Lineage tracking studies have shown that tissue resident macrophages arise from yolk-sac–derived erythromyeloid progenitors and seed the liver and that these cells are termed KCs.^{290,296,316–318} Although there is literature on the subject of defining LM populations, many of the markers used to define this subset are diverse. This leads to a lack of clarity in terms of understanding which LM populations drive the effects outlined above in the viral pathogenesis sections. Questions of whether specific LM populations are more likely to be susceptible to HIV-1 infection, or whether infection biases LMs that previously played a more canonically homeostatic role are currently being investigated—but do not necessarily have an answer at this moment due to a lack of consensus on what defines a LM.

Table 3.

Markers defining populations of hepatic macrophages.

Paper	Marker	Identifies	Source	Method
Mouse models
Klein et al. 2007282	Ly6ClowF4/80high,	KCs	Mouse	Flow cytometry
Kinoshita et al. 2011283	F4/80, CD11b−CD68+, CD11b+CD68−, CD11b+CD68+	KCs	Mouse	Flow cytometry
Ramachandran et al. 201272	CD11bhiF4/80intLy6Clo	KCs	Mouse	Flow cytometry
	Ly6Chi	MDMs
Gautier et al. 2013284	F4/80+MerTK+FcγRI+	LMs	Mouse	Flow cytometry
	C/EBPα, BACH1, CREG-1	LMs		Microarray (RNA)
Ikarashi et al. 2013285	F4/80+CD32+CD11b+CD68+	LMs	Mouse	Flow cytometry
	F4/80+CD32+CD11b−CD68−	LMs		Flow cytometry
Wang et al. 2014286	F4/80, CD11b, CD14, TLR4	KCs	Immortalized mouse KC	Flow cytometry
Zigmond et al. 2014287	CD45+CD11b+MHCII+F4/80+CD64+	KCs	Mouse	Flow cytometry, microarray (RNA)
	Ly6ChiCD11bhiMHCIIneg	MDMs
Morinaga et al. 2015288	F4/80loCD11b+	KCs	Mouse	Flow cytometry
	F4/80highCD11b+	MDMs
Scott et al. 2016289	Clec4F, CD45, F4/80, CD11bint; Clec4f+CD45+F4/80+CD11bint	KCs	Mouse	Flow cytometry, microarray (RNA)
Mass et al. 2016290	F4/80, Timd4, Nr1h3, Id1, Id3	KCs	Mouse	Bulk RNA-seq
van de Garde et al. 2016291	CD45+F4/80lowCD11bhighLy6chig	KCs	Mouse	Flow cytometry
	CD45+F4/80highCD11bint	MDMs
Beattie et al. 2016292	F4/80, CD163, MARCO, RIC3, Colec12, Timd4, Clec4F	KCs	Mouse	Flow, whole-genome array
David et al. 2016293	F4/80, MHC-11, CD11b, CD206, CD317, CD1d	KCs	Mouse	Immunofluorescence, mass cytometry
	F4/80, MHC-11, CD11b, CD11c	KCs
Halpern et al. 2017294	Irf7, Spic, Clec4f	KCs	Mouse	scRNA-seq
Sierro et al. 2017295	F4/80, CX3CR1hi	Liver capsular macrophages	Mouse	Immunofluorescence
	TIM4, F4/80, CD45	KCs
Gomez Perdiguero et al. 2015296	F4/80	KCs	Mouse	Immunofluorescence
Bonnardel et al. 2019297	Clec4F, F4/80	KCs	Mouse	Flow, RNA (bulk RNA-seq), IF
	Ly6Chi	MDMs
Roth et al. 2019298	Cx3cr, F4/80lowCD11bhighCx3Cr1high	MDMs	Mouse	Bead isolation and flow cytometry
	F4/80, F4/80hiCD11blowCx3Cr1low	KCs
Elchaninov et al. 202017	CD68, CD86, CD163, CD206, F4/80	KCs	Mouse	Flow cytometry
	Ly6C, CD11b, F4/80	LMs
Bleriot et al. 2021299	CD206loEXAM−, CD206hiESAM+	KCs	Mouse	scRNA-seq
Su et al. 202114	Clec4F, VSIG4	KCs	Mouse	scRNA-seq, in situ hybridization
	CCR2	MDMs
Jiang et al. 202127	Clec4F	KCs	Mouse	Flow cytometry, western blot, qRT-PCR
De Simone et al. 2021300	CD45+F4/80+CD11bintTIM-4+	KCs	Mouse	Flow cytometry
	CD45+F4/80+CD11bintTIM-4+ CD206+ ESA+	KCs		Flow cytometry, immunofluorescence
Hildebrandt et al. 2021301	F4/80, Clec4f	KCs	Mouse	Immunofluorescence, spatial transcriptomics (mRNA capture)
Ni et al. 2021302	F4/80, CD11b, TIM-4	KCs	Mouse	Flow cytometry, RT-PCR (RNA)
	F4/80+CD11b−TIM-4−	MDMs
Molina et al. 2022303	Clec4F	KCs	Mouse	Immunoflourescence, RT-PCR (RNA), flow cytometry
	IBA1+CLEC4F−	MDMs
Aktories et al. 2022304	CD45+CD11b+F4/80+,TIM-4+	KCs	Mouse	Flow cytometry
Peiseler et al. 2023305	Clec4F, CRIg, TIM-4	KCs	Mouse	Immunofluorescence
Vanderborght et al. 2023306	Clec4F	KCs	Mouse	Flow cytometry
	Ly6C, CD11b, F4/80	MDMs		Flow cytometry
Rat models
Armbrust et al. 199632	ED1+ED2+, ED1+ED2-	KCs	Rat	Immunofluorescence
Peng et al. 2007307	ED1, ED2, ED3, F4/80	KCs	Immortalized rat KC	Immunofluorescence
Nonhuman primate model
Ahsan et al. 2013233	Ham56, CD163	Macrophage	Rhesus macaque	Immunohistochemistry
	CD68, CD163	KCs		Flow cytometry
Human models
Baldus et al. 1998308	CD68	KCs	Human	Immunohistochemistry
Balagopal et al. 2009241	CD68	KCs	Human	IHC
Kegel et al. 201616	CD68	KCs	Human	IF
MacParland et al. 2018309	CD68+, LYZ, CSTA, CD74	KCs—proinflammatory	Human	Flow cytometry, scRNA-seq, immunohistochemistry
	CD68+, CD5L, MARCO, VSIG4, CPVL, CD163, CCDC88A, C5AR1, LIPA, LILRB5, MAF, CTSB, MS4A7, VMO1, RAB31, SLC31A2, TTYH3, VCAM1, KLF4, HMOX1, AIF1l, TMIGD3	KCs—anti-inflammatory
Tasnim et al. 2019310	CD11, CD14, CD68, CD163, CD32, CLEC-4F, ID1, and ID3	KCs	Human	RNA (qPCR), Flow, IHC
Martrus et al. 2019311	CD14, CD16, CD68	Macrophages	Human	Flow cytometry, immunofluorescence
	CD49a, VSIG4, MARCO	KCs
Ramachandran et al. 201971	TIM-4, MARCO, TREM2, CD9	Scar-associated macrophages	Human	scRNA-seq
	TIM-4, MARCO	KCs
Wu et al. 2020312,313	CD32, CD68, CD11b, CD14; CD14hiCD32mid, CD14midCD32hi	LMs	Human	Flow cytometry, sc-RNAseq
	CD206, VCAM1, HMOX1, MARCO, MRC1	KCs		scRNA-seq
Aizarani et al. 2020314	CD163+VSIG4+, LIRB5+CD5L+MARCO+HMOX1hi, CD1C+FCER1A	KCs	Human	scRNA-seq
Zhao et al. 2020315	CD16, CD206, FTL, SELENOP, HMOX1, FABP5, CD5L, APOE, APOC1, FABP4, NUPR1, MARCO, LGMN, SLC40A1, MRC1	LMs	Human	scRNA-seq

Given the various markers used to define LM populations in the literature, the literature was reviewed for the markers used to define LM populations in different models and listed here for convenience. These markers are described here, along with the study in which they were defined, the population and/or subpopulation of LM they define, the model system in which they were defined, and the method used to identify these markers. These are arranged by model system and hence chronologically within the model system.

The literature defines primary LM subgroups as either KCs, which are liver resident macrophages, or infiltrating MDMs, that are recruited to the liver in response to infection or damage (Table 3).^{283,285,289,319} Ontological studies in rodent models show that KCs are yolk-sac derived, similar to many other tissue macrophages,³²⁰ arising from erythromyeloid progenitors that provide tissue resident macrophages to other organ compartments.^290,296 These cells, separated from other cell types in the liver by LSECs, are able to extend processes to contact HSCs and hepatocytes and both regulate and activate these cell types.²⁹⁷ KCs are self-renewing, long-lived, and represent the first line of defense in the liver^{285,291,296,299,308,314,319} and are generally considered to be the primary modulators of inflammation in that organ. They are most often defined by the markers ARG-1, CD163, CD206,³²¹ and secretion of IL-10 and TGF-β,³²² which likely occurs in response to polarization by IL-4 among other cytokines.³²¹ Notably, the markers used to define KC in rodents are not consistent with those seen in human populations, and KC are challenging to obtain, so similarity between rodent and human KC, and the specific functions of the human KC population, remains poorly defined.

As in many tissues, MDMs arise from monocytes recruited to the liver from the bone marrow upon injury or insult and then mature into macrophages. Within the liver, these cells can drive inflammation and phagocytose bacteria to fight off invading pathogens and mediate tissue damage and repair. The infiltrating MDM often display the markers iNOS and CD80 and secrete inflammatory cytokines such as IL-6, TNF-α, and IL-1β in response to polarization that is driven by IFN-γ and LPS.³²² Unlike KC, these proinflammatory MDM can cross the LSEC barrier to interact directly with hepatocytes,²³ which may result in phagocytosis or clearance of dead cells. However, unlike in many organ systems, MDMs in the liver can take on a KC-like phenotype if the tissue-resident KCs are depleted.²⁸⁹ MDMs that repopulate the liver, generally defined as Ly6C positive cells in rodents, result in a LM population that is highly reminiscent of yolk-sac derived LMs at the transcriptomic and epigenetic level but are still distinct from yolk-sac derived KCs, which are typically defined by CLEC4F.^{20,23,288–290,292,296,303,323} In mouse models, the majority of macrophages in the liver have a KC identity. Upon depletion this can swing to favor MDM temporarily, but yolk-sac–derived KCs will eventually repopulate the niche through self-renewal.²⁸⁷

LMs play 2 main roles. The yolk-sac–derived KCs are primarily thought to maintain tolerance and produce anti-inflammatory mediators, while the infiltrating MDM are generally thought to play a role in clearing infection through the production of proinflammatory cytokines. This can be complicated when the MDM assume the role of yolk-sac–derived KCs upon depletion of this population. Further, cell type–specific functions are difficult to ascribe to specific populations due to lack of clarity in the markers that define each population, particularly in human models. This lack of consensus may be due to the fact that macrophages are highly plastic cells and show different types and levels of receptor expression in response to the liver microenvironment throughout their life cycle. Some of these changes in receptor expression may be triggered by inflammatory stimulation.^{23,312,324,325} Many studies do not distinguish between the KCs of a yolk-sac origin vs MDMs and instead refer to all hepatic macrophages as KCs without separating based on cell surface marker or function, especially when utilizing human models (Table 3).^4,60,244

In mice, studies differentiating KCs from other populations of macrophages in the liver typically rely on cell surface markers CLEC4F, CRIg, and TIM-4, the general macrophage markers CD68 and F4/80, and the myeloid lineage markers CD45 and varying levels of CD11b. These studies typically define KCs as containing lower levels of F4/80 and CD11b than other macrophage populations—although there is some variance in this throughout the literature (Table 4).^{283,285,286,289,304,305,319} Based on the heterogeneity of macrophages in culture, it is important to identify shared/consensus markers (Table 4) in order to better understand the nuances between the different populations of LMs, specifically in regards to their function in the liver compartment.

Table 4.

Markers most commonly used to differentiate LM populations.

Macrophage population	Marker(s)	Function	Model	Citations
MDMs	Ly6C—general marker of monocyte/macrophage lineage; high expression associated with proinflammatory activation	Clearance of infection	Mouse	Martinez-Carmona et al. 2021326, Dos Anjos Cassado et al. 2017327, Molina et al. 2022303, Bonnardel et al. 2019297, Morinaga et al. 2015288, Zigmond et al. 2014287, Vanderborght et al. 2023306
	F4/80—unique marker of murine macrophages	Proinflammatory cytokine production
	CD11b—integrin molecule on the surface of macrophages
	CD68—general marker of macrophage identity
KCs	CLEC4F—scavenger receptor	Anti-inflammatory cytokine production	Mouse	Wong et al. 2010328, Yuan et al. 2017186,251, Scott et al. 2016289, Peiseler et al. 2023305, Mass et al. 2016290, Molina et al. 2022303, Aktories et al. 2022304, Bonnardel et al. 2019297, Bleriot et al. 2021299
	TIM-4—mediates clearance of apoptotic cells	Wound repair
	CRIg—complement receptor involved in maintenance of tolerance	Tolerance
	CD206—marker of anti-inflammatory macrophage polarization
Peritoneal macrophages	Gata6—transcription factor	Infiltration of liver postinsult	Mouse	Hossain et al. 2022329, Rosas et al. 2014323, Wang et al. 2016330
Splenic macrophages	CD11b	Contribution to fibrosis	Mouse	Dos Anjos Cassado et al. 2017327, Fukushima et al. 2020331, Li et al. 202158, Zhang et al. 202060
	CD43—T-cell–associated marker
	Ly6C
Liver capsular macrophages	F4/80	Neutrophil recruitment	Mouse	Bleriot et al. 201921, Sierro et al. 2017295
	CXC3CR1—mediates migration, adhesion, and chemotaxis	Resolution of infection
Proinflammatory LMs	CD68—general marker of macrophage identity	Proinflammatory cytokine production	Human	Wu et al. 2020312,313, MacParland et al. 2018309
	CD11b	Resolution of infection
	CD74—surface receptor for cytokine macrophage migration inhibitory facor; plays a role in antigen presentation
	LYZ—bacteriolytic function
Anti-inflammatory macrophages	TIM-4	Production of profibrotic cytokines	Human	Cosma et al. 2023332, Wu et al. 2020312,313, MacParland et al. 2018309, Martrus et al. 2019311, Ramachandran et al. 201971
	MARCO—macroghage scavenger receptor	Tolerance
	CD9—cell adhesion molecule
	CD206
Scar-associated macrophages	TREM2—stimulate phagocytosis and suppress cytokine production	Wound healing	Human	Ramachandran et al. 201971
	TIM-4	Fibrosis
	MARCO

Using markers of LM populations laid out in table one, the most common markers identifying these subsets were identified and defined in order provide possible targets to use in differentiating LM populations from one another.

Both CLEC4F and TIM-4 have KC specific roles, with CLEC4F serving as an important cell receptor for phagocytosis and TIM-4 helping to resolve inflammation.^27,302 Other possible markers of KC identity include Nr1h3, ID3, CD206, MHCII, MerTK, FCyRI, BACH1, CREG-1, CD163, and CD32 (Table 3).^{284,287,289,290,299,305} In mice, this population of CLEC4F+ and TIM-4+ liver resident macrophages is separate from MDMs that are recruited to the liver upon depletion of KCs, or injury to the organ. These MDMs are characterized by high levels of F4/80, as well high levels of Ly6C—both of which are commonly used markers of macrophage identity (Table 4).^17,72 It has been well documented that depletion of KCs in mice results in the engraftment of MDMs, and that these macrophages may take on a KC-like identity and many of the same functions.^{72,282,289,297,302} However, there appear to be key differences in phagocytic activity and proinflammatory cytokine production between these 2 subsets—indicating that after a loss of KCs, MDMs partially, but do not completely, restore this niche, leaving certain functions unfulfilled unless and until the KCs self-renew.^{283,287,291,298,300,305,319}

Markers defining human KCs are not as well defined as in mouse, but recent research using single-cell RNA sequencing (scRNA-seq) and transcriptomics has identified multiple populations of LMs, as seen in mouse models, and provided possible markers for distinguishing between these (Table 3). As in mice, at least 2 subsets of macrophages are present in the human liver.^{309,311,313–315} These populations are defined in the literature as proinflammatory and anti-inflammatory macrophages, delineated by their activity promoting or resolving of inflammation. While these definitions are extremely broad, macrophages are commonly divided into these very large groups and then further subdivided; as they are extremely plastic cells that change based on stimuli and timing, so it would be nearly impossible to define a static population of macrophages in this way. Macrophages in the liver displaying proinflammatory function are most commonly defined by CD68, CD11b, CD74 and LYZ and generally show increased proinflammatory cytokine production (Table 4).^{309,311,313–315} Notably, in humans the markers that have commonly been used to designate LMs as proinflammatory are general macrophage markers, and do not appear to be specific to cells of a myeloid lineage in the liver. The second population of human LMs appears to play a more anti-inflammatory and tolerogenic role in liver function and are defined by the presence of CD206, MARCO, VSIG4, VCAM1, and HMOX1 (Table 4).^{309,311,313–315} Other markers that are used to define this population of macrophages include MRC1, CD163, KLF4, TIMGD3, LIRB5, CD1C, CD5L, and APOC1 (Table 3). This second population typically has more immunosuppressive functions and has been shown to secrete less TNF-α in response to LPS stimulation.^{309,311,313–315} Recently, a third population of TIM-4+MARCO+TREM2+CD9+ macrophages has been identified as producing factors that promote fibrosis. This population, termed scar-associated macrophages, appears to have roles in scarring and wound healing and seem to expand during liver fibrosis, inducing the production of collagen and activating profibrogenic pathways (Table 3).⁷¹

These different populations of macrophages also appear to cluster at different locations in the liver. The VSIG4+MARCO+ macrophages seem to concentrate in periportal areas, are more transcriptionally reminiscent of the canonical yolk-sac KCs, and appear to have a role in resolution of inflammation and promotion of anti-inflammatory cytokines. In contrast, CD68+ MARCO− LYZ+ CD74+ macrophages display a proinflammatory function, are more migratory, and appear to be more transcriptionally similar to MDMs that are defined as inducing inflammation (Table 4).²³ Unlike in mice, in human livers, the macrophages promoting inflammation seem to comprise a greater percentage of the total LMs population.^287,309 However, human samples almost always come from a donor undergoing liver biopsy as a result of investigation for liver disease or a deceased donor, so it is difficult to ascertain the effect these factors would have on LM distribution.

In addition to yolk-sac–derived and monocyte-derived LMs, other populations of macrophages have also been identified in the liver, specifically, peritoneal, splenic, and capsular macrophages that infiltrate the liver after injury or depletion. Peritoneal macrophages have been shown to invade injured organs in response to danger signals.^23,329,330 In mice, these macrophages are Gata6+ and have been shown to infiltrate the liver when it shows metastases in a model of metastatic mesothelioma.²³ Splenic macrophages can to populate the liver and contribute to fibrosis through the promotion of a proinflammatory macrophage phenotype in mouse models.^331,333,334 These splenic macrophages are CD11b+CD43+ and have low levels of Ly6C, but no specific identifying markers have been identified to our knowledge.³³³ Liver capsular macrophages that are involved in neutrophil recruitment and the resolution of infection in the liver have been identified as TIM4-, but coexpressing F4/80 and CX3CR1^hi.²⁹⁵ Currently, little is known about what role these other macrophage subpopulations play in the pathogenesis of liver disease, especially in the context of viral infections, as specific markers differentiating these cells from other LMs have not been well defined.

5. Models of studying LMs

As stated previously, it is important to note that the majority of the studies on LMs have been conducted in mice, which have fewer limitations in regard to techniques such as lineage tracking, providing more information about the origin of cells as well as increasing the ease of obtaining liver cells in order to study them. Because of this gap in accessibility between mouse and human cells, there is a dearth of knowledge regarding the exact markers and the different populations of LMs in humans, although there are a number of groups actively working to rectify this deficit.

5.1. Rodent models

Almost all of the studies that define the markers and functions of LMs have been done in mouse models. These models rely on a number of different techniques to isolate LMs including in situ hybridization, reverse transcription PCR (RT-PCR), immunofluorescence, or differential centrifugation for initial separation from the nonparenchymal cell fraction.^{14,17,303,335} High-throughput methods such as RNA-seq, both single cell and bulk, as well as microarrays are also used.^{14,21,284,287,297,299,301} These studies are often combined with functional assays examining cytokine and ROS production, immune cell recruitment, and phagocytosis.^{27,283,288,291,297,299} Mouse models are also useful for modeling the function of LMs during disease states. For example, use of carbon tetrachloride (CCl4) and high fat diets (HFDs) to model liver fibrosis in mice indicate that macrophages produce proinflammatory cytokines, which drive chronic inflammation and the development of liver disease.^336–338 Fewer studies have used rat models to study LMs, although much of the research done in these models echoes that seen in mice, showing multiple populations of LMs with differential expression of markers ED1 and ED2 based on location.³² There have also been attempts to establish immortalized mouse and rat LM cell lines. An immortalized mouse KC line was shown to express F4/80, CD11b, CD13, and TLR4 and have phagocytic activity and proinflammatory cytokines, but other markers including CLEC4F were not noted,²⁸⁶ and an immortalized rat KC line was shown to express ED1 and ED2 along with F4/80, while also maintaining cytokine production.³⁰⁷

5.2. Zebrafish models

Zebrafish can also provide an avenue for understanding the developmental origin and function of LMs,³³⁹ as LM development in this model parallels development in mammals. Further, the tissue-resident macrophages in zebrafish share many mammalian LM features including self-renewal and maintenance functions at homeostasis.^340–342 The functions of LMs in zebrafish models of liver disease also appear similar to those in mammals, including models of metabolically associated fatty liver disease, acute liver injury models, and HCC.³⁴⁰ Zebrafish models are particularly promising due to the capacity for live cell imaging, but further studies are needed to interrogate the heterogeneity of macrophage populations to establish whether this is a viable representative of human disease and LM function.

5.3. Nonhuman primate models

NHP models of LMs are uncommon, with the majority of these studies taking place in the context of viral disease, rather than doing lineage-tracing or functional assays to establish the role of LMs in this model. However, many of the studies on NHP LM use CD68 as a catch-all for LM identity rather than differentiating specifically based on population as has been done in mouse and some human models.^{233–235,240,343}

5.4. Primary human cell models

The majority of studies involving primary human LMs involve the isolation of LMs from a nonparenchymal cell fraction using magnetic beads or differential centrifugation, or from the dissociation of liver tissue before sorting.^{16,309,311,315} These studies are generally limited by access to human samples and often confounded by the disease state of the tissue. The functions of LMs in human studies are largely synonymous to those that have been found in mouse models, specifically the presence of multiple populations of macrophages that may have either more pro- or anti-inflammatory function.^{292,297,309,311,315} However, the relevance of mouse findings to human models requires further investigation as mouse and human immune systems do not always show similar functions or responses to disease.^344–347 Thus, there is considerable need for a consistent source of human liver samples for individuals interested in LM function and disease, as well as a pressing demand for the development of other models to study human LMs.

5.5. Organoids and induced pluripotent stem cell–derived models

Organoids are 3D, multicellular, artificially grown structures that can be derived from a number of different sources including primary cells and induced pluripotent stem cells (iPSCs).³⁴⁸ Organoids can be useful for modeling complex interactions between multiple cell types, as they have been shown to retain features comparable with their origin tissues.³⁴⁸ These models are still relatively new, but recent studies have suggested that organoids may be a better way to understand in vivo mechanisms of liver disease. Growing primary cell organoids consisting of primary human hepatocytes, KCs (as defined by CD68 expression), HSCs, and sinusoidal endothelial cells for 4 wk, researchers were able to show markers of liver function including albumin synthesis, urea production, and drug metabolism, as well as a miR-122-mediated inhibition that led to increased inflammation in this model.³⁴⁹

Organoids can also be generated using iPSCs, which is of particular interest, as access to primary human liver cells is limited. The use of iPSC-derived models would alleviate these difficulties, but there has been some difficulty in establishing such models. However, a recent study done by Ouchi et al.³⁵⁰ showed that liver organoids generated from iPSCs contained Kupffer-like cells (as defined by CD68 expression) that were transcriptionally similar to in vivo derived tissues. Other groups have described nonorganoid protocols for generating iPSC-derived KCs that express similar morphologies and cell markers to primary LMs including CD14, CD68, CD163, CD11, and CD32.^304,310 Further, macrophages cocultured within a liver organoid have been shown to shift to a more Kupffer-like phenotype, and the results shown using this model are consistent with clinical outcomes.³⁵¹ Although there have been recent advances in the field of iPSC-derived KCs, more work needs to be done to characterize the populations of LMs generated by iPSCs and compare them with primary human cell models.

Each model of LMs offers different advantages and caveats, as well as different insights into the function and breakdown of LM and other liver cell populations. Rodent models, particularly mice, are the most commonly used and have been instrumental in establishing the basic markers and functions of LMs, especially in the context of liver fibrosis and chronic inflammation through utilizing interventions such as HFDs and CCl4. However, rodent models likely do not have direct translation to human research. Zebrafish models provide the opportunity for live-cell imaging, but likely suffer the same issues as rodent models in terms of translatability. While NHP models are likely more applicable to human studies, they are difficult to acquire and require substantial resources to use. Primary human cell models are limited by access to samples, but likely provide the most information about LM populations in a human system. As organoid and iPSC-derived models continue to gain popularity and feasibility, it will be important to use the above models to categorize this generated cell type, as there is still significant work that needs to be done to refine them and compare them with primary human cells.

6. Future directions

LMs of all types play critical role in the development of liver pathogenesis in response to viral infection, especially in the context of liver cancer. However, the field is still limited in the model systems and assays that can be used to define and evaluate the role of distinct LM populations and viruses in the etiology of liver disease. The majority of the models currently used to study LMs to date are rodent systems, which have been critically important in defining the markers and functions of LM populations in health and disease. However, moving forward will require new systems that will allow for better interrogation of the specific molecular processes by which LMs, and specifically virally induced changes in LM function, contributes to the development of liver pathologies, particularly HCC. As immunologic findings in mice are not always analogous to those in humans, and little is known about the specific parallels between mice and human livers, development of human systems is imperative to advance translational research in this area. With the advent of new technology such as scRNA-seq, it is now technically possible to understand LMs in human systems and track how these cell populations change with disease. However, this would require better access to primary liver samples for labs interested in studying LMs and disease, as well as to develop and explore other models including iPSCs, organoids and liver/slice and coculture models. The use of cocultures of either primary or iPSC-derived human LMs, HSCs, and hepatocytes could enable direct study of infection in these cells and provide a more consistent, accessible model of direct coinfection with HIV/HBV or HIV/HCV. This is of particular importance as the current models, which depend on rodents, do not allow for infection with HBV or HIV, as neither of these viruses can infect rodents. Using multiple models of different types of LMs alongside techniques such as scRNA-seq and epigenetic sequencing, we may be better able to define differences in tissue resident macrophages and thereby understand how different cells contribute to development of liver diseases like HCC, especially in the context of viral infection.

Use of large-scale sequencing analyses, combined with both complex and reductionist models will also help to resolve a critical caveat for this review, the intentionally narrow focus on the direct effects of viral infection, particularly viral proteins from HIV-1, HBV, and HCV, on LM function. While this is a critical area of investigation, it also represents only a portion of the interactions that regulate the role of LM during viral infection. In addition to infecting macrophages (HIV-1) and releasing proteins (HIV-1, HCV, and HBV) that influence immunologic phenotype and response, hepatotropic viruses such as HBV and HCV can indirectly influence immune activity via infection and dysregulation of nonimmune cells such as hepatocytes. These infected cells secrete a number of immunomodulatory factors that can alter the liver microenvironment, and this is highly relevant to LM function given the high plasticity of macrophages and their responsiveness to local environmental cues, including signals from both parenchymal and nonparenchymal liver cells.^{129,352–356} For example, transgenic mouse models expressing viral proteins such as HCV NS3/4A in hepatocytes show that the factors released by hepatocytes in response to intracellular protein expression shape immune cell composition and behavior in the liver.^357–359 Additionally, hepatitis viruses can induce the release of chemokines such as CXCL8, which is implicated in HCC pathogenesis³⁶⁰ and may possess IFN-antagonistic or proviral activity in the context of HBV and HCV infection.^361–364 Release of CXCL8 and other CXCR2 ligands that attract and activate not only on macrophages, but also a host of other immune cells including granulocytes, dendritic cells, and T cells. This will, in turn, have effects that spread outside of the cells that are virally infected, which we have primarily identified as macrophages and hepatocytes. While this review does not directly address these indirect pathways, the scope of these interactions is enormous and almost certainly plays an important role in liver pathogenesis. Connecting the direct influence of viral proteins on macrophages with the broader influence of viral infection on different liver cell types, and then assessing how these changes intercellular communication, immunological networks and the hepatic microenvironment is a critical area for future investigation.

Notably, the need to address these complex networks and better define how changes across multiple cell types and throughout the entire liver drive pathogenesis also highlights the need for more refined definitions of LM subpopulations. Without defining the types of LM responding to these changes, it will be extraordinarily difficult to study, discuss and target them. This is similar to the developments in other tissue systems like the CNS, in which there has been an intense focus on defining different types of microglial populations in the brain.^365–367 Leveraging omics technologies scRNA-seq and spatial transcriptomics, we can identify distinct LM populations based on their molecular signatures. This expanded classification would enable alignment of these specific populations with known markers, which can then be used to carry out functional studies. Functional studies tied to precise populations are keys to understanding the roles of different LMs in health and disease, and are an essential step in the development of novel therapeutic strategies, including interventions that selectively target these populations to modulate liver function. It is possible that certain parts of the liver may contain macrophages that drive more or protect against the pathogenic effects associated with fibrosis and the development of HCC, and identification of such populations is a key to targeting them. Such interventions could specifically activate regenerative processes or attenuate harmful inflammation, offering new avenues for treating liver diseases, including fibrosis and chronic liver inflammation. This focus on expanding our knowledge of LM diversity and controlling their specific responses could ultimately revolutionize how we approach liver therapy, much like the shift seen with microglia in neuroinflammation research and expanding this understanding within the field is key to the design of future studies.

7. Conclusion

Current data show that the populations of macrophages in the liver are highly heterogenous and that the function of these cells is altered by viral infections. Many of these changes, such as HBV induced changes in tolerogenicity and anti-inflammatory cytokine production, and HIV mediated increases in LPS sensitivity, NF-κB activation, and production of inflammatory mediators, could contribute to the persistence of viral infection and chronic inflammation in liver, driving the development of fibrosis and HCC. A possible mechanism underlying these changes centers around the fact that HIV, HBV, and HCV all appear to affect TLR signaling. In HBV-infected patients, HBcAg drives proinflammatory cytokine production in a TL2R-dependent manner, while much of HCV structural protein activity is driven by binding to TLR2 to drive expression of inflammatory modulators, and HIV infection drives expression of TLR4, although the specific importance of this and mechanisms underlying this interaction in the context of HIV infection are yet-to-be defined. The effects of viral infection of distinct populations of LMs in the response to liver injury and the development of HCC require further study to better define the functions that are altered to accelerate and exacerbate disease pathogenesis. The need for this is especially acute in the context of coinfection, as the study of HIV–HCV and HIV–HBV coinfection is currently challenging and there is currently little mechanistic research despite the increased risk of disease. Although there are significant limitations to studying LMs, especially in the context of viral infections, recent advances in the field provide insights to how the dual role of LMs in immune activation and tolerance may underlie liver damage and provide a roadmap to novel treatment options.

Funding

This work was supported by funding from the National Institutes of Neurological Disease and Stroke NIH-R01NS089435 to M.R.N. and the National Institutes of Allergy and Infectious Disease R21AI158024 to M.J.B. and M.R.N. and from the National Institutes on Drug Abuse, R01DA057337 and R61DA058051 to P.J.G.