Review article: chemokines as orchestrators of autoimmune hepatitis and potential therapeutic targets


  • A. J. Czaja

    Corresponding author
    1. Division of Gastroenterology and Hepatology, Mayo Clinic College of Medicine, Rochester, MN, USA
    • Correspondence to:

      Dr A. J. Czaja, Division of Gastroenterology and Hepatology, Mayo Clinic College of Medicine, 200 First Street S.W., Rochester, MN 55905, USA.


    Search for more papers by this author

  • This uncommissioned review article was subject to full peer-review.



Chemokines contribute to the pathogenesis of autoimmune hepatitis by directing the migration and positioning of inflammatory and immune cells within the liver.


Describe the liver-infiltrating effector cell populations in autoimmune hepatitis, indicate the chemokines that influence their migration, describe the role of chemokines in hepatic fibrosis and identify chemokine-directed treatment opportunities.


Studies cited in Pub Med from 1972 to 2014 for autoimmune hepatitis, chemokines in liver disease, pathogenesis of autoimmune hepatitis and chemokine therapy were selected.


T helper type 17 lymphocytes expressing CXCR3 and CCR6 are attracted to the liver by the secretion of CXCL9, CXCL10 and CXCL11. These cells recruit pro-inflammatory T helper type 1 lymphocytes expressing CXCR3 and CCR5 by secreting CXCL10. Resident natural killer T cells expressing CXCR6 migrate in response to the local secretion of CXCL16, and they modulate the inflammatory response. T helper type 2 lymphocytes expressing CCR4 are attracted by CCL17 and CCL22, and they dampen the expansion of pro-inflammatory cells. Regulatory T cells expressing CXCR3 are attracted by the secretion of CXCL9, and they help dampen the pro-inflammatory responses. CCL2, CCL3, CCL5, CXCL4, CXCL10 and CXCL16 promote fibrosis by activating or attracting hepatic stellate cells, and CX3CL1 may prevent fibrosis by affecting the apoptosis of monocytes.


Chemokines are requisites for mobilising, directing and positioning the effector cells in immune-mediated liver disease. They are feasible therapeutic targets in autoimmune hepatitis, and the evaluation of monoclonal antibodies that neutralise the pro-inflammatory ligands or designer peptides that block receptor activity are investigational opportunities.


Autoimmune hepatitis is characterised by a dense lymphoplasmocytic infiltrate of the portal tract and interface hepatitis.[1-3] The inflammatory infiltrate may extend into the liver parenchyma and bridge across portal tracts and central veins or be associated with lobular collapse.[4-6] Interface hepatitis is a manifestation of the ongoing apoptosis of hepatocytes,[7-9] and apoptotic bodies derived from the dying hepatocytes can stimulate hepatic stellate cells to transform into myofibroblasts and promote hepatic fibrosis.[10, 11] The molecular events that account for these histological manifestations and affect the clinical severity and outcome of the disease are in turn countered by cellular and molecular responses that promote a return to normalcy.[12, 13]

The main elements that influence the attraction of immune and inflammatory cells into the liver, position them at the critical sites of tissue damage and affect hepatic fibrosis are chemokines.[14-22] Chemokines are small proteins that mediate the migration of inflammatory and immune cells to sites of tissue damage, and they are cytokines that can directly interact with liver cells.[23, 24] Chemokines induce chemotactic activity in diverse effector cell populations by serving as ligands that attract the cells with complementary ligand receptors. The ligands can be produced by injured resident cells within the liver (hepatocytes, endothelial cells, hepatic stellate cells and dendritic cells) and by first responder cells (neutrophils and monocytes).[25] The attracted effector cells can in turn produce cytokines that enhance the production of ligands and the attraction of more effector cells in a positive feedback loop.[22]

The chemokine receptors are expressed mainly on immune and inflammatory cells, but they can also be present on resident cells within the liver.[25] Multiple ligands of the same family can attract different cells expressing the same receptor, and the targeting of the immune response can be promiscuous, redundant, and widespread.[17, 25, 26] These chemoattractant molecules, both the ligands and the ligand receptors, have emerged as key regulators of the pathogenic pathways of diverse chronic liver diseases, including autoimmune hepatitis,[25, 27, 28] and they are feasible therapeutic targets.[22, 29, 30]

The goals of this review are to describe the key effector cell populations in autoimmune hepatitis that are chemokine-responsive, indicate the chemokines that influence the intrahepatic migration of these cell populations, describe the role of chemokines in tissue fibrosis, and identify chemokine-directed treatment opportunities.


The English abstracts cited in Pub Med from 1972 to 2014 for autoimmune hepatitis, chemokines in liver disease, pathogenesis of autoimmune hepatitis and chemokine therapy were reviewed. Abstracts judged pertinent to the review were identified; key aspects were noted; and full-length articles were selected from the abstracts judged germane to the review. A secondary bibliography was developed from the references cited in the selected full-length articles and additional PubMed searches were performed to expand the concepts developed in these articles. Secondary PubMed searches focused mainly on each chemokine and effector cell subset implicated in liver disease. The discovery process involving abstract review and the acquisition of full-length articles was repeated, and a tertiary bibliography was developed after reviewing these selected articles. The tertiary references were then expanded by additional PubMed searches. The number of abstracts cited by PubMed and reviewed for pertinence to this topic during the primary, secondary and tertiary searches exceeded 2000. Those judged most pertinent to the topic exceeded 300 and the number of full-length articles reviewed exceeded 60.

Nomenclature and nature of the chemokine network

Chemokines are designated by the configuration of cysteine (C) residues that are conserved at the amino terminus of the molecule and by the number of variable amino acids (X) that separate the conserved motifs.[25, 31] Four sub-classifications have been described, and they have been designated as C, CC, CXC and CX3C. The chemokines are small structurally related polypeptides (8–12 kDa) that serve as ligands (L) that attract cells with complementary receptors (R).[31-33] The expression of chemokines can be transiently induced during an immune response or they can be constituents of lymphoid tissue that help regulate the development and homoeostasis of immune cells.[17, 31] The ligands are designated by their cysteine motifs followed by the appendage, L and a number indicating their place within a family of ligands.

The receptors (R) are transmembrane molecules at the surface of immune and inflammatory cells that are coupled to G proteins within the cell.[22] The binding of the chemokine ligand with the receptor triggers conformational changes that generate intracellular signals which promote directional cell migration.[17, 31, 33] The receptors are designated by their cysteine motif followed by the appendage, R and a number indicating their place within a family of receptors. The receptors can interact with diverse ligands associated with the inflammatory and immune responses.[17] The chemokine network consists of at least 50 ligands and 19 receptors.[21]

Effector cell populations and chemokine phenotypes in autoimmune hepatitis

T helper type 1 (Th1) lymphocytes and T helper type 2 (Th2) lymphocytes have long been recognised as important contributors to the pathogenesis of autoimmune hepatitis (Table 1).[34-36] T helper type 17 (Th17) cells have also been implicated, and they may actually constitute the first wave of immune cells infiltrating the target tissue (Table 1).[37-44] T regulatory cells (CD4+CD25+) cells that specifically express the transcription factor, Foxp3, are recruited to the liver to attenuate the immune response and restore homoeostasis,[45-47] and natural killer T (NKT) cells with inhibitory and stimulatory actions on the cytokine pathways can either favourably modify the cytopathic process[48-52] or accentuate the inflammatory response and promote hepatic fibrosis.[30, 53, 54] The migration of each of these various cell populations to sites of injury within the liver is directed by the chemokine network.

Table 1. Principal cell types associated with autoimmune hepatitis
Cell typeChemokine receptorsCytokine productionsSalient pathological features
  1. AIH, autoimmune hepatitis; IFN-γ, interferon-gamma; IL, interleukin; NK cells, natural killer cells; NKT cells, natural killer T cells; PPAR-γ, peroxisome proliferator-activated receptor-gamma; TGF-β, transforming growth factor-beta; Th1, T helper 1; Th2, T helper 2; Th17, T helper 17; TLR, Toll-like receptor; TNF-α, tumour necrosis factor-alpha.

Th17 lymphocytes

CXCR3 for hepatic migration[44]

CCR6 for epithelial positioning[44]

IL-17A, IL-17F, IL-21, IL-22, TGF-β, IFN-γ and IL-23 receptor[37, 39-41, 44]

Generated by TGF-β and IL-6[40]

IL-23 essential for maintenance and expansion[29]

Present in diverse liver diseases, including AIH[43, 55-58]

Pro-inflammatory effects and severe disease[43, 44, 56, 57]

Anti-inflammatory counteractions via IL-22[40]

Attracted by ligands CXCL9, CXCL10, CXCL11[61]

Th1 lymphocytes

CXCR3[39, 67, 79]

CCR5[65, 66]

IFN-γ, IL-2, TNF-α[39, 82]

Secretion of IFN-γ and induction by IFN-γ[39]

Maturation promoted by IL-12[39, 82]

Pro-inflammatory and immune-mediated effects[39]

Attracted by CXCL9, CXCL10, CXCL11[22, 25, 67-70]

Antagonised by PPAR-γ and IL-12 suppression[82]

Th2 lymphocytes

CCR4[65], [67],[77]

CCR3[78, 79]


IL-4, IL-5, IL-13, IL-25[39]

Autocrine induction by IL-4[39], [82]

Supports antibody-dependent cell-mediated toxicity[64]

Anti-inflammatory actions counter Th1 cells[39]

Inhibits Th17 expansion by secreting IL-4[37]

Expresses PPAR-γ and secretes IL-10[39, 82]

Sustained by IL-4-based positive feedback loop[39, 82]

T regulatory cellsCXCR3[84, 98]IL-10[84, 85]

Induced from naïve CD4+ cells by TGF-β and IL-2[92]

Anti-inflammatory effects via PPAR-γ and IL-10[82, 83]

Inhibited by IL-6 and TLR-8[39, 43]

Attracted to liver by CXCL9[84]

Reciprocal relationship with Th17 cells[38, 39]

NKT cellsCXCR6[30, 107]IL-4, IL-10, IL-21, TNF-α, IFN-γ[109]

Attracted to CXCL16 ligand[107, 108]

Stimulatory and inhibitory immune actions[103]

Activates B cells, T cells, NK cells, other NKT cells[105]

Promotes regulatory T cells[106]

Activates hepatic stellate cells[53, 54]

Th17 lymphocytes

Th17 cells can be generated within the liver by the intrahepatic production of interleukin (IL)-6 and transforming growth factor-beta (TGF-β), or they can migrate to the liver from the peripheral circulation.[43] Diverse chronic inflammatory diseases of the liver, including autoimmune hepatitis, have been associated with Th17 cells.[41-43, 55] The number of Th17 cells is increased in the peripheral blood and liver of patients with autoimmune hepatitis, and these cells have been associated with severe hepatic inflammation and advanced fibrosis.[43] The number of intrahepatic Th17 cells in alcoholic liver disease has been associated with severe liver disease[56]; Th17 cells congregate near damaged bile ducts in primary biliary cirrhosis (PBC)[57]; and Th17 cells have been associated with severe disease in patients with chronic hepatitis B.[58]

The chemokine receptors, CXCR3 and CCR6, are expressed on intrahepatic Th17 cells, and they are essential for the recruitment and positioning of these cells within the inflamed liver (Table 1).[44, 59] Th17 cells migrate to the liver in response to the intrahepatic production of ligands complementary to CXCR3,[44] and they are positioned at the injured epithelial surface by the attraction of CCR6 to CCL20 (Figure 1).[43, 44, 60] CXCL9 [monokine-induced by IFN-γ (MIG)], CXCL10 [IFN-γ-inducible protein 10 (IP-10)], and CXCL11 [interferon-induced T cell alpha chemo-attractant (ITAC)] are the ligands of CXCR3,[61] and they can orchestrate the migration of Th17 cells into the hepatic sinusoids.

Figure 1.

Chemokine ligands and receptors orchestrating the hepatic infiltration of inflammatory and immune cells. Subsets of T helper (Th) lymphocytes, B lymphocytes, natural killer T (NKT) cells, and monocytes constitute the principal pro-inflammatory components of the cellular response, and regulatory T lymphocytes constitute the main anti-inflammatory cellular response. The migration of these effector cells is regulated by the chemokine ligands (CXCL9, CXCL10, CXCL11, CXCL12, CXCL16, CX3CL1, CCL17, CCL20 and CCL22) secreted by resident cells within the liver [sinusoidal epithelial cells, NKT cells, dendritic cells, cholangiocytes and hepatic stellate cells (HSC)] and nonresident cells that migrate to the liver (macrophages, lymphocytes, monocytes, NKT cells) at the onset of tissue injury. The cognate receptors (CXCR3, CXCR4, CXCR6, CX3CR1, CCR4, CCR5 and CCR6) on the migrating effector cells are attracted to the complementary ligand. The interactions between the ligands and receptors are promiscuous, and multiple ligands can attract different cells expressing the same receptor.

Th1 lymphocytes

Th1 lymphocytes can differentiate and clonally expand into liver-infiltrating CD8+ cytotoxic lymphocytes, and they have been recognised as key effector cells in the development of autoimmune hepatitis (Table 1).[62-64] Th1 cells preferentially express CCR5[65, 66] and CXCR3,[65-67] and they are attracted to the ligand, CXCL10 (Figure 1).[25, 67, 68] Other CXCR3 ligands are CXCL9 (MIG) and CXCL11 (ITAC).[67, 69, 70] Th17 cells induce the production of the pro-inflammatory cytokines, IL-6, IL-1, and tumour necrosis factor-alpha (TNF-α) and the expression of CXCL10. These products recruit the Th1 cells to the sites of liver injury.[39] Th1 cells in turn produce IFN-γ and help modulate expansion of the Th17 cells.[37]

Th2 lymphocytes

Th2 lymphocytes differentiate along a cytokine pathway that promotes the clonal expansion of plasma cells and the production of immunoglobulin in autoimmune hepatitis (Table 1).[34, 35, 64] Th2 cells also exert an inhibitory effect on Th1 cells,[39] and they can suppress the expansion of Th17 cells, monocytes and macrophages.[37, 71] The interactions between the Th1 and Th2 lymphocytes can affect the activity of autoimmune hepatitis, as best exemplified during pregnancy. High blood oestrogen levels during pregnancy mediate a cytokine shift from a pro-inflammatory Th1 profile to an anti-inflammatory Th2 profile, and the activity of autoimmune hepatitis can subside.[72, 73] Following delivery, blood oestrogen levels fall, the cytokine milieu shifts back to a Th1 predominance, and autoimmune hepatitis can exacerbate in 12–87% of patients.[74-76]

Th2 lymphocytes preferentially express CCR4,[65, 67, 77] but they may also express CCR3[78, 79] and CCR8[80] (Figure 1). The CCR4 ligands are CCL17 [thymus-and-activation-regulated chemokine (TARC)] and CCL22 [macrophage-derived chemokine (MDC)].[67, 81] Th1 and Th2 cells express heterogeneous and overlapping receptors that are not subset specific, and the differentiation of these populations by their receptors or their receptor ligands is imprecise.[67, 79]

Regulatory T lymphocytes

Regulatory T cells constitute 5–10% of the CD4+ T cell population,[82] and they suppress immune-mediated responses by expressing peroxisome proliferator-activated receptor-gamma (PPAR-γ),[82, 83] by producing the anti-inflammatory cytokine, IL-10[84, 85] and by inhibiting pathogenic Th17 cells (Table 1).[86] CD4, CD25 and the transcription factor, Foxp3, characterise the regulatory T cells,[45-47] but since other activated CD4+ T cells can transiently express CD25 and Foxp3,[87] CD127 (IL-7 receptor),[88] CD39[89] and site-specific demethylation of the FOXP3 gene[90] have been suggested as additional distinguishing features of this cell population.[91]

Regulatory T cells are induced from naïve CD4+ CD25 T cells by TGF-β and IL-2,[92] and each of these cytokines is essential for the development, survival and function of the regulatory T cell population (Table 1).[84, 93-97] CXCR3 is expressed by the regulatory T cells,[85, 98] and cells bearing this receptor are attracted to the CXCR3 ligand, CXCL9 (MIG) (Figure 1).[84, 85, 99] The production of CXCL9 is increased in inflammatory liver injury by sinusoidal endothelial cells, and regulatory T cells are attracted to these sites of injury.[99] Intra-hepatic regulatory T cells are abundant in the liver of patients with autoimmune hepatitis, and the expression of the CXCR3 ligand, CXCL9, is up-regulated.[84, 100] In contrast, the number of regulatory T cells in the peripheral blood is reduced.[101] Corticosteroid therapy reverses the population densities in the liver and peripheral blood probably because of treatment-induced improvement in the degree of hepatic inflammation.[84]

Regulatory T cells expressing CD39 are decreased in number and function in autoimmune hepatitis.[102] They fail to suppress the production of IL-17 by Th17 cells, and they inadequately hydrolyse pro-inflammatory nucleotides. They also have a propensity to produce IFN-γ or IL-17 when stimulated with recombinant human pro-inflammatory cytokines, IL-6 and IL-1β.[102] Dysfunctional regulatory T cells and phenotypic instability during inflammatory challenge may be key factors that enable autoimmune hepatitis. Importantly, the exact role of regulatory T cells in autoimmune hepatitis remains controversial because some human studies applying highly discriminating markers for this cell population have also demonstrated normal numbers and functions in peripheral blood.[91]

NKT cells

NKT cells are resident cells within the liver that have surface markers and functions that are shared by T lymphocytes and natural killer cells (Table 1).[103, 104] NKT cells activate B cells, T cells, natural killer cells, and other NKT cells,[105] and they can modulate this response by promoting the differentiation of regulatory T cells.[104, 106] NKT cells express CXCR6, and they migrate in hepatic sinusoids in response to the secretion of CXCL16 (Figure 1).[30, 107] CXCL16 is produced by sinusoidal endothelial cells, cholangiocytes, dendritic cells and macrophages.[30, 107, 108] Its production is increased in chronic liver injury, especially in cirrhosis[30] and the expression of CXCL16 is up-regulated by TNF-α and IFN-γ.[109] CXCL16 can also function as an adhesion molecule to anchor NKT cells to dendritic cells.[110] CXCR6 is expressed by CD4+ cells, CD8+ cells and natural killer cells in addition to NKT cells,[107, 110, 111] and CXCL16 can attract a wide array of effector cells that bear this receptor. The secretion of TNF-α and IFN-γ by NKT cells can increase the expression of CXCL16 and attract more NKT cells and other pro-inflammatory cells in a positive feedback loop.[30, 109]

Chemokine ligands and receptors in autoimmune hepatitis

Multiple chemokine ligands are released during liver injury, and the intensity and nature of the chemokine response can reflect the severity and location of the tissue damage.[43, 44, 112] Individual chemokines lack disease- and organ-specificity, but some chemokines may predominant over others in certain types of liver disease.[28] The principal chemokine ligands that influence the hepatic migration of the immune cell populations implicated in immune-mediated liver disease are CXCL9, CXCL10, CXCL11, CCL20, CXCL12 [stromal cell-derived factor 1 (SDF-1)] and CX3CL1 (fractalkine) (Table 2). Preliminary studies in autoimmune hepatitis, PBC and primary sclerosing cholangitis (PSC) have suggested that serum levels of eotaxin-1 (CCL11), eotaxin-3 (CCL26) and CCL11 (MDC) have diagnostic value.[28] Chemokine assessments are unlikely to have the disease specificity that would enhance the yield of currently available diagnostic instruments,[1, 113, 114] and chemokines are likely to emerge mainly as markers of inflammatory activity, disease severity and treatment response and as therapeutic targets.

Table 2. Principal chemokines and receptors associated with hepatic inflammation
ChemokineChemokine sourceChemokine receptorReceptor-bearing cellsPathological features
  1. AIH, autoimmune hepatitis; ELR motif, glutamic acid-leucine-arginine motif in CXCL10; HCV, hepatitis C virus; HSC, hepatic stellate cells; IFN-γ, interferon-gamma; IP-10, interferon-γ inducible protein 10; ITAC, interferon-induced T cell alpha chemoattractant; LARC, liver activation-regulated chemokine; MIG, monokine-induced by IFN-γ; NK cells, natural killer cells; NKT cells, natural killer T cells; PBC, primary biliary cirrhosis; SDF-1, stromal cell-derived factor; Th1, T helper lymphocytes; UDCA, ursodeoxycholic acid.

CXCL9[25] (MIG)Hepatocytes[68]CXCR3[115]

Th1, Th17 cells[39, 44]

Dendritic cells[116]

Chemoattractant for T lymphocytes[115]

Increased in PBC and AIH[27], [115]

Increased in HCV-cryoglobulinemia[117]

Chemoattractant for regulatory T cells[84]

CXCL10[25] (IP-10)



Dendritic cells[25]

NK, NKT cells[22]


CXCR3[22], [25]

Immune cells[25]

Endothelial cells[25]

Chemoattractant for immune cells[25, 59, 118]

Associated with progressive fibrosis[68], [123]

Promotes angiogenesis if ELR+[22, 25, 124]

Participates in ‘amplification loop’[22]

Mixed cryoglobulinemia in HCV[121]

CXCL11[125] (ITAC)




CXCR3[125]Activated T cells[125]

Chemoattractant for activated T cells[117]

Increased in HCV-cryoglobulinemia[117]

Undetectable in PBC[115]

CCL20[128] (LARC)




CCR6[126, 128]

Lymphocytes[126, 127]

Dendritic cells[130]

Pro-inflammatory, pro-fibrotic effects[129]

Up-regulated by TNF-α[128]

Correlates with severity[129, 131, 132]

CXCL12[134] (SDF-1)

Biliary cells[133]

Progenitor cells[133]



T and B cells[134, 140]

Cholangiocytes[139, 140]

Monocytes[134], [140]

Perivascular cells[141]

Increases B cell maturation[134]

Attracts T cells, B cells, monocytes[134]

Activates HSC[137, 138]

Protective effects in mouse model[136]

CX3CL1[159] (fractalkine)


Hepatocytes[152, 153]



Monocytes[159] T lymphocytes[159]

Kupffer cells[156]

Natural killer cells[160]

Migration and adhesion of leucocytes[154]

Anti-apoptotic properties[153, 155]

Anti-fibrotic effects[155], [156]

Increased with cholangitis in PBC[168]

Eotaxin-3[28] (CCL26)Endothelium[169]CCR3[170]Eosinophils[28]

Increased in autoimmune hepatitis[28]

Possible diagnostic specificity[28]


CXCL9 (MIG) is induced by IFN-γ, and it attracts T lymphocytes bearing the receptor, CXCR3 (Table 2).[115] CXCL9 is secreted by monocyte-derived dendritic cells within the liver,[116] and it is a pro-inflammatory cytokine that attracts Th1 and Th17 cells to the liver.[39, 44, 67] CXCL9 and CXCL10 are both increased in patients with PBC compared to normal individuals, whereas CXCL11, which shares the same receptor as CXCL9 and CXCL10 (CXCR3), is undetectable.[115] CXCL9 and CXCL10 are also increased in autoimmune hepatitis, and both chemokines are associated with disease severity.[27] Furthermore, circulating levels of CXCL9 and CXCL11 have been associated with vasculitis in patients with chronic hepatitis C and mixed cryoglobulinemia.[117] The expression of CXCL9 is up-regulated in hepatic sinusoidal endothelial cells in patients with autoimmune hepatitis, and regulatory T cells expressing CXCR3 can migrate to the liver as a protective countermeasure to the immune response.[84]

The pro-inflammatory effects associated with CXCL9 improve with therapy. In PBC, treatment with ursodeoxycholic acid decreases the expression of CXCR3 and the ligands, CXCL9 and CXCL10.[115] In autoimmune hepatitis, improvement in the laboratory indices of liver inflammation after corticosteroid therapy is associated with reductions in the levels of these same ligands.[27] These observations suggest that the pro-inflammatory chemokines might be useful as barometers of inflammatory activity, indices of disease severity, and targets of treatment.


CXCL10 [IFN-γ inducible protein 10 (IP-10)] is a small (10 kDa) protein that is induced by IFN-γ and secreted by monocytes, neutrophils, endothelial cells, fibroblasts, dendritic cells, CD4+ lymphocytes, CD8+ lymphocytes, natural killer cells (NK) and NKT cells (Table 2).[22, 25] CXCL10 is a promoter of the pro-inflammatory Th1 response,[22] and its function is affected by a three amino acid motif consisting of glutamic acid, leucine and arginine that is encoded as ELR. Molecules without the ELR motif are associated with poor neovascularisation during tissue repair.[22, 25]

CXCL10 binds to CXCR3 which is expressed on immune cells and resident endothelial cells,[25] and it regulates the immune response by recruiting monocytes, T lymphocytes, eosinophils and NK cells to sites of tissue damage.[118] Multiple immune-mediated diseases, including systemic lupus erythematosus, rheumatoid arthritis, systemic sclerosis and autoimmune thyroid disease, have increased serum and tissue concentrations of CXCL10,[25] and serum levels of CXCL10 correlate with disease activity in systemic lupus erythematosus.[119]

Serum concentrations of CXCL10 are higher in patients with autoimmune hepatitis, PBC and chronic viral hepatitis than in normal individuals, and these levels have correlated with the serum levels of the laboratory tests reflective of liver inflammation.[120] Patients with chronic hepatitis C and mixed cryoglobulinemia have increased serum levels of CXCL10, and concentrations are highest in individuals with concurrent autoimmune thyroiditis.[121, 122] The serum CXCL10 level has been proposed as a marker of progressive fibrosis in chronic hepatitis C, especially in African-American patients,[68, 123] and serum and intrahepatic levels of CXCL10 have been inversely correlated with the outcome of interferon therapy.[124] The centrality of CXCL10 in diverse inflammatory liver diseases positions it as a prime candidate for therapeutic targeting.[22, 25]


CXCL11 (ITAC) is secreted by cells within the liver, pancreas and spleen, and it attracts activated T cells bearing CXCR3 (Table 2).[125] CXCL11 has a higher affinity for CXCR3 than CXCL9 and CXCL10, and it is a more potent chemo-attractant.[125] Although IL-2 activated T lymphocytes are strongly attracted to this ligand, unstimulated T lymphocytes, neutrophils and monocytes are not.[125] The secretion of CXCL11 is regulated by IFN-γ, and like CXCL9 and CXCL10, it has pro-inflammatory effects. High serum levels of CXCL11 have been demonstrated in patients with chronic hepatitis C with mixed cryoglobulinemia, vasculitis, and autoimmune thyroiditis.[117, 122] CXCL11 is undetectable in PBC in contrast to CXCL9 and CXCL10,[115] and these observations support the hypothesis that patterns of chemokine production may distinguish some liver diseases or predominant in certain clinical phenotypes.[28]


CCL20 [liver activation-regulated chemokine (LARC) or macrophage inflammatory protein-3α (MIP-3α)] is secreted by macrophages, hepatocytes and hepatic stellate cells, and it attracts lymphocytes and dendritic cells expressing CCR6 (Table 2).[126-130] The hepatic expression of CCL20 is up-regulated by TNF-α,[128] and serum levels correlate with disease severity in animal models of toxic liver injury and patients with acute and chronic inflammatory liver diseases.[129, 131, 132] The induction of experimental autoimmune hepatitis in a thymectomised mouse model is dependent on CCL20 and its ability to attract T lymphocytes from the spleen.[128] In patients with alcoholic hepatitis, serum and hepatic levels of CCL20 are increased, and they correlate with the degree of fibrosis, portal hypertension, severity scores, and early mortality.[129] In patients with chronic hepatitis C, serum levels of CCL20 are associated with serum aminotransferase concentrations and histological severity,[132] and in patients with diverse liver diseases (including autoimmune hepatitis), intrahepatic expression of CCL20 and CCR6 are increased.[131] CCL20 promotes fibrosis in cultured hepatic stellate cells,[129] and antibodies to CCL20 prevent experimental toxic hepatitis and improve laboratory indices of liver inflammation.[129] The broad-ranging pro-inflammatory and pro-fibrotic properties of CCL20 make this ligand another attractive candidate for therapeutic targeting.


CXCL12 (SDF-1) is produced by ductal plate cells in the foetal liver, and these cells are progenitors of the biliary epithelium.[133, 134] CXCL12 also promotes the differentiation and maturation of B lymphocytes, and it attracts T lymphocytes and monocytes to sites of tissue injury and inflammation.[134-136] Exposure of hepatic stellate cells to CXCL12 increases their proliferation in a dose-dependent fashion,[137] and CXCL12 also induces hepatic stellate cell contraction which may in turn contribute to hepatic fibrosis and portal hypertension.[138]

CXCR4 and CXCR7 are the receptors for CXCL12 (Table 2).[139-141] CXCR4 is expressed on hematopoietic progenitor cells, cholangiocytes, T and B lymphocytes and monocytes,[139, 140] and CXCR7 is expressed mainly on the endothelium of venules and the smooth muscle cells of arterioles.[141] The expression of CXCR4 on malignant cells, including hepatocellular carcinoma, has implicated the CXCL12/CXCR4 axis as a possible determinant of tumour growth and invasion.[142, 143] The CXCL12/CXCR4 axis has also been ascribed a protective effect on the liver in a murine model of chronic liver injury,[136] and it has been associated with improved clinical outcome in patients with hepatic metastases from colorectal cancer.[144] The deleterious and protective effects ascribed to CXCL12 may relate to the interactions of CXCR4 and CXCR7 on CXCL12 signalling either as single or joint receptors.[145] Both CXCR7 and CXCR4 are required to form a receptor unit for CXCL12 signalling in some cells, whereas the receptor roles of CXCR4 and CXCR7 may be separate or switched in other cells.[145]

Biliary epithelial cells secrete CXCL12, and the increased expression of this ligand in interlobular and septal bile ducts and proliferated bile ductules can recruit inflammatory and immune cells to the inflamed liver.[134] The CXCL12/CXCR4 axis has been implicated in the migration of immune cells to the liver in patients with chronic hepatitis B and C, PBC, PSC and autoimmune hepatitis.[134, 146, 147] CXCL14 has been proposed as a natural inhibitor of CXCL12, and CXCR7 may also naturally modulate this pro-inflammatory axis.[148] Small peptides have already been designed to block CXCR4 affinity for CXCL12,[149] and small neutralising antibodies against CXCL12 can inhibit the chemokine function.[150] The uncertainty and complexity of the CXCL12 signalling pathways in different cells and organs suggest that therapeutic targeting of CXCL12 and its receptors must be well studied before its clinical application in human liver disease.[145]


CX3CL1 (fractalkine) is expressed by hepatocytes, hepatic stellate cells, biliary endothelial cells, neurons, and epithelial cells of the lung, intestine and kidney (Table 2).[151-153] CX3CL1 can act as a soluble or membrane bound ligand, and it is involved in the migration and adhesion of leucocytes, especially monocytes.[153, 154] CX3CL1 may also have anti-apoptotic properties which can favour the survival of inflammatory and immune cells.[153, 155] Like other chemokines, CX3CL1 may have pro-inflammatory and protective effects, and in animal models of liver injury, it prevents hepatic fibrosis.[155, 156] The cognate receptor of CX3CL1 is CX3CR1 which is expressed on monocytes, Kupffer cells, natural killer cells, T lymphocytes and smooth muscle cells.[153, 156-160]

High plasma levels of CX3CL1 have been described in patients with chronic hepatitis C co-infected with human immunodeficiency virus, and the levels have been associated with advanced fibrosis and histological activity.[161] The intrahepatic expression of CX3CR1 is increased in patients with chronic hepatitis C, especially in individuals with advanced hepatic fibrosis, and CX3CL1 may support homoeostatic mechanisms that reduce the extracellular matrix by suppressing the tissue inhibitor of metalloproteinase-1 (TIMP-1).[162]

CX3CL1 and CX3CR1 are both up-regulated during chronic liver injury in portal and lobular areas in human liver tissue, and high levels are also present in regenerating bile ducts during acute severe hepatitis.[163] In PBC, CX3CL1 is a component of the microenvironment recruiting lymphocytes to bile ducts,[152, 159, 164-166] and CX3CL1 expression is up-regulated in injured biliary epithelial cells.[167] Serum levels are high in patients with early stage PBC and severe cholangitis, and the levels parallel the response to therapy with ursodeoxycholic acid.[168] CX3CR1 is expressed on natural killer cells, and the strong expression of CX3CL1 on damaged bile ducts may account for the accumulation of natural killer cells around injured bile ducts in biliary attresia.[160] This observation supports the hypothesis that the innate immune system is involved in some forms of bile duct injury.

The CX3CL1/CX3CR1 axis is another feasible therapeutic target in chronic inflammatory liver disease, especially in PBC.[152, 153, 168] The uncertainty of the full range of effects associated with this axis and its known opposing pro-inflammatory, anti-apoptotic and anti-fibrotic effects challenge the design of an effective intervention.[153] Although the CX3CL1/CX3CR1 axis may attract monocytes to sites of liver injury and enrich the inflammatory infiltrate within the liver, this axis may also protect hepatocytes from apoptosis,[153, 155] suppress TIMP-1 activity,[162] reduce net collagen deposition[162] and inhibit Kupffer cell production of TGF-β and the subsequent activation of hepatic stellate cells.[156]


The chemokine, eotaxin-3 (CCL26), which recruits eosinophils to sites of inflammation, is produced in the vascular endothelium,[169] and it cognate receptor is CCR3 (Table 2).[170] Eotaxin-3 is significantly increased in the serum of patients with autoimmune liver disease (autoimmune hepatitis, PBC and PSC) compared to normal individuals or patients with chronic hepatitis C, whereas the serum level of MDC (CCL22) is significantly lower.[28] Furthermore, PSC can be distinguished from autoimmune hepatitis and primary biliary cirrhosis by having a significantly higher serum level of eotaxin-1 (CCL11), another eosinophil-specific chemokine.[28] These preliminary findings suggest that diagnostic algorithms based on relative serum levels of particular chemokines may be possible. The role of these chemokines in autoimmune liver disease is unknown, but their distinctive serum elevations suggest an allergic component of the pathogenic process that may provide another therapeutic target.[28]

Chemokines and hepatic fibrosis

Key components of tissue repair are neovascularisation and collagen deposition.[21, 171, 172] Liver-infiltrating T lymphocytes activate hepatic stellate cells and promote their transition to myofibroblasts.[10, 11, 21, 173] Hepatic stellate cells secrete numerous chemokines (CCL3, CCL5, CCL12, CXCL8, CXCL9, CXCL10 and CXCL12) that have mainly pro-inflammatory and pro-fibrotic effects.[21, 137, 174] CXC chemokines with the ELR motif induce angiogenesis, whereas CXC chemokines without the ELR motif suppress neovascularisation.[21, 171, 175]

CCL2 production is increased in resident liver cells,[176, 177] and CCL2 attracts monocytes, dendritic cells, Th1 lymphocytes, and hepatic stellate cells that express the receptor, CCR2 (Table 3).[21, 178] CCR2 interacts with the ligands, CCL7, CCL8 and CCL13, and these promiscuous interactions broaden the inflammatory response and intensify the fibrotic process.[21] CCL3 and CCL5 have the same cognate receptors (CCR1 and CCR5), and these chemokine interactions promote fibrosis in murine models by favouring immune cell migration and activation of hepatic stellate cells.[179, 180] Hepatic fibrosis is also affected by the chemokines that govern the migration of Th1 and Th2 lymphocytes.[21] Th1 lymphocytes expressing CXCR3 and CCR5 produce IFN-γ and IL-12 that inhibit hepatic fibrosis, whereas Th2 lymphocytes expressing CCR3 and CCR4 produce IL-4 and IL-13 which increase hepatic fibrosis.[181]

Table 3. Principal chemokines and receptors associated with hepatic fibrosis
ChemokineReceptorFibrotic EffectsMechanisms
  1. HSC, hepatic stellate cells; NKT cells, natural killer T cells; IFN-γ, interferon-gamma; IL, interleukin.


Secreted by resident liver cells[176, 177]

Attracts dendritic cells, HSC[178]

Promotes fibrosis in animal models[178]





Same receptors as CCL5[21]

Promotes hepatic fibrosis in mice[179]

Contributes to HSC activation[179]

Increases immune cell infiltration[179]





Same receptors as CCL3[21]

Promotes hepatic fibrosis in mice[179, 180]

Contributes to HSC activation[179]

Increases immune cell infiltration[179]



Attracts liver-infiltrating CD8+ cells[183]

Induces pro-inflammatory cytokines[183]

Promotes liver fibrosis in mice[183]


Increased in severe fibrosis[68, 123]

Promotes fibrosis in animal model[182]


Increases intrahepatic NKT cells[30]

Promotes NKT cell secretion of IL-4, IFNγ, IL-13[30, 107]

Attracts lymphocytes to bile ducts[111]

Induces lymphocyte migration to damaged hepatocytes[111]

CX3CL1CX3CR1 Anti-fibrotic

Protective against fibrosis by regulating apoptosis[153]

Regulates survival of monocytes[21, 155]

Prevents progressive hepatic fibrosis[155]

The degree of hepatic fibrosis can be modulated in a murine model by neutralising the CXCL10 ligand of CXCR3 and altering the balance between Th1 and Th2 lymphocytes.[182] High serum and tissue levels of CXCL10 have been associated with severe liver inflammation and hepatic fibrosis,[68, 123] and CXCL4, another ligand of CXCR3, promotes hepatic fibrosis in a murine model by attracting liver-infiltrating CD8+ lymphocytes and activating hepatic stellate cells.[183] Other pro-fibrotic chemokines are CCL3 and CCL5 which each attract cells expressing CCR1 or CCR5 (Table 3).[179, 180] CXCL16 and its cognate receptor, CXCR6, control the migration of NKT cells into liver sinusoids and in this fashion, modulate hepatic fibrosis.[30, 107, 111] Invariant NKT cells that accumulate early in liver injury express CXCR6, and they promote hepatic fibrosis by secreting IL-4 and IL-13 and activating hepatic stellate cells.[30, 53, 54] CX3CL1 is the principal cytokine that has been associated with an anti-fibrotic effect, and it does so by modulating the differentiation and apoptosis of liver-infiltrating monocytes in a murine model of hepatic fibrosis.[155]

Chemokine-directed therapy

Liver damage and progressive hepatic fibrosis are consequences of diverse cellular and molecular interactions that are sustained by inflammatory activity.[10, 11, 13, 35] The common feature within these pathogenic pathways is their dependence on the chemokine network.[10] This network is characterised by the redundancy, plasticity and promiscuity of its individual components. These attributes allow a relatively small number of ligands and receptors to participate in the trafficking and positioning of diverse effector cells at sites of tissue injury. They also complicate efforts to identify a single critical chemokine that can be targeted. Furthermore, chemokines can form multimeric and heterodimeric structures that complicate assessments of their biological function,[21] and efforts to neutralise any one ligand or receptor that is involved in multiple homoeostatic mechanisms can result in unanticipated and serious consequences.[31] In autoimmune hepatitis, chemokine therapy is feasible, but the key targets are undefined and the strategy untested.

Current therapies in autoimmune hepatitis succeed by suppressing liver inflammation and tissue damage, and in this fashion they indirectly reduce chemokine production (Figure 2).[10, 11, 13, 184-186] This strategy has already been documented in patients with autoimmune hepatitis who respond to corticosteroid therapy,[27] patients with PBC who respond to treatment with ursodeoxycholic acid,[115] and patients with chronic hepatitis C who respond to anti-viral medication.[124] The next generation purine antagonist, mycophenolate mofetil, is emerging as a frontline and salvage therapy for autoimmune hepatitis,[187-189] and it can induce the apoptosis of activated lymphocytes, suppress the proliferation of lymphocytes by inhibiting the synthesis of deoxyribonucleic acid and impair the expression of adhesion molecules.[104, 190] These combined effects may in turn indirectly diminish the production of chemokines. Mycophenolate mofetil may also preserve the suppressive activity of regulatory T cells and facilitate their expansion after antigenic stimulation (Figure 2).[191-194] A logical extension of broad spectrum immunosuppressive therapy that enhances regulatory T cell function would be the adaptive transfer of these cells.[39, 195-197]

Figure 2.

Feasible therapeutic interventions to reduce chemokine production in autoimmune hepatitis. Treatments that reduce liver inflammation (Conventional Immunosuppression), modulate regulatory T cell (Treg) function (Regulatory T cell Enhancement), directly block or neutralise chemokine actions (Direct Chemokine Blockade or Neutralisation), or indirectly suppress the secretion of chemokine ligands (Indirect Chemokine Suppression) have been used in experimental models or humans with immune-mediated diseases, and they are candidates for study in autoimmune hepatitis. Pharmacological agents can indirectly suppress the production of key chemokine ligands and tumour necrosis factor-alpha (TNFα), whereas molecular agents such as monoclonal antibodies (monoclonal AB) can directly neutralise the chemokines.

Regulatory T cells can be expanded and maintained in cell culture for adaptive transfer,[198] and they have been able to suppress histological activity in a murine model of experimental autoimmune hepatitis (Figure 2).[199] Adaptive transfer of regulatory T cells is not chemokine-directed therapy, but regulatory T cells do modulate chemokine production.[84] Furthermore, their adaptive transfer in autoimmune hepatitis has a strong rationale and investigational basis.[196, 199] Similar modifications in the chemokine network can be anticipated in evolving therapies that are designed to disrupt the cytokine pathways.[104, 200]

Monoclonal antibodies to CD20 (rituximab) target mainly activated B lymphocytes and monoclonal antibodies to TNF-α (infliximab) can inhibit the differentiation and proliferation of liver-infiltrating cytotoxic T lymphocytes.[104, 200] Both interventions may affect the production of cytokines and chemokines, and each has been effective in small observational studies of patients with steroid-refractory autoimmune hepatitis or steroid-intolerance.[201, 202] Therapeutic manipulations of key counter-regulatory homoeostatic mechanisms such as the cytokine pathways also have risks, and these risks include systemic infection and altered immune reactivity.[202, 203] Infliximab has been implicated in the production of a clinical syndrome that resembles autoimmune hepatitis,[204-206] and experiences with this drug indicate the hazards that may be encountered when manipulating complex interactive pathways that can affect host defence mechanisms.[203]

The chemokine-directed therapies which have had preliminary success in treating animal models and patients with diverse immune-mediated diseases have included medications that repress chemokine production, antibodies that neutralise specific chemokines and peptides that block receptor activity (Figure 2). Agonists of PPAR-γ activity, such as rosiglitazone, can reduce CXCL10 production by inhibiting transcriptional activity of nuclear factor-κB in Grave's disease;[22, 207-210] fenofibrate, a ligand of PPAR-γ, can repress the expression of genes encoding CXCL10, CCL2 and CCL20 and improve experimental Crohn's disease in mice;[22, 211] and methimazole can reduce the up-regulation of TNF-α, impair INF-γ production and indirectly decrease CXCL10 secretion in autoimmune thyroid disease.[212]

Neutralising antibodies against individual chemokines provide a direct and highly selective method to interrupt pro-inflammatory and pro-fibrotic pathways (Figure 2). Preliminary studies indicate that monoclonal antibodies to CXCL10 improve the response of rheumatoid arthritis to therapy with methotrexate,[213] and they reduce hepatic fibrosis in a mouse model of chronic toxic liver injury.[182] Neutralising antibodies to CXCL16 increase the survival of mice with immune-mediated acute severe liver injury[214]; and monoclonal antibodies to CCL20 improve the laboratory tests and the hepatic expression of pro-inflammatory and pro-fibrotic genes in experimental models of acute, chronic and acute-on-chronic toxic liver injury.[129]

Nanobodies are single-domain antibody fragments that are devoid of light chains and that bind selectively to their target.[215] Their emergence has increased the repertoire of therapeutic options. Nanobodies directed against CCL2, CCL5, CXCL11 and CXCL12 block receptor binding with high affinity, and they are available for study in a wide range of immune-mediated and inflammatory diseases (Figure 2).[150] Chemokine-directed therapy is ongoing in patients with rheumatoid arthritis, co-infection with human immunodeficiency virus and hepatitis C virus, endometriosis, and cardiovascular disease.[21, 22, 213] A clinical trial of humanised antibodies to CXCL10 is underway in PBC, and the results from this study could generate investigations in autoimmune hepatitis.[29]

Peptides that block the chemokine receptor can disrupt chemokine function, and they constitute another feasible therapeutic option in autoimmune hepatitis (Figure 2). Two peptides have been designed and synthesised using the N-terminal region of CXCL12, and they simulate docking of CXCL12 with its natural receptor, CXCR4. The peptides block CXCL4 binding to CXCL12, and they have specificity and affinity for this receptor.[149] Similarly, treatment with a recombinant analogue to CCL5 that is an antagonist to CCR5 and CCR1 has inhibited the migration and proliferation of cultured hepatic stellate cells and reduced chemokine and collagen production.[180] Treatment has also prevented and reversed hepatic fibrosis in an experimental murine model of chronic liver injury.[180] Strategies based on the inhibition of chemokine function by neutralising antibodies or blockade of a critical chemokine receptor have not been compared, but the availability and variety of antibodies against the chemokine ligands favour their consideration in studies of autoimmune liver disease.


Chemokines are essential for mobilising, directing, and positioning the inflammatory and immune cells involved in the pathogenesis of acute and chronic inflammatory liver disease. Chemokines are also involved in orchestrating the protective and reparative processes within the injured liver by modulating the apoptosis of effector cells, attracting regulatory T cells to sites of tissue damage, activating hepatic stellate cells and facilitating fibrotic repair. They constitute a redundant and promiscuous network in which any single component may have inhibitory and stimulatory functions that are vital for the preservation or restoration of the normal innate and adaptive immune responses. Preliminary animal studies and human experiences with chemokine-directed therapies for liver (viral, alcoholic, toxic and immune-mediated) and nonliver (rheumatic, cardiovascular, gynaecological and malignant) diseases have encouraged further study, and this intervention must be considered in autoimmune hepatitis. The appropriate chemokine target is undefined in autoimmune hepatitis, but CXCL10, CCL20 and CCL5 are prime candidates for study. The caveat to this intervention is to avoid the unwinding of a thread that holds the homoeostatic fabric together.


Guarantor of the article: Albert J. Czaja, MD.

Author contributions: The review article was conceived, researched, designed and written by Dr Albert J. Czaja without other writing assistance, research support or financial aid from a funding agency or institution. The two figures and the three tables were developed, constructed or drawn by Albert J. Czaja, MD for this review. He has reviewed the entire final version of the manuscript prior to its submission.


Declaration of personal and funding interests: None.