Alcohol-associated liver disease (ALD) refers to liver damage caused by the chronic excessive drinking of alcohol (1). ALD is often observed in patients who have pre-existing Alcohol Use Disorder (AUD). AUD is a mental disorder assigned by the Diagnostics and Statistics Manual (DSM) (Fifth Edition) and is characterized by compulsive excessive alcohol consumption over an extended time with associated loss of control and negative emotions (2). Alcohol consumption has been on the rise in the United States and worldwide; in 2019 there were 14.1 million adults in the United States that met the criteria for AUD (3). Interestingly, only 10-20% of chronic and heavy drinkers eventually progress to more severe and permanent forms of ALD, such as hepatitis and cirrhosis (4). There is likely a multitude of disease modifiers responsible for this progression and it still remains a challenge to identify this subset of AUD patients who will develop ALD (5). Such predictive ability could enable earlier interventions that may significantly reduce the proportion and likelihood of AUD patients who develop alcohol-associated hepatitis (AH).

Alcohol-associated hepatitis (AH) is an advanced and acute form of ALD that is attributed to elevated pro-inflammatory status in response to excessive alcohol intake and various other altered domains involved in liver injury (1). Severe alcohol-associated hepatitis (SAH) is characterized by a large systemic inflammatory response syndrome (SIRS) leading to severe liver inflammation and dysfunction with an associated high short-term mortality of 30-50% (6). Understanding the underlying cytokine cascading pathways holds the potential to predict the development and severity of ALD regardless of the changes indicated by the clinical measures. IL-17 and IL-22 are two particular cytokines involved in inflammation and cell death in ALD. While they have been implicated in alcohol associated liver disease, few studies have published on their role in AH (7–9).

Chronic and excessive alcohol intake causes gut dysfunction and increased gut permeability (10). This leads to the release of bacterial endotoxins, such as lipopolysaccharide (LPS), which bind to toll-like receptor 4 (TLR4) on the liver initiating a cascade of pro-and anti-inflammatory events (11–13). Alcohol metabolism in the liver further contributes to this inflammatory response through the production of acetaldehyde and concomitant reactive oxygen species (ROS) (14). This gut dysfunction and alcohol metabolism initiate a cascade of events in the liver involving the production of pro- and anti-inflammatory cytokines, including tumor necrosis factor-α (TNF-a), interleukin 1 (IL-1), interleukin 6 (IL-6), and transforming growth factor-beta (TGF-β) (15–18). These in turn lead to the differentiation of specific CD4⁺ T cells (namely, Th-17), which produce interleukin 17 (IL-17), interleukin 17F (IL-17F), interleukin 21 (IL-21), interleukin 22 (IL-22), and granulocyte-macrophage colony-stimulating factor (GM-CSF) (19–21). These in turn further promote neutrophil recruitment, inflammation, and fibrosis in ALD. This study evaluated the role of IL-17 and IL-22 as markers of inflammation in various stages of alcohol associated disease; spanning AUD without liver injury, AUD with liver injury (or Early Stage ALD), and ultimately non-severe and severe AH compared with the healthy controls.

The functional aspects of so-called pro- and anti-inflammatory cytokines are so interwoven, that it is often difficult to separate a particular cytokine into one or the other of these two classifications. Such is the case with IL-17 and IL-22 cytokines that mostly originate from Th-17 cells. In general, IL-22 is considered beneficial, and IL-17 as detrimental in any inflammatory pathology. We see such activity from our findings also in alcohol-associated liver disease. In this study, we found that at baseline, IL-22 levels significantly differentiated AUD from AH patients whereas IL-17 tended to increase in conjunction with liver injury. Several other studies have found definite increases in IL-17 with ALD. For example, Lemmers et al. found that IL-17 plasma levels increased in patients with alcoholic liver disease (including AH) compared with healthy subjects (7). A 2013 study of 21 AH patients with time points at BL, 14 days (W2) and a month (W4) measured IL-17 and IL-22 in plasma using ELISA (9). They found that IL-17 was 1.5 times more elevated in AH compared to healthy controls, which is consistent with the data in this study. There were several other similarities found with this current study. AH IL-17 increased, and IL-22 decreased at W4 compared to BL, although these differences were not statistically different. In addition, like this study, their AH IL-22 data did not differ from healthy controls. Another group studied IL-22 levels in alcohol associated liver disease and found that they correlated with MELD (8). While we did not see a similarly strong positive correlation of IL-22 with MELD, Parfieniuk-Kowerda’s findings are consistent with our findings of positive relationships with other markers of liver damage.

BL IL-22 correlated negatively with BL IL-1β. Since IL-1β is a key cytokine in the pro-inflammatory response, this was the only correlation to directly support the anti-inflammatory nature of IL-22. It can be surmised that the other IL-22 correlations are indicating a response to liver injury, and/or increased IL-22 resistance ( Figure 5 ).

Lu et al. (2017) used human samples from the TREAT consortium to measure IL-17 and IL-22 in the serum of healthy volunteers, heavy drinkers, and AH subjects using a multiplex protein assay. Unlike our study, they found that AH subjects had an increased IL-22 compared to healthy controls. However, similarly to our study, they found IL-22 increased from heavy drinkers to AH. They found no significant differences in IL-17. The authors thought perhaps increased IL-22 resistance was the reason for increased IL-22 in AH subjects. Other studies have shown increases in the cytokines and growth factors that induce IL-22 with increasing injury. Accordingly, since levels of all these IL-22 inducers are increasing, it is reasonable to expect that IL-22 will also increase. It is possible that this is the result of an overwhelming inflammatory response that increases cytokines, such as IL-6, IL-1β, and IL-23 that in turn signal Th-17 to make increased IL-22 ( Figure 5 ). There is some pivot, however, where the levels of the pro-inflammatory inducing cytokines will begin to decrease, and IL-22 will be able to employ its beneficial effects.

Of these two cytokine measures, IL-22 is easier to detect using standard protein assays due to high levels of expression when triggered. And accordingly, statistically significant relationships are more clearly measurable. BL versus W4 IL-22 measures correlate well, so at least in this study, it consistently increases over the spectrum of liver injury and post-treatment. It correlates negatively with IL-1β (a further downstream pro-inflammatory cytokine), correlates positively with liver damage markers (AST: ALT, AST) as well as shows corresponding initiation of pro-inflammatory cytokines such as TNF-α and IL-8 in our findings.

Additionally, the number of years of alcohol drinking was found to be closely associated with IL-22 in non-severe AH patients and was predictive of W4 IL-22 levels in these same subjects, indicating that increased drinking is associated with increases in IL-22. In turn, as liver injury increases, levels of IL-22 will also increase in an attempt to mitigate the inflammation and initiate regenerative processes (33). Stoy et al. also measured IL-22 expressing T cells using FACs and found that patients with a favorable short-term course, as measured by the Glasgow score, had higher levels of these IL-22 producing Th17 cells (9). Thus, the overall notion is that, even though IL-22 is apparently increasing with markers of liver injury, the ultimate outcome of increased IL-22 may be liver regeneration and survival, although our study did not include these endpoints per se.

In this study, as anticipated in the Th-17 cell response, IL-17 increased overall with liver injury in a step-wise progression per the corresponding increase in liver severity. Whereas plasma IL-22 decreased in healthy volunteers and in AUD, albeit increased again in AH, this observation is important. This response could reflect a subtlety in signaling that occurs during alcohol associated inflammation. The decrease in IL-22 in AUD could be a result of increased production of TGF-β by Kupffer cells in the initial inflammatory response to alcohol overconsumption; TGF-β is known to decrease IL-22 production (34). Kim et al. found that serum TGF-β levels increased in alcohol-dependent subjects compared to healthy controls (35). Increasing concentrations of TGF-β were found to inhibit IL-6-induced production of IL-22 in Th17 cells in a model of autoimmune disease (34). Either IL-23 or IL-6 could stimulate IL-22 production independently of each other. Another study has shown liver IL-6 to decrease, though minimally, in AH (36). The results in our study were consistent with other studies that found blood IL-6 increased with increasing severity of alcoholic liver disease (37, 38). It is possible that the lower levels of IL-22 found in AUD patients could have been due to similar TGF-β signaling as a part of the early inflammatory response. Recently, it was found that both Th17 cells and IL-23R expression are increased in patients with AH and Hepatitis B, respectively (39). An increase in IL-23 activity, possibly independent of IL-6, may contribute to the increase in IL-22 seen in AH patients as compared to AUD subjects. Measurement of plasma IL-23 in AUD and AH patients would be helpful in supporting this hypothesis, which could be a next direction of investigation. Additionally, measurement of suppressor of cytokine signaling 3 (SOCS-3) in AUD and AH might be informative, since it is thought to be involved in a negative feed-back loop for IL-22 production to enable signal transducer and activator of transcription (STAT3) signaling (40). Evidence has shown that STAT3 signaling is indeed deficient in early alcohol associated liver disease (41). It may be that the lower levels of IL-22 in AUD would correlate with increased SOCS-3 activity.

IL-22BP is the endogenous decoy ligand for the IL-22/IL-10 dimer. Its lower levels in AH are associated with increased mortality (42). IL-22 was found to have a feedback loop that reduces IL-22R, thus decreasing IL-22 bioactivity. IL-22BP is thought to negate the IL-22 suppression of IL-22R. This could be one of the reasons a high association of IL-22 levels with liver damage markers is seen. There is also a line of thought in the literature that IL-22 is detrimental in liver disease (43, 44). Recent literature has found, in opposing results, that IL-22 is both beneficial (45) in a mouse model but not beneficial (46) in clinical studies of acute on chronic liver failure, which AH is a subset of. However, use of a mouse model for studies on IL-22/IL-22BP is lacking because humans produce several thousand-fold more IL-22BP than mice do (45, 46) and humans produce three IL-22BP isoforms (47) that mice do not. This signaling system is so complex that it is difficult to draw strong conclusions from in vitro and mice experiments.

Currently, the progression from AUD to AH is not well understood due to a lack of comprehensive study of the evolution of inflammation across the spectrum of AUD and ALD. An understanding of these mechanisms could lead to earlier intervention and hopefully prevention of devastating liver disease and its sequelae. We found a substantial association of IL-22 and AST, as well as AST : ALT ratio (both are markers of liver injury progression) both at baseline and post-SOC in the spectrum of ALD, supporting the role of IL-22 in the progression of ALD. IL-17 binds to the IL-17A/C receptors, which are ubiquitously expressed, but are also induced by stress on hematopoietic cells and fibroblasts. The IL-17 receptors most critical in ALD are most likely those expressed on stellate cells (Lemmers et al., 2009). The signal is propagated by transcription factor STAT3 (Arab et al., 2019; Qu et al., 2013). Upon stimulation with IL-6, IL-23, and IL-1β through the AHR receptor Th-17 cells will also release IL-22. IL-22 then binds to the heterodimeric IL-10R2/IL-22R1. IL-10R2 is ubiquitously expressed, but expression of IL-22R1 receptors is largely limited to epithelial cells, which in the case of the liver are hepatocytes. IL-22R1 in the liver is also found on hepatic stellate cells, liver progenitor cells, and some fibroblasts. IL-22 signaling is involved in tissue repair and antimicrobial activity. It can also inhibit apoptosis and induce proliferation through STAT3 signaling (Kong, Feng, Mathews, & Gao, 2013; Kong et al., 2012; Qu et al., 2013).

Recent studies point to the role of hepatic progenitor cells (HPC) in AH (36, 48). It was found that HPC accumulate in AH patients, but they fail to differentiate into hepatocytes. This group found a decrease in STAT3 and pSTAT3 in AH livers and thought that this might be a contributing factor to the lack of regeneration seen in AH. So even while IL-22 might be increasing in AH, the most important downstream activity of the transcription factor STAT-3 was decreasing. This points to dysfunctional IL-22 signaling in AH by increased IL-22 resistance. IL-22 was shown to inhibit epidermal differentiation of keratinocytes (Boniface 2005), but its effect on differentiation of HPC to hepatocytes is not known. Changes in levels of IL-22, which is known to be associated with liver regeneration (33) could be a part of the dysfunctional regenerative response that seems to occur in AH.

The protective effect of IL-22 on the liver was first reported in 2004 (49, 50). Clinical trials have been undertaken for an IL-22 conjugate for treatment of AH. The Phase 2A trial showed benefits with no adverse events (51). Our data suggests that this treatment might be especially beneficial during AUD, when IL-22 levels are lower, to increase their levels and perhaps offset progression to AH. The treatment results in extremely high levels of IL-22 (~700,000 pg/ml after IL-22Fc compared to 200 pg/ml before in healthy volunteers). Therefore, in addition to interplay with IL-22BP and receptors, there is no doubt a dose-dependent effect for IL-22, that could be a potential medical management modality. This may explain why this treatment works and why we found relatively lower IL-22 levels (60 pg/ml) in AH correlating with liver damage and dysfunction markers.

The extensive range of AUD and alcohol associated liver disease investigated in this study showed that IL-22 levels correlated positively with markers of liver injury. Number of years of alcohol drinking was closely associated with IL-22 in non-severe AH patients and predicted W4 IL-22 levels. The effects of IL-22 are predominantly anti-inflammatory in nature and increased IL-22 production is thought to be in response to liver injury and/or IL-22 resistance. At baseline, IL-22 levels significantly differentiated AH patients from AUD patients; regression analysis showed that IL-22 predicted progression from AUD with liver injury to moderate AH. Overall, in the alcohol associated groups, IL-22 at baseline correlated very strongly with AST: ALT ratio at week 4. It seems that the compensatory mechanism of IL-22 is superseding the IL-17 response when ALD is not severe. This could be attributed to the IL-22’s survival factor response in liver (50). In severe AH patients, IL-17 response is more meaningful and correlated with the severity assessment. IL-17 response shows that the pro-inflammatory pathway is well underway in severe AH; and IL-22’s compensatory response could have receded. The accuracy of these predictions will only be verified by large long-term longitudinal studies that would begin by enrolling AUD patients, and then following them through potential progression to liver injury over a long period of time. The balance of these two cytokines produced by the same cell may be the key to predicting progression from AUD to severe liver injury.

The original contributions presented in the study are included in the article/supplementary materials. Further inquiries can be directed to the corresponding author.

The studies involving humans were approved by University of Louisville IRB, and NIAAA CNS IRB. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

VV is the project PI and developed the study design. JF, MS, MLS, DJ and VV performed the sample handling and analyses. VV, MS, JF, DJ, HH, MLS and MK performed the data analysis. JF, MS, MK, AR, DJ, RB and VV interpreted the data analysis. JF, MS, DJ, AR, RB and VV wrote the manuscript. VV, MK, MLS and AR reviewed and scientifically contributed to the manuscript. All authors contributed to the article and approved the submitted version.