Evidence that extracellular HSPB1 contributes to inflammation in alcohol-associated hepatitis

INTRODUCTION

Alcohol-associated liver disease (ALD) has reached a global prevalence of 4.8% and is now the most common cause of death due to excessive alcohol consumption in the United States (US).^{1, 2} In the US, ALD mortality increased 23% between 2019 and 2020, and claimed an estimated 45,290 American lives in 2022.^{1, 3} ALD has also become the leading indication for liver transplantation in the US, with a 63% increase in the number of patients listed for transplant between 2007 and 2017.^{4, 5} The term ALD broadly refers to any type of alcohol-induced liver damage and encompasses steatosis, steatohepatitis, cirrhosis, and hepatocellular carcinoma.⁶ In general, higher levels of alcohol consumption lead to more severe disease, but transitions between types of disease are unpredictable and ALD may go unrecognized until it becomes severe. These factors make it imperative to understand the pathophysiology of ALD in order to develop reliable diagnostics that predict risk of progression and therapies able to prevent severe forms of disease.

Alcohol-associated hepatitis (AH) is the most severe and life-threatening manifestation of ALD.⁷ Mortality in severe AH approaches 50% within three months of diagnosis and rises to 70% in patients that do not respond to corticosteroids.⁸ Patients with AH present with rapid onset signs of liver failure and systemic inflammation.⁹ Liver biopsy in these patients typically reveals steatosis, hepatocyte ballooning, Mallory-Denk hyaline inclusions, fibrosis, and infiltration by innate and adaptive immune cells.⁹ The inflammatory cell infiltrate is accompanied by high levels of pro-inflammatory cytokines such as TNFα (tumor necrosis factor alpha), IL-1β (interleukin-1 beta), and IL-6 (interleukin-6).⁹ These factors suggest that AH is an acute inflammatory disease superimposed on chronic ALD.⁶ They also argue that targeting inflammation could slow or reverse liver damage and decrease AH-associated mortality.

Inflammation results from both direct and indirect effects of alcohol on the liver. In the normal liver, tissue resident macrophages, called Kupffer cells, line the sinusoids and react to microbes, microbial components, and microbial products transmitted to the liver from the gastrointestinal (GI) tract.¹⁰ These microbe-associated molecular pattern molecules (MAMPs) are detected by innate immune receptors on the surface or in the cytosol of macrophages and activate signaling cascades that lead to production and release of cytokines and chemokines. In the absence of alcohol, Kupffer cells are exposed to low levels of MAMPs, and inflammation is limited in both extent and duration.¹¹ In contrast, alcohol can directly activate production and release of AH-associated cytokines, such as TNFα, from Kupffer cells or macrophages recruited into the liver.¹² Alcohol also damages the gut mucosal barrier, allowing increased translocation of MAMPs to the liver where they activate production of the same suite of pro-inflammatory cytokines and chemokines in Kupffer cells or recruited macrophages.¹³ Finally, ethanol stressed hepatocytes can contribute to macrophage activation through release of damage-associated molecular pattern molecules (DAMPs). DAMPs are intracellular proteins that are naturally found at high levels inside cells or that are rapidly upregulated in response to stress. These danger signals are actively secreted from ethanol stressed cells or released into the extracellular environment secondary to cell death.¹⁴ In the extracellular environment, DAMPs activate pattern recognition receptors (PRR) on macrophages, which leads to production of the same cytokines and chemokines that macrophages produce in response to MAMPs.¹⁵

One DAMP of particular interest in AH inflammation is HSPB1 (Heat Shock Protein Family B (Small) Member 1; also known as Hsp25/27). HSPB1 protein levels are rapidly increased in a wide range of cells subjected to MAMPs or oxidative stress.^16–18 In these cells, HSPB1 directly and indirectly counteracts cell damage and programmed cell death mechanisms, allowing them to survive in otherwise deadly conditions.^{19, 20} Although HSPB1 has not been systematically evaluated in ALD, it has been implicated in alcohol-associated disease in other ways. Hepatocytes are primarily responsible for ethanol metabolism in the liver and this metabolism generates reactive oxygen species that accumulate and cause oxidative stress and cell damage with high or prolonged alcohol use.²¹ There is a large body of literature demonstrating that HSPB1 acts to prevent programmed cell death in cells experiencing oxidative stress.²² A role for HSPB1 in cellular responses to ethanol is further supported by studies in astrocytes, which showed that ethanol challenge leads to increased HSPB1 gene expression in in these cells.²³ Extracellular HSPB1 is of particular interest in ALD since patients with hepatocellular carcinoma have increased levels of HSPB1 and anti-HSPB1 antibodies in their blood; increases that were linked to a history of alcohol use in some studies.^24–28 HSPB1 is reportedly released from stressed cells, including hepatocytes to activate PRR on adjacent cells.^29–32 HSPB1-mediated PRR activation then leads to production and release of the AH-associated cytokines TNFα and IL-1β from macrophages.^{32, 33}

These factors led us to examine HSPB1 in AH, with the goal of understanding the role that this protein plays in inflammation and liver damage caused by alcohol consumption. We found that patients with AH have significantly higher levels of HSPB1 in both their blood and liver than liver disease free controls or patients with alcohol-associated cirrhosis (AC). This suggested that increases in HSPB1 were related to the higher level of inflammation in AH. In a murine model of chronic alcohol consumption (CAC), HSPB1 was increased in the liver prior to significant liver damage, meaning that it does not result from inflammation. Similar to our human and animal studies, HSPB1 protein was rapidly increased in and released from hepatocytes stressed with ethanol in vitro. In the extracellular environment, HSPB1 promoted inflammation by triggering TNFα release from macrophages. The importance of HSPB1 in ethanol-activated inflammation was reinforced by the fact that anti-HSPB1 antibody almost completely mitigated TNFα release from macrophages treated with media conditioned by ethanol stressed hepatocytes. Together, these findings establish a key role for HSPB1 in the inflammation of AH and identify it as a potential biomarker and therapeutic target for this intractable disease.

RESULTS

Circulating HSPB1 is increased in ALD patients with the highest levels in severe AH

Predicting the course of ALD in a given patient is extremely challenging. Nearly all patients consuming high levels of alcohol develop steatosis and many never progress to more severe disease.⁹ However, patients can progress to life-threatening disease quickly and without any prior indication. This makes development of biomarkers that can identify patients at risk of progression vital in ALD. The ideal biomarker would be amenable to non-invasive monitoring, differentiate liver disease caused by alcohol use from other causes, predict risk of acute destabilization, and be targetable to prevent severe disease.

Identification of HSPB1 as a potential DAMP and biomarker of inflammation in AH led us to retrospectively measure HSPB1 in serum from healthy controls (HC), heavy alcohol consumers (HA), patients with AC, and patients with AH (Table 1). Levels of HSPB1 were similar in the groups without liver disease (HC and HA). However, HSPB1 was significantly higher in patients with ALD (AC and AH) versus those without liver disease (Fig. 1A). Comparison of serum HSPB1 concentrations between AC and AH further revealed that HSPB1 was significantly higher in AH than AC. We next investigated relationships between disease severity and HSPB1 in the ALD groups (Table 2). There was no correlation between HSPB1 and disease severity in AC and no difference between HSPB1 in patients with severe versus mild-moderate AC (Fig. 1B, C). However, within the AH group, HSPB1 increased as disease severity increased and levels of HSPB1 were significantly higher in severe versus mild-moderate disease (Fig. 1D, E). HSPB1 levels were not related to age or sex, despite literature describing HSPB1 as an estrogen responsive gene (Fig. 1F, G).³¹ Thus, high circulating HSPB1 is a feature of severe AH.

• Download figure
• Open in new tab

Figure 1. Circulating HSPB1 is increased in ALD patients.

A, Violin plot of HSPB1 concentration in patient sera measured by ELISA. B, Linear regression of HSPB1 concentration and Child-Pugh score of AC patients. C, HSPB1 concentration in AC patients with mild-moderate (≤8) vs. severe (>8) disease scores (Child-Pugh). D, Linear regression of HSPB1 concentration vs. MELD score of AH patients. E, HSPB1 concentration in AH patients with mild-moderate (≤22) vs. severe (>22) disease scores (MELD). F, Linear regression of HSPB1 concentration vs. age in all patients. G, HSPB1 concentration in male vs. female patients. H, Phosphorylated HSPB1 concentration in patient sera measured by ELISA. All data plotted as mean ± SEM. A,H. Kruskal-Wallace test with a Dunn’s multiple comparisons post-hoc test. B, D, F. Simple Linear Regression. C, E, G. Mann-Whitney test. HC, healthy control; HA, high alcohol consumer; AC, alcohol-associated cirrhosis; AH, alcohol-associated hepatitis. n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001.

View this table:

• View inline
• View popup
• Download powerpoint

Table 1. Patient demographics in studies of circulating HSPB1.

View this table:

• View inline
• View popup
• Download powerpoint

Table 2. Patient demographics and clinical data in studies of circulating HSPB1 and disease severity.

Several post-translational modifications have been described for intracellular HSPB1, including phosphorylation. Human HSPB1 contains three serines capable of accepting a phospho-group, serine 15, serine 78, and serine 82.³⁴ Phosphorylation of these serines determines oligomerization and distribution of HSPB1 inside cells, but the role of phosphorylation in extracellular functions is far less clear. Serine 15 phosphorylation is associated with HSPB1 distribution to insoluble cell components during hydrogen peroxide induced oxidative stress, making it unlikely to be found in extracellular HSPB1 during ethanol stress.¹⁸ This made us focus on HSPB1 phosphorylation at serines 78 and 82 in the patient blood (Table 3). Of the tested samples, only 3 of 57 were positive for phosphorylated-HSPB1, two of the positive samples were in the AC group and one was in the AH group (Fig. 1H). All of the samples positive for phosphorylated HSPB1 were also positive for total HSPB1. This suggests that extracellular HSPB1 is only rarely phosphorylated in ALD patients and so, phosphorylation is not likely to be involved in its extracellular functions in AH.

View this table:

• View inline
• View popup
• Download powerpoint

Table 3. Patient demographics and clinical data in studies of phosphorylated HSPB1 in serum.

HSPB1 is increased in the liver of patients with AH

We next considered the source of circulating HSPB1 in patients with severe AH. The liver is responsible for the majority of ethanol metabolism and is subject to high levels of oxidative stress generated by the highly reactive byproducts of this metabolism.³⁵ A large body of literature has demonstrated that HSPB1 is cytoprotective in cells experiencing oxidative stress.³⁶ HSPB1 release from stressed hepatocytes has also been reported.²⁹ This suggested that the liver was the most likely source of circulating HSPB1 in the AH patients.

In the United States, very few ALD patients undergo diagnostic liver biopsy, so large numbers of liver samples from these patients are difficult to obtain. We were fortunate to acquire banked liver lysates from five patients with severe AH and five healthy controls (Table 4). These samples were obtained during liver transplantation from patients with severe AH and healthy donor livers. There was a clear difference in HSPB1 between HC and AH, with HSPB1 significantly higher in AH than HC (Fig. 2A, B). We also utilized existing RNA-seq datasets to evaluate HSPB1 gene expression in the livers of patients with AH of varying severity or other types of liver disease. In non-severe and severe AH (liver biopsies or explanted liver), HSPB1 gene expression was significantly higher than both HC and any other type of liver disease (Fig. 2C, D, E). Interestingly, metabolic dysfunction-associated steatotic liver disease (MASLD) was not associated with increased HSPB1 gene expression. This is similar to reports in the literature that HSPB1 in circulation does not differ between pediatric healthy controls and patients with MASLD and that ballooning degeneration is associated with decreased HSPB1 gene expression and protein in livers from adult MASLD patients.^{37, 38} Taken together, these data identify the liver as an important source of circulating HSPB1 in AH and suggest that high levels of HSPB1 in the liver are linked to ALD, but not MASLD.

• Download figure
• Open in new tab

Figure 2. HSPB1 is increased in the liver of patients with AH

A, Representative immunoblot for HSPB1 and GAPDH in HC and AH liver samples. B, HSPB1 protein concentration relative to GAPDH as measured by densitometry in immunoblots represented in (A). C, HSPB1 gene expression in RNA-seq analysis of liver from HC, AH, AC, MASLD, and HCV. D, Forrest plot depicting Log2 fold changes in HSPB1 expression in samples described in (C). Patients with various types of liver disease were compared to healthy controls. E, Forrest plot depicting Log2 fold changes in HSPB1 expression in samples described in (C). Patients with various types of liver disease were compared to patients with severe AH. Data plotted as mean ± SD, except C,D,E which is mean ± SEM. B. Mann-Whitney test. C, D. DeSeq2. HC, healthy control; AH alcohol-associated hepatitis, MASLD, metabolic dysfunction associated with steatotic liver disease; HCV, hepatitis C virus; TPM, transcripts per million. n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001.

View this table:

• View inline
• View popup
• Download powerpoint

Table 4. Patient demographics and clinical data for samples used for HSPB1 liver lysate immunoblot.

Chronic ethanol consumption increases HSPB1 in murine liver

HSPB1 could be increased in the liver secondary to damage, or it could be contributing to damage. To distinguish between these possibilities, we examined livers using a murine model of CAC. Mice were fed the Lieber-DeCarli liquid diet with increasing amounts of ethanol over 25 days or a calorie-matched diet supplemented with maltodextrin (Fig. 3A).³⁹ The final concentration of ethanol in this model is 6% and that concentration is maintained for one week. Controls were pair-fed, meaning that their calorie intake was adjusted daily to match that of a paired ethanol-consuming mouse. This model replicates early or mild ALD. We chose this model because our goal was to determine whether increased liver HSPB1 results from liver damage or precedes and therefore potentially causes liver pathology seen in AH.

• Download figure
• Open in new tab

Figure 3. Chronic ethanol consumption increases HSPB1 in murine liver.

A, Illustration of murine model of CAC. B, Relative gene expression of HSPB1 in the liver of Pair and EtOH mice measured by qRT-PCR. C, HSPB1 concentration in the liver of Pair and EtOH mice measured by ELISA. D&E, Immunostaining for HSPB1 (yellow) and DNA (Hoechst, blue) in pair- and EtOH-fed mice. Scale bar = 100 μm. Images in (D) are focused on periportal hepatocytes and images in (E) are focused on pericentral hepatocytes. Images are representative of 12 Pair and 11 EtOH mice. F, High-magnification image of HSPB1 (yellow) and DNA (Hoechst, blue) immunofluorescence in the liver of EtOH mice. Scale bar = 50 μm. G, Linear regression analysis of liver HSPB1 and triglyceride levels in Pair and EtOH mice. H, AST levels in serum from Pair and EtOH mice. I, ALT levels in serum from Pair and EtOH mice. Data are plotted as mean ± SD. B, C, H, I. Unpaired T-test. G. Simple Linear Regression. CAC, chronic alcohol consumption; Pair, pair-fed mice; EtOH, ethanol-fed mice; TG, triglycerides; AST, aspartate aminotransferase; ALT, alanine aminotransferase. n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001.

Ethanol-fed (EtOH) mice had significantly higher levels of both HSPB1 mRNA and protein in their livers than pair-fed (Pair) controls (Fig. 3B, C). Immunofluorescence staining of liver sections from these mice identified weak HSPB1 staining in periportal regions and high HSPB1 staining around central veins in EtOH mice (Fig. 3D, E). Periportal hepatocytes are exposed to high levels of gut-derived components, including MAMPs.⁴⁰ Ethanol consumption, especially at high levels, damages the gut epithelium, allowing increased translocation of microbes, microbial components, or microbial products across the barrier and into the liver through portal circulation.⁴¹ HSPB1 production in response to MAMPs such as lipopolysaccharide and flagellin is well-documented and is likely responsible for the weak periportal HSPB1 staining in the EtOH mice.^{16, 17} In contrast, HSPB1 staining was pronounced in pericentral hepatocytes, which are primarily responsible for ethanol metabolism in the liver.⁴² Of note, hepatocytes positive for HSPB1 were also filled with lipid droplets and HSPB1 levels in the liver were positively correlated with triglyceride levels in the blood of these mice (Fig. 3F, G). This is in keeping with a report that HSPB1 knockout mice have lower levels of triglycerides in their blood.⁴³ The presence of HSPB1 in the ethanol-stressed liver prior to damage was reinforced by the fact that circulating aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were similar in EtOH and Pair mice (Fig. 3H, I). Since HSPB1 is increased in the ethanol-stressed liver prior to the onset of significant damage, the increased HSPB1 is more likely to be a cause of damage than the result of damage.

Ethanol stressed hepatocytes release HSPB1 into the extracellular environment

Findings from patients and mice chronically consuming alcohol identified the liver, and specifically hepatocytes, as the likely source of circulating HSPB1 in AH. To verify ethanol activated HSPB1 release from hepatocytes, we performed in vitro studies of ethanol challenge using hepatocyte cell lines.

Ethanol exposure rapidly and robustly increased HSPB1 protein in HepG2 cells, even before HSPB1 mRNA levels increased (Fig. 4A, B, C). This type of rapid upregulation without the need for increased transcription has been reported for heat shock proteins, including HSPB1 with other types of cell stress.^{44, 45} It also fits with the extensively documented roles that HSPB1 plays in cellular survival during many different types of stress, including oxidative stress. Since our primary interest in hepatocyte produced HSPB1 was in HSPB1 release from hepatocytes, we investigated whether ethanol challenge led to increased amounts of HSPB1 in the hepatocyte media. Exposure to ethanol led to an approximately 3-fold increase in extracellular HSPB1 within 1 hr. (Fig. 4D). This increase was triggered in response to ethanol metabolism, not cell membrane damage, since inhibiting ethanol metabolism with 4-methylpyrazole (4-MP) completely eliminated HSPB1 release in response to ethanol challenge (Fig. 4E, F). We performed these experiments in two hepatocyte cell lines, HepG2 (human) and AML12 (murine), to account for any cell-line related effects on ethanol metabolism or HSPB1 production and release. Thus, ethanol metabolism leads to rapid increases in both intracellular HSPB1 and HSPB1 released from hepatocytes.

• Download figure
• Open in new tab

Figure 4. HSPB1 is increased in EtOH-stressed hepatocytes and released into the extracellular environment.

A, Immunoblotting for HSBP1 and GAPDH in HepG2 cells exposed to vehicle control or 2.5 mM (1.5%) EtOH for 1 hr. B, Intracellular HSPB1 concentration measured in cell lysates of HepG2 cells exposed to vehicle control or 2.5 mM EtOH for 1 hr. by ELISA. C, HSPB1 gene expression relative to GAPDH in HepG2 cells exposed to vehicle control or 2.5 mM EtOH for 1 hr. by qRT-PCR. D, Extracellular HSPB1 concentration measured in culture media from HepG2 cells exposed to vehicle control or 2.5 mM EtOH for 1 hr. by ELISA. E, Extracellular HSPB1 concentration measured in cell media from HepG2 cells exposed to vehicle control, 5 mM 4MP, 2.5 mM EtOH, or 2.5 mM EtOH and 5 mM 4MP for 1 hr. by ELISA. F, Extracellular HSPB1 concentration measured in cell media from AML12 cells exposed to vehicle control, 5 mM 4MP, 2.5 mM EtOH, or 2.5 mM EtOH and 5 mM 4MP for 1 hr. by ELISA. Data are plotted as mean ± SD. B, C, D. Unpaired T-test. E, F. One-way ANOVA with a Tukey’s multiple comparisons post-hoc test. n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001.

HSPB1 released from ethanol stressed hepatocytes promotes liver-damaging inflammation

In the extracellular environment, HSPB1 functions as a DAMP to amplify macrophage inflammatory responses, including those involving TNFα.^{32, 33} TNFα is of particular interest since it has been implicated in liver damage and hepatocyte death in patients with AH.⁴⁶ This cytokine is also reportedly increased in plasma from liver disease patients with elevated HSPB1 in circulating lymphomonocytes.⁴⁷ Despite minimal changes to the hepatic parenchyma in our EtOH mice receiving the Lieber-DeCarli diet, there was a higher level of TNFα in the liver of these mice than Pair controls (Fig. 5A). Furthermore, the level of TNFα was related to the level of HSPB1 in the liver in these mice (Fig. 5B).

• Download figure
• Open in new tab

Figure 5. Extracellular HSPB1 promotes liver-damaging inflammation.

A, TNFα concentration in liver lysates from of Pair and EtOH mice measured by ELISA. B, Linear regression analysis of liver HSPB1 and TNFα in Pair and EtOH mice. C, TNFα concentration measured in RAW264.7 cells after 1 hr. of treatment with 3 μM HSPB1, 2.5 mM (1.5%) EtOH, or both 2.5 mM EtOH and 3 μM HSPB1 by ELISA. D, TNFα concentration measured in RAW264.7 cells after 24 hr. of treatment with 3 μM HSPB1, 2.5 mM EtOH, or both 2.5 mM EtOH and 3 μM HSPB1 by ELISA. E, TNFα concentration in RAW264.7 cells after exposure to EtOH stressed HepG2 (2.5 mM EtOH for 24 hr.) conditioned media with or with or without 2.5 mg/ml anti-HSPB1 antibody measured by ELISA. Data are plotted as mean ± SD. A. Unpaired T-test. B. Simple Linear Regression. C, D, E. One-ANOVA with a Tukey’s multiple comparisons post-hoc test. n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001.

Macrophages are the primary producers of pro-inflammatory cytokines, including TNFα in the ethanol-stressed liver.¹² Our finding that ethanol stress triggers HSPB1 release from hepatocytes led us to consider that extracellular HSPB1 could act on nearby macrophages to amplify TNFα secretion during CAC. As previously reported, ethanol robustly increased TNFα secretion by macrophages within one hour of exposure (Fig. 5C).⁴⁸ In contrast, TNFα secretion was unchanged in HSPB1 treated macrophages at 1 hour. However, at 24 hours, TNFα secretion was significantly increased in HSPB1 treated macrophages and levels were essentially identical to ethanol-exposed macrophages (Fig. 5D). The combination of HSPB1 and ethanol increased TNFα release slightly over each component alone. These data suggest that the mechanisms and kinetics of TNFα secretion elicited by HSPB1 and ethanol differ, but not the inflammatory outcome. They also demonstrate that macrophages need never encounter ethanol or MAMPs directly to activate inflammation during CAC. HSPB1 is sufficient to activate TNFα mediated inflammation and has the potential to amplify inflammation triggered by either ethanol or MAMPs.

We next investigated whether blocking extracellular HSPB1 is sufficient to inhibit TNFα release from macrophages exposed to ethanol stressed hepatocytes. Media conditioned by ethanol stressed HepG2 cells was used to replicate the natural form and amount of HSPB1 released from hepatocytes in these experiments. We added anti-HSPB1 antibody to the hepatocyte conditioned media, then added the treated media to macrophages, and measured TNFα release. The addition of anti-HSPB1 antibody reduced TNFα release from macrophages significantly (Fig. 5E). Overall, our data suggest that HSPB1 is an important initiator of TNFα dependent inflammation and hepatocyte death in AH. They also show that anti-HSPB1 antibody is a rational therapeutic strategy to protect the ethanol stressed liver.

DISCUSSION

Inflammation destabilizes ALD, leading to disease progression and AH. In this study, we discovered that HSPB1 is increased in the liver and blood of patients with severe AH. We also demonstrated that HSPB1 is released from ethanol stressed hepatocytes and triggers inflammation linked with progression to AH and increased mortality in ALD. Extracellular HSPB1 has been identified as a DAMP in published literature. For example, cardiomyocytes subjected to ischemia and reperfusion injury have been shown to release HSPB1 to activate toll like receptor 2/4 on endothelial cells.⁴⁹ Extracellular HSPB1 is also reported to trigger TNFα inflammation in macrophages.^{32, 33} Our work shows that HSPB1 fits the classical profile of a DAMP in ALD, in that it activates macrophages when released from ethanol stressed cells. However, it is unlike many other DAMPs, in that it normally present at low levels and is rapidly increased in response to ethanol. In CAC, this means that it is released from cells that have encountered ethanol or evidence of ethanol toxicity (i.e., MAMPs) and upregulated HSPB1. Release of DAMPs is generally considered to be a product of dead cells.¹⁵ However, there is ample evidence that HSPB1 is secreted by live cells, including hepatocytes.^{29, 50}

This work not only adds HSPB1 to the list of DAMPs identified in ALD, it also demonstrates that this protein is a cause of TNFα mediated inflammation associated with the progression to AH. Inflammation was established as a therapeutic target in AH decades ago.⁵¹ Since that time, preclinical and clinical studies have identified inflammatory pathways and mediators that contribute to liver damage and hepatocyte death. The cytokine TNFα is among the inflammatory mediators most strongly linked to AH. Patients with AH have higher levels of TNFα in their blood than healthy controls and TNFα is increased in the liver of animals chronically consuming ethanol versus control animals.^52–54 At the cellular level, ethanol stabilizes TNFα mRNA in macrophages, leading to rapid increases in TNFα production and release from these cells.^{48, 55, 56} Macrophages also produce TNFα in response to MAMPs and DAMPs that are increased in CAC.⁵⁷ Alcohol increases MAMP delivery to the liver by damaging the intestinal epithelium and allowing translocation across the gut mucosal barrier.⁵⁸ Alcohol also increases DAMPs in the liver due to secretion from stressed cells and release from dead or dying cells.¹⁵ The high level of TNFα generated in these conditions activates programmed cell death in hepatocytes both in vitro and in vivo.^{46, 59} In pre-clinical models, TNF receptor 1 knockout or anti-TNFα antibody protect animals from liver damage.^{54, 60} However, clinical studies of anti-TNFα antibodies were halted early due to life-threatening infections in the AH patients.^{61, 62} This highlights the importance of targeting liver damaging “sterile inflammation” while leaving host defense intact in AH.

The first step in targeting liver damaging sterile inflammation is identifying appropriate therapeutic targets and biomarkers for those targets. Our findings suggest that HSPB1 is a component of sterile inflammation in AH and therefore warrants further investigation as a biomarker and/or therapeutic target in this disease. HSPB1 fits several criteria for these applications: 1) Differences in HSPB1 can be measured in the blood. 2) HSPB1 is significantly higher in AH than other forms of liver disease and increases with increasing disease severity. 3) HSPB1 is upstream of liver-damaging inflammation and so potentially indicates risk for inflammation before it occurs. 4) Blocking HSPB1 action blocks TNFα mediated inflammation in vitro.

HSPB1 has already been proposed as a biomarker for disease severity in cancers, most notably hepatocellular carcinoma (HCC).²⁷ In HCC, HSPB1 is increased in lymphomonocytes and tumor tissue, especially in cases associated with a history of alcohol use.^{28, 47} High levels of liver HSPB1 have also been associated with increased immune cell infiltration into tumor tissue and worse prognosis in HCC.^{27, 28} Alcohol-associated liver diseases have also been linked to naturally occurring anti-HSPB1 antibodies.²⁵ These antibodies could be important both diagnostically and therapeutically. Concern has been raised that anti-HSPB1 antibodies could interfere with measurement of HSPB1 in standard, ELISA-based assays.³¹ Additionally, naturally occurring anti-HSPB1 antibodies could influence HSPB1 functions and/or interfere with antibody-based therapeutics directed at HSPB1. Here, we utilized a commercially available ELISA, and results supported that this assay is able to measure the form and amount of HSPB1 released from ethanol stressed cells. We did not measure anti-HSPB1 antibodies in our patient population due to limited sample volume and this limitation also precluded any investigation of effects of naturally occurring anti-HSPB1 antibody on extracellular HSPB1 functions. These types of studies will need to be performed in the future if HSPB1 is developed as a therapeutic target for AH.

Development of HSPB1 as a biomarker for AH will also require prospective studies that measure HSPB1 concentration throughout the course of disease. High levels of HSPB1 may predict which patients are at highest risk for disease exacerbation and most in need of intense monitoring or early liver transplantation. Conversely, low HSPB1 may be useful to stratify patients into a low-risk group, since HSPB1 was no different in our patients consuming large amounts of alcohol without liver disease and healthy controls. Multiple timepoint sampling would reveal whether a single cut-off value of HSPB1 or changing levels of circulating HSPB1 would be most useful clinically.

Aside from its potential to serve as a biomarker in ALD, HSPB1 also shows promise as a therapeutic target in the inflammation of AH. Our data show that HSPB1 is upstream of TNFα mediated inflammation. Extracellular HSPB1 has also been reported to activate a second macrophage inflammatory mediator implicated in AH, IL-1β. Therefore, targeting HSPB1 has the potential to block activation of multiple inflammatory pathways and broadly protect hepatocytes from inflammation activated cell death programs. Our work specifically demonstrates the potential efficacy of an anti-HSPB1 antibody to interrupt inflammation in AH. We would expect this type of agent to be particularly effective at blocking the DAMP function of HSPB1, while leaving intracellular, cytoprotective functions of HSPB1 intact. It would also preserve TNFα activation by microbes and allow TNFα mediated inflammation in host defense mounted by macrophages.

Together, our findings directly link HSPB1 to inflammation in AH and provide a potential explanation for disease exacerbation. We propose that HSPB1 is highly upregulated in hepatocytes in response to ethanol metabolism and released into the extracellular environment where it activates TNFα in macrophages. When ethanol exposure is limited in amount or duration, we presume that HSPB1 is both cytoprotective in hepatocytes and promotes tissue repair through activation of low levels of TNFα at the site of damage. Conversely, when ethanol exposure occurs at high levels or is sustained for long periods of time, the high levels of TNFα could overwhelm HSPB1 cryoprotection, leading to hepatocyte death. Additionally, the dying ethanol stressed hepatocytes would be expected to release high levels of HSPB1, which would amplify TNFα production and causes even more hepatocyte death leading to a feed-forward or vicious cycle. Our data support this model in a number of ways. However, several key questions do remain unanswered. First, does extracellular HSPB1 act on cells other than macrophages during CAC? Hepatic stellate cells and other infiltrating innate and adaptive immune cells could also be influenced by HSPB1 to contribute to disease. Second, does anti-HSPB1antibody inhibit inflammation in the complex context of the intact liver during CAC? To answer this question, we need to test our anti-HSPB1 antibody in an animal model that mimics more severe, inflammatory ALD. This experiment would also reveal any negative effects of anti-HSPB1 antibody on hepatocyte health in this context. Lastly, which inflammatory pathways does extracellular HSPB1 activate in the liver during CAC? We focused on TNFα in this study due to its strong link to AH, but HSPB1 likely also activates IL-1β and potentially other pro-inflammatory cytokines in CAC. Answering these questions could lead to novel diagnostic and therapeutic approaches to ALD based on targeting HSPB1. This type of strategy could identify patients at risk of severe inflammatory liver disease and short-circuit inflammation to protect the liver against life-threatening injury.

MATERIALS AND METHODS

Conflict of Interest

The authors declare no conflicts of interest.

Financial Support

NIDDK 1K08DK114713 (JSM), NIAAA Northeast Ohio Alcohol Research Center Pilot and Feasibility Award (JSM), NIAAA Northeast Ohio Alcohol Research Center 5P50AA024333 (LEN), NIAAA Johns Hopkins Clinical Resource for Alcoholic Hepatitis Investigation 5R24AA025017.

Author Contributions

Conceptualization: JSM

Methodology: JSM, AMCO, MB, DR, AMS, LEN

Investigation: AMCO, MB, AB, MM, DC, EH, VP, CF, RV, SD, JD, DS, NW

Funding acquisition: JSM

Project administration: JSM

Supervision: JSM, AMCO, DR, AMS, LEN

Writing – original draft: JSM

Writing – review & editing: JSM, AMCO, MB, AMS, LEN

Data and materials availability

All data and code used in the analysis are available upon request. Materials used in this study are available upon request and with completion of the appropriate material transfer agreements.