SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

The liver plays a central role in ethanol metabolism, and oxidative stress is implicated in alcohol-mediated liver injury. β-Catenin regulates hepatic metabolic zonation and adaptive response to oxidative stress. We hypothesized that β-catenin regulates the hepatic response to ethanol ingestion. Female liver-specific β-catenin knockout (KO) mice and wild-type (WT) littermates were fed the Lieber-Decarli liquid diet (5% ethanol) in a pairwise fashion. Liver histology, biochemistry, and gene-expression studies were performed. Plasma alcohol and ammonia levels were measured using standard assays. Ethanol-fed (EtOH) KO mice exhibited systemic toxicity and early mortality. KO mice exhibited severe macrovesicular steatosis and 5 to 6-fold higher serum alanine aminotransferase and aspartate aminotransferase levels. KO mice had a modest increase in hepatic oxidative stress, lower expression of mitochondrial superoxide dismutase (SOD2), and lower citrate synthase activity, the first step in the tricarboxylic acid cycle. N-Acetylcysteine did not prevent ethanol-induced mortality in KO mice. In WT livers, β-catenin was found to coprecipitate with forkhead box O3, the upstream regulator of SOD2. Hepatic alcohol dehydrogenase and aldehyde dehydrogenase activities and expression were lower in KO mice. Hepatic cytochrome P450 2E1 protein levels were up-regulated in EtOH WT mice, but were nearly undetectable in KO mice. These changes in ethanol-metabolizing enzymes were associated with 30-fold higher blood alcohol levels in KO mice. Conclusion: β-Catenin is essential for hepatic ethanol metabolism and plays a protective role in alcohol-mediated liver steatosis. Our results strongly suggest that integration of these functions by β-catenin is critical for adaptation to ethanol ingestion in vivo. (HEPATOLOGY 2012;)

The liver plays an essential role in metabolizing ingested ethanol. 1 Excessive alcohol ingestion can lead to fatty liver (i.e., steatosis), inflammation and fibrosis (i.e., steatohepatitis), and development of cirrhosis. 2 Alcohol-related liver injury is a cause for significant morbidity and mortality around the world. 3, 4 The pathogenesis of ethanol-induced (Et-OH) liver injury is complex and involves, among others, gut-derived lipopolysaccharide, cytokines, the innate immune system, and oxidative stress as well as the interactions of these factors with intracellular signaling pathways. 5-9 Few effective treatments exist for alcohol-related liver disease, making it imperative to understand its pathogenesis so that better treatments can be developed. 10

In the liver, the first step in the metabolism of alcohol takes place via the alcohol dehydrogenase (ADH) family of enzymes, of which ADH 1A/B/C is the predominant form in the liver. 11 Acetaldehyde, formed by the action of ADH, is then metabolized to acetate via aldehyde dehydrogenase (ALDH) enzymes, of which ADH2 is the most abundant isoform in the liver. An alternative pathway of metabolism in the liver takes place via the microsomal cytochrome P450 2E1 (Cyp2E1) enzymes, which are up-regulated with chronic alcoholic ingestion. 12, 13 Cyp2E1 is an important source of reactive oxygen species (ROS) generation and contributor to oxidative stress in the liver. 14

Activity of the key alcohol-metabolizing enzymes, ADH and Cyp2E1, is zonated across the liver lobule and is more prominent in the perivenous (zone 3) hepatocytes. 15, 16 Recently, β-catenin was shown to be the master regulator of hepatic metabolic zonation. 17, 18 We and others have previously reported that β-catenin regulates the expression of Cyp2E1, the loss of which makes β-catenin knockout (KO) mice resistant to acetaminophen-induced hepatotoxicity. 19-21 On the other hand, liver-specific loss of β-catenin leads to increased susceptibility to steatohepatitis in the methionine choline-deficient diet model of liver injury. 22 In addition to its metabolic role, β-catenin has also been implicated in the response to oxidative stress. 23 Because alcohol metabolism generates oxidative stress in the liver, we hypothesized that β-catenin may regulate the coordinated response of the liver to alcohol-metabolism and the associated increase in oxidative stress. Thus, this study was undertaken to determine the effect of hepatocyte-specific loss of β-catenin on ethanol metabolism and alcohol-mediated liver injury in vivo in a murine model using the Lieber-DeCarli ethanol diet.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Animal Genotypes, Dietary Intervention, and NAC Treatment.

Liver-specific β-catenin KO mice were generated as previously described. 19 Female KO mice (Ctnnb1−/−, Cre+/−) and wild-type (WT) littermates (Ctnnb1loxp/loxp;Cre−/−; or Ctnnb1loxp/−,Cre−/−; or Ctnnb1loxp/−;Cre+/−) were between the ages of 8 and 12 weeks at the start of the experiments. All three WT genotypes were used in the experiments as controls and showed indistinguishable phenotype among them on both diets. Mice were maintained on 12-hour light-dark cycles and had free access to the diets. The high-fat Lieber-DeCarli liquid diet (5% final ethanol concentration) was used with a 6-day ramp-up period (2 days of control diet, 2 days of 1.8% ethanol, 2 days at 3.4% ethanol, and then 5% ethanol for 1, 6, or 22 days). The control group received an isocaloric maltodextrin-containing diet in a pair-fed (PF) fashion. For collection of blood for plasma ethanol and ammonia levels, mice were fed the high-fat Lieber-DeCarli liquid diet for 7 days (6 days of ramp-up followed by 1 day of 5% ethanol), and blood was collected at the end of the dark cycle at 7 a.m. The University of Pittsburgh Institutional Animal Care and Use Committee approved the study.

Other reagents and methods are described in the Supporting Materials.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

KO Mice Exhibit Systemic Toxic Effects and Rapid Mortality on Ethanol Ingestion.

During the ethanol ramp-up period, both genotypes had similar food intake, weight change, and exhibited normal behavior. However, KO mice began to exhibit signs of acute illness between 3 and 7 days after initiating the 5% ethanol diet, characterized by weight loss, absence of grooming, and decreased activity. These changes were followed by death or distress, necessitating euthanasia within 48 hours. Fifty percent of KO mice died within the first 6 days of initiating the 5% ethanol diet, whereas none died in the WT/ethanol group (Fig. 1A). Food intake was similar in the two EtOH groups, except for just before death in the KO group (Fig. 1B). To avoid confounding results from animals in extremis, we sacrificed the remaining mice after day 6 on 5% ethanol, and the experiments described below were performed on these mice. PF KO and WT mice appeared healthy and gained weight (data not shown). EtOH KO mice were hypoglycemic with 2-fold lower blood-glucose levels than WT mice (Fig. 1C) and had 10% lower body weight (Fig. 1D). EtOH KO mice had cachexia and severely depleted intra-abdominal fat, compared with the WT/ethanol group, likely representing a baseline defect in energy homeostasis and ethanol-induced acute illness and decreased food intake in KO mice (Fig. 1E; Supporting Fig. 1 24). There was no difference in body temperature between the groups. We conclude from these results that KO mice are highly susceptible to systemic toxicity and death after short exposure to ethanol ingestion.

thumbnail image

Figure 1. KO mice exhibit systemic effects and increased mortality on alcohol diet. (A) Kaplan-Meier survival curve showing increased mortality in EtOH KO mice. (B) Food intake in EtOH WT (open squares) and KO (closed circles) mice. (C) Serum-glucose levels at the time of sacrifice. Data include only KO mice alive on day 6 of 5% ethanol treatment (n = 5/group). (D) Percentage weight change on the experimental diets. (E) Necropsy showing intra-abdominal fat in EtOH mice. Note the abundant intra-abdominal fat (white arrows) in the WT animal, compared with almost complete loss of visceral adipose tissue in the KO mouse (black arrow).

Download figure to PowerPoint

EtOH KO Mice Exhibit Striking Hepatic Steatosis.

Both groups of KO mice had lower liver weight (Supporting Fig. 2). However, only PF KO mice had a lower liver:body weight ratio, compared with the corresponding WT group (Supporting Fig. 3). On microscopic examination of the liver, EtOH KO mice exhibited severe micro- and macrovesicular steatosis in all three zones of the liver lobule. In contrast, WT mice developed only mild (predominantly zone 2) microvesicular steatosis (Fig. 2A, upper panel). Similarly, Oil red O staining for neutral lipids confirmed the presence of increased hepatic steatosis in the KO/ethanol group (Fig. 2A, bottom panel). KO mice had approximately 5-fold higher alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels than WT mice on the ethanol diet (Fig. 2B,C). Biochemical assays revealed higher liver triglyceride and cholesterol levels in the KO/ethanol group, compared with WT mice (Fig. 2D,E). Serum triglyceride and total cholesterol levels were similar in WT and KO mice (data not shown). Thus, these results show that KO mice develop severe liver steatosis and moderate transaminase elevation on ethanol ingestion in a time period that causes only mild lipid accumulation and no change in liver injury tests in WT mice.

thumbnail image

Figure 2. EtOH KO mice develop severe liver steatosis. (A) Liver sections stained with hematoxylin and eosin or Oil Red O (ORO) from the four experimental groups, as indicated. (B) Serum ALT levels. (C) Serum AST levels. (D) Liver triglyceride levels. (E) Liver total cholesterol levels (n = 5-7/group).

Download figure to PowerPoint

Ethanol Ingestion Increases Hepatic Oxidative Stress in KO Mice.

Increased hepatic oxidative stress is an important mechanism of ethanol-mediated liver injury, and lipid peroxidation (LPO) is used as an indicator of oxidative stress in tissues. Therefore, we performed an assay for malondialdehyde (MDA) levels as an indicator of LPO in the liver. KO mice had higher hepatic MDA levels than WT mice on the ethanol diet (Fig. 3A). Because a supply of reducing equivalents is essential in the face of oxidative stress, we measured hepatic nicotinamide adenine dinucleotide phosphate-reduced (NADPH) levels. EtOH KO mice had 2-fold lower hepatic NADPH levels than corresponding WT mice (Fig. 3B; Supporting Figure 4). Thus, the oxidative stress generated in KO livers by ethanol ingestion likely depletes hepatic NADPH, an important donor of reducing equivalents in antioxidant defense pathways.

thumbnail image

Figure 3. Increased hepatic oxidative stress in EtOH KO mice. (A) Hepatic MDA levels. (B) Hepatic NADPH levels (n = 5/group).

Download figure to PowerPoint

Changes in Mitochondrial Function in EtOH KO Mice.

Mitochondrial dysfunction resulting from hepatic oxidative stress can mediate alcohol-induced liver injury. To determine whether the increased oxidative stress in EtOH KO mice was associated with mitochondrial dysfunction, we assayed the activities of key enzymes in isolated mitochondria from freshly harvested liver tissue (Fig. 4). KO mice had no change in complex I, II, and IV activities, but had lower activity of citrate synthase, the first enzyme of the tricarboxylic acid (TCA) cycle, than the corresponding WT groups. Additionally, activity of aconitase, which catalyzes the second step in the TCA cycle, was lower in PF KO mice and in both EtOH groups. Citrate synthase activity is known to be susceptible to oxidative damage from peroxyl radicals. 25 We conclude from these results that EtOH KO mice have mitochondrial dysfunction associated with increased oxidative stress.

thumbnail image

Figure 4. Changes in mitochondrial function in KO mice. (A) Complex I activity. (B) Complex II activity. (C) Complex IV activity. (D) Citrate synthase activity. (E) Aconitase activity (n = 4-5/group).

Download figure to PowerPoint

Hepatic Superoxide Dismutase Expression Is Lower in KO Mice.

β-Catenin has been implicated in the oxidative stress response via binding to forkhead box (FOXO) transcription factors and regulating expression of antioxidant genes. 23 Manganese superoxide dismutase (SOD2) is a mitochondrial enzyme that is critical for protection against oxidative stress. We hypothesized that β-catenin participates in protection against alcohol-mediated oxidative stress by regulating the expression of SOD2. Western blotting analysis showed significantly lower SOD2 protein levels in KO mice in both treatment groups (Fig. 5A,B). Real-time polymerase chain reaction (PCR) analysis showed that expression of Sod2 was lower in KO mice, suggesting transcriptional regulation by β-catenin (Fig. 5C). Expression of Sod2 is up-regulated by the forkhead transcription factor, FoxO3a. Therefore, we performed immunoprecipitation studies in WT livers and found protein-protein interaction between FoxO3 and β-catenin (Fig. 5D). Expression of two other targets of FoxO3, Cdkn1b and GADD45, were lower in KO mice (Supporting Fig. 5). We conclude from these results that β-catenin transcriptionally regulates the critical mitochondrial oxidative stress response protein, SOD2, via binding to its upstream regulator, FoxO3.

thumbnail image

Figure 5. KO mice have lower SOD2 expression. (A) Western blotting analysis of SOD2 protein levels in the liver. (B) Quantification of SOD2 protein, normalized to GAPDH. Averages of four samples in each group from two independent experiments are shown. (C) Real-time PCR analysis of SOD2 expression levels in the liver (n = 5/group). (D) β-catenin coimmunoprecipitates with FoxO3, the transcriptional regulator of SOD2 expression. IP, immunoprecipitation; IB, immunoblot; IgG, IP performed with immunoglobulin G as the negative control; input, immunoblot with indicated antibody and total liver lysate as the positive control; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

Download figure to PowerPoint

We then asked whether treatment with the antioxidant, N-acetylcysteine (NAC), could prevent mortality associated with ethanol ingestion in KO mice. KO mice were given NAC via twice-daily intraperitoneal injection (500 mg/kg). We found that NAC could not prevent either death (data not shown) or liver steatosis (Supporting Fig. 6) in EtOH KO mice.

KO Mice Exhibit Defective Ethanol Metabolism.

We then asked whether KO mice would exhibit differences in ethanol metabolism in vivo. To avoid the potentially confounding results obtained from acutely ill animals with severe liver steatosis, we assayed plasma-alcohol levels in mice exposed to the 5% alcohol diet for only 1 day after the initial ramp-up period. At this time point, food intake and weight gain were similar between the two genotypes, and KO mice had only modestly greater hepatic steatosis compared with WT mice (Supporting Fig. 7). Under these conditions, plasma-alcohol levels in KO mice were 30-fold higher than WT mice (Fig. 6A). We conclude that KO mice have defective hepatic ethanol metabolism that is independent of the severity of ethanol-induced liver steatosis.

thumbnail image

Figure 6. KO mice have higher plasma-alcohol and -ammonia levels. (A) Serum-alcohol levels. (B) Serum-ammonia levels (n = 5/group). ND, not determined.

Download figure to PowerPoint

β-Catenin regulates glutamine synthetase expression in the liver, and KO mice develop hyperammonemia in certain conditions. 20 Consistent with those previous results, we found that in freshly collected blood samples, the KO/ethanol group had significantly higher plasma-ammonia levels, compared with the WT/ethanol group (Fig. 6B). This hyperammonemia likely represents an additional source of morbidity in EtOH KO mice.

Changes in Major Hepatic Alcohol-Metabolizing Enzymes in KO Mice.

Given the high blood-alcohol levels in KO mice, we measured the activity of the major enzymes responsible for hepatic ethanol metabolism. Both ADH and ALDH activities were lower in PF KO mice. However, enzyme activities in the two EtOH groups were similar (Fig. 7A). The nicotinamide adenine dinucleotide (NAD)/NADH ratio was similar between KO and WT mice (Supporting Fig. 8). Because of previous reports that ethanol-metabolizing enzymes have a perivenous zone-predominant expression pattern and the role of β-catenin as a transcriptional regulator, we then asked whether β-catenin regulated the expression of major genes involved in alcohol metabolism. Real-time PCR analysis showed lower expression of Adh1 and Aldh2 in KO mice (Fig. 7B). Western blotting analysis revealed lower ADH 1 protein levels in both groups of KO mice, but ALDH2 levels were lower only in EtOH KO mice (Fig. 7C,D). Western blotting analysis for Cyp2E1 protein levels in hepatic microsomal preparations showed increased expression in WT mice on ethanol. However, KO mice had almost no detectable levels of Cyp2E1 protein on either diet (Fig. 7E,F).

thumbnail image

Figure 7. KO mice have decreased expression of major alcohol-metabolizing pathways. (A) ADH and ALDH activity assays (n = 5/group). (B) Real-time PCR analysis of Adh1 and Aldh2 genes in the liver (n = 5-6/group). (C) Western blotting analysis of ADH1 and ALDH2 proteins in the liver. GAPDH is shown as the internal loading control. (D) Densitometric analysis of western blotting shown in (C) with normalization to GAPDH. (E) Representative western blotting analysis showing Cyp2E1 protein levels in hepatic microsomal preparation. (F) Densitometric analysis of microsomal Cyp2E1 levels. GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

Download figure to PowerPoint

We then analyzed the expression pattern of ADH1 in liver sections by immunofluorescence microscopy (Fig. 8). PF and EtOH WT mice exhibited a modest perivenous-predominant staining pattern for ADH1, and diffuse cytoplasmic ADH1 staining was visible within hepatocytes (Fig. 8, panels A-D and I-L, respectively). In contrast, PF KO mice had less-prominent ADH1 staining around the central veins (Fig. 8, panels E-H). Furthermore, EtOH KO mice had a patchy, nonuniform ADH1 staining, with several hepatocytes showing aberrantly stained globules projecting into intracellular vacuoles (Fig. 8, panels M-P). Taken together, these results establish that β-catenin regulates the expression, subcellular and lobular localization, and activity of the major ethanol-metabolizing enzymes in the liver.

thumbnail image

Figure 8. Changes in ADH1 localization in hepatocytes in KO mice. Frozen liver sections were triple-stained with ADH1 antibody (FITC; green), TRITC-phalloidin (for F-actin; red), and DRAQ5 (nuclear counterstain; blue), and images were analyzed by confocal laser-scanning microscopy. (A-D) WT/pair-fed. (E-H) KO/PF. (I-L) WT/EtOH. (M-P) KO/EtOH (400x magnification). Images in (D), (H), (L), and (P) are at 800x magnification). Arrows in (P) show the aberrant localization of ADH1 in vacuolated structures within hepatocytes in the KO/EtOH group. FITC, fluorescein isothiocyanate.

Download figure to PowerPoint

Despite many attempts with using mouse liver tissue or the human hepatoma cell lines, Hep3B and HepG2, we could not demonstrate binding of T-cell factor 4 (TCF4), the transcriptional coactivator for β-catenin, to approximately 5,000-base-pair regions spanning the ADH1A and CYP2E1 transcription start sites (Supporting Fig. 9 and data not shown). These results imply that β-catenin either regulates these genes indirectly or that it binds an enhancer outside the span of DNA targeted in our experiments (see discussion below).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Given the growing research interest in the functional role of β-catenin in adult liver homeostasis, we undertook this study to determine whether β-catenin participates in ethanol metabolism and protection against alcohol-mediated liver pathology. β-Catenin is known to associate with FoxO proteins to regulate the expression of SOD and mediates adaptation to oxidative stress. 23, 26 SOD2, which is present in the mitochondria, is critical for protection against oxidative stress. Mice with homozygous Sod2 disruption exhibit dysfunction in multiple organs and mortality in the perinatal or early neonatal period. 27, 28 Heterozygous disruption of Sod2 causes increased oxidative stress and mitochondrial dysfunction. 29 On the other hand, overexpression of SOD2 protects against alcohol-induced liver injury in rodents. 30 Thus, our results show conservation of the β-catenin-FoxO interaction in the mammalian liver and its relevance to hepatic oxidative stress response. It should be noted that the Sod2 gene has multiple levels of transcriptional, epigenetic, and post-translational regulation. 31 Therefore, regulation by β-catenin represents just one of many inputs of the complex network regulating SOD2 expression.

Despite striking liver steatosis, EtOH KO mice had relatively modest increases in oxidative stress and serum ALT/AST levels and exhibited no survival advantage with NAC treatment. Though NAC does not prevent liver steatosis, it does prevent alcohol-induced oxidative stress. 32 Therefore, factors other than oxidative-stress–mediated liver injury were likely causing mortality in KO mice. We show here that the absence of hepatic β-catenin affects the expression and activity of ethanol-metabolizing enzymes and results in high blood-alcohol levels. This defect in ethanol metabolism, along with the hyperammonemia in β-catenin KO mice resulting from loss of hepatic glutamine synthetase expression, likely results in the acute sickness and mortality observed soon after exposure to a 5% ethanol diet. 20

Several hypotheses have been proposed for the rate-limiting step of the major pathway of alcohol metabolism. Some investigators have proposed that the rate of ethanol metabolism is regulated by the amount of hepatic alcohol dehydrogenase. 33 Others have suggested that the rate at which NADH is reoxidized to NAD+ represents the rate-limiting step in alcohol metabolism. 34 A third point of view proposes that there is not a single rate-limiting step in ethanol metabolism, and that control is shared among several steps. 35 These controversies notwithstanding, Ronis et al. recently demonstrated that a significant proportion of alcohol-mediated liver injury occurs independently of alcohol metabolism. 36 Our results, showing that KO mice develop severe steatosis despite a significant block in hepatic ethanol metabolism, are consistent with their study.

A remarkable feature of the adult liver is the zonation of metabolic function across the liver acinus. 37 β-catenin is a key player in establishing hepatic metabolic zonation and is a regulator of the perivenous program of gene expression. 17, 18, 24 Whether alcohol metabolism is zonated is somewhat controversial, and two opposing views of alcohol metabolism have been proposed. Studies with microquantitative techniques or immunohistochemistry suggested that ADH activity was maximal in the perivenous area. 38, 39 Microquantitative techniques for ADH and ALDH from the human liver similarly showed an increasing periportal to perivenous gradient, although there were gender- and age-related differences. 40 On the other hand, Kashiwagi et al. reported no zonal differences in ADH-dependent ethanol metabolism in hemoglobin-free perfused rat liver. 41 The effect of the hepatic zonal architecture on ethanol metabolism is highlighted by the fact that Cyp2E1 is strongly perivenous in its distribution. 12, 15 Our results suggest a modest increase in perivenous ADH1 staining in WT mice, which is absent in KO livers. Furthermore, in contrast to WT mice, we found that EtOH KO mice exhibited a nonuniform pattern of cytoplasmic ADH staining and vacuoles within hepatocytes where there was intense, localized ADH staining. The functional significance of these findings is currently under investigation, but may represent defective protein trafficking within KO hepatocytes, as previously reported for specific proteins in cells depleted of β-catenin/E-cadherin–based adherens junctions, or stress-induced autophagy. 42, 43

We could not detect binding of TCF4, the transcriptional coactivator of β-catenin, at the ADH1A and CYP2E1 promoters. Though these negative results may imply that these genes are not direct transcriptional targets of β-catenin, it is possible that β-catenin/TCF4 bind to enhancers located outside these approximately 5-kb (kilobase) regions spanning the transcription start sites that were targeted by us. Hatzis et al. showed that TCF4-binding sites could be located at large distances (>100 kb) and could be far upstream, intronic, or downstream of transcription start sites of target genes. 44 Similarly, there are five β-catenin responsive elements located 400 kb upstream from its target gene, MYC, and align with the MYC promoter through long-range chromatin loops. 45 Thus, further studies will be needed to determine whether β-catenin directly or indirectly regulates ethanol-metabolizing genes in the liver.

In summary, we show here that liver-specific loss of β-catenin leads to defective ethanol metabolism, increased hepatic oxidative stress, mitochondrial dysfunction, and severe liver steatosis, which combine to produce rapid mortality in EtOH mice. Thus, β-catenin plays a key role in the integration of ethanol metabolism and oxidative-stress functions of the liver.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Additonal Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
HEP_24766_sm_suppinfo.doc6436KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.