ORIGINAL ARTICLE

Aberrant Centrizonal Features in Chronic Hepatic Venous Outflow Obstruction Centrilobular Mimicry of Portal-based Disease Gregor Krings, MD, PhD,* Bilge Can, MD,w and Linda Ferrell, MD*

Abstract: Chronic hepatic venous outflow obstruction is characterized by centrizonal scarring but may also display features that can lead to portal tract mimicry and misdiagnosis as biliary disease, especially given elevated cholestatic liver profiles in these patients. However, these histopathologic features have not been systematically described. We graded the numbers of centrizonal arterioles, ductules, keratin 7+ hepatocytes, CD34+ microvessels, and capillarized sinusoids in 61 cases of chronic venous outflow obstruction and assessed changes in metabolic zonation by glutamine synthetase staining. Centrizonal arterioles and ductules were present in 82.0% and 72.1% of cases, respectively, and correlated with fibrosis. Centrizonal CD34+ microvessels and sinusoidal capillarization were closely associated and present in 25 (92.6%) and 26 (96.3%) of 27 cases, respectively. Centrizonal capillarized sinusoids and microvessels, which were present in all cases with advanced fibrosis, were demonstrated in 90% and 80% of the cases without significant fibrosis, respectively. The results suggest that capillarization and/or microvessel formation precede and may contribute to centrizonal scarring, whereas arterialization likely reflects vascular remodeling associated with progressive fibrosis. Centrizonal ductules were often immature, being either keratin 7+/ keratin 19  (36.4%) or keratin 7  /keratin 19  (10.0%). Centrizonal keratin 7+ intermediate-phenotype hepatocytes were present in 25 (92.5%) of 27 cases. Lastly, 22 (91.7%) of 24 cases showed loss of metabolic zonation, with reversed zonation in 2 (8.3%) cases. Together, the findings indicate that vascular and lobular reorganization in chronic venous outflow obstruction may result in mimicry of central zones as portal tracts. Recognition of these changes is essential to prevent misdiagnosis of this condition as biliary tract disease.

From the *Department of Pathology, University of California, San Francisco, CA; and wDepartment of Pathology, Ondokuz Mayis University, Samsun, Turkey. G.K. and B.C. contributed equally. Conflicts of Interest and Source of Funding: Funded by the Department of Pathology, University of California, San Francisco, CA. The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article. Correspondence: Gregor Krings, MD, PhD, Department of Pathology, University of California San Francisco, 505 Parnassus Avenue, San Francisco, CA 94143 (e-mail: [email protected]). Copyright r 2013 by Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 38, Number 2, February 2014

Key Words: chronic venous outflow obstruction, centrizonal arterialization, glutamine synthetase, ductular metaplasia, sinusoidal capillarization (Am J Surg Pathol 2014;38:205–214)

T

he obstruction of hepatic venous blood flow at the level of intrahepatic veins, extrahepatic veins, inferior vena cava, or right atrium by a wide variety of disorders, including veno-occlusive disease, Budd-Chiari syndrome, or cardiac diseases such as congestive heart failure, results in both acute and chronic problems in liver function and physiology.1–4 The diagnosis of chronic venous outflow obstruction is typically made on clinical and radiologic grounds but is often confirmed on liver biopsy, which can also provide additional critical information, including extent of progression and fibrosis and the presence or absence of superimposed conditions. Chronic venous outflow obstruction manifests as a set of characteristic histopathologic features in the liver, including centrizonal hepatocyte ischemia with hepatocyte atrophy, necrosis, sinusoidal dilatation and congestion, perivenular sinusoidal fibrosis, and centrizonal (zone 3) scarring, which may progress to cirrhosis.1–3,5–9 In addition to these welldefined lobular features, several studies have described the presence of periportal changes, including mild portal inflammation, fibrosis, and ductular reaction, in patients with chronic venous outflow obstruction.10,11 Interestingly, Pai and Hart10 have also identified cytokeratin (CK) 7+ perivenular hepatocytes of intermediate phenotype in chronic venous outflow obstruction, which correlated with a cholestatic chemistry profile. Indeed, patients with chronic venous outflow obstruction often present with a cholestatic liver profile, including elevated bilirubin and alkaline phosphatase.10,11 Misinterpretation of these histologic features may therefore potentially impact diagnosis and patient management, resulting in misdiagnosis and/or unnecessary evaluation for superimposed biliary tract problems in patients with isolated chronic venous outflow obstruction. Steatohepatitis, which is also associated with centrizonal scarring, often shows architectural remodeling resulting in aberrant centrizonal arterialization and ductule formation.12 The presence of arterioles and ductules near patent or sclerosed central veins may mimic portal www.ajsp.com |

205

Krings et al

Am J Surg Pathol

regions with or without proper biliary ducts and/or an obstructive pattern of injury. Accordingly, this aberrant phenotype may lead to misdiagnosis of steatohepatitis as a biliary lesion with or without ductopenia or result in inappropriate disease staging if scarred central zones are mistaken for fibrotic portal tracts.12 In our experience, chronic venous outflow obstruction also leads to significant centrizonal vascular alterations, including arteriolar ingrowth and microvessel formation in central scars. In this respect, Tanaka and Wanless7 have described a uniquely remodeled lobular vasculature in cases of advanced cirrhosis due to Budd-Chiari syndrome. However, these features have not been systematically evaluated. Furthermore, the effects of vascular and lobular architectural reorganization on hepatocyte function have not been well characterized. Recognition of these changes in biopsies of chronic venous outflow obstruction is essential to prevent misinterpretation of centrizonal lesions as portal tracts and the consequent misdiagnosis of superimposed biliary disease with or without ductopenia, which is a common reason for consultative referral in our experience and may result in unnecessary biliary tract evaluation and/or clinical mismanagement in this patient population. In this study, we address this issue by systematically evaluating the centrizonal histopathologic features of chronic venous outflow obstruction, with specific emphases on the presence and nature of centrizonal ductular differentiation using CK immunoprofiling and on reorganization of the centrilobular vasculature, including arteriolar ingrowth, microvessel formation, and sinusoidal capillarization by CD34 immunohistochemistry. In addition, the effects of these changes on functional lobular organization are assessed by glutamine synthetase (GS) immunohistochemistry as a surrogate marker for lobular metabolic zonation and thereby hepatocyte functional specification.

MATERIALS AND METHODS To identify cases of chronic hepatic venous outflow obstruction, electronic surgical and autopsy pathology files at the University of California San Francisco were searched using keywords “chronic venous outflow obstruction,” “venous outflow obstruction,” or “impairment of venous outflow.” Eight-three consecutive cases diagnosed between 1990 and 2011 were identified. Two additional consecutive autopsy cases from 1982 and 1983 were subsequently identified using identical search parameters and an extended search range (1982 to 2011). Of these initial 85 patients, cases with concomitant steatohepatitis, chronic cholangiopathies (primary biliary cirrhosis, primary sclerosing cholangitis, and overlap syndromes), and/or radiographically documented biliary tract obstruction were subsequently excluded. The resultant final study population comprised 61 cases from 58 patients with clinically documented chronic venous outflow obstruction. Of these, 2 patients had a history of concomitant hepatitis C infection (3.4%), and 1 patient carried a diagnosis of genetic hemochromatosis (1.7%). Only 1 of the 3 patients with concomitant chronic liver

206 | www.ajsp.com



Volume 38, Number 2, February 2014

disease (a hepatitis C-positive patient) showed portal inflammatory activity (grade 3 of 4), and none demonstrated significant portal-based fibrosis. Specimens included 48 needle biopsies, 8 autopsies, 2 explants, 2 wedge biopsies, and 1 partial hepatectomy (the latter performed for a persistent mass-like lesion revealed to be extensive regenerative change due to chronic venous outflow obstruction). Serial specimens were available for 3 patients. All specimens were fixed in 10% buffered formalin and embedded in paraffin. The number of centrizonal arteries (with a visible muscular component of the vessel wall identifiable on hematoxylin and eosin (HE) and/or trichrome staining) was graded quantitatively (scale 0 to 3) by HE and trichrome stains as follows: 0, no artery in central zones; 1, 1 to 2 arteries in 1 to 2 central zones; 2, >2 arteries in 2 arteries in >50% of central zones.12 The numbers of centrizonal and periportal ductules (defined as small, biliary-type cells with scant cytoplasm arranged in ductular configurations) were semiquantitatively scored (scale 0 to 3) as follows: 0, none; 1, mild; 2, moderate; and 3, marked. Owing to lack of a formal staging system for chronic venous outflow obstruction and a recognized similarity of fibrosis progression between this disorder and steatohepatitis, trichrome-stained slides were staged for fibrosis using the Non-Alcoholic Steatohepatitis Clinical Research Network (NASH CRN) system.12,13 Material for immunohistochemical analysis was available in 28 cases from 25 patients (including 3 serial specimens). In some cases, available tissue was insufficient to perform all immunohistochemical stains, and this is noted where applicable. The following antibodies were used: CD34 (dilution 1:400, clone QB-END/10; Novocastra, Leica Biosystems, Buffalo Grove, IL), CK7 (dilution 1:500, clone OV-TL 12/30; Dako, Glostrup, Denmark), CK19 (dilution 1:60, clone RCK108; Dako), and GS (dilution 1:250, clone GS-6; Millipore, Billerica, MA). Antigen retrieval systems used were Dako 6.0 (CK7 and GS), and TRS 9.0 (CD34). No antigen retrieval was used for CK19. Positive controls were used for all antibodies, as follows: CD34: unremarkable skin (dermal blood vessels); CK7: ovarian serous carcinoma; CK19: unremarkable liver (bile ducts); and GS: unremarkable liver (centrizonal hepatocytes, perivenular ring-like pattern). On-slide cell types expected to lack expression of these markers were used as negative controls, and a full-thickness section of normal colon was used as an antibody-negative control. Immunostaining for CK7 and CK19 was performed to characterize centrizonal ductular reaction (CK7 and CK19) and intermediate-phenotype hepatocyte staining (CK7; n = 24). As on HE sections, ductular reaction was defined as the presence of small, biliary-type cells with scant cytoplasm arranged in ductular configurations. Intermediate-phenotype hepatocytes were defined as larger cells with hepatocyte-like morphology, CK7 immunopositivity, and lack of ductular arrangement. Immunostaining for CD34 was performed to identify centrizonal arteries r

2013 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 38, Number 2, February 2014

Central Zones in Hepatic Venous Outflow Obstruction

(defined as above), microvessels (defined as small vascular channels with CD34+ endothelial cell lining but no visible muscular wall), and capillarization of sinusoids (n = 27). The numbers of centrizonal CK7+ and/or CK19+ ductules, CK7+ intermediate-phenotype hepatocytes, CD34+ microvessels, and CD34+ capillarized sinusoids were scored (scale 0 to 3) semiquantitatively as follows: 0, none; 1, mild; 2, moderate; and 3, marked.12 The pattern of GS staining was used as a surrogate for changes in hepatic metabolic zonation (n = 24). The statistical analysis of differences in immunohistochemical staining patterns between groups was performed using the Fisher exact test, and the w2 test was used to assess for statistical correlations.

Notably, relevant liver-related laboratory studies demonstrated moderately to markedly elevated alkaline phosphatase and total bilirubin in 67% (n = 36; mean 254 U/L, range 127 to 1143 U/L) and 58% (n = 33; mean 2.9 mg/dL, range 1.1 to 10.7 mg/dL) of patients, respectively. In fact, levels of either alkaline phosphatase or bilirubin were increased in 71.1% of patients, and both were concomitantly increased in 51.6%. In contrast, aspartate aminotransferase and alanine aminotransferase levels were mildly to moderately elevated in 37.5% (n = 40; mean 52.6 U/L, range 17 to 222 U/L) and 27.5% (n = 40; mean 56.8 U/L, range 9 to 530 U/L), respectively (Table 1).

RESULTS

Selected histologic features of all cases are summarized in Table 2 (n = 61). Sinusoidal dilatation with congestion were present in 58 (95.1%) of 61 cases, and zones of centrizonal atrophic hepatocytic plates were identified in 59 (96.7%) cases. Centrizonal hepatocyte necrosis and/or dropout were present in only 7 (11.5%) cases. Predominantly lymphocytic portal inflammation was present in a minority (19.7%) of cases. Consistent with previous reports,10,11 43 (70.5%) cases demonstrated periportal ductular reaction; periportal ductular reaction was mild (grade 1) in the majority (26 cases, 60.5%) of these. All cases demonstrated centrizonal fibrosis and perivenular scarring. The extent of fibrosis was variable and ranged from minimal centrizonal-based fibrosis to cirrhosis (Table 2).

Study Population and Clinical Parameters The study population comprised 61 specimens from 58 patients with clinically documented chronic venous outflow obstruction in the absence of radiographically documented biliary tract obstruction, chronic cholangiopathy, or steatohepatitis. The underlying etiologies of chronic venous outflow obstruction were grouped into 4 categories as follows: cardiac (29 cases, 50%, including congestive heart failure, valvular disease, congenital heart disease, aortic aneurysm, cardiomyopathy, and chronic pulmonary disease), Budd-Chiari syndrome (18 cases, 31%), veno-occlusive disease (2 cases, 3%), and unknown (9 cases, 16%) (Table 1). The age at diagnosis ranged from 18 to 85 years (mean 52.4 y), with the exception of one 8-month-old child with acute leukemia treatment–related veno-occlusive disease. There were 37 men and 21 women. TABLE 1. Summary of Clinical and Laboratory Features in Study Population Age (y) Male/female Etiology (n [%]) Cardiac Budd-Chiari syndrome Veno-occlusive disease Unknown Alkaline phosphatase, increased (n = 36)w Bilirubin, increased (n = 33)z Alkaline phosphatase/bilirubinwz (n [%]) Either increased (n = 38) Both increased (n = 31) AST, increased (n = 40)y ALT, increased (n = 40)8 Alkaline phosphatase/bilirubinwzy8 Either increased, with normal AST or ALT (n = 37) Either increased, with normal AST and ALT (n = 37)

52.4* 37/21 29 18 2 9 24 19

(50) (31) (3) (16) (67) (58)

27 16 15 11

(71.1) (51.6) (37.5) (27.5)

19 (51.4) 17 (46.0)

n = 58 unless otherwise noted. *Range, 18 to 85 years; excluding one 8-month-old child. wAlkaline phosphatase: mean 254 U/L (range, 127 to 1143 U/L; normal range, 31 to 95 U/L). zBilirubin: mean 2.9 mg/dL (range, 1.1 to 10.7 mg/dL; normal range, 0.2 to 1.3 mg/dL). yAST: mean 52.6 U/L (range, 17 to 222 U/L; normal range, 17 to 42 U/L). 8ALT: mean 56.8 U/L (range, 9 to 530 U/L; normal range, 11 to 50 U/L [female patients] and 12 to 60 U/L [male patients]). ALT indicates alanine aminotransferase; AST, aspartate aminotransferase.

r

2013 Lippincott Williams & Wilkins

Histopathologic Findings in Chronic Venous Outflow Obstruction

Aberrant Centrizonal Vasculature in Chronic Venous Outflow Obstruction Well-formed arterioles demonstrating readily identifiable medial smooth muscle were identified on HE and trichrome-stained slides in centrizonal fibrous scars in 50 (82.0%) of 61 cases and were common (grades 2 to 3) in 30 (49.2%) cases (Fig. 1, Table 2). The number of centrizonal arterioles correlated directly with fibrosis stage (P = 0.009; Table 3). Furthermore, in 2 of 3 patients from whom serial specimens were available, the number of centrizonal arterioles increased temporally with increasing fibrosis stage, whereas the number of arterioles remained unchanged in the third case (Table 4). In addition to centrizonal arteriolar ingrowth, immunostaining for CD34 revealed the presence of distinct but delicate, thin-walled microvessels in centrizonal scars in 25 (92.6%) of 27 cases (Figs. 1, 2). Centrizonal microvessels were common (grades 2 to 3) in 15 (55.6%) cases (Table 2). Immunolabeling for CD34 was also used to assess for centrizonal sinusoidal capillarization, as CD34 expression correlates with loss of the normal fenestrated phenotype of hepatic sinusoidal endothelial cells and may precede fibrosis in certain types of liver injury.14 Indeed, centrizonal sinusoids expressed CD34 in 26 (96.3%) of 27 cases (Figs. 1, 2; Table 2). The degree of centrizonal sinusoidal capillarization strongly correlated with microvessel formation (P < 0.001; Table 5), but neither capillarization nor microvessels correlated with fibrosis stage (P = 0.899 and www.ajsp.com |

207

Krings et al

TABLE 2. Summary of Histopathologic Features in Chronic Venous Outflow Obstruction in Study Population Sinusoidal dilatation/congestion (n = 61) (n [%]) 58/61 (95.1) Hepatocyte plate atrophy (n = 61) (n [%]) 59/61 (96.7) Hepatocyte necrosis/dropout (n = 61) (n [%]) 7/61 (11.5) Portal inflammation (n = 61) (n [%]) 12/61 (19.7) Periportal ductular reaction (n = 61)* (n [%]) Grade 1 26/61 (42.6) Grade 2 12/61 (19.7) Grade 3 5/61 (8.2) Total 43/61 (70.5) Centrilobular ductular metaplasia (n = 61)* (n [%]) Grade 1 29/61 (47.5) Grade 2 7/61 (11.5) Grade 3 8/61 (13.1) Total 44/61 (72.1) Centrizonal arterioles (n = 61)w (n [%]) Grade 1 20/61 (32.8) Grade 2 17/61 (27.9) Grade 3 13/61 (21.3) Total 50/61 (82.0) Combined centrilobular ductular metaplasia and centrizonal arterioles (any grade) as a function of fibrosis stage (n = 61)*wz Stage 1a 0/4 (0) Stage 1b 10/18 (55.6) Stage 2 10/15 (66.7) Stage 3 16/20 (80.0) Stage 4/cirrhosis 4/4 (100) Total 40/61 (65.6) Centrizonal microvessels (n = 27)* Grade 1 10/27 (37.0) Grade 2 11/27 (40.7) Grade 3 4/27 (14.8) Total 25/27 (92.6) Centrizonal sinusoidal capillarization (n = 27)* Grade 1 12/27 (44.4) Grade 2 9/27 (33.3) Grade 3 5/27 (18.5) Total 26/27 (96.3) Centrilobular intermediate-phenotype hepatocytes (n = 27)* Grade 1 8/27 (29.6) Grade 2 12/27 (44.4) Grade 3 5/27 (18.5) Total 25/27 (92.5) Aberrant GS zonation (n = 24) Absent or decreased 20/24 (83.3) Reversed 2/24 (8.3) Total 22/24 (91.7) Fibrosis stage (n = 61)z Stage 1a 4/61 (6.6) Stage 1b 18/61 (29.5) Stage 2 15/61 (24.6) Stage 3 20/61 (32.8) Stage 4/cirrhosis 4/61 (6.6) *Semiquantitative grading of centrizonal capillarization and microvessels by CD34 immunohistochemistry, centrilobular ductular metaplasia, centrilobular intermediate-phenotype hepatocytes by CK7 immunohistochemistry, and periportal ductular reaction as follows: 0, none; 1, mild; 2, moderate; 3, marked. wGrading of arterioles: 0, no arteriole in central zones; 1, 1 to 2 arterioles in 1 to 2 central zones; 2, >2 arterioles in 2 arterioles in >50% of central zones. zFibrosis stage determined using the NASH CRN system.

0.623, respectively) or arterialization (P = 0.271 and 0.806, respectively). However, 9 (90%) of 10 cases demonstrated centrizonal sinusoidal capillarization in the absence of significant fibrosis (NASH CRN stage 1), and 5 (83.3%) of 6 cases with early fibrosis (NASH CRN stage 1) lacking

208 | www.ajsp.com

Am J Surg Pathol



Volume 38, Number 2, February 2014

ductules also showed centrizonal sinusoidal capillarization (Fig. 2). In contrast, all cases with a more advanced fibrosis stage (NASH CRN stage Z2) demonstrated centrizonal sinusoidal capillarization. Similarly, 8 (80%) of 10 cases without significant fibrosis (NASH CRN stage 1) demonstrated centrizonal microvessels, which were present in all cases with established fibrosis (NASH CRN stage Z2). Neither the number of centrizonal microvessels nor capillarization increased with increasing fibrosis stage in the 3 patients for whom serial specimens were available (Table 4).

Aberrant Centrilobular Ductular Reaction in Chronic Venous Outflow Obstruction Centrilobular ductules were present in 44 (72.1%) of 61 cases and were common (grades 2 to 3) in 15 (24.6%) cases (Fig. 3; Table 2). Similar to centrizonal arteriolar ingrowth, the number of centrilobular ductules correlated directly with fibrosis stage (P = 0.003; Table 6). In fact, the number of specimens demonstrating both centrizonal arterioles and ductules increased progressively with increasing fibrosis stage (P = 0.015; Table 2). Interestingly, centrilobular ductules were also closely associated with the extent of periportal ductular reaction (P < 0.001). Whereas 22 (91.7%) of 24 cases demonstrated centrilobular ductules expressing CK7, only 14 (58.3%) cases demonstrated CK19+ ductules. Furthermore, centrizonal ductules in 8 (36.4%) cases expressed only CK7 in the absence of CK19 (n = 22), and 2 (10.0%) cases were both CK7 and CK19 negative (n = 20; Fig. 3). It is noteworthy that these CK7  /CK19  ductules demonstrated a distinct ductular-like morphology and underlying architectural pattern reminiscent of hepatic plates with associated reticulin lining and were immunonegative for CD34 (not shown), supporting their epithelial differentiation. Of cases expressing CK19 in centrizonal ductules, CK19 staining was in general very weak (1+), with 9 (64.3%) of the 14 cases demonstrating more intense CK7 than CK19 ductular staining, and the other 5 (35.6%) cases showing CK7 and CK19 staining of equal intensity. Only 1 (7.1%) case revealed 2+ CK19 (and 2+ CK7) staining, compared with 42.9% of cases with Z2+ CK7 staining. In addition to centrilobular ductular reaction, 25 (92.5%) of 27 cases demonstrated centrilobular hepatocytes with a CK7+ intermediate phenotype (Fig. 3; Table 2). The number of centrilobular CK7+ intermediate-type hepatocytes correlated directly with fibrosis stage (P = 0.044) and trended toward near statistical significance with centrizonal CK7+ ductule formation (P = 0.053); in contrast, no association was identified with periportal ductular reaction (P = 0.321). The only 2 cases lacking intermediate-phenotype hepatocytes demonstrated early-stage fibrosis (NASH CRN stage 1b).

Loss of Normal Lobular Zonation in Chronic Venous Outflow Obstruction To assess the effects of vascular reorganization on functional lobular hepatocyte activity, we performed immunohistochemical analysis for GS as a surrogate marker for lobular metabolic zonation. In a properly zonated r

2013 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 38, Number 2, February 2014

Central Zones in Hepatic Venous Outflow Obstruction

B

A

C

CV CV CV

D

E

F

CV CV

H

G

I

CV

CV

CV

K

J

CV

L

CV

FIGURE 1. A–I, Vascular reorganization of central zones in chronic venous outflow obstruction. Ingrowth of arterioles into central scars and perivenular regions can mimic ductopenic portal tracts. A and B, Several arterioles (black arrows) in a scarred central zone are easily identifiable on HE (A) and trichrome (B) stains. C, Arteriole in perivenular area without significant fibrosis (HE). D and E, Single arteriole in rounded central scar on HE (D) and trichrome (E) stains. Note the presence of perivenular hepatocyte atrophy. F, In this example, the arteriole is at the edge of the scar (HE). G–I, Well-formed arterioles adjacent to a large central venule on HE (G) and trichrome (H) stains. The same central zone also demonstrates sinusoidal capillarization, as determined by CD34 immunohistochemistry (I, white arrowheads). J–L, Although not obvious on HE stain (J), this central zone demonstrates formation of several microvessels (black arrowheads, K and L) in the scar, as well as centrilobular sinusoidal capillarization (white arrowheads, K and L) by CD34 immunohistochemistry. CV indicates central vein.

liver, GS expression is limited to a thin, circumferential, ring-like rim of pericentral hepatocytes and lacking in periportal and mid-zonal hepatocytes. However, 22 (91.7%) of 24 cases of chronic venous outflow obstruction in our study population showed significantly aberrant GS r

2013 Lippincott Williams & Wilkins

staining. Specifically, 20 (83.3%) cases demonstrated decreased, asymmetric, or complete loss of pericentral ringlike GS staining, and 2 (8.3%) additional cases showed periportal (but not pericentral) GS staining, indicative of reversed zonation (Fig. 4; Table 2). www.ajsp.com |

209

Krings et al

Am J Surg Pathol

TABLE 3. Comparison of Fibrosis Stage With Prevalence of Centrizonal Arterioles in Chronic Venous Outflow Obstruction1 Fibrosis Stage* Grade 0 Grade 1 Grade 2 Grade 3 Total (n [%]) 0/1a 1b 2 3 4 Total

4 4 1 2 0 11

0 7 7 4 2 20

0 4 3 9 1 17

0 3 4 5 1 13

0/4 14/18 14/15 18/20 4/4 50/61

(0%) (77.8%) (93.3%) (90.0%) (100%) (82.0%)

Grading of arterioles: 0, no arteriole in central zones, (1) 1-2 arterioles in 1-2 central zones, (2) > 2 arterioles in 2 arterioles in > 50% of central zones. 1 P = 0.009. *Fibrosis stage determined using the NASH CRN system.

DISCUSSION In this study, we have systematically described and characterized the presence of aberrant centrizonal histopathologic features in chronic venous outflow obstruction, which may result in portal tract mimicry. These features include reorganization of the lobular vasculature, such as centrizonal arterialization, microvessel formation, and CD34+ sinusoidal capillarization, as well as centrilobular ductular metaplasia and intermediate-phenotype hepatocyte differentiation. The functional consequences of outflow obstruction are highlighted by the loss of normal liver zonation in the vast majority of cases. Recognition of these features is important to prevent misinterpretation and misdiagnosis of chronic venous outflow obstruction as biliary disease, which may lead to unnecessary diagnostic biliary tract evaluation and/or patient mismanagement. Isolated arterial branches within the lobule of otherwise unremarkable livers were originally described by Rappaport and subsequently by Ekataksin.15,16 However, in humans, these vessels are typically not found immediately adjacent to small central veins, unlike the distinctly centrizonal arterioles described in our study. Whether the centrizonal arterioles of chronic venous outflow obstruction (and/or steatohepatitis; see below) are derived from these physiologically “normal” intrahepatic arterial



Volume 38, Number 2, February 2014

branches or from another source remains an interesting outstanding question. Recent studies have also reported the presence of similar aberrant centrizonal arterioles and microvessels in NASH.12 NASH, like chronic venous outflow obstruction, is also associated with centrizonal scarring, raising the possibility that centrizonal neovascularization and/or remodeling may either contribute to or result from the development of centrizonal fibrosis in these contexts. Importantly, we have shown that centrizonal microvessel formation and sinusoidal capillarization are closely associated with one another and are often found in both early-stage (NASH CRN fibrosis stage 1) and late-stage (NASH CRN fibrosis stage Z2) lesions. These results suggest that these features precede and may therefore contribute to centrizonal fibrosis and scarring in chronic venous outflow obstruction. Indeed, sinusoidal capillarization often precedes fibrosis and has been previously implicated in contributing to centrizonal fibrosis in alcoholic liver disease and sinusoidal obstruction syndrome.17,18 Further supporting this hypothesis, experimental studies in a rat model system have shown that differentiated but not capillarized sinusoidal endothelial cells promote hepatic stellate cell quiescence by means of vascular endothelial growth factor–mediated nitric oxide production, thereby explaining increased fibrogenesis upon induction of sinusoidal capillarization.19 We speculate that the formation of centrizonal microvessels in chronic venous outflow obstruction may stem directly from the progressive entrapment of capillarized sinusoids by the associated evolving fibrotic scar, thereby explaining the strong statistical correlation between the development and extent of these features. In contrast, we hypothesize that ingrowth of centrizonal arterioles presumably proceeds through a process likely related to the vascular remodeling occurring during progressive fibrosis and eventual cirrhosis due to other fibrotic liver diseases, such as steatohepatitis and viral hepatitis, among others.12,20–22 The lack of significant association between centrizonal arterioles and either microvessels or capillarization further supports this interpretation, as does the correlation between the fibrosis

TABLE 4. Selected Histopathologic Features in Serial Specimens of Patients With Chronic Venous Outflow Obstruction

Case

Year

Centrizonal Microvessels (Grade)*

1

2009 2010 2006 2011 1998 2008

2 2 1 1 2 2

2 3

Centrizonal Sinusoidal Capillarization (Grade)* 2 2 1 1 1 1

Centrizonal Arterioles (Grade)w

Ductular Metaplasia in Centrizonal Scar (Grade)*

Fibrosis Stage (NASH CRN System, Scale 0-4)

1 2 0 3 3 3

0 3 0 1 2 1

2 4 1a 1b 1b 3

*Semiquantitative grading of centrizonal capillarization and microvessels by CD34 immunohistochemistry and centrizonal ductular metaplasia as follows: 0, none; 1, mild; 2, moderate; 3, marked. wGrading of arterioles: 0, no arteriole in central zones; 1, 1 to 2 arterioles in 1 to 2 central zones; 2, > 2 arterioles in 2 arterioles in >50% of central zones.

210 | www.ajsp.com

r

2013 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 38, Number 2, February 2014

A

Central Zones in Hepatic Venous Outflow Obstruction

B

CV

C

CV

D

CV

E

F

FIGURE 2. Centrilobular sinusoidal capillarization and microvessel formation in early-stage and advanced fibrosis. A–C, A case of chronic venous outflow obstruction without significant fibrosis (A, HE; B, trichrome) demonstrating well-developed centrilobular sinusoidal capillarization by CD34 immunohistochemistry (arrowheads, C). D–F, Another case with established cirrhosis (D, HE; E, trichrome) demonstrating both centrilobular sinusoidal capillarization (white arrowheads, F) and microvessel formation (black arrowheads, F) by CD34 immunohistochemistry. Black arrow (D and F): centrizonal arteriole. CV indicates central vein.

stage and the degree of centrizonal arteriolar ingrowth but not microvessel formation or capillarization. In addition to arteriolar ingrowth and microvessel formation, central zones in both steatohepatitis10,12,23 and chronic venous outflow obstruction demonstrate ductular reaction and/or CK7+ intermediate-phenotype hepatocytes, suggesting an association between these features and centrilobular hypoxia. Indeed, centrizonal ductular reactions (so-called “type 2B”) are thought to develop secondary to tissue hypoxia in conditions of impaired venous outflow and in centrinodular locations of focal nodular lesions.24 Such ductules may derive from (i) dedifferentiation of mature centrizonal hepatocytes (ductular metaplasia) and/or acquisition of a hepatic progenitor cell phenotype, (ii) induction of hitherto unidentified centrilobular hepatic progenitor cells, or (iii) migration of

hepatic progenitor cells from their known periportal sites, and this remains a matter of significant debate.24,25 Regardless of their derivation, the centrizonal ductules in chronic venous outflow obstruction appear to display an immature phenotype in many cases, as demonstrated by either a CK7+/CK19  or CK7  /CK19  immunophenotype in nearly one-half of the cases in this study. We speculate that these ductules represent latestage atrophic hepatocytes, perhaps in very early metaplastic transition to a biliary phenotype. The histogenesis of associated centrizonal CK7+ hepatocytes is also unclear, as these cells may represent partially differentiated descendants of bipotential progenitor cells or partially dedifferentiated, metaplastic hepatocytes.23,24 It is noteworthy that, in addition to chronic venous outflow obstruction and steatohepatitis, centrilobular CK7+

TABLE 5. Comparison of Prevalence of Centrizonal CD34+ Capillarization With Microvessels in Chronic Venous Outflow Obstruction Centrizonal Microvessels Centrizonal capillarization Grade 0 Grade 1 Grade 2 Grade 3 No. correlated grades (%)1

Grade 0

Grade 1

Grade 2

Grade 3

1 1 0 0 1/2 (50)

0 8 2 0 8/10 (80)

0 3 7 1 7/11 (63.6)

0 0 0 4 4/4 (100)

No. Correlated Grades (%)1 1/1 8/12 7/9 4/5

(100) (66.7) (77.8) (80.0)

Semiquantitative grading of centrizonal capillarization and microvessels by CD34 immunohistochemistry as follows: 0, none; 1, mild; 2, moderate; 3, marked. 1 P < 0.0001.

r

2013 Lippincott Williams & Wilkins

www.ajsp.com |

211

Krings et al

Am J Surg Pathol

A

B



Volume 38, Number 2, February 2014

TABLE 6. Comparison of Fibrosis Stage With Prevalence of Ductular Metaplasia in Chronic Venous Outflow Obstruction1 Fibrosis Stage Grade 0 Grade 1 Grade 2 Grade 3 Total (n [%])

C

0/1a 1b 2 3 4 Total

D

4 7 4 2 0 17

0 10 9 9 1 29

0 1 2 3 1 7

0 0 0 6 2 8

0/4 11/18 11/15 18/20 4/4 44/61

(0) (61.1) (73.3) (90.0) (100) (72.1)

Semiquantitative scoring of ductular metaplasia: 0, none; 1, mild; 2, moderate; and 3, marked. 1 P = 0.003.

E

F

G

H

I

J

that study, which the authors interpret as indirect support for their association with periportal hepatic progenitor cells and therefore possible derivation from centrizonal or migrating progenitors.23 In the current study, however, we did not find an association between CK7+ centrilobular hepatocytes and periportal ductular reaction in chronic venous outflow obstruction. Given the distance of centrilobular CK7+ hepatocytes and associated ductules from the known periportal hepatic progenitor cell niche and the lack of identified perivenular progenitor cells, we favor these features to represent a metaplastic process and/or dedifferentiation of centrilobular hepatocytes with possible acquisition of progenitor cell-like phenotypes. Regardless of their histogenesis, the numbers of centrizonal ductules are closely correlated with the extent of fibrosis in chronic venous outflow obstruction; similar correlations have been identified between centrizonal ductules and fibrosis in NASH.12 We have also discovered

A

B

CV CV

CV

CV

FIGURE 3. Aberrant centrilobular ductular metaplasia with an immature phenotype in chronic venous outflow obstruction. A and B, Two cases demonstrating ductule formation (thin black arrows) at the periphery of fibrotic central scars (HE). C–F, A different case demonstrates few identifiable ductules on HE (C) and trichrome (D) stains. However, the ductules are significantly highlighted by CK7 immunohistochemistry (E), whereas CK19 staining is sparse and weak (F). G and H, Another case shows abundant centrilobular CK7+ (G) but CK19  (H) ductules (same case as Fig. 1J–L). I and J, Two cases with abundant CK7+ hepatocytes in central zones. Thick black arrows (A, C–F) indicates centrizonal arteriole. CV indicates central vein.

hepatocytes have recently also been identified in various other chronic liver diseases, including autoimmune hepatitis, primary biliary cirrhosis, and hepatitis B and C infection.23 Centrilobular CK7+ hepatocytes were reported to correlate with periportal ductular reaction in

212 | www.ajsp.com

C

PT

CV

PT

D

PT

CV

PT

FIGURE 4. Loss of metabolic zonation in chronic venous outflow obstruction. A and B, Centrizonal scar with complete loss of zonal, ring-like GS staining pattern (A: HE, B: GS). C and D, This case demonstrates loss of ring-like GS staining in the central zone with aberrant diffuse staining in periportal areas, indicative of “reverse” zonation (C: HE, D: GS). CV indicates central vein; PT, portal tract. r

2013 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 38, Number 2, February 2014

an association between the numbers of centrizonal CK7+ hepatocytes and fibrosis stage, which likely reflects the histogenetic or developmental relationship of these cells to the associated ductules. Importantly, periportal ductular reactions are directly associated with progression of fibrosis in various conditions, including NASH and hepatitis C, suggesting that the ductular reaction itself may contribute to fibrogenesis.26,27 Indeed, various signaling pathways and mediators induced by hepatobiliary cells of the ductular reaction may directly activate hepatic stellate cells or recruit/activate inflammatory cells, which may subsequently induce fibrosis.28–30 Understanding the mechanism by which ductular reactions may contribute to progression of fibrosis in chronic venous outflow obstruction and other chronic liver diseases carries important implications for the prevention and/or reversal of fibrosis and should be addressed in future studies. The presence within the hepatic lobule of well-defined zones composed of functionally heterogenous hepatocytes displaying different and often complementary metabolic activities is known as metabolic zonation.31 The mechanism underlying this zonation is thought to be related to gradients of oxygen, metabolites, and hormones derived from portal-to-central blood flow within the lobule and/or developmentally determined morphogen-regulated tissue patterning.31,32 In this context, GS expression, which is typically limited to a thin, circumferential rim of perivenular hepatocytes, is often utilized as a surrogate marker for loss of lobular zonation. The loss and/or reversal of GS zonation in the vast majority of cases in our study highlights the detrimental functional consequences of chronic venous outflow obstruction on hepatocyte metabolic activity and liver function. We hypothesize that this loss of zonation is most likely due to aberrant neovascularization and/or vascular reorganization, as described earlier, which is likely to result in functionally relevant alterations in lobular blood flow, as previously described during cirrhotic progression.20,21,33 In conclusion, we have described a collection of aberrant centrizonal histopathologic features in chronic venous outflow obstruction, which are of special relevance to the surgical pathologist with respect to differential diagnosis. The presence of fibrotic central zones displaying arteriolar ingrowth, centrizonal ductular reaction/metaplasia, centrilobular CD34+ sinusoidal capillarization, and/or loss of centrizonal GS staining can easily cause misinterpretation of scarred central zones as portal tracts with or without duct loss, especially in small needle biopsies (Figs. 1 and 3). Specifically, the presence of a fibrotic centrizonal scar with arteriolar ingrowth can mimic the appearance of a hepatic arteriole in a ductopenic portal tract. Furthermore, although hepatic sinusoidal endothelial cells typically lack CD34 expression, periportal CD34 staining may be seen in some cases of chronic liver disease and cirrhosis34–36; the presence of centrizonal capillarization in chronic venous outflow obstruction may therefore further cause misinterpretation of these areas as periportal zones. Central zones lacking typical GS expression and/or periportal zones with aberrant “reversed” GS expression may further contribute to this r

2013 Lippincott Williams & Wilkins

Central Zones in Hepatic Venous Outflow Obstruction

portal tract mimicry. Lastly, the presence of a centrizonal ductular reaction in isolation or in combination with other aberrant centrilobular features can be especially problematic in this context, by potentially resembling portal tract obstructive changes in these patients, who often present with cholestatic liver profiles.10,11 Accordingly, recognition of these aberrant features is critical to prevent misdiagnosis of an isolated or superimposed biliary tract problem in this patient population, which can lead to unnecessary diagnostic procedures and/or clinical mismanagement. In this context, portal tracts can be differentiated from scarred central zones with aberrant features by paying careful attention to architectural and spatial clues. These include characteristic apposition of the portal vein, the arteriole, and the bile duct, which is not usually present in scarred central veins with arteriolar ingrowth, as well as a distinctive, dense pattern of portal fibrosis and elastin deposition typically not seen in scarred central zones.37 Cognizance of the potential diagnostic pitfalls imposed by portal tract mimicry of scarred central zones, as well as the distinctive architectural features of normal portal tracts, can thus help differentiate portal tracts from aberrant central zones in chronic venous outflow obstruction, thereby minimizing misinterpretation and diagnostic errors. REFERENCES 1. Menon KV, Shah V, Kamath PS. The Budd-Chiari syndrome. N Engl J Med. 2004;350:578–585. 2. Aydinli M, Bayraktar Y. Budd-Chiari syndrome: etiology, pathogenesis and diagnosis. World J Gastroenterol. 2007;13:2693–2696. 3. Valla DC. Primary Budd-Chiari syndrome. J Hepatol. 2009;50: 195–203. 4. Naschitz JE, Slobodin G, Lewis RJ, et al. Heart diseases affecting the liver and liver diseases affecting the heart. Am Heart J. 2000; 140:111–120. 5. Ludwig J, Hashimoto E, McGill DB, et al. Classification of hepatic venous outflow obstruction: ambiguous terminology of the BuddChiari syndrome. Mayo Clin Proc. 1990;65:51–55. 6. Hoekstra J, Janssen HL. Vascular liver disorders (I): diagnosis, treatment and prognosis of Budd-Chiari syndrome. Neth J Med. 2008;66:334–339. 7. Tanaka M, Wanless IR. Pathology of the liver in Budd-Chiari syndrome: portal vein thrombosis and the histogenesis of venocentric cirrhosis, veno-portal cirrhosis, and large regenerative nodules. Hepatology. 1998;27:488–496. 8. Iwai M, Kitagawa Y, Nakajima T, et al. Clinical features, image analysis, and laparoscopic and histological liver findings in BuddChiari syndrome. Hepatogastroenterology. 1998;45:2359–2368. 9. Dilawari JB, Bambery P, Chawla Y, et al. Hepatic outflow obstruction (Budd-Chiari syndrome). Experience with 177 patients and a review of the literature. Medicine (Baltimore). 1994;73: 21–36. 10. Pai RK, Hart JA. Aberrant expression of cytokeratin 7 in perivenular hepatocytes correlates with a cholestatic chemistry profile in patients with heart failure. Mod Pathol. 2010;23:1650–1656. 11. Kakar S, Batts KP, Poterucha JJ, et al. Histologic changes mimicking biliary disease in liver biopsies with venous outflow impairment. Mod Pathol. 2004;17:874–878. 12. Gill RM, Belt P, Wilson L, et al. Centrizonal arteries and microvessels in nonalcoholic steatohepatitis. Am J Surg Pathol. 2011;35:1400–1404. 13. Kleiner DE, Brunt EM, Van Natta M, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology. 2005;41:1313–1321. 14. DeLeve LD. Hepatic microvasculature in liver injury. Semin Liver Dis. 2007;27:390–400.

www.ajsp.com |

213

Krings et al

15. Ekataksin W. The isolated artery: an intrahepatic arterial pathway that can bypass the lobular parenchyma in mammalian livers. Hepatology. 2000;31:269–279. 16. Rappaport AM. The microcirculatory acinar concept of normal and pathological hepatic structure. Beitr Pathol. 1976;157:215–243. 17. Horn T, Junge J, Christoffersen P. Early alcoholic liver injury: changes of the Disse space in acinar zone 3. Liver. 1985;5:301–310. 18. Narita M, Oussoultzoglou E, Chenard MP, et al. Liver injury due to chemotherapy-induced sinusoidal obstruction syndrome is associated with sinusoidal capillarization. Ann Surg Oncol. 2012;19: 2230–2237. 19. Deleve LD, Wang X, Guo Y. Sinusoidal endothelial cells prevent rat stellate cell activation and promote reversion to quiescence. Hepatology. 2008;48:920–930. 20. Ohnishi K, Chin N, Sugita S, et al. Quantitative aspects of portalsystemic and arteriovenous shunts within the liver in cirrhosis. Gastroenterology. 1987;93:129–134. 21. Rappaport AM, MacPhee PJ, Fisher MM, et al. The scarring of the liver acini (Cirrhosis). Tridimensional and microcirculatory considerations. Virchows Arch A Pathol Anat Histopathol. 1983;402: 107–137. 22. Popper H. Pathologic aspects of cirrhosis. A review. Am J Pathol. 1977;87:228–264. 23. Matsukuma S, Takeo H, Kono T, et al. Aberrant cytokeratin 7 expression of centrilobular hepatocytes: a clinicopathological study. Histopathology. 2012;61:857–862. 24. Desmet VJ. Ductal plates in hepatic ductular reactions. Hypothesis and implications. I. Types of ductular reaction reconsidered. Virchows Arch. 2011;458:251–259. 25. Gouw AS, Clouston AD, Theise ND. Ductular reactions in human liver: diversity at the interface. Hepatology. 2011;54:1853–1863. 26. Richardson MM, Jonsson JR, Powell EE, et al. Progressive fibrosis in nonalcoholic steatohepatitis: association with altered regeneration and a ductular reaction. Gastroenterology. 2007;133:80–90.

214 | www.ajsp.com

Am J Surg Pathol



Volume 38, Number 2, February 2014

27. Clouston AD, Powell EE, Walsh MJ, et al. Fibrosis correlates with a ductular reaction in hepatitis C: roles of impaired replication, progenitor cells and steatosis. Hepatology. 2005;41:809–818. 28. Alvaro D, Mancino MG, Glaser S, et al. Proliferating cholangiocytes: a neuroendocrine compartment in the diseased liver. Gastroenterology. 2007;132:415–431. 29. Theise ND, Kuwahara R. The tissue biology of ductular reactions in human chronic liver disease. Gastroenterology. 2007;133:350–352. 30. Copple BL, Bustamante JJ, Welch TP, et al. Hypoxia-inducible factor-dependent production of profibrotic mediators by hypoxic hepatocytes. Liver Int. 2009;29:1010–1021. 31. Jungermann K, Kietzmann T. Zonation of parenchymal and nonparenchymal metabolism in liver. Annu Rev Nutr. 1996;16:179–203. 32. Gebhardt R, Baldysiak-Figiel A, Krugel V, et al. Hepatocellular expression of glutamine synthetase: an indicator of morphogen actions as master regulators of zonation in adult liver. Prog Histochem Cytochem. 2007;41:201–266. 33. Aoki T, Imamura H, Kaneko J, et al. Intraoperative direct measurement of hepatic arterial buffer response in patients with or without cirrhosis. Liver Transpl. 2005;11:684–691. 34. Cui S, Hano H, Sakata A, et al. Enhanced CD34 expression of sinusoid-like vascular endothelial cells in hepatocellular carcinoma. Pathol Int. 1996;46:751–756. 35. Ohmori S, Shiraki K, Sugimoto K, et al. High expression of CD34positive sinusoidal endothelial cells is a risk factor for hepatocellular carcinoma in patients with HCV-associated chronic liver diseases. Hum Pathol. 2001;32:1363–1370. 36. Pusztaszeri MP, Seelentag W, Bosman FT. Immunohistochemical expression of endothelial markers CD31, CD34, von Willebrand factor, and Fli-1 in normal human tissues. J Histochem Cytochem. 2006;54:385–395. 37. Ferrell LD, Greenberg MS. Special stains can distinguish hepatic necrosis with regenerative nodules from cirrhosis. Liver Int. 2007;27:681–686.

r

2013 Lippincott Williams & Wilkins

Aberrant centrizonal features in chronic hepatic venous outflow obstruction: centrilobular mimicry of portal-based disease.

Chronic hepatic venous outflow obstruction is characterized by centrizonal scarring but may also display features that can lead to portal tract mimicr...
852KB Sizes 0 Downloads 0 Views