OUP user menu

Hepatocyte Cytokeratin 7 Expression in Chronic Allograft Rejection

Andrew M. Bellizzi MD, Robin D. LeGallo MD, James C. Boyd MD, Julia C. Iezzoni MD
DOI: http://dx.doi.org/10.1309/AJCPNRXCAP92KNOJ 238-244 First published online: 1 February 2011


We examined hepatocyte cytokeratin 7 (CK7) expression in chronic allograft rejection (CR), a ductopenic condition in which this has not been systematically evaluated, in 20 patients with the clinicopathologic diagnosis of CR and age-, sex-, and native-disease–matched control subjects. We also studied baseline biopsy specimens from both groups. Three pathologists independently reviewed H&E- and CK7-stained sections, counting interlobular bile ducts (BDs) and portal tracts (PTs), noting the morphologic pattern of injury and scoring hepatocyte CK7 expression (0, none; 1+, rare; 2+, multifocal, predominantly periportal; 3+, extension into the lobule; 4+, diffuse). Mean BD/PT ratios and CK7 scores were calculated. The mean BD/PT ratio (0.58) and CK7 score (1.01) for the “CR, diagnostic” group were significantly different from all other group means (P < .05); no other comparisons were significant (P > .05). A CK7 score of 1 or more was observed in 9 (56%) of 16 CR specimens and in 3 (7%) of 41 remaining specimens. Hepatocyte CK7 expression is frequently noted in CR, and it would appear to reflect ductopenia. CK7 staining may be a useful diagnostic adjunct in evaluation of transplant liver biopsy specimens.

Key Words:
  • Liver
  • Transplant
  • Chronic rejection
  • Ductopenia
  • Cholestasis
  • Metaplasia
  • Cytokeratin 7
  • Immunohistochemistry

Approximately 6,000 liver transplants are performed in the United States each year for an array of clinical conditions that disrupt the organ’s synthetic and detoxifying function.1 Among the myriad potential complications of this procedure is chronic allograft rejection (CR), which places patients at risk for graft loss. The histopathologic diagnosis of CR is challenging, with only modest interobserver agreement.26 Its recognition is critically important, especially with increasing recognition that, particularly in its “early” phases, CR is a potentially reversible condition.711

The liver contains 2 different epithelia, hepatocytes and cholangiocytes, each expressing cytokeratins (CKs) 8 and 18. Biliary epithelium and hepatic progenitor cells also express CK7 and CK19, frequently referred to as “biliary cytokeratins.”1214 Aberrant expression of CK7 by hepatocytes has been noted in several conditions characterized by ductopenia and/or cholestasis, including primary biliary cirrhosis (PBC), Alagille syndrome, and progressive familial intrahepatic cholestasis.1519 This expression is believed to reflect a metaplastic change in the setting of cholestasis and loss of hepatocyte contact with the biliary tree. While a previous keratin immunohistochemical study in hepatic allograft rejection examined the usefulness of keratin AE1 in counting bile ducts, to our knowledge, the expression of CK7 by hepatocytes in CR, another ductopenic, cholestatic condition, has not been systematically studied.20 We hypothesized that CK7 would be aberrantly expressed, a finding of potential diagnostic usefulness in the evaluation of transplant liver biopsy specimens.

Materials and Methods

Case Selection

Liver specimens with a clinicopathologic diagnosis of CR were identified by searching the surgical pathology database of the University of Virginia Health System, Charlottesville, for the terms “chronic rejection” and “ductopenia” in combination with “allograft.” The included cases had consistent clinical histories. Patient age (at time of diagnosis of CR), sex, native diagnosis, and time from transplant to the diagnosis of CR were recorded. The control group consisted of transplant recipients without CR; control subjects were matched for age, sex, native diagnosis, and interval from transplantation. In addition, liver biopsy specimens obtained within 1 month of transplantation (generally postperfusion biopsy specimens) were examined from both groups.

Histologic Assessment

Three pathologists (A.M.B, R.D.L, and J.C.I.) independently reviewed H&E-stained sections, counting interlobular bile ducts (BDs) and portal tracts (PTs) and noting the morphologic pattern of injury. The study pathologists were blinded to clinical information and the original histologic diagnosis. Counts were performed as described by Snover.21 Care was taken to include only interlobular bile ducts (characterized by an intraportal location, parallel to an arteriole, with a well-defined lumen, and lined by cuboidal epithelium). The morphologic patterns of injury included the following: acute rejection (AR), biliary (characterized by bile stasis with or without bile ductular proliferation), chronic hepatitis (CH), CR, harvest injury (HI), steatosis, and normal. More than 1 pattern of injury could be identified, and pathologists could express uncertainty (eg, CH vs CR).

Immunohistochemical Assessment

CK7 (OV-TL, dilution 1:800; DAKO, Carpinteria, CA) immunohistochemical analysis was performed on a DAKO Autostainer following antigen retrieval. Hepatocyte CK7 immunoreactivity was scored as follows: 0, none; 1+, rare positive cells; 2+, multifocal, predominantly periportal (groups of cells involving multiple portal tracts); 3+, as for 2+ but with extension into the lobule; and 4+, diffuse Table 1 and Image 1. Care was taken to score only hepatocytes and not to include bile ducts, ductules, and extraportal biliary cells (oval or canal of Hering cells). Hepatocytes were characterized by large size, polygonal shape, low nuclear/cytoplasmic ratio, and (if present) modest CK7 staining intensity with membranous accentuation. Mean BD/PT ratios and CK7 scores were calculated for each specimen. CK7 scores of 1 or more were considered “positive.” Specimens with fewer than 6 portal tracts were excluded from further analysis, according to recent recommendations from the Banff Working Group regarding biopsy adequacy.6

View this table:
Table 1

Statistical Analysis

The Mann-Whitney U test was used to compare the age and interval from transplantation for the 2 groups. The Kruskal-Wallis test with post-hoc pairwise comparisons was used to analyze group mean data; P values were adjusted using bootstrap, stepdown Bonferroni, and Hochberg methods (SAS Multitest procedure, SAS, Cary, NC). P values of less than .05 were considered statistically significant.


Twenty-three case patients and matched control subjects were initially identified. Owing to a combination of specimens with too few portal areas and specimen unavailability, 3 case patients and 3 separate control patients fell out of the analysis. The final study population consisted of 20 case and 20 control patients (28 males; 12 females; ages 1–74 years). The mean ages (44.9 and 45.8 years) and intervals to a diagnostic specimen (4 and 3 years) for cases and controls were not significantly different (P = .85 and P = .20, respectively). The clinical characteristics of the patients are summarized in Table 2. We excluded 20 specimens from the 40 patients owing to too few portal areas, and 3 specimens were unavailable, resulting in a final data set derived from 57 specimens. These 57 specimens were divided among 4 groups as follows: “CR, baseline” (n = 16); “CR, diagnostic” (n = 16); “control, baseline” (n = 12); and “control, diagnostic” (n = 13).

The mean CK7 score and BD/PT ratio for the CR, diagnostic group were 1.01 and 0.58, respectively. P values for the overall Kruskal-Wallis tests of CK7 and BD/PT means were P = .0013 and P < .0001. All pairwise comparisons of CK7 and BD/PT means that included the CR, diagnostic group were statistically significant (ie, CR, diagnostic vs CR, baseline; control, baseline; or control, diagnostic) with all P values less than .008. No other comparisons were statistically significant. The mean CK7 score and BD/PT ratio for all 4 groups are given in Table 3, and Table 4 summarizes the prevalence of the various morphologic patterns of injury.

Image 1

Cytokeratin 7 immunohistochemical analysis. A, Score = 0, no hepatocyte staining (immunoperoxidase, ×200); B, Score = 1+, rare hepatocyte staining (circled) (immunoperoxidase, ×200); C, Score = 2+, multifocal, predominantly periportal hepatocyte staining (immunoperoxidase, ×200); D, Score = 3+, as for 2+ but with lobular extension (asterisk indicates a portal vein branch) (immunoperoxidase, ×100). No cases demonstrated 4+ (diffuse) hepatocyte staining. Care must be taken not to misinterpret bile duct, bile ductular, and canal of Hering staining as hepatocyte staining.

A CK7 score of 1 or more was noted in 9 (56%) of 16 specimens from the CR, diagnostic group. Native diagnoses in these patients included hepatitis C (n = 3), hepatitis B (n = 1), alcohol-related liver disease (n = 1), primary sclerosing cholangitis (PSC; n = 2), and cryptogenic cirrhosis (n = 2). Two others had a score of 0.67 (ie, consensus rather than unanimous agreement on a score of 1+). An illustrative case is depicted in Image 2. The 5 other specimens from the CR, diagnostic group had a score of 0: 2 patients with a native diagnosis of biliary atresia and BD/PT ratios of 0.17 and 1.07, 1 patient with a native diagnosis of sclerosing cholangitis demonstrating “chronic vascular rejection” with a BD/PT of 0.84, 1 patient with a native diagnosis of sarcoidosis and a BD/PT of 0, and 1 patient with a native diagnosis of cryptogenic cirrhosis and a BD/PT of 0.29.

A CK7 score of less than 1 was seen in 38 (93%) of 41 specimens from the other groups. Two baseline biopsy specimens demonstrated 1+ staining (BD/PT ratios, 0.87 and 0.97; biliary and AR/HI injury patterns). Only 1 specimen from the control, diagnostic group demonstrated significant staining. It occurred in a patient with a native diagnosis of alcohol-related chronic liver disease with chronic large duct obstruction resulting in ductopenia (BD/PT ratio, 0.63). The positive predictive value and negative predictive value of CK7 immunohistochemical analysis in transplant liver biopsy evaluation are calculated at 75% and 84.4%, respectively.

View this table:
Table 2

In this study, 5 specimens from 4 patients with PSC and 2 specimens from 2 patients with PBC were examined. From the PSC group, 2 were baseline biopsy specimens, one showing AR and the other HI. One was a biopsy obtained at 17 months with mild, nonspecific, portal-based inflammation and a BD/PT of 1.04. Two patients demonstrated CR, with BD/PT ratios of 0.53 and 0.68. These cases lacked features of (recurrent) PSC, including significant bile ductular proliferation and periductal fibrosis. One PBC specimen was a baseline biopsy specimen with HI. The other was a biopsy specimen obtained at 3 years with a CH pattern of injury and moderate steatosis, without granulomas and with a BD/PT of 1.02.


Despite a significant rate of decline in the last 30 years, CR still complicates 3% to 5% of transplants at 5-year follow-up.4,610,22,23 This declining rate is attributed to an increased recognition of antecedent episodes of AR and better immunosuppression.4,6,10,2225 It is worth noting that the designation chronic rejection refers to a clinicopathologic condition rather than a time course. In fact, most cases occur in the first year after transplantation following repeated episodes of AR or in the setting of an AR episode that proves refractory to treatment. A smaller number of cases manifest in an indolent manner, generally in the absence of preceding clinically overt AR.4,6,10,11,2224

While the morphologic hallmarks of CR include interlobular bile duct senescence and eventual loss, obliterative arteriopathy, and secondary parenchymal changes, the histologic diagnosis of CR can be challenging. If not specifically sought after, ductopenia can be overlooked, and formal counting is time-intensive. Foam cell arteriopathy is typically observed at the hilum in explants and, thus, is rarely seen on biopsy. Secondary parenchymal changes are nonspecific. An early investigation of diagnostic reproducibility in liver allograft rejection found interobserver agreement for the diagnosis of chronic rejection to be only “moderate” and for the assessment of ductopenia to be “fair to moderate.”2 Although a subsequent study using expanded histologic criteria achieved “good” interrater agreement among a group of expert hepatopathologists, the authors emphasized the difficulty of differentiating CR from a normal/near normal biopsy, chronic hepatitis, and biliary obstruction.3

View this table:
Table 3
View this table:
Table 4

Aberrant hepatocyte staining for CK7 has been documented in a variety of ductopenic and/or cholestatic diseases. For example, Yabushita et al16 examined CK7 hepatocyte expression in 83 cases of PBC, including cases from all 4 Ludwig stages. They demonstrated a rough correlation between the extent of CK7 expression and histologic stage. They also found that aberrant CK7 expression correlated with copper staining in periportal hepatocytes, although CK7 was more often positive.16 These results have been corroborated by others.17,18 In addition, several pediatric cholestatic conditions including Alagille syndrome (ductopenic), progressive familial intrahepatic cholestasis (often ductopenic), extrahepatic biliary atresia and PSC (variably ductopenic), and inborn errors of bile acid synthesis (not ductopenic) have been shown to demonstrate aberrant CK7 expression, with expression most prominent in ductopenic cases.15,19

This study demonstrates frequent aberrant expression of CK7 by hepatocytes in CR across a spectrum of native diseases. The relative specificity of staining irrespective of age, sex, and posttransplantation interval links the aberrant staining to CR (ie, aberrant staining is not merely a transplant-related phenomenon). Lack of significant staining was observed across a range of other morphologic patterns of injury, including AR (negative in 4 of 5), biliary (negative in 1 of 2), CH (negative in 4 of 4), HI (negative in 15 of 16), and normal (negative in 9 of 9). Among the 7 false-negatives, 2 had CK7 scores of 0.67 (scored as 1+ by 2 pathologists and 0 by 1), emphasizing that focal staining may be overlooked. Both cases of CR in which the native diagnosis was biliary atresia were negative, possibly implicating the native disease on the ability of transplanted hepatocytes to express CK7. One of the negative cases demonstrated chronic vascular rejection without a significant ductopenic component (which occurs in up to a third of CR cases).5,26 Surprisingly, the 3 most ductopenic specimens (with BD/PT ratios of 0, 0.17, and 0.29) failed to demonstrate aberrant CK7 expression. This may reflect the requirement of a critical mass of biliary epithelium to support hepatocyte CK7 expression. Two of the false-positives occurred in patients with large duct obstruction, one in a baseline biopsy specimen and another in a diagnostic biopsy specimen in which chronic large duct obstruction resulted in ductopenia; in the absence of this clinical history, all 3 study pathologists interpreted the morphologic pattern of injury in this second case to be CR. Thus, hepatocyte CK7 expression is not useful in the distinction of CR from large duct obstruction, which should be clinically excluded in any cholestatic transplant liver biopsy. As hepatocyte CK7 expression has been described in PBC and PSC, it would not be useful in the distinction of CR from recurrent disease. The features of recurrent PBC and PSC in the hepatic allograft are similar to those in native livers, including bile ductular proliferation, interface activity, and granulomas in PBC and an abnormal cholangiogram, bile ductular proliferation, bile stasis, and periductal fibrosis in PSC.

Image 2

Illustrative case of a 50-year-old man who received a liver transplant for hepatitis C with chronic rejection in an explanted liver 3 years after transplantation. A, Representative portal area lacking an identifiable interlobular bile duct; the interlobular bile duct to portal tract ratio in this case was 0.41 (H&E, ×400). B, Cytokeratin 7 highlights numerous periportal hepatocytes; the majority of portal areas showed a similar pattern of staining, which was interpreted by all 3 study pathologists as 2+ (immunoperoxidase, ×600).

Two theories have been offered to explain this pattern of staining in cholestasis; aberrant CK7 expression may represent a stem cell response or metaplasia. The first is drawn by analogy from the observed activation of the stem cell compartment seen in liver regeneration after submassive necrosis and in moderate to severe active chronic viral hepatitis. Roskams and colleagues27 stated that “progenitor cells mainly differentiate toward the hepatocytic lineage after hepatocyte loss, whereas they mainly form ductules and regenerate interlobular bile ducts in biliary diseases.” If the stem cell hypothesis is correct, why would stem cells differentiate toward hepatocytes in CR, in which interlobular bile ducts represent the primary site of injury? In fact, there are several lines of evidence suggesting that CR is characterized by a deficient stem cell response. First, CR is characterized by a relative paucity of bile ductular proliferation.6 Van den Heuvel and colleagues28 demonstrated a diminished number of extraportal biliary cells (a population believed to contain the hepatic stem cell) and decreased proliferative activity of this compartment in CR cases vs control cases. Cholangiocytic precursors (ie, biliary stem cells) in CR undergo increased apoptosis by terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate-biotin nick-end labeling (TUNEL) assay.29 Finally, hepatic progenitor cells have been shown to express acetylcholine receptors, and diminished progenitor cell activation may, at least in part, reflect the graft’s denervated state.30 In light of this, we believe the aberrant expression to most likely represent metaplasia, an attempt by hepatocytes to assume a more biliary phenotype in the setting of cholestasis and loss of contact with the biliary tree. This explanation would account for the observation of this phenomenon across the spectrum of cholestatic conditions.

Regardless of the precise mechanism, given challenges in arriving at the histologic diagnosis and the relative specificity of staining in this setting, we believe CK7 immunostaining may serve as a useful diagnostic adjunct. The presence of periportal hepatocellular staining for CK7 in a case histologically “suggestive” of CR would support the morphologic impression. As such, it would function similarly to copper staining in the diagnosis of PBC. Potential further directions of study include the delineation of the time course of hepatocyte CK7 expression in CR and assessment of the usefulness of CK7 immunostaining in supporting the diagnosis of CR in marginally adequate samples.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
View Abstract