Printer Friendly

A histopathologic pattern of centrilobular hepatocyte injury suggests 6-mercaptopurine-induced hepatotoxicity in patients with inflammatory bowel disease.

Drug-induced hepatotoxicity is recognized as a significant complication of thiopurine analog treatment for inflammatory bowel disease (IBD). Hepatotoxicity can occur with either of the two thiopurine analogs in use for IBD treatment, azathioprine (AZA) and 6-mercaptopurine (6-MP), and varies in severity, timing, and clinical presentation. (1) Retrospective studies have reported incidence rates of hepatotoxicity as high as 17%, (1) indicating that hepatotoxicity is an important impediment to the use of these otherwise efficacious medications for treatment of IBD. In addition to performing liver function tests, monitoring of the thiopurine metabolites 6-thioguanine nucleotide (6-TGN) and 6-methylmercaptopurine nucleotide (6-MMP) may aid in identifying thiopurine-induced hepatotoxicity and determining the need for liver biopsy. However, the precise role of metabolite monitoring in patient management is still largely undefined.

A variety of histopathologic findings may be observed in thiopurine-induced hepatotoxicity. The most common abnormality is nodular regenerative hyperplasia (NRH), which may lead to noncirrhotic portal hypertension and cause significant morbidity. (2,3) Development of NRH is most commonly seen with 6-TG, a guanine analog used for treatment of leukemia, with reported incidence rates as high as 54% in biopsy series. (4,5) Other forms of endothelial injury can be seen, including venoocclusive disease, peliosis hepatis, sinusoidal dilatation, and perisinusoidal fibrosis. (1,3) Cholestasis, with or without associated hepatocyte necrosis, may also occur. (1,6) It is important for the pathologist to recognize these patterns of thiopurine-induced hepatotoxicity, given that other disease processes may be clinically favored in this setting. In particular, primary sclerosing cholangitis (PSC) is a common consideration in the IBD patient with abnormal liver function test results.

We report on 3 patients with IBD who underwent liver biopsies due to elevated aminotransferase levels suspected to be due to 6-MP. The 3 biopsy samples exhibited a pattern of centrilobular injury characterized by ceroidladen macrophages, hepatocyte anisonucleosis, and lipofuscin pigment, as well as centrilobular steatosis. Amino-transferase levels returned to normal or remained at borderline elevated levels after the dose of 6-MP was decreased (in 1 patient) or discontinued (in 2 patients). Although histologic findings are not specific and may not enable pathologists to exclude other causes of lobular injury, together they create a distinctive pattern that can allow pathologists to suggest thiopurine-induced hepatotoxicity in IBD patients receiving 6-MP.


Three cases of 6-MP hepatotoxicity were identified from the teaching case files of one of the authors, dating from 2007 to 2011. A natural language computer search of the final diagnoses of liver biopsy specimens processed during the period from 2001 to 2011 in the Department of Pathology at Massachusetts General Hospital (Boston) did not reveal additional cases. Biopsy specimens were fixed in 10% buffered formalin, embedded in paraffin, and stained with hematoxylin-eosin, trichrome stain, periodic-acid Schiff with diastase digestion (PASD), and Prussian blue. The biopsy in case 3 was processed at an outside institution, and slides were reviewed in consultation at Massachusetts General Hospital. Clinical information was obtained from patients' medical records. This study was approved by the Massachusetts General Hospital institutional review board.


Clinical Summaries

Case 1.--A 38-year-old man with ulcerative colitis receiving 6-MP monotherapy underwent a liver biopsy for elevated aminotransferases. The patient had been diagnosed with ulcerative colitis 2 years earlier and had not responded to treatment with mesalamine and prednisone. Two months after being diagnosed, he was admitted for intravenous steroid therapy. At that time, he was started on 6-MP, 50 mg/day. He was weaned from steroids 3 months later and was maintained on 6-MP monotherapy, 100 mg/day, with a complete clinical response. Six months later, he was noted to have an alanine aminotransferase (ALT) level of 82 U/L (reference range, 10-55 U/L), an aspartate aminotransferase (AST) level of 42 U/L (reference range, 10-40 U/L), and an alkaline phosphatase (AP) level of 74 U/L (reference range, 45-115 U/L). The patient's 6-MP dose was decreased to 75 mg/day. His ulcerative colitis continued to be in remission. The elevated aminotransferase levels persisted for 1 year, during which ALT values were 80 to 90 U/L, and AST values were slightly lower. AP level remained in the range of 60 to 80 U/L. Investigation for etiologies of chronic liver disease was undertaken. Hepatitis serology test results were negative, ceruloplasmin level was normal, and anti-mitochondrial antibody and anti-smooth muscle antibodies were negative. Ultrasonography findings suggested fatty liver. The patient's 6TGN level was 312 pmol/8 X [10.sup.8] red blood cells (RBC; reference range, 235-450 pmol/8 X [10.sup.8] RBC, associated with maximum drug efficacy), and his 6-MMP level was approximately 2900 pmol/8 X [10.sup.8] RBC (reference range, >5700 pmol/8 X [10.sup.8] RBC, associated with increased hepatotoxicity). His only other medication was amoxicillin, 250 mg/day, for cystic acne, which he had taken for 13 years, from the age of 22 to 35 years, and which he had restarted approximately 1 year before his aminotransferase elevations were noted. A decision was made to proceed to liver biopsy. At the time of biopsy, due to concerns that the patient was experiencing thiopurine-mediated hepatotoxicity, the 6-MP dose had been decreased to 50 mg/day.

Afterward, the patient remained on 6-MP therapy at the decreased dose of 50 mg/day, and aminotransferase levels normalized 1 month after the biopsy. Four years after the biopsy, the patient's ulcerative colitis was in remission, and his aminotransferase and AP levels were normal at a 6-MP dosage of 50 mg/day.

Case 2.--A 53-year-old woman with Crohn disease receiving 6-MP monotherapy underwent liver biopsy for investigation of elevated aminotransferases. The diagnosis of Crohn disease had been made 30 years earlier. The patient's disease was mainly ileal and perianal and included an ileocolonic fistula. One year and 8 months prior to biopsy, the patient was hospitalized for a Crohn flare and received intravenous steroids. She was discharged on a prednisone taper, and 6-MP at 50 mg/day was initiated. Six months later, prednisone was stopped, and the 6-MP dose was increased to 75 mg/day. The patient had a good response, and her disease was initially well controlled. Seven months later, a Crohn flare required a 4-month oral prednisone taper. At that point, laboratory studies showed a 6-TGN level of 103 pmol/8 X [10.sup.8] RBC and a 6-MMP level of 4035 pmol/8 X [10.sup.8] RBC. The 6-TGN level was believed to be subtherapeutic, and as a result, the 6-MP dose was increased to 100 mg/day. Two weeks later, the patient was noted to have an AST level of 42 U/L, an ALT level of 67 U/L, and an AP level of 75 U/L. The abnormal aminotransferase levels persisted for 6 weeks, at which point her AST level was 56 U/L and her ALT level was 99 U/L. Due to concerns about hepatotoxicity, her physician stopped her 6-MP and started ciprofloxacin therapy. Liver ultrasonography showed fatty liver. Three weeks later, aminotransferases continued to rise, with an AST level of 103 U/L and an ALT level of 184 U/L. Ciprofloxacin therapy was stopped. The patient was not taking any other medications. Results from a hepatitis workup were negative, and autoimmune markers were negative. One week later, her AST level was 96 U/L, her ALT level was 171 U/L, and her AP level was 71 U/L. Two weeks later, aminotransferases had trended downward to an AST level of 57 U/L and an ALT level of 79 U/L. A liver biopsy was performed 2 weeks later (8 weeks after 6-MP was discontinued).

The patient's aminotransferase levels normalized while she was off medication. She was started on mesalamine, 4 mg/day, but 1 month later, she developed a Crohn flare and was started on ciprofloxacin, 500 mg/day, and 6-MP, 50 mg/day, with good response. Her aminotransferase levels remained in the normal-to-borderline-elevated range (~40-50 U/L). One month later, 6-MP dosage was increased to 75 mg/day. Five weeks later, ciprofloxacin was discontinued, and she remained on 6-MP monotherapy. She continued to have good response, with the exception of an episode of diarrhea and elevated aminotransferases (ALT, 84 U/L; and AST, 41 U/L), which was attributed to the use of antibiotic for a tooth abscess. Six months after the biopsy, her Crohn disease was well controlled with 6-MP at 75 mg/day, with borderline elevated levels of ALT (~40-50 U/L) and normal AST. The 6-TGN level was 73 pmol/8 X [10.sup.8] RBC, and the 6-MMP level was 5322 pmol/8 X [10.sup.8] RBC. At that time, in view of the low 6-TGN levels, the patient's physician discontinued 6-MP therapy and started her on allopurinol, 100 mg/day, and AZA, 25 mg/day. The patient's IBD symptoms were well controlled with this regimen. Three months later, the patient's aminotransferase levels had normalized, with an ALT value of 26 U/L, an AST value of 31 U/L, and an AP value of 86 U/L.

Case 3.--A 40-year-old man with ulcerative colitis receiving 6-MP monotherapy underwent a liver biopsy for elevated aminotransferases. The patient had left-sided ulcerative colitis, and his disease was well controlled with 6-MP, 75 mg/day. Eight months prior to biopsy, he was noted to have an ALT level of 98 U/L and an AST level of 57 U/L. Two weeks later, given the patient's persistently elevated aminotransferase levels, the 6-MP dose was decreased to 50 mg/day. Seven months later, his aminotransferase levels were markedly elevated, with an ALT level of 393 U/L and an AST level of 186 U/L; AP level was normal at 96 U/L. Due to concerns about hepatotoxicity, the physician discontinued 6-MP therapy and investigated possible causes of underlying liver disease. Hepatitis serology test results were negative, anti-nuclear antibody was positive at 1:40, and antimitochondrial antibody was negative. Alpha-1 antitrypsin levels and iron study results were normal. Two weeks later, a liver biopsy was performed.

Following the liver biopsy, the patient was started on mesalamine with good response. Aminotransferase levels trended downward over the course of the following 9 months, to AST values in the normal range and ALT values of ~40 to 50 U/L. AP level remained normal. The patient's ulcerative colitis was well-controlled with mesalamine at 4.8 g/day.

Histologic Findings

Biopsy findings were nearly identical in the 3 cases. A prominent feature was centrilobular steatosis, more pronounced in case 1. The steatosis was macrovesicular and confined to the centrilobular region (Figure 1). Lobular inflammation was minimal to absent, and trichrome stains showed no perivenular sinusoidal fibrosis (steatofibrosis). Centrilobular injury was noted in all cases, consisting of relatively numerous ceroid-laden macrophages ("tombstone lesions") localized to the steatotic regions. These could be seen with hematoxylin-eosin stain as lightly pigmented macrophages in the lobules, although they were more easily identified with PASD stain (Figure 2, A and B). Also noted in all cases was hepatocyte anisonucleosis in the region affected by steatosis (Figure 3). Increased pigmentation was noted in centrilobular hepatocytes, accentuated at the canalicular pole of hepatocytes, suggestive of lipofuscin. The pigment was negative on iron stain. No canalicular bile was noted. NRH was absent, and there was no sinusoidal dilatation or fibrosis, venoocclusive disease, or peliosis hepatis. Portal tracts and bile ducts were unremarkable, without changes to suggest PSC or autoimmune hepatitis. An iron stain showed mild hemosiderin deposition in occasional hepatocytes (of 1+ of 4) and in Kupffer cells in case 1.


Biopsy samples from these 3 patients with suspected 6-MP-induced hepatotoxicity exhibited the same histopathologic findings, namely centrilobular steatosis, ceroid-laden macrophages, hepatocyte anisonucleosis, and increased hepatocyte lipofuscin pigment. In case 1, aminotransferase levels normalized after the 6-MP dose was decreased. In cases 2 and 3, aminotransferase levels normalized or were only borderline elevated after discontinuation of 6-MP. Other causes of liver injury were excluded. We believe that in this clinical setting, this recognizable pattern of centrilobular injury is most likely a morphological manifestation of thiopurine-mediated hepatotoxicity.

For pathologists interpreting liver biopsy results, recognition of this pattern of injury and its association with thiopurine-induced hepatotoxicity is important because other processes, such as PSC and autoimmune hepatitis, are likely to figure more prominently in the differential diagnosis of liver biopsies from IBD patients. In particular, PSC is the most common consideration in patients with ulcerative colitis and Crohn disease who present with abnormal liver function tests. The emphasis on diagnosing PSC may cause the pathologist to focus on bile ducts and to overinterpret subtle duct injury and portal fibrosis, particularly if the pattern of liver function test abnormalities (ie, aminotransferase levels versus AP and bilirubin levels) are not communicated to the pathologist.

Conversely, pathologists are likely to dismiss evidence for thiopurine toxicity, given that the histologic findings in our cases are overall nonspecific. Nuclear anisonucleosis and increased lipofuscin pigment in hepatocytes are so common that rarely are they used to generate a differential diagnosis, although both findings can be seen in patients taking various medications, among other reasons. Furthermore, given the prevalence of risk factors for fatty liver disease in the North American population, steatosis on a liver biopsy is often dismissed as a finding of little diagnostic relevance. However, steatosis can be seen in reactions to drugs such as methotrexate, although it has not been reported in thiopurine-induced hepatotoxicity. In IBD patients, steatosis can be attributed to nutritional issues. Unfortunately, we did not obtain biopsies after the transaminase levels had normalized to confirm that the steatosis resolved. Interestingly, however, none of the patients had steatofibrosis to indicate that the fatty liver disease was chronic and progressive. While admittedly we cannot be certain that the steatosis was due to 6-MP effect, the fact that it was present in all 3 patients to a similar degree, it was tightly localized to zone 3, and it colocalized with the anisonucleosis and ceroid-laden macrophages suggested that the fat was due to metabolic disruption in centrilobular hepatocytes secondary to 6 MP.

Another nonspecific but characteristic finding in our cases was ceroid-laden macrophages in the centrilobular region. In isolation, ceroid-laden macrophages, or so-called tombstone lesions, are seen in the resolution phase of hepatocellular injury secondary to drugs, viral hepatitis, or other causes of lobular injury. In our cases, 6-MP was the likeliest culprit. In case 2, the patient was also taking ciprofloxacin during part of the time she had elevated transaminase levels. However, the patient was first noted to have elevated aminotransferases 2 weeks after her 6-MP dose was increased, before she had been exposed to ciprofloxacin. Several months after the biopsy, her ALT became borderline elevated again when she was receiving 6-MP monotherapy, and her aminotransferases normalized only after discontinuation of 6-MP. In case 1, the patient was also taking amoxicillin, but he had a long history of amoxicillin exposure both before and after the episode of transaminase elevations. Therefore, although in both of these cases the patients were taking antibiotics as well as 6-MP, the timing of the transaminase elevations the absence of cholestatic features, and the similarity of the histopathology favor 6-MP as the culprit.




It is worth noting that hepatotoxicity developed in all 3 patients despite the fact that thiopurine metabolite levels were not elevated, underscoring the limitations of metabolite monitoring in detection and management of hepatotoxicity or in correlating thiopurine metabolite levels with histologic findings. Other studies have reported that thiopurine-induced hepatotoxicity usually occurs in the setting of normal metabolite levels. (3) Therefore, determination of hepatic toxicity is based on clinical judgment, and pathologic interpretation of liver biopsies must be done within that context, but awareness of the various patterns of 6-MP toxicity allows for a more confident assessment of liver biopsies.

We are not aware of reports of 6-MP toxicity that describe the constellation of histologic features of our cases. One case report of predominantly centrilobular damage in a patient treated with AZA describes a different pattern of injury, with centrilobular cholestasis, focal feathery degeneration, and occasional ground-glass hepatocytes. (6) The clinical presentation in that case was primarily cholestatic, with near-normal aminotransferase levels and markedly elevated AP and bilirubin, which resolved upon withdrawal of AZA. (6) In contrast, the 3 biopsies reported here exhibited hepatocellular injury without tissue bilirubinostasis. Laboratory findings, namely elevated aminotransferase levels and normal AP and bilirubin levels, are consistent with the biopsy findings. Other patterns of injury previously reported in patients receiving thiopurine analogs include NRH and various forms of endothelial injury, such as venoocclusive disease, peliosis hepatis, sinusoidal congestion, and sinusoidal fibrosis (reviewed by Gisbert et al (1)).

Important insights into the mechanism of thiopurine-mediated hepatotoxicity have been obtained from in vitro studies. Incubation with AZA or 6-MP significantly decreases the viability of cultured rat hepatocytes by depleting intracellular ATP, depleting glutathione levels, and damaging mitochondria. (7,8) In those studies, injury to hepatocytes is hypothesized to result primarily from increased oxidative stress due to generation of reactive oxygen species during oxidation of AZA and 6-MP by xanthine oxidase. (8) Consistent with this hypothesis, the decrease in cell viability can be prevented by allopurinol, which inhibits xanthine oxidase. (8) These observations have led to improved therapeutic approaches, as administration of low-dose allopurinol has been found to reduce hepatotoxicity in patients with IBD treated with AZA or 6MP. (9) Alternatively, AZA and 6-MP can be detoxified via conjugation, which consumes reduced glutathione, and may therefore exacerbate injury due to oxidative stress. (7,8) Under conditions of glutathione depletion, detoxification of thiopurine drugs leads to recruitment of stress-activated kinase pathways and causes cell death. (10)

In view of these observations, the centrilobular distribution of 6-MP-mediated injury we observed may be due to the spatial compartmentalization of xenobiotic metabolism within the liver acinus. Cytochrome P-450 enzymes are preferentially expressed in zone 3 hepatocytes. (11) During xenobiotic metabolism, the generation of toxic electrophiles by cytochrome P-450 enzymes leads to increased levels of reactive oxygen species in zone 3. (11,12) Compounding the problem is the fact that hepatocytes in zone 3 have decreased levels of reduced glutathione compared to hepatocytes in zone 1. (11,13) Given that thiopurine-induced hepatocyte injury is mediated by reactive oxygen species accumulation and glutathione depletion, it would be expected to preferentially affect zone 3 hepatocytes, as observed in our 3 cases. The combination of steatosis and anisonucleosis is also seen in liver biopsies of patients taking methotrexate. The similarities may be fortuitous, but because both drugs interfere with purine synthesis, it is interesting to speculate whether that is the mechanism for anisonucleosis in these two drug toxicities.

In summary, we report 3 patients with IBD and 6-MP-induced hepatotoxicity who were found to have centrilobular steatosis, ceroid-laden macrophages, hepatocyte anisonucleosis, and lipofuscin pigment in liver biopsy specimens. None of these features is specific by itself, and even together, they might be seen after lobular injury from other causes including other drugs. However, they create a distinctive appearance that in the setting of a patient with IBD who is being treated with 6-MP, suggests thiopurine-mediated hepatotoxicity.


(1.) Gisbert JP, Gonzalez-Lama Y, Mate J. Thiopurine-induced liver injury in patients with inflammatory bowel disease: a systematic review. Am J Gastroenterol. 2007; 102:1518-1527.

(2.) Daniel F, Cadranel JF, Seksik P, et al. Azathioprine induced nodular regenerative hyperplasia in IBD patients. Gastroenterol Clin Biol. 2005; 29:600-603.

(3.) Vernier-Massouille G, Cosnes J, Lemann M, et al. Nodular regenerative hyperplasia in patients with inflammatory bowel disease treated with azathioprine. Gut. 2007; 56:1404-1409.

(4.) Geller SA, Dubinsky MC, Poordad FF, et al. Early hepatic nodular hyperplasia and submicroscopic fibrosis associated with 6-thioguanine therapy in inflammatory bowel disease. Am J Surg Path. 2004; 28:1204-1211.

(5.) Teml A, Schwab M, Hommes DW, et al. A systematic survey evaluating 6-thioguanine-related hepatotoxicity in patients with inflammatory bowel disease. Wien Klin Wochenschr. 2007; 119:519-526.

(6.) Romagnuolo J, Sadowski DC, Lalor E, Jewell L, Thomson AB. Cholestatic hepatocellular injury with azathioprine: a case report and review of the mechanisms of hepatotoxicity. Can JG astroenterol. 1998; 12:479-483.

(7.) Lee AU, Farrell GC. Mechanism of azathioprine-induced injury to hepatocytes: roles of glutathione depletion and mitochondrial injury. J Hepatol. 2001; 35:756-764.

(8.) Tapner MJ, Jones BE, Wu WM, Farrell GC. Toxicity of low dose azathioprine and 6-mercaptopurine in rathepatocytes. Roles of xanthine oxidase and mitochondrial injury. J Hepatol. 2004; 40:454-463.

(9.) Ansari A, Elliott T, Baburajan B, et al. Long-term outcome of using allopurinol co-therapy as a strategy for overcoming thiopurine hepatotoxicity in treating inflammatory bowel disease. Aliment Pharmacol Ther. 2008; 28:734-741.

(10.) Menor C, Fernandez-Moreno MD, Fueyo JA, et al. Azathioprine acts upon rat hepatocyte mitochondria and stress-activated protein kinases leading to necrosis: protective role of N-acetyl-L-cysteine. J Pharmacol Exp Ther. 2004; 311: 668-676.

(11.) Jungermann K, Keitzmann T. Zonation of parenchymal and nonparenchymal metabolism in liver. Annu Rev Nutr. 1996; 16:179-203.

(12.) Jungermann K, Katz N. Functional specialization of different hepatocyte populations. Physiol Rev. 1989; 69:708-764.

(13.) Kera Y, Penttila KE, Lindros KO. Glutathione replenishment capacity is lower in isolated perivenous than in periportal hepatocytes. Biochem J. 1988; 254:411-417.

Ricard Masia, MD, PhD; Daniel S. Pratt, MD; Joseph Misdraji, MD

Accepted for publication August 30, 2011.

From the Department of Pathology (Drs Masia and Misdraji) and the Gastrointestinal Unit (Dr Pratt), Massachusetts General Hospital, Boston.

The authors have no relevant financial interest in the products or companies described in this article.

Reprints: Joseph Misdraji, MD, Department of Pathology, Massachusetts General Hospital, 55 Fruit St, Warren 219, Boston, MA 02114 (e-mail:
COPYRIGHT 2012 College of American Pathologists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Masia, Ricard; Pratt, Daniel S.; Misdraji, Joseph
Publication:Archives of Pathology & Laboratory Medicine
Date:Jun 1, 2012
Previous Article:HER2/neu gene amplification and protein overexpression in gastric and gastroesophageal junction adenocarcinoma: a review of histopathology,...
Next Article:Biospecimens and biorepositories for the community pathologist.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters