Nodular regenerative hyperplasia of the liver after IL-2 therapy in an HIV-infected patient

Abstract

IL-2 is a potent stimulator of T-cell proliferation and maturation, used as intermittent therapy of HIV-infected patients in combination with antiretroviral treatment to increase the CD4 T-cell count [1–5]. Acute IL-2 therapy has a dose-dependent hepatotoxicity estimated at between 20 and 33% of cases [1,4]. We report the case of an HIV-infected patient who developed liver nodular regenerative hyperplasia (NRH), revealed by anicteric cholestasis and portal hypertension 33 months after the beginning of six courses of IL-2 therapy and 2 months after rechallenge. A 65-year-old man, seropositive for HIV-1, was referred to our liver unit in July 2003 for chronic anicteric cholestasis. He had no history of illicite drug use, no viral co-infection and no excessive alcohol consumption. HIV infection discovered in 1995 was treated with azidothymidine monotherapy. In 1996, treatment including didanosine, lamivudine and saquinavir was started for a low CD4 T-cell count (Fig. 1). In February 2000, with the emergence of viral escape mutant, lamivudine was stopped, abacavir and ritonavir were introduced. Viral load dropped and remained definitely under the detection level. He received six courses of IL-2 (4.5 MIU twice a day, subcutaneously) for 5 days, stopped in September 2001 for adverse events (’flu-like syndrome, asthenia, rash). He progressively developed an anicteric cholestasis. In July 2003, a first liver biopsy showed mild portal fibrosis without inflammation. Sinusoids were enlarged in the mediolobular region. IL-2 therapy was re-introduced for a persistent low CD4 T-cell count. Two months later, he a developed ascites, and the spleen was enlarged without clinical signs of hepatic insufficiency. Endoscopy revealed oesophageal varices (grade I–II). Liver function tests, including albumin, prothrombin time and bilirubin, remained within the normal range. Liver imaging showed liver dysmorphy, without vascular obstruction, or liver tumour. Transjugular liver biopsy with haemodynamic evaluation showed portal hypertension (10 mmHg, N < 4 mmHg), defined as the difference between wedge hepatic venous pressure (18 mmHg) and free hepatic pressure (4 mmHg) [6]. Liver histology showed no cirrhosis, but modified liver architecture with atrophic and hypertrophic hepatic cells, sinusoidal distension, suggestive of NRH, confirmed by reticuline staining. Eighteen months later the patient remained stable under beta-blocker therapy.Fig. 1: Antiretroviral treatment, hepatic tests, and CD4 cell count during the follow-up period.—▪— Alanine aminoaspartate;Alanine aminotransferase;gamma-glutamyl transferase; —×— Alkaline phosphatase.Liver NRH is thought to result from occlusion of the terminal branches of the hepatic arterioles and portal veinules, secondary to endothelial cell damage [7]. In the present case, the diagnosis of NRH was assessed by the following considerations. First, the classic conditions associated with liver NRH, including systemic and vascular disorders, coagulopathies, myelo and lymphoproliferative diseases, drug-induced endothelial toxicity and diabetes [7], were ruled out. Second, our observation clearly shows a temporal relationship between IL-2 therapy, the progressive development of an anicteric cholestasis, and the appearance of portal hypertension with oesophageal varices revealed by ascites 2 months after IL-2 rechallenge. Third, the diagnosis of liver NRH suggested by the presence of (grade I–II) oesophageal varices despite a modest elevation of the hepatic venous pressure gradient (below 12 mmHg), in the absence of hepatocellular insufficiency was confirmed by sequential liver biopsies showing the progressive appearance of intrahepatic veinule abnormalities followed by remodelling of liver architecture and fibrosis. High dose of IL-2 induced severe systemic toxicity with a capillary leakage syndrome characterized by a loss of intravascular fluids, leading to generalized oedema and haemodynamic instability and organ dysfunction [8]. Elevated transaminases and bilirubin are commonly observed with subcutaneous IL-2 therapy (3–15 MIU/day) [1,4]. However, hepatic tests are not fully described in long-term follow-up studies [2]. Of note was the fact that one series mentioned the occurrence of liver NRH in a patient [3]. In our case, liver toxicity related to highly active antiretroviral therapy could not be excluded, but does not provide an explanation for portal hypertension in the absence of cirrhosis. In animal models, pharmacological doses of IL-2 ranging from 4 × 106 to 36 × 106 IU/kg per day induced vascular leak syndrome associated with increased transaminases, hyperbilirubinemia, lymphocytic infiltration in the livers of C57BL/6 mice as a consequence of an increased production of TNF-α and chemokines by activated Kupffer cells, increased adhesion molecules by endothelial cells (interstitial cell adhesion molecule 1, vascular cell adhesion molecule 1) [9], leading to a decreased hepatic sinusoid blood flow [10,11]. Prolonged studies support a fibrogenic property of IL-2, in rat with human recombinant IL-2 injection after 4 months [12], or in the schistosomia model in mice [13]. The diagnosis of liver NRH needs a careful examination of liver tests, mainly anicteric cholestasis. Our observation clearly shows a temporal relationship between IL-2 therapy and the occurrence of liver NRH, revealed by portal hypertension in an HIV-infected patient.