Nonalcoholic fatty liver disease (NAFLD) is an increasingly important feature of Western countries. Long considered as a benign condition, it now increasingly accounts for a morbidity factor leading to impaired insulin sensitivity, higher cardiovascular death, and constitutes a key step in the development of fibrosis.1 Inflammation is a pathogenic pathway leading to insulin resistance,2 and recruitment of macrophages in the overload adipose tissue plays a crucial role in this process.3, 4 In a similar fashion, Kupffer cells (KCs), the hepatic resident macrophages, could participate in the development of steatosis and hepatic insulin resistance. In a recent study accepted for publication in HEPATOLOGY,5 Stienstra et al. addressed this question. To this aim, the authors used mice fed a high-fat diet (HFD) enriched in palm oil for 20 weeks, a regimen that induced obesity, adiposity, and steatosis. HFD-fed mice received intraperitoneal (i.p.) injections of either clodronate-loaded liposomes or phosphate-buffered saline (PBS)-loaded liposomes twice during the last week of HFD. Clodronate injections depleted the liver of its KCs. This was associated with a decrease in hepatic interleukin 1β (IL-1β) messenger RNA expression and an enhanced fatty acid oxidation mediated by peroxisome proliferator-activated receptor α, resulting in a lower hepatic triglyceride content. On this basis, the authors conclude on a pathogenic role of IL-1β secreted by KCs on HFD-induced steatosis.
We would like to make three comments with respect to data interpretation. First, KCs are not activated in this model. Indeed, hepatic transcript levels for F4/80 and CD68 (markers of KCs) as well as IL-1β are not elevated. Second, in our experience6 like in that of others,7, 8 when clodronate liposomes are injected i.p., they affect not only liver but also intra-abdominal adipose tissue macrophages (ATMs) such as epididymal and omental ATMs. This process may be overturned if clodronate liposomes are injected intravenously.6 This, given the well-established role of ATMs in the pathogenesis of hepatic inflammation,4 calls for caution when restricting the observed effect to the sole KC depletion in this study. Third, contrasting with repeated observations reported in the literature,2–4 the authors fail to observe adipose tissue inflammation in their mice that received a palm oil–rich diet for 20 weeks. This is confirmed by the microarray analysis on epididymal adipose tissue where expression of inflammatory-related genes such as tumor necrosis factor, IL-1β, and IL-6 was similar in PBS-treated HFD-fed animals and in control animals fed a standard chow. Moreover, PBS-treated HFD mice do not appear to exhibit insulin resistance, because glucose and insulin levels are not increased. An explanation to those intriguing observations might be that i.p. injection of empty (PBS-loaded) liposomes induced weight loss. Another explanation is that i.p. PBS-loaded liposomes alter intra-abdominal ATMs, as it is the case in our results (see supporting figure in Lanthier et al.6). So, it would have been important to compare HFD-fed clodronate-treated animals to HFD-fed animals with and without PBS liposome injection.
Therefore, in our view, the direct and strict implication of KCs in the amelioration of steatosis in this study is not demonstrated.