Beyond liver fibrosis: Hepatic stellate cell senescence links obesity to liver cancer by way of the microbiome


  • Nicholas A. Shackel M.B., B.S., Ph.D.,

    1. Liver Injury and Cancer Group, Centenary Institute, Sydney, NSW, Australia
    2. A.W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, Sydney, NSW, Australia
    Search for more papers by this author
  • Mathew A. Vadas M.B., B.S., Ph.D., D.Sc.,

    1. Vascular Biology Programm, Centenary Institute, University of Sydney, Sydney, NSW, Australia
    Search for more papers by this author
  • Jennifer R. Gamble Ph.D.,

    1. Vascular Biology Programm, Centenary Institute, University of Sydney, Sydney, NSW, Australia
    Search for more papers by this author
  • Geoffrey W. McCaughan M.B., B.S., Ph.D.

    1. Liver Injury and Cancer Group, Centenary Institute, Sydney, NSW, Australia
    2. A.W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, Sydney, NSW, Australia
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

Yoshimoto S, Loo TM, Atarashi K, Kanda H, Sato S, Oyadomari S, et al. Obesity-induced gut microbial metabolite promotes liver cancer through senescence secretome. Nature 2013;499:97-101. (Reprinted with permission.)


Obesity has become more prevalent in most developed countries over the past few decades, and is increasingly recognized as a major risk factor for several common types of cancer. As the worldwide obesity epidemic has shown no signs of abating, better understanding of the mechanisms underlying obesity-associated cancer is urgently needed. Although several events were proposed to be involved in obesity-associated cancer, the exact molecular mechanisms that integrate these events have remained largely unclear. Here we show that senescence-associated secretory phenotype (SASP) has crucial roles in promoting obesity-associated hepatocellular carcinoma (HCC) development in mice. Dietary or genetic obesity induces alterations of gut microbiota, thereby increasing the levels of deoxycholic acid (DCA), a gut bacterial metabolite known to cause DNA damage. The enterohepatic circulation of DCA provokes SASP phenotype in hepatic stellate cells (HSCs), which in turn secretes various inflammatory and tumour-promoting factors in the liver, thus facilitating HCC development in mice after exposure to chemical carcinogen. Notably, blocking DCA production or reducing gut bacteria efficiently prevents HCC development in obese mice. Similar results were also observed in mice lacking an SASP inducer or depleted of senescent HSCs, indicating that the DCA-SASP axis in HSCs has key roles in obesity-associated HCC development. Moreover, signs of SASP were also observed in the HSCs in the area of HCC arising in patients with non-alcoholic steatohepatitis, indicating that a similar pathway may contribute to at least certain aspects of obesity-associated HCC development in humans as well. These findings provide valuable new insights into the development of obesity-associated cancer and open up new possibilities for its control.


Cellular senescence is a stable and long-term loss of proliferative capacity, despite continued viability and metabolic activity.[1, 2] It can be induced by replicative exhaustion or by stress, including oncogene activation. Importantly, the last decade has seen the phenotype of cellular senescence evolve from an in vitro phenomenon to being recognized as an important mediator of normal development, in a range of physiological processes and essential in the pathobiology of many diseases. Cellular senesence is now a recognized mediator of aspects of embryonic development,[3] although still the best-recognized pathological role of cellular senescence is as a mechanism suppressing the onset and growth of cancer cells.[2] On the one hand, cellular senescence provides a barrier to cellular transformation by oncogene activation. However, cellular senescence in mesenchymal cells such as fibroblasts or hepatic stellate cells may also induce a hyper-production of a mixture of cytokines that are able to support cancer growth in a paracrine fashion. The senesence cell production of characteristic cytokines, chemokines, and proteases is referred to as the senesence-associated secretory phenotype (SASP) (Fig. 1).[4]

Figure 1.

(A) The schematic of SASP, which was observed to surround HCC nodules and associated with HSC senescence. (B) This phenotype was perpetuated and sustained by the enterohepatic circulation of the bile acid DCA. Further, in (B,C), HCC induction requires an HFD, the carcinogen DMBA, and gram-positive bacteria, especially Clostridium species. The link to the bowel flora perpetuating the HCC phenotype is provided by the enterohepatic circulation of DCA, which arises predominantly from Clostridium species and causes DNA damage resulting in cellular senescence of HSC and then the SASP response.

Inflammatory-associated obesity-related liver injury, nonalcoholic steatohepatitis (NASH) is associated with cirrhosis and hepatocellular carcinoma (HCC) formation.[5] Hepatic stellate cells (HSC) are crucial mediators of the cirrhosis process and responsible for the abnormal matrix that accumulates with progressive liver fibrosis. However, HCC in NASH has been increasingly recognized in the noncirrhosis state but it has been unclear if the mechanisms involved are similar to those that drive HCC in cirrhosis.[6] In cirrhosis, proliferation of the hepatocyte and hepatocyte dysplasia, as a result of direct hepatocyte injury, are thought to be crucial for malignant change. The article in Nature by Yoshimoto et al.[7] provides a new role for HSC in HCC development whereby obesity, whether diet-induced or genetic, leads to cellular senescence of HSCs, which in turn produce an SASP that enables the malignant transformation of liver cells giving rise to HCC. The original hypothesis derived from the knowledge that crucial cytokines induced by the SASP such as interleukin (IL)-6 had previously been linked to cancer development in obesity.[8]

The authors used mice in which p21Waf1/Cip1 up-regulation, which causes cellular senescence, can be monitored by luminescence.[9] These mice were fed a high-fat diet (HFD) and treated with the carcinogen DMBA (7,12-dimethylbenz (a) anthracene). As a result, these mice developed HCC with a p21Waf1/Cip1 up-regulation in HSCs adjacent to the HCC nodules that were detected by luminescence signals. Further, these cells displayed all the relevant features of cellular senescence, damaged DNA foci, and cell cycle arrest. In addition, the HSCs were shown to secrete SASP associated cytokines, including IL-1β. While mice genetically unable to produce IL-1β still display features of cellular senescence and NASH, the incidence of liver cancer was dramatically reduced, suggesting that this component of SASP, rather than cellular senescence-associated cell cycle arrest, was crucial in obesity-induced liver carcinogenesis. The importance of HSC itself was further confirmed by treatment of mice with small interfering RNA (siRNA) to HSP47, which dramatically reduced stellate cell numbers (but not NASH). This also reversed the oncogenic phenotype. The fact that no liver fibrosis was seen in this model of HCC development clearly separated the role of HSC in liver carcinogenesis from a role in liver fibrosis.

The key question arose: How does obesity lead to stellate cell senescence? Previous work has shown that the gut micriobiome by way of Toll-like receptor (TLR)4 signaling promoted HCC development in a diethylnitrosamine (DEN) and carbon tetrachloride (CCl4) model.[10] The authors explored this and made the startling discovery that altering the gut flora in obese mice, e.g., by antibiotic treatment, reversed oncogenesis. However, they did not find a role for lipopolysaccharide (LPS), as mice lacking TLR4 in this setting actually had a slight increase rather than a decrease in HCC development. In further dissecting the role of the gut microbiome, they noted that the antibiotic vancomycin that primarily kills gram-positive bacteria was sufficient to reverse oncogenesis.

The authors then tested serum by way of liquid chromatography mass spectrometry and showed increased serum levels of deoxycholic acid (DCA) in HFD-fed animals. This bile acid is a secondary bile acid produced in the gut mainly by gram-positive organisms of the Clostridium cluster XI and XIVa and is known to cause DNA damage, a crucial inducer of SASP. Suppression of the activity of 7-alpha dehydroxylation of primary bile acids, the metabolic pathway to produce DCA, by treatment of DFAIII reversed the effects of obesity on liver cancer generation. Furthermore, treatment with DCA over a longer period of time was shown to increase HCC development.[7, 11] These changes in HCC development were also associated with increases or decreases in HSC senescent cells, respectively. Analysis of feces for bacterial and associated DCA metabolites identified increased levels of Clostridium cluster XI and genes associated with bile acid dehydroxylation.

Relevance to human disease was partially established by: (1) demonstrating that IL-1β could induce SASP in human HSCs, and (2) finding evidence of cellular senescence and SASP in HSCs adjacent to HCCs in patients with NASH but little fibrosis.

Several issues, however, remain to be answered. The model used by the authors involves treatment of neonatal mice with DMBA, a carcinogen that induces Ras mutations, and DMBA by itself is not sufficient to cause cancer but does so in combination with obesity. This scenario is somewhat removed from the human equivalent situation, where no identified carcinogens have been identified. A second issue is that the model used does not elicit fibrosis and the HFD alone is not sufficient to induce an increase in incidence of HCC. Third, the intermediates in the inflammation/NASH-HCC axis have not been fully clarified. The evidence that anti-IL-1 treatment or the destruction of stellate cells reverses the oncogenic phenotype suggests, but does not prove, that the gut flora do not in part exert their effect through the excess adipose tissue. Indeed, it would have been interesting to see whether DCA treatment (in antibiotic-treated mice) resulted in HCC if stellate cell numbers were depleted.

Intriguingly, this work highlights the potency of a local SASP response when it is combined with the appropriate environmental cofactors such as obesity and other triggers of oncogenesis. Also noteworthy is that, while stellate cells undergo cellular senescence, the very cells undergoing malignant change appear to have escaped the cellular senescence barrier. However, this may be simply due to the fact that neonatal induction of Ras (by DMBA) could have led to other genetic changes to render these cells responsive to local inflammatory stimuli.

Finally, the results raise some exciting questions about human disease pathogenesis and potential therapeutics. These include:

  1. Is there a role for liver biopsy in NAFLD/NASH patients in order to identify HSC senescence?
  2. Could measurements of IL-1β levels in the liver or serum identify patients at increased risk of HCC?
  3. Could serum/fecal DCA levels also help stratify patients?
  4. Is there a role for antibiotics in the management of NAFLD/NASH and the subsequent risk of HCC?

The relevance of these questions and application to patients with NASH cirrhosis-associated HCC also needs to be addressed. In conclusion, the work of Yoshimoto et al. brings us significantly closer to understanding the complexity of the pathogenesis of HCC, but leaves us with continuing questions, overall a great outcome!

  • Nicholas A. Shackel, M.B., B.S., Ph.D.1,2

  • Mathew A. Vadas, M.B., B.S., Ph.D., D.Sc.3

  • Jennifer R. Gamble, Ph.D.3

  • Geoffrey W. McCaughan, M.B., B.S., Ph.D.1,2

  • 1Liver Injury and Cancer Group, Centenary Institute, Sydney, NSW, Australia

  • 2A.W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, Sydney, NSW, Australia

  • 3Vascular Biology Program, Centenary Institute, University of Sydney, Sydney, NSW, Australia