Build to understand biliary oncogenesis via organoids and FGFR2 fusion proteins

Intrahepatic cholangiocarcinoma (ICC) is the second most prevalent primary malignancy in the liver, making up about 3% of all cases of gastrointestinal cancer.¹ It is associated with a median survival of less than 3 years.¹ Recent genetic studies have shown that ICC has a high level of inter- and intra-tumoural heterogeneity, meaning that ICC can rely on a suite of different driver mutations such as KRAS and FGFR-translocations to grow.² This characteristic has made developing effective targeted treatments for ICC particularly challenging. Our lack of effective therapies is exacerbated in part by a lack of models that can mimic human pathologies in a genetically defined setting.

In this issue of Journal of Hepatology, Cristinziano G et al.³ demonstrate a system where liver organoids, derived from adult mouse livers, can be used as a testbed to genetically engineer organoids and direct their oncogenic transformation towards ICC. In this publication, the authors develop a series of organoid lines containing a range of FGFR2-fusion proteins that reveal a necessary role for ERK signaling in the formation of ICC. Importantly, the authors successfully show that mouse liver organoids can act as a viable platform to test genetic factors that are potentially critical in the development of ICC and they can be used to screen for new, targeted therapeutics. To this end, they engineered liver organoids from TP53 null mice with FGFR2 fusion constructs. The engineered liver organoids showed the characteristics of ICC following implantation in mouse models. The authors then demonstrated Erk1/2 activation acts as a central switch downstream of FGFR2-fusion proteins and confers oncogenic signaling. The combination of an FGFR-specific tyrosine kinase inhibitor with MEK1/2 blockade led to an augmented therapeutic response in vitro and in an immunodeficient mouse model (Fig. 1).

Fig. 1. Depiction of study by Cristinziano G et al. showing oncogenic transformation of mouse liver organoids to ICC using FGFR2 fusion proteins.

FGFR2 fusion proteins can induce ICC development in TP53^−/− organoids. The developed tumors are responsive to double inhibition by FGFR-specific tyrosine kinase and MEK1/2 inhibitors (blue arrows on left). The developed organoids show ICC charactristics in vitro and show stromal reactivity and angiogenesis in NOD-SCID mouse models.

Three dimensional self-organized micro-physiological systems derived from stem cells, named organoids, were shown to capture key aspects of native organ function, structure, physiology, and disease.⁴ In the case of the liver, they can span from epithelial-only organoids derived from adult liver progenitors⁵ to multilineage organoids derived from pluripotent stem cells.^6,7 Epithelial-only liver organoid cultures from isolated hepatic progenitors have been developed in recent years and enable expansion of hepatocyte-like populations in vitro. Importantly these organoids maintain their genetic, karyotypic and phenotypic stability.⁸ A modified version of this culture system also supports the derivation of organoids from primary liver cancers that maintain a representative genomic landscape and histology even after long-term expansion in vitro.^9,10 While this model offers an invaluable platform to study tumors, it relies on the generation of organoids from primary human cancers with complex signaling networks and the presence of both driver and passenger mutations at the time of analysis. Additionally, whilst the human origin of these organoids is considered to be widely beneficial, they only represent a single tumor or indeed a sub-clone of that tumor. In order to understand ICC biology, vast numbers of organoids would need to be generated from primary patient samples in order to develop a library of organoids with different genetics that could be used for therapeutic testing. A benefit of the approach detailed here by Cristinziano et al. is that, in principle, using mouse organoids could allow for the rapid generation of organoids containing suites of mutations found in ICC. Additionally, it enables temporal studies of carcinogenesis, helping researchers to identify the key drivers and passenger mutations.

Active angiogenesis, inflammation and fibrosis are all hallmarks of ICC and can play a role in therapeutic response. The mouse ICCs developed in this study also showed active angigenesis together with myofibroblast differentiation and proliferation. Therefore, this system provides an opportunity to examine candidate targets within the cancer niche, such as the stellate cell population. Targeting the cancer niche to modulate the course of liver cancer has already been shown to be an effective therapeutic approach by Seehawer M and colleagues.¹¹ They demonstrated that the type of cell death in the cancer neighborhood shapes cellular fate selection (biliary vs. hepatocyte differentiation) within the tumor cells themselves. In this study, necroptosis-associated cytokines such as IL6 could favor ICC outgrowth from oncogenically transformed hepatocytes.¹¹ It will be valuable to study how the results from these 2 studies can converge to provide a more systematic understanding of the cancer niche and vulnerabilities to therapeutics. However, a successful approach to this question requires the usage of immune-competent models. In fact, in the majority of studies to date, human or mouse organoids are transplanted into immunocompromised animals, which lack a complete immune system. As ICC is known to be a highly stromal tumor with a large immune repertoire, the absence of these cell types in xenograft models could be problematic and we do not currently know what biology we are missing by omitting these cell types from mouse models.¹² This challenge is not insurmountable however, and by using mouse organoids in which oncogenic mutations have been introduced, in place of human ones, the authors have taken a step towards being able to transplant these organoids back into syngeneic hosts (i.e., back into mice of the same genetic background as the organoids). These hosts are immunocompetent and could be used to address the interaction between different pathological mutations and the immune microenvironment. These types of approaches have been particularly powerful in other organoid systems and have proved beneficial in understanding both disease processes and therapeutic development.¹³

Treatment options for ICC are currently limited, and for the majority of patients, curative surgical treatments are not an option. For these patients, standard of care chemotherapy has a relatively modest effect on overall survival, and the treatment that is normally provided is palliative rather than curative. Recently, the FGFR-inhibitor pemigatinib has been approved in the US and Europe for the treatment of ICC in patients with FGFR2-fusions and represents the first targeted therapy in ICC that has been widely licensed to target a particular genetic alteration.^14,15 Pre-clinical and early clinical trials using pemigatinib demonstrated favorable results in this very specific group. Whether this will be the panacea for FGFR-fusion positive ICC remains to be seen – as in other cancers, such as lung adenocarcinoma and melanoma, the targeting of the FGF/MAPK pathway has been fraught with difficulty, not least the rapid evolution of tumor cells that are no longer sensitive to targeted FGFR or BRAF inhibitors.¹⁶ Perhaps this is where the organoids developed by Cristinziano et al.³ have a lasting place in pre-clinical research. They could be used to model the most likely genetic outcomes of sustained FGFR2 inhibition and to screen for compounds that either re-sensitize resistant cancer cells to pemigatinib or can be used as second- or third-line therapies once targeted FGFR2-inhibitors are no longer effective.

How an initial cell state and tissue microenvironment control the ultimate fate after an oncogenic transformation is an ongoing question in cancer research. In this study, the developing tumor exhibited a molecular signature associated with biliary fate, such as upregulation in YAP- and Notch-related pathways, compared to normal liver samples, which suggested a role for the FGFR2-fusion protein in promoting an ICC fate rather than a hepatocellular carcinoma identity. However, a more rigorous analysis and comparison with control Tp53 null organoids will be valuable to understand the degree of biliary fate commitment after oncogenesis by FGFR2-fusions. For instance, hepatic organoids are not fully mature and often display gene regulatory networks from alternative fates, such as progenitor and biliary cells. This may favor ICC development as a default outcome. Additionally, developing tumors can contain undifferentiated cells which may display a degree of similarity to biliary differentiation after comparison against fully mature adult liver samples. However, all in all, the cancer organoid model presented in this study will set the stage for a build-to-understand strategy to study key genetic drivers and therapeutic modalities in ICC.