Intrahepatic Cholangiocarcinoma: A Summative Review of Biomarkers and Targeted Therapies

Simple Summary Intrahepatic cholangiocarcinoma is the second most common primary liver malignancy. Among patients with operable disease, surgical resection is the cornerstone of therapy. Among the majority of patients who present with advanced disease treatment, systemic or targeted therapy is indicated. Recent advancements have provided more novel therapeutic approaches to a subset of patients with intrahepatic cholangiocarcinoma. Abstract Although rare, intrahepatic cholangiocarcinoma (ICC) is the second most common primary hepatic malignancy and the incidence of ICC has increased 14% per year in recent decades. Treatment of ICC remains difficult as most people present with advanced disease not amenable to curative-intent surgical resection. Even among patients with operable disease, margin-negative surgical resection can be difficult to achieve and the incidence of recurrence remains high. As such, there has been considerable interest in systemic chemotherapy and targeted therapy for ICC. Over the last decade, the understanding of the molecular and genetic foundations of ICC has reshaped treatment approaches and strategies. Next-generation sequencing has revealed that most ICC tumors have at least one targetable mutation. These advancements have led to multiple clinical trials to examine the safety and efficacy of novel therapeutics that target tumor-specific molecular and genetic aberrations. While these advancements have demonstrated survival benefit in early phase clinical trials, continued investigation in randomized larger-scale trials is needed to further define the potential clinical impact of such therapy.


Introduction
Cholangiocarcinoma is defined by anatomic location and can arise throughout the biliary tree: distally, peri-hilar region, or intrahepatic. The different anatomic locations of cholangiocarcinoma correspond to varied etiologies of the disease. Specifically, extrahepatic cholangiocarcinoma likely derives from stem cells in peribiliary glands while intrahepatic cholangiocarcinoma (ICC) arises from hepatocyte stem cells [1,2]. In turn, the varied anatomic locations of cholangiocarcinoma can have implications for diagnosis, surgical planning, and resectability. In particular, ICC is the second most common primary hepatic malignancy. ICC is a rare cancer, with approximately 8000 cases diagnosed each year in the United States, comprising only 3% of gastrointestinal malignancies diagnosed globally per year [3,4]. The incidence of ICC has increased in recent decades, with some studies demonstrating a 14% increase in incidence per year since the early 1990s [5]. The increase in ICC incidence is likely due to the rising global prevalence of hepatitis C infection, as well as obesity associated non-alcoholic fatty liver disease and non-alcoholic steatohepatitis, which are known risk factors for ICC [5][6][7]. There are also regional differences in the risk factors associated with ICC. For example, the relative incidence of hepatolithiasis, liver fluke infection, as well as viral hepatitis is markedly higher among patients with ICC in

Pathogenesis of ICC
ICC most often occurs in the setting of chronic biliary inflammation and stasis. Patients with disease processes that promote chronic biliary stasis and inflammation (primary sclerosing cholangitis, hepatolithiasis, liver fluke infection, chronic hepatitis and cirrhosis) are therefore at risk of developing ICC [22,23]. Biliary stasis and inflammation promote excessive generation of nitric oxide, which potentiates oxidative damage to DNA and inhibits DNA repair and cellular apoptosis. Additionally, cyclo-oxygenase-2 (COX-2) is upregulated by induced nitric oxide species and acts to promote cholangiocyte growth and survival. Bile acids have also been demonstrated to facilitate cholangiocarcinogenesis through interference with cell signaling pathways and normal cellular apoptosis [22,23].
Common genetic mutations associated with ICC include KRAS, BRAF, TP53, and epidermal growth factor receptor [24][25][26][27][28][29][30]. KRAS mutations are the most commonly recognized mutation associated with ICC; however, their incidence varies widely from 8 to 53% [30]. BRAF mutations are also common and have a reported incidence of 0-22% in ICC tumors [30]. Interpretation of these data are complicated, however, as studies examin-ing genetic mutations associated with cholangiocarcinoma typically categorize ICC with peri-hilar cholangiocarcinoma and distal cholangiocarcinoma, which may have different genetic mutational patterns [25,31,32]. In a study of isolated ICC, 7% of ICC tumors were associated with KRAS mutations and 7% had BRAF mutations [30]. In this study, tumors associated with KRAS and BRAF mutations also had a more advanced TNM stage mostly due to lymph node metastasis. Tumor size, number, grade, evidence of vascular invasion or perineural invasion, or satellite lesions did not correlate with mutational status. Patients with tumors exhibiting either KRAS or BRAF mutations had worse median survival versus wild-type tumors (23 months versus 34 months, respectively, p = 0.05) [30]. Patients with KRAS mutations also had a worse 5 year overall survival compared with patients who had ICC tumors with BRAF mutations (13.5 months vs. 23.2 months, respectively, p = 0.05) [30].
More recently, studies employing gene expression profiling, high-density singlenucleotide array, and mutation analyses of formalin-fixed ICC samples have identified two predominant biological types of ICC: inflammatory and proliferative ( Figure 1) [32]. These analyses integrate components of previously distinct hypotheses on the epigenetic, molecular, and genetic origins of ICC. Inflammatory type ICC accounts for 38% of ICC and is defined by activation of pro-inflammatory signaling pathways via interleukin-10 (IL-10), interleukin-4 (IL-4), interleukin-6 (IL-6), overexpression of other cytokines, and activation of signal transducer and activator of transcription 3 protein (STAT3) (Figure 1) [32]. STAT3 has been noted to play a role in many types of cancer and has been associated with a worse prognosis in specific cancers. STAT3 is activated by upstream cytokines and growth factors, including IL-6 and human epidermal growth factor (EGFR) [33]. STAT3 relays signals from activated cytokine and growth factor-receptors at the cellular plasma membranes to the nucleus to promote gene transcription that promotes cellular differentiation, proliferation, angiogenesis, and immune response and inhibits apoptosis [33]. Gene expression that is induced by STAT3 produces cytokines that can re-activate STAT3 (i.e., IL-6), leading to an intrinsic and cyclic propagation of this oncogenic pathway [34]. Therefore, in conditions of chronic biliary inflammation and stasis that activate STAT3, STAT3 also acts to propagate local inflammation and oncogenesis.
The proliferative type of ICC accounts for 62% of ICC and is defined by activation of oncogenic signaling pathways, most notably through receptor tyrosine kinases (RTK) [32]. RTKs are cellular plasma membrane receptor proteins (consisting of an extracellular ligand binding domain, transmembrane helix, and intracellular domain) that mediate cellular communication and signaling [35]. There are 58 RTKs that are encoded within the human genome, which are grouped into 20 different classes, differentiated by signaling pathway and function [35]. The extracellular domain of RTKs binds growth factor ligands to signal downstream intracellular processes, including cellular proliferation, differentiation, motility, and metabolism [36]. Aberrant expression and mutations within genes encoding the component parts of RTKs (i.e., the extracellular domain, transmembrane helix, or the intracellular domain) promote various mechanisms of abnormal RTK function. The main mutations types that drive RTK dysfunction are gain of function mutations (EGFR), overexpression and genomic amplification of RTKs (EGFR, MET), chromosomal rearrangements (fusion of genes encoding BCR and ABL tyrosine kinases, or PDGFR tyrosine kinases), and constitutive activation by kinase domain duplication [36]. Furthermore, in subgroup analyses, Sia et al. demonstrated that proliferative-type ICC could be further divided into subclasses (named P1-P3), with distinct prognostic implications reflective of their gene expression predominance [32]. Proliferative-type ICC also demonstrated enrichment of several oncogenic pathways, that are described in hepatocellular cancer (cluster A, G3 proliferation, S1 signature), which may help explain its relatively worse prognosis when compared with inflammatory type ICC [32].  [32]. Summary of characteristics of ICC classes. Specific molecular and clinical characteristics differ between ICC classes. Molecular characteristics such as signatures of poor prognosis (i.e., cluster A, CC-like, G3, S1, S2, and stem-cell like ICC), oncogenic pathways (i.e., IGF1R, MET, EGFR), gene expression (i.e., EGFR, ILs), copy number variations, and oncogenes mutations (KRAS and EGFR) are differentially enriched in the proliferation and inflammation classes. Clinical characteristics such as moderate/poorly differentiated tumors and intraneural invasion are more frequent in the proliferation class. Differences in survival and recurrence were observed.
The proliferative type of ICC accounts for 62% of ICC and is defined by activation of oncogenic signaling pathways, most notably through receptor tyrosine kinases (RTK) [32]. RTKs are cellular plasma membrane receptor proteins (consisting of an extracellular ligand binding domain, transmembrane helix, and intracellular domain) that mediate cellular communication and signaling [35]. There are 58 RTKs that are encoded within the human genome, which are grouped into 20 different classes, differentiated by signaling pathway and function [35]. The extracellular domain of RTKs binds growth factor ligands to signal downstream intracellular processes, including cellular proliferation, differentiation, motility, and metabolism [36]. Aberrant expression and mutations within genes encoding the component parts of RTKs (i.e., the extracellular domain, transmembrane helix, or the intracellular domain) promote various mechanisms of abnormal RTK function. The main mutations types that drive RTK dysfunction are gain of function mutations (EGFR), overexpression and genomic amplification of RTKs (EGFR, MET), chromosomal rearrangements (fusion of genes encoding BCR and ABL tyrosine kinases, or PDGFR tyrosine kinases), and constitutive activation by kinase domain duplication [36]. Furthermore, in subgroup analyses, Sia et al. demonstrated that proliferative-type ICC could be further divided into subclasses (named P1-P3), with distinct prognostic implications reflective of their gene expression predominance [32]. Proliferative-type ICC also demonstrated enrichment of several oncogenic pathways, that are described in hepatocellular cancer (cluster A, G3 proliferation, S1 signature), which may help explain its relatively worse prognosis when compared with inflammatory type ICC [32].
When comparing clinical phenotypes of inflammatory versus proliferative ICC, several important differences have been described. For example, proliferative-type ICC tumors are more likely to be moderately to poorly differentiated, while inflammatory type tumors are more likely to be well differentiated. Survival analyses of proliferative versus  [32]. Summary of characteristics of ICC classes. Specific molecular and clinical characteristics differ between ICC classes. Molecular characteristics such as signatures of poor prognosis (i.e., cluster A, CC-like, G3, S1, S2, and stem-cell like ICC), oncogenic pathways (i.e., IGF1R, MET, EGFR), gene expression (i.e., EGFR, ILs), copy number variations, and oncogenes mutations (KRAS and EGFR) are differentially enriched in the proliferation and inflammation classes. Clinical characteristics such as moderate/poorly differentiated tumors and intraneural invasion are more frequent in the proliferation class. Differences in survival and recurrence were observed.
When comparing clinical phenotypes of inflammatory versus proliferative ICC, several important differences have been described. For example, proliferative-type ICC tumors are more likely to be moderately to poorly differentiated, while inflammatory type tumors are more likely to be well differentiated. Survival analyses of proliferative versus inflammatory type ICC reveal that patients with proliferative-type ICC have shorter time to recurrence (15 vs. 37 months, p = 0.03) and a reduced median survival (24.3 vs. 47.2, p = 0.048) [32]. In addition, the incidence of KRAS mutations was 8% in proliferative-type ICC tumors and 7% in inflammatory type ICC tumors; the incidence of BRAF mutations was 5% among proliferative-type ICC tumors and 2% among inflammatory type ICC tumors [32]. The proliferative class of ICC may share common progenitor cells with HCC, which may explain the similar aggressive phenotypes. On sub-analyses that utilized DNA copy number variation and gene expression-based classes to predict recurrence and survival, certain subclasses of the proliferative-type ICC had poor prognostic signatures similar to signatures associated with HCC ( Figure 1) [32]. This hypothesis may also be supported by the observation that progenitor cells, which give rise to both hepatocytes and cholangiocytes, are activated in adult human livers by hepatocyte replicative senescence and microenvironment inflammation and damage. Therefore, in the right epigenetic context, progenitor cells may give rise to both HCC and ICC or even cases of combined HCC-ICC [37,38].
The tumor microenvironment has also been used to define genomic and molecular subgroupings related to cholangiocarcinoma. To this point, Job et al. described four subtypes of ICC that were distinguished by cell type and functional transcriptomic markers within tumor specific microenvironments [39]. These subtypes correlated with prognosis and overall survival, highlighting that immunogenic ICC subtypes may be more amenable to treatment with immune checkpoint inhibitors than others ICC lesions [39]. Similarly, Andersen et al. profiled transcriptomes from resected cholangiocarcinoma within the context of their tumor microenvironment and identified certain microenvironment characteristics associated with worse prognosis [40]. These investigators also reported better treatment response to lapatinib, a dual inhibitor of EGFR and HER2, as well as trastuzumab, a selective inhibitor of HER2, among specific ICC subtypes [40].
Despite the considerable progress made in understanding the molecular and genetic underpinnings of ICC in recent decades, many of these studies have been only hypothesis generating. The data do, however, suggest a complicated network of microenvironment, molecular, and genetic factors that strongly influence cholangiocarcinogenesis. Importantly, there may be several different ICC phenotypes that reflect differences in etiology and pathogenesis. In turn, better characterization of different phenotypic-genomic ICC tumors has facilitated the investigation of novel and targeted treatment approaches for ICC.

Identified Targetable Mutations
The development of next-generation DNA sequencing has shifted the treatment landscape for many different types of cancer including ICC. In fact, up to 70% of ICC tumors may have at least one targetable gene aberration and on average have anywhere from 1 to 2 targetable mutations per tumor examined [41,42]. Several early small studies observed the most common mutations in ICC within AT-rich interactive domain-containing protein 1A (ARID1A, 36%), isocitrate dehydrogenase 1/2 (IDH1/2, 36%), TP53 (36%), and Myeloid Leukemia and Chlamydia (MCL1, 21%) genes [41]. Common actionable mutations (those with FDA-approved drugs available for treatment) included fibroblast growth factor receptor 2 (FGFR2, 14%), KRAS (11%), phosphatase and tensin homolog (PTEN, 11%),  [42]. After controlling for patient and disease factors, multivariable analyses demonstrate that TP53 aberrations were associated with worse survival (HR 1.68, p = 0.015), while FGFR aberrations were associated with improved overall survival (HR 0.478, p = 0.03) [42].  Programmed Death-Ligand 1 (PD-L1) expression has also been reported to be present in 10-70% of ICC tumor specimens [43][44][45]. PD-L1 positivity has been associated with more aggressive ICC characteristics and worse survival [43]. While uncommon, microsatellite-instability (MSI) has been explored as a biomarker and target for personalized ICC therapy. Due to the rarity of MSI in ICC, definitive conclusions regarding its incidence and prognostic implications have been difficult to decipher. The available data suggest Programmed Death-Ligand 1 (PD-L1) expression has also been reported to be present in 10-70% of ICC tumor specimens [43][44][45]. PD-L1 positivity has been associated with more aggressive ICC characteristics and worse survival [43]. While uncommon, microsatelliteinstability (MSI) has been explored as a biomarker and target for personalized ICC therapy. Due to the rarity of MSI in ICC, definitive conclusions regarding its incidence and prognostic implications have been difficult to decipher. The available data suggest that microsatellite unstable tumors (as defined by loss of DNA mismatch repair proteins MLH1, PMS2, MSH2, MSH6) are uncommon and occur only in a minority of patients with ICC (ranges from 1 to 10%) [43,46].

Results of Targeted Therapies for ICC
The evolution in the understanding of molecular and genetic pathways associated with ICC has facilitated several early phase clinical trials. These studies have generally investigated novel targeted therapies used either as mono-therapy or in combination with other systemic therapies. Novel therapies and their genetic or molecular targets has been well summarized in schematic form by Rizzo et al. [43] (Figure 3). Although various genetic aberrations have been identified in ICC, the incidence of many mutations remain low, which has been made the use of targeted therapy challenging. Early phase clinical trials have been initiated, however, for some of the more commonly occurring genetic aberrations including those involving IDH1/2, FGFR, EGFR, and VEGF [21,43].

Targeted Therapy: Isocitrate Dehydrogenase
Since IDH1/2 mutations are present in approximately 10-20% of ICC lesions, this genetic aberration has been the target of therapeutic intervention [41]. The mechanism by which IDH aberrations promote cholangiocarcinogenesis are still being fully elucidated. Animal models suggest that mutation of IDH induces an abnormal response to hepatocyte injury and inflammation, as well as silencing of HNF4-alpha, a transcription factor crucial to hepatocyte differentiation that is a potent anti-proliferative and tumor suppressor. In mouse models, mutant IDH-associated silencing of HNF4-alpha has resulted in a pro-neoplastic state with a biliary predominance [48].

Targeted Therapy: Isocitrate Dehydrogenase
Since IDH1/2 mutations are present in approximately 10-20% of ICC lesions, this genetic aberration has been the target of therapeutic intervention [41]. The mechanism by which IDH aberrations promote cholangiocarcinogenesis are still being fully elucidated. Animal models suggest that mutation of IDH induces an abnormal response to hepatocyte injury and inflammation, as well as silencing of HNF4-alpha, a transcription factor crucial to hepatocyte differentiation that is a potent anti-proliferative and tumor suppressor. In mouse models, mutant IDH-associated silencing of HNF4-alpha has resulted in a proneoplastic state with a biliary predominance [48].
The safety and efficacy of Ivosidenib, an inhibitor of mutated IDH1, have been investigated in a multicenter randomized double-blinded placebo-controlled phase 3 trial among patients with advanced cholangiocarcinoma who had progressed on either fluorouracil or gemcitabine-based chemotherapy [49]. The trial included patients with all types cholangiocarcinoma (i.e., intrahepatic, hilar, distal); however, the majority of accrued patients had ICC (89% in treatment arm and 77% in placebo arm) [49]. Among the 185 patients (124 in treatment arm and 61 in placebo arm), Ivosidenib was associated with an improved progression-free survival (median 2.7 months vs. 1.4 months, HR 0.37, 95% CI 0.25-0.54, p < 0.0001) and comparable safety profiles (30% of patients in the treatment arm patients had serious adverse events vs. 22% in the control arm) [49]. While other phase 1 and 2 studies have examined the safety profiles of inhibitors of mutated IDH1/2, a definitive benefit in the treatment of ICC has yet to be established [50].

Targeted Therapy: Fibroblast Growth Factor Receptor
FGFR aberrations have been identified in 10-15% of ICC tumors [41,42,51]. FGFR is expressed on multiple cell types and consists of four transmembrane receptors (FGFR1-4) with intracellular tyrosine kinase domains. The binding of these FGFR receptors leads to unregulated activation of several cellular proliferation pathways, including RAS-Raf-MAPK, JAK-STAT, and PI3-AKT-mTOR, leading to angiogenesis and unregulated cellular proliferation [52].
Pemigatinib is an oral kinase inhibitor, which inhibits FGFR 1, FGFR2, and FGFR3. Several Phase 2 trials have examined the safety and antitumor activity of Pemigatinib among patients with advanced cholangiocarcinoma with FGFR rearrangement. In one trial, among 146 enrolled patients, 35% had an objective treatment response, 42% of patients died from disease progression, and 45% of patients had serious adverse events [53]. Pemigatinib is currently being investigated in a phase 3, open-label, randomized, active-controlled multicenter trial that compares efficacy and safety of Pemigatinib versus standard of care (gemcitabine-cisplatin) in patients with advanced or metastatic cholangiocarcinoma with FGFR2 aberration (ClinicalTrials.gov, accessed on 21 September 2021, Identifier: NCT03656536). Futibatinib is a different highly selective irreversible oral inhibitor of FGFR 1-4. Its safety has been studied in a multicenter, phase 2 trial of patients with advanced or metastatic ICC with FGFR2 gene rearrangements who had disease progression after first line therapy (gemcitabine-cisplatin). Among 103 patients who enrolled, 34% had an objective response and the disease control rate was 76% at ≥6 months follow-up. Serious adverse events were experienced by 73% of patients (≥grade 3) [52]. The safety and efficacy of Futibatinib are now being compared with standard of care (gemcitabine-cisplatin) in a multicenter phase 3 study of patients with advanced cholangiocarcinoma with FGFR2 gene rearrangements (NCT04093362). In addition to these more well-known drugs, there are a number of other early phase studies examining drug safety and efficacy of inhibitors of mutant FGFR, as well as several upstream or downstream processes related to FGFR-based pathways [42,50]. Javle et al. published results from a single-arm phase II study that examined efficacy and safety of infigratinib, a FGFR-selective tyrosine kinase inhibitor, in patients with previously treated advanced cholangiocarcinoma [49,51,54]. In a cohort of 108 patients with advanced or metastatic cholangiocarcinoma, the objective response rate was 23%, the median duration of response was 5.0 months, and the median progressionfree survival was 7.3 months (95% CI 5.6-7.6 months) [51,54,55]. The majority of patients experienced hyperphosphatemia despite taking a prophylactic phosphate binder (77%) and non-severe eye disorders (68%), while a minority of patients experienced serious eye disorders (16%, central serous retinopathy and retinal pigment epithelium detachment) [54,55]. Based on these results, a phase III clinical trial of infigratinib versus standard of care gemcitabine/cisplatin is currently being conducted (NCT 03773302). Additional ongoing trials examining treatment efficacy and safety of FGFR targeted therapies are listed in Table 1.

Targeted Therapy: Epidermal Growth Factor Receptor
The HER tyrosine kinase family includes EGFR and HER1-4 aberrations. EGFR aberrations have been identified in 25% of ICC tumors [56]. EGFR is a subclass of the tyrosine kinase transmembrane receptor that binds to epidermal growth factor and activates signaling pathways involved in cell motility, cell adhesion, angiogenesis and invasion [57]. HER/EGFR aberration causes unregulated activation of these pathways.
Several early phase trials examining safety and efficacy of inhibitors of aberrant EGFR and HER (lapatinib, erlotinib, pertuzumab, trastuzumab) in patients with cholangiocarcinoma and other solid tumors are either pending or have failed to show significant objective treatment response. Analyses, have been limited, however, by patient selection and heterogeneity of diagnosis [50,58].

Targeted Therapy: Immune Checkpoint Inhibitors
Programmed Death-Ligand 1 (PD-L1) has been noted in 10-70% of ICC tumor specimens and its expression has been associated with tumor aggressiveness and diminished survival [43]. PD-L1 is expressed on tumor cells and binds PD-1 receptors on activated T cells to inhibit cytotoxic action, thereby resisting the immune mediated defense. The safety and efficacy of a number of PD-L1 inhibitors have been examined in relation to advanced or metastatic PD-L1-positive cholangiocarcinoma ( Table 2) [43]. These early phase trials failed to demonstrate definitive drug related benefits, yet initial results do suggest potential for both treatment efficacy and safety [43]. As such, while MSI-ICC is rare and small study numbers preclude robust analyses, some studies have reported relatively prolonged survival of patients with MSI-ICC treated with targeted immune checkpoint inhibitors. There are a number of ongoing early phase trials evaluating immune checkpoint inhibitor treatment efficacy and safety in patients with advanced MS-ICC (Table 3) [43].

Targeted Therapies: BRAF Mutations
BRAF mutations occur in 5-7% of ICC [63]. BRAF is a tyrosine kinase member of the RAS-RAF-MEK-ERK pathway (mitogen-activated protein kinase, MAPK), which is an integral pathway that mitigates cell proliferation, differentiation, transformation, and apoptosis [64]. BRAF mutations have been associated with higher TNM stage, resistance to systemic chemotherapies, and worse survival [26,30]. Targeted therapies pertaining to BRAF-positive ICC have demonstrated mixed results [65]. Many of studies to date have been "bucket" studies that included many different kinds of solid tumors harboring BRAF-mutations [65]. As such, interpretation of these data and the identification of diseasespecific efficacy has been challenging. While several ongoing trials pertaining to patients with metastatic biliary tract cancer are ongoing, results from these trials are still pending (Table 4) [65].

Conclusions
Considerable progress has been made in understanding the molecular and genetic pathogenesis of ICC in recent decades. This evolution in knowledge has facilitated development and study of novel targeted therapies for patients with advanced ICC. While some targeted therapies have demonstrated significant potential relative to disease progression and even survival, the data are still emerging. Notwithstanding these limitations, patients with advanced ICC should have genetic profiling performed to identify potential targeted therapies. As with many rare disease, it is critical that patients be appropriately enrolled in clinical trials to help better define the role, efficacy and safety of targeted therapies for ICC.