The systemic inflammatory response as a source of biomarkers and therapeutic targets in hepatocellular carcinoma

The pathogenesis of hepatocellular carcinoma (HCC) strongly relates to inflammation, with chronic up‐regulation of pro‐inflammatory mediators standing as a potential unifying mechanism that underscores the origin and progression of HCC independent of aetiology. Activation of the diverse pro‐inflammatory mediators either within the tumour or its microenvironment is part of an active cross‐talk between the progressive HCC and the host, which is known to influence clinical outcomes including recurrence after radical treatments and long‐term survival. A number of clinical biomarkers to measure the severity of cancer‐related inflammation are now available, most of which emerge from routine blood parameters including neutrophil, lymphocyte, platelet counts, as well as albuminaemia and C‐reactive protein levels. In this review, we summarise the body of evidence supporting the biologic qualification of inflammation‐based scores in HCC and review their potential in facilitating the prognostic assessment and treatment allocation in the individual patient. We also discuss the evidence to suggest modulation of tumour‐promoting inflammation may act as a source of novel therapeutic strategies in liver cancer.


| INTRODUC TI ON
The link between inflammation and cancer is well-established, with Rudolf Virchow first suggesting that 'lymphoreticular infiltrate' found next to the cancer site of origin had an active role in the pathogenesis of the disease. 1 Hepatocellular carcinoma (HCC) is the third leading cause of cancer deaths worldwide. 2 Crucially, inflammation has been recognised to have a major role in the tumourigenic process for HCC, with approximately 80% of cases developing as a consequence of chronic liver disease progressing to fibrosis and ultimately malignancy. 3 The different aetiologies of HCC include hepatotropic viral infection, 4 non-alcoholic steatohepatitis 5 and alcohol-induced fibrosis. 6 While these all induce the tumorigenic process via separate mechanisms, inflammation stands out as a unifying mechanism amongst all. The pro-inflammatory response also plays a significant role in facilitating the intercellular cross-talk between the tumour cells and the tumour microenvironment, which includes cancer-associated fibroblasts (CAFs), endothelial and immune cells. 7 The transition from fibrosis to carcinoma is accompanied by a persistent and unopposed release of cytokines such as interleukins (IL) and chemokines, further facilitating HCC pathogenesis. 8 Furthermore, inflammation does not only seem to have a role in the development of HCC but rather recent studies confirm that inflammation actually plays a prognostic role in determining the clinical course of the malignancy.
The prognostic role of inflammation is corroborated by studies demonstrating that within peritumoural tissue, a cytokine shift of Th-1 to Th-2 can impact the likelihood of HCC recurrence and mortality after radical resection. 9 Additionally, many single candidate studies illustrate that unopposed local or systemic release of | 2009 SANGHERA Et Al pro-inflammatory mediators can predict an adverse course of disease. For example, elevated peritumoural expression of IL-2 and IL-5 is predictive for an earlier recurrence and a shorter survival time post-resection. 10 Furthermore, an increased production of IL-10 has been associated with immune dysfunction, resulting in a higher number of myeloid-derived suppressor cells (MDSC) and impaired maturation of dendritic cells. 11 Combined, these molecular factors help explain the role of inflammation in influencing HCC progression. 12 A pro-inflammatory state in the tumour microenvironment has also been linked with augmented angiogenesis in many studies.
For example, in certain tumours the recruitment of IL-17 secreting T-helper cells helps facilitate angiogenesis and consequently negatively impacts prognosis. 13 Collectively, such studies demonstrate that it is the intricate interplay between the tumour itself, its microenvironment, the host's immune response and several concurrent Finally, given the number of studies exploring the prognostic role of cytokines in HCC, it is difficult to determine which candidate should be evaluated for use in the clinical setting.

| Biomarkers of systemic inflammation
Analysis of the inflammation-induced changes found in routine and accessible peripheral blood parameters gives additional insights into cancer-associated inflammation in HCC. For instance, studies have recognised numerous inflammation-related features in the peripheral blood of HCC patients. Such features include thrombocytosis, 14 leucocytosis, 15 hypoalbuminaemia, 16 increased plasma fibrinogen, 17 relative lymphopaenia, 18 hyperferritinaemia 19 and elevated C-reactive protein (CRP). 20 An increasing number of studies support the use of a combination of various acute phase proteins to develop composite, inflammation-based prognostic scores. These scores include the neutrophil-to-lymphocyte ratio (NLR), the platelet-to-lymphocyte ratio (PLR), the prognostic nutritional index, calculated using a nomogram based on hypoalbuminaemia and lymphopaenia (albumin in g/dL × 10 + 0.005 × total lymphocyte count), the prognostic index (PI), derived using elevated CRP (>1 mg/dL) and leukocytosis (>11.000/μL), 21 and finally the modified glasgow prognostic score (mGPS)-named the inflammation based index (IBI) when applied to HCC-which combines elevated CRP (>1 mg/dL) and hypoalbuminaemia (<35 g/L) ( Table 1). 22 The derivation of these inflammation-based scores, as well as an outline of the systemic inflammatory response and its consequences in HCC, are shown in Figure 1.
Throughout the last decade, there has been an unprecedented increase in the volume of research exploring the prognostic power of inflammation-based scores in cancer, with over 70 studies having explored the prognostic role of NLR in patients with solid tumours, 23 and likewise with the evaluation of GPS and mGPS. 24 In particular, the prognostic use of these inflammation-based scores has now expanded to HCC, with studies accessing each biomarker individually or in comparison, in both curative and palliative settings.
In this review, we summarise the current body of evidence surrounding the use of inflammation-based scores in HCC and discuss where they sit in the prognostic assessment of HCC with regard to currently used staging systems and treatment algorithms. Secondly, we deliver an insight into the biological mechanisms underlying the prognostic decline found in those with deranged inflammatory scores and examine if modulation of the cancer-related inflammatory response may provide a novel therapeutic strategy for HCC.
Finally, we look at the criticalities around the optimal clinical use of these inflammation-based scores in HCC.

| Neutrophil-to-lymphocyte ratio
The normal proportion of neutrophil and lymphocytes changes from 50% to 60% and 30% to 40% respectively during the acute phase

Key points
• In hepatocellular carcinoma, chronic upregulation of pro-inflammatory mediators within the tumour or its microenvironment is known to influence clinical outcomes, including recurrence after radical treatment and long-term survival.
• This underlying inflammation that is essential to the pathogenesis of HCC has been extensively studied, with the aim of devising clinical biomarkers to assess the severity of cancer-related inflammation.
• There is evidence supporting the biologic qualification of inflammation-based scores in HCC, which can be used to facilitate prognostic assessment and treatment allocation in patients.
• Modulation of tumour-promoting inflammation may be employed as an emerging therapeutic strategy in the management of HCC. response, giving rise to peripheral blood neutrophilia and relative lymphopenia respectively. In turn, this causes the relative ratio between neutrophils and lymphocytes to increase above the normal value of 2.
In solid tumours, angiogenesis and tumour progression is facilitated through the activation of oncogenes such as RAS and MYC, enabling a cytokine-rich tumour microenvironment and granulocyte recruitment and activation. 25 There is significant evidence that this granulocyte recruitment and activation is predominantly driven by the paracrine and endocrine actions of cytokines derived from hypoxic and necrotic tumorous tissue during a pro-angiogenic signalling cascade. The induction of these pathways involved in the hypoxic response, along with the production of pro-angiogenic cytokines, demonstrates a recognised mechanism of neutrophil migration to the peritumoural tissue. 26 During necrosis, the progression of hypoxia to anoxia in the tumour can also further elicit the innate immune response as a consequence of complement cascade activation, production of damage-associated molecular pattern molecules and opsonin release, leading to an overall rise in the neutrophil count. 27 The local release of pro-angiogenic cytokines, such as IL-17, has been shown to be essential to this process in HCC, where it fosters neutrophil chemotaxis to the tumour via CXC chemokines derived from epithelial cells, resulting in increased production of MMP-9 and angiogenesis. 28 IL-17 has also been shown to promote HCC growth and upregulated concentrations correlate with an elevated NLR score. 29 More recent evidence has linked tumour-derived CXCL5, the secretion of which is dictated by combined TGF-beta and Axl expression, with infiltration of neutrophils and disease-free survival (DFS), verifying the prognostic power of neutrophil infiltration in determining HCC progression. 30,31 Similarly, the CXCR2-CXCL1 axis is able to regulate neutrophil infiltration into HCC tumour tissue, indicating a poorer prognosis. 32 In addition to angiogenesis, neutrophil-mediated tumour promotion can be achieved by facilitating genomic instability, which arises as a consequence of the generation of reactive oxygen species (ROS) as part of the oxidative burst. 27 Secondly, the neutrophilfacilitated acquisition of an invasive phenotype can be mediated through interaction with the extracellular matrix via the production of proteases and initiation of hepatocyte growth factor signalling pathways. 33 Neutrophils are now known to possess an element of plasticity in their response, instead of being terminally differentiated innate immune effector cells. The variability in the sensitivity and response of neutrophils to different cytokines casts an additional layer of complexity in discerning the role they play in cancer-related inflammation. 34 A notable example is transforming growth factor-β (TGF-β), a tumour-derived cytokine involved in the progression and metastatic dissemination of HCC, 35 which also has the ability to polarise the neutrophil response from an antitumoural 'N1' phenotype to a protumoural 'N2' phenotype. Rather than solely relying on exclusive interactions with the tumour, this process instead involves complex inter-cellular cross-talk with additional agents of the tumour microenvironment, which includes lymphocytes, macrophages and stromal cells.
It is also well established that HCC patients with an elevated NLR have a worse prognosis after curative resection. In a ret-  inhibit antitumour CD8+ T cell and NK cell responses, thus promoting tumour progression. 40 Several other mechanisms that could explain the use of NLR as a prognostic factor have also been explored. Foremost, neutrophilia has been linked with inhibition of the cytolytic activity of immune cells, including lymphocytes, activated T cells and NK cells. 41 Neutrophils present in the intratumoural regions in HCC have been associated with increased autophagy activation and the release of growth factors and proteolytic enzymes, enabling poor survival and pro-tumourigenic effects such as invasion, metastasis and angiogenesis. 42,43 A recent study has also revealed the role of HCC derived CAFs in fostering neutrophil mediated immune suppression in HCC.
Specifically, HCC-CAF derived IL-6 was able to induce PD-L1+ neutrophils via the JAK-STAT3 pathway, which in turn impaired T cell function through PD-1/PD-L1 signalling and hence facilitated HCC progression. 44 Across a wide number of malignancies, numerous studies have confirmed that the interaction between the local immune response and systemic inflammation is because of a causal rather than casual relationship. 45 However, the molecular and immunological drivers of this relationship have not been fully explicated in HCC, calling for additional clarification in further studies.

| Platelet-to-lymphocyte ratio
In the context of acute inflammation, the reactive systemic thrombocytosis response enables resolving of tissue damage by stimulating local haemostasis and wound healing via focal production of a variety of platelet-derived humoral signals. However, this physiological response is adversely affected by the systemic release of cytokines in cancer. These cytokines act on platelets to establish an autostimulatory loop and subsequently cause the release of platelet-secreted mediators such as vascular endothelial growth factor (VEGF), platelet derived growth factor (PDGF) and others. 46 The reduction in platelets, associated with HCC treatment, and the rise in platelets, associated with HCC recurrence, indicate that platelets may also be promoted by HCC-derived factors. Such factors include IL-1, IL-6, G-CSF, M-CSF, thrombopoietin, FGF and VEGF, which have been shown to be higher in HCC than cirrhosis patients. 47 In HCC, the presence of underlying cirrhosis can influence platelet counts so absolute thrombocytosis may not be apparent.
This arises as a consequence of portal hypertension secondary to cirrhosis, which can lead to hypersplenism and associated thrombocytopaenia. 48 Studies have now integrated absolute platelet The systemic inflammatory response in hepatocellular carcinoma: relationship between the tumour microenvironment and the systemic consequences of a sustained cancer-related pro-inflammatory state counts with other liver function biomarkers to produce combined prognostic scores. This includes the aspartate aminotransferase (AST)-to-platelet ratio index, which has been suggested to have a prognostic role in assessing the likelihood of HCC recurrence after primary ablation or resection, particularly in hepatitis B virus (HBV) induced HCC. [49][50][51] Platelet counts may also reveal a more aggressive neoplastic phenotype that is independent of liver function, with some studies finding a correlation between thrombocytosis and adverse clinicopathological features. 48,52 Specifically, thrombocytosis has been linked to a strong risk of extrahepatic metastasis. 47 This can be attributed to the elevated levels of VEGF and PDGF associated with thrombocytosis, which in turn promotes angiogenesis, cell proliferation, migration and tumour metastasis.
Platelet-derived serotonin can also promote angiogenesis and the development of an invasive phenotype. 53 60 It has also been suggested that platelets might possess a more complicated immunopathological function in HCC by controlling the hepatic build-up of virus-specific CD8+ lymphocytes and heightening the necro-inflammatory damage locally, predisposing the onset and progression of HCC. 61 In addition to thrombocytosis, the mean platelet volume (MPV) has also been shown to have some prognostic value, with a high MPV being associated with improved survival in patients with advanced HCC. 62 This could be attributed to a low MPV representing functionally exhausted platelets or, alternatively, reduced thrombopoietin production resulting from reduced liver function. 63,64 While inflammation-driven reactive thrombocytosis is strongly Given the ready availability of antiplatelet agents in the clinical environment, there is a strong case for the prioritisation of this avenue of research.

| Inflammation-based index (mGPS)
During acute tissue injury, cells of the innate immune system are locally recruited and activated by chemotactic mediators acting in both a paracrine and autocrine manner, promoting and maintaining an inflammatory response locally. These mediators include IL-1, which is involved in thermoregulation, and IL-6, which acts as a hepatocyte stimulatory factor in the induction of the acute phase response. 68 IL-6 is synthesised acutely after local tissue injury and is released into the bloodstream, where it reduces the production of albumin and stimulates CRP production. 69 CRP is a homopentameric soluble acute-phase protein and is an important regulator of inflammatory processes, 70  Elevated CRP levels, combined with hypoalbuminaemia, were originally utilised to develop the GPS, which has sustained additional modifications (mGPS) to enhance the prognostic accuracy in individuals with early and advanced stage tumours in various treatment modalities. 24 In addition, data of large international cohorts have also illustrated mGPS to have an independent prognostic value irrespective of tumour stage, which correlates with a worse performance status (PS). 74 In addition to underlying liver dysfunction in HCC, systemic inflammation seems to play an equally relevant role in influencing albumin levels. 22 Hence, by modulating the underlying inflammatory response, anticancer therapies have been found to exert a positive effect on albumin levels. 75 In contrast, it remains unclear whether increased CRP secretion promotes tumorigenesis or is merely an innocent bystander that happens to sit downstream of excess systemic cytokine release. Nevertheless, it is apparent that elevated CRP secretion corresponds to poorer HCC prognosis in both the curative and palliative setting. Furthermore, the IBI has been confirmed as a stage independent prognostic indicator that has greater accuracy compared to other inflammatory indices, 22 with changes in its value after loco-regional treatment being predictive of disease modulation and survival in those with intermediate stage HCC. 75

| Inflammation-based prognostic indices in the curative setting
In accordance with the Barcelona Clinic Liver Cancer (BCLC) algorithm, it is recommended that patients with unifocal asymptomatic HCC and intact liver function should have the option of undergoing percutaneous radiofrequency ablation (RFA) or hepatectomy as radical treatment. 76 However, in patients that have a complete response after resection for early stage HCC the predicted 5-year survival rate is only 17% to 53%. 77 In those undergoing radical treatment for HCC the overall lifetime risk of recurrence is nearer to 70%. This figure is thought to take into account the likelihood of primary pro- and des-gamma-carboxyprothrombin. 89

| Inflammation as a therapeutic target in HCC
As HCC progression is closely linked to systemic inflammation, a hypothesised viable therapeutic strategy is to target the cancer-related inflammation pharmacologically, with the potential for integration with other systemic, loco-regional or molecular anticancer therapies. This section aims to review the evidence behind the therapeutic strategies in HCC.

| Targeting the systemic inflammatory response
With growing evidence that anti-inflammatory agents could play an antineoplastic role, there has been a shift towards re-purposing these therapeutics and integrating them as part of an anticancer regime. These anti-inflammatory therapeutics range from broad spectrum corticosteroids and non-steroidal anti-inflammatory drugs, to compounds that directly perturb specific molecular pathways.
However, the mechanism of action of most of these anti-inflammatory strategies remains unelucidated and needs further characterisation. Dexamethasone, for example, is already widely prescribed to counteract cancer-related cachexia and anorexia, particularly in patients with advanced malignancy. However, in murine models of HCC, there has also been evidence that dexamethasone may inhibit tumour growth by encouraging a shift from glycolysis to gluconeogenesis, 103 warranting further evaluation in prospective trials.
Aspirin has also emerged as both a chemopreventative and direct anticancer treatment, which is supported by robust retrospective evidence 104 and a plethora of prospective randomised trials across a range of different tumour types. 105 For example, a retrospective study in a small cohort of patients with unresectable HCC found that aspirin combined with TACE improves OS, 106 while another study proposed a synergistic antitumour effect of aspirin when combined with sorafenib. 107 It is currently unclear whether the anticancer properties of aspirin are mediated through its antiplatelet properties, reducing T-cell mediated necro-inflammation in the liver, 108 or instead relies on maintained inhibition of cyclooxygenase (COX), the expression of which in the tumour microenvironment is associated with a worse prognosis. 109 Foremost, active suppression of the inflammatory NFκB signalling cascade justifies the use of aspirin and other NSAIDS as disease modulating agents, since this pathway regulates both inflammation and the proliferation of tumour cells, and is also closely associated with the pathogenesis and advancement of HCC. 110,111 In addition to aspirin, other antiplatelet therapies also have the potential to suppress hepatic immunopathology and slow the progression of HCC. Platelets have been shown to be present within CD8+ T cell-containing hepatic necro-inflammatory foci and their depletion ameliorates the severity of liver disease. 112 A study evaluating the use of aspirin and clopidogrel in a HBV transgenic mouse model of chronic immune-mediated necro-inflammatory liver disease found that these antiplatelet therapies delayed the development of HCC and improved OS. This was associated with diminished accumulation of intrahepatic virus-specific CD8+ cells, leading to reduced recruitment of non-specific inflammatory cells, hepatocellular inflammation and injury, ultimately attenuating the sustained immune-mediated necro-inflammatory liver dysfunction that leads to HCC. 108 In addition, antiplatelet therapies may also inhibit platelet-derived factors that promote tumour growth independent of CD8+ T cells.
Another potential treatment option is the use of renin-angiotensin system (RAS) inhibitors. 113 In addition to being a therapeutic target in cardiovascular disease, RAS is expressed locally within many tissues where it regulates cellular proliferation and metabolism. Of interest, a number of preclinical studies have implicated the role of RAS in the development, growth and proliferation of cancer. 114 More specifically, RAS has been shown to promote tumour-associated inflammation and infiltration of pro-inflammatory cell through various mechanisms. 114,115 One such mechanism is increased production of pro-inflammatory cytokines through angiotensin II/angiotensin II type 1 receptor (AngII/AT1R) signalling, the main effector pathway of RAS implicated in HCC-related inflammation through the induction of downstream PKC and NF-κB expression. 113,116 These cytokines include TGF-β, IL-6 and IL-8, which can exert an immunosuppressive effect through control of lymphoid and myeloid cell recruitment and differentiation. [117][118][119][120] CRP and COX-II expression is also increased, the former of which is suspected to impede dendritic cell function and the latter reducing antitumour immunity. 121,122 AngII/AT1R can also increase cancer-related inflammation through promoting increased tumour infiltration by TAMs and facilitating the production of ROS, which can improve the function of both TAMs and regulatory T cells, while reducing the function of effector T cells. 119,[123][124][125][126] Significantly, in addition to their use in liver fibrosis and portal hypertension, the direct antitumour effects of RAS inhibitors have been shown to prolong survival outcomes in HCC patients. 127-130

| Molecularly targeted modulation of cancerrelated inflammation
Various targeted anti-inflammatory strategies, including the se- in turn leads to an immunosuppressive effect. However, there is also evidence that sorafenib has an immunomodulatory role. For instance, treatment with sorafenib has been shown to improve the immune response by upregulating the activity of tumour-specific effector T cells, as well as reducing the number of PD-1 expressing CD8+ T cells and regulatory T cells, which has been correlated with improved OS. [134][135][136] In addition to its direct immunomodulatory effect on lymphocytes, sorafenib may also exert an indirect effect on the HCC microenvironment by decreasing the levels of immunosuppressive cytokines such as IL-10 and TGF-β1. 136 There is also evidence that sorafenib can modulate dendric cells 137 and induce the proinflammatory activity of TAMs, stimulating an antitumour natural killer cell response. 101 In contrast, because of its anti-angiogenic nature, sorafenib has also been shown to increase tumour hypoxia and inflammation, which in turn exerts an overall immunosuppressive effect. One possible mechanism of this is the induction of SDF1α and CXCR4 axis by the hypoxic environment, which mediates the infiltration of pro-inflammatory cells including Gr-1+ MDSC, TAMs and regulatory T cells. 138,139 In addition, hypoxia also promotes the expression of immune regulatory proteins such as PD-1/PD-L1, facilitating immunosuppression. However, it may be possible to overcome this effect through the use of SDF1α/CXCR4 inhibitors and PD-1 blockade, which ultimately improves the efficacy of sorafenib and delays disease progression. 139,140 Transforming growth factor-β is another promising therapeutic target for HCC, because of its pleiotropic signalling effects within the tumour microenvironment. TGF-β modulates both tumour progression through stimulating angiogenesis, promoting EMT and inducing immune cell dysfunction. 35 Various strategies can be employed to target the TGF-β signalling cascade, including inhibiting TGF-β synthesis directly, ligand interference with monoclonal antibodies and selectively inhibiting the TGF-β receptor intracellular kinase domain. 141 Lerdelimumab and metelimumab, for example, are recombinant IgG 4 antibodies that target TGF-β1 and TGF-β2 respectively. Within the intracellular kinase inhibitors, the TGF-βR1 inhibitor galunisertib (LY2157299) has proven to be a leading candidate in a phase I study. 142 Consequently it is being evaluated in proof-ofconcept studies in numerous solid tumours, such as HCC, and also in combination with other immunomodulatory strategies, such as PD-L1 receptor blockade. 143 In general, an elevated NLR has been associated with worse outcomes, with higher pre-treatment scores being associated with poorer OS and PFS in patients treated with ICI for melanoma, nonsmall cell lung cancer (NSCLC) and genitourinary cancers. 144 A recent meta-analysis of 17 studies showed that a raised pre-treatment NLR was associated with a 1.81 increased risk of progression and 2.26 increased risk of death across malignancies. 145 This could be attributed to neutrophilia/lymphopenia, indicating an intrinsically immune-tolerogenic microenvironment, and leading to inferior response rates and a poorer prognosis. Interestingly, a number of studies have shown a superior predictive ability of post-treatment rather than pre-treatment NLR in correlating with outcome, suggesting systemic inflammation to be a dynamic rather than static predictive correlate of treatment benefit. 146,147 While PLR has also been linked to poorer OS and PFS in ICItreated patients, no significant difference was found in the context of the aforementioned meta-analysis, possibly because of the small number of relevant studies available. 145,148,149 Similarly, contradictory results have arisen when evaluating the predictive role of mGPS in the setting of ICI treatment. For instance, one study found that mGPS was superior to NLR and other inflammation-based scores in determining outcomes in NSCLC patients treated with nivolumab, while another study refuted such an association. 150,151 In a disease area where anticancer immunotherapy is rapidly In addition to improved treatment allocation on the basis of efficacy, several studies highlight the correlation between systemic inflammation and systemic anticancer treatment toxicity. This is derived from alterations in pharmacokinetic measures such as direct volume of distribution suppression of cytochrome P450 metabolism in the liver, which is a significant detoxification pathway involved in the clearance of sorafenib. 157 The combination of an inflammatory diathesis with sorafenib toxicity has been suggested to predict for prognosis in advanced HCC, a point that should be further investigated in prospective studies. 158 Even though research has helped to unravel the role of systemic inflammation, an enhanced understanding of its importance in the pathogenesis of HCC specifically will undoubtedly enable clinicians and scientists to facilitate the provision of personalised medicine for both early and more advanced stage HCC.

CO N FLI C T O F I NTE R E S T
No potential conflicts of interest. No financial support.

AUTH O R CO NTR I B UTI O N S
All authors equally contributed to this paper with conception and design of this study, literature review and analysis, drafting and critical revision and editing, and approval of the final version.