The Role Of Peritumoral Inflammatory Cells In The Prognosis Of Hepatocellular Carcinoma And Validation Of The "Immune Microenvironment Signature"

Hepatocellular carcinoma(HCC) is one of the most common malignant tumors,which is the fifth cause of cancer death and has been projected to claim over 564000 lives per year in the world. In china,HCC patients account for 50%of all the HCC cases worldwide and HCC is the second cause of cancer death in the nation.In recent years,the understanding of HCC in both clinic and basic research has developed.However,the 5-year survival rate is still 5%and the 5-year recurrence rate is as high as 60-70%no matter in east or west.The high recurrence rate after liver resection has severe effects on the patients' survival time,making the study on recurrence and metastasis of HCC important and imperative.The recurrence and metastasis of HCC is a multi-stage,multigenes-related,motional process. Including local invasion towards and entry into blood vasculature,survive in circulation,"homing" to target organs and growth of metastatic focus,the whole process is affected by the microenvironment,especially the inflammatory/immunological factors.As a hot spot for research recurrently,tumor microenvironment plays an important role in the recurrence and metastasis of HCC.Tumor microenvironment is the microenvironment for carcinogenesis and progression, which consists of tumor cells,fibroblasts,endothelial cells,innate and acquired immune cells, structure cells of tumor vessels,tissue-specific mesenchymas and their metabolites,products. Recent studies have shown that the microenvironment at the tumor site closely relates to the tumor does have vital effects on the carcinogenesis and progression.Some studies showed that immunological status of tumor microenvironment,not restricted by the TNM stages,is the most accurate independent factor for the prognosis of HCC patients.Thus a new concept "Cancer is the disease of microenvironment and immune system" has turned up recurrently.For the primary tumors,our study has shown that immunological imbalance has closed relationship with the potential invasion of tumor cells and the prognosis of the treatment.Furthermore,we have developed "immune microenvironment signature",which can predict the recurrence and metastasis of HCC.First,we study the role of peritumoral inflammatory cells in the HCC recurrence and metastasis.According to the "seed and soil" theory,the state of the target organs is one of the important "speed limitation" factors determining the recurrence and metastasis of the tumor.Since the target organ of recurrence and metastasis after the liver resection is usually the remnant liver, thus the microenvironment in the remnant liver,especially the inflammatory and immune microenvironment directly influences the prognosis after liver resection.Researchers have found that liver tissue microenvironment with chronic poison or inhibition of cell growth can selectively clone the mutational cells and form tumors.The tumor cells transplanted to the normal liver can't form focus,but when transplanted to the liver which is treated with Retrorsine(A drug can inhibit the normal cells growth),they can survive and form the tumor focus.The telomerase of hepatocyte in the cirrhotic liver is shortened and the capacity of regeneration is worse than that of normal hepatocyte.Thus it is a risk factor of HCC and easy to form the nodule.We consider that the recurrence of HCC is just due to selectively clone of the disseminated tumor cells or new mutational cells.Many studies has validated that liver cirrhosis is an independent risk for the recurrence of HCC after liver resection.There is a negative correlation between inflammatory status of remnant liver and the capacity of regeneration.About 15%of human cancer was found to be related to chronic inflammation and liver cancer is just the typical one.Clinical markers of inflammatory status of liver such as ALT,γ-GT et al,have been validated to have a positive correlation with the recurrence of HCC.Inflammatory cells have been found to participate in many courses such as angiogenesis,immune cytotoxicity,immune tolerance et al.But few researches are focused on the inflammatory cells in the peritumor microenvironment of liver.We think peritumoral tissue can represent the state of the remnant liver.Therefore,we studied inflammatory cells in the peritumoral liver tissue and explored the relationship between the inflammatory status of remnant liver tissue and the prognosis of HCC,especially the recurrence and metastasis of HCC.Macrophage(Mφ) and mast cell(MC),which are involved in inflammation extensively,are important components of innate immune system.Nowadays,they have been found to have a close relationship with tumors.Tumor-associated macrophage(TAM),as a classical immune cell,is one of the most important components in tumor microenvironment,and studies on TAM have uncovered its biphasic effect during tumor progression.It is found that TAM can accelerate cell proliferation,extracellular matrix remodeling and angiogenesis to promote tumor progression(M2 type function).However,as is well known,apart from its role as a "matrix cell",macrophage has been long recognized as a classical multifunctional immune cell,which can not only kill tumor cell by direct cell-cell contact through innate immune response,but also perform as an antigen-presenting cell(APC) and activate adaptive immune response resulting in effective antitumor effect(M1 type function).TAM was reported to be associated with favorable prognosis in several cancer types,such as stomach,lung and colorectal cancers.However,in other cases such as breast,prostate,ovarian and cervical cancers,TAM was found to have an active role in tumor invasion and metastasis.The role of TAM in HCC is still controversial and few studies are focused on the relationship between macrophage in peritumor microenvironment and HCC.As one of the primary inflammatory cells in humans,MC can generate lots of inflammatory mediators,and significantly influence multiple features of chronic inflammatory responses,through diverse effects that can either promote or suppress aspects of these responses.MCs also have peculiar function on carcinogenesis and progression.They are recruited to and activated in the microenvironment of a developing tumor for their benefit,and consequently result in an adverse outcome in squamous cell carcinoma of the esophagus and lung adenocarcinoma.Conversely,they have been shown to confer a better prognosis in breast and ovarian cancers.Such discrepancy is probably due to the selective release of beneficial and destructive mediators by MCs which are local environment-dependent.Most previous studies were focused on MCs' contributions to tumor angiogenesis.However,MCs display a diverse range of biological effects besides angiogenesis, among which dampening the immunosurveillance against tumors is of particular importance for certain malignant tumors,especially cutaneous malignancies.In the context of HCC,intratumoral MCs accumulation has been detected,and they are proved to promote HCC cells growth in vitro, while the relevance of the peritumoral counterparts of the intratumoral MCs to the prognosis of HCC still remains speculative.It has been shown that Mφs and MCs are cooperative in inflammation and tumor progression through mediators they secret and direct touch.Mφs and MCs are not only the main participators of chronic inflammation in liver,but also play an important role in tumor progression.Therefore,we think Mφs and MCs can represent the inflammatory status of the remnant liver and by studying them,we can learn the role of peritumor inflammatory status in the prognosis of HCC,especially the recurrence and metastasis of HCC.Then we validated "immune microenvironment signature" which consists of 8 effector immune response-related genes(GNLY,GZMB,TRAV10,CD3Z,TBX21,IFNG,GATA3,IRF1)."Immune microenvironment signature" is a recurrence score model which has the power to accurately stratify patients into high vs.low recurrent risk,based on the expression profiles of 8 effector immune response-related genes.The model also accords with the theory "Metastasis of cancer depends on cancer cells,tumor microenvironment and genetic background of host".In the second part,we validated whether this model was reproducible and powerful for prospective patient prediction in 298 randomly selected patients.Last,we combined the peritumor inflammatory cells with "immune microenvironment signature" to predict the recurrence/metastasis of HCC after liver resection.The primary tumor, tumor microenvironment and tissue microenvironment have different contributions in the different stages of HCC.In the initial stage,the liver tissue microenvironment plays the main role in carcinogenesis.Then the primary tumor and tumor microenvironment have the largest contributions during tumor progression.After liver resection,the primary tumor and its microenvironment removed,the tissue microenvironment of remnant liver is in charge.In the second part of study,we found that some patients in low recurrent risk group quickly reoccurred after resection and some in the high risk group had no recurrence.Therefore we analyzed "immune microenvironment signature" and found that the model was focused on effector and adaptive immune response-related genes in tumor microenvironment,without any relationship with peritumor microenvironment which is in charge after resection.Thus we used inflammatory cells in peritumor microenvironment to explain the phenomena and then combined the peritumor inflammatory cells with "immune microenvironment signature" to predict the recurrence/metastasis of HCC.The role of peritumoral inflammatory cells in the prognosis of hepatocellular carcinomaThis study aims to investigate whether distribution state of Mφand MC is associated with clinical characters and the prognosis of HCC patients.We randomly selected 298 HCC patients,and obtained their wax specimens.Using immunohistochemistry and tissue microarray,we investigated the distribution of peritumoral Mφs and MCs and analyzed the results by correlation and survival analysis.The results:(1)Mφs had a positive correlation with MCs in the peritumor(R=0.442,P＜0.001).(2)MCs were found more in highγ-GT(P=0.001),large tumor size(＞5cm)(P=0.04), thrombosis positive(P=0.019) and BCLC:B-C stages(P=0.007) patients.While Mφs were only more in thrombosis positive patients than those in thrombosis absence patients.(3)On univariate analysis,low number of MC(P=0.004,P=0.001) and Mφ(P=0.003,P=0.027) in the peritumor correlated with prolonged OS and DFS.It was also showed that combination of MC and Mφ(P=0.001,P=0.001) also had a negative correlation with OS and DFS.(4)When we divided 298 patients into groups with the number of MC and Mφ:groupⅠ=both high;groupⅡ=either high;groupⅢ=both low.Using Kaplan-Meirer Survival Analysis,it was showed that groupⅠhad shortened OS(P=0.001) and DFS(P＜0.001),while groupⅢhad prolonged OS(P=0.005) and DFS(P=0.016).(5)On multivariate analysis,we found that MC in the peritumor was the independent predictor on DFS(P=0.002) and Mφhad negative correlation with OS(P=0.003).Combination of MC and Mφwas another independent predictor on OS(P=0.004) and DFS(P=0.048).(6) ROC curve:we found that the area under curve(AUC) of combination of MC and Mφ(AUC_os=0.608,P=0.001) was only smaller than that ofγ-GT(AUC_os=0.646,P＜0.001) on OS,and larger than that of BCLC Stage(AUC_os=0.606,P=0.002),MC(AUC_os=0.586,P=0.011),Mφ(AUC_os=0.573,P=0.031),differentiation(AUC_os=0.600,P=0.003).When it comes to DFS,the AUC of combination of MC and Mφ(AUC_recurrence=0.594,P=0.005) was only under that of BCLC Stage(AUC_recurrence=0.616,P=0.001) and larger than that of MC(AUC_recurrence=0.592,P=0.006), capsule(AUC_recurrence=0.588,P=0.009),and AFP(AUC_recurrence=0.574,P=0.027).In conclusion,the results suggest that as a pivotal component of innate immune system,MC and Mφrepresent the inflammatory status of peritumor microenvironment.There is a negative correlation between the inflammatory status of peritumor microenvironment and progression of HCC.A further investigation on the exact role of inflammatory response in the peritumor microenvironment may substantially promote studies on HCC invasion,metastasis and immune escape,as well as bear fundamental clinical significance.Validation of "immune microenvironment signature"In this study,we designed to validate "immune microenvironment signature",using real time RT-PCR on 298 randomly selected patients.Through comprehensive literature searching,Gao et al in our institute identified 27 marker genes symboling or representing effector immune response (mainly adaptive immunity),immunosuppression and inflammatory response as target genes and enrolled an independent as well as random cohort of surgical HCC patients from Zhongshan Hospital,Fudan University.After optimizing the PCR array platform,the expression profiles of these 27 genes were detected using our specifically customized PCR arrays.2 genes of extremely low expression or unexpected expression were excluded for data analysis.The open source software Cluster and Treeview were used in conducting retrospective unsupervised hierarchical cluster and correlation matrix analysis,followed by survival analysis.Finally,an "immune microenvironment signature" was generated by the well-recognized recurrence score system for prospective patient stratification.That was:rs8=(-3.108×IFNG value) +(-3.534×TBX21 value) +(-2.975×GATA3 value) +(-4.361×GZMB value) +(-5.530×IRF1 value) +(-3.497×GNLY value) +(-3.154×CD3Z value) + (-3.932×TRAV10 value)The higher rs8 value is,the more risk of recurrence/metastasis is.The model demonstrates that an effector immune response(mainly adaptive immunity) dominated microenvironment plays a central role in HCC recurrence and it is clinically feasible and rational.We have already detected the 8 genes(GNLY,GZMB,TRAV10,CD3Z,TBX21,IFNG,GATA3,IRF1) that are used in "immune microenvironment signature" and evaluated the rs8 values. We divided 298 patients into two groups by cutoff value which was made before and analyzed "immune microenvironment signature" by survival analysis and ROC curve.The results:(1)On univariate analysis,it was demonstrated that "immune microenvironment signature" had a negative correlation with both OS(P＜0.001) and DFS(P＜0.001).And on multivariate analysis,the model was still an independent factor for OS(P＜0.001) and DFS(P＜0.001).(2)ROC curve:we took the factors that were significant in multivariate analysis into ROC curve analysis.And we found "immune microenvironment signature" had the largest AUC value (AUC_recurrence=0.620,P＜0.001).Others were BCLC Stage(AUC_recurrence=0.616,P=0.001),capsule (AUC_recurrence=0.588,P=0.009) and AFP(AUC_recurrence=0.574,P=0.027).Taken together,"immune microenvironment signature" has a good prediction on the recurrence/metastasis of HCC and its power is even better than BCLC Stage System and other clinical factors.The model which is steady and feasiable,is of clinical significance.The relationship between peritumoral inflammatory cells and "immune microenvironment signature"In the second part of the study,we found a strange phenomena that some patients in low recurrent risk group quickly reoccurred after resection and some in the high risk group had no recurrence.Therefore we analyzed "immune microenvironment signature" and found that the model was focused on effector and adaptive immune response-related genes in tumor microenvironment, without any relationship with peritumor microenvironment which is in charge after resection.Thus in this part,we used inflammatory cells in peritumor microenvironment to explain the phenomena and then combined the inflammatory cells with "immune microenvironment signature" to predict the recurrence/metastasis of HCC.According to rs8 value and recurrence state,we divided the patients into 4 groups:groupⅠ=low risk but recurrence;groupⅡ=low risk and no recurrence;groupⅢ=high risk and recurrence;groupⅣ=high risk but no recurrence.Then we compared the inflammatory cells(MC and Mφ) in each group and took ROC curve analysis.The results:1) When taking account of all the patientsIt was showed that Mφs(P=0.04) and MCs(P＜0.001) were recruited more in groupⅠthan that in groupⅣand the number of MCs(P=0.006,P=0.007) accumulated in groupⅠandⅢwas more,compared with groupⅡandⅣrespectively.We also found MCs(P=0.044) accumulated less in groupⅣthan that in groupⅡ.2) When taking account of early recurrence patients(1)Cutoff with 2 yearsWe found that Mφs(P=0.024) and MCs(P＜0.001) were recruited more in groupⅠthan that in groupⅣand the number of MCs(P=0.002,P=0.008) accumulated in groupⅠandⅢwas more, compared with groupⅡandⅣrespectively.(2) Cutoff with 1 yearIt was demonstrated that Mφs(P=0.03) and MCs(P＜0.001) were recruited more in groupⅠthan that in groupⅣand the number of MCs(P=0.008,P=0.004) accumulated in groupⅠandⅢwas more,compared with groupⅡandⅣrespectively.We also found Mφs(P=0.032) accumulated less in groupⅣthan that in groupⅢ3) ROC curveIt was showed that the value of AUC in the combination of "immune microenvironment signature" and peritumoral MCs(rs8＆MC)(AUC_recurrence=0.666,P＜0.001) was the largest.Others were combination of model and peritumoral Mφs(rs8＆Mφ)(AUC_recurrence=0.634,P＜0.001),the model(rs8)(AUC_recurrence=0.620,P＜0.001),combination of MC and Mφ(MC&Mφ)(AUC_recurrence=0.594,P=0.005),MC(AUC_recurrence=0.592,P=0.006),BCLC Stage (AUC_recurrence=0.616,P=0.001),Capsule(AUC_recurrence=0.588,P=0.009).Collectively,"immune microenvironment signature" has a good prediction on the recurrence/metastasis of HCC with the value of rs8.But if we take account of the inflammatory status of peritumor microenvironment,the power of prediction will be larger.Conclusion(1) Peritumor Mφhas a negative correlation with OS of HCC patients and MC promotes the recurrence/metastasis of HCC.They are cooperative in the tumor progression and combination of MCs and Mφs is an independent factor for the prognosis of HCC patients.(2) "Immune microenvironment signature" has a good prediction on the recurrence/metastasis of HCC.The model is steady,feasible and of clinical significance.(3) We find that there are more MCs and Mφs recruited in the group of low risk but recurrence than that in the group of high risk but no recurrence.It is demonstrated in the first part that MCs and Mφs could promote the recurrence and metastasis of HCC.That's the reason why the confused phenomena occurred.(4) Combination of "immune microenvironment signature" and peritumoral MCs has a better prediction power on the recurrence/metastasis of HCC,compared with the model itself and the combination of the model and peritumoral Mφs.The novelty of this study(1) For the first time,in a large cohort of HCC patients,we demonstrated that MCs and Mφs in the peritumor microenvironment might cooperate in promoting the recurrence/metastasis of HCC after liver resection.(2) Combination analysis of tumor microenvironment and target organ microenvironment of HCC is a valid method to predict the recurrence/metastasis of HCC.The potential application of this studyThe inflammation in the peritumor microenvironment can promote the recurrence/metastasis of HCC.Thus it can provide a new way to prevent the recurrence/metastasis of HCC.Analysis of both peritumor inflammation and "immune microenvironment signature" has a better prediction for HCC recurrence,which could be applied to select high-risk patients for appropriate prophylactic treatment.