肝細胞癌的免疫組成:從標靶藥物到免疫治療

Immune Contexture of Hepatocellular Carcinoma: From Sorafenib to Immune Checkpoint Inhibitors

呂理駿(Li-Chun Lu)

指導教授 ： 徐志宏 ； 鄭安理

國立臺灣大學/醫學院/腫瘤醫學研究所/博士(2020年)

https://doi.org/10.6342/NTU202000150

摘要

肝細胞癌(肝癌)一直以來都是全球死亡率前幾名的癌症之一。晚期肝癌病患無法接受局部治療，而全身性治療，包括「抗血管新生標靶藥物」和「抗計畫性細胞死亡蛋白-1 (PD-1)為主的免疫檢查點抑制劑(ICIs)」，是標準治療。近來，腫瘤微環境(TME)和全身性治療療效間的關係已成為主要的研究課題。蕾莎瓦 (sorafenib)為一抑制多重激酶的標靶藥物 (包含抗血管新生的療效)，也是第一個被核准在晚期肝癌患者使用的藥物，過去研究指出sorafenib具有多重免疫調節的作用，但是其對晚期肝癌TME內的計畫性細胞死亡配位基-1 (PD-L1)表現的影響並不清楚。另外，TME的免疫組成會影響ICIs的療效，但是在肝癌不同器官的免疫組成，是否會造成ICIs在不同器官轉移腫瘤的差別反應也不清楚。
在本研究的第一部份，要驗證的假設為:sorafenib會對肝癌TME內的PD-L1表現有影響。我們回顧有接受過sorafenib的晚期肝癌病患，收集其sorafenib治療前後的配對腫瘤檢體，利用免疫組織染色，半定量分析檢體內腫瘤細胞(TCs)及免疫細胞(ICs)其細胞膜上PD-L1的表現。在23對可分析的檢體中，我們發現ICs的PD-L1表現在sorafenib治療惡化後的檢體顯著高於治療前的；然而，TCs的PD-L1表現在sorafenib治療前後並無顯著差異。我們進一步證明，這些表現PD-L1的ICs，也同步表現巨噬細胞的標記CD68，暗示這些表現PD-L1的巨噬細胞在sorafenib治療後而惡化的肝細胞癌可能有特殊意義。
在本研究的第二部份，要驗證的假設為:各器官TME不同的免疫組成，會造成在各器官內的肝癌腫瘤對ICIs的差別反應。我們回顧曾接受ICIs治療並有可量測腫瘤大小的晚期肝癌病患，分別評估ICIs對其肝、肺、淋巴結、及腹腔內轉移腫瘤的反應率。在75位接受分析的晚期肝癌病患中，有58、34、19及18位患者有可量測之肝、肺、淋巴結、及腹腔內轉移腫瘤，其相對應的特定器官腫瘤反應率分別為22.4%、41.2%、26.3%、及38.9%。在39位同時有肝臟腫瘤及肝外轉移的患者中，12位其肝外轉移腫瘤有獲得控制，但是肝臟腫瘤卻惡化；反之只有四位患者肝臟腫瘤獲得控制，但是肝外轉移腫瘤有惡化(P = .046, McNemar test)。而在16位治療前只有肝臟腫瘤和肺部轉移的患者中，8位其肺部轉移獲得控制，但是肝臟腫瘤卻惡化；反之沒有一位患者有肝臟腫瘤獲得控制，但是肺部轉移腫瘤卻惡化的情形(P = .005)。在本部分的研究，我們發現肝內的肝癌反應率似乎比肝外轉移的腫瘤來得差，而肺部轉移的肝癌對ICIs的反應率最好，暗示TME的免疫組成，會影響對ICIs的療效。
為了進一步探究上述「ICIs對肝癌之不同器官轉移的療效差別反應」之可能機轉，我們利用BNL小鼠肝癌細胞株，分別建立包括原位肝臟肝癌、皮下、及肺臟植入肝癌的同源小鼠模式。我們分析未接受治療的不同器官肝癌小鼠模式，分離腫瘤浸潤CD45陽性白血球和分析其次族群細胞組成，發現肝臟腫瘤比肺部腫瘤有趨勢上較高比例的巨噬細胞及骨髓抑制細胞，不過CD4及CD8陽性T細胞在肝臟及肺部腫瘤的比例則相近。然而，肝臟腫瘤浸潤的CD8 T細胞，比起其他器官腫瘤的有較高比例會表現衰竭標記。此外，我們發現VEGF和其他一些免疫抑制細胞激素，在未接受治療的原位肝腫瘤內比皮下腫瘤來的高；而在PD-L1抗體治療後，其他免疫抑制激素在原位肝腫瘤內有下降，但VEGF和IDO等的表現仍持續比皮下腫瘤顯著的高。在接受PD-L1抗體治療後，我們發現皮下植入的肝癌腫瘤療效好，而原位肝臟植入的肝癌腫瘤反應較差。整體而言，我們的小鼠模式研究結果發現原位肝癌的TME，比植入其他器官的更傾向免疫抑制，支持「不同器官相異的免疫組成，可能影響ICIs的治療效果」。
總結上述研究，首先我們發現在sorafenib治療後惡化的肝癌檢體，其ICs (特別是巨噬細胞) 的PD-L1會顯著上升。其次，我們利用臨床肝癌病人的資料證實，各器官內的肝癌腫瘤對ICIs有差別反應，肝內的肝癌反應率比肝外轉移的腫瘤來得差。在機轉的研究上，我們分析小鼠原位肝癌，有較高比例的巨噬細胞、有較多的衰竭CD8 T細胞、及較高表現的VEGF，使得肝內TME傾向免疫抑制，因此可造成ICIs的療效比其他器官的肝轉移來得差。我們的研究成果可提供未來晚期肝癌全身性治療發展的參考。

關鍵字

肝細胞癌 ； 蕾莎瓦 ； 計畫性細胞死亡配位基-1 ； 免疫檢查點抑制劑 ； 腫瘤微環境 ； 免疫組成

並列摘要

Hepatocellular carcinoma (HCC) has been the leading cause of cancer mortality globally for decades. In cases of advanced HCC, for which locoregional therapy is unsuitable, systemic anticancer therapy through either antiangiogenic agents or anti-programmed-cell-death protein 1 (PD-1)-based immune checkpoint inhibitors (ICIs) is the standard of care. Recently, the interaction between the tumor microenvironment (TME) and systemic therapy response has been a major focus of research. Sorafenib, a multikinase inhibitor with antiangiogenic properties and the first systemic therapy approved for HCC, has been demonstrated to exhibit various immunomodulatory effects. However, the impact of sorafenib on programmed death ligand-1 (PD-L1) expression in the TME of advanced HCC was unclear. In addition, the antitumor effects of ICIs are influenced by the immune contexture of the TME. However, whether distinct immune contextures in different organ systems contribute to the variable tumor response to ICIs in HCC was unknown.
The first part of the thesis addressed the hypothesis that sorafenib treatment affects the expression of PD-L1 in the TME of HCC. We identified patients who received sorafenib for advanced HCC and who had paired tumor tissues obtained before and after sorafenib treatment. We analyzed the tissue slides using immunohistochemistry to semiquantitatively score the membrane PD-L1 staining in tumor cells (TCs) or tumor-infiltrating immune cells (ICs). In 23 paired HCC tissues, PD-L1 expression in ICs was increased in HCC tissues after sorafenib treatment. However, PD-L1 expression in TCs did not increase significantly after sorafenib treatment. We also demonstrated that PD-L1-expressing ICs were highly co-localized with CD68-positive tumor-infiltrating macrophages, suggesting that PD-L1-expressing macrophages play roles in HCC progression after treatment with sorafenib.
The second part of the thesis explored the hypothesis that the distinct immune contexture of the TME is associated with variable tumor response to ICIs in HCC. We reviewed data from patients with advanced HCC who had received ICIs and had measurable diseases. The objective response to ICIs in tumors located in different organ systems—such as the liver, lung, lymph nodes, and other intra-abdominal sites—was evaluated independently. Among the 75 enrolled patients, 58, 34, 19, and 18 patients had measurable hepatic tumors and lung, lymph node, and other intra-abdominal metastases, respectively, with the corresponding organ-specific objective response rates being 22.4%, 41.2%, 26.3%, and 38.9%. Among the 39 patients who had both hepatic and extrahepatic tumors, 12 had disease control in the extrahepatic tumors but progressive disease (PD) in the hepatic tumors, whereas only 4 exhibited disease control in the hepatic tumors but PD in the extrahepatic tumors (P = .046). Among the 16 patients with evaluable tumors in only the liver and lung at baseline, 8 had disease control in the lung and PD in the liver, and none had disease control in the liver but PD in the lung (P = .005). In the present study, we demonstrated that hepatic tumors of HCC are less responsive to ICIs than extrahepatic lesions, and lung metastases respond most favorably to ICIs, suggesting that the TME influences the efficacy of ICIs.
To explore further the mechanisms underlying the aforementioned organ-specific differential responses to ICI treatment in HCC, we employed a syngeneic mouse liver cancer model in immunocompetent mice. We used BNL mouse liver cancer cells to establish orthotopic, subcutaneous, and lung HCC mouse models independently. We isolated tumor-infiltrating CD45+ leukocytes from untreated tumors at various organ sites. The percentages of overall macrophages and myeloid-derived suppressor cells in liver tumors were numerically higher than those in lung tumors. Conversely, the overall percentages of CD4+ and CD8+ T cells were similar between liver and lung tumors. However, the tumor-infiltrating CD8+ T cells were more exhausted in the liver tumors than in tumors of other organs. In addition, we evaluated the soluble factors in untreated liver and subcutaneous tumors. We observed that the levels of vascular endothelial growth factor (VEGF) and some other immunosuppressive cytokines were higher in the liver than in subcutaneous tumors. After treatment with PD-L1, although the levels of other immunosuppressive cytokines were decreased in the liver, the levels of VEGF and indoleamine 2,3- dioxygenase remained high in liver tumors. Anti-PD-L1 antibody or isotype control was administered to the orthotopic, subcutaneous, or combination models. The results showed that orthotopic HCC tumors responded more poorly to anti-PD-L1 therapy than subcutaneous HCC tumors. Overall, the data demonstrated that BNL mouse liver cancer cells, when implanted in the liver, were associated with a more immunosuppressive TME than those implanted in other organs. The results supported that the distinct immune contextures of different organ sites influenced the efficacy of ICI treatment in mouse HCC models.
Overall, we found that increased PD-L1 expression in ICs, mainly macrophages, was associated with HCC progression after treatment with sorafenib. In addition, we demonstrated organ-specific responses to ICIs in patients with advanced HCC. HCC in the liver was less responsive to ICIs than were extrahepatic metastases. We demonstrated mechanistically that orthotopic liver tumors had higher percentages of macrophages, more exhausted CD8+ T cells, and higher VEGF levels, which could render the TME more immunosuppressive and lead to a poorer ICI response than in other organ systems. Our findings provide insights that could facilitate the development of systemic therapy for advanced HCC in the future.

並列關鍵字

hepatocellular carcinoma ； sorafenib ； PD-L1 ； immune checkpoint inhibitors ； tumor microenvironment ； immune contexture

參考文獻

1. Bray F, Ferlay J, Soerjomataram I, et al: Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 68:394-424, 2018

2. Hsiue EH-CH, Lo W-C, Lin H-H, et al: Abstract 286: Forecast of the incidence and etiology of liver cancer in Taiwan, Japan, United States, and United Kingdom: toward harmonization of East and West. Cancer Research 77:286-286, 2017

3. Forner A, Reig M, Bruix J: Hepatocellular carcinoma. Lancet 391:1301-1314, 2018

4. Llovet JM, Ricci S, Mazzaferro V, et al: Sorafenib in advanced hepatocellular carcinoma. N Engl J Med 359:378-90, 2008

5. Cheng AL, Kang YK, Chen Z, et al: Efficacy and safety of sorafenib in patients in the Asia-Pacific region with advanced hepatocellular carcinoma: a phase III randomised, double-blind, placebo-controlled trial. Lancet Oncol 10:25-34, 2009