肝細胞癌藥物治療之研究

Study of Systemic Therapy for Hepatocellular Carcinoma

許駿(Chiun Hsu)

指導教授 ： 鄭安理 ； 賴明陽

國立臺灣大學/醫學院/臨床醫學研究所/博士(2004年)

https://doi.org/10.6342/NTU.2004.01876

摘要

肝細胞癌﹝hepatocellular carcinoma, HCC﹞是台灣最常見的癌症之一。根治性手術只適合用於少數﹝約20%至30%﹞的病患。大部分的病患都在相當短的時間內死於局部進行性腫瘤或遠處轉移腫瘤。HCC病患往往因為本身肝硬化以及伴隨出現的脾臟功能亢進﹝hypersplenism﹞及血球降低，使得全身性化學治療的副作用太大而無法承受。因此，發展具有療效且副作用低的治療方式，對於HCC病患而言是非常重要的。

在本論文中我們嘗試研究兩種不同的治療廣泛性﹝advanced﹞HCC的方式。第一種方式是抗血管新生治療﹝anti-angiogenesis therapy﹞。癌細胞誘發之血管新生現象在癌症的發生以及惡化過程中均扮演關鍵的角色。傳統的化學治療毒性主要針對癌細胞本身，而抗血管新生治療則是針對經癌細胞誘發而異常增生的血管內皮細胞。與化學治療相比，抗血管新生治療具有藥物動力學上的優勢，因為藥物對於腫瘤血管的可接近性﹝accessibility﹞較腫瘤本身為佳。血管內皮細胞本身是基因正常的細胞，因此產生抗藥性的機會應該也較癌細胞為少。HCC 為一典型的高度血管異常增生的腫瘤，因此抗血管新生治療對於廣泛性HCC病患應該是一個合理的選擇。

第二種方式是細胞訊息傳導路徑﹝signal transduction pathways﹞的調控。 細胞訊息傳導路徑在細胞週期、細胞的活動性、以及細胞凋亡﹝apoptosis﹞門檻的控制機轉中均扮演重要的角色。許多細胞訊息傳導路徑的調控劑，不論是單獨使用或是與傳統化學治療藥劑並用，對於各種腫瘤的可能療效均已被廣泛研究，而且在肺癌、乳癌、及惡性淋巴癌均已有成功的例子。在我們過去的研究中發現，使用高劑量tamoxifen可以增進化療藥物對肝癌細胞引發細胞凋亡的現象，而此一現象可能與細胞protein kinase C活性的調控有關。由於細胞訊息傳遞路徑的調控劑一般而言副作用都比較輕微，因此它們在治療廣泛性肝細胞癌的角色也值得進一步探討。

第一部分 抗血管新生治療對於廣泛性﹝advanced﹞HCC治療之研究

HCC腫瘤細胞及其周圍的間質細胞﹝stroma cells﹞會過度表現多種促進血管增生的因子﹝pro-angiogenic factors﹞，包括vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), 以及matrix metalloproteinases等。這些因子曾被發現與腫瘤細胞的高惡性度，以及腫瘤造成的肝門靜脈栓塞及腫瘤包膜侵犯﹝capsular invasion﹞相關。此外，在肝硬化及HCC的病患血清中也可以偵測到促進血管增生因子的增加。因此，可以想像HCC病患體內的環境處於適合血管新生的狀態，而抗血管新生治療則可望控制HCC的生長。

Thalidomide 在1950年代的歐洲被當作孕婦的鎮靜劑及止吐劑使用，但很快就因為其致畸胎之嚴重毒性﹝會造成海豹肢等﹞禁止使用。D’Amato等學者最早使用corneal micropocket assay證實thalidomide對於由VEGF及bFGF導致的血管新生現象有抑制作用。thalidomide在惡性疾病的可能療效最早則是在多發性骨髓瘤﹝multiple myeloma﹞及與人類免疫不全病毒感染相關的Kaposi’s sarcoma病患中發現。也有一些初步研究顯示thalidomide對於HCC 可能也有療效。

我們首先對於無法手術根除也無法進行血管栓塞治療的廣泛性HCC病患進行了一個compassionate treatment計畫以評估thalidomide的可能療效。Thalidomide的起始劑量是每天口服200 mg，如果沒有明顯的副作用則會將劑量逐漸提高﹝每次提高100 mg﹞至最高劑量每天600 mg。在接受治療的68位病患中有63位可評估thalidomide之療效。根據WHO 療效指標，有一位病患得到完全緩解，三位部分緩解，整體緩解率為6.3 % (95% 信賴區間. 0 to 12.5%)。持續緩解的時間分別為50+, 24.6, 11.6+ 及 8.7+ 星期。這四位病患血液中胎兒蛋白(alpha-fetoprotein,a-FP)含量均明顯降低。在另外42位治療前a-FP升高的病患中，有6位 a-FP數值在接受治療之後下降50%以上。因此總共有10位HCC病患對thalidomide治療產生反應。所有68位病患的整體存活中位數為18.7星期﹝95%信賴區間11.8 至25.6星期﹞，一年存活率為27.6%。10位接受thalidomide治療有反應的病患，其整體存活中位數為62.4星期﹝95%信賴區間31.2 至93.6星期﹞。所有產生反應的病患，其使用thalidomide的劑量均在每天300 mg或以下。此一處方劑量的thalidomide 一般而言副作用均在可接受範圍內：分別有16位，6位及0位病患發生grade 2、3、4的毒性。因此我們認為低劑量﹝200-300 mg per day﹞thalidomide對於少部分廣泛性HCC 病患確實可以產生控制腫瘤的效果。

接下來我們嘗試評估利用power Doppler sonography以及血液中促進血管新生因子的變化，來測量thalidomide 對HCC的抗血管新生治療效果之可行性。Doppler sonography在臨床上使用方便，容易進行一系列的追蹤檢查，而且在過去研究指出利用Doppler sonography來評估包括大腸癌、胃癌、卵巢癌及子宮頸癌的血管新生現象，其結果對於病患的預後評估有重要參考價值。在此一研究中，對於無法手術根除也無法進行血管栓塞治療的廣泛性HCC病患，給予每天thalidomide 200-300 mg的治療。在治療開始前及治療過程中規則用power Doppler sonography追蹤計算腫瘤的血管新生指標﹝vascularity index, VI﹞，並定期抽血測量病患血液中之VEGF, bFGF,及placental growth factor (PlGF) 的變化。在144位接受power Doppler sonography檢查的HCC病患中，有56位的腫瘤血管新生指標為可評估者。無法評估的主要原因包括：(1)腫瘤太大無法完全涵蓋在超音波檢查範圍中；(2)因為浸潤性腫瘤生長﹝diffuse tumor infiltration﹞、嚴重肝硬化、或是過去接受的局部治療，而使得腫瘤的邊界無法清楚界定；(3)待測量的腫瘤在超音波無法偵測的區域﹝例如肺部或是後腹腔淋巴結﹞。在臨床特徵﹝包括年齡、性別、慢性病毒性肝炎種類、腫瘤分期、肝硬化嚴重程度、存活率﹞等方面可用power Doppler sonography評估tumor vascularity與無法評估的病患之間並無顯著差異。在47位同意進入此一試驗的病患中有44位腫瘤反應為可評估，有五位腫瘤對於thalidomide呈現有效反應：包括兩位根據WHO criteria得到腫瘤緩解﹝一位完全緩解及一位部分緩解﹞，另外三位則是持續a-FP數值下降超過50%且腫瘤穩定﹝穩定時間分別為10.4, 5.3，及 3.5 個月﹞。在thalidomide治療前測量的vascularity index，對thalidomide有反應的病患﹝中位數7.42，範圍2.99 to 13.88﹞明顯較無反應的病患﹝中位數2.15，範圍 0 to 25.35﹞為高﹝p=0.03﹞。五位對thalidomide有反應的病患中有四位在接受治療後vascularity index 明顯降低﹝45% 至100%﹞。在另一方面，血液中促進血管新生因子的變化與病患接受thalidomide是否有效無明顯相關。

第二部分 慢性C型肝炎病毒感染與血管異常增生的關係的研究

過去曾有研究指出，慢性C型肝炎與B型肝炎相比，前者比較容易造成血管異常增生，但是慢性C型肝炎與B型肝炎所引發的HCC，其中血管異常增生現象的程度及相關機轉是否不同則並不清楚。因此我們首先試圖比較慢性C型肝炎與B型肝炎所引發的HCC病患，對於thalidomide治療的效果是否有差異。我們將在台大醫院與國家衛生院進行的三個有關thalidomide治療HCC的臨床試驗資料合併分析：在112位病患中，血液檢查B型肝炎抗原﹝HBsAg﹞為陽性而C型肝炎抗體﹝anti-HCV﹞為陰性者為B-HCC﹝共61位﹞，HBsAg為陰性而C型肝炎抗體﹝anti-HCV﹞為陽性者為C-HCC﹝共33位﹞。病患對於thalidomide的療效則根據下列定義：(1)根據WHO標準達到完全或是部份緩解；(2)對開始治療前a-FP升高的病患，如果a-FP下降超過50%且持續8星期以上。B-HCC與C-HCC的各項臨床指標除年齡外均無明顯差異。﹝C-HCC, 67.5 ± 7.6 歲； B-HCC, 53.6 ± 13.7 歲, p< 0.001﹞. 對thalidomide治療有8位 B-HCC 及9位 C-HCC 產生療效，整體反應率對於B-HCC為 13.1% (95% 信賴區間4.4 to 21.8%)，對於C-HCC為 27.3% (95% 信賴區間11.2 to 43.3%) (p=0.09)。C-HCC病患與B-HCC病患相比，C-HCC病患之平均疾病惡化之時間﹝time to disease progression﹞﹝14.1 星期 vs. 8.3 星期, p=0.03﹞及平均存活時間﹝32.6 星期 vs. 21.4 星期, p=0.08﹞均似乎較佳。在多變數分析的結果中： 肝癌分期﹝CLIP 分數﹞、 病患performance status、以及病毒性肝炎種類﹝HBV或HCV﹞在預測整體存活率時均為獨立之影響變項﹝p<0.01﹞，而只有病毒性肝炎種類在預測疾病惡化之時間﹝time to disease progression﹞時為獨立之影響變項﹝p=0.027﹞。thalidomide對於C-HCC可能具有較好的療效，但這些可能的差異是否是由於抑制血管新生現象所造成則仍有待進一步研究。

我們接下來試圖探討慢性C型肝炎感染引發異常血管新生的可能機轉。C型肝炎病毒的基因轉錄後會合成一個單一的蛋白質。C型肝炎的核心蛋白﹝core protein﹞在此一蛋白質的氨基端，經由宿主細胞中的蛋白脢切下後，在宿主細胞 內會參與調控多種病毒以及宿主基因的表現，包括細胞凋亡﹝apoptosis﹞及細胞惡性轉化﹝transformation﹞。至於HCV core protein或是其他HCV病毒蛋白對於血管增生的可能調控機轉則仍不清楚。

我們使用雙槽的共同細胞培養系統﹝two-chamber co-culture system﹞來探討具有HCV core protein表現的癌細胞，對於血管內皮細胞或是巨噬細胞的影響。我們由黃麗華教授處得到一個bicistronic 轉錄病毒載體﹝retroviral vector﹞，可以將HCV core protein 在HeLa細胞內表現。血管內皮細胞是由自然生產或剖腹生產後得到的臍帶製備。將此一表現core protein的癌細胞與血管內皮細胞或是THP-1巨噬細胞株用two-chamber co-culture system共同培養，測量血管內皮細胞或巨噬細胞的細胞增生﹝proliferation assay﹞或是細胞移動﹝migration assay﹞的變化。血管內皮細胞或巨噬細胞種在此一系統的內槽﹝inner chamber﹞而HeLa細胞種在外槽﹝outer chamber﹞。兩槽之間有一polycarbonate membrane，上面有不同大小的孔﹝用於proliferation assay 的為0.4 mM，用於migration assay的為8 mM﹞，膜上舖一層0.1% gelatin作為basement membrane的代替物。
在血管內皮細胞的增生作用部分，可以看出在與帶有C型肝炎核心蛋白的HeLa細胞﹝HeLa-C190﹞共同培養時，血管內皮細胞的增生作用有顯著增加；與僅帶有vector的HeLa細胞﹝HeLa-S2﹞共同培養時則此效應不明顯。在巨噬細胞部分，同樣可以看出在與HeLa-C190細胞共同培養時，巨噬細胞的增生作用均有顯著增加；與HeLa-S2細胞共同培養時則此效應不明顯。而在血管內皮細胞的移行作用部分，雖然與HeLa-C190共同培養時，血管內皮細胞的移行作用也有增加的趨勢，但與HeLa-S2共同培養者相比時則此效應之差異並不明顯。

第三部分 訊息傳導路徑之調控對於廣泛性HCC治療之可行性研究

1. 人類上皮細胞生長因子接受體第二型﹝human epidermal growth factor receptor-2, HER-2/neu﹞在HCC之研究：許多調控細胞生長的生長因子及其接受體，在癌細胞常會有突變或是過度表現的情形。這些因子的異常對於癌細胞的生長會經由自我調控﹝autocrine﹞及鄰近細胞間調控﹝paracrine﹞的效應來刺激癌細胞生長。人類上皮細胞生長因子接受體﹝human epidermal growth factor receptor, HER﹞可以說是這些生長因子中被研究得最廣泛的一種。它在許多種癌症中都有過度表現的現象，而其相關之訊息傳導路徑也已經被廣泛的研究。例如HER-2/neu的過度表現在許多癌症就被發現與腫瘤的惡性度，期別的嚴重性以及病患的預後有關。但是之前對於HER-2/neu在HCC是否也有過度表現的現象則仍不清楚。

HER-2/neu也是一個新的癌症標靶治療方式。Trastuzumab (herceptin)是一個針對HER-2/neu的單株抗體，已經證實對於HER-2/neu過度表現的乳癌病患有明顯的療效。另一重要現象是在乳癌trastuzumab與多種化學治療藥物之間均有顯著的協同增強﹝synergistic﹞療效，即使是病患過去對於化學治療已經產生抗藥性。Trastuzumab或是其他對抗HER-2/neu訊息傳導路徑的藥物是否也可以應用在其他癌症則正在研究中。

在本研究中我們希望能進一步了解在HCC腫瘤組織中HER-2/neu過度表現的情形，也希望探討是否可能使用trastuzumab來治療廣泛性HCC。從過去曾在本院進行的廣泛性HCC之化學治療臨床試驗中，我們找到有36位曾經接受病理組織檢查﹝手術或切片﹞的病患，針對其腫瘤組織進行HER-2/neu免疫染色。使用的抗體﹝A0485, DAKO, Copenhagen﹞，呈色系統(Envision+, DAKO)以及免疫染色評分標準均為目前使用於乳癌病理檢查之標準程序。我們也對12個HCC 細胞株以西方墨點染色方式評估HER-2/neu表現情形，並用trastuzumab及不同化學治療藥物﹝包括doxorubicin, gemcitabine, cisplatin, irinotecan﹞進行藥物毒性測試﹝MTT試驗法﹞來評估單獨使用trastuzumab或合併使用trastuzumab及化療藥物對肝癌細胞的療效。

在免疫染色試驗部分，只有一位病患的組織染色呈現HER-2/neu 微陽性﹝1+﹞，這是一位57歲男性，在接受化學治療後得到兩個月的疾病穩定﹝stable disease﹞。在35位HER-2/neu陰性的病患中有8位在接受化學治療後得到部分緩解﹝partial response﹞﹝p=0.78﹞。在細胞株研究部分，12個HCC細胞株中只有一個細胞株﹝Tong﹞有明顯的HER-2/neu表現。Trastuzumab單獨使用，濃度直到10 mg/ml時對於Tong或其他細胞株都沒有明顯的生長抑制效果；在與化療藥物合併使用時也沒有明顯的加成效果。由於在過去乳癌的研究中，HER-2/neu的過度表現在免疫染色試驗中要達到3+才具有明顯的臨床意義，及可能預測腫瘤對於trastuzumab治療的反應，因此我們的研究結果顯示，HER-2/neu的過度表現情形在HCC應該非常少見，而且使用藥物來調控HER-2/neu訊息傳導路徑的方式來治療進行性HCC可能沒有明顯效果。

2. Peroxisome proliferator-activated receptor g (PPARg)接受體在HCC之研究：PPARg 是屬於細胞核內荷爾蒙接受體﹝nuclear hormone receptor﹞的一種。在接受引子﹝ligand﹞刺激後PPARg 接受體的功能是轉錄調控因子，特別是對於脂肪代謝，葡萄糖代謝平衡，以及發炎反應的調控方面。PPARg接受體或是維生素A酸x接受體﹝retinoid X receptor, RXR﹞在與個別的ligands結合後，兩種接受體會形成「雙體」﹝dimerization﹞而產生轉錄調控的功能。當PPARg 及RXR的活化劑﹝agonists﹞同時存在時，在活化下游基因的功能方面也被發現有協同增強﹝synergistic﹞的效果。

PPARg 活化劑的抗癌效果最先是在惡性脂肪肉瘤﹝liposarcoma﹞中被發現。在1到10 mM濃度下，PPARg 活化劑對於乳癌，大腸癌及攝護腺癌細胞也有抑制細胞生長的效果。PPARg 及RXR的活化劑對於乳癌或脂肪惡性肉瘤細胞的生長抑制作用可能有加成或是協同增強的效果。最近也有研究結果顯示，高劑量的PPARg的活化劑對於肝癌細胞也有生長抑制以及抑制細胞生長週期的效果。但是PPARg活化劑對於肝癌的臨床治療效用則仍不清楚。

在本研究中我們用西方墨點染色法偵測肝癌細胞核中PPARg 及RXR接受體的表現程度。Troglitazone﹝一種PPARg活化劑﹞對於肝癌細胞生長抑制的效果﹝單獨使用或是合併RXR活化劑﹞則用MTT assay來測試。我們也評估troglitazone在臨床應用可及的濃度範圍﹝2到5 mM﹞，是否可以增強化學治療藥劑對於肝癌細胞的毒殺效果。我們發現所有測試的肝癌細胞在細胞核內均有不同程度的PPARg 及RXR接受體表現。Troglitazone在濃度10 mM以下時對任何肝癌細胞均無明顯的生長抑制的效果。生長抑制的效果在troglitazone 濃度20 mM以上時可見到，但是此效果與細胞核內PPARg 及RXR接受體的表現程度並無明顯相關。9-cis retinoic acid﹝一種RXR的活化劑﹞在濃度至10 mM，本身對於肝癌細胞並無明顯生長抑制作用，也不會影響troglitazone對於肝癌細胞的抑制效果。Troglitazone濃度至10 mM時對於常見化學治療藥物，包括gemcitabine，cisplatin，paclitaxel，及irinotecan對於肝癌細胞的毒殺效果均無明顯影響。

綜合上面有關thalidomide的治療研究，我們可以得到下列初步結論：﹝1﹞低劑量﹝每天200至300毫克﹞thalidomide對於無法進行手術根除或其他局部性治療的廣泛性肝細胞癌病患，有少數﹝約5%﹞可以達到腫瘤縮小50%以上，且副作用相當輕微。對於在此劑量無效的病患，進一步增加劑量則無明顯幫助。﹝2﹞使用功率都卜勒超音波檢查對於部分﹝約40%﹞廣泛性肝細胞癌病患，可以作為評估腫瘤血管異常新生現象的參考。﹝3﹞在接受thalidomide治療前經由功率都卜勒超音波檢查計算得的「血管新生指標」較高的肝細胞癌病患，可能比較有機會產生療效﹝包括腫瘤縮小或血中胎兒蛋白數值降低﹞。在接受thalidomide治療後，大部分產生療效的病患，其「血管新生指標」也會隨之降低。﹝4﹞使用血中促進血管新生因子﹝例如VEGF, bFGF, PlGF﹞的變化，無法有效預估病患在接受thalidomide治療時的可能療效。

隨著有關腫瘤血管新生現象的基礎以及臨床研究不斷的進展，許多新的理論以及生物模式都被提出來解釋不同腫瘤中所見的不同血管新生現象。以肝細胞癌來說，在像肝臟這樣血液循環豐富的器官，腫瘤生長與血管異常新生的關係，比較接近所謂的「vessel cooption」。肝臟本身的網狀組織結構在腫瘤區域仍然保留。腫瘤細胞像是「取代」了正常的肝細胞，但兩者之間的接觸非常密切，也沒有明顯的發炎細胞浸潤或是血管新生現象。這種現象在乳癌及部分大腸癌的肝轉移病理組織中均曾被描述。以原發的肝細胞癌來說，許多肝細胞癌其腫瘤細胞與周圍的血管竇內皮細胞﹝sinusoid endothelial cells﹞之間的關係，確實符合vessel cooption的特徵。其次，在vessel cooption現象中扮演重要角色的angiopoietin-Tie-2訊息傳導機轉，也被發現與肝細胞癌的細胞惡性程度與腫瘤血管新生有關。從發生學的角度來看，正常肝臟在發生的初期就需要來自血管內皮細胞所提供的生長訊息才能正常進行器官的發生，這種肝細胞與血管內皮細胞的交互作用是早在內皮細胞形成血管結構之前就開始的。最後，在藥物引起的肝傷害實驗模式中發現，肝臟在受到傷害時會釋出VEGF，進而﹝1﹞透過VEGFR-2 (Flk/KDR)刺激血管竇內皮細胞本身的增生；﹝2﹞透過VEGFR-1 (Flt-1)刺激內皮細胞分泌肝細胞生長因子﹝hepatocyte growth factor﹞及interleukin-6，再進而刺激肝細胞的生長。這個複雜的回饋機制直接發生在肝細胞與內皮細胞之間，阻斷這一交互作用就會使肝臟在受傷害後的再生受到顯著影響。我們假設在腫瘤細胞於其周圍的血管竇內皮細胞之間也有類似的交互作用存在，因此我們接下來的研究，就是希望找出這個交互作用的相關機轉，並進一步設計具有專一性的藥物治療方式。在腫瘤發生的最早期抑制其進展。

在我們的研究中所看到的，C型肝炎引起之肝細胞癌患者在接受thalidomide治療之後有較長的「至疾病惡化時間」以及整體存活時間，以及可能對thalidomide產生療效的機會較高等現象，是否是因為thalidomide的抗血管新生作用所引起的，由於還沒有很好的替代性生物指標來監測抗血管新生治療的生物活性，目前仍然不清楚。除了抗血管新生的作用之外，thalidomide 也被發現可以透過其他機轉產生抑制腫瘤生長的效果，例如免疫調控、抑制腫瘤壞死因子﹝tumor necrosis factor a, TNF-a﹞的表現及活性、以及使癌細胞的細胞凋亡機轉較容易被啟動等。進一步的研究有待更具專一性的抗血管新生藥物以及更準確的生物療效指標來驗證。在慢性C型肝炎病毒與血管新生現象的體外研究部分，雖然目前所得到的結果還非常初步，但我們的試驗結果顯示，會表現C型肝炎病毒核心蛋白的腫瘤細胞確實可能會影響血管內皮細胞以及巨噬細胞的一些生長特性。我們接下來將努力澄清核心蛋白在宿主細胞內是啟動了哪些機轉﹝例如NFKB訊息傳遞路徑或其他﹞，又是分泌了哪些因子﹝VEGF、bFGF等﹞對血管內皮細胞及巨噬細胞產生paracrine效用，以探討使用更具有專一性的藥物來抑制血管新生的治療的可行性。

在探討HER-2/neu以及PPARg 訊息傳導路徑之調控作為進行性肝細胞癌治療方式的研究部分，都沒有我們希望得到的結果。但是這種針對特殊標的的癌症治療模式﹝targeted cancer therapy﹞，卻是未來藥物治療發展的主流。從近年來在轉移性乳癌及肺癌的基礎及臨床研究中的經驗可知，這種治療模式的研發，成功的條件如下，這也是我們未來必須持續努力的方向：﹝1﹞臨床試驗設計時應該預先設計配合的實驗針對可能的替代性生物性指標進行探討。因此病患在進入臨床試驗之前還是應該盡量保留足夠的腫瘤組織。﹝2﹞發展更多的實驗模式來探討藥物作用及抗藥性機轉的機轉，並利用新的基因體及蛋白質體研究方式來做更深入的探討。﹝3﹞在進行大規模的第三期隨機臨床試驗前應該先考慮一些較小規模的試驗來驗證實驗室所觀察到的可能療效。

從過去我們在肝細胞癌的臨床以及基礎研究的經驗，我們希望，未來在進一步開展肝細胞癌藥物治療的研究領域時，能夠從下列四個方向來努力：
﹝1﹞ 針對最近研究所指出，肝細胞癌細胞循環調控機轉以及mitogen-activated protein kinase (MAPK)傳導路徑的異常，測試這兩個訊息傳導路徑作為特殊標的藥物治療﹝targeted cancer therapy﹞的對象之可行性。並且運用在肝細胞癌基因體醫學及蛋白質體醫學的研究成果，持續篩選有潛力作為特殊標的藥物治療的對象。
﹝2﹞ 持續以高水準的臨床試驗，發展針對肝細胞癌的抗血管新生治療以及其他針對特殊標的的癌症藥物治療，並且在這些治療性臨床試驗中探索合適的分子標的，作為開發新的治療性以及預防性藥物的參考﹝nested prevention trial﹞。
﹝3﹞ 利用肝細胞癌的生體內﹝in vivo﹞實驗模式，以找尋值得進一步發展的藥物治療標的﹝包括腫瘤細胞本身以及其周圍血管內皮細胞及間質細胞的特殊分子變化﹞以及其相關作用機轉。
﹝4﹞ 尋找合適的分子標的，作為發展肝細胞癌化學預防﹝chemoprevention﹞的目標。

關鍵字

藥物治療 ； 肝細胞癌 ； 血管新生 ； 特殊分子標的治療

並列摘要

Hepatocellular carcinoma (HCC) is one of the most common cancers in Taiwan. Surgery is curative in only a minority (about 20 to 30%) of patients. Most patients die from locally advanced or metastatic diseases within a relatively short period of time. Systemic chemotherapy is difficult for HCC patients to tolerate because hepatic function is often impaired due to underlying cirrhosis, which is often accompanied by hypersplenism and peripheral cytopenia. New modalities of treatment with a more favorable therapeutic index are needed for patients with advanced HCC.

In this thesis we explore two approaches to the treatment of advanced HCC. The first is anti-angiogenesis therapy. Cancer-induced angiogenesis plays a pivotal role in the development and progression of cancer. Compared with conventional cytotoxic therapy, which targets cancer cells, anti-angiogenesis therapy aiming at the tumor-induced neovasculature has a pharmacokinetic advantage because the tumor vasculature has better drug accessibility than the tumors per se. Besides, the endothelial cells are genetically stable diploid cells, and acquired drug resistance may be rare. HCC is typically a hypervascular tumor. Anti-angiogenesis therapy may thus be a reasonable choice for patients with advanced HCC.

The second is modulation of cellular signal transduction pathways. Cellular signal transduction pathways are involved in the control of cell cycle, cell motility, and apoptosis threshold. Modulators of various cellular signal transduction pathways have been intensively studied for their anti-cancer activity, either alone or in combination with cytotoxic agents, and success has been seen in the treatment of lung and breast cancers and in non-Hodgkin’s lymphoma. In earlier studies we have shown that high-dose tamoxifen may enhance chemotherapy-induced apoptosis of HCC cells by modulating cellular protein kinase C activation. The use of signal transduction modulators in the treatment of advanced HCC is worthy of further evaluation because of the generally favorable toxicity profile of these agents.

Part I. Study of Anti-angiogenesis Therapy for the Treatment of Advanced HCC
HCC tumor cells and their surrounding stroma cells may over-express pro-angiogenic factors such as vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), and matrix metalloproteinases. Some of these pro-angiogenic factors are closely associated with high histological grade, portal venous thrombosis, and tumor capsular invasion. Elevated serum levels of angiogenic factors have also been found in patients with liver cirrhosis or HCC. It appears that a pro-angiogenic milieu exists to facilitate tumor proliferation and invasion in HCC and anti-angiogenesis therapy may be able to control this tumor.

Thalidomide was used as a sedative and anti-emetic agent for pregnant women in the 1950’s in Europe, but was soon withdrawn from this indication after the discovery of serious teratotoxicity (amelia and phocomelia) associated with its use. Using a corneal micropocket assay, D’Amato et al demonstrated that angiogenesis induced by VEGF and bFGF can be inhibited by thalidomide. Its potential therapeutic efficacy for neoplastic disorders has been reported in patients with refractory multiple myeloma and human immunodeficiency virus-associated Kaposi’s sarcoma. Preliminary data also suggest beneficial activity of thalidomide in advanced HCC.

We first started a compassionate treatment program for patients with unresectable and not embolizable HCC to analyze the efficacy of thalidomide in the treatment of advanced HCC. Thalidomide was given at a starting dose of 200 mg per day, and the dose was gradually escalated in 100-mg steps up to 600 mg per day if no limiting toxicities developed. Sixty-three of the 68 patients were evaluable for response. One complete and 3 partial responses, defined by WHO criteria, were seen, with a response rate of 6.3 % (95% C.I. 0 to 12.5%). The duration of response was 50+, 24.6, 11.6+ and 8.7+ weeks, respectively. All 4 responders had dramatic decrease of alpha-fetoprotein (a-FP) levels. Another 6 of the 42 patients with elevated pre-treatment a-FP had more than 50% decrease of their a-FP levels. Totally 10 patients had an objective response to thalidomide. The median overall survival for all of the 68 patients was 18.7 weeks (95% C.I. 11.8 to 25.6 weeks) with a 1-year survival rate of 27.6%. The median overall survival of the 10 patients with an objective response to thalidomide was 62.4 weeks (95% C.I. 31.2 to 93.6 weeks). All responders responded at a dose equal to or less than 300 mg per day. Toxicities of thalidomide were generally manageable, and only 16, 6, and 0 patients developed grade 2, 3, and 4 toxicities, respectively.

This study has demonstrated that low-dose thalidomide induces an unequivocal beneficial tumor response in a minority of patients with advanced HCC. Patients with an objective response to thalidomide treatment had improvement in cancer-related symptoms without developing significant treatment-related toxicities. However, if the tumors are refractory to low-dose (200 to 300 mg per day) thalidomide, escalation to higher dose is not beneficial. This relatively dose-independent effect of thalidomide appears to be a common feature of non-cytotoxic biologic agents. In contrast to the relatively dose-independent anti-cancer effect of thalidomide, the treatment-related toxicities of thalidomide, such as somnolence, fatigue, and constipation, are largely dose-dependent. It is intriguing to note that patients with chronic hepatitis C-related HCC appear to be more likely to benefit from thalidomide therapy in this study, although the possibility of false-positive finding resulting from multiple comparisons can not be excluded. Previous studies have compared the clinicopathological features of hepatitis B- and C- related HCC, but the results are inconclusive [35-39]. However, more and more evidence has suggested that HCC associated with hepatitis B and C infection may involve different genetic mechanisms in the regulation of angiogenesis and tumor microenvironment. It remains to be clarified whether these differential mechanisms of angiogenesis and microenvironment regulation may partly account for the differential response to thalidomide between hepatitis B- and C-related HCC.

We then tried to evaluate the feasibility of using power Doppler sonography (PDS) and circulating angiogenic factors (CAF) to monitor the anti-angiogenic effect of thalidomide in HCC. Doppler sonography has the advantage of non-invasiveness and ease to use in serial follow-up. Angiogenesis detected by Doppler sonography has been shown to provide important prognostic information for patients with colon, gastric, ovarian, and cervical cancers. In this study patients with advanced HCC were treated with thalidomide 200-300 mg/day. Tumor vascularity index (VI), evaluated by PDS, and CAF, including VEGF, bFGF, and placental growth factor (PlGF), were measured at baseline and periodically during thalidomide treatment. Fifty-six of the 144 patients with advanced HCC who were screened by PDS were considered evaluable for tumor vascularity. The main reasons for non-evaluability included (1) huge tumor size that could not be encompassed within the field of the sonographic examination; (2) indiscrete tumor margin resulting from diffuse tumor infiltration, severe cirrhosis, or prior local therapy; (3) location of the index tumor that could not be approached by PDS (e.g., lung, retroperitoneal lymph nodes). There was no significant difference between patients whose tumor vascularity was evaulable or not evaluable in terms of age, sex, types of chronic viral infection, disease stage, severity of liver cirrhosis, and survival. Of the 47 patients with bi-dimensionally evaluable disease enrolled in this study, 44 were evaluable for tumor response. Among these, 5 were classified as having objective response, including 1 complete and 1 partial response by WHO criteria, and 3 other patients with decreased a-FP level for more than 50%, together with stable disease for 10.4, 5.3, and 3.5 months, respectively. The pre-treatment median VI was significantly higher in responders (7.42, range: 2.99 to 13.88) than non-responders (2.15, range: 0 to 25.35) (p=0.03). Four of the 5 responders had a significant decrease of VI (45% to 100%) after starting thalidomide treatment. However, none of the CAF was significantly different between responders and non-responders.

This study demonstrated that evaluation of tumor vascularity by power Doppler sonography is feasible in a subset of patients with advanced HCC, and that higher pre-treatment VI level in these patients may predict a better chance of response to thalidomide. Tumor vascularity detected by conventional color Doppler sonography has shown good correlation with vessel density determined by immunohistochemical staining. Preclinical studies have shown the potential of Doppler sonography as a surrogate endpoint in anti-angiogenesis therapy. Power Doppler sonography has the advantages of low noise variance, relative angular independence, and increased dynamic signal range, which make it more sensitive and specific than the conventional Doppler sonography. However, the use of power Doppler sonography for HCC has several limitations. The first is the reproducibility of the techniques. Inter-observer variation, difference in instrument, angle of evaluation, and setting of color gain may make comparison among patients difficult.. The second limitation is the patient population. The application of Doppler sonography was generally successful for patients with small HCC. However, in this study only about 40% of patients with advanced HCC who were otherwise eligible for thalidomide therapy were considered evaluable by power Doppler sonography. Although in this study no significant differences were found in the clinical characteristics between evaluable and non-evaluable patients, screening by power Doppler sonography may impose a selection bias on patients with anti-angiogenic therapy.
The correlation of pre-treatment VI with the angiogenic factors bFGF and PlGF suggests that VI may reflect the activity of tumor angiogenesis. However, a precise correlation between levels of VI or angiogeneic factors and response to thalidomide could not be concluded in this study. Although the median VI before thalidomide treatment was significantly higher in patients who responded to thalidomide treatment, a clear-cut threshold VI value that may predict the response to treatment in the majority of patients was not found. This may partly be due to the relatively small sample size of the study and the low response rate of HCC to thalidomide. Further, decrease in VI after thalidomide treatment was found in un-responsive patients, and one of the 5 patients with objective tumor response even had a paradoxical increase of VI during thalidomide treatment. It has been argued that microvessel density may not represent the true angiogenic activity of tumors. Change in microvessel density reflects the change of ratio of the vascular component of the tumor to its tumor cell component rather than vascular inhibition per se. Besides, the possibility that thalidomide may exert its anti-tumor effect through mechanisms other than angiogenesis inhibition should be considered as a potential explanation for the anomalous VI increase during thalidomide treatment in one of our patients.
Measurement of circulating angiogenic factors was not useful in either prediction or follow-up of patients undergoing thalidomide treatment.
The relationship between tumor response to thalidomide and changes in circulating angiogenic factors has been extensively investigated in patients with multiple myeloma. VEGF, bFGF, TNFa and interleukin-6 have all been found in some but not other series to be associated with tumor response to thalidomide, while the key regulators of tumor angiogenesis remain undetermined. This study did not find any consistent patterns of change of angiogenic factors correlated with tumor response. The clinical implications of elevation of PlGF level after thalidomide therapy and its correlation with VEGF change, regardless of tumor response, are unknown.

Part II. Study of the relationship between chronic hepatitis C infection and angiogenesis in HCC

Previous studies have suggested that chronic HCV infection may be more angiogenic than chronic HBV infection. It is not yet know whether HCV- and HBV-related HCC is different in terms of the extent of angiogenesis and the underlying mechanisms of angiogenesis induction. We first sought to clarify if HCV- and HBV-related HCC may respond differently to thalidomide. We collected the clinical data from 3 clinical trials conducted for patients with advanced HCC in National Taiwan University Hospital and the National Health Research Institute in which thalidomide was given by a similar protocol. One hundred and twelve patients were enrolled. Patients with positive serum HBV surface antigen (HBsAg) and negative anti-HCV antibody were categorized as HBV-related HCC (B-HCC, n=61), while patients with negative HBsAg and positive anti-HCV were categorized as HCV-related HCC (C-HCC, n=33). Response to thalidomide was defined by either a complete or partial response according to WHO criteria or a more than 50% decrease of serum a-FP level for more than 8 weeks for patients with elevated a-FP level at the start of thalidomide therapy. All major clinicopathological features were comparable between the two groups except for age (C-HCC, 67.5 ± 7.6 years; B-HCC, 53.6 ± 13.7 years, p< 0.001). Eight patients with B-HCC and 9 patients with C-HCC responded to thalidomide treatment. The total response rate to thalidomide was 13.1% (95% C.I. 4.4 to 21.8%) for B-HCC and 27.3% (95% C.I. 11.2 to 43.3%) for C-HCC (p=0.09). Patients with C-HCC appeared to have better median time to disease progression (14.1 weeks vs. 8.3 weeks, p=0.03) and overall survival (32.6 weeks vs. 21.4 weeks, p=0.08) than patients with B-HCC. In multivariate analysis, staging (CLIP score), performance status and types of viral infection were independent predictors for overall survival (p<0.01), while only types of viral infection was independent predictor for time to disease progression (p=0.027). There is a trend favoring better anti-tumor activity of thalidomide in patients with C-HCC. Whether this result is related to angiogenesis inhibition remains to be determined.

We then tried to explore the possible mechanisms of angiogenesis induced by chronic hepatitis C infection. HCV genome encodes a single polyprotein. The core protein lies at the N-terminal end and is cleaved by host cell proteases. Core protein has been shown to regulate the expression of multiple viral and cellular genes. It has also been implicated in control of apoptosis and cellular transformation. The direct effect of core protein or other HCV proteins on angiogenesis regulation remains elusive.

A two-chamber co-culture system was used to examine the effects of core protein-expressing cancer cells on endothelial cells and macrophage. Expression of HCV core protein in HeLa cancer cells was done by using a bicistronic retroviral vector. (Dr Lih-Hwa Hwang, Graduate Institute of Microbiology, National Taiwan University College of Medicine). HUVEC (human umbilical vein endothelial cells) were prepared from umbilical cords obtained from nature or caesarean birth. The macrophage cell line THP-1 was kindly provided by Prof Pan-Chyr Yang. The function of endothelial cells and macrophage were evaluated by proliferation assay and cell migration assay. Briefly, HUVEC or macrophage were plated in the inner chamber and HeLa cells transfected with HCV core gene or mock vectors were plated in the outer chamber. A polycarbonate membrane with different pore size and coated with 0.1% gelatin, which is used as a substitute for basement membrane, were placed between the upper and lower well plates. In proliferation assay the diameter of the pores in the bottom of the inner chamber was 0.4 mm, while in migration assay the pore diameter was 8 mm. We found that co-culture with the core protein-expressing HeLa cells (HeLa-C190 cells) caused significant increase of proliferation of both HUVEC and macrophage, whereas co-culture with the vector-containing HeLa cells caused no evident change. Increase of migration ability of HUVEC after co-culture with HeLa-C190 cells was also noted, although the difference was not statistically significant from co-culture with HeLa-S2 cells.

Part III. Study of Modulation of Signal Transduction Pathways for the Treatment of Advanced HCC
Over-expression of HER (human epidermal growth factor -related receptor) has been found in a variety of human cancers, and the clinical and biological significance of HER-related signal transduction pathways have been extensively studied. For example, in many cancers over-expression of HER-2/neu is correlated with aggressive tumor behavior, advanced stages of disease, and a worse outcome. However, the role of HER-2/neu over-expression in HCC remains inconclusive.

HER-2 is a novel target for cancer therapy. A humanized form of anti-HER-2 antibody, trastuzumab (herceptin), exerts significant activity in the treatment of breast cancer expressing high levels of HER-2/neu. Importantly, trastuzumab has a significantly synergistic effect with multiple cytotoxic agents in breast cancer, even in patients who are refractory to previous chemotherapy. Whether the therapeutic efficacy of trastuzumab or other drugs targeting HER-2/neu signal transduction pathways can be extended to other human malignancies is currently under intense investigation.

In the present study we sought to examine the significance of HER-2/neu expression in HCC and explored the possibility of using trastuzumab in the treatment of HCC. Tissue specimens from 36 HCC patients who had been enrolled in 3 separate prospective clinical trials of systemic chemotherapy were studied by immunohistochemical staining. A polyclonal antibody (A0485, DAKO, Copenhagen) against HER-2/neu and a horseradish peroxidase-based visualization system (Envision+, DAKO) was used. Scoring criteria was in accordance with the manufacturer’s guidelines. Twelve HCC cell lines were examined for HER-2/neu over-expression by Western blotting. Single-agent growth regulatory activity of the anti-HER-2/neu antibody, trastuzumab (Genentech, South San Francisco), and its combinative cytotoxicity with chemotherapeutic agents (doxorubicin, gemcitabine, cisplatin, irinotecan) were determined by a tetrazolium-based colorimetric assay (MTT test).

In the immunohistochemical study, all but one of the HCC tumor tissues were negative for HER-2/neu expression. The only patient with positive (1+) HER-2/neu expression was a 57-year-old man who achieved stabilization of disease for 2 months after chemotherapy. Eight of the 35 patients with negative HER-2/neu expression had had partial remission after chemotherapy (p=0.78). In the cell line study, only one (Tong cells) out of the 12 HCC cell lines had a significant level of HER-2/neu expression. However, trastuzumab up to 10 mg/ml had no discernible growth inhibitory or chemosensitizing effect on Tong cells or any other cell lines.

Previous studies of HER-2/neu expression in HCC have yielded discrepant results. Most of these studies did not use the standardized immunohistochemistry or fluorescent in situ hybridization methods. Our study employed a new visualization technique and standardized scoring criteria that are used in the FDA-approved HercepTest to avoid these problems. Although the disease in patients in our study may not represent HCC in general, the very low percentage of HER-2/neu expression in our patients indicates that HER-2/neu overexpression is not an important prognostic factor for patients with advanced or metastatic HCC.

Previous in vitro study demonstrated a correlation between HER-2/neu expression and susceptibility to antibody-induced growth inhibition in breast cancer cells. In contrast, HER-2/neu overexpressing gastric and colon cancer cell lines were resistant to anti-HER-2 antibodies, i.e., the growth inhibitory and chemosensitizing effects of anti-HER-2 regulation may be tumor-specific. Variation of the response of downstream signaling pathways to anti-HER-2 regulation in different cancer cells may explain the differential effects of anti-HER-2/neu antibodies. Our data suggest that HER-2/neu expression in HCC cells may not affect cell growth and chemosensitivity because of the moderate level of expression and possible disparity in downstream signal transduction.

The peroxisome proliferator-activated receptor g (PPARg), a member of the nuclear hormone receptor superfamily, functions as a ligand-dependent transcription factor and plays an important role in several signaling pathways, including lipid metabolism, glucose homeostasis, and inflammation. PPARg dimerizes with retinoid X receptor (RXR) after binding with either PPARg or RXR ligand. Synergistic activation of downstream genes may occur when both ligands are present.

The anti-cancer activity of PPARg agonists was first demonstrated in a liposarcoma model. PPARg agonists, at concentrations of 1 to 10 mM, may also induce growth inhibition in a variety of cancers, including cancers of breast, colon, and prostate. Synergistic or additive effects of growth inhibition between PPARg and RXRa agonists have been found in liposarcoma and breast cancer cells. Recent studies have demonstrated that PPARg agonists may induce cell cycle arrest and growth inhibition in HCC cells at high concentration. The utility of PPARg agonists for the treatment of HCC remained undetermined.

In this study the nuclear expression of PPARg

並列關鍵字

targeted therapy ； angiogenesis ； systemic therapy ； thalidomide ； hepatocellular carcinoma

參考文獻

American Joint Committee on Cancer: Manual for staging of cancer. 4th ed. Philadelphia: J.B.Lippincott, 1992

Arap W, Pasqualini R, Ruoslahti E: Chemotherapy targeted to tumor vasculature. Curr Opin Oncol 1998; 10: 560-565

Augustin HG, Kozian DH, Johnson RC: Differentiation of endothelial cells” analysis of the constitutive and activated endothelial cell phenotypes. BioEssay 1994; 16: 901-906

Baccala AA, Zhonh H, Clift SM, et al: Serum vascular endothelial growth factor is a candidatebiomarker of metastatic tumor response to ex vivo gene therapy of renal cell cancer. Urology 1998; 51: 327-332

Bahary N, Zon LI: Endothelium-chicken soup for the endoderm. Science 2001; 294: 530-531