讓精準醫療更精準：非小細胞肺癌標靶治療及其抗藥性的研究

Abstract

2009年的IPASS study確認了表皮生長因子接受體酪胺酸酶抑制劑(epidermal growth factor receptor tyrosine kinase inhibitor, EGFR TKI) gefitinib比傳統化療更能控制具有EGFR基因突變的肺癌，開啟了肺癌精準醫療的時代。在台灣，約有一半的肺癌病患具有EGFR突變，遠較西方國家的15-25%為高。目前EGFR的標靶藥物已有一代gefitinib和erlotinib、二代afatinib和dacomitinib及三代的osimertinib。EGFR T790M突變是最常見的第一代和第二代EGFR TKI之抗藥機轉，由於第三代的osimertinib對EGFR T790M特別有效，因此找到具有EGFR T790M的患者就十分重要。我們發現使用第一代或第二代EGFR TKI的時間較長的患者，發生EGFR T790M的機會也較高，而uncommon EGFR mutation的患者發生EGFR T790M突變的機率較低。然而，並不是所有EGFR T790M的患者，使用osimertinib的效果都一樣的好。我們發現osimertinib對complex EGFR mutations with T790M之非小細胞肺癌治療效果不佳：腫瘤反應率只有27%，中位數無惡化存活期(progression-free survival, PFS)只有2.9個月。Osimertinib是否也為這群病患的治療首選，值得存疑。在次世代定序(next generation sequencing, NGS)的年代，腫瘤產生抗藥性後，再次切片用腫瘤組織檢測NGS，或是直接抽血檢測NGS即可，是一個重要的臨床問題。我們前瞻性的在台大醫院研究接受過EGFR TKI抗藥的肺癌病患，同時以腫瘤組織檢體和血液檢體進行NGS檢測。我們發現已接受組織再切片的病患中，仍有30%無法產出組織NGS的報告。在同時有組織及血液NGS報告的患者中，有50%的第一代或第二代EGFR TKI抗藥性的患者產生了EGFR T790M突變：其中48%的T790M只在組織中被檢測到而17%的T790M只在血液中被檢測到。無論從組織中檢測到EGFR T790M突變，或是從血液中檢測到EGFR T790M突變，對後續的osimertinib治療的效果都很好。因此，如果能同時進行組織和血液的NGS檢驗，會發現更多對下一線osimertinib治療有效的患者。利用NGS做檢驗除了EGFR突變以外還可以檢測其他的驅動基因突變，如MET基因擴增(MET amplification)。EGFR TKI抗藥後的MET amplification已在多個臨床試驗中顯示可以被EGFR TKI及MET TKI的合併治療控制。我們的研究中發現10%的病患在EGFR TKI抗藥後產生MET amplification。在EGFR TKI抗藥後使用NGS檢測相較於傳統的EGFR單基因檢測，能夠找到更多能繼續接受標靶治療的患者。

然而，這些病患接受了第二線的MET TKI合併治療後，往往最後仍再次產生抗藥性。對合併兩種TKIs後產生抗藥性的機轉，目前仍不清楚。我們收集台大醫院及台大癌醫使用EGFR TKI後，產生MET amplification，再合併EGFR TKI及MET TKI治療後，再產生抗藥性的病患。我們發現在7位雙重抗藥後的病患中，有3位出現BRAF fusion；有1位出現EGFR T790M；另有1位病患出現了ALK fusion。我們從接受osimertinib後產生抗藥性的病患胸水中培養出具有MET amplification之EGFR TKI抗藥肺癌細胞株(PE5345)，以及病患在同時使用osimertinib及MET TKI – capmatinib治療後，再產生抗藥性的肺癌細胞株(PE5867)。另外，在實驗室中使用一系列逐漸增加藥物濃度治療PE5345後，我們培養出另一株對osimertinib + capmatinib合併治療後有抗藥性的肺癌細胞株(PE5345 os/cp R)。我們發現，PE5867及PE5345 os/cp R這兩株對osimertinib + capmatinib合併治療有抗藥性的細胞株，它們的MET amplification都減少了。然而兩株細胞的抗藥機轉不同。PE5345是經由BRAF fusions，造成下游ERK活化而產生抗藥性。此抗藥性可以被osimertinib合併MEK抑制劑trametinib逆轉；而PE5345 os/cp R細胞的抗藥性可能是EGFR及ERBB2 amplification及overexpression造成抗藥性，而使用afatinib可以抑制這株細胞的生長。腫瘤的異質性在後線腫瘤的治療，扮演重要的角色。使用EGFR TKI合併MET TKI後再產生抗藥性後，而有些患者的抗藥機轉仍可以被標靶治療克服。EGFR及MET雙特異性抗體(bispecific antibody) – amivantamab，對具有MET amplification的PE5345，以及對osimertinib合併capmatinib有抗藥性的PE5867及PE5345 os/cp R細胞，都有很強的抗體依賴性細胞媒介細胞毒性作用(antibody-dependent cell-mediated cytotoxicity, ADCC)。Amivantamab對具有MET amplification的EGFR TKI抗藥細胞，及使用過EGFR及MET抑制劑後仍再次產生抗藥性的雙重抗藥性肺癌，都顯示出治療的潛力。

除了EGFR以外，異生性淋巴癌激酶(anaplastic lymphoma kinase, ALK)的融合基因(ALK fusion)是肺癌另一個重要的驅動基因。目前ALK已有很好的標靶藥物—異生性淋巴癌激酶酪胺酸酶抑制劑(ALK TKI)：一代crizotinib、二代alectinib、ceritinib和brigatinib及三代的lorlatinib可以控制腫瘤，然而哪些因素會影響ALK TKI的效果，仍不太清楚。有文獻報導指出較長的ALK fusion，如第一型和第二型的EML4-ALK fusion，可能對crizotinib的治療效果較佳，而較短的EML4-ALK fusion，如第三型，對crizotinib的治療效果較差。然而，我們分析台大醫院使用crizotinib治療的病患，並沒有看到crizotinib的治療效果和不同型的ALK fusion的相關性。我們和台北榮總及高醫合作，進行了前瞻性的世代研究：將ALK fusion的肺癌病患治療前的腫瘤組織，進行NGS檢驗。我們發現ALK fusion的種類，也和alectinib的治療效果無關。而治療前腫瘤裡發現MYC amplification的話，alectinib治療的效果會較差。然而儘管ALK TKI治療效果很好，但大部分的患者終究會再產生抗藥性。我們和全國7間醫學中心合作，分析使用ALK TKI後產生抗藥性的機轉。我們發現抗藥性的ALK mutations出現在大約1/4的病患。而對lorlatinib有抗藥性的compound ALK mutations在crizotinib, ceritinib及alectinib 治療失敗後都有可能出現。在ALK TKI治療產生抗藥性時，分析抗藥性的機轉可能有助於後線藥物的治療選擇。

在台灣，大多數的肺癌具有驅動基因突變。如何精準的使用標靶藥物治療，是非常重要的課題。我這幾年的研究證實了真實世界中EGFR及ALK標靶藥物的治療效果；藉由了解EGFR及ALK標靶藥物抗藥性的機轉，可以發現不同抗藥性機轉的腫瘤需要給予不同的藥物治療；在合併使用EGFR抑制劑及MET抑制劑後仍再產生的抗藥性的腫瘤，仍然有機會可以再次使用標靶治療。藉由分析病患腫瘤的驅動基因，抗藥性的原因，抗藥後再次產生抗藥性的機轉，我們能讓精準醫療更精準，以改善肺癌病患的預後。

In 2009, the IPASS study demonstrated the superiority of the epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) gefitinib over chemotherapy in treating patients with EGFR mutant [EGFR(+)] non-small cell lung cancer (NSCLC), marking the advent of precision medicine in this field. In Taiwan, approximately half of NSCLC patients exhibit EGFR mutations, a prevalence notably higher than the 15-25% observed in Caucasian populations. Currently, several generations of EGFR TKIs have been approved for advanced EGFR(+) NSCLC, including first-generation (gefitinib and erlotinib), second-generation (afatinib and dacomitinib), and third-generation (osimertinib) inhibitors. The EGFR T790M mutation is the major resistance mechanism for the first and second-generation TKIs, with osimertinib effectively targeting this mutation. Therefore, identifying patients with the EGFR T790M mutation is critical.

Our analysis of patients at National Taiwan University Hospital (NTUH) revealed that those who received first or second-generation EGFR TKIs longer are more likely to develop an EGFR T790M mutation upon acquiring resistance, whereas patients with uncommon EGFR mutations are less likely to acquire the T790M mutation. However, not all patients with EGFR T790M respond to osimertinib well. We found the response rate (RR) and progression-free survival (PFS) of osimertinib in patients with complex EGFR mutations with T790M was only 27% and 2.9 months. Whether osimertinib is the treatment of choice becomes questionable. In the era of next-generation sequencing (NGS), we can sequence dosens to hundards cancer-related gene at a time. After EGFR TKI resistance, NGS can be conducted on tumor tissue (tissue rebiopsy) or plasma (liquid rebiopsy), though the superiority of one method over the other remains debated. In our prospective study of patients who developed resistance to EGFR TKIs, we performed both tissue and liquid rebiopsies concurrently. NGS analysis was not possible in 30% patients received tissue rebiopsy. In patients with tissue and liquid rebiopsy NGS results, EGFR T790M was detected in 50% patients resistant to first or second-generation EGFR TKI. Among them, 48% was detected only by tissue rebiopsy and 17% was detected solely by liquid rebiopsy. Additionally, MET amplification was found in 10% of total patients. To do tissue and liquid NGS in parallel after EGFR-TKI resistance may find more patients with targetable cancers.

MET amplification following resistance to EGFR TKI therapy can be counteracted by incorporating a MET inhibitor; however, resistance can still develop despite such combination treatments. In our review of patients who developed resistance after combining EGFR inhibitors with MET inhibitors, we identified acquired BRAF fusions, the EGFR T790M mutation, and EML-ALK fusions as mechanisms of resistance. We established a patient-derived EGFR L858R and MET amplification cell line (PE5345) and its paired cell line (PE5867) after clinical resistance to osimertinib combined with capmatinib. In addition, PE5345 was exposed to progressively increasing concentrations of osimertinib and capmatinib over seven months, resulting in a drug-tolerant cell line (PE5345 os/cp R) in our laboratory. Both PE5867 and PE5345 os/cp R lost MET amplification. PE5867 acquired GTF2I-BRAF fusions, and resistance was reversed with a combination of osimertinib and trametinib. The drug-tolerant PE5345 os/cp R exhibited EGFR and ERBB2 amplifications, and was inhibited by afatinib in vitro. Moreover, amivantamab significantly induced antibody-dependent cell-mediated cytotoxicity (ADCC) against PE5345, PE5867, and PE5345 os/cp R. Resistance following dual inhibition is heterogeneous, and targeting the newly emerged resistant mechanisms may overcome resistance associated with dual inhibition.

ALK fusion is another targetable driver in NSCLC, accounting for 5-10% of NSCLC patients. First-generation ALK TKI crizotinib, second-generation ALK TKI alectinib, ceritinib and brigatinib and third-generation ALK TKI lorlatinib are approved to treat advanced NSCLC with ALK fusion [ALK(+)].However, factors associated with ALK TKI efficacy are not well-known. While some studies have suggested an association between ALK fusion variants and TKI efficacy, our analysis of data from patients at NTUH did not reveal a correlation between ALK fusion variants and progression-free survival (PFS) to crizotinib. We initiated a prospective multicenter study to evaluate the association between cancer genetic alterations including ALK fusion variants and co-occuring mutations, and alectinib PFS. Similar to our findings with crizotinib, ALK fusion variants were not associated with alectinib PFS; however, the presence of MYC amplification was correlated with shorter PFS on alectinib. In spite of initial good ALK TKI reponse, however, the majority of patients ultimately developed resistance following ALK TKI treatment. We led a multicenter prospective study utilizing NGS to analyze resistance mechanisms to ALK TKIs. We found that ALK kinase domain mutations could be detected in approximately one-fourth of the patients. Compound ALK mutations, which may contribute to lorlatinib resistance, could occur in lung cancers resistant to crizotinib, ceritinib, and alectinib.

In Taiwan, the majority of lung cancers have oncogenic driver mutations. Precision medicine is a crucial issue. My research over the past few years has confirmed the effectiveness of EGFR and ALK targeted therapies in real-world settings. By understanding the mechanisms of resistance to EGFR and ALK inhibitors, we can identify tumors with different resistance mechanisms that require distinct treatment approaches. Moreover, even tumors that develop resistance after combination therapy with EGFR and MET inhibitors still have the potential to benefit from further targeted treatments. By analyzing the driver genes of patients’ tumors, the reasons for resistance, and the mechanisms involved in the re-emergence of resistance, we can enhance the precision of precision medicine to improve the prognosis for lung cancer patients.