Characterization of Hyperprogression after Immunotherapy in a Lung Adenocarcinoma Patient with Strong Expression of Programmed Death Ligand 1.
Yu Peng,Lei Zhang,Tian Zeng,Li Liu,Xingli Liu,Ying Yang,Henghui Zhang,Zhihua Ruan
DOI: https://doi.org/10.1016/j.jtho.2019.08.007
IF: 20.121
2020-01-01
Journal of Thoracic Oncology
Abstract:In the latest National Comprehensive Cancer Network guideline for NSCLC, pembrolizumab is recommended as preferred first-line therapy of nonsquamous NSCLC patients with programmed death ligand 1 (PD-L1) expression positive (≥50%) and no EGFR, ALK receptor tyrosine kinase (ALK) genomic aberrations. However, not all PD-L1–positive patients can respond to immune checkpoint inhibitors (ICIs). Based upon the categorization of tumor immunity in the microenvironment into four distinct types clarified by Kim et al.,1Kim T.K. Herbst R.S. Chen L. Defining and understanding adaptive resistance in cancer immunotherapy.Trends Immunol. 2018; 39: 624-631Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar the expression of PD-L1 is only one of the elements of successful immune checkpoint blockade (ICB) treatment and the presence of tumor infiltrating lymphocytes (TILs) is indispensable to anti–programmed death 1 (PD-1) therapy as well. Recent research has shown that the highest CD8 T cell density was observed in those lung adenocarcinoma patients who harbored tumor protein p53 (TP53) mutations without co-existence of serine/threonine kinase 11 (STK11) or EGFR mutations benefited from PD-1 blockade more remarkably.2Biton J. Mansuet-Lupo A. Pécuchet N. et al.TP53, STK11, and EGFR mutations predict tumor immune profile and the response to anti–PD-1 in lung adenocarcinoma.Clin Cancer Res. 2018; 24: 5710-5723Crossref PubMed Scopus (179) Google Scholar Notwithstanding, we encountered a Chinese lung adenocarcinoma patient who had a strong expression of PD-L1 (≥90%) along with TP53 mutation without EGFR/STK11 alterations but who did not respond to pembrolizumab; this patient exhibited hyperprogressive disease (HPD) after immunotherapy. To explore what is responsible for this HPD, whole-exome sequencing, targeted sequencing, multiplex immunohistochemistry (mIHC), and HLA typing assay were used to examine his genomic features and tumor immunity in the microenvironment signature. At the end of May 2017, a 65-year-old man presenting with hemoptysis for 2 months was taken to hospital for treatment. Positron-emission tomography–computed tomography (CT) scan revealed abnormal increased intake of 18F-flurodeoxyglucose in the inferior lobe of the right lung and several enlarged mediastinal lymph nodes. Then he underwent right lower lobectomy and was diagnosed with a T3N2M0, stage IIIa, poorly differentiated lung adenocarcinoma (Fig. 1A). The resected tumor specimens were detected by a 61-gene panel (unknown genetic-testing company). Only two pathogenic mutations, BRAF p.G469V (24.24%) and TP53 p.R337L (41.76%) were identified. Approximately 25 days after surgery, follow-up magnetic resonance imaging results showed new metastasis in his right frontal lobe with severe edema around the tumor (Fig. 1B). In the next 4 months, he underwent radiotherapy (60 Gy in 20 times) for 1 month and four cycles of chemotherapy (pemetrexed plus nedaplatin) for 3 months, and there was a clear shrinkage of brain lesions. Nonetheless, 5 months after surgery, his chest CT scan revealed new multiple bilateral lung nodules. Previously, a PD-L1 assay (SP142 staining) of his surgical tumor tissue had been performed. The expression level of PD-L1 was up to 98% (Fig. 1C). Based on such high expression of PD-L1 without other actionable alterations, he received pembrolizumab (150 mg, 2 mg/kg) combined with paclitaxel liposome for two cycles. Unfortunately, 1.5 months after initiation of pembrolizumab, he began to experience multiple dysfunction of central nervous system including loss of memory, aphasia, and disorient. Both chest CT scan and encephalic magnetic resonance imaging confirmed a rapid progression in his lung along with brain (Fig. 1B). Thereafter, his needle pulmonary biopsy sample obtained after exposure to pembrolizumab was sent for a second PD-L1 (SP142) assay and next-generation sequencing test, targeting all exons or hotspots of 465 cancer-related genes at a College of American Pathologists–certified laboratory (GeneCast Biotechnology Co., Beijing). It is noteworthy that in addition to preceding BRAF/TP53 mutations, MET amplification (4.8×) was detected in this test (Fig. 2A) whereas no copy number variation of MNNG HOS transforming gene (MET) was reported in the last 61-gene panel test. Besides, there was still an extremely high expression level of PD-L1 (90%) in the recurrent tumor (Fig. 1C). In February 2018, this patient started treatment with oral c-MET inhibitor crizotinib. One month later, he complained that anhelation was alleviated markedly. Chest CT scan displayed a significant regression of bilateral lung nodules and right hilar lymph nodes compared with that before administration with crizotinib (Fig. 1B). After following 2 months of crizotinib treatment, he exhibited sustained partial response according to Response Evaluation Criteria in Solid Tumors version 1.1 criteria and continued this targeted therapy for 5 months. His condition remains stable.Figure 2Genomic alterations and multiplex immunohistochemistry (mIHC) results of tumor specimens from this patient. A, Copy number variations of MET, MDM2, CDK4, CDK6, and (B) two shared pathogenic mutations in his primary and recurrent lesions. C, Representative images of PanCK (pan cytokeratin), CD3, CD8, PD-1, and CD19 shown by mIHC in resected tumor tissues. Nuclei (blue) were counter-stained by DAPI. Original magnification ×200. D, Quantification results of mIHC from 15 random vision fields.View Large Image Figure ViewerDownload Hi-res image Download (PPT) In light of the criterion of HPD described before, this patient’s worse outcome after immunotherapy can be measured as HPD.3Kato S. Goodman A. Walavalkar V. et al.Hyperprogressors after immunotherapy: analysis of genomic alterations associated with accelerated growth rate.Clin Cancer Res. 2017; 23: 4242-4250Crossref PubMed Scopus (574) Google Scholar Whole-exome sequencing results revealed MET amplification (4×) along with above two somatic mutations BRAF p.G469V (20%) and TP53 p.R337L (32%) in his primary cancer (Fig. 2A and B). Additionally, copy number alterations of MDM2, CDK4, and CDK6 were detected in primary tumor as well as recurrent lesions (Fig. 2A). Despite the potential involvement of MDM2/4 amplification in HPD caused by anti–PD-1 therapy, there was no high-level amplification of MDM2 in this case, which seems to be insufficient to explain this HPD.3Kato S. Goodman A. Walavalkar V. et al.Hyperprogressors after immunotherapy: analysis of genomic alterations associated with accelerated growth rate.Clin Cancer Res. 2017; 23: 4242-4250Crossref PubMed Scopus (574) Google Scholar, 4Fuentes-Antrás J. Provencio M. Díaz-Rubio E. Hyperprogression as a distinct outcome after immunotherapy.Cancer Treat Rev. 2018; 70: 16-21Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar Importantly, there is inconsistent evidence about predicting HPD by MDM2/4 amplification. According to Rizvi et al.,5Rizvi H. Sanchez-Vega F. La K. et al.molecular determinants of response to anti-programmed cell death (PD)-1 and anti-programmed death-ligand 1 (PD-L1) blockade in patients with non-small-cell lung cancer profiled with targeted next-generation sequencing.J Clin Oncol. 2018; 36: 633-641Crossref PubMed Scopus (823) Google Scholar progression-free survival of those NSCLC patients with amplifications in MDM2/4 was not mainly different from that of overall patients.5Rizvi H. Sanchez-Vega F. La K. et al.molecular determinants of response to anti-programmed cell death (PD)-1 and anti-programmed death-ligand 1 (PD-L1) blockade in patients with non-small-cell lung cancer profiled with targeted next-generation sequencing.J Clin Oncol. 2018; 36: 633-641Crossref PubMed Scopus (823) Google Scholar In addition, CDK4/6 amplification or overexpression is believed to be closely related to tumorigenesis.6Choi Y.J. Li X. Hydbring P. et al.The requirement for cyclin D function in tumor maintenance.Cancer Cell. 2012; 22: 438-451Abstract Full Text Full Text PDF PubMed Scopus (247) Google Scholar Furthermore, CDK4/6 inhibition was reported to be capable of inducing antitumor immunity through enhancing tumor antigen presentation and suppressing regulatory T cells.7Goel S. DeCristo M.J. Watt A.C. et al.CDK4/6 inhibition triggers anti-tumour immunity.Nature. 2017; 548: 471-475Crossref PubMed Scopus (728) Google Scholar However, little attention has been paid to the role of copy number variation of CDK4/6 in resistance to ICIs or HPD of immunotherapy so far. A most confusing problem of this case is why he harbored such strong PD-L1 expression as well as high frequency of TP53 mutation without EGFR/STK11 alterations but was refractory to anti–PD-1 antibody pembrolizumab. We examined the immune microenvironment of his primary lesions by mIHC and found that there were rather low degrees of infiltration of CD3+/CD8+ lymphocytes as shown in Figure 2C. Quantitative analysis revealed that the percentages of CD8+ and CD3+ TILs in tumor region, stroma region, and whole region were 0.26%, 1.39%, and 0.62% along with 1.27%, 3.87%, and 2.11%, respectively (Fig. 2D). Despite not having enough specimens of recurrent lesions to perform mIHC, it was shown that extremely few lymphocytes were present in the PD-L1 assay (Fig. 1C). On the basis of the principles of ICB treatment, lack of CD8+ TILs is likely responsible for his nonresponse to pembrolizumab, but it perhaps cannot fully illuminate the reason for developing HPD.1Kim T.K. Herbst R.S. Chen L. Defining and understanding adaptive resistance in cancer immunotherapy.Trends Immunol. 2018; 39: 624-631Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar Finally, we performed validated HLA typing assay of this individual’s DNA from peripheral blood cells. Chowell et al.8Chowell D. Morris L.G.T. Grigg C.M. et al.Patient HLA class I genotype influences cancer response to checkpoint blockade immunotherapy.Science. 2018; 359: 582-587Crossref PubMed Scopus (579) Google Scholar discovered that ICB-treated melanoma patients with HLA-B*15:01 allele of HLA-B62 supertype had poor outcomes.8Chowell D. Morris L.G.T. Grigg C.M. et al.Patient HLA class I genotype influences cancer response to checkpoint blockade immunotherapy.Science. 2018; 359: 582-587Crossref PubMed Scopus (579) Google Scholar Herein, both of classical HLA-I loci of this patient are listed in Table 1. Notably, he has two different germline alleles (HLA-B*15:01 and HLA-B*15:02) of HLA-B62 supertype; that is, it is “homozygosity” for the HLA-B62 supertype. Molecular dynamics simulations showed that the structural feature and dynamical factors of HLA-B*15:01 protein likely impair its affinity with peptide for effective neoantigen presentation, which is possibly linked to its negative impact on ICB-treated patients’ survival.8Chowell D. Morris L.G.T. Grigg C.M. et al.Patient HLA class I genotype influences cancer response to checkpoint blockade immunotherapy.Science. 2018; 359: 582-587Crossref PubMed Scopus (579) Google Scholar Theoretically, the HLA-B*15:02 protein might have the same or similar property as HLA-B*15:01 for interaction with antigen peptide. Given this, it may have dual influence upon the cancer-immunity cycle. For one thing, HLA-I genotype-caused weak capacity for antigen presentation in antigen-presenting cells goes against priming and activation of T cells in lymph nodes near tumor lesions. For another, cancer cells take advantage of reduced self-antigen presentation by HLA-I molecules to escape being recognized and attacked by CD8+ TILs. The synergy effects from above two aspects probably involved in scanty CD8+ and CD3+ TILs along with poor response to pembrolizumab in this patient, although his somatic mutation profile was actually in favor of T cell infiltration.Table 1Classical HLA Class I Genotype of the PatientHLA Class IHLA Class I AllelesHeterozygosityHLA Class I SupertypesHLA-AA*11:01 A*24:02YesA03 A24HLA-BB*15:01 B*15:02YesB62 B62HLA-CC*01:02 C*08:01YesNA NA Open table in a new tab With respect to the predictive value of HLA in ICB treatment, we believe that the HLA-I genotype is perhaps not suitable to be used as an independent biomarker for NSCLC, but it can be integrated with other biomarkers to better understand clinical outcome of NSCLC patients receiving PD-1/PD-L1 inhibitors. To date, the exact molecular mechanisms behind HPD of immunotherapy are still poorly understood. In one sense, HPD is a special type of primary resistance to ICB treatment. Briefly, because of a lack of TILs or the defects associated with cancer-immunity response or immunosuppressive microenvironment, ICIs do not always take effect, that is, there is resistance. Cancer cells are generally governed by some key deregulated signaling pathways or pathogenic genomic alterations. Immune checkpoint-mediated immunosuppression is also a deregulated mechanism facilitating tumor growth. When the PD-1/PD-L1 axis is blockaded by ICIs, other innate pro-survival pathways of cancer cells are prone to be aberrantly activated and certain cell subsets with specific driver mutations might gain growth advantage. Under the circumstances, if antitumor activity of TILs restored by ICB is insufficient to cope with such kind of cancer cells, the accelerated tumor growth or rapid progression likely occurs, that is, HPD. As evidenced by the fact that those NSCLC patients with driving EGFR alterations usually have worse outcome after undergoing anti–PD-1 therapy, it is plausible that MET amplification, another important driving alteration of NSCLC, participates in the pathologic process of primary resistance to pembrolizumab, even HPD in this patient3Kato S. Goodman A. Walavalkar V. et al.Hyperprogressors after immunotherapy: analysis of genomic alterations associated with accelerated growth rate.Clin Cancer Res. 2017; 23: 4242-4250Crossref PubMed Scopus (574) Google Scholar, 9Gainor J.F. Shaw A.T. Sequist L.V. et al.EGFR mutations and ALK rearrangements are associated with low response rates to PD-1 pathway blockade in non–small cell lung cancer: a retrospective analysis.Clin Cancer Res. 2016; 22: 4585-4593Crossref PubMed Scopus (816) Google Scholar Collectively, MET amplification may cooperate with low-level infiltration of CD8+ T cells potentially related to HLA-B62 supertype to result in HPD of this patient after ICB treatment. This study was granted by the National Key Sci-Tech Special Project of China (No. 2018ZX10302207).