|
news & views IMMUNOTHERAPY Targeting the cancer mutanome of breast cancer
A metastatic hormone receptor–positive breast cancer, usually resistant to immunotherapy, is successfully treated with tumor-infltrating lymphocytes enriched for neoantigen reactivity, underscoring the broad potential of this immunotherapeutic approach. Laszlo G. Radvanyi Mutations in cancer (the tumor ‘mutanome’) are the source of neoantigens that can be recognized by the immune system as foreign-like peptides called ‘neoepitopes’ presented on major histocompatibility complex (MHC) molecules1 . Although the cancer mutanome is considered to be a possible source of potent tumor antigens for cancer immunotherapy, it remained largely out of reach for decades because of the lack of suitable genomic and proteomic methods to identify actionable mutations. Recent advances in genomic DNA and RNA sequencing as well as the improved software tools for discerning singlenucleotide variants (SNVs) and other relevant genomic alterations together with improved prediction of peptide sequence binding affinities to different subtypes of MHC molecules have made this elusive source of antigens actionable in a realistic and practical way2 .
A number of immunotherapeutic approaches, including vaccination with synthesized long peptides or RNA, DNA plasmid, or viral vector vaccines that encode relevant neoepitope sequences, as well as activated T cell therapies that recognize neoantigens, are now in development3 . One of the most well-established cell therapies for solid tumors that has the longest clinical history entails the use of autologous tumor-infiltrating lymphocytes (TILs) that are expanded ex vivo and reinfused into patients with cancer4,5 . TIL therapy is especially active against tumors with a higher somatic mutation burden, such as cutaneous melanoma and lung cancer linked to smoking. Accumulating data are also showing a strong association of clinical response to other immunotherapies, such as T cell checkpoint blockade, with tumor mutational burden and neoepitope load across a range of different cancers6 . However, data comparing TMB and actual neoepitope presentation to the immune system has found that only a fraction of mutations (<0.1%) actually result in productive or actionable epitopes triggering a T cell response.
Tumors with fewer mutations, such as hormone receptor–positive (ER+) breast cancer and non-microsatellite-instable cancers, have been suggested to be ‘off-limits’ for mutanome-directed therapy7 . It is in this context that an article in this issue of Nature Medicine by Zacharakis et al.8 reporting the successful treatment of a patient with metastatic ER+ (luminal A) breast cancer with multiple disseminated tumors refractory to multiple lines of prior therapy (Patient 4136) using isolated and expanded neoantigen reactive TIL is presented . In their study (summarized in Fig. 1), Zacharakis et al.8 identified a series of point mutations in a recurrent right breast subcutaneous mass using whole-exome sequencing (WES) and RNA-seq.
They then cultured TILs from this isolated tumor mass and screened the expanded T cells for reactivity against a series of neoepitopes using autologous antigenpresenting dendritic cells (DCs) transduced with plasmids encoding strings of these neoepitopes so that the neoepitopes are processed and presented8 . In this manner, they isolated TILs reactive against mutations in four genes (SLC3A2, ECPAS (KIAA0368), CADPS2 and CTSB) and used a number of assays to confirm the reactivity of the different neoepitope-reactive TIL clones. Another innovative approach they used to confirm the reactivity of the T cells was measuring the induced expression of the T cell activation marker 4-1BB (CD137, also known as TNFRS9) in response to neoepitopes using flow cytometry12. The patient was ultimately treated with 8.2 × 1010 TILs expanded from three tumor fragments harboring the highest frequency of T cells against the four mutant genes. As is common in all current TIL transfer therapies, the patient also underwent prior nonmyeloablative lymphodepletion and received a course of high-dose interleukin (IL)-2 to support the expansion and survival of the infused TILs in vivo8 . What was new in this protocol compared with previous autologous TIL transfers was that the patient also received an additional immunotherapy using anti–programmed cell death protein 1 (PD-1) antibody, a so-called ‘immune checkpoint blockade’ (ICB) therapy, administered over four infusions at day –2 and at 3, 6, and 9 weeks after TIL adoptive transfer. Follow-up analysis of target lesions revealed that the patient had a complete response after 42 weeks in multiple subcutaneous and liver metastases with no additional therapy8 , an unprecedented response in such advanced breast cancer. This remarkable response now sets the stage for further testing the effectiveness of neoantigen-specific TIL isolation and therapy on other similar patients with breast cancer. Importantly, this study, along with other recent results in ovarian cancer by Coukos and colleagues9 showing a consistent ability to detect neoepitopespecific TILs in ovarian cancer, is dispelling a growing dogma that tumors with fewer mutations, such as hormone-driven breast cancer, ovarian cancer, prostate cancer and other adenocarcinomas, are not amenable to neoantigen-based targeted therapies. Furthermore, in addition to showing the feasibility of the neoantigen-specific TIL approach in this type of breast cancer, the study also showed that combining ICB with neoantigen-specific TIL is safe and should be further explored in future trials. It is unclear to what degree the ICB contributed to the patient’s response.
However, as no PD-L1 expression was found in pretreatment biopsies from the patient, the results suggest that the neoantigen-reactive TILs were the key factor and that ICB acted synergistically by blocking PD-1 signaling in the infused T cells, enhancing their antitumor activity in the tumor microenvironment.
This hypothesis could not be tested in the current study, but needs to be tested in future trials in which anti-PD-1 is given following TIL infusion and high-dose IL-2. In these future studies, changes in the persistence of infused T cell clones and the endogenous T cell repertoire should be tracked using high-throughput T cell receptor (TCR)-α, TCR-β, complimentary-determining region 3 (CDR3) sequencing and TCR-β clonotyping after TIL infusion and then following anti-PD-1. This will also reveal Nature Medicine | www.nature.com/naturemedicine © 2018 Nature America Inc., part of Springer Nature. All rights reserved. news & views whether additional T cell clones from the endogenous T cell repertoire (not found in the infused TILs) also get activated through an ‘antigen spreading’ or ‘epitope spreading’ process, including T cell clones possibly recognizing other neoantigens not found in the original TIL isolates. The authors were able to track TCR clones in the infused patient and found that a significant amount of the infused T cell clones (8 of 11 clones) persisted in the patient for at least 17 months post-treatment8 . Interestingly, additional clones also expanded that were originally undetectable in the infusion product. A few of these were against two of the mutated genes and were detected at a <0.001% frequency in the infusion product after using an ultra-deep sequencing method not routinely performed, while another dominant clone that expanded in vivo after 6 weeks was against an undefined antigen. These results are too premature to make any definitive conclusions but raise the question as to the relative contributions of dominant and detectable clones at the initial time of tumor isolation and TIL expansion versus undetectable low-frequency neoepitopereactive clones that may expand later in vivo, contributing to the tumor clinical response. In addition, they further raise the question of the role of endogenous reconstituting T cell clones toward durable clinical responses that can be activated through antigen or epitope spreading or to subdominant or newly emerging neoepitopes during the course of therapy. The inability to detect and expand low-frequency neoepitope-reactive TILs significantly contributing to antitumor responses may be a limitation of current TIL approaches for targeting the mutanome. Nevertheless, one key advantage of using TILs versus other immunotherapeutic approaches is that they are a natural repertoire of T cells recognizing a broad array of heterogeneous tumor antigens, including neoantigens as well as overexpressed self-antigens (which still should not be ruled out as viable tumor targets)10. They also contain both CD8+ and CD4+ T cells and thus naturally recognize both MHC class I– and MHC class II–presented neoantigens, thereby offering the ability to target a broader array of heterogeneous neoantigens. Recent data suggest that most tumor neoantigens may in fact be presented on MHC class II molecules, making neoepitope vaccines problematic because of the lack of reliable peptidebinding algorithms for MHC class II versus class I (ref.11). Incidentally, the pooled TIL infusion products against the four mutated breast cancer genes in the Zacharakis et al.8 study were 62.5% CD4+ T cells. However, using TIL as a route to neoantigen-based immunotherapy can have some drawbacks, including the need to isolate and expand T cells from small tumor resections or biopsies, which may not always be feasible. It also depends on the identification and isolation of preexisting T cell clones against neoepitopes and may miss a significant part of the targetable tumor mutanome owing to the low expression of some mutant proteins and their corresponding neoepitopes presented on different MHC molecules by antigen-presenting cells, especially when MHC loci are heterozygous in a patient and the avidity of neoepitope presentation is reduced or when the frequency of the neoepitope-specific TIL clones is too low12. The expansion of undetectable T cell clones in Patient 4136 after infusion suggests that the TIL method used here can miss highly tumor-specific and potent neoepitopes that perhaps another approach, such as a neoepitope vaccine, can target. Thus, perhaps in the future both approaches can be combined for optimal results. As elegantly shown here by Zacharakis et al.8 using neoepitope-reactive TIL, as well as other neoepitope vaccine approaches under study, we are now at the cusp of a major revolution in finally realizing the elusive goal of being able to target the plethora of mutations in cancer through immunotherapy. The challenge in the next decade will be to find creative ways to optimally integrate cancer neoantigen targeting with other therapies to maximize the benefits for patients. We now also have the opportunity to use approaches such as TIL therapy to treat tumors with lower mutational burdens previously thought to be ‘untreatable’ using mutanome-directed immunotherapy (Fig. 1). ❐ Laszlo G. Radvanyi Ontario Institute for Cancer Research, Toronto, Ontario, Canada. e-mail: laszlo.radvanyi@oicr.on.ca Published: xx xx xxxx https://doi.org/10.1038/s41591-018-0065-z References 1. Campbell, B. B. et al. Cell 171, 1042–1056.e1010 (2017). 2. Alvarez, B., Barra, C., Nielsen, M. & Andreatta, M. Proteomics https://doi.org/10.1002/pmic.201700252 (2018). 3. Efremova, M., Finotello, F., Rieder, D. & Trajanoski, Z. Front. Immunol. 8, 1679 (2017). 4. Dudley, M. E. et al. Science 298, 850–854 (2002). 5. Rosenberg, S. A. Nat. Rev. Clin. Oncol. 8, 577–585 (2011). 6. Goodman, A. M. et al. Mol. Cancer Ter. 16, 2598–2608 (2017). 7. Schumacher, T. N. & Hacohen, N. Curr. Opin. Immunol. 41, 98–103 (2016). 8. Zacharakis, N. et al. Nat. Med. https://doi.org/10.1038/s41591- 018-0040-8 (2018). 9. Bobisse, S. et al. Nat. Commun. 9, 1092 (2018). 10. Robbins, P. F. et al. Nat. Med. 19, 747–752 (2013). 11. Kreiter, S. et al. Nature 520, 692–696 (2015). 12. Chowell, D. et al. Science 359, 582–587 (2018). Acknowledgements The author thanks all colleagues and scientists who over the years have shared ideas and concepts and have worked diligently to develop cell therapy and vaccination for cancer, bringing it to the forefront of immunotherapy. Competing interests The author declares no competing interests. Patient Infuse TILs Tumor and blood Tumor and germline DNA/RNA WES RNA-seq Mutation calls Transduce patient TIL outgrowth DC Neoepitope prediction Synthesize neoepitope Clone mutant epitopes into minigenes ASPLGTRNK LMLYSHTYV ASTHLNKYV VLSHGSFVM KLMVLSEGVGSF ELMKLSGNASWKGE ALSMETKLVKLMEGVASE Verify neoepitope reactivity Fig. 1 | Adoptive T cell therapy successfully treats a metastatic hormone receptor–positive breast cancer. Zacharakis et al.8 carried out WES on a metastatic hormone receptor–positive breast cancer. After identification of mutated genes, strings of peptides predicted to bind to MHC (also known as human leukocyte antigen (HLA)) class I and class II were cloned into minigene cassettes and transduced into autologous DCs from the patient. TILs grown out of the isolated tumor were screened for reactivity against the transduced DCs using a number of readouts for T cell activation; the TIL cultures most active against the neoepitopes were selected and expanded to large numbers (billions) and infused back into the patient. Nature Medicine | www.nature.com/naturemedicine © 2018 Nature America Inc., part of Springer Nature. All rights reserved.