cancer cell

Cancer Cell

The 27 November issue of Nature contains a wealth of new studies on how immune checkpoint inhibitors target various types of cancer, and how researchers and physicians might be able to identify the patients who are most likely to benefit from treatment with these agents.

These studies are described in five papers published in that issue of Nature. This issue also contains a “News & Views” commentary on these articles by Drs. Jedd D. Wolchok and Timothy A. Chan (both at the Memorial Sloan Kettering Cancer Center). This article serves as an introduction to the five research articles.

In addition, Science Magazine published a commentary on these articles, entitled “Multiple boosts for cancer immunotherapy”, by contributing correspondent Mitch Leslie.

Checkpoint inhibitors can be used to treat several types of cancer

One important result of these studies is the expansion of the range of cancers that can be treated via immunotherapy beyond melanoma, kidney cancer, and non-small cell lung cancer (NSCLC). The papers by Powles et al. and Herbst et al. contain results from a Phase 1 clinical trial of Genentech’s monoclonal antibody (MAb) PD-L1 blocker MPDL3280A. Herbst et al. reported that MPDL3280A showed therapeutic responses in patients with NSCLC, melanoma, renal cancer, and head and neck cancer. Powles et al. focused on the effects of this agent in a larger group of patients with metastatic urothelial bladder cancer (UBC). In both reports, researchers documented that a subset of patients experienced durable responses, and that the treatment showed low toxicity.

We discussed earlier presentations of the results of the Phase 1 trial of MPDL3280A in our Insight Pharma Report (IPR), Cancer Immunotherapy: Immune Checkpoint Inhibitors, Cancer Vaccines, and Adoptive T-Cell Therapies. As we discussed in this report, the FDA granted breakthrough therapy designation for MPDL3280A for treatment of UBC. Roche/Genentech has initiated a Phase 2 clinical trial (clinical trial number NCT02108652) of MPDL3280A in UBC. UBC is the ninth most common cancer in the world. Metastatic UBC is associated with a poor prognosis, and has few treatment options. There have been no new treatment advances in nearly 30 years.

Checkpoint inhibitors work by reactivating tumor-infiltrating T cells (TILs)

Perhaps the most important finding of the research published in the November 27th issue of Nature is that checkpoint inhibitors work via reactivating endogenous tumor-infiltrating T cells. (These T cells are often called “TILs”, which is an acronym for “tumor-infiltrating lymphocytes”.)

For example, as described in the Powles et al. report, Genentech’s PD-L1 blocker MPDL3280A was found to be especially effective in treating patients whose tumors contained PD-L1-positive TILs. As we discussed in our IPR report, Genentech researchers found that MPDL3280A not only targets PD-L1 on the surface of tumor cells, but also PD-L1 on the surface of TILs. PD-L1 on activated T cells interacts not only with PD-1, but also with B7 on the surface of antigen presenting cells, sending a negative signal to the T cells. MPDL3280A targets the PD-L1-B7 interaction, thus enabling reactivation of PD-L1-bearing TILs so that they can attack the tumor.

As we also discuss in our report, targeting PD-1, PD-L1, and CTLA-4 may also be important in reversing immunosuppression by regulatory T cells (Tregs), which typically heavily infiltrate tumors. This provides another mechanism by which checkpoint inhibitors can reactivate TILs and thus induce anti-tumor immune responses.

As described in Powles et al, MPDL3280A was engineered with a modification in the Fc domain that eliminates antibody-dependent cellular cytotoxicity (ADCC). Genentech researchers did this because PD-L1 is expressed on activated T cells, and they wanted an anti-PD-L1 MAb agent that would reactivate these T cells, not destroy them via ADCC.

In the studies described by Herbst et al., researchers showed that Genentech’s PD-L1 blocker MPDL3280A gives antitumor response across multiple types of cancer, in tumors that expressed high levels of PD-L1. These responses especially occurred when PD-L1 was expressed by TILs. The studies suggest that MPDL3280A is most effective against tumors in which endogenous TILs are suppressed by PD-L1, and are reactivated via anti-PD-L1 MAb targeting.

In the Tumeh et al. study, the researchers found that patients responding to treatment with Merck’s MAb PD-1 blocker pembrolizumab (Keytruda) showed proliferation of intratumoral CD8+ T cells that correlated with reduction in tumor size. Pretreatment tumor samples taken from responding patients showed higher numbers of CD8, PD-1, and PD-L1 expressing cells at the invasive tumor margin and within tumors, with a close proximity between PD-1 and PD-L1, and a clonal TCR repertoire.

Based on this information, the researchers developed a predictive model based on CD8 expression at the invasive tumor margin. They validated this model in an independent 15-patient cohort. The researchers concluded that tumor regression due to treatment with the PD-1 blocker pembrolizumab requires preexisting CD8+ T cells whose activity has been blocked by PD-1/PD-L1 adaptive resistance. This study, like those of Powles et al. and Herbst et al., thus indicate that checkpoint inhibitors work against cancer by reactivating TILs. The Tumeh et al. study also indicates that CD8 expression at the invasive tumor margin is a predictive biomarker for sensitivity of patient tumors to treatment with anti-PD-1 checkpoint inhibitors.

The Powles, Herbst, and Tumeh reports all involved studies in human patients. However, the other two papers—Yadav et al. and Gubin et al. involve studies in mouse tumor models.

In the study of Yadav et al., the researchers used their mouse model to develop a method for discovering immunogenic mutant peptides in cancer cells that can serve as targets for T cells. They sequenced the exomes of two mouse cancer cell lines, and looked for differences with the corresponding normal mouse exomes. They also identified which of the neoantigens that they identified via exome sequencing could bind to histocompatibility complex class I (MHCI) proteins, and thus could be presented to T cells. They then modeled the MHC1/peptide complexes, and used these models to predict which of these neoantigens were likely to be immunogenic.

These methods identified only a few candidate neoantigens. Vaccination of tumor-bearing mice with these neoantigens resulted in therapeutically active T-cell responses. In addition, the researchers developed methods for monitoring the antitumor T cell response to peptide vaccination.

In the study of Gubin et al., the researchers used similar genomic and bioinformatic approaches to those of Yadav et al., and identified two neoantigens that were targeted by T cells following therapy with anti-PD-1 and/or anti-CTLA-4 antibodies. [Human CTLA-4 is the target of the checkpoint blockade inhibitor ipilimumab (Medarex/ Bristol-Myers Squibb’s Yervoy).] As with PD-1 and PD-L1 blockers, we discussed this agent in our IPR report. T cells specific for these neoantigens (in the context of MHCI proteins expressed by the mice) were present in the tumors. These T cells were reactivated by anti-PD-1 and/or anti-CTLA-4 antibodies, enabling the mice to reject the tumors.

As in the study of Yadav et al., the Gubin et al. researchers performed experiments in which they vaccinated tumor-bearing mice with peptides that incorporated the mutant epitopes. This vaccination induced specific tumor rejection that was comparable to treatment with checkpoint blockade inhibitors. As in the case of Yadav et al, the Gubin et al. researchers concluded that specific mutant antigens were targets of checkpoint inhibitor therapy in their mouse models, and that the mutant antigens could also be used to develop personalized cancer vaccines.

Since the studies of Yadav et al. and Gubin et al. were carried out using mouse tumor models, the results are not directly applicable to cancer in human patients. However, the studies suggest that immune checkpoint inhibitors work by reactivating endogenous TILs, and that anti tumor TILs work by attacking specific neoantigens on the tumors.

As we discussed in our IPR report, Dr. Steven Rosenberg (National Cancer Institute, Bethesda, MD) identified specific antigens that were the targets of TILs, both in metastatic melanoma and in metastatic cholangiocarcinoma (a type of epithelial bile duct cancer). However, these target antigens were from human cancers, and they were targets of TILs that has been isolated from patient tumors, cultured and expanded ex vivo, and used in adoptive cellular immunotherapy.

Moreover, the antigens were targets of TIL therapies that resulted in a durable compete remission in the case of the melanoma patient, and long-term tumor regression in the case of the metastatic cholangiocarcinoma patient. The metastatic cholangiocarcinoma case was highlighted in our September 16, 2014 Biopharmconsortium Blog article.

The Yadav et al. paper referenced the Rosenberg group’s work. However, this paper stated that “few mutant epitopes have been described because their discovery required the laborious screening of patient tumour-infiltrating lymphocytes for their ability to recognize antigen libraries constructed following tumour exome sequencing.”

The methods of Yadav et al. (and of Gubin et al.) are thus designed to simplify and accelerate the discovery of immunogenic mutant peptides. They carried out their studies in mouse models, which helped these researchers to develop methods that could potentially discover greater numbers of neoantigens more efficiently. However, it remains to be seen to what extent they can apply their methods to human patients.

Unifying the field of immuno-oncology

As can be seen, for example, from the title of our IPR report, the three major approaches to immuno-oncology in 2014/2015 are development of immune checkpoint inhibitors, of cancer vaccines, and of adoptive T-cell therapies.

In the immuno-oncology papers published in the 27 November issue of Nature, researchers show that checkpoint inhibitors work via reactivating of endogenous TILs. They also (in mouse tumor models) identified neoantigens that are targets of these reactivated TILs, and designed peptide vaccines that were as effective as checkpoint inhibitor therapy in the mouse models. In principle, one can isolate TILs that are reactive to particular neoantigens in the mouse tumors, culture and expand them ex vivo, and infuse them back into the mice to target their tumors. Thus the studies in the 27 November issue of Nature serve as a template for the unification of the immuno-oncology field as it now exists.

However, it will be necessary to apply the methodologies developed by Yadav et al. and Gubin et al. to human patients. And at least so far, peptide vaccines have not been very successful in treating patients, as compared to TIL therapy (in the subset of patients in whom TIL therapy can be done). It is thus possible that once these methods of neoantigen identification are applied to human patients, it will be found that targeting the neoantigens with ex vivo-expanded TILs will be more successful than therapy with peptide vaccines. However, whether this is true awaits the application of the new methodologies to neoantigen identification in human tumors.


As the producers of this blog, and as consultants to the biotechnology and pharmaceutical industry, Haberman Associates would like to hear from you. If you are in a biotech or pharmaceutical company, and would like a 15-20-minute, no-obligation telephone discussion of issues raised by this or other blog articles, or of other issues that are important to your company, please contact us by phone or e-mail. We also welcome your comments on this or any other article on this blog.