Thursday, November 9, 2017

Hot and Cold Tumors

Hot and Cold Tumors

Each human malignancy is uniquely different from every other cancer, but from the point of view of immunotherapy, there are just two types of tumors: Hot and Cold. Likewise, there are two types of immunotherapy treatment: Immune response inducing and anti-immunosuppressive.

There are at least 100 different types of cancers if they are classified based on their tissue of origin. Looking more carefully at the cell type from which the malignant growth originates, there are at least 200 different types of cells in the human body, each of which can give rise to cancer. Even this is a superficial view since a cell type will behave differently when it is a different organ. Also, a tumor can be less or more aggressive, sometimes to the degree where it is classified as a different cancer. Even if two tumors originate from the same cell type in the same organ, their mutation profile will be different and thus in the end every tumor is a unique individual. 

Despite all this complexity, treatments in oncology have traditionally been much less complicated. “Pancreatic cancer” is treated with drugs X, Y and Z without regard for the individual characteristics of the tumor, and the same is true for most cancer types. Perhaps the most glaring example is sarcomas, which lumps more than 50 different tumor types together, but there are only a couple of widely used therapeutic regimens. In the past two decades treatment paradigms have changed somewhat for certain tumors such as lung, breast and colon cancer, and melanoma, where molecular subtypes are treated differently, increasing complexity in treatment choices. 

Immunotherapy has recently made a breakthrough into treatment of many types of cancers, including melanoma, lung, kidney, head and neck, urinary tract cancer, lymphoma and others. Amazingly, what has emerged is that from the point of view of immunotherapy, there seem to be only two types of tumors: Hot and Cold. 

The former type are characterized by the presence of tumor infiltrating lymphocytes (T-cells, a type of white blood cell), which are capable of recognizing aberrant proteins in or on cancer cells. Typically the easiest types of proteins to recognize are neoantigens (antigen = immunological target), which means proteins not normally found in the body. Certain types of mutation such as frameshift, point mutations, in-frame deletions and gene fusions can result in neoantigens. Neoantigens are more immunogenic (=able to cause an immune response) than overexpression or misplaced expression of normal antigens. Logically, the more mutations in the tumor, the more neoantigens, resulting in higher likelihood of T-cells detecting the tumor.

In theory, tumor-recognizing T-cells would be expected to kill the cells they detect. This probably happens all the time in our bodies, and we never knew that we had cancer. However, eventually some tumors are able to escape detection by the immune system, in a three-step process known as the three E’s of immunoevasion (Elimination, Equilibrium, Escape). Every tumor that is diagnosed has been able to escape killing by T-cells. One key mechanism is development of immunosuppression, meaning mechanisms that counteract detection and killing by immunological cells. These immunosuppressed tumors often feature expression of a molecule called PD-L1. For further reading on this please see my book referenced below.

In contrast to Hot tumors, Cold tumors don’t have a lot of mutations, often lack neoantigens and feature low or no T-cells inside the tumors. Since there is little cellular immunity against the tumor, there is less need for immunosuppression and therefore PD-L1 expression is typically low. One variant of a cold tumor is an “immune excluded” tumor where there might be mutations but because of certain biological reasons the T-cells are unable to penetrate the tumor. 

Fitting well with classification of tumors into two categories, Hot and Cold, there are two main types of immunotherapy treatment: 1) immune response inducing and 2) anti-immunosuppressive therapies.

Examples of the former category include cancer vaccines, T-cell therapies such as TIL, TCR and CART therapies, and oncolytic viruses, also known as oncolytic immunotherapy. Of these approaches, there is one CART approved for leukemia (tisagenlecleucel or Kymriah) and one oncolytic virus (talimogene laherparepvec or T-Vec or Imlygic) approved for treatment of melanoma. 

It seems likely many more drugs in these categories will be approved in the next decade. It has not been studied much but it could be that oncolytic viruses work best in Cold tumors, as proposed by Taipale and colleagues in 2016. Viruses such as adenovirus are able to destroy cancer cells in a manner which irritates the immune system, resulting in a T-cell dominated anti-tumor immune response, which can achieve prominent therapeutic effects if there was little or no immunosuppression present. 

A clearer breakthrough has already been seen in the latter category. The immunotherapies most widely used currently (eg nivolumab, pembrolizumab, atezolizumab, avelumab) block interaction between PD-L1 and its receptor PD-1. These “checkpoint inhibiting” drugs only work in Hot tumors characterized by neoantigens, T-cells and PD-L1 expression. They cannot generate new immunity and thus if T-cells were not present the therapy will probably not work. 

Oncologists love to mix and match ingredients to make their witches’ cocktails. Thus it comes as no surprise that oncolytic viruses have already been combined with checkpoint inhibitors, and trials were started even before the difference between Hot and Cold tumors was understood. These trials are still in their early phases or in planning but the initial data emerging is awesome. Ribas and collagues reported that the majority of patients responded emphatically without severe side effects, and one third had a complete response.

Although the combination of oncolytic virus with checkpoint inhibition is very well tolerated, it might represent overtreatment for some patients. It might be better to study the immunological status of tumors and treat Hot tumors with checkpoint inhibitors or their combinations, while Cold tumors could be treated with oncolytic viruses. Impressively, tests allowing distinguishing between Hot and Cold tumors were  rapidly incorporated into commercially available drug sensitivity kits such as OncoSTRAT&GO and Caris Molecular Intelligence. Thus, aficionados can already offer this testing to their patients.

Nevertheless, since generation of an immune response in the body always results in an immunosuppressive counter-response (this is how the body protects itself against auto-immunity), oncolytic viruses might benefit from an anti-immunosuppressive “companion treatment”. This would prevent dampening and attenuation of the anti-tumor response. 

Tumors have tremendous capacity for developing resistance, for example through a target-negative clone. Thus, there might be fewer relapses following checkpoint inhibition if it would be combined with a therapy such as an oncolytic virus that can increase the number and broaden the repertoire of tumor detecting T-cells. The beauty of the oncolytic approach is that it is dynamic and continuous – whichever epitopes are present in the tumor will be presented to the immune system. 

It is not trivial at all how these drugs are sequenced and dosed. Also, there are at least a dozen different types of oncolytic viruses but they are not likely to be equally effective in generating a T-cell dominated anti-tumor immune response. Likewise, there are dozens of potentially druggable immune checkpoints.
In summary, underlying incredible complexity between tumors, in the context of immunotherapy all tumors seems to fall into one of two classes: Hot or Cold. With regard to available immunotherapies, oncolytic viruses appear to work in Cold tumors while checkpoint inhibitors work in Hot tumors. The combination of these approaches is appealing and could become commonly used in the next decade. 


Hemminki A. Crossing the Valley of Death with Advanced Therapy. Published by Nomerta, Turku, Finland, 2015. Available at and several e-book stores globally

Ribas A, Dummer R, Puzanov I, VanderWalde A, Andtbacka RHI, Michielin O, Olszanski AJ, Malvehy J, Cebon J, Fernandez E, Kirkwood JM, Gajewski TF, Chen L, Gorski KS, Anderson AA, Diede SJ, Lassman ME, Gansert J, Hodi FS, Long GV. Oncolytic Virotherapy Promotes Intratumoral T Cell Infiltration and Improves Anti-PD-1 Immunotherapy. Cell. 2017 Sep 7;170(6):1109-1119.e10. doi: 10.1016/j.cell.2017.08.027.

Taipale K, Liikanen I, Koski A, Heiskanen R, Kanerva A, Hemminki O, Oksanen M, Grönberg-Vähä-Koskela S, Hemminki K, Joensuu T, Hemminki A. Predictive and Prognostic Clinical Variables in Cancer Patients Treated With Adenoviral Oncolytic Immunotherapy. Mol Ther. 2016 Aug;24(7):1323-32