The concept of whether cancer is recognized by the immune system has been a topic of intense discussions and experimentation for more than a century. Philosophically speaking, the argument revolves around one central aspect: cancer originates from "self" tissue, therefore why should the immune system attack it? The traditional concept of immunology which teaches that the main purpose of the immune system is to distinguish between "self" and "non-self" suggests that since cancer is "self" there should be no immune response against it. Current-day immunological advances, however, have struck down this notion. However, before going into these advances in detail, we will first overview the history of cancer immunotherapy in order to provide a background for our discussion.
In the late 1800s a physician at Sloan Kettering Cancer Center, William Coley, made the empirical observation that certain types of tumors would go into remission subsequent to bacterial infections. One of the first patients he saw in his career was a young girl who died of a rapidly progressing sarcoma originating in her arm. A different patient with a similar type of sarcoma in the neck had lived for seven years after diagnosis with no detectable signs of cancer. The only noteworthy difference between the two patients, in his mind was that the latter patient had repeated encounters with bacterial infections. This prompted Coley to search the medical literature, where he found that sarcoma remission had previously been documented to be associated with erysipelas, a streptococcal infection of the skin. This prompted Coley to begin purposely inoculating patients with various bacterial extracts with the aim of stimulating an immune response that would somehow "cross-over" and lead to cancer regression. He reported that the first patient purposely inoculated suffered from an inoperable late stage neck sarcoma. This patient was administered extracts of a streptococcal broth that was generated from another patient. According to the published description, this treatment led to a significant remission of a "hen egg"-sized tumour within ten days, and resulted in patient survival for eight years, after which he died of tumour relapse [1, 2]. Eventually, due to the uncontrollable effects of unstandardized bacterial mixtures, Coley generated a combination of heat-killed serratia marcescens and heat-killed streptococci which were eventually named "Coley's Toxins" and sold in the United States by Parke-Davis from 1923 to 1963 . The advent of chemotherapy, as well as unpredictable reactions that patients would have to Coley's Toxin contributed, at least in part, to the discontinuation of this therapy. Nevertheless, William Coley is considered by many the father of modern day cancer immunotherapy . Interestingly, a standardized version of Coley Vaccine is currently being developed by the Canadian biotechnology company MBVax.
Despite the suggestion that immune response to bacterial infections may somehow "re-awaken" the immune system to kill cancer, scientifically, there could have been other explanations for the effect of Coley's Toxins. For example, it may be possible that compounds inside the bacterial extracts had ability to directly kill cancer cells , or to preferentially inhibit tumor angiogenesis . Accordingly, in our discussion of whether the immune response actually inhibits cancer or not, we will turn to animal models.
The era of molecular biology has allowed for gene-specific deletion in animals. This means that genes associated with immune responses can be "knocked-out" of animals so as to study the importance of the specific gene in an in vivo setting. Speaking in very general terms, there are two pathways that the immune system can take when it is activated. The first type is called "Th1", which is involved in destroying cells of the body that are infected from the inside, such as virally infected cells. The second type of immune response is called "Th2", which is responsible for killing targets that reside outside of the cells of the body, such as parasites and certain bacteria [7, 8]. Since cancer consists of cells of the body that have distinguishing properties from the other cells (ie high proliferation, ability to metastasize, etc), it may be possible to rationalize the Th1 path of the immune response would be the one responsible for control of cancer, if the immune response is involved at all. Indeed, Coley's toxin (and its constituents) were identified decades later to be potent inducers of the Th1 cytokine TNF-alpha, as well as activators of this general immune response pathway [9, 10]. The discovery of transcription factors that induce Th1 or Th2 immunity has allowed experimental assessment of the roles of these types of immune responses in control of cancer. Transcription factors controlling Th1 immunity include T-bet , and STAT4 , and those controlling Th2 include GATA-3  and STAT6 [14, 15]. When various tumors are administered to STAT6 knockout animals (therefore having a Th1 predisposition since the Th2 pathway is ablated), these tumors are either spontaneously rejected , or immunity to them is achieved with much higher potency compared to wild-type animals . Furthermore, in STAT6 knockout animals, immunologic resistance to metastasis formation is observed . On the other hand, STAT4 knockout mice lack Th1 capability and therefore have only upregulated Th2 immunity. Such animals allow accelerated cancer formation after treatment with chemical carcinogens .
While the above suggest the importance of the Th1 immune response in controlling tumors, in many cases animal data does not translate efficiently to human disease. Accordingly, we turn our attention to situations where immune suppression is induced either by genetic abnormality or in response to a medical condition. In general, natural killer (NK) cell activity is associated with Th1 immune responses and tumor immunity . Patients with the congenital abnormality Chediak-Higashi Syndrome, are characterized by absent or severely diminished NK function. In this population, the overall incidence of malignant tumors is 200–300 times greater than that in the general population . Another example of an inborn trait associated with immune deviation is patients born with a specific polymorphism of the IL-4 receptor gene that is known to be associated with augmented Th2 responses. Multivariate regression analysis showed that this polymorphism was an independent prognostic factor for shorter cancer survival and more advanced histopathological grade . In addition to inborn genetic abnormalities, the immune suppressive regimens used to prevent transplant rejection are associated with a selective inhibition of Th1 responses [23–25]. In support of the concept that suppression of Th1 immunity is associated with cancer onset, the incidence of cancer in the post-transplant population is markedly increased in comparison to controls living under similar environmental conditions [26–31]. In terms of disease associated immune suppression, HIV infected patients also have a marked predisposition to a variety of tumors, especially, but not limited to lymphomas, as a result of immunodeficiency .
The above examples support a relation between immune suppression (or Th2 deviation) and cancer proliferation, the opposite circumstance, of immune stimulation resulting in anticancer response is also documented. Numerous clinical trials using antigen specific approaches such as vaccination with either tumor antigens alone [33, 34], tumor antigens bound to immunogens [35, 36], tumor antigens delivered alone  or in combination with costimulatory molecules by viral methods , tumor antigens loaded on dendritic cells ex vivo [39–41], or administration of in vitro generated tumor-reactive T cells , have all demonstrated some, albeit modest clinical effects. It is documented that inappropriate immune responses (broadly speaking Th2 responses) can actually stimulate tumor growth [43, 44]. Accordingly, these data all support the presumed recognition of cancer by the immune system and the notion that the immune system, if stimulated properly, may induce cancer regression.
The philosophical question posed at the beginning of this discussion; how can the immune system recognize cancer when cancer is part of self, is answered in the following manner. The immune system is not responsible for seeing only "self" versus "non-self" but actually seeing and responding to different variations of "self". The tumor, in its quest for proliferative advantage, ability to metastasize, and need for formation of new blood supply, actually expresses new molecules at levels that are recognized by the immune system. Immunological recognition of molecules needed for the tumor to have the "cancer phenotype" has been well-documented. We will not provide an overview of these data here but will provide some examples. Specifically, the proliferative advantage of tumors is associated with growth factor receptor upregulation, accordingly immune responses to various such receptors are known to exist naturally or to be inducible [45, 46]. The same is true for matrix metalloproteases involved in tumor extravasation and metastasis [47, 48], as well as for angiogenic factors involved in formation of new blood vessels [49, 50]. The question, of course remains, if the immune system can see cancer, why does it not eradicate it, and why has the clinical implementation of cancer immunotherapy yielded such poor results at the bedside?