Article Text

Download PDFPDF

Dr William Coley and tumour regression: a place in history or in the future
  1. S A Hoption Cann1,
  2. J P van Netten2,
  3. C van Netten1
  1. 1Department of Health Care and Epidemiology, University of British Columbia, Vancouver, British Columbia, Canada
  2. 2Special Development Laboratory, Royal Jubilee Hospital and Department of Biology, University of Victoria, Victoria, British Columbia, Canada
  1. Correspondence to:
 Dr S A Hoption Cann
 Department of Health Care and Epidemiology, University of British Columbia, 5804 Fairview Avenue, Vancouver, BC, Canada V6T 1Z3;


Spontaneous tumour regression has followed bacterial, fungal, viral, and protozoal infections. This phenomenon inspired the development of numerous rudimentary cancer immunotherapies, with a history spanning thousands of years. Coley took advantage of this natural phenomenon, developing a killed bacterial vaccine for cancer in the late 1800s. He observed that inducing a fever was crucial for tumour regression. Unfortunately, at the present time little credence is given to the febrile response in fighting infections—no less cancer.

Rapidly growing tumours contain large numbers of leucocytes. These cells play a part in both defence and repair; however, reparative functions can also support tumour growth. Intratumoural infections may reactivate defensive functions, causing tumour regression.

Can it be a coincidence that this method of immunotherapy has been “rediscovered” repeatedly throughout the centuries? Clearly, Coley’s approach to cancer treatment has a place in the past, present, and future. It offers a rare opportunity for the development of a broadly applicable, relatively inexpensive, yet effective treatment for cancer. Even in cases beyond the reach of conventional therapy, there is hope.

  • Dr William Coley
  • tumour regression
  • fever
  • immunosuppression
  • immunotherapy
  • macrophages
  • neoplasms
  • BCG, bacillus Calmette-Guerin
  • IL, interleukin
  • TNF, tumour necrosis factor

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


“Drugs can only repress symptoms: they cannot eradicate disease. The true remedy for all diseases is Nature’s remedy .... There is at bottom only one genuinely scientific treatment for all diseases, and that is to stimulate the phagocytes. Stimulate the phagocytes. Drugs are a delusion”. So goes the counsel of physician Sir Bloomfield Bonington in George Bernard Shaw’s 1906 play The Doctor’s Dilemma.1 Prophetic words, indeed, that may be the basis for a new paradigm in cancer treatment as we enter the 21st century.

William Coley began his career as a young surgeon at New York Memorial Hospital. Similarly disillusioned with conventional medicine, he wondered whether nature indeed held a cure for cancer. Could he harness the power of the immune system to the benefit of his cancer patients? His search for a new approach began after the loss of his very first patient in 1891. A young woman of 17 had injured her right hand and presented with persistent inflammation and pain. He diagnosed her lesion as a sarcoma of the bone and opted for amputation of her right arm below the elbow. Yet, despite no clinically evident metastases, the patient succumbed to her disease two and a half months after surgery. Shaken by this failure, Coley searched the hospital records for previous cases to learn more about this uncommon disease.

Serendipitously, he discovered the record of an immigrant patient who presented with an egg-size sarcoma on his left cheek.2 The sarcoma was operated on twice and still recurred as a 4.5 inch grape-like cluster below his left ear. The extensive wound after surgery could not be closed and skin grafts were unsuccessful. Ironically, this failure to close the wound would play a key part in the patient’s eventual cure. The tumour progressed and a final operation only partially removed the tumour. The case was considered hopeless. After the last operation, the wound became severely infected with erysipelas (that is, Streptococcus pyogenes) and the patient developed a high fever. Little could be done to stop the infection, yet surprisingly, after each attack of fever the ulcer improved, the tumour shrank, and finally disappeared completely. The patient was discharged four and a half months later. Coley, eager to find this patient, spent weeks searching throughout New York’s lower east side. His efforts were not in vain. The patient, still bearing a large scar from his previous operations, had no trace of cancer and claimed excellent health since his discharge—seven years previously.

Coley suspected that somehow the infection was responsible for this miraculous cure. He resolved to put his theory to the test and infect his next suitable case with erysipelas. In fact, he infected his next 10 patients.2,3 Problems with this approach soon became apparent: sometimes it was difficult to induce an infection, other times there was strong reaction and the disease regressed, however, occasionally the infection was fatal. Due to its unpredictability, he elected to switch to a vaccine containing two killed bacteria:S pyogenes and Serratia marcescens. Experimental work at the time suggested that the latter bacteria increased the virulence of the former.4 In this way, he could simulate an infection (for example, inflammation, chills, fever) without worrying about the risks of an actual infection. This vaccine became known as “Coley’s toxins”. His first case was a success,4 a man bedridden with an inoperable sarcoma involving the abdominal wall, pelvis, and bladder. The disease regressed completely and the patient was followed up until his death from a heart attack 26 years later.5

Coley worked in the Bone Service at the hospital, later becoming its chief in 1915. Thus, the majority of malignancies he treated were sarcomas. Success with Coley’s vaccine, however, was by no means limited to this tumour type. Contrary to what has been suggested by others,6,7 Coley’s vaccine was widely and successfully used by other contemporaries for sarcomas as well as carcinomas, lymphomas, melanomas, and myelomas.5,8–11 A striking feature of his immunotherapy regimen was that even when applied to patients in their final stages of disease some remarkable recoveries were obtained, with patients often outliving their cancer (box 1).5,9–13 Coley himself went on to treat hundreds of patients, and over time, gained an appreciation of how this regimen could be most effectively administered. In fact, Coley was considered to have treated more sarcoma patients than any other physician up to that time.14 Coley considered several points crucial to a patient’s survival.15 First and foremost was to imitate a naturally occurring acute infection, and thus, inducing a fever was essential. Injections were optimally administered daily (or every other day) for the first month or two. To avoid immune tolerance to the vaccine, the dosage was gradually increased over time (depending on patient response). The vaccine was injected directly into the primary tumour and metastases, when accessible. Finally, a minimum six month course of weekly injections was followed to prevent disease recurrence.

Box 1: The country vet

The patient, a veterinary surgeon, was struck over the right superior maxilla by the horn of a bull in February 1901. Several weeks later, he experienced severe pain over the site of injury and to relieve the pain a tooth was pulled. The pain persisted, and in April, the patient was diagnosed as having a mixed cell sarcoma.

Excision of the upper jaw was performed in May. A large tumour was found occupying the maxillary antrum and almost the entire upper jaw. The growth was too extensive for complete removal. There was also an egg-size mass under the left ear. Several days after surgery, the sarcoma began to increase in size, invading the nose and extending along the palate into the pharynx and invading the parotid region.

At this time the patient had difficulty retaining food. His speech was difficult to understand. Although previously able to walk several miles, he now could barely walk from his bed to the door. Jaundice was pronounced; the liver was enlarged. Nausea and vomiting increased. His pulse was between 140–150 beats per minute and was weak and intermittent. Although sceptical, the patient’s physician, O K Winberg agreed to treat him with Coley’s vaccine. In August, starting at a low dose, he gave the patient daily injections while gradually increasing each subsequent dose. Initially, he injected into the upper and lower jaw and later into the abdominal wall. By this time, the patient’s disease progressed to the point where he could no longer take nourishment. He was 113 lbs. The patient had severe abdominal pain; his eyesight began to fail. His teeth became so tightly closed that it was impossible to cleanse his mouth and his speech was almost incomprehensible. He fell into a stupor, but during a brief period of consciousness, he implored to Winberg “either kill me or cure me”.13

Although gradual at first, improvement became increasingly marked after his first fever. Jaundice disappeared after three weeks. By September, he was attending to a large veterinary practice, which often called him away both night and day. By that time the patient weighed 143 lbs. No trace of tumour could be found in the neck, face, or jaw. Abdominal examinations showed nothing abnormal. He continued to receive injections through to January 1902.

The patient remained in good health and free from recurrence until 1907, six years after treatment, when he died of acute nephritis from alcoholic excess.

Shortly before Coley’s death in 1936, Coley’s vaccine received an endorsement in the New and Nonofficial Remedies of the American Medical Association, which stated “its use as a prophylactic in conjunction with conservative or radical surgery” and “inoperable cases may be quite justified”.9 In addition, some textbooks from that period advocated the use of Coley’s vaccine. For example, the 1931 edition of Modern Surgery advised, “after removing a sarcoma in any region, the patient should be given courses of injections of Coley’s fluid”.16 Similarly, in Modern Operative Surgery,17 it is stated for sarcomas that “after amputation, prophylactic injections of Coley’s fluid should be given in doses sufficient to cause a sharp febrile reaction”.

It is often believed that Coley’s vaccine was more effective against sarcomas than carcinomas.18,19 This is due to the fact that Coley primarily treated sarcomas (not surprising considering his specialty) and because Coley initially had less success in treating carcinomas in his early experiments with streptococcus cultures.10 Coley, however, later changed his views as his successes against carcinomas,20–22 and those reported to him by others,5,23 accumulated. In an analysis of 896 patients with microscopically proven malignancy treated with Coley’s vaccine,10 the five year survival for inoperable carcinomas (34%–73%) was similar to inoperable sarcomas (13%–79%), the range varying with tumour subtype.

In an effort to evaluate past successes using Coley’s regimen to that currently observed with modern conventional treatment, a retrospective study compared the 10 year survival rates of patients treated by either method using data from the Surveillance Epidemiology End Result cancer registry.24 Limitations of the study included sample sizes and staging of patients treated with Coley’s vaccine. Still, this study found that despite the billions of dollars spent to develop modern cancer treatments, patients receiving modern conventional therapies did not fare better than patients receiving the treatment initiated by Coley over 100 years ago. Where would cancer treatment be today if equivalent effort and funds had been used to develop a better understanding of the treatment pioneered by Coley? Some answers might be found by taking a closer look at the history of spontaneous regression.


Peregrine Laziosi (1265–1345), having been afflicted with cancer himself, was several centuries later canonised and named the patron saint of cancer patients. In the course of his untiring work preaching, converting and reconciling sinners, he noticed a large growth emerging on his leg. The growth on his tibia was pronounced unanimously by the best physicians of his time to be malignant.25 His only option was to have his leg amputated. The lesion grew to the point where it broke through the skin and became severely infected. In fact, it was stated “such a horrible stench was given off that it could be endured by no one sitting by him”.26 Miraculously, by the time he was due to have his operation, his physician was astonished to observe that there were no signs of the tumour. Saint Peregrine’s cancer never returned. Was this spontaneous regression an isolated event or did the infection play a part?

Once Coley’s interest in tumour regression was kindled, he found that spontaneous regression in association with acute infections was often mentioned in the historical literature2 and he was to record many more cases himself.15 Moreover, he discovered that many past physicians had used these infections to the advantage of their patients. Such coincidental infections had in fact inspired a wide variety of rudimentary cancer immunotherapies. The earliest example of such cancer immunotherapy may be thousands of years old. In the writings of the Ebers Papyrus (c 1550 BC), attributed to the great Egyptian physician Imhotep (c 2600 BC), the recommended treatment for tumours (swellings) was a poultice followed by incision.27 Such a regimen would inevitably lead to an infection at the tumour site. By the 1700 and 1800s AD, crude forms of cancer immunotherapy became widely known and accepted.8 For example, Tanchou in his comprehensive treatise on cancer published in 1844, provides insight on how these immunotherapies came into being:“One knows that often the affected lymph nodes and primary growths disappear during the course of concurrent illness, never to return. It is according to that idea ... that a large number of observers have advised establishing ‘issues’[suppurating sores] on diverse portions of the body and even in the wounds remaining after operation”.28 He goes on to cite many cases from other physicians where “issues” were successfully established. Other strategies from that era included applying septic dressings to ulcerated tumours,28,29 or deliberately introducing infections such as erysipelas,30,31 gangrene,28,32–34 or syphilis.35,36 Until Coley produced his killed vaccine, using live bacteria to initiate an infection was a precarious gamble between life and death.

The tide began to turn against “Nature’s remedy” for cancer during the 20th century. Firstly, cancer surgery, like any other operation, became a sterile procedure after acceptance of Lister’s aseptic techniques in the late 1800s.37 In fact, in a 1909 discussion paper on cancer treatment,38 one surgeon suggested that the postoperative infections that were common in the past improved survival and should be encouraged. Yet in this new era, his suggestion was harshly criticised as “a doctrine that would make surgery go backwards”. Secondly, by the time of Coley’s death in 1936, radiotherapy had become an established cancer treatment, and chemotherapy was rapidly gaining acceptance. These treatments could be more easily standardised than Coley’s approach and the hope that these therapies would eventually lead to a cure for cancer was high. Such therapies ran counter to immunotherapy, as they are highly immunosuppressive. Thirdly, following World War II, antibiotic use during and after surgery became commonplace. Thus, postsurgical infection rates were reduced even further, in addition to diminishing the severity and duration of those infections that did occur. Finally, once the immune system became “redundant” in fighting infections, antipyretics came into routine use to eliminate the discomforting symptoms of an immune response. Hence, reports of spontaneous regression have become less commonplace, although an association with acute infections is often noted when it occurs.39–50 In fact, a retrospective study by Ruckdeschel et al found that patients who developed empyema after lung cancer surgery had a significantly better five year survival (50%v 18%).51 Nature exists in a delicate balance, the immune system being no exception. Attempts to create an increasingly sterile environment may further reduce our innate cancer curing ability, until we may finally convince ourselves that it never existed at all.


Febrile immune response

As mentioned previously, Coley asserted that fever induction was a key aspect of his treatment. In fact, he observed that a strong febrile reaction was the symptom most associated with tumour regression. A retrospective study of patients with inoperable soft tissue sarcomas treated with Coley’s vaccine found a superior five year survival in patients whose fevers averaged 38–40°C, compared with those having little or no fever (<38°C) during treatment (60% v 20%).52

With the current widespread use of antibiotics to treat infections and antipyretics to “manage” symptoms of an infection, the critical part played by fever is often overlooked. In hospital settings, fever is frequently suppressed as a matter of routine.53,54 Many modern immunology texts make little mention of fever,55–57 and may disregard it as being “insignificant”55 or refer to it as a “mystery”.56 Is the febrile immune response inconsequential to the outcome of an infection?

Historically, fevers were not only considered beneficial, but were actively encouraged. For example, Native Americans were known to treat acute febrile diseases with sweat baths.58 In North Africa, it was observed that fevers were treated by sweating induced by hot sand or hot water baths.58 Similarly, the great medical historian Celsus (1st century AD) described how patients with febrile illnesses would be “well covered up to excite at the same time a violent heat and thirst”.58 Such practices, interestingly, would reinforce the immune response to infection as it reduces the considerable amount of energy expended on fever generation and maintenance. In the late 1800s, the meticulous studies of Carl Wunderlich on febrile illnesses (over one million observations) demonstrated the diagnostic and prognostic value of continued temperature observations.59 He also concluded that the intensity of fever was a reliable measure of disease severity, underscoring its importance in fighting disease. Recent studies support his observations,60–62 although the ability to mount a fever in response to infection is diminished in the elderly.63–65

During this past century, the 1927 Nobel Prize for medicine was awarded to Julius Wagner-Jauregg for devising an ingenious fever therapy for dementia paralytica (that is, neurosyphilis). Surprisingly, or maybe not so surprisingly, his adventitious discovery arose in an analogous manner to Coley’s. As a newly practising psychiatrist in 1883, he was to observe a woman patient who “spontaneously” recovered from a severe mental illness after an attack of erysipelas.66 Eager to learn more about this unexpected coincidence, he undertook a comprehensive and exploratory review of the literature.67 He discovered that the spontaneous remission of psychoses had been reported after a wide spectrum of febrile illnesses. Often remissions were temporary, although reported cures were not exceptional. In contrast to Coley who initially used live bacteria, Wagner-Jauregg began experimenting with derivatives of killed bacteria (tuberculin and later staphylococci) in treating neurosyphilis, but was dissatisfied with the reaction to these agents. However, he noted that the treatment was most effective when some febrile illness intervened.68 He then modified his treatment by injecting patients with tertian malaria (associated with recurrent fevers), which could be controlled to a certain degree with quinine and arsenicals. Due to the intricate nature of the treatment, cure rates were variable. Remissions to the extent that patients were able to return to work were in the range of 55%–65%.69 Gross mortality ranged from 1%–10%,69–71 with approximately half being due to neurosyphilis.70 Untreated, death generally occurred within several years of the diagnosis of dementia paralytica.72 Tens of thousands of patients were eventually treated in this manner before penicillin came into common use.73 On his efforts to use an infectious agent as treatment for neurosyphilis, Wagner-Jauregg asserted “We have listened to nature; we have attempted to imitate the method by which nature itself produces cures”.74

Fever is in fact a highly conserved physiological response to infectious stimuli. It is more than just a rise in body temperature and not analogous to hyperthermia (that is, mechanically achieved increase in temperature). Hyperthermia, increasingly being applied in combination with radiotherapy and chemotherapy,75 has been of limited use as it lacks the systemic effects obtained with Coley’s vaccine. In contrast to hyperthermia, fever is accompanied by diverse immunological changes; notably, biochemical reaction rates increase, and leucocyte proliferation, maturation, and activation is enhanced.76,77 Febrile thermogenesis (for example, chills, shivering, etc) is associated with an increase in the metabolic rate by 2–3 times, while fever maintenance has been associated with a 30%–50% increase in the metabolic rate.78,79 Thus, due to the substantial energy expenditure required for a febrile immune response (for example, increased heart rate, oxygen consumption, and metabolism), it is unlikely that such a response would be conserved unless it had considerable adaptive value.79 Even animals that cannot generate a fever (for example, fishes, amphibians, reptiles) exhibit heat seeking behaviour during an infection.80 Moreover, this heat seeking behaviour in infected animals corresponds to significantly improved survival over animals prevented from such behaviour.81 Similarly, in animals that can generate fevers during infection, such as mammals, antipyretic administration impairs pathogen clearance and reduces survival compared with untreated animals.82–89

Human studies conducted in this area provide similar support for the concept that fever enhances host defence against infections. For example, retrospective studies found that paralysis incidence and severity after polio infection was more severe in children who received antipyretics.90 Another study, in children with malaria, found that antipyretic drug therapy significantly impaired plasma clearance of the malarial parasite compared to untreated individuals.91 Studies of children with chickenpox92 and upper respiratory infections93 found that antipyretic use was associated with an increased duration and severity of illness. In studies where subjects were experimentally infected with rhinovirus, antipyretic use was associated with increased viral shedding,94 nasal signs and symptoms,95 and duration of disease,95 while antibody responses were reduced.95 In those subjects infected experimentally with influenza A and Shigella sonnei, antipyretic therapy was associated with a significantly increased duration of illness.96 Finally, several reports have suggested an association between the use of non-steroidal anti-inflammatory drugs and progression of invasive streptococcal infections, particularly necrotising fasciitis,97–99 and the occurrence of empyema after pneumonia in children.100

One unexpected observation by Coley was the salutary effect of fever on cancer pain.5 This beneficial property had been observed by others in association with infection-induced tumour regression.23,101–104 In fact, patients would often reduce or discontinue their use of narcotic pain medications while receiving treatment. This phenomenon appears to be independent of tumour regression, as it often occurred immediately after vaccine injection, preceding such regressions. Lagueux, after many years of experience using Coley’s vaccine, commented that “pain always disappeared after the first injections”.5 Actually, this remarkable analgesic effect has long been noted. The well known description of inflammation by Celsus is followed by a largely unappreciated observation on the benefits of fever:“Now the signs of an inflammation are four: redness and swelling with heat and pain ... if there is pain without inflammation, nothing is to be put on: for the actual fever at once will dissolve the pain”.105

Mechanisms of tumour suppression

Coley’s published papers on cancer regression associated with his mixed bacterial vaccine stimulated others to explore the underlying mechanisms of this phenomenon. Specifically, researchers strived to identify the “active” component of Coley’s vaccine.106,107 This also led to investigations to determine which host factors produced in response to the vaccine could induce tumour regression. Cytokines such as tumour necrosis factor (TNF), interleukins, and interferons were considered as possibilities.108–110 However, the answer is far more complex than ascribing the response to one or another factor. Any immune response to pathogens is associated with a multitude of cytokine cascades, which in turn triggers other cascades and a diversity of cellular responses. This immune cascade was readily evoked through the use of Coley’s crude bacterial vaccine, but difficult to reproduce with single cytokine therapy. For example, the most effective treatment for superficial bladder cancer,111 bacillus Calmette-Guerin (BCG), is presently the only conventional bacterial vaccine in use. Unlike Coley’s treatment, BCG is not administered with the intent of inducing a fever. Furthermore, the BCG vaccine contains a live attenuated strain of bacteria (Mycobacterium bovis), and thus, must necessarily be used more cautiously to avoid disseminated infections.112,113 Yet, similar to Coley’s approach, the vaccine is applied directly to the tumour site and repeated courses following initial therapy (as Coley recognised over 100 years ago)4 reduces recurrence.114 After intravesical administration of this vaccine, a wide range of cytokines become detectable in the urine including interleukin-(IL)-1, IL-2, IL-6, IL-8, IL-10, IL-12, IL-18, interferon-γ, interferon-γ inducible protein-10, macrophage colony stimulating factor, and TNF-α.115–121 Many more cytokines are upregulated, others downregulated, to varying degrees throughout the course of treatment; yet, this illustrates the point that individual immunomodulating cytokines are in fact only one small facet of this complex immunological response to infection, and correspondingly, tumour regression. Thus, it becomes clear why the effectiveness of BCG is limited to superficial bladder cancer. The heat and immune activation associated with local inflammation is analogous to a scaled down systemic febrile response, and correspondingly this local response is only effective in the immediate region where it occurs. To strike at invasive cancer, a systemic response is necessary, although such a strategy would undoubtedly be hazardous as BCG is a live vaccine.

Some interesting facets of spontaneous regression include both the broad diversity of organisms that have been observed in association with this phenomenon (for example, bacterial, fungal, viral, and protozoal pathogens) and the speed at which this reaction can occur.8,10,11,50,122 It has been proposed that the cell mediated (that is, type 1) immune response, rather than the humoral (that is, type 2) response, is a key mediator of cancer regression.123 Yet, many of the cases of spontaneous regression as well as tumour inhibition in animal studies involved infections that elicit a humoral immune response (for example, aspergillus,50 malaria,11 trichinella,124 trypanosoma125). Moreover, while tumour regression was often noted within hours of tumour injection with Coley’s vaccine,5 primary adaptive immune responses are often delayed by several days to a week.126 In fact, Coley’s experience5,15 and an exploratory evaluation of case reports of spontaneous regression8,10,11 support the concept that infection-stimulated tumour regression generally results from a “non-specific” innate immune response. In cases where the regression was partial and the acute or febrile phase of the infection subsided, residual tumour generally regrew.11 Similarly, if the infection recurs or is reintroduced, tumour regression may proceed as before.10,11 Coley stated that daily injections should be given, if the patient could bear it, as discontinuing the vaccine even for a few days would often lead to regrowth of residual tumour15—again suggesting that specific antitumour immunity was not a primary mechanism of this vaccine. An important mediator of the innate immune response is the Toll-like receptor family, primarily expressed on macrophages and dendritic cells.127 These receptors both guide the course of the innate defensive response, as well as shape the reparative response. It would appear that the antigenic complexity of Coley’s vaccine was inadvertently an important factor in its success as it triggered many Toll-like receptors essential for a defensive response.

Observations that a wide spectrum of pathogens can induce tumour regression suggest that there are some unifying characteristics of the innate immune response responsible for this phenomenon. The immune system has an important dual role in maintaining the integrity of the host. The immune system is primarily recognised for its role in defence against foreign pathogens; however, it plays an equally important part in tissue repair. During wound healing, leucocytes are actively engaged in matrix degradation, growth factor production, and the induction of new blood and lymphatic vessels.8,128–131 If the wound is sterile, cytotoxic defensive functions do not become activated.

A tumour, however, being partly “self” and partly “foreign”, can elicit a reparative growth-promoting response from intratumoural leucocytes.8 Based on the results of our previous research on human cancers,128,132,133 we devised a model of this duality in function, illustrating how the immune system may either enhance or inhibit tumour growth (fig 1). Like wounds, expanding tumours release chemokines and other cytokines that attract leucocytes and signal that increased oxygen and nutrients levels are required.8,134 In this way, an aberrant and detrimental reparative response is generated, where the immune system essentially supports tumour growth.128,135 Leucocytes, particularly macrophages, are present in large numbers in many rapidly growing tumours.8,128,132 Macrophages are versatile and resilient phagocytes capable of prolonged survival in the acidic wound environment.136 Moreover, macrophages contribute to the production, mobilisation, activation, and regulation of all immune cells.137 There is even evidence that monocyte/macrophages can differentiate into endothelial progenitor cells138,139 and fibroblasts.128,129 Interestingly, tumour-derived fibroblasts have been shown to stimulate tumour cells in vitro, an effect not observed with normal tissue fibroblasts.140 Thus, macrophages can play a pivotal part in tumour stroma formation. Moreover, macrophages are abundant in areas of tumour cell proliferation, where evidence of macrophage-induced tumour cell killing is rare or absent.136

Figure 1

The model illustrates the double edged nature of the immune system. When a tumour develops, the relative balance of these two arms determines its outcome. Tumour induces the reparative arm and thus subverts leucocyte growth-promoting activities to its own benefit. An exogenous antigenic stimulus such as Coley’s vaccine may shift the balance back to the defensive arm resulting in tumour regression. Dark grey arrows represent a detrimental and light grey arrows a beneficial immune response.

Coley observed more intense reactions in highly vascular tumours.15 In these tumours, rapid degeneration occurred, often with the formation of sloughs. He also noted that less vascular tumours, in contrast, more often regressed through slow absorption without breaking down or sloughing. The tortuous and fragile nature of tumour vasculature compared with ordinary vessels141 makes it more susceptible to febrile immunostimulated collapse, resulting in haemorrhagic necrosis of the dependent tumour mass. The simultaneous suspension of immune reparative functions also counters tumour growth (fig 1).8 For example, macrophages secrete a wide range of factors, some of which stimulate blood and lymphatic vessel development (for example, platelet-derived growth factor, vascular endothelial growth factor-A, -B, -C, -D).131,142,143 A shift to the defensive mode downregulates the production of these factors. A final key factor contributing to tumour regression involves the direct killing of tumour cells by macrophages (for example, production of reactive oxygen and nitrogen metabolites). The fact that macrophages express Toll-like receptors involved in both defence and repair underscores the delicate balance that exists between immune-mediated tumour growth and regression. Thus, the presence of hypoxia or necrosis in an otherwise sterile tumour can induce the release of factors that stimulate tumour growth; while the introduction of Coley’s vaccine, or other bacterial, viral, or fungal products, can shift the balance back towards a defensive immune response (fig 1).

There is often confusion in the literature as to what constitutes an immune response and what constitutes a side effect. For example, in a review on high dose interferon-α for the treatment of melanoma, it states that flu-like symptoms associated with this therapy “are quite manageable ... with prophylactic antipyretics ... to control fever, headache, and myalgia”.144 Thus, the immune system is stimulated on the one hand with interferon-α, even as it is suppressed on the other with antipyretics. Little consideration is given as to whether these flu-like symptoms can improve patient survival. Although severe adverse effects must be avoided, a failure to recognise aspects of the immune response that are critical to disease regression counters the effectiveness of the treatment. Moreover, single cytokine therapies may result in many unique toxicities due to the fact that such treatments present an unnatural challenge. Although Coley’s vaccine and treatment regimen was not free of adverse symptoms, it is crucial to understand that the symptoms arising from this form of treatment (for example, chills, fever, fatigue, etc) are normal adaptive responses to immunostimulation and facilitate disease regression. Furthermore, the beneficial effect of this regimen on cancer pain would have dual advantages—suppression of pain also leads to reduced usage of agents that inhibit key aspects of the immune response, such as antipyretics145 and opioids.146,147

A final point concerns the paradoxical influence of acute and chronic infections on tumour formation. It is now well established that some malignancies arise in association with chronic infections of one type or another. Helicobacter pylori and gastric cancer, Schistosoma haematobium and bladder cancer, and human papilloma virus and cervical cancer are some examples. These infectious diseases generally afflict the organ where the cancer later develops. However, unlike the acute febrile response, a chronic infection generally represents a failed immune response to disease, and many mechanisms have been uncovered to explain the infectious role in tumour promotion.148 Yet, even chronic infections may have temporary periods of benefit as during an acute flare-up or concurrent illness.8,149

Questions (true/false; answers below)

  1. Coley was the first to use live bacteria as an immunotherapy for cancer.

  2. Coley’s vaccine was only effective when directly injected into the primary tumour.

  3. Fever generation increases the metabolic rate to a greater degree than fever maintenance.

  4. An infection in a cold blooded animal induces a “fever” response.

  5. Tumour necrosis factor-α was sole factor responsible for the effects of Coley’s vaccine.

  6. Immune cells are important for tissue repair during wound healing.

  7. Macrophages can induce the formation of both blood and lymphatic vessels.

  8. A non-specific immune response to the cancer can play a major part in tumour regression.

  9. Chronic infections may be associated with the spontaneous regression of cancer.

  10. Coley’s vaccine was more effective for the treatment of sarcomas than carcinomas.


  1. False. There are many examples that precede Coley’s work.

  2. False. Numerous case reports have been published where metastatic lesions regressed without direct injection.

  3. True. By analogy, acceleration of an automobile uses more fuel than maintaining its speed at a higher velocity.

  4. True. Experimental studies demonstrate that cold blooded animals will preferably move to a warmer microclimate after infection to increase their body temperature. Antipyretics suppress this behaviour.

  5. False. As there is not one dietary component responsible for proper nutrition, nor is one cytokine responsible for the effects of Coley’s vaccine.

  6. True. In vivo and in vitro studies have found that immune cells are capable of phagocytising debris, producing cellular growth factors, and stimulating the formation of new blood and lymphatic vessels.

  7. True. Macrophages have been shown to produce both blood and lymphatic-specific growth factors.

  8. True. Although specific immunity may be involved in tumour regression, there is evidence from many sources that the non-specific immune response can cause cancer regression.

  9. True. Cases of spontaneous cancer regression in patients with chronic infections have been reported; generally, in the initial acute phase of the infection, during an acute flare-up, or during a concurrent illness.

  10. False. Although more sarcomas were treated than carcinomas, evidence suggests that survival rates were similar. In fact, bone sarcomas generally showed the poorest survival.


The ability of the immune system to provoke “spontaneous” regressions has been reduced by use of immunosuppressive therapies such as chemotherapy, radiation and antipyretics, as well as the reduction of infections by sterile surgical techniques and antibiotics. Retrospective studies in lung cancer patients,50,150,151 however, suggest that there is a significant benefit when such accidental infections occur. More comprehensive studies in cancers where significant postoperative infections are not uncommon (for example, bladder, lung, colorectal) are required.

After being used for almost 70 years and despite hundreds of publications on its effects on cancer, Coley’s vaccine was assigned “new drug” status in 1963 by the US Food and Drug Administration. This ruling forces anyone interested in this area to go through a considerably expensive series of testing protocols that essentially bar Coley’s treatment from being used on cancer patients. After this date, the opportunity to develop an inexpensive and effective vaccine for the treatment of a wide spectrum of cancers was virtually lost.

One way of testing Coley’s regimen would be to apply it in spontaneously occurring animal tumours. With recruitment from veterinary practices and in consultation with owners, seemingly hopeless cases could be sought to demonstrate the utility of this approach against metastatic disease. Alternatively, a pilot phase I study in terminal cancer patients may be an option. Key points that Coley considered essential for tumour regression could be investigated: inducing fevers, direct tumour injections, and repeated injections; at the same time, the adverse effects of the vaccine could be quantified. Contributing factors that support the induction and maintenance of fever, which appears to be an essential component for treating metastatic cancer, could be further analysed. In addition, treatment effects could be monitored through the measurement of serum tumour markers, cytokine expression, and immune cell number and activity. Once the principles of Coley’s treatment have been re-established and put into perspective in relation to the large amount of human data that are already available, more comprehensive human studies could be undertaken.

A revival of interest and a closer look at the scientific merits of Coley’s treatment may seem to some like a doctrine that would make cancer research go backwards; yet judging from the slow progress of conventional cancer treatments, a step backward may indeed be a giant leap forward.


SAHC is supported by a postdoctoral fellowship from the Canadian Institutes of Health Research.