VIRUSES AND VACCINES

Which cancers in humans are caused by a virus? In the late 1950s, asking this question was fully justified, given the evidence on cancers induced by viruses in several animal species. The first evidence came from Peyton Rous’s pioneering experiments with sarcoma in chickens in 1910 (recognized with the Nobel Prize 56 years later). Since the 1930s, results had followed in rabbits, mice, frogs, ducklings, turkeys, and guinea fowl. Among other findings, these studies showed the possibility of transmission of carcinogenic viruses between species. The first answer to the question of virus-induced cancers in humans came from Africa.

BURKITT LYMPHOMA

The first report of the malignancy that became known as Burkitt lymphoma was published in 1958 (see “The discovery of Burkitt lymphoma in East Africa”). The search for possible causes soon began, and led in 1964 to the identification of a new virus of the herpes family, named Epstein–Barr virus (EBV) after its discoverers. EBV is a DNA virus with versatile properties: it was found to silently infect people from an early age in most countries; it turned out to be the agent of infectious mononucleosis, which is most common among adolescents and young adults; and in the laboratory it became a tool for “immortalizing” lymphoid cell lines, making them continue to grow indefinitely. IARC participated in research on Burkitt lymphoma from its earliest days. It contributed to the standardization of methods to measure anti-EBV antibodies and directed serological surveys to detect the presence of EBV in East Africa. In 1972, IARC initiated a large-scale prospective study in five counties of the West Nile District of Uganda (see “IARC’s prospective study in the West Nile District of Uganda”).

I was studying for my thesis in 1978–1979, and Guy de Thé and Gilbert Lenoir were working on Burkitt lymphoma, just during the period when translocations were identified and the mechanisms started to be understood. We met Denis Burkitt several times at the Agency. There was a very, very strong collaboration. – Thierry Philip, long-term IARC collaborator

The title of the second volume of the IARC Scientific Publications series, Oncogenesis and Herpesviruses, published in 1972, documents IARC’s interest in the wider field of cancer and herpesviruses. The possibility that these agents, and in particular EBV, could be involved in the development of cancers other than Burkitt lymphoma was suggested by findings of elevated concentrations of anti-EBV antibodies in cases of nasopharyngeal cancer, a tumour frequent in populations of Chinese origin.

Evidence for a major role of EBV in Burkitt lymphoma in equatorial Africa emerged from the West Nile investigation in Uganda. In the meantime, Burkitt lymphoma was reported sporadically from continents other than Africa, but in those cases EBV was found in the tumour cells much less frequently. As was highlighted already in 1985 by the numerous papers in IARC Scientific Publication No. 60, Burkitt’s Lymphoma: A Human Cancer Model, EBV acts in cooperation with other co-factors: probably impaired immunity due to malaria infection in cases from equatorial regions and multiple other co-factors in cases from other areas, as well as in nasopharyngeal cancers.

The vertical axis shows the percentage of subjects with different levels of antibodies against the Epstein–Barr virus (EBV), indicated by the numbers on the horizontal axis. High concentrations occur in a larger percentage of subjects with Burkitt lymphoma (BL) or nasopharyngeal carcinoma (NPC) than in subjects with other cancers (OT). GMT, geometric mean titre.

Images

IARC laboratories have continued to study the mechanisms of action of EBV and its interaction with co-factors. Recent contributions have shown the relevance of tobacco smoking and of some genetic variants in nasopharyngeal cancer. In addition, some recent results point to a possible interaction between EBV and aflatoxin, a known carcinogen endemic in the regions where Burkitt lymphoma is commonly found in Africa (see the chapter “Carcinogens in the human environment”). The Working Group responsible for Volume 100 of the IARC Monographs evaluated as sufficient the evidence for the carcinogenicity of EBV for several tumours, including Burkitt lymphoma, Hodgkin lymphoma, non-Hodgkin lymphoma, and nasopharyngeal cancer. Malaria as a co-factor in the etiology of Burkitt lymphoma had been investigated in an IARC study in the United Republic of Tanzania in the late 1970s. In 2012, the Working Group for Volume 104 of the IARC Monographs categorized malaria caused by infection with Plasmodium falciparum, the parasite present in highly endemic areas, as probably carcinogenic to humans.

The initial focus of IARC scientists on EBV also prompted research on the most frequently occurring tumour in women in developing countries: cancer of the uterine cervix. This tumour was a rather obvious candidate for having an infectious cause. Several studies had documented its association with the number of sexual partners of a woman or her husband. As Richard Doll and Richard Peto wrote in 1981, “The present evidence strongly suggests that one of the primary causes of the disease is an agent passed between partners in intercourse, quite possibly a virus.” At IARC and in other laboratories, tests for EBV were conducted on serum samples of patients and on cervical cancer cells, with disappointing results. These investigations pointed to a need for an expanded search for relevant infectious agents, starting with other herpesviruses already known to be sexually transmissible.

The map shows a large global variation in the incidence of cervical cancer in 2012. ASR (world) is the age-standardized rate (new cases per 100 000 women per year), standardized for age using the world standard population.

Images

HUMAN PAPILLOMAVIRUSES AND CANCER, ACT ONE

The rhythms of research on cervical cancer at IARC have been marked by two interwoven developments: field epidemiological studies, periodically interspersed with syntheses of the accumulating evidence, including from IARC, on causes of this cancer. A 1989 IARC Scientific Publication already took stock of the changing direction in research, from fruitless efforts on herpesviruses to promising explorations of the role of human papillomaviruses (HPV) (see “1989: Cervical cancer and infection – growing evidence amid much uncertainty”). In several laboratories, considerable advances were made in the understanding of the molecular mechanisms by which certain types of HPV could transform normal cells into cancerous cells. However, the epidemiology linking HPV to cervical cancer lagged behind.

It is at this juncture that IARC played a key role, mainly through a case–control study involving cervical cancer cases and randomly selected population controls in Colombia and Spain. The frequency of cervical cancer in Colombia was about 8 times that in Spain. Cells from cervical-swab specimens were tested for the presence of HPV DNA. In both countries, the percentage of positivity for HPV was much higher in cells from cases of invasive (i.e. advanced) cancer than in cells from control subjects, and this result was independent of whether the women were sexually active.

Human papillomavirus (HPV) DNA was measured with polymerase chain reaction (PCR) in cervical cells. The percentage of women testing positive for the presence of HPV DNA was much higher among cervical cancer cases (black columns) than among controls (grey columns), regardless of the time since last sexual intercourse.

Images

A second IARC study was conducted, also in Colombia and Spain, with less advanced cases of cervical cancer (carcinoma in situ). Again, a marked excess of HPV positivity was found among cases compared with controls. The strong association between HPV and cervical cancer, both at an early stage (carcinoma in situ) and at a later stage (invasive cancer), indicated that HPV infection precedes the full development of the tumour, strengthening the evidence for a causal rather than a “passenger” role of the virus.

Six years after the 1989 IARC publication, the prevailing views on HPV and cervical cancer had changed radically. The Working Group for Volume 64 of the IARC Monographs, Human Papillomaviruses, concluded that there is sufficient evidence for the carcinogenicity of HPV types 16 and 18, the subtypes that most frequently infect cervical cells. In 2008, Harald zur Hausen, whose laboratory in Heidelberg produced key experimental evidence on the carcinogenicity of HPV, received the Nobel Prize “for his discovery of human papillomaviruses causing cervical cancer.” The crucial epidemiological demonstration that some HPV types indeed cause cervical cancer in humans earned Nubia Muñoz, the lead scientist of the IARC studies, several prestigious recognitions around the world.

Nubia Muñoz received the Gairdner Foundation’s Canada Gairdner Global Health Award in 2009. The award was conferred “for her epidemiological studies that defined the essential role of the human papillomavirus in the etiology of cervical cancer on a global level, which led to the development of successful prophylactic vaccines.” Seen at the awards ceremony are, from left to right, John Dirks, president and scientific director of the Gairdner Foundation, the Ontario Minister of Health, Nubia Muñoz, and the deputy director of the Gairdner Foundation.

Images

This diagram illustrates the progression from normal tissue of the uterine cervix to carcinoma in situ, the earliest stage of cervical cancer, via a series of steps dependent on the persistence of human papillomavirus (HPV). From top to bottom are shown the microscopic appearance of the tissue, the appearance of the exfoliated cells used for the Pap smear screening test, a schematic representation of the progressive changes in the tissue, and the terminology for the stages of progression (CIN, cervical intraepithelial neoplasia; SIL, squamous intraepithelial lesion).

Images

HUMAN PAPILLOMAVIRUSES AND CANCER, ACT TWO

Establishing that HPV causes cervical cancer opened up entirely new perspectives in research, aimed at preventing the occurrence of the cancer or controlling it through early diagnosis and treatment. The scientific basis for the production of preventive vaccines acting against virus-like particles of HPV had been established during the 1990s, and vaccines approved for use in human populations were made available by the pharmaceutical industry starting in 2006. However, deciding who should be given the vaccine, and when, demanded a thorough knowledge of the natural history of HPV infection. Most sexually active individuals of both sexes acquire HPV infection at some time during their life, but more than 90% of new infections regress over 6–18 months. In the remaining 10% of cases, infection persists, and in some women cells may progress to precursor lesions and ultimately to invasive cervical cancer. To sustain this evolution, other factors (viral, host, or environmental) must play a role. Tobacco smoking is the best-established co-factor, as a result of several epidemiological studies from IARC and other research groups.

Which population should be eligible for vaccination? A study by the International Collaboration of Epidemiological Studies of Cervical Cancer was coordinated by IARC and published in 2012. The results showed that although women can be infected by carcinogenic HPV at any age (with most infections occurring soon after first intercourse), the risk of cervical cancer arising from a new infection falls sharply with age and is very low after about age 40, an indication that vaccination efforts should focus on young people.

Another recent collaborative study, initiated by the United States National Cancer Institute, was conducted in Costa Rica. Healthy women aged 18–25 years were randomized to receive vaccination against HPV types 16 and 18 or vaccination against hepatitis A virus. After 4 years of observation, the HPV vaccine showed a high efficacy in preventing the occurrence of high-grade precursor lesions. A third IARC study used a mathematical model of the transmission of HPV 16 and 18 to investigate the impact of different vaccination options in high-income countries. It concluded that maximizing vaccination coverage of girls is currently the most effective option for decreasing HPV infection of the cervix, rather than aiming at vaccination of boys and girls.

The study in Costa Rica and an ongoing IARC-coordinated randomized trial of HPV vaccine in rural India have shown that two doses of the vaccine are as good as the standard three-dose schedule in providing immunity and preventing infection of the cervix with HPV. This important observation will make implementation of HPV vaccination more affordable and has led to World Health Organization (WHO) support for a two-dose schedule. As the HPV vaccine becomes more affordable, more countries are assessing the potential benefits with regard to cervical cancer prevention for their own populations. One of the pieces of evidence needed for evidence-based policy-making is the prevalence of HPV subtypes in cervical tumours from patients in the country concerned. IARC has helped inform such considerations with its HPV prevalence survey, which uses a standardized protocol and analytical methodology applied to different populations worldwide to provide such information. To date, results are available for 27 populations across the world. The frequency of HPV infection varies widely, from 3% in Spain to 15% in Colombia and to the very high prevalence of more than 50% in Guinea.

IARC’s international surveys of prevalence of human papillomavirus (HPV) infection were carried out from 1995 to 2013 in sexually active women aged 15–59 years. N denotes the number of women tested.

Images

Different types of HPV vaccines are commercially available or in preparation. The rapidly increasing availability of these vaccines has not obviated the need for early diagnosis of cervical cancer in unvaccinated women. The Pap smear, the well-established screening method for cervical cancer, may now be preceded or superseded in many countries by testing for the presence of HPV DNA, thus improving the overall quality of a screening programme (see the chapter “Cancer screening and early diagnosis”).

In principle, HPV subtypes other than types 16 and 18 could also be carcinogenic, and organs other than the cervix could also be affected. In 2009, the Working Group for Volume 100B of the IARC Monographs reviewed all the available evidence. As the figure shows, several other subtypes in the alpha species, to which the carcinogenic subtypes HPV 16 and 18 belong, also cause cancer, and in particular cervical cancer. HPV 16 was concluded to cause not only cervical cancer but also cancers of other genital organs (vulva, vagina, penis) and of the anus, oral cavity, oropharynx, and tonsil. A recent IARC study observed wide variations in the proportion of oropharyngeal cancers associated with HPV infection in different populations worldwide. This finding will be important in assessing the future benefits of vaccination on cancers at sites other than the cervix.

This phylogenetic tree of 100 human papillomaviruses shows how the types are genetically interrelated. The types in the alpha species for which the best evidence of carcinogenicity exists are shown in orange.

Images

IARC scientists are currently studying another family of papillomaviruses, the beta subtypes. The beta species may play a role in skin cancer, where there are, once again, possible interactions with environmental co-factors, such as ultraviolet radiation. Comparisons of the similarities and differences with the alpha subtypes assist in understanding how these and other viruses trigger the development of cancer.

Overall, it is estimated that worldwide, more than half a million new cancer cases per year are attributable to infection with HPV, about the same as the number of liver cancer cases caused by infection with hepatitis B and C viruses. Therefore, HPV infection has been proven to be an important cause of cancer. IARC has been part of the remarkable success story in translating that knowledge into the potential to prevent a substantial proportion of those cases.

Number of new cancer cases per year attributable to infections, subdivided by infectious agent and degree of development of world regions.

Images

LIVER CANCER

The spark of Burkitt’s keen clinical observations on lymphoma, followed by the discovery of EBV, was the first time that there was evidence for the carcinogenicity of a virus in animals and humans. For liver cancer in humans, the identification of causal factors took much longer. The first volume of the IARC Scientific Publications series, Liver Cancer, contains the proceedings of a conference held in London in July 1969. The proceedings document the mixture of well-established facts, dubious or fragmentary findings, and mere conjectures that prevailed at the time (see “Liver cancer as seen in 1969”). From the outset, IARC started to take advantage of what was already known to probe the etiology of liver cancer. Several research projects were developed on aflatoxins, experimentally recognized as carcinogens, to investigate their possible role in human cancer (see the chapter “Carcinogens in the human environment”).

As soon as reasonably reliable markers of hepatitis viruses, particularly of hepatitis B virus (HBV) infection, became available, IARC epidemiologists led or participated in case–control studies of liver cancer. The results showed a clear association with the HBV markers, and pointed to the role of tobacco smoke as a subsidiary etiological factor. Indeed, by the early 1980s the accumulating evidence from both case–control and cohort studies had shown that the association between HBV infection and liver cancer was strong, specific, and consistent, but was restricted to chronically persistent forms of HBV infection. Most epidemiologists accepted the association as being causal, although substantial proportions of liver cancer did not exhibit the HBV markers.

The idea rapidly gained ground that a study of subjects vaccinated against HBV would be valuable for several reasons: it would prevent hepatitis B, it would be the acid test of whether HBV was a cause of liver cancer, and – should this prove to be the case – it would prevent the occurrence of this tumour in the population under investigation.

There was not (and still is not) a vaccine against EBV. Hence, the issue at stake was to test the first anticancer vaccine. – Christian Trépo, long-term IARC collaborator

HBV is one of the most common infectious viruses in the world, and vaccines of different types had been available against HBV since 1969. To test the vaccine’s effectiveness in preventing liver cancer, a population was required in which HBV infection rates were high and liver cancer was common. Of the countries in sub-Saharan Africa with these characteristics, The Gambia presented a suitable option. In the mid-1980s, The Gambia had a population of a little more than 1 million people and a reasonable infrastructure. IARC established a multiparty research collaboration – the Gambia Hepatitis Intervention Study – in which it played the leading scientific role, in collaboration with the government of The Gambia, the United Kingdom Medical Research Council (which had a long-established research unit in Fajara, close to the capital city of Banjul), and the Italian government, which provided considerable funds.

The frequency of infection with the hepatitis B virus (HBV), as measured by the presence of the Australia antigen (HBV surface antigen), shows wide variation throughout the world.

Images

A map showing the different regions of The Gambia in West Africa on the two banks of the Gambia River, which flows into the Atlantic Ocean at the Kanifing District. The Gambia Hepatitis Intervention Study covered the whole country, whereas the recently initiated Prevention of Liver Fibrosis and Cancer in Africa (PROLIFICA) study is taking place in the Western Region.

Images

It was known that in The Gambia the highly endemic HBV infection occurs by transmission during childhood, for example in the family or at school. Therefore, a vaccination programme was started in 1986, with a schedule of injections at birth and at the ages of 2, 4, and 9 months. The anti-HBV vaccine was incorporated into the existing Expanded Programme on Immunization recommended by WHO. The original design of the study – a “stepped-wedge” trial – was both scientifically and ethically sound (see the chapter “Innovation in statistical methods”). At the end of the fourth year of the vaccination campaign, two comparable groups, vaccinated and unvaccinated, of more than 60 000 children each had been recruited into the study. A series of logistic problems had to be overcome, including those arising from the staff (at the full complement, more than 80 people) belonging to three different organizations: IARC, the United Kingdom Medical Research Council, and the government of The Gambia. A clear indication of success is the 93% completeness of the vaccination coverage of the targeted population of children.

To follow up the study participants and record the liver cancer cases, the Gambia National Cancer Registry was established in 1986. It has been estimated that results adequate to measure the protective effectiveness of infant vaccination on adult liver cancer will emerge from the comparison of vaccinated and unvaccinated children 30–35 years after immunization, before 2020. Several valuable intermediate results have already been gathered, the most important of which is the 94% protection that the vaccine shows against chronic persistence of the virus even 20 years after vaccination, indicating that there is no need for a booster dose in adolescence.

Assessing the results of the Gambia Hepatitis Intervention Study involves linking the cases of liver cancer (HCC), recorded by the Gambia National Cancer Registry, with the vaccination database containing the listed items. Vaccinated people will be recognizable by the position of the scar from the antituberculosis bacille Calmette–Guérin (BCG) vaccine – given on the left forearm to people vaccinated against hepatitis B virus and on the right forearm to the unvaccinated – and by comparison of their current foot and palm prints with those taken in childhood.

Images

Studies showing combined effects of aflatoxin exposure and HBV carrier status have been informative in illustrating the potential benefit of interventions against aflatoxins, particularly for those individuals already chronically infected with HBV and therefore out of reach of the protection afforded by the vaccine. In the same vein, a new study has been started in The Gambia and Senegal. The Prevention of Liver Fibrosis and Cancer in Africa (PROLIFICA) study aims to assess whether antiviral therapy can reduce the incidence of liver cancer in West Africa by identifying and treating chronic HBV carriers before their disease progresses to malignancy.

WHO has just reviewed the long-term protection from hepatitis B vaccine, and produced a statement on global policy for whether or not a boost was required, and the only first-rate evidence they had was from the Gambia study. So it is still highly influential in terms of global policy-making. – Andrew Hall, former IARC scientist

LEARNING FROM VIRUSES

For research on cancer biology, viruses have proven to be an inexhaustible source of information on mechanisms of carcinogenesis. For instance, the first oncogenes (crucial genes that, when activated, can trigger malignant transformation of the cell) to be discovered were found to result from sequences of viral DNA inserted into the human genome by viruses over the course of evolution.

In the identification of causes of cancer, the discovery of carcinogenic viruses has been the most important breakthrough during the past 50 years, since IARC was established. This discovery was relevant for both science and health. Virally induced cancers represent a substantial burden, particularly in developing countries, and they can be controlled with preventive measures. According to IARC’s most recent estimates of the global burden of cancer associated with infections, about 11% of new cancer cases worldwide are attributable to viral infections, and 16% to infections as a whole; in sub-Saharan Africa, one third of all cancers are infection-related. From its earliest years, IARC has made substantial contributions to the advances aimed at reducing this burden.

For IARC, the successes of its research on viral carcinogenesis and on the control of virus-induced tumours illustrate the value of combining three approaches, which reinforce each other. First, field studies that collect original data with the direct involvement of IARC epidemiologists, support from IARC laboratories, and a large international multidisciplinary collaboration facilitated by IARC’s standing as a research organization within the WHO framework. Second, periodic authoritative reviews of the evidence pertinent to specific issues of importance in relation to causation and prevention – as have been done for the carcinogenicity of EBV, HBV, hepatitis C virus, and HPV – carried out by international experts under the auspices of IARC as an international agency. Third, an investment in the continuity of long-term research programmes that, although they evolve and adapt over the decades, remain fixed in their commitment to the original objective (see “IARC’s long-term research projects in developing countries”).