1Page 1 of 58

Page 1of 58

Evaluation of the Potential Oncogenicity

ofRadiofrequency Fields in Experimental Animal Models

David L. McCormick, Ph.D., D.A.B.T.

IIT Research Institute

Chicago, Illinois 60616 USA

Abstract

It has become increasingly clear that a comprehensive evaluation of the possible health risks of human exposure to radiofrequency radiation (RFR)should be based on a “weight of the evidence” approach that integrates data from epidemiology studies in human populations, hazard identification studies performed in predictive animal models, and mechanistic studiesconducted using in vitro and in vivo model systems. Assessment of the potential carcinogenicity of RFR in animal models is a critical component of this overall evaluation.

The vast majority of carcinogenicity studies of RFR in animal models have failed to identify significant increases in cancer response. Most importantly, the results of seven chronic (two-year)oncogenicity bioassays of RFR in rats and mice were negative. Ten carcinogenicity studies of RFR in tumor-prone mouse models for lymphoma, mammary cancer, and brain cancer were also negative, as were over twenty initiation-promotion and co-carcinogenesis studies performed in animal models for cancers of the brain, breast, skin, liver, and colon.

By contrast, three bioassays of RFR may provide signals of possible carcinogenic activity. The results of the largest and arguably most sensitive two-year bioassay of RFR in rats demonstrated increased incidences of proliferative lesions in the brain and heart of male rats (but not female rats or either sex of mice); these increases may have been at least partially the result of improved survival in male rats versus male sham controls, but persisted after mortality adjustment of study data.

Increased incidences of liver and lung tumors have been reported in studies performed in two independent laboratories using the same multi-stage carcinogenesis model. Although the model used in these studies does not have an extensive history of use is not validated for hazard identification, these data provide a second signal that may suggest possible positive carcinogenic activity of RFR.

No conclusive evidence of RFR carcinogenicity has emerged from either epidemiology or experimental studies. Although the vast majority of carcinogenicity studies of RFR in animal models have failed to identify any significant risks of RFR exposure, the positive signals seen in three studies suggest that reasonable attempts to minimize RFR exposure may be prudent.

  1. Introduction
  2. Possible Health Effects of Radiofrequency Radiation from Mobile Telephones and Other Wireless Communications Devices

Wireless communications devices are ubiquitous components of the modern lifestyle in both industrialized countries and less developed areas of the world. Although mobile telephones and other wireless devices now seem essential for everyday life, their history of use is relatively short. In 1983, the Motorola DynaTAC 8000Xwasthe first mobile phone approved by the Federal Communications Commission for use in the United States [1]. Growth in mobile phone usage since that time has been extraordinary: whereas the number of U.S. cell phone subscriptions in 1990 was equal to only 2% of the American population [2], the number of U.S. cell phone subscriptions in 2017 [3] approaches75% of the American population.

Calendar year 2017 data from the GSM Association, an industry trade group, indicates that more than 2/3 of the world’s population currently has access to a mobile communicationsdevice, and more than five billion subscriptions to wireless services are now active [4]. Primarily as a result of the continuing rapid growth in wireless device use in India and other parts of Asia, the GSM Association predicts that the number of wireless subscriptions worldwide will reach 5.7 billion by 2020 [4].

Because wireless devices communicate through the use of radiofrequency radiation (RFR), a consequence of the nearly universal use of cellular telephones and other wireless devices is that billions of people worldwide receive daily exposure to RFR. Over the past two decades, a substantial number of laboratory and epidemiology studies have been performed to investigate whether exposure to RFR may lead to adverse health outcomes. Both positive and negative findings have been reported from studies in laboratory animals and in humans [reviewed in 5,6,7], but no consistent pattern of RFR health effects has been identified. As a result, no scientific consensus has emerged concerning the possible risks of human exposure to RFR. However, in consideration of the extremely large number of people who are regularly exposed to RFR, even a small increase in the risk of cancer or other disease that may result from the use of cellular telephones or other wireless communications devices could have important public health implications.

A substantial body of literature published over more than 50 years clearly demonstrates that exposure to microwaves or RFR at high field strengths (power levels that are substantially higher than those emitted by wireless communications devices) can induce tissue heating. As a consequence of this tissue heating, exposure to high levels of RFR (as generated by radar equipment, diathermy machines, unshielded microwave devices, and other equipment)may induce temperature-dependent changes in several tissues. Adverse health effects of tissue heating by RFR in animal models includes induction of cataracts [8,9,10] and adverse reproductive outcomes [11,12].

Although the use of mobile phones results in the local deposition of RF energy in the brain and other tissuesin the head [13,14,15], flux densities of RF fields generated by wireless communications devices are well below levels that will induce measurable tissue heating. For this reason, identification of possible adverse effects of exposure to non-thermal levels of RFR has become the central focus of health effects research related to the use of wireless communications devices. The key question to be resolved in RFR health effects research is “does exposure to RFR generated by cellular telephones and other wireless devices induce non-thermal effects that will induce or exacerbate disease?”

1.2.Strategies to Identify Possible Health Effects of Radiofrequency Radiation from Mobile Telephones and Other Wireless Communications Devices

It has become increasingly clear that a comprehensive evaluation of the possible health risks of human exposure to RFR should be based on a “weight of the evidence” approach that integrates data from (a) epidemiology studies in human populations; (b) hazard identification studies in predictive animal models; and (c) mechanistic studies in relevant in vivo and in vitro test systems [16]. At the present time, neither epidemiology alone, experimental bioassays alone, nor mechanistic studies alone are sufficiently informative to support a broadassessment of RFR health effects.

Although epidemiology offersthe obvious advantage of examining health effects in humansreceiving “real world” exposures, RFR epidemiologyhasseveral important limitations. In addition to the strengths and limitations of all epidemiology studies [17], key issues in RFR epidemiology include:

  • the unknown duration of RFR exposure that is required to induce an adverse health effect. Because broad public use of wireless communications devices has a relatively brief history [2], the duration of human exposure to RFR may be too short for epidemiology to identify chronic toxicities or other adverse effects (including some malignancies) with long latent periods. For this reason, definitive epidemiology data for the possible relationship between RFR exposure and some cancers may not be generated for twenty years or more. In consideration of this possibility, important unanswered questions include: (a) what duration of RFR exposure is necessary to induce an adverse health effect, and (b) when should monitoring of exposed populations be initiated to identify such effects?
  • challenges in the quantitation of actual RFR exposuresfor users of mobile telephones and other wireless communications devices. Although a number of approaches to exposure assessment have been used in RFR epidemiology studies [18,19,20], accurate assessment of human exposure to RFR generated by mobile telephones remains a major challenge. Clearly, reliable quantitation of RFR exposure is essential to support epidemiologic findings of adverse health effects [18].
  • the need for adverse health outcomes to have occurred prior to hazard identification, thus delaying the possible identification of a true human hazard for years, if not decades. This is considered to be perhaps the most important limitation of RFR epidemiology. Given the potentially long latency of RFR-induced health effects, identification of such effects through epidemiology alone may require years (or decades) of exposure. Should epidemiology identify significant adverse health effects of RFR exposure, billions of people will have already been exposed to RFR for extended periods and will therefore be at risk of those health effects. Given the truly massive population exposure to RFR from wireless devices, avoidance of a potentially major public health crisis mandates that risks be identified more quickly than can be accomplished through epidemiology alone.

Studies in predictive animal models may identify potential human health hazards years (or decades) earlier than can epidemiology. Thisshorter time required to identify possible healthhazards of RFR increases in importance when those hazards may include neoplasms with long latent periods. Importantly, a large body of evidence has developed to support the predictive power of well-designed animal bioasssays, particularly the two-year oncogenicity bioassay in rodents, to identify human carcinogens [reviewed in 21,22]. Well-designed and conducted animal bioassays also offer the opportunity to evaluate possible health effects under tightly controlled exposure conditions that support the characterization of dose-response (or exposure-response) relationships, and reduce or eliminate factors that could confound or otherwise impact results. That said, however, the need to extrapolate animal bioassay data from rodents to humans, and the common requirement to extrapolate effects of high dose exposures in rodents to possible effects of much lower exposure levels in humans greatly complicate the interpretation of animal bioassay data [23].

Data from mechanistic studies performed using in vivo or in vitro model systems may identify cellular, biochemical, or molecular mechanisms that underlie effects of RFR identified in epidemiology studies or animal bioassays. It should be noted, however, that identification of a mechanism of action is not essential to identify an agent as being hazardous or possibly hazardous to humans. Furthermore, although demonstration of specific cellular, biochemical, or molecular effects of RFR that may be relevant to the induction of cancer or other diseases is clearly relevant to hazard identification, such data arenot, in themselves, sufficient to identify a hazard. Because most critical physiological processes are regulated by redundant mechanisms, identification of an effect on a single mechanism may or may not be biologically significant for the organism. For this reason, findings of effects on disease-related mechanisms are most often considered to be secondary data that can be used to understand, support, and interpret the results of epidemiology studies or animal bioassays. Without evidence of a hazard identified by epidemiology studies or bioassays in experimental animal models, mechanistic data alone cannot be interpreted as definitive evidence that a health hazard exists.

Based on the complementary strengths and limitations of epidemiology studies and animal bioassays, and the secondary role played by mechanistic data in hazard identification, it is clear that integration of data from all three types of investigations provides the most comprehensive approach to identifying possible health hazards that may result from human exposure to RFR. In situations where the epidemiology data are incomplete, inconclusive, or conflicting, the importance of animal bioassays increases. In this chapter, the results of animal bioassays of RFR are reviewed and interpreted in the context of developing a comprehensive assessment of the possible health effects of RFR exposure.

1.3.Investigative Studies of RFR Exposure in Laboratory Animal Models

The scientific literature contains asubstantial body of research addressing the toxicity and potential oncogenicity of RFR in animal models. For the purposes of this review, toxicology and carcinogenesis bioassays of RFR in animal models are summarized in three sections of this chapter.

  • In Section2, chronic studies to investigate the possible oncogenicity of RFRin standard-bred animal models are reviewed. The most comprehensive studies included in this section are two-year oncogenicity bioassays performed in rodents. Over several decades, a substantial database has developed that supports the value of the two-year oncogenicity bioassay inrodents in predicting carcinogenic activity in humans [reviewed in 21]. The design of these studies involves observation of relatively large groups of animals over the majority of their normal lifespan. These studies also include microscopic evaluation of a large list of tissues from all study animals, and therefore support an evaluation of the risk of oncogenesis in all major organs. In most cases, studies are designed to comply with the safety assessment requirements of organizations such as the United States Food and Drug Administration (FDA; the regulatory body responsible for the oversight of the safety of cellular telephones and other wireless communications devices in the United States); the United States Environmental Protection Agency (EPA); and/or the International Council for Harmonisation (ICH, an international organization focused on standardizing safety testing protocols used around the world).
  • In Section3, studies designed to evaluate the potential carcinogenic activity of RFRin tumor-prone mice will be reviewed. Studies discussed in this section include those performed using(a) animals that have been genetically engineered by insertion of an oncogene or deletion of a tumor suppressor gene, resulting in increased sensitivity to neoplastic development, and (b) animal strains that have been selectively bred to increase sensitivity to neoplastic development in specific tissues. In addition to supplementing data from two-year oncogenicity bioassays, these models may identifycarcinogenic effects of RFR that occur in sensitive subpopulations. It is important to note, however, that most tumor-prone animals lack a long history of use in hazard identification; as such, their value in predicting human responses has generally not been established.
  • In Section 4, studies are reviewed in which exposure to RFRis combined with simultaneous or sequential exposure to other chemical or physical agents to characterize the possible activity of RFR as a tumor initiator, a tumor promoter, or a co-carcinogen in one or more organs. These multi-stage carcinogenesis bioassays are most commonly performed using animal models that have been designed for use as research tools to study organ-specific carcinogenesis. As is the case with tumor-prone animals, the value of data from multi-stage carcinogenesis bioassays in predictinghuman cancer responses to exogenous agents has not been established.

2.Assessment of the Possible Oncogenic Activity of RFR using the Rodent Two-Year Bioassay

2.1.Design of the Rodent Two-Year Oncogenicity Bioassay

The chronic (two-year) rodent oncogenicity bioassay is considered to be the “gold standard” protocol for the experimental assessment of carcinogenic activity. Two-year bioassays in rodents have been demonstrated to be useful predictors of human carcinogenic responses [21]; data from these bioassays are accepted by American and international regulatory agencies as providing the most comprehensive and most predictive experimental approach to assess agent carcinogenicity.

In a well-designed chronic rodent bioassay, elements of the study protocol(e.g., group sizes, exposure levels) are designed to maximize the likelihood of identifying an increase in cancer incidence following long-term exposure to agents with carcinogenic activity. The design of these studies was discussed in detail in the chapter in this volume entitled “Evaluation of the Toxicity and Potential Oncogenicity of ELF Magnetic Fields in Experimental Animal Model Systems”, and will not be repeated here. Briefly, however, exposures begin either in utero or when mice or rats are young adults, and continue for two years. The two-year exposure period encompasses the majority of the normal life span of these species, and thereby maximizes the probability of identifying an oncogenic effect. Group sizes (≥ 50 animals per sex per group) are larger than are used in other toxicology bioassays, and are often increased to as many as 100 (or more) animals per sex per group to increase statistical power. Study designs generally include complete histopathologic evaluation of approximately 45 tissues from each study animal, thereby permitting the assessment of oncogenicity in all major organs in the body.

These studies are most often conducted using a standardized study design that is referred to as the “Chronic Oncogenicity Bioassay in Rodents.” This study design has been used widely as the basis to evaluate the potential carcinogenicity of new drugs, agricultural chemicals, occupational chemicals, and a wide range of environmental agents, and is considered to be a useful predictor of human oncogenicity [21]. The International Agency for Research on Cancer (IARC) and United States National Toxicology Program (NTP) have both performed analyses demonstrating the utility of the rodent chronic bioassay as a predictor of predict human carcinogenicity. As a result, it is generally accepted among toxicologists that the chronic rodent bioassay currently provides the best available experimental approach to identify agents that may be carcinogenic in humans.

2.2Results of Two-Year Rodent Oncogenicity Bioassays of RFR

Two-year bioassays of RFR and microwaves of a similar frequency have been performed in three strains of rats (Sprague-Dawley, Fischer [F344], Wistar) and in one strain of mice (B6C3F1).

It is important to note that these studies were performed using several different RFR exposure metrics and delivery systems; study protocols also include substantial differences in the duration of daily exposure. Several chronic bioassays of RFR wereperformed using “head only” or “head first” exposures, in which restrained rodents are exposed using a “Ferris wheel” type of exposure system. This type of exposure system directs RFR to the head of the animal, and in this manner simulates the regional deposition of RF energy that humans may receive from a mobile telephone. On this basis, use of a “head only” exposure system may provide an advantage in terms of the specific parts of the body in which RFR is deposited. However, the need to restrain animals during exposures performed using these systems limits the duration of daily RFR exposure to several hours, and may also induce restraint stress.