CIESIN SEDAC icon NASA logo home page icon
Stratospheric Ozone and Human Health Project

UNEP logo WMO logo
Environmental Effects of Ozone Depletion: 1994 Assessment



CHAPTER 2


EFFECTS OF INCREASED SOLAR ULTRAVIOLET
RADIATION ON HUMAN HEALTH

J.D. Longstreth (USA), F.R. de Gruijl (The Netherlands), M.L. Kripke (USA),
Y. Takizawa (Japan), and J.C. van der Leun (The Netherlands)



Table of Contents

  1. Summary
  2. Introduction
  3. Ocular Effects
  4. Immunological Effects
  5. Infectious and Other Diseases
  6. Skin Cancer
  7. Wavelength Dependancy of Melanoma
  8. Related Issues
  9. References



Summary

The increase in UV-B associated with stratospheric ozone depletion is likely to have a substantial impact on human health. Potential risks include increases in the incidence of, and morbidity from, eye diseases, skin cancer, and infectious diseases. Quantitative estimates of risk are available for some effects (e.g., skin cancer), while for others (e.g., infectious diseases), quantitative estimates are not possible due to a lack of sufficient data.

UV radiation has been shown in experimental systems to damage the cornea and lens of the eye. Chronic exposure to UV-B (resulting in a high, cumulative, lifetime dose) is one of several factors clearly associated with the risk of cataract of the cortical and posterior subcapsular forms. Estimates of the effect of ozone depletion on cataract have been made, but are still highly uncertain. (As stated in the l989 report, [van der Leun, et al., l989] these estimates predict an approximately 0.5 % increase in cataract for each 1% sustained decrease in ozone.)

Some components of the immune system are present in the skin, which makes the immune system accessible to UV radiation. Experiments in animals show that UV exposure decreases the immune response to skin cancers, infectious agents, and other antigens and can lead to unresponsiveness upon repeated challenges. Suppressed immunity may occur either locally in sun-exposed skin or systematically, at non-exposed sites. Studies in human subjects also indicate that exposure to UV-B radiation can suppress the induction of some immune responses and may cause systemic alterations in immune function. The importance of these immune effects for infectious diseases in humans in unknown. However, in areas of the world where infectious diseases already pose a significant challenge to human health, and in persons with impaired immune function, the added insult of UV-B-induced immune suppression could be significant.

In susceptible (light-skinned) populations, the cumulative lifetime exposure to UV-B radiation is the key risk factor for development of non-melanoma skin cancer (NMSC). This knowledge has permitted the development of quantitative risk estimates for increases in the incidence of NMSC resulting from ozone depletion. Using information derived from animal experiments and human epidemiology, it is estimated that a sustained 1% decrease in stratospheric ozone will result in an increase of NMSC incidence of approximately 2.0% The relationship between UV-B exposure and melanoma skin cancer is less well understood and appears to differ fundamentally from that of NMSC in that it is not apparently a function of cumulative lifetime dose but may be related to the accumulation impact of multiple high dose exposures such as those received in sunburns. Epidemiologic data indicate that the risk of melanoma increases with an increase in episodes of intense sunlight exposure, (i.e. sunburn) especially during childhood. There is, however, uncertainty about how the relationship between these exposures and melanoma should be modeled so that the estimates of the increase in melanoma that would result from ozone depletion are much less certain.

[Back to top of page]


Introduction

As presented in detail in chapter 1, solar ultraviolet radiation (UVR) illuminates nearly everything and everyone on the earth's surface not covered or shadowed. With stratospheric ozone depletion, increases in the ambient levels of a particular type of ultraviolet radiation known as UV-B are likely to occur. In humans and animals, the primary (i.e., direct) effects of increases in UV-B on health are manifest through those organs which are exposed to sunlight, i.e., eyes and skin. These effects occur because of the absorption of UV-B photons by molecules in these organs and the resulting tranfer of energy to produce changes that may be either beneficial or adverse. A direct beneficial influence of exposure is the formation of Vitamin D in the skin, a process important to the maintenance of bone tissue. Direct adverse effects of exposure to UV-B include snow blindness, cataract, sunburn, "aging" of the skin, photodermatoses and skin cancer. Some effects may have both beneficial or adverse elements, depending on how they are expressed. This applies, for instance, to the influences of UV-B radiation on the immune system. The resulting suppression of immune reactions in the skin is beneficial in patients suffering from psoriasis, (a hyperproliferative skin disorder) but adverse when it affects the immune defense against skin tumors or infectious agents. In addition, there may be indirect beneficial or adverse effects. An example indirect adverse effect could be the increase in disease associated with the potential decrease in food production discussed in chapter 3.

The question to be addressed by the present chapter is, what will be the human health consequences of an increase in the UV-B radiation reaching the surface of the earth? One cannot simply say that all effects of UV-B radiation will change in proportion to the increase in radiation, because in many cases, the relationship between the amount of exposure and effect is non-linear. For example, exposure of the skin to UV-B radiation results in a hyperproliferation of skin cells and increased pigment production providing the skin with efficient protection against sunburn through the increase in the UV absorbing molecules, keratin and melanin. The hyperproliferation of skin cells may, however, also render the skin more susceptible to cancer. In some instances, the modifying influences may be so strong that greater doses of radiation lead to smaller effects.

For example, photodermatoses are skin diseases where the skin lesions are caused by light. Solar UV-B radiation is the predominant causative agent for several of these diseases. Although many patients and their doctors expect an aggravation of these diseases with a decreased ozone layer, there are reasons to question this expectation. In the first place, these diseases generally occur less frequently and with less severity in sunny areas of the world. Second, many patients with photodermatoses are treated effectively by regular exposures to low-dose UV-B radiation during winter. Because depletion of the ozone layer will increase UV-B irradiance, especially in winter, this may improve the patients' condition [van der Leun and de Gruijl, 1993].

Due to such complications, and due to the limited knowledge we have on some of the effects of UV-B radiation on humans, the overall consequences of an increase of UV-B irradiance on health cannot be predicted in a straightforward manner. This section, therefore, presents both qualitative and quantitative answers to the question of what the consequences of stratospheric ozone depletion and its accompanying increase in UV-B may be for human health. Quantitative estimates of UV-B effects are presented for non-melanoma skin cancer; however, the impact of UV-B exposure on other effects has been treated qualitatively. It should be stressed that the dependence on qualitative estimates for some effects does not imply that these are less significant; indeed, the quantitative estimates should be treated very cautiously because they involve many assumptions, are based on data mainly from the United States, and may not be representative of all regions of the world.

[Back to top of page]


Ocular Effects

Background

Perhaps the best documented short-term ocular effect of exposure to UV radiation (especially UV-B and UV-C) is photokeratoconjunctivitis (`snow blindness' and 'welder eyes'), i.e., an inflammatory reaction (a reddening) of the surface of the eyeball. Extraordinarily painful, one episode should be sufficient to induce behavior modification to prevent recurrences, e.g., the use of proper eye protection. The effects of long-term or chronic exposures, e.g., pterygium or cataract, are less well documented in part because they result after many years of exposure, and, in part, at least for cataract, because many other factors are known to have etiologic role. For such endpoints, as with endpoints such as NMSC, causality has to be inferred from epidemiological studies supported by animal experiments. For a much more detailed review of this subject, the reader is referred to Pitts and Kleinstein [1993].

Epidemiological Data

Epidemiological data indicate that chronic sunlight exposure is associated with pterygium, an outgrowth of the conjunctiva (outermost mucous layer) over the neighboring cornea (overlying the lens), and with climatic droplet keratopathy, a degeneration of the corneal stroma (fibrous layer of tissue of the cornea) with droplet-shaped deposits [Doughty and Cullen, l989, and Hollows, 1989]. Climatic droplet keratopathy can be a major cause of blindness. Both of these conditions are common in certain geographical locations, especially in snowy or sandy areas.

Sun exposure is also thought to be a contributing factor in the development of cataract, an opacity in the crystalline lens of the eye (for an extended analysis, see Dolin 1994). The World Health Organization (WHO) estimated in 1985 that cataract was the main cause of avoidable blindness, responsible for 17 million cases of blindness worldwide (about half of the total) [Maitchouk, 1985].

The etiology of cataract in humans is multifactorial; increased relative risks are associated not only with increased exposure to UV-B, but also with increasing age, diabetes, renal failure, severe diarrhea, heavy smoking, hypertension, high alcohol consumption, excessive heat, and malnutrition [e.g., Harding and Van Heyningen, 1987]. The increase in relative risk associated with increased UV-B exposure is 1.3 to 3.5-fold, compared to diabetes which has a relative risk increase of approximately 10-fold. However, because of the nearly universal exposure of humans to solar UV radiation, the size of the population likely to be affected by this factor is significantly greater than for any of the other factors (except age) listed above.

In comparing epidemiological studies, two things must be kept in mind: 1) there are differences in the precise definition of cataract since the form of the disease can vary from small opacities which do not impair vision to a completely opaque lens with severe loss of vision, (furthermore, there can be variation between clinicians, and techniques; future studies would benefit from standardization of classifications) and 2) there are different types of cataract that show differing associations to the factors discussed above. Exposure to solar UV radiation appears specifically to increase the risk of cortical opacities (including opacities not impairing vision): a clear UV dose-related trend in risk was established in a well designed, cross-sectional study of Chesapeake Bay watermen [Taylor et al., 1988]. This UV exposure-associated increased risk of cortical opacities was confirmed for the male population in the Beaver Dam Eye Study (an odds ratio of 1.36 between numbers with `high versus low exposures', 95 percent confidence interval 1.07-1.79), but the risks of other types of lens opacities, i.e., nuclear and posterior subcapsular cataract, were not detectably increased by UV exposure [Cruickshanks et al., 1992].

The relationship between sunlight exposure and vision-impairing cortical cataract was confirmed in a large case-control study drawn from three Italian opthalmology clinics [Italian-American Study Group, 1991]. Increased odds ratios for cortical cataract were found for working outdoors (1.75, 1.15-2.65) and spending leisure time in sunlight (1.45, 1.09-1.93); these increased odds ratios were also found for the mixed cataract, but not for posterior subcapsular and nuclear cataract. However, posterior subcapsular cataract is relatively rare, making detection of risk in a small population difficult. In a special case-control study in the Chesapeake Bay area on 168 surgical cases involving this type of cataract, a highly significant (p = 0.006) positive trend was established between risk and exposure to ambient UV radiation [Bochow et al., 1989].

The relationship between geographical location (latitude) and prevalence of cataracts has also been attributed to ambient UV radiation; e.g., in Aborigines [Hollow and Moran, 1981] and in the North American population [Hiller et al., 1983]. The latter study provided the basis for the USEPA risk estimates for increases in cataract from ozone depletion [USEPA 1987].

Experimental Data

There is ample experimental evidence that UV radiation can damage the lens and that in rabbits the active portion of the solar spectrum lies in the UV-B region [Pitts et al., 1977]. In albino mice, regular UV-B exposure (1 to 2 months of daily exposure) can disrupt the anterior (frontal) part of the lens, resulting in opacities [Jose and Pitts, 1985]. High dose, daily UV-A exposure also induced opacities in the anterior lens of nocturnal animals and squirrels [Zigman et al., 1991]. However, experimentally-induced opacities in animals are located centrally in the anterior lens (the irradiated zone), whereas human anterior cortical cataracts tend to form in the periphery (the equatorial region) of the lens. Also, microscopic examination of experimental cataract commonly shows marked disruption of the lens epithelium and inward folding of the underlying cortical fibers, which is not seen in human cataract. These differences, while possibly attributable to differences between animals and humans in geometry, level of exposure and/or ocular structure, contribute to the uncertainty associated with extrapolating from data in mice to human cataract risk.

Solar UV Exposure of the Human Eye

The exposure of the unprotected eyes to solar UV radiation is significantly influenced by the shielding from the eyebrows and eyelids (through squinting) and is strongly dependent on the direction of the line of sight. Highly reflecting surfaces increase the exposure dramatically; for example, the risk of photokeratoconjunctivitis is strongly increased over snow surfaces [Sliney, 1987]. This variability in exposure can seriously complicate studies in which eye exposures in different environments are compared unless careful exposure measurements are made. [Rosenthal et al., 1991].

In humans, the anterior eye is not only exposed to light entering along the line of sight, but also to light impinging from very oblique angles. Coroneo [1993] drew attention to the phenomenon in which impinging (UV) light, especially that entering the eye at a very oblique angle, can be focussed onto the overlying periphery of the cornea and the lens; this would intensify the light while it is traversing the cornea and the lens in a direction roughly perpendicular to the line of sight. This phenomenon could explain not only the sites of preference for pterygium and climatic droplet keratopathy, but also why many cortical cataracts occur in the nasal quadrant of the lens [Schein et al., 1994, Adamson et al., 1991, Hollow, 1989, and personal communication with Dr. B. Klein for the Beaver Dam Eye Study].

Impact of Ozone Depletion on Eye Conditions

In view of present data, it is prudent to consider that a depletion of the ozone layer could be associated with an increase in the incidence of cataract and other ocular effects of UV-B, e.g., pterygium and snow blindness. It is, however, difficult to estimate the magnitude of the increase without adequate information on the wavelength dependence of these effects and proper dose-response relationships. In the case of cataract, by assuming a certain wavelength dependence and with some additional assumptions, one can produce an estimate based on epidemiological data (e.g., an estimate from EPA based on Hiller et al., [l985], cited in van der Leun [l989], and one from van der Leun and de Gruijl, [1993], based on Taylor et al., [l988], and Pitts et al. [l977]: 0.3-0.6% and 0.5% increase in cataract, respectively, for every 1% decrease in ozone) but such an estimate has a high degree of uncertainty. A similar quantitative estimate is not yet possible for snow blindness, pterygium and climatic droplet keratopathy because of the inadequacy of epidemiologic information and experimental data. Nevertheless, it is prudent public health policy to indicate to medical personnel and the lay public that these are effects which may well increase with increased exposure to UV-B.

[Back to top of page]


Immunological Effects

Background

Through a variety of complex, delicately balanced mechanisms, the immune system helps maintain health by protecting the host against infectious diseases and some cancers. The two most important of these mechanisms are 1) humoral immunity, involving the production of antibodies that can neutralize toxins, kill microorganisms, prevent infection, and assist in the elimination of infectious agents, and 2) cellular immunity, involving the production of chemical mediators (cytokines) by lymphocytes which activate other cells of the lymphoid system to kill pathogens, virus-infected cells, and cancer cells. These two arms of the immune response, humoral and cellular, are delicately balanced, and some cytokines involved in activating one pathway tend to inhibit the other. A severe imbalance in either direction can lead to pathological conditions, such as allergies and inflammatory and autoimmune diseases.

Because skin is an important immunological organ, the immune system is vulnerable to modification by environmental agents, including UV-B radiation. Demonstrations that systemic immunity can be perturbed by exposing skin to UV-B radiation raise the concern that ozone depletion might adversely influence immunity to infectious diseases.

Immunomodulatory Effects of UV-B Radiation

There is now ample evidence that exposure of humans and experimental animals to UV-B radiation from artificial or natural sources can modify the immune system both at the site of exposure (locally) as well as systemically, mainly by decreasing cellular immune responses [Hersey et al., 1983; Morison, 1989; Kripke, 1984, 1990; DeFabo and Noonan, 1993; Cruz and Bergstresser, 1988]. The immunosuppressive effects of UV-B radiation have been shown to play an important role in determining the outcome of both melanoma [Donawho and Kripke, 1991] and non-melanoma skin cancer [Kripke, 1984; 1990; Fisher and Kripke, 1982], certain infectious diseases [Jeevan and Kripke, 1993], and some forms of autoimmunity [Ansel et al., 1985] and allergy, e.g., delayed type hypersensitivity [Kripke, 1984; Cruz and Bergstresser, 1988, Noonan et al., l981] in laboratory animal models of these diseases. Furthermore, introduction of a foreign substance (antigen) during a critical period after UV irradiation can lead to immunological tolerance, rendering the host unresponsive to re-introduction of the same antigen at a later time [Kripke, 1984].

In addition to these systemic effects on immune responses, UV-B irradiation can also inhibit local inflammatory responses within UV-irradiated skin; thus, the response elicited by injection of antigen into the skin of sensitized individuals (delayed hypersensitivity response) may be diminished in UV-irradiated skin [Morison, 1989]. Such local effects of UV irradiation can also decrease resistance to the growth of cancer cells, including melanomas [Donawho and Kripke, 1991].

Information on the UV-induced alterations in immune function in humans is fairly limited and is drawn principally from experimental studies of the impact of UV exposure on natural killer cell activity, contact allergy (e.g., poison ivy) and delayed hypersensitivity responses (e.g., the tuberculin skin test reaction) [Hersey et al., 1983; Morison, 1989; Yoshikawa et al., 1990; Cooper et al., 1992].

Mechanisms

In recent years, many advances have been made in our understanding of the cellular and molecular mechanisms involved in modulation of immune responses by UV-B exposure. Studies with mice provide strong evidence that at least some of the immunomodulatory effects of UV irradiation result from the production and release of immunologically-active cytokines and other substances from cells in the skin. Interleukins 1 and 10 (IL-1, IL-10), tumor necrosis factor (TNF)-alpha, urocanic acid (UCA), and prostaglandins have all been suggested as contributors to the altered immune responses observed in UV-irradiated animals [DeFabo and Noonan, l983, 1993; Rivas and Ullrich, 1992, 1994a; Noonan et al. 1988; Vermeer and Streilein, 1990; Luger and Schwartz, 1990].

The ability of UV-irradiated keratinocytes to produce IL-10 [Rivas and Ullrich, 1992] is particularly noteworthy, because this cytokine preferentially decreases delayed hypersensitivity responses, leaving humoral responses undiminished [Mossman et al., 1991]. IL-10 is thought to alter the cells that initially take up antigens (mainly macrophages, epidermal Langerhans cells, and dendritic cells in lymphoid organs) and inhibit their ability to stimulate the subset of murine helper T lymphocytes responsible for generating delayed hypersensitivity responses (Th1 cells), while leaving intact their ability to stimulate Th2 lymphocytes, which are involved in antibody formation and suppression of Th1 cells [Ullrich, 1994; Rivas and Ullrich, 1994]. Figure 2.1 presents one possible model for how these various factors and events may operate in UV-B induced immunosuppression.

FIGURE

Fig. 2.1. Model for UV-B Induced Immunosuppression. (See text for details; Cruz and Bergstresser, 1988; Kripke, 1984; DeFabo and Noonan, 1993; Luger and Schwartz, 1990; Rivas and Ullrich, 1994)

There is also evidence that the cis-photoisomer of UCA can alter the ability of antigen-presenting cells to stimulate T lymphocytes [Noonan et al., 1988] and that TNF-alpha [Vermeer and Streilein, 1990] can alter the migration pattern of antigen-presenting cells in the skin (epidermal Langerhans cells). Direct exposure of epidermal Langerhans cells to UV-B radiation may also alter their antigen-presenting activity [Simon et at., 1990; 1991], and UV irradiation of the skin induces an inflammatory response that attracts other types of antigen-presenting cells into the UV-irradiated site [Cooper et al., 1993].

Thus, UV radiation appears to alter the induction of immune responses by perturbing the balance of factors that normally regulate the immune response and by altering the activity and distribution of the cells responsible for triggering these immune responses. The mechanisms involved in the local inhibition of delayed hypersensitivity and tumor resistance are unknown, but they are likely to involve local modifications of cytokines and other mediators [Luger and Schwartz, 1990] and changes in the expression of adhesion molecules on cells of various types in the skin [Krutmann et al., 1992].

The molecular events that initiate photoimmunological effects are incompletely understood. Studies in the mouse [Kripke et al., 1992; Yarosh et al., 1994] and the opossum [Applegate et al., 1989] demonstrated that increasing the repair of UV-specific lesions in DNA abrogated the suppression of contact allergy and delayed hypersensitivity reactions, implying that UV-induced DNA damage is an essential initiating step. These authors hypothesized that DNA damage may trigger the production of immunoregulatory cytokines in the skin [Kripke et al., 1992].

Another mechanism by which UV irradiation could initiate immunosuppression is by converting trans-UCA to its cis-isomer. Trans-UCA is present in large quantity in mammalian epidermis; the more soluble cis-isomer is formed upon absorption of UV radiation. There is considerable evidence that cis-UCA suppresses certain immune reactions in mice, suggesting that UCA may also mediate some of the immunomodulatory effects of UV irradiation [DeFabo and Noonan, 1983, 1993].

Wavelength Dependence of Immune Suppression

The two molecular mechanisms proposed for initiating immune suppression (DNA damage and UCA isomerization) imply different wavelength dependencies, which in turn, give different Radiation Amplification Factors (RAFs); see Chapter 1). If DNA damage were the sole mechanism by which UV radiation caused immune suppression, then the percent increase in innumosuppressive UV radiation per percent ozone depletion (RAF) would be between 1.2 and 1.7. On the other hand, if UCA isomerization were the sole mechanism, the RAF would be between 0.4 and 0.8 (See Table 1, Chapter 1). The most detailed action spectrum for immune suppression was determined for systemic suppression of contact allergy in the BALB/c (albino) mouse strain, using wavelengths between 254 and 320 nm [DeFabo and Noonan, l983] (See Figure 2.2). Another action spectrum for local suppression of contact allergy in a different mouse strain has also been measured [Elmets et al., l985]. However, neither action spectrum can distinguish among the two proposed mechanisms or a composite mechanism with absolute certainty, partly because of a lack of detailed information concerning the effect of wavelengths >320 nm. At present, there is no information available on the action spectrum for immune suppression in humans.

FIGURE

Fig. 2.2. Action spectra of possible importance to immunosuppression: suppression of CH (-BULLET-) and absorption UCA (....) [DeFabo and Noonan, 1983], in vivo trans ---> cis-UCA (*) [Gibbs et al., 1993], pyrimidine dimers in skin [Freeman et al., 1989]

[Back to top of page]


Infectious and Other Diseases

The finding that UV-B irradiation of laboratory animals and humans could impair the induction and elicitation of certain types of immune responses raised concerns that immunity to infectious agents might also be impaired. Theoretically, UV-B irradiation could affect the pathogenesis of infectious diseases by modifying the defense mechanisms of the host to a microbial pathogen or by directly activating an infectious organism present within exposed skin. Evidence for both effects has been provided in model systems, but data in humans are still very limited.

Effects of UV in Experimental Models

Viral diseases

Because of the ability of UV radiation to activate herpes simplex virus (HSV) infection in humans [Perna et al. 1991], experimental models of this infection have been studied extensively. Exposure to UV radiation triggered active disease in mice with latent HSV infection, decreased the delayed hypersensitivity response, induced suppressor T cells to HSV, and impaired resistance to initial infection [reviewed in Jeevan and Kripke, 1993]. UV irradiation also decreased cellular immunity to reovirus infection in mice; however, clearance of systemic virus was unimpaired, suggesting that antibody may be more important than cellular immunity in controlling this infection [Letvin et al., 1981]. Reports that the AIDS virus could be detected in epidermal Langerhans cells in the skin of HIV+ persons [Tschachler et al., 1987] raised concern that UV-B irradiation might accelerate the course of this disease. Although there is no evidence to support this possibility in humans, one study in mice suggested that the course of an AIDS-like immune deficiency was accelerated by chronic UV irradiation [Brozek et al., 1992]. The study showed that UV-irradiated mice developed antibodies against a murine retrovirus earlier than non-irradiated controls, suggestive of more rapid disease progression.

Virus activation by UV

In addition to decreasing host resistance to infection by means of its immunomodulatory effects, UV-B radiation has the ability to activate latent viruses contained within cells exposed directly to UV. This phenomenon, which has been demonstrated in vitro [Zmudzka and Beer, 1990; Schmitt et al., 1989], would be expected to affect viruses present in cells of the skin, such as papillomaviruses, herpes simplex virus, and perhaps HIV, which can be present in epidermal Langerhans cells [Tschachler et al., 1987]. In vivo activation of HIV in the skin has been demonstrated in transgenic mouse models, in which UV irradiation turned on genes involved in virus activation and replication in the skin [Morrey et al., 1991; Vogel et al., 1992]. These studies have raised the concern that exposure of HIV-infected persons to UV radiation early in the course of infection, when the virus is present in the skin, might accelerate the course of AIDS. However, there is at present no experimental evidence to support this possibility. Thus, the likely impact of increased UV-B radiation on HIV infection resulting from viral activation remains unknown.

Parasitic infections

UV-B induced immunosuppression has been shown to have an impact on leishmaniasis [Giannini, 1986; Giannini and De Fabo, 1987], malaria (B. Ward, personal communication) and trichinosis [Goettsch et al., 1994], but not on schistosomiasis [Jeevan et al., 1992]. Leishmaniasis, a tropical parasitical disease has been studied in some depth using a mouse model. In humans, the parasites are deposited intradermally by infected sandflies, where they induce ulcerating, cutaneous lesions. The infection may be limited to the skin or can progress to a systemic disease, which may be fatal. In different mouse strains, the organism can produce either a self-limiting disease controlled by cellular immunity or a progressive, lethal infection. The outcome depends upon what type of immune response the particular strain of mouse mounts against the parasite: a Th1 response leads to immunity, whereas a Th2 response leads to disease progression. UV irradiation of mice before and after infection through exposed tail skin improved the appearance of the resulting skin lesions, but decreased the delayed hypersensitivity (Th1-type) response to the parasites thereby decreasing clearance of the organism [Giannini, 1986]. The mortality rate was increased in infected and UV-irradiated mice, which also exhibited decreased resistance to reinfection with the parasite [Giannini and DeFabo, 1987]. Thus, in this model, although there was a reduction in the size and severity of the skin lesions in UV-irradiated mice, the pathogenicity of the disease was increased. Similar decreases in the clearance of parasites have also been observed in rodent models of both trichinosis and malaria, but not in schistosomiasis [Goettsch et al., 1994; B. Ward, personal communication; Jeevan et al. 1992].

Bacterial infections

The impact of UV-B exposures on immunity to mycobacterial infection in mice has been studied extensively. Infection of UV irradiated mice in the hind footpad with Mycobacterium bovis BCG or M. lepraemurium resulted in a decreased delayed hypersensitivity response and delayed the clearance of bacteria from the lymphoid organs. Furthermore, macrophages from the spleen and peritoneal cavity of UV-irradiated mice had a reduced ability to ingest M. bovis. A single high dose of UV radiation given 3 days before infection accelerated the rate of death from M. lepraemurium [Jeevan and Kripke, 1993]. Similar results were obtained in a mouse model of Lyme disease. UV-irradiated mice infected with Borrelia burgdorferi exhibited a decreased delayed hypersensitivity response, decreases in certain subclasses of anti-Borrelia antibodies, and increased numbers of organisms in the joints [Brown et al., 1994; E. Brown, unpublished data].

Fungal Infections

The only model of fungal infection studied to date is systemic Candida albicans infection in mice. Candida is an opportunistic fungus normally present on the skin, which causes systemic disease in immunosuppressed persons. Exposure of mice to a single high dose of UV-B radiation one day before a lethal intravenous injection of Candida significantly reduced their survival time. Lower doses of UV radiation decreased the delayed hypersensitivity response to this organism, but had no effect on the outcome of systemic disease [Denkins and Kripke, 1993; Denkins et al., 1989].

Autoimmune and other diseases

The possibility exists that UV-B could reduce some forms of autoimmunity, by virtue of its ability to attenuate cell-mediated immunity. However, studies in experimental models of autoimmune diseases are quite limited and only serve to underscore the complexity of the situation. In a study of the effects of UV-B irradiation on the development of autoimmune hemolytic anemia in autoimmune strains of mice, UV-B irradiation accelerated and exacerbated the disease process [Ansel et al., 1985]. In another study using a different strain of autoimmune mice, chronic UV irradiation seemed to have no effect on the pathogenesis of this disease [Strickland, 1984].

In humans, UV-B is used therapeutically for the treatment of certain skin diseases, such as psoriasis, which seem to have an immunological component. On the other hand, one autoimmune disease, systemic lupus erythematosis is aggravated by UV exposure, and UV is involved in the pathogenesis of some photoallergic and photosensitivity diseases. Thus, increased UV-B radiation is likely to have varied, and even opposing, effects on autoimmune and other diseases, and it is presently not possible to make any predictions as to its impact.

Effects of UV on Human Diseases

Of course, the crucial question regarding the significance of increased UV-B radiation is whether the findings in animal studies apply to humans. It is clear that exposure of humans to natural or artificial sources of UV-B radiation impairs the activity of T lymphocytes, decreases the activity of natural killer cells, decreases the number of epidermal Langerhans cells, and abrogates the induction of the contact allergy response to chemicals applied onto the irradiated skin [Morison, et al., 1979; Hersey, et al., 1983; Yoshikawa, et al., 1990; Cooper, et al., 1992]. Furthermore, most of these immunological effects occur even in persons with darkly pigmented skin, implying that pigmentation does not confer complete protection against UV-induced immune suppression [Scheibner, et al., 1986; Vermeer et al., 1991] so that the population potentially at risk for an impact on infectious diseases is far greater than that for skin cancer.

There is evidence both in murine models [Streilein and Bergstresser, l988; Noonan and Hoffman, 1994] and humans [Yoshikawa et al., 1990] for genetic differences in susceptibility to UV-induced immune suppression unrelated to pigmentation. Such differences may be important determinants of risk for skin cancer development and susceptibility to UV-induced modifications of infectious diseases.

The effects of UV-B radiation on infectious disease processes are likely to be complex and unpredictable. Historically, UV radiation was used to treat a variety of skin diseases, most notably skin tuberculosis, and in recent years, UV-B radiation has been used to treat psoriasis, acne, and other cutaneous diseases. On the other hand, sunlight exposure was thought to aggravate pulmonary tuberculosis and scarring from smallpox [van der Leun and de Gruijl, 1993]. There is also considerable evidence that exposure to sunlight and to UV radiation can trigger the appearance of cutaneous lesions caused by herpes simplex virus types 1 and 2 in persons already harboring a latent infection [Klein and Linnemann, 1986; Spruance, 1985; Wheeler, 1975], and exposure to UV radiation has been reported to increase the severity of skin lesions associated with herpes zoster infection [Szigeti et al., 1976]. UV radiation may also be a contributing factor to papillomavirus infections in immunosuppressed patients, since such individuals have a high incidence of viral warts on sun-exposed body sites [Boyle et al., 1984; Dyall-Smith and Varigos, 1985]. Whether these effects are a result of virus activation or immune suppression by UV or both is not clear.

One approach to addressing the effect of UV-B radiation on resistance of humans to infection is to analyze the effect of UV on a cell-mediated immune response against microbial antigens. Healthy contacts of leprosy patients often exhibit a delayed hypersensitivity response upon intradermal injection of lepromin, an antigen prepared from leprosy bacilli. Exposure of the injection site to UV-B radiation before and after inoculation of lepromin markedly reduced the size of the skin test reaction and granulomatous response and decreased the number of T lymphocytes within the reaction site, compared to the reaction site in unexposed skin of the same subject [Cestari et al., 1994]. This study demonstrates that exposure to UV-B radiation can diminish the cutaneous immune response to an infectious microorganism in humans; however, it does not address the question of whether UV-B irradiation would also exacerbate the disease process in persons suffering from leprosy or other mycobacterial infections.

Obviously, it will be very difficult to assess the role of UV-B radiation on natural infections in human populations. Based on current knowledge, we would predict that an effect of UV-B radiation would manifest itself as an increase in the severity or duration of disease and not necessarily as an increase in disease incidence except possibly where reactivation of a latent virus is reflected as incidence. Since infectious diseases are influenced by many host and environmental factors, the effect of UV-B radiation on a given disease process may be difficult to discern from epidemiological studies. Clearly, more information is needed on this subject. The growing evidence that the balance of Th1- and Th2-type immune responses plays an important role in determining the outcome of various infectious diseases, and the suggestion from animal studies that UV irradiation may shift this balance toward a Th2-type response suggests that UV radiation may indeed influence the pathogenesis of some diseases. It may also influence the outcome of vaccination against infections. Whether this influence is beneficial or harmful will probably depend on what type of immune response is most effective in protecting against a particular microorganism. Thus, it is presently difficult to predict both the direction and magnitude of an effect of UV-B radiation on a particular disease process.

This possibility is borne out in a study in mice demonstrating that UV irradiation could prevent the induction of an autoimmune demyelinating disease, experimental allergic encephalomyelitis, in mice [Hauser et al., 1984] which is a Th1-dependent disease.

For example, photodermatoses are skin diseases where the skin lesions are caused by light. Solar UV-B radiation is the predominant causative agent for several of these diseases. Although many patients and their doctors expect an aggravation of these diseases with a decreased ozone layer, there are reasons to question this expectation. In the first place, these diseases agenerally occur less frequently and with less severity in sunny areas of the world. Second, many patients with photodermatoses are treated effectively by regular exposures to low-dose UV-B radiation during winter. Because depletion of the ozone layer will increase UV-B irradiance, especially in winter, this may improve the patients' condition [van der Leun and de Gruijl, 1993].

UV dose-response and wavelength studies for immunosuppressive effects of UV radiation in humans are essential for making quantitative predictions. More important still is understanding the significance of these immunological effects for the pathogenesis of human diseases. In spite of its central importance to the analysis of the consequences of stratospheric ozone depletion for human health, information to address the latter issue is difficult and expensive to obtain and is thus almost totally lacking.

[Back to top of page]


Skin Cancer

Background

UV radiation damages DNA (i.e., is genotoxic). This may lead to faulty replication of DNA in a daughter cell, i.e., fixation of a mutation. Mutations in certain key genes (proto-oncogenes or tumor suppressor genes) that regulate the cell cycle, cell differentiation, and cell death (apoptosis) can lead to formation of a cancer cell. Recent research is beginning to reveal how these steps relate to the course of UV-induced tumor formation.

Animal data and human epidemiologic studies clearly indicate that excessive exposure to UV radiation is associated with skin cancer in humans. (See IARC [1992] for a comprehensive overview). The involvement of DNA damage, for instance, was elegantly demonstrated in opossums (Monodelphis domestica) which have an enzyme (photolyas) that reverses certain types of UV-induced DNA damage (cyclobutane pyrimidine dimers) upon exposure to UV-A and visible light; UV-A plus visible light treatments counteracted the UV induction of skin tumors (both squamous cell carcinomas and melanomas). It is also known that xeroderma pigmentosum (XP) patients whose cells are deficient in the repair of UV-induced DNA damage have dramatically increased risks of skin cancer (both squamous and basal cell carcinomas, and melanoma). [This observation is, however, complicated by the fact that individuals with other kinds of defects in DNA repair (e.g., Cockayne's syndrome and trichothiodystrophy), do not have an increased risk of skin cancer.] One interpretation of this seeming contradiction is that deficiencies in the repair of genes that are normally inactive, e.g., XP, leads to skin cancer, whereas deficient repair of active genes, as in Cockayne's syndrome, is related to developmental and degenerative effects.

Non-Melanoma Skin Cancer

There are two main types of non-melanoma skin cancer (NMSC): basal cell carcinoma (BCC) and squamous cell carcinoma (SCC). In most countries, reporting of these tumors to a cancer registry is either not required or inadequately standardized, resulting in inadequate cancer registry data and an inability to track trends in incidence or mortality. Where good medical care is available, the overall mortality is thought to be less than 1 percent. Although BCC typically represents 80 percent of NMSC, the mortality is mainly due to SCC.

SCC has a convincing and clear-cut relationship to UV-B radiation, whereas that for BCC is somewhat less compelling. The causative role of sunlight in SCC is supported by the observations that a) SCC occurs predominantly on the most sun-exposed parts of the skin, face, neck and hands; and b) in comparable populations the incidence of SCC is highest in geographic areas with the most sunlight. The observation that SCC occurs predominantly in fair-skinned people is consistent with this conclusion. The risk of developing SCC appears strongly related to the total dose of sunlight received in the course of a lifetime. In addition, in animal models, exposure to UV radiation is associated with the development of SCC and not BCC.

The evidence is less clear-cut with BCC, which rarely appears on the well-exposed backs of the hands, but instead occurs more on the face and neck areas, with a fair percentage developing on the trunk. Also, a person's most recent history of sun exposure (the preceding 10 to 20 years) relates better to the risk of developing SCC than BCC. [Vitasa, et al., 1990]. This may be interpreted to mean that UV radiation is somehow related to an "early event" in the development of BCC, after which other "events" (including growth) must occur before the tumor develops.

FIGURE
Fig. 2.3. Action spectra for cancer in animal models: Melanoma in fish [Setlow et al., 1993], non-melanoma skin cancer in mice -SCUP-m [DeGruijl et al., 1993], and in humans -SCUP-h [DeGruijl and van der Leun, 1994]

Important molecular evidence for the role of UV-B in the induction of these tumors is that large percentages (>50 percent) of SCC and BCC in humans bear UV-specific mutations (i.e., at dipyrimidine sites where cytosine is replaced by thymine, a C-to-T transition) in their p53 tumor suppressor gene. This constitutes the most direct evidence that UV radiation causes skin cancer in humans. Furthermore, it was found that in mice these types of mutations already exist in the precursor lesions of SCC, implying that UV irradiation can be an early event in the development of the tumor. Strikingly, it has recently been shown that certain UV-related mutations (tandem transitions of CC-to-TT) can be detected in sun-exposed skin of skin cancer patients (17/24 in Australia), but are virtually non-existent in unexposed skin (1/20). [Nagazawa et al., 1994].

Information on the wavelength dependence of the UV induction of cancer is crucial for quantitative risk assessments; however, it would be both impractical and unethical to derive such information from experiments in humans. Several groups have developed such action spectra based on tumorigenesis in hairless mice, with the most complete spectrum being that published by De Gruijl and his colleagues [1993], and presented as the Skin Cancer Utrecht-Philadelphia (SCUP) action spectra (see Figure 2.3). The carcinogenicity for wavelengths over 340 nm was somewhat higher than expected from constructed action spectra (mutagenicity corrected for epidermal transmission), which could be due to an increased level of indirect, radical-mediated damage in vivo. More recently, the SCUP action spectrum has been corrected for differences in epidermal transmission between mouse and human, thus a SCUP-h was derived from the SCUP-m action spectrum (`h' standing for human and `m' for murine). [De Gruijl and van der Leun, 1994]. This correction must be considered to represent a crude average, because the epidermal transmission will vary both in human and mouse from one individual to another, and it will decrease under regular UV exposure.

In relation to a depletion of the ozone layer, it is important to quantify how much more carcinogenic UV radiation reaches the ground level for each percent decrease in ozone. For NMSC, annual doses are assumed to be an appropriate measure, and personal doses are assumed to be proportional to ambient doses. This increase in carcinogenic dose is expressed by the Radiation Amplification Factor (RAF), which equals 1.4 for SCUP-m-weighted UV doses and 1.2 for SCUP-h-weighted UV doses (i.e., 1.4 or 1.2 percent more carcinogenic UV radiation for each percentage decrease in ozone; see also Chapter 1). Next, the relationship between skin cancer incidence and an increase in carcinogenic UV radiation must be determined. For stationary situations, this is estimated from incidence data in comparable populations living at different geographical locations under different levels of solar UV exposure comparable in genetic composition, lifestyle, etc.). Estimates for the white population in the United States predict that for every 1 percent increase in annual carcinogenic UV radiation, the SCC incidence over a human lifetime will ultimately rise by 2.5+/-0.7 percent, and the BCC by 1.4+/-0.4 percent, based on SCUP-h (for SCUP-m-weighted doses, the numbers are virtually the same). These latter numbers are referred to as the Biological Amplification Factors (BAF). Multiplying the RAF and BAF gives the overall Amplification Factor (AF), the ultimate, predicted increase in skin cancer for each percent decrease in ozone. For SCC, AF equals 3.0+/-0.8 percent; for BCC, AF is 1.7+/-0.5 percent, and for all nonmelanoma skin cancers combined, AF is 2.0+/-0.5 percent. With approximately 1.2 million new cases each year worldwide [Longstreth et al., 1991], this would amount to 250,000 additional cases each year from a sustained 10 percent decrease in average ozone concentration [for more detail, see Madronich and De Gruijl, 1993].

It is also worth noting here that there are subpopulations of individiuals at very high risk of developing NMSC, who will be greatly affected by an increase in ambient UV-B radiation. As noted above, persons with the rare genetic disorder XP are exquisitely susceptible to the development of UV-induced NMSC early in life and often die of this disease. A much larger population at high risk are renal allograft recipients and other immunosuppressed individuals, who tend to develop multiple, aggressive, and often fatal SCC on sun-exposed skin at a relatively young age. Increased UV-B radiation is likely to be especially hazardous for such high-risk persons [Glover et al., 1994].

In reality, the projections suggest not a sustained decrease, but rather a transient seasonal decrease in ozone that will vary substantially over the globe, reaching its maximum in early spring at the poles. The concentration of ozone in the stratosphere is projected to reach a minimum around the year 1998, and will normalize to 1980 levels around the middle of the next century (see Chapter 1). Increases in UV-B associated with these ozone losses are likely to cause a delayed transient increase in skin cancer incidence. Besides knowledge of the UV dose dependency, information on the time dependency of the response after changes in the ambient UV load is also needed. Such information is not available for human populations, and additional hypotheses must be introduced to extend the model to time-dependent responses. To a certain extent, animal experiments can assist in providing this extension of the model [De Gruijl and van der Leun, 1991], however, certain assumptions as to how the animal data translate to human responses are still required. Although more work should be done in this area, a plausible model estimates that a steady increase in nonmelanoma skin cancers will occur even under the most recent international agreement with regard to the phase-out of ozone-depleting substances, reaching a 25 percent higher level in the year 2050 in comparison to l980 at approximately 50 degrees NL [Slaper et al., 1992] see Figure 2.4.

Melanoma

Cutaneous melanoma (CM) is the result of the neoplastic transformation of melanocytes, the pigment-producing cells in mammalian epidermis. There are four different categories of CM in humans: 1) superficial spreading melanoma (SSM) 2) nodular melanoma (NM) 3) lentigo maligna melanoma (LMM) also known as Hutchinson's melanotic freckle) and 4) unclassified melanoma.

The etiology of LMM appears to be similar to that of NMSC in relation to sun exposure, whereas that of SSM and NM appears to be different in that it appears linked more with intermittent, intense exposures (i.e., severe sunburns) and/or exposures in childhood (for a comprehensive overview, see IARC, 1992; USEPA, l987; IBMC, 1992; or Elwood, 1993).

FIGURE
Fig. 2.4. Estimated increase in NMSC incidence under 3 CFC-phaseout scenarios. [Slaper et al., 1993]

In the case of SSM and NM, evidence supporting an etiologie relationship with solar (UV) radiation includes:

  1. the fact that in general, people light-skinned individuals who are more sensitive to sunburn run a higher risk; in such individuals irregularities in pigmentation, i.e., freckles, a high number of moles (nevi), and/or atypical nevi, contribute especially strongly to the risk.

  2. high levels of solar exposure during childhood (before ages of 15-20 years, and possibly especially intense intermittent exposures) is strongly associated with increased risk (e.g., Holman and Armstrong, l984), and increases the number of nevi (potential precursor lesions) that a person develops (Gallagher et al., 1990).

  3. the risk is often higher for indoor workers than for outdoor workers, and can be associated with (intense) intermittent exposure to the sun in leisure time.

  4. positive correlations between ambient UV levels (e.g., latitudinal gradients) and CM in sensitive populations.

  5. high risk in XP patients who are deficient in repairing DNA damage inflicted by UV radiation (see 2.5.1).

  6. mutations in the N-ras gene at di-pyrimidine sites (the target sites of UV-B radiation in DNA) in SSM and NM occurring in sun-exposed skin (Van't Veer et al., l989).

The hypothesis that intermittent UV exposure is important for the risk of CM was largely inferred from the locations of CM over the body (relatively large numbers on irregularly exposed sites) and more occurences in indoor workers than in outdoor workers. Later confirmation of this hypothesis is, however, solely based on studies in which patients and control groups were retrospectively interviewed; thus, these studies are vulnerable to recall-bias--after having contracted a CM, people tend to remember more sunburns than before [Weinstock et al., 1991].

In the U.S., incidence rates for CM among white-skinned populations during the decade from 1974 to 1986 increased at an average yearly rate of 34 percent (varying from -2 to -7 1/4 %. Increases in mortality during that time period showed a similar trend, although a slower rate. More recently, Scotto et al. [1991] have analyzed trends in skin melanoma death rates by cohort for fair-skinned ("white") males and females over a 35-year period (1950- 1984) and observed upward trends for older men and women (over 40) and downward trends for the younger cohorts. Assuming no life-style changes and constant UV radiation levels, these authors project that the 2 to 3 percent upward trend in mortality per annum observed since 1950 will discontinue and bend downward by the second decade of the 21st century.

This information is critical to assessing the risks of stratospheric ozone depletion, which would clearly need to incorporate cohort data and age- specific trend analyses into the baseline data used in such an activity. This same sort of cohort analysis information is also critical to estimating the potential increases in non-melanoma skin cancer; unfortunately most countries are not collecting sufficient data on non-melanoma skin cancer to be able to perform such trend and cohort analyses.

Animal Models for Melanoma

In the past few years, several animal models have been developed in which primary cutaneous melanomas can be induced by UV radiation alone or in combination with cancer-inducing chemicals. The most interesting of these involves the induction of melanoma by UV-B radiation in Monodelphis domestica, the grey, short-tailed, South American opossum [Ley et al., 1989]; see below.

Another animal model in which melanomas can be induced with UV radiation is the tropical fish model of Setlow [Setlow et al., 1989]; see below.

As described in the 1991 UNEP panel report [Longstreth et al., 1991], primary melanomas can be induced in mice using a combination of UV radiation and chemicals, but UV radiation alone has so far been ineffective [Romerdahl et al., 1989]. These and more recent studies [Donawho and Kripke, 1991; Hasan et al., 1992] indicate that UV radiation may play several different roles in the induction of murine melanomas, including that of an initiator, a tumor promoter, and a co-carcinogen that contributes to melanoma development by means of its immunosuppressive effects.

The animal models are instructive because they demonstrate that UV radiation can contribute to the induction of melanotic tumors in a variety of ways. However, they are not very helpful for assessing increases in melanoma incidence in humans exposed to increased UV-B radiation. In the opossum model, the dose-response and wavelength data are inadequate for making such calculations, and extrapolation of the information from the fish model to humans is obviously fraught with difficulties.

[Back to top of page]


Wavelength Dependency of Melanoma

Animal experiments have not yet yielded results that unambiguously indicate mechanism(s) by which the UV radiation may be causing CM in humans. The wavelength dependence of the induction of CM is important for risk assessments, but conclusive data are not available, and the wavelength dependence cannot be confidently constructed from presumed mechanisms.

In hybrid fish of the genus Xiphophorus a single, early in life UV-A exposure is quite efficient in evoking melanomas; a UV-B exposure is only 10 to 50 times more effective (per J/m2, i.e., this action spectrum is relatively flat when compared with the SCUP action spectra, see Figure 2.3; Setlow et al., 1993). These experiments were, however, contaminated by the aquarium lighting. This lighting contributed to a high background occurrence of melanomas (in about 25% of the fish), and possibly even counteracted the induction of melanomas by UV-B radiation the fish have photolyase). In the opossum Monodelphis domestica, melanomas can be induced by chronic broadband combined UV-B/UV-A exposure; such tumors do not develop when the animals are kept solely under yellow light. Visible light exposure after each UV exposure counteracts the development of the UV-induced melanomas, which would indicate that UV-induced di-pyrimidine dimers cause melanomas. If these dimers are the main DNA lesions causing mela-nomas in human skin, then the reversal by visible light is not likely to occur (humans appear not to have photolyase, [Li et al., 1993] although this is a matter of some controversy). The action spectrum in humans could then follow the induction of these DNA lesions in the skin, and would presumably resemble the SCUP action spectra more than the one found for the melanomas in fish (for comparison see Figure 2.3).

It is conceivable that UV radiation may contribute in various ways to the induction of melanomas, and that the specific mechanisms differ in the two animal models. Although it is difficult to induce melanoma in mice by UV irradiation, it can be done quite efficiently with exposure to chemical carcinogens, and concomitant UV exposure can then promote the melanomagenesis.

How these experimental data should be extrapolated to humans is, of course, very much an open question. CM in humans may well have a multifactorial etiology. Although UV radiation is likely to play a dominant role, (e.g., initiating precursor lesions during youth and suppressing immunity to the tumor cells as a result of a sunburn in the final stage of tumor development), other factors may affect expression of the UV effect.

[Back to top of page]


Related Issues

The majority of this chapter has been devoted to a review of the potential human health effect of the increased ultraviolet radiation which may result from stratospheric ozone depletion. There are, however, other health issues associated with this stratospheric ozone depletion. This section addresses four such issues. First, animal populations beyond humans are potentially affected by increased UVR which can lead to impacts on agricultural productivity. Second, the effects on human health of increased UVR, can be potentially mitigated by appropriate changes in behavior. Third, although most ozone depleting substances (ODSs) are scheduled for complete phase out by 1995, methyl bromide is not yet subject to complete phase out. Its toxicity is of some concern, so has been included in this dicussion. Fourth, the chemicals which replace ODSs may themselves have toxicity, so a brief synopsis of information known about them is also presented.

Effects on Animals

Animals of several species develop skin cancer, mainly SCC, in sparsely haired, light-colored areas of the skin. This includes cows, goats, sheep, cats and dogs [Emmett, l973; Dorn et al., l971; Nikula et al., 1992]. The body distribution is consistent with sunlight as the etiologic agent; UV-B is implicated by extrapolation from studies with laboratory animals. Cancers of the eye also occur in many animal species, including horses, sheep, swine, cats, and dogs, and are particularly frequent in cattle [Hargis, l981]. In rats, mice, hamster, and the opossum (Monodelphis domesticus), studies on the induction of skin cancer with mainly UV-B radiation have sometimes induced ocular tumors as well [Ley et al., l989]; [Blum, l943]; [Freeman and Knox, l964]. Photokeratitis and cataract have been induced experimentally in rabbits, with the most effective wavelengths falling mainly within the UV-B range [Pitts et al., l977]. Bovine infectious keratoconjunctivitis, an eye infection caused by the bacterium Moraxella bovis, is triggered and aggravated by UV-B irradiation of the eye [Hughes, et al., l965].

In animals in which these effects occur under natural conditions, an increase in UV-B irradiation would be expected to exacerbate them. However, it is not possible to estimate the magnitude of such effects because of the paucity of information on dose-response and wavelength dependence and on possible behaviorial modifications.

Mitigation/Amelioration

The effects of increased UV-B exposure on human health may, in principle, be mitigated by reducing the exposure time to sunlight. Being indoors gives a practically complete protection. Limiting exposure during the hours of maximal UV-B irradiance, that is, between 2 hours before and 2 hours after solar noon is especially effective.

Even while outdoors, there are still possibilities for protection. An important protection is offered by constitutional skin pigmentation. Dark-skinned people are better protected against skin cancer and sunburn than light-skinned individuals, but their pigment does not seem to protect them against some of the suppressive effects of UV-B radiation on the immune system.

Changes in behavior also offer many possibilities for mitigation. Many clothing fabrics give good protection. Hats are especially effective because they offer some protection of the facial skin and the eyes, two body sites at comparatively high risk. Sun glasses of appropriate material also provide good protection for the eyes. Being in the shade of buildings or trees also reduces the UV-B dose received.

Sunscreens may be a useful addition, especially for occasions when an unusually high exposure is expected, such as a holiday in a sunnier area. Sunscreens are effective against sunburn. To some extent, they may also be protective against skin cancer, as long as they are not used to prolong the exposure. Several experimental results suggest that the effectiveness of sunscreens against the effects of UV-B radiation on the immune system is limited. Continuous use of sunscreens with a high protection factor is sometimes advocated, but it may be counterproductive; such "sunblocks" also block both the formation of Vitamin D3 in the skin and the body's own defense systems, such as adaptation to the UV environment by thickening of the outer skin layers and tanning.

Methyl Bromide

Methyl bromide (MeBr) is discussed here because of its potential contribution to ozone depletion and its adverse effects on living organisms. Despite its ozone-depletion potential, current revisions to the Montreal Protocol do not require phase-out of MeBr, rather production is limited to the amount produced in 1991. A variety of countries have taken unilateral action which range in level of severity; for example, the Netherlands banned all soil uses in 1992, Denmark and Italy plan a total phase out in 1998 and 2000, respectively, the US will ban production and importation in January 2001, and Canada plans a 25% reduction in 1998 [US EPA, 1994].

MeBr is a colorless, generally odorless, ozone-depleting gas which is widely used for fumigation of soil (77%), commodities and quarantine facilities (12%) and sub- surface structures (5%). Current uses result in significant releases to the atmosphere. Methyl bromide is toxic by all routes of exposure. Immediate effects from acute exposures show a dose-dependent increase in severity ranging from dizziness, headache, nausea, contact burns to the eyes and skin, respiratory irritation, venticular fibrillation, pulmonary edema (sometimes delayed for several days), convulsions, coma and death. The Time-Weighted Average occupational limit recommended in the United States by the American Conference of Government Industrial Hygienists is 5 ppm. Signs of toxicity in humans following chronic exposure have included persistent numbness in the hands and legs, impaired superficial sensation, muscle weakness, unsteadiness of gait, and depressed or absent distal tendon reflexes; however, once exposure has ceased, symptoms generally disappear and recovery is complete [HSDB 1994].

Animals show similar signs of toxicity. Studies on horses, goats, cows and cattle fed hay or food contaminated with various levels of MeBr (ranging from 170 to 8400 ppm) displayed symptoms which ranged in a dose-dependent fashion and included difficulty in walking, incoordinated movement and gait, listlessness, inability to rise, and even death. MeBr is also toxic to aquatic species with acute and chronic toxicity to salt water species occurring at levels as low as 11 and 6.4 mg/l, respectively. A similar level for acute toxicity was found for fresh water species [HSDB 1994].

Toxicity of CFC Substitutes

As the production of fully halogenated chlorofluorocarbons (CFCs) ceases pursuant to the provisions of the Montreal Protocol and the London and Copenhagen Amendments, a variety of potential CFC substitutes are likely to be introduced into the environment. The toxicity of these chemicals is of interest: clearly, we do not want to introduce chemicals as replacements which are more problematic than those we are replacing. Table 2.1 summarizes what is currently known about the toxicology of the more common CFC substitutes: hydrofluorocarbons (HFCs) and hydrochlorofluorocarbons (HCFCs), and terpenes. Most of this information has been drawn from references found either in the open literature or submitted to EPA under the requirements of TSCA. In those cases with a sufficient database, EPA has derived a Reference Concentration (RfC), which is the concentration, at which exposure for a lifetime should be without adverse effect. The process of deriving these numbers basically involves selection of a No Observed Adverse Effect Level (NOAEL) and dividing it by an uncertainty factor that reflects differences in individual sensitivity, differences in species responsiveness, and inadequacies in the database (such as the lack of a two-generation reproduction study or a Lowest Observed Adverse Effect Level [LOAEL]).

Several chemicals have been the focus of extensive animal studies. The conclusions of such studies can promote a basis for determining the adverse human effects of these compounds. For example, HCFC-141b was determined to decrease reproductive performance; HCFC-124 has had transient central nervous system (CNS) effects; and HCFC-22 exposure has increased liver, kidney, adrenal, and pituitary weights. On the other hand, HFC-134a, HCFC-142b, and HCFC-152a have shown no signs of adverse effects.

TABLE 2.1. Update on Potential CFC Substitutes
CHEMICAL ADVERSE EFFECTS
according to the EPA documentation and review of the inhalation RfC
d-Limonene
CASRN: 5989-27-5
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). The compound is listed on FDA's Generally Regarded as Safe (GRAS) list and is approved for use as a food additive (Opdyke 1975). A chronic bioassay involving exposure via oral administration resulted in kidney damage in the male rats (NTP 1990); however, the mechanism involves alpha2uglobulin which has been judged an inappropriate endpoint for effects occurring in humans (USEPA 1991)
alpha-Pinene
CASRN: 80-56-8
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). The compound is listed on FDA's Generally Regarded as Safe (GRAS) list and is approved for use as a food additive (Opdyke 1975). EPA has judged that there are no chronic studies by inhalation or oral routes of exposure which are adequate to the development of an RfC. Contact hypersensitivity studies in humans suggest that this compound is a contact allergen. (Cachao et al. 1986)
beta-Pinene
CASRN: 127-9193
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). The compound is listed on FDA's Generally Regarded as Safe (GRAS) list and is approved for use as a food additive (Opdyke 1975). EPA has judged that there are no chronic studies by inhalation or oral routes of exposure which are adequate to the development of an RfC. Contact hypersensitivity studies in humans suggest that this compound is a contact allergen. (Cachao et al. 1986)
1,1-Dichloro-2,2,2-
trifluoroethane
HCFC-123
CASRN: 306-83-2
EPA reviewed the data for this chemical and derived a verified inhalation reference concentration (RfC) of 10 mg/cu m (1.6 ppm)(EPA1992) based on a chronic inhalation bioassay in which rats were administered the compound at 0, 300, 1,000 or 5,000 ppm 6 hours/day 5 days/week for 2 years (Malley 1990). The study identified a NOAEL (1,000 ppm) and a LOAEL (5,000 ppm) with the critical effects being increased relative liver weight, focal histopathology in the liver, an effect on lipid metabolism, and at high concentrations (5,000 ppm), CNS depression. The European Center for Ecotoxicology and Toxicology has completed a review of this chemical (ECETOC 1994).
Chlorodifluoromethan e
HCFC-22
CASRN: 74-45-6
EPA reviewed the data for this chemical and derived a verified inhalation reference concentration (RfC) of 50 mg/cu m (14 ppm)(EPA1992; EPA1994) based on a chronic inhalation bioassay in which rats were administered the compound at 0, 1,000, 10,000 or 50,000 ppm 5 hours/day 5 days/week for slightly more than 2 years (Tinston et al. 1981). The study identified a NOAEL (10,000 ppm) and a LOAEL (50,000) with the critical effects being increased liver, kidney, adrenal and pituitary weights.
2-Chloro-1,1,1,2-
tetrafluoroethane
HCFC-124
CASRN: 2837-89-0
EPA reviewed the data for this chemical and derived a verified inhalation reference concentration (RfC) of 300 mg/cu m (143 ppm)(EPA1992; EPA1994) based on a subchronic inhalation bioassay in which rats were administered the compound at 0, 5,000, 15,000 or 50,000 ppm 6 hours/day, 5 days/week for 90 days (Malley 1991). The study identified a NOAEL (15,000 ppm) and a LOAEL (50,000 ppm) with the critical effects being transient CNS effects. The European Center for Ecotoxicology and Toxicology has completed a recent review of this chemical (ECETOC 1994).
1,1,1,2-Tetrafluoro-ethane
HFC-134a
CASRN: 811-97-2
EPA reviewed the data for this chemical and derived a verified inhalation reference concentration (RfC) of 100 mg/cu m (71 ppm)(EPA1992; EPA1994) based on a chronic inhalation bioassay in which rats were administered the compound at 0, 2,500, 10,000 or 50,000 ppm 6 hours/day, 5 days/week for 2 years (Hext and Mould 1991). The study identified a NOAEL (50,000 ppm) but not a LOAEL.
1,1-Dichloro-1-fluoroethane
HCFC-141b
CASRN: 1717-00-6
EPA reviewed the data for this chemical and derived a verified inhalation reference concentration (RfC) of 100 mg/cu m (21 ppm)(EPA1992; EPA1994) based on the interim report of a 2 generation reproduction study in which rats were administered the compound by inhalation at 0, 2,000, 8,000 or 20,000 ppm 6 hours/day, 7 days/week for 10 weeks prior to first mating, unexposed for a short interval around parturition, then exposed with the same regiment through a second mating and delivery until the F1B offspring were 4 days old (Brooker et al. 1992). The study defined a NOAEL (8,000 ppm) and a LOAEL (20,000 ppm) with the critical effect being decreased reproductive performance.
1-Chloro-1,1-difluoroethane
HCFC-142b
CASRN: 75-68-3
EPA reviewed the data for this chemical and derived a verified inhalation reference concentration (RfC) of 50 mg/cu m (12 ppm)(EPA1992; EPA1994) based on a chronic inhalation bioassay in which rats were administered the compound at 0, 1,000, 10,000 or 20,000 ppm 6 hours/day, 5 days/week for 80 weeks (Seckar et al. 1986). The study identified a NOAEL (20,000 ppm) but not a LOAEL.
1 ,1-Difluoroethane
HCFC-152a
CASRN 75-37-6
EPA reviewed the data for this chemical and derived a verified inhalation reference concentration (RfC) of 40 mg/cu m (15 ppm)(EPA1992; EPA1994) based on a chronic inhalation bioassay in which rats were administered the compound at 0, 2,000, 10,000 or 25,000 ppm 6 hours/day, 5 days/week for 2 years (McAlack and Schneider 1992). The study identified a NOAEL (25,000 ppm) but not a LOAEL.
Pentafluoroethane
HFC-125
CASRN: 354-33-6
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). The acute inhalation toxicity of pentafluoroethane is very low with a 4 hour LC50 of 709,000 ppm (Panepinto 1990). Developmental toxicity studies in rats and rabbits have revealed no toxicity at 15,000 ppm and only slight anesthetic effects in rats at 50,000 ppm (Masters et al. 1992; Masters et al. 1992) The compound is a cardiac sensitizer in dogs at concentration above 100,000 ppm (Hardy et al. 1992). The European Center for Ecotoxicology and Toxicology of Chemicals has completed a recent review of this chemnical (ECETOC 1994).
1,3-Dichloro-1,1,2,2,3-
HCFC 225cb
CASRN: 507-55-1
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to pentafluoropropane develop an inhalation reference concentration (RfC) have not been met (EPA 1993). Several short-term inhalation studies either of HCFC-225cb in rats at concentrations as high as 41,216 ppm for four hours (Jackson et al. 1992) or of the mixed isomers (HCFC-225cb;HCFC-225ca) in mice at concentrations as high as 13,000 ppm for four weeks (Frame et al. 1992) induced changes in lipid and carbohydrate metabolism at several concentrations. However, these effects were reversible after a 14 day recovery period.
1,1-Dichlor- 2,2,3,3,3-
pentafluoropropane
HCFC 225ca
CASRN: 507-55-1
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). Several short-term inhalation studies either of HCFC-225cb in rats at concentrations as high as 46,527 ppm for four hours (Jackson et al. 1992) or of the mixed isomers (HCFC-225cb;HCFC-225ca) in mice at concentrations as high as 13,000 ppm for four weeks (Frame et al. 1992) induced changes in lipid and carbohydrate metabolism at several concentrations. However, these effects were reversible after a 14 day recovery period.
Trifluoromethane
HFC-23
CASRN: 75-46-7
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). A subchronic inhalation study in rats at 10,000 ppm or dogs at 5,000 ppm 6 hours/day 7 days/week for 90 days showed no adverse effects in either species (Leuschner et al. 1983). Cardiac sensitization testing in dogs at concentrations as high as 300,000 ppm revealed no activity (Hardy 1992)
Difluoromethane
R-32; FC-32
CASRN: 75-10-5
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). Exposure of rats to 200,000 ppm 6 hours/day, 5 days/week for 2 weeks revealed no treatment related effects (Moore 1976). Cardiac sensitization testing in dogs revealed that rare individuals were sensitive at concentrations of 250,000 (1/12) (Mullin 1993).
1,1,1-Trifluoroethane
HCFC 143a
CASRN: 420-46-2
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). Exposure of rats to 1, 2,000, 10,000, or 39,000 ppm 6 hours/day 5 days/week for 4 weeks showed testicular effects in animal exposed nose-only but no effects in animals given whole body inhalation exposures leading to the conclusions that the finding with nose-only exposures may be related to stress (Warheit et al. 1991, 1992)
1,1,1,2,3,3,3-
Heptafluoropropane
(HCF-227ea)
CASRN: 43-89-0
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). No data were found.
Perfluorobutane
FC-3-1-10
CASRN: 355-35-9
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). A subchronic (90 day) inhalation exposure of rats to 0, 5,000, 15,000 or 50,000 ppm 6 hours/day, 5 days/week induced no exposure related adverse effects (Kenny et al. 1992) suggesting that 50,000 ppm is a NOAEL for this compound. Cardiac sensitization testing in dogs at concentrations as high as 170,000 ppm reveal no activity as a sensitizer (Hardy and Kieran 1992)
Perfl uorohexane
FC-5-1-14
CASRN: 355-42-0
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). Acute exposures as high as 79,000 ppm for 4 hours were not lethal to rats (Jackson et al. 1992) nor is the compound a cardiac sensitizer in dogs at levels as high as 400,000 ppm (Hardy and Kieran 1992).
Bromodifluoromethane
FM-100
CASRN: 1511-62-2
EPA reviewed the data for this chemical and determined that the requirements for the minimal database necessary to develop an inhalation reference concentration (RfC) have not been met (EPA 1993). However, in a developmental toxicity study of rats exposed to 0, 1,000, 4,000, or 10000 ppm 6 hours/day on gestation days 6-15, a NOAEL of 4,000 and a LOAEL of 10,000 ppm were observed for both developmental and maternal toxicity (Nemec 1991b). A similar NOAEL for rabbits can be derived from Nemec (1991c). Cardiac sensitization potential testing in dogs revealed a NOAEL of 3,000 ppm.

[Back to top of page]

References


TOC Table of
Contents
TOC UNEP Ozone
Secretariat
HOME Ozone/Health
Home Page