sign in sign up
<<
Home , Hormesis, Radiation hormesis

Radiation

Roger M. Macklis and Beverly Beresford

Joint Centerfor Therapy, Department ofRadiation Therapy, Harvard Medical School, Boston, Massachusetts

beneficial level of exposure to has “Radiationhormesis―is the name given to the putative major economic and epidemiological implications (4). stimulatoryeffects of low levelionizingradiation(generally Increasing disenchantment with the dependence on fos in the range of 1—50cGy of low-LET radiation). Based on sil fuel energy and recent advances in historical and pharmacologic principles reminiscent of technology are forcing a reassessment ofthe risk/benefit some of the major tenets of , most of these equation for nuclear energy (5,6). In addition, it is now effects are now generally ascribed to protective feedback apparent that the major source of radiation exposure in systems that, upon exposure to low concentrationsof our society comes not from man-made sources but from toxins, proceed to stimulate metabolic detoxification and the inhalation of natural -222 and its decay repair networks. The activation of these networks may then result in net beneficialeffects on the cell, or daughters ( 7). The costs of effective radon abatement species. Discussions of possible stimulatory effects of low would be extraordinarily high and even partisans have levelsof ionizingradiationhave recently become entangled difficulty rallying support for the costly elimination of withthe separate but related question of whether a thresh this natural radioactive gas (8). Thus, the traditional old dose levelexists on the radiotoxicologic dose-response assumptions that (a) essentially all exposure to ionizing curve. This review summarizes some of the relevant his radiation is measurably detrimental and that (b) the torical and scientific data bearing on the question of radia health effects of low level exposure may be directly tionhormesis.We findthe data in supportof most of the inferred from linearly scaled deleterious high level ef hormesis postulates intriguing but inconclusive. fects, are both multi-billion dollar propositions that bear J NucIMed 1991;32:350-359. careful consideration (9).

HISTORICAL ROOTS OF

5 all nuclear radiation harmful? When this question The elucidation ofthe principles ofradioactive decay was posed by Hugh Henry (1) in 1961, the debate was and ionizing radiation by such pioneers as Bequerel, already over half a century old. Radiation hormesis Roentgen and the Curies captured the imagination of (Greek, “rapidmotion―from hormaein, “toexcite―)is the public. Early investigations into the physiologic a branch of a larger area of toxicologic inquiry that effects of ionizing radiation produced claims of dra traces its roots back to the principles of medicinal matic responses in patients suffering from many types homeopathy and the Arndt-Schultz conjecture: sub ofskin diseases and malignancies. Weart ( 10) has traced stances that inhibit biological processes at high levels the popular conceptions and portrayals of ionizing ra may be expected to stimulate them at lower levels (2). diation as they changed from the early, positive impres In the field of ionizing radiation research, the concept sions to the later, distinctly negative toxic associations. of hormesis is often taken to imply some element of At the turn of the century, radioactivity was portrayed physiologic benefit from low-LET radiation in the range as a vital force capable of producing profound feats of of 1—50cGy total absorbed dose. Data sets offered in rejuvenation. Even within the medical literature, adver support ofthe hormesis hypothesis run the gamut from tisements and testimonials to “quack―radioactive de biochemical to organismal to evolutionary, and lack of vices and nostrums abounded. The oral and parenteral consistency in the experimental endpoints used to use ofmicrogram quantities ofradium-226 and radium quantitate hormesis has contributed greatly to the lack 228 as treatments for various metabolic diseases such ofconsensus surrounding discussions ofthe topic (3). as diabetes, hypertension, infertility and impotence was The question of whether one can define a safe or referred to as “mildradium therapy,― a title meant to distinguish this branch of radiation medicine from the ReceivedOct. 1, 1990;revisionaccepted Oct. 20, 1990. high dose destructive treatments ofthe oncologists ( 11). Forreprintscontact:RogerMacklls,JointCenterforRadiationTherapy, Dept.ofRadiationTherapy,HarvardMedicalSchool,50 BinneySt., Boston, Mild radium therapy was considered an accepted mcd MA 02115. ical subspecialty at the interface of radiology and en

350 The Journal of Nuclear Medicine•Vol.32 •No. 2 •February 1991 docrinology, and thousands of patients were treated tacitly assumes that proportionally scaled high level annually in the years 1910—1925. exposures will accurately reflect the epidemiologic con The widespread use of radium compounds in the sequences of LLIR (27,28). production ofluminous paints for watch dials and other This is the heart of the modern radiation hormesis instrument display panels led to the recognition that controversy. Hormeticists argue that these scaled epi chronic ingestion of microgram amounts of such bone demiologic extrapolations are invalid and that the ma seeking isotopes led to osteoradionecrosis and osteosar jority of the available experimental data on LLIR do comas in significant numbers of factory workers cx not support a significantly adverse health effect; rather posed to these compounds (12). By the early l930s, much of the data suggest some theoretical and experi radium exposure was a well-recognized occupational mental “benefits―accruing from such exposures. This illness ( 13), though the public's perception of radium review will briefly outline some of the major lines of and ionizing radiation remained very positive. The argument and experimental data in the ongoing hor death of the Pittsburgh millionaire industrialist Eben mesis controversy. M. Byers, an internationally know sportsman and play boy who for several years had been taking a radioactive ThE SPECTRUM OF CLAIMS IN RADIATION over-the-counter rejuvenator called RADITHOR (14), HORMESIS finally succeeded in drawing public attention to the Many polemical discussions of radiation hormesis problem ofradium control and resulted in strict radium appear to have their roots in disagreements over ter control laws throughout the U. S. and Europe (I 1). minology (29). As the word was initially used, “hor This highly publicized incident sounded the death knell mesis―was a pharmacologic term implying a paradox for the mild radium therapy movement though the ical stimulatory effect at low concentrations of any public continued to regard nuclear technology as basi substance found to be inhibitory at high concentrations cally positive. (2). This definition was meant to be independent of With the development of the atomic bomb and the assay, endpoint, and health implications. As the mcdi consequent public realization of the potential for large cal, industrial, and military uses of nuclear energy and scale radioactive contamination, the perception of flu radioactive materials have become more highly politi clear energy as a powerful but essentially benign tech cized, much of the radiation hormesis controversy has nology began to change (10). Serious scientific inquiries shifted to disagreements over the positive or negative into potential new uses of ionizing radiation in agricul implications of the putative “stimulatory―effects of ture and biology slowed or stopped, and attention was LLIR for the ecosystem or for the population as a whole. refocused on the occult somatic and genetic effects of Thus, some hormesis investigators claim that the cvi ionizing radiation exposure (15). These investigations dence supports certain narrowly defined stimulatory brought nearly unanimous agreement from radiation effects of radiation without implying that such stimu researchers concerning the potentially serious detrimen latory effects are necessarily beneficial (26) or that the tal effects of cumulative absorbed radiation doses of 1 existence of such hormetic effects would exclude con Gy or more oflow-LET radiation, with correspondingly current detrimental effects within the same dose range. lower toxic thresholds for high-LET sources. Data on Other, more extreme hormeticists argue that LLIR is the health effects oflow-level ionizing radiation (LLIR) not only beneficial but may be absolutely essential for were much more equivocal, and the late biologic con life (30). The hormesis controversy is not one argument sequences of low-LET exposures in the range of 1—50 but many, and even the hormeticist camp is bitterly cGy were generally inferred from large epidemiologic divided over the medical implications of their conjec data sets such as those acquired from study of atomic tures. bomb survivors (16,1 7), nuclear industry workers (18, It is thus useful to divide the hormesis controversy 19), and medically exposed patient cohorts (20). In the into the evidence for and against specific claims that absence of specific evidence that a threshold existed for have been advanced within the hormesis rubric. The radiation damage, it was deemed prudent to assume majority ofthe hormesis data can be broken down into that any measurable dose of ionizing radiation was nine separate but interrelated claims that run the gamut potentially dangerous. This safety provision has led to from narrow biochemical observations to cellular phe divisive and acrimonious disputes between those radia nomenology to epidemiologic oncology (Table 1). tion workers who favor a strict no-threshold linear or For the purposes ofthis analysis, we will review these linear quadratic dose-effect toxicity curve, and those claims separately, concentrating on the data obtained who argue that in the absence of definitive data the over the last three decades. It should be understood that lower end of the dose-effect curve should be left unde not all hormesis proponents would agree with all of the fined (21—26).Most current estimates of the health above claims, and that most hormetic arguments are impact of populational exposures to LLIR continue to assumed to apply only to low-LET forms of radiation use a “riskper person per rem―extrapolation which (74).

Radiation Hormesis •Macklis and Beresford 351 TABLE 1 Proposed HormeticEffects of Low-LevelIonizingRadiation BiochemicalCellularOrganismalStimulates

inhibitoryDNAunscheduledStimulates immune responseSelectively inactivates repair(humoral and cellular)or senescent parts of orga nismsInduces intoxificationfree radicalde Functionsas a vitallifeforceDecreases cancer risk andrepairthat may even be essen chronically exposed popula systemstial requirement Functions as a general met Extendsaveragelife-spanin aboliccatalyst and fertility lightlyexposed populations enhancertions Produces evolutionarily favora ble SeleCtiOnpressure that benefits species as a whole

IonizingRadiationStimulatesUnscheduledDNA effect of LLIR and the radiation-related increase in Repair intracellular free concentration on thymidine Because this claim is one of the most narrowly de kinase activity, an indicator of DNA synthesis. They fined, it is also one of the most easily supportable of found that thymidine kinase activity was temporarily inhibited in murine bone marrow cells exposed to as the hormesis postulates. Wolff has shown evidence that little as several of cGy of cesium-l37 gamma irradia lymphocytes seem able to mount an adaptive response tion. This suppression reached its peak in 5 hr after to the challenge of LLIR. Cells allowed to proliferate in which activity gradually recovered to normal levels the presence of 0.0 1-0. 10 @Ci/ml3H-TdR (and thus within the next 24 hr. Biochemical maneuvers designed “primed―with the low-energy tritium emissions) ap peared to become less sensitive to chromatid damage to increase intracellular free radical concentrations also reproduced this suppressive effect. Inhibition of the by subsequent 150 cOy doses of X-rays (31). This effect thymidine kinase activity was thought to be accom has been attributed to the stimulation of a damage panied by an increase in the concentration of the free inducible repair system analogous to the “SOS―UV radical scavenger glutathione. inducible system found in prokaryotes (26,32). The Feinendegan interpreted these and related data as process takes @@-4—6hr to induce, persists for several cell evidence for a cellular control mechanism that func generations, and is blocked by the protein synthesis tions to temporarily inhibit DNA synthesis in order to inhibitor cycloheximide. Gel electrophoresis studies allow repair of lesions induced by free radical attack. using two-dimensional separations show new protein By temporarily inhibiting critical intracellular processes signals apparently induced by radiation exposure. The and by triggering production of free radical scavengers function and significance of these radiation-inducible in response to LLIR, this feedback mechanism would gene products are not yet established. theoretically confer some degree of protection against Wolffand his colleagues are careful to point out that the deleterious effects of future exposures to ionizing these data are at present only phenomena radiation (35) and other free radical challenges. and their physiologic significance (if any) are not clear. The interpretation of this set of experiments is corn However, the existence ofany reproducible well-defined plex. Critics have pointed out that the transient increase biochemical system induced in response to ionizing radiation (perhaps in response to the free radicals pro in intracellular free radical concentration after several duced during this process) represents an important con cGy of low-LET radiation would be barely detectable ceptual starting point for any claim of a damage-induc above the background level and would, thus, be a poor trigger mechanism for an important homeostatic feed ible feedback loop through which hormetic effects back loop (33). However, evidence is now accumulating might be achieved. Whether the transitory increase in that free radicals are involved in many detrimental the steady state concentration offree-radicals that might cellular processes including chronic diseases and senes be expected to occur after LLIR is enough to activate cence (36) and the putative hormetic radiation impact such a damage-inducible system is entirely conjectural on this feedback loop may be a consequence of a (33). serendipitous utilization of a defense mechanism that evolved in response to entirely different organismal Low-LevelIonizingRadiationInducesFree Radical insults. Detoxificationand RepairSystems Data in support of this contention have been derived Ionizing Radiation as an Immunostimulant from several sources, notably Feinendegan and his col Although high dose ionizing radiation has long been laborators (34). These workers have investigated the known as an immunosuppressant, some data suggest

352 The Journal of Nuclear Medicine •Vol. 32 •No. 2 •February 1991 that LLIR actually may be immunostimulatory. In of antibody production) were both significantly en 1909, Russ reported that mice treated with radiation hanced by single or continuous exposure to 2.5—7.5Gy appeared to have acquired some level of resistance ofgamma radiation. against bacterial disease (37), and Luckey collected The implication in all of this work is that LLIR may several dozen references supporting this contention in produce a beneficial overshoot in the potency of the his exhaustive review (30). Most of these studies are immune system during its recovery from radiation cx poorly designed and statistically unimpressive, though posure (41). This overshoot may initially take the form the data in aggregate are difficult to dismiss out of hand. of increased expression of certain hematopoietic cell Murphy and Morton performed early studies on the growth factors which then induce rapid recovery of effect of radiation on the growth rate of mouse tumors immunocompetence in the irradiated animals. These (38). They excised spontaneously occurring murine data are not inconsistent with the recent studies on the tumors, irradiated the mouse with an undefined “stim recovery of the immune system in bone marrow trans ulating―dose of X-rays from a Coolidge Tube, then plant patients undergoing TB! (42). These TB! patients regrafted the tumors into the groins of the irradiated often shown a supraphysiologic surge in “naturalkiller― animals. Of 29 non-irradiated animals, 28 showed tu (NK) cells during the period of recovery, and it is mor regrowth. In contrast, tumor failed to recur in 26 possible that these cells might have some anti-neoplastic of 52 irradiated mice. The tumors themselves showed potential. Whether radiation doses in the 1—50cGy no effect when irradiated with the same dose of X-rays. range are sufficient to trigger this response is not clear. No data are presented to control for the effects of handling, treatment-related inflammation, etc. Evalua Is IonizingRadiationa Vital Life Force? tion of the white blood cell counts in the irradiated This postulate probably embodies the most contro animals seemed to demonstrate a biphasic response versial premise in the hormesis field. The core of this an initial period ofleukopenia and immunosuppression hypothesis is the assertion that LLIR is a basic biophys followed by myeloid and lymphoid recovery with con ical phenomenon that has been present since the emer sequent reappearance of immunologic competence at gence of life and the beginning of evolution. Such a supranormal levels. proto-stimulus, the argument goes, may well have mi Metcalf attempted to give a physiologic explanation tiated physiologic processes that have become so inter for this recovery overshoot phenomenon by evaluating nalized in the cell life cycle that the deprivation of LLIR the effect of several Gy of total-body irradiation (TB!) would be subtly detrimental to the organism. Thus, on lymphocyte counts in C57BL/6 mice (39). Not only Hickey (4) speaks of a “hormeticdeficiency―implying did he confirm a significant lymphocytosis in the re that life forms deprived of ionizing radiation will show covering animals, he also found elevated levels of deleterious effects. thymic lymphocyte growth factors in the post-radiation Luckey has published a hypothetical “completedose period. Finally, Liu et al. (40) showed that the reactivity response curve―for ionizing radiation embodying this of thymocytes to interleukin-1 and the magnitude of conjecture (30) (Fig. 1). The curve suggests that optimal the “plaque-formingassay―(a gross measure ofthe level organismal proliferation takes place in the presence of some baseline amount of ionizing radiation and that doses substantially above or below this optimal radia 2(SPC@iSL @ I Cortt@ tion level result in suboptimal growth. Data in support

L@ of this dose-response relation are mainly derived from protozoan studies attempting to quantitate amoeba pro I liferation under conditions in which the are /4, I either exposed to LLIR or deprived of >95% of the @ 0(FIC1OC'@ . DUILJTAT$ON I I usual background dose of ionizing radiation by careful @ I I shielding and growth media screening. The data sug I gested an @-40%decrease in cell number after eight days I •@@@_AM$IU@1RADIATION @ @To,@ i&a of such radiation privation (43). This inhibition could ISURAMRIO4T) DOSE be reversed by the addition to the shielded amoeba FIGURE 1 colonies of radioactive in concentrations Luckey's hypothetical “complete―dose response curve for meant to mimic natural levels. chronicirradiationoforganisms.Thecurveis meantto indicate qualitative patterns of change in response to different doses Luckey defined a number of potential facets of the ofionizingradiation.Unitsarenotspecifiedandwoulddepend “radiation deficiency syndrome― (30)—low growth onthe experimentalsystem.The “zeroequivalentpoint―(ZEP) rates, slow development, decreased fecundity, poor indicatesa dose that elicitsa responseindistinguishablefrom health, slow metabolism, and decreased lifespan—while controls. The projection to the left of the ordinate is based upon exposure to protozoa, (Reprintedwithpermissionfrom acknowledging that evidence for many of these phe 1982;43:771—778. nomena are fragmentary.

RadiationHormesis•Macklisand Beresford 353 These postulates and experiments are intriguing but fertility increase. Higher, more cytotoxic doses of radia have been greeted by skepticism over experimental tion negate this beneficial effect. methodology. In addition, the relevance of such single Though the “biopositive―effects of radiation on fer cell organismal requirements to the metabolic needs of tility are still occasionally discussed (primarily in East more highly evolved multicellular organisms is not efli European journals) in reference to the radon spas, clear. In a sense, the hypothesis that some baseline level Broth has criticized these data as inadequate and mis of ionizing radiation is necessary for life has become leading (48). On balance, the data marshalled in sup the litmus test identifying the “true―hormeticists, as port of the claim that radiation acts as a nonspecific distinguished from those individuals who view back biophysical catalyst in higher organisms seem uncon ground levels of ionizing radiation as biophysical phe vincing. nomenon whose presence or absence is largely irrele vant to most physiologic processes. Ionizing Radiation Can Selectively Inactivate Senescent or Inhibitory Structures in Plants and Ionizing Radiation as a Metabolic Catalyst and Higher Organisms Fertility Enhancer This claim is really a mechanismal outgrowth of the The belief that LLIR is a general metabolic catalyst last conjecture. The underlying premise of this claim was one of the underlying tenets of the mild radium involves the putative existence within higher organisms therapy movement in the early decades of the century of specific “inhibitory―centers whose natural role is to (11). Thisbeliefwasoriginallybasedon a largebody conserve the organism's resources and prevent it from of circumstantial evidence concerning the salutary ef outgrowing available nutritional resources. If these in fects of mineral water from radon-rich natural springs hibitory centers are selectively destroyed by radiation, found in many of the European spa centers. Expen the organism may undergo a temporary increase in mental observations on the impact oflow level radium growth or fecundity. exposure on the rate of membrane ruffling and metab This explanation has been invoked most compel olism of human leucocytes and bacteria seemed to lingly as an explanation for the widely reported effect confirm this effect. The German physiologist George of LLIR on plant growth and fruiting patterns. Miller Wendt reported that vitamin-deprived rats could be and Miller have reviewed the data on this subject (49) temporarily rejuvenated by exposure to radium or high and find that most modern studies show a small (<10%) frequency electromagnetic radiation (44) and that increase in such parameters as branches, flowers, linear many cell types responded to such stimuli by “radi growth, and fruit yields after exposure to LLIR. Some otaxis,―or purposeful movement toward the radiation of the older data in the field may be suspect inasmuch source. as horticulture approaches involving LL!R treatment Luckey collected -@@40references suggesting that ver rather than crossbreeding of genetically determined tebrate growth and fecundity could be stimulated by crop traits were widely promoted in the Soviet Lysen LLIR (30). Some of these data are derived from acute koist literature. In addition, the effect seems quite dif exposure experiments and some from chronic, fraction ficult to reproduce, though some investigators with ated exposures on the order of 1 cGy per day or less. many years ofexperience apparently could demonstrate Despite the counter-intuitive ofthis claim, some the effect at will. The phenomenon apparently requires investigators seem to have gone to great pains to control a precise relationship between the radiation dose, tim for all the obvious confounding factors such as animal ing, nutritional status and species ofthe irradiated plant handling time, environment, etc., and continued to and it is doubtful that the small magnitude of the claim a small but significant stimulatory effect. An response would allow practical agricultural exploitation apparently increased rate of fertility associated with (75).Itisintriguingtonote,however,therecentreports LLIR is extensively documented with various strains of of Soviet physicians concerning the lush floral over insects and rodents, and the effect has been reported growth now taking place near Chernobyl. both after external irradiation and after injection of Conceptually, this growth effect has been attributed radiopharmaceuticals. to the selective destruction of the plant's apical meris Clinical data documenting such observations are tern tissues, a part of the organism that appears to sparse. Kaplan followed 644 women treated with 50— exercise control over lateral growth and fruition (49). 100 cGy (ovarian dose) of 200 KvP X-rays for infertil The “pinchingoff―ofthe terminal meristem tissues has ity, reporting that approximately half went on to con long been used by gardeners who wish to create fuller ceive (45). A plausible mechanism for this observation plants and more widely distributed fruiting and flow was reported by Meyer (46,47), who claimed that pre ering patterns in axial-dominant plants (a process that natal radiation kills some oocytes but stimulates the has given rise to the expression “greenthumb―), and it development of the remaining cells, producing a net is not unreasonable to suppose that LL!R or pharma

354 The Journal of Nuclear Medicine •Vol. 32 •No. 2 •February 1991 cologic interventions might reproduce this effect under the hazard function for appears to have certain circumstances. peaked for this cohort, while the risk of solid tumors is Kondo extends this concept of selective radiation continuing to rise. mediated destruction of inhibitory tissues to higher The Biological Effects of Ionizing Radiation (BEIR) organisms, proposing that the undifferentiated primor study group represents the most important U. S. panel dial cells may be the most sensitive to radiation effects of expert opinion on the subject of radiation toxicity because their response to LL!R involves a kind of and . Based in part on the RERF data, “altruisticcell suicide―(50). This theory involves the the BEIR V report (53) suggests that risks of radiation observation that certain cells undergo programmed cell carcinogenesis may be 3—4times greater than previously death () in response to LLIR (51). Kondo believed. This increase is due in part to the realization suggests that this behavior results in a proliferation of that previous neutron flux estimates for the Nagasaki healthy pluripotent stem cells to replace those destroyed bomb were too high, and in part to the panel's skepti by the radiation. The hormetic effect is produced when cism concerning the possibility of a threshold for radia the cellular proliferation more than compensates for tion carcinogenesis and the sparing effects of low dose those cells damaged by the radiation. rate exposures. Though their report makes extensive Kondo (50) attempts to distinguish this “altruistic use of the RERF data, it admits that the confidence cell death―from the “simpleradiation death―discussed bars on the relative risk estimates are too broad to by Miller(49) in reference to the plant menstem tissues. provide unambiguous information on whether radia Kondo believes that the meristem tissue destruction tion in the range of <25 cGy significantly increases (or results in a disinhibition of tissues, while the altruistic decreases) cancer risk. In addition, the atomic bomb cell death phenomenon relies on repopulation to pro data set represents the effects of acute exposure to duce the hormetic effect. In both cases, however, the gamma (and in the case of Hiroshima, neutron) irradi driving principle is the selective destruction of one cell ation and is consequently not an ideal model for the population in order to produce a proliferative effect on sort of chronic, fractionated LLIR that has been pro the organism as a whole. posed as the optimal hormetic regimen ( 77). Because large epidemiologic data sets such as the Low-LevelIonizingRadiationDecreasesCancer atomic bomb data clearly demonstrate radiation carci Risk nogenesis at high absorbed doses, the heart of the hor No aspect of radiation toxicity has received more mesis controversy with respect to cancer induction is attention than the relationship between radiation cx the issue of a threshold on the dose response curve. It posure and the subsequent risk of malignancy. It seems is beyond the scope ofthis review to reiterate the masses clear that radiation exposures at the level of 1 Gy or ofdata bearing on this question. However, the BEIR V more increase the relative risk of acute leukemia and report, unlike some of its predecessors, has come down other malignancies by factors of 2—5or more ( 17). The squarely on the side of a no-threshold interpretation of effect ofcumulative radiation doses ofless than 25 cGy the data, at least for the purposes of calculating cancer are less clear, and some have claimed that the error bars risks for an exposed population (65). The authors of on the relative risk analyses at this low dose are broad the report make clear the oversimplified nature of such enough that they do not exclude a negligible or even an an analysis, but critics have argued that the way in inverse relationship between radiation exposure and which the report is presented will overemphasize the subsequent risk ofcarcinogenesis (52). relevance of the data to the general population ( 77). The most extensive data set bearing on this issue is The hormetic argument that LLIR may actually de the information collected by the Radiation Effects Re crease cancer risk is based primarily on environmental search Foundation (RERF) on the atomic bomb survi and occupational exposure data. The High Background vors. The latest estimates of relative risk of developing Radiation Research Group has investigated thyroid malignant neoplasms reflect the recent changes in ra nodularity and cancer rates in the Guangdong Province diation for the explosions (1 7). For a group in China, an area with a natural radiation level ‘@‘3 of4l,7l9 total subjects with an average shielded Kerma times normal (54). For the years 1975—1978,the high dose of 30 cGy, the latest estimates conclude that about background area had a cancer mortality rate of 36.53 half of all leukemic deaths and -@-8%of the non-leu per 10(5) person-years, compared to 52.85 per iO@ kemic cancer deaths were related to the radiation cx person-years for the control area (4). In the U.S., the posure. These data obviously reflect a range of individ states with the three highest mean radon levels (Cob ual exposures, making exact dose-response extrapola rado, North Dakota, and Iowa) show a lung cancer tions difficult at these low doses. At an exposure level death rate of 41 per i0@,as compared to a rate of 66 of 1 Gy, the estimated relative risk of leukemia is per i0@in the 3 lowest radon level states (Delaware, approximately 5, while the relative risk of non-leukemic Louisiana, and California) (55). Recent analyses of cancer is approximately 1.5. With 45 yr of follow-up, cancer mortality data for workers at the Hanford, Oak

RadiationHormesis•MacklisandBeresford 355 Ridge, and Rocky Flats nuclear weapons facilities also references in which many organisms (primarily insects fail to show a statistically significant increased cancer and rodents) appear to have increased average lifespans risk for workers exposed to cumulative radiation doses (often by as much as 20%) after exposure to LLIR. on the order of 10 cGy (19). Because this effect was driven primarily by a decreased These data and analyses are all controversial, and incidence of early deaths rather than an increased me many have argued that the data are seriously limited by dian lifespan, he and others have concluded that this problems with cancer case records, population and effect may be related to the increased immunologic ambient radiation figures, limited study timespans, and protection against disease (as discussed above). Once confounding variables such as smoking and lifestyle again, most of the primary vertebrate data bearing on differences. Nevertheless, it seems clear that, whatever this question were acquired in poorly controlled studies the effects of LL!R at doses less than 25 cGy, their with relatively small numbers of animals and many epidemiologic consequences are generally not great potentially confounding variables. enough to be statistically detectable against the great Several sets ofepidemiobogic data have been analyzed force of cancer mortality currently observed in indus for their relevance to the question of ionizing radiation trialized societies (56). The animal data (30) offered in and its impact on . In 1956, Warren found support ofthe hormesis “cancerprotection―theory may that lifespans for radiologists were 5 yr shorter than be a reflection of small numbers and the publication those for physicians as a whole (62). However, correct bias toward interesting or desired results, but they cer ing for age, Seltzer et al. used the same data to show tainly appear to be inadequate to prove the point. that radiologists lived on average 2.5% longer than the Conceptually, one might imagine that if LLIR were average age-matched physician (63). Other attempts to able to induce some sort of free-radical protection study the subject have been similarly inconclusive. mechanism then the mutagenic risks of minute doses Recent attempts to evaluate the process of might be more than counterbalanced by the feedback at the cellular level have implicated the repression of protection (34). At present, however, this theory ap growth genes such as c-fos and the reciprocal activation pears unsubstantiated (57). ofantiproliferative genes such as Rb! (64). One might, With the recent report of 7—8-foldincreases in the therefore, invoke radiation as either a rate of leukemia and lymphoma in children of heavily cause or an inhibitor of the fundamental process of exposed Sellafield nuclear workers, the potential genetic genetic aging, though there are few direct data to sup consequences ofLLIR are being re-evaluated (58). This port either claim. Once again, it appears that the lack report is still very controversial and appears to contra of independent reproducible differences in lifespan be dict the accepted notion (derived primarily from the tween high natural background and low natural back atomic bomb data) that LLIR has very little effect on ground locations tends to rule out a very significant the unexposed Fl generation of exposed individuals positive or negative effect of LLIR. (59). As a consequence of the Sellafield report, major epidemiologic data sets are now being reinvestigated, and it seems clear that the question of the carcinogenic IonizingRadiationas an EvolutionaryDrive risks of fractionated, low-LET ionizing radiation has The hypothesis that LLIR produces a net positive not yet been answered. All available evidence concern populational effect by catalyzing evolutionary changes ing radiation in the range of 1—25cGy, however, points in species is supported by very little direct evidence. to an increase (or decrease) in relative cancer risk too Indeed, the Sellafield data notwithstanding, one of the close to negligible to allow direct detection in a cancer most intriguing aspects of clinical radiation late effects prone society (60). For purposes ofsafety, however, the studies has been the virtual absence of anticipated he community has adopted a conserv reditary changes in children of exposed individuals. ative approach that assumes that the upper level esti These data are in accord with the findings of Kaplan mates of carcinogenesis risk after LL!R exposure must (65) who reported no genetic effects in children and be taken as the standard (61,20). grandchildren of women irradiated for infertility. In contrast, Russel et al. found that rodents exposed Low-LevelIonizingRadiationExtendsAverage to LLIR do show increases in mutation frequency (66). Lifespans Wolff has argued that since most nonspontaneous mu Free radicals are thought by many to play an impor tations are not beneficial, any evidence for heritable tant role in aging and senescence (36) and one might, radiation effects should be construed as potentially therefore, conjecture that a feedback mechanism trig dangerous (26). Thus, while it is conceivable that LLIR gered by LLIR that functioned to either reduce intra did play a role in spontaneous evolution from single cellular free radical levels or that eliminated free-radical celled to higher organisms, it is unlikely that any further damaged cells might result in increased longevity. Luck benefits might accrue from continuing exposure of hu ey's review of this subject (30) includes several dozen mans, and this line of argument appears to be of only

356 The Journal of Nuclear Medicine •Vol. 32 •No. 2 •February 1991 theoretical importance since humans appear relatively tion it may be more readily accepted within the context insensitive to the genetic effects of radiation (60). of a broad-based inquiry into the physiologic responses to low levels of various classes of toxins ( 72,73). Until such time that the health risks are authoritatively re CONCLUSIONS evaluated, it seems reasonable to continue to be wary In reviewing the data on the health effects of LL!R, of the health risks oflow-level ionizing radiation ( 76). one is struck by the paucity of convincing data and by the lack of expert consensus concerning the health implications of what data are available. On balance, it ACKNOWLEDGMENTS appears that the large populational data sets such as the The authors thank John Humm, PhD, for helpful discus RERF data and the radon data do not support the sions. Supported in part by PHS grant R-29 CA49O17 from existence of a discernible hormetic effect in the range the National Cancer Institute and a Junior Faculty Research of 1—25cGy total absorbed dose, but neither do they Award from the American Cancer Society to RMM. argue strongly in favor ofclear cut deleterious effects of the magnitude generally assumed from a linear dose response model ( 77). Obviously, one must distinguish REFERENCES between the lack of evidence for a toxic effect and the presence of evidence for a beneficial effect (only the 1. Henry HF. Is all nuclear radiation harmful? JAMA latter would constitute true hormesis); but the available 196l;l76:671—675. data do suggest that at least some physiological proc 2. StebbingARD. Hormesis—thestimulation ofgrowth by low esses are sensitive enough to LL!R to make a radiation levels of inhibitors. The Science of the Total Environment triggered “SOSrepair―process feasible. Many of the l982;22:213—234. hormesis conjectures share a common conceptual 3. SaganLA. What is hormesisand why haven't we heard about it before? Health Phys 1987;52:521—525. framework: a LL!R-induced feedback loop that origi 4. Hickey Ri, Bowers El, Clelland RC. Radiation hormesis, nally evolved to help in radiation (or other free-radical public health, and public policy:a commentary. Health Phys inducer) detoxification, and that under some circum l983;44:207—219. stances overshoots the set point to produce net benefi 5. HafeleW. Energyfrom nuclear power.Sd Am 1990;Septem cial effects on the cell or organism. This sort of system ber.137-144. 6. Council on scientific affairs. Low-level radioactive wastes. has clearly been demonstrated using other types of JAMA 1989;26:669—674. toxins ( 73). 7. Harley, NH, Harley JH. Potential lung cancer risk from At present, society has placed the burden of proof indoor radon exposure. Cancer Journal for Clinicians squarely on those who argue that the regulations con l990;40:265—275. cerning ionizing radiation should be reduced and 8. Orient JM. Radon. JAMA l989;262:702. 9. Moghissi AA, Ray DL. Radiation hormesis and radiation brought more into line with other epidemiologic risks cancer risk. Health Phys. 1988;54:473. of the same demonstrated order of magnitude. Al 10. Weart SR. Nuclearfear: Cambridge, MA: Harvard University though some have argued that current levels of concern Press;1988:3—151. for LL!R are excessive and phobic (55), new data 11. Macklis RM. Radithor and the era of mild radium therapy. continue to appear that argue that the perceived risks JAMA l990;264:614—618. 12. Martland HS. Occupational poisoning in the manufacture of may be accurate reflections of the dangers posed (67). luminous watch dials. JAMA 1929;92:466—473. Many processes that result in LLIR exposure also bear 13. Simpson RE, Shuman FGD, Moghissi AA, Blackburn JA, the concurrent potential for a disastrous high level BaileyED. Occupationalexposureto dial painters and assem radioactive release (68). Although this potential is blers ofradioluminous timepieces. Health Phys 1983;44:SOl clearly a separate issue, the risk of such a high level 506. 14. Macklis RM, Bellerive MR, Humm JL. The radiotoxicology catastrophe plays an obvious role in the formulation of of radithor. JAMA l990;264:619—621. public policy on the issue of LLIR. 15. Walker iS. The controversy over radiation safety. JAMA The health effects and carcinogenic risks of chemical 1989;263:664—668. mutagens and various types of non-ionizing radiation 16. Yoshimoto Y. Cancer risk among children of atomic bomb are currently being re-evaluated (69,70). Perhaps this survivors. JAMA l990;264:596—600. 17. Shimizu Y, Schull WJ, Kato H. Cancer risk among atomic would be an opportune time to reconsider the entire bomb survivors.JAMA l990;264:601—604. question ofLL!R, and to develop a set ofunified criteria 18. Gofman JW. The question of radiation causation of cancer for the assessment of risk for various proposed health inHanfordworkers.HealthPhys1979;37:6l7—639. threats including ionizing radiation, natural and artifi 19. Gilbert ES, Fry SA, WiggsLD, et al. Analysesof combined cial chemical , and electromagnetic fields. mortality data on workers at the Hanford site, Oak Ridge National Laboratory, and Rocky Flats Nuclear Weapons The subject of radiation appears at present Plant. Radiat Res 1989;l20:l9—35. too politicized to allow a dispassionate risk analysis 20. Mole RH. after prenatal exposure to diag (71),andif a hormeticeffectexistsforionizingradia nostic X-ray examinations in Britain. Br J Cancer 1990;

Radiation Hormesis •Macklis and Beresford 357 62:152—168. Radiologyl959;72:518—521. 21. Kohn HW. Comments on radiation, hormesis, 46. Meyer MB, Tonascia JA, Men T. Long-term effectsof pre and public policy. Health Phys l984;46:l 159—1160. natal X-ray on development and fertility of human females. 22. Archer yE. Epidemiological studies of low-level radiation In: Biologicaland environmental effects of low-levelradia efFects.Health Phys 198l;40:129—130. tion. Vol. II, Proc. ofSymposium, Chicago, Nov. 1975;273— 23. Gofman JW. Health effectsof ionizingradiation. Dr. Sagan's 284. (Vienna International Atomic Energy Agency). paradigms.Health Phys 1987:52:679—680. 47. Meyer MB, Tonascia J. Long-term effectsof prenatal X-ray 24. Meinhold CB. President's message—l981. Health Phys of human females.II. Growth and Development. Am J Epi 1981:41:1. demiol198l;l14:317—336. 25. Sagan LA. On radiation, paradigms, and hormesis. Science 48. Broda E. Biopositive effects of ionizing radiation? Health I989;245:574, 621. Phys 1984;46:704—705. 26. Woffi'S.Are radiation-inducedeffectshormetic?Health Phys 49. Miller MW, Miller WM. Radiation hormesis in plants. Health 1989;245:575,621. Phys1987;52:607—616. 27. National ResearchCouncil. Health effectsofexposure to low 50. Kondo S. Altruistic cell suicide in relation to radiation hor levels of ionizing radiation. Washington DC: National Acad mesis. mt J Radiat Biol 1988;53:95—l02. emy Press. 1989. 51. Ucker DS. Cytotoxic T lymphocytesand glucocorticoidsac 28. Dinitz MM, ed. Higher cancer risks found. ACR l990;46: tivate an endogenous suicide process in target cells. Nature I, 5. 1987;327:62—64. 29. Catlin Ri. [Letter].Science 1989:246:311. 52. Delpla M. Let us come back to ‘radiationhormesis.' Health 30. Luckey TD. Physiological benefits from low levels of ionizing Phys l989;57:203—212. radiation. Health Phys l982;43:77 1—789. 53. National Academy of Sciences Committee on the Biological 3 1. Olivieri G, Bodycote J, WolffS. Adaptive response of human Effectsof Ionizing Radiation. Health EffectsofExposure to lymphocytes to low concentrations of radioactive thymidine. Low Levels of Ionizing Radiation (BEIR V). Washington, Science 1984:223:594—597. DC: National Academy Press; 1990. 32. Tuschl H, Altmann H, Kovac R, et al. Effectsof low-dose 54. Wang Z, Boice JD Jr., Wei L, et al. Thyroid nodularity and radiation on repair processes in human lymphocytes. Radiat aberrations among women in areas of high Re@sl980;8 1:1—9. background radiation in China. J Nat Can Inst 1990:82:478— 33. Koppenol WH. Hormesis.Science 1989:246:311. 485. 34. Feinendegen LE. Muhlensiepen H, Bond VP, Sondhaus CA. 55. Yalow RS. Concerns with low level ionizing radiation: Ra Intracellularstimulation of biochemicalcontrol mechanisms tional or phobic? J Nuci Med l990;3 I:17A—18A, 26A. by low-dose, low-LET irradiation. Health Phys 1987:52:663— 56. Marwick C. Low-dose radiation: Latest data renew questions 669. of ‘safe'level. JAMA 1990:264:553—554,557. 35. Held KD. Role of free radical scavangersin radiation protec 57. Kato H, Schull WJ, Awa A, Akiyama M, Otake M. Dose tion of DNA. In: Johnson JE Jr., Walford R, Harman D, responseanalysesamong atomic bomb survivors exposed to Miquel J, eds. Free radicals, aging, and degenerative diseases. low-levelradiation. Health Phys l987;52:645—652. New York, NY: Alan R. Liss, Inc. 1986;527—568. 58. Roberts L. British radiation study throws experts into tizzy. 36. Harman D. Free radical theory of aging. In: Johnson JE Jr., Science 1990;248:24—25. Walford R, Harman D, Miquel J, eds. Free radicals, aging, 59. Fogle S. U.S. Scientists blast “Sellafieldeffect.―J NIH Res and degenerative diseases. New York, NY: Alan R. Liss, Inc. 1990;2:28—30. I986:3—49. 60. Boice JD Jr. Studies of atomic bomb survivors. JAMA 37. RussVK.Consensusofthe effectofX-rays on bacteria.Hygie 1990;264:622—623. I909;56:34I—344. 6 1. Baverstock KF, Papworth D, Vennart J. Risks of radiation at 38. Murphy JB, Morton JJ. The effect of roentgen rays on the low dose rates. Lancet 198l;430—433. rate of growth of spontaneous tumors in mice. J Exp Med 62. Warren S. Longevityand causes of death from irradiation in 1915:800—803. physicians. JAMA 1956;162:464—468. 39. Metcalf D. Long-term effects of whole-body irradiation on 63. Seltzer R, Sartwell PE. Application ofcohort analysis to study lymphocyte in the mouse. Radiat. Res 1959; ofionizing radiation and longevity in physicians. Am JPublic 10:313—322. Healthl959;49:1610—1620. 40. Liu SZ, Liu WH, Sun JB. Radiation hormesis: Its expression 64. Goldstein S. Replicative senescence: The human fibroblast in the immune system. Health Phys 1987;52:579—583. comesofage.Science1990;249:1129—1133. 41. Gidali J, Bojtor I, Feher I. Kinetic basis for compensated 65. Marshal E. Academy panel raises radiation risk estimates. hemopoiesis during continuous irradiation with low doses. Science l990;247:22—23. Radiat Res l979;77:285—29l. 66. RussellWL, Kelly EM. Mutation frequency at low radiation 42. Anderson KC, Ritz J, et al. Hematologic engraftment and intensity [Abstract]. Washington, DC: Science l960;13l: immune reconstitution post-transplantation with anti-B 1- 1320. purgedautologousmarrow. Blood l987;69:597—604. 67. Modan B,et al. Increasedrisk ofbreast cancer after low-dose 43. Soleilhavoup JP, Tixador R, Croute F, Richoilley G. Para irradiation. JAMA 1989;262:690. mecium aurelia as a cellular model used for studies of the 68. Marshall E. Radiation exposure: Hot legacyof the cold war. biological effects of natural ionizing radiation or chronic low Science 1990;249:474. levelirradiation. IAEA-134/17(Vienna:International Atomic 69. Ames BN, Gold LS. Too many rodent carcinogens: Mitogen EnergyAgency)l979;335. esis increases mutagenesis. Science l990;249:970—971. 44. Wendt 0. Observation on the physiologic influence of uni 70. Pool R. Is there an EMF-Cancer connection? Science polar high frequency electric discharges in connection with 1990;249:1096—1098. radium radiation. Presented at the 13th International Con 71. Roberts L. Counting on science at EPA. Science 1990; gress of Physiologists; Boston, MA: August, 1929. 249:616—618. 45. Kaplan I. Genetic effects in children and grandchildren of 72. Calabrese EJ, McCarthy ME, Kenyon E. The occurrence of women treated for infertility and sterility by roentgen therapy. chemicallyinduced hormesis.Health Phys l987;52:53l—54l.

358 The Journal of Nudear Medicine •Vol. 32 •No. 2 •February 1991 73. Stebbing ARD. Growth hormesis: A by-product of control. responsein irradiated seed. Health Phys 1987:52:599—605. Health Phys 1987;52:543—547. 76. Upton AC. Carcinogenic Effectsof Low-LevelIonizing Ra 74. Bond VP, Feinendegen LE, Sondhaus CA. Microdosimetric diation. JNCJ 1990;82:448—449. concepts applied to hormesis.Health Phys 1987;52:659—661. 77. Tilyou S. BEIR V report: experts urgecautious interpretation 75. Sheppard SC, Regitnig PJ. Factors controlling the hormesis of higher risk estimates.J NuciMed l990;31:13A—l9A.

Radiation Hormesis •Macklis and Beresford 359