hall 18_01
DOSE - RESPONSE RELATIONSHIPS
Radiation biology applied to clinical radiotherapy is concerned with the relationship between a given absorbed dose of radiation and the consequent biologic response; of particular interest are factors that modify this relationship. With increasing radiation dose, radiation effects may increase in severity (i.e., grade), in frequency (i.e., incidence), or both. In most eases, it is the relation between dose and incidence that is important. Such dose - response curves have a sigmoid (5) shape, with the incidence tending to zero as dose tends to zero and the incidence tending to 100% at very large doses. This applies to both tumor control and normal-tissue complications.
A simple example is shown in Figure 18.1. Tumor control probability is plotted as a function of total dose, and the incidence of normal-tissue complications is also plotted as a function of dose. What is illustrated is a favorable situation where the tumor is more radiosensitive than the normal tissue. In the case of tumor control, the shape can be explained solely from the random nature of cell killing (or clonogen survival) after irradiation and the need to kill every single cell.
For most normal-tissue end points, the biologic interpretation of the sigmoid shape of the relationship is not obvious. Some researchers have evoked a hypothetical tissue rescue unit (TRU), arguing that tissue breakdown occurs when the number of TRUs falls below a critical level; however, this explanation is questionable.
Therapeutic Ratio (Therapeutic Index)
The ratio of the tumor response for a fixed level of normal-tissue damage has been variously called the therapeutic ratio or therapeutic index. In the hypothetical example in Figure 18.1, there is a favorable therapeutic ratio, because a 30% probability of tumor control is possible for a 5% incidence of complications.
The time factor is the one parameter that has been most often manipulated to increase this ratio; hyperfractionation, for example, produces a greater sparing of late-responding normal tissue than tumor control. Another strategy often quoted, though seldom achieved in practice, is to add a drug or radiosensitizer that potentiates the tumor control without potentiating the radiation damage to normal tissue. In practice, it does not need to be as clear-cot as this; it would suffice for the drug to increase tumor control to a greater extent than it increases normal tissue damage. This would result in a therapeutic gain. This is illustrated in Figure 18.2. The addition of the drug moves the tumor control curve to the left to a greater extent than the normal- tissue damage curve; that is, the drug has a larger cytotoxic effect on the tumor than on the normal tissue. Consequently, with the combined modalities, an improved tumor control probability is possible for the same probability of normal-tissue injury.
MITOTIC DEATH AND APOPTOSIS: HOW AND WHY CELLS DIE
Most cell lines cultured in vitro die a mitotic death after irradiation; that is, they die attempting to divide. This does not necessarily occur at the first postirradiation mitosis; the cell may struggle through one, two, or more mitoses before the damaged chromosomes cause cell death while attempting the complex task of cell division. Time-lapse films of irradiated cells cultured in vitro clearly show this process of mitotic death, which is the dominant cause of death if reproductive integrity is assessed in vitro as described in Chapter 3.
Mitotic death, however, is not the only form of cell death. Programmed cell death, or apoptosis, occurs in normal tissues and neoplasms, in mammals and amphibians, in the embryo and the adult. It is implicated, for example, in tissue involution, such as the regression of the tadpole tail during metamorphosis, and it is common during embryonic development. It also can occur after irradiation. Apoptosis means “falling off,” as of petals from flowers or leaves from trees. (Like the word mitosis, it comes from the Greek.)
Apoptosis is characterized by a stereotyped sequence of morphologic events that take place in two discrete phases. In the first phase, cells condense and bud to produce many membrane-enclosed bodies. In the second phase, these bodies are phagocytized and digested by nearby tissue cells. The characteristic “laddering” of DNA that occurs during apoptotic death is illustrated in Chapter 3. Apoptosis characteristically affects scattered individual cells. If apoptosis affects cells in tissues, the resulting apoptotic bodies are squeezed along the intercellular spaces and are either shed from the epithelial surface or rapidly phagocytized by nearby cells. The cells surrounding those being deleted merely close ranks, and there is no tissue disorganization such as occurs after necrosis.
ASSAYS FOR DOSE - RESPONSE RE LATIONSHIPS
A number of experimental techniques are available to obtain dose - response relationships for the cells of normal tissues. First, there are a limited number of clonogenic assays - techniques in which the end point observed depends directly on the reproductive integrity of individual cells. These systems are directly analogous to cell survival in vitro. The techniques developed by Withers and his colleagues are based on their observation of a clone of cells regenerating in situ in irradiated tissue. Skin colonies, regenerating crypts in the jejunum, testes stem cells, and kidney tubules are described briefly later in this chapter. It is also possible to obtain dose - response curves for the cells of the epithelial lining of the colon or stomach, but the method used is essentially the same as for the jejunum. Kember described a system for scoring regenerating clones in cartilage at about the same time as the Withers’ skin colony system, but it has not been used widely and is not discussed here.
The assay system for the stem cells in the bone marrow or cells of the thyroid and mammary gland depends on the observation of the growth of clones of cells taken from a donor animal and transplanted into a different tissue in a recipient animal. In Till and McCulloch’s bone-marrow assay, colonies of bone marrow cells are counted in the spleens of recipient animals. Dose - response curves for mammary and thyroid cells have been obtained by Gould and Clifton by observing colonies growing from cells transplanted into the fat pads of recipient animals.
Second, dose - response relationships can he obtained that are repeatable and quantitative but that depend on functional end points. These include skin reactions in rodents or pigs (e.g., crythema and desquamation), pneumonitis or fibrosis in mouse lungs reflected in an increased breathing rate, myelopathy of the hind limbs from damage to the spinal cord, and deformities to the feet of mice. The end points observed tend to reflect the minimum number of functional cells remaining in a tissue or organ, rather than the fraction of cells retaining their reproductive integrity.
Finally, one can infer a dose - response curve for a tissue in which it cannot be observed directly by assuming the form of the dose - response curve (linear- quadratic) and performing a series of multifraction experiments. This procedure, first suggested by Douglas and Fowler, has been used widely to infer values for α and β in the dose - response relationships for normal tissues in which the parameters cannot be measured directly.
This chapter includes assays for both early- and late-responding tissues. The skin, intestinal epithelium, and bone-marrow cells, for example, are rapidly dividing self-renewal tissues and respond early to the effects of radiation. The spinal cord, lung, and kidney, by contrast, are late-responding tissues. This reflects the current philosophy that the radiation response of all tissues results from the depletion of the critical parenchymal cells and that the difference in time at which early- and late-responding tissues express radiation damage is a function simply of different cell turnover rates. Many older papers in the literature ascribe the response of late-responding tissues to vascular damage rather than to depletion of parenchymal cells, but this thesis is becoming increasingly difficult to accept.
The various types of normal-tissue assay systems are described briefly. The reader who is content with the summary already given may wish to skip the remainder of this chapter.
CLONOGENIC END POINTS
Clones Regrowing in Situ
Skin Colonies
Withers developed an ingenious technique (Fig. 18.3) to determine the survival curve for mouse skin cells. The hair was plucked from an area on the back of the mouse, and a superficial x-ray machine was used to irradiate an annulus of skin to a massive dose of 30 Gy (3,000 rad). This produced a “moat” of dead cells, in the center of which was an isolated island of intact skin that had been protected during the first exposure to low-voltage x-rays by a small metal sphere. This small area of intnct skin was then given a test dose (D) and subsequently observed for regrowth of skin. If one or more stem cells survived in this small area, nodules of regrowing skin could be seen some days later. If no cells survived in this small area, the skin would heal much later by infiltration of cells crossing the moat. Figure 18.4 shows nodules regrowing in mouse skin. To obtain a survival curve, it was necessary to repeat this operation with a number of different areas of skin. A range of ball bearings was used to shield a small area of skin in the middle of the “moat.” The resulting survival data are shown in Figure 18.5, in which the dose (D) to the control area is plotted against the number of surviving cells per square centimeter of skin.
There are practical limits to the range in which the dose - response relationship can be determined. At one extreme, it is not possible to irradiate too large an area on the back of the mouse to produce the moat of sterilized skin. At the other extreme, the smallest area that can be used is determined by the fact that even 30-ky radiation scatters laterally to some extent. As can be seen in Figure 18.5, the technique results in a single-dose survival curve that extends from about 8 to 25 Gy (800—2,500 rad); over this range, with dose plotted on a linear scale and the number of surviving cells per square centimeter plotted on a logarithmic scale, the survival curve is straight and has aD0 of 1.35 Gy (135 rad). This D0 value is very similar to that obtained with mammalian cells cultured in vitro.
The extrapolation number cannot be obtained directly with this technique; the ordinate is the number of surviving cells per square centimeter of skin, and this cannot be converted to the surviving fraction because it is not known with any accuracy how many skin stem cells there are per unit area. It is, however, possible to make an indirect estimate of the extrapolation number by obtaining the survival curve for doses given in two fractions separated by 24 hours. The survival curve obtained in this way also is shown in Figure 18.5. It is parallel to that obtained for single doses but is displaced from it toward higher doses. As explained in Chapter 3, this lateral displacement in a direction parallel to the dose axis is a measure of Dq, the quasithreshold dose. The Dq for mouse skin is about 3.5 Gy (350 rad), which is very similar to the value for human skin estimated from split-dose experiments.
Crypt Cells of the Mouse Jejunum
A technique perfected by Withers and Elkind makes it possible to obtain the survival characteristics of the crypt cells of the mouse jejunum. The lining of the jejunum is a classic example of a self-renewal system. The cells in the crypts divide rapidly and provide a continuous supply of cells that move up the villi, differentiate, and become the functioning cells. The cells at the top of the folds of the villi are slowly but continuously sloughed off in the normal course of events and are replaced continuously by cells that originate from mitoses in the crypts.
Figure 18.6, an electron micrograph, dramatically shows the three-dimensional structure of the lining of the intestinal epithelium. Mice are given a total-body dose of 11 to 16 Gy (1,100— 1.600 rad). which sterilizes a significant proportion of the dividing cells in the crypts but has essentially no effect on the differentiated cells in the villi. Consequently, crypt degeneration appears early after irradiation, and the villi remain long and their epithelial covering of differentiated cells shows little change. With the further passage of time, the tips of the villi continue to be sloughed away by normal use, but no replacement cells are available from the depopulated crypts, and so the villi begin to shorten and shrink. At sufficiently high doses, the surface lining of the jejunum is completely denuded of villi.
To obtain a survival curve for the jejunal crypt cells, groups of animals are exposed to graded total- body doses of radiation. After 3.5 days. each animal is sacrificed and sections are made of the jejunum (Fig. 18.7A). At this time, crypts are just beginning to regenerate and it is relatively simple to identify them. Figure 18.7B shows a number of regenerating crypts at a higher magnification. These pictures also show the shortened villi and the greatly reduced density of cells lining the surface. The score of radiation damage is the number of regenerating crypts per circumference of the sectioned jejunum. This quantity is plotted as a function of dose and yields a survival curve as shown in Figure 18.8. The single-dose survival curve has a D0 (for y-rays) of about 1.3 Gy (130 rad). Also shown in Figure 18.8 are survival curves for radiation delivered in multiple fractions, from 2 to 20. The separation between the single- and two-dose survival curves gives a measure of Dq, which has the very large value of between 4 and 4.5 Gy (400—450 rad).
This technique has two limitations. First, the quantity plotted on the ordinate is the number of surviving crypts per circumference, not the surviving fraction. Second, experiments can be done only at doses of about 10 Gy (1,000 rad) or more, at which there is a sufficient level of biologic damage for individual regenerating crypts to be identified. The doses can be delivered, however, in a number of smaller fractions, as long as the total results in enough biologic damage to be scored. The shape of the entire survival curve then can be reconstructed from the multifractiec data if it is assumed that in a fractionated regimen each dose produces the same amount of cell killing and if an estimate is made of the number of clonogens at risk per crypt. This has been done by Withers and his colleagues; the resultant survival curve is shown in Figure 18.9.
Testes Stem Cells
A technique to measure the radiation response of testicular cells capable of sustaining spermatogenesis (i.e., the stem cells) was devised by Withers and his colleagues. About 6 weeks after irradiation, mouse testes are sectioned and examined histologically. Sections of normal and inadiated testes are shown in Figure 18.10. The proportion of tubules containing spermatogenic epithelium is counted and plotted as a function of dose in Figure 18.11. As in many in vim assays, relatively high single doses of 8 to 16 Gy (800—1,600 rad) are necessary so that the level of damage is sufficient to be scored. In this dose range, D0 is about 1.68 Gy (168 rad). If the split-dose technique is used, the Dq is about 2.7 Gy (270 rad). It is possible to estimate the effect of small doses and reconstruct a complete survival curve by giving large doses in multiple small fractions and assuming that the response to each fraction is the same. The result of this reconstruction is shown in Figure 18.12.
Kidney Tubules
A technique using kidney tubules, again developed by Withers and his colleagues, is the first clonal assay for a late-responding tissue. One kidney per mouse is irradiated with a small field and removed for histologic examination 60 weeks later. Figure 18.13 shows sections of normal and irradiated kidneys. For ease of scoring, only those tubules touching the renal capsule are scored, and a tubule is considered fully regenerated only if it is lined with well-differentiated cuboidal or columnar cells with a large amount of eosinophilic cytoplasm. By 60 weeks, tubules either have no surviving epithelial cells or are lined completely with epithelium that has regenerated from a small number of surviving cells, usually one. The number of tubules regenerating in an arbitrary number of sections counted is plotted as a function of radiation dose. The result is shown in Figure 18.14; D0 is about 1.53 Gy (153 rad).
The radiosensitivity of the cells of this late- responding tissue is not very different from that of early-responding tissues, such as the skin or intestinal epithelium. The rate of response, however, is quite different. The time required for depletion of the epithelium after a single dose of 14 Gy (1,400 rad) is about 3 days in the jejunum, 12 to 24 days in the skin, and 30 days in the seminiferous tubules of the testes, but 300 days in the kidney tubules. These results argue strongly that radiation injury in the kidney results from depletion of parenchymal cells and that the slow expression of injury merely reflects the slow turnover of this cell population. Vascular injury is unlikely to be the mechanism underlying the destruction of renal tubules.
Cells Transplanted to Another Site
Bone Marrow Stem Cells
Till and McCulloch developed a system to determine a survival curve for colony-forming bone marrow cells (Fig. 18.15). Recipient animals first are irradiated supralethally with a dose of 9 to 10 Gy (900—1,000 rad), which sterilizes their spleens. Nucleated isologous bone marrow cells taken from another animal are then injected intravenously into the recipient animals. Some of these cells lodge in the spleen, where they form nodules, or colonies, 10 to 11 days later, because the spleen cells of the recipient animals have been sterilized previously by the large dose of radiation. At this time, the spleens are removed and the colonies counted. Figure 18.16 is a photograph of a spleen showing the colonies to be counted.
About l0 cells must be injected into a recipient animal to produce one spleen colony, because the majority of the cells in the nucleated isologous bone marrow are fully differentiated cells and would never be capable of forming a colony. To obtain a surviving fraction for bone marrow cells, a donor animal is irradiated to some test dose, and the suspension of cells from the bone marrow is inoculated into groups of recipient animals that previously had been irradiated supralethally. By counting the colonies in the spleens of the recipient animals, and with a knowledge of the number of cells required to produce a colony in an unirradiated animal (plating efficiency), the surviving fraction may be calculated as follows:
Surviving fraction for a dose D = colonies counted / cells inoculated x plating efficiency
This procedure is repeated for a range of doses, and a survival curve is obtained (Fig. 18.17). These bone marrow stem cells are very sensitive with aD0 of about 0.95 Gy (95 rad) and little or no shoulder to the survival curve.
Mammary and Thyroid Cells
Clifton and Gould and their colleagues developed very useful clonogen transplant assays for epithelial cells of the mammary and thyroid glands. They have
been used largely for cell survival studies, described later, but the initial motivation for their development was to study carcinogenesis in a quantitative system. Most in vitro transformation assays involve fibroblasts, and the bulk of human cancers arise in epithelial cells - hence, the importance and interest in these two systems.
The techniques for these two systems are much the same. To generate a survival curve for mammary or thyroid gland cells in the rat, cells may be irradiated in vivo before the gland is removed from donor animals and treated with enzymes to obtain a monodispersed cell suspension. Known numbers of cells are injected into the inguinal or interscapular white fat pads of recipient animals.
Under appropriate host conditions and grafted cell numbers, the injection of mammary cells gives rise to mammary structures that are morphologically and functionally normal. One such mammary structure may develop from a single cell. By 3.5 weeks after the injection of mammary cells, positive growth is indicated by alveolar units. An example of a milk-filled alveolar unit is shown as an inset in Figure 18.18. If thyroid cells are injected, thyroid follicular units develop (Fig. 18.19).
With either type of cell, a larger number must be injected to produce a growing unit if the cells first are irradiated to a given dose. In practice, some fancy statistics are involved, a discussion of which is beyond the scope of this chapter; in essence, the ratio of the number of irradiated to unirradiated cells required to produce one growing unit (thyroid follicular unit or alveolar unit) is a measure of the cell-surviving fraction corresponding to the dose. This procedure must be repeated for a range of graded doses to generate a survival curve. The resultant survival curve for mammary cells is shown in Figure 18.18. The characteristics of the curve are unremarkable: D0 is about 1.27 Gy (127 rad), and the extrapolation number is about 5, quite typical of rodent cells cultured in vitro. The corresponding survival curve for thyroid cells is shown in Figure 1 8.19. D0 is a little larger than for mammary glands assayed in a similar way, implying that the cells are a little more resistant. Figures 18.18 and 18.19 also show data for cells left in situ for 24 hours after irradiation before being removed and assayed. If this is done, the shoulder of the survival curve is larger because of the repair of potentially lethal damage. This is discussed in more detail in Chapter 5.
An interesting use of these clonogen transplant assays is that the physiologic states of either donor or recipient animals can be manipulated hormonally. For the mammary cell assay, cells may be taken from inactive, slowly dividing glands of virgin rats, from rapidly dividing glands of rats in mid- pregnancy, or from milk-producing glands of lactating rats. For the thyroid cell assay, the physiologic states of both donor and recipient can be manipulated by control of the diet or by partial thyroidectomy.
SUMMARY OF DOSE-RESPONSE CURVES FOR CLONOGENIC ASSAYS IN NORMAL TISSUES
The survival curves for all of the clonogenic assays in normal tissues are plotted together io Figure 18.20. There is a substantial range of radiosensitivities, with shoulder width being the principal variable. In vitro curves for cells from patients with ataxia-telangiectasia also are shown because these are probably the most radiosensitive mammalian cells.
DOSE—RESPONSE RELATIONSHIPS FOR FUNCTIONAL END POINTS
Pig Skin
Pig skin has been used widely in radiobiologic studies because it has many features in common with human skin, such as color, hair follicles, sweat glands, and a layer of subcutaneous fat. In view of these structural similarities, it is not surprising that the response of pig skin to radiation closely resembles that of human skin, both qualitatively and quantitatively.
Fowler and his colleagues pioneered the use of pig skin as a radiobiologic test system. A number of small rectangular fields on the pig’s flank were irradiated with graded doses of x-rays, and the reactions were scored daily using the arbitrary scale shown in Table 18.1. After a single dose of radiation, the reaction becomes apparent after about 15 days and develops as shown in Figure 18.21.
Two phases of the reaction can be distinguished. First, an early wave of erythema occurred (at 10—40 days), which was variable from one animal to another. This represents the uncomfortable “acute” reaction sometimes seen in patients on radiotherapy at about the end of a course of treatment. Second. a more gradual increase to a second broad wave of moderately severe reactions took place (at 50—100 days), representing a more permanent kind of damage. This second wave shows the tolerance of skin to a more serious type of long-term damage and is also a more repeatable and consistent index of radiation damage. It was subsequently found to correlate well with longer-term damage (up to 2 years) and with subcutaneous damage.
The “score” of radiation damage is taken to be the average skin reaction occurring between certain time limits that encompass the medium-term reactions. After a single dose, this might be a 35 - day period between 50 and 85 days after irradiation. For a protracted fractionated regimen, this period of reaction may come later, between days 65 and 100. The average skin reaction in the chosen time period then is plotted as a function of dose; examples of dose—response curves obtained this way are shown in Figure 18.22 for single and fractionated doses.
Late effects also have been studied in pig skin by measuring the contraction that results from fibrosis a year or more after irradiation. A square is tattooed on the skin of the animal in the irradiated field, and the dimensions of this square are recorded as a function of dose as the contraction occurs. This is a primitive but effective measure of late effects.
Many of the important early studies on the fractionation effects of x-rays and the comparison of x-rays with fast neutrons were performed with this biologic system. One overwhelming advantage is that data obtained this way can be extrapolated to the human with a high degree of confidence. The disadvantage is that the animals are large and awkward to work with, and their maintenance involves a considerable expense.
Rodent Skin
Because of the inconvenience and expense of using pigs, the skin of the mouse leg and foot is commonly used instead. One hind leg of each animal is irradiated; the other serves as a control. The skin response is observed each day after irradiation and is scored according to the arbitrary scale shown in Table 18.2. Various doses are used. The progressive development of the reaction after ten doses of 6 Gy (600 rad) each is illustrated in Figure 18.23; each point represents the mean of several animals. Reactions appear by about the 10th day, peak by 20 to 25 days, and then subside. The second wave of the reaction, noted for pig skin, is not seen in mice but is observed in rats. A dose - response curve is obtained by averaging the skin reaction over a period of time and plotting this average as a function of dose.
Early and Late Response of the Lung Based on Breathing Rate
Travis and her colleagues developed a noninvasive assay of breathing frequency to assess both early and late damage in mouse lungs. Breathing frequency increases progressively with dose after a threshold of about 11 Gy (1,100 rad) (Fig. 18.24). The increased breathing frequency in rodent lungs at 16 and 36 weeks is associated with the early response (i.e., pneumonitis); by 52 weeks, the elevated breathing frequency is associated with the late response (i.e., fibrosis). This is a simple but highly quantitative and reproducible system.
Spinal Cord Myelopathy
A dose - response relationship can be determined for late damage caused by local irradiation of the spinal cords of rats. A number of investigators have worked with this system, notably van der Kogel. After latent periods of 4 to 12 months, symptoms of myelopathy develop, the first signs of which are palpable muscle atrophy, followed some time later by impaired use of the hind legs. Figure 18.25 shows the steep dose - response relationship for hind-limb paralysis following the irradiation of a section of the spinal cord in rats. These data also show the dramatic sparing that results from fractionation; this is discussed further in another section of this chapter.
The various syndromes of radiation-induced injury in rodent brain and spinal cord are very similar to those described in humans. Lesions observed within approximately the first 6 months after irradiation are limited primarily to the white matter and range between early diffuse or focal demyelination and extensive necrosis. Different pathogenic pathways toward the development of white-matter necrosis have been proposed, with the glial and vascular tissue components the major targets. The most common type of late delayed injury peaks at 1 to years postirradiation and almost certainly has a vascular basis. Another type of late injury that has been described more recently in various species, including humans, is slowly progressive glial atrophy. This lesion is not associated with necrosis but occurs diffusely and at lower doses. With improvements in diagnostic procedures such as magnetic resonance imaging, glial atrophy may become more frequently recognized adverse effect of brain tumor therapy.
Latency
Over a dose range of about 25 to 60 Gy (2,500- 6,000 rad), delivered in single doses, the general tendency is a decreasing latency with increases in dose of approximately 2 days/Gy (2 days/100 rad). There is a considerable variation with animal strain. as well as with the region of the cord irradiated.
In terms of mechanisms, demyelination or slowly progressive atrophy is probably a consequence of interference with the slow continuous turnover of oligodendrocytes by killing of glial progenitor cells. Vascular injury may accelerate, precipitate, or even initiate the white-matter changes leading to necrosis. This is an area of some controversy.
Fractionation and Protraction
The effect of dose fractionation and protraction on tolerance to radiation has been investigated extensively in the rat spinal cord and to a lesser extent in the mouse, monkey, and guinea pig. Because these systems turn over slowly, there is little influence of overall treatment time up to any conventional clinical regimen of 6 to 8 weeks. On the other hand, dose per fraction is very important (Fig. 18.25), with the dose to produce paralysis increasing dramatically with number of fractions. The effect of a large number of very small fractions also has been investigated. Figure 18.26 shows the relation between total dose and dose per fraction to produce paralysis in 50% of rats from irradiation of a short length of cervical spine. The smooth curve is an iso- effect curve calculated for the very low aIjS value of 1,5 Gy (150 rad). The experimental data suggest that the linear-quadratic (L.Q) model overestimates the tolerance for small doses per fraction of less than 2 Gy (200 rad). However, this may be a result of incomplete repair, because in these experiments, the interfraction interval was only 4 hours. There is good reason to believe that repair of sublethal damage takes place slowly in this normal tissue, and indeed, repair may be biphasic, with “fast” and “slow” components. For this reason, if multiple doses per day are used to the spinal cord, the interfraction interval should he at least 6 to 8 hours.
Volume Effects
The total volume of irradiated tissue usually is assumed to have an influence on the development of tissue injury. The spinal cord is perhaps the clearest case in which the functional subunits (FSUs) are arranged in linear fashion, like links in a chain. Figure 18.27 shows the relation between tolerance dose and the length of cord irradiated in the rat. For short lengths of cord, below 1 cm, tolerance in terms of white-matter necrosis shows a marked dependence on the length of cord irradiated. Late vascular injury shows less dependence on cord length. Beyond a few centimeters, the tolerance is virtually independent of the length of cord irradiated. This would be predicted from the linear arrangement of the functional subunits. A chain is broken whether one, two, three, or more links are removed.
Retreatment after Long Time Intervals
The spinal cord does recover to some extent after long time periods following irradiation. The extent of the recovery depends, of course, on the first treatment - that is, what fraction of tolerance was involved. Experiments with rats indicate that after an initial treatment to 50% tolerance, the retreatment tolerance approaches 90% of the tolerance of the untreated control group by about a year after the initial irradiation. If the initial treatment represented a larger fraction of tolerance, the retreatment that can be tolerated is reduced.
INFERRING THE RATIO α/β FROM MULTIFRACTION EXPERIMENTS IN NONCLONOGENIC SYSTEMS
The parameters of the dose - response curve for any normal tissue system for which a functional end point can be observed may be inferred by performing a multifraction experiment. Take, for example, an experiment in which mouse foot skin reaction is scored. Doses that result in the same skin reaction (e.g., moist desquamation over 50% of the area irradiated) if delivered as a single exposure in a multifraction regimen (e.g., 5, 10, or 20 fractions) must be determined experimentally. A number of assumptions must be made:
1. The dose - response relationship is represented adequately by the linear-quadratic formulation:
S = e^ (- αD – βD2)
in which S is the fraction of cells surviving a dose, D, and a and ft are constants.
2. Each dose in a fractionated regimen produces the same biologic effect.
3. Full repair of sublethal damage fakes place between dose fractions, but no cell proliferation occurs.
Suppose the total dose, D, is divided into n equal fractions of dose d. The previous equation then can be rewritten:
If the reciprocal of the total dose (1/nd) is plotted against the dose per fraction (d), a straight line results, as shown in Figure 18.28. The intercept on the ordinate gives α/logeS; the slope gives β/logeS. In general, the value of logeS is not known unless other cell survival studies are available, but the ratio of the intercept to the slope provides an estimate of α/β.
Multifraction experiments have been performed and estimates of a/fl made for essentially all of the normal-tissue end points described in this chapter. One of the important conclusions arrived at is that the value of a/fl tends to be larger for early- responding tissues, about 10 Gy (1,000 rad), than for late-responding tissues, about 2 Gy (200 rad).
Because α/βis the dose at which cell killing by linear and by quadratic components of radiation damage are equal (Chapter 3), the implication is that dose - response relationships for late-responding tissues are “curvier” than for early-responding tissues. The importance of this conclusion becomes evident in the discussion of fractionation in radiotherapy in Chapter 22.
Radiation biology applied to clinical radiotherapy is concerned with the relationship between a given absorbed dose of radiation and the consequent biologic response; of particular interest are factors that modify this relationship. With increasing radiation dose, radiation effects may increase in severity (i.e., grade), in frequency (i.e., incidence), or both. In most eases, it is the relation between dose and incidence that is important. Such dose - response curves have a sigmoid (5) shape, with the incidence tending to zero as dose tends to zero and the incidence tending to 100% at very large doses. This applies to both tumor control and normal-tissue complications.
A simple example is shown in Figure 18.1. Tumor control probability is plotted as a function of total dose, and the incidence of normal-tissue complications is also plotted as a function of dose. What is illustrated is a favorable situation where the tumor is more radiosensitive than the normal tissue. In the case of tumor control, the shape can be explained solely from the random nature of cell killing (or clonogen survival) after irradiation and the need to kill every single cell.
For most normal-tissue end points, the biologic interpretation of the sigmoid shape of the relationship is not obvious. Some researchers have evoked a hypothetical tissue rescue unit (TRU), arguing that tissue breakdown occurs when the number of TRUs falls below a critical level; however, this explanation is questionable.
Therapeutic Ratio (Therapeutic Index)
The ratio of the tumor response for a fixed level of normal-tissue damage has been variously called the therapeutic ratio or therapeutic index. In the hypothetical example in Figure 18.1, there is a favorable therapeutic ratio, because a 30% probability of tumor control is possible for a 5% incidence of complications.
The time factor is the one parameter that has been most often manipulated to increase this ratio; hyperfractionation, for example, produces a greater sparing of late-responding normal tissue than tumor control. Another strategy often quoted, though seldom achieved in practice, is to add a drug or radiosensitizer that potentiates the tumor control without potentiating the radiation damage to normal tissue. In practice, it does not need to be as clear-cot as this; it would suffice for the drug to increase tumor control to a greater extent than it increases normal tissue damage. This would result in a therapeutic gain. This is illustrated in Figure 18.2. The addition of the drug moves the tumor control curve to the left to a greater extent than the normal- tissue damage curve; that is, the drug has a larger cytotoxic effect on the tumor than on the normal tissue. Consequently, with the combined modalities, an improved tumor control probability is possible for the same probability of normal-tissue injury.
MITOTIC DEATH AND APOPTOSIS: HOW AND WHY CELLS DIE
Most cell lines cultured in vitro die a mitotic death after irradiation; that is, they die attempting to divide. This does not necessarily occur at the first postirradiation mitosis; the cell may struggle through one, two, or more mitoses before the damaged chromosomes cause cell death while attempting the complex task of cell division. Time-lapse films of irradiated cells cultured in vitro clearly show this process of mitotic death, which is the dominant cause of death if reproductive integrity is assessed in vitro as described in Chapter 3.
Mitotic death, however, is not the only form of cell death. Programmed cell death, or apoptosis, occurs in normal tissues and neoplasms, in mammals and amphibians, in the embryo and the adult. It is implicated, for example, in tissue involution, such as the regression of the tadpole tail during metamorphosis, and it is common during embryonic development. It also can occur after irradiation. Apoptosis means “falling off,” as of petals from flowers or leaves from trees. (Like the word mitosis, it comes from the Greek.)
Apoptosis is characterized by a stereotyped sequence of morphologic events that take place in two discrete phases. In the first phase, cells condense and bud to produce many membrane-enclosed bodies. In the second phase, these bodies are phagocytized and digested by nearby tissue cells. The characteristic “laddering” of DNA that occurs during apoptotic death is illustrated in Chapter 3. Apoptosis characteristically affects scattered individual cells. If apoptosis affects cells in tissues, the resulting apoptotic bodies are squeezed along the intercellular spaces and are either shed from the epithelial surface or rapidly phagocytized by nearby cells. The cells surrounding those being deleted merely close ranks, and there is no tissue disorganization such as occurs after necrosis.
ASSAYS FOR DOSE - RESPONSE RE LATIONSHIPS
A number of experimental techniques are available to obtain dose - response relationships for the cells of normal tissues. First, there are a limited number of clonogenic assays - techniques in which the end point observed depends directly on the reproductive integrity of individual cells. These systems are directly analogous to cell survival in vitro. The techniques developed by Withers and his colleagues are based on their observation of a clone of cells regenerating in situ in irradiated tissue. Skin colonies, regenerating crypts in the jejunum, testes stem cells, and kidney tubules are described briefly later in this chapter. It is also possible to obtain dose - response curves for the cells of the epithelial lining of the colon or stomach, but the method used is essentially the same as for the jejunum. Kember described a system for scoring regenerating clones in cartilage at about the same time as the Withers’ skin colony system, but it has not been used widely and is not discussed here.
The assay system for the stem cells in the bone marrow or cells of the thyroid and mammary gland depends on the observation of the growth of clones of cells taken from a donor animal and transplanted into a different tissue in a recipient animal. In Till and McCulloch’s bone-marrow assay, colonies of bone marrow cells are counted in the spleens of recipient animals. Dose - response curves for mammary and thyroid cells have been obtained by Gould and Clifton by observing colonies growing from cells transplanted into the fat pads of recipient animals.
Second, dose - response relationships can he obtained that are repeatable and quantitative but that depend on functional end points. These include skin reactions in rodents or pigs (e.g., crythema and desquamation), pneumonitis or fibrosis in mouse lungs reflected in an increased breathing rate, myelopathy of the hind limbs from damage to the spinal cord, and deformities to the feet of mice. The end points observed tend to reflect the minimum number of functional cells remaining in a tissue or organ, rather than the fraction of cells retaining their reproductive integrity.
Finally, one can infer a dose - response curve for a tissue in which it cannot be observed directly by assuming the form of the dose - response curve (linear- quadratic) and performing a series of multifraction experiments. This procedure, first suggested by Douglas and Fowler, has been used widely to infer values for α and β in the dose - response relationships for normal tissues in which the parameters cannot be measured directly.
This chapter includes assays for both early- and late-responding tissues. The skin, intestinal epithelium, and bone-marrow cells, for example, are rapidly dividing self-renewal tissues and respond early to the effects of radiation. The spinal cord, lung, and kidney, by contrast, are late-responding tissues. This reflects the current philosophy that the radiation response of all tissues results from the depletion of the critical parenchymal cells and that the difference in time at which early- and late-responding tissues express radiation damage is a function simply of different cell turnover rates. Many older papers in the literature ascribe the response of late-responding tissues to vascular damage rather than to depletion of parenchymal cells, but this thesis is becoming increasingly difficult to accept.
The various types of normal-tissue assay systems are described briefly. The reader who is content with the summary already given may wish to skip the remainder of this chapter.
CLONOGENIC END POINTS
Clones Regrowing in Situ
Skin Colonies
Withers developed an ingenious technique (Fig. 18.3) to determine the survival curve for mouse skin cells. The hair was plucked from an area on the back of the mouse, and a superficial x-ray machine was used to irradiate an annulus of skin to a massive dose of 30 Gy (3,000 rad). This produced a “moat” of dead cells, in the center of which was an isolated island of intact skin that had been protected during the first exposure to low-voltage x-rays by a small metal sphere. This small area of intnct skin was then given a test dose (D) and subsequently observed for regrowth of skin. If one or more stem cells survived in this small area, nodules of regrowing skin could be seen some days later. If no cells survived in this small area, the skin would heal much later by infiltration of cells crossing the moat. Figure 18.4 shows nodules regrowing in mouse skin. To obtain a survival curve, it was necessary to repeat this operation with a number of different areas of skin. A range of ball bearings was used to shield a small area of skin in the middle of the “moat.” The resulting survival data are shown in Figure 18.5, in which the dose (D) to the control area is plotted against the number of surviving cells per square centimeter of skin.
There are practical limits to the range in which the dose - response relationship can be determined. At one extreme, it is not possible to irradiate too large an area on the back of the mouse to produce the moat of sterilized skin. At the other extreme, the smallest area that can be used is determined by the fact that even 30-ky radiation scatters laterally to some extent. As can be seen in Figure 18.5, the technique results in a single-dose survival curve that extends from about 8 to 25 Gy (800—2,500 rad); over this range, with dose plotted on a linear scale and the number of surviving cells per square centimeter plotted on a logarithmic scale, the survival curve is straight and has aD0 of 1.35 Gy (135 rad). This D0 value is very similar to that obtained with mammalian cells cultured in vitro.
The extrapolation number cannot be obtained directly with this technique; the ordinate is the number of surviving cells per square centimeter of skin, and this cannot be converted to the surviving fraction because it is not known with any accuracy how many skin stem cells there are per unit area. It is, however, possible to make an indirect estimate of the extrapolation number by obtaining the survival curve for doses given in two fractions separated by 24 hours. The survival curve obtained in this way also is shown in Figure 18.5. It is parallel to that obtained for single doses but is displaced from it toward higher doses. As explained in Chapter 3, this lateral displacement in a direction parallel to the dose axis is a measure of Dq, the quasithreshold dose. The Dq for mouse skin is about 3.5 Gy (350 rad), which is very similar to the value for human skin estimated from split-dose experiments.
Crypt Cells of the Mouse Jejunum
A technique perfected by Withers and Elkind makes it possible to obtain the survival characteristics of the crypt cells of the mouse jejunum. The lining of the jejunum is a classic example of a self-renewal system. The cells in the crypts divide rapidly and provide a continuous supply of cells that move up the villi, differentiate, and become the functioning cells. The cells at the top of the folds of the villi are slowly but continuously sloughed off in the normal course of events and are replaced continuously by cells that originate from mitoses in the crypts.
Figure 18.6, an electron micrograph, dramatically shows the three-dimensional structure of the lining of the intestinal epithelium. Mice are given a total-body dose of 11 to 16 Gy (1,100— 1.600 rad). which sterilizes a significant proportion of the dividing cells in the crypts but has essentially no effect on the differentiated cells in the villi. Consequently, crypt degeneration appears early after irradiation, and the villi remain long and their epithelial covering of differentiated cells shows little change. With the further passage of time, the tips of the villi continue to be sloughed away by normal use, but no replacement cells are available from the depopulated crypts, and so the villi begin to shorten and shrink. At sufficiently high doses, the surface lining of the jejunum is completely denuded of villi.
To obtain a survival curve for the jejunal crypt cells, groups of animals are exposed to graded total- body doses of radiation. After 3.5 days. each animal is sacrificed and sections are made of the jejunum (Fig. 18.7A). At this time, crypts are just beginning to regenerate and it is relatively simple to identify them. Figure 18.7B shows a number of regenerating crypts at a higher magnification. These pictures also show the shortened villi and the greatly reduced density of cells lining the surface. The score of radiation damage is the number of regenerating crypts per circumference of the sectioned jejunum. This quantity is plotted as a function of dose and yields a survival curve as shown in Figure 18.8. The single-dose survival curve has a D0 (for y-rays) of about 1.3 Gy (130 rad). Also shown in Figure 18.8 are survival curves for radiation delivered in multiple fractions, from 2 to 20. The separation between the single- and two-dose survival curves gives a measure of Dq, which has the very large value of between 4 and 4.5 Gy (400—450 rad).
This technique has two limitations. First, the quantity plotted on the ordinate is the number of surviving crypts per circumference, not the surviving fraction. Second, experiments can be done only at doses of about 10 Gy (1,000 rad) or more, at which there is a sufficient level of biologic damage for individual regenerating crypts to be identified. The doses can be delivered, however, in a number of smaller fractions, as long as the total results in enough biologic damage to be scored. The shape of the entire survival curve then can be reconstructed from the multifractiec data if it is assumed that in a fractionated regimen each dose produces the same amount of cell killing and if an estimate is made of the number of clonogens at risk per crypt. This has been done by Withers and his colleagues; the resultant survival curve is shown in Figure 18.9.
Testes Stem Cells
A technique to measure the radiation response of testicular cells capable of sustaining spermatogenesis (i.e., the stem cells) was devised by Withers and his colleagues. About 6 weeks after irradiation, mouse testes are sectioned and examined histologically. Sections of normal and inadiated testes are shown in Figure 18.10. The proportion of tubules containing spermatogenic epithelium is counted and plotted as a function of dose in Figure 18.11. As in many in vim assays, relatively high single doses of 8 to 16 Gy (800—1,600 rad) are necessary so that the level of damage is sufficient to be scored. In this dose range, D0 is about 1.68 Gy (168 rad). If the split-dose technique is used, the Dq is about 2.7 Gy (270 rad). It is possible to estimate the effect of small doses and reconstruct a complete survival curve by giving large doses in multiple small fractions and assuming that the response to each fraction is the same. The result of this reconstruction is shown in Figure 18.12.
Kidney Tubules
A technique using kidney tubules, again developed by Withers and his colleagues, is the first clonal assay for a late-responding tissue. One kidney per mouse is irradiated with a small field and removed for histologic examination 60 weeks later. Figure 18.13 shows sections of normal and irradiated kidneys. For ease of scoring, only those tubules touching the renal capsule are scored, and a tubule is considered fully regenerated only if it is lined with well-differentiated cuboidal or columnar cells with a large amount of eosinophilic cytoplasm. By 60 weeks, tubules either have no surviving epithelial cells or are lined completely with epithelium that has regenerated from a small number of surviving cells, usually one. The number of tubules regenerating in an arbitrary number of sections counted is plotted as a function of radiation dose. The result is shown in Figure 18.14; D0 is about 1.53 Gy (153 rad).
The radiosensitivity of the cells of this late- responding tissue is not very different from that of early-responding tissues, such as the skin or intestinal epithelium. The rate of response, however, is quite different. The time required for depletion of the epithelium after a single dose of 14 Gy (1,400 rad) is about 3 days in the jejunum, 12 to 24 days in the skin, and 30 days in the seminiferous tubules of the testes, but 300 days in the kidney tubules. These results argue strongly that radiation injury in the kidney results from depletion of parenchymal cells and that the slow expression of injury merely reflects the slow turnover of this cell population. Vascular injury is unlikely to be the mechanism underlying the destruction of renal tubules.
Cells Transplanted to Another Site
Bone Marrow Stem Cells
Till and McCulloch developed a system to determine a survival curve for colony-forming bone marrow cells (Fig. 18.15). Recipient animals first are irradiated supralethally with a dose of 9 to 10 Gy (900—1,000 rad), which sterilizes their spleens. Nucleated isologous bone marrow cells taken from another animal are then injected intravenously into the recipient animals. Some of these cells lodge in the spleen, where they form nodules, or colonies, 10 to 11 days later, because the spleen cells of the recipient animals have been sterilized previously by the large dose of radiation. At this time, the spleens are removed and the colonies counted. Figure 18.16 is a photograph of a spleen showing the colonies to be counted.
About l0 cells must be injected into a recipient animal to produce one spleen colony, because the majority of the cells in the nucleated isologous bone marrow are fully differentiated cells and would never be capable of forming a colony. To obtain a surviving fraction for bone marrow cells, a donor animal is irradiated to some test dose, and the suspension of cells from the bone marrow is inoculated into groups of recipient animals that previously had been irradiated supralethally. By counting the colonies in the spleens of the recipient animals, and with a knowledge of the number of cells required to produce a colony in an unirradiated animal (plating efficiency), the surviving fraction may be calculated as follows:
Surviving fraction for a dose D = colonies counted / cells inoculated x plating efficiency
This procedure is repeated for a range of doses, and a survival curve is obtained (Fig. 18.17). These bone marrow stem cells are very sensitive with aD0 of about 0.95 Gy (95 rad) and little or no shoulder to the survival curve.
Mammary and Thyroid Cells
Clifton and Gould and their colleagues developed very useful clonogen transplant assays for epithelial cells of the mammary and thyroid glands. They have
been used largely for cell survival studies, described later, but the initial motivation for their development was to study carcinogenesis in a quantitative system. Most in vitro transformation assays involve fibroblasts, and the bulk of human cancers arise in epithelial cells - hence, the importance and interest in these two systems.
The techniques for these two systems are much the same. To generate a survival curve for mammary or thyroid gland cells in the rat, cells may be irradiated in vivo before the gland is removed from donor animals and treated with enzymes to obtain a monodispersed cell suspension. Known numbers of cells are injected into the inguinal or interscapular white fat pads of recipient animals.
Under appropriate host conditions and grafted cell numbers, the injection of mammary cells gives rise to mammary structures that are morphologically and functionally normal. One such mammary structure may develop from a single cell. By 3.5 weeks after the injection of mammary cells, positive growth is indicated by alveolar units. An example of a milk-filled alveolar unit is shown as an inset in Figure 18.18. If thyroid cells are injected, thyroid follicular units develop (Fig. 18.19).
With either type of cell, a larger number must be injected to produce a growing unit if the cells first are irradiated to a given dose. In practice, some fancy statistics are involved, a discussion of which is beyond the scope of this chapter; in essence, the ratio of the number of irradiated to unirradiated cells required to produce one growing unit (thyroid follicular unit or alveolar unit) is a measure of the cell-surviving fraction corresponding to the dose. This procedure must be repeated for a range of graded doses to generate a survival curve. The resultant survival curve for mammary cells is shown in Figure 18.18. The characteristics of the curve are unremarkable: D0 is about 1.27 Gy (127 rad), and the extrapolation number is about 5, quite typical of rodent cells cultured in vitro. The corresponding survival curve for thyroid cells is shown in Figure 1 8.19. D0 is a little larger than for mammary glands assayed in a similar way, implying that the cells are a little more resistant. Figures 18.18 and 18.19 also show data for cells left in situ for 24 hours after irradiation before being removed and assayed. If this is done, the shoulder of the survival curve is larger because of the repair of potentially lethal damage. This is discussed in more detail in Chapter 5.
An interesting use of these clonogen transplant assays is that the physiologic states of either donor or recipient animals can be manipulated hormonally. For the mammary cell assay, cells may be taken from inactive, slowly dividing glands of virgin rats, from rapidly dividing glands of rats in mid- pregnancy, or from milk-producing glands of lactating rats. For the thyroid cell assay, the physiologic states of both donor and recipient can be manipulated by control of the diet or by partial thyroidectomy.
SUMMARY OF DOSE-RESPONSE CURVES FOR CLONOGENIC ASSAYS IN NORMAL TISSUES
The survival curves for all of the clonogenic assays in normal tissues are plotted together io Figure 18.20. There is a substantial range of radiosensitivities, with shoulder width being the principal variable. In vitro curves for cells from patients with ataxia-telangiectasia also are shown because these are probably the most radiosensitive mammalian cells.
DOSE—RESPONSE RELATIONSHIPS FOR FUNCTIONAL END POINTS
Pig Skin
Pig skin has been used widely in radiobiologic studies because it has many features in common with human skin, such as color, hair follicles, sweat glands, and a layer of subcutaneous fat. In view of these structural similarities, it is not surprising that the response of pig skin to radiation closely resembles that of human skin, both qualitatively and quantitatively.
Fowler and his colleagues pioneered the use of pig skin as a radiobiologic test system. A number of small rectangular fields on the pig’s flank were irradiated with graded doses of x-rays, and the reactions were scored daily using the arbitrary scale shown in Table 18.1. After a single dose of radiation, the reaction becomes apparent after about 15 days and develops as shown in Figure 18.21.
Two phases of the reaction can be distinguished. First, an early wave of erythema occurred (at 10—40 days), which was variable from one animal to another. This represents the uncomfortable “acute” reaction sometimes seen in patients on radiotherapy at about the end of a course of treatment. Second. a more gradual increase to a second broad wave of moderately severe reactions took place (at 50—100 days), representing a more permanent kind of damage. This second wave shows the tolerance of skin to a more serious type of long-term damage and is also a more repeatable and consistent index of radiation damage. It was subsequently found to correlate well with longer-term damage (up to 2 years) and with subcutaneous damage.
The “score” of radiation damage is taken to be the average skin reaction occurring between certain time limits that encompass the medium-term reactions. After a single dose, this might be a 35 - day period between 50 and 85 days after irradiation. For a protracted fractionated regimen, this period of reaction may come later, between days 65 and 100. The average skin reaction in the chosen time period then is plotted as a function of dose; examples of dose—response curves obtained this way are shown in Figure 18.22 for single and fractionated doses.
Late effects also have been studied in pig skin by measuring the contraction that results from fibrosis a year or more after irradiation. A square is tattooed on the skin of the animal in the irradiated field, and the dimensions of this square are recorded as a function of dose as the contraction occurs. This is a primitive but effective measure of late effects.
Many of the important early studies on the fractionation effects of x-rays and the comparison of x-rays with fast neutrons were performed with this biologic system. One overwhelming advantage is that data obtained this way can be extrapolated to the human with a high degree of confidence. The disadvantage is that the animals are large and awkward to work with, and their maintenance involves a considerable expense.
Rodent Skin
Because of the inconvenience and expense of using pigs, the skin of the mouse leg and foot is commonly used instead. One hind leg of each animal is irradiated; the other serves as a control. The skin response is observed each day after irradiation and is scored according to the arbitrary scale shown in Table 18.2. Various doses are used. The progressive development of the reaction after ten doses of 6 Gy (600 rad) each is illustrated in Figure 18.23; each point represents the mean of several animals. Reactions appear by about the 10th day, peak by 20 to 25 days, and then subside. The second wave of the reaction, noted for pig skin, is not seen in mice but is observed in rats. A dose - response curve is obtained by averaging the skin reaction over a period of time and plotting this average as a function of dose.
Early and Late Response of the Lung Based on Breathing Rate
Travis and her colleagues developed a noninvasive assay of breathing frequency to assess both early and late damage in mouse lungs. Breathing frequency increases progressively with dose after a threshold of about 11 Gy (1,100 rad) (Fig. 18.24). The increased breathing frequency in rodent lungs at 16 and 36 weeks is associated with the early response (i.e., pneumonitis); by 52 weeks, the elevated breathing frequency is associated with the late response (i.e., fibrosis). This is a simple but highly quantitative and reproducible system.
Spinal Cord Myelopathy
A dose - response relationship can be determined for late damage caused by local irradiation of the spinal cords of rats. A number of investigators have worked with this system, notably van der Kogel. After latent periods of 4 to 12 months, symptoms of myelopathy develop, the first signs of which are palpable muscle atrophy, followed some time later by impaired use of the hind legs. Figure 18.25 shows the steep dose - response relationship for hind-limb paralysis following the irradiation of a section of the spinal cord in rats. These data also show the dramatic sparing that results from fractionation; this is discussed further in another section of this chapter.
The various syndromes of radiation-induced injury in rodent brain and spinal cord are very similar to those described in humans. Lesions observed within approximately the first 6 months after irradiation are limited primarily to the white matter and range between early diffuse or focal demyelination and extensive necrosis. Different pathogenic pathways toward the development of white-matter necrosis have been proposed, with the glial and vascular tissue components the major targets. The most common type of late delayed injury peaks at 1 to years postirradiation and almost certainly has a vascular basis. Another type of late injury that has been described more recently in various species, including humans, is slowly progressive glial atrophy. This lesion is not associated with necrosis but occurs diffusely and at lower doses. With improvements in diagnostic procedures such as magnetic resonance imaging, glial atrophy may become more frequently recognized adverse effect of brain tumor therapy.
Latency
Over a dose range of about 25 to 60 Gy (2,500- 6,000 rad), delivered in single doses, the general tendency is a decreasing latency with increases in dose of approximately 2 days/Gy (2 days/100 rad). There is a considerable variation with animal strain. as well as with the region of the cord irradiated.
In terms of mechanisms, demyelination or slowly progressive atrophy is probably a consequence of interference with the slow continuous turnover of oligodendrocytes by killing of glial progenitor cells. Vascular injury may accelerate, precipitate, or even initiate the white-matter changes leading to necrosis. This is an area of some controversy.
Fractionation and Protraction
The effect of dose fractionation and protraction on tolerance to radiation has been investigated extensively in the rat spinal cord and to a lesser extent in the mouse, monkey, and guinea pig. Because these systems turn over slowly, there is little influence of overall treatment time up to any conventional clinical regimen of 6 to 8 weeks. On the other hand, dose per fraction is very important (Fig. 18.25), with the dose to produce paralysis increasing dramatically with number of fractions. The effect of a large number of very small fractions also has been investigated. Figure 18.26 shows the relation between total dose and dose per fraction to produce paralysis in 50% of rats from irradiation of a short length of cervical spine. The smooth curve is an iso- effect curve calculated for the very low aIjS value of 1,5 Gy (150 rad). The experimental data suggest that the linear-quadratic (L.Q) model overestimates the tolerance for small doses per fraction of less than 2 Gy (200 rad). However, this may be a result of incomplete repair, because in these experiments, the interfraction interval was only 4 hours. There is good reason to believe that repair of sublethal damage takes place slowly in this normal tissue, and indeed, repair may be biphasic, with “fast” and “slow” components. For this reason, if multiple doses per day are used to the spinal cord, the interfraction interval should he at least 6 to 8 hours.
Volume Effects
The total volume of irradiated tissue usually is assumed to have an influence on the development of tissue injury. The spinal cord is perhaps the clearest case in which the functional subunits (FSUs) are arranged in linear fashion, like links in a chain. Figure 18.27 shows the relation between tolerance dose and the length of cord irradiated in the rat. For short lengths of cord, below 1 cm, tolerance in terms of white-matter necrosis shows a marked dependence on the length of cord irradiated. Late vascular injury shows less dependence on cord length. Beyond a few centimeters, the tolerance is virtually independent of the length of cord irradiated. This would be predicted from the linear arrangement of the functional subunits. A chain is broken whether one, two, three, or more links are removed.
Retreatment after Long Time Intervals
The spinal cord does recover to some extent after long time periods following irradiation. The extent of the recovery depends, of course, on the first treatment - that is, what fraction of tolerance was involved. Experiments with rats indicate that after an initial treatment to 50% tolerance, the retreatment tolerance approaches 90% of the tolerance of the untreated control group by about a year after the initial irradiation. If the initial treatment represented a larger fraction of tolerance, the retreatment that can be tolerated is reduced.
INFERRING THE RATIO α/β FROM MULTIFRACTION EXPERIMENTS IN NONCLONOGENIC SYSTEMS
The parameters of the dose - response curve for any normal tissue system for which a functional end point can be observed may be inferred by performing a multifraction experiment. Take, for example, an experiment in which mouse foot skin reaction is scored. Doses that result in the same skin reaction (e.g., moist desquamation over 50% of the area irradiated) if delivered as a single exposure in a multifraction regimen (e.g., 5, 10, or 20 fractions) must be determined experimentally. A number of assumptions must be made:
1. The dose - response relationship is represented adequately by the linear-quadratic formulation:
S = e^ (- αD – βD2)
in which S is the fraction of cells surviving a dose, D, and a and ft are constants.
2. Each dose in a fractionated regimen produces the same biologic effect.
3. Full repair of sublethal damage fakes place between dose fractions, but no cell proliferation occurs.
Suppose the total dose, D, is divided into n equal fractions of dose d. The previous equation then can be rewritten:
If the reciprocal of the total dose (1/nd) is plotted against the dose per fraction (d), a straight line results, as shown in Figure 18.28. The intercept on the ordinate gives α/logeS; the slope gives β/logeS. In general, the value of logeS is not known unless other cell survival studies are available, but the ratio of the intercept to the slope provides an estimate of α/β.
Multifraction experiments have been performed and estimates of a/fl made for essentially all of the normal-tissue end points described in this chapter. One of the important conclusions arrived at is that the value of a/fl tends to be larger for early- responding tissues, about 10 Gy (1,000 rad), than for late-responding tissues, about 2 Gy (200 rad).
Because α/βis the dose at which cell killing by linear and by quadratic components of radiation damage are equal (Chapter 3), the implication is that dose - response relationships for late-responding tissues are “curvier” than for early-responding tissues. The importance of this conclusion becomes evident in the discussion of fractionation in radiotherapy in Chapter 22.
posted by stefanoxx at 8:40 AM
0 comments
