Apoptosis and Response to Radiation: Implications for Radiation Therapy

Introduction

A poptosis, or programmed cell death, is characterized by specific features that allow it to be distinguished from necrosis in tissues.[1] This distinction is necessary because, whereas apoptosis occurs as part of many normal biologic processes, especially during the development of an organism, it may also play a critical role in certain pathologic conditions, including cancer.[2]

The importance of apoptosis in cancer biology has been generally appreciated only in recent years, and this has led to a surge of reports of its occurrence in various normal and tumor cell systems treated with radiation, drugs, or biologicals. Although these reports clearly demonstrate that apoptosis occurs in model cell systems in response to cytotoxic treatments, the relative impact of apoptotic cell death on tumor response to therapy is essentially unknown. This is due largely to the fact that most studies to date have focused on the assessment of apoptosis in cultured cells. Apoptosis in human tumors following therapy has not been systematically evaluated due to technical and logistical constraints.

Nonetheless, research on radiation-induced apoptosis has uncovered new fundamental concepts relating to the basic processes responsible for cell death following irradiation and, most important, is beginning to reveal the details of the biochemical and molecular pathways that regulate these processes. This knowledge will most likely lead to entirely new strategies for modulating radiation-induced cell killing in tumors or normal tissues for therapeutic advantage.

The ability to cure a tumor with radiation may depend on various factors, including tumor cell proliferation kinetics, tumor hypoxia, the number of tumor clonogens, and intrinsic tumor cell radiosensitivity.[3] Apoptosis may play a role in more than one of these factors, but, because apoptosis is known to occur in cells irradiated in vitro, the most obvious role is that apoptotic propensity may dictate intrinsic tumor cell radiosensitivity. Interest in understanding the mechanistic basis for intrinsic cell radiosensitivity has been stimulated by reports that tumor types whose cells display radioresistance in culture are associated with treatment failures.[4] However, the question of what determines a cell's intrinsic radiosensitivity remains a central issue in radiobiology.

As discussed in a recent review, the "classic" view held that the end point relevant to radiation therapy was "reproductive cell death," which results directly from unrepaired DNA strand breaks.[5] Radiobiologists now recognize the need to reevaluate this classic view and consider new concepts, such as effects on cellular membranes, signal transduction pathways, induction of gene expression, cell-cycle regulation, oxidative stress processes, and the influence of growth factors. The need to reevaluate the role of various pathways of cell death in response to radiation has also been recently appreciated. Thus, whereas for many years radiobiologic research focused exclusively on reproductive cell death, much current effort is also being directed toward understanding the role of apoptosis.

This article will briefly review our current understanding of the role of apoptosis in radiation response and will then discuss the implications of this role in radiation therapy.

Model Systems

In Vitro Cell Systems

In vitro cell systems have been used extensively in radiobiology since 1955, when Puck and Marcus[6] developed clonogenic assays for cell survival using cultured mammalian cells. This important advance provided a model system for the study of the relationship of basic cellular processes that modify radiation response (eg, DNA damage and repair mechanisms) to loss of cell reproductive capacity. Although the knowledge gained using these cell systems has been critical to our understanding of the effects of radiation on cells, such cell systems necessarily focused investigators' attention on reproductive cell death, to the detriment of research centered on other modes of cell death, such as apoptosis.

The term "apoptosis" was coined by Kerr et al in 1972 to describe a process of programmed cell death distinct from necrosis.[1] Whereas apoptosis has become a "hot" area of research during the last few years, radiobiologists have been aware of it for at more than 40 years. Prior to 1972, what we now refer to as apoptosis was included in a process called "interphase death," a term that referred to the death of cells before their first postmitotic division. Fundamentals of Radiobiology, a classic textbook written by Bacq and Alexander and published in 1961,[7] clearly distinguishes interphase death from reproductive cell death, which is termed "mitotic death."

The definitions of Bacq and Alexander[7] do not distinguish between necrosis and apoptosis, however. That distinction is one of the major contributions made by Kerr et al in 1972.[1] One can find numerous reports from the 1950s and early '60s describing cellular responses to radiation identical to apoptosis. One of these, a set of cinemicrographic observations by Schrek[8] published in 1955, provides unequivocal photographic illustrations of rabbit lymphocytes undergoing apoptosis within 4 hours after 10 Gy of irradiation.

Apoptosis vs Secondary Apoptosis--The rather large number of papers published during the last few years reporting apoptosis in irradiated cultured cell systems has possibly led to some misconceptions about the role of the apoptotic process in radiation response. In addition to the occurrence of apoptosis as part of interphase cell death, apoptosis has been reported in cultured cell systems in which the primary mode of death is reproductive cell death. In these latter cases, apoptosis occurs after the first postirradiation mitotic division, and therefore, may have consequences that differ from those occurring when cells die before the first postirradiation division.

This issue has been discussed in detail in another review,[9] but some additional clarification may be warranted. Throughout the remainder of this article, I will refer to cell death as "apoptosis" when it occurs before the first postirradiation mitosis. Apoptosis, therefore, occurs rather quickly after irradiation, usually within 4 to 6 hours, and is abundant in those cell systems in which apoptosis is the primary mode of cell death.

Cells with the morphologic features of apoptotic cells are also observed in cultured cell systems in which reproductive cell death is the primary mode of cell demise. These cells appear much later after irradiation, ie, 24 to 96 hours, and usually with less abundance. Such cells, therefore, may be displaying the features of apoptosis subsequent to reproductive cell death. This process will be referred to as "secondary apoptosis."

Loss of Tumor Clonogens--The distinction between apoptosis and secondary apoptosis is not trivial in the context of radiation response, for two reasons. First, cells that die by apoptosis directly contribute to radiation response in appropriate cultured cell systems because, in those cases, there is usually a one-to-one relationship between an apoptotic cell and the loss of cells from the clonogenic pool. Thus, a radiation dose that would induce 50% of the cells to undergo apoptosis would reduce clonogenic survival to 50%.

Secondary apoptosis does not contribute to radiation response in the same way. In cultured cell populations in which secondary apoptosis occurs, cells die primarily by reproductive cell death; this process is quite sufficient, in and of itself, to remove cells from the clonogenic pool, usually by inducing a permanent growth arrest, a senescence-like state. There is no need for apoptosis under these conditions.

Thus, secondary apoptosis probably does not contribute to overall cell killing and radiation response because it most likely occurs in cells that have already permanently lost their reproductive capacity regardless of whether they display the features of apoptosis. In general, there is not a one-to-one relationship between the proportion of cells undergoing secondary apoptosis and cell killing because a much higher proportion of the cells in such a population usually undergo reproductive cell death than display apoptotic-like features.

Cell Radiosensitivity--Second, and most important, cell populations for which apoptosis is the primary mode of cell death following irradiation may be more sensitive to radiation than are cells whose primary mode of death is reproductive cell death. In such cultured cell systems, substantial apoptosis is usually induced by doses of 5 Gy or less, and the apoptotic cells are visualized within a few hours of irradiation. In contrast, secondary apoptosis generally requires doses higher than 5 Gy and appears more than 24 hours after exposure. Ian Radford clearly established these principles in a 1991 paper,[10] and they have been verified in subsequent reports.[11,12]

Thus, the observation of cells with the features of apoptosis is not necessarily indicative of cell radiosensitivity[12] because the relative sensitivity of the cells in question may depend on whether they are undergoing apoptosis or secondary apoptosis. It should also be pointed out that radiation-induced apoptosis appears to be a feature of cultured cells of hematologic or lymphoid origin, whereas fibroblasts and cultured cells of epithelial origin tend to undergo reproductive cell death and secondary apoptosis following irradiation.[12]

In Vivo Cell Systems

Compared to what has been accomplished using cultured cell systems, evaluation of radiation-induced apoptosis in systems irradiated in vivo has been very limited. Nonetheless, several critical evaluations of apoptosis in irradiated normal and tumor tissues have revealed important differences between in vitro and in vivo systems.

Early Studies--Work in this area prior to 1980 was reviewed by Kerr and Searle.[13] They pointed out that Pratt and Sodicoff[14] reported the presence of apoptotic bodies in irradiated salivary glands and that Potten[15] observed apoptosis in the epithelium of intestinal crypts following irradiation. In addition, Kerr and Searle[13] observed apoptosis in a transplantable mouse tumor, a sarcoma, that had received radiation.

Thus, this early work established that, in in vivo systems, radiation-induced apoptosis could occur in cells of nonhematologic origin. Through a meticulous quantitative analysis of radiation-induced apoptosis in the intestinal crypt, Potten[15] established two critical points concerning this mechanism of cell death: (1) Apoptotic cells appeared very quickly following treatment, with their numbers peaking between 3 and 4 hours after irradiation. (2) Only some of the cells in the crypt responded to radiation by undergoing apoptosis, but those that did were exquisitely sensitive to this mode of cell death; apoptotic figures were detected following radiation doses as low as .05 Gy.

M. D. Anderson Studies--Since these early reports, systematic assessments of radiation-induced apoptosis have been carried out in normal and tumor tissues. Our group's interest in apoptosis was stimulated by an observation made in the radiation oncology clinic, where irradiation of the salivary glands is often unavoidable in the treatment of head and neck cancer. The response of these glands to radiation is remarkable, in that some patients develop decreased salivary function during the first week of therapy, having received less than 10 Gy.[16]

To determine the underlying mechanisms responsible for these troublesome low-dose sequelae, adult female monkeys were irradiated with single radiation doses of 2.5 to 15 Gy.[17] Damage to the salivary gland was assessed from sequential biopsy specimens taken 1 to 72 hours after irradiation. The proportion of apoptotic cells in the tissues was determined by counting the cells displaying the morphologic features of apoptosis. Apoptotic cells could be observed as early as 1 hour after irradiation, and most such cells appeared within 24 hours. Moreover, apoptosis was seen even with the lowest dose used, 2.5 Gy.

These findings are similar to those observed for cultured cell systems undergoing apoptosis with respect to dose-responsiveness and kinetics of appearance. They also suggest that, whereas the salivary gland cells are terminally differentiated, cells need not be in the cell cycle or even have reproductive capacity to undergo apoptosis following irradiation.

Heterogeneity of Apoptotic Propensity--Examination of apoptosis in model tumor systems has led to the appreciation that apoptosis may play a role in the response of at least some types of tumors to radiation. The intent of our initial studies[18] was simply to determine whether apoptosis was a feature of irradiated tumors. Two transplantable murine tumors were chosen based on their known radiation response: a very radioresistant (TCD₅₀ more than 80 Gy) hepatocarcinoma, HCa-I, and a moderately sensitive (TCD₅₀ = 53 Gy) ovarian adenocarcinoma, OCa-I. (TCD₅₀ refers to the dose required to cure 50% of the tumors.) These tumors were grown in the hind legs of mice and were given radiation doses of 25 Gy or more; histologic sections were prepared from tumors removed at 6, 24, 96, and 144 hours after treatment.

This first experiment showed that apoptosis occurred in the OCa-I tumor but not in the HCa-I tumor. Moreover, the apoptotic index was highest at 6 hours following 25 Gy of radiation, declined thereafter, and did not increase with doses over 25 Gy.

A subsequent, more detailed study[19] pinned down the time course and dose response for apoptosis induction in the OCa-I tumor. This study showed that the apoptotic index in the OCa-I tumor peaks at about 30% to 35% of the cells in the histologic sections between 3 and 4 hours after irradiation (Figure 1). Doses of 2.5 Gy induced substantial apoptosis, and the dose response actually leveled off at doses higher than about 7.5 Gy (Figure 2), suggesting that only a subset of cells in the tumor have the propensity for radiation-induced apoptosis. Thus, it became clear from our first two studies that apoptotic propensity was heterogeneous both among different tumor types and even among cells within a given tumor.

These instances of heterogeneity were confirmed in a more elaborate assessment of radiation-induced apoptosis in 14 additional types of murine tumors.[20] This analysis illustrated that some types of tumors, such as adenocarcinomas of the mammary gland and ovaries and lymphomas, have an apoptotic response, whereas other types, such as squamous cell carcinomas, hepatocarcinomas, and fibrosarcomas, do not. Moreover, for tumors that displayed an apoptotic response, their dose responses plateaued at an apoptotic index of 30% at doses higher than 10 Gy.

For many of these tumors, we had previously determined their TCD₅₀ and specific growth delay characteristics in response to single doses of radiation. Therefore, we produced correlation plots (Figure 3) of radiation-induced apoptosis vs TCD₅₀ and specific growth delay for the murine tumors included in the analysis described above. These plots showed that tumors that responded to radiation with significant apoptosis tended to have lower TCD₅₀ values (.1 less than P less than .2) and longer specific growth delays (P less than .05).

For each respective tumor, we also plotted the value for spontaneous apoptosis measured in the nonirradiated tumor vs the value for radiation-induced apoptosis (Figure 4). This correlation was highly significant (P less than .001), suggesting that spontaneous levels of apoptosis predict the apoptotic response to treatment.

Impact of Dose Fractionation--The studies described above utilized single doses of radiation, whereas in radiation therapy the dose is given in fractions. The possible influence of dose fractionation was assessed using the OCa-I tumor.[21] Two protocols were tested: (1) two doses of 12.5 Gy separated by various intervals up to 5 days and (2) five daily fractions of 2.5 Gy.

These experiments showed that the protocol using two 12.5-Gy doses produced a net total proportion of apoptotic cells of about 45% when the two doses were separated by 5 days (total dose, 25 Gy). The daily 2.5-Gy protocol produced about 50% net apoptotic cells after 5 days (total dose, 12.5 Gy), and a single dose of 25 Gy produced only 36% apoptotic cells. Thus, an apoptotic subpopulation of cells reemerged between doses in the fractionated protocols, with the daily 2.5-Gy protocol being the most effective.

Role of Apoptosis in Tumor Response

All of the observations discussed thus far are consistent with the possibility that apoptosis plays a significant role in tumor response to radiation therapy. However, these are only correlations; cause-and-effect relationships are more difficult to establish. Moreover, it is hard to envision how apoptotic indexes of 30% to 35% dictate tumor response in light of a discussion by Dewey et al,[9] which illustrated that killing this proportion of cells in a tumor could reduce the TCD₅₀ by only about 4 to 5 Gy. In contrast, the difference in TCD₅₀ between the sensitive and resistant tumors used in our analysis was 20 to 30 Gy. Therefore, the correlations presented above, while intriguing, cannot directly account for the sensitivity of the tumors analyzed in the study.

One possible explanation is that apoptosis indirectly influences tumor response. The spontaneous levels of apoptosis scored in the untreated tumors may be depicting the rates of cell loss in these tumors, and this factor may, in turn, dictate the number of tumor clonogens, a critical parameter governing tumor response to radiation. The number of clonogens in these transplantable tumors can be estimated from the TD₅₀ (a value representing the number of cells required to yield a tumor in 50% of the injection sites in a tumor transplantation assay). In an assessment of the TD₅₀ values for the murine tumors used in our studies,[22] we confirmed that tumors that have a high spontaneous apoptotic index and respond to radiation by undergoing apoptosis tend to have a high TD₅₀, suggesting that their number of tumor clonogens is low compared to those relatively radioresistant tumors that lack apoptosis as a feature.