Chemical radioprotection: A critical review of amifostine as a cytoprotector in radiotherapy☆

Amifostine 

More than 50 years ago, it was realized that the thiol-containing amino acid cysteine exhibits remarkable radioprotective properties.14 During the time of the nuclear arms race, this finding had immense importance and led to a large research program at the Walter Reed Army Institute of Research. More than 4,000 sulfhudryl-containing substances with radioprotective properties were tested, but only the agent WR-2721, later known as amifostine, was found to exhibit acceptable toxicity.15 The idea of using amifostine in oncology was encouraged by preclinical data suggesting that amifostine offers a selective protection of normal tissues from damage induced not only by irradiation but also from chemotherapy.16, 17 During the 1980s, clinical phase 1 to 2 studies provided evidence that the use of amifostine was feasible and that amifostine might protect normal tissues from both irradiation and chemotherapy. In 1996, the Food and Drug Administration (FDA) registered amifostine for use as a cytoprotective agent with cisplatin-based chemotherapy against ovarian cancer. More recently, a phase 3 study has provided strong evidence that amifostine prevents xerostomia in patients treated with radiotherapy for head and neck cancer. Based on this, the FDA approval was extended in 1999 to postoperative radiotherapy for head and neck cancer as well.6 Also, in the European Union, amifostine has been approved for this indication but without distinguishing between definite and postoperative radiotherapy. The American Society of Clinical Oncology (ASCO) has recommended the use of amifostine for prevention of xerostomia in head and neck irradiation.18

Pharmacologic properties of amifostine 

Amifostine is a very hydrophilic compound that does not readily cross cell membranes.17 After intravenous administration, amifostine is rapidly dephosphorylized to its active metabolite WR-1065 and cleared from the plasma with a half-life ranging from 1 to 3 minutes.19 Dephosphorylisation of amifostine is either caused by spontaneous nonenzymatic hydrolysis or by a catalysed process involving alkaline phosphatase with a pH optimum at 8 to 9.16 Studies with radiolabeled drugs injected into mice have shown that WR-1065 accumulates in most normal tissues shortly after amifostine is cleared from the plasma.16, 20 Uptake of WR-1065 varies considerably between different tissues. Extensive uptake is seen in salivary glands, kidneys, and intestinal mucosa, whereas markedly lower uptake generally is seen in tumour tissues.16, 19, 20 Amifostine and metabolites do not cross the blood-brain barrier.21, 22 Inside the cell, WR-1065 is further metabolized to the disulfide, WR-33278 (Fig 2), that may also cause radioprotection, although to a much lesser extent.23, 24

View Large Image Download to PowerPoint

Fig. 2. Metabolism of amifostine (WR-2721). Amofostine is converted to its active metabolite WR-1065 by either spontaneous dephosphorylation or by an enzymatic process involving alkaline phosphatase. Within the cell, WR-1065 is further metabolized to the symmetrical disulfide WR-33278 that has structural similarities to naturally occurring polyamines.

The mechanism of cytoprotection against ionizing radiation is complicated and not entirely understood: WR-1065 and, to a lesser extent, WR 33278 act as free radical scavengers that protect subcellular components like membranes and DNA from damage.15, 25, 26 Because WR-1065 reacts with free radicals in competition with oxygen, the protection obtained by scavenging is highly dependent on oxygen tension. Protection is maximal at intermediate levels of oxygen. At higher oxygen concentrations, the protection is gradually lost. The degree of protection is also diminished at low oxygen tensions when scavenging becomes less important as the lack of oxygen itself provides radioprotection.15, 26 WR-1065 may also react directly with oxygen and thus protect the cell by creating local hypoxia at the target.27 Additional complex mechanisms are undoubtedly involved: thiols may facilitate chemical repair processes by donation of hydrogen28 and may also decrease accessibility of radiolytic attack sites by induction of DNA packaging.23 These mechanisms may yield oxygen independent protection and explain the protection observed with densely ionising radiation such as neutrons.29, 30 Furthermore, the amifostine metabolite WR-33278 exhibits structural similarities to naturally occurring polyamines31 and may affect processes related to DNA synthesis, DNA repair, gene expression, and cell cycle progression.31, 32, 33 These characteristics of WR-33278 may account for the antimutagenic properties exhibited by amifostine even when administered up to 3 hours after irradiation.34

Quantitatively, there is great variability in the radioprotection achieved by amifostine in different normal tissues: protection factors range from 3 in the haematopoietic system and salivary gland to near 1 in lung, kidney, and bladder25, 35, 36, 37, 38 (Table 1).

Table 1. Examples of protection factors achieved by amifostine in different normal tissues and tumors

	Tissue	Protection Factor
	Salivary gland	2.3-3.3
	Bone marrow	1.8-3.0
	Jejunum	1.5-2.1
	Skin	1.4-2.1
	Testis	1.5-1.6
	Lung	1.2-1.4
	Kidney	1.3-1.5
	Bladder	1.3-1.5
	Tumors	1.0-2.8

NOTE. Preclinical data from references 25, 35, and 63.

As mentioned earlier, preclinical studies with tumor bearing mice have shown that the uptake of WR-1065 in most normal tissues far exceeds uptake in tumors.16, 20 Based on this observation, it has been assumed that amifostine should be considered an inactive pro-drug that is preferentially activated in normal tissues. Such normal tissue selectivity is believed to mainly be because of lower activity of alkaline phosphatases in tumor vessels compared with vessels of normal tissues.41 An acidic environment in tumors, causing inhibition of alkaline phosphatases,41 may enhance this selective uptake. Alternatively, the differential uptake could also be explained by the abnormal nature of the vasculature of tumors. Furthermore, the importance of WR-1065 in tumor cells may be reduced because of hypoxia, often present in tumors.25 Nevertheless, the selective uptake of amifostine in normal tissues has recently been brought to question by a study addressing the effect of amifostine on biotransformation and distribution of the platin analogue ormaplatin in sarcoma bearing Fisher rats. This study failed to show any selective distribution of WR-1065 in normal tissues.42

Toxicity profile, administration, and recommended dose of amifostine 

Most side effects related to amifostine are dose dependent. The major dose-limiting toxicity of amifostine is transient hypotension. Even at the low doses used in radiotherapy close monitoring of blood pressure is required. Most patients receiving amifostine with radiotherapy require antiemetics. Other side effects of amifostine include flushing, somnolence, metallic taste, and transient hypocalcaemia.6, 18

During the 1980s, phase 1 trials were conducted with amifostine. A maximum tolerated dose was not reached, but an acceptable tolerated dose for use as a cytoprotective agent in phase 2 trials with chemotherapy was established, in the range of 740 to 910 mg/m2.43 According to guidelines from the FDA and ASCO the recommended dose for reduction of nephrotoxicity in patients receiving cisplatin-based chemotherapy is 910 mg/m2 amifostine administered intravenously (IV) before each chemotherapy session.6, 18 For use in radiotherapy, a variety of doses and schedules have been tested. However, relatively small cohort sizes have hindered dose-response assessments. The dose recommended in the guidelines from the FDA and ASCO for reduction of radiation-induced xerostomia is 200 mg/m2 amifostine given as an IV push over 3 minutes, 15 to 30 minutes before each fraction.6, 18 Higher doses given 5 days a week have been applied in several clinical trials.18 However, in a recent study, twice daily administration of 150 mg/m2 amifostine 5 days a week caused substantial toxicity, and amifostine had to be discontinued in more than one third of the patients.44

The IV route of administration is quite labor intensive and requires repeated vein punctures. Therefore, alternative regimens have been tested. Phase 1 clinical data suggest that subcutaneous administration of 500 mg amifostine results in an area under the concentration-time curve for the bound form of WR-1065 of 68% compared with 200 mg/m2 IV.45 Preliminary data indicate that subcutaneous administration may offer a more favorable toxicity profile with less hypotension. In a randomized phase 2 study that included 140 patients with different tumor sites, subcutaneous injection of amifostine 500 mg caused significant reduction of mucosal reactions, skin, and bladder toxicity. Hypotension was not observed, but nausea and cumulative asthenia were quite frequent.46 A recent phase 2 trial in head and neck cancer showed that a dose of 500 mg amifostine subcutaneously provided comparable protection against xerostomia as 200 mg/m2 amifostine IV. Cutaneous reactions were the most significant side effect.47

Preclinical data on tumor protection 

A larger number of preclinical studies have addressed the crucial question of whether or not amifostine causes tumor protection from cytotoxic therapies. Data from these studies are not entirely consistent. A substantial number of preclinical studies have concluded that amifostine does not cause tumor protection.24, 48, 49, 50 However, several preclinical studies do actually report significant tumor protection, with a protection factor as high as 2.8.51, 52, 53, 54, 55, 56, 57, 58, 59, 60 For instance, in a canine fibrosarcoma model, which in many respects resembles the clinical situation, administration of amifostine was associated with a significant reduction of tumor control. Furthermore, no protection was observed against late damage to skin, muscles, or bones.61 These conflicting findings have occasionally been subjected to controversies within the scientific community.62, 63, 64, 65

The inconsistencies observed in preclinical data are rather difficult to interpret and may possibly be caused by a variety of factors. Certain strains of mice may develop hypothermia and hypotension when given relatively large doses of amifostine.49 However, tumor protection has also been observed in studies using lower doses of amifostine,37, 66 and in a range of mouse strains and other species in which hemodynamic problems were apparently insignificant.61 In this context, it is important to remember that even mild and persistent hypotension, a well-known side effect to amifostine, may actually cause clinically significant radioprotection.67, 68 Tumor size may also affect the degree of radioprotection. More protection has been observed in small and well-oxygenated tumors compared with large and hypoxic tumors,69 possibly reflecting that the degree of radioprotection offered by amifostine is heavily dependent on oxygen tension.25 Also, less tumor protection is observed with large single radiation doses than fractionated dose irradiations.37, 56, 57, 63, 70 This may be explained by postulating that the response of a single dose is predominantly determined by the irradiation effect on hypoxic tumor cells, whereas during fractionated irradiation, the response of oxygenated cells becomes more significant.37 Because most studies in mice have been performed with relatively big tumors exposed to large single doses of radiation, some preclinical data may actually underestimate the tumor protection that might be obtained with clinical fractionated dose schedules. Therefore, even though artifacts may cause some of the tumor protection observed in preclinical studies, the risk of amifostine-induced tumor protection can certainly not be ruled out.

Clinical data 

During the last decade an increasing number of clinical trials combining amifostine and radiotherapy have been performed. Most studies have been either phase 1 to 2 or retrospective in nature.50 Importantly. Only a limited number of studies have directly addressed the influence of amifostine on therapeutic index (Table 2).

Table 2. Randomized clinical trials addressing the influence of amifostine on tumor control and normal tissue complications

	Author	Year	Trial	Site	Treatment	N
	Komaki	2002	Phase 3	Lung	RT + CT	60
	Antonadou	2002	Phase 2	Head and neck	RT + CT	50
	Antonadou	2001	Phase 3	Lung	RT	146
	Bennett	2001	Phase 2	Head and neck	RT + CT	28
	Brizel	2000	Phase 3	Head and neck	RT	315
	Bourhis	1999	Phase 2	Head and neck	RT	26
	Peters	1999	Phase 2	Head and neck	RT + CT	28
	Vacha	1999	Phase 2	Head and neck	RT + CT	47
	Bohuslavizki	1998	Phase 2	Thyroid	I-131	50
	Buntzel	1998	Phase 2	Head and neck	RT + CT	39
	Liu	1992	Phase 2	Rectum	RT	100

Abbreviations: RT, radiotherapy; CT, chemotherapy; I-131, radio-iodine treatment.

Head and neck irradiation 

In head and neck cancer, clinical studies have mainly focused on the ability of amifostine to protect against xerostomia and mucositis. In 1994, it was shown that a 100 mg/m2 dose of amifostine, given before each fraction of definite radiotherapy over 6 to 7 weeks, was tolerable and improved salivary gland function.71 A randomized, double-blinded trial of 50 patients treated with high-dose radioiodine for thyroid carcinoma showed that an amifostine dose of 500 mg/m2 IV before treatment significantly protected salivary glands against radiation-induced damage.72 In a recent phase 2 study, 26 patients with inoperable squamous cell carcinoma of the head and neck were treated with highly accelerated radiotherapy (46 Gy in 22-23 days, 2 Gy per fraction twice a day). Patients were randomized to radiotherapy with or without 150 mg/m2 amifostine IV before each fraction. Both incidence and duration of mucositis was significantly reduced with amifostine. However, even in the group treated with amifostine, 11 out of 13 patients required a feeding tube. Furthermore, the tolerance of the twice daily amifostine schedule was quite poor. In 5 out of 13 patients, amifostine treatment had to be discontinued because of either vomiting, elevated liver enzymes, or generalized erythema. Locoregional control was not different between the 2 arms with a follow-up of 15 months.44 The statistical power of the study is too low to allow a meaningful conclusion with respect to tumor protection.

A larger multicenter trial recruited 315 patients to receive definite or adjuvant radiotherapy with or without amifostine for head and neck cancer.73 In all patients, at least 75% of both parotid glands were in the radiation field. The patients were treated with daily fractions of 1.8 to 2 Gy to a total dose of 50 to 70 Gy and randomized to receive radiotherapy with or without 200 mg/m2 amifostine IV before each fraction. Amifostine significantly reduced the incidence of moderate and severe acute xerostomia from 78% to 51%. At 1 year, the incidence of xerostomia was also significantly reduced in patients receiving amifostine. Hovever, there was no significant difference in the incidence of acute mucositis between the treatment groups. Improved Patient Benefit Questionnaire scores were observed in patients receiving amifostine, perhaps because of a decreased incidence of xerostomia.74 The patients were not evaluated for other normal tissue endpoints. With and without amifostine the 2-year locoregional control rates were 58% versus 63%, disease free survival rates were 53% versus 57%, and overall survival rates were 71% versus 66%.73 Based on this, it has been concluded that amifostine does not compromise antitumor efficacy. However, even this study, being the largest ever conducted with amifostine, does not have sufficient statistical power to detect a 5% to 10% decrease in local tumor control. Furthermore, the use of surgery in 2/3 of patients hinders the evaluation of the influence of amifostine on tumor tissues.64

The use of amifostine in concomitant radiochemotherapy for head and neck cancer has been evaluated in a small number of trials. Three randomized trials, with an accrual of 39, 28, and 28 patients, respectively, have tested the effect of amifostine in regimens consisting of conventional radiotherapy up to 60 Gy (or 63 Gy) in conjunction with carboplatin 70 mg/m2 on days 1 to 5 and days 21 to 25 (or 29-33). A dose of 500 mg amifostine IV (equivalent range of 250-340 mg/m2) was given before each dose of carboplatin. Two studies reported significant reduction of severe mucositis, acute xerostomia, and thrombocytopenia in the amifostine arm,75, 76 whereas 1 study concluded that amifostine in this form of application did not provide any relevant reduction of treatment related toxicity.77 In a similar treatment schedule, the administration of 250 mg amifostine IV before each radiotherapy session did not cause significant protective effects and a higher amifostine dose was suggested.78 Another trial tested the prophylactic properties of amifostine in 50 head and neck cancer patients treated with conventional radiotherapy (60-74 Gy) in conjunction with carboplatin 90 mg/m2 once a week. A dose of 300 mg/m2 amifostine IV was given before each radiotherapy session on a randomized basis. Significant reduction of xerostomia, mucositis, and dysphagia were observed with amifostine. Transient hypotension was reported in 13.6% of patients receiving amifostine.79 Although none of these studies revealed any signs of compromised antitumor efficacy in patients treated with amifostine, the lack of statistical power in these studies hinders any firm conclusions being drawn regarding tumor protection.64

Thoracic irradiation 

A multicenter phase 3 trial has investigated the effect of amifostine in 146 patients given conventional radiotherapy (55-60 Gy) for advanced lung cancer.80 Patients in the experimental arm received amifostine 340 mg/m2 IV before irradiation. The incidence of pneumonitis, lung fibrosis, and esophagitis were significantly reduced with amifostine. Complete response or partial response was observed in 75% of patients given amifostine and in 76% of patients treated with radiotherapy alone. However, a large proportion of the patients were not evaluated for tumor response because they were excluded because of either death or referral to chemotherapy. Most of the excluded patients were in the amifostine group. Amifostine was generally well tolerated, but 7% of patients given amifostine developed transient hypotension.

Preliminary data from a recent randomized phase 3 study with concomitant chemoradiation for non–small-cell lung cancer have shown that 500 mg amifostine IV twice a week significantly reduced acute pneumonitis and esophagitis. However, data on late radiation-induced toxicities and tumor control rates are still not available.7

Pelvic irradiation 

For pelvic radiotherapy, a retrospective analysis of 83 patients with cancer of the cervix did not show any radioprotective effect of amifostine 75 mg/m2 with regard to either tumor control or late normal tissue damage.81 This could possibly be because of the relatively low dose of amifostine administered.

One phase 3 trial with amifostine and radiotherapy for rectal cancer has been published. One hundred patients with inoperable or recurrent cancer of the rectum were treated with total pelvic irradiation (45 Gy/20 fractions) and additional boost (7.2 Gy/4 fractions or 14.4 Gy/8 fractions) depending on tumor response.82 Patients were randomized between irradiation alone or radiotherapy plus amifostine 340 mg/m2 before each fraction of total pelvic irradiation. No reduction of acute toxicity was observed with amifostine. The authors concluded that amifostine significantly reduced late effects without inducing tumor protection. However, there are several problems that may invalidate the conclusions. The radiation dose was not the same in all patients, and some patients were given chemotherapy in a nonrandomized manner. Many patients had tumor resection before or after radiotherapy. Finally, actuarial methods are imperative to estimate the true incidence of late radiation damage in a population with high mortality.

Alternative concepts in radioprotection 

Increasing evidence indicates that both acute and late normal tissue reactions after radiotherapy are closely related to a persistent over production of certain cytokines such as interleukin (IL)-1, Il-6, Il-10, tumor necrosis factor α, and transforming growth factor β 1.83 It is assumed that a cascade reaction exists that is initiated immediately after irradiation and that this cascade remains active months or years after radiotherapy.84, 85 Transforming growth factor β I in particular seems to play a crucial role in the development of radiation-induced fibrosis.86 Modulation of sustained oxidative stress or dysregulated cytokines in irradiated tissues may represent an alternative target for prevention of radiation damage.13 A limited number of agents including lisofylline, essential fatty acids, superoxide dismutase, pentoxyfilline, and alpha-tocopherol have been tested as antagonists of radiation-induced toxicity.13, 86 Available data are still very sparse but suggest that such substances may have a role in prevention and amelioration of normal tissue injury. Furthermore, some of these substances seem to have the ability to reverse established fibrosis, thus providing the basis for a therapeutic strategy.87 This concept is discussed in detail in the article by Moulder in this issue.88

Future perspectives 

Amifostine is without a doubt the most comprehensively tested radioprotective agent. Based on the available clinical data, it is well documented that clinically relevant radiprotection can be achieved in salivary glands with amifostine. From a biological point of view, this is probably a result of the extreme radiosensitivity of the serous acini combined with a relatively high uptake and retention of amifostine and its metabolites in the parotids.16, 24, 89 Thus, amifostine may very well represent a new strategy for reduction of radiotherapy related xerostomia. However, in this context, it is important to keep in mind that protection of salivary glands could also be achieved by using intensity modulated radiotherapy.4

Because of the uneven protection offered to different normal tissues, it seems unlikely that amifostine will allow for an escalation of radiation dose. For instance, the central nervous system is not protected by amifostine, and it seems quite uncertain to what extent amifostine protects against fibrosis and other dose-limiting late reactions.

Importantly, amifostine may have the potential to allow for intensified treatment by ameliorating mucosal reactions, often being a limiting factor in accelerated fractionation. However, data on the ability of amifostine to protect mucosal membranes are still quite sparse. Based on available clinical results, it would appear that the amifostine dose will need to be increased beyond the level presently recommended for radiotherapy. However, it is at present not clear if protection against mucositis can be obtained within a dose range that does not cause unacceptable cumulative toxicity when applied before each radiation fraction for several weeks. For instance, it is uncertain if sufficient doses of amifostine can be safely applied to hyperfractionated schedules with twice daily irradiation.44 It is possible that the subcutaneous route of administration may result in a more favourable toxicity profile with less hypotension.45, 46, 47 Additional studies are required to confirm this.

Because amifostine appears to be a broad-spectrum cytoprotective agent with the ability to protect tissues from damage induced by both irradiation and chemotherapy, it may enable intensified chemoradiation regimens. However, experience with amifostine in this setting is still very limited, and the optimal dosage and schedule of amifostine has not been established. Furthermore, as pointed out by Phillips and Tannock,90 it is important to be aware of the pitfalls related to trials evaluating the protective effect of amifostine with chemotherapy. Trials showing that more intensive treatment can be applied when combined with amifostine certainly show the biological effictiveness of amifostine. Nevertheless, such trials do not address the important pragmatic question, whether or not such intensified regimens actually improves therapeutic outcome compared with standard treatment.

The major concern related to radioprotectors remains the potential hazard of tumor protection. Preclinical data do not provide consistent evidence that amifostine selectively protects normal tissues without offering protection to tumors. To date, a number of clinical studies have reached the conclusion that amifostine does not cause any reduction of anti tumor efficacy. However, not even the trial conducted by Brizel et al,73 which recruited over 300 patients, has had sufficient statistical power to detect and quantify a possible tumor protective effect of amifostine. Furthermore, only about one third of patients enrolled in that trial received definite radiotherapy. Therefore, the next logical step in the exploration of amifostine is to reproduce the trial of Brizel et al73 in a more homogenous patient population with an accrual being sufficient to definitely rule out any adverse effect on anti tumor efficacy. Until such a trial has been conducted, we do not recommend that amifostine should be incorporated in routine radiotherapy practice.

Agents ameliorating radiation-induced damage by manipulating inflammatory and profibrotic cytokines represent an alternative strategy in radioprotection. Agents with such capabilities have so far only been tested to a limited extent. It also is possible that advances in molecular radiobiology will enable more precisely targeted interventions against radiation-induced damage.91, 92 However, depending on the agent, such toxicity antagonists may carry their own toxicities and manipulation of processes related to growth, differentiation, and repair may also potentially affect tumor response. All these aspects will need to be carefully evaluated.