Radium-223 mechanism of action: implications for use in treatment combinations

Abstract

The targeted alpha therapy radium-223 (²²³Ra) can prolong survival in men with castration-resistant prostate cancer (CRPC) who have symptomatic bone metastases and no known visceral metastases. Preclinical studies demonstrate that ²²³Ra preferentially incorporates into newly formed bone matrix within osteoblastic metastatic lesions. The emitted high-energy alpha particles induce DNA double-strand breaks that might be irreparable and lead to cell death in nearby exposed tumour cells, osteoblasts and osteoclasts. Consequently, tumour growth and abnormal bone formation are inhibited by these direct effects and by the disruption of positive-feedback loops between tumour cells and the bone microenvironment. ²²³Ra might also modulate immune responses within the bone. The clinical utility of ²²³Ra has encouraged the development of other anticancer targeted alpha therapies. A thorough understanding of the mechanism of action could inform the design of new combinatorial treatment strategies that might be more efficacious than monotherapy. On the basis of the current mechanistic knowledge and potential clinical benefits, combination therapies of ²²³Ra with microtubule-stabilizing cytotoxic drugs and agents targeting the androgen receptor axis, immune checkpoint receptors or DNA damage response proteins are being explored in patients with CRPC and metastatic bone disease.

Prostate cancer is the second most common cancer in men, with 1.3 million new diagnoses worldwide in 2018 (REF.¹), and is the fifth leading cause of cancer death globally². Age is a major risk factor for prostate cancer, and most diagnoses are made in men aged >65 years³. The incidence of prostate cancer has increased globally by 40% since 2006, owing to an increased number of diagnoses because of prostate-specific antigen (PSA) screening⁴. An increasingly ageing population also contributes to rising prostate cancer rates across the world⁴.

The distribution of metastatic prostate cancer is primarily to bone. Specifically, the bones of the axial skeleton, which contains the body’s bone marrow, are most commonly involved. Treatment of newly diagnosed metastatic disease involves androgen deprivation therapy (ADT), for example, gonadotropin-releasing hormone agonists and antagonists, given either alone or in combination with a next-generation agent, such as abiraterone, that targets the androgen receptor (AR) signalling axis, or the chemotherapeutic agent docetaxel^5–7. Despite an initial response to this treatment, the disease eventually progresses to metastatic castration-resistant prostate cancer (mCRPC). Various treatments improve survival in men with mCRPC, including cytotoxic therapies, AR-directed therapies, an immune checkpoint inhibitor for patients with microsatellite instability (MSI)-high or DNA mismatch repair-deficient tumours, targeted alpha-particle therapy (TAT) and cellular adoptive immunotherapy^8–20.

Radium-223 (²²³Ra), an alpha-particle-emitting radionuclide approved for patients with mCRPC and bone metastases by the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA)^21,22, prolongs overall survival (median 14.9 versus 11.3 months receiving placebo; HR 0.70, 95% CI 0.58–0.83; P < 0.001) in men with mCRPC¹⁸. ²²³Ra is incorporated into newly formed abnormal bone within metastatic lesions^23,24. Because alpha emitters efficiently and lethally break DNA, ²²³Ra might be less susceptible to the development of resistance mechanisms observed with other anticancer therapies, such as chemotherapy and pathway-directed therapy, against which multiple escape and resistance mechanisms exist^24–26. ²²³Ra is also under investigation in other cancers that metastasize to bone^23,27–37.

In this Review, we summarize the mode of action of ²²³Ra and discuss new preclinical studies aimed at elucidating the mechanism through which ²²³Ra exerts its biological effects. Our current understanding of this mechanism and those of other anticancer agents generates hypotheses for potential ²²³Ra combination therapies in patients with bone metastases, which are now being tested in clinical trials. We also discuss future perspectives for TAT to improve outcomes for patients with bone-predominant disease.

Clinical experience with ²²³Ra

Clinical studies.

In the ALSYMPCA (ALpharadin in SYMPtomatic Prostate CAncer) trial, patients with mCRPC and symptomatic bone metastases were treated with the best standard of care and either six injections of ²²³Ra at 55 kBq/kg every 4 weeks or placebo¹⁸. Patients in the ²²³Ra arm had longer overall survival (median 14.9 versus 11.3 months; HR 0.70, 95% CI 0.58–0.83; P < 0.001) and a longer time to first symptomatic skeletal event (SSE) (median 15.6 versus 9.8 months; HR 0.66, 95% CI 0.52–0.83; P < 0.001) than those in the placebo arm^18,38. The short-term and long-term safety profiles were favourable, with improvements in quality-of-life measures and low rates of myelosuppression^18,39,40. Subsequent analyses have not demonstrated that more doses or higher doses of ²²³Ra confer an additional benefit to patients with mCRPC⁴¹. The ALSYMPCA trial led to the approval of ²²³Ra for patients with mCRPC and bone metastases by the FDA in 2013 (REF.²¹) and the EMA in 2018 (REF.²²). Subsequently, in 2018, the EMA restricted the use of ²²³Ra to exclude combination with abiraterone acetate and prednisone or prednisolone owing to increased rates of death and skeletal fractures in the ERA-223 trial and recommended ²²³Ra for patients who have had at least two previous treatments and who cannot receive other treatments^42,43.

The National Comprehensive Cancer Network (NCCN) guidelines recommend treatment with ²²³Ra in patients with mCRPC, symptomatic bone metastases and no metastases in the viscera⁴⁴. The American Urological Association (AUA) guidelines suggest offering ²²³Ra to patients with mCRPC and symptomatic bone metastases, good performance status and no previous docetaxel chemotherapy or known visceral disease⁴⁵. The European Association of Urology (EAU) recommends ²²³Ra as one of the life-prolonging treatment options for patients with mCRPC and progression following docetaxel chemotherapy⁴⁶.

Long-term safety.

The use of any radiopharmaceutical engenders concerns about long-term potential toxicities, such as effects on marrow integrity and viability, as well as the potential inception of second malignancies. Long-term follow-up monitoring of men with mCRPC and symptomatic bone metastases in the ALSYMPCA trial evaluated the safety of ²²³Ra treatment for up to 3 years from the first injection⁴⁷. The 3-year follow-up analysis did not find an association between ²²³Ra treatment and occurrence of second malignancies, such as acute myelogenous leukaemia, myelodysplastic syndrome or new primary bone cancer. Non-treatment-related malignancies occurred at a low incidence at a similar rate in patients treated with ²²³Ra or placebo⁴⁷. ²²³Ra was associated with a low incidence of myelosuppression and low cumulative incidence rates for haematological and non-haematological adverse events⁴⁷.

REASSURE is an ongoing global, prospective, single-arm, observational study of patients with mCRPC and bone metastases who receive ²²³Ra in routine clinical practice⁴⁸. The objective of this study is to evaluate long-term safety and the occurrence of primary second malignancies during a 7-year follow-up period. At this time, the rate and degree of severity of treatment-emergent adverse events revealed by REASSURE⁴⁸ or follow-up data from ALSYMPCA⁴⁷ are not substantially different from those observed in the original ALSYMPCA trial¹⁸. The primary safety concern raised by ²²³Ra regards the use of the drug with novel AR-targeted drugs such as abiraterone for first-line mCRPC, which is associated with bone fractures.

Mode of action of ²²³Ra

Characteristics of bone metastases.

The health and structural integrity of normal bone is maintained by a balanced, continuous cycle of bone resorption by osteoclasts and bone formation by osteoblasts⁴⁹. Metastatic prostate cancer cells interact, via signalling pathways, with different cellular components of the multiple complex bone microenvironments (FIG. 1). Bone metastases are established through complex mechanisms involving interactions with osteoblasts, osteocytes, and osteoclasts, as well as bone marrow stem cells and haematopoietic cells, as demonstrated in preclinical models and clinical studies^49–51. Growth factors, such as bone morphogenetic proteins, transforming growth factor-β, insulin-like growth factor 1, vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF) and endothelin-1 (REF.⁵²), released by prostate cancer cells in the bone, disrupt the tightly regulated crosstalk between osteoblasts and osteoclasts and dysregulate their activity, leading to an abundance of new disorganized (woven) bone. In turn, osteoblasts express growth factors, including VEGF, monocyte chemoattractant protein 1, interleukin-6 and interleukin-8 (REF.⁵³), that promote prostate cancer cell growth and survival. These bidirectional positive-feedback loops among tumour cells, osteoblasts and osteoclasts and the effects on the bone matrix constitute a vicious cycle of osteoblastic bone metastasis, characteristic of prostate cancer^50,51 (FIG. 1). Prostate cancer cell proliferation within the bone is further promoted by osteoblast-mediated intratumoural steroidogenesis⁵⁴ and bone-marrow-induced activation of PDGF receptor-α through a ligand-independent mechanism⁵⁵, whereas physical contact between tum our cells and osteoblasts can disrupt osteoblast alignment and bone organization through heterotypic cell-cell adhesions and gap junctions⁵⁶. Dysregulated bone metabolism in terms of resorption and deposition leads to pathological bone morphology and physiology, which account for considerable morbidity, bone fracture, pain and ultimately death^57,58.

Fig. 1 |. ²²³Ra and prostate cancer bone metastases.

a | Vicious cycle of osteoblastic prostate cancer bone metastases. Growth factors released by prostate cancer cells in bone disrupt the tightly regulated osteoblast and osteoclast activity that is essential for normal bone integrity. Prostate cancer bone metastases mainly promote osteoblast activity, which stimulates osteoclast activity through the secretion of RANKL. In turn, osteoblasts express growth factors that promote prostate cancer cell growth and survival, resulting in a bidirectional positive-feedback loop between tumour cells and osteoblasts. In response to RANKL secreted by osteoblasts, activated osteoclasts release growth factors from the bone and further stimulate tumour growth, leading to the growth of new disorganized (woven) bone.

b | Mechanism of action of radium-223 (²²³Ra). ²²³Ra substitutes for calcium in hydroxyapatite complexes in woven bone surrounding prostate cancer metastatic lesions. ²²³Ra emits high-energy alpha particles over a range of ≤100μm. The alpha radiation leads to localized cytotoxicity through the induction of DNA double-strand breaks in adjacent tumour cells, osteoblasts and osteoclasts, which disrupts the positive-feedback loops between these cells, leading to inhibition of tumour growth and pathological bone reaction and a stabilization of the normal structure. Adapted from REF.²⁴ by permission from the American Association for Cancer Research: Suominen, M. I. et al. Radium-223 inhibits osseous prostate cancer growth by dual targeting of cancer cells and bone microenvironment in mouse models. Clin. Cancer Res. (2017) 23, 15, 4335–4346.

Evidence from preclinical models.

Animal models of metastatic prostate cancer to bone are notoriously difficult to develop owing to the complexity of the metastatic cascade, which requires recapitulation of each step of the metastatic development, tumour heterogeneity and immunological processes⁵⁹. Thus, preclinical models aimed at investigating the mechanism of action and treatment effects of ²²³Ra were developed and published in 2017 (REF.²⁴), well after the results of initial clinical trials of ²²³Ra became available, and even after the drug was approved by the FDA in 2013 following the pivotal randomized phase III ALSYMPCA trial^18,21. Preclinical studies with bone metastasis models of cancers other than prostate cancer demonstrated that ²²³Ra substituted for calcium in hydroxyapatite complexes within areas of osteoblast-mediated new bone formation at sites of metastatic lesions^23,25.

²²³Ra has a half-life of 11.4 days and decays to produce predominantly alpha particles that have a high linear energy transfer (LET) over a short range (100 μM)⁶⁰. This action leads to localized cytotoxicity, possibly through DNA double-strand breaks in adjacent cells^25,26,61 (FIG. 1). The mode of action of ²²³Ra was investigated in an intratibial murine xenograft model of osteoblastic bone metastasis using patient-derived LuCaP-58 cells, in which mice bearing LuCaP-58 tumours were randomized according to the lesion grade and/or serum PSA level. ²²³Ra or vehicle was administered every 4 weeks twice, and soft-tissue tumours and tibiae were analysed using radiographic and micro-CT imaging²⁴. Using this model, the vast majority of ²²³Ra deposits were observed within the surrounding bone matrix and especially in the vicinity of activated osteoblasts²⁴. The researchers identified a dual mode of action consisting of the induction of potentially cytotoxic DNA double-strand breaks in adjacent tumour cells leading to tumour cell death and also in osteoblasts and osteoclasts, which would be expected to further disrupt the positive-feedback loops between these cell types and to suppress tumour-induced bone pathogenesis. This disruption led to an inhibition of tumour growth and of the pathological bone reaction, resulting in stabilization of normal bone structure. These findings were confirmed in a second intratibial bone metastasis model using human LNCaP prostate cancer cell line xenografts²⁴.

In contrast to high-LET radiation, low-LET radiation, including beta and gamma radiation, induces a different pattern of DNA damage comprising single-strand breaks, base-pair deletions and substitutions, as well as DNA crosslinking²⁶. The type of DNA damage from exposure to low LET is more easily repaired than that to high LET, and the likelihood of a single lethal event is much lower. For example, 1,000 DNA hits from beta particles are required for cell death, whereas only 1–4 DNA hits are required from high-LET alpha particles^26,62. Furthermore, the mechanism of action of ²²³Ra is based on the direct non-reparable damage of tumour cell DNA, leading to tumour cell death and becoming less reliant on active cell proliferation, which might be advantageous for the killing of tumours with a low proliferation rate or dormant micrometastases²⁵. In addition, TAT might be less susceptible to the development of tumour resistance mechanisms, including hypoxia, as DNA damage resulting from TAT is largely independent of oxygen levels²⁶, in contrast to the oxygen-dependent DNA damage induced by external beam radiation therapy (EBRT), and use of alternate signalling pathways that become activated in response to other therapies, such as chemotherapy and cell signalling inhibitors^24–26. Compared with beta particle emitters and EBRT, TAT approaches might also have the advantage of minimizing the damage to normal surrounding tissue owing to their comparatively short track length, which corresponds to 3–6 cell diameters⁶³.

²²³Ra biodistribution and interaction with the bone and tumour compartments were investigated in prostate cancer bone metastases mouse models using LNCaP and PC3 cell lines⁶⁴. ²²³Ra was found to be deposited at the bone surface, but not in the tumour mass, and skeletal accumulation was dependent on the local blood vessel density, indicating that vascular access is important for the effective delivery of ²²³Ra (REF.⁶⁴).

²²³Ra might also enhance immune system activity, inhibit immunosuppression and alter the phenotype of cancer cells, rendering them susceptible to immune-mediated cell killing^65,66. In vitro studies using human prostate, breast, and lung cancer cell lines demonstrated that sublethal doses of ²²³Ra enhanced T cell mediated lysis of tumour cells through antigen-specific CD8⁺ cytotoxic T lymphocytes. This effect seemed to be mediated through the induction of the endoplasmic reticulum stress response pathway that involves PERK transmembrane protein mediating the unfolded protein response, which translates into elevated expression of multiple antigen-processing machinery components⁶⁷. ²²³Ra might also augment immune responses via the stimulator of interferon genes (STING) pathway^68,69. STING, an endoplasmic reticulum translocon-associated transmembrane protein, detects DNA secreted from pathogens or damaged host DNA (from apoptosis or necrosis) in the cell cytoplasm, leading to the initiation of innate immune defence mechanisms⁶⁹. The STING pathway is essential for radiation-induced adaptive immune responses, which are dependent on type I interferon signalling in dendritic cells⁷⁰. However, whether the STING pathway is linked to antitumour T cell responses following TAT is currently unknown.

²²³Ra dosimetry.

Evaluation and knowledge of TAT distribution, absorbed radiation doses and radiation exposure in tumour and normal tissues have the potential to optimize treatment plans and to maximize the antitumour effects of TAT, while sparing normal tissues⁷¹. However, the short range of alpha-particle emitters makes the dosimetric calculations in organs difficult when the activity is localized to specific tissue compartments or cell types⁷¹.

In vivo dosimetry relies on ²²³Ra uptake quantification. Specifically, ²²³Ra decay produces a series of six products, before becoming a stable molecule. Of the total emitted energy (28.2 MeV), 95% comes from alpha emissions, 3.2% from beta particles and 2% from gamma emissions. Gamma emissions enable quantitative imaging of ²²³Ra^72,73. However, microscopic energy deposition within the bone matrix presents challenges for internal dosimetry in the marrow and skeleton⁷². Currently, several small-scale modelling and microdosimetric approaches are being developed, with potential application in clinical imaging and dosimetry^71,74–76. For example, for ²²³Ra, the trabecular marrow cavity model calculates the fraction of marrow volume receiving a cytotoxic absorbed dose. These methods might enable improved estimates of the dose of radiation delivered by radium and other alpha emitters in the future.

Some findings demonstrate a correlation between absorbed dose and local lesion response to ²²³Ra. In nine patients with mCRPC and bone metastases, a correlation between the uptakes of ²²³Ra in a metastatic lesion and the scintigraphy agent technetium-99m-methylene diphosphonate was observed⁷³. Thus, conventional radiolabelled diphosphonate scintigraphy might help to define lesion extent⁷³.

Rationale for treatment combinations in mCRPC

Based on the current understanding of the mechanism of action of ²²³Ra, and its efficacy and safety profile in the clinic, several treatment combination strategies have been proposed (TABLE 1). Their goal is to increase treatment efficacy by using agents with different mechanisms of action together with ²²³Ra, while keeping toxic effects to a minimum.

Table 1 |.

Summary of selected ongoing ²²³Ra combination trials in mCRPC

Combination treatment with 223Ra	Mechanism of action	Trial phase and design	Primary outcomes	Patient characteristics (estimated enrolment)	Estimated completion date (primary)	Trial ID
Hormone therapies
Abiraterone149	CYP17 inhibitor	Phase III (vs abiraterone only)	Symptomatic skeletal event-free survival	CRPC with asymptomatic or mildly symptomatic BM (n = 806)	December 2019 (February 2018)	NCT02043678a
Enzalutamide93	AR inhibitor	Phase III (vs enzalutamide only)	Radiological progression-free survival	CRPC with asymptomatic or mildly symptomatic BM (n = 560)	April 2021 (November 2019)	NCT02194842
Enzalutamide150	AR inhibitor	Phase II	Safety (grade ≥3 AEs, by NCI CTCAE version 4)	Progressing mCRPC; BM and VD allowed (n = 44)	April 2021 (July 2017)	NCT02225704
Chemotherapy
Docetaxel109	Microtubule inhibitor	Randomized phase III (vs docetaxel only)	Overall survival	mCRPC (n = 738)	June 2023 (June 2022)	NCT03574571
Immunotherapy
Atezolizumab151	PD-L1 inhibitor	Phase I	DLT; adverse events; objective response (RECIST v1.1)	Progressing mCRPC; BM and VD allowed (n = 44)	March 2020	NCT02814669
Pembrolizumab121	PD-1 inhibitor	Randomized phase II (vs 223Ra only)	Extent of immune cell infiltration	mCRPC with BM (n = 45)	June 2024 (June 2020)	NCT03093428
Sipuleucel-T112	ACT using PA2024-pretreated cells to stimulate tumour-specific immune responses	Randomized phase II (vs sipuleucel-T only)	Immune responses measured by peripheral PA2024 T cell proliferation	CRPC with asymptomatic or minimally symptomatic BM (n = 34)	December 2020 (December 2018)	NCT02463799
DNA damage response
Niraparib131	PARP inhibitor	Phase Ib	Maximum tolerated dose of niraparib to combine with standard dose of 223Ra	mCRPC; BM and VD allowed (n = 27)	November 2020 (November 2021)	NCT03076203
Olaparib132	PARP inhibitor	Randomized phase I/II (vs 223Ra only)	Maximum tolerated dose of olaparib and 223Ra; radiographic progression-free survival	CRPC with BM (n = 138)	April 2020	NCT03317392

²²³Ra, radium-223; ACT, autologous cell therapy; AEs, adverse events; AR, androgen receptor; BM, bone metastases; CYP17, cytochrome P450 c17; DLT, dose-limiting toxicity; mCRPC, metastatic castration-resistant prostate cancer; PA2024, fusion protein of hum an prostatic antigen phosphatase and granulocyte-macrophage colony-stimulating factor; PARP, poly(ADP-ribose) polymerase; PD-1, programmed cell death 1; PD-L1, programmed cell death 1 ligand 1; VD, visceral disease.

^aUnblinded owing to safety concerns in the combination arm.

Androgen receptor-axis inhibitors.

AR signalling regulates prostate cancer growth, and ADT is a standard treatment for patients with advanced or recurrent disease^77,78. Despite initial responses to ADT, the disease invariably progresses to CRPC, which remains dependent on the AR-signalling axis in the majority of patients. Abiraterone acetate (abiraterone) is a selective steroidal inhibitor of cytochrome P450 c17 (CYP17), a key enzyme for testosterone and oestrogen biosynthesis^78,79 (FIG. 2). Inhibition of CYP17 blocks steroid production in the testes, adrenal glands and tumour, limiting the amount of hormones available to stimulate AR signalling⁷⁹. Enzalutamide, a non-steroidal antiandrogen, binds to the AR ligand-binding domain and inhibits the AR signalling pathway by preventing the binding of androgens to AR, by preventing nuclear translocation of activated AR and by impairing the binding of activated AR to DNA⁸⁰ (FIG. 2).

Fig. 2 |. Mechanism of action of anticancer agents in prostate cancer.

Several anticancer agents with differing mechanisms of action are in use in routine clinical practice or are under preclinical or clinical investigation in prostate cancer. Androgen receptor (AR)-axis inhibitors, microtubule-stabilizing agents, DNA damage response inhibitors, immune checkpoint inhibitors and immunostimulants, each with different but complementary mechanisms of action to radium-223 (²²³Ra), might have potential in combination treatments with ²²³Ra in metastatic castration-resistant prostate cancer. APC, antigen-presenting cell; ATR, ataxia telangiectasia and Rad3-related; CTLA, cytotoxic T lymphocyte antigen; DHT, dihydrotestosterone; PARP, poly(ADP-ribose) polymerase; PD-1, programmed cell death 1; PD-L1, programmed cell death 1 ligand 1; PSMA-TTC, prostate-specific membrane antigen-targeted thorium conjugate. *Approved by the FDA. ^§223Ra also inhibits cellular activity within the adjacent bone microenvironment through induction of DNA double-strand breaks in osteoblasts and osteoclasts²⁴.

Data from large randomized phase III clinical trials, COU-AA-301 and COU-AA-302 for abiraterone plus prednisone^12,15, and AFFIRM and PREVAIL for enzalutamide^10,81, led to the recommendation of these agents by the ESMO and NCCN guidelines for use in patients with mCRPC^44,82. In phase III trials in mCRPC, abiraterone was generally well tolerated and adverse effects included increased incidence of mineralocorticoid-related events, hepatotoxicity and cardiac disorders^12,15. The main adverse effects of enzalutamide included fatigue and hypertension^10,81. Subsequently, the combination of abiraterone plus prednisone or prednisolone with primary ADT was shown to significantly extend the median overall survival (not reached versus 34.7 months; HR for death 0.62, 95% CI 0.51–0.76; P < 0.001) in patients with hormone-sensitive metastatic prostate cancer^6,83, leading to guideline recommendation as a standard of care in the hormone-sensitive setting in suitable patients^44,84.

The differing mechanisms of action of ²²³Ra and abiraterone and enzalutamide and their non-overlapping safety profiles provide a rationale for combining ²²³Ra with abiraterone or enzalutamide in patients with metastatic prostate cancer. Preliminary data for these combinations were provided by post hoc analyses of early access programmes^85,86, patients treated in routine clinical practice⁸⁷ and from a phase II study⁸⁸, which suggested that ²²³Ra could be safely combined with abiraterone or enzalutamide.

The multicentre phase III ERA-223 trial evaluated the efficacy and safety of adding ²²³Ra treatment or placebo to abiraterone plus prednisone or prednisolone in patients with asymptomatic or minimally symptomatic mCRPC⁴³ (TABLE 1). Primary analysis revealed no statistically significant difference in overall survival between the groups (33.3 versus 30.7 months; HR 1.195, 95% CI 0.950–1.505; P = 0.13)⁴³. However, patients in the ²²³Ra combination arm had an unexpected increased incidence of fractures (29% versus 11%). Consequently, the combination of ²²³Ra with abiraterone and prednisolone or prednisone is currently contraindicated by the EMA^42,89.

Several reasons for the increase in skeletal fractures in the combination therapy group compared with the control group have been suggested^90–92. Treatment with low-dose prednisone monotherapy has previously been shown to have a negative effect on bone metabolism through the inhibition of osteoblast function and the promotion of osteoclast activity, leading to decreased bone formation⁹⁰. In addition to the potentially harmful effects of glucocorticoids and the potential hormonal changes within the bone microenvironment as a consequence of abiraterone, ionizing radiation can also contribute by augmenting osteoclastogenesis^91,92. Overall, the additive effects of abiraterone and prednisone and continuous androgen depletion, glucocorticoids and ²²³Ra might result in compromised bone health, and, in turn, might lead to more frequent skeletal fractures.

In the ERA-223 trial, concurrent use of bone health agents (BHAs), including bisphosphonates (zoledronic acid) and denosumab from baseline, was associated with a lower incidence of fractures in both the ²²³Ra and placebo groups (15% and 7%, respectively) compared with patients who did not receive BHAs (37% and 15%, respectively)⁴³. The hazard ratio for SSE-free survival for the ²²³Ra and placebo group was 0.932 (95% CI 0.666–1.306) when BHAs were used and 1.252 (95% CI 0.971–1.615) without BHAs⁴³. These findings are underscored by data from the ALSYMPCA trial, in which patients treated with ²²³Ra in combination with BHAs seemed to have a superior outcome with respect to risk of SSEs compared with those who received ²²³Ra without BHAs^18,38. In a univariate analysis of time to first SSE, current use of bisphosphonates at baseline (yes versus no) was an independent predictor of decreased risk of SSEs (HR 0.56, 95% CI 0.44–0.71). In multivariate analysis, the treatment effect of ²²³Ra compared with placebo was maintained (HR 0.65, 95% CI 0.51–0.82) and concurrent use of bisphosphonates at study entry was also independently associated with reduced risk of SSEs (HR 0.49, 95% CI 0.38–0.64)³⁸. Notably, the treatment effect of ²²³Ra versus placebo on the time to first SSE seemed to be greater in patients who received current bisphosphonates at baseline (median 19.6 versus 10.2 months; HR 0.49, 95% CI 0.33–0.74) than those who did not receive bisphosphonates (median 11.8 versus 8.4 months; HR 0.77, 95% CI 0.58–1.02). These data suggest that bisphosphonates should be given with ²²³Ra to maximize treatment benefit, prevent fragility fractures and preserve bone health³⁸.

Whether the increase in fractures will be seen in patients receiving a combination of ²²³Ra and AR-axis inhibitors other than abiraterone remains to be determined. The randomized phase III trial PEACE 3 (REF.⁹³) is currently investigating ²²³Ra in combination with enzalutamide in patients with mCRPC who are asymptomatic or mildly symptomatic (TABLE 1).

Chemotherapy.

Taxanes are antimitotic chemotherapy agents that bind to and abnormally stabilize microtubules in dividing cells^94,95 (FIG. 2). This effect impairs intracellular organelle trafficking, might impair AR signalling and leads to cell cycle arrest and apoptosis^94–98.

Docetaxel was first recommended by the NCCN and ESMO guidelines for patients with mCRPC (particularly those who are symptomatic or have a high disease burden), and cabazitaxel was recommended in patients whose disease progresses on docetaxel treatment^44,82. Results of two phase III trials^5,7 support the recommendation for combining chemotherapy with primary ADT in newly diagnosed patients with metastatic castration-naive disease. The STAMPEDE study showed improved survival in patients treated with long-term hormone therapy combined with docetaxel chemotherapy compared with long-term hormone therapy alone (71 months versus 81 months; HR 0.78, 0.66–0.93; P = 0.006)⁵. The CHAARTED study also demonstrated an advantage of the combination therapy (ADT therapy plus docetaxel) in extending overall survival in patients with metastatic hormone-sensitive prostate cancer (58 months versus 44 months; HR for death 0.61, 95% CI 0.47–0.80; P < 0.001)⁷.

Several factors support the combination of ²²³Ra with docetaxel. First, chemotherapy has been shown to have a radiosensitizing effect through different mechanisms, including elevated radiation-induced damage, reduced DNA repair, increased vulnerability of hypoxic cells to cytotoxic agents, inhibition of prosurvival pathways and reduced ability of tumour cells to repopulate during radiotherapy, which lead to an additive cytotoxic effect when combined with radiotherapy in cancer patients^99–101. However, no data currently exist on radiosensitizing effects when combining chemotherapy with alpha-particle emitters. In addition, the anti-proliferative effects of chemotherapy might reduce the cell repopulation effect associated with radiation therapy, in which rapid regrowth of tum our cells commonly occurs between doses of radiation⁹⁹.

Additional rationale for combining bone-targeted alpha therapy with chemotherapy is provided by the concept of simultaneous targeting of the tum our and the bony compartment of the disease, each agent independently prolonging survival, and by the possibility that each agent might enhance the other by cross-sensitization¹⁰². The potential benefits of combining radiopharmaceuticals with chemotherapy in patients with mCRPC have previously been demonstrated in clinical studies of chemotherapy in combination with palliative beta particle-emitting agents^103–105. For ²²³Ra, exploratory efficacy analysis of a randomized phase I/II clinical trial showed that ²²³Ra (55 kBq/kg every 6 weeks, n = 36) plus docetaxel (60 mg/m² every 3 weeks) might enhance antitumour activity compared with docetaxel alone (75 mg/m² every 3 weeks; n = 17)¹⁰⁶. The combination arm had more durable suppression of PSA (median time to progression 6.6 versus 4.8 months, P = 0.02), alkaline phosphatase (ALP) (9 versus 7 months, P = 0.44) and osteoblastic bone deposition markers (bone ALP and procollagen type I N propeptide)^106–108. Combination therapy was well tolerated and presented no greater safety concerns than treatment with docetaxel alone¹⁰⁶. The most common adverse event was neutropenia, which occurred less frequently in the combination arm (30%) than in the treatment arm with docetaxel alone (38%), and no clinically significant thrombocytopenia was reported. Grade 3/4 treatment-emergent adverse events occurred in 48% and 62% of patients treated with the combination therapy or docetaxel alone, respectively. On the basis of these encouraging results, a phase III trial of this treatment combination is currently recruiting patients¹⁰⁹ (TABLE 1).

Immunotherapy.

Combining ²²³Ra with immunotherapies, particularly those designed to generate and expand endogenous tumour-antigen-specific T cell populations, might be a beneficial approach, particularly when application of a lethal dose of radiation needed to reduce the tum our burden is constrained by toxic effects (FIG. 2). In a preclinical in vitro study, sublethal doses of ²²³Ra have been shown to sensitize the human prostate, breast and lung carcinoma cells to T cell-mediated lysis by CD8⁺ cytotoxic T lymphocytes specific for tumour antigens mucin-1, brachyury and carcinoembryonic antigen¹¹⁰. Sipuleucel-T is an autologous cell therapy, in which patient-enriched monocytic cells are activated with a recombinant fusion protein of hum an prostatic acid phosphatase and granulocyte-macrophage colony-stimulating factor (termed PA2024)¹¹¹. Sipuleucel-T is currently approved by the FDA for the treatment of patients with mCRPC who have few or no symptoms, following data from a phase III clinical trial in which sipuleucel-T compared with placebo was shown to prolong overall survival in patients with mCRPC^19,44. The treatment is generally well tolerated: the most common adverse events are fevers, chills, fatigue, nausea and headache¹⁹. A current randomized phase II study is investigating the combination of ²²³Ra with sipuleucel-T in patients with mCRPC and asymptomatic or mildly symptomatic bone metastases and is due to be completed in November 2019 (REF.¹¹²) (TABLE 1).

Therapeutic intervention with agents targeting programmed cell death 1 (PD-1) or the programmed cell death 1 ligand 1 (PD-L1) seem to be effective in a range of cancer types^113–115. PD-L1 is selectively expressed on many cancer cells and on cells within the tumour microenvironment in response to inflammatory stimuli¹¹³. PD-L1 and PD-1 negatively regulate immune activation, mainly through the inhibition of T cell function. Atezolizumab, a monoclonal PD-L1 antibody, binds to PD-L1 on tumour cells and prevents its interaction with PD-1 and B7.1 receptors on immunosuppressed T cells, resulting in an activation of T cells and tumour cell death through specific immune responses¹¹⁶.

The FDA has approved the PD-1 inhibitory antibody pembrolizumab for MSI-high and DNA mismatch repair-deficient cancers, including mCRPC¹¹⁷. Early analyses showed that responses to PD-1 or PD-L1 therapy might occur in ~50% of patients with defects in DNA mismatch repair genes¹¹⁸. Studies in men with mCRPC not selected for MSI-high disease or DNA mismatch repair deficiencies are ongoing¹¹⁹.

Preclinical studies suggest that the PD-L1-PD-1 axis might dampen the immune response to ionizing radiation¹²⁰. In a mouse model, PD-L1 in the tumour microenvironment was upregulated following treatment with ionizing radiation, leading to the suppression of radiation-induced immune responses¹²⁰. Concurrent administration of inhibitory PD-L1 antibodies enhanced the efficacy of the ionizing radiation through a mechanism that was dependent on cytotoxic T cells. Clinical trials are investigating ²²³Ra in combination with the PD-L1 inhibitory antibody atezolizumab in patients with mCRPC who have progressed on treatment with an androgen pathway inhibitor, and ²²³Ra in combination with pembrolizumab in patients with mCRPC and bone metastases¹²¹ (TABLE 1).

DNA damage response.

The DNA damage response maintains cellular integrity and homeostasis. Cellular recognition of DNA damage activates DNA damage response pathways, resulting in cell cycle arrest and induction of DNA repair or cell death¹²². Genomic defects in DNA repair have been identified in 12–30% of men with advanced CRPC^123–125. In an integrative clinical sequencing analysis of 150 patients with mCRPC, 23% of tumours had aberrations in DNA repair pathways and 8% of patients had germline variants¹²⁴. The incidence of germline mutations in DNA repair genes associated with autosomal dominant cancer predisposition syndromes in a cohort of 692 patients with metastatic prostate cancer was 12%¹²⁵.

Some of these genetic alterations are associated with sensitivity to DNA-damaging anticancer agents, including platinum-based chemotherapy, and sensitivity to inhibitors of DNA damage response proteins, including poly(ADP-ribose) polymerase (PARP)¹²⁶ (FIG. 2). This finding suggests that using PARP inhibitors in patients whose tumours harbour a DNA repair defect (mutation in DNA repair genes) might exploit a synthetic lethal interaction¹²⁷.

Olaparib, a PARP inhibitor, is used in the treatment of patients with advanced ovarian cancer who have germline BRCA1 or BRCA2 gene m utations¹²⁸. In a phase II trial in patients heavily pretreated with mCRPC, olaparib treatment resulted in a response rate of 33%¹²⁷. Next-generation sequencing identified homozygous deletions and deleterious mutations in DNA repair genes in 33% of patients, and 88% of these patients had a response to olaparib. Deleterious BRCA2 mutations or other deleterious alterations and ATM gene mutations were found in the majority of patients showing a response to olaparib¹²⁷. The FDA granted breakthrough therapy designation for olaparib monotherapy for the treatment of BRCA1 or BRCA2 or ATM-mutated mCRPC in patients who have received prior taxane-based chemotherapy and at least one next-generation antihormonal agent (abiraterone or enzalutamide)¹²⁹. A small study in patients with mCRPC treated with ²²³Ra (n = 49) found that 4 of 10 (40%) responders who had undergone DNA sequencing had an alteration to a DNA repair gene¹³⁰. As ²²³Ra induces double-strand DNA breaks, defects in DNA damage repair pathways might confer sensitivity to ²²³Ra, but this hypothesis requires confirmation in a prospective study¹³⁰. Combination of ²²³Ra with PARP inhibitors might be useful, as a PARP inhibitor might further inhibit double-strand break repair capacity in ²²³Ra-irradiated cells, even in those patients whose tumours do not harbour alterations in DNA repair genes. An ongoing phase Ib study is investigating the PARP inhibitor niraparib in combination with ²²³Ra in patients with mCRPC who do not have alterations in DNA repair genes¹³¹, and a randomized phase I/II trial is planned to investigate the combination of olaparib and ²²³Ra (REF.¹³²) (TABLE 1).

The combination of ²²³Ra with other inhibitors of DNA damage response proteins might also be considered. BAY 1895344, an inhibitor of the ataxia telangiectasia and Rad3-related (ATR) kinase (FIG. 2), which is involved in the repair of DNA double-strand breaks, is currently in phase I clinical testing¹³³. A preclinical study demonstrated synergistic antitumour activity in a mouse xenograft model of mCRPC with bone metastases treated with BAY 1895344 in combination with ²²³Ra (REF.¹³⁴). BAY 1895344 doses ranging from 3% to 13% of the single-agent maximum tolerated dose in the combination treatment with ²²³Ra showed antitumour activity, with the highest activity achieved 24 h after administration of the first ²²³Ra dose.

A study published in 2018 evaluated whether pathogenic mutations in homologous recombination DNA repair genes correlate with ²²³Ra efficacy in patients with mCRPC and bone metastases¹³⁵. The study found that patients with homologous recombination deficiency (HRD) who were treated with ²²³Ra showed improved ALP responses (80% versus 39%; P = 0.04) and time to ALP progression (median 10.4 versus 5.8 months, HR 6.4; P = 0.005), and longer overall survival (median 36.9 versus 19.0 months, HR 3.3; P = 0.11) compared with patients without the deficiency¹³⁵. Unfortunately, the patient population was small (n = 28)¹³⁵ and the correlation between the HRD status and response to ²²³Ra needs further validation.

Future perspectives on alpha therapy

²²³Ra specifically targets hydroxyapatite and bone metastases in men with prostate cancer, yielding a clinical benefit in prolonging overall survival and delaying time to first SSE. However, by using a bone-targeting agent such as ²²³Ra, soft-tissue disease is not targeted. Further, because the depth of penetration of an alpha particle is so shallow, disease that is more than a few cell lengths away from newly deposited bone will not be directly affected by the treatment. Future approaches will be focused on delivering alpha-emitting therapy, not so much to newly deposited bone but to the cancer cell itself. Such an approach has the potential to directly kill more tumour tissue, regardless of location. The conjugation of alpha particle-emitting radionuclides with tumour-targeted carrier molecules, including antibodies, peptides and small molecules, has the potential to deliver tumour-specific targeting¹³⁶.

Antibodies targeted against tumour antigens and conjugated to the alpha emitter thorium-227 (²²⁷Th) have been evaluated in preclinical studies and include ²²⁷Th-CD-70 conjugates in renal cell carcinoma models¹³⁷; ²²⁷Th-lintuzumab (CD33 antibody) in acute myeloid leukaemia xenografts¹³⁸; ²²⁷Th-rituximab (CD20 antibody) in lymphoma xenografts¹³⁹; and ²²⁷Th-trastuzumab (HER2 antibody) in breast and ovarian cancer xenografts¹⁴⁰. In addition, a phase I clinical trial of ²²⁷Th-epratuzumab (CD22 antibody) in non-Hodgkin lymphoma is in progress¹⁴¹. Prostate-specific membrane antigen (PSMA) is expressed on both normal and cancerous prostate tissues and PSMA targeting is successfully used in imaging techniques¹⁴². A PSMA-targeted thorium conjugate (²²⁷Th-labeled anti-PSMA IgG1 (PSMA-TTC)) showed statistically significant prevention of tumour growth at a dose of 100 kBq/kg in an orthotopic bone xenograft model (LNCaP-Luc), which supports its clinical development for the treatment of patients with metastatic prostate cancer^143,144. Importantly, applying the principle underlying PSMA-TTC could extend the use of ²²³Ra to patients with mCRPC and visceral disease. Actinium-225 (²²⁵Ac) is also under clinical investigation in mCRPC. In an analysis of patients who received an ²²⁵Ac radiolabelled form of the PSMA-targeting small molecule PSMA-617, a dose of 100 kBq/kg per cycle every 8 weeks was well tolerated¹⁴⁵ and demonstrated a median duration of tumour control of 9 months in 40 patients¹⁴⁶. Similarly, a bismuth-213 (²¹³Bi)-labelled small-molecule PSMA inhibitor (²¹³Bi-PSMA I&T) rapidly targeted PSMA-expressing LNCaP tumour xenografts leading to accumulation of DNA double-strand breaks in the tumour tissue¹⁴⁷. In a patient with mCRPC that progressed under conventional therapy, two cycles of treatment with another ²¹³Bi-labelled small-molecule PSMA inhibitor (²¹³Bi-PSMA-617) achieved a remarkable molecular imaging response and reduction of PSA levels¹⁴⁸.

Conclusions

²²³Ra seems to act through multiple mechanisms, including direct effects via DNA damage to tumour and effector cells of pathological bone remodelling that are in close proximity to newly formed, hydroxyapatite-rich bone matrix, and possibly also through activation of cellular immune defences within the bone microenvironment via the endoplasmic reticulum stress response and STING pathways. Understanding the mechanisms of action of ²²³Ra should be useful for designing combination strategies that include agents that target tumour growth by inhibiting DNA repair, modulating the immune system or directly targeting cell proliferation. Clinical trials to investigate novel ²²³Ra combinations are underway (TABLE 1).

Mechanistically, combining ²²³Ra with other anticancer agents seems feasible and data from preclinical studies are encouraging; however, the results of randomized controlled trials are required before ²²³Ra combinations can be used in clinical practice, as emphasized by the unexpected findings of the ERA-223 trial, despite a strong rationale for combining ²²³Ra with abiraterone in mCRPC.

Our understanding of how ²²³Ra improves patient outcomes might also pave the way for new developments in TAT. Combining the cytotoxic properties of alpha-particle emitters with agents that enable tumour-specific delivery has the potential to expand the spectrum of cancers in which TAT might prove to be an effective treatment option.

Key points.

• Radium-223 (²²³Ra) deposited in the intralesional bone matrix emits high-energy alpha-particles that induce potentially cytotoxic DNA double-strand breaks in cells within a 100-μm distance while sparing surrounding normal tissue.

• ²²³Ra acts via a multi-modal mechanism, killing both tumour cells and the effector cells of pathological bone metabolism, osteoblasts and osteoclasts; ²²³Ra might also promote local antitumour immune responses.

• Understanding the mode of action of ²²³Ra and its interactions with those of other anticancer agents offers potential for new treatment combinations for bone-metastatic cancers.

• Clinical trials are investigating ²²³Ra in combination with agents targeting the androgen receptor signalling axis, cell microtubules, the immune response and the DNA damage response.

• Current combination studies are preliminary and further clinical testing will be required to demonstrate safety, efficacy and clinical benefit.

• ²²³Ra treatment in patients with metastatic castration-resistant prostate cancer has led to the development of other targeted alpha therapies for the treatment of patients with different cancers that also have bone-predominant metastases.