Dosimetric impact of Ac-227 in accelerator-produced Ac-225 for alpha-emitter radiopharmaceutical therapy of patients with hematological malignancies: a pharmacokinetic modeling analysis

Background Actinium-225 is an alpha-particle emitter under investigation for use in radiopharmaceutical therapy. To address limited supply, accelerator-produced 225Ac has been recently made available. Accelerator-produced 225Ac via 232Th irradiation (denoted 225/7Ac) contains a low percentage (0.1–0.3%) of 227Ac (21.77-year half-life) activity at end of bombardment. Using pharmacokinetic modeling, we have examined the dosimetric impact of 227Ac on the use of accelerator-produced 225Ac for radiopharmaceutical therapy. We examine the contribution of 227Ac and its daughters to tissue absorbed doses. The dosimetric analysis was performed for antibody-conjugated 225/7Ac administered intravenously to treat patients with hematological cancers. Published pharmacokinetic models are used to obtain the distribution of 225/7Ac-labeled antibody and also the distribution of either free or antibody-conjugated 227Th. Results Based on our modeling, the tissue specific absorbed dose from 227Ac would be negligible in the context of therapy, less than 0.02 mGy/MBq for the top 6 highest absorbed tissues and less than 0.007 mGy/MBq for all other tissues. Compared to that from 225Ac, the absorbed dose from 227Ac makes up a very small component (less than 0.04%) of the total absorbed dose delivered to the 6 highest dose tissues: red marrow, spleen, endosteal cells, liver, lungs and kidneys when accelerator produced 225/7Ac-conjugated anti-CD33 antibody is used to treat leukemia patients. For all tissues, the dominant contributor to the absorbed dose arising from the 227Ac is 227Th, the first daughter of 227Ac which has the potential to deliver absorbed dose both while it is antibody-bound and while it is free. CONCLUSIONS: These results suggest that the absorbed dose arising from 227Ac to normal organs would be negligible for an 225/7Ac-labeled antibody that targets hematological cancer.


Introduction
Alpha-particle emitter radiopharmaceutical therapy (αRPT) is a promising new approach to cancer therapy. It has been found impervious to conventional resistance mechanism that make traditional therapy ineffective [1]. Encouraging results have been observed in early clinical studies utilizing 225 Ac to deliver alpha-particles for both hematologic and solid tumor treatment, including programs targeting CD33 in acute myeloid leukemia and PSMA in castrate-resistant prostate cancer [2,3]. However, current limitation of available 225 Ac supply, due to the fixed output from 229 Th generator, has been a concern that has impacted preclinical and clinical use of 225 Ac-based αRPT [4]. Accordingly, a number of alternative production methods have been examined as potential sources of large and sustainable quantities of 225 Ac [5][6][7]. Accelerator-produced 225 Ac via 232 Th irradiation (hereafter denoted as 225/7 Ac) contains 0.1 to 0.3% 227 Ac (21.77-year half-life) activity at end of bombardment [8]. To account for the time elapsed for processing, transport and injectate preparation, we consider a scenario where the injected, 225/7 Ac radiolabeled conjugate contains 0.7% 227 Ac. Actinium-227 decays by beta-particle emission primarily (99%) to thorium-227 ( 227 Th; 18.68-d half-life) which in turn decays to radium-223 ( 223 Ra, 11.43-d half-life) and a series of other alpha-and beta-emitting daughters to stable lead-207 ( 207 Pb) (Fig. 1).
Using pharmacokinetic modeling, this work examines the contribution of 227 Ac and its daughters to tissue absorbed doses when 225/7 Ac-labeled antibody is administered intravenously to treat patients with hematological cancers (e.g., acute myeloid leukemia (AML) and/or myelodysplastic syndrome (MDS)).

Overview
Published pharmacokinetic models are used to obtain the distribution of 225/7 Ac-labeled antibody and also the distribution of either free or antibody-conjugated 227 Th. Since 227 Th is obtained from the beta decay branch (99% yield) of 227 Ac rather than a more energetically disruptive alpha-emitter decay, as has been observed with the 212 Pb/ 212 Bi delivery for α-emitter radiopharmaceutical therapy, [9][10][11][12], it is likely that a significant fraction of the 227 Th generated remains antibody-conjugated. A pharmacokinetic model representing the distribution of radiolabeled antibody in patients with hematologically distributed cancer is adapted from reference [13] to obtain the pharmacokinetics for 225/7 Ac and 227 Th-labeled antibody. A model representing the pharmacokinetics of free 227 Th is used to model the distribution of unconjugated 227 Th [14]. Under both circumstances, 223 Ra generated by 227 Th decay is simulated using a pharmacokinetic model that is relevant to free 223 Ra [15]. The 1% of 227 Ac that decays to francium-223 ( 223 Fr, T ½ = 22 min) is considered to have a negligible impact on tissue absorbed dose relative to that from 227 Th which is already expected to be very low because of the low initial amount of 227 Ac in 225/7 Ac. Calculations were performed assuming 1 kg (10 12 antigenpositive cells) in an adult female. The individual model simulations (i.e., Ab model, and the free 227 Th and 223 Ra models) are not coupled to each other. Rather, the biodistribution of free 227 Th or 223 Ra generated in the course of the simulation is assumed distributed throughout the body as it is created according to the kinetics described by the corresponding model (see Eqs. 17 and 18 of the "Appendix"). Table 1 summarizes the various models that were used to simulate the pharmacokinetics (PK) of each radionuclide. All compartmental models were solved using the simulation analysis and modeling software package (SAAM II, The Epsilon Group, Charlottesville, VA). Detailed model equations are listed in the "Appendix".

Antibody pharmacokinetic modeling
The radiolabeled antibody model is depicted in Fig. 2. This model is used to derive the kinetics of antibody-bound radionuclides. Radiolabeled antibody (Ab) is administered Table 1 PK models used for each radionuclide

Fig. 2
Pharmacokinetic model for radiolabeled antibody. The dotted line corresponds to the distribution volume of IV-administered antibody; ECF = extracellular fluid volume. Compartments 1 and 2 represent the free, (Ab) and antigen-bound antibody (AbAg) states, respectively. Compartment 3 represents internalized AbAg. The figure is adapted from reference [13] in the vascular space (compartment 1). It binds to antigen sites via saturable (non-linear) binding represented by the rate parameter k(2,1), which is a function of the affinity constant and the number of free antigen sites available (see equations in the "Appendix"). Antigen-bound antibody (AbAg) in compartment 2 may dissociate to return to the free radiolabeled antibody state by a rate constant, k(1,2) that is equal to the dissociation rate of antigen-bound antibody; AbAg may also internalize via k(3,2) into an intracellular compartment (compartment 3) where it is no longer available for dissociation but is cleared via catabolism at a rate represented by k(0,3). This lumped parameter model neglects aspects related to spatial gradients and transport across vasculature and is, therefore, specific to a radiopharmaceutical therapy of rapidly accessible antigen-positive cells within a vascular space that includes the plasma volume and the extracellular fluid (ECF) volume of the liver, spleen and red marrow (represented by the dotted box).
The essential components of this model have been previously validated using patient data [13].

Thorium biokinetic modeling
Time-activity curves for the fraction of 227 Th that is not antibody-bound following the decay of 227 Ac are given by the biokinetic model shown in Fig. 3. This model was developed and has been validated by Committee 2 of the International Commission on Radiological Protection (ICRP) [14,16].

Radium biokinetic modeling
The biokinetic model for free radium is depicted in Fig. 4. It is based on an ICRP model describing the behavior of alkaline earth elements ("Appendix" of reference [17]), as implemented by Lassmann et al. to calculate normal tissue dosimetry for 223 RaCl 2 [15].

Time-integrated activity coefficients (TIACs)
The time-integrated activity (TIA), for each source region, r i , ( Ã (r i ) ) was obtained by integrating model-derived pharmacokinetic data. The TIAC is given by dividing TIA by the administered activity of 225 Ac or by expressing the pharmacokinetics as a fraction of the administered activity. Equations 4-7 in the "Appendix" were integrated numerically with the substitutions indicated in equations: 8-10; 11-13, and 14-16, to get TIAC for 225 Ac, 227 Ac, and the antibody-bound fraction of 227 Th, respectively. The TIAC for free 227 Th and 223 Ra was obtained by numerically integrating Eqs. 17 and 18. Numerical integration was performed using the trapezoidal method.

TIAC apportionment
Model-derived TIAC was apportioned to tissue parenchyma as specified by the pharmacokinetic models. TIAC calculated for blood (central compartment) was apportioned to all tissue according to their blood volume [18]. The daughters of 225 Ac up to 213 Bi, respectively, were assumed to decay at the site of 225 Ac decay. Likewise the daughters of 213 Bi were assumed to decay at the site of 213 Bi decay. The TIAC, in each case was adjusted by the net yield of each daughter relative to the corresponding parent. The same approach was taken for the daughters of 223 Ra.

Absorbed dose calculations
Absorbed dose calculations were performed using the MIRD Committee S-value based method as described in pamphlet 21 [19]. The International Commission on Radiological Protection (ICRP) recently released absorbed fractions for a new series of phantoms that include far more tissues than were previously available [20]. The new absorbed fractions handle electron emissions far better than prior absorbed fractions which assumed that all or none of the energy associated with electron emissions was absorbed in tissues; absorption of alpha-particle energy is also appropriately considered [18]. A detailed comparison of the results obtained using OLINDA [21] and the new set of ICRP data has been published [22]. The calculations were performed using, newly developed software package, 3D-RD-S (Radiopharmaceutical Imaging and Dosimetry, LLC (Rapid), Baltimore MD), designed to account for the complexity of alpha-particle emitter dosimetry, in particular the differential fate of alpha-particle emitting daughters [23]. Absorbed doses from alpha-particles would ordinarily be multiplied by an RBE value of 5 [24,25]. We have chosen not to use this factor and rather report the absorbed dose for each emission type directly. This approach provides all the information needed to apply an RBE value to the alpha-component of the absorbed dose.

Radionuclide decay scheme data
Decay schemes and half-lives for 225 Ac and 227 Ac and their daughters were obtained from ICRP publication 107 [26]. Table 2 lists the parameter values for the antibody PK model, and the free 227 Th model and the 223 Ra model parameters are available in the publications related to the models that are cited above.

Ab-Ag pharmacokinetic model
The time-activity data obtained from Ab PK model simulations are plotted in Fig. 5A.
The PK for 227 Ac-bound Ab is identical to that shown in Fig. 5, except that all data are scaled by 0.07% (= f Ac227 , Table 2).
Since 99% of 227 Ac decays by beta-particle emission, which is less energetically disruptive than alpha-particle decay, the assumption is made that 70% (= f Ab ) of the daughter radionuclide, 227 Th, remains antibody-bound and obeys PK that is identical to that shown in Fig. 5A, except that all data are scaled by f Ac227 · f Ab . The remaining 30% is assumed to obey the pharmacokinetics of free 227 Th (Fig. 5B); this value was chosen as it is consistent with the retention of 212 Bi following decay of 212 Pb, also a beta-decay transition [27]. Since 227 Th decays by alpha-particle emission to 223 Ra, all of the 223 Ra generated, regardless of whether the 227 Th was Ab-bound or free is assumed to follow free radium kinetics (Fig. 5C).
As indicated in the text, Fig. 5A is scaled relative to an arbitrary amount of administered 225/7 Ac that is antibody-bound. In other words 1 MBq of 225/7 Ac-Ab administered should be multiplied by the fraction of injected activity (FIA) values indicated on the y-axis to obtain the corresponding amount of 227 Ac or 227 Th activity. The PK data plotted on Fig. 5B, C should be multiplied by f Ac227 · (1 − f Ab ) and f Ac227 , respectively, to convert the results to per MBq of 22/75 Ac administered.
The resulting time-activity curves for each "species" were numerically integrated to obtain the TIAC for each of the indicated tissues (Table 3). Table 4 lists absorbed doses for 225 Ac and 227 Ac; the absorbed dose from each particle type is provided separately. Table 5 lists the absorbed dose to selected tissues from 225 Ac, 227 Ac and their respective daughters. (Contributions from the 1% decay of 227 Ac to 223 Fr and daughters with a yield of less than 10 -4 % are not included.)

Absorbed doses
The absorbed dose from 225 Ac and its daughters, along with the 227 Ac to 225 Ac absorbed dose ratio for the top 6 tissues by total absorbed dose, is depicted in Fig. 6.

Discussion
The alpha emitter 225 Ac is a promising radionuclide for the generation of potent radiopharmaceutical agents for hematologic and solid tumor malignancies. However, commercial-scale supply concerns regarding purified 225 Ac have limited more widespread clinical research and development of 225 Ac-based αRPT. As a result, a number of alternative production methods have been examined as potential sources of large and scalable quantities of 225 Ac, including accelerator-produced 225 Ac via 232 Th irradiation. However, accelerator-produced 225 Ac contains 227 Ac as an impurity in the purified material. We undertook this work to investigate the dosimetric impact of the k + (M −1 h −1 ) 0.5 [13] k − (h −1 ) 0.003 [13] T c (h) 40 [13] T i (h) 0.5 Estimated f L1 , f S1 0.18, 0.12 [13] f L2 , f S2 0.08, 0.06 [13] V RMECF , V d (L) 0.22, 3.8 [13] f Ac227 0.007 Based on EOB fraction and assuming two 225 Ac half-lives until time to injection f Ab 0.7 Estimated malignancies. Our modeling results determined that the tissue absorbed dose from 227 Ac would be negligible in the context of therapy, less than 0.02 mGy/MBq for the top 6 highest absorbed dose tissues and less than 0.007 mGy/MBq for all other tissues. Compared to that from 225 Ac, the absorbed dose from 227 Ac would make up a very small component (< 0.04%) of the total absorbed dose delivered to the 6 highest dose tissues: red marrow, spleen, endosteal cells, liver, lungs and kidneys when accelerator produced 225/7 Ac-conjugated anti-CD33 antibody would be used to treat Results of A Ab PK model simulations for 225 Ac-bound Ab. B PK for free 227 Th C PK for free 223 Ra leukemia patients. For all tissues evaluated, the dominant contributor to the absorbed dose arising from the 227 Ac is 227 Th, the first daughter of 227 Ac, which has the potential to deliver absorbed dose both while it is antibody-bound and while it is free. These results suggest that the absorbed dose arising from 227 Ac to normal organs would be negligible for an 225/7 Ac-labeled antibody that targets hematological cancer.
In addition to the models used in these simulations and their related parameters, the following series of assumptions were used to arrive at these conclusion: (1) At time of administration there is 0.7% 227 Ac in the 225/7 Ac-conjugated anti-CD33 antibody. (2) 70% of the 227 Th resulting from 227 Ac decay remains antibody-bound and follows the same kinetics as the actinium-conjugated antibody. (3) 227 Th decay releases free 223 Ra. Under the simulation conditions described above, the spleen, red marrow, endosteal cells and liver would receive the highest absorbed doses from 227 Ac and its daughters. The simulation also assumes high purity in the radiolabeled material so that loss of the labeled Ab also removes the Ac-227 conjugated to the Ab.   It should be noted that different simulation models, parameter values and assumptions will give different results. In particular, these results may not apply to nonantibody carriers. The simulations and absorbed dose calculations were performed assuming 1 kg of antigen-positive cells in an adult female. Within the context of antibody-targeting of hematologic malignancies, alternative assumptions regarding percent 227 Ac in the injectate and the fraction of 227 Th that remains-antibody-bound may be implemented by scaling the listed absorbed dose values by the ratio of the new values with those used in this paper (e.g., by considering the scaling applied in Eqs. 8-16. For example, lower Ab retention of 227 Th following decay of 227 Ac may be obtained by scaling 227 Th and daughter absorbed doses by the new retention fraction divided by 0.7 (= f Ab ). Such scaling can also account for injectate purity.

Conclusions
Using a pharmacokinetic model relevant to treating patients with leukemia and models describing the PK of free thorium and radium, the dose contribution of a 0.7% 227 Ac in accelerator-produced 225 Ac would be negligible in the context of αRPT therapy, less than 0.02 mGy/MBq for the top 6 highest absorbed tissues and less than 0.007 mGy/MBq for all other tissues.
The conclusion above is specific to the parameter values and assumptions outlined and may not apply to lower molecular weight agents or other cancer targets.
Using this model, the amount of radiolabeled antibody in plasma ( Q P (t) ), liver ( Q L (t) ), spleen ( Q S (t) ) and red marrow ( Q RM (t) ) as a function of time, t, may be obtained using the equations below: with f L1 , f S1 , fraction of Ab in the vascular or extracellular fluid space of the liver (L), or spleen (S); f L2 , f S2 , fraction of AbAg in the liver (L), or spleen (S); V RMECF , V d , red marrow ECF volume and total Ab distribution volumes.
Further details regarding this model, including the derivation of Eqs. 1-3, are in reference [13].