Introduction

Non-Hodgkin’s lymphoma (NHL) is the most frequent hematological malignancy in adults. NHL is classified into two groups in clinical practice: the indolent (low-grade) lymphoma and the aggressive (high-grade) lymphoma. Follicular lymphoma (FL) is the most frequent indolent NHL, and diffuse large B-cell lymphoma (DLBCL) represents the most frequent type of aggressive NHL. Antibody-based therapies have been developed in the treatment of NHL [1–3]. The administration of rituximab, a monoclonal antibody (mAb) targeting cluster of differentiation 20 (CD20), has markedly improved the treatment response and disease outcome of NHL [1]. However, indolent NHL is not curable and a certain number of patients with an indolent NHL (about 40% of FL patients) eventually experience relapsed or refractory disease with a more aggressive histology [4]. The aggressive types of NHL are heterogenous and some of them still relapse after chemotherapy (at least one-third of DLBCL patients are refractory to first-line chemotherapy) [5, 6]. Thus, there is a need for alternative strategies that are less toxic and more effective for patients with relapsed or refractory NHL compared with conventional therapy.

Radioimmunotherapy (RIT) is a therapy to selectively deliver therapeutic radionuclides to tumor lesions by conjugating radionuclides to mAbs while normal tissue toxicity is limited. The effectiveness and safety of RIT such as Zevalin [yttrium-90 (⁹⁰Y)-ibritumomab-tiuxetan] and Bexxar [iodine-131 (¹³¹I)-tositumomab] have been proven relative to rituximab in relapsed or refractory NHL. New, effective therapies have been developed and are playing against the spread of RIT in the NHL. Indeed, the sales of Bexxar were discontinued in 2014 for commercial reasons. Zevalin is still available in the market but its sales have reportedly declined compared with a decade ago. However, most of the relapsed or refractory NHL patients are elderly and their therapeutic options are limited. Thus, RIT will be suitable for patients suffering from NHL.

To prevent long-time retention of the radionucleotides in the blood and decrease toxicity, the concept of pre-targeted RIT (PRIT) has been proposed [7]. In PRIT, a non-radioactive antibody is administered and allowed to bind to a tumor antigen. Subsequently, a radioactive payload is injected and captured by a cell-bound antibody. PRIT is currently being evaluated in both preclinical and clinical investigations.

Positron emission tomography (PET) using radiolabeled mAb (immuno-PET) is a new imaging modality that combines mAb and PET radionucleotides and provides pharmacokinetic and pharmacodynamic information. PET using fluorine-18 (¹⁸F)-fluorodeoxyglucose (FDG) has been recommended for relapse detection and therapy assessment in NHL. Nevertheless, ¹⁸F-FDG is not tumor-specific and does not allow the selection of patients for targeted therapy since FDG evaluates only glucose metabolism. To develop immuno-PET using mAb, a positron emitter that shows a long half-life enough for blood clearance of antibodies would be ideal. Recent reports have described long-lived PET radionuclides such as zirconium-89 (⁸⁹Zr) are trapped inside the tumor cell after antibody internalization and thus produce high-resolution, excellent contrast imaging. With current advancements, immuno-PET could be used in the evaluation of NHL as a novel imaging for the selection of patients and treatment response assessment.

This review summarizes the recent preclinical and clinical advances in RIT for treating NHL and discuss future applications of RIT and radiotheragnostic agents.

β-Particle RIT

In RIT, either β- or α-emitting radionucleotide can be linked to the antibodies for delivery of radiation to the tumor cells. Therapeutic radionucleotides are selected on the basis of their particle emission, half-life, energy, and path length. The path length is defined by the distance that the charged particles can travel. The treatable tumor size is determined by path length. Potential β- and α-emitting radionucleotides for RIT of NHL are shown in Table 1.

β-Particle radiation can exert a direct toxic effect on the cell bound by the antibody and eliminate surrounding tumor cells via cross-fire effect. The cross-fire effect can kill cells that have low levels of antigen expression and are not accessible to the antibody [8]. The therapeutic effect of CD20 mAb therapy can be enhanced by the antibody-conjugated β-emitting radionuclide [9]. β-Particle RITs targeting CD20 kill not only CD20⁺ lymphoma cells but also CD20⁺ normal B-cell. However, normal B-cells eventually recover because CD20 is not expressed on B-cell precursors and hematopoietic stem cells. Two β-particle RITs targeting CD20 have been approved for relapsed or refractory NHL. A summary of β-particle RITs evaluated in clinical trials of NHL is shown in Table 2.

The first RIT is ⁹⁰Y-labeled ibritumomab tiuxetan. In February 2002, ⁹⁰Y-ibritumomab tiuxetan (Zevalin) was approved by the USA FDA to treat relapsed or refractory low-grade B-cell NHL. Ibritumomab is a murine variant of rituximab that targets the same CD20 epitope as rituximab. Tiuxetan is covalently bound to ibritumomab and chelates with indium-111 (¹¹¹In, for imaging) or ⁹⁰Y (for therapy). The patients need to receive non-radiolabeled rituximab to block normal B-cells in the blood circulation and in the spleen before ⁹⁰Y-ibritumomab tiuxetan treatment. In a phase I/II study, ⁹⁰Y-ibritumomab tiuxetan resulted in durable responses in patients with NHL including FL, DLBCL, non-follicular low-grade, and mantle cell lymphoma overall response rate (ORR) was 73%. Complete response (CR)/CR unconfirmed (CRu) was 51% and partial response (PR) was 22% [10]. The phase III randomized trial compared a single intravenous dose of ⁹⁰Y-ibritumomab-tiuxetan with four doses of rituximab in 143 patients with relapsed or refractory low-grade, follicular, or transformed CD20⁺ transformed NHL [11]. The ⁹⁰Y-ibritumomab-tiuxetan group showed a statistically significant higher ORR (80% vs. 56%) in comparison to the rituximab alone group. In patients with rituximab-refractory FL (no objective response to rituximab or time to progression of ≤ 6 months), ⁹⁰Y-ibritumomab tiuxetan exhibited excellent effectiveness with 74% ORR (15% CR and 59% PR) [12]. The efficacy of ⁹⁰Y-ibritumomab tiuxetan tended to be lower in the patients who relapsed after treatment with rituximab [13, 14]. In the phase II clinical study of ⁹⁰Y-ibritumomab-tiuxetan as first-line monotherapy for FL, a single injection of ⁹⁰Y-ibritumomab-tiuxetan achieved high response rates (56% CR/CRu and 31% PR) and was well tolerated [15]. Two doses of ⁹⁰Y-ibritumomab-tiuxetan as initial therapy of advanced-stage FL showed excellent response rates (initial ORR was 94.4% and CR/CRu was 58.3%) in a phase II trial [16]. RIT consolidation with ⁹⁰Y-ibritumomab-tiuxetan was highly effective for advanced-stage FL [17, 18] or DLBCL [19]. In addition, myeloablative conditioning with Zevalin (⁹⁰Y-ibritumomab tiuxetan) plus 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU), etoposide, cytarabine, melphalan (BEAM, Z-BEAM) showed improved overall survival (OS) at 4 years with lower toxicity compared with total body irradiation-based conditioning regimens (81.0% vs. 52.7%) [20].

The second is ¹³¹I-labeled tositumomab. Tositumomab is a murine mAb targeting CD20. In June 2003, ¹³¹I-tositumomab (Bexxar) was approved to treat relapsed or refractory low-grade B-cell NHL by the USA FDA. Non-radiolabeled tositumomab is administered intravenously to bind normal B-cells in blood circulation and the spleen before the injection of ¹³¹I-tositumomab. In an integrated analysis of the five clinical studies, a single course of ¹³¹I-tositumomab administration showed high response rates in patients with relapsed/refractory low-grade or transformed low-grade NHL [21]. Response rates ranged from 47% to 68%, CR rates ranged from 20% to 38%, and the 5-year progression free survival (PFS) was 17% [21]. In treatment-naive advanced-stage FL, a single one-week course of ¹³¹I-tositumomab therapy was assessed as an initial treatment [22]. After ¹³¹I-tositumomab therapy, 95% had any response and 75% had a CR. Hematological toxicity was moderate without any transfusions. In a phase III randomized study, cyclophosphamide, hydroxydaunorubicin (adriamycin), oncovin (vincristine), and prednisone (CHOP) followed by ¹³¹I-tositumomab demonstrated significantly better 10-year PFS compared with rituximab-CHOP as an initial treatment for previously untreated patients with FL (56% vs. 42%) [23]. Autologous stem cell transplantation following high-dose RIT using ¹³¹I-tositumomab demonstrated improved OS and PFS in relapsed FL patients compared with conventional high-dose therapy [24]. Given these favorable outcomes, β-particle RIT remains the promising approach for patients with NHL although the marketing of Bexxar was discontinued in February 2014 due to a decline in usage.

CD22 is detected in more than 90% of patients with NHL including DLBCL and FL. Epratuzumab is a humanized mAb to CD22. In a phase I/II study, fractionated anti-CD22 RIT using ⁹⁰Y-epratuzumab tetraxetan showed a high rate of durable CR and ORR in 41% to 73% of the patients with relapsed/refractory NHL [25]. High rates of CR/CRu (92%) and increased median PFS (24.6 months) were shown in patients with relapsed/refractory FL receiving the highest ⁹⁰Y dose levels (> 30 mCi/m²).

CD37-targeting RIT has also been developed because CD37 is detected on not only mature normal B-cells but also the majority of B-cell NHL. Betalutin, ¹⁷⁷Lu-lilotomab satetraxetan, consists of the β-particle ¹⁷⁷Lu chelated to the chemical linker satetraxetan conjugated to the murine anti-CD37 mAb lilotomab. The bone marrow toxicity of ¹⁷⁷Lu is relatively low due to its short β range (Table 1). In addition, γ rays emitted by ¹⁷⁷Lu provide imaging of biodistribution and dosimetry measurements. The safety of ¹⁷⁷Lu-lilotomab satetraxetan has been reported in relapsed CD37⁺ indolent NHL [26–30]. Treatment with ¹⁷⁷Lu-lilotomab satetraxetan shows excellent therapeutic efficacy in NHL preclinical models [31–34]. ¹⁷⁷Lu is also radiolabeled to ofatumumab which is a fully human anti-CD20 mAb and shows more efficient binding to CD20 antigen compared with rituximab. The therapeutic effect of ¹⁷⁷Lu-ofatumumab was tested using the disseminated Raji lymphoma model [35]. When therapy was initiated 4 days after cell injection, 8.51 MBq of ¹⁷⁷Lu-ofatumumab caused the elimination of bioluminescence-detectable tumors and no apparent effect on the whole-body.

The decreased level of CD20 expression is considered to be one of the major contributing factors for anti-CD20 mAb response [36]. Although the loss of CD20 antigen is assumed in the resistant mechanism of CD20-targeted therapy [37, 38], it is difficult to evaluate the frequency of CD20 loss because re-biopsy is not performed in most patients after CD20-targeted therapy. Thus, there is little information about the resistance mechanism of RIT targeting CD20.

In RIT targeting CD20, non-radiolabeled anti-CD20 mAb (rituximab) is injected prior to anti-CD20 radioimmunoconjugate (⁹⁰Y-labeled ibritumomab tiuxetan and ¹³¹I-labeled tositumomab). Several lines of preclinical and clinical studies have shown administering the non-radiolabeled anti-CD20 mAb improves tumor targeting and prolongs the blood residence time of the radioimmunoconjugates. In previous studies using preclinical mouse models, circulating non-radiolabeled antibodies have the possibility to compete with the radioimmunoconjugate compromising the tumor uptake and therapeutic efficacy of the radioimmunoconjugate [39, 40]. To improve the response of RIT, the efficacy of dual-targeted RIT has been tested. Mattes et al. [41] evaluated the combination of an unconjugated humanized anti-CD20 mAb (veltuzumab) with a ⁹⁰Y-epratuzumab tetraxetan in nude mice bearing Burkitt lymphoma (Ramos). In this investigation, tumor response and survival were improved using ⁹⁰Y-epratuzumab tetraxetan along with non-radiolabeled veltuzumab compared with either of them alone. Weber et al. [42] tested combining rituximab with ¹⁷⁷Lu-conjugated humanized anti-CD22 mAb, huRFB4, in a subcutaneous Raji lymphoma model. Treatment with ¹⁷⁷Lu-conjugated huRFB4 significantly attenuated lymphoma growth and prolonged survival in comparison with ¹⁷⁷Lu-conjugated rituximab. In a phase I study including 18 patients with relapsed aggressive B-cell NHL, the combination of two injections of ⁹⁰Y-epratuzumab tetraxetan (222–555 MBq/m²) with four injections of veltuzumab was evaluated [43]. For ⁹⁰Y-epratuzumab tetraxetan, the maximum tolerable dose was 222 MBq/m² because of myelosuppression. Of 17 assessable patients, ORR was 53% including three (18%) CR and six (35%) PR. The combination therapy with ⁹⁰Y-epratuzumab tetraxetan and veltuzumab was tolerable and promising in the elderly or patients who relapse or are not suitable for stem-cell transplantation.

α-Particle RIT

RIT using β-particles has been already approved in patients with NHL, but its clinical application is gradually declining due to some limitations. β-Particles damage surrounding healthy tissues and provide the suboptimal killing of tumor cells because of relatively long tissue penetration ranges and low decay energies. α-Particles, in contrast, have a very short tissue penetration range (50–100 μm) and high linear energy transfer (Table 1); therefore, mAbs labeled with α-particles show high specific killing effects on tumor cells and minimal damage to surrounding normal tissue. α-Particle RIT is an attractive therapy for patients with NHL. A summary of preclinical and clinical investigations in α-particle RIT of NHL is shown in Table 3 and Table 4. Treatment with ²²⁷Th-1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA)-p-benzyl-rituximab suppressed the lymphoma growth in nude mice bearing Burkitt lymphoma (Raji) and caused significantly longer survival than β-emitting ⁹⁰Y-ibritumomab-tiuxetan [44]. ²¹²Pb-rituximab significantly prolonged survival compared with rituximab and the ²¹²Pb-isotypic control in a murine syngeneic lymphoma model [45]. Treatment with ²¹²Pb-1,4,7,10-tetrakis(carbamoylmethyl)-1,4,7,10-tetraazacyclododecane (TCMC)-NNV003 (anti-CD37) demonstrated efficacy and safety in preclinical models of CD37-expressing chronic lymphocytic leukemia and NHL [46]. Treatment with ²¹³Bi-rituximab was more effective in severe combined immunodeficient (SCID) mice bearing Burkitt lymphoma (Raji) than ²¹³Bi or ²¹³Bi anti-human epidermal growth factor receptor 2 (HER2)/neu [47]. Green et al. [48] assessed the anti-CD20 (1F5-B10) mAb labeled with the α-emitting radio-halogen ²¹¹At in both subcutaneous and disseminated lymphoma xenograft models. In a subcutaneous lymphoma xenograft model, high doses of ²¹¹At 1F5-B10 (48 µCi) treatment showed modest attenuation in lymphoma proliferation and slightly longer survival compared with no treatment. In a disseminated lymphoma model, a 15 µCi dose caused complete eradication of the lymphoma in 70% of mice. These results suggest α-particle RIT is more effective for small lymphoma cell clusters in a disseminated model because of its short ranges as expected. In mice bearing subcutaneous Raji lymphomas, ²²⁵Ac-labeled anti-CD20 ofatumumab treatment specifically killed lymphoma cells and showed dose-dependent curative therapeutic efficacy [49]. In a first-in-human dose-escalation phase I study, α-particle emitting ²²⁷Th-labeled anti-CD22 antibody (BAY 1862864) showed clinical safety and tolerability in patients with CD22-positive relapsed/refractory B-cell NHL [50].

There are several advantages in terms of increased cytotoxic effects on tumor cells and reduced non-specific bystander responses on normal tissue in α-particle RIT. However, α-particle RIT would be a challenge in the clinical setting because of the short half-life and low production capability. Further advancements in the method of production will reduce production costs and increase the use of α-particle RIT.

In contrast to β-particle radiation, α-particle radiation doesn’t target surrounding tumor cells and cells showing low levels of antigen expression due to its very short tissue penetration range. To target surrounding tumor cells and tumor heterogeneity, enhancing anti-tumor activities via bystander killing and immunogenic cell death needs to be explored in α-particle RIT [51, 52].

In α-particle RIT, low molecular weight ligands instead of antibodies are desirable for targeting moieties because they migrate into the cell nucleus more easily than antibodies and reduce toxicity. For the therapy of metastatic, castration-resistant prostate cancer, ²²⁵Ac-prostate-specific membrane antigen (PSMA)-617 shows a remarkable therapeutic efficacy against tumor cells [53, 54]. Similar strategies need to be developed in α-particle RIT for NHL.

α-Emitting radionucleotides can be used in combination with diagnostic imaging and therapeutic RIT. Most α-emitting radionucleotides emit photons during their decay. While the biodistribution of α-emitters such as ²¹¹At can be evaluated by single photon emission computed tomography (SPECT) imaging [55], obtaining high-quality SPECT imaging is difficult due to the low intensity of photons. The diagnostic imaging surrogates need to be developed to perform the assessment of pre-therapy in α-particle RIT.

Advantages and disadvantages of RITs compared with antibody-drug conjugates

Theragnostics integrates diagnostic nuclear medicine and RIT, helping to predict the response and toxicity of RIT. The clinicians can manipulate RIT dosage according to the imaging data. In this setting, RIT must be performed in the hospital with the control area for radiation. Therefore, it is difficult to perform RIT in the general hospital. Antibody-drug conjugates (ADCs) are immunoconjugates composed of a mAb tethered to a cytotoxic drug (instead of a radionucleotide in RIT) via a chemical linker [56]. ADCs have bystander effects which are similar to β-RITs and can be administered in the general hospital. However, if the cytotoxic drugs are resistant to tumor cells, the efficacy is significantly attenuated and this is one of the differences from RIT [57]. A comparative analysis between RIT and other upcoming therapies including ADCs is needed for choosing NHL therapy.

Single-domain antibodies

The toxicities including myelotoxicity in full mAb-based RIT are related to the long blood residence time of mAb (several days). In addition, its image acquisition is late because of slow blood clearance. The large molecular weight of mAb (around 150 kDa) impairs tissue penetration and binding to hidden antigens. To improve these disadvantages, smaller mAb-derived fragments have been engineered. Single-domain antibodies (sdAbs) are small antigen-binding fragments generated from heavy chain-only antibodies found in Camelidae [58]. In general, sdAbs bind to their targets with high affinity and specificity. Due to their small molecular weight (10–15 kDa), sdAbs show faster blood clearance and better tissue penetration compared with full-size mAbs. Krasniqi et al. [59] assessed the efficacy of radiolabeled anti-CD20 sdAbs using human CD20⁺ lymphoma mouse models. In this investigation, gallium-68 (⁶⁸Ga)-labeled anti-CD20 sdAb was utilized for PET imaging and ¹⁷⁷Lu-labeled sdAb was administered to treat CD20⁺ lymphoma cells. In the imaging using ⁶⁸Ga-labeled anti-CD20 sdAb, the tumor uptake was specific and the accumulation in nontarget organs (except the kidneys) was low. Although both ¹⁷⁷Lu-labeled anti-CD20 sdAb and ¹⁷⁷Lu-labeled rituximab prolonged the median survival rate of treated mice to the same degree, absorbed doses to healthy organs except the kidneys were much higher for ¹⁷⁷Lu-labeled rituximab in comparison with ¹⁷⁷Lu-labeled anti-CD20 sdAb. These results indicate that radiolabeled anti-CD20 sdAbs are promising theragnostic approaches to target CD20⁺ NHL.

Collectively, sdAb theoretically shows the high-affinity ability. However, there is no clear evidence of superior specificity of sdAb so far and further studies are required for its clinical application.

Conclusions

RIT using anti-CD20 mAbs labeled with β-emitting radionucleotides has shown excellent anti-lymphoma effect and reduced toxicity in the therapy of relapsed/refractory indolent NHL or transformed indolent NHL. Previous clinical studies have shown RIT has the potential to be used as not only a first-line therapy but also consolidation and myeloablative conditioning before stem cell transplantation of advanced-stage indolent or aggressive NHL. Although β-particle RITs including ⁹⁰Y-ibritumomab tiuxetan and ¹³¹I-tositumomab are highly effective for NHL, the number of patients referred for these RITs has gradually decreased in the world over the last two decades due to competing novel therapies and difficulty in performing them in the general hospital. However, targeting surface antigens other than CD20 is also promising and novel therapeutic approaches have continued to be developed in RIT. Recent studies have shown that α-particle RIT is highly effective in preclinical models of NHL. Efficient developments in radiochemistry will reduce production costs and contribute to the increased availability of α-particle RIT. Furthermore, novel treatment strategies such as PRIT have recently made progress and have been expected to be applied clinically to the treatment of NHL. Immuno-PET can provide valuable information about tumor heterogeneity and quantification of antigen expressions as theragnostic approach. The theragnostic approaches including immuno-PET will be vital in monitoring the response of future RIT and other therapies.

Overall, the concept of RIT allows for reduced toxicity and enhances the therapeutic effect of mAbs. RIT is suitable for patients with relapsed or refractory NHL or those who are ineligible for intensive therapies. Further clinical trials will exhibit the potential of future RIT in treating patients with NHL.