A Site-Specifically Labeled Antibody-Drug Conjugate for Simultaneous Therapy and ImmunoPET

Abstract

The conjugation of antibodies with cytotoxic drugs can alter their in vivo pharmacokinetics. As a result, the careful assessment of the in vivo behavior — and specifically the tumor-targeting properties — of antibody-drug conjugates represents a crucial step in their development. In order to facilitate this process, we have created a methodology that facilitates the dual labeling of an antibody with both a toxin and a radionuclide for positron emission tomography (PET). To minimize the impact of these modifications, this chemoenzymatic approach leverages strain-promoted azide-alkyne click chemistry to graft both cargoes to the heavy chain glycans of the immuoglobulin’s F_c domain. As a proof-of-concept, a HER2-targeting trastuzumab immunoconjugate was created bearing both a monomethyl auristatin E (MMAE) toxin as well as the long-lived positron-emitting radiometal ⁸⁹Zr (t_1/2 ~ 3.3 days). Both the tumor targeting and therapeutic efficacy of the ⁸⁹Zr-trastuzumab-MMAE immunoconjugate were validated in vivo using a murine model of HER2-expressing breast cancer. The site-specifically dual-labeled construct enabled the clear visualization of tumor tissue via PET imaging, producing tumoral uptake of ~70 %ID/g. Furthermore, a longitudinal therapy study revealed that the immunoconjugate exerts significant anti-tumor activity, leading to a >90% reduction in tumor volume over the course of 20 days.

Graphical Abstract

INTRODUCTION

The idea of conjugating toxins to antibodies in order to enhance the tumor-specific delivery of chemotherapeutics dates back to the middle of the 20^th century. This field has largely been driven by two parallel trends: (1) the discovery of non-selective yet extremely powerful cytotoxic agents and (2) the development of immunoglobulins with extraordinarily high affinity and selectivity for cancer biomarkers. Over the years, the study of these immunoconjugates — now ubiquitously known as antibody-drug conjugates (ADCs) — has exploded. To wit, as of 2017, there are a remarkable 108 different reviews containing “antibody-drug conjugate” in their title.¹

A sizeable contingent of ADCs has been synthesized via the random conjugation of toxins to the amino acids of the antibody, almost certainly due to the relative ease of this approach. For example, one of the two ADCs currently approved by the United States FDA — KADCYLA® — is synthesized through the random conjugation of emtansine to the lysines of the HER2-targeting antibody trastuzumab.² However, recent years have witnessed a dramatic shift in the field toward ADCs synthesized using site-specific conjugation methods. This pivot toward more well-defined and homogeneous ADCs has been fueled in large part by several studies demonstrating the superior in vivo performance of site-specifically modified immunoconjugates as well as the exigencies of the regulatory review process.^3–7 Yet it is steadily becoming apparent that even the site-specific conjugation of payloads to biomolecules may not be as benign as previously thought. Recent studies have clearly demonstrated that the attachment of cargoes to biomolecular vectors can dramatically alter the pharmacokinetic profiles of the bioconjugates and even impede their ability to reach their target in vivo.^8,9

In light of this data, assessing the tumor targeting and in vivo behavior of ADCs using data obtained for the parent antibody is clearly a shortcut rife with problems. Admittedly, reliable methods do exist for directly determining the concentration of ADCs in blood and serum; however, techniques for quantifying drug concentrations in other tissues remain imperfect. The latter, for example, has typically been performed by analyzing tissues from organs harvested during necropsy, hardly a viable approach with human patients undergoing therapy.

Clearly, the optimal solution is a targeted drug delivery system that can be tracked in vivo using quantitative non-invasive imaging. Non-invasive imaging modalities —primarily magnetic resonance imaging (MRI) or positron emission tomography (PET) — have already been used to facilitate the visualization and quantification of nanomedicine. Grange et al., for example, coated doxorubicin-loaded liposomes with gadolinium complexes and used MRI to verify the tumoral delivery of the liposomes in mice bearing Kaposi’s sarcoma.¹⁰ More recently, a group from King’s College (UK) labeled doxorubicin- and alendronate-containing liposomes with ⁸⁹Zr and used PET imaging to study the biodistribution of the particles in mice bearing mammary carcinoma.¹¹

Yet somewhat surprisingly, this approach has been applied to ADCs remarkably sparingly, especially considering the advent of immunoPET over the last two decades. Indeed, to the best of our knowledge, only two reports exists detailing ADCs that have also been labeled with a radionuclide for nuclear imaging. In this first, Cohen, et al. used a dual radiolabeling approach to create a ⁸⁹Zr-labeled variant of trastuzumab that also bears a ¹³¹I-labeled tubulysin A analogue. The presence of two different radionuclides allowed the authors to track the biodistributions of the antibody and the toxin independently and quantify their delivery to both target and non-target tissues.¹² In the second, Boswell, et al. radiolabeled an anti-TENB2 ThioMab conjugated with monomethyl auristatin E (MMAE) with ¹¹¹In.¹³ The in vivo tracking of the ¹¹¹In-labeled ADC via SPECT imaging allowed the authors to assess the effect of the pre-injection of unconjugated antibody on the ability of the ADC to reach TENB2-expressing xenografts.

The limits of traditional bioconjugation strategies almost certainly play a role in this dearth of imaging-enabled antibody-drug conjugates (^IEADCs). If, as we have noted, the random conjugation of a single payload can create problems, the random attachment of two different moieties only multiplies potential complications, including heterogeneity, impaired immunoreactivity, and suboptimal in vivo performance. In the investigation at hand, we have circumvented this issue by creating what is — to the best of our knowledge — the first immunoconjugate that has been site-specifically labeled with both a toxin and a positron-emitting radiometal. More specifically, we have used a chemoenzymatic approach to create DFO:MMAE-^sstrastuzumab, an ^IEADC in which both monomethyl auristatin A (MMAE) and the radiometal chelator desferrioxamine (DFO) have been conjugated to the heavy chain glycans of the HER2-targeting antibody trastuzumab. Using a murine model of HER2-expressing breast cancer, we clearly illustrate that this ^IEADC is an effective therapeutic agent and can be tracked noninvasively using PET imaging when labeled with the long-lived positron-emitting isotope ⁸⁹Zr (t_1/2 ~3.3 d). Ultimately we envision that dual-labeled radioADCs such as ⁸⁹Zr:MMAE-^sstrastuzumab could play important roles in clinic, both as more accurate scout imaging agents prior to ADC-based therapy and as imaging agents capable of providing real-time dose information during treatment.

EXPERIMENTAL PROCEDURES

Modification of Trastuzumab

5.0 mg of trastuzumab in 750 μL of buffer (50 mM Bis-Tris, 100 mM NaCl, pH 6.0) were incubated with 40 μL of β-(1,4)-galactosidase (2.0 U/mL) at 37°C overnight. A solution of 60 μL of 1 M Tris, 12 μL of 1M solution of MnCl₂, 100 μL of GalT, 30 μL of a 40 mM solution of UDP-GalNAz, and 300 μL of deionized water was then added to the antibody solution, and the resultant mixture was incubated at 30 °C overnight. The excess of UDP-GalNAz was removed by washing the functionalized antibody 4 times with TBS (50 mM Tris.HCl, 150 mM NaCl, pH 7.6) using a 2 mL Amicon Centrifugal Filter with a 50K molecular weight cut-off. The solution of antibody was diluted in TBS to reach a concentration of 2g/L (2200 μL), and the DIBO precursors (in solution in DMSO at 8mM) were added in order to reach a final DIBO concentration of 0.3 mM (DFO-^sstrastuzumab: 82.5 μL of DIBO-DFO; DFO:MMAE-^sstrastuzumab: 41.3 μL of DIBO-DFO, 41.3 μL of DIBO-MMAE) The mixture was incubated overnight at room temperature before purification via size-exclusion chromatography (Sephadex G-25 M, PD-10 column, GE Healthcare; dead volume = 2.5 mL, eluted with 2 × 1 mL fractions of TBS, pH 7.6) and concentration using a 2 mL Amicon Centrifugal Filter with a 50K molecular weight cut-off, ultimately producing the immunoconjugates in a global yield of 70%.

Radiolabeling of Immunoconjugates

For each antibody construct, 500 μg of immunoconjugate solution was diluted to 400μL with PBS, pH 7.4. [⁸⁹Zr]Zr-oxalate (1500 μCi) in 150 μL of 1.0 M oxalic acid was adjusted to pH 7.0–7.5 with 1.0 M Na₂CO₃. After the bubbling of CO₂ stopped, the ⁸⁹Zr solution was added to the antibody solution, and the resulting mixture was incubated at room temperature for 1h. The reaction progress was then assayed using iTLC and an eluent of 50mM EDTA (pH 5). Subsequently, the reaction was quenched with 13 μL of 50mM of EDTA (pH=5), and the antibody construct was purified using size exclusion chromatography (Sephadex G-25 M, PD-10 column, GE Healthcare; dead volume = 2.5 mL, eluted with 500 μL fractions of PBS, pH 7.4) and if necessary concentrated via centrifugal filtration units with a 50,000 molecular weight cut off (Amicon™ Ultra 4 Centrifugal Filtration Units, Millipore Corp. Billerica, MA). The radiochemical purity of the final radiolabeled bioconjugate was assayed by radio-TLC again using 50mM EDTA (pH 5) as an eluent. In the iTLC experiments, free ⁸⁹Zr⁴⁺ cations and [⁸⁹Zr]-EDTA elute with solvent front, while radiolabeled antibody construct remains at the baseline.

PET imaging

PET imaging experiments were conducted on an Inveon microPET-CT (Siemens). Athymic nude mice bearing subcutaneous BT474 xenografts (left shoulder, 60–120 mm³, 25–30 days after inoculation) were administrated with the radioimmunoconjugates (150 μCi, 60–65μg, in solution in 200 μL of saline) via tail vein injection. Approximately 5 min before PET imaging, mice were anesthetized by inhalation of a 2% isoflurane (Baxter Healthcare):oxygen gas mixture and placed on the scanner bed. Anesthesia was maintained using a 1% isoflurane mixture. PET data for each mice were recorded via static scans at 24 h, 72 h and 120 h p.i.

Therapy study

Athymic nude mice bearing subcutaneous BT474 xenografts (left flank; 60–120 mm³) were randomized in 5 groups (n=10) before the study and were administrated with DFO:MMAE-^sstrastuzumab at 2.5 or 10 mg/kg, or with controls (saline solution, 10 mg/kg native trastuzumab, or 0.1 mg/kg free MMAE) via tail vein injection. Following injection, the length (L) and width (W) of the tumors were measured every 2–3 days using calipers, and the tumor volume (V) was assessed from these measured parameters using the formula V=1/2*L*W².

RESULTS AND DISCUSSION

Design and Synthesis

For the synthesis of our site-specifically labeled ^IEADC, we used a reliable and modular approach that our laboratories have developed over the last 5 years: the chemoenzymatic modification of the heavy chain glycans. This strategy — originally inspired by the work of Qasba and Hseih-Wilson — relies on two sequential enzymatic reactions to append terminal azide-bearing monosaccharides to the pair of biantennary sugar chains on the C_H2 domain of the antibody’s F_C region.^14–16 Subsequently, a strain-promoted azide-alkyne cycloaddition (SPAAC) reaction between the azide-bearing antibody and cyclooctyne-modified payloads facilitates the site-specific grafting of cargoes to the immunoglobulin. This approach seemed ideally suited for our application for two reasons. First, we have already successfully demonstrated the utility of this strategy for dual-modification and, just as importantly, demonstrated that the ratio of the two payloads can be altered by modifying the initial ratio of DIBO-bearing precursors in the reaction mixture.^17,18 Second, non-radiolabeled ADCs have recently been synthesized using analog site-specific glycan modification methods and exhibited unimpaired therapeutic potency.¹⁹

The design choices for this proof-of-concept study are straightforward. Trastuzumab was chosen as the targeting vector because the antibody is robust, extraordinarily well characterized, and has been the basis of a variety of different ADCs.^2,20–22 The tubulin inhibitor monomethyl auristatin E (MMAE) was selected as the toxin because ADCs bearing MMAE have been used widely in cancer therapy, with several of these agents currently in clinical trials.²⁸ Finally, ⁸⁹Zr and its gold standard chelator desferrioxamine (DFO) were enlisted due to the advantageous match between the radionuclidic half-life of ⁸⁹Zr (t_1/2 ~3.3 days) and the biological half-life of immunoglobulins as well as the clear clinical utility demonstrated by ⁸⁹Zr-immunoPET.²⁹

The chemoenzymatic modification of trastuzumab was performed according to established procedures via the incubation of the antibody with β−1,4-galactosidase and then the promiscuous galactosyltransferase GalT(Y289L) in conjunction with the azide-modified sugar GalNAz (Fig. 1A). Subsequently, this N₃-^sstrastuzumab intermediate was reacted with cyclooctyne-bearing variants of both the drug and chelator — MMAE-DIBO and DFO-DIBO — in mixture in a 1:1 molar ratio to form the completed ^IEADC: DFO:MMAE-^sstrastuzumab (Fig. 1B). The immunoconjugate was prepared with a global yield of 70% over three steps.

Figure 1.

(A) The chemoenzymatic approach used to create the ^IEADC; (B) Schematic structures of MMAE-DIBO and DFO-DIBO.

Chemical and In Vitro Characterization

The site-specific nature of the bioconjugation reaction was verified via both SDS-PAGE and mass spectrometry. The denaturing electrophoresis gel experiment displayed a clear upward shift in the molecular weight of the heavy chain following the ligation of the payloads to the glycans (Fig. 2). In contrast, no shift was observed for the light chain, demonstrating the site-specific character of the conjugation. To complement this data, mass spectrometry analysis was also performed on the DFO:MMAE-^sstrastuzumab immunoconjugate. After F_c-F(ab´)₂ cleavage, the mass spectrometry shows no modification in the Fab region compared to unmodified trastuzumab (Supp. Fig. S1A). However, a mass increase is observed between the F_c domain of the azide modified trastuzumab – N₃-^sstrastuzumab – and the F_c domain of DFO:MMAE-^sstrastuzumab (Supp. Fig. S1B). These data clearly illustrate that the conjugation occurs site-specifically on the C_H2 domain of the F_c region.

Figure 2.

Denaturing polyacrylamide gel electrophoresis of (1) native trastuzumab, (2) DFO-^sstrastuzumab, and (3) DFO:MMAE-^sstrastuzumab, demonstrating the site-specific conjugation on the heavy chain. HC: heavy chains, LC: light chains.

The mass spectrometry of the whole antibody reveals similar results and allowed us to discriminate the different species constituting DFO:MMAE-^sstrastuzumab (Supp. Fig. S1C). As expected, the predominant species bears 2 DFO and 2 MMAE. However, other species containing different combinations of payloads – ranging from 2 DFO:1 MMAE to 4 MMAE – are also present in the mixture. Overall, the average degree of conjugation for each of the payloads is 2.1 ± 0.2 DFO/mAb and 1.7 ± 0.2 MMAE/mAb. It is important to note that while these data clearly illustrate that DFO:MMAE-^sstrastuzumab is not perfectly homogenous, it is far more homogenous than an analogous randomly conjugated ^IEADC. In all cases, the payloads are site-specifically appended to the biantennary glycans chains of the F_C domain glycans chains, far from the antigen-binding complementarity determining regions (CDRs). This distance between the modification site and the CDRs stands as one of the primary advantages of this bioconjugation approach compared to traditional, non-site-specific bioconjugation methods. In the latter, lysines within the CDRs can be modified, detrimentally affecting — or even completely abrogating — the ability of the antibody to bind to its antigen.

Following this chemical characterization, the in vitro HER2-binding properties of DFO:MMAE-^sstrastuzumab were investigated via flow cytometry and surface plasmon resonance. Not surprisingly, flow cytometry experiments with HER2-expressing BT474 human breast cancer cells revealed that DFO:MMAE-^sstrastuzumab exhibited identical immunoreactivity to both native trastuzumab and a variant of trastuzumab that had been site-specifically modified with DFO alone (DFO-^sstrastuzumab) (Supp. Fig. S2). In addition, surface plasmon resonance experiments revealed that both unmodified trastuzumab and the immunoconjugates demonstrated robust binding with purified HER2. All the three constructs yielded nanomolar K_D values, which were well within range of each other (~ 5 nM) (Supp. Fig. S3). These data clearly indicate the immunoconjugates suffer no loss of immunoreactivity for HER2 in vitro compared to the parent antibody.

Radiochemistry and In Vivo PET Imaging

The ^IEADC DFO:MMAE-^sstrastuzumab and its toxin-lacking cousin DFO-^sstrastuzumab were labeled with ⁸⁹Zr in preparation for in vivo PET imaging. Both immunoconjugates were radiolabeled and purified according to standard published protocols, yielding ⁸⁹Zr:MMAE-^sstrastuzumab and ⁸⁹Zr-^sstrastuzumab in >99% radiochemical purity and with specific activities of 2.7 ± 0.1 and 2.5 ± 0.1 mCi/mg, respectively. Subsequent stability assays revealed both radioimmunoconjugates to be >96% intact after 7 days in human serum at 37 °C (Supp. Fig. S4). The immunoreactive fractions of the radioimmunoconjugates with HER2-positive BT474 human breast cancer cell line were similar for both constructs: 0.92 ± 0.02 and 0.95 ± 0.01 for ⁸⁹Zr-^sstrastuzumab and ⁸⁹Zr:MMAE-^sstrastuzumab, respectively. This confirms that the immunoreactivity of the constructs was not adversely affected by the conjugation or radiolabeling procedures.

The in vivo behavior of ⁸⁹Zr:MMAE-^sstrastuzumab and ⁸⁹Zr-^sstrastuzumab was investigated in athymic nude mice bearing subcutaneous HER2-positive BT474 xenografts. To this end, the mice were administered ~150 μCi of each radioimmunoconjugate and imaged at 24, 72 and 120 h post-injection (Fig. 3). It is important to note up front that we elected not to perform control experiments using a HER2-negative cell line or a blocking dose of trastuzumab. This decision was born not of carelessness but rather a desire to avoid treading on well-trod scientific soil. Indeed, the specificity of trastuzumab-based radiotracers for HER2-positive and HER2-negative xenografts is very well-established.^23–27 In this case, the radioADC and ⁸⁹Zr-^sstrastuzumab exhibited very similar behavior. At 24 h p.i., some tumoral uptake is observed, but the activity concentration in the blood pool also remains relatively high. Over time, however, the radiotracers leave the blood compartment and concentrate overwhelmingly within the tumor, allowing for extremely clear visualization of the tumor by 72 and 120 h after injection. The biodistribution data from 120 h p.i. tell a very similar story (Fig 4. and Supp. Table S5). At this time point, both ⁸⁹Zr:MMAE-^sstrastuzumab and ⁸⁹Zr-^sstrastuzumab display relatively low activity concentrations — less than 8 %ID/g — in the blood while concomitantly producing high activity concentrations in the tumor: 76 ± 13 and 70 ± 7 %ID/g, respectively. Uptake in non-target tissues was also low, with the highest activity concentrations observed in the liver, lungs, kidneys, and bone: each tissue displayed slightly above 3 %ID/g with each of the conjugates. Interestingly, in this case, the presence of the MMAE toxin seems to make very little difference on the in vivo behavior of the two constructs. Finally, autoradiography of tumors excised 120 after the administration of ⁸⁹Zr-^sstrastuzumab revealed significant correlation between the localization of the radiotracer and vascular staining with Hoechst 33342 (Supp. Fig. S6)

Figure 3.

Planar (left) and maximum intensity projection (right) PET images of athymic nude mice bearing HER2-expressing BT474 breast cancer xenografts (white arrow) following the injection of ⁸⁹Zr:MMAE-^sstrastuzumab and ⁸⁹Zr-^sstrastuzumab (150 μCi, 60–65μg). The coronal slices intersect the center of the tumors.

Figure 4.

Biodistribution data collected 120 h after the administration of ⁸⁹Zr:MMAE-^sstrastuzumab and ⁸⁹Zr-^sstrastuzumab (150 μCi; 60–65 μg) to athymic mice bearing HER2-expressing, subcutaneous BT474 human breast cancer xenografts. Stomach, small intestine, and large intestine values include contents.

In Vivo Therapeutic Efficacy

After confirming the tumor targeting properties of the ⁸⁹Zr-^IEADC via PET, we next investigated the therapeutic efficacy of the ^IEADC. To this end, a longitudinal treatment study was performed using athymic nude mice bearing subcutaneous HER2-expressing BT474 xenografts. Based on previous studies, doses of 2.5 and 10 mg/kg DFO:MMAE-^sstrastuzumab were chosen for assessment.^19,30,31 The unlabeled ^IEADC was used for logistical reasons: in order to simplify the experiment and reduce the radiation dose associated with repeated measurements of tumor volume. Moreover, in this context, it is critical to remember that in formulation of the ⁸⁹Zr-labeled ^IEADC — as with any ⁸⁹Zr-labeled antibody — the radionuclide is just a statistical label: >99% of the biomolecules are actually the DFO:MMAE-^sstrastuzumab precursor. Due to the high energies of the γ-photons resulting from its decay — 511 and 909 keV from positron annihilation and the decay of ^89mY, respectively — ⁸⁹Zr has very limited cytotoxic effect at short range. Therefore, the presence or absence of ⁸⁹Zr on the ^IEADC should have no impact on the therapeutic effect of the conjugate. Control cohorts were studied that received doses of vehicle alone, 10 mg/kg of native trastuzumab, and 0.1 mg/kg of MMAE (a dose which provides slightly more MMAE than that contained in a 10 mg/kg dose of DFO:MMAE-^sstrastuzumab). Prior to the experiment, the mice were randomized into cohorts of n = 10. Subsequently, the mice of each cohort received single injections of each agent at day 0 of the therapy study. The tumor volume was monitored via caliper for the next 20 days, an endpoint determined by the last remaining day of efficacy of the estrogen pellets needed to grow the BT474 xenografts.

The mice treated with 0.1 mg/kg MMAE or 10 mg/kg native trastuzumab failed to display a significant reduction in tumor volume over the course of 20 days after injection compared to the vehicle-only control cohort (Fig 5. and Supp. Fig. S7). Importantly, however, the administration of 10 mg/kg DFO:MMAE-^sstrastuzumab resulted in a marked, statistically significant reduction in tumor volume compared to the three control groups. Moreover, this therapeutic effect appears to be dose dependent. To wit, while the administration of 2.5 mg/kg of DFO:MMAE-^sstrastuzumab led to a decrease of tumor size to 50% of its initial volume 20 days after injection, the administration of 10 mg/kg of the unlabeled ^IEADC led to an average reduction of the tumors >90% after the same period of time.

Figure 5.

Relative tumor volume of athymic nude mice (n = 10 per group) bearing subcutaneous HER2-expressing BT474 xenografts following the injection (t = 0) of saline (blue), 10 mg/kg native trastuzumab (red), a 0.1 mg/kg MMAE (grey), 2.5 mg/kg DFO:MMAE-^sstrastuzumab (green), and 10 mg/kg DFO:MMAE-^sstrastuzumab (purple). The solid lines are the moving averages.

CONCLUSION

The chemoenzymatic modification of the heavy chain glycans represents a powerful tool for the construction of site-specifically modified antibodies. In this investigation, we have detailed the synthesis, characterization, and in vivo evaluation of ⁸⁹Zr:MMAE-^sstrastuzumab, a site-specifically modified immunoconjuate bearing both a toxin and a positron-emitting radiometal. Chemical and in vitro characterization assays clearly demonstrated the site-specific nature of the bioconjugation reaction as well as the unimpaired ability of the dual-labeled construct to bind HER2-expressing cells. Subsequently, PET imaging and biodistribution experiments study in mice bearing HER2-expressing BT474 human breast cancer xenografts confirmed that the dual-labeled radioADC produces very high activity concentrations in target tissues and exhibits nearly identical in vivo behavior compared to a variant labeled only with ⁸⁹Zr. Even more importantly, a longitudinal therapy study in the same murine model confirmed the dose-dependent therapeutic efficacy of the dual labeled conjugate. Indeed, treatment with 10 mg/kg DFO:MMAE-^sstrastuzumab reduced tumor volumes by >90% over the course of just 20 days. Even beyond facilitating the construction of well-defined, homogeneous, and highly immunoreactive constructs, one of the main advantages of this chemoenzymatic approach is its modularity. This strategy can be used with any antibody and any set of two toxin and reporter cargoes. Moreover, the ratio of the reporter and toxin can be altered in a straightforward manner to create ^IEADCs bearing a higher DOL of toxin.

In the end, the principal advantage of this approach is that it produces an all-but-identical theranostic pair. Practically speaking, we envision these radioADCs can be employed in three different settings: (1) as probes to study in vivo biodistribution and pharmacokinetics in preclinical investigations, (2) as scout imaging agents administered prior to therapy in the clinic, and (3) as companion imaging agents that can provide real-time voxel-by-voxel drug-dose information during ADC-based therapy. Ultimately, it is our hope that dual-labeled constructs such as these help usher in an era in which true theranostics — constructs that can perform both therapy and imaging — become valuable preclinical and clinical tools.

ABBREVIATIONS

ADC

antibody-drug conjugate

^IEADC

imaging-enabled antibody-drug conjugate

DFO

desferrioxamine

MMAE

monomethyl auristatin E

PET

positron emission tomography