Converting an Anti-Mouse CD4 Monoclonal Antibody into an scFv Positron Emission Tomography Imaging Agent for Longitudinal Monitoring of CD4⁺ T Cells

Understanding an immune response requires knowledge of the whereabouts of the cells and molecules charged with its execution. In preclinical studies, an assessment of the in vivo distribution of immune cells is usually done by excision of secondary lymphoid organs after euthanasia. This makes a longitudinal assessment of responses challenging, an approach mostly limited to partial splenectomy or to the analysis of peripheral blood taken at various time points. To track immune responses against tumors and infectious agents noninvasively, a more dynamic analysis of the distribution of lymphocytes in living animals would be desirable. Especially useful would be methods that do not rely on genetic modification of the cell types to be tracked. This goal is achievable using a noninvasive imaging modality such as positron emission tomography (PET) (1–3).

The development of PET imaging agents concerns two broad categories: small molecules and biologicals. Because of their typically short half-lives, the pharmacokinetics of many small molecules to be imaged benefits from the use of short-lived PET isotopes, such as ¹⁸F (t_1/2 = ∼110 minutes) or ¹¹C (t_1/2 = ∼20 minutes), which poses unique and obvious challenges in terms of their synthesis, downstream processing, and purification (4). In contrast, biologicals, such as Igs, have long circulatory half-lives and therefore require the installation of longer-lived PET isotopes, such as ⁶⁴Cu (t_1/2 = ∼12 hours) or ⁸⁹Zr (t_1/2 = ∼3.3 days) (5, 6). The latter approaches are generally poorly compatible with protocols for same-day imaging. This has inspired the search for smaller Ig-derived formats and other protein-derived scaffolds as imaging agents.

Single-chain variable fragments (scFvs) are widely used as the minimal recognition unit that can be extracted from conventional two-chain Igs. They consist of the V_H and V_L portions connected via a linker. ScFvs have enjoyed popularity as the building blocks for the construction of chimeric Ag receptors and bispecific T cell engagers (7, 8). If it were possible to convert full-sized Igs into scFv-based imaging agents, it would enable a noninvasive assessment of the in vivo distribution of the wide range of targets recognized by the available mAbs. However, the use of monovalent scFv fragments for PET has met with limited success (9–11). From a regulatory perspective, conversions of clinically approved Igs might be preferable to the de novo construction of a suitable nanobody of similar specificity, for which use in humans is contemplated. In this study, we demonstrate the feasibility of converting a mAb into an scFv preparation suitable for PET imaging of CD4⁺ T cells.

Commonly used procedures for labeling of Igs and their fragments rely on maleimide chemistry to target cysteine residues or N-hydroxysuccinimide derivatives to modify lysine side chains (12, 13). Installation of an unpaired cysteine through genetic engineering or mild reduction of existing disulfides are the methods of choice for modification of available -SH groups. In this manner, scFvs equipped with a free Cys at the C terminus can be labeled either fluorescently or with other substituents of choice. Methods of chemical modification can be replaced by chemo-enzymatic methods, which have the advantage of site-specificity, high yield, and homogeneity of the desired modified product (14, 15).

In exploring the properties of nanobodies as PET imaging agents, we discovered that their performance could be improved through site-specific installation of a polyethylene glycol (PEG) moiety in addition to the metal chelator used for labeling with ⁸⁹Zr (16). Nonspecific uptake in organs of elimination, such as kidney, liver, and bladder, was drastically reduced for an ⁸⁹Zr-labeled, PEGylated anti-CD8 nanobody in comparison with its counterpart lacking a PEG substituent.

In this study, we applied the same enzymatic labeling and PEGylation strategy to the modification of an anti-mouse CD4 scFv. A comparison of the PEGylated and non-PEGylated monovalent scFv showed improved performance in PET for the PEGylated agent, with results comparable with those reported for diabodies (17). We used the CD4 and CD8 PET imaging probes to record the in vivo distribution of CD4⁺ and CD8⁺ T cells in wild-type mice, in tumor-bearing animals, and in RAG-deficient mice reconstituted with monoclonal populations of CD4⁺ and CD8⁺ T cells. The approach described in this study for conversion of an intact Ig into a PEGylated scFv should be readily applicable to any clinically useful Ig. It should thus be possible to turn therapeutically useful Abs into scFv-based imaging agents as a means of tracking the distribution of the targets recognized.

Animal procedures were approved by the Boston Children’s Hospital Institutional Animal Care and Use Committee (protocol 19-12-4075R). Female C57BL/6J mice (000664), RAG1-knockout mice (RAG1^−/−; 002216), OT-I TCR transgenic mice (003831) and OT-II TCR transgenic mice (004194) were purchased from The Jackson Laboratory. All mice were housed in a specific pathogen–free environment and used at 6–12 wk of age.

Cells from mouse spleen and lymph nodes were isolated by mashing organs through a 40-µm strainer (Corning) in RPMI 1640 supplemented with 10% heat-inactivated FBS. RBCs were lysed using an ammonium chloride solution (STEMCELL Technologies). Cells were washed and resuspended in FACS buffer (PBS, 2% FBS, and 1 mM EDTA), followed by staining for 30 min at 4°C. The stained cells were then washed and analyzed using a BD LSRFortessa, followed by data analysis using FlowJo v10 software (Tree Star). Fluorescein-labeled anti-CD4 scFv (GK 1.5), prepared as described below, and the following fluorescent dye-conjugated Abs were used for staining: anti-CD3 (145-2C11), anti-CD4 (RM4-5), anti-CD8α (53-6.7), anti-CD19 (eBio1D3), and anti-CD45.1 (A20). Cell viability was measured by an Aqua Dead Cell Stain Kit (Thermo Fisher Scientific).

The published V_H and V_L sequences encoding mouse anti-CD4 (clone GK 1.5) were assembled in the following orientation (18): V_H, a 15-aa linker (G₄S)₃ and V_L, a C-terminal sortase recognition motif (LPETG) and a six histidine tag (Fig. 1A). To clone the anti-CD4 scFv into the pHEN6 expression vector, primers with overlapping complimentary sequences at the 3′ and 5′ ends were used (Table I). The purified anti-CD4 scFv PCR product was cloned into the linearized pHEN6 expression vector using a Gibson Assembly Cloning Kit according to the manufacturer’s instructions (New England Biolabs). Recombinant clones were transformed into competent DH5α Escherichia coli (New England Biolabs). Single colonies were used for isolation of plasmid DNA (Omega Bio-Tek) and GENEWIZ Sanger sequencing. A clone of the correct sequence was transformed into competent WK6 E. coli for recombinant expression of anti-CD4 scFv. WK6 E. coli were then grown in Terrific Broth containing ampicillin (100 µg/ml) at 37°C; when the OD₆₀₀ reached 0.8–0.9, 1 mM IPTG was added, and the cells continued to grow at 30°C overnight. Approximately 20 h later, cells were harvested. The anti-CD4 scFv was released by osmotic shock using TES buffer (24.2 g/l Tris, 0.19 g/l EDTA, and 171.15 g/l sucrose [pH 7.8]) and purified on Ni-NTA beads (QIAGEN), followed by concentrating the scFv in PBS using a 10-kDa cutoff centrifugal filter (Millipore). Size and mass of purified anti-CD4 scFv were confirmed by SDS-PAGE and liquid chromatography– mass spectrometry (LC-MS).

(Gly)₃-Lys-fluorescein and (Gly)₃-Lys-deferoxamine (DFO)-azide, to be used as nucleophiles in sortase reactions, were synthesized by standard solid-phase peptide synthesis as previously described (19). Pentamutant sortase A was used to site specifically modify the C terminus of the anti-CD4 scFv (Fig. 1D). A typical reaction mixture contained anti-CD4 scFv (2 mg/ml), 600 µM nucleophile, 20 µM sortase A, and 10 mM CaCl₂. Reactions were performed for 2 h at 20°C, typically yielding >80% of labeled construct. Residual unreacted scFv, sortase substrate, and excess nucleophile were removed by depletion on Ni-NTA beads, followed by a desalting step on a PD-10 Desalting Column (GE Healthcare) eluted in PBS.

Prior to radiolabeling, DFO-azide–conjugated anti-CD4 scFv was PEGylated to increase circulatory half-life (16). A 5-fold molar excess of 10-kDa or 20-kDa PEG-dibenzylcyclooctyne (Click Chemistry Tools) was added to a solution containing 0.5–1.0 mg of anti-CD4 scFv-DFO-azide that had been chelexed (Sigma-Aldrich) to remove divalent cations. The reaction mixture was incubated while agitating for 20–24 h at 7°C to allow the click reaction to proceed, typically yielding >90% of the PEGylated product. The size of the PEGylated product was confirmed by SDS-PAGE. For radiolabeling, a ⁸⁹Zr⁴⁺ stock solution in 1 M oxalic acid corresponding to 3–5 mCi was adjusted to a pH of 6.8–7.5 with 2 M Na₂CO₃. A volume of ⁸⁹Zr⁴⁺ stock solution containing 1.0–1.5 mCi of radioactivity was added to the chelexed PEGylated anti-CD4 scFv-DFO. The reaction mixture was incubated for 1 h at room temperature, followed by desalting on a PD-10 Desalting Column eluted in PBS. The decay-corrected radiochemical yield of the radiolabeled construct was >80% (0.8–1.3 mCi).

A 20-kDa PEG-modified anti-CD8 variable H chain–only Ab fragment (VHH) was used for immuno-PET imaging of CD8⁺ T cells (16, 20). PET–computed tomography (CT) was performed following procedures described elsewhere (20). Briefly, C57BL/6J mice were anesthetized using isoflurane and injected retro-orbitally with ∼50 µCi (1850 kBq) radiolabeled anti-CD4 scFv or anti-CD8 VHH. The mice were imaged 24 h later using a G8 PET–CT small animal scanner (PerkinElmer). To validate anti-CD4 scFv with and without differently sized PEG moieties, images were acquired at each hour for the first 24 h. Image acquisition for PET and CT was done for 10 min and 1.5 min, respectively. For blocking experiments, C57BL/6J mice were injected with 10 mg/kg unlabeled anti-CD4 scFv (10 kDa PEG) 1 h prior to the injection of ⁸⁹Zr-labeled anti-CD4 scFv (10 kDa PEG). Images were processed using the manufacturer’s image reconstruction software. For further analysis and quantification, data were imported into VivoQuant software (Invicro). PET signal values were expressed as the percentage of injected dose of radioactivity per gram. Regions of interest were set using CT scan overlays.

Singe-cell suspensions of CD4⁺ T cells (from OT-II donor mice) and CD8⁺ T cells (from OT-I donor mice) were prepared from spleen and lymph nodes using the naive CD4⁺ T Cell Isolation Kit and the CD8a⁺ T Cell Isolation Kit, respectively, according to the manufacturer’s instructions (Miltenyi Biotec). The purity of the isolated cells was >95% as determined by cytofluorimetry. Pooled cells were washed and resuspended in PBS, followed by retro-orbital injection of 2 × 10⁶ CD4⁺ and/or CD8a⁺ cells into RAG1^−/− recipients. On day 1 after cell transfer, mice received 100 µg of chicken OVA (Sigma-Aldrich) emulsified in CFA (1:1 ratio), by s.c. injection in the scruff of the neck. Immunized mice received 50 µCi (1850 kBq) radiolabeled anti-CD4 scFv or anti-CD8 VHH by retro-orbital injection on days 4 and 11. PET–CT images were then acquired on days 5 and 12 (Fig. 3A).

Tumor challenge studies were performed with the B16 melanoma (American Type Culture Collection) and B16-OVA melanoma murine cells. B16 melanoma cells were maintained in DMEM supplemented with 10% heat-inactivated FBS and 100 U/ml penicillin/streptomycin. B16-OVA melanoma cells were cultured in RPMI 1640 supplemented with 10% heat-inactivated FBS, 100 U/ml penicillin/streptomycin, 2 mM/l glutamine, MEM nonessential amino acids, 1 mM/l sodium pyruvate, and 50 mM/l 2-ME. Female C57BL/6J mice were inoculated s.c. with 5 × 10⁵ B16 melanoma cells in the right flank, resulting in palpable tumors by approximately day 8. The size of the tumor was measured every 3 d with a caliper. To track CD4⁺ and CD8⁺ T cells, tumor-bearing mice were injected retro-orbitally with ∼50 µCi (1850 kBq) radiolabeled anti-CD4 scFv or anti-CD8 VHH on days 2, 5, 9, and 12, followed by acquisition of PET–CT images the day after injection of the radiotracer (Fig. 4A). For the B16-OVA melanoma model, RAG1^−/− mice were inoculated with 5 × 10⁵ cells on day 0, followed by retro-orbital transfer of 2 × 10⁶ CD4⁺ (OT-II) and/or CD8a⁺ (OT-I) cells on day 1. Mice were primed on day 2 with 100 µg of OVA, followed by tracking of immune cells by immuno-PET on days 5 and 12 (Fig. 5A).

Our strategy was centered on the use of an anti-CD4 scFv as a monovalent imaging agent. Monovalency should not only reduce the possibility of cross-linking the target molecule on the cell surface but also yield an agent smaller in size than the diabodies currently in use. Compared with an intact anti-CD4 Ig, the smaller size of the scFv ought to accelerate clearance from the circulation and possibly improve tissue penetration, both desirable traits for an imaging agent (21). Accordingly, we constructed the anti-CD4 scFv from the published V_H and V_L sequences for mouse anti-CD4 (clone GK 1.5) (Fig. 1A) (18) (Table I). The recombinantly expressed scFv was purified with a yield of 2–3 mg/l of E. coli culture. LC-MS and SDS-PAGE of the purified product confirmed the expected molecular mass of ∼27 kDa (Fig. 1B, 1C).

We avoided N-hydroxysuccinimide– or maleimide-based chemical modification strategies to preclude unwanted side reactions (22, 23). Instead, we opted for site-specific enzymatic installation of the substituent of choice. For this purpose, we have extensively used sortase as a chemo-enzymatic labeling tool (14, 24). Sortase recognizes an LPXTG motif and cleaves the C-terminal Thr residue with concomitant formation of a thioester enzyme substrate intermediate. This is then resolved by addition of a nucleophile, commonly a peptide with an N-terminal Gly residue, to which any payload of interest may be attached. We cloned the anti-CD4 scFv such that it carried a C-terminal extension with a sortase recognition sequence, followed by a (His)₆ tag to enable metal chelate-based affinity purification (Fig. 1D). This scFv expressed well, was readily purified, then was labeled with fluorophores or with other modifications of choice. Because sortase and unreacted scFv contain a His tag, adsorption of the reaction mixture on Ni-NTA agarose separates the desired (unbound) product from the input material. A simple size exclusion chromatographic step then separates the desired modified anti-CD4 scFv from the lower m.w. nucleophile used in excess in the reaction. Anti-CD4 scFv was site-specifically labeled with fluorescein using sortase (Fig. 1D). SDS-PAGE confirmed the fluorescein labeling of scFv (Fig. 1E). Using unfractionated splenocytes as target, fluorescently labeled scFv decorated CD4⁺ T cells as measured by cytofluorimetry (Fig. 1F). The scFv competed efficiently for binding with the RM4-5 mAb, confirming that this scFv construct retained its proper specificity when labeled at its C terminus (Fig. 1F).

To install PET isotopes, we used the linkers diagrammed in (Fig. 2A as nucleophiles for the sortase reaction, with DFO as the chelator for installation of ⁸⁹Zr and PEG spacers of 10 and 20 kDa where applicable. We produced three versions of the labeled scFv, one with no PEG, one modified with 10-kDa PEG, and one modified with 20-kDa PEG. Installation of PEG-bearing linkers proceeded efficiently, with only minimal amounts of unmodified scFv persisting in the final product (Fig. 2B) and unlikely to compete with PEGylated scFvs for binding. The imaging agents were administered i.v. via the retro-orbital plexus, and images were acquired hourly for the first 24 h. Representative images for 1 h, 2 h, 4 h, 8 h, 16 h, and 24 h are shown in (Fig. 2C; videos showing three-dimensional (3D) rendering are given in Supplemental Video 1.

The scFv in its non-PEGylated form was detected in the circulation shortly after injection, disappeared from the bloodstream, and thereafter showed accumulation mostly in liver and kidneys, where it persisted the longest. It was difficult to visualize any lymphoid organs (Fig. 2C1–6). As expected, the versions that carried the 10-kDa (Fig. 2C7–12) and 20-kDa PEG (Fig. 2C13–18) moieties showed increased persistence in the circulation after injection, particularly evident for the 20-kDa PEGylated scFv, with the 20-kDa PEG substituent yielding the longer circulatory half-life. PET imaging at later time points showed progressive loss of label from the circulation. However, at 8 h and 20 h postinjection, both the 10-kDa and 20-kDa PEGylated scFvs showed the presence of label in secondary lymphoid organs, with an improved PET signal-to-noise ratio. Prior administration of unlabeled anti-CD4 scFv (10-kDa PEG) completely blocked accumulation of imaging agent in the secondary lymphoid organs (Fig. 2C20). Ex vivo biodistribution of anti-CD4 scFv (10-kDa PEG) at 24 h after injection in wild-type C57BL/6J mice, CD4-blocked C57BL/6J mice, and RAG1^−/− mice showed the presence of label in secondary lymphoid organs in wild-type mice only, demonstrating CD4 specificity in vivo (Fig. 2C19–21, 2D). We determined that only minimal amounts of label persisted in the kidneys 1 wk after administration of the labeled anti-CD4 scFv, with no detectable label in the secondary lymphoid organs (Supplemental Video 2). The generation of the corresponding VHH-based PEGylated CD8 imaging agent, which we used for comparison, has been described (16), and visualizes CD8⁺ T cells at the expected locations. In terms of PET image quality, the anti-CD8 VHH outperformed the anti-CD4 scFv under all conditions tested (data not shown).

We used TCR transgenic mice as a source of T cells of defined Ag specificity. Splenocytes from OVA-specific OT-I and OT-II TCR transgenic mice were purified to obtain pure populations of CD8⁺ and CD4⁺ T cells, respectively. These were transferred into RAG1^−/− recipients, which lack all B and T cells, so that it would allow us to examine the dynamics of T cell engraftment after adoptive transfer therapy in the absence of a starting, resident population of host CD4⁺ and CD8⁺ T cells (25, 26). Therefore, the observed PET signal would necessarily derive from the transferred cells. Recipient mice were then imaged by immuno-PET. Little or no specific PET signal was detected at 5 d and 12 d posttransfer (Fig. 3B1–4). Only when mice were immunized with OVA in CFA did both OT-I and OT-II T cells persist in secondary lymphoid organs, as visualized at 5 d postimmunization (Fig. 3B5–8). In the absence of immunization, cotransfer of OT-I and OT-II T cells showed no improvement in persistence of either CD4⁺ or CD8⁺ T cells (Fig. 3B9–12). In mice that were immunized with OVA in CFA, persistence of both CD4⁺ and CD8⁺ T cells improved (Fig. 3B13–16). For CD8⁺ T cells in particular, cotransfer of CD4⁺ T cells resulted in appreciable accumulation of CD8⁺ T cells at day 12 at the site of injection, presumably because of local Ag deposits (Fig. 3B8, 16). Little or no such accumulation was seen for CD4⁺ T cells at day 12, which we attribute to the lesser sensitivity of the CD4 imaging agent (Fig. 3B6, 14). Representative videos of 3D renderings are given in Supplemental Video 3.

We injected C57BL/6J mice with 5 × 10⁵ cells of the highly aggressive B16 melanoma line in the right flank. Tumors became palpable at ∼8 d after inoculation. We imaged CD4⁺ and CD8⁺ T cells at days 3, 6, 10, and 13 postinoculation. At the day 3 and 6 time points, the distribution of CD4⁺ and CD8⁺ T cells was comparable with that seen in control animals, with no observable signs of lymphocyte localization to the anticipated sites of the tumor (site of injection) (Fig. 4B1–2, 5–6). The images acquired at 10 d showed clear signs of the tumor by CT and evidence of lymphocyte infiltration by both CD4⁺ and CD8⁺ T cells, representative of a polyclonal response that nonetheless fails to control the tumor (Fig. 4B3, 7). At the day 13 time point, the size of the tumor increased, with enhanced infiltration of both CD4⁺ and CD8⁺ T cells (Fig. 4B4, 8). However, infiltration was less evident for CD4⁺ T cells than for CD8⁺ T cells. Distribution of the CD4⁺ and CD8⁺ T cells over the tumor was not homogeneous and presented as a somewhat patchy image (Fig. 4B, 4C). Representative videos of 3D renderings are given in Supplemental Video 4A.

We next inoculated RAG1^−/− mice with B16-OVA, a variant of the B16 melanoma line expressing an OVA fragment that contains the epitopes recognized by OT-I and OT-II TCR transgenic mice. These mice then received either OT-I T cells, OT-II T cells, or both. Activation of the transferred T cells was accomplished by immunization posttransfer with OVA in CFA at the time points indicated (Fig 5A). At the day 5 time point, the tumor was not palpable, and little or no specific PET signal was detected in mice without immunization (Fig. 5B1–4, 9–12). In contrast, immunization of mice increased the number of CD4⁺ and CD8⁺ T cells at day 5 in secondary lymphoid organs and was followed by increased infiltration of T cells into the tumors at 12 d (Fig. 5B5–8). Cotransfer of OT-I and OT-II as well as immunization expanded the numbers of T cells in both secondary lymphoid organs and in the tumors (Fig. 5B13–16). Representative videos of 3D rendering are shown in Supplemental Video 4B.

The primary goal of this work was to show that established mAbs can be engineered to yield monovalent scFv fragments that are suitable as a PET imaging agent. PET is a noninvasive imaging technique that allows longitudinal monitoring of whole body distribution of the imaging agent at millimeter resolution (2, 27). Clinical use in PET and single-photon emission CT of such smaller Ig-derived formats has been reported for both diabodies (molecular mass ∼50 kDa) and nanobodies (molecular mass ∼15 kDa) (28–30). Diabodies have the inherent advantage of bivalency, but outside of imaging applications, they have not been as widely deployed as other Ig formats (1, 17, 31). Nanobodies labeled with ¹⁸F have yielded acceptable results in preclinical models (32, 33), but the range of available specificities is less extensive than that for conventional Igs. Conversion of existing mAbs into scFvs using well-established methods could thus massively expand the repertoire of imaging agents. In our study, the PEGylated ⁸⁹Zr-labeled anti-CD4 scFv demonstrated specific targeting of both endogenous and tumor-infiltrating CD4⁺ T cells in vivo, as shown by immuno-PET. Different (radio) labeling strategies, such as the choice of metal chelator, radioisotope, conjugation chemistry, and peptide linkers can alter the pharmacokinetics of radiolabeled Ab fragments (31, 34). Site-specific conjugation using sortase A at the C terminus of scFv avoids the complications associated with thiol-specific radiolabeling strategies. The modified scFv retains specificity for its target Ag. PET could suffer from the partial volume effect in detecting activity in small regions of interest, such as lymph nodes, because of the resolution of the scanner (35). Even with this limitation, we have demonstrated that, by immuno-PET, the anti-CD4 scFv can detect secondary lymphoid organs in mice, including in a melanoma tumor model.

We were able to noninvasively track expansion of transferred CD4⁺ and CD8⁺ T cells in an immunodeficient mouse model. Perhaps not surprisingly, prior immunization with the relevant Ag drastically improved the intensity of the observed signal attributable to the expansion of these Ag-specific lymphocytes. Imaging of T cells by direct radiolabeling ex vivo often complicates their detection because of the half-life of the radionuclide used, dilution of probe, and radiotoxicity (36, 37). Use of a reporter gene could circumvent complications associated with direct ex vivo radiolabeling of cells (38, 39) but requires the obvious genetic manipulations and may likewise suffer from a high background signal in organs of elimination (40, 41). Most commonly used metabolic probes for PET imaging, such as [¹⁸F]FDG and [¹⁸F]FAC lack specificity for the detection of infiltrating immune cells (42).

We also examined the distribution of CD4⁺ and CD8⁺ T cells in the setting of an antitumor response by using the B16 melanoma and its OVA-expressing variant as examples. Immuno-PET can detect infiltrating lymphocytes once the tumor becomes palpable but apparently not before. We have previously discussed the sensitivity of immuno-PET; whereas secondary lymphoid organs contain lymphocytes in numbers and at densities that readily allow visualization by immuno-PET, more diffuse tissue distributions remain a challenge (16, 43) for immuno-PET. Unlike immunohistochemistry, which affords resolution at the single-cell and even subcellular level, PET does not. Moreover, positrons do not come in different “colors” or energy levels that would allow the equivalent of multispectral imaging. Notwithstanding these limitations, immuno-PET is currently the only noninvasive method that can visualize the distribution of cell types, such as lymphocytes and other immune cells, in a living mouse without the need for any genetic modifications of the target cell population. The availability of the appropriate imaging agent can reveal aspects of the distribution of the target(s) recognized that might have escaped detection by other means. The expression of PD-L1 on brown adipose tissue detected by PET is a case in point (44). Earlier studies have shown the potential use of imaging the distribution of lymphocytes by immuno-PET as a possible biomarker for a response to immunotherapy (1, 16, 43, 45). Even identification by immuno-PET of Ag-specific CD8⁺ T cells is now a possibility (20).

These results are of practical interest for two main reasons. First, for Abs in clinical use, accurate information for the distribution of the target Ags is not always easy to come by and mostly relies on invasive methods, such as immunohistochemistry, on samples obtained by biopsy or surgical resection. The dynamics of presence and tissue distribution of such Ags can provide important and perhaps actionable data to guide therapy, such as in the case of checkpoint-blocking Abs in the treatment of cancer (1, 16, 43). By converting clinically useful Abs into scFvs and appropriately modifying them for immune-PET, such information could then be obtained noninvasively.

Second, nanobodies, as the smallest Ig fragments that retain Ag-binding properties, are superior as building blocks for PET imaging agents (30, 46), but the generation of such reagents can be time-consuming, might require humanization, and may not always be possible for Ags of interest. To date, no anti-mouse CD4 nanobodies have been reported. To be able to draw on the rich collection of existing mAbs as a possible source of imaging agents may prove to be a powerful alternative, especially in preclinical models and across different species.

We thank Stephen C. Kolifrath for help with the animal experiments and Nicholas McCaul, Justin Lievense, Ryan Alexander, and Thomas Balligand for support and helpful discussions.

The authors have no financial conflict of interest.