DTPA-Receptor – A novel reporter gene system for the specific and sensitive PET imaging of CAR-T cells and AAV transduced cells

doi:10.21203/rs.3.rs-3200226/v1

Download PDF

Article

DTPA-Receptor – A novel reporter gene system for the specific and sensitive PET imaging of CAR-T cells and AAV transduced cells

https://doi.org/10.21203/rs.3.rs-3200226/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Advanced Therapy Medicinal Products (ATMPs), such as cell and gene therapies, necessitate a reliable diagnostic method for quantitative monitoring. We developed a novel reporter gene system for PET imaging consisting of a membrane-anchored Anticalin protein (DTPA-R) that acts as a high-affinity receptor for the radioligand [¹⁸F]F-DTPA•Tb^III. The reporter protein shows high cell surface expression of up to ~1×10⁶ receptors per cell. After systemic administration, the pharmacologically inert radioligand rapidly clears via the renal route and, at t=90 min, generates a strong signal of 22.1 %ID/g for DTPA-R-expressing PC3 cells compared to 0.2 %ID/g for DTPA-R-negative controls (ratio: 125). The detection limit for Jurkat^DTPA-Rcellswas 500 cells in a PET phantom ex vivo and 8,000 if subcutaneously injected. In vivo expansion and migration of CD19-CAR-T^DTPA-R cells was successfully monitored over four weeks with a linear relationship between PET signal and CAR-T cell number. Furthermore, our reporter system allowed quantitative and longitudinal imaging of AAV9 viral vectors with a linear dose-to-signal relation. In summary, DTPA-R shows high potential for in vivo monitoring of ATMP-based therapies.

Biological sciences/Biological techniques/Imaging/Positron-emission tomography

Health sciences/Health care/Medical imaging/Radionuclide imaging

Health sciences/Medical research/Drug development

Biological sciences/Biological techniques/Molecular engineering/Synthetic biology

Health sciences/Medical research/Preclinical research

reporter gene imaging

ATMP

synthetic biomarker

positron emission tomography (PET)

cell tracking

Recent regulatory approvals of cell and gene therapies^1,2 have shown that such advanced therapy medicinal products (ATMPs) can be used in clinical practice and have the potential to cure life-threatening diseases. Despite these exciting breakthroughs, the development of new cell and virus based ATMPs remains hampered by limited knowledge about the biodistribution and persistence of these live drugs³. Technologies to monitor the distribution of Chimeric Antigen Receptor (CAR)-T cells in vivo non-invasively and quantitatively could greatly improve our understanding of their trafficking, therapeutic efficacy, and off-target toxicity⁴. Likewise, Adeno-associated virus (AAV) based gene therapies would benefit from quantifying the location, magnitude, and duration of transgene expression.

Direct ex vivo labeling of cells (e.g., with iron oxide particles) is a sensitive approach for studying the distribution of transplanted cells early after injection. Although sensitive, this approach is limited at later time points because the label is diluted by each cell division, which prevents the assessment of cell proliferation. Furthermore, the label persists even after the cells have died or were phagocytosed. This limits the utility of direct labeling for CAR-T cell tracking, where a small fraction of the graft massively expands following activation.⁵

As an alternative, cell graft proliferation and viability can be imaged by an indirect labeling strategy. This approach relies either on endogenous biomarkers or reporter genes (synthetic biomarkers) that can be detected by imaging⁶. Various luciferase enzymes and fluorescent proteins have been used as reporter genes to study transplanted cells in mice by means of optical imaging techniques⁷. However, light is strongly attenuated and scattered by biological tissues, which limits imaging depth and quantification in optical imaging⁸, particularly in larger animals or humans⁷.

A viable modality for reporter gene imaging should allow for whole body-imaging, be highly sensitive and provide a quantitative, depth-independent signal⁶. Furthermore, imaging should be cross-sectional to allow for precise anatomical localization. PET (positron emission tomography) fulfills all of these requirements, thus enabling the spatiotemporal imaging of transgenes on a whole-body scale while being widely available for preclinical and clinical imaging⁴. The sensitivity of current clinical scanners lies in the picomolar range⁵ and has been further increased by a factor of 40 with the recent introduction of total body PET scanners⁹.

The ideal reporter gene/protein should be small, bio-orthogonal – i.e., not expressed endogenously and not affect cell function – and non-immunogenic. For PET imaging, a straightforward radiolabeling procedure with a commonly available radioisotope such as fluorine-18 would be desirable. The small molecule probe should show low non-specific binding to cells, undergo rapid kidney clearance, but remain in the circulation sufficiently long to ensure quantitative labeling of cells expressing the reporter gene. Notably, despite numerous efforts over three decades¹⁰, no reporter gene system fulfills these requirements in a clinical setting. Herpes simplex thymidine kinase (HSV-tk)¹¹ has been extensively studied as a reporter gene, but this viral protein is highly immunogenic with a high seroprevalence. On the other hand, systems like the sodium/iodide symporter (NIS)¹², somatostatin receptor 2 (SSTR2)¹³, and truncated prostate-specific membrane antigen (tPSMA)^{14, 15} are not immunogenic but are expressed endogenously by various tissues which hampers their application for the specific whole-body imaging of grafted cells.

To address this unmet medical need, we developed and characterized a new reporter gene and radioligand system that combines high sensitivity with exquisite specific. Our approach is based on an Anticalin^{16, 17} that is expressed on the cell surface and binds a bio-orthogonal ¹⁸F-labeled probe. Anticalins are engineered binding proteins based on the lipocalin protein family¹⁸, endogenous plasma proteins comprising a single polypeptide with a robust β-barrel fold¹⁹. We have extensively studied this new reporter system in vivo and demonstrated its suitability to monitor CAR-T cells and AAV-mediated gene transfer in mice.

Design of the reporter protein and radioligand

The PET reporter protein has been designed as an artificial cell surface protein exhibiting a highly specific binding site for a small molecule ligand (Fig. 1A). The protein construct comprises the lipocalin-2 signal peptide, an engineered lipocalin binding protein (Anticalin¹⁸), a peptide linker containing the V5-tag (a multifunctional molecular handle for staining of ATMPs)²⁰ and the α-helical human CD4 transmembrane domain. This fusion protein contains only 257 amino acid residues and is encoded in a 774 base pairs (bp) open reading frame (ORF; for sequences see the Supplementary Information). Finally, for multimodal detection, an optional fluorescent protein was added to the C-terminal cytoplasmic end, such as mRuby3²¹ or miRFP720²² (Fig. 1B). Two Anticalins were chosen to construct different reporter proteins with two ligand specificities: (i) the Anticalin CL31d (DTPAR), which binds CHX-A’’-DTPA•metal complexes with a K_D of ~ 500 pM^{16, 23} and the Anticalin D6.4(Q77E) (ColchiR) which binds colchicine with a K_D of ~ 20 pM²⁴. The cognate reporter probe features CHX-A’’-DTPA•metal (Fig. 1C) or colchicine (Fig. 1D), respectively, a PEG₄-linker to facilitate access to the Anticalin binding pocket without steric hindrance, optionally hydrophilic groups to improve pharmacokinetics and a labeling group for efficient incorporation of the ¹⁸F isotope to enable PET detection (Fig. 1C).

Analysis of reporter protein variants

To compare the two Anticalin-ligand pairs, we generated cell lines expressing the respective reporter genes by retroviral transduction of the human T cell line Jurkat, human embryonic kidney cells (HEK293T), and the prostate carcinoma cell line PC3. Initially, cell surface expression of the DTPA-R or ColchiR reporter proteins on PC3 cells was confirmed by fluorescence microscopy using an anti-V5-tag antibody (Fig. 1E). Membrane localization of mRuby3 in DTPARmRuby3 expressing HEK293T cells indicated efficient transport of the reporter protein through the secretory pathway (Fig. S1). Absolute reporter protein numbers per cell were measured by flow cytometry using an anti-V5-tag antibody and MESF calibration beads. Expression levels on Jurkat cells ranged from ~ 16,000 for DTPA-R-miRFP720 to around 1 million receptors for DTPA-R (Fig. 1F). Omitting the optional fluorescent protein increased expression, e.g., DTPA-R showed 8.3-fold (p < 0.0001) higher expression than DTPA-R-mRuby3 (Fig. 1F). Next, we studied the influence of T cell activation on reporter gene expression by comparing PMA/Ionomycin-activated to untreated Jurkat cells, which did not result in significant differences (Fig. 1F & S3A). Furthermore, the type of Anticalin influenced the expression levels: a 4.9-fold (p < 0.0001) higher expression of DTPA-R was measured compared to Colchi-R (Fig. 1F, S2 & S3A). We also compared the expression level of DTPA-R with a previously described reporter gene that uses the murine single chain variable fragment (scFv) C825²⁵, or its humanized version huC825²⁶, to bind DOTA-metal complexes (Fig. S3A). Using the same expression cassette and membrane anchor domain for these scFvs, DTPAR showed a 5.8 (muC825) or 8.9 (huC825) fold higher median V5-AF488 signal after retroviral transduction. The number of genomically inserted expression cassettes per cell was quantified by ddPCR and showed similar levels for the different constructs (Fig. S3B). However, transgene mRNA levels varied between constructs (Fig. S3C). While protein levels of Anticalin-based reporters correlated with mRNA levels (R² = 0.98), scFv-based reporter genes produced high mRNA levels, which did not translate into elevated protein levels, indicating superior protein folding, stability, and cell-surface presentation for Anticalin-based reporter proteins (Fig. S3D).

A potential metabolic burden of the reporter gene expression was assessed by measuring doubling times of non-activated or PMA/Ionomycin-activated Jurkat cell lines using a carboxyfluorescein succinimidyl ester (CFSE) based proliferation assay (Fig. 1G). The doubling time of wild-type Jurkat cells (non-activated: 18.3 ± 4.3 h / activated: 22.2 ± 4.7 h) was not substantially changed by reporter gene expression (DTPA-R: 17.0 ± 4.1 h / 20.6 ± 4.5 h).

Further applications of the V5-tag as part of the DTPA-R include the isolation of reporter gene expressing cells using magnetic-activated cell sorting. Employing the V5-tag, Jurkat^DTPA−R cells could be isolated by magnetic-activated cell sorting (MACS)²⁷ from a 5:95 cell mixture with high efficacy and purity (> 98%) (Fig. 1H). Finally, the V5-tag also enabled the detection of the reporter protein at the cellular level via immunohistochemistry (IHC) (Fig. S4).

Synthesis & characterization of [¹⁸F]F-colchicine and [¹⁸F]F-DTPA

First, we investigated the binding of NH₂-CHX-A’’-DTPA in complex with ⁹⁰Y^III to Jurkat^DTPA−R cells using a competitive assay, resulting in a half-maximal inhibitory concentration (IC₅₀) of 1.4 nM (Fig. 2A). However, fluorine-18 is the preferred PET radioisotope as it is widely available and almost every decay leads to the emission of a positron (ratio 96.9%) with desirably low energy (E_mean=250 keV, average positron range of 0.6 mm²⁸). On the other hand, positron-emitting radiometals have inferior physical characteristics, and their NH₂-CHX-A’’-DTPA complexes often bind with lower affinity to DTPA-R than Y^III or Tb^III complexes^{16, 23}. For this reason, we designed a DTPA-based radioligand that can be charged with radioactive or non-radioactive metal ions and features a second group that can be easily radiolabeled with fluorine-18 (Fig. 2B & S6). This radiolabeling approach uses a trimethylamine leaving group²⁹, similar to the approved radioligand [¹⁸F]PSMA-1007³⁰. For comparison, we synthesized analogous radioligand precursors for Colchi-R employing different numbers of strongly polar Glu residues and a precursor with two Glu residues for DTPA-R (Fig. S5). After quality control (Fig. S6E), both compounds were radio-fluorinated in a single step and purified by reversed phase HPLC to obtain highly pure radioligands (Fig. 2C & S7A). For [¹⁸F]F-Nic-D-Glu₂-PEG₄-CHX-A’’-DTPA•Tb^III (dubbed [¹⁸F]F-DTPA) a radiochemical yield of ~ 20% and a radiochemical purity of > 98% were achieved. Subsequently, the ¹⁸F-labeled radioligands were tested in ligand binding assays with DTPA-R and Colchi-R expressing PC3 cells (Fig. 2DE & S7B). The binding of both radioligands was highly specific to cells expressing the corresponding receptor (~ 1,000-fold for DTPA-R and ~ 500-fold for Colchi-R) and could be efficiently blocked by competition with non-radioactive ligand (~ 100-fold for DTPA-R and ~ 400-fold for Colchi-R).

Dynamic PET imaging of mice bearing xenograft tumors

As a proof-of-principle for reporter gene imaging, the respective radioligands were injected into CD1-nude mice carrying subcutaneous PC3^{DTPA − R} and PC3^{Colchi − R} xenograft tumors above the right and left shoulder, respectively (Fig. 3AB & S7C). Dynamic [¹⁸F]F-DTPA PET imaging showed fast, almost exclusively renal clearance, accompanied by increasing signals in the kidneys, ureter, and urinary bladder, with no retention in other tissues (Fig. 3A). The radioligand showed rapid accumulation in the PC3^{DTPA − R} xenograft tumor, resulting in a stable activity concentration of 22.8±0.6% Injected Dose (ID)_max/g between 30 and 90 min post injection (p.i.). While both xenograft tumors were clearly visible on MR images (Fig. 3B), only the PC3^DTPAR tumor could be delineated in PET. The maximum activity concentration in the PC3^ColchiR xenograft (0.18%ID_max/g) was 125-fold lower than for the cognate PC3^{DTPA − R} xenograft. [¹⁸F]F-DTPA in the blood pool was quickly cleared with a plasma half-life of 0.5–1.5 min, which resulted in a tumor-to-blood ratio of 261 after 90 min. Furthermore, in a second static PET scan 6 h p.i., still 48% of the initial PC3^{DTPA − R} signal was retained (Fig. 3CD). Overall, these results indicate an exquisite specificity of the DTPA-R reporter gene system in vivo.

The biodistribution of [¹⁸F]F-DTPA at 90 min p.i. in healthy female and male mice was analyzed ex vivo to study sex-specific differences (n = 5♀ + 5♂). Very low retention of below 0.2%ID/g was measured for most tissues (Fig. 3E), except for the kidneys (0.9%ID/g) and the hepato-biliary excretion route (small intestine: 0.2%ID/g; large intestine: 0.1%ID/g). No significant sex-specific differences in biodistribution were observed. (Fig. 3E).

In vivo stability of [¹⁸F]F-DTPA was analyzed after intravenous (i.v.) injection using size-exclusion chromatography (SEC) to separate intact radioligand from fragments with lower molecular weight. [¹⁸F]F-DTPA within urine samples collected from two mice ~ 3 h p.i. was > 97% intact, indicating serum stability (Fig. 3F).

For comparison, a dynamic PET study was conducted for the radioligand [¹⁸F]F-Nic-Glu₂-PEG₄-colchicine ([¹⁸F]F-colchicine). This radioligand showed not only renal (19.8%ID) but also significant hepato-biliary excretion (69.7%ID), which is undesirable due to the strong and variable background PET signals in the gastrointestinal tract (Fig. S7D). The elevated hepato-biliary excretion of [¹⁸F]F-colchicine is likely explained by its lower hydrophilicity, as reflected by the LogD_7.4 of -2.97 vs. -3.58 for [¹⁸F]F-DTPA. Taken together, these pharmacokinetic data, as well as the higher expression levels of the DTPA-R on transfected cells and the much higher radioligand uptake in the xenograft tumor, prompted us to focus on the further development of the DTPA-R reporter system.

Design and evaluation of DTPA-R labeled CAR-T cells

To investigate the application of DTPA-R in the context of human CART cell therapy, we assessed in an animal study if the migration, proliferation, and tissue infiltration of CART^DTPA−R cells can be imaged. To this end, we employed a well-characterized CD19-targeting 2nd -generation CAR based on the scFv FMC63 (Fig. S8A)³¹. The corresponding expression cassette comprised the CAR followed by a 2A self-cleaving sequence separating a second membrane protein, the truncated Epidermal Growth Factor Receptor (EGFRt)³². EGFRt lacks normal EGFR function but is still recognized by cetuximab (Erbitux) for in vivo cell ablation³³. To generate CART^DTPAR cells, we replaced the EGFRt in the retroviral expression cassette with DTPA-R, which enables PET imaging, apart from optionally serving for cell ablation (Fig. S8B). Both CAR-T vectors were used to transduce human peripheral blood mononuclear cells (PBMCs), resulting in a surface expression of approximately 100,000 DTPA-R molecules per CAR-T cell (Fig. S8C).

There was no substantial difference in CD3, CD4, and CXCR3 expression levels or CD8⁺/CD4⁺ ratios between CAR-T^EGFRt and CAR-T^DTPA−R cells. (Fig. S8CD). CAR-T cell function was assessed in a real-time assay of the CAR-mediated killing of HEK^CD19 target cells (xCELLigence; Fig. S8E) as well as a ⁵¹Cr-release assay on Raji cells (Fig. S8F). Specific lysis was comparable between the αCD19CART^EGFRt and αCD19CART^DTPAR cell treated groups in both assays.

PET imaging of CAR-T cell therapy

Both CAR-T cells were used to treat NSG mice engrafted with a systemic Raji tumor, which is known to primarily home to the bone marrow³⁴. After 7 days of Raji engraftment, 2×10⁶ αCD19-CART^DTPAR or reference αCD19-CART^EGFRt cells were infused. This experimental setting ensured constant CART expansion due to non-complete tumor eradication, which was imaged with [¹⁸F]FDTPA on days 4, 8, 15, 22, and 30 (Fig. 4A). PET signals indicated the presence of CART^DTPAR cells at day 4 in the spleen and from day 8 onwards, with increasing intensity, in the bone marrow of the extremities as well as spine, scull, and axillary lymph nodes (Fig. 4B or Fig. S9 for all animals). In contrast, PET signals were absent in mice injected with αCD19-CAR-T^EGFRt cells, demonstrating the specificity of [¹⁸F]F-DTPA (Fig. 4C).

Quantification of the cumulated specific PET signals allowed the assessment of CAR-T cell expansion over time (Fig. 4D). Furthermore, we used flow cytometry at the end of the study, or when the humane endpoint was reached, to quantify the numbers of CAR-T (Fig. 4E) and Raji cells (Fig. 4F). This ex vivo analysis showed differences in CAR-T cell numbers, confirming the high variability between animals and the need for imaging of these therapies. In addition, 3D cross-sectional images of the [¹⁸F]F-DTPA PET/MR allowed the monitoring of local CAR-T infiltration (Fig. 4G–I). Longitudinal PET analysis revealed different trajectories of CAR-T infiltrated lesions (Fig. 4HI). Ex vivo analysis by V5-IHC of sagittal spine sections allowed CAR-T^DTPA−R identification on the cellular level, which confirmed CAR-T infiltration into individual vertebra bodies (~ 1.5×0.5 mm) and, occasionally, invasion through the vertebral bone into the spinal canal (Fig. 4J). Importantly, the pattern of CAR-T infiltration determined by IHC was very well matched by the signals in PET/MR (Fig. 4K). Furthermore, the co-localization of CAR-T^DTPAR with Raji cells was confirmed by IHC co-staining of CD19 and the V5-tag (Fig. 4L). Interestingly, lesions with a CAR-T cell plateau in longitudinal PET imaging (blue and orange, Fig. 4I) showed lower tumor cell density, which indicated better tumor clearance by CART cells compared to neoplasms with constantly increasing PET signal (green, Fig. 4I & S10A). Other organs like the spleen, kidneys, liver, and lung only showed few CAR-T^DTPA−R cells in V5-IHC (Fig. S10B).

Correlation between the actual number of CAR-T^DTPA−R cells detected by flow cytometry after extraction from a hollow bone and the cumulated PET signal obtained for the same extremity (Fig. 5A & S11AB) was performed for mice sacrificed on day 8 or 15 (Fig. 5B, gating see S11C) and on day 16 or 31 (Fig. S11D). A linear relationship (R² = 0.92) indicated that PET imaging of DTPA-R can quantitatively monitor tissue-specific CAR-T cell infiltration. Utilizing the resulting regression line, we calculated a detection limit of around 1,200 CART^DTPAR cells in delineated lesions in the bone (Fig. 5C). To further characterize the relationship between PET signal and the number of DTPA-R expressing cells, we measured the signal of defined numbers of Jurkat^DTPAR cells in a PET phantom in vitro. For cells labeled with [¹⁸F]F-DTPA, we observed a linear correlation (R² = 0.999) with a detection limit of 500 cells (Fig. 5D). In the next step, we determined the detection limit in tissue with low perfusion. Therefore, we injected 4,000 to 64,000 Jurkat^DTPA−R cells, or CART^DTPAR cells, in the dorsal subcutis, immediately followed by i.v. injection of [¹⁸F]F-DTPA (Fig. 5EF). Here, 8,000 cells were clearly detectable by PET. Again, a linear correlation between cell number and PET signal was observed (R² = 0.96 and R² = 0.99, respectively; see Fig. 5EF).

PET imaging of gene transfer mediated by AAV9 viral vectors

In vivo gene therapy represents another relevant application scenario for reporter gene imaging, where PET imaging can help assess the roles of administration routes, dosing regimens, pharmacokinetics of capsid-modified vectors, or endurance of gene expression. As a common example for gene therapy, we have selected vectors based on the Adeno-associated virus (AAV)³⁵. In particular, we have focused on the serotype AAV9, which sparked great interest for transducing muscle and neuronal tissue³⁶ and provided the basis for the FDA-approved drug Zolgemsma³⁵. We constructed a transfer plasmid flanked by AAV2 inverted terminal repeats (ITR) and used it to produce AAV2/9 viral vectors encoding the DTPA-R reporter gene under the control of the CMV promoter (Fig. 6A). First, we investigated the reporter gene expression in mice after systemic injection of AAV9^{DTPA − R}. Immunocompetent C57BL/6 (Fig. 6B) and immunocompromised CD1-nude (Fig. 6C) mice (n = 3 per group) were dosed with 2.5×10¹² viral genomes (vg) per mouse and, on days 14, 21 and 28 [¹⁸F]F-DTPA PET scans were recorded (Fig. 6BC). Although cohorts were treated in parallel using the same virus batch, there were marked differences between the transduction patterns obtained in C57BL/6 and CD1-nude mice. Isocontour segmentation of the heart yielded a value of 15.3 ± 3.8%ID_max/g for C57BL/6 and 28.7 ± 5.4%ID_max/g for CD1-nude mice. Within each cohort, animals showed comparable transduction patterns (Fig. S12). Moreover, we observed consistent trajectories of the repeated PET measurements, indicating stability of the vector expression over time (Fig. 6D).

We then i.v. injected four different titers, ranging from 1×10¹¹ to 2.5×10¹² vg/mouse, into C57BL/6 mice (n = 3 per group). After 7 days, [¹⁸F]F-DTPA PET/MR scans were recorded for all cohorts and an untreated control cohort (Fig. 6E or Fig. S13 for all animals). The transduction patterns were similar to the previous experiment; for example, mice with the highest applied titer showed clear PET signals in dorsal brown adipose tissue (BAT; %ID_max: 7.7 ± 1.5), heart muscle (%ID_max: 6.1 ± 3.6), liver (%ID_max: 11.3 ± 1.4) and adrenals (%ID_max: 29.5 ± 7.0), whereas the background uptake in the upper body of the control mice was ~ 0.5 ± 0.06%ID_max/g. Of note, the high resolution in axial planes provided by the combination of PET with an isomorph MRT allowed the precise analysis of transduction events (Fig. 6F) far beyond previous results obtained by bioluminescence imaging³⁶.

Lower titers resulted in gradually reduced PET signals within the respective tissues. Importantly, PET signals (%ID_max/g) showed a linear correlation with the administered dose in BAT (R² = 0.73), heart muscle (R² = 0.60), liver (R² = 0.92), and adrenals (R² = 0.89), which underlines the quantitative character of DTPA-R PET imaging (Fig. 6G). Well-perfused organs showed a steeper increase of the PET signal with increasing AAV titer. This observation could either be due to a higher transduction rate in these organs or a better local delivery of [¹⁸F]F-DTPA. To test this hypothesis, we compared PET data for different tissues with perfusion values from the literature ^{37, 38}. Indeed, a linear correlation between the perfusion index and the slope of PET signal increase was seen (Fig. 6H).

Next, we stained tissues by V5-IHC and observed transduction of individual cells, for example in the cortex of adrenal glands (Fig. 6I). Quantification using automated positive cell detection showed 2% positive cells for animals transduced with the lowest, 33% for the low, and 72.2% for the high dose, which yielded a perfectly linear relation with the injected vector titer (R² = 1; Fig. 6K). Also, the relationship between the measured PET signal and the injected AAV titer (Fig. 6J), or the transduction levels from histology (Fig. 6L), indicated a linear correlation (R² = 0.98 or 0.96, respectively). Transduction levels were further quantified by ddPCR, revealing a linear correlation between the number of viral genomes and the PET signals in the liver (R² = 0.97; Fig. 6M). A multivariate analysis correlating the measured PET signals with injected AAV9 doses, mRNA levels, and viral genomes in the liver revealed a linear relationship (R² = 1, Fig. 6N).

We have developed a novel reporter gene system based on a membrane-anchored Anticalin that specifically binds a small-molecule radioligand, enabling quantitative and longitudinal PET imaging of ATMPs with remarkable specificity and sensitivity. Based on PET with [¹⁸F]F-DTPA, we have demonstrated the suitability of DTPAR in relevant use cases such as CAR-T cell therapy of CD19 lymphoma³⁹ and AAV9 gene therapy³⁵. CAR-T cell movement and tumor homing in a mouse model of systemic lymphoma could be visualized and quantitatively tracked in vivo over 30 days. Furthermore, the transduction of AAV9 vectors in distinct tissues could be quantitatively analyzed by molecular imaging. This theranostic approach, combining novel cell and gene therapies with a quantitative imaging modality using a universal reporter system that potentially can bridge preclinical development and clinical evaluation, should facilitate the clinical translation of ATMPs⁴⁰. To this end, our PET reporter system offers promising functional features, which are described below.

High expression level of the reporter protein determines strength of the signal, which has been demonstrated for Jurkat^DTPA−R cells displaying ~ 1×10⁶ receptors per cell without measurable negative effects on cellular fitness or T cell function. This number is approximately 10-fold higher compared to well-known lymphocyte receptors, such as the B cell receptor (1.2×10⁵ copies⁴¹) or CD4 (~ 1×10⁵ copies⁴²). Direct comparison of the Anticalin CL31d specific for [¹⁸F]FDTPA with the scFv huC825, which binds DOTA•metal, resulted in an 8.9-fold higher expression of DTPA-R (Fig. S2). Low surface densities for huC825-based reporter proteins are also reflected by transfected HEK293T cells expressing only ~ 1.5×10⁴ receptors per cell²⁶. In contrast to scFv antibody fragments, known for their oligomerization and aggregation tendencies⁴³, Anticalins possess a robust fold, are composed of a single polypeptide chain, and can be expressed as recombinant proteins at high levels¹⁸.

Minimal gene size is of importance due to the limited packaging capacities of viral and non-viral gene shuttles and met by DTPA-R, which is encoded by just 774 bp (~ 26.4 kDa for the mature fusion protein), in contrast to most other PET reporter proteins such as HSV1-tk (1,131 bp)¹¹, NIS (1,932 bp)^{12, 44}, tPSMA^(N9del) (2,226 bp)¹⁵, SStR2 (1,110 bp)¹³, DAbR1-2A-GFP (2,280 bp)²⁵, SNAPtag (846 bp)⁴⁵, eDHFR (480 bp)⁴⁶ or the EGFRt (1,074 bp)³².

Functional inertness, including lack of biological activity, interference, and toxicity of both the reporter gene and probe, are crucial factors as the capability of in vivo imaging is second to the therapeutic function of an ATMP. We have investigated the proliferation kinetics, activation status, receptor expression, and cellular toxicity of CAR-T cells and found no difference due to the DTPA-R expression compared to EGFRt. Given the binding activity of DTPA-R for an exogenous hapten, interferences with cellular processes are much less likely than for reporter proteins that lead to ion transport over the cell membrane (NIS), represent tumor-associated surface antigens (tPSMA, SStR2) or possess catalytic activities (HSV-tk, PSMA, SNAPtag, eDHFR).

Stability of the PET signal over time is important to allow reproducible imaging, which mainly relies on the receptor-ligand affinity. While the soluble recombinant CL31d protein shows a K_D value of 543 pM for the complex with CHX-A’’DTPA•Y¹⁷, the half-maximal inhibitory concentration (IC₅₀) of NH₂-CHX-A’’-DTPA•⁹⁰Y for DTPA-R expressed on the cell surface was slightly higher, with 1.4 nM. Nevertheless, the high stability of the [¹⁸F]F -DTPA signal within the tumor was demonstrated in dynamic PET scans. This is not the case for the NIS reporter gene, for example, due to the naturally occurring efflux of the radioactive iodide⁴⁴. Potential internalization of DTPA-R is very low due to the absence of the cytosolic CD4 domain, which triggers internalization of CD4⁴⁷. In addition, DTPA-R expression levels are independent of T cell activation, which is commonly known to heavily influence expression profiles in T cells⁴⁸.

Linear relation between PET signal and the number of DTPA-R-labelled cells is mandatory to move beyond qualitative imaging. The DTPA-R system allows a strong correlation of the number of CART cells in the bone marrow or the applied AAV9 virus titer with the corresponding PET signal, thus enabling non-invasive quantification. The amount of bound [¹⁸F]F-DTPA ligand is only governed by the local concentration of DTPA-R and the Law of Mass Action, contrasting to the much more complex and unpredictable relationships for transporter- or enzyme-type reporter proteins.

Criteria for a corresponding radioligand ⁴⁹ include its chemical composition as well as the choice of the radioisotope for PET imaging. The choice of ¹⁸F, with its ideal physical half-life, high positron yield, and low positron energy, allows detection with high sensitivity and resolution, especially if compared to isotopes such as ¹¹¹In or ⁸⁶Y, used for DAbR1 / C825 reporter probes^{25, 26}. Other advantages of ¹⁸F are the relatively short half-life of 109 min, which enables repeated serial imaging in a daily interval, while the half-live is long enough to allow multiple PET scans from a single batch of radioligand. Furthermore, the low radiation exposure does not hamper the function of the ATMP nor lead to a significant absorbed biological dose. The stability of [¹⁸F]F-DTPA in blood was demonstrated, together with a stable PC3^{DTPA − R} signal in the time-activity curve. Nevertheless, achieving high specific activity remains a challenge for [¹⁸F]F-DTPA synthesis, as well as for other ¹⁸F-based radioligands, and constitutes a current area of improvement.

The absence of endogenous binding activity was impressively demonstrated by the very low background accumulation of [¹⁸F]F-DTPA in various organs. This is in marked contrast to all fully human reporter proteins described so far, including NIS with its thyroidal, gastric, mucosal, salivary, and lactating mammary gland expression⁵⁰. Furthermore, the biodistribution analysis of [¹⁸F]F-DTPA in female and male mice showed no sex-specific differences. Excretion of [¹⁸F]F-DTPA was mainly seen via the renal route, which led to defined signals in the kidneys and bladder. In contrast, the radioligand for the C825-scFv-baed reporter gene system caused a diffuse signal in the abdomen at t = 30 min p.i., and only ~ 80% of the [⁸⁶Y]Y-DOTA-Bn was cleared via the renal route²⁶.

Furthermore, the ease of probe preparation and the price per imaging day are of importance both in biomedical research and in a clinical setting. In this regard, the DTPA-R system is attractive as the chemical synthesis of the ligand precursor is straight-forward. Furthermore, radiolabeling involves only few steps and the supply of [¹⁸F]F⁻ is neither limiting nor expensive.

Finally, the sensitivity of signal detection is a crucial aspect, which depends on the number of binding sites, radioligand affinity, physical decay characteristics of the isotope and the specific activity of the radioligand. Using a PET phantom, we determined a detection limit of 500 cells while a standard [¹⁸F]F-DTPA imaging protocol allowed the clear detection of as few as 8×10³ Jurkat^DTPA−R or αCD19-CAR-T^DTPA−R cells in vivo. Reported sensitivity for DTPA-R is in line or even higher when compared to other reporter systems⁵⁰. In a comparative study with primary T cells transfected with human reporter genes⁵¹, only the norepinephrine transporter (NET) with [¹⁸F]F-MFBG was able to detect 3 − 4×10⁴ cells, while for HSV-TK/[¹⁸F]F-FEAU and NIS/[¹²⁴I]-iodide a sensitivity of ~ 3×10⁵ and ~ 1×10⁶ was reported, respectively. In direct comparison with those reporter systems DTPA-R/[¹⁸F]F-DTPA not only achieved a higher sensitivity, but also shows the superior excretion profile which leads to lower background signals⁵¹.

In summary, the DTPA-R & [¹⁸F]F-DTPA system meets all relevant design and functional requirements for a universal reporter gene system, which may boost future ATMPs by making them PET-traceable.

Acknowledgments

The authors wish to thank Markus Mittelhäuser, Hannes Rolbieski, Sybille Reder, and Natalie Röder for PET/MR acquisition. Prof. Dr. Franz Schilling and Dr. Geoffrey Topping for help with MR sequence optimization, Olga Seelbach and Marion Mielke for histology, Anja Wolf for production of AAV9 vectors, Dr. Ritu Mishra for FACS sorting, Dr. Andreas Eichinger for help with the generation of the structural model, Prof. Bernhard Küster and Prof. Gil Westmeyer for cell line (all from TU Munich) and Lara Bontadelli (Novartis) for qPCR and ddPCR analysis.

This study was funded by the National Centre for the 3Rs via CRACK IT Challenge 32 (NC/C01905/1 & NC/C019202/1) and SFB824 of the German research foundation (projects A8, Z1 & Z2). Anticalin^® is a registered trademark of Pieris Pharmaceuticals GmbH.

Author Contributions

GlaxoSmithKline and Novartis supported the project with in-kind contributions in the frame of the CRACK IT program of the NC3Rs.

Competing interests

V.M., K.F., A.Sk., and W.W. have applied for intellectual property rights on Anticalins-based reporter genes (WO 20022/101492 A1). A.Sk. is a co-founder and shareholder of Pieris Pharmaceuticals.

Kingwell, K. CAR T therapies drive into new terrain. Nat Rev Drug Discov 16, 301–304 (2017).
Mendell, J.R. et al. Current Clinical Applications of In Vivo Gene Therapy with AAVs. Mol Ther 29, 464–488 (2021).
Minn, I., Rowe, S.P. & Pomper, M.G. Enhancing CAR T-cell therapy through cellular imaging and radiotherapy. Lancet Oncol 20, e443-e451 (2019).
Krebs, S., Dacek, M.M., Carter, L.M., Scheinberg, D.A. & Larson, S.M. CAR Chase: Where Do Engineered Cells Go in Humans? Front Oncol 10, 577773 (2020).
Iafrate, M. & Fruhwirth, G.O. How Non-invasive in vivo Cell Tracking Supports the Development and Translation of Cancer Immunotherapies. Front Physiol 11, 154 (2020).
Ashmore-Harris, C., Iafrate, M., Saleem, A. & Fruhwirth, G.O. Non-invasive Reporter Gene Imaging of Cell Therapies, including T Cells and Stem Cells. Mol Ther 28, 1392–1416 (2020).
Serganova, I. & Blasberg, R.G. Molecular Imaging with Reporter Genes: Has Its Promise Been Delivered? J Nucl Med 60, 1665–1681 (2019).
Volpe, A., Kurtys, E. & Fruhwirth, G.O. Cousins at work: How combining medical with optical imaging enhances in vivo cell tracking. Int J Biochem Cell Biol 102, 40–50 (2018).
Cherry, S.R. et al. Total-Body PET: Maximizing Sensitivity to Create New Opportunities for Clinical Research and Patient Care. J Nucl Med 59, 3–12 (2018).
Tjuvajev, J.G. et al. Imaging the expression of transfected genes in vivo. Cancer Res 55, 6126–6132 (1995).
Yaghoubi, S.S. et al. Noninvasive detection of therapeutic cytolytic T cells with ¹⁸F-FHBG PET in a patient with glioma. Nat Clin Pract Oncol 6, 53–58 (2009).
Merron, A. et al. SPECT/CT imaging of oncolytic adenovirus propagation in tumours in vivo using the Na/I symporter as a reporter gene. Gene Ther 14, 1731–1738 (2007).
Vedvyas, Y. et al. Longitudinal PET imaging demonstrates biphasic CAR T cell responses in survivors. JCI Insight 1, e90064 (2016).
Castanares, M.A. et al. Evaluation of prostate-specific membrane antigen as an imaging reporter. J Nucl Med 55, 805–811 (2014).
Minn, I. et al. Imaging CAR T cell therapy with PSMA-targeted positron emission tomography. Sci Adv 5, eaaw5096 (2019).
Kim, H.J., Eichinger, A. & Skerra, A. High-affinity recognition of lanthanide(III) chelate complexes by a reprogrammed human lipocalin 2. J Am Chem Soc 131, 3565–3576 (2009).
Eggenstein, E., Eichinger, A., Kim, H.J. & Skerra, A. Structure-guided engineering of Anticalins with improved binding behavior and biochemical characteristics for application in radio-immuno imaging and/or therapy. J Struct Biol (2013).
Deuschle, F.C., Ilyukhina, E. & Skerra, A. Anticalin^(R) proteins: from bench to bedside. Expert Opin Biol Ther 21, 509–518 (2021).
Schiefner, A. & Skerra, A. The menagerie of human lipocalins: a natural protein scaffold for molecular recognition of physiological compounds. Acc Chem Res 48, 976–985 (2015).
Dunn, C., O'Dowd, A. & Randall, R.E. Fine mapping of the binding sites of monoclonal antibodies raised against the Pk tag. J Immunol Methods 224, 141–150 (1999).
Bajar, B.T. et al. Improving brightness and photostability of green and red fluorescent proteins for live cell imaging and FRET reporting. Sci Rep 6, 20889 (2016).
Shcherbakova, D.M., Cox Cammer, N., Huisman, T.M., Verkhusha, V.V. & Hodgson, L. Direct multiplex imaging and optogenetics of Rho GTPases enabled by near-infrared FRET. Nat Chem Biol 14, 591–600 (2018).
Eggenstein, E., Eichinger, A., Kim, H.J. & Skerra, A. Structure-guided engineering of Anticalins with improved binding behavior and biochemical characteristics for application in radio-immuno imaging and/or therapy. J Struct Biol 185, 203–214 (2013).
Barkovskiy, M., Ilyukhina, E., Dauner, M., Eichinger, A. & Skerra, A. An engineered lipocalin that tightly complexes the plant poison colchicine for use as antidote and in bioanalytical applications. Biol Chem 400, 351–366 (2019).
Krebs, S. et al. Antibody with infinite affinity for in vivo tracking of genetically engineered lymphocytes. J Nucl Med 59, 1894–1900 (2018).
Dacek, M.M. et al. Engineered Cells as a Test Platform for Radiohaptens in Pretargeted Imaging and Radioimmunotherapy Applications. Bioconjug Chem 32, 649–654 (2021).
Sutermaster, B.A. & Darling, E.M. Considerations for high-yield, high-throughput cell enrichment: fluorescence versus magnetic sorting. Sci Rep 9, 227 (2019).
Sanchez-Crespo, A. Comparison of Gallium-68 and Fluorine-18 imaging characteristics in positron emission tomography. Appl Radiat Isot 76, 55–62 (2013).
Richarz, R. et al. Neither azeotropic drying, nor base nor other additives: a minimalist approach to ¹⁸F-labeling. Org Biomol Chem 12, 8094–8099 (2014).
Cardinale, J. et al. Development of PSMA-1007-Related Series of ¹⁸F-Labeled Glu-Ureido-Type PSMA Inhibitors. J Med Chem 63, 10897–10907 (2020).
Sommermeyer, D. et al. Fully human CD19-specific chimeric antigen receptors for T-cell therapy. Leukemia 31, 2191–2199 (2017).
Wang, X. et al. A transgene-encoded cell surface polypeptide for selection, in vivo tracking, and ablation of engineered cells. Blood 118, 1255–1263 (2011).
Paszkiewicz, P.J. et al. Targeted antibody-mediated depletion of murine CD19 CAR T cells permanently reverses B cell aplasia. J Clin Invest 126, 4262–4272 (2016).
Lee, J.C. et al. In vivo inhibition of human CD19-targeted effector T cells by natural T regulatory cells in a xenotransplant murine model of B cell malignancy. Cancer Res 71, 2871–2881 (2011).
Wang, D., Tai, P.W.L. & Gao, G. Adeno-associated virus vector as a platform for gene therapy delivery. Nat Rev Drug Discov 18, 358–378 (2019).
Zincarelli, C., Soltys, S., Rengo, G. & Rabinowitz, J.E. Analysis of AAV serotypes 1–9 mediated gene expression and tropism in mice after systemic injection. Mol Ther 16, 1073–1080 (2008).
Brown, L.V., Gaffney, E.A., Ager, A., Wagg, J. & Coles, M.C. Quantifying the limits of CAR T-cell delivery in mice and men. J R Soc Interface 18, 20201013 (2021).
Gjedde, S.B. & Gjedde, A. Organ Blood-Flow Rates and Cardiac-Output of the Balb-C Mouse. Comp Biochem Phys A 67, 671–674 (1980).
Amini, L. et al. Preparing for CAR T cell therapy: patient selection, bridging therapies and lymphodepletion. Nat Rev Clin Oncol (2022).
Weber, W.A., Varasteh, Z., Fritschle, K. & Morath, V. A Theranostic Approach for CAR-T Cell Therapy. Clin Cancer Res 28, 5241–5243 (2022).
Yang, J. & Reth, M. The dissociation activation model of B cell antigen receptor triggering. FEBS Lett 584, 4872–4877 (2010).
Wang, M. et al. Quantifying CD4 receptor protein in two human CD4 + lymphocyte preparations for quantitative flow cytometry. Clin Proteomics 11, 43 (2014).
Lee, C.C., Perchiacca, J.M. & Tessier, P.M. Toward aggregation-resistant antibodies by design. Trends Biotechnol 31, 612–620 (2013).
Ravera, S., Reyna-Neyra, A., Ferrandino, G., Amzel, L.M. & Carrasco, N. The Sodium/Iodide Symporter (NIS): Molecular Physiology and Preclinical and Clinical Applications. Annu Rev Physiol 79, 261–289 (2017).
Stotz, S., Bowden, G.D., Cotton, J.M., Pichler, B.J. & Maurer, A. Covalent ¹⁸F-Radiotracers for SNAPTag: A New Toolbox for Reporter Gene Imaging. Pharmaceuticals (Basel) 14 (2021).
Sellmyer, M.A. et al. Imaging CAR T Cell Trafficking with eDHFR as a PET Reporter Gene. Mol Ther 28, 42–51 (2020).
Bojkowska, K. et al. Measuring in vivo protein half-life. Chem Biol 18, 805–815 (2011).
van der Windt, G.J. & Pearce, E.L. Metabolic switching and fuel choice during T-cell differentiation and memory development. Immunol Rev 249, 27–42 (2012).
Volpe, A., Pillarsetty, N.V.K., Lewis, J.S. & Ponomarev, V. Applications of nuclear-based imaging in gene and cell therapy: probe considerations. Mol Ther Oncolytics 20, 447–458 (2021).
Shao, F. et al. Radionuclide-based molecular imaging allows CAR-T cellular visualization and therapeutic monitoring. Theranostics 11, 6800–6817 (2021).
Moroz, M.A. et al. Comparative Analysis of T Cell Imaging with Human Nuclear Reporter Genes. J Nucl Med 56, 1055–1060 (2015).

Yes there is potential Competing Interest. V.M., K.F., A.Sk., and W.W. have applied for intellectual property rights on Anticalins-based reporter genes (WO 20022/101492 A1). A.Sk. is a co-founder and shareholder of Pieris Pharmaceuticals.

SupplementaryDTPARPETreportergeneimagingNatBiotechnol.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

DTPA-Receptor – A novel reporter gene system for the specific and sensitive PET imaging of CAR-T cells and AAV transduced cells

Status:

Version 1

Abstract

Figures

Introduction

Results

Design of the reporter protein and radioligand

Analysis of reporter protein variants

Synthesis & characterization of [¹⁸F]F-colchicine and [¹⁸F]F-DTPA

Dynamic PET imaging of mice bearing xenograft tumors

Design and evaluation of DTPA-R labeled CAR-T cells

PET imaging of CAR-T cell therapy

PET imaging of gene transfer mediated by AAV9 viral vectors

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1

DTPA-Receptor – A novel reporter gene system for the specific and sensitive PET imaging of CAR-T cells and AAV transduced cells

Status:

Version 1

Abstract

Figures

Introduction

Results

Design of the reporter protein and radioligand

Analysis of reporter protein variants

Synthesis & characterization of [18F]F-colchicine and [18F]F-DTPA

Dynamic PET imaging of mice bearing xenograft tumors

Design and evaluation of DTPA-R labeled CAR-T cells

PET imaging of CAR-T cell therapy

PET imaging of gene transfer mediated by AAV9 viral vectors

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1

Synthesis & characterization of [¹⁸F]F-colchicine and [¹⁸F]F-DTPA