In-depth cross-validation of human and mouse CD4-specific minibodies for noninvasive PET imaging of CD4+ cells and response prediction to cancer immunotherapy

Increasing evidence emphasizes the pivotal role of CD4+ T cells in orchestrating cancer immunity. Noninvasive in vivo imaging of the temporal dynamics of CD4+ T cells and their distribution patterns might provide novel insights into their effector and regulator cell functions during cancer immunotherapy (CIT). Methods: We conducted a comparative analysis of 89Zr-labeled anti-mouse (m) and anti-human (h) CD4-targeting minibodies (Mbs) for in vivo positron emission tomography (PET)/magnetic resonance imaging (MRI) of CD4+ T cells in human xenografts, syngeneic tumor-bearing wild-type (WT), and human CD4+ knock-in (hCD4-KI) mouse models. Results: Both 89Zr-CD4-Mbs yielded high radiolabeling efficiencies of >90%, immunoreactivities of >70%, and specific in vitro binding to their target antigens. The specificity of in vivo targeting of 89Zr-hCD4-Mb was confirmed by PET/MRI, revealing ~4-fold greater 89Zr-hCD4-Mb uptake in subcutaneous hCD4+ hematopoietic peripheral blood acute lymphoblastic leukemia tumors (HPB-ALL) than in solid hCD4- diffuse histiocytic lymphomas (DHL) and 89Zr-mCD4-Mb uptake in hCD4+ HPB-ALL tumors. In a comparative cross-validation study in anti-programmed death ligand (αPD-L1)/anti-4-1BB-treated orthotopic PyMT mammary carcinoma-bearing hCD4-KI and WT mice, we detected 2- to 3-fold enhanced species-specific 89Zr-hCD4-Mb or 89Zr-mCD4-Mb uptake within CD4+ cell-enriched secondary lymphatic organs (lymph nodes and spleens). The 89Zr-hCD4-Mb uptake in the PyMT tumors was more pronounced in hCD4-KI mice compared to the WT control littermates. Most importantly, MC38 adenocarcinoma-bearing mice treated with a combination of αPD-L1 and anti-lymphocyte-activation gene 3 (αLag-3) antibodies exhibited ~1.4-fold higher 89Zr-mCD4-Mb uptake than mice that were not responsive to therapy or sham-treated mice. Conclusion: CD4 PET/MRI enabled monitoring of the CD4+ cell distribution in secondary lymphatic organs and the tumor microenvironment, capable of predicting sensitivity to CIT. Our imaging approach will provide deeper insights into the underlying molecular mechanisms of CD4-directed cancer immunotherapies in preclinical mouse models and is applicable for clinical translation.

The absence of reliable biomarkers for the prediction of efficient antitumoral immune responses upon cancer immunotherapy and the growing interest in harnessing CD4 + T cells for their antitumoral potential prompted us to embark on in-depth validation of a minibody-based strategy for tracking endogenous CD4 + T-cell dynamics within tumors and lymphatic organs in vivo by whole-body immune positron emission tomography (immunoPET).This technique offers a direct and noninvasive imaging approach combining the sensitivity of PET isotopes with the specificity of antibodies applicable for monitoring the expression of cancer and immune cell surface proteins [36,37].
To date, only few CD4 + cell-targeting PET tracers have been validated for use in animal models.Kim et al. evaluated a 89 Zr-labeled human CD4-specific therapeutic antibody ibalizumab and a rhesus F(ab´) 2 fragment to noninvasively quantify CD4 + cells in rhesus macaques [38].Kristensen et al. correlated PET uptake of a murine CD4-specific F(ab´) 2 tracer with ex vivo CD4 + T-cell tumor infiltration in seven preclinical syngeneic tumor models and monitored infiltration upon αPD-1-based checkpoint inhibitor therapy (CIT) [39].In addition, Clausen et al. applied an identical tracer to longitudinally investigate collagen-induced arthritis mouse models [40].Wu and colleagues developed a biologically inert murine CD4-specific cys-diabody PET tracer (~50 kDa) [41] to monitor the repopulation of CD4 + T cells after hematopoietic stem cell transplantation [42] and CD4 + T-cell migration into sites of inflammation in DSS-induced experimental colitis [43].Recently, we successfully developed a human CD4-targeting single domain antibody (nanobody)-based immunoPET probe (~15 kDa) to enable CD4 + cell tracking in the clinical setting and validated its biological properties and biodistribution in a human-CD4 knock-in mouse model [44].Nagle et al. implemented a human-specific CD4 scFv-CH3 fragment (minibody, Mb, ~80 kDa) radiolabeled with 64 Cu to follow human CD4 + T cell migration into a patient-derived glioblastoma [45].
In this study, we radiolabeled anti-mouse and anti-human CD4-targeting Mbs with 89 Zr using the chelator desferrioxamine (dfo) and extensively validated both immunotracers in xenograft, syngeneic, and hCD4 knock-in (KI) tumor mouse models to assess their specificity and cross-species comparability.Furthermore, we evaluated the potential of these compounds as imaging biomarkers for predicting the efficacy of ICI-based cancer immunotherapies.To our knowledge, this is the first study to analyze human-and mouse-specific immunotracers comparatively.As a cross-species theranostic approach, CD4-PET might be applied to novel immunotherapies in preclinical mouse models and subsequently used in clinical studies for therapy guidance and response evaluation.Thus, noninvasive in vivo monitoring of CD4 + cells with CD4-specific immunotracers holds the promise of providing valuable insights into the efficacy of cancer immunotherapy and will facilitate a better understanding of T-cell-based immunotherapies and CD4 + T-cell-mediated immune responses in immune-mediated inflammatory diseases.

Zr-hCD4-Mbs and Zr-mCD4-Mbs specifically bind to their target antigens in vitro
For radiolabeling with 89 Zr, hCD4-and mCD4-Mbs were conjugated to the radiochelator dfo, which did not impair the affinity in the (sub)nanomolar range as revealed by ELISA (Figure 1A) and flow cytometry (Figure S1A).Subsequent 89 Zr-radiolabeling yielded >90% radiolabeling efficiency, as determined by HPLC (Figure 1B).Maximum binding assays with increasing numbers of cells (Figure S1B) showed immunoreactive fractions of 78.4% for 89 Zr-hCD4-Mb using CD4-expressing human HPB-ALL cells and 84.3% for 89 Zr-mCD4-Mb using murine CD4 + T cells.We further confirmed the hCD4-expression on HPB-ALL cells (Figure S1D) and the specific binding of both tracers to the target antigen by species-specific CD4 + and CD4 -in vitro cell assays (Figure 1C, Figure S1E) and through CD4 blocking experiments using a 100-fold excess of unlabeled CD4-Mb.Furthermore, we tested for potential cross-reactivity of both tracers by ELISA with hCD4-or mCD4 using both tracers.Here, neither the hCD4-Mb bound to mouse mCD4, nor the mCD4-Mb to human CD4 (Figure S1C).Most importantly, hCD4-Mb did not induce human PBMC proliferation under coated or soluble conditions (Figure S1E) and remained as stable radioimmunoconjugate over 72 hours in serum (Figure S1F), prerequisite for potential clinical application.

hCD4-specific in vivo binding of 89 Zr-hCD4-Mb to hCD4 + HPB-ALL xenografts in NSG mice
After we successfully validated radiolabeling and specific in vitro binding of 89 Zr-hCD4-Mb and 89 Zr-mCD4-Mb to CD4 + cells, we examined the in vivo PET uptake of 89 Zr-hCD4-Mb by HPB-ALL leukemia xenografts (which constitutively express the hCD4 antigen) subcutaneously (s.c.) implanted into NSG mice.The tumor volumes at the tracer injection time point ranged between 50 and 700 mm 3 (Figure S2A).We observed a 3.5-to 5-fold increase in the 89 Zr-hCD4-Mb uptake within hCD4 + HPB-ALL tumors compared with hCD4 -DHL B-cell lymphomas at 6, 24, and 48 h post tracer injection (Figure 2A, 2C-D; Figure S2C).Interestingly, the 89 Zr-mCD4-Mb uptake in hCD4 + HPB-ALL tumors, which served as another control group, was slightly higher than the 89 Zr-hCD4-Mb uptake in CD4 -DHL tumors (Figure 2A, 2C -D; Figure S2C), indicating mCD4 + cell infiltration in T-cell-deficient NSG mice.We detected significantly higher 89 Zr-mCD4-Mb uptake within the spleen (Figure 2B; Figure S2B, in vivo) and in the lymph nodes (Figure 2E, ex vivo biodistribution) compared to both groups injected with 89 Zr-hCD4-Mb.Immunohistochemistry (IHC) of the tumors and spleens confirmed the presence of a limited number of mCD4 + cells, presumably CD4 + myeloid cells, based on their nuclear morphology and absence of CD3 + cells in the spleen (Figure 2F-G).As expected from the Mb-based tracer, most of both 89 Zr-CD4-Mbs cleared from the blood within 48 h with minimal residual uptake in the muscles, lungs, and hearts of all groups (Figure S2C-D).Furthermore, 89 Zr-mCD4-Mb exhibited a higher kidney uptake and lower liver uptake than 89 Zr-hCD4-Mb at the 48-h ex vivo biodistribution time point (Figure S2E), presumably attributed to different excretion kinetics of the radioimmunoconjugates.
In conclusion, this experiment confirmed the specific in vivo binding of 89 Zr-hCD4-Mb to hCD4 + cells and the presence of limited densities of mCD4 + cells 89 Zr-hCD4-Mb in the tumor tissue and lymphatic organs of NSG mice.
We detected 2-3-fold enhanced 89 Zr-hCD4-Mb uptake in the spleens and lymph nodes of PyMT tumor-bearing hCD4-KI mice compared with WT mice (Figure 3B, C; Figure S3A-B, S4B).Likewise, we found a similar 2-3-fold increased PET uptake with 89 Zr-mCD4-Mb in the lymphatic organs of PyMT tumor-bearing mCD4 + WT mice compared to non-species-specific hCD4-KI mice (Figure 3B, C; Figure S3A-B, S4B).Interestingly, the absolute 89 Zr-mCD4-Mb PET uptake was higher in the lymphatic organs of species-specific and non-species-specific groups than the 89 Zr-mCD4-Mb PET uptake (Figure 3B-D).Together with the different excretion kinetics within the NSG mouse models (Figure S2E), these data suggested a slightly different biodistribution between 89 Zr-mCD4-Mb and 89 Zr-hCD4-Mb in a target-unrelated manner.Also, the ex vivo biodistribution analysis of lymphatic organs confirmed our in vivo 89 Zr-hCD4-Mb-PET/MRI results (Figure 3D).In addition, we detected a similar αCD4 PET uptake in the tumor-draining and non-draining lymph nodes (Figure 3B-D).
Next, we focused on 89 Zr-hCD4-Mb and 89 Zr-mCD4-Mb uptake within the TME and detected significantly enhanced 89 Zr-hCD4-Mb uptake in αPD-1/α4-1BB-treated PyMT tumor-bearing hCD4-KI mice when compared to WT mice (3.22 ± 0.23 vs. 2.51 ± 0.06 %ID/ml, Figure 4A, B).Surprisingly, 89 Zr-mCD4-Mb PyMT tumor uptake was even higher in hCD4-KI mice than in target-specific WT mice expressing mCD4 + (Figure 4A, B).To further elaborate on whether target-specific tracer consumption effects in the lymphatic organs occurring only in species-specific mice were responsible for these unexpected observations, we analyzed the in vivo blood uptake in all experimental mice.We found a 1.9-fold lower 89 Zr-mCD4-Mb PET uptake in the blood of WT mice than hCD4-KI control mice (Figure 4C), resulting in a pronounced species-specific 89 Zr-mCD4-Mb PET uptake when corrected for blood background (8.76 ± 0.99 vs. 6.30± 0.77, non-significant, Figure 4D).These findings also applied to the 89 Zr-hCD4-Mb to a lesser extent (Figure 4C-D) and could be confirmed by ex vivo biodistribution analyses of the tumor and blood (Figure 4E-G).
Despite the limitations of knock-in/out mouse models in the absolute quantification of antibodybased immunotracers, we could quantitatively measure increased and comparable PET uptake patterns of the lymphatic organs and the TME in a syngeneic orthotopic tumor mouse model for 89 Zr-hCD4-Mb and 89 Zr-mCD4-Mb.
Consistent with our in vivo results, ex vivo organ biodistribution analysis of the tumors, muscle, and blood revealed increased 89 Zr-mCD4-Mb accumulation in the TME and significantly higher tumor-to-blood and tumor-to-muscle ratios in mice responding to ICI non-responsive MC38 tumors or tumors from sham-treated experimental mice (Figure 5I; Figure S5D).Besides, there were no differences in the ex vivo biodistribution of several organs of interest, such as the spleen, draining lymph nodes, liver, kidney, lung, or blood, between all experimental groups (Figure S5D).To exclude that the differences in tumor volumes between responding and non-responding animals affect the tracer uptake by biological (e.g.tumor penetration and accessibility) or physical (e.g.partial volume effects of the PET detectors) effects, we correlated the in vivo tracer uptake with the tumor volumes and the ex vivo tracer uptake with the tumor weight (Figure 5J).We observed a homogeneous distribution of the uptake values in the sham-treated group (not influenced by therapy-induced tumor growth inhibition) irrespective of the tumor size.Furthermore, comparable linear regressions and R 2 -values between in vivo and ex vivo analyses indicated that the tracer signals were influenced by partial volume or other physical effects.Noteworthy, flow cytometry of the TME revealed no relevant fraction of CD4 + cells when gated for CD45 + CD3 - myeloid cells (Figure S5B).

ICI response-associated infiltration of CD4 + T cells into TME
To better understand the increased 89 Zr-mCD4-Mb uptake within the TME and the intra-tumoral CD4 + T cell distribution of therapy responders, we performed immunofluorescence analyses of the MC38 tumors on day 9 of αPD-L1/αLag-3 therapy after the 48 h PET imaging time point.Representative high-resolution images of the tumor margin and the tumor core (Figure 6A) and CD4 + cell quantification (Figure 6B) revealed about 7-fold increased infiltration of CD3 + CD4 + T cells into the TME of responsive MC38 tumors, whereas no CD3 + CD4 + T cells could be identified in nonresponsive MC38 tumors or tumors from sham-treated mice (Figure 6A, B).Most importantly, we were able to colocalize the 89 Zr-mCD4-Mb distribution to endogenous tumor-infiltrating CD4 + cells by secondary staining of the Mb, highlighting the tracer-specific targeting of CD4 + cells within the TME at the cellular level (Figure 6C).
The lack of clinically translatable human-specific CD4-targeting immunoPET tracers and the promising clinical development of 89 Zr-crefmirlimab berdoxam motivated us to develop Mb-based PET tracers to monitor endogenous CD4 + cell dynamics in the TME and lymphatic organs.First, we verified the in vitro and in vivo target specificity of 89 Zr-hCD4-Mb and 89 Zr-mCD4-Mb.We demonstrated an ~80-fold higher in vitro binding of the 89 Zr-hCD4-Mb in the hCD4 + HBP-ALL cells than in the hCD4-DHL B-cell-lymphoma cells while in vivo accumulation of s.c.injected cells was only ~4-fold higher.In this context, in vivo biodistribution effects, permeability, and unspecific accumulation have to be considered that impair tracer accessibility and specificity, particularly in s.c.injected tumors.
We also monitored the presence of endogenous CD4 + cells in immune cell-enriched lymphatic organs, such as the spleen and lymph nodes, and in the TME of tumor-bearing WT and hCD4-KI mice.Interestingly, even low numbers of endogenous mCD4 + cells present in immunodeficient NSG mice [60][61][62] or in the TME of MC38 tumors could be differentiated by our 89 Zr-CD4-Mb immunoPET approach (Figure 2F), highlighting its superior sensitivity.As for visualizing tumor-infiltrating lymphocytes, enhanced tracer accumulation in the lymphatic organs may be considered an antigen sink, resulting in reduced amounts of radiotracer available to bind to target cells at the tumor site [39,63].This was expected for larger antibody-based molecules and can explain the limitations of head-to-head comparisons with considerable differences in the 89 Zr-CD4-Mb blood biodistribution using knock-out/knock-in mouse models (Figure 4).To avoid this, immune cell-rich organs could be saturated by administering an excess of an unlabeled antibody targeting the same antigen or a higher dose of the immunoPET tracer to increase the availability of the probe at the tumor site [39,63,64].Therefore, optimizing the dose may be essential for obtaining a sufficient imaging signal, while pre-dosing strategies pose a risk of perturbing the binding of tracers to intratumoral immune cells [39].
Besides the target abundance, several factors, including microvascular density, vascular permeability, stromal content, intratumoral pressure, and diffusion, influence tumor accumulation of immunoPET tracers [65].In our study, correcting for 89 Zr-mCD4-Mb and 89 Zr-hCD4-Mb uptake in the TME with the blood uptake enabled us to detect a species-specific Mb signal in experimental PyMT tumor-bearing WT and hCD4-KI mice [39].One of the main concerns in developing new immunotracers is the effects on targeted cells, such as unintentional immune cell activation or the inhibition of immune cell effector functioning.Our study revealed no impact of hCD4-Mb on the proliferation of human PBMCs in vitro.Similarly, the injection of 64 Cu-labeled hCD4-Mb by Nagle et al. did not result in detectable depletion or alterations in the proliferation or polarization of hCD4 + cells in a humanized glioblastoma model [45].These findings hold particular significance for translational studies on tracking CD4 + cells by hCD4-Mbs in patients.Also, Freise et al. could not detect alterations in CD3 + T-cell or CD45 + CD4 + cell compartments in the spleen, lymph nodes, thymus, or blood using a CD4-targeting 89 Zr-radiolabeled cys-diabody despite a mild reduction of CD4-expressing immune cells [41].Interestingly, the investigators reported a dose-dependent downregulation of membranous CD4 expression.Notably, decreased membranous CD4 expression has been reported on activated immune cells [66,67].Nevertheless, Freise et al. could not identify a CD4-cys-diabody-induced activation of CD4-expressing immune cells.
Given the fact that CD4 + cells represent a very heterogenous group with either protumoral or antitumoral function, we investigated whether 89 Zr-mCD4-Mb PET could assess early ICI response or resistance.Strikingly, we could discriminate MC38 tumor-bearing mice that were responsive (higher PET uptake) or nonresponsive (lower PET uptake) to αPD-L1/αLag-3 immunotherapy 7 days after therapy initiation.This finding was consistent with our IF analyses, demonstrating significant CD4 + T cell infiltration into the TME in therapy responders (Figure 6A-C) [70,71].In conclusion, our study presented a comprehensive preclinical crossvalidation of 89 Zr-CD4-Mb as a precision tool for noninvasive monitoring and visualizing endogenous CD4 + cells throughout the body.
Beyond oncological purposes, Mascio et al. performed valuable research on the whole-body PET visualization of CD4 + T cells in non-human primates by using dynamic PET data advanced mathematic modeling approaches.They applied the therapeutic anti-human CD4 full-length antibody ibalizumab and rhesus-specific F(ab') 2 antibody fragments to immunocompetent and Simian Immunodeficiency Virus-infected rhesus macaques.The overall capability of both radioimmunoconjugates to differentiate increased CD4 + T-cell densities of the spleen and lymph nodes appeared comparable to our results.However, the limitations of therapeutic antibodies for diagnostic imaging, observed dose-and blood pool-dependent uptake kinetics, and the use of a primarized F(ab')2 hinder their clinical translatability.
Considering the already demonstrated clinical success of 89 Zr-crefmirlimab berdoxam in targeting human CD8 + cells [51,52], our Mb-based imaging approach offers a remarkable balance between high target specificity and low toxicity due to its rapid biodistribution and clearance from blood.Based on increasing histology-driven evidence from human tumor tissues suggesting CD4 + cells as critical biomarkers to determine therapy efficacy in oncolytic virus [72], CAR T cells [73], and immune checkpoint inhibitor therapy [74][75][76][77][78][79], 89 Zr-CD4-Mb immunoPET holds great promise for clinical translation as a valuable tool for the noninvasive monitoring of CD4 + cells, and the patient-individualized prediction of cancer immunotherapy outcomes.

Enzyme-linked Immunosorbent Assay (ELISA)
Recombinant hCD4-His or mCD4-His (both Sino Biologicals) were diluted to 2 μg/mL in the carbonate bicarbonate buffer (Sigma), and 100 μL were added to each well of a flat-bottom plate and incubated at 2-8 °C overnight.The following day, the plate was washed 3x with Phosphate-Buffered Saline with Tween (PBST 0.05%) and blocked for 1 h at room temperature with PBS containing 1% bovine serum albumin (BSA) (Sigma).The plate was washed 3x with PBST (0.05%).A serial dilution was prepared by diluting the sample 1:2.5 to achieve a standard curve starting from 16 nM for both hCD4-Mb and mCD4-Mb; 100 μL of the sample was added to each well of the test plate and incubated for 1 h at room temperature.The plate was washed 3x with PBST (0.05%).Mouse anti-human IgG Fc-HRP detection antibody (Southern Biotech) was diluted 1:20,000 in assay buffer, and 100 μL was added to each well of the test plate and incubated for 1 h at room temperature.Subsequently, the plate was washed 3x with PBST (0.05%), 100 μL of TMB substrate (BioFx) was added to each well of the test plate and allowed to develop for 10-20 min at room temperature.The reaction was terminated by adding 100 μL of stop reagent (BioFx).The plate was read at 650 nm using a BioTek Synergy II plate reader.

Human PBMC proliferation assay
The antibodies were coated on plates in triplicate at 40 μg/mL in PBS at +4°C overnight.PBMCs were thawed and incubated in Human Serum AB according to the RESTORE protocol before use and plated at 100,000 cells/well (100 mL total).For the soluble assay format, antibodies were added at a concentration of 65 nM in triplicate to the plate coated with PBMCs.The plates were incubated at 37°C and 5% CO2 for 5 days, after which the CTG substrate was added, and the plates were placed on a shaker for 5 min.The luminescence of the plates was subsequently read on a plate reader to determine the extent of proliferation.

Radiolabeling
The dfo-conjugated hCD4-Mb and mCD4-Mb were radiolabeled with 200 MBq Zr-89 oxalate (Perkin Elmer) per mg protein.Briefly, the desired activity was neutralized with a 0.45-fold volume of 2 M sodium carbonate and buffered with a 5-fold volume of 0.5 M ammonium acetate to achieve a pH of 6.5-7.0.After adding the protein, labeling was allowed to proceed for 40 minutes at 25°C, after which the reaction was quenched by adding DTPA (0.2% solution, 40 µL per mg protein) and incubated for an additional 10 min.Incorporation of the radioisotope was confirmed by iTLC analysis (≥95% radiochemical purity (RCP); iTLC-SG (Agilent), mobile phase 10 mM EDTA).The identity of the elution profile with the original protein and radiolabeling of the protein peak was confirmed by high-performance size exclusion chromatography (HPSEC) (BioSep SEC s2000, 300x7.8mm, Phenomenex).In the case of insufficient RCP (<90%), the protein was purified using a Bio-Spin 6 desalting column (Bio-Rad) according to the manufacturer's instructionsto achieve >95% RCP.

Serum stability
Murine blood was obtained from the retrobulbar vein of C57BL/6 mice and collected in 1.5 mL Eppendorf tubes.The blood was kept at room temperature for 2 h and centrifuged at 12.000 rpm for 15 min.The serum was collected as supernatant and incubated with the radiolabeled 89 Zr-hCD4-Mb.HPLC measurements were performed at 0, 24, and 72h after tracer incubation and the % of 89 Zr-hCD4-Mb and 89 Zr-DTPA was determined.
Human CD4 + cells were isolated from whole-blood samples of healthy donors using StraightFrom Whole Blood CD4 MicroBeads (Miltenyi Biotec) immediately before each in vitro experiment.

Tumor cells
The hCD4 + hematopoietic peripheral blood acute lymphoblastic leukemia tumor (HPB-ALL) and hCD4 - diffuse histiocytic lymphoma (DHL) cell lines were purchased from the German Collection of Microorganisms and Cell Cultures (DSMZ) and cultured in RPMI-1640 medium, supplemented with 10% Fetal Calf Serum (FCS) and 1% P/S.The MC38 murine colon adenocarcinoma cell line was purchased from Kerafast and cultured in DMEM (Lonza Biosciences) supplemented with 10% FCS, 1% P/S, and 1% HEPES (Lonza Biosciences).The B16F10 murine melanoma cell line was purchased from ATCC and cultured in DMEM, supplemented with 10% FCS and 1% P/S.The S2WTP3 (PyMT) triple-negative breast cancer cell line was kindly provided by Andreas Moeller (Queensland University, Australia) and cultured in DMEM supplemented with 10% FCS, 1% P/S, and 1% pyruvate (Sigma).

In vitro binding assays
The in vitro binding of 89 Zr-mCD4-Mb or 89 Zr-hCD4-Mb was assessed using freshly isolated or immortalized murine or human CD4 + cells.Different numbers of cells were incubated with 10 ng of 89 Zr-mCD4-Mb or 89 Zr-hCD4-Mb for 90 min.The cells were washed with PBS/2% FCS and resuspended in 200 µL PBS / 2% FCS.For the antigen-blocking experiments, a 100-fold excess of unlabeled CD4-Mb was added to the wells 30 min before tracer incubation.The residual cell-bound radioactivity was measured in a γ-counter (Wallac 1480 WIZARD 3" Gamma Counter; PerkinElmer).The immunoreactive fraction of each radioimmunoconjugate was determined by calculating the maximum specific binding (Bmax) of a one-site nonlinear regression model within an antigen excess assay with an increasing number of cells (0 -32 x 10 6 ).

Image analysis
PET images from list mode were reconstructed using 2-dimensional ordered subset expectation maximization (OSEM-2D), applied to a Gaussian filter of 1.5 mm, and registered to the anatomical T2 MR images using Inveon Research Workplace (Siemens Preclinical Solutions).Volumes of interest (VOIs) of the organs of interest were created based on the anatomical MR images.The uptake values of the respective organs (%ID/mL) were calculated based on the Bq/mL after correction for radioactive decay and normalization to the injected activity.For visual comparison between the PET images, the signal intensity between the groups and the color scale was normalized within one experiment.

Ex vivo biodistribution
The experimental mice were sacrificed by cervical dislocation under deep anesthesia after the final imaging time point.Organs were harvested, and radioactivity was measured by γ-counting using an energy window between 350 and 650 keV.Standardized aliquots of the injected tracer were added to the wells for quantification.The values for each organ are expressed as the percentage of the overall injected dose per g (%ID/g), corrected for radioactive decay, and normalized to the injected activity.

Immunohistochemistry
After γ-counting of the organs, the tumors, spleens, and lymph nodes were fixed in formalin and embedded in paraffin.For histology, 3-5 µm-thick sections were cut and stained with hematoxylin and eosin (H&E).Immunohistochemistry was performed on an automated immunostainer (Ventana Medical Systems, Inc.) according to the manufacturer's protocols for open procedures, with slight modifications.All slides were stained with antibodies against CD4 (SP35; Zytomed, Berlin, Germany) and CD3 (SP7; DCS Innovative Diagnostics-Systeme GmbH & Co., KG).Appropriate positive and negative controls were used to confirm the adequacy of the staining.All samples were scanned with a Ventana DP200 (Roche, Basel, Switzerland) and processed with the Image Viewer MFC Application.Final image preparation was performed with Adobe Photoshop CS6.

Statistical analyses
All data were analyzed using GraphPad Prism, Version 9 or later (GraphPad Software, Inc., San Diego, California, USA).The values were expressed as the arithmetic mean ± standard error of the mean (SEM), if not otherwise stated in the figure legend.For statistical analyses, unpaired t-tests were applied for pairwise comparisons.Ordinary one-way ANOVA or two-way ANOVA was used for multiple group comparisons and was corrected for multiple comparisons using the Tukey post-hoc test.Adjusted P values less than 0.05 were considered significant, and significance levels are indicated as follows: * for ≤0.05, ** for ≤0.01, and *** for ≤0.001.

Figure 4 .
Figure 4. Noninvasive visualization of tumor-infiltrating endogenous CD4 + cells in hCD4-KI and WT PyMT mammary tumor-bearing mice.(A) Representative in vivo PET/MR images of PyMT mammary tumors acquired 48 h post-89 Zr-CD4-Mbs i.v.injection (day 4 post-αPD-1/α4-1BB mAb injection).Tumors are highlighted by white circles.(B) In vivo quantification of 89 Zr-h/mCD4-Mb uptake in PyMT tumors and (C) blood and (D) the tumor-to-blood ratio in hCD4-KI and WT mice 48 h post tracer injection.(E) Ex vivo quantification of 89 Zr-h/mCD4-Mb uptake in PyMT tumors and (F) blood as well as (G) the tumor-to-blood ratio for hCD4-KI and WT mice at 48 h post tracer injection measured by γ-counting.P values were calculated by unpaired t-test.Data are derived from two independent experiments (n = 4-6 per group).*p < 0.05, **p < 0.01, ***p < 0.001.