Advertisement

89Zr-labeled nivolumab for imaging of T-cell infiltration in a humanized murine model of lung cancer

  • Christopher G. England
  • Dawei Jiang
  • Emily B. Ehlerding
  • Brian T. Rekoske
  • Paul A. Ellison
  • Reinier Hernandez
  • Todd E. Barnhart
  • Douglas G. McNeel
  • Peng HuangEmail author
  • Weibo CaiEmail author
Original Article

Abstract

Purpose

Nivolumab is a human monoclonal antibody specific for programmed cell death-1 (PD-1), a negative regulator of T-cell activation and response. Acting as an immune checkpoint inhibitor, nivolumab binds to PD-1 expressed on the surface of many immune cells and prevents ligation by its natural ligands. Nivolumab is only effective in a subset of patients, and there is limited evidence supporting its use for diagnostic, monitoring, or stratification purposes.

Methods

89Zr-Df-nivolumab was synthesized to map the biodistribution of PD-1-expressing tumor infiltrating T-cells in vivo using a humanized murine model of lung cancer. The tracer was developed by radiolabeling the antibody with the positron emitter zirconium-89 (89Zr). Imaging results were validated by ex vivo biodistribution studies, and PD-1 expression was validated by immunohistochemistry. Data obtained from PET imaging were used to determine human dosimetry estimations.

Results

The tracer showed elevated binding to stimulated PD-1 expressing T-cells in vitro and in vivo. PET imaging of 89Zr-Df-nivolumab allowed for clear delineation of subcutaneous tumors through targeting of localized activated T-cells expressing PD-1 in the tumors and salivary glands of humanized A549 tumor-bearing mice. In addition to tumor uptake, salivary and lacrimal gland infiltration of T-cells was noticeably visible and confirmed via histological analysis.

Conclusions

These data support our claim that PD-1-targeted agents allow for tumor imaging in vivo, which may assist in the design and development of new immunotherapies. In the future, noninvasive imaging of immunotherapy biomarkers may assist in disease diagnostics, disease monitoring, and patient stratification.

Keywords

Nivolumab Programmed cell death 1 (PD-1) Positron emission tomography (PET) Immunotherapy Immune checkpoint inhibitor immunoPET 

Introduction

Programmed cell death 1 (PD-1) is an inhibitory costimulatory molecule expressed on the surface of some activated T-cells, B-cells, dendritic cells, and macrophages [1]. It is a member of the B7-CD28 family of proteins and performs fundamental roles in host immunity, tolerance, and tumor evasion. Activation of PD-1 occurs through binding of either of its two natural ligands, programmed death ligand-1 or -2 (PD-L1 or PD-L2), which are expressed on the surface of many antigen presenting cells and tumor cells from various malignancies arising from the breasts, ovaries, kidneys, bladder, and lungs [2]. The interactions between PD-1 and its natural ligands negatively regulates T-cell activation and effector function through downregulation of certain antiapoptotic pathways and proinflammatory cytokines [3], thus diminishing the immune system’s natural defense mechanism and allowing for continued tumor growth and disease progression [4]. Immunotherapies targeting this interaction leads to increased anti-tumor activities, including restoration of T-cell function and expansion, enhancement of natural killer and cytotoxic T-lymphocyte response, and increased cytokine secretion [5, 6, 7, 8]. The use of monoclonal antibodies as immune checkpoint inhibitors has become a common strategy for many treatment-experienced patients and is currently being explored for treatment-naïve patients in some cancers [9]. Despite leading to significant improvements in patient survival and overall outcomes, clinical data reveal that only a small subset of patients respond to anti-PD-1/PD-L1 blockade therapies when used as single agents [10, 11, 12].

Nivolumab (Opdivo®) and pembrolizumab (Keytruda®) are the only anti-PD-1 immunotherapies approved by the Federal Drug Administration in the United States as of 2017 [13]. Both of these therapies have shown durable remission and prolonged survival in some patients. For example, advanced non-small cell lung cancer patients treated with nivolumab showed a 16% overall 5-year survival rate, nearly 4-times higher than that of patients receiving traditional chemotherapeutics [14]. Despite showing efficacy in some patients, severe and sometimes fatal adverse reactions have been associated with nivolumab treatment, including autoimmune-mediated organ toxicities, upper respiratory infections, edema, pruritus, and skin infections [15]. Hence, some organizations have recommended the screening of patients before initiation of PD-1/PD-L1 therapies to alleviate unnecessary pain and suffering caused by adverse reactions in patients unlikely to respond to immune checkpoint therapies. Effective screening techniques have been dampened by the low and heterogeneous expression levels of PD-1 and PD-L1 detectable on tumors and immune cells [16].

Evaluation of the binding and pharmacokinetic properties of immune checkpoint inhibitor antibodies in vivo requires highly complex transgenic or humanized animal models. Because of the innate limitations of many transgenic models, the popularity of humanized animal models has exponentially increased over the last decade. The human peripheral blood lymphocytes-severe combined immunodeficiency (hu-PBL-SCID) model, also known as PBL, is the most readily available and cost-effective humanized murine model [17]. This mouse model is generated through engraftment of human peripheral mononuclear blood cells (hPBMCs) into severally immunodeficient NOD/SCID/IL2γc null (NSG) mice. However, engraftment of hPBMCs induces chronic graft-versus-host disease (GvHD) in the mice at 4–6 weeks post-engraftment, which can limit their usefulness in therapy studies [18, 19, 20]. For PD-1-based studies, the induction of GvHD leads to T-cell activation and increased PD-1 expression, which provides optimal conditions for immune checkpoint blockade imaging-based studies.

Few studies have investigated the role that molecular imaging may play in allowing researchers to visualize immunotherapy biomarkers [9]. For example, Natarajan et al. developed a murine anti-PD-1 tracer for positron emission tomography (PET) imaging of tumor-infiltrating lymphocytes in a murine model of mouse melanoma [21]. While the study revealed that PD-1 was an excellent biomarker for T-cell mapping, the use of a mouse antibody failed to address the clinical scenario and barriers faced in human experiments. In a previous study, our group investigated the pharmacokinetic properties and dosimetry of radiolabeled pembrolizumab in a humanized murine model (22); however, the study was limited to non-tumor-bearing mice and failed to address the imaging of infiltrating T-cells found in humanized murine models of cancer. To improve upon the previous study, we have developed a humanized murine model of lung cancer suitable for imaging the interactions between tumor and cancer cells. In this study, we develop and characterize the anti-PD-1 tracer (89Zr-Df-nivolumab) for imaging of PD-1-expressing T-cell infiltrates in a humanized mouse model of lung cancer. In the future, immunoPET imaging of cancer using immunotherapeutic antibodies may allow physicians to monitor the efficacy of immunotherapy intervention, and specifically tracking PD-1-expressing activated T-cells, while also improving patient stratification and possibly predicting autoimmune toxicities.

Materials and methods

Flow cytometry

This protocol was approved by an institutional review board at the University of Wisconsin–Madison, and all patients consented to the study. Blood samples were collected from patients for isolation of human peripheral blood mononuclear cells (hPBMCs) using the Ficoll-Histopaque (GE Healthcare, Little Chalfont, UK) technique by density gradient centrifugation. The cells were washed and stimulated for 18 h with 40 ng/mL phorbol myristate acetate (Sigma-Aldrich, St. Louis, MO, USA) and 1.3 μg/mL ionomycin (MP Biomedicals, Santa Ana, CA, USA). After stimulation of the hPBMCs, cells were stained with nivolumab or the chelator–antibody conjugate (Df-nivolumab) for 1 h at 4 °C. Next, the cells were before washed and stained with AlexaFluor488-labeled anti-human IgG (Life Technologies, Carlsbad, CA, USA) for 2 h at room temperature. Also, cells were stained with the primary antibodies CD3-v500, CD4-PE, and CD8-APC antibodies (BD Biosciences, San Jose, CA, USA) and GhostDye Red-780 (Tonbo Biosciences, San Diego, CA, USA). Cells were washed and resuspended in phosphate buffered saline (PBS) with 3% fetal calf serum, and analyzed on a BD LSRFortessa cytometer.

Chelation and radiolabeling procedures

Nivolumab was obtained commercially from Bristol-Myers Squibb Company (New York, NY, USA). The antibody was reconstituted in sterile PBS before conjugation with p-SCN-Deferoxamine (Df; Macrocyclics, Dallas, TX, USA) through the exposed lysine residues using methods previously described in the literature [22]. After mixing the antibody and chelator at a molar ratio of 1:5, the pH was adjusted to ~8.5 using Na2CO3 and the mixture was incubated at room temperature for 2 h. After reacting, the solution was purified via PD-10 columns (size exclusion chromatography) with PBS as the mobile phase. The isotope (89Zr) was produced in a GE PETtrace cyclotron by irradiation of natural yttrium foils (250 μm, 99.9%) with 13.8 MeV protons using methods previously described [23]. For radiolabeling, 74–148 MBq of neutralized 89Zr-oxalate was added to the antibody-chelator solution at a ratio of 0.25 mg of Df-nivolumab per 37 MBq of 89Zr. Before incubating at room temperature for 1 h, the volume was adjusted to 1 mL with HEPES buffer. The radiolabeled antibody was purified by PD-10 columns and injected into mice.

Animal models

All animal studies were conducted under a protocol approved by the University of Wisconsin Institutional Animal Care and Use Committee. NOD scid gamma (NSG), also known as NOD-scid IL2Rgammanull mice, were obtained from the UW-Madison Humanized Mouse Core Service. For generation of the hu-PBL-SCID model (PBL), NSG mice at 3–5 weeks of age were subcutaneously injected with 100 μL (1 × 106 cells) of a 1:1 mixture of A549 cells and Matrigel Matrix Basement Membrane (Corning, Corning, NY, USA) into the lower flank of the mice. Tumors were allowed to reach 1–2 mm in diameter before the mice were reconstituted with 0.5 × 106 hPBMCs via retro-orbital injection. The animals were allowed to engraft for 14–16 days before use, ensuring that both PBL and NSG mice of the same age and similar engraftment were used for imaging studies. Engraftment efficiency values were provided by the UW-Madison Humanized Mouse Core Service.

PET imaging and analysis

Mice were injected intravenously with 5–10 MBq of the tracer (89Zr-Df-nivolumab) prior to PET imaging. For imaging, mice were placed in supine position in the Inveon microPET/microCT rodent model scanner (Siemens Medical Solutions, Erlangen, Germany). Scans were performed with 40 million coincidence events static scans being recorded. Images were reconstructed using the 3D ordered subset expectation maximization algorithm and quantified via region-of-interest (ROI) analysis in the Inveon Research Workplace software (Siemens Medical Solutions). Signal quantification was expressed as the percentage of injected dose per gram of tissue (%ID/g).

Ex vivo biodistribution studies

Mice were euthanized via CO2 asphyxiation after the final scan for ex vivo biodistribution and immunohistochemistry studies. Major tissues, organs, and blood were collected for measurement of radioactivity. Each sample was wet-weighed before the radioactivity was counted using the PerkinElmer Wizard2 automatic γ-counter (Waltham, MA, USA). Values were normalized and recorded as %ID/g.

Immunohistochemistry

Tumor, salivary gland, spleen, and lung tissue were embedded in Tissue-Tek optimal cutting temperature compound (Sakura Finetek, Torrance, CA, USA). Frozen tissues were sectioned at 5-μm thickness and fixed with ice-cold acetone for 10 min before being rehydrated in PBS for 5 min. Slides were blocked with 10% normal donkey serum in 0.25% Triton-X in PBS for 1 h at room temperature. For primary antibodies, 1:200 dilutions of rabbit anti-human CD3 (Novus Biologicals, Littleton, CO, USA) and goat anti-human PD-1 (Novus Biologicals) were prepared in the blocking solution and incubated with the tissue for 12 h at 4 °C. Slides were washed three times with PBS for 15 min at room temperature. For secondary antibodies, 1:1000 donkey-anti-rabbit IgG DyLight 550 (Novus Biologicals) and 1:2000 donkey anti-goat IgG DyLight 488 (Novus Biologicals) were made using the blocking solution. The slides were incubated with secondary antibodies for 1 h at room temperature. Next, slides were washed three times with PBS for 10 min before mounting using a mixture of Fluoromount (Novus Biologicals) and Vectashield with DAPI (Vector Labs, Burlingame, CA, USA). The Nikon A1RS system was used to image slides.

Radiation dosimetry extrapolation to humans

The OLINDA/EXM software was used for dosimetry analysis. The %ID/g values obtained from PET imaging of the tracer in mice were used to estimate human dosimetry. The biodistribution of the tracer was assumed to be the same between humans and mice allowing for the use of a monoexponential model. OLINDA provides effective dose outputs and weighting factors from International Commission on Radiological Protection Publication 103 were employed to convert to absorbed dose in each organ [24].

Statistical analysis

Quantitative data were expressed as the mean ± standard deviation with all error bars denoting the standard deviation. Means were compared using the Student t-test and p-values ≤0.05 were considered statistically significant.

Results

Binding of nivolumab to PD-1 expressed on T-cells

Binding of nivolumab to PD-1 expressed on the surface of stimulated and unstimulated CD4+ and CD8+ T-cells was assessed via flow cytometry (Fig. 1). PMA-Ionomycin is a potent stimulator of lymphocytes known to quickly upregulate PD-1 expression in vitro [25]. Increased binding of nivolumab to both CD4+ and CD8+ T-cells was detected when the cells were stimulated; however, this interaction was three- to fourfold lower for unstimulated cells. To demonstrate the specificity further, a nonspecific human IgG was used in place of nivolumab. In return, minimal binding was detected between stimulated/unstimulated T-cells and the non-specific human IgG. Also, flow cytometry was used to evaluate if conjugation of the chelator (Df) would inhibit or alter the binding affinity of nivolumab. Both nivolumab and Df-nivolumab showed similar mean fluorescence intensities when incubated with stimulated CD4+ and CD8+ T-cells, indicating that the chelation and radiolabeling of nivolumab would not affect its binding affinity in vivo.
Fig. 1

Cell binding assay of nivolumab to unstimulated or stimulated CD4+ and CD8+ T-cells. The T-cells were stimulated to express PD-1 before incubation with IgG, nivolumab, or Df-nivolumab. Similar binding was displayed by nivolumab and Df-nivolumab indicating that binding was unaffected by chelation of the antibody. Nivolumab revealed significantly higher binding to stimulated T-cells, and minimal binding was detected with the nonspecific antibody (n=3)

Chelation and radiolabeling of nivolumab

Nivolumab was conjugated with the chelator (Df) for radiolabeling with 89Zr. The purified radiolabeled antibody displayed a specific activity of ~700 mBq per mg antibody. The radiolabeling efficiency was high with more than 85% of the isotope being chelated after 1 h of incubation, as determined by instant thin-layer chromatography.

PET imaging and biodistribution of 89Zr-Df-nivolumab in humanized tumor-bearing mice

After purification, 89Zr-Df-nivolumab was injected and longitudinal PET imaging studies were performed at 3, 6, 12, 24, 48, 72, and 168 h post-injection. Maximum intensity projections were acquired for data analysis. Initial imaging time points revealed similar pharmacokinetic properties of the tracer in NSG and PBL mice (Fig. 2). Differences in the biodistribution profile of the tracer became visible by 24 h post-injection. Quantification of PET data via ROI analysis allowed for comparison between tracer uptake in various organs and tissues, including the tumor, salivary glands, blood, liver, spleen, kidney, and muscle tissues. In both A549-tumor bearing PBL and NSG mice, tumors were identified on the left flank by 6 h post-injection. While tumor uptake in PBL mice continued to increase gradually throughout the study, uptake in NSG mice was significantly lower, reached maximum values by 24 h post-injection, and slowly declined by later time points.
Fig. 2

PET imaging of 89Zr-Df-nivolumab in A549 tumor-bearing PBL and NSG mice. Maximum intensity projections are shown from 3 to 168 h post-injection of the tracer with the scale ranging from 0 to 10%ID/g. SG, salivary gland; H, heart; L, liver; S, spleen; T, A549 tumor

As suggested by the maximum intensity projection images (Fig. 3), tumor accumulation was similar between NSG and PBL mice initially; however, PBL mice displayed higher uptake starting at 72 h after injection and onward (n = 4, p < 0.05; Fig. 3a). More specifically, tumor uptake values of 9.85 ± 2.73 and 3.88 ± 0.38 were obtained 168 h post-injection of 89Zr-Df-nivolumab into PBL and NSG mice, respectively (Table S1). In a previous study, our group verified that development of GvHD results in infiltration and permanent residence of lymphocytes into the lacrimal and salivary glands [22]. As expected, similar results were found in this study with PBL mice displaying significantly higher salivary gland uptake of 89Zr-Df-nivolumab in comparison to NSG mice (n = 4, p < 0.05; Fig. 3b). By 168 h post-injection, there was a twofold difference in salivary gland signal between PBL (8.38 ± 0.98%ID/g) and NSG mice (3.30 ± 0.22%ID/g), further validating the binding specificity of our tracer to its native receptor found on the surface of activated T-cells localized in the tumor and salivary glands.
Fig. 3

Analysis of PET imaging of 89Zr-Df-nivolumab in A549 tumor-bearing NSG and PBL mice. a Tumor uptake of 89Zr-Df-nivolumab was quantified and revealed significant differences after 72 h post-injection (n = 4, p ≤ 0.05). b Salivary gland uptake of 89Zr-Df-nivolumab was quantified and revealed significant differences after 24 h post-injection (n = 4, p ≤ 0.05). (c/d) PET region-of-interest analysis as time-activity curves of the signal in the blood, liver, spleen, kidneys, and muscle after intravenous injection of 89Zr-Df-nivolumab in c NSG and d PBL mice (n = 4)

The tracer showed similar characteristics in reference to blood circulation and off-target accumulation in both PBL and NSG mice. For PBL mice, tracer uptake in the blood pool was 15.5 ± 1.29%ID/g and 2.70 ± 0.42%ID/g at the initial and final time points, respectively (Fig. 3c-d). Similarly, blood pool uptake in NSG mice was 13.1 ± 0.79%ID/g and 1.88 ± 0.10%ID/g (n = 4) at the same time points, suggesting that the tracer remained relatively stable in circulation. Tracer uptake in the kidney and muscle tissues was similar between both groups of mice. In the liver, NSG mice showed higher accumulation likely due to off-target accumulation and antibody metabolism. After the final imaging session, select organs and tissues were extracted for further biodistribution evaluation by ex vivo studies. Similar to PET findings, ex vivo biodistribution studies revealed key differences between tumor and salivary gland uptake, which were two- to threefold higher in PBL mice (Fig. 4). Additionally, bone and spleen uptake was higher in NSG mice, which was not clearly represented by PET imaging. These differences were attributed to spleen size differences and large biases introduced by the partial volume effect [26].
Fig. 4

Ex vivo biodistribution of 89Zr-Df-nivolumab in A549 tumor-bearing NSG and PBL mice at 168 h post-injection (n = 4). SG - salivary gland

To verify the specificity of the tracer, PBL mice were injected with a nonspecific tracer (89Zr-Df-IgG) for imaging (Fig. 5). Imaging studies revealed lower tumor and salivary gland uptake in PBL mice injected with the nonspecific tracer in comparison to 89Zr-Df-nivolumab. For example, tumor signal was significantly lower in PBL mice injected with the nonspecific tracer as compared to 89Zr-Df-nivolumab at 3 h post-injection (0.74 ± 0.32 versus 2.46 ± 0.26%ID/g) and all time points thereafter (n = 4, p < 0.05; Fig. 6a). At the final time point (168 h), 89Zr-Df-nivolumab uptake in the tumors was threefold higher than 89Zr-Df-IgG, with values of 9.85 ± 2.73%ID/g and 2.85 ± 0.39%ID/g, respectively. In respect to the salivary gland uptake, there was a twofold difference between the two tracers with 89Zr-Df-nivolumab (8.38 ± 0.98%ID/g) being significantly higher than 89Zr-Df-IgG (4.65 ± 0.39%ID/g) at 168 h after injection (n = 4, p < 0.05; Fig. 6b). An increase in liver uptake was noticed with the nonspecific group; however, this is a common occurrence that naturally occurs with nonspecific antibodies, yet the cause of this phenomenon remains to be elucidated. Time activity curves show that blood pool uptake steadily decreased after injection, while spleen and muscle accumulation remained steady throughout the study. As demonstrated by PET imaging, the high liver uptake observed in mice injected with the nonspecific tracer was verified by ex vivo biodistribution studies (Fig. 6c-d).
Fig. 5

PET imaging of 89Zr-Df-IgG in A549 tumor-bearing PBL mice. Maximum intensity projections are shown from 3 to 168 h post-injection of the tracer with the scale ranging from 0 to 10%ID/g. SG, salivary gland; H, heart; L, liver; S, spleen; T, A549 tumor

Fig. 6

Analysis of PET imaging of 89Zr-Df-IgG and 89Zr-Df-nivolumab in A549 tumor-bearing PBL mice. (a) Tumor uptake of 89Zr-Df-IgG in PBL mice was quantified and showed significant differences from the first time point onward from PBL mice injected with 89Zr-Df-nivolumab (n = 4, p ≤ 0.05). (b) Salivary gland uptake of 89Zr-Df-IgG was quantified and showed significant differences from PBL mice injected with 89Zr-Df-nivolumab starting at 6 h post-injection (n = 4, p ≤ 0.05). (c) PET region-of-interest analysis as time-activity curves of the signal in the blood, liver, spleen, kidneys, and muscle after intravenous injection of 89Zr-Df-IgG. (d) Ex vivo biodistribution of 89Zr-Df-IgG in PBL mice. SG, Salivary gland

Verification of PD-1 expression by immunohistochemistry

Immunofluorescence imaging was used to further validate the imaging findings (Fig. 7). An anti-cluster of differentiation (CD3) antibody was used as a T-cell marker as shown in green, while PD-1 expression was signified by the red color. PD-1 expression was detected in the tumor, salivary gland, spleen, and lung sections of PBL mice, with the strongest signal being in the tumor and spleen. CD3 signal was also detectable in the four tissue sections with the tumor showing the highest degree of CD3. The nuclei were stained blue to verify the presence of cells. In the NSG tissue sections, minimal PD-1 and CD3 expression were visible.
Fig. 7

Histological staining of tumor, salivary gland, spleen, and lung tissue sections from PBL and NSG mice to verify the localization of PD-1 and CD3+ (T-cell marker). Red - PD-1; Green - CD3; Blue - DAPI. DAPI: 4′,6-diamidino-2-phenylindole, Scale bar: 20 μm

Radiation dosimetry

Dosimetry studies were performed to extrapolate the doses to mice following injection of 89Zr-Df-nivolumab or 89Zr-Df-IgG to adult human females (Table 1). Notable differences in dose were observed for the liver and salivary glands across groups, with other organs providing uniform results. In PBL mice, higher doses to the liver were found with the nonspecific antibody with values of 1.71 ± 0.11 mGy/MBq and 0.98 ± 0.12 mGy/MBq for 89Zr-Df-IgG and 89Zr-Df-nivolumab, respectively. This was attributed to the rapid clearance of nivolumab from circulation when no antigens for attachment were detected in circulation. In addition, notable doses to the salivary glands in all PBL mice (1.61 ± 0.15 mGy/MBq in Nivolumab PBL and 1.0 ± 0.15 mGy/MBq in IgG PBL) were the result of the GvHD process. As shown, low doses to tumors were observed across all groups. For example, doses of 0.025 ± 0.003 mGy/MBq for radiolabeled nivolumab in PBL mice, 0.020 ± 0.002 mGy/MBq for radiolabeled nivolumab in NSG mice, and 0.013 ± 0.001 mGy/MBq for radiolabeled nonspecific IgG in PBL mice, reported for a tumor mass of 2 g. Even considering these differences, similar total body effective doses were calculated across all groups to be about 0.3 mSv/MBq.
Table 1

Estimated radiation absorbed doses to an adult human based on intravenous injection of 89Zr-Df-nivolumab in PBL and NSG mice, along with 89Zr-Df-IgG in PBL mice (n = 4). SD – standard deviation

Organ

89Zr-Df-nivolumab in NSG (mGy/MBq)

89Zr-Df-nivolumab in PBL (mGy/MBq)

89Zr-Df-IgG in PBL (mGy/MBq)

 

Average

SD

Average

SD

Average

SD

Adrenals

0.013

0.001

0.012

0.001

0.015

0.000

Brain

0.064

0.005

0.058

0.005

0.055

0.002

Breasts

0.059

0.004

0.056

0.004

0.059

0.001

Small Intestines

0.008

0.001

0.008

0.001

0.008

0.000

Stomach Wall

0.231

0.017

0.211

0.017

0.231

0.003

Kidneys

0.026

0.004

0.020

0.002

0.024

0.001

Liver

0.973

0.077

0.983

0.121

1.708

0.112

Lungs

0.209

0.014

0.200

0.016

0.228

0.003

Muscle

0.007

0.001

0.007

0.001

0.007

0.000

Ovaries

0.495

0.037

0.452

0.036

0.448

0.013

Pancreas

0.013

0.001

0.012

0.001

0.014

0.001

Red Marrow

0.177

0.013

0.164

0.013

0.171

0.003

Skin

0.119

0.009

0.111

0.009

0.113

0.003

Spleen

0.028

0.002

0.021

0.001

0.018

0.001

Thymus

0.008

0.001

0.008

0.001

0.008

0.000

Salivary Glands

0.844

0.036

1.605

0.154

0.999

0.150

Urinary Bladder

0.180

0.014

0.164

0.013

0.160

0.005

Uterus

0.008

0.001

0.007

0.001

0.007

0.000

Effective Dose (mGy/MBq)

0.387

0.032

0.309

0.019

0.351

0.018

Discussion

Despite being the most preventable malignancy, lung cancer remains the leading cause of cancer-related death worldwide, accounting for nearly 1.6 million estimated deaths in 2017 [27]. Detection of early disease remains critical to successful treatment, yet the lack of physical symptoms results in more than 70% of patients presenting with advanced stage disease (stage III/IV) at the time of diagnosis. While surgical intervention yields the highest cure rates, most patients with advanced stage disease are not candidates for surgery. Hence, these patients rely on radiation therapy, chemotherapy, and newer targeted therapies like immune checkpoint inhibitors. With this in mind, the newly approved immunotherapy agents pembrolizumab and nivolumab have significantly improved the survival rates in a subset of lung cancer patients [28].

In this study, we demonstrated that T-cells expressing PD-1 could be effectively visualized using our novel PD-1 tracer, 89Zr-Df-nivolumab. While most imaging agents are used to directly target the tissue of interest, this work explores the utilization of an imaging agent for non-directly imaging tumor cells. Previously, Natarajan et al. demonstrated the specificity of a murine PD-1 tracer for imaging PD-1-expressing tumor infiltrating lymphocytes using 64Cu in a transgenic mouse model of melanoma [29]. Tumor uptake was 7.4 ± 0.71%ID/g at the final imaging time point (48 h post-injection), which was lower than the tumor accumulation found in this study (9.85 ± 2.73%ID/g). However, our value was obtained at 168 h post-injection, suggesting that the humanized antibody requires more time to accumulate in the tumor as compared to the murine antibody. Additionally, the murine tracer may have continued to accumulate if the study was not halted at 48 h post-injection; however, the short half-life of 64Cu (12 h) was a limiting factor. By employing the long decay half-life properties of 89Zr (3.3 days), this study allowed for the tracking of PD-1-expressing T-cells infiltration into the tumor over the course of 168 h (Fig. 2). In addition, it is important to note that tumor uptake is less than that usually noted with anti-tumor antibodies; however, this is because our imaging agent primarily targets T-cells localized in the tumor (non-direct imaging) and not the actual tumor cells.

Despite an increase in the popularity of immunotherapies, limited studies have characterized the biodistribution and pharmacokinetic properties of immune checkpoint inhibitor antibodies in vivo outside of clinical trials. In a previous study, England et al. demonstrated for the first time that radiolabeled pembrolizumab showed a unique biodistribution profile in humanized mice injected with hPBMCs when compared to non-humanized mice. Radiolabeled pembrolizumab was shown to accumulate in the salivary glands of PBL mice. Similarly, this study revealed increased salivary gland uptake of the tracer (Fig. 3), which has been attributed to T-cells trafficking to the salivary glands as a result of GvHD and other autoimmune diseases [30]. Another interesting finding in this study was the increased spleen uptake of the tracer found in the NSG model (Fig. 2). This is, in part, the result of spleen size as the spleens of NSG mice are significantly smaller than those of immunocompetent mice. Additionally, the spleens of PBL mice can enlarge by 10–50 once injected with hPMBCs [31]; hence, spleen uptake was higher in NSG mice because the smaller spleen led to a large bias introduced by the partial volume effect. Unlike our previous investigation of pembrolizumab, PBL and NSG mice were implanted with tumor cells in this study to assess if targeting T-cells would allow for tumor imaging.

This study demonstrated that humanized murine models of cancer may play important roles in the study and development of future immunotherapy agents. In this study, tracer accumulation is attributed to T-cell infiltration into the tumor, yet tumor cells do not display PD-1 expression. The process of tumor infiltration by lymphocytes has been extensively documented and this type of study may provide additional data regarding the process [32]. Using PBL mice injected with 89Zr-Df-IgG, we verified that the accumulation of our novel tracer was highly specific and not due to off-target accumulation (Fig. 6). In the future, the use of more complex humanized models of cancer may provide added benefits. For example, a limitation of this study was the innate variability caused by the lack of a stable model, as the xenograft efficiency varied between mice. Despite ensuring that all mice met the minimum xenograft efficiency required for study inclusion, the slight variations in xenograft efficiency and GvHD severity could have resulted in small differences we found in mice of the same group.

There are many potential applications likely to benefit from the PD-1 imaging agents. As described earlier, PD-1-based imaging agents may allow physicians to determine which patients are more apt to benefit from anti-PD-1 or anti-PD-L1 therapies in the future. In return, this could save thousands of patients from the unnecessary adverse effects associated with immune checkpoint therapies. Also, PD-1 imaging may be utilized to track T-cells that have become activated by various therapies to gain insight into the interactions between T-cell activation, homing, and tumor residence. These new imaging agents may play an essential role in assisting researchers and physicians determine other potential toxicities that may be associated with newer anti-PD-1 therapies in the future. A limitation for use of immune checkpoint antibodies as imaging agents is that immune checkpoint pathway receptors are heterogeneously expressed in multiple areas of the body, which complicates potential imaging strategies. Hence, innovative approaches are required to target many immunotherapy biomarkers, which create a key limitation in the clinical utility of these agents at this time for imaging-related studies. Lastly, the presence of PD-1 on the surface of tumor cells may also limit the potential of PD-1-targeted imaging agents as a predictive or pharmacodynamics biomarker as this could bias our judgment on treatment prediction. However, PD-1 expression on the surface of T-cells is much higher than that of any tumor cells, suggesting that PD-1 expression from tumor cells would have minimal effect on imaging studies.

Conclusions

Herein, this study demonstrates that radiolabeled nivolumab may be used to noninvasively image PD-1 expression in humanized tumor-bearing mice. More specifically, the infiltration of PD-1-expressing T-cells was effectively imaged in a humanized mouse model implanted with A549 tumors. As the field of cancer immunotherapy is expected to undergo rapid growth during the next decade, improved strategies like noninvasive PET are valuable tools for investigating the biodistribution of newly developed antibody therapies. In the future, imaging of immunotherapy biomarkers may be used for therapeutic monitoring and patient stratification.

Notes

Acknowledgements

This work was supported, in part, by the University of Wisconsin - Madison, the National Institutes of Health (NIBIB/NCI 1R01CA169365, 1R01CA205101, 1R01EB021336, T32CA009206, T32GM008505, 5T32GM08349, P30CA014520), the National Science Foundation (DGE-1256259), the American Cancer Society (125246-RSG-13-099-01-CCE), the National Science Foundation of China (81401465, 51573096), and the Basic Research Program of Shenzhen (JCYJ20170412111100742, JCYJ20160422091238319).

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval

All applicable international, national, and/or institutional guidelines for the care and use of animals were followed.

Supplementary material

259_2017_3803_MOESM1_ESM.docx (14 kb)
ESM 1 (DOCX 14 kb)

References

  1. 1.
    Philips GK, Atkins M. Therapeutic uses of anti-PD-1 and anti-PD-L1 antibodies. Int Immunol. 2015;27(1):39–46. doi: 10.1093/intimm/dxu095.CrossRefGoogle Scholar
  2. 2.
    Patel SP, Kurzrock R. PD-L1 expression as a predictive biomarker in cancer immunotherapy. Mol Cancer Ther. 2015;14(4):847–56. doi: 10.1158/1535-7163.MCT-14-0983.CrossRefGoogle Scholar
  3. 3.
    Keir ME, Butte MJ, Freeman GJ, Sharpe AH. PD-1 and its ligands in tolerance and immunity. Annu Rev Immunol. 2008;26:677–704. doi: 10.1146/annurev.immunol.26.021607.090331.CrossRefPubMedPubMedCentralGoogle Scholar
  4. 4.
    Wang J, Yuan R, Song W, Sun J, Liu D, Li Z. PD-1, PD-L1 (B7-H1) and tumor-site immune modulation therapy: the historical perspective. J Hematol Oncol. 2017;10(1):34. doi: 10.1186/s13045-017-0403-5.CrossRefPubMedPubMedCentralGoogle Scholar
  5. 5.
    Zou W, Wolchok JD, Chen L. PD-L1 (B7-H1) and PD-1 pathway blockade for cancer therapy: mechanisms, response biomarkers, and combinations. Sci Transl Med. 2016;8(328):328rv4. doi: 10.1126/scitranslmed.aad7118.CrossRefPubMedPubMedCentralGoogle Scholar
  6. 6.
    Yuzefpolskiy Y, Baumann FM, Penny LA, Kalia V, Sarkar S. Signaling through PD-1 on CD8 T cells is critical for antigen-independent maintenance of immune memory. J Immunol. 2016;196(1 Supplement):129.6-.6.Google Scholar
  7. 7.
    Guo Y, Feng X, Jiang Y, Shi X, Xing X, Liu X, et al. PD1 blockade enhances cytotoxicity of in vitro expanded natural killer cells towards myeloma cells. Oncotarget. 2016;7(30):48360–74. doi: 10.18632/oncotarget.10235.CrossRefPubMedPubMedCentralGoogle Scholar
  8. 8.
    Dai C, Lin F, Geng R, Ge X, Tang W, Chang J, et al. Implication of combined PD-L1/PD-1 blockade with cytokine-induced killer cells as a synergistic immunotherapy for gastrointestinal cancer. Oncotarget. 2016;7(9):10332–44. doi: 10.18632/oncotarget.7243.CrossRefPubMedPubMedCentralGoogle Scholar
  9. 9.
    Ehlerding EB, England CG, McNeel DG, Cai W. Molecular imaging of immunotherapy targets in cancer. J Nucl Med. 2016;57(10):1487–92. doi: 10.2967/jnumed.116.177493.CrossRefPubMedPubMedCentralGoogle Scholar
  10. 10.
    Heinzerling L, Kirchberger MC, Walter L, Schuler G. Predicting the response to anti-PD1 therapy in metastatic melanoma. Transl Cancer Res. 2016;5(3):S576–S9.CrossRefGoogle Scholar
  11. 11.
    Daud AI, Loo K, Pauli ML, Sanchez-Rodriguez R, Sandoval PM, Taravati K, et al. Tumor immune profiling predicts response to anti-PD-1 therapy in human melanoma. J Clin Invest. 2016;126(9):3447–52. doi: 10.1172/JCI87324.CrossRefPubMedPubMedCentralGoogle Scholar
  12. 12.
    Diggs LP, Hsueh EC. Utility of PD-L1 immunohistochemistry assays for predicting PD-1/PD-L1 inhibitor response. Biomark Res. 2017;5(1):12. doi: 10.1186/s40364-017-0093-8.CrossRefPubMedPubMedCentralGoogle Scholar
  13. 13.
    Gangadhar TC, Salama AK. Clinical applications of PD-1-based therapy: a focus on pembrolizumab (MK-3475) in the management of melanoma and other tumor types. Onco Targets Ther. 2015;8:929–37. doi: 10.2147/OTT.S53164.CrossRefPubMedPubMedCentralGoogle Scholar
  14. 14.
    Gettinger SN, Horn L, Gandhi L, Spigel DR, Antonia SJ, Rizvi NA, et al. Overall survival and long-term safety of nivolumab (anti-programmed death 1 antibody, BMS-936558, ONO-4538) in patients with previously treated advanced non-small cell lung cancer. J Clin Oncol. 2015;33(18):2004–12. doi: 10.1200/JCO.2014.58.3708.CrossRefPubMedPubMedCentralGoogle Scholar
  15. 15.
    Michot JM, Bigenwald C, Champiat S, Collins M, Carbonnel F, Postel-Vinay S, et al. Immune-related adverse events with immune checkpoint blockade: a comprehensive review. Eur J Cancer. 2016;54:139–48. doi: 10.1016/j.ejca.2015.11.016.CrossRefPubMedPubMedCentralGoogle Scholar
  16. 16.
    Gniadek TJ, Li QK, Tully E, Chatterjee S, Nimmagadda S, Gabrielson E. Heterogeneous expression of PD-L1 in pulmonary squamous cell carcinoma and adenocarcinoma: implications for assessment by small biopsy. Mod Pathol. 2017;30(4):530–8. doi: 10.1038/modpathol.2016.213.CrossRefGoogle Scholar
  17. 17.
    Brehm MA, Wiles MV, Greiner DL, Shultz LD. Generation of improved humanized mouse models for human infectious diseases. J Immunol Methods. 2014;410:3–17. doi: 10.1016/j.jim.2014.02.011.CrossRefGoogle Scholar
  18. 18.
    Blazar BR, Carreno BM, Panoskaltsis-Mortari A, Carter L, Iwai Y, Yagita H, et al. Blockade of programmed death-1 engagement accelerates graft-versus-host disease lethality by an IFN-gamma-dependent mechanism. J Immunol. 2003;171(3):1272–7.CrossRefGoogle Scholar
  19. 19.
    Schilbach K, Schick J, Wehrmann M, Wollny G, Simon P, Schlegel PG, et al. PD-1-PD-L1 pathway is involved in suppressing alloreactivity of heart infiltrating t cells during murine gvhd across minor histocompatibility antigen barriers. Transplantation. 2007;84(2):214–22. doi: 10.1097/01.tp.0000268074.77929.54.CrossRefGoogle Scholar
  20. 20.
    Habicht A, Kewalaramani R, Vu MD, Demirci G, Blazar BR, Sayegh MH, et al. Striking dichotomy of PD-L1 and PD-L2 pathways in regulating alloreactive CD4(+) and CD8(+) T cells in vivo. Am J Transplant. 2007;7(12):2683–92. doi: 10.1111/j.1600-6143.2007.01999.x.CrossRefGoogle Scholar
  21. 21.
    Hettich M, Braun F, Bartholoma MD, Schirmbeck R, Niedermann G. High-resolution PET imaging with therapeutic antibody-based PD-1/PD-L1 checkpoint tracers. Theranostics. 2016;6(10):1629–40. doi: 10.7150/thno.15253.CrossRefPubMedPubMedCentralGoogle Scholar
  22. 22.
    England CG, Ehlerding EB, Hernandez R, Rekoske BT, Graves SA, Sun H, et al. Preclinical pharmacokinetics and biodistribution studies of 89Zr-labeled pembrolizumab. J Nucl Med. 2017;58(1):162–8. doi: 10.2967/jnumed.116.177857.CrossRefPubMedPubMedCentralGoogle Scholar
  23. 23.
    Hernandez R, Sun H, England CG, Valdovinos HF, Ehlerding EB, Barnhart TE, et al. CD146-targeted immunoPET and NIRF imaging of hepatocellular carcinoma with a dual-labeled monoclonal antibody. Theranostics. 2016;6(11):1918–33. doi: 10.7150/thno.15568.CrossRefPubMedPubMedCentralGoogle Scholar
  24. 24.
    The 2007 Recommendations of the international commission on radiological protection. ICRP publication 103. Ann ICRP. 2007;37(2–4):1–332. doi: 10.1016/j.icrp.2007.10.003.
  25. 25.
    Ikebuchi R, Konnai S, Okagawa T, Yokoyama K, Nakajima C, Suzuki Y, et al. Blockade of bovine PD-1 increases T cell function and inhibits bovine leukemia virus expression in B cells in vitro. Vet Res. 2013;44(1):59. doi: 10.1186/1297-9716-44-59.CrossRefPubMedPubMedCentralGoogle Scholar
  26. 26.
    Soret M, Bacharach SL, Buvat I. Partial-volume effect in PET tumor imaging. J Nucl Med. 2007;48(6):932–45. doi: 10.2967/jnumed.106.035774.CrossRefPubMedPubMedCentralGoogle Scholar
  27. 27.
    Cancer Facts & Figures 2016. American Cancer Society. 2016;Atlanta, GA.Google Scholar
  28. 28.
    Garon EB, Rizvi NA, Hui R, Leighl N, Balmanoukian AS, Eder JP, et al. Pembrolizumab for the treatment of non-small-cell lung cancer. N Engl J Med. 2015;372(21):2018–28. doi: 10.1056/NEJMoa1501824.CrossRefGoogle Scholar
  29. 29.
    Natarajan A, Mayer AT, Xu L, Reeves RE, Gano J, Gambhir SS. Novel radiotracer for ImmunoPET imaging of PD-1 checkpoint expression on tumor infiltrating lymphocytes. Bioconjug Chem. 2015;26(10):2062–9. doi: 10.1021/acs.bioconjchem.5b00318.CrossRefGoogle Scholar
  30. 30.
    Hayashi T. Dysfunction of lacrimal and salivary glands in Sjogren's syndrome: nonimmunologic injury in preinflammatory phase and mouse model. J Biomed Biotechnol. 2011;2011:407031. doi: 10.1155/2011/407031.CrossRefPubMedPubMedCentralGoogle Scholar
  31. 31.
    Poluektova LY, Makarov E. Human peripheral blood lymphocyte reconstituion as a model of neuroinflammation associated with graft-versus-host disease. In: Xiong H, Gendelman HE, editors. Current laboratory methods in neuroscience research. New York: Springer; 2014. p. 487–90.Google Scholar
  32. 32.
    Reynders K, De Ruysscher D. Tumor infiltrating lymphocytes in lung cancer: a new prognostic parameter. J Thorac Dis. 2016;8(8):E833-5. doi: 10.21037/jtd.2016.07.75.CrossRefPubMedPubMedCentralGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany 2017

Authors and Affiliations

  • Christopher G. England
    • 1
  • Dawei Jiang
    • 2
    • 3
  • Emily B. Ehlerding
    • 1
  • Brian T. Rekoske
    • 4
  • Paul A. Ellison
    • 1
  • Reinier Hernandez
    • 1
  • Todd E. Barnhart
    • 1
  • Douglas G. McNeel
    • 4
    • 5
  • Peng Huang
    • 2
    Email author
  • Weibo Cai
    • 1
    • 3
    • 5
    Email author
  1. 1.Department of Medical PhysicsUniversity of Wisconsin – MadisonMadisonUSA
  2. 2.Guangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, School of Biomedical EngineeringHealth Science Center, Shenzhen UniversityGuangzhouPeople’s Republic of China
  3. 3.Department of RadiologyUniversity of Wisconsin – MadisonMadisonUSA
  4. 4.Department of MedicineUniversity of Wisconsin – MadisonMadisonUSA
  5. 5.University of Wisconsin Carbone Cancer CenterMadisonUSA

Personalised recommendations