You are here

Article Text

Article menu
Immunotherapy biomarkers
Original research
Non-invasive PD-L1 quantification using [18F]DK222-PET imaging in cancer immunotherapy
  1. Akhilesh Mishra1,2,
  2. Kuldeep Gupta1,
  3. Dhiraj Kumar1,
  4. Gabriela Lofland1,
  5. Ajay Kumar Sharma1,
  6. Lilja B Solnes1,
  7. Steven P Rowe1,
  8. Patrick M Forde3,
  9. Martin G Pomper1,3,4,
  10. Edward W Gabrielson3,5 and
  11. Sridhar Nimmagadda1,3,4,6
  1. 1The Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University, Baltimore, Maryland, USA
  2. 2Department of Chemical & Biomolecular Engineering, Johns Hopkins University, Baltimore, Maryland, USA
  3. 3Department of Oncology, The Sidney Kimmel Comprehensive Cancer Center and the Bloomberg–Kimmel Institute for Cancer Immunotherapy, Johns Hopkins University, Baltimore, Maryland, USA
  4. 4Department of Pharmacology and Molecular Sciences, Johns Hopkins University, Baltimore, Maryland, USA
  5. 5Department of Pathology, Johns Hopkins University, Baltimore, Maryland, USA
  6. 6Department of Medicine (Clinical Pharmacology), Johns Hopkins University, Baltimore, Maryland, USA
  1. Correspondence to Dr Sridhar Nimmagadda; snimmag1{at}jhmi.edu

Abstract

Background Combination therapies that aim to improve the clinical efficacy to immune checkpoint inhibitors have led to the need for non-invasive and early pharmacodynamic biomarkers. Positron emission tomography (PET) is a promising non-invasive approach to monitoring target dynamics, and programmed death-ligand 1 (PD-L1) expression is a central component in cancer immunotherapy strategies. [18F]DK222, a peptide-based PD-L1 imaging agent, was investigated in this study using humanized mouse models to explore the relationship between PD-L1 expression and therapy-induced changes in cancer.

Methods Cell lines and xenografts derived from three non-small cell lung cancers (NSCLCs) and three urothelial carcinomas (UCs) were used to validate the specificity of [18F]DK222 for PD-L1. PET was used to quantify anti-programmed cell death protein-1 (PD-1) therapy-induced changes in PD-L1 expression in tumors with and without microsatellite instability (MSI) in humanized mice. Furthermore, [18F]DK222-PET was used to validate PD-L1 pharmacodynamics in the context of monotherapy and combination immunotherapy in humanized mice bearing A375 melanoma xenografts. PET measures of PD-L1 expression were used to establish a relationship between pathological and immunological changes. Lastly, spatial distribution analysis of [18F]DK222-PET was developed to assess the effects of different immunotherapy regimens on tumor heterogeneity.

Results [18F]DK222-PET and biodistribution studies in mice with NSCLC and UC xenografts revealed high but variable tumor uptake at 60 min that correlated with PD-L1 expression. In MSI tumors treated with anti-PD-1, [18F]DK222 uptake was higher than in control tumors. Moreover, [18F]DK222 uptake was higher in A375 tumors treated with combination therapy compared with monotherapy, and negatively correlated with final tumor volumes. In addition, a higher number of PD-L1+ cells and higher CD8+-to-CD4+ cell ratio was observed with combination therapy compared with monotherapy, and positively correlated with PET. Furthermore, spatial distribution analysis showed higher [18F]DK222 uptake towards the core of the tumors in combination therapy, indicating a more robust and distinct pattern of immune cell infiltration.

Conclusion [18F]DK222-PET has potential as a non-invasive tool for monitoring the effects of immunotherapy on tumors. It was able to detect variable PD-L1 expression in tumors of different cancer types and quantify therapy-induced changes in tumors. Moreover, [18F]DK222-PET was able to differentiate the impact of different therapies on tumors.

  • PET
  • Immune Checkpoint Inhibitors
  • Programmed Cell Death 1 Receptor
  • Non-Small Cell Lung Cancer
  • CTLA-4 Antigen

Data availability statement

Data are available upon reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/jitc-2023-007535

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • While cancer immunotherapy with immune checkpoint inhibitors (ICI) has advanced survival rates, it remains effective for only a subset of patients. There is a pressing need to precisely forecast therapeutic results and overall patient survival subsequent to ICI treatments. A common occurrence post-ICI therapy is the surge in programmed death-ligand 1 (PD-L1) expression at the tumor site due to immune cell infiltration. Yet, there is a scarcity of tools designed to track these dynamic changes in PD-L1 expression, both prior to and amidst ICI therapy.

    WHAT THIS STUDY ADDS

    • The findings of this research show the effectiveness of using [18F]DK222, a peptide, in positron emission tomography (PET) imaging to accurately measure the varying levels of PD-L1 expression across different types of tumors and changes induced by therapy. Additionally, the study uncovered that conducting analyses on [18F]DK222-PET images to observe spatial distribution patterns in PD-L1 expression could provide valuable insights into contrasting impacts between combination immunotherapy and monotherapy on tumors.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • [18F]DK222-PET demonstrates promise as a successful and non-invasive method for monitoring PD-L1 changes during cancer immunotherapy. With further confirmation, [18F]DK222-PET has the potential to provide valuable insights for clinical decision-making and contribute significantly to the progress of cancer immunotherapy research and application.

    Background

    Cancer immunotherapy using immune checkpoint inhibitors (ICI) has significantly improved survival across tumor types including melanoma, urothelial carcinoma (UC), colorectal cancer and non-small cell lung cancer (NSCLC).1 Monoclonal antibodies (mAbs) against cytotoxic T lymphocyte antigen 4 (CTLA-4), programmed cell death protein-1 (PD-1) and its ligand (PD-L1) form the backbone of those therapeutics in current practice.2 However, only 20–40% of patients receive benefits from such ICIs and current diagnostic techniques do not accurately predict response to the treatment.3 4 To enhance the efficacy, nearly 5,000 therapeutic combinations are currently being investigated in clinical trials with anti-PD-(L)1 ICIs.5 6 The determination of efficacy for these therapies takes many years and may use immune-related radiographic response criteria, with near-term pharmacodynamic endpoints rarely being used.

    Mechanistically, most ICI therapies incorporate modulation of PD-L1, making PD-L1 expression a crucial factor in their efficacy.1 7 Also, tumors with microsatellite instability (MSI) are frequently treated with anti-PD-1 mAbs and exhibit characteristic influx of immune cells and high PD-L1 positivity.8 PD-L1 positivity by immunohistochemistry (IHC) is a Food and Drug Administration approved diagnostic companion to screen patients that would benefit from ICI therapies (specifically those targeted against PD-1/PD-L1), but challenges remain for evaluating dynamic changes in target expression.9 Therefore, there is a need for more robust tools that can non-invasively and serially monitor PD-L1 expression in tumors, including metastases throughout the body, given the dynamic nature of PD-L1 expression.10

    Evaluation of PD-L1 expression by functional imaging techniques such as positron emission tomography (PET) enables non-invasive measurements throughout the body, addressing the inherent challenges posed by tissue-based IHC assays.11 Early clinical trials with [89Zr]atezolizumab, a PD-L1 binding antibody, have demonstrated the numerous benefits of PD-L1 PET over tissue-based detection methods.12 13 Furthermore, we and others have developed peptide and small protein-based imaging agents that provide high contrast PD-L1 specific images within 60–120 min, offering an advantage over antibody-based agents.14 ,15 16 Unlike antibodies, those smaller agents may detect therapy-induced changes in PD-L1 expression, serving as an early indicator of increased immune cell activity throughout the course of ICI therapies.11 17 Among the peptide-based imaging agents is [18F]DK222, human PD-L1 specific imaging agent developed in our laboratory (online supplemental figure S1).18 The potential of [18F]DK222-PET to detect variable PD-L1 expression throughout treatment is yet to be explored fully; however, preclinical data suggest an accurate and non-invasive quantification of PD-L1 expression.19

    Supplemental material

    Here, we investigated the correlation between PD-L1 expression and therapy induced changes with the hypothesis that [18F]DK222-PET imaging could quantify this relationship. To test that hypothesis, we conducted experiments in multiple models, including six human NSCLC and UC cell lines and xenografts, MSI human colon xenografts in humanized mouse models, and melanoma xenografts treated with anti-CTLA-4 and anti-PD-1 therapeutics in humanized mouse models. We also investigated the relationship between [18F]DK222 uptake measures and therapy-induced tumor growth and immunological changes. Furthermore, we analyzed spatial distribution patterns in [18F]DK222 tumor uptake to gain a better understanding of the impact of monotherapy and combination therapies on the tumors.

    Results

    Evaluation of [18F]DK222 specificity to PD-L1 in NSCLC and UC xenografts

    We evaluated the specificity of [18F]DK222 to detect variable PD-L1 expression in human cancer xenografts. Our goal was to determine whether [18F]DK222 could quantify variable PD-L1 expression in a variety of tumor types. We selected multiple NSCLC and UC cell lines from the publicly available Cancer Cell Line Encyclopedia database (CCLE, https://portals.broadinstitute.org/ccle) which contains gene expression profiles of 1,457 human cell lines (online supplemental table S1). We first performed flow cytometry analysis of three selected NSCLC cell lines (NCI-H2444, NCI-H226, A549) to measure their surface PD-L1 expression.

    We observed the PD-L1 expression to be H2444>H226>A549 (figure 1A). To illustrate this relationship, in vitro binding assays were performed by incubating 1×106 cells of each cell line with [18F]DK222 for 30 min, and the results showed that highest percent of incubated activity (%IA) was taken up by H2444 (1.12±0.09), followed by H226 (0.96±0.05), and A549 showing the lowest uptake (0.62±0.11), which was consistent with their flow cytometry pattern (figure 1B). We next investigated whether those relative differences in expression could be measured in vivo by [18F]DK222-PET and conducted imaging studies in established NSCLC xenografts in NSG mice injected with ~200 µCi of [18F]DK222. [18F]DK222 PET at 60 min post-injection revealed highest uptake in H2444 tumors, followed by H226 and was lowest in A549 (figure 1C). Kidney had the highest uptake in non-specific tissues as renal clearance is a known mechanism of elimination for peptides.20 IHC analysis of the same tumor xenografts for PD-L1 showed varying PD-L1 expression supporting the PD-L1 expression detected by [18F]DK222 PET (figure 1D).

    Figure 1
    Figure 1

    Evaluation of [18F]DK222 specificity towards PD-L1 in NSCLC cancer xenografts. (A) Flow cytometry data of relative surface expression of PD-L1 in different NSCLC cell lines. (B) Incubation of cells with [18F]DK222 showing [18F]DK222 uptake to be PD-L1 expression dependent. (C) PET-CT images of NSG mice bearing A549, H226 and H2444 tumors showing PD-L1 expression dependent uptake. Mice were injected with ~200 µCi [18F]DK222 and imaged 60 min after injection. (D) PD-L1 IHC confirming PD-L1 expression in tumors harvested from corresponding mice. (E) [18F]DK222 tumor uptake and (F) tumor-to-blood and tumor-to-muscle ratios derived using ex vivo biodistribution studies in NSG mice bearing NSCLC xenografts. Mice were injected with ~20 µCi [18F]DK222, sacrificed and tissues were harvested at 60 min after injection. One-way ANOVA was used to calculate p values in B, E and F. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; IHC, immunohistochemistry; NSCLC, non-small cell lung cancer; PD-L1, programmed death-ligand 1; PET, positron emission tomography; %IA, per cent of incubated activity; %ID/cc, per cent of injected dose per volume; %ID/g, per cent of incubated dose per gram of uptake.

    To further validate these observations, we conducted biodistribution studies. We found that the highest percent of incubated dose per gram of uptake (%ID/g) was observed in H2444 tumor (mean: 10.41±2.63), followed by H226 (5.12±2.57) and lowest in A549 tumor (2.41±0.65) (figure 1E). No significant differences in uptake of [18F]DK222 were observed in various non-malignant tissues (online supplemental figure S2A). Supporting the high specificity of [18F]DK222 to PD-L1, tumor-to-muscle and tumor-to-blood ratios were highest for H2444 tumors and lowest for A549 (figure 1F). Collectively these results demonstrate that [18F]DK222 can be used to measure variable PD-L1 expression in NSCLC tumors.

    Extending our studies of [18F]DK222 specificity to other tumor types, we incubated three UC cell lines BFTC909, T24 and SCaBER with [18F]DK222. In vitro binding assay results showed uptake (%IA) in following order: BFTC909 (2.55±0.43)>T24 (0.96±0.09)>SCaBER (0.67±0.13) (figure 2A). Flow cytometry of these cell lines demonstrated that their surface PD-L1 expression correlated with in vitro uptake results (figure 2B). To validate [18F]DK222 specificity in vivo, we performed PET imaging in the established UC xenografts in NSG mice. [18F]DK222 uptake was found to be highest in BFTC909 tumors, followed by T24 and lowest in SCaBER (figure 2C). IHC of the extracted tumors showed the highest PD-L1 intensity in BFTC909 and the lowest in SCaBER, confirming [18F]DK222 uptake in PET images as PD-L1-specific (figure 2D). We then conducted biodistribution studies to validate those imaging observations and observed the highest %ID/g in BFTC909 tumors (mean: 12.58±4.38), followed by T24 (6.16±3.28) and lowest in SCaBER (1.38±0.38) (figure 2E). All other tissues had non-significant differences in uptake of [18F]DK222 (online supplemental figure S2B). Tumor-to-muscle and tumor-to-blood ratios analysis on these xenografts showed [18F]DK222 uptake was PD-L1-specific, and a high contrast was observed even in low PD-L1 expressing UC tumors. Taken together with results from NSCLC xenografts, these data demonstrate that [18F]DK222 quantifies varying PD-L1 expression in different types of tumors.

    Figure 2
    Figure 2

    Evaluation of [18F]DK222 specificity towards PD-L1 in UC xenografts. (A) In vitro uptake study showed [18F]DK222 uptake to be PD-L1 expression dependent. (B) Flow cytometry data of relative surface expression of PD-L1 in different cancer cell lines. (C) PET-CT images of NSG mice bearing SCaBER, T24 and BFTC909 tumors showing PD-L1 expression dependent uptake. Mice were injected with ~200 µCi [18F]DK222 and imaged 60 min after injection. (D) IHC analysis showing PD-L1 relevant immunoreactivity in tumors harvested from corresponding mice. (E) [18F]DK222 tumor uptake and (F) tumor-to-blood and tumor-to-muscle ratios derived using biodistribution studies in NSG mice bearing UC xenografts. Mice were injected with ~20 µCi [18F]DK222, sacrificed and tissues were harvested at 60 min after injection. One-way ANOVA was used to calculate p values in B, E and F. *p<0.05, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; IHC, immunohistochemistry; PD-L1, programmed death-ligand 1; PET, positron emission tomography; UC, urothelial carcinoma; %IA, per cent of incubated activity; %ID/cc, per cent of injected dose per volume; %ID/g, per cent of incubated dose per gram of uptake.

    [18F]DK222 PET quantifies therapy induced PD-L1 expression in MSI and MSS colorectal cancer xenografts

    Given the specificity of [18F]DK222 to quantify varying expression of PD-L1 in vivo, we next investigated its potential to quantify differential PD-L1 expression in colorectal cancer xenografts with MSI. Tumors with DNA mismatch repair enzyme deficiencies exhibit MSI and are characterized by extensive immune cell infiltration, which forms the basis for successful immune checkpoint therapy in these tumors.21 To evaluate the effects of MSI status on [18F]DK222 uptake, we used known MSI and microsatellite stable (MSS) colorectal cancer xenograft models, HCT-116 and HT-29, respectively (figure 3A). These cell lines had low baseline PD-L1 expression (online supplemental figure S3A). [18F]DK222 PET showed low tumor uptake in both MSI and MSS tumors when they were inoculated in immunocompromised NSG mice (figure 3B). In contrast, there was a significant increase in [18F]DK222 uptake in both tumor types when the same tumors were inoculated in human peripheral blood mononuclear cell-engrafted NSG (huPBMC) mice (figure 3C). This divergence from the clinical observations that MSI tumors display higher baseline PD-L1 expression than MSS tumors could potentially stem from the challenges in replicating the complex interplay of immune-tumor interactions within humanized mouse models. An increase in [18F]DK222 uptake was observed when huPBMC mice were treated with anti-PD-1 pembrolizumab (figure 3D). Percentage of injected dose (%ID) per volume (%ID/cc) calculated using region-of-interest analysis on MSI tumors revealed a median [18F]DK222 uptake of 2.29 %ID/cc in NSG mice (95% CI: 1.54 to 3.04), which was increased to 5.56 %ID/cc (95% CI: 3.80 to 7.31) in huPBMC mice treated with saline and further to 9.91 %ID/cc (95% CI: 8.61 to 10.37) in mice treated with anti-PD-1 therapy (figure 3E). For MSS tumors, the %ID/cc of [18F]DK222 was 2.27±0.74 in NSG mice, 7.56±1.64 in huPBMC mice and 7.64±1.35 in anti-PD-1 treated mice, indicating subdued immune cell responses following anti-PD-1 treatment. Interestingly, [18F]DK222 uptake was higher for anti-PD-1 treated MSI tumors than MSS tumors (p=0.2021), whereas it was identical for saline controls.

    Figure 3
    Figure 3

    PD-L1 upregulation correlates with [18F]DK222 uptake in MSI HCT-116 tumors. (A) Schematic showing treatment and imaging schedule. NSG mice or huPBMC mice were inoculated with MSI (HCT-116, right) and MSS (HT-29, left) colorectal cancer xenografts. Mice were treated with anti-PD-1 pembrolizumab (4 mg/kg) or saline on day 4, 6 and 8 after tumor inoculation. PET-CT fused snapshots of mice administered with [18F]DK222 in (B) NSG mice, (C) huPBMC mice treated with saline or (D) huPBMC mice treated with pembrolizumab. (E) [18F]DK222 uptake measured using image analysis represented as %ID/cc. Percentage of (F) PD-L1+ and (G) CD8+ cells as measured by flow cytometry. NSG: Immunodeficient mice treated with anti-PD-1 treatment and saline, Saline: huPBMC mice treated with saline, anti-PD-1: huPBMC mice treated with anti-PD-1. P values by one-way ANOVA. ANOVA, analysis of variance; aPD-1, anti-programmed cell death protein-1; d, day; h, hour; huPBMC, human peripheral blood mononuclear cell-engrafted NSG mice; i.v., intravenous; MSI, microsatellite instability; MSS, microsatellite stable; PD-L1, programmed death-ligand 1; PET, positron emission tomography; %ID/cc, per cent of injected dose per volume.

    To characterize the cells involved in that PD-L1 upregulation, we conducted flow cytometry analysis of extracted tumors. The data revealed that basal mean percentage of PD-L1+ cells in MSI tumors engrafted in NSG mice was 1.19%, which increased to 7.45% when tumors were engrafted in huPBMC mice and to 13.77% on treatment with anti-PD-1 (figure 3F). For MSS tumors PD-L1+ cells in tumors engrafted in NSG mice comprised 1.73% of tumor cells, which increased to 6.27% in huPBMC mice. In contrast to results seen with MSI tumors, treatment with anti-PD-1 did not significantly increase PD-L1 positive cells in MSS tumors, which remained at 6.16%. The PD-L1 positivity pattern observed in flow cytometry analyses resembled that seen with [18F]DK222 PET (figure 3B–D). Further, we observed similar pattern of increase in CD8+ T cells within MSI tumors, which was 0.12% in tumors of NSG mice, 2.74% of huPBMC mice and 5.18% in tumors of huPBMC mice treated with anti-PD-1 therapy (figure 3G). For MSS tumors, we found that CD8+ cells were 0.08% in tumors of NSG mice, 2.42% of huPBMC mice and 2.55% in tumors of huPBMC mice treated with anti-PD-1 therapy. Anti-PD-1 therapy induced CD8+ cells were significantly lower in MSS tumors compared with MSI. No differences were seen between MSI and MSS tumors for CD45, CD3 and CD4 subsets of cells (online supplemental figure S3B). Tumor growth was significantly (p=0.0351) reduced in huPBMC mice compared with NSG. Median volume of tumors in the anti-PD-1 cohort was smaller, but not statistically significant, compared with saline controls, irrespective of MSI status. (online supplemental figure S3C). IHC analysis on tumor sections showed similar pattern to [18F]DK222 PET (online supplemental figure S3D). MSI tumors treated with aPD-1 had higher PD-L1+ cells/mm2 than saline controls (783.1±336.9 vs 495.6±326.3, p=0.0551). On the contrary, MSS tumors had no significant difference in PD-L1+ cells between aPD-1 and saline treated group (410.2±359 vs 402.7±362.4, p=0.9978).

    [18F]DK222 PET quantifies immunotherapy-induced PD-L1 expression and correlates with pathological and immunological changes

    Combination immunotherapies include two or more therapeutics that target different molecular or cellular mechanisms with the goal of acting synergistically to enhance the immune response against cancer. To investigate the ability of [18F]DK222 to differentiate the pharmacodynamic responses of combination immunotherapies targeting distinct pathways in melanoma, we inoculated A375 tumors in NSG mice humanized with huCD34+ cells that robustly mimic the human immune system and treated them with either anti-PD-1 (pembrolizumab) or anti-PD-1 + anti-CTLA-4 (ipilimumab). [18F]DK222-PET was acquired before and 7 days after treatment (figure 4A). [18F]DK222-PET showed that mice treated with combination therapy had a substantial increase in tumor uptake from baseline, and that the change in [18F]DK222 (%ID/cc) was significantly higher in the group that received combination compared with monotherapy (2.90±1.89 vs 0.47±0.41; p=0.045) (figure 4B,C). Although tumor growth measurements were not the primary objective of this study, a delay in tumor growth was observed with combination when compared with monotherapy (figure 4D). Final tumor volumes negatively correlated with [18F]DK222 uptake, irrespective of treatment regimen, with two of the mice with the least tumor growth in the group receiving combination therapy exhibiting the highest increase in tumor uptake of [18F]DK222 compared with baseline (pretreatment). (R2=0.33, figure 4E). Those imaging and tumor growth changes were corroborated by an increase in PD-L1+ tumor cells and immune cell infiltration in tumors in combination treatment mice compared with monotherapy (figure 4F, online supplemental figures S4A and S4B). A higher number of immune cells per milligram of tumor (1,320 vs 230, p=0.032), PD-L1+cells (6,005 vs 2,739/mg of tumor, p=0.2557) and a higher ratio of CD8+ to CD4+ cells (1.94 vs 0.47, p<0.0001) were observed in mice treated with combination therapy compared with monotherapy (figure 4G).

    Figure 4
    Figure 4

    [18F]DK222 uptake negatively correlates with tumor growth. (A) Schematic of the experiment. (B) Representative PET images of mice receiving either 4 mg/kg pembrolizumab (3×/week; monotherapy) or 4 mg/kg pembrolizumab+4 mg/kg ipilimumab (3×/week; combination therapy). Mice treated with combination therapy had increased [18F]DK222 uptake. (C) ROI analysis on tumors showing a difference in pretreatment and post-treatment [18F]DK222 uptake between monotherapy and combination- therapy. (D) Representative tumor growth curves showing reduced tumor growth in mice with high [18F]DK222 uptake. (E) Increase in [18F]DK222 uptake negatively correlated with tumor growth. (F) Representative flow cytometry data from respective mice showing increased immune cell infiltration and upregulated PD-L1 expression. (G) Flow cytometry data showed increased immune cell infiltration, PD-L1 upregulation and CD8/CD4 ratio in combination treatment mice compared with monotherapy. *p<0.05, ***p<0.001 by Student’s t-test. aCTLA-1, anti-cytotoxic T lymphocyte antigen 4; aPD-1, anti-programmed cell death protein-1; PD-L1, programmed death-ligand 1; PET, positron emission tomography; ROI, region-of-interest; s.c., subcutaneous; %ID/cc, per cent of injected dose per volume.

    [18F]DK222 PET as a highly sensitive tool to measure total PD-L1 expression

    In order to reinforce the specificity [18F]DK222 for PD-L1, we investigated whether the uptake of [18F]DK222 correlates with the presence of PD-L1-positive cells across various tumor types and treatment regimens used in this study. We analyzed the correlation for different experiments and found that [18F]DK222 uptake strongly correlated with PD-L1 expression (R2>0.8) for any tumor type in these experiments, regardless of the treatment regimen (figure 5). Additionally, we combined human cancer xenografts from various tumor types, including NSCLC, UC, colorectal cancer and melanoma, and evaluated the relationship in different treatment settings. Across all data sets, [18F]DK222 displayed a robust correlation with total PD-L1 expression in tumors with an R2 of 0.7895 as demonstrated in online supplemental figure S5A, indicating that these PET measurements could be reflective of combined positive score, which is increasingly used as a biomarker in several cancers.22 We also observed that [18F]DK222 correlated better with PD-L1 on cancer cells than immune cells (online supplemental figures S5B,C). This could be due to the absence of myeloid cells, which harbor high PD-L1 expression, in humanized mice models.

    Figure 5
    Figure 5

    [18F]DK222 uptake correlates with PD-L1 expression. Correlations between flow cytometry derived PD-L1 expression and [18F]DK222 %ID/cc derived using PET-CT of (A) NSCLC (n=7) and UC (n=6) xenografts in immunocompromised NSG mice, (B) colorectal cancer xenografts (n=28) in NSG mice (with or without huPBMC cells) and (C) melanoma cancer A375 xenografts in huCD34 mice on anti-PD-1 monotherapy (mono-ICT) or anti-PD-1+anti-CTLA-4 (combo-ICT) (n=6). Simple linear regression and Pearson coefficient were used. huPBMC, human peripheral blood mononuclear cell-engrafted NSG mice; MFI, mean fluorescence intensity; NSCLC, non-small cell lung cancer; PD-1, programmed cell death protein-1; PD-L1, programmed death-ligand 1; PET, positron emission tomography; UC, urothelial carcinoma; %ID/cc, per cent of injected dose per volume.

    Spatial distribution of [18F]DK222 uptake in tumors differentiates the impacts of combination therapy versus monotherapy

    Differential PD-L1 upregulation seen with monotherapy and combination therapy led us to investigate whether there are differences in the spatial distribution profile of [18F]DK222 standardized uptake value (SUVs) in tumors. To assess those changes, we visualized the intensity of [18F]DK222 distribution across the tumor as a measure of PD-L1 expression and observed that tumors before treatment had low [18F]DK222 SUVs (<2), indicating low PD-L1 expression (figure 6A; left panels). Furthermore, we found the majority of the [18F]DK222 signal was localized at the tumor periphery before treatment, suggesting that PD-L1 was expressed primarily on the tumor edges (online supplemental figure S6A,B; left panels). In contrast, we observed a significant increase in [18F]DK222 SUVs in the tumors following immunotherapy, indicating an upregulation of PD-L1, particularly in the core of the tumors (figure 6A; right panels). Furthermore, [18F]DK222 SUVs were higher for combination than that for monotherapy, indicating variation of therapy induced PD-L1 expression by different treatments (online supplemental figures S6A,B; right panels).

    Figure 6
    Figure 6

    Spatial distribution analysis of [18F]DK222 uptake characterizes PD-L1 upregulation on treatment with different immunotherapy regimens. (A) Representative SUV values across the central slice of tumors showing increased expression of PD-L1 as measured by [18F]DK222 uptake before and after immunotherapy. (B) [18F]DK222 SUV characterized in the core (<3 mm) and the periphery (>3 mm) of the tumors. (C) Plot of [18F]DK222 SUV versus distance (mm) showing increased upregulation of PD-L1 in greater proportion of the tumor and the tumor core in combination therapy treated animals than monotherapy. ****p<0.0001 by one-way ANOVA. aCTLA-1, anti-cytotoxic T lymphocyte antigen 4; aPD-1, anti-programmed cell death protein-1; ANOVA, analysis of variance; PD-L1, programmed death-ligand 1; SUV, standardized uptake value.

    We then conducted a quantitative analysis of the spatial distribution of PD-L1 expression within tumors that had undergone monotherapy or combination therapy. To accomplish this, we measured [18F]DK222 uptake at the periphery and core of the tumors. The core was defined as points within a radius of 3 mm from the center of the tumor, and we quantified [18F]DK222 SUV as a function of distance from the core. Here we observed that at the tumor periphery, [18F]DK222 SUVs increased from 0.84±0.28 in the pretreatment group to 1.19±0.49 (p<0.0001) in monotherapy group, and to 2.38±0.75 in the group receiving combination therapy (p<0.0001) (figure 6C; left). In the core, the anti-PD-1 monotherapy group exhibited an SUV of 1.92±0.30, while the group receiving combination therapy had the higher PD-L1 upregulation with an [18F]DK222 SUV of 4.34±0.65 (figure 6C; right). IHC analysis of tumor sections confirmed these findings (online supplemental figure S7A). There was no statistical difference (p>0.95) in PD-L1 immunoreactivity between edges of tumors treated with monotherapy or combination therapy (online supplemental figure S7B). However, significantly higher PD-L1 positivity (cells/mm2) was seen in cores of combination therapy mice compared with monotherapy ones (32.29±22.13 vs 17.34±14.7, p=0.0089) (online supplemental figure S7C). These results indicate that anti-PD-1 + anti-CTLA-4 combination therapy induces more extensive immune responses that reach a greater proportion of the tumor and the tumor core than monotherapy. The results also suggest that [18F]DK222-PET may have the potential to distinguish between the effects of different immune modulating therapies on the tumor.

    Discussion

    Despite tremendous advances seen with immunotherapy for cancer, few non-invasive molecular imaging tools are available to monitor its effects while fitting within standard clinical workflow, which has been built around fluorodeoxyglucose-PET. PD-L1 has emerged as an important target for non-invasive monitoring, given that PD-L1 expression is used as a companion and complementary diagnostic to predict efficacy to various immune checkpoint inhibitors and many immunotherapy agents affect PD-L1 expression.23 24 In this study, we validated the ability of PD-L1 PET to detect varying expression of PD-L1 using a peptide-based PD-L1 imaging agent, [18F]DK222. Our results showed that [18F]DK222-PET detected variable PD-L1 expression in different types of tumors as well as therapy-induced changes to those expressions. We also observed that [18F]DK222-PET signal correlated with therapy-induced pathological and immunological changes in tumors. Furthermore, by analyzing the spatial distribution profile of [18F]DK222-PET signal, we gained insight into how combination-therapy induces a different impact on tumors compared with monotherapy. The results suggest that [18F]DK222-PET has potential as a non-invasive tool for monitoring cancer immunotherapy.

    ICIs have significantly improved the prognosis of patients with cancer, however, their effectiveness varies widely across patients and cancer types.25 26 To optimize the benefits of ICIs, tissue-based analysis of markers like PD-L1 variation, MSI and tumor mutational burden are approved for guiding certain ICI therapies.27 IHC-based tests have limitations such as the inability to capture the dynamic changes that occur in multiple organs during therapy and the difficulty in obtaining samples to account for intra-patient and inter-patient heterogeneity.28 In this context, PD-L1 imaging agents have emerged as a promising tool for capturing PD-L1 heterogeneity but agents that can be used to detect early changes in PD-L1 as a pharmacodynamic response are few.12 29 30 Here, we have demonstrated that [18F]DK222-PET can be used to detect dynamic changes in PD-L1 early during treatment and relate to pathological and immunological changes. By providing a non-invasive, whole-body imaging approach, PD-L1 PET can generate more accurate and dynamic data set on tumor response to ICIs and help overcome many limitations of tissue-based tests. These results demonstrate the potential of [18F]DK222 to measure PD-L1 expression non-invasively, suggesting clinical utility in detecting therapy induced PD-L1 expression in MSI and MSS tumors.

    Although the expression of PD-L1 in tumors is linked to a favorable response to ICT, the predictive value is constrained by its failure to encompass geographic and temporal heterogeneity.21 There is growing evidence, notably in melanoma and other forms of cancer, that an increase in PD-L1 during the early stages of treatment is a strong indicator of ICT’s success.31 Imaging agents such as [18F]DK222, capable of rapidly producing high-contrast images, could be instrumental in capturing these evolving immune responses. Employing PET imaging instead of biopsies may provide critical insights to guide clinical decisions. Also, it is worth noting that variations in PD-L1 levels are frequently observed, both within a single lesion over time and across different lesions or anatomical sites in a single patient.32 Such inconsistencies, where the degree of PD-L1 positivity in one lesion does not foretell the same in another, have far-reaching therapeutic consequences.26 Unlike PD-L1 IHC, which typically relies on a single tumor specimen per patient,28 PD-L1 PET could reveal the heterogeneous nature of PD-L1 status and dynamics throughout the body, thereby giving a broader view of the tumor immune response to ICT.12 The location of the lesion also appears to be a key factor in immune response or its absence.33 These findings imply that a more profound comprehension of tissue-specific immunoregulation is vital for developing effective immunotherapies against tumors and metastases. This deeper understanding could be facilitated by measuring PD-L1 levels repeatedly using PET imaging.

    Although antibodies, antibody conjugates and small proteins have been used to detect PD-L1 expression non-invasively, peptide-based radiotracers provide several unique advantages. 16 ,14 18 The low molecular weight of DK222 facilitates enhanced tissue penetration. Moreover, its radiolabeling with the short half-life isotope, radiofluoride-18 (~110 min), minimizes radiation exposure. The 60-min time from injection to imaging of [18F]DK222-PET allows multiple imaging sessions with the same subject, capturing therapy-induced PD-L1 dynamics. We have demonstrated this advantage with a combination of anti-PD-1 and anti-PD1+anti-CTLA4 therapies in experimental models of colorectal cancer and melanoma. Additionally, repeat imaging with agents such as [18F]DK222 could be highly relevant to understand the effects of new therapeutics and therapeutic combinations due to its potential to provide valuable pharmacodynamic data. Perhaps the most straightforward situation would be to quantify target engagement, that is, by imaging before and after the administration of anti-PD-L1 therapy to identify the degree to which PD-L1 is no longer available for binding by the imaging agent. These peptide-based imaging agents have a unique binding mode, which enables the assessment of anti-PD-L1 drug exposure and target engagement in the tumors without disrupting therapy.19 More broadly, serial imaging for establishing response assessment would likely be the mostly widely applicable use of these imaging agents. Response assessment may be difficult in the complex milieu of the tumor microenvironment, where it may be difficult to demonstrate direct correlations between uptake and response metrics. However, more arcane metrics can be derived from targeted radiotracers that may represent surrogate response markers.34

    PD-L1 expression is influenced by various factors such as interferons and toll-like receptor ligands, which are released by immune cells after radiotherapy, chemotherapy, and/or targeted therapy.7 However, the current use of PD-L1 expression to determine eligibility for ICIs does not consider the heterogeneity in PD-L1 expression across different therapies, which may not be reflected in the initial PD-L1 positivity of tumors. To understand better how different therapies modulate the immune cell infiltration status in the tumor microenvironment, we used the spatial distribution of PD-L1 as a surrogate marker, measured by [18F]DK222-PET. Our findings revealed that this approach can differentiate various immunotherapy treatments by revealing distinct patterns in the spatial distribution of [18F]DK222. Specifically, we observed that a combination of anti-PD-1 and anti-CTLA-4 therapy led to more extensive immune responses that reached a greater proportion of the tumor and the tumor core than monotherapy. As the number of strategies for combination therapy increases, assessing the heterogeneity of PD-L1 expression longitudinally using IHC assays becomes challenging. Longitudinal [18F]DK222-PET may provide a non-invasive alternative. Additionally, analyzing the spatial distribution of the PET signals can offer early insights into the activity of these therapeutics within the tumor microenvironment. Overall, [18F]DK222-PET can quantify dynamic changes in PD-L1 expression induced during combination immunotherapy and is related to pathological and immunological changes, which could be useful in personalized treatment strategies and in early drug development.

    In summary, the findings from these studies suggest the potential utility of [18F]DK222-PET in tracking PD-L1 during treatment. Even though it is not currently in clinical use, tools like [18F]DK222-PET could offer valuable insights into changes in PD-L1 throughout therapy, correlating them to observed responses. This could be advantageous in monitoring immunotherapy in patients with cancer. First-in-human evaluation of [18F]DK222 is underway.

    Materials and methods

    Chemicals

    DK221 was custom synthesized by CPC Scientific (Sunnyvale, California, USA) with >95% purity. (2,2′-(7-(4-isothiocyanatobenzyl)−1,4,7-triazonane-1,4-diyl) diacetic acid) (NCS-MP-NODA) was purchased from CheMatech Macrocycle Design Technologies (catalog # C110; Dijon, France). All other chemicals were purchased from Sigma-Aldrich or Fisher Scientific.

    Cell culture reagents and antibodies

    All cell culture reagents were purchased from Invitrogen (Grand Island, New York, USA). Pembrolizumab (anti-PD-1) and ipilimumab (anti-CTLA-4) were purchased from Johns Hopkins School of Medicine Pharmacy.

    Synthesis of DK222

    DK222 was synthesized as described previously.18 Briefly, NODA conjugation was performed by adding Diisopropylethylamine (5.0 µL) to the precursor molecule, DK221 (4.0 mg, 2.04 µmoles) in Dimethylformamide (1.0 mL), followed by NCS-MP-NODA (1.6 mg, 4.07 µmoles). The reaction mixture was stirred for 4 hours at room temperature and purified on a reversed phase high performance liquid chromatography (RP-HPLC) system using a semi-preparative C-18 Luna column (5 mm, 10×250 mm Phenomenex, Torrance, California, USA). The HPLC conditions for purification were 50–90% methanol (0.1% trifluoroacetic acid) in 30 min at a flow rate of 5 mL/min with H2O (0.1% trifluoroacetic acid) as cosolvent. The desired DK222 was collected at 15.5 min and characterized by matrix assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS).

    [18F]DK222 radiopharmaceutical preparation

    [18F]DK222 was synthesized as described previously.18 Briefly, carrier-free [18F]fluoride obtained from JHU PET Center cyclotron was trapped on a preconditioned Chromafix 30-PS-HCO3 catridge and washed with 5 mL metal-free water. [18F]F was eluted using 100 uL of 0.4 M KHCO3. pH was adjusted to 4.0 using glacial acetic acid. This [18F]F further equilibrated with 20 uL 2 mM AlCl3.6H2O in 0.1 M NaOH buffer, followed by 2–3 min incubation at room temperature to form Al[18F] complex. The precursor DK222 (~ 100 μg, 42 nmol) was dissolved in 300 µL NaOAc (0.1M, pH 4) and added to the vial containing Al[18F]. The resulting reaction mixture was heated at 110°C for 15 min. The radiolabeled product was purified on a RP-HPLC system (Varian ProStar) with an Agilent Technology 1260 Infinity photodiode array detector (Agilent Technologies, Wilmington, Delaware, USA). A semi-preparative C-18 Luna column (5 mm, 10×250 mm Phenomenex, Torrance, California, USA) was used for purification as described above. The radiolabeled product, [18F]DK222, eluted at a retention time of ~16.2 min was collected, evaporated under high vacuum, formulated with saline containing 5% EtOH, sterile filtered, and used for in vitro and in vivo evaluation.

    Cell culture

    All cell lines were purchased from American Type Culture Collection and cultured in the recommended media in an incubator at 37°C in a humid atmosphere containing 5% CO2. H2444 and H226 were maintained in RPMI-1640 medium. BFTC909 and SCaBER were maintained in Dulbecco’s Modified Eagle’s Medium. HCT-116, HT-29 and T24 cells were maintained in McCoy’s 5A medium. A-549 cells were maintained in Ham’s F-12 medium. All cells were supplemented with 10% fetal bovine serum, 1% Penicillin-Streptomycin. All cell lines were authenticated using short tandem repeat profiling. All cell lines were routinely tested for mycoplasma and used within 10–12 passages after thawing.

    Detection of PD-L1 expression by flow cytometry

    Adherent cells were detached using enzyme free cell dissociation buffer (Thermo Fisher Scientific, Waltham, Massachusetts, USA). PD-L1 surface expression was evaluated by direct staining of 1×106 cells in 100 µL flow cytometry (FACS) buffer (Phosphate-Buffered Saline with 0.1% FBS and 2 mM EDTA) with phycoerythrin labeled anti-PD-L1 antibody (clone: MIH1 BD # 558065), for 30 min at 4°C. Cells were then washed and analyzed for mean fluorescence intensity by flow cytometry.

    In vitro binding assays with [18F]DK222

    In vitro binding of [18F]DK222 to H2444, H226, A549, BFTC909, T24 and SCaBER cells was determined by incubating 1×106 cells with approximately 1 µCi of [18F]DK222 for 30 min at 4°C. After incubation, cells were washed three times with ice cold PBS containing 0.1% Tween20 and counted on an automated gamma counter. All cell radioactivity uptake studies were performed in quadruplicate for each cell line and repeated three times.

    Mouse strains and in vivo studies

    All mouse studies were conducted through Johns Hopkins University Animal Care and Use Committee (ACUC) approved protocols. Protocol M021M175 was approved by Brendan J. Canning (Chair, ACUC, 1-877-932-6675). Xenografts were established in 5–6 weeks old, male (for NSCLC/UC tumors) or female (for melanoma and colorectal), non-obese, diabetic, severe-combined immunodeficient gamma (NSG) mice obtained from the Johns Hopkins University Immune Compromised Animal Core. Female huPBMC mice were purchased from Jackson (JAX) laboratories and used for experiments as-is. Male huCD34 mice were purchased from JAX laboratories and used for experiments as-is.

    Xenograft models

    Mice were injected with cancer cells subcutaneously in 100 µL PBS (top right flank, unless otherwise noted) for all tumor models. Following cell numbers were used: H2444 (5 M), H226 (2 M), A-549 (2 M), BFTC909 (1 M), T24(3 M), SCaBER (5 M), HCT-116 (1 M), HT-29 (1 M). For anti-PD-1 dosing experiment, MSIhigh HCT-116 and MSS HT-29 were inoculated in the right and left flank of mice, respectively. Mice with tumor volumes of 100–200 mm3 were used for treatment, imaging, and biodistribution experiments. A minimum of four mice were used for all biodistribution studies. A375 (2M) was inoculated in huCD34 mice. Mice with tumor volumes of 60–100 mm3 were used for treatment (n=4–5/group). One mouse from each group had to be sacrificed for ethical considerations.

    PET-CT imaging of mouse xenografts

    Mice with tumors of size 100–200 mm3 were injected with ~200 µCi (7.4 MBq) of [18F]DK222 in 200 µL of 5% ethanol in saline intravenously and anesthetized under 2.5% isoflurane. PET images were acquired at 60 min after radiotracer injection at 5 min/bed in an ARGUS small-animal PET/CT scanner (Sedecal, Madrid, Spain) as described. The PET data were reconstructed using the two-dimensional ordered subsets-expectation maximization algorithm and corrected for radioactive decay and dead time. The %ID per cc values were calculated based on a calibration factor obtained from a known radioactive quantity. For change in PD-L1 expression following monotherapy or combo-ICT therapy, Δ%ID/cc was calculated by subtracting baseline (pretreatment) uptake from post-therapy uptake of [18F]DK222. Image fusion, visualization, and three-dimensional (3D) rendering were accomplished using Amira V.6.1 (FEI, Hillsboro, Oregon, USA).

    PET data derived PD-L1 spatial distribution analysis

    Co-registered PET and CT NIFTI-47 (.img) files were manually converted to nii files (n=5 for each monotherapy and combination therapy). RStudio package oro.nifti was used to import PET-CT data as data-frames. CT files were used to locate tumors and find the plane having the largest size of tumor and tumor boundary by visual inspection. This location was used to derive [18F]DK222 activity concentration from co-registered PET nii files. Average tumor size was used to divide the tumor into two halves at 3 mm distance from the center. SUV values were calculated by normalizing for net injected activity and mouse weight. RStudio package tidyverse was used to transform data in the manner it could be used by other packages. RStudio packages plotly and plot3D were used to plot2D histograms and 3D plots, respectively. RStudio package imager was used to export SUV and distance from center values from R as comma separated values to be used to plot in GraphPad Prism.

    Ex vivo biodistribution

    To validate imaging studies, ex vivo biodistribution studies were conducted in mice with tumors of size 100–200 mm3. Mice received ~50 µCi (1.85 MBq) of [18F]DK222 in 200 µL of 5% ethanol in saline intravenously and biodistribution studies were conducted at 60 min after [18F]DK222 injection. Selected tissues (tumors, blood, heart, lung, liver, stomach. spleen, pancreas, kidney, small intestines, testicles and muscle) were collected, weighed, counted, and their %ID/g values calculated as described previously.35

    Anti-PD-1 mAb dosing studies

    Female huPBMC and NSG mice were implanted subcutaneously with 1×106 HCT-116 (MSIhigh, right) or 1×106 HT-29 (MSS) cells in the rostral end. On days 6, 8 and 10 after cell inoculation (average tumor volume=75±15 mm3), mice were randomized and treated with three doses of 4 mg/kg dose of pembrolizumab (48 hours apart) injected intravenously (n=8/group) or with saline (n=4–6/group). NSG mice treated with, similar regimen as huPBMC mice, of pembrolizumab were also used as controls. Twenty-four hours following the last dose, [18F]DK222 PET scans were acquired on at least two mice per group. Median tumor volume: 140.15 (IQR: 72.6–210.81) for MSI (HCT-116) and 56.64 (51.67–131.27) for MSS (HT-29). For monotherapy, mice bearing A375 tumors were treated with either 4 mg/kg anti- PD-1; (pembrolizumab) or 4 mg/kg of anti-PD-1 + 4 mg/kg of anti-CTLA-4 (ipilimumab) (n=5) and [18F]DK222 PET was acquired before and 7 days after start of treatment. NSG mice with A375 tumors were used as control.

    Immunohistochemistry

    Formalin-fixed, paraffin-embedded tissue sections were soaked in xylene to remove paraffin and then rehydrated through incubations in xylene (3×5 min), 100% ethanol (2×5 min), 95% ethanol (2×5 min), 80% ethanol (2×5 min) and H2O (1×5 min). Antigen retrieval was performed by heating the sections in pH 8.5 EDTA buffer in a decloaking chamber for 20 min. Endogenous peroxide activity was blocked with 3% hydrogen peroxide for 10 min followed by blocking with 10% goat serum for 1 hour, and then incubated with a primary anti-human PD-L1 antibody (Clone: E1L3N, #13684, Cell Signaling Technology) at 1:200 dilution at 4°C overnight. After washing with PBS, the secondary antibody, Signalstain Boost IHC Detection Reagent (HRP), was applied and incubated for 30 min at room temperature. The slides were washed and developed using ImmPACT DAB substrate. After washing, the slides were counterstained with Mayer’s Hematoxylin for 1 min, dehydrated using alcohol and xylene, and then the cover slipped. For IHC staining quantification for PD-L1 in monotherapy versus combination therapy of A375 tumors, 3 mm distance was chosen from the periphery to define the edge of the tumors. Points further inside were used as core tumor fields. Necrosis region was ignored when selecting fields. All analyses were carried out in QuPath V.0.3.0.

    Flow cytometry analysis

    Harvested tumors were analyzed for PD-L1 expression and immune cell infiltrates. Briefly, tumors were dissociated following manufacturer’s instructions (Miltenyi Biotec catalog# 130-095-929). Each xenograft was cut into small pieces of 3–4 mm and cut pieces were suspended in 2.5 mL RPMI-1640 media containing 100 µL Enzyme H, 50 µL Enzyme R, and 12.5 µL Enzyme A. Recommended programs were run on gentleMACS Octo Dissociator with Heaters (Miltenyi Biotec #130-096-427). A short centrifugation step was performed to collect the sample material at the bottom of the tube. Samples were resuspended and passed through 70 µm strainers, centrifuged 300×g for 7 min, supernatant was discarded, and cells were resuspended in 2.5 mL FACS buffer. Cells were counted and 1×106 cells were resuspended in 100 µL LIVE/DEAD aqua solution (Thermo Fisher #L34965, 2 µL reconstituted with 2 mL PBS) in a 96-well plate. The cells were incubated for 15 min (dark, room temperature) and washed with 150 µL PBS. Fc blocking was performed with BioLegend Tru Stain (#422 301 Fc Block (1 µL in 100 µL FACS buffer)) and samples were incubated for 10 min (dark, 4°C). After washing with 150 µL cold FACS buffer, the samples were stained with mAbs targeting markers of interest in following dilutions in 100 µL FACS buffer:

    View this table:

    Data analyses was performed using FlowJo (BD). At least 100,000 live cells were acquired for each sample and different cell populations were reported as percentage of live cells.

    Statistical analysis

    All statistical analyses were performed using Prism V.9.0 Software (GraphPad Software, La Jolla, California, USA). Unpaired Student’s t-test, one-way or two-way analysis of variance were used for column, multiple column, and grouped analyses, respectively. P values<0.05 were considered statistically significant. Correlation was done using simple linear regression without keeping constant term zero.

    Data availability statement

    Data are available upon reasonable request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Not applicable.

    Acknowledgments

    We thank the Johns Hopkins University PET Center for 18F production.

    References

    1. ↵
      1. Topalian SL,
      2. Taube JM,
      3. Pardoll DM
      . Neoadjuvant Checkpoint blockade for cancer Immunotherapy. Science 2020;367:eaax0182. doi:10.1126/science.aax0182
    2. ↵
      1. Huang AC,
      2. Zappasodi R
      . A decade of Checkpoint blockade Immunotherapy in Melanoma: understanding the molecular basis for immune sensitivity and resistance. Nat Immunol 2022;23:660–70. doi:10.1038/s41590-022-01141-1
      OpenUrlCrossRefPubMed
    3. ↵
      1. Haslam A,
      2. Prasad V
      . Estimation of the percentage of us patients with cancer who are eligible for and respond to Checkpoint inhibitor Immunotherapy drugs. JAMA Netw Open 2019;2:e192535. doi:10.1001/jamanetworkopen.2019.2535
    4. ↵
      1. Pang K,
      2. Shi Z-D,
      3. Wei L-Y, et al
      . Research progress of therapeutic effects and drug resistance of Immunotherapy based on PD-1/PD-L1 blockade. Drug Resist Updat 2023;66:100907. doi:10.1016/j.drup.2022.100907
      OpenUrl
    5. ↵
      1. Tang J,
      2. Pearce L,
      3. O’donnell-Tormey J, et al
      . Erratum: trends in the global immuno-oncology landscape (nature reviews drug discovery. Nat Rev Drug Discov 2018;17:783–4. doi:10.1038/nrd.2018.167
      OpenUrlCrossRefPubMed
    6. ↵
      1. Upadhaya S,
      2. Neftelino ST,
      3. Hodge JP, et al
      . Combinations take centre stage in Pd1/Pdl1 inhibitor clinical trials. Nat Rev Drug Discov 2021;20:168–9. doi:10.1038/d41573-020-00204-y
      OpenUrlPubMed
    7. ↵
      1. Sun C,
      2. Mezzadra R,
      3. Schumacher TN
      . Regulation and function of the PD-L1 checkpoint. Immunity 2018;48:434–52. doi:10.1016/j.immuni.2018.03.014
      OpenUrlCrossRefPubMed
    8. ↵
      1. Diaz LA,
      2. Le DT
      . PD-1 blockade in tumors with mismatch-repair deficiency. N Engl J Med 2015;373:1979. doi:10.1056/NEJMc1510353
    9. ↵
      1. Büttner R,
      2. Gosney JR,
      3. Skov BG, et al
      . Programmed death-ligand 1 immunohistochemistry testing: a review of analytical assays and clinical implementation in non-small-cell lung cancer. J Clin Oncol 2017;35:3867–76. doi:10.1200/JCO.2017.74.7642
      OpenUrlCrossRefPubMed
    10. ↵
      1. McLaughlin J,
      2. Han G,
      3. Schalper KA, et al
      . Quantitative assessment of the heterogeneity of PD-L1 expression in non-small-cell lung cancer. JAMA Oncol 2016;2:46–54. doi:10.1001/jamaoncol.2015.3638
      OpenUrl
    11. ↵
      1. Nimmagadda S
      . Quantifying PD-L1 expression to monitor immune Checkpoint therapy: opportunities and challenges. Cancers (Basel) 2020;12:3173. doi:10.3390/cancers12113173
    12. ↵
      1. Bensch F,
      2. van der Veen EL,
      3. Lub-de Hooge MN, et al
      . 89Zr-Atezolizumab imaging as a non-invasive approach to assess clinical response to PD-L1 blockade in cancer. Nat Med 2018;24:1852–8. doi:10.1038/s41591-018-0255-8
      OpenUrlPubMed
    13. ↵
      1. Heskamp S,
      2. Hobo W,
      3. Molkenboer-Kuenen JDM, et al
      . Noninvasive imaging of tumor PD-L1 expression using radiolabeled anti-PD-L1 antibodies. Cancer Res 2015;75:2928–36. doi:10.1158/0008-5472.CAN-14-3477
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Niemeijer AN,
      2. Leung D,
      3. Huisman MC, et al
      . Whole body PD-1 and PD-L1 positron emission tomography in patients with non-small-cell lung cancer. Nat Commun 2018;9:4664.:4664. doi:10.1038/s41467-018-07131-y
      OpenUrlPubMed
    15. ↵
      1. Ge S,
      2. Jia T,
      3. Li J, et al
      . Molecular imaging of immune checkpoints in oncology: current and future applications. Cancer Lett 2022;548:215896. doi:10.1016/j.canlet.2022.215896
      OpenUrl
    16. ↵
      1. Chatterjee S,
      2. Lesniak WG,
      3. Miller MS, et al
      . Rapid PD-L1 detection in tumors with PET using a highly specific peptide. Biochemical and Biophysical Research Communications 2017;483:258–63.:S0006-291X(16)32216-1. doi:10.1016/j.bbrc.2016.12.156
      OpenUrlPubMed
    17. ↵
      1. Zhou H,
      2. Bao G,
      3. Wang Z, et al
      . PET imaging of an Optimized anti-PD-L1 probe 68Ga-NODAGA-Bms986192 in immunocompetent mice and non-human primates. EJNMMI Res 2022;12:35. doi:10.1186/s13550-022-00906-x
    18. ↵
      1. Kumar D,
      2. Mishra A,
      3. Lisok A, et al
      . Pharmacodynamic measures within tumors expose differential activity of PD(L)-1 antibody Therapeutics. Proc Natl Acad Sci U S A 2021;118:e2107982118. doi:10.1073/pnas.2107982118
    19. ↵
      1. Kumar D,
      2. Lisok A,
      3. Dahmane E, et al
      . Peptide-based PET quantifies target engagement of PD-L1 therapeutics. J Clin Invest 2019;129:616–30. doi:10.1172/JCI122216
      OpenUrlCrossRefPubMed
    20. ↵
      1. Katz AI,
      2. Emmanouel DS
      . Metabolism of polypeptide hormones by the normal kidney and in uremia. Nephron 1978;22:69–80. doi:10.1159/000181425
      OpenUrlPubMed
    21. ↵
      1. Llosa NJ,
      2. Cruise M,
      3. Tam A, et al
      . The vigorous immune microenvironment of microsatellite Instable colon cancer is balanced by multiple counter-inhibitory checkpoints. Cancer Discov 2015;5:43–51. doi:10.1158/2159-8290.CD-14-0863
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Vuky J,
      2. Balar AV,
      3. Castellano D, et al
      . Long-term outcomes in KEYNOTE-052: phase II study investigating first-line Pembrolizumab in cisplatin-ineligible patients with locally advanced or metastatic urothelial cancer. J Clin Oncol 2020;38:2658–66. doi:10.1200/JCO.19.01213
      OpenUrlCrossRefPubMed
    23. ↵
      1. Yamaguchi H,
      2. Hsu JM,
      3. Yang WH, et al
      . Mechanisms regulating PD-L1 expression in cancers and associated opportunities for novel small-molecule therapeutics. Nat Rev Clin Oncol 2022;19:287–305. doi:10.1038/s41571-022-00601-9
      OpenUrl
    24. ↵
      1. Doroshow DB,
      2. Bhalla S,
      3. Beasley MB, et al
      . PD-L1 as a biomarker of response to immune-checkpoint inhibitors. Nat Rev Clin Oncol 2021;18:345–62. doi:10.1038/s41571-021-00473-5
      OpenUrlPubMed
    25. ↵
      1. Forde PM,
      2. Spicer J,
      3. Girard N
      . Neoadjuvant Nivolumab plus chemotherapy in resectable lung cancer. N Engl J Med 2022;387:572–3. doi:10.1056/NEJMc2208133
      OpenUrl
    26. ↵
      1. Topalian SL,
      2. Hodi FS,
      3. Brahmer JR, et al
      . Five-year survival and correlates among patients with advanced melanoma, renal cell carcinoma, or non-small cell lung cancer treated with nivolumab. JAMA Oncol 2019;5:1411–20. doi:10.1001/jamaoncol.2019.2187
      OpenUrl
    27. ↵
      1. Diaz LA,
      2. Shiu K-K,
      3. Kim T-W, et al
      . Pembrolizumab versus chemotherapy for Microsatellite instability-high or mismatch repair-deficient metastatic colorectal cancer (KEYNOTE-177): final analysis of a randomised, open-label, phase 3 study. Lancet Oncol 2022;23:659–70. doi:10.1016/S1470-2045(22)00197-8
      OpenUrlCrossRef
    28. ↵
      1. Ilie M,
      2. Long-Mira E,
      3. Bence C, et al
      . Comparative study of the PD-L1 status between surgically resected specimens and matched biopsies of NSCLC patients reveal major discordances: a potential issue for anti-PD-L1 therapeutic strategies. Annals of Oncology 2016;27:147–53. doi:10.1093/annonc/mdv489
      OpenUrlCrossRefPubMed
    29. ↵
      1. Xing Y,
      2. Chand G,
      3. Liu C, et al
      . Early phase I study of a 99Mtc-labeled anti-programmed death ligand-1 (PD-L1) single-domain antibody in SPECT/CT assessment of PD-L1 expression in non-small cell lung cancer. J Nucl Med 2019;60:1213–20. doi:10.2967/jnumed.118.224170
      OpenUrlAbstract/FREE Full Text
    30. ↵
      1. Zhou X,
      2. Jiang J,
      3. Yang X, et al
      . First-in-humans evaluation of a PD-L1–binding peptide PET Radiotracer in non–small cell lung cancer patients. J Nucl Med 2022;63:536–42. doi:10.2967/jnumed.121.262045
      OpenUrlAbstract/FREE Full Text
    31. ↵
      1. Vilain RE,
      2. Menzies AM,
      3. Wilmott JS, et al
      . Dynamic changes in PD-L1 expression and immune infiltrates early during treatment predict response to PD-1 blockade in melanoma. Clin Cancer Res 2017;23:5024–33. doi:10.1158/1078-0432.CCR-16-0698
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Cercek A,
      2. Lumish M,
      3. Sinopoli J, et al
      . PD-1 blockade in mismatch repair–deficient, locally advanced rectal cancer. N Engl J Med 2022;386:2363–76. doi:10.1056/NEJMoa2201445
      OpenUrlCrossRefPubMed
    33. ↵
      1. Long GV,
      2. Atkinson V,
      3. Lo S, et al
      . Combination nivolumab and ipilimumab or nivolumab alone in melanoma brain metastases: a multicentre randomised phase 2 study. Lancet Oncol 2018;19:672–81. doi:10.1016/S1470-2045(18)30139-6
      OpenUrlCrossRefPubMed
    34. ↵
      1. Zukotynski KA,
      2. Emmenegger U,
      3. Hotte S, et al
      . Prospective, single-arm trial evaluating changes in uptake patterns on prostate-specific membrane antigen (PSMA)-Targeted 18F-Dcfpyl PET/CT in patients with castration-resistant prostate cancer starting abiraterone or enzalutamide. J Nucl Med 2021;62:1430–7. doi:10.2967/jnumed.120.259069
      OpenUrl
    35. ↵
      1. Chatterjee S,
      2. Lesniak WG,
      3. Gabrielson M, et al
      . A humanized antibody for imaging immune checkpoint ligand PD-L1 expression in tumors. Oncotarget 2016;7:10215–27. doi:10.18632/oncotarget.7143
      OpenUrlPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Twitter @shajay0412

    • Contributors AM: Conceptualization, data curation, formal analysis, interpretation of results, writing—original draft, writing—review and editing. KG: Data curation, formal analysis, interpretation of results, writing—original draft, writing—review and editing. DK, AKS, GL: Data curation, writing—review and editing. LBS, SPR, PMF, MGP: interpretation of the results, writing—review and editing. EWG: Data curation, methodology, writing—review and editing. SN: Funding acquisition, methodology, conceptualization, data curation, writing—original draft, writing—review and editing. SN accepted full responsibility for the study, had access to all the data and the final decision to submit for publication.

    • Funding Grant funding for this study was provided NIH1R01CA236616, NIH R01CA269235 and NIH P41EB024495. Core resources (flow cytometry, histology, and imaging) were supported by NIH P30CA006973.

    • Competing interests SN, MGP and DK are co-inventors on pending patents covering [18F]DK222, and as such are entitled to a portion of any licensing fees and royalties generated by this technology. This arrangement was reviewed and approved by Johns Hopkins University in accordance with its conflict-of-interest policies. SN, MGP and SPR received funding and were consultants for Precision Molecular, Inc., the licensee of [18F]DK222. SN, MGP and SPR have equity in D&D Pharmatech, the parent company of Precision Molecular, Inc. All other authors declare no conflicts pertaining to the described work.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.