Immunotherapy is innovating clinical cancer management. Nevertheless, only a small fraction of patient's benefit from current immunotherapies. To improve clinical management of cancer immunotherapy, it is critical to develop strategies for response monitoring and prediction. In this study, we describe inducible T-cell costimulator (ICOS) as a conserved mediator of immune response across multiple therapy strategies. ICOS expression was evaluated by flow cytometry, 89Zr-DFO-ICOS mAb PET/CT imaging was performed on Lewis lung cancer models treated with different immunotherapy strategies, and the change in tumor volume was used as a read-out for therapeutic response. ImmunoPET imaging of ICOS enabled sensitive and specific detection of activated T cells and early benchmarking of immune response. A STING (stimulator of interferon genes) agonist was identified as a promising therapeutic approach in this manner. The STING agonist generated significantly stronger immune responses as measured by ICOS ImmunoPET and delayed tumor growth compared with programmed death-1 checkpoint blockade. More importantly, ICOS ImmunoPET enabled early and robust prediction of therapeutic response across multiple treatment regimens. These data show that ICOS is an indicator of T-cell–mediated immune response and suggests ICOS ImmunoPET as a promising strategy for monitoring, comparing, and predicting immunotherapy success in cancer.

Significance:

ICOS ImmunoPET is a promising strategy to noninvasively predict and monitor immunotherapy response.

See related commentary by Choyke, p. 2975

The field of cancer therapy has benefited greatly from recent breakthroughs in immuno-oncology, most notably the role of the programmed death-1 (PD-1) signaling axis in mediating immunosuppression. PD-1 and programmed death ligand-1 (PD-L1) immune checkpoint inhibitors have been approved by the FDA for the treatment of many cancers, including non—small cell lung cancer (NSCLC) and have shown promising outcomes in clinical trials (1). Generally, these immunotherapy agents activate the host immune system to attack and destroy malignant cells. Due to disease heterogeneity and complex tumor escape mechanisms, the response rates of checkpoint inhibitors in NSCLC vary from 19% to 45% across different phases of clinical trials (1–3). To improve clinical outcomes for patients diagnosed with cancer, there is an urgent need for tools that enable the assessment of new immunotherapies and accurate monitoring of patient response.

The current gold standards for therapy assessment and response monitoring in the clinic are RECIST and immune-related response criteria (irRC) based on CT and MRI (4, 5). These imaging modalities provide remarkably detailed anatomic information, but fail to capture molecular information or underlying immune response. In fact, it has been well documented that these criteria often fail in the immuno-oncology setting due to pseudoprogression caused by tumor immune cell infiltration (6, 7). Another limitation of these approaches is the long lag between treatment initiation and response assessment. Typically, at least 9–12 weeks pass before following up and assessment of therapeutic efficacy.

Due to the shortcomings of these conventional approaches, research efforts have intensified to identify robust biomarkers of response to immunotherapy. Tumor mutational load, gene expression profiling, T-cell repertoire sequencing, and PD-L1 IHC represent several examples of novel biomarkers that have been reported to show efficacy in predicting patient response to immune checkpoint blockade (8–11). Despite the promise of these approaches, they all require invasive biopsies. A recent report examining heterogeneity in patients with NSCLC demonstrated that these classes of metrics were inconsistent among repeated biopsies taken from 17 of 42 tumors in the study (12).

PET represents a theoretically ideal solution. PET has the ability to dynamically and noninvasively quantify molecular information from the whole human body. The most classic PET radiotracer, 2-deoxy-2-(18F)fluoro-d-glucose (18F-FDG), has been widely used in lung cancer diagnosis and patient management in the clinic. Some have even proposed 18F-FDG for monitoring immunotherapy response in clinical trials (13, 14), but the biggest challenge is the lack of specificity. Because 18F-FDG reports on aberrant metabolic processes, it detects both tumors and proliferative immune cells and it has been technically difficult to deconvolve the signals arising from the two scenarios. Other metabolic radiotracers such as 3′-deoxy-3′[18F]-fluorothymidine (18F-FLT), 1-(2′-deoxy-2′-[18F]fluoro-β-d-arabinofuranosyl) cytosine (18F-AraC), and 2′-deoxy-2′-[18F]fluoro-9-β-d-arabinofuranosyl guanine (18F-AraG) have all faced the same issue as 18F-FDG when applied in the immuno-oncology setting (15, 16).

ImmunoPET tracers specifically targeting immune cell biomarkers have the potential to sensitively assess patient-specific immune response. The first reports of PD-1 and PD-L1 ImmunoPET imaging in mice and humans have demonstrated the technique to be feasible and safe in predicting patient response to immune checkpoint blockade (17–19). In fact, PD-L1 ImmunoPET significantly outperformed PD-L1–based IHC and RNA measurements in the small cohort of patients tested thus far (20). Additional ImmunoPET tracers in development are geared for more general use for imaging a variety of immune cell phenotypes including CD3, CD4, CD8, and CD20 (21–24). While these tracers have been demonstrated to accurately quantify immune cell type and distribution within the body, they have failed in most settings to predict or correlate with immunotherapy response at early timepoints.

To predict response at the earliest stages, biomarkers of immune cell activation are requisite. Upon successful immunotherapy, immune cells are activated and then home to the tumor tissue where they can exert their effector function. The secreted markers, IFNγ and granzyme β (25, 26), have shown promising results both for response monitoring and assessment of therapeutic strength. However, these secreted biomarkers pose challenges for imaging including half-life and dilution. These particular biomarkers also fundamentally represent later stages of activation, once a T cell has already reached the tumor and begun to kill. In a previous study, we have successfully demonstrated that OX40, a T-cell surface costimulatory receptor that is only activated upon antigen-specific recognition, represents a compelling paradigm for detecting the earliest stages of activation in the tumor-draining lymph nodes (TDLN), all the way through to the late stages in the tumor, of T-cell–mediated immune response (27).

Here we investigate inducible T-cell costimulatory receptor (ICOS or CD278), a member of the CD28 superfamily, as a new potential biomarker of T-cell–mediated immunotherapy response. As reported before, ICOS is mainly expressed on activated cytotoxic T cells, memory T cells, and regulatory T cells. The initiation of the ICOS pathway begins through ligation of ICOS and its ligand (ICOSL), which is expressed on B cells, macrophages, and dendritic cells. Ligation triggers a downstream pathway that regulates T-cell proliferation and survival, as well as secretion of IL4, IL10, and IFN (28, 29). OX40 and ICOS cooperate in a nonredundant manner to maximize and sustain Th cell immune responses (30). Although OX40 expression has been reported on both CD4 and CD8 cell subsets, in most models it is skewed toward CD4 expression. In contrast, we demonstrate here in a model of Lewis lung cancer that ICOS is highly upregulated on both activated CD4 and CD8 T-cell subsets and may represent an even more sensitive detection paradigm. Based on this compelling evidence, we developed 89Zr-DFO-ICOS mAb as a means to noninvasively quantify activated CD4 and CD8 T-cell responses targeting lung cancer. Utilizing this approach, we gain insights into the lack of therapeutic efficacy of PD-1 immune checkpoint inhibitors in a model of Lewis lung carcinoma, as well as identify STING (stimulator of interferon genes) agonists as a novel and potent treatment strategy. Overall, we demonstrate that ICOS ImmunoPET is a promising strategy for predicting and monitoring T-cell–mediated immune response to cancer immunotherapy.

Study design

The aim of this study was to evaluate ICOS as a candidate biomarker for early prediction and monitoring of immunotherapy response based on noninvasive ImmunoPET imaging. ICOS expression on both resting and activated phorbol 12-myristate 13-acetate/ionomycin (PMA/IONO) T cells was first tested by FACS, then 89Zr-DFO-ICOS mAb was synthesized, followed by characterization and cell uptake studies. For animal experiments, Lewis lung carcinoma LLC1 cell line (obtained from ATCC) and C57BL/6J (Jackson Lab) were used for establishment of the mouse xenograft models; Lewis lung cancer cell line was cultured in T75 flask and in DMEM media (Thermo Fisher Scientific) supplemented with 10% FBS (Thermo Fisher Scientific) and 1% antibiotic-antimycotic, when grows to 90% confluence, cells will be passaged, to avoid the contamination of Mycoplasma, Mycoplasma testing was performed once every month, no extra cell authentication was conducted. PBS, 200 μg anti-mouse PD-1 antibody, 50 μg 2′3′-cGAMP (STING agonist) or 2′3′-cGAMP plus 200 μg anti-mouse PD-1 antibody were administered intraperitoneally (i.p.) or intratumorally (i.t.) to randomized cohorts of mice-bearing Lewis lung tumors. Tumor volume was measured every other day after treatment initiation. 89Zr-DFO-ICOS mAb PET imaging was used for ICOS detection after treatment. A separate group of treated mice was analyzed by FACS to verify ICOS expression levels [FACS antibodies: anti-mouse CD4(APC/cy7, clone: GK1.5 BioLegend), anti-mouse CD8 (APC, clone: 53-6.7, BioLegend), anti-human/rat/mouse ICOS(PerCP/Cy5.5, clone: 398.4A, BioLegend), anti-mouse OX40 (PE, clone: OX-86, BioLegend), anti-mouse CD25 (PE/Cy7, clone: PC61, BioLegend), anti-mouse PD-1 (BV711, clone: 29F.1A12, BioLegend), anti-mouse Foxp3 (Pacific blue, clone: MF-14 BioLegend), anti-human CD3 (FITC, clone: HIT3a, BD Pharmingen), and Live/Dead (LIVE/DEAD Fixable Aqua Dead Cell Stain Kit, Thermo Fisher Scientific)]. All reported data are representative of at least two independent experiments.

PMA/IONO activation and FACS analysis of ICOS expression on both naïve and activated T cells

C57BL/6J mouse spleens were collected after euthanasia. Mouse T-cell isolation was performed according to the manufacturer's protocol (EasySep Mouse T-cell Isolation Kit). T cells were seeded in a 96-well plate at a density of 200,000 cells/well. To confirm ICOS expression on human-activated T cells, human peripheral blood mononuclear cells (PBMC) were seeded in a 96-well plate, 200,000/well. For T-cell activation, PMA and IONO mixture was added to each well at a final concentration of 10 and 100 ng/mL, respectively. Both resting and activated T cells were collected after 72 hours of activation. FACS staining of mouse CD4 (APC/cy7, BioLegend), mouse CD8 (APC, BioLegend), human/rat/mouse ICOS(PerCP/Cy5.5, BioLegend), human CD3 (FITC, BD Pharmingen), and Live/Dead(LIVE/DEAD Fixable Aqua Dead Cell Stain Kit, Thermo Fisher Scientific) were performed and ICOS expression was analyzed by flow cytometry (BD LSRII). Compensation was performed in all experiments utilizing AbC Total Antibody Compensation Bead Kit (Thermo Fisher Scientific).

DFO conjugation

For desferoxamine (DFO) conjugation, ICOS mAb (clone: 7E.17G9), or isotype control (clone: LTF-2) and DFO (p-SCN-Bn-Deferoxamine) were purchased from Bio-X-cell and Macrocyclics, respectively. 1 mg of the unconjugated mAb was maintained in 1 mL PBS (pH = 7.4), then buffer exchanged with pH 8.8–9.0 PBS solution. After the recovery, 10-fold excess DFO was added to mAb solution. After 1-hour incubation under 37°C, the mixture was buffer exchanged/washed by PBS (pH = 7.4) using vivaspin2, 50K cut-off centrifugal concentrator (three washes, 4,000 g, 9 minutes/wash). About 200 μL DFO-ICOS mAb or isotype control solution was recovered, and the concentration was determined by Thermo Fisher Scientific NanoDrop One Microvolume UV-Vis Spectrophotometer.

89Zr-DFO-ICOS mAb radiosynthesis, AF594-ICOS mAb conjugation

89Zr-oxalic acid was adjusted to a pH range of 7.1–7.8 by adding 1 mol/L Na2CO3 solution and subsequently added to DFO-ICOS mAb stock solution. The mixture was incubated under 37°C for 1 hour, then purified by Zeba Spin Desalting Columns, 7K MWCO, 0.5 mL, 1,000 g, 1 minute. To test the stability of the tracer, 50 μCi of 89Zr-DFO-ICOS mAb was added to 1 mL PBS or 1 mL mouse serum, and kept in a shaker under 37°C. The radiochemical purity was tested at different time points. Alexa Fluor 594 NHS ester (Invitrogen, Thermo Fisher Scientific) was dissolved in DMSO to a final concentration of 10 μg/μL, followed by addition to 200 μg ICOS mAb in PBS. The solution was incubated at room temperature for 1 hour, and purified by vivaspin2, 50K cut-off centrifugal concentrator.

Cell uptake study

For cell uptake studies, 3 μCi 89Zr-DFO-ICOS mAb were added to 500,000K PMA/IONO activated, resting or blocked (activated T-cell treated by 100 μg cold ICOS mAb) T cells, and incubated for 1 hour under 37°C in Iscove Modified Dulbecco Media. After three washes with 200 μL PBS, cells were collected and the activity accumulated in cells was determined by a gamma counter. For AF594-ICOS mAb cell uptake, the probe was firstly diluted to a certain concentration from 10−9 to 10−6 mol/L, and added to 200,000K PMA/IONO activated, resting or blocked (activated T cell treated by 100 μg cold ICOS mAb) T cells, and incubated for 1 hour under 37°C. After three washes with 200 μL PBS, cells were collected, stained for viability, and tested for fluorophore probe binding by flow cytometry.

Animal model establishment and treatment

All research involving animal subjects was approved by the Administrative Panel on Laboratory Animal Care. 5 × 105 Lewis cells in 50 μL PBS were injected subcutaneously in the right shoulder of 6–8 weeks old female C57BL/6 mice. After 7–9 days of inoculation, when the tumor volume reached a range of 40–100 mm3, mice received either PBS, 200 μg anti-mouse PD-1 antibody, 50 μg 2′3′-cGAMP VacciGrade (STING i.t.) or 50 μg 2′3′-cGAMP VacciGrade plus 200 μg anti-mouse PD-1 antibody (Combo) via i.t. or i.p. injection. The original day of treatment was denoted as day 0, and the treatment was given on day 0, 2, 4, and/or 6. Tumor volumes were recorded every other day by calipers and tumor volumes were calculated using the formula (pi/6) × length × width × height.

PET/CT imaging and biodistribution study

All PET scans were performed on Siemens Inveon MM-PET/CT. After anesthetization by 1.5%–2% isoflurane gas, 50 μCi (1 μCi/μL) 89Zr-DFO-ICOS mAb or isotype control was administered via mouse tail vein on day 0 right after the treatment. On days 1, 2, 5, and day7/8, a 15-minute static PET scan was acquired right after CT imaging, which was used for providing an anatomic reference and PET signal attenuation correction. Region of interest (ROI) of PET images was drawn using a three dimensional (3D) volume mode. No partial volume correction was applied. For validation of the ROI data, after the last PET scan on day 7/8, Lewis lung cancer bearing mice were euthanized, and blood, tumor, tumor-draining lymph node, spleen, heart, liver, left kidney, small intestine, large intestine, lung, and muscle were collected and weighed. An automatic gamma counter was used for determining the activity in different organs. Both ROI and biodistribution data were normalized to %ID/g, while tumor-draining lymph nodes' biodistribution data were normalized to %ID.

Statistical analysis

All the data analysis was performed on PRISM 5 (GraphPad) and R studio. One- or two-way ANOVA and unpaired two-tailed student t test were used for data analysis where appropriate. P values less than 0.05 were considered as statistically significant. R square and P value were used for evaluation of 2D or 3D linear regression fits.

Development of an ICOS probe for sensitive and specific detection of activated CD4 and CD8 T cells

To confirm reports of ICOS as a candidate biomarker for activated T cells, we performed PMA/IONO activation assays on both murine T cells and human PBMCs, followed by antibody staining 72 hours later. Flow cytometry revealed ICOS to be strongly upregulated on activated compared with resting murine T cells (Fig. 1A and B left, P < 0.0001). Importantly, the same statistically significant result was observed for human PBMCs (Fig. 1B right, P < 0.0001). We next synthesized fluorescent and radiolabeled probes for sensitive and specific ICOS detection. Our candidate ICOS probe was successfully labeled via amine conjugation of an AlexaFluor 594 NHS ester dye (31). We tested the probe to ensure the labeling did not interfere with the active binding site of the antibody scaffold. In an ICOS binding assay utilizing serial dilutions of the probe, we observed AF594-ICOS mAb to have significantly higher uptake in activated T cells compared with ICOS mAb blocked or resting controls and a limit of detection on the order of 10−6 (P < 0.001) to 10−7 (P < 0.01) molar (Fig. 1C). The radioactive version of the probe was also generated via an amine conjugation strategy, this time utilizing a DFO macrocyclic chelator and 4-isothiocyanatophenyl cross-linking. Zirconium89 was selected as the optimal radiometal due to its favorable half-life (78.4 hours) pairing with mAbs and its increasing clinical usage. To confirm sensitive and specific binding of the radioactive ICOS probe, cell uptake studies were performed. Significantly higher uptake of 89Zr-DFO-ICOS mAb was observed for activated T cells as compared with ICOS blocked or resting controls (Fig. 1D, P < 0.0001). Radiolabeling resulted in high radiochemical yield (>70%), purity (>99%), and specific activity (∼28 GBq/μmol) determined from averaging four validation runs (Fig. 1E; Supplementary Table S1). The radioactive probe remained stable in both serum and PBS, measured out to 6 days (Supplementary Fig. S1).

Figure 1.

ICOS is a sensitive and specific biomarker expressed on activated T cells. A, 3D flow cytometry plots of ICOS expression on activated and resting murine CD4 and CD8 T cells. Scale bar, green, high ICOS expression; dark purple, low ICOS expression. B, Mean fluorescence intensity (MFI) quantification of ICOS expression on activated and resting T cells. Left, murine T cells; right, human T cells; n = 3. C, Uptake of AF594-ICOS mAb in PMA/IONO activated, blocked, and resting murine T cells at 1-hour incubation; activated, blocked, and unstained, n = 3; naïve, n = 2; statistical comparison shown between activated and blocked groups. D, Cell uptake of 89Zr-DFO-ICOS mAb in PMA/IONO activated, blocked, and resting murine T cells at 1-hour incubation, n ≥ 3; statistical comparison shown between activated and blocked/resting groups. E, Radio-iTLC of 89Zr-DFO-ICOS mAb and free 89Zr, mobile phase: 50 mmol/L EDTA. All values represent the mean ± SD unless otherwise specified. Unpaired two-tailed Student t test was used for analyses. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 1.

ICOS is a sensitive and specific biomarker expressed on activated T cells. A, 3D flow cytometry plots of ICOS expression on activated and resting murine CD4 and CD8 T cells. Scale bar, green, high ICOS expression; dark purple, low ICOS expression. B, Mean fluorescence intensity (MFI) quantification of ICOS expression on activated and resting T cells. Left, murine T cells; right, human T cells; n = 3. C, Uptake of AF594-ICOS mAb in PMA/IONO activated, blocked, and resting murine T cells at 1-hour incubation; activated, blocked, and unstained, n = 3; naïve, n = 2; statistical comparison shown between activated and blocked groups. D, Cell uptake of 89Zr-DFO-ICOS mAb in PMA/IONO activated, blocked, and resting murine T cells at 1-hour incubation, n ≥ 3; statistical comparison shown between activated and blocked/resting groups. E, Radio-iTLC of 89Zr-DFO-ICOS mAb and free 89Zr, mobile phase: 50 mmol/L EDTA. All values represent the mean ± SD unless otherwise specified. Unpaired two-tailed Student t test was used for analyses. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Close modal

Determination of a conserved role for ICOS+ T cells in multiple immunotherapy strategies targeting lung cancer

PD-1/PD-L1 axis blockade is an immunotherapy strategy that has been tested in various clinical trials for patients with lung cancer. To determine whether ICOS+-activated T cells are implicated in the immune response pathway to PD-1 blockade, we treated mice-bearing Lewis lung carcinoma tumors with 200 μg anti-mouse PD-1 antibody or 100 μL PBS (i.p.) on a standard treatment schedule (days 0, 2, 4, and 6). Subsequent changes in tumor volume were measured every other day (Fig. 2A). After four treatments, the PD-1 group demonstrated slightly delayed tumor growth and a lower overall tumor fold change. While the treatment effect was significant (Fig. 2B, P < 0.01, two-way ANOVA), the therapeutic result was modest (PD-1: 9.61 ± 3.65 folds increase vs. PBS: 14.73 ± 6.38 folds increase). This was in line with other preclinical and clinical studies utilizing PD-1 blockade in lung cancer (32). PD-1 blockade did generate early activated T-cell responses in lymph sites measured on day 2 after just a single treatment (Fig. 2C). The frequencies of ICOS+ CD4 and ICOS+ CD8 T cells in the TDLN were significantly higher in the PD-1 group, compared with the PBS control (P < 0.01). Significant upregulation of ICOS could also be observed on CD4 T cells in the spleen following PD-1 therapy (P < 0.05). t-Distributed Stochastic Neighbor Embedding (tSNE) analysis revealed ICOS is highly restricted to CD4 and CD8 T cells. Despite the weak overall response to PD-1 blockade, ICOS expression in the TLDN and spleen were the strongest early biological correlates of future tumor volume change, out of all the biomarkers we tested (Fig. 2D).

Figure 2.

ICOS is an early indicator of therapy response in Lewis lung cancer models and is highly restricted to T cells. A, Study design of PD-1 i.p. versus PBS i.p. study (n ≥ 5). B, Tumor volume monitoring curve of PD-1 i.p. and PBS i.p. group. C, tSNE plot and flow cytometry analysis of ICOS frequency on CD4 and CD8 T cells in tumor-draining lymph node and spleen. D, Correlogram: linear regression analysis between frequency of different biomarkers and log10 (D2 tumor volume/D0 tumor volume). Color and circle size represent strength of Pearson correlation. Scale bar, yellow, positively correlated; purple, negatively correlated. E, Study design of the STING agonist study (n ≥ 3). F, Tumor volume monitoring curve of STING i.t., PD-1 i.t., PBS i.t., and Combo groups. G, tSNE plot and FACS analysis of ICOS expression on CD4 and CD8 T cells in tumor-draining lymph node and spleen. H, Comparison of ICOS and OX40 expression on CD4 and CD8 T cells in TDLN. All values represent the mean ± SEM unless otherwise specified. Two-way ANOVA was for determination of PD-1 i.p. and PBS i.p. groups tumor volume change, and unpaired two-tailed Student t test was used for others. *, P < 0.05; **, P < 0.01; ****, P < 0.0001; ns, not significant.

Figure 2.

ICOS is an early indicator of therapy response in Lewis lung cancer models and is highly restricted to T cells. A, Study design of PD-1 i.p. versus PBS i.p. study (n ≥ 5). B, Tumor volume monitoring curve of PD-1 i.p. and PBS i.p. group. C, tSNE plot and flow cytometry analysis of ICOS frequency on CD4 and CD8 T cells in tumor-draining lymph node and spleen. D, Correlogram: linear regression analysis between frequency of different biomarkers and log10 (D2 tumor volume/D0 tumor volume). Color and circle size represent strength of Pearson correlation. Scale bar, yellow, positively correlated; purple, negatively correlated. E, Study design of the STING agonist study (n ≥ 3). F, Tumor volume monitoring curve of STING i.t., PD-1 i.t., PBS i.t., and Combo groups. G, tSNE plot and FACS analysis of ICOS expression on CD4 and CD8 T cells in tumor-draining lymph node and spleen. H, Comparison of ICOS and OX40 expression on CD4 and CD8 T cells in TDLN. All values represent the mean ± SEM unless otherwise specified. Two-way ANOVA was for determination of PD-1 i.p. and PBS i.p. groups tumor volume change, and unpaired two-tailed Student t test was used for others. *, P < 0.05; **, P < 0.01; ****, P < 0.0001; ns, not significant.

Close modal

To confirm a ubiquitous role for ICOS+ T cells in immune response to multiple therapy strategies targeting lung cancer, we selected a novel 2′3′-cGAMP (CDN) stimulator of interferon genes (STING) agonist for further investigation. This CDN ligand is known to bind directly to the endoplasmic reticulum-resident STING receptor, which is implicated in spontaneous tumor antigen recognition and induction of tumor-specific T-cell responses (33, 34). Intratumoral injection of STING agonist has shown recent success in multiple preclinical models (35–37). Here we evaluated STING agonist therapy for the first time, to our knowledge, in Lewis lung cancer. On days 0, 2, and 4, mice-bearing Lewis lung carcinoma tumors received either STING i.t., PD-1 i.t., Combo, or PBS i.t. (Fig. 2E). Tumor volume changes in response to treatment were monitored over time (Fig. 2F). Both STING i.t. and Combo groups demonstrated significantly delayed tumor growth compared with PBS i.t. or PD-1 i.t. alone. The STING arm of the combo therapy accounted for the majority of the therapeutic effect, as there was no advantage of combination therapy compared with STING alone at this time point (p = ns). FACS analyses were performed between PBS i.t. and STING i.t. groups on day 2 (Fig. 2G). ICOS expression was again restricted to T-cell subsets by tSNE analysis, STING agonist was observed to generate immune responses against lung cancer, with a higher frequency of ICOS+ CD8 T cells observed in the TDLN (P < 0.0001) and spleen (P < 0.05). We further phenotyped the ICOS+ cells observed in this model of lung cancer, and as previously reported, its expression was nonredundant with OX40 (Fig. 2H). ICOS expression was notably high on both CD4 and CD8 populations, whereas OX40 was restricted to CD4 alone. This observation is highly relevant for a potential biomarker of response to lung cancer immunotherapy, as ICOS+ CD8 cells accounted for the primary difference between STING-treated and PBS control groups.

ImmunoPET imaging of ICOS as a biomarker of adaptive immune response to lung cancer

To evaluate the ability of 89Zr-DFO-ICOS mAb to detect ICOS+-activated T cells in vivo, ImmunoPET imaging studies were performed on day 2 and 7/8 after tracer administration. To quantify the PET signals in major organs, 3D ROIs were drawn (Fig. 3A), and the PET images were shown in (Fig. 3BD; Video 1–6). Time course pharmacokinetic profiles demonstrated clearance of 89Zr-DFO-ICOS mAb from the blood pool mediated by liver uptake and subsequent hepatobiliary excretion typical of an IgG antibody (Fig. 4A and B). At day 2, significant increases in PET signal in the tumor, TDLN and spleen of the STING i.t. and Combo-treated cohorts could be observed compared with PD-1 i.t. or PBS i.t. groups (Fig. 4C). This signal could be attributed to specific increases in ICOS, determined utilizing an 89Zr-DFO-Isotype mAb control. In contrast, PD-1 intraperitoneally treated mice only demonstrated a significant increase in PET signal compared with PBS i.p. control in the TDLN alone (Fig. 4D), corresponding with our previous data suggesting PD-1 treatment leads to a weaker therapeutic immune response in the Lewis lung cancer model. By day 7/8, these trends persisted but overall signal decreases were observed at these sites, potentially linked with a waning immune response or unbound tracer clearance (Fig. 4E and F). Ex vivo biodistribution (BioD) studies confirmed the accuracy of our manually delineated ROI measurements of PET signal from the images (Supplementary Figs. S2 and S3). Good concordance was observed between PET and BioD measurements across major tissue sites (Fig. 4G; tumor: r2 = 0.8447, P < 0.0001; spleen: r2 = 0.4885, P < 0.0001). Our PET imaging studies indicate our approach enables noninvasive and specific detection of ICOS+-activated T cells in vivo.

Figure 3.

PET/CT imaging of 89Zr-DFO-ICOS mAb in Lewis lung cancer models. A, Reference of 3D ROI drawing. B, Day 2 PET/CT imaging of 89Zr-DFO-ICOS mAb in Lewis lung cancer models. C, Day 2 PET/CT imaging of 89Zr-DFO-ICOS mAb and 89Zr-DFO-isotype control in STING agonist i.t.-treated Lewis lung cancer models. D, Day 7/8 PET/CT imaging of 89Zr-DFO-ICOS mAb in Lewis lung cancer models. All images are representative of n = 3–7 mice per group.

Figure 3.

PET/CT imaging of 89Zr-DFO-ICOS mAb in Lewis lung cancer models. A, Reference of 3D ROI drawing. B, Day 2 PET/CT imaging of 89Zr-DFO-ICOS mAb in Lewis lung cancer models. C, Day 2 PET/CT imaging of 89Zr-DFO-ICOS mAb and 89Zr-DFO-isotype control in STING agonist i.t.-treated Lewis lung cancer models. D, Day 7/8 PET/CT imaging of 89Zr-DFO-ICOS mAb in Lewis lung cancer models. All images are representative of n = 3–7 mice per group.

Close modal
Figure 4.

ROI and biodistribution quantification. A and B,89Zr-DFO-ICOS mAb pharmacokinetics at all imaging time points in major organs. C and D, Day 2 PET image ROI quantification of tumor, TDLN, and spleen. E and F, Day 7/8 PET image ROI quantification of tumor, TDLN, and spleen. G, Linear regression analysis of ROI and biodistribution at day 7/8 time point. All values represent the mean ± SD unless otherwise specified. One-way ANOVA and unpaired two-tailed Student t test were used for comparation of ROI values. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, not significant.

Figure 4.

ROI and biodistribution quantification. A and B,89Zr-DFO-ICOS mAb pharmacokinetics at all imaging time points in major organs. C and D, Day 2 PET image ROI quantification of tumor, TDLN, and spleen. E and F, Day 7/8 PET image ROI quantification of tumor, TDLN, and spleen. G, Linear regression analysis of ROI and biodistribution at day 7/8 time point. All values represent the mean ± SD unless otherwise specified. One-way ANOVA and unpaired two-tailed Student t test were used for comparation of ROI values. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, not significant.

Close modal

Evaluation of ICOS imaging for immunotherapy response prediction in lung cancer

The final goal of this study was to assess the ability of 89Zr-DFO-ICOS mAb imaging to predict and monitor immunotherapy response. To identify the strongest response predictors from our image ROIs, we generated a heatmap depicting all ROIs from our early day 2 imaging timepoint and performed unsupervised hierarchical clustering. The ROI profiles distinguished mice treated with STING agonist or combination therapy from the other cohorts (Fig. 5A). Further analysis revealed the PET signal in the tumor and TDLN had the strongest correlation with tumor volume response at various timepoints (Fig. 5B). A multiple linear regression model utilizing both of these ROIs to monitor (day 7/8 PET %ID/g vs. day 7/8 log10 tumor volume mm3 fold change, r2 = 0.6604, P = 3.053e-07) or predict (day 2 PET %ID/g vs. day 7/8 log10 tumor volume mm3 fold change, r2 = 0.7675, P = 2.217e-09) tumor response yielded robust fits (Fig. 5C). This predictor fit therapy response well across therapeutic cohorts (e.g., STING vs. PD-1) and within therapy cohorts (e.g., PD-1; Supplementary Fig. S4). We conclude here that ICOS imaging is a feasible and promising strategy for monitoring and predicting immunotherapy response in lung cancer.

Figure 5.

Evaluation of ICOS imaging metrics for immunotherapy response prediction in the Lewis lung cancer model. A,Z-normalized heat map of %ID/g tracer uptake values in specified ROIs. Row/column order determined by unsupervised hierarchical clustering. Side bars represent log10 (fold-change tumor volume) at day 2 (d2) or day 7/8 (d7/8) posttherapy. B, Correlogram depicting r2 Pearson correlation between %ID/g tracer uptake measured in the indicated ROI, compared with the log10 (fold-change tumor volumes) at various time points. ±, positive or negative correlation. Circle size and color denotes Pearson correlation strength. C, 3D linear regression of log10 (fold-change tumor volumes) at day 7/8 with %ID/g tracer uptake in TDLN and tumor at day 2 (predict) or day 7/8 (monitor).

Figure 5.

Evaluation of ICOS imaging metrics for immunotherapy response prediction in the Lewis lung cancer model. A,Z-normalized heat map of %ID/g tracer uptake values in specified ROIs. Row/column order determined by unsupervised hierarchical clustering. Side bars represent log10 (fold-change tumor volume) at day 2 (d2) or day 7/8 (d7/8) posttherapy. B, Correlogram depicting r2 Pearson correlation between %ID/g tracer uptake measured in the indicated ROI, compared with the log10 (fold-change tumor volumes) at various time points. ±, positive or negative correlation. Circle size and color denotes Pearson correlation strength. C, 3D linear regression of log10 (fold-change tumor volumes) at day 7/8 with %ID/g tracer uptake in TDLN and tumor at day 2 (predict) or day 7/8 (monitor).

Close modal

Immunotherapy has brought new hope to clinical cancer treatment, but due to variance in patient response and modest efficacy of PD-1 immune checkpoint blockade in this therapeutic setting, new tools need to be explored for early assessment of immune response in cancer and identification of new treatment paradigms. We demonstrate that ICOS ImmunoPET imaging can be utilized to achieve these goals in preclinical immunotherapy models of Lewis lung cancer and suggest that there are compelling reasons to further translate and test this approach clinically based on the evidence and rationale presented here.

Imaging ICOS theoretically represents one of the earliest and most sensitive methods for detecting T-cell–mediated immune response to cancer therapy. ICOS is an extracellular activation marker that is turned on following antigen recognition and costimulatory signaling, the earliest known events in the adaptive immune response cascade. Here, ICOS imaging enabled early detection of response prior to changes in tumor volume, offering a potential clinical advantage to MR- and CT-based anatomic imaging approaches. ICOS expression was highly restricted to activated T cells and was expressed on both CD4 and CD8 subsets, making ICOS both a highly sensitive and specific biomarker. While ICOS was also expressed on regulatory T cells, these cells represented a small fraction (<10%, Supplementary Fig. S5) of the total CD4 cell population upon treatment. The change in ICOS PET signal in treated compared with control was thus primarily attributable to an increase in ICOS+-activated effector T cells. This is pertinent given the relatively low specificity for activated T cells of other PET tracers in clinical testing and development (e.g., 18F-FDG, 18F-FLT, 18F-AraC). ICOS demonstrated earlier correlation with tumor response than changes in CD4 and CD8 subsets. This observation reinforces the hypothesis that imaging T-cell activation markers may provide earlier actionable insights than T-cell phenotype markers. ICOS expression was nonredundant with other T-cell activation markers (e.g., OX40, PD-1) and was conserved across multiple immunotherapy strategies. ICOS expression could also be detected by ImmunoPET imaging in both the tumor and TDLN, elucidating the critical interplay between these sites in immunotherapy response. Previous studies of granzyme β and IFNγ imaging (25, 26) only reported signal in the tumor, potentially missing early stages of immune response detected here in the TDLN. In fact, integrating ICOS ImmunoPET signal from both the tumor and TDLN enabled the best monitoring and prediction of eventual tumor response in this preclinical model.

Our results demonstrate a proof of concept for how imaging ICOS may be utilized to streamline preclinical drug development and improve clinical patient management. Given that early ImmunoPET imaging of ICOS correlated strongly with late therapeutic volumetric tumor response in the model systems tested here, this approach has potential for rapid evaluation and benchmarking of new immunotherapy strategies. For example, in this work, we demonstrate that STING, previously not validated in the lung cancer setting, significantly outperforms PD-1 blockade in inducing immune and therapeutic responses. Long survival studies could be shortened, as these differences were evident from ImmunoPET imaging at just 2 days posttreatment initiation. A multiple linear regression model could forecast the predicted tumor growth for each therapeutic cohort many days later. Furthermore, the mechanistic insights gained from ICOS imaging into the strength of an activated immune response following a novel immunotherapy regiment are likely more informative than changes in tumor volume alone when evaluating drugs for translation. Based on the same results, ICOS imaging could improve clinical patient management by allowing the treating physician to make decisions regarding drug dose or switching drug regiments early after treatment initiation. ICOS imaging thus may lead to improved patient outcomes by enabling the identification of the optimal therapy prior to tumor progression, and potentially decrease costs associated with unnecessary therapies not contributing to patient benefit.

Imaging ICOS as presented here does have limitations. For the ex vivo activation assay, we chose the most common PMA/IONO activation assay, which is quite different from ICOS biology in vivo and can't perfectly mimic the mechanisms in vivo. While ICOS has been reported as a biomarker of response in clinical studies (38, 39) and preclinical models (35, 40), here our evaluation of ICOS as an imaging biomarker was limited to a single subcutaneous lung cancer murine xenograft model, and there were only two immunotherapy adjuvants employed. Further work is needed to explore the feasibility of ICOS ImmunoPET in the prediction and evaluation of immunotherapy response across different treatment strategies and orthotopic or spontaneous models that could better mimic the tumor microenvironment. Studies will also need to be performed to determine the best time to assess ICOS activation posttreatment. According to our longitudinal ICOS imaging studies, PET signals in the tumor and TDLN decreased slightly over time. Further FACS characterization in the anti PD-1 antibody treated group showed ICOS expression to wane on CD4 and CD8 T cells as a function of time posttreatment (Supplementary Fig. S6). Therefore, establishing the right imaging window will be critical for proper therapy response assessment. Also, ICOS+ T cells may not be implicated in all models or therapy types. It is for this reason that we are developing a comprehensive Immunoimaging toolbox (41), and evaluating the roles of multiple markers, imaging agents, and cell types, while making decisions about which candidates are the most promising for clinical translation. This study principally evaluated PD-1 and STING agonist as lung cancer therapies, because of reports that PD-1 therapy generates relatively low numbers of ICOS+ T cells (42) while STING agonist generates relatively high numbers of ICOS+ T cells (36). These were, therefore, ideal selections given our primary goal to assess the feasibility of our ICOS imaging approach to detect differences in T-cell–mediated immune response. While we identified that STING agonist approaches may be efficacious in the lung cancer setting for the first time to our knowledge, more work will need to be done to evaluate the optimal dose and therapeutic schedule.

In this proof-of-concept work, we successfully employed an antibody-based probe for PET imaging of ICOS. ImmunoPET imaging of ICOS offers a compelling approach for monitoring and predicting activated T-cell immune responses. In the age of next-generation sequencing and highly multiplexed assays, ImmunoPET imaging of a single biomarker may seem to offer a paucity of information. In opposition to this notion, ImmunoPET gives a full body view of immune response and can help guide these biopsy-dependent techniques, which often suffer from interlesional and intralesional heterogeneity. In fact, ImmunoPET imaging is just beginning to realize its full potential. ImmunoPET imaging is now entering the clinic and being employed to monitor and stratify patients for cancer immunotherapy, with recent successful reports of PD-1 and PD-L1 imaging in humans (20, 43). Our initial data suggest ICOS imaging offers advantages as a general marker of T-cell activation for monitoring immunotherapy response and merits further investigation.

No potential conflicts of interest were disclosed.

Conception and design: Z. Xiao, A.T. Mayer, S.S. Gambhir

Development of methodology: Z. Xiao, A.T. Mayer, S.S. Gambhir

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z. Xiao, A.T. Mayer, T.W. Nobashi

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z. Xiao, A.T. Mayer

Writing, review, and/or revision of the manuscript: Z. Xiao, A.T. Mayer, T.W. Nobashi, S.S. Gambhir

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.S. Gambhir

Study supervision: A.T. Mayer, S.S. Gambhir

This work was supported in part by funding from the Ben and Catherine Ivy Foundation, the Canary Foundation, and the NCI (R01 CA201719-03).

The authors would like to acknowledge Drs. Tim Doyle, Frezghi Habte, and the Stanford Center for Innovation in In-Vivo Imaging (SCI3) for their assistance with the preclinical imaging. We also thank members of the Stanford FACS facility for sharing their expertise. In addition, we thank Kenneth Lau for assistance with mass spectrometry.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Reck
M
,
Rodriguez-Abreu
D
,
Robinson
AG
,
Hui
R
,
Csoszi
T
,
Fulop
A
, et al
Pembrolizumab versus chemotherapy for PD-L1-positive non-small-cell lung cancer
.
N Engl J Med
2016
;
375
:
1823
33
.
2.
Borghaei
H
,
Paz-Ares
L
,
Horn
L
,
Spigel
DR
,
Steins
M
,
Ready
NE
, et al
Nivolumab versus docetaxel in advanced nonsquamous non-small-cell lung cancer
.
N Engl J Med
2015
;
373
:
1627
39
.
3.
Brahmer
J
,
Reckamp
KL
,
Baas
P
,
Crino
L
,
Eberhardt
WE
,
Poddubskaya
E
, et al
Nivolumab versus docetaxel in advanced squamous-cell non-small-cell lung cancer
.
N Engl J Med
2015
;
373
:
123
35
.
4.
Fournier
L
,
Ammari
S
,
Thiam
R
,
Cuenod
CA
. 
Imaging criteria for assessing tumour response: RECIST, mRECIST, Cheson
.
Diagn Interv Imaging
2014
;
95
:
689
703
.
5.
Tazdait
M
,
Mezquita
L
,
Lahmar
J
,
Ferrara
R
,
Bidault
F
,
Ammari
S
, et al
Patterns of responses in metastatic NSCLC during PD-1 or PDL-1 inhibitor therapy: comparison of RECIST 1.1, irRECIST and iRECIST criteria
.
Eur J Cancer
2018
;
88
:
38
47
.
6.
Borcoman
E
,
Nandikolla
A
,
Long
G
,
Goel
S
,
Le Tourneau
C
. 
Patterns of response and progression to immunotherapy
.
Am Soc Clin Oncol Educ Book
2018
;
38
:
169
78
.
7.
Vrankar
M
,
Unk
M
. 
Immune RECIST criteria and symptomatic pseudoprogression in non-small cell lung cancer patients treated with immunotherapy
.
Radiol Oncol
2018
;
52
:
365
9
.
8.
Samstein
RM
,
Lee
CH
,
Shoushtari
AN
,
Hellmann
MD
,
Shen
R
,
Janjigian
YY
, et al
Tumor mutational load predicts survival after immunotherapy across multiple cancer types
.
Nat Genet
2019
;
51
:
202
6
.
9.
Jamieson
NB
,
Maker
AV
. 
Gene-expression profiling to predict responsiveness to immunotherapy
.
Cancer Gene Ther
2017
;
24
:
134
40
.
10.
Reuben
A
,
Gittelman
R
,
Gao
J
,
Zhang
J
,
Yusko
EC
,
Wu
CJ
, et al
TCR repertoire intratumor heterogeneity in localized lung adenocarcinomas: an association with predicted neoantigen heterogeneity and postsurgical recurrence
.
Cancer Discov
2017
;
7
:
1088
97
.
11.
Patel
SP
,
Kurzrock
R
. 
PD-L1 expression as a predictive biomarker in cancer immunotherapy
.
Mol Cancer Ther
2015
;
14
:
847
56
.
12.
Rosenthal
R
,
Cadieux
EL
,
Salgado
R
,
Bakir
MA
,
Moore
DA
,
Hiley
CT
, et al
Neoantigen-directed immune escape in lung cancer evolution
.
Nature
2019
;
567
:
479
85
.
13.
Sheikhbahaei
S
,
Mena
E
,
Yanamadala
A
,
Reddy
S
,
Solnes
LB
,
Wachsmann
J
, et al
The value of FDG PET/CT in treatment response assessment, follow-up, and surveillance of lung cancer
.
AJR Am J Roentgenol
2017
;
208
:
420
33
.
14.
Goldfarb
L
,
Duchemann
B
,
Chouahnia
K
,
Zelek
L
,
Soussan
M
. 
Monitoring anti-PD-1-based immunotherapy in non-small cell lung cancer with FDG PET: introduction of iPERCIST
.
EJNMMI Res
2019
;
9
:
8
.
15.
Nguyen
NC
,
Yee
MK
,
Tuchayi
AM
,
Kirkwood
JM
,
Tawbi
H
,
Mountz
JM
. 
Targeted therapy and immunotherapy response assessment with F-18 fluorothymidine positron-emission tomography/magnetic resonance imaging in melanoma brain metastasis: a pilot study
.
Front Oncol
2018
;
8
:
18
.
16.
Ronald
JA
,
Kim
BS
,
Gowrishankar
G
,
Namavari
M
,
Alam
IS
,
D'Souza
A
, et al
A PET imaging strategy to visualize activated T cells in acute graft-versus-host disease elicited by allogenic hematopoietic cell transplant
.
Cancer Res
2017
;
77
:
2893
902
.
17.
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
:
2062
9
.
18.
Maute
RL
,
Gordon
SR
,
Mayer
AT
,
McCracken
MN
,
Natarajan
A
,
Ring
NG
, et al
Engineering high-affinity PD-1 variants for optimized immunotherapy and immuno-PET imaging
.
Proc Natl Acad Sci U S A
2015
;
112
:
E6506
14
.
19.
Mayer
AT
,
Natarajan
A
,
Gordon
SR
,
Maute
RL
,
McCracken
MN
,
Ring
AM
, et al
Practical immuno-PET radiotracer design considerations for human immune checkpoint imaging
.
J Nucl Med
2017
;
58
:
538
46
.
20.
Bensch
F
,
van der Veen
EL
,
Lub-de Hooge
MN
,
Jorritsma-Smit
A
,
Boellaard
R
,
Kok
IC
, et al
(89)Zr-atezolizumab imaging as a non-invasive approach to assess clinical response to PD-L1 blockade in cancer
.
Nat Med
2018
;
24
:
1852
8
.
21.
Seo
JW
,
Tavare
R
,
Mahakian
LM
,
Silvestrini
MT
,
Tam
S
,
Ingham
ES
, et al
CD8(+) T-cell density imaging with (64)Cu-labeled cys-diabody informs immunotherapy protocols
.
Clin Cancer Res
2018
;
24
:
4976
87
.
22.
Freise
AC
,
Zettlitz
KA
,
Salazar
FB
,
Lu
X
,
Tavare
R
,
Wu
AM
. 
ImmunoPET imaging of murine CD4(+) T cells using anti-CD4 cys-diabody: effects of protein dose on T cell function and imaging
.
Mol Imaging Biol
2017
;
19
:
599
609
.
23.
Beckford Vera
DR
,
Smith
CC
,
Bixby
LM
,
Glatt
DM
,
Dunn
SS
,
Saito
R
, et al
Immuno-PET imaging of tumor-infiltrating lymphocytes using zirconium-89 radiolabeled anti-CD3 antibody in immune-competent mice bearing syngeneic tumors
.
PLoS One
2018
;
13
:
e0193832
.
24.
James
ML
,
Hoehne
A
,
Mayer
AT
,
Lechtenberg
K
,
Moreno
M
,
Gowrishankar
G
, et al
Imaging B cells in a mouse model of multiple sclerosis using (64)Cu-rituximab PET
.
J Nucl Med
2017
;
58
:
1845
51
.
25.
Larimer
BM
,
Bloch
E
,
Nesti
S
,
Austin
EE
,
Wehrenberg-Klee
E
,
Boland
G
, et al
The effectiveness of checkpoint inhibitor combinations and administration timing can be measured by granzyme B PET imaging
.
Clin Cancer Res
2019
;
25
:
1196
205
.
26.
Gibson
HM
,
McKnight
BN
,
Malysa
A
,
Dyson
G
,
Wiesend
WN
,
McCarthy
CE
, et al
IFNgamma PET imaging as a predictive tool for monitoring response to tumor immunotherapy
.
Cancer Res
2018
;
78
:
5706
17
.
27.
Alam
IS
,
Mayer
AT
,
Sagiv-Barfi
I
,
Wang
K
,
Vermesh
O
,
Czerwinski
DK
, et al
Imaging activated T cells predicts response to cancer vaccines
.
J Clin Invest
2018
;
128
:
2569
80
.
28.
Rudd
CE
,
Schneider
H
. 
Unifying concepts in CD28, ICOS and CTLA4 co-receptor signalling
.
Nat Rev Immunol
2003
;
3
:
544
56
.
29.
Hutloff
A
,
Dittrich
AM
,
Beier
KC
,
Eljaschewitsch
B
,
Kraft
R
,
Anagnostopoulos
I
, et al
ICOS is an inducible T-cell co-stimulator structurally and functionally related to CD28
.
Nature
1999
;
397
:
263
6
.
30.
Tahiliani
V
,
Hutchinson
TE
,
Abboud
G
,
Croft
M
,
Salek-Ardakani
S
. 
OX40 cooperates with ICOS to amplify follicular Th cell development and germinal center reactions during infection
.
J Immunol
2017
;
198
:
218
28
.
31.
Whitson
KB
,
Beechem
JM
,
Beth
AH
,
Staros
JV
. 
Preparation and characterization of Alexa Fluor 594-labeled epidermal growth factor for fluorescence resonance energy transfer studies: application to the epidermal growth factor receptor
.
Anal Biochem
2004
;
324
:
227
36
.
32.
Orillion
A
,
Hashimoto
A
,
Damayanti
N
,
Shen
L
,
Adelaiye-Ogala
R
,
Arisa
S
, et al
Entinostat neutralizes myeloid-derived suppressor cells and enhances the antitumor effect of PD-1 inhibition in murine models of lung and renal cell carcinoma
.
Clin Cancer Res
2017
;
23
:
5187
201
.
33.
Danilchanka
O
,
Mekalanos
JJ
. 
Cyclic dinucleotides and the innate immune response
.
Cell
2013
;
154
:
962
70
.
34.
Woo
SR
,
Fuertes
MB
,
Corrales
L
,
Spranger
S
,
Furdyna
MJ
,
Leung
MY
, et al
STING-dependent cytosolic DNA sensing mediates innate immune recognition of immunogenic tumors
.
Immunity
2014
;
41
:
830
42
.
35.
Sallets
A
,
Robinson
S
,
Kardosh
A
,
Levy
R
. 
Enhancing immunotherapy of STING agonist for lymphoma in preclinical models
.
Blood Adv
2018
;
2
:
2230
41
.
36.
Demaria
O
,
De Gassart
A
,
Coso
S
,
Gestermann
N
,
Di Domizio
J
,
Flatz
L
, et al
STING activation of tumor endothelial cells initiates spontaneous and therapeutic antitumor immunity
.
Proc Natl Acad Sci U S A
2015
;
112
:
15408
13
.
37.
Corrales
L
,
Glickman
LH
,
McWhirter
SM
,
Kanne
DB
,
Sivick
KE
,
Katibah
GE
, et al
Direct activation of STING in the tumor microenvironment leads to potent and systemic tumor regression and immunity
.
Cell Rep
2015
;
11
:
1018
30
.
38.
Sharma
P
. 
Immune checkpoint therapy and the search for predictive biomarkers
.
Cancer J
2016
;
22
:
68
72
.
39.
Carthon
BC
,
Wolchok
JD
,
Yuan
J
,
Kamat
A
,
Ng Tang
DS
,
Sun
J
, et al
Preoperative CTLA-4 blockade: tolerability and immune monitoring in the setting of a presurgical clinical trial
.
Clin Cancer Res
2010
;
16
:
2861
71
.
40.
Curran
MA
,
Montalvo
W
,
Yagita
H
,
Allison
JP
. 
PD-1 and CTLA-4 combination blockade expands infiltrating T cells and reduces regulatory T and myeloid cells within B16 melanoma tumors
.
Proc Natl Acad Sci U S A
2010
;
107
:
4275
80
.
41.
Mayer
AT
,
Gambhir
SS
. 
The immunoimaging toolbox
.
J Nucl Med
2018
;
59
:
1174
82
.
42.
Wei
SC
,
Levine
JH
,
Cogdill
AP
,
Zhao
Y
,
Anang
NAS
,
Andrews
MC
, et al
Distinct cellular mechanisms underlie anti-CTLA-4 and anti-PD-1 checkpoint blockade
.
Cell
2017
;
170
:
1120
33
.
43.
Niemeijer
AN
,
Leung
D
,
Huisman
MC
,
Bahce
I
,
Hoekstra
OS
,
van Dongen
G
, 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
.

Supplementary data