Noninvasive interrogation of CD8+ T cell effector function for monitoring early tumor responses to immunotherapy

Abstract

Accurately identifying patients who respond to immunotherapy remains clinically challenging. A noninvasive method that can longitudinally capture information about immune cell function and assist in the early assessment of tumor responses is highly desirable for precision immunotherapy. Here, we show that PET imaging using a granzyme B-targeted radiotracer named 68Ga-grazytracer, could noninvasively and effectively predict tumor responses to immune checkpoint inhibitors and adoptive T cell transfer therapy in multiple tumor models. 68Ga-grazytracer was designed and selected from several radiotracers based on non-aldehyde peptidomimetics, and exhibited excellent in vivo metabolic stability and favorable targeting efficiency to granzyme B secreted by effector CD8+ T cells during immune responses. 68Ga-grazytracer permitted more sensitive discrimination of responders and nonresponders than did 18F-fluorodeoxyglucose, distinguishing between tumor pseudoprogression and true progression upon immune checkpoint blockade therapy in mouse models with varying immunogenicity. In a preliminary clinical trial with 5 patients, no adverse events were observed after 68Ga-grazytracer injection, and clinical responses in cancer patients undergoing immunotherapy were favorably correlated with 68Ga-grazytracer PET results. These results highlight the potential of 68Ga-grazytracer PET to enhance the clinical effectiveness of granzyme B secretion-related immunotherapies by supporting early response assessment and precise patient stratification in a noninvasive and longitudinal manner.

Trial registration: ClinicalTrials.gov NCT05000372.

Keywords: Cancer immunotherapy; Diagnostic imaging; Oncology.

Figures

Figure 1. 18F-FDG PET/CT images of a representative case of tumor pseudoprogression after ICI therapy.

(A) The representative case was a 61-year-old man with lung squamous cell carcinoma (clinical stage cT3N1M0) receiving ipilimumab plus nivolumab. Baseline 18F-FDG PET/CT shows the SUVmax of the mass to be 13.6 and the SULpeak to be 7.1. Interim PET/CT after 1 cycle of immunotherapy (1 month) shows that 18F-FDG uptake of the mass was increased, with a SUVmax of 20.1 and SULpeak of 10.9 (PMD with PERCIST criteria). PET/CT after 3 cycles of therapy (4 months) shows that the 18F-FDG uptake in the mass decreased to a SUVmax of 15.5 and SULpeak of 7.2 (PMR with PERCIST criteria). Tumors are indicated by red arrows. (B) The SUVmax and SULpeak of 18F-FDG PET/CT at different stages of immunotherapy.

Figure 2. In vitro and in vivo characterization of 68Ga-grazytracer.

(A) Schematic of effector T cell activation and granzyme B (GrzmB) secretion following immunotherapy. (B) Chemical structure of 68Ga-grazytracer. (C) Binding specificity of 68Ga-grazytracer with murine (m) or human (h) granzyme B (n = 5). (D) Autoradiography (middle) and granzyme B immunofluorescence staining (2 slides) of tumor serial sections (15 μm thick) harvested from MC38 tumor–bearing mice at 0.5 hours after 64Cu-grazytracer injection. The overlay regions in these 3 serial sections are indicated by dotted red lines. Scale bars: 1 mm. Data are representative of 3 independent experiments. (E) Representative PET images of 68Ga-grazytracer in anti–PD-1–treated mice at 0.5, 1, and 2 hours after injection. (F) Calculated tumor-to-blood and tumor-to-muscle ratios of 68Ga-grazytracer (n = 5). (G) Small-animal PET images of 68Ga-grazytracer and corresponding tumor uptake values in MC38 tumor–bearing mice (16 mice were subjected to PET imaging, and 2 were excluded due to failed tail vein injection). (H) Western blotting of granzyme B in MC38 tumors harvested from MC38 tumor–bearing mice (from G). The lanes were run on the same gel but were noncontiguous. (See supplemental material for full, uncut gels.) (I) Correlation between the tumor uptake of 68Ga-grazytracer quantified by PET imaging and granzyme B/β-tubulin ratios determined by ex vivo Western blotting (r = 0.7168 by Pearson’s correlation analysis). (J) Correlation between the tumor uptake of 68Ga-grazytracer quantified by PET imaging and ex vivo granzyme B levels determined by ELISA (r = 0.7337 by Pearson’s correlation analysis). Tumors are indicated by white arrows in PET images. All numerical data are presented as mean ± SD. **P < 0.01 by unpaired Student’s t test (C).

Figure 3. Small-animal PET imaging of 68Ga-grazytracer to predict tumor responses to anti–PD-1 therapy in MC38 tumor–bearing mice.

(A) Timeline of anti–PD-1 (αPD-1) therapy and PET imaging in MC38 tumor–bearing mice. (B) Representative PET images of 68Ga-grazytracer at 0.5 hours after injection in MC38 tumor–bearing mice treated with PBS (control) or anti–PD-1 with high and low tumor uptake (cutoff of 1.45 %ID/g). Tumors are indicated by white arrows. (C) Individual tumor volumes of MC38 tumor–bearing mice in the control group and treatment groups with high and low tumor uptake. (D) Quantified tumor uptake of 68Ga-grazytracer at 0.5 hours after injection on day 9 in each group of MC38 tumor–bearing mice (n = 7/group). (E) Tumor volumes of MC38 tumor–bearing mice on days 9 and 16 (n = 7/group). (F) Flow cytometric analysis showing the proportion of CD8+ T cells in CD45+ cells, IFN-γ+ or granzyme B+ in CD8+ T cells, and the granzyme B levels in tumors harvested from mice after the indicated treatments (n = 5–6/group). (G and H) Tumor growth curves and body weight of MC38 tumor–bearing mice after the indicated treatments (n = 6–8/group). All numerical data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by 1-way ANOVA with a post hoc Tukey’s test (D–F) and 2-way ANOVA (G).

Figure 4. 68Ga-grazytracer PET imaging in the pseudoprogression and true-progression murine models following treatment with anti–PD-1 and anti–CTLA-4.

(A) Timeline of immunotherapy and PET imaging in MC38 or 4T1 tumor models. (B and C) Individual tumor growth curves of MC38 tumor–bearing mice (pseudoprogression) (B) and 4T1 tumor–bearing mice (true progression) (C) after treatment. (D and E) Body weight of MC38 (D) and 4T1 (E) tumor-bearing mice after treatment. (F and G) Representative PET images (F) and quantified tumor uptake (G) of 18F-FDG and 68Ga-grazytracer in pseudoprogression MC38 tumor–bearing mice on days 0 and 6 (n = 8–9/group). (H and I) Representative PET images (H) and quantified tumor uptake (I) of 18F-FDG and 68Ga-grazytracer in true-progression 4T1 tumor–bearing mice on days 0 and 6 (n = 8–9/group). (J) Representative immunofluorescence staining of granzyme B in MC38 or 4T1 tumor tissues harvested on days 0 and 6. Scale bars: 1 mm. Data are representative of 3 independent experiments. Tumors are indicated by white arrows in PET images. All numerical data are presented as mean ± SD. *P < 0.05, **P < 0.01 by 2-tailed paired Student’s t test (G and I).

Figure 5. PET imaging of 68Ga-grazytracer in mouse models with or without inhibition of immune cell infiltration.

(A) Timeline of PET imaging, anti–PD-1 plus anti–CTLA-4 combinational immunotherapy, and FTY720 treatment in MC38 tumor–bearing mice. (B) Tumor growth curves of MC38 tumor–bearing mice after the indicated treatments: control (PBS), FTY720, anti–PD-1 plus anti–CTLA-4, and anti–PD-1 plus anti–CTLA-4 plus FTY720 (n = 6–9/group). (C and D) Representative PET images (C) and quantified tumor uptake (D) of 18F-FDG on days 0 and 6 in anti–PD-1– plus anti–CTLA-4–treated MC38 tumor–bearing mice with or without FTY720 treatment (n = 8–9/group). (E and F) Representative PET images (E) and quantified tumor uptake (F) of 68Ga-grazytracer on days 0 and 6 in anti–PD-1– plus anti–CTLA-4–treated MC38 tumor–bearing mice with or without FTY720 treatment (n = 8–9/group). (G) Flow cytometric analysis depicting the proportion of NK1.1+ cells, CD4+ T cells, and CD8+ T cells in CD45+ cells in tumors harvested from mice after the indicated treatments (n = 5/group). Tumors are indicated by white arrows in PET images. All numerical data are presented as mean ± SD. *P < 0.05, **P < 0.01, ****P < 0.0001 by 2-way ANOVA (B), 2-tailed paired Student’s t test (D and F), and 2-tailed unpaired Student’s t test (G).

Figure 6. PET imaging of 68Ga-grazytracer in mice treated with ACT.

(A) Schematic of PET imaging and ACT in B16-OVA tumor–bearing mice. (B and C) Tumor growth curves (B) and body weight (C) of B16-OVA tumor–bearing mice after the indicated treatments: control (PBS); and adoptive transfer of WT T cells and OT-I T cells (n = 5–7/group). (D and E) Representative PET images (D) and quantified tumor uptake (E) of 68Ga-grazytracer on days 8 and 12 in B16-OVA tumor–bearing mice treated with PBS or T cells from WT or OT-I mice (n = 5–6/group). Tumors are indicated by white arrows. (F) Flow cytometric analysis depicting the proportion of CD8+ T cells in CD45+ cells and granzyme B+CD8+ T cells in tumors harvested from mice after the indicated treatments on day 12 (n = 5/group). (G) Quantified tumor uptake of 68Ga-grazytracer on day 12 in mice after the indicated treatments (n = 5–6/group). All numerical data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001 by 2-way ANOVA (B), 2-tailed paired Student’s t test (E), and 1-way ANOVA with post hoc Tukey’s test (F and G).

Figure 7. PET/CT imaging of 68Ga-grazytracer and 18F-FDG in patients with lung cancer.

(A) A 66-year-old male (patient 1) with lung adenocarcinoma, clinical stage cT2bN2M0 (IIIa). 18F-FDG PET/CT before 3 cycles of chemotherapy and anti–PD-1 therapy (pemetrexed disodium + cisplatin + toripalimab) showed that the SUVmax was 8.5 and the SULpeak was 5.3. 18F-FDG PET/CT after treatment showed that the SUVmax was 6.5 and the SULpeak was 3.8; this patient was rated as PMR (with EORTC criteria) and SMD (with PERCIST criteria). 68Ga-grazytracer PET/CT after treatment revealed a SUVmax of 4.1 and tumor-to-blood pool SUVmax ratio (T/B ratio) of 1.2, and the patient was assessed as having positive results. (B) IHC staining of granzyme B in the tumor of patient 1 before and after treatment. (C) IHC staining of PD-L1 in the tumor of patient 1 before treatment. Scale bars: 100 μm. (D) A 70-year-old male (patient 3) with sarcomatoid carcinoma of the lung, clinical stage cT4N3M1c (IVb). 18F-FDG PET/CT before 1 cycle of pembrolizumab revealed a SUVmax of 39.3 and SULpeak of 25.0. 18F-FDG PET/CT after treatment revealed a SUVmax of 26.4 and SULpeak of 16.6; this patient was rated as PMD (with EORTC and PERCIST criteria). 68Ga-grazytracer PET/CT after treatment showed that the SUVmax was 2.0 and T/B ratio was 0.8, and the patient was assessed as having negative results. (E and F) IHC staining of granzyme B (E) and PD-L1 (F) in the tumor of patient 3 before treatment. Scale bars: 100 μm. Primary tumors are indicated by the red arrows in PET images.