Whole-body PET Imaging of T-cell Response to Glioblastoma

Abstract

Purpose: Immunotherapy is a promising approach for many oncological malignancies, including glioblastoma, however, there are currently no available tools or biomarkers to accurately assess whole-body immune responses in patients with glioblastoma treated with immunotherapy. Here, the utility of OX40, a costimulatory molecule mainly expressed on activated effector T cells known to play an important role in eliminating cancer cells, was evaluated as a PET imaging biomarker to quantify and track response to immunotherapy.

Experimental design: A subcutaneous vaccination approach of CpG oligodeoxynucleotide, OX40 mAb, and tumor lysate at a remote site in a murine orthotopic glioma model was developed to induce activation of T cells distantly while monitoring their distribution in stimulated lymphoid organs with respect to observed therapeutic effects. To detect OX40-positive T cells, we utilized our in-house-developed ⁸⁹Zr-DFO-OX40 mAb and in vivo PET/CT imaging.

Results: ImmunoPET with ⁸⁹Zr-DFO-OX40 mAb revealed strong OX40-positive responses with high specificity, not only in the nearest lymph node from vaccinated area (mean, 20.8%ID/cc) but also in the spleen (16.7%ID/cc) and the tumor draining lymph node (11.4%ID/cc). When the tumor was small (<10⁶ p/sec/cm²/sr in bioluminescence imaging), a high number of responders and percentage shrinkage in tumor signal was indicated after only a single cycle of vaccination.

Conclusions: The results highlight the promise of clinically translating cancer vaccination as a potential glioma therapy, as well as the benefits of monitoring efficacy of these treatments using immunoPET imaging of T-cell activation.

Fig.1. ⁸⁹Zr-DFO-OX40 mAb PET scan revealed stimulated lymphoid organs with cancer vaccination.

(A) Conceptual sketch of this study. After vaccination therapy in the left shoulder for mice bearing glioma in the right hemisphere, immunoPET with ⁸⁹Zr-DFO-OX40 mAb was performed to visualize the stimulated immune system, especially focusing on LN of vaccinated area and tumor draining lymph node (TDLN), as well as treatment effect. (B) Experimental schedule. After intracranial injection of glioma cells to right hemisphere, mice were monitored by BLI for 10 to 14 days. CpG+tumor lysate and OX40 mAb were injected subcutaneously to left shoulder 5 days and 6 days prior to injection of ⁸⁹Zr-DFO-OX40 mAb, respectively. PET scans were sequentially acquired on day 1, day 2, and day 5 thereafter. (C) Representative ⁸⁹Zr-DFO-OX40 mAb PET/CT scan (day 5) of vaccinated and control mice. Whole-body fused maximum intensity projection (MIP) image shows multiple lymphatic organs including left axillary LN (vaccinated LN), right cervical LN (TDLN), and the spleen. Control mouse did not show any stimulated LNs but instead a large tumor in the right hemisphere. Corresponding axial fusion images verify the exact location of tracer accumulation in each site. (D) Representative 3D fused PET/CT scans of vaccinated and control mice. Multiple LNs and spleen are clearly visualized from day 2 (arrows). Vehicle demonstrated a large tumor in the right brain on day 5 (arrowhead). (E) Time course of the highest PET signal of ⁸⁹Zr-DFO-OX40 mAb in bilateral axillary, cervical LNs, the spleen, and the brain tumor. Compared to the control group, left axillary LN (vaccinated) (p < 0.0001) and the spleen (p = 0.020) showed significantly higher accumulation in vaccinated group on day 5, followed by right cervical LN (TDLN) and right axillary LN with nearly significance. PET signal in the brain tumor was significantly lower in the vaccinated group on day 5 (p = 0.042). Data are shown with mean ± SEM. (F) Spleen volume comparison between vaccinated and control groups. In vaccinated group, significant splenomegaly was observed (p = 0.019). (G) Correlogram of ⁸⁹Zr-DFO-OX40 mAb accumulation on day 5. Each color box indicates the PET signal strength from individual mouse. The difference of mean PET signal between vaccinated and control group was larger in this order: Left axillary LN (vaccinated) > spleen > right cervical (TDLN) > right axillary LN > left axillary LN. *, <0.05; **, <0.01; ***, <0.001; ***, <0.0001.

Fig.2.. FACS analysis in left axillary LN, left cervical LN, the spleen and the brain tumor.

(A) CD4 and CD8 gated windows. No significant difference was seen in the total CD4 positive population ratio between vaccinated and control groups. (B) OX40 and CD4 gated windows. OX40 positive CD4 cells were significantly seen more in vaccinated group in left axillary (p = 0.017), left cervical LN (p = 0.036), and the spleen (p = 0.036). No significant difference was observed in the tumor. Data are shown with mean ± SEM. *, <0.05.

Fig. 3.. Specificity of ⁸⁹Zr-DFO-OX40 mAb PET in lymphoid organs.

(A) Representative PET images of ⁸⁹Zr-DFO-OX40 mAb (mAb) and ⁸⁹Zr-DFO-rat IgG (isotype control), in both vaccinated and control mice, on day 1, 2, and 5. Lymphoid organs were vividly delineated in vaccinated group (arrows) whilst this was less clear with isotype tracer. (B) Time course of ⁸⁹Zr-DFO-OX40 mAb and ⁸⁹Zr-DFO-rat IgG mAb (isotype control) in each lymphoid organ and main solid organs. Bilateral axillary LNs and the spleen showed higher uptake of ⁸⁹Zr-DFO-OX40 mAb than isotype control in vaccinated group. All lymphoid organs showed soaring curves over time with ⁸⁹Zr-DFO-OX40 mAb in vaccinated group, which was seen neither with isotype control nor non-vaccinated group (left axillary node, p = 0.0083; right axillary node, p = 0.028; spleen, p = 0.026). Heart uptake, which represents blood, showed higher uptake on day 1 with ⁸⁹Zr-DFO-OX40 mAb compared to isotype control tracer in both vaccinated and control groups, but this phenomenon disappears over time. Liver uptake on day 5 scan was higher with ⁸⁹Zr-DFO-OX40 mAb in vaccinated group than the control group. (C) Biodistribution results after day 5 scan. ⁸⁹Zr-DFO-OX40 mAb showed significantly higher counts in the liver, spleen and left axillary LN in vaccinated group than control, while blood count was significantly lower in the same manner. Splenic count was significantly higher in vaccinated mice than control in both ⁸⁹Zr-DFO-OX40 and isotype group. (D) Correlation between PET and Biodistribution results. Strong correlations were seen in both solid organs and lymph nodes. Data are shown with mean ± SEM. *, <0.05; **, <0.01; ***, <0.001; ****, <0.0001.

Fig. 4.. ⁸⁹Zr-DFO-OX40 mAb accumulation in the brain tumor was verified by (A) MRI and (B) autoradiograph (ARG).

(A) Coronal scan of ⁸⁹Zr-DFO-OX40 mAb PET/CT (left), T2WI MRI/CT (right), and these fusion (middle) on day 2 post tracer injection. (B) Coronal scan of ⁸⁹Zr-DFO-OX40 mAb PET/CT on day 5 was compared with ARG of the same day. Top row, good response with vaccination; bottom row, huge tumor in control group. In the vaccinated mice, both PET and ARG showed small signals at the right bottom of the skull, while the mice who did not receive vaccine or the vaccine effect was not insufficient, large tumor was visualized in the colocalized right hemisphere. Top, vaccinated; bottom, vehicle. (C) Correlation between BLI signals in the brain tumor and ⁸⁹Zr-DFO-OX40 mAb accumulation in PET on day 5. Significant correlation was seen in both treated (n = 15) and vehicle-treated mice (n = 11).

Fig.5.. Tumor size tracking with BLI to see therapeutic effect of cancer vaccine.

(A) Representative BLI images on 0, 4, and 8 days after cancer vaccination. Top left, vaccinated, low tumor burden (< 10⁶ [p/sec/cm²/sr]); bottom left, vaccinated, high tumor burden (> 10⁶ [p/sec/cm²/sr]); top right, control, low burden; bottom right, control, high burden. (B) Waterfall plot of best treatment response (= the ratio of lowest signal after vaccination to baseline in BLI) in classes of low and high burden. (C) Comparison between vaccinated and control groups in both low and high burden classes. In low burden mice, vaccinated mice indicated significant shrinking of the brain tumor with BLI (p = 0.015). Data are shown with mean ± SEM. (D) Time course of BLI signals after vaccination in ⁸⁹Zr-DFO-OX40 mAb PET group and isotype control PET group of low tumor burden. There was no significant difference in tumor signals between ⁸⁹Zr-DFO-OX40 mAb and isotype control, indicating ⁸⁹Zr-DFO-OX40 mAb itself does not affect the treatment response. *, <0.05.

Fig. 6.. Evaluating the therapeutic effect of vaccination with ⁸⁹Zr-DFO-OX40 mAb PET/CT.

(A) Whole-body lymphoid organs potentially stimulated by vaccine. (B) MIP images of ⁸⁹Zr-DFO-OX40 mAb PET/CT in low tumor burden (n = 4) and high tumor burden (n = 5) (right). (C) Total lesion immune response (TLIR) in ⁸⁹Zr-DFO-OX40 mAb PET in each stimulated lymphoid organs of vaccinated mice (bilateral axillary, brachial, cervical, inguinal LNs and the spleen). Significantly higher uptake was seen in the spleen and right inguinal LN in high burden group. (D) Whole-body total lesion immune response (WB-TLIR) and % decrease of BLI signal in low and high burden of vaccinated mice. In low burden group, higher WB-TLIR (> 1500×10⁻³ %ID) was associated with good treatment response (≥ 99%), whereas only one mouse exhibited weak treatment effect (< 20 %) in high burden group despite moderate to intense immune response (> 2000×10⁻³ %ID). Data are shown with mean ± SEM. (E) Correlation of ⁸⁹Zr-DFO-OX40 mAb accumulation and BLI ratio as treatment response in low and high burden mice. Left axillary LN (vaccinated) showed a good negative correlation in low burden group, whereas no correlation was seen in high burden group. *, p <0.05.