Background In vivo multiphoton autofluorescence microscopy provides label free, single cell imaging of metabolic changes. These metabolic changes are quantified via the metabolic coenzymes NAD(P)H and FAD which are autofluorescent molecules endogenous to all cells. Metabolic reprogramming is a hallmark of cancer and closely associated with cancer progression and immune cell function. We aim to study metabolic changes during administration of an effective, triple-combination immunotherapy regimen in murine melanoma and colon cancer tumors. This therapy includes external beam radiation, intratumoral hu14.18-IL2 immunocytokine (anti-GD2 mAb fused to IL2) or free IL2, and intraperitoneal anti-CTLA-4, leading to in situ vaccination and cure of murine tumors. Previous work showed that a T cell response is critical to the efficacy of this therapy, so we created an mCherry-labeled T cell mouse model to study this response.

Methods We implanted syngeneic B78 (GD2+) melanoma or MC38 (GD2-) colon carcinoma cells into the flanks of mCherry-labeled CD8+ T cell reporter mice (C57Bl/6 background) to induce tumors. Under anesthesia, skin flap surgery was performed and tumors were imaged at several time points during therapy. Multiphoton imaging was performed to collect NAD(P)H, FAD, mCherry, and collagen signal through a 40X objective (figure 1A). Fluorescence lifetime data were collected using time correlated single photon counting electronics. Tissues were harvested and analyzed via flow cytometry and multiplex immunofluorescence to corroborate intravital imaging findings and characterize the immune infiltrate.

Results Here we demonstrate that our in vivo imaging is sensitive to metabolic changes within both B78 melanoma and MC38 colon tumors during our in situ vaccine. We show that CD8 T cells from immunotherapy treated tumors versus control tumors exhibit different metabolic phenotypes including changes in NAD(P)H and FAD protein binding (figure 1B). We observe distinct CD8 T cell metabolic phenotypes across the two different solid tumor types (figure 1B). The tumor cells also exhibit metabolic changes during immunotherapy (data not shown). Additionally, our in vivo imaging can monitor collagen remodeling (figure 1C), a major component of the extracellular matrix and driving force in the tumor microenvironment, during immunotherapy.

Conclusions These results show that in vivo metabolic imaging enables single cell quantification of metabolic changes during therapy – across multiple solid tumors. Combined with other traditional assays, we can elucidate key immune cell populations and the crucial timepoints during therapy where changes are occurring. With continued efforts, this imaging platform may be leveraged to develop new combinations of immunotherapies.

Acknowledgements This work is supported by the Morgridge Institute for Research (Interdisciplinary Fellowship awarded to A.R.H.) and the NIH (R01 CA205101 and R35 CA197078). The authors thank the University of Wisconsin Carbone Cancer Center (UWCCC) Support Grant P30 CA014520, the UWCCC Translational Research Initiatives in Pathology laboratory – supported by the UW Department of Pathology and Laboratory Medicine and the Office of The Director NIH (S10OD023526), the UWCCC Flow Cytometry Laboratory, and the Genome Editing and Animal Models Laboratory for core services.

Ethics Approval All animal work was approved by the University of Wisconsin Institutional Animal Care and Use Committees.

• Download figure
• Open in new tab
• Download powerpoint

Abstract 87 Figure 1

In vivo multiphoton metabolic imaging of CD8 T cell populations during immunotherapy. (A) Experimental workflow showing flank tumor implantation and growth, administration of triple therapy, and representative in vivo fluorescence intensity image of B78 melanoma tumor and MC38 colon carcinoma growing in CD8+ mCherry reporter mice. Images show mCherry-expressing CD8 T cells (red) infiltrating tumor tissue as well as autofluorescent metabolic coenzymes NAD(P)H (blue) and FAD (green) expressed by the tumor and T cells. Scale bar 25 µm. (B) In vivo single cell CD8 T cell fluorescence lifetime data from B78 tumors and MC38 tumors. Infiltrating CD8 T cells in B78 tumors exhibit increased NAD(P)H protein binding but no significant change in FAD protein binding in treated versus control tumors. Infiltrating CD8 T cells in MC38 tumors exhibit decreased and FAD protein binding in treated versus control tumors. We have shown that CD8 T cells are not critical for B78 tumor cures, so we anticipate the increased NAD(P)H protein binding reflects that the B78 CDS T cells are more oxidative in their metabolism, possibly due to exhaustion or a shift towards a memory phenotype. In contrast, we have shown that CD8 T cells are key to curing MC38 tumors, so we anticipate the decreased NAD(P)H protein binding may reflect that these MC38 CD8 T cells are more glycolytlc in their metabolism due to their activated effector state (n = 2 mice per tumor type, CD8 T cells in B78: control = 82 treated = 55, CD8 T cells in MC38: control = 31, treated = 255, mean ± SD, one-way ANOVA with multiple comparisons test). (C) In vivo second harmonic generation images of B78 and MC38 tumors show collagen morphology and alignment. Collagen fibers from control tumors show little change from Day 0 pretreatment to Day 9 of PBS control while collagen fibers from treated tumors show remodeling of collagen towards a healthier tissue phenotype by Day 9 of immunotherapy treatment. Schematic shows the expected trend where healthy tissue collagen is typically very tortuous while tumor tissue collagen is typically very aligned.