A novel probe for the non-invasive detection of tumor-associated inflammation

Abstract

A novel dual-mode contrast agent was formulated through the addition of an optical near infrared (NIR) probe to a perfluorocarbon (PFC)-based ¹⁹F magnetic resonance imaging (MRI) agent, which labels inflammatory cells in situ. A single PFC-NIR imaging agent enables both a qualitative, rapid optical monitoring of an inflammatory state and a quantitative, detailed and tissue-depth independent magnetic resonance imaging (MRI). The feasibility of in vivo optical imaging of the inflammatory response was demonstrated in a subcutaneous murine breast carcinoma model. Ex vivo optical imaging was used to quantify the PFC-NIR signal in the tumor and organs, and results correlated well with quantitative ¹⁹F NMR analyses of intact tissues. ¹⁹F MRI was employed to construct a three-dimensional image of the cellular microenvironment at the tumor site. Flow cytometry of isolated tumor cells was used to identify the cellular localization of the PFC-NIR probe within the tumor microenvironment. Contrast is achieved through the labeling of host cells involved in the immune response, but not tumor cells. The major cellular reservoir of the imaging agent were tumor-infiltrating CD11b⁺ F4/80^low Gr-1^low cells, a cell subset sharing immunophenotypic features with myeloid-derived suppressor cells (MDSCs). These cells are recruited to sites of inflammation and are implicated in immune evasion and tumor progression. This PFC-NIR contrast agent coupled to non-invasive, quantitative imaging techniques could serve as a valuable tool for evaluating novel anticancer agents.

Keywords: 19F MRI; imaging; inflammation; myeloid-derived suppressor cells; near infrared probe; tumor; tumor-infiltrating leukocytes.

Figure 1. Characterization of the PFC-NIR nanoemulsion. (A) Normalized excitation and emission intensity as a function of wavelength for the PFC-NIR nanoemulsion. (B)¹⁹F content measured by NMR spectroscopy as compared with fluorescence intensity of the dual-mode agent. (C) Dynamic light scattering (DLS) of the contrast agent particle size measured over time for three different storage temperatures. (D) Effect of serum-containing media on the short-term stability of the emulsion as measured by DLS. Error bars represent the standard deviation of triplicate measurements.

Figure 2. In vivo optical imaging of tumor-associated inflammation. (A) Typical time-course of live, whole body NIR imaging just prior to and up to 72 h after, the administration of the contrast agent in a naive tumor-free control (top) and 4T1 cell-derived tumor-bearing mouse (bottom). Far-right inserts define areas used for signal analysis. (B) From left to right, signal (I) in the livers of tumor-free mice, the livers of 4T1-bearing mice and in tumors as obtained by serial measurements over time. Time zero signals (I₀) were used to define background and were subtracted from data. Bold lines denote cohort average, error bars are standard deviation, and shaded lines denote individual mice.

Figure 3. In vitro cellular uptake of the PFC-NIR agent. (A) Representative histogram of PFC-NIR detection in RAW 264.7 cells incubated with the PFC-NIR agent for 3 h (dashed line) or 18 h (solid line), as compared with untreated control cells (shaded histogram). (B) Label uptake by RAW 264.7 cells (closed bars) and 4T1 cells (open bars) incubated with the PFC-NIR probe over time as measured by flow cytometry. (C, D) Viability of RAW 264.7 cells after treatment with increasing amounts of the PFC-NIR agent for 18 h, as measured by Trypan blue exclusion (C) or cytofluorometric analysis of calcein AM and 7-aminoactinomycin D (7-AAD) (D). Results in (D) represent the average frequency of cells labeled with either calcein AM (live), 7-AAD (dead) or both (apoptotic), and error bars represent standard deviation of mean of duplicate samples.

Figure 4. Ex vivo PFC-NIR measurements by NMR and fluorescence correlate and reflect in vivo imaging observations. (A) NIR images of excised tumors and other organs were obtained72 h post-administration of the PFC-NIR agent. Quantitative data of the optical signal, normalized by tissue weight, are shown for individual tissues isolated from tumor-bearing (open bars) or control animals (closed bars) (n = 3 per tissue). Error bars represent the standard deviation. (B) A representative NMR spectrum of an excised tumor analyzed in the presence of a chemically shifted reference compound is shown. The signal from the probe in the tissue is spectrally well separated from the reference compound, and the comparison of the intensities of the two signals yields the relative amount of ¹⁹F per tissue. Quantitative data of the NMR signal, normalized by tissue weight, are shown for individual tissues isolated from tumor-bearing (open bars) or control animals (closed bars) (n = 3 per tissue). Error bars represent the standard deviation. (C) Correlation of fluorescent optical and NMR signals of individual tissues (obtained from both control and tumor-bearing animals), plotted on a log scale.

Figure 5. Correspondence of MRI and optical signal differs in the resolution obtained in the distribution of the contrast agent in the tumor. (A and B) Images obtained 72 h after the administration of the contrast agent are shown for control (A) and tumor-bearing (B) mice. In panels (A and B), optical image are provided on the left, while MRI overlays (¹H/¹⁹F) are depicted on the right. ¹⁹F images collected as a coronal projection were rendered in pseudo-color (hot-iron scale), and only pixels containing the ¹⁹F signal were selected and overlaid on a 2 mm slice of a ¹H anatomical pilot image. As the optical image was acquired in the opposite orientation, it was reflected along the x-axis so that the tumor can be observed on the same side of the mouse. (C) Consecutive 0.5-mm thick ¹⁹F/¹H MRI composite images rendered from 3D data sets anterior to posterior coronal views, shown from top to bottom. Only the region containing the tumor is shown and the ¹⁹F signal is overlayed on the corresponding anatomical slice.

Figure 6. PFC probe within tumor infiltrates is retained in cells with a myeloid-derived suppressor cell-like phenotype (CD11b⁺F480^lowGr-1^low). (A and B) Tumor-bearing mice were administered with the PFC-NIR agent and an equivalent dose of PFC-Red, containing a red fluorescent moiety. After 72 h and confirmatory in vivo imaging of the NIR signal, animals were sacrificed and tumors were isolated for further analysis. Single-cell suspensions obtained from tumors were labeled with antibodies specific CD11b, Gr-1 and F4/80. (A) Average frequencies of cells expressing the indicated markers in tumor extracts, error bars represent the standard deviation. (B) Average frequency of Gr-1⁺, F4/80⁺ or CD11b⁺ cells containing the PFC-Red label. (C) Evaluation of cell populations obtained from tumors in control animals (no contrast agent, gated on live cells by forward and side scatter) or contrast agent-treated animals (gated on live, PFC-Red⁺ cells) using multicolor flow cytometry. One representative tumor isolate for each group is shown. The cellular populations of contrast agent-treated animals gated on live cells only were similar to those observed in animals not receiving the contrast agent.

Figure 7. The PFC probe is located within cells at the tumor site. (A–C) Confocal microscopy was performed on portions of the tumors examined in Figure 6. Evaluation of samples was performed directly after cryosectioning and mounting (A) or mounting with gelvatol (B), to prevent any potential loss of the probe from tissue sections. In (C), Hoechst 33342 was added to stain nuclei and aid in the visualization of cells. Each image represents merged overlays of the PFC-Red on DiC (A and B) or Hoechst 33342 (C). All photomicrographs were taken with a 20X objective.