In vivo cytometry of antigen-specific t cells using 19F MRI

Abstract

Noninvasive methods to image the trafficking of phenotypically defined immune cells are paramount as we attempt to understand adaptive immunity. A (19)F MRI-based methodology for tracking and quantifying cells of a defined phenotype is presented. These methods were applied to a murine inflammation model using antigen-specific T cells. The T cells that were intracellularly labeled ex vivo with a perfluoropolyether (PFPE) nanoemulsion and cells were transferred to a host receiving a localized inoculation of antigen. Longitudinal (19)F MRI over 21 days revealed a dynamic accumulation and clearance of T cells in the lymph node (LN) draining the antigen. The apparent T-cell numbers were calculated in the LN from the time-lapse (19)F MRI data. The effect of in vivo T-cell division on the (19)F MRI cell quantification accuracy was investigated using fluorescence assays. Overall, in vivo cytometry using PFPE labeling and (19)F MRI is broadly applicable to studies of whole-body cell biodistribution.

FIG. 1

Microscopy and FACS of PFPE-Alexa647 labeled T cells. a: Confocal image of labeled T cells immediately after labeling. Cell nuclei are stained in blue and the PFPE-Alexa647 is shown in red. Scale bar = 40 μm. The inset displays a single cell at high magnification, showing an apparent cytoplasmic distribution of the PFPE-Alexa647. Scale bar = 8 μm. b: FACS of labeled T cells immediately after PFPE labeling. The panels show the histograms of labeled T cells (light) and untreated control cells (dark) for the PFPE-Alexa647, CD4, CD25, and the ova-specific T-cell receptor, KJ1-26. No phenotypic differences were observed between labeled and untreated cells.

FIG. 2

Longitudinal in vivo MRI of PFPE-labeled T cells in the inflammation model. a: Representative images acquired at day 4 from a mouse that received PFPE-labeled T cells and ova/IFA. Shown are contiguous coronal slices where ¹⁹F is rendered in pseudocolor and ¹H is in grayscale. T cells are detectable primarily in the DLN, and no labeled cells were detected in the CLN. The external reference capillary used for quantification is labeled R. b: Single slice from the same mouse at day 21. Some label is still detectable in the DLN, although the total integrated signal is ≈ 10-fold weaker than the maximum. c: Apparent T-cell quantification in DLN from in vivo MRI data. We calculated mean results for n = 3 animals and calculated errors bars from the standard deviation.

FIG. 3

In vivo MRI of T-cell trafficking controls. Shown are contiguous, coronal slices at day 4, rendered as in Fig. 2. a: Mouse inoculated with PFPE labeled T cells as before, but the host received antigen-free IFA. No T cells are detectable in the DLN. Instead, ¹⁹F signal is detectable in the mesenteric region. The bladder is labeled B. b: Mouse received ova/IFA as before, but received PFPE-labeled, heat-killed T cells. No ¹⁹F is visible in the DLN, but observed near the bladder at day 4.

FIG. 4

Optical images showing PFPE-Alexa647-labeled T cells both in vivo and in excised tissue. a: In vivo fluorescence image at day 4 in a partially shaven mouse. Fluorescence is concentrated in the mesenteric region and the DLN (arrow). A mask was placed over the gut to reduce interference from the autofluorescent tissue. No fluorescence is visible in the CLN. b: Optical image of excised inguinal DLN and CLN. The image is an overlay of the fluorescent image (false color) and a white light image. The DLN (right) fluoresces and is visibly swollen compared to the CLN (left).

FIG. 5

FACS data from mice receiving CFSE-labeled T cells and ova-Alexa647. The FACS data were acquired by gating first on live cells and then either T cells or DCs. Representative data are shown (n = 3 for all timepoints). a: Dot plots showing CD11c and ova-Alexa647 (upper panels), KJ1-26 and CFSE (lower panels) from the DLNs, and tissue from the site of antigen transfer from days 2, 4, and 7. The gate windows used for these plots are shown. b: CFSE mean fluorescence intensity (MFI) for the DLN and site of antigen transfer over a 7-day period. The MFI drops significantly between days 2 to 4, indicative of rapid proliferation. Using an ANOVA Tukey–Kramer test, P < 0.001 for the DLN and injection site for day 2 versus days 4 and 7. c: The number of KJ1-26⁺ T cells in the DLN and site of antigen transfer, estimated from the FACS analyses. The total number of LN cells was counted using a hemocytometer and the percent CD4⁺/KJ1-26⁺ cells was determined by FACS analysis as illustrated in panel (a). The absolute number of KJ1-26⁺ was determined by multiplying the total number of LN cells by the fraction expressing CD4⁺ and KJ1-26⁺.