Abstract 2436

Poster Board II-413

Sensitive in vivo imaging methods have advanced the fields of stem cell transplantation, graft-versus–host disease (GVHD) and graft-versus-tumor responses (GVT). Near-infrared (NIF) fluorescent proteins (FP) appear advantageous for deeper tissue penetration due to minimized absorbance by hemoglobin, water and lipids. Therefore we tested whether a recently published NIF FP (FP635, “Katushka”) could serve as a single reporter for whole body and single cell imaging.

To compare signal intensities of eGFP and FP635 we generated fluorescent MOSEC cell lines (mouse ovarian cancer), titrated them in vitro and subcutaneously (s.c.) in vivo in Balb/c nu/nu mice.

MOSEC FP635 showed twice the signal intensities compared to MOSEC eGFP in vitro by spectral fluorescence imaging (FLI). In vivo the eGFP signal was attenuated >60% in contrast to only 20% for FP635 from subcutaneous sites. However, FP635 signals from deep tissue layers were quenched.

To address whether reduced signal attenuation of FP635 may allow sensitive visualization of immune processes by FLI and multiphoton-laser-scanning-microscopy (MPM) we generated transgenic mice in the genetic C57Bl/6 (B6) background, expressing FP635 under the ubiquitin promoter. Transgenic founders were selected upon signal intensities of leukocyte populations measured by flow cytometry in the PerCP channel. Combination of FP635 with colors other than red were possible for multiparameter flow cytometry.

Next, eGFP, DsRed and FP635 splenocytes from transgenic donors were titrated as described above. In vitro signal intensities of FP635 splenocytes were >5 times lower compared to the other two FPs. FP635 signal absorption in vivo was low (30%) which is consistent with MOSEC titration results. In vivo DsRed detection was most sensitive and signals were similarly attenuated as FP635 in contrast to eGFP (60%).

Subsequently, we aimed to visualize FP635 in a model of GVHD, where alloreactive T cells undergo massive expansion. Balb/c nu/nu mice were lethally irradiated and transplanted with 5×106 B6.WT bone marrow cells plus either 2×107 B6.DsRed+Luciferase+ or 2×107 B6.FP635 splenocytes.

Sensitivity for DsRed cell detection was superior over FP635 cells. FP635 signal was only weakly detectable in lymph nodes (LN) by ex vivo FLI, where DsRed signals were detectable at earlier timepoints and LNs were even visualized by in vivo FLI. DsRed+ Luciferase+ double transgenic splenocytes allowed direct comparison of bioluminescence imaging (BLI) to FLI. Timely in vivo visualization of immune cells in deep tissues was feasible only by BLI.

After whole body imaging the suitability of FP635 for MPM was checked by co-injecting eGFP B cells and either DsRed or FP635 T cells intravenously into RAG-/- mice. As FP635 is a NIF FP we expected to achieve deeper tissue penetration in hemoglobin rich organs, such as the spleen, in single cell microscopy. After 6 weeks of adoptive cell transfer we imaged spleens by MPM. Tissue penetration depths of DsRed or FP635 T cells were compared to eGFP B cells. No advantage in penetration depth of FP635 over DsRed was measured. Photobleaching is an important factor for microscopy, especially if cells are to be tracked over long time. FP635 transfected 293T cells bleached faster (t1/2=108 sec) than 293T cells transfected with eGFP (t1/2>900 sec) or DsRed (t1/2=411 sec).

These experiments indicate that very high expression levels of FP635 need to be achieved for imaging. The signal attenuation of FP635 is low which may increase the sensitivity but in our hands DsRed showed comparable characteristics. Yet, the fast photobleaching of FP635 compared to the broadly established FPs DsRed and eGFP may be disadvantageous for long term microscopic tracking of cells. Our data indicate that BLI is by far superior over FLI in sensitivity and tissue penetration for whole body imaging of immune cells. However, FLI of red or near-infrared clonally selectable tumor cell lines may provide a welcome color addition to study immune cell-tumor interactions in combined models of BLI and FLI.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.