Understanding in vivo variability of NK cells from different sources using multiplexed ultrasound-guided photoacoustic imaging

Introduction: : Natural Killer (NK) cell therapy is a promising form of adoptive cell transfer distinct from T cell-based therapies, offering advantages such as innate recognition and cytotoxicity without prior sensitization, and potential for off-the-shelf availability. NK cells can be obtained from various sources, including peripheral blood mononuclear cells (PBMCs) and placental tissue, enabling their use in standardized therapeutic products. However, the variability in therapeutic efficacy due to differences in NK cell sources or individual donors has not been thoroughly evaluated in vivo and often shows little-to-no correlation with in vitro findings. To address this critical gap, we employed multiplexed ultrasound-guided photoacoustic (US/PA) imaging along with labeled NK cells using nanoparticles (NPs) with different spectral signatures, to non-invasively track and quantify NK cell infiltration and persistence in vivo. Using US/PA along with NPs allows us to monitor the in vivo status of NK cells to assess the efficacy of the cell therapy. This approach enabled simultaneous tracking of different populations of NK cells within the same animal, facilitating direct comparison of their in vivo trafficking behavior in murine models of Ewings sarcoma. Our strategy improves the understanding of source and donor dependent variability, ultimately aiming to improve NK cell therapeutic strategies against solid tumors.

Materials and

Methods: : Two PA contrast agents were used: indocyanine green-loaded PLGA (PLGA-ICG) as well as Epolight-700 dye-loaded PLGA (PLGA-700). The absorption spectra of PLGA nanoparticles (NPs) were measured with the US-vis-NIR spectrophotometer. For NK cell labeling, NK cells were co-incubated with PLGA NPs for 45 min.

All US/PA imaging data were acquired on Vevo F2/LAZR-X system (FujiFilm VisualSonics) equipped with a 25-57 MHz transducer coupled with a 5 ns, 10 Hz tunable pulsed laser (680 to 970 nm). For in vitro US/PA imaging, 8 % gelatin tissue-mimicking phantoms containing dome-shaped inclusions of labeled NK cells were imaged and analyzed with VevoLAB software. For in vivo US/PA imaging, primary NK cells were isolated from PBMC and placenta from two independent donors, expanded for 14 days, and then labeled with NPs as aforementioned. Equal numbers of differently labeled NK cells (source or donor matched to NP type) were mixed and injected via tail vein into NSG mice bearing Ewings sarcoma (A673) tumors. Mice were given RT at 8 Gy 24 hours prior to adoptive transfer. Tumors were US/PA imaged at 24 h and 72 h post adoptive transfer, and PA datasets were processed and spectrally unmixed in MATLAB to assess infiltration and persistence of each NK cell population.

Results, Conclusions, and Discussions:: Both PLGA formulations, PLGA-ICG (lambda-max = 820 nm) and PLGA-700 (lambda-max = 700 nm), exhibited strong, well-separated absorption spectra that enable spectral unmixing confirmed by UV-vis-NIR spectrometry and PA imaging (Fig 1A). In gelatin phantoms, NK cells labeled with PLGA-ICG and PLGA-700 generated distinct PA signals, confirming that the two spectra can be unmixed even after NK cell labeling in tissue-mimicking conditions (Fig. 1B). Building on this, we performed multiplexed PA imaging 24 h after adoptive transfer of labeled NK cells from different sources and donors. NK cells of different populations were labeled with different NPs for spectral unmixing in vivo. Processed US/PA images showed differences in the PA signal attained from different NK cell populations (Fig. 2). The qualitative signal analysis allowed for the comparison of source and donor-dependent variations in NK cell homing (Fig. 2B).

Our findings highlight two key advantages of US/PA tracking of NK cells: 1) non-invasive, real-time visualization of NK cell trafficking in vivo, and 2) simultaneous unmixing of multiple adoptively transferred NK cells from different donors/sources within a single animal. Clinically, such multiplexed in vivo screening could rank NK cell products from multiple sources/donors within the tumor microenvironment and streamline early phase trials to deliver the most effective therapy to patients. While the present study is qualitative, future work will expand the cohort sizes, incorporate quantitative/statistical analyses across multiple animals, and pair in vivo imaging with ex vivo assessments of NK cell phenotype and cytotoxic function to confirm that NP labeling does not compromise therapeutic efficacy.

Acknowledgements and/or References (Optional)::