CRISPR-clear imaging of melanin-rich B16-derived solid tumors

**Background:** Tissue clearing combined with deep imaging has become a powerful tool for three-dimensional histological analysis, but its application has been limited in melanin-rich tissues such as melanoma. Melanin causes strong light absorption and scattering, and its poor solubility in both lipid- and water-based solvents makes it resistant to extraction by standard clearing protocols. The B16 murine melanoma model is one of the most commonly used syngeneic tumor models for immuno-oncology research, yet its heavy pigmentation has precluded deep optical imaging. Previous attempts using hydroquinone to inhibit melanin biosynthesis resulted in incomplete depigmentation and loss of cell viability. This study aimed to develop a CRISPR-based approach to permanently eliminate melanin production in B16 cells while preserving all other tumor properties, enabling deep tissue imaging when combined with the PACT clearing method.

**Methods:** The authors used a poorly immunogenic sub-clone of B16 melanoma (D5) that stably expresses human epidermal growth factor receptor 2 (HER2) as a transgenic marker (B16-D5-HER2), and generated a variant additionally expressing tdTomato fluorescent protein (B16-D5-HER2-TdTom). CRISPR-Cas9 with a single-guide RNA targeting tyrosinase exon 1 was delivered via lipofection. The sgRNA was designed to minimize off-target effects, requiring at least three base mismatches at any potential off-target site. After confirming a mutation rate of ~54% in bulk transfected cells, single-cell clones were isolated and genotyped. One selected clone carried a 1-bp deletion in one allele and a 46-bp deletion in the other, both predicted to cause frameshifts and loss of tyrosinase activity. Off-target analysis by targeted deep sequencing of 13 potential off-target loci showed mutation rates below the detection error level of 0.5%, while on-target mutation rate was at least 99.5%. Functional knockout was confirmed by RT-qPCR (no amplification from KO cells using a primer requiring wild-type sequence) and Western blot (no detectable tyrosinase band in KO cells stimulated with α-MSH for 48 and 72 hours, while detectable in wild-type cells at >80% confluency). Transcriptome analysis by RNA sequencing showed that only four genes had expression differences with potential statistical significance between KO and parental cells, none related to cancer or melanin synthesis. Flow cytometry showed no measurable differences in forward and side scatter profiles, and HER2 and tdTomato expression levels were similar by Western blot and immunohistochemistry. For in vivo studies, bilateral tumors were engrafted subcutaneously on the flanks of immunodeficient RAG1⁻/⁻ mice and fully immunocompetent C57BL/6-Tg(HER2) mice (melanin⁺ on right flank, melanin⁻ on left flank). Tumors were grown for 25-40 days to volumes of up to 1700 mm³. PACT clearing was performed on excised tumors, followed by confocal microscopy using Zeiss LSM 880 with 10x, 25x, and 63x objectives. Vasculature was labeled by intravenous injection of DyLight 649-conjugated lectin prior to sacrifice. For nanoparticle experiments, 50 µl of 300 nm fluorescent silica nanoparticles were intratumorally injected one hour before sacrifice in RAG1⁻/⁻ mice.

**Key Results:** PACT-cleared melanin⁻ tumors showed dramatically improved optical translucence compared to melanin⁺ tumors, with complete absence of dark spots under DIC microscopy. Confocal imaging at 300 µm depth using the 25x objective revealed a 10- to 30-fold increase in signal intensities for melanin⁻ tumors across all fluorophores and labeling agents (DAPI for nuclei, tdTomato for B16 cells, DyLight 649 for vasculature). This improvement was consistent at depths up to 1045 µm below the tumor surface and was independent of mouse background (RAG1⁻/⁻ or C57BL/6). High-resolution imaging with the 63x objective showed that melanin⁺ tumors exhibited a dense mesh of punctate structures (likely nuclear bodies) in DAPI-stained nuclei that were absent in melanin⁻ tumors, with higher signal-to-background ratio at depths up to 50 µm in the latter. Antibody staining for HER2 showed consistent labeling throughout the imaging volume (50-300 µm depth) in melanin⁻ tumors but a strong drop-off in melanin⁺ tumors. Nanoparticle imaging in melanin⁻ tumors enabled visualization of 300 nm fluorescent nanoparticles and their subcellular localization within the vasculature at 500 µm depth using digital reconstruction. No measurable differences in tumor size or morphology were observed between melanin⁺ and melanin⁻ tumors grown on either RAG1⁻/⁻ or C57BL/6-Tg(HER2) background.

**Clinical Implications:** CRISPR-Clear provides a practical solution for deep optical imaging of melanin-rich tumors, which has been a significant technical barrier in melanoma research. By creating a tyrosinase-knockout B16 cell line that is otherwise phenotypically identical to the wild-type, this method enables detailed three-dimensional characterization of tumor architecture, including vasculature, surface markers, and subcellular structures. The approach is compatible with antibody staining and nanoparticle imaging, making it valuable for studying drug delivery and therapeutic interventions in the tumor microenvironment. Since the B16 model is the gold standard for preclinical immuno-oncology studies, CRISPR-Clear could facilitate investigation of immune cell infiltration, antibody penetration, and nanoparticle-based drug distribution in intact melanoma tissue. The authors note that this approach could potentially be expanded to other clearing techniques and future in vivo workflows.