A practical guideline for in vivo bioluminescence imaging.

INTRODUCTION
Bioluminescence imaging (BLI) is a widely used technique in preclinical cancer research for monitoring biological processes in living animals (Imamura et al., 2018, Mezzanotte et al., 2017). It is based on the enzymatic reaction between luciferase and luciferin, producing photon emission detected by a sensitive imaging system (Choy et al., 2003, Close et al., 2011). For in vivo use, tumor cells are engineered to express luciferase, and luciferin is administered systemically. The emitted light enables real-time, noninvasive tracking of cell location, viability, and activity. Initially, BLI was mainly used to monitor tumor growth in xenograft models and evaluate drug responses (Contag et al., 1995, Contag et al., 1997, Thompson et al., 1995). Today, it serves as a powerful tool in cancer research, allowing dynamic visualization of metastatic progression and therapeutic efficacy across tumor models (Close et al., 2011, Khalil et al., 2013). Here, we provide a brief guide for researchers seeking to apply BLI in vivo, including protocols, key considerations, and troubleshooting strategies. An overview of the typical experimental workflow of in vivo BLI is illustrated in Figure 1.

MAIN BODY

Principles of Bioluminescence Imaging
BLI relies on the enzymatic reaction between luciferase and its substrate luciferin, in the presence of adenosine triphosphate and oxygen. This guideline focuses on firefly luciferase, the most used variant in preclinical cancer models (Choy et al., 2003, Close et al., 2011). This reaction generates visible photons, which are detected by a charge-coupled device (CCD) camera integrated into the imaging system (Contag et al., 1995, Contag et al., 1997, Thompson et al., 1995, Zhang et al., 1999). The intensity of the emitted light correlates with both the number and the metabolic activity of viable luciferase-expressing cells. BLI systems allow optimization of image acquisition by adjusting parameters, such as exposure time, binning, and f-stop (Branchini et al., 2010, Branchini et al., 2007, Maguire et al., 2013). Quantitative analysis is typically conducted by measuring the total photon flux (photons/s) from a defined region of interest (ROI) (Niwa et al., 2023).

Tumor Cell and Luciferin Administration Methods
The route of tumor cell injection significantly influences the localization, growth kinetics, and bioluminescent signal of engrafted tumors. Commonly used methods include subcutaneous (SC), orthotopic mammary fat pad (MFP), and intravenous (IV) tail vein injection. SC injection into the right flank allows for straightforward and reproducible monitoring of localized tumor growth with minimal technical demand, making it suitable for evaluating tumor burden and therapeutic responses, although it lacks organ specificity. Mammary fat pad injection targeting the fourth inguinal mammary fat pad provides a physiologically relevant orthotopic breast cancer model, enabling the study of primary tumor development and spontaneous metastasis, but requires surgical expertise. IV injection introduces cells into the circulation via the lateral tail vein, facilitating lung metastasis modeling, though it poses technical challenges and often produces variable signals (Hidalgo et al., 2014, Siolas and Hannon, 2013).
Luciferin delivery, typically administered at 150 mg/kg in phosphate-buffered saline, also affects signal kinetics and intensity (Carceles-Cordon, 2016). Intraperitoneal (IP) injection is widely used for its simplicity and consistency, yielding peak signals within 10 to 20 minutes. IV injection offers a rapid signal onset (2-5 minutes) but demands technical precision. SC delivery is minimally invasive but often results in delayed and inconsistent kinetics (15-25 minutes) (Cui et al., 2008, Inoue et al., 2009, Inoue et al., 2010, Keyaerts et al., 2008, Khalil et al., 2013). A comparative summary of these luciferin administration routes, including their advantages and kinetic charactreistics, is provided in Table 1.

Data Analysis and ROI Quantification
In vivo BLI generates optical signals that reflect the spatial and quantitative distribution of viable, luciferase-expressing cells (Badr and Tannous, 2011, Imamura et al., 2018, Mezzanotte et al., 2017). The emitted light, captured by a cooled CCD camera, is quantified as photon counts, which can be further processed as either radiance (photons/s/cm²/sr) or its sum, total flux (photons/s). The choice between these metrics depends on the study design: total flux is commonly used for comparing overall tumor burden, while radiance can provide spatially resolved intensity per area (Niwa, 2010, Niwa et al., 2023).
Accurate BLI analysis requires consistent ROI definition and standardized acquisition settings. The ROI determines the area from which the signal is quantified and should be drawn with uniform shape, size, and anatomical reference across all samples and time points. ROI shape and size should be carefully selected based on anatomical landmarks and the spatial distribution of the signal—circular or elliptical ROIs for SC tumors, and free-form ROIs for orthotopic or metastatic lesions. Importantly, the ROI dimensions must be consistent across all time points and experimental groups to ensure comparability (Carceles-Cordon, 2016, Peterson,, Revvity,).
While ROI dimensions should remain fixed across samples, imaging parameters such as exposure time, binning, and f-stop can be adjusted per image to capture optimal signal range without saturation. Exposure time should be optimized to capture an adequate signal without causing pixel saturation; overexposed images cannot be accurately quantified. Binning, which controls pixel resolution and sensitivity, affects both image sharpness and signal detection—higher binning improves sensitivity but reduces spatial resolution. The f-stop, or aperture setting, regulates light collection; lower f-stop settings allow more light to reach the detector but may increase background and reduce depth of field (Carceles-Cordon, 2016). When radiance is used as the unit of measurement, these acquisition settings can be adjusted per image to match signal intensity, as long as quantification remains normalized to area and time.
Finally, a consistent ROI strategy and proper documentation of imaging settings are essential to ensure reliable interpretation across replicates. When appropriate, normalization to baseline or tumor volume can further reduce variability. To ensure these comparisons are valid, transparent documentation of both imaging parameters and ROI definitions is critical. Consistent application of these strategies across replicates enhances the reproducibility and interpretability of preclinical BLI studies.

Challenges and Troubleshooting Tips
Absence of a detectable signal in some animals after luciferin administration is a frequent issue in BLI. This problem often results from incorrect luciferin injection, particularly with IP injections, or poor substrate bioavailability. Incomplete IP injection, degradation of luciferin due to improper storage, or inconsistent injection volumes can all contribute to signal loss. To address this issue, confirm accurate IP injection technique, use light-protected luciferin stored at appropriate temperatures, and perform imaging at a consistent time point (typically 10-15 minutes post injection).
Overexposure or signal saturation is another critical concern that hinders accurate quantification. This occurs when exposure time is too long or the f-stop is too low, leading to pixels reaching maximum detection limits. Saturation leads to pixel clipping and loss of quantitative accuracy, rendering data unusable. As a result, the signal cannot be reliably analyzed. To prevent this, optimal imaging settings should be determined in preliminary trials using autoexposure functions and then fixed for all experimental groups. Equally problematic is signal variability across time points or animals. This can arise from inconsistent ROI placement, differences in animal positioning, or variable luciferin bioavailability. Anchoring ROI placement to anatomical landmarks and standardizing injection and imaging protocols can greatly reduce this variability.
Background luminescence—particularly in the abdominal region—is another challenge, often arising from gut autofluorescence or residual luciferin metabolism. This can interfere with tumor signal detection, especially in orthotopic or metastasis models. To reduce background, it is advisable to use appropriate negative controls, perform background subtraction, and consider fasting animals before imaging.
Inconsistent ROI quantification is also a common issue. This often results from variable ROI placement, inconsistent animal positioning, or unstandardized imaging parameters. To overcome this, ROI location and size should be anchored to anatomical landmarks, and imaging settings must remain constant throughout the study. Finally, interanimal variability can be minimized by normalizing bioluminescence data to baseline or control values, enhancing the interpretability of longitudinal studies.
A consolidated summary of common challenges and corresponding practical troubleshooting strategies is presented in Table 2. Proactive management of these recurrent issues through careful experimental planning, standardized imaging protocols and pilot optimization can substantially improve the reproducibility and interpretability of BLI-based studies.

CONCLUDING REMARKS
Here, we provide a concise overview of in vivo BLI, covering its fundamental principles, common experimental approaches, and practical strategies for data acquisition and analysis. Although specific protocols may vary depending on specific tumor models and imaging platforms, the key elements outlined in this MiniResource aim to assist researchers—especially those new to BLI—in designing and interpreting experiments with improved accuracy and reproducibility.

Funding and Support
This work was supported by grants from the 10.13039/501100003725National Research Foundation of Korea (RS-2025-00556619, RS-2025-18362970 to H.W.P.), Ministry of Health & Welfare of Korea (RS-2025-25459146 to H.W.P.), Korea Basic Science Institute (National Research Facilities and Equipment Center) grant funded by the Ministry of Education (RS-2025-02402969), and by the Brain Korea 21 FOUR Program (to H.L.).

Author Contributions
Hyun Woo Park: Writing – review & editing, Supervision, Methodology, Funding acquisition. Hannah Lee: Writing – original draft, Visualization, Validation, Methodology. Ji Soo Chae: Writing – original draft, Visualization, Methodology.

Declaration of Competing Interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The author Hyun Woo Park is an Associate Editor for Molecules and Cells and was not involved in the editorial review or the decision to publish this article.