Info
Background:
Trastuzumab (TZM), a monoclonal antibody targeting HER2-positive (HER2⁺) breast cancer, has substantially improved clinical outcomes; however, its therapeutic efficacy is frequently limited by heterogeneous intratumoral distribution, incomplete target engagement, and resistance1–4. These limitations are compounded by the inability of conventional imaging modalities to distinguish antibody accumulation from functional receptor binding. Near-infrared fluorescence lifetime imaging-based Förster resonance energy transfer (NIR FLI-FRET) provides a quantitative and spatially resolved approach to directly measure TZM-HER2 engagement in vivo by leveraging the nanometer-scale proximity between donor- and acceptor-labeled antibodies bound to HER2 receptors5–9. FLI-FRET imaging of TZM–HER2 interactions requires conjugation of NIR fluorophores to the antibody; however, standard stochastic labeling via N-hydroxysuccinimide (NHS) ester chemistry generates heterogeneous antibody–dye conjugates with variable dye-to-antibody ratios, potentially altering binding affinity and compromising quantitative FRET readouts. We evaluated the impact of conjugation strategy by comparing conventional lysine-based NHS labeling (TZM-NHS) with a site-specific cyclic peptide meditope-based labeling approach (TZM-MDT), which enables controlled and stoichiometric dye conjugation on engineered trastuzumab10,11. Further, understanding how therapeutic antibodies engage their targets within the complex spatial architecture of solid tumors remains a central challenge in oncology.

Hypothesis/Objective:
To quantitatively resolve how antibody conjugation heterogeneity and tumor spatial molecular heterogeneity govern target engagement in vivo using FLI-FRET imaging integrated with spatial transcriptomics.

Methods:
HER2⁺ AU565 tumor xenografts were established in nude mice and administered either NIR-labeled TZM-NHS or TZM-MDT probes (20 µg donor-AF700 / 40 µg acceptor-AF750). At 48 h post-injection, tumors were excised and imaged ex vivo using time-domain mesoscopic fluorescence molecular tomography (TD-MFMT) to obtain fluorescence lifetime–derived FD fraction maps of HER2 engagement. Tumor sections were subsequently processed for H&E and immunohistochemistry (anti-HER2 and anti-TZM), followed by quantitative image analysis and spatial correlation with FLI-FRET maps using ImageJ- and GIMP-based registration. Spatial transcriptomic profiling was performed on matched sections to assess region-specific gene expression and its association with antibody engagement.

Results:
TD-MFMT imaging revealed that TZM-MDT achieves significantly higher FD fraction values compared to TZM-NHS, indicating enhanced HER2 engagement in vivo despite comparable receptor expression levels (Fig. 1). Donor fluorescence intensity alone failed to capture these differences, underscoring the importance of lifetime-based FRET measurements for functional imaging. Pixel-wise FD fraction mapping demonstrated widespread and uniform intratumoral engagement by TZM-MDT, whereas TZM-NHS was largely restricted to tumor peripheries, consistent with limited penetration and heterogeneous binding (Fig. 1A). These findings were validated by immunohistochemistry (IHC), where anti-TZM staining showed strong spatial concordance with FD fraction maps, while anti-HER2 confirmed comparable receptor levels (Fig. 1B). Line-profile analyses along periphery–center–periphery axes further confirmed broader and less heterogeneous TZM-MDT engagement (Fig. 1C–E). Spatial correlation analysis using ImageJ- and GIMP-based registration demonstrated strong and reproducible agreement between FLI-FRET maps and IHC in TZM-MDT tumors (R² > 0.6), whereas TZM-NHS tumors showed weaker and more variable correlations (R² < 0.6) (Fig. 2). Spatial transcriptomic profiling revealed distinct intratumoral transcriptional niches aligned with engagement patterns (Fig. 3). Differential expression analysis identified genes associated with TZM-MDT high-engagement regions, including JUN, DDIT4, CITED2, IGFBP-5, and MUC5B, while DKK1, SCGB2A2, AZGP1, and DHRS2 genes associated with TZM-NHS low-engagement regions. Spatial feature maps confirmed regional localization of these gene programs. Integration of FLI-FRET and transcriptomic data showed that high FD fraction regions correspond to transcriptionally distinct tumor niches, indicating that antibody binding is influenced not only by receptor availability but also by local biological context.

Conclusion:
Together, this multimodal framework establishes TZM-MDT based FLI-FRET imaging as a robust method for quantifying drug–receptor engagement in vivo, decoding the interplay between drug delivery, target engagement, and tumor heterogeneity, with broad implications for optimizing antibody-based therapeutics.

Image/Figure:
Click to View

Image/Figure Caption:
Figure 1. Site-specific antibody labeling enhances spatial HER2 engagement revealed by ex vivo FLI-FRET imaging and histological validation. Ex vivo time-domain mesoscopic fluorescence molecular tomography (TD-MFMT)–based FLI-FRET imaging was used to quantify trastuzumab (TZM)-HER2 engagement in AU565-XTM tumors was assessed using NIR-labeled TZM-NHS or site-specific TZM-MDT FRET pairs (20 µg donor-AF700 / 40 µg acceptor-AF750) at 48 h post-injection. (A) Intensity and FD Fraction maps show higher and more uniform engagement for TZM-MDT, whereas TZM-NHS remains heterogeneous and peripheral. (B) H&E and IHC (anti-HER2, anti-TZM) confirm strong spatial concordance with FD Fraction maps, indicating functional binding rather than receptor abundance. (C-E) Quantitative analysis demonstrates significantly increased FD Fraction and anti-TZM signal for TZM-MDT with comparable HER2 levels (scale bar 1mm; p < 0.05; n = 4). (F-H) Spatial profiling reveals deeper and more uniform intratumoral penetration for TZM-MDT compared to peripheral gradients in TZM-NHS, validated by FD Fraction line profiles and anti-TZM histology across tumor axis from periphery to core to periphery.

Author
Amit Verma
Albany Medical College