In vivo near-infrared imaging of ErbB2 expressing breast tumors with dual-axes confocal endomicroscopy using a targeted peptide

Abstract

ErbB2 expression in early breast cancer can predict tumor aggressiveness and clinical outcomes in large patient populations. Accurate assessment with physical biopsy and conventional pathology can be limited by tumor heterogeneity. We aim to demonstrate real-time optical sectioning using a near-infrared labeled ErbB2 peptide that generates tumor-specific contrast in human xenograft breast tumors in vivo. We used IRDye800CW as the fluorophore, validated performance characteristics for specific peptide binding to cells in vitro, and investigated peak peptide uptake in tumors using photoacoustic tomography. We performed real-time optical imaging using a handheld dual-axes confocal fluorescence endomicroscope that collects light off-axis to reduce tissue scattering for greater imaging depths. Optical sections in either the vertical or horizontal plane were collected with sub-cellular resolution. Also, we found significantly greater peptide binding to pre-clinical xenograft breast cancer in vivo and to human specimens of invasive ductal carcinoma that express ErbB2 ex vivo. We used a scrambled peptide for control. Peptide biodistribution showed high tumor uptake by comparison with other organs to support safety. This novel integrated imaging strategy is promising for visualizing ErbB2 expression in breast tumors and serve as an adjunct during surgery to improve diagnostic accuracy, identify tumor margins, and stage early cancers.

Figure 1

NIR-labeled peptides. (A) Biochemical structure of ErbB2-specific peptide KSPNPRF (black) labeled with IRDye800 (red) via a GGGSC linker (blue), hereafter KSP*-IR800. (B) Scrambled peptide PPSNFKR is used for control, hereafter PPS*-IR800. KSP*-IR800 and PPS*-IR800 have (C) maximum absorbance at λ_ex = 780 nm and (D) peak emission at λ_em = 800 nm.

Figure 2

Specific peptide binding to ErbB2. KSP*-IR800 shows (A) strong staining to the surface (arrow) of BT474 human breast cancer cells but not to that for (B) MDA-MB-231. (C,D) Scrambled control peptide PPS*-IR800 shows minimal binding to either cell. (E) Quantitative comparison is shown in log₂ scale. We found 3.55-fold greater signal for KSP*-IR800 with BT474 versus MDA-MB-231 cells but only a 1.04-fold difference with PPS*-IR800. Using an ANOVA model fit with terms for 4 means to log-transformed data, we found the difference of differences to be significant. Western blot analysis shows ErbB2 expression for (F) BT474 and MDA-MB-231 cell and for (G) SKBR3 human breast cancer cells transfected with siErbB2 targeting siRNA (knockdown) and siCL non-targeting siRNA (control). (H,I) Use of a Cy5.5 label produces results to that found for IRDye800 with SKBR3 cells. (J,K) KSP*-Cy5.5 and anti-ErbB2-Cy5.5 bind significantly greater to the surface (arrows) of siCL (control) SKBR3 cells compared with that for (L,M) siErbB2 (knockdown) cells. (N) We found 4.42-fold greater signal for KSP*-Cy5.5 with siCL treated SkBr3SKBR3 control cells compared with that for siErbB2 knockdown cells and 1.60-fold greater intensities with anti-ErbB2-Cy5.5. Using an ANOVA model, we fit with terms for 4 means to log-transformed data, and found the difference of differences to be significant. (O) On competition, we found the mean fluorescence intensity with KSP*-Cy5.5 to SKBR3 cells decreases significantly in a concentration dependent manner with addition of unlabeled KSP*. By comparison, binding was significantly less affected with addition of unlabeled PPS*. We used a one-way ANOVA to compare the mean intensities, and show P-values at each concentration. (P–S) Binding of KSP*-Cy5.5 (red) and anti-ErbB2-AF488 (green) co-localizes to the surface (arrows) of SKBR3 cells with a Pearson’s correlation coefficient of ρ = 0.70. Each result is an average of 3 independent measurements.

Figure 3

In vivo photoacoustic imaging. (A) Ultrasound (US) and (B) MR (T₁-weighted, contrast-enhanced) images show structure of human xenograft breast (BT474) tumor (arrows) in nude mouse. Photoacoustic images collected at 1 hour post-injection of (C) KSP*-IR800 and (D) PPS*-IR800 show tumor expression of ErbB2 (arrows). (E) The T/B ratios from BT474 and MDA-MB-231 tumors over time show peak peptide uptake at 1 hour and return to baseline levels at ~24 hours post-injection. (F) T/B ratios from n = 11 tumors in n = 8 mice are shown at 1 hour post-injection. We fit an ANOVA model with terms for 4 means, and found 2.0-fold greater signal with KSP*-IR800 in BT474 versus MDA-MB-231 tumors. The difference of differences was significant, P = 1.9 × 10⁻³.

Figure 4

In vivo optical imaging. Whole body fluorescence images show increased uptake of (A) KSP*-IR800 compared with (B) PPS*-IR800 at 1 hour post-injection in human xenograft breast (BT474) tumor implanted in nude mouse. (C) Distal tip of dual-axes confocal endomicroscope was placed in contact (inset) with (D) tumor in live nude mouse. Optical sections collected in the (E) vertical (1000 × 430 μm²) and (F) horizontal (1000 × 1000 μm²) planes, respectively, show strong uptake of KSP*-IR800 in tumor (arrow). (G) At 1 hour post-injection, the mean T/B ratio for KSP*-IR800 was significantly greater than that for PPS*-IR800 in n = 9 tumors from n = 3 mice with a mean fold-difference of 2.0, P = 1.1 × 10⁻³ by paired, two-way t-test.

Figure 5

Peptide biodistribution. (A) White light image shows individual organs from tumor-bearing mouse euthanized 1 hour after injection of KSP*-IR800, including human breast cancer xenografts (tumors), liver, heart, stomach, brain, lung, spleen, kidney, small intestine, cecum, and colon. (B) Fluorescence image shows high peptide uptake in tumors by comparison with other organs. Strong signal from kidneys support renal clearance. (C) Quantified fluorescence intensities are shown for tumor and other organs from n = 5 mice. We fit a two-way ANOVA model to log-transformed data with terms for 11 tissues and 5 mice and tested each tissue mean against the mean for tumors. Kidney had 1.7 times more intense signals on average (P = 0.03), while all other tissues had at least 2-fold lower signals than tumor on average (P = 0.008 for lung was the largest P-value).

Figure 6

Specific peptide binding to human breast cancer ex vivo. (A) On immunofluorescence (IF), KSP*-IR800 shows strong staining to the surface (arrow) of invasive ductal carcinoma (IDC) cells from human specimens that express ErbB2, while minimal signal is seen with control (B) PPS*-IR800. (C) Immunohistochemistry (IHC) with anti-ErbB2 antibody confirms results. (D) Representative histology (H&E) for IDC. (E) Binding by KSP*-IR800 peptide (red) and anti-ErbB2-AF488 antibody (green) co-localizes on ErbB2 + IDC specimen with Pearson’s correlation coefficient of ρ = 0.53. (F) Magnified region from dashed box in (E) shows cell surface staining (arrow). (G) Quantitative comparison of KSP*-IR800 and PPS*-IR800 binding to human IDC (ErbB2+) with normal (ErbB2−) breast tissue. We fit an ANOVA model with terms for 4 conditions and 4 patients to log-transformed data and found a 2.44-fold greater signal for KSP*-IR800 in IDC than normal, but only a 1.30-fold increase for the same comparison with PPS*-IR800 peptide. The difference of differences was not significant, P = 0.09, which results from small number of human specimens. (H) On IF, we observed good staining of KSP*-IR800 to the surface (arrow) of non-invasive human ductal carcinoma in situ (DCIS) and (I) minimal signal with PPS*-IR800. On IF, we observed minimal staining to normal human breast tissue with either (J) KSP*-IR800 or (K) PPS*-IR800. (L) On IHC, we found strong reactivity to the surface (arrow) of DCIS cells. (M) Representative histology for DCIS. (N) Minimal reactivity was seen for normal human breast on IHC. (O) Representative histology for normal breast.