Necrosis avid near infrared fluorescent cyanines for imaging cell death and their use to monitor therapeutic efficacy in mouse tumor models

Abstract

Quantification of tumor necrosis in cancer patients is of diagnostic value as the amount of necrosis is correlated with disease prognosis and it could also be used to predict early efficacy of anti-cancer treatments. In the present study, we identified two near infrared fluorescent (NIRF) carboxylated cyanines, HQ5 and IRDye 800CW (800CW), which possess strong necrosis avidity. In vitro studies showed that both dyes selectively bind to cytoplasmic proteins of dead cells that have lost membrane integrity. Affinity for cytoplasmic proteins was confirmed using quantitative structure activity relations modeling. In vivo results, using NIRF and optoacoustic imaging, confirmed the necrosis avid properties of HQ5 and 800CW in a mouse 4T1 breast cancer tumor model of spontaneous necrosis. Finally, in a mouse EL4 lymphoma tumor model, already 24 h post chemotherapy, a significant increase in 800CW fluorescence intensity was observed in treated compared to untreated tumors. In conclusion, we show, for the first time, that the NIRF carboxylated cyanines HQ5 and 800CW possess strong necrosis avid properties in vitro and in vivo. When translated to the clinic, these dyes may be used for diagnostic or prognostic purposes and for monitoring in vivo tumor response early after the start of treatment.

Keywords: cancer; cell death; cyanines; imaging; necrosis avid contrast agents.

Figure 1. Physicochemical characteristics and in vitro examination of the necrotic avid properties of the near infrared fluorophores (NIRF) HQ5 and 800CW

a. λabs = absorbance wavelength; λem = emission wavelength; Mw = molecular weight; AI = amphiphilic index; CBN = conjugated bond number; logP = log octanol-water partition coefficient; Z = electric charge. b. In the dry ice cell death assay an area of necrotic cells was induced in the center of a confluent monolayer of 4T1-luc2 murine breast cancer cells by applying dry ice for 15 sec to the underside of a culture well. Cells in the periphery of the culture well remained alive (schematically represented as, D = dead cells, L = living cells). Cell viability was confirmed by bioluminescent imaging (BLI). After 15 min incubation of the cells with HQ5 or 800CW and subsequent washing, a strong fluorescent signal was obtained from the area of dead cells. The 800CW-PEG signal was almost absent both in the areas of living and dead cells. c. FACS analyses performed with viable 4T1-luc2 cells and 4T1-luc2 cells treated with GA (4 uM) or Sta 3 uM) treatment, followed by staining with HQ5 or 800CW. The fluorescence intensity of HQ5 or 800CW stained dead cells was significantly increased compared to that of viable cells. FACS analyses were also performed with viable and GA treated cells, subsequently stained with 800CW-EGF, 800CW-2DG and 800CW-PEG. The fluorescence intensity of both 800CW-EGF and 800CW-2DG, in dead cells, was increased compared to viable cells. Using 800CW-PEG, there was no detectable difference in fluorescence intensity between viable, dead and unstained cells.

Figure 2. Confocal microscopic images of the localization of HQ5 in gambogic acid treated 4T1-luc2 cells

a. The bright-field view (BF) image shows the morphology of GA treated cells undergoing cell death. After treatment with GA, cells were stained with HQ5 (red), AVF (green) and PI (blue). A large number of cells stained positive for AVF. The merged image depicts a fluorescence overlay of HQ5, AVF and PI staining, indicating that cells stained with HQ5 were also positive for PI. b. At the level of a single necrotic cell it was shown that HQ5, appearing as a granular cytoplasmatic staining, did not co-localize with membrane AVF nor with PI nuclear staining. c–d. The granular HQ5 staining was further shown to have a great extent colocalization with Mitotracker (Mito) but not with Lysotracker (Lyso) in GA treated 4T1-luc2 cells. The bar represents 20 μm.

Figure 3. SDS-PAGE analyses of HQ5 and 800CW protein binding

a. SDS-PAGE gel electropherogram of cytoplasmatic- and membrane fractions of 4T1-luc2 cell lysate, incubated with HQ5 or 800CW. Protein binding of HQ5 and 800CW was observed in the cytoplasmic but not in the membrane fraction. Coomassie blue staining confirmed the presence of proteins in both fractions. HQ5 and 800CW staining showed a different pattern, albeit with some common features. b. Binding of HQ5 or 800CW, at different concentrations (0.1, 0.5 and 2 μM), to bovine serum albumin (BSA).

Figure 4. In vivo and ex vivo imaging of spontaneous 4T1-luc2 tumor necrosis with HQ5, 800CW, 800CW-EGF, 800CW-2DG and, 800CW-PEG

a. Representative in vivo whole body BLI and FLI images of a 4T1-luc2 tumor injected with HQ5. The BLI signal originated from the periphery of the tumor (red ring), whereas the HQ5 FLI signal mainly originated from the necrotic core (red spot). b. 3D reconstruction of the localization of HQ5 in a tumor using an automated fluorescence camera mounted on a cryo-microtome. c. Images of a HQ5 and corresponding TUNEL stained tumor section. d. Images of a 800CW, 800CW-2DG, 800CW-EGF and a 800CW-PEG and their corresponding TUNEL stained tumor sections. In contrast to HQ5, 800CW, 800CW-2DG and 800CW-EGF, 800CW-PEG did not co-localize with TUNEL staining.

Figure 5. In vivo opto-acoustic imaging of necrosis of spontaneous 4T1-luc2 tumor necrosis with HQ5

MSOT measurements showing the single wavelength anatomical image and unmixed signals of oxygenized haemoglobin (oxi-Hb), deoxygenized haemoglobin (deoxi-Hb) and HQ5. Detail of HQ5 targeting of the necrotic core in the center of the tumor.

Figure 6. Monitoring anti-tumor efficacy in a EL4-CBG99-luc lymphoma mouse model of chemotherapy

a. Representative whole body FLI images of tumor bearing mice treated with a combination of CTX and ETO and 24 hr later injected with 800CW. After another 24 h, in vivo whole body and ex vivo tumor FLI and BLI images were acquired and signal intensities were quantified. b. The BLI signals obtained from the treated animals were significantly lower as compared to those of the untreated controls (*p < 0.05). In contrast, the 800CW signals from the treated animals were significantly higher as compared to those of the untreated controls (***p < 0.001). c–d. Images of 800CW containing and a TUNEL stained tumor section of an (c) untreated tumor and a (d) treated tumor. The fluorescent signal obtained from a section of a treated tumor co-localized with TUNEL staining of the same section. Negligible 800CW fluorescence and TUNEL staining was observed in the untreated tumor. T = treated; U = Untreated.