Fluorescent contrast agents for tumor surgery.

Cancer is a leading cause of cases of mortality worldwide. The most effective method to cure solid tumors is surgery. Every year, >50% of cancer patients receive surgery to remove solid tumors. Surgery may increase the cure rate of most solid tumors by 4-11 fold. Surgery has many challenges, including identifying small lesions, locating metastases and confirming complete tumor removal. Fluorescence guidance describes a new approach to improve surgical accuracy. Near-infrared fluorescence imaging allows for real-time early diagnosis and intraoperative imaging of lesion tissue. The results of previous preclinical studies in the field of near-infrared fluorescence imaging are promising. This review provides examples introducing the three kinds of fluorescent dyes: The passive fluorescent dye indocyanine green, which has been approved by the Food and Drug Administration for clinical use in the USA, the fluorescent prodrug 5-aminolevulinic acid, a porphyrin precursor in the heme synthesis, and biomarker-targeted fluorescent dyes, which allow conjugation to different target sites.

Introduction

Cancer is a leading cause of cases of mortality worldwide (1). Surgery is an effective method used to remove solid tumors, with >50% of cancer patients undergoing surgery each year (2). As a matter of fact, surgical removal of the tumor may increase the cure rate of most solid tumor types by 4–11 fold (3).

Failure to obtain complete disease clearance due to incomplete resection, including positive tumor margins or metastatic cancer cells in lymph nodes, is a major challenge in tumor surgery and occurs in 20–60% of operations (2). Tumor cells may spread to distant host tissues, leading to metastatic disease, a well-known cause of mortality in patients with cancer (4). Following treatment, high levels of metastasis and the recurrence of cancer may be observed due to incomplete removal of the edges of the primary tumor (4). Surgery holds different challenges, including identifying small lesions, locating metastases, as well as confirming complete tumor removal (5–9).

To improve surgical accuracy, fluorescence guidance is an advisable approach. Near-infrared fluorescence (NIRF) imaging displays promising results in preclinical studies, allowing for real-time early diagnosis and intraoperative imaging lesion tissue (10–12). It describes the non-invasive use of near-infrared light to excite the contrast agent, after which the intensity of contrast agent fluorescence can be detected. Thus, the fluorescence represents the transformation of the molecular structure of the contrast agent, which is tissues in different diseases (4). NIRF guidance was introduced to improve the identification of lesions and guide the removal of these lesions (13). Compared with the more traditional approach of molecular imaging, which involves a radioactive tracer at cm resolution, NIRF provides higher resolution, allowing the identification of numerous details on the surface of the tissue (14). The fluorescence guidance technology is limited by the strong attenuation of the signal, meaning the technology would lose be less accurate with increasing depth (15). Intraoperative fluorescent molecular imaging agents have emerged as an innovative approach to guide surgical resection. There are three types of fluorescent molecular imaging agents, which include passive fluorescent dyes, ‘pro-dye’ fluorescent agents and biomarker-targeted fluorescent dyes. The current review introduces these types kinds of fluorescent contrast agents by using examples of each.

Passive fluorescent dye indocyanine green (ICG)

ICG (Fig. 1) is a near-infrared (NIR) contrast dye, which has been approved by the Food and Drug Administration (FDA) for clinical use in the USA (3). ICG is a tricarbocyanine dye and a water-soluble organic compound synthesized in the Kodak Research laboratories in 1955, it has been used to aid in medical diagnoses and to evaluate blood flow (16–18). ICG is able to easily penetrate tissues and cells, with an adverse reaction rate of <0.1%. In the past, several studies have demonstrated that ICG may be accumulating in metastatic tumors in the liver (19–23). As ICG is associated with biliary excretion, accumulation of ICG in cancerous tissue has been demonstrated to provide excellent contrast of intrahepatic nodules during surgery (24).

Figure 1.

Structure of indocyanine green.

The mechanism of ICG accumulation in a tumor remains elusive. Previous studies have demonstrated that ICG undergoes hepatobiliary excretion (25–28). The excretion of ICG into the liver then bile may impact its clearance in different types of tumors. For hepatic tumors, it is assumed that organic-anion transporting polypeptides expressed on liver cells, transporter proteins and intracellular transporter proteins give rise to the tumor contrast (24,29). For non-liver tumors, the enhanced permeability and retention (EPR) effect is the primary mechanism for the accumulation of ICG in solid carcinomas (30–33). The EPR mechanism has been associated with tumor environments, such as blood pressure, pH, vascular endothelial cell separation, differences in local prostaglandins and bradykinin levels and the lack of angiogenesis in lymphatic vessels (24,34).

Described by Matsumura and Maeda (35) for the first time in 1986, the EPR effect described defects in endothelial cells lead to the systematic and passive accumulation of small molecules, such as ICG, into the walls of tumor blood vessels. Once in the tumor microenvironment, the dye molecules are retained due to global properties, including shape, size, charge and polarity, rather than the tumor-specific ligand-receptor interaction mechanism (36).

The molecular structure of ICG comprises hydrophilic and hydrophobic moieties (37). Driven by its inherent chemical structure, ICG interacts with lipoprotein (LP) and phospholipids (38). ICG combines with LP in human blood circulation (39–41). LP interacts with the hydrophilic end of ICG and forms a complex (ICG-LP) with improved affinity for hydrophobic groups. During necrosis, the hydrophobic tails of phospholipids are exposed and changes in the affinity of ICG-LP to the ruptured lipid layer are observed (Fig. 2) (37). In addition, certain diseases, including malignant tumors, inflammation or trauma, may increase vascular permeability, allowing ICG-LP complexes to penetrate the walls of healthy blood vessels (39,42).

Figure 2.

Normal lipid bilayer vs. necrotic lipid bilayer. Necrosis exposes the hydrophobic tails of the phospholipids. Indocyanine green-lipoprotein exhibits a distinct affinity for the phospholipids from the ruptured lipid bilayer.

Onda et al (43) revealed that 30 min after administration, ICG was internalized into tumor cells, where it remained for at least 24 h. In normal tissue rapid clearance occurred. In the vicinity of 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose, ICG exhibited rapid non-tissue specific extravasation, perhaps due to reversible non-covalent interactions with serum proteins, like albumin, displaying EPR effects contributing to ICG accumulation within the tumor tissue (30–33).

It was further indicated that the intracellular accumulation of ICG may increase as the temperature increases; also, the authors suggested that ICG may be absorbed into the cells by binding to the cell membrane (43). Two molecules may influence the process of ICG uptake: Phospholipids and Pitstop2. The ability of ICG to interact with phospholipids allows it to bind to the cell membrane, the cells then uptake ICG; Pitstop2, the grid protein-dependent endocytosis inhibitor, is activated through the binding of extracellular molecules to the cell membrane and inhibits the uptake of ICG (44). Furthermore, the authors recommended that ICG may be absorbed into the cells by binding to the cell membrane (43).

The conflicting results of the fluorescence imaging often depend on tumor type, staging and microenvironment. The fluorescence emitted by ICG only penetrates 5–10 mm into the tissue, so the depth of the tumor influences the imaging result (45). Hill et al (46) stated that human leukocyte antigen (HLA) is a natural, biodegradable substance; ICG (0.0026–0.0052 mmol, 2.0–4.0 mg) loaded into HLA may become a nanoparticle. Hill et al (46) also indicated that tumor contrast with ICG nanoparticles was significantly improved compared with the use of regular ICG. This indicates that the size of ICG may influence the fluorescence image results.

A PubMed analysis of papers published over the last five years using ICG in surgery by tissue or cell type is given in Fig. 3. Liver cancer exhibits the highest publication numbers describing the use of ICG in surgery, followed by breast and cervical cancer. Publications describing the use of ICG in surgery have increased between 2011 and 2016, as Fig. 4 indicates.

Figure 3.

Number of papers published between 2011 and 2016, according to PubMed, describing the use of indocyanine green for surgery guidance for different types of tumor.

Figure 4.

Number of papers published between 2011 and 2016, according to PubMed, describing the use of indocyanine green for surgery guidance over time.

‘Pro-dye’ fluorescent agent 5-ALA

5-aminolevulinic acid (5-ALA; Fig. 5) is a non-protein amino acid, which is a precursor in the porphyrin biosynthesis (47,48). The ability of 5-ALA to act as a fluorescent agent originates from the accumulation of the light-sensitive protoporphyrin-IX (PpIX), which exhibits a distinct fluorescence spectrum when exposed to a certain wavelength (49). Through the insertion of a ferrous iron (Fe2+), PpIX (Fig. 6) forms heme B, the prosthetic group of hemoglobin, myoglobin, cytochromes, catalases and peroxidases (47). PpIX is photosensitive, absorbing blue light (375–440 nm) and emitting red to pink fluorescence (~635 nm) (48–50). Exogenous 5-ALA is the most commonly used molecule as a photosensitizing agent in intra-operative photodynamic detection of tumor tissue (51), and may become a ‘pro-dye’ fluorescence agent in fluorescence-guided surgery (FGS) (52–54).

Figure 5.

Structure of 5-aminolevulinic acid.

Figure 6.

Structure of protoporphyrin-IX.

Improved PpIX fluorescence following 5-ALA treatment is observed in different types of tumor cells and tissues (Fig. 7), validated through a comparison with a control group (55). Extensive research has demonstrated that increased PpIX fluorescence in tumor cells may be the result of influencing various tumor-associated properties, including heme biosynthesis, mitochondrial function and changes in porphyrin transporters (56).

Figure 7.

Number of papers published between 2011 and 2016, according to PubMed, describing the use of 5-aminolevulinic acid for surgery guidance for different types of tumor.

The activity and expression profile of enzymes participating in heme biosynthesis differ between tumor and healthy cells or tissues. Eight enzymes were included in the heme biosynthesis pathway (57). Comparing the expression level of genes or activity of enzymes involved in heme biosynthesis between tumor, normal cells and tissues from studies indicated that the following enzymes exhibited significant differences in activity (21,58–72). The first enzyme in the heme biosynthesis is called ALA synthase, which catalyzes the formation of 5-ALA from glycine and succinyl-coenzyme A (CoA). Following the migration of 5-ALA from the mitochondrial matrix to the cytoplasm, ALA dehydratase, also referred to as porphobilinogen synthase, catalyzes the formation of porphobilinogen (PBG), by combining two molecules of 5-ALA. The connection of four PBG molecules yielding in hydroxymethyl bilane is catalyzed by the porphobilinogen deaminase (also known as hydroxylmethylbilane synthase). Uroporphyrinogen III decarboxylase (UROD) is involved in the fifth enzymatic step of the heme biosynthesis pathway, where uroporphyrinogen III is decarboxylated by UROD giving proporphyrinogen III. Ferrochelatase (FECH) catalyzes the conversion of PpIX to heme b, making it the last enzyme utilized in the heme biosynthesis.

In order to make it clear that these enzymes are differentially expressed in the tumors, Table I (21,58–72) illustrates the changes observed in gene expression and enzyme activity linked to the heme biosynthesis pathway of various tumor tissues compared to normal tissue. The data provided may further be used a guide aiding the decision as to which tumor types may exhibit improved surgical results through the use of 5-ALA-mediated PpIX florescence.

Table I.

Changes in enzymatic activity and gene expression of enzymes participating in the heme biosynthesis pathway in various tumor tissues.

Enzyme	Cancer type	Effects
ALAS	Colorectal cancer (58)	Gene expression significantly lower
	HCC4017 non-small-cell lung cancer (59)	Gene expression and protein level increased
	Lung cancer xenograft tumor (59)	Protein level increased
PBGD	Cervical cancer (60)	Gene expression and enzymatic activity increased
	Prostate cancer (61)
	Breast cancer (62)
	Meningioma (21)
	Bladder cancer (63)	Enzymatic activity increased
	Colon cancer (64)
	Barrett's esophagus (65,66)
	Esophageal cancer (55,65)
UROD	Friend virus-induced erythroleukemia (mice) (67)	Gene expression or enzyme activity increased in initiation an progress
	Breast tumor (62)	Enzyme activity increased
	Head and neck cancer (68)	Gene expression increased
FECH	Liver cancer (69)	Enzyme activity decreased
	Bladder cancer (63)	Enzyme activity decreased
	Colorectal cancer (58,64)	Gene expression decreased
	Esophageal cancer (58)	Gene expression decreased
	Gastric cancer (58)	Gene expression decreased
	Rectal cancer (58)	Gene expression decreased
	Colon cancer (58)	Gene expression decreased
	Urothelial cancer (70)
	Glioma cancer (71)
	Breast cancer (72)

ALAS, 5-aminolevulinic acid synthase; PBGD, porphobilinogen deaminase; UROD, uroporphyrinogen III decarboxylase; FECH, ferrochelatase.

Succinyl-CoA, one of the two starting materials of the PpIX/heme biosynthesis, is a metabolite produced in the tricarboxylic acid (TCA) cycle. In order to prevent the accumulation of metabolites from the TCA cycle, as well as mitochondrial NADH, a connection between the glucose metabolism, TCA cycle and the heme biosynthesis has been established (73). That is to say, Succinyl-CoA, a metabolite produced in the TCA cycle, participated in the first step of PpIX/heme biosynthesis.

In cancer cells, metabolic reprogramming from the TCA cycle into aerobic glycolysis, generating glutamine for energy, may lead to the accumulation of TCA cycle metabolites and the activation of the heme biosynthesis to remove those metabolites (74). Activation of heme biosynthesis may lead to PpIX accumulation due to FECH saturation (56). Effective 5-ALA absorption and the transport of different porphyrin metabolites may further affect the PpIX accumulation in cells. In theory, the improved 5-ALA-PpIX in tumor cells may be triggered by certain processes, including elevated ALA uptake, improved porphyrin activity and reduced PpIX activity. An increase in 5-ALA uptake has been identified through elevated levels of PpIX in tumor cells (75). Furthermore, studies have indicated that high and low PpIX 5-ALA absorption were not significantly different between cell lines (64,76,77). Nakanishi et al (78) demonstrated that there was no correlation between 5-ALA-induced PpIX accumulation and the uptake clearance of 5-ALA. The aforementioned study also revealed that ALA uptake rates were far greater than maximum conversion rates of ALA to PpIX in LS-180, T24, A2780, DU145 and MCF-7 cell lines. ALA uptake is not the only decisive factor to enhance ALA-PpIX fluorescence in tumor cells. A porphyrin transporter that is postulated to be linked to the porphyrin synthesis is the adenosine 5′-triphosphate-binding cassette subgroup B member 6 (ABCB6), which was originally described as a transporter protein on the outer mitochondrial outer membrane (79). ABCB6 interacts with various porphyrins, including coproporphyrinogen III, PpIX and hemoglobin, with the highest affinity recorded for coproporphyrinogen III (46). Therefore, ABCB6 was thought to be primarily involved in the transporting coproporphyrin III into the mitochondria for PpIX/heme b synthesis (79). Increased ABCB6 expression has been linked to an increase in fluorescence in human glioma tissues allowing for better contrast in fluorescence-guided surgery, via more sufficient PpIX accumulation (80). ABCB6 is located in the cell membrane and Golgi apparatus, and transports coproporphyrinogen III between the cellular departments (81–83). An enhanced ABCB6 function may be observed at increased coproporphyrinogen III concentration, reducing the intracellular concentration of PpIX/hemoglobin (46). The net influence of ABCB6 on 5-ALA-PpIX levels in the cells may depend on the relative ABCB6 activity in the mitochondria and cell membranes (56).

In the plasma membrane, ATP-binding cassette sub-family G member 2 (ABCG2), a transporter, serves the most important role in transporting PpIX. Studies have demonstrated that increased ABCG2 activity reduces the intracellular PpIX level following 5-ALA stimulation, and the cell lines with high ABCG2 expression or activity often exhibit decreased 5-ALA-PpIX fluorescence (84,85). Robey et al (84) indicated that the use of ABCG2 transport inhibitors would enhance 5-ALA-PpIX fluorescence.

Biomarker-targeted fluorescent dyes

Several NIR fluorescent dyes have been developed and incorporated, for example with antibodies (86,87), nanoparticles (88) or encapsulated within nanomaterials (89,90), to be used as contrast agents for molecular imaging of different tumors (4). Researchers have identified that elevated levels of fibroblast activation protein (FAP) in stromal fibroblasts are associated with aggressive cancer types (91–95). FAP is a type II salivary glycoprotein with the ability to cleave biological peptides, including collagen and proteolytic enzymes, and serve a central role in the aggressiveness of the solid tumors. FAP is expressed in stromal fibroblasts of several types of cancer, but not in healthy tissue; it is used as a tumor marker that has drawn increasing attention (91,96). Rüger et al (90) linked anti-single-chain variable fragment directed against FAP antibody fragments to quenched liposomes, which became a novel fluorescence diagnostic contrast dye termed anti-FAP-IL. Anti-FAP-IL antibodies were used to ensure the specificity and fluorescence imaging of FAP expression cells and tumor muscle fibroblasts in mice xenotransplantation (96).

Carbohydrate antigen 19.9 (CA19.9) is a ligand of epithelial leukocyte adhesion molecules and its overexpression has been found in some malignancies as well as in some non-malignant conditions (97–101). CA19.9 is an attractive target for pancreatic ductal adenocarcinoma (PDAC) imaging, due to its high expression on tumors, compared with healthy pancreatic tissue (102,103). The usage of CA19.9 as biomarkers for PDAC led to the identification of several antibodies, including the characterization of the fully human monoclonal antibody 5B1, which binds to extracellular epitopes of CA19.9 with low nanomolar affinity (104–106). So Houghton et al (107) generated three modular tools, including 89Zr-ssDFO-5B1, ssFL-5B1 and 89Zr-ssdual-5B1. These modular tools can target CA19.9, which is an important molecule in invasion and metastasis of many cancers, including PDAC (103). The results revealed that the three modular tools evaluated displayed excellent uptake in the CA19.9 positive xenograph model of PDAC, indicating that each of them is likely to improve the detection rates of tumor of patients with PDAC (107).

The fluorescent gold nanoparticles synthesized by Li et al (108) bind to diatrizoic acid and the nucleolin-targeted AS1411 aptamer. This is a type of fluorescence-guided aptamer-targeted probe. Apart from providing visible fluorescence for detecting, the probe also exhibited high water-solubility, good biocompatibility and strong X-ray attenuation used in computed tomography (CT) contrast enhancement. The probes were intravenously injected into CL1-5 tumor-bearing mice and detection experiments, which included CT imaging and fluorescence detection 30 min post injection, were performed. The results demonstrated that fluorescence nanoparticle conjugates, used as molecular imaging agents to indicate the tumor location by CT imaging, may be easily observed on CT images with the naked eye (108).

Cysteine protease is another biomarker that is highly upregulated in the tumor cells and the surrounding matrix of tumor support cells in multiple types of cancers (109). Fluorescent contrast agents that may be helpful for dynamic monitoring in vivo and used as imaging contrast agents for FGS may improve the detection rates for tumors (110). Researchers designed and synthesized a series of NIR fluorescent probes, using the latent lysosomotropic effect to promote the cell retention of protease activation. These probes exhibit tumor-specific retention, rapid activation kinetics and rapid system distribution. Furthermore, they may be used to detect multiple types of cancer, including breast, colon and lung cancer (110).

The most common biomarker used for targeted fluorescence is folate, a B vitamin involved in metabolic processes, including DNA and RNA synthesis, epigenetic processes, cell proliferation and lung adenocarcinoma survival (111). The folate receptor (FR) family consists of four members, with only FR-α and FR-β displaying high affinities for folic acid. When expressed in the cavity surface of polarized epithelial cells, FR-α is able to prevent binding of the serum folate salt (111–114). The FR-α expression in lung adenocarcinoma (1–3,000,000 receptor/cancer cells) appears more connected with serum folic acid than in normal pulmonary epithelial cells (115–118). For the purpose of diagnosis, FR-α provides a reasonable molecular target for lung adenocarcinoma. Two contrast agents, EC17 and OTL38, have been proposed to image ovarian and lung adenocarcinomas during surgery (12,119). These agents are similar in that they target FR-α via a folate ligand. Although EC17 and OTL38 use the same ligand, they have two different fluorochromes: EC17 contains a fluorescein dye and OTL38 contains a cyanine dye (118). Fluorescein is in the visible wavelength and the cyanine is in the NIR range. De Jesus et al (118) revealed that OTL38 appears to have superior sensitivity and brightness compared to EC17 in a preclinical testing. This conclusion is consistent with the accepted belief that NIR dyes exhibit less auto fluorescence and scattering compared with visible wavelength fluorochromes (118).

Conclusions

Fluorescent contrast agents may guide surgeons in making real-time decisions during surgery. ICG is an NIR contrast dye, which was approved by the FDA for clinical use in the USA. It is a water-soluble organic compound, which may easily penetrate tissues and cells with an adverse reaction rate of <0.1%. The EPR influence is the major mechanism by which ICG accumulates in solid cancer. ICG is processed by the excretory pathways of the biliary system and may offer superiority in some tumor nodules during surgery. 5-ALA is a natural amino acid and a natural prodrug that metabolizes to the heme precursor PpIX. 5-ALA, through oral administration, increases the PpIX accumulation in the tumor tissue and subsequent photosensitizing may guide tumor resection. Fluorescent dyes have been developed and combined with antibodies or nanoparticles to function as contrast agents for molecular imaging by increasing the binding to the target site and providing more accurate information during tumor resection. Increasing studies focus on combining these different advantages into one dye, which it is believed will further the development of fluorescent contrast agents. Table II provides a brief summary of targeting abilities and methods of administration of the three types fluorescent contrast agents discussed in the current review. Combining these imaging agents for clinical use may provide more options in tumor surgeries. The application of contrast agents may significantly improve the surgery outcome.

Table II.

Comparison of three different fluorescence-based dyes.

Dye	Targeted	Administration
ICG	No	Injection
5-ALA	No	Injection or oral
Biomarker	Yes	Depends on the dye

ICG, indocyanine green; 5-ALA, 5-aminolevulinic acid.