Introduction

The role of H2O2 as a second messenger, in regulating fundamental biological processes, has been identified not long ago and is increasingly supported by new data [1–7]. H2O2 is involved in therapeutic processes such as wound healing, anti-bacterial defense, stem cell proliferation, and an adaptive response in astrocytes that leads to neuronal protection [1, 5–10]. However, over-production of H2O2 exerts toxic effects on the cell and its surrounding environment. The aberrant production of H2O2 within cellular compartments is connected to serious pathological conditions such as cancer [11], ageing [12–14], diabetes [15], and neurodegenerative diseases [16–18].

A substantial challenge in elucidating the diverse roles of H2O2 in complex biological environments is measuring the spatial and temporal dynamics of this reactive oxygen metabolite. Much of the data implicating H2O2 both pathological and physiological roles in cultured cells and in vivo has been acquired using treatment with exogenous H2O2[19, 20], over expression of hydroperoxide inducing enzymes or H2O2 lysing/scavenging agents [1, 21], or tampering with ROS production cellular machinery such as NADPH oxidase (Nox) expressing cells [22]. Although these studies were critical in establishing crucial biological roles of H2O2, a key to fully understanding the mechanistic bases of redox biology is measuring the amount of H2O2 generated in specific intracellular compartments.

Traditional approaches to measuring H2O2 in living systems suffer from several issues: 1) The probes are often nonspecific and react with other reactive oxygen species such as hydroxyl radicals and superoxide, as well as reactive nitrogen species. 2) H2O2 is generally produced at a low concentration and can have a short half-life due to the activity of enzymes that neutralize it. 3) H2O2 is often produced in specific cellular compartments and rapidly diffuses across the membranes. For example, dichlorofluorescein (DCF), which has served as the workhorse for the redox biology community, detects multiple types of reactive small molecules, such as superoxide (O2•-), hydroperoxy radical (HO2•), singlet oxygen (1O2), peroxy radical (RO2•). DCF is not an ideal tool to determine the localization of H2O2 production. In addition to this lack of specificity of DCF for H2O2 detection, it can directly result in the creation of further ROS and thiol oxidation when exposed to UV radiation, and interacts with cytochrome c, rather than ROS as a consequence of apoptosis [23–25]. To overcome the aforementioned disadvantages for detecting ROS, recent efforts have aimed at 1) increasing the selectivity for H2O2 detection over related ROS, particularly superoxide, nitric oxide, and hydroxyl radical and 2) improving photostability and determining the localization of H2O2 production. Recently, specific and highly sensitive fluorescent H2O2 probes have been developed to circumvent these issues, including chemoselective fluorescent probes, fluorescent proteins, and nanoparticles [26, 27].

Confocal microscopy coupled with chemoselective fluorescent reporters permits the imaging of localized intracellular H2O2 levels. However, due to scattering and tissue penetration, imaging H2O2 levels in whole organisms often requires deep tissue imaging techniques. For example, two-photon microscopy (TPM) [25, 28] and photoluminescence can permit thick tissue imaging and in vivo studies [29, 30].

In this review, we describe and compare various methodologies for detection and imaging of H2O2 production in cells and whole organisms.

Fluorescent probes for H2O2 imaging

Small-molecule fluorescent probes

Small-molecule fluorescence probes for H2O2 are generally based on the oxidation–reduction processes between the H2O2 and reduced probe, which fluoresces upon oxidation. Several novel fluorescent probes capable of detecting H2O2 with high selectivity have been reported, and some of them have been used to monitor intracellular H2O2. A comprehensive list of fluorescent H2O2 probes is listed in Table 1.

Table 1 Small molecule probes for H 2 O 2 imaging
Full size table

One approach for the development of H2O2-selective probes utilizes a boronate-deprotection mechanism [2–4, 31, 38, 42, 45]. Figure 1 that displays the mechanism of hydrogen peroxide-mediated fluorescent enhancement and several example structures of fluorescent probes based on this mechanism. This detection strategy relies on the selective H2O2-mediated transformation of arylboronates to phenols. Arylboronates are appended to profluorescent molecules, such that reaction with H2O2 generates a fluorescent product. The monoboronate-based family of probes (PF-1, PF-2, PF-3, PO1, and PY1) can detect physiological changes in endogenous H2O2 levels. Since a wide color palette of such probes has been developed, various combinations can be selected for multicolor imaging experiments. The addition of acetoxymethyl ester-protected pro-anionic groups gives rise to the dye peroxyfluor-6 acetoxymethyl ester (PF6-AM), which increases cellular retention and further increases sensitivity to H2O2[2, 25]. Although boronate-deprotection-based probes have improved H2O2 localization studies, quantitative analysis of H2O2 generation using fluorescent probes is still challenging. Specifically, the signal from the single-wavelength emitting probes can be affected by the concentration of the probe. To address this, a monoboronate-based probe was synthesized, Ratio Peroxyfluor 1 (RPF)-1 that provides a ratiometric change of two fluorescent signals upon reaction with H2O2, which can potentially permit normalization to probe concentration [36]. Because peroxynitrite has also been shown to react with boronates to create a fluorescent product [52], it is critical to perform proper controls when using a boronate-based fluorescent probe, such as expression of catalase or using a peroxynitite-specific probe.

Figure 1
figure 1

Fluorescent turn-on mechanism and chemical structures of several examples of boronate-based H 2 O 2 fluorescent probes. (A) Lactone-opening mechanism of fluorescence-enhancement for mono-boronate xanthene-based H2O2 probes. (B) Several examples of lactone-opening-based monoboronate H2O2 fluorescent probes.

Full size image

Combining the boronate–based probe strategy with organelle-targetable functional groups has provided probes that can measure H2O2 levels with spatial resolution. In particular, several mitochondria-targeted probes have been generated, including Mitochondrial Peroxy Yellow 1 (MitoPY1) and SHP-Mito [25, 42, 43, 51], which utilize a positively charged phosphonium moiety for mitochondrial targeting. MitoPY1 and SHP-Mito can both be utilized as two-photon imaging probes [25, 51]. Nuclear Peroxy Emerald (NucPE) 1 is nuclear-localized fluorescent probe that utilizes a boronate protecting group to measure nuclear H2O2 levels, which has been shown to function in vivo[45].

All of the boronate-based probes react irreversibly with H2O2, meaning the fluorescent signal is based on the accumulated H2O2 generated. By contrast, Redoxfluor (RF)-1 uses a reversible disulfide-based redox sensing mechanism, allowing multiple reversible redox reactions in the cellular environment to be detected over time [41]. However, one disadvantage is that RF-1 is not selective for one particular ROS.

One limitation of ROS fluorescent probes is their application in in vivo studies. The visible light excitation affects probe photobleaching, tissue and organ penetration, and subsequent imaging detection. To tackle these issues, Peroxy Caged Luciferin (PCL)-1 and near-infrared (NIR) probe quinone Cy-7 (QCy7) were developed [30, 47, 49]. PCL-1 and PCL-2 are prosubstrates for luciferase and have been utilized for noninvasive in vivo H2O2 detection throughout whole mice [30, 47, 48]. NIR imaging of the Cy-7 is another attractive tool in animal studies due to the deep penetration of the NIR photons and low background fluorescence of the tissue. QCy7 enabled monitoring H2O2 signaling upon injection of lipopolysaccharides (LPS) into mice [49].

Nanoprobes

Compared with small-molecule fluorescent probes, nanoparticles offer several advantages [53, 54]: (1) Nanoparticles often have stronger fluorescent emission due to large number of molecular probes loaded in each single nanoprobe. (2) The high surface-area-over-volume ratio provides a higher probability for analyte detection. (3) Encapsulating small molecule fluorescent probes into nanoparticles can improve their stability. (4) The nanoparticle often serves as a protecting device for the sensory contents, protection from external interference present in biological environments (e.g., undesirable enzymatic reactions and nonspecific uptake by proteins). (5) Nanoprobes can possess multifunctionality, and target-specificity by conjugating ligand moieties onto the nanoparticle surface.

A number of nanoprobes were listed in Table 2. Peroxalate-based nanoparticles were recently developed which chemically excites the encapsulated dye, leading to light emission from the nanoparticles and the imaging of H2O2[29, 55–57]. Encapsulating fluorescent probes into nanoparticles can improve their stability. This nanoprobe was recently demonstrated by Lee et al. for in vivo imaging of global H2O2 in mouse model [29]. Semiconducting polymer-based nanoprobe CF-SPN has another advantage that combines chemiluminescence imaging with ratiometric imaging for liver-targeted detection of ONOO- and H2O2 simultaneously in the liver of living mice and in real time [58]. This nanoprobe demonstrated multifunctionality chemiluminescence imaging and ratiometric imaging for ROS sensing. For high sensitivity single-molecule detection, single-walled carbon nanotube (SWNT) embedded nanosensor exhibited high selectivity and sensitivity to single molecules of H2O2[59–61], which raises the potential for a hitherto unseen level of specificity in redox signaling analysis.

Table 2 Nanoparticles for H 2 O 2 detection
Full size table

Genetic fluorescent proteins

In the past, researchers have largely relied on the use of fluorescent dyes for ROS sensing due to their good sensitivity, high signal-to-noise ratio, cell permeability, and ease of measurement. The emerging genetic fluorescent proteins offer another option for high resolution selective H2O2 imaging. Allowing a dynamic measurement for the reversible detection of H2O2, Belousov, et al. developed a group of genetically encoded fluorescent proteins HyPer and their mutants to enable transient live-cell imaging [6, 70–75]. It is a ratiometric fluorescent indicator of H2O2 in which cpYFP is inserted into the regulatory domain of an Escherichia coli peroxide sensor OxyR [71]. HyPer is able to detect nanomolar concentrations of H2O2in vitro, to micromolar levels of H2O2 exogenously added to cells, or changes of intracellular H2O2 upon growth factor stimulation [71]. Due to genetic labeling, the HyPer family of genetically encoded fluorescent proteins have been successfully targeted to several cellular compartments such as the nucleus, cytosol, peroxisomes, mitochondria and the endoplasmic reticulum [70]. Thereby they allow for the intracellular spatial monitoring of H2O2 production, which can further improve H2O2 imaging with high signal-to-background noise from tissue auto-fluorescence in biological systems. HyPer probes are pH sensitive. To address this issue pH-specific probes are used as controls. HyPer-C199S would be an ideal control as a pH-sensitive and H2O2-insensitive version [76].

To improve the dynamic range of half-oxidation and half-reduction responses, HyPer-2 and HyPer-3 were developed which shows an expanded dynamic range. HyPer-3 showed faster oxidation-reduction kinetics and a higher fluorescence ratio (F500/F420) than what was reported for HyPer, demonstrating its advantage for H2O2 detection [77]. Both HyPer and HyPer-3 are applicable for fluorescence lifetime imaging microscopy (FLIM).

Redox-sensitive GFP (roGFP) [72, 73] coupled with yeast H2O2-sensing signaling peroxidase Orp1 [78] is a genetically encoded H2O2 sensor that has been used to detect and quantify physiological levels of H2O2in vivo. In roGFP2-Orp1, Orp1 relays a disulfide bridge to redox-sensitive GFP (roGFP). The redox equilibrium of the engineered cysteines is associated with measurable ratiometric fluorescent changes. Contrary to HyPer probes, this redox-dependent fluorescence is insensitive to pH changes in the physiological range. RoGFP2-Orp1 has been successfully used to measure physiologically relevant changes in H2O2 levels in Drosophila tissues and in living larvae by real-time imaging [79–81].

Optical imaging of H2O2 in biological system

Confocal microscopy

Confocal microscopy uses a scanning laser beam that is focused on the sample for imaging, with a pinhole placed in front of the detector. Confocal microscopy increases optical resolution and contrast by using a pinhole that prevents the out-of-focus photons from reaching the detector. It is the most popular imaging technique for H2O2 detection ex vivo. Most fluorescent probes were developed for confocal microscopy. However, confocal microscopy has limitations, such as optical scattering, photodamage, photobleaching, and limited imaging depth for use in real time in vivo studies. Furthermore, prolonged visible light exposure can result in artifactual ROS generation and signal amplification [82].

The recent advance of in vivo confocal microscopy is based on microendoscopy. By utilizing a miniprobe for confocal microendoscopy, PY1 was demonstrated for in vivo H2O2 imaging in colorectal cancer [83].

Two-photon microscopy

TPM provides sub-micron resolution imaging with lower phototoxicity and deeper tissue penetration than confocal microscopy [28]. In the two-photon process, a molecule simultaneously absorbs two photons whose individual energy is only half of the energy needed to excite that molecule, and then releases the energy to an emission photon. The main differences between confocal microscopy and TPM are the excitation light source and the fluorescence detection unit. TPM, including all commercial versions, is typically implemented in a laser scanning microscope equipped with a NIR ultrafast pulse laser.

TPM was demonstrated for imaging intracellular H2O2 production in live cells and tissues [25, 35, 50, 51]. Figure 2 shows TPM imaging of intracellular H2O2 in rat primary astrocytes using the chemoslective fluorescence probe PF6-AM. Figure 2A shows the H2O2 imaging mechanism of trappable probe PF6-AM [2]. Figure 2B shows TPM imaging of intracellular H2O2. As a comparison, Figure 2C shows confocal microscopy of the same cells using a 488 nm argon laser with the same fluorescence detection. Three arrows indicate strong light scattering in the same cells in Figure 2C. The TPM imaging here demonstrated the advantages of low scattering and low background noise.

Figure 2
figure 2

Fluorescence imaging of intracellular H 2 O 2 production using fluorescence probe PF6-AM (green). (A) Mechanism of Chemoselective H2O2 PF6-AM. (B) TPF imaging of H2O2 in astrocytes, fluorescence excited with a 770 nm Ti:sapphire laser. (C) Confocal microscopy of H2O2 in same astrocytes imaged in panel B, fluorescence excited with a 488 nm laser. The nuclei were stained with Hoechst 33342 (blue).

Full size image

For the deep tissue in vivo TPM, it has motivated new trends of technology development including long wavelength lasers [84–88], fast scanner [89, 90], and microendoscopes [91–93]. These techniques may further extend in vivo TPM for deep tissue H2O2 imaging in real time.

Ratiometric imaging

Ratiometric imaging is the division of one fluorescence channel by another one to derive the ratiometric channel. Ratiometric imaging has been widely used to detect intracellular ion concentrations, protein distributions, voltage or pH changes [94]. Compared to traditional fluorescence intensity imaging, ratiometric imaging relies on measuring a shift in emission instead of merely a change in intensity. It is extremely attractive for quantitative analysis because it corrects for unequal fluorophore labeling and photo-bleaching. Dual wavelength excitation/detection is the key for measuring emission shifts and intensity changes of fluorophores. The recent development of ratiometric H2O2 probes have been used for ratiometric imaging based on wide-field microscopy, confocal microscopy and TPM [35, 50, 51, 71, 77, 94–96].

Figure 3 shows TPM ratiometric image of a fresh rat hippocampal slice treated with H2O2 production. This imaging technique provides a solution for deep tissues H2O2 quantitative analysis.

Figure 3
figure 3

Ratiometric imaging of fresh rat hippocampal slice treated with H 2 O 2 . (A) The reaction between PN1 and H2O2 produced AN1 as the only major fluorescent product. (B) A hippocampal slice labeled with PN1. (C) Fluorescence spectra responses of 3 μM PN1 to 1 mM H2O2. Spectra were acquired at 0, 10, 20, 30, 40, 50, 60, and 120 min after H2O2 was added. (D) A hippocampal slice labeled with PN1 after pretreated with H2O2. Scale bars: 30 μm. The figures were adapted from ref. [50] with permission.

Full size image

FLIM

FLIM is an optical imaging technique based on the differences in the exponential decay rate of the fluorescence from a fluorescent sample [97, 98]. Because fluorescence lifetime τ is independent of indicator concentration, FLIM measurement is essentially insensitive to indicator expression level, non-uniform distribution, and partial photobleaching. FLIM generates absolute quantitative readouts while requiring only a single-wavelength excitation, provided that the indicator is calibrated in situ (e.g., in permeablized cells) or in vitro under conditions closely resembling intracellular environments. Figure 4 shows inflammation driven H2O2 production in zebra fish larvae using representative FLIM of HyPer-3 [77]. The pattern of the fluorescence lifetime changes indicated the gradient of H2O2 with higher concentrations of the oxidant at the wounding site. This imaging technique provides another solution for H2O2 quantitative analysis.

Figure 4
figure 4

FLIM of HyPer-3 response to H 2 O 2 production induced by inflammation in zebrafish larvae. (A) Left and right panels represent fluorescence intensity and FLIM images, respectively. ROI1 highlights the wound margin; ROI2 represents an area distant from the wound. (B) Fluorescence lifetime distribution plot for ROI1 and ROI2 in panel A. The figures were adapted from ref. [77] with permission.

Full size image

Chemi-/bioluminescence imaging

In vivo chemi-/bioluminescence imaging is a popular method to monitor enzymatic light emission by a living organism [99–101]. The detection signal is generally from red to near-infrared (NIR) light that offers deep depth imaging of H2O2 in organs. By using cooled charge-coupled device (CCD) cameras, this optical imaging modality gets high sensitive. The field of view is up to scores of centimeter covering a whole small animal, but the resolution is at the millimeter level. The recent advances of bioluminescence and chemiluminescence probes enable whole animal studies of H2O2 production [29, 30, 47–49, 55–57, 62, 102].

Figure 5 shows representative chemiluminescence images of global H2O2 production in a mouse model using Peroxalate nanoparticles [29]. H2O2 reacts with the peroxalate ester of (1) Peroxalate nanoparticles to produce a high-energy dioxetanedione intermediate within the nanoparticles (2), which then chemically excites the encapsulated dye, leading to light emission from the nanoparticles (3). Peroxalate nanoparticles were mixed with various concentrations of H2O2 and injected, intramuscularly, into the legs in Figure 5B. The concentration is (I) 10 μM H2O2, (II) 1 μM H2O2, (III) peroxalate nanoparticles only, and (IV) negative control. The chemi-/bioluminescence imaging is the only technique for whole animal global H2O2 monitor in real time.

Figure 5
figure 5

In vivo imaging of H 2 O 2 using peroxalate nanoparticles. (A) Peroxalate nanoparticle. (B) Chemiluminescence imaging. The figures were adapted from ref. [29] with permission.

Full size image

Conclusion

It is imperative to develop fluorescent probes that are able to monitor spatio-temporal intracellular H2O2 production in real time for live cells and in vivo studies. Among these fluorescent probes, chemoselective probes offer an attractive approach to H2O2 detection due to their general compatibility with an array of biological systems without external activating enzymes and genetic manipulation. Nanoprobes were recently developed for in vivo imaging of H2O2. A benefit arising from the multifunctional nanotechnology, probe is that it is designed for both chemiluminescence imaging and ratiometric imaging. Furthermore, genetic fluorescent probes that target redox sensitive proteins to specific cellular locations provide high sensitive targeted imaging technology for real time H2O2 imaging.

These emerging probes enable H2O2 detection using: 1) high resolution fluorescence imaging such as confocal microscopy and TPM; 2) large field of view global imaging involving in vivo chemi-/bioluminescence imaging; and 3) ratiometric imaging or FLIM for the quantification of cellular H2O2 levels.