Chemiluminescent probes for imaging cysteine cathepsin activity

Abstract

The cysteine cathepsins are a family of proteases that are widely expressed in diverse cell types. However, because their activity is highly upregulated at sites of inflammation, they are ideal targets for imaging a range of disease pathologies including cancer, atherosclerosis, fibrosis and others. While significant advances have been made in generation of activatable ‘smart’ probes for cathepsin activity, virtually all of these agents use fluorescent reporters which require a laser light source to induce a signal from the probe. We describe here the development and application of chemiluminescent probes for imaging cysteine cathepsin activity in cells and tissues. These probes provide rapid signal generation with low background in cells and tissues without the need for a light source. We demonstrate that a cathepsin probe containing an optimal peptide sequence can be used topically to detect activity in tumor tissues, demonstrating the potential value of this approach for real-time diagnosis of cancer as well as other conditions involving inflammation.

Graphical Abstract

Human cathepsins are lysosomal cysteine proteases involved in protein degradation and processing, and upregulation of cathepsins is often associated with cancer progression and inflammation.¹ As a result, they have been exploited as molecular imaging targets for detecting tumors and sites of inflammation, including fibrosis² and cancer.³ Previously, we developed fluorescent cathepsin probes, which demonstrated high signal-to-noise ratios in ear infection,⁴ fibrosis⁵ and 4T1 breast cancer models.^6–8 One drawback of the application of fluorescent probes is that they require an external light source for signal detection, which limits their application in resource-limited areas and results in overall high background from autofluorescence of tissues.

Chemiluminescence has garnered significant attention as an alternative to fluorescence imaging due to its potential to provide highly sensitive signals with low overall background without the need for specialized equipment.^9–11 In 1987, Paul Schapp and coworkers pioneered the concept of chemiluminescence induced by chemical reactivity, rather than oxidation. Specifically, a phenoxy-adamantyl-1,2-dioxetane compound was synthesized with a protecting group (PG) on the oxygen atom of the phenoxy moiety which could be removed by reaction with an enzyme or an analyte of interest. The removal of the protecting group triggers the release of adamantanone and the formation of an electronically excited benzoate ester intermediate, which decays to the ground state resulting in light emission.¹² However, a major limitation of Schapp’s dioxetane molecule is its low quantum yield in water, thus limiting its biological applicability. In 2017, Shabat and coworkers discovered that introducing electron groups at the ortho-positions of Schapp’s dioxetane can help overcome this issue, thereby dramatically enhancing its capability for fluorescent emission in aqueous media.¹³ This breakthrough opened up new opportunities for chemiluminescent labeling of a diverse array of biochemical targets in physiologically relevant settings. In particular, the chemiluminescence approach has been employed to develop probes for cathepsin B. These probes contain a cleavable peptide substrate specific to cathepsin B (Val-Cit) connected to a variable motif (R1) and a para-toluidine self-immolative linker, in which the methyl group is tethered to the phenoxy moiety of the dioxetane molecule (Scheme 1). When the peptide is cleaved by Cathepsin B, the linker is spontaneously released via 1,6-elimination to yield the unprotected phenolate capable of initiating chemiluminescence. In addition, Schapp’s dioxetane molecule, featuring an adamantly group, exhibits high chemical stability, typically resulting in slow chemiexcitation and, consequently, producing a ‘glow-mode’ of light emission over a time window of minutes to hours. To improve the solubility and cell permeability of the probes, Shabat and co-workers have attached linkers that lead to enhanced signal, enabling imaging of various cell lines.

Scheme 1. Mechanism of chemiluminescent probes.

(a) Cleavage of Cbz-Phe-Arg sequence in the presence of cathepsin. (b) Self-elimination reaction and Chemi oxidation steps during probe activation process. (c) Light emission as the result of energy transfer enables quantification of signal for biochemical and cellular assays.

Inspired by these prior studies, we previously developed an adamantyl-based chemiluminescent probe targeting the Mycobacterium tuberculosis Hip1 protease.¹⁴ This probe demonstrates linear correlation between signal intensity and Hip1 concentration at nanomolar levels, suggesting high sensitivity. The probe could be used to quantify Mtb bacterial levels in biological samples with high selectivity over related bacteria. Thus, chemiluminescent probes have a high potential to be used for applications in clinical diagnostics and image guided surgery.

As an alternative to the ‘glow-type’ slow onset chemiluminescent signals, Shabat and coworkers demonstrated that it is possible to generate a rapid and intense ‘flash-type’ chemiexcitation by spiro-coupling the dioxetane with more chemically labile moieties.¹⁵ However, excessive chemical instability can compromise the utility of probes in biological conditions as a result of loss of stability. Following a screen of dioxetane compounds with a variety of spiro-strained reactive groups, Shabat and co-workers identified a 6-membered sulfone ring that results in a 293-fold increase in chemiexcitation rate compared to the prior adamantane-containing probes while also displaying significantly higher chemical stability compared to cyclobutane. ¹⁶ By diversifying substrate sequences, it is possible to generate a library of chemiluminescent probes targeting distinct enzymes, enabling the identification of bacterial samples through enzymatic fingerprinting, thereby demonstrating the utility of these probes as diagnostic agents.¹⁷

Building upon these findings, we aimed to extend probe utility by incorporating chemical functionalities into peptide-based substrates to achieve high signal intensity that can be used for imaging tumor tissues without the need for an external light source, enabling their potential use in low-resource settings.

We design our probes based on prior studies of the overall substrate specificity of the primary cysteine cathepsins found at sites of inflammation (Figure 1). Specifically, we chose to use a dipeptide containing a Z-FR sequence fused to the chemiluminescent reporter. We synthesized both the adamantane-based (AD) and sulfonyl-based (SO₂) probes (see SI for methods and compound characterization). The principle of signal detection is depicted in Scheme 1, where AD probe refers to Z-FR-adamantane and SO₂ probe refers to Z-FR-sulfonate. When the probe is incubated with cathepsin or cells expressing cathepsin L, the Z-FR moiety is cleaved by the cathepsin, followed by a self-elimination reaction where the product goes through chemiexcitation to emit light, which is monitored with a plate reader or camera system for quantification. Based on previous studies,¹⁵ the traditional AD probe is expected to produce slow glow-mode chemiluminescence, while the SO₂ probe is expected to show faster flash-mode signal with higher overall signal intensity.

Figure 1. Chemiluminescent Probe Design and Structures.

The adamantane-based probe (AD-Probe, left) and a sulfonyl-based probe (SO₂-Probe, right) containing the cathepsin substrate sequence Z-FR.

To begin the validation of the probes, we chose to use cathepsin L as the ‘model’ cathepsin because it is available and highly active in recombinant form and is a good representative of the cysteine cathepsins. Specifically, cathepsin L is upregulated in human cancers¹⁸ and is a key regulator in macrophage polarization during inflammation.¹⁹ In addition, we have previously demonstrated that probes containing the core Z-Phe-Lys peptide sequence are effective substrates for cathepsins B, L and S which are the predominantly active enzymes in activated macrophages which supply most of the cathepsin activity in tumors.⁷ Therefore, we first evaluated the signal intensities and signal-to-noise ratios (SNRs) of the probes in a standard enzymatic assay using purified cathepsin L (Figure 2). In dose-response experiments with fixed probe concentrations (50 μM) and serially diluted enzymes (13.9 nM to 0.003 nM, two-fold), the SO₂ probe exhibited a rapid signal turn-on with signal-to-noise ratio (SNR) exceeding 40 at sub-nanomolar enzyme concentrations. Similarly, the AD probe achieved high SNR values above 20 at nanomolar concentrations; however, the luminescent signal lasted for over an hour. Both probes produced a signal that correlated linearly with the concentration of enzyme (data not shown). While the ‘flash’ mode of the SO₂ probe produced an overall brighter signal, the slower ‘glow mode’ of the AD probe, while potentially less sensitive, is likely more suitable for use in biological assays, where measurements for many biological processes often require extended time points. The clear difference in the “flash’ and ‘glow’ signals from the probes is the result of a large difference in the chemiexcitation rate with the SO₂ chromophore being 358 times faster than the AD chromophore.¹⁶ Because the rate of probe activation tracks with the rates of chemiexcitation of the chromophores, the rate limiting step in signal production is the chemienzymatic reaction rather than the protease mediated hydrolysis reaction. This is further supported by the rate constants for the chemienzymatic process (0.07 s⁻¹ for AD and 1.7 s⁻¹ for SO₂) which are lower than the k_cat values for cathepsin L cleavage of an optimal substrate (usually >2 s⁻¹).²⁰

Figure 2. Dose-Dependent Characterization of SO₂ and AD Chemiluminescent Probes.

(a – b) Relative luminescence units (RLU) measured for the SO₂-based probe (a) and AD-based probe (b) across a range of cathepsin L concentrations. (c–d) Corresponding signal-to-noise ratios (SNR) for the SO₂ probe (c) and AD probe (d) calculated under identical conditions.

To determine whether the probes are effective substrates for cathepsin L, we performed steady-state kinetic assays, where the probe was added at different concentrations with set concentrations of the enzyme. Total signal intensity was integrated and then converted to reaction velocity before being plotted against probe concentration (Figure 3). Both probes demonstrated Michaelis-Menten kinetics, suggesting they are cleaved by cathepsin L as substrates. The V_max of the SO₂ probe is twice that of the AD probe, which is consistent with higher signal intensity and faster reaction onset. As a reference, we measured the steady-state kinetics of the commercially available fluorescent probe Z-LR-AMC and found substrate inhibition, which is not observed from the chemiluminescent AD and SO₂ probes. The K_m values for the SO2 probe and AD probe are 36.2 μM and 54.4 μM, respectively. These rather high values are likely due to the large and potentially sterically hindered chemiluminescent moiety which reduces binding affinity of the substrates in the active site of cathepsin L. Regardless of these overall weak binding constants, the substrates were effectively cleaved by the cathepsin L.

Figure 3. Kinetic Characterization of SO₂ and AD Chemiluminescent Probes.

(a–b) Dose-dependent luminescence (RLU) profiles for the SO₂ probe (a) and AD probe (b) in human cathepsin L (hCatL) enzymatic assays. (c) Summary of steady-state kinetic parameters for both chemiluminescent probes. (d) Steady-state kinetic analysis of a fluorogenic substrate, showing substrate inhibition at higher concentrations.

After basic functional validation of the probes using purified enzymes, we next applied the probes to RAW macrophage cell suspensions. We have performed labeling of active cathepsins in these cells in prior studies and found the predominant activity to be derived from cathepsins B, L and S.⁷ Both probes showed comparable dose dependent signals (Figure 4a, c), suggesting efficient cleavage in RAW cells. To confirm that the probe signal results from cleavage by cathepsins, we pre-incubated the RAW cells with E64d, a cathepsin-specific inhibitor widely applied in cellular assays.²¹ For the AD probe, some decrease in signal was observed immediately after addition to the cell media. This drop in signal is likely due to the change in solvent from the DMSO stock to the aqueous media. For the SO₂ probe, the initial reduction in signal could not be observed because it begins to degrade and produce high background immediately after addition to the media. Preincubation of the cells with E64d to fully block cathepsin activity fully blocked the signal for the AD probe, while it only slightly reduced the signal for the SO₂ probe. This result is consistent with the overall low aqueous stability of the SO₂ probe which results in a high background signal that cannot be blocked by addition of the cathepsin inhibitor. Quantification of the signal from the cellular assay revealed a dose-response trend, consistent with the results from the enzymatic assays. Thus, the AD probe is suitable for in vivo assays where cathepsins are present and upregulated at inflammatory sites.

Figure 4. Cellular Characterization of SO₂ and AD Chemiluminescent Probes.

(a–b) Luminescent signals from RAW 264.7 macrophages treated with or without inhibitor and incubated with the AD probe. (c–d) Luminescent signals from macrophages treated with or without inhibitor and incubated with the SO₂ probe. (e) Linear correlation between cell number and integrated luminescent signal (RLU) measured by plate reader.

In addition to measuring fluorescence with a plate reader, we performed imaging studies using the LagoX imaging system to quantify total signals (Figure 5a). The images demonstrate that the AD probe produces a strong signal that could be detected by a camera system rather than measured using a more sophisticated plate reader. To test whether the probes could be used for direct imaging of whole tissues to identify sites of inflammation, we imaged the tumors harvested from 4T1 mice with normal healthy organs as a control which we have used extensively for validation of fluorescent cathepsin probes.⁷ For these studies, we topically applied the chemiluminescent probes in DMSO solution, followed by incubation for 45 min prior to collecting images. We observed signals in tumors after 30 minutes of incubation of probes that were absent in the PBS treated negative controls (Figure 5b). We observed significant signal over background in the tumor tissues, consistent with previously reported high levels of cathepsins B, L and S in these tumors.⁷ Importantly, the background observed for normal, healthy tissues was minimal for the AD probe, consistent with our results in cells. Given the high non-specific activation of the SO₂ probe, the AD ‘glow’ probe is likely more suitable for applications in ex vivo imaging of tissue samples.

Figure 5. In Cellulo and Ex Vivo Imaging of Chemiluminescent Probes.

(a) Plate image of RAW 264.7 macrophages at indicated cell numbers incubated with chemiluminescent probes, with or without E64d (100 μM for 1 hour) inhibition (b) Ex vivo imaging of mouse organs topically treated with chemiluminescent probes by treatment with 50 μL of 100 μM probe stock solution in PBS solution for 45 min followed by imaging. Organs shown from top to bottom: tumor, adipose tissue, liver, spleen, kidney, heart, and lung. PBS-treated tissues serve as blank controls.

In summary, we synthesized two chemiluminescent probes that contain a recognition peptide sequence that is effectively cleaved by cysteine cathepsins. Enzymatic and cellular assays demonstrate a dose-dependent signal with high SNR ratios that can be measured using a plate-reader assay. The ability to block the signals in cellular assays using a broad-spectrum cysteine cathepsin inhibitor suggests that the probes have high overall target specificity. Furthermore, the probes provide signals in tumor tissues when topically applied that can be measured using camera-based imaging systems. Interestingly, the faster and brighter ‘flash’ SO₂ probe showed overall higher background in normal tissues such as kidney, heart and lung. Although the SO₂ probe has the potential to be faster and brighter, this comes at the cost of lower stability and selectivity. Therefore, the AD-based ‘glow’ probe is likely a more optimal choice for use for imaging cathepsin activity in excised tissue samples after topical application. The probes presented here have the potential to be used for applications in tumor margin detection as well as for imaging of sites of inflammation where cathepsin activity is highly elevated. The AD probe described here also has the potential to be further engineered to increase selectivity and turnover rate through substrate modifications, resulting in overall higher signals. Ultimately, if probes can be engineered to produce high amounts of light, it should be possible to read out diagnostic tests using the naked eye or simple diode-based detectors such as cell phone cameras. This would enable new clinical applications in low resource settings and circumvent the need for expensive detectors for effective use of the probes.

Highlights.

• Development of a cathepsin-targeted chemiluminescent substrate probe

• Probe provides high signal and low background

• Probe signal can be detected without the need for an external light source

• Probe can be used to image cathepsin activity in cells and tissues

• Probe can be applied topically to enable detection of tumors

Appendix A. Supplementary data

Synthesis and characterization of the probes, enzymatic and cellular characterization, and imaging protocols are described fully in supplementary data.

Data availability

Supplementary information is given in full.