Self-Assembled Lanthanide Antenna Glutathione Sensor for the Study of Immune Cells.

The small molecule 8-methoxy-2-oxo-1,2,4,5-tetrahydrocyclopenta[de]quinoline-3-carboxylic acid (2b) behaves as a reactive non-fluorescent Michael acceptor, which after reaction with thiols becomes fluorescent, and an efficient Eu3+ antenna, after self-assembling with this cation in water. This behavior makes 2b a highly selective GSH biosensor, which has demonstrated high potential for studies in murine and human cells of the immune system (CD4+ T, CD8+ T, and B cells) using flow cytometry. GSH can be monitored by the fluorescence of the product of addition to 2b (445 nm) or by the luminescence of Eu3+ (592 nm). 2b was able to capture baseline differences in GSH intracellular levels among murine and human CD4+ T, CD8+ T, and B cells. We also successfully used 2b to monitor intracellular changes in GSH associated with the metabolic variations governing the induction of CD4+ naïve T cells into regulatory T cells (TREG).

Biologically active thiols known as biothiols, which include cysteine (Cys), homocysteine (Hcy), glutathione (GSH), and hydrogen sulfide (H2S), play a central role in the intracellular regulation of redox homeostasis and in the maintenance of cellular functions, such as post-translational modifications, biocatalysis, metal binding, and xenobiotic detoxification.[1,2] Oxidative stress is a key feature of a wide variety of chronic and degenerative diseases, and changes in the levels of biothiols have been associated with various diseases.[3−8] Distinct responses to metabolic stimuli (bioenergetic signatures) have been associated with differences in the immune function.[9,10] In recent years, several studies have shed light on the dynamic and sophisticated connection between metabolic programs and the function of specialized cells in the immune system.[10,11] This crucial role of metabolism in the control of immune processes, including inflammation, has led to the emergence of a new field of immunometabolism.[11−13] It is increasingly recognized that biothiols play a key role in regulating the metabolic adaptability and thereby the function of cells of the immune system.[14−21] One of the latest discoveries in this field is the regulation of functions through the synthesis and release of various biothiols, in particular, GSH, which affects the metabolism and function of the immune system’s effector cells.[12,17,21−24] Consequently, the interest in developing tools to monitor biothiol levels in immune cells in clinical samples has grown exponentially. To this aim, diverse probes and techniques have been developed for the detection of biothiols. Among the methods used, those based on fluorescence emission are among those that provide the greatest advantages due to their simplicity, low detection limits, and ease of use.[25−28] However, selective and sensitive methods to detect and monitor GSH in cells with flow cytometry, a fluorescence-based, gold-standard tool for the identification and classification of cellular populations, remain an unmet need in the immunology field. However, current methods to measure GSH lack selectivity and sensitivity, and their suitability to flow cytometry remains largely unexplored. The few that have been studied with this technique in immune cells include monochloro (bromo) bimane,[29] mercury orange,[30]o-phthaldialdehyde, and chloromethyl fluorescein diacetate,[31] but none of them is selective for GSH.[32]

Many probes for biothiol sensing are based on Michael acceptors, in which following a nucleophilic attack of the sulfhydryl group and its addition to a double bond of the probe, their fluorescence increases notably. Luminescent sensors based on lanthanide complexes present several advantages over classical organic fluorophores, such as a very high luminescence lifetime and narrow emission bands, which allow an increase in the sensitivity and signal-to-noise ratio, avoiding natural background fluorescence in time-resolved luminescence spectroscopy.[33−35] Among the few lanthanide-based biothiol sensors reported in the literature,[36,37] to our knowledge, no lanthanide antenna-based sensors, which self-assemble in water have yet been reported.

In this field, we have recently reported the discovery of the small and simple structure lanthanide antenna in organic solvents 1a (Scheme ).[38] The lanthanide sensitization by 1a is quenched by H2O addition, setting the basis for its demonstrated application as a H2O sensor. In our screening for potential lanthanide antennas, we observed that the free acid 1b(39) (Scheme ) was also able to sensitize the emission of Tb3+ and more extensively Eu3+ in H2O. For the design of a suitable biothiol sensor from 1b, we synthesized the oxidized analogue 2-oxo-1,2-dihydrocyclopenta[de]quinoline-3-carboxylic acid, 2b, which showed high Michael acceptor reactivity against thiols and an excellent fluorogenic behavior upon reaction. Herein, we report the design and synthesis of this free acid, the photophysical properties and lanthanide sensitization of 1b and 2b, the reactivity of 2b, and proof-of-concept studies of the application of this biothiol sensor to study the cells of the immune system. Strikingly, the results described herein demonstrate that 2b is a selective GSH sensor, which after its reaction with this biothiol, self-assembles in water with a lanthanide cation and, as an antenna, transfers its energy to the lanthanide ion resulting in the long-lived luminescence emission of the lanthanide.

Scheme 1

Synthesis of 8-Methoxy-2-oxo-1,2-dihydrocyclopenta[de]quinoline-3-carboxylic Acid (2b)

Results and Discussion

Synthesis of 2b

As shown in Scheme , the free acid 2b was obtained in good yield from methyl ester 1a by oxidation to 2a, with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) in refluxing toluene, followed by saponification to the corresponding free acid 2b by heating with 2 N NaOH.

Photophysical Properties of 1b and 2b

The photophysical properties of 1b and 2b in CH3CN and H2O are shown in Table . The UV/visible spectrum of 1b showed an absorption maximum at 320 nm, with a small shoulder at around 375 nm, while the oxidized analogue 2b showed a broad absorption band centered around 390 nm, with no significant influence of the solvent polarity in both cases. As expected, due to the extension of the conjugation in the chromophore moiety, the absorption maximum of 2b was shifted approximately 70 nm toward longer wavelengths when compared to that of 1b. In fact, solutions of 1b were colorless, while those of 2b were orange as observed by the naked eye. Regarding emission properties, interestingly, oxidized compound 2b showed almost negligible emission (ΦF of 2b was almost 20 times lower than that of 1b).

Table 1

Photophysical Properties of Free Carboxylic Acids 1b and 2b

compda	solvent	λmaxabs (nm)	ε (M–1 cm–1)	λmaxem (nm)	ΦFb
1b	CH3CN	320, 375	4720	450	0.09
	H2O	320, 373	5451	472	0.11
2b	CH3CN	392	2125	462	0.008
	H2O	390	3050	471	0.006

Measured in duplicate at 12 μM concentration.

Quantum yields calculated with reference to quinine sulfate (in 0.1 M H2SO4).

This low fluorescence emission of 2b could be due to an antiaromatic character of its 2-oxo-1,2-dihydrocyclopenta[de]quinoline [4n]π-electron system, according to Hückel’s rules.[40−42] To clarify this hypothesis, TD-DFT calculations were carried out with the B3LYP functional and the 6-31+G(d,p) basis set,[43−46] within the Gaussian-16 package[47] to determine the minimum energy structures of the free carboxylic acids 1b and 2b in their singlet ground energy state (S0) and in the excited states S1 and T1, and their respective harmonic oscillator model of aromaticity (HOMA) values[48−52] for the common fused ring of 2-oxo-quinoline. Four tautomeric/rotamer structures were considered in the study of the geometries of the S0, S1, and T1 energy states of 1b and 2b, one with a keto group at position 2 (1A and 2A in Figure S1 of the Supporting Information) and the other three with an enol group at that position (1B–1D and 2B–2D in Figure S1). Calculations (Table S1) showed that keto tautomer A is the minimum energy form for both 1b and 2b in the ground state S0, and in the T1 state of 2b, while the enol tautomer D was that of minimum energy in the excited states S1 and T1 of 1b and in the S1 of 2b, although in the latter case its energy was very near to that of keto tautomer A (1.6 kJ mol–1). These results indicate that excitation induces tautomerization in 1b and 2b. Calculations of the HOMA index values for the 2-oxo-quinoline-fused ring common in 1b and 2b showed lower values for 2b than for 1a in the three energy states S0, S1, and T1 (Table S2) and therefore, a lower aromatic character in 2b. Interestingly, a small decrease in the aromatic character of the six-membered rings of acenaphthylene compared to naphthalene has been reported.[53] The calculated HOMA values for the peripheral tricyclic skeleton of the 1,2-dihydrocyclopenta[de]quinoline system of 2b (Table S3) indicated an aromatic character for the three energy states.

On the other hand, when comparing NMR data of the oxidized compound 2b with those of 1b (Table S4), the most significant changes with respect to aromaticity were a 0.21 ppm displacement of 7-H toward a higher field in 2b with respect to that of 1b, and the displacements of carbons C3a (29.8 pm), C8a (8.3 pm), and C8b (11.8 pm) also toward a higher field in 2b with respect to those of 1b. These data are indicative of a decrease in the deshielding of the aromatic ring current in 2b with respect to 1b and, therefore, a lower aromatic character, which could explain their photophysical behavior.

The ability of compounds 1b and 2b to directly bind lanthanide ions in the H2O solution (54 μM) and sensitize their emission was spectroscopically analyzed by the addition of 1 and 2 equivalents of TbCl3, EuCl3, DyCl3, and SmCl3. As shown in Figure S2, the free carboxylic acid 1b sensitized the luminescence of the cations Tb3+ and Eu3+ but preferably that of the 5D4 → 7F6 (490 nm) and 5D4 → 7F5 (540–550 nm) Tb3+ bands. However, under the same conditions, the oxidized free carboxylic acid 2b only sensitized the luminescence of Eu3+ but with a much lower intensity than 1b (Figure S3).

2b Behaves as a Selective and Sensitive Biothiol Sensor

The photophysical properties according to the structure of 2b; we hypothesized that this molecule could be a good Michael acceptor in particular against thiols. This hypothesis was confirmed by following the reaction of 2b (5 μM) with Cys (500 μM) in HEPES buffer pH 7.4 by HPLC–MS. As shown in Figure S4, when Cys was added just at the time of injection, the product of the addition of Cys to the double bond of 2b was rapidly detected (t = 4.25 min).

Considering the good reactivity of 2b against Cys and the photophysical properties of 2b and 1b, we propose that 2b could be employed as a fluorogenic biothiol sensor (Figure ), and consequently we studied its time-dependent reactivity toward GSH, Hcy, Cys, and H2S in HEPES buffer (50 mM, pH 7.4) using luminescence spectroscopy. The addition of 100 equivalents of GSH to 2b (5 μM) resulted in a notable fluorescence increase at 445 nm (λex = 320 nm) with a reaction time from 0 to 3 h (Figure A). By contrast, upon the treatment of 2b (5 μM) with 100 equivalents of Cys, Hcy, or H2S, this fluorescent increase was significantly lower (Figure B), which highlighted the selectivity of our sensor for GSH. This selectivity was confirmed upon the addition of 5 or 10 equiv of GSH, Hcy, Cys, and H2S, as only GSH led to a fluorescence increase (Figure C,D). Furthermore, no obvious changes were detected when other amino acids, such as Ala, or potential interferent species (H2O2 and Fe2+) were added (Figure S5), further emphasizing its selectivity toward thiols.

Figure 1

Schematic representation of biosensor 2b. After the addition of GSH to Michael acceptor 2b, the resulting antenna will increase its fluorescence. Moreover, if lanthanide ions are present, the antenna will self-assemble and intramolecularly transfer its energy (ET) to the metal, resulting in a significant increase in the red long-lived luminescence emission of Eu3+.

Figure 2

(A) Time-dependent fluorescence emission spectra of 2b (5 μM, λex = 320 nm) after the addition of 100 equiv of GSH. (B–D) Changes in the fluorescence emission intensity of 2b (5 μM) at 445 nm (λex = 320 nm) over time, after the addition of (B) 100, (C) 10, and (D) 5 equiv of GSH, Hcy, Cys, and H2S.

Given that 1b acted as a suitable lanthanide antenna, an alternative strategy in the design of the biothiol sensor would entail the addition of lanthanide ions to the reaction product of the biothiol to 2b (2b-SR) (Figure ). To explore this strategy, we next studied the ability of product 2b-GSH to directly bind lanthanide ions in the solution sensitizing their luminescence, thus resulting in a red-shifted fluorogenic sensing reaction and with extraordinary potential to apply time-gated luminescence analysis due to the long luminescence lifetime of lanthanide ions.[35] We carried out the reaction of 2b (5 μM) with 100 equivalents of GSH for 3 h (in HEPES 50 mM, pH 7.4), and then a titration of the corresponding addition product with increasing concentrations of TbCl3, EuCl3, DyCl3, and SmCl3 was performed (Figures A and S6). This led to the appearance of significant bands of the sensitized luminescence of the lanthanide cation, mainly the 5D0 → 7F2 Eu3+ band at 615 nm and, in a lower extent, the 5D4 → 7F6 (490 nm) and 5D4 → 7F5 (540–550 nm) Tb3+ bands (Figure B). The luminescence lifetimes (τ) of the Eu3+ and Tb3+ emissions for their complexes with 2b-GSH were 122 ± 5 and 350 ± 1 μs, respectively, indicating great potential to use time-resolved and time-gated analyses in the detection of biothiols. On the other hand, we also prepared the addition products of Hcy, Cys, or H2S to 2b (2b-Hcy, 2b-Cys, and 2b-H2S) and performed a titration with increasing concentrations of EuCl3. Compared to 2b-GSH, which showed a significant energy transfer to the metal (Figure C), Eu3+ titration curves of 2b-Hcy, 2b-Cys, and 2b-H2S led to a modest or negligible luminescent increase (Figures D and S7). Consequently, the τ of the Eu3+ emission for the complex of 2b-GSH was higher than the ones of the complexes 2b-Hcy and 2b-Cys (Figure S8). The experimental data of the titrations fitted adequately to a binding isotherm with a variable Hill slope (see the Supporting Information for details). The fittings provided values for apparent microscopic dissociation constants of 0.213 ± 0.005, 0.235 ± 0.013, and 2.493 ± 0.092 mM, obtained for 2b-GSH, 2b-Hcy, and 2b-Cys, respectively (Figure D). This confirmed the preference of Eu3+ to directly assemble 2b-GSH or 2b-Hcy and with much less affinity to 2b-Cys. However, the higher Eu3+ luminescence intensity and lifetime exhibited by 2b-GSH indicate more effective protection against quenching caused by water molecules in the complex with 2b-GSH than with 2b-Hcy.[54,55] This protection of the lanthanide ion in 2b-GSH is probably favored by the carboxylate group of the Glu residue present in GSH, which could aid in the formation of an extended coordination cage with the ion.[56−58] To demonstrate this, the geometry of a proposed structure of the europium complex with two units of 2b-GSH has been optimized with the RM1 semiempirical method (Figure ).

Figure 3

(A) Tb3+, Eu3+, Sm3+, and Dy3+ luminescence at their maximum emission wavelengths in the presence of 2b-GSH (5 μM of 2b and 100 equiv of GSH, λex = 320 nm) as a function of the added EuCl3, TbCl3, DyCl3, and SmCl3 molar concentration (0.01–5.5 mM). (B) Emission spectra of 2b-GSH (5 μM, λex = 320 nm) after the addition of 100 equiv of EuCl3, TbCl3, DyCl3, and SmCl3. (C) Titration spectra of 2b-GSH (5 μM, λex = 320 nm) with increasing molar concentration of EuCl3 (0.025–3.0 mM, increase indicated by the arrow). (D) Eu3+ luminescence in the presence of reaction products 2b-GSH, 2b-Hcy, or 2b-Cys (5 μM of 2b and 100 equiv of biothiol, λex = 320 nm) at 615 nm as a function of the added EuCl3 molar concentration (0.025–6.0 mM). Lines represent the fittings to a binding isotherm with a variable Hill slope equation model.

Figure 4

Proposed structure of the coordination of GSH-2b with Eu3+. The geometry of the europium complex has been optimized with the RM1 semiempirical method,[59,60] as implemented in MOPAC2016.[61]

The greater reactivity of 2b with GSH and the greater sensitization of the Eu3+ luminescence would enhance the selectivity of the sensor for GSH in subsequent cell studies, apart from the much higher intracellular concentration of GSH (1–10 mM) than that of Cys or Hcy (30–200 μM).[62−64] We thus focused our attention on the reactivity of our sensor 2b against GSH (2b-GSH) and the selective sensitization of Eu3+ ions. We studied the reaction kinetics for the addition of 100 equivalents of GSH to 2b (5 μM, HEPES 50 mM, pH 7.4) in the presence of EuCl3 (1.5 mM). This led to a significant luminescence increase (λex = 320 nm) in both the emission of the antenna (at 480 nm) and Eu3+ (Figure ). Because Eu3+ ions are added at the beginning of the reaction, they can be pre-coordinated with an unreacted 2b probe. This may slightly change the surroundings of the lanthanide ions upon reaction, resulting in different areas of the 5D0 – 7F1 (592 nm) and 5D0 – 7F2 (616 nm) Eu3+ bands, when compared to the situation in which the lanthanide ions are added after the completion of the reaction. These results open the door to the use of 2b as a self-assembled europium sensitizer to selectively report on the levels of GSH in real time, allowing the reaction to be monitored in situ at different wavelengths, broadening the palette for multiplexing applications. Nevertheless, we recommend using the 592 nm Eu3+ emission band because its magnetic dipole nature makes it less sensitive to the environment.[65] The initial rates, obtained by analyzing the enhancement of the luminescence intensity at 592 nm, exhibited an excellent linear relationship with the initial concentration of GSH (in a logarithmic scale) (Figure S9). The linear fitting yielded a slope of 0.52 ± 0.07. This means a reaction order of 1/2 with respect to GSH, which indicates that the reaction mechanism is complex, possibly involving reversibility.[66,67]

Figure 5

(A) Time-dependent luminescence emission spectra of 2b (5 μM, λex = 320 nm) in the presence of EuCl3 (1.5 mM) after the addition of 100 equiv of GSH. (B) Corresponding luminescence intensity at the emission of the antenna (480 nm, blue symbols) and Eu3+ (592 nm, red symbols).

2b Can Be Used to Monitor Intracellular GSH Levels in Murine and Human Immune Cells

To assess the applicability of 2b to study intracellular GSH in primary cells, we focused on the immune system. We first utilized 2b to evaluate the differences in biothiol levels at the baseline within different sub-populations (CD4+ T cells, CD8+ T cells, and B cells) of mouse splenocytes and human peripheral mononuclear cells (PBMCs) using flow cytometry. Of note, the sensor was not toxic and did not affect cell viability at a wide range of concentrations (0–50 μM) (Figure S10). To maximize the dynamic range of the measurements and capture differences within immune cell compartments, we used the sensor at 25 μM. Incubating the cells at this concentration, we were able to capture differences in the intracellular biothiol levels between CD4+ T cells, CD8+ T cells, and B cells, in both murine and human cells (Figure ). Whereas we observed similar mouse intracellular biothiol levels in CD4+ and CD8+ T cells, B cells showed significantly lower levels (Figure A,B), suggesting that these cells might have lower baseline metabolic rates. Interestingly, we observed a different distribution in PBMCs, with baseline biothiol levels of CD4+ T cells and human B cells being similar, and higher levels in CD8+ T cells (Figure C). These results confirm that 2b can be used in combination with flow cytometry to capture differences in intracellular biothiol levels within primary immune cell types.

Figure 6

Study of the intracellular biothiol levels in murine and human immune cells with sensor 2b. (A) Gating flow cytometry strategy to identify different cell sub-populations (CD4+ T cells in green, CD8+ T cells in red, and B cells in blue) and representative mean fluorescence intensities in the PacBlue channel either from murine or human cells (MFI, associated with intracellular biothiol levels); (B,C) flow cytometry quantification of the intracellular biothiol levels measured using the sensor at 25 μM concentration from different sub-populations of immune cells in mouse (MFI normalized to control without 2b) (B) and human (C) (n = 3 animals/group or n = 4 human samples/group from three independent experiments, ANOVA with Tukey’s HSD t-test, *p < 0.05). (D) Time-resolved and time-gated luminescence spectra from splenocytes in the presence or absence of 2b, EuCl3 (250 μM), or both (λex = 320 nm).

Based on our previous outcomes, where 2b also acted as a europium antenna after reacting with biothiols, we applied this alternate version of the sensor to study changes in immune cell intracellular biothiol levels, in this case using a time-resolved and time-gated intensity analysis adapted to detect the long luminescence lifetime of Eu3+. We cultured splenocytes from wild-type mice and studied the time-resolved and time-gated luminescence spectra between 550 and 750 nm after adding either the 2b biosensor (25 μM), europium (EuCl3 at 250 μM), or both (Figure D). As expected, we only observed changes in the luminescence intensity when the 2b sensor was added together with europium with the detected emission bands perfectly matching those of the Eu3+ emission. This result indicates that the sensor was able to intracellularly sensitize europium luminescence, which could only happen if the sensor and the cation Eu3+ successfully entered the cells and reacted with intracellular biothiols. Once the conditions for the time-resolved and time-gated analysis on splenocytes were optimized, we studied the sensitized emission of Eu3+ in splenocytes in response to biothiol levels for 14 h. Europium luminescence reached peak levels at the beginning of the experiment, slowly decreasing with time (within hours) (Figure S11), indicating that the Eu-based version of the sensor is an option for biological questions in which an increased signal-to-noise ratio (SNR) is required.

2b Captures GSH Dynamic Changes in TREG

Regulatory T cells (TREG) are one of the main mediators of central and peripheral tolerance[68−70] and thus play a key role in autoimmune diseases, organ transplant rejection, and also anti-tumor immune responses.[68−70] GSH is vital for T-cell effector function and proliferation and for preserving TREG function,[22] making the levels of this intracellular species in T cells a particularly relevant signal to monitor.

We decided to test the ability of the 2b sensor to measure GSH intracellular changes in TREG induction cultures. To this aim, we isolated naïve splenic CD4+ T cells (defined as CD44lo CD62Lhi purity > 95%) from C57BL/6-Foxp3-YFP mice and set up TREG induction cultures by culturing them for 5 days with αCD3/αCD28 activating beads under TREG polarizing conditions (IL-2 and TGFβ). In these mice, cells express a yellow fluorescent protein (YFP) fused to Foxp3, which can be detected as naïve T cells become TREG (CD4+Foxp3+), allowing us to selectively study GSH levels in this sub-population. We monitored these levels in the culture daily by incubating with the sensor for 30 min and analyzing the cells by flow cytometry (Figure A). As it has been previously well established in these cultures, we observed an increase in the percentage of TREG in the culture as a function of time which peaked at day 5 (Figure B).[71,72] Interestingly, we observed a sharp increase in TREG GSH levels at day 1, followed by a decrease at days 2 and 3 and then increasing again up to day 5 (Figure C). This result suggests that different metabolic processes act at different times in the process of becoming TREG.

Figure 7

Study of biothiol metabolism in TREG with sensor 2b. (A) Schematic of the TREG induction protocol used to monitor intracellular biothiol levels. (B) Representative scatter plots of flow cytometry analysis and quantification (C) of the number of CD4+ Foxp3-YFP+ cells in the culture at different days (1–6); and (D) flow cytometric quantification and statistical analysis of the intracellular biothiol levels in TREG at different days (n = 4 animals/group from three independent experiments, ANOVA with Tukey’s HSD t-test, *p < 0.05, blue: comparison with t = 2 days, orange: comparison with t = 3 days).

Conclusions

In conclusion, the results herein described show that the small non-fluorescent Michael acceptor 2b, after its reaction with biothiols, becomes fluorescent and an efficient Eu3+ antenna, which self-assembles with the cation in water. This property makes 2b a highly selective GSH biosensor, which can be monitored through either the increase of the fluorescence of the antenna or the luminescence of Eu3+, opening the possibility to multiplexing applications. We have demonstrated the potential of 2b as a GSH biosensor to study murine and human cells of the immune system with flow cytometry (CD4+ T, CD8+ T, and B cells), and to monitor changes in their metabolism as naïve CD4+ T cells polarize to TREG. Together, these experiments constitute a proof-of-concept of the use of 2b to monitor biothiols in immune cells, filling the gap for GSH-metabolic studies in flow cytometry to address biological questions and pave the way to its application to study clinical samples.