Fluorescence technique, as a nonionizing radiation method, is one of the most powerful cellular biology tools to study biomedical phenomena, by which imaging and tracing of various biological species and dynamic biological events within living cells or body could be achieved [1-5]. Over the past 30 years, fluorescence techniques have been revolutionarily advanced, typically including confocal laser scanning microscopy (CLSM), two-photon laser scanning fluorescence microscopy (TPLSM), single molecule localization microscopy (SMLM), structure illumination microscopy (SIM), and imaging-guided surgery (IGS) system [6-9]. Meanwhile, these advances also pose a great challenge on developing more robust fluorescent dyes to match the increasing requirements, especially those with absorption and emission wavelengths in the near-infrared (NIR) range (700–900 nm), a favorable spectral window due to the reduced background signal, enhanced tissue penetration, minimum photodamage to biological samples, and invisibility to eyes that ensures the look of surgical field undisturbed [10,11].

Among various fluorescent dyes, cyanine dyes, since their first synthesis in 1856, have become the extremely valuable fluorophores in chemistry, biology, and medical diagnosis [12]. Within this family, heptamethine indocyanine dyes (Cy7), containing seven conjugated carbon atoms in the polyene linker, are particularly attractive due to their NIR absorption and emission wavelengths (commonly > 750 nm) and exceptionally large molar extinction coefficients (commonly > 2.0 × 10⁵ L mol⁻¹ cm⁻¹). The most typical heptamethine indocyanine is indocyanine green (ICG) (Scheme 1A), the only Food and Drug Administration (FDA)-approved NIR optical marker for clinical uses [13]. However, ICG is known for its poor stability and moderate fluorescent properties; moreover, strong self-aggregation due to its hydrophobic polyene core, heavily binding to serum proteins arising from its unbalanced surface charges, and obvious background fluorescence due to the high uptake in liver and gastrointestinal tract also compromise its imaging application [14,15]. More disadvantageous for ICG is the absence of a single reactive site in its molecular skeleton, which greatly limits its derivatization for improving its targetability for various diagnosis applications. Fortunately, the dilemma was broken since the development of the conjugatable meso–chloro–substituted heptamethine indocyanines Cl-Cy7 (Scheme 1B) [16,17], whose post-synthetic modification via the reactive meso C-Cl bond promptly leaded to a variety of meso–position-functionalized heptamethine indocyanines, either as fluorescent probes to image biological species [11], or as NIR photocaging groups to prompt the spatial and temporal control of drug delivery [18,19], or as phototherapeutic agents to ablate tumors [20-23].

From pre-clinical or clinical perspective, a Cl-Cy7-derived representive example is polyanionic CW800 (IRDye CW800^®) (Scheme 1C), a commercially available meso C–O aryl heptamethine indocyanine that has been developed into NIR fluorescent probes for real-time visualization of operating field, allowing accurate identification and resection of tumors [24,25]. However, the polyanionic structure (four -SO₃⁻ groups) makes the dye or its protein conjugates to suffer from the non-optimal pharmacokinetics and obvious background fluorescence due to the non-specific staining with off-target proteins by electrostatic interaction [26-28]. To overcome the limit, the charge-balanced zwitterionic structures, such as ZW800–1 (Scheme 1D), were subsequently exploited, which largely decreases non-specific staining and efficiently promotes renal clearance, thus substantially increasing the tumor-to-background ratio [29-31]. However, both CW800 and ZW800–1 contain a chemically unstable meso C-O aryl bond, which is easy to suffer from the attack of intracellularly rich thiols or amines to cause chemical degradation [32,33]. Soon afterwards, the issue was intelligently addressed by replacing the reactive meso C-O aryl bond with the more robust meso C-O alkyl bond via the electrophile-integrating Smiles rearrangement, as exemplified by UL766 (Scheme 1E) [34,35]. However, due to the strong electron-donating ability of the meso C–O alkyl group, the electron-rich heptamethine indocyanines are susceptible to photobleaching caused by the simultaneously generated singlet oxygen [36]. Recognizing these shortcomings, the meso–aryl heptamethine indocyanines, featured with a more robust meso C–C bond and a less electron-donating meso–aryl group, subsequently aroused researchers’ keen interest, as exemplified by ZWCC (Scheme 1F) [37-39]. Although improved significantly in chemo- and photo-stability, the series of dyes exhibited extensive self-aggregation in water, especially the nonfluorescent H-aggregation, due to their rigid hydrophobic cores [40,41], thus largely retarding their practical imaging application. As is known, dye self-aggregation is detrimental to protein labeling because it drives multiple self-aggregated dyes to covalently bind at proximal lysine positions on protein surface, partially quenching the fluorescence, influencing protein functions, and also promoting undesired protein aggregation [40-43]. Therefore, how to efficiently inhibit the self-aggregation of meso–aryl heptamethine indocyanines in water subsequently becomes a pressing issue. Recently, two sterically shielded meso–aryl heptamethine indocyanines s775z and FNIR-Tag-766 (Schemes 1G and H) [44,45], along with their bioconjugates, were intelligently synthesized, both of which displayed strong anti-aggregation ability in water, indicating that the sterically shielding arm at the ortho positions of the meso–aryl substituent play a critical role in preventing self-aggregation. However, the synthesis of the two dyes is multi-step and technically demanding, which may hinder their uses in preclinical research. Thus, developing a simpler method to access the sterically shielded meso–aryl heptamethine indocyanines is urgently needed.

Generally, the meso–aryl heptamethine indocyanines could be prepared by way of the Pd-catalyzed Suzuki-Miyaura cross-coupling of Cl-Cy7 and arylboronic acids or arylborate esters in aqueous solution [37-39,41,46-48], which was first reported in 2006 [39]. However, the method only works well in aqueous solution, thus being greatly confined to prepare those highly hydrophilic and cell membrane-impermeable ones; moreover, the method is hardly used to fabricate the sterically shielded meso–aryl heptamethine indocyanines presumably due to the Pd-catalyzed cross-coupling incompatible with the arylboronic acid or arylborate ester with the large steric groups at its ortho-positons. The other method is the direct condensation of a custom Schiff base with indolenine [49,50]. However, the method is only used to prepare meso–phenyl and meso–methyl heptamethine indocyanines to date, probably because of the difficult synthesis of the related Schiff base with the bulkyl Aryl substituent in its meso–position. Recently, by employing the ring opening reaction of Zincke salts [44,51] or PyBox [45] salts, two sterically shielded and anti-aggregation meso–aryl heptamethine indocyanines, i.e., s775z and FNIR-Tag-766 (Schemes 1G and H), were synthesized. However, as mentioned above, the two multi-step synthetic methods are technically demanding.

In this work, for the first time we put forward a simple and universal "cyanine ketone method" for fabricating the meso–aryl heptamethine indocyanine NIR fluorescent dyes (Scheme 2A). As seen, the treatment of the easy-to-get cyanine ketones with various aromatic lithium reagents (ArLi), followed by acidification, could directly give rise to meso–aryl heptamethine indocyanines in one-pot way. Importantly, thanks to the strong nucleophilicity of ArLi reagents, various bulky hydrophilic aromatic groups could easily be integrated into the meso–position of heptamethine indocyanine skeletons, enabling the facile one-pot accessing of the sterically shielded and anti-aggregation meso–aryl heptamethine indocyanines. Using one of the anti-aggregation meso–aryl heptamethine indocyanines, we have successfully fabricated a PEGylated protein labeling agent Tag-776-NHS to label monoclonal antibody panitumumab for in vivo imaging tumor (Scheme 2B), and obtained a high tumor-to-normal (T/N) tissue ratio in a tumor-bearing mouse model.

Our design was inspired by the synthesis of meso–aryl rhodamine fluorophores that employed xanthones as the precursors [52-55]. We envisioned that the approach should also be amenable to the synthesis of meso–aryl heptamethine indocyanines, provided that cyanine ketone precursor 1 (Fig. 1) could be easily obtained. In fact, the compound was first synthesized in 1992 by the one-step reaction of Cl-Cy7 with sodium acetate in N,N-dimethylformamide (DMF), showing absorption and emission maxima at 535 nm and 625 nm, respectively [56]. Upon protonation, it could be transformed into meso–hydroxyl cyanine with absorption and emission maxima at 710 nm and 730 nm, respectively [57-61]. The early work is the start point of our subsequent studies. According to the reported procedure [56], we first synthesized the simplest cyanine ketone 1a, and then tested its reaction with phenyl lithium 2a in THF at −78 ℃. To our delight, the reaction, followed by acidification with aqueous 1 mol/L HCl, provided the meso–phenyl heptamethine indocyanine 3a in 82.4% yield (Fig. 1). This result reveals that our cyanine ketone method should be feasible for fabricating various meso–aryl heptamethine indocyanines.

Encouraged by the above result, we subsequently tested the generality of our "cyanine ketone method". As expected, by the reaction of cyanine ketone 1a with various aryl lithium reagents 2, followed by acidification, a series of meso–aryl heptamethine indocyanines 3 with varied steric bulk in the meso–position could conveniently be synthesized, including meso–2-methylphenyl 3b, meso–2,6-dimethylphenyl 3c, meso–pyridin-4-yl 3d, meso–quinolin-4-yl 3e, meso–anthran-9-yl 3f, meso–10-methoxyanthran-9-yl 3g, and meso–4-(phenoxazin-10-yl)−2-methlyphenyl 3h (Fig. 1). The chemical structures of the above compounds were confirmed by ¹H-nuclear magnetic resonance spectroscopy (NMR), ¹³C NMR, and high-resolution mass spectrometry (HRMS) spectra (Supporting information). Notably, a main objective using the "cyanine ketone method" is to construct the water-soluble and sterically shielded meso–aryl heptamethine indocyanines to inhibit the dye self-aggregation in water for effective protein labeling. Actually, previous studies on s775z and FNIR-Tag-766 (Schemes 1G and H) have indicated that the water-soluble and bulky substituents in the ortho positions of meso–aryl group play critical roles in improving water-solubility and preventing dye self-aggregation [44,45]. With the idea in mind, we further synthesized meso–2-methoxyphenyl 3i, meso–2,6-dimethoxyphenyl 3j, meso–2,6-diethoxyphenyl 3k, and meso–2,6-diphenoxyphenyl 3l through our "cyanine ketone method" (Fig. 1), and demonstrated that, among the series, meso–2,6-dimethoxyphenyl 3j has the best water-solubility (octanol/water partition coefficient: LogP = 1.005) and the strongest anti-aggregation ability in phosphate buffer solution (PBS, 10 mmol/L, pH 7.4), as indicated by the exclusive monomer absorption band even in the high concentration of 40 µmol/L (Fig. S1 in Supporting information). By comparison, in the same condition, ICG and meso–2-methoxyphenyl 3i showed obvious H-aggregation band, and meso–2,6-diethoxyphenyl 3k and meso–2,6-diphenoxyphenyl 3l had low and negligible water-solubility, respectively (Fig. S1), due to the more hydrophobic meso–2,6-diethoxyphenyl and meso–2,6-diphenoxyphenyl groups. To further improve the water solubility, we synthesized a series of PEGylated meso–aryl heptamethine indocyanines, including meso–2,6-di(ethylene glycol monomethyl ether)-substituted phenyl 3m, meso–2,6-di(dipolyethylene glycol monomethyl ether)-substituted phenyl 3n, and meso–2,6-di(tripolyethylene glycol monomethyl ether)-substituted phenyl 3o, by our "cyanine ketone method" (Fig. 1), and demonstrated that the introduction of two ethylene glycol monomethyl ether or two polyethylene glycol (PEG) monomethyl ether chains in the 2,6-positions of the meso–aryl group greatly improved the water-solubility of these dyes, as indicated by the gradually decreased LogP values (LogP: 0.849 for 3m, 0.431 for 3n, and 0.403 for 3o) when compared with that of meso–2,6-dimethoxyphenyl 3j (LogP = 1.005); meanwhile, all of them exhibited the strong anti-aggregation ability in PBS due to the presence of two bulky sterically shielded arms in their meso–aryl group (Fig. S1). The story is not over yet, because it is still possible to replace the hydrophobic indolenine N-ethyl group with the water-soluble PEG₃ chain to further improve the water-solubility. Moving on, we synthesized a PEGylated cyanine ketone precusor 1b, by which we obtained a fully PEGylated dye 3p by our "cyanine ketone method" (Fig. 1). As expected, the molecule, featured with four PEG₃ chains in its molecular structure, showed the highest water-solubility (LogP = 0.092) among the series; moreover, no any self-aggregation was observed in PBS in the concentration range of 2–40 µmol/L (Fig. S1). The chemical structures of these dyes were also confirmed by ¹H NMR, ¹³C NMR, and HRMS spectra (Supporting information).

The photophysical properties of 3a–3p were evaluated in dichloromethane (CH₂Cl₂), acetonitrile (MeCN), and PBS (10 mmol/L, pH 7.4, containing 30% MeCN for ensuring all the dyes are soluble), respectively, and compared with a conventional heptamethine indocyanine Cy7 (Table S1 in Supporting information). The absorption and emission spectra were shown in Figs. S2–S4 (Supporting information). As seen, these meso–aryl heptamethine indocyanines all show NIR absorption/emission wavelengths with maxima in the ranges of 757–778 nm/784–811 nm, respectively, longer than those of control dye Cy7 due to the stronger electron-withdrawing property of meso–aryl groups than that of meso–H atom. The fluorescence quantum yields of these dyes (Φ: 0.26–0.41 in CH₂Cl₂; 0.18–0.32 in CH₃CN; 0.12–0.24 in PBS) are comparable or higher than those of control dye Cy7, except for 3h (Φ: 0.05 in CH₂Cl₂; 0.07 in CH₃CN; 0.07 in PBS) whose strongly electron-donating phenoxazine group effectively quenches fluorescence via the photoinduced electron transfer (PeT). Noteworthily, these dyes all exhibit the considerably large molar extinction coefficients (ε: 2.00–3.95 × 10⁵ L mol⁻¹ cm⁻¹), higher than those of control dye Cy7 (2.00–2.48 × 10⁵ L mol⁻¹ cm⁻¹) in the same solvent. As a result, their fluorescence brightnesses (ε × Φ), except for 3h, are obviously higher than those of control dye Cy7 with the maxima up to 126,000 L mol⁻¹ cm⁻¹ in CH₂Cl₂ (for 3b) and 58,000 L mol⁻¹ cm⁻¹ in PBS (for 3a and 3p). In addition, all these meso–aryl heptamethine indocyanines show similar absorption and emission profiles to those of Cy7, manifesting that the sterically shielded arms on the two ortho-positions of the meso–aryl group do not essentially alter their ground- and excited-state conformations. Encouraged by these results, we further evaluated the chemo- and photo-stabilities of these dyes. Previous studies showed that the meso–phenoxy groups in the meso C–O aryl heptamethine indocyanines, such as CW800 and ZW800–1 (Scheme 1), can be rapidly exchanged by biothiol nucleophiles under aqueous conditions due to the chemically unstable meso C-O aryl bond [32,33]. By contrast, the meso–aryl heptamethine indocyanines 3a–3p were considerably stable toward biothiol glutathione (GSH) due to the robust meso C-O aryl bond (Fig. S5 in Supporting information). In addition, previous studies also showed that the electron-rich meso C–O aryl or meso C–O alkyl heptamethine indocyanines, such as UL766 and CW800 (Scheme 1), are vulnerable to singlet oxygen, which results in photobleaching [36]. By comparison, the meso–aryl heptamethine indocyanines 3a–3p even showed higher photostability than control dyes Cy7 and Cl-Cy7 (Fig. S6 in Supporting information), both bearing no electron-donating meso C–O aryl or meso C–O alkyl group in their molecular structures and thus being more photostable than UL766 and CW800.

Having established that our "cyanine ketone method" is very convenient for constructing various robust meso–aryl heptamethine indocyanines, especially those anti-aggregation versions, we, as a proof-of-concept, set out to develop dye-antibody conjugate for in vivo imaging tumor. Indeed, monoclonal antibodies (mAb) labeled with a fluorescence reporter have shown great potentials in diagnostic applications due to their high affinity and specificity for target antigens [62]. The amine-reactive NHS ester chemistry is the most common way to fluorescently label the exposed lysine amines on mAb surface by forming an amide linkage [63,64]. Previously, we have confirmed that the small-sized meso–2,6-dimethoxyphenyl group could effectively prevent the self-aggregation of heptamethine indocyanine, as indicated by the exclusive monomer absorption bands of 3j (2–40 µmol/L) in PBS (Fig. S1). Based on the consideration, we synthesized a meso–4–tert–butyl 2,6-dimethoxyphenyl 3q from the PEGylated cyanine ketone precusor 1b by our "cyanine ketone method" in order to install a conjugable carboxyl handle on the meso–2,6-dimethoxyphenyl group for protein conjugation, and the subsequent t–butyl deprotection by CF₃COOH provided the carboxy-functionalized congener Tag-776 (Fig. 2A). The corresponding amine-reactive NHS ester Tag-776-NHS, featured with two water-soluble (PEG)₃ chains and an anti-aggregation meso–2,6-dimethoxyphenyl group, was prepared conveniently by 1-(3-dimethylaminopropyl)−3-ethylcarbodiimide (EDC) coupling. Their chemical structures were confirmed by ¹H NMR, ¹³C NMR, and HRMS (Supporting information).

The photophysical properties of Tag-776 in CH₂Cl₂, CH₃CN, and PBS (10 mmol/L, pH 7.4) were shown in Table S2 (Supporting information). As seen, Tag-776 shows the absorption/emission maxima in the range of 762–776 nm/793–807 nm, molar extinction coefficients in the range of 2.67–3.55 × 10⁵ L mol⁻¹ cm⁻¹, fluorescence quantum yields in the range of 0.19–0.36, and fluorescence brightness in the range of 51,000–128,000 L mol⁻¹ cm⁻¹, all of which are approximately identical to those of 3j, indicating that the carboxyl modification does not impact its photophysical properties. Subsequently, we evaluated the water-solubility and self-aggregation of Tag-776 in PBS (10 mmol/L, pH 7.4). As shown in Fig. 2B, due to the presence of two (PEG)₃ chains as well as an anti-aggregation meso–2,6-dimethoxyphenyl group, the dye is highly water-soluble (Lop = −0.29) and no any self-aggregation band in PBS was observed even when its concentration reached up to 40 µmol/L. Encouraged by the result, we subsequently tested whether Tag-776 could be used to prepare dye-labeled mAb conjugate. The epidermal growth factor receptor (EGFR)-targeting mAb panitumumab (Pan) was selected for the purpose [35,45,65]. EGFR is a family of transmembrane glycoproteins, which is overexpressed in many cancer cell lines and associated with poor prognosis and high mortality [66]. To our delight, using standard protein labeling method, the corresponding NHS ester Tag-776-NHS could be successfully conjugated to the EGFR-targeting Pan in PBS with pH 8.5, giving rise to the dye-antibody conjugate Pan-Tag-776 (λ_abs/λ_abs = 768/800 nm in pure PBS) after purification with PD-10 column and size-exclusion column (SE-HPLC) in tandem (Fig. S7A in Supporting information). The high purity of PAN-Tag 776 was demonstrated through sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 2C). The degree of labeling (DOL) in the dye-antibody conjugate was determined to be ~3.0 when 10 equiv. of Tag-776-NHS was used (Fig. S7B in Supporting information). Animal experiment was carried out in accordance with the relevant laws and guidelines issued by the Ethical Committee of Shanxi University (No. SXULL2024010). Then, Pan-Tag-776 (50 µg) was intravenously injected into a BALB/c nude mouse bearing A549 tumor, and imaging was carried out at 6, 24, 32, and 48 h post-injection. As shown in Fig. 2D, Pan-Tag-776 exhibited obvious tumor uptake at the time point of 24 h, and reached up to the maximum at the time point of 32 h post-injection with a high T/N tissue ratio of 3.2, higher than the clinically acceptable threshold of 2.0. After 48 h post-injection, the fluorescence in tumor site was almost unobservable. The tumor and major organs, including heart, liver, spleen, lung, and kidney, were then removed and imaged. As shown in Fig. 2E, the residual fluorescence in liver and kidney indicates that Pan-Tag-776 underwent both hepatic and renal clearance. These results reveal that Pan-Tag-776 is a promising probe for the fluorescence-guided resection of tumors.

In summary, we herein have presented a new synthetic method, i.e., "cyanine ketone method", for fabricating the robust meso–aryl heptamethine indocyanine dyes in one-pot. Using the method, various aromatic substituents, including those hydrophilic and sterically shielded ones, could be facilely installed into the meso–position of heptamethine indocyanines. The as-prepared water-soluble and sterically shielded meso–aryl heptamethine indocyanines display strong anti-aggregation abilities in water, thus facilitating their application in labeling biomacromolecules. As a proof-of-concept, we have developed a PEGylated protein labeling agent Tag-776-NHS for labeling monoclonal antibody Panitumumab for in vivo imaging tumor and achieved a high tumor-to-normal tissue ratios of 3.2. Overall, our studies lay a chemical foundation for facilely accessing the robust meso–aryl heptamethine indocyanine NIR fluorophores and will thus facilitate various advanced imaging and therapeutic applications.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Mengxing Liu: Software, Methodology, Investigation, Data curation. Jing Liu: Writing – original draft, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization. Hongxing Zhang: Visualization, Validation, Software, Methodology, Data curation. Jianan Tao: Investigation, Data curation. Peiwen Fan: Software, Investigation. Xin Lv: Visualization, Validation. Wei Guo: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Formal analysis, Data curation, Conceptualization.

This work was supported by National Natural Science Foundation of China (Nos. 22277070, 22274091, 22007061), Youth Talent Support Program of Shanxi Province, Program for the Top Young Academic Leaders of Higher Learning Institutions of Shanxi, and Fundamental Research Program of Shanxi Province (No. 20210302123445).

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.109994.