Expanded porphyrins: functional photoacoustic imaging agents that operate in the NIR-II region

Photoacoustic imaging (PAI) relies on the use of contrast agents with high molar absorptivity in the NIR-I/NIR-II region. Expanded porphyrins, synthetic analogues of natural tetrapyrrolic pigments (e.g. heme and chlorophyll), constitute as potentially attractive platforms due to their NIR-II absorptivity and their ability to respond to stimuli. Here, we evaluate two expanded porphyrins, naphthorosarin (1) and octaphyrin (4), as stimuli responsive PA contrast agents for functional PAI. Both undergo proton-coupled electron transfer to produce species that absorb well in the NIR-II region. Octaphyrin (4) was successfully encapsulated into 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-poly(ethylene glycol) (DSPE-PEG2000) nanoparticles to afford OctaNPs. In combination with PAI, OctaNPs allowed changes in the acidic environment of the stomach to be visualized and cancerous versus healthy tissues to be discriminated.


Introduction
Photoacoustic imaging (PAI) is an imaging modality that combines the high contrast and sensitivity of optical imaging with the tissue penetration depths of ultrasound (US). 1 This "light insound out" approach relies on the light absorption of either an endogenous or exogenous chromophore, typically excited by a pulsed laser, to produce heat and generate acoustic pressure waves (thermoelastic expansion). 2,3 These acoustic signals are then detected using ultrasound transducers and reconstructed to form the photoacoustic (PA) image. One of the most important preclinical and clinical applications of PAI is mapping blood oxygenation within tissue through the excitation of the endogenous chromophore, hemoglobin, (Hb). 1,[4][5][6] Heme is an iron porphyrin (cf. Fig. 1 for generic structure) that displays distinct absorption differences between its unbound and oxygen-bound forms, allowing imaging of both oxygenated and deoxygenated tissues. 4 In recent years, a number of classic tetrapyrrolic pigments, including porphyrins, chlorins and phthalocyanines, have been demonstrated as PA-responsive systems for the detection of biologically important species (i.e. low pH, enzymes, and hydrogen peroxide (H 2 O 2 )). [7][8][9][10] However, most tetrapyrrolic PA systems are limited to the NIR region (NIR-I: 700-950 nm). 7,11 Absorption greater than 1000 nm, termed the second nearinfrared region (NIR-II: 1000-1350 nm), allows for deeper tissue penetration and reduced light scattering. 12-16 Current NIR-II PA contrast agents comprise of organic semiconducting conjugated polymers; very few are solely organic-based. 17,18 In recent years, numerous groups have focused on developing new Fig. 1 Top row: chemical structures of the porphyrin, chlorin and phthalocyanine scaffolds. Bottom row: chemical structures of manganese texaphyrin (R 1 ¼ -CH 2 CH 2 CH 2 OH, R 2 ¼ -(CH 2 CH 2 O) 3 -CH 3 ) and 26 p-electron-conjugated bis-metal (Zn and Cu) dioxohexaphyrin complexes (M ¼ Cu or Zn, R 3 ¼ C 6 F 5 ) reported by Furuta and collaborators. 22 porphyrinoid systems. 19,20 Many of these, particularly the socalled expanded porphyrins, show promise as NIR-I and NIR-II absorbers. [20][21][22][23] Early on, our group reported a penta-aza Schiff base porphyrinoid known as texaphyrin that absorbs >700 nm and forms stable 1 : 1 complexes with a large array of metal cations. Recently, a Mn(II) texaphyrin was shown to be effective as a PA contrast agent using NIR-I light (Fig. 1). 24 To broaden the application of PAI, it is useful to develop systems that are capable of absorbing in the NIR-II region. In 2020, Furuta and collaborators reported 26 p-electron-conjugated bismetal (Zn and Cu) dioxohexaphyrin complexes as potential NIR-II PA contrast agents (Fig. 1). 25 Since then, Furuta and collaborators have reported several other expanded porphyrin platforms that produce a PA signal in the NIR-II and NIR-III region (NIR-III: 1550-1870 nm). 15,[26][27][28] However, no biological studies were carried out nor were these compounds shown responsive to biological stimuli. There thus remains an unmet need for functional PAI agents that function in the NIR-II region. Here, we report the use of two expanded porphyrins, naphthorosarin (1) and octaphyrin (4), as proton-coupled electron transfer (PCET)-activated PA imaging agents for functional PAI. 29,30 As detailed below, octaphyrin (4) proved effective for the imaging of stomach pH and allowed the discrimination between cancerous tissue (HepG2) and healthy tissue in BALB/c nude mice model.

Results and discussion
The two porphyrinoids considered in this study are naphthorosarin 1 and octaphyrin 4 (Schemes 1 and 2). 29,30 Unique to both systems is their propensity to undergo PCET, this process is currently being used to describe any reaction that involves both a proton transfer (PT) and electron transfer (ET). [31][32][33] In the case of naphthorosarin 1 and octaphyrin 4, it is believed that both undergo stepwise proton transfer electron transfer (PTET) or concerted proton electron transfer (CPET) processes depending on the chosen acid and reductant. 29,30 The addition of a proton source and an oxidizable anion, e.g., HCl, to 24 p-electron antiaromatic 1 results in a quasi-stable non-aromatic triprotonated monoradical dication 25 p-electron species (2) being readily formed. This radical is characterized by an absorption wavelength at $900 nm. In the presence of stronger reductants, such as HI, conversion to the two-electron reduced 26 p-electron aromatic species (3), absorbing at ca. 1000 nm, occurs (Scheme 1). In contrast, the 32 p-electron non-aromatic octaphyrin 4 undergoes a concerted two-electron reduction in the presence of a proton source such as HCl. This yields the corresponding 34 p-electron aromatic form (6) with an absorption maximum at ca. 1200 nm (Scheme 2). In the presence of less redox active acid triuoroacetic acid (TFA), a 33 p-electron radical 5 is formed, which can then be further reduced with the addition of reductants. 30 Therefore, a critical feature to PCET is that it requires both a reductant and a proton source. To our knowledge this has not previously been explored in the context of photoacoustic imaging.
The importance of pH in human health is well appreciated. For instance, numerous studies have shown that changes in the upper gastrointestinal tract pH are implicated in pathological processes. 34,35 A high pH has been observed in the stomach of gastric ulcer (pH ¼ 3.4) and gastric cancer patients (pH ¼ 6.6) compared to healthy subjects (pH ¼ 2.9), whereas, a low pH has been seen in the stomach of esophageal ulcer (pH ¼ 1.9) and duodenal ulcer patients (pH ¼ 2.1). 35 Thus, the ability to probe non-invasively the dynamics of stomach pH could prove useful in monitoring stomach health. Cancer, broadly speaking represents another area where monitoring pH via PAI could be benecial in the context of diagnostic and therapeutic applications. 8,36 It is well-known that the extracellular tumor microenvironment is slightly acidic, pH ¼ 6.4-7.0. 37 Most cancer environments are also highly reducing. 37, 38 We thus postulated that stomach and cancer imaging would provide useful testbeds for evaluating 1 and 4 as possible PCET-based PAI agents.
The pH responsiveness of 1 and 4 were evaluated in THF solution through the careful addition of 1 M aqueous HCl. Both 1 and 4 demonstrated a strong PA response when subject to 900 and 1200 nm pulsed laser excitation (15 mJ), respectively, as the apparent 39 pH decreased (see Fig. S1-S3 †). This was considered indicative of the formation of the non-aromatic triprotonated monoradical dication 25 p-electron species (2) and 34 p-electron aromatic (6), respectively. No 26 p-electron naphthorosarin species (3) was observed, presumably reecting its less positive reduction potential ((1: +0.42 V and +0.04 V); (4: +0.56 V and +0.25 V)). 25,26 A nanoparticle (NP) encapsulation strategy was employed to allow studies of 1 and 4 in biological milieus. Both 1 and 4 could be encapsulated using 1,2-distearoyl-sn-glycero-3phosphoethanolamine-poly(ethylene glycol) (DSPE-PEG 2000 ) (compound: DSPE-PEG 2000 ¼ 1 : 750 (w/w)). Unfortunately, the absorption and PA features of NaphthNPs (from 1) proved unresponsive to pH changes (see ESI - Fig. S4 and S5 †). In the case of the OctaNPs (from 4) the mean size (106 nm) and zeta potential (À55 mV) were considered suitable for biological testing (Fig. 2D). The ratio between 4 and DSPE-PEG 2000 was found to be critical for the PCET-mediated conversion of 4 to 6 (see ESI - Fig. S6 †). No changes in acidic media were observed for OctaNPs at ratios 1 : 10 and 1 : 100 (4:DSPE-PEG 2000 ). On the other hand a ratio of 1 : 750 (4:DSPE-PEG 2000 ) was found to provide an optimal formulation for the formation of 6 in acidic media. A similarly responsive formulation was not found in the case of the NaphthNPs. While studies are ongoing in an effort to determine an optimal nanoparticle strategy for 1, this failure is ascribed to the less positive reduction potential of this particular expanded porphyrin. No changes in color were observed for OctaNPs in aqueous solution over the course of 7 days. In contrast, 4 in THF was observed to gradually became darker in color, which we ascribe to partial degradation (Fig. 2G). On this basis, we suggest that the use of this nanoparticle strategy improves the solution phase stability of 4. Moreover, OctaNPs displayed a pH dependent increase in the long-wavelength absorption band expected for a PCET process wherein the components of the medium (e.g., chloride anions) serve as the reductant (Fig. 2E and F). The importance of reductants to this PCET-response was further conrmed using the less-redox active acid triuoroacetic acid (TFA) to adjust the pH of the aqueous test solutions. At pH 5, a minimal increase in absorption at $1200 nm was observed for OctaNPs. In contrast, co-treatment with both TFA and the biological reductant glutathione (GSH, 40 mM) led to a signicant increase in this absorption feature (Fig. S7 †). We thus focused our attention on OctaNPs for the remainder of this study.
Next, the ability of OctaNPs to detect changes in pH was evaluated using PAI. The maximum PA intensity was seen at pH ¼ 3 with the response being essentially immediate on the laboratory time scale. In contrast, a gradual turn-on response was observed over the 5-6 pH range that peaked aer 6 min (Fig. 3). The lower PA signals observed at pH 1 and pH 2 are ascribed to partial degradation of 4. No adverse toxicities were seen for OctaNPs, as inferred from histological analyses and evaluation of hemolysis rates ( Fig. 3B and C).
OctaNPs were then tested as PAI agents in vivo. As shown in Fig. 4, direct injection (intragastric) of OctaNPs (200 mL, 0.1 mg mL À1 ) into the stomach resulted in an immediate PA signal being observed (Fig. 4C and E). This is due to the presence of HCl within the stomach and the stomach having a known pH range of 1.5-3.5. These environment are thus conducive to PCET reduction of 4. 34 To test whether OctaNPs could be used not just to visualize the stomach, but to monitor directly dynamic changes in the stomach pH, mice were pretreated through the injection of a saturated NaHCO 3 solution (50 mL) to raise the pH (Fig. 4D and F). This was followed by the direct injection of OctaNPs (200 mL, 0.1 mg mL À1 , neutral aqueous) into the stomach. No PA signal was initially observed under 1200 nm excitation (15 mJ), a nding ascribed to the bicarbonate-induced neutralization of the gastric acid. As time  progressed the PA signal at 1200 nm increased reecting the presumed secretion of gastric acid by the mice.
Efforts were then made to explore OctaNPs as PCET-triggered PA agents for cancer imaging. As noted above, the extracellular tumor microenvironment is slightly acidic (pH 6.4-7.0). 37 As shown in Fig. 3A, OctaNPs produce a minimal PA "turn on" response at 1200 nm at pH 6. However, we considered it likely that the reducing nature of tumor tissues 38,40 would facilitate PCET and afford a signicant PA response at 1200 nm. To test this hypothesis, increasing concentrations of the endogenous reducing agent GSH were added to a pH 6 solution containing OctaNPs. This resulted in a marked increase in the PA signal intensity (Fig. 5A and B). In contrast, no changes in absorptivity were observed when GSH was added to a neutral pH 7.4 solution of OctaNPs (see ESI - Fig. S8 †). OctaNPs were then injected into the right ank of BALB/c nude mice as well as into a HepG2 mice tumor model. A statistically signicant PA signal at 1200 nm was observed for the tumor region into which the OctaNPs (200 mL, 0.1 mg mL À1 ) had been injected. In contrast, minimal PA signals were observed for healthy tissues. Overall, 2 minutes post-injection, a 42 (AE7)-fold difference in PA signal intensity between cancerous and healthy tissue was observed ( Fig. 5C and D).

Conclusion
The results presented here lend support to the suggestion that expanded porphyrins could prove useful as PAI agents. Systems such as 4, that are known to undergo PCET, provide for environmental responsive PA imaging, while allowing access to the NIR-II spectral region. Encapsulation of 4 in DSPE-PEG afforded biocompatible nanoparticles (OctaNPs) that were shown to be stable for over 7 days. OctaNPs enabled the visualization of acidic environments such as in the stomach, along with changes in the stomach pH. OctaNPs also proved effective at discriminating between cancerous and healthy tissues with a 42-fold difference in the PA intensity being observed. Overall, this work serves to highlight the role that expanded porphyrins may have to play in functional photoacoustic imaging in the NIR-II region.

Ethical statement
All animal experiments and procedures were performed in compliance with the requirements of the National Act on the Use of Experimental Animals (People's Republic of China) and were approved by the Experimental Animal Ethical Committee of Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences. The accreditation number is SIAT-IRB-180205-YYS-CJQ-A0413.

Data availability
All relevant data supporting the key ndings of this study are available within the article and its ESI or from the corresponding author upon request.

Author contributions
J. Chen and A. C. Sedgwick conceived the project and designed the experiments. J. Chen performed the nanoparticles preparation, characterization and biological experiment. A. C. Sedgwick and S. Sen synthesized compounds. A. C. Sedgwick and J.