Contrast-enhanced photoacoustic imaging in the second near-infrared window using semiconducting polymer nanoparticles

Abstract. Photoacoustic imaging (PAI) is a fast growing deep-tissue imaging modality. However, light scattering and absorption in biological tissues limit imaging depth. Short near-infrared wavelengths (650 to 950 nm) are widely used for PAI. Using longer near-infrared wavelengths reduces scattering. We demonstrate deep-tissue contrast-enhanced in vivo photoacoustic imaging at a wavelength of 1064 nm. An ultranarrow bandgap semiconducting polymer poly (thienoisoindigo-alt-diketopyrrolopyrrole) (denoted as PIGD) is designed and demonstrated for imaging at 1064 nm. By embedding colloidal nanoparticles (NPs) of PIGD in chicken-breast tissue, an imaging depth of ∼5  cm is achieved. Intravenous injection of PIGD NPs in living rats showed brain vascular images with ∼2 times higher contrast compared with the brain vascular images without any contrast agent. Thus, PIGD NPs as an NIR-II contrast agent opens new opportunities for both preclinical and clinical imaging of deep tissues with enhanced contrast.


Introduction
Photoacoustic tomography (PAT) is an emerging hybrid imaging modality that has found many demanding applications in both clinical and preclinical studies. [1][2][3][4][5] PAT combines high optical contrast and scalable ultrasound resolution in a single modality. In PAT, short (nanosecond) laser pulses are absorbed by tissue chromophores and converted to transient heating, which is subsequently converted into an acoustic wave due to thermo-elastic expansion. Images are formed by reconstructing the PA signals acquired at various positions around the tissue surface. Conventional PAT systems use Nd:YAG-pumped optical parametric oscillator (OPO) laser as PA excitation source. Nd:YAG lasers typically emit light with a wavelength of 1064 nm, which is then frequency-doubled to obtain 532 nm. The OPO pumped by 532 nm can generate light wavelengths ranging from 650 to 950 nm (NIR-I window). Due to the reduced scattering in biological tissues, NIR-I window is widely used for deep-tissue PA imaging. 6,7 Several contrast agents based on metallic, inorganic, organic nanoparticles, quantum dots, carbon nanotubes, etc., were used for NIR-I window imaging. [8][9][10] Contrast agents are useful for enhancing the image contrast, for targeted molecular imaging, drug delivery, therapy, etc. [11][12][13] Recent studies on optical imaging have found that NIR-II window (1000 to 1700 nm) can provide quality imaging compared with NIR-I window. 14,15 The fundamental of Nd:YAG laser is 1064 nm, it belongs to the NIR-II window. The advantages of using 1064 nm for PA imaging include: (i) according to the ANSI safety limit, the maximum permissible exposure (MPE) for skin is 100 mJ∕cm 2 at 1064 nm, whereas it is ∼30 mJ∕cm 2 at 800 nm. 16 So, more energy can be used at 1064 nm, which may contribute to deeper imaging. (ii) 1064nm Nd:YAG ns-laser is cheaper and more compact compared with the OPO lasers (or dye laser) required for NIR-I window imaging. However, at NIR-II wavelengths the absorption from endogenous contrast agents (blood, lipid, etc.) is relatively low. Thus, there is a need for NIR-II contrast agents. The NIR-II contrast agents reported so far are copper sulfide (CuS) nanoparticles, 17 silver (Ag) nanoplates, 18 gold (Au) nanorods, 19 phosphorous phthalocyanine (P-Pc), 20 and hyperbranched Au plasmonic blackbody (AuPB). 21 In particular, semiconducting polymer nanoparticles (SPNs), composed of macromolecular chromophores with strong light absorption and robust photostability, have appeared as a new generation of PA contrast agent in the NIR-II-region. [22][23][24] These SPNs as PA contrast agents have demonstrated enhanced photostability and good biocompatibility compared with inorganic NPs and small molecular chromophores in imaging of living mice. 8,[25][26][27][28][29] The studies in photoacoustic imaging proved that NIR-II wavelengths can provide better imaging depth compared with NIR-I wavelengths. An increase in the imaging depth is due to reduced tissue scattering at a longer wavelength. Moreover, at 1064-nm wavelengths more excitation energy ∼100 mJ∕cm can be used for imaging, as a result more light can reach deeper compared with the NIR-I window that allows only ∼30 mJ∕cm 2 at 800 nm. This promising behavior motivated us to explore other semiconducting polymers with high absorption coefficient in the NIR-II region.
Herein, we present a new kind of NIR-II absorbing SPNs composed of ultranarrow bandgap (refers to the optical bandgap of the polymer) semiconducting polymers poly (thienoisoindigo-altdiketopyrrolepyrrole) (denoted as PIGD, Fig. 1) for PA imaging in the NIR-II window. Both thienoisoindigo (TIIG) and diketopyrrolopyrrole (DPP) have been employed as building blocks to construct a myriad of narrow bandgap semiconducting polymers owing to their high structural planarity. [30][31][32] Most of them have shown outstanding performances in organic electronic devices, such as polymer solar cells 33,34 and field-effect transistors. 35,36 Very recently, we combined these two building blocks to construct a new semiconducting polymer PIGD, giving average molecular weight of 24.0 kDa and an optical bandgap as narrow as 0.8 eV. 37 The lower the optical bandgap of the SNPs, the longer wavelength of the light absorption by the SNPs. As a consequence, the deeper tissue-penetration depth is expected, which benefits PA imaging. Despite the synthetic chemistry of PIGD has been reported, the processing of PIGD into colloidal nanoparticles in water, the characterization of the nanoparticles, and their application for the second near-infrared window photoacoustic imaging have not been reported. These properties inspired us to explore their application in PA imaging in the NIR-II window.
In this work, we present a Q-switched Nd:YAG 1064-nm laser-based photoacoustic tomography system for NIR-II window imaging. The design and synthesis of the PIGD NPs that has strong absorption in NIR-II window are presented. The optical and PA properties of the PIGD NPs are discussed. Then, to validate the photoacoustic imaging capabilities of PIGD NPs, we performed deep-tissue imaging experiments at 1064-nm wavelength. At last, the proof-of-concept application of PIGD NPs for NIR-II window PA imaging is demonstrated by imaging brain vasculatures in living rats.  buffer solution from the 10 × PBS stock buffer. 1 × PBS consists of NaCl (137 mM), KCl (2.7 mM), Na 2 HPO 4 (10 mM), and KH 2 PO 4 (1.8 mM). Chloroform-D (99%) was purchased from Cambridge Isotope Laboratories, Inc. All other chemicals and reagents were purchased from Aldrich or Merck and used as received unless otherwise specified. Conjugated polymer PIGD was synthesized according to our previous report.

Characterization
UV-vis-NIR absorption spectra of the samples were measured on a SHIMADZU UV-2450 spectrophotometer. TEM measurements were performed with a TEM Carl Zeiss Libra 120 Plus at an acceleration voltage of 120 kV. A 5-μL droplet of diluted samples was directly dropped onto a copper grid (300 mesh) coated with a carbon film, followed by drying at room temperature. The size distribution of resulting nanoparticles was determined by dynamic light scattering (DLS) using a BI-200SM (Brookhaven) with angle detection at 90 deg.

Preparation of Nanoparticles
Conjugated polymer PIGD (5 mg) and macromolecular surfactant Pluronic F-127 (50 mg) were co-dissolved in 0.5 mL of THF. The mixture was rapidly injected into 2.5 mL of 1 × PBS buffer under ultrasonication (Alstron ultrasonic cleaner, model: ALD-40100-38H). After evaporation of excess THF, a well-dispersed solution of PIGD NPs with a concentration of 2 mg∕mL was obtained. Specifically, a mixture of PIGD polymers and a macromolecular surfactant Pluronic F-127 (The mass ratio of PIGD and F127 is 1∶10) codissolved in 1 mL of THF was rapidly injected into 10 mL of water under sonication [ Fig. 1(b)], followed by exposure in air at room temperature overnight to remove THF. The resulting SPNs were well-dispersed in PBS buffer.

PAT System for Imaging at 1064-nm Wavelength
A home-made PAT system was used for PA measurements. The excitation source is 1064 nm Nd:YAG laser (Continuum, Surelite Ex) that can generate 5-ns pulses at 10-Hz repetition rate. The 1064-nm beam was guided to the circular scanner, and homogenized using an optical ground glass (GG imaging, the transducer was driven by a computer-controlled stepper motor (SM) to continuously move in a circular motion. The acquired A-lines were used to reconstruct the PA image of the sample using a delay-and-sum back projection algorithm. 38,39 The imaging system has a spatial resolution of ∼380 μm using 2.25-MHz UST.

Agar Gel Phantom for Deep-Tissue Imaging
To demonstrate the deep-tissue imaging capability of the system, three agar gel cylinders containing PIGD NPs with different concentrations were embedded inside agar gel of 2.5 cm in diameter. The three cylinders s1, s2, and s3 [ Fig. 3(a)] contain PIGD NPs with concentrations 1.0, 0.5, and 0.25 mg∕mL, respectively. Fresh chicken-breast tissue slices were cut and placed on the phantom to determine the maximum imaging depth. The chicken slice fully covered the phantom to avoid leakage of light to the sample. Several slices of chicken-breast tissue were sequentially placed to make the objects at different depths from the laser-illuminated tissue surface.

Imaging Rat Brain Vasculature In Vivo
For brain vascular imaging, NTac:Sprague Dawley ® SD ® healthy female rats of body weight ∼95 AE 3 gm, procured from InVivos Pte. Ltd., Singapore, were used. All the in vivo experiments were performed according to the guidelines and regulations approved by the institutional Animal Care and Use committee of Nanyang Technological University, Singapore (Animal Protocol Number ARF-SBS/NIE-A0263). Before going for brain vascular imaging, the rat was anaesthetized for short period of time to depilate the hair on the scalp and to mount it inside the scanner. This anaesthesia mixture was prepared with 2 mL of Ketamine (100 mg∕mL), 2 mL of xylazine (20 mg∕mL), and 1 mL of saline. 0.2 mL per 100 g rat weight was injected intraperitoneally. This ketamine/xylazine cocktail can make the animal anesthetized for 30 to 40 min. For brain vascular imaging, the hair on the scalp was depilated. The animal was mounted in the system as shown in Fig. 2(a). While acquiring images in post-injection time period (about 2 h), the anaesthesia was achieved by the continuous inhalation of a mixture of 1.0 L∕ min oxygen and 0.75% isoflurane. PAT imaging was performed before the injection of the contrast agent. Afterward, PIGD NPs (0.25 mL, 2 mg∕mL) was injected into the rat tail vein, then PAT images were acquired for about 2 hours. After collecting the data, the animal was euthanized by the intraperitoneal injection of pentobarbital (300 mg∕mL).

Preparation and Characterization of NIR-II Light-Absorbing Polymeric Nanoparticles
The synthesis of PIGD semiconducting polymers is described in Fig. 1(a). Direct arylation polymerization (DAP) was utilized to synthesize the target semiconducting polymers by coupling between the brominated DPP and nonbrominated TIIG monomers under our optimized reaction conditions. [40][41][42] In contrast to conventional cross-coupling techniques, such as Suzuki 43 coupling or Stille 44 coupling, which requires preactivation of C-H bonds, DAP technique enables the synthesis of semiconducting polymers in fewer steps. Moreover, the absence of flammable organometallic reagents and toxic byproducts makes DAP as a better choice to afford more synthetically scalable and biocompatible materials for biological applications. PIGD SPNs were prepared by a traditional method of nanoprecipitation. 45,46 The representative transmission electron microscopy (TEM) image [ Fig. 1(c)] shows that the PIGD SPNs appear spherical, with an average diameter of 21 AE 5 nm in the dry state, which is smaller than that (28.8 nm) in the hydrated state in water as measured with DLS [ Fig. 1(d)]. This dispersion  displayed a broad absorption band, mainly covering the range from 800 to 1200 nm [ Fig. 1(e)] with a maximum (λ abs;max ) at 960 nm. The presentation of the absorption spectrum of PIGD in toluene, in comparison with its colloidal nanoparticle in water, is to understand the aggregation of PIGD inside the core of the nanoparticles. As compared with that of PIGD in dilute toluene (λ abs;max ¼ 1100 nm), a significant blue shift of absorption band was observed. This behavior could be attributed to the formation of H-aggregates of PIGD macromolecules due to their high planarity and strong π − π stacking in the aggregating state. 47 The measured absorption coefficient of the PIGD SNPs (40 μg∕mL) at 1064 nm is ∼1.5 cm −1 . The absorption coefficients of 40 μg∕mL SNPs at 1064 nm reported are ∼1.96 cm −1 , 23 2.2 cm −1 , 22 and that for 0.5 mM copper sulfide is 1.5 cm −1 . 17

PA Signals from Blood and PIGD Samples at 1064 nm
The strong light absorption of PIGD NPs in the second NIR region [ Fig. 1(e)] suggests that the potential of these NPs as a second window contrast agent. To determine the suitable concentration of NPs for deep-tissue and in vivo imaging, the PA signals of PIGD were compared with the signal of rat blood at 1064 nm. The PA signal amplitude from 2 mg∕mL PIGD NPs is ∼10 times stronger than that from the blood. Figure 2(b) shows the PA signals from rat blood and PIGD NPs with 0.5, 1.0, and 2 mg∕mL concentrations. Figure 2(c) shows that the PA signal intensities increased linearly with the concentration of PIGD NPs.

Deep-Tissue Photoacoustic Imaging at 1064 nm
To validate the advantage of PIGD for PA imaging in the NIR II window, deep-tissue imaging was conducted using the system shown in Fig. 2(a). The solutions of PIGD NPs with three different concentrations were embedded in an agar gel phantom [ Fig. 3(a)]. The concentrations of three spots, s1, s2, and s3 are 1.0, 0.5, and 0.25 mg∕mL, respectively. The phantom was placed under chicken-breast tissues with different thicknesses [ Fig. 3(b)]. PA cross-sectional images were acquired with laser energy density ∼25 mJ∕cm 2 on the tissue surface. PA images at different depths are shown in Figs. 3(c)-3(g). The PA signals from all three spots (s1 to s3) detectable at the tissue depth up to 5 cm [ Fig. 3(g)]. The signal-to-noise ratio (SNR) of these images were calculated using the relation, SNR ðdBÞ ¼ 20 × log 10 ð V n Þ, where V is the average of PA signal amplitudes, and n is the average of background noise amplitude. Figure 3(h) shows that with decreased NPs concentration, the SNR gradually decreased. The decrease in SNR is due to the fact that the 1064 nm beam was weakened as the depth of the chicken-breast tissue increased due to the light scattering and absorption by tissue. According to ANSI safety limit, the maximum permissible exposure (MPE) for skin is 100 mJ∕cm 2 for 1064-nm laser. 16 Due to the limitation of laser power, we could use 25 mJ∕cm 2 for deep-tissue imaging. But additional improvement in imaging depth can be achieved using higher energy lasers. These deep imaging results demonstrated that PIGD NPs could act as an excellent PAT agent to enhance the PA contrast for potential use in noninvasive deep-tissue imaging.

In Vivo Brain Vascular Imaging on Rats at 1064 nm
To evaluate the PA imaging capability of PIGD NPs, in vivo imaging of rat brain was performed. PAT imaging was done noninvasively, i.e., with skin and skull intact. As shown in Fig. 4, the cerebral cortex of a rat was imaged by PAT system [ Fig. 2(a)] at 1064 nm before and after a single injection of PIGD NPs (0.25 mL per rat, 2 mg∕mL). The cross-sectional images of the rat brain before injecting NPs are shown in Fig. 4(a). Although the brain vasculature is visible, the contrast of the image is poor due to little absorption of blood at 1064 nm. The SNR of this image is ∼18 dB. After intravenous injection of PIGD NPs (0.25 mL per rat, 2 mg∕mL), the PA signals from the blood vessels were increased. Brain vascular images were collected for about 2 h after injection. From the postinjection images, we observed that compared with the vascular image based on intrinsic contrast agent [ Fig. 4(a)], the images acquired after the administration of NPs show the brain vasculature with better contrast. The NPs circulating in the blood stream enhanced the contrast between the blood vessels and the brain parenchyma, showing the ability of PIGD NPs to generate strong PA signal. Brain cortex images of the rat acquired at 40-and 70-min postinjection time are shown in Figs. 4(b) and 4(c), respectively. The SNR of these images is ∼29 and ∼37 dB, respectively. Figure 4(d) shows the PA signal in the blood vessel area and image SNR as a function of postinjection time. From the graph, it is clear that the PIGD in the blood vessels helped to enhance the PA contrast. After injecting the NPs in to the blood stream, the concentration of NPs increased within the cortex vessels and reached maximum ∼70 min as shown in Fig. 4(d).
The SNR of the image collected at 70 min is ∼2 times higher than that of image collected at 0 min (preinjection image). The increase SNR achieved after the injection of NPs indicates that PIGD NPs could act as an excellent PA contrast agent. After 70-min postinjection, the PA signal remained strong, indicating a sufficient number of NPs circulating in the blood. The NPs circulation was monitored for a relatively short period (∼120 min). But, the actual circulation time of the NPs in the vasculature could be longer. Long-circulation of NPs could be due to their suitable size of ∼25 nm and the PEGylated surfaces of the NPs. The photographs of the brain taken before and after opening the scalp are shown in Figs. 4(e) and 4(f), respectively. An open scalp anatomical photograph of the cortex vasculature was taken after removing the scalp for comparison. This in vivo brain studies proved that the distribution of long-circulating NPs in the vasculature can provide contrast-enhanced PA imaging of brain vessels of rats with intact skin and skull.

Conclusions
We have demonstrated deep-tissue and contrast-enhanced in vivo photoacoustic imaging in NIR-II window using a new type of SPNs as an exogenous contrast agent. An imaging depth of ∼5 cm was achieved at 1064 nm with an excitation laser energy density of 25 mJ∕cm 2 . The contrast of the brain vasculature was enhanced by ∼2 times, 70 min after the injection of PIGD NPs (0.25 mL, 2 mg∕mL) in living rats. The energy density used for in vivo imaging was 5 mJ∕cm 2 . The imaging depth and contrast of brain vasculature could be further enhanced by increasing the excitation energy density, e.g., up to 100 mJ∕cm 2 for 1064 nm imaging. Although the SNPs in the present work were imaged at 1064 nm, they can serve as a PA contrast agent also in the wavelength range of 750 to 1100 nm [ Fig. 1(e)]. Thus, PIGD SNPs as an efficient NIR-II PA contrast agent opens new opportunities for biomedical imaging with enhanced contrast and depth.

Disclosures
Authors have no relevant financial interests in this work and no other potential conflicts of interest to disclose.