Functional Photoacoustic Imaging of Gastric Acid Secretion Using pH-responsive Polyaniline Nanoprobes

Abstract

A stomach functional imaging technique based on photoacoustics presented here achieves non-invasive gastric acid secretory assessment utilizing pH-responsive polyaniline nanoprobes. A testing protocol mimicking clinical practice is established using a mouse model. After imaging, the nanoprobes are excreted outside the body without inducing systematic toxicity. Further optimization and translation of this technology can help alleviate patients’ suffering and side effects.

Graphical abstract

Gastrointestinal disorders affect approximately 70 million individuals and cause more than 236,000 deaths annually in the United States alone.^[1] Gastric acid secretion is a critical function in digestion and assimilation, and gastric acid secretory disability is involved in many gastrointestinal diseases, such as Zollinger-Ellison Syndrome, pernicious anemia, chronic gastritis, atrophic gastritis, and gastric carcinoma.^[2] Therefore, real-time monitoring of pH in the stomach can offer major benefits in both diagnosis and treatment of these diseases. Unfortunately, functional imaging of the stomach and measuring real-time gastric acid secretory ability are difficult due to the lack of a simple, accurate, and non-invasive approach.^[3] Currently, a number of direct and indirect pH measurement strategies exist, with gastrointestinal intubation (insert a plastic tube through the nose or mouth down into the stomach) and radio telemetry capsule (patients swallow a pH-sensing capsule) being the most popular for continuous direct assessment of gastric pH conditions.^[4] However, these tests suffer from complications, poor patient tolerance, and in particular inappropriateness for infant patients (e.g., hyperacidity, achlorhydria, and gastroesophageal reflux disease or GERD). Therefore, innovations in safe and non-invasive imaging technologies enabling functional imaging of the stomach are urgently needed for accurate diagnosis and effective treatment of gastrointestinal diseases.

Photoacoustic (PA) imaging has the potential to address this problem because it is a non-ionizing, non-invasive, and low-cost imaging modality, combining the rich chemistry and spectral tunability of optical contrast agents, and the spatial resolution of ultrasound detection deep within tissue.^[5] Indeed, PA-based imaging and detection have been demonstrated in various disease models, including intestinal diseases.^[6] During PA imaging, light-absorbing dye molecules, polymers, or nanoparticles (mostly absorbing in the near-infrared (NIR) region to maximize light penetration depth and heat generation) are often used to provide image contrast. For example, based on PA imaging, small-molecule organic dyes such as indocyanine green have been utilized for tumor imaging and diagnosis, whereas inorganic nanoparticles such as Au nanorods and nanoshells are commonly used in tumor and vasculature imaging. Similarly, conductive polymers have been proposed as a better alternative because of their excellent absorption property similar to Au nanostructures and their improved photostability over metallic nanostructure counterparts.^{[7, 8, 9]}

For example, polyaniline (PANI), one of the first discovered conductive polymers, has a concentrated absorption peak centered at 800 nm, ideal for photothermal-based imaging and therapy. However, this strength in spectral characteristics has not translated into wide-spread uses in biological systems because a common problem of conductive polymers is their instability at neutral pH.^[9] In an acidic environment, protonated polyaniline (doped) is highly conductive and has a strong NIR absorption peak; but in a neutral or basic environment, polyaniline is deprotonated (dedoped) and consequently loses its conductivity and NIR absorption. This problem limits its use in most biological imaging applications, but can be useful for functional imaging of the stomach, where the pH is as low as 1–2 for humans. Here, we show how this seemingly major drawback (pH-sensitivity) of conductive polymers can be exploited for gastric acid secretory assessment in combination with PA imaging. Converting this disadvantage to an advantage in imaging can be potentially extended to other pH-sensitive polymers and dye molecules, and can help solve the suffering and side effects of gastrointestinal intubation in hyperacidity, achlorhydria, and GERD patients (in particular infants and elderly).^[10]

To simultaneously achieve rapid and sensitive pH responses during gastric acid secretion, a layer of PANI was coated onto iron oxide nanoparticle as a PA contrast agent for oral administration (Fig. 1). The thin layer of PANI allows fast proton diffusion and consequently fast responses to pH fluctuation; whereas the magnetic nanoparticle (MNP) core serves as a structural template for uniform PANI coating, a sacrificial oxidant for aniline polymerization, and potentially a contrast agent for magnetic resonance imaging or magnetomotive photoacoustic imaging.^{[7, 9, 11]} To confirm formation of the core-shell nanostructure, transmission electron microscopy (TEM) and dynamic light scattering (DLS) were used to measure nanoparticle sizes. Figure 1b and S1 show a PANI shell layer of approximately 15 nm around the MNP core (36.5 ± 2.7 nm in diameter). As a result, the overall size of the core-shell nanoparticle is 50.6 ± 7.9 nm. Please note that the nanoparticles in Fig. S1 appears to be connected, but this is just an artifact caused by TEM sample preparation (particles move together as the nanoparticle solution dries on TEM grids). In solution, the particles are well dispersed, which is supported by two lines of evidence. First, if the particles were aggregated in solution, large 3-D aggregates would be seen on TEM grids instead of particles spread in two dimensions. Second, the hydrodynamic size of the particle in aqueous solution was characterized using DLS. As shown in Figure 1c, the hydrodynamic diameter of the core-shell nanoparticle is 56.0 ± 6.1 nm, proving the particle dispersity in solution. The DLS size is slightly larger than the value measured by TEM, likely due to the electrical double layer on the nanoparticle surface. It has been shown that particles of this size are sufficiently large to avoid passive diffusion through the gastric mucosal layer.^[6]

Figure 1.

The design of MNP-PANI nanoprobes for gastric acid secretion measurement. a) Schematic illustrating the doping process of PANI coated MNPs via gastric acid. NIR absorbance of orally administered de-doped MNP-PANI increases due to acid doping of the conductive polymer inside mouse stomach. The signal enhancement is detected through PA imaging. b) TEM showing successful coating of PANI nanoshells on MNPs. Scale bar, 20 nm. c) MNP-PANI nanoparticles’ hydrodynamic size of 56.0 ± 6.1 nm. d) Diagram illustrating the protocol of gastric acid secretory assessment.

To test the pH-sensitive nanoparticles and PA imaging for gastric acid secretion assessment, experiments were designed based on the current clinical practice of stomach function testing.^{[12, 13]} As shown in Figure 1d, MNP-PANI nanoparticles suspended in sodium bicarbonate solution (pH 8) are administered to mice orally. The sodium bicarbonate solution leads to a pH rise inside the stomach from acidic to basic, and this alkali challenge subsequently triggers gastric acid secretion via parietal cells.^[13] For normal stomach function, gastric acid secretion will gradually reduce gastric pH to normal (pH < 4 for human) in approximately 30 min (for human), whereas abnormal acid secretion could take significantly longer or shorter time to restore the original pH in the stomach.^[13] This entire process can be monitored by PA imaging non-invasively, because of MNP-PANI nanoparticles’ pH sensitivity (initially de-doped with no or weak NIR absorption due to sodium bicarbonate solution’s high pH and later re-doped by protonation with strong NIR absorption when gastric pH returns to the normal value). Towards the end of imaging, the PA signal decays as MNP-PANI exits the acidic stomach, and eventually exits the body in feces.

Before applying this technology to animal models, two key factors critical to the technology application must be confirmed: (i) does the thin PANI shell provide instant responses to pH changes through PA imaging; and (ii) is the nanoshell sufficiently stable against intense irradiation by pulsed laser during PA imaging and the extremely low pH in the stomach?

To probe the pH responsiveness, MNP-PANI particles were suspended in buffers with different pH values. Figure 2a shows that the particles had low NIR extinction (largely due to scattering rather than absorption that matters for PA imaging) at pH 8. However, when these de-doped nanoparticles were transferred into simulated gastric fluid (pH 1, Sigma-Aldrich), they were rapidly doped and exhibited strong, concentrated NIR extinction in less than 2s (the fastest we could handle the sample and take an absorption measurement), and the absorption did not further increase with longer incubation time, indicating instant proton doping throughout the nanometer-thick PANI layer. Quantitative measurements of PANI’s extinction at 800 nm vs pH values show a sharp increase between pH 2 and 6. At pH 4, the extinction (absorption + scattering) is 7.5-fold stronger than the signal at pH 8 (Fig. 2b). A similar effect was also observed through PA measurements. As shown in Figure 2c, the PA signal intensity (800 nm excitation) at pH 4 increases by 8.7 times compared to the value at pH 8, reflecting the difference in the absorption component of extinction. As shown in ultrasound (US) and PA images (Fig. 2d), the phantom region with MNP-PANI embedded at various pH can be clearly seen though US over the whole pH range, but only becomes clearly visible in PA images at acidic pH.

Figure 2.

Simulated gastric acid doping of MNP-PANI. a) Extinction spectra of de-doped MNP-PANI (black dotted) and doped with simulated gastric fluid (red solid). b & c) Normalized extinction at 800 nm of MNP-PANI under various pH. Its NIR absorption stays at background levels at basic and neutral pH, but increases dramatically toward low pH. The PA signal intensity at pH 4 increases 8.7 folds compared to it at pH 8. d) US and PA images of MNP-PANI embedded phantoms (hemisphere shape) under various pH conditions. e) Normalized PA signal amplitude of gastric acid doped MNP-PANI under pulsed laser irradiation with a laser fluence of 13.7 mJ cm⁻². f) Retention of MNP-PANI dialyzed in simulated gastric fluid at 37 °C for 24 hours. No appreciable loss of optical extinction was detected.

Next, we investigated particle stability against intense laser irradiation and low pH. Figure 2e and S2 show that with intense pulsed laser excitation (800 nm, 13.7 mJ cm⁻²), PANI exhibits excellent stability. No PA signal attenuation was observed even after 90 min, significantly more stable than Au nanorods and nanoshells, making long-term imaging and tracking possible.^[8] To assess whether MNP-PANI nanoparticles can withstand the harsh conditions of the stomach,^[14] they were dialyzed in simulated gastric fluid. Remarkably, no appreciable loss of optical extinction was detected (Fig. 2f), demonstrating the stability and suitability of MNP-PANI as an orally administered contrast agent.

With the MNP-PANI nanoparticles’ pH responsiveness and stability confirmed, we proceeded to apply them for functional photoacoustic imaging of gastric acid secretion in mouse models. First, we explored whether the mouse gastric pH is sufficiently low to enable PANI proton doping since the mouse stomach pH is higher than human (mouse typically 3–4, human 1–2). De-doped MNP-PANI (suspended in 150 μl D.I. water) was administered to healthy mice via gavage. A custom-built single-element scanning system was used for acquiring PA images from the scanning area (Figure 3a).

Figure 3.

MNP-PANI for PA stomach imaging and gastric acid secretory assessment. a) The scanning area of mice in PA imaging. b) PA images of the mouse before and after gavage of 0.15 ml de-doped MNP-PANI in water. Circle shows the stomach position (region of interest for data analysis). c) PA intensity enhancement calculation before and after agent administration. Signal in the stomach is enhanced by 7.1 times, but other tissues remain virtually unchanged (outside labeled circle). d) PA imaging of mice with open abdomen cavity, confirming that the PA signal enhancement is specific to the stomach. e) Gastric acid secretory assessment. PA signal intensity enhancement over time (normalized by the signal before particle administration). Background signal from skins remote from the stomach area was monitored as an internal standard (images not shown) to help offset the variation of nanoparticle concentration in mouse stomach.

As shown in Figure 3b, PA imaging revealed significant signal enhancement before and after MNP-PANI administration. Without MNP-PANI, no contrast was observed from the stomach compared to other organs. In the presence of MNP-PANI, however, a 7.1 fold signal enhancement was observed in the stomach area, whereas the signals from other organs remain unchanged (Figure 3c), confirming that mouse stomach is sufficiently acidic and is a suitable animal model. To double-check whether the increased signal specifically came from the stomach, mice were sacrificed and imaged again with an open abdominal cavity (Figure 3d). Indeed, strong PA signals were only observed in the stomach, and other organs do not show high PA signals due to the lack of efficient NIR absorbers. It is worth mentioning that PA signal fluctuation is not only affected by the absorption of PANI nanoparticles but also their concentration. This is the reason why background signals from remote tissues were collected as an internal standard to help offset the nanoparticle concentration fluctuation.

To access gastric acid secretion in vivo, MNP-PANI nanoparticles were suspended in sodium bicarbonate solution (150 μl) and administered to healthy mouse via gavage. To monitor PA signal over time, an Endra Nexus imaging system was used to produce PA images of the scanning area. As shown in the transverse slice in Figure 3e, a series of PA images were acquired at the same location before and after nanoparticle administration at various time points. Quantitative signal intensity analysis was applied to the stomach and surrounding tissue background. Compared to administering MNP-PANI alone, PA signal enhancement in the stomach area was not observed immediately within the first 30 min because neutralization of the alkali challenge takes time. Due to gastric acid re-secretion inside the healthy mouse stomach, the PA signal intensity continuously rose and reached the highest value around 90 min. Quantitatively, the PA signal enhancement is 7.2 fold, matching the value of administering MNP-PANI without sodium bicarbonate.

We did not have a mouse model with acid secretion diseases such as achlorhydria and hyperacidity, but we expect that pH restoration processes (pH value and time) in these disease models would be much different. It is also worth mentioning that it would take less time for gastric acid re-secretion in human tests.^[13] Four hours post nanoparticle administration, the gastric PA signal dropped back to the pre-administration level, indicating complete elimination of the nanoparticles. The particle residence time in vivo is likely controlled by the stomach emptying rate as well as nanoparticle interactions with the gastric mucosa.^[15]

Besides MNP-PANI’s performance in acid secretion assessment, we also evaluated its potential toxicity. First, under in vitro conditions, the toxicity of the nanoparticles was assessed in three cell lines. Dose-dependent cytotoxicity was quantified based on cell viability shows that after 24-hour exposure, MNP-PANI nanoparticles are of very low toxicity (>90% viability) in the concentration range of 0–1.5 mg/mL (Fig. 4). For in vivo toxicity, histology examination of tissues from mice sacrificed 12 hours post nanoparticle administration shows no noticeable inflammatory response or cell damage in the stomach and other major organs (Fig. 5a). Iron staining was also conducted, and virtually no nanoparticles were detected, confirming the in vivo PA imaging results that nanoparticles are eliminated (at least below PA and microscopy detection limits, Fig. 5b). Note that lack of iron staining is not due to complete dissolution of the iron oxide nanoparticle core. We have observed that even at pH 1, the core etching of MNP-PANI is a very slow process, and it takes many days to dissolve the core particle (data not shown).

Figure 4.

Dose-dependent cytotoxicity of MNP-PANI nanoparticles in HeLa, LNCaP, and BNL CL.2 cells.

Figure 5.

Histology assessment of MNP-PANI’s potential toxicity in vivo after oral administration. a) Hematoxylin and eosin stained tissue sections of mice after oral administration of MNP-PANI nanoparticles suspended in sodium bicarbonate solution in comparison with tissues from untreated mice. Gastric pits and glands appear intact without inflammation in the stomach. Histology of other organs appears to be normal, too. Scale bar: 200 μm. b) Iron staining of tissue sections. No increased iron content is detected compared to control tissue sections from untreated mice. Scale bar: 200 μm.

In summary, we have developed a pH responsive MNP-PANI nanoprobe and applied it as a PA contract agent for non-invasive functional stomach imaging of gastric acid secretion. In contrast to prior efforts to overcome PANI’s doping instability issue, we convert its drawback in pH sensitivity to a unique strength to address an important clinical problem. The structural, spectral, and chemical properties of the core-shell nanoparticles were systematically characterized, and a gastric acid secretory testing protocol simulating current clinical practice was developed for live mouse imaging. These nanometer-sized particles are sufficiently large to avoid passive diffusion through the gastrointestinal mucosa membranes, yet the PANI nanoshell is thin for fast proton diffusion and penetration. Complete nanoparticle elimination after imaging and no systematic toxicity may potentially enable this technology in humans, particularly for elderly and infants, to help reduce the suffering caused by gastrointestinal intubation. We expect this strategy to be readily extended to other pH sensitive polymers and dye molecules (converting a disadvantage into a strength) for functional stomach imaging.

Experimental Section

Chemical

Unless specified, all chemicals were purchased from Sigma-Aldrich (St. Louis, MO) and used without further purification. Iron oxide magnetic nanoparticles were purchased from Oceannanotech LLC (San Diego, CA). Ultrapure water was obtained from a Milli-Q water purifier (Billerica, MA).

Polyaniline (PANI) shell growth

Aniline (8 μL stored at 4 °C) was added to DI water (860 μL). Oxidant (NH₄)₂S₂O₈ solution (2 mM) was prepared separately in DI water (3 mL), followed by addition of HCl (1 N, 30 μL). The freshly prepared aniline and oxidant solutions were cooled to 4 °C. Iron oxide nanoparticle aqueous solution (1 mL, 1.5 nM) and SDS (40 mM, 600 μL in DI water) were mixed in 2 mL DI water (pH adjusted to 2) and incubated 10 min, followed by addition of the aniline solution (60 μL). After approximately 5 min incubation, the oxidant solution (~3 mL) was added to the solution to promote PANI polymerization. The reaction mixture was kept in dark (at room temperature) with stirring (600 rpm) for 24 hours, and the resulting nanoparticles were purified with repeated centrifugation (7,000 G × 17 min) and magnetic separation (5 h). The nanoparticles were re-dispersed in DI water, where pH was first tuned to 12 by titration with NaOH (1N), and then adjusted back to 8 with HCl (1N).

Nanoparticle characterization

A UV-vis spectrophotometer (Shimadzu Scientific Instruments) was used to measure the extinction profiles of nanoparticles. Nanoparticles in water (pH 8) were titrated with simulated gastric fluid (without enzyme), and its extinction was measured by UV-vis spectrophotometer. To test the pH-dependent photoacoustic (PA) signal of MNP-PANI, PA imaging was used to measure PA signals of a phantom containing nanoparticle-agarose gel (2%) at various pH. A 532-nm pumped wavelength tunable OPO laser (Surelite OPO Plus, Continuum, Santa Clara, CA, USA) delivered laser pulses at 800 nm to the phantom immersed in a water tank. To assess the stability of nanoparticles in simulated gastric fluid, nanoparticles were dialyzed with 10 mL simulated gastric fluid. The absorption of nanoparticle solution (pH 1) was measured at various time points for three times each.

In vivo PA imaging

All experiments were conducted with approval from SYSU’s IACUC committee. A wavelength tunable OPO laser (Vibrant 355 II HE, Opotek, Carlsbad, USA) was used to deliver laser pulses at 800 nm, and a focused ultrasound transducer (10MHz, V315-SU, Olympus IMS, Waltham, USA) was used to detect PA signals. PA images were obtained by scanning the ultrasound transducer over the region of interest during successive firings of the laser. The PA signal of the stomach was measured before and after gavage of 0.15 ml MNP-PANI solution (30 nM) alone in female BALB/c nude mice. An Endra Nexus PA imaging system (Ann Arbor, MI) was used for gastric acid secretion assessment. A total of 0.15 ml of MNP-PANI plus sodium bicarbonate solution was administered via gavage, and the PA signals were monitored at various time points. Reconstruction of the 3D PA images was performed by OsiriX imaging software (OsiriX Lite, Switzerland).

Nanoparticle toxicity evaluation

After PA imaging, mice were sacrificed, and organs were harvested. Blood and debris were removed by rinsing with a 0.9% saline solution. All organs were immersed in 4% paraformaldehyde and fixed over 24 hours. The fixed organs were processed through increasing grades of alcohol, cleared in xylene and infiltrated with paraffin. They were subsequently embedded, cut and stained with hematoxylin and eosin. Finally, the slides were scanned with a single slide scanner (Leica BX51). For iron Prussian blue staining, sections were rinsed with distilled water and then incubated for 30 min in a solution formulated with 2% aqueous potassium ferrocyanide and 2% hydrochloric acid by a ratio of 1:1, rinsed and then counterstained with 1% neutral red.