Imaging Reactive Oxygen Species in Arthritis

Abstract

Reactive oxygen species (ROS) have been shown to play a role in the pathogenesis of arthritides. Luminol was used as the primary reporter of ROS and photons resulting from the chemiluminescence reaction were detected using a super-cooled CCD photon counting system. Luminol was injected intravenously into groups of animals with different models of arthritis. Imaging signal correlated well with the severity of arthritis in focal and pan-arthritis as determined by histological measurement of ROS by formazan. Measurements were highly reproducible, sensitive, and repeatable. In vivo chemiluminescence imaging is expected to become a useful modality to elucidate the role of ROS in the pathogenesis of arthritides and in determining therapeutic efficacy of protective therapies.

Introduction

Degradation of hyaluronic acid is one of the causes of cartilaginous tissue destruction [1,2] and free radical formation through reactive oxygen species (ROS) has been implicated in initiation and progression of joint inflammatory diseases [3,4]. ROS and other highly reactive radical species have also been implicated indirectly, inducing apoptosis of chondrocytes [5]. Collectively, ROS include a number of different types of molecular oxygen including hydrogen peroxide (H₂O₂), hypochlorous acid, singlet oxygen, and superoxide and hydroxyl radicals; however, most of the data to date implicate hydrogen peroxide as the main ROS signaling molecule [6]. Inflammatory tissues contain abundant neutrophils, monocytes, and macrophages, which generate ROS through peroxidases (“respiratory burst”) as part of the cellular defense mechanism [6]. Furthermore, antioxidant therapies with superoxide dismutases (SOD), SOD mimetics, vitamin E, and other newly developed antioxidants have been used as protective therapies [7–9]. Despite the exact mechanism of ROS generation and inhibition, there are no reliable methods of noninvasive quantitation of these biologically important molecules in vivo.

We reasoned that it should be feasible to image ROS levels directly by exploiting a chemiluminescence reaction in which each chemical reaction releases a single photon detectable by hypersensitive photon counting devices. The goal of this study was to demonstrate proof-of-principle of such chemiluminescence imaging. One widely used ROS in vitro assay uses luminol (3-aminophthalic hydrazide), a small molecule photon donor (Figure 1) [10,11]. Luminol is nontoxic at diagnostic concentrations and no adverse effects have been observed in mice injected at doses of 250 mg/kg [12]. Luminol has also been used clinically for the treatment of autoimmune alopecia [12], and for the promotion of wound healing in early 1960s. If feasible, in vivo chemiluminescence imaging would be useful and complementary to structural imaging (e.g., MRI, X-rays).

Figure 1.

Mechanism of chemiluminescence generation.

Materials and Methods

Imaging

A cryogenically cooled high-efficiency charge-coupled device (CCD) photon detector (Roper Scientific, Trenton, NJ) was used to image luminol-enhanced photon generation in murine models. Similar imaging systems are in use for luciferase-generated bioluminescence imaging. Commercially available detection systems vary in sensitivity, and the most sensitive ones are capable of detecting 1–2 photons over background per hour. In vitro experiments of luminol/H₂O₂ dilution series were initially performed in black 96-well plates to determine relationship of concentration to imaging signals, reproducibility and sensitivity (15-min acquisition). Dilution series of H₂O₂ (10%, 5%, 1%, and 0.1%) were prepared to react with different amount of luminol. The dose range of luminol was between 0 and 5 mg due to limited solubility.

Animal Studies

In a first set of experiments, we used a lipopolysaccharide (LPS; Sigma, St. Louis, MO) induced acute focal arthritis model. In subsequent experiments, we used the KBxN arthritis model as a model of diffuse arthritis [13]. For the first set of experiments, anesthetized DBJ1 mice received intra-articular injection of variable amounts of LPS (1 × 10⁻⁷ to 1 × 10⁻⁵ g LPS in 20 μL of saline) as described previously [14]. This model was initially chosen because it allowed us to induce unilateral arthritis, whereas the contralateral joint in each animal served as a negative control. LPS was injected into the right ankle joint through the Achilles tendon using a 30-gauge needle while an identical volume of normal saline was injected into the opposite ankle joint. Two days after LPS administration (chosen because of maximum inflammatory reaction), mice received an intravenous injection of 5 mg of luminol (Fluka, Milwaukee, MI) and each animal was imaged for 15 min immediately thereafter to maximize photon detection. The mean, standard deviation, and sum of the photon counts were then determined over each ankle joint. Images were displayed in transparent pseudo-color overlay, permitting correlation of areas of chemiluminescence activity with anatomy.

Results

In initial experiments we determined the relationship between photon generation (i.e., imaging signal) and luminol/H₂O₂ concentrations (Figure 2). There was a clear concentration dependence of imaging signal and substrates (r² > .91), and measurements were highly reproducible (error <3%). In separate experiments, we determined the detection threshold of H₂O₂ to be in the femtomole range. Assuming peak tissue concentrations of 10% injected dose of luminol/g tissue, these data are indicative that photon generation should indeed be sufficient for in vivo sensing, even at H₂O₂ tissue concentrations at low μM levels.

Figure 2.

In vitro chemiluminescence assay to determine the extent of photon generation as a function of luminol and H₂O₂. Dilution series of H₂O₂ (10%, 5%, 1%, and 0.1%) were prepared to which luminol was added at differing concentrations (0–5 mg).

We next used a LPS (1 × 10⁻⁵ g LPS in 20 μL of saline) induced acute arthritis model. Local ROS production in LPS-injected joints was clearly evident in all animals investigated (n = 5; Figure 3). The measurements were highly reproducible (photon generation within 5% error) in animals upon repeated injections of luminol. Histology of LPS-injected joints showed abundant polymorphonuclear cells in the joint and the surrounding tissues. Nitroblue tetrazolium (NBT) staining for ROS showed sparse blue deposits of formazan in these areas. There was no such staining in the contralateral control joint (Figure 4).

Figure 3.

In vivo imaging reactive oxygen species. Chemiluminescence signal of both ankle joints overlaid onto a white light image shows exclusive photon generation of the affected joint.

Figure 4.

Histology: Staining for ROS shows blue formazan deposits in the affected joint (left) but not in the contralateral ankle joint (right).

We next asked whether variable degrees of inflammation would result in different levels of ROS and thus photon generation. For these experiments, four groups of mice received different amounts of LPS (control group: no LPS; “incipient inflammation group“: 1 × 10⁻⁷ g LPS; “moderate inflammation group”: 1 × 10⁻⁶ g LPS; and “severe inflammation group“: 1 × 10⁻⁵ g LPS) into the right ankle and imaging was performed identical to the above procedure. Figure 5 clearly shows increasing photon flux with increasing amount of inflammation in affected joints. Total photon counts were as follows in the four groups: Normal joint (saline injection; values represent mean ± SD): 6.3 ± 0.3 photons/acquisition; incipient inflammation group: 22 ± 13 photons/acquisition; moderate inflammation group: 262 ± 103 photons/acquisition; and severe inflammation group: 2859 ± 98 photons/acquisition. Control experiments without luminol injection or without LPS injection did not reveal any significant photon counts above baseline level, indicating that light generation was luminol dependent. Preliminary feasibility studies in the above mouse model also indicate that the method is sensitive enough to follow a therapeutic response to protective compounds such as AEOL, an SOD mimetic (50 mg/kg).

Figure 5.

Correlation between photon generation and disease severity. Different doses of LPS (1 × 10⁻⁵, 1 × 10⁻⁶, and 1 × 10⁻⁷ g) were injected into ankle joints. Increasing photon flux was noted with increasing amount of inflammation in affected joints (note the log scale; values in mean ± SD; p > .0001)</i>.

In additional experiments, we used the KBxN arthritis model. Forty-eight hours after serum transfer, joint associated signal was observed, however, in symmetrical fashion (411 ± 124 photons/acquisition).

Discussion

These results indicate that it is indeed feasible to image ROS in murine models directly. Photons emitted from the arthritic joint were significantly higher than those of control joints and were proportional to disease severity. Measurements were reproducible and repeatable in the same animal. Because luminol is administered as a single dose for each individual imaging experiments, it is unlikely to have long-lasting anti-inflammatory effects. The safety profile of luminol has also been shown to be high and no side effects have been in rodents at imaging doses [12].

Although the current study was designed as a proof-of-principle study, we envision that the described technology can be further improved to lower detection thresholds of ROS. With advancement in photon counting technology, it should also be feasible to reduce acquisition times to less than 15 min as used in this study. It should also be feasible to adapt the technology to fiber-optic detection systems to be used during surgical intervention. Furthermore, it has been reported that direct chemiluminescence assays are sensitive in measuring the antioxidation effect of nonsteroid anti-inflammatory drugs or other antioxidant therapies [15] and we indeed observed protective effects with a new SOD mimetic. In vivo chemiluminescence imaging, as developed here, is likely to become an important adjunct to the rapidly expanding in vivo imaging repertoire. It should be particularly suited for following ROS levels real-time in live animals and in testing the efficacy of novel therapies.