Hyaluronic Acid-Modified Magnetic Iron Oxide Nanoparticles for MR Imaging of Surgically Induced Endometriosis Model in Rats

Endometriosis is defined as the presence of endometrial tissue outside the uterine, which may affect nearly 60% of women in reproductive age. Deep infiltrating endometriosis (DIE) defined as an endometriotic lesion penetrating into the retroperitoneal space or the wall of the pelvic organs to a depth of at least 5 mm represents the most diagnostic challenge. Herein, we reported the use of hyaluronic acid (HA)-modified magnetic iron oxide nanoparticles (HA-Fe3O4 NPs) for magnetic resonance (MR) imaging of endometriotic lesions in the rodent model. Sixteen endometriotic lesions were surgically induced in eight rats by autologous transplantation. Four weeks after lesion induction, three rats were scanned via MR imaging after tail vein injection of the HA-Fe3O4 NPs. Accordingly, the remaining five mice were sacrificed in the corresponding time points. The ectopic uterine tissues (EUTs) were confirmed by histological analysis. Quantification of Fe in the EUT was also performed by inductively coupled plasma-optical emission spectroscopy. Our results showed that by using the HA-Fe3O4 NPs, the EUTs were able to be visualized via T2-weighted MR imaging at 2 hours post injection, corroborating the Prussian blue staining results. The developed HA-Fe3O4 NPs could be used as negative contrast agents for sensitively detecting endometriosis in a mouse model and may be applied for future hyperthermia treatment of endometriosis.


Introduction
Endometriosis is defined as the presence of endometrial tissue outside the endometrium and the myometrium, which may affect nearly 60% of women in reproductive age [1]. Till now, the true mechanism of etiology, pathophysiology, and progression in endometriosis still remains unclear [2]. Deep infiltrating endometriosis (DIE) is defined as an endometriotic lesion penetrating into the retroperitoneal space or the wall of the pelvic organs to a depth of at least 5 mm [3]. Some locations including the posterior fornix, uterosacral ligaments, rectum, vagina, and bladder can be affected by DIE lesions. In this condition, accurately preoperative evaluation of both presence and extension of disease may be helpful for the selection of complete surgical excision.
With the advantages of superb tissue contrast resolution and no ionizing radiation, magnetic resonance (MR) imaging is generally performed for problem solving in determining the etiology of indeterminate sonographic lesions in the female genital tract [4][5][6][7][8].
It is well known that the typical MR findings of endometriosis are hyper intensity on both T 1 and T 2 weighted image (T 1 WI/T 2 WI) as a result of hemorrhagic components associated with menstrual bleeding [9]. For diagnosis of DIE, some lesions may be overlooked on both non-contrast and contrast enhanced MR images for on one hand the lesions often occupy the deep aspects of pelvic organs and on the other hand, the small lesion size as well as atypical signal character usually make it difficult to be differentiated from adjacent normal tissues [10,11].
Nowadays, the most common commercially available MR contrast agent is T 1 agent (Gd/chelator complexes) used in clinical settings for cancer diagnosis [12]. However, with respect to endometriosis diagnosis, the application of T 1 agent provides limited information due to the mild enhancement of lesions that may be largely shaded by hyperintensity on both T 1 WI and T 2 WI images. Therefore, it is reasonable to speculate that T 2 -negative contrast agents may be able to improve the detection rate of small lesions as a result of the increased lesion to background contrast. In recent years, magnetic iron oxide nanoparticles (Fe 3 O 4 NPs) have gained increasing attention in various biomedical applications [13], especially in cancer MR imaging owing to their high r 2 relaxivity [14][15][16][17][18][19][20]. However, to our best knowledge, to image the endometriosis using the Fe 3 O 4 NPs using MR imaging has not been reported.
In our previous work [21], we have successfully synthesized hyaluronic acid (HA)-modified Fe 3 O 4 NPs (HA-Fe 3 O 4 NPs) that can be used as negative contrast agents for MR imaging of cancer cells overexpressing CD44 receptors. It is known that, HA is a member of the glycosaminoglycan family composed of repeating disaccharide units of D-glucuronic acid and N-acetyl-D-glucosamine and has been expanded into various biomedical applications due to its biocompatibility, biodegradability and non-immunogenicity. What's more, the hydrothermally synthesized Fe 3 O 4 NPs show good water solubility, colloid stability and biocompatibility after being modified by HA. So it may be a very meaningful attempt to detect the endometriosis by using the HAmodified Fe 3 O 4 NPs as contrast agents.
In this study, we reported our initial results with HA-Fe 3 O 4 NPs in in vivo imaging of endometriotic lesions in rats. The morphology of the HA-Fe 3 O 4 NPs was firstly confirmed via transmission electron microscopy (TEM) imaging. Subsequently, the HA-Fe 3 O 4 NPs were used for MR imaging of experimentally induced endometriotic lesions in rats.

Synthesis and characterization techniques
The HA (Mw = 5,805)-modified Fe 3 O 4 NPs (HA-Fe 3 O 4 NPs) were synthesized and characterized according to our previous protocol [21]. A JEOL 2010F analytical electron microscope (JEOL, Tokyo, Japan) was used to characterize the morphology of the HA-Fe 3 O 4 NPs at an operating voltage of 200 kV. The sample was dispersed in ethanol. A dilute particle suspension (10 mL) of the sample was then deposited onto a carbon-coated copper grid and dried in air before measurements. The Fe content in the particle suspension and induced endometriotic lesions was measured by using a Leeman Prodigy inductively coupled plasma-optical emission spectroscopy (ICP-OES, Hudson, NH, USA). The effect of MR imaging for HA-Fe 3 O 4 NPs was evaluated at 3.0 Tesla MR imaging machine (Siemens, Erlangen, Germany). Samples were diluted with water to have an Fe concentration in the range of 0.28-8.96 mg/mL before measurements. The instrumental parameters were set as follows: point resolution = 156 mm6156 mm, section thickness = 1.5 mm, TR = 4960 ms, TE = 85 ms, and number of excitation = 1.

Animal models
Animal experiments were carried out according to protocols approved by our institutional committee (obstetrics and gynecology hospital, Fudan University) for animal care and also in accordance with the policy of the National Ministry of Health.
Female Sprague-Dawley rats (200-240 g, Shanghai Slac Laboratory Animal Center, Shanghai, China) were used. Eight rats were anesthetized by intraperitoneal injection of pentobarbital sodium (30 mg/kg). The abdomen was shaved and disinfected with a 75% alcohol solution before covering with sterile drapes. A midline open surgery was performed 1 cm cephalad to the symphysis pubis with a 2-3 cm vertical incision. After exposing both uterine horns, the left uterine horn ligated proximally and distally with 560 sutures and then a segment 1.0 cm in length was resected and placed in phosphate buffered saline (PBS) at 37uC. The cutting uterine horn was trimmed to remove excess fat and endometrium was carefully striped and split longitudinally into two segments of rectangles (3615 mm). Each of endometrium segment was fixed using two 560 sutures to the peritoneal side of bilateral abdominal wall 1 cm apart from the incision with the endoluminal side facing abdomen. Lastly, the abdomen was closed layer by layer with a running 560 suture for peritoneum and musculature, and 4-0 nylon sutures for the skin.   Histological analysis and quantification of Fe in the ectopic uterine tissue (EUT) Four mice (four weeks after surgery) with HA-Fe 3 O 4 NPs (Fe mass = 2.0 mg) intravenously delivered via the tail vein were euthanatized at the time points of 0.25, 0.5, 1 and 2 h. When the abdominal cavity was exposed, the induced endometriotic lesions were carefully examined and excised. The largest lesion was assessed with a calliper. The eutopic uterine and the largest EUTs were stripped and cut into 10 cm-length segments, respectively. Tissues were then fixed in buffered formalin, and cut in 5 mm thick sections. For histology analysis, tissues were stained with haematoxylin and eosin (HE) and Prussian blue. In order to quantify Fe concentration in the EUTs, the extracted EUTs were weighed. After being digested by aqua regia (nitric acid/ hydrochloric acid, v/v = 1:3) for 2 days, the Fe content in EUTs was determined by ICP-OES. The EUTs in mouse without intravenous injection of HA-Fe 3 O 4 NPs were used as control.

Statistical Analysis
Quantitative data were expressed as mean 6 standard deviation (SD). Means were compared using Student's t-test. P values ,0.05 were considered statistically significant.

Synthesis and characterization of HA-Fe 3 O 4 NPs
The HA-Fe 3 O 4 NPs were synthesized and characterized according to our previous work [21]. The morphology of the formed HA-Fe 3 O 4 NPs was characterized with TEM ( Figure 1). It can be clearly seen that the particles with a spherical or quasispherical shape have a quite uniform size distribution and the polymer shell on the outer surface of the NPs can be clearly seen. As contrast agents for MR imaging, the T 2 -weighted imaging effect was also evaluated by the 3.0 T Tesla MR scanner (Figure 2). It can be seen that the NPs are able to obviously induce the decrease of MR signal intensity with the increase of Fe concentration. This result suggests that the HA-Fe 3 O 4 NPs may be used as an effective contrast agent for T 2 MR imaging applications.

The gross findings of the EUTs in mice
All surgically induced endometriotic lesions ( Figure 3) were successfully performed and proven by histological analysis (Figure 4). The endometriotic lesions in experimental model appeared to have a tabular, cystic structure filled with fluids. Although the cutting endometrium flaps were fixed abutting abdominal wall, the endometriotic lesions were usually found at the peritoneal cavity when exposed. Only one largest lesion (the average long diameter was about 17.867.5 mm) in each of experimental model was commonly detected in the studied samples.

In vivo MR imaging of surgically induced endometriotic lesions in rats
After intravenous injection of the HA-Fe 3 O 4 NPs (1 mL in PBS, 2 mg Fe/mouse) into the experimental mice via the tail vein, MR scanning was performed ( Figure 5). The ectopic lesions appeared as the ill defined, cystic mass with low signal on T 1 WI ( Figure 5A). On turbo spin-echo FS-T 2 WI, the cystic components displayed the homogeneous high signal, whereas the cystic wall showed the low signal ( Figure 5B). It was clear that the walls of EUTs MR signal for the mice injected with HA-Fe 3 O 4 NPs gradually decreased with the time post injection ( Figure 5C-5F). At 2 h post injection, the particles were able to induce the highest contrast enhancement  ( Figure 5F). Owing to the anatomic structure overlap, the ectopic uterine could not be definitely indicated on both T 1 WI and T 2 WI images. Quantitative analysis of the MR T 2 -weighted signal intensity of the lesion wall at different time points revealed that the lesion wall had a lower signal noise ratio (SNR) value at 2 h after injection ( Figure 6). T 2 -weighted signal intensity of the lesion wall before injection of the particles and at 0.25, 0.5, 1 and 2 h post injection were measured to be 1341642.9, 1069648.4, 1018629.4, 927648.2, and 732625.2, respectively ( Figure 6). The T 2 WI signal intensity differences between mice before injection and at 2 h post injection differed significantly (P = 0.000).

Histological findings and quantitative evaluation of HA-Fe 3 O 4 NPs enhanced EUTs
HE staining results demonstrated that in EUTs, the endometrial stroma was thin and have fibrosis structure without glands and muscular layers ( Figure 4B) when compared with eutopic uterine tissues ( Figure 4A) (Figures 8A and 8B).

Discussion
Endometriosis, especially DIE, remains a challenging condition for either clinicians or radiologists. It was reported that for deep pelvic endometriosis, the sensitivity of MR imaging for the diagnosis of endometriosis ranged from 76% to 88% in some specific anatomic locations [22]. Some small, insidious lesions may be overlooked even by the experienced pelvic radiologists because on one hand, these lesions often appear no specific MR signals and on the other hand, the complex pelvic floor anatomical structure may obscure the underlying deep endometriotic lesions. Therefore, accurately outlining the the suspected lesions with noninvasive method always has priority before the operation.
In a previous study, Schreinemacher et al. reported the endometriosis detection using dynamic contrast-enhanced MR imaging with gadofosveset-trisodium as a contrast agent in a mouse model [23]. They found that the contrast agents persisted longer in endometriotic tissues due to the extensive angiogenesis in induced lesions and thus achieving enhanced MR signal intensity on T 1 WI image. In another study, ultrasmall superparamagnetic iron oxides (USPIO)-enhanced T 2 WI was used to detect the EUT with a size larger than 3 cm [24]. Large-sized endometriosis might   have high tissue concentrations of clumped USPIO particles, resulting in focal regions of hypointensity signal. Overall, studies focusing on the use of newly discovered contrast agents to detect the endometriosis in animal models were still extremely limited.
In this study, we reported the detection of surgically induced endometriotic lesions in rats with HA-Fe 3 O 4 NPs. Our results disclosed that HA-Fe 3 O 4 NPs intravenously delivered via tail vein were able to improve the conspicuity of the EUTs in rodent model.
Iron oxide nanoparticles are known to be non-toxic and eventually biodegraded to form blood hemoglobin. Fe 3 O 4 NPs with appropriate surface chemistry have been widely used for various biomedical applications including cell and protein separation, drug and gene delivery, tissue repair, hyperthermia, and MR imaging [25][26][27][28][29]. Various polymers such as albumin, dextran [30], dendrimers [31], polyethylene glycol (PEG) [18,32,33], or polyethyleneimine (PEI) [34] have been coated onto Fe 3 O 4 NP surfaces to improve their stability or/and decrease the uptake by the reticuloendothelial system (RES). In recent years, many kinds of multifunctional magnetic NPs with the variable synthetic structures (for example, core/shell structure) have also been reported for different biomedical applications [13].
In our previous studies, we reported the facile hydrothermal synthesis and surface functionalization of branched polyethyleneimine (PEI)-coated iron oxide nanoparticles (Fe 3 O 4 2PEI NPs) for biomedical applications, especially for MR imaging of different types of cancer [21,35]. These prior successes lead us to hypothesize that Fe 3 O 4 NPs, as a T 2 negative contrast agent, could also be applied in detecting endometriotic lesions which has high angiogenic activity [36]. The EUTs may be enhanced similarly with the enhanced permeability and retention (EPR) effect as in the solid tumor [35].
Surgically induced endometriotic lesions were well recorded in the literature [37]. The most experienced experimental model is the autotransplantation of uterine tissue into the peritoneal cavity in the rodent model. In this study, all surgical procedures were successfully performed in all experimental mice. The EUTs abutting the abdominal wall always appeared to have the tubular, cystic structure filled with clear fluid. Our findings were also in line with that reported in the literature [24], where about 4 weeks after establishing the animal model, the size of EUTs seemed to be the largest lesion volume and thereafter, it varied. Histologically, the EUTs in rats contain endometrial glands and stroma, which are likely the human endometriotic lesions. In our study, the excellent lesion configuration in the T 2 WI may be attributed to the following reasons: Firstly, the size of the used HA-Fe 3 O 4 NPs was approximately 15.6 nm in diameter [21], which is larger than USPIO used in the literature. The demonstrated relatively high r 2 relaxivity of HA-Fe 3 O 4 NPs may render them to be more sensitive to magnetic susceptibility effects. Secondly, the coating of HA onto the surface of Fe 3 O 4 NPs may retain their longer blood circulation time, accordingly leading to more phagocytotic activities by endothelial cells [38][39][40]. Both these factors may lead to the perfect MR imaging of EUTs.
In our previous study [21], we found that after 24 h postinjection of HA-Fe 3 O 4 NPs in nude mice, the majority of the Fe was uptaken by the liver and spleen while only a quite small amount of Fe remained in the other organs, such as heart, lung, kidney and tumor. The Fe accumulation in the liver and spleen are typical due to the clearance effect of RES located in these organs [41].
Further, after the injection of the HA-Fe 3 O 4 NPs, there was a sharp contrast between negatively enhanced cystic wall and cystic components with high signals in the EUTs in T 2 WI, which made the lesion margin be easily recognized. Such character makes a very important clinical significance. In human endometriosis, the lesions always appear as the high signal on both T 1 WI and T 2 WI images and therefore, the T 1 -contrast agents may add little useful information in accurate identification of the margin for the enhanced components on T 1 WI may be shaded by the majority of high signals of cystic components. From this point of view, T 2 - There were several limitations in this study. First, unlike humans, rodent animals do not shed their endometrial tissue and do not develop endometriosis spontaneously [37]. It is therefore uncertain whether the surgically induced endometriotic lesions are similar to humans. Second, the implantation, adhesion and development may be affected by the estrous level. The size of EUTs may also vary according to the estrous cycle. Under the surveillance of estrous level and at the much longer time interval, HA-Fe 3 O 4 NPs-enabled MR imaging of the EUTs may be more reasonably evaluated.

Conclusion
By using the HA-Fe 3 O 4 NPs, the surgically induced endometriosis lesion in rats can be clearly outlined. HA-Fe 3 O 4 NPs may be used as negative contrast agents for sensitively detecting of endometriosis and be applied for future hyperthermia treatment of endometriosis.