Original paper
DOI: 10.11152/mu.2013.2066.182.eus
Abstract
Aims: Pancreatic cancer and hepatocellular carcinoma are two of the most aggressive types of cancer with limited therapeutic options in stages of advanced disease. Our objective is to assess the safety and feasibility of injecting iron oxide nanoparticles (IONs) via endoscopic ultrasound (EUS)-guidance, both systemically and locally in the liver and pancreas in order to study new potential therapies for liver and pancreatic tumors. Material and methods: Six domestic pigs were used for our study design, and divided into three groups: two were injected in the portal vein, and other four were subjected to local exposure of IONs in the liver and pancreas, two each. The pigs were on a 7 days follow-up and necropsy was performed with their organs harvested.
A 3T MRI scan was also performed. Results: All animals underwent an endoscopic ultrasound fine needle injection (EUS-FNI) procedure without any complications. EUS-FNI procedure had an average time of 5 minutes and 21 seconds and consisted of 2 ml of ION injection. No perforations and no risk of potential bleeding were recorded. Macroscopic changes were observed only after pancreatic EUS-FNI. A significant amount of IONs was observed in the liver after local injection and after vascular EUS-FNI. The imaging results were confirmed by pathological examination with most of the IONs accumulated in Ito-like cells, Kupfer cells, and sinusoids. Conclusions: IONs have been widely studied for both diagnostic and therapeutic purposes. Their injection through EUS-guidance may develop new diagnosis strategies as well as curative or palliative therapies in pancreatic and liver tumors.
Keywords: endoscopic ultrasound–fine needle injection, iron oxide nanoparticles, liver, pancreas
Endoscopic ultrasound guided injection of iron oxide magnetic nano- particles for liver and pancreas: a feasibility study in pigs.
Bogdan Silviu Ungureanu
1, Daniel Pirici
2, Claudiu Mărgăritescu
3, Ioana Andreea Gheonea
4, Florian Nicu Trincu
5, Adrian Fifere
6, Adrian Săftoiu
1,71Research Center of Gastroenterology and Hepatology Craiova, University of Medicine and Pharmacy Craiova, Ro- mania, 2Department of Research Methodology, University of Medicine and Pharmacy Craiova, Romania, 3Depart- ment of Pathology, University of Medicine and Pharmacy Craiova, Romania, 4Department of Radiology, University of Medicine and Pharmacy Craiova, 5Faculty of Pharmacy, University of Medicine and Pharmacy of Craiova, Romania,
6Institute of Macromolecular Chemistry “Petru Poni”, Iasi, Romania, 7Department of Endoscopy, Gastrointestinal Unit, Copenhagen University Hospital Herlev Copenhagen, Denmark
Received 02.11.2015 Accepted 10.12.2015 Med Ultrason
2016, Vol. 18, No 2, 157-162
Corresponding author: Bogdan Silviu Ungureanu
Research Center of Gastroenterology and Hepatology Craiova, University of Medicine and Pharmacy
66 1 Mai Bvd,
200638 Craiova, Romania Phone/fax: +40 251 310.287 E-mail: [email protected]
Introduction
Pancreatic cancer (PAC) and hepatocellular carcino- ma (HCC) are two of the most aggressive malignancies with a grim prognosis, ranking on the 3rd and 4th place in terms of mortality worldwide [1]. Because both diseases are frequently diagnosed in advanced stages (either local
or metastatic), curative options may be no longer avail- able. This has brought forward the need of new potential therapies, which may prolong the patients’ life expectan- cy and improve their conditions [2,3].
Over the years, nanotechnology biomedical applications have opened a window in the research of new diagnostic and therapeutic settings for HCC and PAC [4,5]. Among the different nanoparticles available, iron oxide nanoparti- cles (IONs) have received extensive attention due to their availability and properties. From diagnostic procedures in enhanced magnetic resonance imaging (MRI) to local ther- apies such as magnetic hyperthermia or even as vectors for shipping cytotoxic drugs, IONs have proven to have a good safety profile [6]. A key feature for their biocompatibility and biodistribution is related to their general characteristics of physiochemical design and coating properties [7].
Regardless of their purpose and method of distribu- tion, either oral, local, or intravenously, the main objec- tive is to deliver a number of IONs as large as possible in the targeted tissue. Several organs, along with the pancre- as and liver can become deposit spaces as the iron may be encapsulated within their cells.
Endoscopic ultrasound (EUS) has evolved since its beginning to a routinely used procedure, which further blends interventional radiology techniques with mini- mal invasive surgery options. Real time imaging of the pancreas and liver provides important details about the tumor’s characteristics and may also serve as a precise setting for guiding therapeutic techniques [8]. After in- troducing EUS-guided fine needle aspiration, more pro- cedures have surfaced expanding the therapeutic arma- mentarium of endoscopic therapies [9,10]. EUS-guided interventions pose a great potential with a prospect of successful treatment results as compared to current avail- able techniques.
Development and improvements in cancer research have indicated EUS-guided fine needle injection (EUS- FNI) as a novel technique for local delivery of specific drugs [11]. The aims of the present study are to assess the safety and feasibility of injecting IONs via EUS- guidance, both systemically and locally in the liver and pancreas in order to study new therapies for liver and pancreatic tumors.
Materials and methods
The procedures were performed according to the Eu- ropean Legislations on animal rights, after obtaining a written approval from the Ethics Committee of the Uni- versity of Medicine and Pharmacy of Craiova (UMFCV).
IONs were synthesized at the Institute of Macromolecu- lar Chemistry “Petru Poni”, Iasi, through the co-precipita- tion method with citric acid solution. At first, 0.60 g of Fe- Cl3x6H2O was added to 2 ml deionized water and another solution was prepared by adding 0.21 g of FeCl2 4H20 to 0.5 ml of 2 m solution of HCl. These substances were vig- orously stirred after being added to 10 mL of DI water with citric acid. The resulting solution was titrated with 2 ml of 5 M of sodium hydroxide and stirred for 30 minutes until a black precipitate was formed, resulting in the Fe3O4 nano- particle suspension. The ferrrofluid was heated to 80oC, left for 2 hours and then centrifuged for 5 minutes at 900xg. As a final step the supernatant was added into water, a process which was repeated several times. Before being injected, the magnetic nanoparticle (MNP) solution was sonicated for a proper dispersion for several minutes.
Six domestic pigs were used for our study design and kept in special conditions. The experimental models were
subjected to fasting and liquids for 24 hours, respectively 6 hours before intervention. Premedication was adminis- tered intramuscularly and consisted of Ketamine 20 mg/
kgc (MSD Animal Health, Germany), Xylazine 2mg/
kgc (Bioveta A.S., Czech Republic) and Athropine 0.015 mg/kgc (Biofarm, Romania). Peripheral access was as- sured with an 18 G catheter (WellcathPlusTM, Wellmed, Noida, India) positioned on the marginal vein of the ear.
The pigs were intubated, maintained under general anes- thesia with Propofol 0,5 mg/kgc (Fresenius Kabi Austria GMBH – Austria) continuously, Fentanyl 3 μg/kgc (Ac- tavis Nordis A/S – Denmark) and Pavulone 0,1 mg/kgc (Pancuronium Bromide, Schering-Plough – USA) while all vital signs were monitored.
All procedures were performed with standard equip- ment intended for animal use only. The pigs were divided into three groups: two were injected in the portal vein, and other four were subjected to local exposure of IONs in the liver and pancreas, two each. EUS-guided injection of IONs was carried out with a linear array EUS scope (GFUCT140-AL5, Olympus, America), with a large in- terventional channel, coupled with a corresponding Evis Exera System (Olympus, America) and an AlokaPro- Sound 5500 Ultrasound System (Hitachi-Aloka, Tokio, Japan). EUS-FNI through a 19-gauge needle was pre- ferred for local and portal vein injection. The EUS-scope was passed through an overtube placed into the esopha- gus and advanced to the stomach until a good position of the targeted organs (liver or pancreas) was obtained (fig 1).
A 19-gauge EUS needle (Boston Scientific, USA) was inserted through the biopsy channel, and 2 ml of MNP solution was directly injected either in the por- tal vein or directly in the liver or pancreas using real time EUS guidance. For the liver, MNP were injected in the left lobe, while for the pancreas the MNP were directed to the head region. Portal vein EUS-FNI con- sisted of puncturing the vascular wall under real-time
Fig 1. Study design for EUS-FNI in the pan- creas on a pig model
EUS guidance and releasing the MNP solution into the bloodstream (fig 2).
The pigs were followed for the next 7 days, with close monitoring regarding any change in their behavior, food intake and body temperature. Animals were euthanized with a pentobarbital overdose and necropsy was per- formed with their liver and pancreas and other organs be- ing harvested. Kidneys and spleen were also collected to compare the quantity of additional deposits of MNPs on vascular EUS-FNI to local organ injection. Gross exami- nation was performed and organs were stored in buffered neutral formalin and sent for a 3T MRI (Philips Ingenia 3T, Netherlands) scanning with a special research coil.
After routine processing for paraffin embedding, 4µm-thick sections were cut from the tissue blocks and were further utilized for hematoxylin-eosin and Prus- sian blue staining in order to assess the histopathology and ferrous iron deposition. For Prussian blue staining, the slides were incubated in a 2% potassium ferricya- nide acidic solution for 30 minutes at 37°C, after which
they were counterstained with Nuclear Red, dehydrated, cleared and mounted with a xyle-based mounting me- dium. In order to best separate the Prusian blue and still be able to investigate the morphology of the tissue, the transmission light spectra of the Prussian blue and Nu- clear Red were separated by spectral unmixing using a Nuance FX multispectral camera capable of resolving 420-720nm spectral range with a resolution of 10nm, and the Nuance 3.0.2 software (PerkinElmer, Hopkinton MA, USA) (fig 3). Blue signal areas were next calculated on 40× images centered on either portal spaces or cen- trolobular veins, averaged (10 image captures for each anatomical region) and means plotted and compared (Mi- crosoft Excel, Microsoft Office 2010).
Results
All 6 pigs, with a weight range between 25-35 kg underwent a EUS-FNI procedure without any significant changes in their behavior or any sign of additional com- plications. No difficulties were encountered in identify- ing the vascular structures and the targeted organs under EUS-guidance. EUS-FNI procedure had an average time of 5 minutes and 21 seconds. During local EUS-FNI, a hyperechoic mass was created at the injection point, hav- ing a median diameter of 1.5 cm in the liver and 1.2 cm in the pancreas. After catheterization of the portal vein, the MNP solution was dispersed into the bloodstream with no immediate sign of thrombosis. No perforations of the gastric wall or other organs and no risk of potential bleeding were recorded during the procedures.
Necropsy results showed no signs of local or distant complications. Macroscopic changes were observed only after local injection in the pancreas with a black spot highlighting the concentration point where the MNP were injected (fig 4).
Fig 2. a) EUS-FNI of IONs in the left liver lobe, showing a wide spreading of the ferrofluid solution around the injection area as a hyperechoic mass; b) EUS-FNI in the portal vein pointing out the IONs being dispersed by the bloodstream; c) local EUS-FNI in the head of the pancreas showing hyperecho- ic images as small quantities of IONs.
Fig 3. Image processing for spectral unmixing separates sig- nals by unmixing the visible light into 10nm bands spectra. Red and blue spectra cubes have been defined (a), and these signa- tures have been used through the analysis. An exemplary RGB image (b) is shown in pure red and blue signals (c), as well as with individual red (d) and blue channels (e).
The organs were evaluated by 3 Tesla (3T) MRI scan- ning, which hallmarked MNP concentrations in the liver and pancreas. A significant amount of IONs was observed in the liver after local injection and after vascular EUS- FNI. Portal vein injection showed no sign of thrombosis, even though there were deposits of MNPs even in the pe- riphery area (fig 5a). Regarding the pancreas, MRI images were only relevant after local EUS-FNI procedures (fig 5b).
Pathological assessment showed various deposits within the selected organs. Local injection in the liver showed a large volume of MNPs in the targeted area and several deposits in small quantities further away diffus- ing into the parenchyma. In contrast, EUS-FNI in the portal vein revealed a large amount of IONs scattered in the hepatic lobules, as well as between the lobules and in the hepatic periportal space.
Initial hematoxylin and eosin staining revealed clear- cut iron pigment deposition in the portal veins and some of the peripheral sinusoids (fig 6a), with no visible he- mosiderin accumulation in and around the central veins (fig 6 b,c). More than 80% of all the visualized portal spaces presented this accumulation. However, when we utilized the specific Perl’s Prussian Blue histological staining, it became clear that the extent of deposition was much larger, with diffuse depositions in the lobule around the central vein (fig 6 d,e,f), the density of the de- posits increasing gradually with shorter distances to the portal spaces (fig 6 g,h). The pigment was so dense in the portal vein, that the spectral unmixing visualized the reddish appearance under the blue dense shade of Perl’s staining. None of the intralobular deposits showed the same bi-phasic spectral signature due most probably to
the dilution of the particles downstream the blood flow.
High resolution images revealed that most of the diffuse deposits in the lobule were present in Ito-like cells (fig 6 j1), Kupffer cells (fig 6 j2) or in the sinusoid themselves (fig 6 j3). Only on rare occasions it seems that the depos- its occurred in the hepatocytes.
When we evaluated the blue areas in Perl’s stain- ing, this showed that despite the fact that the densest deposits were situated around the portal spaces, these were not the largest. This analysis showed that blue ar- eas around centrolobular spaces were significantly larger (27019.85±3290.38µm2 and 15589.53±6526.35µm2) than those around the periportal areas (3290.38±195.16µm2 and 6526.35±290.62µm2), and this staining decreased from perihilar regions compared to peripheric lobar ar- eas. Periportal regions had no such variation between the two anatomical sampling sites, and all instances showed significantly higher values compared to the control cor- responding liver areas (fig 7).
Local pancreas EUS-FNI showed a large area of MNP deposits with a localized inflammatory tissue reaction at about 2 cm away from the injected area. A mild inflam- mation of the pancreatic parenchyma with fat necrosis and atrophy characterized the surrounding tissues.
Fig 4. Necropsy reveals a large deposit of IONs as a black spot at the injection point in the head of the pancreas (arrow)
Fig 5. 3T MRI images recorded after organ’s harvesting point- ing out deposits of IONs within the liver a) and the pancreas b).
Fig 6. Pathological assessment of the iron deposition in the liv- er. Only dense hemosiderin pigment is identified on hematoxy- lin and eosin, mainly in and around the portal spaces (a), while almost no pigment can be identified around the centrolobular spaces (b-c). Spectral unmixing on Perl’s stained slides reveals more diffuse depositions around the central lobular areas (d,e,f) compared to peripheral areas (g,h,i). Most of the diffuse deposi- tions appeared to be accumulating in Ito-like cells (j, j1), Kup- fer cells (j, j2) and in the sinusoids (j, j3).
Discussions
Treatment options for patients with liver and pancreatic tumors are known to be relatively limited if the disease is diagnosed in an advanced stage. Lately, magnetic nanopar- ticles have served as diagnostic tools in contrast imaging MRI in humans [12] and as therapeutic methods such as hy- perthermia, drug carriers or gene therapies in experimental models [13]. Tissue reactions to MNP’s physico-chemical properties are known to be influenced by their general prop- erties and coatings, biocompatibility and toxicity and none- theless methods of distribution [14]. Depending on their purpose they have been injected locally within a region of interest, orally after gavage following the absorption route or intravenously trying to direct them to the tumor. How- ever, these techniques are rather difficult to control as the biological response of the targeted tissue requires a large amount of MNPs to fulfill the therapeutic effect [15]. Gen- erally, vascular injection and distribution of MNPs require their passage through three phases: blood stream clearance, extravasation and interstitial space depositing from where they may attack the cancer cells. So far, there are few stud- ies that exploit the potential of nearby vascular structures injection mainly because of their difficult accessibility.
EUS may be able to fill up this gap, because of its great imaging potential and various therapeutic options.
Our study focused on the distribution of MNPs through several methods of injection trying to emphasize their liver and pancreas enhancement. EUS-FNI may be taken into consideration as a palliative option either by local injection or by gaining vascular access. Its potential in diagnostic and therapeutic settings based on real-time
imaging may overcome the flaws of interventional radi- ology. There is no doubt that EUS-guidance provides a more attractive option for vascular therapies, as both ma- jor and smaller vessels near the gastrointestinal tract can be traced and easily accessed.
EUS-FNI has the ability to target the liver and pan- creas using a transgastric and transduodenal approach.
So far this setting has been tested with intratumoral etha- nol injection [16,17] and chemotherapeutic drugs [18].
EUS-FNI ethanol injection in the pancreas in a porcine model has been proven to be technically possible and rather safe with mild side effects [16]. Our results after local MNP injection in the pancreas showed no evidence of an important pancreatic reaction confirming that EUS- guided injection is technically successful because of the nearby position of the pancreas near the gastric wall.
Similar results were noted by Kai et al [19] who injected an OncoGel solution in a porcine pancreas tail and ob- served the deposits 14 days after exposure proving that the technique is feasible and could be a potential therapy for advanced pancreatic cancer.
Liver delivery of MNP was assessed while using different ways of administration. EUS-FNI local de- livery showed a large amount of MNP concentrated in the targeted area with only small concentrations in the surrounding liver tissue. Even though there are concise protocols for patients with HCC tumors, the use of EUS- FNI has been noted so far in animal settings, as well as some case reports [20,21]. On the other hand vascular access in the portal vein and delivery of MNP showed a wider diffusion of MNP in the entire liver, up to the distal branches. This setting seems to be a more appropriate op- tion if a larger number of tumors are targeted.
Local EUS-FNI and portal vein EUS-FNI might be used for different tumor types. While local injection may be addressed to a solitary liver tumor (such as HCC or solitary metastases focusing on a larger concentration of MNP in the region of interest), the potential of portal vein injection could be directed in hepatic metastases cover- ing most of the liver tissue. Also, specific coated MNP with different chemotherapeutic drugs injected through the portal vein could lower the systemic toxic effects caused by peripheral injection. This setting was studied by Faigel et al [22] who compared the level of several chemotherapeutic drugs after injecting them in the portal vein and systemically in several experimental pig mod- els. After comparing the results he observed that irinote- can, doxorubicin loaded microbeads and paclitaxel na- noparticles had higher concentrations in the liver tissue after a portal vein injection as compared to a systemic injection, while toxic levels were almost halved espe- cially in the cardiac tissue. MRI images were consistent Fig 7. Semiquantitative analysis of Perl’s staining shows de-
creasing staining areas from the hilar to peripheral lobar regions.
At the histological level, the areas were much larger around the centrolobular spaces compared to the periportal regions, and the difference was significant only for hilar anatomical regions.
All injection sites demonstrated a larger signal area compared to the control tissue. * represents a significance on Student’s t testing of p<0.05. Bars represent standard error of the means.
with the pathological examination when comparing local EUS-FNI to portal vein access.
This study has several limitations consisting of the small number of animals and the fact that the MNP distribution was assessed only with MRI scanning and pathological examination. While trying to validate the EUS technique we did not measure the iron concentra- tion within the tissues and also did not focus on urinary samples to see the MNP elimination. Nevertheless, the MNP concentration can be easily estimated through im- aging, leading to a better quantification inside the desired region of interest that will be targeted through magnetic hyperthermia.
Conclusions
Our study focused on showing the feasibility of EUS- FNI of MNPs in the liver and pancreas by local or vascular access and their organ distribution. EUS-guidance offers a unique access to these organs and nonetheless to nearby vascular structures, which facilitates desirable therapeutic techniques, which have been the target of interventional radiology over the years. MNPs have been widely studied for HCC and PAC in different therapeutic scenarios such as hyperthermia or coated with different chemotherapeu- tic drugs. Their injection through EUS-guidance may de- velop new strategies of diagnosis or curative or palliative therapies in pancreatic and liver tumors.
Acknowledgement: This article was financed by the Partnership program in priority areas – PN II, imple- mented with support from National Authority of Scien- tific Research (ANCS), CNDI – UEFISCDI, project nr.
2011-3.1-0252 (NANO-ABLATION).
Conflict of interest: none
References
1. Ferlay J, Shin HR, Bray F, et al. GLOBOCAN 2012. Ac- cessed on 10/10/2015.
2. Galun D, Basaric D, Zuvela M, et al. Hepatocellular carci- noma: from clinical practice to evidence-based treatment protocols. World J Hepatol 2015; 7: 2274-2291.
3. Cid-Arregui A, Juarez V. Perspectives in the treatment of pancreatic adenocarcinoma. World J Gastroenterol 2015;
21: 9297-9316.
4. Rosenberger I, Strauss A, Dobiasch S, et al. Targeted diag- nostic magnetic nanoparticles for medical imaging of pan- creatic cancer. J Control Release 2015; 214: 76-84.
5. Yan SY, Chen MM, Fan JG, et al. Therapeutic mechanism of treating SMMC-7721 liver cancer cells with magnetic fluid hyperthermia using Fe2O3 nanoparticles. Braz J Med Biol Res 2014; 47: 947-959.
6. Krishnan KM. Biomedical Nanomagnetics: A spin through possibilities in imaging, diagnostics and therapy. IEEE Trans Magn 2010; 46: 2523-2558.
7. Gobbo OL, Sjaastad K, Radomski MW, Volkov Y, Prina- Mello A. Magnetic nanoparticles in cancer theranostics.
Theranostics 2015; 5: 1249-1263.
8. Săftoiu A. State-of-the-art imaging techniques in endoscop- ic ultrasound. World J Gastroenterol 2011; 17: 691-696.
9. Vilmann P, Jacobsen GK, Henriksen FW, Hancke S. Endo- scopic ultrasonography with guided fine needle aspiration biopsy in pancreatic disease. Gastrointest Endosc 1992; 38:
172–173.
10. Costache MI, Iordache S, Karstensen JG, Săftoiu A, Vilmann P. Endoscopic ultrasound-guided fine needle aspiration: from the past to the future. Endosc Ultrasound 2013; 2: 77-85.
11. Yan BM, Van Dam J. Endoscopic ultrasound-guided intra- tumoural therapy for pancreatic cancer. Can J Gastroenterol 2008; 22: 405-410.
12. Li L, Jiang W, Luo K, et al. Superparamagnetic iron oxide nanoparticles as MRI contrast agents for non-invasive stem cell labeling and tracking. Theranostics 2013; 3: 595–615.
13. Yu Y, Sun D. Super paramagnetic iron oxide nanoparticle
“theranostics” for multimodality tumor imaging, gene de- livery, targeted drug and prodrug delivery. Expert Rev Clin Pharmacol 2010; 3: 117–130.
14. Trincu NF, Balseanu TA, Ungureanu BS, et al. Blood clear- ance of citric acid-coated superparamagnetic iron oxide na- noparticles in rats – a pilot study. Current Health Sciences Journal 2015; 41: 303-306.
15. Antonelli A, Magnani M. Red blood cells as carriers of iron oxide-based contrast agents for diagnostic applications. J Biomed Nanotechnol 2014; 10: 1732-1750.
16. Zhang WY, Li ZS, Jin ZD. Endoscopic ultrasound-guid- ed ethanol ablation therapy for tumors. World J Gastroen- terol 2013; 19: 3397-3403.
17. Park do H, Choi JH, Oh D, et al. Endoscopic ultrasonogra- phy-guided ethanol ablation for small pancreatic neuroen- docrine tumors: results of a pilot study. Clin Endosc 2015;
48: 158-164.
18. Sharma V, Rana SS, Bhasin DK. Endoscopic ultra- sound guided interventional procedures. World J Gastroin- test Endosc 2015; 7: 628-642.
19. Matthes K, Mino-Kenudson M, Sahani DV, et al. EUS guided injection of paclitaxel (OncoGel) provides thera- peutic drug concentrations in the porcine pancreas (with video). Gastrointest Endosc 2007; 65: 448-453.
20. Klapman JB, Chang KJ. Endoscopic ultrasound-guided fine-needle injection. Gastrointest Endosc Clin N Am 2005;
15: 169-177.
21. Verna EC, Dhar V. Endoscopic ultrasound-guided fine nee- dle injection for cancer therapy: the evolving role of thera- peutic endoscopic ultrasound. Therap Adv Gastroenterol 2008; 1: 103-109.
22. Faigel D, Lake D, Landreth T, Kelman C, Marler R. En- doscopic ultrasonography-guided portal injection chemo- therapy for hepatic metastases Gastrointest Endosc 2014;
3(Suppl 1): S1.
