Usefulness of Photocrosslinkable Chitosan for EndoscopicCancer Treatment in Alimentary Tract

Takuya Hayashi,1 Tomokazu Matsuyama,2 Kenji Hanada,3 Kuniaki Nakanishi,4 Maki Uenoyama,1

Masanori Fujita,5 Masayuki Ishihara,6 Makoto Kikuchi,5 Tomosumi Ikeda,1 Hisao Tajiri7

1 Division of Environmental Medicine, National Defense Medical College Research Institute, Tokorozawa, Saitama, Japan

2 2nd Department of Surgery, National Defense Medical College, Tokorozawa, Saitama, Japan

3 Department of Internal Medicine, Japan Self Defense Central Hospital, Tokyo, Japan

4 1st Department of Pathology, National Defense Medical College, Tokorozawa, Saitama, Japan

5 Department of Medical Engineering, National Defense Medical College. Tokorozawa, Saitama, Japan

6 Division of Biomedical Engineering, National Defense Medical College Research Institute, Tokorozawa, Saitama, Japan

7 Department of Endoscopy, Jikei University School of Medicine, Tokyo, Japan

Received 4 September 2003; revised 1 April 2004; accepted 2 April 2004Published online 21 September 2004 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.b.30099

Abstract: Endoscopic mucosal resection (EMR), an established treatment for superficialcancer in the stomach, colon, and esophagus, involves extracting the lesion with the aid ofhigh-frequency current (as in snare polypectomy), following saline injection into the submu-cosal layer. Of interest in this study was the ability of photocrosslinkable chitosan, which isconverted to a hydrogel (like a soft rubber) by 30-s ultraviolet irradiation, to expand thesubmucosal layer when used as an injection material. Photocrosslinkable-chitosan solution ornormal saline (0.3 mL) was injected into the submucosal layer of the rat stomach, and thethickness of this layer was measured at various time points thereafter. In these two groups, thevalues obtained were 3.8 and 2.0 mm, respectively (means; n � 5 each group; p < 0.005), at30 min after the injection. The cumulative blood loss measured at 20 min after mucosalincision was 113 mg (photocrosslinkable-chitosan solution), versus 1682 mg (saline) (means, n� 10 each group; p < 0.001). Histologic observation revealed that chitosan, which wasretained within the submucosal layer, filled the ulcer base and completely surrounded thebleeding focus. Thus, the photocrosslinkable chitosan developed holds promise for use in EMRas a submucosal injection agent that avoids complications. © 2004 Wiley Periodicals, Inc. J BiomedMater Res Part B: Appl Biomater 71B: 367–372, 2004

Keywords: photocrosslinkable chitosan; endoscopic mucosal resection; gastric perforation;hemostasis; submucosal layer

INTRODUCTION

Chitosan, a natural polysaccharide derived by deacetylationof chitin,1 has been shown to have hemostatic2 and antibac-terial3 properties, and to accelerate wound healing.4 Owing totheir poor water solubility, however, only a few biomaterialscontaining chitosan have been studied; these have been in theform of filaments, powders, or sponges, or as a composite

with cotton or polyester.1,2 Recently, Ono et al. made twoimportant advances: They introduced lactose moieties intochitosan, providing much better water solubility, and added aphotoreactive azide group, providing a hydrogel-formingability through crosslinking with amino groups.5 Irradiationof a photocrosslinkable-chitosan (Az-CH-LA) aqueous solu-tion with ultraviolet light (UV) results in the formation of aninsoluble, flexible hydrogel (like a soft rubber). The sealingstrength of 3% Az-CH-LA aqueous solution has been foundto be greater than that of fibrin glue in preventing air leakagefrom a pinhole in the aorta and also from an incision in theisolated pig trachea.6 In fact, it effectively seals air leakagesfrom lung or trachea, and completely stops bleeding in vivo

Correspondence to: Takuya Hayashi (e-mail: hayashitakuya@mips.ne.jp)Contract grant sponsor: Ministry of Health, Labor, and Welfare, Japan; contract

grant number: Grant-in-Aid for Cancer Research 14-18

© 2004 Wiley Periodicals, Inc.

367

from a pinhole in the rabbit carotid artery within 30 s of UVirradiation.6 Furthermore, the gel has been found to enhancewound contraction and healing.7,8

Because of its viscosity and gelatinization ability it wasthought that this biomaterial might, following its submucosalinjection, facilitate endoscopic mucosal resection (EMR) inthe treatment of superficial tumors of the alimentary tract.EMR is a technique in which the lesion is extracted (as insnare polypectomy) after it is elevated with the aid of a salineinjection into the submucosa.9,10 When the submucosal ex-pansion is insufficient, the snaring forceps may grasp themuscle layer, leading to wall perforation.10–12 For that rea-son, there is a need for an aid that allows a sufficient elevationof the mucosal lesion to be achieved more reliably. Anothermajor complication of conventional technique is bleed-ing.10–12

In the present study, basic animal experiments were per-formed to test the ability of this viscous material (a) to keepthe mucosa elevated sufficiently for a prolonged period oftime, and (b) to reduce bleeding by virtue of its gelatinizationand sealing ability following UV irradiation.

MATERIALS AND METHODS

Preparation of Photocrosslinkable-Chitosan Hydrogel

Photocrosslinkable chitosan (Az-CH-LA) was prepared aspreviously described.5 The chitosan used had a molecularweight of 800–1000 kDa, with an 80% degree of deacetyla-tion (Yaizu Suishankagaku Industry Co. Ltd., Shizuoka, Ja-pan). In Az-CH-LA, azide (p-azidebenzoic acid) and lactose(lactobionic acid) moieties have been introduced into thechitosan molecule through a condensation reaction with theamino group, and it has been estimated that 2.5 and 2% of theamino groups in the chitosan are substituted by p-azideben-zoic acid and lactobionic acid, respectively. The introductionof lactose results in a water-soluble chitosan. A viscousAz-CH-LA aqueous solution (20–30 mg/mL) is convertedinto an insoluble hydrogel within 10 s upon UV irradiation ata lamp distance of 2 cm [UV-irradiation system: Spot CureML-251C/A with a guide fiber unit (SF-101BQ) and 250-Wlamp (240–380 nm); Ushio Electrics Co. Ltd., Tokyo, Japan]through crosslinking of the azide and amino groups of theAz-CH-LA molecule.

A 2.5% Az-CH-LA aqueous solution was used because apreliminary study on the rat stomach found this concentrationto be most suitable, in that it can be readily introduced thougha 25-gauge stainless-steel needle; and it is viscous enough toresist being swept away by the bloodstream and to remainwithin the submucosal layer when the mucosa is removed.

Measurement of the Thickness of the Submucosal Layer

Sprague-Dawley rats (15–16 weeks, all male with averagebody weight 325 � 15 g; SLC Japan, Shizuoka, Japan) wereused in this study. After overnight fasting, rats were anesthe-tized with pentobarbital (25 mg/kg intravenously), and the

mucosa of the glandular stomach was exposed by sterileanterior gastrostomy after a laparotomy. Az-CH-LA aqueoussolution (0.3 mL) or saline (0.3 mL) was injected into thesubmucosa of the posterior wall via a 25-gauge stainless-steelneedle. After the injection, the mucosal surface was observedfor 30 min, and five rats from each group were sacrificed byan overdose of anesthetic. In other rats, at 30 min after thetreatment 10mL saline was injected into the subcutaneousspace to prevent dehydration, and the stomachs and abdomenwere closed; these rats were sacrificed either at 6 h (n � 5,each group) or at 24 h (n � 5, each group). The tissues werethen removed and fixed in 10% formalin for 2 days. Thethickness of the submucosal layer was measured by macro-scopic observation of cross-sectional views of fixed speci-mens. Specimens were also embedded in paraffin, sectioned,and stained with hematoxylin-eosin (HE) reagent. The distri-bution of the Az-CH-LA within the submucosal space, thedistinctness of its margin, and any alterations in vessels orfibers within the submucosa were observed.

Blood Loss from the Stomach

Heparin (300 units) was injected intravenously before theinjection of 2.5% Az-CH-LA aqueous solution or saline (bythe method described above). The mucosa (5–6 mm in di-ameter) around the top of the elevated area of mucosa wasincised, with a surgical blade, and removed. In the Az-CH-LA (photocrosslinkable chitosan) group, the incisedwound was immediately irradiated with UV light for 30 swith the UV-irradiation system described above. Bleedingfrom the stomach in each rat was absorbed with a surgicalswab, and blood loss was estimated as the increase in theweight of the swab (n � 10, each group). Measurement wasperformed every 5 min until t � 20 min, and the animals werethen killed with an overdose of anesthetic. Stomachs wereresected for histological examination. The removed tissueswere fixed in 10% formalin for 2 days, embedded in paraffin,sectioned, and stained with hematoxylin-eosin (HE) reagent.Notice was taken of whether chitosan hydrogel persisted inthe ulcer base, and of the existence of any ruptured vessels orblood clots.

The animal experiments in this study were approved bythe National Defense Medical College and were carried outfollowing the guidelines it has issued for animal experimen-tation (which are based on the policies and principles estab-lished by the Animal Welfare Act and the NIH Guide for theCare and Use of Laboratory Animals).

Statistics

Results are expressed as mean � S E, and data were analyzedwith a Mann-Whitney U-test with the significance level set atP � 0.05.

RESULTS

Measurement of the Thickness of the Submucosal Layer

After injection of Az-CH-LA solution or saline, the mucosalsurface was elevated with a distinct margin [Figure 1(a,b)];

368 HAYASHI ET AL.

however, the Az-CH-LA–injected wall exhibited a steeperrise. At 30 min, the margin of the elevated area showed assteep a rise as it had just after the injection in the Az-CH-LA–injected group [Figure 1(c)], but in the saline-injectedgroup the rise had become noticeably less steep within 30 min[Figure 1(d)]. The findings made in cross-sectional views ofboth fixed specimens and histological specimens were inaccordance with those made in the macroscopic views of themucosal surface: steep margin in the Az-CH-LA–injectedgroup (as Az-CH-LA is fragile, the central part of it becamedetached in the process of preparation) [Figure 2(a)], and anambiguous border in the saline-injected group [Figure 2(b)].The thickness of the submucosal layer at various time pointsis shown in Table I. At any time point within 24 h, theAz-CH-LA–injected submucosa was significantly thickerthan the saline-injected one. Histological examination re-vealed that Az-CH-LA, which exhibited homogeneous eosin-ophilic staining, was retained within the submucosal layer,and moreover the Az-CH-LA did not contain within it anyvessels or fibrous tissue from the submucosa, these beingdisplaced to the zone around the Az-CH-LA (Figure 3). Noblood was seen at the injection site. In contrast, in the saline-injected wall the submucosal space was almost flat, and bothvessels and fibrous tissue were homogeneously distributedtherein. There was evidence of bleeding at the injection site insome cases [Figure 2(b)].

Even at 24 h after the injection, the margin of the Az-CH-LA mass, which had become reticulated in part, was stillvery distinct in the Az-CH-LA–injected group [Figure 4(a)].The interior of the mass had become cancellous and inhomo-geneous. There was no detectable change in the submucosaltissue other than the above-mentioned displacement of sub-

mucosal tissue around the Az-CH-LA [Figure 4(b)]. Therewas no noteworthy change in the saline-injected tissue.

Blood Loss from the Stomach

In the Az-CH-LA–injected group, there was bleeding fromthe incised edge of the mucosal layer or from the gap betweenthe mucosal layer and the Az-CH-LA (prior to irradiation).Bleeding was almost completely stopped within 5–10 minafter UV-irradiation as the generated chitosan gel effectivelyacted as a glue to seal the opening [Figure 5(a)]. In thesaline-injected wall, the bleeding continued for over 20 minafter the incision.

The blood loss in the first 5 min after mechanical mucosalincision was 63.1 � 14.2 mg in the Az-CH-LA group versus922.9 � 143.6mg in the saline group (p � 0.001) (Figure 6).The cumulative values in the former group were 93.2 � 10.5mg at 10 min, 106.4 � 15.2 mg at 15 min, and 113.0 � 15.5mg at 20 min. The corresponding values in the saline-treatedgroup were 1268.0 � 105.5 mg, 1537.2 � 125.4 mg, and1682.3 � 95.2 mg. At any time point, the value obtained for

Figure 1. Macroscopic views of elevated lesions just after and at 30min after injection. Just after the injection, the margin of the elevationwas distinct in both specimens; however, the Az-CH-LA–injected wall(a, upper left) showed a steeper rise than the saline-injected wall (b,upper right). After 30 min, the elevation still had a distinct margin inthe Az-CH-LA–injected wall (c, lower left), but an unclear margin in thesaline-injected wall (d, lower right).

Figure 2. (a) Low-power view of Az-CH-LA–injected specimen (tis-sues were removed at 30 min after the injection) shows a steep rise atthe margin of the elevation. Az-CH-LA exhibited homogeneous eo-sinophilic staining (asterisk). As Az-CH-LA is fragile, the central part ofit became detached in the process of preparation (double asterisk). (b)Saline-injected wall shows a more moderate rise (lower).

TABLE I. Thickness of the Submucosal Layer

30 min 6 h 24 h

Az-CH-LA–injected wall 3.8 � 0.1* 4.0 � 0.1* 4.1 � 0.1*Saline-injected wall 2.0 � 0.2 1.9 � 0.3 1.8 � 0.3

Note: Values (mm) are mean � SE (n � 5, each group). Az-CH-LA, pho-tocrosslinkable chitosan.

* p � 0.005 versus saline-injected wall

369PHOTOCROSSLINKABLE CHITOSAN

the Az-CH-LA group was significantly smaller than thatobtained for the saline-treated group.

Histological observation of the Az-CH-LA-injected spec-imens revealed that chitosan hydrogel filled the ulcer baseand completely surrounded the bleeding focus [Figure 5(b)].Ruptured vessels were not seen within the gel; however,intact vessels were seen around it. In the saline-injected wall,vessels and various amounts of blood clot were seen at theincision site.

DISCUSSION

Superficial cancers, which are considered to be free fromlymph-node metastasis, can now be detected in the esopha-gus,13 stomach,9–12 and colon14 owing to advances in endos-copy in the field of gastroenterology. For these lesions, en-doscopic resection is now the therapy of first choice.9–11

However, perforation and bleeding are major complicationsof EMR treatment.12 An insufficient mucosal-elevating effectof submucosal saline injection is the major cause of wallperforation,12 and the desired level of mucosal elevation aftersubmucosal injection needs to persist until the snaring of thelesion has been completed. For these reasons, hypertonicsaline with epinephrine15 or 50% glucose16 have been used as

alternatives to normal saline, and these solutions also offerbetter hemostasis.

In the present study, 2.5% Az-CH-LA aqueous solution(not UV irradiated) effectively swelled the submucosal spaceowing to its viscosity. In contrast, saline easily diffuses away,and the submucosa did not maintain the saline-induced mu-cosal elevation. Indeed, at 30 min after the injection thethickness of the submucosal layer injected with Az-CH-LAaqueous solution was almost two times that of the saline-injected one. In general, the EMR procedure is completed

Figure 4. (a) At 24 h after the injection, the Az-CH-LA (asterisk) hadbecome a little reticulated; however, its margin was still distinct. (b)There was no noteworthy change in the submucosa other than thedisplacement of fibrous tissue (arrows) around the Az-CH-LA.

Figure 3. High-power view of Az-CH-LA–injected specimen showsthat the hydrogel (asterisk) displaced the vessels and fibrous tissue(arrows) to the zone around it (at 30 min after injection).

370 HAYASHI ET AL.

within 30 min; however, it sometimes takes 1 h or more. Thedata indicate that injection of Az-CH-LA aqueous solutioncould maintain the desired level of mucosal elevation for aslong as 24 h, and so could easily cope with such cases.

Furthermore, Az-CH-LA aqueous solution with UV-irra-diation revealed an excellent ability to stop gastric bleeding.In the Az-CH-LA–treated (with UV irradiation) group, bloodloss was extremely low (about 7% of that seen in saline-injected controls). Even before UV irradiation, bleeding wasnot prominent from the Az-CH-LA-treated walls, probably aresult of the sealing effect due to the viscous characteristics ofthis material. In addition, in this group vessels in the submu-cosa seemed to evade incision, because histological exami-nation of Az-Ch-LA–injected specimens [Figure 2(a)] re-vealed that vessels were displaced to the zone around theAz-CH-LA. Even if vessels were ruptured, they might becompressed by the Az-CH-LA. After UV irradiation, bleed-ing had almost completely stopped by 10 min because of thecompletion of Az-CH-LA gelation, even though under thepresent circumstances hemostasis would normally be difficult

to obtain due to heparinization. Thus, photocrosslinkablechitosan seems very promising as an injection material forEMR, even in cases in which the patient has a bleedingdisorder or bleeding tendency (e.g., due to liver cirrhosis oranticoagulant medication).

As for wound healing, conventional chitosan has beenshown not only to reduce ethanol-induced gastric mucosalinjury, but also to accelerate the healing of acetic-acid–induced gastric ulcers in rats.17 In addition, chitosan hasantibacterial and antifungal properties.18 The photocrosslink-able-chitosan hydrogel enhances wound contraction, and ac-celerates wound closure and healing7,8 in the skin on a rat’sback. If the promotion of ulcer healing by photocrosslink-able-chitosan hydrogel treatment after EMR can be verified,it should be very valuable for patients, especially in casescomplicated by such factors as advanced age or concomitantdisease. Indeed, its use could lead to a shortening of hospi-talization time. Animal studies to evaluate the ability ofphotocrosslinkable chitosan hydrogel to promote ulcer heal-ing, especially in artificial ulcers after EMR, are currently inprogress.

So far, Az-CH-LA aqueous solution and its hydrogel,which is formed after UV irradiation, have shown no evi-dence of cytotoxicity in cell-culture tests on human skinfibroblasts, human endothelial cells, or human smooth musclecells.5,6 In addition, tests of toxicity toward organisms, in-cluding those for mutagenicity and cytotoxicity, have dem-onstrated the safety of both Az-CH-LA aqueous solution andits hydrogel. In the present study, there was no noteworthychange in the submucosa at 24 h after the Az-CH-LA injec-tion. These results suggest that chitosan hydrogel is safe formedical application, although more detailed toxicity tests,with appropriate animal models, remain to be carried out. Inaddition, an endoscopic system with glass-fiber UV illumi-nation is currently under development for use in EMR.

In conclusion, the Az-CH-LA aqueous solution and itshydrogel, which is formed following UV irradiation, areeffective (a) at achieving the desired degree of mucosalelevation for lesion removal by EMR (after its submucosalinjection), and (b) at significantly reducing bleeding (with theaid of UV irradiation) after mucosal resection. Hence, the

Figure 6. Cumulative blood loss from the incised wound in eachgroup (n � 10, each group). Upper curve, saline-injected wall; lowercurve, Az-Ch-LA–injected wall. *p � 0.001 versus saline-injected wall.

Figure 5. (a) After UV-light irradiation, chitosan hydrogel covered theopening like a glue (asterisk). (b) Histology shows that hydrogel (as-terisk) completely filled the ulcer base and surrounded the bleedingfocus (double asterisk). Ruptured vessels were not seen within thegel; however, intact vessels were seen around it (arrows).

371PHOTOCROSSLINKABLE CHITOSAN

photocrosslinkable chitosan developed is promising as a sub-mucosal injection agent for EMR.

REFERENCES

1. Chandy T, Sharma CP. Chitosan—as a biomaterial. BiomaterArtif Cells Artif Organs 1990;18:1–24

2. Shigemasa Y, Minami S. Application of chitin and chitosan forbiomaterials. Biotechnol Gen Eng Rev 1995;13:383–420.

3. Fukasawa M, Abe H, Masaoka T, Orita H, Horikawa H, Cam-peau JD, Wahio M. The hemostatic effect of deacetylated chitinmembrane on peritoneal injury in rabbit model. Surg Today1992;22:333–338

4. Nishimura K, Nishimura S, Nishi N, Sakai I, Tokura S, AzumaI. Immunological activity of chitin and its derivatives. Vaccine1984;2:93–99

5. Ono K, Saito Y, Yura H, Isihara K, Kurita A, Ishihara M.Photocrosslinkable chitosan as a biological adhesive. J BiomedMater Res 2000;49:289–295.

6. Ono K, Ishihara M, Ozeki Y, Deguchi H, Sato M, Saito Y, YuraH, Sato M, Kikuchi M, Kurita A, Maehara T. Experimentalevaluation of photocrosslinkable chitosan as a biologic adhesivewith surgical applications. Surgery 2001;130:844–850

7. Ishihara M, Nakanishi K, Ono K, Sato M, Kikuchi M, Saito Y,Yura H, Matsui T, Hattori H, Uenoyama M, Kurita A. Pho-tocrosslinkable chitosan as a dressing for wound occlusion andaccelerator in healing process. Biomaterials 2002;23:833–840.

8. Ishihara M, Ono K, Sato M, Nakanishi K, Saito Y, Yura H,Matsui T, Hattori H, Fujita M, Kikuchi M, Kurita A. Acceler-ation of wound contraction and healing with a photocrosslink-able chitosan hydrogel. Wound Repair Regen 2001;9:513–521.

9. Takekoshi T, Baba Y, Ota H, Kato Y, Yanagisawa A, Takagi K,Noguchi Y. Endoscopic resection of early gastric carcinoma:results of a retrospective analysis of 308 cases. Endoscopy1994;26:352–358.

10. Inoue H et al. Endoscopic mucosal resection with a cap-fittedpanendoscope for esophagus, stomach, and colon mucosal le-sions. Gastrointest Endosc 1993;39:58–62.

11. Ono H, Kondo H, Gotoda T, Shirao K, Yamaguchi H, Saito D,Hosokawa K, Shimoda T, Yoshida S. Endoscopic mucosalresection for treatment of early gastric cancer. Gut 2001;48:225–229.

12. Inoue H, Kawano T, Tani M, Takeshita K, Iwai T. Endoscopicmucosal resection using a cap: techniques for use and prevent-ing perforation. Can J Gastroenterol 1999;13:477–480.

13. Inoue H. Endoscopic mucosal resection for the entire gastroin-testinal mucosal lesions. Gastrointest Endosc Clin N Am 2001;11:459–478.

14. Kudo S, Tamegai Y, Yamano H, Imai Y, Kogure E, Kashida H.Endoscopic mucosal resection of the colon: the Japanese tech-nique. Gastrointest Endosc Clin N Am 2001;11:519–535.

15. Hirao M, Masuda K, Asanuma T, Naka H, Noda K, Matsuura K,Yamaguchi O, Ueda N. Endoscopic resection of early gastriccancer and other tumors with local injection of hypertonicsaline-epinephrine. Gastrointest Endosc 1988;34:264–269.

16. Makuuchi H, Mitomi T, Machimura T, Mizutani S, Shimada H,Sugano K. Endoscopic mucosal resection for early esophagealcancer by EEMR-tube method [in Japanese with English ab-stract]. Stomach Intest 1988;28:153–159.

17. Ito M, Ban A, Ishihara M. Anti-ulcer effects of chitin andchitosan, healthy foods, in rats. Jpn J Pharmacol 2000;82:218–225

18. Yun YS. Kim KS, Lee YN. Antibacterial and antifungal effectof chitosan. J Chitin Chitosan 1999;4:8–14.

372 HAYASHI ET AL.