hyaluronan hydrogel: an appropriate three-dimensional model for evaluation of anticancer drug...

Available online at www.sciencedirect.com

Acta Biomaterialia 4 (2008) 256–263

www.elsevier.com/locate/actabiomat

Hyaluronan hydrogel: An appropriate three-dimensional modelfor evaluation of anticancer drug sensitivity

Laurent David a,*, Virginie Dulong b, Didier Le Cerf b, Lionel Cazin a,Marek Lamacz a, Jean-Pierre Vannier a

a Groupe de Recherche sur le Micro-Environnement et le Renouvellement Cellulaire Integre (MERCI, UPRES EA 3829), Faculte de Medecine Pharmacie,

Universite de Rouen, 22 boulevard Gambetta, 76183 Rouen, Franceb Laboratoire PBM – UMR 6522, CNRS, Universite de Rouen, 76821 Mont Saint-Aignan, France

Received 30 May 2007; received in revised form 18 August 2007; accepted 20 August 2007Available online 11 September 2007

Abstract

The extracellular polysaccharide hyaluronan (HA) controls cell migration, differentiation and proliferation, and contributes to theinvasiveness of human cancers. In order to investigate the sensitivity of cancer cells to antimitotic agents, we developed a cross-linkedHA hydrogel, a three-dimensional matrix in which cells can invade and grow. We have studied three cell lines (SA87, NCI-H460 andH460M), from primary tumors and metastases, that migrated into the HA hydrogel and proliferated giving rise to clusters and colonies.Concurrently, we studied the growth of these cell lines in a usual monolayer culture system. In these two models, increasing concentra-tions of doxorubicin and 5-fluorouracil were evaluated for their ability to inhibit tumor cell growth and colony formation. Takentogether, our data suggest that the cancer cells were more resistant in the three-dimensional model than in monolayer cell systems.The antimitotic drugs were efficient after 24 h of treatment in the monolayer cultures, whereas they were significantly efficient only afterone week of incubation in the HA hydrogels. Herein, we show that this cross-linked matrix provides a three-dimensional model partic-ularly appropriate for investigating mechanisms involved in cancer cell line sensitivity to antimitotic drugs.� 2007 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Hyaluronan; Hydrogel; Cell growth; Cytotoxicity

1. Introduction

Multidrug resistance in cancer cells arises by severalmechanisms, among which are activated repair and detoxi-fying systems [1]. Two-dimensional (2D) cell culture modelsof cancer cell lines are usually employed to assess the efficacyof new anticancer drugs. These monolayer assays determinethe ability of a treatment to inhibit cell growth or the capac-ity to trigger apoptosis. However, in vivo proliferation ofcancer cells in a three-dimensional (3D) structure modifiestheir invasive and proliferative properties [2]. Recently, theintroduction of 3D culture systems gave rise to in vitro stud-ies on cell–extracellular matrix (ECM) interactions favoring

1742-7061/$ - see front matter � 2007 Acta Materialia Inc. Published by Else

doi:10.1016/j.actbio.2007.08.012

* Corresponding author. Tel.: +33 2 35 14 83 50; fax: +33 2 35 14 83 40.E-mail address: [email protected] (L. David).

tumor invasiveness [3]. Since the behavior of tumor cells isdirectly conditioned by cell contacts with ECM, 3D culturemethods appeared to be appropriate for selective evalua-tions of the effects provoked by ECM components [4,5].Hyaluronan (HA), a glycosaminoglycan (GAG) composedof a repeating disaccharide of glucuronic acid and N-acetyl-glucosamine (b1,4-GlcUA-b1,3-GlcNAc-)n, is one of themain ECM constituents [6]. HA was found to be implicatedin tumor cell behavior [7]. A high content of intracellularHA and its accumulation in the ECM was shown to be ableto create a microenvironment favorable for migration, pro-liferation and invasiveness of malignant cells [8,9].

The 2D model of in vitro growth of cancer cells has fre-quently been employed to determine the antimitotic effectsof different drugs. In these types of culture, malignant cellsare only able to proliferate and to increase their number. In

vier Ltd. All rights reserved.

mailto:[email protected]

L. David et al. / Acta Biomaterialia 4 (2008) 256–263 257

contrast, it was reported recently that in 3D cultures, cre-ated in different cross-linked artificial matrices, the cancercells adopt a behavior similar to what is observed in naturalendogenous matrix [2]. Cells are able to form colonies and,subsequently, secondary tumor-like structures situated inthe migratory front, and mimic, in vitro, the characteristicsof in vivo features. For this reason, we compared the evo-lution of malignant cell lines in 2D and 3D culture modelsin order to select optimal ex vivo conditions to better eval-uate the efficacy of anti-proliferative drugs.

2. Experimental procedures

2.1. Synthesis of hyaluronan hydrogels

The HA hydrogels consisted of a long-chain of HAcross-linked with adipic dihydrazide (ADH) (Sigma, SaintQuentin Fallavier, France) as cross-linking agent and 1-ethyl-3-[3-(dimethylamino)-propyl]carbodiimide (EDCI)(Sigma) as reagent. All hydrogels were prepared from highmolecular weight (>1 · 106 Da) HA (Sigma), similar to theprocedure described by Prestwich et al. [10]. Briefly, theratios of ADH to HA and HA to EDCI were adjusted toobtain hydrogels optimized for cell adhesion and culture.The best results were obtained with ratios of ADH toHA equal to 10:1 and ratios of HA to EDCI equal to1:1. HA and hydrazide cross-linker (ADH) were dissolvedin Milli-Q water and the pH adjusted to 4 by adding 0.1 NHCl. The carbodiimide reagent (EDCl) was dissolved inMilli-Q water and added to the mixture, and allowed togel for 2 h with gently agitation. HA hydrogels were equil-ibrated in 0.1 N NaCl for 2 days, then in a water/ethanolmixture (3/1, v/v) for 2 days, and in Milli-Q water for 2days to remove ADH. HA hydrogels were lyophilized, thenstored. Before use for cell culture, squares of the HAhydrogels were cut (to approximately 5 · 5 · 1.5 mm) thensterilized (100 �C for 30 min) and rehydrated with culturemedium and swelled roughly 100-fold in mass. Based onthe ratios and the degree of ADH substitution of theHA–ADH used in this study (55%), the mesh size is about100 nm.

2.2. Cancer-derived cell lines culture conditions

The lung carcinoma cell line H460M [11] was a gift fromDr J.P. Martin (Inserm U295, Rouen, France) and NCI-H460 were obtained from the ATCC (Rockville, MD).The cell line SA87 was derived from a brain metastasisdeveloped in a patient with a gastric adenocarcinoma[12]. All cultures were made in RPMI-1640 medium with2 mM glutamine (Bio-Whittaker) and 10% fetal calf serum(Eurobio, Les Ulis, France) in humidified atmosphereunder 5% CO2 in air at 37 �C. Cell proliferations were car-ried out in 75 cm2 Falcon flasks. The medium was changedtwice weekly, and cells were trypsinized at confluence with0.05% trypsin 0.02% EDTA solution in Ca2+- and Mg2+-free balanced salt solution (Bio-Whittaker).

2.3. Invasiveness of cell lines in hyaluronan hydrogel and

antimitotic treatment

Cancer cells were trypsinized, centrifuged at 1500 rpmfor 10 min and the pellet was resuspended in culture med-ium to disperse cells for counting. 106 cells were seeded in awell containing a HA hydrogel floating in culture medium(2 ml RPMI 10% FCS). Cells were allowed to migrate andcolonize the HA hydrogel for 24 h in humidified atmo-sphere under 5% CO2 in air at 37 �C. Then the HA hydro-gel was transposed in a new well with fresh culture mediumand incubated in a humidified atmosphere under 5% CO2

in air at 37 �C for 5 days. After these 5 days, the growthmedium was completed with 5-fluorouracil (5-FU, between1 mM and 1 nM) or doxorubicin (between 100 lM and1 nM). The HA hydrogels were examined daily using aninverted microscope for 2 weeks. Colony counting was per-formed within the whole gel by a computerized morpho-metric system using the Cyberview 3.0 program (CervusInternational), an inverted microscope and a CCD camera.The program was calibrated for each magnification. Themeasurements were performed by tracing the diameter ofcolonies on digitized images. The progressive colonizationby malignant cells giving rise to clusters (<150 lm diame-ter) and colonies (>150 lm diameter), only the coloniesof more than 150 lm diameter were scored. Therefore,the invasiveness of cancer cell lines was estimated fromthe ability of cells to invade reconstituted HA matrix.The IC50 values for the antimitotic agents (5-FU or doxo-rubicin) were determined from curves with a non-linearregression analysis using the equation for a sigmoid plot.

2.4. Crystal violet assay

Cells were plated at a density of 100,000 cells per well in24-well plates. The growth medium was completed with 5-FU (between 1 mM and 1 nM) or doxorubicin (between100 lM and 1 nM) and with soluble HA (0, 10 or100 lg ml�1). After 1 or 3 days, the cells were fixed withacetic alcohol (ethanol/acetic acid/water, 95/3/2, v/v/v).The number of the cells in each well was determined usingthe crystal violet (CV) assay. Briefly, the cells were incu-bated in 0.1% CV for 30 min at room temperature, excessdye was removed by three brief rinses with ddH2O, theplates were air dried, and the dye was extracted with 10%acetic acid, which was then read using the Dynex Revela-tion program in a plate reader (Dynatech Laboratories)at 570 nm. The IC50 values for the antimitotic agents (5-FU or doxorubicin) were determined from curves with anon-linear regression analysis using the equation for a sig-moid plot.

2.5. Statistical analysis

To determine the significance of the observed differencesin cell cultures, the Mann–Whitney U test was employed,assuming unequal variance and two-tailed distribution.

258 L. David et al. / Acta Biomaterialia 4 (2008) 256–263

All results were reported as mean ± standard error mean(SEM). Differences were considered statistically significantat P < 0.05.

3. Results

The results presented in this work were obtained fromthree cancer cell lines (SA87, NCI-H460 and H460 M) cul-tured according to the above-described protocols. Eachcancer cell line was cultured in the 3D and the 2D methods,in the absence and in the presence of 5-FU or doxorubicinat concentrations from 1 nM to 1 mM and 1 nM to 1 lM,respectively. The 3D model is composed of a matrix ofcross-linked HA, while the 2D system contains solubleHA (0, 10 or 100 lg ml�1).

3.1. Effect of anticancer drugs on 3D model

In the 3D culture model, the three cell lines were able tomigrate into the HA hydrogel (to a depth of 500 lm).These cells have shown their ability to form colonies witha homogeneous distribution in the hydrogel and not toremain at its periphery (Fig. 1A). The colonies werecounted over 14 day periods. Within the control HAhydrogel, these cells formed numerous, large and irregularcolonies as a primary tumor (Fig. 1A and B), and thenmigrated in a malignant-like manner from the hydrogel,and subsequently, formed 2D colonies at the bottom ofthe wells. These 2D colonies, which were spheroid-like in

Fig. 1. Cancer cell lines in the HA hydrogels. (A) General aspect of the distribthe 5th day after migration. Magnification · 250. (B) General aspect of thehydrogels, shown at the 5th day after migration. Magnification · 100. (C) Gewells, shown at the 5th day after treatment. Magnification · 100. (D) General ashown at the 5th day after migration. Magnification · 100. A representative e

shape and became more spherical as they grew, were clearlydifferent from those observed to the HA hydrogel (Fig. 1Cand D).

3.1.1. Effect of anticancer drugs on hydrogel colonization

As shown in Fig. 2, in the 3D culture model, treatmentwith 5-FU or doxorubicin gradually reduced the number ofcolonies in the hydrogel phase, in comparison to untreatedcontrols. This effect is observed for the three cell lines, andis dose dependent. For 5-FU concentrations >10 lM(SA87 and H460M, P < 0.01) or >1 lM (NCI-H460,P < 0.01) and concentrations of doxorubicin >100 nM(NCI-H460 and H460 M, P < 0.01) or >10 nM (SA87,P < 0.01), a progressive and significant decrease in the col-onies is observed in the hydrogel phase after the 5th day oftreatment. After 14 days of culture, hydrogel phase colo-nies have nearly vanished for high doses of 5-FU and doxo-rubicin (1 mM and 100 lM, respectively).

3.1.2. Effect of anticancer drugs on 2D colonies

Concurrently in the wells, for the three cell lines, 2D col-onies were progressively and temporarily observed at thebottom of the wells, but only for low doses of 5-FU(<1 lM) or doxorubicin (<100 nM) (Table 1). These colo-nies were then progressively eliminated and the delay inclearance was inversely related to the antimitotic concen-tration. The presented results are obtained with the SA87cell lines, but the same results have been obtained withthe other cell lines. In all cases, the disappearance of the

ution of lung carcinoma NCI-H460 colonies within the hydrogel, shown atspreading and growth of gastric adenocarcinoma SA87 cells in the HA

neral aspect of lung carcinoma H460M 2D colonies on the bottom of thespect of lung carcinoma NCI-H460 2D colonies on the bottom of the wells,xperiment is shown for each cell type. Bars = 150 lm.

Fig. 2. Antimitotic agent effect on cancer cell invasiveness in the HA hydrogels. Cancer cells from SA87, NCI-H460 or H460M cell lines (106 per well)were incubated with a HA hydrogel floating in culture medium (RPMI 10% FCS) at 37 �C and 5% CO2 for 24 h. Following incubation, the HA hydrogelswere washed to remove the non-invasive cells and were incubated at 37 �C and 5% CO2 for 5 days in RPMI 10% FCS. The number of colonies of morethan 150 lm diameter was determined for each hydrogel at day 5 using an inverted microscope. After these 5 days, the growth medium was completed with5-fluorouracil (between 1 mM and 1 nM) or doxorubicin (between 100 lM and 1 nM). The HA hydrogels were examined daily using an invertedmicroscope for 2 weeks. The results are the number of colonies per hydrogel as the mean ± SEM of values obtained from five experiments.


colonies is faster at the bottom of the wells than in thehydrogels.

3.2. Effect of anticancer drugs on cell growth

In these experiments, cells were counted after 24 or 72 hof culture. In the absence of antimitotics, soluble HA (10and 100 lg ml�1) had a significant stimulatory effect onthe SA87 line after 24 h (P = 0.0022) but not on the othertwo cell lines (Figs. 3 and 4).

After 24 h of treatment, the H460M cell line was themost sensitive to 5-FU (Fig. 3). Indeed, we observed adecrease in the cell growth starting 24 h after the beginningof incubation for doses P100 lM. The addition of HA didnot modify the cell drug resistance. Doxorubicin reducedthe cell number for doses P1 lM. In contrast, for the othertwo cell lines, we observed a significant increase in the cellnumbers for 5-FU concentrations ranging between 1 lMand 1 nM. This increase was more important in the pres-ence of soluble HA. Doxorubicin reduced the cell numberfor doses P100 nM. The addition of HA did not modifythe cell drug sensitivity (Fig. 4).

After 72 h of treatment, the cell number decreased for 5-FU doses as low as 1 lM (Fig. 3). Moreover, the presenceof soluble HA (100 lg ml�1) improved the cell resistanceto 5-FU at low concentrations (1 nM-1 lM, P < 0.05)

(Fig. 3). After 72 h of treatment, doxorubicin reduced thecell growth at each concentration, and the addition ofHA did not modify the drug sensitivity (Fig. 4).

3.3. IC50 comparison between 2D and 3D assays

The IC50 concentration for the antimitotic agents (doxo-rubicin or 5-FU) in monolayer cell growth and in 3Dexperiments was determined as the concentration requiredto reduce the cell number (2D) or the colony number (3D)by 50%. The IC50s for doxorubicin and 5-FU were higherin the 3D cultures than in the 2D cultures (Table 2). Fur-thermore, for the 3D assays, the IC50 of the two drugs werehigher for the colonies within the hydrogel than for the 2Dcolonies at the bottom of the wells.

4. Discussion

It is unanimously recognized that the tissue microenvi-ronment is of crucial importance to metastatic tumor pro-gression. In the natural 3D environment, the malignantcells interact not only with the stromal cells or cross-linkedproteins but are also exposed to the action of several auto-crine and paracrine factors, such as cytokines or proteo-lytic enzymes [5]. In this way, the fate of cancer evolutionis conditioned by complex interactions that are difficult

Table 1Antimitotic agent effect on cancer cell 2D colony formation at the bottomof the wells

Control d6 d7 d10 d12 d14

A + 2+ 2+ 2+ 2+1 nM + 2+ 2+ 2+ +10 nM + 2+ + 0 0100 nM + + 0 0 01 lM + + 0 0 010 lM + + 0 0 0100 lM + 0 0 0 01 mM + 0 0 0 0

d6 d7 d10 d12 d14B + 2+ 2+ 2+ 2+1 nM + 2+ + 0 010 nM + + + 0 0100 nM + + + 0 01 lM + + 0 0 010 lM + + 0 0 0100 lM + 0 0 0 0

Cancer cells from SA87 cell lines (106 per well) were incubated with a HAhydrogel floating in culture medium (RPMI 10% FCS) at 37 �C and 5%CO2 for 24 h. Following incubation, the HA hydrogels were washed toremove the non-invasive cells and were incubated at 37 �C and 5% CO2 for5 days in RPMI 10% FCS. After these 5 days, the growth medium wascompleted with 5-fluorouracil (between 1 mM and 1 nM, part A) ordoxorubicin (between 100 lM and 1 nM, part B). The bottom of the wellswas examined using an inverted microscope and the presence of 2D col-onies was determined daily for each well using an inverted microscope.The presence of 2D colonies was quantified using the following scoringsystem: 0, no 2D colonies; +, occasional 2D colonies; 2+, numerous andextensive 2D colonies.


to reproduce in ex vivo experiments. It is particularly rele-vant in 2D monolayer cultures in which cells are devoid ofnot only the third dimension in which to proliferate andorganize themselves but also of the fundamental survivalsignals that exist in endogenous ECM. Recent progress intissue engineering has revealed the suitability of a poly-meric 3D scaffold for ex vivo cell growth [13,14]. Recently,we have shown that cross-linked synthetic HA matrix ishighly appropriate for the study of cancer cell invasiveness[15]. In fact, this new 3D scaffold offers the cells a hydro-philic matrix that possesses physical properties very similarto those of biological tissue. Moreover, the accumulationof HA in ECM was found to create a microenvironmentfavoring proliferation and invasiveness of malignant cells[9,16]. This hydrogel appears to possess the required qual-ities to allow transposition of in vivo to in vitro conditions.Thus, the HA hydrogel offers an excellent support in whichthe tumor cells proliferate and migrate rapidly in all direc-tions. In the present work, we observed the evolution of cellinvasiveness in the presence of anticancer drugs and wecompared the cellular sensitivity in both 2D and 3D modelsin order to appreciate the efficacy of anticancer drugsin vitro. In this context, HA binds cell surface receptorssuch as CD44 and RHAMM that transduce the intracellu-lar signals influencing cell morphology and behavior[17,18].

Herein, we report that the 3D culture system allows thestudy of two distinct properties of malignant cells: (i) theformation of colonies in HA hydrogel which correspondsto intratumoral cell expansion (primary tumor), and (ii)the outgrowth of cells, migrating from the primary tumorto the liquid culture medium and the plastic bottom ofthe wells (secondary tumor-like structures). The expansionof cancer cells embedded in monolayers was found to belimited, while in the 3D culture system their growthappeared to be markedly facilitated. The main finding ofthe present work was that the HA 3D system appearedto be an appropriate model to evaluate the efficiency ofcytotoxic drugs on the growth and invasion of tumor cells.

A comparison of the kinetics of cell growth between 2Dand 3D models revealed that ECM plays a determining rolein cancer cell protection. Cytostatic effects were observedearlier in 2D cultures in which the cells were practicallyabsent after a 72 h period of incubation. In contrast, in3D cultures, a similar anti-proliferative effect could beobserved only after 14 days of incubation with the highestconcentrations of doxorubicin or 5-FU tested in the pres-ent study. As indicated by the IC50 coefficients, the concen-trations of both anti-metabolites able to induce anti-proliferative effects had to be higher in the 3D than inthe 2D system. These findings are in a good agreement withresults previously reported indicating that the 3D microen-vironment facilitates drug resistance [19]. These observa-tions confirm studies that activation of CD44 andRHAMM receptors is responsible for the mobilization ofanti-apoptotic pathways as reported in colon cancer cells[20]. Altogether, these observations corroborate data sug-gesting that HA is able to increase the cancer cell survivalthrough PI3K- and MAPK-dependent stimulation ofmulti-drug-resistant proteins [21]. The different effects ofsoluble HA on cells after 24 vs. 72 h may be due to degra-dation of HA by the hyaluronidase secreting cells, whichcould increase the protective effect of HA with the productof small HA oligosaccharides. Moreover, St Croix et al.postulated that high concentrations of hyaluronidase,which were found to provoke the perturbation of HA–tumor cell interactions, can be considered as an apparentchemosensitizer as they have the potential to increase theefficiency of chemotherapeutic drugs used in clinical oncol-ogy treatment protocols [22].

Our results might appear to be affected by the diffusionof chemotherapy drugs into hydrogels. Nevertheless, asdescribed by Luo et al., the lower molecular weight mole-cules (Mw < 4000) can freely diffuse within the HA hydro-gel. However, higher molecular weight molecules(Mw > 70,000) cannot freely diffuse; diffusion thus occursonly as a result of enzymatic degradation [23]. The antican-cer drugs used in this study were low molecular weight mol-ecules (5-FU, Mw = 130.8; doxorubicin, Mw = 580). Then,the drug resistance observed in our experiments is a directconsequence of the conformation of the cancer cells in thehydrogel and is not due to a slow diffusion of antimitoticagents.

Fig. 3. 5-Fluorouracil effect on growth of cancer cells in a monolayer culture system. SA87, NCI-H460 or H460 M cell lines were plated at a density of100,000 cells per well in 24-well plates. The growth medium was completed with 5-fluorouracil (between 1 mM and 1 nM) and with soluble HA (0, 10 or100 lg ml�1). After 1 or 3 days, the cells were fixed with acetic alcohol (ethanol/acetic acid/water, 95/3/2, v/v/v). The number of the cells in each well wasdetermined using the crystal violet assay. The results are the number of colonies per hydrogel as the mean ± SEM of values obtained from fiveexperiments. Statistically significant differences from control were determined by the Mann–Whitney U-test. *HA 100 lg ml�1 vs. control, P < 0.05.

Fig. 4. Doxorubicin effect on growth of cancer cells in a monolayer culture system. SA87, NCI-H460 or H460M cell lines were plated at a density of100,000 cells per well in 24-well plates. The growth medium was completed with doxorubicin (between 100 lM and 1 nM) and with soluble HA (0, 10 or100 lg ml�1). After 1 or 3 days, the cells were fixed with acetic alcohol (ethanol/acetic acid/water, 95/3/2, v/v/v). The number of the cells in each well wasdetermined using the crystal violet assay. The results are the number of colonies per hydrogel as the mean ± SEM of values obtained from fiveexperiments.


Table 2Determination of IC50 values for the inhibition of SA87, NCI-H460 and H460M cell growth by 5-fluorouracil and doxorubicin in 2D and 3D assays

IC50 (5-FU) IC50 (Doxo.)

SA87 3D Colonies in hydrogel 8 lM 50 nMColonies on well bottom 5 lM 50 nM

2D (Cell growth) 4 lM 40 nM

NCI-H460 3D Colonies in hydrogel 5 lM 50 nMColonies on well bottom 3 lm 40 nM

2D (Cell growth) 1 lM 30 nM

H460M 3D Colonies in hydrogel 6 lM 40 nMColonies on well bottom 5 lm 10 nM

2D (Cell growth) 1 M 7 nM

Assays were performed as outlined in Section 2. The IC50 concentration for the antimitotic agents (5-FU or doxorubicin) in 3D colonies (in hydrogel) and2D colonies (at the bottom of the wells) and 2D (cell growth) experiments is the concentration required to inhibit of the number of cells (2D) or the numberof colonies (3D) by 50%. IC50 values were determined from curves with a non-linear regression analysis using the equation for a sigmoid plot.


Even if the cancer may be qualified as a truly 3D disease,many investigations have routinely focused on single-cellmechanisms of drug resistance [21]. Monolayer cell culturesystems do not provide evidence that the cell-to-cell andcell-to-ECM interactions modulate key processes impli-cated in the architecture of normal and cancer tissues.The HA hydrogel therefore provides a valid in vitro modelfor tumor cell behavior that reproduces salient features ofthe natural tumor microenvironment. In addition, the per-turbation of HA interactions by hyaluronidase or HA olig-omers may be very interesting tools for inhibiting tumorgrowth and metastasis, as well as sensitizing cancer cellsto antimitotic agents. Study of those events should attestthe predicted efficacy of therapeutic strategies against thetumor microenvironment.

In conclusion, our 3D model may be considered as anattempt to draw up an ‘‘antimitogram’’ by which drugscould be evaluated in their ability to eradicate chemoresis-tant tumor cells. For the time being, use of this HA-basedsystem may be limited to secreting cancer epithelial cellssuch as colon, bronchi and gastric cancer cells, but experi-ments in progress using a novel hydrogel enriched withother ECM components will allow better understandingof the behavior of non-secreting cancer cells in tumor inva-sion and in anticancer drug sensitivity.

Acknowledgments

L.D. is a recipient of a fellowship from the ConseilRegional de Haute Normandie. This study was supportedby grants from la region Haute-Normandie, le reseau MII-AT ‘Matrice Extracellulaire Normale et Pathologique’ and‘l’Association Vie et Espoir’.

References

[1] Filipits M. Mechanisms of cancer: multidrug resistance. Drug DiscovToday: Dis Mech 2004;1:229–34.

[2] Doillon CJ, Gagnon E, Paradis R, Koutsilieris M. Three-dimensionalculture system as a model for studying cancer cell invasion capacityand anticancer drug sensitivity. Anticancer Res 2004;24:2169–77.

[3] Hotary KB, Allen ED, Brooks PC, Datta NS, Long MW, Weiss SJ.Membrane type I matrix metalloproteinase usurps tumor growthcontrol imposed by the three-dimensional extracellular matrix. Cell2003;114:33–45.

[4] Condon MS. The role of the stromal microenvironment in prostatecancer. Semin Cancer Biol 2005;15:132–7.

[5] Egeblad M, Werb Z. New functions for the matrix metalloproteinasesin cancer progression. Nat Rev Cancer 2002;2:161–74.

[6] Tammi MI, Day AJ, Turley EA. Hyaluronan and homeostasis: abalancing act. J Biol Chem 2002;277:4581–4.

[7] Stern R. Hyaluronan metabolism: a major paradox in cancer biology.Pathol-Biol 2005;53:372–82.

[8] Vizoso FJ, del Casar JM, Corte MD, Garcia I, Corte MG, Alvarez A,et al. Significance of cytosolic hyaluronan levels in gastric cancer. EurJ Surg Oncol 2004;30:318–24.

[9] Kim HR, Wheeler MA, Wilson CM, Iida J, Eng D, Simpson MA,et al. Hyaluronan facilitates invasion of colon carcinoma cellsin vitro via interaction with CD44. Cancer Res 2004;64:4569–76.

[10] Prestwich GD, Marecak DM, Marecek JF, Vercruysse KP, ZiebellMR. Controlled chemical modification of hyaluronic acid: synthesis,applications, and biodegradation of hydrazide derivatives. J ControlRelease 1998;53:93–103.

[11] Corti C, Pratesi G, DeCesare M, Pellegrini R, Giardini R, Supino R,et al. Spontaneous lung metastases in a human lung tumor xenograft:a new experimental model. J Cancer Res Clin Oncol 1996;122:154–60.

[12] Maunoury R. Establishment and characterization of 3 human celllines derived from intracerebral metastatic tumors. C R Acad SciHebd Seances Acad Sci D 1977;284:991–4.

[13] Grayson WL, Ma T, Bunnell B. Human mesenchymal stem cells tissuedevelopment in 3D PET matrices. Biotechnol Prog 2004;20:905–12.

[14] Liu H, Roy K. Biomimetic three-dimensional cultures significantlyincrease hematopoietic differentiation efficacy of embryonic stemcells. Tissue Eng 2005;11:319–30.

[15] David L, Dulong V, Le Cerf D, Chauzy C, Norris V, Delpech B,et al. Reticulated hyaluronan hydrogels: a model for examiningcancer cell invasion in 3D. Matrix Biol 2004;23:183–93.

[16] Laurich C, Wheeler MA, Iida J, Neudauer CL, McCarthy JB, BullardKM. Hyaluronan mediates adhesion of metastatic colon carcinomacells. J Surg Res 2004;122:70–4.

[17] Murai T, Miyazaki Y, Nishinakamura H, Sugahara KN, Miyauchi T,Sako Y, et al. Engagement of CD44 promotes Rac activation andCD44 cleavage during tumor cell migration. J Biol Chem2004;279:4541–50.

[18] Rein D, Roehrig K, Schondorf T, Lazar A, Fleisch M, NiederacherD, et al. Expression of the hyaluronan receptor RHAMM inendometrial carcinomas suggests a role in tumour progression andmetastasis. J Cancer Res Clin Oncol 2003;129:161–4.

[19] Morin PJ. Drug resistance and the microenvironment: nature andnurture. Drug Resist Updates 2003;6:169–72.


[20] Lakshman M, Subramaniam V, Rubenthiran U, Jothy S. CD44promotes resistance to apoptosis in human colon cancer cells. ExpMol Pathol 2004;77:18–25.

[21] Misra S, Ghatak S, Zoltan-Jones A, Toole BP. Regulation ofmulti-drug resistance in cancer cells by hyaluronan. J Biol Chem2003;8:8.

[22] St Croix B, Man S, Kerbel RS. Reversal of intrinsic and acquiredforms of drug resistance by hyaluronidase treatment of solid tumors.Cancer Lett 1998;131:35–44.

[23] Luo Y, Kirker KR, Prestwich GD. Cross-linked hyaluronic acidhydrogel films: new biomaterials for drug delivery. J Control Release2000;69:169–84.

hyaluronan hydrogel: an appropriate three-dimensional model for evaluation of anticancer drug...

Documents