Capsules containing mammalian cells

Abstract

A capsule containing at least one mammalian cell, includes a liquid core, and at least one external envelope totally encapsulating the liquid core at its periphery, the external envelope including at least one gelled polyelectrolyte and/or a stiffened biopolymer and being able to retain the liquid core when the capsule is immersed in a gas. The present invention further relates to the method for preparing such a capsule, to a method for screening cosmetic active ingredients as well as a culture method using such capsules.

Claims

1. A microcapsule comprising: a liquid core; a stiff intermediate envelope comprising at least one biopolymer; and at least one external envelope totally encapsulating the liquid core at its periphery, said intermediate envelope being located between the liquid core and the external envelope, said external envelope being able to retain the liquid core when the microcapsule is immersed in a gas and comprising at least one gelled polyelectrolyte and/or one stiffened biopolymer, said microcapsule further comprising at least one eukaryotic mammalian cell, wherein the liquid core comprises at least one keratinocyte and the intermediate envelope comprises at least one fibroblast; and said stiff intermediate envelope having an elastic modulus that is non-zero.

2. The microcapsule according to claim 1, wherein the biopolymer of said intermediate envelope is selected from the group consisting of proteins of the extra-cellular matrix, proteoglycans, glycosaminoglycans, polysaccharides, and non-hydrolysed or partly hydrolysed form thereof.

3. The microcapsule according to claim 1, wherein the microcapsule is obtained by a method comprising the following steps: a) forming a multi-component liquid drop comprising: a liquid core, a liquid intermediate envelope formed with an aqueous composition comprising at least one biopolymer, totally encapsulating at its periphery the liquid core, and a liquid external envelope formed with an aqueous composition, different from the intermediate composition, said aqueous composition comprising at least one polyelectrolyte and at least one surfactant, said liquid external envelope totally encapsulating at its periphery the intermediate envelope, the liquid core and/or the liquid intermediate envelope comprising at least one eukaryotic mammalian cell as set forth in claim 1, b) gelling by immersion of said multi-component liquid drop in a gelling solution containing a reagent capable of gelling the polyelectrolyte of the liquid external envelope, in order to obtain a gelled microcapsule comprising a gelled external envelope, c) stiffening the intermediate composition of the liquid intermediate envelope, in order to obtain a gelled and stiffened microcapsule comprising a stiffened intermediate envelope, said stiffened intermediate envelope having an elastic modulus that is non-zero, and d) recovering said gelled and stiffened microcapsules.

4. The microcapsule according to claim 3, the method further comprising a step for dissolving the gelled external envelope.

5. A method for preparing the microcapsule according to claim 1, the method comprising the following steps: a) forming a multi-component liquid drop comprising: a liquid core, a liquid intermediate envelope formed with an aqueous composition comprising at least one biopolymer, totally encapsulating at its periphery the liquid core, and a liquid external envelope formed with an aqueous composition, different from the intermediate composition, said aqueous composition comprising at least one polyelectrolyte and at least one surfactant, said liquid external envelope totally encapsulating at its periphery the intermediate envelope, the liquid core and/or the liquid intermediate envelope comprising at least one eukaryotic mammalian cell as set forth in claim 1, b) gelling by immersion of said multi-component liquid drop in a gelling solution containing a reagent capable of gelling the polyelectrolyte of the liquid external envelope, in order to obtain a gelled microcapsule comprising a gelled external envelope, c) stiffening the intermediate composition of the liquid intermediate envelope, in order to obtain a gelled and stiffened microcapsule comprising a stiffened intermediate envelope, and d) recovering said gelled and stiffened microcapsules.

6. The method according to claim 5, wherein the method further comprises a step for dissolving the gelled external envelope.

7. An in vitro method for cultivating eukaryotic mammalian cells comprising: a) cultivating a microcapsule under sufficient conditions for cell growth, said microcapsule comprising a liquid core, and at least one external envelope totally encapsulating the liquid core at its periphery, said external envelope being able to retain the liquid core when the microcapsule is immersed in a gas and comprising at least one gelled polyelectrolyte and/or one stiffened biopolymer, said microcapsule further comprising at least one eukaryotic mammalian cell as set forth in the preparation method according to claim 5; and b) harvesting said microcapsule.

8. A method for screening active ingredients comprising: a) cultivating the microcapsule according to claim 1 in the presence and in the absence of a candidate substance, b) detecting a phenotype of interest in the cells of the microcapsule cultivated in the presence of the candidate substance as compared with the cells of the microcapsule cultivated in the absence of the candidate substance, and c) identifying the candidate substance as an active ingredient if a phenotype of interest has been detected.

9. The method according to claim 8, wherein the active ingredient is a cosmetic active ingredient.

10. An in vitro method for cultivating eukaryotic mammalian cells comprising the following steps: a) cultivating the microcapsule according to claim 1 under sufficient conditions for cell growth, and b) harvesting said microcapsule.

11. The method according to claim 10, wherein said eukaryotic mammalian cells are human cells.

12. The microcapsule of claim 1, wherein the stiff intermediate envelope is obtained by a stiffening method selected from the group consisting of: polymerization, precipitation, colloidal aggregation, and a glassy transition caused by a variation in temperature.

13. The microcapsule of claim 1, wherein the stiff intermediate envelope is obtained by coacervation of an intermediate composition of a liquid intermediate envelope formed with an aqueous composition comprising at least one biopolymer.

Description

(1) The invention will be better understood upon reading which follows, only given as an example, and made with reference to the appended drawings, wherein:

(2) FIG. 1 is a large scale view in a section along a middle vertical plane of a gelled and stiffened capsule according to the invention; and

(3) FIG. 2 is a large scale view, in a section along a middle vertical plane of a stiffened (A) or gelled (B) capsule according to the invention.

(4) FIG. 3 is a large scale view, in a section along a middle vertical plane of a stiffened capsule for which the liquid core contains eukaryotic cells (A) and for which the liquid core and the stiffened envelope (B) contain eukaryotic cells.

(5) FIG. 4 is a large scale view, in a section along a middle vertical plane of a gelled capsule for which the liquid core contains eukaryotic cells (A) or for which the liquid core and the gelled envelope (B) contain eukaryotic cells.

(6) FIG. 5 is a large scale view in a section along a middle vertical plane of a gelled and stiffened capsule according to the invention comprising a liquid core, an external envelope and at least one intermediate envelope. In FIG. 5A, the gelled and stiffened capsule comprises eukaryotic cells in the liquid core. In FIG. 5B, the gelled and stiffened capsule comprises eukaryotic cells in the liquid core and in the intermediate envelope. In FIG. 5C, the gelled and stiffened capsule comprises eukaryotic cells in the liquid core, in the intermediate envelope and in the external envelope. In FIG. 5D, the gelled and stiffened capsule comprises two intermediate envelopes, the liquid core as well as the two intermediate envelopes comprising eukaryotic cells.

(7) FIG. 6 deals with the design and the principle for operating the microfluidic device for forming a microcapsule. The microfluidic platform consists of an external fluidic injection system, of a co-extrusion micro-device and of a gelling bath outside the chip (not shown). The enlarged view of the chip shows the 3 way configuration with the cell suspension (CS), the intermediate solution (IS), and the alginate solution (AL) respectively circulating in the most internal, intermediate and the most external capillary. The inlet orifices of the chip are collected to 3 syringes controlled by 2 syringe pumps. The liquid micro-droplets of compound fall into a 100 mm iso-osmotic calcium bath. The gelling of the alginated shell mediated by the calcium sets the structure of the capsule while the internal solutions diffuse and maintain the cells encapsulated. The analysis of the jet at the outlet of the end piece by a high speed camera, shows that at a low flow rate q (total flow rate), formation of droplets of the order of a millimetre is observed. At a high flow rate q, the intact length of the jet is too long as compared with the distance between the end piece and the gelling bath. No formation of droplets is observed when the flow enters into contact with the bath. At an intermediate flow rate q (typically between 50 and 150 ml.Math.h.sup.1), dispersion of the jet and formation of drops occurs before impact.

(8) FIG. 7 deals with the morphometric and mechanical characterisation of the microcapsules of alginate. (A) is a typical 2D point plot of the average radius depending on the circularity of capsules (without cells) directly collected from the gelling bath showing the existence of two populations of capsules: the fraction of the small spherical capsules (R150 m) (45%) and the fraction of the ellipsoidal or deformed largest capsules (55%) following coalescence of the droplets. (B) is a plot of the aspect ratio of the capsule h/R.sub.out versus the ratio between the internal and external flow rates q.sub.in/q.sub.out. The black points are the experimental data. The dotted line is the theoretical curve derived from conservation of the volume (see example 5, section Methods).

(9) FIG. 8 (A) is the plot of the deformation of the capsule AR/R0 versus the osmotic pressure difference 0. (B) is the Young Modulus E of an alginate gel from an osmotic inflation test. The Young modulus E of the alginate gel is derived from the slope of the dotted line which is adjusted to the data of the plot A. Representative histogram of the distribution of the values of the Young modulus (n=26).

(10) FIG. 9 (A) Micro-indentation of alginate gel capsules by using atomic force microscopy. Typical approach force-displacement curves (thin line, upper curve) and retraction curves (thick line, bottom curve) were obtained on a single capsule. (B) Elongation of macroscopic alginate threads. Stress-elongation plot for 5 different cylinders of alginate gel (length at rest of about 0.2 m, diameter of about 1 mm). The stress is derived from the weight of the calibrated masses by assuming a Poisson ratio of the gel v=. The line is a polynomial adjustment to the data of the second order generating a dependency on phenomenological constraints of the Young modulus at a significant deformation.

(11) FIG. 10 (A) Distribution of the average number of cells per capsule adjusted with a Gaussian curve. The count of the cells is obtained from phase contrast micrographs of the individual capsules merged with epifluorescence micrographs of encapsulated cells coloured with colouring agents for living/dead cells. (B) Plot of the stiffness of the MCS versus time from measurements obtained by photographs acquired with a phase contrast microscope, from a spheroid of cells CT26 in expansion inside a capsule until confluence. The time t=0 corresponds to encapsulation.

(12) FIGS. 11 and 12 deal with the quantitative analysis of the growth of spheroids and of the deformation of the capsules. Representative time plots show the influence of the stiffness of the capsule (via the thickness of the shell) on the growth and the mechanical characteristics of the MCS. FIG. 11A illustrates the study of the normalised radius of the spheroid R.sub.MCS relatively to the internal radius of the non-deformed capsule R.sub.0 versus time. FIG. 11B represents the study of the apparent growth rate 3{dot over (R)}.sub.MCS/R.sub.MCS versus time. FIG. 12A represents the study of the aspect ratio h/R.sub.out of the capsule versus time. The points are the experimented data. The lines are the theoretical predictions by assuming that the alginate gel is an incompressible material. FIG. 12B represents the study of the pressure exerted by the spheroid on the wall of the capsule versus time. The points with an anti-centre and the thin lines correspond to a thin capsule (h=8 m). The solid points and the thick black lines correspond to a thick capsule (h=28 m). The different phases discussed in the text are identified with grey rectangles. Confluence is considered as like the reference time t=0.

(13) FIG. 13 deals with the statistical analysis (n=23 for the thin capsules and n=17 for the thick capsules). FIG. 13A represents the study of the apparent growth rate of the spheroid according to a logarithmic scale for cell monolayers (2D), spheroid in free expansion (3D, free), spheroids encapsulated to confluence (3D, t=0), and during the last stages (phase 3) for the two shell thicknesses (3D, thin and thick). The pressure rate increases (13B) just after confluence (phase 2) and during the last stages (phase 3). FIG. 13C represents the phase contrast intensity study versus the radial distance relatively to the centre of the MCS and versus time. The bright line is the external wall of the capsule. Scale bars, 20 hours and 50 m.

(14) FIG. 14 Growth of spheroids inside the alginate capsules and after bursting of the capsules. Representative plots showing the time dependent change of the radius of the spheroid R.sub.MCS normalise relative to the initial internal radius of the capsule R.sub.0 for different spheroids CT26 in expansion in thin capsules. The time t=0 is the moment of confluence. The sudden increase in R.sub.MCS during the last stages corresponds to the bursting of the capsule. The spheroid freely grows at a rate similar to the one observed during the very early stages following encapsulation.

(15) FIG. 15 (A) Schematic illustration of a middle transverse section of a two-way injector (B) Schematic illustration of a middle transverse section of a three-way injector.

EXAMPLES

Example 1

Experimental Conditions for Obtaining Gelled and Stiffened Capsules

(16) Experimental Device

(17) The method for preparing capsules is based on the concentric co-extrusion of compositions via a triple envelope device for forming multi-component drops (FIG. 15B).

(18) A first composition (C1) circulating in a first compartment 21 of a triple envelope forms the first flow.

(19) A second composition (C2) circulating in a second compartment 31 of the triple envelope forms the second flow.

(20) A third composition (C3) circulating in a second compartment 41 of the triple envelope forms the third flow.

(21) Formation of Gelled and Stiffened Capsules

(22) At the outlet of the triple envelope, is then formed a multi-component drop, the first flow forming the liquid core, the second flow forming the liquid intermediate envelope and the third flow forming the liquid external envelope of the multi-component drop.

(23) The size of the liquid core, the thickness of the intermediate envelope and of the external envelope of the formed capsules are controlled by using several independent syringe pumps, by adjusting the injection flow rates of the different compositions C1, C2 and C3.

(24) The flow rate Q1 of the composition C1 is adjusted to 10 mL/h.

(25) The flow rate Q2 of the composition C2 is adjusted to 1 mL/h.

(26) The flow rate Q3 of the composition C3 is adjusted to 1 mL/h, and may be decreased down to 0.1 mL/h.

(27) Each multi-component drop detaches from the triple envelope and falls in a volume of air, before being immersed in a gelling solution of 1M concentrated calcium lactate.

(28) Once the external envelope is gelled, the gelled capsules formed are rinsed in a rinse solution based on water, and are then immersed in a stiffening bath.

(29) Formation of Stiffened Capsules

(30) The thereby formed gelled and stiffened capsules are then immersed in a depolymerisation solution of 10% concentrated citrate.

(31) Once the external envelope is depolymerised and removed, the obtained stiffened capsules are rinsed in a rinse solution based on water and stored in a storage solution based on water.

Example 2: Capsules with a Double Envelope Based on Natural Latex/Alginate

(32) The composition C1 is an aqueous solution of an amaranth colouring agent at 1 mM.

(33) The composition C2 is an aqueous dispersion of natural latex (chemical name Cis 1,4-polyisoprene, family of dienes, example of a commercial natural latex: natural Rubber grade TSR, SRM, SIR, STR, SVR, ADS, RSS, Crepes, DPNR, from Astlett Rubber Inc.) diluted down to a 20% to 40% mass fraction of particulated polymers relatively to the total mass of the natural latex dispersion, also comprising 1% by mass of a surfactant of the ionic or non-ionic type depending on the grade.

(34) In this example, the mass fraction of particles of polymers is set to 30% (the latex dispersion is titrated by gravimetry after washing by centrifugation) and the surfactant SDS (sodium dodecylsulfate) is used.

(35) The composition C3 is an aqueous solution having a 2.0% mass percentage of sodium alginate and a 0.1% mass percentage of SDS.

(36) The obtained capsules, with a standard diameter of few millimetres, are maintained in the gelling solution of calcium ions for one minute, and are then rinsed with distilled water. They are then stored in an isotonic solution relatively to the internal solution. Double coacervation by permeation of the calcium ions through the gelled alginate envelope is thus obtained. The capsules may then be incubated for 10 minutes in a 10% citrate solution in order to dissolve the outer membrane of alginate hydrogel. Capsules are thereby obtained, having an outer envelope of stiffened natural latex.

Example 3: Capsules with a Double Natural Latex/CB Alginate Envelope

(37) Example 3 is obtained under the same conditions as Example 2, except that the composition C2 further comprises carbon black CB: Carbon Black. To do this, a CB solution is prepared (from Carbon Black N234 from CABOT Corporation) in the presence of 2% SDS surfactant, the mass fraction of particles of polymers being still comprised between 20% and 40% based on the total mass of the natural latex dispersion. The CB fraction being comprised from 1% to 15%.

(38) For this example, the mass fraction of particles of polymers is set to 30% and the mass fraction of CB to 5% based on the total mass of the composition C2.

(39) After gelling the alginate envelope, the capsules are incubated in distilled water for about 20 minutes. The surfactant diffuses outwards from the capsules through the alginate envelope and causes coacervation of the mixed natural latex/CB mixture, giving rise to a stiffened envelope of reinforced rubber.

Example 4: Capsules with a Double Natural Latex/Colloidal Silica and Alginate Envelope

(40) Example 4 is produced under the same conditions as Example 2, except that the composition C2 further comprises colloidal silica with an average diameter of 100 nm (Aerosil from Degussa, Ludox from Sigma), according to mass fraction from 1% to 15% based on the total mass of the composition C2.

(41) For this example, the mass fraction of particles of polymers is set to 30% and the mass fraction of colloidal silica to 5% based on the total mass of the composition C2.

(42) Capsules are thereby obtained, including a stiffened envelope of reinforced rubber.

(43) The prepared capsules according to the invention are easy to form, they have a resistant envelope, with a small thickness, which gives the possibility of ensuring efficient de-aggregation of the capsule when the liquid contained in the capsule has to be released.

Example 5: Capsules with a Simple Envelope Based on Alginate

(44) I. Method

(45) I.a. Making the Co-Extrusion Device

(46) The central unit of the microfluidic devices consist in three glass capillary tubes co-aligned in the axial plane. The most external tapered capillary is obtained by stretching a rounded capillary in a transverse section (Vitrocom, internal diameter (i.d.) of 600 m, an external diameter (e.d.) of 840 m) with a micropipette structure (P2000, Sutter Instrument). The most internal capillary (i.d. of 100 m, e.d. of 170 m) and intermediate capillary (i.d. of 300 m, e.d. of 400 m) were maintained according to a cylindrical shape and were cut to the desired length. The ends of the capillaries were polished with micro-abrasive films (1 m grain, 3M) in order to avoid any bevel shape generating perturbations in the flow and for obtaining the desired tip diameter (typically between 130 and 180 m). A hydrophobic coating (1H,1H,2H,2H-perfluorooctyltrimethoxysilane, ABCR) was applied on the walls of the capillaries according to standard procedures (Perret, E., et al. Langmuir 18, 846.-854 (2002)) in order to prevent any humidification of the external walls of the tip of the injector with the alginate solution. The assembling of the co-extrusion device was carried out under a binocular microscope. The most external capillary was first stuck to a glass slide which is used as a support for the device. Next, the two other cylindrical capillaries were inserted and sealed sequentially by using an epoxy resin (Loctite 3430, Radisopares-RS Components). The co-axial and longitudinal alignments were manually checked during the drying of the resin at room temperature. The inlet orifices of the chip were made by sticking syringe needle fittings to a foam end piece (NN-1950R, Terumo) at the top of the free ends of the capillaries.

(47) I.B. Operation of the Co-Extrusion Device.

(48) The three liquid phases (cell suspension CS, intermediate solution IS and alginate solution ALsee FIG. 6) were loaded into syringes (10MDR-LL-GT SGE, Analytical Science) provided with needles connected to Teflon tubes (Bohlender, inner diameter of 0.5 mm). The other ends of the tubes were inserted into suitable inlet orifices of the co-extrusion device, which is vertically clamped to an upright inside a laminary flow hood. The syringes were mounted on syringe pumps (PHD 4400, Harvard Apparatus) which control the injection of the liquids at the desired flow rates. In this work, the inventors mainly used two sets of flow rates: 1) for thin capsules: q.sub.CS=50 ml h.sup.1, q.sub.IS=50 ml h.sup.1, q.sub.AL=40 ml h.sup.1, and 2) for thick capsules: q.sub.CS=20 ml h.sup.1, q.sub.IS=20 ml h.sup.1, q.sub.AL=30 ml h.sup.1. After initiation of the flow rates, the micro-droplets of compounds are directed towards a gelling bath containing 100 mm calcium chloride (VWR) and trace amounts of the surfactant Tween 20 (Merck), and are placed at approximately 0.5 m below the outlet orifice of the device. Operation for a few seconds was sufficient for producing about 10.sup.4 capsules, which were immediately washed in an iso-osmotic sorbitol solution and transferred into a suitable culture medium. After use, the microfluidic device was cleaned with a disinfectant (Biocidal ZF, Biovalley), ethanol and de-ionised water. Before the next use, the chip was rinsed with a sorbitol solution.

(49) I.C. Preparation of Aqueous Solutions and of Cell Suspensions.

(50) The most external phase (AL solution) was prepared by dissolving 2.5% w/v sodium alginate (FMC, Protanal LF200S) in water, and by adding 0.5 mM of sodium dodecylsulfate surfactant (SDS) (VWR). The solution was filtered at 1 m (Pall Life Science) and was stored at 4 C. The intermediate phase (IS) is generally a 300 mM sorbitol solution (Merck). The most internal phase (CS) was obtained by detaching the cells from the walls of the culture flask with a 0.5% EDTA-trypsin (Invitrogen). After washing in the suitable culture medium and delicate centrifugation (300g, 5 minutes, 20 C.), they are re-suspended in a 300 mM sorbitol solution at an approximate concentration of 310.sup.6 cells per ml.

(51) I.D. Cell Lines, Monolayer and Cultures of Encapsulated Cells.

(52) The inventors used carcinoma cells from the murine colon of the wild type CT26 (purchased from the American Tissue Culture Collection, ATCC CRL-2638) and the CT26 cells stably transfected with LifeAct-mCherry. Tests were also conducted with HeLa cells and murine sarcoma cells (S180, kind donation from Chu Yeh-Shiu, IMCB, Singapore).

(53) All the cells were maintained in Dulbecco's Modified Eagle Medium (DMEM, Invitrogen) supplemented with foetal calf serum at 10% (FBS, Invitrogen) and antibiotics (100 g ml.sup.1 of Streptomycin and 100 units ml.sup.1 of Penicillin (Gibco BRL) in a humidified atmosphere containing 5% CO.sub.2 at 37 C. by replacing the medium every 2 days. The cells were cultivated in the form of sub-confluents monolayers in order to prepare the cell suspensions used for encapsulation in hollow alginate spheres.

(54) Once the cell capsules were formed by following the procedure described above, they were placed inside an incubator (37 C., 5% CO.sub.2 at about 100% of relative humidity). Among the thousands of collected cell capsules, the majority was maintained in a Petri dish containing a culture medium and were cultivated under the same conditions as the cell monolayers. In each encapsulation cycle, several tens of capsules were selected for high resolution imaging. Depending on the requirements for configuring the microscope and on the desired duration of the imaging session (from a few hours to a few weeks), these selected cell capsules were transferred into dedicated culture chambers or devices (see the description below).

(55) As a comparison with our method, CT26 spheroids were also cultivated according to the standard technique on a bed of agarose (Hirschhaeuser, F. et al. J. Biotechnol. 148, 3.-15 (2010)).

(56) I.E. Colouration of the Fixed Spheroids.

(57) The spheroids were fixed in 4% PFA in PBS for 1 hour at room temperature. For the colouration of the cortical actin, they were incubated with 0.5 g ml.sup.1 of phalloidin conjugate with Alexa488 (Molecular Probes) in a PBS solution with Triton-X100 at 1% v/v (Sigma) at 4 C. for 2 hours overnight. Immunological marking of fibronectin was carried out according to a standard procedure. Briefly, the spheroids were permeablised by using Triton-X100 at 2% v/v in PBS. The primary antibodies (anti-fibronectin antibodies produced in a rabbit, Sigma) and secondary antibodies (anti-rabbit goat IgG (H+L) conjugate with Alexa Fluor568, Invitrogen) were diluted to 1/250 and are then incubated for 12 hours.

(58) I.F. Evaluation of Cell Viability and Cell Count.

(59) In order to evaluate the effectiveness of the encapsulation and the concentration of cell seeding, phase contrast images of the capsules were immediately taken after encapsulation and a number of cells per capsule was estimated by using standard ImageJ plugins (Schneider, C. A., et al. Nat. Methods 9, 671.-675 (2012)).

(60) The cell viability was characterised at different stages of the cell culture encapsulated with colouring agents for live/dead cells, calcein AM/ethidium homodimer-1 (EthD-1) (Invitrogen). In order to evaluate the potential obnoxiousness of the encapsulation method on suspended cells, the inventors incubated the cell capsules immediately after formation with EthD-1 at 4 M and calcein AM at 2.5 M for 30 minutes to 1 hour. The number of viable cells was determined by counting the red cells (dead) and the green cells (living) by using an inverted epifluorescence microscope (Axiovert-200M, Carl Zeiss) equipped with an Hg lamp and an EM CCD camera (C 9100-02, Hamamatsu Photonics). As a comparison, the same measurements were conducted on the cell suspension before encapsulation. When the spheroids were formed and were subject to expansion within the capsules, the same procedure was adapted by increasing the concentrations of the colouring agents by ten times and the incubation duration from 2 to 4 hours. The equatorial planes of the spheroids were then viewed as confocal microscopy images. While the red cells (dead) may be detected in the core of the spheroids, the living cells were practically not coloured beyond a peripheral edge of a few layers. This layer internal marking is due to the fact that the ubiquitous intracellular esterase activity of the peripheral cells is sufficient for esterifying all the calcein AM molecules permeable to the cells before they may penetrate more deeply into the spheroid.

(61) I.G. Imaging the Growth of Spheroids in the Long Run and of the Deformation of the Capsules.

(62) The growth of MCSes inside the capsules and the deformation of the shell were monitored by phase contrast microscopy. For all the encapsulation sessions, 24 capsules were selected from the entire batch of cell capsules and were individually transferred into each well of a multi-well culture plate (Falcon). Each capsule was viewed in the form of images every 3 hours with an inverted microscope Nikon EZ (dry objectives 10/0.25 NA or 20/0.4 NA) equipped with a colour CCD camera (Nikon Digital sight DS Fi1) driven by the software package NIS Element. Capturing the images was carried out at room temperature and lasted for approximately 5 minutes. Between each acquisition, the 24-well plate containing the capsules was maintained in the incubator. Half of the culture medium was renewed every two days. For acquiring real time sequences at a higher time resolution (1 sequence every 5 minutes) over extended time periods (about 15 days), the inventors also used a reverted microscope (Nikon Eclipse Ti, dry objective 10/NA0.3) equipped with a motor driven stage (Mrzhuser) and a system for controlling the climate (The Brick, Life Imaging Systems). The microscope and the camera (CoolSNAP HQ.sup.2, Photometrics) were driven by the Metamorph software package (Molecular Devices). In order to prevent any displacement or drift of the capsules in the well outside the field of view, the inventors designed an observation chamber custom-made. Phytagel (Sigma) wells of the following type were prepared by using conical moulds in PDMS (polydimethylsiloxane elastomer, Sylgard-184, Dow Corning) adapted to the wells of a 24-well plate with glass bottoms (Radnor, Pa.). This configuration facilitates the loading of the individual capsules, which are directed towards the centre of the well. The orifices (a diameter of 500 m) in the phytagel replicas, which are mainly used as a micro-conduit, also prove to be effective for limiting the movements of the encapsulating spheroids, and this without inducing stressors which may alter the growth of the MCS.

(63) I.H. Imaging the 3D Cell Organisation of the Encapsulated Spheroids.

(64) In order to view the peripheral cell layers and the core of the spheroids in expansion at a subcellular resolution, the inventors used confocal microscopy with single photon or multiphoton fluorescence.

(65) Confocal imaging of living cells was achieved by using an inverted microscope (LSM710, Carl Zeiss) equipped with a climate regulation chamber (Pecon) controlling the CO.sub.2 percentage, the temperature and the humidity. The samples were prepared by immersing the capsules in a solution of agarose with a low melting point at 0.3% (Invitrogen) (a culture medium without any serum, 37 C.) in a Petri dish with a glass bottom tailor made (diameter of the wells of about 2 mm). After gelling granules of agarose (10 minutes, room temperature), the Petri dish was filled with the culture medium. This assembly gave the possibility of immobilising the capsules, a step required for acquiring images with the z-stack acquisition method (automated acquisition of several images XY along the axis Z). The percentage of agarose was selected in order to generate a minimum stress on the MCSes in expansion. A comparison of the growth kinetics between the MCSes moving freely and the MCSes incorporated into the agarose did not reveal any significant difference. In order to monitor the cell dynamics within the spheroids, we used the cell line CT26 stably transfected with LifeAct-mCherry or CT26 cells of the wild type incubated in FM4-64 (Invitrogen, 2 g ml.sup.1). The fluorescence was acquired by using a laser pumped by solid state diode at 561 nm (15 mW) and an objective with immersion in oil 25/0.80 NA. The images of the surface of the fixed spheroids coloured with phalloidin-Alexa488 were viewed with an argon laser at 488 nm (25 mW) and an objective with immersion in oil 6311.4. The individual images and the stacks of images were processed by using the software package Zen 2011 (Carl Zeiss) and ImageJ or Fiji (Schindelin, J. et al. Nature Methods 9, 676.-682 (2012)). Videos online were edited by using After Effects and were then compressed by using Media Encoder (Adobe).

(66) A multi-photon microscope was used for accessing the core of the encapsulated spheroids. Two types of microscopes were used: 1) a vertical two-photon laser scanning microscope (Lavision) equipped with an objective with immersion in water 20/0.95 NA (Olympus); 2) an inverted microscope LSM710 NLO (Carl Zeiss) equipped with objectives with immersion in oil 25/0.80 NA or with immersion in water 40/1 NA (Carl Zeiss). The configurations were coupled with femtosecond lasers (690-1020 nm, from Coherent or Spectra Physics). The images of the inside of the fixed spheroids coloured by phalloidin-Alexa488 were acquired at a laser wavelength of 920 nm. Sulforhodamine B (SRB, Sigma) was added to the medium at a concentration of 40 g ml.sup.1. The best conditions for live imaging of the spheroids in a culture medium supplemented with SRB was obtained for an excitation at 800 nm (Marmottant, P. et al. Proc. Natl. Acad. Sci. U.S.A. 106, 17271.-17275 (2009)). The capsules were mounted as described for the single photon confocal live imaging.

(67) I.I. Morphometric Measurements of the Capsules.

(68) The characterisation of the sizes and of the shapes of the capsules was determined on capsules containing the cells and on empty capsules. The measurements on the empty capsules, which were obtained by replacing the CS phase with an iso-osmotic sorbitol solution, were conducted immediately after encapsulation and after a week of dwelling in the culture medium at 37 C. (in order to take into account potential morphological modifications induced by ageing). No significant difference was observed between these diverse conditions. The images of large fields for viewing densely grouped capsules were acquired with phase contrast microscopy and were analysed by using the ImageJ. The average radius of the capsule is defined as: R={square root over (S/)}, wherein S is the equatorial transverse surface of the capsule. The circularity of the capsule was measured as a ratio of the minor axis over the major axis of the ellipse adjusted to the external edge of the projected equatorial section.

(69) When the spheroids are at confluence, the external and internal walls of the capsule may be easily detected because of the high optical contrast. On the other hand, for empty or partly filled capsules, the internal wall of the capsule is slightly visible by phase contrast microscopy. The measurements of the thickness of the capsule were therefore conducted by doping the alginate solution with 250 g/ml of FITC-dextran with a high molecular weight (2 MDa, Sigma). The images of the capsules were acquired by confocal microscopy and were analysed with ImageJ. The influence of low flow rates on the aspect ratio h/R.sub.out was evaluated by comparing the experimental data with the theoretical value calculated from the conservation of the volume:

(70) $\frac{h}{R_{\mathrm{out}}}=1-{\left(\frac{q_{\mathrm{in}}/q_{\mathrm{out}}}{1+q_{\mathrm{in}}/q_{\mathrm{out}}}\right)}^{1/3}$

(71) I.J. Measurements of the Elasticity of the Alginate Gels.

(72) Three different methods were used for measuring Young's modulus of the alginate gels.

(73) The inventors first conducted measurements of micro-indentation by AFM on empty capsules.

(74) Alginate capsules positioned at the bottom of a Petri dish filled with a culture medium were placed on the sample stage of an AFM system Catalyst (Bruker) mounted on an inverted optical microscope (171, Olympus) in a force mode (FIG. 9A). The inventors used TR400 cantilevers attached to spherical SiO.sub.2 beads (diameter of 5 m) and having a rated stiffness constant k.sub.cantilever=0.06 N/m (Novascan). The sensitivity of the photodiodes was calibrated before and after measurements on a freshly cleaved mica surface in PBS. The stiffness constant was determined by using the method of thermal fluctuations applied in the software package Bruket Nanoscope 7.2. The force-distance curves (F-z) were recorded for displacements of a peak-to-peak amplitude of about 2 m at 0.25-1 Hz. The relative deflection threshold was controlled for attaining a capsule deformation comprised between 200 nm and 500 nm. The data were analysed within the scope of an indentation of the punctual load in hollow spheres. The functional force (F)deformation () relationship (Fery, A. & Weinkamer, R. Polymer 48, 7221.-7235 (2007)) is the following:

(75) $F=\frac{4}{3\sqrt{1-v^2}}E\frac{h^2}{R}.$
The deformation was calculated in terms of a contact point (z.sub.c) and of the shift of the deflection (d.sub.0) as d=zz.sub.c(dd.sub.0). Experimentally, Young's modulus of the alginate gel was derived from adjustment of the force-deformation traces (FIG. 9A) by taking the values measured for the geometrical properties (R and h) of the capsule and v=0.5 for the Poisson ratio. The inventors observed that E=5544 kPa (SD, N=7).

(76) The second method consists of conducting measurements of the traction on macroscopic alginate gel cylinders of the spaghetti type. These threads (length L.sub.0 of about 0.2 m, diameter D.sub.0 of about 1 mm) were formed with a simple 1 way extrusion device provided with an end piece with a size of about 1 mm, by immersing the tip in the calcium bath in order to suppress the instability of the capillary. A controlled stress was applied with a set of calibrated weights m suspended from the alginate cylinders. The elongation L/L.sub.0 of the alginate sample was measured with a ruler. By supposing that v=0.5, the Young modulus was derived from

(77) $=\frac{4\mathrm{mg}}{D_0^2\left(1-L/L_0\right)}=E.Math.\frac{L}{L_0}.$
The inventors observed that E=7112 kPa (SD, N=9).

(78) A third determination of E is based on an osmotic inflation test. For this purpose, the inventors replaced the cell suspension with a sorbitol solution with 5% w/v dextran, P.sub.m=2 MDa and 500 kDa (Sigma Biochemika). The calcium bath solution and the storage culture medium were also supplemented with 5% w/v dextran. Iso-osmotic equilibrium of all the solutions was controlled. In order to obtain a detectable inflation, capsules with very thin walls were prepared (q.sub.in/q.sub.out=10, which corresponds to a shell thickness h of about 5-7 m). Stepwise dilution of dextran caused osmotic inflation of the capsules. The differences in concentration in the dextran were converted into osmotic pressures .sub.0 and the expansion of the capsules R/R.sub.0 was directly measured. To the first order, in the limit of a slight deformation, the Young's modulus of the alginate was derived by balancing the elastic energy of the spherical shell and the effect obtained by the osmotic pressure difference:

(79) $E\frac{1}{4\mathrm{ho}/\mathrm{Ro}}.Math.{\left(\frac{R/\mathrm{Ro}}{o}\right)}^{-1}$

(80) I.K. Determination of Young's Modulus of an Alginate Gel from the Osmotic Inflation of a Capsule

(81) Considering a spherical capsule consisting of an alginate shell containing a high molecular weight dextran solution (P.sub.m=500 kDa or 2 MDa) immersed in a less concentrated dextran solution, given that the shell is permeable to water (estimated porosity of about 6 nm) but not to dextran (Stokes radii between about 15 nm and 27 nm), the water molecules diffused into the capsule, which inflates until the elastic force of the stretched capsule balances the osmotic pressure.

(82) At the beginning of the test, the dextran concentrations inside and outside the capsule are equivalent. The initial radius of the capsule is R.sub.0, and the dextran concentration in the external bath is then reduced by dilution, so that the concentration difference is c.sub.0. During the inflation, the radius of the capsule increases by R=RR.sub.0 and the concentration difference is reduced from c.sub.0 to c:

(83) $\begin{array}{cc}c=c_0.Math.{\left(\frac{R_0}{R}\right)}^3.& \left(1\right)\end{array}$

(84) The stretching elastic energy is given by (Landau, L. D., et al. Theory of Elasticity, Third Edition: Volume 7. (Butterworth-Heinemann: 1986)):

(85) $\begin{array}{cc}{G}_{\mathrm{el}}=4\frac{E}{1-v}{h\left(R-R_0\right)}^2,& \left(2\right)\end{array}$
wherein h is the thickness of the shell and v is the Poisson ratio. For an incompressible material, v= and the shell becomes thinner when the capsule inflates, according to:

(86) $\begin{array}{cc}h=h_o.Math.{\left(\frac{R_o}{R}\right)}^2,& \left(3\right)\end{array}$
wherein h.sub.0 is the thickness of the unstretched capsule.

(87) Given that solutes are very bulky, the osmotic pressure significantly deviates relatively to the rated value (=nk.sub.BT, wherein n is the number of active species from an osmotic point of view and k.sub.BT is the thermal energy) and proves to be independent of their rated osmolality beyond a given threshold (P.sub.m=200 kDa for dextran) (Reid, C. & Rand, R. P Biophys J 73, 1692-1694 (1997)). Different empirical expressions are reported for adjusting the data of the osmotic pressure (Veretout, F Journal of molecular biology 205, 713-728; Bonnet-Gonnet, C. et al. Langmuir 10, 4012-4021 (1994)). For simplicity purposes, we consider the polynomial expression well established for as a function of c (in weight/volume percentage):
=c+c.sup.2+c.sup.3(4),
wherein =286, =57 et =5. The effect generated by the osmotic pressure for inflating the capsule from R.sub.0 to R is given by:
W=.sub.R.sub.o.sup.R.Math.4R.sup.2dR(5).
By taking into account the dilution effect (Eq. 1), we obtain:

(88) $\begin{array}{cc}W=4R_0^3\left(c_0\ln \left(\frac{R}{R_0}\right)+\frac{1}{3}{c}_0^2\left(1-{\left(\frac{R_0}{R}\right)}^3\right)+\frac{1}{6}{c}_0^3\left(1-{\left(\frac{R_0}{R}\right)}^6\right)\right).& \left(6\right)\end{array}$
The radius of the capsule at equilibrium is indicated by the minimum of the total energy G.sub.el+W. Further by assuming small deformations, R/R.sub.01, we reach:

(89) $\begin{array}{cc}\frac{R}{R_0}=\frac{{}_0}{{}_c+4E\left(h_0/R_0\right)},& \left(7\right)\end{array}$
wherein .sub.0 is the osmotic pressure at c.sub.0, and .sub.c=.sub.0+3c.sub.0.sup.2+6c.sub.0.sup.3.

(90) This reveals that the osmotic pressure .sub.0 varies from 0 to 4 kPa within the explored range of differences in concentrations. The approximation indicated above lies on the assumption that the correction introduced by .sub.c remains negligible relatively to the effective Young modulus E4h.sub.0/R.sub.0. By assuming E=68 kPa and h.sub.0/R.sub.0 is about 0.05, this is only valid for c.sub.0<2%. Under our experimental conditions (c.sub.0 varying from 0 to practically 5%), a more accurate determination of E requires the use of Eq. 7.

(91) I.L. Analysis of the Growth of the Spheroids and of the Deformation of the Capsules.

(92) The phase contrast real time images were analysed by using an algorithm for detecting ages based on the gradient and tailor made, applied in Matlab (MathWorks). By beginning from the centre of the capsule, the intensity profiles were acquired in a radial position and were inspected in order to identify the peaks in the first derivative in order to extract the contour of the MCS and of the capsule containing it in each recorded structure. R.sub.out was derived from the projected transverse surface. A similar approach was followed for monitoring R.sub.MCS inside the capsule. The background noise detected before confluence was mainly due to rotary movements of the non-perfectly spherical cell aggregate. The confluence time (t=0) was determined as the time for which the growth of the MCSes exhibits an inflexion point. The inventors checked that this time coincided, in less than 5 minutes, with the visual determination of confluence (on high time resolution videos). The pre- and post-confluence stages were also quantified by a roughness parameter, =P/2{square root over (A)}, P and A respectively being the perimeter and the surface area of an equatorial transverse section. Whereas the time-dependent change in row has a background noise during the first stages of the growth of the MCS, it decreases when the spheroid approaches the wall of the capsule, before it is saturated to a minimum value close to the theoretical value of 1 for perfectly spherical objects.

(93) I.M. Phenomenological Approach for Non-Linear Elasticity of Alginate Capsules at Significant Deformations

(94) In order to confirm the measurement of Young's modulus derived from the osmotic inflation test, the inventors developed a second mechanical test, consisting of directly evaluating the stress ()-deformation () relationship of the alginate gel threads. These threads (diameter of 1 mm) were stretched with calibrated weight to which were welded tiny alginate droplets at one end. Under low deformation conditions, (typically for the relative elongation =L/L.sub.0<10%), the stress-deformation response is linear and the derived Young's modulus is quite compliant with the one measured earlier (E=7112 kPa). In the case of a highly significant deformation greater than (>80%), water formed from the sample and significant plasticity was obvious. For the intermediate deformation, the material has a non-linear stress-deformation response (FIG. 9B). Such a hardening behaviour at a stress is quite common for biopolymer gels and has already been reported for alginate gels (Zhang, J., et al. 2007 Journal of Food Engineering 80, 157-165). Given that the thin capsules (h/R of about 0.1) which were considerably used in this work exhibit a maximum radial deformation R/R.sub.0 of about 30% before bursting, an accurate determination of the pressure exerted by the confined spheroid in expansion require that this effect be taken into account. A standard phenomenological approach for non-linear elasticity consists of considering a corrective term in .sup.2 for the stress (=E.sub.+A.sub..sup.2). Conversely, by adjusting the with a polynomial expression of the second order, the inventors defined an effective elastic modulus depending on the deformation E.sub.eff()=E(1+a) and we observed =1.5. We used this expression for E in order to derive the pressure from deformation data on thin capsules.

(95) I.N. Expansion of a Spherical Container with Thick Walls Subject to Internal Pressure

(96) The inventors have assumed that the alginate gel is isotropic and that the deformations are small (i.e. <10%). On the other hand, if the condition h/R1 is not satisfied, the assumption of a constant tangential stress through the thickness of the container is not valid. In the general case of a Poisson ratio v, the inventors have to resort to expressions for the radial and circumferential stress (Fung, Y. C. Foundations of Solid Mechanics; Prentice Hall: 1965):

(97) 0$\begin{array}{cc}{}_r=\frac{{\mathrm{PR}}_{\mathrm{in}}^3}{R_{\mathrm{out}}^3-R_{\mathrm{in}}^3}\left(1-\frac{R_{\mathrm{out}}^3}{R^3}\right)& \left(1\right)\\ {}_{}=\frac{{\mathrm{PR}}_{\mathrm{in}}^3}{R_{\mathrm{out}}^3-R_{\mathrm{in}}^3}\left(2+\frac{R_{\mathrm{out}}^3}{R^3}\right),& \left(2\right)\end{array}$
wherein R.sub.inRR.sub.out.
The radial displacement u(R) is obtained from Hooke's law:

(98) $\begin{array}{cc}u\left(R\right)=\frac{\left(1-v\right){}_r-v{}_{}}{E}R.& \left(3\right)\end{array}$
By collecting these results, the inventors reached:

(99) $\begin{array}{cc}u\left(R\right)=\frac{P}{E}\frac{{\mathrm{PR}}_{\mathrm{in}}^3}{R_{\mathrm{out}}^3-R_{\mathrm{in}}^3}\left[\left(1-2v\right)R+\frac{\left(1+v\right)}{2}\frac{R_{\mathrm{out}}^3}{R^2}\right].& \left(4\right)\end{array}$

(100) If the material is incompressible, this equation may be simplified and applied for two particular cases of interest, notably R=R=R.sub.in et R=R.sub.out:

(101) $\begin{array}{cc}u\left(R_{\mathrm{in}}\right)=\frac{3}{4}\frac{P}{E}\frac{R_{\mathrm{in}}}{1-{\left(R_{\mathrm{in}}/R_{\mathrm{out}}\right)}^3},& \left(5\right)\\ u\left(R_{\mathrm{out}}\right)=\frac{3}{4}\frac{P}{E}\frac{R_{\mathrm{in}}}{{\left(R_{\mathrm{out}}/R_{\mathrm{in}}\right)}^3-1}.& \left(6\right)\end{array}$

(102) Finally, from the conservation of the volume of the shell, we have:
R.sub.out.sup.3(t)R.sub.in.sup.3(t)=R.sub.out.sup.3(0)R.sub.in.sup.3(0)=(R.sub.0.sup.3)(7).

(103) By using this equation, the two time variables R.sub.in(t) and R.sub.out(t) are separated and the pressure P(t) is written in function either of R.sub.in(t) or R.sub.out(t). Experimentally, only the initial external and internal radii therefore have to be measured and the time dependent change of the internal or external radius of the capsule has to be followed.

(104) $\begin{array}{cc}P\left(t\right)=\frac{4}{3}E\left[1-\frac{1}{1+\left(R_0^3\right)/R_{\mathrm{in}}^3\left(t\right)}\right]\frac{u\left(R_{\mathrm{in}}\left(t\right)\right)}{R_{\mathrm{in}}\left(t\right)},& \left(8\right)\\ P\left(t\right)=\frac{4}{3}E\left[\frac{1}{1-\left(R_0^3\right)/R_{\mathrm{out}}^3\left(t\right)}-1\right]\frac{u\left(R_{\mathrm{out}}\left(t\right)\right)}{R_{\mathrm{out}}\left(t\right)}.& \left(9\right)\end{array}$

(105) Let us note, that by returning to the general case described by Eq. (4) and by constructing the ratio of the displacements at the internal and external surfaces, it is found (Dym, C. L. & Williams, H. E. (2007) International Journal of Mechanical Engineering Education 35, 108-113):

(106) $\begin{array}{cc}\frac{u\left(R_{\mathrm{out}}\right)}{u\left(R_{\mathrm{in}}\right)}=\frac{3\left(1-v\right)}{2\left(1-2v\right)+\left(1+v\right){}^3},& \left(10\right)\end{array}$
wherein p=R.sub.out/R.sub.in>1.

(107) First of all, given that this ratio is always less than one, the displacement at the external radius is smaller than that at the internal radius, which is intuitive and experimentally observed. Next, u(R.sub.out)/u(R.sub.in) provides a direct estimation of the Poisson ratio, which prove to be v=.

(108) II Results

(109) II.A Formation of Alginate Microcapsules Assisted with a Microfluidic Device

(110) The procedure for preparing the cell microcapsules is inspired from the method developed for making liquid pearls of the order of one millimetre and is further adapted for reducing the diameter of the capsules and reaching the requirements of a cell culture. The fundamental operation principle consists of generating a hydrogel shells containing a suspension of cells by co-extrusion (FIG. 6A). More specifically, the microfluid device is assembled by co-centring of three glass capillaries (FIG. 6B). The cell suspension circulates in the most internal capillary while an alginate solution is injected in the most external tapered capillary. Gelling of the alginate shell is achieved out of the chip in a calcium bath. An intermediate capillary filled with a solution without any calcium is used as a barrier to the diffusion of the released from intracellular stocks and thus avoids blocking of the chip. The inventors also modified the mode of formation of the droplets. At flow rates q, the liquid froze drop-wise from the capillary and produces capsules with a size of 2-3 mm (FIG. 6C), as a consequence of the interaction between gravity and surface tension. In contrast, at a higher flow rate q, the liquid emerges as a jet, which is dispersed into droplets downstream because of the instability of the capillary. It is then expected that the size of the droplets be closely associated with the diameter D of the orifice. A lower limit for the flow rate of the liquid is defined by the condition for occurrence of the dropwise-jet transition, i.e. for a critical Weber number pV.sup.2D/4 with a liquid in the non-viscous limit and with low gravity. By neglecting the structure of the flow in three phases and assuming a simple liquid with =10.sup.3 kg m.sup.3, =50 mN m.sup.2, one obtains V.sub.min1 m s.sup.1, and q.sub.min=(D/2).sup.2V.sub.min of the order of 40 ml h.sup.1 for D=130 m. An upper limit for q is controlled by the height of the fall: the distance d between the end piece and the gelling bath surface should be greater than the intact length of the jet, which may attain 10 to 100D (FIG. 6c), depending on q and external perturbations. On the other hand, the inventors have observed that increasing d promoted the coalescence of two consecutive drops before gelling, which finally generated larger capsules of an ellipsoidal shape. Up till now, the inventors had neglected the fact that the core of the droplet, the shell and the gelling bath were aqueous phases therefore are priori miscible. In order to avoid any mixing, the inventors added trace amounts of surfactant to the alginate solution and to the surface of the gelling bath, which reduces the surface tension and imparts transient stiffness to the drop of compound during the impact.

(111) II. B Characterisation of the Microcapsules

(112) In a typical experiment, approximately half of the capsules are spherical (as determined by the circularity parameter >0.8) and monodispersed (FIG. 7A). The production rate of the capsules (>10.sup.4 s.sup.1) is sufficiently high for allowing rapid manual selection of 10-100 capsules of spherical shape. Although it is possible to increase the fraction of the spheres by forcing instability of the capillary by controlled flow perturbations, slight anisotropy will always be present because of the presence of a small tail which is inherent to the impact in the gelling bath. However, this anisotropy has a negligible effect on the mechanical measurements reported below. It is expected that the average size of the droplets of compound be determined by the fastest growth mode 2/ of the Rayleigh instability. Given that is proportional to the diameter of the liquid jet d.sub.jet for a given viscosity contrast, the conservation of the volume between a cylinder of length and of section d.sub.jet.sup.2/4 and a drop of a radius R causes R=d.sub.jet. For most operational conditions, the diameter of the end piece was D=130 md.sub.jet, by producing an average drop size R=14821 m, which is compliant with the theoretical prediction of the first order. The thickness of the shell may be measured in confocal imaging by colouration with fluorescent dextran of high molecular weight of an alginate capsule. Such an observation gives the possibility of observing a clear separation of the shell from the cell suspension and from the intermediate solution of the capsule, a reduced mixture of the constituents of the capsule. Thus, the thickness of the shell h may be measured with accuracy. However, in a more interesting way, h may be adjusted by varying the ratio between the internal flow rate q.sub.in (sum of the flow rates of the cell suspension and of the intermediate solution) and the external flow rate q.sub.out of the alginate solution. Modifications of the ratio q.sub.in/q.sub.out mainly have an effect on the aspect ratio h/R.sub.out (FIG. 2d), R.sub.out being the external radius of the capsule. The production of capsules with very thin walls is limited by the fragility of the shell. However, an increase in the alginate flow rate aiming at producing very thick shells will tend to generate heterogeneous and deformed capsules. In practice, for capsules with a radius of about 150 m, h may vary, completely reliably, between 5 and 35 m (FIG. 7B).

(113) The inventors also studied the mechanical properties of alginate capsules. Quite surprisingly, the rheology of alginate gels is still an object of debate. Except for the discrepancies observed between studies which use distinct techniques, the Young modulus E of alginate gels, which characterise the stiffness of the raw material, depends on many parameters (alginate concentration, chemical composition, nature and concentration of cross-linking cations). In order to avoid any variability depending on the procedure, the inventors directly evaluated the elasticity of the raw gel of the capsules by using an osmotic inflation test. According to the deformation of capsules pre-loaded with high molecular weight dextran and immersed in a solution gradually depleted of dextran, the inventors derived E=6821 kPa (FIGS. 8A and 8B, example 5 Methods). This value was further confirmed by a micro-indentation test with AFM (FIG. 9A) and a macroscopic elongation of raw alginate cylinders (FIG. 9B). Even if the alginate gels have a particular structure illustrated by the egg box model, an approximate relationship valid for cross-linked polymer gels reticules.sup.27, E/3=kT/.sup.3, gives the possibility of estimating the average size of the meshes of the gel =6 nm, which is sufficiently significant so that globular proteins with a P.sub.m of about 150 kDa may diffuse through the latter. No hysteresis was observed during osmotic inflation-shrinking cycles and no time-dependent change in the deformation was detected when the osmotic pressure difference was maintained for longer time periods (data not shown), which suggests that the hydrogel behaves like a purely elastic material.

(114) II.C Quantitative Analysis of the Growth of a Spheroid Confined in an Elastic Environment

(115) In order to obtain a quantitative description of the impact of the elastic confinement on the growth of MCSes, the inventors adjusted the stiffness, k.sub.caps, of the capsules by varying the thickness of the shell (k.sub.capsEh) and monitored the time-dependent change of the average radius of the spheroids, R.sub.MCS(t), by using microscopy/real time video. Three distinct phases were observed. During phase 1, before confluence (t<0), the R.sub.MCS rapidly increases at similar rates in the thick and thin capsules (FIG. 11A). The spheroid freely grows inside the capsule at a constant growth rate, {dot over (V)}/V=3{dot over (R)}.sub.MCS/R.sub.MCS1.25 (jour).sup.1, which is similar to the doubling rate of 2D cell monolayers (FIGS. 11B and 13A). Phase 2 typically begins when the R.sub.MCS approaches the internal radius of the shell R.sub.in within a single cell size (about 10 m). At t=0 (confluence), the apparent growth rate {dot over (V)}/V decreases by about three times. Phase 2 corresponds to the smoothing transition, and approximately last from t=1 day to t=+1 day (FIG. 11B). During phase 3 (t>0), the R.sub.MCS is practically stabilised for thick capsules and continues to slowly increase for thin capsules (FIG. 11A). An in-depth inspection reveals that {dot over (V)}/V drops by more than one order of magnitude as compared with the free growth of the MCS, but never becomes strictly equal or zero (FIG. 11B). The average of approximately 20 capsules indicates that {dot over (V)}/V in phase 3 is of about 0.07 (day).sup.1 for thin capsules and 0.04 (day).sup.1 for thick capsules (FIG. 13A).

(116) From a qualitative point of view, even if a slower growth is expected in the case of confinement in stiffer capsules, a quantitative explanation requires that the pressure exerted by the expanding MCS be derived. As a first approximation, the capsules have to be considered as pressurised containers with thin walls within the scope of isotropic linear elasticity. The pressure which inflates the shell is then given by:

(117) $P=\frac{2E}{1-v}.Math.\frac{h}{R}.Math.\frac{u\left(R\right)}{R},$
wherein u(R) is the radial displacement at a distance R.sub.inRR.sub.out from the centre of the capsule, and v is the Poisson ratio (Landau, L. D., et al 1986, Theory of Elasticity, Third Edition: Volume 7. Butterworth-Heinemann). The slow reduction of h(t) observed (FIG. 12A, symbols) is compliant with a 1/R.sup.2 dependency (FIG. 12A, lines), as expected for an incompressible gel (v=). In practice, the experimental conditions require additional corrections. First of all, for thin capsules, the assumption of linear elasticity cannot be applied given that the deformations exceed 20%. A phenomenological dependency of Young's modulus on the deformation has to be taken into account for non-linear elasticity (see Example 5 point M.). Next, for thick capsules (h/R of about 0.25), the complete formalism of the theory of a container with thick walls has to be used (see Example 5 point N.). Taking into account these corrections, the inventors observed that the pressure curves of thin and thick capsules mainly drop within the experimental error (FIG. 12B) and exhibit two main characteristics. First of all, the pressure rapidly accumulates during the first 24 hours after confluence (FIG. 13B, {dot over (P)}=2.40.5 kPa (day).sup.1). Next, at a threshold pressure P.sub.th=2.20.5 kPa, the transition in phase 3 is indicated by a dramatic drop in the increase of the pressure, which attains a constant value as low as {dot over (P)}=0.20.08 kPa (day).sup.1) (FIG. 13B). The single fact that {dot over (P)} remains positive indicates that the growth of the spheroids is not interrupted, as confirmed by the resumption of rapid growth after dissolution or bursting of the capsule (FIG. 11A and FIG. 14). On the whole, these results demonstrate that the mechanical characteristics of the confined spheroids may be characterised from a quantitative point of view by measuring the deformation of the elastic capsules. On the other hand, in order to obtain a mechanistical understanding of an altered MCS growth under confinement conditions, it is necessary to study the outcome of post-confluence spheroids at a cell and molecular level.

(118) II.D Impact of Elastic Confinement on the Internal Cell Organisation of the Spheroids

(119) As aforementioned, the post-confluence stages of the MCSes are characterised by the clear occurrence of a dark core. The reorganisation of the structure of the MCS seems to be concomitant with the occurrence of phase 3 (FIG. 13C). In order to elucidate the cause of this significant transparency loss of the core of the MCS, the inventors used fluorescent colouring agents non-permeable to the membranes. First of all, they used an agent for staining hydrosoluble proteins, sulforhodamine B (SRB), which accumulates in the extra-cellular space (permeabilised cells or secreted proteins). By two-photon microscopy, the inventors observed that i) a pale core is nucleated a few hours after confluence (at P of about 0.5 kPa), ii) it propagates towards the outside in a fractal type way (as far as PP.sub.th), and iii) at subsequent moments overtime, the marked core occupies the largest fraction of the spheroid while the 3-4 first peripheral cell layers remain colourless (data not shown). As a control, a spheroid of the same size (R150 m), cultivated in a larger capsule and released before confluence, the difficulty reveals any colouration. The organisation of the core sensitive to the SRB colouring agent is for example only induced by the confinement and is not comparable with the formation of the necrotic core observed in larger spheroids (R>400 m), resulting from limited diffusion of oxygen and nutrients. Next, the inventors acquired encapsulated MCS images marked with a colouring agent sensitive to the membranes, FM4-64. Given that the fluorescence of FM4-64 is more intense in a lipophilic environment, the nuclei of living cells are negatively coloured. The necrotic events are revealed by the occurrence of strong fluorescence in the integrality of the cell. The similarity between the profiles of SRB and of FM4-64 confirms that the core induced by the confinement consists of permeabilised cells or cell debris. Nevertheless, an immunocolouration of fixed post-confluence MCSes also reveals the presence of fibronectin (data not shown), which suggests that the core consists in a mixture of dead cells and secreted proteins of the matrix. This nature of the core of the mixture type is consistent with its strong apparent cohesion given that it resist to dissociation following a treatment with trypsin (data not shown).

(120) The imaging of the core of an MCS is a difficult task because of the restricted diffusion of extrinsic colouring agents and of the limited penetration depth of light. On the other hand, the border of the compressed cells between the shell and the core is further sensitive to high-resolution microscopy of living cells. The inventors have obtained images of CT26 cells stably transfected with LifeAct-mCherry for 3 days before and after confluence (data not shown). At the start, the cells are relatively rounded and moderately mobile within the expanding spheroid. Once the confluence is reached, most of the peripheral cells exhibit significant migration and form long and thin protrusions with lamellopodia and filopodia at the ends. Lamellopodia and filopodia were also observed in non-transfected fixed cells coloured with fluorescent phalloidin (data not shown).

(121) On the whole, these imaging data suggests that confinement induced by the capsule causes reorganisation within the spheroid after confluence, which assumes a layered structure at equilibrium consisting of a compact core consisting of cell debris cemented by extra-cellular proteins such as fibronectin, elastin, and a peripheral border of highly motile elongated cells.

Example 6: Encapsulation of CT26 Cells in a Simple Alginate Capsule

(122) I Experimental Conditions

(123) I. A. Encapsulation

(124) The encapsulation of the cells is achieved by forming a jet consisting of two co-axial phases. The internal phase containing the suspended cells in their culture medium, or an iso-osmotic biological buffer compatible with the encapsulation method; this phase will compose the core of the capsules. The external phase consists of a dispersion of sodium alginate at 2% m/v having an L-guluronic/D-mannuronic (G/M) ratio comprised between 65-75%/25-35% and a viscosity for a 1% m/v dispersion at 20 C. comprised between 200 and 400 mPa.Math.s. (i.e. FMC BioPolymer, Protanal LF 200S) and 0.5 mM of sodium dodecyl sulfate (SDS). The external phase will produce in fine the alginate shell of the capsule. Each of the phases is placed in a sterile syringe, the flow rate of which is controlled by a syringe pump. The syringes are connected to a two-way injector schematised in FIG. 15A, giving the possibility of producing a jet. According to this schematic illustration, the internal phase intended to be encapsulated circulates through the compartment 21 so as to be injected in the centre of the capillary C. The external phase intended to form the alginate shell of the capsule circulates through the compartment 41 and is injected at the internal periphery of the capillary C.

(125) The flow rates delivered by the syringe pumps depend on the geometry of the injector, notably on the diameter of the outlet capillary, and on the viscosity of the fluids used. These flow rates are adapted so as to allow the formation of a jet (passing from the drop-wise conditions to a jet) for which the fragmentation in microdroplets is accomplished according to the Plateau-Rayleigh instability. This fragmentation may be controlled by applying to the fluid of the external phase a vibration controlled by a piezo-electric effect with a frequency located between 0 and 2000 Hz. In order to prevent coalescence of the microdroplets formed, a cylindrical electrode is placed at the fragmentation site of the jet; a DC current under 0 to 2000 V is applied and has the effect of electrically charging the surface of the microdroplets thereby ensuring their respective repulsion and preventing their coalescence.

(126) The multi-component microdroplets formed during the fragmentation of the jet, under the effect of gravity, fall into an aqueous solution of 1% (m/v) calcium chloride which has the effect of cross-linking the alginate outside the microdroplets and of thus forming the alginate shell containing in its core the cells of the internal phase. The capsules are reenvelopeed, rinsed in a physiological buffer not depolymerising the alginate (i.e. without any phosphates or chelating agents) and then placed in sterile cell culture flasks with the culture medium used for the cells. The alginate shell of the capsules being semi-permeable, it allows diffusion of the nutrients and of the gases required for cell survival and growth. The capsules are incubated at 37 C. and with 5% of CO.sub.2 in order to allow the growth of the cells.

(127) I.B. Cell Survival and Growth

(128) The cells used are tumoral cells of the CT26 line.

(129) As the alginate capsules are optically transparent, the encapsulated cells in a first phase were observed in optical microscopy in order to determine their morphology and to follow their evolution.

(130) Cell survival may be determined by using conventional colorimetric methods (e.g. MTT, XTT, Resazurin tests) or fluorimetric methods (calcein, fluorescein diacetate, propidium iodide) based on the metabolism and cell physiology.

(131) In this case, simple cell survival of the cells encapsulated in capsules was carried out by marking with calcein and with propidium iodide, was carried out on newly formed capsules according to the following method. The capsules are incubated in the presence of an esterified form of calcein (Calcein-AM, LifeTechnologies) not fluorescent under the conditions prescribed by the manufacturer. This fluorophore diffuses through the capsule and through the plasma membrane and is hydrolysed within cells for which the metabolism is active (i.e. living); the thereby produced calcein is fluorescent in green and remains, because of its charge, in the cytosol of the cells. After incubation of the capsules with calcein-AM, the capsules are put into contact with propidium iodide. This fluorophore, because of its charge, only diffuses into the cells for which the plasma membrane is damaged and binds onto the DNA, which has the effect of increasing its fluorescence by 20 to 30 times. Thus, after exposure of the capsules to these two fluorophores, the observation of the cells under confocal microscopy gives the possibility of distinguishing the living cells, which are fluorescent in green, from dead cells, which are fluorescent in red.

(132) II. Results

(133) The cell viability was controlled by means of the Live/Dead test at D0 and at D15 after the encapsulation showing very good cell survival thus, the cells survive to encapsulation and have good cell growth beyond 15 days.

(134) The observation of the capsules in optical or confocal microscopy confirms the formation of spheroids, i.e. cell aggregates. Such structures are observed with tumoral cells which do not adhere to the walls of the capsule. These capsules are therefore good models for studying metastasis.

(135) These capsules are particularly advantageous for cultivating non-adherent suspended cells such as blood cells.

Example 7: Three-Dimensional Cultivation of Skin Tissue in Structured Alginate Capsules

(136) In order to go beyond simple co-cultivation of cells in alginate capsules, the capsules may be incubated for several days under conditions allowing cell proliferation and then the organisation of the cells into tissue(s) similar to skin tissues. Thus, the fibroblasts disseminated in the intermediate envelope may, depending on the cultivation conditions, proliferate and then synthesise molecules of the extra-cellular matrix. This organisation corresponds to the organisation of the dermis of the skin tissue. Also, the keratinocytes contained in the core of the capsule are intended to adhere to the internal surface of the intermediate envelope, in the core, proliferate until a cell monolayer is organised covering the inside of the capsule. Upon completion and under defined cultivation conditions, the keratinocytes may enter differentiation and form a cohesive stratified tissue similar to the keratinised stratified epithelium forming the epidermis of the skin tissue.

(137) Thus, it is possible to form capsules independently containing reconstructed dermis, reconstructed epidermis and reconstructed skin, and association of the dermis and of the epidermis reconstructed within a same capsule.

(138) I. Experimental Conditions

(139) IA. Collagen

(140) During the development of SkinPearls, several types of collagen were tested. Depending on the extraction methods used, the solutions of collagens are not all capable of forming a gel. As an indication, Collagen I stemming from rat tails from Gibco at 3 mg/ml and the Collagen solution stemming from bovine skin at 3 mg/ml from Sigma allow the formation of a gel.

(141) The gelling kinetics of the collagen solution is increased by the combined effect of the neutralisation of the pH and a rise in the temperature to 37 C. Indeed, collagen is soluble in an acid aqueous solution, generally of acetic acid, and the neutralisation of the pH allows regeneration of the electrostatic interactions between the collagen fibrils in order to form structured fibres within a lattice.

(142) In parallel, collagen was extracted from rat tails according to the following procedure. Briefly, two rat tails soaked beforehand in 70 ethanol were dissected and the tendons were extracted from their fascia. These tendons were then soaked in acetic acid solutions placed at 4 C. regularly stirred until solubilisation. The solutions having become thicker are then centrifuged several times in order to remove the present debris. The supernatents are kept at 4 C. until use. This collagen actually forms a gel when its pH is neutralised and it was substituted for the commercially available collagen for developing skin model capsules.

(143) Procedure for Neutralising the pH

(144) Typically, a buffer with a high ionic force is prepared and then mixed with collagen. The different buffers are added in this order. Each of the solutions, stored at 4 C., is kept in ice in order to maintain the whole at 4 C. and thereby slow down the formation of the gel.

(145) TABLE-US-00001 Buffer Concentration/pH Volume (L) (Vf = 2 mL) HEPES 200 mM/7.5 200 MEM 10X 200 DMEM 1X/7.4 256 NaOH 1M 21 Collagen 3 mg/mL/3.6 1333

(146) Marking the Collagen with Rhodamine

(147) When making three-dimensional culture capsules and more specifically the intermediate envelope forming the capsules, it is interesting to be able to observe the morphology of this layer. Indeed, many factors may have an impact on its formation and it is important to understand the parameters which control its length, its homogeneity as well as its geometry. For this, collagen was marked with a fluorophore, Rhodamine, in order to be able to produce 3D images of the capsules in confocal microscopy. Briefly, 0.2 mg of rhodamine isothiocyanate (RITC) per ml of collagen are incubated at 4 C. away from light for 48 hours. At the end of this incubation, the collagen is dialysed against an aqueous solution of 0.05M acetic acid in order to remove the excess rhodamine. The collagen marked with rhodamine is diluted in a non-marked collagen when this is necessary. This marking has not shown any negative effect on the polymerisation of the collagen.

(148) I.B Alginate

(149) The capsules are formed from an alginate solution (Protanal LF 200S, FMC) at 2% w/v and of 0.5 mM SDS filtered beforehand to 0.8 m. This solution is supplemented with a Streptomycin/Penicillin mixture at 50 U/mL and kept at 4 C. in order to limit development of microorganisms.

(150) The alginate solution used for partly or totally producing the intermediate envelope is a 1% w/v solution, without any SDS, also filtered to 0.8 m before use.

(151) I.C Buffers

(152) The capsules are formed in a 1% calcium chloride bath filtered to 0.2 m, in the presence of a drop of Tween 20 in order to modify the surface tension at the surface of the bath and to optimise the formation of round capsules. In order to remove the excess calcium ions, the capsules, immediately after formation, are rinsed in a HEPES buffer (300 mOsm, pH=7.5) prepared from a Hepes 5 solution (119.15 g Hepes, 3.75 g of NaOH tablets, water qsp 500 mL, pH adjusted to 7.5). Indeed, the buffer used should be compatible with cell survival and should not depolymerise the alginate as this is the case with phosphate or citrate buffers.

(153) I.D Cell Culture

(154) The human dermis fibroblasts come from plastic surgery waste. These cells are cultivated in cell culture flasks of 75 cm.sup.2, in the presence of DMEM (LifeTechnologies) supplemented to 10% v/v with foetal calf serum (FCS, LifeTechnologies). The passages are achieved at 80-90% confluence, 2 to 3 times a week.

(155) The keratinocytes come from adult human epidermis from plastic surgery waste. These cells are cultivated on 75 cm.sup.2 cell culture flasks coated beforehand with collagen of type I from rat tails. The keratinocytes are cultivated in an Epilife medium (LifeTechnologies) completed with Epilife Defined Growth Supplement (EDGS, LifeTechnologies). The passages are achieved at 70-80% confluence and the medium is renewed every two days.

(156) In order to achieve the passages, the culture media are removed beforehand, the cell coat is rinsed with 3 mL of 0.05% Trypsin solution (LifeTechnologies) discarded immediately and then renewed. The flasks are then placed in the incubator for a few minutes so that the cells are detached from the surface. The trypsin is then neutralised by adding 5 ml of trypsin inhibitor (LifeTechnologies). The cell suspension is then centrifuged (180 g, 8 min) and then the sediment is dispersed in 1 ml of the culture medium corresponding to the cell type. This suspension is then used for seeding new flasks, with a ratio of 3 sown flasks for 1 flask at 80% confluence.

(157) I.E Preparation of the Cells for Encapsulation

(158) The cells are treated like during a passage. Once the cell suspension is obtained, the cells are counted by means of counting cells (i.e. KovaSlide) by conducting an exclusion test with trypan blue. Only living cells not marked with trypan blue, are counted. The fibroblasts intended to be localised in the intermediate envelope of the capsules will be dispersed in the phase of the intermediate envelope, described hereafter, in an amount from 0.3 to 0.75 M of cells per ml. The keratinocytes, intended to be encapsulated in the core of the capsules are dispersed in the culture media or the biological buffer in an amount from 0.5-1.5 M of cells/ml.

(159) I.F Preparation of the Intermediate Envelope

(160) The reconstruction of a skin tissue within alginate capsules requires the possibility of compartmentation of these capsules into two areas corresponding to the two sheets making up the skin, i.e. the epidermis and the dermis. From the physico-chemical point of view, the intermediate envelope, at the interface between the alginate shell and the core of the capsule, should have a composition for which the viscosity is less than that of the phase forming the alginate envelope and greater than that of the phase intended to form the core of the capsule. Further, its composition should allow cross-linking or rapid polymerisation during the formation of the capsules in order to prevent flow phenomena and therefore disorganisation of the structure. From the biological point of view, the composition of the intermediate envelope should allow survival and growth of fibroblasts which will be disseminated in its interior, thereby regenerating a matrix similar to the dermal matrix. Finally, the composition of the intermediate envelope should allow adhesion of the keratinocytes to its surface.

(161) In order to solve the whole of these constraints, the intermediate envelope consists of 50 to 80% of Matrigel (BD Biosciences), or 50 to 80% of collagen of type I (Gibco) the pH of which is extemporaneously neutralised by adding a biological buffer and sodium hydroxide. The polymerisation of these matrix compounds being insufficiently rapid, 20 to 50% of a 4% w/v sodium alginate solution are added. Finally, the fibroblasts are dispersed in this mixture in order to obtain a cell concentration comprised between 0.3 and 0.75 M of cells/ml.

(162) I.G Formation of the Capsules

(163) The capsules are formed by fragmentation of a jet as described earlier with modifications in order to allow structuation of the capsules. For this, a 3-way injector is used (FIG. 15B). The internal phase forming the core of the capsule circulates in a first compartment 21 and is injected to the centre of the capillary C. This internal phase is cladded with the intermediate phase circulating in the compartment 31 and which will form the intermediate envelope of the capsule. The external phase circulating in the compartment 41 is intended to form the alginate shell of the capsule and is injected at the internal periphery of the capillary C. Each of the routes 21, 31 and 41 is injected into the outlet capillary C and is intended to form, from the outside to the inside, the alginate shell, the intermediate envelope and the core of the capsule. The flow rate of each of the routes is controlled by an electric syringe pump. The syringe of the internal phase is also equipped with a magnetic system allowing homogenisation of the cell suspension, without shearing, in order to avoid sedimentation of the cells and to ensure homogeneity during the handling. The fragmentation of the jet and the coalescence of the microdroplets may be controlled by piezo-electric vibration and formation of an electric field respectively. The capsules formed are covered as described earlier and then incubated at 37 C. with 5% CO.sub.2.

(164) The type of produced capsule depends on the type of cells present. The capsules for skin models contain both fibroblasts in the intermediate envelope and keratinocytes in the core. The capsules for the epidermis models only contain keratinocytes in the core while the capsules for dermis models only contain fibroblasts in the intermediate envelope.

(165) I.H Morphology of the Capsules

(166) After formation, the capsules are directly observed in suspension in the culture medium by means of an inverted microscope. The shape of the capsules (i.e. circularity) and the polydispersity of the sizes are determined by calibrating micrographs acquired with this microscope. Also, the distribution of the cells within the capsule and their localisation within the intermediate envelope and within the core of the capsules are also checked.

(167) I.I Structure of the Capsules

(168) In order to check the organisation of the different layers of the formed capsules, the marking of the intermediate envelope is achieved by substituting a portion of the collagen of the composition with collagen marked with rhodamine B, a fluorescent marker. After forming the capsules according to the method described earlier, the three-dimensional distribution of the intermediate envelope is determined by acquiring images in confocal microscopy on the whole of the capsule.

(169) I.J Characterisation of the Cell Viability and Proliferation

(170) Cell viability is determined by imaging the capsules marked with calcein and propidium iodide according to the method described earlier. The monitoring is carried out over several days in order to determine the evolution of the proliferation and of the organisation of the cells.

(171) II Results

(172) II.A Three-dimensional culture: Dermis model

(173) The capsules according to the invention were used in order to set into place models of reconstructed dermis. Thus, these capsules contain fibroblasts disseminated in a matrix which is desirably as close as possible to the dermis. The dermis is a connective tissue rich in collagen, elastin, fibronectin and glycosaminoglycan which gives it its mechanical properties. The dermis produces a supporting and nutrient tissue for the epidermis: indeed as the epidermis is avascular, the nutrients and the gas exchanges essentially come from the dermis. The main cells of the dermis are fibroblasts; disseminated in the matrix, they sustain its composition and play an important role during healing phenomena. Moreover, fibroblasts secrete cytokines and growth factors which stimulate and regulate the proliferation of keratinocytes. Also, the dermal matrix has a particular composition promoting adhesion of the keratinocytes of the epidermis and regulating their proliferation and differentiation. Cosmetically, the dermis is the target of anti-ageing treatments: the maintaining and stimulation of the synthesis of the components of the extracellular matrix is the main target of an anti-wrinkle treatment.

(174) Capsule with a Single Envelope

(175) In a first model, the fibroblasts were encapsulated within the alginate envelope making up the external envelope of the capsule. The goal, in fine, is to contain these cells in this compartment, to produce an intermediate layer with a specific composition for promoting adhesion of the keratinocytes in the core of the capsule. However, after encapsulation of the fibroblasts and the carrying out of a survival test (of the Live/Dead Calcein-AM Type and propidium iodide), it appears that all the cells are dead, none seem to survive. This toxicity is due to the presence in the alginate of sodium dodecyl sulfate (SDS) in an amount of 0.5 mM. The SDS is essential for forming the capsules and the concentration used is the lower limit below which it is no longer possible to form homogenous capsules. Now, the establishment of a cytotoxicity test demonstrates the cytotoxic effect of SDS towards fibroblasts and notably towards the 3T3 line. Thus, the inventors have shown that between 0.05 mM and 0.4 mM of SDS in the culture medium, the cell survival is between 75 and 100%, it is lowered to 38% with 0.5 mM of SDS. Consequently, this SDS concentration causes the death of more than 60% of the cells. It is not possible to remove the SDS, however the fibroblasts may be encapsulated directly in the intermediate envelope. This option provides the possibility of allowing a modification of the composition of the latter in order to approach as close as possible to that of the dermal extracellular matrix.

(176) Capsule with an Intermediate Alginate Envelope

(177) The capsules according to this alternative comprise cells only in the intermediate envelope. In a first phase, the inventors produced capsules for which the intermediate envelope consisted of a 1% w/v alginate dispersion containing fibroblasts from the 3T3 line in an amount of 0.75.Math.10.sup.6 cells/ml. Different flow rates of each of these phases forming the different layers of the capsules were tested, and once rinsed, the capsules are suspended in the full culture medium and placed in the incubator at 37 C. with 5% CO.sub.2. Cell viability was controlled by means of the Live/Dead test at D0, D2, D6 and D28 after encapsulation during this test the living cells are marked with calcein and the dead cells with propidium iodide. The acquisition of the signal of the cells through the capsules is carried out by confocal microscopy. The inventors have thus shown that a portion of the cells is dead after encapsulation, because of the stress caused by the enzymatic treatment for their detachment from the culture flasks but also by the shearing during their mixing with the alginate and during the formation of the capsules. However, a portion of the cells remains alive up to one month after encapsulation. The inventors observed that the cells proliferate but remain restricted to a few areas of the capsule in the form of spheroids or rods. This is due to the fact that the fibroblasts do not secrete any enzymes capable of lyzing the alginate thereby preventing any progression of the cells through the alginate network. The cells proliferate by filling the defects present in this matrix.

(178) Capsule with an Intermediate Collagen Envelope

(179) The same experiment was conducted by substituting the intermediate alginate envelope with a collagen layer. The capsules according to this alternative also comprise, cells only in the intermediate envelope. The collagen is one of the major components of the dermal extracellular matrix and may be degraded by the collagenases secreted by the fibroblasts. Cell viability was monitored by means of the Live/Dead test at D0, D4, D8, D15 and D19 after encapsulation during this test, the living cells are marked with calcein and the dead cells with propidium iodide. The results show that in this case, very good cell survival, spreading out of the cells and rapid proliferation covering the integrality of the capsule.

(180) The fibroblasts in this case rapidly proliferate which may be a problem during the formation of skin model capsules subsequently because the keratinocytes proliferate more slowly. Indeed, the keratinocytes take between 14 and 21 days for forming the different layers which make up the epidermis. Thus, the proliferation of the fibroblasts should be controlled. For this, the inventors tested different compositions of intermediate layer by mixing different proportions of alginate and of collagen. As an example, the inventors showed that the growth of fibroblasts within an intermediate layer half consisting of 1% w/v alginate and for the other half collagen is limited to growth in the form of sheets.

(181) Thus, cell viability was monitored by means of a Live/Dead test at D0 and D15 after encapsulation showing very good cell survival. Further, at D0, the collagen marked with RITC is visible, at D15 only the fluorescent signals of the living and dead cells are recorded demonstrating that the whole of the collagen is lyzed by the fibroblasts during their growth.

(182) As a conclusion, it is possible to control the proliferation and the distribution of the fibroblasts by modifying the composition of the intermediate envelope where the cells are localised and more particularly on the ratio between alginate and collagen.

(183) These parameters were characterised by using fibroblasts from the murine line 3T3. In order to determine the viability of this model for subsequent applications, dermis models were produced from fibroblasts of adult human dermis (HDFa). The culture of HDFa cells by the inventors in capsules with an intermediate envelope of collagen shows that the cells survived encapsulation, adhere and rapidly spread out onto the intermediate collagen envelope. Further, the cell viability test (Live/Dead test at D0 and D5 after encapsulation) shows very good cell survival, validating this model for future optimisation and development.

(184) Capsule with an Intermediate Envelope with an Alginate/Collagen Mixture

(185) The same experiment was conducted by substituting the intermediate alginate envelope with a layer comprising 25% of alginate for 75% of collagen.

(186) Morphology of the capsules: The capsules formed according to the method described above are spherical, their smooth external surface and with a diameter neighbouring 500-600 m depending on the flow rates used and on the diameter of the capillary at the outlet of the injector. The cells are absent from the external alginate layer forming the capsule and are restricted to the intermediate layer consisting of a collagen solution, for which the pH was neutralised at 0.2% m/v (75%) and of a 1% m/v alginate solution (25%), slightly more dense optically, as well as in the core of the capsules.

(187) Organisation of the intermediate layer: The intermediate layer may be viewed inside the capsules by confocal imaging of the collagen marked with a fluorophor. The acquired images show a distribution of the collagen of the structured intermediate layer in a homogenous layer through the capsules, coating the inside of the latter and delimiting a liquid core.

(188) Characterisation of the cell viability and proliferation: Cell viability was directly characterised after encapsulation but also after several days of cultivation in an incubator at 37 C. and with 5% of CO.sub.2. The majority of the cells give a green signal and therefore have active metabolism. A few cells give a red signal expressing membrane permeability, therefore a dead cell. On the whole, more than 80% of the cells are viable after encapsulation. After several days of incubation, the cells proliferate and become organised and less and less distinguishable. Proportionally, the living cells remain a majority.

(189) II.B Three-Dimensional Culture: Epidermis Model

(190) The epidermis is a keratinised stratified epithelial tissue. The keratinocytes are the main cells of this tissue. In vivo, the most basal layer, in contact with the dermo-epidermal junction (DEJ) separating the dermis from the epidermis, is the germinative layer. This layer ensures renewal of the surface layers by stimulating the proliferation of keratinocytes. Once the cells lose their contacts with the DEJ, they initiate their differentiation into corneocytes. This cell specialisation ensures strong cohesion forming the epidermal barrier mainly supported by the most external layer: the stratum corneum. Thus, in order to initiate the formation of an epidermis in the capsules according to the invention, a matrix promoting the adhesion of keratinocytes, their proliferation into a confluent layer which will initiate stratification, should be produced. The final keratinisation steps further require a stimulus related to contact with air. Thus, the epidermis models according to the invention, in the absence of the stimulus will only be able to exhibit an epidermis having not completed its terminal differentiation. The capsules according to this alternative comprise cells only in the liquid core.

(191) Collagen is a substrate of choice for adhesion of keratinocytes. In a first phase, the inventors produced capsules with an intermediate layer exclusively made of collagen, the keratinocytes of the HaCaT line are present in the liquid core. Marking the keratinocytes with calcein shows that the keratinocytes survive encapsulation (at D0), adhere to the collagen matrix (at D5) and proliferate by forming patches (at D16), sheets of cohesive cells, which gradually invade the internal surface of the capsule.

(192) Thus, when the intermediate layer exclusively consists of collagen, the structure of the collagen matrix becomes heterogeneous during its gelling inducing cell adhesion in plates of cells. Indeed, during encapsulation, the dispersion of collagen for which the pH was neutralised begins to gel thereby modifying its viscosity. Further gelling is not instantaneous as may be the cross-linking of the alginate in contact with calcium ions. Thus, collagen slips from the walls of the capsule and piles up at the bottom of the latter before totally gelling, or to a lesser extent, generates instabilities at the origin of mixtures of suspended cells in the core with the collagen.

(193) In order to find the remedy to the problems related to the gelling kinetics of collagen, the inventors mixed the collagen with alginate, this having the purpose of generating a template for the gelling of collagen. In a first phase, the inventors determined the ratios of alginate and of collagen which favoured adhesion of the keratinocytes of the HaCaT line. For this alginate gels with increasing collagen concentrations (0%, 0.05%, 0.1%, 0.15% and 0.2% m/v) and decreasing alginate concentrations (1%, 0.75%, 0.5%, 0.25% and 0% m/v) were produced, and then sown with keratinocytes. The control is the adhesion of the cells, under the same conditions to the plastic treated for cell culture. After a few hours, the cells were observed in optical microscopy. The non-adherent cells are perfectly round and refringent while the adherent cells spread out and have a polyhedril shape. FIG. 14 shows these observations.

(194) From these observations, it appears that the adhesion of the keratinocytes begins as soon as 25% v/v of collagen solution in the alginate (final collagen concentration of 0.05% m/v) while these cells are incapable of adhering to the pure alginate gel. The more the collagen content increases, the more the cells are adherent and spread out. Capsules for which the intermediate layer contained a mixture of alginate and of collagen in order to promote adhesion of the keratinocytes were produced with alginate/collagen 0.5% m/v/0.1% m/v, 0.25% m/v/0.15% m/v, 0% m/v/0.2% m/v.

(195) Marking of the keratinocytes with calcein shows that the keratinocytes survive encapsulation (at D0), adhere to the matrix (at D5) for the whole of the capsules. Nevertheless, the cells adhere and better proliferate (D16) in the presence of collagen alone.

(196) It seems that for an equal mixture of alginate solutions (1% m/v) and of collagen solutions (0.2% m/v), the cells adhere very little and a lot of them die. The centripetal cross-linking of the alginate is able to organise the collagen in a particular way and in a less favourable way to the cells comparatively with the observations made on a gel in a culture dish. In the presence of larger volumes of collagen solution (corresponding to final concentrations of collagen from 0.15 to 0.20% m/v), the keratinocytes adhere and proliferate but remain confined to isolated patches after 16 days of cultivation. The distribution of the collagen of the intermediate layer is not homogenous which does not allow the keratinocytes to migrate over the whole of the internal surface of the capsule. In order to resolve this distribution, for an equal concentration of collagen, the final concentration of alginate may be increased to 1.0% m/v allowing better stiffness of the intermediate layer and better distribution of the collagen.