THE ANATOMICAL RECORD PART A 272A:497–502 (2003) Cell and Organ Printing 2: Fusion of Cell Aggregates in Three-Dimensional Gels THOMAS BOLAND,1* VLADIMIR MIRONOV,2 ANNA GUTOWSKA,3 ELISABETH. A. ROTH,1 AND ROGER R. MARKWALD2 1 Department of Bioengineering, Clemson University, Clemson, South Carolina 2 Department of Cell Biology and Anatomy, Medical University of South Carolina, Charleston, South Carolina 3 Paciﬁc Northwest National Laboratory, Richland, Washington ABSTRACT We recently developed a cell printer (Wilson and Boland, 2003) that enables us to place cells in positions that mimic their respective positions in organs. However, this technology was limited to the printing of two-dimensional (2D) tissue constructs. Here we describe the use of thermosensitive gels to generate sequential layers for cell printing. The ability to drop cells on previously printed successive layers provides a real opportunity for the realization of three-dimensional (3D) organ printing. Organ printing will allow us to print complex 3D organs with computer-controlled, exact placing of different cell types, by a process that can be completed in several minutes. To demonstrate the feasibility of this novel technology, we showed that cell aggregates can be placed in the sequential layers of 3D gels close enough for fusion to occur. We estimated the optimum minimal thickness of the gel that can be reproducibly generated by dropping the liquid at room temperature onto a heated substrate. Then we generated cell aggregates with the corresponding (to the minimal thickness of the gel) size to ensure a direct contact between printed cell aggregates during sequential printing cycles. Finally, we demonstrated that these closely-placed cell aggregates could fuse in two types of thermosensitive 3D gels. Taken together, these data strongly support the feasibility of the proposed novel organ-printing technology. Anat Rec Part A 272A:497–502, 2003. © 2003 Wiley-Liss, Inc. Key words: organ printing; cell printing; cell aggregates; cell fusion; hydrogel Several approaches are currently being used to address the shortage of donor organs, including artiﬁcial mechanical organs, xenotransplantation (using animal organs), tissue engineering, and regenerative medicine. Some artiﬁcial organs are already on the market, but they can signiﬁcantly diminish a patient’s quality of life and may produce unwanted side effects. Xenotransplanation is a promising approach, especially when it involves the use of organs from transgenic animals with a reduced capacity to induce acute immune response after transplantation, and when it is combined with emerging methods of immunotolerance management. However, there is still great concern about the potential spreading of animal viruses (Cooper et al., 2002). Regenerative medicine, or the repair of injured or diseased organs in vivo by gene therapy or cell transplantation, is also a very promising and sophisti© 2003 WILEY-LISS, INC. cated approach. However, gene therapy and cell transplantation, as well as drug therapy, are most effective in the early stages of a disease. At terminal stages, or for injured tissue, there is still a need for organ replacement. Thus it is has been proposed that tissue engineering, Grant sponsor: NASA EPSCoR Grant; Grant number: HEDS6. *Correspondence to: Thomas Boland, Department of Bioengineering, Clemson University, Clemson, SC 29634. Fax: (864) 656-4466. Received 28 March 2002; Accepted 17 February 2003 DOI 10.1002/ar.a.10059 498 BOLAND ET AL. which is the growing of organs in vitro from a patient’s own cells, is a reasonable approach to address the shortage of donor organs. The classical tissue engineering approach involves the use of solid, rigid scaffolds from polyglycolic acid (PGA) and isolated cells (Langer and Vacanti, 1993). It is based on the premise that seeding cells in a bioreactor on porous biodegradable scaffolds will be sufﬁcient to generate organs. However, there are at least four problems with this method: 1. Cell penetration and seeding is not very effective. It proceeds on the time scale of months and is not uniform throughout the scaffold. Although signiﬁcant progress has been made in designing scaffolds that enable effective seeding and cell migration (Ma and Zhang, 2001), it is still far from optimal. 2. Organs usually consist of many cell types, and the need to place different cell types in speciﬁc positions is a very challenging technical problem in solid scaffold design. 3. The rigid, solid scaffolds made from PLA are not optimal for engineering contractile tissue, such as heart and vascular tubes. 4. The main problem with using solid scaffold seeding technology (for constructs larger than 200 ) is the absence of vascularization. The lack of vascularization has been addressed by a “rolling” approach, which has proven to be very effective for building blood vessels (L’Heureux et al., 1998), but it cannot be adapted for complex three-dimensional (3D) organs. Embedding technology is also very promising (Nerem and Seliktar, 2001), but it can only be applied to multilayer tubular organs, and also suffers from the absence of a direct placing mechanism. Recently, rapid prototyping (RP) technology has been used to fabricate physical models of hard tissues, tissue scaffolds, and custom-made tissue implant prostheses (Potamianos et al., 1998; Holck et al., 1999; Winder et al., 1999). RP technology has been used to produce novel scaffolds with controllable porosity and channel sizes, potentially allowing for vacularization (Zein et al., 2002). Although the technique is versatile and can reproduce anatomical structures (Sodian et al., 2002), its main limitation is the lack of polymers with suitable mechanical properties for soft-tissue constructs. (Hutmacher et al., 2001). This technique also has the drawback that cells must be placed in exact positions in the 3D printed scaffold. We recently developed a cell printer (Wilson and Boland, 2003) that enables us to place cells in positions that mimic their positions in an organ. The printer can put up to nine solutions of cells or polymers into a speciﬁc place by the use of specially designed software, and print two-dimensional (2D) tissue constructs. However, up to now this technology was limited to the printing of only 2D tissue constructs. A new opportunity for extending the printing technology to three dimensions is created by the use of thermo-reversible gels (Gutowska et al., 2001). Nontoxic, biodegradable, thermo-reversible gels, which are ﬂuid at 20°C and gel above 32°C, are used as a sort of “paper” on which tissue structures can be printed, and the cells are the “ink.” Successive layers could be generated just by dropping another layer of gel onto an already printed surface (see Fig. 1). Fig. 1. Principle of organ printing. [Color ﬁgure can be viewed in the online issue, which is available at www.interscience.wiley.com.] This technology, which we term “organ printing,” enables the printing of complex 3D organs, with exact placing of different cell types in 3D engineered organs, in only a few minutes. To demonstrate the feasibility of this novel technology, we showed that we can place cell aggregates in a 3D gel closely enough for fusion to occur. To do this, we estimated the minimal thickness of gel that can be reproducibly generated by dropping the liquid at room temperature onto heated substrata. Then, on the basis of this experimentally estimated parameter, we designed cell aggregates with appropriate sizes to ensure a direct contact between printed cell aggregates during sequential printing cycles. Finally, we showed that closely-placed cell aggregates could fuse in two types of 3D gels. MATERIALS AND METHODS Cell Printer Preparation A description of the printer is given elsewhere (Wilson and Boland, 2003). Sterile stainless steel needles with luer plastic hubs were kept on ice until they were ﬁlled with the polymer gel of interest. They were then screwed into the ethanol-sterilized luer ﬁttings of the print head. Care was taken to maintain a needle temperature below 4°C. It is important to maintain a low temperature when working with collagen because it will gel irreversibly and block the needles at room temperature. The print head and the printer were placed onto heated petri dishes. The dishes were placed into specially designed aluminum blocks that were kept at 36°C by a heater (VWR Scientiﬁc, Atlanta, GA). The heater, blocks, dishes, and printer were put into a sterile laminar ﬂow hood. Preparation of Thermosensitive Gels A poly[N-isopropylacrylamide-co-2-(N,N-dimethylamino)ethyl acrylate] copolymer, denoted henceforth as K-70, was used for the gel experiments. A molecular weight of 510 kD was estimated by gel permeation chromatography (GPC) with a light-scattering detector. The polymer synthesis and characterization are described elsewhere (Gutowska et al., 2001). A 10 wt % polymer solution in cold deionized water was put into an external ice bath and stirred magnetically overnight. After the polymer was completely dissolved, the pH of the solution was adjusted to 7.0 with 0.1 M NaOH, using accurate pH strips. The polymer solution 499 CELL AND ORGAN PRINTING 2 was sterilized for 30 min in an autoclave, cooled, and then redissolved. The solutions were allowed to become completely clear and were mixed well before further use. The polymer solutions were combined with an equal amount of 2⫻ cell culture medium consisting of Eagle’s minimum medium (MEM) supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic solution. The ﬁnal polymer concentration was not lower than 4 –5 wt %. To visualize the polymer layers, half of the solutions were dyed by adding 30 l of 1% trypan blue dye to each milliliter of polymer solution. Preparation of Collagen Gels Collagen gels were prepared according to the method of Storm and Michalopoulous (1982), with the exception that lyophilized calf skin collagen rather than rat tail collagen was used. A 3.33 mg/ml collagen solution was prepared by slowly dissolving the collagen in acetic acid (1%) at 0°C. The gel was prepared by adding 156.5 l of sterile, ﬁltered 10⫻ DMEM and 30 l of sterile 0.34N NaOH to each milliliter of collagen solution. The resulting solutions were kept on ice until they were used. Estimation of Gel Thickness The thickness of the gel layers was evaluated by depositing a series of drops onto the heated petri dishes. Each drop was allowed to gel before a new, smaller drop was placed on top of the gel. The series of staggered gel layers thus created was placed under a microscope and visualized. The movement (in microns) of the microscope objective was measured as the focal plane shifted from one layer to the next. Preparation of Cell Aggregates Bovine aortal endothelial cells (American Type Culture Collection, Manassas, VA) (passage 10) were expanded in T25 culture ﬂasks in the presence of MEM supplemented with 10% FBS and 1% antibiotic solution in a 5% CO2 incubator maintained at 36°C. The media were changed at day 1 and subsequently every other day. After growing to a conﬂuent monolayer, the cells were washed by replacing the media with isotonic phosphate-buffered saline (PBS). They were then incubated in a PBS solution containing a millimolar solution of the tripeptide arginine-glycine-aspartic acid (RGD) and agitated for 30 – 45 min in the incubator. This caused the cells attached to the edges of the ﬂask to dislodge. Further mechanical scraping of the ﬂask with a sterile glass pipette caused the cells to detach in aggregates. The aggregates were collected, centrifuged at 1,000 rpm for 5 min, and resuspended in 0.5 ml of media. The aggregates were then reseeded on collagen and K-70 gels. Live/Dead Assay The viability of the cells was assessed with a commercially available live/dead assay (Molecular Probes Inc., Eugene, OR). The samples were rinsed with PBS, and incubated for 30 min in a solution of calcein AM and ethidium homodimer-1 in PBS according to the manufacturer’s protocol. Fluorescence was observed in an inverted epiﬂuorescent microscope (Nikon Diaphot 300, Nikon Inc., Melville, NY) using a DAPI/FITC/TRITC triple-band ﬁlter. RESULTS Alternate layers of clear and tryan blue K-70 gel were analyzed under a microscope. Figure 2 shows a top view of the layered gel. The thickness of the layers varied between 200 and 500 . During the addition of the polymer solutions to the already gelled layers, it was obvious that only minimal (if any) mixing of the layers occurred. In the gelled state the polymer is hydrophobic, and one could clearly observe that the liquid solution drops did not spread out on the gel. The size of the aggregates was also measured. We obtained aggregates with an average diameter of 540 ⫾ 183 m (n ⫽ 18). Figure 3 shows a representative image of the endothelial aggregates used in this study. Directed Fusion of Closely-Placed Cell Aggregates To prove that a directed fusion of aggregates could be achieved, we performed an additional series of experiments using collagen and K-70 gels. Our assumption was that directed fusion of closely-placed cell aggregates should be observed with higher affectivity on the collagen gel than the thermo-reversible gel. It was shown that placing cell aggregates in close opposition on the surface of collagen type 1 gel resulted in the fusion of adjacent cell aggregates, with a sequential formation of elongated rodlike tissue constructs (see Fig. 4). Similar results were produced by placing the cell aggregates on the surface of collagen, and sequentially covering the aggregates with a second collagen layer. This proved that fusion of cell aggregates occurs not only on the surface but also within 3D collagen gel. Finally, we attempted to reproduce the same results using the thermo-reversible gel. Although the overall effectiveness of the thermo-reversible gels in promoting cell aggregate fusion was not equal to that of the collagen type 1 gel, we were able to show that fusion does occur in these gels. To show that directed fusion can occur on patterned gels, we loaded the collagen gel into the print head of the printer. The gels were deposited in ring patterns about 1 cm in diameter and 500 wide. After gelling occurred, the BAEC aggregates were added to the gel. Observation after 24 hr revealed that the cells were spread throughout the gel. A live/dead assay image is shown in Figure 5, which shows that the cells within the gel were alive, whereas most of the cells that migrated out of the gel were not viable. DISCUSSION Directed Cell Placing as the Main Problem of Tissue Engineering A very important task of tissue engineering is to put certain cell types in exact places. This cannot be achieved by traditional bioreactor-based cell-seeding technology using porous biodegradable scaffolds (even sophisticated ones created with prototype printing). Other techniques, such as rolling and embedding, are as yet not suitable for engineering complex multicellular 3D organs. Cell-printing technology using thermo-reversible gel provides (at least theoretically) a unique opportunity to solve this most difﬁcult tissue-engineering task, and opens the way for mass organ printing. This may eventually lead to a cost-effective solution to the problem of organ 500 BOLAND ET AL. Fig. 2. Top view of seven alternate layers of clear and tryan bluedyed K-70 gel on a collagen-coated dish. The edges between the layers reveal the amount of mixing that occurs when a cold drop liquiﬁes some of the gel before it is gelled itself. Optimizing gelling kinetics, drop size, and deposition rate may minimize this effect. Fig. 3. Image of a single endothelial cell aggregate of ⬃670 m diameter. shortage. To assess the feasibility of this exciting technology, we attempted to remove some of the most obvious technological barriers to its use. Estimating the Optimum Minimal Thickness of the Gel Layer The presence of the gel layer between the cell aggregates serves two purposes: 1) it provides mechanical strength and stability to the construct, and 2) it serves as a drug-delivery device. For example, porous PLA/PGA/ PEG polymeric microspheres with controlled degradation times could be put loaded into the printer for controlled delivery of growth factors. Using copolymers of different Fig. 4. Cell aggregate fusion on the (A) collagen gel and (B) K-70 gel. Fig. 5. Cell aggregate fusion on a printed collagen gel ring. The image obtained on an epiﬂuorescent microscope shows fused aggregates and cells that migrated across the gel. Cells that migrated toward the center of the ring (in red) proved nonviable, possibly because they became dry outside the gel. PLA/PGA compositions, microspheres with a range of degradation times could be obtained. Porous microspheres could be loaded with desired growth factors and dispersed within an injectable gel matrix. The release proﬁle of bioactive agents could be adjusted to the required tissuespeciﬁc spatial/temporal pattern of expression. Generation of Sequential Thermo-Reversible Gel Layers Apart from tissue-speciﬁc cell–matrix interactions, the following aspects must be considered: gelation kinetics, CELL AND ORGAN PRINTING 2 matrix resorption rate, possible toxicity of degradation products and their elimination routes, and possible interference of the gel matrix with histogenesis. There are several possible mechanisms that lead to in situ gel formation: gelation in response to temperature change (thermal gelation), pH change, ionic cross-linking, or solvent exchange (Guenet, 1992). The kinetics of gelation are directly affected by the mechanism: thermal gelation, which is rate-limited by heat transfer, is faster then ionic cross-linking and pH-triggered gelation, which are rate-limited by mass transfer. The time it takes for the gel to form will have an effect on cell spacing and distribution within the printed gel matrix. Gels that form in response to temperature change may offer speciﬁc advantages related to fast gelation kinetics. The thickness of the gel layers therefore depends on the kinetics of the gel, as well as on the drop size, which is governed by surface tension and needle diameter. To increase the efﬁciency of the technology, the gel thickness will need to be adjusted to match the size of the aggregates. Using a needle with the same internal diameter as the aggregates appears to give satisfactory results. However, as indicated in Figure 2, some distortions in the layers (possibly due to a melting process) are observed. Further improvements in controlling the aggregate size distribution would also require a more rigorous model for gel thickness. This model would need to take into account gelation kinetics, surface tension, and interfacial forces between the liquid and gel states of the solution. Identiﬁcation of the Optimum Size for Cell Aggregates The optimum size for cell aggregates is deﬁned as a diameter that is small enough (usually ⬍1 mm) to allow centrally positioned aggregate cells to survive, and a magnitude that is equal to the thickness of the successively printed gel layers. The second restraint is much more important because it determines the overall feasibility of the technique, whereas apoptosis can be prevented either by using telomerized cells or cells transfected with blc-2 (Yang et al., 2001; Nor et al., 2001). Based on estimations of the optimum minimal thickness, the diameter of the cell aggregates must be no smaller than 600 m. This does not exclude the possibility of variability in cell aggregate size, but it deﬁnitely places a well-deﬁned limit on the smallest technologically acceptable diameter. Role of Gel and Close Positioning in Promoting the Fusion of Cell Aggregates The fusion of cell aggregates is a spontaneous phenomenon in aggregate suspension. Recently, important insights have been gained regarding the possible mechanism of aggregate fusion in the presence of extracellular matrix (Ryan et al., 2001). It was demonstrated that competition exists between cell-to-cell forces vs. cell-to-substrate forces. If the cell-to-substrate force is stronger than the cell-to-cell force, the outcome will be a monolayer. If the force of cell-to-cell contacts is predominant, the cells will form aggregates. This principle was used by us when we generated cell aggregates. The absence of a cell adhesive substrate forced the cells to aggregate in order to interact with each other. Our data demonstrate that collagen gels are more aggregate-friendly and are more “fusogenic” substrates com- 501 pared to thermo-reversible gels. This is a more complex situation than the single-cell or monolayer case described above. It is possible that the tethering effect of collagen mediated by the RGD tripeptide is responsible for keeping cell aggregates together, and for their sequential fusion. We are exploring ways to optimize the thermo-reversible gel by incorporating RGD or other extracellular matrix analogs (Mann et al., 2001; Grifﬁth and Naughton, 2002) without compromising its thermo-reversible properties. Toward Organ Printing Keeping in mind that the adhesiveness of gels for cells can be modiﬁed, our data show that cell aggregate fusion can be achieved in the thermo-reversible gel. Our data also demonstrate that we can reproducibly generate sequential layers of thermo-reversible gels of a thickness corresponding to the cell aggregate diameter. This thickness allows cell aggregates to be laced in closely enough to fuse. Both 3D collagen and thermo-reversible gels were shown to allow fusion of closely-placed cell aggregates. Aggregate fusion causes the 3D structures to shrink somewhat, an effect that will need to be considered in the design of organ blueprints. However, we can also assume that additional extracellular matrix will be formed by the cells, thus providing additional glue for cell aggregate fusion and serving as a sort of tethering device for promoting the self-assembly of printed tissue constructs. Furthermore, this additional extracellular matrix may act as a buffer between adjacent structures, and thus be more forgiving with respect to the exact tolerances of a blueprint. In addition, the use of gel with RGD ligands may enhance adhesion. Alternatively, aggregates trapped inside gel drops can be printed. Because closely placed drops fuse, we expect the aggregates to fuse as well. However, the time frame of cell attachment is very important, since it determines the speed at which successive layers can be printed. Soft-tissue adhesives such as cyanoacrylate esters, ﬁbrin sealant, and gelatin-resorcinol-formaldehyde glues, or other bioadhesives, could dramatically prevent the constructs from being washed off during successive printing cycles. The addition of a corresponding growth factor in slow-release microspheres could accelerate and direct this process. The programmed release of different growth factors in various concentrations and sequences is also technically possible. Finally, the thermo-reversible gel can be optimized or functionalized by the addition of extracellular matrix analogs. This would promote cell adhesion, and result in better cell survival as well as more effective cell aggregate fusion. Poor cell survival in the printed aggregates is a possible drawback to the described technology. We are currently exploring different ways to avoid apoptosis and necrosis inside the aggregates, such as adding survival factors (e.g., basic ﬁbroblast growth factor), using transient genetic modiﬁcations of cells with antiapoptotic (e.g., bcl-2 and telomerase), and blocking apoptotic pathways. It is well known that the survival of cells in cell aggregates can be seen as a result of cell contact-mediated survival stimuli. In this respect, we expect the survival rate of aggregated cells to be higher than that of single printed cells. Furthermore, this technology can tolerate a small percentage of cell death. The goals of this study were to demonstrate the feasibility of organ-printing technology, solve some important 502 BOLAND ET AL. technical problems, and eliminate some critical technological barriers. The actual printing of 3D tissue constructs will be the subject of another paper, in which we will also address issues of cell survival, tissue perfusion, and vascularization. CONCLUSIONS Taken together, our data strongly indicate that the proposed organ-printing technology is feasible using the originally-designed cell printer and thermo-reversible gel. LITERATURE CITED Cooper DK, Gollackner B, Sachs DH. 2002. Will the pig solve the transplantation backlog? Annu Rev Med 53:133–147. Grifﬁth LG, Naughton G. 2002. Tissue engineering— current challenges and expanding opportunities. Science 295:1009 –1014. Guenet JM. 1992. Thermoreversible gelation of polymers and biopolymers. London: Academic Press. Gutowska A, Jeong B, Jasionowski M. 2001. Injectable gels for tissue engineering. Anat Rec 263:342–349. L’Heureux N, Paquet S, Labbe R, Germain L, Auger FA. 1998. A completely biological tissue-engineered human blood vessel. FASEB J 12:47–56. Holck DE, Boyd EM, Ng J, Mauffray RO. 1999. Beneﬁts of stereolithography in orbital reconstruction. Ophthalmology 106: 1214僒1218. Hutmacher DW, Schantz T, Zein I, Ng KW, Teoh SH, Tan KC. 2001. Mechanical properties and cell cultural response of polycaprolactone scaffolds designed and fabricated via fused deposition modeling. J Biomed Mater Res 55:203–216. Langer R, Vacanti JP. 1993. Tissue engineering. Science 260:920 – 926. Ma PX, Zhang R. 2001. Microtubular architecture of biodegradable polymer scaffolds. J Biomed Mater Res 56:469 – 477. Mann BK, Gobin AS, Tsai AT, Schmedlen RH, West JL. 2001. Smooth muscle cell growth in photopolymerized hydrogels with cell adhesive and proteolytically degradable domains: synthetic ECM analogs for tissue engineering. Biomaterials 22:3045–3051. Nerem RM, Seliktar D. 2001. Vascular tissue engineering. Annu Rev Biomed Eng 3:225–243. Nor JE, Christensen J, Liu J, Peters M, Mooney DJ, Strieter RM, Polverini PJ. 2001. Up-Regulation of Bcl-2 in microvascular endothelial cells enhances intratumoral angiogenesis and accelerates tumor growth. Cancer Res 61:2183–2188. Potamianos P, Amis AA, Forester AJ, McGurk M, Bircher M. 1998. Rapid prototyping for orthopedic surgery. Proc Inst Mech Eng 212: 383僒393. Ryan PL, Foty RA, Kohn J, Steinberg MS. 2001. Tissue spreading on implantable substrates is a competitive outcome of cell-cell vs. cell-substratum adhesivity. Proc Natl Acad Sci USA 98:4323– 4327. Sodian R, Loebe M, Hein A, Martin DP, Hoerstrup SP, Potapov EV, Hausmann H, Lueth T, Hetzer R. 2002. Application of stereolithography for scaffold fabrication for tissue engineered heart valves. ASAIO J 48:12–16. Strom SC, Michalopoulos G. 1982. Collagen as a substrate for cell growth and differentiation. Methods Enzymol 82:544 –555. Wilson C, Boland T. 2001. New inexpensive technologies for two- and three-dimensional patterning of proteins on solid supports. Trans Soc Biomater 24:190. Wilson C, Boland T. 2003. Cell and organ printing 1: protein and cell printers. Anat Rec 272A:492– 497. Winder J, Cooke RS, Gary J, Fannin T, Fegans T. 1999. Medical rapid prototyping and 3-D CT in the manufacture of custom made cranial titanium plates. J Med Eng Technol 32:26 –28. Yang J, Nagavarapu U, Relloma K, Sjaastad MD, Moss WC, Passaniti A, Herron GS. 2001. Telomerized human microvasculature is functional in vivo. Nat Biotechnol 19:219 –224. Zein I, Hutmacher DW, Tan KC, Teoh SH. 2002. Fused deposition modeling of novel scaffold architectures for tissue engineering applications. Biomaterials 23:1169 –1185.