**5. Effect of cell density on the rate of glycosaminoglycan accumulation by chondrocytes**

Glycosaminoglycan (GAG) accumulation in constructs is dependent on the rate of GAG production per cell and on the cell density. It seems intuitive, therefore, that increasing cell density should increase rate of GAG deposition, as indeed has been shown in several studies (Almarza & Athanasiou, 2005, Mauck et al., 2002, Mauck et al., 2003b, Saini & Wick, 2003, Williams et al., 2005, Kobayashi et al., 2008). However, it is apparent from these studies that GAG accumulation in the construct does not increase in proportion to cell density and, indeed, GAG production per cell appears to fall at high cell densities.

We used alginate gels in the form of beads as a model system (Fig.9). Cells were isolated from bovine metacarpal phalangeal joints. They were cultured in alginate beads in DMEM contain‐ ing 6% FBS under 21% O2 at cell densities from 1-33 million cells/ml. The amount of GAG accumulated in a typical culture of bovine articular chondrocytes increased with time in culture. After 7 days a bimodal response was evident with the concentration of GAG accu‐ mulated rising as cell density was increased from 1-10.4 million cells/ml and then falling gradually as cell density was further increased (Fig.10A). However the GAG production per million cells fell as cell density was increased (Fig.10B).

Next, we mainly examined two initial seeding cell densities viz. 4 million cells/ml (Fig.9A) and 25 million cells/ml (Fig.9B); these represent cell densities often used in alginate beads and found in vivo in the cell density in adult cartilage from the bovine metacarpal-phalangeal joint.

**Figure 9.** Three-dimensional cell culture system of alginate beads. (A) 4 x 106 cell/ml, (B) 25 x 106 cell/ml

Significantly more GAG was accumulated by cells cultured at high (25 million cells/ml) than low (4 million cells/ml) densities and in agreement with results shown in Fig.11, GAG accumulated also increased with time in culture. At 4 million cells/ml, the concentration of GAG in the bead reached 520.9 ± 62.4μg/ml in 5 days. These concentrations could be increased to 1297.2 ± 115.2 μg/ml by raising cell density to 25 million cells/ml (Fig.11A). The increase in amount of GAG accumulated was not directly proportional to increase in cell density; although the beads at high cell density contained more than 6 times as many cells as those at low cell density, they only produced only 2-3 times as much total GAG. After 5 days culture at 4 million cells/ml, GAG accumulation per cell was 166.3 ± 38.9μg GAG/million cells (Fig.11B). These amounts fell to 70.9 ± 23.9 μg/million cells when cell density was increased to 25 million cells/ ml. Thus, cells cultured at low density were more active and accumulated significantly more GAG per cell than cells cultured at high density. Evidence of greater cellular activity for cells cultured at low cell density was also seen from measurements of lactate production per live cell; lactate production was significantly higher for cells cultured at low density than for those Importance of Extracellular Environment for Regenerative Medicine and Tissue Engineering of Cartilagious Tissue http://dx.doi.org/10.5772/55566 557

GAG accumulation in the construct does not increase in proportion to cell density and, indeed,

We used alginate gels in the form of beads as a model system (Fig.9). Cells were isolated from bovine metacarpal phalangeal joints. They were cultured in alginate beads in DMEM contain‐ ing 6% FBS under 21% O2 at cell densities from 1-33 million cells/ml. The amount of GAG accumulated in a typical culture of bovine articular chondrocytes increased with time in culture. After 7 days a bimodal response was evident with the concentration of GAG accu‐ mulated rising as cell density was increased from 1-10.4 million cells/ml and then falling gradually as cell density was further increased (Fig.10A). However the GAG production per

Next, we mainly examined two initial seeding cell densities viz. 4 million cells/ml (Fig.9A) and 25 million cells/ml (Fig.9B); these represent cell densities often used in alginate beads and found in vivo in the cell density in adult cartilage from the bovine metacarpal-phalangeal joint.

**Figure 9.** Three-dimensional cell culture system of alginate beads. (A) 4 x 106 cell/ml, (B) 25 x 106 cell/ml

Significantly more GAG was accumulated by cells cultured at high (25 million cells/ml) than low (4 million cells/ml) densities and in agreement with results shown in Fig.11, GAG accumulated also increased with time in culture. At 4 million cells/ml, the concentration of GAG in the bead reached 520.9 ± 62.4μg/ml in 5 days. These concentrations could be increased to 1297.2 ± 115.2 μg/ml by raising cell density to 25 million cells/ml (Fig.11A). The increase in amount of GAG accumulated was not directly proportional to increase in cell density; although the beads at high cell density contained more than 6 times as many cells as those at low cell density, they only produced only 2-3 times as much total GAG. After 5 days culture at 4 million cells/ml, GAG accumulation per cell was 166.3 ± 38.9μg GAG/million cells (Fig.11B). These amounts fell to 70.9 ± 23.9 μg/million cells when cell density was increased to 25 million cells/ ml. Thus, cells cultured at low density were more active and accumulated significantly more GAG per cell than cells cultured at high density. Evidence of greater cellular activity for cells cultured at low cell density was also seen from measurements of lactate production per live cell; lactate production was significantly higher for cells cultured at low density than for those

GAG production per cell appears to fall at high cell densities.

556 Regenerative Medicine and Tissue Engineering

million cells fell as cell density was increased (Fig.10B).

**Figure 10.** Typical results showing effect of cell density on GAG deposition (A) and on GAG accumulation per million cells (B) by articular chondrocytes. Cells were isolated, encapsulated in alginate beads at cell densities ranging from 1 to 33 million cells/ml. Beads were cultured for 7 days at 5 wells/bead in 2 ml medium, 2 wells for each cell density and cultured for 7 days in DMEM containing 6% serum. Beads were then dissociated for cell counting and assay of total GAGs.

**Figure 11.** Effect of cell density on GAG concentration (A) and GAG accumulation per million cells (B) by articular chondrocytes after 2 days and 5 days in culture. Cells were encapsulated in alginate beads, cultured in DMEM with 6% serum under air and GAG concentration and cell density measured after 2 and 5 days culture. These figures give pooled data for the two representative cell densities from 3 separate experiments. Values are mean ± standard error. \*, : Significant difference (P<0.05) between the high cell density (25 million cells/ml) and the low cell density (4 million cells/ml) using non-paired t test.

**Figure 12.** Effect of cell density on lactate production rate (A) and 35S-sulphate incorporation rate (B). (A) Cells were cultured under standard conditions in beads containing 4 and 25 million cells/ ml (1.0ml medium, 5 beads/well) for up to 7 days, with complete medium change daily. Representative beads were dissociated for cell counting and viable cell density/bead recorded. Lactate in the medium was measured at days 1,5 and 7, after 24 hours culture and rates per million cells/24 hrs reported. High cell density lead to a fall in cellular metabolism (\*,\*\*,\*\*\*: P<0.05, Paired t test between high [25 million cells/ml] and low cell density [4 million cells/ml]). Lactate production rate fall with time in culture (#: P<0.05, 2 way ANOVA with repeated measures among 1, 5 and 7 days).(B) At days 1,5 and 7, tracer sul‐ phate was added to the fresh medium of 3 wells, the beads were cultured in the radioactive solution for 4 hours, the beads dissociated and cell density and sulphate incorporation measured (Fig 3B). Results are given as means ± s.e.m of 3 independent experiments. Sulphate incorporation rates fall with increase in cell density (\*,\*\*,\*\*\*: P<0.05, Paired t test between high [25 million cells/ml] and low cell density [4 million cells/ml]) and with time in culture (#: P<0.05, 2 way ANOVA with repeated measures among 1, 5 and 7 days).

cultured at high density (Fig 12A). Lactate production also decreased with time in culture, more rapidly at high than at low cell densities. The rate of sulphate incorporation per live cell was also greater at low than at high cell densities (Fig.12B), though the difference was less marked than that seen in Fig.12A; sulphate incorporation fell more steeply than lactate production with time in culture.

The change in percentage of live and dead cells with time in culture at the periphery and centre of beads is shown in Fig.13 for cells cultured at low (4 million cells/ml) and high cell densities (25 million cells/ml), respectively. For cells cultured at 4 million cells/ml, 100% of the cells were viable at the both the periphery (Fig 13A) and in the centre (Fig 13B). It can be seen that by day 2 of culture at high cell densities, while almost all the cells at the periphery were alive (Fig. 13C,E), 30 percent of the cells in the bead centre were dead (Fig.13D,E). Similar percentages were dead at day 5 of culture, suggesting the profile of viable cells across the bead was established early in culture.

From sections of beads cultured for 5 days at 4 and 25 million cells/ml and then stained with safranine O to visualise the sulphated GAGs accumulated (Fig.14A-D), there was a noticea‐ ble difference between staining at the bead periphery (Fig.14A,C) and the bead centre (Fig. 14B,D). From densitometric measurements, GAG accumulated at centre of beads cultured at 25 million cells per ml was only 60-70% of that accumulated at the periphery (Fig.14E). Less staining was seen in beads cultured at 4 million cells per ml as seen also from chemical

Importance of Extracellular Environment for Regenerative Medicine and Tissue Engineering of Cartilagious Tissue http://dx.doi.org/10.5772/55566 559

**Figure 13.** Effect of cell density on cell viability under conforcal microscope.This shows the variation of cell viability at the edge and centre of beads with time and cell density. Cell viability was determined using a live/dead assay kit; live cells (green) and dead cells (red) were counted manually. Results are means and s.e.ms of percentage of viable cell from 4 representative beads. Figs 13A and 13B shows the periphery and central region respectively of a typical bead cultured at 4 million cells/ml after 5 days. Figs 13C and 13D shows the periphery and central region of a bead cultured at 25 million cells/ml after 5 days. Fig 13E shows the variation of cell viability with region (edge versus centre). At high cell density (25 million cells/ml), cell viability is lower in the centre than at the edge (+: P=0.977, ++: P=0.893, \*, \*\*: P<0.05, 2 way ANOVA with repeated measures between edge and centre).

cultured at high density (Fig 12A). Lactate production also decreased with time in culture, more rapidly at high than at low cell densities. The rate of sulphate incorporation per live cell was also greater at low than at high cell densities (Fig.12B), though the difference was less marked than that seen in Fig.12A; sulphate incorporation fell more steeply than lactate

**Figure 12.** Effect of cell density on lactate production rate (A) and 35S-sulphate incorporation rate (B). (A) Cells were cultured under standard conditions in beads containing 4 and 25 million cells/ ml (1.0ml medium, 5 beads/well) for up to 7 days, with complete medium change daily. Representative beads were dissociated for cell counting and viable cell density/bead recorded. Lactate in the medium was measured at days 1,5 and 7, after 24 hours culture and rates per million cells/24 hrs reported. High cell density lead to a fall in cellular metabolism (\*,\*\*,\*\*\*: P<0.05, Paired t test between high [25 million cells/ml] and low cell density [4 million cells/ml]). Lactate production rate fall with time in culture (#: P<0.05, 2 way ANOVA with repeated measures among 1, 5 and 7 days).(B) At days 1,5 and 7, tracer sul‐ phate was added to the fresh medium of 3 wells, the beads were cultured in the radioactive solution for 4 hours, the beads dissociated and cell density and sulphate incorporation measured (Fig 3B). Results are given as means ± s.e.m of 3 independent experiments. Sulphate incorporation rates fall with increase in cell density (\*,\*\*,\*\*\*: P<0.05, Paired t test between high [25 million cells/ml] and low cell density [4 million cells/ml]) and with time in culture (#: P<0.05, 2

The change in percentage of live and dead cells with time in culture at the periphery and centre of beads is shown in Fig.13 for cells cultured at low (4 million cells/ml) and high cell densities (25 million cells/ml), respectively. For cells cultured at 4 million cells/ml, 100% of the cells were viable at the both the periphery (Fig 13A) and in the centre (Fig 13B). It can be seen that by day 2 of culture at high cell densities, while almost all the cells at the periphery were alive (Fig. 13C,E), 30 percent of the cells in the bead centre were dead (Fig.13D,E). Similar percentages were dead at day 5 of culture, suggesting the profile of viable cells across the bead was

From sections of beads cultured for 5 days at 4 and 25 million cells/ml and then stained with safranine O to visualise the sulphated GAGs accumulated (Fig.14A-D), there was a noticea‐ ble difference between staining at the bead periphery (Fig.14A,C) and the bead centre (Fig. 14B,D). From densitometric measurements, GAG accumulated at centre of beads cultured at 25 million cells per ml was only 60-70% of that accumulated at the periphery (Fig.14E). Less staining was seen in beads cultured at 4 million cells per ml as seen also from chemical

production with time in culture.

558 Regenerative Medicine and Tissue Engineering

way ANOVA with repeated measures among 1, 5 and 7 days).

established early in culture.

analysis. At this low density however there was no significant profile of GAG accumula‐ tion across the bead, with the amount accumulated in the centre similar to that accumulat‐ ed at the periphery.

At low cell density, transmission electron micrographs indicated that all cells appeared viable and active (Fig.15A). Chondrocytes cultured at high cell density appeared viable at the bead periphery (Fig.15B). However cells undergoing apoptosis were seen in the centre; the cells and nuclei were reduced in size and chromatin condensation was visible in the nuclei (Fig. 15C,D). These results are in agreement with those of others who have found regions of cell death in the center of constructs or even of microsphere aggregates (Martin et al., 1999b, Mercier et al., 2004, Obradovic et al., 1999), and that glycosaminoglycan accumulation may highest at the construct peripheries. In addition, others have also found that increasing cell density or cell number does not necessarily increase matrix accumulation (Mercier et al., 2004).

These avascular constructs, unless experimentally perfused, rely on diffusion for supply of nutrients to the cells (Obradovic et al., 2000) simulating the condition seen in cartilaginous tissue. In avascular tissues and in constructs, there are steep gradients of oxygen and other nutrients between the surface and center of the tissue or constructs (Kellner et al., 2002, Malda

**Figure 14.** Effect of cell density on GAG deposition by Safranine O staining after 5 days culture.Fig 4A shows a 20mi‐ cron section through a typical bead of chondrocytes cultured at 25 million cells/ml (A,B) and at 4 million cells/ml (C,D) cultured for 5 days. Images at the periphery (A,C) and at the centre (B,D) were captured digitally and the GAG around cells was quantified using image-analysis. Results are reported as the fraction of stained area in the peripheral and central regions of the beads and data normalized to results at 25 million cells/ml (E). At high cell density, the area of the staining was higher at the edge of the bead (\*: P<0.05, Paired t test between edge and centre). At low cell density, however, there was no significant profile of GAG accumulation between edge and centre (P=0.712, Paired t test be‐ tween edge and centre).

et al., 2004). The steepness of these gradients, and hence the nutrient concentrations in the center of the construct, depend not only on the geometry and properties of the tissue or construct but also on the cell density and the cellular activity (Haselgrove et al., 1993, Zhou et al., 2004, Soukane et al., 2005). Thus, in any particular construct or tissue, an increase in cell density will lead to a corresponding fall in the concentration of nutrients such as oxygen and glucose, and an increase of metabolic by-products such as lactic acid (Zhou et al., 2004), leading, once cell density has risen sufficiently, to a fall in rates of cell metabolism and glycosaminoglycan synthesis (Gray et al., 1988, Ysart & Mason, 1994). If cell density is sufficiently great, oxygen and glucose concentrations and pH levels can fall to levels which can no longer sustain viable cells (Horner & Urban, 2001) leading to the necrotic region in the construct center. Diffusional nutrient transport is thus a limitation on the number of viable and active cells which can be maintained in any construct or tissue; indeed, viable cell density is inversely related to diffusion distance both in disc and in constructs (Horner & Urban, 2001, Stockwell, 1971).

#### **6. Physical limitations to biological repair and tissue engineering**

The interrelationships between cell density, cell viability and activity, and diffusion distance resulting from nutrient supply constraints, limit the rate at which GAG can be accumulated in three-dimensional constructs. GAG accumulation depends on GAG production per cell and on cell density. At low cell densities, cells may be functioning optimally but the low cell density

Importance of Extracellular Environment for Regenerative Medicine and Tissue Engineering of Cartilagious Tissue http://dx.doi.org/10.5772/55566 561

**Figure 15.** Electron micrographs of central and peripheral articular chondrocytes cultured at low (A) and high cell density (B-D). These pictures show representative cells from the central and peripheral regions of beads cultured at low (4 million cells/ml) and high (25 million cells/ml) for 5 days. (A) Central region, low cell density, cells appear nor‐ mal. (B-D) High cell density (25 million cells/ml). (B) Bead periphery, some cells appear normal. (C) Central region showing cells undergoing apoptosis (arrow). The cells and nuclei were reduced in size and chromatin condensation was seen in the nuclei in comparison of the cells in the periphery. (D) High power magnification of apototic nucleus pulposus cell.

et al., 2004). The steepness of these gradients, and hence the nutrient concentrations in the center of the construct, depend not only on the geometry and properties of the tissue or construct but also on the cell density and the cellular activity (Haselgrove et al., 1993, Zhou et al., 2004, Soukane et al., 2005). Thus, in any particular construct or tissue, an increase in cell density will lead to a corresponding fall in the concentration of nutrients such as oxygen and glucose, and an increase of metabolic by-products such as lactic acid (Zhou et al., 2004), leading, once cell density has risen sufficiently, to a fall in rates of cell metabolism and glycosaminoglycan synthesis (Gray et al., 1988, Ysart & Mason, 1994). If cell density is sufficiently great, oxygen and glucose concentrations and pH levels can fall to levels which can no longer sustain viable cells (Horner & Urban, 2001) leading to the necrotic region in the construct center. Diffusional nutrient transport is thus a limitation on the number of viable and active cells which can be maintained in any construct or tissue; indeed, viable cell density is inversely related to diffusion distance both in disc and in constructs (Horner &

**Figure 14.** Effect of cell density on GAG deposition by Safranine O staining after 5 days culture.Fig 4A shows a 20mi‐ cron section through a typical bead of chondrocytes cultured at 25 million cells/ml (A,B) and at 4 million cells/ml (C,D) cultured for 5 days. Images at the periphery (A,C) and at the centre (B,D) were captured digitally and the GAG around cells was quantified using image-analysis. Results are reported as the fraction of stained area in the peripheral and central regions of the beads and data normalized to results at 25 million cells/ml (E). At high cell density, the area of the staining was higher at the edge of the bead (\*: P<0.05, Paired t test between edge and centre). At low cell density, however, there was no significant profile of GAG accumulation between edge and centre (P=0.712, Paired t test be‐

**6. Physical limitations to biological repair and tissue engineering**

The interrelationships between cell density, cell viability and activity, and diffusion distance resulting from nutrient supply constraints, limit the rate at which GAG can be accumulated in three-dimensional constructs. GAG accumulation depends on GAG production per cell and on cell density. At low cell densities, cells may be functioning optimally but the low cell density

Urban, 2001, Stockwell, 1971).

tween edge and centre).

560 Regenerative Medicine and Tissue Engineering

limits the rate of GAG accumulation. At high cell densities, more GAG is deposited at least initially, but nutrient gradients particularly in the center of constructs, reduce the rate of GAG deposition per cell and may even lead to a fall in cell number if cells die. GAG accumulation thus appears necessarily slow, and the general finding that cultures of >7 months are required to achieve concentrations of GAG similar to those seen in vivo may not be easily overcome (Kellner et al., 2002, Roughley, 2004). The different maneuvers which have been tried to increase GAG production all have limitations. An increase in GAG production rate per cell can be induced by addition of growth factors, by providing mechanical or ultrasound stimu‐ lation or through alterations to scaffold properties (Blunk et al., 2002, Richmon et al., 2005, van der Kraan et al., 2002, Kuo & Lin, 2006), but the relative increase which can be achieved is limited (usually two–three fold under optimal conditions) and the consequent increase in metabolic demand can lead to a fall in pH in the construct center (Wu, M.H. et al., 2008) and thus severely limit growth factor efficacy. Indeed, addition of anabolic growth factors, such as Transforming growth factor-β (TGF-β) (Morales, et al. 1988, 1991a,b, Luyten, et al. 1988, van Osch, et al. 1998, Diao, H. et al. 2009,), bone morphogenetic protein-2 (BMP-2) (van Beuningen, et al. 1998, Blaney Davidson, et al. 2007), bone morphogenetic protein-7 (BMP-7) (Chubinskaya, et al. 2007, Hayashi , et al. 2008), insulin growth factors-1 (IGF-1) (Luyten, et al. 1988, van Osch, et al. 1998, Fortier, et al. 2002, Goodrich, et al. 2007) and fibroblastic growth factor (FGF) (Martin, et al. 1999a, Maehara, et al. 2010), to constructs was found to have little effect on the concentration of accumulated GAG although it increased construct size. In addition, GAG production rates appear to fall with time in culture in many different systems also limiting GAG accumulation (Mercier et al., 2004). Increasing cell density potentially should increase GAG deposition, but leads to a lower activity per cell, and also, in general, has not been found to increase GAG deposition rates (Panossian et al., 2001). It should also be noted that tissue in vivo cannot support too high a cell density, so in vitro culture of constructs at high cell density could lead to cell death after implantation.

Culture conditions such as stirring or perfusion (Freyria et al., 2000, Seidel et al., 2004) appear able to overcome diffusive transport initially, but as GAG concentrations rise and the hydraulic permeability of the construct falls, convective transport also is reduced and rates of GAG deposition slow. GAG concentrations were reported to reach 5% by wet weight within 2 months but took a further 5 months to increase to 7% GAG. In view of the long culture times which appear necessary to achieve the required GAG composition in vitro, achievement of in vivo concentration before implantation of a construct may be an unrealistic and possibly unnecessary goal for tissue engineered disc. It has been suggested that about 2-3 fold amount of GAG can be produced using anabolic growth factors. Even if such growth factors are used, more than 100 days of culture is thought to be necessary. Furthermore, it has been reported that turnover of GAG in the cartilage tissue takes about 2 - 3 years (Maroudas, 1975). So, GAG is slowly synthesized in the biological condition, and a long time is necessary to construct articular cartilage with adequate mechanical strength even if cells are maintained in active status by three-dimensional culture. Anabolic growth factors were obvious tools to enhance cartilage repair. Recently, Carragee et al (2011). reported that the bone morphogenic protein-2 (BMP-2) has cancer risk associated with the use of BMP-2 in spinal fusion surgery. In the future, orthopaedic surgeons must exercise caution to use the anabolic growth factors in clinics.

At present, the target diseases of treatment utilizing bioengineering (tissue engineering) such as chondrocytes implantation are local lesions such as traumatic cartilage defect and osteo‐ chondritis dissecans (Bittberg, 1999, Minas, 2001, Ochi, et al. 2002, Robert, et al. 2003) (Fig. 16A). For regeneration of extensive degenerated cartilage in OA, it is necessary to secure a large amount of chondrocytes for implantation. If a large graft is implanted, a nutritional problem may occur as explained above. In addition, the subchondral bone needs to be healthy to obtain the normal function of articular cartilage such as dispersion of load. But, lesion of OA is not localized in the cartilage layer but involves the subchondral bone (e.g., osteosclerosis) (Fig.16B). The osteosclerosis which is the subchondral bone, covers the bone-cartilage junction with age and looks to close the nutritional route through these vascular system (Trueta, 1963, Havelka, et al. 1984). Therefore, even if regeneration of cartilage is achieved by means of hyaline cartilage, the regenerated cartilage may sustain overload and may be degenerated

**Figure 16.** Target diseases of treatment utilizing bioengineering or tissue engineering technique.(A) Chondrocytes im‐ plantation is local lesions such as traumatic cartilage defect and osteochondritis dissecans at present. However, GAG is slowly synthesized in the biological condition, and a long time is necessary to construct articular cartilage with ade‐ quate mechanical strength even if cells are maintained in active status by culture.(B) Biological repair depends on the cartilaginous tissue maintaining a population of viable and active cells. If, in some degenerate cartilaginous tissues, nutrient supply is impeded (as it seems to be) hence resident cells are inactive or die. Therefore, we have to consider chemical and ionic environments in the degenerative cartilage before performing the tissue engineering of cartilagi‐ nous tissue.

again unless the mechanical environment is modified together. Cellular repair using autolo‐ gous chondrocyte transplantation appears successful even though chondrocytes are implanted with no matrix at all. Under these conditions, remodeling in vivo appears to produce a cartilage-type matrix under some conditions. Tissue engineered composites implanted with low GAG appeared to accumulate GAG in vivo, withstand physiological loading, and remodel towards a hyaline-type matrix. Perhaps optimization of such processes is a more useful goal.
