Preventing hypoxia-induced cell death

Preventing hypoxia-induced cell death in beta

cells and islets via hydrolytically activated,

oxygen-generating biomaterials

Eileen Pedraza

a,b, Maria M. Coronela,b, Christopher A. Frakera,b, Camillo Ricordia,b,c,d, and Cherie L. Stablera,b,c,1

a

and

Edited by Robert Langer, Massachusetts Institute of Technology, Cambridge, MA, and approved February 1, 2012 (received for review August 17, 2011)

Diabetes Research Institute and Departments of cSurgery and dMedicine, Leonard M. Miller School of Medicine, University of Miami, Miami, FL 33136;bDepartment of Biomedical Engineering, College of Engineering, University of Miami, Coral Gables, FL 33134

A major hindrance in engineering tissues containing highly

metabolically active cells is the insuf

implants, which results in dying or dysfunctional cells in portions

of the graft. The development of methods to increase oxygen

availability within tissue-engineered implants, particularly during

the early engraftment period, would serve to allay hypoxiainduced

cell death. Herein, we designed and developed a hydrolytically

activated oxygen-generating biomaterial in the form of

polydimethylsiloxane (PDMS)-encapsulated solid calcium peroxide,

PDMS-CaO

PDMS resulted in sustained oxygen generation, whereby

a single disk generated oxygen for more than 6 wk at an average

rate of 0.026 mM per day. The ability of this oxygen-generating

material to support cell survival was evaluated using a

and pancreatic rat islets. The presence of a single PDMS-CaO

eliminated hypoxia-induced cell dysfunction and death for both

cell types, resulting in metabolic function and glucose-dependent

insulin secretion comparable to that in normoxic controls. A single

PDMS-CaO

more than 3 wk under hypoxic culture conditions. Incorporation

of these materials within 3D constructs illustrated the bene

these materials to prevent the development of detrimental oxygen

gradients within large implants. Mathematical simulations

permitted accurate prediction of oxygen gradients within 3D constructs

and highlighted conditions under which supplementation

of oxygen tension would serve to bene

the generality of this platform, the translation of these materials

to other cell-based implants, as well as ischemic tissues in general,

is envisioned.

ficient oxygenation of these2. Encapsulation of solid peroxide within hydrophobicβ cell line2 disk2 disk also sustained enhanced β cell proliferation forfits offit cellular viability. Given

tissue engineering

| encapsulation | bioartificial pancreas | diabetes

T

by inadequate oxygen delivery, typically due to the

inevitable delay in angiogenesis after implantation (1

oxygen levels result in cellular necrosis and apoptosis,

as well as a shift to anaerobic metabolism and energy conservation

(3, 5

containing highly metabolically active cells, such as those comprising

pancreatic islets, is particularly challenging (8, 9). Not

only is the oxygen consumption rate for pancreatic islets elevated

compared with many other cell types, they are also susceptible to

functional impairment even at moderate oxygen tensions (9

In the initial days after device implantation, in the absence of

vascular in

the outer boundaries of the implant and progressively decreases

in square proportion to the radial distance inward. High cell

loading densities further exacerbate hypoxia, by increasing the

overall metabolic load. Thus, designing the appropriate device

dimensions and cell loadings to limit hypoxia within a transplant

is an arduous task, commonly resulting in device scales that are

impractical for clinical implantation (2, 4). Limiting or eliminating

the hypoxic period between implantation and the development

of a fully functional, intradevice, vascular network would –19). In situ oxygen–20). Although allfically calcium peroxide, which is hydrolytically1.

2CaO

2 þ 4H2O → 2CaðOHÞ

2

þ

2H2O2

→

2CaðOHÞ

2

þ

2H2O þ O2 [1]

Oxygen generation via hydration of solid peroxide, however,

is too rapid, leading to hyperoxide conditions that result in the

buildup of the reaction intermediate hydrogen peroxide and increased

susceptibility of side reactions such as hydroxyl radical

(OH

solid peroxide within a polymer could serve to temper its hydrolytic

reactivity. A solid peroxide platform using

(lactic-

oxygen-generating potential of these materials was shown

through enhanced

in a skin

however, was short lived, and the reaction kinetics were less

controlled owing to the hydrolytic degradation of the PLGA.

Furthermore, catalase supplementation was required to mitigate

cytotoxicity. Thus, coculture of these materials with cells less resilient

than

stress, such as

We sought to fabricate an oxygen-generating biomaterial

through the encapsulation of solid peroxides within a highly−) generation (details in SI Text) (21). Encapsulation offilms of polyco-glycolic acid) (PLGA) has been published, whereby thefibroblast proliferation and reduced necrosisflap mouse model (16, 17). The benefit of the films,fibroblasts, particularly those sensitive to free radicalβ cells, is highly undesirable (22).

hydrophobic, biostable material, speci

(PDMS) (Fig. 1

migration through the biomaterial would result in greater temporal

modulation of oxygen generation. Furthermore, by selecting

a material with a high oxygen permeability, the diffusion of

oxygen out of the material should be ef

of controlled reactivity and clearance of end products has the

added bene

the presence of the hydrogen peroxide intermediate and the

susceptibility to generate side reactions such as hydroxyl radicals.

Herein we report the design and fabrication of PDMS-encapsulated

calcium peroxide (PDMS-CaO

we sought to evaluate (

peroxide within PDMS on long-term oxygen generation; (

ability of PDMS-CaO

death in culture; and (

to enhance cell viability within 3D tissue engineered constructs.

We also explored the correlation between oxygen availability and

fically polydimethylsiloxaneA). We hypothesized that the restriction of waterficient. The combinationfit of driving the reaction forward, thereby reducing2) disks. In this studyi) the effect of encapsulation of calciumii) the2 materials to mitigate hypoxia-induced celliii) the potential of PDMS-CaO2 materials

β

materials.

cell viability, to permit for the appropriate design of oxygengenerating

Results

Long-Term Oxygen Generation by PDMS-CaO

the effect of encapsulation of solid peroxide within a hydrophobic

material on oxygen generation, solid peroxide was encapsulated

within PDMS at 25% wt/wt (Fig. 1

potential of the system was noninvasively monitored in an incubator

using oxygen sensors. With the target oxygen tension of

the incubator set at 0.05 mM oxygen, the system represents an

open system, in which gas exchange between the solution and air

daily oxygen concentration within solutions containing

PDMS-CaO

control PDMS-only disk, the oxygen remained steady at 0.05

2 disks, as well as for control PDMS-only disk. For the±

0.0001 mM, con

of oxygen production from the PDMS-only controls. The addition

of a single PDMS-CaO

oxygen in the solution. For the

firming the stability of the incubator and the lack2 disk resulted in a significant increase infirst week, oxygen concentration

fl

oxygen levels averaged 0.12

4 wk (17

Thus, during the

that results in the shift from a hypoxic environment to that

close to optimal cell culture conditions.

uctuated, averaging 0.16 ± 0.017 mM. During the second week,± 0.003 mM, whereas for the subsequent–45 d) oxygen levels averaged 0.073 ± 0.007 mM.first 2 wk, these disks generate oxygen to a degree

Mitigation of Hypoxia-Induced

evaluate the ability of a PDMS-CaO

oxygen to minimize hypoxia-induced cell death, MIN6 cells, a

β Cell Death via PDMS-CaO2 Disk. To2 disk to generate sufficientβ

cell line, were cultured overnight under hypoxic conditions (0.01

mM) with or without a PDMS-CaO

assessed and compared with cells cultured at standard oxygen

tension (0.20 mM). As shown in Fig. 2, coculture of hypoxic

MIN6 cells with PDMS-CaO

by increased methylthiazolylphenyl-tetrazolium bromide (MTT)

levels and total protein, with concomitant decreases in lactate

dehydrogenase (LDH) release and caspase activity. Relative to

normoxic (0.20 mM) controls, MIN6 cell survival under hypoxic

conditions (0.01 mM) with the PDMS-CaO

or even enhanced, with statistically identical levels of MTT

and total protein, as well as signi

and caspase activity (

2 disk. Cell survival was2 prevented cell loss, evidenced2 disk was comparableficant decreases in LDH releaseP < 0.05). The addition of a PDMS-CaO2

disk under normoxic culture conditions resulted in a moderate

but statistically signi

in LDH release and caspase activity. These latter results indicate

a bene

under normoxic conditions.

Elevated caspase activity under hypoxia is modest compared with

LDH increases, indicating dominance of necrosis vs. apoptosis at

the time points tested. As expected, hypoxia resulted in dysfunction

in glucose-stimulated insulin secretion (Fig. 3

a drop in the stimulation index (high/low glucose) from strong

(15.30) to virtually unresponsive (1.2). Hypoxic conditions

quickly resulted in islet spheroid fragmentation and increased

cell membrane permeability, as evidenced by live/dead images

(Fig. 3

was observed with the addition of a single PDMS-CaO

with a signi

complementary decreases in LDH release and caspase activity.

Glucose responsiveness of the islets was maintained, with stimulated

insulin secretion comparable to that in controls, but at

a reduced overall stimulation index (6.94) due to increased basal

insulin secretion. This was further supported by live/dead imaging,

which exhibited intact, highly viable islet spheroids.

B), withE). Alternatively, substantial enhancement in islet survival2 disk,ficant increase in overall cell metabolic activity and

Enhanced Long-Term Proliferation of Cells Cocultured with PDMSCaO

2

Disks.

proliferation and viability of cells under low (0.05 mM) oxygen

tensions was investigated by incubating MIN6 cells (3

per well) with a single PDMS-CaO

cell loading and oxygen tension, our multiphysic models predicted

maintenance of overall cell number for controls, whereas

cells supplemented with a PDMS-CaO

owing to the increased oxygen availability. As expected, overall

cell viability for controls remained steady during the course of

the experiment, with no change in MTT metabolic activity (

0.101). Conversely, MIN6 cells cultured under low oxygen with

a PDMS-CaO

the

twofold higher than controls (average 2.32

23 d,

between days 7 and 23 for this group (

P =2 disk quickly increased in metabolic activity withinfirst 3 d and plateaued at a level equivalent to more than± 0.31-fold from 3 toP < 0.001). No significant change in viability was observedP = 0.141). If the PDMSCaO2

disk was removed, MTT measurements made 3 d later

were statistically equivalent to those in low oxygen controls (Fig.

4

days 3, 7, and 17). Overall,

with the PDMS-CaO

cells cultured at standard oxygen tension (0.20 mM) (

A; dashed lines track readings before and after disk removal onβ cells cultured at 0.05 mM oxygen2 disk were statistically equivalent to controlFig. S1;

P

0.2 mM oxygen level, however, did not substantially affect overall

cellular metabolic activity per MTT absorbance (

lack of bene

limiting factor for MIN6 proliferation; rather other factors, such

as glucose availability or contact inhibition, are preventing further

increases in overall cell proliferation. Controls of PDMSonly

disk (no CaO

controls (

PDMS has no effect on cell survival.

DNA content of the cells within these experimental groups

followed similar trends as the MTT results: the total DNA for

wells cultured with the PDMS-CaO

higher than in controls under identical low oxygen conditions

from day 7 to 24 (Fig. 4

sustained increase in overall cell number during the course of the

experiment. Live/dead time course confocal images of MIN6

cells cultured under low oxygen tension (0.05 mM) revealed

enhanced viability and proliferation of cells when a PDMS-CaO

disk was present. In addition, cells incubated with PDMS-CaO

2

disks tended to form larger cellular spheroids, because of either

growth of the cells into clusters or aggregation.

Enhanced

β Cell Viability for 3D Agarose Constructs Containing PDMSCaO2

Disks.

materials within a 3D tissue-engineered implant, MIN6 cells

(25

3 d under hypoxic (0.05 mM) or normoxic (0.20 mM) conditions

with a centrally placed PDMS-only (control) or PDMS-CaO

To evaluate the potential of our oxygen-generating× 105 total cells) were cultured within agarose constructs for2

disk (Fig. 5

metabolic activity of each group, compared with day 0 controls.

At 0.20 mM oxygen, no enhancement in cell viability was observed

for either group after the 3 d culture. For control constructs

cultured at 0.05 mM oxygen, however, the viability of the

cells was substantially reduced (53%). The viability of the cells at

0.05 mM oxygen was signi

A). Fig. 5B summarizes the fold change in MTTficantly improved when a PDMS-CaO2

disk was present (93%), proving to be statistically identical to the

0.20 mM oxygen group (

sections of the constructs exhibited varying cell viability

patterns after 3 d in culture (Fig. 5

a rim of live cells can be discerned on the outer edge of the

construct, surrounding a central core of dead cells. As expected,

the thickness of the viable core is slightly larger for the higher

oxygen culture condition. This pattern is typical for implants,

whereby the highest oxygen tension is at the periphery of the

implant. On the contrary, this viable demarcation is absent when

the PDMS-CaO

number of viable cells may be observed at normoxic conditions,

live and dead cells are intermingled within the slice, and no

spatial preference is visible. These results indicate that the oxygen

gradient that typically leads to inner pockets of dying or

dysfunctional cells is absent when a PDMS-CaO

placed within the construct.

P = 0.52). Live/dead images of transversalC). For control constructs,2 disk is present. Although a slightly higher2 disk is centrally

Effect of Oxygen Availability on

the relationship between

the oxygen availability within agarose constructs was varied via

(

presence or absence of the oxygen-generating material. Simulation

modeling via COMSOL multiphysics modeling was used to

assist in predicting the oxygen gradients within 3D agarose

constructs. Modeling results not only provide a spatial distribution

of oxygen throughout the construct but can also be used to

quantitatively predict the percentage volume of the construct

exposed to severe hypoxia, which may then be used to predict

and/or characterize experimental results.

COMSOL models of constructs were run with a variation in

MIN6 cell loading density from 5 to 25

an external oxygen tension of low (0.05 mM) or standard (0.20

mM) oxygen levels, and with a PDMS-only or a 25% wt/wt

PDMS-CaO

expected, oxygen gradients become more pronounced as the

external oxygen tension is decreased and as the cell loading is

increased, thereby reducing overall oxygen availability. For

control constructs, because oxygen migration is dependent upon

diffusion and consumption by the cells, a precipitous decrease in

oxygen tension within this construct is predicted. This decrease

in oxygen toward the center of the construct is mitigated, if not

eliminated, with the introduction of the PDMS-CaO

these gradients shift to high oxygen tensions both externally

and centrally.

In examining the individual cell loading models, the effects of

these oxygen gradients on cell viability may be predicted. For

25

(61% vol at 0 mM oxygen), with minimum effects under standard

oxygen (5% vol at 0 mM oxygen). The models predict that the

introduction of an oxygen-generating disk signi

areas of hypoxia under low oxygen tension (11% vol at 0 mM

oxygen) and eliminates hypoxia for constructs under standard

oxygen tension. As the cell loading is decreased to 15

oxygen gradients within the constructs become less pronounced.

Under a low external oxygen condition (0.05 mM), a smaller

portion of the construct is severely hypoxic (12% vol at 0 mM

oxygen), indicating that the lower cell loading serves to increase

overall oxygen availability. Simulations predict that the addition

of the PDMS-CaO

0.2 mM external oxygen, severely hypoxic regions are minimal

for this cell loading (

PDMS-CaO

and the oxygen tension within the control constructs is

maintained well above zero oxygen at both low and standard

surrounding oxygen tensions. The addition of the PDMS-CaO

β Cell Viability. To fully characterizeβ cell viability and oxygen availability,i) total cell loading, (ii) external oxygen tension, and (iii) the× 105 cells per construct,2 disk at the center (Fig. 6A and Fig. S3). As2 disk, where× 105 cells, a substantial hypoxic core is formed at low oxygenficantly reduces× 105 cells,2 disk eliminates these hypoxic regions. At<0.5%), thereby limiting the benefit of the2 disk. For 5 × 105 cells, oxygen gradients are minimal,2

disk only further increases oxygen levels. Overall, these models

predict a moderate to substantial enhancement in cell viability

for cell loadings from 15

a low oxygen environment (0.05 mM). Of note, simulation

models do not account for cellular proliferation and remodeling,

thereby restricting their accuracy to initial stage cultures.

To correlate model predictions of oxygen availability to experimental

results, MIN6 cells were loaded within agarose constructs

and incubated at both low (0.05 mM) and standard (0.20

mM) oxygen conditions at the three cell densities tested: 5

15

Results are expressed as the fold increase of each PDMS-CaO

× 105 to 25 × 105 cells, when cultured in× 105,× 105, and 25 × 105 MIN6 cells per construct (Fig. 6B).2

group over its corresponding PDMS-only control, whereby fold

values greater than 1.0 indicate that the PDMS-CaO

a positive effect on cell viability, and values less than 1.0 indicate2

groups over controls is higher for the lower oxygen culture

condition. For relatively low cell loading densities (5

the treated groups have lower viability compared with controls

(

a decrease in cell viability when the PDMS-CaO

present. For 15

at 0.20 mM oxygen but has a neutral effect at 0.05 mM oxygen.

At the highest cell loading tested, 25

treated groups is greater than that of controls, at both oxygen

tensions. Overall, a general trend of enhanced bene

of PDMS-CaO

both low and standard oxygen culture conditions. These results

correlate strongly with multiphysics modeling predictions.

× 105 cells),<1.0) at both low and standard oxygen tensions, exhibiting2 disks are× 105 cells, the PDMS-CaO2 disk is detrimental× 105 cells, the viability offit of the addition2 disk as cell loading increases is observed at

Discussion

In this study, we sought to encapsulate calcium peroxide within

PDMS disks for use as an oxygen-generating biomaterial suitable

for sustaining cell viability and function within hypoxic conditions.

Without encapsulation, peroxide-based compounds quickly

release oxygen upon contact with water, resulting in bursts of oxygen

that are too severe and transient for use (21). By encapsulating

solid peroxide within a highly hydrophobic biomaterial, we

have introduced a diffusional barrier that substantially reduces its

reactivity, consequently modulating the release of oxygen for more

than 40 d. The reactivity of the encapsulated peroxide is heavily

dictated by the rate of water diffusion into the PDMS, because

water permeability in PDMS is very low, although it is likely increased

by the incorporation of solid peroxide particulates (23).

Furthermore, oxygen diffusion and permeability in PDMS is particularly

high (24), thereby ensuring the ef

out of the material and into the surrounding milieu. With this

degree of control, the geometry and dimensions of the disk, as well

as the calcium peroxide loading, can be manipulated to achieve the

desired release kinetics of oxygen. Given that the reactivity of the

peroxide is reversible, in situ oxygen generation is also dependent

upon the surrounding oxygen conditions, thereby permitting further

modulation of the kinetics. Future studies are focused on fully

characterizing the effects of each of these parameters on oxygen

generation within this material platform.

We were particularly interested in evaluating the capacity of

this material to preserve cell viability and function under hypoxic

conditions for

with MIN6

PDMS-CaO

death and dysfunction, limiting activation of cell stress pathways

and the shift to anaerobic metabolism. Furthermore, this effect

can be observed long term, as evidenced by the sustained enhanced

survival of cells over 3 wk in culture. Removal of PDMSCaO

ficient diffusion of oxygenβ cells. Coculture of oxygen-generating materialsβ cells and pancreatic rat islets demonstrated that2 disks dramatically mitigated hypoxia-induced cells2

disks from treated groups reinforces the theory that the

observed enhancements are due solely to oxygen release and not

due to an artifact or initial shielding from hypoxia. It should also

be noted that no free radical or peroxide scavenging materials

were used in this study, in contrast to other published reports

using solid peroxides in which catalase was used for all experimental

studies (17).

Evaluation of PDMS-CaO

also illustrated the bene

materials when cells are exposed to hypoxic conditions. Placement

of the oxygen-generating disks at the center of the construct

resulted in the greatest bene

gradients. This is veri

correlated strongly with experimental results whereby cellular

metabolic activity was comparable to constructs cultured at oxygen

pressures fourfold higher. Imaging of viable

within agarose constructs permitted evaluation of cellular proliferation

and remodeling in response to oxygen availability.

Control constructs exhibited viability patterns typical of limited

oxygen availability, with live cells relegated to a thin rim on the

outer edge of the construct and a core of predominantly dead cells

(25). Constructs containing PDMS-CaO

a uniform distribution of predominantly live cells, indicating

a divergent pattern in cellular remodeling. This further

establishes that the PDMS-CaO

providing an additional source of oxygen originating in the center

of the construct, where it is most needed.

Studies seeking to evaluate the correlation between oxygen

availability and

oxygen-generating materials within 3D constructs when oxygen is

limited. Decreases in cell loading decreased the positive effect of

PDMS-CaO

oxygen consumption, which, when coupled with oxygen release

from the PDMS-CaO

Hyperoxia conditions can be as detrimental to cells as hypoxia,

particularly for cells that are more sensitive to free radical stress,

such as

con

detrimental oxygen gradients are likely to develop. As illustrated

in this study, mathematical modeling can be used to predict what

cell loadings and implant geometries will lead to oxygen limitations.

Hence, the dosing of the oxygen-generating material, per

peroxide loading, geometry, and surface/volume ratio, can be

subsequently tailored.

Because the PDMS-CaO

loading densities, they represent an ideal tool for enhancing

2 disks embedded within 3D hydrogelsfits of these oxygen-generatingfit by decreasing internal oxygenfied via predictive modeling results, whichβ cell distribution2 disks, however, illustrated2 disk alleviates central hypoxia byβ cell viability illustrate the benefits of these2 materials, likely owing to the lower total rate of2 disks, results in a local surplus of oxygen.β cells (22). Therefore, use of these materials withinfined constructs has to be relegated to scenarios whereby2 disks are most beneficial at high cell

Fig. 6.

simulation models of oxygen gradients within 3D agarose constructs containing

either PDMS-only or PDMS-CaO

outlined in Fig. 5

(0.2 mM in

range (0.05 to 0 mM) is shown on far right. Black, 0 mM oxygen tension. (

Experimental outcomes of oxygen availability on MTT metabolic activity.

Oxygen availability varied via total MIN6

cells), external oxygen tension (0.2 mM or 0.05 mM), and the presence of

a PDMS-CaO

expressed as fold increase over control at identical cell loading density and

oxygen concentration (fold = treated group with PDMS-CaO

group without the PDMS-CaO

indicate a positive effect of the PDMS-CaO

values

Effect of oxygen availability on MIN6 β cell viability. (A) Multiphysics2 disk. (Dimensions identical to thoseA). Models run with 0.05 mM external oxygen are shownFig. S3) at the total MIN6 cell loadings indicated Predicted oxygenB)β cell loading (5–25 × 105 total2 disk. Measurements were collected after 3 d culture and2 disk/control2 disk). Fold increase values greater than 1.02 disk on cell viability, whereas<1 indicate a detrimental effect. Error = SD; n = 3.

Pedraza et al. PNAS

| March 13, 2012 | vol. 109 | no. 11 | 4249

ENGINEERING MEDICAL SCIENCES

oxygenation of cell-based engineered tissues, particularly during

the initial engraftment period. By supplementing oxygen during

the precarious vascularization period, hypoxia-induced cell loss

could be reduced. It is envisioned that PDMS-based implants

would be highly desirable to transplant applications, given its

proven clinical safety pro

of agents (26). With low adhesion properties, the PDMS insert can

be easily removed once the embedded solid peroxide is expired or

the local supplementation of oxygen is no longer desired. Although

one potential drawback of this approach is that hypoxia

inducible factor pathways that stimulate release of angiogenic

growth factors would be suppressed and potentially delay implant

vascularization, we anticipate that this effect could be mitigated

through the codelivery of proangiogenic factors (13, 14). Overall,

we have demonstrated the usefulness of our materials to mitigate

hypoxia-induced cell death, which would be highly desirable for

preserving the viability of the transplanted cells after transplant,

particularly for implants for which cell dosage is critical for ef

such as in the transplantation of file and prolific use in the local deliveryficacy,β cells for type 1 diabetes.

a detrimental effect. One trend that can be readily observed is

that for all cell loading densities, the fold increase of PDMSCaO

2 disk has= 0.65). Introduction of the PDMS-CaO2 disk to cells at theFig. S1). Thisfit at normoxia indicates that oxygen is no longer the2) were statistically identical to material-freeFigs. S1 and S2), verifying that the presence of the2 disks was 2.51 ± 0.36-foldB; P < 0.001), indicating a substantial and2The long-term effects of PDMS-CaO2 material on the× 105 cells2 disk for 3 wk (Fig. 4). At this2 disk would proliferateficant increase in total protein and decreasefit of the presence of the oxygen-generating material even2 Materials. To evaluateB), and the oxygen-generating

dramatically reduce hypoxia-induced cell death and permit for

more clinically translatable devices. Such methods include prevascularization

of the transplant site (12), hastening vascularization

through the delivery of growth factors (13, 14), incorporation

of oxygen carriers within biomaterials (15), or the in

situ generation of supplemental oxygen (16

generation is a highly desirable approach, in that it does not

require multiple surgeries and provides supplemental oxygen

immediately upon implantation. Various agents have been investigated

for their oxygen-generating potential, from the electrochemical

decomposition of water to the photosynthesis of

microalgae to the use of oxygen chambers (18

these agents highlight the potential of in situ oxygen generation

to enhance cell viability, most of these complicate the transplant

site by increasing the size of the implantable device and/or possibly

introducing toxic by-products.

One potent source of oxygen generation is the decomposition

of solid peroxides, speci

activated to generate oxygen via the reaction shown

in Eq.

he implantation of cellular-based devices is commonly hampered–4). Insufficient–7). Providing sufficient oxygen delivery for implants–11).filtration, the maximal oxygen concentration occurs at