Stem Cell Bio Engineering

8/2/2019 Stem Cell Bio Engineering

1/33

Annu. Rev. Biomed. Eng. 2001. 3:275305Copyright c 2001 by Annual Reviews. All rights reserved

STEM CELL BIOENGINEERING

Peter W. Zandstra1,2 and Andras Nagy31Institute of Biomaterials and Biomedical Engineering and2Department of Chemical

Engineering and Applied Chemistry, University of Toronto, and 3Samuel Lunenfeld

Research Institute, Mount Sinai Hospital, Toronto, Ontario, Canada M5S 3G9;

e-mail: [email protected], [email protected]

Key Words cellular engineering, tissue engineering, self-renewal anddifferentiation, bioprocess design, applied developmental biology,microenvironment, embryonic and tissue-specific stem cells

s Abstract Tissue engineering and cellular therapies, either on their own or incombination with therapeutic gene delivery, have the potential to significantly impactmedicine. Implementation of technologies based on these approaches requires a readilyavailable source of cells for the generation of cells and tissues outside a living body.Because of their unique capacity to regenerate functional tissue for the lifetime ofan organism, stem cells are an attractive raw material for multiple biotechnological

applications. By definition they are self-renewing because on cell division they cangenerate daughter stem cells. They are also multipotent because they can differentiateinto numerous specialized, functional cells. Recent findings have shown that stem cellsexist in most, if not all, tissues, and that stem cell tissue specificity may be more flexiblethan originally thought. Although the potential for producing novel cell-based prod-ucts from stem cells is large, currently there are no effective technologically relevantmethodologies for culturing stem cells outside the body, or for reproducibly stimu-lating them to differentiate into functional cells. A mechanistic understanding of theparameters important in the control of stem cell self-renewal and lineage commitmentis thus necessary to guide the development of bioprocesses for the ex vivo culture ofstem cells and their derivates.

CONTENTS

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276

STEM CELL POPULATIONS IN EMBRYOS AND ADULTS . . . . . . . . . . . . . . . . . 277

Embryonic Stem Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277

Tissue-Specific Stem Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279

STEM CELL PLASTICITY OR EMBRYONIC LATENCY . . . . . . . . . . . . . . . . . . . . 283STEM CELL SELF-RENEWAL AND LINEAGE COMMITMENT . . . . . . . . . . . . . 284

Control of Stem Cell Self-Renewal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284

Control of Lineage Commitment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288

STEM CELL CULTURE SYSTEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290

15239829/01/0825-0275$14.00 275

StemCells2008.1:275-305.Downloadedfromwww.annualreviews.or

g

byNationalCentreforBiologicalScienc

eson03/26/12.Forpersonaluseonly.


2/33

276 ZANDSTRA NAGY

Cytokines and Growth Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292

Physicochemical Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293

CONCLUSIONS AND FUTURE DIRECTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295

INTRODUCTION

Stem cells are cells that (a) on dividing are capable of recreating themselves in

at least one daughter cell and (b) have the capacity to differentiate into several

linages. Stem cells play a critical and essential role in the human body not only by

providing the starting material for organs and tissues but also for their continual

maintenance, growth, and renewal throughout ontogeny [e.g. hematopoietic stem

cells (HSCs) erythrocytes, neural stem cells (NSCs) neurons, etc]. As theembryo and the fetus develop, stem cells are seeded into the various tissues and

organs where they remain throughout life (Figure 1). Stem cells can also exist

temporarily during embryonic development and can be artificially established ex

vivo from transient stages of differentiation. Embryonic stem (ES) cells of mice,

for example, are isolated by culturing transiently existing inner cell mass cells of

early embryos. ES cells are cells that retain the properties of primitive ectoderm

cells, i.e. they are undifferentiated but have the capacity to differentiate into all

adult tissues.

Because of their potential to provide essential materials fornext-generation ther-apeutics, including gene, cellular, and tissue regeneration therapies, both naturally

occurring and artificially established stem cells are the subjects of intense investi-

gation. The clinical utility of stem cells is limited, however, because no effective

technological methodologies exist to cultivate stem cells in vitro, or to stimulate

them down particular differentiation pathways. The development of bioprocesses

for the generation or ex vivo maintenance of stem cells and their derivatives is

complicated by the biological properties of most stem cell populations. In general

stem cells are rare, quiescent, or slowly cycling cells with complex microenvi-

ronmental requirements. Culture optimization protocols must thus be designed to

specifically target stem cell populations while incorporating feedback from a po-

tentially dynamic mature and maturing cell population. Alternatively, selection or

enrichment processes can be used to enhance the frequency of stem cells in hopes

of being able to more efficiently target their growth. Approaches being used to

enrich stem cell populations are reviewed elsewhere (1). Optimization of culture

conditions for stem cell growth and differentiation has been further hindered by

the need for retrospective assays of cell function to quantify stem cell responses

to candidate culture parameters. Although these assays have been, and continue to

be, exceedingly useful for describing properties of stem cells, their often lengthy

and complicated nature represents a significant bottleneck in stem cell bioprocess

development. It is additionally important to realize that the application and inter-

pretation of many of these functional assays depends on established principles that

are thought to govern the dynamics of all stem cell populations. These principles


g




3/33

STEM CELL BIOENGINEERING 277

include (a) the irreversibility of stem cell differentiation, (b) the progressive com-

mitment to a particular lineage, and (c) a hierarchical framework with increasing

numbers of cells with a progressively decreasing capacity for self-renewal, pro-

liferation, and lineage potential. Although, in general, these principles have beenshown to be true, recent studies that challenge some of these concepts (e.g. 26)

stress the need for the continual reevaluation of established dogma (7).

The purpose of this review is to outline the salient biological properties of stem

cell populations, highlight the main bioengineering challenges in the development

of stem cellbased technologies, and focus on how recent information about pa-

rameters that may influence stem cell self-renewal and differentiation impact the

design of these bioprocesses. In addressing these questions, we draw mainly on

results from embryogenesis and hematopoiesis, although examples from neuroge-

nesis and other stem cell systems are used as appropriate to argue the generalityof our observations. Given the breadth of the information available, an exhaus-

tive analysis of all potential cell systems and culture parameters is not feasible

(or desirable). Our conceptual focus thus represents our perspective on future di-

rections that we believe may be fruitful in addressing the critical bioengineering

challenges that must be overcome to develop clinically relevant stem cellbased

technologies.

STEM CELL POPULATIONS IN EMBRYOS AND ADULTS

Embryonic Stem Cells

ES cells (8, 9) can be cultured for extended periods and genetically manipulated

without loss of their pluripotential capacity (as assessed by their ability to partici-

pate in all aspects of the development of the embryo proper when reintroduced in

vivo) (10). In spite of this remarkable differentiation potential, ES cells cannot dif-

ferentiate into some extraembryonic membranes, such as parietal endoderm, yolk

sac visceral endoderm, and trophoblast of the placenta (1113). ES cells can be

induced to differentiate in vitro into two- or three-dimensional (embryoid bodies)

cultures containing cells from all three embryonic lineages (ectoderm, mesoderm,

and endoderm), including cells of the hematopoietic (14), endothelial (15), cardiac

(16, 17), and neuronal (18, 19) tissues. The sequence of lineage commitment events

that lead to the formation of these cell types in vitro closely mimic that observed

in vivo, suggesting that the in vitro cell culture model follows developmentally

established rules (20).

ES cells were first derived from mice and are now available from a vari-

ety of mammalian systems, including human. Because ES cells can be main-

tained and expanded in an undifferentiated state under defined culture conditions

(21, 22), and while retaining their capacity to differentiate into functional cell

types in vitro, these cells hold promise as an untapped cell source for multiple

clinical and biotechnological applications. The evaluation of the entire range of


g




4/33

278 ZANDSTRA NAGY

developmental potential of human ES cells will be much more tedious and slower

than that of mouse ES cells, where it was possible to perform in vivo experiments

to test their contribution to chimeras (10) or establish completely ES cellderived

animals (12). The potential of mouse ES cells, however, sets our expectations aboutthe human-derived cells, where such ES cells are capable of differentiating into

all the cell types of adult mice and are completely functional.

In vitro assays for detecting undifferentiated ES cells have been developed

based on the cell surface expression of particular antigens (i.e. stage-specific em-

bryonic antigen) (23), in vitro developmental capacity (24), or expression of genes

associated with the properties of undifferentiated ES cells (e.g. Oct4) (25, 26). The

extent to which these assays predict ES cell developmental potential needs to be

further defined. Oct4, perhaps the most established marker of undifferentiated ES

cells, is a mammalian transcription factor expressed exclusively in blastomeres,pluripotent early embryo cells, and germ-line lineage cells (26, 27). Expression of

Oct4 has been shown to be crucial to the establishment of pluripotential identity in

inner-cell mass cells and also for maintaining such identity in ES cells (28). Oct4

is downregulated on differentiation into both primitive ectodermal and trophec-

todermal cells and in vivo during gastrulation (29). A recent study has suggested

that the precise level of Oct4 expression can result in three separate ES cell fates.

A critical threshold of Oct4 is required to maintain ES cell self-renewal. A less

than twofold increase in Oct4 expression causes differentiation into embryonic

ectoderm and mesoderm, whereas repression of Oct4 expression induces loss ofpluripotency and dedifferentiation into trophectoderm (30).

Although relatively robust techniques have been developed to maintain the de-

velopmental potential of ES cells in vitro, ES cellbased technologies are severely

limited by our inability to control their differentiation (31, 32). Studies that have

investigated the effect of cytokines and growth factors on differentiating ES cells

have typically resulted in the altered proportions of specific differentiated cell

types. Likely because of spatial gradients in the differentiating cell aggregates, as

well as complex cell-cell interactions, until recently any success in directing ES cell

differentiation into specific cell types has not been possible. It is encouraging thatrecent studies examining the differentiation of ES cells to dopaminergic neurons,

either after extensive culture manipulations (33) or in response to labile soluble

factors secreted by stromal cells (34), show that investigations into identifying

extrinsic signals that control lineage specific differentiation of stem cells to allow

the generation of specific tissues in vitro is an attainable end goal. It is important

that ES cell aggregates often also contain undifferentiated ES cells, which when

transplanted in vivo may have the capacity to induce teratomas; thus, further study

is necessary to examine the tumorigenic potential of long-term transplants. The

use of tissue-specific promoters driving selectable markers is one approach thatmay overcome some of these limitations (35, 36). This approach has been used

both to select for undifferentiated ES cells in the absence of leukemia inhibitory

factor (LIF) and to isolate specific populations of ES cellderived cells (36, 37).

These techniques, in combination with improved culture systems for the controlled


g




5/33


differentiation of either aggregates or individual ES cells (38), are necessary to

move this technology forward.

Tissue-Specific Stem Cells

HEMATOPOIETIC STEM CELLS HSCs are defined functionally by their ability to

serially sustain multilineage hematopoiesis after transplantation into a hematolog-

ically compromised host. Experimental evidence supporting the existence of such

cells was first provided by Till & McCulloch (39), who injected lethally irradiated

mice with bone marrowderived cells from healthy syngeneic (genetically iden-

tical) donors and observed the formation in the spleen of multilineage colonies

that could be shown to be derived from single injected cells. Subsequent exper-

iments utilizing genetic marking have demonstrated that long-term engraftmentof both the lymphoid and myeloid lineages can be achieved by the progeny of a

single cell (40, 41). In addition, the ability to repopulate multiple secondary re-

cipients with the progeny of a single clone (42, 43), along with their extensive

amplification when limiting numbers are transplanted (44, 45), has demonstrated

that these very primitive cells are capable of extensive (>1000-fold) in vivo ampli-

fication. Although it has been definitively shown that HSCs grown in culture are

also able to repopulate multiple secondary recipients from the progeny of a single

clone, indicating that individual stem cells are able to undergo self-renewal in vitro

(44, 46), the identification of culture conditions that reproducibly support a signif-icant net in vitro expansion of HSC numbers have been elusive. A large number

of studies have shown that a modest expansion of HSC numbers can achieved in

both adherent and nonadherent culture systems using several different cytokine

combinations (47). It is important to note that in most of these studies, HSC ex-

pansion occurred in a relatively transient manner (typically a maximum of two- to

sixfold over 10 days). These and other observations have led to the hypothesis

that in vitro expansion systems are limited by at least two factors: the eventual

extinction of stem cells due to their differentiation, and an inability to get the stem

cell to home to an appropriate in vivo microenvironment (48). It is interestingthat the few studies that have shown sustained repopulating HSC expansions in

vitro are based on small-scale adherent cell culture systems that have been sub-

jected to the continual depletion of differentiated cells (based on their selective

adhesion to stroma) (4951). These results, along with the recent observations

that particular subpopulations of cells potentiate in vitro HSC expansion (52) and

engraftment (53), suggest that current HSC culture systems may be inhibited by

the types of cells and cellular products they generate. The documented differences

in the secretion of both positive (e.g. Flt-3) and negative regulators [transforming

growth factor (TGF)-] of hematopoiesis by subpopulations of progenitors andmature cells (for a review, see 54) suggests that soluble secreted factors may, at

least in part, play a role in these processes. Thus, in contrast to ES cells, where

our ability to control lineage-specific differentiation is still a major challenge, a

major limitation to the development of technologies using HSCs is the lack of


g




6/33

280 ZANDSTRA NAGY

TABLE 1 Some currently available approaches to measuring and enumerating stem cell

populations, and major bioengineering-related challenges that need to be addressed for the

development of stem cellbased technologiesa

Genetic/phenotypicmarkers of Major technical

Stem cell type In vivo assays In vitro assays undifferentiated cells challenges

ES cells Germ line Alkaline phosphatase Cell surface SSEA-1 Control of tissue-

competence activity, in vitro expression, OCT4 specific

developmental gene expression, differentiation,

capacity (EB expression of other growth, and

formation, tissue- early genes differentiation as

specific differ- single cells (not

entiation, etc) aggregates)

NSCs Transplantation Neurosphere assay Nestin and Markers for stem

models of neurotrophin receptor cell populations,

neuronal expression, not well acceleration of in

differentiation defined vitro proliferation

rate

HSCs Hematological Long-term culture- SCL gene expression, In vitro propagation

rescue, by initiating cell, cell surface expression of undifferentiated

(competitive) cobblestone-forming of specific antigens stem cell

repopulation of cell, and (e.g. CD34, SCA-1, populations

(lethally) colonogenic cell c-kit, etc), absence

irradiated host assays of lineage markers

animals

MSCs/other Not well defined CFU-F assay, not well Stro-1, SH-2, Not well Defining lineage

defined defined relationships, assaysof developmental

potential, and

conditions for in

vitro propagation

aES, Embryonic stem; EB, embryoid body; NSC, neural stem cell; HSC, hematopoietic stem cell; MSC, marrow stromal

cell; CFU-F, colong-forming unitfibroblast; SSEA-1, stage-specific embryonic antigen-1; SCL, stem cell leukemia;

SCA-1, stem cell antigen-1; Stro-1, a stromal precursor cell marker; SH-2, amonoclonal antibody raised against human

mesenchymal stem cells.

an appropriate culture system capable of reproducibly expanding HSC numbers

(Table 1).

Accepted techniques for quantifying HSCs have relied on assays that measure

the ability of progenitor cells to give rise to specific types and numbers of mature

blood cells. These assays can be performed in vitro or in vivo. In vitro, for ex-

ample, such assays detect a variety of apparently lineage-restricted or multipotent

progenitors by their ability to differentiate into morphologically distinct colonies

in semisolid media either immediately (colony forming cells) or, in the case of a

subpopulation of HSCs, after five or more weeks of coculture on supportive stro-mal fibroblasts (long-term culture-initiating cells) (55, 56). In vivo, both murine

and human HSCs have been detected by their ability to serially repopulate multiple

blood cell lineages in lethally irradiated host animals at limiting dilution (41, 57).

Both these assay systems build on earlier work that described a quantitative (lin-

ear) relationship between test cell populations and colonies of HSCs formed in


g




7/33


the spleen of recipient animals (58, 59). HSC populations highly enriched in these

functional properties have been isolated from murine and human bone marrow

based on their expression of cell surface antigens (for a recent review, see 60),

and comparisons of in vivo repopulating cell and long-term culture-initiating cellnumbers have indicated that, at least under some conditions, the cells detected

by these two assays are likely to represent closely related, if not overlapping,

populations (61). Despite extensive functional and physical HSC characterization,

definitive correlations between HSC developmental potential and gene expression

profiles have been more difficult to develop. It is clear that several molecules

are important in the development of specific hematopoietic lineages [e.g. expres-

sion of recombination-activating genes are essential for the development of the

adaptive immune system (62), whereas GATA-1 is important for erythroid devel-

opment (63)], or perhaps even in the maintenance of HSC potential [the stem cellleukemia gene encodes a vital stem cell regulator, which when destroyed inhibits

both embryonic (64) and adult (65) hematopoiesis]. It is likely that the properties

of these cells are a consequence of several gene products, as well as of interactions

between their regulatory pathways, an accurate description of which requires a

global dissection of HSC genetic programs (66).

NEURAL STEM CELLS Using the strict definitions outlined above for HSCs, i.e. the

capacity for extensive self-renewal and multilineage differentiation at the single

cell level, stem cells from several other tissues have more recently been described.

Cells that show some of these properties have been isolated from both the central

nervous system (67) and the peripheral nervous system (68). In vivo evidence that

these cells participate in neural development and homeostasis has been shown

using marking studies that have revealed a surprising turnover in many adult neu-

ral cell populations (69, 70). Central nervous system stem cells can be isolated

from at least two areas in adult brain (71, 72) and can be cultured in vitro using

a clonogenic assay (neurosphere assay). Because these cells can be passaged for

many generations and have been shown to maintain their potential to develop intoneurons, astrocytes, and oligodendrocytes, they have been termed neural stem cells

(NSCs). The transplantation of either undifferentiated (73) or differentiated (74)

cultured NSCs into mouse models of multiple sclerosis has shown that these cells

are capable of replacing at least some of the function of the damaged oligodendro-

cytes. NSCs defined using the neurosphere assay can also be derived from ES cells.

It is surprising that this differentiation program seems to occur in the absence of

specific signals and is inhibited by the endogenous activation of receptors from the

TGF- family (18). By combining factors that either directly or indirectly inhibit

TGF- receptors with exposure to either defined [fibroblast growth factor (FGF)-8,Sonic hedgehog, and FGF-2) or undefined (stromal cell-derivedinducing activity)

exogeniously added signals, the efficient differentiation of ES cells into NSCs may

be possible (33, 34).

Although it is becoming clear that adult cells propagated in the neurosphere cul-

ture systems can participate in neurogenesis (as well as in the development of other


g




8/33

282 ZANDSTRA NAGY

tissuessee below) in vivo (5), the soluble or cell-associated factors that promote

NSC self-renewal or lineage commitment are just beginning to be defined. It is

clear that high concentrations of either epidermal growth factor (EGF) or FGF-2

stimulate the growth of NSCs in a variety of systems, although to a varying de-gree (75). The receptors for both these proteins have in vivo expression patterns

that suggest a role in the proliferation of precursor populations in developing and

adult nervous system, although the differential sensitivity of NSC to FGF and EGF

may be indicative of distinct populations of cells (76). Several other factors, in-

cluding insulin-like growth factor-1 and LIF, have also been described to regulate

NSC proliferation (75). Unfortunately, although these factors have clear effects

on NSC responses, human-derived cells divide at a very slow rate (once every

714 days) (77, 78), severely limiting their utility in stem cellbased technologies.

[It is interesting to note that mouse-derived NSCs proliferate with a doubling timeof 1224 h (77).] This observation has led to the investigation of genetically trans-

formed NSC-like cell lines as potential alternatives to NSCs for transplantation

strategies (79).

Thecombinatorial effect of in vitro supplemented growth factors on central NSC

differentiation is not well understood, although recent studies using embryonic-

derived neural crest stem cells (developmentally part of the peripheral nervous sys-

tem) strongly support the existence of factor-directed differentiation mechanisms

(80, 81). Particularly interesting are results that show clear hierarchies between

differentiation-inducing agents (e.g. notch ligands inhibit neuronal differentiationin a manner that is dominant to bone morphogenic protein-2 differentiation) and

that even transient exposure to differentiation-inducing agents (4 h) can instruct

the fate of peripheral nervous system stem cells in 4-day cultures (80). Although

little is known about the expression of genes specific to NSCs, or even necessary

for NSC differentiation, candidate markers exist that may help in the enrichment of

NSC populations. Particularly interesting in this regard is expression in NSCs of the

neuroectodermal marker nestin (82), as well as of neuroshpere-forming cells that

can be isolated from skin (J Troma, F Akhavan, A Fernandez, D Barnabe-Heider,

D Kaplan & F Miller, unpublished results) and may additionally be expressed inadult pancreas (83). It remains to be seen whether these or other recently identified

markers (84) can be used to isolate NSCs from easily accessible adult tissues, an

end goal that would significantly facilitate the development of NSC-based tech-

nologies (Table 1).

OTHER STEM CELL SYSTEMS Marrow stromal cells (MSCs), also referred to as mes-

enchymal stem cells, are a population of nonhematopoietic cells that can be isolated

from the bone marrow and that when reintroduced in vivo are capable of differenti-

ating into osteoblasts, chondrocytes, and adipocytes (85). In vivo analysis of MSC

has been augmented by functional in vitro assays that show their capacity to form a

range of differentiated cell types. Friedenstein et al (86) initially isolated MSCs by

their adherence to tissue culture surfaces, although cells with similar properties can

be isolated from the circulation (87) and, thus, may be able to survive and proliferate


g




9/33


in suspension. Significantly, although efforts are underway to establish lineage re-

lationships for MSCs using both molecular profiles and functional outputs (88),

thus far a definitive demonstration of both self-renewal and multilineage differen-

tiation at a single-cell level is missing. It is surprising that this has not preventedwidespread adoption of the term mesenchymal stem cells, as well as publications

of their in vitro propagation (89). The colony-forming unitfibroblast (CFU-F) as-

say has been used as indicative of the numbers of marrow-derived stomal cells in a

test population; however, it is clear that this assay does not, on its own, directly de-

tect multilineage capacity or self-renewal potential. It is encouraging that cells that

arise from CFU-F have been shown to have the capacity to regenerate specific mes-

enchymal lineages, both in vitro (90) and in vivo (91), while retaining significant

proliferative capacity (92). What is now required is to combine assays for spe-

cific types of progenitors with quantitative approaches [e.g. gene marking (91),limiting dilution analysis (93)] to determine the frequency and biological prop-

erties of their multipotential precursors. In addition to describing the underlying

biology of mesenchymal cell development, these tools would facilitate efforts to

optimize the production of marrow-derived stem cells for a variety of technological

applications.

Stem cell populations have also been reported from a range of other adult

tissues, including gut (94), liver (95), and skin (96). For these cell types, to various

extents, lineage relationships have been established, assays to enumerate stem cell

developmental potential have been developed, and identification of phenotypic andgenetic markers and elucidation of cultures conditions for their in vitro propagation

have occurred. For example, although it is clear that liver regeneration can occur via

liver progenitor cells (oval cells) when hepatocyte proliferation is inhibited (97),

culture conditions for liver progenitor propagation, or even the in vivo location

of such cells, is still under investigation (98). Conversely, although it has been

conclusively shown that human epidermis is renewed throughout life from stem

cells in the basal layer of the epidermis (99), the in vitro conditions to propagate

these cells are just starting to be identified (100).

STEM CELL PLASTICITY OR EMBRYONIC LATENCY

Observations that adult-derived stem cells may be much less lineage restricted

than previously thought has clouded the boundary between the embryonic and

tissue-specific stem cells. NSCs from mouse brain, when injected into irradiated

mice, are able to repopulate the blood-forming system (4) and, indeed, seem to

be able to participate in most developmental programs (5). Bone marrowderived

stromal cells may be capable of forming glial cells and astrocytes on being in-

jected into brain (101). Similarly, bone marrowderived stem cells seem to have

the capacity to regenerate muscle (102, 103), whereas muscle-derived stem cells

repopulate the blood-forming system (2). The mechanisms behind these apparent

transformations are not clear. It is possible that cells with latent embryonic potential


g




10/33

284 ZANDSTRA NAGY

exist in adults, capable of being stimulated to differentiate into specific tissues

after microenviromental perturbations associated with transplantation. Or possi-

bly apparently lineage-restricted stem cells have more plasticity than was previ-

ously recognized (Figure 2). The recent demonstration that highly enriched HSCs(identified and isolated using markers for adult blood stem cells) are able to dif-

ferentiate into functional hepatocytes in vivo (6) supports this latter hypothesis.

The critical observation in many of these studies is that the in vivo microenvi-

ronment, either under normal conditions or when perturbed, expresses signals

that result in the manifestation of an otherwise latent developmental capacity.

The key to capitalizing on this potential will be to understand the mechanisms

behind this transformation, and to develop in vitro culture conditions where the

parameters that influence intratissue stem cell development can be identified and

modulated. Although in vitro plasticity as been observed by mixing differenti-ating ES cells with NSCs (5, 104), as well as with marrow-derived stomal cells

differentiating in neurons (101), the parameters that modulate these processes are

still poorly defined. It is likely that the identification of these parameters will

involve more than simply measuring stem cell differentiation using a variety of

functional assays. Engineering approaches, such as those developed by the tis-

sue engineering community to maintain in vivo tissue function ex vivo, should

play an important role in identifying the critical parameters that control stem cell

plasticity.

STEM CELL SELF-RENEWAL ANDLINEAGE COMMITMENT

The development of stem cellbased bioprocesses critically depends on a quanti-

tative understanding of the mechanisms that control stem cell responses. Although

it is unlikely that these mechanisms are conserved in all tissues at all stages of on-

togeny, some interesting parallels can be made between different stem cell systems.

The fact that many recent studies question notions of stem cell tissue specificity isin itself sufficient motivation to examine common regulatory mechanisms in stem

cell biology (105).

Control of Stem Cell Self-Renewal

Defining the parameters required to maintain and expand the stem cell pool is a

critical issue for future therapeutic use of stem cells because stem cells will only

be useful in the clinic if they can be grown in large numbers. In many situations,

such as the hematopoietic system, we still do not know how to maintain andexpand the stem cell pool in vitro, even though a single cell can repopulate the

whole system and expand indefinitely in vivo (45). Even in situations like that of

mouse ES cells, where apparently indefinite maintenance of the stem cell state can

be achieved, this is not readily obtainable in all strains of mice, nor is it readily


g




11/33


transferred across species. Human ES cells, for example, may not be directly

supported by soluble LIF(106), the factor necessary for mouse ES cell maintenance

(107).

There are obviously certain intrinsic properties that can influence the magnitudeof stem cell amplification detected or achieved [i.e. defects that may compromise

functional outputs, such as stem cell tissue homing (108, 109), or factors that may

limit the ultimate proliferative potential displayed by a given stem cell without

affecting its undifferentiated state (e.g. telomere shortening (110)]. However, an

important message from a large group of studies using a variety of unspecialized

cells is that the outcome of a mitotic cell division (i.e. self-renewal or loss of devel-

opmental capacity, differentiation) depends on the numbers and types of signaling

ligand-receptor complexes that are activated (111). Thus, if a cell expresses suffi-

ciently high numbers of the relevant receptors and is exposed to sufficiently highconcentrations of the cognate ligands, the probability of undergoing a self-renewal

division will be high. Conversely, if a cell expresses too few receptors or if its recep-

tors have been sufficiently downregulated, or if the cognate ligand concentrations

are sufficiently depleted, the number of ligand/receptor signaling complexes may

fall below a threshold, with the consequence that a majority of cells will undergo

commitment to a particular developmental lineage. In either situation, it is likely

that the threshold comparison must be to some time integration of the ligand-

receptor complex numbers and resulting activation of specific signal transduction

pathways. The fact that this conceptual model has been shown to be relevant indetermining the probability of a cell or a cell population undergoing self-renewal

divisions in ES cells (24), hematopoietic cells (112, 113) [including in vivo repop-

ulating HSC (114) and B-cell progenitors (115)], neural progenitors (116), and

a variety of nonmammalian stem cell systems (117, 118) supports the existence

of conserved biological mechanisms for controlling self-renewal and differentia-

tion. One especially interesting prediction from this model, with great potential for

technological applications, is that presentation of the ligand in a mode that min-

imizes ligand depletion and/or receptor downregulation can potentially enhance

the likelihood for suprathreshold complex levels and correspondingly increasedprobability of self-renewal cell division (119) (Figure 3). Examples of approaches

that may be used to accomplish these goals can be found by modulation of ligand

receptor binding affinity (120, 121), or by tethering ligand to insoluble surfaces

(119, 122).

Although it is becoming clear that threshold levels of receptor activation induced

by soluble and bound factors can modulate self-renewal responses of unspecialized

cells, what has not been clarified are the mechanisms by which such cells sense

and differentially respond to different levels of receptor activation. Recent studies

of the regulation of ES cell self-renewal and differentiation suggest that this isa promising system for analyzing this question. It is well established that ES

cells can be maintained in an undifferentiated state in culture in the presence

of high concentrations of cytokines from the interleukin-6type family (107).

Dose-response studies indicate that leukemia inhibitory factor (LIF) concentration


g




12/33

286 ZANDSTRA NAGY

Figure 3 Examples of different mechanisms that may be important in the ligand-

receptor signalling thresholddependent control of stem cell self-renewal and differ-

entiation. Exogenous and endogenous mechanisms can be related by signalling com-

plementarity or by changes in the rate or extent of cell division. ECM, extracellular

matrix; STAT, signal transducer and activator of transcription; SOCS, suppressor of

cytokine signalling.

directly influences the probability that ES cells will maintain an undifferentiated

phenotype independent of the rate of ES cell proliferation (24). In fact, in EScells, as in PC12 neuronal cells, receptor occupancy level appears to determine the

self-renewal or differentiation decision (116).

A critical step in connecting receptor occupancy to the genetic programs in-

volved in ES differentiation is signaling though the JAK/STAT (janus kinase/signal

transducer and activator of transcription) pathway. Compelling indication that

intracellular STAT3 activation is involved in LIF-mediated changes in ES cell

self-renewal comes from the observation that a threshold level of extracellularly

induced STAT3 activation is essential for this response (123). Dimerization of

gp130 by LIF induces both the ras-mitogenactivated protein kinase (MAPK) andJAK/STAT pathways in ES cells (124). The results of Raz et al (123) suggest that

the level of activated STAT3 plays a critical role in maintaining a block of ES

cell differentiation, ras-mitogenactivated protein kinase activity being predomi-

nantly associated with mitogenesis. What is not clear from these studies is how the


g




13/33


potency of the mitogenic stimulus required to maintain ES cell pluripotency may

be related to the activation of LIF receptors during ES cell commitment.

The idea that different levels of stimulation by exogenous cytokines modulate

cell fate decisions through activation of diverse signaling events is not a new one(116). Receptors with intrinsic or associated tyrosine kinase activity are known to

elicit both proliferative and differentiation responses in factor-dependent cell lines,

and it is likely that attaining threshold levels of altered signaling intermediates is

important in determining the final biological outcome. For example, both the du-

ration and magnitude of extracellular signalregulated kinase (Erk) activation by

nerve growth factor and EGF in the PC12 neuronal cell line determines whether

these cells will proliferate or differentiate (125). Transient activation of Erk by

EGF stimulates proliferation of PC12 cells, whereas they differentiate in response

to nerve growth factor, which acts partly by inducing a sustained activation ofErk. Because, through internalization and phosphorylation, the EGF receptor is

more rapidly downregulated than the nerve growth factor receptor (126), these

results suggest that it is the number of active cell surface receptors that determines

whether Erk activation is sustained and differentiation results. It is important to

note that threshold-dependent regulation may also extend to osteogenesis, where

both the magnitude and duration of parathyroid hormone supplementation mod-

ulate bone development (127). These effects, which result at least in part from

the differential stimulation of adenylyl cyclase and phospholipase C, are depen-

dent on the density of parathyroid hormone receptor expression on the cell surface(127129). The role of inductive membrane-associated or soluble concentration

gradients in activating distinct genetic programs during embryonic development

and tissue specification is also well documented (for recent reviews, see 118, 130).

Cell-secreted factorsmorphogensform relatively stable concentration gradi-

ents over distances of more than 300 m and thereby elicit positional information

that dictates tissue patterning. Examples of molecules that form gradients result-

ing in spatially distinct tissue specification during vertebrate development include

members of the transforming growth factor (TGF)- family (i.e. activin, TGF-,

and bone morphogenic proteins). Mesoderm development in Xenopus is inducedby treating presumptive ectoderm with activin. Low concentrations result in the

induction of hematopoietic tissue, whereas very high concentrations induce the de-

velopment of notochord (131). Both the ability of cells to respond to many factors,

or factor complexes with overlapping activities, and the existence of complex inter-

secting concentration gradients of these factors suggest that relative concentrations

of factors, not just their absolute magnitudes, are important. The observation that

lineage determination during neural crest stem cell differentiation is instructively

influenced by the timing and relative dosage of growth factor encountered (105)

further supports this model.Taken together, the above studies support the notion that the relative potency of

a particular family of cytokines or growth factors to elicit stem cell self-renewal

depends exquisitely on the biochemical properties of the receptor-ligand pair. Said

another way, there may be several different cytokines or growth factors that are


g




14/33

288 ZANDSTRA NAGY

capable of eliciting self-renewal division of stem cells (as long as they activate the

right complementary signal transduction pathway), but the cellular response on

addition will depend on the manner in which the factor is presented and maintained

in the cellular microenvironment. An example of this can be found once again inthe ES cell model, where the concentrations of LIF and interleukin-11 required

to prevent differentiation differ significantly, although these cytokines utilize the

same gp130 transmembrane molecule for signaling (132).

Control of Lineage Commitment

Besides learning how to maintain pools of stem cells, the development of stem

cellbased technologies will also depend on our ability to reproducibly drive their

differentiation into specific tissue lineages. This issue is of particular importancefor the development of ES cellbased technologies, where efficiencies of differen-

tiation into particular subsets of cells and tissues is very low. The generation of ES

cellderived cardiomyocytes (currently the only feasible approach for the in vitro

generation of large amounts of cardiac tissue) is one example where significant

improvements in tissue-specific differentiation are needed (under default mecha-

nisms only 1% of ES cellderived cells form cardiac tissue). Control of lineage

specificity is also likely to become critically important if the clinical potential of

adult stem cell plasticity is to be realized.

Regulation of hematopoietic cell proliferation and differentiation in vivo al-lows the continuous and controlled production of large populations of mature and

maturing cells from a small population of stem cells, as required during develop-

ment and later in response to physiological demand. This process, which implies

the existence of direct or indirect feedback control mechanisms (94), involves the

production and action of a large group of cytokines, many of which have been

characterized and produced in recombinant form. Because HSCs are segregated

among different bones and organs throughout the body, for the stem cell pool to

be regulated in a coordinated fashion, at least some of these controlling factors

must act independently of direct cellular contact. These molecules can exert someof their effects as single agents, but they can also act in synergistic or additive

combinations (133).

Differentiation of individual cells involves the exit from the uncommitted state

and a commitment to a particular developmental pathway. A large body of evi-

dence supports this loss of developmental potential being independently regulated

from commitment to a particular lineage (134). Evidence exists, with ES cells be-

ing a prime example, that in vitro differentiation can occur via a default pathway

(and may not require cell division), and the identity of the ensuing cell popula-

tion can be stochastically determined and their subsequent survival dependent onthe microenvironment in which the cells find themselves (18). This process has

been referred to as a permissive/selective mechanism of differentiation control.

Alternatively, cellular commitment to particular lineages may be initiated by the

exposure of an uncommitted cell to a particular signal that, on exit from the stem


g




15/33


cell state, stabilizes the subsequent identity of the progeny. In this situation, stem

cell self-renewal may be the default mechanism, differentiation only occurring in

the presence of inductive factors (135). This so-called directive mechanism clearly

depends on the interactions between the cellular sensing systemstypically cellsurface receptorsand the cellular microenvironment.

Significant evidence exists that both these mechanisms may occur and that the

particular mechanism used may be dependent on the tissue system/cellular mi-

croenvironment. In the hematopoietic system, several in vitro studies (136, 137)

support previously documented evidence showing stochastic commitment pat-

terns in vivo during spleen-CFU production (138) and suggest that the addition

of particular cytokines to culture systems does not influence the range and rel-

ative frequencies of cells generated from multipotent cells. This conclusion is

supported by observations in the multipotent hematopoietic cell line, FDCP-Mix,where transfection of bcl-2 (which blocks apoptotic cell death) yielded multilin-

eage differentiation in the absence of cell division and/or cytokines (139). These

data suggest that exposure to growth factors may not be obligatory for the differen-

tiation of primitive cells and that, at least under certain conditions, the identity of

the differentiated cell population may simply be a function of the developmental

potential of the test cell. Particularly interesting in this regard is the recent demon-

stration that coexpression of multiple lineage-restricted genes occurs, albeit at a

low level, in multipotent progenitors (134, 140). This so-called multilineage prim-

ing process implies that stem cell commitment to a particular pathway may reflectthe stabilization of a particular subset of expressed molecules (Figure 4). This sta-

bilization process may occur stochastically in the absence of an instructive signal,

the relative rates of transcription factor synthesis and chemical stability of the dif-

ferent transcription factors or their precursors determining the lineage distribution

of the differentiated cell types (141). This situation has been convincingly demon-

strated for lymphoid progenitors, where, in the absence of exogenously added

cytokines, the survival of differentiated bcl-2transfected pro-B Pax/ cells de-

pends on the transcription factor profile of the parent cell (142). Alternatively,

on commitment of an undifferentiated cell, an instructive signal may stabilize aparticular set or subset of expressed transcription factors, resulting in the produc-

tion of specific cell types. Evidence that this latter mechanism is also operative

was postulated many years ago to explain the increasing sensitivity of differen-

tiating erythroid progenitors to erythropoietin (143). Experiments using defined

combinations of cytokines with the same pro-B Pax/ cells referred to above

(142), as well other immortalized hematopoietic cell lines (144), also support this

mechanism.

It is important to note that the development of stem cell technologies for the

generation of differentiated cells will require the identification of culture condi-tions that both increase the probability of stem cell differentiation into a particular

lineage and ensure the survival of the resultant (functional) cell population. A

quantitative framework for the analysis of stem cell differentiation will have to

integrate stochastic approaches with mechanistic models. At it simplest level, an


g




16/33

290 ZANDSTRA NAGY

uncommitted cell may express, at a low level, a range of transcription factors as-

sociated with the spectrum of its developmental potential. The relative expression

levels, intracellular stabilities, and exogenous stabilization (either by the addition

of supplementary factors or through feedback mechanisms from other cell popula-tions) will then determine the probability of commitment to one lineage or another.

These gene expression profiles can be assayed directly [i.e. osteocalsin expression

for bone development (145) or brachyury expression for mesoderm induction (30)]

or indirectly by the analysis of cell function and phenotype. One approach that

may be useful to model and understand aspects of stem cell commitment involves

the adaptation of maximum likelihood methods previously used for the analysis

of molecular phylogeny and evolutionary relationships (146). This method incor-

porates the probabilistic nature of branches or differentiation pathways while

allowing the flexibility to incorporate preferred pathways by error minimization.To test whether this model is applicable to stem cell differentiation, we have applied

it to published data of bone cell development (147). Our results have confirmed the

existence of differentiation routes that are consistent with established gene expres-

sion profiles during osteoprogenitor development (i.e. it is known that osteopontin

is expressed early on bone development and that osteocalcin is expressed late in

bone development). It is important to note that they also support experimental

evidence suggesting that multiple differentiation pathways may lead to similar de-

velopmental endpoints (JE Aubin, PW Zandstra & N Madras, unpublished results).

By understanding the processes that influence stem cell lineage commitment, thesestudies should help identify parameters critical for the design of stem cellbased

bioprocesses (i.e. cytokine supplementation regimes, differentiation mechanisms,

etc).

STEM CELL CULTURE SYSTEMS

The need for improved cultivation methods is driven by the fact that many po-

tential therapeutic applications are limited by the availability of stem cells or

their derivatives. To realize the promise of stem cell therapeutics, the culture vol-

umes of basic biology (1 ml) need to be adapted to clinically relevant scales.

Cellular therapy reactor scales may be 1000 liters, mainly depending on

whether single cultures can be used for multiple patients. The biotechnology indus-

try favors suspended or aggregated mammalian cell cultures in stirred bioreactors

readily used for large-scale production. Stirred suspension culture systems offer

attractive advantages of ready scalability and relative simplicity. Their relatively

homogeneous nature makes them uniquely suited for investigations of different

culture parameters (e.g. O2 tension, cytokine concentration, serum components,

medium exchange rates, etc) that may influence the viability and turnover of spe-

cific stages and types of stem cells (Figure 5). Promising results on the growth

of hematopoietic (148), neural (149), and embryonic (150) stem and progeni-

tor cell numbers in stirred suspension bioreactors cultures have been obtained.


g




17/33


Figure 5 Advantages and different classical modes of operation for stirred suspension

bioreactor systems. Although compared with many culture systems being developed for

functional differentiated tissues (e.g. liver, skin, etc) these culture systems are simple, they

may be ideal for the scalable generation of undifferentiated stem cells.

Similar approaches are also being used to show that marrow-derived stromalcells (detected as CFU-F) can also be expanded in stirred suspension cultures

(151). It is interesting that, in most cases, these suspension culture systems do

not mimic the complex three-dimensional microenvironment of adult tissues and

organs (a major goal of most tissue engineering efforts). It is likely that these

adult microenvironments are highly damped in their ability to initiate extensive

stem cell self-renewal, a phenomena that occurs in vivo predominantly during

development or in response to injury. Although many studies have shown that

lineage commitment and mature cell function are stabilized by cell-cell, cell-

extracellular matrix interactions, stem cell self-renewal may be more efficient

in the absence of these signals. Direct evidence of this feedback-induced dif-

ferentiation has been shown in ES cells using selection agents driven by tissue-

specific promoters (36, 37), and the effect of mature cell populations, or their

products, on HSC differentiation is well documented (152). Mathematical models

(153) that examine strategies to grow rare cell populations in mixed cell cultures

suggest using population fluxes as driving forces to maintain equilibrium dis-

tribution of cells, as has been shown for HCS cultures initiated with very differ-

ent input cell populations (154). This Le Chateliers principlelike regulation

implies that a bioprocess where recognizably mature cells are continuously re-

moved from the culture system, returning undifferentiated cells to the culture

vessel, could result in the establishment of conditions that favor particular cul-

ture endpoints (e.g. self-renewal over differentiation, one lineage over another). In

this scenario, growth factors can be thought of as catalysts to enhance particular


g




18/33

292 ZANDSTRA NAGY

microenvironments, with the possibility that once critical numbers of particular

cells are generated, the culture system may no longer require exogenous factors to

maintain itself.

Cytokines and Growth Factors

Stem cell properties are the result of their expression of a specific subset of genes,

changes in the expression of which determine exit from the stem cell compartment

into functional cell lineages. As reviewed above, although there is still much to

learn about the genes involved in such changes (as well as how they are regulated),

it is clear that stem cells interact with many molecules in their extracellular milieu

via transmembrane receptors (or receptor complexes) to maintain their viability,

and to effect change in their cell cycle progression and differentiated state. A keyfeature of any stem cell culture system is the combination of cytokines it delivers to

the microenvironment of the cells, and how the concentrations of these cytokines

and their associated receptors are maintained over time.

Significant efforts have been made to define cytokine and growth factor supple-

mentation strategies to control stem cell responses. The cytokine composition of

the medium is particularly challenging to optimize in stem cell cultures because

multiple cell types compete for several cytokines that each influence stem cell fate

directly or indirectly. Factorial (155) and composite design methods have been

used to make the empirical investigations more efficient and to reveal unexpectedinteractions missed by conventional dose-response analysis. These engineering

process optimization methods can be applied to the analysis of complex systems

whose mechanistic details are not fully understood. Besides optimizing the stem

cell cultures for particular outputs, these systematic empirical methods have re-

vealed unexpected interactions, further defining the mechanistic basis for stem

cell regulation. Examples of results where this approach has been successful in-

clude the following: the identification of particular factors that influence HSC self-

renewal (i.e. flt-3 ligand) or directed differentiation (i.e. thrombopoietin) (156); the

demonstration that the cytokine requirements of HSCs may change during human

ontogeny (155), a result that has also been observed during NSC development (76);

and the observation of a threshold cytokine concentration effect for HSC (and ES

cell) self-renewal and differentiation (111).

In addition to identifying the types and concentrations of cytokines that need to

be supplemented to elicit specific stem cell responses, the dynamic nature of cy-

tokine supplementation needs to be taken into account (112, 157, 158). Formation

of ligand-receptor complexes results in the recruitment and activation of specific

intracellular molecules that then initiate different signaling pathways. At any spe-

cific point in time [and spatial position (159)], the number of ligand-receptor

complexes per cell depends on two variables (the number of unoccupied re-

ceptors available and the ligand concentration) and one parameter (the ligand-

receptor binding affinity) (160). However, the two variables can change as a func-

tion of time and/or spatial position, so that the number or signaling efficiency of


g




19/33


ligand-receptor complexes can consequently change. Thus, mechanisms that

govern cell receptor number and ligand concentration can be predicted to cor-

respondingly govern whether the resulting complex-activated signal(s) remain(s)

at a particular stimulation level(s). For example, if binding affinity is sufficientlygreat so that the ligand/receptor complex remains stable during the usually rela-

tively brief time required to internalize the receptor (10 min) (160), the ligand

may be rapidly depleted from the extracellular milieu by cellular endocytic degra-

dation of receptor-ligand complexes (161). This internalization and degradation

of receptors can also result in their downregulation (158). Simultaneously, on the

timescale of a cell division cycle, newly synthesized and/or recycled receptors can

typically be reexpressed on the cell surface (120). Sustained signal propagation

can, itself, result in protein-mediated desensitization of the complex to further

signaling (162). Hence, by careful quantification of changes in receptor levelsto improve understanding of the parameters that influence these changes, it may

be possible to develop rational strategies to enhance stem cell self-renewal divi-

sions. Of course, in the complex culture systems typical of stem cell cultures, not

only does the dynamics of stimulation of the supplemented cytokine have to be

taken into account, the production of cytokine and growth factors, both known

and both yet to be identified, are likely to effect the long-term outcome of these

cultures.

Physicochemical Parameters

In addition to the protein microenvironment, it is clear that such physicochemi-

cal parameters as pH, oxygen, and glucose have significant effects on stem cell

responses. These responses can be direct or due to transcriptional events initiated

during particular culture conditions. The establishment and subsequent develop-

ment of oocytes and embryos in vitro is inhibited by high glucose concentrations,

although it is not clear whether this effect is due to the generation of reactive oxy-

gen species by glucose oxidation and phospho-ribosyltransferase inhibition or a

direct consequence of glucose (163). The differentiation of ES cells into cardiac

and hematopoietic cells has been shown to be affected by oxygen or its byproducts

(164, 165). Cardiomyocyte differentiation is enhanced by reactive oxygen species,

presumably by the activation of phosphatidyllinositol 3-kinase signaling (165),

whereas low oxygen concentrations may enhance ES cellderived hematopoiesis,

likely by initiating or enhancing the expression of specific cytokines, such as vas-

cular endothelial growth factor (164, 166). It is becoming clear that a majority of

cellular responses to hypoxia, including expression of erythropoietin and vascular

endothelial growth factor, involve the transcriptional activation of gene expression

by hypoxia inducible factor-1 (167).

Numerous studies have documented effects of different oxygen levels on the

proliferation of hematopoietic cells in vitro, both in semisolid and in liquid cul-

tures. These studies have been motivated primarily by the observation that, in

vivo, oxygen tensions in bone marrow are on the order of 2%7%; they have


g




20/33

294 ZANDSTRA NAGY

also been motivated by reports that compared with parallel cultures maintained in

atmospheric oxygen, oxygen tensions ranging from 1% to 10% can enhance

both the size and number of hematopoietic colonies obtained in semisolid media

(168, 169). Increased growth of hematopoietic cells at low oxygen concentrationshas been attributed, in part, to an increased responsiveness of cells to growth fac-

tors and to a decreased generation of hydrogen peroxide, oxygen radicals, and

other oxygen intermediates produced by mature granulocytes (170). The obser-

vation that more mature cells may be less sensitive to changes in their oxygen

environment than their immediate precursors (171) suggests that the definition

of an optimal oxygen tension may depend not only on the cytokines used to

stimulate the cells, but also on the developmental potential of the cells being

stimulated. For example, in conditions where there was little or no net expansion

of HSCs, low oxygen conditions resulted in the maintenance of higher numbersof these cells (171), whereas greater HSC expansions (172) occur under con-

ditions that included higher oxygen concentrations (173). Overall, these results

indicate that oxygen tension is likely an important, although still incompletely

understood, determinant of hematopoietic cell production. Additional investiga-

tions of the effects of controlled manipulations of oxygen tension in parallel with

measurements of oxygen concentrations in the cellular environment should be

informative, particularly in relation to differences in the types of cells present

and/or being stimulated. Low-medium oxygen concentrations have also been ob-

served to inhibit the propagation of neurospheres, potentially containing NSCs,although it is not clear whether the reported effects are due to cell death at

the core of these multicellular aggregates or to a specific effect on stem cell

responses (149).

Traditional static HSC cultures initiated with large inocula of bone marrow

derived cells and fed by periodic replacement of part of the culture medium with

fresh medium can be limited by nutrient and cytokine depletion, and by expo-

sure to inhibitory compounds (112, 174). Increasing the medium exchange rate

in HSC culture systems increases total cell and progenitor cell output (154, 172).

Cell-specific glucose consumption and lactate production both increase with fastermedium exchange rates and increased serum supplementation (175). On the other

hand, the faster introduction of fresh medium also dilutes metabolic by-products

and inhibitory factors more rapidly. Many studies have shown that medium acid-

ification causes growth inhibition. Some indication of differences in the survival

and differentiation of hematopoietic progenitors into the granulocyte-macrophage

(optimal pH 7.27.4) versus erthythroid lineages (optimal 7.6) have been doc-

umented (176). Recent attempts to untangle the effects of lactate production, pH,

and medium utilization on hematopoietic cell cultures have shown that each of

these parameters likely influences stem and progenitor cell responses. The factthat strong pH-associated effects have also been observed in suspension cultures

of neurospheres (149) stresses the importance of measuring and controlling these

parameters in the development of stem cellbased technologies.


g




21/33


CONCLUSIONS AND FUTURE DIRECTIONS

The results presented here have several implications for the design of bioprocesses

that hope to take advantage of the developmental potential of undifferentiatedcells. First, the importance of the mode of cytokine/growth factor presentation to a

factor-dependent cell population indicates that it is unlikely there will be a magic

bullet thatwhen simply supplemented to in vitro cultureswill be capable

of controlling stem cell fate. The corollary to this hypothesis is, of course, that

there may be several different ways to obtain the required level(s) of stimulation.

Ultimately, it is likely that the outcome of the presentation of a mitogenic signal to

a pluripotent stem cell will depend, at least in part, on the following: (a) the type

of signal presented (i.e. the types of cytokines and the physical environment of the

cell), (b) the magnitude of the signal (i.e. the concentration at which cytokines arepresented), and (c) the method of signal presentation (i.e. surface-bound versus

soluble, as indicated by associated differences in mitogenic responses and in the

duration of tyrosine phosphoralization). Each of these parameters represents a

unique opportunity to investigate the role of environmental control on cellular

survival and mitogenic and differentiation responses. An understanding of the

molecular mechanisms that control cellular fate processes through the modeling,

design, and analysis of the kinetics of novel growth factor delivery strategies should

facilitate the rational development of bioreactors capable of culturing primary

human tissues.In addition to supplementation strategies, and because cytokine concentrations

appear to be of major importance, stem cellbased culture systems will need to

take into account the changing rates of depletion of each cytokine required. Thus,

continuous monitoring of cytokine depletion may be essential for maintaining cy-

tokine concentrations at levels that optimize the production of specific types of

undifferentiated cells. Perfusionbioreactor systems for controlled medium replace-

ment may also be necessary to optimize cytokine concentrations over time. These

systems have the additional advantage that, depending on the medium exchange

rate, changes in medium formulations and the removal of inhibitory products canbe performed online.

An additional consequence of stimulating cells with high concentrations of

cytokines is to temporally modulate their capacity to continue to respond to the

same cytokines (e.g. through changes in cell surface receptor density). The design

of soluble or surface-bound ligands that prevent or limit receptor internalization,

along with investigations of the biological consequences of stimulation with such

molecules, may help achieve sustained expansions of self-renewing stem cells.

Finally, the observations that different types and concentrations of cytokines may

influence the production of functionally distinct progenitors, along with the possi-bility that these cytokines may effect cell responses both directly and through their

manipulation of cell cycle rates, underscores the need for more detailed informa-

tion about how these different responses are elicited. Only through an integration


g




22/33

296 ZANDSTRA NAGY

of these parameters, such as is possible by combining experimental strategies with

kinetic modeling approaches, can robust strategies for the development of stem

cellbased technologies be developed.

A particular challenge for the bioengineering community in the culture of stemcells in complex media that contains multiple growth factors remains the cost of

bioprocess development. Developing scalable bioreactor systems will require new

strategies for the efficient and economic delivery of survival, self-renewal, and

differentiation signals to cells in serum-free medium. Evidence that cell popula-

tion fluxes are exquisitely responsive to microenvironmental conditions, including

increasing evidence of environmentally mediated stem cell plasticity, confirms

the importance of carefully selecting environmental conditions when culturing

stem cells, especially because the analytical window is often limited to a partic-

ular tissue type. Process optimization using microarray technologies may be oneapproach in which this may be efficiently accomplished. Finally, in addition to

these developments, it is clear that stem cellbased technologies will have to be

responsive to ongoing developments in therapeutic cloning and gene therapeutics.

In order to make nuclear transfer technologies efficient, reproducible, and predi-

cable, the possibility of artificially generating stem cells from individual patents

to meet particular clinical needs will take a thorough understanding of the culture

requirements.

ACKNOWLEDGMENTS

We thank Derek van der Kooy for helpful suggestions and valuable critique of our

manuscript. This work was funded by by the Natural Sciences and Engineering

Research Council (NSERC) of Canada, the National Aeronautics and Space Ad-

ministration (NASA), and the National Science Foundation (NSF) (Engineering

Research Center grant to the Biotechnology Process Engineering Center at the

Massachusetts Institute of Technology).

Visit the Annual Reviews home page at www.AnnualReviews.org

LITERATURE CITED

1. Nordon RE, Schindhelm K. 1999.Cell se-

paration. See Ref. 177, pp. 21542

2. Jackson KA, Mi T, Goodell MA. 1999.

Hematopoietic potential of stem cells iso-

lated from murine skeletal muscle. Proc.

Natl. Acad. Sci. USA 96:14482863. Eglitis MA, Mezey E. 1997. Hematopoi-

etic cells differentiate into both microglia

and macroglia in the brains of adult mice.

Proc. Natl. Acad. Sci. USA 94:4080

85

4. Bjornson CR, Rietze RL, Reynolds BA,

Magli MC, Vescovi AL. 1999. Turning

brain into blood: a hematopoietic fate ad-

opted by adult neural stem cells in vivo.

Science 283:53437

5. Clarke DL, Johansson CB, Wilbertz J,Veress B, Nilsson E, et al. 2000. Gener-

alized potential of adult neural stem cells.

Science 288:166063

6. Lagasse E, Connors H, Al-Dhalimy M,

Reitsma M, Dohse M, et al. 2000.


g




23/33


Purified hematopoietic stem cells can dif-

ferentiate into hepatocytes in vivo. Nat.

Med. 6:122934

7. Metcalf D. 1999. Stem cells, pre-

progenitor cells and lineage-committed

cells: are our dogmas correct? Ann. NY

Acad. Sci. 872:289304

8. Martin GR. 1981. Isolation of a pluripo-

tent cell line from early mouse embryos

cultured in medium conditioned by tera-

tocarcinoma stem cells. Proc. Natl. Acad.

Sci. USA 78:763438

9. Evans MJ, Kaufman MH. 1981. Estab-

lishment in culture of pluripotential cellsfrom mouse embryos.Nature 292:15456

10. Bradley A, Evans M, Kaufman MH,

Robertson E. 1984. Formation of germ-

line chimaeras from embryo-derived ter-

atocarcinoma cell lines. Nature 309:255

56

11. Nagy A, Gocza E, Diaz EM, Prideaux

VR, Ivanyi E, et al. 1990. Embryonicstem

cells alone are able to support fetal devel-

opment in the mouse. Development 110:81521

12. Nagy A, Rossant J, Nagy R, Abramow-

Newerly W, Roder JC. 1993. Derivation

of completely cell culture-derived mice

from early-passage embryonic stem cells.

Proc. Natl. Acad. Sci. USA 90:842428

13. Tanaka S, Kunath T, Hadjantonakis AK,

Nagy A, Rossant J. 1998. Promotion

of trophoblast stem cell proliferation by

FGF4. Science 282:20727514. Palacios R, Golunski E, Samaridis J.

1995. In vitro generation of hematopoietic

stem cells from an embryonic stem cell

line. Proc. Natl. Acad. Sci. USA 92:7530

34

15. Wang R, Clark R, Bautch VL. 1992. Em-

bryonic stem cell-derived cystic embryoid

bodies form vascular channels: an in vitro

model of blood vessel development. De-

velopment114:30316

16. Fleischmann M, Bloch W, Kolossov E,

Andressen C, Muller M, et al. 1998. Card-

iac specific expression of the green fluo-

rescent protein during early murine em-

bryonic development. FEBS Lett. 440:

37076

17. Rohwedel J, Maltsev V, Bober E, Arnold

HH, Hescheler J, Wobus AM. 1994. Mus-

cle cell differentiation of embryonic stem

cells reflects myogenesis in vivo: devel-

opmentally regulated expression of myo-

genic determination genes and functional

expression of ionic currents. Dev. Biol.

164:87101

18. Tropepe V, Hitoshi S, Sirard C, Mak T,

Rossant J, van der Kooy D. 2001. Direct

neural fate specification from embryonic

stem cells: a premature neural stem cellstage acquired through a default mecha-

nism. Neuron. In press

19. Bain G, Kitchens D, Yao M, Huettner

JE, Gottlieb DI. 1995. Embryonic stem

cells express neuronal properties in vitro.

Dev. Biol. 168:34257

20. Keller GM. 1995. In vitro differentiation

of embryonic stem cells. Curr. Opin. Cell

Biol. 7:86269

21. Schuldiner M, Yanuka O, Itskovitz-EldorJ, Melton DA, Benvenisty N. 2000.

Effects of eight growth factors on the dif-

ferentiation of cells derived from human

embryonic stem cells. Proc. Natl. Acad.

Sci. USA 97:1130712

22. Wiles MV, Johansson BM. 1999. Embry-

onic stem cell development in a chem-

ically defined medium. Exp. Cell Res.

247:24148

23. Solter D, Knowles BB. 1978.Monoclonalantibody defining a stage-specific mouse

embryonic antigen (SSEA-1). Proc. Natl.

Acad. Sci. USA 75:556569

24. Zandstra PW, Le HV, Daley GQ, Grif-

fith LG, Lauffenburger DA. 2000.

Leukemia inhibitory factor (LIF) con-

centration modulates embryonic stem

cell self-renewal and differentiation in-

dependently of proliferation. Biotechnol.

Bioeng. 69:60717

25. Scholer HR, Ciesiolka T, Gruss P. 1991.

A nexus between Oct-4 and E1A: impli-

cations for gene regulation in embryonic

stem cells. Cell 66:291


g




24/33

298 ZANDSTRA NAGY

26. Yeom YI, Fuhrmann G, Ovitt CE, Brehm

A, Ohbo K, et al. 1996. Germline regu-

latory element of Oct-4 specific for the

totipotent cycle of embryonic cells. De-

velopment122:881

27. Pesce M, Gross MK, Scholer H. 1998.

IN line with our ancestors: Oct-4 and the

mammalian germ. Bioassays 20:722

28. Nichols J, Zevnik B, Anastassiadis K,

Niwa H, Klewe-Nebenius D, et al. 1998.

Formation of pluripotentstem cells in the

mammalian embryo depends on the POU

transcription factor Oct4. Cell 95:379

29. Palmieri SL, Peter W, Hess H, ScholerHR. 1994. Oct-4 transcription factor is

differentially expressed in the mouse em-

bryo during establishment of the first 2

extraembryonic cell lineages involved in

implantation. Dev. Biol. 166:259

30. Niwa H, Miyazaki J-I, Smith AG. 2000.

Quantitative expression of Oct3/4 defines

differentiation, dedifferentiation or self-

renewal of ES cells. Nat. Genet. 24:372

31. Thompson S, Clarke AR, Pow AM,Hooper ML, Melton DW. 1989. Germline

transmission and expression of a corrected

HPRT gene produced by gene targeting in

embryonic stem cells. Cell 56:31321

32. Reubinoff BE, Pera MF, Fong CY,

Trounson A, Bongso A. 2000. Embry-

onic stem cell lines from human blasto-

cysts: somatic differentiation in vitro.Nat.

Biotechnol. 18:399404; Erratum. 2000.

Nat. Biotechnol. 18(5):55933. Lee SH, Lumelsky N, Studer L, Auer-

bach JM, McKay RD. 2000. Efficient

generation of midbrain and hindbrain neu-

rons from mouse embryonic stem cells.

Nat. Biotechnol. 18:67579

34. Kawasaki H, Mizuseki K, Nishikawa

S, Kaneko S, Kuwana Y, et al. 2000.

Induction of midbrain dopaminergic neu-

rons from ES cells by stromal cell-derived

inducing activity. Neuron 28:3140

35. Li M, Pevny L, Lovell-Badge R, Smith

A. 1998. Generation of purified neural

precursors from embryonic stem cells by

lineage selection. Curr. Biol. 8:97174

36. Klug MG, Soonpaa MH, Koh GY, Field

LJ. 1996. Genetically selected cardiomy-

ocytes from differentiating embronic stem

cells form stable intracardiac grafts. J.

Clin. Invest. 98:21624

37. Mountford P, Nichols J, Zevnik B,

OBrien C, Smith A. 1998. Maintenance

of pluripotential embryonic stem cells by

stem cell selection. Reprod. Fertil. Dev.

10:52733

38. Rathjen PD, Lake J, Whyatt LM, Bett-

ess MD, Rathjen J. 1998. Properties and

uses of embryonic stem cells: prospects

for application to human biology andgene therapy. Reprod. Fertil. Dev. 10:31

47

39. Till JE, McCulloch EA. 1980. Hemopoi-

etic stem cell differentiation. Biochim.

Biophys. Acta 605:43159

40. Dick JE, Magli MC, Huszar D, Phillips

RA, Bernstein A. 1985. Introduction of

a selectable gene into primitive stem cells

capable of long-term reconstitution of the

hemopoietic system of W/Wv mice. Cell42:7179

41. Szilvassy SJ, Humphries RK, Lansdorp

PM, Eaves AC, Eaves CJ. 1990. Quan-

titative assay for totipotent reconstituting

hematopoietic stem cells by a competitive

repopulation strategy. Proc. Natl. Acad.

Sci. USA 87:873640

42. Keller G, Snodgrass R. 1990. Life span of

multipotential hematopoietic stem cells in

vivo. J. Exp. Med. 171:14071843. Lemischka IR, Raulet DH, Mulligan

RC. 1986. Developmental potential and

dynamic behavior of hematopoietic stem

cells. Cell 45:91727

44. Pawliuk R, Eaves C, Humphries RK.

1996. Evidence of both ontogeny and

transplant dose-regulated expansion of

hematopoietic stem cells in vivo. Blood

88:285258

45. Iscove NN, Nawa K. 1997. Hematopoi-

etic stem cells expand during serial trans-

plantation in vivo without apparent ex-

haustion. Curr. Biol. 7:8058

46. Fraser CC, Szilvassy SJ, Eaves CJ,


g




25/33


Humphries RK. 1992. Proliferation of tot-

ipotent hematopoietic stem cells in vitro

with retention of long-term competitive

in vivo reconstituting ability. Proc. Natl.

Acad. Sci. USA 89:196872

47. Zandstra PW, Eaves CJ, Piret JM. 1999.

Environmental requirements of hemato-

poietic progenitor cells in ex vivo ex-

pansion systems. See Ref. 177, pp. 246

64

48. Hoffman R. 1999. Progress in the devel-

opment of systems for in vitro expansion

of human hematopoietic stem cells. Curr.

Opin. Hematol. 6:1849149. Piacibello W, Sanavio F, Garetto L, Sev-

erino A, Dane A, et al. 1998. Differen-

tial growth factor requirement of primitive

cord blood hematopoietic stem cell for

self-renewal and amplification vs prolif-

eration and differentiation. Leukemia 12:

71827

50. Rappold I, Watt SM, Kusadasi N, Rose-

John S, Hatzfeld J, Ploemacher RE. 1999.

Gp130-signaling synergizes with FL andTPO for the long-term expansion of cord

blood progenitors. Leukemia 13:203648

51. Yagi M, Ritchie KA, Sitnicka E, Storey

C, Roth GJ, Bartelmez S. 1999.Sustained

ex vivo expansion of hematopoietic stem

cells mediated by thrombopoietin. Proc.

Natl. Acad. Sci. USA 96:812631

52. Verstegen MM, van Hennik PB, Terpstra

W, van den Bos C, Wielenga JJ, et al.

1998. Transplantation of human umbilicalcord blood cells in macrophage-depleted

SCID mice: evidence for accessory cell

involvement in expansion of immature

CD34+CD38- cells. Blood91:196676

53. Bonnet D, Bhatia M, Wang JC, Kapp

U, Dick JE. 1999. Cytokine treatment or

accessory cells are required to initiate

engraftment of purified primitive human

hematopoietic cells transplanted at lim-

iting doses into NOD/SCID mice. Bone

Marrow Transplant. 23:2039

54. Zandstra PW. 2000. Challenges involved

in the development of stem cell-based

technologies: hematopoiesis, osteogene-

sis and beyond. In Bone Engineering, ed.

JE Davies. Toronto: Emsquared. pp. 94

106

55. Sutherland HJ, Lansdorp PM, Henkel-

man DH

Stem Cell Bio Engineering

Documents