For personal use. Only reproduce with permission from Elsevier Ltd

Articles

IntroductionThe therapeutic application of human embryonic stem-cell derivatives is hindered by the exposure of existinglines to animal and human cells. In the USA, all NIH-approved human embryonic stem-cell lines wereisolated, and continue to be maintained, with mouseembryonic fibroblasts as feeder layers or as a source ofconditioned medium for propagation of the cells in theundifferentiated state (or both). Exposure of these linesto live animal cells presents a risk of contamination withretroviruses and other pathogens that could betransmitted to the patient and the wider population.Although human-feeder systems1–4 and feeder-freesystems5–7 have been reported, these approachesgenerally require addition of feeder-conditionedmedium (which can contain potential pathogens), or at aminimum require extensive screening of donor sourcesfor derivation of potential pathogen-free humanembryonic stem cells. Use of feeder layers also limitsstem-cell research design since experimental data mightresult from a combined embryonic stem-cell and feeder-cell response to various stimuli. Amit and colleagues7

and Xu and colleagues8 described a system for feeder-free and conditioned medium-free culture of humanembryonic stem cells, although these and all other suchcell lines reported have been exposed to feeder cellsduring derivation; the latter can be attributed to theunique challenges associated with inner cell massoutgrowth and the earliest stages of stem-cell derivationfrom human embryos, and contrasts with themaintenance of previously established humanembryonic stem-cell lines, which can be grown indefined culture systems with almost 100% success.

Experience with organ and tissue allotransplantationhas shown that diseases such as HIV-1 or HIV-2infection, Creutzfeldt-Jakob disease, hepatitis B or Cviruses, and other infectious agents can be transmittedfrom human donor cells to the recipient. Similarly, co-culture of human embryos with live animal cells causesconcern for infection with recognised or as yetunrecognised infectious agents.9,10 We aimed to deriveand establish new human embryonic stem-cell lines incompletely feeder-layer free and serum-free conditions,which would contribute to solving one of the majorproblems associated with the use of human embryonicstem-cell differentiated progeny in the treatment ofhuman medical conditions.

MethodsHuman embryonic stem cells were cultured aspreviously described.11,12 The culture medium consistedof knockout DMEM, supplemented with 50 mg/Lpenicillin, 50 mg/L streptomycin, 2 mmol/L glutamax-I,0·055 mmol/L �-mercaptoethanol, non-essentialaminoacids (1 to 100 dilution of the stock solution), 8%serum replacement, 8% plasmanate, 20 �g/L humanleukaemia inhibitory factor, and 16 �g/L human basicfibroblast growth factor (plasmanate from Bayer,Research Triangle Park, NC, USA; human leukaemiainhibitory factor from Chemicon, Temecula, CA, USA;all other reagents from Invitrogen, Carlsbad, CA, USA).Cells were passaged either mechanically or with 0·05%trypsin with 0·53 mmol/L EDTA (Invitrogen).

Extracellular matrix was prepared by sodiumdeoxycholate (DOC, Sigma, St Louis, MO, USA)extraction of cell monolayers by a modification of the

Published online March 8, 2005http://image.thelancet.com/extras/04art11036web.pdf

Advanced Cell Technology,Worcester, MA, USA(I Klimanskaya PhD,Y Chung PhD, M D West PhD,R Lanza MD); Department ofPathology and LaboratoryMedicine, University ofWisconsin Medical School,Madison, WI, USA(L Meisner PhD); CytogeneticsLaboratory, Wisconsin StateLaboratory of Hygiene,Madison, WI, USA (L Meisner,J Johnson PhD); and Institute forRegenerative Medicine, WakeForest University School ofMedicine, Winston-Salem, NC,USA (R Lanza)

Correspondence to:Dr Robert Lanza, Advanced CellTechnology, 381 PlantationStreet, Worcester, MA 01605,USArlanza@advancedcell.com

www.thelancet.com Published online March 8, 2005 http://image.thelancet.com/extras/04art11036web.pdf 1

Human embryonic stem cells derived without feeder cellsIrina Klimanskaya, Young Chung, Lorraine Meisner, Julie Johnson, Michael D West, Robert Lanza

SummaryBackground Human embryonic stem cells are likely to play an important role in the future of regenerative medicine.

However, exposure of existing human embryonic stem-cell lines to live animal cells and serum risks contamination

with pathogens that could lead to human health risks. We aimed to derive an embryonic stem-cell line without

exposure to cells or serum.

Methods Frozen cleavage-stage embryos were thawed and cultured to the blastocyst stage. Inner cell masses were

isolated by immunosurgery and plated onto extracellular-matrix-coated plates that can be easily sterilised. Six

established human embryonic stem-cell lines were also maintained with this serum and feeder free culture system.

Findings A new stem-cell line was derived from human embryos under completely cell and serum free conditions.

The cells maintained normal karyotype and markers of pluripotency, including octamer binding protein 4 (Oct-4),

stage-specific embryonic antigen (SSEA)-3, SSEA-4, tumour-rejection antigen (TRA)-1–60, TRA-1–81, and alkaline

phosphatase. After more than 6 months of undifferentiated proliferation, these cells retained the potential to form

derivatives of all three embryonic germ layers both in vitro and in teratomas. These properties were also successfully

maintained (for more than 30 passages) with the established stem-cell lines.

Interpretation This system eliminates exposure of human embryonic stem cells and their progeny to animal and human

feeder layers, and thus the risk of contamination with pathogenic agents capable of transmitting diseases to patients.

For personal use. Only reproduce with permission from Elsevier Ltd

Articles

method described by Hedman and others.13 Briefly,mouse embryonic fibroblasts were prepared from12·5-day embryos from pregnant ICR (Institute ofCancer Research) mice. Embryos were eviscerated,rinsed with phosphate-buffered saline (PBS), mincedwith scissors, and incubated at room temperature in0·05% trypsin, 0·53 mmol/L EDTA (Invitrogen) for5–10 min with constant pipetting. The dissociated cellswere centrifuged in mouse embryonic fibroblast growthmedium (high glucose DMEM supplemented with50 mg/L penicillin, 50 mg/L streptomycin, 2 mmol/Lglutamax-I, and 10% fetal bovine serum) and plated onto150-mm gelatin-coated (Sigma) cell culture dishes(about one and a half to two embryos per plate). After2–3 days, the cells were split at a ratio of one to five andfrozen at passage 1 at confluency. For preparation of theextracellular matrix, mouse embryonic fibroblasts wereused at passage 2 after being mitotically inactivated with1 g/L of mitomycin C (Sigma) for 3 h. The cells wereplated at a density of 50–60 thousand cells/cm2 andcultured for7–21 days. After rinsing with PBS, they wereincubated in 0·5% DOC in 10-mmol/L Tris-HCl, pH8·0, on ice for 30 min, rinsed five or six times with PBS,and stored at 4ºC or dried and sterilised with standardmethods.

A new human embryonic stem-cell line was derivedfrom unused human embryos produced by in-vitrofertilisation for clinical purposes and used incompliance with the guidelines of Advanced CellTechnology’s ethics advisory board. Frozen cleavage-stage embryos were thawed and cultured to theblastocyst stage. Imumunosurgery was done asdescribed by Solter and Knowles14 with the antibody tohuman red blood cells (Inter-cell Technologies,Hopewell, NJ, USA). The inner cell masses were placedonto extracellular-matrix-coated 4-well plates in growthmedium; two-thirds of the medium was changed every2–3 days. When the initial outgrowth formed colonieslarge enough for dispersion (�50 cells), they weredispersed with a finely drawn capillary, or incubated intrypsin for 2–3 min and hand-picked under thedissecting microscope. Passage 1 colonies weremechanically dispersed and expanded throughoutpassages seven to nine and routinely passagedmechanically, with trypsin, or both.

In-vitro differentiation experiments were done eitherwith adherent human embryonic stem cells or withembryoid bodies. For adherent differentiation, stemcells were left to overgrow on extracellular matrix, andafter the morphology of the colonies showed signs of

2 www.thelancet.com Published online March 8, 2005 http://image.thelancet.com/extras/04art11036web.pdf

Figure 1: H9 human embryonic stem-cell line growing on extracellular matrix A: single colonies (stereo microscopy). B: colony surrounded by fibroblast-like cells (phase contrast). C: alkaline phosphatase. D–I: immunofluorescence staining formarkers of pluripotency of human embryonic stem cells growing on extracellular matrix. D: Oct-4. E: corresponding DAPI; F: SSEA-3; G: SSEA-4; H: TRA-1–60;. I: TRA-1–81. Scale bars=100 �m.

For personal use. Only reproduce with permission from Elsevier Ltd

Articles

differentiation, they were switched to embryoid bodiesmedium (usually 8–10 days after passaging), which hadthe same ingredients as growth medium except basicfibroblast growth factor, leukaemia inhibitory factor, andplasmanate; the serum replacement concentration was13%. The medium was changed every 1–2 days. Forembryoid body formation, human embryoid stem cellswere trypsinised and cultured in embryonic bodiesmedium on low adherent costar brand plates (Tokyo,Japan). For teratoma formation, 6–8 week-old NOD-SCID (non-obese diabetic/severe-combined immuno-deficient) mice (Jackson Laboratories, Bar Harbor, ME,USA) were injected intramuscularly with 2–4 millionhuman embryonic stem cells. After tumours becamepalpable (7–10 weeks after injection) they were removedand fixed overnight in 4% paraformaldehyde at 4ºC.Paraffin sections were prepared and processed forhistological examination by Mass Histology Services(Worcester, MA, USA).

For immunostaining and localisation of alkalinephosphatase, cells were fixed with 2% paraformaldehyde,permeabilised with 0·1% NP-40, and blocked with 10%goat serum, 10% donkey serum (JacksonImmunoresearch Laboratories, West Grove, PA, USA) in

PBS (Invitrogen) for 1 h or longer; incubation withprimary antibodies was done overnight at 4°C, thebiotinilated secondary antibodies (Jackson Immuno-research Laboratories) were added for 1 h, followed byfluorescently labelled streptavidin (Amersham,Piscataway, NJ, USA) incubation for 15 min. Between allincubations, specimens were washed with 0·1% Tween-20 (Sigma) in PBS three to five times, for 10–15 min eachwash. Specimens were mounted in Vectashield withDAPI (Vector Laboratories, Burlingame, CA, USA) andobserved under fluorescent microscope (Nikon,Kawasaki, Kanagawa, Japan). Alkaline phosphatase wasdetected with Vector Red kit (Vector Laboratories)according to manufacturer’s instructions. The followingantibodies were used: antibody to ocatmer bindingprotein 4 (Oct-4, monoclonal, Santa Cruz Biotechnology,Santa Cruz, CA, USA), antibody to stage-specificembryonic antigen 3 (SSEA-3), antibody to SSEA-4(Developmental Studies Hybridoma Bank, University ofIowa, IA, USA), antibody to tumour-rejection antigen(TRA)1–60, antibody to TRA-1–81 (Chemicon), tubulin �III (BABCO, Berkeley, CA, USA), antibody to alfa-fetoprotein (DACO), antibody to muscle actin (Abcam,Cambridge, MA, USA).

www.thelancet.com Published online March 8, 2005 http://image.thelancet.com/extras/04art11036web.pdf 3 3

A

D E F

G H I

B C

Figure 2: ACT-14, a human embryonic stem-cell line derived in feeder-free conditions on extracellular matrix A: phase contrast of a single colony on extracellular matrix; arrows show fibrils of matrix. B: colony growing on fibroblast-like cells. C: staining for alkalinephosphatase. D–I: immunofluorescence staining for markers of pluripotency. D: Oct-4 and E: corresponding DAPI image, arrows show differentiated feeder-like cells.F: SSEA-3; G: SSEA-4; H: TRA-1–60; I: TRA-1–81. Scale bars=100 �m.

For personal use. Only reproduce with permission from Elsevier Ltd

Articles

For karyotyping, rapidly growing cells (day 2–3 afterone to three split) were incubated with 0·1 mg/Lcolcemid (Invitrogen) for 30 min, trypsinised,centrifuged, and resuspended in 0·075 mol/L potassiumchloride for 10 min, centrifuged and suspended in oneto three acetic acid/methanol fixative for 30 min,centrifuged, resuspended, and stored in this fixative.Cytogenetic analysis was done on metaphase cells withG banding on 20 cells.

Role of the funding sourceThe sponsor coordinated the study. The sponsor had norole in study design, data collection, datat analysis, datainterpretation, or writing of the report. Thecorresponding author had full access to all the data inthe study and had final responsibility for the decision tosubmit for publication.

ResultsHandpicked colonies of human embryonic stem cellswere plated onto extracellular-matrix-coated plates andsubsequently passaged with either mechanical orenzymatic maintenance techniques. Of note, twodifferent types of cultures were apparent: single coloniesof cells like human embryonic stem cells and coloniesgrowing on fibroblast-like cells that differentiated fromthe human embryonic stem cell colonies (figure 1). Bothsingle colonies and those surrounded by feeder-like cellscould be passaged with either trypsin or mechanicaldispersion. The cells maintained by this culture systemappeared very similar to feeder layer-supportedcounterparts in their performance and maintenancerequirements. On average, enzymatically passagedcolonies became confluent and were passaged every5–7 days, and could be successfully frozen and thawedwith a recovery rate of over 10–20%. When the cellcultures were allowed to overgrow, they spontaneouslydifferentiated into the cells of all three germ layers.

After more than 20 passages on extracellular matrix,human embryonic stem cells retained the markers ofpluripotency, including Oct-4, SSEA-3, SSEA-4, TRA-1–60, TRA-1–81, and alkaline phosphatase (figure 1),whereas feeder-like cells were negative for such markers(figure 2 for the line ACT-14, shown with arrows).Occasionally, instead of forming defined colonies, thecells grew as monolayers that were also positive for thesemarkers (figure 3). The cells maintained a normalfemale karyotype (46, XX) and retained their capacity toform differentiated cell types of all three germ layers, asevidenced by immunostaining with antibodies to muscleactin (mesoderm), tubulin � III (ectoderm), and�-fetoprotein (primitive endoderm). This feeder-freesystem was used to successfully maintain six otherexisting human embryonic stem-cell lines: H1, H7, andH9 (NIH-registered as WA01, WA07, WA09), and fourlines derived at Howard Hughes Medical Institute andHarvard University11 and Advanced Cell Technologywith private funds.

The early stages of inner cell mass outgrowth areextremely vulnerable to cell microenvironment andculture conditions. The culture system based onextracellular matrix reported here proved sufficient forderivation and establishment of a new humanembryonic stem-cell line from unused human embryosin completely cell-free and serum-free conditions. Theinner cell masses from five blastocyst-stage embryoswere isolated by immunosurgery and plated onto plates

4 www.thelancet.com Published online March 8, 2005 http://image.thelancet.com/extras/04art11036web.pdf

Figure 3: Typical morphology of human embryonic stem cells growing as a monolayer of (ACT-14 line) A–D: same field. A: phase contrast. B: alkaline phosphatase. C: Oct-4. D: DAPI. Scale bar=100 �m.

Figure 4: Initial stages of stem-cell derivation in feeder-free conditions A–C: outgrowth of inner cell mass ACT-6, day 8. B: outgrowth of inner cell mass ACT-13, day 7; C: outgrowth ofACT-14 (future stem-cell line) day 9; D: passage 1 of ACT-14. Scale bars=100 �m.

For personal use. Only reproduce with permission from Elsevier Ltd

Articles

coated with extracellular matrix. Initial outgrowth wasobserved in three inner cell masses (figure 4) and onestable cell line (ACT-14) was obtained by mechanical andenzymatic passaging as described above.Morphologically, the cells formed well defined coloniesand resembled embryonic stem cells, with a smallcytoplasm-to-nuclear ratio and multiple nucleoli andcytoplasmic lipid bodies (figure 1). The cells could beextensively propagated (more than 6 months so far), andhad the strong expression of molecular markers ofpluripotential cells, including Oct-4, alkalinephosphatase, SSEA-3, SSEA-4, TRA-I-60, TRA-I-81(figure 2). Cytogenetic analysis showed a stable cell linewith a normal female karyotype (46, XX). The humanembryonic stem cells maintained on extracellular matrixdifferentiated into all three germ layers, as detected byindirect immunofluorescence staining with antibodiesto muscle actin, tubulin � III, and �-fetoprotein(figure 3). When injected into NOD-SCID mice, thesecells formed teratomas with the presence of the tissuesof all three germ layers, including bone and cartilage(mesoderm), neural rosettes (ectoderm), and intestinaland respiratory epithelia (endoderm, figure 5).

The extracellular-matrix-coated-plates in our studyproved remarkably stable. Preliminary studies indicatethat the plates could be dehydrated, treated withparaformaldehyde (2–4%), heat-pasteurised (60°C forover 20 h), or gamma-irradiated (10–25 kGy) withoutsubstantial changes in human embryonic stem-cellperformance. These and an array of other conventionaltreatment and sterilisation processes could be used toeliminate the infectivity of viral structures and otherpossible infectious agents to help ensure humanembryonic stem-cell lines and their progeny are safefrom transmission of pathogens.

DiscussionThe system we describe allows cell-free and serum-freederivation of new embryonic stem-cell lines for futureclinical applications. Three of five embryos (60%)generated inner cell mass outgrowth, and one stablestem-cell line (20%) was obtained. This success rate issimilar to that reported by other investigators who usedconventional feeder-based culture systems.11,15–17

Moreover, most existing human embryonic stem-celllines were derived in the presence of serum in the mediaand later adapted to serum-free conditions, whereas ourline (ACT-14) was derived without serum. These cellshad typical human embryonic stem-cell morphology andbehaviour, could be grown and manipulated easily,including enzymatic passaging with trypsin, and hadhigh recovery after freezing and thawing. Our systemalso allowed robust and reliable maintenance ofestablished human embryonic stem cells in feeder-layer-free and serum-free conditions.

The extrinsic factors necessary for maintaining humanembryonic stem-cell pluripotency and self-renewal arepoorly understood, partly because of complex cultureconditions that include both growth-inactivated feedercells and serum, a cocktail of myriad proteins andsoluble factors. Our culture system affords anopportunity to study these mechanisms by replacingcells and serum with adherent molecules. It is alsopossible that the culture medium that we previouslydescribed11,12 has certain advantages that work in synergywith extracellular matrix to provide the necessary cuesfor human embryonic stem cells to proliferate and retainpluripotency. The extracellular matrix is a uniquelyassembled three-dimensional molecular complex thatvaries in composition and diversity, and consists of basiccomponents such as fibronectin, collagens and other

www.thelancet.com Published online March 8, 2005 http://image.thelancet.com/extras/04art11036web.pdf 5

Figure 5: In vitro differentiation and teratoma formation of ACT-14 A–C: immunofluorescence staining for the markers of three germ layers (in-vitro differentiation). A: muscle actin; B: tubulin � III’ C: �-fetoprotein. D–F: histology ofteratomas formed in NOD-SCID mice; D: bone and cartilage; E: neural rosettes; F: intestinal epithelium. Scale bars=100 �m.

For personal use. Only reproduce with permission from Elsevier Ltd

Articles

glycoproteins, hyaluronic acid, proteoglycans, andelastins; however, it also harbours and presentsmolecules such as growth factors and hormones,18,19

although many of these molecules would be inactivatedduring heat pasteurisation or sterilisation. Furtherstudies will be needed to determine which of theseextracellular matrix molecules or their spatialorganisation are crucial for the maintenance of theundifferentiated human embryonic stem-cell phenotype.

A study20 has shown that human embryonic stem cellscan potentially take up Neu5Gc, an immunogenicnonhuman sialic acid. Our preliminary studies indicatethat extracellular-matrix-coated plates derived fromhuman cells can also be used to maintain such cells.

Human embryonic stem cells have the potential totreat a wide range of serious and often life-threateningconditions, but their therapeutic potential is hindered bythe threat of contamination from serum products andlive feeder cells. Serum-free systems, such as that in thisstudy, use a defined protein cocktail that can begenerated under sterile conditions. Similarly, embryocontact with serum-derived antibodies and complementduring immunosurgery can be eliminated bymicrosurgery or no immunosurgery. Our system alsoeliminates exposure of these cells and their progeny toanimal and human feeder layers, and thus the risk ofcontamination with pathogenic agents capable oftransmitting diseases to the patient.

ContributorsI Klimanskaya and Y Chung participated in study design, experimentswith stem cells, data analysis, and writing of the report. L Meisner andJ Johnson did karyotype analysis and reviewed the report. M D Westparticipated in data analysis and reviewed the report. R Lanzaparticipated in study design, data analysis, and preparation of the report.

Conflict of interest statementI Klimanskaya, Y Chung, M D West, and R Lanza are employees ofAdvanced Cell Technology. L Meisner and J Johnson declare that theyhave no conflict of interest.

AcknowledgmentsAdvanced Cell Technology funded this study. This work was iniated atHarvard University in the laboratory of Douglas Melton with supportfrom HHMI and JDRF.

References1 Richards M, Fong CY, Chan WK, Wong PC, Bongso A. Human

feeders support prolonged undifferentiated growth of human innercell masses and embryonic stem cells. Nat Biotechnol 2002; 20:933–36.

2 Amit M, Margulets V, Segev H, et al. Human feeder layers forhuman embryonic stem cells. J Biol Reprod 2003; 68: 2150–56.

6 www.thelancet.com Published online March 8, 2005 http://image.thelancet.com/extras/04art11036web.pdf

3 Lee JB, Lee JE, Park JH, et al. Establishment and maintenance ofhuman embryonic stem cell lines on human feeder cells derivedfrom uterine endometrium under serum-free condition.Biol Reprod 2004; 72: 42–49.

4 Hovatta O, Mikkola M, Gertow K, et al. A culture system usinghuman foreskin fibroblasts as feeder cells allows production ofhuman embryonic stem cells. Hum Reprod 2003; 18: 1404–09.

5 Xu C, Inokuma MS, Denham J, et al. Feeder-free growth ofundifferentiated human embryonic stem cells. Nat Biotechnol 2001;19: 971–74.

6 Rosler ES, Fisk GJ, Ares X, et al. Long-term culture of humanembryonic stem cells in feeder-free conditions. Dev Dyn 2004; 229:259–74.

7 Amit M, Shariki C, Margulets V, Itskovitz-Eldor J. Feeder layer- andserum-free culture of human embryonic stem cells. Biol Reprod2004; 70: 837–45.

8 Xu R, Peck RM, Li DS, Feng X, Ludwig T, Thomson JA. Basic FGFand suppression of BMP signaling sustain undifferentiatedproliferation of human ES cells. Nat Meth 2005; 2: 185–90.

9 US Department of Health and Human Services Food and DrugAdministration (FDA) Center for Biologics Evaluation andResearch (CBER). Information and recommendations forphysicians involved in the co-culture of human embryos withnonhuman animal cells. Rockville MD: FDA, 2002.

10 Cooper, DKC, Lanza, RP. Xeno: the promise of transplantinganimal organs into humans. New York: Oxford University Press,2000.

11 Cowan CA, Klimanskaya IV, McMahon J, et al. Derivation ofembryonic stem cell lines from human blastocysts. N Eng J Med2004; 350: 1353–56.

12 Klimanskaya I, McMahon J. Approaches for derivation andmaintenance of human ES cells: detailed procedures andalternatives. In: Lanza R, Gearhart J, Hogan B, et al, eds. Handbookof Stem Cells. Vol 1: Embryonic Stem Cells. San Diego:Elsevier/Academic Press, 2004: 437–49.

13 Hedman K, Kurkinen M, Alitalo K, Vaheri A, Johansson S,Hook M. Isolation of the pericellular matrix of human fibroblastcultures. J Cell Biol 1979; 81: 83–91.

14 Solter D, Knowles BB. Immunosurgery of mouse blastocyst.Proc Natl Acad Sci USA 1975; 72: 5099–102.

15 Thomson JA, Itskovitz-Eldor J, Shapiro SS, et al. Embryonic stemcell lines derived from human blastocysts. Science 1998; 282:1145–47.

16 Lanzendorf SE, Boyd CA, Wright DL, Muasher S, Oehninger S,Hodgen GD. Use of human gametes obtained from anonymousdonors for the production of human embryonic stem cell lines.Fertil Steril 2001; 76: 132–37.

17 Mitalipova M, Calhoun J, Shin S, et al. Human embryonic stem celllines derived from discarded embryos. Stem Cells 2003; 21: 521–26.

18 Gordon MY, Riley GP, Watt SM, Greaves MF.Compartmentalization of a haematopoietic growth factor(GM-CSF) by glycosaminoglycans in the bone marrowmicroenvironment. Nature 1987; 326: 403–05.

19 Roberts R, Gallagher J, Spooncer E, Allen TD, Bloomfield F,Dexter TM. Heparan sulphate bound growth factors: a mechanismfor stromal cell mediated haemopoiesis. Nature 1988; 332: 376–78.

20 Martin JM, Muotri A, Gage F, Varki A. Human embryonic stemcells express an immunogenic nonhuman sialic acid. Nat Med2005; 11: 228–32.