Enhanced Lung Epithelial Specification of HumanInduced Pluripotent Stem Cells on DecellularizedLung MatrixSarah E. Gilpin, PhD, Xi Ren, PhD, Tatsuya Okamoto, MD, PhD,Jacques P. Guyette, PhD, Hongmei Mou, PhD, Jayaraj Rajagopal, MD,Douglas J. Mathisen, MD, Joseph P. Vacanti, MD, and Harald C. Ott, MDDepartment of Surgery, Massachusetts General Hospital, and Harvard Medical School, Boston, Massachusetts

Background. Whole-lung scaffolds can be created byperfusion decellularization of cadaveric donor lungs. Theresulting matrices can then be recellularized to regeneratefunctional organs. This study evaluated the capacity ofacellular lung scaffolds to support recellularization withlung progenitors derived from human induced pluripo-tent stem cells (iPSCs).

Methods. Whole rat and human lungs were decellular-ized by constant-pressure perfusion with 0.1% sodiumdodecyl sulfate solution. Resulting lung scaffolds werecryosectioned into slices or left intact. Human iPSCs weredifferentiated to definitive endoderm, anteriorized to aforegut fate, and then ventralized to a population ex-pressing NK2 homeobox 1 (Nkx2.1). Cells were seededonto slices and whole lungs, which were maintained un-der constant perfusion biomimetic culture. Lineage speci-fication was assessed by quantitative polymerase chainreaction and immunofluorescent staining. Regenerated leftlungs were transplanted in an orthotopic position.

Results. Activin-A treatment, followed by transform-ing growth factor-b inhibition, induced differentiation

Accepted for publication May 13, 2014.

Presented at the Fiftieth Annual Meeting of The Society of ThoracicSurgeons, Orlando, FL, Jan 25–29, 2014.

Address correspondence to Dr Ott, Massachusetts General Hospital,Department of Surgery, Harvard Medical School, 185 Cambridge St,CPZN 4812, Boston, MA 02114; e-mail: hott@partners.org.

� 2014 by The Society of Thoracic SurgeonsPublished by Elsevier

of human iPSCs to anterior foregut endoderm asconfirmed by forkhead box protein A2 (FOXA2), SRY (SexDetermining Region Y)-Box 17 (SOX17), and SOX2expression. Cells cultured on decellularized lung slicesdemonstrated proliferation and lineage commitment af-ter 5 days. Cells expressing Nkx2.1 were identified at40% to 60% efficiency. Within whole-lung scaffoldsand under perfusion culture, cells further upregulatedNkx2.1 expression. After orthotopic transplantation,grafts were perfused and ventilated by host vasculatureand airways.Conclusions. Decellularized lung matrix supports the

culture and lineage commitment of human iPSC-derivedlung progenitor cells. Whole-organ scaffolds and bio-mimetic culture enable coseeding of iPSC-derived endo-thelial and epithelial progenitors and enhance early lungfate. Orthotopic transplantation may enable furtherin vivo graft maturation.

(Ann Thorac Surg 2014;-:-–-)� 2014 by The Society of Thoracic Surgeons

t present, transplantation is the only therapeutic op-

Ation for many patients with end-stage lung disease;yet, limitations in organ availability and complications ofimmunosuppression and chronic graft dysfunction remain[1]. The successful translation of organ-engineering stra-tegies to regenerate lungs for transplantation would pro-vide a novel organ source and increase the number oflungs available for patients. The ability to create patient-specific organs would be of further benefit, reducing theneed for immunosuppression and potentially improvinggraft survival.

Acellular lung scaffolds can be created from cadavericorgans by multiple approaches, all aiming to removecellular material while retaining the extracellular matrix

scaffold upon which cells can later be reintroduced [2]. Ourdecellularization methodology uses constant-pressuredetergent perfusion through the native lung vasculatureand results in a biocompatible whole-organ scaffold[3]. This scaffold can support primary epithelial andendothelial recellularization to regenerate transplantableconstructs that could be perfused and ventilated in vivofor up to 1 week [4, 5].There is great interest in the use of cell populations

derived from induced pluripotent stem cells (iPSCs) astools for patient-specific lung regeneration [6]. Derivationof these cells requires recapitulation of the sequentialsteps in early lung specification. During development,germ layer restriction first occurs during gastrulation,with cells exposed to strong nodal signalling becoming

The Appendix can be viewed in the online version ofthis article [http://dx.doi.org/10.1016/j.athoracsur.2014.05.080] on http://www.annalsthoracicsurgery.org.

0003-4975/$36.00http://dx.doi.org/10.1016/j.athoracsur.2014.05.080

2 GILPIN ET AL Ann Thorac SurgLUNG RECELLULARIZATION WITH HUMAN IPSCs 2014;-:-–-

specified to definitive endoderm [7]. In culture, nodalsignalling is mimicked with high dose activin-A togenerate an enriched definitive endoderm populationmarked by the transcription factors SRY (Sex Deter-mining Region Y)-Box 17 (Sox17) and forkhead box pro-tein A2 (Foxa2) [8]. The endoderm layer then formsthe primitive gut tube from which the respiratory anddigestive systems develop through an anterior-posteriordivision. Lungs and trachea both arise from the anteriorforegut endoderm, marked by the transcription factorSox2 and preservation of Foxa2 [9, 10]. Timed inhibitionof the transforming growth factor-b signalling pathwayis sufficient to anteriorize iPSC-derived endoderm invitro [11].

The transcription factor NK2 homeobox 1 (Nkx2.1)is the earliest marker of lung-specified endoderm [12],as well as early thyroid and brain. Mouse embryoscarrying a homozygous disruption in the Nkx2.1 locusform a primitive lung structure but fail to undergobranching morphogenesis or epithelial developmentand cannot support ventilation [13]. Specific growthfactors, including Wnt, bone morphogenetic protein(BMP), and fibroblast growth factor (FGF), are critical toventralization and the generation of Nkx2.1-expressingcells [9]. From this ventral wall, the primary lung budsemerge and separate from the dorsal esophagus, drivenby signalling from the surrounding mesenchyme [14].

Directed, step-wise in vitro differentiation protocolshave successfully generated airway epithelium and lungpneumocytes [11, 15–17] from human iPSCs by recapit-ulation of these developmental stages.

The role of matrix–cell interactions in mediatingcellular fate decisions is an important but incompletelyunderstood element of lung development. Basementmembrane attachment can direct type II pneumocytes toadopt a differentiated type I phenotype [18], and matrixinteractions can enhance surfactant protein-C expres-sion by differentiating embryonic stem cells [19]. The3-dimensional structure of lung matrix [20], alongwith mechanical stimuli, such as stretch, also providedevelopmental cues [21, 22]. Moreover, the cross-talkbetween endogenous cells and the extracellular matrixduring tissue repair, involving growth factor productionand matrix modification, further highlights this complexrelationship [23].

By combining principles of lung development andbioengineering, the aim of patient-specific therapies forend-stage lung disease may be realized through theregeneration of whole-organ scaffolds [24]. The use ofpatient-derived stem cells and the elucidation of the keydevelopmental stages for recellularization of native scaf-folds are important milestones toward functional organregeneration and transplantation.

Material and Methods

Human iPSC CultureAll experiments used iPSCs generated by messengerRNA transfection (Kruppel-like factor 4 [Klf4], c-Myc,

octamer-binding transcription factor 4 [Oct4], Sox2,LIN28) of postnatal fibroblasts [25]. Undifferentiated cellswere maintained in feeder-free culture in mTeSR media(STEMCELL Technologies Inc, Vancouver, BC, Canada)on Geltrex-coated plates (Life Technologies Corp, Carls-bad, CA), and passaged with Accutase (STEMCELLTechnologies Inc) at 70% confluency. Differentiation wasinitiated at 80% confluency.

Human iPSC-Derived EndotheliumEndothelial differentiation used a basal media of Iscove’smodified Dulbecco’s medium, BIT 9500 Serum Substitute(STEMCELL Technologies) nonessential amino acids(Gibco, Carlsbad, CA), L-glutamine (Gibco), mono-thioglycerol (10%, Sigma-Aldrich, St. Louis, MO), andpenicillin/streptomycin. Complete differentiation mediawas supplemented with BMP-4 (50 ng/mL, PeproTech,Rocky Hill, NJ), FGF-basic (50 ng/mL, PeproTech),and vascular endothelial growth factor 165 (50 ng/mL,PeproTech) for 6 days. Differentiated endothelial cellswere expanded in endothelial cell growth medium-2(Lonza, Basel, Switzerland) supplemented with 20%fetal bovine serum (HyClone/GE Healthcare Life Sci-ences Laboratories, South Logan, UT) on 0.1% gelatin-coated plates (Appendix Fig 1).

Human iPSC-Derived Epithelial ProgenitorsA basal media of Roswell Park Memorial Institute 1640with Glutamax (Life Technologies) plus 2% B27 serumsupplement minus insulin (Invitrogen, Carlsbad, CA) wassupplemented to mimic three developmental stages: (1)induction of definitive endoderm with 50 ng/mL activin-A(PeproTech) for 4 days, (2) anteriorization of endodermwith 1mM A8301 (Tocris Bioscience, Bristol, UnitedKingdom) for 4 days, and (3) ventralization with 10 ng/mLBMP-4 (PeproTech), 100 ng/mL FGF-2 (PeproTech), and100nM CHIR 99021 (Stemgent, Cambridge, MA) for 4 to 7days.

Histology and ImmunofluorescenceParaformaldehyde-fixed, paraffin-embedded tissue sec-tions were stained by hematoxylin and eosin or under-went antigen retrieval and permeabilization with 0.1%Triton-X100. Cultured cells were fixed with 100%methanol.Immunofluorescent staining included (1) blocking by

5% donkey serum, (2) primary antibodies (1:200) over-night at 4�C, (3) washing with 0.1% Tween-20, (4) sec-ondary antibodies (1:400; AlexFluor 488 or 594, LifeTechnologies) for 1 hour at room temperature, and (5)counterstaining with 40,6-diamidino-2-phenylindole for 5minutes. Images were captured with a Ti-PFS invertedmicroscope (Nikon, Melville, NY).

Quantitative Real-Time Polymerase Chain ReactionSamples were stored in RNALater (Qiagen, Valencia, CA)until RNA isolation by RNeasy-Plus Mini Kit (Qiagen).RNA was transcribed to complementary DNA by Super-Script III Reverse Transcriptase (Life Technologies). Geneexpression was quantified by Taqman assays using the

3Ann Thorac Surg GILPIN ET AL2014;-:-–- LUNG RECELLULARIZATION WITH HUMAN IPSCs

OneStepPlus system (Applied Biosystems, Foster City,CA). Gene expression was analyzed by the DDCt methodwith normalization to 18S gene expression.

Lung Tissue Slice ExperimentsDecellularized human lung tissue was fixed in 30% su-crose, embedded in optimum cutting temperature com-pound, cut at 100 mm, and immobilized on tissue cultureplates. Slices were washed with 100% and 95% ethanoland then air-dried for 30 minutes. Sections were incu-bated overnight in phosphate-buffered saline to removeresidual detergents and then preconditioned with mediafor 3 hours before seeding.

On day 2 of ventralization, cells were reseeded directlyto lung slices and cultured for 5 days. For cell localization,images at original magnification �10 were captured on6 separate cell-matrix cultures. ImageJ software (NationalInstitutes of Health, Bethesda, MD) was used to quantifycells positive for 40,6-diamidino-2-phenylindole localizedon or off the matrix. Counts were normalized to totalmatrix area.

Decellularization, Recellularization, and Whole-OrganCultureAnimal experiments were approved by the Massachu-setts General Hospital Institutional Animal Care and UseCommittee and performed in compliance with the Ani-mal Welfare Act. Cadaveric rat lungs were perfusiondecellularized as previously described [4]. In brief,cadaveric lungs were explanted from Sprague-Dawleyrats (260 to 280 g, Charles River Laboratories Interna-tional Inc, Wilmington, MA). The pulmonary artery (PA)was cannulated with an 18-gauge dispensing needle(McMaster-Carr), and perfused under constant pressure(50 cm H2O) sequentially with heparinized (10 U/mL)saline, 0.1% sodium dodecyl sulfate (2 hours), deionizedwater (15 minutes), and 1% Triton-X-100 (10 minutes).Scaffolds were washed with phosphate-buffered salinecontaining antibiotics for 72 hours (Sigma-Aldrich) toeliminate residual detergent. Removal of cells and resid-ual DNA was routinely confirmed by hematoxylin andeosin staining and by quantitative assay (Quant-iT kit,Invitrogen) [3].

Decellularized lungs were intubated with 16-gaugeintratracheal cannulae and placed in a custom biore-actor facilitating continuous perfusion through the PA(2 mL/min). Human iPSC-derived endothelial cells weredelivered to the PA (15 to 20 � 106) by syringe in a 10-mLvolume at 3 to 5 mL/min. At day 10, a confluent 10-cmplate of ventralized iPSC-derived epithelial progenitorcells was collected and delivered to the airways by gravityin a 30-mL volume.

Orthotopic Lung TransplantationRecipient Sprague-Dawley rats (260 to 300 g, CharlesRiver) were anesthetized with 5% isoflurane (Abbott,Abbott Park, IL), intubated with a 16-gauge endotrachealtube (Becton-Dickinson, San Jose, CA), and ventilatedwith a rodent ventilator (Harvard Apparatus, HollistonMA) supplying 100% supplemental oxygen. Regenerated

lungs (n ¼ 5) were removed from the bioreactor, and theleft lung was prepared for transplantation using a modi-fied nonsuture external cuff technique [26]. In brief, thedonor hilar vessels were isolated and transected tomount cuffs prepared from Venisystems Abbocaths(Hospira, Lake Forrest, IL). Arterial and venous vesselswere secured with 8-0 silk to the 18-gauge cuff, and thebronchus was secured to a 16-gauge cuff.A left anterior thoracotomy was performed on the

nonheparinized recipient rats. The hilar vessels weredissected circumferentially and individually clampedwith micro-Serrefines clamps. Donor vessels and cuffswere inserted into the recipient vessels and secured with6-0 silk.Blood gases were analyzed at 20 and 60 minutes after

left lung reperfusion. Samples were drawn from the leftpulmonary vein at the anastomosis and analyzed using aniStat Portable Analyzer with CG4þ Cartridges (Abbott).Lungs were then explanted and fixed in 10% formalin orstored in RNAlater.

Results

Human fibroblast–derived iPSCs were differentiatedin vitro toward a lung epithelial progenitor phenotypethrough defined developmental stages (Fig 1A). First,activin A treatment generated definitive endoderm atday 4, as confirmed by loss of OCT4 and inductionSOX17 and FOXA2 (Fig 1B, C). Next, inhibition of TGF-bsignalling for 4 days facilitated anteriorization to aforegut fate, indicated by SOX2 upregulation (Fig 1D, E).Stimulation with FGF-2 and BMP-4, in combinationwith glycogen synthase kinase-3 pathway inhibition,successfully ventralized the anteriorized population to-ward a lung epithelial progenitor phenotype. Approxi-mately 40% to 60% of cells expressed Nkx2.1, the markerof lung specification, after 12 days of differentiation(Fig 1F); corresponding to an eightfold to 21-fold increasein messenger RNA expression over that of undifferenti-ated iPSCs (Fig 1G).Next, ventralized iPSCs (day 10) were seeded onto

human decellularized lung slices and maintained in ven-tralizing media for 5 days. More seeded cells were foundlocalized to the matrix slice than on tissue culture plastic(Fig 2A). Cell viability (>90%) was observed (Fig 2B),and maintenance of a proliferative capacity confirmedby Ki67 expression (Fig 2C). Expression of E-cadherindemonstrated cell–cell interaction and predominanceof an epithelial phenotype (Fig 2D). Lung progenitorphenotype was confirmed by Nkx2.1 expression in asubset of cells (Fig 2E). Loss of OCT4 and maintenance ofFOXA2 expression was confirmed for the ventralized cellpopulation on day 5 of cell-matrix culture. In addition,expression of mature alveolar epithelial cell markersT1a and mucin-1 was detected along with the ciliatedcell marker FOXJ1 (Fig 2F). Compared with controls,an increase in Nkx2.1 expression was measured in cell–matrix cocultures (Fig 2G).Whole-lung constructs were prepared for trans-

plantation by first recellularizing the vasculature with

Fig 1. Differentiation of human induced pluripotent stem cells (iPSCs) to lung epithelial progenitors. (A) In vitro differentiation protocol (BMP¼ bonemorphogenic protein; Nkx2.1 ¼ NK2 homeobox 1; RiPSC ¼ RNA induced iPSCs; RPMI ¼ Roswell Park Memorial Institute; TGF-b ¼ transforminggrowth factor b.) (B) Generation of definitive endoderm indicated by loss of octamer-binding transcription factor 4 (OCT4; upper right), SRY (SexDeterminingRegionY)-Box17 (SOX17; lower right), and forkheadbox proteinA2 (FOXA2;upper left), anddual-immunofluorescence indicating single-cell coexpression (yellow, upper left). (DAPI¼ 40,6-diamidino-2-phenylindole.) (C)Quantitative polymerase chain reaction (qPCR) analysis on day 4 ofdifferentiation relative to undifferentiated day 0 cells. (D)Anteriorized endodermpopulation indicated bySOX2 expression on day 8 of differentiation byimmunofluorescent staining (red) and (E) qPCR analysis relative to day 0 (light gray) and day 4 (dark gray) (DE¼ definitive endoderm.) (F)Differentiation to a lung epithelial progenitor population on day 12 indicated by nuclear NK2 homeobox 1 (Nkx2.1) expression (red), and (G) qPCRexpressed as fold-increase from day 0 undifferentiated population (n ¼ 3 separate experiments). Gene expression all normalized to 18S expression byDDCt (n ¼ 3 samples/differentiation in duplicate). The error bars represent the standard deviation. Scale bars ¼ 100 mm.

4 GILPIN ET AL Ann Thorac SurgLUNG RECELLULARIZATION WITH HUMAN IPSCs 2014;-:-–-

Fig 2. In vitro differentiation of inducedpluripotent stem cell (iPSCs)-derived lungprogenitor cells on human decellularizedlung slices. (A) Cell localization relative tomatrix slice and normalized to total matrixarea per image (n ¼ 6). ***P ¼ 0.0006 byStudent t test. Error bars represent the stan-dard deviation. (B) Viability of day 10differentiated iPSCs seeded onto lung slicesand cultured for 5 days. CalceinAM dye(green), matrix autofluorescence (magenta),40,6-diamidino-2-phenylindole (DAPI; blue).Scale bar ¼ 100 mm. (C) Proliferation of day10þ5 cell–matrix cultures by Ki67 (red). (D)Lung epithelial marker E-cadherin (E-cad;red). (E) Lung progenitor marker NK2 ho-meobox 1 (Nkx2.1; red). Matrix slice identi-fied by autofluorescence (green) and nucleusby DAPI (blue). Scale bar ¼ 50 mm. (F andG) Gene expression of day 10þ5 cells onmatrix slices. (F) Fold increase normalized to18S expression and relative to undifferenti-ated cells (DDCt). (G) Fold increase inNkx2.1 normalized to 18S expression andrelative to parallel no-slice cultures (DDCt).Four independent differentiated cultures(labeled a, b, c, d) were analyzed in duplicate.Error bars represent the standard deviation.

5Ann Thorac Surg GILPIN ET AL2014;-:-–- LUNG RECELLULARIZATION WITH HUMAN IPSCs

human iPSC-derived endothelium (Appendix Fig 1).Cells were delivered by the PA and cultured for 3 daysunder constant perfusion with endothelial media.Next, ventralized (day 10) iPSC were delivered to theairways. Biomimetic culture with ventralizing mediacontaining vascular endothelial growth factor wasmaintained for 2 or 5 days (Fig 3A, B). After culture,heterogeneous cell distribution was observed withinthe scaffold, with some areas remaining relativelyacellular and others displaying notable cell attachmentand retention (Fig 3C). Seeding efficiency wasapproximately 40% to 50% scaffold coverage, basedon histologic assessment. Cell proliferation within

whole-lung cultures was confirmed. Conservation ofthe epithelial progenitor phenotype was demonstratedby nuclear Nkx2.1 expression and the endothelialphenotype confirmed by cluster of differentiation (CD)31 expression (Fig 3D).Further specification toward mature lung phenotypes

was noted by expression of the alveolar type I pneumo-cyte marker T1a/podoplanin and Clara cell secretoryprotein/CC10, thereby indicating an upper airwayphenotype (Fig 3E). Quantification of Nkx2.1 revealedincreased expression in cells cultured on whole-lungscaffolds (Fig 3F). For recellularized lungs cultured for2 days in the bioreactor, an average increase in Nkx2.1

Fig 3. Whole-lung scaffold recellularizationand biomimetic culture. (A) Experimentaldesign. (iPS ¼ induced pluripotent stem;RiPSCs ¼ RNA induced pluripotent stemcells; TGF-b ¼ transforming growth factorb.) (B) Reseeded rat lung scaffold in thebioreactor. (C) Cell distribution and retentionin recellularized, cultured lung constructs, asindicated by hematoxylin and eosin staining(scale bar ¼ 25 mm). (D) Cell proliferation byKi67 staining (red, left panel; scale bar ¼ 48mm), NK2 homeobox 1 (Nkx2.1) expression(red, middle panel; scale bar ¼ 24 mm), andendothelial phenotype by cluster of differen-tiation (CD) 31 (red, right panel; scale bar ¼24 mm) within cultured lung constructs.Matrix outlined by autofluorescence (green)and nucleus by 40,6-diamidino-2-phenylindole (DAPI; blue). (E) Expression ofT1a (red, left panel) and Clara cell secretoryprotein (CCSP; red, right panel) in lungconstructs on day 10þ5; scale bar ¼ 32.5 mm.(F) Polymerase chain reaction quantificationof Nkx2.1 expression in lung constructsnormalized to 18S and relative to parallelin vitro cultured cells (DDCt). Four tissuepieces/lung, analyzed in duplicate; n ¼ 3regenerated lungs. Error bars are the stan-dard error.

6 GILPIN ET AL Ann Thorac SurgLUNG RECELLULARIZATION WITH HUMAN IPSCs 2014;-:-–-

expression of 5.8 � 2.8-fold was measured compared withcells maintained in vitro. For lungs cultured for 5 days,the average expression increased 14.2 � 9.2-fold (n ¼ 3lungs/time point).

After bioreactor culture, single orthotopic left lungtransplant was performed (n ¼ 5, Fig 4A, B). Afterreperfusion, blood gases were sampled from the leftpulmonary vein just proximal to the anastomosis,representing a mixed venous sample. Samplesdemonstrated adequate ventilation in the presence ofthe transplanted graft as indicated by a partial pressureof carbon dioxide maintained within normal limits overthe study period; observed partial pressure of oxygenvalues were compatible with animal survival (Fig 4B).Histologic analysis of the explanted lungs showed red

blood cell perfusion throughout the alveolar capillarynetwork, with occasional alveolar hemorrhage (Fig 4C).

Comment

There is continued need for alternate and novel ther-apies for patients with end-stage lung disease.Although advances in organ preservation, recipientmanagement, and immunosuppression represent sig-nificant therapeutic advancements, lung transplantationremains the only effective treatment option [1, 27].Progress in donor management and organ recon-ditioning have increased the usable fraction of thedonor pool to nearly 50% [28], yet half of all donorlungs are not transplanted, often for logistical reasons.

Fig 4. Orthotopic left lung transplantation of recellularized constructs. (A) Experimental flow from cadaveric lung explant to transplantation ofrecellularized construct. (B) Implantation and reperfusion of recellularized lung construct. The right panel identifies the custom cuffs and sites ofanastomosis (PA ¼ pulmonary artery; PV ¼ pulmonary vein.) (C) Blood gases of recipient animal at 20 and 60 minutes after reperfusion withpositive-pressure ventilation with 100% oxygen. Results from two separate transplants (Tx) are presented. (PCO2 ¼ partial pressure of carbondioxode; PO2 ¼ partial pressure of oxygen.) (D) Histologic analysis of the transplanted, recellularized constructs demonstrates blood perfusion of thealveolar capillaries; scale bars ¼ 25 mm.

7Ann Thorac Surg GILPIN ET AL2014;-:-–- LUNG RECELLULARIZATION WITH HUMAN IPSCs

Decellularization and regeneration would impart newvalue to these organs from otherwise healthy donorswith a structurally undamaged matrix.

Stem cell–based therapies for lung repair is an areaofmuch research [29] involving endogenous and exogenousmechanisms of cellular action [30]. However, effectiveengraftment or direct enhancement of stem cellpopulations or their derivatives has not yet beensuccessful in patients. As directed stem cell differentiationtoward mature lung phenotype advances [31], the abilityto therapeutically engraft stem cell–derived epithelial cellpopulations into diseased lungs remains a challenge.We propose that acellular scaffolds can optimallyaccommodate delivered cells and support the requisition

of function. This goal presents many challenges, includingthe generation of sufficient cell numbers for regeneratingclinical-scale organs and spatially delivering them to theregionally distinct proximal and distal lung. For these rea-sons,we hypothesize that recellularizationwithmultipotentprogenitor cells at the stage of lung specification wouldfacilitate ex vivo expansion and localized differentiation af-ter scaffold seeding. The Nkx2.1-expressing population canthen be directed toward mature epithelial cell phenotypesby exogenous growth factors and intrinsic matrix-derivedcues. We demonstrate that induction of the Nkx2.1-expressing progenitor population is enhanced in cell–ma-trix culture compared with differentiation in traditionalculture.

8 GILPIN ET AL Ann Thorac SurgLUNG RECELLULARIZATION WITH HUMAN IPSCs 2014;-:-–-

Advancement toward a transplantable, fully regener-ated organ would likely require seeding of purified pro-genitor cells or improving in situ control of cell fate toexclude off-target lineages. Gene modification of thesource iPSCs may be required for lineage selection, whichhas been reported for embryonic stem cell lines [32]. Yet,these modifications also present the risk of introducingadditional genetic mutations, and alternate techniquesmay be required.

In the advancing field of organ engineering, a require-ment for small-scale models remains. The high-throughputmatrix slice model and rodent-scale whole-organ scaffoldsallow for testingof specific cell populations, variable seedingtechniques, and biomimetic culture conditions in a repro-ducible fashion. Matrix reseeding efficacy at various stagesof differentiation can also be tested to assess expansion re-quirements. An additional benefit of whole-organ seedingis the recapitulation of endothelial and epithelial cross-talkacross their respective physiologic niches—a critical phe-nomenon in lung development [33, 34] and repair [35]. Theefficient differentiation of human iPSCs toward lungcell populations is also beneficial for patient-specific dis-easemodelling [36], facilitating the study of cell biology andthe testing of therapeutics in many diseases that currentlylack strong animal models.

Successful transplantation and reperfusion of the dualepithelial and endothelial recellularized construct is animportant proof of principle. Although these experi-mental results do not directly reflect normal lung graftphysiology, they provide foundational evidence to sup-port the aim of clinical application. After ex vivo cell de-livery and culture, an “in vivo culture” period may berequired for further regeneration. Complete recapitula-tion of the cellular components of the airways and alveoliwill likely be aided by the recipient’s endogenous envi-ronment and repair program [37]. At minimum, theimplanted construct must permit adequate perfusion,with or without ventilation, but could otherwise be“cultured” in vivo for an extended period before thereacquisition of full function. These concepts outline thecritical next steps in the development of a regenerated,transplantable lung for therapeutic use.

This study was supported by the United Therapeutics Corpora-tion, the National Institutes of Health Director’s New InnovatorAward (DP2-OD008749–01), and the National Institutes of HealthResearch Project Grant (5R01-HL-108678-03).

References

1. Yeung JC, Keshavjee S. Overview of clinical lung trans-plantation. Cold Spring Harb Perspect Med 2014;4:a015628.

2. Crapo PM, Gilbert TW, Badylak SF. An overview of tissueand whole organ decellularization processes. Biomaterials2011;32:3233–43.

3. Gilpin SE, Guyette JP, Gonzalez G, et al. Perfusion decellu-larization of human and porcine lungs: bringing the matrix toclinical scale. J Heart Lung Transplant 2014;33:298–308.

4. Ott HC, Clippinger B, Conrad C, et al. Regeneration andorthotopic transplantation of a bioartificial lung. Nat Med2010;16:927–33.

5. Song JJ, Kim SS, Liu Z, et al. Enhanced in vivo function ofbioartificial lungs in rats. Ann Thorac Surg 2011;92:998–1005.

6. Wagner DE, Bonvillain RW, Jensen T, et al. Can stem cells beused to generate new lungs? Ex vivo lung bioengineeringwith decellularized whole lung scaffolds. Respirology2013;18:895–911.

7. Zorn AM, Wells JM. Vertebrate endoderm development andorgan formation. Annu Rev Cell Dev Biol 2009;25:221–51.

8. Nostro MC, Keller G. Generation of beta cells from humanpluripotent stem cells: Potential for regenerative medicine.Semin Cell Dev Biol 2012;23:701–10.

9. Green MD, Chen A, Nostro MC, et al. Generation of anteriorforegut endoderm from human embryonic and inducedpluripotent stem cells. Nat Biotechnol 2011;29:267–72.

10. Que J, Okubo T,Goldenring JR, et al. Multiple dose-dependentroles for Sox2 in the patterning and differentiation of anteriorforegut endoderm. Development 2007;134:2521–31.

11. Mou H, Zhao R, Sherwood R, et al. Generation of multipotentlung and airway progenitors from mouse ESCs and patient-specific cystic fibrosis iPSCs. Cell Stem Cell 2012;10:385–97.

12. Lazzaro D, Price M, de Felice M, Di Lauro R. The tran-scription factor TTF-1 is expressed at the onset of thyroid andlung morphogenesis and in restricted regions of the foetalbrain. Development 1991;113:1093–104.

13. Minoo P, Su G, Drum H, Bringas P, Kimura S. Defects intracheoesophageal and lung morphogenesis in Nkx2.1(-/-)mouse embryos. Dev Biol 1999;209:60–71.

14. Morrisey EE, Hogan BL. Preparing for the first breath:genetic and cellular mechanisms in lung development. DevCell 2010;18:8–23.

15. Longmire TA, Ikonomou L, Hawkins F, et al. Efficientderivation of purified lung and thyroid progenitors fromembryonic stem cells. Cell Stem Cell 2012;10:398–411.

16. Wong AP, Bear CE, Chin S, et al. Directed differentiation ofhuman pluripotent stem cells into mature airway epitheliaexpressing functional CFTR protein. Nat Biotechnol 2012;30:876–82.

17. Ghaedi M, Calle EA, Mendez JJ, et al. Human iPS cell-derived alveolar epithelium repopulates lung extracellularmatrix. J Clin Invest 2013;123:4950–62.

18. Lwebuga-Mukasa JS, Ingbar DH, Madri JA. Repopulation ofa human alveolar matrix by adult rat type II pneumocytesin vitro. A novel system for type II pneumocyte culture. ExpCell Res 1986;162:423–35.

19. Lin YM, Zhang A, Rippon HJ, Bismarck A, Bishop AE. Tissueengineering of lung: the effect of extracellular matrix on thedifferentiation of embryonic stem cells to pneumocytes.Tissue Eng Part A 2010;16:1515–26.

20. Faulk DM, Johnson SA, Zhang L, Badylak SF. Role of theextracellular matrix in whole organ engineering. J CellPhysiol 2014;229:984–9.

21. Joe P, Wallen LD, Chapin CJ, et al. Effects of mechanicalfactors on growth and maturation of the lung in fetal sheep.Am J Physiol 1997;272:L95–105.

22. Sanchez-Esteban J, Cicchiello LA, et al. Mechanical stretchpromotes alveolar epithelial type II cell differentiation. J ApplPhysiol (1985) 2001;91:589–95.

23. Sacco O, Silvestri M, Sabatini F, Sale R, Defilippi AC,Rossi GA. Epithelial cells and fibroblasts: structural repair andremodelling in the airways. Paediatr Respir Rev 2004;5:S35–40.

24. Badylak SF, Taylor D, Uygun K. Whole-organ tissueengineering: decellularization and recellularization ofthree-dimensional matrix scaffolds. Annu Rev Biomed Eng2011;13:27–53.

25. Warren L, Manos PD, Ahfeldt T, et al. Highly efficientreprogramming to pluripotency and directed differentiationof human cells with synthetic modified mRNA. Cell StemCell 2010;7:618–30.

26. Mizobuchi T, Sekine Y, Yasufuku K, Fujisawa T, Wilkes DS.Comparison of surgical procedures for vascular and airwayanastomoses that utilize a modified non-suture external cufftechnique for experimental lung transplantation in rats.J Heart Lung Transplant 2004;23:889–93.

9Ann Thorac Surg GILPIN ET AL2014;-:-–- LUNG RECELLULARIZATION WITH HUMAN IPSCs

27. Christie JD, Edwards LB, Kucheryavaya AY, et al. TheRegistry of the International Society for Heart and LungTransplantation: 29th adult lung and heart-lung transplantreport-2012. J Heart Lung Transplant 2012;31:1073–86.

28. Snell GI, Paraskeva M, Westall GP. Donor selection andmanagement. Semin Respir Crit Care Med 2013;34:361–70.

29. Stripp BR, Shapiro SD. Stem cells in lung disease, repair, andthe potential for therapeutic interventions: state-of-the-artand future challenges. Am J Respir Cell Mol Biol 2006;34:517–8.

30. Rock J, Konigshoff M. Endogenous lung regeneration: po-tential and limitations. Am J Respir Crit Care Med 2012;186:1213–9.

31. Moodley Y, Thompson P, Warburton D. Stem cells: a reca-pitulation of development. Respirology 2013;18:1167–76.

32. Goulburn AL, Alden D, Davis RP, et al. A targeted NKX2.1human embryonic stem cell reporter line enables identifi-cation of human basal forebrain derivatives. Stem Cells2011;29:462–73.

33. Zhao L, Wang K, Ferrara N, Vu TH. Vascular endothelialgrowth factor co-ordinates proper development of lungepithelium and vasculature. Mech Dev 2005;122:877–86.

34. Yamamoto H, Yun EJ, Gerber HP, Ferrara N, Whitsett JA,Vu TH. Epithelial-vascular cross talk mediated by VEGF-A and HGF signaling directs primary septae formationduring distal lung morphogenesis. Dev Biol 2007;308:44–53.

35. Lahm T, Crisostomo PR, Markel TA, Wang M, Lillemoe KD,Meldrum DR. The critical role of vascular endothelial growthfactor in pulmonary vascular remodeling after lung injury.Shock 2007;28:4–14.

36. Somers A, Jean JC, Sommer CA, et al. Generation oftransgene-free lung disease-specific human induced plurip-otent stem cells using a single excisable lentiviral stem cellcassette. Stem Cells 2010;28:1728–40.

37. Xiao H, Li DX, Liu M. Knowledge translation: airwayepithelial cell migration and respiratory diseases. Cell MolLife Sci 2012;69:4149–62.

DISCUSSION

DR THOMAS K. WADDELL (Toronto, Ontario, Canada): So twoquestions, one about proliferation and one about differentiation.You said that the cells are proliferating, but do you actually havequantitative data about what fold-expansion occurred in theslices or in the ex vivo perfused organs? And 1B, how confidentare you that they will stop proliferating when they need to whenyou have got full recellularization?

The differentiation question is about what you showed us thatthey express, all the good things. You did not show us, I think,very much proof that they did not express pancreas markers, forexample, or thyroid markers or some of the wrong turns on thedevelopmental pathway that they might be very prone to.

DR GILPIN: In terms of proliferation, we have Ki67 staining todemonstrate proliferation in vitro, on slices, and on whole scaf-folds. We’re in the process of 5-bromo-2-deoxyuridine (BrdU)labeling in the cultures to get a time course of proliferative ca-pacity throughout differentiation. We also have basic cell countsat different time points in cell expansion prior to recellularizationthat also indicate proliferation.

Whether they would stop proliferating? I cannot definitivelysay when they would stop proliferating. I assume, or I hope, thatas they reach their committed, mature phenotype, they wouldbecome more quiescent. But that is really an assumption and notsomething that I can say will happen for sure. There may beoverproliferation within the scaffold, and that might be an issuethat we will need to address.

In terms of off-target differentiation, that is an absolutelycorrect statement. I showed you the thyroid transcription factor-1(TTF-1) induction. It is not 100%. It is a mixed population. It is aheterogeneous population. The endothelial differentiation ismuch more specific, with greater than 90% endothelial pheno-type. The percentage of TTF-1 expression would be 40% to 60%on average, so the rest of that is everything else. And I would notargue otherwise.

The goals at this point were to assess whether these cells werecapable of whole-scaffold recellularization. Moving forward, Ithink we are going to have to engineer a selection process to havea pure TTF-1 population for recellularization if we are going to

exclude any off-target differentiation, but that is not what wehave done here.

DR SAVERIO LA FRANCESCA (Houston, TX): Very interestingdata. Just as a surgeon, I am trying to wrap my head around yourlung transplant model because I am not quite sure that at thecapillary level you still have an intact endothelium. You cannotput a lung back in unless you have that, and that, of course, is sortof the Holy Grail of the whole recell process.Another point is that you also in your paper showed oxygen-

ation, but you really should do arterial blood gas measurementsfrom the left cuff, because if you get a mixed sample, it is kind ofnice, but it really does not tell you that the recell lung, in thiscase, the left, is really oxygenating, which I do not think it is. So itis kind of premature. It is certainly the right direction, and thework is interesting, but just kind of a word of caution, if you will.

DR GILPIN: Yes. I agree that the mixed blood sample is not fullyrepresentative of the regenerated organ function, but really, whatwe wanted to know is whether we can move towards keepingthese animals alive, so systemically, if are they being oxygenatedto a sufficient level. That was really the goal of sampling themixed population.But I agree. By occluding the other vein and sampling only the

regenerated organ, I do not think we would get adequateoxygenation. It is not a fully regenerated organ, and that is not thestatement I am trying to make. It is a work in progress, and I thinkthe body is really the place that regeneration is going to workbest—in vivo regeneration—and one of our goals of the trans-plantation model was to demonstrate that this would be possible.In terms of construct perfusability, we have scanning electron

microscopy and transmission electron microscopy to show thevasculature network of the scaffold is indeed intact. We have alsoperformed perfusion-leak assays to demonstrate a lack of dam-age down to the capillary level. Using other cell types, we havealso previously published the extended survival after trans-plantation of regenerated lung constructs (Song JJ, Ann ThoracSurg, 2011), further demonstrating the feasibility of thisapproach.