0022-202X/83/81018-()0025$02.00/0
`Tm: JOURNAL OF INVESTIGATIVE DERMATOLOGY, 81:23‐105, 1983
`Copyright © 1983 by The Williams & Wilkins Co.
`
`Vol. 81, No. 1Supplement
`Printedin U S A .
`
`The Reconstitution of Living Skin
`EUGENE BELL, PH.D., STEPHANIE SHER, PH.D., BARBARA HULL, PH.D., CHARLOTTE MERRILL, M.S.,
`SEYMOUR ROSEN, M.D., ANNETTE CHAMSON, PH.D., DANIEL ASSELINEAU, PH.D., LOUIS DUBERTRET, M.D.,
`BERNARD COULOMB, PH.D., CHARLES LAPIERE, M.D., BETTY NUSGENS, PH.D., A N D YVES NEVEUX, M D .
`Department of Biology, Massachusetts Institute of Technology (EB, SS, BH, CM AC, & DA), Cambridge, Massachusetts; Pathology
`Department, Beth IsraelHospital, HarvardMedicalSchool, and the Charles A. Dana Research Institute (SR), Boston, Massachusetts.
`U S A ; Laboratoire de Biochimie, UER de Médecine (AC), St. Etienne Cedex, France; CIRD, Sophia Antirolis (DA), ValbonneCedex,
`France; Service de Dermatologie, HépitalHenriMondor (LD), Creteil, France; Centre de Recherches du Service de Sante des Armees (YN),
`Clamart, France; andHopitalde Bauiere, Clinique Dermatologique, Université de Liege (CL), Liege, Belgium.
`
`and in vivo and to formulate a new approach to skin transplan‑
`A living-skin equivalent useful as a skin replacement
`and as a model system for basic studies has been fabri-
`tation and the problem of graft rejection. We present here a
`cated and tested extensively. It consists of t w o compo-
`review of our work.
`It has longbeen realizedthat tissue cells cultivated onplastic
`nents: (1) a dermal equivalent made up of fibroblasts in
`a collagen matrix that is contracted and modified by the
`or glass m a y not function under the same conditions that
`resident cells, and (2) an epidermis that develops from prevailin vivo. This is especially true of connective-tissue cells,
`such asdermal fibroblasts, that ordinarilyare surrounded by
`keratinocytes "plated” on the dermal equivalent.
`an extracellular matrix. However, it is also true of epidermal
`A multilayered keratinizing epidermis with desmo-
`cells resident on a connective-tissue matrix that interact with
`tonofilaments, and hemidesmosomes forms.
`somes,
`Basement lamella formation occurs Within 2 weeks in
`the products of other cells, since such products diffuse through
`vitro when r a t cells are used. With human cells, crypt or
`the matrix.
`The matrix devised for the skin-equivalent model we have
`pseudofollicular morphogenesis is observed in vitro
`developed is initially relatively simple, but it is changed even in
`within 3 weeks after plating cells on the dermal equiva-
`lent.
`vitro as constituent dermal cells interact with it physically and
`contribute to it biosynthetically. It is also changed as keratin‑
`Autografts and isografts of rat-skin equivalents made
`w i t h cultured cells from biopsies a r e rapidly vascular-
`ocytes attachto it,spread, andform anepidermis and basement
`ized, block wound contraction, and persist essentially membrane. Its form may befurther modified asthe epidermis,
`under certain conditions, undergoes morphogenesis.
`for the lifespanof the host.Sevento 9 days after grafting,
`To reconstitute a dermis, dermal fibroblasts are mixed with
`donor cells becomeactivatedbiosynthetically and mitot-
`ically. By 1 y e a r, the dermal population decreases to a
`collagen, serum, and medium. In the right proportions, a gel
`normal level and the matrix has been extensively r e -
`forms (Fig. 1)and is contracted by the fibroblasts. The mixture
`modeled. The grafts remain free of hair and sebaceous
`can be poured into a mold whose shape determines the geome‑
`try of the contracted tissue. If poured into a petri plate, the
`glands. Grafts to rats have beenin place f o r o v e r 2 years.
`Now, allografts of dermal equivalents have been made
`tissue forms asa disk, of diameter much smaller than that of
`the plate. A fivefold decrease in diameter is usual. The con‑
`across a major histocompatibility barrier and a r e n o t
`rejected. The persistence of cellular elements of
`tracted tissue equivalent floats just below the surface of the
`the
`grafts is monitoredby 'use of a genetic marker. Challenge medium expressed from it. The rate of contraction is propor‑
`tional to cell number and inversely proportional to lattice
`of the allograft with a second skin-equivalent graft after
`1 month does n o t result in rejection of the original graft
`collagen content (Fig. 2). The condensation of fibrils of the
`lattice occurs asthe result of a “collection” process executed by
`0r 0f the second skin-equivalent graft. We propose that
`cells asthey extend andwithdraw cytoplasmic podia that attach
`tissue equivalents are tolerated because
`allografts of
`cells with class II antigens a r e selected against d u r i n g in
`to collagen fibrils. The latter are drawn toward the cell body in
`vitro cultivation and a r e excluded f r o m the graft. Thus
`the course of podial contraction. Moving cells also pick up and
`carry with them some fibrils encountered during translocation.
`the fabrication of skin-equivalent tissues or of other
`The active condensation of fibrils by cells releases or expresses
`equivalent tissues w i t h parenchymal cells that do n o t
`bear class II antigens m a y render transplants of such
`fluid from the matrix, which further reduces interfibrillar dis‑
`tissues immunologically acceptable despite the presence
`tances. There is some degree of fibril ordering, since in the
`vicinity of cells, bundles of fibrils are aligned in parallel arrays
`of allogeneic cells. The capacity to graft across major
`histocompatibility barriers using l i v i n g tissue equiva-
`' (Fig. 3).
`Grossly, the contracted dermal equivalent, after 1to 2 weeks
`lents m a y have important clinical significance.
`in vitro with or without an epidermal overlay, is tissue-like in
`its consistency (Fig. 4). It is slightly opaque and resists length‑
`The capacity to reassemble components of skin into a living
`wise stress without tearing. Cells in the dermal-equivalent
`functional tissue [1‐3] has provided us with an opportunity to
`study aspects of the differentiation and function of keratino- matrix are responsive to their surrounds. After 4 to 7 days in
`cytes and dermal fibroblasts in a defined organ model in vitro
`the matrix in vitro, fibroblasts are growth-regulated; that is,
`they cease to incorporate DNA precursors [4] (Fig. 5). The
`regulation is not due to contact inhibition because cell density
`1510W am? Fells a_re weH-separated from'one' another (F1g. 6).
`The condition is hke that of dermal cells In V l V O that are out of
`cycle but not contactumhibited. However, cells in monolayer
`cultures continue to incorporate label until they are confluent.
`UNIVERSITY OF MASSACHUSETTS - EX. 2001
`
`This work was supported by Flow Laboratories and Grant No.
`DAMD17-80-C-0071 from the US. Army Medical Research and De-
`velopment Command.
`Reprint requests to: Dr. Eugene Bell, Department of Biology, Mas-
`sachusetts Institute of Technology, Cambridge, Massachusetts 02139.
`
`93
`
`

`

`July 1983
`
`35
`RECONSTITUTION OF L I V I N G S K I N
`treatment is needed to make cells permeable to the reagent
`D A B [5] (Fig. 7).
`Monolayered cells also differ from cells in dermal-equivalent
`lattices with respect to collagen processing. In cells in lattices,
`
`cells ( 1 2 x |05)+ 2 25 mg
`colluqen+ medium p H 7 4
`lomake 5m!
`
`FIG 1. Procedure for forming a tissue equivalent: Cultivated cells
`are combined with collagen, medium, and 1.0 ml of fetal calf serum.
`Cells are added after the pH is raised to 7.4. A, The ingredients are
`quickly poured into a bacteriologic petri plate with swirling. Almost
`immediately the mixture gels. B, As cells compact collagen fibrils in a
`period of days, fluid is expressed from the collagen lattice that is
`gradually contracted away from the walls of the dish. It floats below
`the surface of the medium, which is squeezed out of the gel.
`
`l8
`l6
`I4
`12
`
`3 4
`2
`0 1 0 2 0 3 0 4 0 5 0 6
`cells x IO'5
`conc. m g / m t
`FIG 2. The extent of lattice contraction, measured as the diameter
`of the tissue equivalent attained after 10days in vitro, is plotted as a
`function of collagen concentration (left) and cell number (right). Initial
`diameter of the uncontracted lattice is 53.0 mm. At
`the highest cell
`concentration shown, the lattice is fully contracted by 2 days [1].
`
`5
`
`6
`
`|
`
`FIG 3. View of living fibroblasts in a tissue equivalent taken with a
`polarizing microscope. Fibrils of collagen are aligned in bundles closely
`associated with cells (X125).
`
`Since cell cycling is blocked soon after cells are incorporated
`into the collagen matrix, the entire population becomes homo‑
`geneous with respect to DNA synthesis and can be compared
`with cells grown on monolayer, in which DNA synthesis is
`blocked because of contact inhibition. Such comparisons have
`been made with respect to several phenotypic features associ‑
`ated with the biosynthetic repertoire of dermal fibroblasts. Our
`results illustrate that cells in dermal-equivalent tissues are in a
`state of differentiation different from that of cells grown as a
`menolayer on plastic.
`For example, perinuclear peroxidase activity is absent in
`human dermal log-phase cells and in cells grown to a confluent
`monolayeronplasticplates (Fig. 7).Furthermore,cellsattached
`to the plastic substrate are impermeableto diaminobenzine, the
`reagent used to demonstrate oxidase activities. They become
`permeable to the reagent, however, when released from the
`substrate with trypsin. Entry of the reagent can be monitored
`bythe appearance ofmitochondrialcytochrome coxidase stain‑
`ing.However,cells incorporatedinto dermal-equivalent lattices
`rapidly become like some cells of intact skin. By 3 days, many
`cells exhibit perinuclear peroxidase activity, and no special
`
`FIG 4: A contracted dermal equivalent formed from a gel containing
`7.2 X 10"fibroblasts cast in square mold after 5 days. Collagen concen‑
`tration was 1.8 mg/ml.
`
`

`

`4 s
`
`B E L L E T A L .
`
`Vol. 81, No. 1 Supplement
`
`IO
`
`I
`I
`I
`
`--- Grown in T25 flask
`Grown in collagen lattice
`‐
`
`, - , .
`)«
`
`‘3’
`‘‑
`.‘I.h~'
`,
`'
`‘Mofifi Mr
`
`v
`
`”V a
`
`FIG 7. Human cells in a dermal equivalent for 3 days ( a ) and cells
`in a fresh biopsy (b) synthesize perinuclear peroxidase seen as a dark
`band around the nucleus after staining with diaminobenzine (DAB)
`[5]. c, Cells grown as monolayer show no perinuclear staining. Note
`that all cells show staining of mitochondrial associated cytochrome c
`oxidase, indicating that the reagent, DAB, has penetrated the cells (a:
`X7400; b: x6300; c: X5700)
`
`SP ACTIVITY OF PROLY
`HYDROXYLASE IN LATTICLES
`C 6 0 *
`
`SP ACTIVITY OF PROLYL
`HYDROXYLASE IN MONOLAYERS
`SOOI-
`0’
`C
`
`IO
`5
`30
`20
`TIME (MIN)
`TIME (MIN)
`FIG 8. Effect of 0.2 mM ascorbate on the specific activity of prolyl‑
`hydroxylase in cells of low (20 P D L ) and high (39 P D L ) population‑
`doubling levels (PDL) grown in collagen lattices (left) or as monolayer
`(right). Open circles, without ascorbate; closed circles, with ascorbate.
`the low passage, the increase is 65-fold, while at the high passage,
`At
`it is 3.3-fold.
`
`O
`
`l0
`
`O
`
`l5
`/
`o
`
`I5
`
`PDLZO
`
`o
`
`5
`
`PDL39
`
`.
`
`B
`“r
`c 300
`R
`o
`G 200‐‑
`R
`G
`D |O
`N
`A
`
`/
`
`400
`
`o_
`
`O
`0
`c 5 0 0 ‐
`p
`
`BA
`
`y m L ‑
`M|
`E 300'‑
`8R 200-‐
`A
`M Vloot‑
`o
`0
`
`/
`
`5
`“r
`c
`R
`o
`G
`R
`M
`A
`D
`N
`A
`
`5
`
`20
`I5
`IO
`Time ih culture (days)
`FIG 5. AG 1519 human fibroblasts (supplied by the Institute for
`Medical Research, Camden, N J . ) grown as monolayers or in tissue
`equivalents were labeled continuously with 1.0pCi [methyl-':‘H].TdR.
`the time points shown, samples were taken and acid‑
`At each of
`precipitable radioactivity measured after extracts were solubilized [4].
`
`25
`
`FIG 6. Cells incorporated into a dermal equivalent [4] and vitally
`stained with fluorescein diacetate after the tissue has contracted have
`ceased to incorporate DNA precursor even though they are well sepa‑
`rated from one another (X1300).
`
`regardless of population-doubling level, prolylhydroxylase ac‑
`tivity is not enhanced by addition of ascorbate, while in con‑
`fluent monolayers it is enhanced preferentially in cells of low‑
`population‐doubling level [6] (Fig. 8).
`
`10
`
`20
`
`3O
`
`40
`
`50
`
`O
`3 5 F
`
`PDL39
`
`30L : J
`25‘‐
`20f
`
`Is
`
`1 0 ' ‐
`
`5L
`o
`
`0
`
`IO
`
`4o
`
`50
`
`

`

`July 1983
`In addition, collagen is processed differently by cells grown
`in lattices, ascompared with cells grown asa monolayer. While
`newly made collagen is added to the lattice structure of the
`dermal equivalent, in monolayer cultures, 95 percent of newly
`made collagen remains in solution unable to polymerize [7].
`Other proteinssynthesized by fibroblasts for export also become
`associated with the skin-equivalent matrix in vitro.
`Several lines of evidence suggest that a lytic system able to
`degrade the dermal-equivalent tissue can become active in vitro
`[7]. First, chemical amounts of dialyzable labeled hydroxypro‑
`line can be found in the culture medium bathing the tissue
`equivalent; second, over aperiodof 9 days there is a measurable
`decrease in the unlabeled collagen used to make up the tissue
`equivalent; andthird, degradation of newlysynthesizedcollagen
`associated with the dermal equivalent is twice as great as
`degradation of newly made collagen in monolayer cultures.
`Giventhat there is bothpolymerizationanddeposition of newly
`made collagen in the tissue equivalent, as well as collagen
`degradation, the system providesan opportunity to study tissue
`remodeling in vitro.
`Hence a number of enzyme systems, i.e., perinuclear peroxi‑
`dase, prolylhydroxylase, a collagenolytic activity, and probably
`procollagenpeptidases,behave differently in atissue-equivalent
`milieu in vitro than in monolayer cell cultures. These findings
`suggest that cells in the tissue matrix in vitro are led to adopt
`a state of differentiation different from that of conventionally
`cultured cells. They also suggest that conditions in the tissue‑
`equivalent modelm a y bemorelike those which cells experience
`in vivo. This appears to be borne out in studies of more complex
`tissue constructs fabricated in our laboratory.
`Human keratinocytes plated onto contracted dermal equiv‑
`alents made up with human dermal fibroblasts quickly attach
`to the collagen substrate, multiply, and spread to form a con‑
`tinuous sheet [8]. The sheet differentiates, becoming multilay‑
`ered with desmosomes, tonofilaments, and keratohyalin gran‑
`ules (Fig. 9). Rat epidermis on a dermal equivalent made up
`with rat fibroblasts has been shown to elaborate basement
`lamellain 2weeks in vitro [21] (Fig. 10).With humanepidermis,
`conditions that promote development of basementlamella have
`not yet been found. Even after 1month in vitro, only hemides‑
`mosomes have begun to differentiate. Specific antisera to col‑
`lagen type IV have failed to give evidence of a lamellar struc‑
`ture. However, differentiation of keratin proteins by human
`keratinocytes grown on the dermal equivalent is quite normal
`in vitro, particularly when the epidermis israisedslightly above
`the level of the medium so that it
`is exposed to air. Protein
`extracts from labeled cells fractionated on SDS PAGE show
`the presenceof low- but also high-molecular‐weight keratins
`(65 K) (Fig. 11).
`We report for the first time that human epidermis undergoes
`morphogenesis when cultivated for long periods on a dermal
`equivalent substrate in vitro [9]. By 3 weeks, deep crypts 01'
`follicle-like structures containing deposits of keratin are ob‑
`served (Fig. 12a). The presence of fibroblasts is essential for
`crypt formation.
`Irradiation of fibroblast-contracted dermal
`equivalents with 60Co prior to application of keratinocytes
`blocks crypt
`formation (Fig. 12b). When keratinocytes are
`incorporated into a collagen lattice with fibroblasts, they give
`evidence of intense biosynthetic activity when tissue equiva‑
`lents are labeled with radioactive protein precursors (Fig. 12c).
`The absence of living fibroblasts results in a lowered level of
`biosynthetic activity (Fig. 12d). Epidermal differentiation also
`depends on the physical conditions .of culture. When skin‑
`equivalent tissues are cultivated below the surface of the me‑
`dium, the epidermis constitutes itself as a monolayer and little
`crypt formation is observed; however, the epidermis becomes
`multilayered when exposed to a gaseous interface.
`The skin-equivalent modelhas served asan assay system for
`a complex cell
`function, namely, contractility. Fibroblasts,
`smooth-muscle cells, cardiac cells, and even epidermal cells,
`
`RECONSTITUTION OF L I V I N G S K I N
`
`5S
`
`Fm 9. Multilayered human epidermis developed1n21d a y s1nvitro
`after‘‘plating" a suspension of keratinocytes obtained from a biopsy
`onto a contracted dermal equivalent. Desmosomes ( D ) , keratohyalin
`granules ( K ) , and tonofilaments ( T ) are apparent
`(reduced from
`x8800).
`when incorporated into a lattice, will contract it at a character‑
`istic rate. The model has provided a way of comparing cells
`from old and young donors, cells of different population-dou‑
`bling levels [1], or cells from different sources in the organism.
`The results can be precisely quantitated using change in tissue
`equivalent diameter to measure the rate of contraction. Cell
`number c a n be measured by determining DNA content fluoro‑
`metrically [6]. The assay lends itself to study of the effects of
`environmental, pharmaceutical [20], or other agents that come
`into contact with the skin.
`We wish to,suggest that the model can be further elaborated
`(1) by manipulating the composition of the matrix and (2) by
`varying the character of the cell population. In work carried
`out so far, some cell types present in the dermis have been
`selected against, since the dermal equivalent has been consti‑
`tuted with cells passaged a number of
`times in vitro. For
`example, the deliberate inclusion of pigment cells has only
`recently beenundertaken.
`In summary, using skin proteins and skin cells derived from
`biopsies, an organ with many properties of skin has been
`reconstituted and used asa model system for studying biosyn‑
`thesis, matrix modeling, differentiation, morphogenesis, and
`interactions among skin cells
`The physical resemblance of skin-equivalent tissues to skin
`(Fig.4) ledusalmost immediatelyto useit asaskinreplacement
`in experimental animals The first grafts were autografts, then
`isografts, prepared as described1nFig. 13
`
`

`

`B E L L E T A L .
`6 s
`..v m t
`e"'
`,fi‘kfA’_*
`
`'
`
`'
`
`a
`
`Vol. 81, No. 1 Supplement
`
`FIG 11. Autoradiogram of apolyacrylamide gel showing presence of
`a 65,000‐dalton as well as lower-molecular-weight polypeptides. An
`extract of epidermis grown on a dermal equivalent for 21 days was
`layered on the slab gel. The tissue was labeledwith 14pCi/ml of an JH‑
`amino acid mixture (New England Nuclear) for 72 hours prior to
`extraction. The molecular-weight standards were from Biorad; the
`middlebandbeingserum albuminat 66,000 daltons. Track 1,standards;
`2,skin biopsy; 3, primary epidermal culture on plastic; 4,subculture on
`plastic; 5, epidermis from submerged lattice; 6, epidermis from unsub‑
`mergedlattice; 7,epidermis appliedto irradiatedlatticesubmerged (gel
`overloaded); 8, epidermis applied to irradiated lattice not submerged
`(gel overloaded). High-molecular-weight keratin bands are seen in
`tracks 2, 3, 5, and 6. Arrow points to ~65 K band.
`
`b
`FIG 12. a, Crypts or follicle-like structures formed in vitro 3 weeks
`after a suspension of human keratinocytes is “plated" onto a dermal
`equivalent.The crypts formonly when the dermalequivalentis eqused
`to air (reduced from X130). b, Dermal equivalent irradiated with ""Co
`prior to application of keratinocytes. Note absence of crypts.
`
`FIG 10. Basement lamella with hemidesmosomes lies between rat
`epidermis developed from a keratinocyte cell suspension and the con‑
`tracted dermal equivalent containing rat dermal fibroblasts. The base‑
`ment lamella (arrows) developed in vitro 2 weeks after keratinocytes
`were plated on the dermal equivalent (a: X20,000; b: x100,000).
`
`After grafting, the skin equivalent heals within 7 days (Fig.
`14). Eventually, the junction between graft and host becomes
`difficult to detect, as seen in a 5-month graft (Fig. 16b). The
`graft is initially translucent (Fig. 16a), makingit possible to see
`the invading capillary circulation, but gradually it develops an
`opaque pink appearance, remaining soft and pliable. Grafts to
`rats at 1 month (Fig. 15) remain free of hair follicles and
`sebaceous glands, and their absence persists for the lifetime of
`the rat.
`The cells of the dermal equivalent prior to transplantation
`appear to be less active biosynthetically than after transplan‑
`tation. In fibroblasts of a 5-day graft, there is little rough
`endoplasmic reticulum, the majority of polysomes are free in
`the cytoplasm, and the nucleus is heterochromatic (Fig. 170).
`Recall that the dermal fibroblasts are mitotically quiescent
`within a week after being cast in a lattice. By 5 days after
`grafting, cell density in the graft remains low, and cells in the
`immediate vicinity of invading capillary vessels are activated
`[10] (Fig. 17a). By 7 to 9 days after grafting, cell density
`increases dramatically [11] (Fig. 17b); virtually all fibroblasts
`of the dermal equivalent of the graft are activatedand the graft
`as a whole is well-vascularized with a network of capillary
`vessels. Activation of the fibroblasts consists of elaboration of
`a well‐developed rough endoplasmic reticulum, a decrease in
`the number of free polysomes, enlargement of the nucleolus,
`and the development of a euchromatic nucleus (Fig. 17d). Skin
`equivalents seeded with equal numbers of fibroblasts were
`grafted to Fischer rats; actual counts of fibroblasts in 5- and 9‑
`
`

`

`July 1983
`
`‑
`
`!
`v
`
`7'
`@
`
`CONTRACTED
`LAJTICE
`cm
`
`FIG 13. Procedure for fabricating a skin-equivalent autograft or
`isograft. A biopsy (1) taken from a donor is cut into fragments that are
`attachedto atissue‐culture plate (2).Cells growingout of the fragments
`are cultivated (3) and incorporated into a collagen lattice (4) which,
`after contraction, becomes the dermal equivalent. A second biopsy (6)
`provides epidermis that is dissociated (7) and applied as a suspension
`to the contracted dermal equivalent (8). The keratinocytes attach,
`proliferate, and form an epidermalsheet that undergoes differentiation
`(9). The skin equivalent is autografted or isografted to a host (10).
`
`FIG 14. Isograft to a Wistar-Lewis host 7 days after grafting a skin‑
`equivalent isograft showing junction between host skin (left) and the
`skin equivalent graft, which lacks hair follicles (reduced from X100).
`
`78
`RECONSTITUTION 0F L I V I N G S K I N
`of host cells was approached by the use of genetic marking
`[11]. Using isogeneic strains, grafts were assembled with cells
`from female donors and grafted to male hosts. After removal,
`cells were recovered from a graft as from a conventional skin
`biopsy (Fig. 18) and cultivated in vitro to provide a population
`of “mitotics,” that is, cells arrested in metaphase by colchicine.
`These were karyotyped and the ratio of female to male cells
`determined. In the dermal component of 9-day grafts in which
`a threefold increase in cell number is observed, the cells can be
`identified as being primarily of the female genotype. Over 92
`percent of cells cultivated from the graft were keryotyped
`female. Control experiments rule out the possibility of selection
`for the female genotype or for cells grafted in skin-equivalent
`lattices that are then returned to the flask for in vitro cultiva‑
`tion. Controls were the following: (1) when cultivated sepa‑
`rately, the doubling rates of male and female cells were shown
`to be the same; and (2) cells from a 1-month graft were kary‑
`otyped from two different passsges in vitro, the fourth and
`ninth, separated by 10to 12 doublings, and the ratio of female
`the fourth passage, 58
`to male cells remained unchanged. At
`percent of the cells were female, while at the ninth passage, 57
`percent were female. A minimum of 20 karyotypes is used to
`establish each time point. Hence, in addition to metabolic
`activation, returning reconstituted skin to the organism pro‑
`vides a trigger for cell multiplication The initial population
`increase in the dermal equivalent creates a zone of high cell
`density which gradually, by some mechanism, perhaps cell
`migration, becomes less populous and by 13months is like the
`surrounding host tissue in cell density. Histologically, at about
`1 week a mild granulocytic responseis sometimes observed;
`rarely, some mononuclear leukocytes are seen. However, these
`responsesaretransient anddo notcompromise the effectiveness
`of the take.
`The modeling of the matrix begun by cells in vitro soon after
`the dermal equivalent is cast continues after the graft is applied
`to an animal host. We have followed the modeling process by
`examining graft sections histologically with a polarizing micro‑
`scope and observe a steady increase in birefringence of the
`intercellular matrix [12]. By 2 weeks the dermis of a graft is
`stabilizedandhasbundlesof fine, tightly packedcollagenfibrils.
`With time, collagen bundles become organized into a basket‑
`weave pattern similar to that of intact skin (Fig. 19). Of several
`hundred autografts and isografts made to experimental animals
`(guinea pigs, rats, and rabbits), no graft rejection has been
`
`FIG 15. Histologic appearance of an isograft to a Wistar/Lewis rat
`after 1 month (reduced from X13).
`
`day grafts demonstrated a threefold increase in cell density
`[11]. Autoradiographic studies of grafts containing cells labeled
`with3H-TdR showedthat labeledcells persist in grafts for short
`periods, but few are seen after grafts have been in place for 5
`weeks [2]. It was observed as well that some labeled cells had
`wandered into surrounding host
`tissues. The question of
`whether the increase in cell number in grafts to rats seen by 7
`to 9 days wasdue to proliferation of grafted cells or to an influx
`
`FIG 16. a, A 7-day graft to a Sprague-Dawley rat viewed grossly
`appears transparent but already vascularized and healed. b, A 5-month
`autograft to a Sprague-Dawley host viewed grossly. The graft-host
`boundary is difficult to perceive and secondary derivatives are absent.
`Black spots are tattoo marks to measure contraction.
`
`

`

`8 s
`
`B E L L E T AL.
`
`Vol. 81, No. 1Supplement
`
`Fm 17. Light and electron micro‑
`graphs of cells in 5- and 9-day isografts
`made to Wistar/Lewis hosts. The fibro‑
`blasts in the 5-day graft (a) are more
`sparsely distributed and more hetero‑
`chromatic than in the 9-day graft (b). (a:
`X280; b: X280). c, EM of a cell in 5-day
`graft shows many free polysomes but
`little rough endoplasmic reticulum; the
`nucleus is heterochromatic (x12,800). (1,
`EM of a fibroblast in a 9-day graft gives
`evidence of a well-developed rough ER
`and a euchromatic nucleus (x12,800).
`
`observed. The best grafts block wound contraction 80 percent
`or better and remain in place indefinitely, unless removed for
`histology.Grossly,the grafts can berecognizedashairlessareas,
`tinted pink, and neither shiny n o r elevated like hypertrophic
`scars.
`To monitor the long‐term persistence of dermal cells, the
`approach again was through genetic marking of grafted cells by
`applying grafts made up with isogeneic female cells to male
`hosts (Fig. 18). The history of the grafted cell population has
`not been Completely elucidated because we have not deter‑
`minedhow m a n y grafted cells migrate out of the graft, n o r have
`we completely eliminated the possibility that some host tissue
`is taken when the graft is sampled to assay for its genotypic
`constitution. The data for a period of over a year (Table I)
`suggest that equilibriummay beneara 50percent value, which
`we think is conservative because of the sampling problem. We
`nonetheless conclude that with time, the number of male cells
`found in a graft approaches the number of female cells, yielding
`a genotypic chimera. This suggests there is a high degree of cell
`flux in the dermal equivalent and that fibroblast mobility may
`be characteristic of the dermis generally.
`By 5 days after grafting, a skin-equivalent graft 3 cm in
`diameter is covered by epidermis, which by inference appears
`to be the epidermisprovided in vitro. Calculation of the rate of
`spreading required for graft coverage would call for a time
`interval of at least 20to 30days to cover a graft of that size if
`coverage depended on ingrowth of host epidermis from the
`periphery at a rate of 0.5 mm/day. Experiments in which
`dermal equivalents are provided with primary epidermis from
`an allotypic donor have led us to conclude that the epidermis
`persists for at least 7 days, even though by that time it is the
`target of a powerful inflammatory reaction (Fig. 20). Labeling
`experiments to confinn whether the original epidermis applied
`in vitro persists after autografting are in progress, but the most
`rigorous demonstration will depend on the use of appropriate
`serologic or genetic markers.
`
`The epidermis of the graft is initially hypertrophic, but by 13
`months it cannot be distinguished from that of adjacent host
`tissue, except that no secondary derivatives are present (Fig.
`14b). As soon as 7 days after grafting a basement lamella is
`apparent (Fig. 21).
`The skin-equivalent modelhasprovided uswith a vehicle for
`a new approach to the prablem of tissue transplantation. In
`particular,wehave beenreexaminingthe basisfor the rejection
`of allotypic grafts. Very recently, the View has been expressed
`that allograft rejection may be triggered by a subpopulation of
`cells in the skin that carry class II antigens, the dendritic 0r
`Langerhans cells [13,14]. Apart from Langerhans cells, there
`are other cells which carry class II antigens, in particular
`leukocytes associated with the microcirculation [15]. There is
`no hard evidence that keratinocytes or capillary endothelial
`cells carry class II antigens; in fact, they m a y not express them,
`particularly if cultured [16,17],but the proofisnotyet rigorous.
`To assess allograft rejection, there are several issues to con‑
`sider: (1) the features of the graft responsible for triggering the
`immune response, (2) the cells involved in the host response,
`and (3) the actual mechanism of graft elimination. Our concern
`until now has been with the first issue.
`.
`The skin-equivalent model makes it possible to manipulate
`the cellular constituents of a graft, that is, to select cells of
`-particular genotype and phenotype for inclusion in a tissue
`equivalent. Part of the selection process, as others have sug‑
`gested [18],entails cultivationof cells, tissues, or organs in vitro
`to filter out those components which m a y be responsible for
`eliciting animmune response.While it is impossiblesofar to do
`this rigorously with intact organs such as skin, it can be done
`with the skin‐equivalent model assembled in the laboratory.
`Fibroblasts, for example, can be selected for and subcultivated
`in vitro to provide a homogeneous population of cells carrying
`only class I antigens. This has allowed us to ask and answer a
`critical question: Can a dermal equivalent made up with cells
`carrying only class I antigens be allografted?
`
`

`

`95
`RECONSTITUTION OF L I V I N G S K I N
`30days later, neither is rejected. Thus the question we posed is
`answered affirmatively: A dermal equivalent made up only with
`cells bearing class I antigens can be allografted and tolerated.
`Our view is that prior cell cultivation resultsin cell selection
`
`FIG 19. Birefringence in a 2-year graft ( a ) and in normal skin (b).
`Fiber bundles in the graft show atypical basketweave pattern of normal
`dermis but are less coarse (X88).
`
`TABLE 1. Percentage of donor cells in fibroblastsgrown from skin.‑
`equivalent grafts
`Isografts
`91
`64
`
`Allegrafts
`
`58
`52
`44
`50
`
`Age of graft at biopsy (days)
`9
`14
`30
`60
`90
`210
`395
`
`54
`42
`
`FIG 18. Procedure for following fate of cells in grafts by genotypic
`marking. A biopsy from a female donor (1) is cut into fragments to
`recover a population of fibroblasts that are cultivated (2) and used to
`make up a dermal equivalent (3). Epidermis from a male recipient is
`dissociated and keratinocytes are “plated" on the contracted dermal
`equivalent (4). The skin equivalent formed is grafted to the male host
`that provided the epidermis (5). After an interval of time on the host,
`the graft is removed and cut into fragments (6) to provide cells for
`cultivation (7). Cells in metaphase blockedby colchicine are shaken off
`and broken hypotonically on a slide to release chromosomes (8).
`
`We have now shown that dermal-equivalent tissues are ac‑
`cepted as allografts across a major histocompatibility barrier
`[19] (Fig. 22). Dermal fibroblasts from female Brown Norway
`rats (RT1.A") were used to make up graftsimplanted on male
`Fischer (RTLA') hosts. Fischerkeratinocytes were usedfor the
`epidermis. Persistence of cells was monitored as before by
`karyotyping. Results are summarized in Table I. Although a
`transient mononuclear reaction is observed at 7 to 9 days, the
`graft is not rejected. Thirty- and 60‐day grafts