Development 107, 77-86 (1989)Printed in Great Britain © The Company of Biologists Limited 1989

77

Regeneration and pattern formation in planarians

III. Evidence that neoblasts are totipotent stem cells and the source of blastema cells

JAUME BAGUNA1, EMILI SAL61 and CARME AULADELL2

1 Departament de Cenetica, Facultat de Biologia, Universitat de Barcelona, Diagonal 645, 08028 Barcelona, Spain2Unitat de Biologia Cellular, Departament de Bioqulmica i Fisiologia, Facultat de Biologia, Universitat de Barcelona, Diagonal 645,08028 Barcelona, Spain

Summary

In most regenerating systems, blastema cells arise bydedifferentiation of functional tissue cells. In planarians,though, it is still debatable whether dedifferentiated cellsor a population of undifferentiated cells, the neoblasts,are the main source of blastema cells. Moreover, it isunclear whether in the intact organisms neoblasts arequiescent cells 'reserved' for regeneration or if theyserve as functional stem cells of all differentiated celltypes. Both uncertainties partly stem from the failure todistinguish by conventional labelling methods neoblastsfrom differentiated cells.

Here we describe a new approach to these problemsbased on testing the regenerative and stem cell capabili-ties of purified neoblasts and differentiated cells whenintroduced, separately, into irradiated hosts. Introduc-

tion of neoblasts led to resumed mitotic activity, blas-tema formation, and extended or complete survival ofthe host; differentiated cells, in contrast, never did so.

Therefore, planarian neoblasts can be qualified astotipotent stem cells and the main source of blastemacells, while dedifferentiation does not seem to operateeither in intact or regenerating organisms. In addition,these results strengthen the idea that different types ofregeneration and blastema formation, linked to thetissular complexity of the organisms, are present in theanimal kingdom.

Key words: neoblast, planarians, regeneration, blastema,pattern formation, stem cells, totipotency.

Introduction

Although many studies have been published on themechanisms of blastema formation and differentiationin planarians and the cell types that engage in it or donot (see Br0nsted, 1969, and Chandebois, 1976, forgeneral references) a main question remains still un-answered: do blastema cells arise from a stock ofundifferentiated cells (neoblasts) already present in theintact organism or by dedifferentiation of differentiatedfunctional cells? In the last 50 years, this issue has beenpolarized into what is called the 'neoblast versus dedif-ferentiation' controversy, confronting the so-called'neoblast theory' and the 'dedifferentiation theory'(Slack, 1980; Baguna, 1981).

The 'neoblast theory' stems from Wolff and Duboisdemonstration that the lack of regeneration afterX-irradiation in planarians was caused by the progress-ive disappearance of neoblasts and lack of cell turnover,and from the observation that an organism with anunirradiated region (either a shielded region or anunirradiated graft) formed a blastema after a delayproportional to the distance between the wound and thehealthy unirradiated tissue (Wolff & Dubois, 1948).This suggested that a blastema can be formed by

migratory undifferentiated cells (neoblasts) that hadmigrated throughout the irradiated region from theunirradiated area. Hence, neoblasts were qualified astotipotent and migratory cells capable of forming all thetissues of the regenerate (Dubois, 1949; Lender, 1962;Gabriel, 1970).

The 'dedifferentiation theory', mainly based on histo-logical and electron microscopic data (Flickinger, 1964;Woodruff & Burnett, 1965; Hay, 1966; Rose & Shostak,1968; Coward & Hay, 1972; Chandebois, 1976),suggests that neoblasts do not exist at all, or at least arenot involved in regeneration, and that dedifferentiationof specific cells near the wound is the main way ofrecruiting cells to form the blastema. This is analogousto the situation found in regeneration of vertebratesand higher invertebrates.

As presently stated, both theories are incomplete andopen to criticism. The neoblast theory does not excludethe possibility that the migratory neoblasts making theblastema may have resulted from the dedifferentiationof differentiated cells in the unirradiated region beforetheir migration to the wound. Moreover, it is alsopossible that in normal regeneration differentiated cellsnear the wound can dedifferentiate and contribute,albeit slightly, to making the blastema. Finally, it has

78 /. Baguna, E. Said and C. Auladell

been shown that the so-called migration of neoblasts isnot a true cell migration but the result of the slow andprogressive spreading of neoblasts mainly caused byrandom movements linked to cell proliferation (Sal6 &Baguna, 1985). Therefore, the idea of neoblasts astotipotent and migratory cells making exclusively theblastema is still unsettled.

The evidence for dedifferentiation in planarian re-generation is also not well founded. Three main criti-cisms can be raised against it. First, evidence fordedifferentiation of gastrodermal cells to neoblastsclaimed using,vital dyes as markers (Rose & Shostak,1968) is hardly acceptable as the markers used diffusefrom cell to cell and proper controls were not done.Second, recent EM studies have not found, despite acareful search, any evidence of, or role for, dedifferen-tiation in the formation of the regenerative blastema inseveral species of planarians (Bowen etal. 1982; Hori,1983, 1986; Morita & Best, 1984). Finally, mitoticstudies have shown that the early blastema is formedfrom very early dividing (1-8 h of regeneration) G2neoblasts, a period too short to make dedifferentiationlikely (Sal6 & Baguna, 1984).

A case for dedifferentiation in planarians wasbrought up some years ago by Gremigni and co-workersin a very interesting series of experiments (Gremigni &Miceli, 1980; Gremigni etal. 1980, 1982). Using a strainof planarians that are naturally occurring mosaics: thesomatic cells are triploids, the male germ diploids, andthe female germ hexaploid; they showed that regener-ation from a cut surface through the gonadal regiongave rise to blastemas and regenerates that containedmainly triploid cells, but also diploid and/or hexaploidcells from which somatic cells (e.g. pharyngeal musclecells) originated. This suggested that dedifferentiationand transdifferentiation (and hence, metaplasia), how-ever limited, occurred during planarian regeneration.

These results, held as conclusive evidence for dedif-ferentiation in planarians and, hence, for similar mech-anisms of blastema formation in most animal groups(Slack, 1980), can be criticized on the grounds that theydo not demonstrate the occurrence of dedifferentiationand metaplasia but, at the most, suggest the existence ofdedetermination (or transdetermination) (Baguna,1981). This is because the loss of a haploid complementduring spermiogenesis and its doubling duringoogenesis, though one of the first steps from neoblaststo germ cells, is only a small step in cell determinationand occurs in undifferentiated cells of the germinativeepithelium, which are undistinguishable from somaticneoblasts. Moreover, it is known that differentiatingand differentiated germ cells like spermatocytes, sper-matids and spermatozoa, and its counterparts in thefemale germ line, degenerate and lyse after transection(Fedecka-Brunner, 1967; Bowen etal. 1982) and, there-fore, cannot dedifferentiate to give blastema cells.

Overall, the whole issue is clearly unsettled andneeds clarification. A way out from this riddle would beto have clear cell markers to distinguish neoblasts fromdifferentiated cells, making it possible to follow theirfates during blastema formation. As neoblasts are the

only known mitotic cells in planarians, several attemptshave been made to label them with tritiated thymidineto determine whether blastema cells are made oflabelled cells (coming from neoblasts) or unlabelledcells (coming from dedifferentiated cells). However,except for the reported incorporation of this precursorinto neoblasts in the small turbellarian Microstomumlineare (Palmberg & Reuter, 1983) and Convoluta sp.(Drobysheva, 1986), nobody has succeeded so far inlabelling planarian neoblasts with thymidine or otherprecursor (for a discussion, see Coward etal. 1970). Onthe other hand, the very few attempts made to labeldifferentiated cells specifically have also been unsuc-cessful so far (Sal6, 1984).

To overcome these difficulties, we describe here anew approach based on purifying, from total cellsuspensions, neoblasts and differentiated cells, andtesting their regenerative and stem cell capabilitieswhen introduced, separately, into the parenchyma ofirradiated hosts which have neither functional neoblastsnor mitotic activity. The rationale of this experimentalapproach is that if neoblasts are the stem cells of all (ormost) differentiated cell types and the main cell typemaking the blastema, their introduction into a dyingnonregenerating irradiated host will lead to its survivaland regeneration, whereas introduction of differen-tiated cells will neither rescue the host nor make itregenerate. Conversely, if dedifferentiation does occur,introduction of differentiated cells will, eventually,result in their dedifferentiation and, therefore, in therescue of the host and/or in its regeneration.

Whatever the outcome, the results obtained will be ofgeneral significance as they would support either theexistence in all animal groups of a general mechanism toobtain undifferentiated regenerative cells (e.g. dedif-ferentiation), or the presence, in different animalgroups, of different mechanisms that may be linked tocertain structural or functional characteristics.

Materials and methods

SpeciesPlanarians used in this work belong to the fissiparous strain ofthe species Dugesia(G)tigrina, and to the sexual and asexualraces of Dugesia(S)mediterranea (Sal<5 & Baguna, 1985).They were maintained in Petri dishes in the dark at 17 ± 1°Cin planarian saline (PS; Sal6,1984) and fed with Tubifex sp. Inall experiments, one-week-starved organisms were used andthe temperature kept at 17 ± 1°C.

Enrichment of neoblasts and differentiated cellsEnriched fractions of neoblasts and differentiated cells wereprepared, from initial total cell suspensions, by serial fil-tration and Ficoll density gradients, respectively. Suspensionsof total cells were obtained from one-week-starved donororganisms, 8-10mm in length. Briefly, donors were cut intopieces and dissociated by gently pietting into single cells in amodified Holtfreter solution (MHS; Betchaku, 1970) that hadkanamycin sulphate (Sigma, London) at 10 fig ml"1. Cellswere pelleted at 300 g for 5min and resuspended in MHS at106 cells ml"1, this being considered the total cell fraction. Thepercentages of neoblasts and differentiated cells were esti-

Neoblasts in planarian regeneration 79

mated by phase-contrast microscopy (Bagufia & Romero,1981).

Neoblasts were enriched following a serial sieving pro-cedure that separates cells by size. 5 ml samples of the totalcell fraction were serially filtered through nylon meshes ofdecreasing pore sizes (180, 50, 20 and 8jum). Being thesmallest cell type (mean diameter, 7-0 ± 2-2 fim; Auladell andBagufia, unpublished data), neoblast is the main cell typepassing through the last filter. Enriched neoblasts werecentrifuged at 400 g for lOmin, washed and resuspended inMHS.

Differentiated cells were enriched by Ficoll density gradi-ents (Collet et al. 1984; Sal6,1984). 1 ml samples of total cells(106 cells ml"1) were placed on top of a discontinuous Ficolldensity gradient (Ficoll 400, Pharmacia; 10 ml fractions of 3,6, 9 and 12 % Ficoll in MHS) and spun at lOOOg for 1 h at 4°C.Differentiated cells were mainly recovered at the 3-6%interphase. Cells were collected, washed twice and resus-pended in MHS.

The degree of enrichment was determined comparing thepercentage of each type of cell between the initial and theenriched fraction. Viability was assessed by the Trypan blueexclusion test. Yield (in percentage) indicates the number ofneoblasts and differentiated cells present in the enrichedfraction as compared to those present in the initial total cellfraction, all measured with a hemocytometer under phase-contrast microscopy (Bagufia & Romero, 1981).

Host organismsOrganisms to be used as hosts were fully grown (10-12 mmlong) adults of the same species kept under identical con-ditions to donor organisms.

IrradiationOrganisms to be used as irradiated hosts were exposed to8000 rads (1000 rads min"1) using a HT-100 Philips X-raymachine (l'7mm Al filter; 100 kV, 8 mA) and kept in PS withkanamycin sulphate.

Number of neoblasts in nonirradiated and irradiatedorganismsAt different intervals after irradiation, organisms were macer-ated into single cells, and the percentage of neoblasts to totalcells quantified by phase-contrast microscopy (Bagufia &Romero, 1981). As controls, nonirradiated intact and regen-erating organisms were macerated and the percentage ofneoblasts estimated.

Introduction of donor cells into irradiated hosts.Experimental procedureThe experimental and control groups and the injectionprocedure employed are depicted in Figs 1 and 2. From theinitial group of intact organisms, 10-12 individuals were keptaside to serve as external nonirradiated controls (a, Fig. 1).The rest (b, Fig. 1; 50 individuals) were irradiated at 8000 rads

•enXI

>

33momzmX>

O

Fig. 1. Diagrammatic representation of the experimental procedure employed in this work (for a detailed explanation, seetext), a: nonirradiated controls; b: irradiated organisms; b,: irradiated controls; b2: organisms to be injected; b2i: sham(saline)-injected organisms; b22: cell-injected organisms. Hatched: irradiated (X-rays) organisms. Cross-hatching: host bodyregions where donor cells were injected.

80 /. Baguna, E. Said and C. Auladell

AU

\JDONOR

X- rays

ice plate (4°C)

IRRADIATEDHOST

triangular window

Fig. 2. Diagram of the procedure employed to introduce total cells (tc), neoblasts (nb) or differentiated cells (dc) fromnonirradiated donors into the parenchyma of an irradiated host (hatched). For more details, see text.

and, one day after, split in two groups; one (b[, Fig. 1; 10individuals) to serve as noninjected irradiated controls, andthe other (b2, Fig. 1; approx. 40 individuals) to be sham-injected with MHS (b2i, 10 individuals) or to be used as hoststo injected donor cells (b22, 30 individuals; 10 organisms foreach cell fraction).

Cell fractions (total cells, enriched neoblasts, and enricheddifferentiated cells) were introduced, separately, into theparenchyma of irradiated hosts (cross-hatching, Fig. 1) by amicroinjection procedure (see below). 2 days after injection,experimental and control groups were transversally cut 1 mmin front and below the injected areas (or at similar levels alongthe anteroposterior axis in control groups) to stimulateproliferation and regeneration, and kept at 17°C in PS. Allgroups were monitored for a period of 60 days and thepercentage of surviving and regenerating organisms recorded.

Prior to injection, cell fractions were centrifuged for 15tninat 400 g and resuspended in 100/d of MHS (final cell concen-tration 5000 cells /J"1). After immobilizing the irradiated hostby cold (1 h at 4°C), a triangular piece (1 mm side) was cut outfrom the postcephalic area of the host and kept aside (Fig. 2).Using a 80-100 /an (inner diameter) glass micropipette, con-nected via a polyethylene tubing to a screw-drive-controlledHamilton microsyringe, 7-8 /il of cell suspension (aprox.35-40X103 cells in MHS, see below) (b22, Fig. 1) werecarefully introduced into the parenchyma of the lateral andrear sides wound area. After the injection, when the needlewas withdrawn, the triangular piece was placed again in thepostcephalic area of the injected host. Injected organismswere kept, wrapped in cigarette paper and placed on top of

saline moistened Whatman no. 1 paper, in Petri dishes in thedark for 12 h at 4°C and for 24 h at 12°C before beingtransferred to PS at 17°C. As controls, a group of irradiatedorganisms (b2i, Fig. 1) was injected with 7—8 t̂l of MHS usingan identical protocol as in cell-injected groups. The resultspresented are the mean of three different experiments.

Estimates of the number and viability of injected cellsThe actual number and viability of injected cells was esti-mated by an indirect labelling method. Briefly, donor organ-isms were fed an artificial food mixture containing fluorescentlatex beads (Fluoresbrite carboxylated, Polysciences; greenfluorescence, 1-0 ̂ m in diameter) that are incorporatedspecifically into the cytoplasm of differentiated cells (Sal6 &Baguna, 1985), mainly gastrodermal and fixed parenchymacells (Baguna and Romero, unpublished data). One day afterfeeding, organisms were dissociated into single cells, differen-tiated cells enriched by Ficoll density gradients (see above)and resuspended at 5000cells ;tl~' in MHS. The number ofdifferentiated cells labelled with fluorescent latex beads in avolume identical to the injection volume (7—8 yul, having35-40xl03 differentiated cells) was estimated by epifluor-escence (Leitz Dialux 20). The average value found(~7-400± 1-600 labelled cells; that is, one out of every fivecells) was taken as a control zero value.

2h after injecting 7—8/tl of enriched, partially labelled,differentiated cells (~35-40xl03 cells), irradiated hosts weredissociated into single cells and the total number of livelabelled cells estimated and compared to control zero values.

Neoblasts in planarian regeneration 81

Similar measurements were made also at different periods (2,5,10,15,20 and 40 days) after injection. As controls, identicalexperiments were made using as hosts intact nonirradiatedorganisms.

Taking into account that labelled cells represent between 18and 22% of the total differentiated cells injected, andassuming similar losses between labelled and unlabelled cellsduring injection and once within the host, the number oflabelled cells recovered from the host at different times afterinjection serves both to estimate the number of cells actuallyintroduced and kept for different periods of time within thehost, as well as their viability. Besides, though this labellingmethods works only for differentiated cells, it is a reasonableassumption that similar results would apply for total cells andenriched neoblast fractions.

Criteria for mortality, survival and regenerativecapabilities in noninjected and injected organismsLange & Gilbert (1968) showed that supralethal X-ray doses(8000-10000 rads) given to several species of planarianssterilizes all the neoblasts and leads invariably to the death ofthe organism within 20-40 days after irradiation in a species-specific time course. The criterion for mortality was that theentire animal be lysed with no piece left intact. Following thiscriterion, injected planarians are here considered survivors ifthey are alive and fully functional at 60 days postinjection.The regenerative capability of irradiated injected organisms isconsidered to be restored when an anterior blastema bearingeyes appears before 20 days after injection and cutting.

Chromosomal markerTo distinguish injected from host cells we have made use of achromosomal difference (a heteromorphic chromosome pair;Bagufia, 1973) between the sexual and asexual races ofDugesia(S)mediterranea (for more details, see Sal6 &Bagufia, 1985).

Mitotic indexControl and experimental organisms were fixed in 1 N-HCI,stained by a modified Gomori technique and mounted whole(Sal6 & Bagufia, 1984). Mitotic figures and nuclei werecounted, with the aid of an ocular grid divided into 100squares, in a 2 mm strip of tissue along the anteroposterior(cephalocaudal) axis, above and below the wound (seeFig. 2). The number of mitoses were recorded from 20 of thesquares and the mitotic index (number of mitoses/100 nuclei)calculated.

Results

The pattern of mortality in irradiated organismsAfter irradiating at 8000 rads, planarian cells do notproliferate, cell renewal slows down and organisms diewithin 3-5 weeks depending on species, temperature,age and nutritional conditions (Lange, 1968; Chande-dois, 1976). At 17°C and in the culture conditions usedin this work, mean survival times (period in days whenhalf of the irradiated population is still alive) forDugesia(G)tigrina and Dugesia(S)mediterranea were21-8±2-6 (n = 30) and 38-6±5-3 (n = 30) days, respect-ively. The observed pattern of lysis starts in the cephalicand caudal regions and progresses proximally to thecenter until the organism vanishes.

As shown by Lange (1988), death of irradiated

30

25

20

o 15c

o

10

0 1 14days

Fig. 3. Variation in the percentage of neoblasts to totalcells in nonirradiated (dashed lines) and irradiated (X-rays,8000 rads; continuous lines) intact (open circles) andregenerating (filled circles) Dugesia(G)tigrina. The resultsare means ± S.D. (n = 3; 10-12 individuals for eachexperiment). Data for nonirradiated organisms are fromBagufia & Romero (1981).

organisms is due to lack of cell renewal, the latter beinga consequence of lack of proliferation and subsequentdeath of undifferentiated cells (neoblasts). It should beexpected, therefore, that the number of neoblasts willdecrease after irradiation. The results found in bothspecies (see Fig. 3 for Dugesia(G)tigrina) clearlyshows, for both intact and regenerating irradiated andnonirradiated organisms, that the percentage of neo-blasts decreases steadily after irradiation to values lessthan 1% of total cells at two weeks postirradiation.Mitotic figures were never seen in irradiated organisms.

Isolation and partial purification of neoblasts anddifferentiated cellsTable 1 summarizes the main results and Fig. 4 (A-C)ilustrates the appearance of the initial and enriched cellfractions.

Although pure populations of neoblasts and differen-tiated cells were not obtained, differences in percentagefor each class of cells between enriched and initial (totalcells) fractions indicates that a considerable enrichmenthas been achieved, this being particularly relevant forneoblasts (30 to 88%). The viability of enriched cellswas found to be high for all fractions, a fact alsosupported by their appearance under electron mi-croscopy (Auladell & Bagufia, unpublished data). Theyield was, however, rather poor for all fractions, mainlyfor differentiated cells (only 18% of the initial cellsbeing recovered).

82 J. Baguna, E. Said and C. Auladell

Fig. 4. Phase contrast micrographs of initial and enrichedfractions from the planarian Dugesia(G)tigrina. (A) totalcells (initial cell population); (B) enriched neoblasts;(C) enriched differentiated cells, gc: gastrodermal cells; me:muscle cells; nb: neoblasts; pc: parenchyma cells. Bar,20 jun.

Table 1. Enrichment, viability and yield of thedifferent cell fractions obtained from

Dugesia(G)tigrina*

Cell types (% ±s.D.)tViability Yield

Differ Undiffer (%)tt ( % ± S . D . ) *Fraction

Total cells 70 ±5 30 ±7 >95Enriched neoblasts 9 ± 4 88 ± 8 >90 46 ± 14Enriched different. 87 + 7 10 ± 6 >90 18 ± 6

cells

* Numbers (in %) are mean values ± S.D. (n = 4; up to 1000 cellsfor each measurement). For the sake of clarity average values andstandard deviations have been rounded.

t Percentages of differentiated (Differ) and undifferentiated(Undiffer) cells in each fraction were determined by phase-contrastmicroscopy according to the morphological criteria set in Bagunaand Romero (1981).

t t Viability as assessed by the trypan blue exclusion test.I Yield (in %) indicates the number of neoblasts and

differentiated cells present in the enriched fractions are comparedto its number in the initial total cell fraction, both measured with ahemocytometer under phase-contrast microscopy.

i

oX

eel

• o

—

JOra"o

.DB

z

10

9

8

7

6

5

4

3

2

1

-

-'

\

\

•

•

— / / — '1 — ^ ,

•

0 2 2 f

• ~ - — ,

10 15Time

>•

\

\

\

\

20 40 hoursdays

Fig. 5. Number of fluorescent-labelled differentiated cellsrecovered from irradiated hosts of Dugesia(G)tigrina atseveral periods after injection. The results are means ± S.D.(n = 3; 5-6 individuals for each experiment). For moredetails, see text.

Injection of cells into irradiated hosts. Number andsurvival of injected cellsTwo main problems are encountered when trying tointroduce cells by injection into the parenchyma oforganisms, like planarians, made of a solid mass oftissue: (1) to have a reliable estimate of the actualnumber of cells introduced; (2) to check if these cells,mainly differentiated cells, survive for long periodsonce within the host tissues.

To answer both questions, differentiated cellslabelled with fluorescent latex beads were used (seeMethods). From the number of labelled differentiatedcells recovered 2h after injection, we estimate that20-24X103 cells, out of 35-40X103 cells injected, wereactually introduced within the host. This amounts to60-70% of injected cells. To check the long-term

viability of injected differentiated cells once within thehost, similar measurements were made at differentperiods after injection. The results (Fig. 5) show thatdifferentiated cells stay alive and maintain their num-bers at least up to 20 days after injection.

Differential survival and regenerative capacity ofinjected organismsInjection of total cells and enriched fractions of neo-blasts and differentiated cells led to differential survivalof irradiated hosts (Table 2). Whereas organismsinjected with differentiated cells gave similar survivalcurves to control (noninjected) and sham (saline)-injected groups, injection of total cells and enrichedneoblasts increased significantly the mean survival timeand, in the latter, led to complete survival of some

Neoblasts in planarian regeneration 83

Table 2. Survival and regenerative performances of control and injected groups of Dugesia(G)tigrina

Group*

Control (non irradiated)Control (irradiated)Injected (sham-injected)Injected (total cells)Injected (differ, cells)Injected (neoblasts)

Symbol*

abib2ib22b22b22

Numberof

organisms

454130343532

Mean survivaltimet

(days±s.D.)

-22-4 ±3-221-6 ±4-532-6 ±4-624-2 ±2-134-7 ±3-8

Survivorstt(60 days)

4500006

MI*(xlO2)

7-8 ±2-200

0-3 + 0-10

2-0 + 0-4

Regenerationtt(20 days)

+ (45)

---+ (6)

*See legend of Fig. 1.t Mean survival time is the period (in days) when half of the irradiated population is still alive.t t Survival at 60days, taken as a sign of complete recovery, indicates the number of organisms staying alive and fully functional 60 days

after irradiation.t Mitotic index (MI; number of metaphases/lOO cells), determined 12 days after injection using standard procedures (see Methods).Xt Regeneration at 20 days indicates the presence (+) or absence (-) of blastemal structures 20 days after cutting; in parentheses, number

of positive individuals.

organisms. Total cell fractions, where neoblasts rep-resent 30 % of total cells, increased by 10 days the meansurvival time, though all organisms finally die. Incontrast, enriched neoblast fractions, where these cellsrepresent almost 90% of the cell population, increasedthe mean survival time by 12 days and led in 20 % of theorganisms (6 out of 32) to complete recovery of theirradiated hosts.

The regenerative capacity of injected hosts parallelsthe results found for survival (Table 2). Like controland sham-injected irradiated organisms that do notregenerate, organisms injected with differentiated cellsneither regenerate nor build a blastema. Instead, thoseinjected with total cells gave very small transientblastemata that regressed 8-10 days after injection,whereas most organisms injected with enriched neo-blasts gave blastemata of almost normal size, somebearing new eyes, appearing at later periods (c. 12-15days after cutting) than in control nonirradiated groups(c. 4-6 days after cutting).

As these results suggested that survival and regener-ative capacity of irradiated hosts relies on the actualnumber of neoblasts introduced and since these cellsare known to be the only cell type in planariansendowed with mitotic capacity, it should be expectedthat differential survival and regenerative capacitystems from different rates of cell proliferation. Mitoticanalyses in injected hosts showed, accordingly, thatonly those injected with total cells and neoblasts hadsignificant levels of mitoses. Instead, hosts injected withdifferentiated cells never had any mitotic figure(Table 2).

Evidence that neoblasts are the stem cells of alldifferentiated cells in planariansTo establish that donor neoblasts, and not revitalizedhost cells (either neoblasts or differentiated cells), leadto increased or complete survival of the host, injectionof enriched neoblasts into irradiated hosts between thesexual and asexual races of Dugesia(S)mediterranea,which differ in a chromosomal marker (Sal6 & Baguna,1985), were performed following an identical exper-imental procedure (see Figs 1 & 2). This experiment

also had an important additional interest: to ascertain ifinjected neoblasts from one race may replace all thedifferentiated cells of the host and, hence, 'transform'one race into another and vice versa. If this were so,neoblasts could unambiguously qualify as the stem cellsof all differentiated cell types in planarians.

The results found showed, in all combinations tested,that mitotic figures within the host belong always todonor cells and not to revitalized host cells (data notshown). These results agree with earlier reports, usingtissue grafting procedures that had pigmentation,ploidy or chromosomal differences as cytologicalmarkers (see Sal6, 1984; and Sal6 & Baguna, 1985, forgeneral references), in ruling out a possible revitaliza-tion of host cells either by the injection procedure or bythe injected cells.

As expected, injection of neoblasts from the asexualrace of Dugesia(S)mediterranea to irradiated hosts ofthe sexual race rescued the host and transformed it intoan asexual individual able to reproduce by fission butunable to reproduce sexually. Conversely, the introduc-tion of neoblasts from the sexual race into irradiatedasexual hosts transformed the latter into individualsunable to reproduce asexually and capable, after devel-oping germ cells and the copulatory complex, to mateand lay cocoons (Sal6, 1984; data not shown).

Discussion

We have analyzed the capabilities of neoblasts anddifferentiated cells to give blastema cells during regen-eration and to be the stem cells of all differentiated celltypes during the daily cell renewal in intact organisms.The analysis has been based on the differential survivaland regenerative capacity of irradiated hosts wheninjected with different cell fractions. Our main con-clusion is that the ease of recovery and regeneration ofirradiated hosts is proportional to the number ofneoblasts introduced and that, at least under theexperimental conditions employed, differentiated cellsare not able to rescue the host or make it regenerate.Taking into account that neoblasts are the only planar-

84 /. Baguna, E. Said and C. Auladell

ian cell type known to divide (Sal6 & Baguna, 1984;Morita & Best, 1984), these results strongly suggest thatthese cells are totipotent (or at least pluripotent) stemcells in the intact organism and the main source ofblastema cells in regenerating organisms. In turn, theseresults argue against the role of cell dedifferentiation asa means to recruit undifferentiated cells both in intactand regenerating planarians.

Of the fractions tested, enriched neoblasts and totalcells were the only ones giving increased survival times,resumed mitotic activity and, in the former, completesurvival. However, it is at first surprising that neoblastinjection gave a low percentage of total rescue, thepercentage falling to zero when total cells wereinjected. These results could be understood if thenumber of neoblasts actually introduced is considered.Assuming that 25xlO3 donor cells are introduced ateach injection (Fig. 5), it follows that 18-22xlO3 and7-8xlO3 neoblasts are introduced, respectively, fromenriched neoblasts and total cell fractions. If the areawhere cells are injected is estimated to span 2 mm inlength along the anteroposterior axis (see Fig. 1), thenumber of neoblasts introduced is 3-4 times (neoblastfraction) or 8-10 times (total cell fraction) lower thanthe number present in a similar area of a controlnonirradiated organism (~35xlO3 neoblasts mm"1 inlength in an 11 mm long organism; Baguna & Romero,1981; Romero, 1987). Interestingly, the ratio of thenumber of neoblasts in experimental and controls isfairly close to the ratio calculated, from the data ofTable 2, for their respective mitotic indices. This sup-ports again the correlation found between neoblastdensity and mitotic activity and suggests that only fairlyhigh rates of mitotic activity (as seen after injection ofenriched neoblasts) can produce rates of cell replace-ment compatible with extended and complete survival.An additional test backing this proposal, based onincreasing the number of neoblasts injected from totalcells and neoblast fractions by increasing the volume ofcell suspension injected, failed because planarian par-enchyma, being a solid mass of cells with little inter-cellular spaces, did not take up more cells (unpublisheddata).

A similar argument could apply to explain both thepresence of permanent blastemal structures only inthose hosts injected with enriched neoblasts as well astheir delayed appearance when compared to regenerat-ing nonirradiated controls (15-20 days vs 6-8 days).Since rates of blastema formation and differentiationare closely correlated with rates of mitotic activity (Sal6& Baguna, 1984, 1988), it may be expected that onlymitotic rates close to (though lower than) controls, asseen after neoblast injection (Table 2), guarantee asupply of neoblasts compatible with blastema forma-tion. However, their numbers will not be enough tokeep pace with normal (control) regenerative rates;hence, the delayed appearance of differentiated struc-tures in the blastemata.

Which is the main role of neoblasts in planarians?The 'transformation' of sexual to asexual races, and vice

versa, in Dugesia(S)mediterranea via introduction ofneoblasts from one race into irradiated hosts of theother demonstrates the main role of neoblasts inplanarians: namely as the stem cells of all (or most)differentiated cell types in the intact organisms. This'transformation' can be envisaged as due to the slow butcontinuous replacement of host neoblasts and differen-tiated cells, unable to divide, by nonirradiated donorneoblasts capable of division and differentiation. Thisprocess will last until no host cells are left and allneoblasts and differentiated cells are of donor geno-type. In other words, injected neoblasts would use theirradiated host as a sort of three-dimensional 'feeder-layer' where slowly turning over irradiated host cells arereplaced by proliferating neoblasts that probably usehost positional cues and signals to differentiate.

It must be pointed out, however, that earlier studiesusing grafts of nonirradiated donor tissue pieces intowhole irradiated hosts had already showed the 'trans-formation' of hosts into donors (Lender & Gabriel,1965; Teshirogi, 1976; Chandebois, 1976), this beinginterpreted as due to the slow and progressive replace-ment of host cells by proliferating and differentiatingdonor neoblasts. In these experiments, however, thepossibility that dedifferentiated graft cells and notneoblasts were the actual source of new cells could notbe ruled out. The results presented here make thispossibility very unlikely, suggesting in turn that planar-ian neoblasts are true totipotent stem cells.

Cell dedifferentiation in planarians: does it still have arole?Functionality of injected differentiated cells within thehost shown by fluorescent latex beads (Fig. 5), lack ofrevitalization of host differentiated cells shown usingchromosomal and ploidy markers, jointly with the lackof recovery, mitotic activity and regenerative capacityof irradiated organisms injected with enriched differen-tiated cells (Table 2), and the criticisms raised in theIntroduction, makes it very unlikely that cell dediffer-entiation plays a substantial role either in intact orregenerating planarians. This is in line with recent dataon thymidine incorporation in other Turbellaria(acoela, Drobysheva, 1986; polyclads, Drobysheva,1988; rhabdocoela, Palmberg & Reuter, 1983) andCestoda (Wikgren etal. 1971) where intact and regener-ating organisms show specific incorporation of undiffer-entiated cells (neoblasts) whereas there are no signs ofdedifferentiation.

Before cell dedifferentiation in planarians can beruled out, however, a last argument for it can beconsidered. This is to think of blastema and postblas-tema cells appearing after neoblast injection as arisingby dedifferentiation of differentiated cells producedfrom injected neoblasts and not directly from the latter.This argument can also be extended to intact organismsif neoblasts are considered as transient proliferatingcells arising continuously by dedifferentiation fromsome differentiated cell types (e.g. gastrodermal cells).Although this is the strongest argument we can think ofagainst our interpretation, as a clear answer cannot be

Neoblasts in planarian regeneration 85

presently given, we consider it unlikely mainly becauseinjected differentiated cells, despite being fully func-tional, never produced any undifferentiated mitoticcells.

We are fully aware, however, that a definite proof ofthe exclusive (or even a main) role of neoblasts inplanarians must be based on marking differentially, andpermanently, neoblasts and differentiated cells andtracing, in both intact and regenerating organisms, thelineage of these cells. Since freshwater triclads seem tobe impervious to exogenous labelled DNA precursors,other labelling methods such as neoblast-specific mono-clonal antibodies and introduction of gene markers intoneoblasts by retrovirus infection (see Price, 1987, forgeneral references) are presently being tried (Romero,Burgaya, Bueno, Sumoy & Baguna, work in progress).

General implicationsOur results also have two interesting afterthoughts.First, they argue strongly against the long-standingtradition of considering planarian neoblasts as cellsmainly 'reserved' for regeneration (Wolff, 1962; Slack,1980). Since most planarian species reproduce sexuallyand very rarely regenerate in nature, it is hardlyunderstandable, either on economic or evolutionarygrounds, to maintain a population of undifferentiated'reserve cells', amounting to 25-30% of total cells(Bagufia & Romero, 1981), unless they serve a moreimportant function in the adult: to be the stem cell of all(or most) differentiated cell types (Bagufia, 1981;Lange, 1983), as has also been demonstrated in othergroups of Turbellarians (Palmberg & Reuter, 1983;Drobysheva, 1986, 1988) and Cestoda (Wikgren et al.1971). Besides, in most animal groups the so-called'reserve cells' have been found to be either nonexistent(witness the so-called 'neoblasts' in several groups ofAnnelida, mainly Polychaetae; Hill, 1970) or to be thesource of several (e.g. the interstitial cells in hydra;David & Gierer, 1974) or unique (e.g. satellite cells ofmuscle cells in Vertebrata; Cameron et al. 1986) differ-entiated cell types during daily tissue renewal.

Second, and to us more importantly, our results seemto contradict the recent trend to extrapolate to allanimal groups, including planarians, the phenomenonof cell dedifferentiation as a unique mechanism torecruit undifferentiated blastema cells (Slack, 1980). Inour view, this trend results from overlooking thepresent diversity of modalities of regeneration (witnessthe morphallactic processes in some lower groups likeCoelenterata (Cummings & Bode, 1984) and lowerTurbellarians (Palmer & Reuter, 1983), as well as theparticular role of some specific cell types, like satellitecells, as a source of regenerative myoblasts in someVertebrates (Cameron et al. 1986)) and the possiblerelationship between the actual mechanism of regener-ation and the tissular complexity of the species. Indeed,it has been argued (Baguna, 1981) that the planarianneoblast system, based on a unique and totipotentialself-renewing stem cell present everywhere, thoughappropiate to the low level tissular complexity ofplanarians, is inadequate for epimorphic regenerating

organisms like Insecta or Amphibia with static tissueswholly made of non-renewing terminal differentiatedcells and renewing tissues with different determinedstem cells placed in different body regions. Therefore,the dedifferentiation process, necessary for the latter, isnot necessary for organisms like planarians (as well asfor hydras) in a total and continuous state of rapid cellturnover (Bagufia & Romero, 1981; Lange, 1983;Romero, 1987).

Although the present day modalities of regeneration(morphallaxis, epimorphosis, and mixed types combin-ing aspects of both) and the diverse mechanisms torecruit undifferentiated blastema cells may be thoughtto be connected to the structural complexity (coarselymeasured in number of cell types and their organizationinto definite tissues and organs; Kauffman, 1969) andthe phyletic position of these animal groups (Field et al.1988) via the extant mechanisms of cell renewal in theadult, it is nonetheless highly probable that under theseformally diverse mechanisms a general, unique, moda-lity of pattern restoration during regeneration, basedmainly on short-range cell-cell interactions and wherecell origins will be less important, will be uncovered.

We would like to thank Professor Peter Lawrence and twoanonymous referees for improving the first draft of themanuscript. This work was supported by grants from theUniversity of Barcelona and Comisi6n Asesora de Investiga-ci6n Cientffica y Te"cnica (CAICYT) to JB.

References

BAGUNA, J. (1975). Esudios citotaxon6micos, ecol6gicos, ehistofisiologia de la regulaci6n morfogene'tica durante elcrecimiento y la regeneraci6n de la raza asexuada de la planariaDugesia mediterranea n.sp. (Turbellaria, Tricladida, Paludicola).Ph.D. Thesis. Univ. de Barcelona.

BAGUNA, J. (1981). Planarian neoblasts. Nature, Lond. 290, 14-15.BAGUNA, J. & ROMERO, R. (1981). Quantitative analysis of cell

types during growth, degrowth and regeneration in the planarianDugesia mediterranea and Dugesia tigrina. Hydrobiologia 84,181-194.

BETCHAKU, T. (1970). The cellular mechanism of the formation of aregeneration blastema of freshwater planaria Dugesiadorotocephala. I. The behaviour of cells in a tiny body fragmentisolated in vitro. J. exp. Zool. 164, 407-434.

BOWEN, I. D. , DEN HOLLANDER, J. E. & LEWIS, G. H. J. (1982).Cell death and acid phosphatase activity in the regenerationplanarian Polycelis tenuis Iijima. Differentiation 21, 160-167.

BR0NSTED, H. V. (1969). Planarian Regeneration. London:Pergamon Press.

CAMERON, J. A., HILGERS, A. R. & HINTERBERGER, T. J. (1986).Evidence that reserve cells are a source of regenerated adultnewt muscle in vitro. Nature, Lond. 321, 607-610.

CHANDEBOIS, R. (1976). Histogenesis and Morphogenesis inPlanarian Regeneration. Basel: S. Karger.

COLLET, J., AULADELL, M. C. & BAGUNA, J. (1984). Specificprotein markers of neoblasts and differentiated cells in theplanarian Dugesia(G)tigrina J. Embryol. exp. Morph. 82Supplement, 246.

COWARD, S. J. & HAY, E. D. (1972). Fine structure of cellsinvolved in physiological and reconstitutive regeneration inplanarians. Anat. Rec. 172, 296.

COWARD, S. J., HIRSH, F. M. & TAYLOR, J. H. (1970). Thymidinekinase activity during regeneration in the. planarian Dugesiadorotocephala. J. exp. Zool. 173, 269-277.

CUMMINGS, S. G. & BODE, H. R. (1984). Head regeneration and

86 /. Baguna, E. Said and C. Auladell

polarity reversal in Hydra attenuata can occur in absence of DNAsynthesis. Wilhelm Roux's. Arch, devl Biol. 194, 79-86.

DAVID, C. N. & GIERER, A. (1974). Cell cycle kinetics anddevelopment of Hydra attenuata. III. Nerve and nematocytedifferentiation. J. Cell Sci. 16, 359-375.

DROBYSHEVA, I. M. (1986). Physiological regeneration of thedigestive parenchyma in Convoluta convoluta and Oxyposthiapraedator (Turbellaria, Acoela). Hydrobiologia 132, 189-193.

DROBYSHEVA, I. M. (1988). An autoradiographic study of thereplacement of epidermis in polyclad turbellarians. Fortschr.Zool. 36, 97-101.

DUBOIS, F. (1949). Contribution a l'etude de la regeneration chezplanaires dulcicoles. Bull. Biol. Fr. Belg. 83, 213-288.

FEDECKA-BRUNER, B. (1967). Etudes sur la re'ge'ne'ration des organsg6nitaux chez la planaire Dugesia lugubris. I. Regeneration destesticules apres destruction. Bull. Biol. Fr. Belg. 101, 255-319.

FIELD, K. G., OLSEN, G. J., LANE, D. J., GIOVANNONI, S. J.,GHISELIN, M. T., RAFF, E. C , PACE, N. R. & RAFF, R. A.(1988). Molecular phylogeny of the Animal Kingdom. Science239, 748-753.

FLICKINGER, R. A. (1964). lsotopic evidence for a local origin ofblastema cells in regenerating planaria. Expl Cell Res. 34,403-406.

GABRIEL, A. (1970). Etude morphologique et evolutionbiochimique des ndoblastes au cours des premieres phases de laregeneration des planaires d'eau douce. Ann. Embryol etMorphogen. 3, 49-69.

GREMIGNI, V. & MICELI, C. (1980). Cytophotometric evidence forcell 'transdifferentiation' in planarian regeneration. WilhelmRoux's. Arch, devl Biol. 188, 107-113.

GREMIGNI, V., MICELI, C. & PICANO, E. (1980). On the role ofgerm cells in planarian regeneration. II. Cytophotometricanalysis of the nuclear Feulgen-DNA content in cells ofregenerated somatic tissues. J. Embryol. exp. Morph. 55, 65-76.

GREMIGNI, V., NIGRO, M. & PUCCINELLI, I. (1982). Evidence ofmale germ cell redifferentiation into female germ cells inplanarian regeneration. J. Embryol. exp. Morph. 70, 29-36.

HAY, E. D. (1966). Regeneration. New York: Holt, Rinehart andWinston.

HILL, S. D. (1970). Origin of the regeneration blastema inpolychaete annelids. Am. Zool. 10, 101-112.

HORI, I. (1983). Differentiation of myoblasts in the regeneratingplanarian Dugesia japonica. Cell. Differ. 12, 155-163.

HORI, I. (1986). Ultrastructural observations on gastrodermalregeneration in the planarian Dugesia japonica (Turbellaria).Hydrobiologia 132, 217-222.

KAUFFMAN, S. J. (1969). Metabolic stability and epigenesis inrandomly constructed genetic nets. /. theor. Biol. 22, 3.

LANGE, C. S. (1968). Studies on the cellular basis of radiationlethality. I. The pattern of mortality in whole-body irradiatedplanarian (Tricladida, Paludicola). Int. J. Radiat. Biol. 13,511-530.

LANGE, C. S. (1985). Stem cells in planarians. In Stem Cells (ed. C.S. Potten), pp. 29-66. Edinburgh: Churchill Livingstone.

LANGE, C. S. & GILBERT, C. W. (1968). Studies on the cellularbasis of radiation lethality. III. The measurement of stem-cellrepopulation probability. Int. J. Radiat. Biol. 14, 175-388.

LENDER, TH. (1962). Factors in morphogenesis of regenerating

fresh-water planaria. In Advances in Morphogenesis. Vol 2. (ed.M. Abercrombie & J. Brachet), pp. 305-331. London: AcademicPress.

LENDER, TH. & GABRIEL, A. (1965). Les n6oblastes marque's parl'uridine tritie'e migrent et edifient le blasteme de regenerationdes planaires d'eau douce. C. r. hebd. Static. Acad. Sci., Paris.260, 4095-4097.

MORITA, M. & BEST, J. B. (1984). Electron microscopic studies ofplanarian regeneration. IV. Cell division of neoblasts in Dugesiadorotocephala. J. exp. Zool. 229, 425-436.

PALMERG, I. & REUTER, M. (1983). Asexual reproduction inMicrostomum lineare (Turbellaria). I. An autoradiographic andultrastructural study. Int. J. Invert. Reprod. 6 197-206.

PRICE, J. (1987). Retrovirus and the study of cell lineage.Development 101, 409-419.

ROMERO, R. (1987). Analisi cellular quantitativa del creixement i dela reproducci6 a diferentes especies de planaries. Ph.D. Thesis.Univ. de Barcelona.

ROSE, C. & SHOSTAK, S. (1968). The transformation ofgastrodermal cells to neoblasts in regenerating Phagocata gracilis(Leidy). Expl Cell Res. 50, 553-561.

SAL6, E. (1984). Formaci6 del blastema i re-especificaci6 del patr6durant la regeneraci6 de les planaries Dugesia(S)mediterranea iDugesia(G)tigrina: analisi morfolbgica, cellular i bioquimica. Ph.D. Thesis. Univ. Barcelona.

SAL6, E. & BAGUNA, J. (1984). Regeneration and patternformation in planarians. I. The pattern of mitosis in anterior andposterior regeneration in Dugesia(C)tigrina, and a new proposalfor blastema formation. J. Embryol. exp. Morph. 83, 63-80.

SAL6, E. & BAGUNA, J. (1986). Cell movement in intact andregenerating planarians. Quantitation using chromosomal,nuclear and cytoplasmic markers. J. Embryol. exp. Morph. 89,57-70.

SAL6, E. & BAGUNA, J. (1986). Stimulation of cellular proliferationand differentiation in the intact and regenerating planarianDugesia(G)tigrina by the neuropeptide Substance P. J. exp.Zool. 237, 129-135.

SLACK, J. M. W. (1980). The source of cells for regeneration.Nature, Lond. 288, 760.

TESHIROGI, W. (1967). Heteroplastic transplantation in thefreshwater planarians Bdellocephala brunnea andDendrocoelopsis lacteus. Sci. Rep. Tdhoku. Univ. Ser IV (Biol).33, 383-397.

WIKGREN, B. J. P., GUSTAFSSON, M. K. S. & KNUTS, G. M. (1971).Primary anlage formation in dyphyllobothriid tapeworms. Z.Parasitenk. 36, 131-139.

WOLFF, E. (1962). Recent researches on the regeneration ofplanaria. In Regeneration (ed. D. Rudnick), pp. 53-84. NewYork: The Ronald Press.

WOLFF, E. & DUBOIS, F. (1948). Sur la migration des cellules deregeneration chez les planaires. Rev. suisse Zool. 55, 218-227.

WOODRUFF, L. S. & BURNETT, A. L. (1965). The origin of blastemalcells in Dugesia tigrina. Expl Cell. Res. 38, 295-305.

{Accepted 9 June 1989)