Ontogenetic and Phylogenetic Development of the Endocrine

R

OE

JDS

A

TspoaipoTivcatFdbmdpbmgicado

General and Comparative Endocrinology 116, 303–335 (1999)Article ID gcen.1999.7376, available online at http://www.idealibrary.com on

0CA

EVIEW

ntogenetic and Phylogenetic Development of thendocrine Pancreas (Islet Organ) in Fishes

ohn H. Youson and Azza A. Al-Mahroukiepartment of Zoology and Division of Life Sciences, University of Toronto at Scarborough,carborough, Ontario M1C 1A4 Canada

ccepted August 25, 1999

rtbaawmtStopfattcdmctr

fmo

he morphology of the gastroenteropancreatic (GEP)ystem of fishes was reviewed with the objective ofroviding the phylogenetic and ontogenetic developmentf the system in this vertebrate group, which includesgnathans and gnathostome cartilaginous, actinoptyeryg-an, and sarcopterygian fishes. Particular emphasis islaced on the fish homolog of the endocrine pancreas ofther vertebrates, which is referred to as the islet organ.he one-hormone islet organ (B cells) of larval lampreys

s the most basic pattern seen among a free-livingertebrate, with the two-hormone islet organ (B and Dells) of hagfish and the three-hormone islet organ (B, D,nd F cells) of adult lampreys implying a phylogeneticrend toward the classic four-hormone islet tissue (B, D,, and A cells) in most other fishes. An earlier stage in theevelopment of this phylogenetic sequence in verte-rates may have been the restriction of islet-type hor-ones to the alimentary canal, like that seen in protochor-ates. The relationship of the islet organ to exocrineancreatic tissue, or its equivalent, is variable amongony, cartilaginous, and agnathan fishes and is likely aanifestation of the early divergence of these piscine

roups. Variations in pancreatic morphology betweenndividuals of subgroups within both the lamprey andhondrichthyan taxa are consistent with their evolution-ry distance. A comparison of the distribution andegree of concentration of the components of the islet
rgan among teleosts indicates a diffuse distribution of t
303016-6480/99 $30.00opyright r 1999 by Academic Pressll rights of reproduction in any form reserved.

elatively small islets in the generalized euteleosts andhe tendency for the concentration into Brockmannodies of large (principal) islets (with or without second-ry islets) in the more derived forms. The holosteanctinopterygians (Amiiformes and Semiontiformes) shareith the basal teleosts (osteoglossomorphs, elopo-orphs) the diffuse arrangement of the components of

he islet organ that is seen in generalized euteleosts.ince principal islets are also present in adult lampreyshe question arises whether principal islets are a derivedr a generalized feature among teleosts. There is aaucity of studies on the ontogeny of the GEP system inish but it has been noted that the timing of theppearance of the islet cell types parallels the time thathey appear during phylogeny; the theory of recapitula-ion has been revisited. It is stressed that the lamprey lifeycle provides a good opportunity for studying theevelopment of the GEP system. There are now severalarkers of cell differentiation in the mammalian endo-

rine pancreas which would be useful for investigatinghe development of the islet organ and cells of theemaining GEP system in fish. r 1999 Academic Press

Scientific interest in the distribution, structure, andunction of endocrine tissues associated with the ali-

entary canal (gastroenteropancreatic, GEP, system)f fishes can be traced back at least to the beginning of
he 19th century. This interest has heightened as we

anfii(keaaGtSimcab(1sdtpbm(Mfpi(

bfiwt111titupti

asecHtnaevtg

D

sdestctTa(cprirt(vliatfwoiTfii

304 Youson and Al-Mahrouki

CA

pproach the end of the 20th century. Thus, throughearly 200 years of investigation there is still a dualascination with the importance of the GEP tissue tondividual fish species and to how the tissue in fishesn general relates to that found in higher vertebratesFalkmer, 1995). With commercially valuable species,nowledge of this tissue is essential, for hormoneslaborated by the GEP system are critical in intermedi-ry metabolism, which in turn is crucial for fish growthnd eventual maturity (Plisetskaya, 1990a,b); someEP hormones have direct effects even on these latter

wo parameters (Plisetskaya and Mommsen, 1996).ince the hormones of the GEP system have equallymportant roles in higher vertebrates, including hu-

ans, the origins of the cells of the GEP system in theontext of where and how they develop (ontogenesis)nd when they appeared during the history of verte-rates (phylogenesis) have been continuous curiositiesBonner-Weir and Weir, 1979; Epple and Brinn, 1986,987). With the refinement of techniques to isolate andequence peptides or to clone cDNAs, the past 2ecades have seen an explosion in the information on

he molecular structure of fish GEP peptides. Therimary structures of these peptides in fishes haveeen crucial for analyses which attempt to explain theolecular evolution of GEP peptides of vertebrates

e.g., Conlon, 1995; Larhammar, 1996; Plisetskaya andommsen, 1996). The bioactivity of isolated peptides

rom fishes, relative to that of higher animals, hasrovided examples of both a suspected modern special-

zation and a regression of function of GEP hormonesFalkmer, 1995).

The above paragraph emphasizes the value that cane placed on investigations into the GEP system ofshes. This view is certainly not novel, for manyell-known comparative endocrinologists have illus-

rated this point in extensive, earlier reviews (Epple,969; Epple and Lewis, 1973; Epple and Brinn, 1986,987; Falkmer, 1985a,b, 1995; Plisetskaya, 1989a,b,990a,b). In many ways, this review is dedicated tohese pioneers and some of their views will be reexam-ned and placed in the context of more recent litera-ure. It is the objective of this report to provide anpdate of the literature on the GEP system as itertains to the topics of ontogeny and phylogeny of

his system in fish. As it will soon be demonstrated, it
s difficult to separate the endocrine cells within the l
opyright r 1999 by Academic Pressll rights of reproduction in any form reserved.

limentary canal (stomach and intestine) with cells ofimilar type, or elaborating similar hormones, in thendocrine pancreas. The endocrine cells of these threeomponents have an intimate ontogenetic history.owever, the primary focus of this update will be on

he fish homolog of the endocrine pancreas present inonpiscine vertebrates. Furthermore, the focus will becomparison of the distribution, structure, and ontog-

ny of the endocrine pancreatic tissue through thearious groups of fishes, rather than on the compara-ive function and structure of the peptides that theyenerate.

EFINITION OF TERMS

Ontogeny is development of a single individual, or aystem within the individual, from the fertilized egg toeath (Smith, 1960), i.e., a total life history includingmbryonic and postnatal (Gould, 1977). This termhould be distinguished from phylogeny, which is aype of development involving modification of a spe-ies or a group of species, i.e., ‘‘the family history’’ orhe evolutionary history of a lineage (Gould, 1977).his latter type of development is characterized byccompanying changes in structure and functionSmith, 1960). It is the evolutionary trend, or morpho-line (Hildebrand, 1995), of the GEP system of fishes, aarticular phyletic line, that is being considered in thiseview. The word fish has different connotations so it ismportant to establish the terms of reference for thiseview. The interpretation of Nelson (1994) is adopted,hat is, the term ‘‘fish’’ includes both the jawlessAgnatha) and the jawed (Gnathostomata) aquaticertebrates that ‘‘ . . . have gills throughout life and

imbs, if any, in the form of fins.’’ This broad definitionncludes the agnathans, the hagfishes and lampreys,nd the two large extant grades of aquatic gnathos-ome fishes, Chondrichthimorphi and Teleostomi. Theormer grade consists of the class Chondrichthyes,

hich possesses the subclasses Holocephali (singlerder Chimaeriformes) and Elasmobranchi (contain-

ng many orders under superorder Euselachii). Gradeeleostomi contains the classes Sarcopterygii (lobe-nned fishes) and Actinopterygii (ray-finned fishes). It

s the former, and certainly not the latter, class which
ead to the tetrapods. Thus in the context of vertebrate

etmgotNtfafHpprTimsabot(sestsofsaDT

acccmsdttac

atiagivwr

ftBfiiwonRass1twmateaam(sta‘b(tsfspc

Fish Endocrine Pancreas 305

volution, the actinopterygians, and particularly theeleosts, are a highly specialized group which some

ight consider led to a dead end. Among the piscineroups are one-half the approximately 50,000 speciesf living vertebrates and these fishes demonstrateremendous diversity in their form and distribution.

eedless to say, this report cannot take advantage ofhis diversity and likely will be concerned with onlyewer than 1.0% of the living fishes. Despite this,ttempts will be made to provide data on examplesrom as many of the 57 orders of fishes as possible.owever, since a central theme of this review ishylogenetic development and the authors have arime interest in fishes of ancient lineage, particularegard will be given to some of the more ancient orders.he term primitive is used rather sparingly because it

mplies a lack of specialization and some of the ancientembers to be discussed are highly specialized, de-

pite their ancient origins. As seems to be the practicemong fish taxonomists of the day, the terms lower,asal, and generalized reflect close to the starting placef a character or a group and derived implies a charac-er or a group that has been taken from anotherearlier) source. The term gastroenteropancreatic (GEP)ystem is of common use and it has already beenmployed above in the Introduction to emphasize aystem of endocrine cells involving the stomach, intes-ine, and pancreas. The term enteropancreatic (EP)ystem has recently been applied to the endocrine cellsf the intestine and pancreas of lampreys and hagfish,or both of the these agnathans lack a stomach (You-on, 1999). The EP system would seem more appropri-te for other fishes without a stomach (see: Noaillac-epeyre and Hollande, 1981; Rombout and Taverne-hiele, 1982).GEP and EP are considered by some not to be

ll-inclusive terms. Some of the peptides secreted byells of the GEP system are also secreted by cells of theentral nervous system. In fact, Falkmer and hisolleagues have continuously emphasized the develop-ental relationship between the cells of these two

ystems; collectively they are part of a diffuse neuroen-ocrine system (Falkmer, 1985a,b, 1995). To emphasize

his relationship and their common synthetic capacity,he acronym APUD cells, for amine precursor uptakend decarboxylation, has been used for some of the
ells of these two systems (Pearse, 1969). With the fi
pplication of antisera and immunohistochemistry inhe investigations of endocrine cells of the stomach,ntestine, and pancreas of fishes, the use of words suchs argentaffin, argyrophilic, and enterochromaffin isradually disappearing. However, techniques which

dentify any one of these three features in cells are stillaluable in the identity of putative endocrine cellshich show no immunoreactivity to antisera against

ecognized peptides.The endocrine pancreatic homolog in fishes is re-

erred to as the endocrine pancreas, islet tissue, insularissue, islet or insular organ, islets of Langerhans,rockmann bodies, and principal islets. It is surmised

rom reading previous reviews that the term islet organs the endocrine pancreatic homolog of a species,rrespective of its distribution or arrangement, i.e.,

hether the islets which make up the organ are larger small and whether they are concentrated or dissemi-ated (Epple and Brinn, 1986, 1987; Falkmer, 1995).ecently, principal islet has been defined as the largestccumulation of endocrine tissue in the fish pancreas,urrounded by a thin rim of exocrine tissue (Pliset-kaya and Mommsen, 1996). Epple and Brinn (1986,987) have defined Brockmann bodies as large accumula-ions of islet tissue (principal islets) closely associated

ith exocrine pancreatic tissue. However, more com-only, Brockmann bodies and principal islets are used

s synonyms in reference to large bodies of pure isletissue devoid of, or nearly devoid of, strands ofxocrine pancreas (Falkmer and Patent, 1972; Falkmernd Van Noorden, 1983; Falkmer, 1985a,b, 1995; Yangnd Wright, 1995; Maglio and Putti, 1998). The ‘‘Brock-ann body regions’’ from species such as tilapia

Oreochromis niloticus) are under investigation as aource of islet tissue for human transplant because ofhe compacted and pure nature of the ‘‘region’’ (Yangnd Wright, 1995). In a pure sense, in tilapia the

‘region’’ is a Brockmann body or several Brockmannodies and each is composed of some principal isletslarge islets). The Brockmann body(ies) and other isletissue, if present, makes up the islet organ for thispecies. For the purposes of this review, and perhapsor future use, principal islet and Brockmann body areynonyms if the Brockmann body contains a singlerincipal islet. Otherwise, a Brockmann body is aoncentration of pancreatic islets among the viscera of
sh.
Copyright r 1999 by Academic PressAll rights of reproduction in any form reserved.

P

G

tdfiwkib

atspEmotcosageCt(kT

cpafisgidmbdcBw

mpjTbeo

cttclaicacsovirbaoiit(ttisct

C

Geile1(


CA

HYLOGENY OF THE FISH GEP SYSTEM

eneral Morphology of the Teleost Islet Organ

A baseline description of the islet organ in a represen-ative fish species is an essential requirement for theiscussion of the topic of ontogeny and phylogeny ofshes. The species we have chosen is a teleost, thehite sucker (Catastomus commersoni), and to our

nowledge there has not been a detailed description ofts GEP system. McCormick (1925) provided only arief description of this species in his monograph.Teleostei is a division of the subclass Neopterygii

mong actinopterygians. Recent Teleostei are believedo be monophyletic and the division consists of themall subdivisions Osteoglossomorpha, Elopomor-ha, and Clupeomorpha and the large subdivisionuteleostei (De Pinna, 1996). The white sucker is aember of the order Cypriniformes and of the super-

rder Ostariophysi within the largest subdivision (Eu-eleostei) of teleosts (also see below for alternatelassification of ostarioiphysians). Cyprinids rank sec-nd only to the order Perciformes in total number ofpecies; there are 2662 species of cyprinids, all of whichre freshwater (Nelson, 1994). Epple and Brinn (1975)ive only brief reference to the distribution of thexocrine pancreas and the islet size in two species ofypriniformes. More extensive reports of the islet

issue of cyprinids are those on carp and goldfishKobayashi and Takahashi, 1970; Nakamura and Yo-ote, 1971; Rombout et al., 1979; Rombout and Taverne-hiele, 1982).Epple and Brinn (1975) described five morphologi-

al relationships between the exocrine and endocrineancreatic tissues (islet organ) of vertebrates andmong these was the actinopyterygian (‘‘ray-finnedshes’’) type (Fig. 1), which includes most bony fishes,uch as the sucker. As might be expected for this largeroup of fishes, there is great structural diversity in the

slet organs but species share the common feature of aiffuse exocrine pancreas of serous acini within theesentery extending between bile ducts, abdominal

lood vessels, the gastrointestinal tract, the gall blad-er, and the liver. Intrahepatic concentrations of exo-rine tissue are common among the teleosts (Epple andrinn, 1986). The islet tissue can be found anywhere
ithin these areas, except in intrahepatic sites, but is w

ost commonly concentrated near where the extrahe-atic common bile duct enters the intestine at the

unction of the stomach, intestine, and intestinal caeca.his association of the islet tissue with extrahepaticile ducts in teleosts is noteworthy and will be refer-nced at a later point in this review, for there is bothntogenetic and phylogenetic significance.The islet tissue is in the form of highly vascularized

ollections of epithelial cells, usually separated fromhe exocrine acini by a small amount of fibrous connec-ive tissue but generally not confined within a trueapsule (Fig. 2a). The epithelial aggregates (islets) stainightly with hematoxylin and eosin compared to thecinar cells, which have large acidophilic granules. Theslets are of variable size but the cells are arranged asords or small lobules. In the white sucker, islets wereggregated near the gall bladder and some were ofonsiderable size, i.e., principal islets, but among smalltrands of exocrine acini and pancreatic ducts andther smaller (secondary) islets. One principal islet,isible to the naked eye, was located within a concav-

ty of the liver and was surrounded by only a narrowim of exocrine acini (Fig. 2a). The term Brockmannody can be used for both this isolated principal isletnd the larger mass of islet tissue located more posteri-rly. All islets of the white sucker immunostain for

nsulin using anti-mammalian insulin serum, but themmunoreaction of these B cells was weak comparedo that of positive controls. The B cells and D cellsimmunoreactive with anti-somatostatin-25) were dis-ributed throughout the islet. A cells were immunoreac-ive with anti-glucagon and were primarily localizedn the periphery of islets, but with smaller numberscattered throughout more central regions (Fig. 2b). Fells (immunoreactive with anti-PYY) were noted inhe same peripheral locations as some of the A cells.

omparative Morphology of the Euteleost Islet Organ

There have been many studies of the structure of theEP system in euteleosts. The reader is referred to the

xtensive series of reports on GEP cells of the stomach,ntestine, and islet tissue of the sea bass, Dicentrarchusabrax (Carrillo et al., 1986; Beccaria et al., 1990; Lozanot al., 1991a,b; Agulleiro et al., 1993; Gomez-Viscus et al.,996, 1998), and the gilt-head sea bream, Sparus aurataAbad et al., 1986, 1987, 1988, 1992; Elbal et al., 1991),

here reviews of earlier literature are provided. It is

icapqisi1

o(Udtis(t

F

ee(abajTmp


mportant to highlight several other earlier morphologi-al studies of islet tissue from euteleosts, for they aressociated with the isolation of peptides and theroduction of antisera which greatly influenced subse-uent directions of research on regulatory peptides

n fish. These are investigations of islet tissue inalmonids (Wang et al., 1986; Nozaki et al., 1988a,b) andn the anglerfish, Lophius americanus (Johnson et al.,976).

IG. 1. Diagrammatic representation of the distribution of the endosophagus (O), stomach (S), intestine (unlabeled portion of gut), liverpithelial cells of the digestive tube are enlarged relative to the cellshatched) cells. Represented are a protochordate (a), a larval lampreydult lamprey: Northern Hemisphere (d), an adult lamprey: Southernasal actinopterygian (i), and a derived euteleost (j). The protochordand the hagfish have exocrine and endocrine cells in the intestine

unction in larvae and around the bile duct–intestinal junction in the hhe exocrine tissue and islet organs are intermingled in the remaininore diffuse organ in the two bony fishes (i, j); islets are small and

rincipal islet in j.

The most extensive comparative analysis of the islet h

rgan of euteleosts is undoubtedly that of McCormick1925), a member of the Department of Physiology,niversity of Toronto, during the time that insulin wasiscovered in this department. He examined the posi-

ion, size, and number of islets and the extent ofnvasion by acinar tissue and connective tissue in 70pecies of what today would be classified as euteleostsNelson, 1994). McCormick (1925) noted that the isletissue among the species was of variable types, which

issue (islet organ) in fishes relative to the exocrine pancreatic tissue,ll bladder (G), and bile duct (tube leading from the gall bladder). Ther structures to denote the presence of endocrine (dark) and exocrineern Hemisphere (b), a larval lamprey: Southern Hemisphere (c), an

sphere (e), a hagfish (f), an holocephalian (g), an elasmobranch (h), ano pancreas but enteroendocrine cells and larval and adult lampreysslet organ is seen as scattered follicles at the esophagus–intestinal. Islet organs are one (e) or two (d) principal islets in adult lampreys.ies, with a compact organ in the two cartilaginous fishes (g, h) and ae in i but more concentrated (Brockmann body) and with a cranial

crine t(L), ga

of othe: NorthHemi

te has. The i

agfishg specdiffus

e attempted to relate to the fish taxonomy of the day.



CA


H(SitoohmtostrtotaF1ocqpmp

nlttoTtCimp(

cM(t(((staTehooswigmewpdemOmtrt1emo

oai

F

s3

(iaS


is general conclusion was that the LophobranchiiSeries Percomorpha, Order Gasterosteiformes, Familyyngnathidae:pipefishes) showed the first principalslet which was free from acinar tissue. He consideredhis to be an advanced feature, for most of the ‘‘earlierrders’’ had an ‘‘intermediate type,’’ where islets weref variable size among extensive exocrine tissue. Sincee also noted consistencies of structure within com-on groups of teleosts, McCormick (1925) implied that

here may be some relationship between the morphol-gy of islet tissue and the taxonomic position. Thistudy of nearly 75 years ago also examined the isletissue of more basal teleosts and nonteleost actinopte-ygians and thus can be considered the first investiga-ion of the phylogenetic development of the fish isletrgan. Subsequent to the report of McCormick (1925),here was a study of equal magnitude by Siwe (1926)nd then many reviews appeared (e.g., Epple, 1969;alkmer and Patent, 1972; Epple and Brinn, 1975, 1986,987; Falkmer, 1985a,b, 1995) where phylogenetic devel-pment of fish islet tissue was considered in the widerontext of the vertebrate endocrine pancreas. Theuestion arises whether we now have sufficient data torovide a clear picture of whether the comparativeorphology of the islet organ of euteleosts reflects any

hylogenetic development of this system?There are 32 orders of euteleosts and many have

umerous suborders, superfamilies, families, subfami-ies, genera, and subgenera (Nelson, 1994). A review ofhe literature since McCormick (1925) indicates thathe islet tissue has been examined in representatives ofnly 14 of these 32 orders and 24 of the 391 families.hese are Batrachoidiformes (family Batrachoidae,

oadfishes), Characiformes (family Characidae, tetras),yprinodontiformes (families Fundulidae and Poecili-

dae, killifishes and poecilids, respectively), Cyprinifor-es (families Cyprinidae and Gyrinocheilidae, cy-

rinids and algae eaters, respectively), Esociformesfamily Esocidae, pikes), Gadiformes (family Gadidae,

IG. 2. Islet tissue among exocrine acini (arrows). (a) A principalurrounded by a narrow rim of exocrine acini and is located within a150. (b) Portion of a large islet from the white sucker showing imm

c) A small islet among exocrine acini from the bowfin, Amia calva, snsulin; original magnification; 3300. (d) Intermediate-sized islet fronti-mammalian NPY; original magnification; 3300. Sections in b, c,
ystem, San Francisco, CA). Positive staining is indicated by red coloration
ods), Lophiformes (family Lophiidae, goosefishes),ugiliformes (family Mugilidae, mullets), Perciformes

families Blenniidae, Cichlidae, Gobiidae, Helostoma-odae, Moronidae, and Sparidae), Pleuronectiformesfamily Pleuronectidae, flounders), Salmoniformesfamily Salmonidae, salmonids), Scorpaeniformesfamilies Scorpaenidae and Cottidae, rockfishes andculpins, respectively), Siluriformes (families Callilch-hyidae, Heteropneustidae and Ictaluridae, catfishes),nd Tetradontiformes (family Tetradontidae, puffers).his sample represents just a token of the 22, 262uteleostean species, which have highly divergentabits and ancestral origins (Nelson, 1994). Therefore,ne would expect that little could be achieved in termsf phylogenetic patterns in islet tissue through compari-ons of so few species. However, it is noteworthy that,ithout design, the species that have been studied are

n superorders that are often referenced as moreeneralized (Ostariophysi, Protacanthopterygii) andore derived (Paracanthopterygii, Acanthopterygii)

uteleosts; the latter are two of the seven superordershich are considered neoteleosts. Nevertheless, sam-ling of the islet organs in euteleosts has favored theerived orders (8) as opposed to those that are consid-red more generalized (5). Furthermore, it should beentioned that an alternate view is that orders ofstariophysi mentioned above, namely, Characifor-es, Cypriniformes, and Siluriformes, are not part of

he euteleostean subdivision but instead are moreelated to the subdivision Clupeomorpha, lower on theeleostean taxonomic scale (Lecointre and Nelson,996). Thus if these three are removed from theuteleosts, then only the Esociformes and Salmonifor-es are left to represent characteristics of the islet

rgan of the generalized euteleosteans.The small sample size is just one of many limitations

f comparing morphological features of the islet tissuemong generalized and derived euteleosts using exist-ng data. Two other major limitations are the selection

a Brockmann body of the white sucker, Catastomus commersoni, isity of the liver (L). Stain: hematoxylin–eosin; original magnification;ctivity of A cells to anti-glucagon sera; original magnification; 3300.central aggregate of B cells immunoreactive with anti-mammalian

bowfin Amia calva showing peripheral F cells immunoreactive withwere immunolabeled using rabbit histostain-SP kit (Zymed-Lab-SA

islet inconcavunoreahows am theand d

and the counterstain is hematoxylin.


ooitvelwietGbactactlatmt

ctotHteismsdeetstItp

m

rT(amtpbeEdSaK1iasimaafi1v

mwapat(ett(ecsWtamd


CA

f suitable parameters for comparison and the fact thatne has to assume that there is consistency among

nvestigations in both the type and the thoroughness ofhe observations. The types of observations are quiteariable in recent publications of the GEP cells inuteleosts. Needless to say, all three are highly relevantimitations and there is a need for an investigation

here these variables are eliminated. Given thesenherent problems, are there any morphological param-ters that can be used? The following are parametershat could be used in a thorough comparison of theEP system among several groups of euteleosts: distri-ution of islets; size variation of islets; presence orbsence of principal islets; degree of islet–acinar asso-iation; the number of cell types in the islet; distribu-ion of cell types within the islet; and the presence/bsence and distribution of specific immunoreactiveell types in the gastrointestinal system, e.g., somatosta-in or insulin. At the transmission electron microscopicevel, one could compare among species granule shapend density and the presence/absence of colocaliza-ion of peptides within designated cell types but these

orphological parameters have proved to be inconsis-ent even among closely related fishes (Brinn, 1973).

The point to be emphasized is that under the presentircumstances it is not possible to come to any defini-ive conclusions about the phylogenetic developmentf the endocrine pancreatic homolog (islet organ), andhe GEP system in general, within the euteleosts.

aving stated this, are there any suggested trends inhe phylogenetic development of the GEP system inuteleosts? Langer et al. (1979) concluded from theirmmunohistochemical study of the gastrointestinalystem of 11 species of teleosts that since there is souch variation in immunoreactivity that caution

hould be exercised in generalizing about peptideistribution. This opinion had been expressed mucharlier (Falkmer and Patent, 1972). The more recentxample of anti-insulin immunoreactivity of cells inhe stomach of the sea bass, a derived euteleost, isupport for this variation, which may have no relationo phylogenetic development (Gomez-Visus et al., 1996).n contrast, an examination of the distribution of isletissue among the euteleosts seems to show somehylogenetic relatedness (Epple and Brinn, 1986).McCormick (1925, p. 78) stated in relation to fish islet
orphology that, ‘‘In the gross there is some evident b

elationship to the accepted classification of fishes.’’he more derived euteleosts, which Epple and Brinn

1975) reference as Ctenosquamata, have a substantialggregation of islet tissue in extrapancreatic Brock-ann bodies. The phylogenetic development of islet

issue seems to be toward at least a single, largerincipal islet (in this case also called a Brockmannody) independent or nearly independent of exocrinelements and located near the gall bladder or spleen.xamples of this arrangement are present in theerived orders Cyprinodontiformes, Lophiformes, andcorpaeniformes (Boquist and Patent, 1971; Johnson etl., 1976; Klein and Lange, 1977; Patent et al., 1978;lein and Van Noorden, 1980; Stefan and Falkmer,980). Other derived orders have a single principalslet (Brockmann body) with limited exocrine tissuessociated but also several intermediate and small-ized islets among the exocrine acini (Fig. 1j). Thesenclude Mugiliformes, Perciformes, and Tetradontifor-

es (Kobayashi et al., 1976; Abad et al., 1986; Lozano etl., 1991a,b; Maglio and Putti, 1998). It is noteworthy,nd perhaps not unexpected, that Fundulus variedrom another cyprinodontiform (Xiphophorus) in hav-ng this second type of arrangement (Epple and Brinn,975). This illustrates that inevitably, some intraorderariation will exist.An intermediate step in the phylogenetic develop-ent of the euteleost principal islet may be foundithin order Salmoniformes. In both rainbow trout

nd coho salmon there is not a principal islet butancreatic islets, some of the size of principal islets,ggregated in a mass (a Brockmann body?) alongsidehe gall bladder and surrounded by exocrine aciniWagner and McKeown, 1981; Wang et al., 1986; Nozakit al., 1988a,b). Islets are also diffusely distributed inhe adipose tissue that surrounds the pyloric caeca andhe spleen. The mass of islets is called an islet organWagner and McKeown, 1981) or principal body (Wangt al., 1986), with the former described in the trout asonsisting of many islets that fused together with aurrounding capsule of exocrine pancreatic tissue.agner and McKeown (1981) noted that in small trout

he mass is more compact and surrounded by exocrinecini, but in larger fish exocrine tissue invaded theass and divided it into small groups of islets. It is not

ifficult to visualize this arrangement as eventually
eing replaced in more derived euteleosts by a com-

pbOpimodtimobtaosp1adsgoa

Dicandcr(Dhte

E

pmtp

esbsAotoithatmlBsipFt(sit

witdltm1iYaieebdvgaa


act islet organ of a principal islet(s), i.e., a Brockmannody, nearly or totally devoid of exocrine elements.ther orders, such as Cypriniformes and Siluriformes,rovide some support for a more basal arrangement of

slet tissue in the more generalized euteleosts, i.e., aore diffuse arrangement of islet tissue in the islet

rgan. In the cyprinid Carassius, numerous well-efined islets of varying size are scattered throughout

he exocrine pancreas. However, macroscopically vis-ble islets (principal islets?) are also found in the

esentery (Kobayashi and Takahashi, 1970). More-ver, in Barbus conchonius a principal islet is locatedetween the gall bladder and the intestinal bulb, andhere are many small islets more caudally (Rombout etl., 1979). This description would also fit the islet organf the white sucker described herein. Although theilurids (catfishes) have been described as having arincipal islet located close to the gallbladder (Brinn,975; Johnson et al., 1976), they are referenced by Epplend Brinn (1975) as having an intermediate isletistribution like the cyprinids, that is, also extensivemaller islets within the exocrine pancreas. In theseroups, although there may be some large islets whichne may wish to refer to as principal islets, they do notll possess a Brockmann body.The distribution of the four principal cell types (A, B,, F) within the islets of euteleosts has been described

n many species. The most consistent pattern is aentral core of B cells, a diffuse arrangement of D cells,nd A and F cells at the periphery of each islet. It isoteworthy that the white sucker showed a broadistribution of the anti-glucagon immunoreactive Aells (Fig. 2b). The distribution of D cells has particularelevance to their modulation of A and B cell secretionMaglio and Putti, 1998). Two types of D cells, D1 and2, are present in islets of many euteleosts and theyave a specific distribution and a specific immunoreac-

ivity to various types of somatostatin antisera (Nozakit al., 1988a; Abad et al., 1992).

ndocrine Pancreas of Lower Teleostei

For further analysis of the possible existence of somehylogenetic pattern of islet distribution and arrange-ent among teleosts it is possible to include data from

wo of the three other subdivisions: Osteoglossomor-
ha, Elopomorpha, and Clupeomorpha. To our knowl- (
dge there is no collective name for these three subdivi-ions of Teleostei but they are generally accepted aseing the most generalized or low on the taxonomiccale of teleosts. Among the Elopomorpha is the ordernguilliformes, which includes the eels. The pancreasf the eel is compact and has been described asetrapod-like in also having islets scattered through-ut. The general view is that eels do not have principal

slets or aggregates of islets (Brockmann bodies) likehose of more derived teleosts (Kobayashi and Taka-ashi, 1974; Epple and Brinn, 1975). However, inddition to scattered small islets in the head and tail ofhe pancreas, L’Hermite et al. (1985) describe in the

iddle of the pancreas two voluminous islets ineptocephali and a single islet, which they call therockmann body, in the glass eel and in the adulttages. These latter data imply that there is a principalslet in adult eel and that it is formed during metamor-hosis from the fusion of two smaller principal islets.urthermore, large islets become more conspicuous inhe pancreas of adult Anguilla rostrata during lipopexiaA. Epple, personal communication). Thus there isome controversy about the structure of the islet organn eels; however, the general consensus is that isletissue is not organized into Brockmann bodies.

The scattered distribution of islets of variable sizeithout any large islet masses (i.e., a Brockmann body)

s well documented in one of the most basal of extanteleosts, the Osteoglossomorpha (bonytongues). Theistribution of this geologically ancient and monophy-

ectic teleost group was likely influenced by shifts inectonic plates after the Jurassic period and today

embers are found on most continents (Li and Wilson,996). In general, the distribution of islet tissue in theslet organ of osteoglossomorphs (Al-Mahrouki andouson, 1998) is reflective of their basal positionmong the teleosts, that is, no distinct aggregation ofslets but a diffuse dissemination of islet tissue amongxocrine acini (Fig. 1i). This recent study serves as anxample of how morphological features, such as distri-ution and arrangement of islet tissue and types andistribution of gut endocrine cells, can be of somealue in taxonomic characterization even withinroups. There were similarities in islet distribution andrrangement in species within and between familiesnd distinct differences between orders and suborders
Al-Mahrouki and Youson, 1998). One species, Pantodon,

wstaoocotlq

I

tnPtrPtstrwi1oi(sgtrpcmect1

PBtn

ttrattMPBescbagafdtcdw

pgripsonbta1a1tsocsItaH


CA

hich is the most distant of the five species studied,howed the most extreme variation in islet distribu-ion. This species also showed anti-insulin immunore-ctivity in both the stomach and intestine, whereasthers had either no immunoreactivity in either organr immunoreactivity in only the intestine. It wasoncluded that variations in the GEP system of extantsteoglossomorphs may reflect more divergent evolu-ionary histories among species than is presently be-ieved and/or that interspecific differences are a conse-uence of variations in ontogeny of cells of the system.

slet Organ of Basal Actinopterygii

The artificial grouping basal actinopterygians referso the orders Amiiformes and Semiontiformes of theeopterygians and the orders Acipenseriformes andolypteriformes, often included together as Chondros-

ei (see Nelson, 1994). Reference to ‘‘lower’’ actinopte-ygians is avoided because this usually refers to onlyolypteriformes and Acipenseriformes and we wished

o include amids and semionids in the present discus-ion; the latter are often referenced as ‘‘higher’’ actinop-erygians (Gardiner, 1993). It is emphasized that theelationship of these four orders with one another andith the other major actinopterygian group (Teleostei)

s still a matter of great controversy (Gardiner et al.,996; Grande and Bemis, 1996). The name Holostei wasriginally used for a group of actinopterygians which

ncluded Amia calva (the bowfin), Lepisosteus spp.gars), and Polypterus (bichirs). Parsimony analysishows that Amia and the gars are a paraphyleticrouping and that the bowfin is more closely related toeleosts than to the gars (Gardiner et al., 1996). Asecently pointed out by Conlon et al. (1998), theosition of the polypterids is uncertain. However, theomparison made by these latter authors on the pri-ary structures of insulin and glucagon from Polypt-

rus with those from the gar and bowfin showed both alose relatedness among the three species and alsoheir basal actinopterygian position (Conlon et al.,998).Published descriptions of the islet organ in both

olypterus and Amia have been cursory (Epple andrinn, 1975). In both species the islet tissue is found

hroughout the widely dispersed exocrine acini, with
o distinct islet accumulation in Amia but in Polypterus c

here is a tendency for aggregation near the junction ofhe bile duct at the intestinal–stomach junction. In theeedfish Calamoichthys, another Polypteridae, the isletsre more widely dispersed but a large islet is present inhe liver; the islets of this species contain four ultrastruc-urally distinct cell types (Epple and Brinn, 1975;

azzi, 1976). To our knowledge, islet histology forolypterus has not been performed. In Amia, Epple andrinn (1975) used histochemical procedures and somelectron microscopy to identify B and D cells, flame-haped cells, and two other unidentified acidophilicells. Immunohistochemistry has revealed that theowfin islets contain cells immunoreactive to antiseragainst insulin, somatostatin, and peptides of both thelucagon and pancreatic polypeptide families (Figs. 2cnd 2d). Therefore, the islets of the bowfin contain theour cell types characteristic of the islets of moreerived Actinopterygii. However, a unique character

o the islets is that one of the cell types containsytoplasmic crystalline inclusions (Fig. 4). A moreetailed description of the GEP system of the bowfinill appear elsewhere.Epple and Brinn (1975) reported that the islets in the

ancreas of the gar, Lepisosteus spp., are more aggre-ated near the bile duct than in the polypterids. Theyeference an ‘‘islet organ’’ as a large aggregation ofslets within the exocrine pancreas near the lowerortion of the bile duct. A more recent study usederial sections and confirmed the highest distributionf islets (‘‘Brockmann body like islet accumulations’’)ear the papilla of the extrahepatic common bile ductut small islets were also distributed over the length ofhis duct, along the gall bladder, and in the mesenterydjacent to the anterior intestine (Groff and Youson,997). These descriptions resemble those given for isletggregates in Salmoniformes (Wagner and McKeown,981; Wang et al., 1986; Nozaki et al., 1988a,b). It isempting to use this single morphological feature touggest some phylogenetic relationship of this Semi-notiformes and salmonids but present taxonomiclassification based on many features, including fos-ils, would not support this view (Gardiner et al., 1996).mmunohistochemistry of Lepisosteus islets indicatedhe presence of A, B, D, and F cells in a typical higherctinopterygian distribution (Groff and Youson, 1997).owever, correlated fine structural and immunocyto-

hemical observations revealed B and D cells and a

tgsYp

(Pt1nilPsaWmpwdtmdtA

I

AatdtnoriGtgAdr

L

(stl1ae(o(tdmpTilvmimeHtapsAWfTtgfaOsoingfitd(


hird cell type, A/F, which colocalized peptides of thelucagon and pancreatic polypeptide families in theame cell and often in the same granules (Groff andouson, 1998). It is uncertain whether there is anyhylogenetic significance to this latter observation.The distribution of pancreatic tissue in Chondrostei

Order Acipenseriformes, families Acipenseridae andolyodontidae, sturgeons and paddlefishes, respec-

ively) needs some specific attention (Epple and Brinn,987). In the sturgeon, Epple and Brinn (1975) foundo Brockmann bodies and ‘‘mammalian-like’’ (small?)

slets spread throughout the pancreas but perhaps aittle more common near the liver hilus. The islets ofolyodon are more scarce and are of smaller averageize than those in the sturgeon, perhaps because thebsence of any distinct connective tissue capsule.eisel (1972) also emphasized the absence of a Brock-ann body in the paddlefish. Therefore, the present

icture of the chondrostean pancreas is a system ofidely dispersed small islets without any great ten-ency for aggregation (as seen in Fig. 1i). Since the

endency is for aggregated islets or large islets (Brock-ann bodies or principal islets, respectively) in more

erived actinopterygians, the most basal islet distribu-ion seems to be present in one of the most basal ofctinopterygii.

slet Organ of the Sarcopterygii

There are three extant groups of sarcopterygians,ctinistia (e.g., Latimeria), Dipnoiformes (lungfishes),

nd Tetrapoda (Cloutier and Ahlberg, 1996). The firstwo are groups of fishes with lobed fins but it is theipnoans which are the living sister group to the

etrapods (Cloutier and Ahlberg, 1996). There has beeno extensive histological examination of the pancreasf the coelacanth, Latimeria. Epple and Brinn (1975)efute the description of ‘‘groupes d’acini pancret-ques’’ by Millot and Anthony (1972) but do agree withrossner (1968) that the pancreas is a compact, extrain-

estinal organ that they refer to as the primitivenathostome type. Epple (1969) described islet cells of, B, and D varieties along the outside of pancreaticucts and thus, the islet topography of Latimeriaesembles that in cartilaginous fishes.

In the dipnoans the pancreas of both Protopterus and
epidosiren is described as in an intraintestinal location e
Epple and Brinn, 1975) and, more explicitly in a latertudy of Protopterus aethiopicus, as located in the wall ofhe cranial portion of the midgut or adjacent to theayers of the tunica muscularis (Scheuermann et al.,991). In Protopterus annectens the pancreas is associ-ted with the more cephalic and dorsal intestine andither between the intestinal serosa and submucosasmall specimens) or also spreading to the submucosaf the spiral folds in large specimens. Epple and Brinn1975) mention a few encapsulated islets in P. annec-ens. In the more recent study, spherical islets ofifferent sizes are described, with the small- andedium-sized ones located in cephalic regions of the

ancreas near exocrine ducts (Tagliafierro et al., 1996).he endocrine portion of the pancreas of P. aethiopicus

s segregated into a large (2 mm diameter), encapsu-ated primary islet organ (Brockmann body) with aariable number of smaller accessory islets (Scheuer-ann et al., 1991). This description of islet arrangement

n Protopterus spp. would fit well with that found in theore derived euteleosts described above. Tagliafierro

t al. (1996) came to the same general conclusion.owever, given the location of the pancreas relative to

hat of euteleosteans, it does seem fitting that Epplend Brinn (1975) should consider the Protopterus islet–ancreas system a unique entity and give it a separatetatus. The descriptions of islet distribution in theustralian lungfish, Neoceratodus forsteri (Rafn andingstrand, 1981; Hansen et al., 1987), seems different

rom that of Protopterus spp. (Tagliafierro et al., 1996).he distribution and size of islets in Neoceratodus are

etrapod like (Rafn and Wingstrand, 1981). Since theseenera are representatives of two different dipnoanamilies, one wonders whether this is a further ex-mple of the utility of islet morphology as a character.n the other hand, one must be careful about general

peculation concerning the phylogenetic developmentf characters when there has been independent special-

zation within Dipnoiformes. The distribution of immu-oreactivity to antisera against insulin, somatostatin,lucagon, and pancreatic polypeptide (PP) in all lung-sh species is similar to that of other fishes, including

he colocalization of glucagon and PP and differentistribution of cell types between large and small islets

Hansen et al., 1987; Scheuermann et al., 1991; Tagliafi-
rro et al., 1996).


CA


SS

trsptTtbnpdi(it(apaAorfawbobcattapc

ebcct(midfittdpt

spsdmCsTp

Gi

cov(l(ap

F

ppdteaf


ummary: Phylogenetic Development of the GEPystem in Bony Fishes

The EP system has been examined in only a token ofhe total number of nearly 24,000 species of Actinopte-ygii, the ray-finned fishes. The islet tissue of allpecies so far studied has cells possessing four types ofeptides: insulin, somatostatin, glucagon, and pep-

ides of the pancreatic polypeptide family (Fig. 3f).here are variations in the distribution of these pep-

ides within the gastrointestinal component of the GEPut the variations do not reflect any consistent phyloge-etic pattern. In contrast, there seems to be somehylogenetic developmental pattern in islet tissueistribution and arrangement among the actinopteryg-

ans. In the more basal (nonteleost) actinopterygianssturgeon, paddlefish, gar, bowfin, polypterids) theres a tendency for islets of various sizes to be scatteredhroughout the diffusely distributed exocrine tissueFig. 1i). The largest of the islet masses (principal islets)re mostly present near the point where the extrahe-atic common bile duct meets the intestine but therere not concentrated aggregates (Brockmann bodies).mong the lower teleosts, elopomorphs (eels) andsteoglossomorphs (bonytongues), the latter groupeflects the pattern of islet distribution described aboveor the basal actinopterygians. Islet distribution andrrangement among bonytongue species is consistentith taxonomic relatedness. Descriptions of islet distri-

ution in eels are inconsistent but principal islets inne species may represent a phylogenetic trend seenest in the salmonids. Coho salmon and rainbow trout,onsidered more generalized euteleosts, have isletsggregated into a specific area of the exocrine pancreasermed the islet organ or principal body, which may behe forerunner to the Brockmann body. Small islets arelso scattered throughout the remaining pancreas. Thisattern is also seen in other generalized euteoleosts,yprinids, and catfishes. All the more derived eutel-

IG. 3. Diagrammatic representation of endocrine cell types in throtochordate (a), larval (b) and adult (d) lampreys, hagfish (c), anrotochordate, one-hormone islets in the larval lampreys derived fruct in hagfish, and the three-hormone principal islet in the adult lam

o the gut epithelium but in cartilaginous (e) and bony (f) fishes alasmobranch is confined within a compact pancreas and the islet tissprincipal (partially shown) and secondary islets are present amon

rom the gut epithelium; a, c, e, and f were modified from Falkmer (1985a) a

osts have at least one principal islet in the Brockmannody, a large mass of islet tissue with or without aapsule of accompanying exocrine acini (Fig. 1j). Prin-ipal islets (as part of a Brockmann body), in additiono smaller scattered islets, are present in some speciese.g., percids). The ultimate in phylogenetic develop-

ent of islet tissue arrangement in the actinopterygianslet organ may be the presence of all islet tissue in onlyiscrete large bodies, the principal islets (e.g., angler-sh, sculpin). If principal islets are the endpoint, then

heir development can be traced through phylogeny ofhe actinopterygians. In addition, there is some evi-ence to suggest that the form taken by the endocrineancreatic homolog in actinopterygians has utility as a

axonomic character.The compact pancreas of the fishes of Sarcopterygii

eems more tetrapod like in its location but evidence ofrincipal islets in at least some dipnoans is moreimilar to the situation in the derived euteleosts. Theseata are support for the view ‘‘that a lungfish sharesore characters with a cow than a salmon’’ (seeloutier and Ahlberg, 1996) but also suggest that they

hare some characters with higher actinopterygians.his common character my reflect similar ontogeneticatterns in the two divergent groups of bony fishes.

eneral Morphology of the Islet Organn Chondrichthyes

The pancreas of extant chondrichthyans is a distinctompact organ (Figs. 1g, h) which reflects an early statef combined exocrine–endocrine relationship amongertebrates (Epple and Brinn, 1987). In elasmobrachssharks, skates, and rays) the pancreas is of an irregu-ar, often bilobed, (dorsal and ventral portions) shapeSekine and Yui, 1981) and it usually lies below thenterior end of the intestine and extends as far as theyloris. In the holocephalan (chimeras) the pancreas is

elium of the gut, bile duct, pancreatic ducts, and islet tissue of abranch (e), and a teleost (f). Note endocrine cells in the gut of thegut epithelium, the two-hormone islet tissue derived from the bilep to this point in phylogeny, glucagon-family peptides are confinedincluded in the islets (four-hormone islets). The islet organ of thetimate with the endocrine cells of the pancreatic ducts. In the teleost,reatic ducts and acini but were derived during early development

e epithelasmo

om theprey. U

re alsoue is ing panc

nd are reproduced with permission.


F

(F

ioF

a


CA

IG. 4. Electron micrograph of an islet cell of unknown type from the bowfin A. calva, which contains many electron-dense granulesarrowheads) and parallel arrays of crystalline inclusions (arrows); original magnification, 318,000.IG. 5. The islet organ of a larval lamprey, Petromyzon marinus, has insulin-positive follicles both free (arrow) from and attached to (arrow) the

ntestinal epithelium but also intraepithelial aggregates (arrowhead) of similarly stained cells; peroxidase–anti-peroxidase (PAP) technique;riginal magnification; 3300.IG. 6. The cranial principal islet of adult Petromyzon marinus possesses numerous D cells identified by their immunoreactivity to
nti-somatostatin-14 antibody; PAP technique; original magnification; 3300.

m(mt1klohatpft(httidolraBihgosc1ce

emmhtmMheapi

cc(hladbhedfrpdEavotaaaad1cpmts1

eotaasnas

SI


ore band like, lying along the hepatic portal channelJollie, 1973). Epple and Brinn (1986) emphasized the

arked difference in the form of islet tissue withinhese two groups of cartilaginous fishes (Figs. 1g andh). The ancestral group of Chondrichthyes is notnown but the two subclasses are likely a monophy-

etic unit and it is possible that they had commonrigins (Nelson, 1994). Holocephalans are believed toave the most ancient lineage among chondrichthiansnd this fits well with the view of Falkmer (1995) thathey display the first pancreatic gland in gnathostomehylogeny. However, Fujita (1962) had previously

urnished a detailed description of islet morphology ofhe ratfish which would tend to contradict the Falkmer1995) conclusion. It is particularly noteworthy that theolocephalan pancreas has a general morphology like

hat of mammals, with diffusely distributed isletshroughout the exocrine parenchyma. Sometimes theslets are large but they are not principal islets. Aeviation from that of other gnathostome fishes is thatnly insulin, somatostatin, and glucagon peptides are

ocalized in the islets; pancreatic polypeptide immuno-eactivity is present in the intestine and in the pancre-tic ducts (Stefan et al., 1981). In contrast, Epple andrinn (1986) described a few F cells and a X cell in the

slets; the X cells represent 50% of all islet cells inolocephalians (Epple and Brinn, 1987) and store alucagon-like peptide closely related to mammalianxyntomodulin (Conlon et al., 1987). Thus, despite theuggestion of a more ancient lineage of holocephalansompared to elasmobranchs, Epple and Brinn (1986,987) support the view of Fujita (1962) that the holo-ephalan pancreas is intermediate in form to that oflasmobranchs and ‘‘higher’’ fishes.

In contrast to the above, Falkmer (1995) equates thelasmobranch pancreas to be ‘‘of the well-knownammalian type’’ on the basis of both gross andicroscopic features and to be more advanced than the

olocephalan pancreas. In particular, the islets havehe primary four cell types (A, B, D, and F), as seen in

ost bony fishes through to mammals (Jonsson, 1991).oreover, an elegant histochemical study of a shark

as revealed more than just the classical islet cell types,.g., several amphiphil cells (Epple, 1967). Islet cells inskate are immunoreactive to both somatostatin andancreatic polypeptide antisera. Immunoreactivity for

nsulin-like growth factor-2 is present in B cells and is c

onsidered to reflect both the ancient origin and theonserved nature of this molecule among vertebrates.Reinecke et al., 1994). In addition, immunoreactivityas been noted in the islets to antisera against enkepha-

ins, FMRF amide, gastrin/cholecystokinin/caerulein,nd neurotensin (Jonsson, 1991). Collectively theseata suggest, as predicted by Epple (1967), that elasmo-ranchs have many more islet cell types than theolocephalan islets. As in holocephalans, the islets oflasmobranchs have a relationship with pancreaticucts (Fig. 3e), but the relationship is quite variable. In

act, Thomas (1940) described three types of islet–ductelationship in elasmobranchs. These are islet cells asart of the ducts, groups of islet cells budding from theuct, and islets close to, but independent of, the ducts.ven from the small number of descriptions presentlyvailable it seems that there may be considerableariation in the types of islet–duct relationships withinrders and families of elasmobranchs. For example,he pancreatic endocrine cells of the ray Dasyatis akajeire described as representing the most ‘‘primitive’’rrangement among elasmobranchs in being presents a layer beneath an inner ductular epithelium (Sekinend Yui, 1981). In contrast the starry ray, Raja radiata, isescribed as having large and small islets (Jonsson,991). Islets are also described in the sting ray, Rajalavata, but at least some have intimacy with theancreatic ducts (Reinecke et al., 1992, 1994). Isletorphology and the islet–duct relationship also seems

o differ between selachians, e.g., a dogfish and a sharkpecies (Epple and Brinn, 1975; Kobayashi and Ali,981; Jonsson, 1991).The endocrine cells of the gastrointestinal system of

lasmobranchs has received a great deal of attentionver the years and the reader is referred to several ofhe early and recent reports (El-Salhy, 1984; Cimini etl., 1989; Tagliafierro et al., 1989; Yui et al., 1990; Chiba etl., 1995; Chiba, 1998). However, despite the manypecies of elasmobranchs that have been investigatedo phylogenetic pattern involving the presence/bsence or distribution of regulatory peptides in thetomach and intestine has emerged.

ummary of the Phylogenetic Development of theslet Organ in Chondrichthyes

The distribution and form of the pancreatic endo-
rine tissue in cartilaginous fishes seems to bear no

rpcd(a(snasmmetSt(poWpcmbowcowiatfpFfhtg

M

fiaet

cLcmitndbpsiaaapt1ddpohnfi

tlwocbi1tt(LapIwheem


CA

elation to that observed in bony fishes. There is a view,artially based on the presence of three primary isletell types, that the holocephalan pancreas is the firstiscrete pancreas to appear in gnathostome evolution

Falkmer, 1995) but it is remarkably similar in grossnd microscopic appearance to that found in tetrapodsFig. 1g). A contrasting viewpoint is that islet cell typesimplified rather than diversified during the phyloge-etic development of the gnathostome pancreas (Epplend Brinn, 1987). In this latter context, the islets ofelachians among the elasmobranchs may represent aore ancient islet from that seen in holocephalans. Theammalian-type, compact pancreas is also depicted in

lasmobranchs but its microscopic appearance is reflec-ive of a pancreas in a state of morphogenesis of islets.ince the holocephalans are generally believed to havehe more ancient lineage of these two subclassesSchaeffer, 1981), phylogenetic development of theancreas, and in particular the arrangement and formf the islet tissue, is not clear between the two groups.ith this one character it is hard to see an immediate

hylogenetic developmental pattern of the two sub-lasses. On the other hand, variations in islet arrange-ent and form in elasmobranchs (selachians and

atoids) has excellent potential for the demonstrationf the phylogenetic development of an organ systemithin a vertebrate subclass. The relevance of the

hondrichthyan islet organ to the phylogenetic devel-pment of this organ in the vertebrate group as ahole is also not clear. The above data must be

nterpreted in light of results from a recent mtDNAnalysis that show Chondrichthyes representing aerminal position in the piscine tree and divergingrom other gnathostomes much earlier than had beenreviously predicted (Rasmussen and Arnason, 1999).or instance, their cartilaginous skeleton is not theorerunner to the bony skeleton of Osteichthyes origher vertebrates. Thus the islet organ of Chondrich-

hyes is likely not basal to that of other gnathostomeroups.

orphology of the Islet Organ in Agnatha

Lampreys and hagfishes are the only extant jawlessshes or agnathans. They may have shared a commonncestor in the early Cambrian but they diverged veryarly in their evolutionary histories. Hagfish are likely
he most ancient of living vertebrates, since their origin t

an be traced back to close to 550 million years.ampreys, on the other hand, have had a ratheronserved evolution since they first appeared 350illion years ago (Forey and Janvier, 1994). It is

mportant to our discussion that the reader be aware ofhe differences in the life histories of the two ag-athans. Hagfish are entirely marine and have a directevelopment following hatching; that is, the postem-ryonic young resemble the adults. In contrast, lam-reys are represented by anadromous and freshwaterpecies and they have an indirect developmentalnterval in their life cycle. The freshwater larvaeppear after hatching and after 2–7 years they undergometamorphosis to acquire juvenile adult characters

nd to permit adult behavior (Youson, 1988). Thehylogenetic history of the enteropancreatic (EP) sys-

em in agnathans has recently been discussed (Youson,999) and the reader is referred to this article for moreetail and an extensive literature review. The presentescription will concentrate on details of the agnathanancreas which are relevant to the general discussionf the ontogenetic and phylogenetic developmentalistories of the fish GEP system. This discussion wouldot be complete without some reference to these extantshes with such ancient lineages.Hagfish. Falkmer (1995) describes the hagfish as

he first vertebrate to have an islet organ. This organ isocated at the junction of the extrahepatic bile duct

ith the intestine (Fig. 1f) and is an aggregate of isletsr follicles of epithelial cells (Fig. 3c). The epithelialells of the follicles surround a wide lumen in some,ut not all, species and most of the cells (B cells) are

mmunoreactive with only anti-insulin serum; about% of the cells are D cells (Fig. 3c). Cells equivalent tohose of the exocrine acini of other fishes are present inhe intestinal mucosa of the hagfish (Fig. 1f). Falkmer1995) states that this is evidence that the islets ofangerhans of the pancreas are phylogenetically morencient than the exocrine acini and the first ‘‘endocrineancreas’’ was a two-hormone gland (Falkmer, 1985).

n contrast, to the situation in cartilaginous fishes,here islets are associated with pancreatic ducts, the

agfish follicles are intimately associated with thepithelium of the bile duct (Fig. 1f). The bile ductpithelium is also interspersed with insulin- and so-atostatin-immunoreactive cells (Fig. 3c). The other

wo peptides usually found in the islets of other fishes,

npsdbd

lpmadSnasedlctae

sti3scfdcl(hHchlenctpii

igmrplorgh(eP

emuhata1ii(mdpoibcc1pseHblshstti


amely, pancreatic and glucagon family peptides, areresent in the epithelium of the intestine along withomatostatin (Fig. 3c). Thus the EP system in hagfish isispersed among three components, the intestinal andile duct epithelia and an islet organ which is indepen-ent of exocrine acini.Lamprey. Like hagfishes, there is no direct equiva-

ent to a higher vertebrate exocrine pancreas in lam-reys at any stage of their life cycle (Figs. 1b–1e). Inost cases, the zymogen cells are interspersed among

bsorptive cells in the anterior intestine or a smalliverticulum (Barrington, 1972; Youson, 1981) butouthern Hemisphere species have expanded intesti-al diverticuli where these cells are located (Strahannd Maclean, 1969). An enlarged left diverticulum, theo-called protopancreas, of Mordacia mordax is consid-red a derived feature and reflective of the earlyivergence of the family Mordaciidae from the other

amprey families (Gillett et al., 1996). This latter viewontrasts with that of Epple and Brinn (1987) but thewo groups of authors are consistent in their opinion that

compacted mass of exocrine pancreatic-type cellsvolved several times during vertebrate evolution.

In larval lampreys, the endocrine tissue is present asmall islets or follicles in the submucosal connectiveissue near the junction of the esophagus, the anteriorntestine, and the extrahepatic common bile duct (Figs.b and 5) in holoarctic species. In Southern Hemi-phere species the islet aggregate is located near theonfluence of the intestine, diverticula, and esophagus,or the bile duct enters the cephalic part of one of theiverticula (Fig. 1c). The term ‘‘islet organ’’ is not ofommon use to describe the islet aggregate in larvalampreys, but it was applied by Epple and Brinn1986). Islet organ would also seem to be appropriateere, for in some species, such as, in the Southernemisphere, Mordacia mordax, the islets are highly

ompacted (Youson and Potter, 1993), like those in theagfish. Some islets are contiguous with the epithe-

ium of the anterior intestine but all are composedntirely of B cells (Figs. 3b and 5). Anti-insulin immu-oreactivity is also present in some intraepithelial celllusters of the anterior intestine (Fig. 5). Immunoreac-ivity to antisera against somatostatin and PP-familyeptides is present in isolated cells of the anterior

ntestine but intraepithelial cell clusters are also PP
mmunoreactive (Fig. 3B; Cheung et al., 1991b). There c
s some question as to whether peptides of the gluca-on family are present in the larval intestine; theajority of species so far examined show no immuno-

eactivity to either anti-glucagon or -glucagon-likeeptide (Youson and Potter, 1993). Therefore, the larval

amprey EP system possesses a one-hormone, isletrgan and, in general, the remaining most commonegulatory peptides (somatostatin, PP-family, gluca-on?) are confined to the intestinal epithelium. Insulinas been localized in the intestine of larval M. mordaxYouson and Potter, 1993) and just recently in thextrahepatic common bile duct of 35-day-old larvae ofetromyzon marinus (Youson, 1999).The term islet organ has been used to describe the

ndocrine pancreatic homolog in adult lampreys (Falk-er, 1985a; Epple and Brinn, 1986). In the context of its

se in other fishes of the present discussion, includingagfishes and larval lampreys, this term still seemsppropriate for adult lampreys. Islet organ is a generalerm referring to the endocrine pancreatic homolog in

vertebrate, irrespective of its distribution (Falkmer,995). In adult lampreys, the islet tissue is compactednto one or two large bodies of lobules and there is nontervening exocrine acini or much connective tissueFigs. 1d and 1e). The situation in adult lamprey is

uch like the principal islets (Brockmann bodies)escribed in the most derived euteleosts. In particular,rincipal islet usually refers to an isolated, single massf islet tissue. There is a single cranial aggregate of

solated islet tissue in Southern Hemisphere speciesut all Northern Hemisphere species have both aranial and caudal aggregate with an intermediateord of cells between the two (Youson and Cheung,990). We have used the term cranial and caudalancreas for these aggregates, but now admit thatince there is no exocrine element associated withither aggregate, this term is not anatomically correct.enceforth, it is now recommended that principal islet

e used for the endocrine pancreatic homolog in adultampreys. The islet organ in Southern Hemispherepecies is a cranial principal islet (Fig. 1e). Since in theolarctic species the two largest aggregates are bothtructurally and morphogenetically independent andhey are some distance apart, they should be referredo as cranial and caudal principal islets (Fig. 1d). Thentermediate cord is secondary islet tissue. These three
omponents comprise the islet organ in the Northern

Hit

wwismpcisa

ham1GipPicatasc(1sfl1tiIf-tCvets

is(dii1csdapmtaast1imltie

idtdtittgpvp

SI

cmli


CA

emisphere species. This definition of the islet organn adult lampreys would fit well with previous interpre-ations (Falkmer, 1985a, 1995).

It was stated at the outset of this review that thereould be little focus on the structure of peptidesithin the various groups of fishes. However, the

nterfamilial and intergeneric differences in the EPystem of lampreys have proved intriguing. Whenorphological data of principal islets and structure of

eptides of the EP system are compared among spe-ies, a picture is painted of the importance of examin-ng more than one species, and the utility of the EPystem in analysis of phylogenetic development, withingroup of fishes.The principal islets of adult lampreys are a three-

ormone tissue containing B, D, and F cells (Figs. 3dnd 6); no immunoreactivity has been found forembers of the glucagon family (Youson and Elliott,

989; Youson and Cheung, 1990; Cheung et al., 1991a).lucagon and glucagon-like peptide have been local-

zed in the intestine, along with somatostatin andeptides of the pancreatic polypeptide family (Fig. 3d).YY has been conserved among three genera, represent-

ng two families of lampreys (Wang et al., 1999a), butomparisons of amino acid sequences of somatostatinsnd insulin indicate some evolutionary pressure onhese molecules. Extant lampreys share a commonncestral insulin gene, for their insulins all have theame 5-amino-acid extension at the N-terminal of the Bhain; this feature is not found in any other vertebratePlisetskaya et al., 1988; Conlon et al., 1995b; Youson,999). However, insulin of the Southern Hemispherepecies Geotria differs by 17 amino acid substitutionsrom the identical insulins of two genera of holarcticampreys, Lampetra and Petromyzon (Conlon et al.,995b). This difference perhaps accounts for the facthat antisera against Petromyzon insulin would notmmunostain Geotria B cells (Youson and Potter, 1993).n contrast, ‘‘large’’ forms of somatostatin isolatedrom the GEP system of each genus (somatostatin-35,34, -33 for Lampetra, Petromyzon, and Geotria, respec-ively) differ markedly in their N-terminus but the-terminus of each peptide is characterized by aariant form of somatostatin-14, Thr12 = Ser (Andrewst al., 1988; Conlon et al., 1995b). This form of somatosta-in is found only in the lamprey EP system, for brain
omatostatin-14 of both Lampetra and Petromyzon is the l

nvariant form found in the neuroendocrine and GEPystems of all other vertebrates examined to dateSower et al., 1994; Conlon et al., 1995a). A tissue-ependent, differential expression of a gene(s) encod-

ng preprosomatostatin is also found between thentestine and the principal islet of Geotria (Wang et al.,999b). These data serve as documentation for theommon ancestry of the lamprey genera (variantomatostatin-14), that the divergence of Petromyzonti-ae and Geotriidae families is likely ancient (insulin),nd that there has been considerable evolutionaryressure directed against the N-terminal of preproso-atostatin among the lamprey genera. It is notewor-

hy in the context of phylogenetic development of thegnathan EP system that hagfish and lamprey insulinsnd somatostatins show little similarity. However, aomatostatin-34 with an invariant somatostatin-14 athe C-terminus is found in the hagfish (Conlon et al.,988). Thus the agnathan phyletic group is character-zed by a dominant, islet organ somatostatin with a

inimum of 33 amino acids. The persistence of thesearge somatostatins in agnathans and the difference inhe C-terminus between hagfishes and lampreys arentriguing from the points of view of taxonomic relat-dness, molecular processing, and biological activity.

The story of proglucagon in lampreys is still unfold-ng but it seems that there are two genes encoding forifferent glucagons in Geotria intestine and that Lampe-

ra and Petromyzon each express only one, but aifferent, gene (Wang et al., 1999b). These data support

he view of an early duplication of the glucagon genen lamprey evolution. Recently it has been shown thathe two glucagon genes are present in Petromyzon andhat one encodes for glucagon-like peptide-1 and alucagon and a second encodes for glucagon-likeeptide-2 and another glucagon (Irwin et al., 1999). Theiew is that the duplication of the glucagon genereceded the appearance of the first vertebrates.

ummary of the Phylogenetic Development of theslet Organ in Agnatha

Agnathans are unique among fishes in having aomplete separation of their endocrine pancreatic ho-olog (islet organ) from the exocrine pancreatic equiva-

ent which is present either in the mucosa of thentestine or in special intestinal diverticula (larva
ampreys of Southern Hemisphere species). Assuming

ttosomltapiFowmsSHabao1aaea(mCmlisms

PG

tbtpb

Fdta(cdffhsbpmbatnhmtoccig(o1fHtp(at

ttttwrodi


hat hagfish are the oldest of the extant vertebrates,heir endocrine pancreatic homolog is the first isletrgan and it is a noteworthy comparison with gnatho-tome fishes that the lobules of this two-hormonergan are often confluent with the extrahepatic com-on bile duct. However, the insulin-only, islet organ of

arval lampreys may represent the most basal of isletissues present within a free-living vertebrate (Figs. 3bnd 5). The three-hormone islet organ of adult lam-reys may be looked at as an advancement over the

slet organ of the more ancient hagfish (Fig. 3d).amilial variations in the distribution of the islet organf adult lampreys is a specific taxonomic characterhich is a consequence of family-specific differences inorphogenetic events occurring during metamorpho-

is. A single, cranial principal islet is characteristic ofouthern Hemisphere species, whereas the Northernemisphere species have two principal islets, cranial

nd caudal, and a cord of secondary islet tissueetween these two islets (Figs. 1d and 1e). These datare consistent with the views of the earlier separationf the hagfish and lamprey lines (Forey and Janvier,994) and that, although the lamprey families likelyrose from a common stock, divergence of Southernnd Northern Hemisphere species occurred relativelyarly in the evolutionary history of lampreys (Potternd Hilliard, 1987). Moreover, the petromyzontidsRenaud, 1997) had a more conserved evolution than

embers of the two Southern Hemisphere families.omparison of the primary structures of insulin, so-atostatin, and glucagon among members of two

amprey families and between insulins and somatostat-ns of lampreys and hagfishes are providing additionalupport for these views on the phylogenetic develop-ent and variation of the EP system among agnathan

pecies.

HYLOGENETIC ORIGIN OF THE FISHEP SYSTEM

Falkmer and his colleagues have long held the viewhat the origin of the vertebrate endocrine pancreas cane traced to what they call the brain–gut axis withinhe subvertebrate members of Phylum Chordata, therotochordates, and other deuterostomian inverte-
rates (Falkmer and Patent, 1972; Van Noorden, 1984; m
alkmer, 1985a,b, 1995). This view of the phylogeneticevelopment of the vertebrate GEP system contends

hat peptides common to the vertebrate system firstppeared in the nervous system of lower invertebratesVan Noorden, 1984) and then they appeared in endo-rine cells of the gut of some protostomian and manyeuterostomian invertebrates (Falkmer, 1995). Data

rom the protochordates are particularly relevant here,or it is from this group that the first vertebrates mayave arisen (Romer, 1970) and these lower chordateshow dual occurrence of many regulatory peptides inoth the brain and gut, i.e., a culmination of thehylogenetic development of the brain–gut axis (Falk-er, 1995). There is no direct equivalent to the verte-

rate islet organ among the protochordates (Figs. 1and 3a), i.e., Urochordata and Cephalochordata. Al-hough numerous studies of protochordate gut immu-ohistochemistry had preceded that of Reinicke (1981),is investigation of the mapping of polypeptide hor-one immunoreactivity (insulin, glucagon, somatosta-

in, and pancreatic polypeptide) in the digestive tractf Branchiostoma lanceolatum has been one of the mostomprehensive (Fig. 3a). Further immunohistochemi-al, biochemical, and molecular studies have alsondicated that in Branchiostoma spp. there is a commonene encoding insulin and insulin-like growth factorIGF) and they are likely synthesized in the same cellsf the gut and brain (Chan et al., 1990; Reinecke et al.,993a). In fact, IGFs and insulin may have originatedrom the same ancestral molecule (Plisetskaya, 1989a).owever, the two cDNAs recently isolated from the

unicate Chelyosoma productum likely code for separateroducts, a proinsulin molecule and a preproIGF

McRory and Sherwood, 1997). An insulin-like peptidend an IGF-1 coexist in the intestinal cells of anotherunicate, Ciona intestinalis (Reinecke et al., 1999).

The next step in the phylogenetic development ofhe vertebrate endocrine pancreas is the budding ofhese gut endocrine cells from the mucosal epitheliumo produce the first islet organ. The reader is referred tohe schematic presentation of Epple and Lewis (1973),

here several possible developmental pathways, theole of open- and closed-type cells in the gut, and thentogenetic relationship to the exocrine pancreas areiscussed. These various interpretations are important

n light of the facts presented in this review that the
ajor piscine groups diverged early in their evolution

aiaiciifoo(pcfaiio1eitvbeitAdg

O

B

1ftttbdo1(

(aGcncm(

aoppotuaToo1u1lppsrpaw(tnb(p1cfiataa

b


CA

nd their islet organs took on different forms. Fornstance, the compacted pancreas of Chondrichthyes isconsequence of a phylogenetic–ontogenetic pathway

nvolving an intimacy of pancreatic ductular and isletells, whereas agnathans used a pathway involving anntimacy of bile ductular and/or intestinal cells withslet cells. The step as represented in the simplest formrom a hypothetical protochordate-like precursor is thene-hormone islet organ of the larval lamprey whichriginates primarily from the intestinal epitheliumFig. 3b). The hagfish deviated somewhat from thishylogenetic developmental trend by having the endo-rine cells of this two-hormone islet organ originaterom the extrahepatic common bile duct (Fig. 3b),lthough many regulatory peptides are present in thentestine (Van Noorden, 1990). Furthermore, IGF-1mmunoreactivity is present in the intestine, isletrgan, and brain of hagfish (Reinecke et al., 1991,993b) and, as in the protochordate intestine (Reinecket al., 1993a, 1999), insulin and IGF-1 immunoreactivitys colocalized. The above discussion emphasizes howhe study of the phylogenetic development of theertebrate islet organ in some ways contributes to theroader analysis of the origins of vertebrates in gen-ral. However, beyond the simplest islet organ as seenn larval lampreys, it is becoming clear that the defini-ive islet organs of members of Chondrichthyes andctinopterygii are likely a consequence of their earlyivergence and terminal position relative to piscinenathostomes which led to the tetrapods.

NTOGENY OF THE FISH GEP SYSTEM

ackground

In the past decade it was emphasized (e.g., Andrew,984) that it was unfortunate that our knowledge of theactors involved in functional and structural matura-ion of fetal pancreatic endocrine cells was so poor, forhis information had clinical value, particularly inransplant technology (Reddy and Elliott, 1988; Du-ois, 1989). This paucity of information was partiallyue to the fact that the earlier views of a neural crestrigin of islet cells (Pearse, 1969; Pearse and Polak,971) had been refuted through experimental studies
Andrew, 1976; Le Douarin, 1988). Rombout et al. p

1978) used fluorescent microscopy, autoradiography,nd electron microscopy to study development of theEP system in the fish Barbus conchonius and con-

luded that the GEP cells are not derived from theeural crest. Although the endodermal origin of isletells is now accepted, nerve cells and islet cells shareany morphological, functional, and biochemical traits

Fujita, 1989; Teitelmann, 1990; Reddy et al., 1997).Within this present decade there has been a remark-

ble advance in our knowledge of the events of thentogeny of islets of Langerhans in mammals. Inarticular there has been a dramatic shift from thereviously held concept of one hormone/cell duringntogeny to a wider acceptance of a precursor cell forhe four types of principal endocrine cell types whichltimately localize insulin, glucagon, somatostatin,nd pancreatic polypeptide (Hashimoto et al., 1988).hat is, there is a series of stages in the differentiationf the islet cells and one of these stages is that theyften acquire more than one hormone (Zabel et al.,994). Early in development glucagon and insulin aresually localized in the same cell (Teitelmann et al.,993). Somatostatin and pancreatic polypeptide areocalized in cells much later in development but at oneoint in mouse development PYY is present in all fourrincipal islet cell types (Upchurch et al., 1994). Theecond advance has been in the investigation of factorsegulating the development and expression of the fourrincipal hormones. For example, Pax genes 4 and 6re required for development of the mouse pancreasith Pax 6 directing the differentiation of A cells

St-Onge et al., 1997). A new member of the proteinyrosine phosphatase (PTP) family, called PTP-NP (foreural and pancreatic), is a receptor-type transmem-rane molecule and an early marker of developmentChiang and Flanagan, 1996). Adrenomedullin is ca-able of modulating insulin secretion (Martinez et al.,996) and during development it is expressed by isletells (Montuenga et al., 1997). This hormone appearsrst with cells containing glucagon, eventually withinll principal islet cell types, and ultimately with onlyhose containing pancreatic polypeptide (Martinez etl., 1998). It is concluded that early expression ofdrenomedullin is critical for islet development.Studies of the ontogeny of the fish islet organ are few

ut such studies are both plausible and important,
articularly in light of the potential use of fish princi-

p1tdaosd

oetceeaifiEavgdavapsvmeh1tpAtntg

B

gs1p

fttisdtclpaosDps(pcIeiSpte

F

sep


al islets as tissue for xenotransplanting (Wright et al.,998; Yang et al., 1999). Fish can undergo one of twoypes of postembryonic development: indirect, with aramatic metamorphosis of larvae into the juvenile,nd direct, with a gradual differentiation and growthf the fry into the juvenile (Youson, 1988). The readerhould keep these two types of development in minduring the subsequent discussion.The present evidence is that there is a common

rigin of islet tissue in all vertebrates, that is, frommbryonic endoderm, the primitive gut, or their deriva-ives (Epple and Brinn, 1986). In mammals, the pan-reas results following fusion of ventral and dorsalvaginations of the primitive gut. Pancreatic ductsventually produce the cells which create both thecini and endocrine cells (Teitelmann et al., 1987). Earlynvestigations of the development of the islet organ inshes are those of Baron (1935) and Vorstman (1948).pple and Brinn (1986) emphasize that three anlagenre involved in the development of the pancreas inertebrates ‘‘above’’ Chondrichthyes. The ventral anla-en produces a few scattered islets. In contrast, in theorsal anlagen there is an intimacy with the exocrinend endocrine components which may account for theariation in size and distribution of islets in organismsbove Chondrichthyes. A close relationship of the twoancreatic components results in a dissemination ofmall islets, whereas large islets (principal islets) de-elop when intimacy (invasion) of exocrine acini isarginal. It was concluded that this relationship of

xocrine and endocrine components of the pancreasas little functional significance (Epple and Brinn,986). However, these latter authors have emphasizedhe importance of blood circulation and innervation toancreatic ontogeny (Epple and Brinn, 1986, 1987).lthough subsequent studies have described nerve

erminals in definitive islet organs of fish, there isothing further to report on the significance of innerva-

ion to pancreatic morphogenesis in this vertebrateroup.

ony Fish

The development of the teleost pancreas from aross anatomical perspective had been most exten-ively investigated in sea bass, D. labrax (Diaz et al.,989). The primordial pancreas first appears as a
osterodorsal thickening on the right side of the t
oregut of the posthatched fry. Eventually the righthickening buds laterally to produce a pancreaticissue which spreads along blood vessels and thentestinal wall. Light and electron microscopic analy-es (Guyot et al., 1998) have shown that by the seconday after hatching of the gilthead sea bream, S. aurata,

he bud is a cell mass differentiated into two separateell populations, exocrine and endocrine cells, with theatter forming a small islet (Fig. 7). The above view-oint of Epple and Brinn (1986) that dorsal and ventralnlagen have differential participation in developmentf the teleost pancreas was subsequently demon-trated in a study of a four-stage development of. labrax (Garcıa-Hernandez and Agulleiro, 1992). Arincipal islet developed dorsally and later severalmall islets appeared more ventrally. Five types of cellsA, B, D1, D2, and F) have been distinguished in therincipal and smaller islets by both immunohistochemi-al and fine structural features (Agulleiro et al., 1994).n earlier stages of development a ‘‘primordial cord’’ isnclosed in the dorsal wall of the anterior midgut andt was composed of cells immunoreactive to anti-ST-25 (D1) and to anti-insulin (B cell). This cordasses through a primitive islet with the same cell

ypes and eventually into a single islet surrounded byxocrine acinar cells. This islet had a regional distribu-

IG. 7. Transverse section of a 2-day-old prelarva of the giltheadea bream, Sparus aurata, showing islet tissue (arrowhead) andxocrine pancreatic tissue (arrow) which has developed from arimordial cell mass originating from the gut (G); original magnifica-

ion; 31200. (From Guyot et al., 1998.)


ttap(nBsdpecng(slH

ttcpfiatcvTcAppswttaiGtp(

ls

faTfpttip(ostf

oa111cahtrHsteg

r(1sistgtpgtgst


CA

ion of peripheral D1 cells, the most peripheral A cells,he innermost B cells, and the intervening D2 cells. Inddition, ‘‘primordial islets’’ developed in the ventralancreas from the epithelium of pancreatic ductules

Garcıa-Hernandez and Agulleiro, 1992; Garcıa-Her-andez et al., 1994a) and they are composed of D1 and

cells. It was concluded that, in contrast to theituation described above in mammals, each cell typeevelops independently, i.e., not from a commonrecursor cell (Garcıa-Hernandez et al., 1994a). How-ver, there may be a common precursor cell for A and Fells (Garcıa-Hernandez and Agulleiro, 1992). Anotheroteworthy difference to mammals was the absence oflucagon early in development of islets in this speciessee also Beccaria et al., 1990) but perhaps insulin andomatostatin interact early in development (beforearval feeding) to regulate differentiation (Garcıa-

ernandez et al., 1994a).Immediately after hatching of larvae of both the

urbot Scophthalamus maximus (Berwert et al., 1995) andhe gilthead sea bream S. aurata (Guyot et al., 1998),lumps of B cells are the first to appear in a singlerimordial islet located at the epithelium of the undif-

erentiated intestinal tube. In the turbot, insulin-mmunoreactive cells persisted as this islet enlargednd new, smaller islets formed. The onset of feeding inhis species was correlated with the appearance of Dells (these usually first) and A cells but there was aariation in distribution in the different-sized islets.he principal islet had an intermingling of B and Dells but smaller islets had D cells at the periphery with

cells. F cells were the last cell to appear, 12 daysosthatching. Immunoreactivity to anti-IGF-1 ap-eared at day 11 in F, A, in some D cells, and some cellshowed colocalization of the principal peptides; thisas interpreted as corresponding to the beginning of

he growth promoting effect of piscine pancreaticissue. The ordered sequence of appearance of B, D, A,nd F cells in the ontogeny of the turbot pancreas wasnterpreted as reflecting Haeckel’s biogenetic rule (seeould, 1977) that ‘‘ontogeny recapitulates phylogeny’’;

his is the same order that they seem to appear inhylogenetic development of the vertebrate pancreas

Berwert et al., 1995; also see earlier discussion).A recent study used ELISA to show the highest

evels of insulin to be at the hatching stage of gilthead
ea bream when the pancreas primordium is about to

orm (Guyot et al., 1998) and endotrophism (yolkbsorption) is the primary source of nutrition (Fig. 8).his insulin, likely of maternal origin, is probably used

or organ growth and cell differentiation, which arerominent features of developing larvae during the 0-

o 5-day posthatching period. Of particular impor-ance at this time is the development of the liver, whichs the essential organ when the larvae becomes exotro-hic and carbohydrates are the main energy reserve

Guyot et al., 1995). The coordination of developmentf the islet organ, liver, and intestine are critical to theurvival of the larva with respect to insulin produc-ion, glyconeogenesis, and absorption of an exogenousood source, respectively (Guyot et al., 1998).

The development of the gastrointestinal componentf the GEP system in bony fish has received morettention than that of the islet organ (Rombout et al.,978; Connes and Benhalima, 1984; L’Hermite et al.,985; Garcıa-Hernandez et al., 1994b; Reinecke et al.,997). In addition to the differentiation of cells whichontain the classical islet hormones, these studies havelso shown the time and appearance of some of theormones specific to the gastrointestinal component of

he GEP system. For this latter group of hormones, theeader is referred to the most recent articles (Garcıa-ernandez et al., 1994b; Reinecke et al., 1997). For the

ake of both brevity and conformity with the rest ofhis paper we will consider only the ontogeny of cellslaborating insulin, somatostatin, and peptides of thelucagon and pancreatic polypeptide families.The ontogeny of gastroenteroendocrine cells in indi-

ect developing bony fish has been studied in eelsL’Hermite et al., 1985) and the turbot (Reinecke et al.,997). Eel leptocephali just prior to metamorphosishow only slight labeling for glucagon in the smallntestine. However, after metamorphosis the glass eelshow a fairly abundant glucagon immunoreactivity inhe ‘‘duodenal bulbous’’ and scarce staining for gluca-on in the small intestine and rectum. Anti-somatosta-in immunoreactivity was present only in cells of theyloric cecum and stomach at this stage. The pattern oflucagon and somatostatin distribution continued intohe adult eel stage but no insulin was detected in theut at any stage. No physiological or developmentalignificance was attributed to either the distribution oriming of expression of these peptides.

In the study of premetamorphic development of the

tiawmpiiauqatiifitpTves

sTobitpOttaftt

hblc0ae

F

(


urbot gastrointestinal endocrine cells, many morentervals of development were examined (Reinecke etl., 1997). Five prejuvenile stages (Segner et al., 1995)ere described over 24 days of posthatching develop-ent. In the first 4 days (corresponding to endotro-

hism of yolk) there was no presence of any of the fourslet-type hormones. Transient insulin immunoreactiv-ty was present from days 5 to 10 in the stomach anlagend in the upper intestine. IGF-1 appeared in thepper intestine between 8 and 10 days and subse-uently became a conspicuous component of the stom-ch and the two regions of the intestine after 11 dayshrough to the 41-day-old juvenile. Somatostatin-mmunoreactive cells also appeared at the same timen the stomach and upper intestine but were mostlyound in the stomach thereafter. The two regions of thentestine showed immunolabeling for PP-family pep-ides at the 8- to 10-day stages but afterward the mostrominent labeling was present in the upper intestine.he conclusion was that the differentiation of thearious cell types is correlated with the onset orxogenous feeding by the larva and that metamorpho-

IG. 8. A comparison of percent insulin levels (ELISA) in homogenidays and nutritional status) during their development. Comparison

is is important to develop a finetuning of the nervous e

ystem for regulation of the blood flow to the intestine.he transient insulin immunoreactivity was a curiousbservation and was discussed in the context of iteing a phylogenetic relict, i.e., like that seen in the

ntestinal epithelium of larval lampreys or possiblyhat insulin is related to protein synthesis during cellroliferation of the developing mucosal epithelial cells.ther curiosities were the significance of the IGF, the

iming of appearance of glucagon and PP-family pep-ides, the colocalization of anti-SST-14 and -SST-25,nd the lack of colocalization of glucagon and PP-amily peptides. The time of appearance of somatosta-in in the stomach was coincident with the differentia-ion of this organ.

Differentiation of gastrointestinal endocrine cellsas also been examined in a few direct developingony fish. As noted earlier, the enteroendocrine cells of

arval B. conchonius are not derived from the neuralrest (Rombout et al., 1978). In the sea bass (D. labrax),- to 5-day posthatched larvae showed no immunore-ctivity in the undifferentiated gut (Garcıa-Hernandezt al., 1994b). After 9–15 days, at the commencement of

larvae and larva of the gilthead sea bream, S. aurata, at various timesade with prelarvae at hatching as 100%. (From Guyot et al., 1998.)

zed pres are m

xogenous feeding, cells immunoreactive toanti-NPY/


PtiwdaiiftPtpnd

C

tdsoaptspcdapcaoTrdsanrcStoea

iTtit

A

pmofndfeoe(tt1SaoNta1

dtlmabmtattlteil


CA

YY were present in the intestine and by 25–46 dayshe same cells contained glucagon. At this latter time,nsulin, SST-25, PP/glucagon-immunoreactive cells

ere present in the stomach but eventually (55–60ays) the somatostatin cells also stained positive fornti-SST-14. It was suggested that feeding inducesntestinal cell differentiation and that immunoreactiv-ty for both glucagon and insulin represent gut-specificorms of these hormones, i.e., GLP and IGF, respec-ively. There was a colocalization of glucagon with theP-family peptides but glucagon and PP immunoreac-

ivity appeared at the same time and the latter ap-eared before NPY/PYY and became more pro-ounced in the stomach; a reversal of this timing andistribution was found for the intestine.

artilaginous Fish

There is a paucity of information on the ontogeny ofhe GEP system in Chondrichthyes. The broadestescription is that from six embryos of dogfish at theame stage of development, i.e., about one-half the sizef newborn animals (El-Salhy, 1984). Islet-like clusterst this time contained B and D cells, with A cellsresent among the exocrine acini. The intestine con-

ained abundant glucagon and a few insulin, somato-tatin, and PP/PYY-immunoreactive cells; the stomachossessed only a few somatostatin-immunoreactiveells. At 4 months of development of S. stellaris, theifferentiated intestine and the undifferentiated stom-ch possess glucagon and somatostatin; glucagon ap-ears first in the gastric glands before the pyloricrypts (Tagliafierro et al., 1989b). According to the latteruthors, the early appearance of these peptides, andthers (for example, vasoactive intestinal peptide,agliafierro et al., 1988), is support for the view thategulatory peptides play a role in cell growth andifferentiation. Recently, there have been extensivetudies on the ontogeny of NPY within the GEP systemnd vitellointestinal duct (VID) of embryos of Scyliorhi-us torazame (Chiba et al., 1995; Chiba, 1998). Immuno-eactivity for NPY was detected in the sac-like pan-reas at the earliest embryonic stage, 15 mm long.ubsequently, the cells increased in number and even-ually formed clumps like those in adults. As withther peptides in previous studies on elasmobranchmbryos, the intestinal immunoreactivity for NPY
ppeared before that in the stomach; however, gastro- m

ntestinal localization followed that of the pancreas.he NPY and serotonin immunoreactivity in the VID is

ransitory and may indicate that this organ serves as anmportant endocrine organ regulating gut function inhe embryo (Chiba, 1998).

gnathans

We know very little about both embryonic andosthatched development and growth of hagfish (Gorb-an, 1997). Thus, much of what can be stated about

ntogeny of the EP system is based on observationsrom adults (Ostberg, 1976). Barrington (1945) origi-ally proposed that cells in the epithelium of the bileuct of hagfish proliferated and matured into islet

ollicles which eventually became isolated from thepithelium (Fig. 3c). This view was supported bybservations of immunoreactivity for insulin in thepithelia of both the bile duct and the islet organOstberg et al., 1975, 1976). It was suggested at thatime that the ontogenetic history of the insulin cells inhe duct epithelium should be examined (Ostberg,976), but such an investigation has not taken place.omatostatin immunoreactivity has been shown inbout 1% of the islet cells but the ontogeny of this cellr the many other endocrine cells of the intestine (Vanoorden, 1990) awaits investigation. Given that this is

he first adult islet organ in vertebrate phylogeny suchstudy would seem to be very important (Youson,

999).The life cycle of the lamprey lends itself well for

evelopmental investigations from the fertilized egg,hrough embryogenesis and a protracted free-livingarval interval, metamorphosis to a juvenile, a second

etamorphosis during sexual maturation, senescence,nd death after spawning (Youson, 1985). There haseen a recent review of the GEP system throughoutuch of this life cycle (Youson, 1999). Despite the fact

hat embryos and newly hatched larvae are readilyvailable, there has been only cursory attention tohese intervals (Youson and Cheung, 1990). A study ofhis nature is important in light of the fact that thearval islet organ is a one-hormone endocrine gland,he simplest among the vertebrates (see Fig. 3b andarlier discussion). There is also an intimacy of theslets with clusters of cells within the intestinal epithe-ium and only some of these immunostain for insulin;

any of the intraepithelial clusters immunostain for

PftdswottltseiehdcdtiednspTb1

omwptcppiamsclfpt

ti1pebwpdlrHtdipa11fpobTo(ctietcrooel

ttftfial


P-family peptides (Cheung et al., 1991b). Since PP-amily peptides do not appear in the islets, thisransitory immunoreactivity is reflecting a step in theifferentiation of islet B cells (Fig. 3b). If this is theequence of differentiation of larval islet tissue then itould be inconsistent with that observed in many

ther vertebrates, for PP-family peptides are amonghe last to appear (Berwert et al., 1995). It is importanto extend these observations to embryonic intervals inampreys, for peptides such as NPY have been showno appear in embryogenesis of the elasmobranch GEPystem (Chiba et al., 1995). The youngest larvae to bexamined in lampreys (35 days old) already showednsulin-immunoreactive cell clumps in the intestinalpithelium and submucosal area but also in the extra-epatic common bile duct (Youson, 1999). This latteristribution is interesting because, as noted above, Bells of the hagfish islet organ originate from the bileuct epithelium and it has always been assumed that

he intestinal epithelium of larvae yielded all of theslets (Barrington, 1945). Even younger larvae and/ormbryos would be useful in deciphering this step inevelopment which has both phylogenetic and ontoge-etic implications. The timing of appearance of somato-tatin immunoreactivity and whether glucagon ap-ears during development should also be examined.he earliest immunoreactivity for somatostatin haseen noted in a 79-day-old larva (Youson and Cheung,990).A later stage of ontogeny of the EP system is

bserved during lamprey metamorphosis. Unlike theetamorphoses described earlier in the bony fishes,here changes in the GEP system seem to be more of arogression of events which began during larval life,

he metamorphosis in lampreys involves a dramatichange in the islet organ(s) of the EP system. Therogressive development of the cranial and caudalrincipal islets through seven stages of metamorphosis

n the sea lamprey, P. marinus, has been investigated byutoradiography, routine electron microscopy, and im-unologically at both the light and electron micro-

copic levels (Elliott and Youson, 1987; 1993a,b). Theranial pancreas arises through proliferation of thearval islet organ and continued recruitment of cellsrom the intestinal epithelium. The B cells of the caudalrincipal islet (Fig. 1d) have been traced through a

ransdifferentiation (dedifferentiation/redifferentia- i

ion) of cells of the extrahepatic and part of thentrahepatic common bile duct (Elliott and Youson,993b). D cells appear about the midpoint of metamor-hosis but the timing of differentiation of F cells, whichventually sparsely populate the adult islet, has noteen examined. An unknown but transitory cell typeas identified early in development and may be therecursor cell to D and/or F cells. The method ofevelopment of the caudal principal islet in this

amprey, and other Northern Hemisphere species, iseminiscent of that occurring in the adult hagfish.owever, a difference lies in the fact that in lampreys

he bile duct does not persist beyond this stage ofevelopment. As noted earlier, if the bile duct is not

nvolved in production of islet tissue during metamor-hosis there is no caudal principal islet (Fig. 1e), suchs in the Southern Hemisphere species (Hilliard et al.,985; Youson and Cheung, 1990; Youson and Potter,993). Intrahepatic secondary islets arise through trans-ormation of bile ductular epithelia during metamor-hosis in some holarctic species and submucosal sec-ndary islets appear anywhere along the intestineetween the two principal islets (Youson et al., 1988).he submucosal secondary islets seem to form through-ut adult life, including just before they spawn and dieCheung et al., 1990; Youson and Cheung, 1990). Auriosity, which may be developmentally related, ishat B cells in the cranial principal islet have a slightmmunoreactivity to anti-anglerfish peptide Y (Cheungt al., 1991a). There appears to be a continual produc-ion of cells from the intestinal epithelium for theranial principal islet and perhaps this apparent cross-eactivity of antisera is actually reflecting the remnantsf a common precursor for all cell types. The retentionf such an ontogenetic character has been used toxplain the colocalization of insulin and PYY in aizard (Putti and Della Rossa, 1996).

There have been no studies specifically directedoward the ontogeny of the enteroendocrine cells ofhe lamprey EP system. No peptides of the glucagonamily have been detected in the larval intestine or inhe larval or adult islet organs. PP- and glucagon-amily peptides seem to colocalize in adult lampreyntestine (Cheung et al., 1991a). The time of appearancend origin of cells immunoreactive to these two fami-ies of peptides in the adult intestine are important, for
n other vertebrates peptides of these families have

be(

S

cbEdrHeapoainotimHsimepcc

isvtcdfm1rptlop

pwpicrplolos

wurfet1aiba

C

aebibedtmv(dfiTedo


CA

een implicated, either directly or indirectly, in differ-ntiation and growth during intestinal developmentUpchurch et al., 1996).

ummary of Fish GEP Ontogeny

There have been only a few studies of the endocrineells of both the pancreas and alimentary canal duringoth direct and indirect development of bony fishes.ven these few studies showed some variation inevelopmental sequences which likely are a furthereflection of the diversity of this group of vertebrates.owever, as in other vertebrates it seems that the

ndocrine cells are a direct product of the endodermnd/or its derivatives and not the neural crest or anyart of the neuroectoderm. Ventral and dorsal anlagenr corresponding regions of the primitive gut may playrole in directing the type of islet organ that is found

n each species of bony fish. The most recent ontoge-etic study of a bony fish GEP system suggests that anrdered sequence of appearance of insulin, somatosta-in, glucagon, and PP-family peptides in a developingslet organ is also reflecting the phylogenetic develop-

ent of the vertebrate islet organ (Berwert et al., 1995).owever, this sequence of peptides is not found in all

pecies of bony fish examined thus far. Maternalnsulin may play an important role during develop-

ent at the endotrophic phase of some species (Guyott al., 1998). In all species examined so far regulatoryeptide localization in the intestine consistently pre-edes that of the stomach which differentiates near theommencement of exotrophic behavior.

Information on the development of the GEP systemn cartilaginous fishes and an agnathan, the hagfish, ispotty due to the past limitations of sampling. Iniviparous elasmobrachs local glucagon and somatosta-in are critical in early development of the alimentaryanal which has intestinal differentiation preceding theevelopment of the stomach. There has been recent

ocus on NPY as being important to the early develop-ent of the GEP system of elasmobranchs (Chiba,

998). Data on the ontogeny of the hagfish EP system isestricted to observations of adults. Lampreys mayrovide the best opportunity to study the ontogeny of

he GEP system among fishes. Embryogenesis of theamprey EP system results in a one-hormone isletrgan in larvae. Larval life of up to 7 years duration
rovides the possibility of examining the continual o

roduction of islet tissue from cells in the intestinehich may be ontogenetic precursors. Metamorphosisroduces a cranial principal islet derived from the

ntestine, while in addition, in holarctic species, aaudal principal islet is derived from epithelium of theegressing common bile duct. Secondary islets areroduced throughout adult life. Postmetamorphic is-

ets are composed of three principal cell types and twof these types are appearing for the first time in

amprey islet tissue. The origin of these islet cells andf glucagon-immunoreactive cells in the intestinehould be a future consideration.

Research into the ontogeny of the GEP system in fishould benefit from the use of the markers now beingsed in mammalian GEP systems. For instance, theole of PTP-NP is particularly interesting, for one of theorms of this protein tyrosine phosphatase appearsarly in midgut dorsal epithelium, prior to the forma-ion of the pancreatic rudiment (Chiang and Flanagan,996). A second candidate marker is the hormonedrenomedullin, which in islet development of the rats at some time found in all the principal cell typesefore permanently residing in the F cells (Martınez etl., 1998).

ONCLUSION

At the beginning of this present decade, severalrticles appeared which emphasized that fish arexcellent model systems for studies in developmentaliology, neurobiology, and endocrinology and that

nvestigations with fish had already made major contri-utions to these general fields (Powers, 1989; Dickhofft al., 1990; Gorbman, 1990). At the same time a moreirect emphasis was made that experimental investiga-

ions of the endocrine pancreas of fish and its hor-ones are advantageous to a general understanding of

ertebrate pancreatic physiology and biochemistryPlisetskaya, 1990a,b). A retrospective view of this pastecade reveals continued advancements in the areas ofsh GEP morphology, physiology, and biochemistry.he present review has demonstrated and reinforcedarlier views that, because of the large number andiversity of species, fish are an important componentf any attempt to study the phylogenetic development
f a system within the vertebrate body. The GEP

sfigGihaaoi(osl

adecfiidwoocehTapsatipcfoeaitsrl

rfigrCPtisu(

togwftfca1Nm(rgletipwobwG

A

Satro


ystem of only a small token of approximately 25,000sh species have been studied. Despite this, someeneral phylogenetic and ontogenetic patterns in theEP system have been revealed. The one-hormone

slet organ in larval lampreys is followed by a two-ormone gland in hagfish, a three-hormone gland indult lampreys, and eventually, in almost all aquaticnd terrestrial gnathostomes, to a four-hormone isletrgan. The order of sequenced acquisition of hormones

n the islet organs during phylogenetic developmentinsulin, somatostatin, glucagon, and PP) mirrors theirntogenetic developmental sequence and has leadome to revisit Haeckel’s ‘‘ontogeny recapitulates phy-ogeny’’ (Berwert et al., 1995; Falkmer, 1995).

The distribution of the islet tissue with respect to itsssociation with exocrine pancreatic tissue and to itsegree of concentration has undergone considerablevolutionary pressure such that several arrangementsan be found among and within the major groups ofshes. In the largest group, the Teleostei, the tendency

s for the more generalized members to have a diffuseistribution of small islets within the exocrine acini,hereas the more derived members have their islet

rgans with one or more large principal islets with nor little exocrine acini associated. The functional signifi-ance of this apparent specialization is not known, forven the islet organ of adults of the ancient lampreysave principal islets independent of exocrine elements.here appears to be no functional relevance to thessociation of exocrine and endocrine tissues of theancreas (Epple and Brinn, 1986). Therefore, one couldpeculate that the most derived teleosts eventuallycquired, through parallel evolution and natural selec-ion, at least a portion of the pancreas, with principalslets which were similar in general form to thoseresent in an animal, the lamprey, with an ancient andonserved evolution. This leads to several questionsor future consideration. Does the condition of the isletrgan in derived teleosts suggest that the direction ofvolution of the islet organ in teleosts is toward a morencient condition? Is the more ancient condition of theslet organ the more specialized state? Does this meanhat the diffuse islet organ of mammals is the leastpecialized state which is reflected in many of the basalay-finned fishes, the cartilaginous fishes, and theobed-finned fishes?

At the midpoint of this decade, several excellent s

eviews appeared emphasizing the contribution of thesh data, particularly on insulin, neuropeptide Y, andlucagon, to the ontogenetic and phylogenetic histo-ies of the vertebrate GEP system (Falkmer, 1995;onlon, 1995; Plisetskaya, 1995; Larhammar, 1996;lisetskaya and Mommsen, 1996). The bridge between

he vertebrate GEP and the corresponding system innvertebrates, particularly the protochordates, had beenhortened considerably (Falkmer, 1995) and efforts tonite the two systems phylogenetically continue today

McRory and Sherwood, 1997; Reinecke et al., 1999).As the decade began there was general acceptance

hat there were no ontogenetic links between the cellsf the GEP and nervous systems. Now there are evenreater links established between these two systemsith respect to differential expression of genes coding

or members of several of families of regulatory pep-ides, particularly the somatostatin and the NPY/PPamilies. The fish literature on NPY, PP, and PYY hasontributed greatly to our knowledge of the evolution-ry history of the NPY/PP family of peptides (Conlon,995; Larhammar, 1996). As the decade terminates,PY is demonstrated as important in early develop-ent of the GEP system in both fish and mammals

Upchurch et al., 1996; Chiba, 1998). As a result ofesearch activity in the past 10 years on sea bass, theilthead sea bream, the turbot, the dogfish, and the

amprey we have a much better picture of the ontog-ny of the GEP system in fishes. However, if we wisho continue to present fish as models for experimentalnvestigation of regulatory peptides or to use theirrincipal islets as xenotransplants (Wright et al., 1998)e must increase the sample size and relate the

ntogenetic events to those occurring in higher verte-rates. These studies should also include the factorshich regulate the differentiation and growth of theEP system in fish.

CKNOWLEDGMENTS

This study was supported by Grant 5945 from the Naturalciences and Engineering Research Council of Canada to J.H.Y. Theuthors thank many past graduate students for their contributions tohese studies and the many comparative endocrinologists whoseesearch activity is cited in this review. The cooperation and opinionsf Drs. Robert Connes, August Epple, Sture Falkmer, Erika Pliset-
kaya, Manfred Reinecke, and James Wright were very much

atKsN2

R

A

A

A

A

A

A

A

A

A

A

B

B

B

B

B

B

B

B

B

C

C

C

C

C

C

C


CA

ppreciated. J.H.Y. thanks Dr. Frank Moore for the invitation to writehis review and Drs. Stacia Sower, Robert Dores, and Hiroshiawauchi, whose invitation to provide a presentation at a sympo-

ium, ‘‘Molecular Ancestry of Vertebrate Polypeptide Hormones andeuropeptides’’ (Waseda University, Tokyo, Japan, November 23–

4, 1997), was the stimulus for the review.

EFERENCES

bad, M. E., Agulleiro, B., and Rombout, J. H. W. M. (1986). Animmunocytochemical and ultrastructural study of the endocrinepancreas of Sparus auratus L. (Teleostei). Gen. Comp. Endocrinol. 64,1–12.bad, M. E., Peeze Binkhorst, F. M., Elbal, M. T., and Rombout, J. H.W. M. (1987). A comparative immunocytochemical study of thegastro-entero-pancreatic (GEP) endocrine system in a stomachlessand a stomach-containing teleost. Gen. Comp. Endocrinol. 66,123–136.bad, M. E., Taverne-Thiele, J. J., and Rombout, H. W. M. (1988).Immunocytochemical and ultrastructural characterization of coex-istence of pancreatic polypeptide and glucagon-like immunoreac-tivity in the pancreatic endocrine cells of Sparus auratus L.(Teleostei). Gen. Comp. Endocrinol. 70, 9–19.bad, M. E., Garcıa Ayala, A., Lozano, M. T., and Agulleiro, B. (1992).Somatostatin-14- and somatostatin-25-like peptides in pancreaticendocrine cells of Sparus aurataus (Teleost): A light and electronmicroscopic immunocytochemical study. Gen. Comp. Endocrinol.86, 445–452.gulleiro, B., Lozano, M. T., Abad, M. E., and Garcıa Hernandez, M.P. (1993). Electron-microscopic immunocytochemical study of theendocrine pancreas of sea bass (Dicentrarchus labrax). Cell TissueRes. 274, 303–314.gulleiro, B., Garcıa Hernandez, M. P., and Lozano, M. T. (1994).Ontogeny of the endocrine pancreas in sea bass (Dicentrarchuslabrax): an ultrastructural study. II. The big and secondary islets.Cell Tissue Res. 276, 323–331.l-Mahrouki, A., and Youson, J. H. (1998). Immunohistochemicalstudies of the endocrine cells within the gastro-entero-pancreaticsystem of osteoglossomorpha: An ancient teleostean group. Gen.Comp. Endocrinol. 110, 125–139.ndrew, A. (1976). An experimental investigation into the possibleneural crest origin of pancreatic APUD (islet) cells. J. Embryol. Exp.Morphol. 35, 577–593.ndrew, A. (1984). The development of the gastro-entero-pancreaticneuroendocrine system in birds. In ‘‘Evolution and TumourPathology of the Neuroendocrine System’’ (S. Falkmer, R. Hakan-son, and F. Sundler, Eds.), pp. 91–109. Elsevier, Amsterdam.ndrews, P. C., Pollock, H. G., Elliott, W. M., Youson, J. H., andPlisetskaya, E. M. (1988). Isolation and characterization of avariant somatostatin-14 and two related somatostatins of 34 and37 residues from lamprey (Petromyzon marinus). J. Biol. Chem. 263,15809–15814.

aron, H. (1935). Insel-und Zymogengewebe in ihren gegenseitigen


Beziehungen bei Gasterosteus aculeatus und einigen anderen Teleeos-teirn. Z. Wiss. Zool. 146, 653–665.

arrington, E. J. W. (1945). The supposed pancreatic organs ofPetromyzon fluviatilis and Myxine glutinosa. Q. J. Microsc. Sci. 85,391–417.

arrington, E. J. W. (1972). The pancreas and intestine. In ‘‘TheBiology of Lampreys’’ (M. W. Hardisty and I. C. Potter, Eds.), Vol.2, pp. 135–169. Academic Press, London.

eccaria, C., Diaz, J.-P., Gabrion, J., and Connes, R. (1990). Matura-tion of the endocrine pancreas in the sea bass, Dicentrarchus labraxL. (Teleostei): An immunocytochemical and ultrastructural study.I. Glucagon-producing cells. Gen. Comp. Endocrinol. 78, 80–92.

erwert, L., Segner, H., and Reinecke, M. (1995). Ontogeny of IGF-1and the classical islet hormones in the turbot, Scophthalmusmaximus. Peptides 16, 113–122.

oquist, L., and Patent, G. (1971). The pancreatic islets of the teleostScorpaena scropha: An ultrastructural study with particular regardto fibrillar granules. Z. Zellforsch. 115, 416–425.

onner-Weir, S., and Weir, G. C. (1979). The organization of theendocrine pancreas: A hypothetical unifying view of the phyloge-netic differences. Gen. Comp. Endocrinol. 38, 28–37.

rinn, J. E., Jr. (1973). The pancreatic islets of bony fishes. Am. Zool.13, 653–665.

rinn, J. E., Jr. (1975). The pancreatic islet cytology of Ictaluridae(Teleostei). Cell Tissue Res. 162, 357–365.

arrillo, M., Zanuy, S., Duve, H., and Thorpe, A. (1986). Identifica-tion of hormone-producing cells of the endocrine pancreas of thesea bass, Dicentrarchus labrax, by ultrastructural immunocytochem-istry. Gen. Comp. Endocrinol. 61, 287–301.

han, S. J., Cao, Q.-P., and Steiner, D. F. (1990). Evolution of theinsulin superfamily: Cloning of a hybrid insulin/insulin-likegrowth factor cDNA from Amphioxus. Proc. Natl. Acad. Sci. USA 87,9319–9323.

heung, R., Plisetskaya, E. M., and Youson, J. H. (1990). Distributionof two forms of somatostatin in the brain, anterior intestine andpancreas of adult lampreys, Petromyzon marinus. Cell Tissue Res.262, 283–292.

heung, R., Andrews, P. C., Plisetskaya, E. M., and Youson, J. H.(1991a). Immunoreactivity to peptides belonging to the pancreaticpolypeptide family (NPY, aPY, PP, PYY) and to glucagon-likepeptide in the endocrine pancreas and anterior intestine of adultlampreys, Petromyzon marinus: An immunohistochemical study.Gen. Comp. Endocrinol. 81, 51–63.

heung, R., Ferreira, L. C. G., and Youson, J. H. (1991b). Distributionof two forms of somatostatin and peptides belonging to thepancreatic polypeptide family in tissues of larval lampreys,Petromyzon marinus L.: An immunohistochemical study. Gen.Comp. Endocrinol. 82, 93–102.

hiang, M.-K., and Flanagan, J. G. (1996). PTP-NP, a new member ofthe receptor protein tyrosine phosphatase family, implicated indevelopment of nervous system and pancreatic endocrine cells.Development 122, 2239–2250.

hiba, A. (1998). Ontogeny of serotonin-immunoreactive cells in thegut epithelium of the cloudy dogfish, Scyliorhinus torazame, withreference to coexistence of serotonin and neuropeptide Y. Gen.
Comp. Endocrinol. 111, 290–298.

C

C

C

C

C

C

C

C

C

C

D

D

D

D

E

E

E

E

E

E

E

E

E

E

E

F

F

F

F

F

F

F

F


hiba, A., Honma, Y., and Oka, S. (1995). Ontogenetic developmentof neuropeptide Y-like immunoreactive cells in the gastroentero-pancreatic endocrine system of the dogfish. Cell Tissue Res. 282,33–40.

imini, V., Van Noorden, S., and Polak, J. M. (1989). Co-localizationof substance P-, bombesin- and peptide histidine isoleucine(PHI)-like peptides in gut endocrine cells of the dogfish Scyliorhi-nus stellaris. Anat. Embryol. 179, 605–614.

loutier, R., and Ahlberg, P. K. (1996). Morphology, characters, andthe interrelationships of basal sarcopterygians. In ‘‘Interrelation-ships of Fishes’’ (M. L. J. Stiassny, L. R. Parenti, and G. D. Johnson,Eds.), pp. 445–479. Academic Press, San Diego.

onlon, J. M. (1995). Peptide tyrosine-tyrosine (PYY)-An evolution-ary perspective. Am. Zool. 35, 466–473.

onlon, J. M., Dfgard, E., Falkmer, S., and Thim, L. (1987). Aglucagon-like peptide, structurally related to mammalian oxynto-modulin, from the pancreas of a holocephalan fish, Hydrolaguscolliei. Biochem. J. 245, 851–855.

onlon, J. M., Askensten, U., Falkmer, S., and Thim, L. (1988).Primary structures of somatostatins from the islet organ of thehagfish suggest an anomalous pathway of posttranslational pro-cessing of prosomatostatin-1. Endocrinology 122, 1855–1859.

onlon, J. M., Bondareva, V., Rusakov, Y., Plisetskaya, E. M.,Mynarcik, D. C., and Whittaker, J. (1995a). Characterization ofinsulin, glucagon, and somatostatin from the river lamprey,Lampetra fluviatilis. Gen. Comp. Endocrinol. 100, 96–105.

onlon, J. M., Nielsen, P. F., Youson, J. H., and Potter, I. C. (1995b).Proinsulin and somatostatin from the islet organ of the southern-hemisphere lamprey Geotria australis. Gen. Comp. Endocrinol. 100,413–422.

onlon, J. M., Fan, H., and Fritzsh, B. (1998). Purification andstructural characterization of insulin and glucagon from the bichirPolypterus senegalis (Actinopterygii: Polypteriformes). Gen. Comp.Endocrinol. 109, 86–93.

onnes, R., and Benhalima, K. (1984). Ultrastructure de l’intestin duloup Dicentrarchus labrax L. au cours du developpement larvaire.Bull. Soc. Zool. France 109, 19–33.e Pinna, M. C. C. (1996). Teleostean monophyly. In ‘‘Interrelation-ships of Fishes’’ (M. L. J. Stiassny, L. R. Parenti, and G. D. Johnson,Eds.), pp. 147–162. Academic Press, San Diego.iaz, J.-P., Connes, R., Divanach, P., and Barnabe, G. (1989). Devel-oppement du foie et du pancreas du loup, Dicentrarchus labrax: I.Etude de la mise en place des organes au microscope electroniquea balayage. Ann. Sci. Nat. Zool. Paris 10, 87–98.ickoff, W. W., Brown, C. L., Sullivan, C. V., and Bern, H. A. (1990).Fish and amphibian model for developmental endocrinology. J.Exp. Zool. Suppl. 4, 90–97.ubois, P. M. (1989). Ontogeny of the endocrine pancreas. Horm. Res.32, 53–60.

lbal, M. T., Garcia Ayala, A., and Agulleiro, B. (1991). Light andelectron microscopic immunocytochemical demonstration of thecoexistence of somatostatin 14- and somatostatin 25-like peptidesin endocrine cells of the stomach of Sparus aurata (Teleost). Gen.Comp. Endocrinol. 84, 36–43.

lliott, M. W., and Youson, J. H. (1987). Immunohistochemical
demonstration of the development of the endocrine pancreas G
during metamorphosis of the sea lamprey, Petromyzon marinus L.Cell Tissue Res. 247, 351–357.

lliott, M. W., and Youson, J. H. (1993a). Development of the adultendocrine pancreas during metamorphosis in the sea lamprey,Petromyzon marinus L. I. Light microscopy and autoradiography.Anat. Rec. 237, 259–270.

lliott, M. W., and Youson, J. H. (1993b). Development of the adultendocrine pancreas during metamorphosis in the sea lamprey,Petromyzon marinus L. II. Electron microscopy and immunocyto-chemistry. Anat. Rec. 237, 271–290.

l-Salhy, M. (1984). Immunocytochemical investigation of the gastro-enteropancreatic (GEP) neurohormonal peptides in the pancreasand gastrointestinal tract of the dogfish Squalus acanthias. Histochem-istry 80, 193–205.

pple, A. (1967). Further observations on amphiphil cells in thepancreatic islets. Gen. Comp. Endocrinol. 9, 137–142.

pple, A. (1969). The endocrine pancreas. In ‘‘Fish Physiology’’ (W. S.Hoar and D. J. Randall, Eds.), Vol. II, pp. 275–319. Academic Press,New York.

pple, A., and Brinn, J. E. (1975). Islet histophysiology: Evolutionarycorrelations. Gen. Comp. Endocrinol. 27, 320–349.

pple, A., and Brinn, J. E. (1986). Pancreatic islets. In ‘‘VertebrateEndocrinology: Fundamentals and Biochemical Implications’’(P. K. T. Pang and M. Schreibman, Eds.), Vol. 1, pp. 279–317.Academic Press, New York.

pple, A., and Brinn, J. E. (1987). ‘‘The Comparative Physiology ofthe Pancreatic Islets,’’ p. 223. Springer-Verlag, Heidelberg.

pple, A., and Lewis, T. L. (1973). Comparative histophysiology ofthe pancreatic islets. Am. Zool. 13, 567–590.

alkmer, S. (1985a). Comparative morphology of pancreatic islets inanimals. In ‘‘The Diabetic Pancreas’’ (B. W. Volk and E. R. Arquilla,Eds.), pp. 17–52. Plenum, New York.

alkmer, S. (1985b). Phylogenetic aspects of the brain-gut axis, withspecial reference to islet hormones in invertebrates and lowervertebrates. In ‘‘Neurosecretion and the Biology of Neuropep-tides’’ (H. Kobayashi, Ed.), pp. 317–325. Japan Sci. Soc. Press,Tokyo. [Springer-Verlag, Berlin].

alkmer, S. (1995). Origin of the parenchymal cells of the endocrinepancreas: Some phylogenetic and ontogenetic aspects. In ‘‘Endo-crine Tumors of the Pancreas: Frontiers in Gastrointestinal Re-search’’ (M. Mignon and R. T. Jensen, Eds.), Vol. 23, pp. 2–29.Karger, Basel.

alkmer, S., and Patent, G. J. (1972). Comparative and embryologicalaspects of the pancreatic islets. In ‘‘Handbook of Physiology, Vol.1, The Endocrine Pancreas’’ (D. F. Steiner and N. Freinkel, Eds.),pp. 1–23. Williams & Wilkins, Baltimore.

alkmer, S., and Van Noorden, S. (1983). Ontogeny and phylogeny ofthe glucagon cell. Handb. Exp. Pharmacol. 66, 81–119.

orey, P., and Janvier, P. (1994). Evolution of the early vertebrates.Am. Sci. 82, 554–565.

ujita, T. (1962). Uber das Inselsystem des Pankreas von Chimaeramonstrosa. Z. Zellforsch. 57, 487–494.

ujita, T. (1989). Present status of paraneuron concept. Arch. Histol.Cytol. 52(Suppl.), 1–8.
arcıa Hernandez, M. P., and Agulleiro, B. (1992). Ontogeny of the

G

G

G

G

G

G

G

G

GG

G

G

G

G

G

G

H

H

H

H

I

J

JJ

K

K

K

K

K

K

L

L

L

L


CA

endocrine pancreas in sea bass (Dicentrarchus labrax). An immuno-cytochemical study. Cell Tissue Res. 270, 339–352.arcıa Hernandez, M. P., Lozano, M. T., and Agulleiro, B. (1994a).Ontogeny of the endocrine pancreas in sea bass (Dicentrarchuslabrax): an ultrastructural study. I. The primordial cord and theprimitive, single primordial islets. Cell Tissue Res. 276, 309–322.arcıa Hernandez, M. P., Lozano, M. T., an dAgulleiro, B. (1994b).Ontogeny of some endocrine cells of the digestive tract in sea bass(Dicentrarchus labrax): An immunocytochemical study. Cell TissueRes. 277, 373–383.ardiner, B. G. (1993). Osteichthyes: basal actinopterygians. In ‘‘TheFossil Record’’ (M. J. Benton, Ed.), Vol. 2, pp. 611–619. Chapman &Hall, London.ardiner, B. G., Maisey, J. G., and Littlewood, D. T. J. (1996).Interrelationships of basal neopterygians. In ‘‘Interrelationships ofFishes’’ (M. L. J. Stiassny, L. R. Parenti, and G. D. Johnson, Eds.),pp. 117–146. Academic Press, San Diego.illett, E., Cake, M. H., Potter, I. C., and Tajbakhsh, M. (1996).Compact exocrine pancreas of ammocoetes of the southern hemi-sphere lamprey Mordacia mordax contains a trypsin inhibitor:Putative evolutionary considerations. J. Exp. Zool. 274, 227–233.omez-Visus, I., Abad, M. E., Garcıa-Hernandez, M. P., and Agul-leiro, B. (1996). Occurrence of somatostatin and insulin immunore-activities cells in the stomach of sea bass (Dicentrarchus labrax L.):Light and electron microscopic studies. Gen. Comp. Endocrinol. 102,16–27.omez-Visus, I., Garcıa-Hernandez, M. P., Lozano, M. T., andAgulleiro, B. (1998). Glucagon- and NPY-related peptide-immuno-reactive cells in the gut of sea bass (Dicentrarchus labrax L.): A lightand electron microscopic study. Gen. Comp. Endocrinol. 112, 26–37.orbman, A. (1990). Cyclostome models for speculation on evolu-tion of vertebrate endocrine systems. J. Exp. Zool. Suppl. 4, 137–141.orbman, A. (1997). Hagfish development. Zool. Sci. 14, 375–390.ould, S. J. (1977). ‘‘Ontogeny and Phylogeny.’’ Belknap-HarvardUniv. Press, Cambridge, MA.rande, L., and Bemis, W. E. (1996). Interrelationships of Acipenseri-formes with comments on ‘‘Chondrostei.’’ In ‘‘Interrelationshipsof Fishes’’ (M. L. J. Stiassny, L. R. Parenti, and G. D. Johnson, Eds.),pp. 85–115. Academic Press, San Diego.roff, K. E., and Youson, J. H. (1997). An immunohistochemicalstudy of the endocrine cells within the pancreas, intestine andstomach of the gar (Lepisosteus osseus). Gen. Comp. Endocrinol. 106,1–16.roff, K. E., and Youson, J. H. (1998). Fine structure and immunocy-tochemistry of cells within the endocrine pancreas of the gar(Lepisosteus osseus). Can. J. Zool. 76, 6–18.rossner, D. (1968). Das Inselorgan des Crossopterygiers Latimeriachalumnae J. L. B. Smith. Z. Zellforsch. 84, 417–428.uyot, E., Diaz, J. P., and Connes, R. (1995). Organogenesis of liver insea bream, Sparus aurata: J. Fish Biol. 47, 427–437.uyot, E., Diaz, J. P., Romestand, B., and Connes, R. (1998). Insulinduring the early postembryonic development of the gilt-head seabream, Sparus aurata: Ultrastructural, immunohistochemical, andbiochemical studies. Gen. Comp. Endocrinol. 110, 147–156.ansen, G. N., Hansen, B. L., and Jorgensen, P. N. (1987). Insulin-,
glucagon- and somatostatin-like immunoreactivity in the endo- L

crine pancreas of the lungfish, Neoceratodus forsteri. Cell Tissue Res.248, 181–185.ashimoto, T., Kawano, H., Daikoku, S., Shima, K., Taniguchi, H.,and Baba, S. (1988). Transient coappearance of glucagon andinsulin in the progenitor cells of the rat pancreatic islets. Anat.Embryol. 178, 489–497.ildebrand, M. (1995). ‘‘Analysis of Vertebrate Structure,’’ 4th ed.Wiley, New York.illiard, R. W., Epple, A., and Potter, I. C. (1985). The morphologyand histology of the endocrine pancreas of the southern hemi-sphere lamprey, Geotria australis Gray. J. Morphol. 184, 253–261.

rwin, D. M., Huner, O., and Youson, J. H. (1999). Lamprey progluca-gon and the origin of glucagon-like peptides. Mol. Biol. Evol., [inpress].

ohnson, D. E., Torrence, J. L., Elde, R. P., Bauer, G. E., Noe, B. D., andFletcher, D. J. (1976). Immunohistochemical localization of somato-statin, insulin and glucagon in the principal islets of the anglerfish(Lophius americanus) and the channel catfish (Ictalurus punctata).Am. J. Anat. 147, 119–124.

ollie, M. (1973). ‘‘Chordate Morphology.’’ Krieger, Huntington, NY.onsson, A.-C. (1991). Regulatory peptides in the pancreas of two

species of elasmobranchs and in the Brockmann bodies of fourteleost species. Cell Tissue Res. 266, 163–172.

lein, C., and Lange, R. H. (1977). Principal cell types in thepancreatic islet of a teleost fish, Xiphophorus helleri H. Cell TissueRes. 176, 529–551.

lein, C., and Van Noorden, S. (1980). Pancreatic polypeptide (PP)and glucagon cells in the pancreatic islet of Xiphophorus helleri H.(Teleostei). Correlative immunohistochemistry and electron micros-copy. Cell Tissue Res. 205, 187–198.

obayashi, K., and Ali, S. S. (1981). Cell types of the endocrinepancreas in the shark Schliorhinus stellaris as revealed by correla-tive light and electron microscopy. Cell Tissue Res. 215, 475–490.

obayashi, K., and Takahashi, Y. (1970). Light and electron micro-scopic observations on the islets of Langerhans in Carassiuscarassius longsdorfii. Arch. Histol. Jpn. 31, 433–454.

obayashi, K., and Takahashi, Y. (1974). Fine structure of Langer-hans’ islet cells in a marine teleost Conger japonicus Bleeker. Gen.Comp. Endocrinol. 23, 1–18.

obayashi, K., Shibasaki, S., and Takahashi, Y. (1976). Light andelectron microscopic study on the endocrine cells of the pancreasin a marine teleost Fugu rubripes rubripes. Cell Tissue Res. 174,161–182.

anger, M., Van Noorden, S., Polak, J. M., and Pearse, A. G. E. (1979).Peptide hormone-like immunoreactivity in the gastrointestinaltract and endocrine pancreas of eleven teleost species. Cell TissueRes. 199, 493–508.

arhammar, D. (1996). Evolution of neuropeptide Y, peptide YY andpancreatic polypeptide. Regul. Pept. 62, 1–11.

ecointre, G., and Nelson, G. (1996). Clupeomorpha, sister-group ofOstariophysi. In ‘‘Interrelationships of Fishes’’ (M. L. J. Stiassny, L.R. Parenti, and G. D. Johnson, Eds.), pp. 193–207. Academic Press,San Diego.

e Douarin, N. M. (1988). On the origin of pancreatic endocrine cells.Cell 52, 196–171.

i, G.-Q., and Wilson, M. V. H. (1996). Phylogeny of Osteoglossomor-

L

L

L

M

M

M

M

M

M

M

M

N

NN

N

N

O

O

O

P

P

P

P

P

P

P

P

P

P

P

PP

R

R

R


pha. In ‘‘Interrelationships of Fishes’’ (M. L. J. Stiassny, L. R.Parenti, and G. D. Johnson, Eds.), pp. 163–174. Academic Press,San Diego.

’Hermite, A., Ferrano, R., Dubois, M. P., and Andersen, A. C. (1985).Detection of endocrine cells by immunofluorescence method inthe gastroenteropancreatic system of the adult eel, glass eel, andleptocephalic larva (Anguilla anguilla L.). Gen. Comp. Endocrinol. 58,347–359.

ozano, M. T., Ayala, A. G., Abad, M. E., and Agulleiro, B. (1991a).Pancreatic endocrine cells in sea bass (Dicentrarchus labrax L.):I. Immunocytochemical characterization of glucagon- and PP-related peptides. Gen. Comp. Endocrinol. 81, 187–197.

ozano, M. T., Ayala, A. G., Abad, M. E., and Agulleiro, B. (1991b).Pancreatic endocrine cells in sea bass (Dicentrarchus labrax L.):II. Immunocytochemical study of insulin and somatostatin pep-tides. Gen. Comp. Endocrinol. 81, 198–206.aglio, M., and Putti, R. (1998). Morphological basis of the interac-tion between endocrine cell types in the pancreatic islets of theteleost, Blennius gattoruggine. Tissue Cell 30, 672–683.azzi, B. (1976). Note sul pancreas endocrino del polipteriformeCalamoichthys calabarcius. Atti. Acad. Sci. Torino Cl. Sci. Fis., Mat.Nat. 110, 387–392.artinez, A., Weaver, C., Lopez, J., Bhathena, S. J., Elsasser, T. H.,Miller, M. J., Moody, T. W., Unsworth, E. J., and Cuttitta, F. (1996).Regulation of insulin secretion and blood glucose metabolism byadrenomedullin. Endocrinology 137, 2626–2632.artinez, A., Cuttitta, F., and Teitelman, G. (1998). Expressionpattern for adrenomedullin during pancreatic development in therat reveals a common precursor with other endocrine cells. CellTissue Res. 293, 95–100.cCormick, N. A. (1925). The distribution and structure of theislands of Langerhans in certain fresh-water and marine fishes.Trans. R. Can. Inst. 15, 57–81.cRory, J. E., and Sherwood, N. M. (1997). Ancient divergence ofinsulin and insulin-like growth factor. DNA Cell Biol. 16, 939–949.illot, J., and Anthony, J. (1972). Le pancreas des crossopterygienscoelacanthides. Z. Zellforch. 123, 215–223.ontuenga, L. M., Martinez, A., Miller, M. J., Unsworth, E. J., andCuttitta, F. (1997). Expression of adrenomedullin and its receptorduring embryogenesis suggests autocrine or paracrine modes ofaction. Endocrinology 138, 440–451.akamura, M., and Yokote, M. (1971). Ultrastructural studies on theislets of Langerhans of the carp. Z. Anat. Entwicklungsgesch. 134,61–72.elson, J. S. (1994). ‘‘Fishes of the World,’’ 3rd ed. Wiley, New York.oaillac-Depeyre, J., and Hollande, E. (1981). Evidence for somato-statin, gastrin and pancreatic polypeptide-like substances in themucosa cells of the gut in fishes with and without stomach. CellTissue Res. 216, 193–203.ozaki, M., Miyata, K., Oota, Y., Gorbman, A., and Plisetskaya, E. M.(1988a). Different cellular distributions of two somatostatins inbrain and pancreas of salmonids, and their associations withinsulin- and glucagon-secreting cells. Gen. Comp. Endocrinol. 69,267–280.ozaki, M., Miyata, K., Oota, Y., Gorbman, A., and Plisetskaya, E. M.
(1988b). Colocalization of glucagon-like peptide and glucagon
immunoreactivities in pancreatic islets and intestine of salmonids.Cell Tissue Res. 253, 371–375.

¨ stberg, Y. (1976). ‘‘The Entero-Insular Endocrine Organ in aCyclostome, Myxine glutinosa,’’ pp. 1–41. Medical dissertation,Department of Pathology, Umeå University, Umeå, Sweden.stberg, Y., Van Noorden, S., and Pearse, A. G. E. (1975). Cytochemi-cal, immunofluorescence and ultrastructural investigations onpolypeptide hormone localization in the islet parenchyma and bileduct mucosa of a cyclostome, Myxine glutinosa. Gen. Comp.Endocrinol. 25, 274–291.stberg, Y., Boquist, L., Van Noorden, S., and Pearse, A. G. E. (1976).On the origin of islet cells in a cyclostome, Myxine glutinosa. Gen.Comp. Endocrinol. 28, 228–246.

atent, G. J., Kechele, P. O., and Tomichek Carrano, V. (1978).Nonconventional innervation of the pancreatic islets of the teleostfish, Gillichthys mirabilis. Cell Tissue Res. 191, 305–315.

earse, A. G. E. (1969). The cytochemistry and ultrastructure ofpolypeptide hormone-producing cells (the APUD series) and theembryonic, physiologic and pathologic implications of the con-cept. J. Histochem. Cytochem. 17, 303–313.

earse, A. G. E., and Polak, J. M. (1971). Neural crest origin on theendocrine polypeptide (APUD) cells of the gastrointestinal tractand pancreas. Gut 12, 783–788.

lisetskaya, E. M. (1989a). Physiology of fish endocrine pancreas.Fish Physiol. Biochem. 7, 39–48.

lisetskaya, E. M. (1989b). Pancreatic peptides. In ‘‘ComparativePhysiology of Regulatory Peptides’’ (S. Holmgren, Ed.), pp.174–202. Chapman & Hall, London.

lisetskaya, E. M. (1990a). Recent studies of fish pancreatic hor-mones: Selected topics. Zool. Sci. 7, 335–353.

lisetskaya, E. M. (1990b). Endocrine pancreas of teleost fish: Amodel for interaction of islet hormones. J. Exp. Zool. Suppl. 4,53–57.

lisetskaya, E. M. (1995). Peptides of insulin and glucagon superfami-lies in fish. Neth. J. Zool. 45, 181–188.

lisetskaya, E. M., and Mommsen, T. P. (1996). Glucagon andglucagon-like peptides in fishes. Int. Rev. Cytol. 168, 187–257.

lisetskaya, E. M., Andrews, P. C., Pollock, H. G., Elliott, W. M., andYouson, J. H. (1988). Isolation and structure of adult lamprey(Petromyzon marinus) insulin. Gen. Comp. Endocrinol. 69, 46–55.

otter, I. C., and Hilliard, R. W. (1987). A proposal for the functionaland phylogenetic significance of differences in the dentition oflampreys (Agnatha: Petromyzontiformes). J. Zool. 212, 713–737.

owers, D. A. (1989). Fish as model systems. Science 246, 352–357.utti, R., and Della Rossa, A. (1996). Peptide YY and insulin coexistin b-granules in B cells of the Madagascan lizard, Zonosauruslaticaudatus. Gen. Comp. Endocrinol. 103, 249–256.

afn, S., and Wingstrand, K. G. (1981). Structure of intestine,pancreas, and spleen of the Australian lungfish, Neoceratodusforsteri (Krefft). Zool. Scripta 10, 223–239.

asmussen, A.-S., and Arnason, U. (1999). Molecular studies suggestthat cartilaginous fishes have a terminal position in the piscinetree. Proc. Natl. Acad. Sci. USA 96, 2177–2182.

eddy, S., and Elliott, R. B. (1988). Ontogenic development ofpeptide hormones in the mammalian fetal pancreas. Experientia 44,
1–9.

R

R

R

R

R

R

R

R

R

R

R

R

R

RS

S

S

S

SS

S

S

S

S

S

T

T

T

T

T

T

T

T


CA

eddy, S., Elliott, R. B., Poole, C. A., and Ross, J. M. (1997).Double-label immunofluorescence study of glutamic acid decar-oxylase in the fetal and adult ovine pancreas by light and confocalmicroscopy: Evidence for predominant beta-cell coexpression.Gen. Comp. Endocrinol. 106, 301–309.

einecke, M. (1981). Immunohistochemical localization of polypep-tide hormones in endocrine cells of the digestive tract of Branchios-toma lanceolatum. Cell Tissue Res. 219, 445–456.

einecke, M., Drakenburg, K., Falkmer, S., and Sara, V. R. (1991).Presence of IGF-1-like peptides in the neuroendocrine system ofthe Atlantic hagfish, Myxine glutinosa (Cyclostomata): Evidencederived by chromotography, radioimmunoassay and immunohis-tochemistry. Histochemistry 96, 191–196.

einecke, M., Drakenberg, K., Falkmer, S., and Sara, V. R. (1992).Peptides related to insulin-like growth factor 1 in the gastro-entero-pancreatic system of bony and cartilaginous fish. Regul. Pep. 37,155–165.

einecke, M., Betzler, D., Drakenberg, K., Falkmer, S., and Sara, V. R.(1993a). Occurrence of members of the insulin superfamily incentral nervous system and digestive tract of protochordates.Histochemistry 99, 277–285.

einecke, M., Maake, C., Falkmer, S., and Sara, V. R. (1993b). Thebranching of insulin-like growth factor 1 and insulin: An immuno-histochemical analysis during phylogeny. Regl. Pept. 48, 65–76.

einecke, M., Weimar, E., Maake, C., Drakenberg, K., Falkmer, S.,and Sara, V. R. (1994). IGF-2-like peptides are present in insulincells of the elasmobranchian endocrine pancreas: An immunohis-tochemical and chromatographic study. Histochemistry 102, 365–371.

einecke, M., Muller, C., and Segner, H. (1997). An immunohisto-chemical analysis of the ontogeny, distribution and coexistence of12 regulatory peptides and serotonin in endocrine cells and nervefibers of the digestive tract of the turbot, Scophthalmus maximus(Teleostei). Anat. Embryol. 195, 87–102.

einecke, M., Eppler, E., David, I., and Georges, D. (1999). Immuno-histochemical evidence for the presence, localization and partialcoexistence of insulin, insulin-like growth factor I and relaxin inthe protochordate Ciona intestinalis. Cell Tissue Res. 295, 331–338.

enaud, C. B. (1997). Conservation status of northern hemispherelampreys (Petromyzontidae). J. Appl. Ichthyol. 13, 143–148.

ombout, J. H. W. H., Lamers, C. H. J., and Hanstede, J. G. (1978).Enteroendocrine APUD cells in the digestive tract of larval Barbusconchonius (Teleostei, Cyprinidae). J. Embryol. Exp. Morphol. 47,121–135.

ombout, J. H. W. H., Rademakers, L. H. P. M., and van Hess, J. P.(1979). Pancreatic endocrine cells of Barbus conchonius (Teleostei,Cyprinidae). Cell Tissue Res. 203, 9–23.

ombout, J. H. W. H., and Taverne-Thiele, J. J. (1982). An immunocy-tochemical and electron-microscopical study of endocrine cells inthe gut and pancreas of a stomachless teleost fish, Barbus concho-nius (Cyprinidae). Cell Tissue Res. 207, 577–593.

omer, A. S. (1970). ‘‘The Vertebrate Body.’’ Saunders, Philadelphia.cheuermann, D. W., Adriaensen, D., Timmermans, J.-P., andDe Groodt-Lasseel, M. H. A. (1991). Immunohistochemical localiza-tion of polypeptide hormones in pancreatic endocrine cells of a
dipnoan fish, Protopterus aethiopicus. Acta Histochem. 91, 185–192.

egner, H., Storch, V., Reinecke, M., Kloas, W., and Hanke, W. (1995).A tabular overview of organogenesis in larval turbot (Scophthal-mus maximus L.). ICES mar. Sci. Symp. 201, 35–39.

ekine, Y., and Yui, R. (1981). Immunohistochemical study of thepancreatic endocrine cells of the ray, Dasyatis akajei. Arch. Histol.Jpn. 44, 95–101.

chaeffer, B. (1981). The xenacanth shark neurocranium, with com-ments of elasmobranch monophyly. Bull. Am. Mus. Nat. Hist. 169,1–66.

iwe, S. A. (1926). Pancreasstudien. Morphol. Jahrb. 57, 84–307.mith, H. M. (1960). ‘‘Evolution of Chordate Structure. An Introduc-tion to Comparative Anatomy.’’ Holt, Rinehart & Winston, NewYork.

ower, S. A., Chiang, Y.-C., and Conlon, J. M. (1994). Polygenicexpression of somatostatin in lamprey. Peptides 15, 151–154.

tefan, Y., and Falkmer, S. (1980). Identification of four endocrine celltypes in the pancreas of Cottus scorpius (Teleostei) by immunofluo-rescence and electron microscopy. Gen. Comp. Endocrinol. 42,171–178.

tefan, Y., Ravazzola, M., and Orci, L. (1981). Primitive islets containtwo populations of cells with differing glucagon immunoreactiv-ity. Diabetes 30, 192–195.

t-Onge, L., Sosa-Pineda, B., Chowdhury, K., Mansouri, A., andGruss, P. (1997). Pax 6 is required for differentiation of glucagon-producing a-cells in mouse pancreas. Nature 387, 406–409.

trahan, R., and Maclean, J. L. (1969). A pancreas-like organ in thelarva of the lamprey Mordacia mordax. Aust. J. Sci. 32, 54–55.

agliafierro, G., Bonini, E., Faraldi, G., Farina, L., and Rossi, G. G.(1988). Distribution and ontogeny of VIP-like immunoreactivity inthe gastro-entero-pancreatic system of a cartilaginous fish Scyliorhi-nus stellaris. Cell Tissue Res. 253, 23–28.

agliafierro, G., Farina, L., Faraldi, G., Rossi, G. G., and Vacchi, M.(1989a). Distribution of somatostatin and glucagon immunoreac-tive cells in the gastric mucosa of some cartilaginous fishes. Gen.Comp. Endocrinol. 75, 1–9.

agliafierro, G., Rossi, G. G., Bonini, E., Faraldi, G., and Farina, L.(1989b). Ontogeny and differentiation of regulatory peptide- andserotonin-immunoreactivity in the gastrointestinal tract of anelasmobranch. J. Exp. Zool. Suppl. 2, 165–174.

agliafierro, M., Carlini, M., Faraldi, G., Morescalchi, A. M., Putti, R.,Della Rossa, A., Fasulo, S., and Mauceri, A. (1996). Immunocyto-chemical detection of islet hormones in the digestive system ofProtopterus annectens. Gen. Comp. Endocrinol. 102, 288–298.

eitelman, G. (1990). Insulin cells of pancreas extend neurites but donot arise from the neuroectoderm. Dev. Biol. 142, 368–379.

eitelman, G., Lee, J. K., and Alpert, S. (1987). Expression of celltype-specific markers during pancreatic development in the mouse:implications for pancreatic cell lineages. Cell Tissue Res. 250,435–439.

eitelman, G., Alpert, S., Polak, J. M., Martinez, A., and Hanahan, D.(1993). Precursor cells of mouse endocrine pancreas coexpressinsulin, glucagon and the neuronal protein tyrosine hydroxylaseand neuropeptide Y, but not pancreatic polypeptide. Development118, 1031–1039.

homas, T. B. (1940). Islet tissue in the pancreas of the elasmo-
branchii. Anat. Rec. 76, 1–17.

U

U

V

V

V

W

W

W

W

W

W

W

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Z


pchurch, B. H., Aponte, G. W., and Leiter, A. B. (1994). Expressionof peptide YY in all four islet cell types in the developing mousepancreas suggests a common peptide YY-producing progenitor.Development 120, 245–252.pchurch, B. H., Fung, B. P., Rindi, G., Ronco, A., and Leiter, A. B.(1996). Peptide YY expression is an early event in colonic endo-crine cell differentiation: Evidence from normal and transgenicmice. Development 122, 1157–1163.

an Noorden, S. (1984). The neuroendocrine system in Protostomianand deuterostomian invertebrates and lower vertebrates. In ‘‘Evo-lution and Tumour Pathology of the Neuroendocrine System’’(S. Falkmer, R. Hakanson, and F. Sundler, Eds.), pp. 8–38. Elsevier,Amsterdam.

an Noorden, S. (1990). Gut hormones in cyclostomes. Fish Physiol.Biochem. 5, 399–408.

ortsman, A. (1948). The development of the pancreas, the gallbladder, the ductus choledochus and the air bladder of Esox lucius.Biol. Jahrb. Belg. 15, 87–105.agner, G. F., and McKeown, B. A. (1981). Immunocytochemicallocalization of hormone-producing cells within the pancreaticislets of the rainbow trout (Salmo gairdneri). Cell Tissue Res. 221,181–192.agner, G. F., McKeown, B. A., and Popham, D. J. (1981). Theautoradiographic localization of zinc with the pancreatic islets ofthe rainbow trout (Salmo gairdneri). Histochemistry 72, 113–121.ang, Y.-Q., Plisetskaya, E., Baskin, D. G., and Gorbman, A. (1986).Immunocytochemical study of the pancreatic islets of the Pacificsalmon, Oncorhynchuys kisutch. Zool. Sci. 3, 123–129.ang, Y., Nielsen, P. F., Youson, J. H., Potter, I. C., Lance, V. A., andConlon, J. M. (1999a). Molecular evolution of peptide tyrosine-tyrosine: primary structure of PYY from the lampreys Geotriaaustralis and Lampetra fluviatilis, bichir, python and desert tortoise.Regul. Pept. 79, 103–108.ang, Y., Nielsen, P. F., Youson, J. H., Potter, I. C., and Conlon, J. M.(1999b). Multiple forms of glucagon and somatostatin isolatedfrom the intestine of the southern-hemisphere lamprey Geotriaaustralis. Gen. Comp. Endocrinol. 113, 274–282.eisel, G. F. (1972). Anatomy and histology of the digestive systemof the paddlefish (Polyodon spathula). J. Morphol. 140, 243–256.right, J. R., Jr., Yang, H., and Dooley, K. C. (1998). Tilapia—A sourceof hypoxia-resistant islet cells for encapsulation. Cell Transpl. 7,299–307.

ang, H., and Wright, J. R., Jr. (1995). A method for mass harvestingislets (Brockmann bodies) from teleost fish. Cell Transpl. 4, 621–628.

ang, H., Morrison, C. M., Conlon, J. M., Laybolt, K., and Wright, J.R., Jr. (1999). Immunocytochemical characterization of the pancre-atic islet cells of the Nile tilapia (Oreochromis niloticus). Gen. Comp.Endocrinol. 114, 47–56.

ouson, J. H. (1981). The alimentary canal. In ‘‘The Biology ofLampreys’’ (M. W. Hardisty and I. C. Potter, Eds.), Vol. 3, pp.95–189. Academic Press, London.

ouson, J. H. (1985). Organ development and specialization inlamprey species. In ‘‘Evolutionary Biology of Primitive Fishes’’ (R.E. Foreman, A. Gorbman, J. M. Dodd, and R. Olsson, Eds.), pp.141–164. Pergamon, Oxford.

ouson, J. H. (1988). First metamorphosis. In ‘‘Fish Physiology’’(W. Hoar and D. J. Randall, Eds.), Vol. llB, pp. 135–196. AcademicPress, New York.

ouson, J. H. (1999). The agnathan enteropancreatic endocrinesystem: Phylogenetic and ontogenetic histories, structure, andfunction. [submitted for publication].

ouson, J. H., and Cheung, R. (1990). Morphogenesis of somatosta-tin- and insulin-secreting cells in the lamprey endocrine pancreas.Fish Physiol. Biochem. 8, 389–397.

ouson, J. H., and Elliott, W. M. (1989). Morphogenesis and distribu-tion of the endocrine pancreas in adult lampreys. Fish Physiol.Biochem. 7, 125–131.

ouson, J. H., and Potter, I. C. (1993). An immunohistochemicalstudy of enteropancreatic endocrine cells in larvae and juveniles ofthe southern hemisphere lampreys, Geotria australis and Mordaciamordax. Gen. Comp. Endocrinol. 92, 151–167.

ouson, J. H., Elliott, W. M., Beamish, R. J., and Wang, D. W. (1988). Acomparison of endocrine pancreatic tissue in adults of four speciesof lampreys in British Columbia: A morphological and immunohis-tochemical study. Gen. Comp. Endocrinol. 70, 247–261.

ui, R., Shimada, M., and Fujita, T. (1990). Immunohistochemicalstudies on peptide- and amine-containing endocrine cells andnerves in the gut and the rectal gland of the ratfish Chimaeramonstrosa. Cell Tissue Res. 260, 193–201.

abel, M., Surdyk-Zasada, J., Lesisz, I., Jagoda, E., Wysocka, T.,Seidel, J., and Grzeszkowiak, J. (1994). Immunocytochemicalstudies on pancreatic endocrine cells at early stages of develop-ment of the pig. Folia Histochem. Cytochem. 32, 181–185.


Documents

Ontogenetic and Phylogenetic Development of the Endocrine