Development of the Adenohypophysis in theLamprey: Evolution of Epigenetic PatterningPrograms in Organogenesis

KATSUHISA UCHIDAn1, YASUNORI MURAKAMI1, SHIGEHIRO KURAKU1, 2,SHIGEKI HIRANO3, and SHIGERU KURATANI11Laboratory for Evolutionary Morphology, Center for Developmental Biology,RIKEN, Kobe, Hyogo 650-0047, Japan2Department of Biophysics, Graduate School of Science, Kyoto University, Kyoto606-8501, Japan3Faculty of Medicine, Niigata University, Niigata 951-8510, Japan

ABSTRACT In gnathostomes, the adenohypophysis, a component of the hypothalamo-hypophysial complex, is believed to develop through hierarchically organized epigenetic interactionsbased primarily on the topographical relationships between tissues. From a comparison ofdevelopmental processes and gene expression patterns of pituitary-related genes between theagnathan species, lampreys and gnathostomes, we speculate on the evolutionary pathway of thevertebrate adenohypophysis. In the lamprey, this is derived from the nasohypophysial placode(NHP) that develops anterior to the oral ectoderm. The NHP can be identified by the expression ofLjPitxA, before actual histogenesis, but it is initially distant from the future hypothalamic region.Subsequently, the NHP expresses both LjFgf8/17 and LjBmp2/4a gene transcripts, and growscaudally to establish a de novo contact with the hypothalamic region by the mid-pharyngula stage.Later, the NHP gives rise to both the adenohypophysis and an unpaired nasal organ. Thus, thetopographical relationship between the NHP and the hypothalamic region is established secondarilyin the lamprey, unlike gnathostomes in which the equivalent relationship appears early indevelopment. Comparing the developmental pattern of the amphioxus homologue of theadenohypophysis, we hypothesize that a modification of the regulation of the growth factor encodinggene lies behind the evolutionary changes recognized as heterochrony and heterotopy, which leads tothe gnathostome hypophysial developmental pattern. J. Exp. Zool. (Mol. Dev. Evol.) 300B:32–47,2003. r 2003, Wiley-Liss, Inc.

INTRODUCTION

The pituitary gland is an endocrine organspecific to vertebrates. Integrating signals fromthe hypothalamus and periphery, this gland playscentral roles in the regulation of vital processesinvolved in homeostasis, metabolism, growth andreproduction (Bentley, ’98). It consists of twodistinct parts, the anterior adenohypophysis, andthe posterior neurohypophysis. The adenohypo-physis is a derivative of the ectodermal placodeoriginating from the oral ectoderm, whereas theneurohypophysis is derived from the part ofventral forebrain defined as the infundibulum(Baker and Bronner-Fraser, 2001; Etchevers et al.,2001). Fate-mapping data in embryos of variousvertebrates species have demonstrated that theoral ectoderm region that differentiates into the

adenohypophysis is derived from the anteriorneural ridge (ANR; Couly and Le Douarin, ’85;Eagleson et al., ’86; Osumi-Yamashita et al., ’94;Kouki et al., 2001; Whitlock et al., 2003), onceinduced as the neural crest during early embry-ogenesis (Couly et al., ’92). Thus, development ofthe pituitary gland can be viewed as a part of headpatterning, which is potentially associated withthe evolution of the vertebrate head.

In gnathostomes, the developmental sequence ofthe adenohypophysis proceeds in several distinct

Grant Sponsor: Grants-in-Aid to S. K. from the Ministry ofEducation, Science and Culture of Japan.

nCorrespondence to: Katsuhisa Uchida, Laboratory for Evolution-ary Morphology, Center for Developmental Biology, RIKEN, 2-2-3,Minatojima-Minami, Chuo-ku, Kobe, Hyogo 650-0047, Japan.E-mail: k-uchida@cdb.riken.go.jp

Received 25 July 2003; Accepted 2 October 2003Published online in Wiley InterScience (www.interscienc.wiley.

com). DOI: 10.1002/jez.b. 00044

r 2003 WILEY-LISS, INC.

JOURNAL OF EXPERIMENTAL ZOOLOGY (MOL DEV EVOL) 300B:32–47 (2003)

steps (Burgess et al., 2002). It begins afterdisplacement of the midline ANR cells by thegrowth of the forebrain, followed by specificationof the ectoderm and pouch formation in thethickened oral ectoderm (Kawamura et al.,2002). This pouch rudiment, or Rathke’s pouch,invaginates dorsally towards the hypothalamus,and is completely detached from the oral cavitywith the closure of the ventral pouch (Treier andRosenfeld, ’96; Baker and Bronner-Fraser, 2001).Rathke’s pouch gives rise to the adenohypophysis,and the infundibulum differentiates into theneurohypophysis.Recent studies have unraveled general inductive

events for pituitary organogenesis, based on thetopographical relationship defining the interac-tions between the neural and oral ectoderm. Thesecreted factor Shh, emanating from the entireventral diencephalon and oral ectoderm, has acrucial role in the placodal induction and prolif-eration of pouch cells (Treier et al., 2001; Sbrognaet al., 2003). BMP4 and FGF activities from theventral diencephalon are also involved in theformation of the definitive Rathke’s pouch (Eric-son et al., ’98; Takuma et al., ’98). Subsequently, aBMP2 gradient together with an opposing dorso-ventral FGF8 gradient provides polarity to theadenohypophysial primordium; these signalinggradients establish overlapping expression pat-terns of pituitary-related transcription factors aspositional cues enabling adenohypophysial celltypes to differentiate (see review by Scully andRosenfeld, 2002). Hypophysial patterning in themouse is thus based on the hierarchically orga-nized series of epigenetic events that proceedsequentially during development, and topographi-cal relationships of tissues are particularly im-portant for these processes. Such a developmentalprogram raises an intriguing question about itsevolution. When and how has this exquisitedevelopmental system been acquired in phyloge-netic evolution? Solving this problem appears torequire us first to understand the possible primi-tive state of hypophysial development in verte-brates.In the agnathans, the jawless vertebrates, the

pituitary gland is in close vicinity to the hypotha-lamus and is involved in distinct physiologicalfunctions (Gorbman and Tamarin, ’85; Sower andKawauchi, 2001). As in gnathostomes, the glandconsists of three distinct parts, also containingseveral adenohypophysis-specific cell types (Patz-ner et al., ’82; Nozaki et al., 2001), which suggeststhat the hypothalamo-adenohypophysial axes and

their functions might be conserved in the ag-nathans and gnathostomes. Nevertheless, thedevelopmental processes of the agnathans havenot been well studied.

The adenohypophysis in the hagfish originatesfrom the dorsal part of the nasopharyngealepithelium, which is first identified as a thickeningof the archenteric roof in early development(Gorbman, ’83). In the lamprey, the ectoderm onthe ventral aspect of the embryonic forebrainforms an adenohypophysial primordium known asthe nasohypophysial placode (NHP: Gorbman,’95). The posterior part of these epitheliumdifferentiates as the adenohypophysis of thehagfish and lamprey, being the part that actuallycontacts the neurohypophysis, whereas the ante-rior part of the epithelium differentiates early intothe unpaired median nasal epithelium. Thus, thedevelopment of the adenohypophysis in theagnathans is closely related to olfactory organformation, as suggested by embryonic mappingdata in gnathostomes (Couly et al., ’92). Becausethe nasal placode develops in a pair and indepen-dent from the hypophysis in gnathostomes, therewas apparently a systematic topographical changein the developmental plan of this system in theevolutionary transition from agnathan to gnathos-tome. Although this implies a shift in tissueinteractions behind this change, little informationis available on the molecular developmentalmechanisms of the adenohypophysis in agnathans.Thus, it is first necessary to describe the differ-ences in the developmental plans of the pituitarybetween agnathans and gnathostomes. Thepresent paper is thus intended to outline thedevelopmental sequence of the lamprey adenohy-pophysis at the molecular level, and to discuss theevolution of its developmental program in thecontext of vertebrate head evolution.

MATERIALS AND METHODS

Embryos and larvae

Mature male and female lampreys, Lethenteronjaponicum, and ammocoete larvae were collectedin a tributary of the Miomote River, Niigata,Japan, during the breeding season (mid-May) in2002. The eggs were artificially fertilized and keptin 10% Steinberg solution (Steinberg, ’57) at201C. Embryonic stages were assessed morpholo-gically according to Tahara (’88). For histologicalobservation, embryos and larvae were fixed inBouin’s fixative overnight. For whole mountin situ hybridization, embryos were fixed in 4%

EVOLUTION OF PITUITARY DEVELOPMENT 33

paraformaldehyde and 1% methanol in 0.1 Mphosphate buffered saline (PFAM/PBS, pH 7.4).Fixed embryos were dehydrated using a gradedmethanol series and stored at –201C in 100%methanol.

Histology

Specimens were embedded in paraplast. Serialsagittal sections (6 mm) were cut and mounted onglass slides coated with egg white glycerin. Thesections were stained with Carrazzi’s hematoxylin(Wako, Tokyo, Japan) for the observation ofembryos, and with Mayer’s hematoxylin (Wako)and eosin (Sigma, St Louis, MO) for larvae.Stained sections were observed using a lightmicroscope (Olympus BX60, Tokyo).

Isolation of lamprey regulatory genes

We examined the expression patterns of pitui-tary-related genes in lamprey embryos. LjFgf8/17,LjBmp2/4a (Shigetani et al., 2002), LjPax6 (Mur-akami et al., 2001), LjTTF-1 (Ogasawara et al.,2001) and LjOtxA (Ueki et al., ’98) were selected.Partial sequences of lamprey Pitx and hedgehog(Hh) homologs were isolated by reverse transcrip-tion–polymerase chain reaction (RT–PCR) usingtotal RNA extracted from L. japonicum stage20–30 embryos as a template. The lamprey PitxcDNA fragment was amplified using a set ofspecific primers: 50-CAACTGTCGTCGTGTTCT-GA-30 (sense) and 50-CTCGAGTTGCACG-TATCCCT-30 (antisense). The designs of theseprimers were based on the previously identifiednucleotide sequences of the PitxA genes of Petro-myzon marinus and L. planeri (PmPitxA andLpPitxA, Boorman and Shimeld, 2002a). A cDNAfragment of lamprey Hh homolog was amplifiedusing degenerate primers designed on the basis ofthe amino acid sequences of the vertebrate sonichedgehog (Shh) as follows. A sense primer corre-sponding to amino acids KLTPLAYKQ (50-TAGCTGACSCCNYTNGCNTAYAARCA-30) andan antisense primer corresponding to amino acidsAGFDWV(Y/F)YE (50-TCATARTRNACCCART-CRAANCCNGC-30) were used. Amplified cDNAfragments were cloned in a pCRII-TOPO vector(Invitrogen, Carlsbad, CA), and sequenced usingan autosequencer (ABI 3100, Applied Biosystems,Tokyo).

Sequence analysis

After we had confirmed the distinct homology ofnucleotide sequences of lamprey Pitx and Hh

homologs to gnathostome members of Pitx andthe hedgehog family using the BLASTX program,respectively (Altschul et al., ’97), we carried outmultiple alignments of the deduced amino acidsequences of lamprey Pitx and Hh homologs withsequences already reported for other animals. Amolecular phylogenetic tree of the Hh family wasinferred using a distance matrix based on the JTTmodel (Jones et al., ’92) and using the neighbor-joining (Saitou and Nei, ’87) and maximum-like-lihood methods (Felsenstein, ’81; Kishino et al.,’90), and with rate heterogeneity among sitestaken into account. The partial sequences oflamprey Pitx andHh homologs have been assignedto DDBJ/EMBL/GenBank Accession NumbersAB124585 and AB124584, respectively.

Whole mount in situ hybridization

The plasmids containing cDNA of lampreypituitary-related genes were digested with appro-priate restriction enzymes. Antisense and senseRNA probes were generated by in vitro transcrip-tion using a Digoxigenin (DIG)-RNA labeling kit(Boehringer Mannheim, Germany) according tothe manufacturer’s protocol. Whole mount in situhybridization was performed as described inMurakami et al. (2001).

RESULTS

Development of the lampreyadenohypophysis

First, the morphology of the adenohypophysiswas analyzed histologically in the fully grownammocoete larvae. The ammocoete adenohypo-physis is located at the ventral side of thediencephalon, and is in a close association withthe olfactory organ, which opens to the top of thehead through the external nostril (Fig. 1A). As inthe teleost, the adenohypophysis consists of threedistinct parts, the rostral and proximal parsdistalis, and the pars intermedia. The pars inter-media makes close contact with the neurohypo-physis dorsally (Fig. 1B, C). The morphology of thelamprey adenohypophysis, however, appears tohave been established secondarily in embryogen-esis; histological analysis of a series of embryosshowed dynamic changes in the topographicalrelationships between the forebrain, oral ecto-derm, and nasohypophysial plate (Fig. 2).

At stage 22, corresponding to the late neurula,the forebrain and foregut with the first pharyngealpouch are already visible, whereas the definite

K. UCHIDA ET AL.34

nasohypophysial placode (NHP) is not yet identifi-able in the rostral ectoderm covering the forebrain(Fig. 2A, B, I). In stage 24 embryos, which havejust hatched, the oral cavity is identified as aventral depression in the ectoderm (Fig. 2C, D, I).A thickening is first visible on the ventralectoderm anterior to the mouth opening, indicat-ing the initial development of NHP. Althoughclose contact is observed between the NHP andthe ventral forebrain, this site is not the definitivepart of the brain that appears to induce theadenohypophysis of the lamprey (Fig. 2C, D; seebelow). The brain has not yet achieved morpholo-

gical differentiation at this stage, and structuressuch as the epiphysis, optic chiasma and mid-hindbrain boundary could not be discerned.

By stage 26, the upper and lower lips, velum,epiphysis, optic chiasma and mid-hindbrainboundary could be identified by their typical

Fig. 1. Pituitary gland of an ammocoete larva. A. Mid-sagittal section stained with hematoxylin and eosin. The larvaladenohypophysis (arrowheads) is located at the ventral side ofthe diencephalon caudal to the nasal epithelium. Anterior is tothe left. Arrows indicate the remnant of the nasohypophysialduct. B, C. High magnification (B) and a schematic diagram(C). The adenohypophysis consists of the rostral (red) andproximal (blue) pars distalis (rpd and ppd), and the parsintermedia (pi, yellow) that make close contacts with theneurohypophysis (nh, black). Abbreviations: ne, nasal epithe-lium; on, olfactory nerve; ep, epiphysis; hpt, hypothalamus; ch,optic chiasma; me, median eminence.

Fig. 2. Development of the lamprey hypophysis: histologi-cal observations. Histological sections stained with hematox-ylin (A-H) and schematic diagram (I). Anterior is to the left forall the embryos. The panels on the left (A, C, E, G) show lowmagnification views of the embryonic and larval head, and onthe middle (B, D, F, H) high magnification micrographs of thenasohypophysial placode (NHP). A, B. Stage 22 embryo. Theforebrain and foregut are visible, whereas the NHP is not yetvisible. C, D. Stage 24. The oral cavity is formed and the NHPhas begun to form as a thickened ectoderm on the ventralaspect of the head (arrowheads). E, F. Stage 26. NHP isobserved as a fold of ectoderm where close contact isestablished with the ventral forebrain. The NHP has growna caudal process of epithelial cells that extend caudal to thelevel of the optic chiasma (arrowheads). G, H. Stage 30. Theupper lip has lifted dorsally and the well-developed olfactoryepithelium, an NHP derivative, is now in close contact withthe forebrain. The posterior part of the NHP that willdifferentiate into the adenohypophysis (arrowheads) is stillseen as a cord of epithelial cells behind the optic chiasma. I.Schematic diagram showing the NHP development. Thedeveloping NHP is colored pink. The anterior part of theNHP differentiates into the nasal epithelium (ne) and theposterior into the adenohypophysis (pt), respectively. Abbre-viations: n, notochord; mo, mouth; ulp, upper lip; llp, lower lip;ph, pharynx; fb, forebrain; mb; midbrain; hb, hindbrain; tel,telencephalon; hpt, hypothalamus; ch, optic chiasma; ep,epiphysis; nhp, nasohypophysial placode; np, nasal placode;ne, nasal epithelium; oc, oral cavity; fg, foregut; vel, velum; pc,prechordal plate; pm, prechordal mesoderm.

EVOLUTION OF PITUITARY DEVELOPMENT 35

morphological features (Fig. 2E, F, I). The upperlip has grown rostrally beneath the NHP, foldingthe surface ectoderm, and extending caudally toestablish a contact with the post optic level of thediencephalon (Fig. 2F, I). The latter part of theforebrain corresponds to the later hypothalamus.From its topographical relation to the rostral tip ofthe notochord, the NHP at this stage appeared tohave grown more caudally than that of theprevious stage.In the stage 30 embryo, the placode-derived

nasal duct has developed more posteriorly thanthe dorsorostral growth of the upper lip (Fig. 2G,H, I). The rostral part of the NHP differentiates asthe olfactory organ composed of a thick columnarepithelium, covering the rostral aspect of thetelencephalon, whereas the posterior as the futureadenohypophysis, consisting of an epithelium of afew cell layers, extends caudal to the level of theoptic chiasma (Fig. 2H, arrowheads).

Pituitary-related genes in lampreyembryos

Using cDNA that was reverse transcribed fromtotal RNA extracted from the stage 20–30 embryo,we succeeded in isolating a 600 bp-long Pitx cDNAfragment. This nucleotide sequence was comparedwith those of PmPitxA and LpPitxA (data notshown, Boorman and Shimeld, 2002a), and weconcluded that this cDNA fragment is derivedfrom an ortholog of PmPitxA and LpPitxA. Boor-man and Shimeld (2002a) were also unable toassign the lamprey PitxA sequences as orthologs tognathostome Pitx1 or Pitx2 with confidence. Thus,we named this clone LjPitxA (L. japonicum PitxA).We also succeeded in isolating a 358 bp-long Hh

cDNA fragment. The deduced amino acid se-quence of this nucleotide sequence showed strik-ing homology to members of the Hh gene familyalready reported from various animals (Fig. 3A).Subsequent molecular phylogenetic inference wasperformed using neighbor-joining, as well asmaximum-likelihood methods (Fig. 3B). In thehedgehog family, three distinct clusters, sonichedgehog (Shh), indian hedgehog (Ihh) and deserthedgehog (Dhh), have been identified in gnatho-stomes to date (Kumar et al., ’96; Meyer andSchartl, ’99). In our tree, LjHh showed a phyloge-netic affinity to Shh and Ihh, but not to Dhh.However, as is always the case for phylogenetictree analyses of other gene families includingcyclostome sequences (Suga et al., ’99; Kurataniet al., 2002), we could not obtain a robust tree to

support the orthology of LjHh to either of thesesubfamilies. Accordingly, we have named our Hhclone LjHhA (L. japonicum HhA). LjHhA is thefirst member of the hedgehog family reported fromcyclostomes, and further efforts to identify an-other member of the family in this animal groupwill be necessary.

To clarify the involvement of transcriptionfactors in adenohypophysial and olfactory organo-genesis, we examined the expression patterns ofLjPitxA, LjTTF-1, LjPax6 and LjOtxA. Thesegenes potentially serve as markers for specifiedstructures in embryos that are involved inpituitary differentiation (Ueki et al., ’98; Muraka-mi et al., 2001; Ogasawara et al., 2001; Boormanand Shimeld, 2002a).

At pre-pharyngula stages (before Tahara’s stage22; Fig. 4A, B), LjPitxA is expressed in theepithelial ectoderm, corresponding to the futureoral ectoderm. In hatching larvae, LjPitxA is alsoexpressed in the ectoderm around the mouth (Fig.4C). By stage 25, the expression domain of LjPitxAis divided into two ectodermal regions, NHP andoral ectoderm, with each being separated by thedeveloping upper lips that do not express this gene(Fig. 4D). LjPitxA expression is also visible in theposterior brain, anterior to the mid-hindbrainboundary, and pharyngeal endoderm in stage 25embryos. The level of expression becomes moreintense with development and the expressiondomain in the posterior NHP expands towardsthe level corresponding to the rostral tip of thenotochord (Figs. 4E, F). The transcripts are notobserved in the nasal placode per se. By stage 27,the placodal expression has become weak, but isstill restricted within the posterior part of theplacode (Figs. 4G, H).

At stage 22, LjTTF-1 transcripts are alreadyweakly expressed in the ventral diencephalon(Fig. 5A). At stage 25 and 26, this expressiondomain is posterior to the optic chiasma (Figs. 5D,G). Thus, TTF-1 gene expression in the lampreyalso appears to specify the future hypothalamicregion, as in gnathostome embryos. Importantly,the posterior NHP first meets this part of theforebrain at stage 26 as noted above (Figs. 2F, 4F).LjPax6 and LjOtxA are both expressed in theanterior forebrain, whereas no surface ectodermalexpression could be seen at stage 22 (Figs. 5B, C).At stage 25, LjPax6 is expressed in the entire NHP(Fig. 5E), whereas LjOtxA expression is observedonly in its anterior part (Fig. 5F). LjPax6transcripts are identified in the entire NHP atstage 26 (Fig. 5H), whereas LjOtxA expression is

K. UCHIDA ET AL.36

now restricted to the anterior part of the NHP,corresponding to the nasal epithelium (Fig. 5I).To focus on the inductive events involved in the

development of the lamprey adenohypophysis at amolecular level, we further examined the expres-sion patterns of the signaling molecule-encodinggenes LjFgf8/17 and LjBmp2/4a. These are puta-

tive lamprey orthologs for gnathostome Fgf8,Bmp2 and Bmp4, which are involved in gnatho-stome adenohypophysial development (Shigetaniet al., 2002). We also examined the expressionpatterns of the isolated LjHhA to clarify theinvolvement of this molecule in lamprey adenohy-pophysis development.

Fig. 3. Sequence and phylogenetic analyses of hedgehogfamily including LjHhA. A. Deduced amino acid sequence ofLjHhA aligned with sequences already reported from variousgnathostomes and invertebrates. Only amino acid sites wherefragment sequence of LjHhA can be aligned are shown. Aminoacid sequence of LjHhA is highlighted in bold character.

Amino acid numbers are indicated at the both ends of eachsequence. B. Molecular phylogenetic tree of hedgehog familyincluding LjHhA. This tree is inferred with neighbor-joiningmethod using 119 amino acid sites where fragment sequenceof LjHhA can be aligned. For the methods to infer the tree, seeMaterials and Methods.

EVOLUTION OF PITUITARY DEVELOPMENT 37

Fig. 4. Expression patterns of LjPitxA in lamprey embryos.A. Stage 19. LjPitxA transcripts are identified in the ectodermcovering the forebrain and foregut. B. Stage 21. LjPitxA isexpressed in a single ectodermal domain corresponding to thestomodeal ectoderm and in the ventral forebrain. C. Stage 23.LjPitxA expression in the ectoderm covers the future NHPand oral ectoderm (oe). D. Stage 25. LjPitxA expression in theectoderm has been divided into two distinct regions: the NHP(arrowhead) and the oral ectoderm. E, F. Stage 26. Expressionof LjPitxA has become intense and the expression domaincorresponding to the caudal NHP (arrowheads) expandscaudally, whereas the anterior nasal placode (np) does notshow any expression. G, H. Stage 27. Expression in the NHP(arrowheads) has become weak, but is still restricted withinthe posterior region. Note that the LjPitxA-positive epithelialcells now extend caudal to the optic chiasma. F, H: highermagnification views. Anterior is to the left for all the embryos.Abbreviations: see Fig. 2.

K. UCHIDA ET AL.38

Within the domain of ectodermal expression ofLjPitxA, the growth factor encoding genes areupregulated at relatively late stages of develop-ment. At stage 22, LjFgf8/17 transcripts arepresent in the part of the anterior ectoderm thatcorresponds to the future NHP (Fig. 6A). At stage25, LjFgf8/17 transcripts are present anterior partof the NHP (Fig. 6B, lower arrowheads). Inaddition to the NHP domain, LjFgf8/17 is alsoexpressed in the ventral telencephalon covered bythe rostral part of the NHP at this stage (Fig. 6B,upper arrowhead). In the stage 27 embryo,expression in the NHP becomes weak, whereasthat in the ventral preoptic forebrain remains athigh levels (Fig. 6C).LjBmp2/4a transcripts are present in the epi-

dermal ectoderm, which is folding posteriorlybeneath the NHP at stage 25, (Fig. 6D). Other-wise, LjBmp2/4a expression is found in theprechordal plate. In slightly older embryos, theexpression domain of LjBmp2/4a is clearly re-stricted within the posterior part of the ectodermfolding beneath the NHP and the expressiondomain has expanded posteriorly (Fig. 6E). NoLjBmp2/4a expression domain, however, is foundin the anterior nasal placode. In later develop-mental stages, LjBmp2/4a transcripts are foundonly in the pharyngeal endoderm, whereas itsexpression in the posterior NHP has been lost(Fig. 6F).Expression of LjHhA was associated with the

ventral neural tube. In the forebrain, however,transcripts of this gene were not found at earlystages (data not shown). Slightly before hatching,very faint expression of LjHhA is observed in thefloor plate of the neural tube and the ventralforebrain (Fig. 6G). At hatching, LjHhA isstrongly expressed in the floor plate and a partof the ventral diencephalon corresponding to thezona limitans intrathalamica (Fig. 6H). In addi-tion, faint expression is also seen in the hypotha-lamus, which closely contacts the developing NHP(Fig. 6H). By stage 26, no substantial changes are

seen in the expression patterns of LjHhA com-pared with stage 24 (Fig. 6I). These expressionpatterns in the floor plate and ventral forebrainare similar to the Shh expression pattern observedin the central nervous system (CNS) of gnathos-tomes (Ericson et al., ’95). Although no robustorthology of LjHhA to either of three hedgehogsubfamilies can be ascribed with confidence at amolecular level, LjHhA appears to be the putativecognate of gnathostome Shh in terms of itsexpression patterns.

DISCUSSION

Here we have described the developmentalpattern of the lamprey adenohypophysis at bothmorphological and molecular levels. Although thelamprey genome appears to contain the basic set ofregulatory genes required for normal patterningof this organ system, the genes were not expressedin the patterns that comparative embryologybetween this animal and gnathostome wouldpredict. Because the embryonic inductiveevent normally relies on specific topographicalrelationships of cells, such a discrepancy maypossibly provide an important key to the questionof homology of embryonic structures (morpholo-gical homology) and homology of genes (processhomology; Gilbert and Bolker, 2001).

Developmental patterns and processes ofthe vertebrate adenohypophysis

The pituitary gland, consisting of the adenohy-pophysis and the neurohypophysis, is found in allvertebrates, but not in invertebrates. In gnatho-stomes, it is derived from the rostral portion of theANR (Couly and Le Douarin, ’85). Once induced asthe neural crest, this part of the ectoderm givesrise to the cephalic surface ectoderm including theoral ectoderm, a portion of which is later specifiedas Rathke’s pouch (Rathke, 1838). This interactswith the ventral part of the hypothalamic primor-dium, known as the infundibulum. In all the

Fig. 5. Expression of regulatory genes related to hypo-physial development in lamprey embryos. Expression pat-terns of LjTTF-1 (A, D, G), LjPax6 (B, E, H), and LjOtxA (C,F, I), in high magnification micrographs of the nasohypophy-sial region. Anterior is to the left. A, B, C. Stage 22. LjTTF-1transcripts are weakly expressed in the ventral forebrain.LjPax6 and LjOtxA are both expressed in the anteriorforebrain, whereas no ectodermal expression can be seen atthis stage. D, E, F. Stage 25. LjTTF-1 transcripts are presentin the ventral diencephalon caudal to the optic chiasma, and

LjPax6 is expressed in the entire NHP (E: arrowheads).LjOtxA expression is observed only in the anterior part ofNHP (F, arrowheads). G, H, I. Stage 26. LjPax6 transcriptsare identified in the anterior part of the dorsal forebrain andeye (opt). In addition, transcripts are identified in the NHP(H: arrowheads). LjOtxA mRNA is expressed in the anteriorneural tube, and the expression domain is restricted to theanterior part of the NHP, corresponding to the nasal placode(I, arrowheads). Abbreviations: see Fig. 2.

EVOLUTION OF PITUITARY DEVELOPMENT 39

Fig. 7. Summary of expression patterns of pituitary-related genes. The expression patterns of the genes examinedhere are shown in two schemes for each developmental stage.From top to bottom: late neurula, early pharyngula, and post-hatching larva. In the late neurula, although the definitiveNHP is not detected, PitxA expression appears to specify therostral surface ectoderm including the future oral ectoderm.The future hypothalamus expresses TTF-1. Fgf8/17 expres-sion is present in the anterior NHP. At the pharyngula stage,the NHP is visible in the ectoderm anterior to the oral cavity.In the post-hatching larva, the PitxA expression domain hassplit into two regions, of which NHP is specified in the rostraldomain. The NHP has grown caudally by this stage to makecontact with the definitive hypothalamic region specified byTTF-1. The NHP at this stage characteristically expressesboth Fgf8/17 and Bmp2/4a, and the anterior part is identifiedby OtxA expression, while the whole NHP is specified by Pax6.Note that the topographical relationship between the NHPand the ventral diencephalon is established as a secondaryprocess in the lamprey.

K. UCHIDA ET AL.40

examined gnathostomes, the early patterning ofthe head ectoderm is similar, and the topographi-cal relationships between Rathke’s pouch and thehypothalamic anlage is established early in devel-opment, and maintained subsequently (Gleiber-man et al., ’99; Baker and Bronner-Fraser, 2001).Exceptionally, adenohypophysial anlage of teleostis not formed via a structure equivalent toRathke’s pouch, but that appears as a solidstructure that is formed at the anterior edge ofthe head (Dubois and ElAmraoui, ’95; Herzoget al., 2003). However, elasmobranches and stur-geon, which is considered as the basal groups offish, have common morphological feature of theRathke’s pouch of amniotes (Balfour, 1878; Sew-ertzoff, ’28; Gorbman and Bern, ’62; Kurataniet al., 2000). Thus, it seems most likely thatpituitary developmental pattern represents sec-ondary modification introduced into the teleostlineage.Close contact between the Rathke’s pouch and

the hypothalamus is central in the signalingsystem involved in adenohypophysial develop-ment, and a number of signaling moleculesfunction in the inductive events based on this(Dasen and Rosenfeld, 2001). After specification ofthe oral ectoderm by Shh, Pitx1 and Pitx2 geneexpressions, BMP4 derived from the hypothalamicanlage induces the formation of Rathke’s pouch(Ericson et al., ’98; Treier et al., ’98). Subse-quently, FGF8 and BMP4 are both derived fromthe hypothalamic anlage to induce the growth ofthe pouch. Simultaneously, BMP2 in Rathke’spouch, together with the counteracting FGF8 inthe diencephalon constitute gradients of signalingmolecules, which serve as the prepattern ofadenohypophysial cell type distribution (Scullyand Rosenfeld, 2002). All through these develop-mental events, the ventral diencephalon expressesthe homeobox gene, TTF-1, the marker for the

embryonic hypothalamus (Watkins-Chow andCamper, ’98; Sheng and Westphal, ’99). This geneexpression is also an important component inadenohypophysial development (Kimura et al.,’96). Thus, the adenohypophysial patterning inthe gnathostome depends on an early-fixed topo-graphical pattern of embryonic tissues on whichsignaling cascades are exerted as an hierarchicallyorganized epigenetic system.

In the present study, some of the developmentalelements found in gnathostome adenohypophysialpatterning appeared to be conserved in lampreydevelopment. As already reported, for example,LjPax6 is expressed rather ubiquitously in theearly NHP, whereas LjOtxA gene expression isrestricted to the anterior part of this placode,corresponding to the future nasal epithelium(Figs. 5, 7; see also Ueki et al., ’98, and Murakamiet al., 2001). In mouse embryos, Pax6 is alsoexpressed in the oral ectoderm, and the expressionbecomes secondarily restricted to Rathke’s pouch(Kioussi et al., ’99). Murine Otx2 is also requiredfor the same sets of placodes (Matsuo et al., ’95;Mallamaci et al., ’96; Acampora et al., 2000). Thus,these two transcription factor encoding genes mayhave similar developmental functions in thelamprey and mouse (Fig. 8, see below for evolu-tionary consideration).

LjPitxA expression is also seen in the rostralectoderm of the early lamprey embryo (Figs. 4, 7),implying a role in the regional specification of therostral ectoderm in this species, as in the mouse(Dasen and Rosenfeld, 2001). In later develop-ment, however, the LjPitxA expression domain inthe lamprey splits into two regions, the NHP andthe oral ectoderm, and only the former is theninvolved in hypophysial development. In gnathos-tome development, on the other hand, Pitx1 andPitx2 expressions are initially the markers forstomodeal ectoderm, and the same expression

Fig. 6. Expression patterns of signaling molecules relatedto hypophysial development in lamprey embryos. Expressionpatterns of LjFgf8/17 (A, B, C), LjBmp2/4a (D, E, F), andLjHhA (G, H, I), in high magnification micrographs of thenasohypophysial region of embryos. A. Stage 22. A low level ofLjFgf8/17 transcripts is seen in the anterior ectodermcorresponding to the future NHP (arrowhead). B. Stage 25.LjFgf8/17 is expressed at low levels in the NHP (lowerarrowheads), as well as in the ventral portion of the forebrain(upper arrowhead). C. Stage 27. Expression in the NHP hasbecome weak, whereas that in the ventral brain remainsvisible (arrowheads). D. Stage 25. LjBmp2/4a expression ispresent in the posterior part of the ectoderm folding beneaththe NHP (arrowheads), and in the prechordal plate (pc), lowerand upper lip, whereas no expression domain is found in the

hypothalamic anlage (dotted circle). E. Stage 26. Expressiondomain of LjBmp2/4a is restricted to the lower ectoderm ofthe caudally extending NHP (arrowheads). Note the absenceof transcripts in the nasal placode. F. Stage 27. LjBmp2/4a isexpressed in the pharyngeal endoderm, whereas that in theNHP has been lost (arrowheads). G. Stage 23. Very faintexpression of LjHhA is present in the floor plate of the neuraltube, ventral forebrain and pharyngeal endoderm. H, I. Stage24, 26. LjHhA is strongly expressed in the floor plate and apart of the ventral diencephalon corresponding to zonalimitans intrathalamica (zli). Faint expression of this gene isalso present in the future hypothalamus (hpt). Note that noexpression is detected in the NHP (arrowheads) or in the oralectoderm. Abbreviations: see Fig. 2.

EVOLUTION OF PITUITARY DEVELOPMENT 41

Fig. 8. Evolution of the developmental processes ofvertebrate hypophysis: an hypothesis. A. Comparison of thedevelopmental processes of the adenohypophysis in theamphioxus, lamprey and gnathostome. In the amphioxus,Hatschek’s pit originates from the endoderm as Hatschek’sdiverticulum (Hd). The left Hd differentiates into Hatschek’spit (Hp) after metamorphosis. The Hp has an infundibulum-like ventral margin of the brain, which resembles thevertebrate hypothalamo-adenohypophysial axes. Note thatthe Hd expresses both Pitx and Pax6, and that the ventralneural tube expresses TTF-1. The adenohypophysis in thelamprey is derived from the NHP, which appears to bespecified by PitxA. The hypothalamic primordium specificallyexpresses TTF-1. Fgf and Bmp are present in the anterior andposterior portions of the NHP, whereas no signal moleculeshave been identified in the hypothalamic primordium. Notethat the topographical relationship between the adenohypo-physis and hypothalamus is established late in development inboth the amphioxus and lamprey. In gnathostomes, theseparation of the nasal placode (NP) from the adenohypophy-sial complex results in the regionalization of the adenohypo-physial origin (Rathke’s pouch; Rp) in the oral ectoderm (OE).The Rp is induced by BMP derived from the hypothalamic

primordium, following the specification of the ectodermalanlage by the Pitx and Pax6 genes. The pouch is thenexpanded and specific cell types are obtained by the action ofFGF and BMP derived from hypothalamic primordium andthe Rp. The epigenetic interactions between the Rp andhypothalamus are established early in the development ofgnathostomes. B. The evolutionary sequence of the changes ofdevelopmental events involved in the hypophysial organogen-esis during chordate evolution. The events have been placedon the phylogenic tree of the chordates. In the commonancestor of chordates, specification of the rostral endodermand the ventral CNS had already been acquired by someregulatory genes. This study suggests that the vertebrateancestor had already acquired an ectodermal origin of theNHP, and that some regulatory genes may have been involvedin the regionalization and specification of the NHP after thesplit from amphioxus. The topographical relationship betweenthe pituitary primordium and the ventral CNS is secondarilyestablished. After divergence of the jawless vertebrates,separation of NP from NHP may have shifted the adenohy-pophysial developmental field into the oral ectoderm. Thisheterotopic shift is assumed in the lineage to gnathostomes.Abbreviations: see Fig. 2.

K. UCHIDA ET AL.42

domain is secondarily restricted to Rathke’s pouch(Dasen and Rosenfeld, 2001; Suh et al., 2002;Fig. 8). The difference between the two animalgroups is more conspicuous in the spatiotemporalexpression patterns of growth factor encodinggenes; neither of the Fgf8- or Bmp2/4 cognates isexpressed in the hypothalamic anlage of thelamprey. Instead, they are expressed in theectoderm adjacent to the adenohypophysial anlage(Figs. 7, 8A). Such epidermal expression of thesegenes in the lamprey do not appear to beconsistent with the patterns found in gnathos-tomes with respect to the morphological homologyof embryonic tissues (Fig. 8). Interestingly, theexpression patterns are just as inconsistent interms of the anatomical architecture of the fullyformed adenohypophysis (Fig. 1). This inconsis-tency in gene expression patterns betweengnathostomes and lamprey may possibly be linkedin part to the differences in embryonic develop-ment of this organ.Although in the adult lamprey the fully formed

pituitary is identified with the topographicalrelationship similar to that in gnathostomes (Fig.1; Gorbman and Bern, ’62), the development of theadenohypophysis and the expression patterns ofpituitary-related genes are substantially differentbetween the two vertebrate groups. The adenohy-pophysis of the lamprey is derived from theposterior part of the NHP (Figs. 2, 7), whichappears as an independent entity located rostral tothe oral ectoderm. As the upper lip growsrostrally, the posterior part of the NHP growscaudally towards the rostral tip of the notochord toestablish a de novo contact with the postoptic levelof the ventral diencephalon: the hypothalamicprimordium expressing TTF-1 (Figs. 5G, 7). It isat this stage that the lamprey embryo firstacquires the relationship between tissues similarto that found between Rathke’s pouch and thehypothalamus in gnathostomes.During the topographical shift of the NHP, no

signaling molecules can be identified in thehypothalamic anlage, but both FGF and BMPwere identified in the NHP and surroundingectoderm. We could not clarify here whether sucha secondarily established topographical relation-ship in the lamprey is responsible for the spatio-temporal regulation of molecular cascades, orwhether epidermally derived growth factors couldpossibly compensate for similar functions thathave been suggested in the mouse. These ques-tions will require further experimental examina-tion. Thus, the anatomical relationship between

the adenohypophysis and the hypothalamic anlageis not established in the lamprey until a late stageof development. In gnathostome development, onthe other hand, the homologous relationship isobtained very early in embryogenesis, and thesame relationship is responsible for integratedsignaling that epigenetically induces the adenohy-pophysis and its anatomical differentiation intothree components. Irrespective of such a profounddifference in developmental patterns, apparentlyhomologous sets of genes are involved in adeno-hypophysial development in both of theseanimal groups, even though they are expressedin different places and at different times. Theevolutionary developmental biology of the adeno-hypophysis thus appears to provide an intriguingmix of heterochrony and heterotopy (see Hall, ’98).The evolution of the pituitary developmentalprogram from agnathans to gnathostomes couldbe viewed as changes in epigenetic patterningprograms between rostral ectoderm and theventral forebrain.

Evolution of the vertebrate pituitary:epigenetic development changes?

The epigenetic patterning programs in theadenohypophysial development of the gnatho-stomes raises an intriguing question as to howsuch programs might have been acquired inphylogenetic evolution. As exemplified in thegnathostome, the topographical relationshipbetween the hypothalamus and Rathke’s pouchhas two different aspects of biological significance:epigenetic induction during development, and theendocrine network of the adult. The present studyopens questions not simply on the origin of thehypophysis, but also on the mechanisms behindthe integration of developmental systems and thefunction of a fully formed organ system into anembryonic configuration.

Amphioxus, the sister group of vertebrates,shares several common basic features with verte-brates, not only in adult morphology, but also inembryonic developmental mechanisms (reviewedby Holland and Holland, 2001). However, thedevelopment of the hypophysis may be substan-tially different between these animal groups. Inamphioxus, Hatschek’s pit has been consideredthe homolog of the vertebrate adenohypophysis(Nozaki and Gorbman, ’92; Gorbman et al., ’99).This develops from an evagination of the endo-derm, named as Hatschek’s diverticulum, at theanterior end of the pharyngeal endoderm (Fig. 8).

EVOLUTION OF PITUITARY DEVELOPMENT 43

Amphioxus cognates of the Pitx and Pax6 genesappear to specify the rostral endoderm, corre-sponding to the future left Hatschek’s diverticu-lum (Glardon et al., ’98; Yasui et al., 2000;Boorman and Shimeld, 2002b). Amphioxus TTF-1 also seems to identify the ventral part of therostral neural tube, which is closely associatedwith Hatschek’s pit (Venkatesh et al., ’99; Ogasa-wara, 2000). From the spatiotemporal expressionpatterns of these regulatory genes, the basicdevelopmental patterning mechanisms of theHatschek’s pit and the ventral CNS may havealready been established at molecular levels in thecommon ancestor of chordates.Of the pair of Hatschek’s diverticula in the

amphioxus larva, the left one differentiates intoHatschek’s pit, located in the roof of the prebuccalcavity after metamorphosis (Holland and Holland,2001). Simultaneously, the pit makes contact withthe ventral CNS, thus establishing the vertebratehomolog of the hypothalamo-hypophysial complex(Fig. 8). It is not known whether this part of theamphioxus ‘brain’ actually induces the differentia-tion of the Hatschek’s pit, nor are there anyknown genes specifically expressed in this part ofthe neural tube. Nevertheless, the overall anatomyof this system as well as the extensive resemblancein cytological configuration strongly supports theidea that it represents the evolutionary origin ofthe vertebrate adenohypophysis (Gorbman et al.,’99), although there are differences in develop-mental origin; the Hatschek’s pit is of the rostralendoderm derivative, unlike in the vertebrates.Importantly, the topographical relationship be-tween Hatschek’s pit and the ventral CNS issecondarily established in this animal, as observedin the lamprey. It seems highly probable, there-fore, that the close topographical relationshipbetween the CNS and adenohypophysial equiva-lent was originally established in the late phase ofdevelopment in chordates.From a comparison of the gene expression

patterns between amphioxus and vertebrates, theanterior ectoderm in amphioxus may possiblycontain the origin of the sensory placode that ishomologous to the gnathostome olfactory placode(Holland and Holland, 2001). However, in thisspecies, the developmental origins differ betweenthe olfactory placode and Hatschek’s pit. In thelamprey, however, the nasal organ and theadenohypophysis develop from a single placode(Fig. 2). Moreover, as in gnathostomes, somesignal molecules are secreted from the lampreyNHP, suggesting its possible involvement in

differentiation of this region. It seems, therefore,most likely that the ectodermal adenohypophysiswas first acquired in the vertebrate ancestor, orafter the split of the lineage leading to amphioxus,together with the growth factor mediated mole-cular cascades to regionalize the NHP (Fig. 8). Asdefined by Haeckel (1875), such displacement ofthe development of an organ or tissue in space istermed ‘heterotopy’. A heterotopic shift, such asthe differentiation of the olfactory and adenohy-pophysial placodes in the ectoderm, can beassumed in the lineage to agnathans to explainthe evolutionary developmental processes in theearly common ancestor of the vertebrates (Fig. 8).

As the basic developmental plan, the nasal andadenohypophysial placodes are closely related toeach other in vertebrates. For the gnathostomes,the nasal and hypophysial placodes can be fate-mapped very close to each other at early develop-mental stages (Couly and Le Douarin, ’85). Insubsequent developmental stages, however, theyare soon separated from each other, possiblypartly because of the rostral growth of the partof the brain that vertically induces Rathke’s pouch(Kawamura and Kikuyama, ’98; Gleiberman et al.,’99). In gnathostomes, therefore, one crucialevolutionary change introduced into the pattern-ing of the pituitary gland appears to have been theearly embryonic separation of the nasal placodefrom the nasohypophysial complex, resulting inthe regionalization of the adenohypophysis in theoral ectoderm. This event may have resulted in theaccelerated establishment of a topographical re-lationship between the primordia of the adenohy-pophysis and the hypothalamus. This heterotopicshift may also have opened a possibility formesenchymal reorganization.

On the basis of the comparison of the morpho-logical patterning of trigeminal crest cells, whichform the jaw primordia in the lamprey andgnathostome, Kuratani et al. (2001) have postu-lated one possible scenario to explain the evolutionof the gnathostome jaw. Because of the solid NHP,jaw primordia fail to move rostromedially to formthe medial nasal septum as in the gnathostomes,but instead form the upper lip in the lamprey.Therefore, the accelerated separation of thenasohypophysial ectoderm in gnathostomes wouldpossibly be the key innovation allowing jawpatterning during vertebrate evolution.

Our hypothetical scenario for the evolutionarysequence of hypophysial development stronglysuggests that the topographical relationship itselfwas first set up as a functional organ system in the

K. UCHIDA ET AL.44

chordate ancestor, and that embryonic patterninghas secondarily adopted the same topography oftissues towards the earlier phases of development,to utilize it as the basis for epigenetic interactions.Our data on regulatory gene expression patternsimply that regulation of growth factor encodinggenes are more subject to heterotopic changes inregulation, rather than those genes encodingtranscription factors and often serving as specificmarkers. Whether such a change could explain theabove evolutionary reprogramming, often realizedas heterochrony and heterotopy, will be one of thekey questions for evolutionary developmentalbiology. Further studies in the lamprey to eluci-date the function of regulatory genes in this organsystem would be intriguing and contribute tofurther advancements in this field.

ACKNOWLEDGEMENTS

We thank Dr. Kiyokazu Agata and Dr. KaoruKubokawa for valuable discussion on the manu-script. This work was supported by Grants-in-Aidfrom the Ministry of Education, Science andCulture of Japan (Specially Promoted Research).

LITERATURE CITED

Acampora D, Gulisano M, Simeone A. 2000. Genetic molecularroles of Otx homeodomain proteins in head development.Gene 246:23–35.

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z,Miller W, Lipman DJ. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database searchprograms. Nucleic Acids Res 5:3389–3402.

Baker CV, Bronner-Fraser M. 2001. Vertebrate cranialplacodes I. Embryonic induction. Dev Biol 232:1–61.

Balfour FM. 1878. The development of the elasmobranchialfishes. J Anat Physiol 11:405–706.

Bentley PJ. 1998. Comparative Vertebrate Endocrinology.Cambridge: Cambridge University Press.

Boorman CJ, Shimeld SM. 2002a. Cloning and expression of aPitx homeobox gene from the lamprey, a jawless vertebrate.Dev Genes Evol 212:349–353.

Boorman CJ, Shimeld SM. 2002b. Pitx homeobox genes inCiona and amphioxus show left-right asymmetry is aconserved chordate character and define the ascidianadenohypophysis. Evol Dev 4:354–365.

Burgess R, Lunyak V, Rosenfeld M. 2002. Signaling andtranscriptional control of pituitary development. Curr OpinGenet Dev 12:534–539.

Couly GF, Coltey PM, Le Douarin NM. 1992. The develop-mental fate of the cephalic mesoderm in quail-chickchimeras. Development 114:1–15.

Couly GF, Le Douarin NM. 1985. Mapping of the early neuralprimordium in quail–chick chimeras. I. Developmentalrelationships between placodes, facial ectoderm, and pro-sencephalon. Dev Biol 110:422–439.

Dasen JS, Rosenfeld MG. 2001. Signaling and transcriptionalmechanisms in pituitary development. Annu Rev Neurosci24:327–355.

Dubois P, ElAmraoui A. 1995. Embryology of the pituitarygland. Trends Endocrinol Metab 6:1–7.

Eagleson GW, Jenks BG, Van Overbeeke AP. 1986. Thepituitary adrenocorticotropes originate from neural ridgetissue in Xenopus laevis. J Embryol Exp Morphol 95:1–14.

Ericson J, Muhr J, Jessell TM, Edlund T. 1995. Sonichedgehog: a common signal for ventral patterning alongthe rostrocaudal axis of the neural tube. Int J Dev Biol39:809–816.

Ericson J, Norlin S, Jessell TM, Edlund T. 1998. IntegratedFGF and BMP signaling controls the progression ofprogenitor cell differentiation and the emergence of patternin the embryonic anterior pituitary. Development 125:1005–1015.

Etchevers HC, Vincent C, Couly G. 2001. Neural crest andpituitary development. In: Rappaport R, Amselem S,editors. Hypothalamic–Pituitary Development: Geneticand Clinical Aspects. Basel: Karger, p 13–29.

Felsenstein J. 1981. Evolutionary trees from DNAsequences: a maximum likelihood approach. J Mol Evol17:368–376.

Gilbert SF, Bolker JA. 2001. Homologies of process andmodular elements of embryonic construction. In: WagnerGP, editor. The Character Concept in Evolutionary Biology.San Diego: Academic Press, p 435–454.

Glardon S, Holland LZ, Gehring WJ, Holland ND. 1998.Isolation and developmental expression of the amphioxusPax-6 gene (AmphiPax-6): insights into eye and photore-ceptor evolution. Development 125:2701–2710.

Gleiberman AS, Fedtsova NG, Rosenfeld MG. 1999. Tissueinteractions in the induction of anterior pituitary: role of theventral diencephalon, mesenchyme, and notochord. Dev Biol213:340–353.

Gorbman A. 1983. Early development of the hagfish pituitarygland: evidence for the endodermal origin of the adenohy-pophysis. Amer Zool 23:639–654.

Gorbman A. 1995. Olfactory origins and evolution of thebrain-pituitary endocrine system: facts and speculation.Gen Comp Endocrinol 97:171–178.

Gorbman A, Bern HA. 1962. A Textbook of ComparativeEndocrinology. New York: John Wiley and Sons.

Gorbman A, Tamarin A. 1985. Early development of oral,olfactory and adenohypophyseal structures of agnathansand its evolutionary implications. In: Foreman RE,Gorbman A, Dodd JM, Olsson R, editors. EvolutionaryBiology ;of Primitive Fishes. New York and London:Plenum, p 165–185.

Gorbman A, Nozaki M, Kubokawa K. 1999. A brain-Hatschek’s pit connection in amphioxus. Gen Comp En-docrinol 113:251–254.

Haeckel E. 1875. Die Gastrea und die Erforschung der Tiere.Jena Z Naturwiss 9:402–508.

Hall BK. 1998. Evolutionary Developmental Biology, 2nd ed.London: Chapman and Hall.

Herzog W, Zeng X, Lele Z, Sonntag C, Ting JW, Chang CY,Hammerschmidt M. 2003. Adenohypophysis formation inthe zebrafish and its dependence on sonic hedgehog. DevBiol 254:36–49.

Holland LZ, Holland ND. 2001. Evolution of neural crest andplacodes: amphioxus as a model for the ancestral verte-brate? J Anat 199:85–98.

EVOLUTION OF PITUITARY DEVELOPMENT 45

Jones DT, Taylor WR, Thornton JM. 1992. The rapidgeneration of mutation data matrices from protein se-quences. Comput Appl Biosci 8:275–282.

Kawamura K, Kikuyama S. 1998. Morphogenesis of thehypothalamus and hypophysis: their association, dissocia-tion and reassociation before and after ‘Rathke’. Arch HistolCytol 61:189–198.

Kawamura K, Kouki T, Kawahara G, Kikuyama S. 2002.Hypophyseal development in vertebrates from amphibiansto mammals. Gen Comp Endocrinol 126:130–135.

Kimura S, Hara Y, Pineau T, Fernandez-Salguero P, Fox CH,Ward JM, Gonzalez FJ. 1996. The T/ebp null mouse:thyroid-specific enhancer-binding protein is essential forthe organogenesis of the thyroid, lung, ventral forebrain,and pituitary. Genes Dev 10:60–69.

Kioussi C, O’Connell S, St-Onge L, Treier M, Gleiberman AS,Gruss P, Rosenfeld MG. 1999. Pax6 is essential forestablishing ventral-dorsal cell boundaries in pituitarygland development. Proc Nat. Acad Sci USA 96:14378–14382.

Kishino H, Miyata T, Hasegawa M. 1990. Maximum likelihoodinference of protein phylogeny and the origin of chloro-plasts. J Mol Evol 30:151–160.

Kouki T, Imai H, Aoto K, Eto K, Shioda S, Kawamura K,Kikuyama S. 2001. Developmental origin of the ratadenohypophysis prior to the formation of Rathke’s pouch.Development 128:959–963.

Kumar S, Balczarek KA, Lai ZC. 1996. Evolution of thehedgehog gene family. Genetics 142:965–972.

Kuratani S, Nobusada Y, Saito H, Shigetani Y. 2000.Morphological characteristics of the developing cranialnerves and mesodermal head cavities in sturgeon embryosfrom early pharyngula to late larval stages. Zool Sci 17:911–933.

Kuratani S, Nobusada Y, Horigome N, Shigetani Y.2001. Embryology of the lamprey and evolution ofthe vertebrate jaw: insights from molecular and develop-mental perspectives. Phil Trans Roy Soc Lond B 356:1615–1632.

Kuratani S, Kuraku S, Murakami Y. 2002. Lamprey as an evo-devo model: Lessons from comparative embryology andmolecular phylogenetics. Genesis 34:175–183.

Mallamaci A, Di Blas E, Briata P, Boncinelli E, Corte G. 1996.OTX homeoprotein in the developing central nervoussystem and migratory cells of the olfactory area. Mech Dev58:165–178.

Matsuo I, Kuratani S, Kimura C, Takeda N, Aizawa S. 1995.Mouse Otx2 functions in the formation and patterning ofrostral head. Genes Dev 9:2646–2658.

Meyer A, Schartl M. 1999. Gene and genome duplications invertebrates: the one-to-four (-to-eight in fish) rule and theevolution of novel gene functions. Curr Opin Cell Biol11:699–704.

Murakami Y, Ogasawara M, Sugahara F, Hirano S, Satoh N,Kuratani S. 2001. Identification and expression of thelamprey Pax6 gene: evolutionary origin of the segmentedbrain of vertebrates. Development 128:3521–3531.

Nozaki M, Gorbman A. 1992. The question of functionalhomology of Hatschek’s pit of amphioxus (Branchiostomabelcheri). Zool Sci 9:387–395.

Nozaki M, Ominato K, Takahashi A, Kawauchi H, Sower SA.2001. Adenohypophysial cell types in the lampreypituitary: current state of the art. Comp Biochem Physiol B129:303–309.

Ogasawara M. 2000. Overlapping expression of amphioxushomologs of the thyroid transcription factor-1 gene andthyroid peroxidase gene in the endostyle: insightinto evolution of the thyroid gland. Dev Genes Evol210:231–242.

Ogasawara M, Shigetani Y, Suzuki S, Kuratani S, Satoh N.2001. Expression of thyroid transcription factor-1 (TTF-1)gene in the ventral forebrain and endostyle of the agnathanvertebrate, Lampetra japonica. Genesis 30:51–58.

Osumi-Yamashita N, Ninomiya Y, Doi H, Eto K. 1994. Thecontribution of both forebrain and midbrain crest cells tothe mesenchyme in the frontonasal mass of mouse embryos.Dev Biol 164:409–419.

Patzner RA, Erhart G, Adam H. 1982. Cell types in theadenohypophysis of the hagfish, Myxine glutinosa (Cyclos-tomata). Cell Tissue Res 223:583–592.

Rathke H. 1838. Uber die entstehung der glandula pituitaria.Anat Physiol Wissen Med 838:482–485.

Saitou N, Nei M. 1987. The neighbor-joining method: a newmethod for reconstructing phylogenetic trees. Mol Biol Evol4:406–425.

Sbrogna JL, Barresi MJ, Karlstrom RO. 2003. Multiple rolesfor Hedgehog signaling in zebrafish pituitary development.Dev Biol 254:19–35.

Scully KM, Rosenfeld MG. 2002. Pituitary development:regulatory codes in mammalian organogenesis. Science295:2231–2235.

Sewertzoff AN. 1928. The head skeleton and muscles ofAcipenser ruthenus. Acta Zool 9:193-319.

Sheng HZ, Westphal H. 1999. Early steps in pituitaryorganogenesis. Trends Genet 15:236–240.

Shigetani Y, Sugahara F, Kawakami Y, Murakami Y, HiranoS, Kuratani S. 2002. Heterotopic shift of epithelial-mesenchymal interactions in vertebrate jaw evolution.Science 296:1316–1319.

Sower SA, Kawauchi H. 2001. Update: brain and pituitaryhormones of lampreys. Comp Biochem Physiol B 129:291–302.

Steinberg M. 1957. A non-nutrient culture medium foramphibian embryonic tissues. Carnegie Institution ofWashington Year Book 56:347–348.

Suga H, Hoshiyama D, Kuraku S, Katoh K, Kubokawa K,Miyata T. 1999. Protein tyrosine kinase cDNAs fromamphioxus, hagfish, and lamprey: isoform duplicationsaround the divergence of cyclostomes and gnathostomes.J Mol Evol 49:601–608.

Suh H, Gage PJ, Drouin J, Camper SA. 2002. Pitx2 is requiredat multiple stages of pituitary organogenesis: pituitaryprimordium formation and cell specification. Development129:329–337.

Tahara Y. 1988. Normal stages of development in the lamprey,Lampetra reissneri (Dybowski). Zool Sci 5:109–118.

Takuma N, Sheng HZ, Furuta Y, Ward JM, Sharma K,Hogan BL, Pfaff SL, Westphal H, Kimura S, Mahon KA.1998. Formation of Rathke’s pouch requires dualinduction from the diencephalon. Development 125:4835–4840.

Treier M, O’Connell S, Gleiberman A, Price J, Szeto DP,Burgess R, Chuang PT, McMahon AP, Rosenfeld MG. 2001.Hedgehog signaling is required for pituitary gland develop-ment. Development 128:377–386.

Treier M, Rosenfeld MG. 1996. The hypothalamic-pituitaryaxis: co-development of two organs. Curr Opin Cell Biol8:833–843.

K. UCHIDA ET AL.46

Treier M, Gleiberman AS, O’Connell SM, Szeto DP, McMahonJA, McMahon AP, Rosenfeld MG. 1998. Multistep signalingrequirements for pituitary organogenesis in vivo. Genes Dev12:1691–1704.

Ueki T, Kuratani S, Hirano S, Aizawa S. 1998. Otx cognatesin a lamprey, Lampetra japonica. Dev Genes Evol 208:

223–228.Venkatesh TV, Holland ND, Holland LZ, Su MT, Bodmer R.1999. Sequence and developmental expression of amphioxusAmphiNk2-1: insights into the evolutionary origin of thevertebrate thyroid gland and forebrain. Dev Genes Evol209:254–259.

Watkins-Chow DE, Camper SA. 1998. How many homeoboxgenes does it take to make a pituitary gland? Trends Gene14:284–290.

Whitlock KE, Wolf CD, Boyce ML. 2003. Gonado-tropin-releasing hormone (GnRH) cells arise from cranialneural crest and adenohypophyseal regions of theneural plate in the zebrafish, daniorerio. Dev Biol257:140–152.

Yasui K, Zhang S, Uemura M, Saiga H. 2000. Left-rightasymmetric expression of BbPtx, a Ptx-related gene, in alancelet species and the developmental left-sidedness indeuterostomes. Development 127:187–195.

EVOLUTION OF PITUITARY DEVELOPMENT 47