Genetic Analysis of Adenohypophysis Formation in Zebrafish

WIEBKE HERZOG, CARMEN SONNTAG, BRIGITTE WALDERICH,* JO¨ RG ODENTHAL,* HANS-MARTIN MAISCHEIN,* AND MATTHIAS HAMMERSCHMIDT Max-Planck Institute for Immunobiology (W.H., C.S., M.H.), 79108 Freiburg, Germany; Exelixis Germany (B.W., J.O.), 72076 Tuebingen, Germany; and Max-Planck Institute for Developmental Biology (H.-M.M.), 72076 Tuebingen, Germany Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021

The adenohypophysis consists of at least six dif- pophyseal cell types only. The third gene is ze- ferent cell types, somatotropes, lactotropes, thy- brafish pit1 and is required for lactotropes, thyro- rotropes, melanotropes, corticotropes, and gona- tropes, and somatotropes, similar to its mouse dotropes. In mouse, cloning of spontaneous ortholog, whereas the fourth, aal, is required for mutations and gene targeting has revealed multi- corticotropes, melanotropes, thyrotropes, and so- ple genes required for different steps of adenohy- matotropes, but not lactotropes. In conclusion, the pophysis development. Here, we report the results isolated zebrafish mutants confirm principles of of a systematic search for genes required for ad- adenohypophysis development revealed in mouse, enohypophysis formation and patterning in ze- thereby demonstrating the high degree of molec- brafish. By screening F3 offspring of N-ethyl-N- ular and mechanistic conservation among the dif- nitrosourea-mutagenized founder fish, we isolated ferent vertebrate species. In addition, they point to eleven mutants with absent or reduced expression thus far unknown features of adenohypophysis de- of GH, the product of somatotropes, but a normally velopment, such as the existence of a new lineage developing hypothalamus. Of such mutants, eight of pituitary cells, which partially overlaps with the were further analyzed and mapped. They define Pit1 lineage. Positional cloning of the lia, pia, and four genes essential for different steps of adeno- aal genes might reveal novel regulators of verte- hypophysis development. Two of them, lia and pia, brate pituitary development. (Molecular Endocrin- affect the entire adenohypophysis, whereas the ology 18: 1185–1195, 2004) other two are required for a subset of adenohy-

HE PITUITARY GLAND (hypophysis) and the hy- melanocortin (POMC), which also gives rise to endor- Tpothalamus constitute an integrative center of en- phins secreted by specific cells of the brain (2). MSH is docrine control, linking the central nervous system and generated in the intermediate lobe (pars intermedia) of the endocrine system in regulating basic body func- the pituitary, which is rather rudimentary in human (1). tions and homeostasis. The pituitary itself comprises In contrast to the adenohypophysis, the neurohy- two anatomically and functionally distinct systems, the pophysis itself does not contain endocrine cells. anterior lobe (adenohypophysis) and the posterior lobe Rather, it secretes hormones from axonal termini of (neurohypophysis). The adenohypophysis is further hypothalamatic neuroendocrine cells that have inner- subdivided into three regions: the pars anterior (or vated the neurohypophysis. These hormones are oxy- pars distalis), the pars tuberalis, and the pars interme- tocin (isotocin in fish) and vasopressin (vasotocin in dia (1). It contains at least six different cell types that fish), generated in the supraoptic and paraventricular are characterized by the different hormones they pro- nuclei of the hypothalamus. In addition, neuroendo- duce and secrete: lactotropes generating prolactin crine cells from these and other hypothalamic nuclei (PRL), somatotropes generating GH, thyrotropes gen- elaborate peptide hormones (so called releasing or erating TSH, gonatotropes generating FSH and LH, release inhibiting hormones), which control the activity corticotropes generating ACTH, and melanotropes of endocrine cells of the adenohypophysis. In contrast generating MSH. ACTH and MSH are formed via pro- to the neurohypophyseal hormones, these hormones teolytic cleavage from a common proprotein, proopio- do not reach the pituitary via nerve fibers. Rather, they travel in the blood via the hypothalamic-hypophyseal Abbreviations: ENU, N-ethyl-N-nitrosourea; hpf, hours post fertilization; POMC, proopiomelanocortin; PRL, prolac- portal system (1). tin; Prop1, Prophet-of-Pit1. The development of the adenohypophysis and the *B.W., J.O., and H.-M.M. represent the Tuebingen 2000 formation of its different cell types from a common screen consortium. primordium is controlled both by intrinsic and extrinsic Molecular Endocrinology is published monthly by The factors. The adenohypophysis is derived from pla- Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine codal ectoderm at the anterior neural ridge (reviewed community. in Ref. 3), that becomes committed toward an adeno-

1185 1186 Mol Endocrinol, May 2004, 18(5):1185–1195 Herzog et al. • Adenohypophysis Formation in Zebrafish

hypophyseal fate via inductive signals from the ventral the zebrafish is highly suitable for mutant screens. In diencephalon, a subdivision of the forebrain that later the past, three independent large-scale screens have gives rise to hypothalamus, infundibulum and neuro- been carried out, two of which used the chemical hypophysis (4, 5). During further development, the ad- N-ethyl-N-nitrosourea (ENU) to introduce random enohypophyseal anlage remains under the control of point mutations over the entire genome (23, 24). signals from the hypothalamus. These signals, to- Whereas in these screens mutant analyses were gether with intrinsic signals of the adenohypophysis largely restricted to examining embryonic and larval itself, regulate the maintenance, proliferation and dif- morphology at different developmental stages, the ferential specification of adenohypophyseal cells by more recent Tuebingen 2000 large-scale ENU screen sequential activation of different transcription factors, was set up to allow mutant screening with molecular such as the Lim class homeodomain proteins Lhx3 (6) tools (see for cloned mutations isolated in this screen, Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021 or Lhx4 (7), the paired-like homeodomain protein Refs. 25 and 26). As part of the Tuebingen 2000 Prophet-of-Pit1 (Prop1; Ref. 8), the Pou domain pro- screen, we searched for genes required for zebrafish tein Pit1 (9), the T-box factor Tpit/Tbx19 (10–12), the adenohypophysis formation and patterning, carrying zinc finger protein Gata2 (13), and several others (re- out large-scale whole mount in situ hybridizations with viewed in Refs. 14–20). a probe detecting Gh encoding transcripts, a marker Final evidence for the requirement of such factors for somatotropes. The staining was necessary be- for pituitary development requires genetic analyses via loss-of-function mutants. In mouse, such mutants cause the adenohypophysis of zebrafish larvae is mor- have been generated using gene knockout technol- phologically too indistinct to be analyzed in large num- ogy, as in the case of Lhx3, that thereby was shown to bers without molecular tools. be required for all adenohypophyseal cell types except Recent work has revealed both crucial similarities the corticotropes (6). More recently, conditional and differences in the development of the adenohy- knockout techniques allow one to investigate the pi- pophysis between fish and mammals. The zebrafish tuitary-specific role of genes also involved in other pituitary contains the same cell types (lactotropes, processes, as in the case of the steroidogenic factor corticotropes, melanotropes, thyrotropes, somato- Sf1, that is required downstream of Gata2 for the tropes, gonadotropes; Refs. 27–29) as the mammalian formation of gonadotropes (13, 21). However, such pituitary, however, there are crucial differences in the gene targeting approaches are generally biased be- morphogenesis and the architecture of the glands in cause they require a priori knowledge of the molecular the different vertebrate species. Thus, the zebrafish nature of the genes to be analyzed. As an alternative to adenohypophysis maintains its placodal organization this reverse genetics approach, genes essential for and remains in a subepithelial position after oral cavity mouse pituitary development were isolated via posi- formation, whereas no invagination of the oral ecto- tional cloning of spontaneous, viable mutations, such derm equivalent to Rathke’s pouch formation in mam- as Prop1, that is mutated in Ames dwarf mutants and mals takes place (27). Furthermore, there are differ- required for somatotropes, lactotropes, thyrotropes, ences in the patterning of the adenohypophyseal and gonadotropes (8), and Pit1, which is mutated in anlage: in zebrafish, the different cell types are distrib- Snell and Jackson dwarf mutants and required for uted in three distinct domains along the antero- somatotropes, lactotropes and thyrotropes, defining posterior axis of the pituitary, rather than along the the so-called Pit1 lineage of the adenohypophysis (9) dorsoventral axis as in mammals (27), with the Msh- (also see Ref. 20 for recent review). Tpit/Tbx19 re- generating pars intermedia most posteriorly, and the quirement for the Pomc lineage (corticotropes and neurohypophysis located dorsal of adenohypophysis melanotropes) was revealed via the analysis of human (1). In addition, fish and mammals display significant patients with isolated Acth deficiencies caused by re- differences in the onset of the specification of the cessive Tbx19 point mutations (11). However, as in lactotrope lineage, that is the first lineage to specify in other known human pituitary disorders (see Refs. 16 fish, but the last in mouse (27). Despite such differ- and 22 for reviews), there are too few families and ences, crucial mechanisms of adenohypophyseal in- patients to allow a direct positional cloning of the affected human gene. Furthermore, the spontaneous duction and patterning appear to be highly conserved mutation rates in humans and mice are too low to among fish and mammals, as for instance indicated by allow a saturating identification of all genes required the conserved role of Sonic Hedgehog signaling from for adenohypophysis development via forward the ventral diencephalon (27, 28). In light of these data, genetics. the large-scale ENU mutagenesis screen was ex- In light of these limitations of reverse and forward pected to uncover genes regulating fish-specific fea- genetics in mammalian systems, we carried out a for- tures of the adenohypophysis, as well as genes with ward genetic analysis of pituitary mutants in a non- shared functions in mouse and fish. This would fur- mammalian vertebrate, the zebrafish. Due to its extra- ther illuminate the degree of conservation between corporal and rapid development, the transparency of the two vertebrate species, and possibly identify its embryos and larvae, its high fecundity, its relatively novel regulators also involved in mammalian pitu- small size and the ease of high-density maintenance, itary development. Herzog et al. • Adenohypophysis Formation in Zebrafish Mol Endocrinol, May 2004, 18(5):1185–1195 1187

RESULTS absence of gh staining in whole mount in situ hybridization. Other gross abnormalities, such as brain de- Identification of Mutations Affecting Zebrafish generation, which could have secondarily affected the Adenohypophysis Development pituitary, were not detected in these mutants. Only two of the 13 mutants (allele numbers t24594 and t20626) As part of a large-scale diploid F3 ENU mutagenesis showed reduced eye distance and a curled-down tail, screen to uncover recessive zygotic effect mutations, indicative for midline defects. Although t20626 was we screened for mutations affecting zebrafish adeno- not further investigated, genomic mapping placed hypophysis development. In total, F3 clutches of 4584 t24594 31–32 centimorgans from the top of linkage F2 families, representing 4253 mutagenized haploid group 6, just below the marker Z6626. This is the same genomes, were screened at 120 h post fertilization

position as reported for the midline mutation iguana Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021 (hpf) via whole mount in situ hybridization with a probe (Ref. 32; and http://zfin.org/cgi-bin/view_mapplet.cgi, detecting transcripts of gh, a marker of the somato- Ref. 40), and strongly suggests that t24594 is a weak tropes (27). iguana allele that we did not analyze further. The late developmental stage (120 hpf) for mutant Of the remaining 11 adenohypophysis-specific mu- screening was chosen for two reasons: 1) to reveal tations, eight were recovered. A combination of also defects in later steps of pituitary development or complementation crossing and genomic mapping re- patterning that would be missed when mutants were vealed that they fall into four complementation groups, investigated much earlier (see below for lim3 staining defining four genes essential for adenohypophysis de- of pit1 and aal mutants at 32 hpf), and 2) to avoid velopment in the zebrafish. The genomic positions of pituitary phenotypes that are secondary conse- the four genes, the different alleles, and recognized quences of early and general defects, such as early other specific defects of the mutants are listed in Table brain necrosis, that occurs very frequently, but usually leads to embryonic death between 24 and 72 hpf (30). 1. All genes appear to be dispensable for hypothala- Secondary consequences on pituitary formation have mus development, as revealed by the normal expres- also been revealed for midline mutants defective in sion of nkx2.1 (41) in mutant embryos at 32 hpf (Fig. 1, signaling by Nodal and Hedgehog family members B, E, H, K, and N), and the normal number of isotocin- (27, 28, 31–36). In such mutants, eye field separation positive cells at 32 hpf (Fig. 1, C, F, I, L, and O). and hypothalamus formation are compromised, lead- isotocin cDNA was isolated in our laboratory via de- ing to pituitary defects caused by the lack of inductive generate RT-PCR, and is identical with the cDNA re- signaling from the hypothalamus, consistent with the cently reported by Unger and Glasgow (42). It is ex- loss of the pituitary in hypothalamus-deficient mouse pressed in two nuclei of magnocellular neurons in the embryos (4, 37–39). However, in contrast to pituitary- anterior hypothalamus (Fig. 1C), possibly the pre- specific mutants, such zebrafish midline null mutants sumptive supraoptic or paraventricular nuclei (see In- usually die before 120 hpf (32), although secondary troduction), from where the peptide hormone is sup- pituitary defects caused by more subtle, viable midline posed to undergo axonal transport into the more mutations cannot be ruled out. ventrally and posteriorly located region of the neuro- In total, we identified 13 mutants that were viable at hypophysis adjacent to the adenohypophysis (see 120 hpf but displayed a severe reduction or complete Refs. 14 and 43 for review).

Table 1. Identified Genes and Alleles, Their Genomic Localization, and Mutant Phenotypes Mutant Phenotype: Linkage Group Gene Alleles Mutant Phenotype: Adenohypophysis Other Affected (Chromosome) Structures lia t21142 7 32 hpf: no lim3 expression Branchial arches t24149 120 hpf: all pituitary cell types absent Inner ear t24152 t26212 pia t25215 18 32 hpf: lim3 expression reduced — 120 hpf: all pituitary cell types absent aal t22744 24 32 hpf: lim3 expression normal — 120 hpf: all pituitary cell types absent except lactotropes pit1 t21379 9 32 hpf: lim3 expression normal — t22072 120 hpf: thyrotropes, somatotropes, lactotropes absent Allelism of the different mutations was determined by complementation crossing, or by meiotic mapping (see Materials and Methods). Complementation groups consist of all alleles that fail to complement each other in trans-heterozygotes, genetically defining a gene. Linkage groups, defined as units of linked meiotic segregation, are the genetic equivalents of chromosomes (25 in zebrafish; Refs. 58 and 59). —, None. 1188 Mol Endocrinol, May 2004, 18(5):1185–1195 Herzog et al. • Adenohypophysis Formation in Zebrafish Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021

Fig. 1. Early Pituitary and Hypothalamus Development in the Different Zebrafish Mutants First column (A, D, G, J, and M): lim3 expression in eye (marked “eye” in A), interneurons (marked “int” in A) and adenohypophyseal anlage (marked with arrow in A), 32 hpf, lateral view on head, anterior to the left. Second column (B, E, H, K, and N): nkx2.1 expression in the hypothalamus (blue; indicated with “hyp” in B) and prl expression in lactotropes (prl; red; indicated with arrow in B); 32 hpf; lateral view on head; anterior to the left. Third column (C, F, I, L, and O): isotocin expression in magnocellular neurons of the hypothalamus (blue; indicated with “mcn” in C) and prl expression in lactotropes (red; indicated with arrow in C); 72 hpf; ventral view on head; anterior to the left. A–C, Wild-type siblings (wt); D–F, lia mutants; G–I, pia mutants; J–L, pit1 mutants; M–O, aal mutants. All mutants show normal nkx2.1 and isotocin expression, suggesting normal development of the hypothalamus, thereby ruling out that the observed pituitary deficiencies are secondary. Early lim3 expression in the adenohypophyseal anlage is normal in pit1 (J) and aal (M) mutants, reduced in pia mutants (G) and absent in lia (D) mutants. Early prl expession in lactotropes is absent in lia (E and F), pia (H and I), and pit1 (K and L) mutants, but normal in aal mutants (N and O).

In contrast to the normally developing hypothala- posed to be of placodal origin, derived from medial mus and neurohypophysis, mutants in all four genes cells of the anterior neural ridge (Ref. 44), and laterally show specific, but different phenotypes in the adeno- flanked by cells of the olfactory placodes (45). hypophysis. Two of the mutants, lim absent (lia), rep- Whereas a few markers like anf (46) are expressed in resented by four alleles, and pituitary absent (pia), the anterior neural ridge at late gastrula and early represented by one allele, lack the entire adenohy- segmentation stages, none continues to be expressed pophysis. Whereas in pit1 mutants, represented by until cells can be identified as adenohypophyseal. In two alleles, and in all-absent-except-lactotropes (aal) addition, none of them is altered in our mutants (data mutants, represented by one allele, only some of the not shown). In this respect, the currently known earli- pituitary cell types are absent (Fig. 2). est bona fide marker for the presumptive zebrafish Mutations Affecting the Entire Adenohypophysis adenohypophysis is lim3 (31), the homolog of Lhx3 required for Rathke’s pouch formation in mouse (6). In As in mammals and birds (reviewed in Ref. 3), the zebrafish, lim3 starts to be expressed at the 21-somite adenohypophysis of the zebrafish embryo is sup- stage, presumably in all cells of the pituitary anlage, Herzog et al. • Adenohypophysis Formation in Zebrafish Mol Endocrinol, May 2004, 18(5):1185–1195 1189 Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021

Fig. 2. Pituitary Hormone Expression in Mutant Larvae at 120 hpf All panels show ventral views on head region, anterior to the left. First column (A, E, I, M, and Q): prl expression in lactotropes (indicated with arrow in A). Second column (B, F, J, N, and R): expression of pomc in corticotropes and melanotropes of the adenohypohysis (indicated with arrows in B), and in ␤-endorphin-synthesizing cells of the future arcuate nuclei of the hypothalmus (marked “anc” in B). Third column (C, G, K, O, and S): expression of gh in somatotropes (indicated with arrow in C). Fourth column (D, H, L, P, and T): expression of tsh in thyrotropes (indicated with arrow in D). A–D, Wild-type siblings (wt); E–H, lia mutants; I–L, pia mutants; M–P, pit1 mutants; Q–T, aal mutants. All four mutants lack gh (G, K, O, and S) and tsh (H, L, P, and T) expression and show normal pomc expression in the ␤-endorphin-synthesizing cells of the hypothalamus (F, J, N, and R). lia, pia, and aal mutants lack both the anterior and the posterior pomc expression domain of the adenohypophysis, whereas in pit1 mutants, both pomc domains are slightly enlarged. The gap between the anterior and the posterior domain (normally accommodating thyrotropes and somatotropes; Ref. 27) is present but reduced in size. prl-positive cells are absent in lia (E), pia (I), and pit1 (M) mutants, but present in relatively normal numbers in aal mutants (Q). However, although in wild-type larvae, lactotropes are organized in a sharp domain in the anterior region of the adenohypophysis (A; Ref. 27), they are widely dispersed along the anteroposterior axis of the pituitary region of aal mutants (Q). when it is still organized in a placodal fashion in front and the pomc expressing corticotropes and/or mela- of the forming head (31). Differentiation of the first two notropes, occurs very soon after the onset of lim3 pituitary cell lineages, the prl expressing lactotropes expression in lateral regions of the lim3 domain, 1190 Mol Endocrinol, May 2004, 18(5):1185–1195 Herzog et al. • Adenohypophysis Formation in Zebrafish

whereas expression of tsh in thryotropes and gh in tor Lhx3 (6). The complementary lineage, consisting of somatotropes only starts later, after the pituitary has thyrotropes, somatotropes, lactotropes, and gonado- moved inwards (27). At 72 hpf, prl, pomc, tsh, and gh tropes, is characterized by its Lhx3 dependence (6) are expressed in distinct, but partially overlapping do- (and reviewed in Refs. 14, 15, 17, 18, and 20). Another mains along the antero-posterior axis of the adenohy- lineage, comprising a subset of the Lhx3-dependent pophysis (Fig. 2, A–D). cell types, is defined via its dependence on the Pou lia and pia mutants lack the expression of all four domain transcription factor Pit1 (48, 49), and therefore adenohypophysis hormones at 72 hpf (Fig. 1, F and I; called the Pit1 lineage. It consists of thyrotropes, lac- and data not shown) and 120 hpf (Fig. 2, E–L). The totropes, and somatotropes, all of which are absent in remaining expression of pomc in two longitudinal Pit1 mutant mice (9). stripes (Fig. 2, F and J) is confined to ␤-endorphin- In our zebrafish screen, we could identify two genes Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021 synthesizing cells in the ventral base of the dienceph- required for the specification of two different adeno- alon (27) (for mouse see Refs. 2 and 47), whereas hypophyseal cell lineages. Mutants in both genes melanotropes and corticotropes of the adenohypoph- show normal lim3 expression at 32 hpf (Fig. 1, J and M, ysis, as well as lactotropes, thyrotropes, and somato- in comparison to Fig. 1A for wild-type control), indi- tropes, are absent. cating that the pituitary anlage is induced normally and In contrast to their indistinguishable phenotypes at d of normal size. However, already at this early stage (32 3 and 5 after fertilization, lia and pia mutants differ in hpf), the subdivision within the anlage is altered, indi- lim3 expression during earlier stages of development. cated by the loss of prl expression in pit1 mutants (Fig. Although in lia mutants, lim3 expression in the region 1, K and L), whereas prl expression in aal mutants of the adenohypophysis anlage is completely absent appears normal (Fig. 1, N and O, in comparison to Fig. at 32 hpf (Fig. 1D), lim3 expression in pia mutants is 1, B and C, as control). reduced, but clearly present (Fig. 1G in comparison to At 120 hpf, pit1 mutant embryos are characterized Fig. 1A for wild-type control). However, prl expression by loss of lactotropes (prl, Fig. 2M), somatotropes (gh, at 32 hpf is absent in both mutants (Fig. 1, E and H, in Fig. 2K) and thyrotropes (tsh, Fig. 2P), whereas pomc- comparison to Fig. 1B for wild-type control). In sum, positive adenohypophyseal cells, most likely anterior the expression pattern analyses suggest that both lia corticotropes and posterior melanotropes (27, 29), are and pia are required for the specification of all adeno- present at normal or even slightly elevated numbers hypophyseal cell types, with lia acting earlier than pia. (Fig. 2N). The zebrafish pit1 mutants resemble Pit1- lia appears to act upstream of lim3 (but not necessarily deficient mice (9), and indeed could be shown to carry via lim3; see Discussion), whereas pia is more likely to mutations in the zebrafish pit1 gene (50). act at the level, in parallel or downstream of lim3. The aal mutant zebrafish show a very different pheno- phenotype of pia zebrafish mutants is quite similar to type; they fail to generate pomc-, tsh-, and gh- that of Lhx3/Lhx4 double mutant mice (7), suggesting expressing pituitary cells (Fig. 2, R–T), whereas the that it might be caused by a mutation in a zebrafish lim prl-expressing lactotropes are still present both at 32 gene. However, lim3 itself can be ruled out, as it maps hpf (Fig. 1, N and O) and 120 hpf (Fig. 2Q). However, to a different linkage group (LG5; http://zfin.org/ at the later developmental stages, the lactotropes are cgi-bin/view_mapplet.cgi, Ref. 40) than pia (LG18; dispersed along the antero-posterior axis (Fig. 2Q), Table 1). rather than being organized in the sharp anterior domain found in wild-type siblings (Fig. 2A). This pheno- Mutations Affecting Adenohypophyseal type seems to define a new lineage of adenohypophy- Patterning and Lineage Specification seal cells, consisting of corticotropes, melanotropes, thyrotropes, and somatotropes, all of which depend The different cell types of the adenohypophysis derive on the aal gene, whereas the lactotropes appear to from a common primordium, the placode, that initially develop independently of aal function (Fig. 3). In con- consists of identical precursor cells. Subsequently, the trast to pit1, no phenotype similar to that of aal mu- placode becomes patterned, and cells in different re- tants has been described in mouse thus far. Thus, the gions of the primordium undergo differential specifi- aal mutation seems to have revealed a thus far un- cations to form the different cell types characterized known lineage of adenohypophyseal cells that partially by the hormones they produce and secrete. Data from overlaps with the Pit1 lineage. mammalian pituitary development suggest that the patterning process is characterized by subdivision of the common pool of precursor cells into certain cell lineages that then continue to branch until single cell DISCUSSION types are generated. It is assumed that cells of the Pomc lineage, giving rise to corticotropes and mela- By screening approximately 1.5 times more mu- notropes, split off first. In contrast to all other pituitary tagenized haploid genomes than in the first Tuebingen cell types, they express and require the T-box tran- large-scale screen (23), we isolated eleven mutants scription factor Tpit/Tbx19 (10–12), but are indepen- with specific defects in the formation and patterning of dent of the Lim class homeodomain transcription fac- the zebrafish adenohypophysis. Of those 11, three Herzog et al. • Adenohypophysis Formation in Zebrafish Mol Endocrinol, May 2004, 18(5):1185–1195 1191

Fig. 3. Diagram Illustrating the Putative Roles of the lia, pia, pit1, and aal Gene Products during Different Steps of Zebrafish Adenohypophysis Formation and Patterning lia and pia are required for all adenohypophyseal cell types. lia acts upstream of lim3, and might therefore be required for the Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021 formation or maintenance of the entire adenohypophyseal anlage. pia acts after the initiation of lim3 expression, but before lineage segregation. aal and pit1 are required for two distinct, but partly overlapping adenohypophyseal lineages, with pit1 governing lactotrope, somatrope, and thyrotrope development, and aal governing somatotrope, thyrotrope, corticotrope, and melanotrope development.

could not be recovered as yet due to logistic prob- sense morpholino oligonucleotides (51), according to lems. The remaining eight mutations fall into four which inactivation of lim3 leads to the loss of all ad- complementation groups, defining four genes named enohypophyseal cell lineages except some pomc cells lia, pia, pit1, and aal, all of which appear to be required (Herzog, W., and M. Hammerschmidt, unpublished for different steps of adenohypophysis development. data). Final analyses of the epistatic relationships between lia, lim3, and pia will only be possible after the lia Genes Required Upstream and Downstream of and pia genes have been cloned. Lim3 to Drive Development of All To gain further insight into the biological roles of lia Adenohypophyseal Cell Types and pia during the development of the zebrafish adenohypophysis, future experiments must reveal the fate Two genes, lia and pia, are required for all adenohy- of the pituitary precursor cells in the mutants. Thus, pophyseal cell types. In contrast, other placodal de- cell-tracing experiments (44), bromodeoxyuridine in- rivatives appear to develop normally in mutant em- corporation studies, and transferase deoxyuridine bryos, as for instance indicated by the normal triphosphate nick-end labeling or acridine orange expression of marker genes of the olfactory epithelium stainings will be carried out to investigate whether (Herzog, W., and M. Hammerschmidt, unpublished adenohypophyseal cells are lost because of transfat- data), which derives from placodal positions located ing, failed proliferation, or cell death. In addition, anal- left and right of the adenohypophyseal placode at the yses of chimeric embryos generated via cell transplan- anterior neural ridge (45). This indicates that lia and pia tation will help to specify in which cell types the two specifically act on early steps of adenohypophyseal genes are required. placode formation or maintenance, whereas other placodes are regulated differently. During adenohy- Genes Required for Lineage Specification: pophyseal development, lia and pia appear to be re- Evidence for a Conserved and a Novel Lineage of quired for different, maybe subsequent steps, as Adenohypophyseal Cell Types indicated by the differences in the lim3 expression in the two mutants. Thus, lia appears to act at an earlier In contrast to lia and pia, pit1 and aal mutants lose only step, before lim3 expression in the adenohypophyseal some of the adenohypophyseal cell types, indicating placode is initiated, whereas pia acts after the initiation that they are required for adenohypophyseal pattern- of lim3 expression, as indicated in Fig. 3. However, this ing and lineage specification processes during later does not necessarily mean that lia fulfills its indispens- stages of pituitary development. pit1 mutants lack gh, able function via the activation of lim3, nor does it prl, and tsh staining, whereas pomc staining is normal necessarily mean that lim3 is required for pia activa- or even enlarged, particularly in its posterior domain tion. lia could as well activate other essential genes that accommodates both corticotropes and melano- that act in parallel or in addition to lim3, and pia could tropes (29). The phenotype looks very similar to that of as well act in parallel rather than downstream of lim3. mouse Pit1 mutants, characterized by the loss of thy- However, it is interesting to note that pia mutants— rotropes, somatotropes, and lactotropes. Due to the despite the early presence of lim3 transcripts—fail to lack of suitable markers, we could not study the effect express even early adenohypophyseal hormone genes of zebrafish pit1 on the gonadotropes, that according like prl. This indicates that in the absence of pia, lim3 to the mouse mutant should be present in pit1 mutant is not sufficient for early lactotrope specification— fish. One subtle difference in the phenotypes of mouse although it appears to be necessary. The latter is sug- and zebrafish mutants is that zebrafish mutants lack all gested by our preliminary results obtained with anti- tsh expression, whereas in mouse mutants, only the 1192 Mol Endocrinol, May 2004, 18(5):1185–1195 Herzog et al. • Adenohypophysis Formation in Zebrafish

caudo-medial thyrotropes are lost, although the earlier The Cloning of the Mutated Pituitary Genes specifying rostral-tip thyrotropes are present (52) (for review, see Ref. 17). Tsh expression in this special Because the introduction of point mutations during the thyrotrope lineage is independent of Pit1, and most ENU mutagenesis is random, the molecular nature of likely regulated by the leucine zipper transcription fac- the genes affected in the isolated mutants is not tor TEF (thyrotrope embryonic factor; Refs. 52 and 53). known a priori, but has to be determined in subse- The functional importance of the rostral-tip thyrotrope quent steps. However, with the ease of mutant embryo cells is not clear (17). However, our data obtained for collection, together with recent progress in the ze- the zebrafish pit1 mutant suggest that they might be a brafish genome projects, positional cloning of such specialty of mammals. ENU-mutated genes has become relatively easy, and has been successfully applied in multiple cases (55,

In contrast to pit1, the zebrafish aal mutants appear Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021 to define a thus far unknown lineage of adenohy- 56). Using a combination of positional cloning and pophyseal cell types, lacking somatotropes, thyro- candidate testing, we have for instance been able to tropes, corticotropes, and melanotropes, but not the identify the molecular nature of our two zebrafish pit1 lactotropes. Such a combination of lost cell types had alleles (50). Also, future cloning of lia, pia, and aal will not been previously described. In addition, the parallel hopefully identify novel genes in pituitary develop- existence of the Pit1 and Aal lineages indicates mech- ment. In addition, we will be able to investigate anisms of lineage specification beyond the thus far whether such genes are specifically required for pitu- believed subsequent branching off of lineages from a itary development in fish, or whether their roles are common precursor pool. Consistent with such a linear conserved between fish and mammals. mechanism, only two relative patterns of cell lineages had been observed thus far. Lineages were either Possible Reasons for the Low Number of complementary (such as the Pomc lineage and the Identified Pituitary Genes and Perspectives Prop1 lineage), or part of each other (such as Pit1, Based on gene targeting and cloning of spontaneous consisting of somatotropes, lactotropes, and thyro- mutations, a total of approximately 20 genes indis- tropes, and Prop1, consisting of the Pit lineage plus pensable for pituitary development in the mouse were the gonadotropes). In contrast, the zebrafish Pit1 and identified (for reviews, see Refs. 15 and 17). These Aal lineages show a novel relative pattern, with shared include genes required for the combined formation of cell types, as well as cell types specific for one or the hypothalamus, neurohypophysis, and adenohypophy- other lineage (see Fig. 3). Thus, somatotropes and sis (such as Nkx2.5 and Shh; Ref. 4, 38, 39), genes thyrotropes appear to belong both to the Pit1 and the required for hypothalamus and neurohypophysis only Aal lineage. However, corticotropes and melanotropes (such as Brn2; Refs. 57 and 58), and genes required for only belong to the Aal lineage and are independent of the adenohypophysis only (such as Lhx3, Lhx4, Prop1, Pit1, whereas the lactotropes belong to the Pit1 lin- Pit1). Our screen was designed to find adenohypoph- eage only and are independent of Aal. This indicates ysis-specific defects only. Pituitary phenotypes that factors driving cell lineage specification are not caused secondarily due to loss of the hypothalamus only used in a mutually exclusive or consecutive fash- (such as in mouse Nkx2.5 or Shh mutants) would have ion, but can also be recruited in different combinations been missed because screening was performed at a to allow differential parallel cell specifications. late developmental stage, after mutations in such In mammals, no Aal lineage has been identified as genes should have been lethal. yet. However, this does not necessarily mean that the In addition, the zebrafish mutants were screened for mammalian Aal homolog is not required or involved in altered gh expression only—the only adenohypophy- adenohypophysis development. Rather, the exclusive seal hormone gene that had been cloned at the time regulation of prl expression independently of Aal func- the screen was carried out. The gh probe should have tion might be a specialty of fish, consistent with the allowed us to identify genes required for adenohy- much earlier onset of prl expression in fish compared pophysis induction and development in general, as with mouse, and with the earlier and additional func- well as genes specifically required for somatotropes tion of Prl during osmoregulation in water-living larvae and somatotropes-including lineages, such as Lhx3, (1, 27, 54). Along these lines, it is tempting to specu- Prop1, and Pit1. On one hand, gh was a good choice, late that during the evolution of water-living verte- because it allowed us to identify genes that would brates, lactotropes became independent of Aal, or that most likely have been missed with other hormone during evolution of land-based vertebrates, lacto- probes (pit1 mutants still express pomc; aal mutants tropes got under the control of the Aal homolog, sim- still express prl). On the other hand, we might have ilar to the other pituitary cell types. In this case, mu- missed genes that would have been revealed with tations in the Aal homolog in mouse would affect all other probes, such as Tpit/Tbx19 (affecting the Pomc adenohypophyseal cell lineages. Final answers have lineage only), Gata2 (affecting gonatotropes and thy- to await the cloning of the zebrafish aal gene, and the rotropes only), or Sf1 (affecting gonadotropes only). In identification and analysis of possible mammalian light of these possible limitations of the gh probe, we homologues. are currently preparing another large-scale ENU Herzog et al. • Adenohypophysis Formation in Zebrafish Mol Endocrinol, May 2004, 18(5):1185–1195 1193

screen, looking for the expression of pit1, lim3, isoto- of the other mutation. If the mutations are alleles of the cin, pomc, and prl at early and late stages of zebrafish same gene, they fail to complement each other in trans- heterozygous embryos, which show the mutant phenotype development. like homozygotes of either allele. If the mutations are in Another reason for the low number of identified es- different genes, the double-heterozygous offspring show a sential zebrafish genes could be the larger size and wild-type phenotype. complexity of the zebrafish genome. Due to the additional genome duplication that has occurred during Mapping teleost evolution (59), genes controlling pituitary development in zebrafish might display a higher degree Genomic localization of zebrafish mutations was performed as described, using the Tu¨ bingen marker set for genome of functional redundancy than in mammals. However, scans (version 4) on F2 Tuebingen ϫ Wik crosses of the genome analyses suggest that only approximately mutant carriers (55). Primer sequences are available from the Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021 25% of the genes gained in the duplication have re- Massachusetts General Hospital web site (http://zebrafish. mained active, with most of them having evolved a mgh.harvard.edu). The linkage groups (genetic equivalents to different expression pattern than their paralogs (59, chromosones) determined for the different mutations are given in Table 1. 60). Also, parallel searches for zebrafish genes required for other developmental processes, such as Cloning of the Zebrafish Isotocin cDNA angiogenesis (25) and thymus development (Thomas Boehm, Max-Planck Institute for Immunobiology, Isotocin cDNA was cloned via degenerate RT-PCR, using the Freiburg, Germany; personal communication) have following primers: sense 5ЈTGY TAY ATH CAR AAY TGY CC, yielded many more mutants, suggesting that high antisense 5ЈCCR CAR CAD ATN BWN GGN CC, with 35 functional redundancy with a low number of indis- cycles and an annealing temperature of 53 C. To obtain the full-length cDNA sequence, this was followed by a 3Ј-rapid pensable genes might be a feature of some, but not all amplification of cDNA ends (RACE), using the sense primer processes of zebrafish development. On the other and the SMART RACE Kit (CLONTECH, Palo Alto, CA) ac- hand, we cannot rule out general technical problems cording to the manufacturer’s instructions. PCR fragments during our mutagenesis. Comparing our mutagenesis were cloned into pCRII (Invitrogen, Carlsbad, CA). For isotocin in situ antisense probe synthesis, the plasmid was di- conditions with those of previous screens (23), the gested with KpnI and transcribed using T7 RNA polymerase. screening of over 4200 mutagenized genomes should have yielded 50–80% saturation, with an average al- In Situ Hybridizations lele frequency per gene between three and four. How- ever, in case of our identified pituitary genes, the av- For initial screening, F3 clutches were incubated in the pres- erage allele rate was two, with single alleles for two of ence of 0.25 mM 1-phenyl-2-thiourea (Sigma, St. Louis, MO) the four genes. This suggests that the ENU mutagen- to avoid melanin synthesis, and fixed in 4% paraformalde- hyde/PBS at 120 hpf. Whole mount in situ hybridization was esis might have been less efficient than in previous carried out in specially designed 48-well plates (Aldinger, screens. Further complementation testing of other Nagold, Germany) with digoxygenin-labeled probes for gh mutant classes has to be carried out for final conclu- (27), and rag1, a marker for thymic T cells (61), after standard sion about the mutation rates. protocols (62). Red/blue double in situ hybridizations were carried out as described (27). prl, pomc, tsh, gh (27), lim3 (31), and nkx2.1 (41) probes were synthesized as reported (27).

Acknowledgments MATERIALS AND METHODS The large-scale mutant screen was carried out in collabo- ENU Mutagenesis, Inbreeding Schedule, and ration with the Tuebingen 2000 screen consortium, whose Complementation Testing members were: from the Max-Planck Institute for Develop- mental Biology: F. van Bebber, E. Busch-Nentwich, R. Dahm, Adult zebrafish males of the Tu¨ bingen line were incubated in H. G. Frohnho¨ fer, H. Geiger, D. Gilmour, S. Holley, J. Hooge, buffered E3 medium containing ENU as previously reported D. Ju¨ lich, H. Knaut, F. Maderspacher, H.-M. Maischein, C. (30). After mutagenesis, males were repeatedly mated to Neumann, T. Nicolson, C. Nu¨ sslein-Volhard, H. H. Roehl, U. untreated females to generate over 10,000 F1 fish. Such F1 Scho¨ nberger, C. Seiler, C. So¨ llner, M. Sonawane, A. Wehner, fish were crossed to each other to generate over 5000 F2 and C. Weiler; from Exelixis Germany GmbH: P. Erker, H. families that underwent brother-sister-crosses to generate Habeck, U. Hagner, C. Hennen, E. Kaps, A. Kirchner, T. homozygous larvae in the resulting F3 clutches (23, 30). Koblitzek, U. Langheinrich, C. Loeschke, C. Metzger, R. Nor- Twenty to 30 larvae of each F3 clutch were shipped from din, J. Odenthal, M. Pezzuti, K. Schlombs, J. deSatana- Tuebingen to Freiburg, where they were screened via gh in Stamm, T. Trowe, G. Vacun, B. Walderich, A. Walker, and C. situ hybridization (see In Situ Hybridizations). A clutch was Weiler. In addition, we would like to thank the members of the considered positive when 20–30% of the larvae showed a Freiburg screening group (Michael Schorpp, Markus Leicht, similar alteration in the gh expression pattern or intensity. To Elvira Nold, Thorsten Nolting, Carolin Riegger, Dagmar Diek- confirm and recover mutations, F2 pairs producing putative hoff and Tanna Franz) for mutant screening via gh in situ mutants among their F3 offspring were crossed out, and the hybridization. We are very grateful to Xianchun Zeng for his resulting new F3 families underwent the same inbreeding and help during the cloning of isotocin, Donatus Boensch for screening procedure. excellent fish care, and Thomas Boehm for support and To determine whether two mutations causing similar discussions. Many thanks also to Jing-Wen Ting and Chi-Yao phenotypes reside in the same or in two different genes, Chang for sending us the gh plasmid before publication, and complementation analyses were performed, crossing a to Igor Dawid (lim3) and Klaus Rohr (nkx2.1) for published heterozygous fish of one mutation with a heterozygous fish reagents. 1194 Mol Endocrinol, May 2004, 18(5):1185–1195 Herzog et al. • Adenohypophysis Formation in Zebrafish

Received September 26, 2003. Accepted January 21, 16. Kioussi C, Carrie` re C, Rosenfeld MG 1999 A model for 2004. the development of the hypothalamic-pituitary axis: tran- Address all correspondence and requests for reprints to: scribing the hypophysis. Mech Dev 81:23–35 Matthias Hammerschmidt, Max-Planck Institute for Immuno- 17. Burrows HL, Douglas KR, Seasholtz AF, Camper SA biology, Stuebeweg 51, 79108 Freiburg, Germany. E-mail: 1999 Genealogy of the anterior pituitary gland: tracing a [email protected]. family tree. Trends Endocrinol Metab 10:343–352 Work in the laboratory of M.H. was supported by the 18. Sheng HZ, Westphal H 1999 Early steps in pituitary or- Max-Planck Society. ganogenesis. Trends Genet 15:236–240 19. Burgess R, Lunyak V, Rosenfeld MG 2002 Signaling and transcriptional control of pituitary development. Curr Opin Genet Dev 12:534–539 REFERENCES 20. Scully KM, Rosenfeld MG 2002 Pituitary development: regulatory codes in mammalian organogenesis. Science 295:2231–2235 Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021 1. Norris DO 1997 Vertebrate endocrinology. 3rd ed. San 21. Zhao L, Bakke M, Krimkevich Y, Cushman LJ, Parlow AF, Diego: Academic Press Camper SA, Parker KL 2001 Steroidogenic factor 1 (SF1) 2. Young JI, Otero V, Cerda´ n MG, Falzone TL, Chan EC, is essential for pituitary gonadotrope function. Develop- Low MJ, Rubinstein M 1998 Authentic cell-specific and ment 128:147–154 developmentally regulated expression of pro-opiomela- 22. Dorton AM 2000 The pituitary gland: embryology, phys- nocortin genomic fragments in hypothalamic and hind- iology, and pathophysiology. Neonatal Netw 19:9–17 brain neurons of transgenic mice. J Neurosci 18: 23. Haffter P, Granato M, Brand M, Mullins MC, Hammer- 6631–6640 schmidt M, Kane DA, Odenthal J, van Eeden FJM, Jiang 3. Dubois P, ElAmraoui A 1995 Embryology of the pituitary Y-J, Heisenberg C-P, Kelsh RN, Furutani-Seiki M, Warga gland. Trends Endocrinol Metab 6:1–7 RM, Vogelsang E, Beuchle D, Schach U, Fabian C, 4. Kimura S, Hara Y, Pineau T, Fernandez-Salguero P, Fox Nu¨ sslein-Volhard C 1996 The identification of genes with CH, Ward JM, Gonzalez FJ 1996 The T/ebp null mouse: unique and essential functions in the development of the thyroid-specific enhancer-binding protein is essential for zebrafish, Danio rerio. Development 123:1–36 the organogenesis of thyroid, lung, ventral forebrain, and 24. Driever W, Solnica-Krezel L, Schier AF, Neuhauss SCF, pituitary. Genes Dev 10:60–69 Malicki J, Stemple DL, Stanier DYR, Zwartkruis F, Abde- 5. Gleiberman AS, Fedtsova NG, Rosenfeld MG 1999 Tis- liah S, Rangini Z, Belak J, Boggs C 1996 A genetic screen sue interactions in the induction of anterior pituitary: role for mutations affecting embryogenesis in zebrafish. De- of the ventral diencephalon, mesenchyme, and noto- velopment 123:37–46 chord. Dev Biol 213:340–353 25. Habeck H, Odenthal J, Walderich B, Maischein HM, 6. Sheng HZ, Zhadanov AB, Mosinger Jr B, Fujii T, Bertuzzi Schulte-Merker S, Tuebingen screen consortium 2002 S, Grinberg A, Lee EJ, Huang S-P, Mahon KA, Westphal Analysis of a zebrafish VEGF receptor mutant reveals H 1996 Specification of pituitary cell lineages by the LIM specific disruption of angiogenesis. Curr Biol 12: homeobox gene. Lhx3. Science 272:1004–1007 1405–1412 7. Sheng HZ, Moriyama K, Yamashita T, Li H, Potter SS, 26. Knaut H, Werz C, Geisler R, Tuebingen screen consor- Mahon KA, Westphal H 1997 Multistep control of pitu- tium, Nu¨ sslein-Volhard C 2003 A zebrafish homologue of itary organogenesis. Science 278:1809–1812 the chemokine receptor Cxcr4 is a germ-cell guidance 8. Sornson M, Wu W, Dasen J, Flynn A, Norman D, receptor. Nature 421:279–282 O’Connell S, Gukovsky I, Carrie` re C, Ryan A, Miller A, 27. Herzog W, Zeng X, Lele Z, Sonntag C, Ting J-W, Chang Zuo L, Gleiberman A, Andersen B, Beamer W, Rosenfeld C-Y, Hammerschmidt M 2003 Adenohypophysis forma- MG 1996 Pituitary lineage dertermination by the Prophet tion in the zebrafish and its dependence on Sonic Hedge- of Pit-1 homeodomain factor defective in Ames dwarf- hog. Dev Biol 254:36–49 ism. Nature 384:327–333 28. Sbrogna JL, Barresi MJF, Karlstrom RO 2003 Multiple 9. Li S, Crenshaw III EB, Rawson EJ, Simmons DM, Swan- roles for Hedgehog signaling in zebrafish pituitary devel- son LW, Rosenfeld MG 1990 Dwarf locus mutants lack- opment. Dev Biol 254:19–35 ing three pituitary cell types result from mutations in the 29. Liu N-A, Huang H, Yang Z, Herzog W, Hammerschmidt POU-domain gene pit-1. Nature 347:528–533 M, Lin S, Melmed S 2003 Pituitary corticotroph ontogeny 10. Pulichino A-M, Vallette-Kasic S, Tsai JP-Y, Couture C, and regulation in transgenic zebrafish. Mol Endocrinol Gauthier Y, Drouin J 2003 Tpit determines alternate fates 17:959–966 during pituitary cell differentiation. Genes Dev 17: 30. Mullins MC, Hammerschmidt M, Haffter P, Nu¨ sslein- 738–747 Volhard C 1994 Large-scale mutagenesis in the 11. Lamolet B, Pulichino A-M, Lamonerie T, Gauthier Y, Brue zebrafish: in search of genes controlling development in T, Enjalbert A, Drouin J 2001 A pituitary cell-restricted T a vertebrate. Curr Biol 4:189–202 box factor, Tpit, activates POMC transcription in coop- 31. Glasgow E, Karavanov AA, Dawid IB 1997 Neuronal and eration with Pitx homeoproteins. Cell 104:849–859 neuroendocrine expression of lim3, a LIM class ho- 12. Liu J, Lin C, Gleiberman A, Ohgi KA, Herman MA, Huang meobox gene, is altered in mutant zebrafish with axial H, Tsai MJ, Rosenfeld MG 2001 Tbx19, a tissue-selective signaling defects. Dev Biol 192:405–419 regulator of POMC gene expression. Proc Natl Acad Sci 32. Brand M, Heisenberg C-P, Jiang Y-J, Karlstron RO, USA 98:8674–8679 Warga RM, Pelegri F, van Eeden FJM, Furutani-Seiki M, 13. Dasen JS, OЈConnell SM, Flynn SE, Treier M, Gleiberman Granato M, Haffter P, Hammerschmidt M, Kane DA, AS, Szeto DP, Hooshmand F, Aggarwal AK, Rosenfeld Kelsh RN, Mullins MC, Odenthal J, Nu¨ sslein-Volhard C MG 1999 Reciprocal interactions of Pit1 and GATA2 1996 Mutations affecting development of the midline and mediate signaling gradient-induced determination of pi- general body shape during zebrafish embryogenesis. tuitary cell types. Cell 97:587–598 Development 123:129–142 14. Treier M, Rosenfeld MG 1996 The hypothalamic-pituitary 33. Rebagliati MR, Toyoma R, Haffter P, Dawid IB 1998 axis: co-development of two organs. Curr Opin Cell Biol Cyclops encodes a Nodal-related factor involved in mid- 8:833–843 line signaling. Proc Natl Acad Sci USA 95:9932–9937 15. Watkins-Chow DE, Camper SA 1998 How many ho- 34. Sampath K, Rubinstein AL, Cheng AMS, Liang JO, Fe- meobox genes does it take to make a pituitary gland? kany K, Solnica-Krezel L, Korzh V, Halpern ME, Wright Trends Genet 14:284–290 CVE 1998 Induction of the zebrafish ventral brain and Herzog et al. • Adenohypophysis Formation in Zebrafish Mol Endocrinol, May 2004, 18(5):1185–1195 1195

floorplate requires cyclops/nodal sigalling. Nature 395: 49. Ingraham HA, Chen R, Mangalam HJ, Elsholtz HP, Flynn 185–189 SE, Lin CR, Simmons DM, Swanson LW, Rosenfeld MG 35. Kondoh H, Uchikawa M, Yoda H, Takeda H, Furutani- 1988 A tissue-specific transcription factor containing a Seiki M, Karlstrom RO 2000 Zebrafish mutations in Gli- homeodomain specifies pituitary phenotype. Cell 50: mediated hedgehog signaling lead to lens transdifferen- 519–529 tiation from the adenohypophysis anlage. Mech Dev 96: 50. Nica G, Herzog W, Sonntag C, Hammerschmidt M 2004 165–174 Zebrafish pit1 mutants lack three pituitary cell types and 36. Varga ZM, Amores A, Lewis KE, Yan YL, Postlethwait JH, develop severe dwarfism. Mol Endocrinol 18:1196–1209 Eisen JS, Westerfield M 2001 Zebrafish smoothened 51. Nasevicius A, Ekker SC 2000 Effective targeted gene functions in ventral neural tube specification and axon “knockdown” in zebrafish. Nat Genet 26:216–220 tract formation. Development 128:3497–3509 52. Lin S-C, Li S, Drolet DW, Rosenfeld MG 1994 Pituitary 37. Takuma N, Sheng HZ, Furuta Y, Ward JM, Sharma K, ontogeny of the Snell dwarf mouse reveals Pit-1-inde- Hogan BLM, Pfaff SL, Westphal H, Kimura S, Mahon KA pendent and Pit-1-dependent origins of the thyrotrope.

1998 Formation of Rathke’s pouch requires dual in- Development 120:515–522 Downloaded from https://academic.oup.com/mend/article/18/5/1185/2741872 by guest on 27 September 2021 duction from the diencephalon. Development 125: 53. Drolet DW, Scully KM, Simmons DM, Swanson LW, 4835–4840 Wegner M, Chu K, Rosenfeld MG 1991 TEF, a transcrip- 38. Pabst O, Herbrand H, Takuma N, Arnold HH 2000 Nkx2 tional factor expressed specifically in the anterior pitu- gene expression in neuroectoderm, but not in mesend- itary during embryogenesis, defines a new class of leucin odermally derived structures depends on Sonic hedge- zipper proteins. Genes Dev 5:1739–1753 hog in mouse embryos. Dev Genes Evol 210:47–50 54. Kelly PA, Djiane J, Postel-Vinay MC, Edery M 1991 The 39. Treier M, O’Connell S, Glieberman A, Price J, Szeto DP, prolactin/growth hormone receptor family. Endocr Rev Burgess R, Chuang PT, McMahon AP, Rosenfeld MG 12:235–251 2001 Hedgehog signaling is required for pituitary gland 55. Geisler R 2002 Mapping and cloning. In: Nu¨ sslein- development. Development 128:377–386 Volhard C, Dahm R, eds. Zebrafish: a practical approach. 40. Sprague J, Doerry E, Douglas S, Westerfield M 2001 The Oxford, UK: Oxford University Press; vol 261:175–212 Zebrafish Information Network (ZFIN): a resource for ge- 56. Talbot WS, Schier AF 1999 Positional cloning of mutated netic, genomic and developmental research. Nucleic Ac- zebrafish genes. Methods Cell Biol 60:259–286 ids Res 29:87–90 57. Nakai S, Kawano H, Yudate T, Nishi M, Kuno J, Nagata 41. Rohr KB, Concha ML 2000 Expression of nk2.1a during A, Jishage K, Hamada H, Fujii H, Kawamura K, Shiba K, early development of the thyroid gland in zebrafish. Mech Noda T 1995 The POU domain transcription factor Brn-2 Dev 95:267–270 is required for the determination of specific neuronal 42. Unger JL, Glasgow E 2003 Expression of isotocin- lineages in the hypothalamus of the mouse. Genes Dev neurophysin mRNA in developing zebrafish. Gene Expr 9:3109–3121 Patterns 3:105–108 58. Schonemann MD, Ryan AK, McEvilly RJ, O’Connell SM, 43. Gainer H, Wray S 1994 Cellular and molecular biology of Arias CA, Kalla KA, Li P, Sawchenko PE, Rosenfeld MG oxytocin and vasopressin. In: Knobil E, Neill JD, eds. The 1995 Development and survival of the endocrine hypo- physiology of reproduction. New York: Raven Press; thalamus and posterior pituitary gland requires the neu- 1099–1129 ronal POU domain factor Brn-2. Genes Dev 9:3122–3135 44. Whitlock KE, Westerfield M 2000 The olfactory placodes 59. Postlethwait JH, Yan Y-L, Gates MA, Horne S, Amores A, of the zebrafish form by convergence of cellular fields at Brownlie A, Donovan A, Egan ES, Force A, Gong Z, the edge of the neural plate. Development 127: Goutel C, Fritz A, Kelsh R, Knapik E, Liao E, Paw B, 3645–3653 Ransom D, Singer A, Thomson M, Abduljabbar TS, Yelik 45. Whitlock KE, Wolf CD, Boyce, ML 2003 Gonadotropin- P, Beier D, Joly J-S, Larhammar D, Rosa F, Westerfield releasing hormone (GnRH) cells arise from cranial neural M, Zon LI, Johnson SL, Talbot WS 1998 Vertebrate ge- crest and adenohypophyseal regions of the neural plate nome evolution and the zebrafish gene map. Nat Genet in the zebrafish, Danio rerio. Dev Biol 257:140–152 18:345–349 46. Kazanskaya OV, Severtzova EA, Barth AK, Ermakova GV, 60. Postlethwait JH, Woods IG, Ngo-Hazelett P, Yan Y-L, Lukyanov SA, Benyumov AO, Pannese M, Boncinelli E, Kelly PD, Chu F, Huang H, Hill-Force A, Talbot WS 2000 Wilson SW, Zaraisky AG 1997 Anf: a novel class of ver- Zebrafish comparative genomics and the origins of vertebrate homeobox genes expressed at the anterior end tebrate chromosomes. Genome Res 10:1890–9051 of the main embryonic axis. Gene 200:25–34 61. Willett CE, Zapata AG 1997 Expression of zebrafish rag 47. Japo´ n MA, Rubinstein M, Low MJ 1994 In situ hybrid- genes during early development identifies the thymus. ization analysis of anterior pituitary hormone gene ex- Dev Biol 182:331–341 pression during fetal mouse development. J Histochem 62. Hammerschmidt M, Pelegri F, Mullins MC, Kane DA, van Cytochem 42:1117–1125 Eeden FJM, Granato M, Brand M, Furutani-Seiki M, Haff- 48. Bodner M, Castrillo J-L, Theill LE, Deerinck T, Ellisman ter P, Heisenberg C-P, Jiang Y-J, Kelsh RN, Odenthal J, M, Karin M 1988 The pituitary-specific transcription fac- Warga RM, Nu¨ sslein-Volhard C 1996 dino and mercedes, tor GHF-1 is a homeobox-containing protein. Cell 50: two genes regulating dorsal development in the zebrafish 267–275 embryo. Development 123:95–102

Molecular Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.