Shifting boundaries of retinoic acid activity control hindbrain segmental gene expression

Abstract

Retinoic acid (RA) generated by Raldh2 in paraxial mesoderm is required for specification of the posterior hindbrain, including restriction of Hoxb1 expression to presumptive rhombomere 4 (r4). Hoxb1expression requires 3′ and 5′ RA response elements for widespread induction up to r4 and for r3/r5 repression, but RA has previously been detected only from r5-r8, and vHnf1 is required for repression of Hoxb1 posterior to r4 in zebrafish. We demonstrate in mouse embryos that an RA signal initially travels from the paraxial mesoderm to r3, forming a boundary next to the r2 expression domain of Cyp26a1 (which encodes an RA-degrading enzyme). After Hoxb1 induction, the RA boundary quickly shifts to r4/r5, coincident with induction of Cyp26c1 in r4. A functional role for Cyp26c1 in RA degradation was established through examination of RA-treated embryos. Analysis of Raldh2–/– and vHnf1–/– embryos supports a direct role for RA in Hoxb1 induction up to r4 and repression in r3/r5, as well as an indirect role for RA in Hoxb1 repression posterior to r4 via RA induction of vHnf1 up to the r4/r5 boundary. Our findings suggest that Raldh2 and Cyp26 generate shifting boundaries of RA activity, such that r3-r4 receives a short pulse of RA and r5-r8 receives a long pulse of RA. These two pulses of RA activity function to establish expression of Hoxb1 and vHnf1 on opposite sides of the r4/r5 boundary.

Fate of retinoic acid–activated embryonic cell lineages

“…The graded, mosaic distribution of R26R-excised cells observed in r5 through r4/r3 is consistent with notion of a gradient of RA activity, high in the posterior hindbrain (leading to strong RARE-Cre activity and full excision of the reporter), and fading anteriorly; it also supports the model proposed by Sirbu et al (2005) of a short pulse or RA signaling in the r3/r4 prospective territory before the onset of CYP26C1 activity (see the Discussion section).…”

“…ab, alveolar bone; ah, adenohypophysis; ca, cartilage; co, cornea; cp, choroid plexus; de, dermis; dr, dorsal retina; ds, dental sac; ep, epidermis; hf, hair follicle; id, intervertebral disk; in, incisor; le, lens; m1,m2, molar buds; mu, muscle; my, myelencephalon; ng, nasal gland; nh, neurohypophysis; np, nucleus pulposus; ns, nasal septum; oe, olfactory epithelium; ra, radius; re, respiratory epithelium; sb, sphenoid bone; sc, spinal cord; ve, vertebra; vi, vibrissae; vr, ventral retina. have been proposed to account for the RA effects in the developing hindbrain. Sirbu et al (2005) described two phases of RA action, an early phase before overt segmentation with RA reaching the prospective r3, followed by a phase where the RA boundary shifts to the r4-r5 boundary after induction of Cyp26c1 expression in r4. The pattern of reporter excision observed in RARE-Cre;R26R embryos may reflect these two phases of RA activity, considering that during the first phase the pulse of RA may be too short to activate the RARE-Cre transgene and generate sufficient amounts of Cre, except in a few pre-r3/r4 cells.…”

“…Conditions of RA excess through maternal administration (Marshall et al, 1992;Simeone et al, 1995, and references therein) or Cyp26 knock-down Sakai et al, 2001;Emoto et al, 2005;Hernandez et al, 2007), and of RA deficiency through mutation of Rdh10 or Raldh2 (Niederreither et al, 2000;Sandell et al, 2007), have severe effects on hindbrain development and can lead to changes in rhombomeric molecular identities. Several models have been proposed to account for the effects of endogenous RA in controlling hindbrain development, mainly relying on the establishment of a posterior-toanterior RA diffusion gradient, and/or on the sequential action of CYP26 enzymes to generate RA-depleted areas in specific rhombomere territories (Gavalas, 2002;Maden, 2002;Maves and Kimmel, 2005;Sirbu et al, 2005;Hernandez et al, 2007;White et al, 2007;White and Schilling, 2008).…”

“…The rationale for these studies is that most, if not all, tissues express at least one of the RARs (Dolle, 2009 for a review), that should activate the reporter upon binding of endogenous RA. The RARE-hsp68-lacZ transgene has been widely used in subsequent studies (e.g., Malpel et al, 2000;Mic et al, 2002;Niederreither et al, 2002b;Fan et al, 2003;Matt et al, 2003Matt et al, , 2005Yashiro et al, 2004;Sirbu et al, 2005;Molotkov et al, 2006;Sirbu and Duester, 2006;Ribes et al, 2009;Zhao et al, 2009). Three lines of evidence have validated this transgene as a reliable indicator of RA activity: (1) there is a close match between the patterns of lacZ activity and the expression patterns of RA-synthesizing enzymes (Rdh10, Raldhs) at early embryonic stages, (2) activity of the lacZ reporter is severely down-regulated in knockout mice for the above enzymes (Niederreither et al, 1999;Mic et al, 2002;Matt et al, 2005;Molotkov et al, 2006;Sandell et al, 2007), (3) on the other hand, the reporter is almost ubiquitously activated within 6 h after administration of a bolus dose of RA to the pregnant mothers (Rossant et al, 1991), and is ectopically activated when RA-metabolizing enzymes are absent Sakai et al, 2001;Uehara et al, 2009).…”

“…This would provide important information with respect to various developing systems, including the brain (Dupe et al, 1999;Smith et al, 2001;Sirbu et al, 2005;Ribes et al, 2006;Molotkova et al, 2007), sensory organs Matt et al, 2005;Molotkov et al, 2006;Romand et al, 2006), cranial neural crest cells and branchial arch derivatives (Wendling et al, 2000;Niederreither et al, 2003;Mark et al, 2004), and heart (Merki et al, 2005;Ryckebusch et al, 2008;Sirbu et al, 2008). Retinoids are also potent teratogenic compounds in human (Lammer et al, 1985;Hanson and Leachman, 2001;Charakida et al, 2004, and references therein), and excess RA administration affects many morphogenetic processes in mice (e.g., Kessel and Gruss, 1991;Marshall et al, 1992;Shum et al, 1999;Matt et al, 2003).…”