Deana D'Souza - Academia.edu (original) (raw)

Papers by Deana D'Souza

Development, Apr 1, 1996

and may therefore mediate cell contact-dependent interactions. We show that Cek-8 is expressed in... more and may therefore mediate cell contact-dependent interactions. We show that Cek-8 is expressed in distal mesenchyme of early chick limb buds with more pronounced expression posteriorly. Expression decreases as the limb develops but is then upregulated in mesenchymal cell condensations associated with formation of tendons, including their attachment sites to cartilage rudiments. As tendons mature, Cek-8 is downregulated in these structures but up-regulated in other patterned sites, feathers and scales. We have identified factors that influence early expression of Cek-8 in the developing limb and demonstrate that signals originating from the AER are required for Cek-8 induction and maintenance and that FGF-2 and FGF-4 can mimic these signals. Grafts of polarising region, retinoic acid or BMP-2 modulate the Cek-8 expression domain. We also investigated Cek-8 transcript distribution in limb buds of the polydactylous chick mutant talpid 3 to gain insights into its position in the signalling cascade that governs patterning and outgrowth. MATERIALS AND METHODS Chick embryos Fertilised chicken embryos were purchased from Poyndon Farm, Herts, England, and were incubated at 38°C. Embryos were staged according to Hamburger and Hamilton (1951). Following surgical manipulation, embryos were dissected into PBS and processed for whole-mount in situ hybridisation. FGF-2, retinoic acid and BMP-2 application on beads FGF-2 was obtained from R and D Systems and BMP-2 was provided by the Genetics Institute, Cambridge, Massachusetts. All-transretinoic acid was obtained from Sigma (UK). FGF-2 and retinoic acid were applied to heparin acrylic and AG1X2 beads respectively, as described by Niswander et al. (1993) and Tickle et al. (1985). BMP-2 was applied to heparin beads as described by Francis et al. (1994). Experimental manipulation of chick wings All manipulations were performed at stage 20 unless otherwise stated. K. Patel and others R. N. was in receipt of a Wellcome Prize Studentship. C. T. and D. D'S. are supported by the Wellcome Trust. We would like to thank Anne Crawley for help with the histology, and Professor Lewis Wolpert and Professor Paul Brickell for helpful comments on the manuscript.

Developmental Dynamics, 2004

The dorsal and ventral scales of the chick foot can be distinguished morphologically and molecula... more The dorsal and ventral scales of the chick foot can be distinguished morphologically and molecularly: the dorsal oblong overlapping scuta expressing both ␣ and ␤ keratins, and the ventral roundish nonprotruding reticula expressing only ␣ keratins. The question arises how En-1 and Lmx1, whose role in dorsoventral limb patterning has been well established, can affect skin morphogenesis, which occurs 8 to 12 days later. Forced expression of En-1 or of Lmx1 in the hindlimb have, respectively, as expected, a ventralizing or a dorsalizing effect on skin, leading to the formation of either reticula-type or scuta-type scales on both faces. In both cases, however, the scales are abnormal and even glabrous skin without any scales at all may form. The normal inductive interactions between dermis and epidermis are disturbed after En-1 or Lmx1 misexpression. Effectively, while Lmx1 endows the dermal precursors of the ventral region with scuta inducing ability, En-1 blocks the competence of the dorsal epidermis to build scuta.

Current Biology, 2003

and and steroids prevented healing of skin wounds [5]. In the latter study, retarded repair was c... more and and steroids prevented healing of skin wounds [5]. In the latter study, retarded repair was coincident with an Developmental Biology University College London accumulation of cell and matrix debris at the wound site. However, in these depletion studies, steroid treat-Gower Street, London WC1E 6BT United Kingdom ment may have knocked down more than just leukocytes at the wound site. Indeed, it is known that steroids are 2 Cancer Research UK, 61 Lincoln's Inn Fields inhibitory to the upregulation of immediate-early genes in response to growth factor inductive signals [6], and London WC2A 3PX United Kingdom these may be crucial activators of the repair process. The PU.1 null mouse provides an ideal opportunity to 3 The Burnham Institute 10901 North Torrey Pines Road more directly test the role of macrophages and neutrophils in the repair process, essentially by genetic abla-La Jolla, California 92037 tion; this ETS family transcription factor is hematopoietically restricted and essential for maturation and functional competence of several haemopoietic lin-Summary eages, and thus the PU.1 null mouse is born with neither macrophages nor functioning neutrophils [7, 8]. Damage to neonatal and adult tissues always incites an influx of inflammatory neutrophils and macro-We made full thickness 2 mm incisional wounds on the dorsal forepaw of antibiotic-maintained, neonatal phages. Besides clearing the wound of invading microbes, these cells are believed to be crucial coordina-PU.1 null mice derived from PU.1 heterozygote crosses (n ϭ 34). The wound site was labeled by brushing carbon tors of the repair process, acting both as professional phagocytes to clear wound debris and as a major into the incision immediately post-wounding. Wild-type mice from these litters reepithelialize such a wound in source of wound growth factor signals. Here we report wound healing studies in the PU.1 null mouse, which 3 days, sloughing the overlying scab on the fourth day. Staining for a neutrophil-specific enzyme, chloroacetate is genetically incapable of raising the standard inflammatory response because it lacks macrophages and esterase (CAE), reveals a peak of neutrophils at wound sites about 1 day post-wounding (Figure 1A), while in functioning neutrophils. Contrary to dogma, we show that these "macrophageless" mice are able to repair situ expression data for c-fms or immunostaining with the monocyte/macrophage-specific antibody, F4/80[9], skin wounds with similar time course to wild-type siblings, and that repair appears scar-free as in the em-shows an influx of macrophages from 12 hr after wounding, peaking at between 2 and 4 days (Figures bryo, which also heals wounds without raising an inflammatory response. The growth factor and cytokine 1C and 1CЈ), and remaining at the wound site for up to 5 or 6 days post-wounding. profile at the wound site is changed, cell death is reduced, and dying cells are instead engulfed by stand-As expected, PU.1 null littermates show an absence of CAE or F4/80 staining or c-fms expression at equivalent in phagocytic fibroblasts. We also show that hyperinnervation of the wound site, previously believed to be wound sites (Figures 1B, 1D, and 1DЈ). However, contrary to dogma, these mice, while unable to raise a cell a consequence of inflammation, is present in the PU.1 null wound, too. inflammatory response, show no retardation in the repair process, with all wounds examined at 4 days (n ϭ 27) being fully closed and the scab lost, just as for equivalent Results and Discussion wild-type wounds. These data strongly suggest that the inflammatory response is not absolutely essential for The repair of a skin wound is a complex process requirtissue repair. In fact, the healed epidermis at 3 days ing the collaborative efforts of many cell lineages to appeared more mature than in equivalent wild-type shore up the gap and to replace the missing tissues. It wounds (Figures 1E and 1F). Remarkably, at this stage, has long been considered that the inflammatory reresin histology through the PU.1 null wound connective sponse is instrumental in supplying growth factor and tissue showed no obvious difference from the conneccytokine signals that orchestrate the cell and tissue tive tissue in adjacent unwounded dermis (n ϭ 4; Figmovements necessary for repair [1]. Neutrophils are ure 1F). known to express several proinflammatory cytokines at To more fully challenge the wound repair capacity of the wound site [2], and macrophages from wound tisthese mice, we made full-thickness excisional wounds sues also express many of the growth factors that might of 2 mm diameter to the forepaws of a further 18 PU.1 regulate the repair process [3]. Moreover, a classic senull mice. In wild-type wounds of this sort, repair is ries of experiments in the 1970's showed that while achieved by a combination of reepithelialization and granulation tissue contraction [1]. Reepithelialization is

Anatomy and Embryology, 1999

Tendons connect muscle to skeletal elements. Although tendons have been shown to originate from t... more Tendons connect muscle to skeletal elements. Although tendons have been shown to originate from the lateral plate mesoderm, very little is known at the molecular level about how they are formed. We have found that two genes, Follistatin and Eph-A4, are expressed in regions associated with tendon formation in developing chick limbs. Follistatin is expressed near the tip of the digits and subsequently around the tendon, whereas Eph A4 transcripts were localized in a slightly more proximal region and later in the body of the tendon. Previous work has demonstrated that application of TGFβ1 or TGFβ2 to inter-digital regions or the removal of ectoderm in the foot plate induces ectopic cartilage formation, while removal of ectoderm or application of FGF to tips of developing digits leads to truncation. Here we show that TGFβ1 or removal of ectoderm is also able to induce the expression of both Eph-A4 and Follistatin and that manipulations that cause truncations affect these genes. Thus cartilage and tendon development appear to be coordinated. Ectopic application of recombinant human Follistatin, an antgaonist of certain TGFβ super-family proteins including Activin and Bmp-4, results in the loss of tendon, implicating signalling by TGFβ super-family in the development of tendon during chick embryogenesis. Signalling by TGFβ family members, antagonised by Noggin is known to regulate skeletal development. Thus we suggest that parallel pathways govern both skeletal and tendon patterning.

Journal of Anatomy, 1998

 A human polydactylous left foot with 9 toes, amputated from an 11-mo-old child, was exam... more  A human polydactylous left foot with 9 toes, amputated from an 11-mo-old child, was examined by -ray and magnetic resonance imaging and by gross dissection to identify the digits. The normal sequence of toes from medial to lateral is 1, 2, 3, 4, 5. Examination of the morphology of tendons and muscles suggested the toe sequence was 1, 2, 3\4, ?,5, 2, 3\4, 3\4, 5. The 2 toes in the sequence that are underlined were displaced dorsally and were found to have 2 extensor tendons, no flexor tendons and nails that were conical and situated at their tips. These toes resembled those described as ' double-dorsal ' and which develop in paws of mice in which a gene normally expressed ventrally is functionally inactivated (Loomis et al. 1996). Specification of toe formation occurs in leg buds early in embryonic development and later there is rotation of the limb so that the anterior (rostral) part comes to lie medially, i.e. the hallux which was anterior (rostral) now is on the inner (medial) side of the foot. A disruption in the patterning of this foot in both anteroposterior (rostral-caudal) and dorsoventral axes during development could be responsible.

Development, Apr 1, 1996

and may therefore mediate cell contact-dependent interactions. We show that Cek-8 is expressed in... more and may therefore mediate cell contact-dependent interactions. We show that Cek-8 is expressed in distal mesenchyme of early chick limb buds with more pronounced expression posteriorly. Expression decreases as the limb develops but is then upregulated in mesenchymal cell condensations associated with formation of tendons, including their attachment sites to cartilage rudiments. As tendons mature, Cek-8 is downregulated in these structures but up-regulated in other patterned sites, feathers and scales. We have identified factors that influence early expression of Cek-8 in the developing limb and demonstrate that signals originating from the AER are required for Cek-8 induction and maintenance and that FGF-2 and FGF-4 can mimic these signals. Grafts of polarising region, retinoic acid or BMP-2 modulate the Cek-8 expression domain. We also investigated Cek-8 transcript distribution in limb buds of the polydactylous chick mutant talpid 3 to gain insights into its position in the signalling cascade that governs patterning and outgrowth. MATERIALS AND METHODS Chick embryos Fertilised chicken embryos were purchased from Poyndon Farm, Herts, England, and were incubated at 38°C. Embryos were staged according to Hamburger and Hamilton (1951). Following surgical manipulation, embryos were dissected into PBS and processed for whole-mount in situ hybridisation. FGF-2, retinoic acid and BMP-2 application on beads FGF-2 was obtained from R and D Systems and BMP-2 was provided by the Genetics Institute, Cambridge, Massachusetts. All-transretinoic acid was obtained from Sigma (UK). FGF-2 and retinoic acid were applied to heparin acrylic and AG1X2 beads respectively, as described by Niswander et al. (1993) and Tickle et al. (1985). BMP-2 was applied to heparin beads as described by Francis et al. (1994). Experimental manipulation of chick wings All manipulations were performed at stage 20 unless otherwise stated. K. Patel and others R. N. was in receipt of a Wellcome Prize Studentship. C. T. and D. D'S. are supported by the Wellcome Trust. We would like to thank Anne Crawley for help with the histology, and Professor Lewis Wolpert and Professor Paul Brickell for helpful comments on the manuscript.

Developmental Dynamics, 2004

The dorsal and ventral scales of the chick foot can be distinguished morphologically and molecula... more The dorsal and ventral scales of the chick foot can be distinguished morphologically and molecularly: the dorsal oblong overlapping scuta expressing both ␣ and ␤ keratins, and the ventral roundish nonprotruding reticula expressing only ␣ keratins. The question arises how En-1 and Lmx1, whose role in dorsoventral limb patterning has been well established, can affect skin morphogenesis, which occurs 8 to 12 days later. Forced expression of En-1 or of Lmx1 in the hindlimb have, respectively, as expected, a ventralizing or a dorsalizing effect on skin, leading to the formation of either reticula-type or scuta-type scales on both faces. In both cases, however, the scales are abnormal and even glabrous skin without any scales at all may form. The normal inductive interactions between dermis and epidermis are disturbed after En-1 or Lmx1 misexpression. Effectively, while Lmx1 endows the dermal precursors of the ventral region with scuta inducing ability, En-1 blocks the competence of the dorsal epidermis to build scuta.

Current Biology, 2003

and and steroids prevented healing of skin wounds [5]. In the latter study, retarded repair was c... more and and steroids prevented healing of skin wounds [5]. In the latter study, retarded repair was coincident with an Developmental Biology University College London accumulation of cell and matrix debris at the wound site. However, in these depletion studies, steroid treat-Gower Street, London WC1E 6BT United Kingdom ment may have knocked down more than just leukocytes at the wound site. Indeed, it is known that steroids are 2 Cancer Research UK, 61 Lincoln's Inn Fields inhibitory to the upregulation of immediate-early genes in response to growth factor inductive signals [6], and London WC2A 3PX United Kingdom these may be crucial activators of the repair process. The PU.1 null mouse provides an ideal opportunity to 3 The Burnham Institute 10901 North Torrey Pines Road more directly test the role of macrophages and neutrophils in the repair process, essentially by genetic abla-La Jolla, California 92037 tion; this ETS family transcription factor is hematopoietically restricted and essential for maturation and functional competence of several haemopoietic lin-Summary eages, and thus the PU.1 null mouse is born with neither macrophages nor functioning neutrophils [7, 8]. Damage to neonatal and adult tissues always incites an influx of inflammatory neutrophils and macro-We made full thickness 2 mm incisional wounds on the dorsal forepaw of antibiotic-maintained, neonatal phages. Besides clearing the wound of invading microbes, these cells are believed to be crucial coordina-PU.1 null mice derived from PU.1 heterozygote crosses (n ϭ 34). The wound site was labeled by brushing carbon tors of the repair process, acting both as professional phagocytes to clear wound debris and as a major into the incision immediately post-wounding. Wild-type mice from these litters reepithelialize such a wound in source of wound growth factor signals. Here we report wound healing studies in the PU.1 null mouse, which 3 days, sloughing the overlying scab on the fourth day. Staining for a neutrophil-specific enzyme, chloroacetate is genetically incapable of raising the standard inflammatory response because it lacks macrophages and esterase (CAE), reveals a peak of neutrophils at wound sites about 1 day post-wounding (Figure 1A), while in functioning neutrophils. Contrary to dogma, we show that these "macrophageless" mice are able to repair situ expression data for c-fms or immunostaining with the monocyte/macrophage-specific antibody, F4/80[9], skin wounds with similar time course to wild-type siblings, and that repair appears scar-free as in the em-shows an influx of macrophages from 12 hr after wounding, peaking at between 2 and 4 days (Figures bryo, which also heals wounds without raising an inflammatory response. The growth factor and cytokine 1C and 1CЈ), and remaining at the wound site for up to 5 or 6 days post-wounding. profile at the wound site is changed, cell death is reduced, and dying cells are instead engulfed by stand-As expected, PU.1 null littermates show an absence of CAE or F4/80 staining or c-fms expression at equivalent in phagocytic fibroblasts. We also show that hyperinnervation of the wound site, previously believed to be wound sites (Figures 1B, 1D, and 1DЈ). However, contrary to dogma, these mice, while unable to raise a cell a consequence of inflammation, is present in the PU.1 null wound, too. inflammatory response, show no retardation in the repair process, with all wounds examined at 4 days (n ϭ 27) being fully closed and the scab lost, just as for equivalent Results and Discussion wild-type wounds. These data strongly suggest that the inflammatory response is not absolutely essential for The repair of a skin wound is a complex process requirtissue repair. In fact, the healed epidermis at 3 days ing the collaborative efforts of many cell lineages to appeared more mature than in equivalent wild-type shore up the gap and to replace the missing tissues. It wounds (Figures 1E and 1F). Remarkably, at this stage, has long been considered that the inflammatory reresin histology through the PU.1 null wound connective sponse is instrumental in supplying growth factor and tissue showed no obvious difference from the conneccytokine signals that orchestrate the cell and tissue tive tissue in adjacent unwounded dermis (n ϭ 4; Figmovements necessary for repair [1]. Neutrophils are ure 1F). known to express several proinflammatory cytokines at To more fully challenge the wound repair capacity of the wound site [2], and macrophages from wound tisthese mice, we made full-thickness excisional wounds sues also express many of the growth factors that might of 2 mm diameter to the forepaws of a further 18 PU.1 regulate the repair process [3]. Moreover, a classic senull mice. In wild-type wounds of this sort, repair is ries of experiments in the 1970's showed that while achieved by a combination of reepithelialization and granulation tissue contraction [1]. Reepithelialization is

Anatomy and Embryology, 1999

Tendons connect muscle to skeletal elements. Although tendons have been shown to originate from t... more Tendons connect muscle to skeletal elements. Although tendons have been shown to originate from the lateral plate mesoderm, very little is known at the molecular level about how they are formed. We have found that two genes, Follistatin and Eph-A4, are expressed in regions associated with tendon formation in developing chick limbs. Follistatin is expressed near the tip of the digits and subsequently around the tendon, whereas Eph A4 transcripts were localized in a slightly more proximal region and later in the body of the tendon. Previous work has demonstrated that application of TGFβ1 or TGFβ2 to inter-digital regions or the removal of ectoderm in the foot plate induces ectopic cartilage formation, while removal of ectoderm or application of FGF to tips of developing digits leads to truncation. Here we show that TGFβ1 or removal of ectoderm is also able to induce the expression of both Eph-A4 and Follistatin and that manipulations that cause truncations affect these genes. Thus cartilage and tendon development appear to be coordinated. Ectopic application of recombinant human Follistatin, an antgaonist of certain TGFβ super-family proteins including Activin and Bmp-4, results in the loss of tendon, implicating signalling by TGFβ super-family in the development of tendon during chick embryogenesis. Signalling by TGFβ family members, antagonised by Noggin is known to regulate skeletal development. Thus we suggest that parallel pathways govern both skeletal and tendon patterning.

Journal of Anatomy, 1998

 A human polydactylous left foot with 9 toes, amputated from an 11-mo-old child, was exam... more  A human polydactylous left foot with 9 toes, amputated from an 11-mo-old child, was examined by -ray and magnetic resonance imaging and by gross dissection to identify the digits. The normal sequence of toes from medial to lateral is 1, 2, 3, 4, 5. Examination of the morphology of tendons and muscles suggested the toe sequence was 1, 2, 3\4, ?,5, 2, 3\4, 3\4, 5. The 2 toes in the sequence that are underlined were displaced dorsally and were found to have 2 extensor tendons, no flexor tendons and nails that were conical and situated at their tips. These toes resembled those described as ' double-dorsal ' and which develop in paws of mice in which a gene normally expressed ventrally is functionally inactivated (Loomis et al. 1996). Specification of toe formation occurs in leg buds early in embryonic development and later there is rotation of the limb so that the anterior (rostral) part comes to lie medially, i.e. the hallux which was anterior (rostral) now is on the inner (medial) side of the foot. A disruption in the patterning of this foot in both anteroposterior (rostral-caudal) and dorsoventral axes during development could be responsible.