sign in sign up
<<
Home , Ectoderm, Gastrulation, Inner cell mass, Programmed cell death

Histochem Biol (2006) 126:149–158 DOI 10.1007/s00418-006-0214-1

REVIEW

Cell in development: shaping the

Carlos Penaloza · Lin Lin · Richard A. Lockshin · Zahra Zakeri

Accepted: 8 June 2006 / Published online: 1 July 2006 © Springer-Verlag 2006

Abstract in is normally classiWed ter, as in other , cell death plays a major role in as type I (apoptotic), type II (autophagic) or necrotic. shaping and sculpting the embryo. In those situations Of the biologically controlled types of death, in most that have been carefully studied, cell death is under embryos is the most common, although in tight genetic control (including regulation of and in cells with massive products whose function in cell death is not yet known, type II is often seen, and intermediate forms are seen. such as cdk5), with activation of apoptosis sometimes For embryos other than , apopto- regulated by local environmental variables. sis is not seen prior to but thereafter is used to sculpt the organs of the embryo, while overpro- Keywords Cell death · Apoptosis · Development · duction of cells with subsequent death of excess cells is cdk5 · a common means of generating high speciWcity with low information cost. In zebraWsh at least, the inability It is now well understood that cell death is a funda- of embryos prior to the maternal-zygotic transition to mental aspect of development of almost all - undergo apoptosis appears to derive from the inability isms. Historically cell death was mostly noticed in of the cells to resist lysis once apoptosis begins, rather developing systems and for many years considered to than any inhibition of apoptosis. In mammalian be important only in development. In the last 15 years embryos, apoptosis is seen during cavitation. Thereaf- this image has changed, and the importance of cell death has been realized in many other systems and situations that govern our . However, the study of cell death in development still has much to teach us about the embryo and about the mechanisms of cell Histochemistry and Cell Biology Lecture presented at the 48th death. is dynamic and well Symposium of the Society for Histochemistry in Stresa, Lake orchestrated. Cells in very close proximity proliferate, Maggiore, Italy, 7–10 September 2006 diVerentiate, and die, and their decision making X C. Penaloza · L. Lin · Z. Zakeri (&) appears to be in uenced by their microenvironment. Department of Biology, Queens College and Graduate How one cell dies whereas the neighboring cells Center of CUNY, Flushing, NY 11367, USA divide is presumptively regulated by the internal and e-mail: [email protected] external signals that they receive. To understand how R. A. Lockshin these signals are presented and recognized during Department of Biological Sciences, St. John’s University, development we as well as others have used many Jamaica, NY 11439, USA methodologies and systems. Several laboratories had to develop speciWc markers and in doing so learned Present Address: L. Lin much about the steps taken by a dying cell and how Preclinical Department, Medarex Inc., Bloomsbury, developmental cell death resembles and is diVerent NJ 08804, USA from other types of cell death. 123 150 Histochem Cell Biol (2006) 126:149–158

Cell death, its types and its characterization (MZT). However, when zebraWsh embryos are exposed to cycloheximide, both 256-cell embryos Cell death has been classiWed into three classical types: (prior to the MZT) and 1,024-cell embryos (after the apoptosis, (also referred as lysosomal cell MZT) activate -3 after approximately 5 h. Pre- death), and (for review, see Zakeri and Lock- MZT embryos lyse almost immediately thereafter and shin 2002), and all three are seen during development. thus show no morphological signs of apoptosis, Both apoptosis and autophagy are considered to be whereas the older embryos persist for two more hours under biological control and thus programmed (pro- and develop all the signs of classical apoptosis, includ- grammed cell death, PCD). PCD was originally deWned ing phosphatidylserine exposure, DNA fragmentation, by the demonstration of a clear physiological and bio- and margination of the (Negron and Lock- chemical sequence leading to death in a developmental shin 2004; J.F. Negrón and R.A. Lockshin, in press). situation (metamorphosis), but now the term is under- stood to refer to any biologically controlled and man- Lysosomal, autophagic PCD (Type II) aged cell death. The diVerent types of cell death have speciWc features that are used both to identify them and The appearance of large autophagic , lysosomal to characterize them. Although many features are spe- derivatives that consume the bulk of the cytoplasm, is ciWc the boundaries among the diVerent types are not characteristic of this type of PCD. This type of cell death nearly as well deWned as many assume, and there are unlike apoptosis is not characterized by the prompt overlapping features among the diVerent types. It is also destruction of DNA. In most of the cells undergoing noteworthy that there is much crosstalk among the Type II PCD, the cytoplasm is bulky and cannot be easily diVerent ; this crosstalk allows autophagy to shed to be phagocytosed. The function of the is convert to apoptosis or vice versa, and it allows an apop- conducted within the cell. When approximately 80% of tosis typically triggered by ligation of molecules the cytoplasm has been destroyed, the condensation of to follow the pathway of metabolically activated apopto- the cytoplasm and coalescence and migration of the chro- sis. Thus if one pathway is blocked, a cell may still die by matin become apparent, and only then it is possible to a second, highly organized and biological pathway. This detect by electrophoresis a DNA ladder, indicating DNA result is quite common and has led to considerable con- by an apoptosis-type DNase. It is not certain fusion among researchers not attentive to the possibility. that the autophagy is truly a type of cell death, as opposed to the response of a cell under and attempting to Apoptotic PCD (Type I) survive by consuming its resources. Thus, as in other instances in which cells undergo autophagy, there is Apoptosis or type I cell death is characterized by a potential recovery, and the autophagy is a recycling cell’s loss of adherence to neighboring cells as well as mechanism, used under stressful conditions. In the docu- to extracellular matrix. Morphological changes include mented of cells by autophagy, as in meta- cell rounding and condensation of nuclear material. morphosis, organelles are eliminated in waves. Glycogen, The chromatin coalesces into one or a few masses ribosomes, and mitochondria are eliminated in sequence. along the nuclear membrane. PCD type I is also char- In other instances of autophagic death, such as death of acterized by fragmentation of DNA to form a ladder prostatic following castration and possibly when electrophoresed. There is release of cytochrome post-lactational involution of mammary epithelium, the c and other materials from mitochondria and activation actual commitment to die occurs relatively late, and cells of caspases, speciWcally caspase 3. Cells ultimately frag- can revert and recover if given the appropriate hormonal ment into blebs that are taken up by . At support. PC12 cells undergoing starvation from the Wnal stages of apoptosis, phosphatidylserine (PS) is (NGF) withdrawal can recover if the NGF actively extruded from the internal face of the cell is provided. They are capable of recovering until almost membrane of the dying cell; the exteriorized PS serves immediately before they lyse; this point of no return is as one of the markers that identify the cell as a target coincident with the Wnal autophagic destruction of mito- for phagocytosis. This type of cell death can be greatly chondria, presumptively terminating the cells’ ability to aVected by levels of ATP: if levels of energy are not maintain ionic pumps (Xue et al. 1999, 2001). suYcient, cells can undergo partial apoptosis, leading Both type I and type II cell death have been referred to diVerent appearance and possibly misinterpreta- to as . Like proliferation and tions. For instance, it is commonly noted that the cells diVerentiation, cell death, especially PCD, also plays a of early non-mammalian embryos cannot undergo prominent role in normal development. It is essential apoptosis prior to the maternal-zygotic transition for the removal of unwanted cells and is critical both 123 Histochem Cell Biol (2006) 126:149–158 151 for restricting cell numbers and for patterning in early embryos is a solution to the problem of how during development (Chanoine and Hardy 2003; Cou- much genetic information would be needed to con- couvanis et al. 1995; Glücksmann 1951). struct an . To take an easily understood image, the quantity of information to connect 1,000 neurons to Necrosis 1,000 speciWc targets will vary according to several assumptions, but even in the simplest version, the Necrosis characterized as uncontrolled death, occurs as information required would be immense. It is surpris- a result of high levels of stress, such as injury. Rapid and ing but mathematically demonstrable that much less large changes in critical ions, osmolarity, and pH lead to genetic information (DNA sequences) is needed to the lysis of the cell. In the absence of adequate mito- produce an excess number of presynaptic neurons that chondrial function and eVective ion pumps, the high can generally and with little requirement for speciWcity concentration of impermeable as the primary Wnd their way to the target area, and to allow only intracellular anion tends to draw water in, and under those that make successful synapses to survive. Simi- anaerobic conditions lactic acid will also accumulate, larly, to randomly reconWgure DNA to generate a huge pulling in more water. Ultimately the cell membrane variety of cells, each making a diVerent immunoglobu- becomes leaky or bursts, allowing the escape of cell con- lin, and to kill oV those that are inappropriate, requires tents. In most situations, the uncontrolled release of cell only approximately 60 , as opposed to the 100,000 contents activates an inXammatory response, including or so that would be necessary to code individually for the invasion of mast cells, phagocytes, and NK cells. all the antibodies a is capable of producing. Massive necrosis can for many reasons be extremely While these two examples are spectacular, a similar threatening to an organism and, in higher , logic can be deduced for much that is common in the leakage of intracellular materials can lead to autoim- embryonic development, including regionalization, pat- mune reactions. Thus, the provocation of inXammation tern formation, limb and appendage conWguration, is an important, although derivative, distinction between development of axes, and response to location along an necrosis and apoptosis. In fact, there is evidence that the axis. Thus, it is not surprising to Wnd that the cell death ingestion of apoptotic bodies suppresses an inXamma- as a mechanism is highly conserved among a wide vari- tory response (Edinger and Thompson 2004; Haslett ety of embryos and metamorphosing organisms et al. 1994; Lockshin et al. 2001; Schmied et al. 1993). In (Glücksmann 1951). In C. elegans, death of 131 of the contrast to the obvious osmotic mechanisms that lead to original 1,090 cells leaves an adult with 959 somatic the swelling and rupture of necrotic cells, the shrinkage cells. This system has been instrumental in elucidating of apoptotic cells was a mystery for a long time (Lock- genes involved in diVerent pathways of cell death (Hor- shin and Beaulaton 1981). It now appears that the apop- vitz et al. 1994). In , cell death plays an important totic cells shrink by losing intracellular K+ during a role in a variety of aspects from fertilization through period in which suYcient energy is still available to run development as well as in response to environmental pumps (Bortner and Cidlowski 2002). Thus, a key dis- insults. For example, PCD is involved in xylogenesis, tinction between apoptosis and necrosis is the availabil- aerenchyma formation, petal , endosperm ity of intracellular energy. It is not surprising to development, hypersensitive response, and various encounter situations in which cells begin to undergo forms of abiotic stress (for review, see Hoeberichts and apoptosis but ultimately lyse in a necrotic fashion. One Woltering 2003). Although cell death in plants some- such situation is seen when hepatotoxins are given to times resembles apoptosis in many facets, the cells intact animals; at the periphery of the acutely toxic are enclosed in rigid cell walls and cannot round up or zone, some cells initiate apoptosis but ultimately fail and fragment. Thus, the unique cell death found in plants undergo necrosis (Ledda-Columbano et al. 1991). Of depends on vacuolar lytic function, and dying cells dis- course, when cell death is activated in cultured cells, play many functional vacuoles (Fukuda 2000; Kuriyama necrosis is the ultimate fate of the apoptotic fragments, and Fukada 2002). Furthermore, plants process oxygen since no phagocytes are present. in both mitochondria and , and therefore reactive oxygen species (ROS) are important signals in the activation of plant PCD. For instance, in soybean Cell death during development is common to all cells, PCD can be triggered by ROS (Solomon et al. examined embryos 1999). Increasing evidence indicates that the plant PCD proceeds through a cell death mechanism that is func- It was a belated though intellectually necessary realiza- tionally conserved between animals and plants (for tion that the apparently wasteful process of cell death review, see Hoeberichts and Woltering 2003). 123 152 Histochem Cell Biol (2006) 126:149–158

Cell death is known in such as et al. 2001; Williams et al. 2000). Interference with cell (Abrams et al. 1993; Steller and Grether 1994) and sea death by blocking caspase activity produces several dis- urchins (Yüce and Sadler 2001). There is considerable tinct abnormalities, most notably in the and cell death in embryos, occurring in the cen- in skeletal and CNS development (N. Abraham and tral , similar to vertebrate embryos, and R.A. Lockshin, in preparation; Cole and Ross 2001). elsewhere (Abrams et al. 1993). As in Caenorhabditis, These abnormalities are diVerent from those produced these deaths are all controlled by a limited number of in mammals by knocking out apoptosis-promoting genes, although unlike Caenorhabditis, the insect genes genes, such as the initiator or apical caspase-9 appear neither to be nor related to mamma- or apoptosis-activating factor 1 (Apaf-1). However, lian death control sequences. A set of three genes, studies in the zebraWsh have been done by administra- named reaper, grim, and hid, are found within the tion of inhibitors rather than knockout, and it is not 75C1,2 region of the third (Chen et al. certain that the comparisons are valid. 1996; Grether et al. 1995; White et al. 1994). Knocking out these genes blocks all cell deaths in the embryo. Similarly, as judged from appropriate knockouts and Cell death is an early event during mammalian knockins, these genes are required for the death of development ommatidial progenitors in eyeless and other mutants. The deaths are presumed to be apoptotic, though they Mammalian embryos diVer from other vertebrate have not been examined in great morphological detail. embryos in many ways, but for our purposes a central When Drosophila cells are cultured, an apoptosis issue is compaction, cavitation, and formation of the sequence can be easily demonstrated, although the . This deviation from normal vertebrate proteases involved are much less closely related to the development to form a is also characterized caspases of either Caenorhabditis or vertebrates than by an extremely early activation of zygotic genes. Like- is, for instance, the Caenorhabditis caspase ced-3 to wise, cell death starts at a very early stage in mamma- caspases. Furthermore, although there have lian development. While in most vertebrate embryos been various indirect measurements suggesting cas- one sees no cell death prior to gastrulation, in mamma- pase-like enzymes in metamorphosis, these have not lian embryogenesis cell death begins as early as inner been well documented. In fact, when attempting to cell mass diVerentiation in (Hardy et al. document proteases in insect metamorphosis by sub- 1989; for review, see Spanos et al. 2002) (Fig. 1a–c). strate preference and cleavage site or susceptibility to We used the PS to identify cell death in the preimplan- inhibitors, one cannot easily identify a caspase. The tation developing embryo. As described earlier, the bulk of the during metamorphosis is con- phosphatidylserine is a phospholipid normally asym- ducted by other enzymes, such as lysosomal cathepsins, metrically expressed in the inner leaXet of the plasma , or proteasomal proteases (C.O. Facey and membranes in living cells. During apoptotic cell death, R.A. Lockshin, in preparation; Jochová et al. 1997a, b; due to the loss of membrane potential, these molecules Schwartz et al. 1990). Perhaps this is because metamor- are actively translocated to the outer leaXet of the phic cell death is predominantly autophagic (see - plasma membrane. Annexin V is a very speciWc apop- lier). If insect caspases are activated during the Wnal totic marker since it preferentially binds to negatively phases of metamorphic cell death, when the cells charged phospholipids like PS in the presence of Ca2+ Wnally become TUNEL and PS positive and contain and shows minimal binding to phosphatidylcholine and DNA fragmented at nucleosomes, the enzymes are not sphingomyeline. One can detect apoptotic cells as Xuo- well documented. rescent cells against a dark background. Staining is In lower vertebrates there is substantial cell death visualized by Xuorescence or confocal microscopy or during early embryogenesis. Some of this death has by Xuorescence activated cell sorting. been characterized in zebraWsh and established as As early as the two-cell stage we can Wnd cell death important for such developmental processes as forming in the polar bodies. We did not Wnd dying cells in the olfactory pits, separating the lens from the cornea, embryos containing 1 to 8 cells, but a few cells die opening the , shaping the Wns, establishing the between the 16-cell and blastula stages. We also smoothness of the , and sculpting and modeling the showed that generalized inhibitors of caspases aVect and . A cell death-like normal development even at stages where there was no process eliminates the organelles from lens cells. Large cell death (Fig. 1). Inhibition of caspase activity led to the Rohon-Beard neurons form in each segment only to arrest of embryonic development. This cell death in the die at a slightly later stage (Coen et al., 2001; Svoboda developing pre-implantation embryo was not caspase 123 Histochem Cell Biol (2006) 126:149–158 153

Fig. 1 Cell death in developing mouse embryos. a Cultured seen at this time. When there are two dead cells, they are at oppo- mouse embryos at the midblastula and expanded blastula stages site poles at the same latitude. Nile blue sulfate staining of dying observed with light microscopy. b, c Detection of phosphatidyl- cells (arrows) in ED9.5 (d), ED10.5 (e), and ED11.5 (f) mouse by Annexin V on dying cells during normal development. embryos. g Nile blue sulfate staining of dead cells (actually, vacu- The polar bodies exteriorize phosphatidylserine by the two-cell oles of phagocytes) in the interdigital regions of a 13.5 day embry- stage. At compaction and in early blastula, dead cells are present onic mouse hand palette. Some of the most prominent regions are and at expanded blastula, cell death is seen where the inner cell indicated by arrows mass meets the trophectoderm. Either one or two dead cells are dependent. Caspase activation normally induces cell Ding 2002). It has been shown that the development death, but under the conditions used, caspase inhibi- of the rat lens vesicle involves apoptotic elimination of tion led to further cell death and embryonic develop- the cells between the surface and the optic ment inhibition as a result (Zakeri et al. 2005). vesicle to help trigger and facilitate sepa- ration from the ectoderm. Apoptosis also aids in the bowing of the during lens invagination A major part of involves cell death (Mohamed and Amemiya 2003). Apoptotic cell deaths shape the future structure in day 5 chick A crucial role for cell death is seen during organogen- embryo (Avallone et al. 2002). In cardiac morphogen- esis and tissue remodeling. Cell death was originally esis, cell death is essential in generating the overall noticed in metamorphosing animals. Cell death four-chambered architecture of the (Abdelwa- destroys almost all larval tissues during metamorpho- hid et al. 2002). The role of cell death in muscle devel- sis of insects and (Fox 1973). In Drosoph- opment has been well studied. In moths, ila, cell death is induced in larval midgut by the steroid death is required for correct muscle patterning (Fahr- hormone ecdysone (Lee et al. 2002). In tadpoles, bach et al. 1994; Haas et al. 1995). In rat skeletal mus- regression of the tail and destruction of the gills are cle, cell death persists during the Wrst three post-natal attributed to tightly controlled cell death (JeVery weeks, suggesting an indispensable role for cell death 2002). Precisely regulated cell death contributes to the in the development of (de Torres et al. formation of diVerent body parts and multiple organs 2002). In cell death destroys all the primary of an organism. In developing rat lens, apoptosis myotomal myoWbers, which are replaced progressively removes the nuclei in the process termed nuclear by secondary “adult” multi-nucleated myoWbers in the death, and leaves the rest of the cells remaining func- construction of muscles during development (Chano- tional (Gao et al. 1997; Gupta et al. 2002; Wu and ine and Hardy 2003). 123 154 Histochem Cell Biol (2006) 126:149–158

Cell death sculpts the limbs of all , such as cell death in the dying interdigital regions of the limbs. , mouse, and birds by removal of interdigital We Wnd that most cell death is apoptotic in with webs (Ballard and Holt 1968; Zakeri and Ahuja 1994; the classical margination of nuclei and rounding of the Hurle and Colombatti 1996). Cell death in limb devel- cells, which are often seen as apoptotic bodies engulfed opment begins at embryonic gestation day (ED) 10.5 by macrophages (Fig. 2a, d). Although this type of cell and peaks at ED14.5, at which stage there is enough cell death is the major type found we occasionally also Wnd death to permit cellular and molecular analysis necrotic cell death (Fig. 2b) and also cells resembling (Fig. 1d–g). Cell death can be observed in the anterior autophagic cell death (Fig. 2c). and posterior marginal zones of the developing limb One of the characteristics of apoptotic cell death is bud and in later stages in almost all of the interdigital engulfment by macrophages. We have taken advantage mesenchymal tissue. Cell viability in these regions is of this fact and asked if antibodies speciWc for mature dependent on the condensation of core mesenchymal macrophages can recognize the phagocytic cells in cells from the foot or hand palette. In the absence of these regions. Although mature macrophages do not this condensation, as is the case in interdigital area, cells circulate until later during development we can detect undergo apoptosis. This cell death accompanies the for- phagocytic cells in the area of the limb cell death as mation of free and independent digits of birds and early as our Wrst detection of the dying cell, i.e. ED 10.5 mammals. Cell death can be recognized in the embryo and more intensely at ED 13.5 (Fig. 3a, b). The pres- by several techniques. One more eVective way is in vivo ence of these cells is most likely indicative of diVerenti- detection of cell death using vital dye staining. One can ation of the mesenchymal cells to phagocytic cells take advantage of the fact that a characteristic of cell expressing macrophage-like cell surface . In death is the increase in permeability of cell membranes, addition these phagocytic cells are very active in lyso- allowing selective permeability to speciWc vital dyes, somal activity as measured by activation of acid phos- such as Nile blue sulfate, which enters the cell and local- phatase (Zakeri et al. 1994). The activation of acid izes into the acidic compartments of the cells such as the phosphatase can also be found in some of the individ- . We and others have used this type of stain- ual dying cells (Fig. 3c), leading us to infer that maybe ing for identiWcation of cell death in the developing type II cell death is taking place in the cells that are not limbs of mouse embryos (Fig. 1d–g). One of the best found in the macrophages. ways to identify and even more speciWcally examine the A hallmark of apoptotic cell death is the fragmenta- type of cell death is to use electron microscopy. Using tion of DNA to distinct ladders. Although this can be this method we found that there are diVerent types of easily seen using gel electrophoresis of cultured

Fig. 2 Cell death examined by electron microscopy. a Thin section of hand palette showing apoptotic bodies within macrophages. b–d Electron micrographs show- ing diVerent forms of cell death (b) necrotic cell death (left) with disorganized nucle- us and cytoplasm and (right) phagocytosed apoptotic body (c) autophagic cell death with large autophagic vacuoles (d) phagocytosed apoptotic body with condensed nucleus and cytoplasm that is niether ex- tracted nor precipitated

123 Histochem Cell Biol (2006) 126:149–158 155

Fig. 3 Detection of phagocy- tic cells in developmental cell death. a Low magniWcation of the interdigital area of limb stained for phagocytic cells (dark brown) by antibody F4/ 80. b High magniWcation of (a), arrows show the phagocy- tic cells with apoptotic bodies. c Stained section of interdigi- tal region of the limb for the activity of lysosomes. Activity of acid phosphatase in lyso- somes indicated by red precip- itate (arrows). d Detection of DNA fragmentation using TUNEL/DAB-peroxidase reaction. Dead cells (arrows) are revealed in the interdigital area of mouse limb

cells, Wnding it in a situation where only a few cells are determines the patterning of structures in both verte- dying at a time is more challenging. However, one can brates and invertebrates. Precise patterns of the cell use the end labeling TUNEL to identify the dying cells. death are critical for the development of a number of This technique is more applicable to the developing diVerent organs including the nervous system and the embryo, since one can see the fragmented DNA at the . During vertebrate brain development, cellular level without disrupting the cellular arrange- 20 to 80% of the neurons originally produced are elim- ment, so one can measure the levels of cell death very inated by cell death (Gordon 1995), and more than sensitively and with spatial speciWcity. TUNEL is an 80% of ganglion cells in the cat die shortly after acronym for terminal deoxyuridine nucleotide end birth (Barres and RaV 1999). Much of the cell death is labeling, and its method of action is by labeling any free due to limited quantities of trophic factors, such as 3Ј end of DNA. This method detects fragmented DNA NGF; the only neurons that survive are those supplied from necrotic, mitotic as well as apoptotic cells, but the with NGF (Vyas et al. 2002). In the human immune level of DNA fragmentation is so much greater in apop- system, about 95% of developing B cells will die due to totic cells that this technique is among the most reliable faulty gene rearrangement, anti-self receptor expres- assays for apoptotic cell detection. The fragmented sion, or lack of stimulation, and immunological toler- DNA within the dying cell is usually visualized as a dark ance is achieved through highly controlled cell death red to brown staining in the cell. The downside to using (Osborne 1998; Gercel-Taylor et al. 2002). Cell death is this technique is that only late-stage apoptotic cells can also a pivotal part to the development of the germ be reliably scored. Commercial kits are readily avail- cells; in human females, 80% of the original germ cells able for this assay. By this measure we can Wnd many (oocytes) die by apoptosis by the time of birth (for TUNEL positive cells (Fig. 3d) in correlation with the review, see Reynaud and Driancourt 2000). regions of high positive signal for macrophage staining Cell death in the embryo has been studied geneti- and acid phosphatase activity (Fig. 3c). By all these cally for the most part, with some pioneering observa- measurements we can monitor the occurrence and type tions with found that lead to limb of cell death during the development of the limb. malformations, such as Dominant hemimelia (Dh), Luxoid (Lu), Luxate (Lx), Oligosyndactyly (O), Extra toes (Xt), and Hammertoe (Hm). The Hm mutants Cell death sets up the correct patterning of the embryo were used by us to evaluate the relationship between and is under genetic regulation the altered pattern of limb formation with cell death. The Hm mice have webbing between digits 2, 3, 4, and During development, cell death plays an important 5. We Wrst identiWed the pattern of cell death in normal, role in the elimination of redundant cells and thus heterozygote and homozygote embryos during diVerent 123 156 Histochem Cell Biol (2006) 126:149–158 times of gestation from day 11.5 to 15.5. The homozy- number of mutant animals with altered levels of cell gous normal embryo has the correct pattern of cell death gives credence for the regulation of cell death by death, which is partially altered in heterozygotes and speciWc gene activity. Some of the most important stud- completely altered in the homozygous mutant. In the ies crucial for our understanding the mechanism of homozygous mutant hand at day 14.5 when we Wnd the apoptosis came from studies of C. elegans in measuring peak of cell death in the interdigital webs we Wnd cell embryonic or developmental cell death, which led to death only between digit one and two and no cell death the well known bcl-2 like molecules, normally inhibit- between the other digits (Fig. 4a). This lack of cell ing cell death, activation of caspase-like molecules death reXects the seen in the adult where destroying the cells, and the existence of “eat me” sig- there is webbing between all the digits except between nals on the surface of apoptotic cells. The majority of digit one and two (Fig. 4b). This aberrant pattern of cell the information known about cell death in the embryo death does not have any eVect on the correct pattern of came from manipulation of genes and characterization skeleton formation (Fig. 4c) and thus is reXective of soft of the resulting . Further genetic analysis of tissue syndactyly. More interestingly we can correct this cell death will very likely provide much more informa- genetic abnormality by exposing the embryo in utero to tion, which may lead to further understanding of cell all cis-retinoic acid (RA) at a speciWc window of devel- death in development. opment and cause the rescue of the malformation. In One of the sets of genes that may be instrumental this case the homozygous mutant embryos are born in regulating cell death is the cell cycle related genes. with correct limb patterning since we have induced cell One particularly interesting group is that of cyclin death at the correct time by RA. These Wndings indicate dependent . We found that a unique member a direct correlation between pattern formation in the of this family i.e. cyclin dependent 5 (cdk5), limb and correct level and pattern of cell death. identiWed by its to cdc2 and recognized by From genetic studies using C. elegans as well as Dro- its activation during neuronal proliferation, also has sophila and mice it is obvious that developmental cell an activity during cell death (for review, see Dhavan death is under genetic regulation. The presence of a and Tsai, 2001). When antibodies speciWc to Cdk5

Fig. 4 Expression of cell death-related genes in development. a curvature of digits in (m/m) hammertoe mice as opposed to the Cell death (arrows) in interdigital areas of normal developing wild type mice (+/+). d Cdk5 expression (dark spots pointed by ar- mouse limb (+/+), and signiWcantly less cell death (arrows) in the rows) by in interdigital area of ED13.5 same region of hammertoe mutant (m/m). b Hammertoe hetero- mouse limb. e Confocal merge image of red Cy-3 labeling (Cdk5 zygous (m/+) and homozygous (m/m) show no separation be- expression) and green FITC labeling (DNA fragmentation). The tween digits 2 and 5 and increased curvature, while there is yellow signal indicates co-expression of Cdk5 and DNA fragmen- complete separation between digits 2 and 5 in wild type mice (+/ tation in ED13.5 mouse developing limb +). c Skeletal staining of hammertoe mutants, showing increase in

123 Histochem Cell Biol (2006) 126:149–158 157 were used we found a diVuse level of expression in References diVerentiating neurons but an intense signal in dying cells of the hand plate in the interdigital regions Abdelwahid E, Pelliniemi LJ, Jokinen E (2002) Cell death and (Fig. 4d). This expression correlated with the cells diVerentiation in the development of the endocardial cush- positive for DNA fragmentation by TUNEL and ion of the embryonic heart. Microsc Res Tech 58(5):395–403 when confocal images were merged, many cells Abrams JM, White K, Fessler LI, Steller H (1993) Programmed cell death during . Development showed both signals (Fig. 4e). The expression of the 117:29–43 cdk5 is not unique to the dying cells of the developing Avallone B, Balsamo G, Trapani S, Marmo F (2002) Apoptosis limb; its expression is found in all dying cells seen in during chick inner ear development: some observations by diVerent parts correlating with diVerent organs of the TEM and TUNEL techniques. Eur J Histochem 46(1):53–59 Ballard KJ, Holt SJ (1968) Cytological and cytochemical studies developing embryo (Zhang et al., 1997). We found on cell death and digestion in the foetal rat foot: the role of that the expression in the dying cells correlates with macrophages and hydrolytic enzymes. J Cell Sci 3:245–262 the post-translational increase in kinase activity and is Barres BA, RaV MC (1999) Axonal control of oligodendrocyte not due to an increase in transcription or translation development. J Cell Biol 147:1123–1128 Bortner CD, Cidlowski JA (2002) Apoptotic volume decrease and of cdk5 (Zhu et al., 2002). Additionally inducing cell the incredible shrinking cell. Cell Death DiVer 9:1307–1310 death in the day 10.5 embryo with cyclophosphamide Chanoine C, Hardy S (2003) Xenopus muscle development: from (CP), which leads to 90% cell death, we isolated using primary to secondary myogenesis. Dev Dyn 226(1):12–23 both yeast two and pulldown assay a cyclin like Chen P, Nordstrom W, Gish B, Abrams JM (1996) Grim, a novel cell death gene in Drosophila. Genes Dev 10(14):1773–1782 interacting protein (p35) that appears to be cleaved by Coen L, du Pasquier D, Le Mevel S, Brown S, Tata J, Mazabraud a calpain during cell death to p25 and to interact and A (2001) Xenopus Bcl-X(L) selectively protects Rohon- activate CDK5 only during cell death (Zhu et al., Beard neurons from metamorphic degeneration. Proc Natl 2002). Interestingly the activation of CDK5 does not Acad Sci USA 98:7869–7874 Cole LK, Ross LS (2001) Apoptosis in the developing zebraWsh require caspase, functional mitochondria-related bin, embryo. Dev Biol 240:123–142 or (Lin et al., 2006; Ye et al., in preparation). We Coucouvanis ED, Martin GR, Nadeau JH (1995) Genetic ap- can conclude that the activity of CDK5 during diVer- proaches for studying programmed cell death during develop- entiation is regulated by a larger interacting activating ment of the laboratory mouse. Methods Cell Biol 46:387–440 Dhavan R, Tsai LH (2001) A decade of CDK5. Nat Rev Mol Cell protein p35; and during activation of cell death acti- Biol 2(10):749–759 vated cleave the p35 to p25, which interacts Edinger AL, Thompson CB (2004) Death by design: apoptosis, and activates CDK5 which then acts to propagate cell necrosis and autophagy. Curr Opin Cell Biol 16(6):663–669 death. From our studies using dominant negative and Fahrbach SE, Choi MK, Truman JW (1994) Inhibitory eVects of actinomycin D and cycloheximide on neuronal death in adult siRNA and anti-sense RNA we can conclude that Manduca sexta. J Neurobiol 25(1):59–69 CDK5 activation is not the inducer of cell death, but Fox H (1973) Ultrastructure of tail degeneration in Rana tempo- that its activation is a result of cell death. Since it is raria larvae. Folia Morphol (Prague) 21:109–112 seen in a number of situations in which the cells die by Fukuda H (2000) Programmed cell death of tracheary elements V as a paradigm in plants. Plant Mol Biol 44(3):245–253 di erent ways or triggering events its activation is a Gao CY, Zakeri Z, Zhu Y, He H, Zelenka PS (1997) Expression downstream event regulating the process of cell of Cdk5, p35, and Cdk5 associated kinase activity in the death. developing rat lens. Dev Genet 20:267–275 In summary cell death is an instrumental feature of Gercel-Taylor C, O’Connor SM, Lam GK, Taylor DD (2002) Shed membrane fragment modulation of CD3-zeta during development, without which one cannot have an : link with induction of apoptosis. J Reprod Immu- embryo and therefore continuation of a species. The nol 56(1–2):29–44 incidence of this cell death is uniquely regulated and Glücksmann A (1951) Cell deaths in normal vertebrate ontogeny. controlled by speciWc genetic programs that can be Biol Rev Camb Philos Soc 26:59–86 Gordon N (1995) Apoptosis (programmed cell death) and other somewhat manipulated by environmental factors. reasons for elimination of neurons and . Brain Dev Developmental cell death represents a special type of 17:73–77 cell death with common features with many other situ- Grether ME, Abrams JM, Agapite J, White K, Steller H (1995) ations in which we see cell death and special features The head involution defective gene of Drosophila melanog- aster functions in programmed cell death. Genes Dev only seen in development. 9(14):1694–1708 Gupta A, Tsai LH, Wynshaw-Boris A (2002) is a journey: a Acknowledgements The authors wish to thank Drs. Yong Zhu genetic look at neocortical development. Nat Rev Genet and Daniela Quaglino for their technical assistance. We would 3(5):342–355 also like to thank NYC-LSAMP and MAGNET-AGEP, spon- Haas AL, Baboshina O, Williams B, Schwartz LM (1995) Coor- sored by NSF and the National Institutes of Health (MARC dinated induction of the conjugation pathway Minority Access to Research Careers) GM 070387-01 for their accompanies the developmentally programmed death of in- support. sect skeletal muscle. J Biol Chem 270(16):9407–9412 123 158 Histochem Cell Biol (2006) 126:149–158

Hardy K, Handyside AH, Winston RM (1989) The human blas- Schwartz LM, Myer A, Kosz L, Engelstein M, Maier C (1990) tocyst: cell number, death and allocation during late preim- Activation of polyubiquitin gene expression during develop- plantation development in vitro. Development 107:597–604 mentally programmed cell death. Neuron 5:411–419 Haslett C, Savill JS, Whyte MKB, Stern M, DransWeld I, Meagher Solomon M, Belenghi B, Delledonne M, Menachem E, Levine A LC (1994) Granulocyte apoptosis and the control of inXam- (1999) The involvement of cysteine proteases and protease mation. Philos Trans R Soc Lond B Biol Sci 345:327–333 inhibitor genes in the regulation of programmed cell death in Hoeberichts FA, Woltering EJ (2003) Multiple mediators of plants. Plant Cell 11(3):431–444 plant programmed cell death: interplay of conserved cell Spanos S, Rice S, Karagiannis P, Taylor D, Becker DL, Winston death mechanisms and plant-speciWc regulators. Bioessays RM, Hardy K (2002) Caspase activity and expression of cell 25(1):47–57 death genes during development of human preimplantation Horvitz HR, Shaham S, Hengartner MO (1994) The genetics of embryos. 124(3):353–63 programmed cell death in the Caenorhabditis ele- Steller H, Grether ME (1994) Programmed cell death in Dro- gans. Cold Spring Harb Symp Quant Biol 59:377–386 sophila. Neuron 13:1269–1274 Hurle JM, Colombatti A (1996) Extracellular matrix modiWca- Svoboda KR, Linares AE, Ribera AB (2001) Activity regulates tions in the interdigital spaces of the chick embryo leg bud programmed cell death of zebraWsh Rohon-Beard neurons. during the formation of ectopic digits. Anat Embryol (Berl) Development 128:3511–3520 193(4):355–364 de Torres C, Munell F, Roig M, Reventos J, Macaya A (2002) JeVery WR (2002) Programmed cell death in the ascidian em- Naturally occurring cell death during postnatal development bryo: modulation by FoxA5 and Manx and roles in the evo- of rat skeletal muscle. Muscle Nerve 26(6):777–783 lution of larval development. Mech Dev 118(1–2):111–124 Vyas S, Bechade C, Riveau B, Downward J, Triller A (2002) Jochová J, Quaglino D, Zakeri Z, Woo K, Sikorska M, Weaver V, Involvement of survival motor neuron (SMN) protein in cell Lockshin RA (1997a) Protein synthesis, DNA degradation, death. Hum Mol Genet 11(22):2751–2764 and morphological changes during programmed cell death in White K, Grether ME, Abrams JM, Young L, Farrell K, Steller H labial glands of Manduca sexta. Dev Genet 21:249–257 (1994) Genetic control of programmed cell death in Dro- Jochová J, Zakeri Z, Lockshin RA (1997b) Rearrangement of the sophila. Science 264:677–683 tubulin and during programmed cell Williams JA, Barrios A, Gatchalian C, Rubin L, Wilson SW, death in Drosophila salivary glands. Cell Death DiVer 4:140– Holder N (2000) Programmed cell death in zebraWsh rohon 149 beard neurons is inXuenced by TrkC1/NT-3 signaling. Dev Kuriyama H, Fukada H (2002) Developmental programmed cell Biol 226:220–230 death in plants. Curr Opin Plant Biol 5(6):568–573 Wu GS, Ding Z (2002) Caspase 9 is required for p53-dependent Ledda-Columbano GM, Coni P, Curto M, Giacomini L, Faa G, apoptosis and chemosensitivity in a human ovarian Oliverio S, Piacentini M, Columbano A (1991) Induction of cell line. 21(1):1–8 two diVerent modes of cell death, apoptosis and necrosis, in Xue L, Fletcher GC, Tolkovsky AM (1999) Autophagy is acti- rat after a single dose of thioacetamide. Am J Pathol vated by apoptotic signalling in sympathetic neurons: an 139:1099–1109 alternative mechanism of death execution. Mol Cell Neuro- Lee CY, Cooksey BA, Baehrecke EH (2002) Steroid regulation sci 14:180–198 of midgut cell death during Drosophila development. Dev Xue L, Fletcher GC, Tolkovsky AM (2001) Mitochondria are Biol 250(1):101–111 selectively eliminated from eukaryotic cells after blockade of Lin L, Ye YX, Zakeri Z (2006) p53, Apaf-1, caspase-3, and -9 are caspases during apoptosis. Curr Biol 11:361–365 dispensable for Cdk5 activation during cell death. Cell Death Yüce Ó, Sadler KC (2001) Post-meiotic unfertilized starWsh eggs DiVer 13(1):141–150 die by apoptosis. Dev Biol 237:29–44 Lockshin RA, Beaulaton J (1981) Cell death: questions for histo- Zakeri Z, Ahuja HS (1994) Apoptotic cell death in the limb and chemists concerning the causes of the various cytological its relationship to pattern formation. Biochem Cell Biol changes. Histochem J 13:659–666 72:603–613 Lockshin RA, Facey CO, Zakeri Z (2001) Cell death in the heart. Zakeri Z, Lockshin RA (2002) Cell death during development. J Cardiol Clin 19:1–11 Immunol Methods 265(1–2):3–20 Mohamed YH, Amemiya T (2003) Apoptosis and lens vesicle Zakeri Z, Quaglino D, Ahuja HS (1994) Apoptotic cell death in development. Eur J Ophthalmol 13(1):1–10 mouse limb and its suppression in the hammertoe mutant. Negron JF, Lockshin RA (2004) Activation of apoptosis and cas- Dev Biol 165:294–297 pase-3 in zebraWsh early gastrulae. Dev Dyn 231:161–170 Zakeri Z, Lockshin RA, Criado-Rodriguez LM, Martinez-A C Osborne BA (1998) Apoptotic signaling in lymphocytes. In: (2005) A generalized caspase inhibitor disrupts early mam- Lockshin RA, Zakeri Z, Tilly JL (eds) When cells die: a com- malian development. Int J Dev Biol 49(1):43–47 prehensive evaluation of apoptosis and programmed cell Zhang Q, Ahuja HS, Zakeri Z, Wolgemuth DJ (1997) Cyclin- death. Wiley-Liss, New York, pp 245–266 dependent kinase 5 is associated with apoptotic cell death Reynaud K, Driancourt MA (2000) Oocyte attrition. Mol Cell during developmental and tissue remodeling. Dev Biol Endocrinol 163:101–108 183:222–233 Schmied M, Breitschopf H, Gold R, Zischler H, Rothe G, We- Zhu Y, Lin L, Kim S, Quaglino D, Lockshin RA, Zakeri Z (2002) kerle H, Lassmann H (1993) Apoptosis of T lymphocytes in Cyclin dependent kinase 5 and its interacting proteins in cell experimental autoimmune encephalomyelitis: evidence for death induced in vivo by cyclophosphamide in developing programmed cell death as a mechanism to control inXamma- mouse embryos. Cell Death DiVer 9(4):421–430 tion in the brain. Am J Pathol 143:446–452

123