Changes in the Morphology of Mouse Vertebrae Exhibit Specific Patterns Over Limited Numbers of Vertebral Levels

Okajimas Folia Anat. Jpn., 76(1): 17-32, May, 1999

The Mouse Vertebrae: Changes in the Morphology of Mouse Vertebrae Exhibit Specific Patterns Over Limited Numbers of Vertebral Levels

Harumichi SHINOHARA

Division of Human Sciences, Faculty of Medicine, Toyama Medical and Pharmaceutical University, Sugitani 2630, Toyama 930-0194, Japan

-Received for Publication, December 16, 1998-

Key Words: Vertebral column, Metameric body pattern, Genes, Gross anatomy

Summary: The mouse vertebrae from the cervix to the tip of the tail were characterized and anatomical features that have been lacking were added to the classical description. The vertebrae consist of six long-range and fourteen short- range substructures, with the foveal process being a newly identified substructure. The caudal transverse process, cranial hemal process and hemal ridge are substructures that are clearly defined in the mouse. Each long-range and short-range substructure has several specific morphological features such as length, width, area, shape and angle. These features exibit a crescendo, plateau or decrescendo pattern over a limited number of vertebral segments that ranges from just a few to twenty. The variety of substructural combinations and the constant changes in the morphological features lead to the fact that no single vertebra has the same morphology as any other. An analysis of the patterns of changes in morphology provides some insight into the genetic plan for the metameric body axis.

The vertebral column of the mouse has been morphological features of the coccygeal vertebrae studied for almost half a century (Gruneberg, 1950; are described for the first time in detail. The verte- Searle, 1954; Green, 1962). The cited early studies brae of the mouse were found to be composed were focussed on genetic variations into emphasis of a combination of substructures, such as the on changes of the number of vertebrae and in some body, arch and several additional processes. The osteological characteristics, such as the anterior tu- length, shape and other morphological features of bercles of the cervial vertebrae and the lumber rib. the substructures changed with crescendo, plateau In terms of anatomy, such observations were quite and decrescendo patterns. Some of these patterns limited. In more recent decades, systematic de- can be interpreted in the context of molecular scriptions of the vertebral column of the mouse development. were published (Cook, 1965; Hummel et al. , 1966; Cook, 1983). Furthermore, several very recent at- tempts were made to quantitate the morphological Materials and Methods changes in the vertebrae along the antero-posterior axis and some subtle but important features of the Eight C57BL/6 mice, four males and four fe- vertebral column were characterized (Johnson and males, of 15-25 weeks of age were used for the O'Higgins, 1994; O'Higgins et al., 1997). However, present study. They were euthanized with ether and such quantitation is not yet sufficient to define the skinned. The head, anterior and posterior limbs, morphology of the vertebrae from the cervix to the thoracic and abdominal viscera were removed, and tail. the trunk musculature was removed with the costae The primary purpose of the present study was as far as possible. A straight copper wire of 0.2 mm to characterize the entire vertebral column of the in diameter was inserted into the vertebral foramen mouse and to supplement previous incomplete an- of the first cervical vertebra, passed through the atomical descriptions. Moreover, in this report, the vertebral canal and pulled outfrom the interverte-

Part of this work was supported by a Grant-in-Aid for Scientific Research no. 0167-0010 from the Ministry of Education, Japan.

17 18 H. Shinohara

brat space between the sixth lumbar and first sacral lumbar, sacral and proximal three coccygeal verte- vertebrae. The tail was fastened to a straight cop- brae. The vertebral body, the arch and the seven per wire of 0.8 mm in diameter. These wires pre- specific processes were defined as long-range sub- vented the vertebrae from disorganized and kept structures that were present at more than 30 verte- the vertebral column straight during subsequent brae, that is to say, in more than half of the total treatments. The vertebral column was fixed in 10% vertebrae. There were also short-range substruc- formalin for several days and rinsed in tap water. tures that were present at fewer than 30 vertebral Then soft tissues were carefully and thoroughly re- levels (Table 1). The vertebrae at all levels con- moved under a dissecting microscope. The verte- sisted of combinations of these long-range and bral column was kept in distilled water and stained short-range substructures. For example, the ninth with alizarin red (several drops of a 0.1 % solution thoracic vertebra consisted of a body; an arch, of alizarin red in 100 ml of distilled water). Four which was studded with the seven processes; and a specimens were dehydrated in a graded ethanol se- pair of short-range substructures, namely, the fo- ries, cleared in xylene and examined for continuous veal processes. The morphology of the individual changes in vertebral substructures and in their substructures was, however, unstable and there morphological features (Figs. 1, 2 and 3). The ap- were both gradual and sudden changes. For exam- proximate sizes of intervertebral foramina were ple, visual estimation of the craniocaudal length of measured with a micrometer under a dissecting mi- the coccygeal body (Fig. 1) revealed an increase in croscope. In the case of the remaining four speci- the proximal part of the tail and a gradual decrease mens, the individual vertebrae were separated, in the distal part (see also Fig. 5). This change in cleaned and air-dried. The craniocaudal length of length can be described as a crescendo-decrescendo each vertebra was measured with a micrometer pattern. Meticulous observations of the vertebral under a dissecting microscope. The cranial surface column revealed that the vertebrae were composed of each vertebra was photographed (Fig. 4) and the of long-range and short-range substructures whose area of the vertebral foramen was calculated with a morphological features changed from one segment computerized image analyzer (CIA 102; Olympus, to the next, with combinations of crescendo, Japan). plateau and decrescendo patterns. Thus, no single vertebra was morphologically identical to any other. Results

General features Table 1. The long-rangeand short-rangesubstructures that The vertebral column (Figures 1, 2 and 3) con- formthe vertebraeof the mouse sisted of 59 to 61 vertebrae. There were seven cervical vertebrae (C1—C7),thirteen thoracic vertebrae (T1—T13),six lumbar vertebrae (L1—L6),four sacral vertebrae (S1—S4)and from 29 to 31 coccygeal vertebrae (Col—Co31). The vertebral column had two curvatures; lordosis at the cervico-thoracic border and kyphosis at the thoraco-lumbar border. For convenience, the number of coccygeal vertebrae was taken as 29 because five out of eight mice (more than 60%) had 29 such vertebrae. The actual number of coccygeal vertebrae seemed to depend on the extent of ossification of a few segments at the tip of the tail. Thus, older mice tended to have a larger number of coccygeal vertebrae. However, no analysis of differences between young and old or male and female mice was performed in the present study. Each vertebra consisted of a body and an arch (Fig. 4). Seven processes, namely, a pair of transverse processes, a pair of cranial articular processes, a pair of caudal articular processes and an unpaired spinous process, extended from the body and arch and were found in the cervical, thoracic, Morphology of Mouse Vertebrae 19

Fig. 5. The pattern of changes in craniocaudal length of the vertebral body. A crescendo-decrescendopattern with a clear maximum (arrow) is present for the cervical, lumbar, sacral and coccygeal regions, respectively.The thoracic region exhibits a crescendo- plateau-crescendo pattern. The lengths at Cl and C2 are exceptional due to the absence of the vertebral body at Cl and the presence of the dens axis at C2. N: Number of specimensexamined.

Long-range substructures midsagittal line to form the vertebral foramen. The 1. The vertebral body area of the vertebral foramen (Fig. 6) exhibited a The vertebral body was one of the most consis- decrescendo pattern from Cl to T6, with small tent substructures. It was found in all vertebrae fluctuations at C5, C6 and C7. Then it exhibited a with the exception of C1 (atlas). The morphological crescendo-decrescendo pattern from 17 to Co4, features of the vertebral body included the cranio- with a peak at Ll. There was no vertebral foramen caudal length, the shapes and the depth. The mean at Co4 and Co5 (Figs. 1 and 4) as a result of the craniocaudal lengths of the vertebral bodies are abscence of curved laminae (spina bifida). Spina shown in Figure 5. The length exhibited a crescendo- bifida occurs normally at Co4 or CoS; spina bifida is decrescendo pattern in the cervical, lumbar, sacral anomalous only if it occurs at additional cranial and coccygeal regions, respectively (arrows in levels. Fig. 5). The thoracic region exhibited a cresendo- plateau-crescendo pattern. The shape of the cranial 3. The transverse process surface of the body also changed with vertebral The representative bone extension to the lateral levels: it was oval from C3 to T2, bell-shaped from side was designated the transverse process. In a T3 to T10, trapezoid from T12 to L2 and triangular series of thoracic vertebrae, the transverse process from L6 to S3 (Fig. 4). The depth of the vertebral was replaced by three bony extensions. These ex- body was reported by O'Higgins et al. (1997). tensions, namely, the foveal process, cranial tubercle and caudal tubercle, were independent sub- 2. The vertebral arch structures of the transverse process. The vertebral arch consisted of straight pedicles Shape of the transverse process The transverse and curved laminae. The pedicles extended from process was an exceptionally large square mass at the border between the vertebral body and the Cl in dorsal or ventral view (Figs. 1 and 3). The roots of the cranial articular process. The left and process was club-shaped from C2 to T9, but it split right laminae started from the distal ends of the into two bony elevations, the cranial and caudal pedicles and fused with each other on the dorsal tubercles, at T10 (Fig. 2). Thus, the transverse pro- 20 H. Shinohara

Fig. 6. The pattern of changes in the area of the vertebral foramen. The area exhibits a decrescendo pattern from Cl to 17 with small fluctuations from C5 to C7. A crescendo-decrescendo pattern is apparent from C8 to Co4 with a maximum at IA (arrow). N: Number of specimens examined.

cess was not present from T10 to T13. The trans- small, with an area of approximately 0.4 x 0.8 mm2. verse process reappeared at Ll. The process was The area of the foramen in the thoracic and lumbar flat and sickle-shaped from Li to L6, wing-shaped regions increased gradually toward L6, where it was at Si and S2 and resembled a parallelogram from approximately 1.0 x 1.0mm2. The foramina in the S3 to Co4. At Co5, the transverse process split into sacral region decreased in area to as little as cranial and caudal portions, such that the coccygeal 0.5 x 0.5 mm2. The area of the intervertebral fora- vertebrae distal to Co5 had an "hourglass" shape in men between S4 and Col was approximately twice ventrodorsal view (Figs. 1 and 3). The caudal por- that of the sacral foramina, but the size decreased tion was defined as the caudal transverse process in gradually. An intervertebral foramen was found the present study. between Co2 and Co3 (asterisk in Fig. 2) but not Orientation of the transverse process The orien- between Co3 and Co4 (double asterisks). The cau- tation in ventral view (Fig. 3) varied from the cra- dal articular process of Co3 did not make a joint niolateral to the caudolateral side with changes in with the cranial articular process of Co4. vertebral level. The orientation in cranial view (Fig. The joint surface of the cranial articular process 4) also varied from ventrolateral to dorsolateral. was flat from Cl to T10 but curved from T11 to Co3 (Fig. 4). This change in the articular surface oc- 4. The cranial and caudal articular processes curred consistently between T10 and T11 in mice. The cranial and caudal articular processes were By contrast, in humans, the change occurs any- extensions that formed the intervertebral foramen where from T12 to L2 (Shinohara, 1997). with the vertebral body (see for example, the as- The position of the cranial articular process rel- terisk between T6 and T7 in Fig. 2). The inter- ative to the transverse process changed with the vertebral foramina in the cervical region were vertebral level. It was dorsal to the transverse pro- Morphology of Mouse Vertebrae 21 cess from Cl to T1, ventral from T2 to T9 and again in length and had cranial and caudal angles. A dorsal from Li to Co3 (Fig. 4). small and superficial portion of the longus colli muscle arose from the cranial angle and was in- 5. The spinous process serted on the ventral tubercle at Cl. Also, a small The spinous process extended from the vertebral portion of the longus thoracis muscle was inserted arch and was usually present on 34 vertebrae from on the caudal angle of the ventral laminae. Cl to Co4. There was no vertebral arch at Co5 or, occasionally, at Co4 (Figs. 1 and 4). Thus, the spi- 5. Articulation with costae nous process was not present at this and more distal The thoracic vertebrae articulated with costae levels. The spinous process was very short from Cl via the foveal surfaces on the vertebral body and on to T1 but it was exceptionally long, being approxi- the transverse process. The fovea on the body was mately 2 mm in length, at T2 (the vertebra prom- clearly recognized from Ti to T10 but was indistinct inens). Johnson and O'Higgins (1994) reported the at T11, T12 and T13. The changes in the foveal length of the spinous process in the mouse. The surface on the transverse process are described spinous process in dorsal view was literally spinous, below. with a leading angle indicating the direction of the process (Fig. 1, white arrows at T6 and T13). The 6. The foveal processes (tentative), cranial tubercle spinous process was oriented caudally from T3 to and caudal tubercle T10 but cranially at T11 and at more distal verte- The transverse processes of the thoracic vertebrae (Fig. 2, asterisk between T10 and T11). brae were replaced by three bony projections. First, the foveal surface for articulation with the costa Short-range substructures became elevated from the transverse process (ar- 1. The ventral tubercle rows at T5—T10in Fig. 4). This bony elevation was The ventral tubercle was a single hook that ex- defined as the foveal process in this study. The tended from the ventral and midsagittal part of Cl height of the fovea! process exhibited a crescendo- (arrow at Cl in Figs. 2 and 3 and open arrow at Cl decrescendo pattern from T6 to T10, with a maxi- in Fig. 4). A small portion of the longus colli mus- mum at T8. The second and third bony projections cle, which started from the transverse processes of were the cranial and caudal tubercles that appeared the cervical vertebrae and from the ventral laminae at T10 (open and closed arrows at T10 in Fig. 2). of C6, was inserted in the ventral tubercle. The cranial tubercle appeared at T10 but disappeared at T11, as if it had been absorbed by the 2. The alar foramen cranial articular process. The height of the caudal The alar foramen penetrated the transverse tubercle exhibited a crescendo-decrescendo pattern processes dorsoventrally at Cl (arrows at Cl in Fig. from T10 to L5 with a maximum at L2 (closed 1). The foramen communicated with the trans- arrows at T1O—L5in Fig. 2). It should be noted versarium foramen that penetrated the transverse that the caudal tubercle and transverse process co- processes craniocaudally at Cl (circles and arrows existed from Li to L5 (see Discussion). at Cl in Fig. 4). Branches of the vertebral artery and vein passed through the alar foramen and ex- 7. Articulation with coxae tended very small branches to the suboccipital and The transverse processes of Si were wing- neck regions. shaped in ventrodorsal view (Figs. 1 and 3). The transverse processes of Si and S2 tended to fuse 3. The transversarium foramen with each other and become continuous. Only the The transversarium foramen penetrated the root lateral end of the transverse process of Si articu- of the transverse process craniocaudally and was lated with the ilium. present at six levels from C6 to Cl (asterisks at C5 in Fig. 4). Passing through the foramina, the verte- 8. The cranial and caudal hemal processes and bral artery ascended to the atlanto-occipital level hemal ridge (tentative) and entered the foramen magnum (broken line and The hemal processes were bony supports of the arrow at Cl in Fig. 4). coccygeal artery and vein that ran along the ventral sagittal line of the tail. In the rat, there is a pair of 4. The ventral laminae hemal processes at the caudal end of the coccygeal The ventral laminae were bilateral and ventro- body (Hebel and Stromberg, 1986). In the mouse, caudal extensions from the junction between the there were two pairs of hemal processes, one at the body and transverse process at C6 (arrow at C6 in cranial end and the other at the caudal end of the Figs. 2, 3 and 4). They were approximately 1.6 mm coccygeal body (closed and open arrows at Co5 in 22 H. Shinohara

Fig. 3). The cranial hemal processes were cotyle- at the coccygeal level. The sesamoid bones that don-shaped, with the right and left processes ex- were described as "chevron bones" in the rat tending independently. They appeared first at Co5 (Greene, 1968) were also found in the mouse. The (occasionally at Co4; Fig. 4) and were found at long-range and short-range substructures had sev- more distal coccygeal vertebrae. Tendons of the tail eral characteristic morphological features such as terminated at the processes, and the artery and vein length, width, orientation and shape, and these of the tail ran caudally through the valley between features tended to change from one vertebral level the processes. The caudal hemal processes con- to the next. For example, the craniocaudal length of sisted of a pair of small elevations. They were very the vertebral body increases from the second to the small and the right and left processes were so close tenth coccygeal vertebra (= crescendo pattern) and together that they appeared to form a midsagittal then decreases gradually toward the 29th coccygeal ridge (arrow at Co14 in Fig. 4). This ridge was de- vertebra (= decrescendo pattern). Thus, no single fined as the hemal ridge in the present study. The coccygeal vertebra was mophologically identical to hemal ridge was a characteristic feature of the cau- its neighbors. Similarly, no single vertebra was dal end of Co5 and more distal coccygeal vertebrae. identical to any other. Two pairs of processes (i.e., the caudal articular Greene (1968) reported that, in the rat's tail, the and caudal transverse processes) and the hemal chevron bones (= sesamoid bones) extend caudally ridge radiated from the vertebral body formed a as far as the fifth or sixth vertebra from the tip. pentagonal surface (Co14 in Fig. 4). The cranial Wirtschafter and Tsujimura (1961) examined the ends of the coccygeal vertebrae were hexagonal as total number of sesamoid bones in the mouse and a result of six extensions of three pairs of processes reported that there are 43. They counted the (i.e., the cranial articular, cranial transverse and sesamoid bones only in the upper and lower cranial hemal processes). The hexagonal and pen- extremities and apparently ignored the sesamoid tagonal profiles were extremely useful in determi- bones in the tail. The present study found that nations of the cranial and caudal ends of the coc- there are at least 49 sesamoid bones in the tail. cygeal vertebrae. The hemal arch that has been Thus, the total number of sesamoid bones in the described in the rat (Hebei and Stromberg, 1986) mouse is estimated to be more than twice the num- was not found in the mouse. ber proposed by Wirtschafter and Tsujimura. The sesamoid bone between the second and third coc- 9. The sesamoid bones cygeal vertebrae resembles a "wing-nut" and is A pair of sesamoid bones was present at each different, in this respect, from other "sesame-seed- intervertebral space from Col to Co26. Thus, there like" bones. The coccygeal artery and vein pass were 25 pairs or 50 sesamoid bones in a tail. How- under the wing-nut sesamoid bone. In the rat, ac- ever, the members of the second pair of sesamoid cording to Hebel and Stromberg (1986), the coccy- bones often fused with each other to form a wing- geal vertebrae have a pair of hemal processes nut-shaped (sesamoid) bone. Therefore, the total (= processus hemalis). The hemal processes fuse number of sesamoid bones in the tail region was with each other to form an arch (= arcus hemalis or usually 49. Tail tendons converged on the sesamoid hemal arch) at the third coccygeal vertebra and the bones. The sesamoid bones had no bony connec- arch allows the tail's blood vessels to pass through tions to the vertebrae and, therefore, they cannot it. In the mouse, no hemal arch was found at the be considered true substructures of the vertebrae. third coccygeal vertebra. Instead, a wing-nut sesamoid bone allowed passage of the blood vessels. Adjacent vertebrae were joined to one another Discussion by two types of connection: body-to-body connections via intervertebral discs; and connections via The present study revealed the morphology of the caudal and cranial articular processes. The lat- the vertebrae along the entire length of the verte- ter type of connection determines the directions of bral column of the mouse. Six long-range and four- vertebral movements and limits the extent of such teen short-range substructures that characterize the movements (Shinohara, 1997). Connections via the vertebrae were identified. Most of these sub- articular processes disappeared between the third structures have previously been recognized in the and fourth coccygeal vertebrae. Thus, the tail distal rat, but some are new or specific to the mouse. For to the fourth coccygeal vertebra has maximum example, the foveal process appears to be a new flexibility. The tail is defined as a region of the substructure of the presacral level. The caudal coccygeal vertebrae. In terms of movement, the tail transverse process, cranial hemal process and he- seems to be divided into a less flexible base, which mal ridge are new descriptive terms for structures is proximal to the third coccygeal vertebra, and a Morphology of Mouse Vertebrae 23 very flexible body, which is distal to the fourth coc- address represents the combination of active and cygeal vertebra. Wirtschafter and Tsujimura (1961) inactive selector genes that tells the founding cells thought that the wing-nut sesamoid bone in the and their descendents which part of the body they digits of the limbs "limits, confines, shields and should generate. These authors suggested that sim- guides the structures that glide back and forth ilar processes might occur in vertebrates. Let us within the clasp or groove of the sesamoid". The consider the size of the foveal process as a possible coccygeal artery and vein are vulnerable soft struc- example of such a phenomenon. The levels of tures that require protection by the wing-nut ses- morphogens that activate (or inactivate) the selec- amoid bone at the junction between the relatively tor genes to increase (or decrease) the size of the immovable base and the extremely flexible body of foveal process are maximal (or minimal) at the the tail. eighth thoracic level, with increases (or decreases) O'Higgins et al. (1997) measured the depth and from the sixth thoracic sixththoracic level and de- width of the neural canal at the presacral vertebrae creases (or increases) toward the tenth thoracic in two strains of mice and in humans and they level, as judged from the morphological crescendo- found that the dimensions of the neural canal var- decrescendo pattern. ied little within and between species in terms of The concept of structural homology is important patterns of metemeric variation. Therefore, they in classical human osteology. For example, the proposed that the morphology of the vertebral posterior tubercle of the cervical vertebrae is canal might be regulated primarily by the highly thought to be homologous to the transverse process conserved Hox code, with little influence by other of the thoracic vertebrae, the accessory process of genetic and/or enviromental factors. This proposal the lumbar vertebrae and the lateral part of the might be valid for vertebrae at presacral levels but sacrum. Furthermore, the anterior tubercle of the not for those at more distal vertebrae. In humans, cervical vertebrae is thought to be homologous to incomplete fusion of vertebral arches (= sacral the costae at the thoracic level, the transverse (or hiatus) occurs physiologically at sacral levels costal) process at the lumber level and the lateral (Thompson, 1917; Southworth et aL, 1943). Thus, part at the sacral level (Frick et al., 1991; Bannister the neural canal is not formed and the depth of the et aL, 1995). Since the vertebral body, the superior neural canal cannot be measured at sacral and and the inferior articular processes and spinous more distal levels. In mice, spina bifida occurs con- process are all present at all presacral vertebrae, sistently at the coccygeal level, as shown in the the human vertebrae are thought to be organized present study. These observations suggest that, by these common substructures, even if some of the even if the morphology of the neural canal in the substructures are named differently according to sacral and more distal regions might be determined the vertebral level. In the mouse, the transverse by Hox codes, the Hox codes of mice and humans process that is present at the first to ninth thoracic are either different or are expressed in different vertebrae changes into the cranial and caudal ways. tubercles at the tenth thoracic vertebra, and the The present work found that the morphological caudal tubercle exhibited a crescendo-decrescendo features of vertebral substructures changed gradu- pattern at the thoracic and proximal five lumbar ally with crescendo, plateau or decrescendo pat- levels. The caudal tubercle is thought to be a sub- terns over just a few to as many as twenty vertebral structure that is homologous to the thoracic trans- levels along the craniocaudal axis. Similar patterns verse process, as judged from the presence of an have been demonstrated in the human vertebral apparent intermediate structure at the tenth tho- column (Agur et al., 1991). However, no explana- racic vertebra and the positional continuity be- tion in terms of the origin of such patterns has been tween them. The coexistence of the caudal tubercle proposed. The involvement of morphogens is one and a sickle-shaped transverse process at the lum- possible explanation. A morphogen is defined as bar vertebrae suggests that the thoracic and lumbar the gene product of specific cells that participates transverse processes might not be homologous. in the formation of body structures, that diffuses One might speculate that it is the costae and not the and that influences cells in adjacent areas in a thoracic transverse processes that are homologous concentration-dependent manner (Lewis et al., to the lumbar transverse processes. However, there 1977; Jeremy et al., 1990). Lawrence et al. (1996) is no intermediate form between the costae and suggested that, in Drosophila, positional informa- lumbar transverse process, and there is also no tion in the form of morphogen gradients might structural continuity: the costae are separated from assign cells to non-overlapping sets, with each set the body while the lumbar transverse processes are founding a compartment and, moreover, that each not. Therefore, the concept of homology along the such compartment acquires a genetic address. This entire length of the vertebral column does not 24 H. Shinohara

apply to the mouse. In Drosophila, the metameric omy, 38th ed., 1995; pp. 511-537, Churchill Livingstone, body pattern is determined by the hierarchy of ma- London. ternal effect genes, segmentation genes and home- 3) Cook M. Skeleton, In: The Anatomy of the Laboratory Mouse. 1965;pp. 15-52, Academic Press, London. otic genes (Lewis, 1978; McGinnis and Krumlaus, 4) Cook M. Axial skeleton, In: The Mouse in Biomedical Re- 1992). Kessel and Gruss (1991) claimed that the search, vol. 3, 1983;p. 104, eds. Foster HL, Small JD, Fox identity of a vertebral segment is specified by a JG, Academic Press, New York. combination of functionally active Hox genes, 5) Frick H, Leonhardt H and Starck D. Vertbral column, In: namely a Hox code. The Hox code changes from Human Anatomy, vol, 1, 1991; pp. 451-477, Georg Thieme Verlag, Stuttgart. one vertebral segment to the next, an observation 6) Green EL. Quatitative genetics of skeletal variations in the that suggests that the structures of the vertebra at mouse. II. Crosses between four inbred strains (C3H, each level depends on the specific genetic design. DBA, C57BL, BALB/c). J Genetics 1962;47:1085-1096. Although the structure of the vertebral body is de- 7) Greene EC. Skeleton, In: Anatomy of the Rat, 1968;pp. 5— termined by Paxl along the entire length of the 30, Hafner Press. 8) Griineberg H. Genetical studies on the skeleton of the vertebral column in mice (Wallin et aL, 1994), the mouse. I. Minor variations of the vertebral column. J Ge- gene that affects the formation of the transverse netics 1950; 50:112-141. process at each vertebral level has not been identi- 9) Hebei R and Stromberg MW. Osteology. In: Anatomy and fied. The vertebrae should be considered in terms Embryology.Biomed Verlag, WOrthsee, 1986; pp. 9-24. of their structural individuality, which yields a cre- 10) Hummel KP, Richardson FL and Fekete E. Skeleton, In: Biology of Laboratory Mouse, ed. Green EL, 1966; pp. scendo, plateau or decrescendo pattern over a lim- 247-249, Dover Publications Inc. ited number of segments, rather than in terms of 11) Jeremy B, Green A and Smith JCS. Graded changes in any structural uniformity along the entire length of dose of a Xenopus activin A homologue elicit stepwise the craniocaudal axis. transitions in embryonic cell fate. Nature 1990; 347:391— 394. 12) Jonson DR and H'Higgins P. The inheritance of patterns of metameric variation in the mouse (Mus musculus) verte- Acknowledgements bral column. J Zoo!, London 1994;233:473-492. 13) Kessel M and Gruss P. Homeotic transformation of murine The author is extremely grateful for Professor vertebrae and concomitant alteration of Hox codes induced Shigenori Tanaka, Department of Anatomy, School by retinoic acid. Cell 1991;67:89-104. 14) Lawrence PA and Struhl G. Morphogens, compartments, of Medicine, University of Kanazawa, for helpful and pattern: Lessons from Drosophila? Cell 1996; 85:951— discussions and financial assistance. 961. 15) Lewis EB. A gene complex controlling segmentation in Drosophila. Nature 1978;276:565-570. References 16) Lewis J, Slack JMW and Wolpert L. Thresholds in development. J theor Biol 1977;65:579-590. 1) AgurAMR and Lee MJ. The Back,In: Grant'sAtlas of 17) McGinnis W and Krumlauf R. Homeobox genes and axial Anatomy,9th ed., 1991; pp. 199-254,Williams & Wilkins. patterning. Cell 1992;68:283-302. 2) BannisterLH, Berry MM, Collins P, DysonM, Dussek JE 18) O'Higgins P, Milne N, Johnson DR, Runnion CK and Ox- and FergusonMWJ. Vertebral column, In: Gray'sAnat- nard CE. Adaptation in the vertebral column: a compara-

Explanations of Figures

Plate I

Fig. 1. Dorsal view of the vertebral column. Transverseprocess The transverse process is square at Cl, club-shaped at C2—T9,sickle-shaped at L1—L6,wing-shaped at 51 and S2 and a parallelogram at S3—Co4.At Co5, the transverse process splits into the cranial (open arrows) and caudal (closed arrows) portions. Thus, the coccygeal vertebrae distal to Co5 resemble an hourglass. The orientation of the process is caudolateral at C2—05,lateral at C6—T2and craniolateral at T3—T9, L1—L6 and S3—Co4. Articular process The adjacent cranial and caudal articular processes are joined together from C1—C2to Co2—Co3,as indicated by the arrows between Co2 and Co3. However, they are not joined together between Co3 and Co4 (arrows). The vertebrae distal to Co4 are, therefore, connected to one another only via the intervertebral disc. Spinousprocess The spinous process at T2 is exceptionallyhigh (circle).The leading angle of the spinous process at T6 is directed caudally (white arrow) but that at T13 is directed cranially (white arrow). The change in direction occurs consistently between T10 and T11 (asterisk). The spinous process usually has a single spine but becomes bifid at T10 (arrows). The vertebral arch is present at all vertebrae from Cl to Co4 (asterisk) but is absent at Co5 and more distal vertebrae. Thus, the spinous process is absent at the latter vertebrae. Alar foramen The foramen (arrows) opens ventrodorsally at Cl. Bar = 1 mm. Morphology of Mouse Vertebrae 25 Plate I 26 H. Shinohara

tive study of patterns of metameric variation in mice and gesia in surgery. Ann Surg 1943; 118:945-970. men. J Anat 1997; 190:105-113. 22) Thompson J E. An anatomical and experimental study of 19) Searle AG. Genetical studies on the skeleton of the mouse. sacral anaesthesia. Ann Surg 1917;66:718-727. IX. Causes of skeletal variation within pure lines. J Genet- 23) Wallin J, Wilting J, Koseki H, Fritsch R, Christ B and ics 1954; 52:68-102. Balling R. The role of Pax-1 in axial skeleton development. 20) Shinohara H. Changes in the surface of the superior artic- Development 1994; 120:1109-1121. ular jointfromthe lower thoracic to the upper lumbar ver- 24) Wirtschafter ZT and Tsujimura JK. The sesamoid bones in tebrae. J Anat 1997; 190:461-465. the C3H mouse. Anat Rec 1961; 139:399-408. 21) Southworth JL and Hingson RA. Continuous caudal anal-

Plate II

Fig. 2. Left lateral view of the vertebral column. Transverseprocess There is a club-shaped transverse process at T9 (white arrow). At T10, the transverse process is replaced by the cranial (open arrow) and caudal (closed arrow) tubercles. The cranial tubercle is not present at T11. The cranial articular process at T11 is larger than that at T10, as if it has absorbed the cranial tubercle. The length of the caudal tubercle exhibits a crescendo-decrescendo pattern from T10 to L5 (closed arrow) with a maximum at L2. Note that caudal tubercles (closed arrows) and transverse processes (white arrows) are both present from Li to L5. Articular process The connection of the caudal and cranial articular processes forms the intervertebral foramen (for example, between T6 and 11; asterisk). The size of the intervertebral foramen varied with the vertebral level. The articular processes at Co2 and Co3 are connected to form the intervertebral foramen (asterisk), whereas there is no such connection between Co3 and Co4 and no intervertebral foramen is formed (double asterisks). Spinousprocess The process at T2 is exceptionally high (asterisk). The processes from T3 to T10 are triangular with the leading angle oriented caudally in each case. The leading angle at T11 is the opposite of that at T10 (asterisk). A spinous process is present at Co4 (arrow) but absent at Co5 (arrow). Closed arrow at Cl, Ventral tubercle; Open arrow at Cl, Dens axis; Arrow at C6, Ventral lamina. The sesamoid bones between Co8 and Co9 (arrow), for example, are completely isolated from the vertebrae. Bar = 1 mm. Morphology of Mouse Vertebrae 27 Plate II 28 H. Shinohara

Plate III

Fig. 3. Ventral view of the vertebral column. Vertebral body Note that the craniocaudal length chages with the vertebral level (see Figure 5). Hemal process At Co5, both cranial (open arrows) and caudal (closed arrows) hemal processes are present. The hemal arch that found in the coccygeal vertebra of the rat (Hebel and Stromberg, 1986) is absent in the mouse. Sesamoid bones A pair of sesamoid bones appears first between Col and Co2 (arrow) and last between Co25 and Co26 (arrow). The sesamoid bones between Co2 and Co3 have fused to form a wing-nut-shapedbone (arrow). Thus, there are at least 49 sesamoid bones in the tail. As shown in the lateral view (Fig. 2), the sesamoid bones do not have bony connections with the . vertbrae, an observation that suggests that they are not true substructures of the vertebrae. Closed arrow at Cl, Ventral tubercle; Open arrow at Cl, Dens axis; Arrows at C6, Ventral laminae. Bar = 1 mm. Morphology of Mouse Vertebrae 29 Plate HI 30 H. Shinohara

Plate IV

Fig. 4. Cranial views of the vertebrae. Vertebral body There is no vertebral body at Cl. The body at C2 has a dens axis (arrow). The body is oval from 0 to T2, bell- shaped from T3 to T10, trapezoid from T12 to L2 and triangular from L6 to 53. The shape changes gradually and there are intermediate shapes. Transverseprocess This process extends laterally at Cl—T1,dorsolaterally at T2-T8, ventrolaterally at Li—L6,laterally at Si—Col and ventrolaterally at Co2—Co4. Articular process The cranial articular process (CP) is located dorsal to the transverse process (TP) from Cl to Ti. This positional relationship is expressed as CP—TP.The positional relation is TP—CPfrom T2 to T9 and again CP—TPfrom Li to Co3. The joint surface of the cranial articular process is curved at Cl but flat from C2 to T10 (straight arrows at T10). It is again curved at T11 (curved arrows) and remain curved to Co3. Vertebral arch Fusion of the right and left arches results in formation of the vertebral foramen. The changes in area of the vertebral foramen are shown in Figure 6. A vertebral foramen is formed at Co3 (arrow) but not at Co4 (arrow). Transversariumforamen and alar foramen The transversarium foramen is present from Cl to C6 (for example, at Co5, asterisks). The transversarium foramen at Cl (circle) is continuous with the alar foramen that penetrates C1 ventrodorsally (arrows with solid lines). The transversarium foramen is also continuous with the vertebral foramen (arrows with broken lines). Foveal process (tentative) The articular fovea is flat on the transverse process at T5 (arrow) but becomes elevated at T6 (arrow). This elevation (= the foveal process) is present from T6 to T10 (arrow) and its height is maximum at T8. Hemal process and hemal ridge A pair of hemal processes (indicated by H) was usually found at the cranial end of Co5 and occasionally Co4. Members of another pair of hemal processes at the caudal end are so small and so close to each other that they form a slight midsagittal elevation (= hemal ridge; arrow at Co14). The cranial surface of Co5 and more distal vertebrae is hexagonal in shape since three pairs of processes extend from the body, whereas the caudal surface of Co5 and more caudal vertebrae is pentagonal with two pairs of processes extending from the body and an unpaired hemal ridge. Open arrow at Cl, Ventral tubercle; Arrows at C6, Ventral laminae. Bar = 1 mm. Morphology of Mouse Vertebrae 31

Plate IV