Chapter 4
Invertebrate Cartilages, Notochordal
Cartilage and Cartilage Origins
The Cell Theory, which states that all organisms are
composed of cells, is usually traced to observations
made in the seventeenth century by English polymath
Robert Hooke (1635 1703). Using a microscope he had
made, Hooke (1665) described the honeycomb-like
network in thin slices of cork as a network of cellulae,
Latin for small storage rooms. Although what Hooke
saw represented only the thickened walls of previously
living plant tissue cork is dead and dried bark derived
from the cork oak Quercus suber Hookes cellulae are
what we now know as cells.
As introduced in Chapter 2, the Dutch microscopist
Antony van Leeuwenhoek published a description of
the microstructure of bone in 1693. A decade earlier,
Leeuwenhoek was the first to publish on the presence in
pond water of single-celled organisms (now known to be
protozoans). Leeuwenhoek described red blood cells and
sperm as single cells. It was another 150 years, however,
before cells were recognised as the basic units of all life.
That recognition came within the span of 12 months in
1838 1839. It resulted from the research of two professors
at Jena University in Germany, a physiologist named
Theodor Schwann (1810 1882) and a botanist named
Matthias Jakob Schleiden (1804 1881).
Based on his extensive microscopical observations on
plant tissues and plant development, a Cell Theory was
articulated by Schleiden (1838), who asserted that all
plant tissues are composed of cells, that each plant begins
its life as a single cell and that the cell is the basic build-
ing block of all plant matter. Cartilages played a central
role for Schwann (1839), who described cellular struc-
tures in cartilages from animals.
One of the generalities that led to the formulation of
the Cell Theory was the recognition by Schwann that
many invertebrates have a cartilage-like chondroid,
chordoid, mucoid connective tissue, structurally simi-
lar to the parenchymal tissues of plants; both plant cells
and cartilage cells were surrounded by a rigid extra-
cellular matrix or wall. Schwann (1839) concluded that all
Bones and Cartilage. DOI: http://dx.doi.org/10.1016/B978-0-12-416678-3.00004-5
2015 Elsevier Inc. All rights reserved.
organisms consist of one or more cells, that the cell
is the basic unit of structure indeed of life for
all organisms and that every organism, whether plant or
animal, comprises combinations of cells arranged in
predictable patterns.
CHONDROID AND CARTILAGE
Schaffer (1930) exhaustively summarised and reviewed
the nineteenth and early twentieth century studies on the
tissues identified as chondroid, chordoid and mucoid by
Schwann. Schaffer did not regard theses tissues as real
cartilages. Their immature histology, combined with a
scant extracellular matrix and inability to detect glycosa-
minoglycans or collagen in their matrices features diag-
nostic of vertebrate cartilage and their failure to
mineralise, led Schaffer to regard these invertebrate tissues
as chondroid tissues. By chondroid, Schaffer meant carti-
lage-like, implying a less fully evolved skeletal tissue than
cartilage. Schaffer classified chondroid tissues into grades
of greater or lesser resemblance to cartilage without,
however, implying any phylogenetic trends or affiliations1.
Schaffers view prevailed, and skeletal biologists
turned to other problems. Indeed, no research publications
on invertebrate cartilages appeared until 1959. Then in
the 1960s and 1970s, primarily through the work of one
man, Philip Person, several structural studies appeared.
Person was thorough. He used light and electron micros-
copy, biochemical characterisation of extracellular matri-
cial products, histochemistry and mineralisation, and he
provided some results on the regenerative ability of inver-
tebrate cartilage2, With the exception of one study by
Person, no further information on the development of
invertebrate cartilages was available until Alison Cole
began studies in my laboratory this century (Cole and
Hall, 2002, 2004a,b, *2009, and see Chapter 5).
By 1969, sufficient data had accumulated that Person
and Philpott could review the work of the previous
decade to present evidence in favour of regarding these
63
64 PART | II Origins and Types of Skeletal Tissues

invertebrate tissues as cartilages. Evidence included cells G tentacular cartilages of two polychaete annelids, the
in lacunae embedded in a matrix containing collagen feather duster worms Eudistylia polymorpha and
and chondroitin sulphate-A or sulphate-C (chondroitin- Sabella melanostigma; and
4-sulphate and chondroitin-6-sulphate, respectively), G lophophore cartilage in the lampshell, Terebratalia
producing a tissue providing mechanical support and a transversa, an articulate brachiopod7.
site for muscle and ligament attachment (Person and
Philpott, 1969a)3.
A deep homology of extracellular matrices, and so a ODONTOPHORE CARTILAGE
monophyletic origin of the animal kingdom, has been IN CAENOGASTROPODS
proposed on the basis of numerous lines of evidence. In
Caenogastropods (sea snails, whelks), the most diverse
an analysis placing fibrillar collagen in a key position in
and dominant gastropods of marine habitats, comprise
vertebrate evolution fibrillar collagen arose before the
some 60% of all gastropod species. Odontophoral carti-
genome duplication at the base of vertebrate evolution
lages are present in the buccal mass dorsal to the radula
Boot-Handford and Tuckwell (2003) discuss invertebrate
where they support the chitinous and toothed radula
collagens only briefly. Older studies saw collagen as
during feeding. In one of the few functional morpho-
first present in sponges, in part because sponges were
logical studies on invertebrate cartilage, Guralnick and
considered the most basal metazoans4. Although that
Smith (1999) examined the phylogenetic distribution of
phylogenetic position has been challenged, the presence
cartilage in gastropods including criteria to classify
in basement membranes at the divergence of sponges
cartilages and looked at correlations between the evolu-
from cnidarians of type IV collagen cross-linked by
tion of gastropod cartilages and the radula, the cartilage
sulphilimine bonds, the extracellular matrix associated
performing a supporting role.
peroxidase peroxidasin (which catalyses formation of
The channelled whelk, Busycon canaliculatum, has
sulphilimine bonds), and hypohalous acids (which inter-
a large odontophore cartilage consisting of hypertrophic
act with unsaturated phosphatidylcholines) lies at the
cells in capsules separated by scant amounts of extra-
basis of the evolution of tissues such as cartilage (Fidler
cellular matrix (ECM) (Figure 4.1). The scant ECM of
et al., 2013).
these invertebrate cartilages has been a stumbling block
Although type I is the most widespread collagen,
to accepting them as cartilages. However, a number of
invertebrates contain several collagen genes. Type II
vertebrate cartilages are characterised by scant amounts
collagen in cartilage arose with the vertebrates, which
of ECM, including xiphisternal, secondary, and callus car-
Cole and Hall (2004a,b) interpreted as indicating the
tilages in jawed vertebrates and lampreys, and cell-rich
primitiveness of collagen with three identical chains dur-
cartilage in teleosts.
ing evolution. Indeed, a phylogenetic analysis of fibrillar
In general, the glycosaminoglycans of invertebrate
collagen evolution by H. Wada et al. (2006) provided evi-
cartilages are oversulphated in comparison to those from
dence that the common ancestor of all deuterostomes had
vertebrate cartilages. Fitting this pattern, odontophore
three fibrillar collagen genes that were co-opted into for-
cartilage lacks chondroitin sulphate, having instead a
mation of a notochordal sheath with the origin of the
polyglucose sulphate akin to chitin8. Echinoderm connec-
chordates, a topic taken up again in Chapters 41 and 425.
tive tissues have a similar polyfucose sulphate, but also
The distribution of elastin in elastic cartilages and in
have chondroitin sulphate. Odontophore cartilages have
lamprey mucocartilage was discussed in Chapter 3. periodicity and an X-ray dif-
collagen fibres with a 640 A
Previously regarded as absent from all invertebrates
fraction pattern similar to that from vertebrate collagen.
and present in all craniates other than hagfishes and lam-
Busycon cartilage contains myoglobin within its cells.
preys, elastin arose phylogenetically with the evolution of
This is unusual because the blood pigment in molluscs is
closed, high-pressure circulatory systems, although elastin
haemocyanin and cartilages are normally not pigmented9.
is arranged differently from vertebrate to vertebrate6.
Busycon cartilage was the first cartilage from any
The nature of invertebrate cartilages can, perhaps, best
taxon vertebrate or invertebrate found to have cyto-
be reviewed by summarising our knowledge of the carti-
chrome oxidase. Indeed, the low utilisation of oxygen
lages in different taxa. To that end I discuss
by vertebrate cartilage prompted Person et al. (1959) to
G odontophore cartilages in caenogastropods; examine invertebrate cartilages in the first place. They
G branchial (gill book) cartilage in an arthropod, the were prescient. Paradoxically, low oxygen utilisation was
horseshoe crab Limulus polyphemus; coupled with high levels of dehydrogenases requiring
G cranial cartilages of several cephalopods the long- cytochrome oxidase and considerable synthesis of sugars
fin squid Loligo pealii, the common cuttlefish Sepia requiring oxygen, an unexpected mix of metabolic
officinalis, and the common octopus Octopus vulgaris; parameters.
 Chapter | 4 Invertebrate Cartilages, Notochordal Cartilage and Cartilage Origins 65
A B
FIGURE 4.1 The histology of invertebrate cartilages. (A) Odontophoral cartilage of the marine snail, Busycon canaliculatum, with large cells and
scant extracellular matrix. (B) Cranial cartilage of the longfin squid, Loligo pealii, with abundant extracellular matrix. (C) Branchial cartilage of the
horseshoe crab, Limulus polyphemus, again with scant extracellular matrix. All stained with haematoxylin and eosin. Adapted from Person (1960).
Taking advantage of micro-computed tomography
(micro-CT) imaging and computer-based methods of
three-dimensional (3D) reconstruction, Golding et al.
(2009) reconstructed odontophoral cartilages from 18
species of caenogastropods10. The images obtained
illustrate the diversity and complexity of odontophoral
cartilages across the gastropods. Basal species have
a previously unknown cartilage beneath the radula
(a subradular cartilage), which may be a homologue of
the dorsal odontophoral cartilage of more derived taxa.
Little else is known about mollusc cartilages outside gas-
tropods (above) and cephalopods (below) aside from an
intriguing study on regeneration of the proboscis. After
the proboscis is amputated, members of a number of
genera of boring prosobranch molluscs can regenerate a
new proboscis. A single aggregation of cells gives rise to
cartilage and muscle of the new proboscis (Carriker
et al., 1972). A single aggregation is consistent with
muscle and cartilage arising from a single population of
cells as occurs in amphibian limb regeneration, which is
discussed further in Chapter 14.
BRANCHIAL (GILL BOOK) CARTILAGE
IN THE HORSESHOE CRAB, LIMULUS
POLYPHEMUS
The gill books of the horseshoe crab, Limulus polyphe-
mus, contain a cartilaginous endoskeleton (Figure 4.1).
The scant extracellular matrix contains chondroitin-4-
sulphate which the whelk does not and reacts as
does vertebrate cartilage to histochemical tests for glyco-
saminoglycans, although a novel sulphated oligosaccha-
ride containing 3-o-sulphated glucuronic acid was isolated
from chondroitin sulphate K from Limulus cartilage11.
C
Hydroxyproline and hydroxylysine are present in
amounts typical of vertebrate cartilage, and with X-ray
diffraction patterns typical of collagen. Lipid and glyco-
gen are found as storage products, whereas cytochrome
oxidase indicates potential aerobic metabolism. As with
all invertebrate cartilages, gill book cartilages are unmi-
neralised. The presence of lipids, which in vertebrate skel-
etal tissues play a role in mineralisation, is of interest
(see the section on mineralisation below).
As gill book chondrocytes divide, a phragmosome-like
structure12 reminiscent of the membrane formed during
the division of supporting tissue in plants is deposited
between the daughter cells and provides the basis for
new ECM. As the cells mature, the entire chondrocyte is
chondrified; its cytoplasm fills in with matrix products
and the cell dies, a process reminiscent of lignification in
plants. This unusual pattern of chondrification may
explain, in part, the entirely appositional growth of gill
book cartilage, which contrasts with interstitial growth of
cartilages in other cephalopods (below)13.
CRANIAL CARTILAGES IN SQUID,
CUTTLEFISH AND OCTOPUSES
The longfin squid Loligo pealii, the common cuttlefish
Sepia officinalis, and the common octopus Octopus
vulgaris all contain extensive cranial cartilage that pro-
tects the brain and a scleral cartilage that protects the
eye (Figure 4.1). Unlike cartilages in the whelk and horse-
shoe crab, these cranial cartilages have extensive matrix,
sparsely scattered cells and cell densities more typical
of vertebrate bone than cartilages. The matrix may be
fibrous, often with a canalicular system coursing through
it. Cell processes lie within the canals, connecting cells
one to another and giving them the appearance of
66 PART | II Origins and Types of Skeletal Tissues
osteocytes. Some squid have been reported to have carti-
laginous epidermal scales, which would be well worth
further investigation14.
Composition of the Extracellular Matrix
Great diversity of matrix products is seen in cephalopods.
It is unclear if such diversity exists in other invertebrate
groups. Although a number of cephalopod species have
been examined, only one or two species have been stud-
ied in most other groups. Histochemical evidence shows
fewer glycosaminoglycans and a different spectrum of
matrix products in cephalopod cartilage than in Limulus,
emphasising the need for more biochemical characterisa-
tion of matricial products.
Glycosaminoglycans
The predominant glycosaminoglycan is chondroitin-4-
sulphate, as in Limulus. Squid cartilage lacks hyaluronan,
and the core protein differs from vertebrate cartilage
glycosaminoglycan core protein. Squid chondroitin sul-
phate contains novel tetrasaccharide sequences; the
common squid, Ommastrephes sloani pacificus, has a
glucuronic acid containing glycopeptide distinct from
any other glycosaminoglycans15.
Cartilages of the brown or short-finned squid, Ilex ille-
cebrosus coidentii, contain hyaluronan (Box 4.1), a previ-
ously undescribed heavily oversulphated chondroitin
sulphate, and a previously undescribed 39,200 molecular
weight polysaccharide composed mainly of glucuronic
acid, galactose, and mannose in the ratio of 1:2:1.
Extraction and characterisation of proteoglycans from the
Japanese flying squid, Todarodes pacificus, show that one-
fourth of the proteoglycans from cranial cartilages do not
aggregate with hyaluronan. Cranial cartilage is low in pro-
teoglycan content, the proteoglycans do not interact with
hyaluronan, and keratan sulphate is absent (Box 4.1)16.
Cranial cartilage from the arrow squid, Nototodarus gouldi,
contains highly sulphated chondroitin sulphate E. The
epidermis contains unsulphated, but highly glycosylated,
glycosaminoglycans (Falshaw et al., 2000).
Tsilemov et al. (1998) showed that a 35-kDa protein
from squid cartilage species not specified, but the carti-
lage lacks hyaluronan reacts to antiserum against sheep
link protein and binds to the same G1 domain of aggrecan,
which is the globular domain at the N-terminus that
functions as the hyaluronan-binding region. This finding
suggests a role for this protein similar to that carried out
by aggrecan in vertebrates, which is to bind to hyaluronan
to modulate the osmotic properties of cartilage (Box 4.2).
In vertebrates, synthesis of aggregan is enhanced in mes-
enchyme subjected to compression, whereas expressing
aggrecan or aggrecan link protein in cultured fibroblasts or
chondrocytes disrupts cell substrate interaction17.
Collagens
Cranial cartilage collagen in the Japanese flying or com-
mon squid, Todarodes pacificus, resembles vertebrate
type I collagen (Table 4.1).
A combined light, polarising and ultrastructural micro-
scopical study of the head cartilages of Sepia officinalis
and Octopus vulgaris revealed 10 25-nm-diameter colla-
gen fibres similar to vertebrate type I (albeit with some-
what different banding patterns), polymeric aggregates
and a defined perichondrium features typical of verte-
brate cartilages but with chondrocytes connected to
one another by cell processes and blood vessels within
the cartilage, which are features of vertebrate bone and
not cartilage (Figure 4.1). Bairati et al. (1987) proposed
that vascularisation of these cartilages provides a high
level of nutrition, eliminating the need for mineralisa-
tion. This and subsequent studies by Bairati et al. (1989,
*1999) provide the most detailed information on any
squid or octopus cartilage. They undertook
G a transmission electron microscopic (TEM) study of
the perichondrium, demonstrating a fibrous layer that
was discontinuous at connective tissue or muscle
attachments, and remarkably vertebrate-like;
G a TEM study of chondrocytes from the cranial cartilages
of both species, demonstrating many long processes
connecting cells (considered reminiscent of vertebrate
osteocyte processes), absence of secretory granules,
presence of rough endoplasmic reticulum cisternae and
of vesicles opening to the cell surface, and haemocyanin
in many chondrocytes and gap junctions; and
G an immunological study of collagens from Sepia offi-
cinalis using vertebrate and cuttlefish collagen antibo-
dies. Response to antibodies against cuttlefish type I
and rat type V was intense. Reactions to chick anti-
type I and to calf anti-type II were weaker. Indeed,
and surprisingly, suggesting that Sepia collagen is not
highly derived, except for antibodies against mamma-
lian type I collagen, all antibodies tested cross-react
with Sepia cartilage, implicating type I collagen at an
early step in skeletal evolution (Chapter 2).
Sepia cranial cartilage contains a novel collagen with
three distinct chains named C1 C3 consisting of
105, 115 and 130 kDa, respectively. This collagen is more
cross-linked than shark cartilage-type I collagen. Of inter-
est, an antibody against shark cartilage cross-reacts with
Sepia cartilage but a Sepia cartilage antibody does not
cross-react with shark cartilage. Based on rotary shadow-
ing electron microscopy, the matrix collagen fibres have
been compared with those of chick sternal cartilage18.
 Chapter | 4 Invertebrate Cartilages, Notochordal Cartilage and Cartilage Origins 67
BOX 4.1 Hyaluronan
Hyaluronan, a high-molecular-weight polyanion with a highly
negative charge, occurs in low concentrations in many animal
tissues. A single molecule may have a molecular weight of
10 million. Consisting of an alternating polymer of glucuronic
acid and N-acetylglucosamine joined by a 1 3 linkage, hya-
luronan expands enormously in solution to form a net of
extended, random-coiled molecules. Hydrated, it typically
occupies the volume of a sphere 400 nm in diameter a thou-
sand times greater than the volume of the nonhydrated chain.
Hyaluronan then known as hyaluronic acid was iso-
lated in 1934. Hyaluronan is degraded by hyaluronidase.
Much of the research over the 40 years after the discovery of
hyaluronan was biochemical, with few suggestions of associa-
tions with cell proliferation, migration, or aggregation, or of
the presence of hyaluronidase and therefore hyaluronan
removal with subsequent cell differentiation. Perhaps the
first to demonstrate these links with cell behaviour were Mauer
and Hudack (1952), who showed that a large amount of hya-
luronan is synthesised in the early stages of callus formation
during repair of fractured long bones. Hyaluronan plays impor-
tant roles in vertebral and limb chondrogenesis, migration of
neural crest cells, regeneration of amphibian limbs, ectopic
chondrogenesis and osteogenesis (for which, see Chapter 14),
and corneal development, to name only a few examples.
Although some hyaluronan is intracellular, most lies within
extracellular matrices, where it forms the backbone to which
proteochondroitin sulphate is attached. This complex forms
large aggregates with aggrecan within cartilaginous ECM
(for which see Box 4.2). Both hyaluronan and the aggregate
impede transport and help to maintain the greatly hydrated
nature of cartilaginous ECM. Prechondrogenic cells produce
only monomeric proteoglycan. With progressive differentia-
tion, larger and more aggregated forms of proteoglycan are
deposited. Of interest, as they age in vivo or in vitro, chondro-
cytes revert to synthesising small monomer proteoglycans, a
change that may be a form of dedifferentiation, as occurs, to
some extent, in some tumours. With chondrocyte differentia-
tion, hyaluronan is localised preferentially in the hypertrophic
zone of the growth plate, concentration correlating with
lacuna size, reflecting water-binding properties; lacunae fail
to enlarge if chondroblasts are cultured in hyaluronidasea.
Hyaluronan is removed from regions at which joints
between adjacent skeletal elements will form. Spicer and Tien
(2004) review the role played by hyaluronan in morpho-
genesis, especially condensation and joint cavitation during
skeletogenesis. As summarised in Table 1 in their paper,
Spicer and Tien identify 30 genes associated with the synthe-
sis, function and/or degradation of hyaluronan. The enzyme
hyaluronic acid synthase-2, coded by the gene Has2, plays an
important role in removing hyaluronan at presumptive joint
sites through a pathway involving Shh regulation of Gli3,
which binds to a promoter in the Has2 gene (J. Liu et al.,
2013). This study using mouse limb buds sheds light on the
known reduction in Gli3 in limbs buds of talpid2 chick
embryos, discussed in Chapter 39.
40
30
Hexosamine
20
Y = 3.7415x 0.9505
10 r = 0.996
0
0 2 4 6 8 10
Hydroxyproline
FIGURE 4.2 Concentrations of hexosamine and hydroxyproline as
mg/pair of tibiae from chick embryos incubated for 12 days are sig-
nificantly correlated (r 5 0.996) in this study comparing results from
control embryos and from embryos treated with vitamin C or with
the metal thallium. See the text and Hall (1973b) for details of
treatments.
We have some indications of how levels of hyaluronan
and chondroitin sulphate (CS) are correlated (Figure 4.2).
Synthesis of hexosamine is enhanced when tibial cartilage is
treated for 2 days with hyaluronidase and then organ-cultured
for 4 days in the absence of hyaluronidase. Less CS is synthe-
sised than in untreated cartilage; the CS has a shorter chain
length than normal; and CS is uncoupled from the hyaluronan
backbone of the hyaluronan proteochondroitin sulphate aggre-
gate. Synthesis of CS is depressed when hyaluronan is added to
cultures of chondrocytes from embryonic or adult mammals or
birds, which suggests that synthesis of CS is controlled at
least in part by the level of hyaluronan. The effect of hyaluro-
nan is specific. Other glycosaminoglycans chondroitin and
keratan or heparin sulphates do not stimulate synthesis of
CS. Oligosaccharides derived from hyaluronan dob.
Because hyaluronan is primarily extracellular, little atten-
tion has been paid to how hyaluronan might be internalised
by cells. Mammalian articular chondrocytes use the hyaluro-
nan receptor CD-44 . Using rat cranial base synchondroses as
a model, Gakunga et al. (2000) localised hyaluronan to the
hypertrophic but not other chondrocyte zones; autoradio-
graphic and histological analyses of growth of the cartilages in
the midline cranial base of the Norwegian rat Rattus norvegi-
cus to 16 months after birth show that cell proliferation is spe-
cific to individual cartilages. Hyaluronidase and CD-44
colocalise with hyaluronan. H. Jiang et al. (2001) maintain that
hyaluronan is internalised by hypertrophic chondrocytes and
degraded via a CD-44 based mechanism. Synchondroses cul-
tured in Streptomyces hyaluronidase show reduced lacunar
expansion, a finding consistent with hyaluronan-mediated
(Continued )
68 PART | II Origins and Types of Skeletal Tissues

BOX 4.1 (Continued)

hydrostatic pressure playing a role in lacunar expansion during
chondrocyte hypertrophyc.
a. See Larsson and Kuettner (1974) for intracellular hyaluronan; Wiebkin
and Muir (1973), Goldberg and Toole (1984), W. Knudson and Knudson
(1991), Pavasant et al. (1996) and Toole (*1973, *2001) for aggregate for-
mation; and Ovadia et al. (1980) and Pacifici et al. (1981) for synthesis of
monomers by prechondroblasts and ageing chondrocytes. See Pavasant
et al. (1996) for lacunar enlargement, and Toole (*2001) for a review of
hyaluronan and its binding proteins, the hyaladherins. The 2001 paper
deals especially well with the role of hyaluronan in the pericellular matrix
that surrounds chondrocytes.
b. See Hardingham et al. (1972) for synthesis of hexosamine in tibial carti-
lage, and Wiebkin and Muir (1973) and Solursh et al. (1974) for addition of
hyaluronan to cartilage cultures.
c. See H. Jiang et al. (2001) for the CD-44 receptor and see Roberts and
Blackwood (1984) for the study with the Norwegian rat. CD-44 is expressed
in mouse development (9.5 12.5 days) in high levels in the somites, heart,
limb condensations, apical epithelial ridge of limb buds, tooth primordia
and upper and lower jaw mesenchyme (Wheatley et al., 1993).

BOX 4.2 Aggrecan

Aggrecan (cartilage-specific proteoglycan core protein, chon-
droitin sulphate proteoglycan 1, aggregating chondroitin sul-
phate proteoglycan) is the product of the Acan gene and a
member of the aggrecan/versican/proteoglycan family of extra-
cellular matrix proteins. A much as 10% of the dry weight of
some cartilages is aggrecan. As many as 100 aggrecan mono-
mers bind to a single hyaluronan chain via a link protein to
form a high-molecular-weight aggregate, with a globular
domain (G1, the hyaluronan-binding region) at the N-terminus
where the keratan sulphate chains are preferentially located.
Aggrecan shares numerous sequence similarities with versican
(discussed in Box 21.3) and with CD-44 (Ayad et al., 1994).
Aggrecan is a structural proteoglycan, by which I mean
that the aggregation to which aggrecan contributes provides
an important structural property to cartilage, namely to main-
tain an osmotic swelling pressure and the highly hydrated
nature of cartilaginous extracellular matrix. Compressive
forces, such as those acting on articular cartilages, stimulate
synthesis of aggrecan. The core protein of aggrecan is broken
down by proteases, a breakdown that facilitates vascular inva-
sion of cartilage.
Acan mRNA is expressed in periosteal cells of membrane
bones immediately before they begin to differentiate as chon-
droblasts (see the section on evidence for bipotentiality in
Chapter 12). Different mutations in Acan result in dwarfism
in different lineages, including Cartilage matrix deficiency
(Cmd) in mice and Nanomelia (nm) in domestic fowl
(Chapter 27).

TABLE 4.1 Amino Acid Composition of Collagen and Alpha Collagen Chains from Cranial Cartilage of the Japanese
Flying Squid, Todarides pacificus, Compared with Type I Collagen from Rat Tail Tendon and Rat Skina

Squid Rat Tail Tendon Rat Skin

Amino Acid Collagen 1(I) 1(II) Collagen Collagen

3-Hydroxyproline 1 1 1 4.2 1
4-Hydroxyproline 90 92 88 90 92
Alanine 94 102 76 107 106
Arginine 59 58 60 50 51
Aspartic acid 58 53 65 45 46
Glutamic acid 91 92 90 71 71
Glycine 338 341 339 331 331
Histidine 6 6 8 4 5
Hydroxylysine 14 11 24 7 6
Isoleucine 15 11 22 10 11
Leucine 27 25 31 24 24
Lysine 10 12 9 27 28
Methionine 11 11 8 8 8
Phenylalanine 9 10 8 12 11
Proline 90 88 90 122 121
Serine 42 44 36 43 43
Threonine 23 24 20 20 20
Tyrosine 2 3 1 4 2
Valine 20 16 24 23 25
a
From data in Kimura and Karasawa (1985) for squid and in Serafini-Fracassini and Smith (1974) for rat tail tendon and skin. Based on limited pepsin proteolysis
and expressed as residues/1,000 amino acid residues.
 Chapter | 4 Invertebrate Cartilages, Notochordal Cartilage and Cartilage Origins 69
TENTACULAR CARTILAGE IN POLYCHAETE
ANNELIDS
The feather duster worms Eudistylia polymorpha and
Sabella melanostigma have extensive, branched and sub-
divided (segmented?) cartilages at the base of the tentacle
complex, within the tentacles, and in the pinnae that
branch from the tentacles (Figure 4.3). These cartilages
(presumably) give considerable support and flexibility to
the tentacles as a food-gathering organ. The cartilaginous
matrix is sparse and vascularised, causing it superficially
to resemble osteoid. The glycosaminoglycan is a highly
sulphated chondroitin-6-sulphate. Ultrastructural studies
of Sabella cartilage show heterogeneity of matrix
structure19.
All members of the subfamily Sabellinae to which
these two species belong have tentacles (a branchial
A 10 m
B 200 m
FIGURE 4.3 Polychaete and gastropod cartilages. (A) A longitudinal
section of a tentacle of the marine polychaete, the feather duster worm,
Eudistylia polymorpha, showing the central cartilage and the cartilages
of the pinnae. (B) Regenerating odontophoral cartilage of the gastropod,
the marine oyster drill, Urosalpinx cinerea follyensis, 11 days after
amputation of the proboscis. Note the proliferating, blastema-like carti-
lage on the right. Adapted from Person and Mathews (1967) and
Carriker et al. (1972).
crown) supported by large vacuolated cartilage cells,
which are in turn surrounded by a supporting sheath of
extracellular matrix and a vascular supply. This generali-
sation was reinforced by an extensive histological analysis
of the tentacle cartilages in 30 species (18 genera) of
sabellids by Capa et al. (2011). In all species, the base of
the tentacles (dorsal lips) is supported by cartilage, as
described earlier. With elongation, tentacles are supported
either by cartilage with irregularly arranged chondrocytes
or by an acellular extracellular matrix. Capa and collea-
gues review past studies on these tissues, including issues
of terminology, recommending supporting cellular axis
for the basal tissue, supporting acellular axis for the
acellular tissue and hyaline cartilage for the cartilage.
Members of the subfamily Fabriciinae, such as
Fabricia stellaris, on the other hand, support their tenta-
cles with an acellular skeleton, and their tentacles lack
blood vessels. Members of the Serpulidae lack skeletal
support in their tentacles (Randel and Bick, 2012 and
references therein). Such divergence of skeletal structures
in three subfamilies that form a monophyletic clade of
sabellid polychaetes demonstrates the advances that could
be made by future comparative analyses of polychaete
branchial crowns. Indeed, Randel and Bick (2012) con-
sider that the radioles and pinnules that make up the tenta-
cles in these three subfamilies may not be homologous, a
conclusion consistent with the variable nature of the
supporting tissues.
LOPHOPHORE CARTILAGE IN AN
ARTICULATE BRACHIOPOD,
TEREBRATALIA TRANSVERSA
Cartilage in the pedicle of an articulate (two-valve) bra-
chiopod, Terebratalia transversa, was discovered seren-
dipitously during an ultrastructural study of the
development of the pedicle, which is the fleshy stalk that
attaches the animal to the substrate. The cartilage consists
of a weakly metachromatic connective tissue that devel-
ops in the pedicle after metamorphosis. The connective
tissues of the tentacles are acellular and not metachro-
matic after toluidine blue. At the ultrastructural level they
resemble (indeed, appear no different from) vertebrate
fibrous cartilage. The connective tissue in the remainder
of the lophophore is metachromatic and clearly
cartilaginous20.
MINERALISATION OF INVERTEBRATE
CARTILAGES
Box 4.3 contains an overview of the mineralisation of
vertebrate cartilage. Chapter 22 contains a discussion
70 PART | II Origins and Types of Skeletal Tissues

BOX 4.3 Cartilage Mineralisation

Interest in mineralisation of cartilage and bone goes back to
Honor Fells pioneering studies 80 years ago. Because minera- (Figure 4.4), where its distribution coincides with mineralisa-
lisation is so regular, dynamic and precisely controlled tion. The interface between nonmineralised and mineralised
and can be approached from so many different levels and cartilage, which is known as the tidemark, is illustrated in
backgrounds crystallography, mineralogy, endocrinology, Figure 3.11. Tidemarks occur in hyaline and fibrocartilage if a
cell biology, pharmacology, pathology the literature is junction exists between mineralised and nonmineralised carti-
extensive. Throughout the book I touch on mineralisation as a lage (Figure 3.11, and see Havelka et al., 1984, for a review).b
developmental process. Here, I survey mineralisation using Osteonectin (SPARC secreted protein acidic and rich in
the mineralisation of cartilage to highlight some important cysteine) is a 32,000 MW protein that makes up around one-
featuresa. tenth of bone protein. Linking collagen to hydroxyapatite and
acting as a nucleus for mineralisation, osteonectin is present
The Players
in osteoblasts, osteoprogenitor cells, young but not old osteo-
Robin Poole from Montreal contributed greatly to our basic
cytes, chondrocytes of mineralising cartilage, fracture callus
understanding of cartilage mineralisation and its relationship
and ectopic ossifications such as myositis ossificans (illustrated
to growth of long bones at growth plates. Poole et al. (1989)
in Figure 1.6). Osteonectin is expressed in hypertrophic chon-
summarised and reviewed the roles of alkaline phosphatase,
droblasts, in mineralised ECM, and in the perichondrium of
collagen, chondrocalcin and proteoglycans in cartilage miner-
rat mandibular condylar cartilage. Osteonectin is present in
alisation. Here, I outline calmodulin, chondrocalcin and
the cells of all zones of chicken tibial growth cartilage, but
osteonectin.
only in the matrix in the mineralising zoneb.
Calmodulin is so named because this abundant cyto-
The role of proteoglycans in the mineralisation of cartilage
plasmic constituent is a calcium-modulated protein. It func-
is complex, little understood, and cartilage- or taxon-specific.
tions by sensing Ca21 levels and then signalling calcium-
Acridine orange preserves proteoglycans, allowing their visuali-
sensitive enzymes or calcium-sensitive ion channels.
sation in rosettes in the longitudinal septa of the rat growth plate
Calmodulin is present in chondrocytes (detected in mature
at sites where mineralisation occurs. Proteoglycans, monomers
and mineralising chondrocytes of rodent primary and condy-
and link protein also are retained during the mineralisation of
lar cartilages), but not in perichondrial or periosteal cells
mammalian and avian epiphyseal growth plates (Figure 4.7)c.
(Lewinson and Boskey, 1984).
Chondrocalcin (the c-propeptide of type II procollagen) is
a calcium-binding extracellular protein. The c-propeptide
therefore has two distinct roles assembling the triple helix
of type II collagen, and binding calcium in mineralising car-
tilage (A. R. Poole et al., 1984). Although not present in
the proliferative layers of bovine fetal epiphyseal or growth
plate cartilages, chondrocalcin is present in the hypertrophic
zone and in the longitudinal septa of the extracellular matrix

100 m 100 m
A B

A B C D
FIGURE 4.5 The organisation of growth plates. Proximal tibial
FIGURE 4.4 The relative amounts of longitudinal and transverse growth plates from 16- (A) and 25- (B) day-old mice to show organi-
septa (trabeculae) depicted by the thickness of the lines in sation of the cell columns. The articular surface is at the top. P, pro-
normal bone (A), under conditions of rapid growth in length (B), in liferative, H, hypertrophic and R, resting cartilage cell zones. M,
arrested growth (C), and with atrophy (D). The lengths of the bone mineralisation front. Modified from Wikstrom et al. (1984).
segments in A D are proportional to each physiological situation.
From H. A. Harris (1933). (Continued )
 Chapter | 4 Invertebrate Cartilages, Notochordal Cartilage and Cartilage Origins 71
BOX 4.3 (Continued)
100 m
A B
C D
FIGURE 4.6 Reduced development of the proximal tibial growth
plates in Brachymorphic (Bm/Bm) mice at two ages after birth: 16-
day-old normal (A) and bm/bm (B) mice; 25-day-old normal (C)
and Bm/Bm (D) mice. Cartilage and mineralisation (black) is disor-
ganised in the growth plates from the Bm/Bm embryos. The entire
growth plate is reduced in height in mutant embryos of both ages
(all at same scale). P, proliferative, H, hypertrophic and R, resting
cartilage cell zones. M, mineralisation front. Modified from
Wikstrom et al. (1984).
FIGURE 4.7 Link protein. This longitudinal section of a portion of
the hypertrophic zone of a bovine fetal growth plate shows link pro-
tein (green in original) localised to the longitudinal septa (L) and to
spicules of mineralised cartilage matrix (S). Little link protein can be
visualised in the transverse septa. Cytoplasmic staining (orange in
original) reflects a counterstain, not link protein. Modified from A. R.
Poole et al. (1982a).
Cartilage Specificity
Mineralisation of cartilage varies in a cartilage-specific
manner. Mammalian small-celled cartilage normally does not
mineralise. Why? Because mineralisation is the prerogative
of hypertrophic chondrocytes. Secondary cartilages on avian
membrane bones mineralise, but Meckels cartilage (a small-
celled cartilage) does not. Mineralisation and alkaline phos-
phatase are virtual synonyms, but mineralisation of rat
tracheal cartilages is not associated with deposition of
alkaline phosphatase. Guinea pig mandibular condylar, tibial
epiphyseal and articular cartilages exhibit different patterns
of mineralisation in response to scurvy. The jaw cartilages of
(Continued )
72 PART | II Origins and Types of Skeletal Tissues

BOX 4.3 (Continued)

A B C

FIGURE 4.8 Mineralisation of cartilage in vitro. Mandibular mesenchyme removed from chick embryos of H.H. stage 21 (3.5 days of incuba-
tion), dissociated and established in micromass culture (10 L drops of 2 3 107 cells/mL) undergoes chondrogenesis and matrix mineralisation
(as seen with Kossa staining) when cultured in Hams F12:BGJb [3:1] 1 10% fetal calf serum 1 ascorbic acid (150 g/mL). (A) Cartilage nodules
(c) with initial signs of mineralisation (m) after 12 days of culture. (B) Increasing mineralisation (m) after 14 days in vitro; see also Figure 4.9.
(C) After 20 days in vitro, only the central cartilage (c) is unmineralised. Adapted from Ekanayake and Hall (*1994b).

A 40 m B 40 m

C 80 m D 80 m

FIGURE 4.9 Mineralisation of cartilage in vitro. Mandibular mesenchyme, removed from chick embryos of H.H. stage 21 (3.5 days of incuba-
tion), dissociated, and established in micromass culture (10 L drops of 2 3 107 cells/mL) undergoes chondrogenesis with matrix mineralisation.
Culture media were Hams F12:BGJb [3:1] 1 10% fetal calf serum (A); Hams F12:BGJb [3:1] 1 10% fetal calf serum 1 150 g/mL ascorbic
acid, and 1027 M dexamethasone (B D). Culture period was 12 or 14 days; see also Figure 4.8. Mineralisation visualised with the Kossa reac-
tion. (A and B) Histological sections after 12 days of culture. Cartilage (c) is present in both, mineralisation in neither (cf. Figure 4.8). Cultures in
test medium: 12-day (C) and 14-day (D). Onset of mineralisation can be seen in one region of the cartilage (c) after 12 days of culture with ascor-
bic acid and dexamethasone (C), and more extensive and heavier mineralisation after 14 days (D). Adapted from Ekanayake and Hall (*1994b).
(Continued )
 Chapter | 4 Invertebrate Cartilages, Notochordal Cartilage and Cartilage Origins 73
BOX 4.3 (Continued)
clam-crushing myliobatid stingrays are stiffened internally
with mineralised struts and rows of tesserae in a manner seen
only in bone in other vertebratesd. Clement (1992) analysed
the fine structure of mineralisation of elasmobranch skeletons
in 53 species; Summers et al. (2004) correlated CT-scanning
images with measures of stiffness in mineralised cartilages of
the California horn shark, Heterodontus francisci.
Mineralisation In Vitro
Many instances in which cartilage mineralises in vitro are
discussed throughout the text (Figures 4.8 and 4.9). For exam-
ple, in what B. Zimmermann et al. (1990) termed organoid
cultures, explanted mouse limb cells had become almost
completely mineralised after 21 days of culture in medium
containing 10-nM -glycerophosphate. Examination at the
ultrastructural level revealed that almost all chondrocytes in
the mineralised areas had died. In a later study, addition of
osteoblastic calvarial cells enhanced mineralisation, whereas
addition of fibroblasts decreased mineralisation, presumably
by the release of soluble molecules or the disruption of inter-
actions between mineralising cells; addition of dexametha-
sone induces these cells to chondrify (B. Zimmermann and
Cristea, 1993).
Vitamins
Vitamin D plays a critical role in the mineralisation of carti-
lage and bone, both of which are commented on here.
Vitamin D elevates plasma Ca21 and PO432 to levels high
enough for bone to mineralise. Of the two major hormones
that mediate calcium metabolism, parathyroid hormone is
vitamin D dependent and calcitonin is not. The first direct evi-
dence that vitamin D3 stimulates movement of calcium into
bone came from studies on calvariae from Japanese quail
maintained in vitro by Gunasekaran et al. (1986), who para-
doxically found that Ca21 uptake into mouse bone was
reduced by vitamin D3e.
of mineralisation in relation to chondrocyte hypertrophy
and the synthesis of type X collagen.
Although many invertebrates have cartilaginous skele-
tal matrices, no instance of in vivo mineralisation is
known, and no osseous tissue or bone has been found in
any invertebrate. Of interest, many invertebrate taxa
with unmineralised cartilages do have mineralised non-
cartilaginous endoskeletons, raising the question of why
the cartilages do not mineralise. Do they lack the ability
to mineralise in vivo, or are they inhibited from minera-
lising? Would they mineralise in vitro under conditions
discussed in Chapter 3 that permit lamprey cartilage to
mineralise?21
Mineralisation of cartilage in invertebrates is blocked
by temperature, perhaps coupled with the absence of
sufficient concentrations of calcium and phosphorus in
seawater. Hall (*1975a) discusses how the mineralisation
Vitamin D deficiency can result in an inability to mine-
ralise cartilage because hypertrophic chondrocytes fail to
mature, resulting in stunted growth that can lead to rickets in
children and osteomalacia in adults. Rickets and osteomalacia
are two terms for the same condition of softening of bones,
although the consequences differ the earlier the onset.
Collagen synthesis and calcium nucleation are both enhanced
when rachitic bone is exposed to vitamin D. Some of the
action of vitamin D on bone growth and mineralisation may
be indirect through enhancement of plasma Ca21 and PO432 .
Demineralised bone matrix implanted into vitamin
D deficient rats induces normal amounts of cartilage. Less
bone is produced than in controls, and the bone is minera-
lised abnormally unless vitamin D is addedf.
a. For reviews of the microstructure and mineralisation of skeletal tissues,
see J. G. Carter (1990), Francillon-Vieillot et al. (1990) and de Ricqle`s et al.
(1991).
b. See Termine et al. (1981), Stenner et al. (1986), Jundt et al. (1987) and
Nomura et al. (1988) for osteonectin; Copray et al. (1989) and Pacifici et al.
(1990) for expression in condylar cartilage and tibial growth cartilage. See
Figures 4.5 and 4.6 for proximal tibial growth plates in 16- and 25-day-old
mice.
c. See Shepard and Mitchell (1985) for acridine orange, and Davis et al.
(1982) and A. R. Poole et al. (1982c) for retention of proteoglycans during
mineralisation.
d. See Hall (1971b) and Sasano et al. (1993a) for mineralisation of avian
and tracheal cartilages; Irving and Durkin (1965) and Durkin et al. (1969a,
b) for guinea pig; and Summers (2000) and Summers et al. (1998) for mylio-
batid stingrays, which include such species as the cow nose ray,
Rhinoptera bonasus; eagle ray, Aetobatus narinari, and bat ray, Myliobatis
californica. Rhioptera bonasus also develops a fibrocartilage where the ven-
tral intermandibular tendon inserts onto the mineralised jaw cartilage
(Summers et al., 2003).
e. See G. L. Wong (*1984) for the skeletal actions of PTH, and see Pacifici
et al. (1980), Ornoy and Zusman (1983), Narbaitz (1992) and Boskey et al.
(2001) for vitamins and cartilage.
f. See Cousins and DeLuca (1972) and P. H. Stern (1980) for vitamin D
and bone, Underwood and DeLuca (1984) for plasma regulation, and
Van der Steenhoven et al. (1988) for bone implanted into vitamin
D deficient rats.
of invertebrate skeletal tissues relates to the evolution of
vertebrate mineralising skeletal tissues22. Preparations
of cartilage from Loligo, Limulus and Busycon can
mineralise in vitro if incubated in a solution
metastable to hydroxyapatite. Mineralisation occurs only
at a temperature (37 C) well above normal environmental
temperatures of 20 C. Mineralisation was as hydroxyapa-
tite, which is the normal form of calcium phosphate in
vertebrate cartilages. The initial abstract reports accumu-
lation of 18 g Ca21/g cartilage in Loligo. Ca21 accumula-
tion follows formation of perichondrial and matrix
granules and progressive intracellular mineralisation of
the chondrocytes (Figure 4.10). Such a pattern of intracel-
lular mineralisation, normally associated with cell death
in vertebrate chondrocytes (Hall, 1972a), is seen when
lamprey cartilage mineralises in vitro23. The rate of
mineralisation in vitro is correlated positively with the
74 PART | II Origins and Types of Skeletal Tissues
A B
FIGURE 4.10 Squid cartilage can mineralise. Head cartilage of
the longfin squid, Loligo pealii, after incubation for 32 (A) or 40 (B)
hours in a hydroxyapatite mineral solution at 37 C. (A) Phase-contrast
microscopy reveals accumulation of mineral (arrows) beneath the peri-
chondrium. (B) von Kossa staining with toluidine blue counterstain
shows accumulation of mineral granules (black) over the ECM.
Modified from Libbin et al. (1976), which contains details of the incuba-
tion solution.
concentration of phosphatidyl-serine within cartilage,
consistent with lipid involvement in mineralisation of
vertebrate cartilage and bone24.
CARTILAGE ORIGINS
Theories of the evolutionary origin(s) of cartilage abound.
Most start from the premise that cartilage is exclusively a
vertebrate tissue.
Berrill (1955) supported Garstangs earlier conclusion
that vertebrates arose from free-swimming larvae of
Precambrian tunicates. He proposed that cartilage evolved
as an extracellular secretion by tunicate connective tissue
cells to serve a mechanical function, and that bone formed
an outer, impermeable layer because of the tendency for
calcium phosphates to be excreted from the skin. In tele-
ost fish, in fact, calcium is excreted via the gills, and
phosphorus is transported in the intestine with little if any
skeletal involvement25.
Such a premise is either untenable or, if tenable, relates
to the origin of vertebrate cartilage, not to the origin of car-
tilage per se. Why? Because as we have just seen, cartilage
is present in many invertebrates. Consequently, several
important questions remain:
G When and how many times did cartilage evolve as a
tissue?
G What is the precursor of cartilage among the earliest
Metazoans?
G Does the evolution of vertebrate cartilage(s) relate in
any way to the evolution of invertebrate cartilage(s)?
G Can we separate the evolution of cartilage as a tissue
from the evolution of type II (cartilage type) collagen?
Cole and Hall (2004a,b) think we can; see below.
Hemichordates
Often taken as a surrogate vertebrate ancestor, amphioxus
(14 species of hemichordates) such as Branchiostoma lan-
ceolatum, Saccoglossus bromophenolosus, S. kowwalevskii
and Asymmetron spp. have a pharyngeal set of gill bars
that contain skeletal rods. These gill bars contain a fibril-
lar collagen and proteoglycans. Based on a phylogenetic
approach using the triple helical domain, Rychel et al.
(2006) demonstrated that the fibrillar collagen is similar
to both type I and type II vertebrate collagens. Presence
of collagen and proteoglycans in an extracellular matrix is
consistent with this being a cartilaginous tissue but the
gill bar skeleton is acellular the cells remain outside
the extracellular matrix they secrete and is derived
from the endoderm, two pivotal features that remove
these tissues from the vertebrate and invertebrate cartilage
families (Gillis et al., 2012b; Hall and Gillis, *2013).
However, Rychel et al. (2006) hypothesised that the
earliest gill cartilages were acellular, and these were grad-
ually replaced by neural crest cells in vertebrates (p. 543,
and see Chapter 17 for the neural crest origin of pharyn-
geal arch cartilages).
Similar key transcription factors (Pax1/9, Hox1, Eya,
Six1 and FoxC) are expressed in the pharyngeal gill arches
of Saccoglossus kowwalevskii and the pharyngeal arches
of vertebrates. A major difference is that expression that
is in derivatives of all three germ layers in vertebrates is
restricted to the pharyngeal endoderm in S. kowwalevskii,
which lacks Tbx1-expressing mesodermally derived mes-
enchyme in the pharyngeal arches (Gillis et al., 2012b).
The genetic basis for the patterning of the visceral arches
of vertebrates, in which derivatives of all three germ
layers are involved, arose as signalling from pharyngeal
endoderm, the homology of which is well substantiated.
The endodermal pharyngeal pouch supporting skeletal tis-
sue in hemichordates cannot readily be homologised as
cartilage with vertebrate or nonchordate invertebrate
cartilages (Gillis et al., 2012b; Hall and Gillis, *2013).
In addition, it has been proposed that cartilage or a
cartilage-like tissue is present in the tentacle-like oral cirri
around the oral hood of species of hemichordates. The pre-
cise nature of this tissue is unclear from early studies: the
presence of an ECM suggests cartilage. Although retention
of vacuoles in the cells is shared with notochord (see
below), the proposal that the tissue is endodermal in origin
(like the pharyngeal tissue) is not consistent with a rela-
tionship to notochord or to its classification as a cartilage.
Examination of the regeneration of these oral cirri in a
Japanese species Branchiostoma belcheri sheds consider-
able light on the nature of the skeletal supports. Kaneto and
Wada (2011) used expression of fibrillar collagen and SoxE
genes in the regenerating tissues to classify it as cartilage
and propose the following: an evolutionarily conserved
 Chapter | 4 Invertebrate Cartilages, Notochordal Cartilage and Cartilage Origins 75
Chondroid connective tissue/
Mesenchyme
Bone
Type I/V collagen +
Mucopolysaccharides
Cartilage
Type II
? collagen
Notochord
Independent
chondrocyte
specialisation? Invertebrate cartilages
genetic regulatory system is involved in amphioxus
cirrus and vertebrate skeletogenesis (p. 409). Even more
intriguing is the demonstration of known osteogenic mar-
kers, Runx and osteonectin, in the regenerated cartilage,
aligning this tissue with the intermediate tissues discussed in
Chapter 5. Fibrillar collagen and osteonectin genes also are
coexpressed in ascidian notochord and in mineralised bone.
An important generalisation is that molecular and genetic
data provide insights into mechanisms that phenotype that
morphology often cannot. The scenario proposed is that the
common chordate ancestor had a genetic regulatory system
from which both chondrogenesis and osteogenesis emerged
as separate differentiated cell types and tissues.
Living species of amphioxus (hemichordates) clearly
are not vertebrate ancestors and may not even be good can-
didates as surrogate vertebrate ancestors. Fossils, as illus-
trated by amazing fossil finds over the past two decades
and by enhanced methods of palaeohistology, will increas-
ingly come into play in our discussions of the origin(s) of
vertebrate skeletal tissues. Cathaymyrus diadexus and
Cathaymyrus haikoensis have been identified as cephalo-
chordates on the basis of the presence of a notochord and
segmented body muscles. Equally old, however, are two
species of hemichordates, Haikouella lanceolata (over 300
individuals) and Haikouella jianshanensis, from the Early
Cambrian Chengjiang formation in China (530 mya) that
had gills, a tail, six visceral arches presumed to be carti-
laginous and pharyngeal teeth, which may represent the
earliest mineralised tissues26.
No invertebrate has bone, although some invertebrate
cartilages contain such bone-cell markers as osteonectin
FIGURE 4.11 A summary proposal of
the evolutionary relationships among chon-
droid, notochord, cartilage and bone in
relation to invertebrate and vertebrate carti-
lages, presence of type II collagen in noto-
chord and vertebrate cartilages but not in
invertebrate cartilages, and the presence of
several lineages of chondrocyte differentia-
tion among invertebrates. (Reproduced
from Cole, *2011, with permission of the
author.)
Vertebrate
mRNA (Ringuette et al., 1991), a finding that raises fasci-
nating issues: Do invertebrates do with cartilage all that
vertebrates do with cartilage and bone? Cartilaginous
fishes certainly do with cartilage what teleosts and tetra-
pods do with bone. Whatever properties of invertebrate
cartilages brings them into the cartilage family perhaps
superfamily would be a more appropriate term, depending
on issues of homology we must project the origin(s)
of cartilage much further back than the invertebrate
chordates such as hemichordates or tunicates.
We have seen the enormous differences among
lamprey, hagfish, gnathostome and invertebrate cartilages.
Indeed, because extant vertebrates are subdivided into the
three groups (lampreys, hagfishes and gnathostomes),
invertebrates should be subdivided into groups (cephalo-
pods, brachiopods and annelids) to reflect the diversity of
invertebrate cartilaginous tissues. Figure 4.11 shows sev-
eral independent lines of chondrocyte origins in inverte-
brates. Extant agnathans and invertebrates have cartilage
as a tissue but not type II collagen as a matrix product.
Type II collagen, a sine qua non for the identification of
cartilage in jawed vertebrates, was not associated with ini-
tial evolution of cartilage as an invertebrate tissue, shown
in Figure 4.11. Consequently, we must widen our defini-
tion and concepts of what constitutes cartilage. One set of
data causing us to reinvestigate the definition of cartilage
outlined below is the differentiation of cartilage by
notochordal cells. A second set of data, outlined in
Chapter 16, relates to the first steps in formation of verte-
brae in teleost fishes in which mineralisation of the noto-
chord sheath initiates vertebral development through
76 PART | II Origins and Types of Skeletal Tissues
osteogenesis, not chondrogenesis. A third set of data on
similarities between notochord and cartilage, discussed in
Box 42.2, leads us to ask whether cartilage is a form of
notochord or notochord a form of cartilage, or whether
both are forms of a larger set of connective tissues.
NOTOCHORDAL CARTILAGE
Notochord is an unusual tissue with large vacuolated
cells connected by desmosomes and surrounded by an
expanded basement membrane as a perinotochordal
sheath. Notochordal cells synthesise and deposit type II
collagen, which we think of as cartilage-type collagen.
We may have this backward, however. Type II collagen
is notochordal collagen, a notochordal molecule that was
taken over by cartilage and elaborated into an ECM when
vertebrate cartilage originated27.
Even more intriguing issues are whether notochord
should be included in the cartilage superfamily; how
notochord as a tissue relates to connective tissues; and how
it relates to tissues such as chondroid (Figure 4.11) or myx-
oid, typically regarded as intermediates between two other
tissue types (Chapters 1, 5 and 42). The most parsimonious
relationship between these tissues is outlined in
Figure 4.11, in which notochord and vertebrate cartilage
share expression of type II collagen. As discussed in this
chapter, the origin of type I collagen was much older, pre-
dating any of these skeletal tissues (Figure 4.11).
It is well known that the notochord induces sclerotomal
mesenchyme around it to chondrify as vertebral cartilage
(Chapter 42). Metachromatic, glycosaminoglycan-rich
matrix accumulates around the notochord before chondro-
genesis of the adjacent mesenchyme. This sequence was
interpreted as indicating that these notochordal matrix
products progressively become the matrix of the vertebral
cartilage28.
More intriguing, cartilage can arise from the noto-
chord itself. Although not common, notochordal chon-
drogenesis is part of normal vertebral development in
some taxa and is a regenerative response in at least one
taxon of teleost fishes.
D. B. Wake and Lawson (1973) and Lawson (1966)
described the formation of notochordal (intravertebral)
cartilage in the centra of a salamander, Eurycea bislinea-
ta, and in the Frigate Island caecilian, Hypogeophis ros-
tratus. In both species, the cartilage developed from cells
of the notochordal epithelial sheath and not from invad-
ing mesenchymal cells derived from the sclerotome.
Absence of a distinct sclerotome but plentiful notochordal
cartilage also characterises vertebral development in the
northern two-lined lungless salamander, Eurycea bislineata
(D. B. Wake and Lawson, 1973). As Lawson (1966)
emphasised, urodele centra ossify from connective tissue,
whereas anuran centra ossify in cartilage. A more recent
study of notochordal cartilage (chordoid and chondroid)
in the notochords of geckos by Jonasson et al. (2012) is
discussed in Box 42.2 in the context of whether notochord
should be considered a type of cartilage29.
During regeneration after amputation of the tail of the
glass knifefish, Eigenmannia virescens, cartilage develops
at the end of the severed notochord. The chondrocytes do
not hypertrophy, and the matrix does not mineralise.
Cartilage on the ventral surface is removed by multinucle-
ated cells and replaced by perichondral bone (Kirschbaum
and Meunier, *1988).
A further fascinating example, which comes to us
from the fossil record, illustrates the advantages to be had
when an informed palaeontologist incorporates a develop-
mental approach in his/her analysis. The example is a
group of extinct, limbless, enormously elongate reptiles,
the aystopods, with many hundreds of vertebrae. The
challenges in laying down so many vertebrae without
greatly extending ontogeny are considerable. R. L. Carroll
(1989) identified two patterns of vertebral development
in aystopods and in Palaeozoic tetrapods, both of
which reveal unexpected involvement of the notochord.
Vertebrae developed either from a perichondral tube
around the notochord or from a medial notochordal
fibrous sheath, a mechanism that allows more rapid ossifi-
cation than does perichondral ossification, and that is the
mode of origin of vertebral bodies in teleosts (Bensimon-
Britto et al., 2012b, and see Chapter 16).
A final example of extremes of vertebral development
comes from the prolacertiform reptile Tanystropheus
longobardicus from the Middle Triassic (Wild, 1973).
At least three-fourths of the total body length of these
6-m-long ancient reptiles was neck and tail (Figure 4.12).
The enormously elongate neck has been described as
the longest neck possible within the laws of physics.
A
1 cm
B 50 cm
FIGURE 4.12 The longest cervical vertebrae (A) may have belonged
to the middle Triassic reptile, Tanystropheus longobardicus (B). A cervi-
cal vertebra is shown to scale beside the neck in (B). Based on informa-
tion in Wild (1973) and Tschanz (1988).
 Chapter | 4 Invertebrate Cartilages, Notochordal Cartilage and Cartilage Origins 77
Amazingly, it consists of no more than 12 (often only 10)
enormously elongated cervical vertebrae (Figure 4.12). As
determined by Tschanz (1988) from an analysis of eight
individuals from Switzerland that comprise an ontogenetic
series, the largest individuals had cervical vertebrae as
long as 26 cm. Average length of cervical vertebrae 3 11
was 17.5 cm in the largest individual and 3.5 cm in the
smallest, a fivefold increase in size.
The enormous diversity of cartilages in invertebrates
and vertebrates demonstrates the lability of skeletogenic
cells, a lability that is further reinforced when we con-
sider, as we do in the next chapter, the range of intermedi-
ate tissues in vertebrates: tissues with features of two or
more of the four skeletogenic and odontogenic tissues
bone, cartilage, dentine and enamel.
NOTES
1. See W. A. Beresford (*1981) and Dhem et al. (1989) for analyses
of chondroid, its contribution to skeletal growth, and various uses of
the term.
2. See Person et al. (1959), Person (1960, *1983a), Person and Philpott
(1963, 1969a,b), Philpott and Person (1970) and Person in Slavkin
(1972, pp. 140 151, the evolution of cartilage section of a work-
shop held on Santa Catalina Island, CA).
3. See Jeanloz (1960) for a discussion of the nomenclature of the gly-
cosaminoglycans or acid mucopolysaccharides, and Anderson
(1976b), who surveyed the distribution and composition of glycosa-
minoglycans and glycoproteins.
4. See P. J. Morris (1993) for the origin of extracellular matrix, and
Exposito and Garrone (1990), Har-El and Tanzer (1993) and Garrone
(*1998) for collagen appearing with sponges. See Hall (*2003a),
Stone and Hall (2004, 2006) and Hall and Kerney (2012) for analyses
of such issues as deep homology and the complex relationships
between homology at various levels of the biological hierarchy.
5. M. B. Mathews (1968, 1975) and E. Adams (1978) review the compo-
sition of invertebrate collagens, Kobayashi (1971) and Finerty (1981)
the homology of collagens, and Garrone (*1998) the evolution of
metazoan collagens, whereas Cole and Hall (2004a,b) discuss the col-
lagens of invertebrate cartilages and the evolution of type II collagen.
6. See Sage and Gray (*1980) and the analysis by G. M. Wright et al.
(1988).
7. For the latest overviews of invertebrate cartilages see Hall (*1978a),
Cole and Hall (2004a,b, *2009) and Cole (*2011).
8. See Persons (1960) emphasis that invertebrate cartilages possess
chitin but lack chondroitin sulphate.
9. See Person (1960, *1983a) and Person and Philpott (1963) for
light and transmission microscope cytology/histology, Lash and
Whitehouse (1960) and Katzman and Jeanloz (1969) for polyglucose
and polyfucose sulphates, and Lash (1959), Person et al. (1959) and
Person and Philpott (1963) for myoglobin. See Cole and Hall
(2004a,b, *2009) for overviews of gastropod cartilages, Ponder and
Lindberg (1997) for a phylogeny of gastropod molluscs (117 charac-
ters, 40 taxa, mostly prosobranch), including the use of cartilage as
character, and see Guralnick and Smith (1999) and references
therein, for gastropod cartilages.
10. See Metscher (2009a,b) and Wise et al. (2013) for an analysis of
micro-CT methodology, 3D reconstruction, volume rendering and
colour enhancement, illustrated with superb images of unmineralised
embryos and fetal rabbit and rat skeletons.
11. T. Sugahara et al. (1996), who did this study, use the common name
king crab, an older name for the horseshoe crab, but it is clear
that they were examining Limulus. The common name king crab
(red, blue or golden king crab) is more commonly used for the three
species of Alaska king crab, the red king crab Paralithodes
camtschaticus, golden king crab Lithodes aequispinus and the blue
king crab Paralithodes platypus.
12. Descriptions of phragmosomes differ, in part reflecting differences
in different plant taxa. The phragmosome (partition body) is vari-
ously described as the following: (i) a structure within plant cells
a flat membranous vesicle composed of cell-wall components,
microbodies of the cell plate that form after nuclear division, a
raft-like structure composed of actin filaments (see Box 5.1) and
microtubules formed immediately before mitosis at the preprophase
stage; (ii) a region within plant cells the cytoplasmic region con-
taining the nucleus during mitosis; or (iii) by function cytoplasmic
strands that aid the nucleus in its migrate to a central position before
mitosis. All these descriptions fit the membrane formed in dividing
gill-book chondrocytes.
13. See Person (1960), Person and Philpott (1963) and Cowden (1967)
for light and transmission microscopic cytology/histology; Mathews
et al. (1962), Cowden (1967) and Sugahara et al. (1996) for the gly-
cosaminoglycans; Person and Philpott (1969b) for collagen; Philpott
and Person (1970) for cytochrome oxidase; and Person and Philpott
(1963, 1969b) for the pattern of chondrification and its resemblance
to lignification.
14. See Cowden (1967) and Person (1969, *1983a) for cranial, scleral
and epidermal cartilages; Person (1960), Person and Philpott (1963)
and Philpott and Person (1970) for glycosaminoglycan composition;
and the chapters in Wight and Mecham (1987) for the biology of
proteoglycans.
15. See Kinoshita et al. (1997) for the novel sequences, and see Habuchi
et al. (1983) for the distinct glycopeptide.
16. See Hjerpe et al. (1983), Vynios and Tsiganos (*1990) for these
studies. Two mannose-6-phosphate receptors are expressed in carti-
lage during mouse development (Matzner et al., 1992).
17. See Ayad et al. (1994) and Knudson and Knudson (2001) for
detailed reviews of aggrecan and its functions, and see Yang et al.
(1998) for expression in cultured cells.
18. See Sivakumar and Chandrakasan (1998) for C1 C3 collagens, and
Rigo and Bairati (1998) for the rotary shadowing.
19. See Person and Mathews (1967) for structure and biochemistry, and
Cowden and Fitzharris (1975) for ultrastructure.
20. See Stricker and Reed (1985, and references therein) and Reed and
Cloney (1977) for these tissues and the proposals for and against
them as cartilaginous.
21. For descriptions of mineralised tissues in invertebrates, see Watabe
(1965), Travis et al. (1967) and J. G. Carter (1990).
22. See Libbin et al. (1976) and Rabinowitz et al. (1976) for lipid
and in vitro mineralisation of invertebrate cartilages; Irving
(1973), Schuster et al. (1975) and Rogers (2000) for lipids and
mineralisation of bone and vertebrate cartilages; and Roy et al.
(2004) for the influence of dietary phosphorus on bone
mineralisation.
78 PART | II Origins and Types of Skeletal Tissues
23. See Eilberg et al. (1975), Libbin et al. (1976) and Hall (1972a) for
intracellular mineralisation of invertebrate chondrocytes, and Langille
and Hall (1988a, 1993a) and Janvier and Arsenault (2002) for minera-
lisation of lamprey cartilage.
24. Genes that regulate lipid biosynthesis or transport can act dif-
ferentially on cartilage and on lipogenesis. The Gonzo (Goz) mutant
zebrafish, which results from a mutation in the gene Site1 protease
that regulates key enzymes for lipid synthesis and transport, acts
independently to produce cartilage and lipid defects (Scholmbs
et al., 2003). For genes and mutations in zebrafish see the website
http://www.Zfin.org.
25. See Romer (1942), Hall (*1975a, 1999a), Gans and Northcutt
(1983), Northcutt and Gans (1983), Hall and Horstadius (1988),
Smith and Hall (*1990, *1993), Janvier (1996), Northcutt (1996),
Cole and Hall (2004a,b, *2009), Cole (2011) and Hall and Gillis
(*2013) for theories of the origin of cartilage.
26. See Rahr (1981) and de Beer (1937) for gill and oral hood carti-
lages; J.-Y. Chen et al. (1999) for Haikouella lanceolata, Shu et al.
(2003a) for Haikouella jianshanensis; and Mallatt and Chen (2003)
for these species and craniate origins. Haikouichthys ercaicunensis
from the same Lower Cambrian formation in China, with a noto-
chord, segmented muscles, a dorsal fin and cartilaginous hill bars
and olfactory and optic cartilages, was interpreted as a stem group
craniate (an early jawless vertebrate) by Shu et al. (2003b). See
J. A. Long et al. (2010) for the proposal that the evolution of carti-
laginous optic capsules preceded the origin of jaws.
27. See Ekanayake and Hall (1991) for the perinotochordal sheath, and
see Hall (*1998a) and Cole and Hall (2004a,b, *2009) for noto-
chord cartilage affinities. My thanks to Eckhard Witten for com-
ments on this portion of the text.
28. See Frederickson and Low (1971), Strudel (1971) and Corsin (1974)
for discussions of studies on the notochordal extracellular matrix.
29. See R. L. Carroll et al. (1999) and Hall (2003b) for overviews of
evolution and development, respectively, and see R. L. Carroll
(1997), Hall (*1998a) and Hall and Olson (2003) for books that inte-
grate both approaches.