DYNAMICS OF THE PULPCHDENTIN COMPLEX

D.H. Pashley
Department of Oral Biology, School of Dentistry, Medical College of Georgia, Augusta, Georgia 30912-1129, USA

ABSTRACT: Dentin has a relatively high water content due to its tubular structure. Once dentin is exposed, this intratubular
water is free to move in response to thermal, osmotic, evaporative, or tactile stimuli. Fluid shifts across dentin are thought to
cause sufficient shear forces on odontoblasts, nerve endings, nearby fibroblasts, and blood vessels to cause significant
mechanical irritation, disruption, or damage, depending on the magnitude of the fluid shift. Even in the absence of fluid shifts,
the water-filled tubules provide diffusion channels for noxious (i.e., bacterial products) substances which diffuse inward toward
the pulp, where they can activate the immune system, provide chemotactic stimuli, cytokine production, and produce pain and
pulpal inflammation. Viewed from this perspective, dentin is a poor barrier to external irritants.
However, pulpal tissues react to these challenges by increasing the activity of nerves, blood vessels, the immune system,
and interstitial fluid turnover, to make the exposed dentin less permeable either physiologically, via increased outward fluid
flow, or microscopically, by lining tubules with proteins, mineral deposits, or tertiary dentin, thereby enhancing the barrier
properties of dentin, and providing additional protection to pulpal tissues. These reactions involve dentin and pulp, both in
the initiation of the processes and in their resolution. These responses of the dental pulp to irritation of dentin demonstrate
the dynamic nature of the pulpo-dentin complex.

Keywords. Dentin, pulp, collagen, odontoblasts, nerves, pulpal blood flow, inflammation, dentin permeability.

Introduction modified for therapy, changes in pulpal innervation in

he purpose of this review is to integrate a great deal
T of recent information from a wide variety of fields on
how the pulpo-dentin complex responds during non-
physiologic conditions. Indeed, it is during pathologic
stresses that the full range of biologic responses of the
pulp to changes that occur in dentin can be expressed.
These are the conditions with which most dental clini-
cians must contend. It is hoped that insights gained by
the study of pulpal pathobiology will provide new thera-
pies that will be of use to dentistry in the future. Previous
reviews of the biology and pathobiology of the pulpo-
dentin complex were published in 1985 and 1992 (see
Bergenholtz et al, 1985; Pashley et al, 1992a). This review
will not cover the biochemistry of dentin or the pulp (see
Butler, 1995), the molecular biology of tooth develop-
ment (seeThesleff etal, 1995), dentinogenesis (see Linde
and Goldberg, 1993; Linde and Lundgren, 1995), odonto-
blast differentiation (Ruch et al, 1995), dental innervation
(Byers, 1984; Hildebrand et al, 1995), or dentin perme-
ability (Pashley, 1990). These areas have all been recent-
ly reviewed in depth and will not be revisited. Rather, this
review will cover the physiology of dentin, the mechani-
cal properties of dentin, how dentin structure can be
response to exposure of dentin, dentin sensitivity, the
inter-relationships among pulpal nerves, dentin, and
pulpal inflammation, and the reactions of the pulp to
wound healing. These areas will not be reviewed exhaus-
tively. This review attempts to integrate broad areas in
general, instead of a specific topic in depth, and will refer
frequently to more detailed reviews such as those noted
above.
(I) The Concept of the Pulpo-Dentin Complex
The physiologic concept of the "pulp-dentin complex"
has recently been challenged as an oversimplication
(Goldberg and Lasfargues, 1995). These authors correct-
ly point out that there are a number of differences
between the chemistry of dentin and that of pulp. For
instance, the collagen matrix of dentin is composed pri-
marily of type I collagen, while that of the pulp contains
collagen types I, III, and V, among others. Type III colla-
gen is associated with three-dimensional fibrillar net-
works, while type IV collagen is found in basement mem-
branes. Since peripheral dentin has no basement mem-
brane, it is not surprising to find that it has no type IV
collagen. The distribution of noncollagenous proteins is

104 Crit Rev Oral Biol Med 7(2):104-133 (1996)

also different. Phosphophoryn is found in mineralized
dentin and seems to modulate mineralization, but it is
not located in non-mineralizing pulpal soft tissue
(Butler, 1995; Linde and Lundgren, 1995). These differ-
ences led Linde and Goldberg (1993) to conclude that
although pulp and dentin tissues share a common
ancestry, the dental papilla, and although the develop-
ment of the two is closely interrelated, there are no direct
chemical similarities between the pulp and dentin.
Goldberg and Lasfargues (1995) prefer to think of the
"odontoblast-dentin complex" and the pulp as two dis-
tinct tissues. However, this is not very helpful conceptu-
ally. While bone matrix is different from osteoblasts and
fingernails are different from nail beds, these units func-
tion as complexes. The living portion creates and main-
tains the matrix, which, in turn, protects the delicate
cells from injury.
There is a great deal of evidence that dentin and
pulp are functionally coupled and hence are integrated
as a tissue. When normal intact teeth are stimulated
thermally, dentinal fluid expands or contracts faster than
does the volume of the tubules that contain the fluid.
This causes hydrodynamic activation of intradental
nerves and is an example of the functional coupling of
pulp and dentin in an intact tooth. As soon as the exter-
nal tissues that seal dentin peripherally (e.g., enamel and
cementum) are lost for any reason, the normal compart-
mentalization between the two tissues is lost, and they
become functionally continuous. Under these patholog-
ic conditions, there is a fluid-filled continuum from the
dentin surface to the pulp. It is through this medium that
external stimuli are transduced to physical disturbances
in the pulp. It is through this fluid medium that bacteri-
al substances can diffuse across dentin to produce pul-
pal reactions (Bergenholtz, 1977, 1981, 1996; Bergenholtz
and Warfvinge, 1982). The pulp responds to these insults
in the short-term by mounting an inflammatory response
which produces an outward movement of fluid
(Vongsavan and Matthews, 1991, 1992a,b, 1994;
Matthews, 1996) and macromolecules (Raab, 1989; Maita
etal, 1991; Byers, 1996). In the long term, pulpal tissues
produce tertiary dentin as a biologic response to reduce
the permeability of the pulpo-dentin complex and to
restore it to its original sequestration. Radioactive tracer
experiments clearly demonstrate the continuity of the
dentinal fluid-pulpal fluid-circulation in exposed dentin
and the importance of pulpal blood flow in clearing pul-
pal interstitial fluids of exogenous material (Pashley,
1979, 1985; Potts et al, 1985). In summary, there is abun-
dant evidence that dentin and pulp function as an inte-
grated unit which is generally termed the pulpo-dentin
complex. While dentin or pulp can be discussed sepa-
rately for purposes of instruction or to simplify issues, it
should be stressed that the two function together as a
unit.
7(2): 104-133 (1996) Crit Rev Oral Bio! Ued
(II) Dentin
DENTIN MORPHOLOGY
Dentin can be regarded as a porous biologic composite
made up of apatite crystal filler particles in a collagen
matrix (Fig. 1). This mineralized matrix was formed devel-
opmentally by odontoblasts which began secreting col-
lagen at the DEJ and then grew centripitally while trailing
odontoblast processes. The presence of these cellular
processes makes primary and secondary dentin tubular
in nature. Since the circumference of the most peripher-
al part of the crown or root of a tooth is much larger than
the circumference of the final pulp chamber or root canal
space, the odontoblasts are forced closer together as
they continue to lay down dentin, ending up in a colum-
nar layer in parts of the coronal pulp, especially over
pulp horns (Couve, 1986). They are more cuboidal in the
root canal and become flat near the apex. The conver-
gence of dentinal tubules toward the pulp creates a
unique structural organization to dentin which has pro-
found functional consequences, as will be discussed
below. This convergence has been estimated to be 5:1
(HoppeandStiiben, 1965), 4:1 (Walton et a!., 1976), or 3:1
(Fosse et al, 1992) in coronal dentin.
Each individual dentinal tubule is an inverted cone
with the smallest dimensions at the DEJ and the largest
dimensions at the pulp. Originally, each tubule had a
diameter of nearly 3 |xm. However, within each tubule is
a collagen-poor, hypermineralized cuff of intertubular
dentin which is called "peritubular dentin" in many texts.
It is really periluminal dentin or, more accurately,
intratubular dentin (Linde and Goldberg, 1993). Its for-
mation narrows the lumen of the tubule from its original
3 |Jim to as little as 0.6-0.8 |xm in superficial dentin. This
larger amount of peritubular dentin in superficial dentin
Figure 1. Fractured dentin viewed by STEM, showing a single
dentinal tubule surrounded by hypermineralized peritubular
dentin. Notice the mineralized collagen fibers. 19,600X.
Courtesy Dr. J. David Eick, University of Missouri, Kansas City.
105

Figure 2. Schematic representation of prepared cavity showing
the tubule density and diameters in superficial coronal dentin,
deep coronal dentin, and deep radicular dentin. (Modified from
Trowbridge, 1982.)
near the DEI is due, in part, to the fact that it is "older"
than middle or deep dentin, which was more recently
secreted and mineralized. Thus, the width of intratubular
(peritubular) dentin decreases as tubules are followed
inward toward the pulp, with the exception that there is
no peritubular dentin in intraglobular dentin (Blake,
1958). Very close to the pulp, there is no intratubular or
peritubular dentin, and the tubule (luminal) diameter is
about 3 |xm (Garberoglio and Brannstrom, 1976). Thus,
most of the narrowing of the tubule lumen as one
observes dentin more peripherally is due to deposition
of peritubular dentin (Frank and Nalbandian, 1989). The
composition of this dentin is different from that of inter-
tubular dentin, being collagen-poor and mineral-rich.
The mineral is in the form of small calcium-deficient car-
bonate-rich hydroxyapatite crystals which have a higher
crystallinity and is almost 5 times harder than intertubu-
lar dentin (Kinney et al, 1996). Little is known about the
biologic control of peritubular dentin apposition.
Presumably, it is a very slow process, slower than the
incremental formation of secondary dentin in the pulp
chamber. This process can be accelerated by occlusal
abrasion (Mendis and Darling, 1979) and other forms of
pulpal irritation and may be more rapid in deciduous
(Hirayama et al, 1985) than in permanent teeth. When
dentin is exposed by attrition, there is a net outward
movement of dentinal fluid for the first time. Although
the complete composition of dentinal fluid is unknown,
it presumably has an ion product of calcium and phos-
phate which is near or above the solubility product con-
stants for a number of forms of calcium phosphate. This
would tend to form mineral deposits in dentinal tubules
which can take on many forms (Mjor, 1985), since the
outward movement of dentinal fluid would present a
larger amount of mineral ions to the walls of the tubules
than could occur by diffusion in sealed tubules. This
principle has been used experimentally to slow the
106 Crit Rev Oral Biol Med
depth of demineralization of dentin in vivo under simu-
lated caries-forming conditions (Shellis, 1994).
(A) The extent of the odontoblast process
Controversy continues regarding the extent of the odon-
toblast process. Developmentally, odontoblast process-
es do indeed extend from the odontoblast cell body,
through mineralized dentin matrix, to the dentin-enamel
junction at the bell stage of tooth development.
However, as dentin thickens, the cellular processes must
elongate. Since there are no supporting cells or blood
vessels, the ability of the odontoblast cell body to sup-
port a long cytoplasmic process is in question. In human
teeth, the thickness of dentin is about 3-3.5 mm or 3000-
3500 |xm. If the diameter of the process is (on average)
1.0 |xm, then the volume of such a process would be (v =
wr2 x length, or 3.14 [0.5 |jim|2 x 3000 |xm = 2356 |xm3).
The cell body of a columnar odontoblast is about 5 |xm
in diameter and 30 |xm long, or a volume of 589 (Jim3.
Thus, the volume of the cellular process would be four-
fold larger than that of its cell body. This difference is
even larger in cuboidal or flattened odontoblasts. It is
not so much the disparity between the volumes of the
cell body and the cell process that matters, but rather the
diffusion distances involved. An odontoblast process at
the DEJ or CD) would be 3 mm or 3000 |xm from the near-
est capillary, which is a very long diffusion distance.
Although nerve axons are that long or longer, they are
not 3 mm from supporting cells or capillaries. In the
absence of any evidence for cytoplasmic streaming anal-
ogous to axon transport, it sems unlikely that the odon-
toblast could maintain such a long process. The length of
the odontoblast processes of most odontoblasts, regard-
less of dentin thickness, is between 0.1 and 1.0 mm
(Holland, 1976; Byers and Sugaya, 1995). The extent
varies somewhat with species and position within the
tooth (see Holland, 1976, 1985). The extent of the odon-
toblast process is important for a number of reasons. If
odontoblasts participate directly in the mechanism of
dentin sensitivity to surface stimuli, then these stimuli
must interact with a cytoplasmic process. If the process
does not extend to the surface, then the odontoblast can
not be directly involved in dentin sensitivity. When cavi-
ties are prepared in deep dentin in restorative proce-
dures, the odontoblast processes are amputated, there-
by irritating the cell body residing in the pulp. The fact
that many odontoblast processes are found to extend
only about one-third the distance from the tubule has
been explained as being due to a retraction of the
process during extraction (LaFleche et al, 1985). Others
have shown that the process does not extend more than
one-third the length of the tubule under normal condi-
tions (Holland, 1976; Thomas, 1985; Weber and Zaki,
1986) and that LeFleche et al (1985) may have been con-
fused by fixation and tissue preparation artefacts.
7(2):104-133(1996)
 (B) Changes in dentin structure with depth
The area occupied by the lumina of dentinal tubules can
be calculated as the product of the cross-sectional area
of a single tubule, irr2, and N, the number of tubules/cm2.
The term "r" is the radius of the tubule. Since both the
radius of dentinal tubules and their number per unit area
increase (Mjor and Fejerskov, 1979; Fosse et al, 1992;
Olsson and 0ilo, 1993; Dourda et al, 1994) as one exam-
ines dentin from the DE) to the pulp, the area occupied
by tubule lumina also increases. Garberoglio and
Brannstrom (1976) measured the tubule radius and num-
ber very carefully. They were the only authors to correct
for shrinkage artifact. Using their experimental data,
Pashley (1984) calculated the area occupied by tubule
lumina at the DEJ to be approximately 1% of the total
surface area at the DE] and 22% at the pulp (Fig. 2) (Table
1). Since this area is occupied by dentinal fluid, which is
95% water, these areas are also approximately equal to
the tubular water content of these areas. That is, the
water content of dentin near the DEI is about 1% (volume
percent), while that of dentin near the pulp is about 22%.
Textbooks list the water content of dentin at approxi-
mately 10% by weight or 20% by volume (LeGeros, 1991),
but that is an average value. It is clear that the water con-
tent or wetness of dentin is not uniform, but varies 20-
fold from superficial to deep dentin.
The high water content of deep dentin is responsi-
ble, in part, for the difficulty in bonding to deep dentin
(Prati and Pashley, 1992), where water competes with
resin monomers for the surfaces of collagen fibrils (see
reviews by Pashley and Carvalho, 1997; Eick et al, 1996).
The true water content of various dentin regions can
be masked by the presence of a smear layer and smear
plugs (Pashley, 1984; Prati et al, 1991). Smear plugs are
composed of grinding debris whose particle sizes are
smaller than tubule orifices (Figs. 3A,B). They physically
occupy a significant fraction of the luminal area that
could be occupied by water in their absence (Pashley et
al, 1978). It has been calculated that the presence of
smear plugs and the smear layer occupy 78.5% of the
area that would normally be occupied by water (Pashley,
1984). This means that these structures decrease the
water content of those surfaces by 78.5%. The presence of
grinding debris in the tubule orifices and on the dentin
surface also lowers the permeability of dentin (Pashley et
al, 1978).
(C) Physical characteristics of dentin
Dentin has been described as a heterogeneous compos-
ite material that contains micrometer-diameter tubules
surrounded by highly mineralized (ca. 95 vol% mineral
phase) peritubular dentin embedded within a partially
mineralized (ca. 30 vol% mineral phase) collagen matrix
(intertubular dentin) (Marshall, 1993; Marshall et al,
1996).
The bulk of tooth structure is made up of dentin,
which is the vital part of the tooth. Dentin is much soft-
er (Knoop hardness number, KHN = 68) than enamel
(KHN = 343; Craig, 1993) and exhibits much faster wear.

TABLE 1
Dentinal Tubule Density and Diameters at Various Distances from the Pulp a n d the Calculated
Areas of Fluid-filled Tubules, Peritubular Dentin, and in Intertubular Dentin

Areas*
Number of Radius of
Distance Tubules* Tubules* Fluid-filled Peritubular Intertubular
from Pulp x 10 6 cm- 2 xl04cm-4 Tubules Dentin* Dentin

Pulp 4.5 1.25 22.10 66.25t 11.65

0.1-0.5 4.3 0.95 12.19 36.58 51.23
0.6-1.0 3.8 0.80 7.64 22.92 69.44
1.1-1.5 3.5 0.60 3.96 11.89 84.15
1.6-2.0 3.0 0.55 2.85 8.55 88.60
2.1-2.5 2.3 0.45 1.46 4.39 94.15
2.6-3.0 2.0 0.40 1.01 3.01 95.98
3.1-3.5 1.9 0.40 0.96 2.86 96.18

*Data calculated from Garberoglio and Brannstrom (1976).

^ = llr2 N (100), where N = number of tubules per cm2. (Note: A,, although an area, is also the percent of surface area occupied by water.)
A p = IIN(R2 - r2) (100), where R = 2r and r = tubule radius. A; = 100 - (Ap - y\).
*There is no peritubular dentin at the pulpal surface, but it begins close to the pulpal surface.

7(2): 104-133 (1996) Crit Rev Oral Biol Med 107

TABLE 2
Summary of the Ultimate Yield Strength and Modulus of Elasticity of Dentin and Mineralized vs.
Demineralized Dentin

1. Mineralized Strength E Reference

Enamel 10.3 ± 2.6(10) - Bowen & Rodriguez (1962)
- 84.1 Craig (1993) (tensile)
Dentin (human) 51.7± 10.3(9) 19.3 ±5.4 Bowen & Rodriguez (1962)
Dentin (human) 3 6 . 6 ± 1 0 . 9 (12) 11.0 ±5.8 Lehman (1967) (tensile)
Dentin (human) 7 8 . 4 ± 13.3(12) - Watanabe etal. (1996) (shear)
Dentin (human) 93.8 ± 11.1 (6) - Sanoefa/. (1994) (tensile)
Dentin (bovine) 90.6 ±18.9 (8) 14.7 ±5.9 Sanoefa/. (1994) (tensile)
Enamel (human) 93.2 ± 13.9 - Smith & Cooper (1971) (shear)
Dentin, at DEJ 138.3 - Smith & Cooper (1971) (shear)
Dentin, Superficial 131.5 - Smith & Cooper (1971) (shear)
Dentin, Middle 90.3 - Smith & Cooper (1971) (shear)
Dentin, Inner 78.5 - Smith & Cooper (1971) (shear)
Dentin, Deep 49.1 - Smith & Cooper (1971) (shear)
Dentin, Near Pulp 45.1 - Smith & Cooper (1971) (shear)

II. Demineralized
Dentin (bovine) 28.0 ± 3.9(5) - Akimoto (1991) (tensile)
Dentin (bovine) 26.0 ±11.0 (10) 0.26 ±0.12 Sanoefa/. (1994) (tensile)
Dentin (human) 29.6 ± 5.9(10) 0.25 ± 0.07 Sanoefa/. (1994) (tensile)

III. Resin-infiltrated demineralized1 dentin (RIDD)

AB2 121.6 ±20.3 (6) 3.6 ± 0.8 Sanoefa/. (1995) (tensile)
SC-MP 111.5± 14.5(10) 3.1 ±0.7 Sano etol. (1995) (tensile)
SB 117.6 ± 12.2(6) 4.1 ±0.8 Sanoefa/. (1995) (tensile)
CLB 102.6 ± 3.7(6) 2.3 ± 0.3 Sanoefa/. (1995) (tensile)
PB 57.6 ± 16.4(6) 2.1 ±0.5 Sanoefa/. (1995) (tensile)

Abbreviations: AB2, All Bond 2, Bisco, Itasca, IL; SC-MP, Scotchbond Multi-Purpose, 3M Dental Products, St. Paul, MN; SB,
Superbond C&B, Sun Medical Co., Ltd., Kyoto, Japan; CLB, Clearfil Liner Bond 2, Kuraray Co., Ltd., Osaka, Japan; PB, Clearfil
Photobond, Kuraray Co., Ltd., Osaka, Japan. Number in parentheses is the number of specimens tested.

The modulus of elasticity of enamel (Table 1) is about 84
GPa (Craig, 1993) compared with dentin, which has a
modulus of about 13-17 GPa (Table 2). Due to its more
elastic nature, dentin is tougher than enamel and serves
as a stress-breaker or shock absorber for the overlying
enamel.
Regional differences in the shear strength of dentin
were reported by Smith and Cooper (1971). They made
thin ground sections of teeth and then, using 100-juim-
diameter punches, measured the shear strength of enam-
el, the dentino-enamel junction (DEJ), and dentin from
the DE] to the pulp. Their results are shown in Table 2.
The shear strengths of these very small dentin specimens
varied from about 132 MPa in superficial dentin to 45
MPa near the pulp. When a different technique was used,
Watanabe et al. (1996), using much larger specimens,
reported shear strengths of middle dentin between 72.4
and 86.9 MPa, which is intermediate between the
extremes reported by Smith and Cooper (1971).
Early measurements of the ultimate tensile stress
(UTS) and modulus of elasticity (E) of normal human
dentin by Bowen and Rodriguez (1962) gave values of 52
MPa and 19.3 GPa, respectively (Table 2). These were
somewhat higher than those reported by Lehman (1967)
of 36.6 MPa and 11.0 GPa, respectively. Recently, Sano et
al. (1994), using smaller specimens, obtained ultimate
tensile strengths and E for human dentin of 100 MPa and
14 GPa, respectively. In that study, they also measured
these same tensile properties in demineralized dentin,
and reported values of 26 MPa and 0.26 GPa, respective-
ly. This means that dentin collagen is responsible for 26%
of the ultimate tensile strength of dentin but only 1.6% of
its stiffness. The tensile strength of dentin collagen
reported by Sano et al. (1994) was similar for either
human or bovine demineralized dentin, which confirmed
an earlier report by Akimoto (1991) that demineralized
bovine dentin had a tensile strength of 28 MPa. Many
current dentin bonding systems use acidic conditioning

108 Crit Rev Oral Biol Med 7(2):104-133 (1996)

Figure 3. Smear plug creation begins when grinding debris enters open tubules (A) and is then condensed into a solid mass (B) covered
with a highly burnished smear layer (from Pashley et al., 1988; used with permission). Bracket in A = 0.5 |xm.

agents to demineralize the dentin surface and uncover for measurement of both nanohardness and the modu-
collagen fibrils. Adhesive resins are then applied to the lus of elasticity of intertubular and peritubular dentin
demineralized surface in an attempt to envelop the col- (Kinney et al, 1996). Peritubular dentin had a Knoop
lagen fibrils, which then provide micromechanical reten- hardness of 250, while that of intertubular dentin was
tion for the resins. If the resin-dentin bond is stressed to only 52 (KHN).
failure, presumably the weakest link in the adhesive
assembly will break first. Many investigators had thought (D) Dentin permeability
that the weakest link was the collagen fibrils. However, if It is the tubular structure of dentin which provides the
collagen fibrils can withstand a stress of 26-28 MPa, they channels for the permeation of solutes and solvents
may be stronger than many dentin bonds. It has been across dentin. The number of dentinal tubules per mm2
argued that since the cross-sectional area of demineral- varies from 15,000 at the DEI to 65,000 at the pulp
ized dentin that is occupied by collagen fibrils is only 1/3 (Garberoglio and Brannstrom, 1976; Fosse et al, 1992;
of the measured physical cross-sectional area, the true Dourda et al, 1994). Since both the density and diameter
modulus of elasticity and the yield stress of collagen of the tubules increase with dentin depth from the DEI,
should be multiplied by 3. This would give values of the permeability of dentin is lowest at the DEI and high-
about 80 MPa. The infiltration of adhesive resins may est at the pulp. However, at any depth, the permeability
reinforce the strength of the demineralized dentin to
make it even stronger. Sano et al. (1995) reported that
Intradentin permeability
resin-infiltrated demineralized dentin had an ultimate (resin infiltration)
tensile strength of between 60 and 120 MPa, depending
upon the bonding system, indicating that, in terms of
tensile strength, resins can increase the strength of dem-
ineralized dentin from 26 MPa to 120 MPa, which is not
statistically significantly different from the ultimate ten-
sile strength of mineralized dentin (Table 2). Sano et al.
(1995) proposed this as a model for the strength of the
resin-infiltrated hybrid layer that couples many adhesive
resins to dentin (Nakabayashi, 1992). When they mea-
sured the modulus of resin-infiltrated dentin, they found
that it increased from 0.26 GPa to about 3-5 GPa, which
is far below that of mineralized dentin, 13-17 GPa (Table
2).
Recently, modifications have been made to an atom- Figure 4. Schematic representation of an acid-etched dentin sur-
ic force microscope (AFM) by use of a stainless steel can- face which permits transdentinal fluid movement to occur via
tilever and a diamond stylus to permit an AFM to be used tubules and intradentin permeation into intertubular dentin

7(2):104-133 (1996) Crit Rev Oral Biol Med 109


Figure 5. Scanning electron micrograph of dentin following
acid-etching with 6% citric acid for 30 sec. The original smear
layer and smear plug have been dissolved. Remnants of the
original peritubular dentin matrix can be seen in the lumen.
Note the granular substructure of the peritubular dentin.
of dentin in vitro is far below what would be predicted by
the tubule density and diameters (Koutsi etal., 1994), due
to the presence of intratubular material such as collagen
fibrils, mineralized constrictions of the tubules, etc.
The details of methods used to measure dentin per-
meability can be found elsewhere (Pashley, 1990). Dentin
permeability can be divided into two broad categories:
(1) transdentinal movement of substances through
dentinal tubules (such as fluid shifts in response to
hydrodynamic stimuli), or (2) intradentinal movement of
exogenous substances into intertubular dentin, as
occurs during infiltration of hydrophilic adhesive resins
into demineralized dentin surfaces during resin bonding
(Fig. 4) or demineralization of intertubular dentin by bac-
terially derived acids (Kinney et a\., 1995).
The easiest method of measuring transdentinal per-
meability is to quantitate its hydraulic conductance. This
measures the ease with which fluid can filter across a
unit surface area of dentin in a unit time under a unit
pressure gradient (Pashley, 1990). The hydraulic conduc-
tance of dentin is responsive to a number of variables,
the most important of which is the fourth power of the
radius (Pashley, 1990). The presence of smear debris in
tubules (or any intratubular material) lowers the fluid
conductance of dentin and hence its permeability. The
hydraulic conductance of dentin increases as dentin
thickness decreases (Fogel et al, 1988; Koutsi et al, 1994)
in unobstructed dentin. However, the presence of a
smear layer is more important than dentin thickness in
reducing hydraulic conductance. When smear layers are
removed to facilitate dentin bonding (Fig. 5), the
hydraulic conductance increases, and the etched dentin
110 Crit Rev Oral Biol Med
Pulpal Concentration
of Permeating Substances
1. + Smear Layer 1. Perfusion Pressure
2. RDT 2. Tissue Pressure
3. Surface Area 3. Capillary Density
4. Location 4. Location
5. Molec. Size 5. AV Shunting
Figure 6. Balance between influx of substances into the pulp via
dentin permeation and their clearance by the pulpal circulation.
(From Pashley, 1994, with permission.)
tends to become wet with pulpal/dentinal fluid, making
it more difficult for resin monomers to infiltrate wet
dentin (Tao and Pashley, 1989). The severity of pulpal
reactions to restorative procedures is more severe when
smear layers are removed and dentin is made thinner
(Fujitani etal., 1996).
Dentin can be regarded as both a barrier or a perme-
able structure, depending upon its thickness, age, and
other variables (Pashley and Pashley, 1991). Due to its
tubular structure, dentin is very porous. The minimum
porosity of normal peripheral coronal dentin is about
15,000 tubules per mm2. Once uncovered by trauma or
tooth preparation, these tubules provide diffusion chan-
nels from the surface to the pulp. The rate of diffusional
flux of exogenous material across the dentin to the pulp
is highly dependent upon dentin thickness and upon the
hydraulic conductance of dentin (Pashley, 1985, 1990).
Thin dentin permits much more diffusional flux than
does thick dentin. However, there is a competition
between the inward diffusional flux of materials and the
rinsing action of outward convective fluid transport
(Pashley and Matthews, 1993). This may serve a protec-
tive role in mitigating the inward flux of potentially irri-
tating bacterial products into exposed, sensitive dentin.
The permeability of dentin is not uniform but varies
widely, especially on occlusal surfaces, where perhaps
only 30% of the tubules are in free communication with
the pulp (Pashley et al, 1987). Scanning electron micro-
scopic examination of acid-etched occlusal dentin
reveals that all the tubules are exposed, but functional
studies of the distribution of fluid movement across the
occlusal dentin reveal that the tubules that communi-
cate with the pulp are located over pulp horns and that
the central region is relatively impermeable. Apparently,
7(2):104-133 (1996)
there are intratubular materials such as collagen fibrils
(Dai et ai, 1991) and mineralized deposits that restrict
fluid movement, even though the peripheral and central
ends of the tubules are patent. Even microscopically,
within any 100 x 100 fim field, only a few tubules are
open (Hughes etal, 1996).
Axial dentin is much more permeable than occlusal
dentin (Richardson etal., 1991). Thegingival floor of prox-
imal boxes or the gingival extension of finish lines in
crown preparations often ends in regions of high dentin
permeability (Garberoglio, 1994). In a full-crown prepara-
tion on a posterior tooth with a surface area of approxi-
mately 1 cm2, there are as many as 3-4 million exposed
dentinal tubules (Pashley, 1991). Although the tubules
are occluded by smear layers and/or cement following
placement of castings, both cement and smear layers
have finite solubilities and may permit some tubules to
become exposed over time. This is most likely to occur at
the most peripheral extensions of restorations, where
diffusion distances to the external environment are closest.
The permeability of sclerotic dentin is very low
(Tagami et ai, 1992), regardless of whether the sclerosis
was due to caries or was physiologic or pathologic,
because the tubules become filled with mineral deposits.
Indeed, this is a fortuitous reaction that slows the caries
process and tends to protect the pulp. Most pulpal reac-
tions to cavity preparations or restorative materials used
on carious dentin are due to changes that occur across
adjacent normal dentin rather than the almost-imperme-
able caries-affected dentin.
(E) Balance between the permeation of noxious
substances across dentin and their clearance
by the pulpal circulation
There is a balance (Fig. 6) between the rate at which
materials permeate across exposed dentin to the pulp
and the rate at which they are cleared from interstitial
fluid by the pulpal microcirculation (Pashley, 1979).
Procedures which cause reductions in pulpal blood flow
(activation of pulp sympathetic nerves, administration of
vasoconstrictor agents with local anesthetics) upset this
balance, permitting higher interstitial fluid concentra-
tions of extrinsic material to exist than could occur at
normal levels of capillary flow (Pashley and Pashley,
1991).
Electrical stimulation of the inferior alveolar nerve in
cats causes an increase in pulpal blood flow and an
increase in outward dentinal fluid movement (Vongsavan
and Matthews, 1994). This response is interpreted as
being due to axon reflex activity. That is, antidromic stim-
ulation of the inferior alveolar nerve causes simultane-
ous depolarization of all of the nerve terminals providing
sensory innervation of the teeth. This release of neu-
ropeptides from the terminals causes both increased
pulpal blood flow and extravasation of plasma proteins
7(2)104-133 (1996) Crit Rev Oral Biol Med
from the microcirculation (Raab, 1989, 1992; Olgart and
Kerezoudis, 1994). Vongsavan and Matthews (1994) spec-
ulated that exposed dentin that was stimulated hydrody-
namically (for instance, mechanically) causes release of
neuropeptides, increased local blood flow, increased tis-
sue pressure, and increased outward fluid flow. This out-
ward fluid flow might rinse the tubules free of inward-dif-
fusing noxious substances by "solvent drag". Indeed,
Vongsavan and Matthews (1991) demonstrated that
Evans blue dye would not diffuse into exposed dentin in
vivo but would do so in vitro. They speculated that, in vivo,
the outward fluid flow blocked inward diffusion. This
notion was tested in vitro by quantitation of the decrease
in the inward flux of radioactive iodide when a simulated
outwardly directed 15 cm H2O pulpal pressure was
applied to dentin disks. This outward fluid movement
produced a 50-60% reduction in the inward diffusion of
radioactive iodide across acid-etched dentin, in vitro
(Pashley and Matthews, 1993). Thus, stimulated dentin
with open tubules (i.e., hypersensitive dentin) should
have higher rates of outward fluid flow and less inward
diffusion of bacterial toxins than nonstimulated dentin.
This protective effect is diminished in the presence of a
smear layer (Pashley and Matthews, 1993). Pashley
(1992) speculated that bacterial invasion of tubules or
formation of crystalline precipitates in tubules would
interfere more with outward fluid flow than with inward
diffusion of noxious materials, due to the higher sensi-
tivity of bulk fluid movement to changes in tubule radius,
r (which varies with r4), compared with diffusion (which
varies with r2). Under these conditions, the inward flux of
noxious substances may increase and permit higher
interstitial fluid concentrations to be achieved (Pashley,
1979, 1985) than would occur if there were more outward
fluid movement.
In exposed dentin, the outward fluid flow through
tubules is a first line of defense against the inward diffu-
sion of noxious substances. Dentinal fluid also contains
plasma proteins (albumin, globulins) which can bind or
agglutinate some materials which may also serve a pro-
tective role. A second defensive reaction occurs in fresh-
ly exposed dentin that causes the permeability of dentin
to fall following cavity preparation in vital dog teeth but
not in nonvital teeth (Pashley, 1985). Further, in experi-
ments in which there was a continuous inward-directed
bulk fluid movement, the rate of decrease in dentin per-
meability was delayed, suggesting that the reductions in
permeability were due to the slow outward movement of
some unidentified substance. These experiments were
repeated on dogs which were treated with an enzyme
which depleted fibrinogen from their blood. Under these
conditions, cavity preparation produced a much smaller
reduction in dentin permeability. Apparently, the out-
ward movement of fluid is associated with an outward
flux of plasma proteins (Maita et ai, 1991), including fib-
11

Figure 7. Scanning electron micrograph of acid-etched dentin
used to simulate exposed sensitive dentin. The surface was treat-
ed with 30% dipotassium oxalate for 2 min, then rinsed with
water. Most of the tubules are partially occluded beneath the
surface. Sharp angular crystals on the surface are calcium
oxalate dihydrate.
rinogen. Attempts to recover fibrinogen in dentinal fluid
have been unsuccessful (Pashley, unpublished observa-
tion), although labeled fibrinogen has been seen in the
dentin of rat molars (Chiego, 1992). This may be because
the dentin thickness in rat molars following cavity prepa-
ration is only about 100 (Jim. Bergenholtz et al. (1993,
1996) found fibrinogen only in very deep cavities pre-
pared in monkey or human teeth. Pashley et al. (1984)
reported that dentin removed 61% of fibrinogen as it
passed through a 1.0-mm disk. Presumably, in vivo, even
larger amounts of fibrinogen would be removed, because
the tubules contain odontoblast processes at their ter-
minations, which further reduces the sizes of the chan-
nels through which this high-molecular-weight molecule
must move. This molecule may be responsible for reduc-
ing dentin permeability, especially at the pulpal termina-
tions of the tubules, where it may polymerize into fibrin.
The permeability properties of dentinal tubules indi-
cate that, functionally, they have much smaller dimen-
sions than their actual microscopic dimensions (Pashley,
1990). Although the microscopic diameter of dentinal
tubules at the DE) has been reported to be 0.5 to 0.9 |xm,
they function as if they are 0.1 |jim in diameter. Dentin
can remove 99.8% of a bacterial suspension of strepto-
cocci that are approximately 0.5 (xm in diameter
(Pashley, 1985). This tends to prevent infection of the
pulp, even when patients masticate on infected carious
dentin. Fluid shifts across dentin can occur and can
cause sharp, brief pain, but the fluid is virtually sterile.
This is because of the presence of intratubular deposits
112 Crit Rev Oral Biol Med
of mineral and collagen that form multiple constrictions
within the tubule to dimensions less than those of most
micro-organisms. Ten Cate's group recently reported that
65% of the dentinal tubules in occlusal coronal dentin
contain large collagen fibrils (Dai et al., 1991). If we
understand the biologic mechanisms controlling this
reaction, we may be able to reduce the permeability of all
exposed dentinal tubules by collagen secretion into
tubules, thereby decreasing dentin sensitivity and the
potential for pulpal irritation.
(F) Control of dentin permeability
(1) Peripheral Reductions in Dentin Permeability
Whenever dentin is exposed, the pulp is placed at risk
because of the relatively high permeability of normal
dentin. Depending on the magnitude of the permeability
of the exposed dentin (which varies with location, depth,
age, and time (Pashley and Pashley, 1991] and with
whether it is sclerotic, infected with bacteria, etc.), the
types of pulpal irritation would range from hydrodynam-
ic (e.g., fluid shifts in either direction in response to evap-
oration, thermal stimuli, etc.) to immunologic (bacterial
products). The junctional complexes of the odontoblast
layer are disrupted (Turner et al, 1989; Ohshima,1990;
Turner, 1992), permitting relatively large molecules (e.g.,
albumin, globulins, fibrinogen) to exit pulpal blood ves-
sels, gain access to extracellular spaces, and reach the
dentinal tubules, where they can be detected in dentinal
fluid (Bergenholtz et al, 1993, 1996). While it may be dif-
ficult to prevent these changes during restorative dental
procedures, they can be minimized by avoidance of air-
drying of dentin. The subsequent pulpal irritation due to
the diffusion of bacterial products across dentin
(Bergenholtz, 1981) could be avoided if the relatively
high permeability of freshly exposed dentin could be
reduced. This occurs naturally over time via a number of
physicochemical mechanisms (Pashley, 1986), but this
takes days, weeks, or even months to develop, depending
upon the type of subsequent reaction. Dentinal tubules
can be occluded with various crystalline materials (Fig. 7)
following topical application (Pashley and Galloway,
1985; Suge et al, 1995). The use of adhesive resins to
occlude tubules continues to improve in terms of their
sealing ability (Pashley et al, 1992b; Tay et al, 1994).
Ideally, the rapid formation of peritubular dentin would
lower dentin permeability to near zero. Similar low
degrees of dentin permeability have been shown by
Tagami et al (1992) in both aged and caries-affected
dentin, but it took years for sclerosis to develop by accre-
tion of peritubular dentin formation. However, the bio-
logic regulation and control of peritubular dentin forma-
tion are unknown.
7(2):104-133 (1996)
 (2) Internal Reductions in Dentin Permeability
An alternative approach is to apply a biologic growth fac-
tors) to the exposed dentin, allowing it to diffuse across
the dentin to reach the pulp, where it could either stim-
ulate fully differentiated, normal odontoblasts to secrete
additional dentin matrix more rapidly than normal, or
promote the migration and differentiation of mesenchy-
mal cells into odontoblasts if the original odontoblasts
had been destroyed.
As dentin matrix is synthesized, secreted, and min-
eralized, dentin formation appears to trap a number of
growth factors into its structure (Finkelman et al, 1990;
Magloire et al, 1992). That is, growth factors such as
BMP-2, FGF, EGF, IGF-1, and TGF-p may be incorporated
into mineralized dentin matrix. Can these growth factors
be mobilized whenever dentin is demineralized?
Magloire and his colleagues believe that this occurs dur-
ing carious invasion of dentin. The release of these pre-
formed growth factors may permit their diffusion into the
pulp, where they can activate appropriate genes to initi-
ate repair processes.
Transforming growth factors (TGF) include a large
class of molecules that include bone morphogenic pro-
teins (BMPs). D'Souza and Litz (1996) recently reviewed
the importance of TGF-pl in dentinogenesis. The expres-
sion of this growth factor during tooth development
coincides temporally with the initiation of dentin matrix
formation.
The goal of being able to stimulate dentinogenesis
through intact dentin has recently been accomplished.
Smith et al. (1994) created class V cavities in ferret
canines using an atraumatic technique. Control cavities
were treated with either nothing or with rabbit serum
albumin. Experimental cavities were treated with either
lyophilized EDTA-extracted rabbit dentin matrix or with
collagenase-treated insoluble residue remaining after
EDTA extraction. These crude fractions were covered with
a Teflon film and restored with ZOE. In as few as 14 days,
there was significant deposition of reactionary dentin by
normal odontoblasts. In contrast, control cavities lacking
treatment with dentin matrix components showed no
evidence of reactionary dentin deposition. Presumably,
the production of reactionary dentin by normal, fully dif-
ferentiated odontoblasts was due to one or more soluble
growth factors from the dentin matrix. They speculated
that the combination of TGF-p and BMP-2, plus some
unknown factors in the crude extract, might have been
responsible for the reactionary dentinogenesis. The abil-
ity of cells to react to growth factors may be related to
the age of the organism. Smith et al. (1990) reported that
EDTA-soluble matrix proteins from rabbit dentin induced
dentinogenic activity in vivo in pulp exposures in ferret
canine teeth. In young adult ferrets, the reparative dentin
resembled normal primary tubular dentin, while in older
animals it resembled osteodentin in that it was atubular
7(2): 104-1 33 (1996) Crit Rev Oral Bio/ Med
and included cellular inclusions. Smith et al. (1994)
referred to the dentinogenesis that they simulated in
fully differentiated odontoblasts as "reactionary dentino-
genesis", to distinguish it from dentin formation that
occurs following destruction of the original odonto-
blasts, which they termed "reparative dentinogenesis".
The latter is more complex, since it involves the migra-
tion and differentiation of mesenchymal cells into odon-
toblasts.
This was the first unequivocal report of the activa-
tion of dentinogenesis by molecules applied to intact
dentin. The response was less in cavities with thicker
dentin between the cavity floor and the pulp chamber
than if the dentin was thinner. Dentin matrix formation
was sometimes so rapid that cells became trapped with-
in the matrix. The rate of dentinogenesis seemed to
become slower as more dentin was formed. As more
dentin is formed, the amount of growth factors diffusing
from the floor of the cavity to the pulp would be expect-
ed to decrease, which, in turn, should slow matrix secre-
tion. The details of which factors are important, their
dose-response relationships, possible synergism, and
location of receptors (odontoblast cell body) and second
messengers need to be explored. Rutherford et al. (1995)
applied soluble Osteogenic Protein-1 (OP-1) to the floors
of class V cavities prepared in monkey teeth. Control cav-
ities received no treatment, while experimental cavities
received either calcium hydroxide paste or OP-1 at con-
centrations of 0.01, 0.1, 1, or 10 |ULg/|jLL. The lower two
concentrations of OP-1 were found to be ineffective at
stimulating reparative dentin formation when the ani-
mals were killed two months later. In both experimental
groups (calcium hydroxide and OP-1), more reparative
dentin was formed in cavities where the remaining
dentin thickness was 0.2 mm. These results suggest that
receptors for osteogenic proteins are located on the cell
body of the odontoblasts rather than on the odontoblast
process, or that the process is extremely short. These
reports provide additional examples of the biologic cou-
pling of dentin and pulp. While it could be argued that
this is a coupling between dentin and odontoblasts, the
molecular signals, under some conditions, modify non-
odontoblast cells in the pulp to become odontoblasts.
New blood vessels and nerves are then formed to sup-
port the new odontoblasts. Thus, the reactions are
pulpo-dentin in nature.
It would be advantageous to induce odontoblast-like
cells to produce atubular dentin following the topical
application of growth factors to dentin. This would seal
the pulp from hydrodynamic and bacteriologic insult. It
would also interfere with further diffusion of growth fac-
tors to the pulp and would, therefore, be self-limiting.
However, this may not be possible if the odontoblasts
are normal, since the odontoblast process phenotype
requires the production of tubular dentin. Apparently,
113
only less-differentiated odontoblastoid cells without
processes can form atubular dentin.
(Ill) Dynamics of Dentin
Dentin is a living tissue which can and does react to
changes in its environment (Pashley, 1989). It is the only
hard tissue of the tooth which is innervated (Byers,
1984). Normally, dentin is covered coronally by enamel
and, on its root surfaces, by cementum. Whenever dentin
is exposed to the oral environment, it is subjected to
large chemical and mechanical stimuli, in addition to
thermal stimuli and smaller mechanical stimuli which
are effective even in intact teeth. The exposed, fluid-filled
tubules permit minute fluid shifts across dentin whenev-
er dentin is exposed to tactile, thermal, osmotic, or evap-
orative stimuli, which, in turn, activate mechanorecep-
tors in the pulp. These fluid shifts can directly irritate
odontoblasts, pulpal nerves, and subodontoblastic
blood vessels by applying large sheer forces on their sur-
faces as the fluid streams through narrow spaces. In rats,
exposure of dentin irritates the pulp and causes the
release of neuropeptides such as calcitonin gene-related
peptide (CGRP) or substance P (SP) from the pulpal
nerves to create a local neurogenic inflammatory condi-
tion (Kimberlyand Byers, 1988; Byers etal, 1990b; Byers,
1996; Olgart, 1996). These stimuli also produce changes
in pulpal blood vessels, leading to increased flow of plas-
ma fluid and plasma proteins from vessels into pulpal
tissue spaces and out into dentinal tubules (Raab, 1989;
Vongsavan and Matthews, 1992b, 1994; Olgart and
Kerezoudis, 1994). This extravasation of plasma can also
cause increases in local pulpal tissue pressure, which
tends to increase nerve firing (Narhi, 1978). The outward
fluid flow may have a protective, flushing action which
may reduce the inward diffusion of noxious bacterial
products in both exposed cervical dentin (Pashley and
Matthews, 1993) and perhaps even in leaking restora-
tions (Pashley and Pashley, 1991).
(A) REACTIONS TO CAVITY PREPARATIONS:
FLUID SHIFTS
Although most dentists regard cavity preparation as a
minor, routine restorative procedure, from the perspec-
tive of the pulp, it is a crisis. The use of cutting burs in
handpieces produces vibrations, inward fluid shifts
(because of frictional heat generation on the end of a
poorly irrigated cutting bur), outward fluid shifts due to
evaporative water loss (if only air cooling is used), and
slight inward fluid shifts due to osmotic movement of
cooling water into dentin (Horiuchi and Matthews, 1973).
These fluid shifts occur in both directions at various
stages of cavity preparation. Further outward fluid shifts
accompany the application of hypertonic conditioners,
primers, varnishes, or bonding agents (Pashley et al,
1992b), and then, during light-curing of adhesive resins,
114 Crit Rev Oral Biol Ued
additional inward fluid shifts would occur due to heat
generated during polymerization of adhesive resins and
resin composites (Hussey etal., 1995).
All of these fluid shifts create a barrage of hydrody-
namic stimuli across dentin into the pulp. These will
obviously cause pain if the patient is unanesthetized,
and will release sufficient neurotransmitters to cause
local pulpal neurogenic inflammation under the irritated
tubules (Olgart, 1992, 1996) and changes in pulpal blood
flow (Kim and Dorscher-Kim, 1996). Although the fluid
shifts would still occur in patients following administra-
tion of local anesthesia, fewer nerves would fire, there
would be less release of neuropeptides, and there would
be generally less neurogenic inflammation (Raab, 1989,
1992; Hargreaves et al, 1996).
(B) DISRUPTION OF ODONTOBLAST LAYER
Deep cavity preparation in rat molars causes aspiration
of odontoblasts (Byers et al, 1988), while more shallow
cavity preparation causes disruption of junctional com-
plexes between odontoblasts (Ohsima, 1990), which
decreases their barrier properties (Bishop, 1987, 1992),
allowing large molecules (horseradish peroxidase) from
the blood stream to penetrate dentin (Turner, 1989;
Turner et al, 1992). The loss of gap junctions may inter-
fere with the ability of the odontoblasts to secrete a col-
lagen matrix in a synchronous, coordinated manner, due
to the loss of cell signaling between adjacent cells. The
proteins that make up gap junctions are called connexins
(Pinero et al, 1994). How long it takes odontoblasts to re-
establish the continuity of gap junctions and to revert to
tissue rather than cellular function is unknown and
remains a fruitful area for research. The commercial
availability of antibodies to connexins provides the
opportunity for the study of their importance in repara-
tive dentin formation and in pulpal healing following
cavity preparation. Presumably, bacterial substances in
plaque and saliva could easily move from the oral cavity
into the pulp after such operative procedures (Warfvinge
and Bergenholtz, 1986; Pissiotis and Spangberg, 1994;
Bergenholtz, 1996). The mechanism(s) responsible for
the disruption of rat odontoblasts during cavity prepara-
tion has not been studied but is probably due to rapid,
outward movement of dentinal fluid in response to evap-
orative water loss from dentin during cavity preparation.
As sub-odontoblastic capillaries "loop up" into the odon-
toblast layer, disruption of the layer will also cause
microhemorrhages and direct irritation to capillaries that
sustain the odontoblasts. They can leak plasma proteins
(Chiego, 1992) such as fibrinogen out into tissue spaces
which are in communication with dentinal fluid. The per-
meability of dog dentin fell in a time-dependent manner
following cavity preparation (Pashley, 1985). When the
animals were depleted of their plasma fibrinogen, this
reaction was greatly attenuated (Pashley, 1985). Several
7(2):104-133 (1996)
investigators have measured the flux of plasma proteins
across dentin following cavity preparations (Pashley et
al, 1981; Maita et al, 1991; Knutsson et al, 1994). They
reported that several plasma proteins, including albu-
min, could be detected in relatively high concentrations
immediately after cavity preparations in monkeys and
humans (Knutsson et al, 1994) rjut that their concentra-
tion fell over the next few hours to days. These plasma
proteins contain immunoglobulins which may inactivate
some bacterial products. There is the possibility that
bacterial or salivary products may activate complement,
which would also contribute to pulpal inflammation.
Many of these reactions to cavity preparation would
be immediate and short-term. Displacement of odonto-
blasts up into tubules disrupts their internal cytoskele-
ton and causes cell death (Eda and Saito, 1978; Ohsima,
1990). These cells undergo autolysis over the next few
days and are replaced by mesenchymal reserve cells
which begin to differentiate into new odontoblasts. The
cell signals that are responsible for this process are
beginning to be understood (D'Souza and Litz, 1996).
Displacement of odontoblasts and their replacement by
new cells (Fitzgerald, 1979) are associated with very little
pulpal inflammation if the environment is sterile and if
the dentin is well-sealed. It is possible for cavities in
teeth to be prepared with no pulpal inflammation or for-
mation of reparative dentin (Smith et al, 1994). This is
done with copious air-water spray, light, intermittent cut-
ting forces, and the use of sharp burs. However, if the
dentin is not sealed (i.e., cervical abrasion) or if the
restored cavity exhibits microleakage, the pulpo-dentin
complex will undergo long-term reactions which are the
result of continual leakage and permeation of bacterial
substances around gaps in restorations (Pashley and
Pashley, 1991) and through unsealed dentinal tubules
into the pulp. This can convert acute pulpal reactions
associated with cavity preparation or dentin exposure
into chronic pulpal reactions.
Bacteria and their products (Pissiotis and
Spangberg, 1994) have been shown to produce severe
pulpal reactions (Bergenholtz, 1977; Bergenholtz and
Warfvinge, 1982; Warfvinge and Bergenholtz, 1986). The
argument over which irritants, bacterial or chemical,
cause more pulpal inflammation has been going on for
decades. Early germ-free animal studies (Kakehashi et al,
1965; Kobayashi and Yoshida, 1981) demonstrated mini-
mal pulpal reactions to materials. This has been con-
firmed more recently by studies in monkey teeth, where
restorative materials have been placed directly on pulpal
soft tissues followed by surface sealing (Cox et al, 1985;
Cox, 1987; Cox et al, 1987). Inoue and Shimono (1992)
recently reported minimal pulpal reactions in germ-free
rat molars following direct pulp capping with Superbond
C&B resin.
The pulpal reactions seen in most dental materials
7(2) 104-133 (1996) Crit Rev Oral Biol Med
usage studies seem to be the result of the bacterial col-
onization of gaps between restorative materials and
tooth structures. These colonies continually shed bacte-
rial products which result in continual pulpal irritation.
The extensive experimental work supporting this conclu-
sion has been reviewed by Browne and Tobias (1986), by
Cox (1987), by Cox et al (1987), and by Cox and Suzuki
(1994).
(C) PULPAL REACTIONS TO DENTIN EXPOSURE:
NERVE RESPONSES
Sensory nerve fibers innervating dentin respond to cavi-
ty preparation by release of stored materials, including
neuropeptides such as CGRP (Gazelius et al, 1987; Byers
et al, 1988, 1992a; Heyeraas et al, 1994; Byers, 1996),
which has subsequent effects on local blood flow (Olgart
et al, 1991), vessel permeability (Raab, 1989; Olgart and
Kerezoudis, 1994), and local tissue pressure (Heyeraas et
al, 1996). Presumably, the depletion of the neuropeptides
is in response to rapid fluid shifts within the dentinal
tubules produced by evaporative, thermal, and physical
forces that accompany drilling of the tooth dentin. The
exact mechanism causing the release of neuropeptides is
not known but may be due to shear forces on the nerve
terminals. Byers (1996) reported more CGRP-immunore-
activity in sensory nerves in rat molar teeth in function
than in teeth out of function. It is interesting to speculate
how masticatory function might be transduced into neu-
ropeptide activity in nerve terminals. It is possible that
neuropeptides such as CGRP and SP may be continu-
ously released at a slow rate that is unrelated to depo-
larization of the nerves (Olgart, 1990). These peptides
could have local metabolic activities on adjacent fibro-
blasts, odontoblasts, and pulpal blood vessels. The
occlusal surfaces of the rat molars contain exposed
dentin. Thus, during occlusal function, rats grind food on
exposed occlusal dentin. This would be expected to
cause fluid shifts across dentin into the pulps of these
teeth during function (Duncan and Turner, 1995). Byers
(1996) speculated that masticatory activity in rat molars
activates pulpal nerves hydrodynamically, leading to
release of stored CGRP in these nerves. These are rapid-
ly replaced by axonal transport so that, during periods of
inactivity, those teeth that are in heaviest function have
the most CGRP immunoreactivity. A temporal study of
this functional response in young rats might reveal
insights into the mechanism(s) involved. Teeth taken out
of occlusion or placed in hyperfunction might also be
studied.
Whether these responses can be demonstrated in
human teeth remains to be seen. Luthman et al (1992)
could not find many nerve fibers immunoreactive for
CGRP in human teeth. However, Byers et al (1990a) have
shown that neuropeptides in human teeth may be lost
because of nerve injury during extraction and because
115
immersion-fixation is not as rapid as perfusion-fixation
used in animals, leading to proteolysis of neuropeptides.
These releases of CGRP may be sufficient to influence the
mitogenic activity of adjacent cells (Nilsson et al, 1985;
Bernard and Shin, 1990; Trantor et al, 1995) or the secre-
tory activity of odontoblasts (Heyeraas et al, 1996). This
may provide an important coupling between occlusal
function (which in the rat includes loss of tooth structure
through attrition) and dentinogenesis, which is neces-
sary to maintain a mineralized barrier of constant thick-
ness between the pulp and the functioning surfaces of
teeth. In the absence of such coupling in rats, occlusal
function may cause too much dentin to be formed in the
coronal pulp chamber. Denervation experiments have
shown that nerves grow back to innervate the original
terminal nerve distribution pattern in cuspal dentin, sim-
ilar to the neural pattern observed before axotomy
(Berger and Byers, 1983; Heyeraas et al, 1996). What
trophic factors are responsible for such neural activity?
Schwann cells envelop nerve fibers until the nerves reach
the predentin, where the nerve terminals become naked.
Following denervation, the Schwann cells up-regulate
the expression of low-affinity nerve growth factor recep-
tor (p75-NGF receptor) (Byers et al, 1992b,c). Pulpal
injury causes the expression and release of nerve growth
factor (NGF) from pulpal fibroblasts (Byers et al, 1992b).
The signals that cause this are not yet known. This reac-
tion precedes the sprouting of CGRP- and SP-containing
nerves in the pulp. The double-labeling experiments of
Fried and Risling (1991) showed that both CGRP and SP
immunoreactive sensory nerves contain the p75-NGF
receptor and are thus able to respond to NGF (Byers et al,
1990a). As these nerves release CGRP and SP, these neu-
ropeptides may enhance the rate of cell division of the
local fibroblasts (Trantor et al, 1995), providing further
release of NGF and other growth factors. The increased
local production of NGF seems to promote sprouting of
local nerves. The terminal nerve endings in the dentin
are devoid of neural sheaths that guide neuronal regen-
eration in most connective tissue. Does fluid shift across
permeable dentin, or does impermeable dentin com-
press during occlusal function enough to cause fluid
shifts at the dentino-pulpal junction? How are fluid
shifts across tubules sensed in the pulp? Do local fibrob-
lasts secrete NGF in response to shear gradients set up
by the flow of fluid around them through extracellular
spaces (Heyeraas, 1989; Schmid-Schonbein, 1990)?
If the fluid shifts are so severe that the high shear
forces destroy the odontoblasts, the more primitive mes-
enchymal cells that differentiate to produce reparative
dentin tend to form atubular, fibrodentin because they
lack odontoblast processes. Even a thin layer of atubular
dentin matrix may increase the resistance of the tubules
to fluid flow across dentin enough to disrupt the cou-
pling between occlusal function/dentinal fluid shifts.
116 Crit Rev Oral Biol Ued
Reparative dentin is not well-innervated and seems
to form much faster than innervated tubular dentin
(Byers, 1984). This may be due to the fact that atubular
dentin formation disrupts the fluid shifts across dentin
which normally couple occlusal function and dentino-
genesis. In some cases, the rate and direction of repara-
tive dentin formation may be inappropriate and uncon-
trollable. Viewed in this way, the innervation of dentin
functions in a manner that couples tooth function to pul-
pal metabolism rather than as a sensory system (Byers,
1996). A constant thickness of tubular occlusal dentin in
the face of constant abrasion from mastication may pro-
vide a relatively constant hydraulic conductance and/or
modulus of elasticity of the dentin which would provide
relatively constant fluid shifts in response to occlusal
function. If the dentin became too thick, the resistance to
fluid flow (or the modulus of elasticity) would increase
so that occlusal function would move less fluid across
dentin in response to occlusal function. This might pro-
duce less release of CGRP from nerves associated with
these tubules and hence less formation of tubular dentin
(Heyeraas et al, 1996). Excess occlusal attrition would
wear down the occlusal dentin faster than it was being
formed, and the fluid shifts produced by masticatory
function would increase, causing more release of CGRP
and hence faster dentin formation. There seems to be an
association between innervation density and the width
of predentin (Byers, 1984). Those regions of the pulp that
have the highest density of nerves also have the widest
predentin. This could be due either to higher rates of
matrix secretion or to lower rates of mineralization.
While such correlations do not prove cause and effect,
they indicate areas for future research.
(IV) Pulp
(A) PULPAL WOUND HEALING
The therapeutic use of growth factors such as OP-1
bound to insoluble collagen matrix particles (Rutherford
et al, 1994a,b) to stimulate dentinogenesis following
pulpotomy is an example of the recent application of
molecular biology to the treatment of pulpal wounds.
While this combination has been shown to produce
dentin (Anneroth and Bang, 1972; Nakashima, 1989,
1990; Tziafas and Kolokuris, 1990), the distribution of the
dentin is limited to the distribution of the pieces of dem-
ineralized bone collagen matrix. This has both advan-
tages and disadvantages. One advantage is that dentin
forms only where the collagen has been placed. That is
also a disadvantage, in that, once initiated, dentin for-
mation continues until all of the collagen particles are
mineralized. If too much collagen matrix is applied or if
several pieces extend into the root canal, dentin will
form at these sites, in addition to where it is intended to
form, at the orifice of the root canal. The shape of the
7(2):104-133 (1996)
mineralizing mass follows the shape of the collagen
matrix fragments. Due to their irregular shape, there are
spaces between the collagen fragments that tend to
remain, albeit smaller, after dentin formation. This
makes it difficult to create a uniform, hard tissue barrier
to seal off the pulp chamber space from the root canal.
Perhaps in the future, the material can be placed on a
sheet of collagen that can be trimmed to appropriate
dimensions for the surgical site. The type of dentin
formed by OP-1 bound to collagen is primarily atubular
dentin or mixtures of atubular and tubular dentin which
contain cellular inclusions rather than normal tubular
dentin. Long-term (i.e., years) studies in subhuman pri-
mates are not yet published. There is some concern that
dentinogenesis may continue to occur longer than is
necessary. That is, excessive dentinogenesis may obliter-
ate the pulp space and roots, making subsequent
endodontic treatment more difficult if not impossible.
The stability of this osteodentin/bone matrix is also
unknown over a period of 5-10 years. Thus, while promis-
ing, this approach to pulp therapy needs further refine-
ment.
The use of crude extracts of dentin for topical appli-
cation of growth factors to pulpotomy sites has been
shown to induce reparative dentin formation (Smith et
al, 1990), due to the presence of a number of important
growth factors that are present in mineralized dentin
matrix (Finkelman et al, 1990). Dentin bridge formation,
while desirable, often produces an incomplete hard tis-
sue barrier because of numerous "tunnels" that tend to
form as cellular elements of the pulp are incorporated
into the rapidly formed dentin (Clark, 1970; Langeland et
al, 1971; Cox et al, 1996). Thus, even after formation of
"dentin bridges", pulpal inflammation may develop if the
access opening is not well-sealed. Indeed, dentinogene-
sis will not occur in the presence of pulpal inflammation
(Coxetal, 1987).
The use of calcium-hydroxide-containing pulp-cap-
ping agents on pulpal wound healing was reviewed by
Schroder (1985). However, the notion that placement of
calcium-hydroxide-containing varnishes or bases on
intact dentin will stimulate odontoblasts to form repara-
tive dentin is exaggerated (see review by Cox and Suzuki,
1994). The use of calcium hydroxide formulations to pro-
duce dentin bridges on direct pulp exposures is well-
documented (Clark, 1970; Langeland et al, 1971;
Schroder, 1985; Cox et al, 1996), but many investigators
have reported dentin formation against a number of non-
calcium-hydroxide-containing dental materials (Nyborg
etal, 1969; Cox et al, 1985, 1987; Cox, 1987).
An alternative to the use of growth factors absorbed
on collagen or calcium hydroxide as a biologic pulp
dressing is the use of adhesive resins placed directly on
exposed pulps as a dressing/sealing agent. Inoue and
Shimono (1992) demonstrated excellent pulpal healing
7(2): 104-1 33 (1996) Crit Rev Oral Bio! Med
and dentinogenesis against Superbond C&B (Sun
Medical Co., Ltd., Osaka, lapan) in germ-free rat molars.
In a later paper, they reported on infiltration of a 4-META
adhesive resin into tissue spaces when this material was
used as a pulp-capping agent in freshly extracted human
teeth (Miyakoshi et al, 1994). They found that the resin
formed a uniform protective layer of resin on the pulp
surface which was very difficult to remove. A recent
histopathologic study of the pulpal response to placing
Dycal, Light-cured Dycal, or one of two adhesive resin
systems (Clearfil Liner Bond System or Clearfil Liner
Bond II, Kuraray Co., Ltd., Kyoto, Japan) directly on pulp
exposures in monkey teeth was reported by Onoe (1994).
Animals were killed after 3, 90, or 360 days. He reported
more initial pulpal inflammation under the calcium
hydroxide control materials (Dycals) than under the
adhesive resins. The resin-pulpal soft tissue interface
was more porous under the Clearfil Liner Bond System
than under the Clearfil Liner Bond II resin. The author
speculated that it was due to the fact that the latter sys-
tem required no rinsing and hence provided superior
moisture control during placement. When the pulpal
wounds were examined after one year of healing, dentin
bridges had formed under all materials. The calcium-
hydroxide-containing materials seemed to produce more
osteodentin initially which later became more tubular.
Although less dentin was formed under the adhesive
resins, it was all tubular dentin and appeared to have
fewer defects than were seen under calcium hydroxide
dressings. There were no gaps or spaces between the
resins and the reparative dentin that was formed (unlike
Dycal), indicating good biocompatibility. When the inter-
face between pulps treated with Clearfil Liner Bond II
was examined by transmission electron microscopy, the
presence of resin which had infiltrated tissue (or vice
versa) was seen as a unique resin-cell hybrid layer. The
newly formed tubular dentin was immediately beneath
this layer. It is interesting to speculate on the molecular
biology of these pulpal responses to two widely different
chemical agents.
Calcium hydroxide has long been used to induce
dentinogenesis, but the mechanism(s) involved has
remained unknown. Smith et al (1995) recently reported
that calcium hydroxide solutions can solubilize growth
factors, including TGF-p, from dentin. They speculated
that this might be the mechanism responsible for the
success of calcium hydroxide treatments in producing
dentin bridge formation. It should be noted that the
primer used in the Clearfil Liner Bond II bonding system
mentioned above (Onoe, 1994) consists of 20% Phenyl-P,
an acidic adhesive resin with a pH of 1.5. As this priming
agent flows across the wound to the adjacent dentin or if
it comes into contact with any dentin chips or grinding
debris in the wound, it may demineralize the dentin and
release the growth factors known to exist in dentin
117

5-7 D
CHRONIC
INJURY
64 + 16
44116;
28116!
1 the pulps of old rats (Swift and Byers, 1992).
INNERVATED DENTINAL TUBULES
Figure 8. The percentage of innervated dental tubules near the
tip of the pulp horn and the mid-crown, and at the junction of
crown and root in control rat molars and those 5-7 days after
the cusps were ground down. (From Byers et a/., 1988, with
permission.)
(Finkelman et al, 1990). These growth factors may, in
turn, be responsible for the induction of hard tissue for-
mation, with the resin layer providing an excellent seal of
the wound against penetration of bacterial or oral fluids.
It is possible that the interaction of the resins with plas-
ma proteins, tissue proteins, and connective tissue cells
may provide the appropriate substrate for precursor cells
to complete their phenotypic expression to fully differ-
entiated odontoblasts earlier than is the case with calci-
um hydroxide. This model would provide an excellent
opportunity to examine the temporal changes in cell
recruitment, cytokine formation, gene expression
(Takeda et al, 1994), growth factor eleboration, matrix
secretion, and mineralization.
Thus, there is a convergence of research efforts
demonstrating that exogenous growth factors such as
OP-1 (Rutherford, 1996), calcium hydroxide, and adhe-
sive bonding agents (Onoe, 1994) can induce reparative
formation when placed directly on exposed pulps. It
remains to be determined which approach will be most
successful therapeutically.
118 Crit Rev Oral Bio! Med
(B) INNERVATION OF DENTIN
(1) Developmental
What follows is a general overview of the innervation of
dentin. For detailed reviews, see Fearnhead (1957, 1961),
Byers (1980, 1990, 1994), and Hildebrand et al (1995).
Nerves grow into the pulp early in development at
about the bell stage of tooth development (Kollar and
Lumsden, 1979; Byers, 1980; Mohamed and Atkinson,
1983). The innervation includes both afferent neurons
and efferent autonomic nerves that innervate pulpal
blood vessels. The number of myelinated axons in per-
manent teeth increases with age and/or tooth develop-
ment, reaching a plateau value of about 500 myelinated
axons per human premolar at 15 years of age (Johnson,
1990), which remains fairly constant up to age 60.
Unmyelinated axons are more numerous (~ 1500 axons
per premolar) and seem to show a reduction in number
with age (Fried and Hildebrand, 1981; Byers, 1984;
Johnson, 1990). Fried (1992) reported large reductions in
axon number and CGRP, SP, neuropeptide Y (NPY), and
NGF receptor activity in the pulps of cats as a function of
age. Similar reductions in CGRP have been reported in
On histologic examination, the nerves can be seen
entering the apex of the root as one or more dental
nerves in a bundle. Only a few nerves are given off in the
root (< 10%, Byers and Matthews, 1981), while most con-
tinue to the coronal pulp, where they diverge, branch,
and innervate pulpal blood vessels and form the nerve
plexus of Rashkow. From this plexus, some nerves pass
further peripherally to terminate in predentin or dentin,
although few free nerve endings extend more than 100
IJim into tubules. Details of the innervation of pre-
dentin/dentin can be found in reviews by Byers (1984)
and Hildebrand et al (1995). Byers (1984, 1990) extended
the work of Fearnhead (1957, 1961) and summarized the
work of many to show that the percentage of dentinal
tubules innervated by sensory nerves is more than 40% in
the region of coronal pulp horns, 4.1-8.3% in mid-crown,
0.2-1.0% at the CEJ, and 0.02-0.2% in mid-root in humans
(Fig. 8). A qualitatively similar pattern of distribution of
sympathetic nerves was reported by Avery et al (1980).
They found a large number of adrenergic nerve endings
in odontogenic areas near the basal ends of odonto-
blasts, and speculated that they may somehow modify
dentinogenesis. Similar observations have recently been
made by Inoue et al (1995) in human teeth. The authors
speculated that the close association of these nerves
must have some effect on odontoblast cell function.
The exact mechanisms involved in attracting nerves
into the pulp and in regulating their regional distribution
have not been fully elucidated. The importance of NGF in
this process is clear (Byers et al, 1990a). That is, NGF
7(2):104-133 (1996)

Figure 9. Schematic drawing of a tooth containing a class II restoration and exhibiting exposed, sensitive cervical dentin. Cross-section
shows that leakage of bacterial products may lower the threshold of sensory nerves whose branches innervate both regions, making the
cervical dentin more sensitive than normal.
seems to regulate where nerves grow and whether they
sprout in response to pulp injury.
Sensory nerves are divided into subgroups based on
histologic size, conduction velocity, and function. Most
of the myelinated nerves are classified as A-(3 or A-8 sen-
sory nerves. These are relatively large, fast-conducting
fibers that respond to hydrodynamic stimuli such as tac-
tile, evaporative, osmotic, or thermal challenges (Narhi et
al, 1996). Small unmyelinated C-fibers have slow con-
duction velocities and higher thresholds than A-fibers
and respond only to intense stimuli which reach the pulp
rather than to stimulation of dentin surfaces. They do not
respond to hydrodynamic stimuli but are responsive to
bradykinin, histamine, and capsaicin (Narhi et al, 1996).
Other small nerve fibers of the same diameter as C-fibers
(0.2-1.0 |xm) in the pulp are postganglionic sympathetic
nerves. Although parasympathetic nerves in the pulp
have been proposed (Avery and Chiego, 1990), and
acetylcholine and the neuropeptide VIP (vasoactive
intestinal peptide) have been shown to co-exist in
cholinergic parasympathetic nerves, their role in the reg-
ulation of pulpal blood flow is not clear. The vasodilation
of pulpal blood vessels following acetylcholine infusion
is probably due to release of nitric oxide by endothelial
cells (Lohinai etai, 1995). However, Aars et al. (1993) and
Anderson et al. (1994) recently reported evidence that
indicates that human pulpal blood flow may be influ-
enced by a cholinergic system.
(2) Receptive Fields
Individual A-8 fibers have discrete receptive areas on
dentin which can be located by probing. Although these
areas are usually localized to a small spot on the dentin
surface, some single nerve fibers have two or three sepa-
rate receptive areas separated from each other by many
millimeters. If such branching fibers were involved in
7(2):104-133 (1996) Crit Rev Oral Bid Med
Bacterial products from
leaking restoration may
lower threshold on entire
nerve and its branches
neurogenic inflammation, the response could be wide-
spread. The work of Ngassapa (1991) on the receptive
fields in dentin of single-pulpal neurons has raised many
new questions. Are these receptive fields stable or labile
with age? Although the size of the fields does not seem
to change in acute vs. chronic dentin exposure, do their
sensitivities change (Narhi et al, 1996)? Hovgaard et al.
(1991) reported that teeth with cervical dentin sensitivity
which were unresponsive to conventional desensitizing
therapy also contained coronal restorations. While these
restorations were judged to be clinically acceptable,
these investigators removed the restorations and lined
the cavities prior to the insertion of new restorations.
Without any treatment of the cervical dentin sensitivity,
most of these teeth became either insensitive or much
less sensitive over the ensuing few weeks (Fig. 9). Was
the cervical sensitivity due to bacterial irritation of coro-
nal branches of the same neuron, produced by leakage of
the restorations? Ngassapa's thesis work was also impor-
tant in that it demonstrated that chronic exposure of
coronal dentin in dogs did not cause increases in CGRP
activity in pulpal nerves, as it does in rats (Taylor et al,
1989). Further, topical application of CGRP to exposed
dentin did not seem to change the sensitivity of single-
nerve units to hydrodynamic stimuli. Thus, the function-
al correlates of nerve sprouting of CGRP-containing pul-
pal nerves remains to be determined. The recent report
of the mitogenic activity of CGRP and SP to pulpal fibrob-
lasts indicates one possible non-sensory activity of pul-
pal nerves (Trantor et al, 1995).
Matthews (1992) suggested that some axons may
branch such that one branch terminates at a sensory
receptor and another branch of the same nerve termi-
nates on a pulpal arteriole. Whenever the sensory recep-
tor is stimulated, the nerve impulse would travel both
centrally and to the branch innervating vascular smooth
119
muscle, where it could release CGRP or SP and cause
vasodilation. This is an example of an axon reflex
described by Lewis (1927) to explain the flare reaction
which follows stimulation of cutaneous nociceptors.
Matthews (1992) pointed out that A-8 fibers may be a
richer source of vasodilating peptides in cat pulp than C-
fibers, since antidromic stimulation of both A-8 and C-
fibers together produced only a slightly greater vasodila-
tion in cat pulp than did stimulation of the A-8 fibers
alone. This is the opposite of what has been observed in
skin, where C-fibers are the main source of vasodilating
peptides. Matthews' observation (1992) needs to be con-
firmed in both normal and inflamed pulps.
Narhi's group has looked only at the sizes of sensory
receptive fields by recording single-unit activity in
response to stimulation of exposed dentin surfaces. How
many of the nerves that they stimulate send branches to
pulpal blood vessels? What is the size of the "axon reflex
field", and does it change in pulpal inflammation? This
might be studied by looking for extravasation of Evans
blue dye from pulpal blood vessels following stimulation
of a single A-8 fiber on the dentin surface.
(3) Modification of Nerve Activity
Hargreaves et al. (1996) reported that the addition of 2%
lidocaine to superfusion medium of pulpal tissue
blocked the potassium-evoked release of CGRP.
Interestingly, epinephrine alone at a concentration of
1:100,000 (or 0.01 jig/mL) was as effective as 2% lido-
caine in blocking CGRP release. The authors speculated
that adrenergic drugs might suppress the release of pul-
pal neuropeptides. The authors should expand this work
to determine the dose-response relationships of these
two drug classes. Epinephrine is present in both local
anesthetic solutions and in gingival retraction cord.
Ciarlone and Pashley (1992b) found that epinephrine
from retraction cord could diffuse across the full thick-
ness of root dentin in vitro and could be detected in the
pulp chamber. The results of Hargreaves et al (1996) sup-
port in vivo studies, recently summarized by Olgart (1996),
that norepinephrine released from pulpal sympathetic
nerves reduces afferent nerve-induced vasodilation and
plasma extravasation, probably due to inhibition of pep-
tide release.
Hargreaves et al. (1996) have also shown that
extremely low doses (0.7 pmol/L) of recombinant human
NGF lowered capsaicin-stimulated CGRP release from
pulpal tissue by 50%. They speculated that neurotrophic
proteins such as NGF may produce analgesic activity in
chronically inflamed teeth, making them asymptomatic.
If this is confirmed, it may provide an explanation for the
lack of correlation between the signs of pulpal inflam-
mation and patient pain (Hargreaves et al, 1996).
Hannington-Kiff and Dunne (1993) reported that
topical application of 1% guanethidine, an anti-hyper-
120 Crit Rev Oral Biol Med
tensive agent, to hypersensitive dentin caused an imme-
diate and prolonged (two-hour) decrease in sensitivity.
They speculated that it might release norepinephrine
from sympathetic nerves in the pulp. This could reduce
pulpal blood flow, which has been shown to block the
function of A-8 fibers (Olgart and Gazelius, 1977; Narhi,
1985).
(4) Dentin Sensitivity
During the last century, dentists recognized that as soon
as they drilled past enamel and entered dentin, patients
complained of pain. They encouraged their colleagues in
the anatomical sciences to stain dentin for nerves in an
attempt to explain the sensitivity of dentin. However,
these histologic studies failed to identify nerves in
peripheral dentin. Thus, the notion that stimuli applied
to dentin surfaces caused direct activation of intradental
nerves had to be rejected. Others thought that the odon-
toblast process might function as a sensory receptor that
might synapse with pulpal nerves. However, no synapses
could be found between pulpal nerves and odontoblasts.
This theory would also require that the odontoblast
process extend to the DEI, a requirement that is not con-
sistently seen. Prolonged air blasts applied to dentin
cause aspiration of odontoblasts into the tubules, which
destroys them, yet such dentin remains sensitive. Thus,
for many reasons, the odontoblast as a sensory receptor
was rejected. These observations led a number of early
dental scientists to consider that painful stimuli might
produce fluid movement across dentin (Blandy, 1850;
Gysi, 1900; Kramer, 1955), which, in turn, activates pulpal
nerves and causes pain (Fig. 10A). Through a series of
simple but important in vivo and in vitro experiments con-
ducted in the 1960s, Brannstrom and his colleagues (for
review, see Brannstrom, 1992) demonstrated that pain-
producing stimuli induced fluid shifts across dentin in
vitro. They re-introduced the hydrodynamic concept
(Gysi, 1900) but provided experimental data to support
its development into a plausible theory. They reported
that cold stimuli were more painful than hot stimuli of
the same magnitude (Brannstrom et al, 1967). They also
demonstrated that cold stimuli produce an outward fluid
movement, while hot stimuli produce an inward fluid
movement. This was later confirmed by Horiuchi and
Matthews (1973). Apparently, thermal stimuli cause
more rapid volumetric changes in dentinal fluid than in
dentin itself, resulting in an outward (cold) or inward
(hot) movement of fluid. The reason(s) why cold stimuli
are more effective than hot stimuli is unclear. It may be
due to the fact that the resistance to fluid movement
across dentin is different when fluid moves inward vs.
outward. Pashley (1985) measured the hydraulic conduc-
tance of dentin in dog teeth in vivo when positive or neg-
ative hydrostatic pressures were applied. Positive hydro-
static pressures always produced higher hydraulic con-
7(2):104-133 (1996)
 B
-
.<
.V ••
f
K+
•• •
/A
p • • •• •

Figure 10. (A) Schematic diagram of inward fluid movement displacing the odontoblast slightly away from the tubule orifice, which
decreases local velocities of fluid movement across the nerve terminals. Current view of hydrodynamic theory showing branch of sensory
nerve on blood vessel. Activation of nerves causes release of neuropeptides from branches to vessels (stars), causing increases in transu-
dation of fluid (large black arrows) from vessel to tubule. (B) This increase in outward fluid flow could combine with subthreshold hydro-
dynamic stimuli to cause nerves to fire. Note that outward fluid flow displaces the odontoblast cell body slightly, making the dimensions
of spaces for fluid flow smaller, thereby increasing the velocity of fluid and hence the shear forces on odontoblasts and on nerve termi-
nals. Odontoblasts may also modify nerve thresholds by releasing potassium ions (K+) if irritated or injured.

ductances (and higher fluid shifts) than negative pres-
sures. They speculated that the odontoblast process/cell
body can move slightly within the pulpal orifice of the
tubules in response to hydrostatic pressure gradients.
Positive hydrostatic pressures produce an inward fluid
movement that would tend to displace the odontoblasts
slightly away from the orifice (Fig. 10A), while negative
pressures would tend to seat them more firmly in the
tubules (Fig. 10B), thereby changing their hydrodynamic
resistances. Since shear forces depend upon the velocity
of fluid movement, which, in turn, depends upon the
dimensions of the channels available for fluid flow, out-
ward-directed fluid flow may produce higher shear
forces, even though the volume flow is lower. Similar
observations were made by Vongsavan and Matthews
(1994) in cat teeth in vivo. There are other differences
between hot and cold stimuli. As dentinal fluid is cooled,
its viscosity increases significantly (46% increase in vis-
cosity when water is cooled from 25° down to 10°C).
Since dentinal fluid flow through tubules is inversely
related to viscosity (Pashley et al, 1983), decreases in
temperature would decrease fluid flow but increase
shear stresses on dentinal/pulpal nerves. Heating dentin
would do the reverse.
Since the hydrodynamic theory (Fig. 10A) includes
both fluid conductance and neural elements, changes in
either could cause changes in dentinal pain (Pashley,
1992). Upon exposure of dentin to the oral cavity, the
hydraulic conductance of dentin suddenly increases, per-
mitting both a slow outward spontaneous fluid flow (due
to the fact that pulpal tissue pressure is higher than oral
cavity pressure), and rapid fluid shifts in response to the
application of painful stimuli (thermal, mechanical,
osmotic, evaporative). However, several physicochemical
phenomena tend to occlude open tubules, making them
less conductive over time. These processes also decrease
dentin sensitivity over time (Pashley, 1986). The sensitiv-
ity can return if the root surfaces are exposed to grape-
fruit or other citrus fruits, because citric acid would etch
the dentin, demineralizing it (Addy et al, 1987) and mak-
ing it hyperconductive again (Gunji, 1982). As the dentin
becomes hyperconductive, the increased outward fluid
flow (which is the product of the hydraulic conductance
of the dentin and the pulpal tissue pressure) may tend to
rinse the tubules free of any bacterial substances that
may have been diffusing across dentin toward the pulp.
In the absence of this protection, bacterial substances
can diffuse across the dentin, utilizing open tubules to
reach the pulp (Pissiotis and Spangberg, 1994). This may
attract polymorphonuclear leukocytes and/or other
inflammatory cells which may elicit a local inflammatory
response that causes not only elevations in local tissue
pressure and hence more outward fluid flow, but also an
increase in the local concentrations of endogenous
mediators of inflammation sufficient to modify local
nerve thresholds and the distribution of nerve terminals,
or increase the size of receptive fields, thereby increasing
dentin sensitivity (Narhi et al, 1996).
Careful study of dentin sensitivity will ultimately
require simultaneous measurements of fluid flow and
nerve activity. It has been shown in electrophysiological
recordings from dentin in human teeth in vivo that

7(2):104-133 (1996) Crit Rev Oral Bio/ Med 121

intradental nerve activation is related to pain sensations
perceived by the subjects (Edwall and Olgart, 1977;
Ahlquist et al, 1984). Ahlquist et al. (1994) recently
demonstrated that changes in the application of external
hydrostatic pressure were associated with pain more
than were changes in the application of steady pressure,
and that occlusion of the tubules by calcium oxalate
blocked that nerve activity. However, none of these stud-
ies actually measured fluid flow during the pain
response.
The most sensitive method for measuring fluid shifts
across dentin has been developed by Vongsavan and
Matthews (1991, 1992 a,b, 1994), using cat teeth. In
essence, the procession of fat droplets in a micropipette
connected to exposed dentin is followed under micro-
scopic observation, or slight pressure drops across
micropipettes are used to quantitate fluid movements in
response to a variety of manipulations. Using this model,
they have demonstrated that there is normally an out-
ward movement of fluid from the pulp through dentin to
the exposed surface. The rate of this fluid flow depends
upon the hydraulic conductance of dentin (Pashley,
1990) and the pulpal tissue pressure, which they mea-
sured to be 15 cm H2O (Vongsavan and Matthews, 1992a)
in normal cat teeth. Ciucchi etal. (1995) have measured a
similar value in normal human teeth. When Vongsavan
and Matthews (1994) electrically stimulated the cervical
sympathetic trunk in cats, there was a fall in pulpal blood
flow and a fall in outward fluid flow. At high rates of stim-
ulation (2-10/s), the fluid flow actually reversed direction
so that there was an inward fluid flow from dentin into
the pulp. Presumably, as pulpal blood flow fell, local tis-
sue pressure fell. Brown et al. (1969) demonstrated that
the colloid osmotic pressure of plasma proteins could
cause negative tissue pressures under special conditions
of low to zero pulpal blood flow. Thus, under special con-
ditions (intense sympathetic stimulation or adrenergic
drug-induced vasoconstriction), dentinal fluid (and its
contents) can be "aspirated" into the pulp.
Matthews and Vongsavan (1994) and Matthews
(1996) have carefully studied the relationship between
outward fluid flow rates and activation of intradental
nerves. Spontaneous outward fluid movement (18.1 pL
sec"1 mm"2) was not sufficient to cause activation of pul-
pal nerves. The threshold of outward fluid movement
necessary to activate intradental nerves was found to be
1.0-1.5 nL sec"1 mm"2, or 50 times greater than resting or
spontaneous outward fluid flow rates. This rate is 10
times larger than the outward fluid flow rates that they
measured during antidromic stimulation of the inferior
alveolar nerve. They speculated that conditions that
cause more outward fluid movement would bring these
mechanoreceptor nerves closer to threshold.
Presumably, the fluid flow is occurring between the
tubule wall and odontoblasts and/or nerve terminals.
122 Crit Rev Oral Biol Ued
These velocities of fluid movement may produce suffi-
cient shear forces to deform nerve terminals and thereby
excite them. They suggested that the neural elements
responsible for pain may not necessarily involve stretch-
sensitive ion channels but rather simple shear forces on
free nerve endings.
One of their assumptions was that the fluid flow was
uniformly distributed among the tubules within the
exposed dentin surface area. However, the recent work by
Hughes et al. (1996), using a new technique for measur-
ing fluid flow across dentin, indicates that such an
assumption may not be valid. They scanned a micro-
scopic electrode (ca. 1 |xm) near the surface of a disc of
dentin in a raster pattern. Potassium ferrocyanide was
placed in a chamber below the dentin slab and allowed
to diffuse across open dentinal tubules to reach the
scanning electrode tip, where it was electrochemically
oxidized at a transport-controlled rate. After obtaining
the baseline diffusional transport rates across dentin,
the authors then increased the lower (i.e., "pulpal") cham-
ber pressure to 20 cm H2O to induce convective transport
of potassium ferrocyanide along with the diffuse trans-
port. The authors could detect and create images of high-
ly localized (i.e., solitary tubules) fluid flow surrounded
by regions that had no convective transport. They sug-
gested that fluid flow across dentin is far from uniform
and that those tubules that permit fluid flow have veloci-
ties that are three orders of magnitude higher than was
reported by Matthews and Vongsavan (1994). Clearly,
more work is needed in this rapidly evolving research
area.
One of the clinical consequences of the acute expo-
sure of dentin is the development of what has been
termed "hypersensitive dentin". The best example of this
condition is commonly seen following periodontal
surgery, where root surfaces may be planed free of
cementum, thus uncovering large surface areas of
dentin. The level of discomfort develops 5-7 days after
surgery, presumably due to dissolution of the smear
layer that was created on the dentin during root planing
(Kerns et al, 1991). Sensitivity then falls over a period of
2-4 weeks, presumably as a result of natural tubule
occlusion (Pashley, 1986). Anything that can reduce fluid
shifts across dentin should decrease dentin sensitivity.
Presumably, hypersensitive dentin would be hypercon-
ductive. Treatments aimed at making dentin hypocon-
ductive should decrease dentin sensitivity (Ahlquist et al,
1994). Several recent reviews (Curro, 1990; Orchardson et
al, 1994) of dentin sensitivity are available for those
interested in more detail.
(5) Nerve Reactions to Exposure of Dentin
Byers et al. (1990b) recently summarized the effects of
dentin exposure in rat molars. Shallow Class V cavity
preparations just into root dentin leave the primary
7(2):104-133 (1996)
odontoblast layer intact and cause only minimal inflam-
mation. Sensory nerves that exhibit immunoreactivity for
CGRP in the injured region release their CGRP and SP
when the dentin is cut. These agents are known to cause
an increase in pulpal blood flow and may activate local
fibroblast metabolism (Trantor et al, 1995). There is a
rapid expression of NGF by fibroblasts adjacent to the
odontoblast layer within hours of dentin exposure, sub-
siding within two days (Byers et al. 1992a,b). The NGF
seems to bind to p75-NGF receptor protein, which is co-
localized on CGRP-positive nerves (Fried and Risling,
1991). The NGF travels up the axon to the trigeminal gan-
glion, where it stimulates the neuron to increase the syn-
thesis of more CGRP (Lindsay et al, 1989, 1990) and, by
mechanisms not yet completely understood, to cause
branching and sprouting of nerve terminals beneath the
exposed dentin. This sprouting of nerves increases the
number of CGRP-positive nerve terminals in the region
of exposed dentin within 48 hrs, but the response disap-
pears within 7-10 days if there is no further irritation. The
temporal aspects of this reaction may be explained by
the time required for axonal transport.
The permeability properties of rat molar dentin are
unknown but are extremely important in view of all of the
work that has been done on nerve sprouting in response
to exposure of rat dentin. Rat molars contain enamel-
free regions of exposed dentin and enamel-covered
regions of occlusal dentin. The former serves as a chew-
ing surface as soon as the molars erupt, while the latter
rapidly becomes exposed with loss of occlusal enamel
due to attrition. Careful electron microscopy of the
enamel-free region indicates that many tubules become
occluded with mineral crystals and amorphous organic
material, although there were no quantitative data given
on the distribution of patent vs. occluded tubules in the
various regions of the teeth or as a function of post-erup-
tion time (Hawkinson and Eisenmann, 1983). Indeed, rat
dentin under occlusal function (attrition) may become
occluded with both organic and inorganic material in a
manner that is similar to that found in exposed dog
dentin (Hirvonen et al, 1992) or human dentin (Mendis
and Darling, 1979). If, however, the occlusal dentin of rat
molars exhibits as much regional variation as was found
in occlusal human dentin (Pashley et al., 1987), then one
might expect different pulpal responses beneath such
dentin. Most studies on rat molars utilize freshly
exposed cervical root dentin (Byers et al, 1990b). There
are at least two mechanisms that might be responsible
for the observed pulpal reactions. The simplest mecha-
nism would involve the vibratory and thermal stimuli
produced by the dental bur used to expose the dentin.
There would also be other hydrodynamic stimuli pro-
duced during such a procedure. The second mechanism
might involve the diffusion of bacterial substances
7(2):I04-133 i 1996) Crit Rev Oral Biol Med
across the freshly exposed dentin into the pulp. The use
of germ-free animals might evaluate the contribution of
the latter mechanism to the overall pulpal response. The
pulpal reactions to exposure of root dentin vary with
depth of removal of dentin and with time (Byers et al.,
1988). The tubule density and diameters in rat molar
dentin are such that rat dentin should be very permeable
(Forssell-Ahlberg et al, 1975). The observation that the
tubule diameters and numbers of tubules per unit area
are similar in the rat, cat, dog, and monkey (Forssell-
Ahlberg et al, 1975) suggests that major differences
should exist in the hydraulic conductance of the dentin
of these different species. Thin dentin has a higher
hydraulic conductance (and diffusive permeability) than
thick dentin. Rat molar root dentin is only about 300 jim
thick compared with cat, dog, and monkey dentin, whose
thicknesses range from 1 to 3 mm, depending on age.
Given that the diameters and tubule densities are simi-
lar, rat molar dentin should be much more permeable
than the thicker dentin of larger species. This may be
important in the interpretation of published experi-
ments. For instance, the inward diffusion of bacterial
substances in the nerve-sprouting experiments of Taylor
et al (1989), conducted in rat molars, was probably much
higher than that in experiments conducted by Ngassapa
(1991) in the dog or by Matthews (unpublished observa-
tions) in the cat. It would be of considerable interest to
know whether the permeability of that freshly exposed
rat radicular dentin decreases rapidly (hours) as was
found in dog dentin (Pashley, 1985), and how increases
or decreases in rat dentin permeability might modify the
temporal patterns of events that have been seen in rat
pulps following acute exposure of dentin (Byers et al,
1990b; Taylor and Byers, 1990; Byers and Taylor, 1993).
Presumably, large changes in permeability would pre-
cede formation of reparative dentin, which is a response
to odontoblast destruction. Perhaps immunospecific
staining of microscopic sections for albumin and fibrino-
gen as a function of exposure time (minutes to hours)
would demonstrate a progressive outward movement of
these proteins in the exposed tubules. The use of
restorative materials such as adhesive resins to seal the
freshly exposed dentin would be a useful control, but it
is uncertain how well the materials would seal rat dentin
or how much microleakage might occur around these
materials.
It is interesting to speculate that the stimulus for
sprouting is fluid shifts across dentin during cavity
preparation and immediately after exposure. If this is
true, then it should also occur in germ-free rats. This
fluid movement may cause massive release of CGRP
which, in turn, may stimulate local fibroblasts to produce
more NGF, which binds to NGF receptors on sensory
nerves. The ligand-receptor complex may become inter-
123
nalized and then travel upon the nerve to the cell body,
where it up-regulates more CGRP production. The newly
synthesized neuropeptides then travel down the axon to
the nerve terminals, where they are available for release,
proportional to the amount of stimulation that is
received. This type of speculative coupling of dentin
stimulation to neural response provides evidence of the
utility of the concept of a dynamic pulpo-dentin com-
plex.
Alternatively, the stimulus to fibroblasts to produce
and secrete NGF may be the appearance of bacterial sub-
stances (Pissiotis and Spangberg, 1994) or any one of a
number of cytokines produced in response to bacterial
substances. Separation of hydrodynamic fluid shifts from
bacterial substances could be done by repeating the
experiments in germ-free rats or by sealing the exposed
rat dentin with oxalates (Pashley and Galloway, 1985),
which block convective transport better than diffusive
transport.
Another alternative explanation is that as the rat
pulp becomes inflamed, pulpal tissue pressures rise,
causing more outwardly directed dentinal fluid flow,
which may achieve a magnitude high enough to block the
inward diffusion of bacterial products that sustain the
neurogenic inflammation. This reaction might subside
slowly as dentin became less permeable, over time, due
to adsorption of organic material in the tubules, deposi-
tion of mineral at the dentin surface, or formation of
more dentin at the pulpal termination of the tubules.
Then the hydraulic conductance of that dentin may
become too low to permit sufficient fluid shifts during
normal activity to sustain the neurogenic inflammation.
This may explain why the nerve sprouting stops and the
pulps often heal. Similar temporal events have been
described in humans (Lundy and Stanley, 1969) and in
monkeys (Warfvinge and Bergenholtz, 1986). Testing this
hypothesis would require the development of techniques
to measure the permeability of rat molor dentin as a
function of time after dentin exposure.
If cavity preparations are prepared more deeply into
the radicular dentin of rats, and the smear layers are
removed by acid-etching, there is a more severe pulpal
reaction involving destruction of some of the underlying
odontoblasts, monocytic infiltration, micro-abscess for-
mation, and displacement of vital pulp away from the
injured dentin by fluid accumulation (Byers, 1994).
Associated with this is extensive sprouting of CGRP-
immunoreactive nerve fibers in the surrounding pulp,
and in some cases, nerve sprouting fills the associated
root pulp (Taylor and Byers, 1990). This type of injury
usually heals within 3-5 weeks by extensive formation of
reparative dentin and a subsequent loss of nerve sprout-
ing. The reparative dentin is poorly innervated by CGRP-
positive cells.
124 Crit Rev Oral Biol Med
Heyeraas et al. (1996) have recently shown that
direct, long-term electrical stimulation of rat molars
caused not only increases in tissue pressure and pulpal
blood flow, but also accumulation of immunocompetent
CD43+-expressing cells beneath the odontoblast layer.
This did not occur when denervated teeth were similarly
stimulated, suggesting that neuropeptides are chemo-
tactic (Foster et al, 1992: Fristad et al, 1996). Heyeraas et
al. (1996) also measured the rate of dentin formation in
rat molars in control animals compared with those
whose molars were surgically denervated or chemically
denervated (multiple capsaicin injections). Six weeks
after these procedures were done, the rats were killed,
and the rate of dentin formation was assessed as the dis-
tance between Procion Brilliant red-stained uptake lines
in dentin following separate injections of the dye. The
same sections were stained for CGRP and SP.
Significantly less dentin formation was found in the
mesial pulp horn and the central pulpal floor of the den-
ervated rat first molars, but no difference at the mid-root
level, in the experimental animals relative to the con-
trols. This was correlated with reductions in the intensi-
ty of immunoreactivity for CGRP and SP, although the
authors believed that the correlations were not high. For
instance, in the capsaicin-treated animals, while there
was a reduction in staining for CGRP and SP nerve fibers,
they were still present, albeit at a reduced level. Yet, in
some of the capsaicin-treated animals, practically no
dentin formation was observed. The lack of effect on mid-
root dentin may reflect the relatively low innervation
density of that part of the tooth. Stiefel and Raab (1996)
also reported that post-natal injection of rats with cap-
saicin altered the formation and structure of dentin but
not of enamel or cementum. They observed more sclero-
sis of the apical third of root dentin in capsaicin-treated
animals compared with controls.
Thus, these recent reports clearly indicate the
dynamic nature of the interaction of pulpal nerves with
the oral environment and with other constituents of the
pulp. When dentin or pulps are exposed in denervated
teeth, the extent of subsequent pulpal destruction is
much greater than in innervated teeth (Byers and Taylor,
1993; Heyeraas et al, 1996). Presumably, the loss of
nerves prevents the axon reflexes that produce the
increases in pulpal blood flow, transudation of fluid from
vessels (Olgart, 1996), and increased outward fluid flow
(Matthews, 1996) that normally participate in the
defense of the pulp from exposure to the oral cavity. Also
lost would be whatever growth factors might be released
from nerves that would promote pulpal healing (Trantor
et al, 1995). Byers (1996) speculated that the profuse
innervation of teeth may normally provide little sensory
information from the pulp but instead may function as
an interactive neuroendocrine system first proposed by
7(2):104-133 (1996)
Silverman and Kruger (1987). In this paradigm, sensory
nerve activity is utilized for its paracrine effects on the
pulpo-dentin complex rather than for its afferent activity.
(C) Diagnostic Potential
Pashley (1990) suggested the use of dentinal fluid as an
indicator of changes in the composition of pulpal inter-
stitial fluid. Once dentin is exposed, there is a loss of
junctional complexes between odontoblasts. While this
is not normal, it may permit sufficient communication
between pulpal interstitial fluid and dentinal fluid which
may be of diagnostic advantage (Pashley et al, 1981;
Maita et al, 1991). For instance, if there is an acute
inflammatory response occurring under dentin, then
sampling dentinal fluid for lysosomal enzymes found in
PMNs may help detect pulpal inflammation and its reso-
lution (Rauschenberger, 1992; Rauschenberger and
Richardson, 1992a,b). For this approach to work, the
diagnostic markers must not bind to dentin. If SP was
released into pulpal interstitial fluids following an exper-
imental condition but then bound to dentin as the pul-
pal fluid flowed to the surface, then attempts to measure
SP in dentinal fluid would be futile. Complicating this
approach is the high probability that pulpal inflamma-
tion is usually due to dental caries. As bacteria invade
dentin, they produce sclerosis of the tubules, which
blocks fluid flow. However, it is possible that fluid can be
collected from adjacent normal dentin that is in equilib-
rium with inflamed pulp.
Another useful technique for sampling interstitial
fluid which has been used in neuropharmacology is the
technique of microdialysis (Lindefors et al, 1989; Rada et
al., 1993-, O'Brien et al, 1994; Goodis et al, 1996). In this
technique, a small hollow double-barreled catheter is
carefully positioned in the neuronal tissue of the CNS.
The terminal end of the catheter is covered with a mem-
brane that permits small molecules of interest, such as
neurotransmitters, to pass through the membrane into
the catheter while not permitting larger molecules, such
as hydrolytic enzymes or inhibitor molecules, to enter
the catheter. The inner portion of the double-lumen
catheter is continuously perfused with an appropriate
isotonic solution to bring the samples of CNS interstitial
fluid to the surface, where they can be collected in a frac-
tion collector for analysis. Microdialysis probes have
already been used by O'Brien et al. (1994) to sample the
interstitial fluid of periodontal flap sites. Similar studies
should be done in tooth pulps, both normal and
inflamed, as a function of time. If one seals a chamber in
a cavity prepared in dentin, one can perfuse that cham-
ber slowly over time and sample material (Potts et al,
1985) that diffuses across dentin, much as is done in
microdialysis experiments, using dentin as a "dialysis
membrane". This approach holds much promise for
7(2): 104-133 (1996) Crit Rev Oral Biol Med
future research.
(D) Therapeutic Potential
The concept of using the porosity of dentin as an avenue
for drug delivery to the pulp was suggested by Pashley
(1990). Under the proper conditions, it should be possi-
ble to introduce therapeutic agents to pulpal tissues via
dentinal tubules. The problems involved with this
approach have already been discussed (Ciarlone and
Pashley, 1992a). Matthews and his colleagues have
recently demonstrated the ability of 10 or 50% lidocaine
to anesthetize pulpal nerves following topical applica-
tions to dentin (Amess and Matthews, 1996). It is impor-
tant that the therapeutic agent not bind to dentin. For
instance, if CGRP binds to dentin, then its lack of effect
on pulpal nerves following topical application (Narhi,
unpublished observations) would not be surprising.
Rutherford et al. (1995) demonstrated the ability of
osteogenic protein-1 (OP-1) to promote reparative
dentin formation following topical application to dentin
in cavities prepared in monkey teeth. In the future, denti-
nal tubules may provide a route of administration of
many drugs, including growth factors and neuropeptides.
Perhaps in the future, dentists will begin their
restoration of dentin with the topical application of a
mixture of growth factors designed to seal the cut denti-
nal tubules on the pulpal side of dentin via production of
atubular dentin. After a few minutes of "loading" of the
tubules with growth factor(s), the cavity side of the cut
dentin would be sealed with adhesive resins (Tay et al,
1994) to provide a surface seal to the dentin. This dentin
would have an immediate surface seal, and, in several
weeks to months, the dentin would enjoy a double seal
from the pulpal side, which would protect the pulp and
prevent dentin sensitivity.
Some investigators (Pashley et al, 1992; Tay et al,
1994) suggest that clinicians consider sealing all pre-
pared dentin surfaces with adhesive resins in an attempt
to regain the hermetic seal that normal dentin enjoys.
One can marshall experimental support for such a pro-
cedure in view of the relatively poor performance of cal-
cium hydroxide liners (Cox and Suzuki, 1994). This can be
done using adhesive resins, although there may be sim-
pler approaches to occluding dentinal tubules (Pashley
and Galloway, 1985; Suge et al, 1995).
Recently, an intricate network of contact between
pulpal microvessels and nerve fibers, provided by the
cellular processes of dendritic cells (specialized antigen-
presenting cells), was demonstrated by Jontell (1996). He
also showed that low concentrations (10~9 mol/L) of SP
stimulated T-cell mitogenesis, in vitro, when pulp cells
were used as accessory cells, while CGRP produced inhi-
bition under the same conditions. These findings indi-
cate that immunocompetent cells, the nervous system,
and the pulpal vasculature may interact and provide sig-
125
nificant defensive functions at multiple levels in the
defense of the pulp.
Conclusion
The concept of the pulpo-dentin complex continues to
be useful when the pathobiology of dentin and pulp is
considered. Developmental^, pulpal cells produce
dentin, nerves, and blood vessels. Once formed,
although dentin and pulp have different structures and
compositions, they continue to react to stimuli as a func-
tional unit. Exposure of dentin through attrition, trauma,
or caries produces profound pulpal reactions that tend
to reduce dentin permeability and stimulate formation of
additional dentin. These reactions are brought about by
changes in fibroblasts, nerves, blood vessels, odonto-
blasts, leukocytes, and the immune system. Recent dis-
coveries of the effects of nerves on pulpal blood vessles
and vice versa have produced a new appreciation for the
interaction of these two systems in response to stimuli
applied to dentin.
Too often, the individual components of the pulpo-
dentin complex are studied in isolation from each other.
This is often necessary for technical or experimental rea-
sons. However, it is becoming clear that the individual
components are very interactive and that each modifies
the activity of the other. Thus, where we used to speak of
neurovascular interactions, we now must consider
neuro-vascular-immuno-odontoblast (Jontell, 1996) inter-
actions. Although the pulpo-dentin complex is difficult
to study, it offers a unique environment for research. As
more scientists apply multi-disciplinary techniques to
the problems associated with the pulpo-dentin complex,
new insight will develop which will be of therapeutic use
in dentistry.
Acknowledgments
The author is grateful to Shirley \ohnston for outstanding secretarial
support. This work was supported, in part, by grant DE 06427 from
the NIDR and by the Medical College of Georgia Biocompatibility
Program.
REFERENCES
Aars H, Brodin P, Andersen E (1993). A study of choliner-
gic and a-adrenergic components in the regulation of
blood flow in the tooth pulp and gingiva in man. Acta
Physiol Scand 148:441-447.
Addy M, Absi ED, Adams D (1987). Dentin hypersensitiv-
ity: The effects in vitro of acids and dietary substances
on root-planed and burred dentine. J Clin Periodont
14:274-279.
Ahlquist MC, Edwall LGA, Franzen OG, Haegerstam G
(1984). Perception of pulpal pain as a function of
126 Crit Rev Oral Biol Ued
intradental nerve activity. Pain 19:353-366.
Ahlquist M, Franzen O, Coffey 1, Pashley DH (1994).
Dental pain evoked by hydrostatic pressures applied
to exposed dentin in man. A test of the hydrodynam-
ic theory of dentin sensitivity. I Endodont 20:130-134.
Akimoto T (1991). Study of adhesion of MMA-TBBO resin
to dentin. ]pn Soc Dent Mater Devices 10:42-54.
Amess TR, Matthews B (1996). Effects of topical ligno-
caine on exposed dentine in the cat. In: Dentin/pulp
complex. Shimono M, Takahashi K, editors. Tokyo,
lapan: Quintessence Publishing.
Anderson E, Aars H, Brodin P (1994). Effects of cooling
and heating of the tooth on pulpal blood flow in man.
Endod Dent Traurnatol 10:256-259.
Anneroth G, Bang G (1972). The effect of allogenic denv
ineralized dentin as a pulp capping agent in )ava
monkeys. Odontol Revy 23:315-328.
Avery JK, Chiego D] (1990). The cholinergic system and
the dental pulp. In: Dynamic aspects of dental pulp.
Inoki R, Kudo T, Olgart L, editors. New York: Chapman
and Hall, pp. 297-331.
Avery IK, Cox CF, Chiego DJ (1980). Presence and location
of adrenergic nerve endings in the dental pulp of
mouse molars. Anat Rec 198:59-71.
Bergenholtz G (1977). Effects of bacterial products on
inflammatory reactions in the dental pulp. Scand J
Dent Res 85:122-129.
Bergenholtz G (1981). Inflammatory responses of the
dental pulp to bacterial irritation. J Endodont 7:100-
104.
Bergenholtz G (1996). Influence of dentin permeability of
microbial challenges to the dentin/pulp complex. In:
Dentin/pulp complex. Shimono M, Takahashi K, edi-
tors. Tokyo, lapan: Quintessence Publishing.
Bergenholtz G, Warfvinge J (1982). Migration of leuko-
cytes in dental pulp in response to plaque bacteria.
Scand I Dent Res 90:354-362.
Bergenholtz G, Mjor IA, Cotton WR, Hanks CT, Kim S,
Torneck CD, et al. (1985). Consensus report. ) Dent Res
64(SpecIss):631-633.
Bergenholtz G, Jontell M, Tuttle A, Knutsson G (1993).
Inhibition of serum albumin flux across exposed den-
tine following conditioning with GLUMA primer, glu-
taraldehyde or potassium oxalate. I Dent 21:220-227.
Bergenholtz G, Knutsson G, Okiji T, Jontell M (1996).
Albumin flux across dentin of young human premo-
lars following temporary exposure to the oral envi-
ronment. In: Dentin/pulp complex. Shimono M,
Takahashi K, editors. Tokyo, Japan: Quintessence
Publishing.
Berger RL, Byers MR (1983). Dental nerve regeneration in
rats. II. Autoradiographic studies of nerve regenera-
tion to molar pulp and dentin. Pain 15:359-370.
Bernhard GW, Shin C (1990). The osteogenic stimulating
effect of neuroactive calcitonin gene-related peptide.
7 ( 2 ) 1 0 4 - 1 3 3 (1996)
 Peptides 1 1:625-632.
Bishop MA (1987). An investigation of pulp capillaries
and tight junctions between odontoblasts in cats.
Anat Embruol 177:131-138.
Bishop MA (1992). Extracellular fluid movement in the
pulp: the pulp/dentin permeability barrier. Proc Finn
Dent Soc 88(Suppl l):331-336.
Blake GC (1958). The peritubular translucent zones in
human dentine. Br Dent 1 104:57-59.
Blandy AA (1850). On the sensibility of teeth. Am ) Dent
Sri 1:22-28.
Bowen RL, Rodriguez MS (1962). Tensile strength and
modulus of elasticity of tooth structure and several
restorative materials. ] Am Dent Assoc 64:378-387.
Brannstrom M (1992). Etiology of dentin hypersensitivi-
ty. Proc Finn Dent Soc 88(Suppl 1):7-13.
Brannstrom M, Linden LA, Astrom A (1967). The hydro-
dynamics of the dental tubule and of pulp fluid. Caries
Res L310-312.
Brown AC, Barrow BL, Gadd GN, Van Hassel H) (1969).
Tooth pulp transcapillary osmotic pressure in the
dog. Arch Oral Biol 14:491-502.
Browne RM, Tobias RS (1986). Microbial microleakage
and pulpal inflammation: a review. Endod Dent
Traumatol 2:177-183.
Butler WT (1995). The nature and functional significance
of dentin extracellular matrix proteins. \nt ) Dev Biol
39:169-179.
Byers MR (1980). The development of sensory innerva-
tion in dentin. ) Comp Neurol 191:413-428.
Byers MR (1984). Dental sensory receptors. \nt Rev
Neurobiol 25:39-44.
Byers MR (1990). Neuroanatomical studies of dental
innervation and pain. In: Experimental endodontics.
Spangberg L, editor. Boca Raton, FL: CRC Press, pp.
77-113.
Byers MR (1994). Dynamic plasticity of dental sensory
nerve structure and cytochemistry. Arch Oral Biol
39(Suppl):13S-21S.
Byers MR (1996). Neuropeptide irnmunoreactivity in den-
tal sensory nerves: Variations related to primary
odontoblast function and survival. In: Dentin/pulp
complex. Shimono M, Takahashi K, editors. Tokyo,
lapan: Quintessence Publishing.
Byers MR, Matthews B (1981). Autoradiographic demon-
stration of ipsilateral and contralateral sensory nerve
endings in cat dentin, pulp and periodontium. Anat
Rec 201:249-260.
Byers MR, Sugaya A (1995). Odontoblast processes in
dentin revealed by fluorescent Di-I. 1 Histochem
Cytochem 43:159-168.
Byers MR, Taylor PE (1993). Effect of sensory denervation
on the response of rat molar pulp to exposure injury.
) Dent Res 72:613-618.
Byers MR, Narhi MV, Mecifi KB (1988). Acute and chron-
7(2):104-133 (1996) Crit Rev Oral Biol Med
ic reactions of dental sensory nerves to hydrodynam-
ic stimulation or injury. Anat Rec 221:872-883.
Byers MR, Schatteman GC, Bothwell M (1990a). Multiple
functions of NGF receptor in developing, aging and
injured rat teeth are suggested by epithelial, mes-
enchymal and neural immunoreactivity. Development
109:461-471.
Byers MR, Taylor PE, Khayat BG, Kimberly CK (1990b).
Effects of injury and inflammation on pulpal and peri-
apical nerves. ) Endodont 16:78-84.
Byers MR, Swift ML, Wheeler EF (1992a). Reactions of
sensory nerves to dental restorative procedures. Proc
Finn Dent Soc 88(Suppl l):73-82.
Byers MR, Wheeler EF, Bothwell M (1992b). Altered
expression of NGF and p75-NGF-receptor mRNA by
fibroblasts of injured teeth precedes sensory nerve
sprouting. Growth Factors 6:41-52.
Byers MR, Kvinnsland I, Bothwell M (1992c). Analysis of
low affinity nerve growth factor receptor during pulpal
healing and regeneration of myelinated and unmyeli-
nated axons in replanted teeth.) Comp Neurol 326:470-
484.
Chiego DJ (1992). An ultrastructural and autoradiograph-
ic analysis of primary and replacement odontoblasts
following cavity preparation and wound healing in the
rat molar. Proc Finn Dent Soc 88(Suppl):243-256.
CiarloneAE, Pashley DH (1992a). Medication of the den-
tal pulp: a review and proposals. Endodont Dent
Traumatol 8:1-5.
Ciarlone AE, Pashley DH (1992b). Permeability of root
dentin to epinephrine released from gingival retrac-
tion cord. Oper Dent 17:106-111.
Ciucchi B, Bouillaguet S, Holz ), Pashley DH (1995).
Dentinal fluid dynamics in human teeth, in vivo. )
Endodont 21:191-194.
Clark NG (1970). The morphology of the reparative den-
tine bridge. Oral Surg Oral Ued Oral Pathol 29:746-752.
Couve E (1986). Ultrastructural changes during the life
cycle of human odontoblasts. Arch Oral Biol 31:643-
651.
Cox CF (1987). Biocompatibility of dental materials in
the absence of bacterial infection. Oper Dent 12:146-
152.
Cox CF, Suzuki S (1994). Re-evaluating pulp protection:
Calcium hydroxide liners versus cohesive hybridiza-
tion. I Am Dent Assoc 125:823-831.
Cox CF, Bergenholtz G, Heys DR, Syed M, Fitzgerald M,
Heys R) (1985). Pulp capping of dental pulp mechan-
ically exposed to oral microflora: A 1-2 year observa-
tion of wound healing in the monkey. J Oral Pathol
14:156-158.
Cox CF, Keall CL, Keall H), Ostro E, Bergenholtz G (1987).
Biocompatibility of surface sealed dental materials
against exposed pulps. ) Prosthet Dent 57:1-8.
Cox CF, Subay RK, Ostro E, Suzuki S, Suzuki S (1996).
127
 Tunnel defects in dentin bridges: Their formation fol-
lowing direct pulp capping. Oper Dent 21:4-11.
Craig RG (1993). Restorative dental materials. 9th ed. St.
Louis: C.V. Mosby Co., p. 64.
Curro FA (1990). Tooth hypersensitivity. Dent Clin N Am
34:403-581.
D'Souza RN, Litz M (1996). Odontoblast gene regulation
during development and repair. In: Dentin/pulp com-
plex. Shimono M, Takahashi K, editors. Tokyo, Japan:
Quintessence Publishing.
DaiX-F, Ten CateAR, Limeback H (1991). The extent and
distribution of intratubular collagen fibrils in human
dentin. Arch Oral Biol 36:775-778.
Dourda AO, Moule A), Young WG (1994). A morphomet-
ric analysis of the cross-sectional area of dentine
occupied by dentinal tubules in human third molar
teeth. \nt EndodontJ 27:184-189.
Duncan RL, Turner CH (1995). Mechanotransduction and
functional response of bone to mechanical strain.
Calcif Tissue \nt 57:344-358.
Eda S, Saito T (1978). Electron microscopy of cells dis-
placed into the dentinal tubules due to dry cavity
preparation. I Oral Pathol 7:326-335.
Edwall L, Olgart L (1977). A new technique for recording
intradental sensory nerve activity in man. Pain 3:121-
126.
Eick JD, Gwinnett AJ, Pashley DH (1996). Current con-
cepts on adhesion to dentin. Crit Rev Oral Biol Med (in
press).
Fearnhead RW (1957). Histological evidence for the
innervation of human dentin. J Anat 91:267-277.
Fearnhead RW [\96\\r- The neurohistology of human
teeth. Proc R Soc Med 54:877-884.
Finkelman RD, Mohan S, Jennings JC, Taylor AK, Jepsen S,
Baylink DJ (1990). Quantitation of growth factors IGF-
1, SGF/IGF-II and TGF-£ in human dentin. J Bone Min
Res 5:717'-723.
Fitzgerald M (1979). Cellular mechanisms of dentinal
bridge repair using 3H-thymidine. J Dent Res 58:2198-
2206.
Fogel HM, Marshall FJ, Pashley DH (1988). Effects of dis-
tance from the pulp and thickness on the permeabili-
ty of human radicular dentin. J Dent Res 67:1381-1385.
Forssell-Ahlberg K, Edwall L (1977). Influence of local
insults on sympathetic vasoconstrictor control in
feline dental pulp. Acta Odontol Scand 35:103-110.
Forssell-Ahlberg K, Brannstrom M, Edwall L (1975). The
diameter and number of dentinal tubules in rat, cat,
dog and monkey. A comparative scanning electron
microscope study. Acta Odontol Scand 33:243-250.
Fosse G, Saele PK, Eide R (1992). Numerical density and
distributional pattern of dentin tubules. Acta Odontol
Scand 50:201-210.
Foster CA, Mandak B, Kromer E, Rot A (1992). Calcitonin
gene-related peptide is chemotactic for human T lym-
128 Crit Rev Oral Biol Med
phocytes. Ann NY Acad Sci 657:397-404.
Frank RM, Nalbandian J (1989). Structure and ultrastruc-
ture of dentin. In: Teeth. Berkovitz BKB, Boyde A,
Frank RM, Hohlings HJ, Moxham BJ, Nalbandian J, et
al., editors. New York: Springer-Verlag, p. 212.
Fried K (1992). Changes in pulp nerves with aging. Proc
Finn Dent Soc 88(Suppl):517-528.
Fried K, Hildebrand C (1981). Pulpal axons in developing,
mature and aging feline permanent incisors. A study
by electron microscopy. J Comp Neurol 203:23-36.
Fried K, Risling M (1991). Nerve growth factor receptor-
like immunoreactivity in primary and permanent
canine tooth pulps of the cat. Cell Tissue Res 264:321-
328.
Fristad I, Heyeraas KJ, Jonsson R, Kvinnsland I (1996).
Sensory denervation—effect on inflammation and
neuropeptide expression in young rat molars. In.-
Dentin/pulp complex. Shimono M, Takahashi K, edi-
tors. Tokyo, Japan: Quintessence Publishing.
Fujitani M, Otsuki M, Inokoshi S, Shimada Y, Yamada T,
Takatsu T (1996). Effects of RDT and smear layer
removal on pulpal healing. In: Dentin/pulp complex.
Shimono M, Takahashi K, editors. Tokyo, Japan:
Quintessence Publishing.
Garberoglio P (1994). The ratio of the densities of denti-
nal tubules on the cervical and axial walls in cavities.
Ouintessence \nt 25:49-52.
Garberoglio R, Brannstrom M (1976). Scanning electron
microscopic investigation of human dentinal tubules.
Arch Oral Biol 21:355-362.
Gazelius B, Edwall B, Olgart L, Lundberg JM, Hokfelt T,
Fisher JA (1987). Vasodilating effects and co-existence
of calcitonin gene-related peptide (CGRP) and sub-
stance P in sensory nerves of cat dental pulp. Acta
Physiol Scand 130:33-40.
Goldberg M, Lasfargues J-J (1995). Pulp-dentinal complex
revisited. J Dent 23:15-20.
Goodis HE, Sacki K, White JM (1996). Microdialysis as a
unique method of dental pulp fluid collection
(abstract). I Dent Res 75(Spec Iss):270.
Gunji T (1982). Morphological work on the sensitivity of
dentin. Arch Histol ]pn 45:45-67.
Gysi A (1900). An attempt to explain the sensitiveness of
dentine. Br J Dent Sci 43:865-868.
Hannington-Kiff JG, Dunne SM (1993). Topical guanethi-
dine relieves dentinal hypersensitivity and pain. J R
Soc Med 86:514-515.
Hargreaves KM, Bowles WR, Engelstad M, Roszkowski
MT, Richardson JD, Jackson DL (1996). Regulation of
neuropeptide release from pulp. In: Dentin/pulp com-
plex. Shimono M, Takahashi K, editors. Tokyo, Japan:
Quintessence Publishing.
Hawkinson RW, Eisenmann DR (1983). Electron
microscopy of dentinal tubule sclerosis in the enamel-
free region of the rat molar. Arch Oral Biol 28:409-414.
7(2):104~133 (1996)
Heyeraas KJ (1989). Pulpal hemodynamics and intersti-
tial fluid pressure. Balance of transvascular fluid
transport. ] Endodont 15:468-472.
Heyeraas KJ, Kim S, Raab WH-M, Byers MR, Liu M (1994).
Effects of electrical tooth stimulation on blood flow,
interstitial fluid pressure, and substance P and CGRP-
immunoreactive nerve fibers in the low complaint cat
dental pulp. Microvasc Res 47:324-343.
Heyeraas KJ, Jacobsen ED, Fristad 1, Raab WHM (1996).
Vascular and immunoreactive nerve fiber reactions in
the pulp after stimulation and denervation. In:
Dentin/pulp complex. Shimono M, Takahashi K, edi-
tors. Tokyo, Japan: Quintessence Publishing.
Hildebrand C, Fried K, Tuisku U, Johansson CS (1995).
Teeth and tooth nerves. Prog Neurobiol 45:165-222.
Hirayama A, Yamada M, Miake K (1985). Analytical elec-
tron microscope studies on the dentinal tubules of
human deciduous teeth (abstract). J Dent Res 64:743.
Hirvonen T, Ngassapa D, Narhi M (1992). Relation of
dentin sensitivity to histological changes in dog teeth
with exposed and stimulated dentin. Proc Finn Dent Soc
88(Suppl 1): 133-142
Holland GR (1976). The extent of the odontoblast
process in the cat. J Anat 120:133-149.
Holland GR (1985). The odontoblast process: Form and
function. J Dent Res 64(Spec Iss):499-514.
Hoppe WF, Stuben J (1965). Uber die messung des volu-
mens der dentinkanalchen und uber des kanalvolu-
mens zum gesamtdentinvolumen. Stoma 65:38-45.
Horiuchi H, Matthews B (1973). \n vitro observations on
fluid flow through human dentin caused by pain-pro-
ducing stimuli. Arch Oral Biol 18:275-294.
Hovgaard O, Larsen MJ, Fejerskov O (1991). Tooth hyper-
sensitivity in relation to quality of restorations
(abstract). ) Dent Res 70(Spec Iss):474.
Hughes NP, Littlewood D, Macpherson JV, Beeston MA,
Unwin PR (1996). Direct measurements of fluid flow
rate in dentinal tubules using scanning electro-chem-
ical microscopy. In: Dentin/pulp complex. Shimono
M, Takahashi K, editors. Tokyo, Japan: Quintessence
Publishing.
Hussey DL, Biagioni PA, Lamey P-J (1995).
Thermagraphic measurement of temperature change
during resin composite polymerization in vivo. J Dent
23:267-271.
Inoue T, Shimono M (1992). Repair dentinogenesis fol-
lowing transplantation into normal and germ-free
animals. Proc Finn Dent Soc 88|Suppl): 183-194.
Inoue H, Taniguchi K, Okamura K, Izumi T, Tamaura N,
Kajiwara S-I, et al. (1995). Ultrastructural relation
between nerve terminals and dentin bridge formation
after pulpotomy in human teeth. Arch Oral Biol 40:669-
675.
Johnson DC (1990). Innervation of teeth: developmental
aspects In: Dynamic aspects of dental pulp. Inoki R,
7(2) 104-1 33 (1996) Crit Rev Oral Biol Med
Kudo T, Olgart L, editors. New York: Chapman and
Hall, pp. 3-28.
Jontell M (1996). Interaction between immunocompetent
cells, nerves and vasculature in the dental pulp. In:
Dentin/pulp complex. Shimono M, Takahashi K, edi-
tors. Tokyo, Japan: Quintessence Publishing.
Kakehashi S, Stanley HR, Fitzgerald RJ (1965). The effects
of surgical exposure of dental pulps in germ-free and
conventional laboratory rats. Oral Surg 20:340-349.
Kerns DG, Scheldt MJ, Pashley DH, Horner JA, Strong SL,
Van Dyke TE (1991). Dentinal tubule occlusion and
root hypersensitivity. J Periodont 62:421-428.
Kim S, Dorscher-Kim J (1996). Contribution of the low
compliance environment in the pathobiology of the
pulp. In: Dentin/pulp complex. Shimono M, Takahashi
K, editors. Tokyo, Japan: Quintessence Publishing.
Kimberly CL, Byers MR (1988). Response of nerve fibers
containing calcitonin gene-related peptide to inflam-
mation of rat molar pulp and periodontium. Anat Rec
222:289-300.
Kinney JH, Balooch M, Haupt DL, Marshall SJ, Marshall
GW (1995). Mineral distribution and dimensional
changes in human dentin during demineralization. J
Dent Res 74:1179-1184.
Kinney JH, Balooch M, Marshall SJ, Marshall GW, Weihs
TM (1996). Hardness and Young's modulus of per-
itubular and intertubular dentine. Arch Oral Biol 41:9-13.
Knutsson G, Jontell M, Bergenholtz G (1994).
Determination of plasma proteins in dentinal fluid
from cavities prepared in healthy young human teeth.
Arch Oral Biol 39:185-190.
Kobayashi C, Yoshida H (1981). Wound healing of the
exposed pulp in germ-free and conventional rats. ]pn
J Conserv Dent 24:747-753.
Kollar EJ, Lumsden AGS (1979). Tooth morphogenesis:
the role of the innervation during induction and pat-
tern development. J Biol Buccale 7:49-60.
Koutsi V, Noonan RG, Horner JA, Simpson MD, Matthews
WG, Pashley DH (1994). The effect of dentin depth on
the permeability and ultrastructure of primary
molars. Ped Dent 16:29-35.
Kramer IRH (1955). The relationship between dentine
sensitivity and movements in the contents of the
dentinal tubules. Br DentI 98:391-392.
LaFleche RG, Frank RM, Steuer P (1985). The extent of
the human odontoblast process as determined by
transmission electron microscopy: the hypothesis of
a retractable suspensor system. J Biol Buccale 13:293-
305.
Langeland K, Dowdon WE, Tronstad L, Langeland L
(1971). Human pulp changes of iatrogenic origin. Oral
Surg Oral Med Oral Pathol 32:943-980.
LeGeros RZ (1991). Calcium phosphate in oral biology
and medicine Vol. 15. Monogr Oral Sci. Myers HM,
editor. New York: Karger.
129
Lehman ML (1967). Tensile strength of human dentin. )
Dent Res 46:197-201.
Lewis T (1927). The blood vessels of the human skin and
their responses. London: Shaw and Sons.
LindeA, Goldberg M (1993). Dentinogenesis. Crit Rev Oral
Biol Ued 45:679-728.
Linde A, Lundgren T (1995). From serum to the mineral
phase. The role of the odontoblast in calcium trans-
port and mineral formation. \nt I Dev Biol 39:213-222.
Lindefors N, Amberg G, Ungerstedt U (1989).
Intracerebral microdialysis. I. Experimental studies
on diffusion kinetics. I Pharmacol Meth 22:141-156.
Lindsay RM, Lockett C, Sternberg J, Winter 1 (1989).
Neuropeptide expression in cultures of adult sensory
neurons: modulation of substance P and calcitonin
gene-related peptide levels by nerve growth factor.
Neuroscience 33:53-65.
Lindsay RM, Shooter EM, Radeke MJ (1990). Nerve
growth factor regulates expression of the nerve
growth factor receptor gene in adult sensory neurons.
Eur] Neurosci 2:389-396.
Lohinai Z, Balla I, Marczis I, Vass Z, Kovach AGB (1995).
Evidence for the role of nitric oxide in the circulation
of the dental pulp. J Dent Res 74:1501 -1506.
Lundy T, Stanley HR (1969). Correlation of pulpal
histopathology and clinical symptoms in human
teeth subjected to experimental irritation. Oral Surg
27:187-201.
Luthman J, Luthman D, Hokfelt T (1992). Occurrence and
distribution of different neurochemical markers in the
human dental pulp. Arch Oral Biol 37:193-208.
Magloire Hr Bouvier M, Joffre A (1992). Odontoblast
response under carious lesions. Proc Finn Dent Soc
88(Suppl l):257-274.
Maita E, Simpson MD, Tao L, Pashley DH (1991). Fluid
and protein flux across the pulpodentin complex of
the dog, in vivo. Arch Oral Biol 36:103-110.
Marshall GW (1993). Dentin microstructure and charac-
terization. Quintessence \nt 24:606-617.
Marshall GW, Marshall SI, Kinney JH, Balooch M (1996).
Dentin structure and properties. I Dent (in press).
Matthews B (1992). Sensory physiology: a reaction. Proc
Finn Dent Soc 88(Suppl l):529-532.
Matthews B, Vongsavan N (1994). Interactions between
neural and hydrodynamic mechanisms in dentine and
pulp. Arch Oral Biol 39(Suppl):87S-95S.
Matthews R (1996). The functional properties of
intradental nerves. In: Dentin/pulp complex. Shimono
M, Takahashi K, editors. Tokyo, Japan: Quintessence
Publishing.
Mendis BRRN, Darling AI (1979). A scanning electron
microscopic and micrographic study of closure of
human coronal dentinal tubules related to occlusal
attrition and caries. Arch Oral Biol 24:725-733.
Miyakoshi S, Inoue T, Shimono M (1994). Interface inter-
130 Crit Rev Oral Biol Ued
actions of 4-META/MMA-TBB resin and dental pulp
(extended abstract). Arch Oral Biol 39(Suppl):147S.
Mjor IA (1985). Dentin-predentin complex and its perme-
ability: Pathology and treatment overview. I Dent Res
64(SpecIss):621-627.
Mjor IA, Fejerskov O (1979). Histology of the human
tooth. 2nd ed. Copenhagen: Munksgaard.
Mohamed SS, Atkinson ME (1983). A histological study
of the innervation of developing mouse teeth. I Anat
136:735-749.
Nakabayashi N (1992). The hybrid layer: a resin-dentin
composite. Proc Finn Dent Soc 88(Suppl 1):321-330.
Nakashima M (1989). Dentin induction by implants of
autolyzed antigen-extracted allogenic dentin on
amputated pulps of dogs. Endod Dent Traumatol 5:279-
286.
Nakashima M (1990). An ultra structural study of the dif-
ferentiation of mesenchymal cells in implants of allo-
genic dentine matrix on the amputated dental pulp of
the dog. Arch Oral Biol 35:277-281.
Narhi M (1978). Activation of dental-pulp nerves of the
cat and the dog with hydrostatic pressure. Proc Finn
Dent Soc 74(Suppl 1):1-61.
Narhi M (1985). The characteristics of intradental senso-
ry units and their responses to stimulation. I Dent Res
64(SpecIss):564-571.
Narhi M, Yamamoto H, Ngassapa D (1996). Function of
intradental nociceptors in normal and inflamed teeth.
In: Dentin/pulp complex. Shimono M, Takahashi K,
editors. Tokyo, Japan: Quintessence Publishing.
Ngassapa D (1991). Dentine sensitivity.- Factors influenc-
ing intradental nerve activation in dog teeth (PhD
thesis). Kuopio, Finland: University of Kuopio.
Nilsson J, von Euler AM, Dalsgaard C-J (1985).
Stimulation of connective tissue cell growth by sub-
stance P and substance K. Nature 315:61-63.
Nyborg K, Brannstrom M, Nyborg H (1969). Points in the
experimental study of pulpal responses to restorative
materials. Odontol Tidskr 77:421-426.
O'Brien T, Wolff L, Roszkowski M, Hargreaves K (1994).
Tissue levels of PGE2, LTB4 and pain after periodon-
tal surgery (abstract). J Dent Res 73(Spec Iss):379.
Ohshima H (1990). Ultrastructural changes in odonto-
blasts and pulp capillaries following cavity prepara-
tion in rat molars. Arch Histol Cytol 53:423-428.
Olgart L (1990). Concluding remarks. In: Dynamic aspects
of dental pulp. Inoki R, Kudo T, Olgart L, editors. New
York: Chapman and Hall, pp. 363-366.
Olgart L (1992). Involvement of sensory nerves in hemo-
dynamic reactions. Proc Finn Dent Soc 88(Suppl l):403-
410.
Olgart L (1996). Neurogenic components of pulp inflam-
mation. In: Dentin/pulp complex. Shimono M,
Takahashi K, editors. Tokyo, Japan: Quintessence
Publishing.
7(2):104-133 (1996)
Olgart L, Gazelius B (1977). Effect of adrenalin and fely-
pressin (octapressin) on blood flow and sensory
nerve activity in the tooth. Ada Odontol Scand 35:69-75.
Olgart L, Kerezoudis NP (1994). Nerve pulp interactions.
Arch Oral Biol 39(Suppl):47S-54S.
Olgart L, Edwall L, Gazelius B (1991). Involvement of
afferent nerves in pulpal blood flow reactions in
response to clinical and experimental procedures in
the cat. Arch Oral Biol 36:575-581.
Olsson S, 0ilo G (1993). The structure of dentin surfaces
exposed for bond strength measurements. Scand 1
Dent Res 101:180-184.
Onoe N (1994). Study of adhesive bonding systems as a
direct pulp capping agent. }pn ) Conserv Dent 37:429-
466.
Orchardson R, Gangarosa LP, Holland GR, Pashley DH,
Trowbridge HO, Ashley FP, et al (1994). Consensus
report. Dentine hypersensitivity into the 21st century.
Arch Oral Biol 39(Suppl):l 13S-119S.
Pashiey DH (1979). The influence of dentin permeability
and pulpal blood flow on pulpal solute concentra-
tion. | Endodont 5:355-361.
Pashley DH (1984). Smear layer: physiologic considera-
tions. Over Dent (Suppl 3): 13-29.
Pashley DH (1985). Dentin-predentin complex: physio-
logical overview. ) Dent Res 64(Spec Iss):613-620.
Pashley DH (1986). Dentin permeability, dentin sensitiv-
ity and treatment through tubule occlusion.) Endodont
12:465-474.
Pashley DH (1989). Dentin: A dynamic substrate: a
review. Scan Microsc 3:161-176
Pashley DH (1990). Dentin permeability: Theory and
practice. In: Experimental endodontics. Spangberg L,
editor. Boca Raton, FL: CRC Press, pp. 19-49.
Pashley DH (1991). Clinical correlations of dentin struc-
ture and function. ) Prosthet Dent 66:777-781.
Pashley DH (1992). Dentin permeability and dentin sen-
sitivity. Proc Finn Dent Soc 88(Suppl 1):215-224.
Pashley DH (1994). Dentin permeability and its role in
the pathobiology of dentine sensitivity. Arch Oral Biol
39(Suppl):73S-80S.
Pashley DH, Carvalho RM (1997). Dentine permeability
and adhesive bonding. ) Dent (in press).
Pashley DH, Galloway SE (1985). The effects of oxalate
treatment of the smear layer of ground surfaces of
human dentine. Arch Oral Biol 30:731-737.
Pashley DH, Matthews WG (1993). The effects of outward
forced convective flow on inward diffusion in human
dentine, in vitro. Arch Oral Biol 38:577-582.
Pashley DH, Pashley EL (1991). Dentin permeability and
restorative dentistry. Am I Dent 4:5-9.
Pashley DH, Livingston MI, Reeder OW, Horner JA (1978).
Effects of the degree of tubule occlusion on the per-
meability of human dentine, in vitro. Arch Oral Biol
23:1 127-1133.
7(21:104-133 (1996) Crit Rev Oral Biol Med
Pashley DH, Nelson R, Williams EC, Kepler EE (1981).
Use of dentine-fluid protein concentrations to mea-
sure pulp capillary reflection coefficients in dogs. Arch
Oral Biol 26:703-706.
Pashley DH, Thompson SM, Stewart FP (1983). Dentin
permeability: effects of temperature on hydraulic con-
ductance. ) Dent Res 62:456-459.
Pashley DH, Galloway SE, Stewart FP (1984). Effects of
fibrinogen in vivo on dentine permeability in the dog.
Arch Oral Biol 29:725-728.
Pashley DH, Andringa HJ, Derkson GD, Derkson ME,
Kalathoor S (1987). Regional variability in dentine
permeability. Arch Oral Biol 32:519-523.
Pashley DH, Tao L, Boyd L, King GE, Horner JA (1988).
Scanning electron microscopy of the substructure of
smear layers in human dentine. Arch Oral Biol 33:265-
270.
Pashley DH, Kim S, Hanks CT, Trowbridge HO, Holland R,
Walton RE, et al. (1992a). Consensus report. Proc Finn
Dent Soc 88(Suppl 1):571-581.
Pashley DH, Horner JA, Brewer PD (1992b). Interactions
of conditioners on the dentin surfaces. Oper Dent
(Suppl 5):137-15O.
Pashley EL, Comer R, Simpson MD, Horner JA,
Caughman F, Pashley DH (1992). Dentin permeability:
Sealing the dentin in crown preparations. Oper Dent
17:13-20.
Pinero GJ, Parker S, Rundus V, Hertzberg EL, Minkoff R
(1994). Immunolocalization of connexin 43 in the
tooth germ of the neonatal rat. Uistochem ] 26:765-770.
Pissiotis E, Spangberg L (1994). Dentin permeability to
bacterial proteins in vitro. I Endodont 20:118-122.
Potts TV, Cunningham T, Finkelstein M), Silverberg-
Strumfeld L (1985). The movement of radioactive
molecules across dentine in vivo in the dog. Arch Oral
Biol 30:353-357.
Prati C, Pashley DH (1992). Dentin wetness, permeability
and bond strength of adhesive systems. Am ) Dent
5:33-38.
Prati C, Erickson R, Tao L, Simpson MD, Pashley DH
(1991). Measurement of dentin permeability and wet-
ness using the Periotron device. Dent Mater 7:268-273.
Raab WH-M (1989). Zur Entstehung des plasmaex-
travasates in der zahnpulpa. Dtsch Zahnarztl Z 44:686-
Raab WH-M (1992). Temperature related changes in pul-
pal microcirculation. Proc Finn Dent Soc 88(Suppl):469-
479.
Rada P, Parada M, Hernandez L (1993). Flexible intra-
venous microdialysis probe for blood sampling in
freely moving rats. ) Appl Physiol 74:466-469.
Rauschenberger CR (1992). Dentin permeability: the clin-
ical ramifications. Dent Clin N Am 36:527-542.
Rauschenberger C, Richardson B (1992a). The relation-
ship between dentinal fluid and pulp tissue concen-
131
 trations of alpha-2-macroglobulin (A2-M) (abstract).)
Dent Res 71 (Spec Iss):627.
Rauschenberger C, Richardson B (1992b). "In vivo"
assessment of pulp inflammation utilizing dentinal
fluid (abstract). J Endodont 18:197.
Richardson DW, Tao L, Pashley DH (1991). Dentin perme-
ability: effects of crown preparation. \nt I Prosthodont
4:219-225.
Ruch IV, Lesot H, Begue-Kirn C (1995). Odontoblast dif-
ferentiation. \nt 1 Dev Biol 39:51-68.
Rutherford RB (1996). Role of osteogenic protein in
reparative and reactionary dentin. In: Dentin/pulp
complex. Shimono M, Takahashi K, editors. Tokyo,
lapan: Quintessence Publishing.
Rutherford RB, Wahle I, Tucker M, Rueger D, Charette M
(1994a). Induction of reparative dentin formation in
monkeys by recombinant human osteogenic protein-
\.Arch Oral Biol 38:571-576.
Rutherford RB, Spangberg L, Tucker M, Rueger D,
Charette M (1994b). The time course of the induction
of reparative dentin formation in monkeys by recom-
binant human osteogenic protein-1. Arch Oral Biol
39:833-838.
Rutherford RB, Spangberg L, Tucker M, Charette MC
(1995). Transdentinal stimulation of reparative dentin
formation by osteogenic protein-1 in monkeys. Arch
Oral Biol 40:681-683.
Sano H, Ciucchi B, Matthews WG, Pashley DH (1994).
Tensile properties of mineralized and demineralized
human and bovine dentin. 1 Dent Res 73:1205-1211.
Sano H, Takatsu T, Ciucchi B, Russell CM, Pashley DH
(1995). Tensile properties of resin-infiltrated deminer-
alized human dentin. J Dent Res 74:1093-1102.
Sarram S, Byers MR, Lee K-F (1995). p75 NGFR knockout
mice have deformed dental tissues and altered inner-
vation (abstract). ] Endod 21:240.
Schmid-Schonbein GW (1990). Microlymphatics and
lymph flow. Physiol Rev 70:987-1028.
Schroder U (1985). Effects of calcium hydroxide-contain-
ing pulp-capping agents on pulp cell migration, pro-
liferation and differentiation. J Dent Res 64(Spec
Iss):541-548.
Shellis RP (1994). Effects of a supersaturated pulpal fluid
on the formation of caries-like lesions on the roots of
human teeth. Caries Res 28:14-20.
Silverman JE, Kruger L (1987). An interpretation of den-
tal innervation based on the pattern of calcitonin-
gene-related peptide (CGRP) immunoreactive thin
sensory axons. Somatosensory Res 5:157-175.
Smith AJ, Tobias RS, Plant CG, Browne RM, Lesot Hr
Ruch JV (1990). In vivo morphogenic activity of dentine
matrix proteins. J Biol Buccale 18.123-129.
Smith AJ, Tobias RS, Cassidy N, Plant CG, Browne RM,
Begue-Kirn C, et al. (1994). Odontoblast stimulation in
132 Crit Rev Oral Biol Med
ferrets by dentine matrix components. Arch Oral Biol
39:13-22.
Smith AJ, Garde C, Cassidy N, Ruch JV, Lesot H (1995).
Solubilization of dentine extracellular matrix by calci-
um hydroxide (abstract). J Dent Res 74:829.
Smith DC, Cooper WEG (1971). The determination of
shear strength. A method using a micropunch appa-
ratus. BrDent] 130:333-337.
Stiefel A, Raab WH-M (1996). Changes in the formation
of dental hard tissues after desensitization with cap-
saicin. In.- Dentin/pulp complex. Shimono M,
Takahashi K, editors. Tokyo, Japan: Quintessence
Publishing.
Suge T, Ishikawa K, Kawasaki A, Yoshiyama M, Asaoka Kr
Ebisu S (1995). Effect of fluoride on the calcium
phosphate precipitation method for dentinal tubule
occlusion. J Dent Res 74:1079-1085.
Swift ML, Byers MR (1992). Effect of aging on responses
of nerve fibers to pulpal inflammation in rat molars
analyzed by quantitative immunocytochemistry. Arch
Oral Biol 37:901-912.
Tagami J, Hosoda H, Burrow MF, Nakajima M (1992).
Effect of aging and caries on dentin permeability. Proc
Finn Dent Soc 88(Suppl 1): 149-154.
Takeda K, Oida S, Goseki M, Iimura T, Maruoka Y,
Amagasa T, et al. (1994). Expression of bone mor-
phogenic protein genes in human dental pulp cells.
Bone 15:467-470.
Tao L, Pashley DH (1989). Dentin perfusion effects on the
shear bond strengths of bonding agents to dentin.
Dent Mater 5:181-184.
Tay FR, Gwinnett AJ, Pang KM, Wei SHY (1994). Structural
evidence of a sealed tissue interface with a total etch
wet bonding technique, in vivo. J Dent Res 73:629-636.
Taylor PE, Byers MR (1990). An immunocytochemical
study of the morphological reaction of nerves con-
taining calcitonin gene-related peptide to microab-
scess formation and healing in rat molars. Arch Oral
Biol 35:629-638.
Taylor PE, Byers MR, Redd PE (1989). Sprouting of CGRP
nerve fibers in response to dentin injury in rat molars.
Brain Res 461:371-375.
Thesleff I, Vaahtokari A, Partanen A-M (1995). Regulation
of organogenesis. Common molecular mechanisms
regulating the development of teeth and other
organs. \nt J Dev Biol 39:35-50.
Thomas HF (1985). The dentin-predentin complex and its
permeability: anatomical overview. J Dent Res 64(Spec
Iss):607-612.
Trantor IR, Messer HH, Birner R (1995). The effects of
neuropeptides (calcitonin-gene-related peptide and
substance P) on cultured human pulp cells. J Dent Res
74:1066-1071.
Trowbridge HO (1982). Dentistry today. Dentistry 82:22-29.
7(2):104-133 (1996)
Turner DF (1992). Immediate physiological response of
odontoblasts. Proc Finn Dent Soc 88(Suppl l):55-63.
Turner D, Marfurt C, Sattelburg C (1989). Demonstration
of physiological barrier between pulpal odontoblasts
and its perturbation following routine restorative pro-
cedures: A horseradish peroxidase tracing study in
the rat. ] Dent Res 68:1262-1268.
Tziafas D, Kolokuris I (1990). Inductive influences of dem-
ineralized dentin and bone on pulp cells: An
approach of secondary dentinogenesis. J Dent Res
69:75-81.
Vongsavan N, Matthews B (1991). The permeability of cat
dentine in vivo and in vitro. Arch Oral Biol 36:641-646.
Vongsavan N, Matthews B (1992a). Fluid flow through cat
dentine in vivo. Arch Oral Biol 37:175-185.
Vongsavan N, Matthews B (1992b). Changes in pulpal
blood flow and fluid flow through dentine produced
by autonomic and sensory nerve stimulation in the
7(2): 104-133 (1996) Crit Rev Oral Biol Med
cat. Proc Finn Dent Soc 88(Suppl 1):491-497.
Vongsavan N, Matthews B (1994). The relation between
fluid flow through dentine and the discharge of
intradental nerves. Arch Oral Biol 39(Suppl):140S.
Walton RE, Outhwaite WC, Pashley DH (1976).
Magnification, an interesting optical property of
dentin. 1 Dent Res 55:639-642.
Warfvinge I, Bergenholtz G (1986). Healing capacity of
human and monkey dental pulps following experimen-
tally induced pulpitis. Endod Dent Traumatol 2:256-262.
Watanabe LG, Nguyen T, Garner M, Kilbourne AM,
Marshall SI, Marshall GW (1996). Dentin shear
strength relative to tubule orientation. Effects of
tubule orientation and intratooth location. Dent Mater
12:109-115.
Weber DF, Zaki AL (1986). Scanning and transmission
electron microscopy of tubular structures presumed
to be human odontoblast processes. I Dent Res 65:982-
986.
133