Bone 75 (2015) 144150

Contents lists available at ScienceDirect

Bone
journal homepage: www.elsevier.com/locate/bone

Original Full Length Article

Quantifying the osteocyte network in the human skeleton

Pascal R. Buenzli a, Natalie A. Sims b,
a
School of Mathematical Sciences, Monash University, Clayton, VIC 3800, Australia
b
St Vincent's Institute of Medical Research, The University of Melbourne, Fitzroy, VIC 3065, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Osteocytes form an extensive cellular network throughout the hard tissue matrix of the skeleton, which is known to
Received 24 October 2014 regulate skeletal structure. However due to limitations in imaging techniques, the magnitude and complexity of this
Revised 25 January 2015 network remain undened.
Accepted 10 February 2015
We have used data from recent papers obtained by new imaging techniques, in order to estimate absolute and
Available online 20 February 2015
relative quantities of the human osteocyte network and form a more complete understanding of the extent
Edited by: Robert Recker and nature of this network.
We estimate that the total number of osteocytes within the average adult human skeleton is ~42 billion and that
Keywords: the total number of osteocyte dendritic projections from these cells is ~3.7 trillion. Based on prior measurements
Osteocyte of canalicular density and a mathematical model of osteocyte dendritic process branching, we calculate that these
Lacuno-canalicular network cells form a total of 23 trillion connections with each other and with bone surface cells. We estimate the total
Mathematical modelling length of all osteocytic processes connected end-to-end to be 175,000 km. Furthermore, we calculate that the
total surface area of the lacuno-canalicular system is 215 m2. However, the residing osteocytes leave only enough
space for 24 mL of extracellular uid. Calculations based on measurements in lactation-induced murine
osteocytic osteolysis indicate a potential total loss of ~ 16,000 mm3 (16 mL) of bone by this process in the
human skeleton. Finally, based on the average speed of remodelling in the adult, we calculate that 9.1 million os-
teocytes are replenished throughout the skeleton on a daily basis, indicating the dynamic nature of the osteocyte
network.
We conclude that the osteocyte network is a highly complex communication network, and is much more vast
than commonly appreciated. It is at the same order of magnitude as current estimates of the size of the neural
network in the brain, even though the formation of the branched network differs between neurons and osteo-
cytes. Furthermore, continual replenishment of large numbers of osteocytes in the process of remodelling allows
therapeutic changes to the continually renewed osteoblast population to be rapidly incorporated into the
skeleton.
2015 Elsevier Inc. All rights reserved.

Introduction (canaliculi) [4,5]. This cellular network is understood to sense local

Bone is a dynamic organ whose size, shape, and material properties
evolve by lifelong reorganisation and reconstruction through the pro-
cesses of growth, modelling, and remodelling [13]. Within the calcied
matrix generated by these processes lies a dense, organised network of
active interconnected osteocytes.
Osteocytes derive from bone-forming cells that become trapped
in bone matrix during formation. These cells occupy small pores in
the bone called lacunae. They create an extensive network throughout
the mammalian skeleton, formed by connections between multiple
dendrite-like processes contained within uid-lled micro-canals
Corresponding author at: St. Vincent's Institute of Medical Research, 9 Princes St,
Fitzroy VIC 3065, Australia. Fax: +61 3 9416 2676.
E-mail address: nsims@svi.edu.au (N.A. Sims).
mechanical strains on the bone matrix, and to release and receive
signals that control the processes of bone matrix production [6,7],
mineralization [8] and resorption [9,10] such that skeletal strength is
maintained, particularly in response to mechanical stimuli [11,12].
The osteocyte network not only contributes to calcium homeostasis,
but also releases FGF23 to regulate phosphate homeostasis at the kid-
ney [13]. Other postulated roles for the osteocyte network that have
been proposed are the regulation of myelopoiesis [14], glucose metabo-
lism [15] and fertility [16], the latter two by endocrine secretion of
osteocalcin. The osteocyte network is thus an endocrine organ with reg-
ulatory power that must be directly inuenced by the extent and con-
nectivity of its network. While the central role of osteocytes in the
adaptation of bone shape, bone microarchitecture, and the regulation
of bone mineral content is beginning to be appreciated, how these
multi-scale regulations are enabled by this network of cells remains
largely unknown.

http://dx.doi.org/10.1016/j.bone.2015.02.016
8756-3282/ 2015 Elsevier Inc. All rights reserved.
 P.R. Buenzli, N.A. Sims / Bone 75 (2015) 144150 145

Fig. 1. A: lacuno-canalicular network around a Haversian canal (reproduced from [21] with kind permission from the American Association of Physicists in Medicine). B: Direct canalicular
paths between a single osteocyte (white) and (i) nearby osteocytes (blue) or (ii) the vascular pore (top) (reproduced from [22] with kind permission from John Wiley and Sons, Inc.). C: An
osteocyte lacuna (yellow) with 86 emanating canaliculi (red) (reproduced from [24] (Fig. 4C) with kind permission from Springer Science and Business Media). (For interpretation of the
references to colour in this gure legend, the reader is referred to the web version of this article.)

Since osteocytes reside within the hard tissue of bone, their cellular
network is difcult to image in three dimensions. However, recent ad-
vances in bone imaging techniques have enabled examination of the
three-dimensional morphology of the lacuno-canalicular pore network
(LCN) within which the cellular network resides [1724] (Fig. 1). Here,
we mainly use published information on the lacuno-canalicular net-
work to quantify osteocytes and the network formed by their dendritic
processes. This information is used to estimate absolute quantities such
as the total number of osteocytes, osteocyte connections, and the cumu-
lative size of the lacuno-canalicular network within the adult human
skeleton. The aim of this paper is to quantify the extent of the osteocyte
network to provide an appreciation of its magnitude and complexity;
this will inform the eld about the extent of the mechano-sensory sys-
tem in the human skeleton and about the large volume of information
that could be transmitted by such a network.
Methods and results
We combined data from the published literature with arithmetic cal-
culations and mathematical models to provide novel quantications of
the osteocyte network system. Quantities taken from the literature
and used in these quantications are listed in Table 1 with their source.
The major quantities derived in this paper are listed in Table 2 with the
formula used to calculate them. Table 3 summarises how these quanti-
ties compare with other organ systems in the body.
Number of osteocytes and their lacunae
To estimate the total number of osteocytes in the human skeleton
(Tt.N.Ot), we multiplied the total volume of bone matrix (BV) by an es-
timate of the average volumetric osteocyte density (N.Ot/BV).
The total volume of bone in an average adult Caucasian skeleton was
estimated by Partt to be BV = 1.75 L = 1.75 106 mm3. This estimate
is based on a dry fat-free skeletal weight of 4 kg and a tissue density of
2.3 kg/L [25, Table 1; Sec. II.A.3].
Osteocyte lacunar density (N.Lc/BV) depends on the skeletal site,
distance to bone surface, tissue bone volume fraction, species, race
and health status [2631]. Values in recent human studies range from
20, 00030, 000/mm3 [3133]. Since it has been estimated that 5% of
osteocyte lacunae are empty [25,28], this suggests an average osteocyte
density (N.Ot/BV) of 19, 00028, 500/mm3. With these gures, we esti-
mate the total number of osteocytes in the skeleton to be 33 109

Table 1
Measured osteocyte network quantities used in calculations (human unless otherwise indicated).

Description (abbreviation) Value Origin

6 3
Skeletal bone volume (BV) 1.75 10 mm [25; Table 1, Sec II.A.3]
Osteocyte lacuna density (N.Lc/BV) 20,00030,000 mm3 [25,3133]
Fraction of lacunae with live osteocyte (N.Ot/N.Lc) 0.95 [25,28]
Fraction of bone formed per day (remodelling rate, Rem.R) 7.6%/year [25]
364.384 mm3/day
Number of dendritic process per osteocyte (N.DP/Ot) 89 [24]
Radius of a canaliculus (Ca.Rd) 157.5 nm 129.5 nm (murine) [36]; 157.5 nm (human) [24]
Radius of a dendritic process (DP.Rd) 73 nm 50.2 nm (murine) [36]; 73 nm (human, calculated such
that DP cross-section occupies 16% of canaliculus cross-section)
Pericellular/perilacunar gap (g) 0.6 m 0.51 m [37], 0.6 m [22]
Dendritic length density (Tt.DP.L/BV) 0.1 m/m3 0.26 m/m3 (chick calvaria) [18], 0.1 m/m3
(based on Appendix A/text and data from [22])
Lacuna volume (Lc.V) 400 m3 [31,33]
Lacuna surface area (Lc.S) 336.2 m2 [31]
Canalicular network mean node degree (k + 1) 3.25 (ovine) [22]
Canalicular network mean link length (l) 2.15 m (ovine) [22]
Lacuna area control mice (2D section) 38 m2 (murine) [56]
Lacuna area lactating mice (2D section) 46 m2 (murine) [56]
Increase in canaliculus radius in lactating mice 0.05 m (murine) [56]
146 P.R. Buenzli, N.A. Sims / Bone 75 (2015) 144150

Table 2
Derived osteocyte network quantities for the human skeleton.

Description (abbreviation) Formula Value

Osteocyte density (N.Ot/BV) (N. Ot/N. Lc) (N. Lc/BV) 19,00028,500 mm3
Total number of osteocytes (Tt.N.Ot) (N. Ot/BV) BV 41.6 109 (~42 billion)
Total number of lacunae (Tt.N.Lc) (N. Lc/BV) BV 43.8 109 (~44 billion)
Total number of osteocyte dendritic processes (Tt.N.DP) Tt. N. Ot (N. DP/Ot) 3.7 1012 (3.7 trillion)
Total number of osteocyte connections (Tt.N.Ot.Cx) (N. Cx/DP) (N. DP/Ot) Tt. N. Ot/2 23.4 1012 (~23 trillion)
Number of terminal connections per osteocyte (N.Cx/Ot) (N. Cx/DP) (N. DP/Ot) 1128
 
Number of terminal connections per dendritic process (N.Cx/DP) 1
1 h1ki Tt:DP:L=DP 12.7
hki hli
Total cumulated length of osteocyte dendritic processes (Tt.DP.L) (Tt. DP. L/BV) BV 175,000 km
Cumulated length of osteocyte processes per osteocyte (Tt.DP.L/Ot) Tt:DP:L=BV 4.2 mm
N:Ot=BV
Cumulated length of a single osteocyte process (Tt.DP.L/DP) Tt:DP:L=BV 47 m
N:Ot=BVN:DP=Ot
Total lacuno-canalicular system surface area (Tt.LCN.S) Lc. S Tt. N. Lc + 2 Ca. Rd Tt. DP. L 215 m2
Total lacuno-canalicular volume (Tt.LCN.V) Lc. V Tt. N. Lc + Ca. Rd2 Tt. DP. L 35.8 cm3
Lacuno-canalicular porosity Tt:LCN:V
BV
2.05%
Total extracellular lacuno-canalicular volume (Tt.EC.LCN.V) Lc. S g Tt. N. Lc + (Ca. Rd2 DP. Rd2) Tt. DP. L 24.2 cm3
Number of osteocytes replaced per unit time Rem. R (N. Lc/BV) 3.33 109 cells/year
9.1 106 cells/day

50 109 (average of ~42 billion), and the total number of lacunae to be of all osteocyte dendritic processes in the human skeleton (Tt.DP.L) to
35 10953 109 (average of ~44 billion). be 175,000 km.
Based on this calculation, each osteocyte has thus an average total
cumulated length of all its dendritic processes (Tt.DP.L/Ot) of 4.2 mm.
Number and length of osteocytic dendritic processes Based on the measured average of 89 cell processes per osteocyte
as above [24], the average cumulated length of a single cell process
The number of osteocytic dendritic processes varies between species projecting from an osteocyte's cell body including all its sub-branches
[17,18,34]; osteocytes in human lamellar bone have been estimated to (Tt.DP.L/DP) is ~ 47 m. This estimate is probably conservative at this
possess between 18 and 106 dendritic processes per cell in [34]. Recent level as well in view of the average separation between osteocytes in
3D measurements have exhibited osteocyte lacunae with 53 to 126 human bone (~3237 m),1 the highly branched and tortuous canalicu-
emanating canaliculi, with a mean ( SD) of 89 25 (see Fig. 1C) lar structure, and the fact that some dendritic processes may connect
[24]. Incorporating this with our previous estimate of total osteocyte distant osteocytes.
number, this gives a total number of processes projecting from the oste-
ocyte cell bodies in the skeleton of 2.96 10124.44 1012 (average of
3.7 trillion). Connections within the network
Osteocyte processes branch close to the cell body, and interconnect
into a dense interwoven network within bone matrix [17,18,22]. This The total number of connections an osteocyte makes with other
network structure connects distant osteocytes with each other and osteocytes, bone surface osteoblasts, and vascular pores (Tt.N.Ot.Cx) has
with vascular pores both through cellcell contact, and via uid ow to our knowledge not been reported previously. We provide an estimate
within the LCN [22] (see Fig. 1B). The total cumulated length of of this number based on the cumulated cell process length Tt.DP.L/DP es-
osteocytic dendritic processes per unit volume of bone (Tt.DP.L/BV) timated above, and a mathematical branching model taking as input the
was measured at 0.26 0.02 m/m3 in embryonic chick calvariae experimental frequency distributions of canalicular node degree and can-
[18]. In Appendix A, we estimate a lower bound of ~ 0.2 m/m3 for alicular link lengths reported in [22]. This mathematical branching model
the osteocyte dendritic length density by using the fact that 80% of is presented in detail in Appendix B. The model estimates the average
bone matrix resides within 1.4 m of the closest canalicular network number of termini of a single process by successively forking and growing
structure in mature ovine bone [22] and by modelling the lowest possi- subbranches of a single stem attached to the cell body. Subbranches are
ble density of edges of a periodic 3D structure. Since osteocytes in em- created such that node degrees and link lengths occur with the frequen-
bryonic calvariae are more densely packed than in mature bone, and cies observed experimentally in [22]. In this model, new subbranches
since some interstitial bone may be located further from the closest can- are added until 47 m of cumulative length of the process is generated
alicular network structure than measured in [22], we use in this paper (see Fig. 2). The number of termini obtained by this procedure is an
for human bone a more conservative whole bone average of Tt.DP.L/ upper limit of possible termini, because it assumes that dendritic cell pro-
BV of 0.1 m/m3. Multiplying this value by the total volume of bone cesses do not anastomose to form closed loops with their own sub-
of 1.75 L [25, Table 1; Sec. II.A.3], we estimate the total cumulated length branches. Considering previously published data describing canalicular
node degree and link length [22], we nd that a cell process projecting
Table 3 from an osteocyte may form up to 12.7 termini in average. Multiplying
Comparisons with other measured quantities for the human body.
the average number of termini per process by the number of processes
Description (abbreviation) Osteocytes/bone Other organs per cell, we estimate that a single osteocyte may possess up to 1128 ter-
Total tissue volume 1.75 L of bone tissue 1.2 L of brain tissue mini that connect with other cells. Extrapolated to the whole skeleton
Total number of cells 42 billion osteocytes 86 billion neurons [39] by multiplying by the total number of osteocytes and dividing by two
Total dendritic process length 175,000 km 150,000180,000 km (each connection being shared), this corresponds to a total number of
(Tt.DP.L) (brain) [40]
osteocytic connections of ~23.4 1012 (23.4 trillion).
Total number of osteocyte 23 trillion 150 trillion neural
connections (Tt.N.Ot.Cx) cortex synapses
Total lacuno-canalicular system 215 m2 118 m2 (Lung) [62]
surface area (Tt.LCN.S) 32 m2 (Gastrointestinal 1
These values are estimated from the cubic root of the mean territory volume (mean
tract) [63]
volume of bone tissue surrounding each lacuna, BV/N.Lc, i.e. the inverse of lacunar
2 m2 (Skin) [64]
density).
 P.R. Buenzli, N.A. Sims / Bone 75 (2015) 144150
Fig. 2. Left: Branching morphology of osteocytic dendritic processes (DP). Right: The mathematical model of branching dendritic process successively forks subbranches starting from the
DP stem until the total cumulated length of the process and all its subbranches (Tt.DP.L/DP) reaches the value of 47 m. At this point, the total number of termini is counted. The branchings
are generated such that the frequency of node degree and frequency of link length are distributed according to the experimental observations of Ref. [22]. Subbranch colours emphasise the
successive steps of the mathematical procedure (only subbranches of a single branch are coloured for clarity). Background image: courtesy of Dr Lynda Bonewald, adapted by permission
from Macmillan Publishers Ltd: IBMS BoneKEy (2009) 6, 6370 [35], copyright 2009.
Surface area and pore volume of the lacuno-canalicular system
To calculate the surface area and pore volume of the lacuno-
canalicular system, we calculated values for the lacunar system and for
the canalicular system separately. Human canalicular diameter was mea-
sured in [24] from 3D phase nano-tomography scans (Fig. 1C) with a
mean value of 365 130 nm, similar to the previously reported murine
canalicular diameter of 259 129 nm [36]; this highly variable parameter
reects a high level of variability also observed in micrographs from ear-
lier human studies referenced in [36]. Assuming a circular canalicular
cross-section, the cross-sectional perimeter of a canaliculus is
~1.14668 m. Multiplying by the total length of dendritic processes esti-
mated above, the total surface area of the human canalicular system
(Tt.Ca.S) (not including the lacunae in which the osteocytic nuclei reside)
is therefore approximately 201 m2.
The average surface area of a human osteocyte lacuna was reported in
[31] to be 336.2 m2. Multiplying by the total number of osteocyte
lacunae (including empty lacunae), this corresponds to a total lacunar
surface area (Tt.Lc.S) of 11.817.7 m2 (average of 14.7 m2). The total sur-
face area of the human lacuno-canalicular network (Tt.LCN.S) is thus
212.4218.3 m2 (average of ~215 m2).
The total pore volume of the lacuno-canalicular network in the
skeleton can be estimated similarly. Based on a measured average can-
alicular diameter of 365 nm [24], and assuming a circular cross-section,
the cross-sectional area of a canaliculus is ~0.104635 m2. Multiplying
by the total cumulated length of osteocyte dendritic process this gives
a total canalicular volume (Tt.Ca.V) of 18,311 mm3 in the human
skeleton.
The average lacunar volume was reported in [31,33] to be about
400 m3. Multiplying by the total number of osteocyte lacunae
(including empty lacunae), this gives a total lacunar volume in the
skeleton (Tt.Lc.V) of 14,00021,000 mm3. The total volume of the whole
lacuno-canalicular system (Tt.LCN.V) is thus estimated to be
32,31139,311 mm3 (average of 35,811 mm3). With the total skeletal
bone volume (BV) reported above, this corresponds to a lacuno-
canalicular porosity in bone matrix of 1.852.25% (0.8%1.2% lacunar,
1.05% canalicular). These gures are consistent with the 0.76 0.19% la-
cunar porosity determined recently in synchrotron-based measurements
of human cortical bone [31].
Experimental studies comparing dimensions of the lacuno-canalicular
pore space and the osteocyte itself have not yet provided data on the pro-
portion of uid-containing pericellular space to cell bodies in human
147
bone. Direct measurement from transmission electron microscopy on
murine bone [36] indicates an osteocyte process diameter of 104
69 nm with pericellular space surrounding each process of 78 38 nm.
Therefore, on average, murine osteocyte processes ll approximately
16% of the canalicular cross-sectional area. Assuming this proportion to
hold for human bone, and given the measured diameter 365 nm of the
human canaliculi [24], the human osteocyte process diameter can be esti-
mated to be 146 nm, and the total volume of extracellular uid surround-
ing dendritic processes in the canaliculi to be ~ 15,381 mm3. Within the
lacunae, the space occupied by the osteocyte cell body reduces as the
cell matures [5]. Based on a width of the extracellular uid-lled space be-
tween the osteocyte body and the lacunar surface of about 0.6 m [22,37]
and given the surface area 336.2 m2 of a lacuna, the volume of extracel-
lular space in each osteocyte-containing lacuna can be estimated to be
0.6 336.2 = 201.7 m3, i.e., about 50% of the lacuna volume. The total
volume of extracellular uid surrounding the cell body and dendritic pro-
cesses of osteocytes in the whole lacuno-canalicular network is thus esti-
mated to be 22,44225,972 mm3 (average of 24,207 mm3).
Osteocyte production rate
Although some osteocytes, such as those in the inner ear, may exist
for as long as human life [38], the majority are regularly removed and
replaced due to the renewal of the bone matrix during remodelling.
Partt [25, Table 7] estimated the average skeletal remodelling rate
(Rem.R) to be 7.6%/year (combining both cortical and trabecular turn-
over rates). Over the entire human skeleton bone volume (BV), this
corresponds to a volume of renewed bone matrix of 133 cm3/year
(0.36 cm3/day). Assuming this renewed bone to contain the same
density of osteocytes, this corresponds to the replacement of
2.66 1093.99 109 (average 3.33 billion) new osteocytes per year
(9.1 million per day).
Discussion
The number of cells of an organ, their organisation, and the total size
of tissues affect these systems' functional roles. Here we have estimated
the absolute numbers of osteocytes, the connections made by these
cells, the size of the lacuno-canalicular system, and the rate of osteocyte
replacement by the average human body to provide an insight into the
complexity and extent of this network.
148 P.R. Buenzli, N.A. Sims / Bone 75 (2015) 144150
The total number of osteocytes in the human skeleton (42 billion) is
in the same order of magnitude as previous estimates of the number of
neurons in the brain (~86 billion [39]), and the total length of dendritic
processes in bone (175,000 km) falls in the same range as the total
length of nerve bres in the human brain (150,000180,000 km) [40]
(Table 3). These cells are held within a total organ volume (1.75 L)
that is larger than the average brain volume (1.2 L). With a total number
of 150 trillion synapses in the human brain cortex, neurocortical neu-
rons are more densely connected with 7000 synapses on average [40],
compared to 1128 dendritic connections per osteocyte. At the whole
skeleton level, osteocytic dendritic cell processes make about 23.4 tril-
lion connections, thereby providing a highly intricate system by which
bone mass can be regulated at the local level, as well as sufcient net-
works to allow osteocytes to act as an endocrine system, releasing
regulable quantities of FGF23 and osteocalcin into the circulation [13].
The magnitude of this network also helps us to understand how a reduc-
tion in sclerostin production by each osteocyte in the presence of a
pharmacological dose of parathyroid hormone (PTH), may both cause
a dramatic reduction in circulating sclerostin in the patient [41], and re-
sult in a lack of Wnt inhibition in many nearby osteoblasts on the bone
surface [42]. The efciency of such a local, diffusion-based signal trans-
mission depends strongly on the density of osteocytes and their cell
processes, and uid ow within the lacuna-canalicular network [12].
Some calculations herein rely on measurements of the 3D morphol-
ogy of the lacuno-canalicular network rather than directly on the oste-
ocyte network itself. While the lacuno-canalicular pore network is
expected to match the network of the cells themselves, osteocytes un-
dergo apoptosis and both their dendritic processes and the connections
they form may also be dynamic [43]. We accounted for this based on a
95% lacunar occupancy, but canalicular occupancy is not yet known,
so it has not been considered in our calculations. Our estimate of
the total number of connections has also excluded connections that os-
teocytes make with other structures, such as vascular pores, and the
possibility of termini without connection to any structure. As imaging
techniques are improved, and the proportion of cells that make these
connections can be calculated, our estimate of the number of connec-
tions of each type can be rened. Although osteocyte networks are likely
to be different in different individuals, bone remodelling and bone adap-
tation are likely to generate osteocytes within a limited range of densi-
ties. It can therefore be expected that the extent of the osteocyte
network in the adult human skeleton does not vary by more than an
order of magnitude between individuals [44].
The nature of cellcell contact between the termini of osteocyte
processes is not yet known, nor is the number of signals that may be
transmitted per connection and the speed of transmission. Mechanisms
depending on cell contact (sometimes called wire-transmission [45])
are known to regulate osteocyte function, including gap junction com-
munication [46,47] and ephrin/Eph receptor tyrosine kinase interac-
tions [48]. At this stage, imaging modalities in the hard tissue of bone
have not yet allowed specic location of these interactions, exchange
of soluble factors, or juxtacrine signalling specically at the site of
dendrite tip to tip contact. However, the efciency of cell-contact-
dependent wire transmission through the dendritic network may also
depend on absolute number of cells within the network given the pos-
sibility of distant neighbour connections and of direct connections to the
vascular pore surface [22,49].
A large number of osteocytes are replaced rapidly by remodelling
an estimated 9.1 million per day. This level of replacement falls below
that of red blood cells (at 34.6 billion per day) [50], but is far greater
than the level of renewal estimated for hippocampal neurons (700 per
day) [51]. This extensive level of replenishment suggests that therapeu-
tic approaches that change the nature of new osteocytes being deposit-
ed (such as gene therapy approaches that target the osteoblast) may
allow rapid incorporation of modied osteocytes within the skeleton.
It also indicates that a large number of bone surface osteoblasts are in-
corporated into the matrix, in a manner dependent on signals from
within the extensive osteocytic network [49]. Although these propor-
tions are unclear, it has been suggested that 1929% of the osteoblast
population is incorporated into the bone matrix [25,28], suggesting
that the number of osteoblasts produced per day in the skeleton ranges
from 1.14 10111.7 1011.
The total number of osteocytes and extent of the lacuno-canalicular
network also provide an indication of the extent of the inuence of
osteocytic osteolysis on bone mass. Osteocytic osteolysis is a process
whereby osteocytes destroy the bone matrix surrounding the cell [52]
and has been observed in the skeleton during lactation in rodents
[5256], and with PTH treatment [52,5759]. One question raised by
these observations is how much such a mechanism could contribute
to the low bone mass, and increased risk of fracture in lactating
women [55]. Prior direct measurement of 2D sections from tibiae of
lactating mice reported an increase in lacunar size from 38 to 46 m2
[56]. If this amount is removed by every osteocyte in the human skele-
ton, assuming an ellipsoid shape, this would equate to a total volume of
15, 06816, 899 mm3 bone lost, increasing lacuno-canalicular porosity
by at least 30% (from 1.852.25% to 2.702.96%) (see Appendix C).
This indicates the potential of this mechanism to alter the structural
properties of bone matrix and the permeability of the osteocyte
network.
High-resolution 3D imaging of bone has enabled the visualisation of
the lacuno-canalicular network and revealed its complexity, but few
studies have quantied the osteocyte network so far. While we have
provided estimates of the number of possible connections each osteo-
cyte can make, an estimation of the number of distinct cells to which
these connections are actually made is more challenging because imag-
ing studies indicate that multiple connections exist between any pair of
osteocytes [21,22,60]. Immature osteoid-osteocytes in chick calvariae
are connected with 56 different bone surface osteoblasts through
more than 20 dendritic processes, corresponding to a four-fold redun-
dancy [60]. It is likely that cellcell connectivity in the osteocyte net-
work has a similar or higher redundancy. Such redundancy may help
inltrate bone matrix homogeneously, such that little bone resides fur-
ther away than a few micrometres from a dendritic process, providing a
mechano-sensing advantage, as well as a way to control the degree of
mineralisation of bone matrix [22,61].
The surface area of the lacuno-canalicular network (Tt.LCN.S) is
large, an estimated total of 215 m2. This is more than that of a healthy
adult lung (118 m2) [62], and much larger than the surface areas
of the gastrointestinal tract (32 m2) [63] and the skin (2 m2) [64]
(Table 3). This area, and the related but slightly smaller total surface
area of the osteocyte cell bodies, provide a remarkably large interaction
interface for the small volume of extracellular lacuno-canalicular uid
(24 mL). This suggests that osteocytes and their dendritic processes
can interact very efciently with signalling molecules and ions from
the liquid, e.g. to sense uid ow and/or pressure, to release sclerostin
in the circulation, and to regulate the mineralisation of the bone matrix.
The embedment of the osteocyte network in a hard bone matrix
substrate has prevented the systematic analysis of the network until
only recently, in contrast with other network systems of similar numer-
ic and organisational properties such as the brain. In neurobiology,
quantication of the neural network is used to understand the brain's
information-processing capacity across species, and the effect of aging
and diseases such as Alzheimer's disease [40,6567]. Just as the number
of neurons and their connectivity provide insight into the information
processing capacity of the human brain [65,67], the large number of os-
teocytes suggests that the network formed by the osteocytes could also
transmit extensive information via cellcell contact or paracrine mech-
anisms. Continuing work to identify whether the extent of the network
varies across species, including extinct species, may also provide in-
sights into how this network complexity has evolved [6871]. Finally,
as mathematical models of the biomechanical regulation of the skeleton
become more comprehensive [49,7281], valid quantications of the
osteocyte network from the cellular to the whole organ scale will
 P.R. Buenzli, N.A. Sims / Bone 75 (2015) 144150
allow realistic predictions of this complex dynamic system and its emer-
gent behaviours [82,83]. High-resolution 3D imaging of bone is improv-
ing, thus revealing new understandings of the function of the osteocyte
network. As these technologies mature and are used to image bones
from different species, ages, skeletal sites (including trabecular, cortical,
osteonal and interstitial bone), pathologies and variations (e.g. osteopo-
rosis, osteogenesis imperfecta, diabetes, osteoarthritis, glucocorticoid
use), the same approach can be used to understand how the osteocyte
network changes with age, region and skeletal health.
The calculations we have made are based on some assumptions that
may also be resolved as more reports of high resolution imaging of bone
are published. The calculation of the total number of osteocytic connec-
tions assumes that osteocytes form 1:1 connections, and do not form
closed loops. Imaging studies of actual osteocyte dendrites, rather than
the network, are likely to have the greatest impact on these calculations.
In contrast, if other assumptions are proved to be incorrect, there will
not be a substantial change in the values obtained: e.g. if osteocyte den-
drites and canaliculi have circular cross sections, or whether the osteocyte
lacunae are ellipsoid in shape. Similarly, the wide range of values we have
calculated for osteocyte density, and all subsequent parameters, including
total LCN volume, should encompass differences that occur with normal
physiological variation; future imaging studies, will be able to narrow
these ranges.
We conclude that these new estimates of the size and complexity of
the osteocyte dendritic network, and the lacuno-canalicular system in
which they reside, show the extent of osteocyte connection, and possible
paracrine and direct cellcell communication, at a whole organ level. This
should be considered when interpreting new ndings regarding osteo-
cyte behaviour. The high level of complexity of the system and its contin-
ual replenishment provide a mechanism by which the network can
integrate multiple signals from changes in uid ow/pressure, cellcell
contact dependent mechanisms, and contributions from and to both
paracrine and endocrine signalling.
Acknowledgments
We thank Kevin Tetsworth for organising star-gazing at the Eighth
Clare Valley Bone Meeting, Sarah L Rea and Gerald J Atkins for useful dis-
cussions, Astrid Bakker and Egon Perilli for suggesting useful reference
works, Lynda Bonewald for the background image in Fig. 2 and T John
Martin and Ego Seeman for enthusiastically and critically reviewing the
manuscript. PRB is the recipient of an Australian Research Council Discov-
ery Early Career Research Award (DE130101191); NAS is funded by a Na-
tional Health and Medical Research Council Senior Research Fellowship
(1019703); St. Vincent's Institute is supported by the Victorian State Gov-
ernment Operational Infrastructure Support Scheme.
Appendices AC. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.bone.2015.02.016.
References
[1] Seeman E. Modeling and remodeling: the cellular machinery responsible for the gain
and loss of bone's material and structural strength. In: Bilezikian JP, Raisz LG, Martin
TJ, editors. Principles of Bone Biology. 3rd ed. San Diego, California: Elsevier; 2008.
p. 328.
[2] Frost HM. Skeletal structural adaptations to mechanical usage (SATMU): 1.
Redening Wolff's law: the bone modeling problem. Anat Rec 1990;226:40313.
[3] Frost HM. Skeletal structural adaptations to mechanical usage (SATMU): 2.
Redening Wolff's law: the bone remodeling problem. Anat Rec 1990;226:41422.
[4] Bonewald LF. The amazing osteocyte. J Bone Miner Res 2011;26:22938.
[5] Knothe Tate ML, Adamson JR, Tami AE, Bauer TW. The osteocyte. Int J Biochem Cell
Biol 2004;36:18.
[6] van Bezooijen RL, Roelen BA, Visser A, van der Wee-Pals L, de Wilt E, Karperien M,
et al. Sclerostin is an osteocyte-expressed negative regulator of bone formation,
but not a classical BMP antagonist. J Exp Med 2004;199:80514.
149
[7] Johnson RW, Brennan HJ, Vrahnas C, Poulton IJ, McGregor NE, Standal T, et al.
The primary function of gp130 signaling in osteoblasts is to maintain bone
formation and strength, rather than promote osteoclast formation. J Bone
Miner Res 2014;29:1492505.
[8] Feng JQ, Ward LM, Liu S, Lu Y, Xie Y, Yuan B, et al. Loss of DMP1 causes rickets
and osteomalacia and identies a role for osteocytes in mineral metabolism.
Nat Genet 2006;38:13105.
[9] Xiong J, Onal M, Jilka RL, Weinstein RS, Manolagas SC, O'Brien CA. Matrix-embedded
cells control osteoclast formation. Nat Med 2011;17:123541.
[10] Nakashima T, Hayashi M, Fukunaga T, Kurata K, Oh-Hora M, Feng JQ, et al. Ev-
idence for osteocyte regulation of bone homeostasis through RANKL expres-
sion. Nat Med 2011;17:12314.
[11] Turner CH, Forwood MR. What role does the osteocyte network play in bone adap-
tation? Bone 1995;16:2835.
[12] Schafer M, Cheung W-Y, Majeska R, Kennedy O. Osteocytes: master orchestrators
of bone. Calcif Tissue Int 2014;94:524.
[13] Fukumoto S, Martin TJ. Bone as an endocrine organ. Trends Endocrinol Metab 2009;
20:2306.
[14] Fulzele K, Krause DS, Panaroni C, Saini V, Barry KJ, Liu X, et al. Myelopoiesis is
regulated by osteocytes through Gsalpha-dependent signaling. Blood 2013;
121:9309.
[15] Lee NK, Sowa H, Hinoi E, Ferron M, Ahn JD, Confavreux C, et al. Endocrine regulation
of energy metabolism by the skeleton. Cell 2007;130:45669.
[16] Oury F, Sumara G, Sumara O, Ferron M, Chang H, Smith CE, et al. Endocrine regula-
tion of male fertility by the skeleton. Cell 2011;144:796809.
[17] Kamioka H, Honjo T, Takano-Yamamoto T. A three-dimensional distribution of osteo-
cyte processes revealed by the combination of confocal laser scanning microscopy
and differential interference contrast microscopy. Bone 2001;28:1459.
[18] Sugawara Y, Kamioka H, Honjo T, Tezuka K, Takano-Yamamoto T. Three-
dimensional reconstruction of chick calvarial osteocytes and their cell processes
using confocal microscopy. Bone 2005;36:87783.
[19] Wang L, Wang Y, Han Y, Henderson SC, Majeska RJ, Weinbaum S, et al. In situ mea-
surement of solute transport in the bone lacunar-canalicular system. Proc Natl Acad
Sci U S A 2005;102:119116.
[20] Schneider P, Meier M, Wepf R, Muller R. Towards quantitative 3D imaging of the os-
teocyte lacuno-canalicular network. Bone 2010;47:84858.
[21] Pacureanu A, Langer M, Boller E, Tafforeau P, Peyrin F. Nanoscale imaging of
the bone cell network with synchrotron X-ray tomography: optimization of
acquisition setup. Med Phys 2012;39:222938.
[22] Kerschnitzki M, Kollmannsberger P, Burghammer M, Duda GN, Weinkamer R,
Wagermaier W, et al. Architecture of the osteocyte network correlates with bone ma-
terial quality. J Bone Miner Res 2013;28:183745.
[23] Zuluaga MA, Orkisz M, Dong P, Pacureanu A, Gouttenoire PJ, Peyrin F. Bone
canalicular network segmentation in 3D nano-CT images through geodesic
voting and image tessellation. Phys Med Biol 2014;59:215571.
[24] P. Varga, B. Hesse, M. Langer, S. Schrof, N. Mnnicke, H. Suhonen, et al.,
SuhonenSynchrotron X-ray phase nano-tomography-based analysis of the lacunar-
canalicular network morphology and its relation to the strains experienced by osteo-
cytes in situ as predicted by case-specic nite element analysisBiomech Model
Mechanobiol. http://dx.doi.org/10.1007/s10237-014-0601-9 (in press)
[25] Partt AM. The physiologic and clinical signicance of bone histomorphometric data.
In: Recker RR, editor. Bone histomorphometry: techniques and interpretation. Boca
Raton, FL, USA: CRC Press; 1983. p. 143223.
[26] Vashishth D, Gibson G, Kimura J, Schafer MB, Fyhrie DP. Determination of bone vol-
ume by osteocyte population. Anat Rec 2002;267:2925.
[27] Hernandez CJ, Majeska RJ, Schafer MB. Osteocyte density in woven bone. Bone
2004;35:10959.
[28] Qiu S, Rao DS, Palnitkar S, Partt AM. Differences in osteocyte and lacunar density
between Black and White American women. Bone 2006;38:1305.
[29] Hannah KM, Thomas CD, Clement JG, De Carlo F, Peele AG. Bimodal distribu-
tion of osteocyte lacunar size in the human femoral cortex as revealed by
micro-CT. Bone 2010;47:86671.
[30] Zarrinkalam MR, Mulaibrahimovic A, Atkins GJ, Moore RJ. Changes in osteo-
cyte density correspond with changes in osteoblast and osteoclast activity
in an osteoporotic sheep model. Osteoporos Int 2012;23:132936.
[31] Dong P, Haupert S, Hesse B, Langer M, Gouttenoire PJ, Bousson V, et al. 3D osteocyte
lacunar morphometric properties and distributions in human femoral cortical bone
using synchrotron radiation micro-CT images. Bone 2014;60:17285.
[32] Bromage TG, Lacruz RS, Hogg R, Goldman HM, McFarlin SC, Warshaw J, et al. Lamel-
lar bone is an incremental tissue reconciling enamel rhythms, body size, and organ-
ismal life history. Calcif Tissue Int 2009;84:388404.
[33] Carter Y, Thomas CD, Clement JG, Peele AG, Hannah K, Cooper DM. Variation
in osteocyte lacunar morphology and density in the human femura syn-
chrotron radiation micro-CT study. Bone 2013;52:12632.
[34] Beno T, Yoon YJ, Cowin SC, Fritton SP. Estimation of bone permeability using accu-
rate microstructural measurements. J Biomech 2006;39:237887.
[35] Pajevic PD. Regulation of bone resorption and mineral homeostasis by osteocytes.
IBMS BoneKEy 2009;6:6370.
[36] You LD, Weinbaum S, Cowin SC, Schafer MB. Ultrastructure of the osteocyte pro-
cess and its pericellular matrix. Anat Rec A Discov Mol Cell Evol Biol 2004;278:
50513.
[37] McNamara LM, Majeska RJ, Weinbaum S, Friedrich V, Schafer MB. Attachment of
osteocyte cell processes to the bone matrix. Anat Rec (Hoboken) 2009;292:35563.
[38] Bloch SL, Kristensen SL, Sorensen MS. The viability of perilabyrinthine osteo-
cytes: a quantitative study using bulk-stained undecalcied human temporal
bones. Anat Rec (Hoboken) 2012;295:11018.
150 P.R. Buenzli, N.A. Sims / Bone 75 (2015) 144150
[39] Azevedo FA, Carvalho LR, Grinberg LT, Farfel JM, Ferretti RE, Leite RE, et al. Equal
numbers of neuronal and nonneuronal cells make the human brain an isometrically
scaled-up primate brain. J Comp Neurol 2009;513:53241.
[40] Pakkenberg B, Pelvig D, Marner L, Bundgaard MJ, Gundersen HJG, Nyengaard JR,
et al. Aging and the human neocortex. Exp Gerontol 2003;38:959.
[41] Drake MT, Srinivasan B, Modder UI, Peterson JM, McCready LK, Riggs BL, et al. Effects
of parathyroid hormone treatment on circulating sclerostin levels in postmenopaus-
al women. J Clin Endocrinol Metab 2010;95:505662.
[42] Baron R, Kneissel M. WNT signaling in bone homeostasis and disease: from human
mutations to treatments. Nat Med 2013;19:17992.
[43] Dallas S. Osteocytes in motion. J Bone Miner Res 2013;28(Suppl. 1).
[44] Bianconi E, Piovesan A, Facchin F, Beraudi A, Casadei R, Frabetti F, et al. An estimation
of the number of cells in the human body. Ann Hum Biol 2013;40:46371.
[45] Marotti G. The osteocyte as a wiring transmission system. J Musculoskelet Neuronal
Interact 2000;1:133.
[46] Bivi N, Nelson MT, Faillace ME, Li J, Miller LM, Plotkin LI. Deletion of Cx43 from oste-
ocytes results in defective bone material properties but does not decrease extrinsic
strength in cortical bone. Calcif Tissue Int 2012;91:21524.
[47] Palumbo C, Palazzini S, Marotti G. Morphological study of intercellular junctions dur-
ing osteocyte differentiation. Bone 1990;11:4016.
[48] Tonna S, Takyar FM, Vrahnas C, Crimeen-Irwin B, Ho PWM, Poulton I, et al. EphrinB2
signaling in osteoblasts promotes their differentiation by preventing apoptosis.
FASEB J 2015;28:448296.
[49] Buenzli PR. Osteocyte as a record of bone formation dynamics: a mathematical
model of osteocyte generation in bone matrix. J Theor Biol 2015;364:41827.
[50] Sackmann E. Biological membranes architecture and function. In: Lipowsky R,
Sackmann E, editors. Handbook of biological physics, vol. 1. Elsevier; 1995.
[51] Spalding KL, Bergmann O, Alkass K, Bernard S, Salehpour M, Huttner HB, et al. Dy-
namics of hippocampal neurogenesis in adult humans. Cell 2013;153:121927.
[52] Belanger LF. Osteocytic osteolysis. Calcif Tissue Res 1969;4:112.
[53] Qing H, Ardeshirpor L, Dusevich V, Wysolmerski J, Bonewald LF. Osteocyte
perilacunar remodeling is regulated hormonally, but not by mechanical unloading.
J Bone Miner Res 2009;29:S440.
[54] Bonewald L. The holy grail of high bone mass. Nat Med 2011;17:6578.
[55] Kovacs CS, Kronenberg HM. Maternal-fetal calcium and bone metabolism during
pregnancy, puerperium, and lactation. Endocr Rev 1997;18:83272.
[56] Qing H, Ardeshirpour L, Pajevic PD, Dusevich V, Jahn K, Kato S, et al. Demonstration
of osteocytic perilacunar/canalicular remodeling in mice during lactation. J Bone
Miner Res 2012;27:101829.
[57] Heller-Steinberg M. Ground substance, bone salts, and cellular activity in bone for-
mation and destruction. Am J Anat 1951;89:34779.
[58] Belanger LF, Robichon J. Parathormone-induced osteolysis in dogs. A microradio-
graphic and alpharadiographic survey. J Bone Joint Surg Am 1964;46:100812.
[59] Teti A, Zallone A. Do osteocytes contribute to bone mineral homeostasis? Osteocytic
osteolysis revisited. Bone 2009;44:116.
[60] Kamioka H, Honjo T, Takano-Yamamoto T. A three-dimensional distribution of oste-
ocyte processes revealed by the combination of confocal laser scanning microscopy
and differential interference contrast microscopy. Bone 2001;28:1459.
[61] Hesse B, Varga P, Langer M, Pacureanu A, Schrof S, Mnnicke N, et al. Canalicular net-
work morphology is the major determinant of the spatial distribution of mass den-
sity in human bone tissue evidence by means of synchrotron radiation phase-
contrast nano-CT. J Bone Miner Res 2014;30:34656.
[62] Coxson HO, Rogers RM, Whittall KP, D'Yachkova Y, Pare PD, Sciurba FC, et al. A quan-
tication of the lung surface area in emphysema using computed tomography. Am J
Respir Crit Care Med 1999;159:8516.
[63] Helander HF, Fandriks L. Surface area of the digestive tract revisited. Scand J
Gastroenterol 2014;49:6819.
[64] Haycock GB, Schwartz GJ, Wisotsky DH. Geometric method for measuring body sur-
face area: a heightweight formula validated in infants, children, and adults. J
Pediatr 1978;93:626.
[65] Roth G, Dicke U. Evolution of the brain and intelligence. Trends Cogn Sci 2005;9:
2507.
[66] Marois R, Ivanoff J. Capacity limits of information processing in the brain. Trends
Cogn Sci 2005;9:296305.
[67] Herculano-Houzel S, Collins CE, Wong P, Kaas JH. Cellular scaling rules for primate
brains. Proc Natl Acad Sci U S A 2007;104:35627.
[68] Schweitzer MH, Zheng W, Cleland TP, Bern M. Molecular analyses of dinosaur oste-
ocytes support the presence of endogenous molecules. Bone 2013;52:41423.
[69] D'Emic MD, Benson RB. Measurement, variation, and scaling of osteocyte lacunae: a
case study in birds. Bone 2013;57:30010.
[70] Sanchez S, Dupret V, Tafforeau P, Trinajstic KM, Ryll B, Gouttenoire PJ, et al. 3D mi-
crostructural architecture of muscle attachments in extant and fossil vertebrates re-
vealed by synchrotron microtomography. PLoS One 2013;8:e56992.
[71] Stein KW, Werner J. Preliminary analysis of osteocyte lacunar density in long bones
of tetrapods: all measures are bigger in sauropod dinosaurs. PLoS One 2013;8:
e77109.
[72] Martin RB. Does osteocyte formation cause the nonlinear relling of osteons. Bone
2000;26:718.
[73] Hernandez CJ, Beaupr GS, Carter DR. A model of mechanobiologic and metabolic in-
uences on bone adaptation. J Rehabil Res Dev 2000;37:23544.
[74] van Oers RF, Ruimerman R, Tanck E, Hilbers PA, Huiskes R. A unied theory for
osteonal and hemi-osteonal remodeling. Bone 2008;42:2509.
[75] Baiotto S, Labat B, Vico L, Zidi M. Bone remodeling regulation under unloading con-
ditions: numerical investigations. Comput Biol Med 2009;39:4652.
[76] Adachi T, Kameo Y, Hojo M. Trabecular bone remodelling simulation considering
osteocytic response to uid-induced shear stress. Philos Trans A Math Phys Eng
Sci 2010;368:266982.
[77] van Oers RF, van Rietbergen B, Ito K, Hilbers PA, Huiskes R. A sclerostin-based theory
for strain-induced bone formation. Biomech Model Mechanobiol 2011;10:66370.
[78] Buenzli PR, Pivonka P, Gardiner BS, Smith DW. Modelling the anabolic response of
bone using a cell population model. J Theor Biol 2012;307:4252.
[79] Pivonka P, Buenzli PR, Scheiner S, Hellmich C, Dunstan CR. The inuence of bone sur-
face availability in bone remodellinga mathematical model including coupled geo-
metrical and biomechanical regulations of bone cells. Eng Struct 2013;47:13447.
[80] Andreaus U, Colloca M, Iacoviello D. Optimal bone density distributions: numerical
analysis of the osteocyte spatial inuence in bone remodeling. Comput Methods
Programs Biomed 2014;113:8091.
[81] Wang H, Ji B, Liu XS, van Oers RF, Guo XE, Huang Y, et al. Osteocyte-viability-based
simulations of trabecular bone loss and recovery in disuse and reloading. Biomech
Model Mechanobiol 2014;13:15366.
[82] Hopeld JJ. Neural networks and physical systems with emergent collective compu-
tational abilities. Proc Natl Acad Sci U S A 1982;79:25548.
[83] Bhalla US, Iyengar R. Emergent properties of networks of biological signaling path-
ways. Science 1999;283:3817.