Long-term culture of bovine nucleus pulposus explants
in a native environment
Bart G.M. van Dijk, MSc, Esther Potier, PhD, Keita Ito, MD, ScD*
Orthopaedic Biomechanics, Department of Biomedical Engineering, Eindhoven University of Technology, PO. Box 513,
GEM-Z 4.115, 5600 MB, Eindhoven, The Netherlands
Received 8 February 2012; revised 7 September 2012; accepted 9 December 2012
AbstractBACKGROUND CONTEXT: Chronic low back pain is a disease with tremendous financial
and social implications, and it is often caused by intervertebral disc degeneration. Regenerative
therapies for disc repair are promising treatments, but they need to be tested in physiological
PURPOSE: To develop a physiological in vitro explant model that incorporates the native environ-
ment of the intervertebral disc, for example, hypoxia, low glucose, and high tissue osmolarity.
STUDY DESIGN: Bovine nucleus pulposus (NP) explants were cultured for 42 days in conditions
mimicking the native physiological environment. Two different approaches were used to balance
the swelling pressure of the NP: raised medium osmolarity or an artificial annulus.
METHODS: Bovine NP explants were either cultured in media with osmolarity balanced at iso-
tonic and hypertonic levels compared with the native tissue or cultured inside a fiber jacket used as
an artificial annulus. Oxygen and glucose levels were set at either standard (21% O2and 4.5 g/L
glucose) or physiological (5% O2and 1 g/L glucose) levels. Samples were analyzed at Day 0, 3,
and 42 for tissue composition (water, sulfated glycosaminoglycans, DNA, and hydroxyproline con-
tents and fixed charge density), tissue histology, cell viability, and cellular behavior with messenger
RNA (mRNA) expression.
RESULTS: Both the hypertonic culture and the artificial annulus approach maintained the tissue
matrix composition for 42 days. At Day 3, mRNA expressions of aggrecan, collagen Type I, and
collagen Type II in both hypertonic and artificial annulus cultures were not different from Day
0; however, at Day 42, the artificial annulus preserved the mRNA expression closer to Day 0. Gene
expressions of matrix metalloprotease 13, tissue inhibitor of matrix metalloprotease 1, and tissue
inhibitor of matrix metalloprotease 2 were downregulated under physiological O2and glucose
levels, whereas the other parameters analyzed were not affected.
CONCLUSIONS: Although the hypertonic culture and the artificial annulus approach are both
promising models to test regenerative therapies, the artificial annulus was better able to maintain
a cellular behavior closer to the native tissue in longer term cultures.
? 2013 Elsevier Inc.
Open access under the Elsevier OA license.
Keywords:Regenerative therapy; Disc degeneration; Nucleus pulposus; Explant culture; Osmolarity
Up to 80% of the population will suffer at least once from
low back pain [1,2], and when this disease becomes chronic,
its financial and social implications are tremendous. Low
back pain can be caused by different mechanisms but is also
heavily associated with disc degeneration . During this
process, there is a shift from an anabolic to a catabolic envi-
crease their production of the main matrix proteins  and
increase their production of degrading enzymes [5–7]. As
FDA drug/device status: Not applicable.
Author disclosures: BGMvD: Nothing to disclose. EP: Nothing to dis-
close. KI: Grants: BMM (F); Board of Directors: AOSpine (E); Research
Support (Staff/Materials): DSM (amount unknown, paid directly to
The disclosure key can be found on the Table of Contents and at www.
* Corresponding author. Department of Biomedical Engineering, Eind-
hoven University of Technology, PO Box 513, GEM-Z 4.115, 5600 MB,
Eindhoven, The Netherlands. Tel.: (31) 40-2474350; fax: (31) 40-2473744.
E-mail address: K.Ito@tue.nl (K. Ito)
1529-9430 ? 2013 Elsevier Inc.
The Spine Journal 13 (2013) 454–463
Open access under the Elsevier OA license.
a result, the NP loses proteoglycans (PGs) and water  and
over time changes from a gel-like to a fibrous-like structure
. This leads to decreased disc height and loss of function,
which ultimately can cause pain [2,10].
therapy and spinal fusion, do not treat this degeneration and
have a limited long-term success. Regeneration of the disc is
an alternative long-term approach to treat the disc degenera-
tion at an early stage and prevent low back pain from occur-
ring [11,12]. A number of regenerative therapies have been
shown to be promising in animal models [13–16], as they
could delay the onset of degeneration [15,16] but still need
to be further developed because they could not fully restore
the disc to its original healthy state [15–17].
Before these therapies can be used, they need to be tested
ally involve induced degeneration by needle aspiration 
or enzymatic digestion , which is different from the hu-
put, cost intensive, and have ethical considerations [20,21].
Hence, using invitromodelsisappealing.Cellculture isless
costly and high-volume throughput; however, the native tis-
sueenvironment is lostduringcellisolation, which mightaf-
posus explant culture is another suitable model because the
earliest detectable changes during degeneration occur in
the NP  and the native tissue environment can be main-
tained  in such a model.
Previously we have successfully cultured bovine NP ex-
plants for 21 days . In that study we used polyethylene
glycol (PEG) to raise the medium osmolarity to the native
level and showed that we were able to prevent the swelling
and PG loss. However, the cellular behavior was different
from the fresh tissue at the gene level. We hypothesized that
although balancing the osmolarity is important to prevent
PG loss at the tissue level, other factors in the native envi-
ronment (eg, hypoxia, low glucose levels, and low pH )
might be important for the cellular behavior. Indeed, it has
been shown that glycosaminoglycan (GAG) production is
maximal at 5% O2for NP explants cultured for 24 hours
 and also at a pH of 7.1 when cultured for 4 hours
. Therefore, it appears important to culture NP explants
under physiological O2, glucose, and pH .
A second possible explanation for the difference in the
cellular behavior we previously observed is that a balanced
medium osmolarity is different from the physiological situ-
ation in which the annulus fibrosus contains NP swelling.
Our second hypothesis is that an artificial annulus approach
is more physiological than a balanced medium osmolarity
approach and might benefit the cellular behavior in NP
To see the effects of regenerative therapies on the tissue
level, we need a model that is stable for even longer than 21
days of culture; therefore, the culture duration was ex-
tended to 42 days. Furthermore, we analyzed the explants
at Day 3 to assess if the changes in the cellular behavior
were a direct effect of the harvesting and culture initiation
stress or occurred over time in culture.
The aim of this study was to improve our NP explant
model by incorporating more factors of the native physio-
logical environment and to determine the importance of
the separate factors. Therefore, we cultured NP explants
for 6 weeks under physiological O2and glucose levels in
a swelling balanced environment using either an osmotic
balance or an artificial annulus approach. We analyzed
the tissue composition with biochemical assays and histol-
ogy and the cellular behavior with gene expression.
Materials and methods
NP explants were harvested from fresh caudal discs of
24-month-old cows. These were obtained from the abattoir
according to the local regulations, and a total of 70 discs
(CC2–CC5) were harvested from 18 donors. The different
levels were distributed equally and randomly among the
different conditions. The discs were opened transversally
directly underneath the end plate, and NP explants were
punched out with an 8-mm diameter biopsy punch (Kruuse,
Sherburn, UK) from the center of the NP.
Standard medium was prepared from basic Dulbecco’s
Modified Eagle Medium powder (Gibco; Invitrogen, Carls-
bad, CA, USA) in milli-Q filtered water (8.3 g/L), supple-
mented with 15.9 mg/L phenol red (Sigma, Zwijndrecht,
The Netherlands), 2% L-glutamine (Lonza, Basel, Switzer-
land), 1% pyruvate (Gibco), 1% penicillin/streptomycin
(Lonza), 3.7 g/L sodium bicarbonate (Sigma), 50 mg/L as-
corbic acid (Sigma), and 10% fetal bovine serum (Gibco).
Glucose was added to the media for a final concentration of
ical conditions. All media were filter sterilized and the pH
was adjusted to 7.1, the pH of a healthy human disc .
In PEG culture, the medium osmolarity was adjusted to
two levels based on the previous study 
? Isotonic to native in situ NP (430 mOsm/kg H2O):
standard mediumþ8.2% w/v PEG (20kD, Sigma)
? Hypertonic to native in situ NP (570 mOsm/kg H2O):
standard mediumþ13.3% w/v PEG.
NPexplantswere placed inside the dialysistubing (15 kD
molecular weight cut-off, Spectra-Por, Rancho Dominguez,
455B.G.M. van Dijk et al. / The Spine Journal 13 (2013) 454–463
CA, USA) and closed with custom-made plastic rings to
prevent PEG, which may be cytotoxic , from entering
the tissue (Fig. 1, Left). Subsequently, these samples were
immediately placed in the corresponding PEG media.
Artificial annulus culture
We observed in pilot studies that explants placed directly
in the artificial annulus swelled to some extent because of
excess space (necessary for the insertion of tissue). There-
fore, the explants were first placed in dialysis tubing for
100 minutes in a 50% w/v PEG solution in phosphate-
buffered saline to shrink the samples to approximately
20%. Samples were then taken out of the PEG solution
and the clips were removed. The dialysis tubing was cut
to size and the samples were placed inside a fiber jacket.
These jackets were custom knitted (Varodem, Saint-L? eger,
Belgium) from a single fiber of ultra-high molecular weight
polyethylene (Dyneema Purity, DSM, Heerlen, The Nether-
lands) as an 8-mm diameter tube, which was closed at one
end (similar to a sock). To ensure that the jackets could
withstand the swelling of the samples, three layers were
wrapped around the explants (twisted at the open end
and turned inside out) before they were sewn closed with
a Dyneema Purity fiber suture. Then the artificial annulus
samples were placed in a standard medium (Fig. 1, Right).
Standard and physiological levels
Samples were cultured at 37?C and 5% CO2, and the me-
diawere changedtwotimesaweek. Thesamples were either
media (standard conditions) or in a hypoxic incubator at 5%
O2in 1 g/L glucose media (physiological conditions).
Samples were analyzed at Day 3 and Day 42. For every
culture condition, five samples (corresponding to five do-
Theotherhalf was cut in two pieces, oneforgeneexpression
and one for biochemical assays. As a control, Day 0 samples
(fresh native tissue) were analyzed immediately after har-
vesting (n55 discs each from different donors) in the same
way. Because of concerns about infection risk and reproduc-
in duplo (10 samples from 10 donors per group). To investi-
gate the swelling behavior, the weight of all samples was
measured before and after culture.
Biochemical and water content
At Day 0, 3, and 42 the samples were blotted dry,
weighed, and subsequently stored frozen at ?30?C. The
samples were then placed in a lyophilizer (Freezone 2.5;
Labconco, Kansas City, MO, USA) overnight and the dry
weight was measured. The water content was calculated
from the difference between the wet and dry weight,
divided by the wet weight. The dried samples were then di-
gested in papain solution (100 mM phosphate buffer, 5 mM
L-cystein, 5 mM ethylene diamine tetraacetic acid, and
125–140 mg/mL papain, all from Sigma) overnight at
60?C. The digested samples were then used to determine
their content of sulfated glycosaminoglycans (sGAG) as
a measure of PGs, hydroxyproline (HYP) as a measure
for collagen content, and DNA. sGAG content was deter-
mined with the dimethylmethyleneblue assay  using
shark cartilage chondroitin sulfate as a reference (Sigma).
The fixed charge density (FCD) was determined from the
GAG content per wet weight, similar to the method used
by Narmoneva et al. . HYP content was measured using
the chloramine-T assay  and a trans-4-hydroxyproline
(Sigma) reference. DNA content was measured using the
Hoechst dye method , with a calf thymus DNA refer-
ence (Sigma). The amounts of sGAG, HYP, and DNAwere
expressed per mg dry weight of the tissue.
N2and stored at ?80?C until RNA isolation. A 316 stainless
Fig. 1. Culture systems used to prevent swelling. (Left) In polyethylene glycol culture, explants were placed inside a semipermeable membrane, which was
closed with two custom-made closure rings. (Right) In artificial annulus culture, explants were placed inside a semipermeable membrane and subsequently in
a fiber jacket that was sewn closed.
456B.G.M. van Dijk et al. / The Spine Journal 13 (2013) 454–463
steel 8-mm bead and custom-made lid were placed in a 2-mL
tube (Eppendorf, Nijmegen, The Netherlands) and snap
frozen in liquid N2. Frozen samples were placed in between
the bead and the lid and disrupted with a mikrodismembrator
(Sartorius, Goettingen, Germany) for 20 seconds at
1,500 rpm. This was repeated if necessary with the sample
snap frozen between each cycle. After disruption, RNA was
gen mini-kit (Qiagen, Venlo, The Netherlands). The quantity
and purity of the isolated RNAwas measured with a spectro-
photometer (ND-1000; Isogen, de Meern, The Netherlands).
Absence of genomic DNA contamination in the isolated
RNAwas checked using real time polymerase chain reaction
(iCycler; Biorad, Veenendaal, The Netherlands) with glycer-
aldehyde3-phosphate dehydrogenase primers.The VILO-kit
(Invitrogen) was used to reverse-transcribe 60 mg total RNA.
The genes of interest were aggrecan, collagen Type I and II,
matrix metalloprotease 13 (MMP-13), and tissue inhibitor
of matrix metalloprotease 1 and 2 (TIMP-1 and TIMP-2) us-
was selected as a reference gene from three genes (RPL13A,
real time polymerase chain reaction (CFX384, Biorad) and
the expression difference (DCt) was calculated as the differ-
Levels of expression are expressed as 2?(DCt).
Histology and cell viability
Samples were covered in tissue-tek O.C.T. (Sakura,
Alphen a/d Rijn, The Netherlands), snap-frozen in isopen-
tane in liquid N2, and stored at ?30?C. Subsequently,
dorf, Germany) and stored at ?30?C. Sections were either
fixed with 3.7% formalin, stained with Weigert hematoxylin
for cell nuclei, safranin-O for PGs, and Fast Green for colla-
cell viability  and fixed. Images were taken with a com-
bined fluorescence and bright-field microscope (Zeiss
Observer; Zeiss, Sliedrecht, The Netherlands). Fluorescence
and bright field images of the same region of interest were
combined using ImageJ software (National Institutes of
Health, Bethesda, US) for the LDH–stained sections.
On the iso- and hypertonic PEG, the artificial annulus
cultures with physiological conditions, and the Day 0 sam-
ples, one-way analysis of variance was performed, followed
by Dunnett test for differences compared with Day 0. To in-
vestigate the differential effects of the standard and physi-
ological culture conditions, two-way analysis of variance
was performed for all the culture groups, followed by Bon-
ferroni corrected post hoc tests. Statistical significance in
all cases was assumed for p!.05.
The samples in the hypertonic PEG group did not in-
crease in wet weight. In 3 of 10 samples in both artificial
annulus culture groups, the samples increased between
20% and 30% in wet weight. We assumed that the jackets
were not properly closed in these samples and the swelling
was not balanced. These samples were excluded from the
study, leaving n57 for all artificial annulus culture groups
at Day 42.
Is artificial annulus culture beneficial to an osmotic
After 42 days, the water content increased slightly for
isotonic PEG and artificial annulus compared with the fresh
tissue (Day 0) when cultured under physiological condi-
tions (Fig. 2A). The sGAG, DNA, and HYP content were
not different from Day 0 (Fig. 2B, D, and E) for all condi-
tions. The FCD in isotonic PEG decreased compared with
Day 0 (Fig. 2C) but was not different in hypertonic PEG
and artificial annulus condition. Cell viability, assessed
with LDH staining, was maintained in all culture groups af-
ter 42 days in culture with physiological O2and glucose
levels (Fig. 3).The Safranin-O/Fast Green staining showed
no differences in the tissue composition for all culture
groups with physiological O2and glucose levels when com-
pared with the fresh tissue (Fig. 4). Aggrecan and TIMP-1
messenger RNA (mRNA) expressions were significantly
downregulated in all culture groups compared with Day
0 (Fig. 5A and E). Collagen Type I and Type II expressions
were downregulated in both PEG conditions compared with
Day 0 but not in the artificial annulus condition (Fig. 5B
and C). Matrix metalloprotease 13 was not detected at
Day 0 and in 4 of 5 of the hypertonic PEG and 6 of 7 of
the artificial annulus samples. In isotonic PEG culture,
MMP-13 expression was significantly upregulated com-
pared with Day 0 (Fig. 5D). No significant differences were
observed for TIMP-2 expression for all culture groups com-
pared with Day 0 (Fig. 5F).
Are physiological oxygen and glucose beneficial to
Only in isotonic PEG culture with physiological condi-
tions, there was a slight increase in water content compared
with the standard conditions (Fig. 6A). However, there were
no significant effects of the physiological conditions on the
sGAG, FCD, DNA, and HYP content (Fig. 6B–E). After 42
days, there were no differences in the tissue morphology
(Fig. 7) and cell viability (not shown) between physiologi-
cal and standard O2and glucose levels for all groups. Ag-
grecan and collagen Type II mRNA expressions were not
different between physiological and standard O2and glu-
cose levels (Fig. 8A and B). However, there was an effect
457B.G.M. van Dijk et al. / The Spine Journal 13 (2013) 454–463
of culture group: expression of both genes was higher in
artificial annulus condition than in both PEG conditions.
There were no significant differences in collagen Type I
expression in all groups (Fig. 8C). MMP-13, TIMP-1, and
TIMP-2 expressions were significantly downregulated with
physiological O2and glucose levels compared with stan-
dard (Fig. 8D–F).
Do changes in gene expression occur immediately or
gradually over time?
At Day 3, there were no changes in the tissue morphol-
ogy, viability, and biochemical content (not shown). Aggre-
can mRNA expression was downregulated at Day 3 in
isotonic PEG and artificial annulus culture compared with
Fig. 2. (A) Water content, (B) sGAG content expressed per dry weight, (C) FCD, (D) DNA content expressed per dry weight, and (E) HYP content expressed
per dry weight, after 42 days in culture. Values are mean6standard deviation; n55 (7 for artificial annulus); *p!.05 compared with Day 0. sGAG, sulfated
glycosaminoglycans; FCD, fixed charge density; HYP, hydroxyproline; PEG, polyethylene glycol.
Fig. 3. Lactate dehydrogenase (LDH)–stained sections of fresh and 42-days cultured nucleus pulposus explants. (A) Day 0; (B) isotonic polyethylene glycol
(PEG); (C) hypertonic PEG; and (D) artificial annulus with physiological O2and glucose levels. Dark blue indicates living cells (active LDH), and red
indicates cell nuclei (either living or dead). Scale bar5200 mm.
458B.G.M. van Dijk et al. / The Spine Journal 13 (2013) 454–463