www.nature.

com/scientificreports

OPEN Molecular Therapy for Degenerative

Disc Disease: Clues from Secretome
Analysis of the Notochordal Cell-
received: 17 June 2016
accepted: 01 March 2017 Rich Nucleus Pulposus
Published: 30 March 2017
AjayMatta1, M.ZiaKarim1, DavidE.Isenman2 & W.MarkErwin1,3,4

Degenerative disc disease (DDD) is associated with spinal pain often leading to long-term disability.
However, the non-chondrodystrophic canine intervertebral disc is protected from the development
of DDD, ostensibly due to its retention of notochordal cells (NC) in the nucleus pulposus (NP). In this
study, we hypothesized that secretome analysis of the NC-rich NP will lead to the identification of key
proteins that delay the onset of DDD. Using mass-spectrometry, we identified 303 proteins including
components of TGF- and Wnt-signaling, anti-angiogeneic factors and proteins that inhibit axonal
ingrowth in the bioactive fractions of serum free, notochordal cell derived conditioned medium (NCCM).
Ingenuity Pathway Analysis revealed TGF1 and CTGF as major hubs in protein interaction networks. In
vitro treatment with TGF1 and CTGF promoted the synthesis of healthy extra-cellular matrix proteins,
increased cell proliferation and reduced cell death in human degenerative disc NP cells. A single intra-
discal injection of recombinant TGF1 and CTGF proteins in a pre-clinical rat-tail disc injury model
restored the NC and stem cell rich NP. In conclusion, we demonstrate the potential of TGF1 and CTGF
to mitigate the progression of disc degeneration and the potential use of these molecules in a molecular
therapy to treat the degenerative disc.

Degenerative disc disease (DDD) is a predominant contributor (~40%) to the genesis of spinal pain and is a major
cause of disability worldwide, imposing enormous socio-economic burden and clinical costs to society1. The
2010 Global Burden of Disease Study estimated that low back pain is among the top 10 diseases and injuries that
account for the highest number of disability-adjusted life years worldwide. It is estimated that approximately 25%
of the population in United States of America suffers from lower back and neck pain at costs greater than $100
billion annually2. Current surgical treatments including fusion and disc replacement do not attempt to slow the
degenerative process or promote repair3. In fact, a meta-analysis suggested that lumbar fusion might result in a
higher prevalence of adjacent segment degeneration or disease than motion-preserving procedures4. Thus, there
exists a need for the identification of novel, effective therapeutic agents that can be clinically translated as mini-
mally invasive regenerative therapies for disc repair. Several approaches including gene therapy, growth factors,
stem cells and tissue engineering have been proposed for the treatment of disc degeneration510. Growth factors
including insulin-like growth factor-1 (IGF-1), fibroblast growth factor-2 (FGF-2) and morphogens such as bone
morphogenetic protein-2 (BMP-2), osteogenic protein-1(OP-1) and growth differentiation factor-5 (GDF-5) have
been investigated for their potential to stimulate extracellular matrix (ECM) synthesis in intervertebral disc (IVD)
cells6. However, the use of BMP-2 for the treatment of DDD in humans is limited by its osteo-inductive charac-
teristics11. Similarly, the injection of rhBMP-7/OP-1 resulted in exogenous bone formation with no evidence of
disc regeneration in the beagle IVD - NP12. Moreover, administration of anti-inflammatory agents including
interleukin 1 receptor antagonist (IL-1Ra) and synthetic inhibitors targeting tumor necrosis factor (TNF) or
nuclear factor kappa B (NFB) have shown limited efficacy in stimulating the anabolic response13,14. Therefore,
the identification of novel biologically active agents that can restore the homeostatic microenvironment in a
degenerative disc would revolutionize the treatment of DDD.

1
Krembil Research Institute, Toronto Western Hospital, Toronto, Ontario, Canada. 2Department of Biochemistry,
Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada. 3Department of Surgery, University of Toronto,
Toronto, Ontario, Canada. 4Division of Research, Canadian Memorial Chiropractic College, Canada. Correspondence
and requests for materials should be addressed to W.M.E. (email: [email protected])

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 1

www.nature.com/scientificreports/

The intervertebral disc (IVD) is composed of a central nucleus pulposus (NP) surrounded by the concentric
annulus fibrosus and attached to the adjacent vertebrae by thin cartilaginous end plates15,16. The NP plays an
important role in maintaining the biomechanical properties of the spine such that in youth the healthy NP is rich
in large, vacuolated notochordal cells (NCs) and possesses a proteoglycan rich ECM. However, NCs observed
during childhood are gradually replaced by small chondrocyte-like cells (CLCs) by early adolescence in humans,
a NP cellular phenotype often associated with disc degeneration15,16. The degenerative disc NP is a catabolic phe-
notype representing a failure of homeostatic regulation of the microenvironment with the increased expression
of pro-inflammatory cytokines and matrix metalloproteinases (MMPs) leading to progressive cell death, loss of
proteoglycan content and development of an inferior, fibrocartilaginous extra-cellular matrix (ECM). This affects
IVD structural integrity, compromising its biomechanical properties often leading to pain, neurological compro-
mise and disability1519. With respect to the development of DDD, a temporal relationship has been suggested
between the loss of NCs and the progression of DDD1517.
Unlike humans, non-chondrodystrophic (NCD) canines, porcine and rodents preserve large, vacuolated NCs
within their NPs and are relatively resistant to DDD, until much later in their lives19. In our previous studies,
we reported that conditioned medium collected from alginate cultures of notochordal cells (NCs) derived from
NCD-canines reduced cell death and stimulated collagen 2 synthesis in bovine NP cells20,21. Similarly, other stud-
ies have demonstrated increased proteoglycan synthesis in NP cells treated with notochordal cell derived condi-
tioned medium (NCCM) in vitro22. A recent study also demonstrated the anti-angiogeneic effects of NCCM in
vitro23. These studies provide compelling evidence that NCs secrete proteins, which may confer anabolic char-
acteristics, maintain a homeostatic microenvironment in NP and resists DDD. We hypothesized that secretome
analysis of NC-rich NPs will give us insight into the protein networks regulating the development of DDD and
may be exploited for regenerating the degenerative discs in humans. In this study, we identified the proteins
in serum free NCCM using mass spectrometry and, improved and characterized a pre-clinical rodent disc -
injury model of DDD. We used Ingenuity Pathway Analysis (IPA) to identify protein networks that might be
of significance in the homeostatic regulation of the healthy disc NP. Next we evaluated the regenerative and
anti-inflammatory potential of connective tissue growth factor (CTGF) and transforming growth factor beta 1
(TGF1), identified as major hub proteins in network analysis, using in vitro assays and an in vivo pre-clinical
rodent disc - injury model of DDD.

Results
Evaluation of a pre-clinical rat-tail disc injury model of DDD. We characterized a pre-clinical rodent
model of DDD to establish a platform for the evaluation of therapeutic agents. Using fluoroscopic image guid-
ance, we performed needle puncture injuries using a 27-gauge needle in the caudal (tail) discs of 12 week-old,
healthy Wistar rats (n=21, 4 discs per animal). Alterations in the composition of the ECM and the cellular
phenotype were determined in a time dependent manner (72hrs10 weeks, Fig.1AD). Histological analysis
revealed a gradual loss of NCs, increased Safranin-O staining intensity and reduced expression of aggrecan
and collagen 2 by the end of 10 weeks post injury (p.i.) in comparison to uninjured, healthy control IVD-NP
obtained from age matched healthy controls (Fig.1B, SupplementaryFig.1A). Of note, needle puncture injury
increased the expression of the pro-inflammatory cytokines, TNFwithin 72hrs and interleukin-1 (IL-1)
as early as 1 week p.i. (Fig.1C, SupplementaryFig.1B). However, the active form of IL-1(~17kDa) was not
observed until between 810 weeks, coincident with an increased expression of the inflammatory mediator,
cyclooxygenase 2 (COX2) and the ECM degrading enzymes, matrix metalloproteinases-3 (MMP-3) and MMP-
13 as compared to no treatment controls (NTCs, i.e. tissue lysates obtained from uninjured IVD-NPs, Fig.1C,
SupplementaryFig.1B). Interestingly, the loss of tissue inhibitor of metalloproteinases 1 (TIMP1), a natural
inhibitor of MMPs was also observed at the end of 10 weeks p.i. coincident with the onset of increased levels of the
MMPs (Fig.1C, SupplementaryFig.1B). We also observed increased expression of A disintegrin-like and metal-
loprotease with thrombospondin type 1 motif 4 (ADAMTS4), one of the major enzymes responsible for aggrecan
degradation (Fig.1C, SupplementaryFig.1B). In addition to ECM changes, we also determined the expression of
notochordal and stem cell markers in healthy and injured IVD-NP lysates. Our western blotting results demon-
strated the loss of the NC markers (brachyury and galectin 3) and stem cell markers (Oct4 and Nanog) in the
injured disc NPs, 10 weeks p.i. (Fig.1D, SupplementaryFig.1C). These findings clearly demonstrated a shift in the
NP milieu from a healthy, homeostatically regulated environment to a pro-inflammatory, catabolic state with loss
of both NCs and stem cells, similar to the human degenerative disc NP. Comparative analysis of the histological
characteristics of human degenerative disc NPs with bovine, NCD - canine and Wistar rat IVDs is described in
Supplementary Data (SupplementaryFig.2A). Negative control tissue sections (rat healthy/mongrel IVD NP)
wherein the primary antibody was replaced by rabbit or goat isotype IgG controls revealed no immunostaining
thereby demonstrating the specificity of the antibodies (SupplementaryFig.2B).

Intra-discal injection of NCCM promotes regeneration of the degenerative disc nucleus pulpous
in a pre-clinical rodent model of DDD. We injected concentrated, serum free NCCM or control medium
(~8L/disc) into the injured (4-week p.i.) rat-tail disc NPs (n=6 animals/group, 4 discs per animal) using a
specially designed, 32-gauge needle with a 35 degree bevel under fluoroscopic guidance. Ten weeks p.i., histolog-
ical analysis revealed NC-rich NPs with moderate Safranin-O staining in NCCM injected rat-tail injured discs
(Fig.2A). In contrast, rat-tail injured discs injected with control medium showed low cellularity and displayed a
fibrocartilaginous matrix, with intense Safranin-O staining (Fig.2A). Immunohistochemistry and western blot-
ting revealed that discs injected with NCCM displayed restoration of aggrecan, collagen 2, brachyury, Oct4 and
Nanog in contrast to the sham/control medium injected discs (Fig.2A,B, SupplementaryFig.2C). However, pro-
tein extraction using RIPA lysis buffer is a potential limitation for estimation of Collagen2 expression in tissues

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 2

www.nature.com/scientificreports/

Figure 1. Needle puncture injury in rat-tail disc leads to the development of fibrocartilaginous matrix
and loss of notochordal (NC) and stem cells in nucleus pulposus (NP). (A) Representative fluoroscopic
image-guided needle puncture injury in rat-tail disc NP. (B) Histological analysis demonstrating representative
hematoxylin and eosin (H&E) and Safranin O staining images showing development of a fibrocartilaginous
matrix over a period of 10 weeks p.i. in rat NP. Both H&E and Safranin O staining was done in duplicates on
IVD sections obtained from all the animals in each group. Immunohistochemistry showing the loss of the ECM
proteins, aggrecan and collagen 2 in time dependent manner (healthy to 10 weeks p.i., Scale bar 50, n = 2 per
group). (C) Representative Western blot panels showing alterations in the expression of pro-inflammatory
cytokines (IL-1and TNF), inflammation mediator, Cox2 and ECM proteins (MMP-3, MMP-13, TIMP1
and ADAMTS4) in a time dependent manner in post-injury rat NP tissue lysates. (D) Representative Western
blot panels showing loss of NC markers (brachyury, galectin 3) and stem cell markers (Oct4, Nanog) in NPs
obtained from rat-tail injured discs over a period of 10 weeks. -actin was used a loading control in Western
blots. All Western blots were run under the same experimental conditions as described in Materials and
Methods.

using Western blotting. Nonetheless, these results validated our hypothesis and demonstrated that soluble factors
within NCCM have therapeutic potential for a degenerative IVD - NP.

Identification of soluble factors in NCCM using mass spectrometry. Serum-free NCCM was con-
centrated sequentially using 50kDa and 3kDa filters, followed by fractionation using size exclusion chroma-
tography (Fig.3A). Of the 30 fractions collected, only five of these fractions (50PF4, 50PF5, 50PF6, 3PF4 and
3PF5) suppressed etoposide - induced caspase-3/7 activity in bovine NP cells (SupplementaryFig.3AC). Using
mass-spectrometry analysis, we identified 303 non-redundant proteins in these bioactive fractions of NCCM, of
which 31% proteins have been reported within the ECM (Fig.3B, Table1, SupplementaryTableS1). We identified
several growth factors including transforming growth factor beta 1 (TGF1), connective tissue growth factor
(CTGF), Wnt-induced soluble protein 2 (WISP2), insulin-like growth factor binding protein 7, angiopoietin-like
7 and their modulators, including chordin, sclerostin and cartilage intermediate layer protein (CILP, Fig.3C,
Table1). For a complete list of proteins see SupplementaryTableS1. Immunohistochemical analysis demon-
strated moderate to strong immunostaining of CTGF, Wntinducible soluble protein 2 (WISP2) and TGF1 in
NC cells and ECM in healthy rat-tail disc NPs. However, no detectable expression of CTGF, WISP2 or TGF1
was observed in the ECM of degenerative needle puncture injured rat-tail or human disc NPs (Fig.3D). These
findings suggested an association between the loss of CTGF, WISP2 and TGF1 with the development of DDD.

Network Analysis of Proteins Identified in Secretome Analysis. Using Ingenuity Pathway

Analysis (IPA) software, we analyzed the networks inherent to the healthy, notochordal cell rich NP. Among
the major signaling pathways, TGF , Wnt/- catenin, axonal guidance, nitric oxide synthase (eNOS),

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 3

www.nature.com/scientificreports/

Figure 2. NCCM confers anabolic and anti-catabolic characteristics to degenerating NP in rat-tail injured
IVD - NP. (A) Representative Safranin O staining sections showing fibro-cartilaginous matrix and small NP
cells in rat tail injured disc NPs injected with control medium as compared to the injured disc NPs injected with
serum free NCCM showing proteoglycan-rich ECM with large notochordal cells (n=12 tissue sections/group).
Representative panel showing immunohistochemistry of aggrecan, collagen 2, brachyury and Oct4 in paraffin
embedded sections of rat-tail injured discs treated with protein free Hybridoma medium used as control or
NCCM (Scale bar 50). Immunohistochemistry for all the proteins was performed in duplicates in at least 3
different IVD sections obtained from different animals from each group used in this study. (B) Representative
Western blot panels showing expression of collagen 2 and the stem cell markers Oct4 and Nanog in tissue
lysates obtained from healthy controls, rat-tail injured disc NPs left untreated, injected with control medium or
NCCM.

leucocyte extravasation signaling and inflammation associated matrix degradation emerged as significant
pathways (p<0.05, SupplementaryFig.4AE). In a hierarchical model, TGF1 emerged as the top pro-
tein hub interacting or regulating the expression of several other proteins including CTGF, platelet-derived
growth factor (PDGF), secreted protein acidic and rich in cysteine (SPARC), MMPs and collagens identi-
fied in this study (Fig.4). In another model, CTGF appeared as a central hub protein interacting with BMPs,
TGF1, Smads, WISP2, several ECM proteins (aggrecan, collagens, decorin, HAPLN1, fibulin1) and kinases
(SupplementaryFig.4B).

Evaluation of the anabolic effects of CTGF, WISP2 and TGF1 on NP cells. We treated rat-tail disc
NP cells with recombinant human (rh) CTGF, WISP2 or TGF1 proteins in order to evaluate their effect upon cell
viability in a dose and time dependent manner (5ng/mL100ng/mL, 24hr96hrs, SupplementaryFig.5AD).
Treatment with CTGF (100ng/ml) or TGF1 (10ng/ml) alone or in combination increased the viability of NP
cells derived from healthy rat-tail discs and degenerative discs (rat-tail injured discs, bovine and human) in
48hrs72hrs (Fig.5AC, SupplementaryFig.6AC). Treatment with a combination of CTGF (100ng/ml) and
TGF1 (10ng/ml) increased cell viability by 35% within 72hrs in human degenerated disc NP cells (Fig.5C,
SupplementaryFig.6C). However, treatment with WISP2 showed no significant differences in cell viability of NP
cells derived from rat, bovine and human discs (Fig.5AC). This was further confirmed with cell proliferation
assays using bromodeoxyuridine (BrdU) incorporation. We observed a significant increase in DNA synthesis in
rat-tail (healthy/injured) and human degenerated disc NP cells upon treatment with TGF1 (10ng/ml) alone
or in combination with CTGF (100ng/ml) within 72hrs (Fig.5D,E). We also observed increased mRNA levels
of important ECM proteins including collagen type 2, hyaluronan and proteoglycan link protein 1 (HAPLN1),

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 4

www.nature.com/scientificreports/

Figure 3. Analysis of NCCM. (A) Schematic representation of the methodology for identification of proteins
in NCCM using mass - spectrometry. (B) Pie-chart showing distribution of ECM proteins identified in NCCM.
(C) Peptide signature peaks for CTGF, WISP2 and TGF1 observed in mass spectrometry analysis of NCCM.
(D) Immunohistochemistry showing expression of CTGF, WISP-2 and TGF1 in paraffin-embedded sections of
rat NP (healthy and injured discs) and human degenerative disc NP (Scale bar 50).

versican and thrombospondin1 (THBS1) in human degenerative disc NP cells upon treatment with a combina-
tion of CTGF and TGF1 (Fig.5F). Western blotting verified an increase in collagen 2 expression on treatment
with a combination of CTGF and TGF1 within 24hrs (SupplementaryFig.7AandB), supporting their anabolic
roles on human degenerated disc NP cells.

Evaluation of anti-catabolic effects of CTGF, WISP2 and TGF1 on NP cells. Human and rat NP
cells were treated with pro-inflammatory cytokines, IL-1(10ng/ml) alone or a combination of IL-1 (10 ng/ml)
and TNF(50ng/ml) in the presence of CTGF (100ng/ml), WISP2 (100ng/ml) and TGF1 (10ng/ml) alone or
in combination for 2448hrs. Treatment with IL-1alone or a combination of IL-1and TNFinduced apop-
tosis as revealed by increased caspase-3/7 activity in NP cells (Fig.6AC). However, a significant decrease in
inflammationinduced caspase-3/7 activity was observed in degenerative disc NP cells (rat/human) in the pres-
ence of TGF1 alone or combined with CTGF (Fig.6AC). A significant reduction in IL-1and TNF induced
caspase-9 activity was also observed in the presence of CTGF, WISP2 or TGF1 in human degenerative disc NP
cells (Fig.6E,F). Treatment with CTGF, WISP2 or TGF1 reduced expression of MMP-3, MMP-13 and Cox2
proteins in rat NP cells treated with IL-1and TNF(Fig.6G,H, SupplementaryFig.8,AandB). Treatment of
human degenerative NP cells with the pro-inflammatory cytokines, IL-1and TNF, induced expression of Cox2
(H1 and H2) and MMP13 in all 3 cases (H1, H2 and H3) as compared to their respective no treatment controls
(NTC, Fig.6I,J). However, the addition of CTGF and TGF1 reduced Cox2 and MMP-13 mRNA levels induced
by IL-1and TNFin human degenerative discs NP cells demonstrating their anti-catabolic effects (Fig.6I,J).

Intra-discal injection of CTGF and TGF1 mediates progression of degenerative disc disease in
a pre-clinical rodent model. To test the potential of CTGF and TGF1 to mediate progression of DDD in
a pre-clinical rodent model, we performed image guided rat-tail disc injuries (n=24) in two independent experi-
ments. Four weeks following injury, the animals were randomized into four groups (n=6 animals/group) and an
intra-discal injection (~8L) of CTGF (100ng/mL i.e. 0.80ng/disc), TGF1 (10ng/mL, i.e. 0.08ng/disc), a combi-
nation of CTGF (100ng/mL) and TGF1 (10ng/mL) or phosphate buffered saline (PBS, 1X, pH=7.2) as a vehicle
control was administered in each injured disc NPs (4 discs per animal). Histological analysis of discs injected with
PBS (1X, pH=7.2) showed a fibrocartilaginous, acellular matrix with intense Safranin-O staining within the NP
(Fig.7AC). However, injured rat-tail discs treated with CTGF, TGF1 alone or in combination, demonstrated
a healthy disc, rich in NCs, 10 weeks p.i. (Fig.7AC). Immunohistochemical analysis confirmed the restoration
of a healthy NP on treatment with rh CTGF or TGF1, demonstrating strong expression of aggrecan, collagen

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 5

www.nature.com/scientificreports/

S. No. Accession Number Protein Name-FASTA Function/Cell Signaling References

Involved in development of notochord,
1 F1PW10_CANFA (+1) Transforming Growth Factor Beta-1 24,26
Stimulates ECM synthesis
Involved in development of notochord,
2 J9NTX8_CANFA Connective Tissue Growth Factor Stimulates ECM synthesis, Regulates TGF 2731
signaling
3 E2RJ75_CANFA WNT1 Inducible Signaling Pathway Protein 2(CCN5) Negatively regulates TGF signaling 32
4 F1PM26_CANFA CD109 Negatively regulates TGF signaling 3436
5 F6V790_CANFA CILP Regulates TGF signaling 37
6 PGS2_CANFA Decorin Negatively regulates TGF signaling 38
7 E2RJE0_CANFA [2] Cartilage Oligomeric Matrix Protein1 Regulates TGF signaling 39
8 J9P9E4_CANFA sclerostin (BMP antagonist) Negatively regulates TGF signaling 40
9 E2RBE4_CANFA (+1) Chordin Negatively regulates TGF signaling 40
10 F1PY69_CANFA (+1) Follistatin-Like Protein 1 Regulates TGF signaling 51
Promotes matrix mineralization through TGF
11 F6XL96_CANFA Von Willebrand Factor A Domain-Containing Protein 1 52
signaling
12 E2R612_CANFA EGF-Containing Fibulin-Like Extracellular Matrix Protein Regulates cell proliferation and apoptosis 53
Involved in muscle and bone linkage. Activates
13 E2RNR0_CANFA Osteoglycin 54
Samd3/4 independent of TGF.
14 E2R8H9_CANFA (+1) pleiotrophin, Neurite Growth-Promoting Factor 11 Involved in Angiogenesis and neurite growth 55
Involved in ECM assembly, cell invasion,
15 F1P7Y6_CANFA nidogen-2 56
involved in Angiogenesis
Binds to BMP4 enhances the canonical
16 F1PQV4_CANFA Olfactomedin-Like 3 SMAD1/5/8 signaling pathway, involved in 57
Angiogenesis
17 F1PN55_CANFA Frizzled-Related Protein 1 Regulates Wnt signalling 41
18 E2R2G1_CANFA Angiopoietin-Like 7 Involved in Angiogenesis 45
Inhibits angiogenesis and nerve ingrowth in
19 E2R0R3_CANFA Semaphorin 46,47
IVD
20 F1Q3G2_CANFA Slit Homolog 3 (Drosophila) Involved in Angiogenesis 58
Involved in ECM assembly, invasion and
21 E2RMA3_CANFA Secreted Protein, Acidic, Cysteine-Rich (Osteonectin) 59
proliferation
Involved in ECM composition, invasion and
22 E2R161_CANFA (+1) Secreted Phosphoprotein 1 (Osteopontin) 59
migration

Table 1. Representative list of proteins identified in NCCM and their functions.

2, Brachyury and Oct4 as compared to injured disc NPs injected with PBS (1X, pH=7.2) used as sham controls
(Fig.7B). Western blotting validated the restoration of Brachyury and Oct4 expression in rat-tail injured disc NPs
following treatment with rh-CTGF and TGF1 proteins (Fig.7D, SupplementaryFig.9A). Further, treatment
with CTGF and TGF1 also suppressed levels of the ECM degrading enzyme, MMP-13 and inflammation medi-
ator, cyclooxygenase-2 (Cox2) in injured disc NPs (Fig.7C, SupplementaryFig.9A).

Discussion
Degenerative disc disease (DDD) is associated with the progressive loss of notochordal cells (NCs) and the devel-
opment of an inferior NP matrix compromising the structural integrity and biomechanical properties of the
spine1519. In this study, we characterized a pre-clinical rodent disc injury model that mimics progressive disc
degeneration, similar to that observed in humans. Our in vitro and pre-clinical rodent model emphasized the
role of inflammation in deterioration of the NP - ECM and cell death during progressive disc degeneration.
Further, we have identified the key factors secreted by NCD-canine notochordal cells responsible for maintain-
ing a healthy, hydrophilic, proteoglycan rich IVD NP that resists DDD. Our secretome analysis of the NC-rich
nucleus pulposus revealed components of TGF- and Wnt-signaling, anti-angiogeneic factors (semaphorins,
nidogen), several proteogylcans including aggrecan, small leucine rich proteoglycans (SLRPs) and collagens in
serum free NCCM. Both TGF1 and CTGF emerged as the major protein hubs regulating the expression of sev-
eral ECM proteins and interacting with other growth factors in protein network analysis. These protein networks
are important for maintenance of a healthy disc NP and might be involved in resisting the development of DDD.
Both TGF1 and CTGF play a critical role in the development of the IVD and cartilage in the embryonic
stage as well as in post-natal development of the spine24,25. TGF1 belongs to the transforming growth factor beta
(TGF) superfamily of proteins known to play an important role in cartilage and bone development26. CTGF
is an important constituent of the intervertebral disc microenvironment and interacts with several growth fac-
tors and matrix proteins including integrins and heparan sulfate proteoglycans2731. CTGF and WISP2 (CCN5)
belong to the connective tissue growth factor/cysteine-rich 61/nephroblastoma overexpressed (CCN) - family of
matricellular proteins that possess an amino-terminal secretory peptide followed by four conserved domains with
sequence homologies to insulin-like growth factor-binding proteins, von Willebrand factor C (VWC) domain,
thrombospondin type 1 repeat (TSR) and a carboxy-terminal (CT) domain that contains a cysteine-knot motif27.
Interestingly, the von Willebrand factor C (VWC) domain present in CTGF mediates physical interactions with

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 6

www.nature.com/scientificreports/

Figure 4. Ingenuity Pathway Analysis (IPA). IPA was performed using 94 ECM proteins identified in
our secretome analysis to determine the significant protein networks. Hierarchical model showing TGF
1 as a major hub protein, interacting with or regulating expression of several other proteins in the network.
Bold arrows show direct regulation, bold lines show direct interactions while dotted lines represent indirect
interactions among proteins shown in the network.

members of the TGF-superfamily, including bone morphogenetic proteins (BMPs) and other TGF ligands2731.
In contrast, CCN5 negatively regulates TGFsignaling by suppressing the expression of TGFRII. Knocking
down expression of CCN5 in breast cancer cells (MCF-7) resulted in deregulation of several components of the
TGF- signaling pathway including its downstream effector SMAD signaling proteins32.
Our data showed the loss of TGF1, CTGF and WISP2 expression in degenerative disc NPs implicating their
association with the loss of NCs. Notably, treatment with a combination of TGF1 and CTGF restored a healthy,
cellular NP rich in NC and stem cells as compared to sham control discs in our pre-clinical models. In addi-
tion, our results demonstrated that treatment with TGF1 and CTGF increased cell viability, DNA synthesis and
expression of healthy ECM genes in NP cells derived from degenerative human discs. The mitigation of inflam-
mation induced caspase-3/7 activity, Cox2, MMP-3 and MMP-13 expression levels by recombinant TGF1 and
CTGF demonstrated their synergistic action. CTGF has been shown to suppress IL-1- induced mRNA levels of
MMP-3, ADAMTS5, syndecan-4 and prolyl hydroxylase 3 by binding to cell surface integrin receptors v3 and
51 in healthy NP cells33. These finding clearly suggest the potential overlap in the cell signaling activated down-
stream of TGF1 and CTGF regulating their anti-inflammatory and anti-catabolic activities in NP cells. However,
our study is limited to studying the changes in the cellular and biochemical composition of NP in response to
inflammation in vitro and disc injury in vivo but lacks determination of changes in disc height, if any. Further
investigation needs to done in animal models to determine if such treatments have the capacity to retain or gain
increased disc height in addition to biochemical changes in response to these treatments.
Among other proteins, we identified CD109, COMP, CILP, decorin, chordin and sclerostin proteins that are
known activators or modulators of TGF-Smad dependent signaling. Soluble CD109 can bind to all 3 isoforms
of TGF-1/2/3, thus interfering with the activation of TGFdependent signaling34. In fact, membrane bound
CD109 acts as a co-receptor for TGF1 and results in internalization and degradation of TGFreceptors via
binding to Smad735,36. Further, CILP inhibits transcriptional activation of matrix genes in NP cells by binding
to TGF1 and inhibiting the downstream phosphorylation of Smads, thereby promoting disc degeneration37. In
addition, small leucine rich proteoglycans (SLRPs) like decorin sequester multiple growth factors including TGF
1 and directly antagonizes the function of EGFR and IGF-IR38. In contrast, COMP activates TGF- dependent
transcriptional activity39. Ishida et al.39, proposed that COMP may serve as an instructive matrix component in
vivo slowing the diffusion of the growth factors and promoting their biological activity. Similarly, sclerostin binds
to BMP-6 and BMP-7 while chordin binds to BMP-2 and BMP-4 preventing their interaction with BMP recep-
tors40. Taken together, these studies suggest the presence of a delicate balance in TGF- signaling in NC - rich NPs
that is well regulated and responsible for maintaining a healthy, proteoglycan rich NP.
In addition, sclerostin and soluble frizzled-related protein 1 (sFRP1) identified in NCCM are known antag-
onists of Wnt-signaling pathway. The secreted frizzled-related proteins (sFRP) bind to Wnt ligands, whereas the
sclerostin binds to components of the Wnt receptor41. We also identified LRP1, LRP2 and VLDLR, which act as
receptors for activation of Wnt-signaling. Hiyama et al.42, reported that the activation of Wnt signaling upreg-
ulates the pro-inflammatory cytokine TNF-leading to disc degeneration. Thus blocking Wnt signaling might
protect nucleus pulposus cells against degeneration43. Further, Iwata et al.44, demonstrated that Wnt/-catenin
signaling could enhance Runx2 expression in NPs resulting in IVD calcification. Thus, cross talk between TGF-
and Wnt-signaling pathways might play an important role in the development and progression of DDD. However,
these findings require further investigation in order to understand their impact on the development and progres-
sion of DDD.

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 7

www.nature.com/scientificreports/

Figure 5. Anabolic effects of CTGF, WISP2 and TGF1 in an in vitro model of DDD. Effect of CTGF, WISP2
and TGF1 treatment alone or in combination on cell viability (72 hrs) as determined using MTT assays in
NP cells obtained from (A) rat tail (healthy/injured) discs **p= 0.049, *p 0.02), (B) bovine degenerative disc
NPs, *p<0.01 and (C) human (H1, H2) degenerative disc NPs (*p<0.005). Each bar represents mean S.D.
of 3 independent experiments done in triplicates (n=9). Cell proliferation assays (72hrs) in (D) rat NP cells
(healthy/injured discs), *p< 0.001, **p <0.01 and (E) human (H1, H2) degenerative disc NPs treated with
CTGF and TGF1 alone or in combination as determined using colorimetric anti-BrdU ELISA, *p< 0.001.
The p-values were determined using paired Students t-test, for treatment with CTGF, WISP2 or TGF1 alone or
in combination with respect to no treatment control (NTC). (F) Histograms showing alterations in expression
of collagen 2, HAPLN1, versican and thrombospondin1 (Thbs1) on treatment of human degenerative disc NP
cells (H1, H2, H3) with CTGF and TGF1 as revealed by real time PCR analysis. Each bar in the histogram
represents the meanS.D. of 3 independent experiments done in duplicates (n= 6, *p < 0.001).

Another class of soluble factors identified in NCCM included proteins that prevent the ingrowth of blood
vessels and neurites into the NP, a hallmark of DDD. Angiopoeitin like 7 (ANGPTL7) has been proposed as
an anti-angiogeneic protein abundantly expressed in keratocytes and plays a major role in maintaining corneal
avascularity and transparency45. Semaphorin 3E identified in NCCM belongs to class 3 semaphorin family of pro-
teins, known as potent inhibitors of both pathological nerve innervation and vascular proliferation46,47. Another
member of the family, Semaphorin 3A is highly expressed by healthy disc cells, primarily localized to the outer
annulus fibrosus, but decreases significantly in the degenerative disc nucleus pulposus46,47. Further, low expres-
sion of semaphorin 3A showed a significant correlation with overexpression of endothelial cell marker, CD31
and PGP9.5, marker for neurons47. However, the role of ANGPTL7 and Semaphorin 3E in disc degeneration is
currently unknown and requires further investigations.

Conclusions
In conclusion, we have identified the soluble proteins present within NCCM that have the ability to maintain a
healthy, proteoglycan rich nucleus pulpous and delay DDD. We have further identified the proteins networks
important for homeostatic regulation of the healthy IVD - NP. Our findings suggest that the loss of TGF1 , CTGF
and WISP2 is associated with the progression of DDD. Both TGF1 and CTGF represent pivotal protein hubs
involved in cellular signaling and ECM regulation of the healthy IVD NP. We further demonstrated the potential
utility of a combination of TGF1 and CTGF as novel, minimally invasive molecular therapeutic agents for the
biological treatment of degenerative disc disease (Fig.8).

Materials and Methods

Generation of serum free notochordal cell derived conditioned medium (NCCM). Notochordal
cell derived conditioned medium (NCCM) was collected from notcohordal cell- rich nucleus pulpous (NP)

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 8

www.nature.com/scientificreports/

Figure 6. Anti-catabolic effects of CTGF, WISP2 and TGF1 in an in vitro model of DDD. Histograms
showing caspase 3/7 activity in NP cells derived from (A) rat injured IVD (*p= 0.008), (B,C) human
degenerative disc NP treated with pro-inflammatory cytokines, IL-1and TNFalone or in presence of
CTGF, WISP2 and TGF1 (*p <0.05). Histograms showing caspase 9 activity in NP cells (D) rat injured IVD,
(E,F) human degenerative disc NP cells treated with pro-inflammatory cytokines, IL-1and TNFalone or
in presence of CTGF, WISP2 and TGF1(*p < 0.05, **p <0.005). Each bar in the caspase assays is showing
meanS.D. of 2 independent experiments done in quadruplets (n=8). The p-values were determined using
paired Students t-test. For the combination of IL-1and TNFare with respect to no treatment control (NTC),
while p-values for the groups containing growth factors (CTGF, WISP2 or TGF1) are with respect to the group
containing combination of IL-1and TNFonly. Representative Western blot panels showing expression of
MMP-3, MMP-13 and Cox2 in rat healthy IVD NP cells treated with (G) IL-1alone, (H) IL-1and TNF
in combination, and in the presence of CTGF, WISP2 or TGF1. All Western blots were run under the same
experimental conditions as described in Materials and Methods. Histograms showing decreased expression
of (I) Cox2, (J) MMP-13 mRNA levels in human degenerative disc NP cells treated with IL-1and TNF in
presence of CTGF and TGF1 in comparison to IL-1and TNFonly treatments. Each bar in the histogram
represents meanS.D. of 3 independent experiments done in duplicates (n= 6, *p < 0.001).

obtained from IVDs of non-chondrodystrophic (NCD) canines as described earlier20,21. All animals (n= 12)
were obtained in collaboration with a licensed animal facility and all experimental protocols were approved
by the Animal Care Committee, Research Animal care policy and Ethics Approval Board of University Health
Networks, Toronto, Ontario, Canada. All experimental procedures were performed in accordance with the guide-
lines approved by Animal Care Committee and Ethics Approval Board of University Health Networks, Toronto,
Ontario, Canada. All non-chondrodystrophic canines were 8 to 14 months of age and had failed at adoption or
were to be euthanized for other purposes. Deep sedation was achieved using a combination of Acepromazine
(10mg/mL, Atravet-Aerst Pharmaceuticals St. Laurent, Quebec, Canada) mixed with Xylazine 100mg/mL
(Xylomax-Bimeda-NHC Animal Health, Broomhill Road, Tallaght, Dublin, Ireland) at a combined dose of
1mL/15Kg body weight. Once deep sedation had occurred, euthanasia was accomplished using intravenous
sodium pentobarbital (CDMV) (St. Hyacinthe, Quebec, Canada at a dose of 30mL/kg body weight. Within 2hrs
of euthanization, the lumbar spines were removed and nuclei pulposi were isolated under aseptic conditions20,21.
Nuclei pulposi were washed with phosphate buffered saline (PBS, pH=7.2) and three NPs were placed within
tissue culture inserts with 40m-filters in serum free CD Hybridoma media (protein and phenol red free, Cat No
#11279-023, Life Technologies, USA) containing 100 units penicillin/streptomycin in 6 well plates under hypoxic
conditions (3.5% O2, and 5% CO2, NuAire incubators) at 37C. The conditioned medium referred hereafter as
NCCM was collected after 24hrs 48hrs, centrifuged at 8000rpm for 30minutes, filtered through 0.2m syringe-
tip filters and stored in 80C until further use.

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 9

www.nature.com/scientificreports/

Figure 7. Evaluation of the therapeutic potential of CTGF and TGF1 in a pre-clinical in vivo rodent disc
injury model of DDD. (A) Representative Safranin-O staining photomicrographs representing uninjured,
healthy IVD and injured rat IVDs injected with phosphate - buffered saline (PBS, used as a control), CTGF,
TGF1 or a combination of CTGF and TGF1 (Scale bar 500, n =12 tissue sections/group). (B) Representative
Safranin-O and immunohistochemical staining of ECM proteins, aggrecan and collagen 2 in paraffin embedded
sections of rat tail injured IVD - NPs treated with phosphate - buffered saline (PBS, used as a control), CTGF,
TGF1 or a combination of CTGF and TGF1 (Scale bar 50). Immunohistochemistry for all the proteins was
performed in duplicates in atleast 3 different IVD sections obtained from different animals from each group. (C)
Scatter plots showing histological grading scores (meanS.D.) of rat healthy tail disc-NPs and injured tail disc-
NPs following treatment with PBS, CTGF, TGF1 and a combination of CTGF and TGF1 (**p < 0.001). The
p-values were determined for treatment with CTGF, TGF1 alone or in combination with respect to PBS. (D)
Representative Western blot panels showing decreased expression of MMP-13 and Cox2, and restoration of the
NC marker, brachyury and stem cell marker, Oct4 in NP tissue lysates (pooled) obtained from rat tail injured
discs treated with CTGF, TGF1 or a combination of CTGF and TGF1.

Figure 8. Proposed model demonstrating the mechanism of progressive disc degeneration in presence of
pro-inflammatory cytokines (IL-1 and TNF) and effect of intervention by novel, potential therapeutic
agents (CTGF/TGF1) for regeneration of the IVD - NP.

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 10

www.nature.com/scientificreports/

Identification of bioactive fractions of NCCM. Notochordal cell condition medium (NCCM) was
concentrated sequentially using 50kDa and 3kDa spin-ultrafiltration protein concentrators (EMD Millipore,
MA, USA) following manufacturers instructions. The respective concentrated NCCM samples were further
fractionated by size - exclusion on a Superose 12 HR 10/30 fast protein liquid chromatography (FPLC) column
(Pharmacia) in running buffer (10mM sodium phosphate, 150mM NaCl, 1mM EDTA, pH=7.4). Thirty frac-
tions (~1ml) were collected, measured for protein concentration by absorbance at 280nm and stored at 80 C
until further use. Consecutive protein-containing fractions were pooled pairwise and evaluated for their effect
on etoposide (cytotoxic drug) induced caspase-3/7 activity in bovine tail disc NP cells as described below.
Bioactive fractions were defined as protein fractions showing a decrease in caspase-3/7 activity in bovine NP cells
on treatment with etoposide. These bioactive fractions were later analyzed for identification of proteins using
mass-spectrometry.

Mass spectrometry analysis. The LC-MS/MS analysis was performed at the SPARC BioCentre, The
Hospital for Sick Children, Toronto, Canada. Briefly, the bioactive fractions were reduced with dithiothreitol
(DTT), the free cysteine residues alkylated with iodoactetamide and digested overnight with modified bovine
trypsin (Promega, Madison, USA). The tryptic peptides were desalted and loaded onto a 50cm 75 m ID col-
umn containing RSLC 2m C18 packing material (EASY-Spray, Thermo-Fisher, Odense, Denmark) with an
integrated emitter. The peptides were eluted into a Q-Exactive hybrid mass spectrometer (Thermo-Fisher, San
Jose, CA) using an Easy-Spray nLC 1000 chromatography system (Thermo-Fisher, Odense Denmark) with a
90-minute gradient from 0% to 35% acetonitrile in 0.1% formic acid. The mass spectrometer was operated in
a data dependent mode with 1 MS followed by 10 MS/MS spectra. The MS was acquired with a resolution of
70,000 FWHM, a target of 1 106 ions and a maximum scan time of 120ms. The MS/MS scans were acquired
with a resolution of 17,500 FWHM, a target of 1 106 ions and a maximum scan time of 120ms using relative
collision energy of 27%. A dynamic exclusion time of 15seconds was used for the MS/MS scans. The raw data
files were acquired with XCalibur 2.2 (Thermo-Fisher Scientific) and processed with the Sequest search engine
(Thermo-Fisher Scientific) using the UniProt canine database August 12, 2014 version with 28,460 entries and
with X!-Tandem (Beavis Informatics, Winnipeg, Canada). The processed data was imported into Scaffold 3.2
(Proteome Software, Portland, OR). Peptides were considered to be identified if the Scaffold score exceeded the
0.1% false discovery rate (FDR) as determined by searching against the reversed UniProt canine database.

Development and characterization of a pre-clinical rodent injury model of DDD. We used

12-week old female Wistar rats (n=21, Charles River Laboratories Inc., MA, USA) to develop a pre-clinical
rodent tail disc injury model of DDD as follows. Experimental protocols were approved by the Animal Care
Committee and Research Ethics Approval Board of University Health Networks, Toronto, Ontario, Canada. All
experimental procedures were performed in accordance with the guidelines approved by Animal Care Committee
and Ethics Approval Board of University Health Networks, Toronto, Ontario, Canada. Anesthesia was achieved
using isofluorane (5L/min plus 1L/min O2) and maintained at 3L/min. Once deeply anaesthetized, the animal
was affixed on a stereotactic procedure apparatus (Model 900, Kopf Instruments, CA, USA) with nose cone inha-
lation. For disc injury (4 caudal discs per animal), we used a 26-gauge (G) needle (Hamilton Company, USA)
mounted on a Hamilton syringe. The needle was advanced completely through the selected tail IVD to penetrate
the full thickness inclusive of the annulus fibrosus on both sides of the disc. Confirmation of needle placement
was made using fluoroscopy and maintained in position for 2minutes, withdrawn halfway to the center of the NP
and left there for 1minute and then slowly withdrawn completely over a 1-minute period. The animals were then
removed from the stereotactic apparatus and allowed to recover in a warmed cage. At the end of study period i.e.
72hrs - 10 weeks, animals were humanely euthanized using CO2 asphyxiation and each vertebral lumbar/caudal
motion segment was dissected aseptically. Of the four injured IVDs, atleast one IVD per animal was fixed in
formalin as a representative for histological analysis and rest of the injured IVD - nucleus pulposi were harvested
and lysed in RIPA buffer (50mM Tris, pH=7.4, 150mM NaCl, 1% NP-40 and protease inhibitor cocktail) for
western blotting.

Evaluation of NCCM and soluble factors as therapeutic agents in pre-clinical rodent model of
DDD. We used a pre-clinical rat-tail disc injury model of DDD for evaluating the therapeutic potential of
NCCM and soluble factors identified. For the 1st set of experiments involving NCCM, animals (n= 6/group) were
randomized four weeks post-injury and an intra-discal injection (~8L) of either concentrated NCCM (2.2g/l)
or the control medium was given under general anesthesia using fluoroscopic image guidance in 4 injured discs
per animal as described earlier. For the 2nd set of experiments involving growth factors including CTGF and
TGF1, animals (n=6/group) were randomized four weeks post-injury and an intra-discal injection (~8L)
of either phosphate buffered saline (PBS, 1X, pH=7.2), CTGF (100ng/mL), TGF1 (10ng/mL) or a combina-
tion of CTGF (100ng/mL) and TGF1 (10ng/mL) was given under general anesthesia using fluoroscopic image
guidance in 4 injured discs per animal as described above. Further, a group of six animals were left uninjured that
served as age matched healthy controls in each set of experiments. At the end of experiments (i.e. six weeks later),
at least one IVD per animal was fixed in formalin as a representative for histological analysis. The remainder of
the NPs harvested from rat caudal IVDs were pooled together according to their respective groups (healthy/
treated) for lysis in RIPA lysis buffer for western blotting. Each of these experiments was repeated independently
to ensure reproducibility.

Histology and immunohistochemistry (IHC). For formalin fixation, at least one IVD per animal
(healthy/injured following treatment with NCCM, PBS, CTGF or TGF1 alone or in combination) was harvested

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 11

www.nature.com/scientificreports/

for paraffin embedding. The tissues were fixed in 10% formalin followed by decalcification and paraffin embed-
ding. Hemaetoxylin and eosin (H&E) and Safranin-O staining was performed in duplicates on paraffin-embedded
sections (5m) to assess general morphology and proteoglycan content21. Histological grading of IVD-NP (injury
followed by treatment) was carried out on the basis of NP-cellularity in H&E sections as described earlier48.
Following H&E and Safranin O staining, serial tissue sections were deparaffinized in xylene, hydrated in gradient
alcohol followed by antigen retrieval in Tris-EDTA buffer (pH=9.0) for IHC. The sections were incubated with
hydrogen peroxide (0.3% v/v) in methanol for 15minutes, followed by blocking with 10% serum. Thereafter, the
slides were incubated with either rabbit/goat polyclonal or mouse monoclonal primary antibodies overnight
(O/N) at 4C. Protein expression was detected using respective secondary antibodies (rabbit/goat/mouse) from
Vectastain ABC kit and diaminobenzidine (DAB) as a chromogen. In negative controls, the primary antibody was
replaced by isotype-matched IgG. Immunohistochemistry was done in at least 3 different IVD sections per group.
The bright field sections were evaluated by light microscopic examination using a ScanScope XT, Aperio Whole
Slide Scanner and Nikon bright field microscope.

Ingenuity Pathway Analysis (IPA). Ingenuity Pathway Analysis (IPA, Ingenuity Systems, www.ingenuity.
com) was used to identify the canonical pathways, major protein hubs and the associated networks involved
in the maintenance of a healthy disc NP48. Of the 303 proteins identified in the bio-active fractions of NCCM,
only 94 ECM proteins were used for network analysis. The protein accession numbers were used to determine
IPA abbreviated names. These notations were then used to navigate the databases available online. The eligible
proteins served as nodes for generating the canonical and novel proteins interaction networks. All networks were
scored [log (p-value)] based on the number of network eligible molecules contained in these networks. The
higher the score, the lower was the probability of finding the observed number of network eligible molecules by
random chance. The canonical pathways were scored based on their p-values, calculated using the right-tailed
Fisher Exact Test; a p-value0.05 indicates a statistically significant, non-random association.

Cell culture, Western blotting and quantitative real time PCR (qPCR). Healthy rat lumbar/caudal
spine IVDs, injured tail discs and bovine caudal IVD NPs were removed aseptically. Bovine caudal disc NPs
were obtained from six, 3-year old steers. Human degenerative disc nucleus pulposus tissues were obtained
from patients (n=3) undergoing discectomy or fusion surgery at Toronto Western Hospital, University Health
Network (UHN), Toronto. For human degenerative disc NP samples used in this study, informed consent was
obtained from all the patients prior to surgery. All human tissue samples were collected in accordance to the
guidelines approved by the Research Ethics Board, Toronto Western Hospital, UHN, Toronto. The nucleus pul-
posus (NP) was separated and enzymatically digested according to our established methods20,21. The next day,
the cells were filtered with a 70m cell strainer and cultured within a hypoxic incubator (NuAire, MN, USA)
in 3.5% O2, 5% CO2, in Advanced Dulbeccos modified Eagles medium (ADMEM) supplemented with 8% fetal
bovine serum (FBS), penicillin and streptomycin (100U/mL) until passage (P2) as described earlier20,21. The
NP cells obtained from rat IVDs were pooled together for treatment with respective agents as follows. The cells
were either cultured in serum free DMEM (no treatment controls) or treated with IL-1(10ng/mL), TNF
(50ng/mL), CTGF (10100ng/mL), WISP2 (10100ng/mL) or TGF1 (520ng/mL) for various time points
(24hrs72hrs) under hypoxic conditions. The effect of pro-inflammatory cytokines (IL-1and TNF) and solu-
ble factors (CTGF/WISP2/TGF1) on expression of ECM genes and proteins in NP cells were determined using
qPCR and Western blotting respectively (See Supplementary Data for details). For Western blotting, protein esti-
mation was performed using Bradford assay (BioRad, CA). Briefly, equal amounts of whole cell or tissue lysates
prepared using RIPA lysis buffer were subjected to Western blotting as described earlier49. Total lysates (30g)
were resolved on 10% sodium dodecyl sulphate-polyacrylamide gels (SDS-PAGE) under reducing conditions and
then proteins were electro-transferred onto polyvinyledendifluoride (PVDF) membranes (BioRad, CA). After
blocking with 5% non-fat powdered milk in Tris-buffered saline (TBS, 0.1M, pH=7.4), blots were incubated with
rabbit/goat polyclonal or mouse monoclonal primary antibodies at 4C overnight (See Supplementary Data for
antibody details). Membranes were washed three times with Tween (0.1%)-Tris-buffer saline (TTBS) and then
incubated for 2hrs at room temperature (RT) with the respective HRP-conjugated anti-IgG secondary antibodies
(BioRad, CA), diluted as per the manufacturers suggestions in 2% non-fat milk in TBS (pH=7.2, 1X). Blots were
washed three times with TTBS for 15minutes and protein bands were detected by the enhanced chemilumines-
cence method (BioRad, CA) on Kodak Hyperfilm50. -actin was used as a loading control for each experiment.
All Western blots were run under the same experimental conditions and repeated at least 3 times to ensure
reproducibility. All Western blots were quantified using densitometry values calculated using Image J software
(available online). For calculation of fold changes representing expression levels of protein under each treatment
conditions, densitometry values of each band for the protein of interest was normalized with -actin used as a
loading control in each experiment50.

Cell viability, cell proliferation and apoptosis assays. Cell viability was determined using 3-(4,5-dim
ethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma-Aldrich, USA) as described earlier50. We used
a BrdU - ELISA (colorimetric) assay (Cat# ab126556, Abcam) to determine the effect of treatment with CTGF
(100ng/mL) or TGF1 (10ng/mL) on human and rat NP cells (healthy/injured) following the manufacturers
instructions. Apoptosis in NP cells (rat and human) was determined using Caspase-3/7 and Caspase-9 specific
Lumi-Glo assays (Promega, Madison). See Supplementary Data for details.

Statistical analysis. All data for cell viability, proliferation, apoptosis (casase-3/7, caspase-9) assays
and qPCR are expressed as meansSD. Significant differences in test and no treatment controls (NTC) were

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 12

www.nature.com/scientificreports/

determined using the paired Students t-test. Statistical analysis was performed using the Graphpad Prism. Data
analysis for qPCR was carried out using calculation of Ct values, fold changes and p-values using Students
t-test. P-value<0.05 was defined as statistically significant for all tests. Ingenuity Pathway Analysis (IPA) software
calculated p-values for canonical pathways based on the Fisher Exact Test; a p-value<0.05 indicates a statistically
significant, non-random association.

References
1. Global Burden of Disease Study 2013 Collaborators. Global, regional, and national incidence, prevalence, and years lived with
disability for 301 acute and chronic diseases and injuries in 188 countries, 19902013: a systematic analysis for the Global Burden of
Disease Study 2013. Lancet 386, 743800 (2015).
2. Waterman, B. R., Belmont, P. J. Jr. & Schoenfeld, A. J. Low back pain in the United States: incidence and risk factors for presentation
in the emergency setting. Spine J 12, 6370 (2012).
3. Phillips, F. M., Slosar, P. J., Youssef, J. A., Andersson, G. & Papatheofanis, F. Lumbar spine fusion for chronic low back pain due to
degenerative disc disease: a systematic review. Spine (Phila Pa 1976) 38, E40922 (2013).
4. Bydon, M. et al. Lumbar fusion versus non-operative management for treatment of discogenic low back pain: a systematic review
and meta-analysis of randomized controlled trials. J. Spinal Disord Tech. 27, 297304 (2014).
5. Nishida, K. et al. Gene therapy approach for disc degeneration and associated spinal disorders. Eur. Spine J 17, 45966 (2008).
6. Masuda, K. Biological repair of the degenerated intervertebral disc by the injection of growth factors. Eur. Spine J 17, 44151 (2008).
7. Wang, Z. et al. Efficacy of intervertebral disc regeneration with stem cells - a systematic review and meta-analysis of animal
controlled trials. Gene. 564, 18 (2015).
8. Sakai, D. & Andersson, G. B. Stem cell therapy for intervertebral disc regeneration: obstacles and solutions. Nat. Rev. Rheumatol. 11,
24356 (2015).
9. Kandel, R., Roberts, S. & Urban, J. P. Tissue engineering and the intervertebral disc: the challenges. Eur. Spine J. 17, 48091 (2008).
10. Walsh, A. J., Bradford, D. S. & Lotz, J. C. In vivo growth factor treatment of degenerated intervertebral discs. Spine (Phila Pa 1976).
29, 15663 (2004).
11. Malham, G. M. et al. Anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2: a prospective
study of complications. J. Neurosurg. Spine 21, 85160 (2014).
12. Willems, N. et al. Intradiscal application of rhBMP-7 does not induce regeneration in a canine model of spontaneous intervertebral
disc degeneration. Arthritis Res. Ther. 17, 137 (2015).
13. Goupille, P., Mulleman, D. & Chevalier, X. Is interleukin-1 a good target for therapeutic intervention in intervertebral disc
degeneration: lessons from the osteoarthritic experience. Arthritis. Res. Ther. 9, 110 (2007).
14. Sinclair, S. M. et al. Attenuation of inflammatory events in human intervertebral disc cells with a tumor necrosis factor antagonist.
Spine (Phila Pa 1976) 36, 11906 (2011).
15. Risbud, M. V. & Shapiro, I. M. Notochordal cells in the adult intervertebral disc: new perspective on an old question. Crit. Rev.
Eukaryot.Gene Expr. 21, 2941 (2011).
16. Butler, D., et al. Discs degenerate before facets. Spine. 15, 1113, (1990).
17. Aguiar, D. J., Johnson, S. L. & Oegema, T. R. Notochordal cells interact with nucleus pulposus cells: regulation of proteoglycan
synthesis. Experimental Cell. Research 246, 129137 (1999).
18. Risbud, M. V. & Shapiro, I. M. Role of cytokines in intervertebral disc degeneration: pain and disc content. Nat. Rev. Rheumatol. 10,
4456 (2014).
19. Smolders, L. A. et al. Intervertebral disc degeneration in the dog. Part 2: chondrodystrophic and non-chondrodystrophic breeds. Vet.
J 195, 2929 (2013).
20. Erwin, W. M., Islam, D., Inman, R. D., Fehlings, M. G. & Tsui, F. W. Notochordal cells protect nucleus pulposus cells from
degradation and apoptosis: implications for the mechanisms of intervertebral disc degeneration. Arthritis Res. Ther. 13, R215 (2011).
21. Erwin, W. M., Ashman, K., ODonnel, P. & Inman, R. D. Nucleus pulposus notochord cells secrete connective tissue growth factor
and up-regulate proteoglycan expression by intervertebral disc chondrocytes. Arthritis Rheum. 54, 385967 (2006).
22. Bach, F. C. et al. The species-specific regenerative effects of notochordal cell-conditioned medium on chondrocyte-like cells derived
from degenerated human intervertebral discs. Eur. Cell Mater. 30, 13247 (2015).
23. Cornejo, M. C., Cho, S. K., Giannarelli, C., Iatridis, J. C. & Purmessur, D. Soluble factors from the notochordal-rich intervertebral
disc inhibit endothelial cell invasion and vessel formation in the presence and absence of pro-inflammatory cytokines. Osteoarthritis
Cartilage 23, 48796 (2015).
24. Jin, H. et al. TGF-signaling plays an essential role in the growth and maintenance of intervertebral disc tissue. FEBS Lett. 585,
120915 (2011).
25. Bedore, J. et al. Impaired intervertebral disc development and premature disc degeneration in mice with notochord-specific deletion
of CCN2. Arthritis Rheum. 65, 263444 (2013).
26. Li, X. & Cao, X. BMP signaling and skeletogenesis. Ann. N. Y. Acad. Sci. 1068, 2640 (2006).
27. Leask, A. & Abraham, D. J. All in the CCN family: essential matricellular signaling modulators merge from the bunker. J. Cell Sci.
119, 480310 (2006).
28. Erwin, W. M. The Notochord, Notochordal cell and CTGF/CCN-2: ongoing activity from development through maturation. J. Cell
Commun. Signal 2, 5965 (2008).
29. Nishida, T. et al. CCN2 (Connective Tissue Growth Factor) is essential for extracellular matrix production and integrin signaling in
chondrocytes. J. Cell Commun. Signal 1, 4558 (2007).
30. Abreu, J. G., Ketpura, N. I., Reversade, B. & De Robertis, E. M. Connective tissue growth factor (CTGF) modulates cell signalling by
BMP and TGFbeta. Nat. Cell Biol. 4, 599604 (2002).
31. Tran, C. M. et al. Regulation of CCN2/connective tissue growth factor expression in the nucleus pulposus of the intervertebral disc:
role of Smad and activator protein 1 signaling. Arthritis. Rheum. 62, 198392 (2010).
32. Sabbah, M. et al. CCN5, a novel transcriptional repressor of the transforming growth factor signaling pathway. Mol. Cell Biol. 31,
145969 (2011).
33. Tran, C. M. et al. CCN2 suppresses catabolic effects of interleukin-1 through 51 and V3 integrins in nucleus pulposus cells:
implications in intervertebral disc degeneration. J. Biol. Chem. 289, 737487 (2014).
34. Li, C. D. et al. Soluble CD109 binds TGF-and antagonizes TGF-signaling and responses. Biochem. J pii: BJ20141488 (2015).
35. Bizet, A. A. et al. The TGF-co-receptor, CD109, promotes internalization and degradation of TGF- receptors. Biochim. Biophys.
Acta. 1813, 74253 (2011).
36. Bizet, A. A., Tran-Khanh, N., Saksena, A., Liu, K., Buschmann, M. D. & Philip, A. CD109-mediated degradation of TGF- receptors
and inhibition of TGF-responses involve regulation of SMAD7 and Smurf2 localization and function. J. Cell Biochem. 113, 23846
(2012).
37. Seki, S. et al. Cartilage intermediate layer protein promotes lumbar disc degeneration. Biochem. Biophys. Res. Commun. 446, 87681
(2014).
38. Neill, T., Schaefer, L. & Iozzo, R. V. Decorin: a guardian from the matrix. Am. J. Pathol. 181, 3807 (2012).

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 13

www.nature.com/scientificreports/

39. Ishida, K. et al. Cartilage oligomeric matrix protein enhances osteogenesis by directly binding and activating bone morphogenetic
protein-2. Bone. 55, 2335 (2013).
40. Yanagita, M. BMP antagonists: their roles in development and involvement in pathophysiology. Cytokine Growth Factor Rev. 16,
30917 (2005).
41. Mii, Y. & Taira, M. Secreted Wnt inhibitors are not just inhibitors: regulation of extracellular Wnt by secreted Frizzled-related
proteins. Dev. Growth Differ. 53, 91123 (2011).
42. Hiyama, A., Yokoyama, K., Nukaga, T., Sakai, D. & Mochida, J. A complex interaction between Wnt signaling and TNF-in nucleus
pulposus cells. Arthritis Res. Ther. 15, R189 (2013).
43. Hiyama, A. et al. Enhancement of intervertebral disc cell senescence by WNT/-catenin signaling-induced matrix metalloproteinase
expression. Arthritis Rheum. 62, 303647 (2010).
44. Iwata, M. et al. Enhancement of Runx2 expression is potentially linked to -catenin accumulation in canine intervertebral disc
degeneration. J. Cell Physiol. 230, 18090 (2015).
45. Toyono, T. et al. Angiopoietin-like 7 is an anti-angiogenic protein required to prevent vascularization of the cornea. PLoS One. 10,
e0116838 (2015).
46. Yin, P., Lv, H., Zhang, L., Zhang, L. & Tang, P. Semaphorin 3A: A Potential Target for Low Back Pain. Front. Aging Neurosci. 26, 7:216
(2015).
47. Binch, A. L. et al. Class 3 semaphorins expression and association with innervation and angiogenesis within the degenerate human
intervertebral disc. Oncotarget. 6, 1833854 (2015).
48. Han, B. et al. A simple disc degeneration model induced by percutaneous needle puncture in the rat tail. Spine (Phila Pa 1976) 33,
192534 (2008).
49. Matta, A., Masui, O., Siu, K. W. & Ralhan R. Identification of 14-3-3zeta associated protein networks in oral cancer. Proteomics. 16,
107989 (2016).
50. Matta, A., DeSouza, L. V., Ralhan, R. & Siu, K. W. Small interfering RNA targeting 14-3-3increases efficacy of chemotherapeutic
agents in head and neck cancer cells. Mol. Cancer Ther. 9, 267688 (2010).
51. Chaly, Y. et al. Follistatin-like protein 1 regulates chondrocyte proliferation and chondrogenic differentiation of mesenchymal stem
cells. Ann. Rheum. Dis. 74, 146773 (2015).
52. Ohyama, Y. et al. Modulation of matrix mineralization by Vwc2-like protein and its novel splicing isoforms. Biochem. Biophys. Res.
Commun. 418, 126 (2012).
53. Camaj, P. et al. EFEMP1 binds the EGF receptor and activates MAPK and Akt pathways in pancreatic carcinoma cells. Biol. Chem.
390, 1293302 (2009).
54. Tanaka, K. et al. Role of osteoglycin in the linkage between muscle and bone. J. Biol. Chem. 287, 1161628 (2012).
55. Neidlinger-Wilke, C. et al. Mechanical stimulation alters pleiotrophin and aggrecan expression by human intervertebral disc cells
and influences their capacity to stimulate endothelial migration. Spine (Phila Pa 1976). 34, 6639 (2009).
56. Khling, R. et al. Nidogen and nidogen-associated basement membrane proteins and neuronal plasticity. Neurodegener. Dis. 3,
5661 (2006).
57. Miljkovic-Licina, M. et al. Targeting olfactomedin-like 3 inhibits tumor growth by impairing angiogenesis and pericyte coverage.
Mol. Cancer Ther. 11, 258899 (2012).
58. Wang, L. J. et al. Targeting Slit-Roundabout signaling inhibits tumor angiogenesis in chemical-induced squamous cell
carcinogenesis. Cancer Sci. 99, 5107 (2008).
59. Grogan, S. P. et al. Influence of cartilage extracellular matrix molecules on cell phenotype and neocartilage formation. Tissue Eng.
Part A. 20, 26474 (2014).

Acknowledgements
We are thankful to Drs. Steven Lewis, Mohammad Shamji and Michael Fehlings for providing the human surgical
disc samples used in this study and our summer student, Stanley Zhou, BSc for his technical assistance in some
of the preliminary experiments. We also gratefully acknowledge Dr. Robert Inman for manuscript review and
editorial suggestions. We are grateful to The Skoll Family Trust for financial assistance with the production of
this work.

Author Contributions
A.M. and W.M.E. designed experiments, wrote and edited manuscript. A.M., M.Z.K. and W.M.E. performed
experiments and analyzed data. D.E.I. provided reagents and infrastructure for size exclusion chromatography
and edited the manuscript. W.M.E. provided all the reagents, financial support and infrastructure for performing
the experiments.

Additional Information
Supplementary information accompanies this paper at http://www.nature.com/srep
Competing Interests: The authors declare no competing financial interests.
How to cite this article: Matta, A. et al. Molecular Therapy for Degenerative Disc Disease: Clues from
Secretome Analysis of the Notochordal Cell-Rich Nucleus Pulposus. Sci. Rep. 7, 45623; doi: 10.1038/srep45623
(2017).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
This work is licensed under a Creative Commons Attribution 4.0 International License. The images
or other third party material in this article are included in the articles Creative Commons license,
unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license,
users will need to obtain permission from the license holder to reproduce the material. To view a copy of this
license, visit http://creativecommons.org/licenses/by/4.0/

The Author(s) 2017

Scientific Reports | 7:45623 | DOI: 10.1038/srep45623 14