Glucose-dependent insulinotropic polypeptide (GIP) directly affects collagen fibril diameter and collagen cross-linking in osteoblast cultures.

(1)

Original Full Length Article

Glucose-dependent insulinotropic polypeptide (GIP) directly affects collagen ﬁ bril diameter and collagen cross-linking in osteoblast cultures

Aleksandra Mieczkowska

^a

, Beatrice Bouvard

^b

, Daniel Chappard

^a,c

, Guillaume Mabilleau

^a,c,

⁎

aGEROM Groupe Etudes Remodelage Osseux et bioMatériaux—LHEA, IRIS-IBS Institut de Biologie en Santé, CHU d'Angers, LUNAM Université, 49933 Angers Cedex, France

bService de Rhumatologie, CHU d'Angers, 49933 Angers Cedex, France

cSCIAM, Service Commun d'Imagerie et Analyses Microscopiques, IRIS-IBS Institut de Biologie en Santé, CHU d'Angers, LUNAM Université, 49933 Angers Cedex, France

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

Article history:

Received 4 October 2014 Revised 19 December 2014 Accepted 5 January 2015 Available online 10 January 2015 Edited by Ms. J. Aubin Keywords:

GIP Osteoblast Collagen

Glucose-dependent insulinotropic polypeptide (GIP) is absolutely crucial in order to obtain optimal bone strength and collagen quality. However, as the GIPR is expressed in several tissues other than bone, it is difﬁcult to ascertain whether the observed modiﬁcations of collagen maturity, reported in animal studies, were due to direct effects on osteoblasts or indirect through regulation of signals originating from other tissues. The aims of the present study were to investigate whether GIP can directly affect collagen biosynthesis and processing in osteoblast cultures and to decipher which molecular pathways were necessary for such effects.

MC3T3-E1 cells were cultured in the presence of GIP ranged between 10 and 100 pM. Collagenﬁbril diameter was investigated by electron microscopy whilst collagen maturity was determined by Fourier transform infrared microspectroscopy (FTIRM).

GIP treatment resulted in dose-dependent increases in lysyl oxidase activity and collagen maturity. Furthermore, GIP treatment shifted the collagenﬁber diameter towards lower value but did not signiﬁcantly affect collagen heterogeneity. GIP acted directly on osteoblasts by activating the adenylyl cyclase–cAMP pathway.

This study provides evidences that GIP acts directly on osteoblasts and is capable of improving collagen maturity andﬁbril diameter.

Introduction

Glucose-dependent insulinotropic polypeptide (GIP) is an important gastro-intestinal hormone synthesized and secreted into the blood stream by intestinal endocrine K cells after ingestion of a mixed meal [1–3]. In humans, fasting plasma GIP is ranged between 12 and 92 pM and increases to 35–235 pM postprandially[4]. To induce a biological response, GIP binds to a speciﬁc glucose-dependent insulinotropic polypeptide receptor (GIPR) expressed in several tissues including bone tissue[5–9]. However, despite expression of its receptor at the sur- face of osteoblasts, little is known about the osseous effects of the GIP/

GIPR pathway. Previously, we reported that the deletion of theGipr gene in mouse resulted in dramatic alterations of trabecular bone microarchitecture[10]. We also evidenced that bone strength and material properties were markedly reduced in this animal model with a dramatic 16% decrease in collagen maturity[11]. Furthermore, treatment of healthy Copenhagen rats with a stable GIP agonist resulted in a 13% increase in collagen maturity[12]. Taken together, these results suggest that the GIP/GIPR pathway is involved in the control of collagen quality in the bone matrix. However, as the GIPR is expressed in several

tissues other than bone, it is difﬁcult to ascertain whether the observed modiﬁcations in collagen quality are due to direct effects on osteoblasts or indirect through regulation of signals originating from other tissues.

Bone tissue is considered to be a two-component material made primarily of mineral (poorly crystalline hydroxyapatite) and organic (collagen type I) matrices[13]. The mineral component confers stiffness to the bone tissue[14,15]whilst collagen confers a degree of tensile strength, ductility and toughness[16,17]. Collagen type I biosynthesis by osteoblasts is a complex process and includes intracellular post- translational modifications, assembly of procollagen chains and secretion [18]. In the extracellular space, C-terminal and N-terminal propeptides are cleaved resulting in collagenfibril formation. The ultimate processing of collagenfibrils is represented by enzymatic cross-linking. Cross-linking stabilizes the collagen network and is due to oxidative deamination ofε-amino groups of specific lysine and hydroxylysine residues by lysyl oxidase, forming aldehyde moieties [19]. These aldehydes are highly reactive and lead to formation of cova- lent crosslinks that are critically required for the formation of functional mature collagen. The assembly of collagen molecules intofibrils has been described as entropy driven, similar to that occurring in other self-assembling proteins such as microtubules or actinfilaments[20].

However, data obtained from genetically-modiﬁed animals revealed that several other molecules, present in the extracellular bone matrix,

⁎ Corresponding author. Fax: +33 244 688 451.

E-mail address:[email protected](G. Mabilleau).

Contents lists available atScienceDirect

Bone

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / b o n e

(2)

may regulate collagen deposition andfibril formation. Among them, glycosaminoglycan-containing molecules such as biglycan and decorin are suspected to play a major role in collagen deposition andfibril formation as demonstrated by abnormal collagenfibrils in deficient animals[21,22]. Furthermore, evidences have been provided that lysyl oxidase activity is also a pre-requisite for normal collagen deposition andfibril organization[23].

The aims of the present study were to investigate whether GIP directly affects collagen deposition and maturation in osteoblast cultures and to decipher what molecular pathways were necessary for such effects.

Material and methods Reagents

Alpha-Minimum Essential Medium Eagle (α-MEM), fetal bovine serum (FBS), bovine calf serum, penicillin and streptomycin were purchased from Lonza (Levallois-Perret, France).β-Aminopropionitrile (BAPN) was purchased from VWR (Fontenay-sous-bois, France). All other chemicals were purchased from Sigma-Aldrich (Lyon, France) when otherwise stated.

Peptide synthesis

D-ala²-glucose-dependent insulinotropic polypeptide (GIP) was synthesized in our peptide supplier Genecust (Dudelange, Luxembourg).

Brieﬂy, GIP was sequentially synthesized using standard solid-phase Fmoc protocols[24]using Wang resin. Substitution of alanine at the N-terminal position 2 with its (D) isomer was performed to ensure bet- ter resistance to serum dipeptidylpeptidase-4 as described elsewhere [25]. The peptide was then puriﬁed by reverse-phase HPLC (purity N95%) using a SinoChrom ODS-BP column (4.6 × 250 mm) and subsequently characterized by electrospray ionization mass spectroscopy (ESI-MS) using a LCMS-2010 device (Shimadzu Corporation, Duisburg, Germany). Mass spectra were collected using full ion scan mode over the mass-to-charge (m/z) range of 500–1800. Observed molecular mass for GIP was 4982.6 Da corresponding to theoretical value of 4983.6 Da[26]. GIP was then dissolved in phosphate buffer saline at a concentration of 10⁻⁵M, aliquoted and stored at−80 °C until use.

Cell culture

MC3T3-E1 cells are phenotypically normal osteoblasts derived from fetal mouse calvaria and differentiate in culture resulting ultimately to the formation of a bone-like extracellular matrix[27]. MC3T3-E1 cells were purchased from American type culture collection (ATCC, Teddington, UK). Cells were grown and expanded at a ratio 1:4 in propagation medium containingαMEM supplemented with 5% FBS, 5% bovine calf serum, 100 U/mL penicillin, and 100μg/mL streptomycin in a humidified atmosphere enriched with 5% CO2at 37 °C. For differentiation studies, cells were detached with trypsin-EDTA and plated at a density of 1.5 × 10⁴cells/cm²and grown to confluence in propagation medium. At confluence, the propagation medium was replaced by the differentiation medium containingαMEM supplemented with 5% FBS, 5% bovine calf serum, 100 U/mL penicillin, 100μg/mL streptomycin, 50μg/mL ascorbic acid, 2.5 mM βglycerophosphate and various concentrations of GIP. This day was considered as day 1. 2′,5′- Dideoxyadenosine (DDA) was used as an adenylyl cyclase inhibitor at a concentration of 50μM. This concentration has been chosen based on preliminary results showing 95% reduction in adenylyl cyclase activity and cAMP production without affecting osteoblast viability. The differentiation medium was replenished every two days.

cAMP determination

Cells were plated at a density of 3 × 10⁵cells/cm²and grown for 48 h inαMEM supplemented with 5% FBS, 5% bovine calf serum, 100 U/mL penicillin, and 100μg/mL streptomycin in a humidiﬁed atmosphere enriched with 5% CO₂at 37 °C. Then cells were starved inαMEM supplemented with 0.5% BSA for 16 h and incubated for 15 min inαMEM supplemented with 0.5% bovine serum albumin and 1 mM IBMX prior to stimulation with GIP. After 45 min, cells were washed with ice-cold PBS, incubated in lysis buffer and centrifuged at 12,000gfor 10 min. Su- pernatants were collected and stored at−80 °C until use. Cyclic AMP determination was performed with aﬂuorometric commercially available kit (R&D Systems Europe, Abingdon, UK) according to the manufacturer's recommendations.

Lysyl oxidase (LOX) activity

LOX activity was assessed in the cell culture supernatant and in the cell layer after 13 days of culture, as this period of time corresponds to the highest expression of lysyl oxidase[23]using aﬂuorometric assay developed by Palamakumbura et al.[28]. At day 12, the differentiation medium was replaced by phenol red-free DMEM containing 0.5%

bovine serum albumin, 50μg/mL ascorbic acid, 2.5 mMβ-glycerophosphate and various concentrations of GIP. After 24 h, cell culture supernatants were collected, centrifuged at 3000 rpm for 30 min at 4 °C, aliquoted and stored at−80 °C until use. The cell layer was extracted in urea buffer as reported in[29]. Then, supernatants were incubated at 37

°C for 30 min in the presence of 1.2 M urea, 50 mM sodium borate (pH 8.2), 1.3 nmol H₂O₂, 1 UI/mL horseradish peroxidase, 10 mM diaminopentane, 10μM Ampliflu^TMred and ± 0.5 mM BAPN in opaque 96 well plates. At the end of the incubation period, the plate was placed on ice andfluorescence was read using a M2 microplate reader (Molecular devices, St Gregoire, France) with excitation and emission wavelength set up at 563 nm and 587 nm, respectively. As BAPN is a specific lysyl oxidase inhibitor, the residual amine oxidase activity evidenced in the presence of BAPN was subtracted from the activity observed in the absence of BAPN. Lysyl oxidase activity was reported as the fold change versus CTRL of the pooled supernatant and cell layer fractions.

Western blotting

MC3T3-E1 cells were were washed in cold PBS and lysates were made by using a lysis buffer containing 50 mM Tris–HCl pH 7.5, 100 mM NaCl, 50 mM NaF, 3 mM Na3VO4, protease inhibitor cocktail and 1% Nonidet P-40. Samples were spun at 13,000 rpm for 30 min at 4 °C, the supernatant was collected and protein concentration was determined by BCA assay (ThermoScientiﬁc, Brebières, France). Samples (20μg per lane) were run on a 10% acrylamide gel and blotted onto a PVDF membrane.

The membranes were washed in Tris buffered saline (TBS) and blocked with 5% bovine serum albumin. Samples were incubated overnight with the anti-GIPR (reference sc-98795, Santa Cruz Biotechnology, Heidelberg, Germany) orβ-actin (Sigma-Aldrich). Subsequently membranes were washed in TBS and incubated with the appropriate second- ary antibodies coupled to HRP (R&D Systems Europe). Immunoreactive bands were visualized using an ECL kit (Amersham, UK).

Fourier transformed infrared microscopy (FTIRM)

For collagen cross-link analysis, the same osteoblast cultures used for LOX activity were processed. Osteoblast cultures wereﬁxed in absolute ethanol, scrapped off the culture dish and transferred onto BaF2windows where they were air-dried. The integrity of the collagen extracellular matrix was veriﬁed by comparing the obtained FTIRM spectrum with those of commercial collagen. Spectral analysis was performed using a Bruker Vertex 70 spectrometer (Bruker optics, Ettlingen, Germany) interfaced

(3)

with a Bruker Hyperion 3000 infrared microscope equipped with a standard single element Mercury Cadmium Telluride (MCT) detector. Infra- red spectra were recorded at a resolution of 4 cm⁻¹, with an average of 32 scans in transmission mode. Background spectral images were collected under identical conditions from the same BaF2windows at the be- ginning and end of each experiment to ensure instrument stability. For FTIRM analysis, 20 spectra were acquired for each condition and analyzed with the Opus Software (release 6.0, Bruker). Water vapor was corrected prior to baseline correction. Spectra were then subjected to second derivative spectroscopy and curveﬁtting routines using a commercial available software package (Grams/AI 8.0, Thermoﬁsher scientif- ic, Villebon sur Yvette, France) as described previously[30]. The collagen maturity index was determined as the relative ratio of pyridinium trivalent (Pyr, mature collagen) to dehydrodihydroxylysinonorleucine divalent (deH-DHLNL, new collagen) collagen cross-links using their re- spective subbands located at 1660 cm⁻¹and 1690 cm⁻¹of the amide I peak[31,32]. The proteoglycan-to-matrix ratio was calculated as the ratio of the 1376 cm⁻¹peak, corresponding to the CH3symmetric bend- ing vibration of GAGs[33], to the Amide I band (1585–1725 cm⁻¹). The sulfated proteoglycan-to-matrix ratio was calculated as the ratio of the 1064 cm⁻¹peak, corresponding to SO3−symmetric stretching vibration of sulfated proteoglycan[34]to the Amide I band.

Transmission electron microscopy (TEM)

After 13 days of culture, osteoblasts and their produced extracellular matrix were rinsed with 0.1 M cacodylate buffer (pH 7.4) andfixed in 2.5% glutaraldehyde in cacodylate buffer. Then cultures were post- fixed with 1% osmium tetroxide/1% potassium ferrocyanide and dehydrated in a graded series of ethanol prior to embedding in epoxy resin. Ultrathin sections were cut with a diamond knife, contrasted with 3% uranyl acetate and observed with a Jeol JEM 1400 (Jeol, Croisy sur Seine, France) operating with an accelerating voltage of 120 KeV and a magnification of ×30,000 (field overview) or ×100,000 (collagen network).

Characterization of collagenﬁbril diameter

In order to avoid confounding factors due to angle incidence during sectioning, image analysis with shape descriptors was used to determine the diameter of cross-sectioned collagenfibrils. TEM photographs were subjected to image analysis using ImageJ 1.45s. Fibril sections werefitted with ellipses to evidence the presence of local tilting. The as- pect ratio (AR), defined by the following equation,

AR¼Diameter major axis diameter minor axis

was computed and onlyfibrils with an ARb1.2 were considered with low probability of local tilting. The minor diameter offitted ellipse was selected as the value offibril diameter as described by Starborg et al.[35]. More than 800fibrils were measured per condition. The diameter distribution was then analyzed with a lab-made routine using ImageJ 1.45s and Excel 2010. Rapidly, the frequency of occurrence of anidiameter was calculated as follows:

Fi¼100N_i N_t

where Nirepresents the number ofﬁbrils with theidiameter and Nt

represents the total number offibrils. The frequency distribution, as a function offibril diameter, was plotted and the following parameters were deduced from this distribution: Dpeakcorresponding to the most frequently observedfibril diameter, D_mean, the averagefibril diameter and Dwidththe width of the histogram at half maximum of the peak.

Statistical analysis

Each experiment has been replicated at least 3 times. Results were expressed as mean ± standard error of the mean (SEM). Non- parametric Mann–WhitneyU-test was used to compare the differences between the groups using the Systat statistical software release 13.0 (Systat software Inc., San Jose, CA). Differences atpb0.05 were considered to be signiﬁcant.

Results

GIP affects lysyl oxidase (LOX) activity and collagen maturity

As representedFig. 1A, GIPR is expressed in MC3T3-E1 osteoblast cells. GIP dose dependently induced a signiﬁcant cAMP stimulation to reach 4 pmoles/well at concentrations of 1 nM and above and an EC50

of 96pM (Fig. 1B). We next investigated LOX activity in the presence of GIP ranged between 10 and 100 pM (Fig. 1C). LOX activity was augmented in cultures treated with as low as 10 pM GIP (1.1-fold increase, pb0.05) as compared with untreated cells. Furthermore, LOX activity was signiﬁcantly increased by 2.2-fold (pb0.001), 3.6-fold (pb0.001) and 3.9-fold (pb0.001) in the presence of 20 pM, 50 pM and 100 pM GIP, respectively.

We then examined the extent of mature and immature collagen cross-links in the extracellular matrices (Fig. 2). The area of the 1690 cm⁻¹subband, representative of the deH-DHLNL immature collagen cross-link, was dose-dependently augmented in GIP-treated culture (Fig. 2A). A similar pattern was observed with the area of the 1660 cm⁻¹, a peak representative of pyridinoline collagen cross-links.

The augmentation of the 1660 cm⁻¹area was even more marked

(C) (B)

(A)

Fig. 1.GIP affects LOX activity and collagen maturity. (A) GIPR expression has been assessed by western blotting in three independent MC3T3 cell culture. (B) cAMP production in response to increasing doses of GIP. (C) LOX activity is dose-dependently increased in the presence of GIP. *:pb0.05 and ***:pb0.001 vs. untreated cultures.

(4)

at higher dose of GIP (Fig. 2B). The collagen maturity index was not sig- niﬁcantly different between vehicle- and 10 pM GIP-treated cultures (p= 0.075) (Fig. 2C). But, at higher concentrations of GIP, signiﬁcant augmentations in this index were observed with 73% (p= 0.047), 127% (p= 0.009) and 131% (p= 0.009) increases in this parameter with 20 pM, 50 pM and 100 pM, respectively. We also assessed whether GIP treatment affected the proteoglycan content of the extracellular matrix. The PG-to-matrix and sulfated PG-to-matrix ratios were unaffected by the presence of GIP (Figs. 2D and E).

GIP affects collagenﬁbril diameters

As represented inFig. 3A, it seemed thatﬁbril diameters were reduced in the presence of 100 pM GIP. The frequency distribution ofﬁbril diameters also appeared shifted towards lower diameter in the

presence of 100 pM GIP (Fig. 3B). Indeed, Dmean, the mean diameter, and Dpeak, the most often observed diameter, were dramatically reduced, in a dose-dependent manner, in the presence of GIP with lower diameters observed at the highest tested concentrations (Table 1). On the other hand, Dwidththat characterizes heterogeneity inﬁbril diameters was not signiﬁcantly different between untreated and GIP-treated cultures at all tested concentrations (Table 1).

Adenylyl cyclase activity is required for GIP-mediated increases in LOX activity and collagen crosslinking

As binding of GIP to its receptor in osteoblasts leads to activation of adenylyl cyclase and production of cAMP, we ascertained whether the observed increases in LOX activity and collagen cross-linking required a functional adenylyl cyclase. The use of the adenylyl cyclase inhibitor

(C) (B)

(A)

(E) (D)

Fig. 2.GIP affects collagen maturity. (A) Amount of deH-DHLNL, (B) amount of Pyr and (C) collagen maturity index, expressed as the ratio of Pyr/deH-DHLNL (1660 cm⁻¹/1690 cm⁻¹), are dose-dependently increased in the presence of GIP. (D) Total proteoglycan content and (E) sulfated proteoglycan content in the extracellular bone matrix. *:pb0.05, **:pb0.01 and ***:

pb0.001 vs. untreated cultures.

(B) (A)

Fig. 3.GIP affects collagenfibril diameter. (A) TEM photographs of untreated (left) and 100 pM GIP-treated (right) cells. Bar = 200 nm. Inset represents a higher magnification of collagen fibrils. Bar in inset = 100 nm. (B) Distribution frequency offibril diameter between untreated (black) and 100 pM GIP-treated (gray) cells. Note the shift toward lower values forfibril diameter in GIP-treated cells.

(5)

2′,5′-dideoxyadenosine (DDA) significantly decreased the production of cAMP in GIP-treated cultures (Fig. 4A). Furthermore, DDA had no influ- ence on LOX activity in vehicle treated-cultures (p= 0.507,Fig. 4B). On the other hand, this activity was reduced by 25% in the presence of DDA in GIP-treated cultures (pb0.01,Fig. 4B). As a consequence, addition of DDA in 100 pM GIP-treated cultures resulted in reductions in both 1690 cm⁻¹and 1660 cm⁻¹area, suggesting lower amount of immature and mature collagen cross-link, respectively (Figs. 4C and D). The collagen maturity index was also significantly reduced in the presence of DDA and 100 pM GIP (−36%,p= 0.016,Fig. 4E) as compared with GIP alone-treated cultures.

Adenylyl cyclase activity is required for GIP-mediated modiﬁcations of the collagen matrix

As representedFig. 5, thefibril diameter seemed unmodified in control cultures treated with or without DDA. On the other hand, thefibril diameter in GIP-treated cultures seemed higher in the presence of DDA. The frequency distribution offibril diameter was unaffected in control cultures by the presence of DDA whilst the frequency distribution in GIP-treated cultures in the presence of DDA was broadened and shifted toward higher values. Not surprisingly, GIP treatment led to a decrease in Dpeakand Dmeanas reported above whilst Dwidthwas unmodified (Table 2). In control cultures, the addition of DDA did not affect any of the distribution frequency parameters. However, addition of DDA in GIP-treated cultures resulted to significant increases in Dmeanand Dwidth

but not Dpeak. The addition of DDA in GIP-treated cultures led to values for D_peakand D_widthnot signiﬁcantly different from those observed in control cultures.

Discussion

Previously, in animal models of GIPR deficiency or stable GIP analog treatment, evidences have been provided that maturity of the collagen bone matrix was modified by the GIP/GIPR pathway[11,12]. However due to the pleiotropic expression of GIPR, it was unclear whether this modification of collagen maturity resulted from direct action of the GIP/GIPR pathway on bone cells. In the present study, using a validated model of osteoblast cultures, we evidenced a significant stimulation of cAMP production in response to increasing doses of GIP, significant augmentations in LOX activity and collagen maturity in GIP-treated cultures and significant reductions in collagenfiber diameters. These modifications of LOX activity and collagen properties were partly due to stimulation of adenylyl cyclase and production of cAMP as evidenced by alterations of these modifications in the presence of the adenylyl cyclase inhibitor, 2′,5′-dideoxyadenosine.

Table 1

Properties of collagenﬁbrils in the presence of GIP.

GIP (pM) Dpeak(nm) Dmean(nm) Dwidth(nm)

– 32.7 ± 0.3 33.3 ± 0.5 10.7 ± 0.8

10 30.3 ± 0.7* 29.7 ± 0.5** 12.3 ± 2.6

20 28.5 ± 0.3** 29.3 ± 0.8** 10.4 ± 1.2

50 27.9 ± 0.8** 27.9 ± 0.2** 9.6 ± 0.5

100 27.8 ± 0.7** 27.8 ± 0.5** 9.5 ± 0.7

*:pb0.05 and **:pb0.01 vs. cells cultured in the absence of GIP.

Table 2

Properties of collagenﬁbrils in the presence of an adenylyl cyclase inhibitor. Osteoblast cultures were treated with vehicle or 100 pM GIP in the presence of 50μM DDA.

Treatment Dpeak(nm) Dmean(nm) Dwidth(nm)

Vehicle 32.0 ± 0.7 32.7 ± 0.6 11.8 ± 1.3

Vehicle + DDA 32.0 ± 1.2 32.7 ± 1.1 12.5 ± 0.8

GIP 28.0 ± 0.2^# 27.7 ± 0.2^# 10.4 ± 0.8

GIP + DDA 29.1 ± 0.8 30.2 ± 0.2**^# 13.8 ± 0.6 *

*:pb0.05 and **:pb0.01 vs. cells cultured in the absence of DDA.

# pb0.05 vs. vehicle.

(B) (A)

(E)

(C) (D)

Fig. 4.GIP effects on LOX activity and collagen maturity are due to a cAMP-dependent mechanism. (A) cAMP stimulation in 100 pM GIP culture is abolished by the use of 50μM 2′,5′- dideoxyadenosine (DDA). (B) Increase in LOX activity observed in the presence of 100 pM GIP is signiﬁcantly reduced in DDA-treated cultures. White bars = untreated or GIP-treated cultures in the absence of DDA, black bars = untreated or GIP-treated cultures in the presence of DDA. (C) Amount of deH-DHLNL, (D) amount of Pyr and (E) collagen maturity index are signiﬁcantly reduced in 100 pM GIP-treated cultures in the presence of DDA. White bars = untreated or GIP-treated cultures in the absence of DDA, black bars = untreated or GIP- treated cultures in the presence of DDA. *:pb0.05 vs. cells cultured in the absence of DDA, **:pb0.01 and ***:pb0.001 vs. cells cultured in the absence of DDA.

(6)

Collagen must be cross-linked to exhibit the normal physical properties of tensile strength. Theﬁnal enzymatic step of collagen cross-linking is ensured by LOX, an enzyme with amine oxidase activity, generating peptidyl-δ-hydroxy-δ-aminoadipic-δ-semialdehyde and peptidyl-α- aminoadipic-δ-semialdehyde from hydroxylysine and lysine residues, respectively [36]. These aldehydes then spontaneously react with other aldehydes or unmodiﬁed lysine and hydroxylysine residues to form a variety of intra- and intermolecular cross-links in collagens.

LOX is synthesized in osteoblast as a pro-enzyme and excreted in the extracellular environment as a pro-enzyme. The pro-enzyme is then processed by procollagen C-proteinases in the extracellular environment to give rise to the active 32 kDa enzyme and an 18 kDa propeptide[37, 38]. Additionally, four LOX-like enzymes (LOXL1-4) have been identified and present significant sequence identity with mature LOX[39–44]. In MC3T3-E1 cells, LOX, LOXL1, LOXL3 and LOXL4 have been shown to be the prominent form at the RNA level[45]. Furthermore, LOXL1, LOXL3 and LOXL4 also present amine oxidase activities[46–48]. In the present study, the LOX activity was determined in conditioned media of MC3T3- E1 cells stimulated with various concentrations of GIP. We observed a significant increase in this parameter; however it is unclear whether this observed increased is related to increases in only LOX or LOXL activities. Furthermore, it is also unclear whether the inhibitory action of DDA on LOX activity reflects direct inhibition of LOX or LOXLs activities or a combination of both. However, it is worth noting that inlox−/− animals, significant reductions in LOXLs at the RNA levels were observed [49]. Nevertheless, the increase in LOX activity after GIP stimulation was accompanied by an augmentation in collagen cross-linking as determined spectroscopically. The ratio between the subbands located at 1660 cm⁻¹and 1690 cm⁻¹is a cheap and validated methodology for the determination of collagen cross-linking[32]. This methodology required second-derivative spectral deconvolution and curvefitting in order to resolve underlying bands of the amide I peak. However, it can be applied either on tissue sections or cell culture material as we performed in the present study[50–56].

Not only collagen maturation but also collagenfibril diameter was affected by GIP treatment. Indeed, we observed a significant 17% reduction in the meanfibril diameter in the presence of 100 pM GIP. However, the same heterogeneity infibril diameter distribution was observed between control and GIP-treated cultures as determined by the Dwidthpa- rameter. Collagenfibril assembly was thought to be an entropy-driven process as observed for microtubules and actinfilament. On the other hand, some non-collagenous proteins have been implicated infibrils assembly. Among them, biglycan and decorin, two proteoglycans of the extracellular matrix have been suggested. Biglycan and decorin are composed of a protein core with leucin-reach repeats attached to chains of chondroitin sulfate. To determine the content of total proteoglycans in the extracellular matrix we chose a spectroscopic approach and we used the 1376 cm⁻¹band instead of the conventional 940– 1140 cm⁻¹region. Indeed, although the 940–1140 cm⁻¹area com- prises the C\OH and C\C ring vibrations of the carbohydrate in proteoglycans, this region might also contain information from plastic particles detached during extracellular matrix scrapping and ethanol residues (mainly between 940 and 1050 cm⁻¹). Recently, Rieppo et al. demonstrated that the ratio of the 1376 cm⁻¹band over the entire integrated area of the Amide I peak represented a good indicator of total proteoglycan content[33]. Furthermore, these authors made also the demonstra- tion of the usefulness of the 1064 cm⁻¹band as a marker of sulfated proteoglycans[33]. The overall proteoglycan content and the sulfated proteoglycan content of the extracellular matrix were determined and appeared unaffected by GIP treatment. As such, it is unlikely, although we did not determine the expression of biglycan and decorin in the extracellular matrix, that the observed reductions in collagenfibril diameters were due to a modified pattern of proteoglycan expression.

The GIP receptor has been shown to be expressed and functional in several osteoblastic cell lines (Saos-2, TE-85, MG-63) and primary murine osteoblasts[6,8]. The GIP receptor is also expressed on MC3T3-E1 cells and addition of GIP in the cultures resulted in a signiﬁcant stimulation of cAMP with an EC50 around 100 pM. The observed EC50is

(B)

(C) (A)

Fig. 5.GIP effects on collagenfibril diameter are due to a cAMP-dependent mechanism. (A) TEM photographs of extracellular matrices in vehicle, vehicle + DDA, GIP and GIP + DDA conditions. Bar = 100 nm. (B) Frequency distribution offibril diameter in vehicle (black) and vehicle + DDA (gray) cultures. (C) Frequency distribution offibril diameter in GIP (black) and GIP + DDA (gray) cultures.

(7)

relatively close to the physiological circulating plasma levels of GIP en- countered after a meal. The use of a speciﬁc adenylyl cyclase inhibitor, 2′,5′-dideoxyadenosine was capable of reducing by 88%, the production of cAMP in MC3T3-E1 cells in response to GIP. Moreover, DDA did not only abolish the cAMP response but also contributed to signiﬁcant reductions in LOX activity (−25%) and collagen cross-linking (−36%).

Furthermore, collagen deposition was also affected by the use of DDA in GIP-treated cultures and signiﬁcant augmentations in the meanﬁbril diameter and distribution heterogeneity were evidenced. Taken together, these data support a direct role for a GIPR-adenylyl cyclase-cAMP axis in the control of extracellular matrix properties in response to GIP. Noteworthy is the residual cAMP activity in the presence of GIP/

DDA. Indeed, in this condition, MC3T3-E1 exhibited still a 7% increase in cAMP as compared with untreated cells. One could wonder why we did not use higher dose of DDA to completely abolish the rise in cAMP.

As a preliminary experiment, we used several doses of DDA and 50μM was the highest dose usable as higher concentration led to cell death in the culture. Of course, we cannot rule out that other intracellular pathways, independent of adenylyl cyclase activation, might also be stimulated by the binding of GIP to the GIPR and as such might contrib- ute to the observed modiﬁcations of extracellular matrix.

In conclusion, the present study supports a direct role for the GIPR- adenylyl cyclase-cAMP in the control of extracellular matrix deposition and properties in osteoblasts. GIP-treated osteoblasts presented with augmented LOX activity and collagen cross-linking as well as modiﬁca- tion of collagen deposition.

References

[1]Andersen DK, Elahi D, Brown JC, Tobin JD, Andres R. Oral glucose augmentation of insulin secretion. Interactions of gastric inhibitory polypeptide with ambient glucose and insulin levels. J Clin Invest 1978;62:152–61.

[2]Schirra J, Katschinski M, Weidmann C, Schafer T, Wank U, Arnold R, et al. Gastric emptying and release of incretin hormones after glucose ingestion in humans.

J Clin Invest 1996;97:92–103.

[3]Varner AA, Isenberg JI, Elashoff JD, Lamers CB, Maxwell V, Shulkes AA. Effect of intravenous lipid on gastric acid secretion stimulated by intravenous amino acids.

Gastroenterology 1980;79:873–6.

[4]Jorde R, Burhol PG, Schulz TB. Fasting and postprandial plasma GIP values in man measured with seven different antisera. Regul Pept 1983;7:87–94.

[5]Baggio LL, Drucker DJ. Biology of incretins: GLP-1 and GIP. Gastroenterology 2007;

132:2131–57.

[6]Bollag RJ, Zhong Q, Phillips P, Min L, Zhong L, Cameron R, et al. Osteoblast-derived cells express functional glucose-dependent insulinotropic peptide receptors.

Endocrinology 2000;141:1228–35.

[7]Mabilleau G, Gaudin-Audrain C, Irwin N, Flatt PR, Basle MF, Chappard D. Deﬁciency in glucose-dependent insulinotropic peptide receptor results in higher bone mass in male mice. Osteoporos Int 2012;23:S407–8.

[8]Pacheco-Pantoja EL, Ranganath LR, Gallagher JA, Wilson PJ, Fraser WD. Receptors and effects of gut hormones in three osteoblastic cell lines. BMC Physiol 2011;11:12.

[9]Zhong Q, Itokawa T, Sridhar S, Ding KH, Xie D, Kang B, et al. Effects of glucose- dependent insulinotropic peptide on osteoclast function. Am J Physiol Endocrinol Metab 2007;292:E543–8.

[10]Gaudin-Audrain C, Irwin N, Mansur S, Flatt PR, Thorens B, Basle M, et al. Glucose- dependent insulinotropic polypeptide receptor deﬁciency leads to modiﬁcations of trabecular bone volume and quality in mice. Bone 2013;53:221–30.

[11]Mieczkowska A, Irwin N, Flatt PR, Chappard D, Mabilleau G. Glucose-dependent insulinotropic polypeptide (GIP) receptor deletion leads to reduced bone strength and quality. Bone 2013;56:337–42.

[12]Mabilleau G, Mieczkowska A, Irwin N, Simon Y, Audran M, Flatt PR, et al. Beneﬁcial effects of a N-terminally modiﬁed GIP agonist on tissue-level bone material properties. Bone 2014;63C:61–8.

[13]Chappard D, Basle MF, Legrand E, Audran M. New laboratory tools in the assessment of bone quality. Osteoporos Int 2011;22:2225–40.

[14]Currey JD. The mechanical consequences of variation in the mineral content of bone.

J Biomech 1969;2:1–11.

[15]Currey JD. The effect of porosity and mineral content on the Young's modulus of elasticity of compact bone. J Biomech 1988;21:131–9.

[16]Wang X, Bank RA, TeKoppele JM, Agrawal CM. The role of collagen in determining bone mechanical properties. J Orthop Res 2001;19:1021–6.

[17]Wang X, Shen X, Li X, Agrawal CM. Age-related changes in the collagen network and toughness of bone. Bone 2002;31:1–7.

[18]Prockop DJ, Kivirikko KI. Collagens: molecular biology, diseases, and potentials for therapy. Annu Rev Biochem 1995;64:403–34.

[19]Yamauchi M, Sricholpech M. Lysine post-translational modiﬁcations of collagen.

Essays Biochem 2012;52:113–33.

[20]Kadler KE, Holmes DF, Trotter JA, Chapman JA. Collagenﬁbril formation. Biochem J 1996;316(Pt 1):1–11.

[21]Corsi A, Xu T, Chen XD, Boyde A, Liang J, Mankani M, et al. Phenotypic effects of biglycan deficiency are linked to collagenfibril abnormalities, are synergized by decorin deficiency, and mimic Ehlers–Danlos-like changes in bone and other con- nective tissues. J Bone Miner Res 2002;17:1180–9.

[22]Danielson KG, Baribault H, Holmes DF, Graham H, Kadler KE, Iozzo RV. Targeted disruption of decorin leads to abnormal collagenﬁbril morphology and skin fragility.

J Cell Biol 1997;136:729–43.

[23]Hong HH, Pischon N, Santana RB, Palamakumbura AH, Chase HB, Gantz D, et al. A role for lysyl oxidase regulation in the control of normal collagen deposition in dif- ferentiating osteoblast cultures. J Cell Physiol 2004;200:53–62.

[24]Fields GB, Noble RL. Solid phase peptide synthesis utilizing 9-ﬂuorenylmethoxycarbonyl amino acids. Int J Pept Protein Res 1990;35:161–214.

[25]Kuhn-Wache K, Manhart S, Hoffmann T, Hinke SA, Gelling R, Pederson RA, et al. An- alogs of glucose-dependent insulinotropic polypeptide with increased dipeptidyl peptidase IV resistance. Adv Exp Med Biol 2000;477:187–95.

[26]McIntosh CH, Widenmaier S, Kim SJ. Glucose-dependent insulinotropic polypeptide (Gastric Inhibitory Polypeptide; GIP). Vitam Horm 2009;80:409–71.

[27]Franceschi RT, Iyer BS, Cui Y. Effects of ascorbic acid on collagen matrix formation and osteoblast differentiation in murine MC3T3-E1 cells. J Bone Miner Res 1994;9:

843–54.

[28]Palamakumbura AH, Trackman PC. Aﬂuorometric assay for detection of lysyl oxidase enzyme activity in biological samples. Anal Biochem 2002;300:245–51.

[29]Bedell-Hogan D, Trackman P, Abrams W, Rosenbloom J, Kagan H. Oxidation, cross- linking, and insolubilization of recombinant tropoelastin by puriﬁed lysyl oxidase.

J Biol Chem 1993;268:10345–50.

[30]Mabilleau G, Mieczkowska A, Irwin N, Flatt PR, Chappard D. Optimal bone mechanical and material properties require a functional GLP-1 receptor. J Endocrinol 2013;

219:59–68.

[31] Paschalis EP, Gamsjaeger S, Tatakis DN, Hassler N, Robins SP, Klaushofer K. Fourier transform infrared spectroscopic characterization of mineralizing type I collagen enzymatic trivalent cross-links. Calcif Tissue Int 2014.http://dx.doi.org/10.1007/

s00223-014-9933-9.

[32]Paschalis EP, Verdelis K, Doty SB, Boskey AL, Mendelsohn R, Yamauchi M. Spectroscop- ic characterization of collagen cross-links in bone. J Bone Miner Res 2001;16:1821–8.

[33]Rieppo L, Saarakkala S, Narhi T, Helminen HJ, Jurvelin JS, Rieppo J. Application of second derivative spectroscopy for increasing molecular speciﬁcity of Fourier transform infrared spectroscopic imaging of articular cartilage. Osteoarthritis Cartilage 2012;

20:451–9.

[34]Servaty R, Schiller J, Binder H, Arnold K. Hydration of polymeric components of cartilage—an infrared spectroscopic study on hyaluronic acid and chondroitin sulfate.

Int J Biol Macromol 2001;28:121–7.

[35]Starborg T, Kalson NS, Lu Y, Mironov A, Cootes TF, Holmes DF, et al. Using transmission electron microscopy and 3View to determine collagenﬁbril size and three- dimensional organization. Nat Protoc 2013;8:1433–48.

[36]Trackman PC. Diverse biological functions of extracellular collagen processing enzymes. J Cell Biochem 2005;96:927–37.

[37]Panchenko MV, Stetler-Stevenson WG, Trubetskoy OV, Gacheru SN, Kagan HM.

Metalloproteinase activity secreted byﬁbrogenic cells in the processing of prolysyl oxidase. Potential role of procollagen C-proteinase. J Biol Chem 1996;271:7113–9.

[38]Uzel MI, Scott IC, Babakhanlou-Chase H, Palamakumbura AH, Pappano WN, Hong HH, et al. Multiple bone morphogenetic protein 1-related mammalian metallopro- teinases process pro-lysyl oxidase at the correct physiological site and control lysyl oxidase activation in mouse embryoﬁbroblast cultures. J Biol Chem 2001;

276:22537–43.

[39]Huang Y, Dai J, Tang R, Zhao W, Zhou Z, Wang W, et al. Cloning and characterization of a human lysyl oxidase-like 3 gene (hLOXL3). Matrix Biol 2001;20:153–7.

[40]Ito H, Akiyama H, Iguchi H, Iyama K, Miyamoto M, Ohsawa K, et al. Molecular cloning and biological activity of a novel lysyl oxidase-related gene expressed in cartilage.

J Biol Chem 2001;276:24023–9.

[41]Jourdan-Le Saux C, Le Saux O, Donlon T, Boyd CD, Csiszar K. The human lysyl oxidase-related gene (LOXL2) maps between markers D8S280 and D8S278 on chromosome 8p21.2–p21.3. Genomics 1998;51:305–7.

[42]Kenyon K, Modi WS, Contente S, Friedman RM. A novel human cDNA with a predict- ed protein similar to lysyl oxidase maps to chromosome 15q24–q25. J Biol Chem 1993;268:18435–7.

[43]Maki JM, Kivirikko KI. Cloning and characterization of a fourth human lysyl oxidase isoenzyme. Biochem J 2001;355:381–7.

[44]Maki JM, Tikkanen H, Kivirikko KI. Cloning and characterization of aﬁfth human lysyl oxidase isoenzyme: the third member of the lysyl oxidase-related subfamily with four scavenger receptor cysteine-rich domains. Matrix Biol 2001;20:493–6.

[45]Atsawasuwan P, Mochida Y, Parisuthiman D, Yamauchi M. Expression of lysyl oxidase isoforms in MC3T3-E1 osteoblastic cells. Biochem Biophys Res Commun 2005;327:1042–6.

[46]Kim MS, Kim SS, Jung ST, Park JY, Yoo HW, Ko J, et al. Expression and puriﬁcation of enzymatically active forms of the human lysyl oxidase-like protein 4. J Biol Chem 2003;278:52071–4.

[47]Kim S, Kim Y. Variations in LOXL1 associated with exfoliation glaucoma do not affect amine oxidase activity. Mol Vis 2012;18:265–70.

[48]Lee JE, Kim Y. A tissue-speciﬁc variant of the human lysyl oxidase-like protein 3 (LOXL3) functions as an amine oxidase with substrate speciﬁcity. J Biol Chem 2006;281:37282–90.

[49]Pischon N, Maki JM, Weisshaupt P, Heng N, Palamakumbura AH, N'Guessan P, et al.

Lysyl oxidase (lox) gene deﬁciency affects osteoblastic phenotype. Calcif Tissue Int 2009;85:119–26.

(8)

[50]Mourant JR, Yamada YR, Carpenter S, Dominique LR, Freyer JP. FTIR spectroscopy demonstrates biochemical differences in mammalian cell cultures at different growth stages. Biophys J 2003;85:1938–47.

[51]Paschalis EP, Mendelsohn R, Boskey AL. Infrared assessment of bone quality:

a review. Clin Orthop Relat Res 2011;469:2170–8.

[52]Ramesh J, Salman A, Hammody Z, Cohen B, Gopas J, Grossman N, et al. FTIR micro- scopic studies on normal and H-ras oncogene transfected cultured mouseﬁbro- blasts. Eur Biophys J 2001;30:250–5.

[53]Thaler R, Spitzer S, Rumpler M, Fratzl-Zelman N, Klaushofer K, Paschalis EP, et al. Dif- ferential effects of homocysteine and beta aminopropionitrile on preosteoblastic MC3T3-E1 cells. Bone 2010;46:703–9.

[54]Thaler R, Zwerina J, Rumpler M, Spitzer S, Gamsjaeger S, Paschalis EP, et al. Homo- cysteine induces serum amyloid A3 in osteoblasts via unlocking RGD-motifs in collagen. FASEB J 2013;27:446–63.

[55]Turecek C, Fratzl-Zelman N, Rumpler M, Buchinger B, Spitzer S, Zoehrer R, et al. Col- lagen cross-linking inﬂuences osteoblastic differentiation. Calcif Tissue Int 2008;82:

392–400.

[56]Varga F, Rumpler M, Zoehrer R, Turecek C, Spitzer S, Thaler R, et al. T3 affects expression of collagen I and collagen cross-linking in bone cell cultures. Biochem Biophys Res Commun 2010;402:180–5.