Real-Time Spectroscopic Analysis of Extracellular Matrix Produced by MC3T3-E1 Preosteoblastic Cells Cultured Under Dynamic Conditions NORBERT HASSLER,* MONIKA RUMPLER, ROMAN THALER, RICHARD MENDELSOHN, ROGER PHIPPS, FRANZ VARGA, URS P. FRINGELI, KLAUS KLAUSHOFER, and ELEFTHERIOS P. PASCHALIS Ludwig Boltzmann Institute of Osteology at the Hanusch Hospital of WGKK and AUVA Trauma Centre Meidling, 1st Med. Dept., Hanusch Hospital, Vienna, Austria (N.H., M.R., R.T., F.V., K.K., E.P.P.); Chemistry Dept., Rutgers University, Newark, New Jersey (R.M.); Husson University School of Pharmacy, Bangor, Maine (R.P.); Institute of Biophysical Chemistry, University of Vienna, Austria (U.P.F.)
The deposition of extracellular matrix (ECM) produced by osteoblasts is one of the first steps in bone formation. Composition and structure of the ECM influence the development and strength of bone, as well as the onset of its mineralization. Since ECM is secreted onto the surface where the cells attach, Fourier transform infrared (FT-IR) attenuated total reflection (ATR), as a surface sensitive technique, provides a useful tool for its investigation, as ECM instead of the cells is predominantly detected by the IR beam. The purpose of the present study was to develop the FT-IR ATR technique so that real-time measurements of the ECM produced by MC3T3-E1 osteoblasts could be obtained in situ. Measurements were performed using polarized incident IR light to apply a procedure for solvent compensation which reduces the influence of culture medium on the evaluation of the amide I and II bands. The formation of ECM took place in a flow-through chamber under dynamic conditions by applying a constant flow of culture medium and was tracked over a time period of two weeks by evaluation of the integrated absorbance of amide I and II bands reflecting the amount and isotropic arrangement of amide bonds in the ECM. Cultures without ascorbic acid had a reduced protein concentration that enabled the analysis of cell-mediated matrix accumulation. Presence and proliferation of cells after two weeks of permanent flow-through of culture medium was shown by cell counting exhibiting a 67% increase in cell number as well as by crystal violet and live/dead staining. These results demonstrate the application of FT-IR ATR spectroscopy for monitoring matrix formation. Index Headings: Fourier transform infrared spectroscopy; Attenuated total reflection; FT-IR ATR; Osteoblast; Extracellular matrix; Dynamic cell culture.
INTRODUCTION The extracellular matrix (ECM), secreted by osteoblasts, fulfills many functions beyond providing structural support for cell adhesion, proliferation, and differentiation. There is a tight bidirectional interaction between the ECM and the cells. On one hand, environmental signals are transmitted from the ECM into the cell via adhesion molecules, such as integrins, followed by an activation of specific intracellular signaling mechanisms, where they contribute to the regulation of cell responses such as proliferation and differentiation. On the other hand, the maturation status of cells affects composition of the ECM, since all single components present in the ECM are produced by the cells. The ECM is a complex protein network composed Received 28 February 2011; accepted 11 October 2011. * Author to whom correspondence should be sent. E-mail: norbert.
[email protected]. Deceased on October 6th, 2008. DOI: 10.1366/11-06282
40
Volume 66, Number 1, 2012
of a diversity of proteins that includes collagens, fibronectin, proteoglycans, growth factors, cytokines, etc. A recent study has identified more than 150 component molecules that cooperate and mediate the link between cells and their ECM.1,2 Many ECM proteins act as autocrine and paracrine factors to regulate the commitment and differentiation of the cells surrounded by it, through a kind of feedback mechanism.3–8 Fourier transform infrared (FT-IR) spectra provide information on the organic matrix (mainly collagen) through three intense peptide bond vibrations termed amide I, II, and III. Amide I arises from the motion of the peptide bond C=O stretch, while amide II and III result from mixed motions of C–N stretch and N–H in-plane bend, with amide III possessing in addition a significant contribution from coupled N–H/C–H deformations.9 Attenuated total reflection (ATR) spectroscopy is an application of FT-IR spectroscopy used for investigations of processes within several microns of a surface. It provides the sensitive and nondestructive technique that is needed to study biomembranes, monolayers, or thin films with respect to surface concentration and molecular structure.10,11 FT-IR ATR spectra measured with parallel and perpendicular polarized IR light permit access to information about functional groups of molecules, their quantity, and their spatial orientations.12,13 The purpose of the present experiments was to develop a method to monitor ECM producing osteoblasts (MC3T3-E1) on germanium (Ge) plates, in real time, under dynamic conditions, employing FT-IR ATR spectroscopy coupled to a flow-through chamber.
MATERIALS AND METHODS Cell Culture. The preosteoblastic cell line MC3T3-E1 derived from newborn mouse calvariae (kindly donated by Dr. Kumegawa, Meikai University, Department of Oral Anatomy, Sakado, Japan) was used for matrix production. Cells were cultured in a-MEM (Minimum Essential Medium; Biochrom, Berlin, Germany) containing 5% fetal calf serum (PAA laboratories, Pasching, Austria), 4.5 g/L glucose, and 10 mg/ L gentamycin (Sigma) in humidified air under 5% CO2 at 37 8C. For propagation cells were subcultured twice a week before reaching confluence using 0.001% pronase E (Roche, Mannheim, Germany) and 0.02% EDTA in phosphate-buffered saline (PBS). For the dynamic culture in the flow-through chamber, a HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid)
0003-7028/12/6601-0040$2.00/0 Ó 2012 Society for Applied Spectroscopy
APPLIED SPECTROSCOPY
//Xinet/production/a/apls/live_jobs/apls-66-01/apls-66-01-03/layouts/apls-66-01-03.3d Thursday, 1 December 2011 9:02 am Allen Press, Inc. Page 40
FIG. 1. Schematic view of the flow-through chamber with attached plates for temperature control after mounting of the Ge MIRE. Both compartments are independently accessible for liquids. Each of them is sealed by a Viton O-ring and has a volume of about 100 lL.
buffered culture medium with pH 7.4 consisting of a-MEM, 25 mM HEPES, 4.5 g/L glucose, 5% fetal calf serum, 10 mg/L gentamycin, and 50 mg/L ascorbic acid14 (denoted in this study as complete medium) was used. In the experiment without ascorbic acid the latter was omitted. Fourier Transform Infrared Attenuated Total Reflection Setup. The FT-IR spectra were recorded on a Bruker Equinox 55 spectrometer at 4 cm1 resolution using double-sided interferograms, Blackman–Harris three-term apodization, a zero filling factor of 4, and power spectrum as Fourier transformation. The spectrometer, purged with dry carbondioxide-free air, was equipped with an external ATR attachment (OPTISPEC, Neerach, Switzerland), a mercurycadmium telluride (MCT) detector, and an aluminum grid polarizer on a KRS-5 substrate (Specac, Orpington, U.K.), thus enabling measurements with computer-controlled switching between parallel (pp) and perpendicular (vp) polarized IR light. A total of 800 scans were collected per measurement with pp and vp IR light, respectively. A trapezoidal germanium multiple internal reflection element (Ge MIRE) with dimensions of 52 3 30 3 2.0 mm3 and an angle of incidence of h = 458 was mounted in a hydrodynamically optimized flow-through chamber made of PEEK (polyether ether ketone). A sketch of the chamber is shown in Fig. 1. Two compartments with a volume of about 100 lL each were placed on the MIRE. One compartment was used for cell growing, the other for real-time reference measurements of the culture medium. The mean length corresponded to N ; 10 active internal reflections. The compartments were connected with PTFE (polytetrafluorethylene) tubes to an external peristaltic pump (Ismatec, Glattbrugg, Switzerland). The flow-through chamber was kept at (37 6 0.5) 8C by attached thermostatization plates. Experimental Procedure. Treatment of Germanium (Ge) Plates. Before use the measuring element was subjected to several cleaning steps. Polishing of the surface with diamond paste (0.25 lm grain size) guaranteed absence of deposited ECM and adsorbed proteins. Subsequent washing with acetone, water, and ethanol resulted in a surface free of soluble inorganic and organic substances. Prior to assembly, the flow-
through chamber and measuring element were placed for at least 2 hours in ethanol (70% in water) for sterilization purposes. All steps including the manipulation of culture medium were performed under sterile conditions. Dynamic Cell Culture. After assembly of the flow-through chamber and filling of the compartments with HEPES buffered medium, a suspension of MC3T3-E1 cells cultured in the absence of ascorbic acid was prepared. One compartment was filled with the suspension containing 1.2 3 109 cells/L (resulting in 24 000 cells per cm2 on the Ge plate). Following this, the chamber was stored in horizontal position at 37 8C to enable adhesion of cells to the surface of the reflection element. After 4 hours, the suspension was replaced by fresh medium to remove cells not attached to the measuring element. The chamber was thereafter mounted in the spectrometer and connected to a continuous medium supply at a flow rate of (670 6 15) lL/h, corresponding to a 6.7-fold exchange of medium in the compartment per hour. Formation of extracellular matrix (ECM) was monitored in situ over a time period of two weeks by recording spectra. The time length of two weeks was chosen as we are interested in the properties of the ECM and culturing for longer periods of time would potentially result in mineral deposition of these osteoblastic cells,15,16 thus potentially leading to spectral interference from the v1 and v3 stretching vibrations of the (PO4)3 groups of bone apatite with bands representative of ECM constituents such as proteoglycans.17 Afterwards, adhered cells were treated according to one of the following procedures. (i) Cells were killed by pumping of ethanol (70% in water) into the chamber for 5 min. After replacement of ethanol by medium, the latter was pumped through the chamber for another 3 days. Spectra were recorded applying the same conditions as in the presence of living cells. This would enable us to determine whether the infrared bands monitored were due to the ECM produced by living cells or random physical chemical adsorption of the culturing media components without the requirement of living cells. (ii) After disassembling the chamber, cells adhered to the Ge plate were washed with PBS and detached by 0.001% pronase E/0.02% EDTA in PBS treatment. The number of cells was determined by a cell counter (CASY, Schaerfe Systems, Reutlingen, Germany). (iii) Cells were stained and imaged on the measuring element (see next section) to determine viability under flow conditions. Cell Staining. After disassembling the flow-through chamber, cells adhered to the Ge plate were washed with PBS and subjected to following staining procedures consecutively at room temperature. Photos were taken on a Nikon Eclipse 80i microscope in the fluorescence (viability staining) and reflection (crystal-violet staining) mode, respectively. For assessment of dead and living cells, a live/dead viability staining (Invitrogen Corp.) was performed. Cells were incubated with ethidium homodimer and calcein AM (calcein acetoxymethylester) using a concentration of 8 lM each in medium for 30 min and subsequently washed with PBS. The viability staining is based on the fact that non-fluorescent cellpermeant calcein AM is cleaved by intracellular esterase activity resulting in fluorescent calcein (green), whereas ethidium homodimer (red) is able to pass disturbed membranes, an earmark of dead or heavily damaged cells. Before staining with crystal-violet (Sigma) the cell layer was fixed with paraformaldehyde (4% in PBS) for 15 min and then
APPLIED SPECTROSCOPY
41
//Xinet/production/a/apls/live_jobs/apls-66-01/apls-66-01-03/layouts/apls-66-01-03.3d Thursday, 1 December 2011 9:02 am Allen Press, Inc. Page 41
washed with PBS. Afterwards, cells were incubated with crystal-violet (1% in PBS) for 10 min. Determination of Amount and Orientation of Functional Groups. In ATR spectroscopy the absorbance of vibrational bands is not only dependent on the number but also on the arrangement/orientation of the transition dipole moments of the respective functional groups.12 For that reason, it is necessary to perform measurements in at least two different light polarizations; generally pp and vp infrared light is used. The relevant parameter for analysis of orientation is the socalled dichroic ratio R. It represents the ratio of the integrated absorbances (Rexp = app/avp) of vibrational bands determined experimentally by measurements with pp and vp incident infrared light. It contains information about the orientation of functional groups with respect to the surface of the reflection element. It should be noted that peak absorbances can also be used. A theoretical dichroic ratio Riso of isotropically arranged functional groups is accessible by calculation according to ATR theory.12 Riso is dependent on the angle of incidence h of the IR beam onto the reflection element (in our case h = 458) and on the refractive indices n1 and n2 of reflection element and layer, respectively; if the thickness d of a layer is much smaller than the penetration depth dp of the infrared light (d ,, dp), the dependency is extended to the refractive index n3 of the liquid enclosing the layer, e.g., Riso of the amide I band in an aqueous environment like culture medium is 1.60. If d . dp, Riso equals 2 for all functional groups. Deviation of Rexp from Riso indicates a certain alignment of the considered functional group concerning the surface of the reflection element. If Rexp equals Riso, the integrated absorbances of both polarizations reflect the same information about the amount of a functional group but differ by the factor of Riso. Spectra recorded with vp light are not shown in this study, but the evaluation is considered in the relevant Rexp. Error in evaluation was estimated to be 65% in case of integrated absorbances .1 and 610% in case of integrated absorbances ,1. Solvent Compensation. FT-IR spectroscopy in the presence of solvents is faced with the following problem: in order to evaluate spectra of a sample, a reference spectrum in the absence of the sample has to be recorded. The advantage is that all device-specific parameters and influences of the surroundings are included in that reference and are thus eliminated from the calculated sample spectrum. However, introducing the sample into the solvent replaces a corresponding amount of the latter and causes an overcompensation of the solvent. This feature can be corrected by a scaled subtraction of the solvent spectrum from the sample spectrum. Recently, Marcsisin et al.18 presented a solvent compensation method for spectra measured in the transmittance mode. In this chapter we present a compensation method for ATR spectra measured with polarized IR light. The method can be applied to any sample layer, not only to a layer composed of cells and ECM, as investigated in this work. In ATR spectroscopy an absorption spectrum of a layer deposited on the surface of the reflection element is usually calculated by using an intensity spectrum recorded in the absence of the layer as a reference. Spectra measured in the presence of solvents provide adequate results if the evaluated spectral region of the layer is unaffected by the solvent. If there is interference, two cases have to be distinguished: (i) As long
42
as the layer is thin compared to the penetration depth dp of the IR beam into the solvent, for instance in case of a protein monolayer, the error resulting from solvent replacement may be neglected. (ii) Since films of cells embedded in ECM reach thicknesses in the micrometer range, imperfect solvent compensation caused by the aqueous medium should be considered, especially if a precise evaluation of the amide I band is required. Since according to Eq. 1 below, the evanescent field on the surface of the reflection element, which results from the totally reflected IR beam, penetrates into the rarer medium proportionally to the wavelength of the beam, basic ATR theory12 has to be applied for accurate compensation of the solvent. Thus, dp ¼
k qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2p n1 sin2 h n221
ð1Þ
where k denotes the wavelength of the IR beam, h is the angle of incidence of the beam into the reflection element (in our case 458) and n21 = n2/n1 denotes the ratio of the refractive indices outside (n2) and inside the reflection element (n1), respectively. As long as the solvent is weakly absorbing, i.e., the Napierian absorption index j of the complex refractive index ^ n = n þ ij fulfills the condition j , 0.1, analysis of spectra may be performed with the concept of effective thickness. The effective thickness de represents a hypothetical thickness for a transmission cell in order to result in the same absorbance as obtained per one internal reflection in the ATR experiment. A general expression of the effective thickness for a layer of thickness d of a weakly absorbing solvent ranging from zi to zf (d = zf zi) is given by Eq. 2: 2z 2z dp n2 i f e dp e dp ðE r02 Þ2 ð2Þ de ¼ n1 cosh 2 where ðE r02 Þ2 represents the relative electric field on the surface of the reflection element in case of either pp or vp IR incident light. It should be noted that such a simple expression for the effective thickness is only available for polarized light. During measurement of the ATR intensity spectrum of the solvent, which serves as a reference for calculation of the layer spectrum, the solvent is in direct contact with the reflection element. Now the thickness d can be considered as infinite with zi = 0 and zf = ‘. Therefore, for bulk medium, Eq. 2 results in debulk ¼
1 n 2 dp r 2 ðE Þ cosh n1 2 02
ð3Þ
From Eqs. 2 and 3 it follows that the corresponding absorption spectrum of the solvent is given by: mÞ Asolvent ð~mÞ ¼ N eð~mÞ c debulk ð~
ð4Þ
where A(~m) denotes absorbance, e(~m) is the molar (decadic) absorption coefficient, c denotes molarity, and N is the number of active internal reflections. After adhesion of cells to the substrate, they and the ECM will displace the solvent. Considering Eq. 2, the solvent starts now at zi = d and zf = ‘. Therefore, the effective thickness de of the solvent in the presence of a layer is described by de ¼
dp n2 2d e dp Er2 0 n1 cosh 2
ð5Þ
Volume 66, Number 1, 2012
//Xinet/production/a/apls/live_jobs/apls-66-01/apls-66-01-03/layouts/apls-66-01-03.3d Thursday, 1 December 2011 9:02 am Allen Press, Inc. Page 42
Equations 3 and 5 differ in the factor eð2d=dp Þ only. Therefore, the overcompensated absorption spectrum of the solvent Aovercomp(~ m) taking replacement by the layer into account is given by mÞ ¼ N eð~ mÞ c debulk ð~ mÞ e Aovercomp ð~ ¼ Asolvent ð~ mÞ e
d2dp
d2dp
ð6Þ
Subtracting Aovercomp from the reference spectrum of medium measured in the absence of cells results in the spectrum of the thin medium layer (spectrum B of Fig. 2). The latter can now be used for solvent compensation. In the present work, layer thickness d was adjusted by the operator in such a way that the resulting spectrum of the layer showed a reasonable compensation in the spectral region of the stretching vibration m of the aqueous medium at 3100–3400 cm1 (see spectrum C of Fig. 2). It should be noted that thickness d does not reflect the real thickness of the layer, but the amount of replaced solvent. Standard MATLAB software was used for calculation of solvent compensated spectra. For completeness, the following comments should be taken into account. The refractive indices used in these calculations were 4.0 for n1 (Ge) and 1.31 for n2 (culture medium). The latter was determined by averaging the refractive index of water in the spectral region of the amide I and amide II bands (1710–1480 cm1).19 Since medium can be viewed as a low concentration solution, the influence of dissolved substances on the refractive index of water was neglected. Further, the anomalous dispersion of water was not taken into account for calculating the penetration depth of the evanescent field.
FIG. 2. ATR spectra of MC3T3-E1 cells recorded with pp infrared light (A) before and (C) after solvent compensation, and (B) of culture medium. (A) MC3T3-E1 cells measured in the presence of medium after 6 h of seeding before the procedure of solvent compensation was applied. Denoted bands are assigned in the text with further explanations. The negative band in regions of water absorption (m(H2O): 3350 cm1) resulted from medium displacement by the cells in the sample cuvette of the flow-through chamber, while in the reference cuvette medium was in direct contact with the surface of the Ge MIRE. (B) ATR spectrum of medium scaled by visual inspection to achieve optimum water compensation of (A). The corresponding layer thickness, as ˚. calculated according to the procedure of solvent compensation, was 29 A Reference: air in empty reference cuvette. (C) ATR spectrum of MC3T3-E1 cells after solvent compensation. A perfect compensation of bands resulting from OH stretching and bending vibrations is not possible because of the presence of water bonded to protein molecules and free OH groups of amino acid residues as a part of the cells.
RESULTS An ATR spectrum of live MC3T3-E1 cells measured with pp incident infrared light after 6 hours of seeding is shown in Fig. 2A. The most prominent bands are the amide I and amide II bands (denoted as amide I & II bands) at 1649 cm1 and 1547 cm1, respectively, reflecting the protein components. Bands in the fingerprint region between 1200 cm1 and 900 cm1 belong mostly to C–O vibrations of polysaccharide components as well as to phosphate groups of the phosphodiester linkage present in DNA, RNA, phospholipids, etc., and to sulfate groups. Figure 2A also reveals a negative band in the region of water absorption at 3100–3400 cm1. This overcompensation results from the fact that the complete surface of the measuring element was covered by culture medium during the reference measurement before cell seeding, while in the presence of the MC3T3 cells the medium was partly displaced by the cell bodies. For correction of this feature, the procedure of solvent compensation (as described in the Materials and Methods subsection Solvent Compensation) was applied. Figure 2B represents the ATR spectrum of the medium, measured with pp incident IR light and scaled to a ˚ , which was found to result in optimum layer thickness of 29 A solvent compensation of the spectrum shown in Fig. 2A. It is dominated by water absorption bands because air was used as the reference. The sum of the spectra of Fig. 2A and Fig. 2B resulted in the solvent compensated spectrum, shown in Fig. ˚ reflects an average height of the 2C. The thickness of 29 A replaced medium layer because in between the cells the medium is still in contact with the surface. Moreover, water adsorbed to cellular molecules such as proteins and carbohydrates as well as free OH groups of amino acid residues and sugars contributed to these spectral regions. As a consequence, a perfect compensation of the OH stretching and bending vibration is not possible. The presence of MC3T3-E1 cells on the Ge plate after 14 days of permanent flow-through of medium is shown in Fig. 3A by crystal violet staining. The dark blue areas resulted from confluent cells near the chamber outlet. The condition of cells is presented in Fig. 3B by a live/dead staining. Cells reached confluence and principally exhibited a normal shape typical of healthy cells without accumulation of ethidium homodimer. Moreover, the abundance of calcein in the cytoplasm indicated that the cells were viable and metabolically active. Figure 4 represents the time course of ECM formation by MC3T3-E1 osteoblastic cells onto the Ge plate based on the sum of integrated absorbances (int. abs.) of the amide I and II bands. Evaluation of spectra was performed after the reference measurement representing medium in contact with the Ge plate was taken into account. Results of measurements with pp (filled circles) and vp (open circles) infrared light are shown. The original data of the latter were multiplied by Riso = 2 because this value represents the theoretical dichroic ratio of isotropically arranged functional groups inside a layer thicker than the penetration depth of the IR beam, a situation encountered in the case of deposited ECM and cells adhered to a surface. Therefore, this presentation format was chosen to demonstrate the deviation of experimentally determined data from theoretical Riso. Results of both polarizations of the experiment presented in Fig. 4 showed accordance within the estimated 65% error. The corresponding experimentally determined Rexp was in the range of 1.90 to 1.97 pointing to an isotropic arrangement of the amide bonds. Analysis of the
APPLIED SPECTROSCOPY
43
//Xinet/production/a/apls/live_jobs/apls-66-01/apls-66-01-03/layouts/apls-66-01-03.3d Thursday, 1 December 2011 9:02 am Allen Press, Inc. Page 43
FIG. 5. ATR spectra recorded with pp infrared light (A) 6 h, (B) 166 h, and (C) 334 h after seeding of MC3T3-E1 cells on the Ge reflection element. Reference: Medium in contact with the Ge plate. FIG. 3. Staining of MC3T3-E1 cells adhered on the Ge plate after 14 days of permanent flow-through of culture medium. (A) Crystal-violet staining. Dark blue areas represent confluent cells near the outlet of the chamber. (B) Live/ dead staining. Absence of ethidium homodimer in the nuclei and presence of calcein in the cytoplasm indicated that the cells were viable and metabolically active.
other experiments presented in this study exhibited only temporary deviations from the expected range, i.e., the first twelve hours in the second experiment with complete culture medium (1.80 , Rexp , 1.90), and the last five days in the experiment without ascorbic acid (1.85 , Rexp , 1.90). For that reason, only results of pp measurements are shown in the following evaluations of ECM deposition because these results already reflect the time course of protein formation. Spectra measured with pp light were also subjected to the procedure of solvent compensation as described previously to demonstrate
FIG. 4. Time course of ECM formation by MC3T3-E1 cells recorded with pp (filled circles) and vp (open circles) infrared light. The evaluation is based on the sum of integrated absorbances of the amide I and II bands (range of integration: 1702–1483 cm1). Results of the vp measurement were multiplied by Riso = 2. Accordance of data within the error limits pointed to an isotropic arrangement of the amide bonds. The procedure of solvent compensation was applied on the results of the pp measurement (filled squares) to demonstrate the effect on integrated absorbances of the amide I and II bands.
44
the effect on the integrated absorbance of the amide I and II bands (filled squares). As can be clearly seen, the procedure causes a shift up to 13% in the absolute value. Attenuated total reflection spectra representing MC3T3-E1 cells and ECM after 6 h (A), 166 h (B), and 334 h (C) of seeding of cells on the reflection element are shown in Fig. 5. They are the results of spectral evaluation without applying the procedure of solvent compensation. To verify that the organic matrix monitored by these experiments is due to ECM produced by osteoblasts rather than non-specific adsorption of culture medium proteins onto the Ge plate, two experiments were run at identical conditions showing the reproducibility of the system. Results are shown in Fig. 6. In one of the two experiments (open circles), at an arbitrarily chosen time point the culture medium was replaced by ethanol (70% in water) for 5 min in order to kill all living cells. Following this, ethanol was replaced by culture medium, which was pumped through
FIG. 6. Time course of ECM formation by MC3T3-E1 cells resulting from two separate experiments. In one experiment (open circles), cells were killed by replacing culture medium with ethanol after 357 h incubation time followed by pumping of culture medium through the chamber for another three days to check the influence of flowing medium on integrated absorbances of amide I and II. The sudden increase of the amide I and II bands after contact with ethanol is due to altered physical-chemical properties of the ECM.
Volume 66, Number 1, 2012
//Xinet/production/a/apls/live_jobs/apls-66-01/apls-66-01-03/layouts/apls-66-01-03.3d Thursday, 1 December 2011 9:02 am Allen Press, Inc. Page 44
FIG. 7. Time course of ECM formation by MC3T3-E1 cells in the presence (open and filled circles) and absence (triangles) of ascorbic acid in the culture medium. The lack of ascorbic acid resulted in a reduced production of protein components in the ECM.
the chamber for another three days applying the same conditions as in the presence of living cells, so as to determine the influence of flowing medium on the integrated absorbance of the amide bands. In a second experiment cells grown on the Ge plate for 14 days during permanent flow-through of medium were detached by pronase/EDTA solution, and the number was determined by cell counting, resulting in a 67% increase compared to the number seeded. To confirm that cells anchored onto the Ge plate respond in a similar fashion with their counterparts under typical cell culture conditions, an experiment was performed without ascorbic acid in the culture medium and compared with the two previous experiments run at control conditions, i.e., in the presence of ascorbic acid.14 Results are shown in Fig. 7, in which it may be seen that in the absence of ascorbic acid (filled triangles) there was less ECM produced (monitored through evaluation of the amide I and II bands) compared to the control cultures (filled and open circles).
DISCUSSION Growth and differentiation of cells are mediated through signals from the local environment. In addition to growth and differentiation factors as well as hormones, proper cell–matrix interactions play an important role directing differentiation and allowing survival of cells by preventing cell death.20–22 The extracellular matrix (ECM), beyond signaling to cells, provides structural support for organs and tissues, and, in the case of bone, it is the main component of tissue itself. ECM composition, by impacting sub- and supra-molecular structure, controls development of functional bone tissue. This is achieved in several ways on several structural levels: collagen type I, the main component of bone tissue, regulates cell multiplication, drives differentiation, and also prevents apoptosis by epigenetic mechanisms.22 At this level, primary protein structure (amino acid (AA) sequence), i.e., the RGDtripeptide sequence, via integrins, conveys intracellular signals.23 Changes in AA-sequence lead to distortions of the triple helical structure resulting in severe malformation of bone tissue as observed in osteogenesis imperfecta.24 Additionally, proper triple helix formation as well as formation of the supra-
molecular structure of bone matrix depend on coordinated intra- and inter-molecular cross-linking of collagen.25 The fact that proper collagen cross-linking directly affects osteoblastic proliferation and differentiation was recently demonstrated in an in vitro culture system of mouse osteoblasts.26–28 It becomes evident then that coordinated formation of the structure of the bone matrix is a dynamic process and its understanding requires the study of its temporal events. In the present study we developed a method to spectroscopically monitor the formation of ECM in real time, with no extra processing steps involved such as cell or ECM separation that may introduce artifacts. Moreover, the cells are grown in a dynamic environment (culture media is flowing rather than static), a situation much closer to the in vivo one compared to the classical culture dish experiments. The produced ECM may be analyzed either quantitatively (through integration of various characteristic bands as demonstrated in the present work), or in the future, in a qualitative manner through further analysis of spectral bands to derive outcomes pertaining to collagen cross-links and protein structure.29 Whereas in transmission spectroscopy the sample is penetrated totally by the light beam, in the case of ATR spectroscopy the beam is reflected at the surface where cells attach. Therefore, the signal is influenced by culture medium and, to a lesser extent, by the cellular body itself and to a greater extent by the ECM produced by osteoblasts. For evaluation, two aspects have to be considered: (i) Deposited ECM and cells adhering on the surface of the reflection element possess a thickness larger than the penetration depth of the IR beam (; 0.8 lm). (ii) Because of the high number of diverse proteins with differentially orientated secondary structures, isotropic arrangement of amide bonds is to be expected. In such a system the experimentally determined dichroic ratio Rexp of amide I and II equals to Riso = 2, as was also the case in our experiments (1.90 , Rexp , 2.01) taking into account error limits of 5%. Temporary deviations from Riso were more a result of unequal cell distribution on the reflection element rather than of oriented amide bonds. Furthermore, we herein present a procedure denoted as solvent compensation, enabling correction of spectra measured in the presence of cells and deposited ECM. In the case of evaluating the amide I band, such a correction is necessary because the reference measurement recorded at the onset of the experiment cannot account for the growing layer of proliferating cells. The procedure increases the absolute value of integrated absorbances due to the contribution of the water bending vibration of the thin medium layer replaced by cells to the spectral region of the amide I band. Since in this study only time courses of different experiments were compared, the effect of solvent compensation was not taken into account for further evaluation, but it should be noted that results of band shape analysis of the amide I band are affected to a significant extent by applying this procedure. Results of the present study show that MC3T3-E1 osteoblasts can be cultured on germanium prisms and remain alive, as shown by crystal-violet and live/dead staining, and proliferate, as determined by cell counting. Interestingly, cell density is higher at the outlet of the culturing chamber compared to the inlet, most likely due to differential local pressure (since culture medium is flowing over the cells). This aspect will be the focus of future studies employing the setup
APPLIED SPECTROSCOPY
45
//Xinet/production/a/apls/live_jobs/apls-66-01/apls-66-01-03/layouts/apls-66-01-03.3d Thursday, 1 December 2011 9:02 am Allen Press, Inc. Page 45
presented herein, since cells are known to respond to hydrodynamics and mechanical stress.30–41 Continued supply with fresh culture medium as well as removal of metabolites of cells cultured on germanium prisms under continuous flow introduces the possibility of studying matrix deposition with reduced autocrine and paracrine effects on cells and curtailed influences of their own metabolites; likewise, pulsing effects of frequent medium changes are prevented. Furthermore, adsorption of proteins or drugs on the germanium prisms enables studying effects on cell adhesion and early influences on growth, function, and differentiation of cells. The applicability of fluorescence techniques on the germanium prisms enables immunologic protein analysis and in situ hybridization for mRNA expression. Moreover, the size of the prisms allows sufficient material to be produced for genome-wide expression analysis, quantitative reversed transcribed PCR (polymerase chain reaction)42 and protein analysis techniques, e.g., matrix-assisted laser desorption ionization– time of flight (MALDI-TOF) mass spectrometry.43 To further ensure that the organic matrix monitored in these experiments was due to the ECM synthesized by osteoblasts rather than merely physical-chemical adsorption of proteins present in the culture media, cells were killed using alcohol during the flow-through experiment, followed by re-introduction of culture medium. Once alcohol was introduced, the increasing trend in the integrated absorbance of the amide I and II bands ceased, and re-introduction of culture medium did not result in an increase thereafter, strongly suggesting that the organic matrix monitored through the amide I and II bands was due to ECM synthesized by living cells rather than random protein adsorption. The fact that after alcohol introduction the integrated absorbances were higher in value is most likely due to the fact that alcohol displaces water, resulting in an ECM smaller and denser than before, thus increasing the protein (amide I and II) signal. Comparison of results indicated that MC3T3-E1 cells effectively produced ECM during experiments in the flow-through chamber as well as that omission of ascorbic acid in the culture medium, a vitamin responsible for proper collagen synthesis,44 reduced ATR signals of the ECM. The introduction of alcohol indicated that the amide I and II signals monitored are mainly due to active deposition of ECM by osteoblasts as opposed to physical-chemical adsorption of organic moieties in the culture media, while experiments involving inclusion or exclusion of ascorbic acid indicate that cells in the flow-through environment react like their counterparts in classical cell culture static conditions, where a factor regulating matrix production and deposition is present or absent. No direct comparison between dynamic and static cell culture conditions was attempted, as spectroscopic analysis in the latter case involves killing of cells, scraping off the matrix in cell culture dishes, and transferring them to BaF2 windows,45 a procedure that may lead to incomplete transfer of the entire amount of matrix produced. In conclusion, the results of the present study demonstrate the application of ATR spectroscopy for real-time monitoring of matrix formation by live osteoblasts. The identification of absorption bands specific for, i.e., collagen cross-links25 offers the possibility of studying drug effects on the dynamics of ECM formation. Moreover, it is feasible that matrix degrada-
46
tion can be studied as well. All these possibilities make ATR a promising tool to study dynamic cellular processes. ACKNOWLEDGMENTS This study was supported by the Fonds zur Foederung der wissenschaftlichen Forschung (FWF; The Austrian Science Fund) Project P20646-B11, the WGKK (Vienna Health Insurance Fund), the AUVA (Austrian Workers’ Compensation Board), and a research grant by The Alliance for Better Bone Health.
1. P. J. Marie, Osteoporosis Int. 20, 1037 (2009). 2. R. Zaidel-Bar, S. Itzkovitz, A. Ma’ayan, R. Iyengar, and B. Geiger, Nat. Cell Biol. 9, 858 (2007). 3. R. O. Hynes, Science 326, 1216 (2009). 4. A. L. Berrier and K. M. Yamada, J. Cell. Physiol. 213, 565 (2007). 5. D. E. Discher, D. J. Mooney, and P. W. Zandstra, Science 324, 1673 (2009). 6. B. Geiger, J. P. Spatz, and A. D. Bershadsky, Nat. Rev. Mol. Cell Biol. 10, 21 (2009). 7. R. O. Hynes, Cell 110, 673 (2002). 8. K. R. Legate, S. A. Wickstrom, and R. Fassler, Genes Dev. 23, 397 (2009). 9. J. L. R. Arrondo, A. Muga, J. Castresana, and F. M. Goni, Prog. Biophys. Mol. Biol. 59, 23 (1993). 10. G. Reiter, M. Siam, D. Falkenhagen, W. Gollneritsch, D. Baurecht, and U. P. Fringeli, Langmuir 18, 5761 (2002). 11. G. Reiter, N. Hassler, V. Weber, D. Falkenhagen, and U. P. Fringeli, Biochim. Biophys. Acta 1699, 253 (2004). 12. U. P. Fringeli, ‘‘In situ infrared attenuated total reflection membrane spectroscopy’’ in Internal Reflection Spectroscopy, F. M. Mirabella, Ed. (Marcel Dekker, New York, 1992), Chap. 10, p. 255. 13. U. P. Fringeli, D. Baurecht, T. Bu¨rgi, M. Siam, G. Reiter, M. Schwarzott, and P. Bru¨esch, ‘‘ATR Spectroscopy of Thin Films’’ in Handbook of Thin Film Materials, H. S. Nalwa, Ed. (Academic Press, New York, 2002), Vol. 2, Chap. 4, p. 191. 14. R. T. Franceschi, B. S. Iyer, and Y. Cui, J. Bone Miner. Res. 9, 843 (1994). 15. T. Matsumoto, C. Igarashi, Y. Takeuchi, S. Harada, T. Kikuchi, H. Yamato, and E. Ogata, Bone 12, 27 (1991). 16. N. Fratzl-Zelman, P. Fratzl, H. Horandner, B. Grabner, F. Varga, A. Ellinger, and K. Klaushofer, Bone 23, 511 (1998). 17. L. Rieppo, S. Saarakkala, T. Na¨rhi, J. Holopainen, M. Lammi, H. J. Helminen, J. S. Jurvelin, and J. Rieppo, Microsc. Res. Technol. 73, 503 (2010). 18. E. J. S. Marcsisin, C. M. Uttero, M. Miljkovic, and M. Diem, Analyst 135, 3227 (2010). 19. H. D. Downing and W. Dudley, J. Geophys. Res. 80, 1656 (1975). 20. C. Streuli, Curr. Opin. Cell Biol. 11, 634 (1999). 21. S. M. Frisch and R. A. Screaton, Curr. Opin. Cell Biol. 13, 555 (2001). 22. R. Thaler, H. Karlic, S. Spitzer, K. Klaushofer, and F. Varga, Apoptosis 15, 728 (2010). 23. B. Ganss, R. H. Kim, and J. Sodek, Crit. Rev. Oral Biol. Medicine 10, 79 (1999). 24. D. Basel and R. D. Steiner, Genet. Med. 11, 375 (2009). 25. E. P. Paschalis, K. Verdelis, S. B. Doty, A. L. Boskey, R. Mendelsohn, and M. Yamauchi, J. Bone Miner. Res. 16, 1821 (2001). 26. C. Turecek, N. Fratzl-Zelman, M. Rumpler, B. Buchinger, S. Spitzer, R. Zoehrer, E. Durchschlag, K. Klaushofer, E. P. Paschalis, and F. Varga, Calcif. Tissue Int. 82, 392 (2008). 27. R. Thaler, S. Spitzer, M. Rumpler, N. Fratzl-Zelman, K. Klaushofer, E. P. Paschalis, and F. Varga, Bone 46, 703 (2010). 28. H. Fernandes, K. Dechering, E. Van Someren, I. Steeghs, M. Apotheker, A. Leusink, R. Bank, K. Janeczek, C. Van Blitterswijk, and J. de Boer, Tissue Eng., Part A 15, 3857 (2009). 29. E. P. Paschalis, Osteoporos Int 20, 1043 (2009). 30. E. A. Botchwey, S. R. Pollack, S. El-Amin, E. M. Levine, R. S. Tuan, and C. T. Laurencin, Biorheology 40, 299 (2003). 31. D. Du, K. S. Furukawa, and T. Ushida, Biotechnol. Bioeng. 102, 1670 (2009). 32. S. S. Kohles, N. Neve, J. D. Zimmerman, and D. C. Tretheway, J. Biomech. Eng. 131, 121006 (2009). 33. P. A. Kokkinos, I. K. Zarkadis, T. T. Panidis, and D. D. Deligianni, J. Mater. Sci. Mater. Med. 20, 655 (2009). 34. M. R. Kreke and A. S. Goldstein, Tissue Eng. 10, 780 (2004). 35. K. Mukundakrishnan, P. S. Ayyaswamy, M. Risbud, H. H. Hu, and I. M. Shapiro, Ann. N. Y. Acad. Sci. 1027, 85 (2004).
Volume 66, Number 1, 2012
//Xinet/production/a/apls/live_jobs/apls-66-01/apls-66-01-03/layouts/apls-66-01-03.3d Thursday, 1 December 2011 9:02 am Allen Press, Inc. Page 46
36. L. A. Sharp, Y. W. Lee, and A. S. Goldstein, Ann. Biomed. Eng. 37, 445 (2009). 37. X. F. Walboomers, S. E. Elder, J. D. Bumgardner, and J. A. Jansen, J. Biomed. Mater. Res., Part A 76, 16 (2006). 38. S. Weinbaum, P. Guo, and L. You, Biorheology 38, 119 (2001). 39. F. Zhao, R. Chella, and T. Ma, Biotechnol. Bioeng. 96, 584 (2007). 40. F. Zhao, W. L. Grayson, T. Ma, and A. Irsigler, J. Cell. Physiol. 219, 421 (2009). 41. B. Grabner, F. Varga, N. Fratzl-Zelman, E. Luegmayr, H. Glantschnig, M.
42. 43. 44. 45.
Rumpler, A. Tatschl, P. Fratzl, and K. Klaushofer, Rev. Sci. Instrum. 71, 3522 (2000). L. K. McCauley, A. J. Koh, C. A. Beecher, Y. Cui, J. D. Decker, and R. T. Franceschi, J. Bone Miner. Res. 10, 1243 (1995). V. Egelhofer, J. Gobom, H. Seitz, P. Giavalisco, H. Lehrach, and E. Nordhoff, Anal. Chem. 74, 1760 (2002). B. Peterkofsky, Am. J. Clin. Nutr. 54, 1135S (1991). Y. D. Halvorsen, D. Franklin, A. L. Bond, D. C. Hitt, C. Auchter, A. L. Boskey, E. P. Paschalis, W. O. Wilkison, and J. M. Gimble, Tissue Eng. 7, 729 (2001).
APPLIED SPECTROSCOPY
47
//Xinet/production/a/apls/live_jobs/apls-66-01/apls-66-01-03/layouts/apls-66-01-03.3d Thursday, 1 December 2011 9:02 am Allen Press, Inc. Page 47
