Acta Biomaterialia 7 (2011) 515–525

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /actabiomat

Fetal bovine serum xenoproteins modulate human monocyte adhesion andprotein release on biomaterials in vitro

David Schmidt a,b, Evan James Joyce b, Weiyuan John Kao a,b,c,⇑a School of Pharmacy, University of Wisconsin-Madison, Madison, WI 53705, USAb Department of Biomedical Engineering, College of Engineering, University of Wisconsin-Madison, Madison, WI 53705, USAc Department of Surgery, School of Medicine and Public Health, University of Wisconsin-Madison, Madison, WI 53705, USA

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

Article history:Received 7 June 2010Received in revised form 25 August 2010Accepted 25 August 2010Available online 15 September 2010

Keywords:MonocyteMacrophageFetal bovine serumPolyethylene glycolPDMS

1742-7061/$ - see front matter � 2010 Acta Materialdoi:10.1016/j.actbio.2010.08.022

⇑ Corresponding author at: School of Pharmacy, Unson, Madison, WI 53705, USA. Tel.: +1 6082632998; f

E-mail address: wjkao@wisc.edu (W.J. Kao).

Monocyte-derived macrophages are critical in the host–foreign body response to biomaterials and havebeen studied extensively in various culture conditions in vitro, such as medium supplemented with fetalbovine serum (FBS) or autologous human serum (AHS). Since monocyte maturation into macrophages ishighly plastic and may vary considerably depending on the surface, isolation procedures and in vitro cul-ture conditions, we hypothesize that variations in protein adsorption and serum type will greatly impactmonocyte behavior in a surface-dependent manner. The impact of xenoproteins on monocyte-surfaceinteractions has not been well studied methodically and the use of AHS rather than FBS for macro-phage-biomaterials studies in vitro is far from universal. The commonly used reference materials – tissueculture polystyrene (TCPS), polyethylene glycol (PEG) and polydimethylsiloxane (PDMS) – wereemployed in this study and we found a 3-fold higher adherent monocyte density on TCPS when AHSwas used vs. FBS-supplemented medium. On PEG hydrogels, an 8- to 10-fold higher adhesion densitywas observed when AHS was employed vs. FBS, while on PDMS no difference in adhesion density wasobserved between the two sera conditions. Additionally, the presence of lipopolysaccharide abrogatedthe serum-dependent effect on cell adhesion on TCPS. Significantly different variations in protein releasewere observed between the serum conditions on these surfaces; in particular, there was a 100-fold higherconcentration of growth-related oncogene for the AHS condition on PDMS even though the adhesion lev-els were comparable between the two serum conditions. These results emphasize the combined impactof the surface type and FBS xenoproteins in mediating the observed monocyte response to biomaterialsin vitro.

� 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Monocyte-derived macrophages function in multiple phases ofthe host response, promoting both pro-inflammatory and anti-inflammatory activities through the release of numerous chemicalmediators and influencing the adaptive immune response throughthe presentation of antigens [1]. Thus, monocyte-derived macro-phages are a significant cell type in the investigation of material-mediated inflammation and wound healing. Multiple differencesbetween transformed macrophage cell lines and primary macro-phages matured from peripheral blood monocytes have beenobserved; consequently, primary macrophages are likely moreclinically relevant to biomaterials research [2–4]. Monocytes maybe isolated via a variety of means, including flow cytometry, den-sity gradients, and surface adhesion; however, due to the plasticity

ia Inc. Published by Elsevier Ltd. A

iversity of Wisconsin-Madi-ax: +1 6082625345.

of monocytes, each of these methods may lead to variation in thefinal isolated monocyte phenotype [5,6]. Much like the plasticityobserved in vivo with organ-specialized monocyte-derived macro-phages, variable protein adsorption to biomaterial surfaces maylead to surface-specific macrophage phenotypes [7,8].

Similarly to the effects of the surface on monocyte maturationinto macrophages, culture conditions can dramatically affect thematuration process. The phenotype can be modified in vitro byexposure to various cytokines or exogenous stimuli such as lipo-polysaccharide (LPS) during the maturation process [9–11]. Bothfetal bovine serum (FBS) and autologous human serum (AHS) areused as additions to in vitro culture conditions in biomaterials re-search [7,12–15]. The role of AHS and FBS in modifying monocytematuration (via surface antigens and protein release) has beenstudied extensively in suspended monocyte-derived dendritic cells(DC) for clinical applications [16–20]; however, the impact of FBSxenoproteins on human non-DC monocytes, and more specificallyon biomaterial–monocyte interactions, has not been well studied[21,22]. Depending on the surface properties, biomaterials

ll rights reserved.

516 D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525

promote differential adsorption of certain serum proteins, and thuswe hypothesized that the effect of serum type may vary by surface[1]. For example, polyethylene glycol (PEG) hydrogels limit adsorp-tion of common serum proteins but may sustain significant com-plement component 3 (C3) activity comparable to other surfaces,potentially allowing this protein to take on a more influential rolein mediating monocyte/macrophage responses [23]. Proteins inhigh abundance in FBS include: bovine serum albumin, alpha-1-antiproteinase, plasminogen, lactoperoxidase, kniogen, alpha-2-HS-glycoprotein, hemiferrin, prothrombin, apolipoprotein A-I,integrin b1, IGFBP2, insulin-like growth factor (IGF) II, transform-ing growth factor (TGF)-b1 [24]. Proteins in high abundance inAHS include: albumin, complement factor H, angiotensinogenand prostate-specific antigen [25]. Thus, it is likely that the surfaceprotein concentration and identity would differ significantly be-tween AHS and FBS. Furthermore, there is a lack of extensive pro-teomic analysis of AHS and FBS with a specific focus on adhesionproteins for monocytes.

In this study, human blood-derived monocytes were obtainedfrom multiple donors and cultured in FBS- or AHS-supplementedmedium on PEG hydrogels, tissue culture polystyrene (TCPS) andpolydimethylsiloxane (PDMS). TCPS and PDMS were selected be-cause they are both commonly used reference materials with differ-ent surface properties, and PEG hydrogels were selected becausethey have been shown to support monocyte adhesion while limitingprotein adsorption along with having substantially differing surfaceproperties from TCPS and PDMS [1,15,23,26,27]. PEG hydrogels arehighly hydrophilic swollen polymer networks, PDMS disks arehighly hydrophobic films and TCPS is a rigid, slightly hydrophobicsurface. Exogenous LPS was employed as a positive control for theinnately activated highly pro-inflammatory macrophage matura-tion state, which bears similarities to the classically activated M1macrophage cell type but does not require concurrent addition ofinterferon-gamma (IFN-c) [28]. To initially assess the extent of cellactivation, selected critical proteins were quantified: the broadlyacting inflammatory cytokines interleukin (IL)-1b, IL-1a and tumornecrosis factor alpha (TNF-a); growth factors, including granulocytemacrophage-colony stimulating factor (GM-CSF) and TGF-a; andchemokines, including induced protein-10 (IP-10) and monocytechemoattractant protein-1 (MCP-1). To further assess the polariza-tion state of the adherent cells, both surface CD86 and CD163markers and soluble pro-inflammatory and anti-inflammatoryprotein markers were quantified: interleukin-12 (IL-12(p40) andIL-12(p70)), growth-related oncogene-2 (GRO-2 or CXCL2), TNF-a,IFN-c, IL-10 and IL-1b [28]. These molecules are among the manybiomarkers for the pro-inflammatory M1 and more diverse/anti-inflammatory M2 states and thus may be used for comparison tothese molecularly induced states [28]. The alternatively activatedM2 phenotype is characterized by high levels of the anti-inflamma-tory signaling factors IL-1 receptor antagonist IL-10 and IGF-1. Whileadhering to strict human subject safety standards, multiple donorswere used for each part of this study to garner statistical significancesince large donor variability has been reported when using primarymonocytes [29–33].

2. Materials and methods

2.1. Material preparation

PEG diacrylate (PEGdA) was synthesized from PEG-diol (MW3400; Sigma) following previously established methods [15].Briefly, PEG-diol was dissolved in dried THF at a ratio of PEG:THFratio of 1 g:5 ml. Acryloyl chloride (ACl) and triethylamine (TEA)were added at a PEG:ACl:TEA molar ratio of 1:4:6, stirred in thedark for 3 h, filtered twice, precipitated dropwise in a large volume

of cold hexane (at least 100 ml of hexane per 1 g of starting PEG-diol), filtered and dried in a vacuum oven. The PEGdA productwas characterized via high-performance liquid chromatographywith purity typically >97%. To obtain PEG hydrogels, a mixture of10 wt.% PEGdA and 0.1 wt.% ultraviolet (UV) photoinitiator(Irgacure 2959�) was dissolved in phosphate-buffered saline(PBS) at 60 �C. The heated solution was pipetted into a polytetra-fluoroethylene mold (7 mm diameter and 0.75 mm thickness) thenclamped between two glass coverslips that were first cleaned in asulfuric acid and nitric acid wash (3:1 respectively; 1 N:1 N). Theentire assembly was placed under a UV lamp (UVP Model B100AP) and polymerized for 3 min per side to form the PEGhydrogel. Samples were cold sterilized twice for 30 min with 70%ethanol and stored in sterile PBS for at least 24 h for equilibration.Immediately before cell culture, samples were washed at least fivetimes in sterile PBS to remove any residual traces of ethanol.

Medical grade PDMS (Specialty Manufacturing Inc.; non-rein-forced vulcanized gloss/gloss), �0.76 mm thick, was punched into7.9 mm diameter disks using a die and hammer punch. This diam-eter was chosen for consistency with the increase in size of the PEGhydrogels due to swelling. The PEG hydrogel diameter was mea-sured after 24 h of swelling in PBS and found to be 7.9 mm, whileno swelling was observed for the PDMS disks. PDMS disks weresonicated and vortexed in a 0.5% solution of Triton X-100 for clean-ing, then cold sterilized in 70% ethanol twice for 30 min. The PDMSdisks were stored in sterile PBS for at least 24 h for equilibrationand washed at least five times in sterile PBS prior to use. Forty-eight-well culture plates were used as the TCPS surface (BD Fal-con). TCPS well plates were kept at room temperature with sterilePBS prior to use in culture. Both the PEG hydrogels and the PDMSdisks were placed in 48-well culture plates prior to cell seeding.

2.2. Monocyte culture and analyses

Human subject protocol guidelines were followed for blood col-lection; however, these guidelines strictly limited the amount ofblood that could be safely taken and severely limited the numberof different assays that could be performed. Consequently, thesame culture wells were tracked continuously for the length ofthese studies rather than stopped at each time point. Three donorswere used for each of the two major parts of this study; however,as the same three donors were not available for both parts, no sta-tistical comparisons are made between the two parts. Neverthe-less, the high level of similarity in the results between the twodonor groups suggests good reproducibility of these findings.

In the initial part of the study we evaluated the monocyte/mac-rophage response to AHS/FBS on PEG and TCPS. Primary monocyteswere isolated from the citrated whole blood of three medication-free healthy adult human volunteers via a well-established non-adherent density gradient procedure [14,15]. At the time of eachdonation, additional non-citrated blood was collected to obtainAHS. Briefly, the blood was incubated at 37 �C for 3 h and thencentrifuged. Serum was collected from the top of the tube awayfrom the thrombus and care was taken to avoid collecting anyred blood cells or blood clots. The AHS was immediately dividedinto smaller volumes for replenishment of the cell culture at everytime point and frozen at �20 �C. Immediately after isolation, cellswere resuspended in either 10% FBS (Atlanta Biologicals, triple0.1 lm filtered, not heat inactivated) or 10% AHS (not heat inacti-vated) in RPMI 1640 medium, then statically seeded in 48-wellTCPS plates with or without the PEG hydrogels at a concentrationof 1 � 106 cells ml�1. Cell cultures were maintained at 5% CO2

and 37 �C. At each culture time point (2 and 24 h and 4, 7 and14 days), culture supernatant was collected for protein analysis(described below). Following supernatant collection, the cells werewashed twice and the medium was replenished. Between the 7 and

D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525 517

14 day time points, the cells were washed twice and the mediumreplenished every 2 days. The cultures were continued for 14 days,and thus the images used for calculating adhesion densities fol-lowed the same wells for the entire 14 days.

The wells were imaged using an inverted light microscope(Nikon, Eclipse TE300) coupled to a computer-assisted videoanalysis system (MetaMorph v4.1) at 2 and 24 h and 4, 7 and14 days. A minimum of three images were taken of each well at amagnification of 20� (0.36 mm2), from which the adherent celldensity was determined. Protein analysis was performed on the cul-ture supernatant via a microsphere-based multiplex assay (Bio-Rad,Millipore). Three samples were analyzed per donor at each timepoint. The microsphere assays were performed according to themanufacturers’ instructions for all analytes: IL-1b, IL-1a, TNF-a,GM-CSF, TGF-a, IP-10 and MCP-1 (Millipore Millplex Map). Standardcurves were determined with 10% FBS and 10% AHS for each donor toensure elimination of the background presence of the protein targetsin the serum. The standard curve for the microsphere assay utilizedstandards up to 1 � 105 pg ml�1 and was a five parameter logisticregression curve that accounts for asymmetry in the low and highconcentration ends of the assay. Thus, it is S-shaped with linearityin the mid concentration levels and non-linearity at the high andlow ends of the curve. All extrapolated data is above the higheststandard concentration of 1 � 105 pg ml�1, but is still below theupper asymptote of the regression curve equation. While thesuppliers’ technical documents make no claim on a measure ofconfidence for extrapolated values below the maximum asymptoteof the curve, the confidence of those values is in line with theconfidence of the interpolated data (within the standard range).In this study, all values above 2 � 105 pg ml�1 are reported as2 � 105 pg ml�1 to limit reliance on extrapolated data beyond twotimes the upper limit of the highest standard concentration.

Following the differential adhesion results on PEG and TCPS, weexpanded the study to include an additional surface and used differ-ent biomarkers to better characterize differences in the macrophageresponses to serum type and surface. Further, LPS was used as apositive control for pro-inflammatory innately activated macro-phages. Cells were isolated as described above and suspended in10% AHS in RPMI 1640, 10% FBS in RPMI 1640, 10% AHS with100 ng ml�1 LPS or 10% FBS with 100 ng ml�1 LPS. The suspendedcells were then statically seeded at (1 � 106 cells ml�1) on TCPS,PDMS and PEG hydrogels. The cells were imaged as in part one at2 and 24 h and 4 and 7 days. Surfaces were also washed and themedium replenished as before. Supernatant was collected at eachtime point to quantify IL-1b, IL-10, IL-12(p40), IL-12(p70), TNF-a,GRO-2 and IFN-c as described above. At day 7 after imaging andsupernatant collection, the adherent cells were fixed in 1% parafor-maldehyde (PFA; Polysciences, Inc.) and stained with antibodiesdirected against CD86 conjugated to Alexa Fluor 488 (Biolegend)and CD163 conjugated to Alexa Fluor 647 (Biolegend) at themanufacturer’s recommended dilution. Cells were imaged viafluorescent confocal microscope (Bio-Rad MRC-1024, Kr/Ar: 41, iris:4.0, gain: 1284, offset: -5.0, room temperature, 0.30 numerical aper-ture for 10� and 0.50 for 20�, immersion medium: air, Lasersharp5.2 software). Autofluorescence was minimal, as was backgroundstaining with control antibodies (BioLegend). The brightness of theimages was enhanced to improve image quality; however, no quan-titative conclusions were drawn based on differential brightness.Autofluorescence and background staining with control antibodieswere found to be negligible and these black images are not shown.

2.3. Statistical analysis

Cell density and protein concentration data were analyzed viatwo-way analysis of variance and Tukey’s post-testing (SigmaStatv2.03), with p values <0.05 considered as significant.

3. Results

3.1. Initial monocyte culture with AHS or FBS on PEG or TCPS

The morphologies of the adherent cells were very different,with enhanced spreading observed for the AHS-supplemented cul-tures on both TCPS and PEG (Fig. 1). When cultured in AHS but notwith FBS, the cells took on more of a spindle shape on TCPS, incontrast to a more rounded shape on PEG. This difference in shapebecame evident at the 4 day time point and beyond, but was notobservable at 24 h. The adherent cell density was vastly differenton TCPS and on the PEG hydrogel when either AHS- or FBS-supple-mented medium was used. For instance, at 2 h on TCPS, the adher-ent cell density was 2- to 3-fold higher for the AHS condition thanfor the FBS condition, while on the PEG hydrogels the differencewas nearly 10-fold higher (Fig. 2). This trend continued at 24 hfor both surfaces and remained consistent and statistically signifi-cant on PEG up to day 7, after which there was no longer a statis-tical difference in adhesion between the FBS and AHS conditions.IL-1b and TNF-a concentrations were significantly different be-tween the AHS and FBS conditions in the presence of PEG hydro-gels, particularly at 24 h (Fig. 3), while no difference wasobserved for IL-1a, GM-CSF, TGF-a or IP-10 (not shown). The che-mokine MCP-1 was generally high for both surfaces and serumconditions from 24 h on (Fig. 4). This is consistent with other stud-ies reported where TCPS produced increases in pro-inflammatorymarkers over PDMS [44]. Taken together, these results indicatethat selected monocyte/macrophage responses (i.e. adhesion,IL-1b and TNF-a release) were modulated by both the serum pro-teins present and the substrate, while other monocyte/macrophageresponses (i.e. GM-CSF release) were not.

3.2. Monocyte maturation modulated by material and xenoproteins

Primary monocytes were statically seeded on TCPS with orwithout exogenous LPS, PDMS and PEG hydrogels using AHS- orFBS-supplement medium. At 2 and 24 h, the adherent cell densityon TCPS was significantly higher when cultured with AHS thanwhen cultured with FBS (Fig. 5). Further, for the entire 7 day cul-ture, the adherent monocyte density on PEG was significantlyhigher (approximately 10-fold) when cultured with AHS thanwhen cultured with FBS. The adherent monocyte density on PEGfor the AHS condition was also higher than that on PDMS or TCPSwith or without LPS at 4 days. In contrast to TCPS without LPS andPEG, monocyte adhesion to PDMS and TCPS cultured with LPS didnot differ between the two serum conditions for the entire 7 dayculture. The morphology of the adherent monocytes on the varioussurfaces differed depending on the serum condition (Fig. 1). Great-er cell spreading was observed for PEG and TCPS surfaces when theculture medium was supplemented with AHS instead of FBS, butnot for PDMS and TCPS cultured with LPS. Adherent monocyteswere stained for CD86 and CD163 at 7 days (Fig. 6) when a maturemacrophage phenotype would be expected [34]. Monocytes/mac-rophages adherent on all surfaces that were cultured with eitherAHS- or FBS-supplementation stained positive for CD86; in con-trast, little to no staining was observed for CD163, a surface anti-gen involved in endocytosis [28].

Various protein expressions showed dependency on serum typeand surface type (Fig. 7). Specifically, the GRO-2 concentration wassignificantly higher by nearly 100-fold at 24 h for monocytes adher-ent on PDMS with AHS-supplemented medium compared to thosewith FBS. The GRO-2 concentration was also significantly higherfor monocytes adherent on PEG with AHS-supplemented mediumcompared to FBS at 24 h and 4 days. The IFN-c concentration formonocytes cultured with FBS was significantly higher than that with

Fig. 1. Light microscopy photomicrograph of adherent cells on TCPS and PEG hydrogels: (20�magnification) 4 days in 10% FBS-supplemented medium on TCPS (A), 4 days in10% AHS-supplemented medium on TCPS (B), 4 days in 10% FBS-supplemented medium on PEG hydrogel (C), 4 days in 10% AHS-supplemented medium on PEG hydrogel (D),24 h in 10% FBS-supplemented medium on TCPS (E), 24 h in 10% AHS-supplemented medium on TCPS (F), 24 h in 10% FBS and 100 ng ml�1 LPS supplemented medium onTCPS (G) and 24 h in 10% AHS and 100 ng ml�1 LPS supplemented medium on TCPS (H); and on PDMS and PEG: 24 h in 10% FBS-supplemented medium on PDMS (I), 24 h in10% AHS-supplemented medium on PDMS (J), 24 h in 10% FBS on PEG (K) and 24 h in 10% AHS on PEG (L).

518 D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525

AHS. The IL-1b concentration was significantly higher for monocytescultured on TCPS with LPS and AHS at 24 h than on all other surfacessupplemented with AHS. The IL-1b concentration was also signifi-cantly higher for monocytes on PEG cultured with AHS than withFBS at 24 h. The IL-12 p40 fraction concentration was significantly

higher for monocytes adherent on PEG supplemented with AHScompared to TCPS or PDMS. The p70 fraction of IL-12 was virtuallyundetected on all surfaces under all conditions. The TNF-a concen-tration was significantly higher for monocytes adherent on PEG withculture supplemented with AHS compared to FBS. Lastly, as

Fig. 2. Adherent cell density on TCPS or PEG hydrogels cultured with FBS ( ) or AHS ( ). All data are presented as average ± standard deviation (n = 3 human blooddonors). *Differs significantly from the other serum type on the same surface at the same time point (p < 0.05). �Differs significantly from the other surface with the sameserum type at the same time point (p < 0.05).

D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525 519

expected, the protein concentrations for the LPS positive controlwere significantly higher at various time points for GRO-2, TNF-aand IL-10. Additionally, IFN-c was measurable at 24 h for the LPSconditions but not for the others (Fig. 7).

4. Discussion

The increased spreading of the cells for the AHS conditions onPEG and TCPS suggests that the combination of soluble and ad-sorbed AHS proteins may have better allowed for greater cytoplas-mic extension in comparison to the xenogenic bovine proteins.Greater cell spreading on TCPS for the AHS condition has previ-ously been observed in monocytes being driven to form foam cells,and our observations are in line with these published images [35].Cells adhere via interaction with adsorbed proteins, thus the differ-ence in both cell spreading and adhesion levels may indicate thatthe human serum proteins were more readily available for cellinteraction and that the AHS proteins were likely more recogniz-able to receptors on the cell surface compared to proteins foundin FBS [36]. The different adhesion levels between the AHS andFBS (3-fold for TCPS, 10-fold for PEG) suggest that the material sur-face can greatly influence the effect of serum on monocyte adhe-sion. Given that the only difference between the FBS and AHSconditions are the proteins present, this phenomenon of differen-tial adhesion may have an explanation rooted in protein adsorptionto the biomaterials. For example, PEG hydrogels are well knownto be resistant to protein adsorption; however, a study on thePEG-like surface tetraglyme indicated that complement proteinC3 is the primary mediator of monocyte adhesion to tetraglyme[37]. Additionally, C3 activity on PEG hydrogels is comparable toTCPS, and monocyte adhesion to PEG hydrogels is lowered when

using C3-depleted serum [23]. C3 is present only at very low,nearly undetectable levels in FBS, while adult humans have a C3concentration of approximately 75–135 mg dl�1 [38,39]. Thus, ifmore C3 were present in the AHS condition, this could potentiallyhave promoted more cell adhesion once C3 was adsorbed/absorbedby the PEG hydrogel, although it is likely that C3 was not the onlyfactor as the difference in adhesion in this case is 10-fold and in theC3-depleted study reduced cell adhesion at 24 h was closer to half[23]. Monocyte adhesion to TCPS is mediated by a number of pro-teins, including C3, and thus is not subjected to such substantialdifferences in cell adhesion between FBS and AHS due to a depen-dence on a single protein for adhesion [14,32]. To further corrobo-rate this, activated C3 (C3b and iC3b) has been shown to complexwith IgG, thereby masking the presence of surface-bound IgG orreleasing it from a material’s surface [40]; also, as immobilized,insoluble IgG is a potent inducer of the anti-inflammatory cytokineIL-1ra [41], the highly activated C3 on PEG in the AHS conditioncould partially explain the observed inflammatory response. Thesefindings raise important concerns when comparing monocytebiomaterial studies that utilize different types of serum. The studyalso indicates the potential role of PEG in influencing monocytebinding to a single adsorbed protein (i.e. C3).

Differences in protein concentrations were also observed be-tween the AHS- and FBS-treated groups. The higher concentrationson the PEG surface of IL-1b and TNF-a compared to TCPS suggest aheightened inflammatory response to PEG, as has been demon-strated previously and may be due to adsorbed C3 [15,42,43].However, the increased amount of inflammatory protein secretionfor the AHS condition on PEG compared to the FBS condition mayalso be the result of the substantial difference in cell adhesion.While not statistically significant, differences in the GM-CSF,TGF-a and IP-10 concentrations between serum and surface types

Fig. 3. Concentration (pg ml�1) of IL-1b and TNF-a in the presence of TCPS or PEG cultured with FBS ( ) or AHS ( ). All data are presented as average ± standard deviation(n = 3 human blood donors). *Differs significantly from the other serum type on the same surface at the same time point (p < 0.05). �Differs significantly from the other surfacewith the same serum type at the same time point (p < 0.05).

520 D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525

may have been masked by the high level of deviations that arecommonly associated with donor variability, while the concentra-tion of the chemokine MCP-1 detected in the cultures may simplybe an overall response to the stress of cell culture where the MCP-1concentrations on PDMS and TCPS were both comparably high. Thestress on cells induced by culture conditions has been reportedpreviously by another group [44].

As with the first part of the study, the difference in adhesion be-tween AHS/FBS was much greater for PEG than TCPS, showing thatthis finding is consistent. However, the lack of observed differencesin the adhesion level on TCPS between serum types in the presenceof LPS is likely due to LPS activation. LPS has been shown previ-ously to enhance monocyte adhesion, and this was true for theFBS condition at 2 and 24 h [45]. This enhanced monocyte adhe-sion was not observed for the AHS condition probably becausethe adherent monocyte density on TCPS with AHS supplementa-tion was already very high. Thus, any increase in adhesion due toLPS activation may not have been observed due to the unavailabil-ity of space on the surface for additional cells to adhere. Similarly,no difference in cell adhesion was observed for all time points be-

tween AHS- and FBS-supplementation on PDMS. PDMS is a hydro-phobic surface that can mediate a high level of protein adsorption,which may have negated any possible contribution by the func-tionality of the adsorbed proteins (i.e. AHS vs. FBS) in promotingcell adhesion [46,47]. This explanation is also consistent with theobserved lack of a difference in cell spreading between the AHSand FBS conditions on PDMS and TCPS cultures with LPS.

The adherent cells on all surfaces for both serum conditionsstained for CD86, a surface antigen involved in the co-stimulationof T cell activation and proliferation that is indicative of a pro-inflammatory M1 macrophage phenotype [30,48]. In contrast, evenon extremely high laser settings, little to no staining was observedfor CD163, a scavenger receptor more indicative of a deactivated,pro-healing maturation state [30]. The consistent CD86 expressionsuggests that the cells are maturing and also that a basal level ofstress from cell culture is plausible [44]. It is also plausible thatthe culture system was lacking a specific signal to trigger the cellsto shift from an M1 to an M2 phenotype.

We further quantified the concentration of specific proteinsassociated with pro-inflammatory and anti-inflammatory states

Fig. 4. Concentration (pg ml�1) of MCP-1 in the presence of TCPS or PEG cultured with FBS ( ) or AHS ( ). All data are presented as average ± standard deviation (n = 3human blood donors). *Differs significantly from the other serum type on the same surface at the same time point (p < 0.05). �Differs significantly from the other surface withthe same serum type at the same time point (p < 0.05).

Fig. 5. Adherent cell density on TCPS, TCPS supplemented with 100 ng ml�1 LPS (LPS), PDMS or PEG hydrogels (PEG) cultured with FBS ( ) or AHS ( ). All data are presentedas average ± standard deviation (n = 3 human blood donors). T, differs significantly from TCPS (p < 0.05); L, differs significantly from TCPS with added LPS (p < 0.05); M, differssignificantly from PDMS (p < 0.05); P, differs significantly from PEG (p < 0.05); �, differs significantly from other serum type on same surface at same time point (p < 0.05).

D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525 521

Fig. 6. Confocal microscopy photomicrograph of adherent cells (20� magnifica-tion): 7 days in 10% FBS-supplemented medium on TCPS (A), 7 days in 10% AHS-supplemented medium on TCPS (B), 7 days in 10% FBS and 100 ng ml�1 LPSsupplemented medium on TCPS (C), 7 days in 10% AHS and 100 ng ml�1 LPSsupplemented medium on TCPS (D), 24 h in 10% FBS-supplemented medium onPDMS (E), 24 h in 10% AHS-supplemented medium on PDMS (F), 24 h in 10% FBS onPEG (G) and 24 days in 10% AHS on PEG (H). Green staining is of CD86 while redstaining is of CD163.

522 D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525

to better characterize the state of the adherent cells. The extremelyhigh concentrations of the chemokine GRO-2 and the pro-inflammatory cytokine IL-1b under the +LPS condition indicatesthe expected highly pro-inflammatory response from the cells[30,49,50]. In sharp contrast, the 100-fold higher GRO-2 proteinconcentration for the AHS culture condition on PDMS over theFBS culture was unexpected, especially considering that the adhe-sion levels were very comparable. Macrophages have previouslybeen shown to respond in a somewhat inflammatory manner toPDMS, which is a hydrophobic polymer and thus promotes proteinadsorption [44]. Consequently, it is possible that the adsorbed AHSproteins, in contrast to the adsorbed FBS proteins, were in a confor-mation that better stimulated a large release of GRO-2. However,GRO-2 proteins can also be implicated in M2 activities and, whenconsidered with the low concentration of other inflammatory pro-teins at 24 h for the AHS culture on PDMS, may indicate similaritieswith macrophages that are not strongly activated by M1 [28].There are very few, if any, examples of monocyte studies on bioma-terials that measure GRO-2; this finding may thus suggest theimportance of GRO-2 as a marker for monocyte/macrophage re-sponses to biomaterials. Lastly, the higher GRO-2 concentrationfor the AHS condition on PEG is likely explained by the higheradherent cell density for the AHS condition and not a novel phe-nomenon as observed on PDMS. However, the higher GRO-2 maybe indicative of a pro-inflammatory cellular response to the PEGhydrogels for both the AHS and FBS conditions.

IFN-c is a pro-inflammatory cytokine that is rarely produced bymacrophages and only under certain pro-inflammatory conditions,such as LPS supplementation [51]. Additionally, IL-10 is released inresponse to high concentrations of LPS to aid in the prevention ofendotoxic shock, indicating that the innately activated macrophagecontrol cells responded fairly predictably [52,53]. Conversely, theconcentration of IL-10 was fairly low for all other conditions, indi-cating a lack of this type of anti-inflammatory/regulatory activity.This high IL-10 concentration for the LPS condition may at first ap-pear to be counterintuitive as IL-10 is a potent anti-inflammatorymolecule [28]. However, IL-10 is also released in response tohigh concentrations of LPS to aid in the prevention of endotoxicshock [52,53]. Thus, the concurrent up-regulation of pro- andanti-inflammatory cytokines, especially under the LPS-challengedcondition, would not be unexpected. Further, although the concen-tration of IL-12(p40) tended to be higher for the +LPS condition,high levels of error rendered this result not statistically significant.

In contrast to the initial part of the study, high, though not sta-tistically significant, levels of IL-1b were observed on TCPS for theFBS condition. This was also observed for the FBS condition onPDMS. However, the standard deviation was very large. This wasbecause the extremely high levels of IL-1b came from one blooddonor and were not observed in the other donors, potentially indi-cating that this donor’s macrophages may have some sensitivity tothe bovine proteins in FBS. Contamination of the FBS is unlikely asFBS from the same bottle was used for the other donors in thestudy without similar response. Additionally, we have observedthis phenomenon previously in our laboratory (unpublished data).Colli et al. [35] previously demonstrated higher IL-1b productionfor macrophages cultured on TCPS with FBS compared to those cul-tured with AHS; however, when considering the present results,the higher level of IL-1b may be a donor-specific response. Donorvariability can be substantial in primary monocyte cultures but isnot always explicitly documented in the biomaterials literature.Surprisingly, a higher IL-1b concentration for the FBS conditionon PEG was not observed, suggesting that the surface also plays arole in mediating release of IL-1b from FBS-sensitive macrophages.The lack of many surface proteins to bind to for the FBS condition

Fig. 7. TNF-a (A), IL-1b (B), IFN-c (C), IL-10 (D), GRO-2 (E), IL-12(p40) (F) and IL-12(p70) (G) concentrations for TCPS, TCPS supplemented with 100 ng ml�1 LPS (LPS), PDMS orPEG hydrogels (PEG) cultured with FBS ( ) or AHS ( ). All data are presented as average ± standard deviation (n = 3 human blood donors). T, differs significantly fromTCPS (p < 0.05); L, differs significantly from TCPS with added LPS (p < 0.05); M, differs significantly from PDMS (p < 0.05); P, differs significantly from PEG (p < 0.05); �, differssignificantly from other serum type on same surface at same time point (p < 0.05).

D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525 523

524 D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525

on PEG may have influenced this low IL-1b release. The TNF-a con-centration paralleled the results for IL-1b, and again were primarilydue to one blood donor with a strong response.

5. Conclusions

The impact of xenoproteins in FBS on the interaction betweenmaterials and primary human cells is not well documented, andthus this study raises important considerations for the culturemedium when comparing the literature on monocyte adhesionand inflammatory responses to biomaterials. Our study implicatesthe critical role of the biomaterial surface in tandem with theculture serum in mediating the adhesion and protein release ofhuman blood-derived monocytes. For example, monocyte adhe-sion was approximately 3-fold higher on TCPS in the presence ofAHS compared to FBS, while on PEG hydrogels this differencewas much greater (10-fold). Further, the addition of LPS or theuse of PDMS eliminated any differential cell adhesion betweenthe two serum types. Through various markers, we also observedthat most of the cells exhibited a pro-inflammatory behavior,although the response on PEG tended to be the highest; further,GRO-2 protein was secreted at a 100-fold higher concentrationfor the AHS condition on PDMS compared to FBS, and this phenom-enon was unique to this surface.

Acknowledgement

This study was supported in part by NIH Grants HL77825 andEB6613.

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figure 6, are difficultto interpret in black and white. The full colour images can be foundin the on-line version, at doi:10.1016/j.actbio.2010.08.022.

References

[1] Anderson JM. Inflammation, wound healing, and the foreign-body response. In:Ratner BD, Hoffman AS, Schoen FJ, Lemons JE, editors. Biomaterials science. Anintroduction to materials in medicine. San Diego, CA: Elsevier Academic Press;2004. p. 296–304.

[2] Katsikis J, Yu H, Maurer F, Medcalf R. The molecular basis for the aberrantproduction of plasminogen activator inhibitor type 2 in THP-1 monocytes.Thromb Haemost 2000;84:468–73.

[3] Sundstrom C, Nilsson K. Establishment and characterization of a humanhistiocytic lymphoma cell line. Int J Cancer 1976;17:565–77.

[4] Rothlein R, Springer TA. The requirements for lymphocyte function-associatedantigen in homotypic leukocyte adhesion stimulated by phorbol ester. J ExpMed 1986;163:1132–49.

[5] Elkord E, Williams PE, Kynaston H, Rowbottom AW. Human monocyteisolation methods influence cytokine production from in vitro generatedendritic cells. Immunology 2005;144:204–12.

[6] Lehner M, Holter W. Endotoxin-free purification of monocytes for dendritic cellgeneration via discontinuous density gradient centrifugation based on dilutedFicoll-Paque Plus. Int Arch Allergy Immunol 2002;128:73–6.

[7] Labow RS, Meek E, Santerre JP. Hydrolytic degradation of poly(carbonate)-urethanes by monocyte-derived macrophages. Biomaterials 2001;22:3025–33.

[8] Collier TO, Anderson JM. Protein and surface effects on monocyte andmacrophage adhesion, maturation, and survival. J Biomed Mater Res2002;60:487–96.

[9] Shi C, Simon DI. Integrin signals, transcription factors, and monocytedifferentiation. Trends Cardiovasc Med 2006;16:146–52.

[10] Watson RWG, Redmond HP, Bouchierhayes D. Role of endotoxin inmononuclear phagocyte-mediated inflammatory responses. J Leukoc Biol1994;56(1):95–103.

[11] Figueiredo MD, Salter CE, Hurley DJ, Moore JN. A comparison of equine andbovine sera as sources of lipopolysaccharide-binding protein activity in equinemonocytes incubated with lipopolysaccharide. Vet Immunol Immunopathol2008;121:275–80.

[12] McBane JE, Matheson LA, Sharifpoor S, Santerre JP, Labow RS. Effect ofpolyurethane chemistry and protein coating on monocyte differentiationtowards a wound healing phenotype macrophage. Biomaterials 2009;30:5497–504.

[13] Boynton EL, Waddell J, Meek E, Labow RS, Edwards V, Santerre JP. The effect ofpolyethylene particle chemistry on human monocyte–macrophage functionin vitro. J Biomed Mater Res 2000;52:239–45.

[14] McNally AK, Anderson JM. Interleukin-4 induces foreign body giant cells fromhuman monocytes/macrophages. Am J Pathol 1995;147(5):1487–99.

[15] Schmidt DR, Kao WJ. Monocyte activation in response to PEG hydrogels graftedwith RGD and PHSRN separated by interpositional spacers of various lengths. JBiomed Mater Res 2007;83A(3):617–25.

[16] Moldenhauer A, Nociari MM, Dias S, Lalezari P, Moore MAS. Optimized cultureconditions for the generation of dendritic cells from peripheral bloodmonocytes. Vox Sang 2003;84:228–36.

[17] Pietschmann P, Stockl J, Draxler S, Majdic O, Knapp W. Functional andphenotypic characteristics of dendritic cells generated in human plasmasupplemented medium. Scand J Immunol 2000;51:377–83.

[18] Royer PJ, Tanguy-Royer S, Ebstein F, Sapede C, Simon T, Barbieux I, et al.Culture medium and protein supplementation in the generation andmaturation of dendritic cells. Scand J Immunol 2005;63:401–9.

[19] Duperrier K, Eljaafari A, Dezutter-Dambuyant C, Bardin C, Jacquet C, Yoneda K,et al. Distinct subsets of dendritic cells resembling DCs can be generatedin vitro from monocytes, in the presence of different serum supplements. JImmunol Methods 2000;238:119–31.

[20] Duperrier K, Eljaafari A, Dezutter-Dambuyant C, Schmitt D, Gebuhrer L, RigalD. Human serum or autologous plasma, allows generation of more maturedendritic cells than fetal calf serum, or serum-free medium. Exp Hematol1999;27(7):46.

[21] Mazlyzam AL, Aminuddin BS, Saim L, Ruszymah BH. Human serum is anadvantageous supplement for human dermal fibroblast expansion: clinicalimplications for tissue engineering of skin. Arch Med Res 2008;39:743–52.

[22] Shahdadfar A, Fronsdal K, Haug T, et al. In vitro expansion of humanmesenchymal stem cells: choice of serum is a determinant of cellproliferation, differentiation, gene expression, and transcriptome stability.Stem Cells 2005;23:1357–66.

[23] Wang X, Schmidt DR, Joyce EJ, Kao WJ. Application of MS-based proteomics tostudy serum protein adsorption/absorption and complement C3 activation onpolyethylene glycol hydrogels. J. Biomater. Sci. Polym. Ed.

[24] Xiaoyang Z, Baker H, Hancock WS, Fawaz F, McCaman M, Pungor E. Proteomicanalysis for the assessment of different lots of fetal bovine serum as a rawmaterial for cell culture. Part IV. Application of proteomics to manufacture ofbiological drugs. Biotechnol. Prog. 2006;22:1294–300.

[25] Adkins JN, Varnum SM, Auberry KJ, Moore RJ, Angell NH, Smith RD, et al.Toward a human blood serum proteome. Mol Cell Proteomics 2002;1:947–55.

[26] Dewez JL, Doren A, Schneider YJ, Rouxhet PG. Competitive adsorption ofproteins: key of the relationship between substratum surface properties andadhesion of epithelial cells. Biomaterials 1999;20:547–59.

[27] Naim JO, Van Oss CJ, Ippolito KML, Zhang JW, Jin LP, Fortuna R, et al. In vitroactivation of human monocytes by silicones. Colloids Surf B Biointerfaces1998;11:79–86.

[28] Martinez FO, Sica A, Mantovani A, Locati M. Macrophage activation andpolarization. Front Biosci 2008;13:453–61.

[29] Proudfoot AEI, Buser R, Borlat F, Alouani S, Soler D, Offord RE, et al. Amino-terminally modified RANTES analogues demonstrate differential effects onRANTES receptors. J Biol Chem 1999;274(45):32478–85.

[30] Trkola A, Matthews J, Gordon C, Ketas T, Moore JP. A cell line-basedneutralization assay for primary human immunodeficiency virus type 1isolates that use either the CCR5 or the CXCR4 coreceptor. J Virol 1999;73(11):8966–74.

[31] Gardiner DL, Michinton R. The macrophage assay. Aust J Med Lab Sci1987;8(2):19–24.

[32] Atkinson S, Valadas E, Smith SM, Lukey PT, Dockrell HM. Monocyte derivedmacrophage cytokine responses induced by M. bovis. BCG 2000;80(4–5):197–207.

[33] Bol SM, Van Remmerden Y, Sietzema JG, Kootstra NA, Schuitemaker H, Van’tWout AB. Donor variation in in vitro HIV-1 susceptibility of monocyte-derivedmacrophages. Virology 2009;390(2):205–11.

[34] Dinnes DLM, Marcal H, Mahler SM, Santerre JP, Labow RS. Material surfacesaffect the protein expression patterns of human macrophages: a proteomicsapproach. J Biomed Mater Res 2007;80A:895–908.

[35] Colli S, Lalli M, Rise P, Mussoni L, Eliginia S, Galli C, et al. Increasedthrombogenic potential of human monocyte-derived macrophagesspontaneously transformed into foam cells. Thromb Haemost 1999;81:576–781.

[36] Horbett TA. The role of adsorbed proteins in tissue response to biomaterials.In: Ratner BD, Hoffman AS, Schoen FJ, Lemons JE, editors. Biomaterials science.An introduction to materials in medicine. San Diego, CA: Elsevier AcademicPress; 2004. p. 237–46.

[37] Mayorga L, Ratner BD, Horbett TA. The role of complement adsorption andactivation in monocyte adhesion to ultralow protein adsorption surfaces madeby RF plasma deposition and activation in monocyte adhesion to ultralowprotein adsorption surfaces made by RF plasma deposition of PEO-liketetrathylene dimethyl ether (tetraglyme). World Biomaterials Congress;2008. p. 1162.

[38] Linscott WD, Triglia RP. The bovine complement system. Adv Exp Med Biol1981;137:413–30.

[39] Borigini MJ. Medical Encylopedia: Complement Componenet 3 (C3). Bethesda,MD: National Institutes of Health.

D. Schmidt et al. / Acta Biomaterialia 7 (2011) 515–525 525

[40] Nilsson UR. Deposition of C3b/iC3b leads to the concealment of antigens,immunoglobulin and bound C1q in complement-activating immunecomplexes. Mol Immunol 2001;38:151–60.

[41] Clinchy B, Gunneras M, Hakansson A, Hakansson L. Production of IL-1Ra byhuman mononuclear blood cells in vitro: influence of serum factors. Cytokine2006;34:320–30.

[42] Johnson RJ. The complement system. In: Ratner BD, Hoffman AS, Schoen FJ,Lemons JE, editors. Biomaterials science. An introduction to materials inmedicine. San Diego, CA: Elsevier Academic Press; 2004. p. 318–28.

[43] Griffin FM, Mullinax PJ. In vivo activation of macrophage C3 receptors forphagocytosis. J Exp Med 1985;162:352–7.

[44] Matheson LA, McBane JE, Malowany JI, Santerre JP, Labow RS. Is cell culturestressful? Effects of degradable and nondegradable culture surfaces on U937cell function. Biotechniques 2007;42:744–50.

[45] Kang YH, Lee CH, Brummel SE, Newball HH, Forrester J. Effects of endotoxin onexpression of VLA integrins by human bronchoalveolar lavage macrophages. JLeukoc Biol 1995;57:624–34.

[46] Schmidt DR, Waldeck H, Kao WJ. Protein adsorption to biomaterials. In: BiziosR, Puleo D, editors. Biological interactions on materials surfaces:understanding and controlling protein, cell, and tissue. New York: Springer;2009. p. 1–18.

[47] Chen H, Brook MA, Sheardown H. Silicone elastomers for reduced proteinadsorption. Biomaterials 2004;25:2273–82.

[48] Boehm U, Klamp T, Groot M, Howard JC. Cellular responses to interferon-gamma. Annu Rev Immunol 1997;15:749–95.

[49] Dinarello CA. Biologic basis for interleukin-1 in disease. Blood1996;87(6):2095–147.

[50] Semple BD, Kossmann T, Morganti-Kossmann MC. Role of chemokines in CNShealth and pathology: a focus on the CCL2/CCR2 and CXCL8/CXCR2 networks. JCereb Blood Flow Metab 2010;30(3):459–73.

[51] Carvalho-Pinto CE, Garcia MI, Mellado M, Rodriguez-Frade JM, Martin-Caballero J, Flores J, et al. Autocrine production of IFN-gamma bymacrophages controls their recruitment to the kidney and the developmentof glomerulonephritis in MRL/lpr mice. J Immunol 2002;169:1058–67.

[52] Kwan WH, Boix C, Gougelet N, Fridman WH, Mueller CG. LPS induces rapid IL-10 release by M-CSF-conditioned, tolerogenic dendritic cell precursors. JLeukoc Biol 2007;82:1–9.

[53] Berg DJ, Kuhn R, Rajewsky K, Muller W, Menon S, Davidson N, Grunig G, et al.Interleukin-10 is a central regulator of the response to LPS in murine models ofendotoxic shock and the Schwartzman reaction but not endotoxin tolerance. JClin Invest 1995;96:2339–47.