The effect of collagen I mimetic peptides on mesenchymal stem cell adhesion and differentiation, and on bone formation at hydroxyapatite surfaces

Abstract

Integrin-binding peptides increase cell adhesion to naive hydroxyapatite (HA), however, in the body, HA becomes rapidly modified by protein adsorption. Previously we reported that, when combined with an adsorbed protein layer, RGD peptides interfered with cell adhesion to HA. In the current study we evaluated mesenchymal stem cell (MSC) interactions with HA disks coated with the collagen-mimetic peptides, DGEA, P15 and GFOGER. MSCs adhered equally well to disks coated with DGEA, P15, or collagen I, and all three substrates, but not GFOGER, supported greater cell adhesion than uncoated HA. When peptide-coated disks were overcoated with proteins from serum or the tibial microenvironment, collagen mimetics did not inhibit MSC adhesion, as was observed with RGD, however neither did they enhance adhesion. Given that activation of collagen-selective integrins stimulates osteoblastic differentiation, we monitored osteocalcin secretion and alkaline phosphatase activity from MSCs adherent to DGEA or P15-coated disks. Both of these osteoblastic markers were upregulated by DGEA and P15, in the presence and absence of differentiation-inducing media. Finally, bone formation on HA tibial implants was increased by the collagen mimetics. Collectively these results suggest that collagen-mimetic peptides improve osseointegration of HA, most probably by stimulating osteoblastic differentiation, rather than adhesion, of MSCs.

Figure 1. DGEA and P15 increase cell adhesion and cell spreading on HA

A) MSCs labeled with Cell Tracker dye were allowed to adhere for 1 hour to HA disks pre-coated with collagen I (Col I), DGEA, P15 or GFOGER. The adherent cells were subsequently lysed and adhesion was quantified by measuring fluorescence. Values were folded to uncoated HA (dotted line) (*=p<0.05 to uncoated HA). B) Representative images of MSCs allowed to adhere for 1 hour to collagen I (Col I), DGEA, P15 or GFOGER-coated HA. Cells were labeled with Alexa-488 phalloidin (Cell Signaling Technologies).

Figure 2. DGEA and P15 do not inhibit cell adhesion when presented in the context of adsorbed serum proteins

A) HA disks were coated with collagen I (Col I), DGEA, P15, RGD or RGE followed by an over-coating with serum (FBS) to allow protein adsorption. As a control, some disks were coated with FBS only. Fluorescently-labeled MSCs were then seeded onto the disks and monitored for cell adhesion as previously described. Values were folded to uncoated HA (dotted line) (*=p<0.05 to FBS). B) Representative images of Alexa-488 phalloidin stained MSCs adherent to HA disks coated with FBS alone or to disks coated first with either collagen I (Col I), DGEA, P15, RGD or RGE, followed by an FBS over-coating.

Figure 3. DGEA and P15 do not inhibit cell adhesion when presented in the context of adsorbed proteins from the tibial microenvironment

A) HA disks were either left uncoated (UNC, panels 1-2) or coated with DGEA (panels 3-4) or P15 (panels 5-6). The disks were then implanted into the rat tibiae for 30 minutes to allow protein adsorption, retrieved, and washed extensively with PBS. MSCs were seeded onto the surfaces, and allowed to adhere for 1 hour. Cells were then fixed and labeled with Alexa-488 phalloidin (1,3,5) and DAPI (2,4,6). Representative images are shown. B) Adherent cells were quantified by counting cells from multiple microscopic fields from 5 separate implants per group. Data are reported as fold to the uncoated HA implants. No significant difference was observed in the number of adherent cells for the 3 surface treatments.

Figure 4. DGEA and P15 pre-coatings do not affect protein adsorption from serum or the tibial microenvironment

A) Representative western blots of FN and VN desorbed from disks coated with either FBS alone, or with DGEA or P15, followed by FBS. B) Western blots of FN, VN and Fbg desorbed from uncoated, DGEA-coated or P15-coated, HA disks implanted for 30 minutes into rat tibiae.

Figure 5. DGEA and P15 increase osteoblastic markers in the presence of osteogenic media

A) MSCs were seeded onto disks coated with FBS alone, or on disks sequentially coated with either DGEA/FBS or P15/FBS. Cells were allowed to grow in osteogenic media (OS media) for 2 weeks. A representative image of an alkaline phosphatase (ALP) activity assay is shown. B) MSCs adherent to peptide/protein-coated disks were grown in OS media for 2 weeks as previously described. As a control, some cells were grown on FBS-coated disks in growth media rather than OS media. Following a 2-week incubation in either OS or growth media, culture supernatants were collected and evaluated for osteocalcin (OCN) secretion by ELISA. Values are reported as fold to the samples incubated in growth media (*=p<0.05 relative to FBS samples in growth media and †=p<0.05 relative to FBS in OS media).

Figure 6. DGEA and P15 increase osteogenic markers in the absence of osteogenic media

A) Representative image of an alkaline phosphatase activity assay for MSCs incubated in growth media on HA disks pre-coated with FBS, DGEA/FBS, or P15/FBS. B) Quantification of osteocalcin (OCN) secretion from samples incubated in standard growth media for two weeks. (*=p<0.05 to FBS).

Figure 7. DGEA and P15 increase bone formation around HA implants

A) Representative images of Goldner's Trichrome-stained sections from tibiae implanted with uncoated (UNC), DGEA-coated, or P15-coated, HA disks. B) Quantification of the amount of new bone formed (black bars) and the amount of bone directly contacting the implant perimeter (white bars) around uncoated (UNC), DGEA-coated or P15-coated implants (*=p<0.05 to UNC)

Figure 8. Model describing the effects of integrin-binding peptides on osseointegration of HA implants

A) Integrin activation by adsorbed proteins, such as FN and VN, plays a key role in MSC adhesion, survival and osteoblastic differentiation. When RGD is present on the HA surface, we hypothesize that integrins such as αvβ3 bind the RGD rather than full-length FN or VN, leading to poor cell adhesion and survival. Collagen-binding integrins such as α2β1 would not likely be engaged with ligand, given that minimal amounts of collagen I would adsorb to the HA surface from blood (given that fibrillar collagen I is not abundant in blood). The combination of weak signaling from RGD-dependent integrins (e.g. αvβ3, α5β1, αIIbβ3), and a lack of signaling from collagen-selective integrins, is proposed to contribute to poor implant integration. B) Conversely, the presence of either DGEA or P15 on the HA surface provides a ligand for collagen-selective integrins that, upon activation, initiate signaling mechanisms promoting osteoblastic differentiation. As well, RGD-dependent integrins would engage the native FN, VN or Fbg adsorbed from blood, resulting in strong adhesive and survival signaling. Collectively, signaling from these multiple integrin species is hypothesized to enhance osseointegration of HA biomaterials.