Protein Polymer MRI Contrast Agents: Longitudinal Analysis of Biomaterials in Vivo

Abstract

Despite recent advances in tissue engineering to regenerate biological function by combining cells with material supports, development is hindered by inadequate techniques for characterizing biomaterials in vivo. Magnetic Resonance Imaging (MRI) is a tomographic technique with high temporal and spatial resolution and represents an excellent imaging modality for longitudinal non-invasive assessment of biomaterials in vivo. To distinguish biomaterials from surrounding tissues for MR imaging, protein polymer contrast agents (PPCAs) were developed and incorporated into hydrogels. In vitro and in vivo images of protein polymer hydrogels, with and without covalently incorporated PPCAs, were acquired by MRI. T₁ values of the labeled gels were consistently lower when PPCAs were included. As a result, the PPCA hydrogels facilitated fate tracking, quantification of degradation, and detection of immune response in vivo. For the duration of the in vivo study, the PPCA-containing hydrogels could be distinguished from adjacent tissues and from the foreign body response surrounding the gels. The hydrogels containing PPCA have a contrast-to-noise ratio two-fold greater than hydrogels without PPCA. In the absence of the PPCA, hydrogels cannot be distinguished by the end of the gel lifetime.

Introduction

A primary goal of tissue engineering is to regenerate biological function by combining cells and a biocompatible material with appropriate physical and biochemical stimuli to guide growth. Despite recent advances, development is hindered by inadequate techniques for material characterization (1). In in vivo studies where a biomaterial is implanted into an animal for a prescribed time, animals are sacrificed and tissues excised for histological examination. Drawbacks of this approach include the use of numerous animals, a lack of three-dimensional data, artifacts from the animal not being alive, and the inability to perform longitudinal studies over time (2). Non-invasive imaging techniques are needed for a thorough evaluation of the properties and performance of biomaterials in vivo.

Magnetic Resonance Imaging (MRI) is a tomographic technique with high temporal and spatial resolution and a favorable safety profile. It is an excellent imaging modality for longitudinal non-invasive assessment of biomaterials in vivo. Since MRI is sensitive to the chemical and physical environment of water molecules, it can provide information about both engineered hydrogels and the surrounding tissue. Previously, MRI has been used to image multiple properties of hydrogels in tissue engineering that include promotion of angiogenesis (3), porosity and mass transport (4–5), surrounding tissue characteristics (6–8), immune response (9–11), degradation (12–14), and drug release (15–16).

Typically, there is insufficient contrast-to-noise ratio (CNR) between an area of interest and its surroundings in an MR image. Contrast agents (CAs) address this shortcoming by increasing the relaxation rate of neighboring water protons (17). Clinically used CAs, such as gadolinium-diethylenetriaminepentaacetic acid [Gd(III)-DTPA] and gadolinium-1,4,7,10-tetraazacyctododecome-1,4,7,10-tetraacetic acid [Gd(III)-DOTA] have limited sensitivity (18). Attachment of multiple Gd(III) ions to a macromolecule amplifies the signal by increasing Gd(III) concentration and slowing molecular tumbling (19–22).

We have reported a family of multivalent, macromolecular protein polymer contrast agents (PPCAs) that have high relaxivity, and are biodegradable and nontoxic (23). These PPCAs can be covalently crosslinked into the protein polymer hydrogel we have designed for tissue engineering applications (24). Since the PPCAs have high relaxivity and degrade at the same rate as the unmodified hydrogel, we hypothesized that we could non-invasively monitor the degradation of the hydrogels by MR imaging.

Here, we describe our results of imaging protein polymer hydrogels with covalently incorporated PPCAs that were implanted into mice. The substantially higher contrast-to-noise ratio of the hydrogel containing PPCA facilitates fate tracking, quantifying degradation, and the ability to distinguish the hydrogel from host tissues and in vivo immune response. These studies demonstrate that the inclusion of PPCA within the hydrogels is necessary. Hydrogels without PPCA could not be visualized toward the end of the gel lifetime, prohibiting longitudinal, non-invasive evaluation.

Experimental Methods

Contrast Agent Synthesis

Protein polymer contrast agents were synthesized as described previously (23). The backbone of the CA was the protein polymer designated K8-120, with a sequence of GH₁₀SSGHIDDDDKHM(GKAGTGSA)₁₂₀G. The K8-120 PPCA has 120 repeats of an amino acid sequence containing lysines spaced eight amino acids apart. It was synthesized using genetic engineering and recombinant protein expression and purification with standard techniques (25). The conjugation reaction between the protein polymer and the Gd(III)-1,4,7-tris(carboxymethyl)-10-carboxybutyl-1,4,7,10-tetraazacyclododecane [Gd(III)-DO3A] chelators was performed in an aqueous buffer with 1-ethyl-3-carbodiimide hydrochloride (EDC, Fisher Scientific) and N-hydroxysulfosuccinimide (sulfo-NHS, Fisher Scientific). The reaction was dialyzed and lyophilized to obtain the PPCA conjugate.

Relaxivity

T₁ measurements were performed in triplicate using a Bruker mq60 NMR Analyzer (Bruker Canada, Milton, Ont., Canada) at 60 MHz (1.5 T) and 37 °C at three concentrations. Inductively coupled plasma atomic emission spectrometry (ICP-AES) on a Varian VISTA-MPX ICP spectrometer (Palo Alto, CA) at Northwestern University’s IMSERC was used to measure Gd(III) concentration. Relaxivity was determined by fitting the slope of 1/T₁ versus Gd(III) concentration.

Hydrogel Formation

Hydrogels were prepared by enzymatic crosslinking of two protein polymers, K8-30 and Q-6 (GH₁₀SSGHIDDDDKHM [GH₁₀SSGHIDDDDKHM(GKAGTGSA)30G] [(GQQQLGGAGTGSA)₂(GAGQGEA)₃]₆G) with and without the K8-120 PPCA. The K8-30 and Q-6 proteins were cloned, expressed, and purified using the same methods as described for the K8-120 protein polymer (25), but were further purified by phase separation to remove endotoxins. It was not necessary to remove endotoxins for the K8-120 PPCA because the small concentrations used in the experiments constituted a negligible amount.

Endotoxins were removed through multiple rounds of a phase separation method optimized from a literature protocol (26). Triton X-114 (Sigma) was added at 1% to protein dissolved at 10 mg/mL in endotoxin-free water and the pH was adjusted to ~9.5. The solution was stirred for 30 min at 4 °C, placed in a 37 °C water bath for 10 min and centrifuged at 10,000g at 37 °C for 10 min. The supernatant containing the protein was placed into a new conical tube and the process was repeated multiple times, with pH readjustment to ~9.5 after every four rounds. Following the last round of phase separation, the solution was placed on degassed Bio-beads SM2 Adsorbents (Bio-rad Laboratories, Hercules, CA) to remove any remaining Triton X-114. Samples were dialyzed against endotoxin free water and lyophilized. The endotoxin levels were tested using a QCL-1000 Endpoint Chromogenic LAL assay (Lonza, Walkersville, MD). Proteins were only used if the endotoxins (EU) per mL of protein when dissolved at concentrations used in the hydrogels were less than 20 EU/mL.

Tissue transglutaminase (tTG) from guinea pig liver (Sigma) was used to crosslink the protein polymers into hydrogels. tTG was dissolved at 0.04 units/μL in 2 mM Ethylenediaminetetraacetic acid (EDTA), 20 mM Dithiothreitol (DTT), pH 7.7. The lysine-containing protein, K8-30, was dissolved at 10 wt% in 200 mM 4-Morpholinepropanesulfonic acid (MOPS), 20 mM CaCl₂, pH 7.6. The glutamine-containing protein, Q-6, was resuspended at 15 wt% in 2 mM EDTA, pH 7.3. The three components were combined at a ratio of 1:1.5:1.5 for tTG:K8-30:Q-6 solutions. K8-120 CA was dissolved at 10 wt% in 200 mM MOPS, 20 mM CaCl₂, pH 7.6. When K8-120 CA was included, the K8-30 volume was decreased by a corresponding amount to maintain constant hydrogel volume. Protein polymer precursors with enzyme were mixed with brief vortexing and incubated at 37 °C until gelation occurred (minutes).

Fluorophore-labeled K8-30 was included at 1% of the total hydrogel volume to validate the presence of the hydrogel with histology. To synthesize this conjugate, the Alexa-Fluor 488 5-TFP dye (excitation wavelength = 496 nm, emission wavelength = 519 nm, Invitrogen) was coupled with the K8-30 protein polymer. The protein polymer was dissolved at 1.16 mg/mL in 0.1 M sodium bicarbonate, pH 9.0, and the AlexaFluor 488 5-TFP was dissolved at 1 mg/μL in DMF. A total of 50 μL of the AlexaFluor solution was added to 5 mL of protein polymer solution. The reaction was incubated at room temperature with continuous stirring for 1 h. Unreacted fluorophore and salt were removed using a CENTRI-SEP spin column (Princeton Separations, Adelphia, NJ). The average number of conjugated fluorophores was estimated by MALDI-TOF spectrometry. This conjugate was dissolved at 10 wt% in 200 mM MOPS, 20 mM CaCl₂, pH 7.6 for inclusion in the hydrogel.

The Gd(III) concentration in the hydrogel samples was analyzed by inductively coupled plasma mass spectrometry (ICP-MS) on a Thermo Electron Corporation (Waltham, MA) XSeriesII ICP-MS with Thermo PlasmaLab software. All Gd(III) standards and samples contained 5 ng/mL of a multielement internal standard (Spex CertiPrep, Metuchen, NJ) consisting of Bi, Ho, In, Li, Sc, Tb, Y, and 3% nitric acid (v/v).

In Vitro Hydrogel Imaging

Gels with 0.1 mM Gd(III) of K8-120 PPCA were imaged on a 4.7 T Bruker Biospec 4740 MR system using a 38 mm diameter birdcage coil (Rapid MRI, Columbus, OH) without temperature control. A spin echo pulse sequence was used with TR = 100 ms, TE = 14.5 ms, 20 signal averages, and a 0.12 × 0.12 × 1 mm³ resolution. To calculate T₁, an accelerated variable TR pulse sequence (RARE-VTR) was used with TE = 9.8 ms, TR = 131.5, 250, 400, 550, 750, 1050, 1550, and 3750 ms, two signal averages, and a 0.12 × 0.12 × 1 mm³ resolution. T₂ values were measured using a multiecho sequence with TR = 2000 ms, TE = 15 – 240 ms (intervals of 15 ms), one signal average and a 0.12 × 0.12 × 1 mm³ resolution.

In Vivo Studies

All animal procedures were approved by Northwestern University’s and Northshore University HealthSystem’s Institutional Animal Care and Use Committees. Hydrogels with a total volume of 50 μL were preformed and surgically implanted subcutaneously on the hind flanks of age-matched C57/BL6 mice. Mice were continuously anesthetized with 2% isoflurane and the surgical area was shaved and wiped with alcohol. Incisions were parallel to the spine and a pocket was created by blunt dissection to allow insertion of the hydrogel. Monocryl 5-0 sutures (Esutures, Mokena, IL) were used to close incisions.

In Vivo MR Imaging

For each imaging experiment, mice were anesthetized using 2% isoflurane and respiration was monitored using a pressure sensor. Anesthesia was adjusted as needed to maintain a respiratory rate of 60–90 breaths/minute. The mice were serially imaged over a period of up to 22 days after surgery using the same imaging protocol employed for the in vitro image acquisition. After preliminary coronal and axial images were acquired to localize the graft, a series of fat-suppressed images were measured.

To compare relaxation times in the hydrogels (pre- and post-implantation), mice were implanted with the same hydrogels that were imaged in vitro. Mice were imaged immediately after surgery and again one day later. To measure T₁s, a RARE-VTR pulse sequence with 8 TRs from 115 to 3750 ms was used with TE = 12.5 ms, one signal average, and a 0.16 × 0.16 × 1 mm³ resolution. T₂ values were measured using a multiecho sequence with TR = 2000 ms and 16 echoes with TE = 20 to 400 ms, one signal average and a 0.16 × 0.16 × 1 mm³ resolution.

In Vivo Detailed Degradation Curve

For a detailed assessment of hydrogel degradation, 0.3 mM Gd(III) PPCA-containing hydrogels were implanted into three mice and they were imaged every 3–5 days for up to 23 days after implantation. A MSME spin echo pulse sequence was used with TR/TE = 500/14.5 ms, 4 cm FOV, 256 × 256 matrix, 1 mm slice thickness, and 2 signal averages. Both coronal and axial stacks were acquired to fully cover the grafts.

In Vivo Time Course

Eight mice were implanted with two hydrogels each: a hydrogel without PPCA on the mouse’s left side, and with 0.2 mM Gd(III) PPCA on the right. All imaging was performed with a 4 cm FOV and 256 × 256 matrix and gain settings were constant for each type of scan for each mouse over the entire time. Imaging was conducted to evaluate hydrogel degradation, contrast to noise ratio, and T₁ values over time. Stacks of axial slices with a MSME spin echo pulse sequence with 8–10 0.5 mm thick slices and TR/TE = 285/14.6 ms were acquired to analyze graft volumes. Gradient echo images were acquired to analyze contrast to noise ratio and were run with the following parameters: TR/TE = 100/4.1 ms, two axial 1 mm slices (one slice through the center of each hydrogel), 4 cm FOV, 256 × 256 matrix, and 20 signal averages. T₁ values were measured using a RARE-VTR sequence with TE = 9.8 ms, TR = 255, 370, 505, 670, 881, 1177, 1673, 3750 ms, two 1 mm slices (the same slices used for CNR measurements), 4 cm FOV, 256 × 256 matrix, and two signal averages.

Image Analysis

Images were analyzed using ImageJ software (National Institutes of Health, Bethesda, MD). Additionally, 3-D reconstructions of the grafts were constructed in Amira 4.0 software (Visage Imaging, Inc., Andover, MD). To calculate the volume, the graft was manually planimetered on each slice and the areas were summed and multiplied by the slice thickness. Only bright central areas were included in volume calculations, omitting any surrounding ring of lighter intensity (which was interpreted as inflammatory response).

Data from multiple TR and multiple TE scans (to calculate T₁ and T₂ values, respectively) were fit with Origin 7 (OriginLab Corporation, Northampton, MA) using the following equations:

$$S_i=S_0(1-exp)\left(-TR/T_1\right))$$ (1)

$$S_i=S_{0}exp\left(-TE/T_2\right)$$ (2)

where S₀ is a constant and S_i is the signal intensity at a particular TR or TE.

Gradient echo images were analyzed for the signal intensity of the hydrogel with PPCA, the hydrogel without PPCA, and skeletal muscle as an internal control (constant across mice and time). Additionally, the standard deviation (SD) of the signal intensity of the background was measured. Contrast-to-noise ratio (CNR) was calculated relative to skeletal muscle using the following equation:

$$\mathit{CNR}=\frac{{SI}_H-{SI}_M}{SD}$$ (3)

where SI_H is the signal intensity of the hydrogel and SI_M is the signal intensity of skeletal muscle.

Histology

At each time point, 1 or 2 mice were sacrificed for histological examination. Implants were excised and placed immediately in 10% neutral buffered formalin (Fisher Scientific). Samples were processed on a Leica TP 1050 tissue processor (Leica Microsystems, Germany) and paraffin embedded on a Sakura Tissue Tek embedding unit (Sakura Finetek U.S.A., Torrance, CA). Sections were cut at 5 μm thickness on a Leica 2135 microtome (Leica Microsytems). Slides were stained with hematoxylin and eosin (H&E) with standard procedures and all materials were purchased from Sigma. Masson’s trichrome staining (all supplies from Sigma) was performed according to standard protocols to detect collagen capsule formation. Sections were examined by immunohistochemistry to distinguish macrophages using a Mac-3 antibody, clone M3/84, (BD Pharmingen, San Jose, CA), dilution 1:200, with a biotinylated secondary antibody/streptavidin horseradish peroxidase/3,3′-diaminobenzidine system (Dako, Carpinteria, CA). Slides were imaged using a Nikon Eclipse 50i upright microscope (Nikon, Melville, NY) and pictures were taken with Spot Advanced software (Diagnostic Instruments, Sterling Heights, MI) in Northwestern University’s Institute for Bionanotechnology and Medicine (IBNAM). A mercury lamp with a 492–518 nm excitation wavelength and a 532–536 emission filter was used to image fluorescently labeled gels on unstained slides.

Statistical Analysis

Statistics were performed using Origin 7 (OriginLab Corporation, Northampton, MA) with a one-tailed, two-sample T test. A p value less than 0.05 was considered significant.

Results

In Vitro Hydrogel Imaging

The synthesized K8-120 PPCA contained ~39 Gd(III)-DO3A chelators, representing a conjugation efficiency of 32%. The relaxivity was 15.1 mM⁻¹s⁻¹ per Gd(III) and 584.4 mM⁻¹s⁻¹ per molecule at 37 °C and 1.5 T. Protein polymer hydrogels, with and without 0.1 mM Gd(III) of the K8-120 PPCA, were imaged on a 4.7 T magnet. The hydrogels containing the PPCA have greater signal intensity on a T₁ weighted image (Figure 1). Signal intensity differences correlate with T₁ relaxation times at 4.7 T, with the PPCA-containing hydrogel having a T₁ relaxation time of 354 ± 2 ms, significantly shorter than the non-PPCA-containing hydrogel’s T₁ of 716 ± 4 ms (P < 0.01) (Table 1). The T₂ relaxation times are also significantly different (P < 0.01), with the PPCA-containing hydrogel’s value 2.7 times lower.

Figure 1.

MR images at 4.7 T of hydrogels containing 0 mM Gd(III) and 0.1 mM Gd(III) of the K8-120 PPCA. Imaging parameters: TR/TE = 100/14.6 ms, 128x128 matrix, 3 cm FOV, 1 mm slice thickness, and 20 signal averages.

Table 1.

T₁ and T₂ relaxation times for hydrogels as measured on a 4.7 T imager.

	T1a		T2a
	CA	No CA	CA	No CA
Hydrogel prior to implantation	354 +/− 2	716 +/− 4	45.9 +/− 1.1	124.1 +/− 1.4
Immediately after implantation	452 +/− 12	910 +/− 13	45.6 +/− 0.8b	157.9 +/− 2.5b
One day after implantation	601 +/− 60	1152 +/− 6	52.1 +/− 1.8b	149.3 +/− 2.2b

^aAll times are in ms. n = 2 except where noted.

^bn=1.

T₁ and T₂ Values of Hydrogels Before and After Implantation

The hydrogels imaged in vitro were subcutaneously implanted into mice and imaged immediately and again one day later. As shown in Table 1, the T₁ values significantly increase from prior to implantation to immediately after implantation and again one day after implantation (an overall 69% increase). On the other hand, T₂ values vary less, with only a 21% overall increase.

Detailed Degradation Curve

To obtain a detailed degradation curve, 50 μL PPCA-containing hydrogels were implanted subcutaneously in three mice and imaged at a high frequency, every 3–5 days. The hydrogel implanted into Mouse 1 fractured prior to implantation, resulting in a smaller gel. The hydrogels almost completely degrade over the course of 20 days (Figure 2). By measuring the area of the hydrogel in each slice, followed by multiplying by the slice thickness (and summing over all slices), the implant volume is obtained and degradation can be quantified (Figure 2B). For each mouse, there is a lag time of approximately 6 days where no apparent degradation occurs, followed by a linear degradation rate of 4.2 ± 0.6 mm³ per day. Toward the end of the experiment, it appears that the degradation rate slows.

Figure 2.

Detailed degradation curve of a PPCA-containing hydrogel implanted subcutaneously in three mice. Images were acquired every 3–5 days on a 4.7 T magnet with a MSME spin echo pulse sequence. a) axial spin echo images of Mouse 2 b) 3-D reconstructions of the grafts in Mouse 2 c) Calculated volume versus time using an average of coronal and axial images for Mouse 1(▲), Mouse 2(), and Mouse 3 ().

Time Course

A time course series study was performed to obtain statistically relevant data for degradation, contrast-to-noise ratio, and T₁ relaxation times. Imaging time points immediately after implantation, and corresponding to the beginning, middle, and end of degradation were determined from the degradation curve. These points corresponded to 1, 8, 15, and 22 days after implantation. Both the implant volume and the contrast decreased over time, as shown in Figure 3. The volume was calculated for the hydrogels containing PPCA using 0.5 mm slices (Figure 4).

Figure 3.

Gradient echo images on a 4.7 T magnet of the same mouse at 1, 8, 15, and 22 days after hydrogels were implanted. Arrowheads indicate the hydrogel with PPCA and arrows indicate the hydrogel without PPCA. Volume and CNR of implants decrease over time; hydrogels with PPCA have greater CNRs throughout the time series. Imaging parameters: TR/TE = 100/4.1 ms, 1 mm slice thickness, 256 matrix, 4 cm FOV, and 20 averages.

Figure 4.

Volume of hydrogel implants with 0.2 mM K8-120 PPCA as measured with MR images at 4.7 T. Each volume is statistically greater than the subsequent volume with P < 0.01. Images were acquired with 0.5 mm slices, TR/TE = 285/14.6 ms, 4 cm FOV, 256 matrix, and 4 signal averages.

Contrast-to-noise ratio was compared between the hydrogels with and without PPCA. Hydrogels with PPCA have a much higher CNR for all time points (P < 0.05) (Figure 5). With PPCA, hydrogels initially have a CNR of 74.8, which drops to 23.5 by day 22. Hydrogels without PPCA have an initial CNR of 33.0, less than half of the value with PPCA; it decreases to 3.4 at day 22.

Figure 5.

Contrast-to-noise ratio of hydrogel to skeletal muscle. Gray bars: hydrogels with PPCA. White bars: hydrogel without PPCA. A statistically significant difference with P < 0.01 is denoted with ** and P < 0.05 is denoted by *.

T₁ relaxation times corresponded to intensities seen in the images. A ring was observed around the hydrogel in the MR images that was hypothesized to be inflammation (discussed in detail in the next section). Measurements were taken of both the hydrogel and this surrounding area. At day 1 there was no apparent inflammation surrounding the gels and thus, no measurement was taken. At each time point the T₁ of the hydrogel with PPCA was significantly lower than without PPCA and the T₁ of both hydrogels increased from day 1 to day 22 (Figure 6). The inflammation surrounding the gels had a significantly longer T₁ than all the gels except the non-PPCA gel at day 22. At days 8 and 15, the T₁ of the area surrounding the PPCA- containing hydrogels is significantly shorter than the area adjacent to the non-PPCA-containing gels. Between day 1 and day 8, the T₁ of the PPCA-containing gel remains approximately the same, but is significantly longer at each of the next two time points.

Figure 6.

Comparison of T₁ relaxation times over 22 days for hydrogel graft with PPCA (white), hydrogel graft without PPCA (grey), inflammatory area surrounding hydrogel with PPCA (white hashed), and inflammatory area surrounding hydrogel without PPCA (grey hashed). A statistically significant difference with P < 0.01 is denoted with ** and P < 0.05 is denoted by *.

Imaging Immune Response

During the course of the in vivo studies, none of the animals appeared sick or died unexpectedly, indicating the implants were non-toxic. However, there was a visual indication of a local immune response from the implants, with round protrusions at the site in some animals. At each time point, at least one animal was sacrificed and both implants were excised for histological examination and validation of the MR images. H&E stained histology samples that show significant immune cell infiltration corresponds well to acquired MR images of the implanted gels with a surrounding layer (Figure 7a–h).

Figure 7.

Histology and MR images of excised implants. a&e: day 1, no PPCA; b,f,i: day 8, with PPCA; c: day 15, no PPCA; g: day 15, with PPCA; d,h,j,k,l: day 22, with PPCA. a–d,g: H&E stained images with 4x magnification; e,f,h: Corresponding gradient echo MR images with TR/TE = 100/4.1 ms and with red arrow marking lighter area of inflammation; i: unstained image confirming hydrogel; j: Mac-3 staining at 4x magnification of PPCA implant at 22 days; k: Masson’s trichrome staining at 4x magnification of PPCA implant at 22 days; l: Enlarged close-up of Masson’s trichrome staining. * indicates the implant, arrowheads mark polymorphonuclear leukocyte infiltration, red arrow marks lighter area of inflammation, and black arrow points to the collagen capsule.

The hydrogels could be positively identified in histology images by including fluorophore-conjugated K8-30 (Figure 7i). After one day, immune cell infiltration was minimal, with a small layer of cells and no fibrinous exudate, and there is no ring of lighter intensity in the corresponding MR image. However, at the next time point (8 days after implantation), a dense polymorphonuclear leukocyte infiltration and a ring of fibrinous exudate appear that persisted through the last time point at 22 days after implantation. The corresponding MR images at 8, 15, and 22 days had a lighter area around the gels, whether or not they contain contrast agent. In addition to H&E staining, both Mac-3 staining for macrophages, and Masson’s trichrome staining for collagen all demonstrated a significant immune response to the hydrogels. Over time, there is greater macrophage presence (indicated by a brown color) and the formation of a collagen capsule (indicated by a blue ring), as demonstrated with histological images at day 22 (Figure 7j–l).

Discussion

MRI has the potential to non-invasively image biomaterials if sufficient contrast difference between the biomaterial and the surrounding tissues is evident. We have demonstrated that incorporating PPCAs into a hydrogel provides enhanced contrast over the in vivo lifetime of the materials. We have shown the ability to track the hydrogel, quantify degradation, and detect and observe an immune response over time. Although the hydrogel is initially visible without PPCA, it becomes indistinguishable from the surrounding tissue after degradation. Therefore, it is necessary to include PPCA to image implanted hydrogels for in vivo fate mapping.

The K8-120 PPCA used in these experiments, a bioconjugate of a protein polymer with Gd(III) chelators covalently attached through lysine side chains, has previously been characterized (23). The protein polymer contains 120 repeats of the amino acid sequence, GKAGTGSA, providing 121 (120 lysines in the repeats and one in a 10x histidine tag) possible sites for the hepta-coordinated Gd(III)-DO3A chelator. An amide bond formation reaction between the free amines on the lysine side chain and a carboxyl group on the chelator yields the conjugate. Since the conjugation reaction is performed in an aqueous buffer, there was a low of efficiency that resulted in only 39 chelators in the conjugate PPCA.

Although the uncrosslinked K8-120 PPCA had a very high relaxivity in water at 1.5 T and 37 °C, it was difficult to predict a priori how much contrast would be evident in a hydrogel at room temperature and 4.7 T. As demonstrated in the in vitro hydrogel images, the K8-120 PPCA enhanced the contrast within an enzymatically crosslinked protein polymer-based hydrogel when imaged at 4.7 T and room temperature (field strength, temperature, macromolecular content, and the effects of crosslinking into a hydrogel all affect contrast in a MR image).

Both the high concentration of the protein polymers and the resulting elevated viscosity affect observed contrast because macromolecular content and greater viscosity both increase relaxivity (2,27). As shown in Figure 1 and Table 1, the hydrogel without contrast agent has significant contrast and a relatively low T₁ and is most likely due to the high protein content and the gel properties. The relaxation properties of the PPCA in a gel can differ in solution due to multiple factors, including rate of water exchange, water access, and rotational correlation time. The contrast observed in the phantom images correlates well to the T₁ relaxation times. The PPCA-containing hydrogel’s T₁ is approximately half that of the non-PPCA containing hydrogel, however both values are significantly lower than water (T₁ ≥ 2500 ms).

Imaging the hydrogels before and immediately after implantation shows that the T₁ values increase between the in vitro and in vivo placement. The T₁ of both hydrogels increased approximately 27% on implantation and an additional 30% one day later. The T₁ of the hydrogel with PPCA was consistently lower, with the ratio of the unlabeled to labeled hydrogel T₁s remaining at approximately 2. Changes in temperature and solvent and solute flux may explain these differences in T₁ over time. Other data indicate that degradation and inflammation are not significant after 1 day and should only minimally affect T₁ values.

The results demonstrate the ability to track hydrogels in vivo over time by incorporating PPCAs. Imaging mice every 3–5 days allowed preparation of a detailed degradation curve by calculating the graft volume at each time point. These data were used to choose appropriate imaging times for subsequent studies to capture initial, intermediate, and end points of the apparent degradation curve. The gel volumes were tracked using the hydrogels with PPCA. Although both hydrogels have enough contrast at earlier time points, the gels without PPCA are difficult to distinguish from surrounding tissue at later time points. We have assembled a degradation curve using volumes averaged across all mice with relatively low standard deviations. The average volume at each time point remains statistically greater than the subsequent volume (P < 0.01).

The degradation rate of biomaterials is a critical parameter in tissue engineering and influences the tissue morphogenesis since the physical properties of the hydrogel change as it breaks down. Both mesh size and swelling increase as degradation progresses, which concomitantly influence diffusion and mechanical strength (28). Several studies have highlighted the importance of hydrogel degradation rate on the formation and assembly of extracellular matrix (ECM) and structurally sound tissues (29–31). Ideally, the material degradation synchronizes with the tissue growth; a deposited ECM can replace the biomaterial (32). The degradation rate can also have an impact on the immune response through changes in shape, porosity, release of harmful degradation products, and surface roughness (33). For in situ gelling materials, it is also advantageous to be able to track, spatially, the initial deposition of the hydrogel (34).

The importance of including the PPCA within the hydrogel is illustrated by the difference in CNR. At each time point, the CNR for PPCA-containing hydrogels is at least 1.6 times greater than non-PPCA containing hydrogels. By 22 days after implantation, the CNR without PPCA is so low that it is not possible to distinguish between the hydrogel and the surroundings. By incorporating the PPCA, the difference in signal strength between the hydrogel and the surrounding tissue is augmented, providing more distinct boundaries to properly identify the hydrogel.

The contrast difference is confirmed by the analysis of T₁ relaxation times of the hydrogels and the surrounding tissue. Throughout the duration of the study, the T₁ of the PPCA- containing hydrogel is significantly different from the surrounding inflamed tissue. However, without the PPCA, the T₁ of the hydrogel and the surrounding area are the same by day 22.

At days 8 and 15, the T₁ of the surrounding tissue is greater when no PPCA is present, so it is possible that some degraded PPCA has diffused out of the gel, lowering the T₁ of the area adjacent to the PPCA-containing hydrogel. The difference in T₁ over time also indicates how the degradation affects the PPCA within the center of the hydrogel, since the T₁ increases over time. As demonstrated previously, the high relaxivity of the PPCA is due to the long length of its protein backbone; thus degradation would increase the T₁ (23).

In the assessment of all parameters, it must be noted that since it becomes difficult to distinguish the hydrogel, the measured regions of interest may not be accurately defined without PPCA. This shortcoming also highlights the necessity of the PPCAs.

MRI has previously been used to image the host response of a biomaterial, correlating relaxation times and signal intensity to an inflammatory response (2,9–11,35). However, to the best of our knowledge, this is the first time that the foreign body response of a biomaterial has been detected using a Gd(III)-based contrast agent. By shortening the T₁ of the hydrogel, the PPCA distinguishes the implant from the surrounding inflammatory tissue. Histology shown here confirms that a ring surrounding the protein polymer hydrogels in MR images is due to an immune response that is characterized by polymorphonuclear leukocyte infiltration and a fibrinous exudate (Figure 7).

As illustrated in the histology and corresponding MR images, there was a significant inflammatory response to the protein polymer hydrogels both with and without PPCA. Since the PPCAs were not purified for endotoxin removal, it is possible that the small amount of endotoxins contributed from the PPCAs increased the immune response in the PPCA-containing hydrogels. However, this potential effect was overwhelmed by other influences. These results indicate that these hydrogels are immunogenic in the formulation used in these experiments.

This immune response persists through the end of the study and later time points show increased macrophage presence and the formation of a collagen capsule. These results are typical of a foreign body response to a biomaterial (36). As demonstrated here, the PPCA enables clear distinction of the hydrogel from the surrounding inflammatory tissue. The biocompatibility of a biomaterial is a critical property, and discriminating between the hydrogel and the tissue in an MR image improves the accuracy of the assessment.

Conclusion

The inclusion of the K8-120 PPCA in protein polymer hydrogels enables in vivo analysis of critical properties of the biomaterial over time through MRI. Fate mapping, quantification of degradation, and detection of a foreign body response are facilitated through the incorporation of PPCA. By displaying a CNR that is two-fold greater than hydrogels that do not contain PPCA, the PPCA-containing hydrogels can be distinguished from the surrounding tissue throughout the duration of the study. Without PPCA, hydrogels have similar signal intensities to surrounding tissue at the end of the gel lifetime. T₁ relaxation times support CNR data. This is the first report of use of a Gd(III)-based CA for detection of foreign body response to biomaterials, an essential property to assess before further material development. With incorporation of PPCAs, biomaterial improvement can be advanced through non-invasive in vivo assessment.