Reduced fibrous capsule elastic fibers from biologic ECM-enveloped CIEDs in minipigs, supported with a novel compression mechanics model

BACKGROUND: Fibrous capsules (Fb) in response to cardiovascular implantable electronic devices (CIEDs), including a pacemaker (P) system, can produce patient discomfort and difficulties in revision surgery due partially to their increased compressive strength, previously linked to elevated tissue fibers. OBJECTIVE: A preliminary study to quantify structural proteins, determine if biologic extracellular matrix-enveloped CIEDs (PECM) caused differential Fb properties, and to implement a realistic mechanical model. METHODS: Retrieved Fb (-P and -PECM) from minipigs were subjected to biomechanical (shear oscillation and uniaxial compression) and histological (collagen I and elastin) analyses. RESULTS: Fb-PECM showed significant decreases compared to Fb-P in: low strain-loss modulus (390 vs. 541 Pa) across angular frequencies, high strain-compressive elastic modulus (1043 vs. 2042 kPa), and elastic fiber content (1.92 vs. 3.15 μg/mg tissue). Decreases in elastin were particularly noted closer to the implant’s surface (Fb-PECM = 71% vs. Fb-P = 143% relative to dermal elastin at mid-tangential sections) and verified with a solid mechanics hyperelasticity with direction-dependent fiber viscoelasticity compression simulation (r2 ≥ 98.9%). CONCLUSIONS: The biologic envelope composed of decellularized porcine small intestine submucosa ECM for CIEDs promoted fibrous tissues with less elastic fibers. Novel compression modeling analyses directly correlated this singular reduction to more desirable subcutaneous tissue mechanics.


Introduction
The formation of fibrous capsules is a foreign body reaction (FBR) against nondegradable materials, long-term implanted into the host system, dependent on factors including the material's physical and chemical surface properties [1][2][3][4]. This tough tissue is induced to wall-off and systemically-isolate foreign objects (like cardiovascular implantable electronic devices (CIEDs) [5] with titanium housing and their polymer-coated leads) and is a general indicator for implant biocompatibility: the thinner the layer, the more compatible the material [3,6,7]. The fibrous capsule (also called fibrotic scar tissue or fibrosis) is distinct from other tissue types in the subcutaneous space and is mainly composed of compact collagen fibers produced by activated fibroblasts. The capsule can also contain layers of loose granulation tissue, immune cells, vessels, and a variety of enzymes and signaling molecules [1,8,9]. Aside from collagen (mainly type I, with occasional types II and III) [10][11][12], fibrous capsules have other extracellular matrix (ECM) structures, most notably elastic fibers [1,9,13,14]. Elastic fibers are responsible for elastic recoil after stretching and compression for reversible recovery after deformation caused by applied loads, particularly beneficial for skin and blood vessels [15][16][17]. However, in fibrous capsules surrounding medical implants, their presence in excess may be detrimental to the tissue mechanics. An elastic fiber's major component (∼90%) is elastin. Initially, the relatively small (M w ∼ 60 kDa) soluble precursor tropoelastin is secreted by cells (such as fibroblasts), which then assembles and is crosslinked (forming covalent bonds) into elastic fibers via elastogenesis. Lysyl oxidase (and related enzymes) oxidize lysine residues for crosslinking with other tropoelastins and ECM proteins including collagens.
The fibrous capsule FBR is dynamic (time-dependent) and constantly remodeled depending on the stimuli. CIED-induced fibrous tissues can be used to an advantage: to subcutaneously affix and prevent device migration [18][19][20][21]. Conversely, too much fibrosis causes unwanted scar appearance, local tissue hardness, high compressive strength, and difficulties during CIED generator changes and pacing lead revision surgeries [22][23][24][25][26]. In patients with silicone breast implants, contracture, which induces increased stiffness, discomfort, and pain, can be attributed in part to increased elastin expression [13,16,27]. Similarly, issues encountered with CIEDs may be due to the overabundance of elastic fibers. Biologic ECM envelope products composed of decellularized porcine small intestine submucosa (SIS) were developed [28][29][30] to potentially mitigate the fibrous response. Besides acting as a cover for the CIED surface, SIS ECM implants are rich in bioactive growth factors, glycosaminoglycans, and structural proteins that can affect the subcutaneous microenvironment and FBR [31][32][33][34][35]. To investigate their underlying effect, this study compared fibrous capsule biomechanical and histological properties and associated elastic fiber distribution from pacemakers without and with biologic ECM envelopes in a shortterm minipig subcutaneous model. A computer model was proposed and fitted to explain and verify the observed compressive stress-strain responses and correlate them to elastin amounts and arrangements.

Animals and fibrous dissection
Surgical procedures and animal care were conducted at an independent research organization (American Preclinical Services, Minneapolis, MN, USA). Protocols were approved by the Institutional Animal Care and Use Committee (IACUC), and animals received humane care in compliance with NIH guidelines [36]. Human pacemakers (pulse generators) with an attached, coiled pacing lead (Edora 8 SR-T and Solia JT 45, Biotronik, Lake Oswego, OR, USA), without (P) and with decellularized SIS ECM (PECM) envelopes (CanGaroo ® Envelope (medium: 6.9 cm × 6.5 cm = area of 44.85 cm 2 ), Aziyo Biologics, Silver Spring, MD, USA) were implanted subcutaneously (at 8 per group) in contralateral ventral neck-chest areas of adult miniature pigs (minipigs), then sacrificed and necropsied for a total of 3 months duration (based on [19]). Implants that failed before the endpoint were excluded from this study. After pacemaker retrieval ( Fig. 1), the surrounding fibrous capsule (Fb) and adjacent adipose (Ad) with occasional underlying muscle tissue were excised, trimmed, and assigned as groups (Fig. 2): Fb-P (fibrous tissue from pacemaker only, n = 5) and Fb-PECM (fibrous tissue from pacemaker in ECM envelopes, n = 2), with Ad (adipose, n = 4) control, where n = biological replicates (individual animals). Multiple technical replicates (≥3) were obtained per biological replicate per test to account for individual variability [37]. These groups and sample sizes were used in subsequent experiments.
Images were captured in Cytation 5 (Agilent, Santa Clara, CA, USA) and processed with ImageJ (NIH, Bethesda, MD, USA). IHC area ratios were quantified in thresholded and leveled images using ImageJ's analyze particle area sum over polygonal tissue area selection. Individual slide (-) signals were subtracted.

Modeling of compression
COMSOL Multiphysics (COMSOL, Stockholm, Sweden) using the solid mechanics interface with nonlinear structural materials module was employed to generate a 3D time-dependent compression model matching the experimental using parameters: Fb (E = 295 kPa (determined through parametric sweep), density ( ) = 1.05 g/mL, and Poisson's ratio ( ) = 0.4999) and EF properties (E = 400 kPa (calculated based on [50]) and = 1.2 g/mL). Fb was assumed as neo-Hookean hyperelastic incompressible [51]. EFs were modeled with direction-dependent (tangential and axial) domain node fibers (linear elastic with a volume fraction or ratio (r v )). The tissue was divided into sections (in, mid, and out) and each assigned with fibers with viscoelasticity of a single-branch generalized Maxwell containing: energy factor (f e ) and relaxation time (t r ) [50]. A rigid indenter at 0.1 mm/s prescribed velocity compressed Fb with constrained bottom boundary to 40%. Unstructured uniformly-distributed quadrilateral meshes along h were used. Average of the domain's top boundary in contact with the probe was evaluated for stress = Gauss point von Mises stress and strain = displacement/h.

Data analysis
Data were processed in Excel (Microsoft) and MATLAB (Mathworks, Natick, MA, USA). Technical replicate values were averaged, and biological replicates' values again averaged then reported as mean ± standard deviation. Student's t-test and analysis of variance (ANOVA, 1 and 2-factor) with Tukey-Kramer were used for comparison using a probability (p) < 0.05 deemed significantly different and represented as: ***p < 0.001, **p < 0.01, and *p < 0.05.

Effect of compression
Macroscopic thicknesses (for strain calculation) using a caliper yielded 4.0 ± 1.1 for Fb-P and 3.3 ± 0.5 mm for Fb-PECM, but still statistically similar (p = 0.59).
Compression to high strains (some up to ∼60%) led to reversible deformation within the elastic region even with nonlinearity of the -curves. Ad had drastically lower and its derivative (d ∕d ) compared to both Fb's ( Fig. 4a-b), which rendered to lower (***p < 0.001 at 0-10% and at least *p < 0.05 at 30-40% ) compressive E (Fig. 4c). It is remarkable that Ad and Fb-PECM's curves had generally similar semi-log shapes (with adipose just translated downwards), while Fb-P's exhibited a distinct pattern (Fig. 4a) and more noticeable in numerical differentiation (Fig. 4b blue curves). Quantitatively at high (30-40% or higher), Fb-P had significantly greater (*p = 0.026) slope or stiffer E than Fb-PECM (Fig. 4c, Table 1).

Histology analysis
Sections of the samples from both Fb groups showed collagen secreted by activated fibroblasts, filling the entire thickness with interspersed blood vessels (Fig. 6). Fb-PECM stained their cell nuclei faint and ECM lighter even after several repeats, perhaps due to differences in chemistry. Additionally for Fb-PECM, there were no signs of remnant ECM envelope suggesting full remodeling of the SIS. No foreign body giant cell was noticed indicating the absence of chronic inflammation. In some sections, adipose tissue can be seen in the mid to outer zones, which indirectly indicated fibrous formation originating from the implant's surface. Regions with hematoma demonstrated reddish coloration in MT with round nuclei and brownish in VGE, likely as macrophages and blood, found in the inner surface.
Elastic fibers (EFs) were detected as black strands and visually more present in cross-sections of Fb-P, especially in regions closer to the inner surface ( Fig. 6 VGE inset). Immunostaining confirmed higher (*p = 0.011) differentially-expressed EF elastin in Fb-P vs. Fb-PECM (Fig. 7, Table 1, 1.49 ± 0.12 and 0.98 ± 0.25 relative to dermis control levels). In tangential cuts (Fig. 7a boxed), there appeared to be more elastin in Fb-P (and in dermis) vs. Fb-PECM at inner and mid zones. Statistically, Fb-P > Fb-PECM (*p = 0.024) at mid-tangential (Fig. 7b, Table 1). Fb sections showed various stages of elastogenesis,  ranging from diffuse elastin staining to assembled EFs with different orientations. The biologic ECM envelope displayed lower (*at least p < 0.05) background elastin compared to others.
Collagen type I is prominent among all tissue groups: Fb-P, Fb-PECM, ECM envelope, and dermis controls were found to be statistically similar (p = 0.81) (Fig. 8). Its density was qualitatively more intense in bundles observed in the dermis (Figs 6, 8a). Conversely, collagen I in fibrous capsules did not appear bundled but as separate fibers, suggesting less organization and maturation. EFs did not colocalize with collagen I in Fb samples.

Computer simulation
Simulation with 6-fold axially-oriented fibers at the inner and mid-Fb (in + 1/3 mid) relative to 12% r v of tangential whole-tissue fibers (with consistent EF properties) resulted in a compressive -curve that agreed well (r 2 = 99.7%) compared to experimental Fb-P observation, including elevated von Mises stress seen across tissue equivalent to higher compressive strength (Fig. 9a). When these axial elastic fibers were removed (at 0 fiber, Fig. 9b-c), the plot fitted accurately to Fb-PECM experimental (Fig. 9d, r 2 = 98.9%). Different iterations of domain material models were simulated (Fig. 9e), but ultimately the hyperelastic with regional viscoelastic direction-dependent fibers showed the best for modeling Fb-P and Fb-PECM which agreed with experimental trends relating compression (Figs 4a, 9d) to differential elastic fiber expression (Figs 5b, 6, 7) using the viscoelasticity parameters: f e = 2% and t r = 50 s.

Biologic ECM envelopes mitigate elastic fiber formation and fibrous tissue stiffness
The fibrous capsule response was influenced by host contact to implant surface materials: titanium housing of pulse generator, epoxy resin of connector receptacle, silicone rubber of sealing plug, silicone rubber and polyurethane cover of pacing lead, and decellularized porcine SIS ECM of the envelope [28,52]. It is noted that pacing leads were coiled around the pulse generator in this study (together influencing the fibrous capsule formation), while in clinical usage, leads are extended and implanted towards the heart muscles. In this scenario, the pulse generator and pacing leads will induce separate fibrotic responses. Fibrous capsules are generated from activated fibroblasts which can originate from different locations within the skin including as hypodermal fibroblast residing in the adipose tissue [53].
Macro and micro-histologic features of recovered and processed fibrous tissues, without or with the biologic ECM envelope treatment, indicate differing degrees of FBR biocompatibility, previously observed in a rabbit model [54]. As expected from prior animal studies, both implant groups induced varying levels of fibrous tissues prominent in collagen I [10][11][12]55]. Despite similarities in appearance, thickness, and collagen type I morphology and content, we found five properties with statistically less magnitudes from Fb-PECM samples, summarized in Table 1. Compared to biopsied tissue thickness measurements of 4.0 and 3.3 mm for Fb-P and Fb-PECM, respectively, histological samples (at 2.7 and 2.2 mm) were apparently thinner because of shrinkage during the ethanol dehydration steps.
Changes in biomechanics at specific strain conditions, as seen at low-shear oscillation and highcompression, can be explained by the anisotropic [56] levels of tissue elastin, which constitute the mature functional form of load-bearing elastic fibers. Elastic fibers are abundant in the capsular tissue from several other different subcutaneous implant studies [1,9,13,14]). High-strain E of Fb-P tissue samples compressed at ∼30% were well within the 1.6-5.7 MPa of silicone implant fibrous capsules collected from breast capsular contracture patients [13]. E values for Ad control tissue fell within the acquired spread from human adipose [57], supporting that our tests were sound and accurate. Use of the biologic ECM envelope significantly lowered Fb-ECM levels of elastin and elastic fibers versus Fb-P, particularly in tissue regions close to the pacemaker and leads. The reduction strongly correlated with a drop in E below Fb-P and the previously reported range of undesirable tissue from breast capsular contracture. Since this study only had a short 3-month duration, mechanical differences between the Fb-ECM and Fb-P groups should widen over time as elastic fibers continue to accumulate at a faster rate in the Fb-P group, pushing the mechanical properties to match tissue from the most severe cases of contracture. These data suggest biologic ECM envelopes could be used to alter the FBR for subcutaneous implants and mitigate excess accumulation of structural components, like elastic fibers, in the fibrous tissue that increase tissue elastic modulus and cause contraction.
Reducing tissue stiffness following subcutaneous implantation surgeries may improve patient comfort and long-term clinical outcomes. For patients receiving CIEDs, a hard fibrotic tissue can cause lead dislodgment [58,59]. Tissue with higher stiffness values encasing generator and/or leads also increases the difficulty of revision procedures and contributes to higher complication rates observed for revision procedures compared to those reported for de novo implantations [24,25,60,61]. Using biologic ECM envelopes to wrap CIEDs and modify the FBR could be an effective method to reduce these complications and improve patient health.

Novel compression modeling effectively simulates fibrous tissue from CIED FBRs
In compression testing of fibrous capsules reported in previous studies, excess collagen levels confounded the mechanical effects of elastic fibers. With statistically similar levels and orientation of collagen between this study's two tissue groups, the role of elastic fibers could be isolated in computer modeling.
In the proposed hyperelastic model with viscoelastic fibers (Fig. 9), Fb-PECM was set as the starting fibrous tissue baseline, assuming at the early timeline of Fb-P's development that ultimately produced more elastic fibers. For this domain baseline, the simulated compressive stress-strain curve agreed with experimental once elastic fiber orientation was assigned as force-orthogonal (tangential or radial) and its tissue volume fraction to 12%, within acceptable degree of variability to the observed 8.6% (in crosssection immunohistochemistry). Addition and successive increase of force-parallel (axial) fibers to 6 times greater in the inner 44% region then created the accurate Fb-P compression model. The fiber ratio ultimately plateaued to around 7-fold maximum capacity, indicating peak resemblance to the dermal control and correlating with the similarities in elastin quantity and histology, which possibly signifies a similar tissue genesis. It is noted that fibers placed closer to the implant provided better simulation outcome with higher stress-strain influence than the mid and outer zones. Overall, evidence considering statistics, potential sources of error, visual, and modeling analyses suggest that more elastic fibers are found toward the CIED surface in Fb-P. This also resulted in higher loss modulus taken in the rheology experiment due to elastic fiber-dependent increased water absorption and tissue retention [16,62].
The biological importance of elastic fiber's localized expression and orientation is highlighted in a few skin studies involving normal versus diseased or altered histological states. In healthy young skin, elastic fibers are normally arranged perpendicular (or axial) to the surface, specifically in the papillary or upper dermis directly underneath the epidermis [63,64]. Aging caused significant decrease in these organized structures which led to changes in skin's biomechanical performance [63]. Malignant melanoma associated with fibrosis shifted and increased the quantity of elastic fibers, instead, more towards the reticular or lower dermis, with prevalent parallel (or tangent) directionality [64]. Similarly, in patients with morphea, characterized by thick and hardened skin patches, there is an increased parallel elastic fibers in the reticular dermal tissue [65], possibly for stronger shear resistance. The organizational orientation of groups of ECM fibers is naturally to resist directional loads. The inferred trend in the dermis-like fibrous capsule, supported by our computer model (in Fb-P), is that the closer to the hard surface (such as the titanium pulse generator CIED case), the higher the axially-directed elastic fibers to withstand the dominant compressive stimuli. Addition of the polymeric SIS (in Fb-PECM) that absorbs water provided damping and shock-absorbing effects, lessening the need to resist compression and thus the amount of surface-perpendicular elastic fibers.

Conclusions
CIEDs wrapped in the biologic ECM envelope reduced elastic fibers in the fibrous capsule region close to the implant, which directly led to a more subtle biomechanical response including lower compressive strength with some likeness to surrounding adipose tissue. Novel mechanical modeling highlighted the potential of this envelope for improved clinical outcomes of CIED implantations. A long-term study with more biological replicates for investigating elastic fiber anisotropy and associated cells and structures can be conducted in the future.