4.1. Intraindividual Comparative Proteomic Profiling in Plasma, Wound, and SMI-Adhesive Proteome
Any implanted foreign body elicits an immediate inflammatory response resulting in either fibrous encapsulation by excessive ECM turn-over or incorporation of the implanted biomaterial [
57]. The controlled method of sample collection, as well as the analysis used, is an integral step in biological significance in the diagnostic research process of capsular fibrosis etiology.
Blood biomarkers are widely applied for the screening and diagnosis of inflammation because of their availability. However, the extent of the soft tissue reaction surrounding the implant depends on several factors, although to study wound infection directly, wound bed fluid is the prerequisite sample to choose. Wound bed fluid is not only tied to the wound bed environment but is easily available. Proteins in wound fluid obtained by a sterile, closed-suction drain placed in the subcutaneous tissue (following mastectomy) [
47] are not only exposed to proteinase 3 [
58], but also to plasminogen activity [
59]. Moreover, drain collection from day 1 to day 5 post-implantation in flasks stored at room temperature establishes a 24 h time window of proteomic degradation.
The identification of protein repertoires in the wound bed fluid directly after SMI implantation is notoriously challenging for the discovery of potential biomarkers for the switch and progression from the normal healing response towards chronic fibrosis and the excessive encapsulation of the silicon device. Moreover, protein adsorption at the implant surface is a key driver of local reactions to silicone; thus, we aimed to identify proteins that surround the device non-specifically in the acute wound and adhere to the silicon surface.
In this study, we investigated the proteomes of pre-operative plasma (−1 d), wounds that are derived from surgical drainages following NSME (d1–d5) and associated with/adhered to the SMI surface approx. 8 months post-op.
The etiology of this exaggerated capsule formation is multifactorial but primarily induced by immune mechanisms towards the foreign material silicone. Performed in vivo for the first time in human patients, we decided to intra-individually identify, characterize, and quantify the global FBR against SMI, common to two commercially available tissue expanders with varied surface roughness and topography: the
SMIa with reduced surface roughness (4 µM of Ra; SmoothSilk
®, Motiva) and conventional
SMIb (60 µM of Ra; CPX
® 4, Mentor). Sample collection was performed simultaneously for both devices in a sterile environment from the same individuum at the same time point (
n = 7) (
Figure 1 and
Figure 2). Here, we provide insight into a set of diagnostic biomarkers applicable to silicone devices with varying surface topography roughness.
4.4. Pathogen Binding and Activation of Inflammasome in the Wound
Immediately following the placement of a breast implant, the destruction of the tissue initiates the host response to reactive oxygen species (ROS), nitric oxide, and toxic and mechanical stress [
11,
29]. The first remarkable result to emerge from the data is 34 plasma-derived DEPs with high significance, such as superoxide dismutase (periredoxin, extracellular and mitochondrial) and catalase (
Supplementary Table S9), which are actively involved in the oxidative response to NO and ROS. However, we found no expression of additional ROS or NO disbalance responders in the local wound compartment.
The increased expression of plasma-derived PAL-1 during the first five days after implantation was observed. This factor plays an essential role in blood coagulation, by inhibiting the conversion of plasminogen to plasmin, thereby preventing fibrinolysis [
62,
63]. Moreover, we observed elevated levels of plasma-derived HBB, HBA2, MB, HBD, HBG1, and, locally expressed in the wound, HBG2, which reflects the first response to tissue damage, bleeding, and clotting. Damage to the epithelial cells leads to the aggregation of erythrocytes and platelets that form a blood clot to contain the spread of damage.
The destruction of the tissue and accumulation of altered self-components as well as reactions to noninfectious foreign material triggers inflammation and the release of damage-associated molecular patterns (DAMPs), or allows the entrance of microorganisms and thereby pathogen-associated molecular patterns (PAMPs) [
29]. To our surprise, we found that PGLYRP1 was differentially expressed in the local environment. This 196aa innate immunity protein acts as a bactericidal pattern receptor that binds to murein peptidoglycans (PGN) of Gram-positive bacteria [
64]. Moreover, in the present investigation, we could confirm the abundance of LBP, BPI, CFL1, CAMP, LYZ, and B2M in the wound. Fifteen plasma-derived DEPs with high significance, such as CAMP, LTF, DEFA1B, and ITGB2, are actively associated with host–pathogen interaction and antimicrobicidal response. CAMP, a 177aa cationic antimicrobial protein with a molecular mass of 18kDa, is an acute-phase reactant that binds to bacterial lipopolysaccharides (LPS) [
65,
66] and acts via neutrophil N-formyl peptide receptors to enhance the release of CXCL2 [
67].
4.5. FBR and Inflammation: The Response in Inflammatory Matrisome
Local cells such as dendritic cells and macrophages are activated by DAMPs and PAMPs through toll-like receptors (TLRs) and protease-activated receptors (PARs), and they produce the first round of pro-inflammatory mediators [
11]. Interestingly, thePGRP1, which is detected in the local wound environment, acts in complex with the Ca
2+-binding protein S100A4 as a chemoattractant that is able to induce lymphocyte movement [
68]. The abundance of plasma-derived S100A4, as well as differentially expressed CD44 and MNDA, could be confirmed in the wound environment (
Supplementary Materials Table S8).
In our study, strong evidence was found in the GO enrichment analysis of plasma-derived and local wound proteins, revealing highly significant involvement in the biological processes of neutrophil, granulocyte, and leucocyte activation, mediation, and degranulation (
Figure 4a), which confirms an infiltration of neutrophils and monocytes that respond to the pro-inflammatory stimuli at the implant surface. Phagocytes such as neutrophils and macrophages remove tissue debris and potentially threatening particles, whereas neutrophil and granulocyte degranulation induce the first inflammatory storm [
11,
29].
The granules of neutrophils contain numerous enzymes, such as matrix metalloproteinases (MMPs), elastase, and cathepsins that can specifically cleave collagenous and non-collagenous connective tissue components that are mandatory for pro-fibrotic ECM turn-over. Correspondingly, we identified plasma-derived DEPs involved in ECM remodeling, such as ELANE, leucocyte elastase inhibitor (SERPINB1), and NCF1B, as well as MMP2, MMP8, and MMP9 (
Supplementary Materials Table S8). Moreover, neutrophils play an indirect role by activating further innate cellular components [
30,
60], reflected in highly significant DEPs such as CSF1R and BCAP31 (
Supplementary Materials Table S8).
In a pro-inflammatory environment, activated monocytes and resident macrophages(M1) produce pro-inflammatory mediators such as IL-6, TNF-α, and IL-1 [
69]. We detected IL1RN, IL1RAP, IL1R2 in plasma (
Supplementary Table S8), and IL6ST in local wound proteome (
Supplementary Materials Table S14). Our findings confirm the presence of activated macrophages in the wound environment that produce IL-1 to activate fibroblasts and induce the overproduction of ECM proteins. Although we did not identify TNF-α components in the wound proteome, we detected TNF-α-triggered signal transduction via the NF-κB pathway, which was reflected in the local expression of tumor necrosis factor receptor superfamily member 10C and NFKB1 (
Supplementary Materials Table S14).
Remarkably, traces of antifibrotic effected by interferon (IFN)-γ produced ls were visible in plasma-derived and local interferon-inducible components IFIT3 (
Supplementary Table S8), IFIT2, IFI30, IFI35, and IFI16 (
Supplementary Materials Table S14). Effective healing is usually characterized by a dominant T helper 1 (Th1) cell response, whereas a predominant T helper 2 (Th2) response and an increase in T helper 17 (Th17) cells lead to chronic inflammation, which can ultimately result in fibrosis [
29]. On the other hand, Th2 cells mediate the adaptive immune response to injury by producing pro-fibrotic (anti-inflammatory) cytokines (e.g., IL-4, IL-13, IL-10); Th1 cells mediate tissue damage response by Th1-related proinflammatory cytokines (IFN-y, IL-12) that suppress fibroblast-induced collagen synthesis and attenuate fibrosis [
70]. In addition, IFN-γ production up-regulates the expression of matrix metalloproteinases (MMPs), whose proteolytic activity helps alter ECM remodeling. As a commonly recognized opponent of Th1 cells, Th2 cells can alter Th1-associated IFN-γ expression levels, and high levels of Th2 cytokines have been reported in several fibrotic diseases [
70].
Taken together, we obtained comprehensive results demonstrating two main implant-induced pro-fibrotic progressions—a systemic (plasma-derived) proinflammatory response and local proinflammatory mediation in the wound. The wound bed proteome analysis post-silicon implantation provides detailed insights into wound healing stages characterized by various inflammatory matrisome proteins.
4.6. From Wound to Fibrosis
In the course of wound healing, inflammatory responses subside and turn into repair responses through the effects of anti-inflammatory and pro-repair cytokines. In accordance, we found transforming growth factor-beta-induced protein (TGFB1,
Supplementary Materials Table S8) as a compound of the plasma-derived protein and locally expressed late-transforming growth factor binding proteins LTBP4 and LTBP2, as well as LTBP1 (
Supplementary Table S14), the latter being differentially expressed over the first five days after SMI implantation (
Figure 5g,
Supplementary Materials Table S15). LTBPs are involved in fibroblast proliferation and migration; thus, with the finding of plasma-derived SLC9A3R1 and local CRYAB, we provide further evidence for the stage of fibrogenesis and fibroblast proliferation into the more contractile myofibroblasts [
29]. Myofibroblasts produce an extracellular matrix (ECM) to close the open wound and form a scar [
71]. As already reported, active ECM turnover could primarily be suspected through the detection of MMP2, 8, and 9, whereas MMP8 was upregulated in the wound over the total first five days post-surgery. Further, we identified and characterized 71 plasma-derived and 30 local wound proteins as core matrisome, as well as 142 plasma-derived and 48 local wound proteins as matrisome-affiliated proteins [
72,
73]. Strikingly, we could identify not only pro-fibrotic marker COL1A (plasma-derived;
Supplementary Materials Table S9) and ELANE as being upregulated over the first five days post-op, but also the fibrosis drivers S100A8 and S100A9 derived from plasma. Moreover, COL6 and COL3 were detected; however, these were not significantly increased over the total 5-day period (
Supplementary Materials Table S8). The set of profibrotic markers was supplemented with the locally expressed keratins (
Supplementary Table S15: KRT-8, -36, -82, -75, -76 and -79), as well as collagen alpha chains (
Supplementary Materials Table S15: COL-12A1, -14A1, -15A1, and -18A1).
Regularly, in the last phase of repair excess, ECM is degraded for the incorporation of new cells as the tissue regains its structure and function. However, in fibrosis, aberrant wound healing results in the excessive accumulation of collagen and ECM components.
To identify potential biomarkers for the switch and progression from the normal healing response towards chronic fibrosis and the excessive encapsulation of the silicon device, we further focused not on serum, as previously reported [
20,
74], but on wound bed fluid proteins associated with both silicone surfaces and that were still adhered to them after 8 months lingering in the body under pro-inflammatory and pro-fibrotic conditions.
Our data reveals not only a common SMI-associated proteome (
Figure 6), independent of surface topography, after 8 months residing in a human body, but also a common implant-enclosing wound proteome adhered to SMI surfaces (
Figure 7). Enrichment analyses of the common (on
SMIa and
SMIb) adhesive proteome confirmed the main involvement in ECM turnover reflecting early-stage fibrosis.
Aside from the wound-derived proteome, we identified an additional 54 proteins, which were found to be adhered to SMI surfaces but not detected in wound bed fluid. This early-stage fibrotic set of proteins was characterized by several 60 S and 40 S ribosomal proteins, a major group of proinflammatory mediators [
75,
76,
77]. Moreover, we detected CAPG and CD9 on SMI surfaces. This favorably correlates with a recent study in fibrotic adipose tissue, where Rabhi and colleagues [
78] showed that CD9
+ senescent macrophages activate a fibrotic transcriptional program in adipocyte progenitors.
Strikingly, among the SMI adhesive proteome, we found various collagens and S100A family members (
Supplementary Table S17), all previously identified as part of the plasma-derived inflammatory matrisome in the acute wound. In healthy epidermal tissue, collagen type I is more prevalent compared to type III, which comprises about 30% of the total collagen content [
79,
80]. Due to ECM turnover during wound healing, COL3 can increase up to 90% and a decrease in type III can lead to excessive scar formation [
71]. In contrast with earlier reports, we did not observe a significant upregulation of COL3 but of COL1 and COL6 in the acute wound (
Supplementary Materials Figure S5). However, COL3 was detected on the SMI surface, significantly upregulated on days 3 and 4 post-op, and we found higher levels of COL3 compared to COL1 in the acute wound proteome (
Supplementary Materials Table S20). Here we manifest the role of COL1 in late fibrosis [
81,
82], moreover, our new data may also provide information on the chronological production of other collagens and non-collagenous proteins.
Unexpectedly, S100A8/A9, a major fibrogenesis marker and fibrosis driver (fibroblast proliferation, differentiation, and activation of collagen production) [
83] was found only in the wound environment, whereas S100A4, A10, and A11 were found to be associated with the silicon surface (
Figure 7 and
Supplementary Table S17). The latter was not surprising, as S100A4 was identified as a fibrosis driver and a useful biomarker for diagnosis and monitoring disease progression [
68]. Our experiments do not only corroborate with previous studies; here, we provide the finding of different roles for the S100A family members in fibrosis, as for example S100A8/9 marks the acute inflammatory profibrotic process, and S100A4 exerts its pro-inflammatory role in early-stage fibrosis 8 months after SMI implantation. Furthermore, multiple processes involved in fibroblast activation on the SMI surface were reflected by the confirmation of HSP90 adhesion to the implant site (
Figure 7 and
Supplementary Materials Table S17). HSPs are involved in the activation or inhibition of epithelial-mesenchymal transition signaling pathways and the generation and degradation of the extracellular matrix to regulate fibrotic diseases; thus, we were not surprised to find HSPA1B, HSPA1A, HSPA8, HSPB1 (plasma-derived;
Supplementary Materials Table S17), and HSPD1(local;
Supplementary Materials Table S18) attached to the surface [
21].
Moreover, we can demonstrate a chronic fibrotic signaling axis, with PGLYRP1, part of the ant microbicidal response in a wound, that forms an equimolar complex with heat shock protein HSPA1A (on expanders) and activates the TNFR1 receptor (expressed locally TNF superfamily 10) from the immediate inflammatory response to chronic inflammation 6–8 months after SMI implantation.
Strikingly, HSP60, a part of the antimicrobial inflammatory response [
84] in the acute wound, was found attached to the silicon shell surface. The stress-induced protein is involved in the bystander activation of T cells. By inducing secretion of the proinflammatory cytokine IFN-γ [
85], it acts directly on activating bacterial HSP-responsive gamma delta T cells [
84]. It can also act as a target for autoreactive hsp60-specific T-cell responses and directly contribute to chronicity at the site of inflammation. Hsp60-responsive T cells are pro-inflammatory cells with an increased secretion of IFN-γ [
85] and low IL-10 profile. Involved in the human chronic inflammatory diseases of Th1 and Th2 cells, it can majorly distract wound healing [
86]. Our values correlate favorably well with a series of reports and further support the idea that HSP60 is an essential homeostatic antigen with both immunoregulatory and inflammatory properties. Its abundance confirms the presence of antimicrobial antigens and response in the wound after SMI implantation. However, the signaling of an endogenous protein expressed in low amounts chronically attached to SMI surfaces may lead to distinct patterns of activation of TRL-expressing cells, such as dendritic cells, macrophages, and T cells.
Derived from wound proteome, the detection of CCT8 (local) and T
REG receptor CD44 (plasma-derived), both adhered to SMI surfaces (
Supplementary Table S18), clearly confirm T cell response at and directed to the implant site. Only scarce data about specific local side effects (local immune response, activity of immune cells) focusing on lymphocytes isolated from fibrous capsules have been reported so far. In a previous attempt, we characterized the cellular composition of fibrous capsules formed around SMIs, by showing that macrophages and fibroblasts were the most predominant cell populations in the region abutting the silicone surface (designated as “pseudo synovium”) [
21]. Strikingly, among T cells, Treg numbers, in peri-SMI fibrotic capsules, were inversely proportional to the degree of fibrosis (Baker scores I to IV). Most interestingly, we showed that Tregs were decreased in those capsules removed from patients with clinically severe symptoms of capsular contracture (Baker scores III to IV) [
21].
Thus, we deciphered the three-dimensional composition of a silicon implant surface associated/adhesive proteome (
Figure 8) by identifying the adhesion of (i) plasma and local tissue-derived proteins in the acute wound and (ii) components expressed later in the early stages of fibrosis. Among these, we identified long-term capsular fibrosis markers after a simultaneous NSME and implant-based breast reconstruction, providing novel diagnostic targets for the long-term tracking of capsular fibrosis and potential capsular contracture.
As also shown here, the wound and SMI-adhesive proteomes of individuals vary greatly; therefore, investigating specific differences in the groups would benefit from larger sets of data to exclude unreliable findings and intensify the differences. Nevertheless, the fact that significant differences still emerged between the different proteomes, despite the limited patient number, illustrates the power of the intraindividual comparison of two different silicone breast implants in detecting subtle qualitative differences in protein patterns. Moreover, the detection of mammary gland-specific breast cancer marker Mammaglobin A in local wound tissue after bilateral mastectomy in all tested breast cancer patients (no tumor diagnosis yet) strengthens our data integrity immensely and confirms the high specificity but applicability of our approach. Although limited, we demonstrate the power of our combined proteomics and bioinformatics approaches in a detailed immunomic picture of capsular fibrosis etiology, from the immediate inflammatory response to early-stage fibrosis.
Finally, we would like to highlight our “secondary finding” of the SMI surface topography exclusive acute wound and adhesive proteome. Biomaterial surface chemistry, mechanical properties, and topography have been shown to influence the immune response [
17,
33] and implant surface-associated biofilm formation, especially due to antimicrobic (antibiotic)-resistant microbial strains, which can lead to chronic immune system activation [
87,
88]. However, depending on the topography of these surfaces, varying degrees of capsular contracture have been reported [
22,
89,
90], and recent studies in rodents and preserved capsular tissue samples have confirmed a reduction in inflammation and foreign body response on implants with an average roughness of 4 μm [
50]. These results propose a further investigation of the effect of surface topography on microbiome and proteome composition after breast implant insertion in a controlled clinical setting with breast cancer patients, as demonstrated here.
Our innovative, intraindividual, comparative molecular identification of biomarkers and their chronological progression generates information of great importance for a basic understanding of the fibrotic side effects of SMIs, in terms of diagnostics prevention, potential new therapeutic approaches, as well as the improved biocompatibility of SMIs.