Extracellular Matrix Remodeling: The Common Denominator in Connective Tissue Diseases: Possibilities for Evaluation and Current Understanding of the Matrix as More Than a Passive Architecture, but a Key Player in Tissue Failure

Collagens

Collagens are a family of proteins made up of three α-chains supercoiled around each other completely or partially in a triple helix with a characteristic Gly-X-Y repeat. Intra- and intermolecular cross-links bring stability to the collagen molecules, contributing to the characteristically high tensile strength and minimal extensibility of collagen. Type I, II, III, and V collagens belong to the group of fibrillar collagens, which are the most abundant collagen group in the body. In addition to the triple-helical domain, they also contain N- and C-terminal propeptide domains that are cleaved off by N- and C-procollagenases, respectively, before fibril assembly.⁴⁷ Type I–VI collagens are the most well described at present and are the focus of this section.

Type I collagen is composed of the heterotrimer α₁α₁α₂(I) and is the most abundant type of collagen that is ubiquitously expressed. It provides tensile stiffness in bone and has important load-bearing, tensile strength, and stress-carrying properties in other tissues as well. In tendons, type I collagen fibrils are arranged in parallel to form bundles, whereas in skin, the arrangement is more random, forming a complex network of interlaced fibrils. These different arrangements contribute to the different properties of the tissues. Type I collagen is often incorporated into fibrils with either type III⁴⁸ or type V collagen.⁴⁹ The synthesis, concentration, and circulating levels (serum concentration) of degradation products of type I collagen have been proven to be increased during breast, bone, lung, ovarian, prostate, and skin malignancy.^50–55

Type II collagen is the major component of hyaline cartilage, but is also found in the vitreous body of the eye, the corneal epithelium, the notochord, the nucleus pulposus of invertebral discs, and embryonic epithelial-to-mesenchymal transitions (EMTs).⁴⁷ Type II collagen is a homotrimer consisting of three α₁(II) chains, and the primary sequence has a high content of hydroxylysine and glycosyl residues, which mediate interactions with proteoglycans, another important component of hyaline cartilage. Type II collagen degradation is mainly associated with rheulatological diseases such as osteoarthritis and rheumatoid artiritis.⁵⁶

Type III collagen is mainly present in association with type I collagen and is an important component of the interstitial tissues of the lung, liver, dermis, spleen, and vessels. Type III collagen is a homotrimer consisting of three α₁(III) chains. A characteristic feature of type III collagen is that it is correlated to extensibility of tissues, and that it may contribute to elasticity, a property that is uniquely connected to this type of collagen.⁵⁷ Type III collagen has been mostly assiociated with various fibrotic diseases.^58–61

Type IV collagen is the main component of the BM, a specialized type of ECM that separates the epithelium from the stroma in all tissues in the body. It consists of three domains: N-terminal 7S domain, a central triple helix, and a large C-terminal NC1 globular domain. Its triple helix is ∼25% longer than those seen in the fibrillar collagens, and the Gly-X-Y repeat is frequently interrupted, accounting for the relatively high flexibility of this type of collagen.²⁸ Instead of fibrils, type IV collagen molecules assemble into a flexible 3-D network. The most abundant isoform of type IV collagen is α₁α₁α₂(IV), but tissue-specific isoforms also exist: α₃α₄α₅(IV) heterotrimers are found in the lung, glomeruli of the kidney, cochlea, eyes, and testis, whereas α₅α₅α₆(IV) is found in skin, Bowman's capsule of the kidney, the esophagus, and the knee joint.⁶² Turnover of the basement membrane is associated with a range of diseases.

Type V collagen is expressed in tissues containing type I collagen, but is a quantitatively minor component.⁶³ The most common structure of type V collagen is α₁α₁α₂(V), although homotrimers of three α₁(V) chains and heterotrimers of the α₁α₂α₃(V) isoforms have also been detected.⁶⁴ It typically forms heterofibrils with type I collagen,^49,63 where it makes up the core structure of these heterotypic fibrils. Type V collagen is of special importance for the structure of tissues. It has been shown to be essential for the correct assembly of collagen fibrils and to regulate their size and organization.⁶⁵ This characteristic makes type V collagen especially unique and interesting to study. The N-terminal domain contains a high level of tyrosine sulfated residues that contribute to the strong interactions that type V collagen has with triple-helical domains of other collagen types. This enhances the stability of fibrils.⁶⁶

Type VI collagen is a heterotrimeric molecule with the isoform α₁α₂α₃(VI) and consists of a short triple helix flanked by two extended globular domains, and it is expressed, albeit variably, in virtually all tissues. The primary fibrils are arranged in overlapping dimers in an antiparallel manner and form parallel tetramers that are stabilized by intermolecular disulfide bonds. They aggregate to form filaments and an independent microfibrillar network. Type VI collagen molecules have a uniquely beaded appearance and interact with several ECM components such as type I collagen and fibronectin.⁶⁷

Proteoglycans

Proteoglycans are ECM macromolecules formed by a protein core with one or more glycosaminoglycans (GAGs) bound covalently. Due to the negative charge and structural conformation of GAGs, proteoglycans can interact with a large variety of macromolecules.⁶⁸ Proteoglycans can be divided into five families according to the structural properties of their core protein.^69,70

The small leucine-rich proteoglycan (SLRP) family is formed by proteoglycans that bind specifically to other ECM constituents and contribute to the structural framework of connective tissues. SLRPs are small molecules, with core proteins of 40 kDa, and possess characteristic 6–10 leucine residuces at conserved locations between the flanking cystein-rich disulfide-bonded domains at the N- and C-terminus that participate in protein–protein interactions with collagens, matrix glycoproteins, and cell membrane components.^70,71 Based on several parameters, including gene organization and amino acid homologies, SLRPs are further divided into five classes: class I includes decorin, biglycan, and asporin; class II includes fibromodulin, lumican, keratocan, proline arginine-rich end leucine-rich repeat protein (PRELP), and osteoadherin; class III includes epiphycan, mimecan, and opticin; class IV includes chondroadherin and nyctalopin; and class V includes podocan.^69,72

Decorin, fibromodulin, asporin, lumican, PRELP, and chondroadherin can interact with collagen and influence collagen fibril formation and interaction.⁶⁹ In addition to their ECM functions in tissue hydration and collagen fibrillogenesis, proteoglycans are able to influence tissue repair and tumor growth, to facilitate cellular adhesion, proliferation, and migration, and to modulate growth factors and cytokine activities. For this reason, they are referred to as matricellular protein, with the ability to modulate cell–matrix interactions and cell functions.⁷² In particular, decorin, biglycan, and lumican exert many modulation roles in different biological processes. These functions highlight the important effect of ECM components in the cellular phenotype by influencing cell communication through, that is, signal transduction, cytokine modulation, adhesion, and migration.⁷² All the different matricellular functions exerted by these three important SLRPs are detailed in Table 1.^3,73–118

Some of the most important proteoglycans expressed in the ECM are briefly described in the following paragraphs.

Aggrecan is the major proteoglycan of the cartilage, and it is the most highly glycosylated, with 150 chondroitin sulfate and keratan sulfate GAGs bound to a large central core protein. Through its specific binding with hyaluronan and link protein, it forms a supramolecular structure whose characteristics make it able to retain water molecules in the cartilage, providing the tissue with the property of resisting compressional forces with minimal deformation.⁶⁹

Versican is a large interstitial chondroitin sulfate proteoglycan. It is present in many tissues, and it is one of the principal ECM components of normal blood vessels where it influences the assembly of ECM and controls elastic fiber fibrillogenesis.¹¹⁹ It is present in the intima and adventitia of most arteries and veins, and it is synthesized by vascular smooth muscle cells as well as endothelial cells, myofibroblasts, and macrophages.¹²⁰ Versican interacts with hyaluronan and link protein to form high molecular weight stable aggregates. These complexes create a reversibly compressive compartment and provide a swelling pressure within the ECM that is compensated by collagen and elastic fibers. Dramatically increased levels of versican have been observed in atherosclerosis and restenosis, implying that this proteoglycan is a specific component of developing lesions and contributes to their progression in atherosclerosis and restenosis.¹¹⁹

Perlecan is a heparan sulfate proteoglycan widely distributed in BMs, and it has the largest core protein found in proteoglycans. It is able to self-associate or interact with several other BM macromolecules, including laminin and type IV collagen.⁶⁸

Decorin is the most abundant SLRP in cartilage. It contains one GAG chain, often dermatan sulfate, which can adopt complex secondary structures and form specific interactions with matrix molecules. Its level increases with age. Its main function is to regulate collagen fibrillogenesis and to maintain tissue integrity by its binding with fibronectin and thrombospondin.⁶⁹ However, decorin also exerts important matricellular functions, favoring the cell–matrix interactions and influencing cell phenotype (see Table 1). Decorin is an important antifibrotic agent: it influences fibrogenesis in different organs by inhibiting TGF-β; it regulates ECM synthesis and turnover, and it is involved in regulation of cell death, adhesion, and migration.⁷²

Biglycan is a small SLRP. It is found in many connective tissues, such as skin, bones, and blood vessels. Within the hyaline cartilage tissue, biglycan is localized mainly pericellularly.⁷⁰ Together with decorin, biglycan is a key regulator of the lateral assembly of collagen fibres, and it interacts primarily with type VI collagen.⁶⁹ Biglycan is thought to have a role also in fibrogenesis and in assembly of elastin fibers.¹²¹ Moreover, this proteoglycan is able to bind to the membrane-bound proteoglycan, dystroglycan, and to a wide variety of proteins. It is involved, for instance, in cell signal transduction during cell growth and differentiation and in regulating cytokine activity through its capacity to bind TGF-β and tumor necrosis factor (TNF) α (see Table 1).⁹⁹

Mimecan is a keratan sulfate proteoglycan belonging to the SLRP family, and it is the product of the gene that encodes for osteoglycin.¹²² Its main role consists in regulating the collagen fibril diameter.¹²³ Apart from corneal tissue, where it has been first identified, mimecan is expressed also in other tissues, such as medullary bone,¹²⁴ amniotic membrane,¹²⁵ cartilage¹²⁶ and pituitary.¹²⁷ In the lung, mimecan mRNA expression is correlated to nonsmall-cell lung cancers,¹²⁸ and in arteries, there is an indication that it can be involved in arterial remodeling during atherosclerosis.¹²⁹

Fibromodulin is one predominant SLRP in cartilage. It contains up to four keratan sulfate chains, and it is able to influence collagen fibril formation and maintain a sustained interaction with the formed fibrils.^69,130

Lumican is a highly biologically active SLRP. It can exist as proteoglycan (with GAG chains) and as glycoprotein (with mono- or polysaccharide chains). In the human adult cornea, it is present in the first form, whereas it is in a glycoprotein form in embryonic cornea and in skin. Lumican expression in cornea has been widely studied. In this tissue, it exerts its main role in controlling the polymerization of collagen into small-diameter fibrils.¹³¹ Lumican is also highly present in skeletal muscle, kidneys, placenta, heart, intervertebral discs, blood vessels, intestine, uterus, and pancreas,⁷¹ and it has a widespread distribution in connective tissues, including cartilage, where it modulates collagen fibrillogenesis and regulates the assembly and diameter of collagen fibers and interfibrillar spacing, enhancing collagen fibril stability.¹³² Together with decorin and biglycan, it is an important component of the ECM exerting matricellular functions (see Table 1).

Other proteins: The glycoproteins fibronectin and tenascin C modulate the integrin-mediated adhesion of cells to other ECM proteins, such as collagens, and as such play a key role in cancer invasion. A single gene encodes fibronectin, but alternative splicing allows formation of multiple isoforms from which some are tumor specific.¹³³ The fibulins, Galectin-1 and Fibulin-1, function as intramolecular bridges in the organization of ECM supramolecular structures, such as elastic fibres and BMs.¹³⁴ Galectin-1 and Fibulin-1 can bind ECM components, that is, laminin and fibronectin, and therefore modifies the adhesive properties of cancer cells.^134–136

Effect of structural proteins on cellular phenotype: selected examples

There is growing evidence that ECM molecules have functions other than structural roles, but as integrated players in the structure and functional homeostasis of tissue. A nonexhaustive list of these proteins is given in Table 1. This highlights that ECM proteins are beginning to be recognized as paracrine-signaling molecules, with profound effects on cellular phenotypes that until recently was restricted to cytokines, growth factors, and hormones.¹³⁷ Of particular relevance, which will be discussed later in this review, some proteins do not change cellular phenotypes in their native conformation, whereas subsequent to a specific PTM, a highly potent and novel function of that protein is revealed. A well-thought example of such cryptic sites is RGD sequences that are either exposed by protease digestion in most collagen species⁸⁷ or even more scholarly exemplified by endostatin, which is a fragment of collagen type XVIII that by cleavage becomes possibly the most powerful anti-angiogenic molecule to date.⁸³

Table 1 lists examples from outstanding research groups which serve to highlight that the matrix encompasses strong signaling motifs that may be revealed during the pathological process. Consequently, the matrix molecules themselves, in addition to cytokines, growth factors, and hormones, become essential players in tissue homeostasis. As the ECM molecules both anchor cells in the right spatial distribution and cell orientation, these structural components may have a dual effect due to their emerging signaling roles.

One gene, 1,000 protein subtypes

The importance of PTMs is best described by the fact that one gene may result in 1,000 different and unique proteins with different functional implications. This is illustrated in Figure 3. Here, modifications to amino acids by specific PTMs or degradation of the protein result in both immunologically different as well as functionally different proteins. Measurement of the same protein may provide highly different information, such as either protein formation or protein degradation, which obviously entails opposite information. Some pathologies may further modify the protein specifically, and thus give a specific protein fingerprint of pathology such as glycosylation of hemoglobin resulting in HbA1c type I diabetes.²⁷³ A long line of evidence suggests that measurement of the intact protein provides some information, and measurement of one of these subforms of a protein provides different and more pathologically-relevant information. These have been carefully described for C-reactive protein (CRP),²⁷⁴ type I and XVI collagens,^275,276 osteocalcin,²⁷⁷ and a selection of other analytes. Thus, it is becoming increasingly clear that measurement of a given pathologically-modified protein enables refinement of clinical chemistry and diagnostic procedures. Most likely the best example is hemoglobin, in which the modification of glycosylation is the gold standard marker of diabetes; thus, the intact protein is a necessity of life, whereas the PTM-modified protein is a pathology-specific marker.

As will later be described in larger molecular detail, many amino acids are amenable to specific modifications (citrullinations, phosphorylations, acetylations, methylations, nitrosylations, and glycosylations¹⁷). These modifications can have both positive and negative impacts on the function of the protein, and even target the protein for degradation. In addition, many proteins are born with a propeptide that needs to be cleaved before the protein is in the active configuration that being enzymatic activity or structural enablement. Both N- and C-terminal propeptides are present, which may be further modified, and thus is a waste underestimation of the complexity of these peptides. The degradation products of proteins may be the specific action of pathological-specific enzymes, and there is an accumulating amount of evidence suggesting that different fragments of the same protein may have different physiological and pathophysiological meanings.²⁷⁸ Lastly, polymerization may both be understood as aggregates of the same protein such as hyperphosphorylated Tau or cross-linked collagens, but also pentameric CRP. Each of these subpools obviously holds unique information.

Cross-linking

Cross-linking plays an important role in the ECM meshwork and thereby in tissue integrity. Cross-linking between different ECM components or between different protein chains can result from enzymatic and nonenzymatic pathways. Enzymatic cross-linking is often processed by members of the LOX enzyme family, whose members have been shown to promote the linearization of interstitial collagens which stiffen the tissues, thus leading to neoplastic progression of tumor cells.^280–283 Interestingly, this matrix stiffness was associated with different phenotypes and enhanced mechanoresponsiveness of the epithelium.^280,281 Therefore, cross-linking plays an important part in both the initiation and progression of metastasis. Similarly, in fibrotic disease, increases in tissue tension mediated by cross-linking can lead to activation of TGF-β signaling from the latency-associated complex and other signaling changes, driving a fibrogenic feed-forward loop.

Valuable assays for evaluation of bone- and cartilage-related diseases have been developed using antibodies highly specific for protease-cleaved sites in type I collagen²⁷⁶ and type II collagen,²⁸⁴ respectively. The antibodies in these assays also assess the cross-linking between the lysines in the epitopes. C-terminal telopeptide of type I collagen (CTX-I) is an 8-amino acid fragment from the C-telopeptide of type I collagen generated by cathepsin K activity, and the rate of its release from bone is a useful reflection of the resorbing activity of osteoclasts.²⁸⁵ Measuring this fragment is useful for the evaluation of treatment efficacy in bone diseases such as osteoporosis.²⁸⁶ The CTX epitope contains an aspartylglycine motif (DG) that is prone to spontaneous isomerization. In other words, EKAHD(α)GGR epitopes are released during degradation of newly synthesized type I collagen, whereas EKAHD(β)GGR epitopes are released from matured type I collagen. It has been established that the α/β ratio is a useful measure of the age of bone tissue;^287–289 the lower the ratio, the older the bone tissue.²⁹⁰ Further, the lysine residue of the CTX residue is cross-linked. Figure 4 outlines schematically how assessment of both a cross-linked and cathepsin K-degraded epitope may be undertaken through the use of sandwich enzyme-linked immunosorbent assay (ELISA) technology. Additional ECM assays may be constructed by a similar approach, to include as much possible information of protein subtype as possible. Resorption rates of newly synthesized collagen type I can be assessed by specific immunoassays targeting the detection of αCTX in urine samples.²⁹¹ Degradation rate of matured, isomerized collagen can be estimated by another specific assay targeting βCTX in both urine and serum samples.²⁷⁶

Another way of cross-linking is through the actions of tissue transglutaminases (TGs). They play a fundamental role in tissue stabilization by transamidation of glutamine residues of one protein chain to the amino group of a lysine residue in a second protein chain. This results in the formation of the covalent N-γ-glutaminyl-ɛ-lysyl-isopeptide bond, which is resistant to proteolytic degradation.²⁹² As several ECM proteins, such as collagens, fibronectin, laminin, and vimentin, act as substrates for TGs, they are involved in physiologic tissue integrity while being associated with various pathologies, including neurodegenerative diseases, cancer, inflammation, and fibrosis. In fibrosis, TGF-β promotes activation of TG cross-linking, thereby reducing the ECM turnover, leading to deposition and accumulation of ECM proteins, and thus stabilizing the ECM network and facilitating proteolytic resistance. In cancer, intracellular cross-linking by TG2 has been shown to be both pro- and antiapoptotic, and favoring cell survival, invasion, and motility by the close association with surface integrins.^293,294

Oxidations and hydroxylations

Oxidative damage to proteins is often caused by the action of free radicals, reactive oxygen species (ROS) and reactive nitrogen species such as hydrogen peroxide and nitric oxide, generated in cells by the mitochondrial respiratory chain.²⁹⁵ Oxidizing PTMs have been implicated in several pathological and healthy tissue turnover processes. Although many amino acids can be attacked by ROS, some seem more likely to undergo oxidation than others. For example, lysine and proline are readily oxidized to aldehydes; methionine is sulfoxidized; and tyrosines are nitrosylated.²⁹⁶ Under normal conditions, these ROS are strictly regulated by antioxidants, such as peroxidases and dimutases among others.²⁹⁷ However, under pathological conditions, oxidation may be implicated in tissue destruction. The role of ROS in almost all aspects of cancer initiation and development^{139,295,298–303} is still debated. Measurement of specific components of the ECM that hold these oxidized PTMs may be useful for both early diagnosis and prognosis of cancer.

Protease-generated neoepitopes

Matrix remodeling at specific disease stages results in both elevated levels of, and uniquely modified, proteins. Endopeptidases, such as MMPs and cysteine proteases, play major roles in the degradation of extracellular macromolecules such as collagens and proteoglycans. Specific proteolytic activities are a prerequisite for a range of cellular functions and interactions with the ECM, resulting in the generation of specific cleavage fragments. Even though many components of the ECM, as well as enzymes responsible for remodeling, are present in different tissues, the combination of a specific peptidase and specific ECM protein may provide a unique combination that elucidates activity in a particular tissue or a specific disease mechanism.

One often-taught example of protease degradation of a given tissue is that of joint degenerative diseases. Joint degenerative diseases lead to alterations in the metabolism of the articular cartilage and subchondral bone.^{278,304–309} Cartilage is for the most part composed of collagen type II, which accounts for 60%–70% of the dry weight of cartilage, and proteoglycans accounting for 10% of the dry weight, of which aggrecan is the most abundant.³¹⁰ Since type II collagen is the most abundant protein in cartilage, several different degradation fragments of collagen type II have been indicated as useful for monitoring degenerative diseases of the cartilage.^272,311 C-terminal telopeptide of type II collagen (CTX-II) is an MMP-generated neoepitope derived from the C-terminal part of type II collagen,³¹⁰ and measurement of CTX-II is highly useful for monitoring degradation of type II collagen in experimental setups assessing cartilage degradation.^278,312 Examples of a range of protease-generated neoepitopes have already been described in the literature, but they have not been utilized by applied science to produce quantifiable methods of disease assessment. Assays detecting a few neoepitopes that have been developed and that are used in both clinical and preclinical studies were reviewed recently.³¹³

To some extent, C-terminal telopeptide of type I collagen (ICTP) and MMP-derived fragments of type I collagen assays^{53,54,314–316} as an indicator of cancer progression have been developed and used in prognosis of lung and ovarian cancers. A range of biochemical markers based on degradation products of the ECM, particularly collagen, may be identified and used in cancer. The collagen composition of the BM and interstitial matrix may be relevant for the development of the given marker for the ECMR associated with soft tissue metastasis.

Isomerization: Advanced glycation end product of ECM proteins

Proteins containing aspartate (D), asparagine (N), glutamate (E), or glutamine (Q) residues linked to a low–molecular-weight amino acid, such as glycine (G), can undergo spontaneous nonenzymatic isomerization.¹⁵ This isomerization introduces a kink in the conformation of the molecule, as the peptide backbone is redirected from the γ-carboxyl group in the native newly synthesized form to the side chain γ-carboxyl.²⁹⁰ Peptides that contain amino acid isomerizations are often resistant to proteolysis.^317,318 This feature affects the processing of antigens for presentation on the major histocompatibility complex II during the immune response signaling for the production of T-cells and antibodies.¹⁵ In preclinical studies, it has been shown that various known autoantigens contain sites prone to deamidation and isomerization. These autoantigens are involved in type I diabetes, rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), and experimental autoimmune encephalomyelitis.^{317,319–322}

The C-telopeptide of type I collagen marker CTX-I is a marker of bone resorption. It has been shown that assessment of the nonisomerized epitope (αCTX-I) is more sensitive as a marker for bone metastases secondary to breast and prostate cancer than the isomerized epitope (βCTX-I).²⁰⁷ This is due to the high ECMR of type I collagen in the bone area invaded by cancer cells, and thus a high amount of newly formed nonisomerized collagen type I undergoes resorption by osteoclasts.

Nonenzymatic glycosylation

Nonenzymatic glycosylation is also called the Maillard reaction, and leads to PTMs of proteins, nucleic acids, and lipids.²⁷³ A common cause of nonenzymatic glycosylation is increased blood glucose levels, and accordingly, most knowledge about nonenzymatic glycosylation arises from studies performed in diabetics.²⁷³ The marker HbA1c is an established PTM marker in type II diabetes. Recently, advanced glycation end products (AGEs) have been implicated in cancers. The nicotine-induced accumulation of AGEs is a cause of cancer.³²³ The receptor for AGEs, called RAGE, is currently under intense investigation as both a marker and an inducer of cancer³²⁴ and to assess whether there is a link between chronic inflammation and cancer, since inflammatory mediators can both be pro- and antitumorigenic.^{139,216,324,325}

Citrullination

Citrullination or deimination is the term used for the PTM of the amino acid arginine, which can transform into the amino acid citrulline. The change is facilitated by peptidylarginine deiminases (PADs).^326,327 The conversion of arginine into citrulline can have important consequences for the structure and function of proteins, since arginine is positively charged at a neutral pH, whereas citrulline is uncharged. The positive charge increases the hydrophobicity of the protein, leading to changes in protein folding.

Histone deacetylase 1 (HDAC1) inhibitors are currently under development for the treatment of certain cancers, particularly breast cancer.²¹ Histone lysine and arginine residues contain a wide array of PTM-producing processes, including methylation, citrullination, acetylation, ubquitination, and sumoylation. The combined action of these modifications regulates critical DNA processes, including replication, repair, and transcription. In addition, enzymes that modify histone lysine and arginine residues have been correlated with not only cancer but also arthritis, heart disease, diabetes, and neurodegenerative disorders.^328,329

Histone methylation plays a key role in regulating the chromatin structure and function. The recent identification of enzymes that antagonize or remove histone methylation offers new insights into histone methylation plasticity in the regulation of epigenetic pathways. Peptidylarginine deiminase 4 (PADI4; also known as PAD4) was the first enzyme shown to antagonize histone methylation. PADI4 functions as a histone deiminase, converting a methylarginine residue to citrulline at specific sites on the tails of histones H3 and H4. PADI4 associates with HDAC1.^328–330

An example of a combined aged, cross-linked, and cleaved neoepitope for the evaluation of bone metastases

The relationship between skeletal tumor load and elevations in serum or urine levels of αCTX and seven other biomarkers related to bone turnover has been investigated in a pooled group of breast and prostate cancer patients.²⁰⁷ Patients were stratified according to the Soloway score:

• • Score 0: 0 bone metastases

• • Score 1: <6 bone metastases

• • Score 2: 6–20 bone metastases

• • Score 3: >20 bone metastases

• • Score 4: superscan showing that >75% ribs, vertebrae, and pelvic bone are affected.

In breast cancer patients, a strong linear association was observed between bone metastases and all biomarkers, except osteoprotegerin and receptor activator of nuclear factor κB ligand (Figure 5). All six remaining markers were significantly elevated in patients with a Soloway score of 1. The relative percent increases in biomarker levels in the presence of bone metastases were most pronounced for αCTX-I, which was elevated by more than 600% in patients with Soloway score of 3. The next highest increases were in bone-specific alkaline phosphatase and N-telopeptide of type I collagen (NTX), which were elevated by 470% and 440% at Soloway score 3, respectively. These findings were supported by observations in prostate cancer patients, which showed that of the seven biomarkers, αCTX-I was the most sensitive for bone metastases.³⁴⁸ The higher sensitivity of αCTX-I could be explained by the fact that this epitope is released from sites of high bone remodeling, where collagen fibrils do not have time to mature and undergo β-isomerization. The αCTX-1 epitope was located by immunostaining adjacent sections of bones invaded by breast cancer or prostate cancer,²⁰⁸ and at the sites of high bone remodeling.

These data support that careful selection of matrix constituents and, in particular, those that carry one or more PTMs such as isomerization in a type I collagen fragment generated by cathepsin K as described in this example may be superior markers reflecting pathological, including malignant, events in the ECM.

An example of MMP-degraded type III collagen for the assessment of liver fibrosis

The central pathological change in fibrosis is uncontrolled ECMR.^349,350 During fibrogenesis, the quantity and composition of matrix proteins in the liver change, resulting in excessive accumulation of fibrous (scar) tissue and an overall increase in the ECM density.³⁵¹ ECM matrix proteins in a normal liver are distributed mainly in the portal tracts, whereas a BM-like matrix is located in the perisinusoidal space of Disse. The most abundant collagens in the liver are type I and III collagens, which by immunohistochemistry are found predominantly in the perisinusoidal spaces, in portal tracts, and in subcapsular areas.^5,352 The ECM of the cirrhotic liver contains approximately six times as much matrix as the normal liver,³⁵³ which is a result of increased levels of type I, III, and IV collagens.³⁵⁴ However, levels of MMPs such as MMP-9 also increase in cirrhosis.^349,355 The combination of active and overexpressed MMP-9 with the accumulation of type III collagen poses the interesting hypothesis that an MMP-9-generated fragment of type III collagen could be used as a biochemical marker of liver fibrosis.

Type III collagen degradation by MMPs, and even MMP-9 exclusively, may result in many unique fragments, such as those derived from type II collagen and previously published.⁸¹ The CO3-610 (C3M) fragment (KNGETGPQGP) is one of those, and is exclusively derived from MMP-9. When this C3M fragment was assessed in two separate animal models of liver fibrosis, the BDL and CCl4 animal models, a >200% fold upregulation was observed, as well as a highly significant correlation to portal pressure.^60,356,357 These data strongly suggest that liver fibrosis is not merely an accumulation of ECM proteins, but a dynamic condition with accelerated ECM turnover, in which both tissue formation and tissue degradation are highly upregulated. In the case of liver fibrosis, ECM tissue formation outstrips tissue degradation, leading to a net accumulation of scar tissue over time. This example also suggests that PTMs released by protease degradation of proteins may in some cases be more sensitive markers for pathology than intact proteins. This idea is receiving increased attention.¹⁷ This approach has been recently been described as the protein fingerprint technology, in which the different subpools of the total pool of information about one protein during formation or degradation provide distinct and important data.²⁰