Keloid progression: a stiffness gap hypothesis

ABSTRACT

Keloids are fibroproliferative skin disorders characterised clinically by continuous horizontal progression and post‐surgical recurrence and histologically by the accumulation of collagen and fibroblast ingredients. Till now, their aetiology remains clear, which may cover genetic, environmental and metabolic factors. Evidence in the involvement of local mechanics (e.g. predilection site and typical shape) and the progress in mechanobiology have incubated our stiffness gap hypotheses in illustrating the chronic but constant development in keloid. We put forward that the enlarged gap between extracellular matrix (ECM) stiffness and cellular stiffness potentiates keloid progression. Matrix stiffness itself provides organisational guidance cues to regulate the mechanosensitive resident cells (e.g. proliferation, migration and apoptosis). During this dynamic process, the ECM stiffness and cell stiffness are not well balanced, and the continuously enlarged stiffness gap between them potentiates keloid progression. The cushion factors, such as prestress for cell stiffness and topology for ECM stiffness, serve as compensations, the decompensation of which aggravates keloid development. It can well explain the typical shape of keloids, their progression in a horizontal but not vertical direction and the post‐surgical recurrence, which were evidenced by our clinical cases. Such a stiffness gap hypothesis might be bridged to mechanotherapeutic approaches for keloid progression.

Introduction

Keloids, the commonly seen fibroproliferative skin disorders, are notorious for the accumulation of collagen and fibroblast ingredients and the post‐surgical recurrence, whose aetiologies still remain unclear. Although commonly accepted as excessive wound healings aberrant from normal ones, they are widely and actively explored in terms of the mechanism of their development and progression, in genetics, biochemistry, endocrinology, immunology and nutrition. The potential genetic loci 1, the injury‐wound tension theory 2, sebum/sebocyte hypothesis 3, 4 and neurogenic inflammation hypothesis 5, as well as others involving nitric oxide 6, have been put forward under such backgrounds. In particular, allowing mechanics to be considered in the formation and progression of keloids dawns new lights and opens a promising new era. Epidemiologically, hypertrophic scars (HSs) and keloids prefer areas exposed to frequent mobility and/or high stretching tension, such as anterior chest and scapular regions, but seldom in areas where stretching/contraction is rare, such as parietal region or anterior lower leg 7. Clinically, the typical shapes of keloid, butterfly, crab's claw or dumbbell, are largely determined by the local mechanics on the skin 8 (Figure 1), and fundamentally, that evidence is increasingly directed to mechanotransduction 9. However, it is still difficult to draw a universal picture of the mechanic‐oriented mechanisms in the complex micro‐environment in vivo. Thus, from a holistic perspective based on the function‐follows‐shape phenomenon, we introduce our ‘stiffness gap hypothesis’ in this article and suggest that the enlarged gap between extracellular matrix (ECM) stiffness and cellular stiffness potentiate keloid progression, beyond all the compensations and together with other changes such as local inflammation.

Figure 1.

Typical shapes of keloids. Clinically, in the predilection sites of keloids, such as chest or scapula, that are constantly exposed to cutaneous stretch, the typical shapes of keloids are butterfly, crab's claw or dumbbell, which are largely determined by the local mechanics on the skin. A. Butterfly in the chest B. Crab's claw in the chest C. Dumbbell on the shoulder

Why stiffness gap

Roles of ECM stiffness: the function‐follows‐shape phenomenon

Matrix stiffness is an important micro‐environmental cue that regulates cell behaviour and function, the role of which has been proven in orienting cell division, maintaining tissue homeostasis, driving cell migration and regulating differentiation 10, 11. In addition, in vivo tissues are viscoelastic and are made up of cells and ECM. When matrix mechanics become imbalanced, disease progression may ensue 12.

The surging development of mechanobiology has refreshed the views of in vivo cell fates that have been dominated by genes and biochemistry. Recently, the old chicken‐and‐egg problem of shape and function is being oriented into ‘change the shape, change the function’ in face of ECM stiffness, by providing organisational guidance cues for the cellular decision making in migration, differentiation, proliferation 13 contractility 14 and even cancer cell tumorigenicity 15.

As indicated in the phenomenon of durotaxis, the moving speed and direction of NIH 3T3 or primary heart fibroblasts are ECM stiffness‐dependent 16, 17. The transition of fibroblasts from soft to rigid ECM induces the expansion of lamellipodia and lamella and the subsequent cell migration across the boundary towards the stiffer one, while the reverse stiffness gradient resulted in their retractions 17. Such rigidity‐guided movement is echoed in inflammatory cells such as neutrophils, which also favour migration towards stiffer fibrils 18. Moreover, the ECM stiffness‐driving specification in mesenchymal stem cells demonstrates targeted cellular differentiation induced by simply altering substrate stiffness, the compliance of which matches well with the final tissue lineages, and the soluble induction factors demonstrate less selectiveness than matrix stiffness in driving specifications 19. Additionally, human dermal fibroblasts preferentially proliferate on stiffer substrates while reducing their proliferation rate markedly on softer substrates 20. Increasing ECM stiffness can also reduce the chemotherapy‐induced apoptosis of hepatocellular carcinoma cells 21. Furthermore, fibroblasts in collagen gels with higher boundary stiffness generate higher traction forces 22. That being the case, the ECM is increasingly reconsidered being a dynamic, mobile and multifunctional regulator of cell behaviours, rather than merely the cellular scaffold and cytokines storehouse 23, 24.

However, the significance of functional ECM lies in its dynamic interactions with the cells that reside in it. As the active part of the tissue, cells interact with their ECM to perform specific functions, but still, varied types of cells have different responses to the same ECM components. That being the case, we put forward the hypothesis that stiffness gap‐mediated biomechanical signalling between ECM and the constituent cells in keloids, rather than merely ECM‐mediated biochemical signalling, is involved in the continuous progression and deterioration of primary keloids or their recurrence after excision. It allows the cell heterogeneities in their mechanosensitivity to ECM on the one hand while on the other hand integrates constituent cells' compensations to the ECM stiffness stimuli.

Lessons learned from tumour invasion

The ECM stiffness gap during keloid progression is to some extent similar to the phenomenon during the deterioration and invasion of malignant tumours. Take breast tumors for an example. Tumors usually demonstrate to be more rigid (stiffer) than their surrounding normal tissue. Such a tumour stiffening can be sensed in vivo by patients and surgeons subjectively as an early diagnostic sign and can further be confirmed objectively by the increased matrix deposition and cross‐linking under cell culture in vitro 25, 26. However, a single cancer cell is more compliant (softer) than normal cells 27, 28. The stiffness peak in densly packed tumour cells, normal glandular epithelium and benign solid lesions is 0.5–0.8 kPa, ∼1 kPa and ∼3 kPa, respectively 29. Considering the fact that the ECM ingredients are derived from its constituent cells, such an enlarged stiffness gap between ECM and cellular stiffness is aggravated during the tumour progression, which subsequently forms a vicious cycle during the cancer progression in breasts or other tissues.

The benign keloids appear to share several similar features with tumours. (i) Clinically, keloids can horizontally progress, or ‘invade’, into their adjacent healthy tissue without spontaneous regression, shown initially as expanding inflammation and subsequently as accumulations of collagens and dermal fibroblasts, and they tend to recur after surgical excision. Furthermore, the distribution of the tensions that is applied to the skin around the wound site determines the shape of the keloid and its ‘invasion’ in greater extent along the x‐axis (in the direction of stretching) than the y‐axis (perpendicular to the direction of stretching) 30. (ii) Histologically, keloids show strong resistance to apoptosis 31. Fortunately, such an ‘invasion’ is only in a horizontal direction. In the direction of the z‐axis, a keloid can confine itself vertically within the dermis 32 unless a further ulcer occurs with a subsequent malignant Majolin's ulcer. The shared features of invasion with tumours not only wins keloids the title of ‘skin tumour’ but also compels us to explain the progression of keloids using the cellular and ECM stiffness in keloids, borrowed from tumour invasion.

What are the stiffness gap and its decompensation in keloids

During the long period of keloid progression, changes in ECM and its constituent cells are dynamic, and the relationship between them facilitates the keloid pathogenesis. In healthy human skin, stiffness of its dermis in terms of Young's modulus is 1–5 kPa 33, 34, whereas in fibrotic tissue, its Young's modulus increases to 20–100 kPa, similar to that in collagen‐dense tissue, such as a tendon or osteoid structure 35. However, the dermis in fibrosis builds to be stiff gradually in the chronic process as the initial provisional ECM of an early wound is reported to be as soft as the newly polymerised collagen gel (elastic modulus = 0·01–0·1 kPa) 34. Obviously, the interactions between cells and their derivative ECM participated actively in the long run.

We put forward here the stiffness gap hypothesis to illustrate the mechanism of the progression of keloids, which aggravates their horizontal expansion: (i) the ECM stiffness (the dominant ECM ingredients in keloids are collagen type I 36) and cell stiffness (the dominant cell constituents in keloids are dermal fibroblasts) are not well balanced in keloids, with the ECM stiffness increasingly higher than cell stiffness, and such a continuously enlarged stiffness gap potentiates the cellular changes towards the keloid progression, and (ii) either ECM or cell stiffness is influenced by cushion factors (e.g. prestress for cell stiffness and topology for ECM stiffness) as compensations to the drastic environmental changes, the decompensation of which aggravates keloid development (Figure 2).

Figure 2.

Stiffness gap hypothesis, the balance between cell and extracellular matrix (ECM) stiffness and their cushion factors. Cutaneous resident cells in keloids can sense (mechanosensitivity) and respond (mechanoresponsiveness) to the stiffness gap between ECM and themselves by mechanotransduction where intracellular biochemistry and gene expression leads to cell function changes. The input stiffness gap can somewhat be cushioned by cellular factors such as prestress, structural organisation and composition, as well as ECM factors, such as its topology, thickness and composition. If failed, the tilted balance induced by enlarged stiffness gap between ECM and cell will guide the keloid progression through the output of changes in cellular functions, such as reduced apoptosis and increased proliferation.

Concomitant with the omnipresent environmental changes, such as the local mechanics commonly seen in keloids, cutaneous cells can sense (mechanosensitivity) and respond (mechanoresponsiveness) to stiffness gaps between ECM and themselves by mechanotransduction, where intracellular biochemistry and gene expression leads to cell functions changes 37, 38. It is through ECM, where cells are fully embedded in and surrounded by, can the external forms of environmental signals be influential to the microenvironment and finally to cells within it to take effects. The higher ECM stiffness is, compared to that of cell stiffness, the easier it is for a keloid to progress with active cellular functions of proliferation, migration and collagen synthesis that characterises keloid pathology and the subsequent clinical symptoms of mass and invasion.

Just as ECM stiffness is regulated by its composition, thickness and topology, cell stiffness has its cushioning factors in front of the enlarged gap before final cellular functions can be induced towards fibrosis. These factors may include cell composition, structural organisation, prestress 39 and traction force. In particular, prestress, which is generated in the contractile actin filaments before exposure to external forces 39, is a central factor determining cell stiffness 40 and is closely associated with cell stiffness independent of cell geometry, through the polymerisation of actin lattice, myosin cross‐bridge recruitment 41 and tensegrity that integrates homogeneity within the cell through cytoskeletons 42. In living 3T3 fibroblasts, the average shear modulus correlates linearly with the prestress 39, and significant stiffening occurs only when the prestress exceeds the threshold beyond 0·1–1 kPa 39, 43. Besides, cells can exert the actomyosin‐ and cytoskeleton‐dependent traction forces on ECM during mechanoreciprocity 44. In the rigid collagen I fibres, fibroblasts generate maximal traction forces to stabilise integrin adhesions 45. If those cushions are compromised, for example, traction forces are abnormally high in a compliant matrix as seen in the transformed cells, the consequences will be disrupted cell–cell junction integrity, compromised tissue polarity, promoted anchorage‐independent survival and enhanced invasion 44. Conversely, if those cushions are over‐challenged, mechanosensitivity to ECM stiffness will be downregulated, as seen in the significant high prestress and the lack of ECM stiffness sensitivity in alveolar epithelial cells 46. Although ingredient cells and their mechanical/biochemical micro‐environment cannot be identically the same among varied in vivo situations, the gap between ECM and its derivatives of constituent cells can contribute to fibrosis or aggravate the vicious cycle, if not initiate it.

It is well accepted that the formation and development of keloid takes time after evident wounding or not. Take a burn wound for example. Healing within 10 days or beyond 21 days will result in risks of 4% or ≥70%, respectively, of developing into a hypertrophic scar as 47. Besides, although the skin that covers our body will be somewhat exposed to mechanics, not all lesion sites or every patient after wounding or surgery will develop keloids. Indeed, the prevalence of keloid varies from 0·09% in UK to 16% in Congo 48. To some extent, the decompensation of cells to the enlarged stiffness gap between ECM and cells accrue with time and are amplified through the vicious cycle during the chronic process and thereby gradually manifest as evident keloid progression, during which period, the cell's susceptibility and tolerance varies among individuals.

How is the stiffness gap manifested in keloid progression

Stiffness gap hypothesis can find supporting evidence at the tissue, cell, and molecular levels in explaining the clinical symptoms and cellular changes in keloids and, promisingly, in molecular events (Figure 3).

Figure 3.

Molecular‐ and tissue‐level illustration of stiffness gap hypothesis in keloid progression. At the tissue level, during the vertical advancement of keloids, the sudden drop of the extracellular matrix (ECM) stiffness from collagen type I in the dermis to the fat in the underlying subcutaneous level will reduce the stiffness gap of ECM and cells. Although the dermis and subcutaneous adipose are different in many aspects, such as niche components, we believe that the stiffness gap contributes, at least partially, to the confinement of keloid progress. Together with the reactive compensations from cells, vertical keloid progression stops beyond the subcutaneous layer. At the molecular level, the mechanosignalling pathways involved in pathological scarring cover transforming growth factor‐β, integrin and G proteins and may provide mechanobiological evidence for stiffness gap hypothesis in explaining keloid progression.

At the tissue level, keloid progression is clinically characterised with its horizontal expansion and vertical limitation. It can be sourced to the different stiffness gaps between ECM and cells (dominated by fibroblasts) in varied local settings. Let us take the anterior chest keloid invasion in a butterfly shape (Figure 1) as an example. On one hand, the local ECM stiffness is determined by the upper limb movements. The exposure of the local skin to steady and vectorial stretching brings anisotropy of ECM stiffness in a manner of axisymmetric distribution along the midline, and the nearer it is to the stretching sources of arms, the higher the stretching/tensions on the local skin will be. It is further supported by the finite element analysis that stiffness of the skin at the circumference of a keloid correlates directly with the degree of skin tension there 30. Additionally, as evidenced in our chest keloid clinical case, if such a symmetrical stretch between the midline is accidently broken, the stretch from the side of dominant‐hand will facilitate the anisotropy of ECM stiffness and the subsequent enlarged stiffness gap, which was followed by the continuous horizontal growth of the half‐butterfly on this side, while on the other side, the volume of the original half‐butterfly tended to be much smaller, which indicated the somewhat ameliorated progression of the keloid on that side (Figure 4). Thus, the elevated local ECM stiffness by stretching anisotropy paves the way for the progress of the butterfly wings in the horizontal direction. In particular, after keloid excision, the tensions in situ on suturing, together with the unavoidable daily stretch, enlarged the stiffness gap far out of what the cushion factors can allow and compensate, which further facilitates the not‐uncommonly seen recurrence after surgery. On the other hand, fibroblasts try to match their compliance with surrounding ECM 49, although this failed in keloids. Usually, in a static cell, the cell stiffness associates with the tensile stress level within the cytoskeleton as the traction force at the cell–substrate interface has to be balanced by the internal stress 41. When placed on a softer environment, the cellular contractile forces produced by actin tend to be reduced, together with the decreased focal adhesion area 50, 51. However, in keloids, the migratory feature of keloid fibroblasts and their characteristic to migrate outside the original wound margin 52, 53 are concomitant with a softer population of filaments and disassembling focal adhesions in migrating keloid fibroblasts 51. Furthermore, the atomic force microscopy (AFM)‐based evidence indicates changes in the mechanical properties of cytoskeletons in response to cell polarisation and migration. In polarised fibroblasts, the microtubules are distributed towards their rear end, in contrast to the even distribution of microtubules in non‐polarised cells 53. Additionally, the filament elasticity of the lateral rear (5·20 ± 0·66 kPa) is lower than the front (15·2 ± 2·5 kPa) and the lateral‐front edge (21·0 ± 3·1 kPa) in polarised cells and even significantly lower than the filament elasticity throughout the cell periphery (31·8 ± 1·9 kPa) in non‐polarised cells 53. Thus, the softer migrating keloid fibroblasts even enlarged the stiffness gap between ECM and cells, which may aggravate the vicious cycle. Cells near the stiff matrix even exhibit greater cell motility yet remained locally adherent 17. Although the leading edge within a migrating fibroblast can be nearly twice stiffer than the trailing edge, fibroblasts migrate towards stiffer substrates in vitro and thereby enrich the proliferating fibroblast pool in turn towards keloid progression 17, 54.

Figure 4.

Chest keloid case demonstrating the change in local extracellular matrix (ECM) stiffness by stretching anisotropy and subsequent shape modification. The patient was an 84‐year‐old male with a chest keloid in the shape of a butterfly. Two years ago, he accidently broke the keloid from the central, and the wound healed automatically without medical intervention. After that, the right half‐butterfly on his dominant‐hand side progressed as usual under the vectorial stretching and the subsequent anisotropy of ECM stiffness, while the left half‐butterfly ameliorated. Consequently, the butterfly was not symmetrical any longer, with the left half much smaller than the right half.

If the stiffness gap between ECM and fibroblasts is large enough to facilitate the advancement in the horizontal direction, the status in the vertical direction confirms it from the opposite side. The aggressive advancement of keloid fibroblasts in the vertical direction will meet the much softer subcutaneous layer during their dermal expansion. The sudden drop of the ECM stiffness from collagen type I in the dermis to the fat in the underlying subcutaneous level will reduce the stiffness gap of ECM and cells, which contributes, at the least partially, to the confinement of keloid along the z‐axis. It is such a vertical confinement within the dermal layer that excludes keloids from the malignant invasion. In fact, in human metastatic cancer cells, the cell stiffness can be >70% softer than the control benign cells 27. And the stroma at the front of the‐invading tumour is substantially stiffer than the non‐invasive edge 45.

At the cellular level, in response to the steady and vectorial mechanical stimulation and the consequent enlarged stiffness gap between ECM and cells, the advancing edges of keloids are characterised with an activated proliferating phenotype of fibroblasts with abundant ECM synthesis, protease inhibitors and fibrogenic mediators 9. If failed in maintaining the balance under the chronic and steady mechanical challenges, by reduced extensibility, reorientation or increased traction force of a single cell and by the redistribution of cells, keloid progression will be present. In fact, those cell function decisions influenced by the stiffness of the surrounding matrix echo each other. The increase in cellular traction force in response to stiff substrates expects an increase of adhesion strength 20 by inducing growth of the focal adhesion area 55 and reduces cellular migration 20. The apoptosis is induced when cell spreading is restricted 56.

It is now well‐known that local mechanics is a key factor in keloid development and progression 57. However, the molecular mechanisms of stiffness gap in keloid progression are still under investigation. The mechanosignalling pathways involved in keloids may mediate the external mechanical signals into intracellular transcriptional changes to induce keloid formation and progression through the transforming growth factor (TGF)‐ β, integrin, G protein and other signalling pathways 9. Moreover, the stimulus‐triggered acquisition of pluripotency and nuclear reprogramming has been demonstrated in CD45⁺ cells, which required neither nuclear transfer nor transcription factor introductions 58. If so, it might also be induced by the strong external stimuli of stiffness gap as long as the somatic cells (e.g. adult human fibroblasts) can be induced into pluipotent stem cells 59. Besides, the fibroblast‐secreted proteins such as the matrix cross‐linker lysyl oxidases may be involved to set up a positive feedback, which in turn increases ECM stiffness and deteriorates the fibrotic status 60.

If what has been described above centres to some extent on the main cell constituents of fibroblasts in keloids, what cannot be neglected are the effects of ECM stiffness on other resident cells in keloids, such as endothelial cells (ECs) and inflammation cells. ECs are mechanosensitive to ECM stiffness. Increased ECM stiffness, as typically seen in keloid progression, can inhibit sprouting of ECs 61 and decrease vacuole size 62. Besides, high matrix stress increases lumen size and decreases EC elongation. The following structures exhibit less branching and have larger lumens 63. Moreover, ECM stiffness inhibits EC branching to enhance migration speed and thereby promotes ECM invasion 64. Actually, ECM stiffness can influence the endothelial network formation through altering traction forces 65, 66. The increased ECM stiffness evokes focal adhesion reinforcement, which is followed by high traction force and elongated cell shapes 67. This evidence may have important implications for the redness and progression of the advancing hard edge in keloids. Indeed, when cultured on stretched collagen gels, ECs prefer the outgrowth along the tensional stress direction in a dose‐dependent manner 68, and VEGFR2, mRNA and protein levels are higher in stiffer gels 69. Similarly, the substrate stiffness still has an immediate impact on the morphology and spreading function of neutrophils. On stiffer substrate, the neutrophils migrate more slowly but more persistently on stiffer ECM, with greater distance over time 70. It fits well with such a typical manifestation of keloids as the sustained and chronic inflammation at its hard advancing edge, in accordance with the refractory itching and redness in situ.

What does stiffness gap mean to the diagnosis and treatment of keloids

The enlarged stiffness gap between ECM and cells in keloid progression, if confirmed through producing keloid‐like tissues by artificially modulating cell‐ECM stiffness gap, can be bridged to meaningful and feasible diagnostic and therapeutic approaches for keloid progression.

For diagnostic purposes, the stiffness can be developed as a biomarker for the diagnosis of keloid progression. Techniques based on ultrasound or magnetic resonance elastography pave the way for ECM stiffness evaluation 71. Similarly, the atomic force microscopy is ready for direct measurement of cellular stiffness if cells were first separated from their ECMs 72. Surely, thresholds for the enlarged gap between ECM stiffness and cell stiffness are needed for varied populations as required.

For therapeutic purposes, to directly reduce ECM stiffness or to indirectly weaken the cushioning factors can be the targets. In fact, softening the ECM stiffness has been proven effective in keloid treatment, consciously or not, as a way to reduce the stiffness gap between ECM and the cell constituents. Silicon dressing is a good example for softening scars and decreasing their volumes 73.

To sum up, we put forward in this article the stiffness gap hypothesis and suggest that the enlarged stiffness gap between ECM and cell constituents facilitate keloid progression, after compensation by the respective cushioning factors in the balance. Such a mechanic‐oriented hypothesis echoes the burgeoning development of molecular mechanobiology from a holistic perspective on the basis of the function‐follows‐shape principle and tensegrity theory. It also explains the clinical invasion and progression of keloids in the horizontal direction and its confinement beyond subcutaneous layer in the vertical direction, thereby separating keloid invasion from a malignant one. As keloid aetiology remains unclear, our stiffness gap hypothesis enriches the understanding of keloid progression but does not exclude roles of other contributing factors in keloid formation. Ultimately, systematic mechanobiological experiments have to be designed and performed to verify or adjust this hypothesis. A better understanding of mechanosensitive pathological entities will help to develop new diagnostic and therapeutic strategies for keloids that can prevent, reduce or even reverse pathological scar progression in clinics.