Abstract
Background:
Chronic wounds have been associated with elevated burden of cellular senescence, a state of essentially irreversible cell cycle arrest, resistance to apoptosis, and a secretory phenotype. However, whether senescent cells contribute to wound chronicity in humans remains unclear. The objective of this article is to assess the role of clinicopathological characteristics and cellular senescence in time-to-healing of chronic wounds.
Methods:
A cohort of 79 patients with chronic wounds were evaluated in a single-center academic practice from February 1, 2005, to February 28, 2015, and followed for up to 36 months. Clinical characteristics and wound biopsies were obtained at baseline, and time-to-healing was assessed. Wound biopsies were analyzed histologically for pathological characteristics and molecularly for markers of cellular senescence. In addition, biopsy slides were stained for p16INK4a expression.
Results:
No clinical or pathological characteristics were found to have significant associations with time-to-healing. A Cox proportional hazard ratio model revealed increased CDKN1A (p21CIP1/WAF1) expression to predict longer time-to-healing, and a model adjusted for sex and epidermal hyperplasia revealed increased CDKN1A expression and decreased PAPPA expression to predict longer time-to-healing. Increased p16INK4a staining was observed in diabetic wounds compared to non-diabetic wounds, and the same association was observed in the context of high dermal fibrosis.
Conclusions:
The findings of this pilot study suggest that senescent cells contribute to wound chronicity in humans, especially in diabetic wounds.
Keywords: geriatric dermatology, wound healing, ulcer
Introduction
Chronic wounds are non-healing, stalled wounds that cause significant morbidity, mortality, and healthcare burden.1, 2 They typically require complex and labor-intensive management, including wound debridement, irrigation, use of antibiotics or antiseptics, frequent dressing changes, and pressure offloading or compression.1, 3 Despite intensive management, most chronic wounds recur2, 4 and a significant proportion become complicated by cellulitis, chronic osteomyelitis, and gangrene, requiring resection, revascularization surgeries, skin grafts, and amputation.1, 2, 5 Additionally, chronic wounds are associated with substantial mortality: foot ulcers are associated with 2.5 times higher mortality at 5 years in patients with diabetes mellitus.2 Diabetic foot ulcers are estimated to affect 1 to 3.5 million adults in the United States and cost $9 to $13 billion annually.2, 6 Hence, it is crucial to better characterize non-healing wounds and develop root-cause targeting wound management strategies.
Chronic wounds result from a failure to regulate several biological processes that lead to repair of structural and functional defects.1, 7 These processes include hemostasis, inflammation, angiogenesis, cellular senescence, and extracellular matrix remodeling, all of which are coordinated by cytokines, proteases, and other chemical messengers.7 Disruptions in any of these processes or signals could impair wound healing, resulting in non-healing, stalled wounds. For instance, endothelial disruption and persistent inflammation are thought to contribute to chronic vascular wounds.1 Diabetic wounds involve disruption of multiple pathways, including neuropathy, vascular damage, dysfunctional immune responses, and diminished tissue integrity.5, 8
Cellular senescence is a state of essentially irreversible cell cycle arrest induced by intrinsic or extrinsic stressors, such as oxidative or genotoxic stress, nutrient deprivation, mitochondrial dysfunction, and oncogene activation.9, 10 Senescent cells are associated with a senescence-associated secretory phenotype (SASP), or the secretion of various factors involved in inflammation, tissue remodeling, angiogenesis, fibrosis, and other cellular processes.9, 10 Transient cellular senescence induces an acute SASP, which has been postulated to benefit acute wound repair by contributing to the pro-inflammatory wound cascade, promoting wound closure, and preventing excessive fibrosis.11–15 However, accumulation of senescent cells and chronic SASP have been implicated in pathological healing, contributing to disordered inflammation, excessive dermal proteolysis, and reduced regenerative capability, especially in the context of diabetes mellitus.16–19 However, more evidence is needed to establish the role of senescent cells and the SASP in chronic wound healing.
Currently, standard treatments do not target molecular mechanisms underlying non-healing wounds. Given that senescent cells are hypothesized to contribute to wound chronicity, characterizing the role of cellular senescence in non-healing wounds could aid in predicting prognosis and identifying therapeutic targets. To account for the heterogeneity of chronic wounds, identifying clinicopathological risk factors is also essential. This pilot study was a prospective cohort study that evaluated associations between clinicopathological factors or cellular senescence-associated molecular markers and chronicity of stalled wounds. Skin biopsies were obtained from patients with chronic, stalled wounds, who were followed for up to 36 months. The aim of this study was to assess and characterize clinicopathological factors and cellular senescence markers in chronic wound biopsies of patients with and without type 2 diabetes mellitus.
Methods
Patients evaluated at the Mayo Clinic Wound Care Clinic in Rochester, MN for chronic wound management from February 1, 2005, to February 28, 2015 were recruited (Fig. 1). To characterize chronic stalled wounds, inclusion criteria included wound duration of at least 6 weeks, wound surface not less than 2 cm2 or greater than 25 cm2 without active infection at evaluation, and less than 50% reduction in wound size after 30 days of standard wound care, defined as adequate off-loading for diabetic ulcers and compression for venous ulcers. Exclusion criteria included acute wounds misclassified as chronic wounds, wounds resulting in amputation, or wounds without tissue available for pathological or molecular analysis. For participants enrolled in the study, clinical characteristics were collected. Skin biopsies were obtained for histological analysis and molecular analysis by reverse transcription quantitative PCR (RT-qPCR).
Figure 1. Study Flow Diagram.
Enrollment of patients, inclusion and exclusion criteria, and other potential sources of retrospective bias, such as the unavailability of tissue for molecular analysis and no tissue collection within one week of presentation to wound care clinic. *Standard-of-care defined as adequate off-loading for diabetic ulcers and compression for venous ulcers along with wound care.
Categorical patient characteristics are reported as frequencies and percentages, and continuous characteristics are reported as mean ± standard deviation. Univariate Cox proportional hazards models were used to assess clinical and pathological characteristics for an association with wound healing. Multivariable Cox proportional hazards models were used to assess cellular senescence markers after adjusting for sex and epidermal hyperplasia. Hazard ratios (and 95% confidence intervals [CIs]) were reported and P-values were calculated using the Wald test. Cumulative incidence of wound healing following wound care presentation (and 95% CIs) were calculated using the Kaplan–Meier method, and p-values were calculated with the log-rank test. Analyses were performed with SAS version 9.4 (SAS Institute, Cary, North Carolina, United States). P-values <.05 was considered statistically significant.
Reverse transcription-quantitative PCR (RT-qPCR) analysis for this study was performed in a batched process as previously described.20 RNA purification was from FFPE tissue (Qiagen). RT-qPCR was performed using the BioMark HD System and dynamic integrated fluid circuits (Fluidigm). All cDNA was pre-amplified (TaqMan Preamp Master Mix, Applied Biosystems). TaqMan Gene Expression Master Mix (Applied Biosystems) was used to help array-based quantitative PCR. After thermal cycling, raw cycle threshold (Ct) data, defined as PCR amplification cycles at which the RNA signal exceeded background noise, were checked for linear amplification. Higher Ct value indicated lower RNA expression. For this study, gene expression was corrected by the mean obtained from housekeeping genes (RLP0, RLP8, and ACTB) using the ΔCt method. A total of 29 genes were quantified in the wound biopsies.
Immunohistochemical staining of CDKN2A/p16INK4a and haematoxylin and eosin (H&E) was performed on formalin-fixed, paraffin-embedded ulcer biopsy sections. Imaging and quantification were performed by one individual to ensure consistency in the analysis. Prior to imaging, each slide was assigned an identification code by an independent source that masked group assignment (diabetic or non-diabetic) until the code was matched to a clinic number during analysis. Images were acquired using a Zeiss Axio Scan Z1 digital slide scanner and analyzed using FIJI software.21 H&E-stained sections were used to help identify the epidermal-dermal junction (EDJ) in adjacent p16INK4a-stained sections, where each EDJ length was measured using FIJI and converted to micrometers. Total numbers of cells that were positive for p16INK4a along the EDJ were counted and used to calculate the number of p16INK4a cells per millimeter of the EDJ. For skin biopsies with multiple sections (up to 3 sections), the EDJ length and p16INK4a+ count per section were added together. GraphPad Prism 9 was used for graphing and statistical analysis of the mean p16INK4a density along the EDJ (Mann-Whitney test).
Results
Among participants enrolled in the study, cumulative incidence of wound healing was poor, consistent with stalled, chronic wounds (Fig. 2). By 3, 6, and 9 months after clinical presentation, the cumulative incidence of being healed was 20.2% (95% CI, 9.2–29.9%), 35.3% (95% CI, 20.8–47.2%), and 47.8% (95% CI, 31.1–60.4%), respectively. Even with intensive management, including primary and secondary dressings, irrigation, and debridement with wound vacuum-assisted closure (VAC) placement, skin graft, and offloading of pressure for diabetic wounds or compression for venous ulcers, approximately half of the wounds were not healed by one year following wound care presentation.
Figure 2. Cumulative incidence of wound healing following the date of wound care presentation (A) and wound chronicity (B).
A: Using the Kaplan-Meier method and adjusting for follow-up duration, cumulative incidence of wound healing following wound care presentation is poor. By 3, 6, and 9 months after presentation, cumulative incidence of being healed was 20.2% (95% CI=9.2–29.9%), 35.3% (95% CI=20.8–47.2%), and 47.8% (95% CI=31.1–60.4%), respectively. B: Example of wound chronicity in 79-year-old male with non-vascular ulcer treated with standard primary and secondary dressings, irrigation, and debridement with wound vac placement and subsequent split thickness skin graft (not shown) for right lower extremity wounds.
Clinicopathological characteristics were evaluated for associations with wound healing or chronicity (Fig. 3). None of the clinical characteristics evaluated were associated with statistically significant (P<.05) differences in incidence of wound healing (Table 1). However, characteristics that approached significance included greater incidence of healing with negative steroid history (P=.06), male gender (P=.08), lower extremity wound location (P=.09), and unilateral compared to bilateral wounds (P=.12). Clinical characteristics associated with the highest probabilities of unhealed wounds by 9 months were wound location in an upper extremity (67%; 95% CI = 30 to 100%), bilateral wounds (68%; 95% CI=50–93%), peripheral vascular disease (68%; 95% CI=50–92%), steroid history (81%; 95% CI=63–100%), and chemotherapy and/or radiation (67%; 95% CI=30–100%).
Figure 3. Cumulative incidence of wound healing following the date of wound care presentation, by (A) sex, (B) steroid history, (C) wound laterality, and (D) presence of epidermal hyperplasia.
Based on two-sided log-rank test, the listed factors approached statistical significance: sex (P=.08), steroid history (P=.06), bilateral location (P=.12), and epidermal hyperplasia (P=.11). Factors supporting wound healing included male sex, absence of steroid history, unilateral wound, and presence of epidermal hyperplasia.
Table 1.
Summary of univariate results evaluating clinical characteristics for an association with wound healing.
| Wound characteristics | Total (n = 79) | Level | No. of patients | No. healed in 9 months | P-value | Probability of unhealed wound by 9 months (95% CI)a |
|---|---|---|---|---|---|---|
| Age at wound care presentation (years) | Age (tertiles) at presentation (years) | |||||
| Mean (SD)a | 62.5 (13.1) | < 58.5 | 26 | 10 | 0.18 | 50.00% (31.84, 78.49) |
| Range | 25.9–80.4 | 58.5–70 | 26 | 3 | 64.29% (38.52, 100.0) | |
| ≥ 71 | 27 | 11 | 46.24% (28.63, 74.70) | |||
| Gender | ||||||
| Male | 30 (38.0%) | Male | 30 | 13 | 0.08 | 35.21% (18.90, 65.57) |
| Female | 49 (62.0%) | Female | 49 | 11 | 64.24% (48.91, 84.37) | |
| Wound location | ||||||
| Lower extremity | 73 (92.4%) | Lower extremity | 73 | 21 | 0.09 | 54.03% (40.89, 71.39) |
| Upper extremity | 3 (3.8%) | Upper extremity | 3 | 1 | 66.67% (29.95, 100.0) | |
| Trunk | 3 (3.8%) | Trunk | 3 | 2 | -- | |
| Bilateral wound | 29 (36.7%) | No | 50 | 18 | 0.12 | 42.08% (27.08, 65.39) |
| Yes | 29 | 6 | 68.47% (50.34, 93.14) | |||
| Infection within 2 weeks | 22 (27.8%) | No | 57 | 18 | 0.39 | 53.97% (40.12, 72.59) |
| Yes | 22 | 6 | 52.52% (30.00, 91.95) | |||
| Vascular type | ||||||
| None | 40 (50.6%) | None | 40 | 15 | 0.36 | 39.46% (23.97, 64.97) |
| Peripheral vascular disease | 31 (39.2%) | Peripheral vascular disease | 31 | 7 | 67.90% (49.94, 92.33) | |
| Vasculitis | 8 (10.1%) | Vasculitis | 8 | 2 | 53.33% (21.42, 100.0) | |
| Steroid history | 27 (34.2%) | No | 52 | 21 | 0.06 | 43.26% (29.55, 63.33) |
| Yes | 27 | 3 | 80.64% (62.94, 100.0) | |||
| Chemotherapy and/or radiation | 3 (3.8%) | No | 76 | 23 | 0.71 | 51.03% (38.01, 68.51) |
| Yes | 3 | 1 | 66.67% (29.95, 100.0) | |||
| Type 2 diabetes | 28 (35.4%) | No | 51 | 17 | 0.61 | 48.74% (33.91, 70.04) |
| Yes | 28 | 7 | 61.59% (41.83, 90.69) | |||
| Smoking history | 24 (30.4%) | No | 55 | 19 | 0.77 | 50.35% (36.44, 69.57) |
| Yes | 24 | 5 | 59.67% (35.21, 100.0) |
CI: confidence interval, SD: standard deviation
Wound biopsy characteristics evaluated histologically included dermal fibrosis, epidermal hyperplasia, and neutrophilic inflammation (Fig. S1). Similarly, no pathological findings evaluated were associated with statistically significant differences in wound healing (Table 2). An increased incidence of wound healing was observed with epidermal hyperplasia (P=.11), which approached statistical significance. Biopsy characteristics associated with the highest probabilities of unhealed wounds by 9 months were mild dermal fibrosis (68%; 95% CI=46–100%), no inflammation (71%; 95% CI=45–100%), and no epidermal hyperplasia (67%; 95% CI=48–95%).
Table 2.
Summary of univariate results evaluating biopsy characteristics for an association with wound healing.
| Biopsy characteristics | Total (n = 79) | Level | No. of patients | No. healed in 9 months | P-value | Probability of unhealed wound by 9 months (95% CI)a |
|---|---|---|---|---|---|---|
| Wound diameter at biopsy (cm, n = 62) | Diameter (tertiles) at biopsy (cm) | |||||
| Mean (SD)a | 8.4 (7.1) | < 3.5 | 20 | 7 | 0.79 | 58.74% (39.22, 87.98) |
| Range | (0.3–30.0) | 3.5–9.9 | 20 | 4 | 62.07% (37.76, 100.0) | |
| ≥ 10 | 22 | 6 | 51.69% (28.97, 92.23) | |||
| Biopsy depth | ||||||
| Superficial dermis | 16 (20.3%) | Superficial dermis | 16 | 4 | 0.77 | 64.29% (41.19, 100.0) |
| Deep dermis | 30 (38.0%) | Deep dermis | 30 | 10 | 50.09% (31.85, 78.77) | |
| Fat | 33 (41.8%) | Fat | 33 | 10 | 46.77% (27.84, 78.58) | |
| Degree of dermal fibrosis | ||||||
| None | 2 (2.5%) | None | 2 | 0 | 0.48 | -- |
| Mild | 22 (27.8%) | Mild | 22 | 4 | 68.02% (45.72, 100.0) | |
| Severe | 55 (69.6%) | Severe | 55 | 20 | 47.00% (33.03, 66.89) | |
| Dermal angioplasia | 61 (77.2%) | No | 18 | 5 | 0.49 | 57.54% (35.00, 94.59) |
| Yes | 61 | 19 | 50.90% (36.64, 70.70) | |||
| Acute inflammation | 32 (40.5%) | No | 47 | 13 | 0.50 | 55.66% (39.82, 77.80) |
| Yes | 32 | 11 | 47.71% (29.80, 76.38) | |||
| Chronic inflammation | 54 (68.4%) | No | 25 | 6 | 0.25 | 60.61% (40.21, 91.35) |
| Yes | 54 | 18 | 48.30% (33.44, 69.78) | |||
| Degree of inflammation | ||||||
| None | 11 (13.9%) | None | 11 | 2 | 0.47 | 71.43% (44.71, 100.0) |
| Sparse | 27 (34.2%) | Sparse | 27 | 11 | 49.61% (32.27, 76.24) | |
| Moderate | 16 (20.3%) | Moderate | 16 | 5 | -- | |
| Diffuse | 25 (31.6%) | Diffuse | 25 | 6 | 57.15% (34.81, 93.84) | |
| Epidermal hyperplasia | 44 (55.7%) | No | 35 | 6 | 0.11 | 67.32% (47.85, 94.71) |
| Yes | 44 | 18 | 43.93% (29.26, 65.95) | |||
| Degree of fibroblast abundance | ||||||
| Low | 19 (24.1%) | Low | 19 | 6 | 0.90 | 42.31% (19.53, 91.67) |
| Intermediate | 16 (20.3%) | Intermediate | 16 | 7 | 43.51% (22.77, 83.14) | |
| High | 44 (55.7%) | High | 44 | 11 | 61.55% (45.53, 83.22) | |
| Blood vessel thickening | 13 (16.5%) | No | 66 | 17 | 0.21 | 60.90% (47.41, 78.22) |
| Yes | 13 | 7 | 22.22% (6.55, 75.44) |
CI: confidence interval, SD: standard deviation
In addition to established clinicopathological characteristics, associations between molecular markers of cellular senescence and wound healing outcomes were evaluated (Table 3). Cox proportional hazard regression models demonstrated decreased probability of healing with higher expression of CDKN1A (hazard ratio (HR)=0.65, P=.02). Other associations that approached significance included increased probability of healing with higher expression of PAPPA (HR=1.41, P=.07) and VCAM1 (HR=1.45, P=.08), and decreased healing with higher expression of IL-8 (HR=0.69, P=.08) and TNC (HR=0.64, P=.06). When the models were adjusted for sex and epidermal hyperplasia, an increased probability of healing was associated with higher expression of PAPPA (HR=1.43, P=.05) and decreased healing with higher expression of CDKN1A (HR=0.66, P=.04). Associations that approached significance in the adjusted model included increased healing with higher expression of VCAM1 (HR=1.49, P=.06) and MMP3 (HR=1.48, P=.07).
Table 3.
Results from evaluating senescence-associated genes for association with healing.
| Gene | No. | Unadjusted analysis | Adjusted analysisc | ||
|---|---|---|---|---|---|
| Unadjusted HR (95% CI)a,d |
P-value | Adjusted HR (95% CI)a,d |
P-value | ||
| KRT14 | 70 | 1.01 (0.70,1.46) | 0.95 | 0.90 (0.61,1.31) | 0.58 |
| MLANA | 68 | 0.91 (0.59,1.43) | 0.69 | 0.91 (0.57,1.45) | 0.69 |
| MITF | 69 | 0.68 (0.43,1.08) | 0.10 | 0.75 (0.46,1.21) | 0.24 |
| TP53 | 69 | 1.02 (0.63,1.64) | 0.94 | 1.01 (0.65,1.57) | 0.97 |
| PLAT | 69 | 0.83 (0.51,1.35) | 0.44 | 0.93 (0.56,1.56) | 0.79 |
| CDKN2A2 | 64 | 1.07 (0.66,1.71) | 0.79 | 1.12 (0.69,1.82) | 0.65 |
| CDKN1A | 70 | 0.65 (0.45,0.94) | 0.02 | 0.66 (0.45,0.98) | 0.04 |
| CDKN2A1 | 69 | 0.83 (0.46,1.48) | 0.52 | 0.81 (0.45,1.44) | 0.47 |
| IL8 | 70 | 0.69 (0.45,1.04) | 0.08 | 0.76 (0.49,1.20) | 0.24 |
| IL1A | 70 | 0.78 (0.45,1.37) | 0.39 | 0.82 (0.44,1.53) | 0.54 |
| IL6 | 70 | 0.90 (0.59,1.37) | 0.61 | 1.00(0.61,1.63) | 0.99 |
| IGFBP4 | 70 | 0.96 (0.54,1.70) | 0.89 | 1.06 (0.56,1.99) | 0.86 |
| IL1B | 70 | 0.91 (0.58,1.42) | 0.67 | 1.00 (0.60,1.67) | 0.99 |
| INHBA | 70 | 1.32 (0.86,2.04) | 0.21 | 1.26 (0.79,2.03) | 0.34 |
| PAPPA | 70 | 1.41 (0.97,2.05) | 0.07 | 1.43 (1.00,2.04) | 0.05 |
| VCAM1 | 70 | 1.45 (0.96,2.18) | 0.08 | 1.49 (0.98,2.25) | 0.06 |
| CXCL1 | 70 | 1.36 (0.91,2.03) | 0.14 | 1.36 (0.92,2.02) | 0.12 |
| SERPINEB2 | 70 | 1.06 (0.70,1.61) | 0.77 | 1.05 (0.68,1.63) | 0.82 |
| SPP1 | 70 | 0.98 (0.63,1.51) | 0.91 | 1.07 (0.66,1.73) | 0.79 |
| SERPINE1 | 70 | 1.08 (0.70,1.65) | 0.74 | 1.32 (0.85,2.07) | 0.22 |
| TNC | 69 | 0.64 (0.40,1.01) | 0.06 | 0.70 (0.42,1.17) | 0.18 |
| IGFBP2 | 70 | 1.10 (0.71,1.72) | 0.66 | 1.04 (0.66,1.64) | 0.86 |
| MMP2 | 70 | 1.02 (0.64,1.61) | 0.94 | 1.19 (0.71,2.00) | 0.51 |
| MMP3 | 70 | 1.40 (0.91,2.15) | 0.13 | 1.48 (0.97,2.25) | 0.07 |
| MMP12 | 64 | 1.02 (0.61,1.69) | 0.95 | 1.15 (0.65,2.03) | 0.64 |
| CCL2 | 69 | 0.78 (0.49,1.23) | 0.29 | 0.91 (0.54,1.54) | 0.73 |
| MMP10 | 70 | 1.27 (0.81,2.00) | 0.29 | 1.40 (0.88,2.22) | 0.15 |
| TGFA | 69 | 1.15 (0.76,1.73) | 0.50 | 1.15 (0.75,1.78) | 0.52 |
| MMP9 | 70 | 1.14 (0.71,1.83) | 0.58 | 1.31 (0.81,2.13) | 0.26 |
CI: confidence interval, HR: hazard ratio.
Bolded results have a p-value < 0.05.
Each gene was evaluated as a z-score in a separate Cox proportional hazards regression model, adjusted for sex and epidermal hyperplasia.
Each hazard ratio is per a 1-unit decrease in each z-score.
Chronic wound biopsies of patients with and without type 2 diabetes mellitus were compared based on p16INK4a staining, a senescence marker. The epidermal-dermal junctions of skin surrounding diabetic wounds showed significantly increased density of p16INK4a+ cells compared to non-diabetic wounds (P=.007; 11 non-diabetic vs. 20 diabetic) (Fig. 4). Diabetic wounds were further subcategorized by level of dermal fibrosis. Among wounds with high levels of dermal fibrosis, diabetic skin had significantly higher densities of p16INK4a+ cells compared to non-diabetic skin (P=.008). Staining of p16INK4a was also present in papillary and reticular dermis (Fig. S2).
Figure 4. A. Hematoxylin-eosin (H&E) and p16INK4a immunohistochemistry staining in diabetic and non-diabetic wound biopsies.
Red dotted line shows epidermal-dermal junction (EDJ). Black dotted line shows wound edge. E = epidermis. D = dermis. Arrows point to subset of p16INK4a+ cells along the EDJ. B. Quantification of p16INK4a+ cells along the EDJ. Wound biopsies in patients with diabetes mellitus have significantly more p16INK4a+ cells along the EDJ (Mann-Whitney U-test P=.007; 11 non-diabetic, 20 diabetic). C. Quantification of p16INK4a+ cells along EDJ in non-diabetic and diabetic patients with low and high dermal fibrosis. Among wound biopsies with high dermal fibrosis, patients with diabetes had significantly higher density of p16INK4a+ cells along the EDJ compared to patients without diabetes (P=.008). Among patients with diabetes, differences in EDJ staining for p16INK4a+ cells between those with low and high dermal fibrosis approached significance (P=.09). All other comparisons were not statistically significant. Mann-Whitney U-test was performed for all comparisons. aH&E: hematoxylin and eosin. EDJ: epidermal-dermal junction.
Discussion
Not only do chronic wounds adversely affect patient quality of life, but their management also requires substantial provider time, cost, and physical restrictions.1Even with intensive management, including multiple dressings, irrigation, debridement, pressure offloading or compression, and skin grafts, chronic wounds can become stalled, taking months to years to heal. Hence, more investigation is needed to understand the pathophysiology of stalled wounds and identify molecular targets, such as senescent cells or SASP factors, and to advance translation of therapeutic targets into clinical practice.7
Molecular markers of cellular senescence predicted wound chronicity with statistical significance, even when the clinicopathological characteristics examined in this study could not. Specifically, increased expression of CDKN1A/p21, a canonical marker of senescence, was associated with delayed healing. In addition, when adjusted for sex and epidermal hyperplasia, increased expression of PAPPA was associated with earlier healing. Comparing diabetic and non-diabetic chronic wounds, diabetic wounds had significantly higher densities of p16INK4a+ cells at the EDJ, suggesting elevated abundance of senescent cells. The subset of wounds with dermal fibrosis revealed similar differences, with significantly higher concentrations of p16INK4a+ cells at the epidermal-dermal junction in diabetic vs. non-diabetic chronic wounds.
These findings are consistent with previous studies of cellular senescence in chronic wounds. Stanley and Osler (2001) examined human venous ulcers for senescent cells using X-gal staining for SA-β-galactosidase, a senescence marker.19 They found that the proportion of cells that were senescent correlated with time-to-healing. Similarly, Wilkinson, et al. (2019) investigated cellular senescence in wounds of diabetic and aged mice.16 They observed increased accumulation of SA-β-galactosidase in diabetic and old mice compared to young healthy controls. In diabetic mice specifically, increased p16INK4a staining was observed, as well as increased expression of CDKN1A, CDKN1B, CDKN2A, and TP53 by quantitative RT-PCR. This study included diabetic and non-diabetic chronic human wounds, allowing us to analyze clinical characteristics and compare wound etiologies.
The association between increased CDKN1A/p21CIP1/WAF1 expression and wound chronicity is especially pertinent because CDKN1A/p21CIP1/WAF1 is hypothesized to cause G1 cell cycle arrest in senescent cells.22, 23 p21CIP1/WAF1, the protein product of CDKN1A, has been identified as a primary mediator of senescence in vitro. Although PAPPA has been identified as a SASP gene, its roles are more heterogeneous. PAPPA produces a metalloproteinase that regulates insulin-like growth factors and is associated with cellular and chronological aging24, 25 as well as inflammation and wound healing.26, 27 The association between increased PAPPA expression and earlier wound healing is consistent with observations that upregulation pro-inflammatory cytokines was observed to induce PAPPA,26 and PAPPA was expressed locally in acute wound healing in skin.27
Skin aging is characterized by changes to the EDJ, including flattening, so senescent cells were postulated to be concentrated in the region.28, 29 Like p21CIP1/WAF1, p16INK4a is a cell cycle inhibitor and canonical senescence marker.30 Hence, the increase in p16INK4a+ cells at the EDJ of skin surrounding diabetic wounds suggests an increased senescent cell burden in diabetic skin. This is consistent with previous results demonstrating increased cellular senescence in patients with type 2 diabetes mellitus.31 Furthermore, accumulation of senescent cells in diabetic skin could be mediated by repetitive trauma or injury, which are associated with dermal fibrosis. Among biopsies with high dermal fibrosis, diabetic skin had significantly more p16INK4a+ cells at the EDJ than non-diabetic skin, suggesting that patients with diabetes mellitus could have reduced clearance of senescent cells following skin trauma or injury leading to their gradual accumulation. Moreover, this association suggests that cellular senescence could be implicated in the predisposition for patients with diabetes mellitus to develop chronic wounds.2
Our findings indicate that molecular markers of cellular senescence could predict wound healing of chronic wounds. In clinical practice, prognosticating chronic wound outcomes using molecular markers could aid in care planning, or monitoring management and progression. The association between cellular senescence and wound chronicity also presents a therapeutic target. Our findings support the investigation of oral or topical senolytics, which specifically eliminate senescent cells, and senomorphics, which inhibit SASP factors, to improve chronic wound healing outcomes or prevent chronic wounds, especially in patients with type 2 diabetes mellitus.
As a pilot study, this study was limited in sample size and number of molecular markers. Despite the limited number of patients and genes evaluated, the RT-qPCR and biopsy staining suggest an association between cellular senescence and wound chronicity that warrants further investigation. Future studies could aim to include a larger range of molecular markers or unbiased analysis by obtaining transcriptomic, proteomic, or epigenomic data. Studies could also capture trends over time or with management by obtaining wound biopsies at additional timepoints.
Supplementary Material
Supplementary Figure 1. Histological wound biopsy characteristics. Shown are images of H&E-stained slides for low and advanced degree of dermal fibrosis (A, B), epidermal hyperplasia (C, D), and degree of inflammation (E, F). Black dotted lines represent fibrosis. Blue dotted lines represent epidermal hyperplasia. Black triangles represent neutrophilic inflammation. Objectives: A = 4x, B = 2x, and C-F = 10x.
Supplementary Figure 2. p16INK4a immunohistochemical staining in the papillary (P) and reticular (R) dermis. Black dashed line shows approximate boundary between papillary (superior) and reticular (inferior) dermis. Arrows point to subsets of p16INK4a+ cells. Scale bar is shown.
Acknowledgements
We gratefully acknowledge Amy Weaver and Austin Todd for their statistical expertise, Julia Tomtschik for her illustration, and Traci Paulson for her help with formatting.
Funding source:
G. Yu was supported by the Mayo Clinic Medical Scientist Training Program institutional training grant (T32 GM065841). J. L. Kirkland and T. Tchkonia were supported by NIH grants R37AG013925, R33AG061456, and P01AG062413, the Connor Fund, Robert J. and Theresa W. Ryan, and the Noaber Foundation. S. P. Wyles was supported by the Mayo Clinic Department of Dermatology and NIH grant R03AG082919-01.
List of Abbreviations
- CI
confidence interval
- Ct
cycle threshold
- EDJ
epidermal-dermal junction
- H&E
haematoxylin and eosin
- HR
hazard ratio
- RT-qPCR
reverse transcription quantitative polymerase chain reaction
- SASP
senescence-associated secretory phenotype
- VAC
vacuum-assisted closure
Footnotes
Conflict of interest: T. Tchkonia and J. L. Kirkland have a financial interest related to this research, including patents, and pending patents covering senolytic drugs and their uses that are held by Mayo Clinic. This research has been reviewed by the Mayo Clinic Conflict of Interest Review Board and was conducted in compliance with Mayo Clinic conflict of interest policies.
Abstract Presentation: Mayo Clinic Symposium on Regenerative Medicine and Surgery - Hybrid meeting November 4–7, 2021, in Phoenix, AZ.
References
- 1.Singer AJ, Tassiopoulos A, Kirsner RS. Evaluation and Management of Lower-Extremity Ulcers. New Engl J Med 2017; 377(16): 1559–67. [DOI] [PubMed] [Google Scholar]
- 2.Armstrong DG, Boulton AJM, Bus SA. Diabetic Foot Ulcers and Their Recurrence. New Engl J Med 2017; 376(24): 2367–75. [DOI] [PubMed] [Google Scholar]
- 3.Baum CL, Arpey CJ. Normal cutaneous wound healing: clinical correlation with cellular and molecular events. Dermatol Surg 2005; 31(6): 674–86; discussion 86. [DOI] [PubMed] [Google Scholar]
- 4.Abbade LPF, Lastória S. Venous ulcer: epidemiology, physiopathology, diagnosis and treatment. Int J Dermatol 2005; 44(6): 449–56. [DOI] [PubMed] [Google Scholar]
- 5.Noor S, Khan RU, Ahmad J. Understanding Diabetic Foot Infection and its Management. Diabetes Metab Syndr 2017; 11(2): 149–56. [DOI] [PubMed] [Google Scholar]
- 6.Rice JB, Desai U, Cummings AKG, Birnbaum HG, Skornicki M, Parsons NB. Burden of Diabetic Foot Ulcers for Medicare and Private Insurers. Diabetes Care 2014; 37(3): 651–58. [DOI] [PubMed] [Google Scholar]
- 7.Eming SA, Martin P, Tomic-Canic M. Wound repair and regeneration: Mechanisms, signaling, and translation. Sci Transl Med 2014; 6(265): 265sr6–65sr6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sawaya AP, Stone RC, Brooks SR, Pastar I, Jozic I, Hasneen K, et al. Deregulated immune cell recruitment orchestrated by FOXM1 impairs human diabetic wound healing. Nat Commun 2020; 11(1): 4678. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Gorgoulis V, Adams PD, Alimonti A, Bennett DC, Bischof O, Bishop C, et al. Cellular Senescence: Defining a Path Forward. Cell 2019; 179(4): 813–27. [DOI] [PubMed] [Google Scholar]
- 10.Herranz N, Gil J. Mechanisms and functions of cellular senescence. J Clin Invest 2018; 128(4): 1238–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jun J-I, Lau LF. The matricellular protein CCN1 induces fibroblast senescence and restricts fibrosis in cutaneous wound healing. Nat Cell Biol 2010; 12(7): 676–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Krizhanovsky V, Yon M, Dickins RA, Hearn S, Simon J, Miething C, et al. Senescence of Activated Stellate Cells Limits Liver Fibrosis. Cell 2008; 134(4): 657–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Demaria M, Ohtani N, Youssef SA, Rodier F, Toussaint W, Mitchell JR, et al. An essential role for senescent cells in optimal wound healing through secretion of PDGF-AA. Dev Cell 2014; 31(6): 722–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hiebert P, Wietecha MS, Cangkrama M, Haertel E, Mavrogonatou E, Stumpe M, et al. Nrf2-Mediated Fibroblast Reprogramming Drives Cellular Senescence by Targeting the Matrisome. Dev Cell 2018; 46(2): 145–61.e10. [DOI] [PubMed] [Google Scholar]
- 15.Rhinn M, Ritschka B, Keyes WM. Cellular senescence in development, regeneration and disease. Development 2019; 146(20). [DOI] [PubMed] [Google Scholar]
- 16.Wilkinson HN, Clowes C, Banyard KL, Matteuci P, Mace KA, Hardman MJ. Elevated Local Senescence in Diabetic Wound Healing Is Linked to Pathological Repair via CXCR2. J Invest Dermatol 2019; 139(5): 1171–81.e6. [DOI] [PubMed] [Google Scholar]
- 17.Wilkinson HN, Hardman MJ. Senescence in Wound Repair: Emerging Strategies to Target Chronic Healing Wounds. Front Cell Dev Biol 2020; 8: 773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Tomic-Canic M, DiPietro LA. Cellular Senescence in Diabetic Wounds: When Too Many Retirees Stress the System. J Invest Dermatol 2019; 139(5): 997–99. [DOI] [PubMed] [Google Scholar]
- 19.Stanley A, Osler T. Senescence and the healing rates of venous ulcers. J Vasc Surg 2001; 33(6): 1206–11. [DOI] [PubMed] [Google Scholar]
- 20.Bellomo D, Arias-Mejias SM, Ramana C, Heim JB, Quattrocchi E, Sominidi-Damodaran S, et al. Model Combining Tumor Molecular and Clinicopathologic Risk Factors Predicts Sentinel Lymph Node Metastasis in Primary Cutaneous Melanoma. JCO Precis Oncol 2020; 4: 319–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods 2012; 9(7): 676–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ju Z, Choudhury AR, Rudolph KL. A Dual Role of p21 in Stem Cell Aging. Ann N Y Acad Sci 2007; 1100(1): 333–44. [DOI] [PubMed] [Google Scholar]
- 23.Georgakilas AG, Martin OA, Bonner WM. p21: A Two-Faced Genome Guardian. Trends Mol Med 2017; 23(4): 310–19. [DOI] [PubMed] [Google Scholar]
- 24.Conover CA, Bale LK. Loss of pregnancy-associated plasma protein A extends lifespan in mice. Aging Cell 2007; 6(5): 727–29. [DOI] [PubMed] [Google Scholar]
- 25.Saul D, Kosinsky RL, Atkinson EJ, Doolittle ML, Zhang X, LeBrasseur NK, et al. A new gene set identifies senescent cells and predicts senescence-associated pathways across tissues. Nat Commun 2022; 13(1): 4827. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Resch ZT, Chen B-K, Bale LK, Oxvig C, Overgaard MT, Conover CA. Pregnancy-Associated Plasma Protein A Gene Expression as a Target of Inflammatory Cytokines. Endocrinology 2004; 145(3): 1124–29. [DOI] [PubMed] [Google Scholar]
- 27.Chen B-K, Leiferman KM, Pittelkow MR, Overgaard MT, Oxvig C, Conover CA. Localization and Regulation of Pregnancy-Associated Plasma Protein A Expression in Healing Human Skin. J Clin Endocrinol Metab 2003; 88(9): 4465–71. [DOI] [PubMed] [Google Scholar]
- 28.Langton AK, Halai P, Griffiths CE, Sherratt MJ, Watson RE. The impact of intrinsic ageing on the protein composition of the dermal-epidermal junction. Mech Ageing Dev 2016; 156: 14–6. [DOI] [PubMed] [Google Scholar]
- 29.Low E, Alimohammadiha G, Smith LA, Costello LF, Przyborski SA, von Zglinicki T, et al. How good is the evidence that cellular senescence causes skin ageing? Ageing Res Rev 2021; 71: 101456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Krishnamurthy J, Torrice C, Ramsey MR, Kovalev GI, Al-Regaiey K, Su L, et al. Ink4a/Arf expression is a biomarker of aging. J Clin Invest 2004; 114(9): 1299–307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Palmer AK, Gustafson B, Kirkland JL, Smith U. Cellular senescence: at the nexus between ageing and diabetes. Diabetologia 2019; 62(10): 1835–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure 1. Histological wound biopsy characteristics. Shown are images of H&E-stained slides for low and advanced degree of dermal fibrosis (A, B), epidermal hyperplasia (C, D), and degree of inflammation (E, F). Black dotted lines represent fibrosis. Blue dotted lines represent epidermal hyperplasia. Black triangles represent neutrophilic inflammation. Objectives: A = 4x, B = 2x, and C-F = 10x.
Supplementary Figure 2. p16INK4a immunohistochemical staining in the papillary (P) and reticular (R) dermis. Black dashed line shows approximate boundary between papillary (superior) and reticular (inferior) dermis. Arrows point to subsets of p16INK4a+ cells. Scale bar is shown.