A pilot study evaluating protein abundance in pressure ulcer fluid from people with and without spinal cord injury

Abstract

Objective

To determine whether the biochemistry of chronic pressure ulcers differs between patients with and without chronic spinal cord injury (SCI) through measurement and comparison of the concentration of wound fluid inflammatory mediators, growth factors, cytokines, acute phase proteins, and proteases.

Design

Survey.

Setting

Tertiary spinal cord rehabilitation center and skilled nursing facilities.

Participants

Twenty-nine subjects with SCI and nine subjects without SCI (>18 years) with at least one chronic pressure ulcer Stage II, III, or IV were enrolled.

Outcome measures

Total protein and 22 target analyte concentrations including inflammatory mediators, growth factors, cytokines, acute phase proteins, and proteases were quantified in the wound fluid and blood serum samples. Blood samples were tested for complete blood count, albumin, hemoglobin A1c, total iron binding capacity, iron, percent (%) saturation, C-reactive protein, and erythrocyte sedimentation rate.

Results

Wound fluid concentrations were significantly different between subjects with SCI and subjects without SCI for total protein concentration and nine analytes, MMP-9, S100A12, S100A8, S100A9, FGF2, IL-1b, TIMP-1, TIMP-2, and TGF-b1. Subjects without SCI had higher values for all significantly different analytes measured in wound fluid except FGF2, TGF-b1, and wound fluid total protein. Subject-matched circulating levels of analytes and the standardized local concentration of the same proteins in the wound fluid were weakly or not correlated.

Conclusions

The biochemical profile of chronic pressure ulcers is different between SCI and non-SCI populations. These differences should be considered when selecting treatment options. Systemic blood serum properties may not represent the local wound environment.

Introduction

People with a spinal cord injury (SCI) are at increased risk of developing pressure ulcers with one-third of those with SCI reporting a pressure ulcer.^1,2 In the first year after SCI, pressure ulcers are the most frequent secondary medical complication.³ The risk of pressure ulcer development increases with the increase in duration of SCI.^3,4 Pressure ulcers in patients with SCI are often refractory to treatment and particularly slow to close.⁵ Post-SCI, many physiological factors may contribute to impaired wound healing, including mechanical and metabolic factors. Mechanical factors include excessive pressure during seating/lying, shear forces during dressing or transfers, and moisture, due to neurogenic bladder dysfunction and associated incontinence which can increase the impact of these mechanical factors.⁶ Metabolic factors include, but are not limited to anemia, glucose intolerance, elevations in inflammatory markers (e.g. C-reactive protein (CRP)), autonomic dysfunction below the level of injury, altered collagen metabolism, and altered vascular response.^7–10

Much research using experimental models has been done to characterize the complex sequence of biochemical and cellular events involved in healing an acute surgical or traumatic wound. Studies examining the biochemical composition of wound fluids have revealed that growth factors, cytokines, and proteases are present during the wound healing process.^11–17 Platelets are a source of growth factors including platelet-derived growth factor, epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), and transforming growth factor beta 1 (TGF-b1) secreted during the inflammatory phase.¹¹ Lymphocytes, neutrophils, mast cells, and macrophages contribute a number of factors that participate in wound healing and help debride the wound of necrotic tissue and bacteria, and degrade and remove damaged extracellular matrix components.^11,18,19 These include pro-inflammatory cytokines such as tumor necrosis factor alpha (TNF-a),¹⁸ interleukins (IL-1, IL-2, and IL-6),^20,21 interleukin-2 receptor subunit alpha (IL-2RA),²⁰ fibroblast growth factor (FGF) acidic^13,22 and basic growth factors,^11,18 and proteases such as elastase and collagenase. Lymphocytes and mast cells also secrete nerve growth factor (NGF), which affects fibroblast functionality and influences repair of inflammation-induced tissue damage.²³ Fibroblasts are a key cell during remodeling, releasing both matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs).^11,18,24,25 MMPs are a family of zinc-dependent endopeptidases secreted as inactive zymogens capable of degrading extracellular components and basement membrane proteins at neutral pH. MMPs 1, 2, 8, and 9 preferentially target specific extracellular matrix proteins. TIMPs (TIMPs 1, 2, 3, and 4) are small glycoproteins produced by fibroblasts that bind non-covalently and irreversibly to MMPs, resulting in their inactivation. The ratio of MMP-9/TIMP-1 has been reported to decrease as chronic pressure ulcers heal.²⁶ A balance between MMPs and TIMPs is prerequisite for wound healing.

Chronic wounds are defined as wounds that: “fail to progress through a normal, orderly, and timely sequence of repair or wounds that pass through the repair process without restoring anatomical and functional results”.²⁷ They include pressure, diabetic foot, venous, and arterial ulcers. Both in vivo and in vitro studies have detected differing levels of cytokines, growth factors, and proteases in acute versus chronic wounds and much effort has been expended by researchers to identify wound healing biomarkers.^{11–15,28–33} Wound fluid, blood serum, urine, sweat, and tears are body fluids being evaluated for biomarkers of healing. Potential wound biomarkers identified from pressure ulcer wound fluid include chemokine CXC ligand 9 (CXCL9), and calgranulins, members of the S100 family of proteins.³⁴ S100A8, A9, and A12 also known as calgranulin A, B, and C respectively, have antimicrobial and anti-inflammatory functions.³⁵ While it is recognized that an altered biochemical environment of a wound is a key underlying factor in chronic leg wounds and diabetic foot ulcers, less is known about pressure ulcers. Further, the biochemical environment in pressure ulcers of people with chronic SCI has not been investigated. People with SCI have specific immunological, inflammatory, and impaired autonomic nervous system control that predisposes them to pressure ulcers.³⁶

The majority of evidence for the treatment of pressure ulcers is based on research conducted in elderly individuals without SCI.⁶ A number of studies have sought to evaluate the differences in the blood/serum of people with SCI and pressure ulcers versus those without SCI and pressure ulcers in order to identify proteins that may be responsible for delayed wound closure.^20,37–39 This is the first report of a comparison of wound fluid and blood serum in these two groups. The purpose of this study was to determine whether the biochemistry of chronic pressure ulcers differs between the patients with and without SCI through measurement and comparison of the concentration of wound fluid inflammatory mediators, growth factors, cytokines, acute phase proteins, and proteases.

Methods

Subjects

Subjects were recruited via a poster campaign and referrals from treating clinicians at the participating sites. Participants were enrolled at sites in Buffalo, New York, Toronto and London, Ontario. Institutional Review Board approval for the study and the consent form was granted by the Catholic Health System and the Human Subjects Research Review Committee at Daemen College for the subjects without SCI. The Health Sciences Research Ethics Review Board, University of Western Ontario and the Toronto Rehab Institute Research Ethics Review Board granted approval for enrollment of the subjects with SCI. Adult individuals (>18 years) with Stages II, III, and IV chronic pressure ulcers with and without SCI provided written consent for participation. Potential subjects were excluded from the study if their pressure ulcer was being treated with negative pressure wound therapy, topical growth factors, or biological dressings containing proteins, including enzymatic debridement and collagen-containing dressings.

The subjects with SCI had chronic (>1 year) SCI (C1-T12 American Spinal Injury Association (ASIA) Impairment Scale (AIS) A–D) and at least one non-healing pressure ulcer. Subjects were current inpatients or outpatients affiliated with a tertiary spinal cord rehabilitation center in Toronto, Ontario, and inpatients outpatients, or residents of a skilled nursing facility in London, Ontario. Subjects without SCI were all inpatients recruited from long-term skilled nursing facilities in Buffalo, NY, USA and the surrounding area.

Wounds were photographed and the wound surface area was measured using the Visitrak system (Smith & Nephew, Hull, England). Depth was determined using a metal probe and measured at the deepest portion of the wound bed.

Wound fluid proteins

Real-time proteins present in pressure ulcers were collected by passive adsorption onto 15.2 cm sterile polyester tipped applicators (Puritan Medical Products Company Inc., Guilford, Maine). A new applicator was applied after saturation, until the entire wound bed was sampled. Care was taken to prevent bleeding by rolling the applicator without applying pressure. The wound bed was not cleaned prior to fluid collection. After sampling, the tip of each saturated applicator was broken off and contained within a sterile 2 ml screw cap microcentrifuge tube containing 155 µl of a freshly prepared sterile solution of phosphate-buffered saline (PBS) (Fluka 79378 BioUltra, Sigma-Aldrich, St. Louis, MO, USA) containing 1X protease and phosphatase inhibitors (HALT Phosphatase Inhibitor Cocktail 78428 and HALT Protease Inhibitor Cocktail 78430, Thermo Fisher Scientific Inc., Rockford, IL, USA). Swabs were kept on ice until transfer to a −20°C freezer. Samples were transferred weekly from −20°C freezers to an ultracold freezer (−80°C). Samples procured from Canadian facilities were shipped overnight with dry ice to Daemen College, Natural and Health Sciences Research Center, Center for Wound Healing Research, Amherst, NY, USA.

Swab samples were thawed to 4°C and proteins resuspended from the polyester tip by the addition of 100 µl PBS. Protease and phosphatase inhibitors were supplemented to a final 2× concentration and tubes were vortexed for 60 seconds. The swabs were inverted and liquid removed from the polyester tip via centrifugation for 2 minutes at 14 000 rpm (18 620 rcf) in a 4°C refrigerated centrifuge. The swabs were removed from the vial and cellular debris pelleted by repeating the centrifugation for 10 minutes. Supernatants were collected and pooled for each wound and collection date prior to protein quantification using the method of Bradford in a microplate format (BioRad, Catalog number 500-0006, Hercules, CA, USA) with Bovine IgG (BioRad, Catalog number 500-0005) as the protein standard. After protein quantification, samples were aliquoted into sterile 2 ml screw top conical microcentrifuge tubes and vials stored at −80°C. Protein was expressed as mg/ml (total protein concentration) and the wound protein yield from all the swabs was calculated using the concentration and total fluid volume.

Blood chemistry

Blood serum samples were tested for complete blood count, albumin (ALB), hemoglobin A1c (HbA1c), total iron binding capacity (TIBC), iron, percent (%) saturation, CRP, and erythrocyte sedimentation rate (ESR) by Quest Diagnostics, Pittsburgh, PA, USA (non-SCI) or LifeLabs Inc., Toronto, ON, Canada (SCI) subjects.

Antibody microarray

Multiplexed microsphere-based suspension microarrays were used to assay the concentration of 22 protein analytes in the wound fluid and the blood serum samples. Analyte concentrations for wound fluid samples were standardized using the wound fluid total protein and expressed as pg or ng analyte/mg total protein.

Statistical analysis

Demographic data were presented descriptively and continuous variables were expressed as mean ± standard deviation. Student's t-test was used to evaluate the differences in initial wound surface area and depth, subject age and wound fluid protein concentration, and the yield between SCI and non-SCI populations for the initial sample collected for each subject. Correlations between the observed and predicted levels were assessed using linear regression analysis (SPSS, version 21, IBM, Armonk, NY, USA). Demographic data comparisons with P < 0.01 were considered statistically significant.

Wound fluid samples, protein concentration, and 22 analyte concentrations in the serum and wound fluid were tested for differences between the SCI and non-SCI subjects using a two-tail Mann–Whitney test with correction for multiple testing (q-values) according to Benjamini–Hochberg. For SCI and non-SCI comparisons, proteins expressed below the assay sensitivity were assigned the minimum detectable concentration. Averages were calculated prior to analyses for individuals with multiple baseline samples. A comparison of subjects with SCI between the level of injury, AIS A and B, for whole blood tests, total protein in wound fluid, and 22 analyte concentrations in the serum and wound fluid was completed as described for SCI and non-SCI comparisons. Proteins with concentrations below the assay sensitivity were removed from the AIS analysis. The Wistar Institute Bioinformatics Facility (Philadelphia, PA, USA) completed the data analysis using R Foundation for Statistical Computing (v 2.11, Vienna, Austria) and Matlab^® (v 7.3, MathWorks Inc., Natick, MA, USA) software. Samples for each subject were combined and the averages compared and modeled using regression analyses and t-tests. For subjects with SCI, the serum analyte values and standardized wound fluid analyte values for the same subject and time point were paired. Spearman correlation was performed in R to quantify the linear relationship between the serum and wound fluid analyte concentrations. The median analyte concentration was determined by combining values obtained for all wound fluid samples and was compared to the median value for the same analyte in all serum samples.

Because this was a pilot study with a relatively low number of subjects in the SCI (n = 29) and non-SCI (n = 9) groups, statistical tests involving multiple testing differences with P ≤ 0.1 and q ≤ 0.5 (false discovery rate) were considered statistically significant.

Results

Demographics and wound characteristics

A total of 29 subjects with SCI and 9 subjects without SCI were enrolled in the study. Participant demographics and brief health history are presented in Table 1. Among subjects with SCI, there was a history of pressure ulcers with 13 subjects having previous ulcers and 10 subjects with recurrent ulcers. Previous ulcers were defined as wounds that existed prior, but in a different location than the monitored ulcer. Ulcers with a history of opening and closing repeatedly were defined as recurrent. Six subjects were particularly prone to pressure ulcers, reporting both prior and recurrent pressure ulcers. More than half of the SCI population had failed conservative wound closure treatments previously. Subjects with SCI were equally sampled among tetraplegia and paraplegia and the mean duration of the current wound was 15.6 ± 18 months. The median wound number was two per person. AIS classification of A, B, C, or D scores were collected for a subset of the subjects with SCI with AIS A predominating. Of the nine subjects without SCI, five had a history of prior pressure ulcers. The average age of the non-SCI group was 76 ± 11.5 years; significantly greater than the average age of the SCI group, 48 ± 12.6 years (P < 0.001, Table 1). Wound duration for non-SCI ulcers ranged from 7 weeks to 14 months. Pressure ulcers for both the SCI and non-SCI groups included Stages II, III, and IV⁶ with deeper Stages III and IV most prevalent for both groups. The ischial tuberosity was the most common place for SCI pressure ulcers and the sacrum was the most common among the non-SCI subjects.

Table 1 .

Patient demographics and brief health history

	SCI (n = 27)			Non-SCI (n = 9)
Age ± SD*	48.1 ± 12.6 years			75.6 ± 11.5 years
SCI duration ± SD (n = 27)	15 ± 11.3 years
Sex (n)	Female     Male			Female    Male
	11        18			7         2
SCI level	14 tetraplegia  15 paraplegia			–
ASIA Impairment Scale (n = 12)	AIS A = 9    AIS B = 3			–
Wound site	13 = Ischial tuberosity			6 = Sacrum
	7 = Sacrum			3 = Foot/Heel
	4 = Trochanter
	5 = Foot
Wound severity	Stage II Stage III Stage IV			Stage II Stage III Stage IV
mean ± SD	1    22    6			2    4    3
History of pressure ulcers (n)	Previous = 14			Previous = 5
	Recurrent = 10
	Both = 5
	None = 7
Wound surface area (cm2)	6.2 ± 10.1			5.0 ± 6.5
Wound depth (cm)	1.79 ± 1.4			1.13 ± 1.29
Wound fluid protein (mg/ml)*	8.65 ± 8.02			2.55 ± 1.60
Total protein yield (g)*	11.11 ± 12.48			2.81 ± 2.98

*Significantly different, P < 0.001.

Average initial wound surface area and depth did not differ between the SCI and non-SCI groups but the group age and wound fluid protein were significantly different (Table 1). The depth and wound protein were positively correlated with an increasing depth yielding significantly greater protein concentrations (as measured by mg/ml) or total protein yield (P < 0.001). Neither the population age nor the wound surface area affected the protein concentration or yield.

Blood parameters

Hematocrit, iron, red blood cell distribution width (RDW), ALB, and iron % saturation were significantly different (P ≤ 0.10, q = 0.28) and red blood cell count (RBC) marginally significant (P = 0.11, q = 0.28) between the subjects with and without SCI (Table 2). SCI subjects had higher values for all except RDW. Hemoglobin, hematocrit, and RBC were significantly different (P ≤ 0.10, q = 0.47) and the mean cell hemoglobin concentration was marginally significant (P = 0.11, q = 0.57) between the subjects with SCIs AIS A and B classifications (Table 3). Subjects with AIS A classification had higher values for all parameters.

Table 2 .

Mean and standard error (SE) for blood parameters of subjects with SCI and without SCI (non-SCI)

	Non-SCI (USA)					SCI (Canada)
	Normal	Mean	SE	Abnormal (n)		Normal	Mean	SE	Abnormal (n)
				High	Low				High	Low
Hemoglobin (HGB) (g/dl)	13.2–15.5	11.1	0.73	–	1	12.0–16.0	12.62	1.97	1	13
Hematocrit (HCT) (%)*	38.5–45.0	33.9	1.27	–	1	35–45	38.81	4.76	1	3
White Blood Cell Count (WBC) (×109/l)	3.8–10.8	8.4	2.7	1		4.0–11.0	7.31	1.97	2	–
RBC (×1012/l)	4.2–5.1	4	0.14	–	4	4–5.1	4.46	0.59	2	6
Mean cell volume (MCV) (fl)	80–100	84.5	5.09	–	1	80–100	87.14	5.02	–	1
Mean cell hemoglobin (MCH) (pg)	27.0–33.0	27.7	2.22	–	1	27.5–33.0	28.26	1.98	–	7
MCHC (g/dl)	32.0–36.0	32.8	0.99	–	1	30.5–36.0	32.03	2.49	–	4
RDW (%)*	11.0–15.0	17.8	3.15	3	–	11.5–14.5	15.14	1.93	15	–
Platelet count (×109/l)	150–400	311.8	91.42	1	–	150–400	300.44	73.78	3	–
HbA1C (%)	<6.0	6.1	1.22	1	–	0.04–0.06	5.39	0.56	2	–
ALB (g/dl)*	3.6–5.1	3.4	0.47	–	3	3.5–5.0	3.86	0.39	–	6
ESR, Westergren (mm/hour)	0–15	33.3	19.62	3	–	up to 20	34.42	31.29	13	–
CRP (mg/dl)	<0.8	6.8	10.15	3	–	<0.7	2.53	2.15	21
Iron total (mcg/dl)*	40–170	28.3	10.75	–	4	61.4–150.7	45.31	17	–	19
TIBC (mcg/dl)	250–450	278.3	47.95	–	1	251.1–401.8	291.52	46.56	–	5
Iron (%) saturation (%)*	15–50	10.5	3.7	–	4	20–50	15.52	5.53	–	17

*Significantly different with Mann–Whitney rank sum P ≤ 0.10 and false discovery rate (FDR) q-value = 0.28.

Non-SCI (n = 4), SCI (n = 27) except ESR, CRP (n = 26), total iron, TIBC, Iron %, RDW, platelet count (n = 25).

MCHC, Mean cell hemoglobin concentration.

Table 3 .

Blood test parameters differing between subjects with SCI AIS A and B

				Subjects with SCI
Parameter	P-value*	q-value**	Directionality	AIS A	AIS B
Hemoglobin (HGB)	0.05	0.47	AIS A > B	8	3
Hematocrit (HCT)	0.06	0.47	AIS A > B	8	3
RBC	0.06	0.47	AIS A > B	8	3
Mean cell hemoglobin concentration (MCHC)	0.11	0.57	AIS A > B	8	3

*Mann–Whitney rank sum P-value.

**False discovery rate (FDR) q-value.

Blood serum samples of SCI subjects

Fibroblast growth factor 1 (acidic) (FGF1) was found in only 17% of blood serum samples tested. Interleukin-2 (IL-2) was detected in 57% of the samples. IL-2RA was measurable in 83% of the serum samples, considerably more than the wound fluid. AIS B subjects had higher concentrations of MMP-9, TGF-b1, and TIMP-1 while subjects with an AIS A classification had higher concentrations of MMP-2 (P = 0.06, q = 0.47) (Table 4).

Table 4 .

Serum protein parameters differing between subjects with SCI AIS A and B

				Subjects with SCI
Analyte	P-value*	q-value**	Directionality	AIS A	AIS B
MMP-9	0.06	0.47	AIS B > A	7	2
TGF-b1	0.06	0.47	AIS B > A	7	2
TIMP-1	0.06	0.47	AIS B > A	7	2
MMP-2	0.06	0.47	AIS A > B	7	2

*Mann–Whitney rank sum P-value.

**False discovery rate (FDR) q-value.

Wound fluid in the subjects with SCI versus those without SCI

The median wound fluid concentrations were significantly different between the subjects with and without SCI for total protein and MMP-9, S100A12, S100A8, S100A9, fibroblast growth factor 2 (FGF2), interleukin-1 beta (IL-1b), TIMP-1 and TIMP-2, and TGF-b1 (Table 5). Subjects without SCI had higher values for all significantly different analytes measured in the wound fluid except FGF2, TGF-b1, and wound fluid total protein. Analytes with marginally significant differences include TIMP-4 (P = 0.14, q = 0.29) and TNF-a (P = 0.2, q = 0.39). Collectively, IL-2 was detected in only 45% of the 65 samples tested and its receptor, IL-2RA, was measurable in 23% of the samples (Table 6).

Table 5 .

Pressure ulcer wound fluid protein comparison between subjects with SCI and without SCI (non-SCI)

				Subjects
Analyte	P-value*	q-value**	Directionality	SCI	Non-SCI
FGF2	0.004	0.03	SCI > non-SCI	29	9
Total protein (µg/µl)	0.02	0.07	SCI > non-SCI	29	9
TGF-b1	0.08	0.2	SCI > non-SCI	29	9
VEGF	0.45	0.61	SCI > non-SCI	29	9
MMP-2	0.45	0.61	SCI > non-SCI	29	9
IL-6	0.54	0.64	SCI > non-SCI	29	9
MMP-1	0.86	0.9	SCI > non-SCI	29	9
MIG	1	1	SCI > non-SCI	20	9
MMP-9	0.0007	0.01	Non-SCI > SCI	29	9
S100A12	0.001	0.01	Non-SCI > SCI	20	9
TIMP-2	0.007	0.04	Non-SCI > SCI	29	9
S100A8	0.02	0.07	Non-SCI > SCI	20	9
S100A9	0.03	0.08	Non-SCI > SCI	20	9
IL-1b	0.08	0.2	Non-SCI > SCI	29	9
TIMP-1	0.1	0.23	Non-SCI > SCI	29	9
TIMP-4	0.14	0.29	Non-SCI > SCI	29	9
TNF-a	0.2	0.39	Non-SCI > SCI	29	9
IL-2	0.32	0.57	Non-SCI > SCI	29	9
EGF	0.39	0.61	Non-SCI > SCI	29	9
TIMP-3	0.41	0.61	Non-SCI > SCI	29	9
FGF1	0.54	0.64	Non-SCI > SCI	29	9
NGF	0.56	0.64	Non-SCI > SCI	29	9
IL-2RA	0.63	0.69	Non-SCI > SCI	29	9

*Mann–Whitney rank sum P-value.

**False discovery rate (FDR) q-value.

Table 6 .

Microarray analytes, corresponding detection limit, and sample prevalence

			Prevalence (%)
			Serum	Wound Fluid
Analyte		Detection limit (pg/ml)	SCI	SCI	Non-SCI
EGF	Epidermal growth factor	5	87	98.1	92.3
FGF2	Fibroblast growth factor 2 (basic)	3.2	87	98.1	100
IL-1b	Interleukin-1 beta	0.4	73.9	100	100
IL-2	Interleukin-2	0.3	56.5	48.1	30.8
IL-6	Interleukin-6	0.6	100	100	100
IL-2RA	Interleukin-2 receptor alpha	2	82.6	28.8	0
TNF-a	Tumor necrosis factor alpha	1	100	100	100
VEGF	Vascular endothelial growth factor	5.6	100	100	100
MMP-1	Matrix metalloproteinase-1	4.4	100	100	100
MMP-2	Matrix metalloproteinase-2	6.8	100	100	100
MMP-9	Matrix metalloproteinase-9	7.5	100	100	100
NGF	Nerve growth factor	2.5	100	78.8	46.2
FGF1	Fibroblast growth factor 1(acidic)	2	17.4	80.8	84.6
TGF-b1	Transforming growth factor beta 1	9.8	100	100	76.9
TIMP-1	Tissue inhibitor of metalloproteinase-1	1.6	100	100	100
TIMP-2	Tissue inhibitor of metalloproteinase-2	14.7	100	100	100
TIMP-3	Tissue inhibitor of metalloproteinase-3	86	100	100	100
TIMP-4	Tissue inhibitor of metalloproteinase-4	1.3	100	100	100
CXCL9/MIG	Chemokine CXC ligand 9	18.2	nt*	100	100
S100A8	S100 calcium binding protein A8	298	nt	100	100
S100A9	S100 calcium binding protein A9	213	nt	100	100
S100A12	RAGE-binding protein	28.9	nt	100	100

*Not tested.

Wound fluid and serum

Pairwise comparison of serum analyte concentration and standardized wound fluid concentration for samples collected on the same day for 16 subjects resulted in only one analyte, interleukin-6 (IL-6), with a marginal positive correlation (r = 0.4) and two analytes, TGF-b1, and TIMP-3, with a marginal negative correlation (r = −0.4, −0.3, respectively). When all the subjects’ median serum and wound fluid protein were compared, median baseline serum protein levels reflected median baseline wound fluid protein levels (when one rose the other did also) across a range in scale from 0 to 10⁵ in abundance (r = 0.7) for 18 microarray analytes.

Wound fluid AIS score

TIMP-2 and MMP-9 were identified as different (P ≤ 0.1, q = 0.47) between SCI AIS A and B wound fluid (Table 7). IL-1b, VEGF, and TIMP-4 were marginally significant between the subjects with SCI AIS A and B classifications (P = 0.12, q = 0.57). Subjects with an AIS B classification had higher values for all analytes.

Table 7 .

Wound fluid protein parameters differing between subjects with SCI AIS A and B

Analyte	P-value*	q-value**	Directionality	AIS A	AIS B
TIMP-2	0.03	0.47	AIS B > A	7	3
MMP-9	0.07	0.47	AIS B > A	7	3
IL-1b	0.12	0.57	AIS B > A	7	3
VEGF	0.12	0.57	AIS B > A	7	3
TIMP-4	0.12	0.57	AIS B > A	7	3

*Mann–Whitney rank sum P-value.

**False discovery rate (FDR) q-value.

Discussion

This report represents a pilot investigation and first description of the biochemical profile of wound fluid from pressure ulcers occurring in people with SCI, and, the first comparison of wound fluid biochemistry between pressure ulcers from patients with and without SCI. Wound fluid was collected from subjects with SCI who had long-standing, recurrent, and frequent Stages II–IV pressure ulcers. Protein microarray analysis of wound fluid samples taken from pressure ulcers revealed differences between the people with and without SCI. The wound fluid protein concentrations of MMP-9, S100A8, S100A9, S100A12, TIMP-1, TIMP-2, and IL-1b were significantly greater for the subjects without SCI, while the wound fluid concentration of FGF-2 and TGF-b1 were significantly greater for the subjects with SCI. Weak or no correlation between the circulating levels of these proteins and the standardized local concentration of the same proteins in the wound fluid was found. Analysis examining the differences between subjects with AIS A and AIS B suggest that the degree of motor and sensory impairment, particularly of the sacral segments, may affect the analyte concentration in pressure ulcer wound fluid.

Blood chemistry

Both subjects with and without SCI had anemia, iron deficiency, and low circulating levels of serum ALB. Anemia and low serum ALB levels are prevalent in individuals with advanced age.^36,40,41 Hematocrit, iron, ALB, and % saturation were all significantly greater and RDW was significantly lower in the SCI group versus the non-SCI group, similar to the findings reported by Scivoletto et al.³⁹

Vaziri et al.³⁸ evaluated blood serum from SCI individuals with and without pressure ulcers and able-bodied individuals without pressure ulcers and found no significant differences for ALB and hematocrit. Hayes et al.⁴⁰ compared the subjects with SCI and healthy subjects who were able-bodied and found no significant differences in total serum protein, ALB, or white or red blood cell count values. Davies et al.³⁷ compared people with SCI and able-bodied, age-matched controls and found significantly reduced serum protein and ALB levels in subjects with tetraplegia versus controls, but only serum ALB was reduced in people with paraplegia versus controls. Hemoglobin values did not significantly differ versus controls.³⁷

Anemia and low circulating levels of iron and ALB have been well-documented in people with SCI and pressure ulcers,³⁹ but comparisons between the groups of people with pressure ulcers with and without SCI have not been previously documented. In this study, all subjects without SCI had below normal levels of RBC and most also had low ALB values. The differences observed in the circulating level red blood cells and serum proteins could be explained, in part, by advanced age and institutionalization of the non-SCI group.

In our study, the subjects without SCI were biased towards higher blood serum concentrations of ESR, RDW, and CRP when compared to the subjects with SCI. ESR and CRP are reported to increase with age.^42,43 CRP is an acute phase protein that participates in response to inflammation, infection, and tissue damage. Increased values observed for subjects without SCI in this study may be correlated with comorbidities common to the elderly.⁴⁴ Hirsch et al.⁴⁵ reported hemoglobin and hematocrit values below normal in 70% of post-acute subjects with SCI. In our sample, hemoglobin, hematocrit, and RBC values were all significantly greater in subjects with AIS A compared to AIS B. These results are in contrast to reports by Davies et al.³⁷ where no differences were observed by AIS classification. The authors recognize that the number of subjects in this sample is low; however, comparison of data across the impairment groups allows for hypothesis generation for future studies.

Serum and ASIA impairment scale (AIS)

Few comparisons and even fewer differences have been documented for serological markers between AIS classes.^37,40 Among them, IL-6 was found to be significantly higher in subjects with paraplegia, intermediate in subjects with tetraplegia, and lowest in subjects without SCI.³⁷ Although our subject numbers are low (n = 9), subjects with an AIS B classification had higher concentrations of MMP-9, TGF-b1, and TIMP-1 while subjects with an AIS A classification had higher concentrations of MMP-2. Further research is required to confirm whether serum protein levels are related to neurological level, degree, or type of impairment.

Wound fluid

Wound fluid protein concentration was significantly higher for the SCI group compared to the non-SCI group in the current study. As observed previously, wound fluid protein concentration was affected by the wound depth but not surface area.^34,46 Neither depth nor surface area was significantly different between these two populations indicating additional factors contributing to the differences in total protein observed. The elderly non-SCI population was recruited from skilled nursing facilities and all had comorbidities typical of the population, including low serum ALB. The low serum ALB is likely reflected in the low protein present in the wound bed as well.⁴⁷ Serum ALB levels naturally decrease with increasing age and nutritional challenges for aged populations in nursing facilities may lead to reduced serum protein production and slow wound healing.^48–50 Healing wound fluid contains higher levels of total protein and ALB.^30,47,51 Lower protein in the wound fluid of subjects without SCI can be explained in part by the increased age compared to subjects with SCI and chronic condition of the ulcers sampled.

TIMPs/MMPs

All the subjects included in this study had chronic, non-healing pressure ulcers with some subjects having had their ulcers for several years. Researchers comparing chronic and acute wounds have reported elevated concentrations of MMP-9 in chronic wounds.⁵² The median MMP-9 measured in the subjects without SCI in our study (67.9 ng/100 µg) was greater than the acute wound values previously reported and within the range of chronic pressure ulcer values reported by Yager et al.⁵² (186.3 ± 59.1 ng/100 µg). Median MMP levels measured in this SCI population (35.1 ng/100 µg) were low compared to other populations with chronic wounds.⁵² Evaluating concentrations of MMPs between studies is complicated by differing sample collection methods, analysis techniques, and reporting and normalization methods. Importantly, we did not evaluate MMP levels in acute wounds of SCI individuals whose wounds were actively healing. Rather, all the subjects included in the present report had had their pressure ulcers for at least 4 weeks.

TIMPs regulate MMPs and the ratio of MMP 9/TIMP-1 has been reported to decrease as chronic pressure ulcers heal.²⁶ Ladwig et al.²⁶ reported that the greatest ratios seen in poor healers were 120:1 and in good healers were 30:1. In our study, ratios of MMP-9/TIMP-1 for SCI were 42:1 and non-SCI were 52:1. Therefore, it seems that while MMP values were lower in subjects with SCI compared to levels reported previously, the MMP/TIMP ratios were similar for both populations. This suggests that overall MMP and TIMP abundance was reduced in the SCI population.

In this study, MMP-9 levels in the wound fluid were significantly lower in subjects with SCI. Elevated MMPs in chronic wounds have been reported,²⁵ but comparisons of MMP concentrations from pressure ulcer wound fluid of people with and without SCI have not been previously reported. Xia et al.⁵³ reported that tissue from aged donors had a higher basal level of MMP expression compared to younger donors. Therefore, it is possible that the differences we observed between the subjects with SCI and the elderly subjects without SCI in the present study are age-related.

S100s

Previously, we reported differences in S100 protein abundance depending on the location within the wound bed.³⁴ Advanced glycosylation end product receptor (AGER) isoform sRAGE, S100A6, and S100A7 were present at higher levels in the periphery of chronic pressure ulcers versus the interior of the wounds.³⁴ Calgranulin A (S100A8), calgranulin B (S100A9), and calgranulin C (S100A12 or ENRAGE) have anti-inflammatory and antimicrobial functions and S100A8 and A9 have been shown in inflammation sites in a murine model to increase neutrophil migration.⁵⁴ The S100 proteins are dynamically regulated during tissue regeneration and scarring.^55,56 S100A8 and S100A9 have been reported in numerous inflammatory diseases.^57–59 Both increase after epidermal injury^60,61 and S100A9 has been reported to be elevated in non-healing wounds.^62,63 Our study showed significantly higher levels of S100A12, S100A8, and S100A9 in non-SCI compared to SCI pressure ulcer wound fluid suggesting substantial differences in inflammatory and antimicrobial function between these populations.

The role of advanced glycation end products (AGEs) on MMP-9 production was evaluated by Zhu et al.⁶⁴ in a keratinocyte cell culture model. The AGE receptor, AGER, was shown to have a role in MMP-9 expression. It is accepted that lymphocytes have an important role for wound healing in humans.²¹ Lymphocyte function is reported to be reduced in people with SCI.^36,65,66 Lymphocytes stimulate secretion of MMP-9⁶⁷ and the reduced function of these cells in people with SCI may contribute to the findings in this study.

Wound fluid cytokines, growth factors, and receptors

Because this study is the first to compare pressure ulcer wound fluid between the subjects with and without SCI, comparison of wound fluid results to literature findings was extended to include blood serum. IL-1b is a pro-inflammatory interleukin, which has been reported to be elevated in chronic wounds.⁶⁸ In this study, subjects without SCI had lower IL-1b concentrations in pressure ulcer wound fluid when compared to subjects with SCI. Davies et al.³⁷ found elevated levels of some pro-inflammatory cytokines in the blood serum of subjects with SCI and further elevated values in subjects with SCI and urinary tract infection (UTI), neuropathic pain, or pressure ulcers but no detectable levels of IL-1b.

Segal et al.²⁰ found greater amounts of circulating plasma levels of IL-6 and IL-2RA, in patients with SCI and pressure ulcers versus able-bodied subjects without pressure ulcers. Individuals with SCI and slow healing pressure ulcers had the highest concentrations. The mean IL-6 plasma concentration was higher in subjects with SCI compared to able-bodied controls, but not significantly. Pressure ulcer wound fluid concentrations of IL-6 and IL2RA were not different between the SCI and non-SCI populations in our study.

For the current study, FGF2 was measured in wound fluid at higher concentrations in subjects with SCI and FGF1 was not significantly different between subjects with and without SCI. Clarke et al.⁶⁹ studied FGF in subjects during bed rest with and without exercise and found bed rest alone caused a significant decreased levels of FGF1, while bed rest plus exercise resulted in greater levels of FGF1. In the current study, subjects without SCI were less active than subjects with SCI, in part due to age but also living situation. Corresponding FGF1 levels in the wound fluid of subjects without SCI correlated with the physical activity level and were significantly lower than found in subjects with SCI. TGF-b1 levels were greater in the subjects with SCI. Although TGF-b1 has a significant role in wound healing, elevated levels can lead to decreased healing.⁷⁰

Wound fluid AIS

This is the first report of wound fluid and AIS category analysis. TIMP-2, MMP-9, IL-1b, VEGF, and TIMP-4 were significantly greater in subjects with an AIS B classification. With the limitation of small sample size, the significance of these differences needs to be confirmed and further investigations are warranted.

Correlation between the serum and wound fluid samples

Two of 22 protein analytes had a weak correlation between blood serum and standardized pressure ulcer wound fluid concentration for samples taken from the same individual at the same time. Yager et al.⁵² also observed no correlation between the serum and pressure ulcer wound fluid for MMP-9 and MMP-2. These results imply that we need to exercise caution when assuming that circulating levels of inflammatory cytokines or other bioactive proteins will reach and impact the wound environment. It also calls into question the hypothesis that elevated circulating levels of pro-inflammatory mediators contribute to delayed healing in the SCI population.

Study limitations

In order to evaluate differences in protein abundance between people with pressure ulcers and with or without SCI, it was necessary to find a population of people without SCI and with pressure ulcers. In the able-bodied population, this is primarily the elderly. Our subjects were not age-matched due to this limitation. Both of these groups are known to be poor healers. The site of wounds varied and was not matched between populations, which may contribute to the observed biochemical differences. This is a pilot study with low subject numbers and includes nearly three times as many subjects with SCI as those without. The comparison of these groups is relevant in that the vast majority of treatments for pressure ulcers are developed in the non-SCI population, unlike prevention options that are often evaluated in people with SCI.

Conclusions

This pilot study has demonstrated for the first time that the biochemical profile of the wound environment is markedly different in pressure ulcers of people with SCI compared to a group of elderly individuals without SCI. We found significant differences between individuals with and without SCI in all categories of proteins known to be involved in wound healing (growth factors, MMPs, TIMPs, and inflammatory mediators). Weak or no association between the serum and wound fluid analyte concentration was found, suggesting blood samples may not be representative of the local wound environment. Further research is required to confirm if wound fluid protein levels are related to neurological level, degree, or type of impairment, metabolic status or other health complications, or explained by age and living situation or other known/unknown factors that differentiate between the two groups. The altered immune response seen in people with SCI may be responsible for many of the results found in this study. It is important to consider these differences as we choose and develop therapies for treatment of chronic wounds in the SCI population.

Acknowledgments

We thank Catholic Health System Partners in Rehab and ElderWood Health Care for allowing us to enroll subjects from their facilities. We thank Sherry Green, RN, Mohammed Ghotbi, MD, Kirsten Allen, RN, Diane Leber, RN, Christa Lodwig, and Jennifer Schaffstall, DPT for their data and sample collection efforts. Statistical analysis and expertise was provided by Andrew Koss, Wistar Institute (Philadelphia, PA, USA). Dr Craven acknowledges the support of Toronto Rehabilitation Institute, which receives funding under the provincial rehabilitation research program from the Ministry of Health and Long-Term Care in Ontario. Rajna Ogrin acknowledges the support of Ontario Neurotrauma Foundation and the Ministry of Research and Innovation, Ontario in the form of Post Doctoral Fellowship funding.

Disclaimer statements

Contributors LE, CC, RO and PH designed the study. LE, RO and JW collected and interpreted data. LE and JW created the original draft and LE, JW, RO, CC and PH revised the article. All authors approved the manuscript version to be published.

Funding This work was supported by a grant received from Ontario Neurotrauma Foundation (ONF) - 2007-SCI-BIOELEC-487.

Conflicts of interest None.

Ethics approval Participants were enrolled at sites in Buffalo, New York, Toronto, Ontario, and London, Ontario. Institutional Review Board (IRB) approval for the study and consent form was granted by the Catholic Health System and the Human Subjects Research Review Committee at Daemen College for Non-SCI subjects. The Health Sciences Research Ethics Review Board, University of Western Ontario and the Toronto Rehab Institute Research Ethics Review Board granted approval for enrollment of the SCI subjects. Adult individuals with Stage II, III and IV pressure ulcers with and without SCI provided written consent for study participation.