Page 1
A Molecular and Culture-Based Assessment of the 1
Microbial Diversity of Diabetic Chronic Foot Wounds 2
and Contralateral Skin Sites 3 4
Angela Oatesa, Frank L. Bowlingb, Andrew JM Boultonb, 5 Andrew J McBaina* 6
7 School of Pharmaceutical Sciencesa and Department of Medicine Manchester Royal Infirmaryb, The 8
University of Manchester, Manchester, U.K, 9 10
11
12
13
Key words: Chronic wounds, bacteria, DGGE and culture. 14
15
16
17
18
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 *Corresponding author: Andrew McBain, School of Pharmacy, The University of Manchester, 38 Oxford Road, Manchester M13 9PT, UK. Tel: 00 44 161 275 2361; 39 Fax: 00 44 161 275 2396; Email: [email protected] 40 41
Copyright © 2012, American Society for Microbiology. All Rights Reserved.J. Clin. Microbiol. doi:10.1128/JCM.06599-11 JCM Accepts, published online ahead of print on 2 May 2012
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 2
2
ABSTRACT 42
Wound debridement samples and contralateral (healthy) skin swabs, acquired from 26 patients 43 attending a specialist foot clinic were analysed by differential isolation and eubacterial-specific 44 PCR-DGGE, in conjunction with DNA sequencing. Thirteen out of twenty-six wounds harboured 45 pathogens according to culture analyses, of which Staphylococcus aureus was the most common 46 (13/13). Candida (1/13), pseudomonas (1/13) and streptococci (7/13) were less prevalent. 47 Contralateral skin was associated with comparatively low densities of bacteria and overt pathogens 48 were not detected. According to DGGE analyses, all wounds were associated with significantly 49 greater eubacterial diversity than contralateral skin (p<0.05), although no significant difference in 50 total eubacterial diversity was detected between wounds from which known pathogens had been 51 isolated and those that were putatively uninfected. DGGE amplicons with homology to 52 Staphylococcus sp. (8/13) and S. aureus (2/13) were detected in putatively infected wound samples, 53 whilst Staphylococcus sp. amplicons were detected in 11/13 non-infected wounds; S. aureus was 54 not detected in these samples. Whilst a majority of skin-derived DGGE consortial fingerprints could 55 be differentiated from wound profiles through principal component analysis (PCA), a large minority 56 could not. Furthermore, wounds from which pathogens had been isolated could not be distinguished 57 from putatively uninfected wounds on this basis. In conclusion, whilst chronic wounds generally 58 harboured greater eubacterial diversity than healthy skin, the isolation of known pathogens was not 59 associated with qualitatively distinct consortial profiles or otherwise altered diversity. Data 60 generated support the utility of both culture and DGGE for the microbial characterization of chronic 61 wounds. 62 63 64 65
66
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 3
3
INTRODUCTION 67
Chronic wounds occur in approximately 2% of the population in developed countries (20) where 68
they cause considerable morbidity and mortality (23, 24, 34). In 2008, over two hundred thousand 69
patients in the UK were affected, associated with an economic burden of c. £3 Billion (34). 70
71
Common forms of chronic wound include pressure sores and diabetic/venous ulcers. Risk factors for 72
the development of diabetic ulcers include conditions associated with neuropathy and/or venous 73
insufficiency, which restrict oxygen supply and impair the transport and integration of leukocytes 74
and macrophages to tissues, leading to ischemic necrosis and subsequent ulceration (2, 12, 43). This 75
creates a portal of entry for bacteria and thus, increases susceptibility to infections. Infection can 76
then lead to further tissue damage and impaired healing by exacerbating the inflammatory state (15, 77
26, 37). 78
79
Clinical laboratory investigations of chronic wound infections commonly rely upon bacterial 80
isolation by culture, which most efficiently detects numerically dominant organisms amenable to 81
growth on laboratory media. Whilst this is a useful and well-established approach for the detection 82
of many common pathogenic bacteria associated with wound infections, it may underestimate 83
microbial diversity (45). Thus, various culture-independent methods have been assessed in a limited 84
number of studies as potential adjuncts or replacements of culture for the microbial characterization 85
of chronic wounds (11, 18, 23). Culture-independent investigations of the bacterial diversity 86
utilizing pyrosequencing (11), PCR-denaturing gradient gel electrophoresis (DGGE) (8), other DNA 87
fingerprinting techniques (39) and qPCR (31) have generally identified a greater range of bacteria 88
than traditional culture techniques, and taxa not previously detected in wounds have been reported. 89
For example, Hill et al. (23) used 16S rDNA clone sequence analysis and culture to assess the 90
microbial composition of a chronic venous leg ulcer. Acinetobacter sp. was detected by culture in 91
both swabs and tissue samples; swab samples yielded Proteus sp. and Candida tropicalis, whereas 92
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 4
4
Staphylococcus epidermidis was only isolated from tissue samples. Importantly however, molecular 93
analysis of the same samples identified clones that were closely related to the cultured organisms, 94
together with species that has not been isolated from the samples (Morganella morganii, 95
Bacteroides urelyticus, Enterococcus faecalis and Peptostreptococcus octavius). A study conducted 96
by Davies et al., (8) which assessed the microbiota of healing and non-healing chronic venous leg 97
ulcers also reported greater eubacterial diversity according to PCR-DGGE than culture. Of the 98
sequences obtained in the Davies study, 40% were organisms which had not been isolated from the 99
same samples using culture (8). These are among a limited number of reports relating to the ability 100
of molecular techniques to identify a distinct range of organisms in wounds and also illustrate the 101
importance of sampling techniques and sample site on the outcome of analyses (23). 102
103
Whilst it is clear that culture-independent methods may provide deeper characterization of microbial 104
diversity, the role that taxa thus identified play in infection remains poorly understood. This 105
contrasts with isolation methods where the pathogenicity of prominent culturable organisms such as 106
Staphylococcus aureus and Pseudomonas aeruginosa has been well documented (3, 10, 19, 29). 107
The current study used clinical diagnostic isolation techniques in combination with eubacterial-108
specific PCR-DGGE to compare bacterial consortia associated with chronic wound debridement 109
samples to healthy skin (17, 21), to assess the utility of PCR-DGGE to culture and to determine 110
whether samples from which pathogens could be isolated were otherwise compositionally distinct. 111
112
METHODS 113
Chemicals and media. Unless otherwise stated, chemicals used were obtained from Sigma 114
(Poole, Dorset, U.K.). Dehydrated bacteriological media was obtained from Oxoid (Basingstoke, 115
Hampshire, U.K.) and prepared according to instructions supplied by the manufacturer. 116
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 5
5
Ethical approval. This study was reviewed by the North Manchester Research Ethics 117
Committee and Central Manchester University Hospital Research and Development department. 118
Reference number: 09/H1006/41, protocol number 1.0. 119
Collection of chronic wound tissue and contralateral skin swabs. A total of 26 wound 120
tissue debridement samples from chronic diabetic foot wounds (defined as distal to the medial and 121
lateral malleoli, with a known duration greater than four weeks) and 26 contralateral skin swabs, 122
were obtained from patients with diabetic chronic foot wounds attending a specialist foot clinic, 123
Manchester, U.K between 2/2/2010 and 2/2/2011 as follows: Wound tissue samples were taken 124
from the wound bed and surrounding tissue using a sterile scalpel by the attending clinician and 125
placed in sterile 0.85% (w/v; 5 ml) saline for transportation. Skin swabs of an area of intact 126
contralateral skin measuring 40 cm2 were also collected using Dual Amies transport swabs (Duo 127
Transwab, MWE, Wiltshire, U.K). Swabs were moistened with sterile saline and then used to 128
thoroughly scrub skin sites within the contralateral area. All samples were transported to the 129
laboratory and processed within 3 h of collection, as detailed in the following section. 130
Semi-quantitative culture and differential bacteriological identification. Tissue samples 131
were dissected with a sterile scalpel in the laboratory, weighed and homogenised using a sterile 132
tissue pulpier (VWR, Leicestershire, U.K) in 3 ml of sterile saline. Dual Amies swabs were 133
aseptically separated with sterile scissors, with one swab archived at -80°C for bacterial DNA 134
extraction in 1.5 ml microcentrifuge tubes. The remaining skin swabs and samples of homogenized 135
tissue were streaked for isolation onto four quadrants of Health Protection Agency (HPA)-136
recommended agars and atmospheres as shown in Table 1, to isolate clinically-relevant organisms, 137
according to standardized methods (25). Residual samples of homogenized tissue were archived at -138
80°C for bacterial DNA extraction. Bacterial species isolated from the four quadrants were reported 139
as scant (< 10 colonies), light (first quadrant), moderate (second quadrant), or heavy growth (third-140
fourth quadrant) according to methods outlined and validated by Angel et al., (1) and Healy and 141
Freedman (22). Bacterial identification of was based upon colony morphology, Gram staining, 142
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 6
6
catalase reaction, latex coagulase reaction tests, Lancefield group reaction to identify β haemolytic 143
streptococci (Prolex Streptococcal grouping latex kits, Pro-lab diagnostic, Cheshire, U.K.) and 144
subculture onto brilliance UTI media (25, 40) 145
Chronic wound tissue and skin swab bacterial community profiling using PCR-DGGE. 146
DNA was extracted from archived macerated tissue samples and swab samples using a DNeasy 147
blood and tissue kit (Qiagen Ltd., West Sussex, U.K.) in accordance with the manufacturers’ 148
instructions. The V2-V3 region of the 16S ribosomal DNA was amplified using the eubacterium-149
specific primers HDA1 (with additional GC clamp) (5’-CGC CCG GGG CGC GCC CCG GGC 150
GGG GCG GGG GCA CGG GGG GAC TCC TAC GGG AGG CAG CAG T-3’) and HDA2 (5'-151
GTA TTA CCG CGG CTG CTG GCA C-3') (44). The reaction mix was as follows: Red Taq DNA 152
polymerase ready mix (25 µl), HDA primers (2 µl of each (5 µM)), nanopure water (16 µl), and 153
extracted DNA template. The reactions were performed in 0.2 ml DNA free PCR tubes with a T-154
Gradient DNA thermal cycler (Biometra, Germany). The thermal amplification program was as 155
follows: 94°C (4 min), followed by 30 thermal cycles of 94°C (30 sec), 56°C (30 sec), and 68°C (60 156
sec). The final cycle incorporated a 7 minute chain elongation step (68°C). Positive and negative 157
controls (5 µl of microbial DNA extracted from saliva and nanopure water respectively) were run 158
concurrently with each reaction run. 159
Polyacrylamide electrophoresis was done using 30% and 60% denaturing concentrations 160
using the DCode Universal Mutation Detection System (Bio-Rad, Hemel Hempstead, UK) 161
according to manufacturers’ recommendations for perpendicular DGGE. Parallel denaturing 162
gradient gels of 10% acrylamide:bisacrylamide (37:1:5) (Sigma, Poole, Dorset, U.K.) were cast 163
with a linear gradient of urea and formamide ranging from 60% at the base to 30% at the top (100% 164
denaturants corresponds to 7M urea and 40% formamide). The gels were left to equilibrate at room 165
temperature overnight in the tank containing 7L of 1 x Tris-acetate buffer solution. Gel 166
Electrophoresis was carried out at 140 volts at a constant 60°C for 5.5 h (16, 30). All gels were 167
stained using SYBR® Gold stain (Molecular Probes, Leiden, The Netherlands) for 30 min with 168
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 7
7
occasional agitation after which they were transferred to a UV transilluminator (UVP, California, 169
USA), visualised under UV light at 312 nm, and photographed using a Canon D60 digital single 170
lens reflex (DSLR) camera (Canon, Surrey, UK). 171
Construction of dendrograms. Gel images were processed and aligned (with the aid of the 172
positive controls of saliva as internal controls used on each gel) using Adobe Photoshop Elements 173
version 7 and analysed using the BioNumerics Fingerprint package (Applied Maths, Belgium). 174
Bands were detected automatically and then checked manually. Dendrograms were constructed 175
using cluster comparison; employing unweighted pair group method with arithmetic mean 176
(UPGMA) algorithm (28). 177
178
Diversity indices were determined for each individual samples and also for each specimen group 179
(i.e. the presence or absence of cultured wound pathogens) and using the Shannon-Weaver index of 180
diversity (H’) using the following equation where s is the number of species and Pi is the proportion 181
of species in the sample i (13, 16). 182
183
184
185
The resultant indices were compared between wound and skin samples, and presence or absence of 186
cultured wound pathogens cohorts by the Mann-Whitney U test performed with SPSS version 16 187
(SPSS, Chicago, IL). 188
DGGE band excision and re-amplification for sequence identification. Replicated 189
DGGE bands (visualised on a UV transilluminator) i.e. those which were present across several 190
samples and unique bands were excised using a sterile scalpel and placed in nuclease-free tubes with 191
20 µl nanopure water. The bands were then stored at 4°C for 24 h before archiving at -80°C. Before 192
sequence analysis, the tubes were vortexed for 30s, then centrifuged (MSE Microcentaur; Sanyo, 193
Loughborough, U.K.) for 10min at 14,462 x g. Extracts (5µl) were then be used as a template for 194
PCR using the protocol outlined in the bacterial community profiling section. PCR products derived 195
s H’ = - ∑ Pi ln (Pi) i = 1
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 8
8
from excised DGGE bands were purified using QIAquickTM PCR purification kit (Qiagen Ltd., 196
West Sussex, U.K.) in accordance with the manufacturers’ instructions. PCR products were 197
sequenced using the non GC clamp (reverse) HDA primer. The sequencing reaction was as follows: 198
94°C (4 min) followed by 25 cycles of 96°C (30 sec), 50°C (15 sec), and 60°C (4 min). Once chain 199
termination was complete, sequencing was carried out in a Perkin-Elmer ABI 377 sequencer. DNA 200
sequences were compiled using CHROMAS-LITE (Technelysium Pty Ltd, Australia). Sequence 201
matching was undertaken using the bioinformatics program: Basic Local Alignment Search Tool 202
(http://www.ncbi.nlm.nih.gov/blast) to mine the prokaryotic database for matching sequences. 203
Principal component analysis. To produce graphical representations in which clusters can 204
be differentiated, principal component analysis was used. Similarity matrix data, derived from 205
UPGMA algorithms of DNA fingerprints of chronic wound tissue and intact skin swabs were 206
utilised to determine principal component data. Briefly, similarity matrix data of correlated variables 207
was reduced using factor analysis (SPSS, SPSS Inc., Chicago, Illinois) in which variances between 208
groups i.e. the different band position for each sample when compared to another, are maximised to 209
produce three overall uncorrelated variables (principal components). The first principal component 210
accounts for the greatest degree of variance in the overall group with each succeeding components 211
representing the remaining variances. The resulting principal component data was plotted on a three 212
axes scatter plot. 213
214
RESULTS 215
Semi-quantitative bacterial counts and identification of bacteria from wound and skin 216
samples. Data in Tables 2 and 3 show semi-quantitative growth data from chronic wound tissue 217
samples and intact skin swabs, respectively. Tissue samples were grouped based upon the isolation 218
or absence of pathogenic wound organisms (1, 22, 25). From the organisms cultured, S. aureus was 219
the only overt pathogen identified using the defined assessment parameters (1, 22, 25). Intact skin 220
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 9
9
swabs produced comparatively low numbers of skin-associated bacteria with no sample producing 221
moderate to heavy growth of any bacterial isolate. 222
DGGE analysis of the microbial diversity of chronic wound and contralateral skin 223
swabs. To determine the homology between DGGE profiles of bacterial communities associated 224
with chronic wounds and contralateral controls of healthy skin, a UPGMA dendrogram was 225
constructed to compare the overall eubacterial DNA fingerprints of wound and skin communities 226
derived from chronic wound tissue samples and contralateral intact skin swabs. Similarities scores 227
ranged from 10-60%, with the average similarity score below 50% indicating that generally skin 228
surface and wounds were colonised with highly divergent consortial profiles. However, clustering 229
can be seen in Figure 1, for a minority of DGGE profiles derived from wound and skin consortia. 230
This is explored further via principal component analysis of similarity matrix scores (Figure 2), in 231
which two major clusters are apparent. Whilst one of these represents a combination of skin and 232
wound profiles, the other one is composed primarily of skin-derived consortial profiles. 233
Comparisons between diversity indices derived from individual wound and skin samples (Table 4) 234
revealed marked differences in eubacterial diversity between all the wound and skin samples and 235
those where no pathogens were cultured (p<0.05), although significant differences were not 236
detected between wounds and skin cohorts where pathogens were isolated. 237
Comparisons of 16S DNA sequence data between chronic wound tissue samples and 238
contralateral intact skin swabs for each patient. Comparisons were made between bands with 239
matching positions or taxonomic affiliation and sequences across the wound and the contralateral 240
control skin swab for each patient. Examples are given from Patient A and G (no pathogens 241
isolated) are presented in Figures 3 and 4, and Patient D and I; Figures 5 and 6 (wound pathogens 242
isolated). In general, a greater proportion of skin-associated bacteria were detected in contralateral 243
wound sites (two or more correlated bands in n=8 samples) where no wound pathogens were 244
cultured when compared to the proportion of skin associated bacteria found in wounds were 245
pathogens were cultured (no correlated bands in eleven samples). 246
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 10
10
Corroboration of isolation data by DGGE sequence analyses. DGGE-derived sequence 247
identities were compared to isolation data for each patient and between sample cohorts. Sequence 248
analyses of DGGE amplicons suggested the presence of Staphylococcus sp. in 8/13 and S. aureus in 249
2/13 wound samples from which S. aureus was cultured. PCR amplicons with homology to 250
coagulase negative staphylococci were detected in 8/13 and Staphylococcus sp. in 3/13 of the wound 251
samples where no pathogens were isolated. For both tissue and skin samples, the most prevalent 252
genera were staphylococcus and bacillus; the latter were not detected by culture. Additionally, a 253
greater number of obligate anaerobic organisms were identified in both the skin and tissue isolates 254
by DGGE in comparison to culture. According to PCR-DGGE analyses, of the 22 genera identified 255
in the wound tissue and 21 genera in skin swabs, four were unique to the wounds (Klebsiella sp. 256
Abiotrophia sp. Escherichia coli and Peptoniphilus sp.) and three were unique to the intact skin 257
swabs (Kocuria rhizophilia, Morexellaceae sp. and Rhodocyclaceae sp.). 258
259
DISCUSSION 260
The aetiology of chronic wounds commonly relates to underlying pathologies; initiation is often 261
associated with primary tissue damage which creates a portal of entry for microorganisms in which 262
complex microbial communities can develop and infection may occur (15, 26, 37). The progression 263
and chronicity of wounds can be correlated with infection which, from a microbiological 264
perspective is dependent upon the types of bacteria present and their relationship with the host 265
immune responses. Wound infection is commonly defined according to the presence of pathogens 266
and colonization densities exceeding ≥ 106 organisms per gram of tissue and where significant 267
tissue damage and clinical signs of infection become evident (32, 36, 37). 268
269
There has been considerable speculation regarding to the potential aetiological role of the 270
taxonomically diverse microbial populations which commonly develop within chronic wounds (14, 271
27, 37, 38) and the potential role that culture-independent techniques could play in research and 272
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 11
11
diagnosis (42). The current study investigated the relationship between the isolation of overt 273
pathogens from wounds, as defined by established culture methods, and eubacterial diversity, 274
assessed using PCR-DGGE. These techniques were also used compare the bacterial composition of 275
wounds and contralateral health skin sites. Semi-quantitative culture, often used clinically as an 276
indication of infection severity (22, 25, 35, 41) was adopted, providing a means by which culturable 277
pathogenic organisms could be detected and a relevant comparator for PCR-DGGE. 278
279
Of the 26 chronic wound tissue samples investigated, common wound pathogens could be isolated 280
from 13; S. aureus was detected in all of these in addition to enteric species and various 281
representatives of the regional skin microbiota. Additionally, two out 13 samples (E and M) were 282
also associated with Candida sp., Pseudomonas sp. and haemolytic group G streptococci. Coliforms 283
were isolated from seven samples, of which six also harboured skin and/or enteric flora, indicating 284
putative colonization or contamination. A moderate growth of coliforms was noted for patient 285
sample Z which was not deemed pathogenic based upon clinical details. A scant culture of E. coli 286
was also isolated from sample A. Whilst E. coli and other coliforms may be considered pathogenic 287
and thus significant in specific wound cultures such as gastrointestinal surgical wound sites, in the 288
context of the current study, sample A was classified as not harbouring wound pathogens due to the 289
low numbers of E. coli isolated and the wound type from which it was isolated. 290
291
Comparisons of the bacteriological composition of wounds from which pathogens had or had not 292
been isolated using eubacterial-specific PCR-DGGE detected S. aureus in 2/13 and Staphylococcus 293
sp. in 8/13 of wound samples associated with pathogens. Several sequences obtained using DGGE 294
analysis could not be identified to species level, which is commonly associated with this method. 295
Isolation methods detected a more limited range of taxa from both wound tissue and skin swabs, in 296
comparison to DGGE. Additionally, a greater number of obligate anaerobic organisms were 297
identified in both the skin and tissue isolates using the PCR-based technique, in comparison to the 298
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 12
12
culture, despite the fact that validated anaerobic isolation methods had been used. Interestingly, a 299
higher proportion of those taxa present on the contralateral (control) skin sites occurred in wounds 300
which did not harbour overt pathogens compared to those from which pathogens had been isolated. 301
Within the 13 tissue samples where pathogens had been isolated, 11 produced no bands (i.e. PCR 302
amplicons) which matched to the contralateral skin swab profile, whereas all tissue samples were no 303
pathogens had been isolated were associated with least one or more matching skin swab bands. 304
Previous diversity profiling studies of the human skin microbiota by Gao et a. (17) and Grice et al, 305
(21) suggest that whilst there is comparatively little inter-individual compositional similarity in 306
healthy skin microbiotas; high levels of conservation between the contralateral skin sites in 307
individuals can be demonstrated (17, 21). On this basis, contralateral intact skin samples may 308
provide an insight into the normal microbiota of the site and thus, the compositions of skin prior to 309
wounding. 310
311
Primary colonisers of wounds are reportedly often members of the autochthonous skin microbiota 312
due to their proximity to the tissue injury (4, 5). However, delayed healing may enable adventitious 313
bacteria to proliferate and thus compete against autochthonous species. Additionally, since the 314
microbiota of healthy skin is likely to be water and nutrient-limited (6), hydration and nutrient 315
availability may have a marked influence on the microbial composition of wounds. Restricted 316
nutrient and moisture content of healthy skin may limit the proliferation of fastidious organisms and 317
thus, select for Gram positive bacteria such as coagulase-negative staphylococci, corynebacteria and 318
propionibacteria (6, 7). In contrast, the comparatively nutrient-rich environment of a wound may 319
facilitate the growth of a wider variety of organisms including S. aureus, P. aeruginosa, 320
streptococci, enterobacteriaceae and other facultative anaerobic species (6, 9). The transition from 321
healthy skin to colonised/infected wound may therefore be associated with the clonal expansion of 322
bacteria not normally associated with health. In most cases, wounds were colonized by more diverse 323
microbial communities than healthy skin, but the overall eubacterial diversity of wounds which 324
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 13
13
harboured pathogens and those which no pathogens were isolated did not significantly differ. It 325
could be agued that this observation highlights the utility of culture as a straightforward means of 326
detecting wound pathogens. However, it also indicates the need for further investigations of 327
potential associations between microbial profiles and clinical outcome because the contribution of 328
altered microbial colonization in causality remains poorly understood, in contrast to the 329
involvement of overtly pathogenic species which are more readily detectable by culture (25). 330
Previous analysis of chronic wound diversity by Dowd et al. (11) using pyrosequencing, DGGE and 331
rRNA gene shotgun sequencing identified several genera and species not isolated upon culture, a 332
method which in general failed to identify the primary bacterial populations of the wounds tested 333
(11). Importantly, sequence analyses of DGGE amplicons within the current study indicated that 334
wounds which did not harbour pathogens were associated with a greater proportion of normal skin 335
bacteria then were infected wounds. 336
337
This study provides a cross-sectional assessment of the bacterial diversity of wounds, which were 338
initially assessed according to the presence or absence of culture-defined pathogenic species. It also 339
provides an opportunity to compare isolation methods to a culture-independent DNA profiling 340
technique. In some cases, pathogens detected by isolation were not detected by PCR-DGGE and 341
conversely, bacterial diversity indicated on DGGE gels was not readily detectable by culture. Both 342
DGGE and culture have distinct characteristics: culture is relatively simple to implement; cost-343
effective and can detect numerically dominant culturable pathogens. It is however limited by the 344
culturability of target bacteria. Conversely, DGGE can be used to analyse complex communities 345
whilst facilitating the identification of unculturable organisms which may only represent as little as 346
1% of the total bacterial population (33). Due to the length of DNA sequences that DGGE produces 347
however, taxonomic associations made may not be categorical. 348
349
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 14
14
Analysis of the microbiotas of wound and contralateral skin sites indicated that in general, healthy 350
skin-associated organisms were underrepresented in wounds from which pathogens were cultured, 351
but that significant alterations in total eubacterial diversity were not detected. Therefore, whilst 352
DGGE is a useful tool for the reproducible culture-independent profiling of bacterial consortia, data 353
presented in the current investigation also highlight the utility of culture. On this basis, the two 354
analytical approaches are complimentary. 355
356
ACKNOWLEDGEMENT 357
This work was partially supported by grants from the BBSRC and Convatec. 358
359
REFERENCES 360
361 1. Angel, D. E., P. Lloyd, K. Carville, and N. Santamaria. 2011. The clinical efficacy of two 362
semi-quantitative wound-swabbing techniques in identifying the causative organism(s) in infected 363
cutaneous wounds. International Wound Journal 8:176-185. 364
2. Bass, M. J., and L. G. Phillips. 2007. Pressure Sores. Current Problems in Surgery 44:101-365
143. 366
3. Bourke, W. J., C. M. O'Connor, M. X. FitzGerald, and T. J. McDonnell. 1994. 367
Pseudomonas aeruginosa exotoxin A induces pulmonary endothelial cytotoxicity: protection by 368
dibutyryl-cAMP. Eur Respir J 7:1754-1758. 369
4. Bowler, P. G., B. I. Duerden, and D. G. Armstrong. 2001. Wound microbiology and 370
associated approaches to wound management. Clin Microbiol Rev 14:244 - 269. 371
5. Brook, I., and E. H. Frazier. 1990. Aerobic and Anaerobic Bacteriology of Wounds and 372
Cutaneous Abscesses. Arch Surg 125:1445-1451. 373
6. Chiller, K., B. Selkin, and G. Murakawa. 2001. Skin Microflora and Bacterial Infections 374
of the Skin. Journal of Investigative Dermatology Symposium Proceedings 6:170-174. 375
7. Cogen, A. L., V. Nizet, and R. L. Gallo. 2008. Skin microbiota: a source of disease or 376
defence? British Journal of Dermatology 158:442-455. 377
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 15
15
8. Davies, C. E., K. E. Hill, M. J. Wilson, P. Stephens, C. M. Hill, K. G. Harding, and D. 378
W. Thomas. 2004. Use of 16S Ribosomal DNA PCR and Denaturing Gradient Gel Electrophoresis 379
for Analysis of the Microfloras of Healing and Nonhealing Chronic Venous Leg Ulcers. J. Clin. 380
Microbiol. 42:3549-3557. 381
9. Davies, C. E., M. J. Wilson, K. E. Hill, P. Stephens, C. M. Hill, K. G. Harding, and D. 382
W. Thomas. 2001. Use of molecular techniques to study microbial diversity in the skin: chronic 383
wounds reevaluated. Wound Repair Regen 9:332 - 340. 384
10. Dinges, M. M., P. M. Orwin, and P. M. Schlievert. 2000. Exotoxins of Staphylococcus 385
aureus. Clin. Microbiol. Rev. 13:16-34. 386
11. Dowd, S., Y. Sun, P. Secor, D. Rhoads, B. Wolcott, G. James, and R. Wolcott. 2008. 387
Survey of bacterial diversity in chronic wounds using Pyrosequencing, DGGE, and full ribosome 388
shotgun sequencing. BMC Microbiology 8:43. 389
12. Eaglstein, W. H., and V. Falanga. 1997. Chronic Wounds. Surgical Clinics of North 390
America 77:689-700. 391
13. Edwards, M. L., A. K. Lilley, T. H. Timms-Wilson, I. P. Thompson, and I. Cooper. 392
2001. Characterisation of the culturable heterotrophic bacterial community in a small eutrophic lake 393
(Priest Pot). FEMS Microbiology Ecology 35:295-304. 394
14. Edwards, R. H., Keith G. 2004. Bacteria and wound healing. Skin and soft tissue 395
infections 17:91-96. 396
15. Fazli, M., T. Bjarnsholt, K. Kirketerp-Moller, B. Jorgensen, A. S. Andersen, K. A. 397
Krogfelt, M. Givskov, and T. Tolker-Nielsen. 2009. Nonrandom Distribution of Pseudomonas 398
aeruginosa and Staphylococcus aureus in Chronic Wounds. J. Clin. Microbiol. 47:4084-4089. 399
16. Gafan, G. P., V. S. Lucas, G. J. Roberts, A. Petrie, M. Wilson, and D. A. Spratt. 2005. 400
Statistical Analyses of Complex Denaturing Gradient Gel Electrophoresis Profiles. J. Clin. 401
Microbiol. 43:3971-3978. 402
17. Gao, Z., C.-h. Tseng, Z. Pei, and M. J. Blaser. 2007. Molecular analysis of human 403
forearm superficial skin bacterial biota. Proceedings of the National Academy of Sciences 404
104:2927-2932. 405
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 16
16
18. Gontcharova, V., E. Youn, Y. Sun, R. D. Wolcott, and S. E. Dowd. 2010. A comparison 406
of bacterial composition in diabetic ulcers and contralateral intact skin. Open Microbiol J 4:8-19. 407
19. Gorbet, M. B., and M. V. Sefton. 2005. Endotoxin: The uninvited guest. Biomaterials 408
26:6811-6817. 409
20. Gottrup, F. 2004. A specialized wound-healing center concept: importance of a 410
multidisciplinary department structure and surgical treatment facilities in the treatment of chronic 411
wounds. Am J Surg 187:38S-43S. 412
21. Grice, E. A., H. H. Kong, G. Renaud, A. C. Young, N. C. S. Program, G. G. Bouffard, 413
R. W. Blakesley, T. G. Wolfsberg, M. L. Turner, and J. A. Segre. 2008. A diversity profile of 414
the human skin microbiota. Genome Res. 18:1043-1050. 415
22. Healy, B., and A. Freedman. 2006. ABC of wound healing: Infections. BMJ 332:838-841. 416
23. Hill, K. E., C. E. Davies, M. J. Wilson, P. Stephens, K. G. Harding, and D. W. Thomas. 417
2003. Molecular analysis of the microflora in chronic venous leg ulceration. J Med Microbiol 418
52:365-369. 419
24. Howell-Jones, R. S., M. J. Wilson, K. E. Hill, A. J. Howard, P. E. Price, and D. W. 420
Thomas. 2005. A review of the microbiology, antibiotic usage and resistance in chronic skin 421
wounds. J Antimicrob Chemother 55:143 - 149. 422
25. HPA. 2009. Investigation of Skin and Superficial and Non-surgical Wound Swabs, UK 423
Standards for Microbiology Investigations, vol. 11. V11 Issue 5 http://www.hpa-424
standardmethods.org.uk/documents/bsop/pdf/bsop11.pdf. 425
26. Jones, S. G., R. Edwards, and D. W. Thomas. 2004. Inflammation and Wound Healing: 426
The Role of Bacteria in the Immuno-Regulation of Wound Healing. International Journal of Lower 427
Extremity Wounds 3:201-208. 428
27. Kirketerp-Moller, K., P. O. Jensen, M. Fazli, K. G. Madsen, J. Pedersen, C. Moser, T. 429
Tolker-Nielsen, N. Hoiby, M. Givskov, and T. Bjarnsholt. 2008. The distribution, organization 430
and ecology of bacteria in chronic wounds. J. Clin. Microbiol.:JCM.00501-08. 431
28. Ledder, R. G., P. Gilbert, A. Pluen, P. K. Sreenivasan, W. De Vizio, and A. J. McBain. 432
2006. Individual microflora beget unique oral microcosms. J Appl Microbiol 100:1123-31. 433
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 17
17
29. Lyczak, J. B., C. L. Cannon, and G. B. Pier. 2000. Establishment of Pseudomonas 434
aeruginosa infection: lessons from a versatile opportunist. Microbes Infect 2:1051-60. 435
30. McBain, A. J., R. G. Bartolo, C. E. Catrenich, D. Charbonneau, R. G. Ledder, and P. 436
Gilbert. 2003. Growth and molecular characterization of dental plaque microcosms. J Appl 437
Microbiol 94:655-64. 438
31. Melendez, J. H., Y. M. Frankel, A. T. An, L. Williams, L. B. Price, N. Y. Wang, G. S. 439
Lazarus, and J. M. Zenilman. 2010. Real-time PCR assays compared to culture-based approaches 440
for identification of aerobic bacteria in chronic wounds. Clinical Microbiology and Infection 441
16:1762-1769. 442
32. Murphy, R. C., M. C. Robson, J. P. Heggers, and M. Kadowaki. 1986. The effect of 443
microbial contamination on musculocutaneous and random flaps. J Surg Res 41:75-80. 444
33. Muyzer, G., E. C. de Waal, and A. G. Uitterlinden. 1993. Profiling of complex microbial 445
populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-446
amplified genes coding for 16S rRNA. Appl Environ Microbiol 59:695-700. 447
34. Posnett, J., and P. J. Franks. 2008. The burden of chronic wounds in the UK. Nurs Times 448
104:44-5. 449
35. RNAO. 2005. Assessment and management of foot ulcers for people with diabetes. 450
Canada:Registered Nurses' Association of Ontario. 451
36. Robson, M. C. 1979. Infection in the surgical patient: an imbalance in the normal 452
equilibrium. Clin Plast Surg 6:493-503. 453
37. Robson, M. C. 1997. Wound Infection: A Failure of Wound Healing Caused by an 454
Imbalance of Bacteria. Surgical Clinics of North America 77:637-650. 455
38. Robson, M. C., B. D. Stenberg, and J. P. Heggers. 1990. Wound healing alterations 456
caused by infection. Clinics in Plastic Surgery 17:485-492. 457
39. Singh, S. K., K. Gupta, S. Tiwari, S. K. Shahi, S. Kumar, A. Kumar, and S. K. Gupta. 458
2009. Detecting Aerobic Bacterial Diversity in Patients With Diabetic Foot Wounds Using ERIC-459
PCR: A Preliminary Communication. The International Journal of Lower Extremity Wounds 8:203-460
208. 461
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 18
18
40. Steel, K. J., G. I. Barrow, R. K. A. Feltham, and S. T. Cowan. 1993. Cowan and Steel's 462
manual for the identification of medical bacteria. Cambridge : Cambridge University Press 463
41. Stone, N. D., D. R. Lewis, H. K. Lowery, L. A. Darrow, C. M. Kroll, R. P. Gaynes, J. A. 464
Jernigan, J. E. McGowan, Jr., F. C. Tenover, and C. L. Richards, Jr. 2008. Importance of 465
bacterial burden among methicillin-resistant Staphylococcus aureus carriers in a long-term care 466
facility. Infect Control Hosp Epidemiol 29:143-8. 467
42. Thomsen, T. R., Aasholm, M. S., Rudkjøbing, V. B., Saunders, A. M., Bjarnsholt, T., 468
Givskov, M., Kirketerp-Møller, K., et al. . 2008. The bacteriology of chronic venous leg ulcer 469
examined by culture-independent molecular methods. Wound repair and regeneration 18:38-49. 470
43. Vasconez, L. O., W. J. Schneider, and M. J. Jurkiewicz. 1977. Pressure sores. Current 471
Problems in Surgery 14:1-62. 472
44. Walter, J., G. W. Tannock, A. Tilsala-Timisjarvi, S. Rodtong, D. M. Loach, K. Munro, 473
and T. Alatossava. 2000. Detection and identification of gastrointestinal Lactobacillus species by 474
using denaturing gradient gel electrophoresis and species-specific PCR primers. Appl Environ 475
Microbiol 66:297-303. 476
45. Wilson, M. J., A. J. Weightman, and W. G. Wade. 1997. Applications of molecular 477
ecology in the characterization of uncultured microorganisms associated with human disease. 478
Reviews in Medical Microbiology 8:91-102. 479
480
481
482
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 19
19
483 FIGURE LEGENDS 484
FIGURE 1. A UPGMA dendrogram for patients A – Z, showing percentage matching of wound 485 DGGE fingerprints. , debridement samples from wounds in which pathogens were isolated; , 486 debridement samples from wounds where no pathogens were isolated; , skin samples from 487 individuals wounds which harboured pathogens; , skin samples from individuals with no 488 pathogens isolated in wound tissue. 489 490 FIGURE 2. Principal component analysis of DGGE fingerprints of chronic wound samples and 491 intact skin swab (patients A – Z). See legend to Figure 1. 492 493 FIGURE 3. Characterization major taxa in wound and skin samples based on dominant PCR 494 amplicons and matched bands derived from DGGE gels (patient A; no pathogens isolated). 495 496 FIGURE 4. Characterization major taxa in wound and skin samples based on dominant PCR 497 amplicons and matched bands derived from DGGE gels (patient G; no pathogens isolated). 498 499 FIGURE 5. Characterization major taxa in wound and skin samples based on dominant PCR 500 amplicons and matched bands derived from DGGE gels (patient D; pathogens isolated). 501 502 FIGURE 6. Characterization major taxa in wound and skin samples based on dominant PCR 503 amplicons and matched bands derived from DGGE gels (Patient I; pathogens isolated). 504 505 506 507
508
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 20
20
509
TABLE 1. Bacteriological agars used in the study 510
511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
Medium Incubation conditions Target Bacterial group*
Cysteine lactose electrolyte deficient agar 5% CO2, 37°C Enterobacteriaceae
Pseudomonads
5% Horse Blood Agar/Vacomycin**
Anaerobic, 37°C Gram Negative Anaerobes
5% Horse Blood Agar/Neomycin**
Anaerobic, 37°C Gram Positive Anaerobes
5% Horse Blood Agar 5% CO2, 37°C Streptococci (Lancefield Groups A, C and G) Pasteurella spp. S. aureus Vibrio spp. Aeromonas spp.
*Commonly isolated wound associated pathogens which may be indicative of infection. (25) **Addition of 5µg Metronidazole Disc
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 21
21
TABLE 2. Microbial characterization of chronic wounds by semi-quantitative culture.
CNS: Coagulase negative staphylococci, GBS: β haemolytic streptococci (Lancefield Group B), GDS: Enterococcus faecalis (Lancefield Group D), GGS: β haemolytic streptococci (Lancefield Group G), GPC: Gram positive cocci. 1+, Light growth, 2+ Moderate Growth, 3+ Heavy Growth (see methods section for definitions). Shading indicates isolation of wound pathogens from tissue according to established culture-based criteria. Blank cells, none detected. *Genus, **Species identified on DGGE sequence analysis from bands derived from the sample or aligned sequenced bands
Bacterial group Patient Code
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
S. aureus 3+(*) 2+(**) 2+ 2+ 2+(*) 3+(*) 2+(*) 1+(*) 1+(**) 1+ 2+(**) 1+(*) 2+(*)
CNS 3+(*) 3+(*) 3+(**) 2+ 2+(*) 2+(**) 3+(**) 3+(*) 2+(**) 3+(**) 3+ 3+ 1+(**) 1+(**) 3+(*) 1+(**) 2+(**) 3+(**) 1+ 3+(**)
E. coli 1+ 3+ 3+
Coliform 2+(**) 3+(*) 2+ 1+ 3+ 3+(*) 3+(**)
GDS 1+(**) 2+(*) 2+(*) 1+
GBS 1+ 2+
GGS 2+(*)
Corynebacterium
sp. 2+ 2+ 2+
Pseudomonas sp. 1+
Micrococcus sp. 2+
Anaerobes
(GPC) 1+(**)
Candida sp. 2+
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 22
22
TABLE 3. Microbial characterization of contralateral skin sites by semi-quantitative culture.
Bacterial group Patient Code
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
CNS S(**) S(**) S(*) S(**) S(**) S 1+(*) S S(**) S(**) 1+(**) 1+(**) 1+(**) 1+(**) 1+(**) 1+(**) 1+(**) 1+(**) 1+(*) 1+(**) S(*) S S
Corynebacterium spp. S S S 1+
Micrococcus spp. S 1+
CNS: Coagulase-negative staphylococci. S, scant growth; 1+, light growth (see methods section for definitions). Shading indicates skin samples from contra lateral sites of patients with wounds which cultured pathogens, defined according to established culture-based criteria. Additional bacterial groups shown in Table 2 were not detected. See legend to Table 2.
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 23
23
TABLE 4. Eubacterial diversity indices and proportion of skin amplicons also detected in wounds. Patient code
ShannonWeaver diversity indices Shared amplicons* Wound skin
A 0.583 0.901 4/11
B 0.583 0.410 0/5
C 1.166 0.328 3/4
D 0.510 0.328 0/4
E 0.364 0.164 0/2
F 0.219 0.819 0/10
G 0.656 0.492 2/7
H 0.874 0.246 2/3
I 0.656 0.410 0/5
J 0.801 0.492 1/6
K 0.510 0.410 0/5
L 0.583 0.655 0/8
M 0.729 0.328 1/4
N 0.364 0.573 0/6
O 0.364 0.328 2/4
P 0.801 0.082 0/1
Q 0.219 0.246 0/3
R 0.437 0.492 2/6
S 0.510 0.819 3/10
T 0.146 0.328 1/4
U 0.801 0.410 1/5
V 0.510 0.328 2/3
W 1.239 0.410 0/5
X 0.364 0.328 0/4
Y 0.583 0.492 0/6
Z 0.729 0.492 1/6
Shaded rows indicate wounds which harboured pathogens based on established culture criteria. *Proportion of bands present in intact skin DGGE analysis found in chronic wound DGGE analysis. Diversity indices were compared between wound and skin, and infected and non-infected cohorts by the Mann-Whitney U test. A significant difference was found between diversity indices of wound and skin samples and between wound and skin samples when grouped into non-infected cohorts P<0.05. No significant difference was found between wound and skin when grouped into infected cohorts
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 24
24
FIGURE 1. Oates et al.
X Y G I M V V N F R X A K L D Z E F Y A C Q W H Z P U R S N S K T G E D H B C Q P B O O M T U J J I L
W
100
908070605040302010
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 25
25
FIGURE 2. Oates et al.
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
-0.8-0.6
-0.4-0.2
0.00.2
0.40.6
0.81.0
-0.8-0.6-0.4-0.20.00.20.40.60.8
PC3
(9%
var
ianc
e)
PC1 (
21%
varia
nce)
PC2 (14% variance)Wound (infected group)Intact skin (infected group)Wound (non-infected group)Intact skin (non-infected group)
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 26
26
FIGURE 3. Oates et al.
Putative identity of nearest database match and accession numbers. ih, sequence has insufficient homology to enable identification; na, no corresponding band.
, uninfected wound sample; , contralateral intact skin. Superscript number, where matched between columns indicate band matched on the basis of comparative position and/or taxonomic affiliation.
Band Wound Skin 1 ih Bacteroidales bacterium HM079538
2 Staphylococcus epidermidis HM452104 Staphylococcus simulans HM4620531
3 Staphylococcus simulans HM4520001 Staphylococcus sp. HM0748362
4 Staphylococcus sp. HM0748362 Enterococcus faecalis HQ2597273
5 Enterococcus faecalis HQ2597273 ih
6 Streptococcus sp. AY806239 Staphylococcus epidermidis AM697667
7 Bacterium HM338822 Bacterium HM344790
8 Micrococcus yunnanensis HQ2857734 Bacillaceae bacterium EU596919
9 na Micrococcus yunnanensis HQ2857734
10 na ih
11 na ih
1
2
3 4
5 6
7
8
1
2
3 4 5
6 7 8
9 10
11
A
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 27
27
FIGURE 4. Oates et al.
See legend to Figs. 1 and 3.
Band Wound Skin 1 Clostridium sp. DQ169781 Staphylococcus sciuri HQ1545801
2 Staphylococcus sp. HM0743051 Variovorax spp. HM113661
3 Staphylococcus aureus HM209755 Staphylococcus simulans HM462053
4 Prevotella bivia AB547674 Sphingomonas sp. HM3462052
5 Streptococcus dysgalactiae HM359249 ih
6 Stenotrophomonas sp. FJ609992 ih
7 Sphingomonas sp. HM3462052 Kocuria rhizophila FR687213
8 ih na
9 ih na
10 ih na
G
6
3
1
2
4
5
6 7
8
9
10
1
2
3
4
5
7
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 28
28
FIGURE 5. Oates et al.
Band Wound Skin 1 Bacteroidales bacterium HM079538 Staphylococcus sp. HM074899
2 Staphylococcus sp. HM074836) bacterium HM338822 3 Staphylococcus epidermidis HM452104
ih 4 ih Enterobacter sp. HM461227 5 Acinetobacter sp. HQ246291 na
6 Staphylococcus sp. HM074305 na
7 Bacillus pumilus strain AF260751 na
8 Klebsiella sp. HQ264073 na
9 Klebsiella sp. GU294294 na
10 ih na
See legend to Figs. 1 and 3.
4
1
2 3
5 6
7
8
9
10
1
2
3
4
D
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from
Page 29
29
3
1
2
4
5
6
7
8
9
2
3
4
5
I
FIGURE 6. Oates et al.
Band Wound Skin 1 Bacteroides fragilis FQ312004 Bacillus pumilus FM179663
2 Clostridiales sp. HM076639 Sphingomonas sp. HM346205
3 Enterococcaceae sp. EU572465 Stenotrophomonas maltophilia AY321966
4 Streptococcus dysgalactiae HM359249 Bacillus pumilus (FM179663)
5 ih Actinomycetales sp. HM077215
6 Enterobacter sp. HM461227 na 7 Bacillaceae bacterium EU596919 na 8 Enterobacter sp. FJ609991 na 9 Enterobacter sp. FJ609991 na See legend to Figs. 1 and 3.
on February 13, 2018 by guest
http://jcm.asm
.org/D
ownloaded from