Understanding the Burn Wound Microbiome: Comparing Traditional Wound Cultures to Next Generation Sequencing Technology

1. - INTRODUCTION: The human microbiome consists of microorganisms including bacteria, fungi, archaea and viruses that live in association with the body. In the average person, bacteria alone outnumber host cells ten-fold and contain a thousand more genes than the human genome. Therefore, despite consisting of \<3% of the body's biomass, the human microbiome plays a crucial role in establishing and maintaining the complex biomechanics of our health and development. In the last two decades, the NIH Human Microbiome Project has spotlighted a number of newly discovered influences the microbiome has on human physiology (Turnbaugh et al., 2007). Despite the remarkable progress achieved in the last two decades, the diverse ecosystem of microbiota present in cutaneous wounds and their influence on host regeneration, immune response and wound healing have only recently begun to be deciphered. To date we have only a superficial understanding of the burn wound microbiome and its impact on infection, systemic inflammation, wound healing and ultimately patient prognosis.

2. - BACKGROUND: Severe burns are a major global source of morbidity and mortality. According to the National Center for Injury Prevention and Control and the American Burn Association, 450,000 patients suffer thermal burns each year in the United States alone; of those admitted for treatment, 3% will not survive mostly due to infection and sepsis. After severe burn, the pattern of microbial colonization in burn wounds follows a relatively predictable sequence of supplementation beginning with gram-positive bacteria followed by gram-negative bacteria and subsequently more virulent pathogens. Several studies showed length of hospitalisation is associated with the microbial species isolated from burn wounds. Numerous studies explored the microbiome of chronic cutaneous wounds with findings suggesting healing is influenced by both pathologic and beneficial microbial populations. Traditional microbiology culture analysis is a surrogate for the actual wound microbiome but fails to identify the full breath of the microbial ecosystem present in burn wounds, as many species are difficult to cultivate under standard conditions and the culture environment itself differs substantially from the wound environment. Furthermore, the wound microbiome is dynamic and changing influenced by environmental factors, wound care agents, the host immune response, and medications. In addition, traditional wound cultures take several days to finalise and over the course of such time may no longer reflect the current microbial residents of the wound itself. Lastly, a great deal of attention is focussed on known pathogenic species and multi-drug resistant strains (Pseudomonas aeruginosa, Haemophilus influenzae, carbapenem-resistant Enterobacteriaceae to name a few), but the presence and influence of beneficial species is often overlooked.

   Culture-independent methods using Next Generation Sequencing (NGS) technology, such as the 16S and 18S amplicon sequencing and metagenome shotgun sequencing, have proven to be faster and more sensitive than traditional microbiology cultures. 16S NGS technology utilises 16S rDNA to encode for ribosomal rRNA sequences of prokaryotes. Nine variable regions are present within the 16S rDNA sequence, which target genetic differences between bacterial species. In a similar manner 18S NGS technology encodes for ribosomal rRNA sequences of eukaryotes and the variable regions are used to classify species differences among eukaryotic micro-organisms in a sample.

3. - STUDY PROCEDURES Specimen and Data Collection: The Burn Wound Data/Bio-Repository will collect biospecimens from excised burn wound tissues to be discarded by the surgical team. No more than two progenitor specimens will be collected from any individual operative case. Each specimen will be divided at the time of collection into two sub-specimens that each weigh approximately 0.5-2 grams. Each specimen must be from only one specific body region classified as upper extremities (either arm or hand), lower extremities (either leg or foot), anterior thorax, posterior thorax or head/neck. In addition, no specimens will be collected from the genitalia or perineum due to the high risk of specimen contamination.

The sub-specimens will be transported immediately on ice and temporarily stored in a -20ºC freezer, then a portion of the tissue (less than 0.5 cm) will be placed in RNAlater(R) RNA Stabilization Solution (RNAlater(R)) in order for nucleotide preservation/stabilization. Tissue placed in RNAlater(R) will be refrigerated at 4ºC overnight to allow sufficient tissue penetration. Afterwards preserved specimens will be stored at -20ºC to -80ºC. This process will be performed within 1 hour of collection. Each sub-specimen will generate two additional specimens to be placed in RNAlater(R). After the initial period of refrigeration, RNAlater(R) specimens will be stored in the Second Bio-Repository in a -80ºC freezer. Un-used tissue will be labeled with the same de-identified sequence and barcode and stored in a -80ºC freezer for a period of no more than 12 months and no less than 6 months (First Bio-Repository).

To address the other labeled sub-specimen, it will be transported to the Clinical Microbiology Laboratory (CML) in a 15mL specimen collection vial to undergo microbiology culture analysis. After completion of the microbiology culture analysis, specific bacterial colonies from the culture plates at the CML will be isolated and processed in a nucleotide preserving solution (RNAlater(R)). Similar to their counterparts from the Second Bio-Repository, the specimens will be batched for NGS analysis. When sufficiently batched, specimens from the Third Bio-Repository will undergo nucleotide extraction by the NGS Facility staff prior to NGS analysis using Shotgun Whole Genome Sequencing analysis. Patients/LARs will be approached for informed consent after introduction to the study by the clinical staff and confirmed that more information regarding informed consent is acceptable. Once signed informed consent is obtained, clinically pertinent wound-specific data will be collected for the First Database under the same de-identified label. If informed consent is not obtained, the specimens will be discarded and not considered further. All collected information will be linked to the de-identified unique identifiers, which will be specific to each progenitor specimen. All collected data will be independent of the clinical record and stored securely.

The following patient-specific data will be collected:

• Age of patient in years at time of burn injury (input is numerical in years)
• Gender of patient at time of burn injury (categories are Male, Female, Other)
• Ethnicity of the patient (categories are White, Black, Hispanic, Asian, Native American, Pacific Islander, and other)
• Weight and BMI of the patient at the time of surgery (input is numerical)
• Pregnancy status (categories are positive, negative or unknown)
• COVID-19 status (categories are positive, negative or unknown)
• Oral or IV antibiotics received within 24 hours of surgery (categories are vancomycin, meropenem, daptomycin, Zosyn, doxycycline, Ancef, Bactrim, Augmentin, any anti-fungal or any anti-viral)
• Medication Class - home and inpatient (categories are diabetic therapy, steroids, immunotherapy, or chemotherapy)
• Any previous de-identified sequence associated with the subject from the current admission
• TBSA burned (input is numerical: % total body surface area)

The following wound-specific data will be collected:

• Days since admission and hospital days on the day of collection (input is numerical)
• Days since burn injury (input is numerical)
• Number of operative interventions during the current admission including the surgery at time of specimen collection (input is numerical)
• Number of total operative interventions associated with index burn injury including operations performed at other facilities, operations performed on previous admissions to UTMB and the surgery at time of specimen collection (input is numerical)
• Body location from which the specimen is collected (categories are upper extremity either arm or hand, lower extremity either leg or foot, anterior thorax, posterior thorax or head/neck). Tissues from the genitalia and groin will not be collected.
• Total body surface area burned (TBSA) percentage as determined on admission (input is numerical)
• Concern for full thickness or 3rd degree burn injury depth from the site where specimen is being collected (categories are yes or no)
• Presumed type of tissue from the site where specimen is being collected (categories are cutaneous, subcutaneous or myofascial)
• Etiology of burn injury (categories are flame, scald, contact, chemical, electrical or other)
• Antimicrobial wound care immediately prior to surgery on the wound from with tissue is being collected (categories include antibiotic ointment, Dakins' solution irrigation, silver nitrate irrigation, mafenide acetate irrigation, amphotericin soaks, silver sulfadiazine cream, silver-based synthetic dressing, non-silver-based synthetic dressing).

BIOSTATISTICS

We are proposing to gather at least 300 progenitor samples to include in the Bio-Repository to analyze data to answer (at least) the following questions:

• What is the correlation between the NGS burn wound microbiome and traditional culture techniques?
• What is the relative distribution of microbe species in the burn wound?
• What is the relative distribution of microbe number and concentration in burn wounds compared to traditional techniques?

All of these questions can be answered solely with de-identified discarded samples. For additional questions, those samples with associated clinical and wound-specific data can be used to answer the following:

• Does age in years of the patient engender different burn wound microbiome? Is this reflected with classical culture techniques?
• Does gender have any effect on burn wound microbiome? Is this reflected with classic culture techniques?
• Does ethnicity affect the burn wound microbiome? Weight and BMI? Pregnancy? COVID-19? Burn size?
• Does topical and/or systemic antibiotic use affect the burn wound microbiome?
• Does age of the burn wound affect the burn wound microbiome? Does number of previous or subsequent operative interventions have any effect?
• Does the body region injured have any effect on the burn wound microbiome?

STATISTICAL ANALYSIS:

Descriptive statistics for continuous variables will be presented as mean ± SD, and for categorical variables as frequencies and proportions. Bivariate analysis will include correlation testing with Pearson correlation coefficient where the normality assumption is not violated, and Spearman's rho when normality is not assumed. Using multiple linear regression models, we will assess the relationship between NGS burn wound microbiome and traditional culture techniques as well as determine the predictors of burn wound microbiome in collected biospecimens from adult burn patients. Diagnostic testing will be performed on the full multiple regression models. Two-sided statistical significance will be considered for alpha at the 0.01 level.

Power analysis: Power analysis for our prediction model was based on an effect size f2 of 0.136 calculated from anticipated r2 of 0.12 with 20 predictors at alpha=0.01 and for a power of the f test to reject the null hypothesis at 90%. The sample size needed for such a model with 20 predictors to explain at least 12% of the variation on the dependent variable (burn wound microbiome) at 90% power and at alpha=0.01 is 271 progenitor samples. We aim to collect 300 progenitor samples to account for 10% analytical failure. We used G*power (version 3.1.9.7) under the f test family to estimate this sample size suggested.