Metagenome Analysis of Cholesteatoma-associated Bacteria: A Pilot Study

Abstract

Introduction

Cholesteatoma usually harbors a poly-microbial infection. As the diversity of bacterial pathogens in the Indian COM is unknown, we set out to identify the bacteria associated with cholesteatoma disease in different patients of North India using targeted metagenomic analysis of the 16 S rRNA gene.

Methods

We recruited 15 patients of cholesteatomatous chronic otitis media (COM), who underwent surgical disease clearance. We divided these patients into four groups based on the four clinic-radiological stages categorized as per the EAONO/JOS joint consensus statement classification. Representative samples were extracted during the surgery and sent for bacterial culture and sensitivity and 16 S rRNA gene metagenomic analysis.

Results

While 12 (80%) of the patients belonged to clinical Stage I/II; one patient had an extracranial complication (stage III) and two patients had an intracranial complication (stage IV). Our detailed bacterial metagenomics analyses showed that while phylum Proteobacteria was most abundant (reads up to ∼ 95%) in specimens from nine patients, phylum Firmicutes was most abundant (up to ∼ 80%) in specimens from four patients. Gamma (γ) Proteobacteria and Epsilon (ε) Proteobacteria were the most abundant class amongst Proteobacteria. Class Tissierellia stood out as the most abundant Firmicutes (40–60%), followed by Clostridia (20%) and Bacilli (10%). There was negligible difference in the bacterial profiles across all four clinical stages.

Conclusion

Cholesteatoma is primarily associated with Proteobacteria and Firmicutes phyla, even in complicated disease. Further studies with a larger sample size are required to validate our findings.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12070-024-04678-9.

Introduction

Chronic otitis media (COM) is one of the chief causes of hearing impairment in children and young adults in developing countries including India [1]. COM is primarily of two types- mucosal and squamosal. Squamosal COM occurs due to epithelium growth inside the middle ear cavity due to infections, metaplasia, or tympanic membrane retraction due to eustachian tube dysfunction [2]. This epithelium is known as cholesteatoma. Cholesteatoma is a cystic, non-tumorous, destructive lesion of the temporal bone primarily consisting of misplaced epithelium that may induce severe complications due to continuous growth at the expense of underlying bone resorption. Lytic enzymes, microorganisms and several inflammatory mediators are an integral part of cholesteatoma matrix [3].

Chole et al. have shown the presence of bacterial biofilms in cholesteatoma specimens hence, antibiotic resistance during clinical treatment [4]. Hence, knowing the associated bacteria with the disease can aid in prescribing tailor-made antibiotic therapy to a patient, in case of extra-temporal spread of disease or otherwise, in a healing post-surgical cavity. Cholesteatoma usually harbors a poly-microbial infection. Pseudomonas aeruginosa, Staphylococcus aureus, Proteus mirabilis, Escherichia coli and anaerobes have been reported to be the most abundant microbes in various studies [5–7].

As the diversity of bacterial pathogens in the Indian COM is unknown, we set out here to identify the bacteria associated with cholesteatoma disease in different patients of North India and to look for any observable differences in patients with aggressive disease. Typical culturing of the infected tissue and surface-scrapings is inadequate to provide a holistic view of bacterial diversity as the media used for culturing may be unsupportive to their in vitro growth [8].

Therefore, with our cholesteatoma specimens, by performing targeted metagenomic analysis of the 16 S rRNA gene, we (i) profiled and determined the relative abundance of the bacterial diversity, and (ii) evaluated if the identified profile and abundance pattern(s) correlate(s) with the aggressive form of the disease. To our best knowledge, our analyses is the first report from India to provide insights into COMs-associated bacterial profile and their relative abundance.

Methodology

Study Duration and Location

This prospective study was conducted over three years (2019–2021) at our tertiary care center in North India.

Study Population

With an appropriate informed written consent, we recruited 15 patients with squamosal chronic otitis media (COM) with cholesteatoma. Their disease was clinically categorized as per the EAONO/JOS joint consensus statement classification (Table 1). All patients underwent high-resolution computed tomography (HRCT) of temporal bone to radiologically stage the disease. The patients were recruited and operated at our tertiary care center. The patients who had non-cholesteatomatous COM and the patients who did not consent to participate in the study were excluded. The study was approved by the institutional ethics committee (IEC-621/02.11.2018, RP-35/2018).

Table 1.

EAONO/JOS joint consensus statement classification for chronic otitis media with cholesteatoma

Staging systems for respective cholesteatoma types:
1) Pars flaccida cholesteatoma (attic cholesteatoma)	2) Pars tensa and congenital cholesteatoma
Stage 1: Disease localized in the attic	Stage I: localized in the tympanic cavity
Stage II: Cholesteatoma involving two or more sites	Stage II: Cholesteatoma involving two or more sites
Stage III: Cholesteatoma with extracranial complications*	Stage III: Cholesteatoma with extracranial complications*
Stage IV: Cholesteatoma with intracranial complications#	Stage IV: Cholesteatoma with intracranial complications#

• Facial palsy, labyrinthine fistula, labyrinthitis, Postauricular abscess or fistula, zygomatic abscess, neck abscess

# Purulent meningitis, epidural abscess, subdural abscess, brain abscess, sinus thrombosis, otitic hydrocephalus

Sample Collection

All enrolled patients underwent surgical clearance of the disease (mastoidectomy). To ensure adequate representation of the microbiome, we collected samples from the most infected part of the disease. Collected specimens were processed for NGS and in vitro culturing. They are labelled as Sample_1 to Sample_15.

Gram Staining and Culture

Cholesteatoma samples collected during mastoid surgery were first collected in saline filled specimen bottles and immediately transported at RT to the clinical microbiology lab of our hospital. In the microbiology lab, specimens were cultured on Blood agar and MacConkey agar plates and smear prepared. The specimens were also simultaneously inoculated in brain heart infusion broth (BHI) for bacterial enrichment. Microscopy was employed to identify the bacteria and the associated pus cells. All three, i.e. one plate each of blood agar, MacConkey agar, and BHI broth were incubated O/N (16–18 h) at 37 °C and then observed for growth/turbidity.

Upon detecting growth in either plates or broth we processed the bacterial colonies for identification with MALDI-TOF Mass Spectrometer (Matrix-assisted laser desorption/ionization-time of flight- Bimearux, Germany) and Antibiotic sensitivity testing (AST) performed using disc diffusion test. Culture plates without any growth in 24 h were incubated for an additional 24 h. Specimen with no growth in the solid media were evaluated for turbidity in the BHI broth. Turbid + ve samples were processed further as per SOPs. Those specimens that did not show growth even after 2 d of incubation were labelled as ‘Sterile’.

Next-generation Genome Sequencing (NGS)

For analyzing associated bacteria present in each of the cholesteatoma samples, we outsourced our 16SrRNA metagenomic sequencing to Genotypic Technology Private Limited, Bengaluru, Karnataka, India.

DNA Extraction and Quality Control

Total DNA from the received cholesteatoma samples were extracted and purified using Qiagen DNeasy Blood and tissue Kit (Cat # 69,506; Qiagen, USA) as per the manufacturer’s instructions. Samples were homogenised in liquid N₂ using Tommy Microsmash (GeneCrafts Labs, Indonesia) and treated with 300 µl Lysozyme (10 mg/ml) for 30 min at 37 °C. A total of 180 µl ATL (buffer; supplied by manufacturer as part of the kit) was then added to each sample and incubated for an additional 15 min at 56^oC. Then, 20 µl of Proteinase K (20 mg/ml; Qiagen) was added and incubation extended until lysate turned clear. The lysate was then subjected to RNase A (20 mg/ml; MP Biomedicals Cat # 2,101,076) treatment for 20 min at 65^oC. Then, 200 µl AL buffer (supplied in the kit) was added and the total lysate, mixed with 200 µl of absolute Ethanol (Analytical CSS reagent, India) and loaded onto columns placed in collection tubes. Column washes and elutions were done as per manufacturer’s protocol. DNA was eluted in 30 µl of 10 mM Tris-HCl, pH 8.0 and quantified using Nanodrop 2000 (Thermo Fisher Scientific, USA). Eluted DNA (100 ng) was electrophoresed in a 1% agarose gel (Mupid exugel system, Takara, Japan).

PCR Amplification

The obtained DNA were subjected to complete 16s rRNA gene amplification using region-specific primers (rRNA Forward primer AGAGTTTGATCCTGGCTCAG; 16s rRNA Reverse primer: TACGGCTACCTTGTTACGACTT). PCR reaction was performed in a final volume of 20 µl containing LongAmp Taq 2X master mix (NEB, USA), primers (Bioserve Biotechnologies India Pvt Ltd, India), and DNA template (50 ng/µl). PCR amplification was confirmed by electrophoresing aliquots of the amplicons on 1% agarose. The remaining amplicons DNA were purified using 1.6x AMPure XP beads (Beckmann Coulter, USA) as per manufacturer’s recommendations.

Amplicon Library Preparation and Sequencing

Amplicon library preparation was performed using sequencing Ligation kit (Oxford Nanopore Technologies, UK) and PCR barcoding kit (Oxford Nanopore Technologies). A total of 50 ng purified amplicon DNA from each sample was end-repaired (NEBnext ultra II end repair kit) and cleaned up with 1x AMPure beads. Barcode adapter ligation was performed with NEB blunt/TA ligase and cleaned with 1x AMPure beads. Qubit (Thermo Fisher Scientific, USA)-quantified adapter-ligated DNA samples were then barcoded using PCR reaction (LongAmp® Taq 2X Master Mix; NEB) and cleaned up with 1.6x AMPure beads. Barcoded samples were pooled at equimolar concentration and end-repaired using NEB next ultra II end repair kit (NEB). End-repaired DNA was cleaned up with 1x AMPure beads and sequencing adapter ligation was performed for 15 min using NEB blunt/TA ligase (NEB). The library mix was cleaned up using 0.4x AMPure beads and finally eluted in 15 µl of elution buffer (as supplied in the kit). Sequencing was performed on GridION X5 (Oxford Nanopore Technologies) using SpotON flow cell R9.4 (FLO-MIN106) in a 48 hsequencing protocol. Nanopore raw reads (‘fast5’ format) were base-called (‘fastq5’ format) and demultiplexed using Guppy v2.3.4 [9].

Data analysis

The nanopore raw reads were processed using Porechop for removal of adapter sequences. Adapter free reads were then aligned against centrifuge bacterial database using Centrifuge tool [10]. Using the genome indexing technique (which is based on Burrows-Wheeler transform) and the Ferragina-Manzini index, a fast and effective comparison was performed [11, 12]. The taxonomic classification obtained by Centrifuge were analysed using the Pavian interactive browser application [13].

Comparative Analysis

For comparative analysis, the sample specific absolute abundance was normalized to obtain relative abundance and used for comparative studies.

Results

The clinico-demographic details of the 15 patients are shown in Table 2. Thirteen (86.6%) patients were males. Eight (53%) patients were < 18 years old. All our patients had complaints of ear discharge and conductive hearing loss at the time of presentation. While 12 (80%) of the patients belonged to clinical Stage I/II, three patients had a complication at the time of presentation; one patient had an extracranial complication (post-aural abscess; patient #3) and two patients had concurrent extracranial complication with an intracranial complication (one patient had a LSCC fistula with lateral sinus thrombosis (patient #8), and the other had a LSCC fistula with cerebellar abscess (patient #12)). Both the patients with intracranial complications were assigned stage IV. All three patients with complicated disease were ≤18 years. All patients underwent canal-wall down mastoidectomy for surgical clearance of the disease. None of the patients had history of a previous surgery or comorbidity.

Table 2.

Clinico-demographic characteristics of the recruited 15 patients

Characteristic	Number of patients, n (%)
Sex
Female  Male	2 (13.3%) 13 (86.6%)
Age
Mean (in years)  Median (in years)  Age (< 18 years)  Age (> 18 years)	19 17 8 (53%) 7 (46%)
Symptoms
Ear discharge  Hearing impairment  Dizziness and vertigo  Facial weakness  Extracranial complications  Intracranial complications	15 (100%) 15 (100%) 2 (13.3%) 0 2 (13.3%) 3 (20%)
Type of hearing loss
Conductive hearing loss  Sensorineural hearing loss  Profound SNHL	15 (100%) 2 (13.3%) 0
Clinico-radiological stage (EAONO/JOS)
I  II  III  IV	4 (26.6%) 8 (53%) 1 (6.6%) 2 (13.3%)
Surgery performed
Canal wall-up mastoidectomy  Canal wall-down mastoidectomy	0 15 (100%)

Gram Staining and Culture Findings

The summarised results from the Gram staining and culture are tabulated in Table 3. Pseudomonas aeruginosa was the most common bacterium grown in culture (53%). One specimen exhibited growth of Staphylococcus aureus. Five of the remaining specimens grew ≥ three Gram negative bacilli colonies and one other specimen grew no bacteria.

Table 3.

The most abundant bacteria grown in culture along with the respective number of samples it yielded the growth

S.no.	Culture result	Number of samples
1	Pseudomonas aeruginosa	8
2	Staphylococcus aureus	1
3	> 3 gram negative bacteria	5
4	Sterile	1

Metagenomics Analyses for Bacterial Diversity and Abundance

Our detailed bacterial metagenomics analyses of 15 COM samples showed that while phylum Proteobacteria was most abundant (reads up to ∼ 95%) in specimens from nine patients (patient # 1–3, 5–8, 11, 14), phylum Firmicutes was most abundant (up to ∼ 80%) in specimens from four patients (# 4, 9, 10, 12) (Fig. 1; Table S1). The second most abundant phyla in the above nine and four patients was Firmicutes (up to ∼ 23% and Proteobacteria (up to ∼ 22%) respectively. In these nine and four samples, the other phyla among the top five included Synergistetes, Actinobacteria and Bacteroidetes (Fig. 1; Table S1). One patient (# 15) exhibited equal abundance of Synergistetes and Proteobacteria, while another patient (# 13) exhibited Actinobacteria and Proteobacteria respectively (Fig. 1; Table S1). Thus, overall, in all 15 samples Proteobacteria and Firmicutes phyla dominated (Fig. 1).

Fig. 1.

The cholesteatoma-associated microbiome abundantly constitutes bacteria from Proteobacteria and Firmicutes phyla. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate top 5 bacterial phyla

Upon analysing Proteobacteria for its classes, six (among the nine) patients (i.e. patients # 2, 3, 5, 6, 8, 14) and patient # 15 revealed Gamma (γ) Proteobacteria as their dominant class while two patients (# 7 and 11) exhibited Epsilon (ε) Proteobacteria as the most abundant class (Fig. 2A). The remaining patient (#1) sample had abundant Alpha (α) Proteobacteria (Fig. 2A). Delta (δ) Proteobacteria were barely present in any of the samples. Similarly, upon further analyses of four Firmicutes dominant COM-samples (patient # 4,9,10,12) for classes, Tissierellia stood out as the most abundant (40–60%) class in all of them, followed by Clostridia (up to 20%) and Bacilli (< 10%; Fig. 2B, Table S1) classes. Bacilli was upto 40% abundance only in one patient (# 4; Fig. 2B). Class Negativicutes was barely present in any of the samples. Among the four patients who showed abundance of Firmicutes over Proteobacteria (Fig. 1; # 4, 9, 10, 12), two also exhibited small amounts of γ-Proteobacteria (#9 and 10), one each exhibited β- (# 4) and ε- Proteobacteria (# 12) (Fig. 2A). Similarly, among the patients (nine - # 1–3, 5–8, 11 and 14) that showed abundance of Proteobacteria over Firmicutes (Fig. 1), three (# 7, 8, and 11), one (# 2), and five patients (# 1,3,5,6, and 14) exhibited small amounts of Tissierellia, Clostridia and Bacilli (Fig. 2B) genera. Patients (# 2, 7, 8, and 11) exhibited mixtures of Tissierellia, Clostridia and Bacilli (Fig. 2B).

Fig. 2.

(a) The cholesteatoma-associated microbiome is abundant with γ- and ε- Proteobacterial classes. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate top 5 classes of Proteobacteria. (b) The cholesteatoma-associated microbiome is abundant with Tissierellia and Clostridia classes of Firmicutes. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate top 5 classes of Firmicutes

In the seven patients who exhibited abundant Gamma (γ) Proteobacteria, (Fig. 2A), the abundant families included either Pseudomonadaceae (three patients - # 3, 5, 15) or Enterobacteriaceae (four patients - # 2, 6, 8, 14; Fig. 3A). The class epsilonproteobacteria showed an almost exclusive presence of Campylobacteraceae across four samples (patient # 7, 11–13; Fig. 3B). The most common genera of Pseudomonadaceae and Enterobacteriaceae families were Pseudomonas (Fig. 4A) and Klebsiella respectively (Fig. 4B). Interestingly, patients # 8 and 14 exhibited more than one genus of Enterobacteriaceae (Fig. 4B).

Fig. 3.

(a) Pseudomonadaceae and Enterobacteriaceae families dominantly represent γ-Proteobacteria class in the cholesteatoma-associated microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate the γ-Proteobacteria families. (b) Campylobacteraceae family dominantly represents ε-Proteobacteria class in the cholesteatoma-associated microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate the ε-Proteobacteria families

Fig. 4.

(a) The genera Pseudomonas dominantly represents Pseudomonadaceae in the cholesteatoma-associated microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate major genera of Pseudomonadaceae. (b) Genera Klebsiella and Escherichia dominantly represent Enterobacteriaceae in the cholesteatoma-associated microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate major genera of Enterobacteriaceae

Similarly, when we analysed classes Campylobacteraceae and Helicobacteraceae, the most abundant genera in the samples turned to be Campylobacter (Fig. 5A) and Helicobacter (Fig. 5B). Where genetic signatures of Firmicutes phyla were present, Peptoniphilaceae family was predominant (Fig. 6). Except in one patient (# 4), in general Staphylococcaceae family was present in traces (Fig. 6). The most common genera amongst Peptoniphilaceae were Parvimonas, Anaerococcus and Peptoniphilus (in decreasing order) (Fig. 7A), while for Staphylococcaceae, it was Staphylococcus (Fig. 7B). Patients numbered 7, 8, 9, 10, 11, 12 and 13 exhibited more than one genus of Peptoniphilaceae especially when predominated by Anerococcus or Peptoniphilus (Fig. 7B).

Fig. 5.

(a) Genera Campylobacter dominantly represents Campylobacteraceae in the cholesteatoma-associated microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate major genera of Campylobacteraceae. (b) Genera Helicobacter and Sulfuricurvum dominantly represent Helicobacteraceae in the cholesteatoma-associated microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate major genera of Helicobacteraceae

Fig. 6.

Peptoniphilaceae family dominantly represents Firmicutes in the cholesteatoma-associated microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate the Firmicutes families

Fig. 7.

(a) Genera Parvimonas and Anaerococcus dominantly represent Petoniphilaceae in the cholesteatoma microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate major genera of Petoniphilaceae. (b) Genera Staphylococcus dominantly represents Staphylococcaceae in the cholesteatoma-associated microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples (X-axis), were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized to relative abundance and plotted on Y-axis. Different colored plots indicate major genera of Staphylococcaceae

The phyla distribution was varied in the three patients with complications, i.e., patient number 3 (stage III), 8 and 12 (stage IV). While patient number 3 primarily harboured Proteobacteria, patient number 8 harboured Proteobacteria, Firmicutes and Bacteroidetes and patient number 12 had Firmicutes and Bacteroidetes. For patients aged < 18 years, the most found phyla were Proteobacteria Firmicutes and Bacteroidetes, like the most found phyla in adults. Actinobacteria and Synergistetes were primarily found in adult patients only.

We have attached a detailed supplementary file (Table S1) which details the individual results for each patient, in Phylum-class-order-family-genera-species format (in separate sheets), citing the percentage abundance in each category for all patients.

Discussion

Despite COM being a major contributing factor towards hearing impairment, metagenomics analysis of biopsy samples of COM with cholesteatoma has not been reported thus far in the Indian subcontinent. Given the prominence of bacterial pathogens getting cultured in most of the COM samples that we routinely diagnose and treat, employing metagenomics on a small sample set (n = 15), we not only cultured the 15 biopsy samples (routine practice), but also profiled and determined the genetic diversity and relative abundance of bacterial genera in small portion of each sample (Table S1). Our immediate future objective was to go beyond the boundaries of culturing of the biopsy tissues to obtain a true sense of bacterial diversity and the treatment challenges that the compositions might impose on our treatment strategies and regimens.

Our observations in the COMs samples indicate the following: (i) COMs with cholesteatoma exhibit polymicrobial association; (ii) Abundant genome signatures in COM specimens are from γ- and ε-Proteobacteria and Tissierellia, and Clostridia classes of Firmicutes (Figs. 1 and 2; Table S1); (iii) Prominent Proteobacteria include Pseudomonas, Klebsiella, Campylobacter, Escherichia and Salmonella (in decreasing order of abundance; Fig. 8A; Table 4); and (iv) Prominent Firmicutes include Parvimonas, Anaerococcus, Peptoniphilus, Staphylococcus and Finegoldia (in decreasing order of abundance; Fig. 8B; Table 5). Our observations nicely align with reported literature from other parts of the globe that state that Pseudomonas, Staphylococcus, Escherichia, Proteus and few anaerobes as most abundant genera in cholesteatoma samples [4–7]. Most of these microbes are potent biofilm producers, which enhances their ability to develop resistance to antibiotics and cause chronic infections [14–17]. However, it remains unclear whether these organisms are a causative factor of the disease or they simply thrive in COM-type diseased environment, making it a ‘chicken and egg’ question. Our work too cannot assign causality to these microbes, but opens avenues for further research in this regard, specially from the Indian subcontinent and other areas where this disease has high prevalence. We are cautious in not ignoring the possible role of host immune responses and previous antibiotic usage in altering the COMs-associated microbiome.

Fig. 8.

Diversity and abundance of bacterial genera from Proteobacteria (A) and Firmicutes (B) in cholesteatoma-associated microbiome. Total DNA extracted from 15 cholesteatoma patients’ samples were analyzed for bacterial diversity and abundance by metagenomic sequencing with nanopore technology. Raw data was processed using Porechop⁽¹⁰⁾ and centrifuge⁽¹¹⁾ tools. Comparative data obtained was normalized and different genera (X-axis) plotted against relative abundance in Log₁₀ scale on Y-axis. Different colored scatter bar plots indicate different genera. Each circle/square/triangle/diamond indicate each patient in the clinical cohort

Table 4.

The most abundant species of the phylum Proteobacteria, found collectively in the 15 samples, listed under their respective genus

Pseudomonas	Azotobacter	Klebsiella	Escherichia	Salmonella	Campylobacter
Pseudomonas aeruginosa	Azotobacter vinelandii	Klebsiella pneumoniae	Escherichia coli	Salmonella enterica	Campylobacter concisus
Pseudomonas aeruginosa group	Azotobacter chroococcum				Campylobacter ureolyticus
Pseudomonas alcaligenes					Campylobacter curvus
Pseudomonas chlororaphis					Campylobacter gracilis
Pseudomonas chlororaphis group					Campylobacter pinnipediorum
Pseudomonas citronellolis					Campylobacter sputorum
Pseudomonas furukawaii
Pseudomonas koreensis
Pseudomonas mosselii
Pseudomonas putida
Pseudomonas putida group
Pseudomonas sp. SWI6
Pseudomonas stutzeri
Pseudomonas stutzeri group
Pseudomonas stutzeri subgroup

Table 5.

The most abundant species of the phylum Firmicutes, found collectively in the 15 samples, listed under their respective genus

Parvimonas	Anaerococcus	Peptoniphilus	Staphylococcus
Parvimonas micra	Anaerococcus mediterranensis	Peptoniphilus sp. ING2-D1G	Staphylococcus aureus
	Anaerococcus prevotii		Staphylococcus epidermidis
			Staphylococcus xylosus
			Staphylococcus lugdunensis
			Staphylococcus agnetis
			Staphylococcus argenteus
			Staphylococcus capitis
			Staphylococcus hominis
			Staphylococcus sp. AntiMn-1
			Staphylococcus haemolyticus
			Staphylococcus cohnii
			Staphylococcus simiae
			Staphylococcus simulans

Unlike metagenomic analyses, upon direct culturing of a small portion of the COMs samples, in atleast 50% of them, we observed growth of Pseudomonas aeruginosa (Table 3). A third of the remaining samples exhibited more than one type of G-ve bacterium in their cultures. Only one sample was Staphylococcus aureus positive (Table 3). In contrast, by metagenomic analyses, in four patients we observed Klebsiella sp., three had Escherichia sp., four had Campylobacter sp., and eight had Helicobacter sp. Thus, our observations clearly indicate the severe limitation of direct culturing and potential hierarchy in the survival of specific pathogens in the media that supports growth of bacteria. Such direct culturing is almost a norm in most LMICs that have limited funding and almost no access for day-to-day metagenomic analyses. Stretching metagenomic analyses to detailed genetic profiling of antibiotics resistome perhaps will aid in targeted and personalized therapeutics regimen post-operatively in cases of complications. Though, we understand that parallel processing of all clinical samples for both in vitro culturing and metagenomic analyses routinely should be the way ahead and it would provide us a more educated handle for administering a personalized antibiotic treatment to these patients, at this juncture, with no ‘in house’ metagenomic lab and NGS costs still being high, it is beyond our reach at this moment.

Though limited by our samples, we did not observe any significant difference in the bacterial profile of the patients with complicated disease (patient #3,8,12) when compared to the patients with relatively less extensive disease (rest of patients). All our patients with complicated disease were aged < = 18 years. Lima et al., also reported the more aggressive pattern of cholesteatoma in patients aged < 16 years [18].

Our study has few limitations. One, we did not include healthy human ears as controls in our study. Healthy ears might be harbouring several (if not all) of the identified bacteria as commensals which is not possible to evaluate without recruiting controls. Taking middle ear samples from healthy controls would have raised ethical concerns and was thus, avoided. Second, we had a small sample size due to limited research funds.

Conclusion

Employing metagenomics clearly provides a comprehensive view of the microbiome associated with chronic diseases like cholesteatoma. Cholesteatoma is primarily associated with Proteobacteria and Firmicutes phyla, even in complicated disease. There is little difference in the paediatric and adult population. Further studies with a larger sample size targeting a more detailed phylogenetic picture with age-matched controls are required to validate our findings, especially from the subcontinent. Such studies may provide further insights into the associated-microbiota and possible interventions/treatments especially when empirical treatment seems the current norm.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author Contributions

Conceptualization: Kapil Sikka, Krishnamohan Atmakuri, Rabia Monga. Data curation: Narayana Sudhao Rao, Wagh Vaibhav Prakash, Vyas Meenal. Formal analysis: Anupam Kanodia, Krishnamohan Atmakuri, Mohd Ilyas, Yash Verma. Investigation: Narayana Sudhao Rao, Wagh Vaibhav Prakash, Vyas Meenal. Methodology: Kapil Sikka, Narayana Sudhao Rao, Krishnamohan Atmakuri. Project administration: Kapil Sikka, Krishnamohan Atmakuri. Resources: Kapil Sikka, Krishnamohan Atmakuri, Rabia Monga. Software: Anupam Kanodia, Mohd Ilyas, Narayana Sudhao Rao. Supervision: Krishnamohan Atmakuri, Kapil Sikka, Narayana Sudhao Rao. Validation: Narayana Sudhao Rao. Visualization: Kapil Sikka, Krishnamohan Atmakuri. Writing-original draft: Anupam Kanodia, Krishnamohan Atmakuri. Writing-review & editing: Anupam Kanodia, Krishnamohan Atmakuri, Kapil Sikka. Approval of final manuscript: All authors.

Funding

AIIMS-ICMR intramural funding.

Declarations

Disclosures

None.

Conflict of Interest

The authors declare that they have no conflict of interest to disclose.

Consent

An informed written consent was obtained from all participating patients included in the study.

Ethics Approval

The study was approved by the institutional ethics committee (ref. ID: IEC-621/02.11.2018, RP-35/2018). All procedures performed in studies involving human participants were in accordance with the ethical standards of the institution and with the 1964 Helsinki declaration and its later amendments (latest 2013) or comparable ethical standards.