3.4 Fungal microbiome: taxonomic analysis
At the phylum level, Ascomycota (mean abundance in VKC 79.2%; HC 71.9%) and Basidiomycota (VKC 19.4%; HC 27.4%) were the two dominant phyla accounting for > 98% of all sequences in both groups. At the family level, Saccharomycetaceae (VKC 55.8%; HC 62.7%), Malasseziaceae (VKC 15.6%; HC 3.3%),Pleosporaceae (VKC 12.4%; HC 2.6%) and Cladosporiaceae(VKC 4.0%; HC 3.0%) accounted for the vast majority of sequences (Figure 6). OTUs referred to Malasseziaceae were significantly increased in VKC children compared to HC (W=42).Saccharomycetaceae, Malasseziaceae , and Dipodascaceae were present in all subjects, thus constituting the fungal core microbiome of both HC and VKC patients. At the genus level, Saccharomyces (VKC 55.8%; HC 62.7%), Malassezia (VKC 15.6%; HC 3.3%),Alternaria (VKC 9.9%; HC 2.4%) and Cladosporidium (VKC 4.0%; HC 3.0%) were the prevalent genera.
Discussion
The application of high-throughput sequencing (HTS) techniques to metagenomics has drastically contributed to reveal a level of microbial diversity that was previously hidden by the selectivity of cultivable based methods (5). In our study, VKC patient showed a significantly higher bacterial α diversity than HC. Similar results have been already observed in patients with both ocular and extra-ocular inflammatory diseases (14, 33-35). Differently from bacteria, no significant difference in fungal α diversity metrics was observed between VKC and HC. Alpha diversity metrics were also evaluated according to other clinical variables showing that both bacterial and fungal richness and diversity were significantly higher in IgE-negative than IgE-positive VKC suggesting that atopic patients may have a different conjunctival microbiome. There is an accumulating evidence that dysbiosis precedes the development of allergic manifestations at least for food allergy and asthma (36). Interestingly, in our study children who had been formula-fed showed a significantly higher bacterial α diversity than those who had been breastfed as well as children with and without history of atopic dermatitis during the first year of life. Results of the 10-item questionnaire revealed that breastfeeding had been significantly more common among healthy children than VKC patients. Breastfeeding has been shown to be associated with a reduced risk of childhood asthma, atopic dermatitis and rhinitis (37). It has been suggested that breastfeeding shapes the early-life gut microbiota directly by transfer of the human milk microbiota (including lactobacilli and bifidobacteria) and indirectly by exposure to human milk oligosaccharides (HMOs), short-chain fatty acids (SCFAs), secretory IgA and components of the innate immune system which collectively influence and direct microbial growth and metabolism (36, 38, 39). If this can exert a remote and persistent influence on the composition of a local microbiota is unknown.
When exploring the effect of sex on conjunctival microbiome, we found that males had a significantly higher number of observed fungal OTUs than females. The preponderance of VKC in males suggested a potential role of sex hormones in its pathogenesis (40, 41), but a possible role of microbiome on the male prevalence in VKC hasn’t been highlighted. Although the microbiome is known to be affected by age and sex, the influence of these factors on ocular surface microbiota remains controversial (9, 42, 43).
A normal conjunctival microbiome could be described by a core microbiome that contributes to maintain the functional stability and homeostasis of the ocular surface. It has been proposed the existence of a variable ocular surface microbiome, indicating that certain microbial types may always be present, while others are transient depending on age, sex, ethnicity, environment, lifestyle or food habits as well as tearing and blinking(6, 9). The temporal stability of ocular microbiota over a three months period has been assessed showing that, even though the microbial community was relatively stable, the ocular surface is not colonized consistently by any given OTU (6). We tried to define a core microbiome including in this definition all those taxa who were present with a minimum abundance >0.01% in at least 80% of samples.Proteobacteria, Firmicutes and Actinobacteria were identified as the 3 major bacterial phyla in all samples, while Pseudomonassp., Staphylococcus sp., Streptococcus sp. andPropionibacterium sp. were ubiquitous among all the examined subjects at the genus level. All these core taxa are well-known potential ocular pathogens suggesting that the ocular surface is regularly exposed to opportunistic pathogens but the ocular surface can activate mechanisms suppressing microbial pathogenicity. Many other bacterial families (Neisseriaceae, Fusobacteriaceae ,Moraxellaceae, Micrococcaceae, Gemellaceae, Paenibacillaceae , andVeilonnelaceae ) and genera (Prevotella sp., Bacillus sp., Rothia sp., Haemophilus sp. ) met the core microbiome’s definition criteria only in the VKC group. The VKC core microbiome included different species of gram-negative bacteria such as Proteobacteria (Neisseria sp., Moraxellasp., Haemophilus sp. ) and Bacteriodetes(Prevotella sp.) able to induce an LPS-driven inflammatory response through the Toll-Like Receptor (TLR)4 /NF-κB pathway that may trigger the inflammation in these patients (24).
Differently from bacteria, the fungal core microbiome was similar in HC and VKC patients and included Saccharomycetaceae, Malasseziaceae , and Dipodascaceae. Interestingly, a significantly higher abundance of Malasseziaceae was found in VKC patients.Malassezia spp. have already been associated with many different disorders including bacterial keratitis (44, 45), dandruff, atopic dermatitis (46) and inflammatory bowel diseases (47). Atopic dermatitis patients are often sensitized to Malassezia with Th2- and IgE-specific responses or a mixed Th2/Th17 response, which is a possible mechanism also in VKC (24, 48). Malassezia spp.directly interacts with several PRRs, such as Dectin-1, Dectin-2, Mincle, TLR2 and TLR4 activating multiple signaling pathways (49-51). All these PRRs are over-expressed at gene and protein level in VKC patients (24). Therefore, we believe that glycan, phospholipid carbohydrate residues of allergens and microbes, can engage innate receptors on epithelial and dendritic cells priming a Th2/Th17 type response in VKC.
Fourteen of the 22 VKC children (64%) were under one or more topical treatments (cyclosporine A, antihistamines and/or mast cells stabilizers) at the time of swabbing, creating a potential bias in microbiome’s analysis (Table 1). However, only 5 of the 10 VKC patients whose swabs were not sequenced because of the absence of the amplicons, were treated with topical medications. Furthermore, β diversity metrics didn’t showed any significant difference of both bacterial and fungal communities according to the use and type of topical medications (data not shown) suggesting that factors other than topical eyedrops may alter the conjunctival microbiota. Similar results have been previously reported showing that topical cyclosporine didn’t affect the richness and the diversity of the ocular bacterial flora (9, 12).
The main limitation of our study is the low number of patients and sequenced samples. VKC is a relatively rare disease in Western Europe with a prevalence of 1.1-10.5 cases per 10,000 inhabitants (52). On the other hand, in order to minimize the possible biases of age, gender and other variables, the two study groups were as homogeneous as possible. Only 4 out of 20 conjunctival swabs from healthy controls yielded 16S rRNA and ITS2 amplicons and allowed for the microbiome’s analysis; on the contrary, a much higher proportion of swabs from VKC patients yielded sufficient genetic material to be analyzed. Even though this may represent the main concern, the minimum more common recommended sampling size to make a comparison between a pathological sample and normal samples is three to four, therefore sufficient in this case considering the very low microbial load in the healthy conjunctiva. A higher difficulty to obtain amplicon products from healthy subjects has been already described and attributed to the physiological antimicrobial activity and to the lower microbial load of healthy subjects (6, 53).
In conclusion, we showed that VKC patients have a different and more diverse ocular microbiota compared to normal subjects. The major challenge in microbiome studies is to translate metagenomic data into biological knowledge in order to understand how the microbiome may affect the host or vice versa. The comprehension of the relations between dysbiosis and human diseases will pave the way to more focused studies and new therapies.
Acknowledgments
a. Funding/Support: Supported by MIUR DOR1952345/19 and DOR2092417/20 from Italian Institute of Health.
b. Conflicts of Interest:
Andrea Leonardi: No Conflicts of Interest.
Rocco Luigi Modugno: No Conflicts of Interest
Fabiano Cavarzeran: No Conflicts of Interest
Umberto Rosani: No Conflicts of Interest
c. Contributions to Authors in each of these areas: Andrea Leonardi (AL), Rocco Luigi Modugno (RLM), Fabiano Cavarzeran (FC), Umberto Rosani (UR)
Conception and Design: AL, UR
Analysis and interpretation: AL, RLM, UR
Writing the article: AL, RLM
Critical revision of the article: UR
Final approval of the article: AL, RLM, UR
Data Collection: RLM, UR, FC
Provision of materials, patients, or resources: AL, UR
Statistical expertise: FC
Obtaining funding: AL
Literature search: RLM, AL, UR
Administrative, technical or logistic support: AL, FC
d. Statement about Conformity with Author Information: none
e. Other Acknowledgments: none