RNA Sequencing and Gene Ontology Analysis in Acute Invasive Fungal Sinusitis

Introduction

Acute invasive fungal sinusitis (AIFS) remains an aggressive disease with high morbidity and mortality. ¹ Immunocompromised patients, such as those with poorly controlled diabetes or undergoing chemotherapy, are at risk for AIFS. While a large number of patients are immunocompromised, only a small percentage of these patients will develop AIFS. Host risk factors for development of disease systemically, and at the local mucosal level of tissue fungal invasion, remain unclear. Additionally, the relationship of AIFS to other forms of fungal sinusitis is not well understood. Sinonasal fungal elements are found in many patients, however fungal sinusitis is most common in a non-invasive form. ² Recent case series have suggested the possibility that fungal sinusitis could be a spectrum of disease.^3,4 The possibility of progression from non-invasive fungal sinusitis to invasive disease is concerning given our lack of understanding of specific risk factors for disease development.

Gene expression in AIFS patients has not been previously performed. Prior work in this area has focused on patients with chronic rhinosinusitis (CRS). Using RNA sequencing of sinonasal surgical samples, specific differences in gene expression of inflammatory markers have been identified between CRS and control patients. ⁵ Transcriptomic analysis in AIFS patients may similarly provide insight into pathophysiology of disease. These findings would lead to an understanding of deficient immune responses at the level of sinonasal mucosa in at-risk patients. To date, there are no published studies evaluating the transcriptome in the sinonasal tissues of AIFS patients. Therefore, the goal of this work was to compare surgical tissue samples from AIFS patients to control patients using bulk RNA sequencing techniques.

Methods

Results

Tissue samples were available from 4 patients with active AIFS diagnoses, 2 patients with recovered AIFS, 1 patient with a diagnosis of non-invasive fungal ball and 4 healthy controls. Patients with recovered AIFS were taken to the operating room for revision surgeries several months following active AIFS diagnosis, and no evidence of AIFS was seen on endoscopic exam, nor in pathology specimens. Healthy controls did not have fungal findings of any form on pathology examination. Patient characteristics and demographic information are shown in Table 1.

OPEN IN VIEWER

Table 1.

Demographic Patient Information.

Age	Gender	Diagnosis	Comorbidities	Fungal pathogen
35	M	AIFS	B-cell lymphoma	Rhizomucor
47	F	AIFS	Uncontrolled diabetes	Mucorales
58	F	Control	Sinusitis	n/a
71	M	Control	Epistaxis	n/a
34	M	AIFS	Pre-B ALL	Rhizopus microsporus
28	M	Control	Sinusitis	n/a
73	M	AIFS recovered	Sinusitis	n/a
63	F	Control	CSF leak	n/a
43	M	Fungal ball	Sinusitis	Rhizopus
52	M	AIFS	Uncontrolled Diabetes	Rhizopus arrhizus
48	F	AIFS recovered	Diabetes	n/a

In total, 255 genes were identified as either up-regulated or down-regulated in AIFS patients as compared to control patients at a statistically significant level of p < 0.05 (Figure 1). The top five most significantly AIFS enriched genes (NXF2, CCL18, PTPRN, ATP6V1G2-DDX39B, INMT-FMA188B) and Control enriched genes (C6orf58, PRB3, SCGB2A2, PRB4, PRR4) are labeled. A list of all genes in the volcano plot are found in Supplemental Data 1. Sequencing data is available through the National Center for Biotechnology Information Gene Expression Omnibus (Accession Number PRJNA882041). There were no statistically significant differences in other group comparisons. The heat map seen in Figure 2 visually illustrates the differences in patterns of gene expression changes over all four conditions. Each sample is organized into condition in columns, with each individual gene organized by row. The cell color indicates increased RNA expression (red) or decreased RNA expression (blue). Genes underwent k-means clustering, which identified three cluster patterns, and the top ten genes in each cluster are listed to the right. A full listing of genes are found in Supplemental Data 2. Specific GO biological processes that were identified as differentially enriched in AIFS patients as compared to controls included 1. GO:0007155 (cell adhesion), 2. GO:0030199 (collagen fibril organization) and 3. GO:0001525 (angiogenesis).

OPEN IN VIEWER

Figure 1.

Differential gene expression shown as a volcano plot where each dot represents a gene. The genes that were enriched in the Control group (light blue) and enriched in AIFS (red) are depicted. The top 5 differentially expressed genes for Controls and AIFS are labeled. The x-axis shows the normalized log₂ expression values while the y-axis shows the log-10 transformed false discovery rate (adjusted p-value). The top five differentially expressed genes in each group are highlighted. The vertical dotted line indicates log₂ fold change of 2, and the horizontal dotted line depicts FDR<0.05. See Supplemental File 1 for a full listing of genes.

OPEN IN VIEWER

Figure 2.

Heatmap of differentially expressed genes showing marker genes for each condition annotated at the top. Each row represents an individual gene. After k-means clustering, there were 3 clusters identified. The top 10 genes in each cluster are listed on the right. See Supplemental File 2 for a full listing of genes.

Discussion

To date, published work on AIFS has focused on clinical outcomes, prognostic factors and epidemiologic analyses.^8,9 The challenges of translational work in this field have led to only limited investigations of blood testing for detection of fungal antigens in the AIFS population, which have not yielded conclusive findings. ¹⁰ Sinonasal tissue studies have been lacking and represent an area of important future work for improved understanding of AIFS.

In this study, we have identified several genes which are differentially expressed in AIFS patients as compared to control patients. Additionally, we have identified differential expression of 3 GO biological processes which may contribute to defects in the sinonasal tissue cell membrane integrity of AIFS patients. This work suggests several future pathways for potential exploration of AIFS translational research. Specifically, the GO processes identified in this work support tissue membrane structural properties, which may aid in prevention of direct fungal extension. While the role of angiogenesis is unclear in AIFS, angiogenesis is known to be associated with immune cell properties. ¹¹ Tissue angiogenesis may be lacking in AIFS patients, further limiting systemic immune cell signaling and transport to affected tissues, allowing fungal invasion. Treatments which stimulate angiogenesis may be beneficial if directed towards the sinonasal mucosa in these patients. We believe that future investigations are critical to improving our understanding of the pathophysiology of fungal invasion and spread.

Work investigating non-sinus invasive fungal infections, such as pulmonary infections, have focused on the innate immune system and have identified pattern recognition receptors (PRRs), neutrophil, macrophage, natural killer and dendritic cell presence and function as critical to control of infection.^12–14 In our study, a highly significant down-regulated gene in AIFS was SCGB2A2, which encodes mammaglobin-A, a protein involved in cell-signaling and immune response. ¹⁵ CCL18, which is secreted from innate immune cells, and supports tolerance and immunosuppression, ¹⁶ was enriched in AIFS, possibly leading to susceptibility to disease. Another gene that was found to be enriched in the AIFS condition was CAPN14 (Figure 2). This gene has been shown to play an important regulatory role in epithelial structure. In eosinophilic esophagitis, higher levels of CAPN14 protease are associated with epithelial barrier impairment. ¹⁷ We also identified LMO4 as down-regulated in AIFS sinonasal tissue as compared to control tissue. This gene is responsible for macrophage and natural killer cell activation, important components of the innate immune system. ¹⁸ LMO4 has also been shown to also be important in epithelial proliferation, which is an additional component of the innate immune system. ¹⁹ Further investigation into the innate immune cellular response is indicated as targeted treatment for at-risk patients may involve stimulating specific aspects of the innate immune system.

Our work is limited by small sample size and the availability of tissue from patients with active AIFS diagnoses from a single institution. We are also limited by patients with Mucor infections, while Aspergillus is the most common pathogen responsible for AIFS in the severely immunosuppressed non-diabetic population. ⁸ Our data may not be applicable to other AIFS populations. More extensive analyses of recovered AIFS and fungal ball patients were also limited by small sample size. These analyses may be particularly interesting in future studies to determine the role of specific gene expression patterns in understanding recovery of disease and/or progression in the potential spectrum of non-invasive to invasive fungal sinus disease. Additional future directions include the use of single cell RNA sequencing to overcome the limitations of unbiased RNA sequencing, and mycobiome/microbiome analyses. Advances in this area of work may also be possible through a multi-institutional approach for this relatively rare, but potentially fatal, disease.

Conclusion

We have identified 255 genes and 3 GO biological processes which are differentially expressed in AIFS patients as compared to control patients. Future work will be focused on single cell RNA sequencing for a more in depth understanding of the findings presented here. We are also interested in investigating the differential expression between AIFS and other cohorts of patients, specifically those who are immunosuppressed, or with preexisting non-invasive forms of fungal sinusitis. This work will ultimately lead to treatments and prevention for AIFS.

Supplemental Material

Supplemental Material