The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of coronavirus disease 19 (COVID-19), continues to pose a global threat due to its ability to evade the immune system through genetic mutations. Despite the pandemic slowdown, immunized individuals remain susceptible to SARS-CoV-2, which is believed to keep circulating for many years (Kissler et al., 2020; Rossman et al., 2021; Barreiro and San Román, 2022).

In COVID-19, both immune suppression and runaway inflammation have been observed and can result in more severe disease, while protective immunity consists of the induction of both humoral and cell-mediated responses (Merenstein et al., 2022). The SARS-CoV-2 inhibits Interferons I and III (IFN-I and IFN-III) and pro-inflammatory responses, promoting high viral replication in the respiratory tract (Wei et al., 2020). Paradoxically, this immune suppression leads to increased circulating cytokines in COVID-19-associated pneumonia. Severe lung damage from viral replication causes epithelial barrier breakdown, triggering a ‘cytokine storm’ with both pro- and anti-inflammatory actions (Tisoncik et al., 2012; Chen et al., 2020; Hall et al., 2022), which impair lung function promoting immunoparalysis, resulting in irreversible damage (Wong et al., 2019).

In the upper respiratory tract (URT), which comprises the nasal cavity, nasopharynx, and oropharynx, high bacterial densities are found (10^3–10⁶ U⁻¹), while in the lung, which belongs to the lower respiratory tract (LRT), bacterial densities decrease (~10² U⁻¹). Genera characteristic of a healthy URT include Staphylococcus sp., Propionibacterium sp., Leptotrichia sp., Rothia sp., Dolosigranulum sp., Haemophilus sp., Moraxella sp., Veillonella sp., Prevotella sp., Streptococcus sp. and Corynebacterium sp (Man et al., 2017; Wypych et al., 2019). In eubiosis, these microorganisms act as ‘guardians of respiratory health’ (Wong et al., 2019; Wypych et al., 2019), avoiding colonization of the URT by pathogens, including respiratory viruses, suggesting a potential role in resisting SARS-CoV-2 infection (Bogaert et al., 2004; Wang et al., 2022).

The interactions between the microbiome and its host extend to crucial functions such as carbohydrate and protein decomposition, nutrient absorption, vitamin biosynthesis, and modulation of the immune system (Rooks and Garrett, 2016; Zheng et al., 2020). Metabolites and components from these interactions can modulate immune cells through various mechanisms, including signaling of mucosal-associated T cells (MAIT) (Smith et al., 2013; Godfrey et al., 2019; Legoux et al., 2020;, IFN-I (Flerlage et al., 2021), and dendritic cells (Schaupp et al., 2020) beyond the anti-inflammatory activity of Short-Chain Fatty Acids (SCFA) (Fukuda et al., 2011) and regulation of immunoglobulin expression (Corbett et al., 2014; Kim et al., 2016; Dalile et al., 2019; Parrot et al., 2020; Wang et al., 2022).

Recent investigations have revealed substantial dysbiosis in both intestinal and respiratory microbiomes during COVID-19, and their specific compositions are correlated with the severity of the disease and its outcomes (Ferreira et al., 2020; Lloréns-Rico et al., 2021). Additionally, SARS-CoV-2 infection has been found to heighten the susceptibility of patients to secondary pathogens, thereby playing a significant role in exacerbating morbidity and mortality associated with COVID-19 (Cox et al., 2020; Yeoh et al., 2021). Besides, the microbiome has been identified as a modulator of immune responses and diseases in the airways, increasing interest in the interaction between COVID-19 and this microbiome (Merenstein et al., 2022). This study aimed to delve into these interactions at a functional level, identifying taxa and pathways in the URT associated with COVID-19 severity.

Among the 88 selected samples, nine were excluded due to the impossibility of extracting high-quality DNA or unsatisfactory sequencing results. The samples and their respective experimental groups that remained for the analysis were M-COV, n = 22; NO-COV, n = 19; S-COV, n = 20; and NC, n = 18. Data Sheet 1 presents a flowchart representing experimental design and the bioinformatics workflow.

The quality check as well as the error rates analysis from the preprocessing and the complete track of the reads are shown in Supplementary Material ( Data Sheet 2 , Table 1 - Supplementary Data 1 ). After the filtering step ( Data Sheet 3- 0.1 Data preparation), 11 phyla, 200 genera, 364 species, and 50,347 ASVs were identified. The complete description, representative proportions, and overall relative abundance of these taxa are presented in Supplementary Material ( Table 1 - Supplementary Data 2 - Table 1 - Supplementary Data 5 ).

Our study evaluated the upper respiratory tract microbiome in COVID-19 patients, revealing that severe cases (S-CoV) exhibit a statistically significant reduction in α-diversity compared to mild cases (M-CoV) and non-COVID patients (NC). This observation is consistent with previous reports linking reduced diversity to increased COVID-19 severity (Mostafa et al., 2020; Soffritti et al., 2021; Ke et al., 2022). Although β-diversity did not show significant differences - likely due to confounding factors such as antibiotic use and comorbidities (Maes et al.; Soffritti et al.; Nardelli et al., 2021) - the overall disturbance in microbial homeostasis appears to be associated with a more severe disease phenotype.

The URT microbiome was dominated by genera such as Streptococcus sp., Prevotella sp., Staphylococcus sp., Propionibacterium sp., Leptotrichia sp., Rothia sp., Dolosigranulum sp., Haemophilus sp., Moraxella sp., Veillonella sp., and Corynebacterium sp (Man et al., 2017; Wypych et al., 2019). Among these, certain opportunistic species - such as S. aureus and H. influenzae - are well-known contributors to pneumonia in critically ill COVID-19 patients (Bai et al., 2022). Notably, Abiotrophia sp. was enriched exclusively in the S-CoV group, which is positively correlated with COVID-19 severity (Hernández-Terán et al., 2021; Soffritti et al., 2021). Additionally, taxa like Saccharibacteria (specifically, Saccharibacteria_[(TM7]_(G1), (G3) and (G6))) were found in higher abundance in NC and paucisymptomatic patients, whereas ‘non-Saccharibacteria’ underrepresented taxa such as Mesorhizobium sp., Oribacterium sp., and Dolosigranulum sp. were reduced in S-CoV and associated with inflammatory biomarkers, suggesting a protective effect (Bosch et al., 2016; Brugger et al., 2016).

A group of beneficial taxa - particularly members of the Lachnospiraceae family (e.g., Catonella sp., Lachnoanaerobaculum sp., Butyrivibrio sp.) - were enriched in the NC group and inversely related to disease severity. These butyrate-producing bacteria have recognized roles in immune modulation and anti-inflammatory activity (Al Bataineh et al., 2021), suggesting that their depletion in S-CoV could impair protective mechanisms against SARS-CoV-2 infection. Similarly, Neisseria sp. - which may enhance anti-SARS-CoV-2 antibody production through interactions with T- and B-cells (Hung and Christodoulides, 2013; Santana-Mederos et al., 2022) - displayed a reduced differential abundance in S-CoV, further linking microbial composition to immune functionality.

In parallel with taxonomic alterations, our metabolic predictions revealed significant changes in key pathways that appear to reflect the functional impact of microbial dysbiosis in COVID-19. One set of interrelated findings involves the modulation of steroid hormone synthesis and vitamin metabolism. The S-CoV group showed a reduced abundance of pathways involved in steroid hormone synthesis (including androstenedione degradation), a process regulated by ACTH and the renin-angiotensin-aldosterone system (RAAS) that is linked to COVID-19 severity (Pal, 2020; Sezer et al., 2022). This observation is paralleled by the depletion of eubiotic taxa from the order Corynebacteriales (related to vitamin B12 biosynthesis) and the family Propionibacteriaceae (linked to vitamin B12, vitamin K biosynthesis, and SCFA metabolism) in S-CoV. In contrast, taxa such as Novosphingobium sp. were enriched only in M-CoV and correlated with critical inflammatory biomarkers, suggesting their possible role in mitigating severe inflammation.

Another functionally connected group of pathways relates to mucin metabolism and immune modulation. Several genera within the Peptostreptococcaceae family, known for cleaving and transporting mucin-associated monosaccharides (Flynn et al., 2016), were reduced in S-CoV and NC compared to M-CoV. This reduction aligns with observed decreases in the superpathway of fucose and rhamnose degradation—processes that influence mucin O-glycosylation and subsequent IgG Fc fucosylation (Bennett et al., 2012; Flynn et al., 2016; Wlodarska et al., 2017; Wang, 2019; Chakraborty et al., 2021). Such alterations may disrupt immune system activation and inflammatory signaling, further exacerbating disease severity.

Further, metabolic pathways involved in glycosaminoglycan (GAG) biosynthesis - evidenced by decreased superpathway of glucose and xylose degradation - may enhance the availability of xylose for GAG synthesis. This is of particular interest given that both Enterococcus sp. and Lactobacillus sp. (which carry glycosaminoglycan genetic clusters; (Kawai et al., 2018)) were enriched in S-CoV, suggesting a potential role in viral attachment via the S protein and consequent lung inflammation (Cagno et al., 2019; Kim et al., 2020).

Additional metabolic disruptions include reductions in pathways related to aromatic compound degradation, purine nucleobase metabolism, and amino acid degradation (e.g., L-tryptophan, L-glutamate, and L-lysine). These changes have been implicated in altered inflammatory responses and immune regulation (Chakraborty et al., 2021; Wu et al., 2021; Casas-Sanchez et al., 2022; Kasahara et al., 2023). The diminished capacity for polyamine and arginine metabolism, which are linked to enhanced IgA responses and vascular function (Durante, 2013, Durante, 2022; Rodriguez et al., 2017; Falck-Jones et al., 2021; Martí I Líndez and Reith, 2021; Zhao et al., 2023), further underscores the potential for metabolic impairment to contribute to severe COVID-19 outcomes.

Moreover, the decreased predicted abundance of pathways related to the biosynthesis of key microbial-derived components, such as vitamin B12 (Tamura et al., 1999; Yoshii et al., 2019) and vitamin K (Ragab et al., 2020; Kudelko et al., 2021), as well as SCFA metabolism (Dalile et al., 2019; Ashique et al., 2022; Batista et al., 2022), suggests that the loss of these beneficial metabolites may impair immune regulation and anti-inflammatory responses. This is reinforced by the observed reduction in taxa associated with SCFA production, including members of Lachnospiraceae and Propionibacteriaceae.

Lastly, the observed decreases in lipopolysaccharide (LPS) biosynthesis pathways (Maldonado et al., 2016; Bradley et al., 2022), in NAD metabolism (Zheng et al., 2022) and tRNA processing (Katanski et al., 2022) further indicate that key inflammatory and metabolic responses are disrupted in severe COVID-19.

Collectively, the close interrelation between specific taxa and their metabolic functions suggests that dysbiosis in the URT may directly impact host immune responses. The depletion of beneficial microbes such as Dolosigranulum sp. and butyrate-producing Lachnospiraceae, along with impaired vitamin and SCFA biosynthesis, could compromise mucosal immunity and facilitate an exaggerated inflammatory response. In addition, alterations in steroid hormone synthesis pathways, mucin glycosylation, and polyamine metabolism likely contribute to the cytokine dysregulation observed in severe COVID-19. This complex network of taxon - metabolic interactions underscores the potential of the URT microbiome as a modifiable factor in the progression of COVID-19.

This study demonstrates that the primary differences in URT microbiome composition and functionality are associated with COVID-19 severity rather than with SARS-CoV-2 infection. Severe COVID-19 is characterized by a significant reduction in α-diversity, a depletion of protective eubiotic taxa, and widespread alterations in metabolic pathways essential for immune modulation and host homeostasis. Despite inherent limitations (cross-sectional design, not controlling for confounding factors as age, antibiotic use, comorbidities), these findings suggest that specific microbial taxa – such as Lachnospiraceae, Propionibacteriaceae, Dolosigranulum sp. - and their metabolites - such as SCFAs, vitamins B12 and K, and key amino acids - could serve as promising targets for therapeutic intervention. Despite inherent limitations, our results provide a foundation for future studies aimed at restoring microbial homeostasis to mitigate severe respiratory outcomes.