2.1. Study population and sample collection

Infected and non-infected study groups were formed from the patients who applied to Acıbadem Health Group hospitals with suspected SARS-CoV-2 infection. Cases were diagnosed on the basis of the interim guidance of the World Health Organization (WHO) (World Health Organization; Geneva: 2019. Clinical Management of Severe Acute Respiratory Infection When Middle East Respiratory Syndrome Coronavirus (MERS-CoV) Infection is Suspected: Interim Guidance.) and diagnosis and treatment guidelines of COVID-19 in Turkey (COVID-19 (SARS-CoV-2 INFECTION) GUIDE Republic of Turkey, Ministry of Health, April 14th 2020, Ankara). Selected patients did not have any other known infections at the time of diagnosis. A total of 30 patients were enrolled in the study: 15 patients PCR positive for SARS-CoV-2 (mean age: 38.9±13.8) and 15 patients PCR negative for SARS-CoV-2 (mean age: 36.6±15.9). Infected group was equally distributed regarding gender. Most of the non-infected patients were men (60%). As a routine application of our clinical laboratory, WBC count of the patients were also performed (Neutrophil count x10⁹/L mean: 0.6±0.5 for non-infected and 5.6±10.2 for infected).

Samples were collected using a nasal and oropharyngeal (NUCLISWAB) swab (Salubris, Turkey) and were placed in a tube with Universal Transport Media. After collection the samples were stored at -80°C before proteomics analysis. No minors were included in the study and a written informed consent was obtained from each patient that was enrolled in the study. Ethical approval for the conduct of the study was given by Acibadem Mehmet Ali Aydinlar University Human Scientific and Ethical Review Committee (Approval ID: 2020-07/9).

2.2. COVID-19 RT-PCR test

For molecular testing of SARS-CoV-2; the extraction of nucleic acids from the samples was performed by a manual liquid phase method using Bio-Speedy Nucleic Acid Isolation Kit (Bioeksen, Turkey). Nucleic acid amplification test (NAT) was carried out by Bio-Speedy COVID-19 RT-qPCR Detection Kit (Bioeksen, Turkey) according to the manufacturer's instructions on RotorGene (Qiagen, Germany) Real-Time PCR instrument. The test briefly achieves a one-step reverse transcription (RT) and real-time PCR (qPCR) targeting SARS-CoV-2 specific RdRp (RNA-dependent RNA polymerase) gene fragment.

2.3. LC-MSMS analysis

A shorter sample preparation approach was applied to obtain the tryptic peptides for analyses. 50 ul of the naso-oropharyngeal transport solution was taken and lyophilized. The powder was reconstituted in 20ul of 50 mM Ammonium bicarbonate solution with 1 ul of protease inhibitor mixture (Thermo Scientific). The mixture was centrifuged at 14000 xg for 10 min and the supernatant was transferred to a clean Eppendorf tube. DTT was added to 10 mM final concentration and incubated at 60˚C for 10 min. The mixture was alkylated in dark for 30 min with 20mM IAA. To the resulting mixture 1 ug of sequencing grade trypsin (Promega Gold) was added and incubated at 55˚C for 1.5 hrs. The digest was acidified with 1 ul formic acid and transferred to an LC vial for injection. The samples were analyzed by the protocols in our previous studies [1]. Briefly, tryptic peptides were trapped on a Symmetry C18 (5μm,180μm i.d. × 20 mm) column and eluted with ACN gradient (4% to 40% ACN, 0.3 ul/min flow rate) with a total run time of 60 min on a CSH C18 (1.7 μm, 75 μm i.d. × 250 mm) analytical nano column. Data were collected in positive ion sensitivity mode using a novel data-independent acquisition mode (DIA) coined as SONAR [2] with a quadrupole transmission width of 24 Da. Progenesis-QIP (V.2.4 Waters) was used for data processing.

2.4. PPI network and pathway analysis

Protein-protein interaction networks functional enrichment analysis was carried out using the STRING (http://string-db.org/, v11.0) database with the highest confidence interaction score level to identify possible pathways related to the identified proteins. Textmining, experiments, databases co-expression, neighborhood, gene fusion, and co-occurrence selected as active interaction sources. The minimum interaction score was set to high (high confidence = 0.700). REACTOME (http://www.reactome.org) pathways analysis tool was also used for processing.

3.1. Label-free proteomics

Proteins were extracted from a small amount of naso-oropharyngeal samples, fast tryptic digestion was applied and followed with a 60 min reverse-phase separation. NanoLC-MSMS analysis provided the identification of 207 protein groups with high confidence (<1% FDR) (S1 and S2 Tables). Statistical analysis done in Progenesis QIP software identified 17 proteins to be statistically significantly expressed in patients’ naso-oropharyngeal samples (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Significantly altered protein identification list.

https://doi.org/10.1371/journal.pone.0240012.t001

3.2. Bioinformatic analysis

Pathway analysis of the significantly altered protein levels between COVID-19 positive and negative patients’ naso-oropharyngeal swab samples were analyzed using the STRING online database. The PPI network obtained, contained 17 differentially expressed proteins with 15 nodes (disconnected nodes were not shown) and 14 edges as shown in Fig 1A. The main cluster includes Neutrophil Elastase (ELANE), Azurocidin (AZU1), Myeloperoxidase (MPO), Myeloblastin (PRTN3), Cathepsin G (CTSG) and Transcobalamine-1 (TCN1). The abundance of these proteins were found to be increased in COVID-19 positive patient samples compared to negative ones. REACTOME pathway enrichment analyses of the differentially expressed proteins were performed. Nine of the proteins were primarily associated with the immune system pathway. Proteins clustered in the PPI network were mainly enriched in the neutrophil degranulation pathway (HSA-6798695, FDR = 2.01E-7) which is a subpathway of the innate immune system (Fig 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. STRING and Reactome analysis of identified proteins.

A-Protein-protein interaction network. B-Diagram of Initiate Immune System pathway (HSA-168249, FDR = 2.3E-5) from Reactome, showing a significant enrichment for neutrophil degranulation pathway (HSA-6798695, FDR = 2.01E-7) having the most hits.

https://doi.org/10.1371/journal.pone.0240012.g001

4.1. SARS-CoV-2 associated dysregulation in neutrophil degranulation

It is known that neutrophils are the first line of defense against the onset of a viral infection and they begin the process of defending against microorganisms by releasing antiviral enzymes and toxins stored in their granule. These azurophilic neutrophils undergo limited exocytosis when activated [5] and their primary role is believed to be killing and degradation of engulfed microbes in the phagolysosome [6]. Therefore our observation of upregulated proteins related to neutrophil degranulation in SARS-CoV-2 patients’ naso-oropharyngeal samples is not suprising. Although neutrophil degranulation having a positive impact on clearance of SARS-CoV-2 from the nasopharynx, there is still a debate in the inflammatory microenvironment the process is dysregulated and may lead to tissue damage by the secreted molecules from the granules [7].

Myeloperoxidase (MPO), is the most abundant protein in neutrophils and represents 5% of their total protein content [8]. Lau et al. reported the association of MPO and neutrophil activation via CD11b/CD18 integrins which is an indicator of MPO’s possible contribution to neutrophil recruitment to the site of inflammation [9]. Our data supports this hypothesis for SARS-CoV-2 patients which we detected significantly high expressions of MPO (≈4 fold, p<0.05) at the nasopharynx region. Furthermore it is tempting to speculate this neutrophil burst and even more overexpressed MPO may cause production of excess hypochlorous acid (HOCl) and other reactive oxidants that also damages the nasopharynx tissue. So MPO may be an important factor where the protective inflammation can become pathological in SARS-CoV-2 cases.

The resolution phase of inflammation is essential to curtail inflammation and restore tissue homeostasis [10]. Neutrophil apoptosis and the cytokines released from macrophages during phagocytosis of the neutrophils are necessary in the resolution of inflammation [11]. The resolution of inflammation and tissue repair processes are aided by the cytokines released from macrophages during phagocytosis of the neutrophils. The dysregulation in the macrophage orchestrated phagocytosis and neutrophil apoptosis mechanisms leads to inflammation [12]. PRTN3, another azurophilic granule (AG) protein that we identified, is a serine protease enzyme that is involved in granulocyte differentiation and expressed in the neutrophil granulocytes [13]. Increased PRTN3 was shown to have a negative effect on the resolution of inflammation that causes immune system deregulation [14] The strikingly high PRTN3 expression that we observed in SARS-CoV-2 positive patients compared to negative group (over 29 fold, p<0.05), may be the indicator of such phenomenon.

Multifunctional protease CTSG is also thought to be critically important in the maintenance of the delicate balance between tissue protection and destruction during inflammatory responses [15]. As a component of neutrophil proteolytic machinery CTSG regulates the inflammatory responses by stimulating the production of cytokines and chemokines, which are responsible for the activation and mobilization of immune cells to the site of pathogen and/or tissue damage [16,17]. CTSG activates metalloproteases and cleaves extracellular matrix proteins, contributing to neutrophil migration [18]. CTGS upregulation (more than 3 fold, p<0.05) that we observed in SARS-CoV-2 patients, possibly is another indicator for abnormal neutrophil accumulation in nasopharynx.

ELANE has a physiological function as a powerful host defense, but is also known as one of the most destructive enzymes in the body. An overwhelming release of enzymatically active ELANE can cause local tissue injury [19]. Addition to that it is also reported ELANE can activate the spike (S) protein of coronaviruses and shift the viral entry to a low pH-independent route [20]. Therefore, the highly expressed ELANE (≈3 fold, p<0.05) that we observed in SARS-CoV-2 infected patients is supporting these findings.

SARS-CoV-2 infected group appeared to be highly expressing the AZU1 protein (heparin-binding protein/cationic antimicrobial protein of 37 kD) which is mobilized rapidly from emigrating polymorphonuclear leukocytes (PMN). Initially, this inactive serine protease was recognized for its antimicrobial effects. However, it soon became apparent that azurocidin may act to alarm the immune system in different ways and thus serve as an important mediator during the initiation of the immune response. Azurocidin, released from PMN secretory vesicles or primary granules, acts as a chemoattractant and activator of monocytes and macrophages. The functional consequence is enhancement of cytokine release and bacterial phagocytosis, allowing for a more efficient bacterial clearance. Leukocyte activation by azurocidin is mediated via beta(2)-integrins, and azurocidin-induced chemotaxis is dependent on formyl-peptide receptors. In addition, azurocidin activates endothelial cells leading to vascular leakage and edema formation.

4.2. AG proteins and respiratory tract diseases

AG proteins’ active role in neutrophil-associated lung inflammatory and tissue-destructive diseases has been reported [21]. Increased expressions of PRTN3, ELANE, and CTSG was reported in COPD patients [22]. In a mouse model study ELANE or PRTN3 was introduced to the trachea which caused tissue destruction and enlargement in airspace [23]. ELANE expression was also implicated in the impairment of host defense resulting in a decrease in mucociliary clearance of bacteria and also pathogens phagocytosis [24]. CD2, CD4, and CD8 can be cleaved on the surface of T-cells by ELANE and CTSG that dysregulate T-cell function [25]. It was reported that patients that lack alpha-1 antitrypsin (α1-Pi) which is the physiological inhibitor of PRTN3 and ELANE carries a high risk of developing emphysema [26]. Azurophil granules also carry Cathepsin D which was reported to be up-regulated in the pulmonary macrophages in a mouse model of cigarette smoking [27]. Severe COPD cases exhibited MPO positive cells as a signal of neutrophil activation [28]. On the other hand in a mouse model of influenza it was shown that the inflammation damage was reduced by the absence of MPO [29]. Regarding the high expression results of these protein markers we may suggest that during the progression of SARS-CoV-2 infection the same molecular mechanisms are most likely to be induced.

4.3. AG proteins and cytokine driven immune dysregulation

Among the AG proteins that we have identified, especially PRTN3, ELANE and CTSG were mostly associated with cytokine driven immune dysregulation. IL-32, a proinflammatory cytokine with four isoforms, cleavage by PRTN3 propagates cytokine activity and triggers IL-1beta, TNF-alpha, IL-6, and chemokines. It was argued that the targeted inhibition of PR3 or silencing of IL-32 by an inactive form of PRTN3 may halt the IL-32 driven immune dysregulation [30]. ELANE, CTSG, and PRTN3 can cleave pro-IL-1beta to bioactive IL-1beta [31]. Caspase-1/Interleukin-1 converting enzyme (ICE) cleaves proteins like precursors of the inflammatory cytokines interleukin 1β and interleukin 18 into their mature biologically active forms [32] and CTSG regulates Caspase-1 in this pathway [33]. ICE has an active role in cell immunity as an initiator of inflammatory response so once activated it triggers the formation of active IL-1beta which is secreted from the cell that induces inflammation in the neighboring cells [34]. It was recently reported that infection in COVID-19 patients with acute respiratory syndrome showed release of the pro-inflammatory cytokines like IL-1beta and IL-6 [35].

4.4. AG proteins and their role in NETosis

NETosis is a type of programmed cell death where neutrophil extracellular traps are formed [22]. NETs were identified in 2004 and they are often overlooked as drivers of severe pathogenic inflammation [36]. NETs have pathogen killing properties and include strands of DNA wrapped with histones and are enriched with neutrophil proteins like MPO, ELANE, PRTN3 and AZU1 [37] which were also present in our SARS-CoV-2 (+) samples with high amounts. Although NETosis is a very powerful mechanism fighting for the infection, the ability of NETs to damage tissues is well-documented in infection and sterile disease. NETs directly kill epithelial and endothelial cells [11, 38], and excessive NETosis damages the epithelium in pulmonary fungal infection [12] and the endothelium in transfusion-related acute lung injury [39].