Lower peripheral blood Toll-like receptor 3 expression is associated with an unfavorable outcome in severe COVID-19 patients

The role of innate immunity in COVID-19 is not completely understood. Therefore, this study explored the impact of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) infection on the expression of Pattern Recognition Receptors (PRRs) in peripheral blood cells and their correlated cytokines. Seventy-nine patients with severe COVID-19 on admission, according to World Health Organization (WHO) classification, were divided into two groups: patients who needed mechanical ventilation and/or deceased (SEVERE, n = 50) and patients who used supplementary oxygen but not mechanical ventilation and survived (MILD, n = 29); a control group (CONTROL, n = 17) was also enrolled. In the peripheral blood, gene expression (mRNA) of Toll-like receptors (TLRs) 3, 4, 7, 8, and 9, retinoic-acid inducible gene I (RIGI), NOD-like receptor family pyrin domain containing 3 (NLRP3), interferon alpha (IFN-α), interferon beta (IFN-β), interferon gamma (IFN-γ), interferon lambda (IFN-λ), pro-interleukin(IL)-1β (pro-IL-1β), and IL-18 was determined on admission, between 5–9 days, and between 10–15 days. Circulating cytokines in plasma were also measured. When compared to the COVID-19 MILD group, the COVID-19 SEVERE group had lower expression of TLR3 and overexpression of TLR4.


Abbreviations
, were enrolled. COVID-19 was diagnosed by a positive SARS-CoV-2 RT-PCR test in nasopharyngeal and throat swabs and by typical chest CT-scan findings. All enrolled patients had hypoxemia (as defined by peripheral oxygen saturation less than 92%) and were using supplemental oxygen. Patients under 18 years old, pregnant women, and patients in end-of-life protocols were not included.
Initial assessment. All patients were evaluated and treated according to the Institution's protocol. They were submitted to CT, routine laboratory tests, and an additional 20 ml blood sample was collected in a tube with trisodium citrate that was immediately sent to the laboratory. The whole blood was separated by centrifugation, and small aliquots of leukocytes and plasma were stored at − 80 °C until the measurements.
Clinical outcomes and study groups. After emergency department assessment, patients were transferred to hospital wards or intensive care units. Clinical, radiologic, and routine laboratory tests were obtained from the electronic medical records. Enrolled COVID-19 patients (n = 79), had severe COVID-19 on admission, according to WHO guidelines 21 and were further classified into two groups: patients who needed mechanical ventilation and/or deceased (SEVERE, n = 50) and patients who used supplementary oxygen but not mechanical ventilation and survived (MILD, n = 29). Demographic and clinical data from these groups are shown in Table 1. A control group (CON-TROL, n = 17) composed of healthy patients with no history of COVID-19 or active respiratory disease and similar demographic characteristics was also enrolled.
To determine the distribution of TLR3 in different types of blood leukocytes, another set of 20 patients were included in the study, ten with our criteria for SEVERE disease and ten with MILD disease. From these additional patients, blood cells were separated by the Ficoll method in mononuclear and polymorphonuclear cells. RNA was further obtained from these cells for analyses, as described below.
RNA isolation and RT-PCR. Samples with whole white blood cells were thawed and diluted in a solution with Trizol (Thermo Fisher Scientific) and DNAse, according to the supplier's instructions. Extracted RNA was quantified by a photometer (Eppendorf) at 260 nm wavelength. The quantitative real-time polymerase chain reaction was carried out with SuperScript III Platinum One-Step qRT-PCR Kit with ROX (Thermo Fisher Scientific -#11745100) in thermocycler Step One Real-Time PCR System (Applied Biosystems) for TLR 3, 4, 7, 8 and 9, RIGI, NLRP3, IL18, IFN-α, IFN-β, IFN-γ, IFN-λ and pro-IL-1β. B2M was used as a housekeeping gene. Primer sequences (Supplementary Table 2

Results
Patients demographic and clinical characteristics. There was no difference between MILD and SEVERE groups regarding demographic and clinical features, and routine laboratory tests, except circulating C-reactive protein, leukocytes and neutrophils, which were higher in the SEVERE group (Table 1).  Table 3).

Innate immunity receptors and cytokines at enrollment.
An analysis of cytokines mRNA transcription was also performed. IFN-α, IFN-β, IFN-λ, and IL-18 were increased in COVID-19 patients when compared to CONTROL subjects. Pro-IL-1β mRNA was decreased in COVID-19 patients when compared to CONTROL subjects. No difference in IFN-γ mRNA was observed between CONTROL and COVID-19 patients (Fig. 2).
IFN-γ mRNA expression was lower in the COVID-19 SEVERE when compared to COVID-19 MILD. IL-18 transcription was higher in the COVID-19 SEVERE when compared to COVID-19 MILD (Fig. 2).   www.nature.com/scientificreports/ The association between the TLR3 and TLR4 mRNA expression and the antiviral cytokines was investigated. There was a positive correlation between IFN-β and IFN-γ mRNA expression and TLR3 mRNA expression (Fig. 3, Panels A and B), and a positive correlation between pro-IL-1β and IL-18 and TLR4 expression (Fig. 3, Panels C and D).
Innate immunity receptors during the hospital stay. During hospitalization, blood samples were collected every five days as feasible. The difference between COVID-19 MILD and COVID-19 SEVERE TLR3 and TLR4 mRNA expression at admission widened during follow-up. The differential expression of IFN-γ between COVID-19 MILD and COVID-19 SEVERE also widened (Figs. 4 and 5).
Circulating cytokines in plasma from the COVID-19 patients were also measured, and their profile was similar to that observed for mRNA. IFN-γ was lower in the SEVERE COVID-19 group when compared to COVID-19 MILD, while pro-inflammatory cytokine levels (IL-6, TNF-α, and MCP-1) were higher (Fig. 6).

Discussion
Sensing the invading pathogens by PRRs is the beginning of the antiviral defense line 24 . In this study, we evaluated the gene expression of PRRs engaged in antiviral response (TLRs 3, 7, 8, 9, RIGI and NLRP3).
Most PRRs were upregulated in COVID-19 patients compared to CONTROL. The only exception was TLR8. It is known that 1,25-dihydroxyvitamin D3 may impair TLR8 expression 25 , and vitamin D deficiency was associated with susceptibility to COVID-19 26 . However, we did not pursue this line of reasoning in this study since our primary interest was to identify the differential expression of PRRs when comparing COVID-19 MILD and COVID-19 SEVERE (TLR8 expression did not differ between those groups). When comparing COVID-19 MILD and COVID-19 SEVERE groups, a decreased expression of TLR3, accompanied by a deficient IFN-γ and an enhanced TLR4 expression were associated with an unfavorable outcome.
TLR3 has been described as an integral part of the immune response to several double-stranded RNA viruses 27 ; however, it was demonstrated that single-stranded RNA segments harboring stem structures with bulge/internal loops are also potent TLR3 agonists 28 . Moreover, TLR3-/-mice are more susceptible to SARS-CoV infection than wild-type animals 17 , and a porcine coronavirus was able to activate this receptor 29 .
In humans, Zhang et al. found autosomal-dominant deficiencies in patients with life-threatening COVID-19 that included mutations in the TLR3 gene 30 . In the same study, the authors showed that TLR3-/-fibroblasts   www.nature.com/scientificreports/ were more susceptible to SARS-CoV-2 infection in vitro. These data point out a prominent role for TLR3 in the immune response to coronavirus also in humans. The distinct TLR3 expression between MILD and SEVERE groups could be secondary to increased neutrophils in the SEVERE group (9.15 VS 6.11, p 0.004, Supplementary Table 3) as in peripheral blood there is evidence suggesting that TLR3 expression is enhanced in T-cells 31 . Consequently, the lower TLR3 expression in the COVID-19 SEVERE group could reflect a diminished fraction of lymphocytes in their peripheral blood samples. However, the evidence mentioned above 31 was obtained from naïve cells. As observed in Fig. 1, peripheral blood cells from COVID-19 patients present a distinct mRNA profile, with increased expression of all PRRs compared to cells from healthy subjects. Hence, an investigation of TLR3 expression among mononuclear and polymorphonuclear cells in the peripheral blood of COVID-19 patients was performed (Supplementary Table 3). The results demonstrated that TLR3 expression is similar in mononuclear cells and polymorphonuclear cells. TLR3 expression was slightly higher in the MILD group than in the SEVERE group but not statistically significant which may be related to the small number of samples evaluated in the experiment. Altogether, these data suggest that lower TLR3 expression in SEVERE patients was not related only to the increased number of neutrophils in these individuals. However, further studies are needed to find the source of the difference in TLR3 expression if it is due to an overall decrease in TLR3 expression in all cell subsets or a specific cell subset.
Also, there was a statistically significant, positive correlation between TLR3 and IFN-γ expression. Engaging TLR3 by double-stranded RNA viruses has been shown to induce a robust cytokine response, including IFN-γ 32 . However, this relationship is complex since IFN-γ can also increase TLR3 expression via signal transducer and activator of transcription 1 (STAT1) in cultured normal human epidermal keratinocytes 33 . Proving any causal relationship between TLR3 and IFN-γ expression requires a different experimental model.
An enhanced TLR4 expression was associated with the COVID-19 SEVERE group when compared to COVID-19 MILD. TLR4 recognizes lipopolysaccharide from gram-negative bacteria 34 but in silico studies suggested that it can also contribute to the response to the SARS-CoV-2 infection 35 .
Contrary to TLR3, the mRNA transcription for TLR4 was higher in the COVID-19 SEVERE group. This finding may represent a compensatory response by leukocytes when TLR3 expression is deficient. TLR3 and TLR4 often work synergistically against a viral challenge, utilizing the same network of intracellular signaling pathways 36 . Therefore, it seems reasonable to hypothesize that an impaired TLR3 would lead to a compensatory enhancement in TLR4 mRNA translation. The increased TLR4 mRNA expression may lead to a prominent inflammatory response, characteristic of the life-threatening COVID-19 disease.
Some limitations must be considered. Gene transcription was measured, but not protein expression. Also, although some correlations between PRRs and cytokines were described, as this is an observational study it is not an appropriate method to uncover causal relationships. On the other hand, the data was obtained from real patients and allowed in vivo measurements. Moreover, differences in peripheral blood cell type subsets between groups can interfere in the differences of PRRs mRNA expression measured. Finally, only the cells in the peripheral blood were analyzed. At baseline, the TLR expression profile among blood and lung resident cells is not identical 31 . In the present study, the inflammatory cells that migrated to the alveoli during COVID-19 induced ARDS were not evaluated. Conclusions about the expression of PRRs inside the lungs during COVID-19 cannot be inferred.
In summary, our research indicated that patients with an unfavorable outcome presented lower TLR3 expression and enhanced expression of TLR4 which may be a compensatory mechanism that leads to a prominent inflammatory response that is characteristic of severely ill COVID-19 patients. Our data support the possibility of using TLR3 expression to identify patients at risk of life-threatening COVID-19. Further research is necessary to understand the potential of modulating TLR3 expression as a method to prevent or treat COVID-19.

Data availability
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.