Untargeted saliva metabolomics reveals COVID-19 severity: Saliva Metabolomics for SARS-COV-2 Prognosis

TL;DR: In this paper, the authors demonstrate that salivary metabolomics can reveal COVID-19 severity using liquid chromatography-mass spectrometry (LCS) and RT-PCR testing.

Abstract: Background The COVID-19 pandemic is likely to represent an ongoing global health issue given the potential for vaccine escape and the low likelihood of eliminating all reservoirs of the disease. Whilst diagnostic testing has progressed at pace, there is an unmet clinical need to develop tests that are prognostic, to triage the high volumes of patients arriving in hospital settings. Recent research has shown that serum metabolomics has potential for prognosis of disease progression. 1 In a hospital setting, collection of saliva samples is more convenient for both staff and patients, and therefore offers an alternative sampling matrix to serum. We demonstrate here for the first time that saliva metabolomics can reveal COVID-19 severity. Methods 88 saliva samples were collected from hospitalised patients with clinical suspicion of COVID-19, alongside clinical metadata. COVID-19 diagnosis was confirmed using RT-PCR testing. COVID severity was classified using clinical descriptors first proposed by SR Knight et al. Metabolites were extracted from saliva samples and analysed using liquid chromatography mass spectrometry. Results In this work, positive percent agreement of 1.00 between a PLS-DA metabolomics model and the clinical diagnosis of COVID severity was achieved. The negative percent agreement with the clinical severity diagnosis was also 1.00, for overall percent agreement of 1.00. Conclusions This research demonstrates that liquid chromatography-mass spectrometry can identify salivary biomarkers capable of separating high severity COVID-19 patients from low severity COVID-19 patients in a small cohort study.

Summary

Introduction

• Halide perovskites have been at the forefront of new-emerging energy materials, due to their great potential in design and fabrication of new-generation optoelectronic devices, such as solar cells, 1 photodetectors, 2-5 light-emitting devices, 6,7 fieldeffect transistors 8,9 and lasers.
• For the all-inorganic counterparts, a PLQY of 19.4 % was realized in Cs3Bi2Br9, exhibiting quite good photostability and moisture stability as well.
• 31 This might also lead to the loss of the desired materials due to the mismatch with the current database.
• Moreover, it enriches the structural family of non-lead perovskites as well as holds the potential for optoelectronic applications.

Synthesis and Crystal Growth of Cs3BiBr6

• During the synthesis of the single crystal, if the amount of CsBr was much lower than the BiBr3, Cs3Bi2Br9 was obtained.
• Therefore, controlling the ratio between the two raw materials is a key factor to obtain the single crystals of Cs3BiBr6.

X-ray Crystallographic Studies

• Data collections were performed on a Bruker APEX-II CCD diffractometer equipped with graphite monochromated Mo Kα radiation (λ = 0.71073 Å) at 293 K.
• The final refined structural parameters were checked by the PLATON program.
• 34 Crystallographic data and structural refinements are summarized in Table S1 in Electronic Supplementary Information (ESI).
• Rietveld refinement of the powder XRD was performed by the Fullprof program.

Material characterization

• X-ray photodetection spectroscopy (XPS) (Thermo Scientific) was performed to analyze the compositions of samples, with Al-Kα (1486 eV) as the excitation X-ray source.
• The pressure of the analysis chamber was maintained at 2 × 10 -10 mbar during measurements.
• All characterizations were carried out at room temperature.
• Scanning electron Microscope (SEM) and Energy dispersive X-ray spectroscopy (EDX) characterizations were carried out on an FEI Quanta FEG 200 ESEM.
• The absorption behaviours of the samples were studied by the UV-vis spectrometer from Agilent.

Technologies (Santa Clara, USA). Absorption (α/S) data was

• Converted from diffuse reflectance spectra using the Kubelka−Munk function, α/S = (1 − R) 2 /2R, where R is the reflectance coefficient and α, S are the absorption and scattering coefficient.
• 35 All the photoelectrochemical characterizations were performed on the Autolab workstation.
• The phases and purities of all samples were also characterized by powder XRD with Cu Kα1 (λ = 1.5406 Å) radiation.
• Thermogravimetric analysis (TGA) was performed in the Mettler Toledo, Star E System (Columbus, OH, USA) with a GC 100 gas controller in a nitrogen atmosphere at a heating rate of 10 °C/min.
• It should be noted that all the measurements were performed.

Fabrication of Photodetectors

• ITO electrodes were prepared by the ultrasonic cleaning in acetone, ethanol, and Milli-Q water successfully.
• Cs3BiBr6 single crystals were dissolved into DMF and then a certain amount of solution was deposited on the device surface by drop-casting, followed by annealing at 100 °C.
• Bromidecoordination environment for (a) Bi1 and (b) Bi2. (c) 3D framework on the bc plane.
• Fig. 2 Chemical and structural relationship among a series of perovskites.

Results and discussion

• The 3D framework of Cs3BiBr6 is displayed in Fig. 1c, where all BiBr6 octahedra are isolated from each other forming 0D perovskite structure.
• The purity of samples was confirmed by powder X-ray diffraction (Fig. 3a).
• This indicates a higher thermal stability of Cs3BiBr6 than in the CH3NH3PbBr3.
• Fig. 5e and 5f show the responsivity and detectivity of the photodetector versus voltage and light density, respectively.

Conclusions

• In summary, the authors have successfully synthesized a non-toxic single crystal of perovskite, Cs3BiBr6.
• And it is the first time to analyse its crystallography structure, which features as an isolated BiBr6 polyhedra.
• Photodetector based on this perovskite is fabricated, exhibiting a good detectivity of 0.8*10 9 Jones under the 400 nm light illumination, demonstrating the potential for application into optoelectronic devices.
• Besides, this material displays high stability against thermal and moisture.
• It further facilitates the development of lead-free perovskites toward environment friendly energy conversion devices.

Figures

Figure 2: Volcano plot of statistical significance versus effect size for MS/MS validated 282 features separating participants presenting with high severity versus low severity 283 COVID-19 284

Figure 1: Saliva metabolomics analysis for COVID-19 diagnosis and prognosis via 268 LC-MS, showing: 269 A PLS-DA plot for 75 participants, COVID-19 positive / negative 270 B PLS-DA plot for 44 participants, high severity / low severity 271

Table 2: Confusion matrix for PLS-DA model projected on to COVID-19 negative 304 participants 305

Table 1: Summary of clinical characteristics by participant cohort 222

Figure 3: Boxplots of features down regulated with Covid-19 severity 286 [corresponding p-values: 0.066 ; 0.004 ; 0.051] 287 288

Full text

Page 1
Untargeted saliva metabolomics reveals COVID-19 severity 1
Running head: Saliva Metabolomics for SARS-COV-2 Prognosis 2
Cecile F. Frampas
1
, Katie Longman
1
, Matt P. Spick
1
,
Holly M. Lewis
1
, Catia D. S. 3
Costa
2
, Alex Stewart
3
, Deborah Dunn-Walters
3
, Danni Greener
4
, George E. 4
Evetts
4
, Debra Skene
3
, Drupad Trivedi
5
, Andrew R. Pitt
5
, Katherine Hollywood
5
, 5
Perdita Barran
5
and Melanie J. Bailey
1,2
6
1
Faculty of Engineering and Physical Sciences, University of Surrey, Guildford, GU2 7
7XH, UK 8
2
Surrey Ion Beam Centre, University of Surrey, Guildford, GU2 7XH, UK 9
3
Faculty of Health and Medical Sciences, University of Surrey, Guildford, GU2 7XH, 10
UK 11
4
Frimley Park Hospital, Frimley Health NHS Trust, GU16 7UJ, UK 12
5
Manchester Institute of Biotechnology, University of Manchester, M1 7DN, United 13
Kingdom 14
Corresponding Author: Dr Melanie Bailey, tel +44 (0)1483 682593, 15
m.bailey@surrey.ac.uk, Faculty of Engineering and Physical Sciences, University of
16
Surrey, Guildford, GU2 7XH, UK 17
Preprint location: MedRxiv 18
CF, KL and MS contributed equally to this paper. 19
Keywords: Saliva, Metabolomics, COVID-19, Prognostic Test, Liquid Chromatography 20
Mass Spectrometry 21
22
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted July 9, 2021. ; https://doi.org/10.1101/2021.07.06.21260080doi: medRxiv preprint
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.

Page 2
Abbreviations 23
COVID-19 Coronavirus disease 19
CRP C-reactive protein
HTN Hypertension
IHD Ischemic heart disease
KEGG Kyoto Encyclopedia of Genes and Genomes
LC Liquid chromatography
LC-MS Liquid chromatography mass spectrometry
LOOCV Leave-one-out cross validation
MS Mass spectrometry
MS/MS or MS
2
Tandem mass spectrometry
NPA Negative percent agreement
PCA Principal components analysis
PCR Polymerase chain reaction
PLS-DA Partial least squares-discriminant analysis
PPA Positive percent agreement
QC Quality control
RT-PCR Reverse transcription polymerase chain reaction
SARS-CoV-2 Severe acute respiratory syndrome coronavirus 2
T2DM Type 2 diabetes mellitus
VIP Variable importance in projection
24
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted July 9, 2021. ; https://doi.org/10.1101/2021.07.06.21260080doi: medRxiv preprint

1
ABSTRACT 25
Background 26
The COVID-19 pandemic is likely to represent an ongoing global health issue given 27
the potential for vaccine escape and the low likelihood of eliminating all reservoirs of 28
the disease. Whilst diagnostic testing has progressed at pace, there is an unmet 29
clinical need to develop tests that are prognostic, to triage the high volumes of patients 30
arriving in hospital settings. Recent research has shown that serum metabolomics has 31
potential for prognosis of disease progression.
1
In a hospital setting, collection of 32
saliva samples is more convenient for both staff and patients, and therefore offers an 33
alternative sampling matrix to serum. We demonstrate here for the first time that saliva 34
metabolomics can reveal COVID-19 severity. 35
Methods 36
88 saliva samples were collected from hospitalised patients with clinical suspicion of 37
COVID-19, alongside clinical metadata. COVID-19 diagnosis was confirmed using RT-38
PCR testing. COVID severity was classified using clinical descriptors first proposed by 39
SR Knight et al. Metabolites were extracted from saliva samples and analysed using 40
liquid chromatography mass spectrometry. 41
Results 42
In this work, positive percent agreement of 1.00 between a PLS-DA metabolomics 43
model and the clinical diagnosis of COVID severity was achieved. The negative 44
percent agreement with the clinical severity diagnosis was also 1.00, for overall 45
percent agreement of 1.00. 46
Conclusions 47
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted July 9, 2021. ; https://doi.org/10.1101/2021.07.06.21260080doi: medRxiv preprint

2
This research demonstrates that liquid chromatography-mass spectrometry can 48
identify salivary biomarkers capable of separating high severity COVID-19 patients 49
from low severity COVID-19 patients in a small cohort study. 50
51
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted July 9, 2021. ; https://doi.org/10.1101/2021.07.06.21260080doi: medRxiv preprint

3
1. Introduction 52
The SARS-CoV-2 pandemic has caused a sustained threat to global health since the 53
discovery of the virus in 2019.
2
Whilst great strides have been made in both treatment 54
and vaccination development,
3,4
the disease has inflicted multiple waves of infection 55
throughout the world during 2020 and into 2021.
5,6
COVID-19 has higher fatality rates 56
than seasonal influenza,
7
and in addition, new variants are constantly evolving with 57
the potential for either reduced vaccine effectiveness or altered lethality.
8
As a 58
consequence, there is a continuing need both for better understanding of the impact 59
of COVID-19 on the host metabolism as well as for prognostic tests that can be used 60
to triage the high volumes of patients arriving in hospital settings. 61
Nasopharyngeal swabs followed by polymerase chain reaction (PCR) have been 62
adopted worldwide for SARS-CoV-2 detection. However, supply chains for swabs 63
rapidly collapsed amongst exponential increases in demand for testing, highlighting 64
the urgency for alternative sample types and testing approaches. Furthermore, whilst 65
PCR tests are easily deployable and highly selective for the virus, these approaches 66
yield no prognostic information and cannot easily be delivered for rapid turnaround at 67
the point of care, for example during a hospital admissions process. In contrast, tests 68
based on mass spectrometry can be provided in minutes, with mass spectrometry 69
instrumentation typically available in hospital pathology laboratories. Prognostic tests, 70
whilst challenging due to the varied phenotypes that may present themselves,
9
could 71
be used to manage demand for hospitalisation and treatment, especially should 72
vaccine escape lead to future waves of COVID-19 infection. 73
Metabolic biomarkers in serum have been identified that carry prognostic information, 74
10,11
but sampling blood is invasive. Our experience in collecting and analysing patient 75
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted July 9, 2021. ; https://doi.org/10.1101/2021.07.06.21260080doi: medRxiv preprint

Frequently asked questions

Q1. What was the use of the cross-validation test?

26,27 Leave-one-out cross-validation was used for model validation test 197 accuracy, sensitivity and specificity; variable importance in projection (VIP) scores 198 were used to assess feature significance alongside p-values and effect sizes (fold 199 count). 

Q2. What contributions have the authors mentioned in the paper "Untargeted saliva metabolomics reveals covid-19 severity" ?

In this paper, the authors explore the potential of saliva to distinguish between severe and mild COVID-19 infection, with a view to providing a prognostic test that can be used to triage the high volumes of patients arriving in hospital settings. 

Q3. How did supply chains for swabs collapse?

supply chains for swabs 63 rapidly collapsed amongst exponential increases in demand for testing, highlighting 64 the urgency for alternative sample types and testing approaches. 

Q4. How can the authors measure the severity of COVID-19?

Prognostic tests, 70 whilst challenging due to the varied phenotypes that may present themselves, 9 could 71 be used to manage demand for hospitalisation and treatment, especially should 72 vaccine escape lead to future waves of COVID-19 infection. 

Q5. What was the way to determine the significance of COVID-19?

No 292 pathways met the criteria for both meaningful impact and statistical significance, 293 possibly due to the number features identified in each pathway being notably smaller 294 than typically achieved in serum or plasma, consistent with saliva being a filtrate and 295 in general featuring lower metabolic concentrations. 

Q6. What was the population analysis of COVID-19?

Population metadata overview 215The study population analysed in this work included 75 participants, comprising 47 216 participants presenting with a positive COVID-19 RT-PCR test and 28 participants 217 presenting without. 

Q7. What was the significance of the COVID-19 features?

Of these 360 features, 37 were identified as related to medical 249 interventions or food and were excluded, leaving 323 for statistical analysis. 

Q8. How long did the sample sit on ice?

The samples were agitated for one hour, sonicated three times 137 for 30 seconds, with resting on ice for 30 seconds between each sonication. 

Q9. who was responsible for supervision of the research team?

MB oversaw all aspects of 370 this work, including obtaining funding for the study, clinical access, experimental 371 design, analysis and was responsible for supervision of the research team.