key: cord-0848430-txlrh7zi
authors: Shan, D.; Hsiung, J.; Bliden, K. P.; Zhao, S.; Liao, T.; Wang, G.; Tan, S.; Liu, T.; Sreedhar, D.; Kost, J.; Chang, S. T.; Yuan, W. P.; Tantry, U.; Gurbel, P.; Tang, M.; Dai, H.
title: A New Saliva-Based Lateral-Flow SARS-CoV-2 IgG Antibody Test for mRNA Vaccination
date: 2021-06-16
journal: nan
DOI: 10.1101/2021.06.11.21258769
sha: e0411b3037f1720f215bc507017ab0267ff17542
doc_id: 848430
cord_uid: txlrh7zi

Sensitive detection of IgG antibodies against SARS-CoV-2 is important to assessing immune responses to viral infection or vaccination and immunity duration. Antibody assays using non-invasive body fluids such as saliva could facilitate mass testing including young children, elderly and those who resist blood draws, and easily allowing longitudinal testing/monitoring of antibodies over time. Here, we developed a new lateral flow (nLF) assay that sensitively detects SARS-CoV-2 IgG antibodies in the saliva samples of vaccinated individuals and previous COVID-19 patients. The 25-minute nLF assay detected anti-spike protein (anti-S1) IgG in saliva samples with 100% specificity and high sensitivity from both vaccinated (99.51% for samples [≥] 19 days post 1st Pfizer/BioNTech or Moderna mRNA vaccine dose) and infected individuals. Antibodies against nucleocapsid protein (anti-NCP) was detected only in the saliva samples of COVID-19 patients and not in vaccinated samples, allowing facile differentiation of vaccination from infection. SARS-CoV-2 anti-S1 IgG antibody in saliva measured by nLF demonstrated similar evolution trends post vaccination to that in matching dried blood spot (DBS) samples measured by a quantitative pGOLD lab-test, enabling the nLF to be a valid tool for non-invasive personalized monitoring of SARS-CoV-2 antibody persistence. The new salivary rapid test platform can be applied for non-invasive detection of antibodies against infection and vaccination in a wide range of diseases.

The COVID-19 pandemic by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been presenting a major global health crisis [1] [2] [3] [4] [5] [6] . It is clear that mass vaccination offers a viable path back to normalcy 7, 8 . As of May 2021, 17 different COVID-19 vaccines across four platforms are rolled out across the globe, with more than 300 additional vaccine candidates still in development. Over 1.8 billion vaccine doses have been administered worldwide, raising the hope of emerging from the pandemic 9 . However, questions linger on the duration of the protection offered by vaccination against COVID-19. Since antibody response and persistence vary among individuals and between the different vaccines in circulation, antibody assessment at the individual level could help to evaluate antibody persistence and the necessity of booster doses. Various antibody tests have been developed using enzyme-linked immunosorbent assay (ELISA) 10 , nanostructured plasmonic gold (pGOLD) assay 11 , chemiluminescence (CLIA) 12 and lateral flow (LF) technologies 13 . Antibodies against spike proteins have shown high correlation with neutralizing antibodies for COVID-19, offering a simple, first-line assessment of immunity [14] [15] [16] [17] [18] [19] [20] [21] [22] . Currently, most SARS-CoV-2 antibody tests require professional operations and invasive blood/serum processes, limiting the accessibility and frequency of testing. To facilitate personalized monitoring of SARS-CoV-2 antibody persistence, non-invasive and easy testing strategies are desired.

Saliva represents an appealing biofluid for non-invasive SARS-CoV-2 antibody detection, with the ease of self-collection and is pain-free. Previous work has shown that IgG antibodies in saliva correlated well with those in blood for various infectious diseases including SARS-CoV-2 11,23-32 . However, a challenge of saliva sample testing is the much lower concentration of antibodies present than in serum, plasma and blood, demanding high analytical sensitivity of detection methods 33 . The high viscosity of saliva samples also presents a challenge to flow-based rapid testing. At this time, no saliva-based antibody LF assay exist for SARS-CoV-2 vaccination or infection assessment at the point-of-care (POC).

Here, we developed a new lateral flow (nLF) test that is highly sensitive for SARS-CoV-2 IgG antibody detection in saliva. In conventional antibody LF assays, when a biofluid sample containing targeting antibodies is added, the sample flows through a conjugate pad containing gold (Au)-antigen complexes to form Au-antigen-antibody complexes, which migrate towards the detection zone and bind to the test line containing immobilized anti-human antibodies. A readable test line (whether by visual or reader-based detection) appears due to the complex-bound Au nanoparticles suggests a positive test result 34 . A limitation to this LF approach is that biofluid samples always contain abundant non-specific human antibodies that compete with the specific antibodies captured by the Au-antigen complex for the antihuman antibody binding sites on the test line, causing low analytical sensitivity and non-linearity of the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint assay. Moreover, saliva samples are viscous, requiring high degrees of dilution in order to run on conventional LF strips, further lowering the analytical sensitivity of saliva-based LF tests.

For our nLF approach, we first developed and optimized Au nanoparticle-antigen (or magnetic Fe 3 O 4 core/Au shell nanoparticle-antigen) complexes, then incubated with a saliva sample in a reaction tube to capture antibodies specific to the antigen, removed the non-specific antibodies (by micro-centrifuging or magnetic purification), and then applied the Au-antigen-antibody complexes to a LF test card containing anti-human IgG signal line for detection. The entire assay time was about 25 minutes. A portable, economic micro-centrifuge or magnetic rack was employed to remove the non-specific antibodies, making the assay user-friendly in POC settings. The nLF assay affords high anti-SARS-CoV-2 S1 IgG sensitivity (> 99%) with vaccinated saliva samples collected ≥ 19 days post the first Pfizer/BioNTech or Moderna mRNA vaccine dose accompanied by 100% specificity. By applying both SARS-CoV-2 spike protein S1 subunit and nucleocapsid (also known as nucleoprotein, NCP) to make Au-antigen complexes (Au-S1 and Au-NCP) for nLF assays, we differentiated saliva samples of SARS-CoV-2 vaccinated subjects from COVID-19 infected subjects. Importantly, anti-S1 antibodies in saliva showed a similar trend as those measured by a quantitative serology test performed on the pGOLD lab-test platform 11 .

Overall, the rapid nLF assay platform can be easily implemented at various POC sites including clinical laboratories, pharmacies, or eventually at home for non-invasive, highly accurate SARS-CoV-2 IgG monitoring.

The concentration of IgG antibodies in human saliva is much lower (~15 µg mL -1 ) than in serum or blood (~15 mg mL -1 ), so antibodies against SARS-CoV-2 spike protein in saliva are more difficult to detect than in sera of COVID-19 infected individuals 11, 24, 27, 33 . Several lab-based assays have been developed to meet this challenge, but not for saliva-based POC LF tests.

To develop our nLF SARS-CoV-2 IgG antibody saliva test, we first prepared and optimized Au nanoparticle-antigen complexes. The Au nanoparticles were either pure gold (~ 50 nm) or magnetic gold comprised an iron-oxide Fe 3 O 4 core and an Au shell (see Methods). When the Au-antigen (antigen = S1 or NCP of SARS-CoV-2) complexes were incubated with positive saliva samples, IgG antibodies specific to the viral antigen were captured to form Au-antigen-IgG complexes (Figure 1a (i-iii) ). A centrifuging . CC-BY-NC-ND 4.0 International license It 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 June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint or magnetic purification step was applied afterwards to collect the Au-antigen-IgG complexes and remove the unbound, non-specific antibodies. The resulting Au-antigen-IgG complexes were dispersed in a running buffer and added to a specially designed LF testing card to flow through the signal and control lines (Figure 1a (iv) ). The card was comprised of a sample pad, a conjugate pad with pre-deposited colloidal gold nanoparticle-biotinylated BSA (Au-biotin-BSA) complexes, a test line immobilized with anti-human IgG, a control line immobilized with streptavidin, and a terminal absorbent pad (Figure 1b) .

When antigen-specific IgG positive samples were used for the nLF, both test and control lines were detectable on the card as red lines, indicating a positive test, while negative samples or buffer alone only produced a red line in the control line (Figure 1c ). The step of removing unbound non-specific antibodies by micro-centrifuging or magnetic purification was critical to preventing competing antibodies in the sample from binding to the test line, drastically enhancing the detection sensitivity of specific antibodies over the traditional LF assay. The total nLF assay time was ~ 25 minutes.

We first compared the detection limit of nLF to that of a conventional LF test with FDA Emergency Use Authorization (MidaSpot TM COVID-19 Antibody Combo Detection Kit, or MidaSpot, developed by Nirmidas Biotech Inc.). We spiked purified recombinant anti-SARS-CoV-2 spike protein IgG into a negative healthy serum sample at various concentrations to determin the limit of detections by nLF and MidaSpot. MidaSpot showed positive signals for 4X and lower dilutions ( Figure S1a ). The nLF detected anti-S1 IgG even in the 256X diluted sample, affording a ~ 64 times lower detection limit than the conventional LF approach (Figure S1b), owing to the removal of excess, non-specific serum IgG by centrifugation or magnetic suction.

We also spiked purified recombinant anti-SARS-CoV-2 spike protein IgG into a negative healthy saliva sample and performed serial dilutions to establish the nLF assay detection limit for anti-S1 IgG in saliva.

By quantifying the signal intensity on the test line (see Methods), we obtained an estimated antibody detection limit of ~ 0.30 µg mL -1 , ~ 50-fold lower than the total antibody concentration (~15 µg mL -1 ) in human saliva (Figure 2a) 33 .

We next focused on nLF assays of SARS-CoV-2 antibodies in saliva samples from vaccinated and COVID-19 infected cohorts. The major SARS-CoV-2 antigenic targets of human IgG are the spike protein and nucleocapsid (NCP). The SARS-CoV-2 spike protein is used as the target antigen and encoded in the mRNA vaccines from Pfizer/BioNTech and Moderna. Individuals who have been successfully vaccinated by the one of the two mRNA vaccines, without prior COVID-19 infection, should . CC-BY-NC-ND 4.0 International license It 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 June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint only test positive for anti-S1 IgG but not for anti-NCP IgG [35] [36] [37] . However, COVID-19 infected patients are known to test positive for both anti-S1 IgG and anti-NCP IgG 38 . The nLF assay platform is flexible for testing human IgG antibodies against any protein target by simply changing Au-antigen complexes while using the same test card. Using Au-S1 and Au-NCP complexes, we readily developed nLF tests for evaluating IgG antibodies against SARS-CoV-2 spike protein and NCP in saliva samples (Figure 2b ).

With these tests, we observed a typical vaccinated saliva sample (42 days post 1st RNA vaccine dose from Moderna) exhibiting positive anti-S1 IgG and negative anti-NCP IgG, while an infected saliva sample (81 days post symptom onset) showing both positive anti-S1 IgG and positive anti-NCP IgG ( Figure 2c ). The same pattern was observed for larger numbers of vaccinated (20) and infected samples (20) tested.

To fully establish the nLF assay for saliva anti-S1 IgG testing, we tested 34 unvaccinated healthy saliva For the anti-NCP IgG nLF assay, a cut-off value was determined by testing saliva samples from 12 unvaccinated COVID-19 free individuals and 20 COVID-19 infected individuals. The ROC analysis gave a cutoff value of 1.004 intensity unit under the criteria of 100% specificity and 100% sensitivity ( Figure   S2b ). All of the 20 COVID-19 infected saliva samples were tested positive for both anti-S1 IgG (100%) and anti-NCP IgG (100%) (Figure 3a ). In contrast, of the 20 vaccinated saliva samples tested by nLF on day 42-55 post 1st vaccine dose, 0 out of 20 (0%) was tested positive for anti-NCP IgG (Figure 3b) , and all 20 (100%) were tested positive for anti-S1 IgG. This result suggests that salivary anti-S1 and anti-NCP IgG testing with the same nLF platform could offer an effective and non-invasive approach to . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint differentiate immune responses by vaccination (for mRNA vaccines targeting spike protein) or prior COVID-19 infections.

To further validate our nLF assay for monitoring SARS-CoV-2 IgG trends post vaccination, we compared nLF IgG saliva test results with dried blood spot (DBS) serology test results using a previously established quantitative lab-test, the pGOLD assay 11 . We longitudinally measured 64 previously COVID- (9/9) for the 6 groups respectively, corroborating with anti-S1 IgG positivity measured in matched DBS samples by the quantitative pGOLD assay (Figure 4a ). Only one vaccinated saliva sample from group D19-28 was not detected positive to anti-S1 IgG by the nLF assay, with its matched DBS sample also showing a much weaker signal than other samples in the same group by the pGOLD assay ( Figure S3) .

Follow-up studies of saliva and DBS samples from the same participant at the subsequent timepoint (D33-55) were tested anti-S1 IgG positive by nLF and pGOLD assays (data is not presented).

The evolution of salivary anti-S1 IgG and anti-S1 IgG levels in matched DBS samples were monitored by nLF and pGOLD respectively over ~ 4 months post vaccination, showing similar trends (Figure 4b-c) .

Generally, anti-S1 IgG in both saliva and DBS showed moderate levels on day 19-28 after the first dose of mRNA vaccines, and further increased to higher levels after the second vaccine dose. The anti-S1 IgG levels reached a maximum on day 33-55, then remained elevated with a slow declining trend from the peak during the follow up timepoints (day 61-83, 89-108, and 112-133). This trend was consistent with a recent report on antibody persistence over 6 months after the 2nd dose of the Moderna mRNA1273 vaccine 39 . One disadvantage of the nLF assay, when compared to NIR fluorescence based pGOLD assay, is the narrower dynamic range of the test line intensity, making it less differentiating in the high end of antibody concentrations. Nevertheless, we observed a similar trend of anti-S1 IgG in saliva detected by nLF assay and anti-S1 IgG in DBS detected by pGOLD assay (Figure 4) , suggesting the saliva-based . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint assay using the nLF platform could represent a rapid, noninvasive alternative for SARS-CoV-2 IgG testing and longitudinal monitoring.

To explore personalized monitoring of SARS-CoV-2 antibody persistence, we followed several vaccinated participants closely (5 from Pfizer/BioNTech, 2 from Moderna) by collecting their saliva samples weekly or bi-weekly and measuring their salivary anti-S1 IgG levels up to 13 weeks using our nLF assay. The antibody levels and details of evolution were highly variable among all participants ( Figure S4 ). Importantly however, the salivary and blood anti-S1 levels for individuals followed the same trend of evolution post vaccinations ( Figure S5) , suggesting the feasibility of non-invasive salivary antibody measurements by nLF for longitudinal assessment of vaccination effects. For example, the anti-S1 IgG of participant A (Pfizer/BioNTech vaccine) was not detected until week 3. After the 2nd vaccine dose in week 3, the anti-S1 IgG declined in week 4 and then boosted in week 6 and 7, and declined later with some fluctuations (Figure 5a ). In the case of participant B, a moderate anti-S1 IgG was detected in week 2 post the 1st vaccine dose (Moderna), and the IgG level declined gradually in week 3 and 4. After the 2nd vaccine dose in week4, the IgG level jumped drastically in week 5, then was sustained at a relatively high level and sustained with modest decline from the peak during the following weeks ( Figure   5b ).

A simple but important feature of the nLF platform for rapid antibody detection is the removal of abundant, non-specific antibodies from a sample to afford higher analytical sensitivity (by ~ 50-fold over conventional LF) and signal linearity. This added step can be done in POC settings using simple microcentrifuging (handheld, battery operated one available) or magnetic separation by employing magnetic nanoparticles residing in the core of a gold shell and permanent magnets for rapid washing ( Figure S6) .

Although an additional step is required relative to conventional LF antibody test, it is crucial to obtain IgG signals on the test lines that reflect the specific antibody concentrations in the sample due to the elimination of competing antibodies that would always lower the true signal.

The nLF approach is applicable to detecting antibodies in all major body fluids, including serum, plasma, venous and capillary whole blood and dried blood spot, for assessing immune responses to a wide range of infectious diseases or vaccines in POC settings within 25 minutes. It is also useful for detecting . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint autoantibodies for various autoimmune diseases. The rapid test can be semi-quantitative especially with the integration of an LF reader that detects light scattering of gold nanoparticles. A wide range of fluorescent or luminescent particles can also be employed for the nLF platform, boosting the nLF assay dynamic range by orders of magnitude for biomarker quantification.

We presented findings of longitudinal testing of salivary and blood IgG antibodies for individuals at various timepoints post vaccination up to ~ 4 months. The study will continue for > 1 year to correlate salivary and blood antibody levels and their relations with immunity to COVID-19 infection. A limitation of our work is that rigorous immunity assessments require serological neutralizing antibody quantification, raising the question of relevance of anti-SARS-CoV-2 spike protein levels in saliva and blood. However, neutralization assays are time-consuming and employ live viruses, which are unsuitable for mass vaccination studies. Several simplified neutralization assays have been developed without using live viruses 20, [40] [41] [42] . Importantly, there have been mounting evidence that neutralizing antibody levels in COVID-19 patients are well correlated with antibodies against spike protein on the surface of SARS-CoV-2 14, 16, [20] [21] [22] . Further, our results here showed that salivary anti-S1 IgG was well correlated with that in the blood for the mRNA vaccinated cohort as a whole ( Figure 4) and at the individual level ( Figure 5 , Figure S4 , and Figure S5 ). Such correlation has been seen recently for COVID-19 patients 23,24,27 and in various other infectious diseases in the past few decades 30, 31 . All of these results present a compelling case for monitoring salivary IgG against spike protein in vaccinated population as a non-invasive approach to assessing antibody levels and immunity. We clearly observed similar evolution trends and patterns for anti-S1 IgG in saliva and in blood. Since these antibodies are resulted entirely from vaccination, and when immunity wanes, salivary antibodies could also diminish. Continual monitoring of antibody persistence is crucial as COVID-19 vaccines circulate and aid the path to normalcy.

Achieving > 99.5% sensitivity for anti-SARS-CoV-2 spike protein IgG detection in vaccinated saliva samples (19 days post first vaccine shot) at 100% specificity, the rapid nLF assay provides an efficient and easy tool to evaluate population immunity to COVID-19, identifying potential immunity gaps and susceptible populations to inform targeted vaccination/prevention strategies like a booster vaccine dose.

The flexible designability of the nLF assay platform not only enables users to easily differentiate vaccination from infection, but also gives the potential to detect other infected and autoimmune diseases by adjusting the Au-detecting agent complexes. With further improvements and modifications of the nLF . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint platform such as introducing magnetic particles and a LF test reader, the nLF assay will be more user friendly for POC and can be translated into home testing.

The study was performed in accordance with standard ethical principles and approved by the LifeBridge SARS-CoV-2 spike (S1) (40591-V08H) was purchased from Sino Biological. SARS-CoV-2 nucleocapsid protein (NCP) (ncov-PS-Ag6) was purchased from FAPON Biotech Inc.

All chemicals were purchased from Sigma-Aldrich, and were used as received without further purification, except where mentioned otherwise.

A colloidal gold (Au) nanoparticle solution was prepared by adding 0.7 ml 1 wt.% sodium citrate solution to 80 ml boiling gold chloride solution (0.0125 wt.%). After a 30-minute reaction, the colloidal Au nanoparticles were prepared and cooled to room temperature for Au-antigen preparation. The magnetic Au (mAu) nanoparticle containing a Fe 3 O 4 core and an Au shell are products of Nirmidas Biotech. Inc.

For Au-antigen or mAu-antigen complex preparation, 10 μg of S1 or NCP was mixed with 1.25 ml colloidal Au or mAu solution for 30 minutes at room temperature, leading to the ionic adsorption of SARS-CoV-2 antigens on the surface of the colloidal Au or mAu particles. After blocking with a solution containing BSA, the mixture was centrifuged at 10000 rpm for 5 minutes. The supernatant was discarded . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint 1 0 and the pellet containing Au-antigen/mAu-antigen complex was suspended in 60 μ L Tris buffer containing a surfactant and stored at -20 ˚C for later usage. With magnetic mAu particles, a magnetic rack was used to attract and hold the particle for removing unbound species and washing the particles with captured specific antibodies (see Figure S6 ).

The immunochromatographic strip was composed of five components, a plastic backing, a sample pad, a conjugate pad, a nitrocellulose membrane and an absorbent pad. The sample pads and the conjugate pads were treated with 20 mM Tris buffer containing blocking agent and dried at 37°C and 25% relative humidity. The mouse anti human IgG (2 mg/ml) or the streptavidin (1 mg/ml) in PBS (1% wt. trehalose) was dispensed at the test or the control line on the nitrocellulose membrane, using a dispenser at a rate of 1.0 μ L/cm and then dried at 37°C. The gold nanoparticles conjugated with biotin-BSA were applied to the treated conjugate pad and then dried completely. The absorption pad, nitrocellulose membrane, pretreated conjugate pad, and sample pad were attached to a plastic backing and assembled as a strip with a 1.5 mm overlap, sequentially. The assembled plate was cut into 4-mm-wide pieces, using an automatic strip cutter ZQ2000. The generated strip products were packaged in a plastic bag with desiccant and stored at room temperature.

To test a saliva sample, 200 

A 2 mm circle was punched out from a DBS sample (comprised of 10 μ l dried finger stick whole blood) and dissolved in 160 µL sample buffer for 1 h. After centrifugation at 3000 g for 5 minUTES, the DBS-. CC-BY-NC-ND 4.0 International license It 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) 1 1 dissolved sample solution was heat inactivated at 56 º C for 30 minutes. The pGOLD antibody assay was performed in a 64-well format following a protocol similar to our previous work 11 .

This work was supported by Platelet and Thrombosis Research, LLC, Baltimore, MD, USA. Device fabrication was supported by Shenzhen Grant No. JSGG20191231141403880 and Guangdong Grant No. 2020B1111160003.

H.D. worked on this project as a consultant to Nirmidas Biotech Inc., independent of his research at Stanford.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint

formation. After saliva sample collection (i), Au-antigen (antigen = S1 or NCP of SARS-CoV-2) complexes were incubated with the sample (ii) and capture antigen-specific IgG antibodies to form Auantigen-IgG complexes (iii). A centrifuging or magnetic purification step was applied afterwards to collect the Au-antigen-IgG complexes and remove non-specific antibodies. The obtained Au-antigen-IgG complexes were dispersed in a PBS buffer before being added to a specially designed LF testing card. b, Design of the nLF testing card. The card is comprised of a sample pad, a conjugate pad with predeposited colloidal gold nanoparticle-biotinylated BSA (Au-biotin-BSA) complexes, a test line immobilized with anti-human IgG, a control line immobilized with streptavidin, and a terminal adsorbent pad. c, Operation of the nLF assay. When antigen-specific IgG positive samples are used for the nLF, both test and control lines are detectable on the card, indicating a positive test, while negative samples or buffer alone only produced a red line in the control line.

CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint Figure 2 . Highly sensitive and multifunctional nLF test for SARS-CoV-2 antibody testing in saliva. a, Detection limit of anti-SARS-CoV-2 spike protein IgG in saliva. Purified recombinant anti-SARS-CoV-2 spike protein IgG were spiked into a negative healthy saliva sample and serial dilutions were performed to establish the nLF assay detection limit for anti-S1 IgG in saliva. By quantifying the signal intensity on the test line (see Methods), we obtained an estimated antibody detection limit of ~ 0.30 µg mL -1 . b, Schematic representation of nLF for anti-S1 IgG detection (i) and anti-NCP IgG detection (ii). c, Representative testing results and IgG signals obtained from nLF for vaccinated and infected saliva samples. A typical vaccinated saliva sample (42 days post 1st RNA vaccine dose from Moderna) shows positive anti-S1 IgG and negative anti-NCP IgG, while an infected saliva sample (81 days post symptom onset) shows both positive anti-S1 IgG and positive anti-NCP IgG.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint in matched dried blood spot (DBS) samples by the pGOLD assay are 0% for D0 samples and 100% for all of the vaccinated groups. b, The evolution of salivary anti-S1 IgG monitored by nLF. c, The evolution of anti-S1 IgG levels in matched DBS samples monitored by pGOLD assay. Generally, anti-S1 IgG in both saliva and DBS showed moderate levels on day 19-28 after the first dose of mRNA vaccines, and further increased to higher levels after the second vaccine dose. The anti-S1 IgG levels reached a maximum on day 33-55, then remained elevated with a slow declining trend from the peak during the follow up timepoints (day 61-83, 89-108, and 112-133). Figure 5 . Personalized monitoring of SARS-CoV-2 antibody persistence. a, Monitoring of anti-S1 IgG over 13 weeks post vaccination of a representative participant who received Pfizer/BioNTech vaccine Salivary anti-S1 IgG were collected and tested by nLF weekly or bi-weekly. Matched DBS samples were collected and tested by pGOLD assay at selected timepoints. c, Monitoring of anti-S1 IgG over 13 weeks post vaccination of a representative participant who received Moderna vaccine. Salivary anti-S1 IgG were collected and tested by nLF weekly or bi-weekly. Matched DBS samples were collected and tested by pGOLD assay at selected timepoints. 8 all ne.

re . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint Supplementary information Figure S1 . Detection limit of anti-spike IgG in serum by nLF and MidaSpot. Purified recombinant anti-SARS-CoV-2 spike protein IgG was spiked into a negative serum sample and conducted serial dilutions (4X to 1024X dilutions) to determine the detection limit by nLF and MidaSpot. a, Detection limit of anti-spike IgG in serum by MidaSpot. Nirmidas' MidaSpot™ COVID-19 rapid Antibody Combo Detection Kit was approved by FDA EUA for POC testing. Based on the testing instruction, 10 µl serum sample was applied to the sample window followed by adding 4 drops (~ 100 µl) of running buffer, and after 20 min the result was recorded. MidaSpot presents positive signals for the 4X and lower dilutions. b, Detection limit of anti-spike IgG in serum by nLF. The nLF (see Method) is able to detect anti-spike IgG even in the 256x diluted serum sample, affording a ~ 64 times lower detection limit than the conventional LF approach owing to the removal of excess, non-specific serum IgG by nLF. 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 June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint Figure S3 . pGOLD results for anti-S1 IgG of DBS samples from group D19-28. 28 saliva samples from the group D19-28 are tested by nLF, and their matched DBS samples are tested by pGOLD assay. 1 out of 28 vaccinated saliva samples was not detected positive to anti-S1 IgG by the nLF, with its matched DBS sample (red square) also showing a much weaker signal than other samples in the same group by the pGOLD assay. 1 

. CC-BY-NC-ND 4.0 International license It 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 June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint Figure S4 . Personalized monitoring of SARS-CoV-2 antibody persistence. Personalized monitoring of SARS-CoV-2 anti-S1 IgG levels for 7 vaccinated participants (participant A-G) (5 from Pfizer/BioNTech, 2 from Moderna) up to 13 weeks using the nLF assay.

2 of h, . CC-BY-NC-ND 4.0 International license It 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 June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint Figure S5 . Comparison of salivary and blood anti-S1 IgG post vaccination for individuals. a-g, Comparison of salivary and blood anti-S1 IgG for 7 individuals (5 from Pfizer/BioNTech, 2 from Moderna) at selected timepoints post vaccination. Salivary anti-S1 IgG signals were detected by nLF. DBS anti-S1 IgG signals were detected by pGOLD assay using DBS samples. 3 . CC-BY-NC-ND 4.0 International license It 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 June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint Figure S6 . Magnetic Au nanoparticle (mAu) based nLF antibody assay. a, A magnetic rack for removing unbound non-specific antibodies. b, mAu-S1-IgG complex collection by the magnetic rack. c, Representative results from mAu based nLF for anti-S1 IgG positive and negative saliva samples. 4 . CC-BY-NC-ND 4.0 International license It 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 June 16, 2021. ; https://doi.org/10.1101/2021.06.11.21258769 doi: medRxiv preprint

Impact of the COVID-19 pandemic on quality of life and mental health in children and adolescents in Germany

SARS-CoV-2 variants and ending the COVID-19 pandemic

The first and second waves of the COVID-19 pandemic in Africa: a crosssectional study

The economics of COVID-19 pandemic: A survey

Future scenarios for the COVID-19 pandemic

Increase in suicide following an initial decline during the COVID-19 pandemic in Japan

Measuring the impact of COVID-19 vaccine misinformation on vaccination intent in the UK and USA

SARS-CoV-2 infection after vaccination in health care workers in California

Keeping track of the SARS-CoV-2 vaccine pipeline

Diagnostic performance of seven rapid IgG/IgM antibody tests and the Euroimmun IgA/IgG ELISA in COVID-19 patients

Quantification of antibody avidities and accurate detection of SARS-CoV-2 antibodies in serum and saliva on plasmonic substrates

A peptide-based magnetic chemiluminescence enzyme immunoassay for serological diagnosis of coronavirus Disease

Four point-of-care lateral flow immunoassays for diagnosis of COVID-19 and for assessing dynamics of antibody responses to SARS-CoV-2

Robust neutralizing antibodies to SARS-CoV-2 infection persist for months

Temporal profiles of viral load in posterior oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: an observational cohort study

SARS-CoV-2 spike protein: a key target for eliciting persistent neutralizing antibodies

Subunit Vaccines Against Emerging Pathogenic Human Coronaviruses

Convalescent plasma anti-SARS-CoV-2 spike protein ectodomain and receptorbinding domain IgG correlate with virus neutralization

Persistence and decay of human antibody responses to the receptor binding domain of SARS-CoV-2 spike protein in COVID-19 patients

Quantification of SARS-CoV-2 neutralizing antibody by a pseudotyped virus-based assay

A SARS-CoV-2 surrogate virus neutralization test based on antibody-mediated blockage of ACE2-spike protein-protein interaction

Competitive SARS-CoV-2 Serology Reveals Most Antibodies Targeting the Spike Receptor-Binding Domain Compete for ACE2 Binding

ELISA detection of SARS-CoV-2 antibodies in saliva

COVID-19 serology at population scale: SARS-CoV-2-specific antibody responses in saliva

Ultrasensitive measurement of both SARS-CoV-2 RNA and antibodies from saliva

High similarity of IgG antibody profiles in blood and saliva opens opportunities for saliva based serology

Persistence of serum and saliva antibody responses to SARS-CoV-2 spike antigens in COVID-19 patients

Evidence for sustained mucosal and systemic antibody responses to SARSCoV-2 antigens in COVID-19 patients

Immunodiagnosis of dengue virus infection using saliva

Plasmonic gold chips for the diagnosis of Toxoplasma gondii, CMV, and rubella infections using saliva with serum detection precision

Salivary biomarkers: toward future clinical and diagnostic utilities

Salivary diagnostics of the novel coronavirus SARS-CoV-2 (COVID-19)

Antibody responses to SARS-CoV-2 mRNA vaccines are detectable in saliva

Lateral flow assays

Antibody responses to the BNT162b2 mRNA vaccine in individuals previously infected with SARS-CoV-2

COVID-19 vaccine BNT162b1 elicits human antibody and TH1 T cell responses

mRNA vaccine-elicited antibodies to SARS-CoV-2 and circulating variants

Kinetics of nucleo-and spike protein-specific immunoglobulin G and of virusneutralizing antibodies after SARS-CoV-2 infection

Antibody persistence through 6 months after the second dose of mRNA-1273 vaccine for Covid-19

The role of pseudotype neutralization assays in understanding SARS CoV-2

Evaluation of SARS-CoV-2 neutralizing antibodies using a vesicular stomatitis virus possessing SARS-CoV-2 spike protein

Neutralization assay with SARS-CoV-1 and SARS-CoV-2 spike pseudotyped murine leukemia virions