Low Innate Immunity and Lagged Adaptive Immune Response in the Re-Tested Viral RNA Positivity of a COVID-19 Patient.

Recent studies have highlighted observations regarding re-tested positivity (RP) of SARS-CoV-2 RNA in discharged COVID-19 patients, however, the immune mechanisms underlying SARS-CoV-2 RNA RP in immunocompetent patients remain elusive. Herein, we describe the case of an immunocompetent COVID-19 patient with moderate symptoms who was twice re-tested as positive for SARS-CoV-2 RNA, and the period between first and third viral RNA positivity was 95 days, longer than previously reported (18-25 days). The chest computed tomography findings, plasma anti-SARS-CoV-2 antibody, neutralizing antibodies (NAbs) titer, and whole blood transcriptic characteristics in the viral RNA RP patient and other COVID-19 patients were analyzed. During the SARS-CoV-2 RNA RP period, new lung lesions were observed. The COVID-19 patient with viral RNA RP had delayed seroconversion of anti-spike/receptor-binding domain (RBD) IgA antibody and NAbs and were accompanied with disappearance of the lung lesions. Further experimental data validated that NAbs titer was significantly associated with anti-RBD IgA and IgG, and anti-spike IgG. The RP patient had lower interferon-, T cells- and B cell-related genes expression than non-RP patients with mild-to-moderate symptoms, and displayed lower cytokines and chemokines gene expression than severe patients. Interestingly, the RP patient had low expression of antigen presentation-related genes and low B cell counts which might have contributed to the delayed anti-RBD specific antibody and low CD8+ cell response. Collectively, delayed antigen presentation-related gene expression was found related to delayed adaptive immune response and contributed to the SARS-CoV-2 RNA RP in this described immunocompetent patient.

Introduction

Recent reports have highlighted the recurrence of SARS-CoV-2 RNA positivity in discharged COVID-19 patients (1) and their risk of being infectious (2). The causes for recurrent positive detection of SARS-CoV-2 RNA might be low virus load in the respiratory samples of discharged COVID-19 patients and/or low assay-sensitivity of detection methods; leading to false negative results at the time of discharge (3). Therefore, the “turned positive” phenomenon of SARS-CoV-2 RNA could be indicating prolonged viral shedding rather than “recurrence” (4). More importantly, SARS-CoV-2 could accelerate its mutation during prolonged infection in immunocompromised COVID-19 patients, particularly, the mutation of receptor-binding domain (RBD) region of spike gene could lead to immunologic escape, then bring about novel higher infectious virus isolates emergence, leading to recurrence of infection or fatality (5). Moreover, infectious virus could be shedded in few SARS-CoV-2 RNA redetectable patients (6). Therefore, there is great urgency to promptly identify these patients to implement timely therapy and quarantine measures.

Prolonged SARS-CoV-2 RNA shedding has been mostly reported in immunocompromised patients (7–10). These studies demonstrated that a decline or delay in antibody response could lead to relapse or recovery of COVID-19 (7). However, in a recent report involving a lymphoma patient with low immunity response, the patients were re-tested as positive for SARS-CoV-2 RNA after administration of COVID-19 convalescent plasma transfusion (11). Additionally, prolonged SARS-CoV-2 RNA shedding was observed after seroconversion of anti-spike antibody in a COVID-19 patient (3), suggesting that other mechanism attributes, apart from antibody, could also be playing key roles in virus clearance (12). Multiple evidences suggest that cellular immunity also played critical roles in SARS-CoV-2 clearance (13); thereby, relating the reduction of CD8+ T-cell as a potential risk factor for the longer duration of SARS-CoV-2 RNA positivity (14). Herein, we explored the clinical characteristics of a COVID-19 patient who was twice re-tested as positive for SARS-CoV-2 RNA and the immune mechanisms underlying re-tested viral RNA positivity in this immunocompetent patients.

Methods

Patients and Samples

The study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Maoming People’s Hospital (Approval No. 2020012). The patient’s clinical data were extracted from the hospital records. The severity of the disease was assessed using the Seventh Version of the Novel Coronavirus Pneumonia Diagnosis and Treatment Guidance from the National Health Commission of China (15). Plasma at viral RNA positive and negative stage of each patient was collected for antibodies detection. In addition, 1mL of whole blood was added to 4 ml trizol solution (Invitrogen, USA), then stored at −80°C for further transcriptome analysis.

Purification of IgA and IgG From Anti-RBD Antibody-Positive Plasma

The IgA was isolated using Peptide M/Agarose (Invivogen, USA). Briefly, 1 ml of Peptide M/Agarose was packed into a suitable column, then equilibrated with 5 ml equilibration and wash buffer (EWB). Afterward, 0.9 ml of plasma was loaded into the column and was washed with 10 ml of EWB. The binding antibody IgA was eluted with 10 ml of elution buffer. Subsequently, neutralization buffer was added into the elution buffer containing IgA to adjust pH to 7.5. Then, the plasma with IgA was loaded onto the Protein A column (GE, 17508001) and chromatography were performed at a flow rate of 5 ml/min using the AKTA Z100 (GE). Next, the columns were washed with PBS. IgG was eluted with 10 ml of 0.1 M glycine (Scientific Chemical, VA13110) and the pH was immediately adjusted to 7.5 with 1 M of Tris (Sigma-Aldrich, T6606). Finally, PBS exchange was achieved using Amicon Ultra centrifugal filters (Merck Millipore, UFC9050) through a 50-kD membrane according to the manufacturer’s instructions. The absorbance value of anti-RBD/anti-spike IgM, IgG, and IgA were determined using the above ELISA method in a dilution of 1:200.

Anti-NP, Anti-RBD, Anti-Spike Antibody Detection

Antibody detection was performed as previously described (16). Briefly, 100 μl of PBS containing 50 ng of NP, RBD or S1/S2 mixture (Sino Biological, Beijing, China) was added into the wells of 96-well polystyrene plates and were incubated at 4°C overnight. After two washes with PBST, 100 μl of phosphate-buffered saline with 0.1% tween®20 (PBST) containing 5% non-fat milk was added and incubated at 37°C; for 1 h. Diluted plasma (1:200) was then added into the wells after three washings with PBST and incubated at 37°C for 2 h. Some 100 μl of diluted Anti-Human IgG-HRP (1:10,000, Sigma Aldrich, USA), Anti-Human IgM-HRP (1:2,000, Sigma Aldrich, USA), or Goat Anti-Human IgA-HRP (1:40,000, SouthernBiotech, USA) was then added into the wells after four washings with PBST and incubated at 37°C for 1 h. The 96-well polystyrene plates were washed five times with PBST, then 50 μl of TMB (3, 3′, 5,5′-Tetramethylbenzidine) One Solution (Guangzhou Jingxin Biotechnology Co., Ltd, China) was added and incubated at room temperature for 5–15 min. Next, 50 μl of 1 M H2SO4 solution was added to terminate the chromogenic reaction. The ratios of absorbance were calculated by the absorbance value dividing 2.1 × mean absorbance value of four negative control pooled plasma samples (10 SARS-CoV-2 RNA negative plasma samples were pooled as one pool sample).

Pseudotyped Virus-Based Neutralization Assay

Neutralization assay was performed according to our previously described methods (17). In brief, lentiviral vectors pseudotyped by an S protein were produced by co-transfecting three plasmids: a transfer plasmid, a packaging plasmid, and an envelope plasmid expressing SARS-CoV-2 S protein, into 293 T cells (ATCC® CRL-3216™). The plasma and immunoglobulins were serially diluted and incubated with pseudotyped virus for 1 h at 37°C. Subsequently, the mixture was added to 293 T/ACE2 in 96-well plates for infection of cells. After 48 h, luciferase activity in the cell lysates was examined with a luciferase assay system (Vazyme, DD1201). The titers of neutralizing antibodies (Nabs) were calculated as 50% inhibitory concentration (IC50), expressed as the highest dilution of plasma or immunoglobulins which resulted in a 50% reduction of luciferase luminescence compared with virus control wells.

RNA Extraction and RNA Sequencing

The total RNA of whole blood from 1 COVID-19 patient with SARS-CoV-2 RNA retested positivity, nine mild-to-moderate (MM) patients, two severe patients, and four healthy donors were extracted using Trizol reagents (Invitrogen, USA). The sequencing libraries were prepared according to instructions of the Illumina TruSeq RNA Sample Preparation kit (Illumina, USA). Sequencing was performed on a NovaSeq 6000 system (Illumina, USA). These experiments were performed by DianLi Bio-Tec Co., Ltd (Guangzhou, China). The mRNA sequencing datasets have been deposited in the Genome Sequence Archive for Human (https://bigd.big.ac.cn/gsa-human/) with the accession number HRA000661.

Pre-Processing of Raw RNA-Seq data

Raw RNA-seq reads were filtered according to their base qualities, and read sequences were trimmed at the 3’-end after reaching a 2-base sliding window with a PHRED quality score lower than 20. Following filtering, Illumina adapter sequences at the 3’-end were removed using Trimmomatic v0.36 (18). After low-quality filtering and adapter trimming, reads of more than 50 nt in length were retained for further analysis. Next, the trimmed reads were mapped to the human genome reference hg38 using HISAT v2.1 (19) with corresponding gene annotations (Gencode GRCh37/V32 for the human genome) at default settings. Total counts per mapped gene were determined using featureCounts function in SubReads package v1.5.3 (20) with default parameter.

Global Transcriptome Analysis and Digital Cytometry

To determine the difference of whole blood transcriptome between COVID-19 patients with different severity, normalized counts matrix was used to calculate pairwise person coefficients using the build-in function cor in R software. Hierarchical clustering across all samples was based on pairwise Pearson correlation coefficients matrix. The build-in function prcomp in R software was used for principal component analysis (RCA). We inferred the immune cell quantities in each blood sample using the CIBERSORT server (https://cibersortx.stanford.edu/).

Identification of Differentially Expressed Genes and Functional Enrichment Analysis

Raw counts matrix obtained from featureCounts were used as input for differentially expressed gene analysis using the Bioconductor package DESeq2 v1.26 (21) in R v3.6. Gene counts of more than five reads in a single sample or more than 100 total reads across all samples were retained for further analysis. Filtered counts matrix was normalized using the DESeq2 method to remove the library-specific artefacts. Genes with log2 fold change >1 or <−1 and adjusted p-value <0.05 corrected for multiple testing using the Benjamini–Hochberg (BH) method were considered significant. Gene ontology analysis was performed to assess their biological relevance using R package clusterprofiler v3.14.3 (22).

Statistical Analysis

Statistical analyses and graphical presentations were conducted using GraphPad Prism v5.0 (GraphPad Software, Inc., CA, USA). Correlation analyses were performed using the Spearman correlation test. Differences of immune cell counts between the re-tested positivity (RP) patient, non-RP MM patients, and severe patients, NAbs titer between anti-spike/anti-RBD-antibody IgA or IgG positive and negative group were determined using the Wilcoxon signed-rank test. p <0.05 was considered statistically significant.

Results

The Clinical Characteristics of COVID-19 Patients

From January 27, 2020, to March 5, 2020, there were 15 patients with confirmed COVID-19 in Maoming city ( ). By March 5, 2020, all the patients were discharged and continued self-isolation based on the discharge criteria of China’s COVID-19 epidemic prevention and control strategy (15). Among them, one patient was re-tested positive for SARS-CoV-2 RNA and was re-admitted to the hospital twice during post-discharge surveillance (from March 5, 2020, to July 1, 2020) ( and ). However, all the close contact family members of the re-admitted patient were negative for SARS-CoV-2 RNA and antibody following long surveillance (from January 21, 2020, to July 1, 2020), implying the re-tested SARS-CoV-2 RNA positivity of this patient was not caused by re-infection from his close contacts, and representative of SARS-CoV-2 RNA prolonged positivity.

Table 1

The clinical characteristics of the COVID-19 patients in Maoming City.

No.	Gender	Age	Clinical Classification	The time from Wuhan to Maoming	Diagnosis date	Discharge date	RNA positive duration (days)	Comorbidities	Symptoms	Pneumonia	PaO2/FiO2 (mmHg)
S1	Male	53	Severe	Jan.23,2020*	Feb.2,2020	Mar.1,2020	25	Hypertension	Chills and fever	Yes	235
S2	Male	53	Severe	Jan.21,2020	Feb.2,2020	Feb.18,2020	14	Hypertension	Fever, chills, cough	Yes	295
S3	Male	41	Severe	Jan.22,2020	Jan.27,2020	Feb.12,2020	13	No	Fever, stuffy nose and runny nose, muscle soreness, chills,	Yes	265
RP	Male	60	Moderate	Jan.22,2020	Feb.5,2020	Mar.5,2020	95	CAE, Hypertension	Fever	Yes	318
M1	Male	23	Moderate	Jan.21,2020	Feb.3,2020	Feb.27,2020	21	No	Cough, Weakness	Yes	476
M2	Male	40	Moderate	Jan.22,2020	Jan.26,2020	Mar.5,2020	37	No	Fever, Cough, Runny nose, Stuffy nose	Yes	391
M3	Female	68	Moderate	Jan.21,2020	Feb.7,2020	Mar.5,2020	25	Hypertension, DM	No	Yes	326
M4	Female	18	Moderate	Jan.20,2020	Feb.4,2020	Feb.18,2020	12	No	No	Yes	450
M5	Female	33	Moderate	Jan.19,2020	Feb.6,2020	Feb.18,2020	11	No	No	Yes	515
M6	Male	38	Moderate	Jan.21,2020	Feb.6,2020	Feb.27,2020	20	No	No	Yes	417
M7	Male	22	Moderate	Jan.22,2020	Feb.3,2020	Feb.14,2020	9	No	No	Yes	396
M8	Male	34	Mild	Jan.22,2020	Feb.13,2020	Mar.3,2020	16	No	Weakness	No	460
M9	Male	26	Mild	Jan.21,2020	Feb.3,2020	Feb.24,2020	30	No	Fever	No	396
M10	Male	16	Mild	Jan.22,2020	Jan.29,2020	Feb.7,2020	8	No	Fever	No	620
AI	Male	22	AI	Jan.22,2020	Jan.26,2020	Feb.7,2020	8	No	No	No	No data

*infected by YXY(S2); #infected by his sun LWX(M6); S, severe; M, mild-to-moderate; AI, asymptomatic infection; CAE, Coronary artery embolism; DM, Diabetes mellitus.

Figure 1

The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) reverse transcriptase-PCR test results and clinical therapy of the SARS-CoV-2 RNA re-tested positivity patient during hospitalization period.

The RP patient had a history of hypertension and coronary artery interventional therapy and was diagnosed with COVID-19 on February 5, 2020 ( ). On admission, the RP patient had no apparent symptoms and was diagnosed as moderate COVID-19 infection. He was treated with daily doses of anti-interferon α-2b, lopinavir/tropivir tablets, arbidol dispersible tablets and high flow oxygen supply from February 9, 2020, to February 21, 2020. On February 16, 2020, he had a transient high fever (38.8°C) and was treated with cefoperazone sulbactam and linezolid until the lesions in the right lower lung, which were first observed on February 8, 2020, was absorbed (February 21, 2020). The RP patient was discharged on March 5, 2020, based on discharged criteria guidelines (15). However, the patient was re-tested positive for SARS-CoV-2 RNA and was re-admitted to the hospital on March 14, 2020 ( and ). After treatment using Chinese traditional medicines, he was then discharged (March 18, 2020) based on continuous negative SARS-CoV-2 RNA results for 3 days. Appallingly, SARS-CoV-2 RNA was redetectable from oropharyngeal swab sample on April 30, 2020. Collectively, the intermittent viral RNA positivity of this patient lasted from February 5, 2020, to May 10, 2020 (~95 days), and was significantly longer than that of previously reported cases (23). The plasma IgA, IgG and IgM concentration of the RP patient was consistently normal during the SARS-CoV-2 RNA positive and negative stage ( ). Analysis of immune cell types and their proportions based on transcriptome data indicated that CD4+ and CD8+ T cells, and B cells of the patient at the recovery stage were similar with healthy controls (24) ( ). In addition, the RP patient was tested negative for HIV, HBV, HCV, syphilis, tuberculosis, and cancer; indicating that the patient was immunocompetent.

Longitudinal chest computed tomography (CT) demonstrated ground-glass opacities in the posterior segments of the right upper lobe, posterior segments of the right lower lobe, left lower lobe, and the lingual segment of the left upper lobe were observed in the early stage of the first hospitalization (from February 8, 2020, to February 16, 2020), disappeared gradually at the first discharge (February 25, 2020) ( ). However, the ground-glass opacities in the lower part of the posterior segments of the right upper lobe reappeared in the second (first SARS-CoV-2 RNA RP, March 14, 2020) and third admission (second SARS-CoV-2 RNA RP, May 1, 2020) and disappeared by the last discharge date (May 13, 2020) ( ). The patchy, linear, and ground-glass lesion in the right middle lobe was present from February 8, 2020, to February 25, 2020, disappeared in the second hospitalization (March 14, 2020) but re-appeared and was larger during the third hospitalization (from May 1, 2020, to May 13, 2020), then decreased by the return visit (June 24, 2020) ( ). Interestingly, new lesions in the basal segment of his left lower lobe emerged on the second admission (May 1, 2020) but vanished by the last discharge date (May 13, 2020) ( ).

Figure 2

Chest computed tomography scans of the COVID-19 patient with SARS-CoV-2 RNA re-tested positivity during the hospitalization period. (A) The ground-glass opacities in the upper part of the posterior segments of the right upper lobe. (B) The ground-glass opacity in the posterior segments of the right lower lobe. (C) The ground-glass opacity in the lower part of posterior segments of the right upper lobe disappeared in the first hospitalization, re-appeared in the second and third hospitalization (the first recurrent SARS-CoV-2 RNA positive and the second recurrent SARS-CoV-2 RNA positive). (D) The ground-glass opacity in the right upper lobe, left lower lobe, and lingual segment of the left upper lobe. (E) The lesion of the right middle lung displayed as patchy, ground-glass and linear lesion during the first hospitalization, disappeared during the second hospitalization, re-appeared and was enlarged during the third hospitalization, but was absorbed on the return visit (June, 24, 2020). (F) Appearance of the ground-glass opacity in the basal segment of the left lower lung near the diaphragmatic dome at third hospitalization (the second recurrent SARS-CoV-2 RNA positive).

Delayed NAbs Conversion and Anti-Nucleocapsid (NP) IgM Persistent Positive Are Related to Prolonged SARS-CoV-2 RNA Positive

The antibody levels in 13 SARS-CoV-2 RNA positive patients including one RP patient, nine MM patients (M1–M9), and three severe patients (S1–S3) were comparatively explored. Our data demonstrated that the anti-NP antibody IgG in three MM patients (M3, M4, and M9) were persistently negative. The M6 patient had early seroconversion of anti-NP antibody IgG but turned negative ten days later, while for the remaining patients, anti-NP antibody IgG was persistently positive ( ). M3, M8, and S3 patients had early seroconversion of anti-NP antibody IgA. The RP patient had delayed seroconversion of anti-NP antibody IgA while the remaining patients were persistently negative ( ). Three patients (S1, S2, and RP) had seroconversion of anti-NP antibody IgM, S1, and S2 patients, which turned negative in 7 and 10 days. The RP patient had persistently detectable anti-NP antibody IgM for 110 days (February 9, 2020, to May 25, 2020), while the remaining patients remained persistently negative ( ). Given the key role of NP in the SARS-CoV-2 genome replication (25), the long duration for anti-NP IgM positivity could suggest persistent proliferation of SARS-CoV-2 in the RP patient.

Figure 3

The dynamic change of anti-SARS-CoV-2 antibody in all COVID-19 patients. The absorbance ratio was calculated by the absorbance value of COVID-19 patients divided by 2.1 × mean absorbance value of four SARS-CoV-2 RNA negative plasma pool samples. (A) Anti-nucleocapsid (NP) IgG; (B) Anti-NP IgA; (C) Anti-NP IgM; (D) Anti-spike IgG; (E) Anti-spike IgA; (F) Anti-spike IgM; (G) Anti-receptor-binding domain (RBD) IgG; (H) Anti-RBD IgA; (I) Anti-RBD IgM. RP, re-tested positivity; M, Mild-to-moderate; S, Severe.

In addition, all the patients were persistently positive for anti-spike antibody IgG during their first hospitalization ( ). However, the anti-spike antibody IgA level varied between different patients, of whom two MM patients (M4 and M6) were persistently negative, other MM patients (M1, M2, M3, M5, M7, M8) had persistent positivity, two severe patients (S2 and S3) were persistently positive, and one severe patient (S1) was persistently negative during hospitalization. The RP patient had delayed seroconversion of anti-spike antibody IgA, compared to the other IgA positive patients ( ). As for anti-RBD antibody IgA and IgG, six MM(M1, M2, M3, M5, M7, M8) and one severe patient (S3) had early seroconversion ( ). Three MM (M4, M6, M9) patients and one severe patient (S1) were persistently negative for anti-RBD antibody IgA and IgG. One severe (S2) patient displayed early seroconversion of anti-RBD antibody IgG but was persistently negative for anti-RBD antibody IgA. The RP patient was positive for anti-RBD antibody IgG during the early infection stage but then had delayed seroconversion of anti-RBD antibody IgA ( ). Additionally, the majority of the COVID-19 patients had early seroconversion of anti-spike IgM antibody, with three MM (M3, M5, M8) patients being marginally positive ( ). Among them, three severe patients (S1, S2, and S3) and two MM (M6, M7) patients were positive for anti-RBD IgM antibodies. The RP patient was positive at the early infection stage and turned negative at the convalescent stage, while the others were negative for anti-RBD IgM antibody ( ).

The NAbs level of eleven patients on admission and discharge was further evaluated. Our results showed that the NAbs level varied between different patients. Particularly, the RP patient had low NAbs levels at the first and second discharge but had high NAbs levels at the third admission ( ). Interestingly, the high NAbs levels in the RP patient was accompanied with a sharp increase of anti-RBD/anti-Spike IgA and gradual rise of anti-RBD/anti-spike IgG ( ). The high NAbs levels of the RP patient lasted at least 78 days (from April 30, 2020, to July 17, 2020) ( ) and the NAbs levels in the RP patient was positively associated with anti-spike/anti-RBD IgA and IgG, and was negatively associated with anti-spike/anti-RBD IgM ( ). Correlation analysis of all COVID-19 patients further suggested that their plasma NAbs level was positively associated with anti-RBD or anti-spike IgG and IgA, but not with anti-spike/anti-RBD IgM ( ). Next, total IgA and IgG were extracted from the anti-RBD antibody IgA positive plasma ( ). Neutralization assay demonstrated that anti-RBD/anti-spike IgA and anti-RBD IgG antibodies had SARS-CoV-2 virus neutralizing capacity but the anti-spike IgG antibody positivity could not represent SARS-CoV-2 virus NAbs levels ( ).

Figure 4

The kinetic and correlation of anti-SARS-CoV-2 antibody and neutralizing antibody (NAbs) of all COVID-19 patients. (A) The dynamic change of the neutralizing antibody (NAbs) of all COVID-19 patients. (B–G) The correlation between NAbs level and anti-spike IgA(B), anti-spike IgG (C), anti-spike IgM (D), anti-receptor-binding domain (RBD) IgA (E), anti-RBD IgG (F), anti-RBD IgM (G) in all COVID-19 patients. The cut-off for the pseudovirus neutralization assay is 1:50. (H–I) The absorbance ratios of anti-spike an anti-receptor binding domain (RBD) specific antibody in the isolated total IgA (H) and IgG (I), that were calculated by the absorbance value of antibody isolated samples divided by 2.1 × mean absorbance value of anti-SARS antibody-negative samples (M6). (J) The neutralization antibody (NAbs) titers of the anti-spike/anti-RBD antibody IgA positive and negative groups. (K) The neutralization antibody (NAbs) titers of anti-spike/anti-RBD antibody IgG positive and negative groups. RP, re-tested positivity; MM, Mild-to-moderate; S, Severe.

Impaired T Cells and Delayed B Cells Response Are Related to Prolonged SARS-CoV-2 RNA Positivity

To investigate the immunological mechanisms underlying prolonged SARS-CoV-2 persistence, we further compared the immune signatures of one RP patient, nine MM patients, two severe patients and four healthy donors via RNA-seq of their whole blood ( ).

Figure 5

Comparison of immune signatures between the SARS-CoV-2 RNA re-tested positivity (RP) patient and Mild-to-moderate (MM) patients or severe patients. (A) Schema of the RNA-seq experimental design. A total of 12 COVID-19 patients and four healthy controls (HC) were included in this study. (B) Unsupervised principal component analysis based on global gene expression. (C) Heatmap of a subset of innate immunity (II) genes, cytokine and chemokine (CC)-related genes, B cells-related genes, T cells-related genes, antigen presentation (AP) genes (including MHC class I-related genes and MHC class II-related genes) and apoptosis-related genes. (D) Comparison of interferon-stimulated genes (ISGs) scores between the PP patient and MM patients or severe patients. Asterisks represent significant differences between groups (**p-value < 0.01, ***p-value < 0.001, unpaired t-test). (E) Receiver-operating characteristic curve of the clinical course prediction based on the expression of B cells-related genes and antigen presentation (AP)-related genes.

Unsupervised PCA revealed that the gene expression profile of whole blood in healthy controls formed a distinct cluster whereas that in COVID-19 patients formed clusters that were more widely dispersed and overlapped, indicating the immune heterogeneity in different COVID-19 patients ( ). Further, our results demonstrated that the RP patient and severe patients had more markedly declined expression of interferon-stimulated genes (ISGs) in the acute infection stage ( ). Additionally, the ISG score, i.e. the mean expression of six ISGs (IFI27, IFI44L, IFIT1, RSAD2, SIGLEC1, and ISG15) defining a type I IFN response was significantly lower in the RP patient and severe patients compared to MM patients ( ) (26). Consistent with previous report (27), we also observed overexpression of cytokines and chemokines (CC)-related genes in the severe COVID-19 patients during their early stage of infection, including pro-inflammatory cytokines or chemokines such as IL1A, IL1B, IL10, TNF, CCL20, CCL3, CCL4, CXCL1, CXCL2, CXCL3, and CXCL8 ( ). In contrast to severe COVID-19 patients, the MM patients or RP patient exhibited a lower expression of CC-related genes ( ).

Interestingly, a delayed B cells response was found in the RP patient, contrasting to a robust rapidly induced B cells response in MM patients. As for severe patients, the B cells-related genes were at a consistently low expression ( ). Notably, the expression of several immunoglobulin heavy chain variable region genes such as IGHV3-23, IGHV3-7, IGHV1-18, IGHV3-53, and IGHV5-51 were significantly lower in the RP patient and severe patients than in MM patients ( ). Particularly, IGHV3-53, the most frequently used IGHV gene for targeting the receptor-binding domain (RBD) of the spike protein (28), was lowly expressed in the RP patient. T cells-related genes in the RP patient were lowly expressed, compared to healthy controls, until SARS-CoV-2-negative time points. Unlike the RP patient, lower expression of the T cells-related genes only lasted for 10–12 days in the severe patients, whereas a continuous relative normal T cells response was observed in most MM patients. In addition, T cells exhaustion was observed in severe COVID-19 patients but not in the RP or MM COVID-19 patient, as supported by the distinct expression pattern of T cells exhaustion-related genes such as programmed cell death 1 (PD-1, PDCD1) and Tim-3 (HAVCR2) ( ).

Compared to healthy controls, we observed an up-regulation in both MHC class I and MHC class II gene in the RP patient during the SARS-CoV-2-negative period, whereas this observation was not found during the SARS-CoV-2-positive period. In contrast to the RP patient, severe patients had decreased expression of MHC class II genes ( ). Notably, we observed an increased expression of GZMB and PRF1 in the RP patient during his convalescent stage ( ) which acted as cytotoxic effectors of CD8+ T cells, CD4+ T cells, and NK cell (29). In addition, genes related to Fas- or TNF-linked apoptosis pathways showed a higher expression in the RP and severe patients compared to MM patients during the SARS-CoV-2-positive period ( ). Moreover, the B cells proportion at the SARS-CoV-2-positive stage in the RP patient was significantly lower than healthy controls, MM and severe patients ( ). The CD4+ T cells subsets proportions were increased in all three clinical-classification patients at viral RNA positive stage, and returned to normal in the RP patient during viral RNA negative stage ( ). The CD8+ T cells subsets showed different degrees of decline in all three clinical-classification patients at viral RNA positive stage, and returned to normal in the RP patient during viral RNA negative stage ( ).

To determine whether these gene expression patterns could predict clinical course, we performed a prediction analysis based on the expression of the six gene modules. We found that the expression of B cells-related genes could distinguish the RP patient from MM patients, but not from the severe patients ( ). Interestingly, the expression of MHC-related genes could distinguish the RP patient from MM or severe patients, i.e. the downregulation of MHC-related genes was related to severe COVID-19, while the upregulation of MHC-related genes was associated with MM COVID-19 and unaltered expression of MHC-related genes was linked to prolonged SARS-CoV-2 RNA shedding ( ).

Longitudinal Transcriptome Analysis Revealed the Dynamic Immune Response in the Patient With Prolonged SARS-CoV-2 RNA Positive

We then investigated the kinetic immune response in the RP patient by RNA sequencing (RNA-seq) of his whole blood samples covering five-time points (TP1-TP5) ( ). Correlation analyses showed a distinct global gene expression profile between SARS-CoV-2-positive (TP1-TP3) and SARS-CoV-2-negative (TP4-TP5) samples ( ). Compared to TP4 and TP5, we identified 1,418 different expression genes (DEGs) at TP1, 2461 DEGs at TP2, 4,092 DEGs at TP3 respectively, indicating a dramatically disturbed transcriptome homeostasis of peripheral blood cells at SARS-CoV-2 RNA positive stage ( ). Functional enrichment analysis showed that the up-regulated genes were mainly associated with cytokine related signaling pathway and mRNA processing whereas the down-regulated genes were related to adaptive immune and humoral immune response ( ).

Figure 6

Longitudinal transcriptional profile of whole blood in the SARS-CoV-2 RNA re-tested positivity (RP) patient. (A) Samples of whole blood collected at five different time points were used for RNA sequencing. (B) Heatmap of the correlation coefficient matrix across different time points. (C) Heatmap of the unique different expression genes (DEG) identified over time. (D) Functional enrichment analysis of up-regulated (top) or down-regulated gene (bottom) at TP1-TP3 comparing to TP4 and TP5. Color denotes statistical significance as measured by minus logarithm of p-values. (E–H) Heatmap of DEGs related to innate immunity response (E), cytokines or chemokines (F), B cells and T cells (G), and antigen presentation (H) during viral positive and negative period.

The innate immunity genes involving in virus detection (DDX58, IFIH1, TLR2, and TLR4) and induction of antiviral effectors (IFIT1, IFIT2, IFIT3, IFI44L, and RSAD2) were upregulated in the RP patient during viral RNA positive stage ( ). This result indicated that IFN-mediated broad-spectrum antiviral functions were induced in the RP patient. In addition, an increased production of pro-inflammatory mediators such as CXCL8, IL1B, CXCL1, CXCL2, CCL3, and TNF were observed during the viral RNA positive stage ( ), suggesting an unquenched inflammation response during viral persistence in the RP patient.

Compared to SARS-CoV-2-negative time points (TP4 and TP5), we observed a lower expression of B cells-related genes, including B cells transcription and activation regulators related genes (PAX5, MZB1, CD27, CD79A, and CD79B), and immunoglobulin producible genes (IGHG1, IGHA1, IGKC1, and IGLC1) in SARS-CoV-2-positive time points (TP1-TP3) ( ). In addition, the expression of T cells-related genes (CD4, CD3E, CD8A, and CD8B) were lower in the viral RNA positive stage ( ). Furthermore, the expression of MHC class I genes (HLA-A, HLA-B, HLA-C, HLA-E, and HLA-H) and MHC class II genes (HLA-DPB1, HLA-DQA1, HLA-DAQ2, HLA-DQB1, HLA-DRB1, and HLA-DRB5) were lower in SARS-CoV-2-positive samples (TP1–TP3) ( ). Further, we found that the expression of MHC, B cells and T cells-related genes was negatively correlated with viral RNA concentration and the CC-related genes were positively associated with viral RNA concentration ( ).

Discussion

In general, it is reported that SARS-CoV-2 RNA shedding last for 18–25 days in immunocompetent patients (30). In this present study, the immunocompetent RP patient experienced ~95 days of prolonged virus RNA shedding; longer than the longest previously reported 56 days period (23). The RP patient had a new lesion in his lung during the re-detectable RNA positive stage (the second and third admissions) but all the close contact family members of the patient were not infected and were tested negative for SARS-CoV-2 RNA and antibody following a long surveillance; suggesting the previous infected SARS-CoV-2 isolate was not completely cleared during the first discharge. Nevertheless, the RP patient could not be excluded for re-infection of novel mutated SARS-CoV-2 isolates because prolonged infection of SARS-CoV-2 could accelerate its mutation in immunocompromised hosts (5). In contrast to the RP patient in this study, previous study showed that 44 SARS-CoV-2 RNA re-positive cases had no ground glass opacities during viral RNA re-positive period (15), implying the immunity mechanisam underlying viral RNA re-tested positivity in RP patient with novel lung lessions in our study might not be generalized to other RP patients without novel lung lessions.

Although, an upregulation of anti-spike antibody IgA, anti-RBD antibody IgA and NAbs was followed by the complete clearance of SARS-CoV-2 RNA and the lung lesion in the immunocompetent patient during the last hospitolization, the times of seroconversion of anti-RBD IgA and NAbs were indefinite because of a long sampling void between the second (March 5, 2020) and third hospitalization (May 1, 2020). If seroconversion of NAbs occurred during the sampling viod period, suggesting NAbs was not crucial in the clearance of the lung lesions. However, previous data (31) and our experimental data confirmed that anti-RBD antibody IgG and the anti-RBD/anti-spike antibody IgA content was correlated with virus-neutralizing capacity. A recent study demonstrated an inverse correlation between the nasopharyngeal swabs viral load and the plasma level of anti-S1 IgA in a severe patient (32). Moreover, the plasma titer of anti-S1 IgA was a protective factor of day-28 mortality of severe patients (32). Together, the anti-RBD antibody IgA and IgG level are important immunity response evaluation biomarker for SARS-CoV-2 vaccine development (33).

Of note, the delayed plasma anti-SARS-CoV-2 antibody IgA conversion and NAbs production occurred in the RP patient. Moreover, lower levels and slower generation of anti-RBD-specific IgA and IgG were reported to contribute to SARS-CoV-2 RNA persistence in the gastrointestinal tract (34). This unique phenomenon may be related to lagged activation of the MHC class II-mediated antigen presentation-related genes upon SARS-CoV-2 infection in the RP patient. In addition, the unchanged expression of antigen presentation-related genes in the early stage of infection could be a prediction of delay SARS-CoV-2 clearance. On the other hand, the RP patient had lower B cells during his acute infection stage than non-RP MM patients. All these could have contributed to the observed delayed humoral immune response.

The IFN response occurred in the RP patient was lower than in the MM COVID-19 patients and higher than in severe COVID-19 patients, suggesting that limited IFN response was insufficient to timely clear the virus. Moreover, the CC-related gene expression of the RP patient was lower than severe patients, not different from non-RP MM patients. The low CC-related genes explain why the RP patient did not progress to severe disease (35).

Herein, evidences indicated severe patients had the lowest T cells gene expression and recovered in a short time. Accordingly, exhaustion-related genes were most upregulated in severe patients as compared to healthy controls, implying that the dysregulated function of T cells could be related to disease severity (36). However, a sustained low expression of T cells-related gene and overexpression of exhaustion-related genes in the SARS-CoV-2 positive stage of the RP patient, indicating that impaired T cell responses also contributed to viral persistence. Theoretically, the impaired T cell response may be attributed to the downregulated MHC class I genes, upregulated Fas- or TNF-linked apoptosis pathways, and decreased B cells and CD8+ T cells in the RP patient. In the convalescent stage, the RP patient demonstrated higher expression of T cells, MHC class I and II related genes, and upregulation of NAbs; suggesting coordinated T cells immunity and humoral immunity response played a crucial role in viral clearance of the RP patient at the convalescent stage (13).

In conclusion, our study showed that low innate immune response and prolonged low antigen presentation capacity were related to the re-tested SARS-CoV-2 RNA positivity in this immunocompetent patient. In the convalescent stage of this RP patient, upregulation of the antigen presentation gene might have contributed to the coordinated humoral and cellular immune response and led to SARS-CoV-2 clearance.

Data Availability Statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: Genome Sequence Archive for Human, accession no: HRA000661.

Ethics Statement

The studies involving human participants were reviewed and approved by the Ethics Committee of Maoming People’s Hospital (No. 2020012). The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

Author Contributions

Conceptualization: TJ. Clinical sample and data collection: CL, HL, LZ, YC, YFC, WH, FX, LHH, YL, JH, LL and, HZ. Methodology devising and Performing experiments and Data analysis: TJ, XL, QY, LBH, ZL, and YX. Writing original draft: TJ, QY, XL, GW, and QZ. Funding acquisition: TJ, CC, QZ, GW, and HL. Supervision: LC, ZZ, CC, and TJ. All authors contributed to the article and approved the submitted version.

Funding

This study received the following funding: Guangdong Science and Technology Project (2014A020212331), Guangdong Provincial Department of Education Youth innovative talents Project (2017KQNCX168), Guangzhou Health Science and technology project (20201A011078), Guangzhou Science and Technology Project (201904010214, 202102010094), Emergent Science and Technology Project for Prevention and Treatment of COVID-19 disease (#mmkj003) and the High-level Hospital Construction Research Project of Maoming People’s Hospital (zx2020012), Maoming Science and Technology Project (2020007), Guangdong Basic and Applied Basic Research Foundation (2021A1515012550), Natural Science Foundation of Guangdong Province (2019A1515011681), and Guangzhou Institute of Respiratory Health Open Project (2020GIRHHMS01).

Conflict of Interest

YX was employed by Guangdong South China Vaccine Co. Ltd.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.