PD-1 blockade counteracts post–COVID-19 immune abnormalities and stimulates the anti–SARS-CoV-2 immune response

A substantial proportion of patients who have recovered from coronavirus disease-2019 (COVID-19) experience COVID-19–related symptoms even months after hospital discharge. We extensively immunologically characterized patients who recovered from COVID-19. In these patients, T cells were exhausted, with increased PD-1+ T cells, as compared with healthy controls. Plasma levels of IL-1β, IL-1RA, and IL-8, among others, were also increased in patients who recovered from COVID-19. This altered immunophenotype was mirrored by a reduced ex vivo T cell response to both nonspecific and specific stimulation, revealing a dysfunctional status of T cells, including a poor response to SARS-CoV-2 antigens. Altered levels of plasma soluble PD-L1, as well as of PD1 promoter methylation and PD1-targeting miR–15-5p, in CD8+ T cells were also observed, suggesting abnormal function of the PD-1/PD-L1 immune checkpoint axis. Notably, ex vivo blockade of PD-1 nearly normalized the aforementioned immunophenotype and restored T cell function, reverting the observed post–COVID-19 immune abnormalities; indeed, we also noted an increased T cell–mediated response to SARS-CoV-2 peptides. Finally, in a neutralization assay, PD-1 blockade did not alter the ability of T cells to neutralize SARS-CoV-2 spike pseudotyped lentivirus infection. Immune checkpoint blockade ameliorates post–COVID-19 immune abnormalities and stimulates an anti–SARS-CoV-2 immune response.

KGaA, Darmstadt, Germany), influenza vaccine 1% v/v (FLU, Vaxigrip, Sanofi Pasteur, Paris, France), or a trivalent diphtheria, tetanus and pertussis vaccine 1% v/v (DTaP, Polio boostrix, GSK, Brentford, UK), either individually or with the addition of anti-human PD-1 monoclonal antibody (mAb) pembrolizumab (Keytruda, MSD, Kenilworth, NJ, USA) at the working concentration of 5 μg/ml. Cells were then collected and plated on immobilon-P white plates (Merck) precoated with an anti-IFN-g capture antibody (Ab) (BD Biosciences). After 24 hours of incubation at 37 o C, cells were discarded, and a biotinylated anti-IFN-g detection Ab (BD Biosciences) was added to each well of the plate, which was then incubated overnight at 4 o C.
IFN-g release was detected by adding horseradish peroxidase-conjugated streptavidin and 3amino-9-ethylcarbazole (AEC) substrate (BD Biosciences) as specified in the manufacturer's instructions. Finally, the number of spots generated by IFN-g-releasing cells were counted using an ImmunoSpot Reader (CTL Europe GmbH, Bonn, Germany). The number of spots observed in unstimulated cells were considered background and were subtracted from the number of spots observed in stimulated cells. Values are presented as number of spots normalized to 10x10 6 PBMCs. In order to test for the anti-SARS-CoV-2 specific ex vivo T cell response, 3x10 5 PBMCs freshly isolated from patients and healthy controls were cultured for 48h in RPMI 1640 plus 10% FBS complete medium in the presence of a pool of SARS-Cov-2 spike and nucleocapsid proteins (Novatein Biosciences, Hudson, MA, USA; 0.5 µg/ml each) for 48 hours, with or without pembrolizumab (5 μg/ml) and using an anti-human IgG antibody as a negative control (Yumab, Braunschweig, Germany). Cells were then collected and processed as described above.

Cytokine treatment of PBMCs in vitro
The exhaustion and costimulatory profiles of PBMCs isolated from healthy controls were analyzed following exposure to the cytokines IL-1β, IL-1RA, IL-6, IL-8 and IP-10. 1x10 6 PBMCs isolated from 5 healthy individuals were cultured in 96-well plates in RPMI 1640 medium containing 10% human serum comprised of a pool of 5 healthy individuals' heatinactivated human sera. Human recombinant IL-1β (15 pg/ml), IL-6 (10 ng/ml), IL-8 (25 pg/ml) and IP-10 (1 ng/ml) (all from R&D, Minneapolis, Min, USA) were added to the culture either individually or as a pool. The cytokine concentration in each treatment was chosen according to the average plasma concentration levels observed in patients with COVID-19. Each treatment was performed in 5 replicates using PBMCs obtained from 5 different healthy controls. After 24 hours of incubation, cells were recovered and prepared for flow cytometric analysis using a panel of costimulatory molecules and exhaustion markers as described above.

In vitro cytokine inhibition
To investigate the immunological effect of the blockade of cytokines that are overrepresented in the serum of patients with COVID-19, PBMCs were isolated from 3 patients with COVID-19 and then cultured in complete medium containing a pool of heat-inactivated sera obtained from 5 patients with COVID-19, selected on the basis of their high serum cytokine levels. Then, antibodies directed against IL-1β, IL-1RA, IL-6, IL-8, IP-10, IL-7, and TNF-α (each at 10 μg/ml) and IL-1RA (20 μg/ml) (all from R&D) were added to cultured PBMCs, either individually or as a pool. In parallel, PBMCs were also cultured with a pool of heat inactivated sera obtained from 5 healthy individuals. After 24 hours of incubation, cells were recovered and prepared for flow cytometric analysis of a panel of costimulatory molecules and exhaustion markers, as detailed above.

Quantitative RT-qPCR analysis
The T cell exhaustion profile of CD4 + and CD8 + T cells was investigated at the transcriptional level using RNA extracted from CD4 + and CD8 + cells isolated by magnetic bead-based positive selection kits (Miltenyi Biotec, Bergisch Gladbach, Germany), according to the manufacturer's instructions. RNA was extracted from each cell population using the Direct-zol RNA Miniprep kit (Zymo Research, Irvin, CA, USA), samples were then quality-checked for RNA integrity by agarose gel electrophoresis, and RNA was then quantified and quality-checked for the presence of impurities using a Multiskan Go spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). 200 ng RNA was reverse-transcribed using a Quantinova Reverse Transcription Kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. cDNA for each sample was then analyzed by quantitative real-time PCR analysis using a SYBR greenbased Quantinova LNA Human T Cell Anergy & Immune Tolerance focus panel (Qiagen) according to the manufacturer's instructions. Relative quantification of each gene in the panel was calculated by an in house-developed and delta-delta-Ct method-based Qiagen online tool (www.qiagen.com/geneglobe), using averaged levels of ACTB, B2M, GAPDH, HPRT1 and RPLP0 genes for data normalization. For heatmap generation, values for each gene were normalized to the mean across groups, log-transformed and then mean-polished, normalized, and clustered using the Spearman rank correlation method with Cluster 3.0 software (http://bonsai.hgc.jp/).

Soluble PD-1 and soluble PD-L1 assessment
Soluble PD-1 (sPD-1) and PD-L1 (sPD-L1) plasma levels for patients with COVID-19, patients who recovered from COVID-19 and healthy controls were measured quantitatively by an ELISA immunoassay. Levels of sPD-1 and sPD-L1 were measured using 100 µl aliquots of plasma with two commercially available sandwich enzyme-linked immunosorbent assays (Cloud-Clone Corp., Houston, TX, USA) according to the manufacturer's recommendations.
The minimum detectable level for both sPD-1 and sPD-L1 was 0.156 ng/ml; samples for which signal was below the detectable level were assigned a value of 0 ng/ml.

PD1 promoter methylation analysis
PD1 promoter methylation status in CD4 + and CD8 + T cells was assessed by a bisulfite conversion-based quantitative real-time PCR method. CD4 + and CD8 + T cells were isolated from PBMCs by immunomagnetic cell separation methods as described above. DNA was extracted and bisulfite-converted from 50,000 cells of each sample using the EZ DNA Methylation-Direct Kit (Zymo Research). 25 ng converted DNA was then used as template in a SYBR green-based qPCR on a QuantStudio 7 Flex qPCR system (Thermo Fisher Scientific) using PowerUP SYBR Green reagent (Thermo Fisher Scientific) and primers targeting a PD1 promoter CpG methylation site located in chr2:241859917-241859918 (hg38 assembly) that is known to affect PD-1 function/expression (4,5). Relative DNA methylation of the PD1 locus was determined by comparison to total DNA as determined via ACTB reference using primers targeting a DNA region devoid of CpG methylation sites. Primer sequences were designed using the MethPrimer online tool (6)  Effector T cells