Cytokine release syndrome in a patient with colorectal cancer following vaccination with BNT162b2 (tozinameran)

We present a case of cytokine release syndrome (CRS) that occurred ve days after vaccination with BTN162b2 (tozinameran), an mRNA COVID-19 vaccine, in a patient with colorectal cancer on long-standing anti-PD-1 monotherapy. The CRS was evidenced by raised inammatory markers, thrombocytopenia, elevated cytokine levels (IFN-y/IL-2R/IL-18/IL-16/IL-10), and steroid responsiveness. BNT162b2

urine cultures were negative, as was SARS-CoV-2 RT-PCR of serial naso-pharyngeal swabs (Figure 1a). Computed tomography of thorax, abdomen and pelvis revealed no nidus of infection and stable disease.
Over the next ve days, fevers up to 39.8°C continued, with worsening thrombocytopenia (28x10 9 cells/l), increasing in ammatory markers (CRP: 317 mg/l, LDH 849 U/l), including signi cantly elevated ferritin (6,010 µg/l [normal range: 18-464]) (Figure 1b). At this point ( ve days post-admission), CRS was suspected and he was commenced on 1 mg/kg intravenous methylprednisolone (IVMP), and antibiotics were ceased three days later. Biochemical and haematological indices normalised within seven days of IVMP initiation (Figure 1b), and the patient was afebrile and asymptomatic upon discharge home with a weaning corticosteroid regimen. He remained well and was re-challenged with anti-PD-1 on 8 February 2021 (36 days after initial presentation) without any adverse events (Figure 1a).
To explore the features of his presentation further, we performed longitudinal cytokine analysis, pre-and post-IVMP. An exaggerated type 1 helper T-cell (Th1) response is a frequent feature of CRS, 1 and the initial pro le (day 3 of admission; Figure 2a To identify antibody-binding epitopes we performed a serum epitope repertoire analysis (SERA) and protein-based immunome wide association study (PIWAS), using a bacterial display system coupled with next-generation sequencing. Post-vaccine pro le was comparable to that of healthcare workers after COVID-19 mRNA vaccine (data not shown), with a rise in positive signal against spike but not non-spike proteins (vs SARS-CoV-2 positive subjects) 6 ( Figure S2). This was consistent with the lack of prior SARS-CoV-2 infection as a potentially contributor to the clinical presentation.
The laboratory ndings in CRS are variable and relate to the underlying cause, although CRP elevation is universal and correlates with severity. 1 Elevated ferritin and thrombocytopenia are also common abnormalities. 1 Whilst there are no de ned cytokine pro les that con rm CRS, raised IFN-y, IL-2R, IL-18, IL-6, and IL-10 are considered key in establishing the diagnosis. 1,2 All bar IL-6 was were elevated in this case. Whereas transient cytokine elevation (IFN-a, IFN-y, IL-6, IFN-inducible protein-10, and IL-12 p70) was observed following mRNA cancer vaccines co-administered with ICI in melanoma patients, 7 they manifest as self-limiting mild u-like symptoms.
Less than 0.01% of irAEs reported in the context of anti-PD-1 monotherapy involve CRS 4 and to date no CRS events have been reported following BNT162b2. ICI-related CRS typically develops a median of 4 weeks after ICI initiation (range: 1-18 weeks), 4 making CPI as the sole cause of CRS unlikely in this patient who commenced anti-PD1 treatment 22 months prior. The close temporal association of vaccination and clinical presentation favours the vaccine as potential trigger of CRS in this case.
Receptor occupancy associated with anti-PD-1 agents is 2-3 months 8 and it remains possible that CRS was triggered by the vaccine on a background of immune activation secondary to PD1 blockade which results in T-cell proliferation and increased effector function 9 . We did not detect S-reactive T-cells in the periphery, however vaccine-activated T-cells that contributed CRS could be resident within tissue or lymph nodes and evade detection in the blood. 10 T-cell cross-reactivity, as a results of sequence similarity between spike protein and tumour neoantigens is an alternative though a less likely cause of CRS in this case. Cross-reactivity to cardiac tissue was reported as a mechanism of ICI-related myocarditis 11 , and this patient's history of irAEs and the high neoantigen load (typical of MMRD CRC) 12 could in theory increase the likelihood of T-cell cross-reactivity. 13 Given patients with cancer were excluded from SARS-CoV-2 vaccine studies and are currently prioritised in COVID-19 vaccination programs globally, this case motivates prospective pharmacovigilance regarding the safety pro le of COVID-19 vaccines in cancer patients. However, this being an isolated case, the bene t-risk pro le for COVID-19 vaccination remains strongly in favour of vaccination in the cancer population who are genrally more vulnerable to COVID-19. 14,15 Current empirical reccommendations regarding the timing of COVID-19 vaccination suggest administering "on availability" in cancer patients on systemic anticancer treatments including ICI, cytotoxic chemotherapy, and hormone therapy; and avoiding vaccination within 48-72 h of investigational porducts to minimise misattribution of adverse event causation. 16 Methods CAPTURE design, study schedule, and follow-up During admission, the patient was enrolled in CAPTURE (NCT03226886), an observational prospective study of the immune response to SARS-CoV-2 in cancer patients that opened for recruitment in May 2020 at the Royal Marsden NHS Foundation Trust. The study design has been previously published. 17 Adult patients with current or history of invasive cancer are eligible for enrolment, irrespective of cancer type, stage, or treatment. Primary and secondary endpoints relate to patient characteristics of those with and without SARS-CoV-2 infection, and the impact of COVID-19 on long-term survival and ICU admission rates. Exploratory endpoints pertain to characterising clinical and immunological determinants of COVID-19 and vaccine response in cancer patients. Clinical data and sample collection for participating cancer patients are performed at baseline, and at clinical visits per standard-of-care management during the rst year of follow-up; frequency varies depending on in-or outpatient status and systemic anti-cancer treatment regimens. CAPTURE was approved as a substudy of TRACERx Renal (NCT03226886). TRACERx Renal was initially approved by the NRES Committee London -Fulham on January 17, 2012. The TRACERx Renal sub-study CAPTURE was submitted as part of Substantial Amendment 9 and approved by the Health Research Authority on April 30, 2020 and the NRES Committee London -Fulham on May 1, 2020. CAPTURE is conducted in accordance with the ethical principles of the Declaration of Helsinki, Good Clinical Practice and applicable regulatory requirements.

Handling of whole blood samples
For indicated experiments, serum or plasma samples were heat-inactivated at 56ºC for 30 minutes prior to use.

Plasma and PBMC isolation
Whole blood was collected in EDTA tubes (VWR) and stored at 4ºC until processing. All samples were processed within 24 hours. Time of blood draw, processing, and freezing was recorded for each sample. Prior to processing tubes were brought to room temperature (RT). PBMC and plasma were isolated by density-gradient centrifugation using pre-lled centrifugation tubes (pluriSelect). Up to 30 ml of undiluted blood was added on top of the sponge and centrifuged for 30 minutes at 1000 x g at RT. Plasma was carefully removed then centrifuged for 10 minutes at 4000 x g to remove debris, aliquoted and stored at -80ºC. The cell layer was then collected and washed twice in PBS by centrifugation for 10 minutes at 300 x g at RT. PBMC were resuspended in Recovery cell culture freezing medium (Fisher Scienti c) containing 10% DMSO, placed overnight in CoolCell freezing containers (Corning) at -80ºC and then stored in liquid nitrogen.

Serum isolation
Whole blood was collected in serum coagulation tubes (Vacuette CAT tubes, Greiner) for serum isolation and stored at 4ºC until processing. All samples were processed within 24 hrs. Time of blood draw, processing, and freezing was recorded for each sample. Tubes were centrifuged for 10 minutes at 2000 x g at 4ºC. Serum was separated from the clotted portion, aliquoted and stored at -80ºC. S1-reactive IgG ELISA Ninety-six-well MaxiSorp plates (Thermo Fisher Scienti c) were coated overnight at 4°C with puri ed S1 protein in PBS (3 μg/ml per well in 50 μl) and blocked for 1 hour in blocking buffer (PBS, 5% milk, 0.05% Tween 20, and 0.01% sodium azide). Sera were diluted in blocking buffer (1:50). Fifty microliters of serum were then added to the wells and incubated for 2 hours at RT. After washing four times with PBS-T (PBS, 0.05% Tween 20), plates were incubated with alkaline phosphatase-conjugated goat anti-human IgG (1:1000, Jackson ImmunoResearch) for 1 hour. Plates were developed by adding 50 μl of alkaline phosphatase substrate (Sigma Aldrich) for 15-30 minutes after six washes with PBS-T. Optical densities were measured at 405 nm on a microplate reader (Tecan). CR3022 (Absolute Antibodies) was used as a positive control. The cut-off for a positive response was de ned as the mean negative value multiplied by 0.35 times the mean positive value.
Neutralising antibody assay Con uent monolayers of Vero E6 cells were incubated with SARS-CoV-2 virus and two-fold serial dilutions of heat-treated serum or plasma samples starting at 1:40 for 4 hrs at 37ºC, 5% CO 2 , in duplicates. The inoculum was then removed and cells were overlaid with viral growth medium. Cells were incubated at 37ºC, 5% CO 2 . At 24 hours post-infection, cells were xed in 4% paraformaldehyde and permeabilized with 0.2% Triton X-100/PBS. Virus plaques were visualized by immunostaining, as described previously 18 for the neutralisation of in uenza viruses using a rabbit polyclonal anti-NSP8 antibody used at 1:1000 dilution and anti-rabbit-HRP conjugated antibody at 1:1000 dilution and detected by action of HRP on a tetramethyl benzidine-based substrate. Virus plaques were quanti ed and ID50 was calculated.
T-cell stimulation PBMC for in vitro stimulation were thawed at 37 ºC and resuspended in 10 ml of warm complete medium (RPMI, 5% human AB serum) containing 0.02% benzonase. Viable cells were counted and 2x10 6 cells were seeded in 200 µl complete medium per well of a 96-well plate. Cells were stimulated with 4 µl/well PepTivator SARS-CoV-2 S, M, or N pools (representing 1µg/ml nal concentration per peptide; Miltenyi Biotec, Surrey, UK). Staphylococcal enterotoxin B (Merck, UK) was used as a positive control at 0.5µg/ml nal concentration, negative control was PBS containing DMSO at 0.002% nal concentration. PBMC were cultured for 24 hrs at 37 o C, 5% CO 2 .
Activation-induced marker assay Cells were washed twice in warm PBMC. Dead cells were stained with 0.5 µl/well Zombie dye V500 for 15 minutes at RT in the dark, then washed once with PBS containing 2% FCS (FACS buffer). A surface staining mix was prepared per well, containing 2 µl/well of each antibody for surface staining (Supplementary Table 1) in 50:50 brilliant stain buffer (BD) and FACS buffer. PBMC were stained with 50 µl surface staining mix per well for 30 minutes at RT in the dark. Cells were washed once in FACS buffer and xed in 1% PFA in FACS buffer for 20 min, then washed once and resuspended in 200 µl PBS. All samples were acquired on a Bio-Rad Ze5 ow cytometer running Bio-Rad Everest software v2.4 and analysed using FlowJo v10 (Tree Star Inc.) analysis software. Compensation was performed with 20 µl antibody-stained anti-mouse Ig, k / negative control compensation particle set (BD Biosciences, UK). Up to 1x10 6 live CD19-/CD14-cells were acquired per sample. Gates were drawn relative to the unstimulated control for each donor. Gating strategy is shown in Figure S2c. T-cell response is displayed as a stimulation index by dividing the percentage of AIM-positive cells by the percentage of cells in the negative control. When S, M, and N stimulation were combined the sum of AIM-positive cells was divided by the three times the percentage of positive cells in the negative control.