Cellular and humoral immunity after the third dose of SARS-CoV-2 vaccine in patients treated with rituximab

Humoral and cell-mediated immune responses are blunted after SARS-CoV-2 vaccination in patients with a history of CD20 B-cell-depleting treatment.¹ However, vaccination induces SARS-CoV-2-specific antibodies in patients treated with rituximab once peripheral B cells at least partially repopulate.² Moreover, SARS-CoV-2-specific T cells, which have been found in 58% of patients who have had two doses of SARS-CoV-2 vaccine (either mRNA-1273 [Moderna] or BNT162b2 [tozinameran; Pfizer–BioNTech),² might exert protective effects independent of antibody responses. The question of repeat vaccine doses for serological non-responders to induce more robust immunological responses has been raised.³ However, little is known to date regarding response to a third dose of vaccine in patients treated with rituximab, who are particularly prone to develop severe COVID-19.⁴

In France, a systematic third dose of vaccine with mRNA-1273, BNT162b2, or ChAdOx1 nCoV-19 (Oxford–AstraZeneca) was recommended in highly immunocompromised patients (including patients treated with rituximab, mycophenolate mofetil, or cyclophosphamide) at least 1 month after the second dose, with no requirement for assessing serological response before the third dose. Given the uncertainties about the effectiveness of this measure in patients in whom B lymphocytes are not repopulated at the time of the third dose, we investigated the course of humoral and cellular immunity against SARS-CoV-2 in ten patients treated with rituximab after two and three vaccine doses. These patients were recruited from our day hospital, and the only inclusion criterion was that they had not received a third dose of vaccine. Blood samples were taken from patients twice: just before the third vaccine dose (reflecting their immunity after two doses) and 1 month after the third dose. Samples were stored at −150°C. Analyses of all samples from the same patients were done at the virology laboratory of the University Hospital of Strasbourg at the same time by virologists (FG and SF-K) who were masked to the patients' characteristics and to the date of sampling. To explore the SARS-CoV-2-specific T-cell response, IFN-γ enzyme-linked immunospot (ELISPOT) assay was performed (appendix p 3).⁵ For each sample tested, we had a negative control sample against which T-cell response was determined. Positivity was defined as a T-cell response at least 3 SDs larger than that of the negative control against which it was tested. Serology was done using the SARS-CoV-2 IgG II Quant (Spike) commercial assay (Abbott Architect, Chicago IL, USA) that quantifies anti-RBD IgG with 7·1 binding arbitrary units per mL (50 arbitrary units per mL) as a positive cutoff (appendix p 3). Neutralising antibody titres were measured for each serum sample using an in-house viral pseudoparticle-based assay. Results were expressed as the log10 of the sample dilutions that yielded 50% inhibition of pseudoparticle infectivity (log10 IC50). Serum samples were considered neutralising if the 1/40 dilution (1·60 log10) mediated at least a 50% luminometric signal reduction relative to the control condition without serum sample. B-cell depletion was assessed in each patient by phenotyping lymphocyte subpopulations just before the third vaccine dose using routine flow cytometry.Verbal informed consent was obtained from patients. The study was approved by the ethics review board of Strasbourg medical faculty (number CE-2021-103).

Between May 10 and June 22, 2021, we recruited ten patients, of whom eight (80%) were women and two (20%) were men, with a median age of 72·0 years (IQR 68·5-78·0). All patients were treated with rituximab for rheumatoid arthritis except one, who was treated for stiff-person syndrome. Five (50%) had concomitant methotrexate and three (30%) had concomitant steroids. No patient had a change in concomitant disease modifying antirheumatic drug (DMARD) or corticosteroid treatment between the first and third dose of vaccine. Patients had previously received a median of 5 cycles (IQR 3–11) of rituximab. Median time between last rituximab infusion and first dose of vaccine was 227 days (IQR 140–278). Median time between second and third dose of vaccine was 65 days (IQR 56–70). For their two first doses, five patients were vaccinated with BNT162b2, four with mRNA-1273, and one with ChAdOx1 nCoV-19. All received a third dose of either mRNA-1273 or BNT162b2. Patients' demographic and clinical characteristics are shown in the table and virological results are in the appendix (p 2).

After two doses of vaccine, three (30%) of ten patients were seronegative (patients 2, 6, and 8), and one had no detectable T-cell response (patient 10; appendix p 1). Only one patient had detectable neutralising antibodies (patient 4; table ). As previously reported,² humoral and cellular responses were dissociated, with two seronegative patients (patients 2 and 6) having a detectable T-cell response, of whom one had more than 1000 spot forming units per million peripheral blood mononuclear cells (patient 6). Conversely, five seropositive patients (patients 3, 5, 7, 9, and 10) had a quite weak T-cell responses; although no protective T-cell protective threshold could be defined (appendix p 1).

Table.

Patients' characteristics and laboratory results

		Patient 1	Patient 2	Patient 3	Patient 4	Patient 5	Patient 6	Patient 7	Patient 8	Patient 9	Patient 10
Sex		Female	Male	Female	Female	Female	Female	Female	Female	Female	Male
Age, years		82	74	73	77	82	71	70	39	70	64
Biologic treatment		Rituximab	Rituximab	Rituximab	Rituximab	Rituximab	Rituximab	Rituximab	Rituximab	Rituximab	Rituximab
Dose of last rituximab therapy		500 mg	500 mg	500 mg (× 2)	1000 mg	1000 mg	500 mg (× 2)	500 mg	500 mg (× 2)	500 mg	1000 mg
Associated conventional synthetic DMARDs		NA	Oral methotrexate 20 mg per week	Subcutaneous methotrexate 15 mg per week	NA	NA	NA	NA	Subcutaneous methotrexate 25 mg per week	Subcutaneous methotrexate 12·5 mg per week	Subcutaneous methotrexate 25 mg per week
Corticosteroids		Prednisone 10 mg per day	Prednisone 10 mg per day	NA	NA	NA	NA	NA	Prednisone 20 mg per day	NA	NA
Condition		Rheumatoid arthritis	Rheumatoid arthritis	Rheumatoid arthritis	Rheumatoid arthritis	Rheumatoid arthritis	Stiff-person syndrome	Rheumatoid arthritis	Rheumatoid arthritis	Rheumatoid arthritis	Rheumatoid arthritis
Number of anterior rituximab therapy cycles		5	5	6	11	4	2	14	2	12	3
Vaccine administered for first two doses		mRNA-1273 (Moderna)	mRNA-1273 (Moderna)	mRNA-1273 (Moderna)	mRNA-1273 (Moderna)	BNT162b2 (Pfizer–BioNTech)	BNT162b2 (Pfizer–BioNTech)	BNT162b2 (Pfizer–BioNTech)	BNT162b2 (Pfizer–BioNTech)	BNT162b2 (Pfizer-–BioNTech)	ChAdOx1 nCoV-19 (Oxford–AstraZeneca)
Vaccine administered as third dose		mRNA-1273 (Moderna)	mRNA-1273 (Moderna)	mRNA-1273 (Moderna)	mRNA-1273 (Moderna)	BNT162b2 (Pfizer–BioNTech)	BNT162b2 (Pfizer–BioNTech)	BNT162b2 (Pfizer–BioNTech)	BNT162b2 (Pfizer–BioNTech)	BNT162b2 (Pfizer–BioNTech)	BNT162b2 (Pfizer–BioNTech)
Time between last infusion and first dose of vaccine (days)		218	6	384	183	277	84	235	158	279	277
Time between second and third dose of vaccine (days)		110	105	98	98	91	91	74	74	94	81
At the time of third dose
	IgG concentration, g/L	7·8	10·5	6·0	13·1	15·3	11·5	6·8	7·5	4·6	17·5
	B lymphocyte, cells per μL	36	0	20	0	1	0	86	1	81	95
	T lymphocyte, cells per μL	1917	740	906	1310	573	1227	1195	1253	677	581
	CD4 T lymphocyte, cells per μL	1054	551	847	740	354	866	936	868	562	504
	CD8 T lymphocyte, cells per μL	792	172	50	525	210	337	244	393	84	86
Anti-SARS-CoV-2 serology
	Day of third dose	Positive	Negative	Positive	Positive	Positive	Negative	Positive	Negative	Positive	Positive
	1 month after third dose	Positive	Negative	Positive	Positive	Positive	Negative	Positive	Negative	Positive	Positive
Neutralising antibodies
	Day of third dose	Negative	Negative	Negative	Positive	Negative	Negative	Negative	Negative	Negative	Negative
	1 month after third dose	Positive	Negative	Negative	Positive	Negative	Negative	Positive	Negative	Positive*	Positive

DMARDs=disease-modifying antirheumatic drugs. NA=not applicable.

^*Only for D614G Strain and negative for B.1.351.

After the third vaccine dose, all three previously seronegative patients (patients 2, 6, and 8) remained seronegative and had only a slightly increased T-cell response after three doses compared with before the third dose (table; appendix pp 1–2). All three non-responders to the third dose had complete B cell depletion at the time of this dose. The only patient (patient 10) who was previously seropositive but did not have a detectable T-cell response developed a T-cell response after the third dose. After the third dose, a neutralising antibody response developed in three additional patients who were previously seropositive but without detectable neutralising antibodies after two doses (table; appendix p 2).

We found that some patients who have been treated with rituximab can develop anti-SARS-CoV-2 humoral response after two vaccine doses. The third dose might help some of these patients to acquire a neutralising antibody response or to develop a T-cell response, or both. All seronegative patients and those with B-cell depletion after two doses remained seronegative after the third dose. The seven patients who were seropositive after three doses generally had more circulating B cells at the time of the third dose than the three seronegative patients (median of 36 cells per μL [IQR 1–86] vs 0 cells per μL [0–1]; Wilcoxon test p=0·064). We found a significant correlation between B-cell count at the time of the third dose of vaccine and neutralising antibody concentrations 1 month after the third dose (Spearman's r ² 0·81; p=0·0009). We found no statistical correlation between T-cell count or T-cell subsets in terms of humoral or cellular response (appendix p 4).

Our results, which must still be confirmed in ongoing larger studies due to the small sample size of the present study, support delaying the third dose of SARS-CoV-2 vaccine until B-cell repopulation. In seronegative patients, assessment of T-cell responses might be useful to identify patients with neither humoral nor cellular immunity. In seronegative patients treated with rituximab, the crucial questions to be further assessed concern the protection conferred by isolated cellular responses and the optimal strategy between waiting for B-cell repopulation, prophylactive anti-SARS-CoV-2 monoclonal antibody therapy in those who are highly immunocompromised (currently recommended in France⁶), or curative anti-SARS-CoV-2 monoclonal antibody therapy in cases of SARS-CoV-2 infection.

There was no funding for this study. LA declares consulting fees from AstraZeneca, Janssen, and BMS, unrelated to the current work. All other authors declare no competing interests. RF and FG contributed equally and have directly accessed and verified the underlying data. All authors contributed to the concept, design and drafting of the study and have approved the final version.