Potent high-avidity neutralizing antibodies and T cell responses after COVID-19 vaccination in individuals with B cell lymphoma and multiple myeloma

Individuals with hematologic malignancies are at increased risk for severe coronavirus disease 2019 (COVID-19), yet profound analyses of COVID-19 vaccine-induced immunity are scarce. Here we present an observational study with expanded methodological analysis of a longitudinal, primarily BNT162b2 mRNA-vaccinated cohort of 60 infection-naive individuals with B cell lymphomas and multiple myeloma. We show that many of these individuals, despite markedly lower anti-spike IgG titers, rapidly develop potent infection neutralization capacities against several severe acute respiratory syndrome coronavirus 2 variants of concern (VoCs). The observed increased neutralization capacity per anti-spike antibody unit was paralleled by an early step increase in antibody avidity between the second and third vaccination. All individuals with hematologic malignancies, including those depleted of B cells and individuals with multiple myeloma, exhibited a robust T cell response to peptides derived from the spike protein of VoCs Delta and Omicron (BA.1). Consistently, breakthrough infections were mainly of mild to moderate severity. We conclude that COVID-19 vaccination can induce broad antiviral immunity including ultrapotent neutralizing antibodies with high avidity in different hematologic malignancies.

Individuals with hematologic malignancies are at increased risk for severe coronavirus disease 2019 (COVID- 19), yet profound analyses of COVID-19 vaccine-induced immunity are scarce. Here we present an observational study with expanded methodological analysis of a longitudinal, primarily BNT162b2 mRNA-vaccinated cohort of 60 infection-naive individuals with B cell lymphomas and multiple myeloma. We show that many of these individuals, despite markedly lower anti-spike IgG titers, rapidly develop potent infection neutralization capacities against several severe acute respiratory syndrome coronavirus 2 variants of concern (VoCs). The observed increased neutralization capacity per anti-spike antibody unit was paralleled by an early step increase in antibody avidity between the second and third vaccination. All individuals with hematologic malignancies, including those depleted of B cells and individuals with multiple myeloma, exhibited a robust T cell response to peptides derived from the spike protein of VoCs Delta and Omicron (BA.1). Consistently, breakthrough infections were mainly of mild to moderate severity. We conclude that COVID-19 vaccination can induce broad antiviral immunity including ultrapotent neutralizing antibodies with high avidity in different hematologic malignancies.
Coronavirus disease 2019 (COVID-19) results in increased morbidity and mortality in individuals with cancer [1][2][3][4][5][6] . Hematologic malignancies are frequently associated with secondary immunodeficiency, and affected individuals have a higher risk of experiencing severe COVID-19 than individuals with solid cancer, with reported odds ratios of 1.6 to 3. 3 (refs. 1 , 2). In a group of individuals with hematologic malignancies, those who had recent chemotherapy were at increased risk of death during COVID-19-associated hospital admission, with an odds ratio of 2.09 (ref. 1).
Authorized vaccines against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) are effective in preventing and mitigating the course of COVID-19 and in partially reducing viral transmission in immunocompetent individuals, inducing both robust humoral and T cell responses [7][8][9][10] . Recent data indicate that the presence of both binding and neutralizing antibodies may be predictive of protection against symptomatic disease 11 . A high-throughput neutralization assay using authentic, replication-competent viruses allowed us to gain deeper insight into humoral immunity against variants of concern (VoCs) in a longitudinal cohort of healthy individuals 12 . In this study, three timely spaced exposures to the spike protein of SARS-CoV-2 from either vaccination or infection resulted in increases in neutralization capacity per anti-spike antibody unit and increases in antibody

Longitudinal dynamics of anti-SARS-CoV-2 spike IgG
We first quantified vaccine-induced spike S1 domain-reactive antibodies before vaccination 1 (visit 1), at 2-8 weeks (visit 2) or 4-5 months (visit 3) after vaccination 2 and at 2-8 weeks after vaccination 3 (booster shot; visit 4) in infection-naive individuals with cancer. Individuals testing positive for nucleocapsid antibodies were excluded (Table  1). Baseline spike-specific antibody levels were low or undetectable (Extended Data Fig. 1a). Early after the second vaccination (visit 2), the median IgG titer of individuals with hematologic malignancies was 49.1-fold lower than that of healthy individuals (Extended Data Fig. 1b). At visit 3, a 23.6-fold difference was still observed between these two groups (Extended Data Fig. 1b). Notably, at visit 4, the factor of difference decreased to 9.2-fold (Extended Data Fig. 1b). A considerable interparticipant variability was observed, likely due to different Rx treatment regimens; at 2-8 weeks after vaccination 2, individuals with untreated LY and individuals with a longer interval since the last administration of Rx (Rx 12-60 months) showed reduced vaccine-induced median IgG levels that were 33-fold and 11.9-fold lower, respectively, than in matched healthy individuals (Fig. 1b,c). However, the 14 Rx < 12 individuals with LY showed low or negative anti-spike IgG titers after both the second and third vaccinations, consistent with recent studies 24,28,46 , whereas untreated individuals with LY mostly presented with measurable anti-spike IgG titers (Fig. 1c). A trend toward lower median anti-spike IgG concentrations was observed at visit 3 than at visit 2 in both healthy individuals (4.6-fold reduction) and individuals with hematologic malignancies (2.1-fold reduction; Fig. 1b).
Anti-spike IgG titers at visit 4 were markedly elevated in both groups compared to at visit 3 (Fig. 1b). Individuals with hematologic malignancies showed a 9.3-fold rise at visit 4 compared to titers observed at visit 3, and healthy individuals displayed a 4.9-fold increase. At visit 4, median anti-spike IgG concentrations in healthy individuals were 10.8-fold higher than in individuals with cancer (Extended Data Fig. 1b). In individuals with MM, the dynamics of vaccine-induced anti-spike IgG responses were comparable to those observed in healthy individuals, with a marked increase of anti-spike IgG titers from visit 3 to visit 4 (Fig. 1b,c). In these individuals, ongoing therapy had no significant influence on anti-spike IgG levels (Fig. 1c). Among the few individuals with chronic lymphocytic leukemia treated with novel agents, such as the Bruton's tyrosine kinase (BTK) inhibitor ibrutinib (N = 3) or the B cell leukemia/ lymphoma-2 (BCL-2) inhibitor venetoclax (N = 1), only one individual, an individual treated with ibrutinib, had a vaccine-induced anti-spike IgG response above the cutoff of 10 binding antibody units (BAU) per ml.
Taken together, 2-8 weeks after vaccination 2, 22 individuals with hematologic cancer showed negative or low-level anti-spike IgG levels, corresponding to 39% of the 57 individuals included in the statistical analysis, whereas all 25 matched healthy individuals displayed robust anti-spike IgG responses. None of the individuals with an Rx < 12 regimen developed an anti-spike IgG response markedly above background. Of the 35 individuals with hematologic cancer with detectable anti-spike IgG concentrations at visit 2, all still had a positive IgG response at visit 3 and a rise in antibody titers after booster vaccination.

Early increase in avidity of anti-spike IgG in individuals with hematologic cancer
Recently, we reported a maturation of anti-spike IgG avidity in healthy individuals over time and after each encounter with the SARS-CoV-2 spike protein (vaccine induced or infection related), especially after a third exposure to the spike protein 12 . Here, we quantified the avidity of serum IgGs binding to the SARS-CoV-2 spike 1 (S1) and spike 2 (S2) ectodomains. For individuals with hematologic malignancies, we observed avidity 12 . This suggested that the quality rather than the mere quantity of anti-spike IgG may be critical for predicting the humoral vaccine response and possibly also the protection against symptomatic disease. In particular, the maturation of antibodies to the spike protein may play an important role in the development of potent neutralizing responses 12,13 . Interestingly, in individuals who recovered from recurrent COVID-19, the presence of low-avidity IgG molecules targeting the receptor-binding domain of the SARS-CoV-2 spike protein during reinfection was shown to be a negative prognostic factor for developing severe COVID- 19 (ref. 14). The development of cellular immunity following COVID-19 vaccination in individuals with cancer has been addressed in recent studies 2,[15][16][17][18][19][20] ; however, these studies often lack prevaccination samples or fail to assess specific or cross-reactive T cell responses to seasonal human β-coronaviruses (HCoVs) or more recent variants of SARS-CoV-2.
Highlighted in a recent review 21 , considerable uncertainty still remains regarding the efficacy of COVID-19 vaccination in individuals with cancer. Laboratory-based studies on individuals with hematologic malignancies have started to focus on the assessment of longitudinal immune responses induced by COVID-19 vaccines in subgroups with different B cell lymphomas (LY), multiple myeloma (MM) and their related treatment regimens 17,[22][23][24][25] . These reports have provided more differentiated insight yet have also highlighted the challenges of analyses of COVID-19 vaccine responses in these diverse groups of individuals with hematologic cancers. Humoral vaccine responses in individuals with cancer are frequently still explored solely by determining the quantity of anti-spike IgG [26][27][28][29][30][31][32] or by the use of surrogate neutralization assays 33-37 that may not faithfully reflect results from live virus neutralization assays, which are considered the gold standard for sensitive quantification of functional, infection-blocking antibodies with in vivo relevance 38,39 . The detailed assessment of COVID-19 vaccine responses in immunocompromised individuals is also important to define the clinical need for early pharmacological interventions with oral antivirals, including paxlovid or molnupiravir 40 , or the potential use of neutralizing monoclonal antibodies [41][42][43][44][45] .
In the current study, we examined the dynamics of humoral and cellular immune responses in a longitudinal cohort of 60 individuals with either LY or MM (vaccinated primarily with BNT162b2 mRNA) by quantifying antibodies to the SARS-CoV-2 spike protein, antibody avidity and neutralization capacity in serum. Moreover, we performed an enzyme-linked immunospot (ELISpot) assay stimulating interferon-γ (IFNγ) release by peptide pools of SARS-CoV-2 spike S1 and S2 domains of Delta and partially of Omicron (BA.1) to quantitatively assess the SARS-CoV-2-and seasonal HCoV OC43-directed T cell responses to COVID-19 in 53 of these individuals with cancer.

Cohort characteristics and study design
We characterized the SARS-CoV-2-specific humoral and cellular immune responses after two and three vaccinations with mostly BNT162b2 mRNA in 60 individuals with various LYs and MM. Individuals were treated at the Freiburg University Medical Center, Germany (baseline participant characteristics are in Table 1) and were followed up from before their initial COVID-19 vaccination (mostly in March or April 2021) to January 2022, the latter time point, on average, 41 d after the third vaccination (booster). Of note, regardless of the subgroup, the remission state was 'stable disease' or better in almost all individuals at the time of first vaccination. Only two individuals diagnosed with MM showed 'active disease' (initial diagnosis or progressive disease, respectively). The remission state of each individual in the LY subgroup treated with the monoclonal antibody (mAb) to CD20 rituximab (Rx) less than 12 months before the first vaccination (LY Rx < 12) and the time between last Rx infusion and the respective COVID-19 vaccinations 1-3 are shown in Extended Data Table 1. The time chart of vaccinations 1-3 and the time points of sample collection (visits 1-4) are shown in Fig. 1a. a 20% increase in antibody avidity from visit 2 to visit 3, a significant increase of 16% from visit 3 to visit 4 and a highly significant increase from visit 2 to visit 4 ( Fig. 1d). Comparing the two cohorts, we noted that individuals with hematologic malignancies showed significantly higher antibody avidities at visit 2 than healthy individuals at 14 d and even 103 d after vaccination 2. Moreover, individuals with cancer showed a drastically higher anti-spike IgG avidity at visit 3 than healthy individuals (Extended Data Fig. 1c), while avidity levels were comparable in both study groups at visit 4 ( Fig. 1d and Extended Data Fig. 1c). In sum, our findings unexpectedly reveal that in those vaccinated individuals with hematologic cancer who were able to mount a humoral response, the avidity of anti-spike IgG rose rapidly after the second vaccination and was markedly higher after the second and before the third vaccination than in healthy individuals, contrasting the concurrently reduced anti-spike IgG titers in individuals with cancer.

Broad infection-neutralizing activity after COVID-19 vaccination
Next, we quantified infection-neutralizing antibodies in sera collected before and longitudinally after COVID-19 vaccination (Fig. 1a) from individuals with cancer and compared these results to those from matched healthy individuals, reported recently 12 . We used a live virus neutralization assay against six SARS-CoV-2 variants, reflecting the evolutionary course of SARS-CoV-2 during the first 2 years of the pandemic. We assessed the serum neutralization titers by  Anti-spike IgG (BAU per ml) measuring half-maximal inhibitory concentrations (IC 50 ) for preventing virus-mediated cytotoxicity 12,47 . The prevaccination baseline neutralization activity is depicted in Extended Data Fig. 2a. The dynamics of a SARS-CoV-2 variant-centered comparison of neutralization activities in longitudinal serum specimens are shown in Fig. 2a-c. In individuals with cancer and healthy individuals, the infection neutralization capacities for VoC Omicron (BA.1) and, albeit less pronounced, for VoCs Beta and Delta were mostly lower than for the other SARS-CoV-2 variants at all time points investigated (Extended Data Fig. 2b-d), confirming their immune escape properties [48][49][50] .
At visit 2, sera from our cohort had lower neutralization activity against the different SARS-CoV-2 variants than sera from healthy individuals by, on average, 2.3-to 5.4-fold, reaching statistical significance for VoCs Beta, Delta and Omicron (Fig. 2a). At visit 3, the cancer cohort showed a 2.3-to 21.3-fold reduced neutralization capacity, which was statistically significant for VoCs Beta and Delta (Fig. 2b), EU1 (Fig. 2b) and Gamma (Fig. 2b) but not for VoCs Alpha and Omicron. After the booster vaccination, individuals with hematologic malignancies displayed a moderately reduced level of neutralization activity for VoCs Omicron and Beta compared to healthy individuals (that is, by 27.6-and 9.1-fold, respectively), whereas neutralization differences for the other VoCs were less pronounced (1.7-to 4.9-fold; Fig. 2c).
Next, we determined the ratio between neutralization IC 50 values for EU1 and anti-spike IgG concentrations to obtain insight into the relative efficacy of spike-targeting antibodies in serum for live virus neutralization. Most remarkably, shortly after the second vaccination, anti-spike IgG-positive individuals with cancer displayed a 20.8-fold higher neutralization capacity per BAU than healthy individuals (Fig. 2d). Over time, the most pronounced increase in neutralization capacity of antibodies to SARS-CoV-2 was observed in healthy donors (Extended Data Fig. 2e). Notably, the neutralization potency per anti-spike IgG in individuals with hematologic malignancies at day 35 after the second vaccination was significantly higher than at day 103 and even day 210 in the group of healthy individuals. In line with this finding, this ratio was significantly lower in healthy individuals 5 months after the second vaccination (visit 3) than in individuals with hematologic malignancies and reached comparable levels only after the third vaccination (Fig. 2d).
As already observed in the longitudinal assessment of anti-SARS-CoV-2 spike IgG titers, analyses of neutralizing capacities also revealed considerable variability among the different subgroups of individuals with cancer according to disease and treatment administered. In untreated individuals with LY, little change in serum-neutralizing titers against all VoCs was noted from visit 2 to 4 ( Taken together, individuals with malignant hematologic diseases who responded to COVID-19 mRNA vaccination with a humoral response displayed a serum neutralization activity that was, on average, only threefold lower than healthy individuals with time-, variant-and participant subgroup-specific patterns. Remarkably, a drastic increase of the neutralizing capacity per anti-spike IgG unit, which coincided with an enhanced antibody avidity, was observed early after the second COVID-19 vaccination in these vaccine-responsive individuals with hematologic malignancies compared to healthy individuals. Together, this underscores the importance of quantifying not only anti-spike IgG titers but also neutralization capacity and antibody avidity for a refined assessment of potential humoral correlates of clinical protection in these individuals with cancer.

Vaccinated individuals mount SARS-CoV-2 spike-specific T cell responses
To examine the vaccine-induced SARS-CoV-2-specific T cell response, we applied an IFNγ ELISpot assay. To assess T cell responses before and after two-dose standard vaccination in individuals with cancer and matched healthy controls (that is, at visits 1 and 2), peripheral blood mononuclear cells (PBMCs) were stimulated with pools of overlapping peptides derived from the spike protein of the Delta VoC and a set of reference antigens in a total of 53 individuals with hematologic malignancies and 12 healthy individuals. Because some individuals with cancer, especially those with chronic lymphocytic leukemia, had high absolute and relative numbers of malignant lymphocytes in peripheral blood, we expressed the results as T cell-normalized spot-forming units (SFU) per million PBMCs.
Delta spike-specific T cells increased in the majority of individuals with cancer and healthy individuals after vaccination 2 (Fig. 4a,c,e). At visit 2, 85% (45 of 53 participants) of all individuals with cancer had an increase in the frequency of Delta spike-specific IFNγ-secreting T cells compared to prevaccination levels. In 76% of individuals with cancer (40 of 53 participants), this increase amounted to 10 or more SFU per million T cells, and the median response increased from 5 to 50 SFU per million cells compared to the prevaccination baseline (Fig. 4e). The overall response was equally distributed to the S1 and S2 moieties of the Delta spike antigen (Fig. 4a,c), and these two responses were strongly correlated (Extended Data Fig. 4a). This suggests that the vaccination response was not predominantly shaped by individual limitations in T cell diversity. In comparison, 100% of healthy individuals (12 of 12) showed an enhanced frequency of Delta spike-specific IFNγ-secreting T cells at visit 2 compared to prevaccination visit 1, and their median responses to S1 and S2 spike peptides at visit 2 were 4.6-and 2.2-fold higher than those observed in individuals with cancer ( Fig. 4a,c). Overall, this demonstrates a robust development of a cellular immune response after two mRNA vaccinations in most individuals with hematologic cancers.
We also analyzed the concomitant T cell response to the spike antigen of the endemic, seasonal β-coronavirus OC43. At visit 1, OC43 spike-specific T cells were detectable at slightly higher levels than SARS-CoV-2 spike-specific T cells (Fig. 4e,g), consistent with OC43 endemicity [51][52][53] . OC43-specific T cells were slightly increased at visit 2 in individuals with cancer and healthy individuals (1.4-fold and 2.1-fold higher median response, respectively; Fig. 4g), but this increase was expectedly less pronounced than the vaccine-induced increase in SARS-CoV-2 spike-specific T cells (Fig. 4e). Visit 2 responses to SARS-CoV-2 and OC43 spike were correlated only to a limited extent in individuals with cancer (Extended Data Fig. 4b). Moreover, the prevaccination response to OC43 in individuals with hematologic malignancies did not predict the response to SARS-CoV-2 after vaccination (Extended Data Fig. 4c). Thus, T cell cross-reactivity among these β-coronaviruses appears to be present but plays a minor role in shaping outcomes of COVID-19 vaccination. The response to SARS-CoV-2 nucleocapsid (Extended Data Fig. 5a) remained low before and after vaccination (median of 1.8 SFU at visits 1 and 2), indicating the absence of intermittent SARS-CoV-2 infection and corresponding to the observed negative results for the detection of anti-nucleocapsid IgG in all sera (Table 1). T cell responses to adenovirus hexon protein were detectable in most individuals with cancer and only slightly increased from visit 1 to visit 2 (Extended Data Fig. 5b). In contrast to some previous studies 17,20,23,54,55 , uncorrected background signals in our ELISpot approach were very low (0 spots in 83% of all individuals with cancer and in 91% in the subgroup of individuals with LY at visit 1), underscoring that bona fide antigen-specific T cells were detected in our analyses with high confidence. Article https://doi.org/10.1038/s43018-022-00502-x The T cell response was separately analyzed in 49 individuals with cancer in this cohort who could be allocated to one of the five subgroups with hematologic malignancies, as described above. An increase in the median T cell response was apparent in all five subgroups and was statistically significant in four of five subgroups for SARS-CoV-2 Delta S1, S2 and the total spike-specific response (Fig. 4b,d,f). An increase in OC43 spike-specific T cells was detected in two of five subgroups (Fig. 4h). The weakest response to the Delta spike protein was observed in untreated individuals with LY, but even in this group, an increase in total spike-specific T cells from a median of 0 SFU (range of 0-35) to 16 SFU (range of 6-224) was observed, with six of nine individuals with cancer showing increased spike-specific T cells (Fig. 4f). Of particular note, all 13 individuals with cancer from the Rx < 12 LY subgroup who were analyzed had detectable T cells against SARS-CoV-2 Delta spike protein (median of 53 SFU, range of 5-1,021; Fig. 4f), although none of these individuals had a vaccine-induced humoral immune response above background (Figs. 1c and 3). In the other four subgroups of individuals with cancer, a solid vaccination-induced T cell response was observed, representing an increase from a visit 1 subgroup-specific median baseline of 3-11 SFU to a visit 2 median response of 36-78 SFU (Fig. 4b,d,f). It is worth mentioning that the SARS-CoV-2 spike-specific T cell response after vaccination was very similar in individuals with  There was a trend of a positive correlation between spike-specific IgG and T cell responses in individuals with MM (Extended Data Fig. 6a) but not in individuals with LY (Extended Data Fig. 6b). The latter appeared to be able to mount a spike-specific T cell response even in the absence of B cells or a corresponding antibody response. Of note, after the booster dose at visit 4, T cell responses to Delta and Omicron (BA.1) spike peptides appeared closely correlated in seven of eight donors, with one individual scoring as an outlier who recognized Delta much better than Omicron (BA.1; Extended Data Fig. 6c).
To assess the sustainability of the T cell response and the effect of a third vaccination, we performed longitudinal (visits 2, 3 and 4) ELISpot analyses in eight individuals with LY who had received their last Rx treatment within 12-60 months before the first vaccination. For this analysis, T cells were tested against spike peptide pools from the VoCs Delta or Omicron (BA.1), because BA.1 had started to become the dominant VoC in Germany at the time of visit 4. Stimulation with the Delta peptide pools (S1 and S2) generally elicited a marginally higher T cell response than the Omicron (BA.1) peptide pools (S1 and S2) at visits 2 and 4, although these differences did not reach statistical significance (Extended Data Fig. 7a). When analyzing T cell responses to Delta and Omicron (BA.1) VoCs between visit 3 and visit 4 in a combined fashion, a positive trend of the booster vaccination was observed (Extended Data Fig. 7b).
In summary, COVID-19 vaccination of individuals with hematologic malignancies elicits potent T cell responses even in the absence of humoral responses in individuals recently treated with Rx. This response may contribute to the control of variants with pronounced humoral escape properties, including VoCs Omicron and Beta. However, in individual cases, the vaccine-induced T cell response might target epitopes in SARS-CoV-2 VoCs that are not conserved in Omicron (BA.1).

Correlative analyses of vaccine responses and blood values
Next, we performed correlation analyses of vaccine-induced humoral and T cell responses to the total IgG concentrations and the counts of certain peripheral blood cell populations obtained from health records from individuals with hematologic malignancies.
At all time points after the second and third COVID-19 vaccinations, anti-SARS-CoV-2 levels and neutralization responses against the majority of SARS-CoV-2 variants correlated significantly with both B cell numbers and total IgG concentrations (Fig. 5a-c). At visit 4, the correlation of total IgG and infection-neutralizing titers was highly significant for all variants tested, including immune escape VoCs Omicron and Beta (Fig. 5c). Shortly after the second vaccination (visit 2), there was a trend toward a negative correlation between infection neutralization activities and CD8 + T cell counts that was statistically significant for VoCs Alpha, Beta, Gamma and Omicron (Fig. 5a). However, this negative correlation could not be observed at visits 3 (Fig. 5b) or 4 (Fig. 5c).
Shortly after the second and third vaccinations, antibody avidities were correlated with neutralizing activities in sera from individuals with hematologic malignancies but surprisingly not in samples from healthy individuals (Extended Data Fig. 8a). By contrast, antibody avidities measured in sera from healthy individuals correlated significantly with neutralization responses only at visit 3 (Extended Data Fig.  8a). Together with the observation that individuals with hematologic malignancies show higher neutralization potencies per anti-spike IgG unit shortly after the second vaccination than healthy individuals (Fig. 2d), these results indicate that the exposure to COVID-19 mRNA vaccines leads to a rather rapid development of highly efficient antibodies to spike in this group of individuals with cancer. Of note, anti-spike IgG levels, avidities and neutralization responses against SARS-CoV-2 in individuals with hematologic malignancies did not or only weakly correlate with T cell responses against SARS-CoV-2 spike, HCoV OC43 spike and adenovirus hexon protein at visit 2 (Extended Data Fig. 8b).
In conclusion, our analyses show a general positive correlation between B cell numbers and concentrations of antibodies to SARS-CoV-2 as well as neutralization responses in vaccinated individuals with hematologic malignancies. High CD8 + T cell numbers coincided with lower infection neutralization capacities but only shortly after the second vaccination. Furthermore, anti-spike IgG avidities correlated strongly with neutralization responses in individuals with hematologic malignancies shortly after the second and third vaccinations.

Clinical presentation during breakthrough infections
Between 16 August 2021 and 25 July 2022, 13 vaccinated individuals in our hematologic cohort were diagnosed with a SARS-CoV-2 breakthrough infection. Seven of these individuals were LY Rx < 12 individuals, of which two had severe COVID-19 and were hospitalized, required oxygen (but no mechanical ventilation) and received either a combined treatment with casirivimab, imdevimab and tocilizumab or sotrovimab, two experienced moderate symptoms, and three experienced mild symptoms. Mild COVID-19 was also diagnosed in 2 untreated individuals with LY, 2 treated individuals with LY, and 2 untreated individuals with MM. Additionally, four individuals had a second breakthrough infection, one with moderate COVID-19 and three with mild COVID-19 (Extended Data Table 2). Based on prevalence data for VoCs during the summer and fall of 2021 and winter and spring 2022 in Germany, 4 of the breakthrough infections were likely caused by VoC Delta, and 13 were likely caused by VoC Omicron subvariants BA.1, BA.2, BA.4 and BA.5. In particular, the group of individuals with LY with Rx < 12 with virtually absent humoral vaccine responses was intriguing, because five of seven individuals did not require hospitalization for this infection event. No COVID-19-related death was seen in our group of individuals with hematologic malignancies.

Discussion
Using a longitudinal approach, we studied the humoral and T cell immune responses elicited by two and three vaccinations (mostly with BNT162b2 mRNA) in a cohort of individuals with different B cell LYs and MM. Time-resolved, parallel assessment of titers of antibody to SARS-CoV-2 spike protein, antibody avidity and neutralization capacity to six authentic, replication-competent viruses, and SARS-CoV-2-and HCoV OC43-directed T cell responses allowed us to obtain a comprehensive picture of COVID-19 vaccine-induced immune responses in these individuals with hematologic cancers relative to a group of matched healthy individuals.
In the current study, we report six key findings. First, most individuals with a hematologic malignancy who are capable of developing a vaccine-induced antibody response mount an infection neutralization capacity against SARS-CoV-2 variants that is only slightly lower than in healthy individuals (reduced, on average, by 2.5-, 6.4-and 4.1-fold) despite markedly lower titers of antibody to the spike protein in the former group (reduced 49.1-, 23.6-and 9.2-fold at visits 2, 3 and 4, respectively). Second, in individuals with hematologic malignancies, the neutralizing potency per anti-spike antibody unit is drastically enhanced (20.8-fold on average) early after the second vaccination compared to healthy individuals. Third, the avidity of serum IgG binding to the SARS-CoV-2 spike protein in these individuals is higher before the third COVID-19 vaccination than in healthy individuals. Fourth, in a direct comparison of all VoCs, Omicron (BA.1) and, to a lesser extent, Beta and Delta displayed the most pronounced humoral immune escape, in line with results recently described for healthy individuals following either vaccination or infection 12 . Fifth, the majority of the participants in our study, including indivi duals with LY receiving Rx treatment, mounted a robust vaccine-induced T cell response to two recent VoCs. Sixth, clinical presentation during breakthrough infections is consistent with a partially protective effect of vaccination in immunocompromised individuals with hematologic malignancies with no report of a COVID-19-related death in our cohort.
Nevertheless, our observations indicate that individuals with LY without a humoral immune response due to recent Rx treatment have a higher risk for symptomatic breakthrough infections. This group of highly immunosuppressed individuals might benefit from passive immunization through a preexposure prophylaxis with neutralizing mAbs to the spike protein. This approach seems feasible, although many of the currently available mAbs have experienced a loss of antiviral efficacy for recent VoCs [43][44][45] .
Analyzing vaccine-induced antibody responses in at-risk individuals with hematologic malignancies is especially important because those individuals often display an imbalance between humoral and cellular immunity, which can be attributed to disease-related lineage defects, Rx treatment or BTK inhibition. Consequently, it is critical to define parameters that best reflect vaccine-induced immunity.
Our multiparametric longitudinal analysis of vaccine-induced humoral immunity revealed that the quality of antibodies to the spike protein after the second vaccination was remarkably high in individuals with hematologic malignancies who developed antibodies of superior neutralizing quality, as reflected by the ratio of neutralizing titer per antibody unit. By contrast, healthy individuals required a third vaccination to reach a comparable neutralizing capacity. Consistently, increasing antibody avidities correlated strongly with neutralization responses shortly after the second vaccination in individuals with hematologic malignancies. A sudden increase in antibody avidity was noted in healthy individuals only following the third vaccination compared to a steady enhancement of avidity between visits 2 and 4 in our cancer cohort. While sampling time points do not perfectly match between these two non-contemporaneous cohorts, we demonstrate that even 68 additional days, on average, after the second vaccination in healthy individuals did not result in the same high avidity as in

Fig. 4 | SARS-CoV-2-and OC43-specific T cell responses in individuals with hematologic malignancies before and after two-dose COVID-19 mRNA vaccination. T cell responses analyzed by IFNγ
ELISpot and expressed as SFU per 10 6 T cells are shown as box plots with median, bounds between upper and lower quartiles and whiskers between the 10th and 90th percentiles. Samples were obtained before vaccination 1 (visit 1) or at 2-8 weeks after vaccination 2 (visit 2). Differences between time points were analyzed for statistical significance using the Wilcoxon matched-pairs signed-rank test, and differences between healthy donors and individuals with hematologic malignancies were analyzed using the Mann-Whitney test. Brackets show statistically significant differences, and precise numerical P values are indicated. Absence of brackets or P values indicates absence of statistical significance. a,b, T cell responses to SARS-CoV-2 spike S1. c,d, T cell responses to SARS-CoV-2 spike S2. e,f, Sum of T cell responses to SARS-CoV-2 spike S1 and S2. g,h, Sum of T cell responses to human coronavirus Article https://doi.org/10.1038/s43018-022-00502-x individuals with hematologic malignancies shortly after the second vaccination. In line with this finding, the neutralizing potency per anti-spike IgG unit was also significantly higher in individuals with hematologic malignancies at 2-8 weeks after the second vaccination than at 4-5 months and even 7 months in healthy individuals. Collectively, this highlights that differences in anti-spike IgG avidity and SARS-CoV-2 S1 neutralizing capacity observed between the two cohorts cannot be accounted for by small differences in sampling time points. We speculate that in vaccinated individuals with LY and MM, following only two exposures to the spike protein, the affinity maturation of memory B cells may be accelerated or the breadth of antibodies may be expanded by yet undefined mechanisms. In hematologic malignancies, a deregulation and constitutive expression or activation of proteins of the apolipoprotein B mRNA-editing enzyme catalytic polypeptide (APOBEC) family of deaminases, which are involved in oncogenesis, has been reported [56][57][58][59][60][61] . Canonical functions of these types of enzymes, particularly activation-induced cytidine deaminase (AID), are key to somatic hypermutation in B cells and thus the generation of high-affinity antibodies. We propose that higher activities of AID/ APOBEC family members in B cells of these types of hematologic cancer could promote antibody maturation.
While the third vaccination was of importance in both groups to elevate antibody titers, we propose that the quality rather than the mere quantity of vaccine-induced spike-targeting antibodies is of exceptional importance and has a specific signature in individuals with hematologic malignancies. Anti-spike titers alone may thus underestimate the potency of the humoral response in COVID-19 vaccinees with LY or MM, and neutralization and antibody avidity provide additional depth to defining potential functional correlates of protection from severe COVID-19. Consistently, recent studies indicate that neutralizing antibody levels may be highly predictive of immune protection from symptomatic COVID- 19 (refs. 11 , 62).
mRNA-based COVID-19 vaccines are known to induce spike-specific T cell responses in the majority of healthy recipients 36,63,64 . In individuals with cancer, more heterogeneous results have been reported after two vaccinations 15,17,20,23,24 . Our findings on vaccine-induced T cell responses in 85% of individuals with hematologic cancers are in line with a large British cancer cohort study 54 that showed detectable responses in up to 77% of individuals and a cohort of individuals with hematologic cancers with response rates of 80% (refs. 20). Liebers et al. found spike-specific T cells in only 58% of individuals with B cell malignancies with Rx treatment 24 . These authors used low numbers of input cells in their ELISpot analysis, which may have led to an underestimation of functional spike-specific T cells 24 . Our finding of functional T cell responses in 12 of 13 individuals with hematologic cancer (92%) currently or recently treated with Rx supports the notion that two-dose COVID-19 mRNA vaccination may indeed be of high benefit even for individuals who are unable to develop antibody responses. Our limited clinical observations during breakthrough infections indicate that even individuals with LY who are B cell deficient were largely protected from severe COVID-19, likely due to T cell-mediated immunity.
The majority of our two-dose-vaccinated individuals with MM (89%, 16 of 18) developed a detectable SARS-CoV-2 spike-specific T cell response, while Enßle et al. reported considerably lower numbers (34%) 23 . Of note, only 71% of their vaccinated healthy individuals (controls) scored positive, which stands in contrast to findings by us and  [63][64][65] . Possible explanations for these discrepancies include a high background signal in their IFNγ ELISpot assay and their use of incomplete peptide pools 23 . Our results suggest that also the majority of vaccinated individuals with MM raises both efficient spike-specific T cell responses and neutralizing antibody responses.
The strengths of our study include the longitudinal sampling of the clinically well-characterized cohort of individuals with cancer and the matched group of healthcare workers and the expanded methodological approach, including live virus neutralization assays against multiple VoCs and the quantification of anti-spike IgG avidity. Limitations of our study include the fairly small number of participating individuals with hematologic cancer (N = 60). The study lacks a contemporaneous control set with other hematologic malignancies or solid tumors. For correlation analyses, adjustments for multiple testing were not performed, making it difficult to exclude potential confounders.
Based on our current results, the multiparametric assessment of quantity and quality of COVID-19 immunity elicited by vaccines and infection should be the future standard in larger longitudinal studies seeking to strengthen the role of individual quantitative markers as correlates of protection in individuals with hematologic malignancies in light of newly emerging VoCs of this pandemic virus.

Participants and samples
We conducted a monocentric, observational cohort study that commenced recruitment in March 2021 and continued to enroll participants until May 2021. We report results generated during extensive analyses of participant samples received within this study. Because the course of the pandemic and the sudden availability of COVID-19 vaccines for individuals with cancer was not predictable initially, the consent of our participants for an additional draw of blood during a regular outpatient visit was obtained through signing the informed consent form of the Biobank (FREEZE) of the Medical Center of the Freiburg University. Subsequently, all individuals signed the informed consent form of this study. Participants who signed the Biobank (FREEZE) consent before the study-specific informed consent were considered 'retrospective' in this specific case. The timing of blood draws in relation to vaccination date was 'prospective' in all cases. The study was initiated at the Biobank (FREEZE) on 8 March 2021; the local ethics committee received all documents on 1 June 2021.
The first inclusion date of a study participant was 9 March 2021, and the last inclusion date was 17 May 2021.
The following were the study inclusion criteria: Participants must not have received any prior COVID-19 vaccination and must have one of the following diagnoses (and therapy): B cell Non-Hodgkin's lymphoma (follicular lymphoma, mantle cell lymphoma, marginal zone lymphoma, diffuse large B cell lymphoma and chronic lymphatic leukemia) without treatment ('watch and wait'), B cell Non-Hodgkin's lymphoma less than 12 months after therapy with mAbs to CD20 (Rx, obinutuzumab), B cell Non-Hodgkin's lymphoma 12-60 months after therapy with mAbs to CD20, chronic lymphathic leukemia with BTK or BCL-2 inhibition (venetoclax, ibrutinib), MM (watch and wait, treatment with lenalidomide and bortezomib or treatment with daratumumab, ixazomib and carfilzomib) or non-small cell lung cancer under PD1/ PD-L1 checkpoint inhibition (durvalumab, pembrolizumab and nivolumab).
All individuals fulfilling these inclusion criteria and who visited our outpatient clinics between March and May 2021 were given the chance to participate.
The primary endpoint was defined as evaluation of the immunogenicity of COVID-19 vaccination 2-8 weeks after the second vaccination, as confirmed by the presence of spike-specific antibodies to SARS-CoV-2 (Fig.1b). The secondary endpoints focused on a profound analysis of the vaccine-induced immune response. When we started our sample collection, no prior published data on COVID-19 vaccine responses in individuals with cancer were available, and, therefore, no precise sample size could be calculated. Having 80 participants as the goal for recruitment was a best guess estimate from how many individuals with the respective diagnosis were seen in our outpatient clinics from March to May 2021. Because our study design required a blood draw before the first COVID-19 vaccination and the scheduling of vaccination was not part of this study (participants were vaccinated by their primary care physician or national vaccination centers), no further participants could be included in the study after May 2021, resulting in the inclusion of 60 individuals with hematologic malignancies in total. The few recruited individuals with non-small cell lung cancer were disregarded in this report focusing on hematologic malignancies. Addition of further time points for blood draws as well as further analyses with emerging VoCs were in agreement with the local ethics committee.
This study was conducted in accordance with the ethical principles of the Declaration of Helsinki, Good Clinical Practice and applicable regulatory requirements. Adult participants provided written consent to participate in this study or to contribute samples to the Biobank, including agreement to the deposition of pseudonymized data. Participants were not financially compensated. Due to the observational design of this study, no randomization was performed.
This study was approved by the local Ethics Committee (21-1386) and registered at the Paul-Ehrlich Institute (NIS599) and Deutsches Register Klinischer Studien (DRKS00025901). Clinical information regarding participant medical histories was obtained from the hospital digital medical file system, allowing for accurate follow-up. Cancer type was defined according to the International Classification of Disease (10th revision) diagnostic codes. The registered data were abstracted from medical files by board-approved hematologists/oncologists. Details of time points of vaccination and blood sample collection are depicted in Fig. 1a and Table 1 Baseline data included age, sex, ethnicity, cancer and treatment type at time of first dose of vaccination, dates of vaccinations, type of vaccine and interval of the last treatment in relation to the first vaccine dose. Autologous stem cell transplantation at any time before vaccination was also registered ( Table 1).
We considered individuals to have a breakthrough SARS-CoV-2 infection if they tested positive for SARS-CoV-2 RNA by real-time PCR in a respiratory swab.
We collected 36 ml of EDTA blood and 7.5 ml of serum from 60 individuals with hematologic neoplasia (for detailed participant characteristics, see Table 1). Plasma was derived from peripheral blood samples by centrifugation and frozen immediately at −80 °C. PBMCs were isolated using a lymphocyte separation density gradient and immediately frozen in liquid nitrogen.
Data for antibody responses and PBMCs for T cell assays from vaccinated, infection-naive healthy individuals were obtained from a previously described cohort 12 . This cohort included healthcare workers vaccinated with BNT162b2 mRNA. This study was approved by the local ethics committee (ethics vote 476/20 and 26/21S-SR), and Article https://doi.org/10.1038/s43018-022-00502-x participants provided written informed consent to study participation and biobanking. To account for the heterogeneity of both cohorts, we selected cases from the latter group based on age and time intervals to vaccinations and sample collections.

Quantitative antibody detection assays
IgG-type antibody responses to the S1 domain of SARS-CoV-2 spike antigen were quantified using the commercial, automized SARS-CoV-2 IgG II quant assay (Abbott). Levels of IgG-type antibodies to SARS-CoV-2 spike S1 domain in healthy individuals were previously measured using the anti-SARS-CoV-2 QuantiVac enzyme-linked immunosorbent assay (ELISA; IgG; EuroImmun) 12 . To investigate possible differences in the quantitative detection of antibody levels between the two assays, we compared the BAU per ml in serum samples of 20 individuals with COVID-19 measured with both assays. Pearson correlation analysis showed strong (r = 0.85) and significant (P ≤ 0.0001) correlation, and linear regression analysis revealed a conversion factor of 0.71 for quantitative results in BAU per ml between the two assays. This conversion factor was used to normalize and compare data for quantitative antibody measurements shown in Fig. 1b,c and Extended Data Fig. 1b,c.

Antibody avidity assay
Binding strength of the SARS-Cov-2 IgG antibodies to the spike antigen of SARS-CoV-2 strain Wuhan-hu-1 were quantified by adaptation of the commercial IgG agile SARS-CoV-2 ELISA (Virion/Serion) using ammonium thiocyanate (Roth) as a chaotropic agent as described previously 66 . The following formula was used to calculate relative avidity: percent avidity = (IgG concentration, ammonium thiocyanate treated)/(IgG concentration, PBS treated) × 100.

SARS-CoV-2 neutralization assay
High-titer virus stocks were generated and characterized as reported recently 12,67

IFNγ ELISpot assays
IFNγ ELISpot assays were used to determine the frequency of virus peptide-specific T cells 68 . ELISPOT 96-well plates (MultiScreen MSIPN4510, Millipore) were pretreated with 35% ethanol in water, washed and coated overnight with anti-IFNγ according to the manufacturer's protocol (antibody 1-D1K, Mabtech). The next day, cryopreserved PBMC samples were thawed and used in the assay within 2 h of thawing. PBMCs (250,000 cells per well) were co-incubated with pools of 15-mer peptides at 0.5 μg ml -1 per peptide (PepMix, JPT) in 200 μl of medium (RPMI 1640 (Gibco) with 5% human male AB serum (PAN Biotech)). Reactions were generally set up in triplicate wells; duplicate reactions were used when cell numbers were insufficient. Plates were developed by incubation with anti-IFNγ-biotin and streptavidin-alkaline phosphatase according to the manufacturer's instructions (Mabtech). Spots were developed using the AP Conjugate Substrate kit (Bio-Rad). Spots were counted in an automated ELISpot reader (C.T.L. ImmunoSpot).
Peptide pools used in the ELISpot assays were protein-covering pools of 15-mer peptides overlapping in 11 amino acids: SARS-CoV-2 Delta variant spike protein parts 1 and 2 (first and second half of the amino acid sequence), SARS-CoV-2 (Wuhan-Hu-1) nucleoprotein, hCoV OC43 spike protein parts 1 and 2 and human adenovirus 5 hexon protein (all from JPT).
T cell responses expressed as SFU were normalized to the frequency of T cells in the individual PMBC samples. Normalization was mandatory because PBMCs collected from the different groups of individuals with hematologic cancers showed heterogeneous T cell abundance due to the high B cell lymphocytosis characteristic of certain lymphomas (for example, chronic lymphocytic leukemia).

Statistics and reproducibility
Data and statistical analyses were performed in Prism 9 (GraphPad Software). Doses of virus required to infect 90% of cells and IC 50 values for neutralization were calculated after normalized, sigmoidal dose-response curve approximation of the respective data. Precise numerical P values for all statistical analyses can be found in the respective figures and the Supplementary Information.
Data distribution was assumed to be not normal, but this was not formally tested.
No statistical method was used to predetermine sample size. When we started our participant recruitment and sample collection, no published data on COVID-19 vaccine responses in individuals with hematologic cancer were available.
Data from individuals positive for antibodies to SARS-CoV-2 nucleocapsid were excluded. The investigators were blinded to allocation during experiments and outcome assessment.

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability
Pseudonymized participant data, including participant record data and all primary data from measurements conducted, are available in a public repository (https://data.mendeley.com/datasets/z6dw96y8sw/1). The SARS-CoV-2 sequences and protein data are available under accession codes 6VXX (Protein Data Bank), MW717675.1 and MZ945494 (GenBank) and EPI_ISL_412971, EPI_ISL_2557176 and EPI_ISL_8768822.2 (GISAID). Source data are provided with this paper. All other data supporting the findings of this study are available from the corresponding author on reasonable request. Data Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A description of any restrictions on data availability -For clinical datasets or third party data, please ensure that the statement adheres to our policy

Human research participants
Policy information about studies involving human research participants and Sex and Gender in Research.
Reporting on sex and gender Population characteristics

Recruitment
Ethics oversight Note that full information on the approval of the study protocol must also be provided in the manuscript.

Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design
All studies must disclose on these points even when the disclosure is negative.
Sample size

Data exclusions
Replication Pseudonymized participant data, including patient record data and all primary data from measurements conducted, are available in a public repository (DOI: 10.17632/z6dw96y8sw.1). The source data for all figures and extended data figures are available in the supplementary information. The SARS-CoV-2 sequences and protein data are available under accession codes PDB ID 6VXX, GenBank IDs MW717675.1 and MZ945494, and GISAID IDs EPI_ISL_412971, EPI_ISL_2557176, and EPI_ISL_8768822.2. Written informed consent to the publication of pseudonymized data has been been obtained from study participants. All other data supporting the findings of this study are available from the corresponding author on reasonable request.
The sex is described in Table 1 of the manuscript, and in the pseudonymized participant data table (DOI: 10.17632/ z6dw96y8sw.1).
All patients aged 18 years or older and had a confirmed diagnosis of B-cell-Lymphoma or Multiple Myeloma were eligible. The age, sex, disease remission status and treatment history are described in Table 1 of the manuscript , and in the pseudonymized participant data table (DOI: 10.17632/z6dw96y8sw.1). All participants provided a written consent to participation in this study or to sample contribution to the Biobank FREEZE including the agreement to deposition of pseudonymized data.
All patients were recruited while visiting the outpatient center of the Medical Center -University of Freiburg, Faculty of Medicine, University of Freiburg, Germany. No sample-size calculation was performed. When we started our sample collection no prior published data on COVID-19 vaccine response in cancer patients was available. During the recruiting period all patients with either a B-cell lymphoma or multiple myeloma were invited to participate in the study, regardless of sex, age,and comorbidities. For accurate experimental analyses we aimed to receive patient material from before the first vaccination. After the recruiting period all patients in our outpatient center wanting to participate in the study were already vaccinated against COVID-19 and therefore no longer eligible for being included into our study.
Patients were excluded, of whom no samples sufficient for evaluation could be collected. All serum samples were characterized for the prescence of anti-SARS-CoV2 anti-nucleocapsid antibodies.
In case of a positive anti-nucleocapsid titer, patients were excluded from data analysis (except for breakthrough infections).
The assays to determine binding antibody units (BAU) were performed using commercial, diagnostic well-validated tests that make use of calibrators, negative and positive controls. Titers were determined according to WHO standards BAU assuring high standardization. Antibody avidity was characterized using an established modified commercial assay that was validated previously showing low variance between results