Exposure of progressive immune dysfunction by SARS-CoV-2 mRNA vaccination in patients with chronic lymphocytic leukemia: A prospective cohort study

Abstract

Background Patients with chronic lymphocytic leukemia (CLL) have reduced seroconversion rates and lower binding antibody (Ab) and neutralizing antibody (NAb) titers than healthy individuals following Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) mRNA vaccination. Here, we dissected vaccine-mediated humoral and cellular responses to understand the mechanisms underlying CLL-induced immune dysfunction. Methods and findings We performed a prospective observational study in SARS-CoV-2 infection-naïve CLL patients (n = 95) and healthy controls (n = 30) who were vaccinated between December 2020 and June 2021. Sixty-one CLL patients and 27 healthy controls received 2 doses of the Pfizer-BioNTech BNT162b2 vaccine, while 34 CLL patients and 3 healthy controls received 2 doses of the Moderna mRNA-1273 vaccine. The median time to analysis was 38 days (IQR, 27 to 83) for CLL patients and 36 days (IQR, 28 to 57) for healthy controls. Testing plasma samples for SARS-CoV-2 anti-spike and receptor-binding domain Abs by enzyme-linked immunosorbent assay (ELISA), we found that all healthy controls seroconverted to both antigens, while CLL patients had lower response rates (68% and 54%) as well as lower median titers (23-fold and 30-fold; both p < 0.001). Similarly, NAb responses against the then prevalent D614G and Delta SARS-CoV-2 variants were detected in 97% and 93% of controls, respectively, but in only 42% and 38% of CLL patients, who also exhibited >23-fold and >17-fold lower median NAb titers (both p < 0.001). Interestingly, 26% of CLL patients failed to develop NAbs but had high-titer binding Abs that preferentially reacted with the S2 subunit of the SARS-CoV-2 spike. Since these patients were also seropositive for endemic human coronaviruses (HCoVs), these responses likely reflect cross-reactive HCoV Abs rather than vaccine-induced de novo responses. CLL disease status, advanced Rai stage (III-IV), elevated serum beta-2 microglobulin levels (β2m >2.4 mg/L), prior therapy, anti-CD20 immunotherapy (<12 months), and intravenous immunoglobulin (IVIg) prophylaxis were all predictive of an inability to mount SARS-CoV-2 NAbs (all p ≤ 0.03). T cell response rates determined for a subset of participants were 2.8-fold lower for CLL patients compared to healthy controls (0.05, 95% CI 0.01 to 0.27, p < 0.001), with reduced intracellular IFNγ staining (p = 0.03) and effector polyfunctionality (p < 0.001) observed in CD4+ but not in CD8+ T cells. Surprisingly, in treatment-naïve CLL patients, BNT162b2 vaccination was identified as an independent negative risk factor for NAb generation (5.8, 95% CI 1.6 to 27, p = 0.006). CLL patients who received mRNA-1273 had 12-fold higher (p < 0.001) NAb titers and 1.7-fold higher (6.5, 95% CI 1.3 to 32, p = 0.02) response rates than BNT162b2 vaccinees despite similar disease characteristics. The absence of detectable NAbs in CLL patients was associated with reduced naïve CD4+ T cells (p = 0.03) and increased CD8+ effector memory T cells (p = 0.006). Limitations of the study were that not all participants were subjected to the same immune analyses and that pre-vaccination samples were not available. Conclusions CLL pathogenesis is characterized by a progressive loss of adaptive immune functions, including in most treatment-naïve patients, with preexisting memory being preserved longer than the capacity to mount responses to new antigens. In addition, higher NAb titers and response rates identify mRNA-1273 as a superior vaccine for CLL patients.