1139 - Initial anti-CD20 monoclonal antibody therapy in CLL quickly induces cytokine release syndrome with first dose infusion reaction correlating only with high IL-6, IL-8 and IP-10

64% of CLL patients have a potentially life-threatening first dose infusion reaction (FDIR) within 1–2 h after initiating intravenous (IV) rituximab anti-CD20 monoclonal antibody (mAb) therapy. Once FDIR is observed, the infusion must be interrupted and patient symptoms must be managed before restarting. If FDIR is not treated, serious complications and fatalities may occur.
To better understand FDIR, we studied blood samples collected from 37 treatment-naïve CLL patients in our clinical trial NCT03788291 undergoing IV treatment with low dose (50 mg) and slow infusion (25 mg/h) anti-CD20 mAb (rituximab). All human specimen collection and usage was conducted with written informed consent after approval of the University of Rochester Research Subjects Review Board according to the ethical guidelines of the Declaration of Helsinki. We hypothesized that slow infusion of low dose mAb compared to standard infusion and dose (50 mg/h, 375 mg/m2) may produce fewer FDIRs. We analyzed four time points: baseline (0 h (pre), prior to infusion), 1 h later (during infusion), at the end of infusion (~3 h (post)) and at 48 h (Fig.1). Based on previous reports, we anticipated that CLL counts, CD20 levels and consumption of mAbs/complement would correlate with FDIRs. Furthermore, we expected cytokine release syndrome would be measurable in FDIR sera.
24 (65%) patients had a FDIR as defined by Common Terminology Criteria for Adverse Events version 5 (CTCAE) grade 2 or higher and required intervention within the first two hours. This FDIR frequency is comparable to the rate seen in standard rituximab therapy (64%), implying that dose and infusion rate of anti-CD20 mAb are not key components of FDIR, which further unexpectedly suggests that antigen/antibody levels may not correlate with FDIR.
To test this idea, flow cytometry of CLL blood samples was performed to measure CLL cell counts and CD20 levels. We further measured serum rituximab and complement levels. None of these measurements correlated with the occurrence of FDIR in CLL patients.
To explore factors outside of antigen/antibody levels, we examined CLL patient characteristics (IGHV mutation, Rai stage, cytogenetic defects/mutations) and none of these correlated with induction of FDIR.
Previous studies of rituximab FDIR noticed a cytokine release syndrome, although non-FDIR patients were not examined. In our study, we studied human sera from both FDIR and non-FDIR patients using a Luminex xMAP multiplex assay. We found cytokine release occurred in both types of patients with a ≥ 4-fold increase in one or more cytokines compared to pre-treatment levels in 95% of patients at the end of infusion (~3 hr (post)).
Because overall cytokine induction was similar between patients, we examined specific cytokines for differences in induction that might correlate with FDIR. IL-6, IL-8, and especially IP-10 induction ~3 hr (post) correlated with FDIR.
IP-10 (interferon gamma inducible protein-10), also known as CXCL10 (C-X-C motif chemokine ligand 10), is mainly secreted by monocytes and macrophages. To study the source of IP-10, we performed single-cell RNA sequencing on blood samples from two patients with FDIR at three timepoints: baseline (0 h (pre)), at the end of infusion (~3 h (post)) and at 48 h. IP-10 was detected at very low levels only in monocytes and no other cell types. This suggests that IP-10 is produced in tissue-resident cells that are monocytes or derived from monocytes, such as macrophages.
To test this, we prepared human monocyte-derived macrophages (hMDM) and added CLL cells opsonized with anti-CD20 mAb. Compared to no mAb, we found that IL-6, IL-8 and IP-10 were induced, although not necessarily at comparable levels.
To model FDIR in mice, IV anti-CD20 (5D2 mAb) treatment of C57BL/6 wildtype mice induced IL-6 and IP-10 at 2 h, which decreased at 18 h to levels similar to IgG control. This time course is similar to that seen in FDIR in human patients.
In conclusion, FDIR does NOT correlate with anticipated CLL cell characteristics or levels of antigen, antibody, or complement. Infusion of anti-CD20 mAb induces cytokines in most patients, with FDIR correlating with higher levels of IL-6, IL-8, and IP 10. Single cell RNA sequence data suggests tissue resident monocytes or macrophages, are responsible for IP-10 production. Modeling by in vitro assays and a mouse model reproduce mAb induction of cytokines, which is currently under investigation.