Analysis of the IgA Heavy Chain Repertoire in Multiple Myeloma: Implications for Host-Microbiome Influences

Background: Advances in understanding the biology of multiple myeloma (MM) have translated into novel treatment paradigms that have prolonged survival. Nevertheless, relapse remains common even in the absence of measurable residual disease. Therefore, further research is required to identify additional mechanisms underlying MM pathogenesis, particularly in high-risk subtypes. IgA MM represents approximately 20% of cases and for unclear reasons confers a worse prognosis. Given the physiologic inter-relationships of host IgA with microbiota, we first characterized IgA levels in MM patients and then used a highly sensitive next generation sequencing (NGS) technique for genetic-structural analyses to provide possible insights into selection biases in IgA heavy chain usage that may suggest common antigenic exposures.

Methods: Total RNA was obtained from unsorted peripheral blood (PB) and bone marrow (BM) mononuclear cells from 14 newly diagnosed MM patients, as well as PB from 9 healthy controls. A heavy chain isotype-specific ELISA was used to characterize serum IgA levels. For highly sensitive unbiased IgA antibody gene repertoire surveys, we developed a PCR-based technique (El Bannoudi et al. Methods Mol Biol. 2017) using 5'-rapid amplification of cDNA ends to amplify the IgA heavy chain rearrangements for MiSeq (Illumina, San Diego, CA) next generation sequencing (NGS). Immunoglobulin heavy chain (IGH) VH, DH and JH sequences were aligned to their corresponding germline alleles using IgSCUEAL, a previously described sequence alignment algorithm (Frost et al. Philos Trans R Soc Lond B Biol Sci. 2015). To quantify stool microbiota bound to serum IgA, FACS analysis with human anti-IgA-Fc secondary antibodies (Abcam, Cambridge, MA) was performed using primary antibodies from MM or control serum mixed with stool from the same individual.

Results: In healthy adults, IgA1 predominates 4-fold in the circulation compared to IgA2 which is produced mostly in the mucosa, yet more extreme skewing was found in IgA1 in MM patients, with suppression of serum IgA2 as assessed by ELISA. For NGS, there was a mean 2.1x10⁵ sequences read which identified an estimated 7.4 x10⁵ clonotypes, with highly reproducible results among technical replicates. When comparing PB to BM compartments, there were differences in the dominant IgA repertoires, particularly in IgA MM where PB showed a preponderance of a VH3-JH5 (mostly VH3-74) clone. In PB the majority of IGH sequences were derived from subclones of VH1, VH3, and VH4, and JH1, JH4, and JH5 clones, whereas in BM VH1, VH3 and JH4, JH5 subclones predominated. These were distinct from the PB repertoires of healthy controls. Lastly, pilot experiments quantifying IgA bound to stool bacteria using MM serum antibodies and FACS analysis yielded nearly 11-fold total cells, compared to results from a parallel experiment in a healthy control (2.7x10⁶ vs 2.3x10⁵ cells, respectively). Ongoing work is aimed at commensal taxa determination of sorted cells using 16S rRNA sequencing.

Conclusions: We present initial studies that reveal alterations of the total IgA heavy chain repertoire in IgA MM using an unbiased NGS technique, which can suggest influences on B cell selection during clonal evolution. These preliminary studies show that distinct disease-related and background IgA heavy chain repertoires exist in MM compared to healthy individuals. Future analyses are aimed at clustering IgA heavy chain sequences into clonal subsets to define their evolution from MM precursors. In addition, differences in the degrees of IgA-bound gut flora compared to a healthy individual suggest dysbiosis in the MM microbiome, warranting further investigation.