siRNA for the Ig Light Chain Constant Region Reduces Light Chain Production and Secretion By Human Plasma Cells and in a Murine Xenograft Model

Background Ig light chain (LC) diseases such as AL amyloidosis and monoclonal light-chain deposition disease are caused by pathologic free LC. Treatment is aimed at eliminating LC production but success is limited. RNA interference (RNAi) can stop LC production but the diversity of LC variable region sequences poses a challenge that targeting consensus sequences in the constant region (CR) of LC mRNA may overcome (Blood 2014;123:3440). We have developed siRNA pools designed to target the κ or λ LC CR mRNA in human plasma cells and impair LC production and secretion, and have shown that the pool targeting the λ LC CR can do so, and can also trigger a terminal unfolded protein response in clones producing intact Ig due to intracellular accumulation of unpaired heavy chains (ibid). Here we report the results of continued in vitro and in vivo testing of these pools in patient specimens and in a murine xenograft model.

Methods Pools of siRNA for the κ or λ LC CR (si[IGLCκ_CR], si[IGLCλ_CR]) were custom produced with a non-target control (si[-]). They were introduced in vitro into human plasma cells by an optimized streptolysin O-based method (SLO) and in a NOD.SCID xenograft flank plasmacytoma model by in vivo electroporation as per Gene Therapy 2011;18:1150. In vitro we evaluated LC gene expression, production and secretion at 24 hours in human myeloma cell lines and CD138-selected specimens from patients with plasma cell neoplasms, using real-time PCR (qPCR) for LC mRNA, flow cytometry for intracellular LC mean fluorescence intensity (MFI) and ELISA (Bethyl Laboratories) for LC secretion in 24-hour suspension cultures (10⁶ cells/ml). In vivo we inoculated each of the flanks of NOD.SCID mice with 10⁷ human myeloma cells (ALMC-1 or ALMC-2). When plasmacytomas were 0.5cm³ we injected si[IGLCλ_CR] or si[-] one time to each flank plasmacytoma respectively, allowing each mouse to serve as its own control. Two days later, the mice were sacrificed and the plasmacytomas excised for qPCR for λ LC mRNA and serum was obtained to measure human λ LC levels by ELISA.

Results We have previously described results with siRNA targeting the λ LC CR in human cell lines that make λ LC (ALMC-1, ALMC-2, EJM, OPM2, MM.1S, and MM.1R) and in 16 AL λ patient specimens. We demonstrated significant decreases in LC mRNA, intracellular LC MFI, and λ LC secretion by cell lines (Blood 2014;123:3220); moreover, transcriptional profiling indicated minimal off-target effects (ibid; Supplement). We now report that in vitro secretion of λ LC by CD138-selected plasma cells from AL patients (n=3, newly diagnosed λ) treated with si[IGLCλ_CR] was reduced by 65% from a mean of 3.1 to 1.0µg/ml and that the residual λ LC mRNA was 49% of control. Similarly we treated κ LC secreting human myeloma cell lines with si[IGLCκ_CR] and si[-] (IM9, H929, JJN-3, and ARH77). By qPCR the residual κ LC mRNA was 13%, by flow cytometry the MFI was reduced by a median of 67.3% (22.5-90.8), and by ELISA mean κ LC secretion was reduced from 3.7 to 0.8µg/ml (P = 0.055, paired t test). We treated CD138-selected κ patient samples (AL 3, LCDD 1, MM 6) in the same way. By qPCR the residual κ LC mRNA was 57% control, by flow cytometry the MFI was reduced by a median of 37.5% (14-69.8), and by ELISA secretion was reduced from 9.4 to 6.5µg/ml (P = 0.02, paired t test). In the murine dual-flank xenograft model employing λ secreting cells, by qPCR there was a reduction in λ LC mRNA with si[IGLCλ_CR] treatment in 13 of 16 mice (ALMC-1 11/114, ALMC-2 2/2). In these 14, the median λ LC expression was 66% of control (range, 17-97). In 6/13 the average reduction in λ LC expression was 59%. Of note, measurable levels of human λ LC were found in the blood of all mice at sacrifice.

Conclusion With one pool of siRNA targeting the constant region of the κ or λ LC we can significantly reduce production and secretion of LC by clonal human plasma cells, including patient cells, and also reduce the expression of LC in xenograft plasmacytomas in vivo. Two methods of siRNA delivery have been employed in this work thus far, SLO and in vivo electroporation, neither of which require endosomal escape. The specificity of the siRNA pools for plasma cell LC genes and the possible receptivity of plasma cells to RNAi are important positive aspects of this work. Further pre-clinical development of Ig LC CR RNAi employing lipid-based nanoparticle platforms is warranted in order to optimize cell-specific delivery, delivery efficiency and siRNA targeting.