Myeloma Cells Can Permanently Lose CD38 to Escape Daratumumab Therapy
Genomic study reveals 20% of post-CD38-therapy myeloma patients show CD38 loss, including biallelic escape conferring permanent drug resistance.
Summary
A new genomic study in Blood analyzed 50 multiple myeloma patients who relapsed after CD38-targeted antibody therapy (daratumumab or isatuximab). Researchers found that 20% (10/50) had lost CD38 expression due to genomic alterations, with 6% (3/50) showing full biallelic loss of both gene copies, representing permanent resistance. Remarkably, some patients showed convergent evolution, where independent tumor subclones each developed separate CD38-inactivating mutations. Functional testing identified specific missense mutations (L153H, C275Y, R140G) that reduce antibody binding and tumor killing. One mutation, R140G, selectively blocked daratumumab while preserving isatuximab sensitivity — suggesting genomic testing could guide which CD38 antibody to use on retreatment. No biallelic events were seen in 701 newly diagnosed or 67 antibody-naïve relapsed patients.
Detailed Summary
Multiple myeloma, a cancer of plasma cells in the bone marrow, is now routinely treated with CD38-targeting antibodies like daratumumab and isatuximab. These drugs work through multiple immune mechanisms, including antibody-dependent cellular cytotoxicity (ADCC), phagocytosis, and complement activation. Despite impressive early responses, virtually all patients eventually relapse. Previous research identified transient CD38 downregulation as one escape mechanism, but whether tumors can acquire permanent, genomic-level resistance by disrupting the CD38 gene was unknown — a gap this study directly addressed.
Researchers from institutions including the University of Miami, Mayo Clinic, University of Calgary, and Heidelberg University conducted whole genome sequencing (WGS, 60–100x coverage) and whole exome sequencing (WES) on CD138+ myeloma cells from 50 patients who relapsed following anti-CD38 antibody therapy (28 from Mayo Clinic, 22 from University of Calgary). They compared findings against 701 newly diagnosed MM patients from the CoMMpass trial and 67 anti-CD38-naïve relapsed/refractory patients. Clonal architecture was resolved using DPClust and phylogenetic trees, with structural variants called by SvABA, Manta, and DELLY.
The headline result: 20% (10/50) of post-CD38-therapy patients showed CD38 locus loss, with 6% (3/50) demonstrating full biallelic disruption — meaning both copies of the CD38 gene were inactivated, creating potentially permanent resistance. No biallelic events were detected in either the 701 newly diagnosed or 67 CD38-naïve relapsed patients, strongly implicating anti-CD38 therapy as the selective pressure driving these events. Two of the three biallelic cases displayed convergent evolution, in which distinct subclones independently acquired different inactivating CD38 alterations, underscoring strong clonal selection under antibody pressure.
Functional studies were conducted by engineering CD38 missense variants (L153H, C275Y, R140G) into K562 cells and measuring daratumumab and isatuximab binding affinity and ADCC. Both L153H and C275Y substantially reduced binding and killing by both antibodies. Critically, R140G selectively abrogated daratumumab binding and ADCC while retaining near-normal sensitivity to isatuximab — a finding corroborated by three-dimensional protein modeling in PyMOL showing that R140G alters the daratumumab-specific epitope without disrupting the isatuximab-binding region. This differential sensitivity has direct clinical implications for antibody selection on retreatment.
Flow cytometry using a polyclonal anti-CD38 antibody targeting multiple epitopes (CYT-38F2, Cytognos) confirmed that patients with biallelic CD38 events showed absent or dramatically reduced surface CD38 expression on tumor cells. In cases with monoallelic loss, partial CD38 reduction was observed. RNA-sequencing further validated that genomic CD38 alterations correlated with reduced transcript levels. The authors acknowledge that the cohort is heterogeneous (variable treatment histories, median 2 prior lines beyond CD38 therapy before sample collection) and relatively small, limiting precise prevalence estimates. Nevertheless, the complete absence of biallelic events in over 768 control patients makes the association with anti-CD38 therapy compelling and clinically actionable.
Key Findings
- 20% (10/50) of patients relapsing after anti-CD38 antibody therapy showed CD38 locus loss; 6% (3/50) had complete biallelic disruption of the CD38 gene
- Zero biallelic CD38 events were detected in 701 newly diagnosed MM patients or 67 anti-CD38-naïve relapsed/refractory patients, implicating therapy as the selective pressure
- Two of three biallelic-loss cases showed convergent evolution — independent subclones separately acquiring distinct CD38-inactivating mutations under antibody pressure
- Missense mutations L153H and C275Y significantly reduced both daratumumab and isatuximab binding affinity and ADCC killing in K562 cell functional assays
- Mutation R140G selectively abolished daratumumab binding and ADCC while preserving isatuximab sensitivity, confirmed by 3D protein epitope modeling in PyMOL
- Monoallelic CD38 locus loss (without full biallelic inactivation) was found in an additional 14% (7/50) of post-CD38-therapy patients
- Flow cytometry with a polyclonal multi-epitope CD38 antibody confirmed absent or severely reduced surface CD38 protein expression in biallelic-loss patients
Methodology
This was a retrospective genomic study analyzing WGS (60–100x) and WES from CD138+ plasma cells of 50 post-anti-CD38-therapy relapsed MM patients (Mayo Clinic n=28, University of Calgary n=22), compared against 701 newly diagnosed and 67 CD38-naïve relapsed MM patients as controls. SNVs were called using Mutect2, Strelka, and Lancet; copy number alterations via GATK4/ASCAT; structural variants via SvABA, Manta, and DELLY; clonal architecture via DPClust. Functional validation used CD38 mutant-expressing K562 cells co-cultured with PBMCs at 20:1 effector-to-target ratio with daratumumab or isatuximab, measuring ADCC by flow cytometry, and structural modeling via PyMOL.
Study Limitations
The post-CD38-therapy cohort is relatively small (n=50) and heterogeneous, with patients having received variable numbers of prior treatment lines and sample collection often occurring after additional therapies beyond CD38 antibodies (median 2 prior lines post-CD38), which could confound attributing genomic events specifically to CD38-antibody pressure. The study design is retrospective and lacks paired pre/post-CD38-therapy samples for most patients, limiting definitive proof of clonal emergence under therapy in the majority of cases. Several authors have affiliations with Sanofi (manufacturer of isatuximab), representing a potential conflict of interest in the functional comparisons between daratumumab and isatuximab.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.