Losing CD38 in myeloma Free

In this issue of Blood, Diamond et al¹ have illuminated a new mechanism of how multiple myeloma (MM) occasionally escapes anti-CD38 monoclonal antibodies by deleting or mutating the CD38 gene, yet their data simultaneously underscore that most clinical resistance still stems from subtler antigen downregulation and immune dysfunction rather than genetic absence of the target.

Anti-CD38 antibodies such as daratumumab and isatuximab have become foundational in myeloma therapy, from frontline therapy through each relapse.^2-5 Consequently, every patient eventually confronts resistance, making its biological basis a question of pressing translational importance. Early work catalogued multiple nongenetic obstacles, such as rapid internalization, trogocytosis, complement pathway insufficiency, and a progressively exhausted immune effector pool, which collectively blunt antibody cytotoxicity.⁶ 

Yet the field lacked a systematic examination of whether myeloma can dodge therapy by simply discarding CD38 itself, a mode of escape well described for CD19 in B-cell malignancies.⁷ Diamond et al set out to answer this by performing whole-genome and whole-exome sequencing on 701 newly diagnosed MM cases, 67 relapsed cases naïve to anti-CD38 therapy, and 50 relapsed cases previously treated with anti-CD38 antibodies. Their elegant genomic survey registered complete, biallelic inactivation of CD38 in 6% of post–anti-CD38 therapy specimens.

The authors went beyond enumeration. They demonstrated that monoallelic lesions, although insufficient to eliminate the antigen, suppress transcription from the remaining allele to levels below the therapeutic threshold, providing a mechanistic link between heterozygous hits and functional escape. CRISPR-engineered plasma cell lines carrying the identified missense variants showed reduced surface staining, and binding assays revealed nuanced specificity: substitutions L153H and C275Y abrogated recognition by both daratumumab and isatuximab, whereas the R140G variant selectively impeded daratumumab but left isatuximab cytotoxicity intact, raising the prospect that sequencing might guide antibody choice at relapse.

Placed in clinical context, these results recalibrate our view of how often MM deploys “hard” genetic escape. Resistance to anti-CD38 therapy is near universal, yet the genotype Diamond et al describe is demonstrably rare, implying that most failures occur despite intact antigen coding sequences. This conclusion is concordant with emerging immunoprofiling data indicating that, in many patients with progressing disease, malignant plasma cells continue to express appreciable CD38, whereas the surrounding immune compartment displays characteristic signatures of functional exhaustion and paralysis: diminished interferon-γ production, exhausted T-cell signatures, and depletion of CD38⁺ natural killer cells that normally mediate antibody-dependent cellular cytotoxicity.⁸ 

The novelty of this study resides in 3 interlocking insights. Foremost, it furnishes the first rigorous estimate of the prevalence of complete CD38 ablation, demonstrating that, although rare, this event represents a definitive route to therapeutic failure. Second, it establishes that monoallelic events can be almost as deleterious as biallelic events because they suppress transcription and hence protein density, eroding antibody binding without total gene loss. Third, it uncovers mutation-specific differences in antibody recognition, suggesting that daratumumab and isatuximab, although both targeting CD38, do not share identical epitopes and therefore may be sequenced rationally on the basis of mutational fingerprints. Together these findings refine our therapeutic calculus: although genetic sequencing will not explain most cases of resistance, it can provide decisive guidance for a minority in whom antigen integrity is compromised.

As with any investigation, this study is subject to certain limitations. The clinical cohort, although deeply analyzed, remains small and heterogeneous in previous therapies, disease stage, and sample timing. Not every patient had a paired pretherapy genome, so the possibility that rare CD38-null subclones existed before treatment cannot be completely excluded. Functional validation relied on overexpression systems, leaving open whether primary plasma cells tolerate total loss of CD38 enzymatic activity in vivo or incur a fitness penalty that shapes clonal competition. Moreover, the authors did not integrate single-cell immune-microenvironment data, precluding analysis of how gene loss interacts with effector exhaustion, complement deficiencies, or cytokine milieu.

The translational message, nevertheless, is clear. For the vast majority of patients with myeloma, resistance is likely to hinge on reversible CD38 downregulation and an eroding immune response rather than irrevocable gene deletion. Therapeutic strategies addressing these liabilities, such as agents that upregulate CD38 transcription such as all-trans retinoic acid and hypomethylating drugs,⁹ therapies that reinvigorate exhausted cytotoxic T lymphocytes, or graft in fresh effector populations¹⁰; and approaches that bypass CD38 altogether, exemplified by B-cell maturation antigen chimeric antigen receptor T cells and bispecific antibodies, promise broader benefit than simply switching from one anti-CD38 antibody to another. Nevertheless, for the minority with demonstrable antigen mutations, genomic data can now inform the choice and sequencing of available antibodies or even the engineering of next-generation molecules with dual-epitope binding capable of accommodating those mutations.

Looking forward, the field would profit from prospective trials that serially integrate single-cell genomics, epigenomics, and high-parameter cytometry from diagnosis through each therapeutic line. Such efforts could ascertain whether CD38-null or CD38-low clones arise only after immune surveillance falters, determine the evolutionary fitness cost of antigen loss, and pinpoint combination regimens that prevent convergent evolution. Parallel antibody design, arming CD38 antibodies with enhanced Fc receptor engagement or fusing them to T-cell redirecting domains, may drive the tumor into an evolutionary corner in which escape becomes biologically untenable.

In summary, Diamond et al expose a dramatic but uncommon genetic escape hatch, reveal that monoallelic lesions can depress antigen density enough to erode efficacy, and show that discrete mutations differentially cripple our 2 licensed antibodies. The broader challenge, however, lies in sustaining robust CD38 expression and a competent immune effector compartment so that the antibodies retain their lethal edge against a disease that, for most patients, still harbors the target even at relapse.

Conflict-of-interest disclosure: M.M. declares no competing financial interests.