Multiple Myeloma Mouse Models Need More Than Phenotype Checklists
Multiple myeloma remains incurable despite major gains from proteasome inhibitors, immunomodulatory drugs, anti-CD38 antibodies, bispecific antibodies, and CAR-T therapy. The central problem for preclinical science is not the shortage of mouse models, but the lack of a consistent way to match each model to the biological question at hand. A model that supports engraftment may still fail at immune fidelity, while a model that preserves host immunity may miss the genomic diversity and clonal instability that define human disease.
The review frames this gap with unusual clarity. Rather than treating all mouse systems as interchangeable, it separates them into four major categories — syngeneic, xenograft, genetically engineered, and humanized — and measures each against the realities of myeloma biology: marrow tropism, immune escape, osteolytic injury, and disease progression. The need for that discipline is obvious in a malignancy that presents with hypercalcemia, renal dysfunction, anemia, and bone disease, yet evolves over months to years through clonal selection and treatment pressure.
One model rarely captures all of that at once.
Syngeneic Systems Preserve Immunity, But Not Human Heterogeneity
Syngeneic models remain a workhorse because they combine reliable engraftment with intact host immunity. Cells from an inbred strain are transplanted into mice of the same strain, so the graft is not rejected and innate and adaptive immunity stay functional. For myeloma immunology, that matters. It allows direct study of microenvironmental remodeling, immune surveillance, and treatment response in a living immune system rather than in a stripped-down host.
The C57BL/KaLwRij-5TMM series illustrates why these systems still matter. The KaLwRij mouse carries a susceptibility linked to a 180-kb deletion of Samsn1, with loss of SAMSN1 function favoring B-cell proliferation and higher immunoglobulin levels. Aged mice can spontaneously develop monoclonal gammopathy resembling MGUS and smoldering myeloma, including marrow plasma cell expansion and rising serum M-protein. The biologic relevance extends to humans because SAMSN1 is also downregulated in human myeloma.
Among the 5T derivatives, each line offers a different clinical window. The 5T33MM line behaves as an aggressive and rapidly progressive disease, reaching liver and spleen infiltration within 3 to 4 weeks after intravenous inoculation, but it does not produce osteolysis. The 5TGM1 line, by contrast, developed after in vivo passage and does generate marked osteolytic disease with high serum M-protein levels. An eGFP-tagged version enables in vivo imaging, and ibandronate reduced vertebral and long-bone osteolysis in this setting. The 5T2MM line moves in a slower, marrow-dependent fashion over about 12 weeks and is better suited to bone disease and slow-progression work than to frantic tumor kinetics.
These are not interchangeable tools. They answer different questions, and each leaves an important gap behind.
When Stable Disease Reveals Immune Surveillance
The C57BL/6 myeloma stable/resistant control model is especially useful because it exposes host control mechanisms that disappear once disease accelerates. After intravenous 5TGM1 injection, only about half of C57BL/6 mice develop disease, and many remain stable for up to 5 weeks before progression. During that stable period, natural killer cells, type 1 innate lymphoid cells, and CD8+ T cells remain activated and proliferative. As disease advances, immune activity wanes.
Interferon biology sits at the center of that distinction. Single-cell and bulk RNA sequencing identified interferon-response activation as the hallmark of the stable phase, and neutralization of IFN-γ accelerated disease onset. For translational hematology, the lesson is direct: local marrow immunity can suppress myeloma early, and immune failure is not an abstract concept but an observable shift in host state. This makes the model valuable for studying immune surveillance in immunocompetent, gene-editable mice, especially when the clinical question involves escape from endogenous antitumor control.
Bone Disease Still Separates Useful Models from Convenient Ones
The bone phenotype remains one of the sharpest filters in model selection. Human myeloma is defined in part by osteolytic injury, yet many classic systems either produce incomplete bone lesions or none at all. The MOPC315 model, for example, originates from a long-standing plasmacytoma induction system and is useful historically, but the tumors are confined to the peritoneal cavity and poorly reproduce skeletal involvement. Its bone marrow-homing subclone, MOPC315.BM, improves that limitation substantially. After tail vein injection, it colonizes marrow and spleen and produces osteolytic destruction, hypercalcemia, an IgA-λ M-spike, and paraplegia, offering a more CRAB-like phenotype.
Bone destruction in myeloma is not a cosmetic feature of the disease. It reflects the collision between malignant plasma cells and the marrow niche, where osteoclast activation, marrow remodeling, and tumor support feed each other. Models that do not reproduce that circuit may still be useful for pharmacology, but they are weaker platforms for dissecting skeletal complications or for testing agents that are meant to interrupt the tumor-bone axis.
Xenograft and Humanized Models Trade Immune Fidelity for Human Material
Xenograft models bring human cells into immunodeficient mice, most often NSG hosts. That makes them highly valuable for pharmacodynamic work and biomarker validation because they retain important genomic characteristics and, in some settings, clonal heterogeneity from the donor tumor. Their weakness is structural: the immune microenvironment is missing. For any therapy that depends on T cells, NK cells, cytokine circuits, or antigen presentation, that absence can distort the result.
Humanized models try to close that gap by introducing human immune elements, human marrow niche components, or both. In principle, these systems move closest to clinical myeloma biology because they reduce species mismatch. In practice, they are expensive, technically demanding, and still not complete substitutes for the patient setting. The review treats them as valuable but demanding tools, best reserved for questions that truly require human immune context.
A 2 + 1 Cross-Model Rule Could Reduce False-Positive Signals
The proposed 2 + 1 cross-model validation framework is the most operational part of the review. Any candidate therapy should be tested in at least two mechanistically distinct models, and one of them must be immunocompetent. That requirement is meant to blunt the problem of false-positive efficacy signals, which often arise when a drug performs well in one narrow system but fails when the immune context or marrow biology changes.
This approach is clinically sensible because myeloma therapy is no longer limited to direct cytotoxicity. Modern treatment includes immune redirection, antibody-based therapy, and cellular immunotherapy. A candidate that appears active in an immunodeficient xenograft may fail once host immunity becomes relevant, and a therapy that works in a syngeneic bone model may still stumble when confronted with human tumor heterogeneity.
The review also introduces a Clinical Translation Scoring System that weights genomic fidelity at 25%, immune integrity at 25%, bone disease reproducibility at 20%, and disease course dynamics at 30%. The scoring tool is explicitly exploratory and still needs multicenter validation, but the structure is useful because it forces investigators to say what matters most in a given experiment. A marrow-homing immunotherapy study does not need the same model hierarchy as a project focused on skeletal destruction or clonal evolution.
Model Choice Should Follow the Clinical Question
The practical value of this review lies in restraint. No single mouse model can capture all of myeloma: chronicity, relapse, drug resistance, bone disease, and immune escape do not travel together in every system. Syngeneic models preserve immunity but narrow human genetic diversity. Xenografts expand human relevance but shed immune function. Genetically engineered models reproduce endogenous tumor development but often progress slowly and can produce incomplete bone lesions. Humanized models come closest to the patient, yet remain costly and demanding.
That tradeoff is the point. Preclinical work in myeloma becomes more believable when the model is chosen for the exact biological question rather than for convenience alone. The review leaves one clear boundary in place: its scoring system is promising, but it remains an exploratory tool until tested across centers and applied against real translational outcomes.