The Bone Marrow's Division of Labor in Cardiovascular Disease

Corresponding Author

The architecture and activity of the bone marrow fuel our immune system and shape inflammatory responses across a wide range of diseases, from cancer to autoimmune disease to cardiovascular disease. Leukocytes derived from hematopoietic stem cells are mobilized to meet physiological and pathophysiological demands, as illustrated by the emergency hematopoiesis that follows myocardial infarction.¹ This process is tightly regulated and operates at different spatial scales. Most attention to date has focused on the bone marrow niche, the local microenvironment in which hematopoiesis occurs and is regulated, shaped by, for example, innervation and vascularization.² Beyond this microarchitecture, however, our understanding of how anatomically distinct bone marrow sites contribute to disease remains incomplete. Neuroscience studies have shown that, for example, skull marrow is functionally distinct: comparatively protected from the hallmarks of aging, it maintains a unique molecular profile and connects directly to the brain surface through vascular channels that shape its response to stroke.³ Whether comparable site-specific specialization exists in cardiovascular disease has remained largely unexplored.

In this issue of JACC: Basic to Translational Science, Deckers et al⁴ propose a method to capture this site-level diversity, focusing specifically on metabolic activity. Their approach, MarrowMet, uses uptake of the common positron emission tomography tracer ¹⁸F-fluorodeoxyglucose as a readout. After intravenous injection, the tracer accumulates preferentially in metabolically active tissue; bones are then harvested and tracer content quantified ex vivo by gamma counting. This design sidesteps the limited spatial resolution of in vivo positron emission tomography imaging and yields a continuous, quantitative value per bone marrow site, at the cost of being a terminal, coarse-grained, single-timepoint measurement, a trade-off the investigators themselves acknowledge.

The investigators apply MarrowMet across a range of mouse models of immunologically relevant conditions, including inflammatory atherosclerosis, acute ischemia, acute respiratory distress syndrome, metabolic syndrome, and aging. The central finding is that activity differs across bone marrow sites, and that which sites are most active depends on the disease studied. This is a useful provocation: long-standing field conventions, such as the near-exclusive reliance on femoral marrow, deserve to be questioned when practicality rather than biology has driven the choice of sampling site.

MarrowMet makes this point by providing a protocol for assessing bone marrow metabolic shifts across sites side by side, but the inherent limitations of the tool warrant careful consideration. Most pressingly, changes in glucose uptake do not always correspond to changes in immune cell activation. In the investigators' aging cohort, ¹⁸F-fluorodeoxyglucose uptake declines with age even as the tumor necrosis factor responsiveness of bone marrow–derived macrophages rises, an inverse relationship that complicates any reading of tracer signal as a stand-in for inflammatory capacity. Beyond this, the method captures a bulk signal per site that cannot distinguish increased per-cell metabolic activity from a greater number of metabolically active cells. Glucose-independent immune activity, such as a shift in lymphoid-to-myeloid balance driven by progenitor-level changes or by altered maturation state rather than overt metabolic reprogramming, may not register in the assay at all. These constraints are inherent to the design, and they position MarrowMet as a screening tool for glucose-dependent, population-level metabolic shifts. Once such a shift is identified, the assay's interpretive power depends on pairing it with orthogonal data, as the authors demonstrate through in-depth profiling via cytometry by time of flight. This pipeline is elaborate, but it offers an unbiased route to regions of interest that might otherwise be overlooked.

The sternal findings illustrate this well. The investigators identify the sternum as the most metabolically active marrow site in the atherosclerotic mouse, suggesting it may play a disproportionate role in sustaining the chronic myeloid output that drives disease progression. This observation should invite the question: What does sternotomy do to this process? Cardiac surgery requiring median sternotomy transects and traumatizes precisely this marrow compartment, in patients who carry the full burden of the atherosclerotic disease that rendered the sternal marrow active in the first place. Whether that disruption attenuates, redirects, or ultimately reconstitutes the inflammatory output of sternal marrow, and on what timescale, is unknown. Because poststernotomy sternal fluorodeoxyglucose uptake can remain elevated for months to over a year in noninfected patients,⁵ the marrow's response to surgical trauma is itself prolonged, and its interaction with the underlying chronic disease is difficult to disentangle from existing imaging data. Yet the same scenario presents an unusual opportunity: sternal marrow is directly accessible during cardiac surgery, at a moment when the patient's atherosclerotic burden and systemic immune state are well characterized. Sampling, profiling, and, in principle, modulating this compartment at the time of surgery is a translational avenue the field has not yet explored, and one that the spatial reasoning of this study helps to frame.

In this context, the study contributes to a more holistic appreciation of the bone marrow as a scattered yet organized organ whose division of labor shapes the immune response in cardiovascular disease. A central question it leaves open is why certain sites are preferentially tasked. Does a unique molecular constitution, intrinsic to each site, drive the heterogeneity, or does proximity to the inflammatory focus? The relationship between skull marrow and the ischemic brain after stroke, mediated by direct vascular channels, suggests that anatomy and adjacency can be decisive. Whether analogous principles govern sternal dominance in atherosclerosis in mice, or the site-specific patterns seen in metabolic syndrome and aging, remains to be determined. Unwinding these principles and translating them to humans could reveal intervention targets that modify the immune response further upstream, and with greater precision than current systemic approaches permit.

The tools to pursue this are within reach. Spatially resolved multiomics, applied to the marrow sites flagged as most disease-relevant, can reveal the cellular and molecular interplays that govern disease progression on site⁶ and thus possibly elucidate how coordination between marrow sites is achieved. Coupled with multicohort data integration, these in-depth molecular profiles can be used to derive generalizable patient maps representing bone marrow disease trajectories, as has been pursued in heart failure.⁷ This could move the field from descriptive observation toward generalizable insights that pave the way for mechanistic understanding, and ultimately toward intervention.

The therapeutic rationale is already supported by clinical evidence. Anti-inflammatory therapies that dampen the inflammatory cascade downstream of hematopoietic activation, including interleukin-1β blockade with canakinumab, have reduced recurrent cardiovascular events in randomized trials. Notably, the benefit of canakinumab appeared concentrated in carriers of TET2-mutant clonal hematopoiesis, suggesting that interrupting interleukin-1β signaling may act on a marrow-originating inflammatory program.⁸ Gene editing of hematopoietic stem cells has crossed into clinical practice with the approval of Casgevy for sickle cell disease,⁹ and in vivo base editing of proprotein convertase subtilisin/kexin type 9 has entered phase 1 trials in familial hypercholesterolemia.¹⁰ The conceptual distance between these achievements and site-specific marrow intervention in atherosclerosis is narrowing. Identifying which compartments are most active, most accessible, and most causally linked to disease progression is a promising direction of future work that studies like this one can support and enable.

Funding Support and Author Disclosures

The authors have reported that they have no relationships relevant to the contents of this paper to disclose.