Our models:
advantages and silk properties

The Silk4B 3D bone marrow models reproduce the architecture, mechanics and biochemical signalling of the human niche.

Silk4B's models are the first to reproduce both the biology and the mechanics of the bone marrow in vitro. They are built on silk fibroin, a purified biocompatible protein.

• The material behind the model

  Silk fibroin is a purified, well-defined protein with a combination of properties that makes it well-suited to modelling bone marrow as a whole, rather than any single lineage within it.
  
  It can be processed across a broad, tunable stiffness range, including values that match healthy human marrow. This is what allows the model to reproduce the mechanical cues that shape hematopoiesis. Silk fibroin is also well tolerated by the range of cell types the marrow niche contains, from HSPCs to stromal and vascular cells, supporting long-term culture without the material itself confounding the biology.

  
  Together, these properties make it possible to build a single platform that supports the full range of marrow biology: HSPC maintenance, multi-lineage differentiation, stromal and vascular co-culture, and disease modelling.

• Mechanics That Match Human Marrow

  The marrow microenvironment regulates hematopoiesis through matrix stiffness and structure, not biochemistry alone. Take HSPCs out of that mechanical context and they behave differently, losing stemness and narrowing their differentiation potential.
  

  Silk4B's softer formulations sit within the stiffness range reported for healthy human marrow, 0.1 to 10 kPa by AFM. This isn't just for HSPC maintenance: other marrow cells depend on it too. Megakaryocytes use mechanosensors to read matrix stiffness and adjust their behaviour accordingly. On soft silk, megakaryocytes extend long, branched proplatelets; as stiffness rises, proplatelet formation declines.
  

  Because stiffness is tuneable, the same platform can reproduce the mechanical progression from healthy marrow through early and late fibrosis, letting researchers isolate whether a phenotype comes from the cells themselves or from the matrix around them.

• One Model, Any Level of Complexity

  The model scales to fit the question you're asking. Use a single cell type for high-throughput screening or to dissect a specific pathway, such as platelet formation or erythropoiesis. Add a second cell type, such as megakaryocytes with neutrophils, to study how two populations interact. Or build a three-plus cell system, such as HSPCs with megakaryocytes, endothelial cells, and stromal cells, to recreate more of the niche at once. This flexibility makes it possible to isolate the contribution of individual cell types and cell-cell interactions in both normal and diseased states.

• Functional Cells

  Silk4B models mimic the properties of in vivo bone marrow, supporting successful differentiation into mature cells as well as enhanced preservation and functional incorporation of progenitor cells collected from patients or animal models.
  
  

  Platelets released from megakaryocytes in the niche pass activation, adhesion, aggregation, and clot-participation testing.

  
  Red blood cells are produced fully enucleated, the point where most ex vivo systems stop.

  
  Neutrophils introduced from peripheral blood survive better than in standard pellet-based preservation, remain functional, and boost platelet production.

• A Model That Predicts, Not Just Describes

  Because Silk4B models reproduce the mechanical and biochemical environment cells actually respond to, drug response measured in the model tracks what happens in the clinic. In one validated example, platelet output measured in the model correctly identified which thrombocytopenia patients would go on to respond to treatment, demonstrating that the platform can be used to predict clinical drug response, not just describe cell behaviour in a dish. The same system works with patient peripheral blood progenitors and with iPSC-derived cells, making it applicable across a range of disease and drug-testing programmes.

• Disease, Built to Order

  Silk4B models aren't limited to healthy physiology: they can be pushed to reproduce disease states. Matrix and cellular composition can be adjusted to mimic fibrotic, malignant, or otherwise pathological marrow, letting researchers study disease progression and test candidate drugs in a setting that behaves like real marrow rather than a simplified stand-in.

  
  Malignant or patient-derived cells can be incorporated in any configuration: a fully diseased bone-marrow niche, or a single diseased cell type set against a healthy-donor niche, making the platform suited to both mechanistic disease research and target identification for drug discovery.

• Standardised and validated

  The underlying culture protocols were established across thousands of blood samples and published as a reference standard for the field. Step-by-step protocols for 3D culture, differentiation, and blood cell production are already published; adoption doesn't mean building an assay from scratch.