A humanized ossicle model of myelofibrosis reveals THPO-driven fibrosis, osteosclerosis and SPP1-dependent microenvironmental remodeling

This study establishes a humanized ossicle model of myelofibrosis driven by THPO overexpression that recapitulates key disease features like fibrosis and osteosclerosis, identifying SPP1 as a critical mediator of niche remodeling and a promising therapeutic target.

The Gist

The Big Picture: Building a "Human Bone Marrow" in a Mouse

Imagine your bone marrow as a bustling factory where your blood cells are manufactured. In a disease called Myelofibrosis (MF), this factory gets ruined. Instead of making healthy blood, the factory floor gets clogged with scar tissue (fibrosis), the walls get thick and hard (osteosclerosis), and the workers (blood cells) get kicked out of the building to work in the wrong places (like the spleen).

The problem is that doctors don't have a great way to test new cures for this. Mouse models are like trying to fix a human car using a toy car; the parts don't quite match. Human cell models in a petri dish are like a blueprint; they look right, but they don't show how the engine actually runs in a real environment.

The Solution: The scientists in this paper built a "Human Bone Marrow in a Mouse."
They took human bone-building cells and grew them under the skin of special mice (who have no immune system to reject them) to create a tiny, artificial human bone called an ossicle. Think of this as building a miniature, human-sized apartment complex inside a mouse. Once the complex was built, they moved in human blood stem cells to see what would happen.

---

The Experiment: What Happened When They Added "Too Much" Signal?

In Myelofibrosis, a specific signal called THPO (Thrombopoietin) goes haywire. It's like a foreman in the factory who won't stop shouting orders.

The scientists took human blood stem cells and gave them a genetic "superpower": they made these cells produce way too much THPO. They then moved these "shouting" cells into the human bone marrow apartment complex inside the mice.

The Result: The apartment complex quickly turned into a Myelofibrosis factory.

1. The Scar Tissue: The floor got covered in a thick net of scar tissue (fibrosis), just like in human patients.
2. The Hard Walls: The walls of the apartment got thicker and harder (osteosclerosis).
3. The Eviction: The blood cells stopped staying in the apartment and started fleeing to the mouse's spleen and other organs (extramedullary hematopoiesis).

This proved that their new model works perfectly. It's a realistic simulator that mimics the human disease, allowing scientists to test new drugs in a living system that actually looks and acts like a human patient.

---

The Discovery: Finding the "Glue" Holding the Disaster Together

Once they had this perfect model, the scientists asked: "What is the glue holding this scar tissue together? Can we cut that glue to fix the factory?"

They discovered a protein called SPP1 (also known as Osteopontin).

• The Analogy: Imagine SPP1 as super-strong construction glue. In a healthy factory, a little glue is fine. But in this disease, the "shouting foreman" (THPO) tells the construction workers (bone cells) to dump buckets of this glue everywhere.
• This glue sticks the scar tissue together and tells the bone cells to build too much hard wall.
• The scientists found that this same "glue" was also present in high amounts in the bone marrow of actual human Myelofibrosis patients.

---

The Cure: Dissolving the Glue

To test if this "glue" was the real villain, the scientists gave the mice an anti-SPP1 antibody.

• The Analogy: Think of this antibody as a chemical solvent that dissolves the super-strong glue.

The Result: When they dissolved the glue:

1. The scar tissue (fibrosis) started to break down.
2. The blood cells stopped acting crazy and returned to a more normal balance.
3. The "eviction" of cells to the spleen slowed down.
4. The excessive hardening of the bone walls decreased.

Why This Matters

This paper is a two-step victory:

1. The Tool: They built a better "test drive" car (the humanized mouse model) that lets them see how human diseases really work, which is a huge leap forward from old mouse models.
2. The Target: They found a specific "glue" (SPP1) that drives the disease. By neutralizing it, they showed that it's possible to reverse the damage.

In simple terms: They built a human factory inside a mouse, broke it on purpose to study the damage, found the specific "glue" causing the mess, and showed that dissolving that glue can start to clean up the factory. This opens the door for new drugs that could actually cure or reverse Myelofibrosis, rather than just treating the symptoms.

Technical Summary

1. Problem Statement

Myelofibrosis (MF) is the most severe form of myeloproliferative neoplasms (MPNs), characterized by bone marrow (BM) fibrosis, extramedullary hematopoiesis, cytopenias, and osteosclerosis.

• Current Limitations: Existing therapies (e.g., JAK inhibitors) provide symptomatic relief but rarely reverse fibrosis or halt disease progression. Allogeneic stem cell transplantation is the only cure but is limited by patient age and high morbidity.
• Model Deficiencies:
  • Murine Models: While useful, they suffer from species-specific differences in cytokine signaling, megakaryocyte biology, and stromal responses, limiting translational relevance.
  • Xenotransplantation Models: Transplanting human mutated cells into immunodeficient mice often fails to induce robust fibrosis due to a lack of interaction between human hematopoietic cells and murine stromal cells.
  • In Vitro Models: Organoids lack the spatial organization, vascularization, and long-term remodeling dynamics of in vivo bone marrow.

Objective: To establish a humanized in vivo model that faithfully recapitulates human MF pathology, specifically focusing on the interaction between human hematopoietic cells and a human stromal microenvironment, to identify and validate novel therapeutic targets.

2. Methodology

The authors developed a humanized ossicle model in immunodeficient NSG mice using a two-step transplantation strategy:

• Step 1: Generation of Humanized Ossicles (huOssicles):
  • Human bone marrow mesenchymal stromal cells (MSCs) from healthy donors were expanded in vitro.
  • MSCs were mixed with a Matrigel-equivalent matrix and injected subcutaneously into NSG mice.
  • Mice received daily human parathyroid hormone (PTH) injections to promote bone formation.
  • Over 8 weeks, this formed ectopic bony ossicles containing human cortical/trabecular bone, adipocytes, and vasculature.
• Step 2: Induction of Myelofibrosis (huMF Ossicles):
  • Human CD34⁺ hematopoietic stem and progenitor cells (HSPCs) were isolated from healthy donors.
  • Cells were transduced with lentivirus to overexpress Thrombopoietin (THPO) (THPO-GFP) or a control (GFP).
  • These cells were injected directly into the ossicles (intra-ossicle, i.o.) of the mice.
• Experimental Groups & Interventions:
  • Control: GFP-transduced CD34⁺ cells.
  • Disease Model: THPO-overexpressing CD34⁺ cells.
  • Therapeutic Intervention: A subset of huMF mice received intraperitoneal injections of an anti-human SPP1/OPN antibody (neutralizing Osteopontin) starting at week 4, compared to isotype controls.
• Analytical Techniques:
  • Histology: Reticulin staining (Gordon & Sweet's) for fibrosis grading; H&E for cellularity; Trichrome RGB staining to distinguish mineralized bone vs. unmineralized osteoid.
  • Flow Cytometry: Analysis of human CD45⁺ engraftment, myeloid/lymphoid ratios, and megakaryocyte markers (CD41a, CD42b) in ossicles, spleen, femur, and liver.
  • Immunohistochemistry (IHC) & Immunofluorescence (IF): Staining for human mitochondria (to confirm human origin), CD61 (megakaryocytes), CD56 (osteoprogenitors), and SPP1/OPN.
  • Quantification: Automated image analysis (QuPath) for bone area, megakaryocyte clustering, and fibrosis grading.

3. Key Contributions

1. Novel Humanized Platform: Successfully established a humanized ossicle model where human stromal cells and human HSPCs co-exist, allowing for the study of human-specific hematopoietic-stromal crosstalk in vivo.
2. THPO-Driven Pathogenesis: Demonstrated that THPO overexpression in human CD34⁺ cells is sufficient to drive the full spectrum of MF pathology (fibrosis, osteosclerosis, extramedullary hematopoiesis) within a human stromal niche.
3. Identification of SPP1/OPN: Identified Secreted Phosphoprotein 1 (SPP1/Osteopontin) as a critical mediator of the pathological microenvironment, linking osteoprogenitor activation to fibrosis.
4. Therapeutic Validation: Provided in vivo proof-of-concept that neutralizing SPP1/OPN can partially reverse MF phenotypes, suggesting a new therapeutic avenue beyond JAK inhibition.

4. Key Results

A. Establishment of the Model

• Humanized ossicles successfully supported robust engraftment of human CD45⁺ cells (48.9% in ossicles vs. 2.3% in murine femurs).
• The ossicles contained human-derived osteocytes, stromal elements, and distinct anatomical regions (endosteal, periosteal).

B. Recapitulation of MF Features in huMF Ossicles

• Myeloid Skewing: THPO-overexpressing ossicles showed a significantly higher myeloid-to-lymphoid ratio compared to controls.
• Megakaryocyte Dysplasia: Increased frequency of CD41a⁺CD42b⁺ megakaryocytes and significant clustering (a hallmark of MF).
• Progressive Fibrosis:
  • Fibrosis was absent at 4 weeks but emerged by week 5.
  • By week 8, THPO ossicles exhibited dense reticulin fiber networks (Grade 2-3), extending between hematopoietic areas and along trabecular bone, mimicking human MF.
• Extramedullary Hematopoiesis (EMH): Human hematopoietic cells migrated from the ossicles to murine hematopoietic organs, with a significant increase in the spleen (mimicking splenomegaly/EMH) and a trend in the femur.
• Osteosclerosis:
  • THPO ossicles showed increased total bone area and trabecular bone formation.
  • Trichrome RGB staining revealed an accumulation of unmineralized osteoid, indicating active but dysregulated bone remodeling.

C. Mechanistic Role of SPP1/OPN

• Upregulation: SPP1/OPN expression was significantly elevated in THPO ossicles and in bone marrow biopsies from primary MF patients compared to healthy controls.
• Cellular Source: SPP1 was highly expressed in CD56⁺ osteochondroprogenitors, suggesting these cells drive the fibrotic and osteogenic remodeling.
• Therapeutic Intervention (Anti-SPP1):
  • Reduced Fibrosis: Antibody treatment significantly decreased reticulin fiber deposition.
  • Restored Homeostasis: Attenuated myeloid skewing and reduced megakaryocyte expansion.
  • Trends: Observed trends toward reduced splenic engraftment and bone area, though not statistically significant in this short-term study.

5. Significance

• Translational Relevance: This model bridges the gap between murine models (species mismatch) and in vitro systems (lack of physiological complexity). It allows for the testing of therapies targeting the human microenvironment, which is often missed in standard mouse models.
• Novel Therapeutic Target: The study identifies SPP1/OPN not just as a biomarker, but as a functional driver of MF. Neutralizing SPP1 partially rescued the disease phenotype, proposing a new class of anti-fibrotic therapies that target the niche rather than just the malignant clone.
• Pathophysiological Insight: The model confirms that aberrant THPO signaling in hematopoietic cells can reprogram human stromal progenitors (specifically CD56⁺ osteoprogenitors) to drive fibrosis and osteosclerosis, validating the concept of hematopoietic-stromal crosstalk in MF pathogenesis.

Conclusion: The authors have created a robust, humanized in vivo platform that faithfully models the complex interplay of fibrosis, osteosclerosis, and extramedullary hematopoiesis in myelofibrosis. The validation of SPP1/OPN as a therapeutic target offers a promising strategy for developing disease-modifying treatments for MF.