How Knee Stem Cell Injections May Influence Native Repair Cells: What the Research Shows

Marc Darrow, MD · Updated November 11, 2026

Your knee already contains cells with repair potential. The unanswered question is how reliably an injection can help those cells produce a meaningful, lasting improvement.

When patients hear “stem cell therapy,” they often picture injected cells becoming new cartilage and replacing damaged tissue. That is one proposed biological pathway, but it is an incomplete explanation of the research—and it can create expectations the clinical evidence does not support.

My interest in this subject includes the environment inside the knee: inflammation, the remaining cartilage, the joint lining, and communication between cells. Researchers are studying whether biologic treatments can influence that environment and support the activity of the knee’s own cells. We need to distinguish this possibility from proof that a particular injection rebuilds a patient’s knee.

Does the knee already have its own stem cells?

The knee contains several populations of cells described as stem or progenitor cells. A progenitor cell has the potential to produce certain specialized cells under suitable conditions; it is a cell, rather than a growth factor. These populations have been identified in the synovium, synovial fluid, cartilage, bone beneath the cartilage, and fat-containing tissues. Their roles in normal maintenance and osteoarthritis remain incompletely understood. (1)

The synovium is the lining inside the joint capsule. It helps maintain synovial fluid and a functioning joint environment. In a foundational 2011 mouse study, investigators identified resident mesenchymal cell niches in the knee synovium and observed a proliferative and cartilage-forming response after injury. This supports an intrinsic repair response in that experimental setting. It was not a trial showing that an injection regenerates human knee cartilage. (2)

A later review discussed synovial-fluid-derived cells and their potential in knee arthritis and cartilage-defect research. Finding cells capable of forming cartilage-like tissue under laboratory conditions is a useful scientific starting point. It does not show that those cells will reconstruct a worn joint surface on their own. (3)

If repair cells are present, why does osteoarthritis progress?

Having cells with repair potential is different from having the conditions needed for effective repair. Osteoarthritis involves the entire joint, and the ability of cells to respond may change with disease.

A 2012 study compared progenitor cells from normal and osteoarthritic synovial fluid. Their cartilage-forming behavior differed under the culture conditions studied, raising questions about cell aggregation, communication, and the joint environment. It provides a reason to investigate impaired repair responses. It does not establish that all arthritic knees simply need an injection to “restart” their cells. (4)

Single-cell research has also shown different functional states among synovial fibroblasts in osteoarthritis and rheumatoid arthritis. Fibroblasts are connective-tissue cells; they should not all be described as stem cells. This work concerns signaling pathways and cell states, rather than proof that native stem cells cause knee swelling. Rheumatoid arthritis and osteoarthritis are distinct diseases despite some shared biological pathways. (5)

In clinical terms, an impaired repair response may involve several interacting problems. A useful explanation should leave room for joint loading, tissue loss, inflammation, and cell function. The metaphor of “confused cells” can help introduce communication, but it should not replace this more complex biology.

How might injected cells communicate with cells already in the knee?

One proposed mechanism is paracrine signaling: cells release molecules that influence nearby cells. Researchers have studied effects on inflammatory activity, cartilage-cell behavior, and the production or breakdown of the matrix surrounding cells. Direct contact between cells may also influence their behavior. A 2016 review discussed these potential reparative pathways. (6)

A May 2026 review emphasizes the shift from describing mesenchymal cells mainly as replacement building blocks toward examining their signaling and immune-modulating functions. It also explains why terminology and preparation methods matter. These are proposed mechanisms across a varied research field; they do not establish that every marketed preparation has the same biological activity. (7)

This is why I would describe the treatment rationale as an attempt to influence the local environment and host response. I would not tell a patient that an injection has been proven to organize the knee’s native cells into a new cartilage surface.

A plausible mechanism explains why a treatment deserves study. Clinical trials must establish whether patients benefit, how much, and for how long.

Bone marrow concentrate, PRP, and purified cells are different preparations

Bone marrow aspirate concentrate—BMAC or BMC—is prepared from the patient’s own bone marrow. It contains a mixture of cells and signaling components, including a small population of progenitor cells. It is not a purified dose of culture-expanded mesenchymal stromal cells.

Platelet-rich plasma, or PRP, is prepared from blood and concentrates platelets and their associated signaling molecules. PRP is not a stem-cell preparation. Donor-derived tissue products and isolated extracellular vesicles are different again.

These differences limit how confidently a study can be applied to another treatment. A laboratory experiment using expanded cells cannot establish what non-expanded bone marrow concentrate does inside a human knee. Likewise, a total nucleated-cell count should not be presented as a count of mesenchymal stem cells.

Can native cells help repair a torn meniscus?

The meniscus is another important part of the original scientific question. Its outer region has a better blood supply than its inner region, but blood supply is not the only factor influencing repair.

Research published in 2017 examined cells migrating from injured meniscal tissue. The investigators found progenitor-like properties and evidence of an intrinsic response to injury, including the ability to populate experimental defects. This was experimental tissue and animal work, rather than a clinical trial of an injection healing human meniscus tears. (8)

The useful learning point is that the meniscus has biological activity worth studying. The finding does not mean every degenerative tear will heal after a stem-cell injection or that a repaired meniscus has been demonstrated simply because pain improves.

For the condition-specific discussion, see Can stem cell therapy regenerate meniscus tissue?

What does clinical research show about bone marrow injections for knee arthritis?

Our published observations: encouraging symptoms, without proof of the mechanism

In a 2018 paper I coauthored, we reported short-term outcomes in 15 patients involving 20 knees through four bone marrow concentrate treatments. The study reported improvements in patient-reported pain and function. It was a small, uncontrolled study, so it could not determine how much improvement was attributable to the treatment or establish that native cells were activated, cartilage regrew, or four injections were better than another schedule. (9)

These observations matter as preliminary clinical information. They should be considered alongside controlled research rather than used as proof of tissue regeneration.

Placebo-controlled research shows why comparison groups matter

In a trial of 25 patients with arthritis in both knees, one knee received BMAC and the other saline. Both knees improved, but pain relief did not differ significantly between the treatments at the reported follow-up through six months. The study was small, but it demonstrates why improvement after treatment does not, by itself, establish a specific biological effect. (10)

The larger MILES trial randomized 480 people with knee osteoarthritis to cell-based treatments or corticosteroid injection. For the BMAC comparison relevant here, there was no superiority over corticosteroid at one year. The trial also did not demonstrate significant improvement in MRI osteoarthritis scores. These findings do not prove that no individual benefits, but they challenge claims of a reliably superior or cartilage-restoring treatment. (11)

2024–2026 evidence: possible symptom benefits, uncertain superiority

A 2024 systematic review included eight randomized studies involving 937 patients. BMAC groups often improved from baseline. However, differences from hyaluronic acid did not exceed the review’s threshold for a clinically important difference, and a clear advantage over other injections was not evident. (12)

An October 2026 review of 21 randomized trials across orthopedic conditions also described inconsistent superiority over comparison treatments. It identified signals of benefit in selected settings, including early knee osteoarthritis, while emphasizing variation between conditions and treatment strategies. Results from other joints or from surgical augmentation should not be treated as proof of an intra-articular knee injection’s effect. (13)

Does less pain mean the cartilage has regenerated?

No. Pain, function, and tissue structure are different outcomes. A person may walk more comfortably without a measurable increase in cartilage. Conversely, an imaging change needs to be interpreted in relation to symptoms and meaningful function.

To establish regeneration, research needs appropriate structural measures, comparison groups, and sufficient follow-up. Patient-reported improvement cannot tell us whether injected cells became cartilage, changed inflammatory signaling, or stimulated resident progenitors. These explanations require their own evidence.

For the broader treatment discussion, see stem cell and bone marrow research for knee osteoarthritis.

How does this research fit with clinical guidelines?

The 2019 American College of Rheumatology/Arthritis Foundation guideline, published in 2020, strongly recommended against stem-cell injections for knee and hip osteoarthritis because of concerns about heterogeneous, poorly standardized preparations and techniques. It strongly supported measures including exercise and, when appropriate, weight reduction. (14)

The May 2026 VA/DOD guideline reached a different recommendation: insufficient evidence to recommend for or against knee stem-cell injections. It also emphasizes that proposed orthobiologic effects commonly involve signaling and modulation of inflammation rather than demonstrated durable tissue regeneration. This is uncertainty, rather than an endorsement of cartilage-regrowth claims. (15)

My clinical takeaway is that a mechanism discussion should support careful evaluation, realistic expectations, and an overall treatment plan. It should not replace assessment of where the pain is coming from or established approaches to maintaining strength and function.

Wharton’s jelly: what is promising, and what remains unproven?

Wharton’s jelly is connective tissue within the umbilical cord. Tissue matrix preparations, isolated donor cells, culture-expanded cells, and vesicles derived from those cells are distinct materials. A study of one cannot validate the others.

A laboratory characterization study identified growth factors, cytokines, hyaluronic acid, and extracellular vesicles in a particular Wharton’s jelly formulation. These findings provide a biological rationale for investigation. Detecting such components does not establish that a processed product contains viable stem cells, delivers an effective dose, or repairs cartilage after injection. (16)

A published knee-osteoarthritis trial protocol proposed comparing Wharton’s jelly with hyaluronic acid and saline. A protocol describes a planned investigation; it is not a report of successful treatment outcomes. It should not be cited as proof that the treatment works. (17)

Exosomes: a promising signaling pathway, with mixed early clinical evidence

Exosomes are a type of small extracellular vesicle carrying biological material between cells. “Extracellular vesicle” is a broader term; not every preparation labeled as exosomes is necessarily equivalent or equally characterized.

In a 2025 rat study of surgically created cartilage defects, investigators compared Wharton’s jelly-derived cells, their exosomes, hyaluronic acid, and combinations. The exosome group showed favorable repair characteristics on tissue assessment. This supports continued investigation of cell-derived signals, but it was a short-term animal experiment involving a focal defect—not proof that an exosome injection treats established human knee osteoarthritis. (18)

A 2024 triple-blind trial tested placental-cell-derived extracellular vesicles against saline in patients with arthritis in both knees. Thirty-one patients were enrolled and 29 were included in the final analysis. There was no significant advantage in symptoms or MRI findings at the assessed follow-up through six months. No severe local reactions or systemic complications were reported, but that small study cannot establish safety across products or over the long term. (19)

These studies involved different preparations and settings. The animal findings do not cancel the negative clinical result; the clinical result also does not answer every future question about vesicle therapy.

The FDA states that there are no FDA-approved exosome products and that the regenerative products addressed in its consumer alert, including Wharton’s jelly products, are not approved for orthopedic conditions. The discussion here describes research, rather than a claim that any particular donor-cell or exosome product is offered by Dr. Darrow. (20)

What should patients take from the cell-communication research?

The knee contains cells capable of responding to injury. Researchers are investigating how inflammation, signaling, and the surrounding tissue environment influence that response. Biologic injections may affect some of these pathways, but the exact contribution in an individual patient remains uncertain.

I want patients to understand both the rationale and the limits. An MRI finding alone does not establish every source of knee pain, and a promising laboratory mechanism does not establish the best treatment. The clinical question remains whether a specific intervention improves the outcomes that matter to that patient.

Supporting a biological repair response and proving that a knee has regenerated are different claims. Our explanation of treatment should make that distinction clear.