Natural killer cells kill tumour cells — that part is not in doubt. Whether infusing expanded copies of your own NK cells does anything for you is a different question, and the trial that tested it answers it in its title.
What NK Cells Are
NK cells are lymphocytes of the innate immune system. Unlike T cells, they kill target cells without prior sensitisation — no antigen presentation, no priming period [1].
How NK cells choose targets: the "missing self" model
Healthy cells display MHC class I molecules. NK cells read these through inhibitory receptors as a signal not to attack. Many cancerous and virus-infected cells lose or reduce MHC class I expression — precisely to evade T cells — and simultaneously express stress ligands. NK cells detect that imbalance [2].
When activating signals outweigh inhibitory signals, the NK cell kills.
⚠️ Hold on to this mechanism. It is elegant, it is correct, and it is the reason the central claim made for autologous NK therapy does not work. We return to it below.
How NK cells kill
Four mechanisms [3]:
- Perforin/granzyme — perforin creates pores in the target membrane; granzymes enter and trigger apoptosis
- Death receptors — FasL and TRAIL on the NK surface bind death receptors on the target
- Antibody-dependent cellular cytotoxicity (ADCC) — CD16 recognises antibody-coated targets. This is how several antibody cancer drugs achieve part of their effect
- Cytokine production — IFN-γ and TNF-α, recruiting and activating other immune cells
Two NK subsets
Not all NK cells do the same job [4]:
⚠️ Location is not incidental. Most of what is known about human NK cells comes from blood, but NK cells residing in tissues — liver, lung, uterus, gut, lymphoid organs — are phenotypically and functionally distinct from circulating ones, with their own residency programmes and receptor repertoires [13]. An intravenous infusion delivers cells to the circulation. Whether, and in what state, they reach the tissue compartments where tumour surveillance actually happens is an open question, and one of the reasons in-vitro killing assays translate poorly to patient outcomes.
What NKT Cells Are
NKT cells are a distinct lineage sharing features of NK and T cells [5]:
- Express a T cell receptor, but recognise lipid antigens presented by CD1d rather than peptides on MHC
- Express NK markers (CD56, CD161)
- Respond within hours, like innate cells
Type I (invariant) NKT cells carry an invariant TCR (Vα24-Jα18 in humans) and recognise α-galactosylceramide; they are the best characterised. Type II NKT cells have a more diverse repertoire and are less well understood.
On activation, NKT cells rapidly release large amounts of cytokines — IFN-γ, IL-4, IL-17 and IL-10⁶. ⚠️ Note what that list contains: both pro-inflammatory and anti-inflammatory signals. NKT cells can amplify or suppress immune responses depending on context. They are regulators, not simply activators, and any description of them as purely anti-tumour is incomplete.
Immunosurveillance, and Its Limits
The immune system does patrol for transformed cells. The evidence is solid [7][8]: cancer incidence rises in immunosuppressed individuals, tumours evolve specific mechanisms to evade immune detection, immune infiltration correlates with prognosis, and immunotherapy can produce durable remissions.
The three phases of cancer immunoediting [8]describe why this is not the whole story:
- Elimination — nascent transformed cells are destroyed
- Equilibrium — contained but not eliminated, sometimes for years
- Escape — the tumour evolves past immune control
Escape mechanisms include losing MHC class I, secreting TGF-β and IL-10, expressing checkpoint ligands such as PD-L1, and building an immunosuppressive microenvironment.
NK and NKT function does decline with age and disease — reduced cytotoxicity, altered receptor expression, fewer NKT cells [9][10].
⚠️ The step that does not follow. From "immune function declines" it does not follow that raising a laboratory measure of immune function reduces disease. That inference is the entire commercial proposition of immune-optimisation therapy, and it has not been demonstrated.
What the Trials Actually Tested
This is where precision matters most, because the trials and the products sold to well people differ on nearly every variable.
The trial record
Three things are true of every successful trial above: the cells came from a donor, the patient received lymphodepleting chemotherapy first, and IL-2 or equivalent cytokine support was given afterwards. None of the three is part of a wellness infusion.
Even after two decades, the field's own summary is cautious. Rubnitz's 2015 conclusion, ten years after NKAML: "NK cell therapy is safe, feasible, and should be further investigated." [19]
⛔ The autologous problem
The one trial testing autologous expanded NK cells is titled:
"Adoptive transfer of autologous natural killer cells leads to high levels of circulating natural killer cells but does not mediate tumour regression." [17]
Eight patients received an average of 4.7 × 10¹ [0]activated autologous NK cells after lymphodepleting chemotherapy. The infused cells "exhibited high levels of lytic activity in vitro." They persisted in the circulation for at least a week and in some patients for months.
"Although no clinical responses were observed…"
And the trial found why: the persisting NK cells downregulated the key activating receptor NKG2D and could not lyse tumour targets unless reactivated with IL-2. The cells were in the patient, in large numbers, and functionally switched off.
⚠️ The mechanism was already known, and it is the "missing self" model from earlier in this article. The landmark 2005 trial opens by stating it [15]:
"We previously demonstrated that autologous NK-cell therapy after hematopoietic cell transplantation is safe but does not provide an antitumour effect. We hypothesize that this is due to a lack of NK-cell inhibitory receptor mismatching with autologous tumour cells, which may be overcome by allogeneic NK-cell infusions."
Your NK cells and your tumour cells share your MHC class I. The inhibitory KIR–ligand match is intact, so the inhibitory signal wins and the cell does not kill — no matter how many of them there are or how activated they looked in the flask. This is why the entire field moved to haploidentical, KIR-mismatched donors.
The 2021 Nature Reviews Clinical Oncology survey of the field lists the NK cell sources under clinical investigation: haploidentical NK cells, umbilical cord blood NK cells, stem cell-derived NK cells, NK cell lines, adaptive NK cells, cytokine-induced memory-like NK cells, and CAR-NK cells [12]. Autologous expanded NK cells are not among them.
"Immune Optimisation": What It Can and Cannot Mean
Laboratory expansion works. Starting from a blood draw, NK cells are isolated by density gradient, magnetic bead selection or flow sorting, then cultured with IL-2, IL-15 and IL-21, sometimes with irradiated K562 feeder cells, achieving thousand-fold or greater expansion over 14–21 days [11]. Activated cells reliably show increased cytotoxic granule content, enhanced cytokine production and upregulated activating receptors.
⚠️ All of that is a biomarker, and Parkhurst 2011 is the demonstration of why that matters. Those cells had "high levels of lytic activity in vitro", persisted for months, and produced no clinical responses. A cell that kills well in a flask and a treatment that helps a patient are different claims, and only the second one is an outcome.
On "boosting immunity" generally, an analysis of 185 webpages returned for that search term found that "the only evidence-based approach to this is vaccination", and that commercial websites are a major component of the information the public receives — and contribute biased information [21]. We are a commercial website, and that finding applies to us.
⚠️ It is worth knowing the scale of the problem in this sector. An analysis of 896 practice websites advertising stem cell therapy direct to consumers found 95.9% contained at least one statement of misinformation, averaging 4.65 statements per site, across false general claims, inaccurate mechanisms of action, unfounded results, and scare tactics [22].
The longevity rationale, stated as its authors state it
There is a serious research idea here. NK cells help clear senescent cells and the inflammatory factors they secrete; immunosenescence impairs that clearance, allowing senescent cells to accumulate. Restoring NK function is therefore of genuine interest in ageing biology [20].
The authors' own framing: this work "provides the proof-of-concept of adoptive NK cell therapy for senotherapy and holds promise as an emerging revolution in longevity therapeutics."
Proof-of-concept is not evidence of benefit. It means the idea is worth testing. It is not a reason to buy an infusion.
NK/NKT Therapy and MSC Therapy Are Different Things
They are frequently sold together. They work in opposite directions, and it matters.
⚠️ A claim commonly made for this combination does not hold up: that MSC therapy "improves immune function" and therefore combines naturally with NK therapy. The trial usually cited for this is a frailty study whose immunological findings were a reduction in serum TNF-α and in B-cell intracellular TNF-α — inflammatory markers going down. That is anti-inflammatory activity, entirely consistent with what MSCs do, and it is not evidence of immune enhancement.
Nor is there a coherent rationale for giving an immunosuppressive therapy to "prepare" for an immune-activating one. If you are offered both, they should be justified separately, on their own evidence, for your own reasons.
What We Offer, and What the Evidence Is
We offer autologous expanded NK/NKT cell preparation and infusion. Here is the honest position on it.
What is well established: the laboratory process works — cells can be isolated, expanded thousand-fold, activated, and quality-controlled for count, viability, phenotype, sterility, endotoxin and mycoplasma before return. The infusion has a good safety record. Autologous cells carry no graft-versus-host or rejection risk, and reported effects are generally mild — fever, fatigue, flu-like symptoms, occasional infusion reactions [14]. Cytokine release syndrome is a consideration with high-dose cytokine protocols and is usually manageable; neurotoxicity is rare and largely associated with CAR-T rather than NK cells.
What is not established, and this is the part that matters:
- ⛔ No trial has shown that autologous expanded NK cell infusion produces clinical benefit. The one trial designed to test it found none [17]
- No trial has tested NK infusion in healthy people for cancer prevention, immune optimisation, or longevity
- No trial has tested it for recurrence prevention in cancer survivors
- The successful oncology trials used donor cells, lymphodepletion and IL-2 — a different intervention in a different population [15][16][18][19]
If you are considering this for cancer, the appropriate route is a registered clinical trial. There are over 300 worldwide, and they are testing the cell sources that have actually shown activity. Discuss it with your oncologist.
If you are considering this as a well person, you should know you would be buying a biomarker change with no demonstrated clinical benefit, at a price, and that we would rather tell you that than not.
Regulatory Status
- No NK cell product has FDA approval for any cancer indication as of 2026
- Expanded and activated NK cells are substantially manipulated, which places them under drug and advanced-therapy frameworks in the US (FDA), EU (EMA) and Australia (TGA). The ISSCR guidelines are explicit that such products require formal clinical evaluation rather than direct-to-consumer provision
- ⚠️ Availability in a permissive jurisdiction is not regulatory approval, and the two should never be described interchangeably
- Multiple products are in late-stage trials; CAR-NK cells are an active and promising area, still investigational
Frequently Asked Questions
Does NK cell therapy work?
For donor-derived NK cells in specific blood cancers, with conditioning chemotherapy: there is real evidence of activity, including complete remissions [15][18][19]— in trials, not as approved therapy. For autologous expanded NK cells, the kind sold to individuals: the one trial found no clinical responses [17].
Then why is autologous therapy offered at all?
Because the laboratory process is real, the cells are genuinely expanded and activated, and the infusion is safe. Those facts are about the process, not about the outcome. We think you should know the difference before deciding.
Is this the same as CAR-T?
No. CAR-T involves genetically engineering T cells with synthetic receptors — a more complex process with significant toxicity risks including cytokine release syndrome and neurotoxicity. Expanded NK/NKT therapy uses unmodified cells. It is generally better tolerated. It is also far less proven.
Will my body reject the cells?
No. Autologous cells are your own. ⚠️ But the same "self" recognition that prevents rejection is what prevents them from attacking your tumour cells — see the autologous problem above. The property that makes them safe is the property that limits them.
Can NK/NKT therapy cure cancer?
We cannot claim cures. NK cell therapy has shown anti-tumour activity in clinical trials, principally in blood cancers, principally with donor cells. For solid tumours the record is much weaker.
Can healthy people benefit?
There is no clinical evidence that they do. The idea has a coherent rationale — NK cells clear senescent cells, and that clearance declines with age [20]— but its own authors describe it as proof-of-concept. No trial has tested NK infusion in healthy people for any outcome that matters to them.
Should I have MSC therapy first, to prepare?
There is no evidence-based reason to sequence them that way. MSCs are immunosuppressive; NK cells are immune-activating. If you want both, each should be justified on its own.
The Bottom Line
NK and NKT cell biology is one of the more elegant parts of immunology, and the oncology trial programme built on it is real, serious and producing genuine results — complete remissions in patients with poor-prognosis leukaemia, achieved with donor cells, conditioning chemotherapy and cytokine support.
The product sold to individuals shares the vocabulary of that programme and almost none of its variables. It is autologous where the trials are allogeneic, unconditioned where the trials use lymphodepletion, and given to well people where the trials enrol patients with active disease. The single trial that tested the autologous version reported its result in its title.
None of that makes the biology less interesting. It makes the biology the reason for scepticism rather than the reason for confidence — because "missing self" recognition, correctly understood, predicts exactly what Parkhurst found.