NK cells, B cells and T cells: immune cells 101

Lymphocytes are the white blood cells that give the immune system its precision and its memory. Human blood carries three main kinds — T cells, B cells and natural killer (NK) cells — and they share one lineage, one ancestor, and under an ordinary microscope very nearly one appearance. Small, round, unremarkable. What separates them is invisible: what each one can recognise, and whether it had to learn how.

This page is a census of the three — what each cell is, how immunologists tell them apart, what each recognises, where each keeps its memory, and which of them the published research on Transfer Factor Max has and has not measured. Related pages cover what a transfer factor is, the two arms of immunity, and neutrophils. This one is about the cells.

One clarification first, because it decides how you read every study in this field: immune cell and lymphocyte are not synonyms. Neutrophils, monocytes, macrophages and dendritic cells are all immune cells and none is a lymphocyte. These three are a specific family — and not the most numerous cells in your blood.

One family, three jobs

Every blood cell descends from a hematopoietic stem cell in the bone marrow — the population described in our piece on immune stem cells. Early on, that lineage forks. The myeloid branch produces neutrophils, monocytes, macrophages and platelets. The lymphoid branch produces exactly three things of interest here: T cells, B cells and NK cells.

That fork is not a technicality. It is why a study of neutrophil behaviour says nothing directly about T cells, and why a measurement of stem cell mobilisation is a statement about the factory rather than about any product coming off it.

Where each lymphocyte finishes its training is written into its name — in one case misleadingly. The T is the thymus, the small organ behind the breastbone where T cells are educated and mostly eliminated. The B is a historical accident: it comes from the bursa of Fabricius, an organ birds have and humans do not, where the lineage was first identified. In humans the equivalent development happens in bone marrow — a tidy coincidence of initials, not the origin of the letter. As with transfer factor from hen egg yolk, avian immunology taught human immunology something it could not see in itself.

How immunologists actually tell them apart

A T cell and a B cell look alike, and for most of the twentieth century that was a real obstacle: you cannot sort cells you cannot distinguish. The solution — and the reason modern lab reports read like algebra — was to give up on appearance and name cells by the molecules on their surface.

The system is called CD, for cluster of differentiation, and its origin is unusually literal. At the first Human Leucocyte Differentiation Antigens workshop in 1982, laboratories compared antibodies they had each raised against white blood cells. Whenever antibodies from different labs turned out to be binding the same surface molecule, that group was declared a cluster and given a number — and the molecule inherited it. CD4 and CD8 are simply the fourth and eighth molecules catalogued that way; the numbers carry no ranking.

What that gives you is a definition instead of a picture. A T cell displays CD3. A B cell displays CD19 or CD20. An NK cell displays CD56 and, crucially, not CD3 — identified partly by an absence. Flow cytometry reads those markers thousands of cells per second, which is how a tube of blood becomes a list of populations.

T cells: the ones that go to school

A T cell's receptor is assembled by cutting and rejoining gene segments semi-randomly — recent estimates put the human naive repertoire at at least a hundred million distinct clonotypes. Randomness has an obvious hazard: some of those receptors will fit the body that made them. The thymus solves that by attrition.

Developing T cells face two tests. Positive selection asks whether the receptor can read MHC molecules at all; one that binds nothing is useless, and the cell dies by neglect. Negative selection asks the opposite — whether it reacts strongly to the body's own material — and eliminates the cells that do. Roughly five percent of developing thymocytes are commonly said to graduate, an accounting drawn largely from mouse work. The direction is not in doubt: the thymus destroys far more lymphocytes than it releases.

The cells that graduate sort into three jobs.

CD4-positive helper T cells

Helpers kill nothing. They read a fragment displayed on another cell and issue instructions: licensing a B cell to improve and switch its antibody, activating macrophages, summoning others with cytokines. Remove the coordinator and work that other cells remain physically capable of doing stops happening — which is why a CD4 count is such a sensitive indicator of immune competence.

CD8-positive cytotoxic T cells

Killers inspect. Nearly every cell continuously displays fragments of its own contents on MHC class I molecules — a running declaration of what is going on inside. A CD8 T cell reads those declarations, and when it finds a fragment matching its receptor, it instructs the cell to die.

Regulatory T cells

Regulatory T cells — commonly reported as five to ten percent of CD4 T cells, identified largely by the transcription factor FOXP3 — exist to hold responses back. Their importance shows in their absence: mutations disabling FOXP3 cause IPEX syndrome, a severe inflammatory disease in which the problem is too little restraint, not too little activity. That is the biological core of our argument about why “boost” is the wrong verb: the immune system keeps a cell type whose whole job is to say no.

One historical note ties the T cell to the origin of this subject. The 1949 experiment that gave transfer factor its name transferred delayed-type hypersensitivity — a tuberculin skin reaction — between people using an extract of white blood cells. Delayed-type hypersensitivity is by definition T-cell driven, not antibody driven. The founding observation of the field was a T cell observation, developed further in the two arms.

B cells: the only cells that make antibody

A B cell carries a receptor built by the same gene-shuffling machinery — but its receptor is a membrane-anchored antibody, and that changes everything. A T cell can only read fragments already displayed by another cell. A B cell binds the intact, three-dimensional surface of an intruder directly, which is why antibodies can recognise a bacterial coat or a viral spike as an object rather than as a text.

Given its target and permission from a helper T cell, a B cell enters a germinal centre and does something no other cell does: it deliberately mutates the gene for its own receptor, and the versions that bind better are selected to divide. The same process can switch the antibody's class — IgM to IgG, IgA or IgE — changing where it goes and what it recruits without changing what it recognises.

The output takes two forms. Some activated B cells become plasma cells, abandoning most other functions to secrete antibody at industrial rates; figures in the thousands of molecules per second are commonly cited. Others become memory B cells and wait. Readers of our page on colostrum have met this machinery's product already: the antibodies in first milk are plasma-cell output, which is precisely why colostrum antibodies and transfer factor are different things arriving in the same fluid.

NK cells: the lymphocyte that never went to school

Natural killer cells are the family's interesting exception — lymphoid by ancestry, innate by behaviour. They never rearrange genes to build an antigen-specific receptor and never pass through thymic education. They are ready immediately, and ready for a category rather than for a target.

How they choose is the elegant part. Healthy cells display MHC class I molecules, and NK cells carry receptors — the KIR family and the NKG2 group among them — that read that display as identification. The missing-self idea, proposed by Kärre and colleagues in the 1980s, is that a cell which has stopped displaying class I becomes a target. Not for what it shows. For what it has stopped showing.

This produces one of immunology's genuinely satisfying results. Viruses and some tumour cells suppress MHC class I display specifically to avoid inspection by CD8 T cells — and in doing so make themselves conspicuous to NK cells. The trick that hides a cell from one lymphocyte exposes it to another.

NK cells also read antibody. Their CD16 receptor binds antibody already attached to a target and kills what it finds there — antibody-dependent cellular cytotoxicity — which means B cell output can aim NK cell killing. Two broad NK subsets are distinguished in the laboratory by how much CD56 and CD16 they carry, and whether NK cells possess a memory of their own is an active research question rather than a settled fact.

The census, on one page

On proportions: published reference ranges for healthy adults put CD3-positive T cells at the clear majority of blood lymphocytes — one dual-platform flow cytometry study reported roughly 65 to 88 percent — with B cells and NK cells each in a much smaller and wider band. A separate cohort reported different bands again: reference ranges are properties of a population and a method, not universal constants.

Where immune memory actually lives

The most striking measurements in this field are the durations, and the cleanest natural experiment is smallpox. Routine vaccination was discontinued during the 1970s and the disease declared eradicated in 1980 — so any smallpox-specific immunity found today has had no opportunity for a booster from the world. Memory with no maintenance.

The findings are remarkable. Anti-vaccinia memory B cells have been followed in human blood up to roughly 65 years after vaccination. Smallpox-specific antibody has been measured in serum more than 70 years afterwards. And plasma cells secreting vaccinia-specific antibody have been recovered from human bone marrow more than 35 years after the disease itself ceased to exist — cells still manufacturing a defence against something that is gone.

That is what remember means when an immunologist uses it: not a stored impression but a maintained population of cells. It is also, to be explicit, a description of vaccination. No supplement is part of that record, and none is offered here as a route to it.

Which of these has the Transfer Factor Max research measured?

Almost none of them, and the cleanest thing this page can do is say so.

The three are examined in our piece on the three clinical studies. Mapped onto the census above: Yu 2024 was a feasibility trial examining immune-related biomarkers and cytokine changes — laboratory measurements, not subset counts. Jensen 2026 concerned neutrophil activation, and neutrophils are myeloid: the bottom row of the table, not any row above it. Gardner 2026 concerned stem cell mobilisation, upstream of all three lineages — the factory rather than its products.

So: no published study on the finished product reports T, B or NK cell subset counts. None reports antibody titres, a gap already noted on this site. None reports illness frequency. Two of the three are 4Life Research clinical reports rather than peer-reviewed papers. The three cells this article is about are, with one partial exception upstream, not the cells that research looked at.*

What this article does not claim

• This is immunology, not a mechanism for a supplement. Nothing above describes what any product does inside a body.
• No effect on any of these populations is claimed. No supplement discussed on this site is shown to raise, lower or rearrange T cell, B cell or NK cell counts.*
• The smallpox findings describe vaccination — cited to show what durable immune memory looks like when measured, and for no other reason.
• The census is a simplification. Gamma delta T cells, NKT cells, innate lymphoid cells, tissue-resident memory populations and the finer B cell subsets are all real and all omitted.
• Reference ranges are not diagnostic tools. Subset counts shift with infection, medication, age, stress, sleep and time of day. A value outside a published band is a conversation with a clinician, not a verdict.

One brand note, offered as a map rather than a claim. Recognize is the T cell receptor, the B cell receptor and the NK cell's reading of MHC class I. Respond is killing, secreting and signalling. Remember is memory T cells, memory B cells and long-lived plasma cells. Renew is the marrow, upstream of all of them. Four verbs describing compartments of a healthy immune system — not an effect.*

Frequently asked questions

Citations & sources

Keep reading

*These statements have not been evaluated by the Food and Drug Administration. Transfer Factor Max and the other products mentioned are dietary supplements and are not intended to diagnose, treat, cure, or prevent any disease. Individual results vary. This article is general immunology education compiled from published scientific sources; it is not medical advice, it is not a description of what any product does in the body, and nothing in it should be taken as a claim that any product affects any disease, any immune cell population, or any laboratory value. Do not use this page to interpret a blood test. Follow label directions; do not exceed the recommended daily amount. Consult your healthcare provider before beginning any new supplement program, particularly if you take medication, have an autoimmune condition, have an egg or dairy allergy, or are pregnant or nursing.