There are several differences. But first, you have to differentiate between different experiments with different objectives in different species.

a) Mice are not man: Standard anti-mouse anti-CD3 antibody (hamster-anti-mouse 145-2C11) has intermediate affinity, plate-bound use is necessary to induce strong T cell proliferation. In contrast: standard anti-human CD3 (mouse-anti-human OKT3) has very strong affinity, soluble OKT-3 induce strong T cell receptor stimulation

b) What is the objective? Examples: kinome analysis needs strong but short T cell stimulation (use plate-bound anti-CD3), coculture with regulatory T cells needs suboptimal T cell stimulation (strong signal overcome suppressive activity)

c) Use and concentration of anti-CD3 stimulation is strictly dependent on the individual antibody (affinity, species, cross linking activity)

Simply put, soluble does not cross-link CD3 while immobiized does. Soluble anti-CD3 has been associated with the induction of apoptois since there is no second signal, vis-a-vis, CD28. You can cross-link soluble CD3 with an antibody to the host of the anti-CD3 antibody but that isn't really the best way to go. If you are trying to mimick "in vivo" you should really do the combination of soluble anti-CD3 + anti-CD28, or go with immobilized anti-CD3. You should probably compare the literature as to what the authors were looking for in a response. Hope this helps.

Immobilized anti-CD3 crosslinks the T cells to a surface (as physiological T-APC interaction the links the T cells to the APC) and thus provides a stronger signal.So it is true that coated CD3 induces a stronger response than soluble does.

Soluble anti-CD3 are present in the culture medium during T cell differentiation whereas immobilized anti-CD3 are coated on plate surface (dilution in PBS, and at least 2H at 37°C).

Rather than use a double antibody approach you can use biotinylated anti-CD3 followed by avidin for crosslinking. Use the biotinylated antibody and avidin at 4:1 (each avidin crosslinks 4 biotin molecules). Dr. Lapchak described the difference exactly.

There are several differences. But first, you have to differentiate between different experiments with different objectives in different species.

a) Mice are not man: Standard anti-mouse anti-CD3 antibody (hamster-anti-mouse 145-2C11) has intermediate affinity, plate-bound use is necessary to induce strong T cell proliferation. In contrast: standard anti-human CD3 (mouse-anti-human OKT3) has very strong affinity, soluble OKT-3 induce strong T cell receptor stimulation

b) What is the objective? Examples: kinome analysis needs strong but short T cell stimulation (use plate-bound anti-CD3), coculture with regulatory T cells needs suboptimal T cell stimulation (strong signal overcome suppressive activity)

c) Use and concentration of anti-CD3 stimulation is strictly dependent on the individual antibody (affinity, species, cross linking activity)

Well, it is true that immobilized anti-CD3 (on plate surface or on microbeads) srosslinks the CD3 and thus is a better stimulant than soluble antibody. However, it cannot be consider a replacement for addition of antiCD28, as I believe Peter suggested above. If you would not have the CD28 signal, your T cells would not respond optimally, especially if you purify them prior to culture (i.e. remove monocyte-macrophages from say initial PBMC suspension). Also, if you a re interested in the question of "maximal available response" from your T cells, you would use immobilized antiCD3 plus soluble antiCD28 even if stimulating the PBMC. Technical remark: unless you use the stimulatory antibodies you or your lab already used and titrated, it is prudent to to do a pilot with some combinations of concentrations of both Abs around that recommended (in our hands we had to dilute one of the antiCD3s for immobilization at least twice from the recommended, or we had mostly apoptosis),

And I agree with David above, that biotin/avidin xlinking would do, but still it would NOT replace the CD28 signal.

Hi Jacek, you are quite correct! One can use the same strategy to crosslink CD28 too. Its more 'efficient' than the plate bound method.

Hi David, Enrico - both of you are absolutely correct; however, as I pointed above, this approach (crosslinked or immobilized antibodies on plate of beads would yield the maximal available response from stimulated cells. Myself, if I want something closer to physiology, I stimulate the PBMC with immobilized antiCD3 only (accessory signals from monocyte/macrophages), and - for comparison and assessment of the maximal stimulation level, with both the immobilized antibodies. I judge the proliferation using CFSE staining and flow cytometry - can get information from any population that way without need to purify them.

All comments above are correct. The reason you achieve optimal stimulation is because by immobilizing anti-CD3 Abs on a surface you cross-link these antibodies around your cells. Thus, the stimulation is targeted rather than random. If you then add soluble anti-CD28 Abs you can stimulate your T cells optimally without making them anergic. I have also used both anti-CD3/28 bound on a surface with very good results in terms of T cell activation and proliferation.

Soluble antibodies can also be used in assays but it is likely that due to their chaotic behaviour inside the plate well, they might not achieve maximum stimulation.

Dear Theodore,

What means optimal stimulation? The objective of T cell stimulation is essential. Anti-CD3 plus anti-CD28 mAb stimulation is useful for genomic and proteomic analysis of T cell subsets but far away from a physiologic antigen-specific stimulation. In presence of an antigen-presenting mature dendritic cell only a few T cell receptors will be cross-linked but enough to induce optimal activation, proliferation and differentiation into T effector cells. Furthermore, affinity of mAb used for stimulation direct the result of T cell stimulation. OKT-3 has induced a fatal cytokine storm in patients without the need of any costimulation as well as the anti-CD28 superagonistic mAb of Tegenero. T cells integrate all signals. A strong TCR signal can bypass suboptimal costimulation.

Isolated human CD4+ T cells proliferate also after stimulation with soluble anti-CD3 mAb (OKT-3) since OKT-3 has a high affinity and cross link all CD3 molecules on the surface (you can easily control this fact by flow cytometry or fluorescence microscopy). Therefore, in most functional in vitro assays, soluble OKT-3 is used.

Additionally, high amounts of soluble or plate-bound OKT-3 induce activation-induced apoptosis in isolated human T cells in absence of exogenous IL-2.

In conclusion, objective of the experiment, individual mAb with specific affinity used for stimulation, length of T cell stimulation, and T cell subset used for stimulation (naïve versus differentiated, isolated or not…) direct the result.

By optimal stimulation, I am referring to the maximum stimulation indirectly observed by cytokine secretion and cell proliferation in the cell culture system I used. As you say, this may differ from experiment to experiment. But optimal stimulation as a concept can be achieved in every system after trial and error with many factors, e.g. Ab concentration, number od cells, addition of exogenous cytokines, etc.

Jiawu, from my humble experience, proliferation levels of mouse CD4+CD25- splenic T cells stimulated by Invitrogen CD3/CD28 beads are few times higher compared to the stimulation of the same cells with plate bound a-CD3 and soluble a-CD28. If you are interested in maximal cell proliferation, I'd stronly recommend you CD3/CD28 beads.

Indeed it has been demonstrated that anti-CD3 antibodies alone are insufficient to induce T cell proliferation. A second signal is needed, e.g. in the form of an antigen-presenting cell or a co-stimulatory antibody such as anti-CD28. However, when cross-linked, e.g. by binding to a solid surface such as a tissue culture well, an anti CD3 McAb can directly stimulate T cells to proliferation. Furthermore, it has been

found that CD3 McAbs of the IgE isotype are able to stimulate T cells directly, without the need for a second signal. At Sanquin (formerly CLB), a whole-blood lymphocyte culture system has been developed as a simple, reliable and fast method to test for T cell reactivity in patients, in which anti CD3 McAb of the IgE class is used as the stimulant. This IgE anti CD3 antibody was developed at the CLB as a switch variant originating from an IgG1 anti CD3 McAb.

Re Iurii's comment: right, it is even logical from the cytological/machnistical point of view. If we consider immobilized anti-CD3, a lymphocyte has a chance to contact the Ab molecules only with part of its membrane (bottom) and with soluble anti-CD28 throughout, but mostly using its free top part. While with beads its different - they may be on the bottom of the well and in suspension, they are smaller than the cell, so many may be attached (and migrating along) the membrane, xlinking and stimulating consecutive sets of receptors. Thus, using beads should lead to generation of many pseudo immune synapses on a sigle T cell, leading to optimal/maximal stimulation.

Stimulation of human T cells works fine with anti-CD3/CD28 beads, as mentioned by several people before (OKT3 affinity is indeed high enough and plastic mobilization is not necessary). More "physiological" (if I may say so) conditions can be achieved in "feeder mix" cultures, where T cells are stimulated in a mixture of irradiated PBMC and JY cells, in presence of IL-2 (alloreactive stimulation).

For technical details, see how we used this "feeder mix" approach here:

Functional human antigen-specific T cells produced in vitro using retroviral T cell receptor transfer into hematopoietic progenitors.

van Lent AU, Nagasawa M, van Loenen MM, Schotte R, Schumacher TN, Heemskerk MH, Spits H, Legrand N.

J Immunol. 2007 Oct 15;179(8):4959-68

Hello Jiawu Zhao

In the mouse system the difference is that soluble anti-CD3 (and Con A) needs an APC to stimulate T cells whereas plate-bound anti-CD3 does not. So if you are working with very pure T cell populations or with a T cell line and want to stimulate them polyclonally without adding other cells as APC to the system ,, use plate-bound-anti-CD3.

Ises A. Abrahamsohn

São Paulo, Brazil.

The answer actually gets more complicated still. Turns out to get a strong Th1 response you need to stimulate T cells with anti-CD40. If you want a more "regulatory' profile stimulate with anti-CD28. Interstingly CD3 induces CD40L on the T cells, so they self-stimulate through CD40.

Hi David! Fascinating as they are, te CD4loCD40+ T cells are only a subpopulation of the T's. Are you not generalizing too much? You really think there would be no Th1 response without CD40 ligation on T cells? And that this pertains to human immune system? Especially, considering that only your very good Clin Imm 2007 paper concerns human T cells, and other are based on studies in NOD mice. Very interesting anyway, but IMHO we cannot directly translate mouse model results to humans (e.g. the Tregs, which really do behave prety different in mice and humans. This is mostly (again IMHO) because human beings are under constant environmental antigenic challenge and our Treg response is showing all the time (as final part of normal immune reactivity to any Ag; of course averaged by the multitude of challenging Ags and very diverse timing of each individual reaction); while in mice under lab conditions practically any Treg response must be induced and stands alone. Of course I am disregarding here the nTregs which might be similar in mice and man; however again, they would show much more in the lab mice than in man, for abovementioned reasons...

So back to the original Jiawu's question - it all depends what model you use and what questions you ask, yet in human PBMC model stimulation with just immobilized antiCD3 gives submaximal response (as the costimulation is only via monocyte/macrophages present in) while addition of antiCD28 provides this accessory signal an maximalizes the response (regardless the effect of macrophages present); the latter would also give maximal stimulation if purified T or their subpopulations were used.

Hi Jacek! That actually is a really good point. I think that I am generalizing too much. You are correct about CD4CD40 being a subset; AND you are very correct about the extreme difficulty translating between human and mouse....both directions actually.

Hi everybody,

I was wondering what do you recommend for Jurkat cell activation.

I need to have Jurkat cells stimulated by only CD3mAb or (CD3+CD28mAbs).

Do I need to plate-bound the mAbs. If yes, do I need to do that just for CD3mAb or both? How about using the Dynabeads conjugated by one or both of the mAbs?

If I understood the discussion well, it seems that stimulation with soluble CD3mAb in the absence of CD28mAb may trigger the apoptotic pathways. Is that right?

And which mAb cat#s do you recommend for each?

soluble anti-CD3 of the IgG isotype will induce apoptosis. However, if it is of the IgM isotype, it won't due to more efficient cross-linking. Combining anti-CD3 and anti-CD28 provides optimal stimulation. The best way to know which Ab work is to read the papers which use these antibodies. You will notice a trend (company and type) that the majority of researchers use.

Actually I've checked most of the current papers and all of the anti-CD3s are IgG (OKT3, TR66, UCHT1 ....)

Most of the papers mention that they do the plate-bounding, cross-linking with goat anti mouse Ab or using antiCD3/antiCD28 conjugated Dynabeads. But like the thing that Helmut has mentioned in the first answer, some believe that soluble OKT-3 (IgG) stimulate strong T cell receptor stimulation!

But, what is my objective? I am looking for possible changes in some spicific protein-DNA interactions with and without activation of the Jurkat cells. And it seems to me, using the Dynabeads coated with anti-CD3/anti-CD28 could be one of the most reliable ways for activating the Jurkat cells for this experiment.

I also appreciate to know, why mono anti-CD3 could trigger the apoptotic pathway, while the cross-linked couldn't!

strength of signal: crosslinking brings ligands together in a specific area of the cell membrane leading to a strong point signal while little "tickles" don't. Partial activation in this case leads to apoptosis. There are volumes of papers describing the differences. By the way, Nima, could you send me an invitation on LinkedIn. www.linkedin.com/in/peterhlapchaklifescientist/

It seems that I need your Email address to be able to invite you in LinkedIn. (I'm not that much active in LinkedIn, but it would be of my pleasure to join you there)

I'd certainly first ask the question of what you hope to gain or learn from the activation. The strong induced signal is probably useful for some goals - to overexpress a protein that is induced by T cell activation - but I would use it with great caution to study a physiologic process as it is an extremely strong signal that nonetheless lacks a number of other modulating signals as compared to the natural process. One reason that plate-bound versus soluble crosslinked (or bead crosslinked) signals differently is that the cell can't internalize the receptors during the plate-bound experiments very well (they probably do, a little). The signal goes on longer.

Thanks Lisa, my goal is to induce TCR dependent transcription factors when jurkat cells are activated by antiCD3 alone or along with antiCD28. One of my questions is that if there is any DNA sequence specificity for couple of transcription factors upon TCR dependent Jurkat activation. It seems that people like Dynabeads for optimal activation of the T cells , but as you thoughtfully mentioned, I need to know if it is the appropriate way according to my goal! I would appreciate your opinion.

I know that Jurkat have been acceptable in transcription studies and certainly bead stimulation would work well to stimulate them. Try some limiting dilution of beads and see if you get similar results at high and low signalling densities. Consulting and comparing within the transcription field (among your future reviewers) is probably the best way to check so you don't run into trouble when you try to publish. Here is an interesting review from a few years back focused on these cells: http://www.nature.com/nri/journal/v4/n4/full/nri1330.html I personally believe that results are best corroborated in a non-leukemic T cell line under more physiologic conditions (using APC and antigen). Best of luck!

Nice review! Thanks very much Lisa.

We use coated CD3/CD28 for most human / mouse experiments. Miltenyi CD2/CD3/CD28 beads are fantastic for growth curves, CD4 and CD8 cells proliferate quite strongly over 4 weeks. However, even when trying to cut off beads at the end of the experiments, we were not able to perform any PCR-based assay or Western Blot or FACS with the cells, assuming that the beads interfere massively. They are actually much bigger than the isolation beads.

Thanks Ioakim. Do you mean that you failed to detach the cells from beads?

Are these beads different from Invitrogen's Dynabeads?

If the cells were purified T cells, you have to used coated CD3, anti-CD28 could be either soluble or coated, it does not matter;

If the cells were total splenocytes, which contain the cells expressing FcR, such as B cell and macrophages, then soluble anti-CD3 works well. Because, the cells with FcR will cross-link the 2C11.

When you use anti CD3 plate bound or soluble anti CD3 and cross link the antibody dependent on if the T cells are really naive T cells or not really naive T cells you can get a T cell response without costimulation signasl of either CD28 or CD40/80/70.

if you take soluble anti CD3 and do not cross link the antibody you do not get much T cell stimulation.

It is difficult to get real naive T cells from mice that are going in normal animal houses in cages. Often part of these T cells are pre acitvated, which give you a very high background in your t cell responses with anti CD3.

The immobilized anti CD3 would be imitating the recognition of the antigen by the T Cell on the surface of the APC. So, I think is more "realistic" than the soluble anti CD3.

Helmut (Jonuleit) is pretty much the correct expert to answer this question and what he says is correct.

Anthony Asmar is right. immobilized anti cd3 is a better choice for stimulation lalitha kabilan

So how long do you stimulate your cells with the CD3/CD28 beads to study TCR signaling ?

I too have a question, slightly related to the feed here: I'm trying to develop an assay using PBMC, stimulated by a cytokine, etc...to produce either TNFalpha or IL-1/6 response in a renal fibrosis model. Maybe ELISA readout or FACS...the data output is the least of my concerns, as i'm having trouble finding a good cytokine stimulation protocol.  I'm looking for a CYTOKINE, NOT LPS, NOT PMA, NOT IONOMYCIN, etc....thanks!

Usually TNF induces TNF and IL-1

I think you can achieve some level of stimulation using anti-cd28 and anti-cd3 depending on your experimental plan. The co-stimulation triggered by anti-cd28 will prevent T cell apoptosis regardless of how strong the activation from anti-cd3 is.

Thank you for invaluable comments, I have a question about T cell stimulation by soluble anti- CD3 and CD28 . I am using these antibodies from ebbioscence company. What is the optimal concentration  for 1 million PMBC per ml?