Short answer: Conventional wisdom characterized Th1 as T cells that secrete mainly IFN-g, TNF-a, LT-b and Th2 as those that secrete mainly IL-4, IL-5, IL-6, IL-9, IL-10, IL-13 (1). However, it’s becoming increasingly clear that the Th1-Th2 paradigm is perhaps outmoded, that CD4 helper T cells remain highly plastic in their effector functions, and that, following activation, they evolve particular sets of overlapping functionalities with distinct underlying transcriptional programs, and that they retain capacity for reprogramming.
In order to to fully understand how we ended up with our current understanding of CD4 helper T cell biology, we need to start from the beginning. How did the Th1-Th2 paradigm start? Why did it catch on? How did it evolve and how do we use this paradigm today? Finally, is this paradigm still useful or do its flaws make it more an impediment?
How did the Th1-Th2 paradigm start?
This paradigm starts with Mosmann et al’s 1986 Journal of Immunology paper (2). First, they created a large number of CD4 T cell lines in vitro by repeatedly stimulating them with antigens such as FGG (Fowl Gamma Globulin) and KLH (Keyhole Limpet Hemocyanin), and then cloned them. With these CD4 T cell clones in hand and using a variety of bioassays, Mosmann et al showed two things.
- Dichotomy in the cytokines secreted by these CD4 T cell clones.Those that secreted IL-2 did not secrete IL-4 and vice-versa. They called the former Th1 and the latter Th2.
- Apparently stable imprinting of the mutually exclusive cytokine secretion pattern. They stimulated the CD4 T cell clones with a polyclonal stimulus called Concanavalin A, and found similar dichotomy in the pattern of cytokines secreted, i.e. even when these T cells were stimulated very strongly with something else, they retained the same pattern of cytokine secretion.
At this time, ability to detect cytokines was very limited because there were hardly any cytokine-specific monoclonal antibody (Mab) pairs for ELISA (Enzyme Linked ImmunoSorbent Assay). Subsequent expansion of such tools led to improved characterization of CD4 helper T cell subsets.
So far, so good. However, the question remained as to the biological relevance of this apparent in vitro dichotomy in CD4 helper T cell biology.
Why did the Th1-Th2 paradigm catch on?
The Th1-Th2 paradigm was originally an in vitro observation. It caught on when Heinzel et all showed an apparently similar dichotomy in vivo in their 1989 Journal of Experimental Medicine paper (3). They examined CD4 T cells using the experimental mouse model of leishmaniasis (a protozoan parasite with a complex life cycle spanning sand flies and mammals). Since the 1970s, it was known that a particular inbred mouse strain called BALB/c is exquisitely susceptible to Leishmania major infection while other inbred strains such as C57Bl/6 are relatively resistant. Heinzel et al observed T cells from infected susceptible BALB/c mice secreted abundant IL-4 and little IFN-g, a pattern opposite to those from infected resistant C57Bl/6. With this in vivo mouse model observation, here was proof positive about the existence of Th1 and Th2 subsets!
From reference 4.
The early promise of this so-called in vivo proof of the Th1-Th2 paradigm never quite panned out. For example, one of the early predictions from this paradigm was that the IL-4 gene knock out would convert the susceptible BALB/c mouse to resistance. Two groups independently created BALB/c IL-4 knock-outs and duly challenged them with Leishmania major. Utter confusion ensued. One IL-4 knock out BALb/c mouse strain apparently remained susceptible (5) while another apparently became resistant (6). What gave? All sorts of things including different gene knock out strategies, different parasite strains, different parasite doses used for infection, and on and on. In short, biology isn’t easily boxed into neat dichotomies. As it turns out, far from being a puzzle so easily cracked, the Leishmania major susceptibility of the BALB/c mouse strain remains an active area of research till date.
How did the Th1-Th2 paradigm evolve and how is it used today?
The primary influence of the Th1-Th2 paradigm was in imposing a particular mental construct with respect to our understanding of CD4 helper T cell biology. For example, in the mid-1990s, while examining gut-associated CD4 helper T cells, Howard Weiner observed some of them secreted neither IFN-g nor IL-4 but rather abundant TGF-b1. Naturally, he assigned the label Th3 to such cells (7). Few years later, other immunologists identified CD4 T cells that primarily secreted IL-17A. No prizes for guessing the name we bestowed upon them (8). Th17 of course!
At this point it’s worth our while to take stock of the technological advances that helped evolve and solidify the paradigm of Th1, Th2 and other CD4 T helper cell subsets.
- In 1993, the long-sought CD4 T cell co-stimulator molecule was identified as CD28 (9).
- Starting in the 1990s and continuing till date, Mab pairs against all kinds of protein molecules started to become easily available.
- Similarly, biotech companies such as RandD, Invivogen, Invitrogen, etc started to synthesize and make available to the research community a wide variety of recombinant cytokine molecules.
- Capitalizing on the increasing availability of Mabs against a plethora of protein molecules, biotech companies such as , , and more recently developed a variety of immunomagnetic cell isolation procedures that enabled easy enrichment of CD4 T cells from secondary lymphoid organs and blood.
- If purity of cells enriched by immunomagnetic separation were insufficient, there was the alternative Flow cytometry Sorting approach using the Flow cytometer and Mabs which yielded highly pure cell subsets.
- 1, 2, 3, 4 and 5 helped launch a plethora of imaginative in vitro CD4 helper T cell culture techniques. How? Seemingly simple. Isolate T cells using methods 4 or 5, and stimulate them in vitro with anti-CD3 and anti-CD28 Mabs plus the magic sauce to drive them to differentiate into either Th1 or Th2. T cells require two signals, Signal 1 (targeting the T cell receptor) and Signal 2 (co-stimulator), for activation. In this in vitro approach, a Mab against the CD3 molecule serves as a surrogate for Signal 1, while a Mab against the CD28 molecule serves as a surrogate for Signal 2. The magic sauce? Over the years, by adding a variety of recombinant cytokines and Mabs, labs around the world evolved a variety of complex protocols to stably generate in vitro Th1 and Th2 cells. For example, stimulate with anti-CD3 and anti-CD28 Mabs plus anti-IL-4 Mab (to inhibit differentiation to Th2) and recombinant IL-12p70 (or IFN-g) to promote Th1 differentiation, to give one example.
- In concert with cell culture techniques described in 6, molecular biology approaches evolved apace. Such were applied to these so-called Th1 and Th2 cells to explicate their molecular signatures. Thus, the transcription factor T-bet was found primarily associated with capacity of T cells to secrete IFN-g (10) while another transcription factor GATA-3 was found primarily associated with capacity of T cells to secrete IL-4 (11).
- Mouse gene knock-out studies followed, for both cytokines and their purported molecular transcription factor drivers. Such confirmed the association of specific transcription factors with specific CD4 T cell cytokines. In addition to IFN-g with T-bet and IL-4 with GATA-3, today we have additional associations such as IL-17 with ROR-gt and T regulatory cells (Treg) with FoxP3.
In sum, the Th1-Th2 paradigm helped establish a seemingly impregnable mental construct whereby we envisaged CD4 helper T cell function to be non-overlapping or mutually exclusive or even antagonistic secretion of particular groups of cytokines, with different effector functions. This is why today in addition to Th1 and Th2, we have Th3, Th9, Th17, Treg, and so on.
Is the Th1-Th2 paradigm still useful or do its flaws make it more an impediment?
Almost from the beginning, there was pushback to the simplistic subsetting of CD4 helper T cell function along the lines of Th1, Th2 et al.
- Already in the 1990s, Anne Kelso, an Australian immunologist, developed some exquisite in vitro micromanipulation approaches (12, 13) to ask a really interesting question about CD4 helper T cells. Did daughter CD4 helper T cells display the same cytokine expression pattern as the parent? She cultured single CD4 helper T cells in Terasaki wells, stimulated them to proliferate, split the daughter cells into separate wells by micromanipulation, and compared parent and daughter cell cytokine expression patterns. Her results? Contrary to the established Th1-Th2 paradigm, she found daughter CD4 T cell cytokine patterns were not imprinted but rather stochastic. In other words, under the culture conditions she used, a parent CD4 T cell predominantly expressing IFN-g was no guarantee that its daughter T cells would behave the same. However, so powerful had the Th1-Th2 paradigm become that she switched to studying CD8 T cells instead. Ah, such is the impartiality and objectivity that undergirds our modern scientific enterprise!
- Almost from the beginning, the neat dichotomies of mouse CD4 T cell subsets refused to pan out so cleanly for human CD4 helper T cells. As a result, human immunologists accepted the Th1-Th2 paradigm either grudgingly or skeptically or not at all.
- Ah, those deceptively powerful in vitro cell culture techniques that helped define all these CD4 T cell subsets. Let’s examine some of their flaws.
a) Most studies used anti-CD3 (Signal 1) and anti-CD28 (Signal 2) Mabs to stimulate T cells in vitro. Obviously, this is not how our T cells get activated in vivo. So which in vivo scenario do such antigen presenting cell (APC)-free stimulations simulate? For one thing, these Mabs drive a much higher avidity interaction, something we rarely observe when T cells interact with their cognate antigen peptide presented inside an MHC class II molecule by APCs like dendritic cells (DCs), a cell-replete stimulation. Let’s add to this other variables such as different Mab clones, different Mab concentrations, soluble, plate-bound or bead-bound Mabs. You see where I’m going? Given the variety of approaches used by different labs, it’s practically impossible to compare data. At any rate, Mab stimulation of T cells is incontrovertibly an artificial system and exactly how it relates to in vivo CD4 T cell stimulation is unknown and perhaps more importantly, unknowable. Yet this has been the dominant approach in identifying and defining the variety of CD4 helper T cell subsets we are today familiar with from Th1 and Th2 to Treg, Th17 and TFH (more about this subset below).
b) We use a variety of recombinant cytokines to drive CD4 helper T cells to differentiate in vitro into Th1, Th2, Th17, Treg and so on. Again, obviously, recombinant cytokines are artificial and absent in vivo.
c) The absurdity of using anti-cytokine Mabs such as anti-IFN-g or anti-IL-4 in in vitro CD4 helper T cell differentiation protocols. Again, how b and c relate to the in vivo CD4 T cell differentiation program is unknown and perhaps more importantly, unknowable. And yet, we have thousands if not tens of thousands of peer-reviewed papers that have used precisely such cell culture approaches to help define Th1, Th2 et al.
- Where in the CD4 helper T cell subset universe do we slot the T Follicular Helper Cell (TFH)? T Follicular what? Well, you see, given our zeal for CD4 helper T cell subsetting, we immunologists found yet another CD4 T helper subset in the T cells that move into the B cell follicle to help B cells secrete high-affinity, class switched antibodies. After all, we refer to CD4 T cells as helper T cells precisely because they help B cells, CD8 T cells and macrophages develop and manifest their respective effector functions right?
Expressing unique cell surface molecules such as ICOS, CXCR5 among others and unique transcription factors such as Bcl-6, c-maf, etc, and secreting additional cytokines such as IL-5, IL-6 and IL-21, the biggest surprise about Tfh cells was their plasticity (14, 15), in terms of capacity to express cytokines long corralled behind the mutually exclusive Th subset constructs. Could Tfh express IFN-g? Why yes! Could they express IL-4? Why yes! And so it went.
In hindsight, it’s hardly surprising that the so-called Tfh retained the capacity to secrete a wide variety of cytokines. After all, we know from mouse gene knock-out studies that IFN-g is a switch factor for IgG2a. In other words, a B cell switches from IgM to IgG2a after interacting with a cognate (specific for same antigen) CD4 T cell that secretes IFN-g. So much for the erroneous classifications that persisted for years and insisted on defining Th1 cells as indicative of so-called Cellular immune responses and Th2 cells as indicative of so-called Humoral immune responses!
Different CD4 helper T cell cytokines are necessary to help B cells switch from secreting IgM to other high-affinity antibodies including a variety of IgGs, IgA and IgE. Examined in this fashion, is it not reasonable to assume that migrating into a B cell follicle to help a cognate B cell, an activated CD4 helper T cell should retain its capacity to secrete a variety of cytokines, IFN-g to help the B cell switch to IgG2a, IL-4 to switch to IgG1 and IgE, IL-17A to switch to IgG2b, TGF-b1 and IL-6 to switch to IgA, IL-5 to enhance capacity for antibody secretion per se, and IL-21 to help the B cell become a plasma cell, etc? Seen in this light, don’t the non-overlapping CD4 T cell subsets we envisaged appear rather incongruent?
Well, how do we envisage CD4 helper T cells develop their effector functions? One hypothesis (16) could be that the location (the tissue site) of the immune response is a key determinant in directing and guiding the ensuing CD4 helper T cell effector function. This, for two reasons primarily. One, the tissue has a vested interest in ensuring, over evolutionary time, that immune responses taking place within its environs do minimal damage to it. Two, most pathogens perforce evolve select portals of entry and tissue tropism. However, the picture is not complete with these two reasons alone. We need to include microbiota. Most tissues sites harbor a variety of unique bacteria, viruses and fungi, some such as our skin, GI tract, female reproductive tract abundantly so. Surely, our microbiota participate not only in every immune response that ensues in the tissue where they reside but perhaps also in maintaining that tissue’s homeostasis, i.e. in immunoregulation? Thus, gut-, skin- or eye-associated CD4 T cell immune responses to the same antigen could be anticipated to be different and indeed are.
So here’s where we are. We immunologists identified Th1 and Th2 using a variety of highly artificial in vitro cell culture techniques, we observed some apparent evidence of this dichotomy in CD4 helper T cell function in vivo with the experimental mouse model of leishmaniasis, the early promise of this mouse model did not pan out along predicted lines, the mental construct imposed by this dichotomy nevertheless persisted and we began to classify an increasing plethora of CD4 T cell subsets such as Th3, Th9, Th17, Treg and Tfh over the past two decades. In other words, we let our increasing technological proficiency dictate the course for understanding the underlying biology rather than the other way around. Doing so, it took years before we came full circle to appreciating CD4 helper T cell plasticity (17), i.e. that CD4 helper T cells do not actually differentiate terminally into non-overlapping or mutually exclusive or even antagonistic secretion of particular groups of cytokines with different effector functions such as Th1, Th2, Th3, Th17, etc but rather they evolve particular overlapping features of functionalities depending on the activating cell(s), site(s) and microbiota, i.e. the CD4 T cell’s environment. Coming full circle in this fashion? Now that’s what I call irony.
- Noben-Trauth, Nancy, Pascale Kropf, and Ingrid Müller. “Susceptibility to Leishmania major infection in interleukin-4-deficient mice.” Science 271.5251 (1996): 987-990.
- Kelso, Anne, et al. “Heterogeneity in lymphokine profiles of CD4+ and CD8+ T cells and clones activated in vivo and in vitro.” Immunological reviews 123.1 (1991): 85-114
- Kelso, Anne. “Th1 and Th2 subsets: paradigms lost?.” Immunology today 16.8 (1995): 374-379)
- Matzinger, Polly, and Tirumalai Kamala. “Tissue-based class control: the other side of tolerance.” Nature Reviews Immunology 11.3 (2011): 221-230.