Further notes about the morphostasis concept – split files

(06) Reply to Melvin Cohn (dated August 2003

(Added here because it exemplifies and clarifies some of the arguments.)

First!

I came to a need to understand the immune system from an unusual point. It is probably helpful to understand the sequence so you can see where I feel the extant perception is flawed. There is a file that describes this in more detail but here is a short synopsis. (C=comment,Q=question, A=answer).

Well, at this point, and knowing little (molecular) immunology, I decided that the IS was not observing a rigid horror autotoxicus to self tissues. The pathophysiology of TB and common diseases like acne, mouth ulcers, psoriasis and a host of other component disorders (seen in syndrome form in the SNAs) was – to my mind – "all due to immune auto–aggression" directed at self tissues.

So! That is the starting block. Even if "self" is recognised as such by the immune system it is still a long way from avoiding immune auto–aggression to it. So, there is no substantial horror autotoxicus to self antigens. Indeed, as with acne, the major attack seems to be on self and there is, simultaneously, little (and certainly ineffectual) clearance of the pathogen.

Several articles at my web site expand on Behcet's syndrome, clinical morphostasis and the evolving perception of morphostasis.

Second!

Here are some general points about successive "paradigms" of explanation. Let's call them models for simplicity. All previously popular models probably have some parochial value – parochial in both temporal and focused qualities. Thus, Aristotelian explanations seem totally off beam to us today whereas they had value in their own eras in beginning to focus our reasoning into better and better models. To call them "wrong" is as absurd as saying Newton was wrong and that Einstein will prove to be wrong (string theory seems, already, to be highlighting the parochial limits of his theories – it certainly loses definition at Planck distances and with quantum effects). They were appropriate for their era.

When you look back on earlier models, you can begin to see that, within their own bubble of applicability, the details of many theories have continuing value. New models, to be successful, rarely annihilate earlier models. However, they do highlight the presumptions that were made in extrapolating from these parochial observations to implications of a more universal significance. The true facts do not change with models. However, many of these "facts" have been burdened with interpretations that have stretched them into fables. The interpretation of the facts may be radically altered.

One salient point – that is perfectly obvious from both the "Sense of self" debate and the recent Seminars debate on self/non–self – is that there are many models of how the Immune System (IS) works and there are virtual science wars in progress between the competing conceptual camps. Now, history would dictate that they are all, to varying degrees, right. What is missing is a conceptual glue that will weld them all together into a whole. In doing so, such a perception will highlight the presumptions that have stopped their protagonists from moving forward and together. It will also highlight the parochial virtues of the older views. The older models will be subsumed as special cases of the emergent successor. A more universal perception will have been achieved.

That is history. And it will repeat itself. Perhaps not yet – but it will sometime.

Now let's get started on the specifics.

Let me start with some simple questions. Is it likely that the body uses a homeostatic system to maintain the order and form of its tissues? Even the most hardened antagonist of my view of the IS should find it hard to avoid a "yes". Is the IS likely to play a role in tissue homeostasis? If you are tempted to say "no", then step back one. Is it likely that inflammation and its retinue of inflammatory cells have a role to play in tissue homeostasis? To answer "no" to this one would, I submit, fly in the face of a tautology.

Another question. Do simple multicellular organisms, as well as mammals, have a tissue homeostatic mechanism? Do invertebrates, as well as mammals, have inflammatory cells? Do invertebrates, as well as mammals, have an anamnestic immune system?

A last set of questions. Do lymphocytes originate from a stem cell population that is closer to heart (or muscle or skin or what–you–like) than monocytes (macrophage precursors)? Do the surface markers of myeloid–lymphoid cells show a progression from phagocytes to lymphocytes? Could lymphocytes be regarded as amplifiers of the inflammatory response?

Your specific questions.

"How does the IS assay 'maintenance' and 'health'?" Essentially, it doesn't. "Assay" is a teleology that may come from a conviction that lymphocytes patrol the body "looking" for invaders. As David Lo's team have pointed out (1), a block of tissue cells form a network that effectively monitors the health of its own cells and tissue in general. Inflammation does not start until this network has cried "foul" and released chemicals that invite an inflammatory ingress. Individual cells monitor their own health and respond accordingly. Should you doubt these views – and the parallels in various multicellulates that range from plants to mammals – then it is worth getting a copy of a synoptic article by Trisha Gura in the New Scientist (Evolution and comparative immunology section in Supportive Research).

I think the preceding paragraphs have put up a good case to answer "how was it selected for?"

"It might be well to recall that your assumption has a long history . . ." I will answer this with a quotation: "Originality is nothing but judicious imitation." [Voltaire.] I am not staking a claim to intellectual property – I do have well defined views of what constitutes a better perception.

These have gone nowhere because they have not been mapped out onto a reasonable mechanism. I guess that you may not have read or, at least, properly assimilated the "Phlogiston" article at my web site. I do not think that the mechanism is unclear there.

"All that we can do is argue about the semantics of the terms self and non–self." Well, that's not a bad place to start. Now, one thing I think is worth considering (ignoring the teleology of "discrimination" for the moment) is what logical pairs we can consider. The body may discriminate on the basis of one or many of the following opposites.

We can look at each of these in turn and think "how much does this one contribute and when?" However, it is not – to my view – permissible to look at crossed pairs. For example, it is illogical to look at "self–animal/pathogen" discrimination or "non–foreign–organism/pathogen" discrimination or "self–peptide/pathogen–peptide" discrimination. To be logical, we must tie our parameters down as tightly as possible. Then we must consider how the system might feasibly carry out each tightly–defined discrimination.

I perceive that the way you have freely interchanged foreign–organism and pathogen is inadmissible and illogical. And I suspect it also leads to confusion.

"The straightforward assumption that the IS was selected to protect against parasitism, infectious disease and the like is so obvious. . . ." And the world is flat, the earth is at the centre of the universe, there has to be a creator, time and distance are rigidly fixed throughout the universe – and others. No one in their right mind used to doubt these obviously obvious assumptions. Revolution is science has regularly hinged around counterintuitive proposals. The immune system may be no exception. Indeed, when agreement cannot be obtained it is very likely that there is a false assumption somewhere in the prior conceptualisations. I think this would fall under the general description "Bacon's idols".

Now that does NOT dictate that the IS does NOT tend to produce the effect of protecting against parasitism, infectious disease and the like. It hinges around the danger of the teleology "selected to protect". We cannot help talking and explaining with teleologies – the closest we can get to avoiding them is to use a mathematical rather than a social language for description. because we can't avoid them we must be extremely sensitive of their dangers. My point is that the IS is quite mindless of the need to protect against parasitism etc. The effect occurs because it is tidying up the mess that these agents cause.

". . . tied into the regulation of the biodestruction function." So what is happening? Agreed, and without question, is the principle that the IS is capable of biodestruction. But most of this biodestruction is focused around subsets of self cells. Why? – for the following reasons.

Let's start with complement. First, C3 is secreted by phagocytes and actively concentrated in the plasma (by liver phagocytes). It is designed to constantly "condense" onto biological surfaces. Once there a cascade begins that results in the formation of C3b. If you are a healthy self cell then you secrete a C3–inhibitor that averts the C3b from triggering a pro–inflammatory cascade by moving it along to an inactive (non-aggressive) form of C3. This also applies to C4 when a cell gets coated with Ig. Two things might lead to self–cell lysis or fatal opsonisation; reduced C3–Inhibitors (do these downregulate just like Mhc Class I in sick cells?) or excessive Ig attachment. Either way, it is clear to see that there will still be a differentially greater opsonisation of other–than–healthy–self cells compared with healthy–self cells. And this can only start in the presence of plasma and/or phagocytes (ie, after an inflammatory ingress). Coating with Ig results only in bypassing the early stages of the alternative cascade. It attaches a marker to say "get on with complement assessment".

Now let's look at the biodestructive functions of cytotoxic and helper T–cells. Cytotoxic T–cells (Tnk and Tc) are designed to encourage self cells into early apoptosis. Unless some foreign organism could manage to incorporate self–class–I/peptide into its membrane, Tc cells will have no part to play in their disposal. So, cytotoxic T–cells are designed to encourage the melting away (by apoptosis) of self cells.

So what about T helper cells? Now, I think there may be a difference of belief here. You seem to hold that helper T cells are encouraged into aggression by primer T–cells whereas the Lafferty/Cunningham view is that APCs are the primers. Let me assume that APCs are the primers. What is it that the Th1 cell will respond to on epitope re–encounter. It will be another APC presenting a similar peptide in the jaws of its Class II Mhc antigen. It is responding to the presence of a self cell and sets off an intense inflammation in its proximity – an "inflammatory bomb". The biodestructive effect is caused by a cascade that intensifies inflammation at the location of this TCR/APC encounter; nothing more, nothing less. So where is the direct attack on the pathogen? It is, I submit, non–existent. It boils down, in effect, to a hard time for both healthy–self and other–than–healthy–self cells but it is differentially more destructive with other–than–healthy–self cells. But this is differentially cruel to organisms that are outside self cells. For organisms that reside inside self cells, the most effective response is to differentially encourage these flawed cells to die early. Otherwise, it must rely on a differential reduction in the infected cell's healthy self signature.

There is a tendency to look on the biodestructive mechanism of the IS as something that is directed at foreign organisms. For the reasons given above, this simplistic perception is far too naïve.

"You have given me no argument that Ig/TCR have a role in regulating normal physiology." Now, I think that this is your caricaturisation of what I have said. Strictly it is right but you go on, a few sentences later, to demonstrate that you equate this normal physiology with the normal physiology of chemical homeostasis (like glucose control). That is a very liberal and literal stretch of the original thesis. However, since you raise it then it is worth noting that the insulin receptor is closely related to growth factors and that islet cell physiology is intimately affected by gap–junctional communication. So you are, perhaps, not that far out in extending it to be a morphostatic mechanism: but I guess, the effect you sought was to question the perspective. (Gamma-delta T-cells now generally recognised to be involved in tissue homeostasis.)

"This is to be contrasted with their roles in guiding phagocytosis, cytotoxicity, C' lysis, 'transmitter and cytokine' production, etc. all biodestructive." I think that I have already shown how a different perspective can be laid out for these – and in favour of a morphostasis perspective. The facts don't change: the interpretation baggage does. (Both your view and the morphostasis view equally carry such baggage: as will any fresh, usurping view.)

"What is there about that you can't buy?" I buy all the facts! But I don't buy the interpretation baggage. I suspect other views more than I suspect my own but I don't even believe my own represents "the truth" or that it is secure.

"Try to answer for me; what will it take to have you change your mind?" Mel, I think I see your point of view. It is not that I cannot appreciate your perspective. It is that I think another perspective is possible and, once realised, it is better because it opens up a new vista on things. I don't feel tempted to abandon my views and replace them with yours. You quote Occam's razor in support of simplicity. But, I think that we should not be seeking simple mechanisms (biology is replete with the opposite) but simple principles that follow a limited number of common themes. Find a principle or theme that is simple to state, then see that it makes sense of ideas across a whole raft of previously disparate areas of study, and you truly have found an Occam's razor.

"The existence of debilitating autoimmunity shows that Ig/TCR mediated effector functions can destroy self cells and tissues." I have acceded this and even added my contention is that this is but the tip of a very large iceberg.

The Ig/TCR mediated ability to destroy cancer cells only confirms this, as the difference between cancer and non–cancer as seen by the I.S. is the presence of a "foreign" antigen. Sorry but I would say that they are not foreign (in the way we humans might conceive to be foreign). They are antigens expressed during retrodifferentiation and they are germ line encoded. The only exception might be if the mutant oncogene were a surface protein that gave rise to an entirely new peptide species. Macrophages have no trouble in recognising cancer cells in vitro. It is the anergy, necessitated by the T–cell system, that has led to their apparent in vivo invisibility as sick cells.

"The destruction of virally infected cells is another illustration." I think that you are providing further arguments that come to my concept's rescue, not to its demise – see above.

". . . when there is an insufficiency of eTh (delivery of Signal[2])." So, you are out on a committed limb here in persisting to believe that signal 2 is delivered by an eTh, NOT an APC (à la Lafferty/Cunningham). It might, of course, be the right limb. You claim this is simple, an Occam's razor solution: but you risk danger. Is it simplistic rather than simple? – and where is the objective evidence to support your contention? As tacit support for the perspective that I hold, it is instructive to follow the argument through carefully with both aggression or/tolerance selection and with thymic selection. By this view they end up as being the product of a repeating and much more general theme.

Let me set out, again, how I think that T–cell commitment is being made. It is important to exorcise the "tainted think" that attaches some quality to the Mhc/peptide epitope. My impression is that it is hard for people to cast it off. Many still need to (conceptually) award the antigen a quality (a relic of self–epitope/nonself–epitope perspectives). It is JUST an epitope. All epitopes – self and non–self – are just epitopes: nothing more: nothing less: all equal: it is a truly egalitarian system. It is only their context of presentation that leads to aggressive or tolerant responses. I think this last sentence cannot be repeated often enough to ensure that it is appreciated. (Nb, "context" subsumes your "time" as a special case). Let me, for emphasis of function, create the terms Step 1 and Step 2 that represent the precursor T–cell's response to Signal 1 and Signal 2 respectively. In the thymus in step 1 (part A) the T–cell spontaneously develops a particular specificity. This is the bit that ensures the Class I or Class II molecular clasp (not the peptide) is selected for. I suspect this is promoted by a gap junctional "kiss of life" and it turns the precursor T–cell into a selectable T–cell. Effectively, it is going to restrict the recognition of peptides to those that are carried in a self Mhc clasp. Then in Step 1 (part B) the lymphocyte refines its specificity to add in specificity for the peptide groove. Both parts of Step 1 are randomly generated. From now on, selection for aggression or tolerance is the same in the thymus as it is in the periphery. Thymic presentation is almost always non––inflammatory (there may be actual or experimental exceptions). In selecting only self Mhc carriers for the "kiss of life", a lot of precursor T–cells go on to apoptosis by neglect. These apoptosing cells are presented to the carrier specific and peptide specific (but uncommitted) T–cells. The result is the demise of both CD8 +ve T–cells capable of recognising any peptide presented by a dying T–cell: or the demise of any CD4+ve T–cell capable of recognising the debris derived from an apoptosing T–cell that is ingested and then presented by a thymic APC. Effectively this step amounts to the same process as Step 2. Other thymic cells might join in this apoptosis and also be presented to encourage tolerance (thymic epithelial cells, thymic APCs, thymic macrophages). Whatever, the result is tolerance to any peptide/Mhc antigen presented after efficient apoptosis in the thymus. I don't think that we have finally excluded the possibility of suppressor T–cells that can amplify tolerance on re–encounter – so these are also possible complicators of the tolerance mechanism. Now, this process of tolerance will use the same mechanism in the periphery as it does is in the thymus – efficient apoptosis. Inefficient apoptosis (due either to some mutant defect in its silent clearance or overwhelming apoptosis where apoptotic bodies – packed with Il–1 – spill their contents before ingestion) may invoke an aggressive response. SLE is a good example.

Aggression is induced when an APC presents (almost certainly in the lymph nodes) in an inflammatory condition (eg, rich in split Il–1, eicosanoids and interferons). Step 2 has nothing to do with any property of the antigen OTHER than in what inflammatory or non–inflammatory environment it made its debut. The lymphocytes are subsequently (or simultaneously) awarded a location address so that they are preferentially sticky to venules with a similar address code (part of Zinkernagel's localisation). So why do self cell signatures not provoke aggression? Part of the reason is that they share many peptide signatures with the apoptosing, precursor lymphocytes (and etc?) in the thymus. But the other part, and what gets left out of the reckoning, is the staggering volume of apoptotic cell death that is the daily consequence of housekeeping. All self epitopes are far more likely to be met in a non–inflammatory (tolerance favouring and paratope "mopping up") environment than in an inflammatory environment. Chance favours the prepared. The commitment to aggression is (for self/nonself) a non discriminatory, mindless event. Simply make older precursor T–cells more susceptible to aggressive commitment than young precursor T–cells, add the massive mopping up by apoptosis and you strongly augment the favouring of strange over common epitopes as fodder for aggression. The response to strange epitopes will be fast – like lighting a bonfire of thin dry timber: the response to self epitopes will be slow: like trying to start a sodden wet bonfire in a downpour.

In a way, this accords with your idea that older i–Th cells spontaneously move towards aggression. This is supported by experimental evidence (chapter and verse I would have to find). However, in denying that it is the APC that provokes Step 2, you distance yourself from my view.

"Do you wish to conclude that germline selected recognition of nonself markers like Danger, Morphostasis, integrity, localization etc. can be translated into decisions made somatically at the level of recognition of epitopes (which include MHC bound peptide)?" I am a little bemused about what you have in mind here. The germline "markers" of tissue mess will be analogous to the mess you and I see on our streets. First there will be the bits that get presented, nicely wrapped up in plastic bags, concentrated for quick collection. However, one or two of these gets ripped open by the foxes and the rubbish is progressively scattered all over the road. When the garbage collectors go on strike then entropy ensures that these neatly ordered packages progressively spill their contents (foxes, storms, vandals). Normally, the garbage collectors know what the bags look like and can sort them into paper waste, food waste and etc on their bag coding. It is plainly obvious for all what spilt rubbish looks like and I guess, that from time immemorial, animal cells have had a fair idea what constitutes the equivalent of connective tissue and cellular garbage. Macrophages (amoebocytes) are particularly attuned to mopping this up. Inside out membranes, spilt mitochondrial markers, spilt cytoplasmic markers will all have recognisable signatures and then there are the spilt chemicals – like Il–1 and eicosanoids – that shriek "mess ahoy".

I think that the morphostasis concept can subsume aspects of the AAR (provided you can drop your objection to APCs being the deliverers of Signal 2), stranger, integrity (order disorder discrimination), localisation and the immune homunculus. The views I find most distant from my own are yours, those of Medzitov and Janaway, Peter Bretscher and, a bit less distant, Coutinho. The main reasons for the difference with the first three is their absolute requirement that self–nonself discrimination by lymphocytes is effectively seen as a concrete reality. With Coutinho's view, the network is king and I feel this would result in a remarkably Rube Goldberg system. The network/homunculus ideas may be rewarded with an observable shift as described but they would, for me, be the passive consequence of morphostasis and not the active cause of discrimination. I think that morphostasis can subsume them all. It will show up the warts of presumption that have driven them up a blind alley and also, so far, prevented their integration.

I am sure that you will want to disagree – at least on first pass. But that is the fun of modeling. We must all accept the potential frailty of our conceptions.

  1. Lo D, Feng L, Li L, Carson, Crowley M, Pauza M, Nguyen A & Reilly CR. "Integrating innate and adaptive immunity in the whole animal." Immunological Reviews, 1999, Vol. 169;225-239. The authors reference a lovely quote about LPS in this article; it is worth reading this.
  2. "Roots of Immunity" by Trisha Gura in the New Scientist (Feb 18th 2000) pp24–28

Nb, anything in non-bold italics in this reproduced e-mail has been added later.