The genetic basis of immune and autoimmune responses

GF Bottazzo, M Locatelli, A Fierabracci and D Fruci

Scientific Directorate, Autoimmunity and Immunogenetics Laboratories, Ospedale Pediatrico Bambino Gesu, Scientific Institute (IRCCS),Rome, Italy

Bottazzo GF, Locatelli M, Fierabracci A, Fruci D. The genetic basis of immune and autoimmuneresponses. Acta Pædiatr 2004; Suppl 445: 38–42. Stockholm. ISSN 0803-5326

HLA class I and class II molecules play a major role in the presentation of short, pathogen-derivedpeptides to T cells, a process that initiates the adaptive cellular and humoral immune responses.However, the factors governing a cell’s ability to respond or not to particular peptides are still notcompletely understood. Taking the example of a viral infection, in tissues infected with a virus,viral particles are taken up by antigen-presenting cells and uncoated. The viral DNA or RNAenters the nucleus, where it replicates. mRNA enters the cytosol and is transcribed into proteins.These proteins are degraded in proteasomes and the resulting peptides (8–10 residues) are loadedonto class I molecules for export to the surface of the cells. In the meantime, the groove of theclass II molecules is also preparing to accommodate peptides (12–24 residues) generated by theendocytic protein-processing pathway. The surface of the infected cell then becomes adorned withpeptide-loaded human leukocyte antigen (HLA) molecules. CD4� T helper lymphocytes engageclass II molecules and elicit responses from B cells, which will ultimately lead to antibodyproduction, whereas CD8� T lymphocytes become cytotoxic T cells. As a consequence, the virusis eliminated from the body. However, certain mysteries and challenges remain. How can, as anexception to this rule, an autoimmune response be the escape from the perfect machinery? Thisreview offers some hypotheses on how to see the problem through to its solution.

Key words: Autoimmunity, HLA class I and II pathways, self-tolerance

GF Bottazzo, Direzione Scientifica, Ospedale Pediatrico Bambino Gesu, (IRCCS), PiazzaSant’Onofrio 4, I-00165 Rome, Italy (Tel. �39 06 6859 2127, fax. �39 06 6859 2101, e-mail.bottazzo@opbg.net)

Throughout evolution, the immune system has devel-oped a sophisticated mechanism, which protects theindividual against intracellular and extracellular patho-gens, such as viruses and bacteria. Once the pathogenhas entered the body, the immune system is confrontedby a series of challenges. First in line are the innate (ornatural) responses of infected cells, followed by theresponse of cellular elements and adaptive (or acquired)immunity, in which B and T cells detect foreignantigens by means of their receptors. This providesflexible protection against the antigenic universe ofviruses, bacteria, parasites, etc. Each B and T cell has itsown antigen receptor and, upon antigen recognition,undergoes clonal expansion to form effector andmemory cells, thus providing specific acquired immu-nity. The B cells bind native antigens through immuno-globulins (Ig) expressed on their surface; uponmaturation, the B cells will be converted into plasmacells, which secrete antibodies. Through their T cellreceptors (TCRs), cytotoxic and helper T lymphocytesrecognize small antigenic peptide fragments in associa-tion with molecules of the major histocompatibilitycomplex (MHC). The MHC gene complex encodes twomajor classes of peptide receptors, human leukocyteantigen (HLA) class I and class II molecules (1). HLA

class I molecules are expressed on virtually allnucleated cells and are essential to detect pathogensthat replicate intracellularly, such as viruses. CD8�

cytotoxic T lymphocytes (CTLs) screen cell surfaces forHLA class I-bound peptides. HLA class II moleculesare expressed only on the surface of professionalantigen-presenting cells (APCs), such as macrophages,monocytes, dendritic cells (DCs), B cells, Langerhans’cells, activated T cells and epithelial cells in the thymus,and present exogenously derived proteins to CD4� Thelper (Th) cells. Differences in assembly, intracellulartransport and cooperation with specialized chaperonsallow HLA class I and class II molecules to presentpeptides coming from distinct intracellular compart-ments, and therefore from different sources. However,there are exceptions, and class I may handle exogenousantigens and class II may present endogenous peptidesnot coming from endosomes (see below).

Although T cells respond to antigens with highspecificity, they are not able to drive their response inthe absence of professional APCs. In fact, it is necessaryto distinguish between the affector and the effectorphases of T cell triggering. In the affector phase, CTLsencounter antigen for the first time. Activation of suchnaive T cells requires the interaction of HLA–peptide

2004 Taylor & Francis. ISSN 0803-5326

���� ������ ��� ���� ������ ����

DOI 10.0180/08035320310021237

and co-stimulatory molecules expressed on the surfaceof a specialized APC, most likely DCs and macro-phages. Immature DCs are considered immunologicalsensors, alert for potentially dangerous microbesthroughout the body. DCs capture the microbe or itsproducts by several mechanisms and leave the site ofinfection migrating towards the T cell areas of theproximal lymph nodes, via the afferent lymphaticsystem. During their migration, DCs undergo a matura-tion process, where they up-regulate their T cell co-stimulatory functions, becoming efficient activators ofnaive T cells. This occurs presenting both HLA peptidecomplexes and additional co-stimulatory signals, i.e.molecules of the B7 family, to T cells (2). The inter-action between these co-stimulatory signals and theircorresponding ligands on T cells results in the up-regulation of other ligands and secretion of variouscytokines that play different roles in the immuneresponse. Once activated by their encounter with anantigen, naive T cells proliferate, differentiate andacquire the capacity to enter peripheral tissues and tofight the invading pathogen, the so-called effectorphase.

While it is easy to figure out how DCs can captureexogenous pathogen products and present them to HLAclass II-restricted Th cells, this is more complicated forHLA class I-restricted responses, generally thought totarget endogenous antigens. However, another physio-logical pathway does exist, termed cross-presentation.For a long time, this was considered to be just amarginal phenomenon, but it has now been recognizedas playing an essential role in priming CTL responses toviruses. Indeed, cross-presentation refers to any situa-tion in which antigens synthesized in one cell can becaptured as exogenous antigens by APCs, processed inthe HLA class I antigen-presentation pathway and usedto prime CTLs (3).

HLA class I pathwayCells constitutively display peptides derived from allproteins, self-normal, non-self (e.g. viruses) or mutant(e.g. tumors), since they are not able to distinguishbetween them. It is essential that HLA class I moleculesdisplay the largest possible repertoire of peptides. It hasbeen estimated that approximately one third of newlysynthesized proteins are degraded within minutes ofbeing synthesized, thus representing the major source ofpeptides for HLA class I molecules.

How does this occur? First, a protein is covalentlylabeled by linkage to the ubiquitin molecules andsubsequently fed into the cell’s “mincer”—the protea-some—to be degraded into peptides of up to 15 aminoacids. Over 70% of generated peptides are too small tobind HLA molecules and are thus rapidly hydrolyzed tosingle amino acids recycled for the general metabolismof the cells. The ubiquitin–proteasome system is the

major mechanism used by eukaryotic cells to degrademisfolded and damaged proteins at the end of their life.Two types of proteasome have been described: theconstitutive (or standard) proteasome and the immuno-proteasome. The first can be expressed in any cell type,while the immunoproteasome (generated as a conse-quence of interferon-�) is expressed exclusively in cellsof the lymphoid tissue, such as thymocytes, splenocytesand DCs. Peptides generated by these two types ofproteasome seem to be different, but their exact role hasyet to be fully elucidated.

Once generated, these peptides enter the endoplasmicreticulum (ER) via TAP proteins and, after interactionwith novel aminopeptidases, they are appropriatelytrimmed to 8–10 amino acids to fit one of the emptygrooves of HLA class I molecules. Bound peptidestabilizes the HLA class I molecule, which can beefficiently transported to the cell surface. Since onlyHLA class I molecules loaded with peptides are stable,the spectrum of class I-presented peptides accuratelyreflects the proteins expressed within a cell (6). In thisway, the CTLs continuously sample a large number ofpeptides in both healthy and infected cells. Theinhibition of any component of this pathway substan-tially decreases HLA class I expression on the cellsurface.

In humans, the three different HLA class I molecules(-A, -B and -C) are characterized by an elevatedpolymorphism (up to 800 different alleles). Each ofthese molecules is coded by alleles derived frompaternal and maternal chromosomes, which means thateach individual can express up to six different HLAclass I molecules. If one considers that each allele canbind a unique set of peptides, a single cell will simul-taneously express a large number of HLA class Imolecules each with a different set of peptides, thusproviding the basis for efficient immune surveillance byCTLs. The function of an HLA class I molecule isinextricably related to its structure. The peptide bindinggroove contains six pockets, A–F, that accommodatepeptide side-chains of the anchor residues. Thesepockets vary between allelic variants and thus deter-mine the set of peptides that can be bound to each HLAclass I groove.

HLA class II pathwayUnlike HLA class I molecules, HLA class II molecules(DR, DQ and DP) generally bind peptides derived frominternalization and processing of extracellular patho-gens in the endocytic pathway of APCs. The generationof a full set of antigenic peptides and their loading ontoclass II molecules requires the action of severalendocytic proteases (6). After their biosynthesis in theER, class II molecules are diverted into the endocyticpathway by the invariant chain. The non-classical classII molecule HLA-DM acts as a dedicated chaperon in

ACTA PÆDIATR SUPPL 445 (2004) The genetics of immunity and autoimmunity 39

the lysosomal compartment to prevent the functionalinactivation and aggregation of empty HLA class IIdimers, thus enabling the antigen-processing system torespond promptly to the challenge presented by newlyentering antigens. The structure of the grooves of HLAclass II molecules is similar to that of HLA class Imolecules, but with subtle functional differences.Antigenic peptides of 12–24 amino acids in length bindto the cleft and extend on either side.

Immune and autoimmune responsesIn reviewing the mechanism involved in immune andautoimmune responses, certain concepts emerge. Ifcells have to manufacture their own specific proteins,they also have to recycle the ones that, for one reason oranother, are no longer useful for the general economy ofthe cell. So, proteins are reduced to peptides of a distinctlength by the proteasome. Only those of a certain length(around nine amino acids) and with particular electro-chemical characteristics ultimately fit into the groove ofthe corresponding HLA class I molecules and areexpressed as self-peptides on the surface of virtuallyall cells. As mentioned, in ordinary circumstances thesepeptides are not recognized by the TCRs of CTLs,otherwise the cell with the exposed self-peptide wouldbe killed.

What is the functional significance of having self-peptides exposed with HLA class I molecules on thesurface of a cell, if then nothing happens? Apparently,the peptides need to feed the HLA class I groovescontinuously to stabilize them. The system has to keepgoing efficiently, because it must be ready to substitutethe self-peptides with viral peptides derived, this time,from the chopping of the viral proteins by theproteasome. Owing to the abundance of the viralproteins produced by the rapid replication of the virus,the viral peptides compete with the self-peptides tooccupy the grooves that will be exposed on the surfaceof the infected cells. The incredible situation here isthat, while the previously exposed self-peptides on thesame cells were not recognized by CTLs, the CTLs—via their TCRs—now immediately recognize the pep-tides derived from the foreign agent (i.e. virus) and killthe virally infected cells.

At this point, an intriguing question arises: how doesone explain that a person, with an integral immunesystem, is capable of eliminating a virus (by killing thevirus and producing antibodies against it) within oneweek of its invasion of his body, but that the sameperson does not kill his own cells or produce auto-antibodies? After all, intrinsic self-peptides are alsoexposed, like the viral peptides, on the grooves of HLAmolecules. The easiest explanation is that there are noautoreactive CTLs in the peripheral lymphoid organs,but only a large battery of CTLs, ready to recognize allsorts of viral peptides, derived from viruses that the

individual has never encountered before. While thelatter concept is widely accepted, the former is not: notall autoreactive T cells are eliminated in the thymus,some manage to escape from the gland and, as a conse-quence, there are autoreactive CTLs in the peripherallymphoid organs. It is here that the thymus becomespivotal in deciding which CTLs must leave the glandand migrate to the peripheral lymphoid organs, andwhich must otherwise die and be buried there. Toperform these particular tasks, the thymus controls thecentral immunological tolerance, while the peripheralimmunological tolerance is carried out in the otherlymphoid organs. Incredibly, more than 98% of T cellsdie in the thymus: a real hecatomb!

When the concept of central tolerance grew inconsistency, the general belief was that only ubiquitousself-proteins (e.g. albumin, fibrin) were expressed in thethymus, whereas those manufactured by specializedcells, e.g. thyroglobulin and thyroperoxidase in thethyroid, or insulin and glutamic acid decarboxylase(GAD) in the islet� cells were not. Consequently, notall, but the great majority, of autoreactive T cells againstubiquitous self-peptides are eliminated in the thymus,while the T cells potentially recognizing the specializedself-peptides will escape intrathymic apoptotic deathand, after leaving the gland, end up populating theperipheral lymphoid organs. Nothing is static inscience; what is accepted as firm evidence today maybe less certain tomorrow. In fact, somewhat unexpect-edly, it has been reported that self-peptides produced inspecialized cells, those involved in endocrine organ-specific autoimmune diseases, were also expressed bycells in the thymus [for some authors by the epithelialcells (7), for others by the APCs (8) present in the gland]and thus potentially able, like the ubiquitous self-proteins, to eliminate the corresponding autoreactive Tcells.

Despite the lack of experimental evidence, twopossibilities have been suggested to explain how Tcells escape central tolerance in the thymus. The first isthat TCRs of certain T cells have a low affinity forparticular self-peptides expressed by the appropriatecells in the gland. The second possibility is that thesame thymic cells do not express certain self-peptides insufficient amount, so that the TCRs of T cells will notbind the corresponding processed self-peptides, whichin normal circumstances should be exposed on thegrooves of HLA molecules. As a consequence, theapoptotic cascade is not triggered and the autoreactive Tcells will not die (9). Hence, there are T cells in theperipheral lymphoid organs—“drifting mines”—which,if appropriately triggered, can be extremely dangerous.

Some qualification is appropriate at this point. Tocontinue the analogy, there are drifting mines that arenot so dangerous, but there are others that mayconstitute a real threat to the individual’s cells. In thefirst category, are the T cells whose TCRs have a lowaffinity for the self-peptide and, even if they encounter it

40 GF Bottazzo et al. ACTA PÆDIATR SUPPL 445 (2004)

expressed on the groove of HLA molecules of certaincells, are unable to bind it with sufficient avidity, so thechance of ultimately creating problems for these cells isquite remote. In contrast, the potentially dangerous Tcells are those that have escaped negative selection inthe thymus, because particular self-peptides were notsufficiently expressed by the epithelial/APCs cellspresent in the gland. Why do we keep calling thempotentially dangerous cells? It is because they cannotharm any cell expressing that particular self-peptide,despite not being tolerized against it. To become reallydangerous, autoreactive T cells must be activated by theAPCs, whose job is to transform autoreactive T cellsinto committed cells against that particular autoantigen.These T cells, if they are CTLs, are potential killers; ifthey are Th cells, they are potential inducers ofautoantibody production.

A practical example may help to clarify this concept.An infective agent or a toxic substance injures some�cells in the pancreatic islets, which then releaseautoantigens including, for example, GAD. Like otherreleased autoantigens, GAD should be captured fromAPCs, then processed by them and its self-peptidesultimately exposed on the grooves of their HLA class IIand class I molecules, the latter throughout the cross-presentation pathway. APCs travel to the draininglymph nodes, where they find distinct types of T cells.If an individual has T cells that are tolerized to thatparticular self-peptide (e.g. from GAD), they willrecognize it exposed on the surface of the APCs and,as expected, they will not be activated. In contrast, if inthe lymph node of another individual there are auto-reactive T cells, i.e. those that have escaped from thethymus, they will recognize the self-peptide as foreign.At this point, their interaction with the APCs is decisivebecause, via this interaction, the naive autoreactive Tcells will change from being not committed to becom-ing committed to that particular self-peptide. The keyelements for the activation process are the co-stimulat-ing molecules expressed on the APCs that, by bindingthe corresponding receptors on the T cells, make theiractivation possible.

The drifting mines are now triggered: autoreactive Thcells can induce the production of autoantibodies (e.g.anti-GAD) and autoreactive CTLs become lethal, ifthey encounter the same GAD peptide exposed on thesurface of the� cells, those spared by the previousinfectious or toxic attack. After all, that particular GADpeptide on the groove of the HLA class I molecule willnot be different from any viral peptide.

In general, there is no HLA allele association withinfections in humans (10). This lack of association givesalmost every individual the possibility to fight againstthe same virus with an almost equal chance of success.In fact, if one virus-infected individual produces sixviral peptides, which exactly fit into the six grooves ofthe HLA class I molecules, another individual infectedby the same virus may produce six different viral

peptides which also fit the grooves of six distinct HLAclass I molecules. This diversity of combinationsapplies to the multitude of individuals, who may beinfected by the same virus. Not only that, but allindividuals infected by the same virus possess their ownarray of CTLs ready to recognize, via their TCR, thoseparticular six viral peptide–HLA class I complexes and,despite the almost infinite number of complex combina-tions, the same virus will ultimately be defeated by anequally almost infinite number of CTL TCRs.

As mentioned, all of this is possible because of thehighly polymorphic nature of the HLA class I mol-ecules, which indicates that there is an incrediblenumber of HLA class I grooves across the humanspecies, each able to accommodate its own nonapeptidewith its particular shape and physicochemical charac-teristics. From a teleological viewpoint, this simplymeans that each individual has an army (the immunesystem) ready to fight against foreign invaders withtanks (e.g. T cells) loaded with bullets (i.e. TCRs) thatwill precisely hit the alien shape (i.e. the nona viralpeptide), as if they knew perfectly well who the enemywas, but whom, in reality, they had never seen before.The alien shapes must have a strong affinity for thegrooves of the HLA molecules, but, most incredibly, thenona viral peptides change with every new viralinfection and, despite this, the individual is ready tostrike with his own bullets (i.e. TCRs), which, everytime, specifically recognize distinct nona peptide shapesderived from each of the new infecting virus.

Unlike infections, human autoimmune diseases arein general HLA associated (11). The phenomenonseems to be restricted to certain HLA alleles, which arefound predominantly in Caucasians, e.g. HLA-DR3and -DR4. This could explain why certain autoimmunediseases rarely affect Asians and black Africans, inwhom these predisposing HLA alleles are rarely found.Why do certain HLA alleles predispose to and othersprotect against autoimmune diseases? There must befunctional differences between HLA alleles, someconferring susceptibility to human autoimmunediseases and others conferring resistance. A plausibleexplanation could be that HLA alleles that predisposeto autoimmune diseases bind self-peptides with lowaffinity, similarly to what is observed in NOD mice(12). It has also to be considered that, in individualswith the same HLA alleles, self-peptides might beexpressed in different amounts on the surface of theirthymic cells. With regard to self-peptides, the conceptof their affinity and the amount expressed openspossibilities for different scenarios. In an individualwithout predisposing HLA alleles, the high affinity of theself-peptides would induce the elimination of auto-reactive T cells, whether the expression of the same self-peptides was high or low. Conversely, in an individualwith HLA predisposing alleles with low affinity for self-peptides, high amounts of expression would also causeelimination of autoreactive T cells in the thymus,

ACTA PÆDIATR SUPPL 445 (2004) The genetics of immunity and autoimmunity 41

whereas low expression would induce their escape (thedrifting mines).

ConclusionAutoimmunity is a complex chain of events that beginsin the thymus (escape from central tolerance) and con-tinues in the lymphoid organs (escape from peripheraltolerance); the result is the release of drifting mines (i.e.,autoreactive T cells), which are just waiting to betriggered, thus becoming the effector arms of theautoimmune attack, which ultimately damage self-tissue cells. Although the jigsaw puzzle continues totake shape, some relevant pieces are still missing. Thefollowing questions are still unanswered.

� Why do autoreactive T cells die in the thymus, whenthey come into contact with thymic cells, whereas thesame autoreactive T cells, when they meet APCs inthe periphery, are activated?

� What is the repertoire of molecular receptors thatdiscriminates activated (committed) from naive Tcells?

� What is the mechanism that allows an activated(committed) CTL to kill an epithelial cell, presentingeither a viral peptide or a self-peptide on the groove ofits HLA class I molecule?

� Why does only a small proportion of individuals whopossess HLA alleles conferring susceptibility to auto-immune diseases finally develop the clinical mani-festations of these disorders?

� What creates the conditions in the thymus for a self-peptide to bind with poor affinity to the groove of theHLA molecules, which predispose to autoimmunityor to a poor expression of certain self-peptides (or noexpression at all) by the thymic epithelial cell?

� Why does the activation of autoreactive Th cells ingeneral precede that of autoreactive CTLs, as shown

by the presence of circulating autoantibodies, some-times even for many years before the individualprogresses to overt clinical autoimmune disease?

If answers can be found to these questions, then the fieldwill certainly move forwards, and we will be able,hopefully in the not-too-distant future, to learn why theimmune system, instead of defending us, attacks anddestroys our cell components.

References1. Klein J, Sato A. The HLA system. N Engl J Med 2000; 343:

702–92. Kamradt T, Mitchison NA. Tolerance and autoimmunity. N Engl

J Med 2001; 344: 655–643. Heath WR, Carbone FR. Cross-presentation in viral immunity

and self-tolerance. Nat Rev Immunol 2001; 1: 126–344. Rock KL, Goldberg AL. Degradation of cell proteins and the

generation of MHC class I-presented peptides. Annu RevImmunol 1999; 17: 739–9

5. Kloetzel PM. Antigen processing by the proteasome. Nat RevMol Cell Biol 2001; 2: 179–87

6. Janeway CA, Jr, Medzhitov R. Innate immune recognition. AnnuRev Immunol 2002; 20: 197–216

7. Sospedra M, Ferrer-Francesch X, Dominguez O, Juan M, For-Sala M, Pujol-Borrell R. Transcription of a broad range of self-antigens in human thymus suggests a role for central mechanismsin tolerance toward peripheral antigens. J Immunol 1998; 161:5918–29

8. Pugliese A, Brown D, Garza D, Murchison D, Zeller M, RedondoM, et al. Self-antigen-presenting cells expressing diabetes-associated autoantigens exist in both thymus and peripherallymphoid organs. J Clin Invest 2001; 107: 555–64

9. Goodnow CC. Pathways for self-tolerance and the treatment ofautoimmune diseases. Lancet 2001; 357: 2115–21

10. Kaufman J, Volk H, Wallny HJ. A “minimal essential MHC” andan “unrecognized MHC”: two extremes in selection for poly-morphism. Immunol Rev 1995; 143: 63–88

11. Davidson A, Diamond B. Autoimmune diseases. N Engl J Med2001; 345: 340–50

12. Nepom GT, Kwok WW. Molecular basis for HLA-DQ associa-tions with IDDM. Diabetes 1998; 47: 1177–84

42 GF Bottazzo et al. ACTA PÆDIATR SUPPL 445 (2004)