Central Tolerance

Central tolerance is the primary immunological checkpoint operating inside the generative (primary) lymphoid organs—the thymus for developing T lymphocytes and the bone marrow for developing B lymphocytes—that purges, reprograms, or functionally silences immature clones expressing self-reactive antigen receptors before they achieve functional maturity and enter the peripheral bloodstream. Because the vertebrate adaptive immune system generates its antigen-binding repertoire through stochastic somatic V(D)J recombination paired with random junctional nucleotide insertions by terminal deoxynucleotidyl transferase (TdT), an estimated 50% to 75% of newly assembled B-cell receptors (BCRs) and 20% to 50% of MHC-reactive T-cell receptors (TCRs) exhibit measurable autoreactivity against host self-proteins, nucleic acids, or glycans. Foreshadowed by Frank Macfarlane Burnet's Clonal Selection Theory (1957) and Peter Medawar's experimental induction of neonatal transplantation tolerance (1953), central tolerance prevents systemic autoimmunity by coupling receptor affinity thresholds during early lymphocyte ontogeny to apoptotic clonal deletion, secondary receptor editing, or agonist diversion into immunosuppressive Foxp3⁺ regulatory T cells (T_reg). Clones that escape central tolerance are subsequently policed in secondary lymphoid organs by peripheral tolerance.

Anatomic Compartmentalization: Cortex vs. Medulla

Hematopoietic progenitor cells migrate from the bone marrow into the thymus at the corticomedullary junction as CD4⁻CD8⁻ double-negative (DN1–DN4) thymocytes. After rearranging their TRB (β-chain) and TRA (α-chain) loci via RAG1/RAG2 recombinase in the outer thymic cortex, developing cells express a complete αβ T-cell receptor alongside both CD4 and CD8 co-receptors, becoming CD4⁺CD8⁺ double-positive (DP) thymocytes. Across this developmental trajectory, less than 2% to 5% of all thymocytes survive to exit the thymus as mature naive CD4⁺ or CD8⁺ single-positive (SP) T cells; over 95% die inside the thymus due to sequential quality-control selection enforced by specialized stromal microenvironments:

1. Cortical Positive Selection and Death by Neglect (Outer Thymic Cortex): Double-positive thymocytes crawl through a three-dimensional meshwork of cortical thymic epithelial cells (cTECs) displaying self-peptides bound to self-major histocompatibility complex (MHC) class I and class II molecules. Notably, cTECs express a unique catalytic subunit of the 20S proteasome termed the thymoproteasome (β5t / PSMB11) for MHC class I peptides, alongside cathepsin L and thymus-specific serine protease (TSSP / PRSS16) for MHC class II peptides, generating a specialized array of low-affinity cortical self-peptides.
   • Death by Neglect (~85% to 90% of DP thymocytes): If a thymocyte's randomly assembled TCR cannot bind any self-MHC molecule on cTECs, it fails to receive tonic pro-survival calcineurin and ERK signaling and undergoes mitochondrial apoptosis within 3 to 4 days.
   • Positive Selection and Lineage Commitment: Thymocytes whose TCRs bind self-peptide–MHC complexes on cTECs with low-to-intermediate affinity receive survival signals (upregulating Bcl-2 and the chemokine receptor CCR7) and commit to either the CD8⁺ cytotoxic lineage (if selected on MHC class I, mediated by the transcription factor Runx3) or the CD4⁺ helper lineage (if selected on MHC class II, mediated by ThPOK / Zbtb7b).
2. Medullary Negative Selection and Clonal Deletion (Thymic Medulla): Positively selected CCR7⁺ thymocytes migrate inward along a CCL19/CCL21 chemotactic gradient across the corticomedullary junction into the thymic medulla, where they spend 4 to 5 days scanning a dense network of medullary thymic epithelial cells (mTECs) and thymic dendritic cells (DCs). If a thymocyte's TCR binds a self-peptide–MHC complex in the medulla (or cortex) with high affinity/avidity, strong Ca²⁺ flux and JNK/p38 MAPK signaling induce the pro-apoptotic BH3-only protein Bim (BCL2L11) and the orphan nuclear receptor Nur77 (NR4A1), triggering caspase-dependent apoptotic clonal deletion. Thymic macrophages rapidly clear the apoptotic bodies via MerTK-mediated efferocytosis.

How the Thymus Mirrors Peripheral Organ Proteomes

A central paradox of thymic negative selection persisted for decades: how can thymocytes maturing inside the thorax be tested against tissue-restricted antigens (TRAs)—such as pancreatic insulin, thyroid thyroglobulin, ocular interphotoreceptor retinoid-binding protein (IRBP), gastric H⁺/K⁺-ATPase, or central nervous system myelin basic protein—that are otherwise synthesized exclusively in specialized peripheral organs?

In 1997–2002, positional cloning of the gene mutated in human Autoimmune Polyendocrinopathy-Candidiasis-Ectodermal Dystrophy (APECED / APS-1) and targeted mouse knockouts by Diane Mathis and Christophe Benoist solved this puzzle by identifying the Autoimmune Regulator (AIRE) on human chromosome 21q22.3. Expressed specifically inside mature MHC-II^highCD80^high medullary thymic epithelial cells (mTECs), AIRE is a non-canonical transcriptional regulator that drives promiscuous gene expression (PGE) of more than 3,000 to 4,000 peripheral tissue-restricted genes, enabling the thymic medulla to display a molecular "immunological mirror" of almost the entire body's proteome (~85% to 90% of all protein-coding genes).

Biochemical Mechanism of AIRE and Complementarity with FEZF2

Unlike classical transcription factors that recognize a specific DNA sequence motif at gene promoters, the 545-amino-acid AIRE protein operates as an epigenetic reader and transcriptional elongation release factor at poised, silenced chromatin:

1. Epigenetic Recognition of Unmethylated H3K4me0: AIRE contains a SAND domain, two plant homeodomain (PHD1 and PHD2) zinc fingers, and a caspase-recruitment domain (CARD) that mediates homomultimerization inside nuclear speckles. The PHD1 finger specifically binds the unmethylated N-terminal tail of histone H3 (H3K4me0)—an epigenetic mark characteristic of transcriptionally inactive tissue-restricted genes in thymic epithelium.
2. Release of Paused RNA Polymerase II and Topoisomerase Tortional Relief: At tissue-restricted loci, RNA polymerase II is stalled shortly downstream of the transcription start site by negative elongation factors. AIRE recruits the Positive Transcription Elongation Factor b (P-TEFb, CDK9/Cyclin T1), which phosphorylates Ser2 of the RNA Polymerase II C-terminal domain (CTD) and DSIF/NELF to release productive transcript elongation. Simultaneously, AIRE binds DNA Topoisomerase IIα (TOP2A) and the DNA-PK–Ku70/80–PARP1 complex to relax supercoiled DNA topological barriers and suppress cryptic enhancer RNAs in favor of coding TRA mRNAs.
3. Cooperative Antigen Hand-Off to Thymic Dendritic Cells: Any individual mTEC expresses a stochastic mosaic of 1% to 3% of the TRA repertoire at a given time. While mTECs can directly present endogenous TRAs to thymocytes via autophagy (loading cytoplasmic proteins onto MHC class II), mTECs also shed TRA-loaded exosomes and undergo rapid programmed turnover, transferring peripheral self-antigens to neighboring Sirpα⁺ and XCR1⁺ thymic dendritic cells for cross-presentation on both MHC class I and MHC class II.
4. The Independent FEZF2 Pathway and Thymic Mimetic Cells: In 2015, Hiroshi Takayanagi and colleagues demonstrated that a second transcription factor, Forebrain Embryonic Zinc Finger-Like Protein 2 (FEZF2), operates independently of AIRE (regulated via the Lymphotoxin-β Receptor, LTβR, pathway rather than RANK–NF-κB signaling) to transcribe a distinct, non-overlapping set of hundreds of peripheral tissue-restricted antigens (such as Fabp9, Krt10, and Resp18). Furthermore, recent single-cell chromatin atlas studies (Michelson, Benoist, & Mathis, 2022) revealed that post-AIRE mTECs co-opt peripheral lineage-defining transcription factors (such as Foxa2, Hnf4a, Myog, and Sox10) to differentiate into specialized thymic mimetic cells—including thymic tuft cells, microfold (M) cells, ciliated cells, enterocyte-like cells, and neuroendocrine-like cells—that faithfully reproduce peripheral post-translational modifications and splice isoforms inside the thymic medulla.

When AIRE is mutated in human APECED (APS-1), developing thymocytes reactive against endocrine and mucosal self-proteins fail to undergo medullary negative selection or tT_reg conversion, escaping into the periphery to cause the classic clinical triad of chronic mucocutaneous candidiasis (driven by neutralizing autoantibodies against Th17 cytokines IL-17A, IL-17F, and IL-22), hypoparathyroidism (autoimmunity against NALP5 and the calcium-sensing receptor), and autoimmune adrenal insufficiency (Addison's disease) (autoantibodies against steroid 21-hydroxylase, CYP21A2).

Not all self-reactive thymocytes encountering cognate self-antigen in the thymic medulla are destroyed by clonal deletion. Alexander Rudensky, Shimon Sakaguchi, and colleagues established that CD4⁺SP thymocytes whose TCRs bind medullary self-peptide–MHC class II complexes with intermediate-to-high affinity—falling inside a narrow avidity window above the threshold for positive selection but just below the threshold for Bim-mediated apoptosis—are diverted via agonist selection into the thymus-derived regulatory T-cell (tT_reg, formerly natural nT_reg) lineage.

This diversion proceeds via a two-step cytokine-instructed relay:

• Step 1 (TCR + CD28 Co-stimulation): High-affinity TCR ligation paired with CD28 engagement of CD80/CD86 on mTECs or thymic DCs activates the NF-κB subunit c-Rel and NR4A nuclear receptors, which bind the conserved non-coding sequences (CNS2 and CNS3) of the X-linked FOXP3 locus and strongly upregulate the high-affinity interleukin-2 receptor α-chain (CD25 / IL2RA) and CD122, generating a CD4⁺CD25⁺Foxp3⁻ tT_reg precursor.
• Step 2 (TCR-Independent IL-2 / IL-15 STAT5 Signaling): Subsequent stimulation of these CD25⁺ precursors by IL-2 (secreted by neighboring self-reactive thymocytes and DCs) or IL-15 (trans-presented by mTECs) activates STAT5 phosphorylation, which binds the FOXP3 promoter and CNS2 demethylated region (TSDR) to lock in stable Foxp3 transcription. Exported to the periphery, these self-antigen-specific tT_reg cells actively suppress escaped autoreactive effector T cells via CTLA-4 trogocytosis of CD80/CD86, IL-2 consumption, and IL-10/TGF-β secretion.

In the bone marrow, pro-B and pre-B lymphocytes assemble their immunoglobulin heavy (μ) and light (κ or λ) chains to express monomeric surface IgM (sIgM) as immature B cells. David Nemazee and Martin Weigert (1989–1993) discovered that when an immature B cell's sIgM is cross-linked with high avidity by a multivalent membrane-bound or matrix-associated self-antigen in the bone marrow (such as MHC class I molecules, cell-surface glycans, or dsDNA exposed on apoptotic erythroblasts), the cell does not immediately die. Instead, its initial primary tolerance mechanism is receptor editing:

1. Reactivation of RAG1 and RAG2: Strong sIgM cross-linking in the absence of BAFF survival signals arrests B-cell maturation, internalizes the autoreactive sIgM receptor, and sustains transcription of RAG1 and RAG2 (via FoxO1 nuclear retention).
2. Secondary Light-Chain Leapfrogging (IGK → IGL): Because the human IGK locus retains unused upstream V_κ segments and downstream J_κ segments on the same chromosome, RAG1/RAG2 catalyzes a nested secondary V_κ-to-J_κ rearrangement that excises the original autoreactive V_κJ_κ exon as an episomal circle and replaces it with a brand-new light chain. If all J_κ segments on that allele are exhausted, the recombinase cleaves the κ-deleting element (KDE) downstream of the C_κ exon to silence the κ allele, rearrangements proceed to the second homologous IGK chromosome, and if still self-reactive, advance to the λ light-chain (IGL) loci on chromosome 22.
3. Clonal Deletion or Clonal Anergy as Backup: Receptor editing successfully rescues 25% to 50% of all self-reactive immature B cells by pairing the original V_HDJ_H heavy chain with a non-autoreactive light chain. Only if all available IGK and IGL alleles are exhausted without extinguishing self-reactivity does the immature B cell undergo Bim-dependent clonal deletion in the bone marrow. Conversely, if an immature B cell encounters a low-valency soluble monomeric self-antigen (such as soluble serum proteins) that cross-links sIgM weakly, it downregulates surface IgM by 90% while maintaining high surface IgD, exiting into the spleen in a functionally silenced, short-lived state termed clonal anergy (characterized by chronic Lyn/SHIP-1 phosphatase activation and shortened 3-to-5-day survival due to inability to compete with normal B cells for follicular BAFF / BLyS).

1. Kappler, J. W., Roehm, N., & Marrack, P. (1987). "T cell tolerance by clonal elimination in the thymus." Cell, 49(2), 273–280. https://doi.org/10.1016/0092-8674(87)90568-X90568-X)
2. Nemazee, D. A., & Bürki, K. (1989). "Clonal deletion of B lymphocytes in a transgenic mouse bearing anti-MHC class I antibody genes." Nature, 337(6207), 562–566. https://doi.org/10.1038/337562a0
3. Tiegs, S. L., Russell, D. M., & Nemazee, D. (1993). "Receptor editing in self-reactive bone marrow B cells." Journal of Experimental Medicine, 177(4), 1009–1020. https://doi.org/10.1084/jem.177.4.1009
4. Anderson, M. S., Venanzi, E. S., Klein, L., Chen, Z., Berzins, S. P., Turley, S. J., von Boehmer, H., Bronson, R., Dierich, A., Benoist, C., & Mathis, D. (2002). "Projection of an immunological self shadow within the thymus by the aire protein." Science, 298(5597), 1395–1401. https://doi.org/10.1126/science.1075958
5. Klein, L., Kyewski, B., Allen, P. M., & Hogquist, K. A. (2014). "Positive and negative selection of the T cell repertoire: what thymocytes see (and don't see)." Nature Reviews Immunology, 14(6), 377–391. https://doi.org/10.1038/nri3667