Do we already understand all aspects connecting clearance and autoimmunity?

Just glimpsing at the clearance of apoptotic cells the issue seems rather simple: it works automatically, silently, efficiently and almost without witnesses. The scientists working in the field of embryology, hematology, immunology and histology teach that their disciplines assume that each day grams of cells have to die; this guaranties that organ development and tissue homeostasis are performed properly. Asking pathologists, if they see many apoptotic cells in tissue sections of healthy individuals, wherever they look at, they usually answer: “Yes, we have heard about that, however, we seldom observe such funny cells. But we see some necrotic cells in clinical conditions like inflamed skin after extended sun bathing or in ischemic areas after myocardial infarction and in the center of tumors that are not sufficiently vascularised; and sometimes they can be found in tissues from patients with autoimmunity!” So, do apoptotic cells really exist? To answer this question, we now interview an experimental biologist working with cell cultures, who proudly presents hundreds of cell culture dishes some of them just containing cells looking like raspberries, “I just irradiated them with UVB, 12 h ago!” Apoptosis—only an in vitro artefact? Not at all: the differences between these in vivo and in vitro observations—that's clearance!

Why to busy oneself with the cadavers of dead cells—if you can also work with viable ones? The main reason for that is apoptosis serves important functions besides just making room for a fresh and viable new tissue member and because we feel rather uncomfortable with the idea that our tissues may be covered with corpses over the years. When immunologists investigate tolerance, they face a serious problem: how is tolerance achieved for autoantigens that are not expressed in thymus or bone marrow? But we have a perfect source of peripheral autoantigens: the apoptotic cells permanently produced during normal tissue turn-over. Dendritic cells, the prototype of professional antigen presenters can take up these apoptotic cells, migrate to the lymph nodes and present their antigens to T cells in a tolerogenic fashion. However, if any cell, either viable or apoptotic, is infected by bacteria or viruses, the dendritic cells have to change their program. In this case they have to elicit a protective immune response against the pathogen. The instruction what to do is imprinted on the dendritic cells at the feeding ground, where antigen uptake takes place. Here, the microenvironment determines the outcome of antigen presentation (Table 1: danger and all-clear signals).

Table 1.  Danger and all-clear signals.

	Signals	Receptors
Bacteria	Danger: all bacteria contain various classical toll ligands via IL-12, IL-1β, TNFα, etc.	TLR 1, 2, 4, 5, 6 and 9
	All-clear: some expose PS (e.g. Treponema pallidum) via IL-10, TGFβ	Under investigation
Viruses	Danger: lysosomal nucleic acids cytoplasmic DNA via α/β-Interferon	TLR 3, 7, and 8 RIG-I and MDA5
	All-clear: enveloped viruses may expose PS (e.g. HIV-1) via IL-10, TGFβ	Under investigation
Parasites	Danger: lysosomal nucleic acids cytoplasmic DNA via IL-12, IL-1β, TNFα α/β-Interferon	TLR 3, 7 and 8 RIG-I and MDA5
	All-clear: some parasites expose PS (e.g. Leishmania) via IL-10, TGFβ	Under investigation
Early apoptotic cells	Danger: not known
	All-clear: PS via IL-10, TGFβ	Under investigation
Late apoptotic cells	Danger: monosodium urate crystals via IL-12, IL-1β, TNFα	TLR 2 and 4? Others?
	All-clear: PS via IL-10, TGFβ	Under investigation
Necrotic cells	Danger: ATP via Cl^− currents, Ca^++; cAMP	Purinergic receptors
	All-clear: PS via IL-10, TGFβ	Under investigation

What is the predicted consequence of defective clearance? For definition of “Clearance deficiency” see Table 2. This seemingly simple question turns out to be a rather complicated matter since the tissue where clearance deficiency occurs, the type of apoptotic cells that accumulate, the phagocytes and several cofactors involved are able to strongly influence the outcome (Tables 3-6). When we talk about macrophages we have to consider that we often use a single word for a plethora of very different cells, which, in addition, have several ways and stages of activation. Furthermore, it is almost impossible to isolate pure early apoptotic cells and if injected the cells continue their dying process and continuously undergo secondary necrosis. The necrotic or apoptotic cell-derived internal danger molecules are fugacious (e.g. ATP), require cofactors (e.g. HMGB-1), and thin out by diffusion (e.g. monosodium urate crystals). For sure, many of the molecules, signalling internal danger, do not “survive” the freeze-thaw procedure commonly used to produce necrotic cells. The best defined internal danger molecule released by secondary necrotic cells is monosodium urate, which has to crystallize to cause inflammation. The crystallization process requires an above-threshold concentration. Consequently, the uric acid released by scattered secondary necrotic cells is diluted by diffusion to a concentration below the crystallization threshold, resulting in physical detoxification of this potentially harmful compound. In in vitro assays secondary necrotic cell-derived monosodium urate will never reach the pro-inflammatory concentration. The scenario drastically changes in vivo, if a bulk of cells undergoes secondary necrosis, e.g. after chemotherapy. Taken together, the common denominator for most phagocytosis assays is that minor changes in the protocol may severely influence the read-out—some problems of in vitro and in vivo clearance assays are depicted in Table 7.

Table 2.  What does “clearance deficiency” mean?

The steps of the clearance process:
1. Find apoptotic cells by recognizing apoptotic cell-derived chemoattractants
2. Eat apoptotic cells by recognizing apoptotic cell-derived surface changes
3. Inspect apoptotic cells for infection with virus and intracellular bacteria
4. Digest apoptotic cells
5. Kill internal invaders if present
6. Send appropriate anti-inflammatory (IL-10 and TGFß) or pro-inflammatory (TNFα, IL-12, IFNα) signals
Clearance deficiency means: any of the functions 1–6 (and probably others) fail in at least one tissue

Table 3.  Clearance of apoptotic cells is heterologous (examples): tissues.

Tissue	Consequences
Any tissue: A defective clearance results in the accumulation of apoptotic cells, where they are supposed to progress to secondary necrosis within a couple of hours	It can be assumed that (I) internal autoantigens including nuclear material get accessible for autoantibodies, (II) the release of internal danger signals constitutes a pro-inflammatory micromilieu, and (III) toxic molecules cause damage to the surrounding tissue
Blood: Circulating phagocytes (monocytes and granulocytes) are lousy eaters. Blood-borne apoptotic or necrotic cells are mainly cleared during passage of the blood through liver and spleen by sessile macrophages. DNase-1 and C1q co-operate in predigesting circulating chromatin accidentally released by necrotic cells	Circulating cellular debris and dead cell-derived autoantigens e.g. nucleosomes are prone to be deposited as immune complexes in various tissues, e.g. kidneys. This is a generally accepted feature in the pathogenesis of kidney disease in SLE
Germinal centers of lymph nodes and spleen: a huge number of B cells undergo apoptosis—and are usually immediately cleared by specialized phagocytes, referred to as tingible body macrophages. In healthy individuals, non-ingested apoptotic cells can hardly be detected. This prevents secondary necrosis with release of (nuclear) autoantigens	Autoantigens released from secondary necrotic cells are prone to activate complement, bind to follicular dendritic cells and serve as autoantigen able to select autoreactive B cells that have accidentally gained anti-nuclear reactivity due to random mutations during somatic mutations. This mechanism may participate in the etiology of nuclear autoimmunity
Skin: After UVB exposure of skin many epidermal keratinocytes undergo apoptosis, whereas macrophages and dendritic cells escape into deeper layers of the skin. In healthy persons the phagocytes return and clear the apoptotic keratinocytes within one to two days	Clearance may fail in different ways: (I) apoptotic cells may accumulate and disintegrate directly leading to tissue inflammation due to release of internal danger signals. (II) The phagocyte responsible for the clearance may “misinterpret” the scenery and may call for help from inflammatory cells. There are published examples for both conditions—and there may be further possibilities
Cartilage: Chondrocytes live in small cages, referred to as lacunae. Single cells are spatially separated from other cells (e.g. macrophages). If chondrocytes die they cannot be phagocytosed by any other cell—they die the death of a hermit	If the “lonely death in the lacunae” is cause or consequence of cartilage destruction is currently not known

Table 4.  Phagocytes for the clearance of apoptotic cells.

Location	Phagocytes
Throughout the body	There is a plethora of macrophages and other phagocytes. Almost all organs contain tissue-specific macrophages that are equipped with different receptors, produce different cytokines, and endowed with specific functions adopted to the requirements in the respective organs
Tissues where high amounts of apoptotic cells are permanently generated (e.g. thymus)	Specialized professional macrophages persist that are assigned with the clearance function (referred to as tingible body macrophages [TBM]). They anticipate apoptosis and efficiently engulf dying cells already in the early phase of apoptosis
Tissues where huge amounts of apoptotic cells are generated temporarily (limb bud/development, germinal centers/affinity maturation)	TBM are recruited that efficiently engulf early apoptotic cells
Tissues where apoptosis is a seldom (e.g. inducible) event	Apoptosing cells are often just taken up by their neighbours. Alternatively, they are able to attract macrophages by the chemoattractant lysophosphatidylcholine
Throughout the body	Immature dendritic cells efficiently phagocytose apoptotic cells and present them in a tolerogenic or immune priming fashion, when the uptake is in an anti- or pro-inflammatory micromilieu, respectively

Table 5.  Adapter proteins opsonising apoptotic cells.

Autoantibodies	There are two types of antibodies that augment or inhibit the clearance of apoptotic cells. There are many receptors for the Fc portions of IgG and IgA
Annexin-1	Annexin-1 is required for uptake by macrophages of apoptotic cells
Gas-6	Gas-6 is the adaptor, connecting surface molecules of (early) apoptotic cells with the Mer-Kinase
C1q	C1q is bound by late apoptotic and especially by necrotic cells and chromatin. Calreticulin/CD91 serves as receptor for C1q
C4, C3	C1q gets activated by binding to late apoptotic cells or to dead cell-derived chromatin and causes the opsonisation with C4b and C3b of apoptotic cells. The C5 convertase and the lytic membrane attack complex are inhibited. There are several receptors for bound fragments of C3 and C4: CR1 (CD35), CR2 (CD21), CR3 (integrin αMβ2: CD11b/CD18), and CR4 (integrin αXβ2: CD11c/CD18)
Mannose binding lectin (MBL)	MBL binds late apoptotic and especially necrotic cells. Calreticulin/CD91 serves as a receptor for MBL
Long pentraxin (PTX-3)	Binding to late apoptotic cells of PTX-3 inhibits C1q-mediated phagocytosis by dendritic cells of apoptotic cells
Thrombospondin-1 (Tsp-1)	The affinity for apoptotic cells of Tsp-1 is low and requires supernormal concentrations of Tsp-1 for efficient opsonisation achieved only during acute phase responses. The integrin αvβ3 (CD51/CD61) serves as receptor for bound Tsp-1
Lectins	In the late phase of apoptotic cell death internal membranes containing immature glycoproteins are exposed, which may be targets for various lectins

Table 6.  Some receptors binding apoptotic cells.

Ligand	Receptor
Receptors for (opsonized) apoptotic cells	CR1 (CD35); CR2 (CD21); CR3 (integrin αMβ2: CD11b/CD18); CR4 (integrin αXβ2: CD11c/CD18); CD36; integrin αvβ3: CD51/CD61; integrin αvβ5; Calreticulin/CD91; Mer-Kinase
Further receptors involved in apoptotic cells recognition	ABC transporter; CD14; CD31; CD47
Receptor involved in recognition of the apoptotic cells-derived chemoattractant LPC	G2A

Table 7.  Problems of the clearance assays.

Assay	Limitations
General problems	Compared to the in vivo situation the in vitro phagocytosis of apoptotic/necrotic cells is generally rather lousy. Temperature, pH, calcium, and other soluble factors (immunoglobulins, opsonins, DNases, and other proteins) strongly influence the phagocytic capabilities of macrophages causing further variability of the assays. Adherence of phagocytes strongly increases their phagocytic capabilities causing further errors. There are several different amateur and professional phagocytes cooperating in the in vivo clearance, which is usually badly reflected in the in vitro assay often employing only a distinct population of phagocytes. Most assays do not investigate the chemoattraction of apoptotic/necrotic cells, which is of major importance in the in vivo situation
Short term assays	Short term assays are usually performed in huge excess of apoptotic/necrotic cells barely reflecting the in vivo situation. There is often a lag phase before phagocytes start eating causing large errors in short term assays. There are basically two types of assays (I) fluorescence labelled cells get incubated with phagocytes. The uptake of the prey is quantified by cytofluorometry or (II) non labelled cells get incubated with phagocytes. The uptake of the prey is quantified counting phagocytes containing condensed nuclei in light microscopy
Long term assays	In long term assays the apoptotic prey is degraded inside the macrophages, resulting in a “false low” clearance estimate. This can partially be overcome by measuring the “food” consumption
In vivo assays	The in vivo assays for clearance of apoptotic cells are complicated, hard to standardize and expensive. Furthermore, it is complicated to monitor the exact stage of the apoptotic cells at the moment of uptake
Ex vivo assays	A possibility to circumvent some of the problems of in vivo and in vitro assays is the use of whole blood ex vivo assays, containing all humoral (besides coagulation proteins) and cellular components of blood
Important	In patients with SLE the clearance of apoptotic cells may only be disturbed in certain tissues and under certain conditions

What do we observe in patients with SLE? There have been many reports on a delayed, impaired or altered clearance of apoptotic cells in humans with SLE (and in murine lupus). In some, but not all cases, an accumulation of apoptotic cells was reported. Alternatively, a more pro-inflammatory outcome of the clearance process or the infiltration with leukocytes of areas harboring apoptotic cells was observed. Table 8 lists some clearance-related facts reported for patients with SLE. Some, but by far not all models for murine lupus show an accumulation of apoptotic cells in various tissues, including the germinal centers of lymph nodes (MFG-E8 knockout).

Table 8.  What's about clearance of apoptotic cells in patients with SLE?

Observation	Interpretation
Dead cells accumulate in in vitro cultured PBMC from some patients with SLE	This may reflect clearance of apoptotic cells in vivo
There are patients with SLE displaying a generally low phagocytotic capability from different phagocytes affecting uptake of several types of prey in vitro	This is most likely due to general malfunctions of the uptake machinery
There are patients with SLE displaying isolated clearance deficiencies (e.g. in vitro generated macrophages insufficiently take up early apoptotic cells)	This is more likely due to a specialized malfunction of a component of the uptake machinery or to a defect in phagocyte development or maturation
Apoptotic cells accumulate in vivo in the germinal centers of some patients with SLE	Nuclear debris may serve as autoantigen providing survival signals for B cells that have accidentally acquired nuclear autoreactivity due to random somatic mutation
After UVB challenge apoptotic cells persist longer in the skin of some patients with CLE in vivo	Uncleared apoptotic cells may undergo secondary necrosis resulting in the release of toxic internal compounds

Coming back to the title questions: “Do we really know everything about clearance of apoptotic cells and autoimmunity? Is the pathway already brightly illuminated?”, our answer is: “We do not know whether there is a light at the end of the tunnel, but we are pretty convinced that there is a tunnel!”

Acknowledgements

Preparing this manuscript, we immediately noticed it was impossible to cite all the important contributors to this field. Therefore, we completely forwent citations, and refer to the usual databases for references.