Round 18: Lupus and Complement Deficiency: Insights From Knockout Mice

By: Marina Botto, MD

Dr. Botto has no significant financial interest or relationships to disclose.

Release Date: February 24, 2009
Expiration Date: January 1, 2011


What I will try to discuss with you today is the controversial and paradoxical role that complement plays in lupus or Systemic Lupus Erythematosus (SLE).


  • To explain the links between complement and break of B cell tolerance
  • To state the links between complement and clearance of dying cells
  • To analyse the role of complement in tissue damage
  • To discuss the background strain effects in gene-targeted mice

We are only going to discuss one part of the complement, the classical part, which is mainly activated by immune complexes. The initial component of the classical pathway is called C1q. The complement cascade includes C4, C2, and C3; and then there is the formation of a membrane attack complex.

If you are a clinician, you are very well used to the idea that when a patient starts to have low complement levels, the patient is in trouble. You also know that complement increases when the patient is treated and doing well. Tissue biopsy shows complement deposition. These clinical observations led to the traditional view, which is: autoantibodies form immune complex with autoantigens, these fix complement and complement activation causes tissue damage. Therefore, the general view of clinicians is that the complement is the bad guy, the one that generates the damage. However, if one goes back to the clinical evidence, one realizes that the clinical evidence does not quite fit with this traditional view. What are the clues from the clinical observation?


















SLE (%)






Sex (F:M)





Sibling concordance (%)





The table is a summary of the patients who have inherited deficiency of one of the components of the early classical pathway activation. These are rare patients but very informative. If you look at the patients with C1q deficiency, you find that 93%, basically all of them—with the exception of one to my knowledge—developeded SLE. These patients developed SLE, and it is usually a very severe disease. Among C4 deficient patients, the percentage developing SLE is slightly lower. C2 deficient patients could be in your clinic because C2 is the most common complement deficiency among the Caucasian population. Only 10% of these patients develop lupus-like disease, and it is not particularly severe, not comparable to the disease that we find in C1q deficient patients.

When you consider C3 deficiency—in reality, only two patients have a lupus-like disease. These patients thus do not develop SLE. What emerges from this analysis is that there is a hierarchy of disease severity and prevalence, within the classical pathway of activation, with C1q deficiency being the strongest disease susceptibility gene identified in humans for the development of SLE. You could argue that this could be an ascertainment—these are rare patients—they develop the disease, come to the clinic and people discover it. If you look at sibling concordance in families with more than one individual C1q deficient, you find a sibling concordance of 90%. If you compare it with monozygotic twin studies that have been done in SLE, there is a sibling concordance of 20% to 40%. This is clearly telling us that the early part of the classical pathway plays an important role in protecting us from the development of this disease.

There is another paradoxical story, however. Complement acts as interface between the innate and the adaptive immune system. Complement (C3 and C4 fragments bound to antigens and immune complexes) reduces the threshold of B cell activation. Therefore, it helps in the development of antibodies. In addition, it enhances our immunological memory. Mark Pepys in 1974 used cobra venom factor to deplete complement. Mice were then immunized with sheep cells, and when these mice had been decomplemented by cobra venom factor, their immunoresponse was much lower. Approximately twenty years later, Prof D Fearon and colleagues showed the same point by attaching the HEL antigen to a C3 fragment (C3d), demonstrating again that if you bind more C3 fragments you have an immunoresponse to lower doses of antigen. Complement acts as an adjuvant.

Thus we have a paradoxical situation here. Complement deficiency is associated with reduced antibody responses and defective humoral immunity. I also showed you clear evidence from the clinical observations indicating that complement deficiencies are associated with SLE, which is a disease characterized by the production of autoantibodies, so increased antibody response. If you consider the well known functions of the complements system, (opsonin, anaphylatoxin, leucocyte activation and immune complex modification) you then realize that all or most of these functions clearly require C3, which is at the heart of the complement system. However, only patients with deficiency of one of the early components of the classical pathway develop autoimmunity. These observations indicate that there must be another function of the early part of the classical pathway that protect us from the development of SLE.

The steps where complement may play a role in SLE development:

  1. Complement and pathological survival of autoreactive lymphocytes.
  2. Complement and failure of clearing dying cells.
  3. Complement and tissue damage.

Complement and pathological survival of autoreactive lymphocytes. Does C1q have a direct role in the maintenance of B cell tolerance? We initially addressed this question by crossing the C1q deficient mice with hen egg lysozyme (HEL)-transgenic mice expressing: (a) Ig H and L chains with high affinity for HEL; and (b) sHEL (10–30ng) that is sufficient to trigger anergy but not deletion of transgenic B cells.

B Cell Development in Bone Marrow and Periphery Lymph Node

Figure 1. B cell development in the bone marrow and in the periphery lymph node, in the absence of the antigen, in the presence or absence of C1q: there is no difference. The lack of C1q did not alter the maturation steps of B cell development. Anti-HEL Ab in circulation in the presence of the antigen: C1q-deficient mice did not have anti-HEL antibody in circulation. There is normal cell tolerance to the soluble antigen in the C1q deficient mice.

More recently we looked at a more physiological model. We used the anti-ssDNA knock-in transgenic animal known as 3H9R/Vκ8R. The advantage of using this model is that we can also look at receptor affinity, editing etc. Again, we found that the C1q deficiency did not alter the production of autoantibodies in these transgenic models. Anergy induction in both models was not altered by C1q.

Complement and failure of clearing dying cells.

Figure 2

Lupus autoantigens concentrated on the surface of apoptotic blebs and bodies could become the drive of the immune response (Figure 2).

Does complement contribute to the clearance of dying cells? The other key observation that led us to ask this question was the finding by Ahearn and colleagues that apoptotic keratinocytes are able to bind C1q in vitro, in the absence of immunoglobulins. Therefore, you have a direct binding of C1q without the requirement of an antibody. Was this observation valid also in in vivo situation? We exposed the C1q deficient mice to UV radiation, which basically generated the so-called sunburn cells which are apoptotic cells, as you can see from Figure 3 (below).

Figure 3

Figure 3. We stained with an anti-C1q antibody and found that these apoptotic cells were positive. The C1q deficient mouse was the negative control.

The key observation came when we looked at the kidneys of these mice. These mice tend to develop glomerulonephritis. We looked at the kidney of the mice that did not have histological evidence of an ongoing inflammation. We found that the kidneys of these mice had multiple apoptotic bodies.

Figure 4

This was the first in vivo observation in a mammalian system that there was a possible defect in the clearance of these apoptotic cells. This was just a histological observation; can we prove that this was true in vivo? What we did at that time (a method that is now widely used) was to look at the clearance of apoptotic cells in a sterile peritonitis model. We recruited the macrophage in the peritoneum of the mice by injecting thyoglycolate. Then we injected the apoptotic cells in the peritoneum. Then we did a peritoneum lavage:
Figures 5 & 6

What you see above (Figures 5) is a delay in the clearance of apoptotic cells in the C1q deficient mice compared to the wild type mice. The figure shows the percentage of macrophages ingesting apoptotic cells. Two important points that emerged from this experiment are: (1) we proved in vivo that there was a defect in the clearance of these cells in the C1q deficient mice; (2) we proved that it is only a delay–it has never been a situation where there is total absence of clearance. Timing is also important. If we had done lavage only after an hour, we would have had the opposite observation. At the time point in the C1q deficient macrophages we had even more macrophage containing apoptotic cells compared to the wild type cells.

Are the observations from the mice biologically relevant to the human patients? In a sense, do we observe something that tells us something about the patients as well? We isolated macrophages from the peripheral blood of C1q deficient patients and fed them with apoptotic cells. We carried out a time course experiment as we did in vivo in the mice. What you observe here (figure 6) is that the macrophages from the C1q deficient patient had a delay in the uptake of apoptotic cells compared to the macrophages from normal subjects mirroring the observation that we had in vivo in the mice. In vitro, when we added the C1q in the culture, we could rectify the defect, so you see from the red bar. Since our experiment, there has been a lot of data in the literature regarding the role of complement and C1q in the clearance of apoptotic cells. To summarize briefly what the evidence shows:

  • The C1q binds directly to apoptotic cells. C1q bound to apoptotic cells:
    • Can activate the classical pathway with subsequent C4b and C3b/bi deposition.
    • Is recognized by calreticulin on phagocytes and this in conjunction with CD91 stimulates the uptake of apoptotic cells by pinocytosis.
  • C1q binding to apoptotic cells is mediated by IgM antibodies (see figure 7) and/or pentraxin and this causes activation of the classical pathway.

Figure 7

Fig 7. We used mice deficient of the soluble IgM and crossed them with the C1q-deficient mice, and obtained a double deficient serum; so we had serum that lacked C1q and IgM. This is an in vitro assay The most important observation came from the double deficient serum: if just C1q is added the defect is not rectified. If IgM is added, the defect is not rectified. However, if both C1q and IgM are added, the defect is rectified. What the assays shows is that. IgM antibodies, mainly natural antibodies, mediate one of the major mechanisms by which C1q recognizes apoptotic cells.

Does the impaired clearance of complement deficiency have an effect on B cells?
C1q and B cell tolerance to intracellular antigens:

  • Transgenic mice that express an intracellular membrane-bound form of HEL (mHEL-KK) that is sequestered in the endoplasmic reticulum of all MHC 1 expressing cells.
  • the same antigen (HEL) membrane bound induces deletion of the B cells. If you have soluble HEL, it induces tolerance by anergy.
  • With intracellular membrane-bound HEL, the autoantigen positively selects B1 cells and increases the titer of IgM autoantibodies. More importantly, these intracellular self-antigens are immunogenic, and can eventually stimulate the conventional B cells.
  • What happens if we take this model and cross it with the C1q-deficient mice? Essentially, we observe an increase in positive selection of the B1 cells and the level of IgM antibodies.

Figure 8

Fig 8. Serum anti-HEL IgMa, IgMa derives from the transgenic cells. The transgenic mice are the gray area; white area signifies the C1q-deficient mice. There is an increase in the concentration of the anti-HEL IgMaas well as an increase in plasma cells or colony-forming cells. Peritoneum: an increase in the B1 cells in the C1q-deficient mice. The lack of C1q has induced an increase in anti-HEL IgMa antibodies and in the B1 cells.

Figure 9

Fig 9. Conventional B cells. In the spleen and the lymph nodes in C1q-deficient mice, there is an increase in the number of conventional HEL-binding B cells. When these transgenic cells with antigen that is intracellular-membrane bound are made apoptotic (late apoptotic) the HEL antigen appears on the surface of the apoptotic cells. Essentially when you have an intracellular membrane-bound autoantigen, this autoantigen can become expressed on the surface of apoptotic cells, and then the natural IgM antibody and C1q helps in the clearance of these autoantigens.

The intracellular, membrane-bound antigen has the capacity of inducing high levels of IgM and high levels of B1 cells. If C1q is lacking, the level of IgM is even higher. Normally, in homeostatic situations, B1 cells generate natural antibodies that recognize self-sequestered antigens expressed on apoptotic cells. This, with the help of complement, mediates the disposal of dying cells. However, if the C1q is lacking (complement deficiency), there will be a defect in the clearance of apoptotic cells. As a consequence of this, these apoptotic cells may induce an antigen-driven stimulation of B1 cells, which then generate more natural IgM antibodies. The problem is that high titers of B1 cell derived IgM autoantibodies could be indirectly pathogenic if they were to mask autoantigens that would normally induce negative selection of bone marrow derived conventional B cells. . By masking the autoantigen on the apoptotic cells, the escape to autoreactive B cells from the bone marrow in the periphery is favored. Therefore, adequate amount of IgM derived from B1 cells is physiologically helpful. However, when increased, it becomes pathogenic because it would mask the autoantigen, and favor the escape of autoreactive B cells from the bone marrow in the periphery. That is what we think it is happening in this system when you have C1q deficiency. A defective clearance of apoptotic cells would favor the possibility that these cells may stimulate the conventional B cells. Thus the generation of natural antibodies and presence of complement regulates a steady-state clearance of dying cells and in this way affect the antigen-driven selection of B1 cells  However, B1 derived autoantibodies are low affinity and there is little evidence that they are commonly associated with autoimmunity.  Though we have not yet shown what is the link between our observations and the production of the class switched IgG antibodies present in SLE patients, we think this could be an important mechanism by which C1q deficiency drives autoimmunity

The role of C1q can be clearly different according to the phagocytes studied. When, for example, we looked at mesangial cells (a non-professional phagocytic cell) we could not see any effect of the C1q in the clearance of apoptotic cells.. The major source of C1q is macrophages and dendritic cells.

What is the role of C1q deficient dendritic cells in terms of the antigen presentation and T cell response or T cell tolerance? Data are in progress and thus I cannot give you a definitive answer, but I would say that we see differences in the way in which C1q-deficient dendritic cells present antigens to T cells.


Updated: August 16, 2012

Arthritis Center

About Arthritis Center

Founded in 1998, the Arthritis Center at Johns Hopkins is dedicated to providing quality education to patients and healthcare providers alike.