The immunopathogenesis of SLE is complex and is characterized by multiple T- and B-cell abnormalities. Central to these changes are believed to be altered functions of DCs, the most important APCs [3
Peripheral tolerance is believed to be broken in SLE [26
]. DCs, which have a significant role in maintaining peripheral tolerance, have been found to be defective and proposed to be important in the development of autoimmunity in SLE [3
]. Of the two DC subsets, pDCs are thought to have a central role in SLE pathogenesis through the production of IFN-α, which has a pivotal role in inducing SLE [27
]. Although controversial, the number of pDCs in peripheral blood is aberrant when compared with that in normal controls [4
mDCs also have been found to be abnormal in SLE [24
]. Patients with this condition have deficient number of mDCs [4
] and monocyte-derived DCs that exhibit abnormal phenotypes and functions [32
]. Decker et al
] reported that monocyte-derived DCs from SLE patients expressed high levels of CD86 and produced increased quantities of IL-6 on stimulation.
Previous studies focused mostly on PBMC-derived DCs or DCs that are directly isolated from peripheral blood [24
]. It is not known whether defects of these DCs are secondary to (a) DC precursor deficiency; (b) microenvironmental changes in the bone marrow during DC development; or (c) microenvironmental changes or after Ag capture in the peripheral blood and the site of tissue injury. Although murine BMDCs have been studied previously, data on the characteristics and function of human BMDCs in patients with SLE is scarce. It is for this reason that we compared the phenotypic and functional characteristics of BMDCs from SLE patients and healthy individuals.
Traditionally, DCs are generated in vitro
by using GM-CSF/IL-4 [34
]. However, this method induces only mDC generation. In mouse studies, FL has been reported to be capable of inducing pDC development [21
]. Therefore, in our experiments, besides using the traditional GM-CSF/IL-4 culture method to study BM-derived mDCs, we also applied FL to induce BM cells to develop into DCs (which we defined as BM FLDCs), which showed features of both mDCs and pDCs and allowed us to study BM-derived pDCs indirectly.
In the present study, we confirmed that human CD3-
BMCs could be induced to become DCs, with FL as the only growth factor. Consistent with previous studies, BM FLDCs had an increased expression of DC-SIGN, a DC marker, and some costimulator molecules including CD40, CD80, and CD86 when compared with BMCs and the classic DC-culture system involving GM-CSF/IL-4 [35
] which induced mainly mDC development. FL appeared to induce both mDC and pDC development. During differentiation, some of the BM FLDCs expressed phenotypic characteristics (BDCA-2, CD123) similar to those identified in pDCs, whereas others expressed CD11c, which is normally seen in mDCs. In addition, we found that FL-generated mDCs and pDCs existed in a ratio of 1:1. This is consistent with findings reported in previous studies on murine FLDCs derived from BM and peripheral blood [21
To study the phenotypic and functional characteristics of BMDCs at different stages of differentiation, various agents were used to stimulate the maturation of these cells. For immature BM mDCs, TNF-α/LPS were used to stimulate their maturation. However, TNF-α/LPS have not been used to stimulate immature pDCs previously. In this study, therefore, we used CPG ODN2006/CPG ODN2216 plus TNF-α/LPS to stimulate BM FLDCs. After stimulation, BM FLDCs showed increased expression of CD80, CD86, CD40, and CD83, indicating that these cells could be stimulated to maturity efficiently by this method.
Results from our study showed that SLE BMDCs have defective phenotypic expression and function when compared with those from healthy subjects. CCR7 is a chemokine receptor that is preferentially expressed by mature DCs and is important for DC migration [41
]. In our study, we found that immature BM FLDCs from SLE patients expressed higher levels of CCR7 than did those from normal controls, indicating that these cells may have a stronger ability to migrate. Because no obvious differences in CCR7 expression were found between SLE and normal immature BM mDCs, the higher expression of this chemokine receptor on SLE immature BM FLDCs should have been contributed by the pDC population among these cells. The higher CCR7 expression may allow SLE pDCs to migrate into lymph nodes where they could interact with T lymphocytes. This may also partly explain the low number of pDCs found in the peripheral blood of SLE patients in some previous studies [43
]. However, to confirm that SLE pDCs have a higher ability to migrate, further studies using an in vitro
migration assay are needed.
During DC maturation, HLA-DR expression is upregulated. However, in patients with SLE, both BM FLDCs and mDCs expressed lower levels of HLA-DR when compared with controls. Previous studies have suggested that deficiency in HLA-DR expression might be the cause of increased susceptibility of patients with SLE to various infections [32
]. In our study, we found that SLE immature and mature BM mDCs failed to stimulate T-cell proliferation as efficiently as did those obtained from normal controls. This may be explained by their lower expression of HLA-DR. However, this was not true for BM FLDCs. Both immature and mature BM FLDCs stimulated higher T-cell proliferation compared with controls. As BM FLDCs include a mixed population of pDCs and mDCs and because mDCs did not stimulate T-cell proliferation efficiently, the effects of BM FLDCs on T-cell proliferation may be attributed to the pDC subpopulation of BM FLDCs. This effect may be related to the higher expression of CD40, CD80, and CD86 on SLE BM FLDCs than in controls.
To evaluate whether BM FLDCs comprise a subpopulation of pDCs and whether SLE BM FLDCs had higher pDC activity, we measured the level of IFN-α by using ELISA in the supernatants of BM FLDC and mDC cultures. IFN-α is produced mainly by pDCs, and its serum level has been reported to be higher in patients with SLE [27
]. In this preliminary analysis, we found that SLE BM FLDCs produced detectable IFN-α, whereas normal BM FLDCs did not. Furthermore, mature SLE BM FLDCs produced higher levels of IFN-α than did immature SLE BM FLDCs. Neither SLE nor control BM mDCs produced detectable IFN-α. These findings further confirmed that BM FLDCs consisted of both mDCs and pDCs, as per earlier suggestion. It also confirmed that pDCs were the more active of the two types of DCs in SLE and may be the major culprit in inducing autoimmunity in this condition. It should be noted that IFN-α measurement was performed only in the BMDC culture supernatants from a few subjects; further studies are needed to confirm this finding. It is interesting to note that a recent study showed that peripheral-blood pDCs from patients with chronic SLE had decreased in vitro
IFN-α-producing capacity and were desensitized to TLR9 stimulation [13
]. These data, plus those reported previously [3
] and our current data on BMDCs provide further important insight into the role(s) of pDCs in SLE pathogenesis. We hypothesize that pDCs are the dominant DCs during their development in the BM. These IFN-α-producing cells induce the development of SLE. However, they may subsequently become deficient, with reduced IFN-α producing capacity and tolerance to TLR9 stimulation, probably as a result of chronic and persistent exposure to DNA-containing immune complexes in the peripheral environment, which are a hallmark of SLE.
Some limitations to our study exist. First, the number of subjects studied was small. Second, our findings may not be generalized to all patients with SLE, as the patients recruited in this study all had some form of cytopenia or fever requiring further investigations, including a BM examination. Patients with other lupus manifestations were not recruited, as we considered it unethical to perform a BM examination in these subjects. Third, most of the patients studied were receiving some form of treatment, including immunosuppressive agents. It is, therefore, not possible to confirm whether the BMDC changes were a result of the underlying disease or that of the various lupus medications. It should, however, be noted that the majority of these patients had active lupus disease-related cytopenia or fever despite drug treatment; it is therefore tempting to suggest that our findings reflect the true role of BMDCs in lupus disease pathogenesis. Future studies should aim to recruit treatment-naïve or newly diagnosed patients with SLE. However, this will have to involve the collaboration of multiple lupus research units. It has taken the authors more than 2 years to recruit 13 suitable patients from a cohort of more than 500 patients for the purpose of this study. Alternately, future studies may include culturing control BMDCs in vitro with the various immunosuppressive drugs to evaluate whether they acquire a similar phenotype to the one described in this study.
Our study also did not examine whether the BMDC changes were intrinsic defects or secondary to microenvironmental changes in the BM, including the cytokine milieu in our SLE patients. This should be evaluated in future studies. DC precursors in the bone marrow are mainly CD34+
stem cells [35
]. Sun and colleagues [46
] recently reported that CD34+
stem cells from patients with SLE had abnormal expression of CD166 and CD123 and that these abnormalities correlated with the overall lupus disease activity. Mesenchymal stem cells (MSCs), an important compartment in the BM, are believed to be able to affect DC generation, although previous findings have been controversial [47
]. Deficient MSCs from patients with SLE have been reported [49
], but whether MSC may affect BMDC generation and functions requires further detailed studies. In addition, the phenotypes and functions of DCs from patients with SLE could be altered by genetic defects in cell lineage, or as a result of factors capable of inducing their differentiation and maturation. Previous studies have shown higher levels of multiple cytokines in the BM, some of which may be pathogenic in SLE [50
]. DCs from patients with SLE could bear genetic alterations that made them prone to maturation under abnormal conditions, or they may be normal cells with an abnormal phenotype and behavior induced by the bizarre microenvironment from which they were obtained. Further investigations are required.