Search tips
Search criteria 


Logo of ctsLink to Publisher's site
Clin Transl Sci. 2009 August; 2(4): 300–308.
Published online 2009 July 31. doi:  10.1111/j.1752-8062.2009.00135.x
PMCID: PMC2790176

Lymphocyte‐Bound Complement Activation Products as Biomarkers for Diagnosis of Systemic Lupus Erythematosus


Systemic lupus erythematosus (SLE) is frequently misdiagnosed due to the lack of definitive diagnostic tests. The purpose of this study was to determine specifically whether complement activation products (CAP) are deposited on lymphocytes of SLE patients and whether lymphocyte‐bound CAP (LB‐CAP) may serve as novel biomarkers for the diagnosis of SLE. We conducted a cross‐sectional study of 224 patients with SLE, 179 patients with other diseases, and 114 healthy controls. LB‐CAP on peripheral blood lymphocytes was measured by flow cytometry. Diagnostic utility of LB‐CAP was determined by receiver operating characteristic (ROC) analysis. Significantly elevated levels of C4d and C3d were detected specifically on T and B lymphocytes (designated T‐C4d, T‐C3d, B‐C4d, and B‐C3d) of SLE patients. As diagnostic tools, T‐C4d and B‐C4d, respectively, were 56% sensitive/80% specific and 60% sensitive/82% specific in differentiating SLE from other diseases. Moreover, compared with measurement of anti‐dsDNA, serum C3, or serum C4, measurement of T‐C4d/B‐C4d was significantly more sensitive in identifying SLE patients during a single clinic visit. This is the first investigation of lymphocytes bearing complement activation products in human disease. T‐C4d and B‐C4d have high diagnostic sensitivity and specificity for SLE and may have added value to current laboratory tests for SLE diagnosis.

Keywords: systemic lupus erythematosus, SLE, lupus, complement, biomarker


Systemic lupus erythematosus (SLE) is arguably the most clinically and serologically diverse autoimmune disease, with more than 100 autoantibodies found in patients and disease spectra ranging from subtle symptoms to life‐threatening multi‐organ failure. 1 , 2 , 3 Owing to its complex etiopathogenesis, heterogeneous presentation, and unpredictable course, SLE remains one of the greatest diagnostic challenges to physicians, including rheumatologists. 4 , 5 Currently, the diagnosis of SLE is primarily based upon American College of Rheumatology (ACR) criteria, 6 , 7 many of which are subject to interpretation and may require years to evolve. The lack of specific, reliable, and validated biomarkers for SLE not only leads to misdiagnosis and misguided therapy, but also may result in flawed clinical trials if patients in “lupus” treatment arms include false‐positive diagnosis.

Serum C3 and C4 levels have been measured for decades in attempts to monitor disease activity in patients with SLE; however, these complement assays are not considered useful for the diagnosis of SLE. 8 , 9 , 10 , 11 , 12 , 13 We have previously revisited the complement system as a source of SLE biomarkers and discovered that cell‐bound complement activation products (CB‐CAP) hold significant promise as diagnostic biomarkers for SLE. Specifically, erythrocyte‐bound C4d (E‐C4d), erythrocyte CR1 (E‐CR1), and platelet‐bound C4d (P‐C4d) are highly sensitive and specific biomarkers for lupus diagnosis. 14 , 15 Initial studies indicate that these CB‐CAP provide significant added value to current diagnostic tests for SLE, and are capable of capturing the majority of patients who test negative for anti‐double stranded DNA (dsDNA). 14 During these investigations, we found that patients with abnormal levels of E‐C4d would not necessarily have abnormal levels of P‐C4d and vice versa, suggesting a surprising and intriguing hematopoietic lineage specificity. These observations lead to the hypothesis that lymphocyte‐bound complement activation products (LB‐CAP) may also serve as biomarkers for lupus diagnosis. This hypothesis was addressed in a cross‐sectional study to determine and compare levels of LB‐CAP in patients with SLE, patients with other diseases and healthy individuals.

Patients and Methods

Study participants

All study participants were 18 years of age or older and provided written informed consent. No one was excluded based on gender or ethnicity. Ethnicity was self‐reported by study participants. The University of Pittsburgh Institutional Review Board approved this study.

Patients with SLE who met the ACR 1982 6 or 1997 7 revised classification criteria were recruited for this study during routine visits to the University of Pittsburgh Lupus Patient Care and Translational Research Center. A total of 224 patients were studied from June 2004 through August 2007. As part of their routine care, all patients with SLE underwent routine blood work including complete blood count, erythrocyte sedimentation rate, serum levels of C3 and C4, antinuclear autoantibodies (ANA), and anti‐dsDNA level. Tests for ANA (fluorescent assay) and anti‐dsDNA (fluorescent assay using Crithidia lucillae or enzyme‐linked immunosorbent assay) were performed by certified clinical pathology laboratories. In addition, each of these patients underwent a history and physical examination by a physician (AHK or SM), who was blinded to the LB‐CAP results. Disease activity was assessed at the time of the visit using the Safety of Estrogens in Lupus Erythematosus: National Assessment (SELENA) version of the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) 16 and the Systemic Lupus Activity Measure (SLAM). 17

One hundred seventy‐nine patients with non‐SLE autoimmune or inflammatory diseases, including scleroderma, idiopathic inflammatory myositis, Sjögren's syndrome, rheumatoid arthritis, Wegener's granulomatosis, hepatitis C, vasculitis, primary Raynaud's phenomenon, psoriatic arthritis, osteoarthritis, antiphospholipid syndrome, cutaneous lupus, and undifferentiated connective tissue diseases, were recruited during the period from June 2004 through August 2007. The diagnosis was confirmed by their treating specialist physicians from various outpatient facilities at the University of Pittsburgh Medical Center. A total of 114 healthy individuals were recruited through local advertisements posted around the University of Pittsburgh campus. To confirm their healthy status, participants completed a brief questionnaire regarding existing medical conditions.

Flow cytometric assay for measurements of lymphocyte‐bound complement activation products (LB‐CAP)

At the time of each participant's visit, a 3 mL sample of blood was collected into a Vacutainer® tube containing EDTA as an anticoagulant (Becton Dickinson, Franklin Lakes, NJ, USA). Blood samples were stored at 4°C and analyzed within 24 hours after collection. Peripheral blood leukocytes were isolated using the following protocol. Briefly, the blood sample was centrifuged at 200 ×g; the buffy coat was carefully transferred into a fresh tube and contaminating erythrocytes were hypotonically lysed. The leukocyte suspension was washed extensively with phosphate buffered saline (PBS) to remove lysed erythrocytes, resuspended, and divided into equal‐volume portions and stained for different cell surface markers and CAP. T and B lymphocytes, monocytes, and granulocytes within isolated leukocytes were distinguished based on their unique features of forward/side scattering and expression of characteristic surface molecules. Cells isolated using this protocol exhibited characteristics similar to those isolated using the conventional Ficoll gradient centrifugation method (not shown).

The levels of LB‐CAP, specifically T cell‐bound C4d and C3d (T‐C4d and T‐C3d) and B cell‐bound C4d and C3d (B‐C4d and B‐C3d), were measured using a three‐color flow cytometric assay. Briefly, phycoerythrin‐ or phycoerythrin Cy5‐conjugated mouse monoclonal antibodies (mAb) reactive with lineage‐specific cell surface markers (CD3, CD4, and CD8 for T lymphocytes; CD19 for B lymphocytes; BD Biosciences, San Diego, CA) were used in conjunction with either anti‐human C4d mAb (mouse IgG1; reactive with C4d‐containing fragments of C4; Quidel, San Diego, CA, USA) or anti‐human C3d mAb (mouse IgG1; reactive with C3d‐containing fragments of C3; Quidel) that had been labeled with Alexa Fluor 488 using the Zenon antibody labeling kit (Invitrogen, Carisbad, CA, USA). After staining, cells were analyzed using a FACS Calibur™ flow cytometer and Cell Quest software (Becton Dickinson Immunocytometry Systems). To ensure the specificity of the antibody staining detected, leukocyte aliquots from each patient stained with mouse IgG of appropriate isotypes were routinely included in all experiments. All mAb were used at a concentration of 5 μg/mL. Levels of C3d or C4d on the surface of lymphocytes were expressed as specific median fluorescence intensity (SMFI), which was calculated as the C4d (or C3d)‐specific median fluorescence intensity minus the isotype control median fluorescence intensity. To ensure the day‐to‐day reliability of LB‐CAP measurements, the FACS Calibur flow cytometer was calibrated daily using CaliBrite 3 beads and FACSComp software (Becton Dickinson Immunocytometry Systems). The instrument settings were also calibrated daily using PBL stained with an isotype control IgG labeled with Alexa Fluor 488 to ensure that the background fluorescence intensity remained constantly less than 3.5.

Statistical analysis

Descriptive statistics, including means, standard deviations, medians, and interquartile range (IQR: 25th to 75th percentile), were computed for continuous data. Normality of data distribution was checked using Shapiro‐Wilk test. Differences in the levels of T‐C4d, T‐C3d, B‐C4d, and B‐C3d among the three study groups were compared by Kruskal–Wallis test. Two‐sample Wilcoxon rank‐sum (Mann‐Whitney) test was performed to determine the statistical significance of the differences between each of the paired study groups. Correlations between C4d levels on T cells and B cells as well as between levels of C4d on CD4+ and CD8+ T cells within individual SLE patients were determined using the Spearman's rank correlation technique. Utility of T‐C4d and B‐C4d as diagnostic tests for SLE was assessed using the receiver‐operating characteristic (ROC) analysis. 18 , 19 , 20


Characteristics of the study participants

The study population consisted of 224 patients with SLE, 179 patients with other autoimmune or inflammatory diseases, and 114 healthy controls. Demographic and clinical characteristics of the patients with SLE are summarized in Table 1 . Unless otherwise specified, data shown were ACR criteria‐defined clinical manifestations that had been present at any time point during the course of disease. This cohort included patients with new‐onset as well as long‐standing disease, representing a broad range of disease activity with a wide spectrum of organ involvement. The group of patients with other diseases had a mean age of 46.1 ± 14.9 years (range 18–83), and were 92.2% Caucasian and 86.6% female. The healthy control group had a mean age of 45.5 ± 13.7 years (range 18–80), and were 84.2% Caucasian and 87.7% female.

Table 1
Demographic and clinical characteristics of the SLE patient cohort.

Specific presence of CAP on lymphocytes of SLE patients

To investigate the possibility that CAP are specifically present on the surface of lymphocytes in SLE patients, we conducted a cross‐sectional study utilizing a three‐color flow cytometric assay to measure and compare the levels of CAP deposited on the surface of peripheral blood T and B lymphocytes of SLE patients, patients with other inflammatory/autoimmune diseases, and healthy individuals. Initial studies revealed that C4d and C3d were routinely detectable on the surface of T and B lymphocytes ( Figure 1A ), but C4b, C3b, iC3b, and C5b‐9 were not detectable (data not shown). Therefore, the present study of lymphocyte‐bound CAP (LB‐CAP) was focused on C4d and C3d.

Figure 1
Specific deposition of C4d and C3d on the surface of lymphocytes in patients with SLE. The presence of C4d and C3d on the surface of peripheral blood lymphocytes in patients with SLE, patients with other diseases, and healthy controls was examined using ...

The experimental results obtained from the three study groups demonstrated that significantly elevated levels of C4d and C3d were detected on both T lymphocytes (designated T‐C4d and T‐C3d) and B lymphocytes (designated B‐C4d and B‐C3d) in a significant fraction of SLE patients ( Figure 1B ). In contrast, variable, yet generally low, levels of C4d and C3d were detected on T and B lymphocytes prepared from healthy controls and patients with other diseases. When the LB‐CAP levels were compiled for the entire study population of SLE, other diseases and healthy subjects, the mean ± SD levels of T‐C4d (12.1 ± 20.5; 2.5 ± 3.0; 1.7 ± 1.0), T‐C3d (2.8 ± 3.4; 1.5 ± 1.5; 1.0 ± 0.7), B‐C4d (49.0 ± 73.2; 14.7 ± 26.8; 8.1 ± 5.8), and B‐C3d (17.1 ± 12.4; 13.7 ± 11.7; 8.8 ± 4.8) in patients with SLE were significantly higher than those in healthy controls (all p < 0.0001) as well as those in patients with other diseases (p < 0.0001) ( Table 2A ). Notably, levels of B‐C4d and B‐C3d were considerably higher than those of T‐C4d and T‐C3d in all three study groups. Even though T‐C3d and B‐C3d levels (2.8 ± 3.4 and 17.1 ± 12.4, respectively) were significantly elevated in patients with SLE, they were noticeably lower than those of T‐C4d and B‐C4d (12.1 ± 20.5 and 49.0 ± 73.2, respectively). Subsequent studies of the utility of LB‐CAP as SLE biomarkers, therefore, were concentrated on T‐C4d and B‐C4d.

Table 2
Comparison of lymphocyte‐bound CAP levels in patients with SLE, patients with other diseases, and healthy controls.

Initial studies showed that lymphocyte‐bound C4d and C3d measures were not influenced by lymphocyte count in SLE patients (data not shown). Initial studies also indicated that the binding of C4d and C3d to lymphocytes was stable, as repeated measurements of the same blood sample stored at 4°C showed insignificant variation over a 3‐day period (data not shown). This was supported by the observation that exposure to acidic or low‐ionic strength buffers did not influence LB‐CAP levels (data not shown).

Comparison of the LB‐CAP phenotypes of SLE lymphocytes

We next examined the relationships among the LB‐CAP phenotypes of different lymphocyte populations within each blood sample. Several observations were made. First, within each of the four specific assays, the phenotype of the T and/or B cells was homogeneous among all cells in a given patient. Histograms rarely contained shoulders or multiple peaks (representing subsets of cells with distinctly different levels of C4d and/or C3d), but rather were uniformly positive as demonstrated by complete shift of the fluorescent intensity histograms ( Figure 1A ). Second, in general, levels of T‐C4d and B‐C4d within a given patient were significantly correlated (Spearman's rank coefficient (rs) = 0.747, p < 0.0001) ( Figure 1C ). However, discordant levels were observed in some patients, who had elevated B‐C4d levels but low or even normal T‐C4d levels ( Figure 1C , red‐box).

This cellular specificity for LB‐CAP phenotype was further investigated with studies of CD4 and CD8 T lymphocyte subsets. As shown in Table 2B , C4d levels of CD4 T cells (CD4 T‐C4d) were significantly elevated in patients with SLE, compared to those of other disease, and healthy subjects (both p < 0.0001). Likewise, C4d levels of CD8 T cells (CD8 T‐C4d) were also significantly elevated in the SLE group compared to those of the other two groups (both p < 0.0001). Levels of C3d on CD4 T cells (CD4 T‐C3d) and CD8 T cells (CD8 T‐C3d) were lower when compared to CD4 T‐C4d and CD8 T‐C4d, but still differed significantly between SLE patients and patients with other diseases (p= 0.0006; p= 0.002) and between SLE patients and healthy individuals (p < 0.0001; p= 0.02), respectively. In general, CD4 T‐C4d and CD8 T‐C4d were significantly correlated within a given patient (r= 0.743, p < 0.0001) ( Figure 1D ). Although in most SLE patients, similar levels of C4d and C3d were detected simultaneously on CD4 and CD8 T cells, we were able to identify a fraction of patients in which distinct levels of C4d and C3d were present on CD4 versus CD8 T cells ( Figure 1D ; data shown only for C4d).

Sensitivity and specificity of T‐C4d and B‐C4d measures for SLE diagnosis

The statistically significant and specific elevation of LB‐CAP, particularly T‐C4d and B‐C4d, in patients with SLE suggests that these measures may serve as unique biomarkers for SLE and may be of added value to current tests for earlier and more accurate diagnosis of SLE. To explore this possibility, diagnostic sensitivity and specificity of T‐C4d and B‐C4d were first calculated at different cutoff levels using patients with SLE as the diagnostic group and patients with other diseases as the reference group. Diagnostic performance of T‐C4d and B‐C4d in distinguishing patients with SLE from patients with other inflammatory or autoimmune diseases was further assessed using the ROC analysis. 18 The ROC curve was constructed by measuring the sensitivity (true positive) and 1‐specificity (false positive) across various cutoff values that define the positivity of T‐C4d or B‐C4d. The differentiating power of T‐C4d and B‐C4d was then estimated using the area under the ROC curve (AUC). 19 , 20 As shown in Figure 2 , the AUC for the T‐C4d assay and the B‐C4d assay were 0.727 and 0.770, respectively. Based on these data, it was estimated that T‐C4d and B‐C4d are 56% sensitive/80% specific and 60% sensitive/82% specific in differentiating SLE from other diseases, respectively.

Figure 2
ROC‐AUC analysis of the performance of T‐C4d and B‐C4d levels as diagnostic markers for SLE.

Comparison of T‐C4d and B‐C4d with other laboratory tests for SLE diagnosis

We next sought to determine whether T‐C4d and B‐C4d would provide added value to ANA and anti‐dsDNA, laboratory tests conventionally used for SLE diagnosis. Our SLE study cohort consisted of patients with a broad range of disease activity (SLAM and SLEDAI scores ranging from 0 to 21 and 0 to 14, respectively) and disease duration (0–40 years). Given the insidious and heterogeneous course of SLE, a laboratory test that is capable of facilitating the diagnosis during a single clinic visit would be extremely useful. The capability of T‐C4d, B‐C4d, and anti‐dsDNA measures, when used separately or in combination, for identifying SLE patients at the time of the study visit was investigated and compared using this study cohort ( Figure 3 ). At the time of the study visit, 223 of 224 SLE patients (99.6%) were positive for ANA. The analysis was performed with two different decision trees. First, ANA test results were followed by T‐C4d and B‐C4d assays and then with the anti‐dsDNA test ( Figure 3A ). Of the 223 ANA‐positive patients, 141 (63%) patients had abnormally elevated levels of T‐C4d and/or B‐C4d (cut points defined as: the mean level + 2 SD of the healthy control group; 2.71 for T‐C4d and 19.6 for B‐C4d; ref. Table 2 ) ( Figure 3A ). Moreover, the anti‐dsDNA test captured only 9 of the 82 ANA positive/T‐C4d/B‐C4d‐negative patients. The lone ANA‐negative patient tested negative for both anti‐dsDNA and T‐C4d/B‐C4d.

Figure 3Figure 3
Comparison of the capacity of T‐C4d/B‐C4d, anti‐dsDNA, serum C3, and serum C4 measures for SLE diagnosis. Tests for T‐C4d/B‐C4d, anti‐dsDNA, serum C3, and serum C4 were performed on samples obtained on ...

In the second decision tree, ANA‐tested patients were followed by the anti‐dsDNA test and then with T‐C4d and B‐C4d assays ( Figure 3B ). Of the 223 ANA‐positive patients, anti‐dsDNA test was performed on 209. Only 59 of these 209 ANA‐positive patients (28%) exhibited positive anti‐dsDNA at the study visit (compared with 63% positive for T‐C4d and/or B‐C4d). Most patients who were anti‐dsDNA‐positive also tested positive for T‐C4d and/or B‐C4d (50/59; 85%). Of the 150 patients who were ANA‐positive but anti‐dsDNA‐negative, the majority (83; 55%) were positive for T‐C4d/B‐C4d.

In the third and fourth decision trees ( Figure 3C and D ), we compared T‐C4d and B‐C4d with the traditional measures of complement activation—serum C3 and C4. The results indicated that of 221 patients who were tested for serum C3 and C4, only 67 (30%) and 82 (37%) had abnormally low levels of serum C3 and C4, respectively. Most patients with abnormal levels of serum C3 and C4 also tested positive for T‐C4d/B‐C4d (56/67; 84% and 67/82; 82%, respectively). In addition, the majority (82/154; 53%) of patients who had normal serum C3 levels were positive for T‐C4d/B‐C4d, and the majority (71/139; 51%) of patients with normal serum C4 levels were positive for T‐C4d/B‐C4d. All three patients who were not tested for serum C3 and C4 were also positive for T‐C4d/B‐C4d.

Collectively, these results obtained using the four different decision trees indicate that: (1) when used independently, the T‐C4d/B‐C4d assay alone is more sensitive than the anti‐dsDNA test alone, serum C3 measure alone, or serum C4 measure alone in confirming a diagnosis of SLE, and (2) by combining these assays, the T‐C4d/B‐C4d assays identify SLE patients who are negative for anti‐dsDNA and/or have normall serum C3/C4 levels at a single visit.


The current study was undertaken to investigate the potential for LB‐CAP to serve as biomarkers for SLE diagnosis. The results indicate that T‐C4d and B‐C4d are highly sensitive and specific biomarkers for diagnoses of SLE, T‐C4d and B‐C4d may be more sensitive biomarkers for SLE diagnosis as compared with anti‐dsDNA at one clinic visit, and simultaneous determination of anti‐dsDNA and T‐C4d/B‐C4d may provide the highest combined sensitivity‐specificity for SLE diagnosis reported to date.

There are several strengths of this study design. First, it is prospective, dictated by the need for fresh blood samples to perform LB‐CAP assays. Second, the SLE patient cohort studied spanned a broad range of ages, disease activity and clinical manifestations, as the design was to enroll consecutive patients simply based upon willingness to participate. Third, all patients were evaluated at the same site by the same lupologists, ensuring that the diagnosis was as accurate as possible. Fourth, assays were performed on the same flow cytometer by the same operator, and calibrations and standards were established prior to the study. Fifth, the other disease control group was enriched for rheumatologic and autoimmune diseases that are frequently misdiagnosed as SLE, to ensure a most stringent test and the worst‐case scenario for ROC analysis. Sixth, we performed a head‐to‐head comparison between our candidate biomarker(s) and the current gold standards for SLE, ANA, anti‐dsDNA and serum C3/ C4, at a single visit to emulate a real practice scenario.

The results of this study demonstrate significant potential for LB‐CAP as diagnostic biomarkers for SLE. T‐C4d and B‐C4d were 56% sensitive/80% specific and 60% sensitive/82% specific for the diagnosis of SLE, respectively. The sensitivity of ANA in this study was 99.6%. However, it is generally held that ANA‐positivity is highly nonspecific for SLE with a positive‐predictive value as low as 11% in some studies. 21 , 22 Conversely, anti‐dsDNA testing has been shown to be highly specific for confirming a diagnosis of SLE, but the mean sensitivity of anti‐dsDNA testing for SLE among published studies is only 57%. 22 Although 60% of the patients with SLE in this study were anti‐dsDNA‐positive at some point during the course of their disease, the sensitivity of the test at one visit in the current study was only 29% ( Table 1 ). Another laboratory measure commonly used by physicians to aid in diagnosing SLE, despite not being among the ACR classification criteria, is abnormally low levels of serum C3 and/or C4. In the current study, the sensitivity of serum C3 and C4 measurements at one visit was only 30% (67 of 221 patients tested) and 37% (82 of 221 patients), respectively ( Table 1 ). Thus, based upon evaluation at a single visit in this study, T‐C4d and B‐C4d can be considered more specific than ANA and more sensitive than anti‐dsDNA, serum C3, and serum C4.

Most importantly, LB‐CAP testing was diagnostic for SLE in 55% (83/150) of patients who were negative for anti‐dsDNA ( Figure 3B ), 53% (82/154) of patients who had normal serum C3 levels ( Figure 3C ), and 51% (71/139) of patients who had normal serum C4 levels ( Figure 3D ). It is also important to note that the sensitivity and specificity of the LB‐CAP assay are determined by the points on the ROC that are selected. Thus, if LB‐CAP are included in a diagnostic panel with other biomarkers, the cutoff values can be selected based upon the sensitivity and specificity of the other components in the panel and the value desired from the LB‐CAP assays. For example, a B‐C4d cutoff value could be selected to provide nearly 100% specificity with 20% sensitivity based upon the ROC analysis.

The multi‐ethnic cohort of SLE patients studied in the current study consisted predominantly of Caucasians (83.5%; 13.8% Africa Americans; 2.7% Others; Table 1 ). However, preliminary analyses have indicated that a higher frequency of non‐Caucasian SLE patients (28/37; 77%) tested positive for T‐C4d/B‐C4d than did Caucasian SLE patients (113/187; 60%) at a single study visit. This result suggests that T‐C4d/B‐C4d as a diagnostic biomarker may be as sensitive, if not more sensitive, in non‐Caucasian patients as in Caucasian patients. The impact of ethnicity on T‐C4d/B‐C4d as diagnostic lupus biomarkers warrants further investigation.

In addition to demonstrating utility as diagnostic biomarkers, this study generated data that may suggest clues to the mechanisms responsible for generation of LB‐CAP. Although we had previously noted that E‐C4d, 14 R‐C4d, 23 and P‐C4d 15 phenotypes do not necessarily correlate within a given patient, it was still surprising to discover that different LB‐CAP assays such as T‐C4d and B‐C4d levels do not correlate in the same blood sample prepared from some SLE patients ( Figure 1C ). Even more surprising was the observation that levels of C4d and C3d on T‐cell subsets do not necessarily correlate in a given patient at one point in time ( Figure 1D ). These observations further support the hypothesis that there is a tissue‐specific mechanism responsible for generation of cell‐bound complement activation products (CB‐CAP), and elevations of CB‐CAP are not simply the result of systemic activation of the complement system with nonspecific deposition on circulating cells. The findings also suggest that a panel composed of multiple CB‐CAP assays will have the greatest diagnostic value, with a combined sensitivity and specificity greater than any single assay can provide. Further investigation may also produce insight into whether patterns of CAP deposition have a clinical subsetting potential.

Another potential insight into the mechanism of LB‐CAP generation can be gleaned by comparison of the relative levels of C4d and C3d observed on each cell and cell type. Although C4d and C3d were generally detected simultaneously on lymphocyte surfaces ( Figure 1A and B ), the levels of C4d were generally higher than those of C3d ( Figure 1B and Table 2 ). Because C3 activation is more distal in the enzymatic amplification cascade of complement activation, we initially considered that C3d would be a more sensitive LB‐CAP biomarker. However, the unexpected “C4d > C3d” phenomenon was consistently observed for both T and B lymphocytes in study participants from all three study groups ( Table 2 ). Moreover, in spite of high T‐C4d and B‐C4d levels detected in SLE patients, CAP of the terminal step of complement activation, such as the C5b‐9 membrane attack complex, were not detectable (data not shown). These results not only indicate that CAP present on the surface of lymphocytes of SLE patients are products of the classical, and possibly the lectin, pathway of complement activation, but also suggest that specific regulatory mechanisms are in place to harness the activation cascade to prevent generation of C3‐ and C5‐convertases and overt lysis of lymphocytes. The attachment of nonlytic forms/levels of CAP to surface molecules of lymphocytes may potentially alter the physiological functions of lymphocytes and lead to significant consequences in (auto) immune responses in SLE patients.

In summary, this is the first report of lymphocytes bearing complement activation products in human health and disease. The results suggest strong potential for B‐C4d and T‐C4d as diagnostic biomarkers for SLE and provide clues to potential pathogenic mechanisms. Future investigations including multicenter validation are warranted.


This study was supported by grants from the National Institutes of Health (RO1 AI‐077591, RO1 HL‐074335, K23 AR‐051044, and K24 AR‐02213), Department of Defense (Peer Reviewed Medical Research Grant W81XWH‐06–2‐0038), the Lupus Research Institute, and Cellatope, Inc. The authors gratefully acknowledge the following colleagues for providing patient blood samples and clinical information for this study: Dr. Kathleen McKinnon, Dr. Dana Ascherman, Dr. Brian Berk, Dr. Thomas Medsger, Dr. Chester Oddis, Dr. William Ridgway, and Dr. Mary Chester Wasko.


1. Sherer Y, Gorstein A, Fritzler MJ, Shoenfeld Y. Autoantibody explosion in systemic lupus erythematosus: more than 100 different antibodies found in SLE patients.[see comment]. Semin Arthritis Rheum. 2004; 34(2): 501–537. [PubMed]
2. D’Cruz DP, Khamashta MA, Hughes GR. Systemic lupus erythematosus.[see comment]. Lancet. 2007; 369(9561): 587–596. [PubMed]
3. Rahman A, Isenberg DA. Systemic lupus erythematosus. N Eng J Med 2008; 358(9): 929–939. [PubMed]
4. Gamez‐Nava JI, Gonzalez‐Lopez L, Davis P, Suarez‐Almazor ME. Referral and diagnosis of common rheumatic diseases by primary care physicians. Br J Rheumotoi. 1998; 37(11): 1215–1219. [PubMed]
5. Narain S, Richards HB, Satoh M, Sarmiento M, Davidson R, Shuster J, Sobel E, Hahn P, Reeves WH. Diagnostic accuracy for lupus and other systemic autoimmune diseases in the community setting. [see comment]. Arch intern Med 2004; 164(22): 2435–2441. [PubMed]
6. Tan EM, Cohen AS, Fries JF. The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum. 1982; 25: 1271–1277. [PubMed]
7. Hochberg MC. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum. 1997; 40: 1725. [PubMed]
8. Schur PH, Sandson J. Immunological factors and clinical activity in systemic lupus erythematosus. N Engl J Med. 1968; 278: 533–538. [PubMed]
9. Valentijn RM, Van Overhagen H, Hazevoet HM, Hermans J, Cats A, Daha MR, Van Es LA. The value of complement and immune complex determinations in monitoring disease activity in patients with systemic lupus erythematosus. Arthritis Rheum. 1985; 28: 904–913. [PubMed]
10. Sturfelt G, Jonasson L, Sjoholm AG. Sequential studies of complement activation in systemic lupus erythematosus. Scand J Rheumatoi. 1985; 14: 184–196. [PubMed]
11. Porcel JM, Ordi J, Castro‐Salomo A, Vilardell M, Rodrigo MJ, Gene T, Warburton F, Kraus M, Vergani D. The value of complement activation products in the assessment of systmeic lupus erythematosus flares. Gin immunoi Immunopathoi. 1995; 74: 283–288. [PubMed]
12. Esdaile JM, Abrahamowicz M, Joseph L, MacKenzie T, Li Y, Danoff D. Laboratory tests as predictors of disease exacerbations in systemic lupus erythematosus: why some tests fail . Arthritis Rheum. 1996; 39: 370–378. [PubMed]
13. Liu C‐C, Ahearn JM, Manzi S. Complement as a source of biomarkers in systemic lupus erythematosus: past, present, and future. Curr Rheumotoi Rep. 2004; 6: 85–88. [PubMed]
14. Manzi S, Navratil JS, Ruffing MJ, Liu C‐C, Danchenko N, Nilson SE, Krishnaswami S, King DES, Kao AH, Ahearn JM. Measurement of erythrocyte C4d and complement receptor 1 in the diagnosis of systemic lupus erythematosus. Arthritis Rheum. 2004; 50: 3596–3604. [PubMed]
15. Navratil JS, Manzi S, Kao AH, Krishnaswami S, Liu CC, Ruffing MJ, Shaw PS, Nilson AC, Dryden ER, Johnson JJ, Ahearn JM. Platelet C4d is highly specific for systemic lupus erythematosus. Arthritis Rheum. 2006; 54(2): 670–674. [PubMed]
16. Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang CH, Austin A, Bell A, Bloch DA, Corey PN, Decker JL, Esdaile J, Fries JF, Ginzler EM, Goldsmith CH, Hochberg MC, Jones JV, Le Riche NGH, Liang MH, Lockshin MD, Muenz LR, Sackett DL, Schur PH. Derivation of the SLEDAI. A disease activity index for lupus patients. The Committee on Prognosis Studies in SLE. Arthritis Rheum. 1992; 35: 630–640. [PubMed]
17. Liang MH, Socher SA, Larson MG, Schur PH. Reliability and validity of six systems for the clinical assessment of disease activity in systemic lupus erythematosus. Arthritis Rheum. 1989; 32: 1107–1118. [PubMed]
18. Zweig MH, Campbell G. Receiver‐operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine.[erratum appears in Clin Chem 1993 Aug; 39(8): 1589]. Gin Chem. 1993; 39(4): 561–577. [PubMed]
19. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982; 143(1): 29–36. [PubMed]
20. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology. 1983; 148(3): 839–843. [PubMed]
21. Slater CA, Davis RB, Shmerling RH. Antinuclear antibody testing. A study of clinical utility. Arch Intern Med. 1996; 156(13): 1421–1425. [PubMed]
22. Solomon DH, Kavanaugh AJ, Schur PH. American College of Rheumatology Ad Hoc Committee on Immunologic Testing G. Evidence‐based guidelines for the use of immunologic tests: antinuclear antibody testing. [see comment]. Arthritis Rheum. 2002; 47(4): 434–444. [PubMed]
23. Liu CC, Manzi S, Kao AH, Ruffing MJ, Navratil JS, Ahearn JM. Reticulocytes bearing C4d as biomarkers of disease activity for systemic lupus eryhematosus. Arthritis Rheum. 2005; 52: 3087–3099. [PubMed]

Articles from Clinical and Translational Science are provided here courtesy of Wiley-Blackwell