Detection of low level cryoglobulins by flow cytometry

Authors

  • Rüdiger B. Müller,

    1. Department of Internal Medicine 3, University of Erlangen-Nürnberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany
    2. Department of Rheumatology, Kantonsspital St. Gallen, Rorschacherstr. 95, 9007 St. Gallen, Switzerland
    Search for more papers by this author
  • Birgit Vogt,

    1. Department of Internal Medicine 3, University of Erlangen-Nürnberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany
    2. Department of Nephrology and Internal Intensive Care Medicine, Campus Virchow Klinikum, Charité Berlin, Augustenburger Platz 1, D-13353 Berlin, Germany
    Search for more papers by this author
  • Silke Winkler,

    1. Department of Internal Medicine 3, University of Erlangen-Nürnberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany
    Search for more papers by this author
  • Luis E. Muñoz,

    1. Department of Internal Medicine 3, University of Erlangen-Nürnberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany
    Search for more papers by this author
  • Sandra Franz,

    1. Department of Internal Medicine 3, University of Erlangen-Nürnberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany
    Search for more papers by this author
  • Peter Kern,

    1. Department of Internal Medicine 4, Rheumatology, Immunology, and Osteology, Fulda Medical Centre, Fulda, Germany
    Search for more papers by this author
  • Christian Maihöfner,

    1. Department of Neurology, Friedrich-Alexander University of Erlangen-Nürnberg, 91054 Erlangen, Germany
    Search for more papers by this author
  • Ahmed Sheriff,

    1. Department of Internal Medicine 3, University of Erlangen-Nürnberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany
    2. Department of Nephrology and Internal Intensive Care Medicine, Campus Virchow Klinikum, Charité Berlin, Augustenburger Platz 1, D-13353 Berlin, Germany
    Search for more papers by this author
  • Johannes von Kempis,

    1. Department of Rheumatology, Kantonsspital St. Gallen, Rorschacherstr. 95, 9007 St. Gallen, Switzerland
    Search for more papers by this author
  • Georg Schett,

    1. Department of Internal Medicine 3, University of Erlangen-Nürnberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany
    Search for more papers by this author
  • Martin Herrmann

    Corresponding author
    1. Department of Internal Medicine 3, University of Erlangen-Nürnberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany
    • Department of Internal Medicine 3, Rheumatology and Immunology, University of Erlangen-Nürnberg, Kranken-hausstrasse 12, D-91054 Erlangen, Germany
    Search for more papers by this author

Abstract

Several patients with cryoglobulin (CG) associated symptoms are seronegative for CG and other potentially causative biomarkers. We analyzed whether it is possible to detect cryoprecipitates by flow cytometry and whether the sensitivity of their demonstration can be increased as compared to visual inspection. Sera from 91 patients with suspected CG associated symptoms and 33 healthy controls were examined for the presence of CG by conventional visual testing and by flow cytometry for small diffracting particles. For calibration purposes we tested lipid micelle dilutions (positive controls) by both methods. The minimum concentrations of lipid micelles to be detected by visual inspection and flow cytometry were 128.5 and 2.0 pg ml−1, respectively. Among the 91 patients and 33 controls, only 1 patient serum was positive for CG by conventional testing. This sample was also positive on flow cytometry. In the serum of a patient known to be positive for CG, laser diffracting particles were quantified by flow cytometry after keeping serum at 4°C for 3 days. Of the 91 patients, 14 additional samples displayed cold precipitates which redissolved after rewarming during flow cytometry. All 15 (1 + 14) patients positive for CG on flow cytometry suffered from symptoms usually associated with CG. Some precipitates were labeled with anti IgG and IgM antibodies confirming that the particles detected by flow cytometry contained immunoglobulins. No small diffracting particles were detected in the sera of the 33 healthy controls. Flow cytometry is equally specific but much more sensitive in the detection of CG than visual inspection. © 2012 International Society for Advancement of Cytometry

Cryoglobulins (CG) are associated with various diseases, such as immunoproliferative disorders (e.g., multiple myeloma) and several autoimmune diseases (e.g., vasculitis) (1). Especially during hepatitis C virus (HCV) infection, CG are associated with a variety of extrahepatic manifestations including vasculitis, glomerulonephritis, arthritis, and peripheral neuropathy (2, 3). CG are serum immunoglobulins that precipitate if the temperature drops substantially below body temperature. CG dissolves during rewarming. After in vivo precipitation of CG in small vessels, neutrophils and mononuclear cells home to these sites and lead to the induction of inflammation of small vessels. Consequently, a sensitive detection of CG is an important contribution to the diagnostic procedures for vasculitis.

Today, the detection of CG is performed by visual inspection of precipitates formed in cooled serum as compared with serum kept at 37°C. Redissolving of the precipitate after rewarming to 37°C excludes unspecific aggregates and confirms the CG nature of the precipitates (4). Interestingly, it has been reported that CG can also be detected in blood smears (5, 6). Two major disadvantages of the visual inspection method are their low detection rate and its semiquantitative nature. Alternatively, CG can be quantified by the method of Hauke et al. employing cryocrit tubes (7). Higher sensitivity and differentiation of the precipitate is reached by application of a gel diffusion procedure, the Ouchterlony method (8), with detectable levels at 50 mg l−1 (9). On the basis of this rather high detection threshold, we suspected that low CG levels may escape detection by the current diagnostic methods.

Weiner et al. showed that the capacity of CG to activate complement was rather pathognomonic than on the CG level (3). Recently, new therapeutic options such as B cell depletion by Rituximab (10, 11) have been introduced as treatment of CG with promising results. As the detection limits of CG currently seems to be rather high, a more sensitive method identifying CG in more patients may extend the diagnostic armamentarium in chronic inflammatory diseases associated with CG.

Material and Methods

Patients

Ninety one consecutive samples of patients routinely tested for CG and of 33 healthy subjects were collected. In 20 of these 91 patients (non CG patients) another not CG associated diagnosis was later established. Thus, CG was only a differential diagnosis and a not CG associated diagnosis for the symptoms could be established.

Patient's characteristics are depicted in Table 1. The University of Erlangen-Nürnberg Institutional Review Board approved the study and written informed consent was obtained from all individuals before they entered the study. The study did not affect the diagnostic routine.

Table 1. Patient data
 NumberAge (years)Sex (f/m)
All patients9150.762/29
Neoplastic disease650.86/0
Autoimmune disease5051.333/17
Polyneuropathy1551.69/6
Chronic infections747.95/2
Not Cg associated2053.014/6

Serum Tests for Cryoglobulines (CG)

Blood was collected in prewarmed serum tubes (37°C, Sarstedt, Nümbrecht, Germany), immediately stored at 37°C and sera were obtained by centrifugation (2000g) at 37°C. (I) For routine testing for CG, two identical serum samples were incubated for 3 days at 4 and 37°C, respectively. These samples were visually inspected and considered positive for CG if more visible precipitate was formed in the 4°C tube when compared to the 37°C one and if the 4°C precipitate resolved after incubation for 24 h at 37°C. When the precipitate resolves within 24 h, a reduction of the particles can already be observed as early as 30 min after warming of the samples. If it was ambiguous whether a precipitate could be detected a centrifugation at 4°C, 2000G over 10 min was performed to better detect potential cryoprecipitates. (II) For the alternative method, 200 μl serum were added to 1800 μl PBS prewarmed to 37°C, split into two tubes and incubated for 3 days at 4 or 37°C. After quantifying the particles in the tubes employing a flow cytometer (EPICS, Beckman Coulter, Cambridge, MA), the 4 and the 37°C tubes were further incubated at 37 and 4°C, respectively. Twelve to twenty-four hours later the tubes were reassessed by flow cytometry (Supporting Information). For some samples, 2 μl of fluorochrome-labeled antibodies against human IgG and IgM (Pharmingen) were added to the diluted CG samples (30 min at ambient temperature) and quantified by flow cytometry, as described (12).

Lipid Micelles

Two-fold serial dilutions of lipid micelles ranging from 4.5 μg l−1 to 0.4 pg l−1 were assessed visually and by flow cytometry (EPICS, Beckman Coulter, Cambridge, MA) for detectable particle counts. The detection limits for visual inspection and flow cytometry were determined (Supporting Information).

Results

Determination of the Detection Levels for Micelles by the Classical Visual and the New Flow Cytometry Method

CG are usually detected by visual inspection of serum stored at 4°C for up to 3 days and resolving the precipitate after rewarming to 37°C. As this method is dependent on the individual visual judgement, differences in the results depend on the assessor. We therefore addressed the question whether precipitates can more objectively be analyzed by flow cytometry than by visual inspection. First, we analyzed the detection limit for lipid micelles with both methods. The minimal concentrations of lipid micelle preparations detectable by visual inspection and flow cytometry were 128.5 and 2.0 pg ml−1, respectively. The detection level of small lipid micelles employing flow cytometry thus was 62.5 times more sensitive than the conventional visual method. The particle counts quantified by flow cytometry were between 4 and 60 counts/20 s for PBS alone and 200–10,000 counts/20 s for lipid micelles in PBS, depending on their concentration. Higher dilution was associated with lower particle counts in good linear correlation (Fig. 1). In conclusion, small particles can reliably be quantified by flow cytometry.

Figure 1.

Dilutions of lipid micelles analyzed by flow cytometry. Lipid micelles were diluted in PBS in different concentrations 5.8 × 10−8 to 1.5 × 10−4. Samples were assessed by flow cytometry for the presence of laser diffracting particles for 20 s (A), by spectrophotometry at 280 nm (B), and by visual inspection (vertical dotted lines; left: visually detectable, right: not detectable by the unarmed eye). The concentration of lipid micelles (%) is depicted as negative logarithm. Mean values of lipid droplets measured by flow cytometric events (A) and optical density (B) with their standard deviations of four independent experiments are depicted by horizontal lines for each dilution. Lipid suspensions containing more than 3.8 × 10−6% lipid micelles showed a visible turbidity (left from the vertical dotted lines). Particles of fetal calf serum (FCS) are displayed also as mean values of flow cytometric events of four independent experiments. Number of particles (A) and optical density (B) of PBS are depicted as dashed lines, respectively.

Predictive Testing of Patients with Suspected CG

Because flow cytometry displayed a lower detection limit for lipid micelles when compared with visual inspection (Fig. 1), we addressed the question whether it can also be employed for the quantification of CG in serum samples. A cohort of 91 patients with suspected CG associated symptoms was assessed by flow cytometry in parallel to the conventional visual testing (Fig. 2). Serum of a patient with CG detected by visual inspection served as positive control.

Figure 2.

Detection of laser diffracting particles after incubation of a CG positive (visual inspection) serum sample at 37°C (A) and 4°C (B). Dot plot obtained after incubation of the 4°C sample for 15 min and 12 h at 37°C are shown in the panels C and D, respectively.

Only 1 patient of these 91 was confirmed positive for CG by the conventional method. Laser-diffracting particles, which redissolved after rewarming, were also detected in this patient by flow cytometry (asterisk in Fig. 3).

Figure 3.

Detection of CG in human serum samples by flow cytometry: 100 μl serum was added to 900 μl PBS, incubated for 3 days in two fractions at 37 or 4°C. After 3 days, precipitated molecules were assessed by flow cytometry. Particle counts are depicted for patients with autoimmune diseases, chronic infections, polyneuropathy, neoplasm with symptoms often associated with CG and for patients with no CG associated symptoms. Particle counts are depicted as the ratio of laser diffracting particle counts after 3 days at 4°C (cold; C) and 37°C (warm; W), respectively. Sera with a C/W ratio > 4 (= particle count 4°C/particle count 37°C) were considered positive and are represented by black circles. Abbreviations: s, seconds; AI, autoimmune diseases; PNP, polyneuropathy; Neopl, Neoplastic diseases; no CG, not cryoglobulin associated diseases; NHD, normal healthy donors.

After analysis by flow cytometry of the patients' sera kept at 4°C (further referred to as cold) or 37°C for 72 h (warm) by flow cytometry, the ratio of cold/warm (C/W) particle counts was calculated. The experiments were performed in duplicates. A C/W ratio above 4 was considered positive. Nineteen samples showed ratios in the range between 4.3 and 91 (Fig. 3) and were considered positive for CG. In all sera which displayed laser diffracting particles, the latter were reduced after rewarming (average reduction over 24 h—47.8%).

Consequently, the method based on flow cytometry detected 14 more patients with laser diffracting particles in a cohort of 91 patients clinically suspected to suffer from CG associated symptoms than with the conventional visual method. Of this total of 15 patients (14 + 1 also conventionally detected patient) 13 of these in total 15 patients suffered from autoimmune disease (ANCA associated vasculitis, 3× Raynaud'phenomenon, noninfectious papillitis, systemic lupus erythematosus, antiphospholipid syndrome, arthritis, 2× Sjogren's syndrome, cerebral vasculitis), 1 from neoplasm, 3 from chronic infections (HIV, HBV, HCV), and 1 from inflammatory polyneuropathy. Some patients suffered from symptoms associated with two diseases, e.g., polyneuropathy and autoimmune diseases. Consequently, more associated diseases than patients are listed.

To demonstrate whether the new by flow cytometry detected laser defracting structures contain antibodies, we performed staining with anti IgG or IgM antibodies of random positive samples. In all samples (n = 3), positive antibody staining for IgG, IgM or both was detected (not shown).

Discussion

As demonstrated by the analysis of serial dilutions of lipid micelles as surrogate for cryoglobulins, flow cytometry was at least 50 times more sensitive than visual assessment. Lipid micelles could be detected even in a concentration as low as 2 versus 128 ng l−1 by visual inspection. The gel diffusion procedure (8) used for the detection of CG has a threshold of 50 mg l−1. After demonstrating the improved sensitivity of flow cytometry for the detection of laser diffracting particles such as lipid micelles, we hypothesized that flow cytometry may also be more sensitive for the detection of CG than the conventional methods. In a cohort of 91 patients with suspected CG associated symptoms where the visual inspection only identified a single CG positive sample, the flow cytometry detected CG in 14 additional sera. We therefore conclude that flow cytometry is much more sensitive than the conventional visual method for the analysis of CG.

Detection by flow cytometry of CG was sensitive and specific (Tables 2 and 3). First, flow cytometry detected CG in the only serum positive by the conventional method. Second, all of 15 CG positive sera were derived from patients suffering from autoimmune diseases potentially associated with low-level CG. Third, no positive results were found in a cohort of 33 healthy controls. Fourth, only 1 of 20 disease controls tested (weakly) positive for CG. This patient suffered from a cerebral insult. The primary hypothesized vasculitic origin, leading to testing Cg could not be confirmed.

Table 2. Sensitivity of cryoglobuline (CG) detection in patients with CG-related symptoms (n = 78)
 VisualFlow cytometry
  • Note that the detection with flow cytometry is much more sensitive than the visual inspection.

  • a

    The true sensitivity may be much higher since not all sera from patients with CG-related symptoms will indeed have detectable CG.

Sensitivity in % (CI)1.3a (0.0–4.4)23.1a (13.1–33.1)
Table 3. Specificity of CG detection by flow cytometry
 CG versus nonCGCG versus NHD
  1. Patients with CG-related symptoms (CG; n = 78) were compared to patients without CG-related symptoms (nonCG; n = 20) and normal healthy donors (NHD; n = 33).

Specificity in % (CI)95.0 (83.0–100.0)100.0 (98.5–100.0)

Even though the calculated sensitivity of 23.1% appears to be low it is 17.77 times higher than calculated with the conventional visual method. The true sensitivity may be much higher since not all sera from patients with CG-related symptoms will indeed have detectable CG.

Recently, Shihabi et al. stated that in comparison to the high incidence of cryoglobulins in diseases such as hepatitis C virus (HCV) infection it is clear that testing for CG is otherwise underutilized in clinical practice (13). CG testing is often neglected by clinicians because of the long duration of the testing procedure and the rare positive results. Low sensitivity of assays for CG is a relevant diagnostic and clinical problem since CG play an important role in the pathogenesis of chronic conditions such as autoimmune and lymphoproliferative diseases. They may also be involved in the immune complex mediated pathologies of systemic autoimmunity (14). The time frame of the method described in this manuscript may even be shortened further since we observed a detectable (by flow cytometry) cryoprecipitate formation as soon as 6 h after assay start in selected samples (not shown).

In conclusion, the described method for detection of CG has a higher sensitivity compared to the old standard assays visual method and allows better quantification of CG. It represents a reliable method to detect low levels of CG and to unravel occult CG-associated pathologies.

Ancillary