Journal of Neuroimmunology 273 (2014) 103–110
Contents lists available at ScienceDirect
Journal of Neuroimmunology journal homepage: www.elsevier.com/locate/jneuroim
Development of a cell-based assay for the detection of anti-aquaporin 1 antibodies in neuromyelitis optica spectrum disorders Youming Long a,b, Yangbo Zheng a,b, Fulan Shan a,b, Mengyu Chen a,b, Yongxiang Fan a,b, Bin Zhang a,b, Cong Gao a,b,⁎, Qingchun Gao a,b, Ning Yang c a Key Laboratory of Neurogenetics and Channelopathies of Guangdong Province and The Ministry of Education of China, Institute of Neuroscience and the Second Affiliated Hospital of GuangZhou Medical University, 250# Changgang East Road, GuangZhou, 510260 Guangdong Province, China b Department of Neurology, The Second Affiliated Hospital of GuangZhou Medical University, 250# Changgang East Road, GuangZhou, 510260 Guangdong Province, China c Department of Neurology, The Fifth Affiliated Hospital of GuangZhou Medical University, 621# Gangwan Road, GuangZhou, 510700 Guangdong Province, China
a r t i c l e
i n f o
Article history: Received 8 March 2014 Received in revised form 14 May 2014 Accepted 4 June 2014 Keywords: Neuromyelitis optica Multiple sclerosis Cell-based assay Aquaporin 1 antibody
a b s t r a c t Objective: To develop a cell-based assay (CBA) to detect aquaporin 1 (AQP1) antibodies and determine sensitivity/specificity in patients with neuromyelitis optica (NMO) spectrum disorders. Methods: A HEK-293T transfected cell model expressing AQP1 was established and detected to be serum AQP1 antibodies. Results: AQP1 antibodies were present in 73/98 (74.5%) AQP4 antibody-positive patients. Some AQP4 antibodynegative patients were also AQP1 antibody-positive. Test sensitivity was 74.5% in 98 AQP4 antibody-positive patients. Test specificity was 79.6% in 67 multiple sclerosis (MS) patients and 31 controls. Conclusion: A sensitive and simple CBA was developed to detect serum AQP1 antibodies. AQP1 antibodies were mainly present in NMO and its high-risk syndrome, but also in some MS patients. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Neuromyelitis optica immunoglobulin G (NMO-IgG) is used as a biomarker to differentiate neuromyelitis optica (NMO) from multiple sclerosis (MS). Several pathological studies revealed that this autoantibody was an important contributor to NMO pathology (Misu et al., 2007, 2013; Saji et al., 2013). The target antigen of NMO-IgG was identified as aquaporin-4 (AQP4), the main water channel protein in the central nervous system (CNS), which is mainly expressed on astrocyte endfeet at the blood–brain barrier (BBB) (Lennon et al., 2004, 2005). However, although a number of highly sensitive immunoassays for the detection of AQP4 antibody in patients with NMO have been developed, NMO patients negative for anti-AQP4 antibodies have been identified, indicating that other autoantibodies might be involved in NMO pathogenesis (Kitley et al., 2012). However, it is unclear how anti-AQP4 antibodies cross the intact BBB from the
⁎ Corresponding author at: Key Laboratory of Neurogenetics and Channelopathies of Guangdong Province and The Ministry of Education of China, Institute of Neuroscience and The Second Affiliated Hospital of GuangZhou Medical University, 250# Changgang East Road, GuangZhou, 510260 Guangdong Province, China. Tel./fax: +86 20 3415 3147. E-mail address: [email protected] (C. Gao).
http://dx.doi.org/10.1016/j.jneuroim.2014.06.003 0165-5728/© 2014 Elsevier B.V. All rights reserved.
serum into the CNS to cause inflammatory responses. In a previous study, Shimizu et al. (2012) demonstrated that anti-brain microvascular endothelial cell antibodies in the sera from NMO patients disrupted the BBB, providing a new pathological explanation of the triggers for BBB breakdown and trafficking of AQP4 antibodies into the CNS during the acute stage of NMO. This indicates that sera antibodies against microvessels other than anti-AQP4 antibodies are necessary to disrupt the BBB. AQP1, another AQP family member, is highly expressed in human CNS astrocytes (Misu et al., 2013) and AQP1 may also selectively be lost around active NMO lesions, reflecting a role for AQP1 in NMO pathology (Misu et al., 2013). Furthermore, AQP1 is also highly expressed in microvascular endothelia (Verkman, 2002), although the functional significance of AQP1 is unclear. These studies suggested that AQP1 antibodies might induce BBB destruction and cause astrocyte injury. Recently, Tzartos et al. (2013) found that anti-AQP1 auto-antibodies were highly sensitive and specific to NMO spectrum disorders (NMOSD) using a radioimmunoprecipitation assay (RIPA). However, they failed to develop a useful cell-based assay (CBA) using AQP1-transfected HEK293 cells. Here, according to a previous study (Gao et al., 2005), we aimed to establish a CBA for anti-AQP1 antibody detection in NMOSD patients. We also analyzed the diagnostic value and clinical significance of AQP1 antibodies in a large number of Chinese patients with NMOSD and controls.
104
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
2. Patients and methods The study protocol was approved by the Ethics Committee of the Second Affiliated Hospital of Guangzhou Medical University. Written informed consent was provided by all participants. NMO/NMOSD was diagnosed as previously described (Wingerchuk et al., 2006; Fujihara and Sato, 2013). MS was diagnosed according to the 2010 criteria (Polman et al., 2011). Patients with transverse myelitis (TM) fulfilled the inclusion and exclusion criteria proposed by the Transverse Myelitis Consortium Working Group (2002) (Anon., 2002). Longitudinally extensive transverse myelitis (LETM) patients were certified by magnetic resonance imaging (Scott et al., 2006). Optic neuritis (ON) was defined by acute or subacute visual loss in unilateral or bilateral eyes. Other neurological disorders were diagnosed according to their criteria. Until December 2013, 249 samples from consecutive patients (showed in Table 1) who had undergone anti-AQP4 antibody determination were used to detect AQP1 antibodies. Patient data were retrospectively evaluated by medical record reviews and recent interviews. According to the serum profiles, 98 patients positive for AQP4 antibodies were diagnosed with NMOSD (mean age: 34.5 ± 12.5 years [12–70]). This group consisted of 89 women and 9 men (female/ male = 9.9). Furthermore, 63 patients had NMO, 21 patients had recurrent LETM (rLETM), five patients had monophasic LETM, five patients had recurrent optic neuritis (RON), and four patients without ON and TM had brainstem or brain NMO lesions. AQP4-negative patients consisted of 85 women and 66 men (female/ male = 1.3). This group consisted of 11 patients with NMO, 67 patients with MS, 8 patients with rLETM, 8 patients with monophasic LETM, 12 patients with non-LETM, 14 patients with RON, and 31 controls. The control group consisted of 31 patients with other neurological diseases including cerebral infarction (n = 20), CNS infectious disorders (n = 5), and other disorders (n = 6). Data acquired from each patient's previous record and recent interviews included age, sex (shown in Table 1), medication, number of demyelinating events, clinical characteristics, and their Expanded Disability Status Scale score (Kurtzke, 1983). 2.1. Cell culture Human embryonic kidney (HEK) 293T cells were obtained from ATTC. According to a previous study (Gao et al., 2005), HEK-293T cells were cultured in Dulbecco's modified Eagle's medium (DMEM) nutrient mixture supplemented with 10% heat inactivated fetal bovine serum (HyClone, Logan, UT, USA), 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in a humidified atmosphere of 5% CO2. 2.2. Transfection of HEK-293T cells with a construct containing the human aquaporin-1 gene The HEK-293T cell line was transduced with pcDNA3.1 Vectors (Invitrogen, IA, USA). AQP1 was amplified from a human gene cDNA library (Yingshen Company, Guangzhou, China) by polymerase chain reaction using the following primers: forward 5′-CGGGATCCGCCACCAT
GGCCAGCGAGTTCAAGAAG-3′, and reverse 5′-GGAATTCCTATTTGGGCT TCATCTCCACC-3′ and cloned into a vector. HEK-293T cells were transfected with a vector carrying the human AQP1 gene or a control vector without human AQP1 by lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). In brief, HEK-293T cells were plated and grown until 70–80% confluent. For each 35-mm well, 1 μg AQP1 or empty vector was diluted with 100 μL serum-free DMEM in one tube, and 1 μL lipofectamine 2000 was diluted with 100 μL serum-free DMEM in another tube. The DNA solutions and lipofectamine 2000 were then mixed from the two tubes and incubated at room temperature for 20 min. The DNA-lipofectamine 2000 complex mixture was added to each well. After 24 h, the medium containing the DNA-lipofectamine 2000 complex mixture was removed and replaced with complete growth medium.
2.3. Reverse-transcription polymerase chain reaction Total RNA was isolated from HEK293/AQP1 or HEK293/vector control cells using Trizol reagent (Invitrogen). First-strand cDNA was synthesized from total RNA using a reagent kit (Invitrogen) according to the manufacturer's instructions. The primer sequences were: AQP1-F: 5′-ACCTCCTGGCTATTGACTACA-3′ and AQP1-R: 5′-CAGAAAATCCAGTG GTTGCT-3′. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), as an internal control, was amplified using primers F: 5′-GGGAAACTGT GGCGTGAT-3′ and R: 5′-GAGTGGGTGTCGCTGTTGA-3′.
2.4. Western blot HEK-293T cells were lysed and centrifuged at 10,000 ×g for 10 min. For western blot analysis, a 50-μg sample was solubilized at 60 °C for 15 min, resolved by 12% polyacrylamide gel electrophoresis, and electrotransferred to a polyvinylidene difluoride membrane. The membrane with blotted protein was blocked for 1 h with blocking buffer containing 5% nonfat dry milk, followed by incubation with a rabbit anti-human AQP1 antibody (Bioss, Beijing, China) diluted 1:200 in blocking buffer at 4 °C overnight. After three 5-min washes with blocking buffer, the membrane was incubated with a goat antirabbit antibody (Southern Biotech) diluted 1:500 in blocking buffer for 2 h at room temperature. The stained bands were scanned and pixel density was quantified using a Gel Image System (Bio-Rad).
2.5. Expression of AQP1 on the surface of transfected cells and tissues AQP1-transfected cells and empty cells were fixed and sectioned. Monkey brain and stomach tissues were purchased from Euroimmun Company (Lübeck, Germany). After incubation with primary antibodies for 1 h, the substrates were washed three times in phosphate-buffered saline (PBS) and incubated with anti-human or anti-rabbit IgG (Bioss Company, Beijing, China) for 30 min, washed in PBS, and stained with 4′,6-diamidino-2-phenylindole. Images were captured using a Leica microscope.
Table 1 Comparison among NMOSD, MS and others patients. Characteristic
Female/male Median age, y (range) AQP1 antibody, n (%)
NMOSD (n = 98)
89/9 36 (12–70) 73/98 (74.5%)
pa
AQP4 antibody negative patients N-NMO (n = 11)
MS (n = 67)
RON (n = 14)
LETM (n = 16)
Non-LETM (n = 12)
Control (n = 31)
6/5 39 (28–65) 7/11 (63.6%)
37/30 32 (10–68) 19/67 (28.4%)
8/6 19 (14–46) 11/14 (78.6%)
8/8 31 (12–55) 7/16 (43.8%)
8/4 32 (4–72) 4/12 (33.3%)
14/17 40 (17–72) 1/31 (3.2%)
b0.0001 0.521 b0.0001
NMO, neuromyelitis optica; NMOSD, NMO spectrum disorders; N-NMO, AQP4 antibodies negative NMO; MS, multiple sclerosis; RON, recurrent optic neuritis; LETM, longitudinal extensive transverse myelitis. a Compared between NMOSD and MS.
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
105
2.6. Anti-AQP4 antibody testing in serum Anti-AQP4 antibodies were determined by an anti-AQP4 antibody assay as described previously (Long et al., 2012).
2.7. Anti-AQP1 antibody testing in serum Serum (1:10) was diluted in PBS. AQP1-transfected cells and nontransfected cells were fixed, sectioned, and divided into two groups: pretreatment with Triton X-100 (2%) or untreated. Pretreatment slices were incubated with Triton X-100 for 10 min and were rinsed three times with PBS for 5 min. Then, the slices were incubated with 5% normal goat serum for blocking (Boster, Wuhan, China) for 20 min. The diluted sample was combined with AQP1-transfected cells and nontransfected cells on the slides for 1 h at room temperature. The slides were then rinsed three times with PBS before incubation with fluorescein-conjugated goat anti-human IgG for 0.5 h. Finally, the slides were rinsed with PBS and the fluorescence was measured under a microscope. A positive reaction was confirmed by staining on the cell surface and in the cytoplasm of the AQP1-transfected cells, and no staining in non-transfected cells in the same sample. Sera were scored as negative or positive by two independent assessors (Youming Long and Yangbo Zheng), both of whom were blind to the clinical diagnosis. Positive samples were titrated to final dilution.
2.8. Statistical analysis All statistical analyses were performed using SPSS version 11.0 software (SPSS, Inc., Chicago, IL, USA). The McNemar test and the kappa test were used to evaluate the agreement between the two assays. p b 0.05 was considered statistically significant.
3. Results 3.1. Expression of AQP1 in transfected cells To investigate the expression of AQP1 in transfected cells, AQP1 mRNA was determined by quantitative polymerase chain reaction. The relative expression of AQP1 to the internal control revealed higher AQP1 expression in HEK-293T/AQP1 cells than in HEK-293T/vector control cells (Fig. 1). Western blot analysis detected AQP1 protein in HEK293T/AQP1 cells but not in HEK-293T/vector cell lines and 293T cells. An AQP1 band was detected with a molecular size of approximately 29 kDa (Fig. 2). An indirect immunofluorescence method was used to detect AQP1 protein on the surface of transfected cells, but not on HEK-293T/vector cells and HEK-293T cells (Fig. 3).
Fig. 2. Protein level of AQP1 in HEK-293T cells (cell), HEK-293T/vector cells (control) and HEK-293T/AQP1 cells (AQP1). AQP1 band was detected in a molecular size of approx. 29 kDa in HEK-293T/AQP1 cells.
3.2. Development of a cell-based assay to detect AQP1 antibodies In the assay without Triton X-100 pretreatment, we confirmed the low efficiency of detection of AQP1-positive cells. Microscopy confirmed that the specific expression of AQP1 on the surface of HEK-293T-AQP1 cells was colocalized with the binding sites of sero-antibodies (Fig. 4A, B, C, a, b,c). Interestingly, the positive cellular outline forms were round (Fig. 4a, b, c) and obvious background immunofluorescence was observed in most samples (data not shown). Using pretreatment with Triton X-100 as a nonionic surfactant, we confirmed the high efficiency of detection of positive cells and microscopy confirmed that the specific expression of AQP1 on the surface and cytoplasm of HEK293T-AQP1 cells was colocalized with binding sites of sero-antibodies (Fig. 4D, E, F, d, e, f). The most positive cellular outline forms were irregular (Fig. 4d, e, f) and a clear background was observed for all samples. None of the samples had specific IgG bound to HEK-293T/vector cells and HEK-293T cells. Distribution of AQP1 titers in serum samples of different patients is shown in Fig. 5. 3.3. Sensitivity and specificity of AQP1 antibody detection The characteristic surface and cytoplasm staining pattern in HEK293T cells were observed in 73/98 (74.5%) patient's positive for AQP4 antibodies. In the AQP4 antibody-negative group, 7/11 (63.6%) patients with NMO, 11/14 (78.6%) patients with RON, 7/16 (43.8%) patients with LETM, 4/12 (33.3%) patients with non-LETM, 19/67 (28.4%) patients with MS, and 1/31 (3.2%) controls were positive for AQP1 antibodies. When the cut-off point was defined as 1:10, the sensitivity of the test was 74.5% (95% confidence interval [CI] = 64.7–82.8) in 98 patients positive for AQP4 antibodies. The specificity in the control group was 79.6% (95% CI = 70.3–87.1) in 67 MS patients and 31 controls. There was a 78.5% (95% CI = 68.8–86.3) positive predictive value and 75.7% (95% CI = 66.3–83.6) negative predictive value. If patients with NMOSD, NMO without AQP4 antibodies, RON, and LETM were included in one group (n = 139) and the remaining patients (n = 110) were considered controls, the sensitivity of this assay would be 70.5% (95% CI = 62.2–77.9) and the specificity would be 78.2% (95% CI = 69.3–85.5) with an 80.3% (95% CI = 76.2–87.0) positive predictive value and 67.7% (95% CI = 58.9–75.7) negative predictive value. Table 2 summarizes the diagnostic accuracy for NMO/NMOSD diagnosis of models according to AQP1 antibody status. 3.4. Concordance between AQP1 and AQP4 antibodies The McNemar test showed that data from these two biomarkers were significantly different (p = 0.015). The kappa test showed that the data from these two biomarkers were poorly concordant (kappa = 0.385, p b 0.00001). 3.5. Representative case
Fig. 1. Relative expression of AQP1 mRNA in HEK-293T/AQP1 cells and HEK-293T/vector cells.
Patient 1, a Chinese girl, whose age at onset was 16 years, developed a decrease in bilateral vision, intractable hiccups and nausea (IHN) in July 2004. She was diagnosed with optic neuritis. After treatment with methylprednisolone and intravenous immunoglobulin, her vision and
106
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
Fig. 3. AQP1 expressions in HEK-293T/AQP1 cells by indirect immunofluorescence. A: AQP1 protein could be expressed on the surface of transfected cells (arrow). B–C: AQP1 could not be detected in HEK-293T/vector cells and HEK-293T cell.
IHN recovered. Ten months later, she experienced a second attack with decreased vision and sensory disturbance. From April 2005 to October 2012, she experienced five relapses with sensory disturbance, IHN and weakness. She was treated with low-dose oral methylprednisolone (4–8 mg/day) at remission. T2-weighed brain magnetic resonance imaging during follow-up is shown in Fig. 6. Her baseline Expanded Disability Status Scale was two at the most recent interview. Laboratory tests indicated that she was negative for antibodies to AQP4, cardiolipin, extractable nuclear antigen, thyroglobulin, and thyroid peroxidase. Anti-AQP-1 antibodies were detected in serum (N 1:10,000) using our CBA method (Fig. 7). To determine whether the AQP1 antibodies could bind to microvessels in the cerebellum and stomach, the main AQP1 antigen-expressing location, we tested her
sera by indirect immunofluorescence on sections of normal monkey cerebellum and stomach tissues. In contrast to the characteristic intense staining of microvascular elements in the brain and stomach, her serum IgG did bind to the microvessels. Dual immunostaining with AQP1specific rabbit IgG demonstrated that the antigen to which anti-AQP1 IgG binds, colocalized with AQP1 in these tissues and in human umbilical endothelial cells (Fig. 7). 4. Discussion This study describes the development of a sensitive CBA to detect serum AQP1 antibodies that are regarded as an important biomarker for NMO (Tzartos et al., 2013). It was reported that RIPAs could be
Fig. 4. Immunofluorescence patterns of cell-based assay to AQP1 antibody by different pretreatments. A–C (a–c): low efficiency of detection of positive cells in assay without Triton X-100 pretreatment. A (a): green fluorescence on AQP1-transfected cells with anti-AQP1 positive patient serum. B (b): AQP1 protein expressed on the surface of transfected cells (red); C (c): merged picture from A (a) and B (b). D–F (d–f): high efficiency of detection of positive cells in assay with Triton X-100 pretreatment. D (d): green fluorescence on AQP1-transfected cells with anti-AQP1 positive patient serum. E (e): AQP1 protein expressed on the surface of transfected cells (red); F (f): merged picture from D (d) and E (e).
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
107
Fig. 5. Distribution of AQP1 titers in serum samples of different patients. Titers are expressed as the reciprocal of tenfold serum dilutions. Serum samples below the horizontal line are negative. Lowest positive value is 1 in 10; highest is 1 in 100,000. NMOSD: neuromyelitis optica spectrum disorders; NMO: neuromyelitis; RON: recurrent optic neuritis; LETM: longitudinally extensive transverse myelitis; Non-LETM: myelitis with spinal cord lesions did not meet LETM features; MS: multiple sclerosis; AQP: aquaporin.
used as a highly sensitive tool to detect serum AQP1 antibodies in patients with NMOSD. As for the AQP4 detection system, the disadvantage of RIPA is its lack of simplicity, especially for single samples. A previous study demonstrated that RIPA does not have high sensitivity or specificity to AQP4 antibodies (Fazio et al., 2009). Currently, CBA appears to have the highest sensitivity and specificity for AQP4 antibody detection (Jarius and Wildemann, 2013). Another advantage of CBA is its simplicity (Takahashi et al., 2006). Because previous studies successfully established an HEK-293 cell line transfected with AQP1 to measure osmotic water permeability (Gao et al., 2005), we wondered whether sero-AQP1 antibodies could be detected by this assay. In the present study, we successfully established an HEK-293T cell line overexpressing AQP1, confirmed by quantitative polymerase chain reaction and western blot analysis. However, it is difficult to develop a reproducible and reliable CBA because of low concentrations of cell surface-expressed AQP1, as recently described (Tzartos et al., 2013). When using pretreatment with Triton X-100 (2%), we found a dramatic increase in the determinable AQP1 antibody by transfected HEK-293T cells and decreased background immunofluorescence. Triton treatment might result in antigen retrieval through tissue permeabilization or through alterations in conformation of denatured proteins (Mundegar et al., 2008). The method described above for background reduction and increased sensitivity in transfected HEK-293T cells is a simple alternative to RIPA that is quick to perform, and does not require additional reagents. As for CBA to AQP4 antibody detection (Takahashi et al., 2006), no staining was observed in cells without transfected AQP1 and, samples were judged to be either positive or negative
without merging staining of the sera or using the commercially available anti-AQP1 antibody. Consequently, we developed a simple inexpensive method for serum AQP1 antibody detection and this method should facilitate the detection of AQP1 antibodies in patients with NMOSD. In this study, we evaluated AQP1 antibody sensitivity in NMOSD and its specificity when compared with a control group containing different patient populations. In our series of patients with NMO and related disorders, we showed that antibodies to AQP1 were present in most patients with NMO or who were at high risk of disease, as well as in some patients with MS, but were rare in patients with other neurological diseases. Therefore, our findings demonstrated that the specificity was lower than that observed in a previous AQP1 antibody study (Tzartos et al., 2013). In the present study, we explored the AQP1 antibody diagnostic value in NMOSD, and compared AQP1 and AQP4 antibodies' status. However, we did not observe an enhanced diagnostic value when evaluating both AQP1 and AQP4 serum antibodies. Some NMO, RON, LETM, and even MS patients were seronegative for antiAQP4 antibodies but seropositive for anti-AQP1 antibodies. In addition, some sera with AQP4 positive status were negative for AQP1 antibodies using the CBA. We compared our data of AQP1 and AQP4 antibodies using the McNemar test, which looks for a systematic difference between these different biomarkers. We observed a significant difference, suggesting a diagnostic bias between the two markers. We then calculated the agreement level using a kappa test, which suggested that the two assays had a low agreement (kappa = 0.385). Thus, based on our analysis using a large sample number, the AQP1 antibody had lower
Table 2 Diagnostic accuracy for NMO/NMOSD diagnosis of models according to AQP1 antibody status. Model
Sensitivity (%)
Specificity (%)
PLR
NLR
PPV (%)
NPV (%)
NMOSD vs (MS + control) NMO vs MS (NMOSD + N-NMO) vs MS (NMOSD + N-NMO + RON + LETM) vs (MS + non-LETM)
74.49 (64.69–82.76) 70.27 (58.52–80.34) 73.39 (64.07–81.40) 85.96 (78.21–91.76)
79.59 (70.26–87.07) 71.64 (59.31–81.99) 71.64 (59.31–81.99) 76.53 (66.89–84.50)
3.65 (2.43–5.49) 2.48 (1.65–3.73) 2.59 (1.74–3.85) 3.66 (2.54–5.28)
0.32 (0.23–0.46) 0.41 (0.28–0.61) 0.37 (0.26–0.53) 0.18 (0.11–0.29)
78.49 (68.76–86.34) 73.24 (61.41–83.06) 80.81 (71.66–88.03) 80.99 (72.86–87.55)
75.73 (66.29–83.64) 68.57 (56.37–79.15) 62.34 (50.56–73.13) 82.42 (73.02–89.60)
NMO, neuromyelitis optica; NMOSD, NMO spectrum disorders; N-NMO, AQP4 antibodies negative NMO; MS, multiple sclerosis; RON, recurrent optic neuritis; LETM, longitudinal extensive transverse myelitis; PPV: positive predictive value; NPV: negative predictive value; PLR: positive likelihood ratio.
108
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
Fig. 6. MRI follow-up from a representative case with high AQP1 antibody. A: MRI shows lesion in aqueduct of median brain in 2004 (arrow); B: a NMO characteristic lesion in left hypothalamus in 2005 (arrow); C: abnormality in left optic chiasm in 2005 (arrow); D: symmetrical hyperintense lesions in the brain in 2008 (arrow); E: lesions in medulla in 2008 (arrow); F: lesion in aqueduct of median brain in 2009 (arrow); G: symmetrical hypointensity lesions in the brain in 2010 (arrow); H: symmetrical hyperintense lesions in the brain (arrow) and a new lesion (dot arrow) in 2012; I: new non-specific lesions in the brain in 2012 (arrow).
diagnostic value than the AQP4 antibody in NMO. This indicates a discrepancy in antibody status that exists between AQP1 and AQP4 and may reflect their different detection sensitivity and immune mechanisms. Although the present data do not support the use of anti-AQP1 antibodies as a highly-specific diagnostic biomarker for NMO, there are important factors that should be taken into account. First, anti-AQP4 antibodies are not the only contributor to NMO, and other pathogenic antibodies in serum may orchestrate pathogenesis in some seronegative patients (Kitley et al., 2012). AQP1 is strongly expressed by human CNS astrocytes (Misu et al., 2013) and AQP1 was selectively lost around active NMO lesions, indicating that AQP1 might be involved in NMO pathology (Misu et al., 2013). Furthermore, seronegative anti-AQP4 NMO and high-risk syndrome (RON and LETM) patients were positive for anti-AQP1 antibodies in the present study and a previous study
(Tzartos et al., 2013). As a representative patient, a NMO-IgG negative girl with very high serum titers of the AQP1 antibody experienced relapsing ON and brainstem syndrome. Although no critical evidence demonstrated that very high levels of AQP1 antibodies were involved in a similar pathogenesis to NMO-IgG, it indicated that such antibodies might play a role in the relapsing disease course. Second, it is interesting to note that AQP1 and AQP4 antibodies coexist in most patients with NMOSD, especially when there are high titers of AQP1 antibodies. Previous data strongly support a central role for AQP4 antibodies in NMO pathogenesis (Jarius et al., 2008). However, how AQP4 antibodies cross the intact BBB from serum to the CNS to cause inflammatory responses is unclear. Intravenous administration of the AQP4 antibody cannot be observed elsewhere in the brain, spinal cord, optic nerve, or retina (Ratelade et al., 2011) and can induce an NMO model without disrupting the BBB (Kinoshita et al., 2009; Ratelade et al., 2011). The
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
109
Fig. 7. Auto-antibodies in the representative case. A: AQP1 antibodies bind to transfected HEK-293T cell (green); B: AQP1 protein expressed in transfected cells (red); C: merged picture from A and B; D: serum autoantibodies bind to microvessel in cerebellum (green); E: AQP1 protein expressed in microvessel in cerebellum (green); F: merged picture from D and E; G: serum autoantibodies bind to microvessel in stomach (green); H: AQP1 protein expressed in microvessel in stomach (green); I: merged picture from G and H; J: serum autoantibodies bind to human umbilical endothelial cells (HUEVC) (green); K: AQP1 protein expressed in surface and cytoplasm of HUEVC (red); L: merged picture from J and K. Dual immunostaining demonstrated: (1) part of serum IgG that colocalizes with AQP1 in HUEVC (short arrow); (2) some unclear serum antibodies do not colocalize with AQP1 in HUEVC (long arrow).
destruction of the BBB is considered the initial step of NMO (Shimizu et al., 2012). A recent report (Long et al., 2012) indicated that autoantibodies specific for human brain microvascular endothelial cells, other than anti-AQP4 antibodies, in NMO patients may disrupt the BBB. A previous study also showed that antibodies targeting microvessels were observed in idiopathic inflammatory-demyelinating diseases (Long et al., 2013). AQP1 is also highly expressed in microvascular endothelia (Verkman, 2002), although the functional significance of AQP1 is unclear. In the present study, microscopy using dual immunostaining with AQP1-specific rabbit IgG demonstrated that the AQP1-positive IgG colocalized with AQP1 in the microvascular endothelia of tissues and in HUEVCs in vitro. Therefore, AQP1 antibodies might induce BBB destruction and astrocyte injury. This should be confirmed in further studies. Third, 28.4% (19/67) of patients with MS were positive for AQP1 antibodies, although they were present at very low titers. As previously described, the presence of anti-AQP1 autoantibodies in patients diagnosed with MS is not surprising (Tzartos et al., 2013). In a previous study, serum auto-antibodies were observed to bind to brain microvessels differing from the classical NMO-IgG staining (Long et al., 2013). Therefore, in our opinion, it may be that one antigen recognized by such
antibodies is AQP1, indicating that it may be associated with the disruption of the BBB in MS. In conclusion, we developed a sensitive and simple CBA to detect serum AQP1 antibodies. Our study confirmed that: 1) AQP1 antibodies are mainly present in NMO and its high-risk syndrome, especially in patients positive for anti-AQP4 antibodies; and 2) AQP1 antibodies are present in some MS patients. Funding/support This project was supported by the plan of the China Postdoctoral Science Foundation Grant (2012M521586), Guangdong Natural Science Foundation (S2013010016262), the Science and Technology plan project of Guangdong Province (2012B031800240, 2012B061700042), and the plan of GuangZhou Medical University (2012C46). Acknowledgment We thank miss Jinmei Lin (MSSbio company, Guangzhou, China) for the help of transfection of HEK-293T cells.
110
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
References Anon., 2002. Proposed diagnostic criteria and nosology of acute transverse myelitis. Neurology 59 (4), 499–505. Fazio, R., Malosio, M.L., Lampasona, V., et al., 2009. Antiacquaporin 4 antibodies detection by different techniques in neuromyelitis optica patients. Mult. Scler. 15 (10), 1153–1163. Fujihara, K., Sato, D.K., 2013. AQP4 antibody serostatus: is its luster being lost in the management and pathogenesis of NMO. Neurology 81 (14), 1186–1188. Gao, J., Yu, H., Song, Q., Li, X., 2005. Establishment of HEK293 cell line expressing green fluorescent protein–aquaporin-1 to determine osmotic water permeability. Anal. Biochem. 342 (1), 53–58. Jarius, S., Wildemann, B., 2013. Aquaporin-4 antibodies (NMO-IgG) as a serological marker of neuromyelitis optica: a critical review of the literature. Brain Pathol. 23 (6), 661–683. Jarius, S., Paul, F., Franciotta, D., et al., 2008. Mechanisms of disease: aquaporin-4 antibodies in neuromyelitis optica. Nat. Clin. Pract. Neurol. 4 (4), 202–214. Kinoshita, M., Nakatsuji, Y., Kimura, T., et al., 2009. Neuromyelitis optica: passive transfer to rats by human immunoglobulin. Biochem. Biophys. Res. Commun. 386 (4), 623–627. Kitley, J., Woodhall, M., Waters, P., et al., 2012. Myelin-oligodendrocyte glycoprotein antibodies in adults with a neuromyelitis optica phenotype. Neurology 79 (12), 1273–1277. Kurtzke, J.F., 1983. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 33 (11), 1444–1452. Lennon, V.A., Wingerchuk, D.M., Kryzer, T.J., et al., 2004. A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis. Lancet 364 (9451), 2106–2112. Lennon, V.A., Kryzer, T.J., Pittock, S.J., Verkman, A.S., Hinson, S.R., 2005. IgG marker of optic-spinal multiple sclerosis binds to the aquaporin-4 water channel. J. Exp. Med. 202 (4), 473–477. Long, Y., Qiu, W., Lu, Z., et al., 2012. Aquaporin 4 antibodies in the cerebrospinal fluid are helpful in diagnosing Chinese patients with neuromyelitis optica. Neuroimmunomodulation 19 (2), 96–102.
Long, Y., Gao, C., Qiu, W., Hu, X., Peng, F., Lu, Z., 2013. Antibodies target microvessels in neuromyelitis optica and multiple sclerosis patients. Neurol. Res. 35 (9), 922–929. Misu, T., Fujihara, K., Kakita, A., et al., 2007. Loss of aquaporin 4 in lesions of neuromyelitis optica: distinction from multiple sclerosis. Brain 130 (Pt 5), 1224–1234. Misu, T., Hoftberger, R., Fujihara, K., et al., 2013. Presence of six different lesion types suggests diverse mechanisms of tissue injury in neuromyelitis optica. Acta Neuropathol. 125 (6), 815–827. Mundegar, R.R., Franke, E., Schafer, R., Zweyer, M., Wernig, A., 2008. Reduction of high background staining by heating unfixed mouse skeletal muscle tissue sections allows for detection of thermostable antigens with murine monoclonal antibodies. J. Histochem. Cytochem. 56 (11), 969–975. Polman, C.H., Reingold, S.C., Banwell, B., et al., 2011. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann. Neurol. 69 (2), 292–302. Ratelade, J., Bennett, J.L., Verkman, A.S., 2011. Intravenous neuromyelitis optica autoantibody in mice targets aquaporin-4 in peripheral organs and area postrema. PLoS One 6 (11), e27412. Saji, E., Arakawa, M., Yanagawa, K., et al., 2013. Cognitive impairment and cortical degeneration in neuromyelitis optica. Ann. Neurol. 73 (1), 65–76. Scott, T.F., Kassab, S.L., Pittock, S.J., 2006. Neuromyelitis optica IgG status in acute partial transverse myelitis. Arch. Neurol. 63 (10), 1398–1400. Shimizu, F., Sano, Y., Takahashi, T., et al., 2012. Sera from neuromyelitis optica patients disrupt the blood–brain barrier. J. Neurol. Neurosurg. Psychiatry 83 (3), 288–297. Takahashi, T., Fujihara, K., Nakashima, I., et al., 2006. Establishment of a new sensitive assay for anti-human aquaporin-4 antibody in neuromyelitis optica. Tohoku J. Exp. Med. 210 (4), 307–313. Tzartos, J.S., Stergiou, C., Kilidireas, K., Zisimopoulou, P., Thomaidis, T., Tzartos, S.J., 2013. Anti-aquaporin-1 autoantibodies in patients with neuromyelitis optica spectrum disorders. PLoS One 8 (9), e74773. Verkman, A.S., 2002. Aquaporin water channels and endothelial cell function. J. Anat. 200 (6), 617–627. Wingerchuk, D.M., Lennon, V.A., Pittock, S.J., Lucchinetti, C.F., Weinshenker, B.G., 2006. Revised diagnostic criteria for neuromyelitis optica. Neurology 66 (10), 1485–1489.
Contents lists available at ScienceDirect
Journal of Neuroimmunology journal homepage: www.elsevier.com/locate/jneuroim
Development of a cell-based assay for the detection of anti-aquaporin 1 antibodies in neuromyelitis optica spectrum disorders Youming Long a,b, Yangbo Zheng a,b, Fulan Shan a,b, Mengyu Chen a,b, Yongxiang Fan a,b, Bin Zhang a,b, Cong Gao a,b,⁎, Qingchun Gao a,b, Ning Yang c a Key Laboratory of Neurogenetics and Channelopathies of Guangdong Province and The Ministry of Education of China, Institute of Neuroscience and the Second Affiliated Hospital of GuangZhou Medical University, 250# Changgang East Road, GuangZhou, 510260 Guangdong Province, China b Department of Neurology, The Second Affiliated Hospital of GuangZhou Medical University, 250# Changgang East Road, GuangZhou, 510260 Guangdong Province, China c Department of Neurology, The Fifth Affiliated Hospital of GuangZhou Medical University, 621# Gangwan Road, GuangZhou, 510700 Guangdong Province, China
a r t i c l e
i n f o
Article history: Received 8 March 2014 Received in revised form 14 May 2014 Accepted 4 June 2014 Keywords: Neuromyelitis optica Multiple sclerosis Cell-based assay Aquaporin 1 antibody
a b s t r a c t Objective: To develop a cell-based assay (CBA) to detect aquaporin 1 (AQP1) antibodies and determine sensitivity/specificity in patients with neuromyelitis optica (NMO) spectrum disorders. Methods: A HEK-293T transfected cell model expressing AQP1 was established and detected to be serum AQP1 antibodies. Results: AQP1 antibodies were present in 73/98 (74.5%) AQP4 antibody-positive patients. Some AQP4 antibodynegative patients were also AQP1 antibody-positive. Test sensitivity was 74.5% in 98 AQP4 antibody-positive patients. Test specificity was 79.6% in 67 multiple sclerosis (MS) patients and 31 controls. Conclusion: A sensitive and simple CBA was developed to detect serum AQP1 antibodies. AQP1 antibodies were mainly present in NMO and its high-risk syndrome, but also in some MS patients. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Neuromyelitis optica immunoglobulin G (NMO-IgG) is used as a biomarker to differentiate neuromyelitis optica (NMO) from multiple sclerosis (MS). Several pathological studies revealed that this autoantibody was an important contributor to NMO pathology (Misu et al., 2007, 2013; Saji et al., 2013). The target antigen of NMO-IgG was identified as aquaporin-4 (AQP4), the main water channel protein in the central nervous system (CNS), which is mainly expressed on astrocyte endfeet at the blood–brain barrier (BBB) (Lennon et al., 2004, 2005). However, although a number of highly sensitive immunoassays for the detection of AQP4 antibody in patients with NMO have been developed, NMO patients negative for anti-AQP4 antibodies have been identified, indicating that other autoantibodies might be involved in NMO pathogenesis (Kitley et al., 2012). However, it is unclear how anti-AQP4 antibodies cross the intact BBB from the
⁎ Corresponding author at: Key Laboratory of Neurogenetics and Channelopathies of Guangdong Province and The Ministry of Education of China, Institute of Neuroscience and The Second Affiliated Hospital of GuangZhou Medical University, 250# Changgang East Road, GuangZhou, 510260 Guangdong Province, China. Tel./fax: +86 20 3415 3147. E-mail address: [email protected] (C. Gao).
http://dx.doi.org/10.1016/j.jneuroim.2014.06.003 0165-5728/© 2014 Elsevier B.V. All rights reserved.
serum into the CNS to cause inflammatory responses. In a previous study, Shimizu et al. (2012) demonstrated that anti-brain microvascular endothelial cell antibodies in the sera from NMO patients disrupted the BBB, providing a new pathological explanation of the triggers for BBB breakdown and trafficking of AQP4 antibodies into the CNS during the acute stage of NMO. This indicates that sera antibodies against microvessels other than anti-AQP4 antibodies are necessary to disrupt the BBB. AQP1, another AQP family member, is highly expressed in human CNS astrocytes (Misu et al., 2013) and AQP1 may also selectively be lost around active NMO lesions, reflecting a role for AQP1 in NMO pathology (Misu et al., 2013). Furthermore, AQP1 is also highly expressed in microvascular endothelia (Verkman, 2002), although the functional significance of AQP1 is unclear. These studies suggested that AQP1 antibodies might induce BBB destruction and cause astrocyte injury. Recently, Tzartos et al. (2013) found that anti-AQP1 auto-antibodies were highly sensitive and specific to NMO spectrum disorders (NMOSD) using a radioimmunoprecipitation assay (RIPA). However, they failed to develop a useful cell-based assay (CBA) using AQP1-transfected HEK293 cells. Here, according to a previous study (Gao et al., 2005), we aimed to establish a CBA for anti-AQP1 antibody detection in NMOSD patients. We also analyzed the diagnostic value and clinical significance of AQP1 antibodies in a large number of Chinese patients with NMOSD and controls.
104
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
2. Patients and methods The study protocol was approved by the Ethics Committee of the Second Affiliated Hospital of Guangzhou Medical University. Written informed consent was provided by all participants. NMO/NMOSD was diagnosed as previously described (Wingerchuk et al., 2006; Fujihara and Sato, 2013). MS was diagnosed according to the 2010 criteria (Polman et al., 2011). Patients with transverse myelitis (TM) fulfilled the inclusion and exclusion criteria proposed by the Transverse Myelitis Consortium Working Group (2002) (Anon., 2002). Longitudinally extensive transverse myelitis (LETM) patients were certified by magnetic resonance imaging (Scott et al., 2006). Optic neuritis (ON) was defined by acute or subacute visual loss in unilateral or bilateral eyes. Other neurological disorders were diagnosed according to their criteria. Until December 2013, 249 samples from consecutive patients (showed in Table 1) who had undergone anti-AQP4 antibody determination were used to detect AQP1 antibodies. Patient data were retrospectively evaluated by medical record reviews and recent interviews. According to the serum profiles, 98 patients positive for AQP4 antibodies were diagnosed with NMOSD (mean age: 34.5 ± 12.5 years [12–70]). This group consisted of 89 women and 9 men (female/ male = 9.9). Furthermore, 63 patients had NMO, 21 patients had recurrent LETM (rLETM), five patients had monophasic LETM, five patients had recurrent optic neuritis (RON), and four patients without ON and TM had brainstem or brain NMO lesions. AQP4-negative patients consisted of 85 women and 66 men (female/ male = 1.3). This group consisted of 11 patients with NMO, 67 patients with MS, 8 patients with rLETM, 8 patients with monophasic LETM, 12 patients with non-LETM, 14 patients with RON, and 31 controls. The control group consisted of 31 patients with other neurological diseases including cerebral infarction (n = 20), CNS infectious disorders (n = 5), and other disorders (n = 6). Data acquired from each patient's previous record and recent interviews included age, sex (shown in Table 1), medication, number of demyelinating events, clinical characteristics, and their Expanded Disability Status Scale score (Kurtzke, 1983). 2.1. Cell culture Human embryonic kidney (HEK) 293T cells were obtained from ATTC. According to a previous study (Gao et al., 2005), HEK-293T cells were cultured in Dulbecco's modified Eagle's medium (DMEM) nutrient mixture supplemented with 10% heat inactivated fetal bovine serum (HyClone, Logan, UT, USA), 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in a humidified atmosphere of 5% CO2. 2.2. Transfection of HEK-293T cells with a construct containing the human aquaporin-1 gene The HEK-293T cell line was transduced with pcDNA3.1 Vectors (Invitrogen, IA, USA). AQP1 was amplified from a human gene cDNA library (Yingshen Company, Guangzhou, China) by polymerase chain reaction using the following primers: forward 5′-CGGGATCCGCCACCAT
GGCCAGCGAGTTCAAGAAG-3′, and reverse 5′-GGAATTCCTATTTGGGCT TCATCTCCACC-3′ and cloned into a vector. HEK-293T cells were transfected with a vector carrying the human AQP1 gene or a control vector without human AQP1 by lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). In brief, HEK-293T cells were plated and grown until 70–80% confluent. For each 35-mm well, 1 μg AQP1 or empty vector was diluted with 100 μL serum-free DMEM in one tube, and 1 μL lipofectamine 2000 was diluted with 100 μL serum-free DMEM in another tube. The DNA solutions and lipofectamine 2000 were then mixed from the two tubes and incubated at room temperature for 20 min. The DNA-lipofectamine 2000 complex mixture was added to each well. After 24 h, the medium containing the DNA-lipofectamine 2000 complex mixture was removed and replaced with complete growth medium.
2.3. Reverse-transcription polymerase chain reaction Total RNA was isolated from HEK293/AQP1 or HEK293/vector control cells using Trizol reagent (Invitrogen). First-strand cDNA was synthesized from total RNA using a reagent kit (Invitrogen) according to the manufacturer's instructions. The primer sequences were: AQP1-F: 5′-ACCTCCTGGCTATTGACTACA-3′ and AQP1-R: 5′-CAGAAAATCCAGTG GTTGCT-3′. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), as an internal control, was amplified using primers F: 5′-GGGAAACTGT GGCGTGAT-3′ and R: 5′-GAGTGGGTGTCGCTGTTGA-3′.
2.4. Western blot HEK-293T cells were lysed and centrifuged at 10,000 ×g for 10 min. For western blot analysis, a 50-μg sample was solubilized at 60 °C for 15 min, resolved by 12% polyacrylamide gel electrophoresis, and electrotransferred to a polyvinylidene difluoride membrane. The membrane with blotted protein was blocked for 1 h with blocking buffer containing 5% nonfat dry milk, followed by incubation with a rabbit anti-human AQP1 antibody (Bioss, Beijing, China) diluted 1:200 in blocking buffer at 4 °C overnight. After three 5-min washes with blocking buffer, the membrane was incubated with a goat antirabbit antibody (Southern Biotech) diluted 1:500 in blocking buffer for 2 h at room temperature. The stained bands were scanned and pixel density was quantified using a Gel Image System (Bio-Rad).
2.5. Expression of AQP1 on the surface of transfected cells and tissues AQP1-transfected cells and empty cells were fixed and sectioned. Monkey brain and stomach tissues were purchased from Euroimmun Company (Lübeck, Germany). After incubation with primary antibodies for 1 h, the substrates were washed three times in phosphate-buffered saline (PBS) and incubated with anti-human or anti-rabbit IgG (Bioss Company, Beijing, China) for 30 min, washed in PBS, and stained with 4′,6-diamidino-2-phenylindole. Images were captured using a Leica microscope.
Table 1 Comparison among NMOSD, MS and others patients. Characteristic
Female/male Median age, y (range) AQP1 antibody, n (%)
NMOSD (n = 98)
89/9 36 (12–70) 73/98 (74.5%)
pa
AQP4 antibody negative patients N-NMO (n = 11)
MS (n = 67)
RON (n = 14)
LETM (n = 16)
Non-LETM (n = 12)
Control (n = 31)
6/5 39 (28–65) 7/11 (63.6%)
37/30 32 (10–68) 19/67 (28.4%)
8/6 19 (14–46) 11/14 (78.6%)
8/8 31 (12–55) 7/16 (43.8%)
8/4 32 (4–72) 4/12 (33.3%)
14/17 40 (17–72) 1/31 (3.2%)
b0.0001 0.521 b0.0001
NMO, neuromyelitis optica; NMOSD, NMO spectrum disorders; N-NMO, AQP4 antibodies negative NMO; MS, multiple sclerosis; RON, recurrent optic neuritis; LETM, longitudinal extensive transverse myelitis. a Compared between NMOSD and MS.
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
105
2.6. Anti-AQP4 antibody testing in serum Anti-AQP4 antibodies were determined by an anti-AQP4 antibody assay as described previously (Long et al., 2012).
2.7. Anti-AQP1 antibody testing in serum Serum (1:10) was diluted in PBS. AQP1-transfected cells and nontransfected cells were fixed, sectioned, and divided into two groups: pretreatment with Triton X-100 (2%) or untreated. Pretreatment slices were incubated with Triton X-100 for 10 min and were rinsed three times with PBS for 5 min. Then, the slices were incubated with 5% normal goat serum for blocking (Boster, Wuhan, China) for 20 min. The diluted sample was combined with AQP1-transfected cells and nontransfected cells on the slides for 1 h at room temperature. The slides were then rinsed three times with PBS before incubation with fluorescein-conjugated goat anti-human IgG for 0.5 h. Finally, the slides were rinsed with PBS and the fluorescence was measured under a microscope. A positive reaction was confirmed by staining on the cell surface and in the cytoplasm of the AQP1-transfected cells, and no staining in non-transfected cells in the same sample. Sera were scored as negative or positive by two independent assessors (Youming Long and Yangbo Zheng), both of whom were blind to the clinical diagnosis. Positive samples were titrated to final dilution.
2.8. Statistical analysis All statistical analyses were performed using SPSS version 11.0 software (SPSS, Inc., Chicago, IL, USA). The McNemar test and the kappa test were used to evaluate the agreement between the two assays. p b 0.05 was considered statistically significant.
3. Results 3.1. Expression of AQP1 in transfected cells To investigate the expression of AQP1 in transfected cells, AQP1 mRNA was determined by quantitative polymerase chain reaction. The relative expression of AQP1 to the internal control revealed higher AQP1 expression in HEK-293T/AQP1 cells than in HEK-293T/vector control cells (Fig. 1). Western blot analysis detected AQP1 protein in HEK293T/AQP1 cells but not in HEK-293T/vector cell lines and 293T cells. An AQP1 band was detected with a molecular size of approximately 29 kDa (Fig. 2). An indirect immunofluorescence method was used to detect AQP1 protein on the surface of transfected cells, but not on HEK-293T/vector cells and HEK-293T cells (Fig. 3).
Fig. 2. Protein level of AQP1 in HEK-293T cells (cell), HEK-293T/vector cells (control) and HEK-293T/AQP1 cells (AQP1). AQP1 band was detected in a molecular size of approx. 29 kDa in HEK-293T/AQP1 cells.
3.2. Development of a cell-based assay to detect AQP1 antibodies In the assay without Triton X-100 pretreatment, we confirmed the low efficiency of detection of AQP1-positive cells. Microscopy confirmed that the specific expression of AQP1 on the surface of HEK-293T-AQP1 cells was colocalized with the binding sites of sero-antibodies (Fig. 4A, B, C, a, b,c). Interestingly, the positive cellular outline forms were round (Fig. 4a, b, c) and obvious background immunofluorescence was observed in most samples (data not shown). Using pretreatment with Triton X-100 as a nonionic surfactant, we confirmed the high efficiency of detection of positive cells and microscopy confirmed that the specific expression of AQP1 on the surface and cytoplasm of HEK293T-AQP1 cells was colocalized with binding sites of sero-antibodies (Fig. 4D, E, F, d, e, f). The most positive cellular outline forms were irregular (Fig. 4d, e, f) and a clear background was observed for all samples. None of the samples had specific IgG bound to HEK-293T/vector cells and HEK-293T cells. Distribution of AQP1 titers in serum samples of different patients is shown in Fig. 5. 3.3. Sensitivity and specificity of AQP1 antibody detection The characteristic surface and cytoplasm staining pattern in HEK293T cells were observed in 73/98 (74.5%) patient's positive for AQP4 antibodies. In the AQP4 antibody-negative group, 7/11 (63.6%) patients with NMO, 11/14 (78.6%) patients with RON, 7/16 (43.8%) patients with LETM, 4/12 (33.3%) patients with non-LETM, 19/67 (28.4%) patients with MS, and 1/31 (3.2%) controls were positive for AQP1 antibodies. When the cut-off point was defined as 1:10, the sensitivity of the test was 74.5% (95% confidence interval [CI] = 64.7–82.8) in 98 patients positive for AQP4 antibodies. The specificity in the control group was 79.6% (95% CI = 70.3–87.1) in 67 MS patients and 31 controls. There was a 78.5% (95% CI = 68.8–86.3) positive predictive value and 75.7% (95% CI = 66.3–83.6) negative predictive value. If patients with NMOSD, NMO without AQP4 antibodies, RON, and LETM were included in one group (n = 139) and the remaining patients (n = 110) were considered controls, the sensitivity of this assay would be 70.5% (95% CI = 62.2–77.9) and the specificity would be 78.2% (95% CI = 69.3–85.5) with an 80.3% (95% CI = 76.2–87.0) positive predictive value and 67.7% (95% CI = 58.9–75.7) negative predictive value. Table 2 summarizes the diagnostic accuracy for NMO/NMOSD diagnosis of models according to AQP1 antibody status. 3.4. Concordance between AQP1 and AQP4 antibodies The McNemar test showed that data from these two biomarkers were significantly different (p = 0.015). The kappa test showed that the data from these two biomarkers were poorly concordant (kappa = 0.385, p b 0.00001). 3.5. Representative case
Fig. 1. Relative expression of AQP1 mRNA in HEK-293T/AQP1 cells and HEK-293T/vector cells.
Patient 1, a Chinese girl, whose age at onset was 16 years, developed a decrease in bilateral vision, intractable hiccups and nausea (IHN) in July 2004. She was diagnosed with optic neuritis. After treatment with methylprednisolone and intravenous immunoglobulin, her vision and
106
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
Fig. 3. AQP1 expressions in HEK-293T/AQP1 cells by indirect immunofluorescence. A: AQP1 protein could be expressed on the surface of transfected cells (arrow). B–C: AQP1 could not be detected in HEK-293T/vector cells and HEK-293T cell.
IHN recovered. Ten months later, she experienced a second attack with decreased vision and sensory disturbance. From April 2005 to October 2012, she experienced five relapses with sensory disturbance, IHN and weakness. She was treated with low-dose oral methylprednisolone (4–8 mg/day) at remission. T2-weighed brain magnetic resonance imaging during follow-up is shown in Fig. 6. Her baseline Expanded Disability Status Scale was two at the most recent interview. Laboratory tests indicated that she was negative for antibodies to AQP4, cardiolipin, extractable nuclear antigen, thyroglobulin, and thyroid peroxidase. Anti-AQP-1 antibodies were detected in serum (N 1:10,000) using our CBA method (Fig. 7). To determine whether the AQP1 antibodies could bind to microvessels in the cerebellum and stomach, the main AQP1 antigen-expressing location, we tested her
sera by indirect immunofluorescence on sections of normal monkey cerebellum and stomach tissues. In contrast to the characteristic intense staining of microvascular elements in the brain and stomach, her serum IgG did bind to the microvessels. Dual immunostaining with AQP1specific rabbit IgG demonstrated that the antigen to which anti-AQP1 IgG binds, colocalized with AQP1 in these tissues and in human umbilical endothelial cells (Fig. 7). 4. Discussion This study describes the development of a sensitive CBA to detect serum AQP1 antibodies that are regarded as an important biomarker for NMO (Tzartos et al., 2013). It was reported that RIPAs could be
Fig. 4. Immunofluorescence patterns of cell-based assay to AQP1 antibody by different pretreatments. A–C (a–c): low efficiency of detection of positive cells in assay without Triton X-100 pretreatment. A (a): green fluorescence on AQP1-transfected cells with anti-AQP1 positive patient serum. B (b): AQP1 protein expressed on the surface of transfected cells (red); C (c): merged picture from A (a) and B (b). D–F (d–f): high efficiency of detection of positive cells in assay with Triton X-100 pretreatment. D (d): green fluorescence on AQP1-transfected cells with anti-AQP1 positive patient serum. E (e): AQP1 protein expressed on the surface of transfected cells (red); F (f): merged picture from D (d) and E (e).
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
107
Fig. 5. Distribution of AQP1 titers in serum samples of different patients. Titers are expressed as the reciprocal of tenfold serum dilutions. Serum samples below the horizontal line are negative. Lowest positive value is 1 in 10; highest is 1 in 100,000. NMOSD: neuromyelitis optica spectrum disorders; NMO: neuromyelitis; RON: recurrent optic neuritis; LETM: longitudinally extensive transverse myelitis; Non-LETM: myelitis with spinal cord lesions did not meet LETM features; MS: multiple sclerosis; AQP: aquaporin.
used as a highly sensitive tool to detect serum AQP1 antibodies in patients with NMOSD. As for the AQP4 detection system, the disadvantage of RIPA is its lack of simplicity, especially for single samples. A previous study demonstrated that RIPA does not have high sensitivity or specificity to AQP4 antibodies (Fazio et al., 2009). Currently, CBA appears to have the highest sensitivity and specificity for AQP4 antibody detection (Jarius and Wildemann, 2013). Another advantage of CBA is its simplicity (Takahashi et al., 2006). Because previous studies successfully established an HEK-293 cell line transfected with AQP1 to measure osmotic water permeability (Gao et al., 2005), we wondered whether sero-AQP1 antibodies could be detected by this assay. In the present study, we successfully established an HEK-293T cell line overexpressing AQP1, confirmed by quantitative polymerase chain reaction and western blot analysis. However, it is difficult to develop a reproducible and reliable CBA because of low concentrations of cell surface-expressed AQP1, as recently described (Tzartos et al., 2013). When using pretreatment with Triton X-100 (2%), we found a dramatic increase in the determinable AQP1 antibody by transfected HEK-293T cells and decreased background immunofluorescence. Triton treatment might result in antigen retrieval through tissue permeabilization or through alterations in conformation of denatured proteins (Mundegar et al., 2008). The method described above for background reduction and increased sensitivity in transfected HEK-293T cells is a simple alternative to RIPA that is quick to perform, and does not require additional reagents. As for CBA to AQP4 antibody detection (Takahashi et al., 2006), no staining was observed in cells without transfected AQP1 and, samples were judged to be either positive or negative
without merging staining of the sera or using the commercially available anti-AQP1 antibody. Consequently, we developed a simple inexpensive method for serum AQP1 antibody detection and this method should facilitate the detection of AQP1 antibodies in patients with NMOSD. In this study, we evaluated AQP1 antibody sensitivity in NMOSD and its specificity when compared with a control group containing different patient populations. In our series of patients with NMO and related disorders, we showed that antibodies to AQP1 were present in most patients with NMO or who were at high risk of disease, as well as in some patients with MS, but were rare in patients with other neurological diseases. Therefore, our findings demonstrated that the specificity was lower than that observed in a previous AQP1 antibody study (Tzartos et al., 2013). In the present study, we explored the AQP1 antibody diagnostic value in NMOSD, and compared AQP1 and AQP4 antibodies' status. However, we did not observe an enhanced diagnostic value when evaluating both AQP1 and AQP4 serum antibodies. Some NMO, RON, LETM, and even MS patients were seronegative for antiAQP4 antibodies but seropositive for anti-AQP1 antibodies. In addition, some sera with AQP4 positive status were negative for AQP1 antibodies using the CBA. We compared our data of AQP1 and AQP4 antibodies using the McNemar test, which looks for a systematic difference between these different biomarkers. We observed a significant difference, suggesting a diagnostic bias between the two markers. We then calculated the agreement level using a kappa test, which suggested that the two assays had a low agreement (kappa = 0.385). Thus, based on our analysis using a large sample number, the AQP1 antibody had lower
Table 2 Diagnostic accuracy for NMO/NMOSD diagnosis of models according to AQP1 antibody status. Model
Sensitivity (%)
Specificity (%)
PLR
NLR
PPV (%)
NPV (%)
NMOSD vs (MS + control) NMO vs MS (NMOSD + N-NMO) vs MS (NMOSD + N-NMO + RON + LETM) vs (MS + non-LETM)
74.49 (64.69–82.76) 70.27 (58.52–80.34) 73.39 (64.07–81.40) 85.96 (78.21–91.76)
79.59 (70.26–87.07) 71.64 (59.31–81.99) 71.64 (59.31–81.99) 76.53 (66.89–84.50)
3.65 (2.43–5.49) 2.48 (1.65–3.73) 2.59 (1.74–3.85) 3.66 (2.54–5.28)
0.32 (0.23–0.46) 0.41 (0.28–0.61) 0.37 (0.26–0.53) 0.18 (0.11–0.29)
78.49 (68.76–86.34) 73.24 (61.41–83.06) 80.81 (71.66–88.03) 80.99 (72.86–87.55)
75.73 (66.29–83.64) 68.57 (56.37–79.15) 62.34 (50.56–73.13) 82.42 (73.02–89.60)
NMO, neuromyelitis optica; NMOSD, NMO spectrum disorders; N-NMO, AQP4 antibodies negative NMO; MS, multiple sclerosis; RON, recurrent optic neuritis; LETM, longitudinal extensive transverse myelitis; PPV: positive predictive value; NPV: negative predictive value; PLR: positive likelihood ratio.
108
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
Fig. 6. MRI follow-up from a representative case with high AQP1 antibody. A: MRI shows lesion in aqueduct of median brain in 2004 (arrow); B: a NMO characteristic lesion in left hypothalamus in 2005 (arrow); C: abnormality in left optic chiasm in 2005 (arrow); D: symmetrical hyperintense lesions in the brain in 2008 (arrow); E: lesions in medulla in 2008 (arrow); F: lesion in aqueduct of median brain in 2009 (arrow); G: symmetrical hypointensity lesions in the brain in 2010 (arrow); H: symmetrical hyperintense lesions in the brain (arrow) and a new lesion (dot arrow) in 2012; I: new non-specific lesions in the brain in 2012 (arrow).
diagnostic value than the AQP4 antibody in NMO. This indicates a discrepancy in antibody status that exists between AQP1 and AQP4 and may reflect their different detection sensitivity and immune mechanisms. Although the present data do not support the use of anti-AQP1 antibodies as a highly-specific diagnostic biomarker for NMO, there are important factors that should be taken into account. First, anti-AQP4 antibodies are not the only contributor to NMO, and other pathogenic antibodies in serum may orchestrate pathogenesis in some seronegative patients (Kitley et al., 2012). AQP1 is strongly expressed by human CNS astrocytes (Misu et al., 2013) and AQP1 was selectively lost around active NMO lesions, indicating that AQP1 might be involved in NMO pathology (Misu et al., 2013). Furthermore, seronegative anti-AQP4 NMO and high-risk syndrome (RON and LETM) patients were positive for anti-AQP1 antibodies in the present study and a previous study
(Tzartos et al., 2013). As a representative patient, a NMO-IgG negative girl with very high serum titers of the AQP1 antibody experienced relapsing ON and brainstem syndrome. Although no critical evidence demonstrated that very high levels of AQP1 antibodies were involved in a similar pathogenesis to NMO-IgG, it indicated that such antibodies might play a role in the relapsing disease course. Second, it is interesting to note that AQP1 and AQP4 antibodies coexist in most patients with NMOSD, especially when there are high titers of AQP1 antibodies. Previous data strongly support a central role for AQP4 antibodies in NMO pathogenesis (Jarius et al., 2008). However, how AQP4 antibodies cross the intact BBB from serum to the CNS to cause inflammatory responses is unclear. Intravenous administration of the AQP4 antibody cannot be observed elsewhere in the brain, spinal cord, optic nerve, or retina (Ratelade et al., 2011) and can induce an NMO model without disrupting the BBB (Kinoshita et al., 2009; Ratelade et al., 2011). The
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
109
Fig. 7. Auto-antibodies in the representative case. A: AQP1 antibodies bind to transfected HEK-293T cell (green); B: AQP1 protein expressed in transfected cells (red); C: merged picture from A and B; D: serum autoantibodies bind to microvessel in cerebellum (green); E: AQP1 protein expressed in microvessel in cerebellum (green); F: merged picture from D and E; G: serum autoantibodies bind to microvessel in stomach (green); H: AQP1 protein expressed in microvessel in stomach (green); I: merged picture from G and H; J: serum autoantibodies bind to human umbilical endothelial cells (HUEVC) (green); K: AQP1 protein expressed in surface and cytoplasm of HUEVC (red); L: merged picture from J and K. Dual immunostaining demonstrated: (1) part of serum IgG that colocalizes with AQP1 in HUEVC (short arrow); (2) some unclear serum antibodies do not colocalize with AQP1 in HUEVC (long arrow).
destruction of the BBB is considered the initial step of NMO (Shimizu et al., 2012). A recent report (Long et al., 2012) indicated that autoantibodies specific for human brain microvascular endothelial cells, other than anti-AQP4 antibodies, in NMO patients may disrupt the BBB. A previous study also showed that antibodies targeting microvessels were observed in idiopathic inflammatory-demyelinating diseases (Long et al., 2013). AQP1 is also highly expressed in microvascular endothelia (Verkman, 2002), although the functional significance of AQP1 is unclear. In the present study, microscopy using dual immunostaining with AQP1-specific rabbit IgG demonstrated that the AQP1-positive IgG colocalized with AQP1 in the microvascular endothelia of tissues and in HUEVCs in vitro. Therefore, AQP1 antibodies might induce BBB destruction and astrocyte injury. This should be confirmed in further studies. Third, 28.4% (19/67) of patients with MS were positive for AQP1 antibodies, although they were present at very low titers. As previously described, the presence of anti-AQP1 autoantibodies in patients diagnosed with MS is not surprising (Tzartos et al., 2013). In a previous study, serum auto-antibodies were observed to bind to brain microvessels differing from the classical NMO-IgG staining (Long et al., 2013). Therefore, in our opinion, it may be that one antigen recognized by such
antibodies is AQP1, indicating that it may be associated with the disruption of the BBB in MS. In conclusion, we developed a sensitive and simple CBA to detect serum AQP1 antibodies. Our study confirmed that: 1) AQP1 antibodies are mainly present in NMO and its high-risk syndrome, especially in patients positive for anti-AQP4 antibodies; and 2) AQP1 antibodies are present in some MS patients. Funding/support This project was supported by the plan of the China Postdoctoral Science Foundation Grant (2012M521586), Guangdong Natural Science Foundation (S2013010016262), the Science and Technology plan project of Guangdong Province (2012B031800240, 2012B061700042), and the plan of GuangZhou Medical University (2012C46). Acknowledgment We thank miss Jinmei Lin (MSSbio company, Guangzhou, China) for the help of transfection of HEK-293T cells.
110
Y. Long et al. / Journal of Neuroimmunology 273 (2014) 103–110
References Anon., 2002. Proposed diagnostic criteria and nosology of acute transverse myelitis. Neurology 59 (4), 499–505. Fazio, R., Malosio, M.L., Lampasona, V., et al., 2009. Antiacquaporin 4 antibodies detection by different techniques in neuromyelitis optica patients. Mult. Scler. 15 (10), 1153–1163. Fujihara, K., Sato, D.K., 2013. AQP4 antibody serostatus: is its luster being lost in the management and pathogenesis of NMO. Neurology 81 (14), 1186–1188. Gao, J., Yu, H., Song, Q., Li, X., 2005. Establishment of HEK293 cell line expressing green fluorescent protein–aquaporin-1 to determine osmotic water permeability. Anal. Biochem. 342 (1), 53–58. Jarius, S., Wildemann, B., 2013. Aquaporin-4 antibodies (NMO-IgG) as a serological marker of neuromyelitis optica: a critical review of the literature. Brain Pathol. 23 (6), 661–683. Jarius, S., Paul, F., Franciotta, D., et al., 2008. Mechanisms of disease: aquaporin-4 antibodies in neuromyelitis optica. Nat. Clin. Pract. Neurol. 4 (4), 202–214. Kinoshita, M., Nakatsuji, Y., Kimura, T., et al., 2009. Neuromyelitis optica: passive transfer to rats by human immunoglobulin. Biochem. Biophys. Res. Commun. 386 (4), 623–627. Kitley, J., Woodhall, M., Waters, P., et al., 2012. Myelin-oligodendrocyte glycoprotein antibodies in adults with a neuromyelitis optica phenotype. Neurology 79 (12), 1273–1277. Kurtzke, J.F., 1983. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 33 (11), 1444–1452. Lennon, V.A., Wingerchuk, D.M., Kryzer, T.J., et al., 2004. A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis. Lancet 364 (9451), 2106–2112. Lennon, V.A., Kryzer, T.J., Pittock, S.J., Verkman, A.S., Hinson, S.R., 2005. IgG marker of optic-spinal multiple sclerosis binds to the aquaporin-4 water channel. J. Exp. Med. 202 (4), 473–477. Long, Y., Qiu, W., Lu, Z., et al., 2012. Aquaporin 4 antibodies in the cerebrospinal fluid are helpful in diagnosing Chinese patients with neuromyelitis optica. Neuroimmunomodulation 19 (2), 96–102.
Long, Y., Gao, C., Qiu, W., Hu, X., Peng, F., Lu, Z., 2013. Antibodies target microvessels in neuromyelitis optica and multiple sclerosis patients. Neurol. Res. 35 (9), 922–929. Misu, T., Fujihara, K., Kakita, A., et al., 2007. Loss of aquaporin 4 in lesions of neuromyelitis optica: distinction from multiple sclerosis. Brain 130 (Pt 5), 1224–1234. Misu, T., Hoftberger, R., Fujihara, K., et al., 2013. Presence of six different lesion types suggests diverse mechanisms of tissue injury in neuromyelitis optica. Acta Neuropathol. 125 (6), 815–827. Mundegar, R.R., Franke, E., Schafer, R., Zweyer, M., Wernig, A., 2008. Reduction of high background staining by heating unfixed mouse skeletal muscle tissue sections allows for detection of thermostable antigens with murine monoclonal antibodies. J. Histochem. Cytochem. 56 (11), 969–975. Polman, C.H., Reingold, S.C., Banwell, B., et al., 2011. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann. Neurol. 69 (2), 292–302. Ratelade, J., Bennett, J.L., Verkman, A.S., 2011. Intravenous neuromyelitis optica autoantibody in mice targets aquaporin-4 in peripheral organs and area postrema. PLoS One 6 (11), e27412. Saji, E., Arakawa, M., Yanagawa, K., et al., 2013. Cognitive impairment and cortical degeneration in neuromyelitis optica. Ann. Neurol. 73 (1), 65–76. Scott, T.F., Kassab, S.L., Pittock, S.J., 2006. Neuromyelitis optica IgG status in acute partial transverse myelitis. Arch. Neurol. 63 (10), 1398–1400. Shimizu, F., Sano, Y., Takahashi, T., et al., 2012. Sera from neuromyelitis optica patients disrupt the blood–brain barrier. J. Neurol. Neurosurg. Psychiatry 83 (3), 288–297. Takahashi, T., Fujihara, K., Nakashima, I., et al., 2006. Establishment of a new sensitive assay for anti-human aquaporin-4 antibody in neuromyelitis optica. Tohoku J. Exp. Med. 210 (4), 307–313. Tzartos, J.S., Stergiou, C., Kilidireas, K., Zisimopoulou, P., Thomaidis, T., Tzartos, S.J., 2013. Anti-aquaporin-1 autoantibodies in patients with neuromyelitis optica spectrum disorders. PLoS One 8 (9), e74773. Verkman, A.S., 2002. Aquaporin water channels and endothelial cell function. J. Anat. 200 (6), 617–627. Wingerchuk, D.M., Lennon, V.A., Pittock, S.J., Lucchinetti, C.F., Weinshenker, B.G., 2006. Revised diagnostic criteria for neuromyelitis optica. Neurology 66 (10), 1485–1489.
