During allograft failure, many transplant patients have antibodies against endothelial cells (1, 2). To identify antigens associated with allograft failure, we obtained sera after rejection and performed immunoprecipitation with lysates of human umbilical vein endothelial cells (HUVEC) followed by mass spectrometry to identify the proteins binding to the immune complexes. The sera were absorbed with pooled human platelets and verified to be free of detectable antibodies against human leukocyte antigen (HLA) class I antigens. Freshly isolated HUVEC do not express HLA class II antigens, so those would not be detected (3). Antibodies from patient WW were found to react with a protein antigen identified as nucleolin. Nucleolin is a multifunctional protein in the nucleolus and is also found on the cell surface of different types of cells (4). The human nucleolin gene consists of 14 exons with 13 introns and spans approximately 11 kb (5). Anti-nucleolin antibodies (6–10) detected nucleolin on the surface of proliferating tumor cells and on endothelial cells of blood vessels undergoing angiogenesis in vivo. (4, 11).
In these experiments, we developed an enzyme-linked immunosorbent assay (ELISA) with recombinant nucleolin to determine the occurrence of autoantibodies against nucleolin in patients before and after organ transplantation. Kidney transplant and heart transplant recipients have been studied and both groups have been found to produce antibodies against recombinant nucleolin. The possible functional significance of antibodies against nucleolin was investigated by analyzing their effect on endothelial cells in culture. These antibodies in the serum of transplant patients inhibited the growth of endothelial cells and seemed to cause cell death by apoptosis.
Antibodies in Serum From Organ Transplant Recipients After Immunological Rejection
Many sera obtained from organ transplant recipients produced staining of the surface of HUVEC. When sera obtained from transplant recipients after rejection were screened on a panel of endothelial cells, we observed (1) a polymorphic pattern, in which certain sera reacted with some target cells and not with others and (2) a monomorphic pattern in which antibodies bound to every HUVEC in the panel. Serum from patient WW produced a monomorphic pattern of panel reactions and was therefore thought to most likely contain autoantibodies.
Identification of Target Antigens by Immunoprecipitation and Mass Spectrometry
To identify the antigens recognized by antibodies from patient WW, we performed immunoprecipitation followed by mass spectrometry and compared the amino acid sequences obtained with those in a protein database. After immunoprecipitation and elution, a band with a mobility of approximately 100 kDa was observed by Western blot and silver staining (Fig. 1A). This band was cut from the gel and submitted for mass spectrometry. Control experiments were performed with segments of similar mobility retrieved from gels obtained using normal human serum instead of the patient serum. No specific proteins were identified. In the case of serum WW, peptides corresponding to different regions of the protein nucleolin were found in each of three experiments. The location of the peptides identified in the amino acid sequence of nucleolin is shown in Figure 1(B). We concluded that the antibodies in serum from patient WW reacted with nucleolin.
Nucleolin on the Surface of HUVEC
HUVEC were stained with 4E2, a monoclonal antibody specific for nucleolin. It produced staining on fresh HUVEC (Fig. 1C) and of the endothelial cell line EA.hy 926 (data not shown), indicating that nucleolin was detected on the surface of endothelial cells, as previously reported (12).
Production of Recombinant Nucleolin
Recombinant nucleolin (r-nucleolin) was produced to detect antibodies against nucleolin by ELISA. The construct used included the four RNA-binding domains and the C-terminal RGG boxes of nucleolin and a His-tag marker to be used for purification (11). After production of r-nucleolin in Escherichia coli strain BL21 (DE3), lysates containing the protein were purified in two steps: First on a Ni charged His.Band resin column and subsequently using another column with immobilized anti-nucleolin antibody MS-3. After the purification, a single band of approximately 50 kDa was obtained by SDS-PAGE and staining with Coomasie blue. This material was shown to be nucleolin by Western blot with the monoclonal antibody MS-3, against nucleolin.
Determination of Antibodies Against Nucleolin by ELISA With r-Nucleolin and Specificity of the Assay
Sera from 49 normal persons were used to establish the positive threshold for the ELISA assay. The mean of the optical density (OD) readings with the normal sera was 0.185±0.1. We used the mean plus 2SD to determine the threshold, at a reading of 0.4 OD at 450 nm. To confirm the specificity of the ELISA assay with r-nucleolin, absorption and elution experiments were performed. Serum from patient 4, containing antibodies reactive with r-nucleolin by ELISA, was absorbed by incubation with HUVEC. When the serum was tested after absorption, it was found that almost all of the anti-nucleolin activity had been removed. Moreover, the eluate was found to contain much of the anti-r-nucleolin antibody activity that was removed by the absorption.
Antibodies Against Nucleolin Found in Organ Transplant Recipients
We tested 66 sera from patients waiting for a kidney transplant and 51 specimens from recipients that had been transplanted with a kidney from an unrelated donor and who had subsequently rejected the organ. As can be seen in Table 1 and Figure 2, six of the 66 (9.1%) patients on the kidney transplant waiting list were found to have antibodies against r-nucleolin. Among the 51 kidney recipients tested after rejection waiting for a re-transplant, 13 or 25.5% had antibodies against r-nucleolin and strong reactions were observed (Table 1, Fig. 2). We also tested 218 recipients of heart transplants. Chronic rejection manifested as transplant-related coronary artery disease (TCAD) was present in 89 (41%) of these patients. Antibodies against nucleolin were detected by ELISA in 61 (28%) of the heart transplant patients. In the 89 recipients with TCAD, 39 were found to have antibodies against nucleolin, giving a frequency of 43.8%. In heart allograft recipients without TCAD, only 22 of 129 (17.1%) had antibodies to nucleolin. Among 61 patients with antibodies against nucleolin 39 developed TCAD (63.9%) and this lesion was detected in only 50 of 157 patients without antibodies against nucleolin (31.8%). The difference between the two groups was statistically significant (P<0.0001), suggesting an association between antibodies against nucleolin and development of TCAD in this group of patients. Antibodies against donor HLA antigens were found in many of these patients (Table 1). However, six patients who had antibodies against nucleolin did not have antibodies against donor HLA at the time of detection of TCAD. Autoantibodies against lymphocytes by CDC in 25 and by flow cytometry in five patients that had antibodies against nucleolin, including patient WW, were all negative probably because surface nucleolin on resting T cells is too low to detect.
Endothelial Cell Function After Culture With Antibodies Against Nucleolin From Transplant Patients
To determine the effect of cross-linking of surface nucleolin with anti-nucleolin antibody, we used the mouse monoclonal antibody C23 (MS-3) against human nucleolin and IgG fractions of human sera known to have anti-nucleolin activity by Western blot. As shown in Figure 3(A), when endothelial cells were incubated with the monoclonal antibody against nucleolin, a marked decrease in MTT staining occurred compared with cultures incubated with normal mouse IgG (P<0.01). These results agree with observations by Fogal et al. (12) who found that NCL3, a polyclonal rabbit antibody against a nucleolin peptide, can inhibit endothelial proliferation in cultures and inhibit angiogenesis in experimental tumor grafts in vivo. The effect of IgG obtained from sera of four patients found to have autoantibodies against nucleolin is shown in the Figure 3(A). Sera from patients 1 and 4 decreased the survival of human endothelial cells (P<0.05).
To further analyze the effect of the anti-nucleolin antibodies on the function of endothelial cells, we investigated the formation of capillary-like tubes in vitro. HUVEC cultured on dishes covered with Matrigel formed tubular structures resembling capillaries. Similar activity was observed when these cultures were incubated with normal mouse IgG or with normal human serum (Fig. 4). When HUVEC cultures on Matrigel were incubated with the monoclonal antibody C23 (MS-3) against nucleolin, tubule formation was inhibited (Fig. 4A). Moreover, the two sera from patients who had antibodies against nucleolin also markedly inhibited endothelial cell tube formation (Fig. 4B).
Next, we tested for apoptosis by staining for Annexin V by flow cytometry. In these experiments, HUVEC were incubated with nucleolin monoclonal antibody or patient serum containing antibodies against nucleolin. As shown in Figure 3(B), after treatment with the monoclonal antibody 4E2 to nucleolin or patient serum containing nucleolin autoantibodies, 23% or 15% of the cells stained for Annexin V, respectively. In contrast, after incubation with a mouse IgG control, staining was only 6.5% and incubation with normal human serum resulted in 2.5% of the cells staining for Annexin. These results suggest that the antibodies against nucleolin produced apoptosis of endothelial cells with a marked increase in cells staining for Annexin V after incubation with these antibodies.
The results with serum from patient WW, who had previously rejected a kidney allograft, showed that the antigen precipitated by serum WW was nucleolin. This was subsequently confirmed by using antibodies against nucleolin from two different sources by Western blot and by ELISA with r-nucleolin.
Using ELISA, we found that a small number of recipients (9.1%) waiting for a kidney transplant had antibodies to nucleolin. The frequency in normal subjects was only 2% (Table 1). After rejection of a kidney allograft, three times more patients (25.5%) were found to have autoantibodies to nucleolin in their blood. In addition, autoantibodies against nucleolin were also found in patients with heart transplantation. We studied 218 human recipients of heart allografts, many of whom were being followed for more than 10 years, and 89 (40.8%) had developed chronic rejection with documented TCAD by angiography and ultrasound. Antibodies against nucleolin, determined by the ELISA assay with r-nucleolin, and chronic rejection with TCAD were associated. TCAD was observed in 63.9% of patients with antibodies against nucleolin and in only 31.8% of recipients without antibodies against nucleolin (P<0.0001). It is of interest that among the patients studied, there were six who were found to have antibodies against nucleolin and no antibodies against donor HLA antigens at the time that TCAD was detected.
It is not known whether antibodies against nucleolin are harmful to the grafted organs or their vasculature. Our results with mouse monoclonal antibodies and with some antibodies against nucleolin that develop in patients showed that mouse monoclonal antibodies and human serum antibodies against nucleolin can inhibit proliferation of endothelial cells, can decrease the formation of capillary-like tubules in cultures of endothelial cells incubated over a Matrigel matrix, and can cause death of endothelial cells by apoptosis. Whether such effects also occur in vivo is not known, but seems possible. It is also not known whether autoantibodies against nucleolin have a deleterious effect on the survival of tissue or organ transplants. Other autoantibodies, including those against vimentin (13), alpha-tubulin (14), and collagen V (15), have been thought by other investigators to play a role in rejection or to contribute to organ transplant failure (1, 2).
The strong association between antibodies against nucleolin and the most common vascular lesion in heart transplant recipients affecting the coronary arteries is of considerable interest. More work will be needed to determine the time course of the immune response against nucleolin in relation to the lesions developing in the vasculature of the transplant. More detailed studies will determine the role of other antibodies such as donor HLA antigens and donor MICA/MICB antigens in individual cases of a transplant recipient developing the vascular lesions in the coronary arteries. In the present study, there were six patients who had antibodies against nucleolin at the time of development of TCAD in whom antibodies against donor HLA antigens were not detectable using Luminex beads. Because antibodies against nucleolin inhibit angiogenesis and cause apoptosis of endothelial cells in vitro, the possibility of similar effects contributing to the vascular lesions that develop in patients with heart transplants should be studied further.
MATERIALS AND METHODS
Sera from 49 normal human subjects, 66 patients on the waiting list for kidney transplantation, 51 patients that had rejected a kidney allograft and 218 recipients of heart transplants at our institution were included in this study with approval of the Institutional Review Board. The sera were absorbed with washed pooled human platelets to remove antibodies against HLA antigens and verified by testing with single HLA antigen Luminex beads (One Lambda, Canoga Park, CA).
Endothelial Cell Isolation and Culture
Cord veins were cannulated and washed with phosphate-buffered saline solution and treated with collagenase at 23°C for 20 min. Endothelial cells were collected and grown at 37°C in M199 Medium with 20% fetal calf serum in a humid atmosphere of 5% CO2 in air (16).
The pRSET nucleolin plasmid was obtained from Addgene (Cambridge, MA). The mouse monoclonal antibody (MS-3) to human nucleolin was purchased from Santa Cruz Biotechnolgy (Santa Cruz, CA) and the antibody from clone 4E2 from Enzo Life Sciences (Plymouth Meeting, PA). Endothelial cell growth supplement, PE-conjugated monoclonal anti-CD31, no. 555446, Matrigel Matrix basement membrane and Annexin V-FITC apoptosis detection kits were obtained from BD Biosciences (San Diego, CA). 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (TACS MTT) was obtained from R&D systems (Minneapolis, MN). nProtein A Sepharose 4 Fast Flow was purchased from GE Healthcare (Buckinghamshire HP8 4SP, UK). ProFound Mammalian Co-Immunoprecipitation Kit, Micro BCA Protein Assay Kit, and Slide-A-Lyzer Dialysis Cassettes were purchased from Thermo Scientific (Pierce, Rockford, IL) Collagenase, M199 medium, isopropyl β-d-thiogalactoside (IPTG) were purchased from Sigma (St. Louis, MO). HRP-conjugated Goat anti-Human IgG Fc, HRP-conjugated Goat anti-Mouse IgG Fc were obtained from Bethyl Laboratories (Montgomery, TX).
HUVEC were removed from culture dishes with 0.25% Trypsin-EDTA solution and approximately 3×108 cells were incubated with 8 mL of patient serum known to react with the endothelial cells by flow cytometry, diluted 1:4, for 1 hr at room temperature with gentle shaking. After incubation, the cells were washed three times and then eluted with citric acid buffer pH 3.0 for 5 min and neutralized by mixing with pH 9.0 phosphate buffer. IgG antibodies purified by a protein A Sepharose column were used for immunoprecipitation with a Pierce Direct Immunoprecipitation Kit according to the instructions provided by the manufacturer.
Protein Identification by Mass Spectrometry
After performing SDS-PAGE, the bands were visualized by silver staining. Bands were excised, digested with modified porcine trypsin overnight and analyzed by nano-HPLC-MS/MS (Protein Chemistry Technology Center, UT Southwestern Medical Center, Dallas, TX). The files containing the data were submitted for a search against the NCBI protein sequence database using the Mascot software. Protein matches with scores greater than 150 were considered significant and the one with the highest score was considered to be the predominant protein in the sample.
Cell Proliferation Assay
A total of 6×104 HUVEC were incubated for 30 min on ice with MS-3 an anti-nucleolin monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA) or patient serum IgG. After removal of the antibodies, the cells were plated on 96 well plates. After 24 hr of culture, 10 μL of MTT reagent (R&D system, Minneapolis, MN) was added to each well. The plate was incubated at 37°C for 2 to 4 hr. One hundred microliters of detergent reagent was added to the wells, including control wells. Plates were left covered in the dark at 37°C for at least 2 hr and the absorbance was measured at a wavelength of 550 to 600 nm.
Staining for Annexin V as an Indicator of Apoptosis
Primary HUVEC were incubated for 18 hr in 2% fetal calf serum medium containing 100 μg/mL of nucleolin mAb or purified mouse IgG isotype as negative control and patient serum IgG or normal human serum (NHS) IgG, respectively. Percentages of Annexin V-positive cells were determined within the viable population of cells by flow cytometry analysis.
Capillary Tube Formation Assay
These cultures were performed in eight-well chamber slides (Nalge Nunc International Corp., Naperville, IL) coated with 150 μL of Matrigel (BD Biosciences) per well. HUVEC were preincubated for 30 min with 100 μg/mL of nucleolin mAb or patient serum IgG or normal control IgG and placed on the Matrigel for 24 hr. Gels were examined in a Nikon TE300 microscope equipped with Hoffman modulation optics and a Hamamatsu cooled CCD camera (Technical Instruments, Burlingame, CA). Tube formation was quantified by measuring the long axis of each tube in three random fields per well by bright field microscopy at 20× magnification.
Preparation of Recombinant Nucleolin
pRSET nucleolin plasmid containing 6-His-tagged truncated nucleolin, (residues 274-707), which contains all four RNA-binding domains and the C-terminal RGG boxes was obtained from Addgene (Cambridge, MA) and was originally produced by Wu et al. (11). The plasmid was used to transform E. Coli strain BL21(DE3) carrying the plasmid pLysE (11). Nucleolin expression was induced and cells sonicated. Recombinant nucleolin was purified first on a Ni charged His·Band Resin column (Novagen, Madison, WI) and subsequently with an immobilized-nucleolin antibody column.
Ninety-six well plates (Nunc Apogent, Rochester, NY) were coated with 5 μg/mL nucleolin and were blocked with 5% nonfat dry milk/phosphate-buffered saline. The serum samples were diluted at 1:100 and duplicates incubated for 1 hr. After washing, plates were incubated with HRP- conjugated goat anti-human IgG. The cutoff value obtained from 49 normal sera was 0.4 OD at 450 nm.
Means±standard deviations (SD) were calculated and differences between groups were determined by the Student's t test. For three or more groups, statistical analysis was performed using one-way analysis of variance, followed by the Bonferroni correction, as appropriate. Correlation was tested using Pearson or Spearman tests. P less than 0.05 was considered significant.
1.Bharat A, Saini D, Steward N, et al. Antibodies to self-antigens predispose to primary lung allograft dysfunction and chronic rejection. Ann Thorac Surg
2010; 90: 1094.
2.Win TS, Pettigrew GJ. Humoral autoimmunity and transplant vasculopathy: When allo is not enough. Transplantation
2010; 90: 113.
3.Nunez G, Ball EJ, Stastny P. Accessory cell function of human endothelial cells
. I. A subpopulation of Ia positive cells is required for antigen presentation. J Immunol
1983; 131: 666.
4.Hovanessian AG, Soundaramourty C, El Khoury D, et al. Surface expressed nucleolin is constantly induced in tumor cells to mediate calcium-dependent ligand internalization. PLoS One
2010; 5: e15787.
5.Srivastava M, McBride OW, Fleming PJ, et al. Genomic organization and chromosomal localization of the human nucleolin gene. J Biol Chem
1990; 265: 14922.
6.Bell SA, Faust H, Mittermuller J, et al. Specificity of antinuclear antibodies in scleroderma-like chronic graft-versus-host disease: Clinical correlation and histocompatibility locus antigen association. Br J Dermatol
1996; 134: 848.
7.Deng JS, Ballou B, Hofmeister JK. Internalization of anti-nucleolin antibody into viable HEp-2 cells. Mol Biol Rep
1996; 23: 191.
8.Hirata D, Iwamoto M, Yoshio T, et al. Nucleolin as the earliest target molecule of autoantibodies
produced in MRL/lpr lupus-prone mice. Clin Immunol
2000; 97: 50.
9.Minota S, Jarjour WN, Suzuki N, et al. Autoantibodies
to nucleolin in systemic lupus erythematosus and other diseases. J Immunol
1991; 146: 2249.
10.Wesierska-Gadek J, Penner E, Hitchman E, et al. Nucleolar proteins B23 and C23 as target antigens in chronic graft-versus-host disease. Blood
1992; 79: 1081.
11.Wu R, Shen Q, Newburger PE. Recognition and binding of the human selenocysteine insertion sequence by nucleolin. J Cell Biochem
2000; 77: 507.
12.Fogal V, Sugahara KN, Ruoslahti E, et al. Cell surface nucleolin antagonist causes endothelial cell apoptosis and normalization of tumor vasculature. Angiogenesis
2009; 12: 91.
13.Mahesh B, Leong HS, McCormack A, et al. Autoantibodies
to vimentin cause accelerated rejection of cardiac allografts. Am J Pathol
2007; 170: 1415.
14.Goers TA, Ramachandran S, Aloush A, et al. De novo production of K-alpha1 tubulin-specific antibodies: Role in chronic lung allograft rejection. J Immunol
2008; 180: 4487.
15.Nath DS, Tiriveedhi V, Basha HI, et al. A role for antibodies to human leukocyte antigens, collagen-V, and K-alpha1-tubulin in antibody- mediated rejection and cardiac allograft vasculopathy. Transplantation
2011; 2011: 4.
16.Moraes JR, Stastny P. A new antigen system expressed in human endothelial cells
. J Clin Invest
1977; 60: 449.