Upregulation of pro-inflammatory bradykinin peptides in systemic lupus erythematosus and rheumatoid arthritis

Abstract

Our recent study has implicated bradykinin signaling as being of pathogenic importance in lupus. This study aims to investigate the biomarker potential of bradykinin peptides, BK and BK-des-arg-9, in lupus and other rheumatic autoimmune diseases. Sera form SLE patients and healthy subjects were screened for BK and BK-des-arg-9 by LC-MS metabolomics. Serum from 6-month old C57BL/6 mice and three murine lupus strains were also screened for the two peptides by metabolomics. Given the promising initial screening results, validation of these 2 peptides was next conducted using multiple reaction monitoring (MRM) in larger patient cohorts. In initial metabolomics screening, BK-des-arg-9 was 22-fold higher in SLE serum and 106-fold higher in mouse lupus serum compared to healthy controls. In validation assays using MRM and Q-TOF mass spectrometry, BK and BK-des-arg-9 showed significant elevations in SLE serum compared to controls (P<0.0001; AUC = 0.79–0.88), with a similar but less pronounced increase being noted in RA serum. Interestingly, increased renal SLEDAI in lupus patients was associated with reduced circulating BK-des-arg-9, and the reasons for this remain to be explored. To sum, increased conversion of BK to the pro-inflammatory metabolite, BK-des-arg-9 appears to be a common theme in systemic rheumatic diseases. Besides serving as an early marker for systemic autoimmunity, independent studies also show that this metabolic axis may also be a pathogenic driver and therapeutic target in lupus.

Introduction:

Kinins are oligopeptides (9–11 amino acid peptides) generated from kininogens by the action of serine proteases, called kallikreins. There are two types of kininogens, heavy molecular weight kininogen (HMWK) and low molecular weight kininogen (LMWK). Plasma and tissue kallikreins act on HMWK and LMWK producing bradykinin and kallidin (kinins), respectively. Bradykinin (BK) is a nano peptide that is acted upon by kininase I / carboxypeptidase N (CPN) to generate bradykinin-des-arg-9 (BK-des-arg-9), an octapeptide.

BK and BK-des-arg-9 exert their biological effects through their receptors, bradykinin receptor 2 (BK2R) and bradykinin receptor 1 (BK1R), respectively. Both BK1R and BK2R bear seven helix transmembrane domains with a configuration shared with rhodopsin family of G protein coupled receptors (1). While BK2R is constitutively expressed in sensory nerve fibers, leukocytes, mast cells and endothelial cells (2), the expression of BK1R is very low in normal tissue unless stimulated by tissue injury, cytokines, LPS or inflammation (3). Upon tissue injury or exposure to cytokines like IL-1B and TNF alpha, the expression of BK1R is believed to be upregulated under the control of the MAP kinase pathway among others (4). Sequencing of the human BK1R has shown the existence of a transcriptional regulation site for NFKB in the promoter region (5). BK1R is also reported to enhance the production of secondary mediators like prostanoids, tachykinins, nitric oxide and mast cell derived products which help in propagating inflammatory and nociceptive processes (2).

This pathway has also been shown to be pathogenic in a couple of autoimmune and inflammatory diseases. A study conducted by Marceau et al., showed the involvement of BK1R in inflammatory bowel disease and reported the reduction in disease upon blocking BK1R using an antagonist (6). Similar results have been reported by Gobel et al., in multiple sclerosis and experimentally induced encephalomyelitis (7). Reduced glomerular and tubular lesions and improved renal function have been reported by Klien et al in experimentally induced nephritis (8). We have recently shown that blocking the bradykinin B1 receptor subdues lupus nephritis and blood pressure in MRL/lpr lupus-prone mice (9).

Given the role of BK receptors in inflammation and autoimmune diseases, this study focuses on the kinins that bind to these receptors and upregulate downstream pathways. Thus far, the precise levels of BK peptides have never been examined in systemic rheumatic diseases. The current study is designed based on initial semi-quantitative metabolomics data on BK and BK-des-arg-9 in both human and mouse lupus serum samples. Based on the initial data, a novel highly sensitive and quantitative multiple reaction monitoring (MRM) based assay was established to quantitate serum BK and BK-des- arg-9 in two rheumatic autoimmune diseases, SLE and RA.

Methods:

Biological samples:

Archived sera were obtained from SLE patients and healthy controls previously recruited from the Renal Clinic at UT Southwestern Medical Center, Dallas, TX. Disease activity was assessed using SLEDAI (SLE disease activity index) and renal SLEDAI (rSLEDAI), which captures just the renal domains of SLEDAI. SLE patients were sub-classified into active SLE and inactive SLE, where active SLE was defined as patients having a SLEDAI ≥ 5. Thus, inactive patients had SLEDAI ≤ 4. All human research was approved by the UT Southwestern Institutional Review Board (IRB) at Dallas, TX. All RA sera were obtained from the Rheumatology clinics at Albert Einstein college of Medicine, Bronx, New York, with approval of the IRB. All SLE and RA patients satisfied the ACR classification criteria for SLE and RA respectively. The demographics, treatment history and co-morbidities of all SLE and RA patients used for validation are detailed in Tables 1 and 2, respectively. The healthy controls used for validation had a mean age 32.3 (SD = 7.0), with 61% being female, 61% being African American, 19% being White, and 16% being Hispanic.

Table 1:

Clinical and Demographic Characteristics of SLE Patients¹

Characteristics	Variables	All SLE (N=37)	Active (N=24)	Inactive (N=13)
Age		32.4±12.5	29.8±9.6	37.3±15.5
Sex	Female	32 (86.5%)	21 (87.5%)	11 (84.6%)
	Male	5 (13.5%)	3 (12.5%)	2 (15.4%)
Race	Black	14 (37.8%)	9 (37.5%)	5 (38.5%)
	Hispanic	22 (59.5%)	14 (58.3%)	8 (61.5%)
	White	1 (2.7%)	1 (4.2%)	0 (0.00%)
	SLEDAI	9.4±7.1	13.3±5.8****	2.3±1.6
	Renal SLEDAI	5.3±4.6	8±3.4 ****	0.3±1.1
	eGFR (ml/min/m^2)	87.7±50.3	90.0±57.8	83.58±32.1
Disease manifestations	Renal	37 (100.0%)	24 (100.0%)***	1 (7.6%)
	NPSLE	3(8.1%)	3 (12.5%)	0 (0.0%)
	Anemia	13 (35.1%)	8 (33.3%)	5 (38.5%)
	APS	3 (8.1%)	0 (0.0%)**	3 (23.1%)
	PE/DVT	4 (10.8%)	1 (4.2%)	3 (23.1%)
Co-morbidities	DM	0 (0.0%)	0 (0.0%)	0 (0.0%)
	HTN	13 (35.1%)	8 (33.3%)	5 (38.5%)
	HLP	13 (35.1%)	10 (41.7%)	3 (23.1%)
Clinical Features	Serum Creatinine (mg/dl)	1.2±0.8	1.3±0.9	1.1±0.7
	Pr/Cr Ratio	1.7±2.4	2.2±2.6	0.7±1.6
	ANA	16 (43.2%)	13 (54.2%)	3 (23.1%)
	Double-Strand DNA	19 (51.4%)	11 (45.8%)	8 (61.5%)
	Low Complement	13 (35.1%)	9 (37.5%)	4 (30.8%)
Medications	Prednisone	24 (64.9%)	15 (62.5%)	9 (69.0%)
	HCQ	21 (56.8%)	13 (54.2%)	8 (61.5%)
	CTX	3 (8.1%)	2 (8.33%)	1 (7.7%)
	MMF	13 (35.1%)	8 (33.3%)	5 (38.5%)
	AZA	2 (5.4%)	0 (0.0%)	2 (15.4%)
	MTX	1 (2.7%)	1 (4.2%)	0 (0.0%)
	CsA/Tacrolimus	1 (2.7%)	0 (0.0%)	1 (7.7%)

^1:Values are mean ± SD or n (%). SLE, systemic lupus erythematosus; SLEDAI, SLE disease activity index; Renal SLEDAI, renal SLE disease activity index; eGFR, estimated glomerular filtration rate; NPSLE, neuropsychiatric SLE; APS, antiphospholipid syndrome; PE/DVT, pulmonary embolism/deep vein thrombosis; DM, diabetes mellitus; HTN, hypertension; HLP, hyperlipidemia; Pr/Cr Ratio, protein/creatinine ratio; ANA, antinuclear antibody; CTX, cyclophosphamide; HCQ, hydroxychloroquine; MMF, mycophenolate mofetil; AZA, azathioprine; MTX, methotrexate; CsA, cyclosporine. Significant differences between the active and inactive groups are indicated using asterisks

^**,p<0.01;

^***,p<0.001;

^****,p<0.0001).

Table 2:

Clinical and Demographic Characteristics of RA Patients¹

Characteristics	Variables	All RA (N=36)	Active (N=21)	Inactive (N=15)
Age		55.0±12.0	55.5±12.1	53.5±12.6
Sex	Female	33 (91.7%)	19 (90.5%)	14 (93.3%)
	Male	3 (8.3%)	2 (9.5%)	1 (6.7%)
Race	African American	10 (27.8%)	6 (28.6%)	4 (26.7%)
	African Caribbean	1 (2.8%)	0 (0.0%)	1 (6.7%)
	More than one race- (Ecuadorean)	1 (2.8%)	1 (4.8%)	0 (0.0%)
	More than one race-(Peruvian)	1 (2.8%)	0 (0.0%)	1 (6.7%)
	Other-Mexico	4 (11.1%)	2 (9.5%)	2 (13.3%)
	Other-Nicaragua	1 (2.8%)	1 (4.8%)	0 (0.0%)
	Other- Puerto Rico	4 (11.1%)	2 (9.5%)	2 (13.3%)
	White	7 (19.4%)	5 (23.8%)	2 (13.3%)
	Other	7 (19.45)	4 (19.0%)	3 (20.0%)
Clinical Features	RF Positive	29 (80.6%)	16 (76.2%)	13 (86.7%)
	CCP Positive	35 (97.2%)	20 (95.2%)	15 (100.0%)
	CDAI	7.3±7.9	11.3±8.1****	1.7±2.0
	ESR	33.4±24.2	25.3±17.7***	51.3±23.0
	CRP	7.5±12.7	8.4±14.0	7.1±11.2
Medications	Prednisone	22(61%)	13(61.9%)	9(60.0%)
	Arava	4(11.1%)	2 (9.5%)	2 (13.3%)
	Humira	3 (8.3%)	3 (14.3%)	0 (0.0%)
	Methotrexate	26(72.2%)	17(80.9%)	9(60.0%)
	Enbrel	12 (33.3%)	5 (23.8%)	7 (46.7%)
	Orencia	2 (5.6%)	1 (4.8%)	1 (6.7%)
	Plaquenil	5(13.8%)	2 (9.5%)	3 (20.0%)
	Celebrex	1 (2.8%)	1 (4.8%)	0 (0.0%)
	Metoprolol	1 (2.8%)	1 (4.8%)	0 (0.0%)

^1:Values shown are mean ± SD or n (%). RA, Rhuematoid Arthritis; RF, Rhuematoid Factor; CCP, Cyclic Citrullinated Positive; CDAI, Clinical Disease Activity Index; ESR, Erythrocyte Sedimentation Rate; CRP, C Reactive Protein. Significant differences between the active and inactive groups are indicated using asterisks

^***,p<0.001;

^****,p<0.0001

Serum from 6-month-old females from four different mouse strains, C57BL/6 (N=4), BWF1 (N=5), NZM2410 (N=3), and MRL/lpr (N=3) were collected for metabolomic analysis. Blood collected from these mice was clarified by centrifugation at 2500 RPM for 10 minutes at room temperature. The samples were further processed using the MicroLab STAR® system (Hamilton Company) and stored at −80 ° C until further processing.

Sample preparation for UPLC-mass spectrometry:

Quality control standards were added to the serum samples prior to sample preparation. To extract metabolites, proteins in the serum samples were precipitated by mixing the samples with methanol. After adding methanol, serum samples were subjected to vigorous shaking at 650 strokes per minute for 2 minutes (using Genogrinder 2000) followed by centrifugation at 2500 rpm for 10 minutes. Supernatant was collected and the organic solvent was removed by lyophilization and reconstituted for further analysis.

Ultrahigh Performance Liquid Chromatography-Tandem Mass Spectroscopy (UPLC-MS/MS):

UPLC-MS/MS on dried samples was carried out using an LTQ Mass spectrometer with Electron spray ionization (ESI) as ion source and an ion trap mass analyzer by Metabolon (North Carolina). Two aliquots of dried samples were reconstituted in acidic and basic LC-compatible solvents. Extracts reconstituted in acidic conditions were gradient eluted in a C18 (2.1mm*100mm BEH column) column using 0.1% formic acid in water followed by 0.1% formic acid in methanol. Basic extracts were similarly eluted from the C18 column using 6.5mM ammonium bicarbonate in water followed by 6.5 mM ammonium bicarbonate in methanol. The LTQ mass spectrometer captured full scan spectra ranging from 99–1000m/z and exchanged polarities to scan both positive and negative ions. Metabolites in human and mouse serum samples were analyzed by comparison of the chromatographic readouts (retention time, ion spectra, and m/z value) to that of bradykinin peptide standards included in the metabolomics library. Quantitation of the observed peaks was achieved using area under curve.

Statistical analysis:

Statistical analysis was performed using R. Data was log transformed to minimize the effect of outliers. Welsh’s T test was used to measure statistical significance between the groups.

Multiple Reaction monitoring (MRM) assay:

The MRM assay is a novel target specific proteomic approach for the reliable quantitation of low abundant peptides in complex sample mixtures using a QQQ mass spectrometer (10–12). In the QQQ mass spectrometer, the first and third quadrupoles act as mass filters for the peptide and its specific fragment ion respectively, based on predefined specific m/z values, thus increasing the selectivity of the process. Unlike other mass spectrometric methods, the MRM assay does not record the full mass spectra, but narrows it to the selected m/z values corresponding to the peptide/fragment ion, thus increasing the sensitivity by at least two orders of magnitude (12). The MRM assay for this study was carried out at the Proteomics core facility at the UT Health Science Center, Houston, TX.

Standardization/Optimization:

Custom stable isotopically labeled AQUA peptides for bradykinin (RPPGFSPF) and bradykinin des arg 9 (RPPGFSP) were synthesized from Sigma Aldrich. These two peptides are referred to as BK and BK-des-arg-9 respectively, in this report. These isotopically labeled standards are identical in amino acid sequence and share all of the chemical properties of the endogenous peptides but differ only in mass. BK and BK-des-arg-9 AQUA peptides were tested for peptide purity by chromatography and sequencing by MS/MS before using them as internal reference in the assays. Specific m/z values for peptide and fragment ions were chosen and incorporated as mass filters in the QQQ mass spectrometer. Both the LC and MS parameters on the Agilent 6430 QQQ mass spectrometer were optimized based on standard peptide runs.

Sample preparation for MRM assay:

In order to quantitate the absolute concentration of bradykinin peptides in the serum samples, proteins were precipitated using methanol extraction. Serum samples were spiked with equal concentration of isotopically labeled BK nano peptide and BK-des-arg-9 internal standards (25ul of serum +0.5 picomole of BK nano peptide + 0.5 picomole of BK-des-arg-9) and mixed with methanol and vortexed. The sample mixture was then centrifuged at 2500 RPM for 10 minutes after which the supernatant was collected. Organic solvent in the supernatant was removed by lyophilization.

Peptide detection by MRM:

The sample extract was passed through a C18 column and eluents were injected into a QQQ mass spectrometer using an electron spray ionization source. In the QQQ mass spectrometer, the first and third quadrupoles act as mass filters while the second acts as a collision cell. This increases the specificity of peptide detection. Absolute levels of BK nano peptide and BK- des-arg-9 were measured using the included BK and BK-des-arg-9 standards as internal calibrators. Unlabeled peptides co-elute with labelled peptides and the signal to noise ratios of unlabeled peptides is similar to that of labeled peptides. This increases the detection of low abundance peptides with higher accuracy and ensures reproducible quantitation.

All MRM data was analyzed using the Agilent Triple Quad system software and in-house software for signal quantitation. The absolute concentration of the target peptide in the samples was deduced from the ratio of the signal intensity of the peptide of interest in the test serum sample to the signal intensity of the labeled isopeptide included for calibration.

Results:

Semi quantitative mass spectrometric screen reveals significant upregulation of BK-des-arg-9 in serum of SLE patients and murine lupus models

An initial metabolite screen was performed on sera from SLE patients (active N=10, inactive N=10) and healthy controls (N=9) using UPLC-MS/MS LTQ mass spectrometry (13). All active SLE patients in this initial metabolomic screen had nephritis (13). The demographics, co-morbidities and treatment history of the subjects used for this initial screen have previously been detailed in Table 1 of Wu, et al, 2012 (13). Briefly, SLE patients included in the metabolomic screen had a median age of 33.8 (range: 18–40), a median SLEDAI of 5 (range: 0–18), with 75% being females, 50% being Hispanics and 45% being African Americans (13). Inactive and active SLE patients showed significantly higher BK-des-arg-9 levels when compared to healthy controls. Interestingly, serum BK-des-arg-9 was elevated about 21-fold in SLE serum when compared to healthy controls with p<0.001 (Figure 1A) although no association with disease activity was noted. Interestingly, BK peptide was not detected in this assay, although it was screened for (13).

Figure 1: Measurement of BK-des-arg-9 in human and mouse serum samples using LC/MS and GC/MS.

A) BK-des-arg-9 levels were measured in healthy (N=9) and SLE serum samples (active SLE=10, inactive SLE=10) using an LC-MS-based metabolomic scan. B) A similar metabolomic approach was applied for measuring BK-des-arg-9 levels in 4 different mouse strains (C57BL/6 N=4 (control); BWF1 N=5; NZW2410 N=3; MRL/lpr N=3). Levels of BK-des-arg-9 were significantly increased in each of the 3 lupus strains when compared to controls. AU= arbitrary units from LC/MS. The mean line is shown. P < 0.05=*, P < 0.01=**, and P < 0.001=***.

Adapting the same protocol as above, serum from four different mouse strains, C57BL/6, BWF1, NZM2410, MRL/lpr, (all from 6-month old females) were examined for BK-des-arg-9 using UPLC-MS/MS and LTQ mass spectrometry. Compared to the C57BL/6 control strain, the other three strains are spontaneous lupus prone strains, all of which were proteinuric at the time of sampling. Consistent with the human metabolomics data, BK-des-arg-9 was also significantly elevated in the sera of lupus prone mouse strains, compared to the healthy C57BL/6 control (P=0.001) (Figure 1B).

Establishment of a novel MRM assay for the quantitative detection of BK and BK-des-arg-9 peptides:

Based on the promising observation that the pro-inflammatory BK-des-arg-9 peptide was increased in human and murine lupus, a more sensitive and quantitative mass spectrometry platform was adopted. A novel MRM assay was established to detect BK and BK-des-arg-9 in a larger cohort of samples. This assay allows for quantification since standards of isotopically labeled AQUA peptides, bradykinin (RPPGFSPF) and BK-des-arg-9 (RPPGFSP) are included in each run to generate standard curves, over 4 – 4000 fM range. As shown in Figure 2A and 2B, the elution chromatograms of the isotopically labeled standard peptides indicate the purity of the peptide, with an acquisition time of 4.28 and 4.82 minutes for BK and BK-des-arg-9, respectively. The MRM mass spectrometric relative abundance graph indicates the parent ion with +3 charge and most abundant peak of fragment ion with +2 charge. The dynamic ranges of the assays spanned over 3 orders of magnitude (4–4000 fm) with broad linearity as shown in Figure 2C and 2D. Increasing the concentration of the added standards correlated well with the readout on the mass spectrometer, as indicated by the calibration line in Figure 2C and 2D.

Figure 2: Ion chromatogram and MS/MS spectra of isotopes labelled BK and BK-des-arg-9 from human serum.

A, B) Illustration of the retention/acquisition time and MS/MS spectra of heavy isotope labelled BK and BK-des-arg-9. Retention times are measured as 4.248 and 4.821 seconds, respectively. The corresponding MS/MS spectra validate peptide identification by providing sequences of the peptides. C, D) Graphical representation of the broad linear response of the metabolites, demonstrated by calibration curves for BK and BK-des-arg-9, respectively.

Validation of elevated BK peptide in SLE serum using the MRM assay

In order to validate the initially observed increase in BK-des-arg-9 in lupus, levels of BK and BK-des-arg-9 were measured in additional SLE patients (N=37) and healthy controls (N=31), using the more quantitative MRM platform. All patients were age and ethnicity matched. 97% of the patients were African American or Hispanic with a median age of 32.4 years, as summarized in Table 1. As shown in Figure 3A and 3B, BK and BK-des-arg-9 were both significantly elevated in SLE patients when compared to healthy controls (p<0.0001). ROC curve analysis was performed to assess the accuracy of serum BK and BK-des-arg-9 in distinguishing SLE patients from healthy controls (Figure 3C). The ROC curve AUC values for BK-des-arg-9 and BK were 0.88 (P<0.0001) and 0.79 (P<0.0001), respectively, for distinguishing SLE patients from healthy controls. For the generated ROC curves, the cut off values for BK-des-arg-9 and BK were 48.5 nmol/L and 50.8 nmol/L for distinguishing SLE patients from healthy controls.

Figure 3: BK and BK-des-arg-9 concentration, sensitivity, and specificity in SLE and RA in human serum.

A, B) BK and BK-des-arg-9 levels, quantified using the MRM assay in healthy control (N=31), SLE (N=37), and RA (N=36) patients. C, D) ROC plots displaying the ability of BK and BK-des-arg-9 to distinguish SLE and RA from healthy controls, respectively. The solid line indicates BK-des-arg-9 and dotted line indicates BK. E) Dot plot representation of BK-des-arg-9: BK ratios in all 3 groups (healthy controls, SLE, and RA patients). F) ROC plot depicting the sensitivity and specificity of BK-des-arg-9: BK ratio in distinguishing the two disease groups (SLE and RA). Ur Pr/Cr ratio is the ratio of urine protein to urine creatinine. P < 0.05=*, P < 0.01=**, P < 0.001=***, P < 0.0001=****.

Circulating BK peptide in other rheumatic autoimmune diseases

As a disease control, sera from 36 RA patients were also interrogated for bradykinin peptides using the novel MRM platform. Detailed information about the demographics and clinical characteristics of these patients are provided in Table 2. 33 of the 36 RA patients were female, with a median age of 55 years, median ESR of 33.4 mm/h, median CRP of 7.5 mg/l and median disease activity index of 7.3, as summarized in Table 2. BK and BK-des-arg-9 were also significantly elevated in RA serum compared to healthy controls (Fig 3A, 3B). The ROC curve AUC values for BK and BK-des-arg-9 were 0.68 (P = 0.011), and 0.71 (P=0.003), respectively, for distinguishing RA patients from healthy controls (Fig. 3D). For the generated ROC curves, the cut off values for BK-des-arg-9 and BK were 1.8 nmol/L and 0.01 nmol/L for distinguishing RA from healthy controls. Serum BK and BK-des-arg-9 levels did not show any statistically significant difference between SLE and RA patients.

Serum BK-des-arg-9: BK ratios

Given that pro-inflammatory BK-des-arg-9 is derived from BK, we reasoned that BK-des-arg-9: BK ratios may be an interesting metric to examine. However, in SLE and RA patients, serum BK-des-arg-9: BK ratios did not show any statistically significant difference when compared to healthy controls (Fig 3E). The ROC AUC values for BK-des-arg-9: BK ratios in SLE patients vs. healthy controls and RA patients vs. healthy controls were 0.59 (p=0.17) and 0.52 (p = 0.76), respectively, as shown in Figure 3F. These results indicate that the BK-des-arg-9: BK ratio has poor accuracy in distinguishing SLE and RA patients from healthy controls.

Association and Correlation of serum BK and des-Arg9-BK levels with clinical parameters

To examine any possible relationships between serum BK and BK-des-arg-9 levels with clinical parameters in SLE and RA patients, association and correlation analyses were performed in both groups. As shown in Figure 4, BK and BK-des-arg-9 levels were similar in patients with inactive SLE and active SLE. Interestingly, BK-des-arg-9 levels were inversely correlated with rSLEDAI (Fig. 4D), but not other clinical parameters. Likewise, BK and BK-des-arg-9 levels were similar in patients with inactive RA and active RA and failed to show any significant correlation with disease activity index or ESR. (Fig. 5), Both the assayed metabolites were not significantly associated with the use of particular medications (data not shown). Thus, SLE patients dichotomized by prednisone, plaquenil, azathioprine, Cytoxan or MMF treatment did not differ in the levels of these metabolites. Likewise, RA patients dichotomized by prednisone, plaquenil, MMF or TNF inhibitor treatment did not differ in the levels of these metabolites (data not shown).

Figure 4: Clinical significance of serum BK and BK-des-arg-9 and its correlation with disease activity in SLE patients.

A, B) Dot plots illustrating the levels of BK and BK-des-arg-9 in three subject groups (healthy controls N= 31, inactive SLE, N= 16, and active SLE, N= 21). The mean line is shown. C, D) Pearson correlation of BK and BK-des-arg-9 concertation with renal SLE disease activity index (rSLEDAI). E, F) Pearson correlation of BK and BK-des-arg-9 levels with urine Pr/Cr ratio. Dashed line represents the 95% confidence interval and the solid line represents the correlation line. P < 0.05=*, P < 0.01=**, P < 0.001=***, P < 0.0001=****.

Figure 5: Clinical significance of serum BK and BK-des-arg-9 and its correlation with disease activity in RA patients.

A, B) Dot plots illustrating the levels of BK and BK-des-arg-9 in three subject groups (healthy controls, N= 31, inactive RA, N= 15, and active RA, N= 21). The mean line is shown. C, D) Pearson correlation of BK and BK-des-arg-9 level with clinical disease activity index (CDAI). E, F) Pearson correlation of BK and BK-des-arg-9 level with Erythrocyte Sedimentation Rate (ESR). Dashed line represents the 95% confidence interval and the solid line represents the correlation line. P < 0.05=*, P < 0.01=**, P < 0.001=***, P < 0.0001=****.

Discussion:

In the 1930’s Frey et al., observed a reduction in blood pressure upon injection of human urine into dogs. He attributed this change to a hypotensive substance in urine rather than its toxicity (14). As the pancreas was a rich source of this substance, it was named Kallikrein (Kallikreas, greek synonym for pancreas). In 1947, Werle et al., identified that kallikrein generates active bradykinin/kinins from inactive kininogens in plasma. Bradykinin was further investigated and reported to reduce blood pressure and cause the slow contraction of the gut and was hence named bradykinin. (brady= slow and kinin = move)(15).

Previously, increased levels of BK were reported in patients with acute autonomic neuropathy (16) and the authors suggested its association with SLE. Several studies reported the role of angiotensin converting enzyme (ACE) gene polymorphisms in SLE, but thus far no study has directly measured the levels of BK and BK-des-arg-9 in SLE or RA (17). ACE is one of the 3 enzymes that can metabolize BK, as discussed further below. Sheikh et al., reported increased expression of ACE and increased degradation rates of BK and BK-des-arg-9 in untreated SLE and RA patients when compared to patients with degenerative joint disease and healthy subjects (18). No significant difference was found in serum and synovial fluid levels of ACE in untreated RA patients and patients with polyarthritis (19). Polymorphisms in the ACE gene have been reported in RA patients in some studies (20), but not others (21, 22). Similarly in SLE patients, ACE polymorphisms have been associated with disease in certain ethnicities, associated with progression and severity of SLE, but not in other studies (23–27). The current study is the first to directly measure the levels of BK and BK-des-arg-9 peptides in the serum of SLE and RA patients.

MRM (multiple reaction monitoring) is a non-scanning method that utilizes QQQ mass spectrometry to measure low abundance analytes of interest in complex sample mixtures. Unlike other shotgun proteomic techniques full mass spectra is not obtained in MRM (9, 12). MRM can detect multiple peptides (multiplexing) in nanogram concentrations with high sensitivity and reproducibility. This multiplexing, and the ability to detect peptides in low concentration in complex mixtures has made MRM one of the most advanced MS techniques for its application in clinical trials (9, 10, 28–30). For example, MRM has been successfully used to detect 54 plasma proteins simultaneously in 1090 individuals by Garcia et al (31) while Charrier JP et al have monitored L–FABP in about 230 subjects using MRM (32). Although mass spectrometry-based MRM techniques have shown utility in multiple diseases, sample treatment, fractionation, requirement of expertise, and relative lack of automation have been major limitations for transferring this technology to routine clinical use. Although several studies have made significant progress in this transition (33, 34), MRM still needs to be tuned and adapted for routine clinical use in biomarker detection and diagnostics.

This is the first study to explore serum BK and BK-des-arg-9 levels in SLE and RA patients and their correlation with disease activity using two different platforms including the highly sensitive MRM platform. Our results indicate that serum BK-des-arg-9 levels are significantly elevated in both SLE patients and murine lupus models. BK-des-arg-9 also demonstrated good sensitivity and specificity in distinguishing SLE and RA patients from healthy controls. These findings are unlikely to be the consequence of the drugs the patients were on (e.g. ACE inhibitors) since similar findings were noted in murine model of lupus.

To explore possible reasons for the elevated BK-des-arg-9 levels in both rheumatic diseases, one has to consider the underlying metabolic pathway, as diagramed in Fig. 5. One reason that may account for this is increased metabolism of BK to BK-des-arg-9 by the enzyme kininase (Fig. 6). Indeed, previous studies have reported an increase in kininase activity in inflamed vascular tissues along with increase in B1 receptors in the aorta following inflammation (35). An alternate explanation for these observations may be decreased activity of ACE or APP, which are two enzymes that metabolize BK (Fig. 6). Reductions in ACE may be genetically encoded (36), or represent the consequence of ACE inhibitors being prescribed to the patients. However, the latter cannot be the explanation for the BK profiles noted in this study since patients on ACE inhibitors and patients not on ACE inhibitors did not differ in their BK or BK-des-arg-9 levels (data not shown); moreover, similar findings were seen in murine lupus, where medications are not at play. Clearly, further studies are warranted to unravel the mechanistic bases of these findings.

Figure 6: An overview of the Kallikrein-bradykinin pathway:

Kallikreins are serine proteases that catalyze the generation of bradykinin from kininogen. Bradykinin is further metabolized into BK-des-arg-9 by the kininase enzyme. Angiotensin converting enzyme (ACE) and aminopeptidase P (APP) convert active BK and BK-des-arg-9 to inactive peptides by cleaving peptide bonds at the 7–8 and 1–2 positions, respectively. The respective elevations of intermediates in this pathway in SLE and RA serum are also depicted.

Increased expression of BK-des-arg-9 plays a potential role in nociception and accumulation of inflammatory mediators at inflamed sites by binding to B1 receptors (37) (38, 39). Evidence from previous studies supports our current results, which shows increase in influx of neutrophils and mononuclear cells upon injection of BK-des-arg-9, while injection of its antagonists neutralizes the inflammatory effects of BK-des-arg-9 (40). We have recently shown that blocking the bradykinin B1 receptor ameliorates murine lupus nephritis (9), suggesting that elevated BK-des-arg9 in SLE may be driving inflammation by engaging the B1 receptors. Although we observed significant elevations in BK-des-arg-9 in disease vs controls, we did not see any difference in active versus inactive disease; in fact, there was a negative correlation with rSLEDAI. One scenario that could potentially account for this is increased deposition of BK-des-arg9 within the inflamed kidneys, as disease progresses, akin to the changes in the complement system. Clearly, this needs experimental verification.

This study has some limitations. One confounding factor is the increased percentage of females in the patient groups (85.6% among SLE patients and 91.7% among RA patients, compared to 61% among healthy controls). Although females in this study did not exhibit increased BK peptide levels compared to males, it would be important to plan future studies with larger, gender-matched cohorts, given literature reports suggesting that bradykinins and response to bradykinins my exhibit gender dimorphism (41). Also, assaying circulating and intracellular kininogens, kininases, APP and ACE simultaneously may provide further insight on why BK-des-arg-9 may be elevated in rheumatic diseases. Finally, ascertaining if BK-des-arg-9 exhibits elevated renal deposition in lupus nephritis patients also warrants investigation.

Key Points:

Serum BK-des-arg-9 is significantly elevated in mice and patients with lupus.

A similar but less pronounced increase is also noted in RA serum.

Increased conversion of BK to pro-inflammatory BK-des-arg-9 may mediate nephritis.