International Immunology, Vol. 21, No. 3, pp. 237–243doi:10.1093/intimm/dxn142

ª The Japanese Society for Immunology. 2009. All rights reserved.For permissions, please e-mail: journals.permissions@oxfordjournals.org

Serum DNase I, soluble Fas/FasL levels and cellsurface Fas expression in patients with SLE:a possible explanation for the lack of efficacy ofhrDNase I treatment

Elisa Tinazzi1,*, Antonio Puccetti2,3,*, Roberto Gerli4, Antonella Rigo5, Paola Migliorini6,Sara Simeoni1, Ruggero Beri1, Marzia Dolcino2, Nicola Martinelli1, Roberto Corrocher1 andClaudio Lunardi1

1Department of Clinical and Experimental Medicine, Section of Internal Medicine, University of Verona, 37134 Verona, Italy2Institute G. Gaslini, Department of Immunology, Genova, Italy3Department of Experimental Medicine, Section of Histology, University of Genova, Genova, Italy4Department of Clinical and Experimental Medicine, Section of Rheumatology, University of Perugia, Perugia, Italy5Department of Clinical and Experimental Medicine, Section of Hematology, University of Verona, Verona, Italy6Department of Internal Medicine, Section of Clinical Immunology, University of Pisa, Pisa, Italy

Keywords: DNase I, human recombinant DNase I, SLE, soluble Fas/FasL

Abstract

The objectives of the study are to evaluate DNase I serum levels and their correlation with soluble Fas(sFas) and soluble Fas ligand (sFasL) and with cell surface Fas expression in patients with systemiclupus erythematosus (SLE), thus contributing to the dysregulated apoptosis typical of the disease.The methods include the following: Serum DNase I levels in patients and in controls were detectedusing the dot blot method and quantified by densitometry; sFas and sFasL were quantified using anELISA system. Cell surface Fas expression was evaluated by FACS analysis. Apoptosis was studiedby means of internucleosomal DNA degradation using a commercially available kit. The results are asfollows: We found a significant difference in DNase I, sFas and sFasL serum levels between patientsand controls. Levels of DNase I <7.79 ng ml21 are more represented in patients with SLE. Active SLE isstrongly associated with high sFas levels and detectable sFasL. DNase I does not correlate with sFasor sFasL, whereas it correlates with T cell surface Fas expression that is higher in patients with activeSLE than in healthy controls. Finally, administration of exogenous human recombinant DNase(hrDNase) I to freshly isolated T cells up-regulates cell surface Fas expression and induces increasedsusceptibility to Fas-mediated apoptosis. In conclusion, our findings confirm that DNase I is low inSLE and suggest that it may play a role in apoptosis in SLE by regulating the surface expression ofthe cell death molecule Fas. This role may contribute to explain the inefficacy of hrDNase I in SLE,a treatment proposed for the ability of DNase I to remove DNA from auto-antigenic nucleoproteincomplexes.

Introduction

Systemic lupus erythematosus (SLE) is characterized by thepresence of anti-nuclear antibodies (ANAs) directed againstnaked DNA and nucleosomes. The etiology of SLE is un-known, but several studies suggest that increased produc-tion and/or inadequate clearance of nuclear DNA–proteincomplexes after cell death may initiate and/or propagate thedisease (1).

DNases are the primary enzymes involved in the metabolismand clearance of DNA. Low serum and urine DNase activity hasbeen observed in some patients with SLE (2–4), as well as inlupus-prone mice (5); indeed DNase I-deficient mice have beenshown to develop a lupus-like phenotype, with nephritis andserum auto-antibodies directed against nuclear antigens (6).Reduction or loss of DNase I activity, such as in case of gene

*These authors contributed equally to this study.

Correspondence to: C. Lunardi; E-mail: claudio.lunardi@univr.it Received 8 September 2008, accepted 8 December 2008

Transmitting editor: L. Moretta Advance Access publication 30 January 2009

by guest on Novem

ber 26, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

mutation of DNase I (7), may result in a high risk to produceANAs as a potential prerequisite to develop a SLE-like disease.

Besides the removal of DNA from auto-antigenic nucleopro-tein complexes to prevent immune stimulation, DNase I playsa role also in apoptosis (8–10). We have previously demon-strated that DNase I is responsible for internucleosomal DNAdegradation in human cells undergoing apoptosis in re-sponse to pharmacological stimuli (11) and that this endonu-clease behaves as a transcription factor that regulates cellsurface Fas expression (12). Moreover, exogenous humanrecombinant DNase (hrDNase) I is endocytosed into cellsup-regulating cell surface Fas expression (12).

Fas is an apoptosis signaling receptor belonging to thetumor necrosis factor (TNF) receptor family, while FasL is a typeII membrane protein of the TNF family, which behaves as a deathfactor, transducing death signals to Fas-positive cells. The hy-pothesis that perturbation in the Fas/FasL system may be re-sponsible for the development of autoimmunity has promptedstudies in patients with autoimmune disorders leading to contro-versial results (13–15). An abnormal regulation of the Fas/FasLapoptotic signaling pathway has been associated with thepathogenesis of SLE (16, 17). In particular, the extensive apo-ptosis of T lymphocytes (18), of bone marrow CD34+ cells(19), of neutrophils (20) and of monocytes/maturing macro-phages (21), as well as the resistance of autoimmune Th cellsdriving pathogenic auto-antibody production (22), seem to re-sult from impaired activation of Fas/FasL system.

In light of the results of our previous study showing a directlink between DNase I and fas gene expression (12), we de-cided first to evaluate DNase I serum levels in SLE patients,second to evaluate whether DNase I correlates with solubleFas (sFas) and soluble Fas ligand (sFasL) and third whetherDNase I correlates with cell surface Fas expression. More-over, we wanted to evaluate whether these molecules maybe useful in predicting either the development of SLE or theorgan involvement or in discriminating different clinicalstages of the disease.

Finally, we wanted to confirm whether exogenous hrDNase Iis able to up-regulate cell surface Fas expression and facilitateFas-mediated apoptosis in freshly isolated peripheral blood(PB) T cells obtained from SLE patients and from healthy sub-jects, as already shown in human cell lines (12).

Materials and methods

Patients

Seventy-three SLE patients fulfilling the revised criteria for SLE ofthe American College of Rheumatology (23) and 60 age- andsex-matched healthy controls were enrolled. Twenty-two patientswere in remission, 29 had moderate active and 22 had activeSLE according to the SLE disease activity index (24).

Blood samples were drawn from patients and controls af-ter informed oral or written consent.

Serum levels of DNase I

Patients’ and controls’ sera were blotted onto a nitrocellulosemembrane (Hybond-C Extra, Amersham Biosciences, Frei-burg, Germany), incubated overnight with a rabbit anti-DNase I antibody developed in our laboratory (11) and thenrevealed with a peroxidase-conjugated mouse anti-rabbit

IgG antibody (Sigma, Milan, Italy). DNase I level in serumsamples was calculated comparing the density of each dotwith dots containing known concentrations of hrDNase I(purchased from Hoffman–La Roche, Grenzach-Wyhlen,Germany) using a VersaDoc image analyzer apparatus (Bio-Rad, Hercules, CA, USA).

Measurement of human sFas and sFasL

Serum levels of sFas and sFasL were quantified using com-mercially available ELISA kits (Bender MedSystems, Wien,Austria) following the manufacturer’s instructions. The detec-tion limit was of 20 pg ml�1 and 0.1 ng ml�1 for sFas andsFasL, respectively.

Cells preparation and culture

Peripheral blood mononuclear cells (PBMC) were isolatedfrom freshly heparinized PB of 18 patients and 10 controlsby density-gradient centrifugation on Lymphoprep (NycomedPharma, Oslo, Norway) and analyzed by flow cytometry.

In four patients and two controls, PBMC were suspendedat 2 3 106 ml�1 in RPMI-1640 + 10% FCS incubated with 80ng ml�1 of hrDNase I for 96 h at 37�C. Then, they were ana-lyzed by flow cytometry.

Flow cytometry analysis

FITC-conjugated anti-CD3 (anti-CD3–FITC), R-PE-conjugatedanti-CD95/Fas (anti-CD95–PE) mAb or isotype/fluorochrome-matched controls were purchased from Becton Dickinson(San Jose, CA, USA). Data were collected on a FACSCalibur(Becton Dickinson) using CellQuest software and analyzedby FlowJo software (TreeStar).

Measurement of apoptosis

The extent of internucleosomal DNA fragmentation wasquantified using a commercially available kit (Roche Bio-chemical, Indianapolis, IN, USA) according to the manufac-turer’s instructions. The principle of this test is based on thedetection of mono- and oligonucleosomes in the cytoplasmicfractions of cell lysates. The enrichment of mono- and oligo-nucleosomes released into the cytoplasm is calculated bydividing the absorbance value obtained in the cells exposedto the apoptotic stimuli with the absorbance value of the un-treated cells (absorbance of sample cells/absorbance ofcontrol cells). The enrichment factor was used as an indexof apoptosis (apoptotic index) and is shown on the verticalaxis as mean 6 SD of triplicates.

For the induction of apoptosis, cells were incubated witheither the anti-Fas IgG3 mAb (Alexis, Launsen, Switzerland)or the recombinant human sFasL (Alexis) and cell deathassessed at several time points. To inhibit Fas-mediated ap-optosis, cells were incubated with the blocking anti-FasIgG2b mAb (Alexis).

Statistical analysis

Calculations were performed with SPSS 14 statistical pack-age. Quantitative data with a normal distribution wereexpressed as mean 6 SD and were analyzed with Student’st-test. DNase I and sFas, which presented a not-normal dis-tribution, were expressed as median with 25–75� percentile

238 DNase I and SLE

by guest on Novem

ber 26, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

intervals and were analyzed with non-parametric tests(Mann–Whitney or Kruskal–Wallis with all pairwise compari-sons). Qualitative data were analyzed with the Chi-squaretest or Fisher’s exact test when indicated. Because a highrate of subjects did not present a detectable level of sFasL,data are expressed as dichotomous variable (detectable ornot detectable). Low DNase I or sFas levels are defined asthose below the arbitrary threshold value of the 20th percen-tile in the control group.

A multiple logistic regression analysis that simultaneouslycontrolled DNase I, sFas, sFasL status and also sex andage was performed to evaluate the independent associationbetween these variables and SLE. The study population wasthen stratified into many groups on the basis of DNase, sFasand sFasL status and the association of these subgroupsand SLE was analyzed.

DNase I, sFas and sFasL status were also analyzed inSLE subjects divided on the basis of disease activity or renalinvolvement.

Correlations between DNase I, sFas, sFasL, cell surface Faslevels, erythrocyte sedimentation rate (ESR) and C reactiveprotein (CRP) were analyzed with Kendall’s tau test.

A value of P <0.05 was considered significant.

Results

Serum DNase I levels are low in SLE

The characteristics of the study population and the levels ofDNase I, sFas and sFasL are shown in Table 1. SerumDNase I levels were lower in SLE compared with controls(11.75 ng ml�1 in SLE patients versus 13.11 ng ml�1 in con-trol subjects; P = 0.002 by Mann–Whitney test). Moreover,considering low levels of DNase I (<7.79 ng ml�1: below thearbitrary threshold value of the 20th percentile in the healthycontrols), subjects with low DNase I levels were significantlymore represented in SLE patients than in controls [42.4 ver-sus 18.3%; P = 0.003 by Chi-square test; odds ratio (OR)

3.34 with confidence interval (CI) 95% 1.40–7.49]. Thesedata indicate that serum DNase I is reduced in patients withSLE when compared with healthy donors and that levels ofDNase I <7.79 ng ml�1 are associated with SLE.

sFas and sFas ligand levels in SLE

Also sFas levels were lower in SLE patients than in controlsubjects (sFas: 134.7 versus 211.7 pg ml�1; P = 0.002 byMann–Whitney test) and subjects with low sFas levels(<176.9 pg ml�1) were more represented in SLE patients(63 versus 18.3% of controls; P < 0.001 by Chi-square test;OR 7.22 with CI 95% 2.8–18.6).

Considering sFasL, 67.1% of patients and only 20% ofcontrols have a dosable sFasL (P < 0.001 by Chi-squaretest; OR 8 with CI 95% 3.07–20.88). Therefore, low sFas lev-els and dosable sFasL are typical features of SLE.

We then stratified the study population on the basis of lowsFas status and of dosable sFasL (Table 2). A highly signifi-cant asymmetry was observed (total Chi-square = 49.46,P < 0.001; Chi-square for linear trend = 43.04, P < 0.001;remaining Chi-square = 6.42, P = 0.04): remarkably, onlyone of the 73 SLE patients presented high levels of sFasand not dosable sFasL. On the other hand, only two of 60control subjects had low levels of sFas and dosable sFasL.If we consider the SLE diagnosis on the basis of the pres-ence of either low sFas levels or dosable sFasL, we havea test with high sensibility (98.1%) and low specificity(67.5%). If we consider the SLE diagnosis on the basis ofthe concomitant presence of both sFas low levels and dos-able sFasL, we have a test with high specificity (97.5%) andlow sensibility (31.5%).

When low DNase I status, low sFas status, dosable sFasLand also age and sex were simultaneously controlled ina multiple logistic regression analysis, low sFas levels(P < 0.001; OR 69.7 with CI 95% 9.2–526.2) and dosablesFasL (P = 0.001; OR 18.4 with CI 95% 3.1–111.1) maintaineda significant association with SLE, whereas this associationwas lost for low DNase I (P = 0.517; OR 1.76 with CI 95%0.32–9.76). Moreover, there was a significant direct correlationbetween sFas and sFasL (correlation coefficient 0.29;P = 0.003 by Kendall’s rank correlation), while there was nocorrelation between sFas or sFasL and DNase I. These dataindicate that, among the three variables studied (sFas, sFasLand DNase I), sFas and sFasL have a better ability thanDNase I in discriminating SLE patients from controls.

DNase I, sFas, sFasL and SLE disease activity

Next we evaluated whether the three variables studied(DNase I, sFas and sFasL) correlate with SLE serologicaland clinical parameters.

DNase I, sFas and sFasL levels did not correlate with ei-ther inflammation markers (ESR and CRP), presence or ab-sence of anti-double strand DNA antibodies and of renalinvolvement (data not shown).

SLE patients were divided into three groups on the basisof disease activity (quiescent, mild active and active): thelevels of sFas and sFasL were significantly different in thethree patients’ subsets. On the contrary, no significant differ-ence was found in DNase I levels (Table 3).

Table 1. General characteristics of the study population

Controls (n = 60) SLE subjects (n = 73)

Male/female 9/51 (19.1%) 7/66 (11%)Age (years) 38 6 10.9 40.5 6 12.9Disease activity

Quiescent — 22/73 (30.1%)Mild active — 29/73 (39.7%)Active — 22/73 (30.1%)

Renal involvement 25/73 (34.2%)ANA positive 0 65/73 (89%)Anti-nativeDNA+ 0 23/73 (31.5%)DNase I (ng ml�1)a 13.11 (8.99–23.2) 11.75 (3.69–18.53)b

Subjects with DNaseI level <7.79 ng ml�1

11/60 (18.3%) 31/73 (42.4%)c

sFas (pg ml�1)a 211.7 (183.7–262.4) 134.7 (91–276.6)b

Subjects with sFaslevel <176.9 pg ml�1

11/60 (18.3%) 46/73 (63%)c

Subjects with dosablesFasL

12/60 (20%) 49/73 (67.1%)c

aData are expressed as median with 25–75� percentile range.bP < 0.05 by Mann–Whitney test.cP < 0.05 by Chi-square test.

DNase I and SLE 239

by guest on Novem

ber 26, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

Remarkably, SLE patients with quiescent disease resulted tohave lower levels of serum sFas (108.5 pg ml�1; P < 0.05 byKruskal–Wallis all pairwise comparisons) when compared notonly with control subjects (211.7 pg ml�1) but also with SLEpatients with active disease (286.1 pg ml�1). No significantdifference was found between controls and SLE patients withactive disease; however, the latter group presented a trendtoward higher sFas levels. Low levels of sFas (<176.9 pgml�1) was therefore mainly a characteristic of SLE patientswith quiescent disease (21/22–95.5% versus 8/22–36.4% inSLE patients with active disease; P < 0.001 by Chi-squaretest). On the other hand, the proportion of subjects with dos-able sFasL increased significantly and progressively fromquiescent to active disease (18/22–81.8% versus 8/22–36.4% in SLE patients with quiescent disease; P = 0.011 byChi-square test) (Fig. 1). These data show that the combinedutilization of sFas and sFasL levels allows a better discrimi-nation among quiescent, mild active and active SLE.

DNase I levels, cell surface Fas expression and apoptosis

Since we have previously shown that over-expressed or ex-ogenously administered DNase I up-regulates Fas expres-sion in different cell lines, we analyzed the cell surfaceexpression of Fas (Fig. 2A) in a group of 28 subjects (18patients with active SLE and 10 controls). There was a mild

significant direct correlation between DNase I level and Fas-positive-CD3+ lymphocytes (correlation coefficient 0.275;P = 0.045 by Kendall’s rank correlation). Interestingly, the per-centage of Fas-positive-CD3+ lymphocytes was significantlyhigher in SLE patients than in controls (P < 0.05) (Fig. 2B).

Finally, we wanted to verify whether exogenous hrDNase Ican modulate Fas expression in freshly isolated cells. Forthis purpose, PBMC obtained from four patients with SLEand two healthy donors were incubated with hrDNase I andstained with anti-Fas mAb. Indeed, an increased expressionof Fas was detectable on CD3+ cells by FACS analysis afterincubation with hrDNase I in cells obtained from both SLEpatients and healthy donors (Fig. 3A). These data show thatexogeneously administered hrDNase I is able to modulateFas expression also in freshly isolated cells.

Since Fas/FasL-mediated pathway is particularly important inT cells undergoing apoptosis in SLE (18), we evaluatedwhether cells exposed to hrDNase I had an increased sus-ceptibility to Fas-mediated apoptosis. We observed thatmononuclear cells, obtained from the same subjects analyzedfor cell surface Fas induction and incubated with hrDNase Ias above described, showed an increased internucleosomalDNA degradation upon engagement of Fas using either ananti-Fas IgG3 mAb or recombinant FasL (Fig. 3B and C).

The apoptosis observed was truly Fas dependent sincethis phenomenon was not observed in the presence of ananti-Fas-blocking antibody (data not shown). These resultsshow that exogenously administered hrDnase I can up-regulateFas expression and induce Fas-mediated apoptosis also infreshly isolated cells derived from both normal donors andSLE patients.

Discussion

Clearance of nucleosomes from the body is crucial for the de-velopment of SLE; DNase I is responsible for nucleosome deg-radation and therefore low enzyme activity may contribute tothe anti-nuclear auto-antibody production typical of SLE (25,26). The decreased activity may be related to mutation in theDNase I gene as shown in 2 of 200 Japanese patients stud-ied (7). However, this mutation is not present in the three ma-jor ethnic groups: Caucasian, African and Asian (27). Single-nucleotide polymorphism (SNP) may affect or not the serumDNase I activity (28, 29) and from this effect may dependthe described association of a SNP with colorectal (30) andgastric carcinomas (31) and with myocardial infarction (32).

The measurement of serum DNase I activity may be influ-enced by a number of factors such as (i) the presence ofnatural inhibitors of the enzyme (actin), (ii) the presence ofanti-DNA antibodies able to bind the catalytic site of DNaseI and to interfere with DNase I activity (33) and (iii) the

Table 2. Distribution of sFas and sFasL status stratification subgroups in SLE and controls

High sFas andnot dosable sFasL

High sFas anddosable sFasL

Low sFas andnot dosable sFasL

Low sFas anddosable sFasL

Controls (n = 60) 40 (66.6%) 11 (18.3%) 7 (11.7%) 2 (3.3%)SLE (n = 73) 1 (1.4%) 26 (35.6%) 23 (31.5%) 23 (31.5%)

Total Chi-square = 49.46; P < 0.001; Chi-square for linear trend = 43.04; P < 0.001; remaining Chi-square = 6.42; P = 0.04.

Fig. 1. Distribution of subjects with low sFas levels (<176.9 pg ml�1;gray bars) and dosable sFasL (white bars) in controls and in SLEpatients divided on the basis of disease activity. *: Significantdifference from control group; P < 0.05 by Chi-square test. #:Significant difference from SLE patients with active disease; P < 0.05by Chi-square test. §: Significant difference from SLE patients withquiescent disease; P < 0.05 by Chi-square test.

240 DNase I and SLE

by guest on Novem

ber 26, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

measurement of other endonucleases activity, using commer-cially available kits. Because of these problems and becauseof our previous findings on the ability of exogenous DNase Imolecule to enter into living cells by endocytosis and to modu-late surface Fas expression, we decided to measure the DN-ase I protein level. We found that DNase I molecule isdecreased in SLE patients, as already reported for the enzymeactivity (2–4). Indeed, a DNase I level <7.79 ng ml�1 is morerepresented in SLE patients, with an OR of above threetimes. However, DNase I levels did not correlate with organinvolvement or disease activity, as already reported (4), andwith sFas and sFasL. Increased levels of sFas and detect-able sFasL were typical of active SLE, in line with the aber-rant levels of apoptosis present in the active phase of thedisease. Indeed, our results are in line with previous reportsdescribing that in active lupus serum levels of both sFasand sFasL are increased compared with those of patientswith inactive disease and correlate with over-expression ofcell surface Fas by T cells and with T cell apoptosis (34, 35).

Serum DNase I correlates with surface expression of Fasin PB T lymphocytes obtained from patients with active SLE.

Most importantly, we confirmed the ability of hrDNase I to enterinto freshly isolated Tcells and to increase surface Fas expres-sion in PB T cells as already shown in the erythroleukemia cellline K562 (12). A possible explanation for the direct correla-tion between DNase I levels and cell surface Fas expressionin SLE patients is that circulant DNase I enters into the cellsthrough the binding to the cation-independent mannose 6-phosphate/insulin-like growth factor (CI-MPR) receptor andinduces surface expression of the death receptor Fas, behav-ing as a transcription factor for the fas gene. In this way, DN-ase I would contribute to the pathogenesis of SLE not onlythrough removal of DNA from auto-antigenic nucleoproteincomplexes but also by modulating Fas expression and ulti-mately cell apoptosis. Indeed, the results obtained incubatingPBMC from patients and controls with hrDNase I on the ex-pression of Fas and on increased susceptibility to apoptosissuggest that such a mechanism may be present also in vivo.

hrDNase I has been proposed as a therapeutic agent inSLE (26, 36); however, no changes both in clinical mani-festations and in serum markers of disease activity havebeen observed in humans. Moreover, mice over-expressingDNase I showed a reduction in auto-antibodies but were notprotected from developing lupus nephritis (37).

We believe that the lack of effect of DNase I treatment in SLEmay be related to its internalization into cells and to the in-creased expression of cell surface Fas eventually leading to in-creased Fas-mediated apoptosis. Therefore, we think thatDNase I is not a beneficial therapy in SLE because of its abilityto increase cell susceptibility to apoptosis. On the contrary, thistreatment could be detrimental, since CI-MPR is an endocyto-sis receptor ubiquitously expressed in any cell types and par-ticularly abundant in phagocytes. Therefore, DNase I deliverywithin cells like neutrophils and CD34+ cells may even result inexacerbation of some disease features like neutropenia andleukopenia (19–21). This aspect would be useful in cancertherapy since DNase I increases tumor cells’ susceptibility toapoptosis induced by chemotherapeutical agents (11). Suchhypothesis is sustained by the association of a SNP with lowDNase I activity with particular types of carcinoma (30, 31).

In conclusion, our findings show that DNase I enzymelevel is low in patients with SLE and that it can be consid-ered a marker of the disease but not an indicator of diseaseseverity. Moreover, our data using PBMC from patients and

Fig. 2. Surface expression of CD95/Fas on T lymphocytes from SLEand normal subjects. Panel (A): PBMC were labeled with anti-CD3–FITC and anti-CD95–PE mAbs. Representative histograms of threeSLE patients (red) and three controls (blue) are shown. Samples wereanalyzed by FlowJo software (TreeStar). Panel (B): percentage ofCD3+ Fas+ cells in peripheral blood of 18 SLE patients and (redcircles) and 10 healthy controls (blue circles). The mean values 6SDs were 36 6 19 and 24 6 10, respectively (P = 0.046).

Table 3. DNase I, sFas and sFasL levels in SLE patients divided on the basis of disease activity

Controls (n = 60) SLE (n = 73)

Quiescent (n = 22) Mild active (n = 29) Active (n = 22)

DNase I (ng ml�1)a 13.11 (8.99–23.2) 12.3 (6.6–18.5) 7.5 (2.9–18.3) 12.1 (3.5–18.1)Subjects with DNase I <7.79 ng ml�1 9/47 (19.1%) 7/22 (31.8%) 15/29 (51.7%) 7/22 (31.8%)sFas (pg ml�1)a,b 211.7 (183.7–262.4) 108.5 (78.6–129.8)c 154.3 (88.9–251.4) 286.1 (120.8–317.1)Subjects with sFas<176.9 pg ml�1d 12/60 (20%) 21/22 (95.5%)e 18/29 (62%)e 8/22 (36.4%)Subjects with dosable sFasLd 12/60 (20%) 8/22 (36.4%) 22/29 (75.9%)e 18/22 (81.8%)e

aData are expressed as median with 25th–75th percentile range.bP < 0.001 by Kruskal–Wallis.cSignificant difference from controls (P < 0.001) and from SLE patients with active disease (P = 0.03) by Kruskal–Wallis all pairwise comparisons(Dwass–Steel–Chritchlow–Fligner).dP < 0.001 Chi-square test.eSignificant difference from control group; P < 0.05 by Chi-square test.

DNase I and SLE 241

by guest on Novem

ber 26, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

controls confirm previous data obtained with cell lines on theability of DNase I to modulate Fas expression and cell sus-ceptibility to apoptosis. This finding, together with the meas-urements of DNase I and Fas expression ex vivo, suggeststhat a similar mechanism may occur also in vivo, providing

one possible explanation for the lack of efficacy of hrDNaseI in the treatment of SLE.

Abbreviations

ANA anti-nuclear antibodyCI confidence intervalCI-MPR cation-independent mannose 6-phosphate/insulin-like

growth factorCRP C reactive proteinESR erythrocyte sedimentation ratehrDNase human recombinant DNaseOR odds ratioPB peripheral bloodPBMC peripheral blood mononuclear cellssFAS soluble FassFASL soluble Fas ligandSLE systemic lupus erythematosusSNP single-nucleotide polymorphismTNF tumor necrosis factor

References

1 Carroll, M. C. 1999. The lupus paradox. Nat. Genet. 19:3.2 Chitrabamrung, S., Rubin, R. L. and Tan, E. M. 1981. Serum

deoxyribonuclease I and clinical activity in systemic lupuserythematosus. Rheumatol. Int. 1:55.

3 Tew, M. B., Johnson, R. W., Reveille, J. D. and Tan, F. K. 2001. Amolecular analysis of the low serum deoxyribonuclease activity inlupus patients. Arthritis Rheum. 44:2446.

4 Sallai, K., Nagy, E., Derfalvy, B., Muzes, G. and Gergely, P. 2005.Antinucleosome antibodies and decrease deoxyribonucleaseactivity in sera of patients with systemic lupus erythematosus.Clin. Diagn. Lab. Immunol. 12:56.

5 Macanovic, M. and Lachmann, P. J. 1997. Measurement ofdeoxyribonuclease I (DNase) in the serum and urine of systemiclupus erythematosus (SLE)-prone NZB/NZW mice by a new radialenzyme diffusion assay. Clin. Exp. Immunol. 108:220.

6 Napirei, M., Karsunky, H., Zevnik, B., Stephan, H., Mannherz, H.G. and Moroy, T. 2000. Features of systemic lupus erythematosusin DNase I deficient mice. Nat. Genet. 25:177.

7 Yasutomo, K., Horiuchi, T., Kagami, S. et al. 2001. Mutation ofDNase I in people with systemic lupus erythematosus. Nat. Genet.28:313.

8 Polzar, B., Peitsch, M. C., Loos, R., Tschopp, J. and Mannherz, H.G. 1993. Overexpression of DNase I transfected into COS-cells: itsdistribution during apoptotic cell death. Eur. J. Cell Biol. 62:397.

9 Polzar, B., Zanotti, S., Stephan, H. et al. 1994. Distribution ofdeoxyribonuclease I in rat tissue and its correlation to cellularturnover and apoptosis. Eur. J. Cell Biol. 64:200.

10 Peitsch, M. C., Polzar, B., Stephan, H. et al. 1993. Characterizationof the endogenous deoxyribonuclease involved in nuclear DNAdegradation during apoptosis (programmed cell death). EMBO J.12:371.

11 Oliveri, M., Daga, A., Cantoni, C., Lunardi, C., Millo, R. andPuccetti, A. 2001. DNase I mediates internucleosomal DNAdegradation in human cells undergoing drug-induced apoptosis.Eur. J. Immunol. 31:743.

12 Oliveri, M., Daga, A., Lunardi, C., Millo, R. and Puccetti, A. 2004.DNase I behaves as a transcription factor which modulates Fasexpression in human cells. Eur. J. Immunol. 34:273.

13 Jodo, S., Pidiyar, V. J., Xiao, S. et al. 2005. Fas ligand (CD178)cytoplasmic tail is a positive regulator of Fas ligand-mediatedcytotoxicity. J. Immunol. 174:4470.

14 Knipping, E., Krammer, P. H., Onel, K. B., Lehman, T. J., Mysler, E.and Elkon, K. B. 1995. Levels of soluble Fas/APO-1/CD95 insystemic lupus erythematosus and juvenile rheumatoid arthritis.Arthritis Rheum. 38:1735.

15 Mc Nally, J., Yoo, D. H., Drappa, J. et al. 1997. Fas ligandexpression and function in systemic lupus erythematosus. J.Immunol. 159:4628.

Fig. 3. Up-regulation of surface Fas expression and increasedsusceptibility of CD3+ lymphocytes to apoptosis following exoge-neous administration of hrDNAse I. Panel (A): Mean percentage 6 SDof Fas-expressing CD3+ cells before (white bars) and after incubationwith 80 ng of hrDNase I (gray bars). Bars 1–4: cells derived from fourSLE patients. Bars 5 and 6: cells obtained from healthy donors. Meanvalues of four independent experiments performed on each individualsubject are shown. PBMC from patients and controls were stainedwith anti-CD3–FITC and anti-CD95/Fas–PE mAbs. Panels (B and C):extent of internucleosomal DNA fragmentation before (white bars)and after (gray bars) incubation with hrDNase I. Bars 1–4: cellsobtained from SLE patients; bars 5 and 6: cells obtained from healthydonors. Data represent the means 6 SDs of triplicate samples ofthree independent experiments (x axis). Panel (B): apoptosis inducedby an anti-Fas mAb. Panel (C): apoptosis induced by human sFasL.

242 DNase I and SLE

by guest on Novem

ber 26, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

16 Courtney, P. A., Crockard, A. D., Williamson, K., McConnell, J.,Kennedy, R. J. and Bell, A. L. 1999. Lymphocyte apoptosis insystemic lupus erythematosus: relationships with Fas expression,serum soluble Fas and disease activity. Lupus 8:508.

17 Nozawa, K., Kayagaki, N., Tokano, Y., Yagita, H., Okumura, K. andHasitomo, H. 1997. Soluble Fas and soluble Fas ligand inrheumatic diseases. Arthritis Rheum. 40:1126.

18 Xue, C., Lan-Lan, W., Bei, C., Jie, C. and Wei-Hua, F. 2006.Abnormal Fas/FasL and caspase-3-mediated apoptotic signalingpathways of T lymphocyte subset in patients with systemic lupuserythematosus. Cell. Immunol. 239:121.

19 Papadaki, H. A., Boumpas, D. T., Gibson, F. M. et al. 2001.Increased apoptosis of bone marrow CD34(+) cells and impairedfunction of bone marrow stromal cells in patients with systemiclupus erythematosus. Br. J. Haematol. 115:167.

20 Courtney, P. A., Crockard, A. D., Williamson, K., Irvine, A. E.,Kennedy, R. J. and Bell, A. L. 1999. Increased apoptoticperipheral blood neutrophils in systemic lupus erythematosus:relations with disease activity, antibodies to double stranded DNA,and neutropenia. Ann. Rheum. Dis. 58:309.

21 Shoshan, Y., Shapira, I., Toubi, E., Frolkis, I., Yaron, M. andMerovach, D. 2001. Accelerated Fas-mediated apoptosis ofmonocytes and maturing macrophages from patients with systemiclupus erythematosus: relevance to in vitro impairment of interactionwith iC3b-opsonized apoptotic cells. J. Immunol. 167:5963.

22 Xu, L., Zhang, L., Yi, Y., Kang, H. K. and Datta, S. K. 2004. Humanlupus T cells resist inactivation and escape death by upregulatingCOX-2. Nat. Med. 10:411.

23 Tan, E. M., Cohen, A. S., Fries, A. et al. 1982. The 1982 revisedcriteria for the classification of systemic lupus erythematosus.Arthritis Rheum. 25:1271.

24 Bombardier, C., Gladman, D. D., Urowitz, M. B., Caron, D. andChang, C. H. 1992. Derivation of SLEDAI: a disease activity indexfor lupus patients. Arthritis Rheum. 35:630.

25 Tsukumo, S. and Yasutomo, K. 2004. DNase I in the pathogenesisof systemic lupus erythematosus. Clin. Immunol. 113:14.

26 MartinezValle, F., Balda, E., Ordi-Ros, J. and Vilardell-Tarres, M. 2008.DNase I and systemic lupus erythematosus. Autoimmun. Rev. 7:359.

27 Shin, H. D., Park, B. L., Kim, L. H., Lee, H. S., Kim, T. Y. and Bae,S. C. 2004. Common DNase I polymorphism associated withautoimmunity production among systemic lupus erythematosuspatients. Hum. Mol. Genet. 13:2343.

28 Fujihara, J., Takatsuka, H., Kataoka, K., Xue, Y. and Takeshita, H.2007. Two deoxyribonuclease I gene polymorphisms and corre-lation between genotype and its activity in Japanese population.Leg. Med. 9:233.

29 Bodano, A., Gonzales, A., Ferreiros-Vidal, I. et al. 2006.Association of a non-synonymous single-nucleotide polymor-phism of DNASEI with SLE susceptibility. Rheumatology 45:819.

30 Tsutsumi, S., Takeshita, H., Yasuda, T., Kuwano, H. and Kishi, K.2000. Association of DNase I phenotype 2 with colorectalcarcinoma in Japanese populations. Cancer Lett. 159:109.

31 Tsutsumi, S., Asao, T., Nagamachi, Y., Nakajiama, T., Yasuda, T.and Kishi, K. 1998. Phenotype 2 of deoxyribonuclease I may beused as a risk factor for gastric carcinoma. Cancer 82:1621.

32 Kumamoto, T., Kawai, Y., Arakawa, K. et al. 2006. Association ofGln222Arg polymorphism in the deoxyribonuclease I (DNase I)gene with myocardial infarction in Japanese patients. Eur. Heart J.27:2081.

33 Puccetti, A., Madaio, M. P., Bellese, G. and Migliorini, P. 1995.Anti-DNA antibodies bind DNase I. J. Exp. Med. 181:1797.

34 Silvestris, F., Williams, R. C., Calvari, N. et al. 2002. Serumelevations of soluble Fas (CD95(apo-I) concur in deregulatingT cell apoptosis during active lupus disease. Clin. Exp. Med. 2:13.

35 Silvestris, F., Granello, D., Tucci, M., Cafforio, P. and Dammacco, F.2003. Enhancement of T cell apoptosis correlates with in-creased serum levels of soluble Fas (CD95/Apo-I) in active lupus.Lupus 12:8.

36 Davie, J. C., Manzi, S., Yarboro, C. et al. 1999. Recombinanthuman DNase I (rhDNase I) in patients with lupus nephritis. Lupus8:68.

37 Manderson, A. P., Carlucci, F., Lachmann, P. J. et al. 2006. The invivo expression of actin/salt-resistant hyperactive DNase I inhibitsthe development of anti-ssDNA and anti-histone autoantibodies ina murine model of systemic lupus erythematosus. Arthritis Res.Ther. 8:R68.

DNase I and SLE 243

by guest on Novem

ber 26, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from