Normal circulating serum amyloid P component concentration in systemic sclerosis




The observation of reduced circulating concentrations of the constitutive plasma pentraxin protein, serum amyloid P component (SAP), in serum samples obtained from a small number of patients with systemic sclerosis (SSc) has been reported as confirmation of an antifibrotic role of this protein. Because neither sustained SAP depletion in humans nor SAP deficiency in mice is associated with fibrosis, we sought to establish rigorously the serum SAP concentration in well-characterized patients with SSc.


Serum concentrations of SAP were measured by electroimmunoassay in a cross-sectional cohort of 20 patients with diffuse cutaneous SSc and 12 patients with limited cutaneous SSc, and in a separate 12-month longitudinal cohort of 13 patients with diffuse disease and 37 patients with limited disease. The extent and severity of disease were characterized in detail at the time of serum sampling. Serum concentrations of the classic acute-phase reactants, C-reactive protein and serum amyloid A protein, were measured by immunonephelometric assays.


SAP values were entirely within the normal range, regardless of the extent and severity of disease, apart from a very few isolated raised values associated with acute intercurrent complications causing major acute-phase responses.


We observed no reduced circulating concentrations of SAP in patients with SSc, nor any evidence of an association between SAP levels and the extent or severity of fibrosis.

In 2003, Pilling et al (1) reported that serum amyloid P component (SAP) inhibits fibrocyte differentiation in vitro, that patients with systemic sclerosis (SSc; scleroderma) have low circulating SAP concentrations, and that a reduced quantity of SAP may contribute to the pathogenesis of fibrosis in scleroderma. However, these suggestions are neither supported robustly by the published observations nor are they consistent with established knowledge of SAP.

SAP and C-reactive protein (CRP) are closely related plasma proteins that comprise the highly conserved pentraxin family of homopentameric molecules that have specific calcium-dependent ligand-binding properties and which belong to the lectin fold superfamily (2, 3). Human CRP is the classic nonspecific acute-phase reactant, while human SAP is a stable constitutive plasma protein, the level of which does not increase during the early acute-phase response but may rise modestly during chronic inflammation (4). Despite much information about the properties and behavior of these proteins in humans and other species, neither their normal physiologic functions nor their actual roles in the pathophysiology of disease are known for certain. No deficiency state of either protein in humans or even any structural polymorphism has yet been reported, and the glycan moiety of human SAP is remarkably invariant (5).

No CRP-knockout mouse has been described, but the SAP-knockout mouse is fertile, healthy, and has a normal life span (6). The main phenotypic features of SAP-deficient mice are as follows: 1) impaired innate immunity against certain pathogens to which SAP does not bind and enhanced resistance to bacterial pathogens to which SAP binds, reflecting the fact that SAP is a potent antiopsonin (7), and 2) reduced and retarded deposition of AA amyloid induced by chronic inflammation (6). The increased incidence of antinuclear autoimmunity observed in SAP-deficient C57BL/6 mice (8) appeared consistent with the avid in vitro and in vivo binding of human SAP to DNA and chromatin (for review, see ref.2) but was actually attributable to chimeric chromosome 1 genetic effects unrelated to SAP deficiency (9).

Importantly, in the present context, there is no unusual fibrosis or other manifestation related to extracellular matrix or fibroblast function in mice without SAP. Furthermore, patients who received the novel SAP-depleting drug, R-1-[6-[R-2-carboxy-pyrrolidin-1-yl]-6-oxo-hexanoyl]pyrrolidine-2-carboxylic acid (CPHPC) (10), for as long as 2 years, during which time their plasma SAP concentration was persistently reduced by >90% (i.e., to 1–2 mg/liter), also showed no sign of fibrosis or any of the other manifestations of SSc.

Here, we report finding no evidence of reduced circulating SAP concentrations in a large group of well-characterized patients with SSc, nor did we observe any relationship between SAP values and disease extent or activity in 2 separate studies, one cross-sectional (32 patients) and one longitudinal followup study (50 patients).


Patients and sera.

Serum samples were obtained from outpatients (Table 1) attending the Royal Free Hospital, all of whom fulfilled the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) preliminary criteria for the classification of SSc (11). All patients gave informed consent in accordance with the Declaration of Helsinki, and the study was approved by the research ethics committee of the Royal Free Hampstead NHS Trust. Disease duration was defined as the time from onset of the first non–Raynaud's disease manifestation of SSc. Skin sclerosis was assessed using a modified Rodnan skin thickness score (12), by physicians experienced in the technique, concurrently with blood sampling. Major visceral complications were sought according to current standard practice, including pulmonary function testing, electrocardiography, Doppler echocardiography, creatinine clearance, and high-resolution computed tomography (CT) scanning or right heart catheterization if indicated. Pulmonary hypertension was defined as a mean pulmonary artery pressure >25 mm Hg at rest or >30 mm Hg during exercise. Lung fibrosis was detected by high-resolution CT, renal involvement was determined by creatinine clearance of <60 ml/minute or a history of scleroderma renal crisis, gastrointestinal tract involvement was identified by typical history or abnormal results of investigations prompted by symptoms, and skeletal muscle involvement was determined by a creatinine kinase level >4-fold the upper limit of normal. Control serum samples were obtained from 13 age- and sex-matched healthy volunteers.

Table 1. Demographic and clinical features of the study populations*
 Limited cutaneous SScDiffuse cutaneous SScHealthy controls
  • *

    Except where indicated otherwise, values are the median (range). The cross-sectional study comprised 12 patients with limited cutaneous systemic sclerosis (lcSSc), 20 patients with diffuse cutaneous SSc (dcSSc), and 13 healthy control subjects. The longitudinal cohort comprised 37 patients with lcSSc and 13 patients with dcSSc. Visceral involvement was determined as moderate to severe, by standard criteria. NA = not applicable.

  • Summation of physical examination ratings from 17 different skin sites, each assessed on a scale of 0–3 (uninvolved to severe) for skin thickness (maximum possible score 51).

  • Three samples from each patient were analyzed at least 6 months apart.

Cross-sectional cohort   
 Age, years55 (38–69)56 (29–65)34 (22–60)
 No. of women/no. of men11/115/510/3
 Disease duration, years10 (5–20)2 (0.5–15)NA
 Modified Rodnan skin thickness score5 (3–7)28 (8–38)NA
 Visceral involvement, no. (%)   
  Gastrointestinal5 (42)9 (45)NA
  Lung4 (33)8 (40)NA
  Pulmonary arterial hypertension4 (33)4 (20)NA
  Renal1 (8)5 (25)NA
Longitudinal cohort   
 Age, years56 (41–80)52 (20–70) 
 No. of women/no. of men32/59/4 
 Disease duration, years3.5 (0.5–20)2 (0.5–10) 
 Peak modified Rodnan skin thickness score6 (3–14)27 (8–38) 
 Visceral involvement, no. (%)   
  Gastrointestinal20 (54)5 (38) 
  Lung4 (11)6 (46) 
  Pulmonary arterial hypertension2 (5)5 (38) 
  Renal1 (3)1 (8) 

Protein assays.

SAP concentrations, in samples stored frozen (−20°C) until used, were measured by electroimmunoassay (180 V constant for 6 hours at 15°C), using monospecific polyclonal sheep anti-human SAP antibodies in 1% weight/volume agarose gels (Indubiose A37; BioSepra, Borehamwood, UK) and calibrated with standards of normal serum spiked with isolated pure SAP, as described previously (13). Standard curves (by nonlinear curve fitting) were produced in Origin version 7.0 (OriginLab, Northampton, MA), with calibrators from 8–126 mg/liter and assay sensitivity increased by inclusion of a zero-concentration standard to allow extrapolation of values from 1–8 mg/liter. CRP and serum amyloid A (SAA) levels were determined by particle-enhanced immunonephelometric assays on the Dade-Behring BNII autoanalyzer (Dade-Behring, Milton Keynes, UK), as previously described (14). Only SAP and CRP levels were assayed in the longitudinal study.

Statistical analysis.

The significance of differences between groups was tested by one-way analysis of variance, and pairwise group comparisons were performed using t-tests and Mann-Whitney tests. Correlation was sought using Pearson's or Spearman's coefficients as appropriate. P values less than or equal to 0.05 were considered significant.


In a cross-sectional study of 12 patients with limited cutaneous SSc (lcSSc) and 20 patients with diffuse cutaneous disease (dcSSc), almost all SAP concentrations were within the normal range seen in healthy subjects, with no abnormally low values (Figure 1A). As expected, increased SAP concentrations were present in samples from 3 patients with active intercurrent conditions (13) (1 patient with lcSSc had arthritis, 1 patient with lcSSc had digital ulceration, and 1 patient with dcSSc had laryngeal carcinoma), in whom there were major acute-phase responses for both CRP and SAA (Figures 1B and C). The majority of patients with SSc were women (Table 1), reflecting the usual sex distribution in this disease, but the results for men and women were analyzed together, because of the small number of men.

Figure 1.

Circulating concentrations of A, serum amyloid P component (SAP), B, C-reactive protein (CRP), and C, serum amyloid A protein (SAA) in a cross-sectional cohort of 12 patients with limited cutaneous systemic sclerosis (lcSSc) and 20 patients with diffuse cutaneous SSc (dcSSc) compared with 13 healthy controls (HCs). SAP concentrations in the 3 groups were not significantly different (P = 0.19), and there was no correlation between the SAP concentration and either the CRP level (ρ = 0.14, P = 0.47) or the SAA concentration (ρ = 0.19, P = 0.31). As expected, concentrations of CRP and SAA showed positive correlation (P < 0.0001 by one-way analysis of variance), and patients with dcSSc had SAA values that were significantly elevated (P = 0.027) compared with those in normal controls, reflecting an acute-phase response associated with disease complications. Each point represents a single measurement. Horizontals lines show the medians.

In a study conducted to establish a normal reference interval for SAP concentrations in the serum of 500 healthy adult individuals (13), SAP concentrations were significantly lower in normal women (mean ± SD 24 ± 8 mg/liter [range 8–55]; n = 274) than in normal men (32 ± 7 mg/liter [range 12–50]; n = 226). In our study, the values were as follows: for patients with lcSSc, median 22 mg/liter (mean ± SD 29 ± 15 mg/liter [range 10–62]); for patients with dcSSc, median 33 mg/liter (mean ± SD 37 ± 19 mg/liter [range 15–92]). Apart from the few elevated values that were associated with substantial complications, there was no significant correlation (P > 0.05) between SAP values and the modified Rodnan skin thickness score (ρ = 0.40), disease duration (ρ = 0.23), specific organ involvement (ρ = 0.25), or age (ρ = 0.64), nor between the SAP concentration and either the CRP level (ρ = 0.14) or the SAA concentration (ρ = 0.19). Samples from a small group of healthy control subjects (10 women and 3 men; median age 34 years [range 22–60 years]) were tested in the same assays (Figure 1), and the results (median 25 mg/liter, mean ± SD 28 ± 9 mg/liter [range 14–44]) were typical of those in the larger series (13, 15, 16) on which the reference ranges for these analytes are based.

In a longitudinal study in a separate cohort, comprising up to 3 serial samples obtained over a 12-month period from each of 37 patients with lcSSc and 13 patients with dcSSc, the SAP values (median 36 mg/liter [range 13–72]) were within the normal range and showed no consistent pattern within individuals over time. There was no association between SAP values and the modified Rodnan skin thickness score (Figure 2), a well-validated measure of fibrosis specific for SSc, nor with validated measures of internal organ involvement.

Figure 2.

No correlation of serum amyloid P component (SAP) concentrations and the extent of skin fibrosis, as assessed by the modified Rodnan skin thickness score, in a longitudinal cohort of patients with systemic sclerosis (SSc). A, Results of assessments of all patients for whom SAP concentrations and concurrent skin scores were available (ρ = 0.151, P = 0.137). B, Results of assessments of patients with diffuse cutaneous SSc for whom SAP concentrations and concurrent skin scores were available. Each point represents the individual value from each serial sample collected from each patient over a 12-month period of disease activity; all patients exhibited widespread variability in the skin score (ρ = −0.079, P = 0.732). Separate plots for men and/or women and for patients with limited cutaneous SSc also showed no correlation with SAP concentrations.


We did not detect reduced levels of circulating SAP in patients with SSc, nor did we observe any association between SAP concentrations and the extent or severity of SSc. Our results do not substantiate the limited preliminary observations of Pilling et al (1) based on single serum samples obtained from 15 patients with SSc and 10 patients with mixed connective tissue disease, about whom no clinical data were provided other than the information that they fulfilled the 1987 revised ACR criteria.

Neither SAP-deficient mice (6) nor patients with sustained SAP depletion produced by CPHPC treatment (10) develop increased fibrosis. Furthermore, several features of the report by Pilling et al (1) cast doubt on their identification of SAP as an inhibitor of fibrocyte differentiation.

First, they reported ∼90% inactivation of the factor by heating serum at 56°C for 30 minutes, whereas SAP is absolutely stable under such conditions. The most exquisitely sensitive test for protein integrity, namely its turnover after intravenous injection in vivo, is absolutely normal for SAP isolated from serum heated in that way (17).

Second, the commercial SAP that they used (Calbiochem) is lyophilized from buffer containing sodium azide and EDTA, but no mention was made of whether these bioactive additives were removed before use in the fibrocyte differentiation assays. In addition, lyophilization of SAP damages the protein so that, for example, it is instantly cleared and degraded after injection in vivo (Pepys MB, Hawkins PN: unpublished observations), and the properties of SAP that has been lyophilized may therefore no longer be physiologic in vitro.

Third, the SAP depletion and reconstitution experiments showed only modest differences that scarcely attained statistical significance, and thus the results are inconsistent with the dramatic potency reported for the isolated commercial SAP.

Fourth, the activities of the different clinical sera varied by 3-fold, but their SAP concentrations differed by only ∼25%, and there was poor correlation between SAP values and fibrocyte inhibitor activity. The possibility of qualitative differences in SAP concentrations between individuals is extremely speculative, because there is no known genetic or protein sequence polymorphism of SAP, and its glycan is among the most invariant of any known glycoprotein (5).

In conclusion, there was no evidence of any reduction in the circulating SAP concentration in our series of well-characterized patients with SSc, nor is there any rigorous or compelling clinical or experimental evidence to date to support the idea that SAP has a role in normal or abnormal fibrosis.


Dr. Pepys had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Denton, Pepys.

Acquisition of data. Tennent, Dziadzio, Triantafillidou, Davies, Gallimore, Denton, Pepys.

Analysis and interpretation of data. Tennent, Dziadzio, Triantafillidou, Denton, Pepys.

Manuscript preparation. Tennent, Dziadzio, Denton, Pepys.

Statistical analysis. Denton.


We thank Beth Jones for preparation of the manuscript.