Purification of Porcine Aminoterminal Propeptide of Type III Procollagen from Lymph and Use for Lymphatic Clearance Studies in Pigs

Matrix Vol. 1111992,pp. 73-79 © 1992 by Gustav Fischer Verlag, Stuttgart

Purification of Porcine Aminoterminal Propeptide of Type III Procollagen from Lymph and Use for Lymphatic Clearance Studies in Pigs LARS THORBJ0RN JENSEN 1 , JUHA RISTEU 2 , META DAMKJJER NIELSEN 3 , JENS H. HENRIKSEN 1 , HENNING PETER OLESEN 4 , LEILA RISTEU 2 and IB LORENZEN 1 1

2

3

4

Departments of Rheumatology and Clinical Physiology, University of Copenhagen, Hvidovre Hospital, DK-2650 Hvidovre, Denmark, Collagen Research Unit, Departments of Medical Biochemistry and Clinical Chemistry, University of Oulu, SF-90220 Oulu, Finland, Department of Clinical Physiology, University of Copenhagen, Copenhagen County Hospital, DK-2600 Glostrup, Denmark and Institute of Experimental Research in Surgery, University of Copenhagen, Panum Institute, DK-2200 Copenhagen, Denmark.

Abstract To investigate the lymphatic transport of the aminoterminal propeptide of type III procollagen (PIllNP) we established a thoracic duct-venous shunt in 6 pigs. Porcine PIIINP was purified, characterised, and compared with human PIllNP to ensure the suitability of the radioimmunoassay of human PIllNP for measurements in pigs. SDS-PAGE and radioimmunoinhibition assays show human and porcine PIllNP to be similar, thus indicating that the assay of human PIllNP is also reliable for determinations on pig serum and lymph. Intact PIllNP, as identified by gel filtration, accounted for 60% and 40% of the total PIllNP immunoreactivity in lymph and serum, respectively. The higher amount of total immunoreactivity and proportion of intact PIllNP in lymph compared with serum support the hypothesis that intact PIllNP is transported from pripheral tissue into the circulation by lymph. Two days after the shunt was established, the lymph was collected quantitatively hour-byhour for 24h. The flow was higher during the light periods than in the dark (p < 0.01). The PIllNP concentration varied inversely with the flow, being higher in the dark hours (p < 0.03). However, the total collected amount of PIllNP in lymph did not differ during the light and dark periods. Serum PIllNP remained unchanged over the 24 h. The lymphatic clearance of total PIllNP immunoreactive components was 6.2 ml serum/min and the lymphatic clearance of intact PIIINP was 9.1 ml serum/min, equal to 7 and 10 times the plasma volume/24 h, respectively. We conclude that the total amount of PIIINP transported by the thoracic duct is constant over 24 h and that the greater part is transported by lymphatics. On the other hand, the constant serum level of PIIINP despite the removal of 1 mg PIllNP from the body, indicates additional sources of circulating PIllNP. Key words: clearance, lymph, pig, procollagen.

Meta Damkjrer Nielsen died in the summer 1989.

74

L. T. Jensen et al.

Introduction Type III collagen is one of the main constituents in loose connective tissue and granulation tissue. With the introduction of radioimmunoassays capable of detecting the aminoterminal propeptide of type III procollagen (PIIINP), a new aspect has been added to clinical and experimental investigations of fibrosis (reviewed by Risteli and Risteli, 1990). Despite intensive studies of the PIIINP, little attention has been paid to the fact that serum levels of PIIINP depend not only on supply, but also on degradation and excretion. The metabolic routes have been investigated only to a minor degree. Studies of wound healing in humans have shown the concentration of PIIINP to be 1000-fold higher in wound fluid than in serum (Haukipuro et aI., 1987; Haukipuro et aI., 1990). A similar observation has been noted in rat wound fluid (Jensen, L. T., unpublished results). The concentration of PIIINP in synovial fluid is 500-1000-fold higher than in serum (Horslev-Petersen et aI., 1988; Gressner and Neu, 1984). In a study of lymph from sheep lungs, the levels of PIIINP was found to reflect directly the ongoing fibrosis (Nerlich et aI., 1984). Tissue fluid and synovial fluid are known to be cleared by lymph (Wallis et aI., 1987; Weinberger and Simkin, 1989). It is therefore probable that a macromolecule like PIIINP (M r = 42000) is transported into the circulation by the lymphatics and only to a small extent by direct transcapillary tissue to blood transport. The aim of the present study was to purify and characterise porcine PIIINP in relation to human PIIINP, to investigate the profiles of the PIIINP immunoreactive components in lymph and serum, and to determine the diurnal supply of PIIINP into the circulation by lymph. On these data we assessed the mean 24 h lymphatic clearance of total PIIINP immunoreactive components and intact PIIINP. For this purpose we invented an experimental model in conscious pigs, which permitted long-term sampling of thoracic duct lymph (Jensen et aI., 1990).

Materials and Methods Purification ofporcine PIIINP Four litres of porcine lymph collected during and after surgery were precipitated with solid (NH 4hS04 (40% saturation). The precipitated proteins were collected by centrifugation at 15000 g for 30 min and dissolved in 50 mM Tris/HCI, pH 8.6, 2 M urea containing protease inhibitors (3 mg/l of phenylmethanesulphonyl fluoride, N-ethylmaleimide and p-hydroxymercuribenzoate and 10 mM EDTA) and dialysed against this buffer. The dialysed sample was then chromatographed on a DEAE-Sephacel column (5 x 50 cm) equilibrated in the above buffer. Elution was carried out with a linear gradient of NaCI (0-0.4 M NaCl, 4000 + 4000 ml). The propeptide was located by

radioimmunoassay for human PIIINP and by adding a small amount of iodinated human propeptide to the preparation. Antigenic material was pooled, dialysed against 0.2 M NH4HC0 3 , and chromatographed on a Sephacryl S300 column (1.5 x 110 cm) equilibrated in this solution. The fractions containing the propeptide were pooled and dialysed against 50 mM ammonium acetate buffer, pH 5.5, and chromatographed (flow rate 1 ml/min) on a Protein Pak DEAE 5PV column using a Kontron HPLC instrument (Pump 420, Gradient former 425 and Detector 432) coupled to a Gilson 203 fraction collector. The propeptide was eluted with a linear gradient of NaCI (0-0.5 M NaCl, 0-60 min). The fractions containing the propeptide were directly injected into the same HPLC system for reverse phase separation (Vydac 201 TP1 04 column) in 0.1 % trifluoroacetic acid containing 10% 2-isopropanol and eluting the bound propeptide with increasing concentrations of 2-isopropanol (10-70%, 0-45 min). The fractions containing the propeptide were pooled and lyophilised for analysis. The purity of the propeptide was verified by SDS-PAGE in 125 g/l polyacrylamide slab gels stained with Coomassie Brillant Blue. Immunoblotting with purified anti-PIIINP was performed after transfer to a nitrocellulose membrane. Gel-electrophoresis was performed with unreduced and 2mercaptoethanol reduced porcine and human PIIINP. Immunoinhibition curves were established with iodinated human and porcine PIIINP, inhibited by human and porcine PIIINP in dilutions of 1:2500 to 1:10240000, corresponding to PIIINP concentrations of 5.12 f.tg/l to 1.25 ng/1. The antibodies have been described elsewhere (Risteli et aI., 1988). Porcine serum and lymph in dilutions of 1:2 to 1:256 were compared with standard curves of human PIIINP, standardised by quantitative analysis for amino acids after acid hydrolysis (Niemela et aI., 1985).

Animals and surgical procedure The animals were six female pigs (Danish Landrace/ Yorkshire), weighing 34-48 kg. The pigs were fasted overnight before surgery, but permitted free access to water as dehydration reduces the lymph flow (Staub et aI., 1975). Three days before experimentation, the animals were anaesthetised with pentobarbital sodium (15 - 20 mg/kg) injected intravenously, orotracheal intubated, and ventilated mechanically. The anaesthesia was maintained during surgery by repeated doses of pentobarbital sodium (2-3 mg/kg). The thoracic duct was cannulated through a thoracotomy. The tip of the catheter was placed at the level of the 10th thoracic vertebral body, and the catheter was connected through an external valve system to a vein catheter whose tip was positioned in the cranial part of the vena cava superior. Details of the technique are given elsewhere (Jensen et aI., 1990). Anticoagulant therapy was maintained with 10000 IU heparin/day. All lymph delivered in

P III NP in Lymph the first 5 h of thoracic duct cannulisation was collected for later purification of PIIINP. After anaesthesia, the lymph contains high concentrations of PIIINP (up to 14000 ""gil) compared with normal pig serum (100 ""gil). The high concentration and the high ratio of PIIINP/total protein after anaesthesia makes lymph a suitable source for PIIINP purification. Protocol

The shunt was left open for 2 days before experimentation. The pigs were individually penned (2 x 2.5 m), with no restriction of movement. They were fed at 7 am and 1 pm and allowed free access to water. The light/dark cycle was 12: 12 h, with light from 7 am to 7 pm. All delivered lymph was collected quantitatively hour-by-hour for 24 h. No lymph was re-injected. The lymph volume was measured by weighing. Blood samples were taken every hour. The venous catheter of the thoracic duct-venous shunt was flushed with saline between sampling. Handling ofsamples

The blood and lymph samples were collected in dry glass test tubes and allowed to clot for 60 min at room temperature. The clotting time was prolonged owing to the heparin anticoagulation therapy. The samples were centrifuged for 10 min at 1500 g and the supernatants were stored at - 20°C until analysed. Determination ofporcine PIIINP

The PIIINP content of porcine lymph and serum was determined by a commercial assay (PIIINP-RIA kit, Farmos Diagnostica, Oulunsalo, Finland) (Risteli et aI., 1988). We used sequential saturation at room temperature, with 16 h of preincubation without tracer followed by 20 h of incubation with tracer. The maximum sample volume was 300 ",,1, reduced with increasing concentrations, e. g. 10",,1 for the analysis of porcine serum. Buffers were added to all samples to obtain the volume of 300 ",,1. All dilutions were made with phosphate buffered saline, pH 7.2, containing human serum albumin 1 gil (RIA-buffer). Prediluted anti-PIIINP from the kit was used in a volume of 100 ",,1. Antibody bound radioactivity was separated from unbound radioactivity by a solid phase second antibody according to the description of the kit. We labeled the tracer with 125Iodine by use of the iodogen method (Salacinski et aI., 1981), separating the labeleled antigen from free iodine by gel filtration on a 0.6 x 12 cm column of Sephadex G-25 (Pharmacia, Uppsala, Sweden). The calculated specific radioactivity averaged 58 Ci/g PIIINP (N = 7). lntra- and inter-assay coefficients of variation were 3.8% and 5.9%, respectively. Non specific binding was 3.6%. Sensitivity, defined as the 80% intercept of the standard curve, was 0.04 ""gil.

75

Gel filtration ofporcine lymph and serum

The molecular weight distribution of PIIINP antigen holding peptides was investigated by gel filtration of corresponding lymph and serum samples: 1.0 ml serum or 0.1 ml lymph were applied to a Sephacryl S-300 column (1.6 x 90 cm), equilibrated in RIA-buffer and eluted at a flow rate of 14 mVh. Fractions of 1.9 ml were collected. The column was calibrated with radioiodinated intact human PIIINP. The samples from the six pigs were supplemented by samples from two pigs used in a pilot study. Statistics and calculations

The parallelism of the human PIIINP standard curve and dilution curves of porcine lymph and serum samples were tested after logit-log transformation (Rodbard and Lewald, 1970). Changes in serum PIIINP during the observational period were tested by linear regression. Pratt's matched pairs signed rank test for related two samples (Rahe, 1974) was used to test differences between light and dark periods and for the evaluation of differences in the relative distributions of PIIINP immunoreactive components in lymph and serum. P values less than 0.05 were considered statistically significant.

Hu Po St Hu Po

-R-R

+R +R

Fig. 1. Slab-gel electrophoresis of porcine (Po) and human (Hu) PIIINP. PIIINP was tested in unreduced (-R) and in 2-mercaptoethanol reduced (+ R) form. Globular standards (St) were, from lowest to highest Mr : 14400,20100, 30000, 43000, 67000 and 94000. Note that the mobility of PIIINP is slower than that calculated from the real M r (42000) owing to its collagenous structure.

L. T. Jensen et a1.

76

The total lymphatic clearance of PIIINP immunoreactivity (Choral), expressed in ml serum/min, was calculated from the total 24-hour PIIINP in lymph (PIIINP 24h ) and the mean serum concentration of PIIINP (S-PIIINP) by equation 1. The lymphatic clearance of intact PIIINP (Cllnracr) was calculated by correcting the Choral with the amount of intact PIIINP of total immunoreactivity in serum (S-intact) and lymph (L-intact) (equation 2). Choral = C

l

= Intact

S-PIIINP24h (f.-tg)

--------"-----'--=----S-PIIINP (f.-tg/mL) X 24 h X 60 min/h

(1)

Choral xL-intact S-intact

(2)

A 80

g 60 ,.... ><

o

,

III 40 III

20

PIIINP i

i

10 20

.1

80

DILUTION

(X1,OOO>

I i i

320

i

1,280 5,120

Results Purification of PIIINP from 4litres of porcine lymph yielded 4.8 mg, which corresponds to a recovery of 37%. Gel electrophoresis of human and porcine PIIINP showed a single band with a relative molecular mass of 42000 daltons and, after reduction, a single band with a relative molecular mass of 14000 daltons (Fig. 1). Immunoblotting with rabbit polyclonal antibodies against human PIIINP showed staining of the porcine and human bands of intact PIIINP. The inhibition curves of human or porcine PIIINP tracers inhibited by human or porcine PIIINP, were superimposable (Fig. 2). Dilution curves of porcine lymph (N = 6) and serum (N = 6) had a slope of (mean +/- 2 SD) -1.129 +/- 0.028 and -1.129 +/- 0.023, respectively. The slopes were not significantly different from that of the standard curve for human PIIINP (slope -1.107, 50% inhibition = 0.44 f.-tg/l). Gel filtration of porcine thoracic duct lymph and serum showed different patterns of PIIINP antigen holding peptides (Fig. 3). The intact PIIINP proportion of the total PIIINP immunoreactivity was higher (p = 0.01) in lymph than in serum, that is averaging 60% versus 40% (TableI). Table 1. Relative distribution of intact PIIINP and high molecular weight fractions of PIIINP immunoreactive forms. Peak A is eluted corresponding to void volume and probably represents pN-collagen (PIIINP still attached type III collagen) or a conjugate of PIIINP and fibrinogen. Peak B may be a conjugate of PIIINP and another molecule. The origins is unknown. Peak C represents the intact PIIINP. For graphic presentation of the peaks, see Fig. 3. N = 8 pigs. Pigs. Nos 7 and 8 were previously investigated in the pilot. P values of differences between serum and lymph and means are given at the bottom of the table. Pig No.

B

1

80 0 0,....

><

60

0

III

;0

40 20

PIIINP i

I

10 20

i

80

i

320

i

DILUTION

(X1,OOO) i

i

1,280 5,120

Fig. 2. Radioimmunoinhibition curves of human and porcine PIIINP. A: Human 125I_PIIINP. B: Porcine 125I_PIIINP. Inhibited by human PIIINP (e-e) and porcine PIIINP (0-0). Purified human and porcine PIIINP were tested in dilutions of 1:2500 to 1:10240000.

Percent of total PIIINP immunoreactive Peak A Peak B Peak C Serum Serum Serum Lymph Lymph Lymph

7

56 13

2

6

3

7

4

5

5

10

6

5

7

8

8

9

7 15 8 4 16 5 11

7

10

Means NS

53 57 49 50 52 57 48 53 p

28 29 27 37 27 32 27 29 30

= 0.01

37 41 35 46 40 43 35 43 40 p

59 64 58 55 69 52 68 60 60

= 0.01

PIlI NP in Lymph

Serum PIIINP Relative conc. 2.0

Serum PIIINP

200j jJgII 100

o

<.... g/l)

77

+

Cone.

i

3,000 2,000

1.0

1,000

Lymph PIIINP

Cone.

0

00 40

50

70 Fractions

60

Lymph PIIINP Relative conc. 4.0

+

C

80

ml

60

(pg/l)

Lymph Volume

40

3.0

20

2.0

0

+ F

1.0

100

00 40

50

60

70 Fractions

Fig. 3. Gel filtrations of porcine serum and lymph. The column (5epharcryl 5-300, 1.6 x 90 em) was calibrated with labelled human PIIINP. 1.0 ml serum or 0.1 mllymph was applied. Flow rate 14 mllh, fraction volume 1.9 m!. The A peak corrresponds to pN-collagen (PIIINP still attached to the type III collagen). The B peak corresponds to a high molecular weight PIIINPrelated antigen (probably a conjugate of PIIINP and another molecule) and the C peak is the intact PIIINP.

The lymph flow varied widely over 24 h (Fig. 4), being higher in the light period (median 376 mV12 h) than in the dark (median 175 mV12h) (p < 0.01). The concentration of PIIINP varied inversely to the flow: significantly higher (median 1400 I-lg/I) in the dark period than in the light (1010 I-lg/I) (p < 0.03). There was also an inverse correlation between the volume of the lymph and its PIIINP concentration in the individual animals, i. e. a small 24-h lymph volume had a high concentration and vice versa (Table II). The total 24-h amount of PIIINP thus tended to level out (Fig. 4) and no significant difference was observed between the light and dark periods. Serum PIIINP remained constant for each pig, except for a transient increase in the first hour of light. Serum PIIINP did not differ during light and dark periods and no decline was found during the observational period (slope of regression line: range - 0.362.04 I-lg/I X h I , which is not signficantly different from zero). The total lymphatic clearance of PIIINP immunoreactive components (Chotal) was calculated to be 6.2 ml serum/min (range4.2-12.6ml serum/min), corresponding to a diur-

50

o

F

~

~

"gM~Total

PIIINP in lymph

+

0800 12.00 16.00 2000 24.00 04.00 08.00 Time

c:::======J IigM

O

dark

Fig. 4. Hour-by-hour variation of thoracic duct lymph and serum over 24h. Concentrations of PIIINP in serum and lymph, lymph volume and total PIIINP (lymph volume x PIIINP concentration in lymph) transported by the thoracic duct. The arrows on the left of each figure represent the median of the 24-hour observational period. F = Time of feeding. The bars at the b( )trom indicate the time of light and dark periods. Data are expressed as medians, vertical bars indicate range. N = 6.

nal clearance of 8.9 V24 h (7 times the plasma volume). The lymphatic clearance of intact PIIINP (Chntact) was 9.1 mV min (range5.2-21.7), corresponding to a diurnal clearance of 13.1 V24 h (10 times the plasma volume).

Discussion The aminoterminal propeptide of type III procollagen is synthesised in all organs that contain type III collagen, chiefly in loose connective tissue and granulation tissue. Clearance from tissue is believed to take place via the lymph. We have previously reported a ten-fold higher concentration of PIIINP in porcine lymph than in porcine serum (Jensen et al., 1990). This observation prompted us

78

L. T. Jensen et al.

Table II. Lymphatic clearance of PIIINP. The 24-h thoracic duct lymph volume, the mean concentration of the aminoterminal propeptide of type III procollagen (PIIINP) in lymph and serum, the total amount of PIIINP collected (Total PIIINP) and the total lymphatic clearance of PIIINP immunoreactive components (CITotal). The Choral corrected for the difference in distribution ofPIIINP immunoreactive components is given for intact PIIINP (Cllntact). Clearances are expressed as ml serum/min. Means are given at rhe bottom of the table. Toral PIIINP (fig/24h)

PIIINP (fig/I)

Serum Chota I (ml/min)

Cllntact (ml/min)

1580 1540 1720 1100 850 1320

710 630 660 570 1180 740

110 60 110 87 65 120

4.5 7.3 4.2 4.5 12.6 4.3

7.1 11.4 6.9 5.4 21.7 5.2

1350

750

92

6.2

9.1

Pig No.

Volume (ml)

PIIINP (fig/I)

1 2 3 4 5 6

449 408 386 520 1388 562 619

Lymph

to determine the lymphatic clearance of PIIINP, to characterise the porcine PIIINP, and to compare PIIINP antigen holding peptides in lymph and serum. Purification and characterisation of porcine PIIINP showed that the molecular weights of unreduced and reduced PIIINP were identical with those of human PIIINP. Immunoblotting and radioimmunoinhibition curves showed that human and porcine PIIINP share the same immunological properties. The inhibition and dilution curves for porcine lymph and serum were parallel with those for the human standard. We therefore conclude that the assay for human PIIINP can be used for measurements of samples from pigs. Assays with other antibodies should, however, be tested before application. Gel filtration showed the distibution of PIIINP immunoreactive components to be different in serum and lymph. The relative concentration of intact PIIINP in lymph was higher than in serum. The relative distribution was constant in all the animals. In the interstitial fluid of a healing wound, in which the PIIINP concentration is 1000 times that of serum, the intact PIIINP accounts for almost 100% of PIIINP immunoreactivity (Haukipuro et al., 1987). This suggests that the lymph concentration of PIIINP reflects the fibrogenetic activity in areas drained by the lymph, thereby giving rise to the higher proportion of intact PIIINP in the lymph. The high molecular weight peak A (Fig. 3), which probably represents type III pN-collagen or a conjugate of PIIINP and lipoproteins (Bowness et al., 1989), could be released to the lymph directly from the loose connective tissue. Peak B, probably representing aggregated PIIINP and fibrinogen (Bowness et al., 1990), might be formed by the action of the liver (see below), the organ in which intact PIIINP is most probably metabolised (Smedsrod, 1988; Bentsen et al., 1988; Bentsen et al., 1989). The hour-by-hour lymph flow varied widely during the 24-h observational period. It was particularly increased in the hours after feeding and in the first hour of light. The flow was significantly higher in the light than in dark periods, which may be explained by differences in the ani-

rnaIs physical activity or contributions of lymph from the gut after feeding, or both. The concentration of PIIINP varied inversely and to some extent counterbalanced the changes in lymph flow. Owing to this simple dilution, the net amount of PIIINP released by the lymph into the circulation was equalised and consequently there was no difference in the amount found in the light and in the dark periods. The findings suggest that the biosynthesis of type III collagen is constant. All lymph delivered through the thoracic duct during the course of 24 h was collected. If the lymph supply into the circulation accounts for most of the total synthesised PIIINP, one would expect a decline in serum concentration. Surprisingly, we found no such change, even though almost 1 mg of PIIINP had been removed. This suggests that the turn-over of circulating PIIINP is very slow or the presence of other sources of PIIINP superior to thoracic duct lymph. The thoracic duct receives lymph from at least half the body, including the liver, kidneys and gut. Other sources of PIIINP that bypass the thoracic duct are the lungs (lymphatic drainage to the right duct (Warren and Drinker, 1942)) and tissue without lymphatics, e. g. bone, and to some extent muscle. For these reasons, the calculated lymphatic clearance of PIIINP is a minimum estimate of the whole body clearance. The lymhatic clearance of intact PIIINP (9.1 ml serum/min) was higher than that of total PIIINP immunoreactivity (6.2 ml serum/min) owing to the differences in the relative distribution of PIIINP immunoreactive components in lymph and serum. The hepatic clearance ofPIIINP in pigs, reported to be 91 (13-163) mUmin (Bentsen et al., 1989), is IS-fold higher than what we found by the lymphatic clearance. The clearance was calculated from the hepatic extraction of total PIIINP immunoreactivity, i.e. peaks A, Band C (Fig. 2). It is probably an overestimation of the metabolic clearance, because arterio-venous extraction coefficients include propeptides shunted, for example to the lymph. We have recently demonstrated that the liver shunts part of the circulating PIIINP directly to the lymph (unpublished results).

PIlI NP in Lymph We conclude that porcine and human PIIINP are identical with respect to immunoreactivity and behaviour in SDSPAGE. The diurnal variation in lymph flow and lymph concentration of PIIINP were inversely related, thus resulting in a relatively constant net supply of PIIINP into the circulation. This indicates that the biosynthesis of type III collagen over 24 h is constant. The profiles of PIIINP immunoreactive components in lymph were dominated by intact PIIINP, whereas high molecular weight forms were predominant in serum. These findings lend support to the concept that intact PIIINP is drained from peripheral tissue via the lymph and suggest that the high molecular weight forms originate from circulating PIIINP. The lymphatic clearance of PIIINP was low compared with the liver clearance, previously reported. Acknowledgements The authors gratefully acknowledge the technical assistance of Mrs. Annegrete Pedersen, Mrs. Irmelin Krabbe, Mrs. Kristiina Pekkala and Mrs. Aira Harju. This study was supported by the Danish Rheumatism Association (grant No233-654) and by Frolund Nielsens Foundation.

References Bentsen, K. D., Boesby, S., Kirkegaard, P., Hansen, C. P., Jensen, S. L., H",rslev-Petersen, K. and Lorenzen, I.: Is the aminoterminal propeptide of type III procollagen degraded in the liver? ]. Hepatol. 6: 144-150, 1988. Bentsen, K. D., Henriksen, J, H., Boesby, S., Hmslev-Petersen, K. and Lorenzen, I.: Hepatic and renal extraction of circulating type III procollagen amino-terminal propeptide and hyaluronan in pig.]. Hepatol. 9: 177-183, 1989. Bowness, J,M., Tarr, A.H. and Wiebe, R.I.: Transglutaminasecatalysed cross-linking: A potential mechanism for the interaction of fibrinogen, low density lipoprotein and arterial type III procollagen. Thromb. Res. 54: 357-367, 1989. Bowness, J, M. and Tarr, A. H.: Lipoprotein binding of crosslinked type III collagen aminopropeptide and fractions of its antigen in blood. Biochem. Biophys. Res. Commun. 170: 519-525, 1990. Gressner, A.M. and Neu, H.H.: N-terminal procollagen peptide and Ih-microglobulin in synovial fluids from inflammatory and non-inflammatory joints diseases. Clin. Chim. Acta 141: 241-245,1984. Haukipuro, K., Risteli, L., Kairaluoma, M.1. and Risteli, J,: Aminoterminal pro peptide of type III procollagen in healing wound in humans. Ann. Surg. 6: 752-756, 1987. Haukipuro, K., Risteli, L., Kairaluoma, M. I. and Risteli, J.:

79

Aminoterminal pro peptide of type III procollagen in serum during wound healing in human beings. Surgery 107: 381-388,1990. H",rslev-Petersen, K., Saxne, T., Haar, D., Thomsen, B.S., Bentsen, K.D., Junker, P. and Lorenzen, I.: Aminoterminal type III procollagen peptide and proteoglycans in serum and synovial fluid in patients with rheumatoid arthritis or reactive arthritis. Reumatol. Int. 8: 1-9,1988. Jensen, L. T., Olesen, H. P., Risteli, J, and Lorenzen, I.: External thoracic duct-venous shunt in conscious pigs for long-term studies of connective tissue metabolites in lymph. Lab. Animal Sci. 40: 620-624, 1990. Nerlich, A. G., Nerlich, M. L., Langer, I. and Demling, R. H.: Release of aminoterminal procollagen peptides in paraquatinduced pulmonary fibrosis. Exp. Mol. Pathol. 40: 311-319, 1984. Niemela, 0., Risteli, L., Parkkinen, J, and Risteli, J,: Purification and characterization of the N-terminal propeptide of human type III procollagen. Biochem.]. 232: 145 -150,1985. Rahe, A. F.: Tables of critical values for the Pratt matched pair signed rank statistics.]. Am. Assoc. 69: 368 - 373, 1974. Risteli, J., Niemi, S., Trivedi, P., Maentausta, 0., Mowat, A. P. and Risteli, L.: Rapid equilibrium radioimmunoassay for the amino-terminal propeptide of human type III procollagen. Clin. Chem. 34: 715 -718,1988. Risteli, L. and Risteli, J,: Non-invasive methods for detection of organ fibrosis. In: Focus on connective tissue in health and disease, ed. by Rojkind, M., CRC Press Inc., Boca Raton, Florida, 1990, pp. 61-98. Rodbard, D. and Lewald, J,E.: Computer analysis of radioligand assay and radioimmunoassay data. Acta Endocrinol. (Copenhagen) 147: 79-102, 1970. Salacinski, P. R. P., McLean, c., Sykes, J, E. c., Clement-Jones, V. V. and Lowry, P.J,: Iodination of proteins, glycoproteins and peptides using a solid-phase oxidizing agent, 1,3,4,6-tetrachloro-3a,6a-diphenylglucuril (iodogen). Anal. Biochem. 117: 136-146,1981. Smedsmd, B.: Aminoterminal propeptide of type III procollagen is cleared from the circulation by receptor mediated endocytosis in liver endothelial cells. Collagen ReI. Res. 8: 375 -388,1988. Staub, N. c., Bland, R. D., Brigham, K. L., Demling, R., Erdmann, A.J, and Wool Verton, W. c.: Preparation of chronic lung lymph fistulas in sheep.]. Surg. Res. 19: 315-320, 1975. Wallis, W.J., Simkin, P.A. and Nelp, W.B.: Protein traffic in human synovial effusions. Arthritis Rheum. 30: 57-63, 1987. Warren, F.M. and Drinker, C.K.: The flow of lymph from the lungs of the dog. Am.]. Physiol. 136: 207-221, 1942. Weinberger, A. and Simkin, P.A.: Plasma proteins in synovial fluids of normal human joints. Semin. Arthritis Rheum. 19: 66-76,1.989. ble 1. Lars Thorbj",rn Jensen, MD, Department of Rheumatology 232, Hvidovre Hospital, Kettegaard Aile 30, DK-2650 Hvidovre, Denmark.