Serum and Urine Analysis of the Aminoterminal Procollagen Peptide Type III by Radioimmunoassay with Antibody Fab Fragments

Collagen Rel. Res. Vol. 3/1983, pp. 371-379

Serum and Urine Analysis of the Aminoterminal Procollagen Peptide Type III by Radioimmunoassay with Antibody Fab Fragments HEILWIG ROHDP, IRMHILD LANGER!, THOMAS KRIEG2 and RUPERT TIMPU 1 2

Max-Planck-Institut für Biochemie, Martinsried and Department of Dermatology, University of Munich, FRG.

Abstract A radioimmunoassay based on antibody Fab fragments was developed for the aminoterminal peptide Col 1-3 of bovine type III procollagen. This assay does not distinguish the intaet aminopropeptide Col 1-3 from its globular fragment Col 1. Parallel inhibition profiles were observed with human serum and urine allowing the simultaneous quantitative determination of intaet and fragmented antigens in these sam pies. Most of the material has a size similar to that of fragment Col 1 indieating that the aminopropeptide is degraded under physiologie conditions. The eoneentration of aminopropeptide in normal sera was in the range 15-63 ng/ml. Daily exeretion was found to be in the range 30-110 Ilg. More than 50 Ufo of patients with alcoholie hepatitis and liver eirrhosis showed elevated serum levels of aminopropeptide by the Fab assay. Elevated eoneentrations were deteeted more frequently with an antibody radioimmunoassay whieh measures mainly the intact form of the aminopropeptide. It is suggested that analysis of patients material by both assays could improve thei r diagnostic applieation. Key words: pro collagen peptide type III, clinieal radioimmunoassay, serum, and urine analysis. Introcluetion Aminopropeptides are preeursor-specific segments of interstitial procollagens which are released during conversion into collagen (Fessler and Fessler, 1978; Timpl and Glanville, 1981). These peptides persist for some time in tissues (Fleischmajer et al., 1981) or in the cireulation (Rohde et al., 1976; 1979) as demonstrated by immunologieal methods. A radioimmunoassay for the aminopropeptide of procollagen type III has been particularly useful in clinical studies, and has demonstrated inereased serum levels of the peptide in patients with aeute and alcoholic hepatitis, liver eirrhosis and careinoma (Rohde et al., 1979;

372

Heilwig Rohde, Irmhild Langer, Thomas Krieg and Rupert Timpl

Ackermann et a1. , 1981; Bolarin et a1., 1982; Hahn et a1., 1982; Pierard et a1., 1982). The circulating antigen appeared heterogenous, part of it resembling the intact, triple-stranded aminopropeptide in size (M r = 45,000) while another component was sm aller (M r = 10,000) and resembled the globular domain Col 1 of the peptide (Rohde et a1., 1979; Niemelä et a1., 1982). Only the small form could be detected in urine, suggesting that proteolytic degradation precedes the excretion of the peptide. Arecent immunochemical study (Rohde et al., 1983) showed that both the intact aminopropeptide (Col 1-3) and its fragment Col 1 could completely inhibit the antibody in radioimmunoassay but the latter antigen has a ten-fold lower affinity. This difference creates some problems in the quantitative analysis of biological sampies which contain both forms of the peptide. Depending on antibody affinity the assay may measure either the intact antigen rather exclusively (Rohde et al., 1979) or some of the degraded forms in addition (Pierard et al., 1982; Heynen et al., 1982). Such assays would also be of low precision and sensitivity for urine analysis. The different affinities of the antigens could be due to multiple interactions in binding since monovalent Fab fragments do not discriminate between intact aminopropeptide and fragment Col 1 (Rohde et al., 1983). We have now used the Fab radioimmunoassay for the analysis of human serum and urine and compared the data with those obtained previously with an antibody assay (Rohde et al., 1979). Materials and Methods Antigens, antibodies and biological sampies

Peptides Col 1-3 and Col 1 were obtained from pro collagen type III of bovine fetal skin (Bruckner et al., 1978). These peptides were used in radioimmunoassays and also for calibrating a Sephacryl 5-300 column (1.5 X 110 cm) equilibrated in phosphate-buffered saline, pH 7.2 (PBS), 0.04 0 10 Tween 20. Antibodies against procollagen type III were raised in rabbits and purified by affinity chromatography (Nowack et al., 1976). Fab' fragments were prepared from purified antibodies by pepsin digestion and reduction (Nisonoff et al., 1960) and Fab fragments were prepared by papain digestion (Mage, 1980). Serum sampies were obtained from healthy volunteers and stored at -20 oe. Urine was collected over 24 h from hospitalized patients with no obvious connective tissue or liver disease and with normal creatinine clearance. Serum and ascites sampIes oE patients wirh acute, chronic or alcoholic hepatitis or with liver cirrhosis where those characterized in a previous study (Rohde et al., 1979) and were kindly provided b yDr. E. Hahn, Berlin. Radioimmunoassays

The aminopropeptide Col 1- 3 was labeled with [1251] by the chloramine T proced ure yielding material with a specific activity of 10-15 X 103 cpm/ng (Timpl and Risteli, 1982). For inhibition assays (sequential saturation) a fixed amount of Fab fragment (40-100 ng) was preincubated with non-Iabeled inhibitors (15 h, 4 °C) followed by labeled antigen (1 ng) and second antibody (Rohde et a1., 1979). Biological sampIes were routinely analyzed by using duplicates of

Serum assay for procollagen III

373

three different dilutions. Inhibition assays under equilibrium conditions followed established procedures (Timpl and Risteli, 1982). The radioimmunoassay with antibodies instead of Fab fragments was that used previously (Rohde et al., 1979). Results Radioimmuno-inhibition assays based on sequential saturation of Fab' or Fab fragments against the aminopropeptide Col 1-3 showed a high reproducibility and sensitivity (Table 1). Ir was possible to detect as little as 0.1-0.2 ng/ml of antigen (Fig. 1 a). Both the intact aminopropeptide Col 1-3 and its fragment Serum or urine inhibitor ().lI)

100

10 100

0

t:

0

:;::

60

p/

..c:

E

40 20

.0 /

eo

I

/;

j.

.0

!

~l

/0 0

6

10

b

eo

.j.

80 ~ c::

10

CD

I

"-

CD

-=

0,5 0 -0,5

.~

~

100 I

I

I

o

\~ \\\\

·0

ß

\\

0\\

-1 100

10

\\,

100

Aminopropeptide inhi bitor (ng/ml)

Fig, 1. Inhibition profiles in radioimmunoassays with Fab' fragments against aminopropeptide type III using the sequential saturation (a) or equilibrium (b) technique. The labeled test antigen was peptide Col 1-3. Inhibitors of the reactions were peptide Col 1-3 (e), peptide Col1 (0), normal human serum (0) and normal urine (Li). The data of the equilibrium assay were analyzed by a logit plot B = BxlB o with Bx as the amount of bound labeled antigen in the presence of inhibitor and B o as the amount of bound labeled antigen in the absence of inhibitor. The total reaction volume was 0.4 ml.

Col 1 produced almost superimposable inhibition curves indicating that the latter has retained at least 70 Ofo of the original antigenic activity (Table 1). This is in contrast to a radioimmunoassay with intact antibody where fragment Col 1 possesses 10 Ofo or less activity and pro duces a distinctly less steep inhibition profile (Rohde et al., 1979, 1983). Fab radioimmunoassays also showed parallel inhibition profiles with normal human serum and urine (Fig. 1 a), demonstrating that the assay is useful for quantitative analysis. Identical observations were made when the assay was carried out under equilibrium conditions (Fig. 1 b) except for a 5 to 10-fold lower sensitivity in comparison to a sequential saturation

374

Heilwig Rohde, Irmhild Langer, Thomas Krieg and Rupert Timpl

Table 1. Sensitivity and interassay variability of Fab radioimmunoassays for aminopropeptide type III a Antibody fragment

Fab' b Fab

2.2 2.7

C

a

b C

50 Ofo inhibitory level (ng/ml with Coll-3

± S.D.)

± 0.9

Coll 3.1 2.9

± 1.4

Assays with the same antibody fragment and batch of inhibitor but including different batches of labeled Coll-3. Average from 8 assays with pepsin-derived Fab'. Single assay with papain-derived Fab.

assay. All further analysis of biological sampies were therefore done with the sequential saturation assay. The amounts of aminopropeptide in biological specimens were measured with Fab' assays using peptide Col 1-3 as reference inhibitor, and were calculated assuming that the unknown antigens in these sampies have the same affinity for Fab' as the reference inhibitor. The analysis of serum from 53 normal, adult persons (age range 20-65 years) and of 7 urine sampies showed good reproducibility usually within 10-20 Ofo variation (Table 2). The average serum conTable 2. Reproducibility of data obtained with the Fab radioimmunoassay for two randomly selected serum and urine sampIes SampIe (test A or B) a Serum No. 23 Serum No. 24 Urine No.l Urine No. 14

A B A B A B A B

Amount of pro collagen peptide measured at dilution b 1:5 1:10 1:20

(ng/ml) mean ± S.D.

26 30 30 27 59 93 47 62

32 ± 7 30 ± 2 29 ± 1 26 ± 4 85 ± 25 94 ± 5 52 ± 5 56 ± 5

30 31 28 21 88 89 54 55

39 28 29 29 109 99 56 52

• Sequential saturation tests carried out at different occasions with two different batches of labeled antigen but the same Fab fragment and reference inhibitor. b Mean va lues of duplicates at each dilution.

centration was 39 ± 12 ng/ml (± S.D.) indicating a normal range (mean ± 2 S.D.) of 15-63 ng/ml. All the individual values fell within this range except for two borderline cases (Fig.2). Among the group of adults analyzed the serum levels of the peptide appeared to be independent of sex or age. However, slightly higher values were found for the 10-19 years age group (Fig.2), which is in agreement with other data (unpublished) showing 2 to 10-fold high er concentrations of serum aminopropeptide in young people (below 5 years). Urine sampies also contained the aminopropeptide at concentrations ranging from 29-187 ng/ml,

Serum assay for procollagen III

E

BO

"-

Cl

c:

Cl.>

-c

a. Cl.>

e

0-

60 40

0-

'"c: E

«

•• •

~--.------------

• • • 0

0

20

375

• •

·0

••

•••• o. 00. ••

0"

.0 "0 .0

• •• 0

••• •• •

0 ..

••

•

•• • •

0

••

--------~--------

10-19 20-29 30-39 40-65 Age groups (y)

Fig. 2. Serum concentrations of aminopropeptide in normal subjects determined by Fab' radioimmunoassays. Each point represents individual values from female (.) or male (0) volunteers. The dashed line indicates the normal range calculated for the adult group (20-65 years). indicating a daily excretion rate of 30-110 {lg peptide. Approximately 97 Ofo of the antigenie activity could be recovered after dialysis, indicating that low molecular weight components of urine (e. g. urea) do not interfere with the assay. The molecular size of the antigens detected by the Fab' assay in serum was separated into two major peaks which coeluted either with intact aminoproanalyzed by molecular sieve chromatography (Fig.3). The serum antigen was peptide Col 1-3 or fragment Col 1. The average ratio between the small and large form of the serum antigen was about 4: 1. Serum and ascites when examined by an antibody assay contain also distinct amounts of a larger aggregated aminopropeptide (Rohde et al., 1979; Niemelä et al., 1982). Our studies with the Fab' assay on three sera from young people (14-19 years) detected only little of this material. Whether this reflects an inherent difference between both assays or a particular feature of the age group remains to be established. Urine contained only the small form of the antigen resembling Col 1 as has been shown previously (Rohde et al., 1979). In order to evaluate the clinical potential of the Fab' assay we have assayed serum sampies from 95 patients with liver disease who had been previously examined with the antibody radioimmunoassay (Rohde et al., 1979; Hahn et al., 1982). Patients with alcoholic hepatitis or liver cirrhosis frequently showed increased serum concentrations up to 300 ng/ml peptide wirh the Fab' assay (Fig. 4). The antibody assay revealed a slightly higher number of patients (66 Ofo) with elevated serum levels than the Fab' assay (54 Ofo). Ir was remarkable, however, that about 15 Ofo of the patients showed pathological concentrations wirh the Fab' assay and appeared normal with the antibody assay (Table 3). The analysis of serum sampies horn patients wirh acute or chronic hepatitis showed a high score (72 Ofo) of elevated levels with the antibody assay and a distinct1y lower score (26 Ofo) with the Fab' assay (Table 3). We have also analyzed ascites fluid from 10 patients with liver cirrhosis by the Fab' assay and found concentrations from 320 to 1570 ng/ml. A comparison 25

Collagen 3/5

376

Heilwig Rohde, Irmhild Langer, Thomas Krieg and Rupert Timpl

, , ,

VV

2,0 E c:

1.6

w u

1,2

=

= '" c: d .c

::; V>

.c

Co11-3

Col1

E

= c:

"-

cu

6

0,8

,0\

0,4

°°

«

,0,

ß

I

I

..

0,

' .0-

·· •

ä.

cu a...

4

\

2

0,

<0,2

'0,

150

125

100

75

I I I I I I I

"'0

Effluent volume, ml

Fig, 3, Distribution of aminoproptide antigen in aserum sam pIe after molecular sieve chromatogeaphy on 5ephacryl 5-300, The column (1.5 X 110 cm) was equilibrated in PB5, 0,04 % Tween 20 and calibrated with authentie sampIes of the aminopropeptides Col 1-3 and Col 1 (arrows), The serum (2 ml, 18 year donor) was monitored by absorbance measurement for total protein (full line) and using individual fractions by Fab' radioimmunoassay for aminopropeptide (dashed line)). Total recovery of aminopropeptide was 81 Ofo. VV denotes the total volume of the column.

NR

: I

E

= c:

"-

I

150

I I

100

>.

"'0 0

.c c:

«

•

I

I I I I I I

>.

d V> V> d

••

I

•

I

50 •

•

I •• I

I

•

•

.

•

•

_!~ ••~~~~~--~--~~ - --~- ~---- - ~~.!.- •• --d _. .. 50

100

150

200

250

NnD

300

Fab' assay (ng/ml)

Fig, 4, Comparison of serum concentrations of aminopropeptide determined by antibody

oe Fab' radioimmunoassay in patents with alcoholic hepatitis or liver cirrhosis, NR

denotes the normal ranges determined by antibody (2-15 ng/ml) or Fab' (15-63 ng/ml) assay,

Serum assay for procollagen III

377

Table 3. Proportions of increased serum aminopropeptide levels detected by either antibody or Fab radioimmunoassay in patients with liver disease Fab assay

Patient group

antibody assay

increased

Alcoholic hepatitis and cirrhosis

increased 37 (66 Ofo) 19 (340f0) normal

22 8

15 11

total

30 (54 Ofo)

26 (46 Ofo)

Acute and chronic hepatitis

increased 28 (72 Ofo) normal 11 (28 Ofo) total

56

39

9 1 10 (26 Ofo)

normal

19 10 29 (74 Ofo)

of the values with those obtained by the antibody assay showed either a moderate increase (factor 1.1-1.4) which is within the error limits of both assays or 2 to 4-fold higher amounts when analyzed by the Fab' assay.

Discussion We have developed a radioimmunoassay for the aminopropeptide Col 1-3 on the basis of Fab fragments which could be similarly inhibited by Col 1-3 and the fragment Col 1. Quite large differences in the inhibitory capa city of both components are usually observed in radioimmunoassays based on intact antibodies (Rohde et al., 1979; 1983). This observation is not fully understood but could be related to the fact that peptide Col 1-3 is a triple-stranded structure held together by a non-antigenic triple helix; while fragment Col 1 is singlestranded, since it lacks the triple helix (Bruckner et al., 1978). It appears possible that both antigen-binding sites of an intact antibody are able to react with two determinants on a single or two different Col 1-3 molecule thus increasing affinity by multiple or cooperative interactions while this effect is lost with the singlestranded Col 1 or with a monovalent antibody fragment (Rohde et al., 1983). As a practical consequence the Fab radioimmunoassay is more suitable for the analysis of complex biological sampies which may contain both intact and fragmented forms of the aminopropeptide. A disadvantage of the Fab assay is its 5 to 10-fold lower sensitivity when compared to an antibody assay (Rohde et al., 1983). Application of the Fab assay to normal human sera showed 5 to 6-fold higher concentrations of aminopropeptide when compared to values determined by an antibody assay (Rohde et al., 1979; Bolarin et al., 1982; Pierard et al., 1982). This is very likely due to the fact that the Fab assay efficiently measures both intact and degraded forms of the aminopropeptide. The interpretation was supported by molecular sieve chromatography showing a 1: 4 ratio between large and small forms of the serum antigen. Depending on their different affinity for Col 1-3 and Col 1 and the relative amount of degraded forms in the biological samples, antibody assays may either measure primarily the intact aminopropeptide (Rohde et al., 1979) or in addition increasing amounts of the sm aller

378

Heilwig Rohde, Irmhild Langer, Thomas Krieg and Rupert Timpl

form (Pierard et al., 1982; Niemelä et al., 1982; Heynen et al., 1981). This correlates to either a parallel or non-parallel shift in inhibition curves observed between the reference aminopropeptide Col 1-3 and the biological sampie. Interestingly, Heynen et al. (1981) found both types of curve shifts by using either serum or synovial fluid, respectively, indicating that the latter source contains mainly intact aminopropeptide or procollagen. Variations in the proportions of different sized-antigens was also indicated in our study of ascites fluid. The Fab assay was also valuable for analysis of urine which contains degraded aminopropeptide. The data indicate that approximately 20-30 % of the amount present in blood is excreted over 24 h suggesting a rather fast catabolism of circulating aminopropeptide. Alternatively, the excreted peptide may originate from pro collagen present in extracellular fibrils (Fleischmajer et al., 1981) and thus reflect tissue remodelling or necrosis. Degradation presumably occurs in the flexible peptide sequence which joins fragment Col 1 with the tripIe helix (Timpl and Glanville, 1981), while segment Col 1 appears more stable against proteolysis due to its compact nature. The enzymes involved in this degradation are not known. At present we cannot completely eliminate the more remote possibility that degradation is a non-physiologie event which occurs during sampling and storage of sera. Degradation of aminopropeptides has also been observed in cell culture studies (Pontz et al., 1973; Wiestner et al., 1982). The assays described both here and in previous reports (Rohde et al., 1979) offer the potential to search for proteases specifically involved in these reactions. Preliminary studies on liver patients suggest that the Fab assay is as useful in detecting fibrotic activity as the antibody assay (Rohde et al., 1979; Pierard et al., 1982; Hahn et al., 1982). A combined analysis with both assays could perhaps distinguish between enhanced release and degradation of aminopropeptide (see Table 3). The full potential of the Fab assay for differential clinical diagnosis (e. g. acute versus alcoholic hepatitis) and its value for studying other diseases remains to be explored.

References Ackermann, W., Pott, G., Voss, B., Müller, K.-M. and Gerlach, U.: Serum concentration of procollagen-III-peptide in comparison with the serum activity of N-acetyl-ß-glucosaminidase for diagnosis of the activity of liver fibrosis in patients with chronic active liver diseases. Clin. Chim. Acta 112: 365-369, 1981. Bolarin, D. M., Savolainen, E.-R. and Kivirikko, K. I.: Enzymes of collagen synthesis and type III procollagen amino-propeptide in serum from Nigerians with hepatocellular carcinoma and other malignant diseases. Int. }. Cancer 29: 401-405, 1982. Bruckner, P., Bächinger, H. P., TimpI, R. and Engel, J.: Three conformationally distinct domains in the aminoterminal segment of type III procollagen and its rapid tripie helix +coiI transition. Eur. }. Biochem. 90: 595-603,1978. Fessler, J. H. and Fessler, L. I.: Biosynthesis of procollagen. Ann. Rev. Bioehem. 47: 129-162,1978.

Fleischmajer, R., Timpi, R., Tuderman, L., Raisher, L., Wiestner, M., Perlish, J. S. and Graves, P. N.: Ultrastructural identification of extension aminopropeptides of type I and III collagens in human skin. Proe. Natl. Aead. Sei. USA 78: 7360-7364, 1981.

Serum assay for pro collagen III

379

Hahn, E. G., Pencev, D., Pittner, P., Kalbfleiseh, H., Bruguera, M. and Timpi, R.: Radioimmunological determination of pro collagen type III peptide in sera of patients with chronic liver disease: correlation with clinical and histologie data. In: Connective Tissue 0/ the Normal and Fibrotic Human Liver, ed. by Gerlach, U., Pott, G., Rauterberg, J. and Voss, B. Georg Thieme, Stuttgart, 1982, pp. 205-206. Heynen, G., Gaspar, S., Gynen, Ph., Plomteux, G. and Franchimont, P.: Le dosage radioimmunologique de peptide amino-terminal du procollagene de type III. Note technique. C R. Soc. Biol. 175: 533-537,1981. Mage, M. G.: Preparation of Fab fragments from IgGs of different animal species. Methods Enzymol. 70: 142-150, 1980. Niemelä, 0., Risteli, L., Sotaniemi, E. A. and Risteli, J.: Heterogeneity of the antigens related to the aminoterminal propeptide of type III procollagen in human serum. Clin. Chim. Acta 124: 39-44, 1982. Nisonoff, A., Wissler, F. C, Lipman, L. N. and Woernley, D. L.: Separation of univalent fragments from the bivalent rabbit antibody moleeule by reduction of disulfide bonds. Arch. Biochem. Biophys. 89: 230-244, 1960. Nowack, H., Gay, S., Wiek, G., Becker, U. and Timpi, R.: Preparation and use in immunohistology of antibodies specific for type I and type III collagen and procollagen. J. Immuno!. Meth. 12: 117-124,1976. Pierard, D., Plomteux, G., Amrani, N., Robin, M., Gielen, J. and Lapiere, C M.: Type III collagen precursor sequences in sera of patients with hepatic diseases. In: Connective Tissue 0/ the Normal and Fibrotic Human Liver, ed. by Gerlach, U., Pott, G., Rauterberg, J. and Voss, B. Georg Thieme, Stuttgart, 1982, pp. 217-218. Pontz, B. F., Müller, P. K. and Meigel, W. N.: A study on the conversion of pro collagen. Release and recovery of pro collagen peptides in the culture medium. J. Biol. Chem. 248: 7558-7564, 1973. Rohde, H., Nowack, H., Becker, U. and Timpi, R.: Radioimmunoassay for the aminoterminal peptide of procollagen pa1(I)-chain. J. Immunol. Meth. 11: 135-145, 1976. Rohde, H., Vargas, L., Hahn, E., Kalbfleiseh, H., Bruguera, M. and Timpi, R.: Radioimmunoassay for type III procollagen peptide and its application to human liver disease. Eur. J. Clin. Invest. 9: 451-459, 1979. Rohde, H., Bruckner, P. and Timpi, R.: Immunochemical properties of the aminopropeptide of procollagen Type III. Eur. J. Biochem., in press. Timpi, R., and Glanville, R. W.: The aminopropeptide of collagen. Clin. Orthopaed. 158: 224-242,1981. Timpi, R. and Risteli, L.: Radioimmunoassays in studies of connective tissue proteins. In: Immunochemistry 0/ the Extracellular Matrix, ed. by Furthmayr, H. CRC Press, Boca Raton, 1982, Vol. I, pp. 199-235. Wiestner, M., Rohde, H., Helle, 0., Krieg, T., Timpi, R. and Müller, P. K.: Low rate of procollagen conversion in dermatosparactic sheep fibroblasts is parralleled by lllcreased synthesis of type I and type III collagens. EMBO J. 1: 513-516, 1982. Dr. Rupert Timpi, Max-Planck-Institut für Biochemie, Abtl. für Bindegewebsforschung, 8033 Martinsried, FRG.