Desmosines in aneurysms of the ascending aorta (annulo-aortic ectasia)

Biochimica et Biophysica Acta, 717 (1982) 105-110

105

Elsevier Biomedical Press BBA21159

DESMOSINES IN ANEURYSMS OF THE ASCENDING AORTA (ANNULO-AORTIC ECTASIA) TAPIO HALME a TIMO VIHERSAARI a TIMO SAVUNEN b, JUHA NIINIKOSKI b MARKKU INBERG b and RISTO PENTTINEN a ,

o Department of Medical Chemistry, University of Turku, b Department of Surgery, University of Turku, SF-20520 Turku 52 (Finland) (Received December 31st, 1981)

Key words: Desmosine," Elastin; Collagen," Marfan syndrome," (Aorta)

Amino acid chromatography was used for determination of the elastin-specific amino acids desmosine and isodesmosine in acid hydrolyzates of intima-medial samples taken intraoperatively from aneurysms of human ascending aorta. Elastin concentration of the specimens was also estimated by hot alkali extraction followed by nitrogen determination of the extracted material and the insoluble residue. All patients studied had annulo-aortic ectasia i.e., dilatation of the aortic annulus and the ascending aorta. Two patients with the Marfan syndrome had low aortic elastin concentration determined by both methods. A third Marfan syndrome patient, youngest of the three, also had a slightly reduced concentration of elastin in the aorta. Aortic samples were studied from five patients who did not have the classical Marfan syndrome. Two patients of those five had decreased aortic elastin concentration. The change in elastin concentration was accompanied by high hydroxyproline/proline or hydroxylysine/lysine ratios which indicates that the proteins of the aneurysmatic aortic wall contained more collagen than the proteins of the control aortic wall. These findings point to a change in the structure or metabolism of elastin in the aortic wall in the Marfan syndrome and at least in some other patients with annulo-aortic ectasia.

Introduction The Marfan syndrome is a dominantly inherited disorder of connective tissue which causes a variety of symptoms in musculoskeletal, ocular and cardiovascular systems. These include excessive height, scoliosis, deformation of thorax and other parts of the skeleton, and dislocation of lenses [1]. Typical for the syndrome is an aneurysm of the ascending aorta combined with dilatation of the annulus which leads to severe aortic regurgitation [ 1]. This entity has been called annulo-aortic ectasia [2]. Aortic aneurysm can also develop in patients who do not have the Marfan syndrome. The etiol* To whom correspondence should be addressed Abbreviation: CMN, cystic medial necrosis. 0304-4165/82/0000-0000/$02.75 © 1982 Elsevier Biomedical Press

ogy of the aneurysm in such cases is unknown with a few exceptions, e.g., syphilitic, mycotic or atherosclerotic aneurysms. Cystic medial necrosis (CMN) of the aorta is a pathological anatomic description including disorders such as Erdheim's disease (see Ref. 1 for review) in which the aortic wall shows fragmentation of elastic fibers, irregular whorls of hyperplastic smooth muscle cells, increase of collagenous tissue and appearance of cystic spaces occupied by metachromatic material [1]. Cystic medial necrosis is frequently observed in the Marfan syndrome. Changes which simulate CMN can be detected in aortas of apparently healthy persons and the validity of this morphological diagnosis has been questioned by Schlatmann and Becker [3]. Scheck et al. [4] and Byers et al. [5] have suggested that the defect in at least one Marfan

106

cardiovascular and mild skeletal findings without other signs of the Marfan syndrome. Among the non-Marfan patients two were histologically diagnosed as 'cystic medial necrosis' of the aorta, one was described as 'degeneration and fibrosis' of the aorta. In the specimen of one patient taken close to the rupture site of the aortic aneurysm there were no histological alterations from normal aortic wall. Control aortas were obtained from autopsies carried out in the Department of Forensic Medicine, University of Turku. The specimens were taken from ascending aortas of healthy persons died in accidents.

patient was due to an altered type I collagen a2-chain. Other studies have pointed to altered synthesis of collagen types (decrease in type I collagen) in the Marfan aorta [6]. Even though the Marfan syndrome is generally considered to be a defect in the mechanical-chemical properties of collagen fibrils [7] there are also histological changes, e.g., degradation and even complete destruction of elastic lamellae in the aorta [1,8]. On the basis of desmosine analysis and hot alkali extraction we confirm here that the elastin concentration is reduced in the wall of the aortic aneurysm in the Marfan syndrome. We suggest that the reduction is due either to altered properties of elastin or negative balance in its turnover. A similar pathogenetic mechanism might dilate the annulus and the ascending aorta in at least some 'non-Marfan' patients.

Separation of desmosines and other amino acids The intima-media of the aortic samples were separated from the adventitia, washed with cold water to remove the remaining blood and kept at - 2 0 ° C until analyzed. The samples (40-50 mg wet weight) were cut into small pieces and hydrolyzed for 48 h at 105°C in 6 N HC1 under nitrogen atmosphere. Desmosines, isodesmosines and other amino acids were separated as described by Starcher [9] with slight modifications.

Materials and Methods

A ortas Aortic samples were obtained from patients subjected to corrective surgery because of annuloaortic ectasia. The clinical and histopathological diagnoses of the patients are shown in Table I. Three patients had the classical Marfan syndrome. One patient (indicated as MS?) suffered from a dominantly inherited syndrome consisting of severe

Separation of elastin from collagen by hot alkali treatment Small (3-10 mg wet weight) samples were cut

TABLE I A O R T I C SPECIMENS Clinical diagnosis

Annulo-aortic ectasia Marfan syndrome

Inherited aneurysm Other aneurysm of ascending aorta

Age

Sex

Histopathological diagnosis

Abbreviation in the text

10 13 35 12 48 48 50 55

F F M M M M M M

CMN CMN CMN CMN Histologically normal CMN CMN Degeneration and fibrosis of the aorta

MS 1 MS 2 MS 3 MS? OA I OA 2 OA 3 OA 4

3 17 46 52

M M M M

-

CI C 2 C3 C 4

Controls

1 2 3 4

107

into pieces at 4°C and homogenized with an U1tra-Turrax ® homogenizer in ice bath in 2 ml cold distilled water and lyophilized. The samples were allowed to swell overnight in 5.0 ml cold distilled water and centrifuged at 2000 × g for 15 min. The supernatants contained practically no hydroxyproline and were discarded. The sediments were extracted in 5.0 ml boiling 0.1 M NaOH three times for 45 min followed by centrifugation as before. The combined supernatants and the sediments were analyzed for total nitrogen and hydroxyproline [10]. Extraction of tissue with hot alkali is known to solubilize collagen and most other l~roteins leaving mainly elastin in the insoluble residue [11]. Results

Amino acid analysis Table II shows the amounts of glycine, 4-hydroxyproline and valine estimated from chromatograms. Collagen and elastin are rich in glycine [ 12,13], thus glycine concentration reflects the concentration of these proteins. Hydroxyproline correlates with collagen concentration and valine (five times more frequent in elastin [13] than in col-

lagen) with elastin concentration. These amino acids in tissue hydrolyzates can be used as rough estimates of the concentration of the corresponding protein. Glycine values were low in the Marfan specimens MS 2, MS 3 and MS? and in the non-Marfan specimens OA 1 and OA 3 suggesting an increase in proteins other than collagen and elastin. Hydroxyproline concentration was clearly increased in hydrolyzates of all classical Marfan syndrome patients and in the non-Marfan syndrome specimen OA 4 histologically described as fibrosis of the aorta. Valine concentration was clearly smaller in the Marfan syndrome group and in the non-Marfan syndrome specimen OA 4 than in the control specimens. Cystic medial necrosis patients (OA 2 and OA 3) had normal values whereas OA 1 specimen contained a high concentration of valine. Chromatograms of the control samples showed four peaks between phenylalanine and hydroxyl-

2.0

-~

1.0

-

.8

-

.6

-

.5

-

.4

-

.2

-

z .J

z m

TABLE II THE CONCENTRATIONS OF GLYCINE, 4-HYDROXYPROLINE A N D VALINE IN THE H Y D R O L Y Z A T E S OF AORTIC SPECIMENS EXPRESSED AS RESIDUES/1000 AMINO ACID RESIDUES

~ m <

.I

-

0

-

Specimen OA 3 b was hydrolyzed for 72 h instead of the usual 48 hours. Specimen

Glycine

Hydroxyproline

MS 1 MS 2 MS 3 MS?

250 225 190 211

27 24 23 18

53 31 39 42

OA 1 OA2 OA3 a OA3 b OA4

205 238 200 236 224

12 15 12 15 38

122 64 79 57 18

C C C C

260 248 229 220

18 13 16 16

82 84 63 53

• In w

w

o o

>~ >-

.j z

30

60

z

go

120

J

J

150

180

2.0

Valine

1.0 .8

1 2 3 4

.6 w

]c

.6

~.4

< m E o m

.3

o

w

z

z

)

i

J

i

30

60

90

i

120

J 150

180

MINUTES

Fig. 1. Amino acid chromatograms of the basic amino acids of samples from the Marfan syndrome patients MS 3 (lower panel) and from the control patient C 3 (upper panel).

108

• dllm,/I

yz

• lO 2 o hySyl/lye

!i

"•

°

2i

J 0 ~-

,~? MS

I

i 23t OA

•

i

1 zaa~b4 C

• 10

l

'i 0

•

o ~

"" s2a MS

o LI2a OA

s 2 a, ab4 C

Fig. 2. The sum of desmosines in comparison with lysine (11) (expressed as absorbance of desmosines/4× absorbance of lysine) (left panel), and the molar ratios of 4-hydroxyprolineto proline (0) and hydroxylysine to lysine (O) (right panel) in aortic samples of the Marfan syndrome (MS), other aneurysm patients (OA) and in the controls (C).

ysine. Two first of them are known to be isodesmosine and desmosine [9] and the fourth is probably lysinonorleucine. As shown in Fig. 1 very little desmosines can be detected in the hydrolyzates of the Marfan sample (MS 3) (upper panel) whereas they are distinctly present in the control sample (C 3) (lower panel). The amounts of desmosines compared with those of lysine and the ratios of hydroxylated proline and lysine to their unhydroxylated forms are shown in Fig. 2. Control specimen from a 3-year-old boy had the highest value of the ratio desmosines/lysine (7.6- 10 - 2 ) whereas the others varied between 3 . 8 - 5 . 4 . 1 0 -2 (desmosines expressed as the absorbance/4). The hydroxyproline/proline and hydroxylysine/lysine ratios averaged 0.1 and were closely equal in all control samples. Desmosine concentrations in the Marfan syndrome specimens seemed to decrease with the age of the patient. The ratio desmosines/lysine was only a little smaller than the corresponding control in the sample of the youngest Marfan syndrome patient (MS l) but only half of the control value in the second case (MS 2). No desmosines were detected in the oldest case (MS 3). The patient with an inherited Marfan-like syndrome (MS?) also had reduced amounts of desmosines (about one fourth of the controls) in his aortic specimen. One specimen (OA 4) of the non-Marfan aortic aneurysms contained no desmosines. Another specimen (OA 3) had slightly reduced amounts, and the other cystic medial necrosis patient (OA 2) and the histologically normal case (OA 1) showed values which were comparable with the controls.

The ratio of hydroxyproline to proline, an indication of the relative collagen concentration of the specimens, was about two times higher in the Marfan syndrome patient samples than in the controls. The specimen from the non-Marfan case lacking desmosines (OA 4) also had a high h y p r o / p r o value. The ratio hydroxylysine/lysine (which also reflects relative collagen concentration of the sampies) was more variable. In two specimens of the Marfan syndrome group (MS 2 and MS?) the ratios were higher than the control values. In a similar way, the non-Marfan specimen which contained no desmosines (OA 4) had high degree of lysine hydroxylation. Hot alkali extraction The results of the hot alkali extraction of the aortic samples are presented in Table III. The amount of nitrogen in the insoluble residues of the control samples (expressed as per cent) decreased with age. In the Marfan specimens relatively less nitrogenous components were insoluble after the

TABLE III HOT ALKALI EXTRACTION OF THE AORTIC SAMPLES After centrifugation the sediments and supernatants were hydrolyzed in 6 M HCI at 130°C for 3 h followed by determination of nitrogen and hydroxyproline[10]. Specimen

Nitrogen in the sediment (% of total)

Hydroxyproline in the sediment (% of total)

MS 1 MS 2 MS 3 MS? OA 1 OA 2 a OA 3 OA 4 CI

33.4 11.7 8.1 22.5 27.5

15.2 2.1 2.1 7.3

44.5 18.4 47.1

C2 C3 C4

42.8 38.6 34.9

C 5 (32 years,m)

46.6

C 6 (36 years, f)

44.0

15.6 1.0 17.3 13.6 11.9 11.0 15.9 13.4

a Not enough available

6.6

109 extractions than in the controls. Hot alkali treatment did not completely solubilize hydroxyproline containing proteins from which about 11-17% in the control samples remained in the insoluble residue (Table III). The relative content of hydroxyproline in the residues followed that of nitrogen. Hydroxyproline values of the residues were very. low in two Marfan syndrome cases (MS 2 and MS 3) and in the OA specimen 4. Interestingly, aorta of the patient OA 1 (histologically normal) also had a reduced amount of insoluble elastin whereas the total hydrolyzates contained high valine concentration (Table II). Discussion

Amino acid analyses and results from hot alkali extraction of the aortic samples suggest that the concentration of elastin is reduced in many cases of aneurysms of the ascending aorta. Collagen concentration (estimated from hydroxyproline) is high in the same specimens. Results from analysis of elastin-specific cross-links, desmosine and isodesmosine (Fig. 1) agree well with those results. The aortic desmosine concentration was observed to decrease gradually with the age of the control person (Fig. 2). Our results suggest that elastin in the classical Marfan syndrome disappears abnormally rapidly from the aortic wall. While the aortic desmosine values of a 10-year-old Marfan syndrome patient were only slightly smaller than those in the aorta of a 17-year-old control person, a marked difference was noted in specimens of older patients. Collagen concentration, estimated from hydroxyproline values, was elevated from the control level but did not increase with age. The results can be explained in several different ways. The primary defect in the Marfan syndrome can be reflected by a decreased synthesis, increased degradation or decreased cross-linking of elastin. A mutation in the amino acid sequence of elastin near or at the cross-linking site can cause formation of abnormal elastic fibers. Such abnormally cross-linked elastin can also be expected to be solubilized more easily by the hot alkali extraction. Another possibility is a mutation in the structure of the microfibrillar component associated with elastin [14]. These alterations can

make the elastic fibers more susceptible to elastolytic processes, e.g., to proteases of the circulating blood or to endogenous aortic elastase [15,16]. A structural mutation in elastase could also change its substrate or antiprotease affinity. That collagen might not be the weak point in the connective tissue of the Marfan syndrome is supported by the normal stability of polymeric skin collagen observed by others [17]. Boucek et al. [18] however, reported altered cross-linking of collagen in several Marfan syndrome patients. How the pathological findings in elastin correlate with the other well-known features of the Marfan syndrome, e.g., excessive height, scoliosis and other skeletal abnormalities, remains unexplained in the light of the present results. While elastin fibers can stretch 130-140% of their resting length before breaking [13] collagen fibrils are rigid and have only weak viscoelastic properties. The maximal elongation of collagen fibrils under tension is in the range of 15-30% [19]. Pulsating blood pressure caused by cardiac activity can gradually dilate the rigid aortic wall and lead to aneurysm formation, dissection or rupture of the elastin-deficient aortic wall. Results of the patient 'MS?' with a dominantly inherited aneurysm of the ascending aorta but without other clinical findings of the Marfan syndrome fit to the 'Marfan group'. Patients expressing only part of the typical Marfan syndrome are well documented [1]. Aortic aneurysms other than those of the Marfan syndrome consisted of a heterogenous group, two of which were histologically defined as cystic medial necrosis of the aorta, one as degeneration and fibrosis of the aorta and one was histologically normal. The specimens were taken at the apparent site of the vascular pathology but obviously can show variations. The r non-Marfan patients can be examples of one or several diseases. Clinically the aorta of patient OA 4 seemed to have undergone fibrotic degeneration after dissection of aortic aneurysm but the biochemical results agree well with those of the Marfan patient aortas. The weakness of the aorta cannot be estimated on morphological basis because even the histologically normal-looking aorta (OA 1) was ruptured. The results of the chemical analyses suggest, however, that elastin is affected in all Marfan syn-

110

drome patients and at least in part of the nonMarfan aneurysm patients.

Acknowledgements The authors are grateful to Dr. Allen J. Bailey, ARC Meat research Institute, Langford, Bristol, U.K., for comments during the preparation of the manuscript. Drs. Tauno Ekfors and Raino Kunnas, Department of Pathology, University of Turku, kindly helped in pathological anatomic studies of the patient samples, This study was supported by grants from the fund 'Operation Red Heart' of the Lions International, Finland, The Turku University Foundation, The Research and Science Foundation of L~i~ike/Farmos Ltd., Turku, and The Finnish Cultural Foundation. We thank Prof. Jyrki Raekallio, M.D., Department of Forensic Medicine for control aortas.

References 1 McKusick, V.A. (1972) Heritable Disorders of Connective Tissue, pp. 61-223, Mosby Co., St. Louis 2 Ellis, P.R., Cooley, D.A. and DeBakey, M.E. (1961) J. Thor. Cardiovasc. Surg. 42, 363-370 3 Schlatmann, T.J.M. and lk~ker, A.E. (1977) Am. J. Cardiol. 39, 13-20

4 Scheck, M., Siegel, R.C., Parker, J., Chang, Y.-H. and Fu, J.C.C. (1979) J. Anat. 129, 645-657 5 Byers P.H., Siegel, R.C., Peterson, K.E. et al. Proc. natl. Acad. Sci. U.S.A. 78, 7745-7749 6 Krieg, T. and M011er, P.K. (1977) Exp. Cell Biol. 45, 207 -221 7 Pyeritz, R.E. and McKusick, V.A. (1981) N. Engl. J. Med. 305, 1011-1012 8 Kadar, A. (1980) in Front. Matrix Biol. (Robert, A.M. and Robert, L., eAs.), Vol. 8, pp. 54-68, Karger, Basel 9 Starcher, B.C. (1968) J. Chromatog. 38, 293-295 10 Pikkarainen, J. (1968) Acta Physiol. Scan& Suppl. 309, 1-72

11 Lansing, A.I., Rosenthal, T.B., Alex, M. and Dempsey, E.W. (1952) Anat. Record. 114, 555-575 12 Prockop, D.J., K.ivirikko, K.I., Tudermann, L. and Guzman, N.A. (1979) N. Engl. J. Mcd. 301, 13-23 13 Sandberg, L.B., Soskel, N.T. and Leslie, J.G. (1981) N. Engl. J. Mcd. 304, 566-579 14 Ross, R. and Bornstein, P. (1969) J. Cell Biol. 40, 366-381 15 Bieth, J. (1978) in Front. Matrix Biol. (Bieth, J., CoilinLapinet, G.M. and Robert, L., eds.), Vol. 6, pp. 1-82, Karger, Basel 16 Bellon, G., Ooyama, T., Hornebeck, W. and Robert, L. (1980) Artery 7, 290-302 17 Francis, M.J.O., Sanderson, M.C. and Smith, R. (1976) Clin. Sci. Mol. Mcd. 51,467-474 18 Boucek, R.J., Noble, N.L., Gtmja-Smith, Z. and Butler, W.T. (1981) N. Engl. J. Med. 305, 988-991 19 Viidik, A. (1980) in Biology of Collagen (Viidik, A. and Vuust, J., exis.), pp, 237-255, Academic Press, London