British Journal of Haematofogy. 1992, 82, 32-37

Monitoring of multiple myeloma and bone marrow fibrosis with aminoterminal propeptide of type I11 collagen (PIIINP) TAINATAUBE,KAARLEFRANSSILA, LEILA RISTELI,*J U H A RISTELIt AND INKERI ELOMAA Department of Radiotherapy and Oncology, University of Helsinki, and Departments of *Medical Biochemistry and fClinica1 Chemistry, University of Oulu, Finland

Received 6 February 1992: acceptedfor publication 27 April 1992

Summary. Bone marrow fibrosis is known in myelomatosis and depends on the extent of plasma cell infiltration. The serum concentration of the aminoterminal propeptide of type III procollagen (PIIINP) has previously been reported to reflect fibrogenesis in the marrow in myelofibrosis. Here we followed 15 consecutive patients with newly diagnosed multiple myeloma with repeated PIIINP measurements during treatment with intermittent courses of melphalan and prednisolone. PIIINP was found to change with clinical

behaviour of the disease, nonresponders and patients with recurrent disease having elevated values and the values in responders decreasing to the normal level or remaining there. Collagen fibres in plasma cell infiltrates of biopsies from bone marrow or skeletal tumours of these patients stained heavily with antibodies against PIIINP. Our results suggest that PIIINP works as a noninvasive indicator of bone marrow fibrogenesis. In multiple myeloma PIIINP is a sensitive, but not specific marker of disease course.

Fibrosis in bone marrow biopsies is one of the histologic characteristics that have prognostic significance in myelomatosis (Bartl et al, 1987; Krzyzaniak et al. 1988; Riccardi et al, 1988). Myeloma cells release lymphokines, which probably stimulate the formation of fibrosis, the degree of which is related to the magnitude of plasma cell infiltration (Bartl et nl. 1987: Riccardi et al. 1988). The production of the Mcomponent reflects the plasma cell burden, which is the best single parameter to indicate the clinical stage of the disease and also has prognostic significance (Dune & Salmon, 1975). During recent years, serum based methods have been developed for noninvasive measurement of fibrogenesis (for review see Risteli & Risteli. 1990). In myelofibrotic disorders the serum concentrations of aminoterminal propeptide of type 111 procollagen (PIIINP) have been reported to be increased (Hochweiss et a!, 1983: Hasselbalch et al. 1990) and immunohistochemically the fibrous tissue stains heavily for a precursor of type 111 collagen (Apaja-Sarkkinen et al. 1986). In a preliminary study of patients with multiple myeloma, we have found changes in serum PIIINP levels (Elomaa et al, 199 1). Our purpose now was to clarify whether changes in serum concentration of PIIINP reflect formation of fibrosis in myelomatosis, whether the fibres seen in the plasma cell infiltrates contain aminoterminal propeptide of type 111 procollagen immunohistochemically. and whether

the concentration of PIIINP in serum is indicative for disease course. PATIENTS Fifteen consecutive patients (six females and nine males, mean age 59 years, range 33-80 years) with untreated multiple myeloma were prospectively studied. The diagnosis was based on the detection of an M-component in serum or urine and on the presence of abnormal plasma cells in bone marrow biopsy (eight patients) or on the histological diagnosis of plasmacytoma of a skeletal tumour (seven patients). The clinical staging system of Durie & Salmon (1975) was applied. The clinical characteristics of the patients are given in Table I. All patients had widespread focal skeletal disease at radiography. The serum calcium concentration and transaminase activities were normal. Only one patient had an increased serum creatinine level (180 pmol/l, reference interval 60-1 15 pmol/l) at the start of therapy. The first-line chemotherapy consisted of melphalan 9-12 mg/mz and prednisolone 1 mg/kg during four consecutive days every fourth week. If no response was detected or if the disease recurred, MOCCA-regimen (melphalan, vincristine, cyclophosphamide, lomustine, methylprednisolone) was started as a second-line therapy. At the beginning, three patients were operated on for a pathological fracture and 1 4 patients received radiotherapy. Radiotherapy was also given during

Correspondence: Dr Taina Taube. Pajalahdentie 8 A 5. SF 00200 Helsinki. Finland.

32

PllZNP is a Sensitive Marker in Myeloma Table I. Clinical characteristics and pretreatment laboratory values of 1 5 patients with multiple myeloma. M immunohistochemistry. a = alive, d = dead.

=

33

male, F = female, IHC

=

M-component Patient

Age/Sex

Para-protein

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

63 M 74 F 57 M 71 M 56 M 57 M 33 M 45 M 67 F 63 F 51 F 55 M 80 F 66 M 46 F

IgG-lambda IgA-kappa IgG-lambda IgA-kappa IgG-lambda IgG-kappa EJ-lambda IgG-lambda B-J-lambda IgA-lambda IgA-kappa

IgG-kappa IgA-kappa IgA-kappa B-J-kappa

Clinical stage

No. of bone lesions

I1 A I1 A I1 A I1 A I1 A I1 A I1 A I1 A I1 A I1 A 111 A 111 A 111 A 111 A 111 A

6 6 4 5 6 5 5 5 6 9 12 28 38 21 21

s-PIIINP S g/l

U mg/d

890 1890 190 1.33 1150 57 0 1220 600 -

4680

4.9 9.1 2.9 2.3 5.1 2.7 3.7 4.3 8.2 6.3 3.5 9.5 7.6 8.2 8.3

MC of PIIINP

Treatment response

FOIIOW-UP

+

Responder Responder Responder Responder Responder Responder Responder Recurrence Recurrence Recurrence Recurrence Nonresponder Nonresponder Nonresponder Nonresponder

11 a 35 a 32 a 26 a 11 a 37 a 8a 31 a 37 d

+ +

+ + + + +

+ -

+ + + -

(months)

22 d 10 d

4d 10 d 4d

Reference interval: S-PIIINP 1.7-4.2 pg/l. the study for painful bone lesions or impeding fracture. The patients were followed up for 4-3 7 months. METHODS

Radioimmunoassay. The concentration of aminoterminal propeptide of type I11 procollagen (PIIINP)was analysed with an equilibrium type of radioimmunoassay based on human antigen (Risteli et al, 1988). The assay detects the authentic propeptide and another, somewhat larger related antigen, but it is not sensitive to the smaller degradation products of the propeptide. In this assay modification the standards and serum samples give parallel inhibition curves. Serum samples were stored at -2OOC until analysed. The normal level of PIIINP in adult Finnish blood donors ( n= 88) is 1.74.2 pgJ1. The intra- and interassay coefficientsof variation of the method are about 5%. Histological assessment. Bone marrow biopsies or tissue samples of a skeletal tumour were available from seven and five of the patients, respectively. Bone marrow biopsies from the iliac crest were fixed in 10%phosphate buffered formalin, decalcified in EDTA for 3 6 h and embedded in paraffin. The sections were stained with haematoxylin and eosin and with Gomori’s reticulin stain. The aminoterminal propeptide of human type In procollagen was purified from malignant ascitic fluid (Niemela et al, 1985) and the antibodies were cross-absorbed with both basement membrane proteins and the interstitial procollagens I and 111 (Apaja-Sarkkinenet al, 1986).The sections for immunohistochemical staining were deparaffinized and treated with 0.4% pepsin to enhance the availability of the antigenic determinants. The sections were exposed to a 0.3% solution of hydrogen peroxide in absolute methanol to inactivate the endogenous peroxidase. then stained with

anti-PIIINP antibodies by using Avidin-Biotin method (Apaja-Sarkkinen et al, 1986). Effect of treatment was followed-upby measuring the serum and urinary M-components, serum calcium, creatinine, alkaline phosphatase (AP)and transaminase as well as blood count and sedimentation rate. Samples for determining the serum PIIINP concentration like those for the other measurements mentioned above, were collected before treatment and thereafter 1 d before starting each treatment course. The whole skeleton was investigated using radiography every 6 months and additionally when needed. Only the measurable lesions exceeding 1.5 cm in diameter are included into Table I. Statistical analysis. The usual linear correlation coefficients were calculated between the PIIINP and M-component. Kaplan-Mayer product limit estimator was used for survival analysis. RESULTS

Serum PIllNP concentration and M-components Fig l ( a ) shows the behaviour of corresponding serum PIIINP and M-component concentrations, and Fig l(b) that of corresponding serum PIIINP concentration and urinary Mcomponent excretion in the individual patients during follow-up. Although the behaviour is quite similar, there were no statistically significant correlations between serum PIIINP and M-component concentrations or between serum PIIINP concentrations and urinary M-component excretion before or during the treatment. Serum PIlINP concentration and treatment response The patients were grouped according to their response to the therapy (Durie & Salmon, 1975). Patients with a reduction

34

Taina Taube et a1

1

12.0 26.0

7

14.0

1

Iz

/

i

/

&

8.0

h

8.0 4.0

8.0

5

6.0

2.0

1

0

0 0

0

) 0

5

I

8

12

18

24

30

36

4500

5000

2000

t;:::

-1 \

I

30

36

6

12 18 24 T h e (months)

30

36

1000 600

1000

0

0 0 (a)

24

18

1500

20.0

jj

12

3

5000

60,O

6

6

18

12

30

24

36

0

T h e (months)

(b)

Fig 1 . The behaviour of corresponding serum PIIIUP and M-component concentrations ( a )and that of corresponding serum PIIINP concentration and urinary M-component excretion ( b ) in the individual patients during follow-up. Each symbol represents a n individual patient. The area between the dotted lines indicates the reference interval.

Table 11. Amount of reticulin and PIIINP positive fibres in bone marrow biopsies and tumour samples. Amount of positive fibres stained by: Reticulin Plasma cell infiltration Tumour % ' free bone Tumour

Pat. no. Bone marrow biopsy 2 3 4 6 9

10 12

Tumour sample 1

-

8 13

14

+ = normal:

PIIINP

3

0 0 0

50 20 0

+ + + + + + +

+ +++ +++ +++

++ +++ +++ +++

Tumour free bone Tumour

+

+ + + + + +

++

+++

+++

+++ ++ +++ +++ +++

+ + = moderately increased: + + + = highly increased.

PIIINP is a Sensitive Marker in Myeloma by more that 50% in the serum or urinary M-component were defined as responders. If less than a 50% reduction in the M-component was achieved during 6 months, the therapy was considered to have failed (nonresponders). Using these criteria, four patients were nonresponders and 11 responders. Four responders later developed a recurrence. The median survival of the patients was 30 months.

Nonresponders In all four patients bone disease progressed leading to death. Two patients developed hypercalcaemia. Serum PIIINP level was high and remained clearly above the normal level during the whole follow-up (Table I, Fig 2). Responders The disease showed a sustained regression in seven patients. In three patients initial serum PIIINP concentrations were above the normal levels. but decreased into the normal range

g

14.0 12.0

n 10,O z 8,O

i

..............................

2.0

..................................................................................................

0 0

0 0 26.0

I

I

I

I

I

I

8

12

18

24

30

38

I

I

I

I

I

1

6

12

18

24

30

38

I

I

1

i o !

I

0

6

I

12 18 24 T h e (monthe)

I

1

30

36

Fig 2. Changes in the concentrationof the amino terminal propeptide of type 111procollagen (PIIINP)in serum during treatment of multiple myeloma in nonresponders (top), in responders (middle) and in patients with recurrent disease (bottom). The area between the dotted lines indicates the reference interval.

35

during the treatment. In the others, all the PIIINP concentrations measured were within the normal range (Table I, Fig 2 ) . Patients with recurrence Four patients first responded to the therapy with a decrease of M-protein. The pretreatment levels of PIIINP were elevated in three of them (Table I, Fig 2 ) . At the time of the recurrence at 12, 18. 24 and 30 months, respectively, the bone lesions were found to have progressed and the concentrations of Mprotein as well as PIIINP increased. Three patients died having hypercalcaemia in the terminal phase.

Irnrnunohistochernistry

No abnormal plasma cell infiltration was found in four bone marrow biopsies (Table 11). In those biopsies type 111collagen showed a normal distribution in the endosteum lining bone trabeculae and in blood vessels. The network of bone marrow collagen was very fine and resembled that seen in reticulin staining. In three biopsies the plasma cell infiltration varied from 5% to 50% of the marrow cells. Tumour free areas showed a normal staining for type III collagen. However, similarly to all tumour samples, the areas infiltrated by plasma cells showed a marked increase in fibrous tissue, which was stained heavily by antibodies to PIIINP (Fig 3 ) . In reticulin staining the architecture of fibrous tissue appeared coarse. DISCUSSION Severai reports associate bone marrow fibrosis with multiple myeloma (Vandermolen e t al, 1985; Bart1 et al, 1987; Krzyzaniak et al, 1988; Riccardi et al, 1988). Our study supports the finding that marrow fibrosis is restricted to the areas of myelomatous involvement, and that the degree of fibrosis is related to the magnitude of plasma cell infiltration (Krzyzaniaket al, 1988: Riccardi et al, 1988). In fact. Riccardi et aJ(1988) have shown that patients having more plasma cells in their bone marrow have a higher degree of fibrosis than those with lower plasmacytosis, and in individual patients, the bone marrow areas most infiltrated by plasma cells also have the heaviest fibrosis. No clear correlation has been reported between marrow fibrosis and clinical stage. although stages I and I1 tended to have less marrow fibrosis than stage 111 (Riccardi et al, 1988). Our study shows that changes in type 111 collagen metabolism are common in myeloma. Since the serum concentration of PIIINP is often elevated in myelofibrosis (Hochweisset al, 1983; Hasselbalch et al, 1990). it is probable that also in myelomatosis the serum changes are related to a fibrogenesis induced by plasma cells. Indeed, we show that fibrous tissue in areas infiltrated by plasma cells stains heavily with antibodies to the aminoterminal propeptide of type 111 procollagen. indicating the presence of partially processed type 111procollagen in the fibres. Furthermore, it seems that in patients with myelomatosis the amount of the related PIIINP antigenicity in serum reflects the extent of fibrosis, since the highest PIIINP concentrations were seen in nonres-

36

Taina Taube et a1

Fig 3. Immunoperoxidase staining for the aminoterminal propeptide of type 111procollagen (PIIINP) in skeletal plasmacytoma.The fibrous tissue + 1. as well as the walls of blood vessels (smallarrows).The section was counter-stainedwith Mayer’shaematoxylin (thickarrows)is positive ( ( x 200)

++

ponders and in patients with recurrent disease and the lowest in responders. The more the myeloma cell mass increases, the more activating factors for fibrogenesis are released. Thus, fibrogenesis in bone marrow or in the tumours seems to be a dynamic process and its degree depends on the activity of myelomatosis. This provides the rationale for the use of PIIINP as a prognostic and noninvasive marker for fibrogenesis in multiple myeloma. Although serum PIIINP is not a specific marker in myelomatosis like the M component, it seems to be quite sensitive. Particularly. it could be ofvalue in cases of nonsecretory myeloma. However, since our study has a limited number of patients, we will investigate further the usefulness of PIIINP as a prognostic marker in a large follow-up study of 3 50 myeloma patients. Serum PIIINP values may also increase because of other fibroproliferative reactions. such as recent major surgery or liver and rheumatic diseases (Risteli &: Risteli. 1990). In this study, the serum samples were taken before treatment, and thus the initially high values are not due to treatment. None of the patients had a history of liver or rheumatic diseases. Reduced renal function, which is often associated with myelomatosis. has no effects on PIIINP levels. since small degradation products of PIIINP are not detected by the assay used (Risteliet al. 1988).This assay measures the large serum forms of PIIINP. wbich are probably cleared from the blood by receptor-mediated endocytosis in the endothelial cells of the liver (Smedsrod. 1988). We conclude that the fibrous tissue seen in plasma cell

infiltrates in bone marrow biopsies contains type I11 collagen, and that changes in the serum concentration ofPIIINP reflect the degree of fibrogenesis. which depends on plasma cell burden. This provides the rationale for using serum PIIINP as a n additional indicator for the course of disease in myelomatosis. ACKNOWLEDGMENTS We are grateful to the Finnish Cancer Foundation for the support of this work and to Ms Kristiina Pekkala and Ms Mirja Loikkanen for expert technical assistance. REFERENCES Apaja-Sarkkinen,M., Autio-Hamainen. H., Alavaikko, M.. Risteli,J. & Risteli. L. (1986) Irnmunohistochemical study of basement membrane proteins and type 111 procollagen in myelofibrosis. Brirish Joitrnalo/ Haernatology, 63, 571-580. Bartl. R.. Frisch. B.. Fateh-Moghadam. A,. Kettner. G.. Jaeger. K. & Somrnerfeld, W. (1 987) Histologic classification and staging of multiple myeloma. A retrospective and prospective study of 674 cases. American lourha1 of Clinical Pathology, 87, 342-355. Durie. B.M.G. & Salmon, S.E. (1975) A clinical staging system for multiple myeloma. Cancer, 36, 842-854. Elomaa. I.. Risteli, L. & Risteli. J. ( 1 991) Procollagens as markers in multiple myeloma. Calc$ed Tissue International, 48,210. Hasselbalch,H.. lurker. P.. Hoerslev-Petersen,K., Lisse, J. & Bentsen, K.D. (1990)Procollagen type 111 aminoterminal peptide in serum

PIIINP is a Sensitive Marker in Myeloma in idiopathic myelofibrosis and allied conditions. Relation to disease activity and effect of chemotherapy. American Iournal of Hematology. 33, 18-25. Hochweiss. S., Fruchtman. S.. Hahn, E.G., Gilbert, H., Donovan, P.B., Johnson, J., Goldberg, J.D. & Berk, P.D. (1983) Increased serum procollagen 111 aminoterminal peptide in myelofibrosis. American Journal offfematology, 15, 343-351. Krzyxaniak, R.L., Buss, D.H.. Cooper, M.R. & Well, H.B. (1988) Marrow fibrosis and multipte myeloma. American JournaI of Clinical Pathology, 89, 63-68. Niemek, 0.. Risteli. L., Parkkinen, J. & Risteli, J. (1985) Purification and characterization of the N-terminal propeptide of human type I11 procollagen. Biochemical Journal. 232, 145-1 50. Riccardi. A.. Ucci, G.. Cocci, A. & Ascari, E. (1988) Bone marrow fibrosis in multiple myeloma. American Iournal ofClinical Pathology, 90.753.

37

Risteli, J., Niemi. S., Trivedi, P., Maetausta, O., Mowat, A.P. & Risteli. L. (1 988) Rapid equilibrium radioirnmunoassay for the aminoterminal propeptide of human type 111procollagen. Clinical Chemistry. 34, 715-718. Risteli. L. & Risteli, J. (1990) Noninvasive methods for detection of organ fibrosis. Connective Tissue in Health and Disease (ed. by M. Rojkind). pp. 61-98. CRC Press, Boca Raton. Smedsrod. B. (1988) Aminoterminal propeptide of type 111 procollagen is cleared from the circulation by receptor mediated endocytosis in liver endothelial cells. Collagen and Related Research 8, 3 7 5 . Vandermolen, L., Rice, L. &Lynch. E.C. (1985) Plasma cell dyscrasia with marrow fibrosis. Clinicopathologicsyndrome. American Journal of Medicine, 79, 297-302.