WO1999050666A1

WO1999050666A1 - Determination of the probability of bone metastases in patients with primary carcinomas

Info

Publication number: WO1999050666A1
Application number: PCT/EP1999/001835
Authority: WO
Inventors: Markus Seibel; Ingo J. Diel
Original assignee: Bio-Rad Laboratories Gmbh
Priority date: 1998-03-27
Filing date: 1999-03-19
Publication date: 1999-10-07
Also published as: AU3035699A; EP1064554A1

Abstract

A method to a) determine the probability of a future occurrence of bone metastases in patients with primary carcinomas and, if appropriate, the necessity for prophylactic treatment that constitutes the perioperative determination of the content of BSP (bone sialoprotein) and/or OPN (osteopontin) or their fragments or isoforms from one or more body fluids using polyclonal or monoclonal antibodies; b) improve the differentiation of patients with multiple myeloma from healthy persons or osteoporotic patients, to determine the response of patients with multiple myeloma to specific therapy or treatments, and to predict the outcome of the disease or the survival of the patient with multiple myeloma, all of them characterized by the determination of the amount of BSP (bone sialoprotein) and/or OPN (osteopontin) or their fragments or isoforms from one or more body fluids, using polyclonal or monoclonal antibodies.

Description

DETERMINATION OF THE PROBABILITY OF BONE METASTASES IN PATIENTS WITH PRIMARY CARCINOMAS

The subject of the present invention is i) a procedure to determine the probability of a future occurrence of bone metastases in patients with primary carcinomas and if appropriate, the necessity for a prophylactic treatment; ii) a procedure to estimate the probability of survival in patients with multiple myeloma.

Breast cancer is one of the most frequently occurring, malignant tumors of all. According to the latest WHO analysis, it is likely that just more than 10% of all women in the western world will develop breast cancer during their life span. Breast cancer is the most frequent cause of death in women aged between 33 and 50. As a rule, death is actually caused by metastasis. The hallmark of breast cancers is that a prognostic evaluation is usually difficult because of their biological heterogenity and their tendency to early hematogenic metastases. As such, the death rate of afflicted patients lies between 25 and 30%. This figure could also not be improved by the widespread use of the adjuvant systemic therapy procedure. This means that although 95 -_.98% of all patients with primary breast cancer are treated with adjuvant therapy, the survival rate is only improved by 10% at the most.

Breast cancer tumors metastase in bone with extraordinary frequency. About 75% of all patients that die of breast cancer demonstrate bone metastases upon autopsy. Currently, the best prognostic factors for an eventual metastasis (also in bone) is the status of af ected lymph nodes in the arm pits and the detection of disseminated tumor cells in the bone marrow. Despite several attempts to discover prognostic factors for a later metastasis in bone, establishing a suitable and reliable marker has not yet been possible.

The development of malignant bone disease is the major cause of morbidity in patients with multiple myeloma (MM). The associated pathologies range from osteoporosis-like changes to the development of osteolytic bone lesions and frank hypercalcemia. The latter complications are known to be associated with poor survival. The process leading to MM-induced bone disease most probably involves osteoclast-stimulating cytokines that are secreted by the myeloma cells and lead to a paracrine activation of osteoclasts. Moreover, it has been shown that adhesion molecules, i.e. laminin, vitronectin and fibronectin, are involved in processing the ability of myeloma cells to interact with the bone marrow stroma.

Attempts have been made to quantify biochemical markers of bone turnover as diagnostic and prognostic indices in plasma cell dyscrasias. As shown previously, the detection of collagen breakdown products in body fluids as markers of bone resorption may contribute to clinical guidance in identifying patients who are likely to benefit from antiresorptive drugs. An ideal marker, however, would reflect both the activity of the tumor itself, and disease related complications such as neoplastic bone involvement. Due to the nature of the collagen metabolites, only the latter could potentially be accomplished upon their quantification.

Normal tissue turnover as well as particular diseases of bone can be identified and monitored by the measurement of macromolecular components from body fluids. In evaluating bone formation for example, besides total alkaline phosphatase (AP) in serum, other parameters with higher tissue specificity are brought into use such as the bone specific AP (BAP), osteocalcin (OC) or the collagen propeptides (PINP, PICP). The rate of bone resorption can be obtained through the determination of calcium and hydroxyproline in urine as well as pyridinium crosslinks in urine and serum. In this way, a range of metabolic osteopathies, such as osteomalacia or Paget's disease, progress with characteristic and diagnostically indicative changes in these macromolecules. Usually, these components are of use because changes in their concentrations in body fluids indicate particular changes in the metabolism of bone or connective tissue.

The prediction of or early diagnosis of a bone metastasis is extremely important because of the therapeutic consequences. As such, the severity and lethality of the appearance of bone metastases, although not fully prevented, are clearly lowered by early therapy with bisphosphonates.

Serological tumor markers such as CEA and CA 15-3 are of great use in the progressive evaluation of malignant diseases but they do not exhibit any specificity for bone or connective tissue. In contrast, it could be shown that the serum and urine concentrations of particular macromolecules of bone origin are clearly altered in an existing, i.e., already established, metastasis in bone or connective tissue. Thus, discrete to clearly pronounced changes of the total alkaline phosphatase (TAP) or the bone specific AP (BAP) are found in serum in the osteoplastic metastases of breast or prostate cancer, depending on the extent and progress of the finding. In normocalcemic patients with bone metastases, elevated serum osteocaicin (OC) levels are frequently found that can drop to, or even below, the lower end of normal range because of the development of a metastatic hypercalcemia. In patients with manifested bone metastases, the hydroxyproline (OHP) excretion in urine is usually increased, and is thereby employed as an index to measure the success of treatment with a chemo or hormone therapy. Similarly, an increased secretion of pyridinium crosslinks is seen in 80 to 95% of all tumor patients with bone metastases. Bone sialoprotein (BSP) in serum is clearly elevated in established bone metastases as well as in multiple myeloma, a further malignant bone disease. To summarize, already existing bone metastases cause an elevation in the serum and urine concentrations of known and new types of markers of bone turnover.

This is not so in patients in which a malignant tumor has been diagnosed but where there is no evidence of metastasis. Until now, there has been no reliable procedure for these patients, which allows a statement to be made at the time of primary tumor diagnosis about the probability of a subsequent metastasis. This prediction is very important however, because firstly, it affects the overall prognosis of the patient and secondly, it could form the basis of an early and effective therapy with cytostatic substances and inhibitors of bone turnover.

The task of this invention was to develop a method to determine the probability of a future occurrence of a bone metastasis in patients with a primary carcinoma and, if appropriate, the necessity for prophylactic treatment.

The fundamental work by Castronovo and Bellahcene's group from Liege in Belgium could show that the immunohistochemical detection of bone sialoprotein in primary tumors of patients with breast cancer is retrospectively associated with a metastasis of the tumor in bone (see Bellahcene A., Kroll M., Liebens F., Castronovo V. [1996], Bone silaoprotein expression in primary human breast cancer is associated with bone metastases development, J. Bone Mineral Res. H: 665-670, and Bellahcene A., Merville M.P., Castronovo V. [1994], Expression of bone sialoprotein, a bone matrix protein in human breast cancer, Cancer Res. 54: 2823-2826).

It has now been discovered that the task of determining the probability of a future occurrence of bone metastases in patients with primary breast cancer and, if so, the necessity for prophylactic therapy can now be fulfilled when the content of BSP (bone sialoprotein) and/or OPN (osteopontin) or fragments or isoforms thereof are determined perioperatively in serum. The whole protein, protein fragments and isoforms of the protein (each including its posttranslational modifications) are meant by the terms bone sialoprotein or osteopontin when used in the following text.

The period described as perioperative particularly covers the time shortly before to shortly after breast cancer surgery, where the information derived from the determination is at best when the content of BSP and/or OPN in serum are determined shortly before surgery. The situations that all have a predictive value for the later occurrence of bone metastases can be distinguished as follows:

a) The pre- or immediately postoperative determination of BSP and/or OPN in serum of patients with diagnosed breast cancer enables the prediction of the probability of the occurrence of bone metastases at a later time point. When these values are available, the postoperative monitoring of serum levels of BSP and/or OPN is particularly useful because a stark decrease in the serum levels and maintenance of low serum levels permits a good prognosis.

b) A serum level that remains above the low normal base level indicates that either the entire primary breast carcinoma was not removed and/or that bone metastases already exist.

c) A serum level of BSP and/or OPN that rises again after an initial drop is an indication that metastasis has occurred nonetheless and/or a relapse at the site of the original breast carcinoma. Further diagnostic steps should then be undertaken and, if necessary, a very early preventive therapy of the bone metastases should be started before it is manifest or can be otherwise diagnosed. Apart from serum and urine, elevated concentrations of the mentioned markers are observed in other body fluids. Plasma, ascites, cerebrospinal fluid, synovial fluid and pleural effusion as well as the detection as BSP/OPN in bone marrow, or its cellular portion, appear suitable for the innovative procedure.

As far as multiple myelomas are concerned, presently used clinical staging systems are helpful in identifying high-risk patients with multiple myeloma due to renal failure or high tumor burden, but are lacking the biochemical quantification of the actual bone resorbing activity. However, excessive bone resorption and hypercalcemia are among the most important determinants of quality of life and survival in multiple myeloma patients. The inclusion of markers of bone turnover into the diagnostic panel therefore has clinical relevance. We and others have shown that the urinary excretion of hydroxypyhdinium crosslinks" and related telopeptides of type I collagen are helpful indices in the management of myeloma patients. However, increased collagen breakdown products are the consequence, but not the cause of accelerated bone resorption. We therefore put forward the hypothesis that the quantification of BSP in serum may improve the understanding of the underlying pathophysiological processes, the clinical assessment of MM- induced bone disease, and the prediction of individual survival.

The bone marrow represents the almost exclusive microenvironment of myeloma cells and appears to be crucial for the differentiation of the precursor cells into the typical malignant cell clones. The current understanding of this MM-specific selection of its homing involves the interaction of a number of adhesion molecules and cytokines, some of which are produced and expressed by the myeloma cells themselves, and others already present in the bone marrow meshwork. The existence of osteoclast-βtimulating proteins produced by myeloma cells, most likely bone-resorbing cytokines, have been suggested as major factors for the development of osteoclastic bone destruction in multiple myeloma. Osteoclasts express the α_vβ₃ integrin which binds to a common tripeptide sequence, RGD (Arg-Gly-Asp) as found in a number of adhesive proteins, but also in bone matrix proteins such as osteopontin and BSP. Since BSP is present in the bone marrow environment and is also expressed by the myeloma cells, it is a likely candidate to play a an integrative role in this functional interplay of osteoclast migration to the MM cells and the chemotactic attraction of the tumor cells into the bone marrow.

The determination of BSP and/or OPN in the body fluids is preferably accomplished immunologically with the aid of polyclonal or monoclonal antibodies. The polyclonal or monoclonal antibodies can be used in an established way as radioimmunoassays (RIA) or enzyme-linked immuno assays (ELISA) or immunoiuminescence. The current results were obtained with the aid of a radioimmunoassay developed in the innovator's group (Seibel M. J., Woitge H. W., Pechersdorfer M., Karmatschek M. , Horn E., Ludwig M., Armbruster F.P., Ziegler R. [1996] Serum immunoreactive bone sialoprotein as a new marker of bone turnover in metabolic and malignant bone disease. J. Clin. Endocrinol. Metab. 81: 3289-3294 and Karmatschek M., Woitge H. W., Armbruster F.P., Ziegler R., Seibel M. J.- [1997] Improved purification of human bone sialoprotein and development of a homologous radioimmunoassay. Clin. Chem. 43/11 : 2076- 2082). Clinical findings show that patients with elevated bone turnover also show elevated values for serum BSP. The same applies to patients with established bone metastases in breast cancer (Seibel et al. [1996], loc. cit). A new and diagnostically valuable finding is that in perioperative determinations of BSP and/or OPN, valuable information can be obtained about the determination of the probability of a future occurrence of bone metastases in patients with primary breast cancer and, if so, the necessity for prophylactic treatment. Furthermore, determinations of the serum levels of BSP or OPN have value in the therapeutical follow-up of individual patients and their future prognosis.

Bone sialoprotein is a highly glycosylated and phosphorylated protein with an apparent molecular weight of around 70-80 kDa (protein core, 34 kDa) that constitutes roughly 5-10% of the noncollagen, organic bone matrix (1-4). BSP is expressed in significant amounts in active osteoblasts and also to a slight extent in osteoclast-like cells and in odontoblasts (5-7). BSP shows a relatively restricted tissue distribution in comparison to other noncollagen macromolecules. The protein or its RNA is detected under physiological (tiealthy) conditions, especially in mineralized tissues such as bone or dentin as well as at the boundary between calcified cartilage and bone. BSP is otherwise immunohistochemically detectable in placental trophoblasts, in thrombocytes as well as in certain benign, but especially in malignant, breast carcinomas (Bellahcene et al. [1994, 1996], loc. cit.) and in myeloma cells.

Besides its acidic residues, BSP contains an Arg-Gly-Asp (R-G-D) amino acid sequence that serves to recognize and bind the so-called integrins. This property enables BSP to mediate and improve the adhesion of osteoblasts and osteoclasts to certain natural and artificial surfaces, to collagen type I or to calcium. In vitro, BSP promotes the nudeation of hydroxylapatite, the main component of mineralized bone matrix. In vivo, BSP improves and accelerates bone reabsorption mediated by osteoclasts (see Ross F. P., Chappel J., Alvarez J. I. [1993], Interactions between bone matrix proteins, osteopontin and bone sialoprotein and the osteoclast integrin ?_v ?₃ potentiate bone resorption, J. Biol. Chem. 268: 9901-9907; Fujisawa R., Nodasaka Y., Kuboki Y. [1995], Further characterization of the interaction between bone sialoprotein (BSP) and collagen, Calcif. Tissue Int. 56: 140-144 and Hunter G. K., Goldberg H. A. [1994], Modulation of crystal formation by bone phosphoproteins: role of glutamic acid- rich sequences in the nudeation of hydroxylapatite by bone sialoprotein, Biochem. J. 302: 175-179). From this, it can be concluded that BSP plays an important role in cell to cell and cell to matrix interactions in mineralized or mineralizing tissues, especially in the interaction between osteoclast and matrix (bone resorption). y

Osteopontin is also a structurally complex glycoprotein of bone matrix shown to influence the interaction between osteoclasts and matrix positively. Various groups have produced polyclonal antibodies against osteopontin.

Osteopontin potentially plays the same role as BSP.

Moreover, determining the probability of a future occurrence of bone metastases appears possible, in principle, not only in breast cancer but also in other primary cancers by determining the perioperative content of BSP and/or OPN in one or more body fluids. This potential is exemplified by prostate cancer, lung cancer, multiple myeloma, thyroid cancer and kidney cancer. In patients with multiple myeloma, our observations also indicate that the levels of serum BSP at any time point of the disease are predictive of individual survival.

In the clinical setting, two different scenarios are distinguishable:

1 ) The normal production of BSP or OPN in bone and mineralizing tissue that leads to normal levels of circulating BSP in body fluids, especially in serum.

2) The pathological production of BSP or OPN in malignant tissue such as within breast cancer (Bellahcene 1994, 1996 loc. cit.) or myeloma cells. In this instance, a large portion of the circulating BSP or OPN stemmed from the tumor, whereas only a certain 'background level' stemmed from the bone. This is especially so when there is no bone metastasis at the time of cancer diagnosis, i.e., a normal bone metabolism is expected.

If the determination of BSP and/or OPN is solely obtained with polyclonal antibodies, no distinction can be made between normal BSP and/or OPN and the additional amounts that originate from the tumor tissue. Fundamentally, the possibility exists to detect isoforms of BSP and/or OPN with more specific detection methods such as with special monoclonal antibodies. Isoform determination provides the ability to differentiate between the natural continual levels and the additional levels obtained from the carcinoma, myeloma and/ or metastases, where accordingly, a relapse and metastases can also be differentiated.

Finally, it was considered that patients showing a high bone turnover more quickly and more frequently develop bone metastases than patients showing a normal or even lowered bone turnover. This infers that patients with a) elevated perioperative BSP and/or OPN plus b) elevated bone turnover markers, possibly represent a separate risk group, that, in turn, require special care and therapy. The values obtained with the method according to this envention for BSP and/or OPN would then be used relative to measurements of conventional bone turnover markers such as pyridinium crosslinks or the higher molecular weight derivatives like NTX or CTX. Other bone formation markers could also be introduced such as AP, BAP and osteocalcin to identify patients at especially high risk.

The method according to this envention referring to the prediction of future bone metastasis in primary carcinomas is illustrated further by the attached figures and tables. The following can be found in particular:

Figure 1 :

Box/Whisker-Plot of preoperative serum BSP levels in patients with and without bone metastases.

Figure 2:

Kaplan-Meyer analysis organized according to the log-rank test: the probability of the occurrence of bone metastases after a median progress observation of 24 months of patients with a serum BSP value above and below 24 ng/ml at the time of primary diagnosis of a malignant breast carcinoma. Table 1 :

Characterization of the study group of 388 women with primary breast cancer at the time of diagnosis. The table distinguishes patient groups with serum BSP values above and below 24 ng/ml (calculated cutoff). The relationships between serum BSP value and established prognostic parameters were calculated using the chi-square test.

Table 2:

List of patients with resulting metastases with a median of 24 months after primary treatment of breast cancer (surgery).

Abbreviations in line 1 :

ID = identification number

Meno = Pre: premenopausal, post: postmenopausal

T = Tumor stage 1-4

N = Lymph node status

ER = Estrogen receptor

PR = Progesterone receptor

Grad = Histological grading

Meta = Site of metastasis (o: bone, v+o: liver/lungs and bone, v: liver and lungs) value = BSP level in serum

The patients in rows 1-14 experienced purely bone metastasis, the following five patients showed a combination of bone and visceral (liver and lungs) metastasis while the patients in rows 20-28 suffered purely visceral metastasis (liver and lungs). Table 3:

An overview of the median serum BSP values in different patient groups including patients with nonmalignant (benign) tumors of the breast.

These data were obtained from the following study population:

Since January 1995, preoperative whole blood was taken from all consecutive patients admitted into the University Women Hospital for therapy of primary breast cancer. Over the same period, 30 patients with benign breast tumors underwent the same procedure. In all patients with malignant tumors, a bone scan, upper abdominal sonogram and a thoracic X-ray were performed as a conclusion to surgery to exclude any metastases. All of the patients included in the present study were free from any type of metastasis at the type of initial surgery.

The characteristics of the patients with breast cancer are collated in table 1. The histopathological, biochemical and clinical prognostic factors were taken from standardized procedures and were individually:

Tumor size:

T1 = 0-2 cm in diameter

T2 = >2-5 cm in diameter

T3 = >5 cm in diameter

T4 = all tumors with infiltration

Lymph node status:

NO = No lymph nodes affected by metastasis detectable

N+ = Lymph nodes affected by metastasis detectable

Estrogen and Progesterone receptor:

Positive (ER pos/PR pos) or negative (ER neg/PR neg) detection of the respective receptors in the tumor tissue by means of dextran coated charcoal and Scatchard analysis. A positive detection comprised a concentration of > 20 fmol/mg cytosolic protein.

Menopausal status:

Premenopausal (pre) = women with regular menstrual cycles at the time of primary diagnosis.

Postmenopausal (post) = women without regular menstrual cycles at the time of primary diagnosis.

Grading:

Histological grading of the differentiation of the primary tumor.

Gl and Gil = well-differentiated tumor

Gill = poorly differentiated tumor

S-Phase:

Proportion of the cells undergoing cell division as a measure of the rate of growth of the tumor.

All patients underwent surgery and were subsequently treated with polychemo and hormone therapy. All patients were reexamined at three-monthly intervals at the University Women's Hospital and checked for the occurrence of metastasis. Except for clinical questioning (history) and examination, the individual tests (bone scan, upper abdominal sonogram and chest X-ray) were conducted according to risk in different intervals from three to twelve months. In 28 of 388 patients a metastasis in the bone, in the visceral or in both was diagnosed within a median observation time of 24 months by means of a bone scan, upper abdominal sonogram or chest X-ray (see table 2).

Determination of bone sialoprotein (BSP) and osteopontin (OPN) in serum: The concentration of BSP and OPN in serum and other body fluids were determined by immunological and biochemical methods in all patients. Whole blood was obtained from the cubital vein and the serum separated by centrifugation according to standard methods and frozen until analyzed. Urine samples were similarly obtained before surgery and stored frozen.

Various biochemical assays were used to detect BSP. An already published radioimmunoassay was employed that uses a polyclonal anti-human BSP antiserum (Kamatschek et al. [1997]). Here, 100 ul of the serum or the BSP standard were mixed with 100 ul of the tracer (¹²⁵l-labeled BSP, 1.5 ng/ml) and 100 ul of a chicken anti-BSP antiserum and incubated for 24 hours at 4°C. Then, 100 ul of donkey anti-chicken antiserum (1 :15) was added as second antibody, and incubated further for two hours at 4°C. Finally, the bound BSP-chicken- donkey antibody complex was centrifuged for ten minutes at 2000 x g. After various washing steps the radioactivity contained in the pellet was measured. The unknown concentration in the sample could then be calculated using a standard curve of known concentrations of BSP. Similarly, OPN is measured from human serum samples in the same fashion where OPN is in place of BSP and specific anti-OPN antisera replace the anti-BSP antisera. The measuring protocol is otherwise identical.

The following conclusions can be drawn from the figures and tables.

Figure 1 :

The serum levels of BSP measured preoperatively, i.e., in the mean 2 years prior to the actual metastasis, are significantly higher in those patients with later bone metastases than in those that do not show bone metastases within 24-30 months.

Figure 2:

By means of the Kaplan-Meyer curve, a relative risk of 94.07 can be calculated for a later metastasis of the tumor in bone for preoperative values of BSP of > 24 ng/ml. The difference to patients with values <24ng/ml is highly significant (p =

<0.001 ). Table 1 :

A significant relation exists between serum BSP values and tumor size (T1-4) where no relation exists to any other conventional risk parameter of breast cancer.

Table 2:

The results are shown individually.

Table 3:

Summarizes the median results for each group studied and shows that patients with bone metastases have the highest preoperative BSP values.

Taken together, it can be stated that, for the first time, the new method offers the opportunity to determine the probability of a future occurrence of bine metastasis in patients with primary breast cancer with specificity, sensitivity, speed ease and at reasonable cost. It is also possible to, if necessary, to begin a prophylactic therapy very much sooner and more efficiently. This increases the life expectancy by increasing the patients' quality of life.

The method accroding to the invention referring to the diagnosis and foilow-up, as well as to the prediction of survival in patients with multiple myeloma (MM) is illustrated by figures 3-7 and tables 4-7. These show in particular:

Table 4 and Figures 3-5 concern the diagnostic value of serum BSP measurements in multiple myeloma. When compared to healthy controls, age-adjusted serum BSP levels are significantly higher in patients with newly diagnosed multiple myeloma (p< .001 ). Seventy-three percent of patients with multiple myeloma have serum levels above the normal range as compared to 48% of the patients with MGUS, and 27% of patients with OPO (Table 4). With regard to the differentiation between healthy - 16 -

subjects and patients with multiple myeloma, results of ROC analyses were similar for serum BSP and the urinary crosslinks. For serum BSP, the area under the curve (AUC) was 0.922, whereas for urinary PYD and DPD the respective numbers were 0.895 and 0.882. At a chosen specificity of 80%, serum immunoreactive BSP had a sensitivity of 87.5%, and similar values were seen at a specificity of 95% (Figure 3).

Since the differentiation between benign and neoplastic osteoporosis-like changes is of particular clinical interest, an additional subanalysis was performed. Eleven patients with multiple myeloma were diagnosed with osteoporosis-like bone involvement without any evidence for osteolytic bone lesions. Comparing these patients with age-and sex-matched subjects of the OPO group, serum BSP was significantly higher in the multiple myeloma group (p< .05, respectively).

When patients with multiple myeloma were stratified according to the type of bone disease present at the fime of diagnosis, a similar pattern was seen for serum BSP and the urinary crosslinks (Figure 4). Highest values of all three components were found in patients with hypercalcemia of malignancy (p< .001 vs. healthy controls). When stratifying the data set according to the stage classification of Durie and Salmon, serum BSP was highest in patients with stage II (p< .001 vs. healthy controls, and vs. stage I) and stage III disease (p< .001 vs. healthy controls, and p< .01 vs. stage I) (Figure 5).

Figure 6

Concern the value of serum BSP measurements in the follow-up of multiple myeloma after chemotherapy. A subgroup of 21 patients was analyzed according to bone turnover marker levels in response to chemotherapy. In patients with a > 25% decrease in the monoclonal protein, serum BSP was reduced in all 12 patients (on average 53%; p< .001 vs. baseline) after a median observation period of 5 months (range: 1 -10 months). In contrast, in patients with a reduction of < - 17 -

25% or an increase in the monoclonal protein, serum BSP values did not change significantly after a median observation period of 6 months (range: 1-17 months). The percent change in serum BSP was significantly associated with the percent change in monoclonal protein (r=0.55, p< .001 ).

Table 5-7 and Fig 7 concern the value of serum BSP measurements for the prediction of survival of patients with multiple myeloma. Using Kaplan-Meier estimates, the probability of survival was significantly greater in the patients with initially normal serum BSP values (< 21.9 ng/mL) compared to the patients with BSP values outside the normal range (i.e. > 21.9 ng/mL; p< .001 , logrank test) (Figure 7). All other clinical and laboratory variables did not show statistically significant differences when stratifying the data set according to serum BSP using the upper limit of normal as a cut-off point. Interestingly, in a subgroup of patients with baseline levels of BSP > 70 ng/mL, none of these patients survived the observation period (median survival time: 92 days).

Univariate analyses using the Cox proportional hazards model identified the monoclonal gradient (p= .0061), β2-microglobulin (p= .0069), the plasma cell content of the bone marrow (p= .0073), and serum BSP (p= .0126) as prognostic indices for shortened survival. However, a stepwise multivariate regression analysis (including all variables with a p-value of < 0.1 in the univariate analysis) revealed that only BSP and the monoclonal gradient independently contributed to the prediction of surviva Table 6 and 7).

The data on the value of BSP in multiple myeloma (MM) were obtained from the following study population. A total of 388 individuals was included in the study. Routine biochemistry was performed in all participants, including serum calcium, creatinine, albumine, β2-microglobulin, and serum and urine electrophoresis. The control group was an ambulatory population of 139 subjects without any evidence of significant skeletal or non-skeletal disease. All subjects had radiographs of the lumbar and thoracic spine (Multiplanimat Siemens, Erlangen, Germany), and bone mineral density was determined in an anterior-posterior position at the lumbar spine by dual x-ray absorptiometry (QDR 1000, Hologic, Waltham, MA). Subjects with vertebral fractures, a bone mineral density of less than 2 SD of the age- and sex-matched mean, significant degenerative disease of the spine, or abnormal laboratory results, were excluded. None of the controls was taking any medication known to affect bone metabolism, including bisphosphonates, glucocorticoids, HRT, osteotropic vitamins or calcium supplements.

Patients with multiple myeloma (MM): A total number of 69 patients with multiple myeloma participated in the study. Sixty-two of the patients were newly diagnosed and had not been treated with chemotherapy, glucocorticoids, or radiotherapy prior to the study enrollment. Only these patients were analyzed in the cross- sectional study, and in the evaluation of serum BSP and other biochemical markers of bone turnover as prognostic factors. The diagnosis of multiple myeloma was based on i) bone marrow plasma cell density, ii) radiological evidence of bone involvement and/or iii) the presence of hematological impairment or renal failure. All patients were characterized by the presence of a monoclonal paraproteinemia and/or paraproteinuria (74% IgG, 18% IgA, 8% others). In all subjects, plain radiographs were taken of the skull, spine, pelvis, and of painful regions outside the axial skeleton. Bone marrow biopsies were performed in Yamshidi technique. In all patients, serum creatinine was less than 4 mg/dL. Patients were followed over 4 years, with a median observation time of 427 days (range: 1-1550 days).

The staging of multiple myeloma was based upon the system of Durie and Salmon. At the time of diagnosis, 14 patients (23%) were diagnosed with stage I MM, 15 patients (24%) with stage II, and 33 patients (53%) presented with stage III disease (n=33). One patient with stage I, 2 patients with stage II, and 6 patients with stage III disease had serum creatinine levels of > 2 mg/dL (substage B). For further analyses, patients were subdassified according to bone disease: i) no bone disease; ii) osteoporosis-like bone involvement without hypercalcemia; iii) osteolysis without hypercalcemia; and iv) hypercalcemia of malignancy.

A protocol including vincristin, melphaian, cyclophosphamide, and prednisolone was applied in the vast majority of patients, except for three cases with indolent stage I multiple myeloma and one patient who refused cytostatic treatment. None of the patients achieved a complete remission, e.g. a monoclonal protein undetectable by immunofixation. 21 of these patients had follow-up measurements on bone turnover markers and were included in the analyses of therapeutic monitoring.

For validation of serial BSP measurements, we performed follow-up sample collection in 14 patients with MGUS after a median observation time of 18 months (range: 3-32 months). On average, serum BSP levels were unchanged on follow- up compared to the baseline values.

Patients with vertebral osteoporosis (OPO): Seventy-one patients with newly diagnosed and untreated osteoporosis were included. Of the 48 women in this group, all were postmenopausal with a median duration of menopause of 18 years (range: 2-41 years). The diagnosis of OPO was based upon the presence of at least one vertebral fracture (wedge, compression, or biconcave) that was not attributable to adequate spinal trauma, plus a lumbar bone mineral density below 2.5 SD of the age- and sex-matched mean. None of the subjects in this group had a history of malignant disease, nor were there any signs of plasma cell dyscrasia or secondary osteoporosis. At the time of study enrollment, all patients had normal renal and hepatic function, and none was given any medication known to interfere with bone turnover, including bisphosphonates, glucocorticoids, HRT, osteotropic vitamins or calcium supplements. - 20 -

Determination of bone sialoprotein in serum was identical with the procedure described above.

The following conclusions can be drawn from the figures and tables.

Figure 3

Discriminative power of serum BSP, serum osteocalcin (OC) and of urinary pyridinoline (PYD) and deoxypyridinoline (DPD) in patients with multiple myeloma versus healthy subjects (graph A) and patients with MGUS (graph B) or primary vertebral osteoporosis (graph C). Results of ROC analyses are provided as the respective area under the curve (AUC), i.e. the mean sensitivity across the range of possible specificities. Serum BSP has highest diagnostic power when comparing healthy versus myeloma patients (graph A). At a chosen specificity of 80%, serum immunoreactive BSP had a sensitivity of 87.5%, and similar values were seen at a specificity of 95%. In contrast, and just for comparison, serum osteocalcin has no value in this respect.

Figure 4

Serum BSP and osteocalcin (OC), and urinary pyridinoline (PYD) and deoxypyridinoline (DPD) in patients with multiple myeloma stratified according to the type of bone disease diagnosed. Values are expressed as z scores. Full lines within boxes represent the group mean. The 0-line is the mean of the normal (healthy) group, dotted lines are +/- 2 SD of the normal value, i.e. the upper and lower limit of normal. It is evident that all forms of multiple myeloma are associated with elevated serum BSP levels, and that patients with hypercalcemia have highest values.

Abbreviations: No BD, no bone disease visible at examination; O'po, osteoporosis like disease; O'ly, osteolysis visible; HOM, hypercalcemia. P-values are < 0.05 (a), < 0.01 (b), < 0.001 (c) versus healthy controls. Figure 5

Serum BSP and osteocalcin (OC), and urinary pyridinoline (PYD) and deoxypyridinoline (DPD) in patients with multiple myeloma stratified according to stage of disease. Values are expressed as z scores. Full lines within boxes represent the group mean. The 0-line is the mean of the normal (healthy) group, dotted lines are +/- 2 SD of the normal value, i.e. the upper and lower limit of normal. It is evident that stage III multiple myeloma is associated with elevated serum BSP levels.

P-values are < 0.05 (a, d), < 0.01 (b, e), < 0.001 (c,f ) versus healthy controls (a-c) or stage I MM (d-f) respectively.

Figure 6

Serum BSP in patients with multiple myeloma before and after chemotherapy (VMCP regimen) stratified according to response (left panel, > 25% decrease) and non-response (right panel, < 25% decrease) as determined by quantification of monoclonal protein. n= 21 , the median observation period was 6 months. Filled circles are group means, small diamonds connected by dotted lines are individual patients. It is evident that a response to chemotherapy is associated with a significant drop in serum BSP levels. In patients with a > 25% decrease in the monoclonal protein, serum BSP was on average reduced by 53% (p< .001 vs. baseline). In contrast, in patients with a reduction of < 25% or an increase in the monoclonal protein, serum BSP values did not change significantly. The percent change in serum BSP was significantly associated with the percent change in monoclonal protein (r=0.55, p< .001 ).

Figure 7

By means of the Kaplan-Meyer curve, this graph shows the predictive value of serum BSP measurements for the survival of patients with multiple myeloma (any stage). A serum BSP level above 21.9 ng/mL is associated with a significantly shorter survival of the patient. Differences in survival probabilities from the time of diagnosis until death or latest follow-up were estimated using logrank tests. Thus, the probability of survival was significantly greater in the patients with initially normal serum BSP values (< 21.9 ng/mL) compared to the patients with BSP values outside the normal range (i.e. > 21.9 ng/mL; p< .001 ) All other clinical and laboratory variables did not show statistically significant differences when stratifying the data set according to serum BSP using the upper limit of normal as a cut-off point.

Table 4:

Mean levels (95% confidence intervals in brackets) of serum BSP, serum osteocalcin (OC), urinary pyridinoline (PYD) and urinary deoxypyridinoline (DPD) in healthy adults, patients with multiple myeloma (MM), monoclonal gammopathy of unknown significance (MGUS) or osteoporosis (OPO). When compared to healthy controls, age-adjusted serum BSP levels are significantly higher in patients with newly diagnosed multiple myeloma (p< .001 ). Seventy-three percent of patients with multiple myeloma have serum levels above the normal range as compared to 48% of the patients with MGUS, and 27% of patients with osteoporosis.

Abbreviation: ULN, upper limit of normal.

Table 5:

Differences in disease characteristics according to BSP levels in patients with multiple myeloma (median, range in brackets). This table compliments figure 7, which shows the predictive value of elevated serum BSP levels in patients with multiple myeloma.

Table 6 and 7:

Univariate analyses using the Cox proportional hazards model identified the monoclonal gradient (p= .0061 ), β2-microglobulin (p= .0069), the plasma cell - 23 -

content of the bone marrow (p= .0073), and serum BSP (p= .0126) as prognostic indices for shortened survival. Stepwise multivariate regression analysis (including all variables with a p-value of < 0.1 in the univariate analysis) revealed that only BSP and the monoclonal gradient independently contributed to the prediction of survival

Taken together, it can be stated that, for the first time, the new procedure offers the opportunity to i) improve the differentiation of multiple myeloma from healthy and osteoporotic patients ii) improve the follow-up of patients with multiple myeloma after chemotherapy, independently of conventional parameters of disease activity iii) improve the prediction of the outcome of the disease (survival of the patient).

24 -

Tab. 1 : Characterization of the Population

Prognostic Serum BSP Factor n > 24 % of < 24 p-value ng/mL n ng/mL

Tumor Size 0.0001

T1 195 8 4 187

T2 156 15 10 141

T3 25 2 8 23

T4 12 4 33 8

LN status 0.251

NO 252 16 6 236

N+ 136 13 9 123

Estrogenreceptor 0.885

ER + 226 18 8 208

ER - 146 1 1 7 120

Progesteronrec. 0.831

PR+ 228 19 8 209

PR- 130 10 8 120

Menopausal Status 0.539 pre 127 8 6 1 19 post 261 21 8 240

Grading 1 .000 l + ll 247 19 8 228

III 130 10 8 120

S-Phase Proportion 0.243

< 5% 106 1 1 10 95

> 5% 252 17 7 235 Tab. 3Serum BSP values in patients with benign and metastasized tumors

Group n Median serum BSP before surgery

(at time of diagnosis) (ng/mL)

Patients with benign tumors 30 9.7

Patients with visceral metastases 9 12.1

Patients with mixed vizcerale (lung, liver) and osseous metastases 5 27.5

Patients with purely osseous metastases 14 44.6

- 26 -

Table 4: Biochemical markers of bone turnover in the disease groups [median (range)]

Healthy MM MGUS OPO

Adults

ULN >ULN >UL >ULN

N

BSP 10.1 20.9 27.6 73% 18.6 48% 15.8 27%

(ng/mL) (3.4- (4.8- (3.7-76.1 )^c'^f (7.4-82.2)^c'^f

27.4) 120)^c

OC 16 34.4 24 30% 19.1 4% 22.6 10%

(ng/mL) (6.9- (6.1- (4.5-46)^d (12.2-45)³

37.1 ) 134)^b

PYD 23.8 50.6 50.7 50% 27.6 11% 29.8 13%

(nM/mM (9.9- (15- (8.9-63.6)^a'^e (11-101)^{b e}

Cr) 79.7) 774)^c

DPD 5.5 12.0 13.6 57% 6.1 14% 7.8 13%

(nM/mM (2- (3.9- (2.2-21.8)^{a e} (2.7-21.5) ^e

Cr) 18.7) 314)^c

^a p< .05, p< .01 , ^c p< .001 vs. healthy adults. ^d p< .05, ^e p< .01 , ^f p< .001 vs. MM. Abbreviations: see Table 1 . ULN, upper limit of normal (95% confidence interval); S-BSP, serum immunoreactive bone sialoprotein; S-OC, serum osteocalcin; U- PYD, urinary pyridinoline; U-DPD, urinary deoxypyridinoline. — ^, / ^—

Table 5: Differences in disease characteristics according to BSP levels in patients with MM [median (range)]

BSP < 21.9 ng/ml (I) BSP > 21.9 ng/ml (II)

Sex (M/F)^* 6/9 20/27

Age (yrs) 70 (28-90) 67 (39-87)

Paraproteinemia igG 80% 78% igA 20% 15%

Others 0% 7%

S-Cr (mg/dL) 1.1 (0.6-3.3) 1.0(0.41-3.9)

S-Ca (mM/L) 2.28 (2.03-2.88) 2.39(1.99-3.52)^p=0077

U-Ca (mM/mM Cr) 0.3S (0.07-1.02) 0.51 (0.01-23.4)^p=0086 β2 (mg/L) 3.8(1.6-13.4) 4.4(1.4-16.3) t-Prot (mg/dL) 10.1 (5.7-14.2) 8.7 (6.0-14.3)

CRP (mg/dL) 0.52(0.5-10) 0.73(0.5-7.1)

S-Fer (μg/L) 88 (30-1483) 223(10-2315)

Hb (g/dL) 11.3(7.9-13.8) 11 (5.1-16.5)

MG (%) 42.8(17.5-64.7) 35.6 (0-65.7)

PCC (%) 15(0-80) 50 (0-95)^p=0062

^aρ< .05; p< .01;^cp<.001 vs. I. Abbreviations: see Table y. - 2o -

Table 6: Univariate Cox Proportional Hazards Analysis*

Variable Coefficient SE χ² P=

MG 0.0327 0.0121 7.534 0.0061 β2 0.1174 0.0443 7.299 0.0069

PCC 0.0180 0.0069 7.208 0.0073

BSP 0.0135 0.0056 6.219 0.0126 t-Prot 0.1467 0.0768 3.720 0.0538

Stage³ 0.4550 0.2470 3.521 0.0606

^*Univariate survival analyses were performed for all variables listed in table yy.

Only those variables (continuous, except ^a) are presented that reached p-values of

< 0.1.

Abbreviations: see Table y.

Table 7: Multivariate Cox Proportional Hazards Analysis*

Variable Coefficient SE χ² P=

BSP 0.0133 0.0054 6.382 0.0115 MG 0.0937 0.0346 4.826 0.0280

*AII variables listed in table y were included in the stepwise multivariate regression analysis. Only those variables are presented that reached p-values of < 0.05. Abbreviations: see Table y.

Claims

1. A method to determine the probability of a future occurrence of bone metastases in patents with primary carcinomas and, if appropriate, the necessity for prophylactic treatment, characterized by the perioperative determination of the amount of BSP (bone sialoprotein) and/or OPN (osteopontin) or their fragments or isoforms from one or more body fluids.

2. The method as described in claim 1 characterized by the postoperative monitoring of BSP and/or OPN or their fragments or isoforms from one or more body fluids.

3. A method to^" improve the differentiation of multiple myeloma from healthy and osteoporotic patients, characterized by the determination of the amount of BSP (bone sialoprotein) and/or OPN (osteopontin) or their fragments or isoforms from one or more body fluids.

4. A method to determine the response of patients with multiple myeloma to specific therapy or treatments, characterized by the determination of the amount of BSP (bone sialoprotein) and/or OPN (osteopontin) or their fragments or isoforms from one or more body fluids.

5. A method to predict the outcome of the disease and the survival of the patient with multiple myeloma, characterized by the determination of the amount of BSP (bone sialoprotein) and/or OPN (osteopontin) or their fragments or isoforms from one or more body fluids.

6. The method as described in claims 1-5 characterized by the immunological determination of BSP and/or OPN or their fragments or isoforms using polyclonal or monoclonal antibodies.