Low kindlin-3 levels in osteoclasts of kindlin-3 hypomorphic mice result in osteopetrosis due to leaky sealing zones.

Osteoclasts form special integrin-mediated adhesion structures called sealing zones that enable them to adhere to and resorb bone. Sealing zones consist of densely packed podosomes tightly interconnected by actin fibers. Their formation requires the presence of the hematopoietic integrin regulator kindlin-3 (also known as Fermt3). In this study, we investigated osteoclasts and their adhesion structures in kindlin-3 hypomorphic mice expressing only 5-10% of the kindlin-3 level of wild-type mice. Low kindlin-3 expression reduces integrin activity, results in impaired osteoclast adhesion and signaling, and delays cell spreading. Despite these defects, in vitro-generated kindlin-3-hypomorphic osteoclast-like cells arrange their podosomes into adhesion patches and belts, but their podosome and actin organization is abnormal. Remarkably, kindlin-3-hypomorphic osteoclasts form sealing zones when cultured on calcified matrix in vitro and on bone surface in vivo. However, functional assays, immunohistochemical staining and electron micrographs of bone sections showed that they fail to seal the resorption lacunae properly, which is required for secreted proteinases to digest bone matrix. This results in mild osteopetrosis. Our study reveals a new, hitherto understudied function of kindlin-3 as an essential organizer of integrin-mediated adhesion structures, such as sealing zones.

Cdc42 in osterix-expressing cells alters osteoblast behavior and myeloid lineage commitment.

Osteoblasts are not only responsible for bone formation. They also support hematopoiesis. This requires responding to cues originating from several signaling pathways, a task performed by Rho GTPases. We therefore examined several transgenic mouse models and used inhibitors of Cdc42 in vitro. Deletion of Cdc42 in vivo using the Osterix promoter suppressed osteoblast function, while its deletion in differentiating osteoblasts using the Collagen-α1(I) promoter decreased osteoblast numbers. In both cases, bone mineral density diminished confirming the importance of Cdc42. Evaluation of hematopoiesis revealed that deletion of Cdc42 using the Osterix, but not the Collagen-α1(I) promoter increased the common myeloid progenitors (CMPs) in the bone marrow as well as the erythrocytes and the thrombocytes/platelets in peripheral blood. Causality between Cdc42 loss in early osteoblasts and increased myelopoiesis was confirmed in vitro. Work in vitro supported the conclusion that interleukin-4 mediated the increase in myelopoiesis. Thus, Cdc42 is required for healthy bone through regulation of bone formation in Osterix-expressing osteoblasts and the number of osteoblasts in differentiating osteoblasts. In addition, its expression in early osteoblasts/stromal cells modulates myelopoiesis. This highlights the importance of osteoblasts in regulating hematopoiesis.

Inhibition of fibronectin accumulation suppresses tumor growth.

Understanding how the extracellular matrix affects cancer development constitutes an emerging research field. Fibronectin and collagen are two intriguing matrix components found in cancer. Large concentrations of fibronectin or collagen type I have been implicated in poor prognosis in patients. In a mouse model, we had shown that genetically decreasing circulating fibronectin resulted in smaller tumors. We therefore aimed to manipulate fibronectin pharmacologically and determine how cancer development is affected. Deletion of fibronectin in human breast cancer cells (MDA-MB-231) using shRNA (knockdown: Kd) improved survival and diminished tumor burden in a model of metastatic lesions and in a model of local growth. Based on these findings, it seemed reasonable to attempt to prevent fibronectin accumulation using a bacterial derived peptide called pUR4. Treatment with this peptide for 10 days in the breast cancer local growth model or for 5 days in a melanoma skin cancer model (B16) was associated with a significant suppression of cancer growth. Treatment aimed at inhibiting collagen type I accumulation without interfering with fibronectin could not affect any changes in vivo. In the absence of fibronectin, diminished cancer progression was due to inhibition of proliferation, even though changes in blood vessels were also detected. Decreased proliferation could be attributed to decreased ERK phosphorylation and diminished YAP expression. In summary, manipulating fibronectin diminishes cancer progression, mostly by suppressing cell proliferation. This suggests that matrix modulation could be used as an adjuvant to conventional therapy as long as a decrease in fibronectin is obtained.

Fibronectin and Its Receptors in Hematopoiesis.

Fibronectin is a ubiquitous extracellular matrix protein that is produced by many cell types in the bone marrow and distributed throughout it. Cells of the stem cell niche produce the various isoforms of this protein. Fibronectin not only provides the cells a scaffold to bind to, but it also modulates their behavior by binding to receptors on the adjacent hematopoietic stem cells and stromal cells. These receptors, which include integrins such as α4ß1, α9ß1, α4ß7, α5ß1, αvß3, Toll-like receptor-4 (TLR-4), and CD44, are found on the hematopoietic stem cell. Because the knockout of fibronectin is lethal during embryonal development and because fibronectin is produced by almost all cell types in mammals, the study of its role in hematopoiesis is difficult. Nevertheless, strong and direct evidence exists for its stimulation of myelopoiesis and thrombopoiesis using in vivo models. Other reviewed effects can be deduced from the study of fibronectin receptors, which showed their activation modifies the behavior of hematopoietic stem cells. Erythropoiesis was only stimulated under hemolytic stress, and mostly late stages of lymphocytic differentiation were modulated. Because fibronectin is ubiquitously expressed, these interactions in health and disease need to be taken into account whenever any molecule is evaluated in hematopoiesis.

The Rho GTPase RAC1 in Osteoblasts Controls Their Function.

The regulation of the differentiation of the bone-forming cells, the osteoblasts, is complex. Many signaling pathways converge on the master regulator of osteoblast differentiation Runx2. The role of molecules that integrate several signaling pathways such as the Rho GTPases need to be better understood. We, therefore, asked at which stage Rac1, one of the Rho GTPase, is needed for osteoblast differentiation and whether it is involved in two pathways, the anabolic response to parathyroid hormone and the stimulatory effect of fibronectin isoforms on integrins. Genetic deletion of Rac1 in preosteoblasts using the osterix promoter diminished osteoblast differentiation in vitro. This effect was however similar to the presence of the promoter by itself. We, therefore, applied a Rac1 inhibitor and confirmed a decrease in differentiation. In vivo, Rac1 deletion using the osterix promoter decreased bone mineral density as well as histomorphometric measures of osteoblast function. In contrast, deleting Rac1 in differentiating osteoblasts using the collagen α1(I) promoter had no effects. We then evaluated whether intermittent parathyroid hormone (PTH) was able to affect bone mineral density in the absence of Rac1 in preosteoblasts. The increase in bone mineral density was similar in control animals and in mice in which Rac1 was deleted using the osterix promoter. Furthermore, stimulation of integrin by integrin isoforms was able to enhance osteoblast differentiation, despite the deletion of Rac1. In summary, Rac1 in preosteoblasts is required for normal osteoblast function and bone density, but it is neither needed for PTH-mediated anabolic effects nor for integrin-mediated enhancement of differentiation.

A Subpopulation of Stromal Cells Controls Cancer Cell Homing to the Bone Marrow.

Breast and prostate cancer cells home to the bone marrow, where they presumably hijack the hematopoietic stem cell niche. We characterize here the elusive premetastatic niche by examining the role of mesenchymal stromal cells (MSC) in cancer cell homing. Decreasing the number of MSC pharmacologically enhanced cancer cell homing to the bone marrow in mice. In contrast, increasing the number of these MSCs by various interventions including G-CSF administration diminished cancer cell homing. The MSC subpopulation that correlated best with cancer cells expressed stem, endothelial, and pericytic cell markers, suggesting these cells represent an undifferentiated component of the niche with vascular commitment. In humans, a MSC subpopulation carrying markers for endothelial and pericytic cells was lower in the presence of cytokeratin+ cells in bone marrow. Taken together, our data show that a subpopulation of MSC with both endothelial and pericytic cell surface markers suppresses the homing of cancer cells to the bone marrow. Similar to the presence of cytokeratin+ cells in the bone marrow, this MSC subpopulation could prove useful in determining the risk of metastatic disease, and its manipulation might offer a new possibility for diminishing bone metastasis formation.Significance: These findings establish an inverse relationship between a subpopulation of mesenchymal stromal cells and cancer cells in the bone marrow. Cancer Res; 78(1); 129-42. ©2017 AACR.

An O-Glycosylation of Fibronectin Mediates Hepatic Osteodystrophy Through α4ß1 Integrin.

Patients with cholestatic liver disease experience increased fracture risk. Higher circulating levels of a fibronectin isoform called oncofetal fibronectin (oFN) were detected in a subset of such patients. Administering this isoform to mice suppresses osteoblast differentiation and diminishes bone mineral density in vivo, suggesting it is responsible for bone loss in cholestatic liver disease. The aim of this study was to define the mechanism by which oFN affects osteoblast function and evaluate possible modifiers in experimental hepatic osteodystrophy. The fibronectin isoform oFN is characterized by the presence of various glycosylations. In line with this, adding oFN that underwent enzymatic O-deglycosylation to osteoblasts normalized nodule formation in vitro. Of three possible O-glycosylation sites in oFN, only a mutation at AA 33 of the variable region or binding of this glycosylated site with an antibody normalized osteoblast differentiation. Because the responsible site is located in the variable region of fibronectin, which binds to α4ß1 or α4ß7 integrins, these integrins were evaluated. We show that integrin α4ß1 mediates the inhibitory effect of oFN both in vitro as well as in vivo. In a hepatic osteodystrophy mouse model, we demonstrate that liver fibrosis is associated with increased circulating oFN and diminished BMD. In addition, trabecular bone loss induced by oFN injection or fibrosis induction could be prevented by either administering an antibody that binds to α4 integrin (PS/2) or the CS1 peptide, which contains a binding site for α4ß1 integrin. In summary, oFN inhibits osteoblast activity. This is because of an O-glycosylation in the variable region that results in decreased integrin-mediated signaling. This deleterious effect can be thwarted by binding α4ß1 integrin. Thus, we have characterized the defect and the receptor mediating bone loss in patients with hepatic osteodystrophy and evaluated possible therapeutic interventions in a murine model. © 2016 American Society for Bone and Mineral Research.

EDA-Fibronectin Originating from Osteoblasts Inhibits the Immune Response against Cancer.

Osteoblasts lining the inner surface of bone support hematopoietic stem cell differentiation by virtue of proximity to the bone marrow. The osteoblasts also modify their own differentiation by producing various isoforms of fibronectin (FN). Despite evidence for immune regulation by osteoblasts, there is limited knowledge of how osteoblasts modulate cells of the immune system. Here, we show that extra domain A (EDA)-FN produced by osteoblasts increases arginase production in myeloid-derived cells, and we identify α5ß1 as the mediating receptor. In different mouse models of cancer, osteoblasts or EDA-FN was found to up-regulate arginase-1 expression in myeloid-derived cells, resulting in increased cancer growth. This harmful effect can be reduced by interfering with the integrin α5ß1 receptor or inhibiting arginase. Conversely, in tissue injury, the expression of arginase-1 is normally beneficial as it dampens the immune response to allow wound healing. We show that EDA-FN protects against excessive fibrotic tissue formation in a liver fibrosis model. Our results establish an immune regulatory function for EDA-FN originating from the osteoblasts and identify new avenues for enhancing the immune reaction against cancer.

Inhibition of fibronectin deposition improves experimental liver fibrosis.

BACKGROUND & AIMS: Common pathogenic steps in liver fibrosis are inflammation and accumulation of extracellular matrix proteins including collagen, which lead to disruption of tissue microarchitecture and liver dysfunction. Adequate fibronectin fibril formation is required for collagen matrix deposition in several cell types in vitro. We therefore hypothesized that preventing fibronectin fibril assembly will result in decreased collagen matrix accumulation, and hence diminish liver injury associated with fibrosis. METHODS: In vitro studies on hepatic stellate cells and in vivo studies in mice were performed. RESULTS: In vitro studies on hepatic stellate cells confirmed that a fibronectin assembly inhibitor, pUR4 diminishes the amount of both fibronectin and collagen, accumulating in the extracellular matrix, without affecting their production. Induction of fibrosis using CCl4 or DMN was therefore combined with pUR4-treatment. pUR4 normalized the amount of fibrotic tissue that accumulated with injury, and improved liver function. Specifically, pUR4-treatment decreased collagen accumulation, without changing its mRNA expression. Most interestingly, we did not detect any changes in Kupffer cell numbers (F4/80+) or α-smooth muscle actin expressing hepatic stellate cell numbers. Further, there was no impact on TGF-ß or TNF-α. Thus, in line with the in vitro findings, decreased fibrosis is due to inhibition of matrix accumulation and not a direct effect on these cells. CONCLUSIONS: In summary, a peptide that blocks fibronectin deposition results in decreased collagen accumulation and improved liver function during liver fibrogenesis. Thus, fibronectin matrix modulation offers a therapeutic benefit in preclinical models of liver fibrosis.

Blood clot formation does not affect metastasis formation or tumor growth in a murine model of breast cancer.

Cancer is associated with increased fracture risk, due either to metastasis or associated osteoporosis. After a fracture, blood clots form. Because proteins of the coagulation cascade and activated platelets promote cancer development, a fracture in patients with cancer often raises the question whether it is a pathologic fracture or whether the fracture itself might promote the formation of metastatic lesions. We therefore examined whether blood clot formation results in increased metastasis in a murine model of experimental breast cancer metastasis. For this purpose, a clot was surgically induced in the bone marrow of the left tibia of immundeficient mice. Either one minute prior to or five minutes after clot induction, human cancer cells were introduced in the circulation by intracardiac injection. The number of cancer cells that homed to the intervention site was determined by quantitative real-time PCR and flow cytometry. Metastasis formation and longitudinal growth were evaluated by bioluminescence imaging. The number of cancer cells that homed to the intervention site after 24 hours was similar to the number of cells in the opposite tibia that did not undergo clot induction. This effect was confirmed using two more cancer cell lines. Furthermore, no difference in the number of macroscopic lesions or their growth could be detected. In the control group 72% developed a lesion in the left tibia. In the experimental groups with clot formation 79% and 65% developed lesions in the left tibia (p = ns when comparing each experimental group with the controls). Survival was similar too. In summary, the growth factors accumulating in a clot/hematoma are neither enough to promote cancer cell homing nor support growth in an experimental model of breast cancer bone metastasis. This suggests that blood clot formation, as occurs in traumatic fractures, surgical interventions, and bruises, does not increase the risk of metastasis formation.

Metastasis of prostate cancer and melanoma cells in a preclinical in vivo mouse model is enhanced by L-plastin expression and phosphorylation.

BACKGROUND: Tumor cell migration and metastasis require dynamic rearrangements of the actin cytoskeleton. Interestingly, the F-actin cross-linking and stabilizing protein L-plastin, originally described as a leukocyte specific protein, is aberrantly expressed in several non-hematopoietic malignant tumors. Therefore, it has been discussed as a tumor marker. However, systematic in vivo analyses of the functional relevance of L-plastin for tumor cell metastasis were so far lacking. METHODS: We investigated the relevance of L-plastin expression and phosphorylation by ectopical expression of L-plastin in human melanoma cells (MV3) and knock-down of endogenous L-plastin in prostate cancer (PC3M). The growth and metastatic potential of tumor cells expressing no L-plastin, phosphorylatable or non-phosphorylatable L-plastin was analyzed in a preclinical mouse model after subcutaneous and intracardial injection of the tumor cells. RESULTS: Knock-down of endogenous L-plastin in human prostate carcinoma cells led to reduced tumor cell growth and metastasis. Vice versa, and in line with these findings, ectopic expression of L-plastin in L-plastin negative melanoma cells significantly increased the number of metastases. Strikingly, the metastasis promoting effect of L-plastin was not observed if a non-phosphorylatable L-plastin mutant was expressed. CONCLUSIONS: Our data provide the first in vivo evidence that expression of L-plastin promotes tumor metastasis and, importantly, that this effect depends on an additionally required phosphorylation of L-plastin. In conclusion, these findings imply that for determining the importance of tumor-associated proteins like L-plastin a characterization of posttranslational modifications is indispensable.

Complement activation correlates with liver necrosis and fibrosis in chronic hepatitis C.

Chronic hepatitis C viral infection modulates complement. The aim of this study was to determine whether complement analysis predicts liver inflammation and fibrosis in patients with chronic hepatitis C. 50 chronic hepatitis C patients who underwent a liver biopsy were compared to 50 healthy controls and 35 patients with various liver diseases. Total plasma complement activity (CH50) in plasma was diminished in hepatitis C patients suggesting complement activation. This decrease correlated with increased necrosis (r = -0.24, p < 0.05), and patients with levels below the normal range had a higher METAVIR activity score reflecting enhanced inflammation. SC5b-9, a marker of complement activation, correlated with inflammation (r = 0.40, p < 0.05), activity (r = 0.42, p < 0.05), and fibrosis scores (r = 0.49, p < 0.05). Finally, the prevalence of C1q auto-antibodies was higher in hepatitis C patients, and their presence was associated with increased inflammation and seemed to affect fibrosis. We conclude that complement-induced liver inflammation contributes to fibrosis in patients with chronic hepatitis C.

Circulating fibronectin controls tumor growth.

Fibronectin is ubiquitously expressed in the extracellular matrix, and experimental evidence has shown that it modulates blood vessel formation. The relative contribution of local and circulating fibronectin to blood vessel formation in vivo remains unknown despite evidence for unexpected roles of circulating fibronectin in various diseases. Using transgenic mouse models, we established that circulating fibronectin facilitates the growth of bone metastases by enhancing blood vessel formation and maturation. This effect is more relevant than that of fibronectin produced by endothelial cells and pericytes, which only exert a small additive effect on vessel maturation. Circulating fibronectin enhances its local production in tumors through a positive feedback loop and increases the amount of vascular endothelial growth factor (VEGF) retained in the matrix. Both fibronectin and VEGF then cooperate to stimulate blood vessel formation. Fibronectin content in the tumor correlates with the number of blood vessels and tumor growth in the mouse models. Consistent with these results, examination of three separate arrays from patients with breast and prostate cancers revealed that a high staining intensity for fibronectin in tumors is associated with increased mortality. These results establish that circulating fibronectin modulates blood vessel formation and tumor growth by modifying the amount of and the response to VEGF. Furthermore, determination of the fibronectin content can serve as a prognostic biomarker for breast and prostate cancers and possibly other cancers.

Circulating fibronectin isoforms predict the degree of fibrosis in chronic hepatitis C.

OBJECTIVE: Hepatic stellate cells only produce fibronectin isoforms in disease states. The isoform-defining domains can be detected in the blood circulation. This study examines whether circulating levels of fibronectin isoforms show a relationship with liver fibrosis on histology in patients with chronic hepatitis C. MATERIAL AND METHODS: In a prospective study, 50 patients with chronic hepatitis C who underwent a liver biopsy were compared to 50 matched controls and 35 patients with other liver conditions. RESULTS: Circulating levels of the fibronectin isoforms were significantly higher in patients with chronic hepatitis C compared to healthy controls [oncofetal fibronectin (oFN) 2.45 +/- 0.17 versus 1.76 +/- 0.16 mg/l, P < 0.005; extra domain-A (EDA) 1.05 +/- 0.06 versus 0.86 +/- 0.06 mg/l, P < 0.05; and extra domain-B (EDB) 14.55 +/- 0.74 versus 9.31 +/- 0.58 mg/l, P < 0.001], even though total fibronectin was lower (198.9 +/- 3.5 versus 343.6 +/- 14.5 mg/l, P < 0.001). A correlation with the fibrosis score was found for both oFN (r = 0.46, P < 0.005) and EDA (r = 0.51, P < 0.001). The combination of an elevation in both markers (oFN and EDA) in the upper quartile was associated with a specificity of > 99% for predicting significant fibrosis (stages 2-4) and 95% for predicting advanced fibrosis (stages 3-4). A combination of decreased values in the lowest tertile for both markers had a specificity of 94% for excluding significant fibrosis. Based on these findings, 30% of the patients scheduled for a liver biopsy could be correctly classified as having or not having significant fibrosis. The remainder would have to proceed with a biopsy. CONCLUSION: Circulating fibronectin isoforms produced by activated stellate cells represent a viable marker for the presence of significant fibrosis or a lack thereof.

Prospective study on warfarin and regional chemotherapy in patients with pancreatic carcinoma.

AIMS: The aim is to prospectively examine the effect of regional gemcitabine and mitomycin-C with systemic gemcitabine together with warfarin in patients with inoperable pancreatic carcinoma, and compare the effect to systemic gemcitabine alone. METHODS: Seventeen patients received 1.25 mg of warfarin daily, gemcitabine 800 mg/m2 on day 1 and mitomycin-C 8 mg/m2 on day 2 regionally and gemcitabine 800 mg/m2 on day 14 peripherally. The cycle was repeated every 4 weeks. RESULTS: Median survival since presentation was 6.8 months, while median total survival was 9.6 months. Excluding the 3 patients who died before receiving any therapy, the median survival since presentation resulted in 10.7 months and the median total survival, 12.7 months. One patient developed bleeding that required transfusion and 2 patients developed anemia (Grades III/IV). Comparing these data to historical controls of large cohorts supports the notion that this regimen offers a viable alternative to systemic gemcitabine alone. CONCLUSION: A regimen consisting of regional gemcitabine and mitomycin-C with systemic gemcitabine and low-dose warfarin compares favorably to the gold standard of systemic gemcitabine. These data suggest the feasibility of a large prospective study on the use of warfarin and combined regional and systemic chemotherapy in patients with pancreatic carcinoma.

Coagulation, anticoagulation and pancreatic carcinoma.

An increase in coagulation in patients with pancreatic carcinoma has long been documented. In this Review, we present what is known about the pathophysiology of increased coagulation in cancer and how it applies to pancreatic carcinoma. The relationship between the activation of coagulation or symptomatic thromboembolic disease and the development of pancreatic carcinoma is explored. Data on the relationship between thromboembolic disease and the behavior of pancreatic cancer before, during or after a diagnosis is made are also reviewed. Finally, the rationale and evidence for the use of oral anticoagulants or heparin in patients with pancreatic carcinoma is presented. This Review is a critical appraisal of what is known, and when the evidence is acceptable, on the subject of thromboembolism, anticoagulation, and treatment with anticoagulants in patients with pancreatic carcinoma.

The role of the receptor activator of nuclear factor-kappaB ligand/osteoprotegerin cytokine system in primary hyperparathyroidism.

CONTEXT: The mechanisms of action of PTH on bone in vivo remain incompletely understood. The objective of this investigation was to examine changes in serum levels of receptor activator of nuclear factor-kappaB ligand and osteoprotegerin (OPG) in primary hyperparathyroidism and their relationship to bone loss. PATIENTS AND METHODS: Twenty-nine patients with primary hyperparathyroidism had baseline circulating soluble receptor activator of nuclear factor-kappaB ligand (sRANKL) and OPG measured. The relationship to biochemical markers of bone turnover and changes in bone mineral density over 2 yr was examined. RESULTS: Baseline sRANKL levels were elevated (1.7+/-0.1 pmol/liter), whereas OPG remained in the normal range (5.6+/-0.4 pmol/liter). Circulating sRANKL did not correlate with PTH but did correlate with markers of bone resorption (urine deoxypyridinoline cross-links: r=0.51, P<0.01; serum N-telopeptide of type I collagen: r=0.37, P<0.05). Furthermore, sRANKL correlated with both IL-6 and IL-6 soluble receptor (IL-6sR) (r=0.47, P<0.05 and r=0.55, P<0.005, respectively). Serum sRANKL levels also correlated with bone loss at the total femur (r=-0.53, P<0.01). Lastly, a high value of sRANKL in combination with values of IL-6 and IL-6sR in the upper quartile (sRANKL>or=1.81 pg/ml, IL -6>or=11.8 pg/ml, and IL-6sR>or=45.6 ng/ml) defined a group of four women with significantly greater rates of bone loss at the total femur than the remaining patients (-2.7+/-1.7% vs. +0.5+/-0.3%; n=4 vs. n=19, P<0.05). CONCLUSION: Determination of circulating levels of sRANKL may be useful in identifying patients with mild primary hyperparathyroidism at greater risk for bone loss. The fact that circulating sRANKL did not correlate with PTH but did correlate with markers of bone resorption suggests that skeletal responsiveness to PTH may differ in this disease.

Effects of low-dose warfarin and regional chemotherapy on survival in patients with pancreatic carcinoma.

OBJECTIVE: To report the effect of regional combination chemotherapy in a cohort of patients with inoperable pancreatic carcinoma treated with or without low-dose warfarin. MATERIAL AND METHODS: A retrospective analysis was performed on 180 patients with pancreatic carcinoma. Patients received one of seven regimens of chemotherapy. Unrelated to the type of chemotherapy, some patients received 1.25 mg warfarin daily. The primary end-point was median survival. RESULTS: Treatment with warfarin resulted in improved median survival from the start of regional therapy (warfarin versus no warfarin: 5.0 versus 2.3 months, n = 111 versus 69; p < 0.0001). This effect was not dependent on the type of chemotherapy used. Among the seven regimens examined, the one consisting of regional gemcitabine and mitomycin-C with systemic gemcitabine was associated with the longest median survival of 5.1 months from the start of regional therapy (p = 0.006) and 12.7 months from diagnosis. This regimen combined with warfarin was associated with improved median survival (7.1 months, n = 32). CONCLUSIONS: Treatment with low-dose warfarin improved survival irrespective of the chemotherapy received. Of the regimens examined, the combination of regional gemcitabine and mitomycin-C with systemic gemcitabine was associated with the longest survival time. Survival was increased further by the addition of warfarin. These data provide a rationale, based on safety and efficacy, for a definitive study on the use of warfarin and combined regional and systemic chemotherapy in patients with pancreatic carcinoma.

Role of the interleukin-6/interleukin-6 soluble receptor cytokine system in mediating increased skeletal sensitivity to parathyroid hormone in perimenopausal women.

We have observed a strong correlation between circulating levels of both interleukin-6 (IL-6) and interleukin-6 soluble receptor (IL-6sR) and rates of bone turnover in patients with primary hyperparathyroidism. Furthermore, we have found that serum levels of IL-6sR predict rates of bone loss in postmenopausal women with this disease. Estrogen modulates parathyroid hormone (PTH)-induced increases in serum IL-6/IL-6sR, such that, in the estrogen-deficient state, there is an exaggerated release of these cytokines. We therefore propose that the perimenopausal period represents a time when skeletal sensitivity to the resorbing actions of PTH increases because of augmented release of IL-6 and IL-6sR. To test this hypothesis, we retrospectively examined data from 91 women with primary hyperparathyroidism who were seen over the last 5 years at our institution. Women were categorized, based on their age, as premenopausal (n = 20, 41 +/- 2 years), perimenopausal (n = 17, 54 +/- 1 years), or postmenopausal (n = 54, 64 +/- 1 years). Despite having similar mean values for PTH, perimenopausal women had a mean serum IL-6 value that was significantly higher than that in the premenopausal group (13 +/- 2 vs. 8 +/- 2 pg/ml; p = 0.03). This difference in cytokine profile was mirrored by higher mean values for urine N telopeptides of type I collagen (NTX) in the perimenopausal group compared with premenopausal women (114 +/- 9 vs. 80 +/- 11 nM bone collagen equivalents (BCE)/mM creatinine, p = 0.01). Of the three groups of patients, values for IL-6 and urine NTX were highest in the postmenopausal group. We conclude that the perimenopausal period may be a time of increased risk for the skeletal complications of hyperparathyroidism. This is because of increased skeletal sensitivity to the resorbing actions of PTH, mediated in part, by the IL-6/IL-6sR cytokine system.

Circulating levels of interleukin-6 soluble receptor predict rates of bone loss in patients with primary hyperparathyroidism.

It remains unclear whether mild primary hyperparathyroidism results in accelerated bone loss, with recent studies reaching different conclusions. This could be due to intrinsic differences in disease activity not captured by the classical biochemical markers of the disease. Because circulating levels of IL-6 and IL-6 soluble receptor (IL-6sR) are reportedly elevated in patients with hyperparathyroidism, we sought to determine whether measures of this cytokine axis could be helpful in determining the risk for bone loss in hyperparathyroidism. We prospectively followed 23 patients with hyperparathyroidism for 22 +/- 1.5 months and found that baseline circulating levels of IL-6sR correlated significantly with rates of bone loss at the total femur (r = -0.53, P < 0.01). Furthermore, the combination of a serum IL-6sR in the upper tertile (> or=45.6 ng/ml) and IL-6 in the upper half (> or =11.8 pg/ml) of values in the whole group defined a subset of patients with a significantly greater rate of yearly bone loss at the total femur than the remainder of the group (-2.6 +/- 1.3% vs. +0.4 +/- 0.3%, P < 0.05). We conclude that the combined measurements of serum IL-6sR and IL-6 may be helpful in identifying patients with untreated hyperparathyroidism who are more likely to experience bone loss at the total femur.