Section of Oncology
hypercalciemia complicating breast cancer using these agents. High doses failed to reduce the serum calcium or the urinary hydroxyproline: creatinine ratio in ten patients with skeletal metastases, four of whom had hypercalcxmia. Acknowledgments: We are indebted to Professor A M Neville, Professor I Maclntyre, Dr P B Greenberg and Dr M K Ward for their help and advice, and also to Dr B Tulloch for the urinary cyclic AMP measurements and Mrs C J Hillyard for the parathyroid hormone estimations. R C Coombes is in receipt of a Medical Research Council Clinical Research Fellowship. REFERENCES Brereton H D, Halushka P V, Alexander R W et al. (1974) New England Journal of Medicine 291, 83 Berson S A & Yalow R S (1971) American Journal of Medicine 50, 623 Brown B L, Ekins R P & Tampion R S (1970) Biochemical Journal 120, 8 Coombes R C, Hillyard C J, Greenberg P B et al. (1974) Lancet i, 1080 Coombes R C, Neville A M, Bondy P K et al (1976) Prostaglandins 12, 1027 Coombes R C, Ward M K, Greenberg P B et al. (I1976) Cancer, 38, 21 1 1 Galasko C S B & Burns J I (1971) British Medical Journal iii, 573 Gordan G S & Schachter D (1963) Proceedings of the Society of Experimental Biology and Medicine 113, 760 Laszlo D, Schulman C A, Bellin J et al. (1952) Journal of the American Medical Association 148, 1027 Melvin K E W, Hepner G W, Bordier P et al. (1970) Quarterly Journal of Medicine 39, 83 Prockop D J & Udenfriend S (1960) Analytical Biochemistry 1, 228 Powles T J, Clark S A, Easty D M et al. (1973) British Journal of Cancer 28, 316 Powles T J, Leese C & Bondy P K (1975) British Medical Journal ii, 164 Reiner M, Naderajah A, Leese B et al. (1970) Calcified Tissue Research 4 (Suppt), 95 Seyberth H W, Segre G V, Morgan J L et al. (1975) New England Journal of Medicine 293, 1278 Silva 0 L, Chisholm C & Becker K L (1975) Clinical Research 23, 596A Szymendera J (1970) In: Recent Results in Cancer Research, vol 27. Heinemann, London
Dr T J Powles (Royal Marsden Hospital, Sutton, Surrey)
Mechanisms for the Development of Bone Metastases and Hypercalcmmia in Patients with Breast Cancer
Patients with breast cancer frequently develop bone metastases and hypercalcemia. This may be related to the ability of the primary tumour cells to release osteolytically active substances which could
199
facilitate the development of tumour in bone. To test this hypothesis various groups have undertaken experiments in the laboratory with human and experimental tumours, and clinical studies on patients with breast cancer, but unfortunately the results have produced a somewhat confusing picture. This paper is therefore a review of the work in this field, in an attempt to clarify the possible mechanisms involved. In vitro Osteolysis, Prostaglandin Synthesis and Aspirin We have shown that 60 % of human breast tumours are able to cause breakdown of bone in organ culture (Powles, Dowsett, Easty et al. 1976), a property which is related to the ability of the tumour cells to release prostaglandins (PGs) and other nondialysable osteolytic agents (Dowsett, Easty, Powles et al. 1976). PG synthesis and release by the tumour cells can be inhibited by introducing aspirin into the organ-culture system. Production of the nondialysable osteolytic agent(s) is also inhibited, suggesting that PG synthesis may be involved in its release. In vitro osteolysis is, in part, inhibited by aspirin, which affects the tumour, the bone or both. Patients whose tumours possess in vitro osteolytic activity are more likely to have or develop bone metastases and hypercalcmemia (Powles, Dowsett, Easty et al. 1976), suggesting that development of tumour metastases in bone may depend on the ability of tumours to release substances which can break down bone. This concept is supported by the work of Bennett et al. (1975), who have described how some breast tumours, particularly those from patients with bone metastases, are able to synthesize materials with PG-like biological activity in vitro. The nondialysable osteolytic agent(s) has not yet been identified, although it has been shown that it is probably not protein-bound PG or parathyroid hormone (PTH) (Dowsett, Easty, Powles et al. 1976). Osteoclast activating factor reported to be released by myeloma cells (Mundy et al. 1974) has not been excluded, and the possibility that osteolytic enzymes are involved is at present under investigation. However, it seems that osteolysis is probably a multifactorial process, and that production and/or release of the non-PG osteolytic agent(s) is related to PG synthesis by the tumour cells. As Easty warns (p 191), it is not possible to evaluate the relative importance in vivo of these various agents. Calcium Metabolic Studies in Humans with Breast Cancer Patients with hypercalcemia and breast cancer usually have widespread osteolytic bone secondaries (Galasko & Burn 1971) associated with release of calcium from the skeleton in excess of
200
Proc. roy. Soc. Med. Volume 70 March 1977
that which can be cleared by the kidney (Laszlo et al. 1952). This idea is supported by the data of Coombes, who has shown how patients with breast cancer and bone metastases have reduced calcium absorption from the gut (Coombes, Ward, Greenberg et al. 1976). This, coupled with the observation that these patients have increased urine and endogenous fecal calcium excretion, suggests that hypercalcwemia in breast cancer may result from direct breakdown of bone by tumour metastases and not from abnormalities of calcium absorption or excretion. The organic bone matrix is also destroyed by the bone metastases and Cuschieri has shown how increased excretion of hydroxyproline in the urine reflects this process (Cuschieri 1973). If bone metastases from breast cancer depend on PG synthesis for osteolysis, as the in vitro evidence suggests, and if osteolysis by bone metastases is associated with hypercalcemia and hydroxyprolinuria, it might be expected that inhibition of PG synthesis by drugs like aspirin and indomethacin would reduce these abnormalities. However, Coombes has shown that these drugs do not affect hypercalc2emia or hydroxyprolinuria in patients with breast cancer and bone metastases (Coombes, Neville, Bondy et al. 1976). Thus, these in vitro and in vivo results appear at first sight difficult to reconcile with one another.
Hypercalcatmia in Humans with Non-Mammary Tumours in the Absence of Obvious Skeletal Metastases The observation by several groups (Brereton et al. 1974, Ito et al. 1975, Seyberth et al. 1975) that indomethacin will reduce the serum calcium in some patients with hypercalctmia associated with bronchial, pancreatic and renal malignancies is even more difficult to explain. Hypercalcxemia that complicates these tumours is not usually associated with widespread bone metastases and has usually been considered to be a result of 'ectopic' production by the tumour of a circulating osteolytic material. It is possible that these tumours produce PGs which are carried by the circulation to the bones, but although this idea is supported by some animal experiments (Tashjian et al. 1974), it is more probable that circulating PGs are inactivated as they pass through the lungs. It seems unlikely that circulating PGs can cause hypercalcemia in vivo because many tissues (including some benign breast tumours) can synthesize immunoreactive PGs and possess marked osteolytic activity in vitro (Dowsett, Powles, Easty et al. 1976) with no associated hypercalcemia. An alternative hypothesis is that circulating osteolytic agents (a PTH-like substance, for example) released by the tumour may stimulate osteolysis by stimulating PG synthesis in bone.
This is supported by the observation described by Easty (Dowsett, Eastman, Easty et al. 1976) that collagenase, an osteolytic enzyme, can stimulate PG synthesis and osteolysis in bone in organ culture. Furthermore, PTH-stimulated osteolysis can be inhibited in vitro with aspirin (Powles, Easty, Easty & Bondy 1973) and successful treatment of hypercalctmia has been reported in a patient with a parathyroid adenoma (Blum 1975), suggesting that the action of PTH may be mediated in part by PG synthesis. Release of a non-PG osteolytic agent by the tumour may depend on PG synthesis by the tumour cells in a manner similar to that suggested for breast tumour cells (see above). PG metabolites in the urine do not necessarily originate from the tumour; they may come from PG synthesis by bone, kidney, or other tissue.
Hypercalcamia and Bone Metastases in Experimental Animals Many animal tumour models have been used for studies of tumour-associated hypercalcemia. As in the human situation, there appear to be two distinct types: those in which hypercalcemia is associated with spread of tumour to bone, and those in which it occurs in the absence of metastases. Tashjian has chosen the non-metastasizing HSDM mouse tumour to study the role ofcirculating factors and has clearly shown that PG synthesis is involved, and that the hypercalcemia can be corrected with indomethacin medication (Tashjian et al. 1974). In contrast, we have used a model (the Walker tumour in the rat, which gives rise to osteolytic bone metastases and hypercalcemia) in order to study the type of problem which arises in breast cancer. We found that in vitro osteolysis by this tumour could be partly inhibited by aspirinlike drugs. Development of bone metastases and hypercalcemia could be prevented by treatment with these drugs (Powles, Easty, Easty & Neville 1973) provided that it was started before obvious lytic bone deposits were present (unpublished observation). These experiments support the hypothesis that development of bone metastases and hypercalcaemia may depend on release of osteolytic agents by the tumour cells and that this can be inhibited by drugs which inhibit PG synthesis. Galasko and Bennett (1976) have described two phases of osteolysis for development of the VX2 tumour transplanted directly into the femur of the rabbit. The first is associated with active osteoclastic bpne resorption and can be prevented by the administration of indomethacin. The later phase appears to be a direct effect of the tumour cells and indomethacin is unable to reverse this process. The part played by PG synthesis by tumour and bone cells in these two phases is purely speculative at this stage, but it seems likely that it is only part of a
Section of Oncology
multifactorial, multiphasic process in the development of bone metastases. In conclusion I suggest that there are two basic types of hypercalcemia, the first an ectopic humoral type associated with bronchial, renal and the experimental HSDM tumour, in which PG synthesis by the tumour and/or bone may be involved; and the second, a type associated with widespread osteolytic bone deposits which depend on the release of osteolytic substances from the tumour cells. In the early stages, the development of osteolytic deposits is associated with increased osteoclastic activity and may be prevented with PG synthetase inhibitors such as indomethacin. In the later stages, however, the development of established osteolytic deposits is not associated with osteoclastic activity and is not prevented by indomethacin. This suggests that in the later phase, either PG synthesis is not involved or the inhibitory effect of these drugs is exceeded. Patients with hypercalcemia and breast cancer usually have established osteolytic deposits and little evidence of circulating osteolytic factors, whereas patients with other tumours may have little or no bone metastasis and circulating osteolytic factors may be a major factor. In patients with breast cancer, although aspirin may not reverse established osteolysis, it may prevent the establishment and development of early bone metastases.
20
REFERENCES Bennett A, McDonald A M, Simpson J S et al. (1975) Lancet i, 1218 Blum I (1975) Lancet i, 866 Brereton H D, Halushka P V, Alexander R W et al. (1974) New England Journal of Medicine 291, 83 Coombes R C, Neville A M, Bondy P K et al. (1976) Prostaglandins 12, 1027 Coombes R C, Ward M K, Greenberg P B et al. (1976) Cancer, 38, 2111 Cuschieri A (1973) British Journal of Surgery 60, 800 Dowsett M, Eastman A R, Easty D M et al. (1976) Nature (London) 263, 72 Dowsett M, Easty, G C, Powles T J et al. (1976) Prostaglandins 2, 447 Dowsett M, Powles T J, Easty G C et al. (1976) Lancet i, 970 Galasko C S B & Burn J I (1971) British Medical Journal iii, 573 Galasko G & Bennett A (1976) Acta orthopawdica Scandinavica, in press Ito H, Sanada T, Katayama T et al. (1975) New England Journal of Medicine 293, 558 Laszlo D, Schulman C A, Bellin J et al. (1952) Journal of the American Medical Association 148, 1027 Mundy G R, Raisz L G, Cooper R A et al. (1974) New England Journal of Medicine 291, 1041 Powles T J, Dowsett M, Easty G C et al. (1976) Lancet i, 608 Powles T J, Easty G C, Easty D M & Bondy P K (1973) Nature (New Biology) 245, 83 Powles T J, Easty G C, Easty D M & Neville A M (1973) British Journal of Cancer 28, 316 Seyberth H W, Sagre G V, Morgan J L et al. (1975) New England Journal of Medicine 293, 1278 Tashjian A H, Voelkel E F, Goldhaber P et al. (1974) Federation Proceedings 33, 81.
hypercalciemia complicating breast cancer using these agents. High doses failed to reduce the serum calcium or the urinary hydroxyproline: creatinine ratio in ten patients with skeletal metastases, four of whom had hypercalcxmia. Acknowledgments: We are indebted to Professor A M Neville, Professor I Maclntyre, Dr P B Greenberg and Dr M K Ward for their help and advice, and also to Dr B Tulloch for the urinary cyclic AMP measurements and Mrs C J Hillyard for the parathyroid hormone estimations. R C Coombes is in receipt of a Medical Research Council Clinical Research Fellowship. REFERENCES Brereton H D, Halushka P V, Alexander R W et al. (1974) New England Journal of Medicine 291, 83 Berson S A & Yalow R S (1971) American Journal of Medicine 50, 623 Brown B L, Ekins R P & Tampion R S (1970) Biochemical Journal 120, 8 Coombes R C, Hillyard C J, Greenberg P B et al. (1974) Lancet i, 1080 Coombes R C, Neville A M, Bondy P K et al (1976) Prostaglandins 12, 1027 Coombes R C, Ward M K, Greenberg P B et al. (I1976) Cancer, 38, 21 1 1 Galasko C S B & Burns J I (1971) British Medical Journal iii, 573 Gordan G S & Schachter D (1963) Proceedings of the Society of Experimental Biology and Medicine 113, 760 Laszlo D, Schulman C A, Bellin J et al. (1952) Journal of the American Medical Association 148, 1027 Melvin K E W, Hepner G W, Bordier P et al. (1970) Quarterly Journal of Medicine 39, 83 Prockop D J & Udenfriend S (1960) Analytical Biochemistry 1, 228 Powles T J, Clark S A, Easty D M et al. (1973) British Journal of Cancer 28, 316 Powles T J, Leese C & Bondy P K (1975) British Medical Journal ii, 164 Reiner M, Naderajah A, Leese B et al. (1970) Calcified Tissue Research 4 (Suppt), 95 Seyberth H W, Segre G V, Morgan J L et al. (1975) New England Journal of Medicine 293, 1278 Silva 0 L, Chisholm C & Becker K L (1975) Clinical Research 23, 596A Szymendera J (1970) In: Recent Results in Cancer Research, vol 27. Heinemann, London
Dr T J Powles (Royal Marsden Hospital, Sutton, Surrey)
Mechanisms for the Development of Bone Metastases and Hypercalcmmia in Patients with Breast Cancer
Patients with breast cancer frequently develop bone metastases and hypercalcemia. This may be related to the ability of the primary tumour cells to release osteolytically active substances which could
199
facilitate the development of tumour in bone. To test this hypothesis various groups have undertaken experiments in the laboratory with human and experimental tumours, and clinical studies on patients with breast cancer, but unfortunately the results have produced a somewhat confusing picture. This paper is therefore a review of the work in this field, in an attempt to clarify the possible mechanisms involved. In vitro Osteolysis, Prostaglandin Synthesis and Aspirin We have shown that 60 % of human breast tumours are able to cause breakdown of bone in organ culture (Powles, Dowsett, Easty et al. 1976), a property which is related to the ability of the tumour cells to release prostaglandins (PGs) and other nondialysable osteolytic agents (Dowsett, Easty, Powles et al. 1976). PG synthesis and release by the tumour cells can be inhibited by introducing aspirin into the organ-culture system. Production of the nondialysable osteolytic agent(s) is also inhibited, suggesting that PG synthesis may be involved in its release. In vitro osteolysis is, in part, inhibited by aspirin, which affects the tumour, the bone or both. Patients whose tumours possess in vitro osteolytic activity are more likely to have or develop bone metastases and hypercalcmemia (Powles, Dowsett, Easty et al. 1976), suggesting that development of tumour metastases in bone may depend on the ability of tumours to release substances which can break down bone. This concept is supported by the work of Bennett et al. (1975), who have described how some breast tumours, particularly those from patients with bone metastases, are able to synthesize materials with PG-like biological activity in vitro. The nondialysable osteolytic agent(s) has not yet been identified, although it has been shown that it is probably not protein-bound PG or parathyroid hormone (PTH) (Dowsett, Easty, Powles et al. 1976). Osteoclast activating factor reported to be released by myeloma cells (Mundy et al. 1974) has not been excluded, and the possibility that osteolytic enzymes are involved is at present under investigation. However, it seems that osteolysis is probably a multifactorial process, and that production and/or release of the non-PG osteolytic agent(s) is related to PG synthesis by the tumour cells. As Easty warns (p 191), it is not possible to evaluate the relative importance in vivo of these various agents. Calcium Metabolic Studies in Humans with Breast Cancer Patients with hypercalcemia and breast cancer usually have widespread osteolytic bone secondaries (Galasko & Burn 1971) associated with release of calcium from the skeleton in excess of
200
Proc. roy. Soc. Med. Volume 70 March 1977
that which can be cleared by the kidney (Laszlo et al. 1952). This idea is supported by the data of Coombes, who has shown how patients with breast cancer and bone metastases have reduced calcium absorption from the gut (Coombes, Ward, Greenberg et al. 1976). This, coupled with the observation that these patients have increased urine and endogenous fecal calcium excretion, suggests that hypercalcwemia in breast cancer may result from direct breakdown of bone by tumour metastases and not from abnormalities of calcium absorption or excretion. The organic bone matrix is also destroyed by the bone metastases and Cuschieri has shown how increased excretion of hydroxyproline in the urine reflects this process (Cuschieri 1973). If bone metastases from breast cancer depend on PG synthesis for osteolysis, as the in vitro evidence suggests, and if osteolysis by bone metastases is associated with hypercalcemia and hydroxyprolinuria, it might be expected that inhibition of PG synthesis by drugs like aspirin and indomethacin would reduce these abnormalities. However, Coombes has shown that these drugs do not affect hypercalc2emia or hydroxyprolinuria in patients with breast cancer and bone metastases (Coombes, Neville, Bondy et al. 1976). Thus, these in vitro and in vivo results appear at first sight difficult to reconcile with one another.
Hypercalcatmia in Humans with Non-Mammary Tumours in the Absence of Obvious Skeletal Metastases The observation by several groups (Brereton et al. 1974, Ito et al. 1975, Seyberth et al. 1975) that indomethacin will reduce the serum calcium in some patients with hypercalctmia associated with bronchial, pancreatic and renal malignancies is even more difficult to explain. Hypercalcxemia that complicates these tumours is not usually associated with widespread bone metastases and has usually been considered to be a result of 'ectopic' production by the tumour of a circulating osteolytic material. It is possible that these tumours produce PGs which are carried by the circulation to the bones, but although this idea is supported by some animal experiments (Tashjian et al. 1974), it is more probable that circulating PGs are inactivated as they pass through the lungs. It seems unlikely that circulating PGs can cause hypercalcemia in vivo because many tissues (including some benign breast tumours) can synthesize immunoreactive PGs and possess marked osteolytic activity in vitro (Dowsett, Powles, Easty et al. 1976) with no associated hypercalcemia. An alternative hypothesis is that circulating osteolytic agents (a PTH-like substance, for example) released by the tumour may stimulate osteolysis by stimulating PG synthesis in bone.
This is supported by the observation described by Easty (Dowsett, Eastman, Easty et al. 1976) that collagenase, an osteolytic enzyme, can stimulate PG synthesis and osteolysis in bone in organ culture. Furthermore, PTH-stimulated osteolysis can be inhibited in vitro with aspirin (Powles, Easty, Easty & Bondy 1973) and successful treatment of hypercalctmia has been reported in a patient with a parathyroid adenoma (Blum 1975), suggesting that the action of PTH may be mediated in part by PG synthesis. Release of a non-PG osteolytic agent by the tumour may depend on PG synthesis by the tumour cells in a manner similar to that suggested for breast tumour cells (see above). PG metabolites in the urine do not necessarily originate from the tumour; they may come from PG synthesis by bone, kidney, or other tissue.
Hypercalcamia and Bone Metastases in Experimental Animals Many animal tumour models have been used for studies of tumour-associated hypercalcemia. As in the human situation, there appear to be two distinct types: those in which hypercalcemia is associated with spread of tumour to bone, and those in which it occurs in the absence of metastases. Tashjian has chosen the non-metastasizing HSDM mouse tumour to study the role ofcirculating factors and has clearly shown that PG synthesis is involved, and that the hypercalcemia can be corrected with indomethacin medication (Tashjian et al. 1974). In contrast, we have used a model (the Walker tumour in the rat, which gives rise to osteolytic bone metastases and hypercalcemia) in order to study the type of problem which arises in breast cancer. We found that in vitro osteolysis by this tumour could be partly inhibited by aspirinlike drugs. Development of bone metastases and hypercalcemia could be prevented by treatment with these drugs (Powles, Easty, Easty & Neville 1973) provided that it was started before obvious lytic bone deposits were present (unpublished observation). These experiments support the hypothesis that development of bone metastases and hypercalcaemia may depend on release of osteolytic agents by the tumour cells and that this can be inhibited by drugs which inhibit PG synthesis. Galasko and Bennett (1976) have described two phases of osteolysis for development of the VX2 tumour transplanted directly into the femur of the rabbit. The first is associated with active osteoclastic bpne resorption and can be prevented by the administration of indomethacin. The later phase appears to be a direct effect of the tumour cells and indomethacin is unable to reverse this process. The part played by PG synthesis by tumour and bone cells in these two phases is purely speculative at this stage, but it seems likely that it is only part of a
Section of Oncology
multifactorial, multiphasic process in the development of bone metastases. In conclusion I suggest that there are two basic types of hypercalcemia, the first an ectopic humoral type associated with bronchial, renal and the experimental HSDM tumour, in which PG synthesis by the tumour and/or bone may be involved; and the second, a type associated with widespread osteolytic bone deposits which depend on the release of osteolytic substances from the tumour cells. In the early stages, the development of osteolytic deposits is associated with increased osteoclastic activity and may be prevented with PG synthetase inhibitors such as indomethacin. In the later stages, however, the development of established osteolytic deposits is not associated with osteoclastic activity and is not prevented by indomethacin. This suggests that in the later phase, either PG synthesis is not involved or the inhibitory effect of these drugs is exceeded. Patients with hypercalcemia and breast cancer usually have established osteolytic deposits and little evidence of circulating osteolytic factors, whereas patients with other tumours may have little or no bone metastasis and circulating osteolytic factors may be a major factor. In patients with breast cancer, although aspirin may not reverse established osteolysis, it may prevent the establishment and development of early bone metastases.
20
REFERENCES Bennett A, McDonald A M, Simpson J S et al. (1975) Lancet i, 1218 Blum I (1975) Lancet i, 866 Brereton H D, Halushka P V, Alexander R W et al. (1974) New England Journal of Medicine 291, 83 Coombes R C, Neville A M, Bondy P K et al. (1976) Prostaglandins 12, 1027 Coombes R C, Ward M K, Greenberg P B et al. (1976) Cancer, 38, 2111 Cuschieri A (1973) British Journal of Surgery 60, 800 Dowsett M, Eastman A R, Easty D M et al. (1976) Nature (London) 263, 72 Dowsett M, Easty, G C, Powles T J et al. (1976) Prostaglandins 2, 447 Dowsett M, Powles T J, Easty G C et al. (1976) Lancet i, 970 Galasko C S B & Burn J I (1971) British Medical Journal iii, 573 Galasko G & Bennett A (1976) Acta orthopawdica Scandinavica, in press Ito H, Sanada T, Katayama T et al. (1975) New England Journal of Medicine 293, 558 Laszlo D, Schulman C A, Bellin J et al. (1952) Journal of the American Medical Association 148, 1027 Mundy G R, Raisz L G, Cooper R A et al. (1974) New England Journal of Medicine 291, 1041 Powles T J, Dowsett M, Easty G C et al. (1976) Lancet i, 608 Powles T J, Easty G C, Easty D M & Bondy P K (1973) Nature (New Biology) 245, 83 Powles T J, Easty G C, Easty D M & Neville A M (1973) British Journal of Cancer 28, 316 Seyberth H W, Sagre G V, Morgan J L et al. (1975) New England Journal of Medicine 293, 1278 Tashjian A H, Voelkel E F, Goldhaber P et al. (1974) Federation Proceedings 33, 81.
