Cancer Metastasis Rev DOI 10.1007/s10555-013-9449-1
CLINICAL
Concomitant resistance and early-breast cancer: should we change treatment strategies? Carlos M. Galmarini & Olivier Tredan & Felipe C. Galmarini
# Springer Science+Business Media New York 2013
Abstract The dynamics of disease recurrence shows a bimodal pattern with a fairly broad dominant peak at about 1.5–2 years after surgery followed by a second peak at about 5 years. Nowadays, this clinical pattern is explained by assuming that primary breast tumours as well as their metastases have phases of both arrested (tumour dormancy) and active Gompertzian growth. Tumour dormancy at metastatic sites is currently ascribed to biological particularities of local tissue microenvironments that inhibit the growth of tumour cells. However, in some patients, tumour dormancy appears to also depend on the direct interplay between the primary tumour and those metastases, a biological phenomenon called “concomitant resistance”. Concomitant resistance is related to three biological processes: concomitant immunity, tumourinduced angiogenesis and athrepsia. Concomitant resistance can explain the bimodal relapse pattern of breast cancer patients as well as many other clinical phenomena such as the better clinical outcome among patients surgically treated during the putative early luteal phase, or the worse clinical outcome of African-American premenopausal women. Any therapeutic interventions (even surgery) can affect concomitant C. M. Galmarini : F. C. Galmarini Fundación Marcel Dargent—Escuela Sudamericana de Oncología, Buenos Aires, Argentina F. C. Galmarini e-mail: [email protected] O. Tredan Département de Cancérologie Médicale, Centre de Recherche en Cancérologie de Lyon, Centre Léon Bérard, 69373 Lyon, France e-mail: [email protected] Present Address: C. M. Galmarini (*) Fundación Marcel Dargent—Escuela Sudamericana de Oncología, Montevideo, 1686 Buenos Aires, Argentina e-mail: [email protected]
resistance with the potential to induce a worse as well as a better clinical outcome. This should be taken into account when planning new treatment strategies. Keywords Breast cancer . Metastasis . Tumour dormancy . Concomitant resistance
1 Introduction Until the mid-twentieth century, breast cancer was considered to be a loco-regional disease that could only be treated by surgical resection. Various attempts to increase the extent of primary surgery did not lead to improvements in survival. This experience led to breast-conserving surgery, and several controlled clinical trials demonstrated that excision of the primary tumour followed by whole breast irradiation was as effective as total mastectomy for control of disease in women with early-stage cases [1, 2]. Later, sentinel lymph node mapping was introduced to reduce the extent of axillary dissection, and there is now evidence that this procedure does not improve survival of women even if they have a small number of sentinel nodes involved by tumour [3]. Evidence that the extent of local treatment did not affect survival suggested that breast cancer was often a systemic disease at diagnosis and led to the concept that systemic therapy, given before or after surgery, might eradicate micrometastases and increase the probability of long-term survival. Numerous trials have shown that adjuvant hormonal therapy for estrogen-receptor (ER) positive disease (mostly using tamoxifen or aromatase inhibitors), adjuvant chemotherapy, and adjuvant trastuzumab for HER2-expressing disease can all lead to substantial improvement in survival of women with apparently localised breast cancer [4]. Overall 5- and 10-year relative survival rates for women with breast cancer after state of the art treatment are approximately 85 and 78 %,
Cancer Metastasis Rev
Comparison of untreated patients with those undergoing surgery demonstrated that surgery improved survival. When results of the series of untreated women reported by Bloom were analysed by calculating the risk of death as a function of time after mastectomy, there was a single peak at about the third to fourth year, followed by a near constant plateau (Fig. 1a) [6, 7]. In contrast, the risk of death from 1,173 patients who underwent mastectomy in Milan studies showed a first peak at about the third to fourth year after surgery and a later peak near the eighth year [7]. A double-peaked distribution of mortality was also observed in other published reports that used different surgical procedures with or without systemic adjuvant treatment [8–13]. Thus, there appears to be a single peak in the hazard of death for untreated patients and two peaks for the treated patients, indicating that the survival benefit of any surgery was a delay of death for some patients by around 4–5 years or more. The hazard rate for any relapse after surgery also present a bimodal pattern: an early peak of relapse at about 1.5–2 years after surgery followed by a second peak at almost 5 years and then a tapered tail of relapse extending up to 15 years [7] (Fig. 1b). Patients with large tumours, N+, or premenopausal
status relapsed more frequently in the first peak with almost 27 % of distant relapses of premenopausal N+ patients occurring within the first year [14]. The dynamics of recurrence were quite different depending upon whether the tumour did or did not express the ER [7]. The overall risk of early recurrence was much higher for patients bearing ERnegative tumours than for those bearing ER-positive tumours. The second later recurrence peak was, in contrast, higher for ER-positive and lower for ER-negative tumours [7]. As in the Milan series, others found an early sharp peak of relapse between 1 and 3 years after surgery, and a second broad peak at 5 years that extended over the next 20 years [8, 15–20]. Thus, the bimodal pattern of recurrence emerges from different clinical data sets; although the amplitudes of the peaks vary from study to study, their timing is virtually identical. When related to death risks dynamics, it was seen that the first mortality peak (third to fourth years after surgery) included patients coming from the first recurrence peak or that did not recurred [21]. On the contrary, the women that died at the second peak (eighth years after surgery) included patients who suffered recurrence at the first recurrence peak (mainly loco-regional disease) and all those from the second recurrence peak. An analysis was performed by Demicheli and colleagues for women with early-breast cancer who were given 6 or 12 courses of adjuvant cyclophosphamide, methotrexate, and fluorouracil (CMF) after mastectomy [22, 23]. The risk of recurrence of CMF-treated patients was lower than that of women undergoing only surgery (Fig. 2b) [7]. Most of the difference apparently reflected reduction in early recurrence. Metastases that developed later were presumably derived from tumour cells that were refractory to adjuvant therapy.
Fig. 1 Natural history of untreated and treated breast cancer patients. a Hazard rate for death of breast cancer in untreated patients and patients undergoing mastectomy without or with adjuvant chemotherapy. Solid line Death-specific hazard rate of untreated patients in the Bloom series; dotted line death-specific hazard rates for patients undergoing mastectomy alone in the Milan series; dashed line death-specific hazard rates of patients treated
with surgery followed by adjuvant chemotherapy in the Milan series. b Recurrence hazard rate values in patients undergoing surgery alone or followed by adjuvant chemotherapy. Solid line Recurrence-specific hazard rates for patients undergoing mastectomy alone in the Milan series; dotted line recurrence-specific hazard rate of patients treated with surgery followed by adjuvant chemotherapy in the Milan series. Adapted from [23, 130]
respectively [5]. In an effort to understand how further improvements might be achieved, we review here the natural history of breast cancer and the biological and clinical framework that has been and is currently used to explain its clinical evolution. This will help to develop new treatment strategies more adapted to the clinical reality.
2 Effects of surgery and adjuvant chemotherapy
Cancer Metastasis Rev Fig. 2 Examples of tumour dormancy in breast cancer patients. Examples of micrometastatic foci in bone tissue as detected by haematoxylin/eosin staining (a) and immunohistochemistry for ER+ tumour cells (b). Synchronic presence of breast tumour cells in the primary (c) and in an axillar node (d). Microphotographs showing metachronic tumour cells in a primary (e) as well as in its lung metastases (f); this last were detected by immunohistochemistry for HER2+ cells. All images were obtained under microscope at ×20
In addition, although the recurrence-free survival rates were equal for patients receiving 6 or 12 CMF cycles, the risk peaks occurred at different times. Women receiving six CMF cycles displayed a single symmetrical recurrence peak at 1.4– 1.6 years after surgery, while those receiving 12 CMF cycles had a peak at 2–2.4 years after surgery: Thus, 12 CMF cycles may delay appearance of metastases due to the influence of treatment during the second 6-month period of therapy.
3 The concomitant resistance evidence The double-peaked distribution of mortality and relapse after surgery is currently explained by assuming those primary breast tumours as well as their metastases alternate phases of
continuous Gompertzian-type cellular growth and tumour dormancy [7, 24] (Fig. 2). For example, it is already known that about 30–40 % of primary breast cancer patients have disseminated tumour cells in the bone marrow without any evidence of clinical metastases [25–30]. After completion of adjuvant treatment, about 15–20 % of those patients still have persistent tumour cells in the bone marrow; however, not all patients will develop recurrent disease during extended follow-up [25, 26, 31–33]. Similarly, 59 % of women who had no evidence of clinical disease present circulating tumour cells (CTCs) up to 22 years after mastectomy, indicating that clinically silent tumour foci may exist and continuously shed CTCs into the bloodstream [34, 35]. Tumour dormancy is mostly attributed to specific local tissue microenvironments that should maintain the metastatic
Cancer Metastasis Rev
niche in a cell cycle arrest state. However, experimental data showed that tumour dormancy at metastatic sites appears to also depend on the direct interplay between the primary tumour and those metastases. This phenomenon is called “concomitant resistance” and describes a biological situation in which, upon certain circumstances, a primary tumour exerts a controlling and inhibitory action on the growth of its metastases, while, paradoxically, it continues to grow [36]. Concomitant resistance has received the following explanations (Fig. 3): 1. The growth of a tumour might generate a specific antitumour immune response, which, even though not strong enough to inhibit the primary tumour, is capable of preventing the progression of small secondary tumour implants [37, 38]. At some point, dormant tumour cells in metastatic foci escape the immune system and start to grow. This phenomenon is called “concomitant immunity” and is tumour specific and associated with a typical Tcell-dependent antitumour immune reaction. One possible mechanism of concomitant immunity at micrometastatic foci is active suppression of cytotoxic T cell lymphocytes (CTLs) through specific depletion or secretion of inhibitory cytokines (e.g., interferon gamma (IFN- γ) or interleukin-12 (IL-12), respectively) [39]. Among stromal cells, myeloid-derived suppressor cells (MSDCs), which include immature myeloid cells and macrophages, contribute to CTL inhibition [40, 41]. Thus, tumour dormancy can be established through equilibrium not only between
Fig. 3 Mechanism of concomitant resistance in experimental models. The first mechanism is called concomitant immunity. In mice models, concomitant immunity is mainly observed in small immunogenic tumours (2,000 mm3). Surgical removal of the primary tumour release all these inhibitory effects of concomitant resistance inducing the acceleration of tumour growth in the secondary inoculum
dormant tumour cells and CTLs but also between stromal cells and CTLs through the killing of MSDCs. This equilibrium can be broken by immunosuppressive enzymes. This gradual development of resistance suggests that tumours do not merely survive passively, but that there is a continuous struggle between the host immune response and the dormant tumour cells that can be broken up and thus induce metastatic growth. 2. Primary tumours produce antiproliferative or antiangiogenic molecules such as thrombospondin, angiostatin, or endostatin that suppress vascularisation and growth of small metastases [42, 43]. As described by Folkman and colleagues, in the microenvironment of primary tumours, angiogenic factors are sufficient to overcome the effects of angiogenic inhibitors; however, when inducers and inhibitors are shed from the tumour bed into the circulation, levels of the more labile inducers fall off rapidly, whereas levels of the more stable inhibitors create a systemic antiangiogenic environment that prevents neoangiogenesis in small nests of metastatic cells inducing a dormant state [42]. There is also the potential for primary tumours to release serum factors (e.g., cytokines) that have a direct inhibitory effect on the tumour cells populating subclinical metastases. 3. Another is the “athrepsia theory”, which states that nutrients essential for tumour growth are consumed by the primary tumour, making it difficult for a second implant to grow [36]. The high metabolic rate in the primary tumour induces the release into the bloodstream of
Cancer Metastasis Rev
metabolism byproducts that inhibits the proliferation of tumour cells. Recently, Ruggiero et al. have identified two of these serum factors as being meta-tyrosine and orthotyrosine [36, 44]. The antitumour effects of the tyrosine isomers are apparently mediated by early inhibition of the mitogen-activated protein/extracellular-signal regulated kinase pathway and inactivation of STAT3, potentially driving tumour cells into a state of dormancy. On this basis, a secondary tumour can be inhibited by circulating m - and o -tyrosine at the same time as the primary tumour is protected from their inhibitory effects by the presence of counteracting amino acids (phenylalanine, glutamic acid, aspartic acid, glutamine, and histidine) and thus could continue to grow [43, 45]. Regulation of oxidative stress in tumour microenvironments might also be critical to control concomitant resistance by athrepsia [46, 47]. For instance, it is assumed that the oxidised forms of tyrosine are generated post-translationally when phenylalanine present in proteins is exposed to hydroxyl radicals during oxidative damage released by myeloid-derived suppresor cells present in tumours.
4 Can surgery affect concomitant resistance? Thus, according to the “concomitant resistance” model, in some patients, the primary tumour can exert some kind of homeostatic growth inhibitory effect on micrometastases. By affecting any of the mechanisms of concomitant resistance, surgical removal of the primary tumour in such patients would have the potential to promote and accelerate the growth of micrometastases and, thus, tumour recurrence [48, 49]. For instance, surgical removal of the primary tumour can release the angiogenesis blockade on micrometastatic foci or the proliferation inhibition induced by metabolism byproducts produced by the primary tumour. Surgical removal of the primary tumour would also avoid the competition between the primary tumour and its metastases for essential hostderived nutrients that are required for tumour proliferation. Local treatment not only alter the inhibitory regulation that the primary tumour may exert on micrometastases, but also wound healing and local inflammation induced by surgery might promote the growth of tumour cells, which is further enhanced by immunosuppression induced by surgical stress [50, 51]. Different studies determined that fluid from mastectomy wounds contains high levels of mitogenic and angiogenic factors and low levels of antiangiogenic factors that are capable of stimulating the growth of both endothelial and cancer cells [52–55]. Surgical stress also induces an immunosuppression that favours tumour escape from immunemediated rejection at micrometastatic foci [50, 51]. Although
humoral immunity remains relatively intact, major surgery suppresses cellular immunity for several days [56]. After surgery, there is a measurable decrease in the number of circulating natural killer (NK) cells, cytotoxic T lymphocytes, dendritic cells, and T-helper cells. The production of cytokines that favour cellular-mediated immunity (e.g. IL-2, IL-12 and IFN-γ) is also decreased while an increase in the production of cytokines that interfere with cell-mediated immunity (e.g. IL10) is observed. A peak in immunosuppression is said to occur at day 3, and this may be a window of opportunity during which minimal residual disease can grow and spread [57]. All these phenomena are temporary, diminishing as the wounds heal. Thus, by modifying the homeostatic processes among the host, the primary tumour and its metastases, surgery could favour the rapid resumption of tumour cell growth and the appearance of recurrences at the clinical level. This can explain the first peak of relapse observed in different series of patients after surgical treatment. In summary, the time to transition between dormancy and active metastatic growth may be governed by a variety of local (specific tumour microenvironments) or systemic (concomitant resistance) mechanisms [7]. According to this, the first peak of relapse might be predominantly the result of surgerydriven acceleration of micrometastases growth from dormant states by liberating them from concomitant resistance, while the second peak of relapse may be more reflective of natural stochastic transitions from dormant to active states. Thus, surgical treatment has the potential to induce a worse, as well as a better, clinical outcome depending on whether tumour biology is governed by concomitant resistance or not (Fig. 4).
5 Other clinical evidences Other clinical data also indicate that concomitant resistance phenomena are involved in breast cancer tumour biology. Firstly, and although a controversial subject, several reports have noted lower recurrences and longer disease-free or overall survival among patients surgically treated during the putative early luteal phase (days 14–20) when progesterone levels are elevated [58, 59]. This phenomenon was ascribed to the antiproliferative, antiangiogenic and immunostimulatory effects of circulating progesterone [60–62]. Humoral effects of wounding and/or primary tumour removal should be thus counterbalanced by high progesterone levels found during the early luteal phase, and micrometastases would be more likely to remain dormant or to be eliminated by the immune system. Secondly, and paradoxically, the early detection of breast cancer provided by screening mammography seems to be harmful in women aged
CLINICAL
Concomitant resistance and early-breast cancer: should we change treatment strategies? Carlos M. Galmarini & Olivier Tredan & Felipe C. Galmarini
# Springer Science+Business Media New York 2013
Abstract The dynamics of disease recurrence shows a bimodal pattern with a fairly broad dominant peak at about 1.5–2 years after surgery followed by a second peak at about 5 years. Nowadays, this clinical pattern is explained by assuming that primary breast tumours as well as their metastases have phases of both arrested (tumour dormancy) and active Gompertzian growth. Tumour dormancy at metastatic sites is currently ascribed to biological particularities of local tissue microenvironments that inhibit the growth of tumour cells. However, in some patients, tumour dormancy appears to also depend on the direct interplay between the primary tumour and those metastases, a biological phenomenon called “concomitant resistance”. Concomitant resistance is related to three biological processes: concomitant immunity, tumourinduced angiogenesis and athrepsia. Concomitant resistance can explain the bimodal relapse pattern of breast cancer patients as well as many other clinical phenomena such as the better clinical outcome among patients surgically treated during the putative early luteal phase, or the worse clinical outcome of African-American premenopausal women. Any therapeutic interventions (even surgery) can affect concomitant C. M. Galmarini : F. C. Galmarini Fundación Marcel Dargent—Escuela Sudamericana de Oncología, Buenos Aires, Argentina F. C. Galmarini e-mail: [email protected] O. Tredan Département de Cancérologie Médicale, Centre de Recherche en Cancérologie de Lyon, Centre Léon Bérard, 69373 Lyon, France e-mail: [email protected] Present Address: C. M. Galmarini (*) Fundación Marcel Dargent—Escuela Sudamericana de Oncología, Montevideo, 1686 Buenos Aires, Argentina e-mail: [email protected]
resistance with the potential to induce a worse as well as a better clinical outcome. This should be taken into account when planning new treatment strategies. Keywords Breast cancer . Metastasis . Tumour dormancy . Concomitant resistance
1 Introduction Until the mid-twentieth century, breast cancer was considered to be a loco-regional disease that could only be treated by surgical resection. Various attempts to increase the extent of primary surgery did not lead to improvements in survival. This experience led to breast-conserving surgery, and several controlled clinical trials demonstrated that excision of the primary tumour followed by whole breast irradiation was as effective as total mastectomy for control of disease in women with early-stage cases [1, 2]. Later, sentinel lymph node mapping was introduced to reduce the extent of axillary dissection, and there is now evidence that this procedure does not improve survival of women even if they have a small number of sentinel nodes involved by tumour [3]. Evidence that the extent of local treatment did not affect survival suggested that breast cancer was often a systemic disease at diagnosis and led to the concept that systemic therapy, given before or after surgery, might eradicate micrometastases and increase the probability of long-term survival. Numerous trials have shown that adjuvant hormonal therapy for estrogen-receptor (ER) positive disease (mostly using tamoxifen or aromatase inhibitors), adjuvant chemotherapy, and adjuvant trastuzumab for HER2-expressing disease can all lead to substantial improvement in survival of women with apparently localised breast cancer [4]. Overall 5- and 10-year relative survival rates for women with breast cancer after state of the art treatment are approximately 85 and 78 %,
Cancer Metastasis Rev
Comparison of untreated patients with those undergoing surgery demonstrated that surgery improved survival. When results of the series of untreated women reported by Bloom were analysed by calculating the risk of death as a function of time after mastectomy, there was a single peak at about the third to fourth year, followed by a near constant plateau (Fig. 1a) [6, 7]. In contrast, the risk of death from 1,173 patients who underwent mastectomy in Milan studies showed a first peak at about the third to fourth year after surgery and a later peak near the eighth year [7]. A double-peaked distribution of mortality was also observed in other published reports that used different surgical procedures with or without systemic adjuvant treatment [8–13]. Thus, there appears to be a single peak in the hazard of death for untreated patients and two peaks for the treated patients, indicating that the survival benefit of any surgery was a delay of death for some patients by around 4–5 years or more. The hazard rate for any relapse after surgery also present a bimodal pattern: an early peak of relapse at about 1.5–2 years after surgery followed by a second peak at almost 5 years and then a tapered tail of relapse extending up to 15 years [7] (Fig. 1b). Patients with large tumours, N+, or premenopausal
status relapsed more frequently in the first peak with almost 27 % of distant relapses of premenopausal N+ patients occurring within the first year [14]. The dynamics of recurrence were quite different depending upon whether the tumour did or did not express the ER [7]. The overall risk of early recurrence was much higher for patients bearing ERnegative tumours than for those bearing ER-positive tumours. The second later recurrence peak was, in contrast, higher for ER-positive and lower for ER-negative tumours [7]. As in the Milan series, others found an early sharp peak of relapse between 1 and 3 years after surgery, and a second broad peak at 5 years that extended over the next 20 years [8, 15–20]. Thus, the bimodal pattern of recurrence emerges from different clinical data sets; although the amplitudes of the peaks vary from study to study, their timing is virtually identical. When related to death risks dynamics, it was seen that the first mortality peak (third to fourth years after surgery) included patients coming from the first recurrence peak or that did not recurred [21]. On the contrary, the women that died at the second peak (eighth years after surgery) included patients who suffered recurrence at the first recurrence peak (mainly loco-regional disease) and all those from the second recurrence peak. An analysis was performed by Demicheli and colleagues for women with early-breast cancer who were given 6 or 12 courses of adjuvant cyclophosphamide, methotrexate, and fluorouracil (CMF) after mastectomy [22, 23]. The risk of recurrence of CMF-treated patients was lower than that of women undergoing only surgery (Fig. 2b) [7]. Most of the difference apparently reflected reduction in early recurrence. Metastases that developed later were presumably derived from tumour cells that were refractory to adjuvant therapy.
Fig. 1 Natural history of untreated and treated breast cancer patients. a Hazard rate for death of breast cancer in untreated patients and patients undergoing mastectomy without or with adjuvant chemotherapy. Solid line Death-specific hazard rate of untreated patients in the Bloom series; dotted line death-specific hazard rates for patients undergoing mastectomy alone in the Milan series; dashed line death-specific hazard rates of patients treated
with surgery followed by adjuvant chemotherapy in the Milan series. b Recurrence hazard rate values in patients undergoing surgery alone or followed by adjuvant chemotherapy. Solid line Recurrence-specific hazard rates for patients undergoing mastectomy alone in the Milan series; dotted line recurrence-specific hazard rate of patients treated with surgery followed by adjuvant chemotherapy in the Milan series. Adapted from [23, 130]
respectively [5]. In an effort to understand how further improvements might be achieved, we review here the natural history of breast cancer and the biological and clinical framework that has been and is currently used to explain its clinical evolution. This will help to develop new treatment strategies more adapted to the clinical reality.
2 Effects of surgery and adjuvant chemotherapy
Cancer Metastasis Rev Fig. 2 Examples of tumour dormancy in breast cancer patients. Examples of micrometastatic foci in bone tissue as detected by haematoxylin/eosin staining (a) and immunohistochemistry for ER+ tumour cells (b). Synchronic presence of breast tumour cells in the primary (c) and in an axillar node (d). Microphotographs showing metachronic tumour cells in a primary (e) as well as in its lung metastases (f); this last were detected by immunohistochemistry for HER2+ cells. All images were obtained under microscope at ×20
In addition, although the recurrence-free survival rates were equal for patients receiving 6 or 12 CMF cycles, the risk peaks occurred at different times. Women receiving six CMF cycles displayed a single symmetrical recurrence peak at 1.4– 1.6 years after surgery, while those receiving 12 CMF cycles had a peak at 2–2.4 years after surgery: Thus, 12 CMF cycles may delay appearance of metastases due to the influence of treatment during the second 6-month period of therapy.
3 The concomitant resistance evidence The double-peaked distribution of mortality and relapse after surgery is currently explained by assuming those primary breast tumours as well as their metastases alternate phases of
continuous Gompertzian-type cellular growth and tumour dormancy [7, 24] (Fig. 2). For example, it is already known that about 30–40 % of primary breast cancer patients have disseminated tumour cells in the bone marrow without any evidence of clinical metastases [25–30]. After completion of adjuvant treatment, about 15–20 % of those patients still have persistent tumour cells in the bone marrow; however, not all patients will develop recurrent disease during extended follow-up [25, 26, 31–33]. Similarly, 59 % of women who had no evidence of clinical disease present circulating tumour cells (CTCs) up to 22 years after mastectomy, indicating that clinically silent tumour foci may exist and continuously shed CTCs into the bloodstream [34, 35]. Tumour dormancy is mostly attributed to specific local tissue microenvironments that should maintain the metastatic
Cancer Metastasis Rev
niche in a cell cycle arrest state. However, experimental data showed that tumour dormancy at metastatic sites appears to also depend on the direct interplay between the primary tumour and those metastases. This phenomenon is called “concomitant resistance” and describes a biological situation in which, upon certain circumstances, a primary tumour exerts a controlling and inhibitory action on the growth of its metastases, while, paradoxically, it continues to grow [36]. Concomitant resistance has received the following explanations (Fig. 3): 1. The growth of a tumour might generate a specific antitumour immune response, which, even though not strong enough to inhibit the primary tumour, is capable of preventing the progression of small secondary tumour implants [37, 38]. At some point, dormant tumour cells in metastatic foci escape the immune system and start to grow. This phenomenon is called “concomitant immunity” and is tumour specific and associated with a typical Tcell-dependent antitumour immune reaction. One possible mechanism of concomitant immunity at micrometastatic foci is active suppression of cytotoxic T cell lymphocytes (CTLs) through specific depletion or secretion of inhibitory cytokines (e.g., interferon gamma (IFN- γ) or interleukin-12 (IL-12), respectively) [39]. Among stromal cells, myeloid-derived suppressor cells (MSDCs), which include immature myeloid cells and macrophages, contribute to CTL inhibition [40, 41]. Thus, tumour dormancy can be established through equilibrium not only between
Fig. 3 Mechanism of concomitant resistance in experimental models. The first mechanism is called concomitant immunity. In mice models, concomitant immunity is mainly observed in small immunogenic tumours (2,000 mm3). Surgical removal of the primary tumour release all these inhibitory effects of concomitant resistance inducing the acceleration of tumour growth in the secondary inoculum
dormant tumour cells and CTLs but also between stromal cells and CTLs through the killing of MSDCs. This equilibrium can be broken by immunosuppressive enzymes. This gradual development of resistance suggests that tumours do not merely survive passively, but that there is a continuous struggle between the host immune response and the dormant tumour cells that can be broken up and thus induce metastatic growth. 2. Primary tumours produce antiproliferative or antiangiogenic molecules such as thrombospondin, angiostatin, or endostatin that suppress vascularisation and growth of small metastases [42, 43]. As described by Folkman and colleagues, in the microenvironment of primary tumours, angiogenic factors are sufficient to overcome the effects of angiogenic inhibitors; however, when inducers and inhibitors are shed from the tumour bed into the circulation, levels of the more labile inducers fall off rapidly, whereas levels of the more stable inhibitors create a systemic antiangiogenic environment that prevents neoangiogenesis in small nests of metastatic cells inducing a dormant state [42]. There is also the potential for primary tumours to release serum factors (e.g., cytokines) that have a direct inhibitory effect on the tumour cells populating subclinical metastases. 3. Another is the “athrepsia theory”, which states that nutrients essential for tumour growth are consumed by the primary tumour, making it difficult for a second implant to grow [36]. The high metabolic rate in the primary tumour induces the release into the bloodstream of
Cancer Metastasis Rev
metabolism byproducts that inhibits the proliferation of tumour cells. Recently, Ruggiero et al. have identified two of these serum factors as being meta-tyrosine and orthotyrosine [36, 44]. The antitumour effects of the tyrosine isomers are apparently mediated by early inhibition of the mitogen-activated protein/extracellular-signal regulated kinase pathway and inactivation of STAT3, potentially driving tumour cells into a state of dormancy. On this basis, a secondary tumour can be inhibited by circulating m - and o -tyrosine at the same time as the primary tumour is protected from their inhibitory effects by the presence of counteracting amino acids (phenylalanine, glutamic acid, aspartic acid, glutamine, and histidine) and thus could continue to grow [43, 45]. Regulation of oxidative stress in tumour microenvironments might also be critical to control concomitant resistance by athrepsia [46, 47]. For instance, it is assumed that the oxidised forms of tyrosine are generated post-translationally when phenylalanine present in proteins is exposed to hydroxyl radicals during oxidative damage released by myeloid-derived suppresor cells present in tumours.
4 Can surgery affect concomitant resistance? Thus, according to the “concomitant resistance” model, in some patients, the primary tumour can exert some kind of homeostatic growth inhibitory effect on micrometastases. By affecting any of the mechanisms of concomitant resistance, surgical removal of the primary tumour in such patients would have the potential to promote and accelerate the growth of micrometastases and, thus, tumour recurrence [48, 49]. For instance, surgical removal of the primary tumour can release the angiogenesis blockade on micrometastatic foci or the proliferation inhibition induced by metabolism byproducts produced by the primary tumour. Surgical removal of the primary tumour would also avoid the competition between the primary tumour and its metastases for essential hostderived nutrients that are required for tumour proliferation. Local treatment not only alter the inhibitory regulation that the primary tumour may exert on micrometastases, but also wound healing and local inflammation induced by surgery might promote the growth of tumour cells, which is further enhanced by immunosuppression induced by surgical stress [50, 51]. Different studies determined that fluid from mastectomy wounds contains high levels of mitogenic and angiogenic factors and low levels of antiangiogenic factors that are capable of stimulating the growth of both endothelial and cancer cells [52–55]. Surgical stress also induces an immunosuppression that favours tumour escape from immunemediated rejection at micrometastatic foci [50, 51]. Although
humoral immunity remains relatively intact, major surgery suppresses cellular immunity for several days [56]. After surgery, there is a measurable decrease in the number of circulating natural killer (NK) cells, cytotoxic T lymphocytes, dendritic cells, and T-helper cells. The production of cytokines that favour cellular-mediated immunity (e.g. IL-2, IL-12 and IFN-γ) is also decreased while an increase in the production of cytokines that interfere with cell-mediated immunity (e.g. IL10) is observed. A peak in immunosuppression is said to occur at day 3, and this may be a window of opportunity during which minimal residual disease can grow and spread [57]. All these phenomena are temporary, diminishing as the wounds heal. Thus, by modifying the homeostatic processes among the host, the primary tumour and its metastases, surgery could favour the rapid resumption of tumour cell growth and the appearance of recurrences at the clinical level. This can explain the first peak of relapse observed in different series of patients after surgical treatment. In summary, the time to transition between dormancy and active metastatic growth may be governed by a variety of local (specific tumour microenvironments) or systemic (concomitant resistance) mechanisms [7]. According to this, the first peak of relapse might be predominantly the result of surgerydriven acceleration of micrometastases growth from dormant states by liberating them from concomitant resistance, while the second peak of relapse may be more reflective of natural stochastic transitions from dormant to active states. Thus, surgical treatment has the potential to induce a worse, as well as a better, clinical outcome depending on whether tumour biology is governed by concomitant resistance or not (Fig. 4).
5 Other clinical evidences Other clinical data also indicate that concomitant resistance phenomena are involved in breast cancer tumour biology. Firstly, and although a controversial subject, several reports have noted lower recurrences and longer disease-free or overall survival among patients surgically treated during the putative early luteal phase (days 14–20) when progesterone levels are elevated [58, 59]. This phenomenon was ascribed to the antiproliferative, antiangiogenic and immunostimulatory effects of circulating progesterone [60–62]. Humoral effects of wounding and/or primary tumour removal should be thus counterbalanced by high progesterone levels found during the early luteal phase, and micrometastases would be more likely to remain dormant or to be eliminated by the immune system. Secondly, and paradoxically, the early detection of breast cancer provided by screening mammography seems to be harmful in women aged
