Tumors spontaneously develop central necroses due to inadequate blood supply. Recent data indicate that dead cells and their products are immunogenic to the host. We hypothesized that macrophage tumor-dependent reactions can be mediated differentially by factors released from live or dead tumor cells. In this study, functional activity of resident peritoneal macrophages was investigated in parallel with tumor morphology during the growth of syngeneic nonimmunogenic hepatoma 22a. Morphometrical analysis of tumor necroses, mitoses and leukocyte infiltration was performed in histological sections. We found that inflammatory potential of peritoneal macrophages in tumor-bearing mice significantly varied depending on the stage of tumor growth and exhibited two peaks of activation as assessed by nitroxide and superoxide anion production, 5′-nucleotidase activity and pinocytosis. Increased inflammatory reactions were not followed by the enhancement of angiogenic potential as assessed by Vascular Endothelial Growth Factor mRNA expression. Phases of macrophage activity corresponded to the stages of tumor growth characterized by high proliferative potential. The appearance and further development of necrotic tissue inside the tumor did not coincide with changes in macrophage behavior and therefore indirectly indicated that activation of macrophages was a reaction mostly to the signals produced by live tumor cells.
It is now abundantly evident that innate immune response plays important role in antitumor defense [
Macrophages represent blood-borne-derived descendants of mononuclear cells which migrate from the circulation into tissues and display a high degree of plasticity, which is tuned by the tissue microenvironments where they reside. It was shown that macrophages play a dual role in tumor growth and can possess antitumor as well as protumor activities. Therefore they were subdivided into M1 (classically activated) or M2 (alternatively activated) phenotypes [
In order to be fully activated macrophages need to be exposed to two signals: IFN-
On the contrary, M2 cell phenotype is induced in response to Th2 cytokines IL-4, IL-13, and IL-10, as well as by apoptotic cells and immune complexes. Several data indicate that under the influence of tumor environment macrophages exhibit predominantly M2 phenotype which is characterized by increased production of arginase-1 and reduced NO production, enhanced release of anti-inflammatory cytokines (IL-10), ability to support angiogenesis, and tissue remodeling [
However, several other data have indicated that tumor-associated macrophages express both M1- and M2-like characteristics thus demonstrating a “mixed” phenotype [
These data indicate that M1/M2 classification needs to be specified. Moreover, it is now well accepted that there exist at least four different types of adaptive immune responses, namely, Th1, Th2, Th17, and T-regulatory cells type; and innate immune reactions provide a special microenvironment for each of them [
The contribution of macrophages from extratumor sites to the host immunity has been investigated in a very few papers and much less is known on their proangiogenic functions. Several tumor-derived factors are known to stimulate production of angiogenic factors by macrophages. For example, TNF-
Finally, there is now a growing interest to the data indicating that dead cells and their products are immunogenic to the host and can activate innate immune response [
In this study, resident peritoneal macrophages were investigated in parallel with tumor morphology during the growth of syngeneic nonimmunogenic hepatoma 22a. Several parameters attributed to inflammatory and angiogenic macrophage potential were studied: the production of cytotoxic factors such as NO and ROI [
We found that inflammatory potential of peritoneal macrophages in tumor-bearing mice significantly varied depending on the stage of tumor growth and exhibited two peaks of activation as assessed by abovementioned parameters. These phases corresponded to increased activity of the tumor growth characterized by high proliferative potential. The appearance and further development of necrotic tissue inside the tumor did not coincide with changes in macrophage behaviour and therefore indirectly indicated that activation of macrophages was a reaction mostly to the signals produced by live tumor cells.
C3HA male mice weighing 18–20 g, three months old, were purchased from “Rappolovo” Animal Farm, Russian Academy of Medical Sciences, St. Petersburg, Russia. All experiments were performed using protocols approved by the Russian Animal Ethics Committee and followed institutional animal use and care guidelines.
The cell line of hepatoma 22a (MH-22a) was obtained from tissue culture collection of Institute of Cytology, Russian Academy of Sciences, St. Petersburg, and was passaged in vitro in DMEM supplemented with 10% fetal calf serum (Sigma). Cells originated from the solid nonmetastasizing murine tumor induced by 3-methylcholantrene in C3HA mice [
In our studies hepatoma 22a was characterized as a nonimmunogenic tumor by transplantation test. Ten days after i.p. immunization with 107 lethally irradiated (100 Gy) cells, mice were s.c. inoculated with 103–106 live hepatoma cells, five animals per group. During 4 weeks the dynamics of tumor growth in previously immunized animals was compared to control mice. Tumor volume was determined by using the formula
Results of the transplantation test.
Vaccinations | Challenge with live hepatoma cells | |||
103 | 104 | 105 | 106 | |
Cotrol (no vaccination): | ||||
(i) Number of tumor takes | 1/5 | 2/5 | 5/5 | 5/5 |
(ii) Mean tumor volume, cm3 | 0.01 | 0.29 ± 0.12 | 2.21 ± 1.20 | 7.66 ± 0.94 |
107 irradiated hepatoma cells i.p.: | ||||
(i) Number of tumor takes | 1/5 | 3/5 | 5/5 | 5/5 |
(ii) Mean tumor volume, cm3 | 0.03 | 0.04 ± 0.01 | 3.10 ± 1.4 | 9.40 ± 2.60 |
Number of tumor takes (numerator), number of animals per group (denominator).
Tumor volume was measured on 28 day of tumor growth (M ± SEM). Control mice received 0,5 mL of PBS i.p.
In all further experiments we used only live hepatoma cells for the injections. For the development of solid tumors 105 live hepatoma cells in 0.5 mL PBS were inoculated s.c. on the back. Control mice received injection of PBS only. Mice were sacrificed by cervical dislocation on 3, 7, 14, 21, 28, and 35 day after starting experiment.
Resident peritoneal macrophages were obtained by peritoneal lavage with injecting 5 mL of PBS and pooled from 5 mice in each group. Peritoneal cells were left for 2 hours to adhere to the bottom of 96-well plates and nonadherent cells were discarded. Macrophages were cultured in RPMI 1640 media supplemented with 10% FCS and kept in incubator at 37°C, 5% CO2 during 24 hours. Nitrite production was measured as spontaneous or induced by adding 10 ng/mL LPS (E. Coli B5-055, Sigma) [
It was determined in Giemsa-stained smears. Blood was collected by retroorbital puncture and the number of nucleated cells was counted with the help of hemocytometer.
They were isolated and weighed and number of nucleated cells per organ counted.
The level of VEGF, VEGFR-1, VEGFR-2, VEGFR-3, and TSP-1 mRNA expression was studied by semiquantitative reverse transcription followed by polymerase chain reaction (RT-PCR). Total RNA was isolated from macrophages by a single-step method using TRI reagent (Sigma) from each mouse individually. In some experiments thymocytes were also studied. Synthesis of cDNA templates was carried out using 2
At different time points when mice were sacrified they were bled to collect peripheral blood. Then, sera were obtained by centrifugation and stored at −20°C until being analyzed. Serum levels of TNF-
Excised tumor nodules were fixed in 10% formalin and embedded in paraffin. Microscopic sections (5
Data were analyzed using Student’s
To evaluate the inflammatory phenotype of peritoneal macrophages we studied six different parameters: spontaneous and LPS-induced NO2-production, spontaneous and TPA-induced superoxide anion production (NBT-test), pinocytosis, and 5′-N activity.
NO2-production fluctuated up and down with two points of decrease (3 and 14 days) and two points of increase (7 and 28 days). The difference between stimulated and spontaneous production on 7 day of tumor growth was 3 times more than in control mice, and on day 28 was about 1.5 times more than in non-tumor bearers (Figure
Dynamics of functional activity of peritoneal macrophages from mice bearing hepatoma 22a.
The most specific changes were found in the activity of membrane enzyme 5′-N in macrophages. This was the only one characteristic that altered at all time-points of the study. This parameter showed significant fluctuations and was downregulated at 3 and 14 days and upregulated at 7, 21, 28 and 35 days (Figure
We next examined how these inflammatory changes corresponded to macrophage mRNA synthesis of angiogenic or anti-angiogenic factors.
We expected that inflammatory changes of macrophage activity would be accompanied by certain alterations in their angiogenic properties. Therefore we studied mRNA expression of main angiogenic factor—VEGF and its receptors, as well as anti-angiogenic factor—TSP-1. Unexpectedly, we did not found any difference in VEGF mRNA expression in peritoneal macrophages of tumor-bearing mice at any time of the study (3, 7, 14, 21, 28, or 35 days) as compared to control animals (data not shown). There was slight enhancement of VEGFR-1 mRNA expression at 7 and 35 days of tumor growth and VEGFR-2 mRNA increase at 35 day (Figure
Hepatoma growth increases VEGFR-1 (a) and VEGFR-2 mRNA expression (b) in peritoneal macrophages. Total RNA was isolated from peritoneal cells from each mouse individually. Two
Opposite to VEGF, TSP-1 mRNA expression in macrophages was shown to increase at the late stages of tumor growth (Figures
Hepatoma growth increases TSP-1 mRNA expression in peritoneal macrophages (a) and thymocytes (b). Total RNA was isolated from peritoneal cells or thymocytes from each mouse individually. Two
Tumors remained invisible several days after inoculation. At 7 days tumors represented a small palpable nodules about 2-3 mm in diameter. The tumor grew fast within 3 weeks and reached a steady state after 28 day (Figure
Morphological and histological parameters of hepatoma 22a growth (a–c). (a) Tumor volume after s.c. inoculation of 105 live hepatoma cells was measured on days of sacrifice; bars represent areas of viable (empty parts of the bar chart) and necrotized (dashed parts of the bar chart) tumor tissue in percent to the whole tumor volume. Areas of viable and necrotic tissues were calculated in histological sections per field of vision, magnification ×200; ten fields of vision were counted for each mice (
Histological study revealed that 3 days after inoculation tumor cells were located separately, isolated from each other, and were polymorphous, without mitoses or perifocal infiltration (Figures
Microphotographs of hepatoma 22a (a) Tumor at 3 day after inoculation ×200 (b) higher magnification ×600: Polymorphous, separately lying tumor cells, no reaction from peripheral tissue. Tumor tissue is shown by arrows. Hematoxylin-eosin staining (c) tumor at 7 day after inoculation ×200 (d) higher magnification ×600. The appearance of new vessels in the tumor tissue shown by arrows. (Azure-II-eosin staining).
At the 7th day, the tumor tissue was more compact, organized, and started to be vascularized from periphery (Figures
Microphotographs of hepatoma 22a. (a) Tumor at 7 day after tumor inoculation ×900. Mononuclear infiltration at the periphery of tumor nodule is shown by arrows. Azure-II-eosin staining. (b) Tumor at 14 day after inoculation ×600. The appearance of first central necrosis is shown by arrow. Azure-II-eosin staining. (c) Tumor at 21 day after inoculation ×200. Necrotic tissue (shown by arrows). Azure-II-eosin staining. (d) Tumor at 28 day after inoculation ×900; Mitosis in the remaining viable tumor tissue. Hematoxylin-eosin staining.
At 14 days tumor diameter was about 10 mm. Examination of sections showed that tumor cells became more homogeneous, moderately differentiated, and typical for trabecular hepatoma. Mononulear cells at the periphery of the tumor nodule were almost absent. In the central parts adjacent to necrotic areas there were visible foci of white blood cell infiltration (Figure
Rate of tumor growth per week.
Part of tumor tissue | Increase in the tumor tissue per week, cm3 | |||
7 day–14 day | 14 day–21 day | 21 day–28 day | 28 day–35 day | |
Necrotic | 0.01 ± 0.005 | 0.10 ± 0.02 | 1.20 ± 0.30 | 1.59 ± 0.35 |
Viable | 0.29 ± 0.06 | 0.65 ± 0.11 | 0.73 ± 0.15 | −0.34 ± 0.10 |
Tumor volume (M ± SEM) within the previous week (e.g., day 7) was subtracted from the tumor volume within the next week (e.g., day 14). Volume of viable tumor tissue within the last week decreased.
At the late stages of tumor growth the necrotized areas were almost homogenous with no live tumor or infiltrating cells evident (Figure
The number of nucleated cells in the peritoneum was not significantly changed during the time of observation (Figure
The influence of tumor growth on blood leukocytes, cytokine serum level, spleen, and thymus. (a)—blood leukocyte (upper curve) and peritoneal cell (lower curve) counts during the growth of hepatoma 22a. (b)—absolute cell numbers of leukocytes in the blood of tumor-bearing mice: polymorphonucleus leukocytes (1), lymphocytes (2), and monocytes (3). (c)—percent of total number of leukocytes: polymorphonucleus leukocytes (1), lymphocytes (2), and monocytes (3). Each point represents average values of 10 mice (M ± SEM),
Nevertheless, the growth of hepatoma influenced the amount and composition of peripheral white blood cells. Total leukocyte count was increased on day 7 and 28 (Figure
Control male mice of C3HA strain used in our experiments possessed about 45.0 ± 1.53% lymphocytes, 38.7 ± 1.6% polymorphonuclear cells, and 16.3 ± 0.9% monocytes of total number of peripheral blood leukocytes that differed from the Jackson laboratory Phenome database for male C3H/HeJ mice that have 68% lymphocytes, 32% polymorphonuclear cells (including 24% neutrophils, 4% eosinophils and 4% basophils), and 2% monocytes.
Since we observed significant changes in blood leukocyte counts, we were also interested to find some quantitative changes in spleen and thymus. Starting from the 21 day of hepatoma growth thymus weight was dramatically decreased which was also followed by significant reduction of the total number of thymocytes per organ (Figure
Serum levels of proinflammatory cytokine TNF-
Neither TNF-
During the past decade, there appeared significant amount of new data indicating that dying cells produce variety of signals inducing innate immune responses [
Spontaneous development of necrosis is an inherent property of solid tumors and is remarkably evident especially in fast-growing animal neoplasms. This phenomenon occurs due to inadequate blood supply, coincides with oxygen depletion, nutrient, and energy deprivation, and was studied in details [
The notion that necrotic tumor cells could provide signals to enhance the growth of remaining viable ones has been considered over 50 years [
While producing many chemokines, live tumor cells have a strategy to recruit inflammatory cells from the circulation in order to have additional resource of growth factors and angiogenic activity [
Histological study of nonimmunogenic solid hepatoma 22a transplants revealed the appearance of first necrosis on day 14, and later on, this process developed progressively up to 35 day (the end of experiment). We may suppose that factors released from dead cells will be increased constantly in the circulation and will influence immune cells in a monotone manner. Unlike this, the behaviour of viable tumor tissue had a discrete character with periods of fast and slow growth. Histological examination discovered two dramatic time-points for the growth of murine hepatoma: the 7th and 28th days.
Day 7 of the tumor growth was characterized by the appearance of first small blood vessels and mononuclear infiltration around the tumor. It is assumed that as soon as tumor nodule reaches the size beyond 2-3 mm, its development entirely depends on formation of a unique tumor vasculature to meet increased metabolic demands of fast growing malignant tissue [
The 4th week of hepatoma growth was characterized by more moderate rate of growth but still at 28 day the volume of viable tumor tissue and rate of the tumor growth within this week were the largest during experiment. On day 28 we also observed maximal numbers of mitosis in live tumor tissue. Therefore, we suggested that both 7 and 28 days may be characterized by high angiogenic demands of the tumor. Later on, day 35 showed decreased mitotic activity, loss of viable tumor tissue, and dramatic increase in necrosis.
While summarizing all tumor-dependent effects, immunological changes during the growth of hepatoma fell into two categories: those that went in parallel steadily to the development of necrosis and others that differed significantly and showed pulsed dynamics. To the first type belonged thymic involution, lymphopenia, splenomegaly, and leukemoid reaction; to the second-changes in peripheral macrophage functional activity.
The existence of a common mechanism for the development of the first group of changes is not proved. However, the coincidence of these alterations was first mentioned during the study of several mammary adenocarcinomas in mice [
Changes in the activity of peritoneal macrophages did not resemble the development of any of the abovementioned processes. In contrast to evident quantitative changes in the numbers of thymocytes and splenocytes as well as in blood granulocytes and lymphocytes, there was not any numerical difference in the number of peritoneal macrophages as well as in the number of blood monocytes during the whole experiment. Opposite to these we observed certain modifications of macrophage inflammatory phenotype every week. Several features, including NO2-production and 5′-N activity, changed up and down, that may possibly reflect the prevalence of different types of factors. One specific phenotype was repeated on 7 and 28 days of tumor growth and represented two peaks of macrophage activation when all inflammatory parameters were increased (Table
Changes in macrophage activity during the growth of hepatoma 22a.
↑ increased parameter,
Several changes on 7 and 28 days, like elevation of nitroxide as well as superoxide production (assessed by NBT-test), are known to be a feature of classically activated macrophages or M1 phenotype [
Similar changes of ecto-enzyme activity in macrophages were also shown in mice bearing lymphomas [
Decrease in 5′-N activity in macrophages was also observed in our experiments on 3 and 14 days of hepatoma growth. But other parameters investigated at this period of time did not allow to attribute these activities to M1 type. Downregulation of NO2 production (3 and 14 days) and increased pinocytosis (14 day) are considered to characterize alternatively activated or M2 macrophages [
Independently of their interpretation these data allow to suggest that changes of 5′-N activity in tissue macrophages are a specific feature of tumor growth. CD73 (5′-N) is a GPI-linked cell-surface enzyme of purine metabolitic pathway which dephosphorylates AMP to adenosine [
Taken together alterations in macrophage activity on days 3, 7, 14, and 28 could be characterized by “mixed” phenotype. The day 21 was characterized by almost no changes in macrophage functional activity with the exception for increased 5′-N activity.
At the late stage of tumor growth (day 35) we did not observe any signs of inflammatory response in macrophages. On the contrary, there was a decrease in superoxide anion production that can be considered as a sign of exhaustion after prolonged activation. Nevertheless, we cannot assume these changes as depression because they were followed by increase in VEGFR-1, VEGFR-2, and TSP-1 mRNA expression. These activities may be regarded as functions related to chemotaxis via VEGF receptors [
Additionally we did not found any increase in the mRNA synthesis of proangiogenic cytokine VEGF in macrophages. Apart from this, we found, for the first time, increased expression of TSP-1 mRNA in peritoneal macrophages and thymocytes, the factor known to have antiangiogenic properties [
To conclude we found that the s.c. growth of nonimmunogenic hepatoma in mice induced in peripheral macrophages diverse forms of inflammatory responses. In available literature we found contradictory results concerning activity of macrophages isolated from sites distal to tumor. Different tumor models revealed specific changes in macrophage activity. In particular, in case of syngeneic fibrosarcoma or AK-5 tumor, NO-production by peritoneal macrophages was increased [
The most intriguing question was as follows: how did tumor cells activate distant macrophages? We suggested that at 7 and 28 days the macrophages could be activated by factors produced by live tumor cells and associated with angiogenic potential of the tumor cells. To elucidate this situation we examined the level of several cytokines in the blood of hepatoma-bearing mice throughout the time of experiment. For this study we selected angiogenic factor VEGF, chemotactic for blood monocytes [
Unfortunately, we did not find elevation of TNF-
We can also suggest that macrophage activity may be modulated by the molecules released from dying tumor. In particular, abundant tumor cell-derived substances like hyaluronan fragments together with heat shock proteins are endogenous ligands to TLR2 and TLR4 and trigger M2-like cytokine profile in macrophages [
An important role in modulation of macrophage behaviour may also play purine metabolites. When any cell dies and loses membrane integrity, nucleotides and metabolites, such as ATP, adenosine, and uric acid, will be released to potentially alert the immune system, not only activating but also suppressing mononuclear phagocytes. Thus, ATP released from dying cells is known to activate macrophages and potentiate their superoxide generation [
The elevation of 5′-N activity in macrophages, observed in our experiments, presumably, may lead to increased concentration of extracellular adenosine and serve as an autoregulatory mechanism. On one hand, enhanced concentration of adenosine may dampen inflammatory reactions; on the other hand, intracellular adenosine after entering the cell becomes an important source of superoxide radical production via desamination process in activated macrophages [
In summary, our studies demonstrate that nonimmunogenic syngeneic tumor affects functional activity of distant macrophages, this influence being mostly of stimulatory character. We propose that peritoneal macrophages can receive two types of immunomodulatory signals produced by either live or dead tumor cells. Peritoneal macrophages do not directly participate in the dramatic events occurring at the tumor site. According to the modern knowledge these cells can only play a role of bystanders but their type of reaction reflects systemic immune response to tumor growth as a whole.
The authors are grateful to Dr. Dmitry Isakov for helpful discussion of the results and valuable comments on this manuscript. They also thank Dr. Kristina Shainidze for her assistance in making histology images. This work is supported by Grant of Russian Foundation for Basic Research no. 09-04-00429a.