Beyond foreign-body-induced carcinogenesis: Impact of reactive oxygen species derived from inflammatory cells in tumorigenic conversion and tumor progression


  • Futoshi Okada

    Corresponding author
    1. Department of Biomolecular Function, Graduate School of Medical Science, Yamagata University, Yamagata, Japan
    • Department of Biomolecular Function, Graduate School of Medical Science, Yamagata University, 2-2-2, Iidanishi, Yamagata 990-9585, Japan
    Search for more papers by this author
    • Fax: 81-23-628-5230


Foreign-body-induced carcinogenesis is a traditional, maybe old, way of understanding cancer development. A number of novel approaches are available today to elucidate cancer development. However, there are things we learn from the old, and thus I bring out some examples of various clinical cases and experimental models of foreign-body-induced tumorigenesis. What is notable is that the foreign bodies themselves are unrelated to each other, whereas commonly underlying in them is to induce inflammatory reaction, especially stromal proliferation, where those exogenous materials are incorporated and undigested. Such foreign-body-induced carcinogenesis is also recognized in the step of tumor progression, the final step of carcinogenesis that tumor cells acquire malignant phenotypes including metastatic properties. And the phenomenon is universally observed in several cell lines of different origins. In this review I would like to show the evidence that tumor development and progression are accelerated inevitably by inflammation caused from foreign bodies, and that reactive oxygen species derived from inflammatory cells are one of the most important genotoxic mediators to accelerate the process. © 2007 Wiley-Liss, Inc.

The cancer research advanced in the last decade has pointed out that at least 5 cancer-causing factors exist: spontaneous replication errors in DNA, cytotoxic and/or inflammatory carcinogenic substances, genotoxic (direct DNA injurious) carcinogenic substances, irradiation including ultra violet and transduction of viral oncogenes. Infection/inflammation is unquestionably among them. In 1981, cancer epidemiologists estimated that around 10% of cancers was due to infection/inflammation in the Unites States.1 And the most recent statistics show the major involvement of inflammation in the total cancer death, at various proportions among countries; for example, in 2005, the ratio of infection/inflammation to all the cancer death causes was around 25% in sub-Saharan Africa while that in Europe was around 6%.2 In the year 2000, collectively, ∼20% of all cancer deaths was attributable to inflammatory reactions due to chronic infections caused by infectious agents.3

In the same line of the evidence, foreign bodies incorporated into body for medical reasons or accidentally appear to lead to inflammation-based carcinogenesis. One of the epoch-making experimental studies in this aspect was established by Dr. Boone and his colleagues.4, 5, 6, 7, 8, 9 They observed tumorigenic conversion of immortalized cell lines after implantation of the cells attached to foreign bodies such as a piece of plastic plate or glass bead. Following their finding, we expanded the experiment to tumor progression, namely, the final stage of carcinogenesis acquiring malignant properties, and revealed that inflammatory-cell-derived reactive oxygen and nitrogen species were actually involved in the malignant progression. We also confirmed the universality of the phenomena in the cell lines of rat, mouse and human.

In the industrialized world, including developing and undeveloped countries alike, foreign-body-induced carcinogenesis tends to be poorly managed. A typical example is pleural mesothelioma among the particular workers in Japan; the disease is now belatedly being recognized after their years of exposure to asbestos in the past. Besides those, parasite-related or environmental carcinogenesis is among the foreign-body-induced carcinogenesis; however, in this article, I focus on the materials related with industrial products. I will present the past and present evidence for foreign-body-carcinogenesis and demonstrate that host inflammatory reaction is one of the major factors promoting carcinogenesis.

Foreign-body-induced carcinogenesis in human

Exogenously incorporated foreign bodies can induce tumors in human. Various materials are implanted in human body for prosthetic, reconstructive or cosmetic purposes today; or accidental implantation occurs, such as bullets or shrapnel at war; further, nondigestible particles and scarring may be added to this category of foreign body.10, 11, 12 Tumor latency period from such implantation/injury till its detection is extremely long. It is estimated that over 25% of tumors appears in the span of 15 years and 50%, 25 years,11 and the overall latent period in human is said to be 20 years.10 It should be noted, however, that despite the increasing use of medical implants over the last 6 decades the frequency of tumor development there is extremely low. The critical difference of foreign-body-induced carcinogenesis from inflammation-based carcinogenesis is the low or rare tumor incidence, although the incidence depends on the composition of foreign body.

Particulate carcinogens

Particulate carcinogens are carbon black, asbestos, diesel exhaust, coal particle, acid aerosols, tobacco smoke and reactive gases such as ozone, sulfur dioxide and nitrogen oxides, and the most steady mechanism of particulate carcinogens is to elicit intense inflammatory responses. The conspicuous feature of inflammation in particle-induced carcinogenesis is the release of inflammatory cytokines and reactive oxygen species (ROS).13 On the other hand, other mechanisms are also reported13, 14; for example, such particles can act as carriers of carcinogen such as polycyclic aromatic hydrocarbons to increase retention period.15

Carbon black, a product of incomplete combustion of carbonaceous fuels and usually regarded as soot, was provisioned by Sir Percivall Pott in 1775 as a dreadful example of environmental carcinogen to cause scrotum tumors in chimney sweepers.16 Soot itself appears to be carcinogenic on skin but possibly not on those of intracorporal organs such as respiratory tract epithelium.17 A long-term cohort study revealed that carbon black or related by-products cause lung cancers in the workers at carbon black manufacturing factories.18 Chronic exposure to particulate carcinogen increases tumor incidences to around 20%, with the majority of tumors appearing after 18 months of exposure.19 While we are blessed with numerous products by technological innovations, regrettably some products are double-edged, contrary to the original aim of their production. The common features of the foreign bodies that induce carcinogenesis are listed in the later section. We need to pay attention to newly developed products before they appear in the market. Nano-scale particles, for example, are a great worriment because of the nature of their structures and those products are now spreading into every aspect of our daily life.20

Asbestos fibers

Asbestos is the commercial name for a group of hydrated magnesium silicate fibrous minerals, which is used as materials of cement, ceiling, automobile brake linings, and shipbuilding for their characteristic resistance to heat and combustion.21 Asbestos is in 3 types of crystals: chrysotile (white asbestos), crocidolite (blue asbestos) and amosite (brown asbestos). The carcinogenicity of asbestos is believed to be related with the structural type rather than physiochemical properties.22 Imports of asbestos increased in Japan from early 1960 and had a peak in 1974,23 although there was general warning in 1973 that inhalation of asbestos causes pneumoconiosis, bronchial carcinoma, carcinoma of the gastrointestinal tract and mesotheliomas.13, 24 In parallel with asbestos consumption, the risk of developing mesothelioma (commonly pleural, and also peritoneal and pericardial mesotheliomas, depending on the distribution of mesothelium in body cavities) was belatedly recognized in people who had worked with asbestos. The mean latency of mesothelioma development is estimated at 20–40 years after the first exposure to asbestos.21, 25

When asbestos is experimentally injected into a subcutaneous space in mice, extensive inflammatory reaction develops; surprisingly the subcutaneously injected asbestos fibers are transported into submesothelium of thorax and abdomen.26 The conspicuous feature of asbestos-associated mesotheliomas is persistent inflammation27 and active fibrosis28 evoked by asbestos fibers since immune system cannot remove the nondigestible particles that lead to chronic inflammation (also termed “sterile inflammation”).29, 30 It is considered that possible causes of asbestos-induced mutation are ROS, reactive nitrogen species (RNS) and their by-products. Asbestos fibers themselves catalyze the formation of hydroxyl radicals.31, 32 The production of hydroxyl radicals in cells after treated with crocidolite asbestos may result in the formationof premutagenic DNA basis, namely 8-hydroxy-2′-deoxyguanosine (8-OHdG).33 Moreover, asbestos fibers stimulate infiltrated phagocytes to produce nitric oxide.34 Thus in the environment of asbestos deposit site, RNS, in addition to generated ROS, may react with superoxide anion to produce peroxynitrites, which are able to oxidize guanosine to produce 8-OHdG. Unfried et al. have confirmed the evidence in in vivo situation that 8-OHdG- dependent mutagenesis in the mesothelium (i.e., mostly G to T transversions) occurred in rats after injection of crocidolite asbestos.35

In Japan, an exponential increase of deaths from pleural mesothelioma was observed in the early 1990's. The mean annual number of deaths from malignant pleural mesothelioma was estimated at 500 between 1995 and 1999, and double-fold in 2005.21 Considering the long latency period of mesothelioma, the incidence is expected to increase further in the coming decades.36


Kawanishi et al. have proposed the hypothesis that carcinogenicity of metal compounds is in 3 ways of inducing oxidative DNA damage.37 First, metal generates ROS directly; second, ROS are generated through induction of inflammatory reaction by metal; third, metal activates carcinogenic chemicals. Inflammation-inducible metals are chromium (VI), cadmium, lead, cobalt (II), iron (III) and nickel compounds such as NiSO4, NiO, Ni3S2 while Ni3S2 has a role to induce oxidative DNA directly in the presence of hydrogen peroxide.37, 38, 39

Medical devices

A wide variety of synthetic materials have been implanted in human since 1940's. In the middle to late 1960's, there were a few anecdotal reports, but they seldom mentioned the association between carcinogenicity and implantation of plastic materials. A long-term investigation of cosmetic breast surgery (breast augmentation) was the most typical example. Of 40,000 cases of implantation of polyvinyl alcohol sponges, breast cancer developed only in 6 cases (0.015%).40 A cohort study revealed that the overall mortality rate was lower in women with breast implants than women with other plastic surgery or general population.41 And in breast cancer statistics, no differences were found between women with breast implants and those without them in the disease stage at diagnosis, cancer recurrence or survival time.42, 43 Moreover, breast implant after mastectomy did not affect cancer-related mortality.44 It is acceptable and certainly safe to implant foreign body for medical purposes by today's technology. Nevertheless, there are chances that implantation of foreign body is not always safe; e.g., after arthroplasty of hip chromosomal aberrations are frequently found in the adjacent bone marrow.45

Scar cancer and mechanical trauma/wound-healing-associated cancer

Scars reactive to or encapsulation of foreign bodies is the fundamental and common host reactions (also known as fibrosis, desmoplasia or stromal proliferation). In early 1940's, G. Friedrich and R. Rössle found that lung carcinomas grew at the sites of fibrous scars in patients with smoking habit.46, 47 They observed that the scars had preceded the cancer, and thus termed it “scar cancer (Narbenkrebs).” Encapsulation is the oldest host adaptive immune system, and all the living things, from invertebrates to vertebrates, are equipped with the system.48 In man, it is commonly associated with pathological situations and the typically described is fibrous tissue capsule or scar which surrounds foreign bodies. Histological examination shows that fibrous capsule is generally 300-μm thick around the implanted materials; however, no association is observed between the implanted materials and the degree of capsule reaction.49 The initial pathognomonic sign is an emergence of extensively proliferating spots with a population of atypical cells, usually observed adjacent to a foreign body, and the cells are of polygonal and/or spindle cell types, expressing proliferating cell nuclear antigen.49 These lesions can be considered as transitive to preneoplastic lesions.

There are 2 theories as for the induction of scar cancer. One is that the cells composing scar tissues convert themselves directly into tumorigenic ones; the other is that scar tissues secrete soluble factors that stimulate carcinogenic conversion or proliferation of tumor cells pre-existing around the scar. The former theory is supported by the evidence that tumor nodules arise in the serial transplantation of scar tissues which have originally been formed in the mice carrying T6 marker chromosomes; therefore by monitoring the chromosome as tag, one could detect the initial changes when the scar tissue exhibits conversion into malignant tumors at various degrees in the scar environment.50, 51 At this point, the latter theory is more acknowledged, because scar formation occurs along with tumor development, rather than before tumor formation, as seen in lung adenocarcinoma development, for instance.52 Scar can promote malignant growth of tumor cells; for this reason, the minimum take dose of scarred tissue is much smaller than that of unscarred tissue, for lethal growth of transplanted tumor cells.53

Mechanical trauma (post-traumatic inflammation) or wound-healing processes show a similar trend to scar cancer. In 1920's a number of reports were published concerning the association between trauma/wound-healing and cancer, and some of them were the reports of case studies, showing that the incidence of trauma could cause cancer.54, 55 Trauma is considered not to act as initiator, but to act as promoter of carcinogenesis.56 Fibrosis associated with tumor growth provides the environment to stimulate progression of tumor cells. Most of the clinical local recurrences of gastrointestinal tumors after resection of the primary tumor occur at the operated area of incision or anastomosis.57, 58 Whether fibrosis is accompanied with tumor or not depends on the types of collagen it produces.59 Tumor-associated fibrosis is characterized by increased collagen Type III content; such fibrotic lesions are mostly in the early immature stage. On the other hand, fibrosis unaccompanied with tumor contains decreased content of Type III and increased contents of Types I and IV collagen; such fibrotic lesions are mostly in the late mature stage.60

Foreign-body-induced experimental carcinogenesis

In 1941, Turner described the first experimental evidence for foreign body-induced sarcoma in rat using a disk of Bakelite61 which was the first plastic made from synthetic components.62 Brand et al. verified that carcinogenic potential depended on the properties of foreign body such as shape/size, smoothness, hardness, porosity and electrostatic load,63 and was also influenced by gender and strain of the host.64 Oppenheimer revealed that the tumor-forming material was not of degradable nature and had enduring smooth surface,65 which was thereafter referred to as “Oppenheimer effect.”66, 67 The effect is also called “solid-state tumorigenesis” because the carcinogenic conversion is brought about on solid surface of a foreign body.

The shape necessary for carcinogenesis as found in foreign bodies must be in vivo as well. An easily absorbable liquid would provide little or no chance for the conversion; however, if the liquid turns viscous, it can produce fibrosarcoma at the site of implantation.67 The shape also influences tumor frequency; it is much higher in mice implanted with concave plastic discs than in those with convex ones because concave discs evoke more intense fibroblastic reaction.10 Moreover, tumor incidence correlates with surface area of implants because the size of foreign bodies determines the degree of the inflammatory reaction.10 Tumor formation seldom occurs when a small material has been implanted, but a large foreign body produces tumors constantly.68 Textile materials, rough-surfaced implants and minced materials have little potential to induce tumors, whereas, if they are implanted, uninterrupted and smoothly-surfaced, they develop tumors.10, 28, 69

Porosity of the materials is influential. When implants are perforated and the holes are large, tumor incidence is reduced.10, 68 For instance, filters with pores of 0.22 μm or larger do not induce tumors whereas those of the same thickness with smaller pores will do. This is because the implants poorly develop connective tissue capsules and the pore is infiltrated with phagocytes in the former,10 but the pore sizes below 0.22 μm are surrounded by thick collagen capsules and the filter pores are not infiltrated with phagocytes. The latter fibrous capsules are highly vascularized and the capillary loops develop close to the implant surface (less than 30 μm); on the other hand, angiogenic response is suppressed in the area of nontumorigenic filter implantation and the mean distances between the implant surface and capillaries remain 300μm.10 The filter with large porosity was found to have less cell proliferation, apoptosis, and fibrosis.70 Moreover, hydrophobic filters develop more tumors than mice implanted with hydrophilic ones.10 The electrical effects, interfacial forces or electrostatic and/or electrokinetic imbalance of implanted materials affect carcinogenic incidence.53, 69 Foreign materials positive for electric charge spontaneously attract inflammatory cells or easily form thrombus because most inflammatory cells, especially neutrophils, are negative for electric charge.28

It is of interest to know what cell types are responsible for foreign-body-induced carcinogenesis. There are 3 candidates, macrophages, fibroblasts and pericytes.10 However, it would be appropriate to exclude macrophage as the early experiment suggested its irrelevance. Namely, in the experiment of bone marrow macrophage transplantation using distinguishable syngeneic mice, genetically marked sarcoma was not found in the mice whose bone marrow had been replaced to genetically tagged marrow cells.10 Histologically the predominant cell type to form capsules around implants is fibroblast while foreign body-induced tumors are predominantly sarcomas. Fibroblast is a dubious entity. While development of fibrosarcoma is dominant in chemically-induced carcinogenesis, a variety of sarcomas appear in foreign-body-induced carcinogenesis: fibrosarcomas, malignant fibrous histiocytomas, pleomorphic sarcomas, myxosarcomas, hemangiosarcomas, rhabdomyosarcomas, osteosarcomas, leiomyosarcomas and mixed type.49 This may illustrate that the initial mesenchymal stem cell reactive to foreign body is pluripotent. Pericyte is another candidate since its initial role is to support local angiogenesis. However, because of their pluripotent potential as a local mesenchymal cell population, they can convert themselves into sarcomas with various histological characteristics.10

Genetic alterations in the tumors arisen in the foreign body implantation have extensively been analyzed initially by karyotyping. Derivations from the normal chromosome number are seen in all the tumors studied10; morphological aberrations of chromosomes such as metacentrics or double minutes, and conspicuous abnormalities are recognized. Most karyotypes of foreign-body-induced tumors are exhibited in the hyperdiploid or the hypo- to hyper-tetraploid ranges.10 Tazawa et al. identified some genes possibly responsible for the development of foreign-body-induced sarcomas.71 They found that in 79% heterozygous p53-deficient (Trp53+/−) mice with only 1 functional allele in the p53 gene developed spontaneous sarcomas after plastic plate implantation, whereas only 10% of mice with wild-type p53 developed sarcomas. They further demonstrated that the arising tumors have lost the remaining wild-type p53 allele, meaning complete loss of p53 function; this appears to be the underlying molecular mechanism during the development of sarcomas. That is, the tumor suppressor gene p53 is one of the genes responsible for foreign-body-induced carcinogenesis. p53 allele is inactivated by an increase of inflammation-mediated RNS.71 Association between RNS and p53 mutations is evidenced in inflamed lesions of the colon with ulcerative colitis,72 stomach,73 brain74 and breast.75

The susceptibility to foreign-body-related carcinogenesis does exist across species, and in strain- and gender-dependent manner.10 Carcinogenicity of foreign bodies in human is rare, especially that of clinically used materials such as polyethylene, polyurethane, polyvinyl chloride, polymethylmethacrylate, silicone, titanium, nickel chromium, cobalt-chromium alloy and aluminum oxide; however, the same materials are indeed responsible for 25.8% incidence of sarcomas in 490 rats examined.49 There is no association between histological types of sarcoma and implanted foreign materials, although polyurethane is the only material which tends to form hemangiosarcomas.49

Experimental rodent animals, except guinea pigs, exhibit high incidences of foreign body-induced carcinogenesis.11 Yet differences do exist among the strains of mice in the incidence and the latency period.64 And individuals genetically sensitive to foreign-body-induced tumors do exist in certain areas,76 indicating differences in the sensitivity among races or groups of inhabitants. It is expected that there are labile gene(s) for this. Gender differences are also exhibited in mice. Latency period of tumor development is shorter in female mice than in male mice; however, the malignancy of the arising tumors is much enhanced in male mice.10 By comparing the differences in sensitivity or resistance to foreign body-related carcinogenesis we might be able to find a clue to identify the molecules that control foreign-body-related carcinogenesis.

Experimental models of foreign-body-induced carcinogenesis and tumor progression

Serendipitous discovery of the foreign-body-induced carcinogenesis was made by Dr. Boone's group.4, 5, 6 They showed that cells of nontumorigenic but immortalized mouse cell lines or freshly isolated connective tissues from mice were converted into lethally growing tumors of monoclonal origin in mice after they were implanted, being attached to foreign bodies.4, 5, 8, 9 They suggested at least 5 causes for the conversion: (i) immortalized cells acquired preneoplastic phenotype for the lack of anchorage-independent growth property, and thus the substrate-attached cells grew exponentially in vitro and lethally in vivo4, 5; (ii) activation of endogenous oncornaviruses. However, this possibility was low6; (iii) culture medium contained carcinogenic substances7; (iv) culture dish secreted carcinogenic substances.8 Heppner, one of their contemporary researchers, speculated that inflammation was one of the possible causative factors for foreign-body-induced tumorigenesis.77 Since then, we have expanded the experimental systems and demonstrated that foreign-body-induced inflammation and its-derived ROS are definitely the cause for the conversion.

To establish the experimental model of conversion, we chose the cells which were weakly- or nontumorigenic and nonmetastatic, but would grow in vitro, bearing the idea of xenogenization in mind. Xenogenization of tumor cells is the term meaning immunological spontaneous regression of tumor cells which had been infected with xenogeneic viruses,78 transfected with the genes coding allogeneic antigen,79 or exposed to mutagenic chemicals,80 after injected into normal syngeneic host. Another approach to establish the model is to obtain the culture cell lines from a precancerous lesion.

Before I detail our unique animal models in which we can consistently observe not only tumor development but also malignant progression of regressive tumors or precancerous cells originated from rat, mouse and human, I should briefly trace the history of our experiments. Following Dr. Boone's work, Drs. Takeichi and Hamada expanded the concept of foreign-body-induced carcinogenesis to foreign-body-induced tumor progression. They obtained weakly tumorigenic and nonmetastatic clonal ER-1 cells by exposing SST-2 culture cell line, which had been established from a spontaneous mammary adenocarcinoma developed in SHR rat, to ethyl methanesulfonate in vitro.81 The ER-1 cells regressed spontaneously after injected into syngeneic normal rats; however, if they were implanted, being attached to plastic plates (polystyrene, used as culture dish), into a subcutaneous space in rats, they acquired not only tumorigenicity but also metastatic ability.81 The tumors arisen in rats showed various malignant phenotypes and their acquired phenotypes were stable. Thus they found that chronic inflammation was required for the progression of ER-1 cells. This model mimics inflammatory breast cancer in human, the most aggressive form of primary breast carcinoma with a dismal outcome despite multimodal treatment approaches.82

They also determined that fibrous stroma secreted soluble factors such as epidermal growth factor (EGF) and transforming growth factor-ß (TGF-ß).83 The ER-1 cells were turned into invasive and metastatic tumor cells by EGF continuously added into the culture.84 Those phenotypic changes depended on the duration of the EGF treatment; the malignant features were reversible during the 24-hr exposure to EGF, and after more than 4 weeks of exposure the acquired malignant phenotypes were completely fixed.84 It has been confirmed that acquisition of malignant phenotypes can be prevented by addition of antioxidants, N-acetylcystein or selenium, into the medium,84, 85 and that 8-OHdG, a marker of ROS-mediated nucleic acid damage, is formed in the EGF-treated ER-1 cells; therefore, it is strongly suggested that the fundamental cause for the malignant progression is ROS generated by the fibrous stroma-derived growth factors.84

The mouse model has thoroughly been investigated, and reveals a definite connection between foreign-body-induced inflammation and carcinogenesis. For further experiments, we obtained regressive clonal QR cells by exposing clonal fibrosarcoma cells, BMT-11 cl-9, to a mutagen, quercetin in vitro.80 Since QR cells grew progressively in immunosuppressed hosts, their regression was mediated by host immunity.86 QR cells did not form tumors or metastasis after subcutaneous (2 × 105 cells) or intravenous (1 × 106 cells) injection into mice.80 However, implantation of 1 × 105 QR cells attached to plastic plate87 or injection into the preinserted gelatin sponge88 in subcutaneous space of mice induced lethal tumors. Moreover, the arising tumor exhibited metastatic properties. The acquired various malignant phenotypes remained stable as far as examined for 1 year at least under cultivation in vitro.

We detected several gene alterations through the malignant conversion in this model.89 The level of thymosin ß4 gene, which is known as an actin-regulating protein and to function for angiogenesis and wound healing,90 was elevated in all of the arising tumor cells. From the results of sense and antisense cDNA transfection experiments, we determined that thymosin ß4 gene was responsible for tumor metastasis through regulating cell motility.91 The expression of E1AF, a member of the ets oncogenic transcription factor, was found high in the arising tumor lines.92 E1AF regulates tumor cell motility and invasive activities through induction of membrane-type 1-matrix metalloproteinase which converts the latent form of matrix metalloproteinase-2 into active form.92 Thus E1AF makes tumor cells invasive.92

We used two foreign bodies in our decades of experiments. One was a piece of plastic plate and the other a piece of gelatin sponge which is used as hemostasis material during surgical operation. The difference between the foreign bodies is in the capacity to elicit inflammation. Plastic plate initially induces acute inflammation, which then changes to chronic inflammation, whereas gelatin sponge induces mild inflammation, and about 4 weeks after implantation it is naturally absorbed93; therefore such transition from acute to chronic inflammation is unlikely to occur in the use of sponge.88 By using those foreign bodies, we modulated the quality and duration of the inflammation, and found that the type of inflammation needed for QR cells' growth and progression was acute-phase inflammatory reaction.87, 88, 94 By histological examination, we found that neutrophils predominantly infiltrated the sponge.93, 94 In fact, one of the benefits of using gelatin sponge is that it is possible to collect the infiltrated inflammatory cells by treating the sponge with collagenase; the inflamed cells separated from the sponge can convert QR cells into tumorigenic ones if both cells are mixed and injected.88, 94 Namely, our primary observation was that foreign-body-reactive, early-phase inflammatory cells contribute to augment malignancy of QR cells.87, 88, 94

To test the role of neutrophils during inflammation, we eliminated neutrophils by administering anti-neutrophils antibody (RB6). Nearly all the arising tumors in the mice, nontreated or treated with control rat IgG, acquired malignant phenotypes. On the other hand, RB6 antibody-administered mice did not acquire malignant phenotypes.94 We confirmed the results in integrin-β-2 knockout mice (C57BL/6JItgb2tm1Bay equivalent to CD18- deficient).94 Integrin-β-2 is the key adhesion molecule for the migration of neutrophils into an inflammatory region. Neutrophil infiltration into gelatin sponge is abolished and acquisition of malignant phenotypes is suppressed in the integrin-β-2 knockout mice.94 These findings show that neutrophils are one of the main components of inflammation-associated tumor development and progression. Interestingly, the capability of neutrophils to accelerate tumor cell malignancy depends on their activation phase since circulating or bone marrow neutrophils do not convert regressive tumor cells into malignant ones.94

It is reported that tumor-infiltrated neutrophils have the role to induce angiogenesis via MMP-9 secretion.95 We assumed that the neutrophils also produced genotoxic substances since we detected somatic gene mutation in the coculture of QR cells and the infiltrated neutrophils.96 The somatic mutation was inhibited in the presence of a radical scavenger, mannitol96; and in our previous study, immunostaining of 8-OHdG in the tumor tissues evidenced neutrophils' infiltration.93, 94 Therefore we examined the roles of neutrophil-derived ROS in this model. To determine the direct contribution of ROS to carcinogenesis, we used gp91phox gene knockout mice. Bactericidal function of neutrophils brought about generation of superoxide anions by forming NADPH oxidase complex (gp22phox, gp40phox, gp47phox, gp67phox, gp91phox and Rac1/Rac2) from interaction with cytochrome b558.97 The frequency of tumor development from the QR cells coimplanted with gelatin sponge was decreased in the gp91phox−/− mice. Moreover, acquisition of metastatic ability was reduced in the mice.98 Thymosin ß4, a genetic marker of QR cell progression,91 was not detected in the arising tumors in gp91phox−/− mice. To confirm whether phagocyte-derived ROS were actually involved in tumor progression, we isolated phagocytes from wild-type mice and transferred them into gp91phox−/− mice. As a result, wild-type-derived phagocytes increased the frequencies of tumor development and metastasis. In contrast, the phagocytes obtained from knockout mice did not have such activity.98 Moreover, administration of aminoguanidine, a broad inhibitor for inducible nitric oxide synthase, partially but significantly suppressed malignant conversion in the model93; thus we concluded that RNS were also involved in the process. These results show that ROS and RNS, derived from foreign-body-induced neutrophils, are an intrinsic factor in the conversion of regressive tumors to more malignant ones. The mouse model may recapitulate the typical inflammation-based carcinogenesis, and thus be suitable for analyzing biological causes of the process. Gelatin sponge implantation induces massive and persistent infiltration of activated leucocytes; such situation may provide pathogenic resemblance to the continuous infiltration of activated inflammatory cells into target organs by bacteria (Helicobacter pylori99) or parasites (Opistorchis sp., Chlonorchis and Schistosoma100) infections in human.

Since we used rodent cells previously, we have been eager to prove the phenomena by using cells of human and epithelial origin. For this purpose, instead of establishing regressive tumors from lethally growing tumors, we have picked up an available cell line derived from precancerous tissues of colon. In patients with ulcerative colitis or Crohn's disease, a typical inflammation-related carcinogenesis is seen; that is, continuous infiltration of disordered immune cells into the autologous colon tissue precedes the development of colorectal cancer.101, 102 A culture cell line of adenoma cells (FPCK-1 cells) was established by Dr. Kawaguchi from a colonic polyp of a patient with familial adenomatous polyposis.103 The phenotype of the cell line was stable and no spontaneous conversion was observed during the maintenance under regular cultivation at least for 1 and half years.

We used several sister lines of adenoma FPCK-1; they did not grow in nude mice when injected at 5 × 106 cells in a suspension form. However, adenoma cells of FPCK-1-1 formed palpable tumors in about 5 months when they were attached to plastic plate and implanted into subcutaneous space of mice (Fig. 1a).104 Their tumorigenic growth always started from 1 colony (monoclonal origin) (Fig. 1b). Serial implantation of the culture cell line with plastic plate, established originally from the arising tumor after implantation of the adenoma cells with plastic plate, shortened the latency period, in correlation with the number of passage (Fig. 1c).

Figure 1.

Features of the arising adenocarcinoma after implantation of human colonic adenoma cells being attached to plastic plate. (a) Adenoma cells (1 × 105 cells per plate) were implanted into a subcutaneous space of nude mouse. Bar, 5 mm. (b) Around 5 months later, a palpable mass with forming angiogenesis appeared; the emergence was always monoclonal growth. Bar, 5 mm. (c) Latency period was shortened after serial transplantation of the arising adenocarcinoma cells with plastic plate attached. (d) Histologically the arising tumors were moderately differentiated adenocarcinoma with proliferating stroma (Azan staining). Bar, 100 μm.

Histologic examination revealed that the arising tumors were moderately differentiated adenocarcinoma, and they were surrounded by highly collagenic fibrous stroma (Fig. 1d). The fibrous tissue, rather than attachment to the plastic plate substrate, was considered essential for the conversion, because we observed tumor growth after injection of the adenoma cells directly at the site of proliferating stromal tissues where the plastic plate had been implanted for about 5 months and then removed. In contrast, there was no tumor development in nontreated or sham-operated mice, and tumor seldom arose after coimplantation with gelatin sponge in those, which indicated that acute inflammation did not suffice to convert the adenoma cells.104 The stromas at the chronic inflammatory region secrete specific soluble factor(s), which have not been identified yet; we assume that the factor(s) could stimulate adenoma cell growth but not the growth of adenocarcinoma cells. Such factors could not be produced either from normal subcutaneous fibroblasts or immortalized mouse fibroblast cell lines.104

In our foreign-body-induced carcinogenesis and tumor progression models, we can observe that inflammation promotes malignancy of various species of cells. As summarized in Table I, the phenomenon is universally observed among species and in tissues of different origins.

Table I. Universality of Carcinogenesis and its Progression Induced by Foreign Bodies in the Cells Derived from Immortalized Cells, Regressive Tumor Cells or Precancerous Lesions
CellsOrigin of the cellImplantation hostForeign bodyCharacteristics of arising tumors (reference)Year
BALB/3T3mouse, vascular endotheliummouseglass beadsmalignant hemangioendothelioma (4)1975
NCTC 8467mouse lung tissuemouseglass helicessarcoma (105)1980
BALB/3T3mouse, vascular endotheliummousepolycarbonate plastic platevasoformative sarcoma (5) 1976 
C3H/10T1/2mouse, embryomousepolycarbonate plastic platefibrosarcoma (5)1976
C3H/10T1/2mouse, embryomousepolycarbonate plastic plateinvasive fibrosarcoma (9)1979
Connective tissues obtained from adult mice mousepolycarbonate plastic plateundifferentiated sarcoma (8)1979
ER-1rat, regressive mammary adenocarcinomaratpolystyrene plastic platemalignant, sarcomatoid histology (81)1992
QR-32mouse, regressive fibrosarcomamousepolystyrene plastic platemalignant fibrosarcoma (87)1993
FPCK-1-1human, colonic adenomamousepolystyrene plastic platehigh tumor incidence, adenocarcinoma (104)2000
Benzo(a)pyrene-induced sarcoma mousepolystyrene plastic plateshortened latency period (112)2002
Subcutaneous space of mice p53 heterozygous mousepolystyrene plastic platehigh tumor incidence, fibrosarcoma (72)2007
Subcutaneous space of mice p53 wild-type mousepolystyrene plastic platelow tumor incidence, fibrosarcoma (72)2007
QR-32mouse, regressive fibrosarcomamousegelatin spongemalignant fibrosarcoma (88)1992
Lk9dLrat, nonmetastatic renal carcinomaratgelatin spongemetastatic renal carcinoma (107)1998
FPCK-1-1human, colonic adenomamousegelatin spongelow tumor incidence, adenocarcinoma (104)2000

Perspective of prevention of foreign-body-induced carcinogenesis

Inflammatory environments due to the existence of foreign body cause a variety of biological responses as they contain increased growth/survival factors, chemotactic cytokines (chemokines), matrix metalloproteases, adhesion molecules, extracellular matrix, inflammatory mediators (i.e., histamine, eicosanoids, inflammatory cytokines and proteases), DNA-damage-promoting agents (i.e., ROS and RNS) and augmented angiogenesis108 (Fig. 2). I have been emphasizing the roles of inflammatory-cell-derived ROS in the foreign-body-induced carcinogenesis. In this section, I would like to suggest some strategies to prevent it. If the inflammatory-cell-derived ROS are critical to promote foreign-body-induced carcinogenesis, the sensitivity of cells to the conversion should be controlled by intracellular or tissue antioxidative potentials. Our earlier study demonstrated that a variation was observed among the QR clones in the frequency of tumor progression by coimplantation with foreign body.96 We determined that the variation was due to intracellular antioxidative enzymes, manganese-superoxide dismutase, Mn-SOD and glutathione peroxidase, GPx level in the QR clone. An inverse correlation was observed between the contents/activities of those enzymes and the sensitivity of QR clones to progress under foreign-body-induced inflammatory environment.96 Thus, we determined that cells with low antioxidative levels are prone to convert themselves into more malignant ones in the presence of inflammation and the inflammatory cell-derived ROS. In other words, foreign-body-induced carcinogenesis in QR cells will be prevented if the cells have an adequate amount of antioxidative enzymes or induction of the enzymes at the implantation site. The prevention was actually achieved by induction of Mn-SOD at the coimplantation site by administering Mn-SOD-inducible biological response modifier109 or orally available superoxide dismutase in this system.110 Moreover, it is assumed that RNS are partially involved in the model, because administration of a broad inhibitor for inducible nitric oxide significantly inhibited inflammation-induced tumor progression.93

Figure 2.

Key biologic features of foreign-body-induced carcinogenesis. Malignant conversion of precancerous cells or progression of regressive tumor cells requires foreign body-induced inflammatory soluble mediators. The mediators can be classified as (i) growth promotion, (ii) inflammation and host immune response to foreign body and (iii) genotoxic substances (i.e., reactive oxygen and nitrogen species). Each precancerous cell or regressive tumor cell possibly carries a variety of gene alterations. And only one of the cells may undergo the final crucial molecular changes that convert/progress to malignant one. Such clone-originated cell acquires various malignant phenotypes during proliferation under the influence of inflammatory soluble mediators.

By extensively analyzing human tissue materials obtained from typical inflammation-based carcinogenesis, it has been revealed that ROS and RNS are inevitably involved in the development and progression of tumors.111, 112 Considering those clinical studies and our results, we conclude that the proneness of tumor cells to be more malignant depends on the balance between antioxidative enzyme activities in themselves and the duration/amount of ROS and RNS generated by inflammatory cells, and that disturbing the balance will be the most effective strategy for the prevention of foreign-body-induced carcinogenesis.