Endogenous type I interferons as a defense against tumors

Cytokine & Growth Factor Reviews 13 (2002) 111–118

Endogenous type I interferons as a defense against tumors Ion Gresser a,∗ , Filippo Belardelli b b

a INSERM U255-Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France Laboratory of Virology, Istituto Superiore di Sanità, Viale Regina Elena 299, 00161 Rome, Italy

Abstract We have reviewed the experimental results which indicate that endogenous type I interferon (IFN) present either constitutively or possibly induced by the tumor plays an important role in limiting the development of transplantable tumors in mice. Thus, treatment with potent polyclonal neutralizing antibodies to IFN ␣/␤ markedly enhanced the subcutaneous growth, invasiveness and metastases of xenogeneic tumor cells (uninfected or infected with RNA or DNA viruses) in athymic nude mice; enhanced the intraperitoneal transplantability of six different syngeneic murine tumors in three strains of immunocompetent mice; and completely abrogated the resistance of allogeneic C57Bl/6 (H-2b ) or C3H (H-2k ) mice to the multiplication of Friend erythroleukemia cells (H-2d ) in the liver and spleen resulting in the death of most mice. The mechanisms by which mice respond to the injection of relatively few tumor cells appear to be multiple, to depend on the site of tumor growth, to occur early and prior to an immunologic response. Endogenous type I IFN appears to constitute an essential component of these defense mechanisms enabling the host to restrict tumor growth. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Endogenous type I interferon; Immune surveillance; Tumor

1. Introduction It has been suggested that the immune system plays a role in inhibiting the emergence and development of autochthonous tumors [1–3]. Although the immune system is clearly important as a defense mechanism of the host against a variety of infectious pathogens, it is still uncertain whether it is important in antitumor defense. Patients with an impaired immune system either because of genetic or acquired deficiencies (infection, immunosuppressive agents, etc.) often show an increased frequency of certain malignancies such as lymphomas, carcinomas of the vulva, perineum, skin, lips, and hepatobiliary tract, all of which are often associated with viruses such as EBV, papilloma virus, hepatitis B or C viruses [4]. The incidence of cancers usually observed in the general population (i.e. cancers of the lung, prostate, colon, rectum and breast) is not increased [4]. Likewise, the tumors developing in immuno-compromised nude mice have been almost exclusively lymphomas rather than epithelial cancers [5]. Thus, an immune surveillance system against tumors may be directed against viral or ∗ Corresponding author. Present address: UNITE’INSERM 255 Immunologie Cellulaire et Clinique, Centre de Recherches Biomedicales des Cordeliers, UPMC, 15 Rue de l’Ecole de Medicine, 75270 Paris Cedex 06, France. Tel.: +33-1-45035182; fax: +33-1-40510420. E-mail address: [email protected] (I. Gresser).

other infectious agents rather than against the proliferation of neoplastic cells themselves. Our experimental work in the past two decades has indicated that endogenous type I interferon (IFN) either present constitutively or possibly induced by the tumor plays an important role in limiting the development of tumors in mice. We will review the results essentially from our laboratories involving type I IFN (␣/␤) and discuss possible mechanisms by which these IFNs may interact with the host to control tumor development. More recent work by other groups using different experimental models has also indicated a role for type II IFN (IFN␥) in tumor surveillance and this topic will be covered elsewhere in this issue. 1.1. Background Three lines of research in the past years on the biologic effects of type I IFNs have led us to the hypothesis that endogenous IFNs may play a constitutive role in restricting the emergence and development of tumors. 1. Type I IFNs can exert a marked antitumor activity in experimental animals [6,7] and in patients [8]. The mechanisms of this antitumor effect appear to be multiple and complex. Whereas in some instances IFN may act directly on the tumor itself, in many other instances IFN appears to induce or enhance host mechanisms— including immune-mediated effects—to achieve inhibi-

1359-6101/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 9 - 6 1 0 1 ( 0 1 ) 0 0 0 3 5 - 1

112

I. Gresser, F. Belardelli / Cytokine & Growth Factor Reviews 13 (2002) 111–118

tion or rejection of the primary tumor and tumor metastases. 2. We and others have shown that IFN is produced constitutively in the normal and axenic mouse and that this endogenous type I IFN is responsible for maintaining some cells (macrophages [9], and lymphocytes [10]) in an antiviral state. 3. We produced a potent polyclonal neutralizing antibody to type I IFN, and using this antibody demonstrated the importance of type I IFN as a primary defense in numerous experimental viral infections of mice [11,12]. Using this antibody to type I IFN, we sought to determine whether there was a baseline resistance in immunocompetent and immunodeficient mice to the development of various transplantable tumors and whether this resistance was mediated by endogenous type I IFN. 2. Evidence that type I IFNs are important as a defense against experimental tumors 2.1. Antibody to type I IFN enhances the growth and spread of virus-infected xenogeneic tumor cells in nude mice Athymic nude (nu/nu) mice have been widely used as experimental animals for transplantation of tumors and tumor cell lines. Thus, a variety of allogeneic and xenogeneic cell lines injected s.c. into nude mice grow and form tumors that are usually encapsulated and neither invade the surrounding tissues nor metastasize. In our first study with B. Bloom’s group, we found that injection of nude mice with antibody to type I IFN markedly enhanced the growth of transplanted human HeLa or hamster BHK cells [13] (Table 1). Histopathologic examination of the tumors in mice injected with antibody to IFN showed: (a) absence of a stromal cap-

sule; (b) absence of a significant host cell infiltrate; (c) invasiveness into surrounding tissues; (d) increased incidence of metastases. The implication from these observations was that even in nude mice there were IFN-induced host mechanisms (probably not T cell immune-mediated) that limited tumor invasiveness and metastases. (A direct effect of IFN on the tumor would seem to be excluded as mouse IFN does not act to any significant degree on either human or hamster cells.) When these xenogeneic cell lines were pre-infected with mumps or measles viruses [13], the tumor cells grew poorly or not at all in nude mice. In contrast, when mice injected with these virus-infected xenogeneic cells were also treated with antibody to IFN␣/␤, the virus-infected cells grew rapidly, were invasive in all mice and even metastasized in some mice (Table 1). One of the most surprising observations from this study was the significant proportion of anti-IFN treated mice developing tumor metastases [13]. Thus, in a study of over 1000 nude mice, metastases were rarely observed, although as few as 10–100 highly tumorigenic BHK and HeLa cells formed local progressive tumors [13]. In contrast, all the anti-IFN␣/␤ treated mice injected with uninfected or virus-infected HeLa or BHK cells, showed invasiveness of adjacent tissues by tumor cells and approximately 30% of the animals developed tumor metastases [13] (Table 1). Antibody treatment of Balb/c mice injected with HBV-infected PLC/PRF/5 cells also resulted in a marked increase in tumorigenicity; the tumors appeared more rapidly, the minimal number of cells necessary to induce a tumor was less, and the tumors were all invasive [14] (Table 1). It was also of interest that treatment with antibody to IFN␣/␤ resulted in a far greater increase in tumorigenicity than in mice treated with either X-ray or anti-lymphocytic sera [14].

Table 1 Results indicating that type 1 IFN plays a role in the defense against the development of xenogeneic transplantable tumors (subcutaneous) in immunodeficient nude (nu/nu) micea Mouse strain immunocompetence

Tumor cells transplanted

Virusinfected

Effect of injection of antibody to IFN␣/␤ compared with control injected mice

Reference

Tumorigenicity (increased tumor growth and/or increase in % mice with tumor)

Invasiveness (in all mice)

Metastases (% mice)

Balb/c (nu/nu)

BHK BHK HeLa HeLa HeLa

No Mumps No Measles Mumps

+ + + + +

+ + + + +

+(37) +(25) +(27) − −

[13]

Balb/c (nu/nu)

Surgical specimens of human prostatic adeno-carcinoma PLC/PRF/5 PLC/PRF/5 U 937

No

+

+

−

[13]

HBV HBV HIV

+ + +

+ + −

− − −

[14]

Balb/c (nu/nu) CBA/n (nu/nu) Swiss (nu/nu) a

[15]

BHK: baby hamster kidney; HeLa: human cell line derived from cervical carcinoma; PLC/PRF/5: human hepatoma cell line which synthesizes HbsAg in vitro; U 937: human histiocytic lymphoma cell line; HBV: hepatitis B virus; HIV: human immunodeficiency virus.

I. Gresser, F. Belardelli / Cytokine & Growth Factor Reviews 13 (2002) 111–118

Splenic lymphocytes from mice injected with virusinfected xenogeneic cells in these experiments [13,14] had elevated levels of natural killer (NK) cell activity for extended periods of time against these cells. This increase in NK cell activity observed in mice injected with virus-infected cells was abrogated in the mice receiving antibody to IFN␣/␤, suggesting that endogenous type I IFN was responsible for the increased NK cell activity, which in turn may have been related not only to inhibition of the primary tumor but also to inhibition of metastases. Large solid tumors developed in nude mice injected with human U 937 cells [15]. In contrast, subcutaneous tumors did not develop in the majority of mice after injection of even 107 HIV-infected U 937 cells. Some tumor growth was observed initially in some mice (10–30%) injected with 2 × 107 HIV-infected cells but these tumors generally regressed 3–4 weeks after tumor cell injection. In contrast, treatment of mice with antibodies to IFN␣/␤ resulted in the development of rapidly growing solid tumors in all mice injected with HIV infected U 937 cells [15] (Table 1). It was of interest that all of the anti-IFN treated mice with established U 937 HIV positive tumors exhibited very high serum levels of HIV p24 levels. As neither tumor growth nor p24 antigenemia was observed in nude mice treated with control sheep globulins and injected with HIV-infected U 937 cells, these results suggested that endogenous mouse

113

IFN␣/␤ was involved not only in the restriction of tumor growth but also in the in vivo inhibition of HIV multiplication [15]. 2.2. Enhanced growth of syngeneic transplantable tumors in immunocompetent mice injected with antibody to type I IFN The results obtained in nude mice injected with a variety of xenogeneic tumor cells and treated with antibody to type I IFN prompted us to determine the effect of this antibody on the intraperitoneal (i.p.) transplantability of eight different syngeneic cell lines in three strains of immunocompetent mice [16]. The results were clear-cut. Injection of this antibody resulted in an increase in the percentage of mice developing tumors for a given cell inoculum; the tumors grew more rapidly and the percentage of mice dying with tumors was increased [16] (Fig. 1). The tumors appeared to spread extensively in the peritoneum of antibody treated mice, but metastases were not observed. It is important to emphasize that the enhancement of tumor growth in antibody treated mice was also observed when IFN␣/␤ resistant tumor cells lines (L1210R [17] or 3Cl8 FLC [18]) were injected, indicating that this endogenous IFN acted via host mechanisms to restrain tumor growth and not on the tumor cell itself. (We had previously shown that administration of IFN␣/␤ was

Fig. 1. Effect of antibody to type I IFN on survival time of mice injected i.p. with different syngeneic mouse tumor cells. Experimental details given in [16].

114

I. Gresser, F. Belardelli / Cytokine & Growth Factor Reviews 13 (2002) 111–118

Table 2 Effect of antibody to IFN-␣/␤ on the multiplication of FLC (H-2d ) in the liver of adult C57Bl/6 (H-2b ) micea Treatment

Number of experiments

Number of mice in which FLC multiplied in the liver/total number of mice injected

None Control globulins Polyclonal antibody to IFN ␣/␤

22 7 23

1/107 0/36 100/102

a

Reprinted from [21].

equally effective in inhibiting the multiplication of IFN␣/␤ sensitive and resistant L1210 and FLC cells after i.p. inoculation.) In these experiments, we used hyperimmune anti-mouse IFN␣/␤ globulins from three sheep and one goat (produced in three laboratories). The control preparations consisted of the Ig from a sheep immunized with the “impurities” present in the partially purified IFN, and normal sheep and goat serum globulins. There seems little doubt, therefore, that antibody to mouse IFN␣/␤ was indeed responsible for the enhancement of transplantability of these tumors in mice. That the presence of low levels of endogenous type I IFN can modulate the phenotype of tumor cells in the mouse was shown by the following series of experiments. Friend erythroleukemia cells, both the IFN␣/␤ sensitive 745 and IFN resistant 3Cl8 clones, produce Friend virus in vitro. Serial i.p. passage of the IFN␣/␤ sensitive clone in syngeneic mice resulted in a clear-cut reduction of the virus producer phenotype of tumor cells, and after further in vivo passage no further virus was detected after cultivation of cells in vitro [19]. This virus nonproducer phenotype was a stable characteristic of clones derived from in vivo passaged IFN sensitive FLC. (This serial in vivo passage of the IFN␣/␤ sensitive line did not result in the selection of an IFN-resistant phenotype as the cells were still sensitive to the antiviral effects of IFN␣/␤ using vesicular stomatitis virus as test virus.) In contrast, in vivo passaged cloned cell lines of IFN␣/␤ resistant FLC cells did not result in any decrease in the capacity of these tumor cells to produce virus [19]. When IFN␣/␤ sensitive cells were passaged i.p. in mice treated with antibodies to IFN␣/␤, it was found that 96% of the tumor cell lines recovered continued to produce virus, indicating that endogenous IFN␣/␤ was indeed an important host factor for the in vivo selection of virus—non-producer tumor cell variants [19]. Similar results were obtained using

other retrovirus producer cells: the IFN␣/␤ sensitive and resistant clones of L1210 cells and IFN sensitive RBL-5 cells [19]. 2.3. Endogenous type I IFN is an essential factor in determining the resistance of allogeneic mice to FLC tumor metastases in liver and spleen When FLC (H-2d ) are injected intravenously (i.v.) into syngeneic DBA/2 mice, they multiply rapidly in the liver and spleen and kill all mice [20]. In contrast, FLC injected into adult C57Bl/6 (H-2b ) or C3H (H-2k ) mice lodged in the liver and spleen (as in syngeneic mice) but failed to multiply and were no longer recovered from these organs in the ensuing days [21]. When the same number of FLC were injected into C57Bl/6 mice treated with antibody to IFN␣/␤, the FLC multiplied rapidly in the liver and spleen (Table 2) as in syngeneic mice, resulting in death of most mice (Table 3). Immunoglobulins prepared from three different sheep immunized against different preparations of partially purified IFN␣/␤ and from one sheep immunized with highly purified IFN␣/␤ [22] were all effective in abrogating the resistance of C57Bl/6 mice to multiplication of FLC in the liver and spleen, whereas a potent monoclonal antibody against IFN␥ [23] proved ineffective in both C57Bl/6 and C3H mice [21]. Eight hours after i.v. inoculation of FLC, we found the presence of poly(A)+ RNA hybridizable with specific DNA probes for mouse IFN␣ and IFN␤, but not for IFN␥ in the livers of C57Bl/6 mice [21]. We therefore measured the serum IFN levels in C57Bl/6 mice shortly after i.v. inoculation of FLC. There was an inverse correlation between the presence of IFN in the serum and the capacity of FLC to multiply in the liver [24]. (Using specific monoclonal antibodies to IFN ␣, ␤, or ␥ we could show that this IFN was IFN ␣ [24].)

Table 3 Mortality in allogeneic mice treated with anti-IFN-␣/␤ globulin and injected intravenously with FLCa Mouse strain

Number of experiments

Treatment

Number of mice dead/total number injected

C57B1/6

6

Anti-IFN ␣/␤ Control Ig

13/15 0/25

C3H

1

Anti-IFN ␣/␤ Control Ig

2/3 0/7

a

Reprinted from [21].

Mean day of death ± S.E. 9.7 ± 0.3 – 11.5 ± 1.5 –

I. Gresser, F. Belardelli / Cytokine & Growth Factor Reviews 13 (2002) 111–118

Thus, all 44 FLC-injected adult C57Bl/6 mice had circulating IFN-␣ and FLC did not multiply in the liver of any of the mice [24]. When FLC-injected C57Bl/6 mice were X-ray-irradiated or treated with cyclophosphamide, IFN was no longer present in the serum, and FLC did multiply in the liver. Likewise, when FLC were injected into newborn C57Bl/6 mice, IFN was not present in the serum, and FLC multiplied in the liver [24]. (It should be noted that i.v. inoculation of syngeneic DBA/2 mice does not usually result in detectable serum levels of IFN, and FLC multiply exponentially in the liver.) We injected FLC i.v. into DBA/2 mice and C57Bl/10 mice congenic with DBA/2 mice (B10D2) and found that type 1 IFN was present in the serum of B10D2 mice (but not in DBA/2 mice) and FLC did not multiply in the liver of B10D2 (but did of course in the liver of the syngeneic DBA/2 mice). It is usually accepted that immune mechanisms are responsible for this tumor cell rejection across a strong MHC barrier [25]. The behavior of FLC injected i.v. into hybrid or congenic mice [21,24] suggested, however, that factors in addition to MHC may be important in restricting tumor growth as first pointed out by Snell and Stevens [26,27] and then extended by the Hellström [28,29]. Our results using either neutralizing antibody to IFN␣/␤ or means to block IFN␣/␤ production indicate that IFN␣/␤ is one of the essential factors in restricting FLC multiplication in the liver after i.v. inoculation of allogeneic mice.

3. Discussion 3.1. How do endogenous type I IFNs act to limit the growth of transplantable tumors in mice? In most of the above experiments, we used neutralizing antibody to type I IFN to show the importance of this endogenous IFN either constitutive or tumor-induced in restricting tumor growth in various experimental systems. It was therefore of great interest to us that in a series of recent experiments, Seif, Picaud and De Maeyer determined the resistance of a strain of type I IFN receptor KO mice backcrossed on C3H mice to several syngeneic and allogeneic tumors (article submitted for publication). They showed that subcutaneous inoculation of these mice with a syngeneic murine melanoma resulted in an increased incidence of mice with tumor, an increased growth of the tumor and an increased percentage of mice dying with tumor (compared to suitable control groups of mice). Similar results but with even more marked effects were observed in the KO mice using two allogeneic tumors injected subcutaneously. What is the source of the endogenous type I IFN? To the extent that we tested, the xenogeneic or murine tumor cells did not release mouse IFN spontaneously in vitro. With the exception of the experiments in which we injected allogeneic mice i.v. with FLC, we seldom detected the presence of IFN

115

in various body fluids or tissue extracts prior to or after tumor cell inoculation. Although we cannot rule out the possibility that the presence of a growing tumor induces type I IFN in the host, we favor the interpretation, consistent with our previous results [9,10] and those of others [30,31], that there are low levels of type I IFN produced constitutively that exert an antitumor effect in these experimental models. It is possible that type I IFN either constitutive or tumor-induced acts through the same host mechanisms operative when immunocompetent tumor inoculated mice were treated with type I IFN and clear-cut antitumor effects were observed [7,32]. If this assumption has some validity, we may briefly summarize some hypotheses as to the mechanisms by which type I IFN inhibits tumor growth. It has been known for many years that the antiviral activity of IFN is only one manifestation of the effect of IFN on cells [33,34]. IFN can inhibit normal and tumor cell division, induce cell surface alterations, enhance or inhibit specialized cellular functions [33]. Thus, IFNs are capable of interacting with virtually all cells of the body. Experimentally, the antitumor mechanisms of IFN seem to depend on many factors, among which the site of tumor growth appears to be of considerable importance. We have often chosen to use type I IFN resistant tumor cells (either L1210R [17], or the 3Cl8 line of FLC [18]) to enable us to separate the direct effects of IFN on the tumor cell from IFN-induced host mechanisms. For example, when IFN resistant FLC were injected i.p. into syngeneic mice, IFN treatment resulted in a very rapid elimination of the tumor cells [35]. The host’s immune system did not seem to be involved and the usual measures of decreasing the activity of host peritoneal cells, that might be expected to be involved in eliminating tumor cells such as macrophages or NK cells, failed to abrogate consistently the antitumor activity of IFN in the peritoneum [36]. Despite many years work in several laboratories, there has not been to our knowledge any convincing explanation as to how a few injections of IFN induces the host to eliminate syngeneic IFN resistant tumor cells so efficiently. We have wondered whether the IFN treated peritoneal epithelium exerted the antitumor effect directly on the tumor cells through contact, or by the liberation of soluble factors, but have never been able to substantiate these hypotheses [37]. When the tumor cells were injected s.c., IFN may have enhanced restrictive interactions between stroma and tumor. As mentioned above (Section 2), when nude mice were injected s.c. with xenogeneic tumor cells and treated with antibody to IFN␣/␤, the tumors were invasive and were not encapsulated. IFN has been shown to inhibit tumor angiogenesis [38], and we found that IFN induced a hemorrhagic necrosis of subcutaneous FLC tumors by acting on the endothelial cells of the fragile blood vessels that grow into the rapidly expanding tumor [39]. When FLC were injected i.v., IFN enhanced innate defense mechanisms such as macrophages and NK cells.

116

I. Gresser, F. Belardelli / Cytokine & Growth Factor Reviews 13 (2002) 111–118

When these cells were lacking as in beige mutant mice or their function depressed, IFN did not inhibit the development of liver metastases and the mice died [40]. If, however, the IFN treated mice survived initially, an intact immune system was essential to ensure full expression of IFN’s anti tumor effect [41,42]. Not only does type I IFN increase the humoral and cell mediated immune response [43], but it inhibits the generation of suppressor T cells [44], and favors the persistence of long term memory T cells [45]. Furthermore, we have shown that endogenous IFN is essential to the antitumor activity of immune CD4+ and CD8+ lymphocytes [46]. Thus, adaptive transfer of sensitized T lymphocytes inhibited the development of visceral FLC metastases and increased mouse survival time, but these immune T cells were totally ineffective when FLC injected mice were pretreated with antibody to IFN␣/␤ [46]. 3.2. Hypothesis: role of endogenous type I IFN in the defense systems of the mouse Bocci proposed that there was a “physiological interferon response” characterized by the production of IFN at “specific sites” which “can maintain active defense systems essential for survival” [30], and the use of antibody to type I IFN in our experiments has reinforced his hypothesis. Several examples: (1) the resistance of macrophages to viral multiplication is not an “inherent” property of these cells as is usually thought, but stems from their interaction in the mouse with an endogenous factor-namely IFN. When this endogenous IFN is neutralized by antibody, the macrophages become fully susceptible to viral multiplication [9,10]; (2) the response of mice immunized with chicken ␥ globulin is higher than in mice treated with antibody to type I IFN or in IFN I R KO mice [43]; (3) we have alluded above (Section 3.1) to the finding that immune T lymphocytes failed to exert an antitumor activity in mice treated with antibody to type I IFN [46]. 3.3. Hypothesis: essential role of type I IFN in allogeneic inhibition of tumor cells The Hellströms and Klein proposed that there were host mechanisms of tumor surveillance referred to as “allogeneic inhibition” “capable of eliminating neoplastic cell variants whose surface structure differs from that of the surrounding cells within a tissue without need for immunological reactions of the classic type” [47,48]. When xenogeneic tumor cells were injected s.c. into nude mice (Section 2) there was a clear-cut host reaction: the tumor cells were encapsulated and there was a host cellular infiltrate. When FLC were injected i.v. into allogeneic mice, they seeded the liver but failed to multiply and could no longer be recovered in the ensuing days (Section 2.2). When histocompatible (i.e. syngeneic) but antigenically different in vitro passaged FLC were first injected into the peritoneum

of syngeneic DBA 2 mice, they either failed to multiply or formed discrete tumor nodules on the peritoneal surface, but tumor ascites did not develop (Section 2.1). These are all experimental situations in which the tissues or organ (liver or spleen) is effective in restricting or eliminating relatively few “foreign” cells and can be considered as examples of “allogeneic inhibition”. By what mechanisms were these tumor cells restricted in their growth and eliminated? Hellström postulated “that target cell contact with foreign surface structures (such as H-2 antigens) can lead to their destruction” [47,49]. Without invoking mechanisms of active destruction of tumor cells, it may be that these tumor cells are unable to multiply and therefore survive in the unfavorable tissue environment of an incompatible host. To illustrate this point: when we cultivated murine L1210 tumor cells in the chemostat, we found that there was a minimum growth rate for these tumor cells, and growth rates below this minimum rate were not compatible with cell viability [33,50]. When IFN␣/␤ was added to the chemostat cultures, we found that cells died, not by a direct toxic effect of IFN per se, but secondary to the slowing of the cell growth rate [33]. We suggest that endogenous type I IFN can be perceived as an essential mediator of allogeneic inhibition. Administration of type I IFN enhances different mechanisms that may underlie allogeneic inhibition such as enhancement of the expression of histocompatibility antigens [51,52], key molecules in recognition of “foreigness”, and the activity of NK cells [53,54] which may constitute one of the effector mechanisms. Conversely, when endogenous type I IFN’s activity is neutralized by antibody, or IFN␣/␤ production is blocked (Section 2.2) or IFN is unable to bind to receptors as in IFN I R KO mice (Section 3.1), these particular mechanisms of host defense referred to collectively as “allogeneic inhibition” are no longer operative, and the tumor can grow, invade and even metastasize. We have previously emphasized [21,24] the similarity between the mechanisms responsible for allogeneic resistance when FLC were injected i.v. in our experiments [24] and the mechanisms underlying the resistance of X-irradiated F1 hybrid or allogeneic mice to parental bone marrow grafts. It is noteworthy that in both instances only treatment of mice with antibody to IFN␣/␤, but not antibody to IFN ␥, abrogated both allogeneic resistance to FLC multiplication in the liver [21] and the resistance of X-ray-irradiated F1 hybrid and allogeneic mice to bone marrow grafts [55]. The Hellströms speculated that allogeneic inhibition may have an “early phylogenetic origin in its role for growth surveillance” [47]. In this context too, IFNs seem to have early phylogenetic origins (even 300 million years ago [56]) and IFNs have been detected through-out the animal kingdom. The interactions [57,58] of IFNs with the phylogenetically recent development of the specialized immune system may derive from their partnership with more ancient mechanisms of defense evolved to maintain the integrity of the organism.

I. Gresser, F. Belardelli / Cytokine & Growth Factor Reviews 13 (2002) 111–118

3.4. Does endogenous type I IFN constitute a defense against autochthonous tumors in man? As in mice, biologically active IFNs have been detected only irregularly and in small amounts in human tissues in the absence of induction, although IFN ␣ mRNA was present in the spleen, kidney, liver and peripheral leucocytes of normal individuals [59]. Families have been described which lack the IFN ␥ receptor [60,61]. These individuals appear to be highly susceptible to mycobacterial infection, but to date no increase in the frequency of malignancies has been noted. We are not aware of any reports of individuals lacking the genes for type I IFN receptor (similar to KO mice lacking the type I IFN receptor [62]). This finding might suggest that type I IFNs are essential molecules for survival. In this regard, the existence of a surprising number of IFN-␣ subtypes with comparable biological activities (the redundancy of IFN-␣ genes) may reflect the advantages of providing the host with alternative defense pathways. As so much experimental oncology is based on the use of animal tumor models, it is always difficult to determine to what extent experimental results can be extrapolated to the development of autochthonous tumors in man. Nevertheless, the striking importance of endogenous type I IFN in the defense of mice against experimental tumors suggests to us that the presence of endogenous type 1 IFN produced constitutively (or tumor-induced) also constitutes an important defense against tumors in man.

Acknowledgements One of us (I.G.) gratefully acknowledges the help of Mme. C. Maury, M.-T. Bandu, F. Vignaux, F. Zambetti and Dr. M. Tovey in the realization of these experiments. One of us (F.B.) is grateful to the Italian Association for Cancer Research (AIRC) for financial support. We are grateful to Drs. J. Vilcek, K.E. Hellstrom and M. Ferrantini for their critical reading of this article and to Drs. I. Seif, S. Pacaud and E. De Maeyer for permitting us to quote their unpublished work on KO mice (Section 3.1).

References [1] Burnet FM. The concept of immunological surveillance. Progr Exp Tumor Res 1970;13:1–27. [2] Thomas L. Discussion. In: Lawrence HS, editor. Cellular and humoral aspects of the hypersensitive states, London: Cassell, 1959. p. 529–32. [3] Smyth MJ, Godfrey DI, Trapani JA. A fresh look at tumor immunosurveillance and immunotherapy. Nature Immunol 2001;2: 293–9. [4] Penn I. Tumors of the immunocompromised patient. Ann Rev Med 1988;39:63–71. [5] Stutman O. Immunodepression and malignancy. Adv Cancer Res 1975;22:261–422.

117

[6] Gresser I, Bourali C, Lévy J-P, Fontaine-Brouty-Boyé D, Thomas MT. Increased survival in mice inoculated with tumor cells and treated with interferon preparations. Proc Natl Acad Sci USA 1969;63:51–7. [7] Gresser I. How does interferon inhibit tumor growth? Phil Trans R Soc London B 1982;299:69–76. [8] Strander H. Interferon treatment of human neoplasia. Adv Cancer Res 1986;46:1–265. [9] Belardelli F, Vignaux F, Proietti E, Gresser I. Injection of mice with antibody to interferon renders peritoneal macrophages permissive for vesicular stomatitis virus and encephalomyocarditis virus. Proc Natl Acad Sci USA 1984;81:602–6. [10] Gresser I, Vignaux F, Belardelli F, Tovey MG, Maunoury M-T. Injection of mice with antibody to mouse interferon ␣/␤ decreases the level of 2-5 -oligo-adenylate synthetase in peritoneal macrophages. J Virol 1985;53:221–7. [11] Gresser I, Tovey MG, Bandu M-T, Maury C, Brouty-Boyé D. Role of interferon in the pathogenesis of virus diseases in mice as demonstrated by the use of anti-interferon serum. I. Rapid evolution of encephalomyocarditis virus infection. J Exp Med 1976;144:1305– 15. [12] Gresser I, Tovey MG, Maury C, Bandu M-T. Role of interferon in the pathogenesis of virus diseases in mice as demonstrated by the use of anti-interferon serum. II. Studies with herpes simplex, moloney sarcoma, vesicular stomatitis, newcastle disease, and influenza viruses. J Exp Med 1976;144:1316–23. [13] Reid LM, Minato N, Gresser I, Holland J, Kadish A, Bloom BR. Influence of anti-mouse interferon serum on the growth and metastasis of tumor cells persistently infected with virus and of human prostatic tumors in athymic nude mice. Proc Natl Acad Sci USA 1981;78:1171–5. [14] Shouval D, Rager-Zisman B, Quan P, Shafritz DA, Bloom BR, Reid LM. Role in nude mice of interferon and natural killer cells in inhibiting the tumorigenicity of human hepatocellular carcinoma cells infected with hepatitis B virus. J Clin Invest 1983;72:707–17. [15] Puddu P, Locardi C, Sestili P, Varano F, Petrini C, Modesti A, Masuelli L, Gresser I, Belardelli F. HIV-infected tumor xenografts as an vivo model for antiviral therapy. Role of interferon ␣/␤ in the restriction of tumor growth in nude mice injected with HIV-infected U937 tumor cells. J Virol 1991;65:2245–53. [16] Gresser I, Belardelli F, Maury C, Maunoury M-T, Tovey MG. Injection of mice with antibody to interferon enhances the growth of transplantable murine tumors. J Exp Med 1983;158:2095–107. [17] Gresser I, Bandu M-T, Brouty-Boyé D. Interferon and cell division. IX. Interferon-resistant L1210 cells: characteristics and origin. J Natl Cancer Inst 1974;52:553–9. [18] Affabris E, Jemma C, Rossi GB. Isolation of interferon-resistant variants of Friend erythroleukemia cells: effect of interferon and ouabain. Virology 1982;120:441–52. [19] Ferrantini M, Belardelli F, Locardi C. Role of endogenous ␣/␤ interferon in the selection of virus nonproducer Friend leukemia cells after serial intraperitoneal passages in syngeneic mice. J Virol 1988;62:600–5. [20] Gresser I, Maury C, Woodrow D, Moss J, Grütter MG, Vignaux F, Belardelli F, Maunoury M-T. Interferon treatment markedly inhibits the development of tumor metastases in the liver and spleen and increases survival time of mice after intravenous inoculation of Friend erythroleukemia cells. Int J Cancer 1988;41:135–42. [21] Gresser I, Maury C, Vignaux F, Haller O, Belardelli F, Tovey MG. Antibody to mouse interferon ␣/␤ abrogates resistance to the multiplication of Friend erythroleukemia cells in the livers of allogeneic mice. J Exp Med 1988;168:1271–91. [22] De Maeyer-Guignard J, Tovey MG, Gresser I, De Maeyer E. Purification of mouse interferon by sequential affinity chromatography on poly(U)- and antibody-agarose columns. Nature (London) 1978;271:622–5. [23] Spitalny GL, Havell EA. Monoclonal antibody to murine ␥ interferon inhibits lymphokine-induced antiviral and macrophage tumoricidal activities. J Exp Med 1984;159:1560–5.

118

I. Gresser, F. Belardelli / Cytokine & Growth Factor Reviews 13 (2002) 111–118

[24] Gresser I, Maury C, Bandu M-T, Belardelli F. Importance of interferon in the resistance of allogeneic C57Bl/6 mice to the multiplication of Friend Erythroleukemia cells in the liver. Int J Cancer 1990;45:364–71. [25] Bach FH, Sachs DH. Current concepts: immunology. Transplantation immunology. New Engl J Med 1987;317:489–92. [26] Snell GD. Histocompatibility genes in the mouse. II. Production and analysis of isogenic resistant lines. J Natl Cancer Inst 1958;21:843– 77. [27] Snell GD, Stevens LC. Histocompatibility genes of mice. III. H-1 and H-4, two histocompatibility loci in the first linkage group. Immunology 1961;4:366–79. [28] Hellström KE. Differential behavior of transplanted mouse lymphoma lines in genetically compatible homozygous and F1 hybrid mice. Nature 1963;199:614–5. [29] Hellström KE, Hellström I. In: Palm J, editor, Isoantigens and cell interactions, Philadelphia (PA): Wistar Inst. Press, 1965. p. 79– 90. [30] Bocci V. Production and role of interferon in physiological conditions. Biol Rev 1981;56:49–85. [31] Bocci V. The physiological interferon response. Immunol Today 1985;6:7–9. [32] Gresser I. Antitumor effects of interferons past, present and future. Br J Haematol 1991;79:1–5. [33] Gresser I, Tovey MG. Antitumor effects of interferon. Biochim Biosphys Acta 1978;516:231–47. [34] Gresser I. Wherefore IFN? J Leukocyte Biol 1997;61:567–74. [35] Belardelli F, Gresser I, Maury C, Maunoury M-T. Antitumor effects of interferon in mice injected with interferon-sensitive and interferon-resistant Friend leukemia cells. Int J Cancer 1982;30:813– 20. [36] Belardelli F, Gresser I, Maury C, Maunoury M-T. Antitumor effects of interferon in mice injected with interferon-sensitive and interferon-resistant Friend leukemia cells. II. Role of host mechanisms. Int J Cancer 1982;31:821–5. [37] Gresser I. How does interferon inhibit tumor growth? In: Gresser I, editor. Interferon, vol. 6. New York/London: Academic Press, 1985. p. 93–126. [38] Sidky YA, Borden EC. Inhibition of angiogenesis by interferons: effects on tumor and lymphocyte-induced vascular responses. Cancer Res 1987;47:5155–61. [39] Dvorak H, Gresser I. Microvascular injury in the pathogenesis of interferon-induced necrosis of subcutaneous tumors in mice. J Natl Cancer Inst 1989;81:497–502. [40] Gresser I, Kaido T, Maury C, Woodrow D, Moss J, Belardelli F. Interaction of IFN ␣/␤ with host cells essential to the early inhibition of Friend erythroleukemia visceral metastases in mice. Int J Cancer 1994;57:604–11. [41] Gresser I, Maury C, Carnaud C, De Maeyer E, Maunoury M-T, Belardelli F. Anti-tumor effects of interferon in mice injected with interferon-sensitive and interferon-resistant Friend erythroleukemia cells. VIII. Role of the immune system in the inhibition of visceral metastases. Int J Cancer 1990;46:468–74. [42] Gresser I, Carnaud C, Maury C, Sala A, Eid P, Woodrow D, Maunoury M-T, Belardelli F. Host humoral and cellular immune mechanisms in the continued suppression of Friend erythroleukemia metastases after interferon ␣/␤ treatment in mice. J Exp Med 1991;173:1193–203. [43] Le Bon A, Schiavoni G, D’Agostino G, Gresser I, Belardelli F, Tough DF. Type I interferons potently enhance humoral immunity

[44]

[45]

[46]

[47] [48] [49] [50]

[51]

[52]

[53]

[54]

[55]

[56]

[57]

[58] [59]

[60]

[61]

[62]

and can promote isotype switching by stimulating dendritic cells in vivo. Immunity 2001;14:461–70. Fradelizi D, Gresser I. Interferon inhibits the generation of allospecific suppressor T lymphocytes. J Exp Med 1982;155:1610– 22. Tough DF, Borrow P, Sprent J. Induction of bystander T cell proliferation by viruses and type I interferon in vivo. Science 1996;272:1947–50. Gresser I, Maury C, Kaido T, Bandu M-T, Tovey MG, Maunoury M-T, Fantuzzi L, Gessani S, Greco G, Belardelli F. The essential role of endogenous IFN ␣/␤ in the antimetastatic action of sensitized T lymphocytes in mice injected with Friend erythroleukemia cells. Int J Cancer 1995;63:726–31. Hellström KE, Hellström I. Cellular immunity against tumor antigens. Adv Cancer Res 1969;12:167–223. Hellström KE, Hellström I. Allogeneic inhibition of transplanted tumor cells. Progr Exp Tumor Res 1967;9:40–76. Hellstrom KE, Hellstrom I, Haughton G. Demonstration of syngeneic preference in vitro. Nature 1964;204:661–4. Tovey M, Brouty Boyé D, Gresser I. Early effect of interferon on mouse leukemia cells cultivated in a chemostat. Proc Natl Acad Sci USA 1973;70:2785–8. Lindahl P, Leary P, Gresser I. Enhancement by interferon of the expression of surface antigens on murine leukemia L120 cells. Proc Natl Acad Sci USA 1973;70:2785–8. Lindahl P, Leary P, Gresser I. Enhancement by interferon of the expression of histocompatibility antigens of mouse lymphoid cells in vitro. Eur J Immunol 1974;4:779–84. Gidlund M, Örn A, Wigzell H, Senik A, Gresser I. Enhanced NK cell activity in mice injected with interferon and interferon inducers. Nature 1978;273:759–61. Djeu JY, Heinbaugh JA, Holden HT, Herberman RB. Augmentation of mouse natural killer cell activity by interferon and interferon inducers. J Immunol 1979;122:175–81. Afifi MS, Kumar V, Bennett M. Stimulation of genetic resistance to marrow grafts in mice by interferon ␣/␤. J Immunol 1985;134:3739– 45. Miyata T, Hayashida H, Kikuno R, Toh H, Kawade Y. In: Gresser I, editor. Evolution of interferon genes in Interferon 6. London: Academic Press, 1985. p. 1–30. Belardelli F, Gresser I. The neglected role of type I interferon in the T cell response: implications for its clinical use. Immunol Today 1996;17:369–72. Belardelli F. Role of interferons and other cytokines in the regulation of the immune response. APMIS 1995;103:161–79. Tovey MG, Streuli M, Gresser I, Gugenheim J, Blanchard B, Guymarho J, Vignaux F, Gigou M. Interferon messenger RNA is produced constitutively in the organs of normal individuals. Proc Natl Acad Sci USA 1987;84:5038–42. Newport MJ, Huxley C, Huston S, Hawrylowicz CM, Oostra BA, Williamson R, Levin M. A mutation in the interferon-␥-receptor gene and susceptibility to mycobacterial infection. New Engl J Med 1996;335:1941–9. Jouanguy E, Altare F, Lamhamedi S, Revy P, Emile JF, Newport M, Levin M, Blanche S, Seboun E, Fischer A, Casanova JL. Interferon-␥-receptor deficiency in an infant with fatal bacille Calmette–Guerin infection. New Engl J Med 1996;335:1956–61. Müller U, Steinhoff U, Reis LF, Hemmi S, Pavlovic J, Zinkernagel RM, Aguet M. Functional role of type I and type II interferons in antiviral defense. Science 1994;264:1918–21.