Interferon-α as Antiviral and Antitumor Vaccine Adjuvants: Mechanisms of Action and Response Signature

Introduction

I nterferons (IFNs)-α are cytokines exhibiting multiple biological activities, including effects on cells of the immune system (Vilcek 2006). IFN-α have a long record of clinical use and are still currently utilized in medical practice in some patients with certain types of cancer and infectious diseases. For a long time, the importance of the effects of IFN-α on cells of the immune system was neglected in spite of early studies showing the crucial role of the immune system in the antitumor response elicited by the cytokine in mice (Belardelli and Gresser 1996). Studies published over the last 2 decades have shown that IFN-α can effectively promote the differentiation and the activity of certain subsets of T and B lymphocytes and monocyte-derived dendritic cells (DC) in both mouse and human models, supporting the concept that IFN-α can act as important factors in linking innate and adaptive immunity (reviewed by Ferrantini and others 2007; Rizza and others 2010). In particular, an ensemble of data stemming from early and recent studies conducted on both mouse and human models suggest that IFN-α can act as effective vaccine adjuvants for inducing protective antiviral and antitumor immunity.

The potential effects of IFN as vaccine adjuvants have been the specific topic of some review articles since the last few years (Tovey and others 2006, 2008; Toporovski and others 2010; Rizza and others 2011). In the first part of this review, we summarize some studies conducted on mouse models with a special focus on the data recently obtained by our laboratory, including results of studies performed in a gammaherpesvirus model of infection, such as mice infected with murine herpesvirus 68 (MHV-68). We also review the latest studies from different groups, including ours, on the effects of IFN-α on the differentiation and activity of human DC, as well as the results of the few clinical trials based on the use of IFN-α as adjuvant, when the cytokine was injected at the same time and site of a reference vaccine. Lastly, we thoroughly review the ensemble of studies published in the last decade that describe the IFN-α molecular signature detected in cells derived from IFN-treated patients under different modalities and clinical settings, as these data can be instrumental in understanding the molecular mechanisms of the immune adjuvant activity of these cytokines and may provide important insights into the design of the novel strategies of the clinical use of IFN-α in patients.

Importance of Mouse Models for Understanding the Immune Adjuvant Activity of IFN-α and Their Mechanisms of Action

Mouse models have often been used to address the immune adjuvant role of IFN-α in several different in vivo settings. A first important step in understanding the importance of immune host-mediated mechanisms in the response to type I IFN (IFN-I) is represented by an ensemble of studies where IFN was administered in mice transplanted with syngeneic IFN-resistant tumor cells (reviewed by Belardelli and Gresser 1996; Ferrantini and others 2007). Thus, in early studies performed by one of us (F.B.) together with Ion Gresser in Paris, we showed that host-mediated mechanisms played a major role in the antitumor response observed in mice transplanted with IFN-resistant Friend leukemia cells (FLC) (reviewed by Belardelli and Gresser 1996; Ferrantini and others 2007). Of note, we could show the crucial importance of T cells and antibodies (especially IgG2a) in the IFN-mediated suppression of metastatic FLC growth (Gresser and others 1991). Subsequently, several studies carried out by our group as well as by others demonstrated that the transduction of the IFN-α1 gene into several unrelated types of metastatic tumor cells resulted in tumor rejection and host-mediated antitumor responses (reviewed by Ferrantini and Belardelli 2000). Notably, the tumor rejection occurred only in immunocompetent mice and was not observed when immonodeficient SCID mice were transplanted with IFN-producing tumor cells. The tumor rejection observed in immunocompetent syngenic animals resulted in the protection from the subsequent controlateral injection of parental cells, suggesting that a tumor-specific antitumor immunity was induced when immunocompetent mice were injected with IFN-α-producing cells (Ferrantini and others 1994; Ferrantini and Belardelli 2000). In particular, in mice transplanted with IFN-α-producing TSA tumor cells, we could also demonstrate that CD8⁺ T cells were the major effector cells responsible for the rejection of the metastatic tumor cells as a result of the local cytokine production (Ferrantini and others 1994). Notably, the injection of mice with IFN-α-producing cells (but not with control tumor cells producing unrelated cytokines) resulted in the in vivo proliferation and long-term survival of CD8⁺ T cells exhibiting the memory phenotype (Belardelli and others 1998). Thus, the set of these studies in mice transplanted with genetically modified tumor cells producing IFN represented an important step in understanding the immune adjuvant activity of a specific IFN-α subtype in inducing a protective antitumor response. Moreover, these studies strongly suggested that IFN-α could be used as an effective adjuvant for the development of cancer vaccines (Ferrantini and Belardelli 2000; Ferrantini and others 2007).

In the subsequent years, several groups, including ours, provided further evidence on the immune adjuvant activity of IFN-I using mouse models not involving the transplantation of tumor cells, and some important mechanisms have been identified. In 1999, Gallucci and co-authors showed that IFN-α could be considered among the “endogenous activating substances” capable of acting as natural adjuvants and stimulating a primary immune response when administered to naïve mice together with a soluble protein antigen (Gallucci and others 1999). A few years later, the work by David Tough's group together with our group in Rome demonstrated the efficacy of IFN-I used as a vaccine adjuvant in enhancing both T-cell and antibody responses and promoting immunological memory against a soluble protein such as chicken gamma globulin (CGG) (Le Bon and others 2001). Such a basic in vivo setting was also used to provide a direct evidence of the link between IFN-I and the initiation of immune responses by DC, as the antibody response enhancement could also be observed when only DC were able to respond to IFN-I (Le Bon and others 2001).

The availability of mice lacking a functional IFN-α/β receptor (Müller and others 1994), thus being genetically unresponsive to the effects of the cytokine, offered the possibility of demonstrating the adjuvant role of IFN-α in a more complex in vivo setting. A work comparing the effects of several Th1 adjuvant compounds in wt versus IFNARI−/− mice proved that endogenous IFN-I was essential for the promotion of both the IgG2α antibody and T-cell responses by a variety of substances, such as poly I:C and CpG (Proietti and others 2002). Moreover, IFN-I turned out to be a powerful adjuvant when administered in association with the human influenza vaccine, inducing a strong IgG2α and IgA production and conferring protection from virus challenge (Proietti and others 2002).

Recent Studies in Mouse Models on the Immune Adjuvant Activity of IFN-α in Inducing a Protective Antiviral (MHV-68 Model) and Antitumor Response

Over the last decade, several additional studies on the immune adjuvant activity of IFN-I in inducing a protective antiviral and antitumor response in mouse models have been reported. We will mention here only some studies recently performed by our group as well as by other laboratories not previously reviewed or poorly considered in recent review articles (Tovey and others 2008; Rizza and others 2010, 2011).

The adjuvant role of IFN-I, either endogenously produced in response to viral infection or administered as a vaccine adjuvant in prophylactic vaccination settings, has been investigated in the viral model of mice infected with the MHV-68. The similarities between the genetic and pathogenic features of MHV-68 and human gammaherpesviruses (Blaskovic and others 1980) render the MHV-68 model not only an amenable small animal model that studies gammaherpesvirus biology in their natural host, but also a versatile, highly defined experimental setting for exploring the immune response against gammaherpesviruses and defining the requirements for effective control against their infection (Aricò and others 2011b, and references cited therein). In our study (Aricò and others 2002), we showed that the virus-specific humoral response induced by a heat-inactivated MHV-68 vaccine was significantly enhanced when IFN-I was co-administered as a vaccine adjuvant with the same treatment schedule proved to be effective in a previous study based on the use of the reference CGG antigen (Le Bon and others 2001). We observed an IFN-mediated shift toward a Th1 type of immune response, as the antibody enhancement was associated with a more evident increase in the serum levels of anti-MHV-68 IgG1 and IgG2b in mice receiving the cytokine with regard to the mice vaccinated with the inactivated virus alone, and with a significant induction of IgG2α observed only in mice receiving IFN-I together with the vaccine (Aricò and others 2002).

The mice infected with MHV-68 represent a useful in vivo model that explores the role of endogenous IFN-I in the natural response against viral infection. In fact, as expected on the basis of its antiviral activity, endogenous IFN-I turned out to play a significant role in limiting MHV-68 acute replication at early stages of infection and before the onset of the adaptive immune response (Dutia and others 1999). Moreover, it has been demonstrated that IFN-I also counteracts MHV-68 latency, by contributing to the development and maintenance of an anti-MHV-68 innate immune response and by controlling virus reactivation from latency (Liang and others 2004). Notably, similar to many herpesviruses, MHV-68 can evolve several immune evasion strategies that neutralize IFN-I-mediated control of acute and latent infection (Aricò and others 2011b, and references cited therein).

In a more comprehensive study aimed both at investigating the role of IFN-I in the natural course of MHV-68 infection and at exploiting its adjuvant activity to develop an effective vaccination against gammaherpesviruses, we recently described the effects of the continuous production of IFN-α1 on the course of infection with a clone of MHV-68 genetically modified by the insertion of the mouse IFN-α1 coding gene (MHV-68mIFN-α1) (Aricò and others 2011b). Of interest, the high amount of the cytokine released in the supernatants of cells infected with the recombinant virus did not have any inhibitory effect on the in vitro replication of MHV-68mIFN-α1, which was undistinguishable from that exhibited by wtMHV-68. Nevertheless, a significant attenuation of the pathological features of MHV-68mIFN-α1 could be clearly observed in vivo, as mice carrying a functional IFN-I system experienced a dramatic reduction in all the parameters of acute virus infection with regard to IFNARI−/− animals. Taken together, these results strongly suggest that the in vivo attenuation of MHV-68mIFN-α1 was not caused by a direct antiviral effect elicited by the cytokine produced during virus replication in vivo, but was rather related to the immunomodulatory effects exerted by IFN-α1 and to the potent the antiviral immune response primed in the presence of the cytokine. We, thus, suggested that the local production of IFN-α1 at the same site of virus replication and antigen release can subvert MHV-68-antiviral immune evasion mechanisms, and compensate for the virus-mediated inhibition of DC differentiation (Weslow-Schmidt and others 2007), thus possibly triggering a more efficient DC activation and a protective antiviral immune response. This scenario may also explain the efficacy of MHV-68mIFN-α1-based vaccines, administered as either live attenuated or partially inactivated viruses, in protecting naïve mice against the acute and latent phase of MHV-68 infection (Aricò and others 2011b). The fact that the inactivated MHV-68mIFN-α1 vaccine was more effective in counteracting the establishment of long-term latency by lower than higher doses of challenging virus, together with the dramatic reduction of the acute infection parameters in MHV-68mIFN-α1 vaccinated mice, suggests that these vaccines may, at least in part, act through the block of the primary infection, by priming the immune system to effectively recognize free infecting virions as well as infected cells expressing lytic viral antigens. Interestingly, in spite of the dramatic reduction of the MHV-68 virus replication and antigen release after virus challenge, mice vaccinated with MHV-68mIFN-α1 reached levels of IgG2α and IgG2b, similar to the ones induced by the massive virus replication in unvaccinated control mice, thus also confirming in this setting the IFN-α-mediated enhancement of the anti-viral Th1-like antibody response. It is worth remarking that our studies showed that the adjuvant activity of IFN-α was much more protective when the cytokine was associated to vaccines based on the viruses inactivated with a procedure preserving the viral proteins antigenicity, such as psoralen-mediated UV crosslinking (Aricò and others 2011b), rather than with heat-inactivated viral particles (Aricò and others 2002), thus once again confirming the immunostimulatory role exerted by this cytokine when used as a vaccine adjuvant.

Some early and recent studies conducted on mouse tumor models strongly support the importance of endogenous IFN-1 in restricting tumor growth by means of host-mediated mechanisms (Gresser and others 1983; Picaud and others 2002; Dunn and others 2005; Aricò and others, manuscript in preparation). In 1983, Gresser and co-workers had shown that the injection of mice with a potent anti-IFN-I antibody markedly enhanced tumor growth in syngeneic animals transplanted with different types of tumor cells, including IFN-resistant FLC, suggesting that endogenous IFN played an important role in restricting tumor growth by host-mediated mechanisms. In subsequent years, the importance of endogenous IFN-I in the host-mediated control of tumor growth was demonstrated by a set of studies in mice knock-out for the IFN-I receptor (Picaud and others 2002; Dunn and others 2005). Recently, in a mouse model of spontaneous carcinogenesis, we observed that the mice lacking a functional IFN-I system experienced an earlier and more aggressive development of mammary tumors, which was associated with an increased vascularization of tumor lesions when compared with immunocompetent mice. These observations corroborate the importance of IFN-I as an anti-angiogenic factor and further support the role of these cytokines in the immune surveillance of spontaneous tumors (Aricò and others, manuscript in preparation).

The concept stemming from our early in vivo studies with IFN-producing cells (reviewed by Ferrantini and Belardelli 2000) has led to additional reports supporting the role of IFN-α as an effective adjuvant of cancer vaccines (reviewed by Ferrantini and others 2007), which has been lately confirmed in additional tumor models. As an example, Sikora and co-workers recently demonstrated that IFN-α strongly enhances vaccine-induced CD8⁺ T cell numbers, effector functions, and antitumor activity in mice transplanted with metastatic B16 melanoma cells (Sikora and others 2009). Likewise, Hance and colleagues (2009) recently reported a strong immunoadjuvant effect of IFN-α in combination with recombinant poxvirus vaccines in mice with CEA⁺ adenocarcinomas. However, the overall information stemming from mouse tumor models has been poorly translated so far into clinical studies to evaluate the possible efficacy of IFN as an adjuvant for cancer vaccines in humans. In a subsequent section of this article, we review the few clinical trials performed by our group as well as by others aimed at evaluating the possible immune adjuvant activity of IFN-α in humans.

Interactions Between IFN-α and DC and Their Role in the Generation of a Protective Immunity

Over the last decade, an ensemble of studies conducted on both mouse and human models have provided convincing evidence on the role of IFN-α in inducing the differentiation and activation of DC and its possible involvement in antiviral and antitumor immunity (reviewed by Rizza and others 2010, 2011). We will first mention here some recent studies conducted on mouse models that provide relevant information on the importance of IFN-I-DC interactions and then summarize the overall effects of IFN-α on the differentiation/activation of human DC, discussing their relevance for the induction of antitumor immunity and possible clinical use.

Recently, Lorenzi and co-workers (2011) have shown that IFN-I controls antigen retention and survival of mouse CD8a(+) DC after the uptake of tumor apoptotic cells, leading to cross-priming. Notably, 2 recent studies performed on different mouse tumor models have provided, for the first time, convincing direct evidence on the crucial importance of the interactions between type I IFN and DC in the generation of a protective antitumor immune response (Diamond and others 2011; Fuertes and others 2011). In particular, Diamond and colleagues (2011) have demonstrated that mice lacking the IFN-I receptor in DC are unable to reject highly immunogenic tumor cells. Likewise, by using mice lacking the IFN-I receptor or Stat1, Fuertes and colleagues (2011) have shown that IFN signaling is required in the hematopoietic compartment at the level of host antigen-presenting cells and selectively for intratumoral accumulation of CD8α+ DC.

Some groups, including our laboratory, have shown that IFN-α markedly promotes the differentiation and activation of human DC (Paquette and others 1998; Santini and others 2000; for a comprehensive review, see Santini and others 2009, and references quoted therein). In particular, we demonstrated that the addition of IFN-α to granulocyte-macrophage colony-stimulating factor (GM-CSF)-treated human monocytes resulted in the rapid differentiation of these cells into highly active DC (Santini and others 2000). The DC generated after 3 days of treatment with IFN-α showed the phenotype of partially mature DC, as revealed by the expression of DC activation markers and migratory responses (Santini and others 2000; Parlato and others 2001). Notably, these DC, designed as IFN-DC, were endowed with potent functional activities, not only in vitro but also in vivo, as evaluated by their strong capability to induce the generation of a primary human antibody response and CTL expansion both in vitro and in vivo, after injection into SCID mice reconstituted with human peripheral blood lymphocytes (Santini and others 2000; Santodonato and others 2003; Lapenta and others 2003). IFN-DC proved to be superior with regard to the reference monocyte-derived DC generated in the presence of IL-4 and GM-CSF in inducing a CTL response and the in vivo cross-priming of CD8⁺ T cells against exogenous viral antigens (Lapenta and others 2003, 2006). Notably, IFN-DC were also endowed with direct cytotoxic activity against tumor cells (Santini and others 2000). Of interest, the in vitro experiments performed with IFN-DC and apoptotic cells showed that IFN-DC can take up apoptotic cell-derived antigens and cross-present them to CD8⁺ T cells, thus leading to an efficient cross-priming of these cells against tumor-associated antigens (Parlato and others 2010). Notably, using this 3-day IFN-DC system, Lattanzi and colleagues (2011) have recently shown that the IFN-α boosted epitope cross-presentation by DC via the modulation of proteasome activity, while Spadaro and co-workers (2012) have recently demonstrated that the same cytokine markedly enhances cross-presentation in human DC by modulating antigen survival, endocytic routing, and processing. In addition, Santini and co-workers (2011) have reported that IFN-DC combined a Th1-orienting attitude with the induction of autologous Th17 responses, demonstrating a role of IL-23 and IL-12 in such responses. Since Th17 responses have been correlated with autoimmunity, these recent results, together with previous data by J. Banchereau's group describing the IFN-DC phenotype of monocytes cultured in the presence of serum from patients with systemic lupus erythematosus (SLE) (Blanco and others 2001), further support the hypothesis of a role of IFN-α and IFN-DC interactions in the pathogenesis of certain human autoimmune diseases (reviewed by Banchereau and Pascual 2006; Rizza and others 2010).

Of interest, a DC population strongly resembling IFN-DC has recently been identified as a typical cell infiltrate during the spontaneous regression of highly immunogenic Molluscum contagiosum virus-induced skin lesions (Vermi and others 2010), supporting the concept that IFN-DC can indeed represent naturally occurring DC promptly generated in vivo during the response to IFN-I induced by viruses and other natural danger signals. Likewise, cells with a phenotype similar to IFN-DC have recently been identified in patients with some autoimmune diseases (reviewed by Farkas and Kemény 2011), further supporting a possible role of these cells in both antiviral/antitumor responses and autoimmunity (Rizza and others 2010).

Over the last decade, DC have been extensively considered as valuable cellular adjuvants for the development of therapeutic cancer vaccines (Steinman and Banchereau 2007). In the light of all the data just discussed, it is possible to speculate that (i) IFN-DC can play a relevant role in the antitumor/antiviral immune response observed in cytokine-treated patients as well as in subjects undergoing certain autoimmune responses; (ii) in vitro cultured IFN-α-conditioned DC can exhibit some advantages for the development of more effective DC-based strategies of cancer immunotherapy.

IFN-α as Potential Antiviral and Antitumor Vaccine Adjuvants in Humans

Some early and recent clinical trials have attempted to evaluate the potential efficacy of IFN-α as an adjuvant of viral vaccines in humans. Most of these studies have recently been reviewed elsewhere (Tovey and others 2008; Toporovski and others 2010; Rizza and others 2011). We will restrict our focus here to some results recently obtained by our group in both immunocompromised and healthy individuals subjected to hepatitis B virus (HBV) vaccination, in collaboration with other groups in Italy (Rizza and others 2008) and with the group of C. Garcia-Monzon in Madrid (Miquilena-Colina and others 2009), as in both these studies the modalities of IFN and vaccine administration were similar to those utilized in our previous IFN/vaccination studies in mouse models (Proietti and others 2002). In fact, the major aim of both clinical trials was to evaluate whether the results obtained by our group in mouse models showing the potent vaccine adjuvant activity of IFN-I in an influenza infection system (Proietti and others 2002) could be translated into proof-of-concept in humans.

In a clinical study performed by using natural IFN-α as an adjuvant for HBV vaccine in healthy individuals, no enhancement of the antibody response was observed in IFN-treated patients, at the dose of both 1 and 3 MU of IFN given intramuscularly (i.m.) at the site of the vaccine injection, in spite of a significant increase in biological and molecular IFN-induced markers observed on CD14⁺ monocytes at an early time after the IFN injection (Rizza and others 2008). One possible explanation for the lack of vaccine adjuvant activity observed in this study may consist of the use of natural leukocyte IFN-α (a partially purified mixture of different IFN subtypes), which could exert different effects and mask a possible IFN-α stimulatory component because of the concomitant presence of other cytokines/factors endowed with some immunosuppressive activity.

In a recent clinical trial conducted on hemodialysis (HD) patients, recombinant IFN-α-2b has proved beneficial in enhancing the immune response to HBV vaccination (Miquilena-Colina and others 2009). HD patients suffer from a generalized impairment of the immune function, largely due to a decreased co-stimulatory activity of DC, leading to T-cell dysfunctions. Since they are patients at a high risk of HBV infection, measures for improving their immune response to the HBV vaccine are highly needed. Recently, Miquilena-Colina and co-workers (2009) found a significantly higher and earlier seroprotection rate in HD patients undergoing a vaccination cycle by adding IFN-α, at the dose of 1 and 3 MU, to each vaccine injection. Furthermore, patients showing earlier seroprotective antibody titers also showed a qualitatively better immune response, in terms of IgG1 isotype switching and restored altered Th1/Th2 cell balance. We may assume that, in HD patients, IFN-α can act on an immune system that is somehow defective, thus providing an essential help for inducing an otherwise ineffective immune response. We will not review here some other important clinical studies aimed at exploring the immune adjuvant activity of IFN when the cytokine was administered by modalities different from the i.m. injection in close proximity of the viral vaccine (for instance, sublingual or intranasal administration), as these reports have been mentioned in a previous review (Rizza and others 2011), while we will briefly discuss some critical issues in considering IFN-α as an adjuvant of human cancer vaccines.

While the use of potent immune adjuvants is strictly required for the development of cancer vaccines because of the need of overcoming the cancer-induced immune suppression environment, IFN-α has rarely been considered a potential immune adjuvant and analyzed as such in specifically designed clinical studies. In fact, much more than for the IFN-α potential capability to stimulate a protective immune response against cancer cells (Belardelli and Gresser 1996; Ferrantini and others 2007), the use of these cytokines in clinical oncology has been mostly based on their cytostatic activity, with the dosage and timing of administration different from those usually utilized for factors acting as immune adjuvants of putative cancer vaccines. For these reasons, the significant correlation between clinical response to high doses of IFN-α and autoimmune manifestations, reported by Gogas and co-authors (2006) in melanoma patients, was considered surprising by the scientific community. However, these observations, which shed new light on the possible use of IFN-α in association with cancer vaccines, are in line with many previous studies that reported the correlations between the clinical response to IFN and the stimulation of both humoral and cellular responses against defined tumor-associated antigens (reviewed by Ferrantini and others 2007), and are consistent with the possible involvement of IFN-DC in the pathogenesis of autoimmune responses, possibly leading to new perspectives for identifying categories of cancer patients responding to the IFN-α therapy (Rizza and others 2010).

It is worth mentioning that the results of a pilot phase I-II trial in melanoma patients, reported by our group, showed that IFN-α, administered as an adjuvant for Melan-A/MART-1:26-35(27L) and gp100:209-217(210M) peptides, induced a surprisingly high level of blood CD8⁺ T cells recognizing modified and native peptides (Di Pucchio and others 2006). These data strongly support the use of IFN-α as an effective adjuvant for cancer vaccines in humans. Notably, a similar approach tested by the Kirkwood group in a large clinical trial based on the use of high-dose IFN-α (or GM-CSF) administered in combination with multiepitope melanoma vaccines could only show a trend toward an augmentation of the immunologic response to the vaccine by IFN-α (43.2% versus 25% response rate with or without IFN, respectively), which proved to be not statistically significant (Kirkwood and others 2009). In this regard, however, the different IFN regimens used in the 2 studies should be considered. Overall, these results strongly underline the critical importance exerted by the dosage and timing of IFN-α administration with regard to the vaccine and its possible impact on both clinical and immunological outcomes.

At present, there are still uncertainties on whether IFN-α can indeed represent a valuable vaccine adjuvant in humans and, in particular, on which are eventually the best IFN subtypes and the optimal modalities of cytokine treatment. Notably, some of the clinical studies based on the use of IFN-α as an adjuvant (Di Pucchio and others 2006; Rizza and others 2008) have included the characterization of the molecular patterns of gene expression in cells from either health subjects or cancer patients treated with IFN (Aricò and others 2011a) in an attempt to clarify the molecular mechanisms of action of the cytokine in vivo. The results of some studies performed by our group will be summarized in the next section and discussed in the context of a general overview of the main microarray data generated by the analysis of cells/tissues from patients treated with IFN or endogenously exposed to the cytokine.

Studies on the IFN-α Signature and Their Potential Importance for Understanding the Immune-Related Mechanisms Involved in the IFN Response in Patients

In an attempt to understand in greater detail the mechanisms of action of IFN-α in vivo, in both clinical and physiopathological settings, several studies have recently exploited the microarray technologies that detect and possibly analyze the IFN-α-specific signature in tissue samples isolated from individuals exposed to IFN-α (Table 1). In general, a microarray analysis of gene expression allows the description and discrimination of disease states, and provides the opportunity for both diagnostic and prognostic marker discovery. In the specific case of IFN-α, defining with acceptable accuracy the pool of genes considered the signature of IFN-α in vivo has the double advantage of providing new insights into the in vivo molecular mechanisms of IFN action, and also of allowing a possible definition of the transcription patterns induced by the cytokine correlated with the treatment efficacy and predictions of the IFN-α therapy outcome.

The significant enrichment of the expression of IFN-regulated genes, together with granulopoiesis-specific genes, has been observed in blood mononuclear cells isolated from active SLE patients, suggesting that the overexpression of a specific set of genes can represent the hallmark of in vivo cell exposure to IFN-α, which is commonly detected in the sera of these patients (Bennet and others 2003). Moreover, the observation that IFN signature is extinguished by high-dose intravenous steroids treatment further confirms the role of this cytokine in SLE, and indicates the IFN-α as a specific target for therapeutic intervention (Bennet and others 2003).

As expected on the basis of its antiviral activity, IFN-α signatures have been identified as prominent aspects of many transcriptional profiles involved in the response to pathogens in primates (Rubins and others 2004). More recently, the involvement of IFN-α in the delicate balance between immunity and autoimmunity, initially suggested by the autoimmune-like phenomena reported in melanoma patients responding to IFN-α therapy (Gogas and others 2006), has been corroborated by the finding of a clear IFN-α signature in patients experiencing several autoimmune disorders, including psoriasis, multiple sclerosis, rheumatoid arthritis, dermatomyositis, primary biliary cirrhosis, and insulin-dependent diabetes mellitus (Baechler and others 2006).

The availability of a well-characterized list of genes, known to be the signature of IFN-α, opens the possibility to finely dissect the involvement of this cytokine in the mechanisms of action of some therapeutic interventions, whose effects are assumed to be mediated by IFN-α. For example, the characterization of the early transcriptional events induced by imiquimod administration into basal cell carcinoma (BCC) lesions, performed by microarray gene profiling analysis on tumor lesions taken at early times after the treatment, identified a typical IFN-α signature, which was consistent with the acknowledged role of the cytokine in the tumor rejection phenomenon induced by this compound (Panelli and others 2007). In addition, the global transcriptional microarray approach used in this study allowed the identification of a set of imiquimod-induced up-regulated genes different from the known interferon stimulated genes (ISG), such as granzymes, perforin, NKG-5, more closely related to activation cytotoxic and other potent proinflammatory innate immune effector functions, and more tightly associated with rejection, thus suggesting a consistent but not exclusive role of IFN-α in the imiquimod-mediated BCC rejection.

In the clinical setting of subjects affected by chronic hepatitis C, in which IFN-α is commonly administered both as an antiviral and an immunomodulatory drug in combination with ribavirin (Aghemo and others 2010), Thomas and co-authors (2011) applied microarray technology to cell-culture models to demonstrate that ribavirin itself can induce a distinct set of ISG, including IFN regulatory factors 7 and 9. These data represent a significant element of knowledge that may explain the efficacy of ribavirin and IFN-α combination therapy with the synergistic instead of additive action of the 2 drugs.

The possible contribution of IFN-α to the immunomodulatory effects of dacarbazin (DTIC) in melanoma patients undergoing immunotherapy has been suggested by our group in a study reporting that several of the immunomodulatory genes found to be up-regulated in the peripheral blood mononuclear cells (PBMC) of patients 24 h after DTIC administration were well known ISG (Nisticò and others 2009). Although the role of IFN-α in DTIC-induced stimulation needs to be clarified and confirmed in bigger patient cohorts, the presence of the IFN-α signature in patients undergoing chemotherapy is in line with the evidence obtained by our group in a mouse model, showing that cyclophosphamide can induce IFN-I production and that the marked therapeutic efficacy of this drug when combined with an adoptive cell immunotherapy is mediated by IFN-I (Proietti and others 1998).

A recent study by Berry and others (2010) provided the first comprehensive and systematic description of the human blood transcriptional signature of tuberculosis (TB) by microarray gene profiling analysis. The reported increase in type I and II IFN-inducible transcripts in the blood of patients with active TB, correlating with disease severity, clearly demonstrated a hitherto underappreciated role of IFN-I signaling in the pathogenesis of TB, with important implications for diagnosis, vaccine, and therapeutic tools future development.

The ensemble of studies characterized with regard to IFN-α signature in different in vivo and in vitro settings significantly contributed toward an increment of the list of genes known to be induced by IFN-α (ISG). First categorized by de Veer and co-authors (2001), on the basis of the microarray platforms and studies available in 2001, the complex network of ISG (comprising also genes induced by IFN-γ) is now organized into a searchable, much more complete, and comprehensive database, named Interferome (Samarajiwa and others 2009). In some cases, novel and unexpected ISG were added to the list of possible in vivo mediators of IFN-α immunomodulatory and/or antitumor activity, as in the case of the study by Parlato and colleagues (2010), who analyzed the specific signature of IFN-conditioned DC to better understand the molecular mechanisms underlying the potent activity exhibited by these cells in vitro and in vivo. Stemming from the results of the global gene profiling analysis by microarray on IFN-DC, as compared with classical IL-4-DC, these authors discovered the marked overexpression of the scavenger receptor oxidized low-density lipoprotein receptor 1 (LOX-1) in IFN-α-conditioned DC, which was associated with increased levels of genes belonging to immune response families and high competence in inducing T-cell immunity against antigens derived from allogeneic apoptotic lymphocytes. LOX-1 receptor was shown to mediate most of the peculiar features of IFN-DC in presenting antigens associated with apoptotic cells, thus unraveling a novel, natural IFN-α-mediated pathway in DC that can explain the effective apoptotic cell uptake and subsequent induction of T-cell immunity by these cells.

Besides the application of gene profiling analysis by microarray to unravel the mechanisms of action of IFN in vivo, another important area of investigation is the search for patterns of gene expression profiles significantly associated with IFN-α treatment efficacy, in view of their potential use as candidate biomarker predictors of response to therapy. This aim is of particular interest in consideration of the side effects and/or the social costs of IFN-α treatment, as it is considered greatly beneficial to count on reliable tools that make appropriate therapeutic decisions and avoid unnecessary unethical treatment in individuals unlikely to respond. In spite of many efforts, the literature in this field suffers from a lack of consistency among the results obtained from patients with different diseases and receiving different IFN-α preparations. The majority of these studies have been performed in patients chronically infected with Hepatitis C virus (HCV), while attempting to identify one or more consensus blood biomarkers predictive of IFN-α/ribavirin efficacy (Bennet and others 2003; Ji and others 2003; Tateno and others 2007; Taylor and others 2007, 2008; Younossi and others 2009; Berry and others 2010; Asselah and others 2008). The more likely explanation for IFN-α resistance reported by these studies, reporting gene profiling analysis performed in both liver biopsies and PBMC, is the up-regulation of ISG in nonresponder patients before treatment (reviewed by Asselah and others 2010). Moreover, the expression of some ISG induced 24 h after the treatment has been correlated with the probability to achieve a sustained virological response (Younossi and others 2009). Since many of the genes found to be differentially expressed in nonresponders versus responders encode cytokines possibly secreted in the serum, it is tempting to expect that the amount of data generated by these gene profiling studies will evolve in the development of easily accessible serum marker predictors of response to HCV treatment.

Despite the accumulating information on the IFN-α-induced genes and of their possible in vivo role in HCV infection and other clinical settings, little information is currently available on the transient and long-term effect of low doses of IFN-α used with modalities typical of a vaccine adjuvant. In a study recently reported by us, we applied microarray technology to profile the gene expression in human PBMC from individuals treated in vivo with IFN-α, administered at a low dose as a vaccine adjuvant in the context of two separate clinical trials, performed on melanoma patients and healthy subjects, following a similar treatment schedule (Aricò and others 2011a). Notably, a clear-cut signature of IFN-α in vivo could be observed in human PBMC 24 h after the cytokine administration in both clinical studies and was consistently induced after each repeated administration of the cytokine. The extent of the modulations induced by the cytokine at the transcriptional level was mainly transient, and did not reach a steady-state level refractory to further stimulations. Moreover, the transcriptional modulations observed turned out to be fairly homogeneous among the different subjects analyzed, and no major differences between groups of subjects receiving 2 different doses (1 or 3 millions) of the cytokine could be observed.

The results of our transcriptional profiling on PBMC provided the molecular basis supporting a predominant immunomodulatory role of IFN-α when administered as a vaccine adjuvant. In fact, the immunological pathways influenced by IFN-α in vivo seems to embrace the main steps for the generation of a specific immune response, comprising not only the classical antiviral ISG (OAS, MX), but also genes encoding the factors involved in inflammation (TLR7, NMI, CXCL10, and MYD88), recruitment of immune cells (CXCL10, C3AR1, and CX3CR1), antigen processing and presentation (PSMB9, HLA-DOA), and effector-specific immune responses (SERPING1, C2, BST2, MYD88, and TNFSF13B/BAFF). Interestingly, although a rigorous comparison among the results of different microarray studies is impaired by the bias possibly induced by different platforms and statistical approaches, the core IFN-α signature identified by us in subjects receiving the cytokine as a vaccine adjuvant was not considerably different, in terms of modulated genes and Gene Ontology categories, from the one reported by studies investigating the same issue in melanoma or HCV-infected patients treated with IFN (IFN-α2b or PEG-IFN-α2b) and ribavirin (Ji and others 2003; Tateno and others 2007; Taylor and others 2007, 2008; Zimmerer and others 2008a, 2008b), or in PBMC treated in vitro with Peg-IFN-α2b and ribavirin (Taylor and others 2004). Overall, these observations strongly suggest that a similar signature occurs both in vivo and in vitro (at least in PBMC), regardless of the dose or type of IFN-α used or even of the condition of the subjects receiving the cytokine (healthy donors and HCV-infected or cancer patients).

Of interest, the comparison of the modulations observed in human PBMC isolated after the in vivo administration of the cytokine with the changes occurring in PBMC or purified monocytes isolated from healthy donors and exposed in vitro to the cytokine revealed a significant correlation among the IFN-α-up-modulated genes in the various groups. This comparative approach allowed us to define a “core” signature of IFN-α administered as a vaccine adjuvant consistently observed in all the in vivo and in vitro settings tested in our study (Table 2). Even though further studies are needed to define the role of monocytes in IFN-α activity in vivo, the significant overlapping between IFN-α signature in monocytes and PBMC in vitro and in vivo suggests that monocytes contribute to the transcriptional modulation seen in total PBMC, in line with previous observations from our group and others on IFN-α linking innate and adaptive immunity by affecting monocytes differentiation into DC (reviewed by Rizza and others 2010). This assumption was further supported by our proteomic analysis performed on monocytes, which confirmed at the protein level the effect of IFN-α on chemoattraction and inflammation observed at the transcription level in vitro and in vivo, thus corroborating the results of the gene ontology analysis on the immunomodulatory role of IFN-α in vivo. In particular, our data showed an up-regulation of CXCL10 expression after a local low-dose IFN-α injection, previously suggested to be a marker predictive of the final treatment outcome of HCV patients treated with PEG-IFN-α2b (Lagging and others 2006) and reported to experience an increase in plasma levels in melanoma patients treated with relatively low doses of IFN-α in parallel with augmented levels of CD16⁺ monocytes (Mohty and others 2010).

Of interest, the detection of the typical IFN-α signature in PBMC was paralleled by an increase in circulating CD14⁺ and CD14⁺/CD16⁺ monocytes, both expressing high levels of costimulatory and HLA-DR molecules (Aricò and others 2011a). Notably, monocytes expressing the CD14⁺ marker were reported by Ziegler–Heitbrock to be precursors of DC in response to danger signals (Krutzik and others 2005; Ziegler-Heitbrock 2007). Of interest, the kinetics of increase of CD14⁺ and CD14⁺/CD16⁺ was identical to the kinetics of appearance of the IFN-α signature, as both phenomena were observed 24 h after the first and all the subsequent IFN-α treatments. Overall, the ensemble of our data suggest that the transient up-regulation of costimulatory and HLA-DR molecules in CD16⁺ monocytes, together with the observation of a PBMC IFN-α molecular signature, highly involving the up-regulation of immune-related cytokines/chemokines, may be considered a marker of the biologic response to local IFN-α treatment. It is tempting to speculate that in these conditions the rapid generation of active DC, similar to those naturally generated from this monocyte subset in response to infections and danger signals, may be effectively stimulated.

Concluding Remarks

The recent progress on the biology of the IFN system has revealed that IFN-I play a crucial role in linking innate and adaptive immunity and that IFN-α, the cytokines with the longest record of clinical use, can be used as immune adjuvants for the development of more effective vaccines in humans. In particular, we have begun to understand that the effects of IFN-α on the differentiation and activation of DC are important for the induction of antiviral and antitumor immunity. While an ensemble of studies on the effects of IFN on the differentiation/activation of human monoctyte-derived DC published over the last decade have provided evidence on the potential efficacy of IFN-α-conditioned DC as cellular adjuvants for the development of therapeutic vaccines, the direct proof-of-concept in patients is still missing, and the optimal modalities for exploiting the immune adjuvant activity of these cytokines in humans are still to be defined. So far, most of the data on the vaccine adjuvant activity of IFN-I in mouse models have been obtained by using either specific IFN-α subtypes or IFN-α/β preparations, and little information is available on the specific role of IFN-β. Notably, the remarkable immune adjuvant activity of IFN-α observed in mouse models has not yet been translated into similar clear-cut evidence in humans. All this emphasizes the importance of well-defined clinical studies specifically designed to exploit the potential clinical efficacy of IFN-α used as an immune adjuvants in patients. Comparative studies conducted both on experimental models and in clinical settings based on the use of different IFNs would be useful for understanding the possible importance of different IFN-I preparations and cytokine formulations as well as of various administration modalities in inducing immune protective effects. Recent microarray studies on the IFN signatures in subjects treated with IFN-α as an adjuvant can be instrumental in underscoring some of the mechanisms involved in the immune adjuvant activity of these cytokines. Figure 1 illustrates how gene profiling studies can contribute to the identification of molecular response markers in patients and to the clarification of the mechanisms relevant for the IFN-α activity when the cytokine is used either in vivo (as a vaccine adjuvant) or in vitro (to generate highly active DC for therapeutic vaccination strategies). Of note, a broad set of cytokines/chemokines and other factors/markers are promptly induced in the peripheral blood of individuals treated with IFN-α as vaccine adjuvants (Aricò and others 2011a); of interest, a similar set of molecules are induced in vitro by the addition of the cytokine to cultured monocytes, which are also similar to those typically expressed in monocyte-derived IFN-DC (Parlato and others 2010) (Table 2). All this reveals that monocytes can represent important targets for the immune adjuvant activity of IFN-α and that the rapid differentiation/activation of DC can play a role in the adjuvant response. This knowledge can be instrumental for developing well-defined molecular markers of the immune adjuvant activity of IFN and can lead to a more selective and effective clinical use of these cytokines.

OPEN IN VIEWER

FIG. 1.

Gene expression profiling for unraveling the mechanisms of action and discovering potential response signatures of IFNα adjuvant activity. IFNa, interferon alpha.