Background: Genital infection with certain strains of human papillomavirus (HPV) is associated with a high risk of malignant transformation, and HPV-associated cervical intraepithelial neoplasia (CIN) can become invasive cancer. Host factors are critical in regulating tumor growth, and cytokines that modulate immunologic control may be of particular importance. The type 1 cytokines interleukin 2 (IL-2) and interferon gamma (IFN γ) are immunostimulatory and are thus capable of limiting tumor growth. The type 2 cytokines interleukin 4 (IL-4) and interleukin 10 (IL-10) are immunoinhibitory and are thus capable of stimulating tumor growth. Purpose: We analyzed the production of cytokines by peripheral blood mononuclear cells (PBMCs) in women with CIN associated with localized or extensively spread HPV infection. Methods: Thirty women diagnosed with CIN and 10 ageand sex-matched healthy control subjects were enrolled in the study conducted at Istituto Nazionale Tumori, Milan, Italy. The following parameters were analyzed: 1) HPV infection of the cervix and other sites of the lower genital tract by colposcopic, cytologic, and histologic examinations; 2) HPV typing; 3) in vitro production of IL-2 by PBMCs in response to stimulation with soluble antigen (influenza [FLU] antigen) or to cell-associated human leukocyte antigen (HLA) alloantigen; and 4) in vitro production of the type 1 cytokines IL-2 and IFN γ and of the type 2 cytokines IL-4 and IL-10 by PBMCs in response to mitogen stimulation. Statistical significance was determined by nonparametric tests (two-sided). Results: High-grade CIN associated with HPV infection was detected in all case patients, and HPV type 16 or 18 infection was detected in cervical tissue of 21 (70%) of 30 case patients. HPV infection that had spread to other sites of the lower genital tract, thus resulting in more extensive disease, was detected in 16 (53%) of the 30 individuals with CIN, whereas HPV infection was limited to the portio in 14 (47%). IL-2 production by PBMCs in response to stimulation with soluble antigen or HLA alloantigen was reduced in the group with extensive disease compared with that in the group with localized disease or with that in healthy control subjects. In contrast, IL-4 and IL-10 production in response to mitogen stimulation was elevated in the group with extensive disease compared with that in the group with localized disease or with that in healthy control subjects. The highest production of IL-4 and IL-10 was detected in patients with HPV infection that had extended beyond the genital tract. Conclusions: CIN is characterized by different immunologic profiles, in which HPV infection is or is not confined to the portio. Production of cytokines that mainly enhance potentially protective cell-mediated immunity is defective in the women in whom extended HPV infection was observed. A pronounced shift from type 1 to type 2 cytokine production is associated with more extensive HPV infection. Implications: These data reinforce the need for detailed analyses of immune dysregulation in CIN patients. They also suggest the potential usefulness of the cytokine assays for determining prognosis or deciding whether cytokine-based therapy is indicated.
Infection with human papillomavirus (HPV) is frequently observed in sexually active individuals ( 1 ). HPV infection can be asymptomatic, can induce genital warts, or can result in carcinomas of the uterine cervix, the second most common cancer of the female reproductive tract ( 2 , 3 ). Although more than 23 different strains of genital HPV have been described, infection with only a few of these strains is associated with a high risk of malignant transformation ( 4 ). HPV is among those viruses (e.g., hepatitis B virus, Epstein-Barr virus, and human Tlymphocyte virus type 1) that are capable of inducing tumors in humans. Tumor growth is likely to be favored or permitted as a result of an immune response that is quantitatively and/or qualitatively inadequate. Qualitative analysis of the immune response in human tumors has been well clarified by the Th1/Th2 model of immune regulation that was developed in the mouse ( 5 ). This model was later extended to include humans ( 6 ) and to examine immune dysregulation induced by infection with human immunodeficiency virus (HIV). We ( 7 , 8 ) introduced a functional definition of type 1 and type 2 cytokines and defined type 1 cytokines as those that mainly induce cell-mediated immunity and type 2 cytokines as those that predominantly stimulate humoral immunity. We observed that a decline in type 1 cytokine production and an increase in type 2 cytokine production are associated with, and predictive for, progression of HIV infection to acquired immunodeficiency syndrome (AIDS) [reviewed in ( 7 , 8 )]. Type 1 and type 2 cytokines are generated by multiple cell types and can exert autocrine and paracrine regulatory effects under different physiologic and pathologic conditions, ranging from pregnancy ( 9 ) to several viral and parasitic infections to neoplastic diseases [reviewed in ( 10 )].
Type 1 cytokines include interleukin (IL) 2, IL-12, and interferon gamma (IFN γ). IL-12, produced by monocytes-macrophages, is a potent activator of cellular immunity and has been shown to have antitumor as well as antimetastatic activity against murine tumors ( 11 , 12 ). In addition, IL-12 can up-regulate (i.e., increase) IFN γ production by different cell types ( 13 , 14 ). IL-2, IL-12, and IFN γ activate cytotoxic T lymphocyte (CTL)-and natural killer (NK) cell-mediated cytolytic functions that have been proposed to provide effective antitumor defense mechanisms. Moreover, IL-2 can induce the transformation of NK cells into lymphokine-activated killer cells that have been associated with improved capacity to destroy tumor cells. In contrast, type 2 cytokines (IL-10 in particular) have been shown to be associated with enhanced tumor growth. Thus, IL-10 production was reported to be elevated in certain tumors, including B lymphocytic cancers ( 15 ) and cutaneous basal and squamous cell carcinomas ( 16 ). In addition, tumor-derived IL-10 was suggested to modulate antitumor responses by inhibiting tumor antigen presentation by professional ( 17 ) and nonprofessional ( 18 ) antigen-presenting cells (APCs).
Finally, T-helper cell function can be analyzed in depth by stimulating peripheral blood mononuclear cells (PBMCs) with antigens that activate the immune response through different T-helper (Th)-APC pathways. These Th-APC interactions can be divided into three distinct categories based on the T-cell subsets and source of APCs utilized. Thus, Th responses to soluble antigens such as influenza virus are dependent on CD4 + T cells and autologous APCs. In contrast, responses to human leukocyte antigen (HLA) alloantigens can be generated by three different pathways of help: 1) CD4 + T cells and autologous APCs (similar to the requirement for recall antigens), 2) CD4 + T cells and allogeneic APCs, and 3) CD8 + T cells and allogeneic APCs ( 19 ). The first of these pathways requires antigen processing, whereas the second and third pathways involve direct recognition of HLA antigens on the allogeneic APCs. In this study, we analyzed the immune response in women diagnosed with cervical intraepithelial neoplasia (CIN) in whom HPV infection was limited to the cervix or involved other sites of the lower genital tract. In particular, we evaluated in these patients cytokine production of PBMCs, e.g., IL-2, IFN γ, IL-4, and IL-10. In addition, to verify whether Th defects were present in this population of patients, we measured IL-2 production by PBMCs after in vitro stimulation with influenza virus and HLA alloantigens.
Subjects and Methods
Patients and Control Subjects
Thirty Italian HIV-seronegative women (median age, 34.5 years) with grade III CIN were included in the study. The study protocol was approved by the Research Ethics Committee, Istituto Nazionale Tumori, Milan, Italy. Written informed consent was obtained from all patients and control subjects before enrollment. The diagnosis of HPV infection was made by use of colposcopic, cytologic, and histologic methods; in situ biopsy was performed on all sites of the lower genital tract in which HPV infection was suspected. HPV infection was classified by use of the revised Bethesda System for reporting cervical/vaginal cytologic diagnoses ( 20 ). HPV infection was characterized before any therapy by use of the chemiluminescence molecular hybridization assay method applied to cervicovaginal swabs (VIRAPAP; Diagnostic Digene Hybrid Capture System, Silver Spring, MD). This chemiluminescent assay can discern between DNA of HPV strains imparting low risk (HPV types 6, 11, 42, 43, and 44) and DNA of HPV strains imparting high risk (HPV types 16, 18, 31, 33, 35, 45, 51, 52, and 56) of malignant transformation. Ten healthy Italian women (median age, 34 years) who had undergone routine gynecologic examinations and had tested negative for HPV infection by colposcopy, cytology, and VIRAPAP test, volunteered to have their blood drawn for immunologic tests and served as control subjects.
Cigarette smoking was reported by 55% of the patients with HPV infection limited to the cervix, by 60% of the patients with HPV infection extending beyond the cervix, and by 51% of the healthy control subjects. Oral contraceptive use was reported by 29% of the patients with HPV infection limited to the cervix, by 37% of the patients with HPV infection extending beyond the cervix, and by 40% of the healthy control subjects. All the specimens were coded, and laboratory personnel were blinded to the code throughout the experimental procedures.
IL-2 Production by PBMCs in Response to Stimulation by Influenza (FLU) and HLA Alloantigen
Whole blood was collected in vacutainer tubes containing preservative-free heparin (Becton Dickinson, Rutherford, NJ) under protocols approved by the Research Ethics Committee, Istituto Nazionale Tumori. PBMCs were separated on lymphocyte separation medium (Organon Teknika Co., Durham, NC) and were washed twice in phosphate-buffered saline, and the number of viable leukocytes was determined by trypan blue dye exclusion.
The PBMCs were resuspended at 3 × 10 6 /mL in RPMI-1640 medium (HyClone, Cramlington, U.K.) containing 100 U/mL penicillin and 2 mM glutamine. To determine the ability of PBMCs to produce IL-2 in antigen-stimulated culture, we cultured the PBMCs at 37 °C in a moist, 7% CO 2 atmosphere in 96-well, flat-bottomed tissue culture plates (Corning Costar Corp., Cambridge, MA). PBMCs were either unstimulated or stimulated with a soluble antigen (influenza A whole virus, strain A/Bangkok RX73 H3N2; Istituto Sieroterapico Milanese, Milan) at a final concentration of 1:500 or with 2 × 105 irradiated (50 Gy for 5 minutes) allogeneic PBMCs per well from unrelated donors (ALLO). The HLA alloantigen on these cells stimulates immune response and hence IL-2 production. PBMCs were cultured for 7 days in the presence of 2 µg/mL of the human anti-IL-2 receptor antibody anti-Tac to prevent IL-2 consumption by the stimulated cells. Culture supernatants were then harvested, frozen, and stored at −20 °C until they were assayed for IL-2 content. The IL-2 assays consisted of culturing 8 × 10 3 of the IL-2-dependent cell line, CTLL, per well in 96-well, flat-bottomed microtiter plates, in the presence of four twofold dilutions of unstimulated or antigen-stimulated culture supernatants, as previously described ( 21 ). Then 24 hours later, the cultures were pulsed with 1 mCi of [ 3 H]thymidine (Du Pont de Nemours & Co., Milan) and harvested after 18 hours by use of a 96-well cell harvester (TOMTEC Inc., Orange, CT). Determinations of 3H incorporation were made with an LKB Beta-plate spectrometer (Pharmacia LKB Biotechnology, Piscataway, NJ). The method used to calculate IL-2 units was described previously ( 22 ). Briefly, the units were calculated as a constant multiplied by the reciprocal supernatant dilution corresponding to half-maximal CTLL proliferation. This dilution was computed by extrapolation of the line generated by linear regression analysis of the counts per minute for CTLL proliferation as a function of the supernatant dilution, by use of off-plateau values. All values were multiplied by a factor of 100 to obtain relative units of IL-2. An IL-2 response to FLU or ALLO was considered positive when the stimulation index was 3 or more. The response was calculated as follows: proliferation of CTLL (in counts per minute) in supernatants of antigen-stimulated cultures divided by the proliferation of CTLL in supernatants of unstimulated cultures.
Production of IL-2, IFN γ, IL-4, and IL-10 by Mitogen (Phytohemagglutinin)-Stimulated PBMCs
PBMCs, resuspended at 3 × 106/mL in RPMI-1640 medium, were either unstimulated or stimulated for 2 days with phytohemagglutinin (PHA) (Life Technologies, Inc. [GIBCO BRL], Gaithersburg, MD) diluted 1:100 (final concentration) at 37 °C in a moist, 7% CO 2 atmosphere. Supernatants were harvested after 48 hours because kinetic studies previously indicated that 48 hours was the optimal time for assessing lymphokine production after PHA stimulation. Production of IL-2, IFN γ, IL-4, and IL-10 by PBMCs was evaluated with commercially available enzyme-linked immunosorbent assay kits: for IL-2, human IL-2 Intertest-2 (Genzyme, Cambridge, MA); for IFN γ, human IFN γ Intertest-g (Genzyme); for IL-4, human IL-4 Intertest-4 (Genzyme); and for IL-10, human IL-10 Intertest-10 (Genzyme). All test kits were used in accordance with the procedures suggested by the manufacturer. Cytokine production was calculated from a standard curve of the corresponding recombinant human cytokine in each case.
The statistical analysis was based on a nonparametric Jonckheere-Terpstra test for trends. All data were also analyzed by a nonparametric Kruskal-Wallis test. All P values were two-sided.
Detection of HPV Infection and Clinical Characterization of Patients
HPV infection was evaluated by colposcopic, cytologic, and histologic methods in 30 patients. HPV infection with virus strains that impart a high risk of malignant transformation was detected by VIRAPAP in 21 (70%) of 30 patients ( Table 1 ). HPV infection was confined to the cervix in 14 patients, whereas it extended beyond the cervix in 16 patients. Microinvasion of the tumor was observed in three of the 16 patients in whom HPV infection was extended beyond the cervix, but it was not observed in the 14 patients in whom HPV infection was limited to the cervix. In addition, relapsing (reactivation and/or reinfection) HPV infection occurred in three patients (twice in two patients and thrice in one patient) with extended HPV infection. Such relapsing virus infections were not observed in women with HPV infection limited to the cervix. The clinical characteristics of the patients and control subjects are summarized in Table 1.
IL-2 Production by PBMCs After Antigen Stimulation
IL-2 production by antigen-stimulated PBMCs was evaluated in all patients and control subjects. PBMCs were stimulated with FLU and ALLO because these two antigens induce IL-2 production by different Th-APC pathways, thereby permitting better evaluation of the functional integrity of the immune response. FLUstimulated IL-2 production was defective (<3 U) in four (29%) of 14 women with HPV infection localized to the cervix, in nine (56%) of 16 women in whom HPV infection extended beyond the cervix, and in one (10%) of 10 healthy control subjects. ALLO-stimulated IL-2 production was defective in three (19%) of 16 women with HPV infection extended beyond the cervix. We detected a weak trend in ALLO-stimulated IL-2 production by comparing data obtained in the three different groups ( P = .013 by Jonckheere-Terpstra test). We observed a stronger trend in FLU-stimulated IL-2 production ( P = .0016); this result was mainly secondary to the defective IL-2 production observed in patients with HPV infection that extended beyond the cervix ( P = .0004 by Kruskal-Wallis test) ( Table 2 ).
Production of IL-2, IFN γ, IL-4, and IL-10 by Mitogen-Stimulated PBMCs
To evaluate mitogen-induced type 1 and type 2 cytokine production, we stimulated PBMCs of patients and of healthy control subjects with PHA. Cytokine production was measured by use of unfractionated PBMCs to approximate closely the in vivo situation and because cells other than CD4 + T lymphocytes in blood can also secrete type 1 and type 2 cytokines. Because of our experience using PHA to assess T-cell function in HIVinfected individuals, we stimulated PBMCs with PHA, which allowed us to compare these results with those obtained in previous studies [reviewed in ( 7 , 8 )].
PHA-stimulated IL-2 production ( Fig. 1, A and B ) and IFN γ production ( Fig. 1, C and D ) were reduced in patients compared with those in healthy control subjects and in patients with diffuse HPV infection compared with those both in patients with HPV infection localized to the cervix and in control subjects. This trend did not reach statistical significance for IL-2 production, whereas it was marginally significant for IFN γ production ( P = .044 by Jonckheere-Terpstra test). PHA-stimulated IL-4 production was augmented in patients with HPV extended beyond the cervix compared with that in either patients with HPV infection limited to the cervix or healthy control subjects ( P = .043) ( Fig. 2, A and B ). The same phenomenon was detected when PHAstimulated IL-10 production was measured. Differences in IL-10 production were highly significant when data obtained in the three groups were compared ( P = .00003) ( Fig. 2, C and D ).
We analyzed cytokine production profiles in a group of women with CIN associated with HPV infection. These women were divided into those in whom HPV infection was limited to the cervix and those in whom HPV infection extended to other areas of the lower genital tract. A different pattern of cytokine production was observed in the former (higher production of IL-2 after antigen stimulation and lower production of IL-4 and IL-10 after mitogen stimulation) compared with the latter patients (lower production of IL-2 after antigen stimulation and higher production of IL-4 and IL-10 after mitogen stimulation). In particular, both defective FLU-and ALLO-stimulated IL-2 production was observed in women with HPV infection that extended beyond the cervix, compared with women with HPV infection confined to the cervix or with healthy control subjects. These data imply that a more extended and aggressive HPV infection is associated with defective type 1 cytokine production and augmented type 2 cytokine production.
It has been suggested that protection against intracellular pathogens and tumors is mainly influenced by cellmediated immunity ( 4 – 8 ); it was also observed that cell-mediated immunity is preferentially activated by type 1 cytokines, whereas type 2 cytokines (and IL-10 in particular) have a suppressive action on cell-mediated immunity ( 4 – 8 ). Both the serum concentration of IL-10 and its production by PBMCs have been reported to be elevated in a variety of human neoplasias, including bronchogenic carcinoma ( 23 , 24 ), renal cell cancer ( 25 – 27 ), basal and squamous cell carcinomas ( 16 ), lymphomas ( 15 ), gliomas ( 28 ), and melanomas ( 29 , 30 ). The data from this study add HPV-associated CIN to the list of tumors characterized by secretion of high levels of IL-10. It is interesting that increased IL-10 production was supported by immunocompetent cells such as tumor-infiltrating lymphocytes in some cases ( 24 ) and directly by tumor cells in other cases ( 16 ). This generalized impairment of cytokine production might be explained by the observation that Tlymphocyte priming in the presence of high concentrations of IL-4 and IL-10 (and presumably of low concentrations of IL-12) favors the maturation of Th2 lymphocytes ( 31 , 32 ). Thus, type 2 cytokine-secreting tumor cells could induce a preferential priming of type 2 cytokinesecreting lymphocytes with the generation of a self-amplifying loop.
Results from different experiments suggest that augmented production of IL-10 may be a mechanism by which tumor cells escape recognition by the immune system (particularly as a result of suppression of T-lymphocyte function), thus favoring tumor growth. The mechanisms by which IL-10 production can facilitate tumor growth are still not well understood. One explanation for this phenomenon is that IL-10 down-modulates (i.e., suppresses) expression of major histocompatibility complex (MHC) class I expression, preventing tumor antigen presentation to CD8 + CTLs ( 17 , 18 ). Consistent with this hypothesis is the observation that preincubation of human melanoma cells with IL-10 results in a dose-dependent inhibition of CTL-mediated tumor-specific lysis secondary to the reduction of MHC class I expression on the surface of the tumor target cells ( 18 ). In addition, transfection of a mouse lymphoma cell line with the IL-10 gene dramatically reduced the expression of class II and resulted in the conversion of the lymphoma line to a CTL-resistant phenotype ( 33 ). IL-10 is also known to down-modulate MHC class II expression and CD4-dependent antigen presentation ( 34 ). It is noteworthy that, in the same group of CIN patients in whom HPV infection was extended to other areas of the lower genital tract and in whom elevated IL-10 production was present, we observed a defect in MHC class II CD4-dependent IL-2 production in response to soluble antigens.
It is interesting that defective antigenstimulated IL-2 production was shown in HIV infection to be progressive (i.e., loss of IL-2 production in response to soluble antigens precedes loss of IL-2 production in response to ALLO) ( 21 ). Moreover, it was shown to be predictive for clinically relevant AIDS-related events, such as 1) reduction in the number of CD4 + lymphocytes ( 35 ), 2) progression to AIDS (36), and 3) augmented incidence of recurrent bacterial and opportunistic infections ( 37 ).
Although it is not yet well established whether tumor prognosis is associated with low type 1 cytokine and/or high type 2 cytokine production, a report ( 38 ) suggested that low-dose melphalan treatment induces a type 2-to-type 1 cytokine shift in mice in which experimental, large MOPC-315 tumors could be eradicated. We ( 39 ) observed that defective IL-2 production after antigen stimulation tends to be associated with more advanced clinical presentation and more compromised hematologic parameters in patients with newly diagnosed Hodgkin's lymphoma ( 39 ). Thus, in two cases of different cancers, we verified that defective antigenstimulated IL-2 production is a correlate of a more extended tumor and more advanced clinical presentation. Among the patients in this preliminary study, we observed that more severe immune impairment and a more pronounced type 1-totype 2 shift in cytokine production profiles are associated with more extensive disease. These data reinforce the need for in-depth analyses of immune dysregulation in these patients and point to the potential usefulness of cytokinebased therapy in the treatment of cancer patients.